Language Contact : A Multidimensional Perspective [1 ed.] 9781443867429, 9781443844017

Citation preview

Language Contact

Language Contact: A Multidimensional Perspective

Edited by

Kelechukwu U Ihemere

Language Contact: A Multidimensional Perspective, Edited by Kelechukwu U Ihemere This book first published 2013 Cambridge Scholars Publishing 12 Back Chapman Street, Newcastle upon Tyne, NE6 2XX, UK

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

Copyright © 2013 by Kelechukwu U Ihemere and contributors All rights for this book reserved. No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. ISBN (10): 1-4438-4401-2, ISBN (13): 978-1-4438-4401-7

For Ekaterina – because your lamp burns ever brighter in our hearts

TABLE OF CONTENTS

List of Figures.............................................................................................. x List of Tables............................................................................................. xii Acknowledgements ................................................................................... xv Chapter One................................................................................................. 1 Language Contact Kelechukwu Ihemere Part I: Contact, Change and Competence Editor’s Introduction to Part One .............................................................. 24 Chapter Two .............................................................................................. 29 Cost Makes a Difference: Codeswitching Frequencies in Bilingual Language Production Carol Myers-Scotton Chapter Three ............................................................................................ 52 Analysing Moroccan Arabic and English Diminutives in Codeswitched Discourse Najat Benchiba-Savenius Chapter Four.............................................................................................. 59 The Gray Area of Acceptability Judgements: Clefts and Exhaustivity in Cypriot Greek Evelina Leivada, Elena Papadopoulou, and Natalia Pavlou Chapter Five .............................................................................................. 87 Contact Induced Language Changes in Modern Uyghur Aminem Memtimin

viii

Table of Contents

Chapter Six .............................................................................................. 102 The Role of the Input in the Early Development of a Heritage Language Francesca LaMorgia Part II: Language Shift and Language Maintenance Editor’s Introduction to Part Two............................................................ 126 Chapter Seven.......................................................................................... 129 Language Shift as an Outcome of Expansive and Integrative Language Contact in Africa Herman Batibo Chapter Eight........................................................................................... 140 Language Maintenance and Shift in Migration: A Case Study of Bhojpuris in Kolkata Aditi Ghosh Chapter Nine............................................................................................ 158 Languages in Contact in Minority Communities: The Case of Francoprovençal in Italy Carmela Perta Chapter Ten ............................................................................................. 176 Language Maintenance and Language Shift among Afro-Costa Rican Women Lisbeth Philip Chapter Eleven ........................................................................................ 208 Survival Against All Odds: Longevity of Sri Lanka Portuguese Creole Shihan de Silva Jayasuriya Part III: Language Attitudes and Identity Editor’s Introduction to Part Three.......................................................... 232 Chapter Twelve ....................................................................................... 235 Multilingualism and National Identity in Mixed-Marriage Families Marina Ronþeviü, Branka Blaževiü, and Tatjana Šepiü

Language Contact: A Multidimensional Perspective

ix

Chapter Thirteen ...................................................................................... 248 Language Attrition and Loss of Identity: Azeri-Farsi Codeswitching in Tabriz Iran Thomas Hawes and Seyed Hadi Mirvahedi Chapter Fourteen ..................................................................................... 266 Who is Afraid of Anglo-Nigerian Pidgin? Voices from Urban Nigeria Charles Mann Contributors ............................................................................................. 285 Index ........................................................................................................ 292

LIST OF FIGURES

Chapter Four Figure 4-1: Overall acceptance in all stories Figure 4-2: Overall acceptance in all stories (Alternative presentation) Figure 4-3: Embu 1 Figure 4-4: Embu 2 Figure 4-5: Cleft 1 Figure 4-6: Cleft 2 Figure 4-7: Story 1 Figure 4-8: Story 6 Figure 4-9: Story 3 Figure 4-10: Story 5 Figure 4-11: Story 2 Figure 4-12: Story 4 Figure 4-13: Change in progress Chapter Six Figure 6-1: Average number of days spent in Italy and Ireland every Year Figure 6-2: Average percentage of weekly exposure to each language in Ireland Figure 6-3: Average percentage of exposure to each language in one Year Chapter Eight Figure 8-1: Language preference in the home domain Figure 8-2: Language use in the formal domain Figure 8-3: Language use in other domains Figure 8-4: Preference for Bhojpuri and Hindi across generations Figure 8-5: Comparative language preference for Bhojpuri among male and female members in the family

Language Contact: A Multidimensional Perspective

Chapter Nine Figure 9-1: Faeto and Celle San Vito Figure 9-2: Dimension 1, positive variables Figure 9-3: Dimension 1, negative variables Figure 9-4: Dimension 2, positive and negative variables Figure 9-5: Dimensions and ‘age’ Figure 9-6: Dimensions and subjects Figure 9-7: Francoprovençal vs. Italian Figure 9-8: Francoprovençal vs. Dialect Figure 9-9: Dimensions and ‘age’ Figure 9-10: Dimensions and speakers Figure 9-11: Representation of the repertoire in Celle Figure 9-12: Representation of the repertoire in Faeto Chapter Ten Figure 10-1: Place and language of interview Chapter Fourteen Figure 14-1: Map of Nigeria

xi

LIST OF TABLES

Chapter Three Table 3-1: Moroccan Arabic diminutives Table 3-1: Further Moroccan Arabic diminutives Table 3-1: Moroccan Arabic feminine diminutives Table 3-1: Further Moroccan Arabic French origin diminutives Table 3-1: Further Moroccan Arabic diminutives (second root consonantal repetition) Table 3-1: Modified Moroccan Arabic diminutives Chapter Four Table 4-1: Participants Table 4-2: Conditions Table 4-3: Randomization Chapter Five Table 5-1: The fate of Arabic b and f at the end of stems Table 5-2: The distribution of y and ž Table 5-3: The distribution of Ћ Table 5-4: Numbers in Chinese and Uyghur Chapter Six Table 6-1: Mothers’ output Table 6-2: Mothers’ vocabulary Table 6-3: Mothers’ MLU Table 6-4: Summary of results Table 6-5: Results at MLU 1.5-2 Table 6-6: Results at MLU 2.5-3 Chapter Seven Table 7-1: The position of language endangerment in Africa Table 7-2: The quadriglossic structure of languages in Africa Table 7-3: Attitudes towards the use of mother tongue by the various ethnic groups

Language Contact: A Multidimensional Perspective

xiii

Chapter Eight Table 8-1: Most preferred language for communication Chapter Nine Table 9-1: Repertoire in Celle Table 9-2: Repertoire in Faeto Chapter Ten Table 10-1: Language(s) participants speak Table 10-2: Participants’ attitudes towards Spanish, Limón English and its varieties Table 10-3: Ethnicity of participants Table 10-4: Social group ethnic composition Table 10-5: Do you consider the majority of your contacts to be with people outside your ethnic group? Table 10-6: Language attitudes associated with LMLS Table 10-7: Ethnicity and inter/intra ethnic relations associated with LMLS Chapter Twelve Table 12-1: Tendency of the overall sample as to the self-evaluation of one's own identity with respect to various factors Table 12-2: Knowledge of the minority language Table 12-3: Emotions evoked by using the minority language Table 12-4: Knowledge of the minority language with respect to the gender of the respondents Table 12-5: Evaluation of the importance of the minority language and the official language with respect to the gender of the respondents Table 12-6: Frequency of using the minority language with respect to the gender Table 12-7: Frequency of using the minority language with respect to the gender – bilingual respondents Table 12-8: Influence of the language of one of the parents, where one of the two belongs to a minority language community Table 12-9: Influence of the minority language on the formation of clear identity and/or shared identity Table 12-10: Arithmetic middle and statistically significant differences in the self-evaluation of personal importance of the minority language and the official language in respondents with clear or shared identity

xiv

List of Tables

Chapter Thirteen Table 13-1a: Azeri Table 13-1b: Farsi Table 13-2a: Azeri Table 13-2b: Farsi Chapter Fourteen Table 14-1: Sample profiles and ANP contact details: Overall national data Table 14-2: Attitudinal dispositions towards ANP: Overall national data Table 14-3: Macro categories of attitudinal dispositions towards ANP: Overall national data

ACKNOWLEDGEMENTS

I would like to gratefully acknowledge the support and assistance of Sharon Sinclair for her professionalism in ensuring that everything was in place to make this project a success. My thanks go to all the contributors to this volume for their generosity and patience. I would like to express my gratitude to Martin Ehala, Afaf Nash, Steve Nicolle and Tope Omoniyi who read the manuscript and provided me with challenging and supportive commentary on content and structure. To all the participants at the 2011 second Westminster University Linguistics conference thank you for your input. Thank you to Carol Koulikourdi, Amanda Millar, and Emily Surrey at Cambridge Scholars Publishing for doing a great job. To Ira, Chris and Tasha thank you for putting up with me - Ȼɨɥɶɲɨɟ ɫɩɚɫɢɛɨ.

CHAPTER ONE LANGUAGE CONTACT KELECHUKWU IHEMERE

1. Introduction Since the inception of modern contact linguistics through the works of Weinreich (1953) and Haugen (1953), numerous investigators have studied the manifestations of language contact across different disciplines, naturally adopting varied perspectives and approaches relevant to their particular field of inquiry. Thus, for example, historical linguists have studied language contact phenomena from the point of view of its use as an explanation for certain changes in language (Andersen 1995; Blake and Burridge 2003; Campbell 2004; Lehmann 1992; Thomason and Kaufman 1988; Trask 1996). For psycholinguists it is the explication of the acquisition of proficiency in another language (Ellis 1985; Gleason and Ratner 1998; Kess 1976; Klein 1986; Steinberg, Nagata, and Aline 2001). Others have investigated the effects of contact on mental processes, while sociologists have treated it as an element in culture conflict and have looked at some of the consequences of linguistic heterogeneity as a societal phenomenon (Fishman 1964, 1965, 1968; Greenfield 1972; Parasher 1980; Rumbaut and Portes 2001). In the field of education, language contact has been studied in connection with public policy and to address questions relating to the alleged link between bilingualism and intelligence (Cummins 2003; Lippi-Green 1997; Mufwene, Rickford, Bailey, and Baugh 1998), whether certain types of bilingualism are good or bad, and the circumstances under which they arise are the foci of interest. In spite of the many approaches and interests, quite simply when speakers of different languages interact closely it is typical for their languages to influence each other. The influence could be as common as the exchange of words or what is termed vocabulary borrowing in the literature. It can also go deeper, extending to the exchange of even basic characteristics of a language such as morphology and grammar. In some

2

Chapter One

cases, the result of the contact of two languages can be the replacement of one by the other. This is most common in asymmetric relationship between languages, and sometimes lead to language shift and death (see Ihemere 2007, 2011).

2. Possible outcomes of language contact Here, a brief discussion of some of the outcomes of language contact is given (for more detailed discussions, see the subsequent chapters in this volume and the listed references). When speakers of different languages interact closely, a number of things are likely to happen as a result of the contact:

2. 1 Borrowing One group will actually take into its language some words from the other group’s language to refer to objects, activities, or concepts that the other group has, largely for those things that are new to the first group – lexical borrowings. This exchange is mostly asymmetrical, one group tends to do more taking and the group that takes the most is the one with less prestige in some vital public area, such as socio-economic status or political control. For example, when the Norman French conquered England in 1066, they not only had the political power, but their mode of living was considered more civilised, more sophisticated (i.e. it had cachet). (Myers-Scotton 2006, p. 209, 211) Today, as Myers-Scotton (ibid., p. 212) correctly observes, English is the laeding source of borrowings into many languages, other languages are also currently important as donor languages as well, if on a more limited level. Borrowed words are almost always adapted to the recipient language in morphology. For instance, according to Myers-Scotton (ibid., p. 224), when a Spanish speaker, speaking Spanish, uses an English borrowed word, such as weekend, it receives a Spanish determiner (realised as a singular with masculine gender, the default gender for Spanish) and it is therefore el weekend. And when a Swahili speaker borrows a noun from English, it receives the noun class agreement prefixes of one of the Swahili noun classes. Thus, the borrowed noun sweta ‘sweater’ receives the demonstrative form (hii ‘this’) that integrates it into the noun class that is called class nine by Swahili grammarians. Myers-Scotton explains that class nine is the default class for singular items that are borrowed into Swahili. Sweta is phonologically integrated into Swahili by ending in a vowel (Swahili has a usual CVCV pattern, with words ending in a vowel):

Language Contact

3

1. Swahili-English (Myers-Scotton, unpublished Nairobi corpus, 1988) Hii sweta Mummy ndiyo alishona. ‘This sweater, Mummy indeed sewed it.’

In example (2), from Türker (2000, p. 174 cited in Myers-Scotton 2006, pp. 224-225), a borrowed word, Norwegian matpakke ‘lunch bag’ is integrated into a Turkish frame. Note the Turkish particle de that modifies matpakke ‘lunch bag/box’ and the Turkish verb for ‘spread’ sür. 2. Turkish-Norwegian (Türker 2000, p. 174) Bi de matpakke sür de-di one particle lunch bag/box spread say-PAST3SING ‘He also said [to me], “prepare a lunch bag/box”.’

Myers-Scotton (ibid., p. 224), observes that the speaker has borrowed not only the word for lunch bag/box, but the entire Norwegian expression as a loan translation in which the verb – with a Norwegian meaning – has been realised in Turkish (Norwegian smøre matpakke ‘prepare a lunch box/bag’ (which literally means ‘butter [your] lunch bag’). In Standard Turkish the verb for ‘prepare’ would be hazirla-. (Myers-Scotton 2006, p. 224)

2. 2 Codeswitching Perhaps, the most studied and discussed outcome of language contact phenomena is codeswitching - the ability to use more than one language within a single utterance. Much earlier, Gumperz (1982, pp. 75-84) put forward a number of discourse functions that codeswitching is seen to realise such as: quotations; addressee specification; interjections; reiteration; message qualification and personalisation versus objectivisation. An important part of Gumperz’s approach relies on the symbolic distinction between we vs. they embodied in the choice of codes. In other words, the tendency is for the minority language to be regarded as the ‘we’ and the majority language as the ‘they’ code. The ‘we’ is the in-group, informal, personalised activities, while the ‘they’ code typically signifies out-group, more formal relations. Thus, in example (3) below (Ikwerre-Nigerian Pidgin English (NPE) codeswitching, Ihemere 2007), Ikwerre serves to mark the in-group talk and NPE the out-group:

4

Chapter One 3. Ikwerre-NPE codeswitching (Ihemere 2007, p. 87) A: ‘Anyi nde na asu asusu anyi (.) ma ha na e yuzu Pidgin for skul.’ ‘We’ [their parents] speak our language (the Ikwerre language), but ‘they’ [the younger generation] speak NPE at school.’

The switch from Ikwerre to NPE in the above example stresses the boundaries between ‘them’ (the younger generation) and ‘us’ (the parent generation). This generalisation is, however, very problematic and not as clear-cut as Gumperz suggests. In fact, either Ikwerre or NPE can function as we-code, depending on the occasion and the age group concerned. This notwithstanding, stemming from Gumperz’s (1982) work, the general view is that codeswitching acts as a contextualisation cue (Auer 1995, 1998). Codeswitching “contextualises” by highlighting in a certain context against which inferences are drawn. For Chan (2003, 2004) the communicative role of codeswitching may go even further beyond contextualisation cues in Gumperz’s sense. He observes that codeswitching is quintessentially a textualisation cue, which “frames” elements in a discourse that are to be interpreted in some way different from the preceding text. The implication being that the act of switching rather than the switched code is the most essential cue. Under this analysis, contextualisation is only one of the pragmatic functions of codeswitching. In making this argument Chan proposes two taxonomies, namely, the three-fold motivations of codeswitching (i.e. social, pragmatic, and processing) and three levels of pragmatic meanings that codeswitching may convey (i.e. ideational, interpersonal, and textual). On the other hand, Myers-Scotton (1983, 1993) observed that in many of the world’s bilingual communities, fluent bilinguals sometimes engage in codeswitching by producing discourses which, in the same conversational turn or in consecutive turns, include morphemes from two or more of the varieties in their linguistic repertoire. In various publications, MyersScotton developed the theme of codeswitching as a tool for the speaker and an index for the addressee of the negotiation of interpersonal relationships, with participants cast within a “rational actor” framework, weighing costs and rewards of choices made against a backdrop of awareness for all interaction types of “unmarked” vs. “marked” choices (see also Myers-Scotton this volume). Her argument was based on what she termed the Markedness Model (MM). The MM is more centred on the notion that speakers make choices because of their own goals. Of course, they cannot ignore some consideration for listeners. After all, without listeners, there is no conversation (Myers-Scotton 2006, p. 158). Further, the MM is an attempt to establish a principled procedure that both

Language Contact

5

speakers and listeners use to judge any linguistic choice that they might make or hear as more or less marked, given the interaction in which it occurs. Thus, unmarked choices are those that are more or less expected, given the ingredients of the interaction (participants, topic, setting, and so on.). For example, Myers-Scotton (1993) refers to a Rights and Obligations set (RO) as part of the normative expectations for each interaction type. These expectations illustrate an unmarked way to behave. Concerning language, the unmarked choice is the linguistic reflection of any specific RO set, but only in a specific interaction type (Myers-Scotton 2006, p. 159). Therefore, in example 3 above, for Ikwerre-NPE bilinguals the unmarked choice to use to elderly relatives at family gatherings is Ikwerre. This linguistic choice is indexical of the RO set. Accordingly, in the words of Myers-Scotton (ibid.), when a speaker makes the unmarked choice, he/she is causing no social ripples because participants expect such a choice, based on experience. The MM presupposes that as part of their general cognitive architecture all speakers have a markedness evaluator. This abstract component underlies the capacity to conceptualise markedness. Specifically, as a sociolinguistic construct, markedness refers to the capacity to develop the following three abilities: i. Most important is the perception that relevant linguistic choices for a specific interaction type fall along a multidimensional continuum from more socially unmarked to more marked. ii. Additionally, speakers learn to recognise that the markedness ordering of choices is dynamic. It depends on the specific interaction type, as well as on how the individual interaction develops. iii. Lastly, speakers develop the ability to provide relevant interpretations for all choices, marked as well as unmarked, given the interaction type. What the markedness evaluator offers is not a set of rules, but rather a process for evaluating potential choices. The interpretations that speakers attach to linguistic choices have to do with the speaker’s projection of his/her own persona and relations with other participants. Hence, any choice a speaker makes is perceived as indexing a desired (RO) set between participants. All participants interpret a choice against the backdrop of those choices that index the more unmarked RO sets for a specific interaction type. As a corollary, this means that they also recognise some choices as indexing more marked RO sets (see MyersScotton and Bolonyai 2001, pp. 8-11).

6

Chapter One

Finally, the main thrust of the MM is a reliance on the notion of intentionality in human actions (Levinson 1995). That is, actors intend their actions to reflect goals or attitudes, and observers attribute intentions to actions. They also give at least a nod toward the notion that innate architectures coordinate readings of cost-benefit analyses of competing choices. In interpersonal contexts, such architectures can also be seen to coordinate readings of intentionality. It is equally important to note that codeswitching is not only studied from a sociolinguistic perspective. In fact, a host of other investigators including Carol Myers-Scotton (this volume) have studied the syntax of codeswitching. A number of researchers (Poplack 1980; Poplack, Wheeler, and Westwood 1987; Poplack and Meechan 1995; Sankoff, Poplack, and Vanniarajan 1991) are interested with formulating constraints on points in a sentence where codeswitching can take place on the grounds of surfacelevel linear differences between the languages concerned. These researchers view restrictions on codeswitching along the lines of dissimilarities in word order, either across clauses (inter-sentential) or on phrases within clauses. For instance, Poplack’s (1980) Equivalence Constraint is based on this premise that switching is not permitted when the syntax of two languages does not match at a potential switch point. This thinking still pervades much of current research in the field (Adalar and Tagliamonte 1998; Budzhak-Jones 1998; Eze 1998; Samar and Meechan 1998; Turpin 1998). Muysken (2000, p. 118) says that “the evidence overwhelmingly supports more surface-oriented constraints”. However, others (Bhatt 1997a, b, 1999, 2001a, b; Kupisch 2003; Möhring and Meisel 2003) have put forward many counter examples to such models. It is very possible to observe that when the surface word order of two languages does not match, as in the examples below from Igbo-English bilingual discourse, switching is still possible. 4. Ӑ-zөrө efè na 3S-PAST-buy shirt-NOUN at-PREP ‘He bought [the] shirt at [the] shop

shop shop

5. Ӑ ma-ghi ajө-I ajө-jө mgbe ӑ ga èbido ime 3S will-NEG ask-VERB-you question-NOUN when-ADV 3S start do interview interview ‘He will not ask you [any] questions when he will start the interview’

Language Contact

7

In these examples Igbo is the Matrix Language. This is evident from the fact that every element in the clause follows Igbo word order rather than English, indicating that only one language supplies morpheme order (i.e. Igbo). Some researchers within this group interested in devising constraints on codeswitching on the basis of surface-level linear differences dismiss such counter-examples to the Equivalence Constraint by asserting that any “out of order” Embedded Language words are borrowings, not codeswitching. For instance, Shin (2002) who worked on Korean-English codeswitching took this viewpoint to account for a number of English words in what would be accusative (objective) case position in Korean. According to Myers-Scotton (2006, p. 251) Korean is a language with the verb in final position and a preceding direct object (that sometimes receives an overt accusative case marker and sometimes does not). Shin (2002) reports that the percentages of English- and Korean-origin nouns that occur with an overt accusative case marking suffix or with no suffix are similar. Because of this, she argues that the English nouns are borrowings into Korean and so their “wrong order” in Korean-English codeswitching does not count. Additionally, there is another group of researchers who disregard singly occurring Embedded Language words as codeswitches. They look for explanations at a more abstract level than linear structure. These researchers (Belazi, Rubin, and Toribio 1994; DiSciullo, Muysken, and Singh 1986; Halmari 1997) do this by structuring their explanations along the lines of what are considered generative theories of syntax. That is, they assert that the grammatical organisation of codeswitching can be accounted for in terms of the principles of current syntactic theories, even though these theories were initially formulated to explain monolingual data. They do not recognise any theoretical (or useful) value in recognising the asymmetry between a Matrix Language and an Embedded Language. Thus, with Myers-Scotton (2006, p. 251), I argue that in attempting to rework syntactic theories intended to explain monolingual utterance structure, these researchers are looking for explanations at an abstract level of grammatical structure. Further, given that these researchers do not consider single content words as part of codeswitching, they can view them as borrowings, although they generally receive inflections from the language the Matrix Language Frame (MLF) model categorises as the Matrix Language. These single content words can justifiably be viewed as borrowings since established borrowings also generally accept inflections from the recipient language corresponding to the Matrix Language. However, the main drawback of this view is that these researchers provide

8

Chapter One

no theory to account for how the production and use of these forms vary from that of established loan words that are in the lexicon of the Matrix Language. Moreover, these researchers attempt to explain code-switching within the terms of the syntactic theory of Binding (Belazi, Rubin, and Toribio 1994) and Government (DiSciullo, Muysken, and Singh 1986). Chomsky’s (1995) Minimalist Programme has been adopted by MacSwan (1999, 2000), Toribio and Rubin (1996) and Bhatia and Ritchie (1996). Increasingly, serious counter-examples have been suggested to Government constraints on codeswitching (for example, Bhatt 1997a, b, 1999, 2001a, b; Bhatt and Hancin-Bhatt 2002; Ezeizabarrena 2003; Hinzelin 2003; Kupisch 2003; Möhring and Meisel 2003). The view taken in this paper is similar to that in Jake, Myers-Scotton, and Gross (2002) that the Minimalist Programme on its own without the Matrix Language versus Embedded Language distinction is not adequate to explain what happens in codeswitching. What is more, there are other investigators who differ from the earlier groups because they usually take into account all Embedded Language material including singly occurring forms and full phrases and sometimes full clauses or sentences in their analyses. For example, Clyne (2003) considers different levels of structure and is more interested in what in the words of Myers-Scotton (2006, p. 242) is termed composite codeswitching, but which he includes under the label transference. Hence, these approaches differ from the MLF/4-Morpheme (4-M) model, which considers codeswitching as contact between languages only within the clause. The MLF/4-M (Myers-Scotton 1993, 2002, 2006) model makes the case for a distinction between the Matrix Language and the Embedded Language. The Matrix Language plays a dominant role in providing more morphemes in a given codeswitching discourse and in shaping the overall morpho-syntactic properties of codeswitched utterances. In other words, the model posits two hierarchies in reference to mixed constituents: both languages do not participate equally; only one language is the source of the abstract morpho-syntactic frame. This language (and the frame) is called the Matrix Language and the other language is called the Embedded Language. This idea is formalised as the Morpheme Order principle and the System Morpheme principle of the MLF model. These are testable hypotheses referring to the existence of asymmetry between the languages implicated in codeswitching. On the basis of these principles only one language (the source of the frame) supplies both morpheme order and frame-building system morphemes to the frame. Such morphemes are

Language Contact

9

referred to as outsider late system morphemes under the 4-M model (Myers-Scotton 1993, p. 83; 2002, p. 59).

2. 3 Convergence and Attrition According to Myers-Scotton (2006), convergence and attrition occur when one language becomes more like another. Convergence is speech by bilinguals that has all the surface-level forms from one language, but with part of the abstract lexical structures that underlies the surface-level patterns coming from another language (or languages). Attrition is generally thought of as language change within the speech of one individual. Both tend to happen to the L1 of bilinguals when they live in a community where the “invading” or other language is socially and politically dominant. Codeswitching often precedes (or combines with) convergence or attrition, but either can occur without extensive codeswitching, too (p. 271; see Hawes and Mirvahedi; and La Morgia this volume). An example of syntactic convergence is found in (6a). The extract is from Dimitrijeviü-Saviü’s (2008, p. 58) study of a variety of Serbian spoken in a migrant community in Melbourne, Australia. 6a. AS (Australian Serbian) a svako and everyoneNOM

poþne da beginPRES.3.SG that

smeje mi laughPRES.3.SG IDAT

b. HS (Homeland Serbian (spoken in Serbia)) a svi poþnu da mi beginPRES.3.pL that IDAT and everyoneNOM.PL ‘and everyone begins to laugh at me’

se smeju EFL laughPRES.3.SG

Dimitrijeviü-Saviü observes that (6a) exemplifies similarity between syntactic structures in one language (Serbian) in contact with another (English) increasing at the expense of difference. Convergence has variously been defined as the increasing agreement of languages in regards to features of their overall structure (Hock 1986, pp. 492–512), or as increasing similarity between two or more languages in a particular area of grammar (Silva-Corvalán 1994, pp. 4–5). McMahon (1994) has argued for three dimensions of difference between convergence and borrowing: (i) convergence requires long-term bilingualism; borrowing can occur with only limited bilingualism; (ii) convergence tends to affect syntax and morphology; borrowing tends to be limited to the lexicon; (iii)

10

Chapter One

convergence is a mutual process, i.e., it affects both languages; borrowing is unilateral.

2. 4 Pidgins and Creoles Pidgins and Creoles are languages that develop out of contact between the speakers of different languages. There are in most cases quite marked social conditions associated with that contact. For example, Meyerhoff (2011, p. 259), the speakers may only be in contact in a reduced set of social interactions, such as trading or work. Because of the limited social contact between the speakers, they seldom have extensive access to native speaker models of each other’s languages. Traditionally, linguists distinguish pidgins and creoles on the grounds of how they are learnt, with a pidgin being defined as a contact language that is nobody’s first language, and a creole as a contact language that does have native speakers. Generally, there is one language that provides most of the vocabulary in pidgin/creole – this is known as the lexifier because it provides the building blocks of the lexicon (vocabulary). Although other languages may not dominate the surface structure of the pidgin/creole so much, they may still have profound and subtle effects on the way the words are used and how the sentences are structured – that is, on the semantics and syntax. These languages are known as the substrate, and their effects are displayed in underlying structure (ibid, p. 262; see further, Mann this volume). Fromkin et al. (2007, pp. 435), observe that one of the characteristics of pidgins is that speakers from different linguistic backgrounds may have different sets of rules, giving the language a more haphazard feel than one gets from a fully developed language. They cite the example of a Japanese (and SOV language) speaker of an English-based pidgin who may put the verb last (in accordance with Japanese word order), as in The poor people all potato eat. Whereas a Filipino speaker of Tagalog (a VSO language), may put the verb first as in Work hard these people. Furthermore, the set of pronouns is often simpler in pidgins. In an English-based pidgin spoken in Cameroon (CP), the pronoun system does not show gender or all the case differences that exist in Standard English (SE):

Language Contact 7.

CP a yu i i wi wuna d‫ܭ‬m

mi yu i/am i/am wi wuna d‫ܭ‬m/am

ma yu i i wi wuna d‫ܭ‬m

11

SE I me my you you your he him his she her her we us our you you your they them their Fromkin et al. (2007, p. 435)

Pidgins have not been viewed positively in many places. In fact, through massive education, English replaced a pidgin spoken on New Zealand by the Maoris. Pidgins have been unjustly maligned according to Fromkin et al (2007, p. 436); they may serve a useful function. For example, a person can learn an English-based pidgin well enough in six moths to commence varied types of semiprofessional training. To learn English for the same purpose might require ten times as long. In multiethnic/cultural areas with mutually unintelligible languages, a pidgin could bridge the communicative and cultural gaps. (ibid., p. 436) The study of pidgins and creoles is significant to our understanding of the nature of human language and the genetically determined constraints on grammars. (ibid., p. 437)

2. 5 Language Endangerment Another likely outcome of language contact is language endangerment. Batibo (2005) observes that the term endangered language denotes a language that is threatened by extinction. The threat may arise because the pool of speakers is declining rapidly to small numbers – younger speakers are not learning to speak it or the domains in which the language is used have diminished so much that it is not routinely used in the language community. The linguistic structures of a language are so eroded and simplified that the language is progressively becoming non-functional. Perhaps, it is more accurate to regard language endangerment as a continuum, or a sliding scale, with ‘safe’ languages at one end and ‘dying’ languages at the other. Broadly, the minority languages would be the ones which would be found in the ‘endangered’ zone, as many of them would manifest the main features of endangered languages. Brenzinger and de Graf (2006, pp. 3-4) add that language endangerment may be caused primarily by external forces such as military, economic, religious, cultural, or educational subjugation. It may also be caused by internal forces, such as a community’s negative attitude towards its own

12

Chapter One

language or by a general decline of group identity. Internal pressures always derive from external factors. Together, they halt the intergenerational transmission of linguistic and cultural traditions. Many minority communities associate their disadvantaged social and economic position with their ancestral culture and language. They have come to believe that their languages are of no use anymore and not worth retaining. Speakers of minority languages abandon their languages and cultures in the hope of overcoming discrimination, to secure a livelihood and enhance social mobility for themselves and their children. (ibid., pp. 3-4) Furthermore, they explain that endangered languages are not necessarily languages with few speakers. Even though small communities are more vulnerable to external threats, the size of a group does not always matters. The viability of a language is determined first and foremost by the general attitude of its speakers towards their heritage culture, of which their language may be considered the most important component. In this respect, they cite the case of the Suruaha, a small Indian community that lives in a remote area of Amazonia in Brazil, consists of approximately 150 members and all of them – including the children – were monolingual in Suruaha at the time of first contact with linguists. Despite the small size of the population, the community holds on to its language and traditional way of life in all domains (Brenzinger and de Graf 2006, pp. 3-4). Simply put, in normal circumstances, no community would like to see its language die as a language provides a communicative and interactive lifeline for its speakers (Batibo 2005, p. 63; see also Batibo this volume). It is important to point out that the discussions in the above sections are by no means exhaustive of all the possible outcomes of language contact. For instance, one can add that language contact engenders all sorts of attitudes towards the speakers and languages in a speech community. On both a societal and an interpersonal level, the languages that a person is able to speak are an index of that person’s position in society – or what Pierre Bourdieu (2005) refers to as a person’s symbolic power. The value of a language variety as an “asset” is related to other forms of capital associated with the specific variety in question (economic capital, cultural capital). Altogether, these forms of capital define the place of an individual in social “space” (i.e., the person who speaks that specific variety). (Myers-Scotton 2006, p. 114)

3. The Structure of this Book Though a great number of book volumes have appeared over the years exploring various aspects of language contact and associated phenomena,

Language Contact

13

the present volume is unique in that it brings together research by distinguished scholars and other highly talented researchers from across the world to offer a multidimensional exploration of the field. The individual chapters present contemporary discussions and analyses of the topics grouped in three parts:

Part I: Contact, Change and Competence In Chapter Two, Carol Myers-Scotton explores the likely “costs” in language production in codeswitching corpora. Specifically, she argues that the Matrix Language (ML) frame can accept Embedded Language (EL) verbs as lexical items in many Codeswitching (CS) pairs, but not the structure-building verbal morphology of the EL. In Chapter Three, Najat Benchiba-Savenius examines a grammatical category which is barely touched upon in current sociolinguistics literature, namely that of the placement of Moroccan Arabic and English diminutives within codeswitched data. In Chapter Four, Elena Papadopoulou, Evelina Leivada, and Natalia Pavlou investigate the hypothesis that embu ‘{it-}is- {it-}that’ is an underlying form of clefts as well as deciphering exhaustivity effects between cleft and embu-structures in Cypriot Greek. In Chapter Five, using empirical examples from the lexical, phonological, morphological, syntactic and pragmatic domains and aspects of the present sociolinguistic situation, Aminem Mementin illustrates contact induced language changes in Modern Uyghur, a language spoken by a Muslim Turkic nation living in Western China. In Chapter Six, Frances La Morgia tests the assumption that language acquisition cannot take place without sufficient access to linguistic input and that successful bilingual acquisition is only achievable when the child is exposed to the two languages and uses them both frequently.

Part II: Language Shift and Language Maintenance In Chapter Seven, Herman Batibo explores the cultural incompatibilities that have been brought about by the use of one code in a cultural and social context of the other and the effects of this development on the cultural future of the minority language speakers or their mastery of the cultural norms of the major languages. In Chapter Eight, Aditi Ghosh proposes in her paper that migration is one of the major factors for influencing language maintenance and shift. Using language use and attitudes data from a cross section of Bhojpuri speakers in Kolkata she opines that a migrated language is almost always faced with dominant

14

Chapter One

varieties, which motivates language shift. In Chapter Nine, Carmela Perta analyses some contact phenomena in two francoprovençal communities in Southern Italy – Faeto and Celle San Vito. On the basis of her overall findings, it emerges that the minority language in Faeto is largely diffused among the population: francoprovençal is used in all ‘inner' domains (such as family and friends domains) and gradually extends itself to contexts once reserved to Italian; whereas, in Celle it is possible to observe a gradual decline of the minority language and its shift to the regional variety of Italian and to the local Italo-romance dialect - a default minority language situation, where the linguistic loyalty is associated to old, rural areas and primary sector employment. In Chapter Ten, Lisbeth Philip focuses on the role of attitudes and identity in the maintenance and shift of Limón English spoken by 127 Afro-Costa Rican women in two bilingual communities in the province of Limón, Costa Rica. In Chapter Eleven, Shihan de Silva Jayasuriya explores how trades of the mother-tongue speakers, who ascribed importance to their Portuguese identity and the significance of music as a centripetal force to bind the community, have all helped in the longevity of Sri Lanka Portuguese Creole (SLPC).

Part III: Language Attitudes and Identity In Chapter Twelve, Marina Ronþeviü, Branka Blaževiü, and Tatjana Šepiü investigate the existence of multilayered identity of children in mixed-marriage bilingual families by assessing and determining the extent to which the usage of mother's minority language and the perception of domination by one parent influence the shaping of one's identity, be it a single identity or a multilayered one. In Chapter Thirteen, Thomas Hawes and Seyed Mirvahedi examine the case of language attrition in Tabriz, Iran as indexical of membership of mainstream society and “loyalty” to the central government. In Chapter Fourteen, Charles Mann uses the voices of questionnaire and interview respondents to reflect attitudes toward AngloNigeria Pidgin (ANP) (variously called ‘Nigerian Pidgin English’, ‘Nigerian Pidgin’, ‘Broken’, and more recently ‘Naijá’) in twelve urban centres in Nigeria in a context of non-formal, non-standard, socially-marked varieties of language. Finally, I hope that teachers and students will find this volume useful on courses in language contact and change, language shift and maintenance, language attitudes and identity, and broadly language in society. The volume also includes exercises designed to give students sound understanding of the various topics.

Language Contact

15

References and Further Reading Adalar, N. and Tagliamonte, S. 1998. Borrowed nouns; bilingual people: The case of the “Londrali” in Northern Cyprus. In: S. Poplack, and M. Meechan (eds.), Instant loans, easy conditions: The productivity of bilingual borrowing (Special issue of The International Journal of Bilingualism), 139-159. London: Kingston Press Services. Anderson, H. 1995. Historical linguistics. Amsterdam: John Benjamins. Auer, P. 1995. The pragmatics of code-switching: A sequential approach. In: L. Milroy and P. Muysken (eds.), One speaker, two languages, cross disciplinary perspectives, pp. 115-135. Cambridge: Cambridge University Press. —. (ed.) 1998. Code-switching in conversation: Language, interaction and identity. London: Routledge. Batibo, H. M. 2005. Language decline and death in Africa. Clevedon: Multilingual Matters. Belazi, H. M., Rubin, E. J., and Toribio, A. J. 1994. Code-switching and X-bar theory: The functional head constraint. Linguistic Inquiry, 25, pp. 221-237. Bokamba, E. G. 1987. Are there syntactic constraints on code-switching? In: K. M. Denning, S. Inkelas, S. C. McNair-Knox, and J. R. Rickford (eds.), Variation in language: Proceedings of the fifteenth annual conference on new ways of analyzing variation, pp. 35-51. Stanford, Department of Linguistics, Stanford University. Bhatt, R. M. 1997a. Constraints, code-switching and optimal grammars. Lingua, 4 (102), pp. 123-151. —. 1997b. Code-Switching and the Functional Head Constraint. World Englishes, 1 (16), pp. 171-176. —. 1999. Verb Movement and the Syntax of Kashmiri, in the series Studies in Natural Language and Linguistic Theory, Kluwer Academic Press. —. 2001a. Constraint demotion and null-subjects in Spanish L2 acquisition. In J. Camps and C. Wiltshire (eds.), Romance Syntax, Semantics and their L2 Acquisition, pp. 121-135. Philadelphia: John Benjamin. —. 2001b. Code-switching: Structural models. In R. Mesthrie (ed.), Concise Encyclopedia of Sociolinguistics, pp. 456-461. Amsterdam: Elsevier Science. Bhatt, R. M. and Hancin-Bhatt, B. 2002. Structural minimality, CP and the initial state in second language acquisition. Second Language Research, 4 (18), pp. 348-392.

16

Chapter One

Bhatia, T. K. and Ritchie, W. C. 1996. Bilingual language mixing, universal grammar, and second language acquisition. In W. C. Ritchie and T. K. Bhatia (eds.), Handbook of second language acquisition, pp. 627-88. San Diego: Academic Press. Blake, B. J. and Burridge, K. 2003. Historical linguistics. Philadelphia: John Benjamins. Budzhak-Jones, S. 1998. Against word-internal code-switching: Evidence from Ukrainian-English bilingualism. In: S. Poplack, and M. Meechan (eds.), Instant loans, easy conditions: The productivity of bilingual borrowing (Special issue of The International Journal of Bilingualism), pp. 161-182. London: Kingston Press Services. Bourdieu, P. 2005. Language and symbolic power. Oxford: Blackwell Publishing. Brenzinger, M. and de Graf, T. 2006. Language Documentation and Maintenance. In: Encyclopedia of Life Support Systems (EOLSS 6.20B.10.3), UNESCO. Campbell, L. 2004. Historical linguistics: An introduction. Edinburgh: Edinburgh University Press. Chan, B, H-S. 1998. Functional heads, Cantonese phrase structure and Cantonese-English code-switching. UCL Working Papers in Linguistics, 10, 1-34. Chan, B. 2003. Aspects of the syntax, the pragmatics and the production of codeswitching: Cantonese and English. New York: Peter Long. —. 2004. Beyond contextualisation: Code-switching as a textualisation cue. Journal of Language and Social Psychology, 23 (1), pp. 7-27. Chomsky, N. 1995. The minimalist programme. Cambridge, MA: MIT Press.Clyne, M. 2003. Dynamics of language contact. Cambridge: Cambridge University Press. Cummins, J. 2003. BICS and CALP: Origins and rational for the distinction. In: C. B. Paulston and G. R. Tucker (eds.), Sociolinguistics: The essential readings, pp. 322-328. Oxford: Blackwell Publishing. Dimitrijeviü-Saviü, J. 2008. Convergence and Attrition: Serbian in Contact with English in Australia. Journal of Slavic Linguistics, Vol 16 (1), pp. 57-90. DiSciullo, A. M. (2002. The asymmetry of morphology. In: P. Boucher (ed.), Many morphologies, 1-33. Somerville, MA: Cascadilla Press. DiSciullo, A. M., Muysken, P., and Singh, R. 1986. Code-mixing and government. Journal of Linguistics, 22, 1-24. Eliasson, S. 1989. English-Maori language contact: Code-switching and the free morpheme constraint. Reports from Uppsala University Department of Linguistics, 18, pp. 1-28.

Language Contact

17

Ellis, R. 1985. Understanding second language acquisition. Oxford: Oxford University Press. El-Noory, A. 1985. Code-switching and the search for universals: A study of (Egyptian) Arabic/English bilingualism. Paper Presented at the 16th Annual Conference on African Linguistics. Yale, New Haven. Eze, E. 1998. Lending credence to a borrowing analysis: Lone Englishorigin incorporations in Igbo discourse. In: S. Poplack, and M. Meechan (eds.), Instant loans, easy conditions: The productivity of bilingual borrowing (Special issue of The International Journal of Bilingualism), 183-201. London: Kingston Press Services. Ezeizabarrena, M-J. 2003. Null Subjects and optional infinitives in Basque. In N. Müller (ed.), (In)vulnerable Domains in Multilingualism, pp. 83–106. Fishman, J. A. 1964. Language maintenance and language shift as a field of inquiry. Linguistics, 9, pp. 32-70. —. 1965. Who speaks what language to whom and when? Linguistics, 2, pp. 67-8. —. 1968. Sociolinguistic perspectives in the study of bilingualism. Linguistics, 39, pp. 21-48. Gleason, J. B. and Ratner, B. N. 1998. Psycholinguistics. Forth Worth: Harcourt Bruce College Publishers. Greenfield, L. 1972. Situational measures of normative language views in relation to person, place and topic among Puerto Rican bilinguals. In: J. Fishman (ed.), Advances in the sociology of language, pp. 1735.Gumperz, J. J. 1982. Discourse strategies. Cambridge: Cambridge University Press. Halmari, H. 1997. Government and code switching: Explaining AmericanFinnish. Amsterdam: John Benjamins. Hinzelin, M-O. 2003. The acquisition of subjects in bilingual children: Pronoun use in Portuguese-German children. In: N, Müller (ed.), (In)vulnerable Domains in Multilingualism, pp. 107–137. Haugen, E. 1953 [1969]. The Norwegian language in the Americas: A study in bilingual behavior. Bloomington: Indiana University Press. Hock, H. H. 1986. Principles of historical linguistics. Berlin: Mouton de Gruyter. Ihemere, K. U. 2007. A Tri-Generational Study of Language Choice and Shift in Port Harcourt. Boca Raton, Florida: Universal Publishers, USA. —. (ed.) 2011. Language Contact and Language Shift: Grammatical and Sociolinguistic Perspectives. Munich, Germany: Lincom Europa Academic Publishers.

18

Chapter One

Jake, J. L., Myers-Scotton, C., and Gross, S. 2002. Making a minimalist approach to code-switching work: Adding the matrix language. Bilingualism: Language and Cognition, 5, pp. 69-91. Kamwangamalu, N. M. 1989. Some morpho-syntactic aspects of French/English-Bantu code-mixing: Evidence for universal constraints. In: B. Music, R. Groezyk, and C. Wiltshire (eds.), Parasession on language contact: Regional meeting of the Chicago Linguistic Society, 2, 157-170. Chicago: Chicago Linguistic Society. Kess, J. F. 1976. Psycholinguistics: An introductory perspectives. New York: Academic Press. Klein, W. 1986. Second language acquisition. Cambridge: Cambridge University Press. Kupisch, T. 2003. The DP, a vulnerable domain?: Evidence from the acquisition of French. In N. Müller (ed.), (In)vulnerable Domains in Multilingualism, 1–39. Lehman, W. P. 1992. Historical linguistics: An introduction. London: Routledge. Levinson, S. C. 1995. Three levels of meaning. In: F. Palmer (ed.), Grammar and Meaning, pp. 90-115. Cambridge: Cambridge University Press. Lippi-Green, R. L. 1989. Social network integration and language change in progress in a rural alpine village. Language in Society, 18 (2), pp. 213-34. Matras, Y. 2009. Language Contact. Cambridge: Cambridge University Press. McMahon, A. M. S. 1994. Understanding language change. Cambridge: Cambridge University Press. MacSwan, J. 1999. A minimalist approach to intra-sentential codeswitching. New York: Garland Publishing. —. 2000. The threshold hypothesis: Semilingualism, and other contributions to a deficit view of linguistic minorities. Hispanic Journal of Behavioural Science, 22 (1), pp. 3-45. Meyerhoff, M. 2011. Introducing sociolinguistics. London: Routledge Möhring, A. and Meisel, J. M. 2003. The verb–object parameter in simultaneous and successive acquisition of bilingualism. In: N. Müller (ed.), (In)vulnerable Domains in Multilingualism, pp. 295–334. Mufwene, S., Baugh, J., and Bailey, G. (eds.) 1998. African-American English: Structure, History and Use. London: Routledge. Muysken, P. 1987. Neutrality in code-mixing. In Eigen en vreemd: Identiteit en ontlening in taal, literatuur en beeldende kunst: Handelingen van het 39ste Nederlands Filologencongres, Vrije

Language Contact

19

Universiteit, Amsterdam, 18-19 December, 1986, 359-373. Amsterdam: VU, Uitgeverji. —. 2000. Bilingual speech: A typology of code-mixing. Cambridge: Cambridge University Press. —. 2005. A modular approach to sociolinguistic variation in syntax: The gerund in Ecuadorian Spanish. In: L. Cornips and P.Corrigan (eds.), Syntax and Variation, 31-53. Amsterdam: John Benjamins. —. 2006. Review of An introduction to contact linguistics’. Journal of Pidgin and Creole Languages, 1 (21), pp. 192–195. Myers-Scotton, C. M. 1983. The negotiation of identities in conversation: A theory of markedness and code choice. International Journal of the Sociology of Language, 44, pp. 115-36. —. 1992. Comparing code-switching and borrowing. Journal of Multilingual and Multicultural Development, 13, pp. 19-39. —. 1993. Duelling languages: Grammatical structure in codeswitching. Oxford: Oxford University Press. —. 2002. Contact linguistics: Bilingual Encounters and grammatical outcomes. Oxford: Oxford University Press. Myers-Scotton, C. 2006. Multiple voices: An introduction to bilingualism. Oxford: Blackwell Publishing. Myers-Scotton, C. M. and Jake, J. L. 2000. Four times of morpheme: Evidence from aphasia, code-switching and second language acquisition. Linguistics, 38, pp. 1053-100. —. 2001. Explaining aspects of code-switching and their implications. In: J. Nicol (ed.), One mind, two languages: Bilingual language processing, pp. 84-116. Oxford: Blackwell. Myers-Scotton, C. and Bolonyai, A. 2001. Calculating speakers: Codeswitching in a rational choice model. Language in Society, 30, pp. 128. Nartey, J. A. 1982. Code-switching: Interference or Faddism? Language use among educated Ghanaians. Anthropological Linguistics, 24, pp. 183-192. Parasher, S. N. 1980. Mother-tongue-English diglossia: A case study of educated Indian bilinguals’ language use. Anthropological Linguistics, 22 (4), pp. 151-68. Poplack, S. 1980. Sometimes I’ll start a sentence in Spanish y termino en ESPANOL: Towards a typology of code-switching. Linguistics, 18, 581–618. Poplack, S. 1981. Syntactic structure and social function of codeswitching. In: R. Duran (ed.), Latino discourse and communicative behaviour, 169-84. New Jersey: Ablex.

20

Chapter One

Poplack, S., Wheeler, S. and Westwood, A. 1987. Distinguishing language contact phenomena: Evidence from Finnish-English bilingualism. In: P. Lilius and M. Saari (eds.), The Nordic languages and modern linguistics, 6, pp. 33-56. Helsinki: University of Helsinki Press. Poplack, S. and Meechan, M. 1995. Patterns of language mixture: Nominal structure in Wolof-French and Fongbe-French bilingual discourse. In: L. Milroy and P. Muysken (eds.), One speaker, two languages: Cross-disciplinary perspectives on code-switching, pp. 199232. Cambridge: Cambridge University Press. Rumbaut, R. G. and Portes, A. (eds.) 2001. Ethnicities: Children of immigrants in America. Berkley: University of California Press. Samar, G. R. and Meechan, M. 1998. In: S. Poplack, and M. Meechan (eds.), Instant loans, easy conditions: The productivity of bilingual borrowing (Special issue of The International Journal of Bilingualism), 203-219. London: Kingston Press Services. Sankoff, D., Poplack, S., and Vanniarajan, S. 1991. The empirical study of code-switching. In European Science Foundation (ed.), pp. 181-206. Strasbourg: European Science Foundation. Shin, S. J. 2002. Differentiating language contact phenomena: Evidence from Korean-English bilingual children. Applied Psycholinguistics, 23, 337-60. Silva-Corvalán, C. 1994. Language contact and change: Spanish in Los Angeles. Oxford: Clarendon Press. Steinberg, D. D., Nagata, H., and Aline, D. P. 2001. Psycholinguistics: Language, mind, and world. Harlow: Longman. Thomason, S. and Terrence, K. 1988. Language Contact, Creolization and Genetic Linguistics. California: University of California Press. Thomason, S. 2001. Language Contact: An Introduction. Edinburgh: Edinburgh University Press. Toribio, A. J. and Rubin, E. J. 1996. Code-switching in generative grammar. In: A. Roca and J. B. Jensen (eds.), Spanish in contact: Issues in bilingualism, 203-26. Somerville, MA: Cascadilla Press. Trask, R. L. 1996. Historical linguistics. London: Arnold. Turpin, D. 1998. Le français, c’est le last frontier’: The status of Englishorigin nouns in Acadian French. In: S. Poplack, and M. Meechan (eds.), Instant loans, easy conditions: The productivity of bilingual borrowing (Special issue of The International Journal of Bilingualism), 221-233. London: Kingston Press Services. Weinreich, U. 1963. Languages in Contact. Berlin: Mouton. Winford, D. 2002. An Introduction to Contact Linguistics. Oxford: Blackwell Publishing.

Language Contact

21

Exercises 1. What is language contact? 2. Identify and discuss the likely outcomes of language contact. 3. Explain how the effects of language contact can be both sociological and linguistic. 4. Conduct a small-scale study of your community and identify any influences on your language due to past or recent language contact.

PART I: CONTACT, CHANGE AND COMPETENCE

EDITOR’S INTRODUCTION TO PART ONE

Investigators working with mixed language data are cognisant of the fact that there are constraints of some kind at major constituent boundaries or semantic unit boundaries, on the switch point (although few have managed to put forward constraints, to which counter-examples have not been increasingly suggested, see for instance Bhatt 1997a, b, 1999, 2001a, b; Bhatt and Hancin-Bhatt, 2002). Muysken and his colleagues (DiSciullo 2002; DiSciullo, Muysken and Singh 1986; Muysken 1990, 1991, 1995, 2000, 2005, 2006) have sought to capture the evident but elusive regularities in codeswitching behaviour by positing various kinds of borrowability hierarchies and switched constituent hierarchies. On the other hand, Myers-Scotton (1983, 1993) observed that in many of the world’s bilingual communities, fluent bilinguals sometimes engage in codeswitching by producing discourses which, in the same conversational turn or in consecutive turns, include morphemes from two or more of the varieties in their linguistic repertoire. In various publications, MyersScotton developed the theme of codeswitching as a tool for the speaker and an index for the addressee of the negotiation of interpersonal relationships, with participants cast within a “rational actor’’ framework, weighing costs and rewards of choices made against a backdrop of awareness for all interaction types of “unmarked versus marked’’ choices. Her argument was based on what she termed the Markedness Model (MM). In this volume, she considers likely “costs” in language production in codeswitching corpora. Although no costs are actually measured here, psycholinguistic experiments that track Response Time could do this, counting costs as the time it takes a speaker in natural discourse to initiate the morpheme or structure in question. Asymmetry between the participating languages in supplying important grammatical structure is one way production costs are reduced in codeswitching. When uniformity of structure under the Uniform Structure Principle is paired with asymmetry, this is another way in which production costs are evaded; structures of the Matrix Language (ML) are preferred (Myers-Scotton 2002; Myers-Scotton and Jake 2009). The new argument in her contribution, however, is to suggest that the robustness of some Embedded Language (EL) morphemes in codeswitching is related to their low production costs. The hypothesis is that production cost is reduced if the

Editor’s Introduction to Part One

25

abstract lexical structures of a particular EL morpheme type require relatively little checking for congruence with the structure that the ML frame projects. In turn, reduced cost promotes frequent employment of the EL morpheme type in question. (The premise is that even though the Abstract Level model assumes three levels of abstract lexical structure in linguistic elements, not all are equally salient or even present in all morpheme types.) Employing the structural framework of the Matrix Language Frame Model (MLF, Myers-Scotton 2002) and its application to intra-sentential codeswitched data from English to Moroccan Arabic, Benchiba-Savenius examines a grammatical category which is barely touched upon in current sociolinguistic literature, namely that of the placement of Moroccan (MA) Arabic and English diminutives within codeswitched data. This is evidenced whereby speakers employ certain discourse strategies particularly in terms of implications of translations and conveyance of meaning. In most cases, there is little lexical equivalence in translating MA diminutives into accurate English other than adding a string of further adjectives to make meaning more transparent. Such cultural substitution is a translation technique employed by most MA/English bilingual speakers. In this regard, the grammatical categories of diminutives and lexical insertion in general make the issue of accurate translation from L1 to L2 highly relevant particularly where gender assignment of each MA diminutive plays a major role in noun-adjective agreement and mutual intelligibility. The third paper in this section by Papadopoulou, Leivada, and Pavlou investigates the hypothesis that embu ‘{it-}is- {it-}that’ is an underlying form of cleft as well as deciphering exhaustivity effects between cleft and embu-structures in Cypriot Greek (CG). The latter is a variety that lacks recognition as official language in Cyprus and, as a result, it is heavily influenced by the use of Standard Modern Greek, which is one of the two official languages in Cyprus. The results of their research challenge the availability of bona fide clefts in CG, since all participants in around 50% of all stories in all conditions allow for non-exhaustive interpretation with ‘it is XP that YP’ clefts, despite the fact that this construction has been argued to carry typical properties attributed to clefts, such as exhaustivity. What the results reveal is wide intra-dialectal variation, since all sentences elicited mixed responses with respect to their status as true or false. This variation they claim can be attributed to morpho-syntactic change in progress. Following this paper is that by Memtimin which explores contact induced changes in Modern Uyghur (an Subject Object Verb (SOV)

26

Editor’s Introduction to Part One

language), presently using Arabic based writing system, with agglutinative suffixing and vowel harmony. Ample examples are cited from the lexical, phonological, morpho-syntactic and pragmatic domains and aspects of the present sociolinguistic situation to illustrate contact induced changes in modern Uyghur. Francesca La Morgia in her contribution presents findings from an investigation on the role of the input in heritage language acquisition. The first assumption in her paper is that language acquisition cannot take place without sufficient access to linguistic input and that successful bilingual acquisition is only achievable when the child is exposed to the two languages and uses them both frequently. The quantity and quality of input and the subjects’ development are measured by analysing spontaneous conversations between Italian mothers and their children over one year. The results of the study show that frequent exposure to qualitatively rich input allows the child to develop the heritage language faster and more similarly to an L1, while exposure to more limited quality or quantity of input can result in slow rate of development, cross-linguistic influence and fossilisation errors, and may ultimately be responsible for incomplete acquisition. These papers offer clearly presented theoretical and experimental models that will be of interest to both teachers and students of contact linguistics.

References and Further Reading Bhatt, R. M. 1997a. Constraints, code-switching and optimal grammars. Lingua, 4 (102), pp. 123-151. —. 1997b. Code-Switching and the Functional Head Constraint. World Englishes, 1 (16), pp. 171-176. —. 1999. Verb Movement and the Syntax of Kashmiri, in the series Studies in Natural Language and Linguistic Theory, Kluwer Academic Press. —. 2001a. Constraint demotion and null-subjects in Spanish L2 acquisition. In: J. Camps and C. Wiltshire (eds.), Romance Syntax, Semantics and their L2 Acquisition, pp. 121-135. Philadelphia: John Benjamin. —. 2001b. Code-switching: Structural models. In: R. Mesthrie (ed.), Concise Encyclopedia of Sociolinguistics, pp. 456-461. Amsterdam: Elsevier Science. Bhatt, R. M. and Hancin-Bhatt, B. 2002. Structural minimality, CP and the initial state in second language acquisition. Second Language Research, 4 (18), pp. 348-392.

Editor’s Introduction to Part One

27

DiSciullo, A. M. 2002. The asymmetry of morphology. In: P. Boucher (ed.), Many morphologies, pp. 1-33. Somerville, MA: Cascadilla Press. DiSciullo, A. M., Muysken, P., and Singh, R. 1986. Code-mixing and government. Journal of Linguistics, 22, pp. 1-24. Ezeizabarrena, M-J. 2003 Null Subjects and optional infinitives in Basque. In: N. Müller (ed.), (In)vulnerable Domains in Multilingualism, pp. 83–106. González-Vilbazo, K. and López, L. 2011. Some properties of light verbs in code-switching. Lingua, 121: pp. 832-850. Hinzelin, M-O. 2003. The acquisition of subjects in bilingual children: Pronoun use in Portuguese-German children. In: N, Müller (ed.), (In)vulnerable Domains in Multilingualism, pp. 107–137. Holm, A., Holm, W., and Spolsky, B. 1971. “English loan words in the speech of six-year-old Navajo chidren.” Navajo Reading Study Progress Report 16. University of New Mexico: Albuquerque, NM. Kupisch, T. 2003. The DP, a vulnerable domain?: Evidence from the acquisition of French. In: N. Müller (ed.), (In)vulnerable Domains in Multilingualism, pp. 1–39. LaPolla, R. J. 2000. Valency-changing derivations in Dulong/Rawang. In: R. M. W. Dixon and Alexandra Y. Aikhenvald (eds), Changing Valency: Case Studies in Transitivity, pp. 282-311. Cambridge: Cambridge University Press. MacSwan, J. 1999. A Minimalist approach to intrasentential code switching. New York: Garland Press. Möhring, A. and Meisel, J. M. 2003. The verb–object parameter in simultaneous and successive acquisition of bilingualism. In: N. Müller (ed.), (In)vulnerable Domains in Multilingualism, pp. 295–334. Muysken, P. 1990. Ten remarks from the perspective of grammatical theory. In: European Science Foundation (ed.), pp. 15-30. Strasbourg: European Science Foundation. —. 1991. Needed: A comparative approach. In: European Science Foundation (ed.), pp. 253-72. Strasbourg: European Science Foundation. —. 1995. Code-switching and grammatical theory. In: P. Muysken and L. Milroy (eds.), One speaker two languages: Cross disciplinary perspectives on code-switching. Cambridge: Cambridge University Press. —. 2000. Bilingual speech: A typology of code-mixing. Cambridge: Cambridge University Press. —. 2005. A modular approach to sociolinguistic variation in syntax: The gerund in Ecuadorian Spanish. In: L. Cornips and P.

28

Editor’s Introduction to Part One

Corrigan (eds.), Syntax and Variation, pp. 31-53. Amsterdam: John Benjamins. —. 2006. Review of An introduction to contact linguistics. Journal of Pidgin and Creole Languages, 1 (21), pp. 192–195. Myers-Scotton, C. M. 1983. The negotiation of identities in conversation: A theory of markedness and code choice. International Journal of the Sociology of Language, 44, pp. 115-36. Myers-Scotton, C. M. 1993. Duelling languages: Grammatical structure in codeswitching. Oxford: Oxford University Press. Myers-Scotton, C. 2002. Contact Linguistics: Bilingual Encounters and Grammatical Outcomes. Oxford: Oxford University Press. Myers-Scotton, C. and Jake, J., 2009. A universal model of codeswitching and bilingual language processing and production. In: Bullock, B., Toribio, J.A. (eds.), The Cambridge Handbook of Linguistic Codeswitching. Cambridge University Press, Cambridge, pp. 336–357. Ross, M. 2003. Diagnosing prehistoric language contact. In: R. Hickey (ed.), Motivations for language change, pp. 174-198. Cambridge: Cambridge University Press.

CHAPTER TWO COST MAKES A DIFFERENCE: CODESWITCHING FREQUENCIES IN BILINGUAL LANGUAGE PRODUCTION CAROL MYERS-SCOTTON

1. Introduction The goal of this paper is to show how codeswitching data provide implicational evidence about the costs of language production. Psycholinguistic experimentation provides ample evidence from various lexical tasks that both languages are available before speaking whether or not they are present in speaking. When speaking actually occurs, naturallyoccurring codeswitching (CS) shows that both languages can appear in mixed constituents bilingual clauses. Thus, both languages are available to contribute phrasal constituents and individual lexical items in the same clause. But the way that the two languages do supply linguistic material shows that there is an asymmetry in the roles of the two languages. Consistently, across diverse language pairs, only one language projects the grammatical frame of a bilingual clause. Under the Matrix Language (MLF) model, the language providing basic grammatical structure is called the Matrix Language (ML) and the other language(s) is called the Embedded Language (EL) (Myers-Scotton 1993; 1995 inter alia). Further, there is asymmetry regarding the occurrence of EL morpheme types; not all of them occur with the same frequency in bilingual clauses; some of them do not seem to occur at all, at least in some data sets. Example (1) illustrates the prevalence of one language, Spanish, in structuring the clause. (Bold face marks the EL throughout this chapter.)

30

Chapter Two 1. Spanish-English CS (Jake, Myers-Scotton and Gross 2002, p. 81) Me dieron medium y yo necessito large 1S.OBJ give.3PL.PRET medium CONJ 1S.SUBJ need.1S.PRES large ‘They gave me [a] medium when I need [a] large.’

1. 1 Explaining Asymmetries This paper is organized around two explanations, both having to do with production costs. The first is an explanation for the unequal distribution of the ML and the EL in CS. The second is an explanation for why some EL elements appear regularly, seemingly with no hesitation or other cost, but other ones do not. By “cost”, I mean the time and energy that underlies the measure of Response Time (RT) when elements are produced. There already is an extensive psycholinguistic literature on experiments comparing RT when speakers are directed to switch between lexical items from two different languages (e.g., Kroll and DeGroot 2005; Kroll, Bogulski and McClain (in press). Providing experimental evidence of comparative processing time between CS clauses and monolingual ones is beyond the scope of this paper, but I hope psycholinguists will be motivated to devise experiments to make more accurate measures of cost in CS, especially in naturally-occurring corpora (see Dussias, Guzzardo, Valdes-Kroff, and Gerfen (in press) for using eye tracking to study some aspects of CS experimentally). Here, I suggest that production costs must be a factor affecting how certain EL elements participate or not in mixed constituents in bilingual clauses. A bilingual clause is defined as any clause that contains surface level morphemes from two or more languages. I begin with a general explanation of asymmetry that seems obvious: The costs of having structure-building contributions from both participating languages is too high to warrant having two sources for these aspects of morpho-syntax. The two principles structuring the MLF model are, in effect, hypotheses about limits on the structure of CS clauses. They refer to roles that only one of the participating languages fulfils; in so doing, they identify the language that is to be called the ML. Just as important, they limit EL participation, and this limits production costs. Briefly stated, the Morpheme Order Principle limits morpheme order to that of one of the participating languages. The System Morpheme Principle limits one language as the source of what are now called the “late” system morphemes under the 4-M model; these are the ones that co-index relationships and indicate other hierarchical relationships (see MyersScotton 1993; 1995, p. 88 for the original statements of the principles and Chapter One (this volume) for an overview). Examples such as (1) and

Cost Makes a Difference

31

much other empirical evidence indicate that the principles are accurate in identifying what these parts of CS clauses would look like. In this example and in most CS data sets, the EL’s main contribution consists of content words, especially nouns, not grammatical structure. Also, EL participation was limited in another way - in regard to congruence. The first book-length treatment dealing with the MLF model discusses congruence; it states that in order for EL elements to appear in mixed constituents they are expected to show congruence to ML counterparts, although the notion of congruence was never formalized (Myers-Scotton 1993). Congruence was discussed more extensively in Myers-Scotton and Jake (1995) inter alia. Exactly what was meant by congruence was never stated, although Myers-Scotton (1993, 1995, p. 120) did include a Blocking Hypothesis that referred to “three levels of abstraction regarding subcategorisation.” Later, Myers-Scotton and Jake began referring to the Abstract Level Model and three levels of abstract lexical information in relation to congruence (e.g., Myers-Scotton and Jake 2001; Myers-Scotton 2002). Since then, our thinking has evolved. Some of the ways are discussed below when the Abstract Level Model is introduced. The second explanation proposed here about asymmetries in CS and production costs is a hypothesis about further ways that the EL is limited in CS. The hypothesis is that differences in the salience of the three levels of abstract structure across EL morpheme type are related to the production costs they carry; in turn, the costs are related to frequency of EL morphemes in CS.

1. 2 Data to be Considered To support this hypothesis, I discuss the robustness of some EL morpheme types, such as nouns. I also develop an argument about EL verbs, asserting that it is EL non-finite verb forms, but not EL finite verbs, that occur in bilingual clauses. EL verbs are found in two types of ML verb phrases. First, in many data sets, but not all, EL verbs receive grammatical inflections for agreement and tense/aspect from the ML, as does the EL verb buy in (2). All the inflections come from Swahili, but note especially those that co-index agreement (the subject prefix for ‘he’, a- and the object prefix -ni- ‘me’). This is in line with the System Morpheme Principle of the MLF model discussed above. Second, less often, but still frequently and across diverse corpora, EL verbs appear in “do” constructions as infinitives that carry the intended meaning of the verb phrase, as does the Dutch infinitive kijen in (3). That is, an ML “do”

32

Chapter Two

verb receives all relevant verbal inflections and an EL verb appears as an infinitival form. In (3) Turkish-Dutch CS, the Turkish verb for ‘do’ (yap) is inflected with morphology conveying subject agreement and is followed by a Dutch infinitive. These two types of verb phrases will be discussed further later in the paper. 2. Swahili-English CS (Myers-Scotton 2002, p. 98) Hau ku-on-a a-ki-ni-buy-i-a beer siku hi-yo? 2S.NEG-NEG-see-FV 3s-PROG-1S.OBJ-buy-INSTR-FV beer day CL9.DEM-that ‘Didn’t you see him buying beer for me that day?’ 3. Turkish-Dutch CS (Backus 1996, cited in Myers-Scotton 2002, p. 134) ja, maar toch, millet kijk-en yap-iyor yes but still everybody watch-INF do-PROG.3S ‘Yeah, but still everybody is watching you’

1. 3 Where the EL provides Structure Note that in neither of these cases does the EL provide basic grammatical structure in mixed constituents in the bilingual clause. Does this mean that EL verbs can never provide structure in bilingual corpora? No, but it means that when this happens, in effect, what had been the EL in earlier clauses in the discourse, becomes the ML of a new clause. What makes the new clause a clause in its own right is that it includes finite verbal inflections from the former EL. This new clause is generally a monolingual clause (example 4), but it may be bilingual, as is (5). Some CS data sets do show such switches of the ML from one clause to the next, although many data sets maintain one language as the ML (i.e., all participants employ the same ML). Both examples are from a XhosaEnglish data set with a number of switches of the ML by the same speaker. In (4), a monolingual English clause, and I have been in other towns, follows a bilingual clause in which English is the EL with Xhosa as the ML. Example (5) is preceded and followed by monolingual clauses in English, although monolingual Xhosa clauses or bilingual clauses with Xhosa as the ML dominate in this particular corpus.

Cost Makes a Difference

33

4. Xhosa-English CS (Myers-Scotton 2004 South African corpus) i-Komani a-yi-kho developed like … CL9-Queenstown NEG-CL9-CL9-LOC developed [filler] ‘Queenstown, it is not developed, like, … And I have been in other towns … Queenstown, it is not developed, like … And I have been in other towns …’ 5. Xhosa-English CS (Myers-Scotton, 2004 South African corpus) But every time I go back to

e-khaya LOC-home

area ‘But every time I go back to [my] home area, …’

Another strategy to introduce EL material into a bilingual clause is to produce what is called an EL island. This is a phrase that occurs within a clause with a grammatical frame from the ML and therefore the ML does not change. However, such islands are usually noun phrases or prepositional phrases. Especially EL islands that are PPs tend to be adjuncts (as is for that party in example 6), rather than part of the core arguments of the clause. To my knowledge, no data set shows an EL verb with finite inflections in an EL island. Two EL islands are illustrated here; the one in (7) is a set phrase/collocation; such islands are relatively common. EL islands are not discussed further here (but see Myers-Scotton 2002 or 2006 for more detail). 6. Swahili-English CS (Myers-Scotton 1988, Nairobi corpus) ...na-tak-a tu ku-inquire just INF-inquire 1S.NON-PST-want-FV a-li-ku-j-a for that party. 3S-PST-INF-come-FV for that party ‘I just want to inquire if he came for that party.’

kama if

7. Xhosa-English CS (Myers-Scotton 2004 South African corpus) ... at the end of the day i-yez-a i-zo-ku-galet-a CL9-FUT-OBJ-infect-FV ‘... at the end of the day, CL9-come-FV `‘...at the end of the day, it [AIDS] comes it will infect you.’

34

Chapter Two

2. Theoretical Background Two universals are relevant to discussions in this paper, the Uniform Structure Principle (USP) and the Asymmetry Principle (AP). The USP is relevant to all language production, monolingual as well as bilingual data. That is, the USP recognizes the universality of what constitutes language. Here is the principle in brief: Any given constituent type in any language has a uniform abstract structure. The requirements of this constituent type must be observed when the constituent appears. In bilingual speech, the structures for constituent types of the ML are always preferred (see MyersScotton 2002 or Myers-Scotton and Jake 2009 for more details). In CS, the USP predicts that the ML will supply most or all critical grammatical structure in the bilingual clause. This principle explains why EL inflectional morphemes do not provide structure in the VPs to be discussed below. However, the problem of why EL verbs seemingly easily receive ML versions of these morphemes still needs an explanation. The Asymmetry Principle is specific to bilingual data. It refers to the structural inequality of several types in bilingual contact data, not just in CS. First, asymmetry is found between the participating languages in regard to the ways they participate in a bilingual clause. At the surface level, this asymmetry is reflected in CS in a minimum of two ways that are empirically obvious: (a) the morpho-syntactic frame of the clause largely or entirely reflects that of only one of the participating languages. That is, the critical grammatical morphemes in mixed constituents come only from one language, called the ML in the MLF model. The other participating language(s) is called the Embedded Language (EL). (b) Word order follows that of the ML. Second, the Asymmetry Principle is also reflected in the division between content and system morphemes, as well as between different types of system morphemes. This division is empirically evident in CS data sets in the distribution of morpheme types in any data set. The most obvious way is that the EL regularly supplies content morphemes, especially nouns, but also verbs and other morphemes that carry the speaker’s intentions, but only one type of system morpheme. This type is called an early system morpheme under the 4-M model discussed below. Just as important, there is an asymmetry in morpheme type at the abstract level; this is the division between conceptually-activated morphemes and structurally-assigned morphemes. This division is an essential part of the 4-M model, to be discussed shortly.

Cost Makes a Difference

35

3. Relevant Linguistic Models Three models that are relevant to CS are important to the discussion in this paper of EL morphemes. They are the Abstract Level Model, the 4-M Model, and the MLF Model. My original formulation of the MLF Model (Myers-Scotton 1993; 1995) implied that, on its own, it could explain the nature of CS data. However, except for recognizing the asymmetry between the participating languages and that between content and system morphemes, its explanation applied largely only to the surface structure. The other two models add additional explanatory power to the MLF model because of their emphasis on abstract levels of structure. They also differ from the MLF model in that neither is specific to CS; they have relevance to linguistic data in general.

3. 1 The Abstract Level Model The premise of the Abstract Level Model is that all lemmas in the mental lexicon (in any production model) may include three levels of abstract structure. These levels contain all the semantic and grammatical information necessary for the surface realization of a lexical entry. The levels are: (i) lexical-conceptual structure; (ii) predicate-argument structure; and (iii) morphological realisation patterns. Other linguists discuss lexical structure, some with ideas similar to this model. What is innovative about the Abstract Level model is the notion that, in bilingual clauses, the three levels are not necessarily equally salient. This flexibility in salience is relevant to CS. It implies that the three levels in one morpheme type can be split and combined with levels in another morpheme type. This can make a difference in CS when the issue of inserting an EL element into an ML frame arises. It is especially relevant to explaining how EL verbs can participate in ML frames with ML verbal inflections.

3. 2 The 4-M Model The 4-M Model classifies morphemes into four types. It is a model of morpheme classification in general; it is not an extension of the MLF model as some have assumed. A useful feature of the model is that it distinguishes three types of system morpheme; under the MLF Model, system morphemes were just discussed as a single type. But the model’s key feature is the two-part division it makes of morpheme types at the abstract level into conceptually-activated morphemes and structurallyassigned ones. (This division is discussed briefly here; for more information

36

Chapter Two

see Myers-Scotton and Jake 2000; Myers-Scotton 2002; Myers-Scotton and Jake 2009; and Jake and Myers-Scotton 2009). Conceptually-activated morphemes reflect the speaker’s intentions about conveying semantic and pragmatic information. Content morphemes and one type of system morpheme are conceptually-activated. Content morphemes are directly-elected by speaker intentions; that is, they are primary in indicating what the speaker wants to convey. Early system morphemes are indirectly-elected by their content morpheme heads to add specificity to meaning. They are called “early” because, in contrast to structurally-assigned system morphemes, they are salient in the mental lexicon along with their content morpheme heads. Examples of early system morphemes include determiners, plural markers, and derivational elements such as satellite prepositions/particles in phrasal verbs. All conceptually-activated morphemes are salient at the level of the mental lexicon. Structurally-assigned system morphemes are called “late” because the hypothesis is that they are not salient in production until the level of the formulator. (For more on this hypothesis, see Myers-Scotton 2002; 2005a). The formulator is the production mechanism that puts together the larger constituents that are critical in indicating the hierarchical structure of the clause and other relationship among elements. The formulator receives directions from the lemmas in the mental lexicon about how these constituents are to be assembled. There are two types of late system morphemes. “Bridges” are the most responsible for hierarchical structure. They include the associative elements that join together two NPs (e.g., book of the student) and complementizers that join together two clauses. “Outsiders” make relationships more transparent in various ways. They include agreement markers and many of the morphemes marking case, although, across languages, not all morphemes that are identified as labelling case are outsider morphemes. Outsider morphemes are called outsiders because they occur outside the element that calls them; for example, case is often assigned by verbs or prepositions, but occurs on other elements, such as determiners or nouns. The 4-M Model is not of as much relevance in this chapter as is the Abstract Level Model, as will become clear. Still, a main subject of this paper is the finite verb construction in which the verb comes from the EL, but the grammatical elements that provide structure come from the ML. These are examples of late outsider system morphemes. For example, the agreement markers that co-index the verb with a nominal element that is elsewhere in the clause (or elsewhere in the discourse) are such

Cost Makes a Difference

37

morphemes. These are the type of morpheme to which the System Morpheme Principle of the MLF Model refers; this model is discussed above in the introduction and also in relation to example (2). The 4-M Model gives a name (late outside system morphemes), and a more precise description, to these morphemes than the MLF Model does.

3. 3 A Production Model The production model followed in this paper is derived from Levelt’s model in his oft-cited 1989 book that considers language production from intention to articulation. The activities of language production are not linear, but can happen simultaneously. The model has these features: x Conceptual Level: Based on sociolinguistic/psycholinguistic information, the speaker selects an ML. Bundles of semantic/pragmatic features are identified that satisfy the speaker’s intentions. x Mental Lexicon: Lemmas supporting language-specific content morphemes are directly-elected and they, in turn, indirectly elect lemmas for those early system morphemes that make intended meanings more precise. The lemmas encode information in the bundles from the conceptual level, but go beyond doing this: what defines lemmas is that they may include information at all three levels of abstract lexical structure (cf. see the discussion above of the Abstract Level Model). x Formulator Level. Structurally-assigned system morphemes are activated. These are the late system morphemes of the 4-M Model. The formulator uses them to build hierarchical structures and check relationships/dependencies between lemmas from the mental lexicon. x Surface/Functional Level. Surface-level structures, including phonological features are assembled.

4. Research Questions In this paper, the main research questions addressed are these: 1. How is it that EL nouns are so quickly and seemingly easily inserted into ML frames?

38

Chapter Two

2. How is it that two different EL early system morphemes (determiners and prepositional satellites/particles in verb phrases) differ in their frequency in CS? 3. How is it that coordinators and subordinating conjunctions appear similar to complementizers, but have different patterns of occurrence in CS? 4. How is it that EL verbs can fit into an ML frame with ML grammatical morphemes (a frame of late system morphemes)? 5. How is it that “do” verb constructions seem to have a complex surface structure, but are still so widely available in many CS corpora?

5. EL nouns Nouns from the EL are freely accepted in ML frames. Singlyoccurring nouns are the most frequent instances of EL material to appear in bilingual constituents across diverse CS data sets. Some singlyoccurring nouns are “bare”, meaning they do not receive the requisite affixes that a well-formed ML counterpart would have. For example, in example (1) from Spanish-English CS, both medium and large are bare forms—in Spanish, ‘medium’ would be una mediana and ‘large’ would be la grande. That bare EL nouns can occur indicates that the frame that the ML projects is not identical with the frame of the source language of the ML. However, many EL nouns are framed by ML material. For example, in some data sets, an EL noun appears with ML modifiers, as in tus coworkers ‘your coworkers’ in the sentence, Pero tú te refieres à tus coworkers ‘But are you talking about your coworkers? (Spanish-English CS, Jake, Myers-Scotton and Gross 2002, p. 81). There are many similar examples when Arabic is the ML, such as fi el statistics ‘in the statistics’ (Okasha 1999, p. 61). EL nouns sometimes are marked as plural with ML affixes; however, plural is sometimes marked on EL nouns with an EL affix, as is plate in the Swahili-English NP, plate-s tatu (plates three) ‘three plates’ (MyersScotton 1988, p. 23). Plural may also be marked with both an ML affix and an EL one. An example from Acholi-English CS is lu-civilian-s, which shows an Acholi plural determiner and an English plural suffix (see Myers-Scotton 2002, pp. 91-93 on “double morphology”). In order to find a place in CS, it seems that all that is required is that such EL nouns have the lexical-conceptual structure that fits the speaker’s intentions. Presumably, these intentions would not be very different if the

Cost Makes a Difference

39

speaker were using an ML noun; therefore, congruence seems easy to establish at this level of abstract structure. It seems likely that the EL noun is used, not the ML one, because of nuances in the message it carries. Further, one can hypothesize that the need to match congruence with the ML for the other two levels of abstract lexical structure is weak—as long as lexical-conceptual structure matches the speaker’s intentions. Thus, singly-occurring EL nouns or even nouns with EL modifiers “receive a pass” from any filter that checks for congruence across the ML and EL for the levels of predicate-argument structure or morphological realization patterns. At least, the frequency of EL nouns across CS data sets is empirical evidence that the “production cost” of inserting such nouns or noun phrases into an ML frame is minor. The story is somewhat different for compound nouns, probably because hierarchical structure is involved and therefore the levels of predicate-argument structure and morphological realization patterns must be checked for congruence with the levels of the ML. For example, in a corpus of Xhosa-English CS, what would be a compound noun in English (crime rate) can only appear in an associative construction of NP+ associative element + NP, which is how Xhosa, the ML, treats two nouns in a hierarchical relationship. This is the case, even though the nouns themselves may come from English. In this corpus from 48 Xhosa-English bilinguals, the only time an English compound noun appears is when English frames the clause. Compare examples (8a) with business person and (8b) with rate ye crime, both from this corpus. 8a and 8b. Xhosa-English CS (Myers-Scotton 2004 South African corpus) 8a. … Let’s say I am a business person. 8b. … i-rate y-e crime CL9-rate CL9-ASSOC crime ‘… [the] rate of crime is high.’

i-nyuk-ile CL9-go up-PERF

6. How Early System Morphemes Differ Determiners in noun phrases and prepositional satellites in verb phrases both count as early system morphemes because of their semantic features. (Such satellites are called particles by some linguists.) Both add specification to the meaning of their heads. However, satellites are derivational in the sense that they extend or change the meaning of their heads. For this reason, it may be they are entirely salient along with their

40

Chapter Two

heads when the heads (verbs) are selected at the conceptual level. For example, they can change argument structure. Prepositional verb satellites occur quite frequently in the EL with either EL or ML verbs as their heads. Example (9) shows an ML verb (vini ‘come’ from Haiti) and a satellite (from English) and (10) shows an ML verb (cambian ‘change’) with around from English as the satellite. 9. Haitian Creole-English CS (Hebblethwaite 2007, p. 296) … yo djus vini from 3PL just come from ‘…they just came from Haiti.’

Ayiti. H

10. Spanish-English CS (Pfaff 1979, p. 303) Sabes los cambia-n know.2s 3PL.OBJ change-3PL.PRES ‘You know they change them around.’

around around

In contrast, in example (11) both the satellite preposition and the verb come from the EL (find out), although Ewe is the ML. 11. Ewe-English CS (Amuzu 2010, p. 160) E- dze be wó-a-find out first, wó-á-find out be 3s be proper COMP 3PL-FUT-find out 3PL-FUT-find out COMP e- li hã 3S-be present EMPH ‘They have to [it is proper] that they find out first, they find out that he be present.’

Altogether, examples of prepositional verb satellites from the EL are easy to find in CS corpora. EL determiners are quite another story regarding their frequency of occurrence. They hardly occur at all except under two conditions: x EL determiners do occur when they accompany an EL noun as their head, as is the case in many examples in Arabic-French data sets (e.g., les histoires ‘the stories’.) In effect, the combination of EL Det + N is an EL island. Often (but depending on the languages involved) in addition to adding specificity to the head noun, such determiners encode grammatical gender and plural. (Grammatical gender is an inherent feature projected by the head noun involved

Cost Makes a Difference

41

and plural is an indirectly-elected early system morpheme.) Example (12) from Wolof-French CS offers another example of a determiner + noun from the EL (ces mots ‘those words’; but in this case the noun has an ML modifier, too (tubaab ‘foreign’). 12. Wolof-French CS (Swigart 1992 cited in Myers-Scotton and Jake 2001, p. 85) … mën naa la tontu … ci Wolof sans boole ces mots tubaab … able 1S.PERF 2OBJ answer… in Wolof without mix DET.PL words foreign ‘I can answer you … in Wolof without mixing in those foreign words [French].’

x The determiner is part of an EL island that is a set collocation. For example, in (13a) all the clothing can be considered an instance of the collocation all the X. 13a. Swahili-English CS (Myers-Scotton 1993; 1995, p. 80) … ni-ka-wash ‘… and I washed

all all

the the

clothing … clothing’

In addition to collocations which are EL islands, even more likely are collocations which are mixed; that is, they are headed by an ML determiner, as is la real thing in (13b). 13b. French-English CS (King 2000, cited in Myers-Scotton 2006, p. 265) J’étais certain que c’était pas la 1S.be.PST certain COMP it.be.PST NEG DET.F ‘I was certain that it was not the real thing.’

real real

thing thing

Given these patterns, it is not surprising for ML determiners to occur with single EL nouns frequently. For example, in Spanish-English corpora there are many instances of Spanish determiners with English nouns, with Spanish as the ML (e.g., el half chicken). In her corpus of Palestinian Arabic-English CS, Okasha (1999, p. 60) found 110 Arabic determiners and 20 demonstratives with English nouns (e.g., el charity), but no English determiners with Arabic nouns. Such examples make it clear that to switch from the ML for determiners is a costly matter. The reason EL determiners do not occur otherwise in mixed constituents may be that the complexity that they, or ML counterparts, contain in regard to the level of predicate-argument structure, causing

42

Chapter Two

congruence problems. In some languages, determiners are multimorphemic and contain information relevant to argument structure; that is, they contain other morphemes in addition to the early system morpheme in which lexical-conceptual structure is most salient. Multi-morphemic determiners may be subject to the “drag down principle” if they include a morpheme encoding case (Myers-Scotton 2002, p. 305; see also MyersScotton and Jake 2001, pp. 113-114). Case is an outsider late system morpheme; as such it is structurally-assigned at the level of the formulator. Therefore, a determiner, which is multi-morphemic and contains case, is not salient at the level of the mental lexicon when the ML is assembling noun phrases. Such determiners are only salient when structurallyassigned morphemes are activated and it is “too late” to look to the EL for a determiner. Prepositional verb satellites are a much simpler matter. True, their presence can affect the degree to which their head verb is congruent with the ML frame’s requirements for predicate-argument structure. But because the suggestion is that they are already salient at the conceptual level, it is the verb that “absorbs” any changes in argument structure that can be attributed to the satellite. For example, when look occurs on its own, it is intransitive, but when it occurs with certain satellites, as for example, look for, it is transitive, taking a direct object. The empirical evidence is that that mixed constructions of verb + an EL satellite are frequent and NPs with only the determiner from the EL are not.

7. Complementizers vs. Subordinating and Coordinating Conjunctions These lexical categories are similar in that they all join together two clauses. However, they differ in their status as morphemes according to the 4-M model’s classification. The difference is that complementizers qualify as bridges that join together two clauses in a hierarchical relation. Hierarchical structure is a feature of the abstract level of predicateargument structure. Complementizers index the notion that clause 2 has a structural relationship with clause 1. In contrast, conjunctions are content morphemes and the clauses are in a sister relationship. What is most salient about them is their level of lexical-conceptual structure; that is, the meanings they encode. It is true that subordinating conjunctions do carry procedural meaning, as well; as such, they have an important discourse function because they enable the listener/reader to anticipate/predict the pragmatic import of the main clause. For example, in example (14) the subordinating conjunction from French d’ailleurs ‘besides’, has the

Cost Makes a Difference

43

pragmatic message that might be something like this: ‘what follows offers another explanation that reinforces what has previously been said in the discourse.’ 14. Moroccan Arabic-French CS (Bentahila and Davies 1998, p. 37) D’ailleurs, hadi ma tadxul š ƒ s serval. Besides, this not enters not in the trousers ‘Besides, this one does not tuck inside the trousers.’

Both subordinating conjunctions (such as equivalents of because or but) and coordinating conjunctions from the EL are relatively frequent in CS. The fact that they often represent a switch in languages makes their message stand out, a reason why an EL version of the conjunction is employed. Further, it seems that the level of lexical-conceptual structure is most salient in coordinators—no surprisingly given their status as content morphemes. In contrast, complementizers often come only from the ML in some data sets. This is the case in very diverse corpora, everything from ArabicEnglish to Chicheǒa-English. Still, in other data sets, such as SpanishEnglish, there are examples of the complementizer coming from the EL (English), but these are found only occasionally and seem to be more likely if the second clause is in the EL. However, four examples from Spanish-English are cited in Myers-Scotton and Jake (2009, p. 352), showing that Spanish que ‘that’ can introduce either an English clause or a Spanish clause if it seems to be the dominant ML in the discourse (see example 15a). Example (15b) comes from a language pair in which the complementizer is always from the ML, Arabic. Note that such complementizers as ?innu in Arabic take person and number inflections. This difference in frequency of EL coordinators (quite frequent) and EL complementizers (not frequent or not at all) implies that the difference in morpheme type (content morpheme or bridge morpheme) makes quite difference. EL complementizers may be more costly, just because there is more grammatical structure that has to be check (i.e., the level of predicate-argument structure is more salient in complementizers. 15a. Spanish-English CS (Paff 1979, p. 315, cited in Myers-Scotton and Jake 2009, p. 352) Si, but the thing is que empiezan bien recio and … Yes, but the thing is COMP start.3PL very fast and … ‘Yes, but the thing is that they start pretty fast and …’

Chapter Two

44

15b. Palestinian Arabic-English (Okasha 1999, p. 71, cited in MyersScotton and Jake 2009, p. 353) Kaan el-doctor yisuk ?innu PERF.3MASC.be the doctor IMPERF.3Mdoubt COMP.3.MASC is not very reliable is not very reliable ‘[he]was, the doctor, doubting that it was not reliable.’

8. EL Verbs in Finite Frames Finally, this paper considers the question, what makes CS involving EL verbs in clauses framed by the ML so frequent and seemingly effortless? Relevant examples, from those data sets that accept this pattern, include two from Swahili-English, a-ki-ni-buy-i-a beer ‘he was buying beer for me’, which occurs in example (2) above and tu-li-po-watch ‘when we were watching’ in (16). In (17) there is an example from Ewe-English, dona-a, ‘lend habitually’ and in (18) English ‘clean’ receives Finnish suffixes. 16. Swahili-English CS (Myers-Scotton 1988 Nairobi corpus) … a-li-vyo kuwa karibu ku-fainti wakati 3S-PST-ADVcopula near INFIN-faint time tu-li-po-watch fili y-a Dracula … 1PL-PST-CONDIT-watch CL9-ASSOC ‘How she was near to fainting when we watched [the] film of Dracula.’ 17. Ewe-English CS (Amuzu 2010, p. 165) Nyonu-a lend-na ga na vi-a Woman-DEF lend-HAB money to child-DEF ‘The woman regularly lends money to her child.’ 18. Finnish-English CS (Halmari 1997, p. 113) Mina clean+as+i+n 1s clean+VM+PAST+IS ‘I cleaned the building.’

building+in building-ACC

Cost Makes a Difference

45

8. 1 Explaining EL Verbs in ML Constructions The puzzle is the robustness of EL verbs in light of the expected production cost in checking congruence between EL verbs and their ML counterparts. Presumably, the EL verb is selected because its semantic/pragmatic features, though similar to those of an ML counterpart, better satisfy the speaker’s intentions. Establishing congruence at the level of lexical-conceptual structure would not be costly. However, verbs would surely have to be checked across the levels of predicate-argument structure and morphological realization patterns. In morphologically-rich languages, such as Swahili and Finnish, agreement (co-indexing of subjects and objects) is also involved. Thus, the ease with which EL verbs appear in these ML frames—ease characterized by what analysts perceive as speed without hesitations--motivates a hypothesis that avoids a scenario involving checking EL finite verbs. Non-finite EL verbs offer a less costly scenario. Non-finite verbs differ from finite verbs in that they have little or nothing to do with argument structure and agreement; that is, the abstract level of predicateargument structure is not salient in their makeup. Thus, an EL non-finite verb only needs to be checked for congruence with ML counterparts at the level of lexical-conceptual structure.

8. 2 Inflected EL Non-finite Verbs In addition to low production costs, three pieces of evidence motivate the hypothesis that those EL verbs appearing in a ML finite construction are non-finite. First, the phonological shape of both finite verbs and infinitives in English can be identical and mislead the analyst. Even though English infinitives often appear with the infinitival marker ‘to’ in monolingual English, they also appear without it. For example, consider the English verb ‘go’ in let him go. Second, independent evidence when other languages serve as the EL offers even more compelling evidence that even though English non-finite verbs have the same surface form as English finite verbs, those appearing in CS in an ML finite construction are non-finite. Consider examples when French is the EL. Examples (19) and (20) show clearly that the French verb that receives ML inflections is an infinitive, so marked by the suffix of its verb class (apparten-ir in (19) and –comprend-re in (20). In (20), comprendre does seem to function as an infinitive; it receives the Lingala prefix that marks infinitives (ko-), but the French verb also occurs with its infinitive class marker, the –re suffix.

46

Chapter Two 19. Shaba Swahili-French CS (De Rooij 1996, p. 186) Donc, (h)ii richesse y-ote (h)I i-na-tu-appartenir shi Therefore, CL9.dem CL9-all CL9.dem CL9-NONPAST-OBJ.1PL-belong.INF us ba-toto y-ake CL2-child CL9-his ‘Therefore, all these riches, it belongs to us, his children.’ 20. Lingala-French CS (Bokamba 1988, cited in Myers-Scotton, 1993, 1995, pp. 110-111) l-heure ya kala trois quarts ya ba-jeune-s DET-hour of past three-quarters ASSOC CL2-youth-PL baza-ko-comprendre avenir … CL2.NEG-INF-understand future … ‘In the past, three quarters of the young people did not understand [their] future…’

Third, a similar piece of evidence comes from Acholi-English CS, where another English non-finite verb, the present participle receives verbal inflections. See example (21). A reason for employing the present participle and not the infinitive may be that Aspect is salient in monolingual Acholi verb phrases; the English present participle encodes Aspect, but the English infinitive does not. 21. Acholi-English CS (Myers-Scotton 2005b, p. 12) Chances me accident pol Chances of accident many ka i-boarding taxi if 2S-boading taxi '[the] chances of [an] accident [are] many if you board [a] taxi'

8. 3 Accessing Non-finite EL Verbs in Another Way Not all CS corpora seem to accept EL non-finite verbs in constructions with ML inflections. For example, in corpora with Arabic as the ML and English as the EL, English verbs do not occur in these constructions. However, in many data sets, another strategy that still employs EL infinitives is very frequent; it may be even more frequent than EL infinitives with ML VP inflections—although, again, it does not seem to be available in all corpora. This strategy is the “do” verb construction. That is, the ML verb meaning ‘do’ receives ML finite construction

Cost Makes a Difference

47

inflections. Inflecting this ML verb, which has little meaning of its own, avoids the issue of checking congruence entirely. Also, because the role of the EL infinitive is only to convey the speaker’s semantic/pragmatic intentions regarding the verb phrase, checking congruence at the levels of predicate-argument structure and morphological realization patterns is avoided. Such “do verb constructions” are found in CS corpora all over the world, from Asia to Africa to Europe. Example (3) above illustrates such a construction from Turkish-Dutch CS (with the Dutch infinitive kijk-en ‘watch-infinitive marker’) preceding the Turkish ‘do’ verb yap-iyor ‘doprogressive-3s’. (Turkish, of course, is a verb-final language.) Example (22) shows such a construction from Chicheǒa-English CS with the ML verb –chita ‘do’ and the EL infinitive destroy. 22. Chicheǒa-English CS (Simango 1996 corpus) Atata, ta-bwera-ni ….nd-a-chi-chit-a destroy Daddy, IMPER-come-HON … 1S-PERF-CL7.OBJ-do-FV destroy ‘Daddy, come … I have destroyed it [a class 7 object].’

9. Conclusion The main goal in this chapter has been to develop an explanation for why some EL elements occur in codeswitching corpora and others do not. The explanation has to do with differences in possible production “costs” when different types of EL morphemes are accessed. The production costs are linked to the nature of congruence checks that happen when EL elements appear in a mixed constituent in a bilingual clause. Under the Abstract Level model, three levels of abstract lexical structure may require checking for congruence between an EL element and its ML counterpart, or with the ML frame for the bilingual clause. Not only does congruence between the ML and EL matter; also, the Uniform Structure Principle prejudices the ML as the source of grammatical structure in the bilingual clause. That is, asymmetry in favour of the ML is expected. Both of these factors may be the source of some production costs attached to EL elements at the outset. However, in some contact phenomena, such as CS, some levels of an EL morpheme’s abstract lexical structure seem to be open to modification when in contact with the grammatical frame of another language (i.e. the ML). That is, some splitting and recombining of levels in the EL morpheme can happen in the interaction between the EL and ML frame. Or, the salience of some levels in EL morphemes may not be relevant

48

Chapter Two

when CS is the medium. This may explain the robustness of content morphemes and prepositional verb satellites in CS. In most detail, this chapter suggests a hypothesis that EL verbs that appear in ML frames, specifically those that receive ML verbal inflections, are non-finite verbs, not finite verbs. The only level of abstract structure that is inherently salient in non-finite verbs is that of lexical-conceptual structure (i.e. semantics/pragmatics). Thus, the cost of congruence checking between the EL verb and an ML counterpart (or the ML frame itself) is reduced to checking the EL verb against the speaker’s intentions regarding meaning. Another possibility is that the ML itself accommodates non-finite EL verbs in its frame. This provides an explanation for the robustness of these EL verbs. A hypothesis is that the ML frame itself supplies all the relevant information about the other two levels of abstract lexical structure, predicate-argument structure and morphological realization patterns; remember that these levels are missing in the nonfinite verb. Thus, production costs for EL non-finite verbs may be minimal. In sum, this paper has suggested some ways to explain why some EL morphemes are frequent in ML frames and others are not. It has related this frequency to the production cost that different morpheme types may incur. In turn, this cost is related to the particular makeup of a morpheme type at abstract levels of lexical structure.

References and Further Reading Amuzu, E. 2010. Composite codeswitching in West Africa: The case of Ewe-English codeswitching. Saarbrucken: Lambert Academic Publishing. Backus, Ad. 1996. Two in one: Bilingual speech of Turkish immigrants I in the Netherlands. Tilburg: The Netherlands. Bentahila, A and Eirlys D. 1998. Codeswitching: An unequal partnership? In: R. Jacobson (ed.), Codeswitching worldwide, pp. 25-50. Amsterdam: Benjamins. Bokamba, E. 1988. Code-mixing, language variation, and linguisitic theory: Evidence from Bantu languages. Lingua 76: pp. 21-62. DeRooij, V. 1996. Cohesion through contrast, discourse structure in Shaba Swahili/French conversations. Amsterdam: IFOT (Institute for Functional Research into Language and Language Use). Dussias, P. E., R. E. Guzzardo Tamargo, J. R. Valdes-Kroff, and C.Gerfen. In Press. Looking into comprehension of Spanish-English code-switched sentences: Evidence from eye movements. To appear in F-H Liu and J. Huang (eds.), Volume dedicated to Rudolph Troike.

Cost Makes a Difference

49

Halmari, H. 1997. Government and codeswitching, explaining American Finnish. Amsterdam: Benjamins. Hebblethwaite, B. 2007. Intra-sentential code-switching among Miami Haitian Creole-English bilinguals. PhD dissertation (unpublished). Bloomington, Indiana: Department of French and Italian, Indiana University. Jake, J., Myers-Scotton, C., and Gross, S. 2002. Making a minimalist approach to codeswitching work: Adding the Matrix Language. Bilingualism, Language and Cognition 5: pp. 69-91. —. 2009. Which language? Participation potentials across lexical categories in codeswitching. In: L. Isurin, D. Winford, and K. de Bott (eds.), Interdisciplinary approaches to code-switching. Amsterdam: Benjamins. pp. 206-240. King, R. 2000. The lexical basis of grammatical borrowing: A Prince Edward Island French case study. Amsterdam: Benjamins. Kroll, J. and de Groot, A. (eds.). 2005. Handbook of bilingualism, psycholinguistic approaches. New York: Oxford: Oxford University Press. Kroll, J., Bogulski C.A., and McClain, R. In Press. Psycholinguistic perspectives on second language learning and bilingualism: the course and consequences of cross-language competition. Linguistic Approches to Bilingualism. Levelt, Willem J.M. 1989. Speaking, from intention to articulation. Cambridge, Massachusetts: MIT Press. Myers-Scotton, C. 1988. Nairobi corpus, Swahili-English (unpublished) —. 1992. Comparing codeswitching and borrowing. Journal of Multilingual and Multicultural Development 13: pp. 19-39. —. 1993. Duelling languages, grammatical structure in codeswitching. Oxford: Oxford University Press (Clarendon Press). 2nd ed., 1995, with new Afterword. pp. 240-259. —. 2002. Contact linguistics, bilingual encounters and grammatical outcomes. Oxford: Oxford University Press. —. 2004. Xhosa-English corpus (unpublished). —. 2005a. Supporting a differential access hypothesis: Code switching and other contact data. In: J. Kroll, and A. de Groot (eds.), Handbook of bilingualism, psycholinguistic approaches. New York: Oxford University Press. pp. 326-349. —. 2005b. Embedded Language elements in Acholi/English codeswitching: What’s going on? Language Matters 36: pp. 3-18. —. 2006. Multiple voices, an introduction to bilingualism. Walden, Massachusetts: Blackwell.

50

Chapter Two

Myers-Scotton, C., and Jake, J. 1995. Matching lemmas in a bilingual competence and production model: Evidence from intrasentential code switching. Linguistics 33: pp. 981-1024. —. 2000. Four types of morpheme: Evidence from aphasia, code switching, and second language acquisition. Linguistics 38: pp. 10531100. —. 2001. Explaining aspects of code-switching and their implications. In: J. Nicol (ed.), One mind, two language, bilingual language processing. Malden, Massachusetts: Blackwell Publishers. pp. 84-116. —. 2009. A universal model of code-switching and bilingual processing and production. In: B. Bullock, and A. J. Toribio (eds.), Cambridge handbook of linguistic code-switching. Cambridge, UK: Cambridge University Press. pp. 336-357. Okasha, M. 1999. Structural constraints on Arabic-English codeswitching: Two generations. PhD dissertation (unpublished). Columbia, South Carolina: University of South Carolina Linguistics Program. Pfaff, C. 1979. Constraints on language mixing: Intrasentential codeswitching and borrowing in Spanish/English. Language: pp. 291-318. Simango, R. 1996. Chichewa-English corpus (unpublished). Swigart, L. 2002. Two codes or one? The insider’s view and the description of codeswitching in Dakar. Journal of Multilingual and Multicultural Development 13: pp. 83-102.

Exercises 1. Myers-Scotton claims that in bilingual data, there is always an asymmetry between the roles of the languages involved. She specifically refers to codeswitching as an example of bilingual data. Can you identify two asymmetries in codeswitching data? One involves the source of grammatical structure in the clause that shows codeswitching. The other asymmetry has to do with morpheme types. 2. Can you identify an asymmetry that exists when speakers of one language take in words from another language? For example, what types of words (that is, what lexical categories) are generally taken in when such borrowing occurs? 3. When bilingual speakers engage in codeswitching, they are using elements from two or more languages. But is the result a "mixing" of the languages--in the sense that there is no principled way to tell which language supplies which elements? Myers-Scotton would

Cost Makes a Difference

51

say "no, there is not mixing". What is the evidence that there is no "mixing"? 4. Do you think that allowing some codeswitching to go on in the classroom when the subject being studied is a language (switching between the speakers' first language and the language that is being studied in the class) interferes with the students' attempts to learn the target language? Or, is allowing some codeswitching useful in any ways?

CHAPTER THREE ANALYSING MOROCCAN ARABIC AND ENGLISH DIMINUTIVES IN CODESWITCHED DISCOURSE NAJAT BENCHIBA-SAVENIUS

1. Introduction Diminutives in Moroccan Arabic are used quite freely in natural discourse however, they are rarely referred to in linguistic literature and this is surprising given the frequent occurrence in both monolingual and bilingual discourse.1 This paper then seeks to both highlight and present the use of diminutives in natural parsing among Moroccan Arabic speakers. The diminutive, which is a slight variation of the original root form, in essence cites the smallness of the object in terms of its size, or is used as a term of endearment.2 In Moroccan Arabic it is almost always the latter that is often used with both masculine and feminine objects and objects of affection in general. Other MA nouns which are masculine can be modified to the feminine in terms of usage are the commonly used diminutives which are normally terms of endearment, speaking to children or sweetening phrases: 1. Noun Gloss ƫalnjf pig galb heart

Gender MASC MASC

Diminutive ƫlilifa gliliba

Gender FEM FEM

Gloss little piglet little heart of mine

Note also the chameleon-type nouns which are sometimes masculine and sometimes feminine. This shows that the use of diminutives at times, sharpens the distinction and gives absolute grammatical gender assignment. Given the hypocorism, the norm is to add a suffixation to the noun or adjective to render it a diminutive.3 Harrell (2004) notes that a distinguishing characteristic of Moroccan Arabic diminutives is the fact

Analysing Moroccan Arabic and English Diminutives

53

that they are formed by the affixation of the morpheme –i- after the second segment of the base, after an initial cluster of two consonants followed by i (2004: 81).4 Some examples of Moroccan Arabic diminutives follow, commencing with monosyllables with trilateral roots where the diminutive particle –iyye- is inserted between the second and third consonants:5 Base kelb xubz ‫ܒ‬erf

Gloss ‘dog’ ‘bread’ ‘piece’

Diminutive kliyyeb xbiyyez ‫ܒ‬riyyef

Table 3-1: Moroccan Arabic diminutives Moroccan Arabic verbs with middle-weak tri-literal roots usually have the diminutive patterning FwiyyeL as shown in Table 3-2 below: Base bƗb fƗr kƗs jƯb

Gloss ‘door’ ‘rat’ ‘glass’ ‘pocket’

Diminutive bwiyyyeb fwiyyer kwiyyes jwiyyeb

Table 3-2: Further Moroccan Arabic diminutives Base bint rjel ωsel semdž ωƯn

Gloss ‘daughter, girl’ ‘foot, leg’ ‘honey’ ‘sun’ ‘eye’

Diminutive bnita rjila ωsila smidža ωwina

Table 3-3: Moroccan Arabic feminine diminutives Finally, some monosyllabic Moroccan Arabic verbs with trilateral verbs and the vowel e show the diminutive pattern FωiLa as shown in Table 3-3 above. In terms of codeswitching, how do Moroccan Arabic verbs fit in? Can English verbs also adopt the same diminutive forms? The following section analyzes the data collated starting with calques or loanwords in Moroccan Arabic.

Chapter Three

54

2. English Diminutives Given the formation of diminutives as described above, there are many English integrational forms in MA which have been assimilated and rendered ‘diminutive’ by the consonantal shape and form. A few examples elicited are as follows in Table 3-4: Base limuna banana sandala blasa b‫ܒ‬at‫ܒ‬a

Gloss ‘lemon ‘banana’ ‘sandal’ ‘place’ ‘potato’

Diminutive Lwimina Bwinina Snidila Bliyssa b‫ܒ‬iyy‫ܒ‬a

Table 3-4: Further Moroccan Arabic French origin diminutives By extension, English diminutives in a Moroccan Arabic environment can be formed as long as the noun adheres to the MA monosyllabic shape. In terms of codeswitching, insofar as the morpheme does not contravene Myers-Scotton’s MLF or Uniform Structure Principle, diminutive forms are constructed with relative ease and regularity as detailed below: 2. džufi ωla rwibitta ki mazint –ha ta-run fi l garden See PREP rabbit how pretty FEM run 3SG in DEF garden ‘Look at the pretty little rabbit running in the garden’

In (2) above, the noun ‘rabbit’ is given the diminutive Moroccan Arabic pattern in line with the affixation of the morpheme –i- as described above. Crucially, note how the modifying adjective mazint-ha is given the feminine suffix –ha in line with the matrix language principle where the embedded forms must align themselves with the grammar of the matrix frame; in this case, the noun ‘rabbit’ is rendered feminine and therefore, the adjective must also be made feminine in line with Moroccan Arabic grammar.6 3. kan lƗbes waƫΩd al jwikitta brown ta-ƫemeq mωa Was wearing one DEF jacket DEM brown go mad 3SG with blue jeans ‘He was wearing a little brown jacket with blue jeans’

The use of diminutives in Moroccan Arabic and English codeswitched discourse further validates the MLF as even derived English forms strictly

Analysing Moroccan Arabic and English Diminutives

55

adhere to the MLF where the embedded morpheme tallies with the grammar provided by the syntactic frame of the matrix language. This is where the ML is the language projecting the morpho-syntactic frame for the entire CP which shows intra-sentential CS. The above example shows how the Moroccan Arabic syntax which is the ML is structured so that the Noun + Adjective construction is adhered to as opposed to English syntax which is strictly Adjective + Noun. Not only are MA nouns used in the diminutive form but also adjectives of colour and defect.7 These are regularly used in the diminutive form and the patterning is that of FωiωeL showing a repetition of the second root consonant (Harrell 2004, p. 82). Some examples are as follows: Base bkem ƫmer bxƯl kbir mliƫ ‫ܒ‬wil kƫel

Gloss ‘mute, dumb’ ‘red’ ‘miserly’ ‘big’ ‘good, excellent’ ‘tall’ ‘black’

MA Diminutive bkikem ƫmimer bxixel kbiber Mlileƫ ğwiwel kƫiƫel

Table 3-5: Further Moroccan Arabic diminutives (second root consonantal repetition) The interesting point about Moroccan Arabic diminutives is the way in which bilingual speakers attach suffixes to the diminutive stem or make either modifications to the diminutive to make it sound even more affectionate or more endearing. The below are some examples of modified MA diminutive forms: Base Rajul ƪuta Fouad Najat Gelb Bussa

Gloss ‘man’ ‘fish’ ‘Fouad’ ‘Najat’ ‘heart’ ‘kiss’

Diminutive A rwijjul ƫwita Fiw Nwijit gliyyeb bwissa

Diminutive B rwiwij ƫwittita Fwiwiw Nwijitta glilib bwiwissa

Table 3-6: Modified Moroccan Arabic diminutives MA suffixation acts in the normal manner where possessives for example can be added as in:

56

Chapter Three 4. Masculine suffixation to MA diminutive B stem: huwwa glilib dyalli w na-mut ωli -h He is heart FEMIN POSS and die 1Sg on him ‘He is my little love and I am crazy about him’ 5. Feminine suffixation to MA diminutive B stem: Hiya gliliba8 dyalli w na-mut ωli –ha She is heart FEMIN POSS and die 1SG on her ‘She is my little love and I am crazy about her’

As has been shown, MA diminutives as well as the modified English MA diminutive are derived from a great number of nouns and adjectives with common distinguishing characteristics. Further, use of such diminutives is also a discourse strategy as they streamline grammatical gender assignment as the subsequent adjectival agreement must adhere to the gender of the initial diminutive and this, in bilingual discourse, is agreed upon by both speakers if the first speaker refers to a noun in the feminine by adding the feminine suffix, the second speaker will not then refer to it in the masculine form. This is a process I term ‘semantic levelling’, where both speakers agree upon the grammatical gender of a noun.

3. Concluding Remarks The use of diminutives in codeswitched discourse transcends generational groups and is not specific to one generation. Rather it can be concluded that all three generations make good use of Moroccan Arabic diminutives as well as the hybrid English forms. Given the data presented, it is evident that the Matrix language Frame Model is a suitable vehicle through which to describe the diminutive phenomenon in general monolingual data as well as that of bilingual intra-sentential data where the structural configurations are uniformly aligned with the principles and parameters of the MLF.

Analysing Moroccan Arabic and English Diminutives

57

Notes 1

See also Watson (2006) on the use of diminutives in Yemeni Arabic. Moroccan Arabic diminutives differ from the morphological formation of diminutives in Classical Arabic. 3 The process of affixation is often accompanied by the labialization of the first segment of the base if it happens to be one of the dorsal consonants (k, g, x, etc) or labial consonants (l, b, m). 4 Harrell states that there are several different classes of stem types, with a number of irregularities and the exact meaning is not uniformly predictable from the meaning of the derivational base (2004, p. 81). 5 It can be said that in disyllabic and tri-syllabic forms, the same number of syllables surface in the diminutive forms. However, a full and circumscriptive analysis of Moroccan Arabic diminutives is beyond the scope of this thesis due to space and time constraints. cf. Al Ghadi (1990) gives a full account of the phonological and morphological processes involved in Arabic diminutives. 6 Most diminutive nouns are made feminine to render them more affectionate and endearing and this is clearly the case with rwibitta. There are numerous cases where baby boys are referred to using feminine diminutive forms to optimize the endearment. 7 cf. Harrell 2004 for more on adjectives of defect. 8 The feminine suffixation can also be glilibti thereby excluding the requirement for the possessive marker ‘dyal’. 2

References and Further Reading Auer, J, C. P. 1998. Code Switching in Conversation: Language, Interaction and Identity. London: Routledge. Backus A. 1996b. Two in one: bilingual speech of Turkish immigrants in the Netherlands. Doctoral dissertation. Katholieke Universiteit Brabant, Tilburg, Studies in Multilingualism 1. Tilburg University Press. Benchiba, N. 2008. A Structural Analysis of Moroccan Arabic and English intra-sentential code switching. PhD Thesis: University of London Bentahila, A., and Davies, E.E. 1983. The Syntax of Arabic-French codeswitching. Lingua 59, pp. 301-330. Boumans, L. 1998. The Syntax of Codeswitching: Analysing Morccan Arabic/Dutch Conversation. Tilburg: Tilburh University Press. DiScullo A.M., Muysken P., and Singh R. 1986. Government and codeMixing. Journal of Linguistics: 22, pp. 1-24. Gardner-Chloros, P. 1991. Language Selection and Switching in Strasbourg. Oxford: Clarendon Press. Harrell, R.S. 2004. A Short Reference Grammar of Moroccan Arabic. Georgetown University Press.

58

Chapter Three

Heath, J. 1989. From Code-Switching to Borrowing: Foreign and Diglossic Mixing in Moroccan Arabic. London: Kegan Paul International. Joshi, A. 1985. Processing of Sentences with Intrasentential Code Switching. D. R. Dowty, L. Karttunen and A. Zwicky (eds.), Natural Language Parsing, pp. 190-205. Cambridge: Cambridge University Press. Myers-Scotton, C. 1983. The Negation of Identities in Conversation: A Theory of Markedness and Code Choice. International Journal of the Sociology of Language, 44: pp. 115-36 —. 1992. Comparing Code-Switching and Borrowing. Journal of Multilingual and Multicultural Development, 13 (1&2), pp. 19-39. —. 1993a. Social Motivations for Code-Switching: Evidence from Africa. Oxford: Clarendon Press. —. 1993a. Duelling languages : grammatical structures in codeswitching. Clarendon Press, Oxford —. 2002. Contact Linguistics. Bilingual Encounters and Grammatical Outcomes. Oxford Nortier, J. 1990. Dutch-Moroccan Arabic Codeswitching. Dordrecht: Foris. Poplack, S. 1980. Sometimes I’ll Start a Sentence in English Y Termino en Español: Towards a Typology of Codeswitching. Linguistics: International Journal of Human Communication, 18, pp. 582-618. Watson, J. 2006. Arabic morphology: Diminutive verbs and diminutive nouns in San’ani Arabic. Morphology 16: 189-204.

Exercises 1. What is the Matrix Language Frame Model? 2. In grammatically mapping Moroccan Arabic onto English, can you forsee any barriers to codeswitching? 3. Give an example of a Moroccan Arabic diminutive in codeswitched discourse. How is gender mapped?

CHAPTER FOUR THE GRAY AREA OF ACCEPTABILITY JUDGEMENTS: CLEFTS AND EXHAUSTIVITY IN CYPRIOT GREEK1 ELENA PAPADOPOULOU, EVELINA LEIVADA AND NATALIA PAVLOU

1. Introduction Cypriot Greek (henceforth, CG) is the variety of Modern Greek that is spoken on the island of Cyprus. Lacking the status of an official and codified variety, the use of CG co-exists with the sociolinguistically ‘high’ variety of Greek that is spoken in Cyprus; Standard Modern Greek (henceforth, SMG). SMG is the official language and also the variety spoken in mainland Greece and the different values attached to each variety as well as the different registers that facilitate the use of each one of them affect the way they are eventually put in use, since there is a twoway relation between attitudes towards different varieties and language change (Tsiplakou 2004). The simultaneous use of SMG and CG as well as the status of the latter in relation to the former in sociolinguistic terms has been the object of inquiry in many studies (see inter alios, Papapavlou 1998, Papapavlou and Pavlou 1998, Tsiplakou 2004) and certainly the dynamics behind such co-existence are in a position to influence the way speakers of CG adjust their linguistic performance in certain tasks and registers and, in the long run, modify the use of the most marked aspects of their native variety towards the direction of the standard. The co-existence of two2 varieties that are linguistically related results in a complex interplay between certain morphosyntactic properties that belong to the two varieties and, as a result, it also fades away the boundaries between the different varieties and their respective grammars

60

Chapter Four

giving rise to a linguistic continuum. Similar to what is reported in Grohmann et al. (in press), for the acquisition of object clitic placement in different child populations residing in Cyprus, where the results showed a mixed3 placement pattern in both child and adult populations, focus strategies and complementizers in CG elicited mixed results in terms of acceptability judgements (i.e. in terms of semantic acceptability qua their status as true or false) in the present experiment. The attested focus constructions and complementizers are CG specific and as such they are not found in the performance of a monolingual speaker of SMG. In CG, however, their existence is discussed in the literature (Grohmann et al. 2006, Fotiou 2009, Panagidou 2009) and their presence makes available a range of different cleft sentences, some of which prima facie resemble the typical cleft construction in English ‘it is XP that YP’ and as such, they have been linked to typical properties attributed to clefts cross-linguistically, such as exhaustivity (following Kiss 1998). Under standard assumptions the interpretation of a subject cleft (1) and an object cleft (2) should be exhaustive and presume that only the denotation of XP participates in the YP event. 1. En ton andra pu is.3SG the.ACC man.ACC that ‘It is the man that you saw.’

ides. saw.2SG

2. En o andras pu pezi mappa. is.3SG the.NOM man.ACC that play.2SG football.ACC ‘It is the man who plays football.’

Embu ‘{it-}is-{it-}that’, an element that appears optionally in whquestions and declarative sentences (3)–(4), has received two different syntactic accounts. In interrogative environments, Grohmann et al. (2006) adopt a split-CP analysis with a focus projection FocP whose specifier is filled by the cleft where the matrix clause is the complement of the Chead. The CP-domain remains empty and pu ‘that’ introduces the matrix clause in declarative contexts. As argued by Papadopoulou (in progress), this analysis becomes problematic when we consider that embu cannot inflect for tense (*itabu ‘was-(it-)that’) or be negated (*ennembu ‘not-is(it-)that’) in wh-questions, even though the copula in cleft sentences can. Papadopoulou suggests that has been grammaticalized as a fossilized focus element merged directly in C0. The different properties of clefts and embu-structures are presented in more detail in subsections 2.1 and 2.2 respectively.

The Gray Area of Acceptability Judgements

61

3. O Yannis embu eklotsisen tin mappan. the.NOM John embu kick.3SG the.ACC ball.ACC ‘It is John that kicked the ball.’ 4. Tin mappan embu eklotsisen o Yannis. the.ACC ball.ACC embu kick.3SG the.ACC John ‘It is the ball that John kicked.’

A recent account on the semantic properties of CG clefts suggests that the clefted XP is not always linked to an exhaustive interpretation (Panagidou 2009: 18), following similar claims made by Doetjes et al. (2004) and Prince (1978) for French and English. Following Prince’s terminology, Panagidou provides examples of “informative-presupposition clefts” that intend to present statements as facts without an exhaustive interpretation. However, this absence of exhaustivity is not an argument to be linked to the inexistence of clefts proper in CG, because Panagidou’s examples of non-exhaustive clefts involve PP rather than DP as the clefted constituent. Yet, these PPs denote properties of entities and not entities in the discourse world; therefore the non-exhaustive interpretation in these clefts can be explained away by the syntactic nature of the clefted XP constituent. Alongside with exploring the different syntactic and semantic properties of the embu-structures and the ‘it is XP that YP’ form of clefts in CG, the two main goals of this chapter are to test the hypothesis that embu is an underlying form of cleft and to decipher exhaustivity effects between cleft and embu structures. 187 participants were asked to judge whether 12 declarative sentences were true in relation to each of the 6 stories provided (3 = subject, and 3 = object). The task was written and administered through “Facebook writing” with the aim of differentiating possible effects of written language, since CG is mainly used orally as it lacks orthographic codification4 and some of its sounds correspond to letters not existing in the Greek alphabet and therefore are written in different ways by different people. When CG is used in a written form, usually the Latin alphabet is employed together with the simplest phonological adaptation (i.e. in Facebook, text messages or other social networks). The working hypotheses of the experiment give rise to four possible scenarios: i.

If embu is a focus Complementizer, it should allow for nonexhaustive interpretation. In case it unequivocally allows for a nonexhaustive interpretation, then it should be analysed as a

62

ii. iii. iv.

Chapter Four

grammaticalized focus Complementizer in line with what Papadopoulou (in progress) has proposed. If embu is an underlying form of cleft, it should only allow for an exhaustive interpretation. If ‘it is XP that YP’ is bona fide ‘English type’ cleft, it should only allow for exhaustive interpretation. If both ‘it is XP that YP’ and embu allow for non-exhaustive interpretation, then neither of them can be analysed as bona fide cleft.

2. Exhaustivity in Cypriot Greek In this section, we review different structures that according to the claims made so far in the literature should elicit an exhaustive reading in CG, paying special emphasis to the two kinds of syntactic environments that were attested in our experiment in relation to exhaustivity effects.

2. 1 Cleft Structures Clefts were firstly discussed by Jespersen (1927), who notes that the DP in the cleft is so definite that it cannot be further restricted so as to call the ‘that’-clause a relative clause. In later work, Jespersen (1937) analyzes this relative clause as a special kind of “parenthetic clause”, namely a cleft. Semantically, clefts express a single proposition in a bi-clausal structure. Jackendoff (1972) transforms Chomsky’s term of “natural” responses to utterances with meaning and defines them as “focus” and “presupposition”. Even though more specific and recent suggestions have been made in the literature illustrating syntactic explanations for clefts, we will not provide a theoretical analysis of clefts at this stage, but instead focus on their distribution with the exhaustivity condition. Cypriot clefts are argued to be focus structures or existential clauses with a non-restrictive relative clause (for a more detailed discussion see Grohmann et al. 2006, Gryllia and Lekakou 2006, Fotiou 2009, Agouraki 2010). In CG cleft structures, there is fronting of the focused constituent expressing either a person or an object or a concept and embedding of a secondary clause introduced with pu ‘that’, as shown in (5) and (6) below. 5. En ton andra pu ides. is.3SG the.ACC man that saw.2SG ‘It is the man that you saw’

The Gray Area of Acceptability Judgements

63

6. En o andras pu pezi mappa. is.3SG the.NOM man that play.2SG football.ACC ‘It is the man who plays football.’

Clefts have been argued to display exhaustivity properties (Kiss 1998), which means that they exhaust the fronted constituent with the relevant property given by the secondary clause. 7. En to milon pu efaen o andras. is.3SG the apple.ACC that ate.3SG the man.NOM ‘It is the apple that the man ate.’

In this case, the exhaustive interpretation of the cleft requires that the man has eaten only an apple. In the case that the man has also eaten a banana and a peach, then this sentence should be odd for the given context and the clause in (8) should be produced: 8. En to milon, tin mbanana tzie to rodakinon is the.ACC apple, the.ACC banana.ACC and the.ACC peach.ACC pu efaen o andras. that ate.3SG the.NOM man.NOM ‘It is the apple, the banana and the peach that the man ate.’

The exhaustivity property appears to be acquired late by children, who go through a non-exhaustive stage in questions (Seymour et al. 2005, Roeper et al. 2006a, 2006b) and clefts (Heizmann 2007). Following previous assumptions for the exhaustivity expressed by clefts, we have included cleft sentences in our experiment to compare them with the potentially focused particle embu (Papadopoulou, in progress) that is presented further on. There are no previous studies regarding the exhaustive interpretation of clefts by Greek Cypriot adults or children, except simple production of clefts by children in the context of a more complex experiment (Pavlou 2010), so this study is firstly examining whether these assumptions also hold for CG and secondly whether the same effect appearing in clefts, also appears in the embu-particle.

2. 2 Embu CG question and cleft formation bear morphological resemblance to SMG counterparts with minor pragmatic-semantic and morpho-phonological differences (Newton 1972, Arvaniti 2002) as well as substantial formation

64

Chapter Four

differences with respect to the embu ‘is-it-that’ strategy analysed later in this chapter and the different wh-words employed. Embu is mostly found in wh-questions and it optionally appears after the wh-word. CG wh-words involve the quantifiers pcos ‘who/which’, posos ‘how much/many’, ti ‘what’, and inda ‘what’ as well as the adverbs pote ‘when’, pu ‘where’, jati ‘why’, pos5 ‘how’, inda ‘why’, and indalo(i)s ‘how’ (Simeonidis 2006, p. 217; cf. Holton et al. 1997 for SMG). The quantifier inda ‘what’, and the adverbs inda ‘why’ and indalo(i)s ‘how’, are dialect-specific to CG (for a more detailed description see Grohmann and Papadopoulou 2010, 2011, Papadopoulou, in progress) and as discussed in Simeonidis (2006, p. 217), the wh-quantifier inda initially derived from the interrogative pronoun tinda ‘what’ used in Asizes,6 literally ti ine afta ‘what are these’. The theoretical account for the status of the embu-strategy, as Grohmann et al. (2006) refer to, stands in contrast to Papadopoulou’s claim (in progress) for the fossilization of embu in questions. This contrast brings together two opposing views on how embu-questions are formed. According to Grohmann et al. (2006), the embu-strategy involves a splitCP analysis with a focus projection (FocP), whose specifier is filled by the cleft and a C-projection, which takes the matrix clause as its complement deriving questions such as (9) and (10) below. On the other hand, Papadopoulou (in progress) argues for the existence of only the Cposition, arguing for a fossilized embu.7 9. Ti embu șcavazi o andras? what is-it-that read.3SG the.NOM man ‘What is the man reading?’ 10. Pco vivlio embu șcavazi o andras? which book is-it-that read.3SG the.NOM man ‘Which book is the man reading?’

Embu also appears in declarative clauses, where it is argued to express focus: 11. Ton Yannin embu agapa i Maria. the.ACC John embu loves the.NOM Mary ‘It is John that Mary loves.’

The fronted constituent needs to appear with a determiner, expressing contrastive focus and suggesting that embu might be a focus particle. Identical to clefts (Section 2.1), the focus expressed by embu denotes an

The Gray Area of Acceptability Judgements

65

exhaustivity condition, which sets the fronted constituent as the only item that has been eaten by John: 12. To psomin embu efaen o Yannis. the.ACC bread embu ate.3SG the.NOM John ‘It is the bread that John ate.’

We have presented two structures that share the property of focus and as a result they are assumed to also share the property of exhaustivity. Clefts have been cross-linguistically portrayed as exhaustive (see Kiss 1998 and references cited therein) and in the absence of argumentation suggesting otherwise are assumed to be exhaustive in CG as well. In parallel, embu-declaratives, which appear to express focus can be argued to be exhaustive. In our study, we have included both clefts and embudeclaratives in an attempt to compare the two structures based on the same criterion; the exhaustivity condition.

3. The Present Study This section introduces and describes in detail the Cypriot Greek Exhaustive (*Embu) Clefts (CyGEEC) experiment designed to tease apart the interpretative quirks mentioned in section (2.2) above.

3. 1 Participants CyGEEC was administered to 187 monolingual Greek Cypriot adults from all over Cyprus with the aim to be administered and adapted later on to child populations. Participants were divided into three age groups namely 18 – 30 years old for AG1, 30 – 45 years old for AG2 and 45 years old and above for AG3 (Table 4-1). Age group

Age range

participants

AG1

18 – 30

AG2 AG3

Education

Gender M

F

Lyceum

College

University

148

33

115

39

5

104

30 – 45

25

32

68

1

3

21

45 +

14

7

7

10

1

3

Table 4-1: Participants

Chapter Four

66

Most participants across all three age groups were female and have received university education with the majority being 18 – 30 years old, hence not allowing for a ‘valid’ comparison between and within groups for gender and education. Even though number of participants across AGs was not balanced a proportional comparison is provided in section 4.3 of the results.

3. 2 Methodology For the investigation of embu and cleft dis -or association in CG CyGEEC was conducted. The task involved a total of 40 test items and 12 controls, divided in three object (O) and three subject (S) stories (see pages 68 and 69 for a story example), across three pairs of verbs, agents and nouns (Table 4-2). O/S

Pair 1

O

2 3 1

S

2 3

Verb

Noun

krato hold

petasso throw

mappan ball

kasettinan pencil case

pulukkuin teddy bear

kashian box

vrisko find agorazo buy

pino drink troo eat

nero water vazanin aubergine

gala milk peponin watermelon

krasin wine kapira toast

ximos juice portokalin orange

krato hold

petasso throw

mappan ball

kasettinan pencil case

pulukkuin teddy bear

kashian box

vrisko find agorazo buy

pino drink troo eat

nero water vazanin aubergine

gala milk peponin watermelon

krasin wine kapira toast

ximos juice portokalin orange

Table 4-2: Conditions All conditions, embu and cleft exhaustive and non-exhaustive interpretations along with S and O stories and verb-noun (V/N) pairs were distributed within groups and randomised as depicted in Table 4-3.

The Gray Area of Acceptability Judgements

Story

S/O

V/N Pair

Embu / Cleft

1

O

1

E

Items 1

2

C

3

4

T

2

S

3

C+E

1

4

C

3

O

2

C

4

S

1

E+C

5 6

O S

3 2

C E

67

3

C

C

T

1

4C

1

C

C

T

3

2

2C

F

1C

2

5

C

3C 1

T

CT = True control (referring to N3) CF = False control (referring to N4) 1 = embu (non exhaustive interpretation with N1) 2 = embu (non exhaustive interpretation with N1 and N2) 3 = embu exhaustivity (all 3 Ns) 4 = embu (non exhaustive interpretation with N2 and N3) 5 = embu (non exhaustive interpretation with N2)

5

F

C

2C

2C

C

5C C F

1C C F

5

1 5C

2

T

3

5C

5

3

4 C

C

2

F

C

C

C

C

4

C 4C

3

3

C

C

T

4

5 C

4

F

1c = cleft (non exhaustive interpretation with N1) 2c = cleft (non exhaustive interpretation with N1 and N2) 3c = cleft exhaustivity (all 3 Ns) 4c = cleft (non exhaustive interpretation with N2 and N3) 5c = cleft (non exhaustive interpretation with N2)

Table 4-3: Randomisation The six stories created involved either embu only structures and/or (only) cleft structures as those in (5) and (6) respectively (repeated as (13) and (14) below) resulting in some stories having 5 test items and 2 controls and the others having 10 test items and 2 controls. 13. En ton andra pu ides. is.3SG the.ACC man that saw.2SG ‘It is the man that you saw’ 14. En o andras pu pezi mappa. is.3SG the.NOM man that play.2SG football.ACC ‘It is the man who plays football.’

The stories were presented to each participant in the exact order presented in Table 4-3. The task was administered online through Survey Monkey (http://www.surveymonkey.com), a research tool for creating online surveys. The instructions given at the beginning of the task were in

68

Chapter Four

CG. Investigating preference rather than only grammar in the interpretation of the attested structures, we were interested in the spontaneous reaction to the story; hence participants were not allowed to change their answer once answering a question. This was considered necessary since following items could have triggered a possibly different answer to the items already presented. With respect to the representation of the stimuli, the Latin alphabet was used and the simplest phonological adaptation was applied across test structures. For instance ‘ball’ in CG was written as mappa rather than ȝȐʌʌĮ. Also ‘teddy bear’ was written as pulukkuin rather than ʌȠȣȜȠȣțȠȪȚȞ which if it was to be written orthographically using the Latin alphabet, should have been pouloukkouin. In order to establish this Facebook ‘norm’ we have asked 4 persons aged 21 – 28 to write one of the stories as if they were writing in Facebook. Such a norm was considered necessary since CG does not have an official unified writing coding system – while there has been a recent attempt by Papadima et al. (2011) to create one – and in order to avoid possible effects of SMG writing on CG (see discussion in section 1 for SMG influence on CG and the linguistic environment of Cyprus). An example of a story as presented to participants translated into English can be found below (the translation is in bold type). This is the first story provided with embu. Lena had a ball, a pencil case, a teddy bear and a box. She threw the ball in the dustbin. She threw the pencil case in the dustbin. She threw the teddy bear in the dustbin. I Lena epiaen mian mappan, mian kasettinan, enan pulukkuin tze mian kashian. Epetaksen mes ton kalathon tin mappan. Epetaksen mes ton kalathon tin kasettinan. Epetaksen mes ton kalathon tze to pulukkuin. 1. The ball (embu) Lena threw in the dustbin. Tin mappan embu epetaksen mes ton kalathon i Lena. 2.

The ball and the pencil case (embu) Lena threw in the dustbin. Tin mappan tze tin kasetinan embu epetaksen mes ton kalathon i Lena.

3.

The teddy bear Lena threw in the dustbin. To pulukkuin epetaksen mes ton kalathon i Lena.

4.

The ball, the pencil case and the teddy bear (embu) Lena threw in the dustbin. Tin mappan, tin kasettinan tze to pulukkuin embu epetaksen mes ton kalathon i Lena.

The Gray Area of Acceptability Judgements

69

5.

The pencil case and the teddy bear (embu) Lena threw in the dustbin. Tin kasettinan tze to pulukkuin embu epetaksen mes ton kalathon i Lena.

6.

The box Lena threw in the dustbin. Tin kashian epetaksen mes ton kalathon i Lena.

7.

The pencil case (embu) Lena threw in the dustbin. Tin kasettinan embu epetaksen mes ton kalathon i Lena.

Participants were given the story (first two lines) and then each numbered option, one at a time. They were not allowed to trace back and change any answers facilitating in this way spontaneous response to each item rather than comparison between all possible conditions.

4. Results This section provides a description and analysis of the results in three age groups. All scores correspond to “Correct according to the story” answers given by the participants, divided in the three age groups, namely AG1 18-30, AG2 30-45 and AG3 45 and above. Since not all stories (see Table 4-3 above) had the same number and type -cleft or embu- of items the number of the story is always provided and the results are presented not in the exact order of presentation during the experiment but rather in pairs depending on the type of items. Precisely, Story 1 is compared to Story 6, Story 2 compared to Story 4 and Story 3 to Story 5. Following Figure 1 and the options available, within story items 1-7 were re-coded depending on the number of nouns involved in the action. In particular, when one noun was involved the item was renamed as embu 1 or cleft 1, when two nouns were mentioned embu 2 or cleft 2 and when three nouns were used they were renamed as embu 3 and cleft 3. All participants answered all questions since they were not allowed to continue unless they answered whether the sentence was true (option 1), false (option2), or don’t know (option 3) according to the story. Astonishing seems the fact that all participants went either for the true or false option. Overall, in all stories all participants accepted embu 3 and cleft 3 to be the most ‘correct’ answer according to the story allowing for an exhaustive interpretation and accepted much less the embu 1 and 2 as well as the cleft 1 and 2 (Figure 4-1 and for a different presentation see Figure 4-2). Participants aged 45 and above accepted less all possibilities except the exhaustive interpretations.

70

Chapter Four

Figure 4-1: Overall acceptance in all stories

Figure 4-2: Overall acceptance in all stories (Alterntive presentation).

Once separating embu 1 and embu 2 in different stories, results become more apparent. Acceptance rates for both embu 1 and embu 2 are higher for story 1 for all AGs (Figures 4-3 and 4-4) but the younger the group the higher the acceptance, with AG1 and AG2 accepting at a higher rate potentially non-exhaustive interpretations. AG3 (45 +) are much more conservative throughout both embu 1 and embu 2 with only an interesting high acceptance of the second appearance of embu 1 in story 2 (Figure 43). First appearance of embu 1 and embu 2 in story 1 are much more

The Gray Area of Acceptability Judgements

71

highly accepted than any other but weakens when clefts are introduced in story 2. That is, participants are much more eager to accept a nonexhaustive interpretation for embu if there are no clefts in the story.

Figure 4-3: Embu 1

Figure 4-4: Embu 2

Interestingly, acceptance rates for cleft 1 and cleft 2 do not exceed 39% in all cases for all AGs in contrast to embu 1 and embu 2 possibly projecting a potential difference between embu and cleft structures (Figures 4-5 and 4-6). AG3 is much more conservative across all stories

72

Chapter Four

whereas AG1 and AG2 could easily be fused into one AG since they function very alike with AG2 being more open to acceptance.

Figure 4-5: Cleft 1

Figure 4-6: Cleft 2

Within story, embu analysis (Figures 4-7 and 4-8) confirms the observations made for Figures 4-3 and 4-4 above, demonstrating a high acceptance of the first appearance of embu 1 in story 1, while it is reduced almost by 50% in story 6 – which is the last story of the experiment. Acceptance rates for first appearance of embu 1 and embu 2 are above and

The Gray Area of Acceptability Judgements

73

near 50-60 % in story 1, but less than 30% in story 6. AG 3 (45 +) is again the most conservative in relation to the other two group’s acceptances at a higher rate in embu 1 and 2.

Figure 4-7: Story 1

Figure 4-8: Story 6

A totally different pattern is observed for within story cleft analysis. Clefts are less accepted as “true according to the story” by all AGs when compared to embu items. Strikingly, AG3 nearly accepts cleft 1 and 2 with the highest acceptance rate being no more than 12 % (Figures 4-9 and 4-

74

Chapter Four

10). Cleft 3 in story 5 gets a very low rate of acceptance since the wrong noun was used as N3 hence the item was false indicating that participants did not develop a strategy of accepting any item with three nouns but rather indeed concentrate on the interpretation provided.

Figure 4-9: Story 3

Figure 4-10: Story 5

The Gray Area of Acceptability Judgements

75

In the stories with both embu and cleft structures the results did not deviate from the previous observations with clefts 1 and 2 as well as embu 1 and 2 being accepted less than cleft 3 and embu 3 (Figures 4-11 and 412).

Figure 4-11: Story 2

Figure 4-12: Story 4

76

Chapter Four

As it can be inferred from the results above, AG1 and AG2 are less ‘conservative’ than AG3 possibly suggesting a language change. Both embu and cleft structures were most probably bona fide clefts in the past (see Grohmann et al. 2006, Pavlou 2010), but now they seem to allow for different interpretations in the specific experiment (see section below for further discussion), hence supporting a complementiser analysis for embu (Papadopoulou, in progress).

4. Discussion Teasing apart the relation of the above findings to the arguments for the existence of bona fide clefts in CG, the interpretative quirks in the responses elicited by the test structures point out three things: First, there is great microdialectal variation across speakers of CG for both embu and cleft structures. All test structures across stories and conditions elicited mixed responses. This is suggestive of the so-called heterogeneity of the dialect and further supports the argument for the existence of a linguistic continuum in Cyprus rather than a mere coexistence of two varieties, CG, SMG and/or Standard Cypriot Greek. Second, the importance of priming effects of such experiments should be highlighted. The first test structure in the first story elicited quite high percentages of acceptance as being correct according to the story that preceded it. These percentages dropped significantly once the really exhaustive test structure – that is, the one that involved all three items or individuals – was made available. In other words, the performance of the participants in the first story and particularly in the first test structure of the first story is probably the most credible and important one with respect to not attributing an exhaustive interpretation to embu structures, since participants were unaffected at this point and had no knowledge of whether a sentence that made available all three items/individuals (exhaustive) was following. The same accounts for the first appearance of cleft 1 in Story 2. That is, this instance shows the highest acceptance of cleft 1 as being ‘true according to the story’. What is more, apart from the age effect observed, is the within story carry over effect of exhaustive interpretation on any other option following. Only embu 1, following cleft 3 in Story 2 (Figure 4-11) shows the reverse pattern. The effect observed suggests participants’ answers were affected by the fully exhaustive items (cleft 3 and embu 3) provided. Accordingly, any conclusions drawn regarding the acceptability or not of (non)exhaustive cleft and embu-declaratives should rely initially

The Gray Area of Acceptability Judgements

77

on the first appearance of embu 1 and embu 2 in Story 1 and cleft 1 and 2 before item 7 in Story 2. Third, it should be noted that acceptability judgments8 form a gray area and individuals might not be consistent in their responses when asked whether an utterance is acceptable in their native variety or not. In our study, participants belonging to different groups show diverse behaviour across stories. This could receive a sociolinguistic explanation that boils down to the phenomenon of syntactic change in progress, yet this explanation would account only for interspeaker variation, whereas our findings point out to the existence of intraspeaker variation as well. The conceptualization of E(xternal)-language in natural languages lacks precision when viewed in comparison to the one existing in formal languages. If in the case of the latter, the contents of an E-language are unequivocally accepted and defined by stipulation, in the case of the former it is emprically impossible to precisely define the contents of an Elanguage that corresponds to any natural language, so sometimes, it is equally impossible to classify an utterance as unequivocally unacceptable. Assuming that E-language is viewed as a set of well-formed formulas/sentences, it is a relatively uncontroversial fact that an individual (qua Internal-language) is in a position to generate a set of sentences but it is less clear whether s/he is also in a position to generate a distinguished Elanguage. If s/he could, the area of acceptability judgements would not appear to be so dubious and would not elicit different answers across speakers of the same language. Following Chomsky (1955), generative linguists often assume varying degrees of acceptability of judgments and they accordingly classify linguistic categories in non-discrete scales “more often than not” (Fanselow et al. 2006: 1). Different examples in our stories, come with different degrees of acceptance, despite the fact that they correspond to the same syntactic structure, and such gradience illegitimizes a view of E-language as a set of well-formed formulas/sentences. It seems that natural languages lack a notion of well-formedness that corresponds fairly to the one found in formal languages: Chomsky and Lasnik (1993) portray this as an empirical fact in the sense that when we think of Mary as a speaker of English, it is easy to translate this into the part that relates to I-language (i.e. Mary has a mental state L, and L=English), but is much more difficult to flesh out its empirical implications (i.e. L corresponds or gives rise to a set of acceptable expressions which is conventionally called English). The lack of a distinguished E-language behind the label ‘English’ is proven by the lack of consensus with respect to acceptability judgements across speakers of English. At this point, the notion of ‘idiolect’ enters the equation, but

78

Chapter Four

even if our conceptualization of English is narrowed down in reflecting only Mary’s idiolect, the concept of Mary’s distinguished E-language would still lack a solid basis because the idiolect itself is subject to change and might result to elicitation of different acceptability judgments across different occasions, especially if one targets a sentence that is neither wellformed or ill-formed. The standard anecdotal piece of evidence for this that linguists use is their own inability to provide judgements for such sentences: Having heard them so many times, they start looking more and more acceptable (see Snyder 2005 for issues pertinent to this phenomenon of ‘satiation’). The findings of this study add the factor of priming in the equation, since our participants modified their responses according to the stimuli they received: First appearance of embu 1 and embu 2 in story 1 are more accepted as exhaustive than the ones that follow them, even if the structure behind the test item is identical. One of the main points explored in this paper is the validity of judgments provided by native speakers and the extent to which these can be taken into account for the given analysis. As claimed in section 3.3 the implications of the data collected may suggest that language change may be in progress for the true nature of the dialectal element embu in Cypriot Greek. Embu may have started out with a more complex structure, such as a cleft (Grohmann et al. 2006), but the judgments of the native speakers in the CyGEEC suggested that younger populations prefer the interpretation of a simple syntactic illustration of embu, that of a Complementiser (Papadopoulou, in progress). Early sociolinguistic research (Labov 1984) has dealt with the methods for approaching a speaker and the goals of an interview so as to achieve an objective response. This has brought to surface the doubt of objectivity in the collection of data labeled as the ‘observer’s paradox’ or otherwise the ‘experimenter effect’. The CyGEEC experiment was given in the form of a link on Facebook, so there was no contact between the experimenters and the participants. The design of the experiment did not allow the participants to go back to their previous judgments and change them once they had proceeded to the next test question, so there could not be any effect related to their change of their first decisions. It is, however, possible that such a change could take place in the absence of a certain syntactic structure, as seen in the results. As already pointed out appearance of embu 1 and embu 2 in story 1 are much more highly accepted than any other, but this high acceptance fades away when clefts are introduced in story 2. If no cleft sentences were introduced to the experiment, then the results could have been very different from the ones presented in this paper. What is also challenged is the fact that populations like the one participating in CyGEEC and involving a wide

The Gray Area of Acceptability Judgements

79

range of possible parameters (rural vs. urban speech, possible effect from SMG etc.), are difficult to give a clear-cut distinction with regard to their acceptability of structures like embu, which may be the result of an ongoing change. At this point, the suggestion of ongoing language change is the closest scenario to explain the acceptability rates in the data discussed. Given the Cypriot sociolinguistic context as well as previous studies suggesting that Cypriot Greek undergoes syntactic changes or moves towards a linguistic continuum, we believe that several changes in all the levels of the specific variety could appear at this stage. Similarly to this case, Pavlou (2011) argues that the Cypriot-specific wh-word inda mbu ‘what/why’ undergoes a syntactic change, which appears in young populations as inda mbuallomoprhs. More particularly, the wh-words nambu, innambu, tambu and ambu have been identified as allomorphs of the inda mbu, but show restricted syntactic distribution in comparison with inda mbu. In this study, questionnaires were given to elicit acceptability judgments in a 5grade scale and the results indicated a higher preference of inda mbu in older populations than younger populations. This led to the conclusion of a syntactic change in progress on the basis of the acceptability judgements collected. We base this study on the argument that acceptability judgments can suggest language change and in some cases they can possibly depict the distinction of the age groups to suggest change in progress, or otherwise known ‘the apparent-time construct’. But we acknowledge the fact that variation in the different age groups can either indicate change in progress or simply ‘stable variation’ based on other sociolinguistic factors. Attempting a comparison of the performance across the different age groups, a language change seems to be in course. According to the responses elicited, embu in older ages seems to be analyzed as a cleft (in line with Grohmann et al. 2006), whereas in younger groups it seems to be analyzed as a fossilized lexical item (Papadopoulou, in progress). It should be made clear that this claim assumes that clefts in Cypriot Greek show exhaustivity properties and that based on the embu-analysis (Grohmann et al. 2006) presented above which sets embu as a form of cleft, exhaustivity was expressed in both clefts and in embu in older generations. This will only be once a larger number of participants is tested for all groups as well as compared to a younger group of children who are still in the course of acquisition.

80

Chapter Four

5. Conclusion This paper aimed to address the hypothesis that embu ‘(it-)is-(it-)that’ is an underlying form of cleft as well as deciphering exhaustivity effects between cleft and embu-structures in CG. 187 participants were asked to judge, in an online written task, whether 12 declarative sentences were true in relation to each of 6 stories provided. The four attested hypotheses are the following: i.

ii. iii. iv.

If embu is a focus Complementizer, it should allow for a nonexhaustive interpretation. In case it unequivocally allows for a nonexhaustive interpretation then it should be analysed as a grammaticalised focus Complementiser in line with what Papadopoulou (in progress) has proposed. If embu is an underlying form of cleft, it should only allow for an exhaustive interpretation. If ‘it is XP that YP’ is bona fide ‘English type’ cleft, it should only allow for an exhaustive interpretation. If both ‘it is XP that YP’ and embu allow for non-exhaustive interpretation, then neither of them can be analyzed as a bona fide cleft.

Starting off from (iv), both ‘it is XP that YP’ and embu allow for nonexhaustive interpretation, hence neither can be analyzed as a bona fide cleft. This finding relates to (iii); since ‘it is XP that YP’ is not a bona fide ‘English type’ cleft, it does not only allow for an exhaustive interpretation. The same is observed with respect to (ii); relating this observation to (i), embu can be analysed as a grammatical focus complementiser (Papadopoulou, in progress). The claim set out in the current paper focuses on the sociolinguistic assumption of language change, and more specifically change in progress or apparent-time construct. Change follows prototypically a path where some variant in the speech of an older group in a community appears more frequent in the speech of the middle generation and even more in the youngest generation. Following the Labovian (1984) model, it is expected that the analyzed results should be similar to the prototypical S-curve line for language change. An S-curve is assumed to illustrate the three different stages met in an assumed change of a linguistic feature (morphology, syntax or phonology) by identifying the initial stasis, the rapid rise and tailing off of the phenomenon. Even though, the S-curve is often discussed in quantitative studies, it can be rarely found as it needs to capture all

The Gray Area of Acceptability Judgements

81

different stages based on the different age groups studied. For our study, the results are summarized in Figure 4-13 below:

Figure 4-13: Change in Progress

Figure 4-13 shows the mean of the percentages for each age group regarding the two different syntactic structures, that of embu-clauses and that of clefts. As indicated, an S-curve does not appear, but based on the low performance and percentages observed, it could be argued that this is only the beginning of the assumed change in progress. In fact, after comparing similar studies in CG (Pavlou 2010), one can easily observe that the same pattern as in Graph 12 also appears in other structures which are assumed to undergo language change. In Pavlou (2010), inda mbu ‘what’ is the structure undergoing change to nambu. The data show that the same pattern is observed with the only difference being that the small increase is observed in the group of 45-60, and not 30-45 as in this study. It should be pointed out though that the population in that study was divided into four age groups, namely ‘60+’, ‘45-60’, ‘30-45’ and ‘18-30’. Given also that the structure studied in Pavlou (2010) is a wh-word, it can be argued that frequency is a significant factor for determining the ‘when’ of the beginning of language change. Embu, on the other hand, is an optional element. It is therefore expected that the sociolinguistic factors driving language change will first target those elements that are frequent in speech and in later stages, any other elements appearing less frequently.

82

Chapter Four

What is, however, striking is the similarity of the two phenomena studied in different points in time and still showing the same pattern of the first stage or otherwise called ‘initial stasis’. Precisely for this reason, the outlook of this paper is that not only older populations should be tested to complete our pool of data, but also populations younger than 18, since these are expected to show a rise of exhaustivity interpretations with embu and provide further support for a claim of change in progress in CG. Pending further testing rounds that will involve child populations, we take the results of this experiment to reveal intra-dialectal variation, since all sentences elicited mixed responses with respect to their status as true or false. This variation is expected if one takes into account that the dialect is largely heterogeneous. The cause of this variation can be related to morphosyntactic change in progress, a hypothesis that as mentioned above, seems to receive initial support from the different behavior of different age-groups. Older forms of CG, for example the poem “9 July 1821” (Michaelides 1873), involve both clefts of the sort ‘it is XP that YP’ and embu-clefts possibly due to language contact with French and English, which both display the first type of clefts. The change towards a nonexhaustive reading of CG clefts could be the effect of language contact with Standard Modern Greek, which does not have either bona fide or the embu-type clefts.

Notes 1 We are grateful to the audience of the 2nd Westminster Linguistics Conference of valuable feedback. Thanks are also due to members of the Cyprus Acquisition Team, and especially to Kleanthes Grohmann, for comments and suggestions. We also acknowledge financial support from the Gen-CHILD Project (no. 8037– 61017, awarded to Dr. Kleanthes K. Grohmann by the University of Cyprus). 2 There is no consensus with respect to the variety that one finds when moving towards the acrolectal pole of the Cypriot linguistic continuum. According to standard assumptions one would expect this variety to be SMG, however there are indications that what gets classified as Standard is in many cases much different from the SMG that one finds in Greece (Leivada et al. 2012). Therefore, the acrolectal pole could be taken to involve “Cypriot Standard Greek” (Arvaniti 2002, 2006) rather than SMG. 3 The attested environment was indicative sentences which are proclitic in SMG but enclitic in CG. Mixed performance boils down to the gradient nature of the morphosyntactic properties that pertain to each variety (Grohmann and Leivada, in press), viewing gradience from the perspective of syntactic variants existent within and affected by a dialect-standard continuum (cf. Cornips 2006 for the case of Standard and Heerlen Dutch).

The Gray Area of Acceptability Judgements

83

4

For a recent take on the need to standardize orthographic codification, see Armostis et al. (forthcoming). 5 Pcos ‘who/which’ is also found as pjos ‘who/which’ in literature (Grohmann et al. 2010) 6 Asizes is series of texts referring to the dialect which used to be spoken on the island around the 10th–11th century (see Simeonidis 2006: 150 for further details). 7 See Grohmann et al. 2006 and Papadopoulou (in progress) for further discussion. 8 We employ the term ‘acceptability judgments’ rather than the commonly used ‘grammaticality judgments’ first because we tested the semantic interpretation of the attested structures and not their syntactic status (although a difference in semantics possibly reflects a different syntactic analysis of embu) and second because the term ‘grammaticality judgments’ is a misnomer, since no individual has a precise knowledge of what part of her system of grammar accepts or rules out an utterance (Boeckx 2010).

References and Further Reading Agouraki, Y. 2010. It-clefts and stressed operators in the preverbal field of Cypriot Greek. Lingua 120, pp. 527-554. Arvaniti, A. 2002. The maintenance of diglossia in Cyprus and the emergence of Cypriot Standard Greek. Ms., University of California, San Diego. —. 2006. Linguistic practices in Cyprus and the emergence of Cypriot Standard Greek. San Diego Linguistic Papers 2, pp. 1–24. Boeckx, C. 2010. Language in Cognition: Uncovering Mental Structures and the Rules Behind them. Malden, Massachusetts: Wiley-Blackwell. Cornips, L. 2006. Intermediate syntactic variants in a dialect-standard speech repertoire and relative acceptability. In Gisbert Fanselow, Caroline Féry, Ralf Vogel and Matthias Schlesewsky (eds.), Gradience in Grammar: Generative Perspectives, pp. 85–105. Oxford: Oxford University Press. Doetjes, J, George R., and Rialland, A. 2004. Cleft sentences. In Francis Corblin and Henriëtte de Swart (eds.), Handbook of French Semantics, pp. 529–552. Stanford, California: CSLI. Fotiou, C. 2009. Focusing strategies in Cypriot Greek. Selected Papers from the 2006 Cyprus Syntaxfest. Chapter 3: pp. 63-91. Grohmann, K. K. and Leivada, E. To appear. Interface ingredients of dialect design: Bi-x, socio-syntax of development, and the grammar of Cypriot Greek. In: Anna Maria Di Sciullo (ed.), Towards a Biolinguistic Understanding of Grammar. Amsterdam: John Benjamins. Grohmann, K. K., Panagiotidis, P., and Tsiplakou, S. 2006. Properties of wh-question formation in Cypriot Greek. In Mark Janse, Brian Joseph and Angela Ralli (eds.), Proceedings of the Second International

84

Chapter Four

Conference of Modern Greek Dialects and Linguistic Theory, Patra: University of Patras. Grohmann, Kleanthes K. and Elena Papadopoulou. 2011. Question(able) issues in Cypriot Greek. Linguistic Analysis 37, pp. 1-35. Grohmann, K. K. and Papadopoulou, E. 2010. Cypriot anomalies in wh-in situ structures. In: A. Ralli, B. D. Joseph, M. Janse and A. Karasimos (eds.), e-Proceedings of the 4th International Conference on Modern Greek Dialects and Linguistic Theory (Chios, Greece: 11–14 June 2009), pp. 74–105. Grohmann, K. K., Theodorou, E., Pavlou, N., Leivada, E., Papadopoulou, E. and Silvia Martínez-Ferreiro. In press. The development of object Clitic placement in Cypriot Greek and the Romance connection. In Sadrine Ferré, Phillipe Prévost, Laurie Tuller and Rasha Zebib (eds.), The Romance Turn. Newcastle-upon-Tyne: Cambridge Scholars Publishing. Gryllia, S. and Lekakou, M. 2006. Syntax reflects information structure: Evidence from Cypriot Greek. Paper presented at the 27th Annual Meeting of the Department of Linguistics, Aristotle University of Thessaloniki. Heizmann, T. 2007. Child acquisition of exhaustivity in clefts. In Heather Caunt-Nulton, Samantha Kulatilake and I-hao Woo. Proceedings of the 31st Annual Boston University Conference on Language Development, 298-309. Somerville, Massachusetts: Cascadilla. Holton, D., Mackridge, P., and Philippaki-Warburton, I. 1997. Greek: A Comprehensive Grammar of the Modern Language. London: Routledge. Jackendoff, R. 1972. Semantic Interpretation in Generative Grammar. Cambridge, Massachusetts: MIT Press. Jespersen, O. 1927. A Modern English Grammar on Historical Principles, Part III, Syntax, vol. 2. Heidelberg: Winter. —. 1937. Analytic Syntax. London: Allen and Unwin. Kiss, Katalin E. 1998. Identificational focus versus information focus. Language 74(2), pp. 245–273. Leivada, E., Kambanaros, M., and Grohmann, K. K. 2012. Evaluating teachers’ grammatical competence in dialect and standard language. Paper to be presented to Bilingual and Multilingual Interaction, ESRC Centre for Research on Bilingualism in Theory & Practice, Bangor University. Michaelides, V. 1873. Ǿ 9Ș ǿȠȣȜȓȠȣ IJȠȣ 1821 [9th July 1821]. Retrieved from http://kypros.org/Poetry/Enati/. Newton, Brian. 1972. Cypriot Greek: Its Phonology and Inflections. The Hague: Mouton.

The Gray Area of Acceptability Judgements

85

Panagidou, F. 2009. Clefts and wh-question formation in Cypriot Greek. MA thesis, University of Utrecht. Papadima, A., Ayiomamitou, I., and Kyriacou, S. 2011. A harmonized orthographic system for the Greek Cypriot dialect. Paper presented at the 20th International Symposium on Theoretical and Applied Linguistics (ISTAL) 20, Thessaloniki, Aristotle University. Papadopoulou, E. In progress. Acquisition of wh-questions: Evidence from Cypriot Greek. Doctoral dissertation, University of Essex, Colchester. Papapavlou, A. 1998. Attitudes towards the Greek Cypriot dialect: Sociocultural implications. International Journal of the Sociology of Language 13, pp. 15–28. Papapavlou, A. and Pavlou, P. 1998. A review of the sociolinguistic aspects of the Greek Cypriot dialect. Journal of Multilingual and Multicultural Development 19, pp. 212–220. Pavlou, N. 2010. Inda mbu, nambu or ine ti pu? Acquiring complex structures in Cypriot Greek. In Antonis Botinis (ed.), Proceedings of Tutorial and Research Workshop in Experimental Linguistics, pp. 141144. Athens: University of Athens, University of Adger & University Of Wisconsin-Madison. Prince, Ellen F. 1978. A comparison of wh-clefts and it-clefts in discourse. Language 54(4), pp. 883–906. Roeper, T., Schulz, P., Pearson, B. Z., and Reckling, I. 2006a. From singleton to exhaustive: The acquisition of Wh-. In M. Becker and A. McKenzie (eds.), Proceedings of SULA 3. GLSA, University of Massachusetts. Roeper, T., Pearson, B. Z., and Strauss, U. 2006b. The acquisition path of the determiner quantifier every: Two kinds of spreading. In Tanja Heizmann (ed.), UMOP (University of Massachusetts Occasional Papers) 34. GLSA, University of Massachusetts, Amherst. Seymour, H., Roeper, T., and de Villiers, J. 2005. Development and validation of a language test for children speaking non-standard English: A study of children who speak Black English. National Institute on Deafness and Other Communication Disorders under Contract # N01 DC8-2104 to Harry N. Seymour, P.I. Simeonidis, C. P. 2006. ǿıIJȠȡȓĮ IJȘȢ ȀȣʌȡȚĮțȒȢ ǻȚĮȜȑțIJȠȣ [History of the Cypriot Greek Dialect]. Nicosia: Holy Monastery of Kykkos Centre of Research. Tsiplakou, S. 2004. ȈIJȐıİȚȢ ĮʌȑȞĮȞIJȚ ıIJȘ ȖȜȫııĮ țĮȚ ȖȜȦııȚțȒ ĮȜȜĮȖȒ: ȂȚĮ ĮȝijȓįȡȠȝȘ ıȤȑıȘ; [Attitudes towards language and language change: A two-way relation?] In Georgia Catsimali, Alexis

86

Chapter Four

Kalokairinos, Elena Anagnostopoulou and Ioanna Kappa (eds.), Proceedings of the 6th International Conference on Greek Linguistics. Rethymno: Linguistics Lab e-book.

Exercises 1. Are ‘it is XP that YP’ sentences bona fide ‘English type’ clefts in Cypriot Greek? Provide evidence based on the CyGEEC results. 2. Which factors influence participants’ judgements so as to classify sentences as 'true/false according to the story'? 3. Can acceptability judgement tasks show language change or change in progress? 4. Are there any carry over effects between the six stories described in CyGEEC? If yes, explain their impact on participants’ judgments. 5. As it is observed AG3 (45+) allows more exhaustive than nonexhaustive interpretations between cleft 1 and 2 (Group A) as well as embu 1 and embu 2 (Group B) on the one hand and cleft 3 and embu 3 (Group 3) on the other. Focus on the results and identify differences, if any, between Groups A and B. Provide possible explanations for such differences.

CHAPTER FIVE CONTACT INDUCED LANGUAGE CHANGES IN MODERN UYGHUR1 AMINEM MEMTIMIN

1. Introduction The Uyghur language belongs to the southeast group of the Turkic language family together with Özbek. It is spoken in the Xinjiang Uyghur Autonomous region of the Peoples’ Republic of China, bordered on the north by Russia and Kazakhstan, on the west by Kirgizstan and Tajikistan, in the south by Tibet and in the east by inner China. Today approximately 10 million people use Uyghur in China, beside speakers in other countries. Other languages have strongly influenced Uyghur in the past and in the present. This influence is reflected not only in the lexicon but also in morphology, syntax and pragmatics. According to the Annotated Dictionary of Uyghur (1999), approximately 20% of the lexicon consists of nonoriginal words, mainly Arabic-Persian words which entered after Islamisation, Mongolic words coming with the 13th century Mongol invasion, new Russian vocabulary from the 20th century and modern Chinese with the emergence of the Chinese Republic. A few words from Turkish were brought in by some intellectuals. Influences other than Chinese practically stopped in the second half of the 20th century; those have already become part of Uyghur. Features not typical of Turkic, like prefixes in morphology, IndoEuropean type right branching clauses in syntax and some new sounds and the new position of some phonemes in phonology have introduced some changes into the structure of Uyghur. Some new pragmatically induced influences can be found in the area of proverbs, idioms and new expressions. Code switching with Chinese, showing the significant influence of this language, frequently appears in everyday speech but not in the written language.

88

Chapter Five

1. 1 Main Linguistic Features of Uyghur Uyghur (an SOV language) is presently using an Arabic based writing system, with agglutinative suffixing and vowel harmony. It is a typical left branching language; attributes precede the head. It has subject predicate agreement, genitive-possessive government and uses the singular after cardinal numbers. Subjects can be expressed by personal suffixes of verb forms: 1. (Män) ottura mäktäp-tä oqu-y-män. (I) middle school-LOC read-PRS-1.sg ‘I am studying at a secondary school.’

1. 2 The Contact Background of Uyghur Uyghur has, during its history, had contact with different languages, among them Iranian languages, Arabic, Russian, Mongolic and an isolating language like Chinese, through geographical, religious, political, social and economic reasons. These contacts caused changes not only in the inventory of the lexicon, but also in morpho-phonological and syntactical structures. The changes took place to different degrees in different dialects, but are not wholly reflected in spelling. Chinese borrowings of the last hundred years appear mainly in the spoken language.2 The contact situation of Uyghur is still one of maintenance, although there was and is wide-spread bilingualism; this means that the potential of social motivation for lexical and structural borrowing is strong. Although code switching is popular among Uyghur bilinguals, it has not been studied systematically and seriously till now. There are different, conflicting attitudes towards code switching: In principle, Uyghurs would like to keep the language as pure as possible, without known foreign influence. (By “known foreign influence” I mean the elements concerning which one is aware that they are not original native material; this does not include foreign elements whose origin is not known to the speaker, which are so deeply rooted in the native speakers language that they are not felt to be foreign.) On the other hand, code switching is felt to be a sign of good Chinese language skills. These conflicting attitudes are well worth studying.

Contact Induced Language Changes in Modern Uyghur

89

2. The Impact of Lexical Borrowing on Phonology The most common and prominent way that languages influence each other is in the lexicon. Lexical borrowing can bring changes in phonology, as shown in Modern Uyghur.

2. 1 Introduction of New Phonemes In the phonological system of Modern Uyghur, the fricative ž is allowed to appear only in word onset before high vowels as an allophone of the y phoneme, e.g. yigit / žigit ‘young man’, yür- / žür- ‘to go’, yiraq / žiraq ‘far’, yu- / žu- ‘to wash’. But the introduction of foreign lexemes, especially from Russian, broke this rule: In the words žanir ‘genre’, žandarma ‘gendarme’ or žornal ‘journal’ it appears in front of low vowels and, e.g. in garaž ‘garage’, at the end of the syllable. Thus, ž is not a new sound in the language but, from being an allophone, it has become a new phoneme.

2. 2 Redistribution of Sounds The sounds /l, r, z, n, m, h, d, š , x, g, w, f, gh/ could originally not appear in word onset but got introduced into this position through lexical borrowing. Examples for this are lämpung < Chin. liang fen ‘a cold dish made from beans’, rayun < Russ. rayon ‘region’, nƗn < Pers. nƗn ‘bread’ or dos ‘friend’< Pers. dost ‘friend’.

2. 3 The Breaking of Phonotactic Rules In the original Turkic sound system, uvular q and its voiced fricative counterpart gh can be combined in one syllable only with back vowels like a, o or u, while the velar stops k and g can be combined only with front vowels like ä, ö or ü. Some of the borrowed words follow this rule, but not all of them, e.g. qäghäz from the Arabic word for ‘paper’, qäläm from Arabic ‘pencil’, ghäm from the Arabic word for ‘worry’ or kadir from the Russian word itself coming from French ‘cadre’. In some cases we find fluctuation: Persian goš ‘meat’ gave Uyghur göš by adapting the vowel to the velar stop, while the Qašghar variant ƥoš ‘meat’ retained Persian o but changed the velar to a uvular stop to accommodate it to the vowel.

Chapter Five

90

3. The Impact of Lexical Borrowing on Morphology 3. 1 Prefix-based New Word Formation Turkic languages use suffixes or analytical means to form new lexical elements. Originally there were no prefixes or infixes, but Persian influence has affected the Uyghur system to some degree. During the Chaghatay period of the language, which extended from the 14th to the 19th century (and even during the preceding 2-3 centuries), Persian was the language of culture and civilization for much of the Turkic world, including the Uyghurs; there also was some Turkic-Persian bilingualism and there still is widespread Turkic-Tajik bilingualism in Uzbekistan (Tajik essentially being a Persian dialect). Many Persian words which were adopted into Chaghatay Turkic are not in use any longer among Uyghurs, but many others are still being used at present, especially in the literary written language. Uyghur does not have prefixation in the derivational domain, and some Persian words were adopted with and without prefixes: In this way, the specific meaning and use of the prefixes became clear to Uyghur speakers. Some such words which were adopted together with prefixes were and are being used only with those stems; in a few cases, however, prefixes began to function independently, being added to native stems as well. In this way the original, canonical way of affixation was broken to a certain degree: 2a. Negation with na: na-ümid ‘hopeless’ with a Persian adjective but na-toghra ‘wrong’ with a Turkic one 2b. bät ‘bad’: bät-niyät ‘having bad intentions, bät-qiliq ‘behaving badly’ 2c. privative bi: bi-aram ‘uncomfortable’, bi-soraq ‘without asking, without permission’.

Among these instances, na-toghra, bisoraq and bät-qiliq are formed by adding prefixes to native Uyghur words; the content of bisoraq can also be expressed by soraqsiz with the Turkic privative suffix +siz.

Contact Induced Language Changes in Modern Uyghur

91

3. 2 Weak Gender Difference There is no gender category in the Turkic languages; biological gender can be expressed by lexical means. The Arabic feminine suffix +a was introduced with Arabic words like šƗҍira ‘female poet’ corresponding to šƗҍir ‘poet’. In Uyghur, šƗyirä is still in the use for referring to female poets, šƗyir for masculine poets, though the suffix has not been applied to original Uyghur stems.

4. Semantics Foreign words not only affected the lexicon but also had some influence on semantics. There are different types of result from this:

4. 1 Division within Semantic Fields The introduction of new elements sometimes changed the structure of semantic fields. et, the original Turkic word for ‘meat’, is used together with göš coming from Persian (see above), but the semantics of et got narrowed: While the general term for ‘meat’ is göš, et only denotes ‘human muscles’ or ‘meat with less bone’ and is used for positive reference to thin people putting on some meat: 3a. Bir kilo göš ber-ing. one kilo meat give-IMP.2.SG ‘Please give me one kilo meat’. 3b. Et-ing bäk ching-ken. muscle-POSS.2SG very tight-COP.EVID ‘Look, your muscle is very tight!’ 3c. Bu bala ämdi heli et elip qaptu. This child now quite meat take-CVB ACT COP.3SG ‘This child has now become quite fat.’

nƗn ‘bread’, borrowed from Persian, is another such case. Original Turkic ekmek (having undergone some sound changes) is in use only in some Uyghur dialects for denoting some specific types of bread, but nƗn is the general term. Words of Arabic origin are mostly abstract concepts mainly used in the written language. Words coming from Persian, on the other hand, denote not only abstract content but also everyday matters; words from Persian occupy an important position even in the basic

92

Chapter Five

vocabulary. Words coming from Russian serve mainly as technical, scientific and cultural terms. Chinese words are relatively numerous, but many among them are mostly restricted to the spoken domain: Speakers make efforts to exclude them from formal written and spoken language, as (perfect or weakly qualified) bilinguals know that they are not native; speakers may not be aware of the foreign origin older borrowings, already incorporated into Uyghur.

4. 2 Phrases Following Foreign Models Uyghurs have several possibilities for greeting each other, among them yaxšimu siz? ‘Are you fine?’, Tinchliqmu? ‘Are (you) at peace?’, qandaq ehwalingiz? ‘How is your state?’ or Essalamu äläykum, which is borrowed from Arabic as a whole. The first of these, a greeting which in fact does not expect a literal answer, does not have a long history: This is a calque which appeared under the influence of the Chinese popular greeting ni xao ‘Hallo!’ consisting of ni ‘you’ and xao ‘good’; this comes from ni xao ma which – literally – is a question, as it ends with the question particle ma (similar to the Uyghur question particle mu of the corresponding Uyghur expression). Uyghur women, especially in Urumqi, the capital of Xinjiang, have recently begun to use the expression Ћanim as the term of address ‘my dear’. Uyghur Ћan ‘soul’ changes its stem vowel and becomes Ћenim when the possessive suffix is added, and then signifies ‘my soul, my dear’. Turkish films being popular in Urumqi, Ћanim has been taken over from Turkish, where it took the Turkish shape and same semantic expansion.

5. Structural Changes 5. 1 Verb Stem Modifying Noun Imitation and translation of new Chinese cultural and scientific terms resulted in some unusual structures in Modern Uyghur. For some time, the Uyghur term for ‘supermarket’, a relatively recent institution, was deriЋidin tašqiri bazar, literally ‘market outside the degree’. This corresponds to Chin. chao shi (㉸ᕷ) ‘supermarket’, with Chin. chao ‘extraordinary’ intended as a literal translation of ‘super’. Later, when chao shi was expanded to zi xuan chao shi (⮬䘹䎵ᐲ) ‘self chose super market’ Uyghurs switched to talla bazar, literally ‘choose market’. talla and bazar are loan translations from Chinese. But talla bazar does not

Contact Induced Language Changes in Modern Uyghur

93

have a proper Uyghur structure, as verbs cannot modify nouns in Turkic languages; this widely used term follows Chinese grammar.

5. 2 The Appearance of Circumlocutions Another matter which has been discussed among Uyghur intellectuals is the expression elip bar- ‘to carry out’, which appears in documents, news and political reports translated from Chinese and in other formal texts. Chinese uses the verb jinxing ‘to carry out something’ with action nouns, e.g. jinxing diaocha (䘋㹼䈳ḕ) ‘to investigate’, literally ‘carry out an investigation’. Uyghur would normally have used the verb täkšür- ‘to investigate’ for expressing this content but instead replaces this by täkšürüš elip bar- ‘to carry out an investigation’, mechanically using elip bar- for translating jinxing. A similar case is munazire elip bar- ‘to carry out a discussion’ appearing in formal speech and formal documents and reports instead of ‘to discuss’, this being a literal translation of jinxing taolun (䘋㹼䇘䇪) ‘carry out a discussion’.

5. 3 Right-branching Structures Replacing Left-branching Ones In Turkic, relative and complement clauses are typically left-branching like other modifiers, whereas they are right-branching in most IndoEuropean languages: 4a. käl-gän adäm come-PRFP person ‘the person who came’, 4b. biz ular-ning sözliš-iwatqan-liq-i-ni kör-duq. we they-GEN talk-IPRG-DER-POSS-ACC see-PST ‘We saw that they are talking’.

But right-branching structures also exist in Uyghur as a result of contact with Persian, borrowing the Persian conjunction ki; e.g. 5a. Bu šundaq bir gül-ki, yaz-di-mu echil-idu, this such one flower CONJ summer-LOC-PRTC open-PRS.3SG qiš-ti-mu echil-idu. winter-LOC-PRTC open-PRS.3SG ‘This is such a flower that it opens both in summer and in winter’

The corresponding canonical relative clause would be:

94

Chapter Five 5b. Bu qiš-ti-mu, yaz-di-mu echil-idighan gül. This winter-LOC-PRTC summer-LOC-PRTC open-PART flower ‘This is a flower which opens in winter as well as in summer’.

5. 4 Words Consisting of Initials Some sequences of initials have been taken over in the shape they have in source languages, being spelled with Latin characters unlike Uyghur itself; such are VCD < Eng. / Chin., CD < Eng. / Chin., SARS < Eng. / Chin., BRT < Eng. / Chin., ‘Bus Rapid Transit’, HSK < Chin. abbreviation of Hanyu shuiping kaoshi ‘Chinese Level Test and WSK < Chin. abbreviation of waiguoyu shuiping kaoshi ‘foreign level test’. All these are important ingredients of Uyghur life: VCD is available in many families; the SARS virus frightened many when it spread in China; BRT entered Uyghur life towards the end of 2011; HSK is an important Chinese level test, one of the criteria for getting the bachelor’s degree. Being very short and having core function in people’s lives, these words got accepted quickly and are widely used. Uyghur-based terms were introduced instead but could not compete with them successfully.

6. The integration of Loanwords 6. 1 Phonological Integration Lexical borrowings tend to be adapted to the phonology and morphology of the recipient language. In Uyghur, borrowed items are assimilated to the sound and morphemic system.

6. 2 Consonant Clusters Most consonant clusters are not welcomed in Turkic languages; they change by means of prothesis, anaptyxis, syncope or apocope. In onset position, where they are completely barred, we find e.g. teraktur / tiraktur < Russ. traktor ‘tractor’, deramma < Russ. drama ‘drama’ or istil < Russ. stil’ ‘morality, way of life’. At syllable end we get e.g. bäxit < Pers. baxt ‘luck’, dos < Pers. dost ‘friend’ qäst ‘plot’ < Ar. qa‫܈‬d ‘intention’ or, with unvoicing, ghärp < Ar. gharb ‘west’. At the end of syllables consonant clusters are allowed only if they consist of a sonorant and a stop, e.g. änt ‘revenge, promise’; note, however, that alp ‘hero’ is pronounced as alip and qirq ‘40’ as qiriq. In semunt / semunit < Russ. tsement ‘cement’ and qän / qänt < Pers. qant ‘sugar’, we find that even the consonant cluster nt

Contact Induced Language Changes in Modern Uyghur

95

can appear either unchanged or with sound changes like apocope or anaptyxis.

6. 3 Hiatus, Diphthongs and Initial Pharyngeal Consonants Second syllables of Uyghur words do not start with vowels except in some borrowings in which the second syllable starts with a pharyngeal or with a glottal stop. There are two types of these. In one type, the first syllable ends with a consonant, e.g. änänä / ännänä < Ar. ҍanҍana ‘tradition’, or qälä / qällä < Ar. qalҍa ‘fort’. Note that the syllable border in the Arabic word is after the nasal, but in Uyghur the second syllable either begins with the nasal as ä-nä-nä, or gemination gives the pronunciation än-nä-nä; the original Arabic pharyngeal appears in the spelling but only a few intellectuals or people who know Arabic pronounce it. In the second type the first syllable ends with a vowel; Uyghur here avoids hiatus by adding a glide between the syllables: radiyo < Russ. radio ‘radio’, riyalizim < Russ. rializm ‘realism’, piyoner < Russ. pion’er ‘pioneer’. Not only hiatus but also diphthongs do not exist in Uyghur; in Chinese loanwords diphthongs are adapted as monophthongs. If the Chinese diphthong has a labial first component, a weak bilabial semivowel can be inserted after it, e.g. Guwang dung < guang dong, the name of a province in China (ᗅь), bing tüwän < Chin. bing tuan ‘corps’. In other cases, the diphthong assimilation can take a different course: xa Ћo < Chin. xua jiao ‘black pepper’; bä säy < Chin. bai cai ‘cabbage’; so+la- with the Turkic deverbal formative +lA- < Chin. suo ‘to lock’.

6. 4 The Fate of Arabic B and F at the End of Stems The voiced stop b becomes voiceless p when appearing at the end of a stem and a fricative w when followed by a vowel, e.g. the possessive suffix: Uyghur

Arabic

with 3SG possessive

meaning

1 2 3

ghäzäp Tärtip mänsäp

gha਌ab tartƯb manৢab

ghäziw-i tertiw-i mänsiw-i

anger order official title

4

Katip

kƗtib

katiw-i

secretary

Table 5-1: The fate of Arabic b and f at the end of stems

Chapter Five

96

When the Arabic source has syllable final f this also changes to the stop p, but this p does not become w in intervocalic position; e.g. Uyghur mƗrip Sinip Täräp

1 2 3

Arabic ma‫ݧ‬Ɨrif ৢinf ৬araf

with 3SG possessive mƗrip-i sinip-i tärip-i

meaning education class side

Table 5-2: The distribution of y and ž Above we spoke about the distribution of y and ž, and confirm that ž is allophone of y in conditioned circumstance. Ž occurs at the end of syllable result of fricativization of ‫ۥ‬. There is no Ћ in the syllable final position in Uyghur, so we can say safely this sort of new position for ž is the outcome of language contact. Here are some examples:

1 2 3

Uyghur Mäžbur Säždä Taž

Arabic ma‫ۥ‬bur sa‫ۥ‬da ta‫ۥ‬

meaning forced prostration crown

Table 5-3: The distribution of Ћ

6. 5 Plural as Singular Some Arabic words in plural form have been accepted in Uyghur as singulars and receive the native plural suffix +lAr when the plural is intended; e.g. haywan < Ar. ‫ۊ‬aywƗn (‘animal’ in both languages) and haywanat < ‫ۊ‬aywƗnƗt, the Arabic plural form. The Uyghur plural is regular haywanlar, but the double coding haywanat-lar is used in the name of a zoo and of a television program called Haywanatlar Dunyasi ‘the world of animals. One other example is ehwal, coming from a‫ۊ‬wƗl, the Arabic plural form of ‫ۊ‬Ɨl ‘state, situation’. In Uyghur both hal and ehwal came to be used in different contexts as singulars, and plurals are formed by adding the Turkic plural marker. We find hal-ehwal, a binomial consisting of the combination of these, signifying ‘the general state of affairs’. For greeting and to ask the health or well-being of a person, ehwal is generally used, also with the plural marker -lAr; hal, however, is only used as singular and mainly for the negative aspect of some person or event, like uning hali chong ‘He is not modest’ or hali xarap ‘His/Her situation is bad’. Arabic

Contact Induced Language Changes in Modern Uyghur

97

inflectional forms have thus undergone a lexical and semantic split in Modern Uyghur.

6. 6 Verbs Borrowed by Means of Suffix or Auxiliary Some Chinese verbs have been taken into the Uyghur lexicon by means of the very common denominal verbal suffix +lA- or the auxiliary verb qil- ‘to make’, e.g. süylä- < Chin. cui ദ both ‘to hasten’ or sola- < Chin. suo 䬱 both ‘to lock’; this latter verb already entered Uyghur a thousand years ago and has stayed transparent, as both the Chinese base and the suffix +lA- remained unchanged. chisili qil- ‘to make very angry’ < Chin. qisile ẻṚ஢ ‘got very angry’ is even more interesting, as it incorporated a complete verb phrase including the perfective morpheme le.

6. 7 Competition between Borrowings Among the languages which influenced Uyghur, Mongolian influence was limited to the period of Mongolian rule. Arabic and Persian influence started earlier, lasted much longer and was much deeper because of Islamic culture, but direct contact of the Xinjiang Uyghur population with the speakers of these languages ceased a number of centuries ago. Elements of these languages have been deeply rooted in Uyghur for some time and there is no Persian and hardly any Arabic education, so that Uyghurs cannot now recognise them. Russian became very popular around the beginning of the 20th century and then after the Chinese Cultural Revolution, when many Russian political, cultural and scientific terms were introduced. The influence of Russian in business, technology and education later weakened and has now practically disappeared; it has slowly been replaced by English and enthusiasm towards English is growing. Chinese influence is, however, clearly the strongest and is steadily becoming stronger; this trend is supported by the media, by general education and by the fact that bilingualism has became the necessary means to a better education and a better job. Close contact has made Chinese a live language for Uyghurs not only in its written form but also as a spoken language; it is thus no wonder that there are many Chinese borrowings in the oral language, though not so many in the written one. In the cities, many new cultural concepts are introduced first in Chinese, in the way Russian was being used before. For some of these terms Uyghur replacements are then coined by applying Uyghur word formation to native words, but the Chinese terms often stay on or loan translations emerge.

98

Chapter Five

Elements from these languages are often in competition with each other and are sometimes represented differently in the spoken and in the written language. Examples for such competitions among Uyghurs are Ћoza < Chin., üstäl < Russ. and širä < Mong. all signifying ‘table’, Turkic orunduq and bending < Chin. ‘chair’, universitet < Russ. and aliy mäktäp < Ar. and dašö < Chin. signifying ‘university’. Chinese diyen nao and eliktronluq mengä, literally ‘electronic brain’, a literal translation of the Chinese term, compete against Western kompiyuter. Among these words, Ћoza, bending and dašö are normally used in the spoken language, üstäl, orunduq, universitet and aliy mäktep mostly in the written and/or the formal language. diyen nao is, like the other Chinese terms, typical for the spoken language while, in the written language, kompiyuter has mostly replaced eliktironluq mengä. eliktironluq mengä is a typical example for eologisms proposed by some intellectuals and nationalists which did not gain popularity, partly because of their coming late and partly because Uyghurs just did not get to know them: kompiyuter and diyen nao are now preferred. Similarly, the spoken language prefers šoji/šouji < shouji. Chin. a Chinese borrowing for the ‘cell phone’, to yan telephun/ yan telepun, literally ‘side telephone’, which was considered to be more Uyghur. Some of the borrowings stayed in use even though quite proper Uyghur words were introduced after them. like ünalghu ‘tape recorder’, yangratqu ‘sounder’, kir alghu ‘washing machine’, tonglatqu ‘refrigerator’ and issitqu ‘water heater’, all formed by native means, are considered to be successful. The reason may be that some of them are in fact loan translations. Take Chinese luyinji ‘tape recorder’: It literally means ‘sound taking tool’ as it consists of lu ‘to record’, yin ‘sound’ and ji machine, tool’. In Uyghur ün ‘sound’ is governed by al- ‘to take, get’ and the derivational suffix -ghu forms nouns denoting purpose. Nevertheless, the former borrowed words luyinЋi / loyungЋi / loyung < Chin. luyinji, yingšang < Chin. yinxiang, šiyiЋi< Chin. xiyiji, bingšang < Chin. bing xiang and reshüychi < Chin. reshuiqi still kept their status in the spoken language in spite of the formal status of the neologisms.

7. Codeswitching Uyghurs commonly use Chinese numbers when mentioning a telephone number unless they are illiterate in Chinese or if they want to tell them on purpose in Uyghur. Here is an Uyghur dialog: 6. A: Telifun numur-ingiz qanche? telephone number-POSS.2SG how.many ‘How is your telephone number?’

Contact Induced Language Changes in Modern Uyghur

99

B: yao san wu qi jiu ling yao er jiu qi. siz-ning-chu? You-GEN-PRTC ‘1357901297. How about yours?’ A: yao wu san ge jiu san ge yao si yao qi, ‘15999111417’

This happens also with street numbers of houses; in the following conversation the words si ‘four’, hao ‘number’ and lo ‘building’ are Chinese while däyyän comes from Chinese danyuan: 7. A: Öy-ingiz qaysi bina-da? house-POSS.2SG which building-LOC ‘In which building is your house?’ B: Si hao lo-di-ki üch-ünchi däyyän-dä. four number building-LOC-CONV three-ORD unit-LOC ‘In the third unit of building number four.’

Such conversations with Chinese telephone and house numbers are very common among Uyghurs, but Uyghur numbers are used in all other occasions. In the next conversation säy ‘vegetable’ comes from Chinese while koy (< Chin. kuai) and mo (< Chin. mao) are monetary units: 8 A: Ma säy-ning kilo-si qanchä pul? this vegetable GEN kilo-POSS.3SG how.many money ‘How much does this vegetable cost by kilogram?’ B: Töt koy bäš mo. four kuai five mao ‘Four yuan and a half.’

Why are Chinese numbers limited to those specific contexts? For an attempt at an explanation compare the telephone number ‘1357906298’ in Chinese and Uyghur: Num Uyg Chin

1 Bir Yao

3 üš san

5 bäš wu

7 yättä qi

9 toqquz jiu

0 nöl ling

Table 5-4: Numbers in Chinese and Uyghur

6 altä liu

2 iškki er

9 toqquz jiu

8 sekkiz ba

100

Chapter Five

One explanation might be economy: As you see above, Chinese numbers are generally shorter than Uyghur numbers. Some Uyghur speakers say they find a number easier to remember if they say it in Chinese; they might also be motivated by the need to show off their Chinese capability. There is no doubt, at any rate, that Chinese, the official and important language at work and in the educational domain, used in the vast majority of media like television films and advertisements, has enhanced and facilitated this trend.

8. Conclusion Language contact is a very complicated topic, as it is not only related to the past but also the present, and touches not only upon language but also on all aspects of social life. It is determined by many complicated factors. What we have commented on above are only some parts of the contact phenomena around Uyghur; a more systematic description and analysis would need more space and time.

Notes 1

I would like to express my thanks to my Professors Metrihim Sayit and Marcel Erdal who gave their valuable comments on some of the topics. 2 Some of the examples come from the publications of Abduraxman and Abduqadir and Kuresh and from the Annotated Dictionary of Uyghur (Uyghur Tilining Izahliq Lughiti) published by the Language and Writing Committee of the Xinjiang Autonomous Region (Xinjing Uyghur Aptunum Rayunluq Til we Yeziq Kometiti).

References Abduraxman, H and Abduqadir. E. 2001. Uyghur Tiligha Chättin KirgänSözlärning Izahliq Lughiti, Xinjiang Xälq Näshriyati, Xinjiang. Küräsh, R. 2003. Hazirqi Zaman Uyghur Tili. Xinjiang Xälq Näshriyati, Xinjiang. Xinjiang Uyghur Aptunum Rayunluq Til wä Yeziq Kometiti 1999. Uyghur Tilining Izahliq Lughiti. Xinjiang Xälq Näshriyati, Xinjiang.

Contact Induced Language Changes in Modern Uyghur

101

Abbreviations ACC AUX ACT CONJ CONV COP CVB DER GEN EVID IMP LOC NG ORD POSS PRFP

accusative auxiliary actionality marker conjunction converter copula converb derivation genitive evidential imperative locative negation ordinal suffix possessive perfect participle

IPRG intra-terminal progressive PRS present PART prospective participle PST past PTCL participle PRTC particle SG singular 1 first person 2 second person 3 third person Ar. Arabic Chin. Chinese Mong. Mongolian Pers. Persian Russ. Russian Tur. Turkic

Exercises 1. Is the present language situation of Uyghur one of maintenance or of shift? Explain. 2. How was the Uyghur language influenced in the past and what is taking place with its interactions at present? 3. Can you give an example for contact induced phonological change in Uyghur? 4. What factors could be causing the consistent use of Chinese numbers rather than Uyghur ones in certain contexts in the spoken language? 5. Did contact bring changes to the grammatical system of Uyghur?

CHAPTER SIX THE ROLE OF THE INPUT IN THE EARLY DEVELOPMENT OF A HERITAGE LANGUAGE1 FRANCESCA LA MORGIA

1. Introduction One of the major difficulties in assessing linguistic competence in bilingual children lies in accounting for the different language internal and language external factors that could explain variation in performance and in ultimate attainment. Variation in bilingual development is associated with factors such as the contexts of acquisition and use, the attitude of the child, the amount of use of the two languages and their role in the child’s daily routine (De Houwer 2007; Paradis 2011). Accounting for these and other variables is essential for understanding the different phenomena which characterise the linguistic development of bilingual children. The variable under examination in this study is the linguistic input to which the child is exposed. As several studies have demonstrated, bilingual children generally acquire the majority language easily because they have the opportunity to hear it and use it frequently (Gathercole and Thomas 2009). However, the other language, which may be spoken by one or more members of the family and generally by a minority group, may be acquired more slowly, it may not be fully acquired, or it may present features of a non-native language (Schlyter 1993). This study examines the relationship between the development of a heritage language and the quantity and quality of input the child receives. In the context under examination, Italian is a minority language spoken by a socially integrated community group within the Irish society. The children who participated in this study have an Italian mother and an Irish father, and they acquired Italian mainly by being exposed to input provided by their mother. For this reason, throughout this chapter, Italian will be referred to as heritage language (henceforth HL), following Valdès’ (2000) definition, according to which a HL speaker is a bilingual

The Role of the Input in the Early Development of a Heritage Language 103

individual raised in a home where a language other than the majority one is spoken (see also Polinsky and Kagan 2007; Kagan and Dillon 2008; Benmamoun, Montrul and Polinsky 2010). In the context under examination, the majority language, English, is the family language, while the HL is used by the mother and by a limited number of people in the child’s environment. In terms of input and interaction, this means that the child has limited opportunities to hear and use the HL. Ultimately, the lack of access to quantitatively and qualitatively rich linguistic evidence may result in difficulties in acquiring and maintaining the HL (Montrul 2008; Rothman 2009). Therefore, this study aims to determine whether there is a correlation between exposure to input and development of the HL. The implications of this study are both practical and theoretical: first of all, there is growing awareness of the fact that the acquisition of two or more languages is increasingly becoming the norm: understanding the role of input and interaction for the development of a language that is spoken only at home can help to stress the importance of parents, who are responsible for providing the necessary conditions for successful bilingual language acquisition and maintenance. Secondly, examining the possible limitations and difficulties encountered by HL children has implications for understanding typical and atypical bilingual language acquisition. Moreover, it is important to examine the role of the input in relation to language development to understand the causes of phenomena such as attrition and incomplete acquisition.

1.1 The Role of the Input in Language Acquisition The HL is acquired within the family environment and it is generally used for a limited amount of functions (Rothman 2007; Benmamoun et al. 2010). Given the variation in the availability of opportunities to use the HL, we can expect to find different outcomes in its acquisition and maintenance (Place and Hoff 2011). It is therefore essential to gain a better understanding of the kind of input children are exposed to in order to fully understand HL development and to find out whether lack of input can result in incomplete acquisition or in anomalies in the children's linguistic production. It has been widely demonstrated that language development can be significantly affected by the quality and quantity of interaction with adults (Vigil, Hodges and Klee 2005; Pancsofar and Vernon-Feagans 2006; Pan, Rowe, Singer and Snow 2005). Research has shown that the input provided by the adult affects the child’s lexical and morphosytanctic development (Hoff and Naigles 2002) and that the adult’s engagement in conversation has a significant impact on the child’s understanding of word

104

Chapter Six

meaning (Akhatar and Tomasello 2000). The lack of cohesive and pragmatically appropriate input has also been found to affect the rate of linguistic development. In a study of parent-child interaction, Vigil et al. (2005) found that parents of late-talking toddlers had been providing insufficient feedback to their children, using less expansions, responses and self-directed speech than parents of typically developing toddlers. Evidence of the importance of the input has also been highlighted by studies on language delay and deprivation. Windsor, Glaze and Koga (2007) found that children who were raised in a deprived environment within institutions showed more limited language skills compared to their peers living with foster families. As these studies have demonstrated, the quality of input has a significant impact on language development. Similar results have been found in studies involving bilingual children. Research has shown that bilinguals who are not sufficiently exposed to the HL or who are exposed to qualitatively poor input may not develop this language at a normal rate and they may fail to acquire it fully (Montrul 2008). It is also necessary to consider that while the provision of rich input may constitute part of the basic prerequisites for successful acquisition, other non-linguistic factors may play an important role. These include the parents’ linguistic strategies, the attitudes towards the two (or more) languages as well as the number of opportunities to use them (Juan Garau and Pérez-Vidal 2001). Through a survey of over 1000 families in Flanders, De Houwer (2007) found that all children acquired the majority language, but those raised in families in which both parents spoke both languages were less likely to use the minority language. The main factor affecting the use of the two languages by the children proved to be the combination of languages spoken by the parents, confirming the importance of input and interaction within the family. Yamamoto (2001) also argued that children who “engage” with the HL and are often exposed to it will be more likely to use it frequently. It is therefore extremely important to measure the quality and quantity of input in order to determine the effects of limited exposure on the development of a HL.

1. 2 Measuring Quantity and Quality of Input The initial assumption tested in this study is that both quantity and quality of input affect the development of the HL. The quantity of input is generally measured on the basis of questionnaires which gather information on the amount of time the child spends with each parent and with other interlocutors. While the quantity of input is essential for the child to gather the necessary linguistic evidence, the quality of the

The Role of the Input in the Early Development of a Heritage Language 105

language used by the parent and the type of interaction between the adult and the child have a significant impact on the acquisition process (La Morgia 2011). The input can be considered qualitatively rich when it is comprehensible, informative and relevant to the context, and when it contains rich vocabulary as well as a variety of comprehensible syntactic structures. Non-linguistic communicative behaviour used to establish joint attention, such as gesture, gazing and pointing, is also an important aspect of parent-child interaction (Zammit and Schafer 2011). The assessment of the quality of the adult’s input is frequently based on the analysis of vocabulary richness or density, which is generally considered to be a predictor of the child’s lexical development. Word production and comprehension have also been found to be enhanced in children who are exposed to many different words in a variety of contexts (Bornstein, Haines and Painter 1998; Hoff and Naigles 2002). Other criteria which have been tested for the assessment of input quality are morpho-syntactic complexity, measured on the basis of the mean length of utterance (MLU), turn-taking, types of utterances (questions, statements, commands) and amount of speech produced by the parent within a set period of time (Pancsofar and Vernon-Feagans 2006). In this study, the quantity of input will be measured on the basis of the results from the Questionnaire on the Linguistic Background of the Bilingual Child and the quality on the basis of audio-recordings of spontaneous spoken data. The qualitative criteria selected, which were adapted from Pancsofar and Vernon-Feagans’ study (2006), are lexical variety (vocabulary), amount of speech produced (output), and morphosyntactic complexity (MLU calculated in morphemes). In order to determine whether there is a relationship between exposure to input and linguistic development, the children’s production is measured on the basis of MLU and vocabulary, which are common criteria for the assessment of language development. Two other criteria have been included, namely production of errors and use of Italian, since it has been previously shown that bilingual children mix more when using the non-dominant language (Genesee, Nicoladis and Paradis 1995) and that they make non-native like errors (Schlyter 1993; Bonnesen 2009). The child’s choice to switch to English when addressing Italian speakers may be seen as a compensatory strategy employed to fill a knowledge gap or to use the language with which they feel more comfortable. The analysis of errors is carried out on the basis of the assumption that children who are exposed to more limited input will be more likely to encounter some difficulties in the HL, due to the fact that the access to positive evidence is insufficient to acquire the language and to “correct” common developmental errors. Bilingual children

106

Chapter Six

have been found to make more errors than monolinguals, and to make non-native like errors in morphology and syntax (Bonnesen 2009; Döpke 2000). We can therefore assume that children who do not have sufficient access to linguistic input as well as the opportunity to use the HL will make errors for a prolonged period, and some of these non-target forms could become fossilised.

2. Method 2.1 Participants The families involved in this study were composed of an Italianspeaking mother, an English-speaking father and one child. The parents used the one parent-one language strategy, with English as the family language. The families were resident in the greater Dublin area, and the mothers had been living in Dublin for 5-7 years. All parents had been employed since they completed higher education and their children attended English-speaking nurseries on a part-time basis. The mothers were all highly proficient in English, while the fathers had different levels of proficiency in Italian and rarely used it. Parents claimed to have a positive attitude towards the use of Italian at home. This emerged from conversations which took place prior to the start of the research and through responses to the Questionnaire on the Linguistic Background of the Bilingual Child. The three children, Paolo, Francesca and Costanza, acquired English and Italian from birth. At the start of the longitudinal study, the children were starting to produce two-word utterances in Italian. Conversations between children and their mother were recorder monthly over one year. Paolo was 3;1 at the time of the first recording. He started attending nursery since he was 12 months old for 4 days a week for a total of 36 hours a week. He usually spends about 15-20 hours a week with his mother and 10-15 with his father and he is exposed to a daily average of 23 hours of Italian and 10-13 of English. He speaks mostly English and sometimes he uses it when speaking to his mother. He speaks Italian daily with his mother and sometimes on the phone to the family in Italy. He also spends at least one month in Italy every year. Francesca was 2;4 at the time of the first recording. She started attending nursery since she was 9 months old for 5 days a week for a total of 25 hours a week. She spends about 25 hours a week with her mother and 20 with her father. She uses English in the nursery, with her father and his family and Italian with her mother and her family. She has also spent at least one month in Italy every

The Role of the Input in the Early Development of a Heritage Language 107

year since her birth. Costanza was 1;11 at the time of the first recording. She started attending nursery when she was 6 months old for 2 days a week for a total of 16 hours a week. When the parents work, she is taken care of by members of the extended family who only speak English. She spends about 24 hours a week with her mother and 8 with her father and she is exposed to about 4 hours of Italian and 8 of English every day. She uses English in the nursery, with her father and his family and Italian with her mother and her family. She spends 1 to 2 months in Italy every year.

2.2 Procedure The children’s exposure to Italian was assessed examining both quantity and quality of maternal input.The quantity of input was measured by administering the Questionnaire on the Linguistic Background of the Bilingual Child. The questionnaire included questions concerning the amount of time spent by the child with Italian speakers on a weekly and yearly basis, taking into account the child’s daily routine and also the time spent in Italy every year. It was compiled by the parents, who were asked to observe their family’s linguistic behaviour during a normal week. The responses gathered through the questionnaires were used to estimate the average quantity of exposure to Italian during a typical week and over a year. The qualitative analysis of the input was based on audio-recordings of mother-child spontaneous interactions in three free play sessions, each lasting 1-2 hours. Samples of one hundred utterances were chosen on the basis of the spontaneity of the interaction; they were not extracted from the beginning of the recording sessions and they did not involve reading or storytelling. The criteria used for analysing the quality of input were output, vocabulary and complexity (Pancsofar and Vernon-Feagans 2006). Output accounted for the total number of words produced by the parent over a period of 10 minutes during the three sessions. Vocabulary richness was calculated by computing the number of different word roots, omitting unintelligible words and fillers. Complexity was measured by computing the parents’ MLU in morphemes. Finally, the children’s production was analysed on the basis of vocabulary, number of errors of verbal morphology, word order, gender and number agreement and percentage of use of Italian utterances.

Chapter Six

108

3. Results 3. 1 Quantity of Input The analysis of the quantity of input was based on the parents’ responses to the questionnaire. Figures 6-1 and 6-2 illustrate the percentage of weekly and yearly exposure to the two languages, while Figure 6-3 illustrates the amount of exposure over one year, taking into account the increased exposure during the periods spent in Italy. The results based on the total yearly amount of Italian input show that all children are exposed to Italian less than to English and that Costanza’s exposure to Italian is twice as frequent as Paolo’s. 320 200

305 60

45

320 45

0 Costanza

Francesca Italy Ireland

Paolo

Figure 6-1: Average number of days spent in Italy and Ireland every year 

100 50

33

85

75

67 25

15

0 Costanza

Francesca Italian English

Paolo

Figure 6-2: Average percentage of weekly exposure to each language in Ireland

The Role of the Input in the Early Development of a Heritage Language 109

100 50

76

64

59 41

36

24

0 Costanza

Francesca Italian English

Paolo

Figure 6-3: Average percentage of exposure to each language in one year

3. 2 Quality of Input The results of the analysis of the input show that when they engage in conversation or play with their children, the mothers produce between about 300 and 700 words in each 10 minute sample (as displayed in Table 6-1). Mother of

Number of words produced in 10’ Sample 1 364

Number of words produced in 10’ Sample 2 370

Number of words produced in 10’ Sample 3 280

Average

Paolo

400

450

379

409.6

Costanza

786

725

653

721.3

Francesca

338

Table 6-1: Mothers’ output The output produced by Costanza’s mother is almost twice as lexically dense as that of the other two mothers. The output was calculated by computing the number of utterances produced by the parent over a period of 10 minutes during 3 free play sessions, which are displayed in chronological order (Table 6-2).

Chapter Six

110 Mother of

Sample 1

Sample 2

Sample 3

Average

Francesca

190

226

182

199.3

Paolo

286

202

228

238.6

Costanza

272

280

260

270.6

Table 6-2: Mothers’ vocabulary Mother of

Sample 1

Sample 2

Sample 3

Average

Francesca

4.52

6.4

7.64

6.18

7

7

7.5

7.16

8.92

8.92

8.94

8.92

Paolo Costanza

Table 6-3: Mothers’ MLU Child

weekly exposure to Italian (days) 25

yearly exposure to Italian (days) 36

output

vocabulary

complexity

338

199.3

6.18

Paolo

15

24

409.6

238.6

7.16

Costanza

33

41

721.3

270.6

8.92

Francesca

Table 6-4: Summary of results As shown in Table 6-4, Costanza’s scores were higher both for the quantity and quality of input. The initial assumption was that input can affect the child’s linguistic development, so we would predict that Costanza’s development of Italian will benefit from the high quantity and quality of exposure to the language and that she may have an advantage over Paolo and Francesca. The results also create the premise for addressing a different question: since Paolo and Francesca receive respectively more qualitative and more quantitative input, the analysis of their linguistic development would allow us to infer whether quality has a

The Role of the Input in the Early Development of a Heritage Language 111

more significant impact than quality. Even though the amount of participants in this study is not sufficient to make reliable generalisations, the patterns discovered could constitute the basis for further investigations.

3. 3 The Children’s Linguistic Development The study of the input shows that overall Costanza is exposed to a higher quantity and quality of input, while Paolo and Francesca’s parents respectively achieve higher scores in the qualitative and quantitative analyses. The diversity among the three children allows us to determine whether both quality and quantity are essential for successful language acquisition of a heritage language, and which of the two affects development more significantly. The children were matched for MLU calculated in morphemes in order to compare them at similar developmental stages. Two MLU stages were chosen, because they were available for all children and they represented the transitional stages from one to two-word utterances (MLUm 1.5-2) and from two to multi-word (MLUm 2.5-3). The children were compared at these two stages on the basis of their vocabulary size, number of errors (morphology and syntax) and percentage of Italian utterances. The initial prediction was that children who were exposed to a “reduced input” were more likely to have a smaller vocabulary, to make more errors and to use English as a compensatory strategy. The results show that Costanza, who is exposed to the highest quantity and quality of input, reaches the two MLU stages the earliest (Tables 6-5 and 6-6). Her vocabulary is still limited at the age of 1;11, but 4 months later its size has increased. At both stages Costanza makes few or no errors and she rarely switches to English. Francesca and Paolo display a different pattern. If we observe Paolo’s results, we see that his score for vocabulary is high even with a low MLU, but 8 months later it has not significantly increased. However, his use of Italian increases. Francesca is the child who makes most errors and who uses English the most. Even though the quality of input Francesca is exposed to is lower that Paolo’s, her vocabulary increases at a rate similar to Costanza’s, and her percentage of use of English decreases. What we can infer, given the small number of participants, is that exposure to a limited quantity of Italian input results in slower development, as demonstrated by Paolo’s MLU, while the results suggest that exposure to a lower input quality may cause the production of a higher number of syntactic and morphological errors.

Chapter Six

112 Child

Age

Vocabulary

Number of errors

Italian utterances

Francesca

2;6

17

5

71.4%

Paolo

3;3

31

4

89%

Costanza

1;11

13

0

98%

Table 6-5: Results at MLU 1.5-2 Child

Age

Vocabulary

Number of errors

Italian utterances

Francesca

3;1

44

10

92.4%

Paolo

3;11

37

4

96%

Costanza

2;2

32

2

98.6%

Table 6-6: Results at MLU 2.5-3

3. 4 Futher Insights into the Children’s Language Use This section provides a qualitative analysis of the children’s data, extracting some features which can be considered relevant for the understanding of linguistic phenomena typical of development of Italian as a HL.

3. 5 Features of Francesca’s Linguistic Development Up to the age of 2;6 Francesca produces mostly one-word utterances, which are generally the repetition of her mother’s last word (1). She also produces a number of English and mixed utterances (2, 3). 1.

MOT MOT FRA

questo libro non lo puoi leggere tu. you can’t read this book. è per grandi. it’s for grown-ups. grandi. grown-ups.

The Role of the Input in the Early Development of a Heritage Language 113 2.

MOT FRA

3.

FRA

questa è l’oca! this is the duck! duck!

2;4.20

open questo. open this.

2;5.10

Francesca’s English utterances are longer and more complex than the Italian ones and they all contain verbs. The first Italian verbs emerge at 2;7. Before the emergence of inflection, the only verbs are produced in the imperative form fai ‘do’, guarda ‘look’, apri ‘open’, lascia ‘leave’, siediti ‘sit down’. The first inflected verbs used productively by Francesca are the third person form of essere ‘to be’ and avere ‘to have’. She also starts using the verb volere ‘want’ frequently from the age of 2;9. However, she constantly uses the second person form vuoi instead of the first voglio and sometimes also instead of the third (4a-4f). This error is very frequent in Francesca’s data, and it is documented up until the last recording, mostly with the verb volere ‘want’. Francesca’s mother often corrects her use of this verb, but she continues to use the second person2. Mothers often use the verb want, and, especially when addressing their child, they are likely to use phrases such as ‘do you want’ (vuoi) very frequently. Unlike want, volere is an irregular verb and every person in the present tense is marked by inflection. Francesca’s mother frequently uses vuoi when addressing her child and no instances of voglio are found in the recordings of the mother’s speech. Therefore, the misuse of first person verb is probably a fossilised error that could be attributed to frequency of input. The occurrence of such phenomenon for a prolonged period of time is typical of heritage language acquisition and it is likely to be related to the lack of access to sufficient evidence for establishing the correct form. Examples 4a, b and c show the use of the second person vuoi instead of the first voglio, both followed by a noun (4a) and by a verb (4b). Francesca uses vuoi with null subjects, but also with overt ones, as shown in example 4e. Example 4d shows the extension of vuoi to the third person.

Chapter Six

114 4a.

4b.

4c.

FRA mamma, vuoi l’acqua ancora! mum, I want water again! FRA vuoi acqua ancora! I want water again! MOT che vuoi? what do you want? FRA acqua. water. FRA

FRA FRA

4d.

4e.

4f.

FRA

FRA

FRA

2;9.07

vuoi colorare! I want to colour.

2;10.17

lo vuoi! I want it. vuoi televisione! I want TV!

3;0.17

perchè vuoi mangi i pesci? why do you want to eat fish? (referring to an animal who wants to eat fish)

3;1.17

io vuoi togliermi scarpe. I want to take off my shoes.

3;2.27

vuoi yoghurt! I want yoghurt!

3;5.0

From age 2;9 we can observe an increase in the use of longer sentences and inflected verbs. 5.

FRA

6.

FRA

7.

FRA

che c’è dentro questa? what is inside this? do’ (dove) sta casa lupo? where is the wolf’s house? vhe fai, ti metti giacca? what are you doing, are you putting on (your) jacket?

2;9.7

There is little evidence of deviant word order in Francesca’s data. The few examples are found in the use of the genitive.

The Role of the Input in the Early Development of a Heritage Language 115 8.

9.

FRA

FRA

Marta macchina questa. MARTA CAR THIS this is Marta’s car Target: questa è la macchina di Marta/è la macchina di Marta questa

2;9.7

questo bambino’s letto THIS CHILD’S BED this (is the) child’s bed Target: questo è il letto del bambino

2;10.17

However, this deviant word order is found together with the correct word order. 10. FRA do’ (dove) sta passeggino Marta? where is Marta’s pram?

2;9.7

After turning 3, Francesca produces longer sentences and exhibits a more varied lexicon. She makes only few gender and number marking errors, but she still has difficulties with verb inflection and the use of articles (11). 11. FR A Io vai alla scimmiette. 3;2.27 I go(2nd sg) to the(sg) monkeys(pl).

The data collected over 12 months show that Francesca’s Italian has been developing within the norm. Her MLU and vocabulary size have increased, but she exhibits difficulty with some linguistic features. Francesca’s use of vuoi appears to be a fossilised error, which can be attributed to the lack of input, suggesting that the evidence she is exposed to is not sufficient for her to acquire the correct form at a fast pace. It is likely that Francesca, given sufficient input, will start distinguishing between vuoi and voglio, since volere (want) is a high frequency verb. However, if access to input has implications for fossilization or delay in acquisition, we could expect that more complex or less frequent linguistic features may be acquired later or may not be acquired at all.

3. 6 Features of Paolo’s Linguistic Development Paolo is the oldest of the four children who participated in this study. His Italian is initially considerably less developed than his English, and also less developed in comparison to his bilingual peers. His Italian

Chapter Six

116

production at 3;2 still shows some of the characteristics of the early stage of language acquisition. Initially most of his sentences are very short, and he produces incomplete words and utterances. Paolo exhibits several examples of multiple switching within the same sentence. The switches occur mostly at the VP, as shown in the examples below. 12a.

PAO guarda, he can’t get out. look, he can’t get out. INV non riesce a uscire. (he) can’t get out PAO no, can’t uscire là. no, can’t get out there.

12b.

MOT PAO

12c.

MOT PAO

12d.

PAO

dopo cena se sei buono. after dinner if you are good. I’m buono. I’m good.

P 3;3.23

e loro cosa fanno? and what do they do? gone via. gone away (they go away). ma what’s he got in mano? but what’s he got in his hand?

3;4.25

Just before turning 4, Paolo starts producing more complex utterances. During this phase, most of his sentences are interrogatives. At age 4 Paolo still omits determiners and makes gender and number errors. Some of his utterances show an influence from English. In example 13, Paolo responds to his mother using a typical English contrastive expression, no, you are!, which bears emphasis on the pronoun you. This is a case of crosslinguistic influence, showing that Paolo is using English syntax with Italian lexicon. 13.

MOT stai facendo un po’ di confusione, sai? you are getting a little confused, you know? PAO no, tu sei! no, you are

3;11.17

The Role of the Input in the Early Development of a Heritage Language 117

Unlike the other children, Paolo produces non-target word order that deviates from the target in both languages. However, these cases are rare, and most of his deviant word order utterances seem to be modelled on the English word order. 14a.

14b.

14c.

14d.

PAO

PAO

lupo mano! wolf hand (the wolf’s hand) Target: la mano del lupo

3;3.23

aranciona barca piccolo. orange boat small Target: barca arancione piccola/ Piccola barca arancione/barca piccola arancione

3;11.17

PAO è l’acqua andato. (he) is the water gone (he is gone into the water) Target: è andato in acqua/nell’acqua

3;10.19

PAO cosa questo è? what this is? Target: cosa è questo?

4;0.29

Overall, the development of Paolo’s Italian is slower than the other two children’s. There also seems to be a delay in the production of structures that are generally acquired early in Italian (see La Morgia 2011). Paolo’s sentences are often incomplete and his frequent mixing often makes his utterances hard to understand. He also seems to have difficulties in comprehension, and often asks his interlocutor to repeat. The step towards the production of multi-word utterances takes place just before the age of 4.

3. 7 Features of Costanza’s Linguistic Development The analysis of the data collected during the first two recording sessions captures the early stage of language production. Costanza is the youngest child involved in this study. Since the first session, she showed to have full understanding of every sentence addressed to her, and she has always responded with no hesitation to everything she was asked, displaying the ability to talk about a variety of topics with confidence and with age-appropriate vocabulary. Unlike the other children, she seems to

Chapter Six

118

be very similar to a monolingual Italian child. Costanza’s early sentences show instances of article and verb omission within the norm, and of correct gender and number marking. Unlike Francesca and Paolo, since the early stages Costanza spontaneously starts a new conversational topic. After the age of two, we can observe an increase in her production of new vocabulary, as well as the emergence of tense and of more complex structures. At this stage, Costanza masters gender and number marking, and the use of articles (15a and 15b). 15a.

COS

eccolo il cane, eccola la nonna. here is the dog (masc.sing.), here is the granny (fem.sing).

15b.

MOT

e tu di che colore ce li hai i calzini? and what colour are your socks?

COS

bianchi. white (masc., plur.).

16.

COS

mamma, di’ a Francesca devo mette questa. mum, tell Francesca (that) I have to put on this.

MOT

no, mettiti la scarpa che scivoli. no, put on the shoe or you’ll slip.

17.

COS

Franci, vieni perchè ci nascondiamo! Franci, come here because we hide!

18.

MOT

lasciamo la bambolina qua? we leave the dolly here?

COS

no, perchè se lei piange non possiamo uscire. no, because if she cries we can’t go out.

2;2.17

2;6.7

2;9.14

In the last session, which took place when Costanza was 2;9, we can observe an increase in MLU and in the use of subordinate and coordinate clauses (16-18). Finally, it is worth noting that Costanza often speaks Italian when playing alone.3 This shows that her use of Italian is not only restricted to the interaction with her mother, but it is her natural linguistic choice in other contexts.

The Role of the Input in the Early Development of a Heritage Language 119

4. Discussion This study investigated the effects of maternal input on the development of a heritage language in young bilingual children. The three families selected constituted a rather homogeneous sample: all mothers were Italian, they had been living in Dublin for 5 to 7 years and they were employed part-time. The children were exposed to Italian and English daily, and the parents used the one parent-one language strategy. The amount of input was quantified on the basis of the number of hours of exposure to the HL during a typical week and over a year, including the periods of time spent in Italy, in which children have access to a greater amount of input in a variety of contexts and with different speakers. The quality of the input was measured on the basis of the analysis of vocabulary, MLU and output. The results showed that Costanza was exposed to Italian more frequently than the other two children. The qualitative analysis also showed that Costanza’s mother provided the most qualitatively rich input. On the basis of the factors analysed, Costanza was therefore the child who was exposed to the highest quantity and quality of input. Francesca and Paolo differed in the quantity and quality of input: Francesca was exposed to Italian more frequently than Paolo, but her mother’s output, vocabulary and MLU scores were on average lower. The initial assumption was that exposure to limited input would have an impact on the development of the HL, as shown in previous studies. When analysing bilinguals, language-external factors need to be accounted for, especially because the HL is likely to be spoken by a minority (sometimes just by one parent) and the child is likely to use it less frequently and in a more limited range of contexts than the majority language. The bilingual children who participated in this study developed Italian at different rates, and their language showed features that are uncommon in normally developing monolingual children. In terms of rate of acquisition, Paolo and Francesca reached an MLU of 1.5-2 respectively at 3;3 and 2;6, while Costanza reached it at 1;11. Moreover, Costanza reached the multiword stage sooner than Paolo and Francesca, showing that her morpho-syntactic development was faster and more native-like. The results from the analysis of vocabulary show that Costanza’s and Francesca’s lexical richness rapidly increased, but Paolo’s remained stable. The errors, which included morphological and syntactic deviant forms, were more frequent in Francesca’s data. Francesca displayed particular difficulty with verbal morphology. The morphological errors she made are not uncommon in monolinguals, but they tend to persist for a longer time in some bilinguals. These fossilization errors can be attributed

120

Chapter Six

to the lack of access to positive evidence which would allow the child to acquire the correct form in a limited amount of time. This example shows that children need to be exposed to sufficient input in order to acquire the HL fully and at a normal rate. The quality and quantity of interaction in each language is essential for the child’s linguistic development. But how much input does a child need in order to develop and maintain a HL? The analysis of the data from three bilingual Italian-English children has shown that Costanza, who was exposed to the highest quantity and quality of input, rapidly progressed from the two-word to the multi-word stage, her vocabulary developed at a normal rate, her production of target-deviant forms was minimal and she consistently used Italian. Therefore, Costanza represents evidence of a correlation between frequent exposure to qualitatively rich input and faster development rate. The other two children show interesting results. Francesca, who was exposed to a qualitatively poorer input made many errors (some of which were frequently repeated) and switched to English more often than the other two children. Costanza’s Italian seems to be more similar to that of a monolingual child, while Francesca and Paolo produce sentences exhibiting cross-linguistic influence. This study constituted an initial step in the examination of the relationship between input and HL development. It aimed to gain understanding of what could cause the difficulties HL speakers encounter, focusing on the role of the input, and providing an overview of some interesting features which characterise the early development of Italian as a HL. To conclude, it has emerged that the access to input can be one of the main factors affecting variable success in HL acquisition during childhood and language maintenance in adulthood, as argued by Rothman: While there is little doubt that incomplete acquisition and attrition exist and are explanatory for many native/heritage differences, it would take longitudinal studies that are currently unavailable to really corroborate such assumptions, let alone distinguish properly between attrition (what was acquired and then lost), incomplete acquisition (what was available but simply not acquired) and issues related to input type (what could not be acquired given its absence in the input). (Rothman 2007: 361-362)

HL speakers represent an interesting challenge for bilingualism research. The many studies emerging in this area have demonstrated that identifying features typical of HL development is essential for the evaluation of the children’s needs, both in terms of language instruction and of assessment of language pathologies.

The Role of the Input in the Early Development of a Heritage Language 121

Notes 1

This research was funded by the Irish Research Council for the Humanities and Social Sciences (IRCHSS). 2 Francesca also uses vuoi appropriately for second person marking. 3 This fact is confirmed by the parents and it is also documented in the last recording session, during which Costanza goes to play by herself in her tent and talks to her toys in Italian.

References and Further Reading Akhtar, N., and Tomasello, M. 2000. The social nature of words and word learning. In: R. M. Golinkoff, et al. (eds.), Becoming a word learner: A debate on lexical acquisition. Oxford: Oxford University Press, pp. 115–135. Benmamoun, E., Montrul, S., and Polinsky, M. 2010. White Paper: Prolegomena to heritage linguistics. Harvard University. Bonnesen, M. 2009. The status of the “weaker” language in unbalanced French/German bilingual language acquisition. Bilingualism, Language and Cognition 12 (2), pp. 177-192. Bornstein , M. H., Haynes, M. O., and Painter, K. M. 1998. Sources of child vocabulary competence: a multivariate model. Journal of Child Language 25, pp. 367-393. De Houwer, A. 2007. Parental language input patterns and children’s bilingual use. Applied Psycholinguistics 28 (3), pp. 411-424. Döpke, S. 2000. The weaker language in simultaneous bilingualism: why it is not like L2. Unpublished Manuscript, Monash University. Gathercole, V., and Thomas, E. M. 2009. Bilingual first-language development: Dominant language takeover, threatened minority language take-up. Bilingualism, Language and Cognition 12 (2), pp. 213-217. Hoff, E., and Naigles, L. 2002. How children use input to acquire a lexicon. Child Development 73 (2), pp. 418-433. Juan-Garau, M., and Pérez-Vidal, C. 2001. Mixing and pragmatic parental strategies in early bilingual acquisition. Journal of Child Language 28 (1), pp. 59-86. Kagan, O., and Dillon, K. 2008. Issues in heritage language learning in the United States. In: N. Van Deusen-Scholl and Hornberger, N. H. (eds.), Encyclopedia of Language and Education (2nd ed., Vol 4), pp. 143156. New York: Springer.

122

Chapter Six

La Morgia, F. 2011. Bilingual first language acquisition: The nature of the weak language and the role of the input. Unpublished PhD Thesis. Dublin City University. Montrul, S. 2008. Incomplete acquisition in bilingualism. Re-examining the age factor. Amsterdamo: John Benjamins. Pan, B. A., Rowe, M. L, Singer, J. D., and Snow, C. E. 2005. Maternal correlates of growth in toddler vocabulary production in low-income families. Child Development 76 (4), pp. 763-782. Pancsofar, N., and Vernon-Feagans, L. 2006 Mother and father language input to young children: Contributions to later language development, Journal of Applied Developmental Psychology, 27 (6), pp. 571–587. Paradis, J. 2009. Oral language development in French and English and the role of home input factors. Report for the Conseil scolaire centrenord, Edmonton, Alberta. —. 2011. Individual differences in child English second language acquisition. Comparing child-internal and child-external factors. Linguistic Approaches to Bilingualism 1 (3), pp. 213-237. Polinsky, M., and Kagan, O. 2007. Heritage languages: In the 'wild' and in the classroom. Language and Linguistics Compass 1(5), pp. 368–395. Rothman, J. 2007. Heritage speaker competence differences, language change, and input type: inflected infinitives in Heritage Brazilian Portuguese. International Journal of Bilingualism 11 (4), pp. 359389. —. 2009. Pragmatic deficits with syntactic consequences?: L2 pronominal subjects and the syntax-pragmatics interface. Journal of Pragmatics 41 (5), pp. 951-973. Schlyter, S. 1993. The weaker language in bilingual Swedish-French children. In: K. Hyltenstam and Vilberg, A. (eds.), Progression and regression in language. Cambridge: Cambridge University, pp. 289-308. Valdés, G. 2000. The teaching of heritage languages: an introduction for Slavic-teaching professionals. In: O. Kagan and Rifkin, B. (eds.), The learning and teaching of Slavic languages and cultures. Amsterdam: John Benjamins, pp. 375–403. Vigil, D. C., Hodges, J., and Klee, T. 2005. Quantity and quality of parental language input to late-talking toddlers during play. Child Language Teaching and Therapy 21, pp. 107-122. Windsor, J., Glaze, L. E., and Koga, S. F. 2007. Language acquisition with limited input: Romanian institution and foster care. Journal of Speech, Language, and Hearing Research 50 (5), pp. 1365-1381.

The Role of the Input in the Early Development of a Heritage Language 123

Yamamoto, M. 2001. Language use in interlingual families: A Japanese– English sociolinguistic study. Clevedon: Multilingual Matters. Zammit, M., and Graham, S. 2011. Maternal label and gesture use affects acquisition of specific object names. Journal of Child Language 38, pp. 201-221.

Exercises 1. This paper argues that input has a significant impact on the acquisition of a heritage language. Consider how linguistic and extra-linguistic factors may affect the success in acquiring a language that is spoken only at home. 2. Discuss possible alternative ways to measure the quantity and quality of input. 3. This study focuses on parent-child interaction. Do you think that the educational setting also plays a role in the child’s language development and use? Discuss.

PART II: LANGUAGE SHIFT AND LANGUAGE MAINTENANCE

EDITOR’S INTRODUCTION TO PART TWO

Language shift results when speakers abandon their language, willingly or under pressure, in favour of another language, which then takes over as their means of communication and socialisation. The papers in this section explore the causes and consequences of language shift and measures taken to maintain languages. The discussions and analyses are based on data from locales in Africa, Asia, Europe, and Central America. In the opening paper of this section, Herman Batibo highlights that language contact between the indigenous languages of Africa has been a common phenomenon for many years, given the many movements that took place often motivated by search for greener pastures, political conflicts, demographic pressure and personal reasons (Nurse 2000). As a result, languages of all sizes, dynamism, prestige and socio-economic standing came into contact and co-existed for many years. However, this cohabitation often gave rise to different forms of language relationship, namely adjacency, expansion, domination, integration and intermixture. The paper also demonstrates that, although it is common for languages to co-exist in one community or territory, language shift depends largely on the type of language contact that prevails between the languages that are in contact. He argues that language shift occurs mainly in situations, where there is expansive or integrative language contact situation. Migration is one of the major factors influencing language maintenance and shift. A migrated language is frequently faced with dominant varieties, which motivates language shift with varying results. In her paper, Ghosh presents a study of the Bhojpuri community in Kolkata. Bhojpuri is one of the various linguistic communities with a long history of migration to Kolkata. Bhojpuri speech, which is historically classified as a close sister to Bengali, is also generally considered to be a dialect of Hindi and recognised in the Indian Census as such. Bhojpuris in Kolkata are therefore in regular contact with three dominant languages: Hindi, one of the official languages of India and considered to be the standard variety of a number of speech forms including Bhojpuri; Bengali, one of the 22 scheduled languages of India, the local official language of the state of West Bengal, the language of the majority in the city and it also happens to be linguistically close to Bhojpuri; and English, the other official language of India, the former colonial language with undeniable significance

Editor’s Introduction to Part Two

127

and prestige in everyday life of an average city-dweller in an Indian Metropolis. Perta’s paper begins by analysing the vitality of a minority language Francoprovençal - through describing the speakers’ competence and use of the language, and an exploration of language distribution in the social repertoire of two communities where Francoprovençal is spoken. In her paper, Philip describes the role of attitudes and identity as factors associated with the maintenance and shift of Limón English spoken by 127 Afro-Costa Rican women in two bilingual communities in the province of Limón, Costa Rica. She treats language maintenance and language shift from a sociolinguistic perspective to explain how the women in her study perceive Limón English and their need to maintain or cease speaking it. The final paper in this section by Shihan de Silva Jayasuriya explores the factors that have contributed to the longevity of Sri Lanka Portuguese Creole (SLPC). She theorises that trades of the mother-tongue speakers, who ascribed importance to their Portuguese identity and the significance of music as a centripetal force to bind the community have all helped in the longevity of this language. She also offers detailed linguistic descriptions of the language so to establish that SLPC is a language in its own right apart from European Portuguese the lexifier language. From these papers, the reader can infer the role of attitudes in language maintenance, just as in language shift is critical because the speakers of a language hold the key to the continuation or abandonment of their language; the inter-transmission or dis-transmission of the language to their children; the expansion or reduction of the domains in which it is used. However, the speakers’ attitudes depend heavily on the status and prestige of their language. Such prestige results from their perception of its symbolic or utilitarian value.

References and Further Reading Anderson, B. 1991. Imagined communities. London/New York: Verso. Anonby, S. J. 1997. Reversing Language Shift:Can Kwak'wala be Revived? Unpublished M.A. thesis, University of North Dakota. Brandt, E. A., and Youngman, V. A. 1989. Language renewal and language maintenance: a practical guide. Canadian Journal of Native Education 16 (2), pp. 42-77. Dauenhauer, N. M., and Dauenhauer, R. 1998. Technical, emotional and ideological issues in reversing language shift: examples from Southeast Alaska. In: L. A. Grenoble and Whaley, L. J. (eds.), Endangered

128

Editor’s Introduction to Part Two

Languages: current issues and future prospects. Cambridge: Cambridge University Press, pp. 57-98. Dorian, N. C. (ed.) 1989. Investigating Obsolescence: Studies in Language Contraction and Death. Cambridge: Cambridge University Press. Douglas, E. T. K. 1992. Maori Language Nests (Kohanga Reo) - Their Impact on New Zealand Communities. Journal of Indigenous Studies 3 (1), pp. 13-31. Edwards, J. 1985. Language, society and identity. Oxford: Blackwell. Fishman, J. A. 1991. Reversing Language Shift: Theoretical and Empirical Foundations of Assistance to Threatened Languages. Philadelphia: Multilingual Matters. —. 1997. Maintaining Languages: What Works? What Doesn't? In: G. Cantoni (ed.), Stabilizing Indigenous Languages. Flagstaff: Northern Arizona University, Center for Excellence in Education. pp. 186-198. Fleras, A. 1989. Te Kohanga Reo: A Maori renewal program in New Zealand. Canadian Journal of Native Education 16 (2), pp. 78-85. Harrison, K. D. 2006. When languages die: the extinction of the world’s languages and the erosion of human knowledge. Oxford: Oxford University Press. Hinton, L. 1994. Flutes of Fire. Berkeley, California: Heyday Books. Rigsby, B. J. 1987. Indigenous Language Shift and Maintenance in Fourth World Settings. Multilingua 6 (4), pp. 359-378.

CHAPTER SEVEN LANGUAGE SHIFT AS AN OUTCOME OF EXPANSIVE AND INTEGRATIVE LANGUAGE CONTACT IN AFRICA HERMAN BATIBO

1. Introduction The process of language shift from one language to another is very widespread in Africa, where languages are not only in constant contact, but also differ in size, status, prestige and function (Brenzinger 1992; Smieja 1996, 1998; Sommer and Vossen 2000). However, the process of language shift correlates with the phenomenon of language endangerment, given that when speakers of a given language shift to another language, the language they spoke becomes distinct. According to Batibo (2005), the position of language endangerment in Africa is as shown in Table 7-1. Category of language

1 2 3 4

Relatively safe Relatively endangered Severely endangered Extinct or nearly extinct Total

Number of languages

Percentage of category

336 1,287 308 201 2,132

15.8% 60.4% 14.4% 9.4% 100%

Table 7-1: The position of language endangerment in Africa (after Batibo 2005:155) The indicators of critical language endangerment include the development of negative attitudes towards ones language, ambivalent language loyalty, association of mother-tongue with low economic and

130

Chapter Seven

social status inactive language transmission to children, reduction of domains of use, diminishing number of speakers, limited stylistic variation, structural erosion and simplification, and lexical reduction (Batibo, 2005:65). Usually, language shift takes place in favour of the more dominant, prestigious or socio-economically viable language, which acts as an attraction. The aim of this paper is to demonstrate that, although it is common for languages to co-exist in one community or territory, language shift depends on the type of language contact that prevails between languages that co-exist. It is argued in this paper that language shift occurs mainly in situations where there is expansive or integrative language contacts.

2. Types of Language Contact in Africa Through constant migration and expansion in search of better environments for grazing, cultivating, hunting and other socio-economic activities, people tend to move and come into contact with other groups. Language contact has therefore taken place in Africa on regular basis, between the indigenous languages themselves, even before the dawn of colonialism (Nurse 2000; Thomason and Kauffman 1988). These language contacts were of several types. The first type of language contact was that of language adjacency, that is, of languages which existed side by side, in the same territory or area. Such contact was said to be interactive, as the speakers of these languages interacted and learnt each other’s language and became bilingual, just to be able to interact as neighbours and collaborate or share in their social, economic and cultural activities. Usually, the speakers of these languages were relatively equal in socio-political, economic or demographic strengths. Only few members of these linguistic groups were attracted to shift to the other languages. This would happen where intermarriages or family adoptions took place. This type of contact was very common. The second type of language contact occurred when groups migrated to settle in other territories, which were already inhabited by other much larger groups. This type of encounter could be described as integrative contact, given that the migrant groups were usually weaker demographically or socially. Often they were fugitive or off-shoot groups looking for greener pastures or safer habitation. Such groups were, in most cases, absorbed or integrated into the already existing groups in the area, leaving behind sub-stratum elements or no traces at all, except their ethnonyms or some vague historical or cultural accounts.

Expansive and Integrative Language Contact in Africa

131

The third type of contact situation took place when predominant groups expanded into wider territories, for socio-political, economic or cultural reasons or to ease demographic pressure, subjugating the other groups on their way. This happened with large languages, such as Arabic, Hausa, Peul, Kiswahili, Wolof, IsiZulu or Amharic. The smaller groups were forced to learn the dominant languages and often used them as lingua franca or language of wider communication. The language use pattern that prevailed could be described as diglossic, as the dominant language would be at the top with a High (H) status, and prestige, while the subjugated language would be at the bottom with the lower (L) status, having less status and use in public functions. The fourth type of contact involves communities which intermingled through interaction or marriage, thus, resulting in mixed or hybrid languages. This is the case of Mbugu in East Africa, which is a mixture of Pare (a Bantu language) and Ma’a (a Cushitic language) (Mous 1994). Moreover, through the southern expansion of Kiswahili, some mixed languages have emerged between Kiswahili and some of the Mozambican Bantu languages, creating hybrids like Ekoti and Kimwani (Schadeberg 2004a, 2004b). Also, many of the Central Khoesan languages, such as /Gwi, //Gana and Naro, can be considered as hybrids of Khoe and San groups which have intermixed through marriage and interaction over the last 4,000 years, in which they have been in constant contact and interaction. The last type of contact took place when groups of people came into contact with other groups with a special goal, such as trade, labour, tourism or evangelism. The two parties had no common language, but, in order to be able to communicate, a form of communication had to be established, involving gestural signs, noises and simple or generalized words, without grammar or complex sounds. This was the genesis of Pidgin forms, like Fanagalo and Tsotsitaal in South Africa or English Pidgin in Nigeria. Some of these forms developed to become complex linguistic systems with mother-tongue speakers. Hence they became fullfledged languages, known as Creoles. In Africa, a number of languages have evolved as Creoles. They include Krio, in Sierra Leone, and Afrikaans in South Africa. Language contact has therefore been a regular feature in all parts of Africa (Nurse 2000). The interactive, expansive and integrative language contacts were the most common and they have given rise to many outcomes, including multilingualism, multiculturalism, language dominance, marginalization and language endangerment.

132

Chapter Seven

3. Expansive and Integrative Contacts as the Most Dynamic Processes It is argued in this study that the expansive and integrative types of language contact are the most common causes of language shift in Africa. The expansive language contact usually involves a major language expanding into the territory of another language. Such a language would normally have demographic, socio-political, economic or technological superiority over the language found in the area. Even where no military conquest takes place, the speakers of the language that is found in the area will not only be dominated or pressurized by the newly arrived language, but also attracted to shift to it. This is what is happening in most African countries where the 336 nationally or regionally dominant languages, like Arabic, Kiswahili, Chishona, Hausa, Yoruba, Igbo, Wolof, Akan, Amharic, Chishona, Chinyanja, IsiZulu, IsiXhosa, Setswana, Dinka, More-Mossi, Kituba and Somali are expanding across their respective countries or territory, while the areally dominant languages, like Umbundu, Kimbundu (Angola) Duala Bamileke (Cameroon), Gbaya, Banda (Central African Republic), Lingala, Ciluba, Kikongo (DRC), Fang (Gabon), Gikuyu, Luo (Kenya), Chiyao, Chitumbuka (Malawi), Emakhuwa, Tsonga (Mozambique), Serer-Sine (Senegal), Shisukuma (Tanzania), Kabiye (Togo), Sindebele (Zimbabwe), Tiv, Ibibo, Izon, Kanuri, Efik (Nigeria), Tigrinya, Tigre (Ethiopia) and so on, are expanding in their respective zones, as areal languages. In fact, one can talk of a quadriglossic situation in many African countries (instead of the classic triglossic model proposed by AbdulazizMkilifi, 1978), in which, ex-colonial languages are at the top (H) as official or global languages, then the major nationally dominant languages would follow at a lower (L) status in relation to the ex-colonial languages, but at a higher (H) status compared to the areally dominant languages. On their part, the areally dominant languages would be of lower (L) status, compared to the nationally dominant languages, but predominant in relation to the minority languages (L). This quadriglossic structure is shown in Table 7-2.

Expansive and Integrative Language Contact in Africa

133

Language status H

Ex-colonial language as Official language

L

Nationally dominant language as National language

H

H

Areally dominant language as Areal language

L

L

Localized language as Minority language

Table 7-2: The quadriglossic structure of languages in Africa (after Batibo, 2005) As a result of this marked language status hierarchy, a gradual language shift process is taking place, particularly with the languages at the lowest level, which in most cases have no any public functions. Hence, little or no active intergenerational transmission of the languages is happening. This is because of the strong pressure and attraction from both areally and nationally dominant languages, as also remarked by Mufwene (2005). Moreover, the integrative type of contact usually involves small groups that have migrated to live in other territories, normally as run-away communities, whose movement is motivated by political, economic or personal reasons. In most cases, they would be in a weak position and would be ready to be absorbed or integrated in other communities. This is the case that Dorian (1977) has described as suicide language death, given that the relevant speakers would abandon their language for strategic reasons (Winter 1979). Sometimes substratum elements would be infused in the dominant language as in the case of the Khoekhoe or Nama communities in southern Africa, which disappeared, but left behind rich fauna and flora vocabulary and a set of clicks as sub-stratum elements in the contact Bantu languages, like IsiXhosa, IsiZulu and Sesotho. But no traces of Otjiherero were left in the Ovaherero communities which shifted to Nama language in the southern Kgalagarhi area of Botswana, except for their ethnonym and autonymic traditions which they have preserved. However, some migrant groups have managed to maintain their linguistic and cultural identity. This is mainly due to strong cultural attachment, high sense of identity or strong resistance to integration into the mainstream groups. This is the case of the Zanzibari group in Durban, South Africa, whose members are highly attached to their culture and Islamic way of life. Although they are only about 1,000, they have

134

Chapter Seven

managed to maintain their language, traditions and religion from their original home in northern Mozambique. This is also true of the Indian communities in South Africa, which have maintained their languages, culture and traditions, due to their strong cultural and identity attachment (Mestrie 2007). Equally, the Sonjo of northern Tanzania (Nurse and Rottland 1991/92) have managed to preserve their language and culture in the heart of Maasai domination. Such cases are rare, given the overwhelming pressure of the already existing languages. The first contacts involved the major language families, such as the spread of Niger-Congo in West Africa and its Bantu branch in the south of the equator. The expansion of the Niger Congo family displaced earlier groups in both West and Central Africa. Equally, the Bantu groups absorbed or displaced the groups from the other African language families, including the Khoesan, Pygmy, Nilotic and Cushitic communities. Some contacts in the earlier times were triggered by movements, such as the Maghrebian conquest during the spread of Islam in the 6th Century, in which the speakers of Egyptian and Berber languages were eliminated or absorbed in Arabic, which was introduced in the region. Also, many movements took place in southern Africa during the regional conflicts, such as the Chaka Zulu and Mfacane wars which displaced many communities and languages.

4. The Process of Language Shift Language shift usually presupposes that the people affected are bilingual in their own languages and in the language to which they are shifting. Given the unequal status of the two languages, the speakers become more attracted to shift to the new language or to transmit the new language to their children, so as open more socio-economic opportunities for them. Usually the parents would develop negative attitudes towards their own languages, as shown in Table 7-3 of a study carried out in Botswana (Batibo, 1998).

Expansive and Integrative Language Contact in Africa Respondent’s preferences 1

2

3

4

5

6

Would prefer to use mother tongue for family interaction Would prefer to use mother tongue for village communication Would prefer to use mother tongue exclusively at school Would prefer to use mother tongue with Setswana and English at school Would prefer to be called by the ethnic name When adult will transmit mother tongue to own children

135

Set.

Otj.

Thi.

She.

Ika.

Shi.

Kho.

96.0

81.0

77.2

58.9

83.3

57.3

38.2

90.4

76.8

74.6

55.4

84.6

30.1

10.4

93.0

56.0

47.2

28.3

65.6

27.4

3.9

100.0

82.3

70.5

63.2

80.8

46.4

7.5

98.7

95.4

87.9

69.4

93.5

67.5

74.3

100.0

90.1

87.5

66.0

88.4

21.4

14.8

Table 7-3: Attitudes towards the use of mother tongue by the various ethnic groups (in % of total number of respondents in each group) (after Batibo, 1998) From the figures manifested by Table 7-3 above, it is clear that the respondents from communities that speak nationally dominant language, like Setswana and areally dominant languages like Ikalanga, Otjiherero and Thimbukushu, have generally positive attitudes towards their languages, as the majority of them would like to see their mother tongue used at home, in the village and in the school. Such communities would like as future adults, to transmit their languages to their children. On the other hand, only a small percentage of the respondents from the communities which speak minority languages like Shekgalarhi, Shiyeyi and the Khoesan languages, wanted to see their languages used in the family, in the village and in the school or be transmitted to their children. What was remarkable in this study is hat all groups preferred to be known by their ethnic names. The need to preserve ethnic names, even where there is linguistic and cultural loss, has been remarked in a number of

136

Chapter Seven

similar studies, such as Batibo (1998, 2008), Chebanne and Nthapelelang (2000), Nyati-Ramahobo (2000), and Vossen (1988). At the same time, most of the respondents were more favourable to the idea of bilingual education in which the mother tongue would be used alongside Setswana and /or English rather than the use of mother tongue alone. The process of language shift therefore involves a progressive shift from one language to the other. The first shift concerns language, in which A acquires B in order to enable him or her to communicate and have access to the opportunities of subject B, such as education and employment. Then subject A will acquire the cultural features of subject B in order to be accepted as part of the relevant community. The next step is for subject A to adopt the name of subject B or alternatively to name his children the names of B so as to authenticize himself in the new community for identification purpose. Finally, he or she will abandon his ethnic identity in favour of that of B. This is normally done as a last resort in order to A to fully integrate in community B. In this way, Subject A will have lost all his original traits to fully shift into the new linguistic community (Lamy 1979; Pool, 1979).

5. Conclusion As rightly observed by Mufwene (1997, 2005), it is the major indigenous languages, which are the major causes of language shift in Africa, this is because they are the ones that are in the state of expansion to the grassroots of the affected communities, and are also responsible for the integration of many of the migrant languages. According to the Marked Bilingualism Model proposed by Batibo (2006), language shift is accelerated in situations where there is a sharp gradient of imbalance between the dominant and the dominated language. Given that many of the nationally and even the areally dominant languages have been accorded socio-political status, as national, official or languages of local or lower education, they have multiplied their weight and attraction, as languages. Unless some public domains are accorded to the minority languages, they will continue being vulnerable, as their speakers will see them as entities that have no utilitarian value, as they are not used in key areas like education, administration, judiciary, the media or in accessing job opportunities (Batibo 2001; Chebanne 2004; NyatiRamahobo 2004; Visser 2000).

Expansive and Integrative Language Contact in Africa

137

References and Further Reading Abdulaziz-Mkilifi, M. H. 1978. Triglossia and Swahili-English bilingualism in Tanzania. In: J. A. Fishman (ed.), Advances in the Study of Societal Multilingualism. pp. 129-149. The Hague: Mouton de Gruyter. Batibo, H. M. 1998. The fate of the Khoesan languages of Botswana. In: M. Brenzinger (ed.), Endangered Languages in Africa. pp. 267-284. Frankfurt/Main: Peter Lang. —. 2001. Empowerment of minority languages for education and development. In: R. Trenby and S. Fitchat (eds.), Language and Development in Southern Africa: Making the Right Choice. pp. 123135. Windhoek: Gamsberg-Macmillan. Batibo, H. M. 2005. Language Decline and Death in Africa: Causes, Consequences and challenges. Cleverdon: Multilingual Matters. —. 2006. The Bilingualism Markedness Model. In Malilime. 1:1 – 26. —. 2008. Circumstances of language shift and death in Southern Africa. In: C. Vigouroun and S. Mufwene (eds.),Globalization and Language Vitality: Perspectives from Africa. pp. 51 – 69. New York: Continuum International Publishing Group. Brenzinger, M. (ed.). 1992. Language Death: Factual and Theoretical Explorations, with Special Reference to East Africa. Berlin: Mouton de Gruyter. Chebanne, A. 2004. Language policy and ethnicity in Botswana: National policy and relationship of autochthonous linguistic minorities in educational and cultural practices. University of Duisburg-Essen, Series A: General and Theoretical Papers, no. 589. Chebanne, A., and Nthapelelang, M. 2000. The sociolinguistic survey of the Eastern Khoe in Boteti and Makgadikgadipans area of Botswana. In: H. Batibo and B. Smieja (eds.), Botswana: The Future of the Minority Languages. pp. 79 - 94. Frankfurt/Main: Peter Lang. Dorian, N. C. 1977. The problem of the semi-speaker in language death. In: W. U. Dressler and R. Wodak-Kodotter (eds.), Language Death (International Journal of the Sociology of Language 12). pp. 23-32. Paris: Mouton Publishers Lamy, P. 1979. Language and Ethno-linguistic identity: The bilingualism question. International Journal of the Sociology of Language. 20: pp. 23–36. Mestrie, R. 2007. Language shift, cultural change and identity retention: Indian South Africans in the 1960s and beyond. South African Historical Journal 57: pp. 134 – 152.

138

Chapter Seven

Mous, M. 1994. Ma’a or Mbugu. In: P. Bakker and M. Mous (eds.), Mixed Languages: Fifteen Case Studies in Language Intervening. pp. 175– 200. Amsterdam: IFOTT. Mufwene, S. 1997. Kituba. In: S. G. Thomason (ed.) Contact Languages: A Wider Perspective. pp. 175-208. Amsterdam: John Benjamins. —. 2005. Globalization and the myth of killer languages: What’s really going on? In: G. Huggan and S. Klassen (eds.), Perspectives on Endangerment. pp. 19 – 48. Hildesheim/New York: Georg Olms Verlag. Nurse, D. 2000. Language Contact in Africa: Inheritance, Contact and Change in two East African Languages. Cologne: Ruediger Koeppe Verlag. Nurse, D., and Rottland, F. 1991/92. Sonjo: Description, classification, history. In Sprache und Geschichte in Afrika. 2/13: pp. 171 – 239. Nyati-Ramahobo, L. 2000. Linguistic and cultural domination: The case of the Wayeyi of Botswana. In: H. Batibo and B. Smieja (eds.), Botswana: The Future of the Minority Languages. pp. 217-234. Frankfurt/Main: Peter Lang. —. 2004. The language policy, cultural rights and the law in Botswana. LAUD Series on General and Theoretical Papers, no. 613. University of Duisburg-Essen.Pool, J. 1979. Language planning and identity planning. International Journal of the Sociology of Language, 20: pp. 5–25. Schadeberg, Th. C. 1994a. Ekoti: Between Kiswahili and Emokhua Ms Leiden. —. 1994b. Kimwani at the Southern Fringe of Kiswahili. In: P. Bakker and M. Mous (eds.), Mixed Languages: Fifteen Case Studies in Language Intertwining. pp. 239-145. Amsterdam: IFOTT. Smieja, B. 1996. Language shift, cultural identity and language loyalty in Botswana. (Ms) —. 1998. Language shift and its implications for minority languages in Botswana. (Ms). Sommer, G. and Vossen R. 2000. Language gain and language loss: The spread of Setswana in Ngamiland. In: H. Batibo and B. Smieja (eds.), Botswana: The Future of the Minority Languages. pp. 129-146. Frankfurt/Main: Peter Lang. Thomason, S. G. and Kauffman, T. 1988. Language contact, Creolization and Genetic Linguistics. Berkeley: University of California Press. Visser, H. 2000. Language and cultural empowerment of the Khoesan people. The Naro experience. In: H. Batibo and B. Smieja (eds.),

Expansive and Integrative Language Contact in Africa

139

Botswana: The Future of Minority Languages. pp. 193-216. Frankfurt/ Main: Peter Lang. Vossen, R. 1988. Patterns of Languages Use and Attitudes in the Ngamiland in Botswana. (Bayreuth African Studies, 13). Bayreuth: Bayreuth University. Winter, J. C. 1979. Language shift among the Aasax: A hunter-gatherer tribe in Tanzania. Sprache und Geschichte in Afrika. 1: pp. 175–204.

Exercises 1. It is argued in this study that the expansive and integrative types of language contact are the most common causes of language shift in Africa. With examples from African languages that you are familiar with, discuss the processes that occur in each type of contact and the extent to which language shift takes place. 2. Explain why the phenomenon of language endangerment is more critical in Africa than in the other continents of the world? 3. According to Salikoko Mufwene (2005), it is the major African indigenous languages, rather than the ex-colonial languages, which are the most “deadly killers” of other languages. Discuss the validity of this statement in the light of the hierarchy of language ecology in Africa. 4. Select a major African language from one part of Africa. Make a list of the smaller languages that are in contact or dominated by this language. Choose one of these languages and then consult the Ethnologue to determine the population of speakers, attitudes towards their language, patterns of language use, any signs of language shift and any other information that you may find. Using this information, describe the degree to which the smaller language is endangered.

CHAPTER EIGHT LANGUAGE MAINTENANCE AND SHIFT IN MIGRATION: A CASE STUDY OF BHOJPURIS IN KOLKATA1 ADITI GHOSH

1. Introduction Migration is a common phenomenon presently and it has obvious and undeniable influence on language. It can engender cases of borrowing, second language acquisition, creation of new language varieties, language shift and even language change. Kerswill (2006) notes that migration is one of the ‘key extra-linguistic factor(s)’ of language change. In this paper, I look into the pattern of language use and attitude of a section of a migrated community in Kolkata, namely the Bhojpuris. Kolkata is one of the most multicultural and multilingual cities of India. It attracted migrants from all over India for a very long time. According to the 1991 Indian Census 36.4% of Kolkata’s residents speak a language other than Bengali, the dominant local language. But according to a report submitted by the state government before the assembly in 2003 the number is a lot higher.2 The Bhojpuris have a long history of migrating to and settling in various parts of the state of West Bengal including Kolkata. Bengali, the most widely spoken local language, is of the same genealogical sub branch as Bhojpuri and has a great deal of linguistic similarities with it. However, Bhojpuri is understood both officially (as reflected in the census) and by the common people as a type of Hindi. Bhojpuri is classified officially as a kind of Hindi, one of the two official languages of India and in states like Uttar Pradesh and Bihar, where Bhojpuri is widely spoken, Hindi is the state official language and Khariboli, the standard Hindi, used for all official interaction and education. Bhojpuri, in these states are used for informal communications creating a diaglossic situation. Commenting on

Language Maintenance and Shift in Migration

141

this history of migration and the nature of the relationship between the local Bengalis and the Bhojpuris, Tiwari (1960, p. xxvi) comments: ‘…neither the Bhojpuri who is of a humble rank, usually going to earn a living in Bengal as a watchman or labourer or a small trader, nor the Bengali who employs him or with whom he has dealings, has any idea that their speeches, Bengali and Bhojpuri have a great deal in common. Both think that the speech of the Bhojpuri is a kind of Hindi or Hindusthani, particularly, because the Bhojpuri if literate is literate in Hindi and not in Bhojpuri as a separate language.’ Tiwari (1960, p. xxvi)

It is difficult to determine the exact number of Bhojpuri speakers in Kolkata. Bhojpuri is not recognised as a separate language in the census but as a mother tongue under Hindi and, as the above comment suggests, is understood by the general public as such. Besides, Bhojpuri is one of the many language communities of the north central zone of India which is characterised by a dynamic linguistic identity. Khubchandani (1978, 1985) calls this area ‘fluid zone’ and notes that there are at least five distinct heterogeneous linguistic groups that claim Hindi as their mother tongue including monolinguals (mostly rural) speaking regional vernaculars or languages other than Khariboli, but giving Hindi as their mother tongue, since they regard themselves as part of the great Hindi tradition (1985, p. 204). According to the census reports (1991), Hindi is spoken by 21% in Kolkata. One can assume a good number of these are Bhojpuri speakers. In this paper, I will analyse the responses of a cross section of these residents, their pattern of usage of different languages in their repertoire and their attitude towards those languages.

2. Standard Language Ideology To analyse the respondents’ attitude towards the speech varieties in their repertoire, I will use the concept of language ideology (cf., Woolard and Schieffelin 1994), in particular standard language ideology (Milroy and Milroy 1985). The standard language in this context can be understood as an idea in the mind rather than a reality – a set of abstract norms to which actual usage may conform to a greater or lesser extent. (Milroy and Milroy 1985, p. 23) Put simply, standard language ideology is a set of beliefs (e.g., of linguistic correctness, status or purity) as reflected in the attitudes (e.g., in favourable evaluation of a speech variety) shared by a community towards that idea. Such ideological positioning, as pointed out

142

Chapter Eight

by Milroy (2007, p. 133), is usually not a deliberate one, but it is understood as natural or normal. Speakers are not usually conscious that they are conditioned by these ideological positions: they usually believe their attitudes to language to be common sense and assume that virtually everyone agrees with them (ibid, p. 133). Lippi-Green (1994) cites a number of court cases where discriminatory practices are allowed under the influence of standard language ideology. Wiley and Lukes (1996) argue that adherence to standard language ideology adversely affects the education system in the United States. Milroy (2001) shows, linguists, including sociolinguists, in spite of the long tradition of vocalising equality of all language forms, are strongly affected by standard language ideology in their work. The case of Bhojpuri is important here not only because it lacks the status of a language, but also because Bhojpuri is one of the many mother tongues under Hindi that are linguistically deviant enough for linguists like Chatterji (1963, p. 111) to question their grouping as the same language. Moreover, the ideology of national language is an important factor in this case. In India, which according to census reports has 122 distinct languages spoken by at least 10,000 speakers, Hindi and English are two official languages. Twenty-two languages are listed under the eighth schedule of the Indian constitution and are accorded special official status. But no single language is picked out either in the constitution or in any act of parliament as a ‘national language’. This notwithstanding, Hindi is very widely believed or ‘imagined’ (cf., Chand 2011) to be the national language of India and English similarly is widely seen as an international or foreign language. The spread of Hindi as a second language is also considerable. Even though it has not quite acquired the prestige of English, it scores higher than English for its ‘Indianness’. Dasgupta and Fishman’s (1971) analysis of subsidiary language claims in India shows that though in overall results English is claimed as subsidiary language more often than Hindi or any other languages, since it is more functional for social mobility, among migrants a different tendency is noticed. Hindi tends to be more frequently claimed by migrants than non-migrants (Dasgupta and Fishman 1971, p. 237). The results can be different for non-migrants. For example, Prasher’s (1980) case study of educated Indian bilinguals reported ‘very little use of regional language and Hindi’ (p. 152). The responses in this study, however, reflect extensive usage of Hindi even in family and home domains and the opinions reflected by the respondents also show that they accept Hindi as truly representative of India and are quite proud to be associated with it as dialect speakers.

Language Maintenance and Shift in Migration

143

3. The Bhojpuri speech As mentioned earlier, Bhojpuri is listed in the census as one of the fifty mother tongues classified under Hindi. It is the second largest speech variety3 under Hindi with 33,099,497 speakers.4 The largest speech variety is Hindi and Khariboli, the so-called standard variety, has only 47,730 speakers according to the 2001 census data. Bhojpuri is spoken over a large area in India covering the states of Uttar Pradesh, Bihar, Jharkhand, and due to large scale migration and settlement it is also spoken widely in Assam, West Bengal and Delhi. Outside India, it is spoken extensively in Nepal, Mauritius, Fiji, Surinam, and Trinidad and Tobago. Deliberating on the relationship between language and dialect, Trudgill (2004) points out that, language can become dialects, through loss of autonomy and dialects can become languages, through the opposite process. The relationship between Bhojpuri (and some other varieties listed under Hindi) and Khariboli may well demonstrate this point. Bhojpuri is traditionally classified as one of the speech varieties of the eastern groups of Aryan languages, its closest sisters being Maithili and Magahi (c.f., Grierson 1883-87, Chatterjee 1970). The other languages genealogically close to this group are Assamese, Bengali and Oriya, three separate independent languages listed in the 8th schedule of the constitution. Magahi, like Bhojopuri, is also classified as another mother tongue under Hindi but Maithili is recognised as an independent language since 2003 and it is now one of the twenty two languages listed in the 8th schedule of the constitution. Echoing the views of Kloss (1967) and Haugen (1966), the present situation demonstrates that language is more of a politicalcultural construct than linguistic.5

4. Result of the Sample Survey: Language Use Patterns Using questionnaire based survey in Kolkata, 289 respondents who have lived in the city for no less than ten years and educated to the tenth class have been interviewed so far. These two criteria ensure that they have been in the city long enough with sufficient exposure to different dominant languages including English, which is only learned through school. Bengali is the most widely spoken dominant language in Kolkata. Apart from Bengali, the two other dominant languages for the respondents are Hindi (because Bhojpuri is officially a variety of Hindi and also because it is the frequently used communication language in multilingual situations) and English (because of the undeniable prestige that it has in the country). Only five respondents responded to the question ‘what is

Chapter Eight

144

your mother tongue?’ as Bhojpuri. The survey did not provide any definition of mother tongue to the respondents. Their responses to the question may be taken as their projected language identity in the given context. Among those who responded to this question as ‘Hindi’, eighteen have either reported that their parents’ mother tongues are Bhojpuri or they have named Bhojpuri as one of preferred languages while communicating with family members. The sample for this study consists of these 23 respondents. All the respondents are residents of Kolkata for at least ten years. Sixteen of them were born and brought up in Kolkata. However, as it is common in the case of many in-migrants, most of them are in regular contact with the place of their origin. Ten of the respondents are male, twelve female in the age group of 18 to 42 with diverse occupational backgrounds (students, teachers, businessmen and housewives). As reported by the respondents, the overall pattern of the preferred language for communicating at home is as follows: Hindi and Bhojpuri are used almost equally in the home domain as can be seen in Figure 8-1. 50

46

43

40 30 20

11

10 0 Bhojpuri

Hindi

other

Figure 8-1: Language preference in the home domain

Bhojpuri is used in no other domains. In the formal domain also Hindi is used most frequently –68.2%, followed by Bengali 17.6% and English 13.46% (as in Figure 8-2 below).

Language Maintenance and Shift in Migration

80 70 60 50 40 30 20 10 0

145

68.2 Bengali Hindi English

17.6

Bhojpuri

13.46 1.9

rs

others

ot

he

ri pu oj Bh

is h gl En

Be

Hi

ng

al

nd

i

i

0

Figure 8-2: Language use in the formal domain

In other domains such as friend and neighbourhood, Bengali, the locally dominant language is used more frequently at 38.11% though Hindi is mostly used at 52.9%. 60 50.9 50 40

38.11

30 20 10

6.5

4.5 0

0 Bengali

Hindi

English

Figure 8-3: Language use in other domains

Bhojpuri

others

Chapter Eight

146

The language preferences while interacting with different members of the family or within the home is as follows: Total Grand father N % Grand mother N % Father N % Mother N % Brother N % Sister N % Husband N % Wife N % Son N % Daughter N % Domestic Help N %

Bhojpuri

Hindi

English

Bengali

Other

13 100

12 92.31

01 7.69

0 0

0 0

0 0

18 100

16 88.89

2 11.11

0 0

0 0

0 0

21 100

11 58.38

10 47.62

0 0

0 0

0 0

22 100

14 63.64

08 36.36

0 0

0 0

0 0

20 100

7 35

11 55

2 10

0 0

0 0

18 100

8 44.44

10 55.56

0 0

0 0

0 0

4 100

0 0

3 75

0 0

1 25

0 0

3 100

1 33.33

2 66.67

0 0

0 0

0 0

4 100

0 0

3 75

1 25

0 0

0 0

3 100

0 0

2 66.67

1 33.38

0 0

0 0

16 100

3 18.75

7 43.75

0 0

6 37.50

0 0

Table 8-1: Most preferred language for communication One can notice a gradual decrease in the preference for Bhojpuri with generation for example twenty eight respondents out of thirty one reported that Bhojpuri is the preferred language of communication while interacting

Language Maintenance and Shift in Migration

147

with their grandparents. But one out of the seven respondents (who have children) reported it as a preferred language for communicating with their children. At the same time one can notice an increase in the choice of Hindi while interacting with children. This is demonstrated in Figure 8-4. Bhojpuri Hindi

100

92.31 88.89

80

75

60

58.38 47.62

40

0

er

n

r

Da

ug

ht

So

Si

ot

he

ste

r

r he ot

M

Br

r

Fa

th

he

er

ot

at h

dm

df

0

Gr

an

an

er

7.69 11.11

0

66.67

55.56 44.44

55 36.36 35

20

Gr

63.64

Figure 8-4: Preference for Bhojpuri and Hindi across generations

There also seems to be slightly high preference for Bhojpuri while communicating with the female members of the family though this trend is not maintained in case of reported preferred language for communicating with grandmother (see Figure 8-5). 92.31 88.89

100 80 60 40 20 0

58.38 63.64

er

n ug

ht

So Da

er st Si

r he ot

he ot

th

r

er M

Br

nd

0

G

ra

Fa

m

fa

ot

th

he

r

er

0

nd ra G

44.44

35

Figure 8-5: Comparative language preference for Bhojpuri among male and female members in the family

148

Chapter Eight

5. Result of the Sample Survey: Attitude towards Language During the interviews, the respondents were encouraged to give their opinion on different languages and language communities in Kolkata. Some common expressions and views were reflected in their opinions. x The most common metaphor used for Bengali was ‘sweet’. x The same adjective is used for Bhojpuri by a section of the respondents along with terms like ‘home’ ‘our own’, and so on. Another group preferred to distance themselves from the language saying that they do not speak it well. It may be mentioned here that since Bhojpuri is not an officially recognised independent language, there is no scope for learning the language formally in school. Except five, all the respondents had Hindi as their first language at school, four had English and one had Bengali as first language. x For Hindi, the recurrent point of discussion was that of a national language, a language that can ‘connect’ India. x For English, having accepted its usefulness, almost everyone expresses their displeasure about the position that English enjoys on the ground that it is ‘not ours’. To demonstrate these attitudes, I present here a few excerpts from some of the interviews. Respondent no 1 is a 26-year-old male who is a third generation Bhojpuri speaker in Kolkata. Like many other Bhojpuri speakers in the city, they are in regular contact with their village. He reasoned that though he grew up in Kolkata and is working there now, he does not consider himself a migrant, as the village is still “alive” in him. He spoke emotionally about Bhojpuri. But later he complained about Bhojpuri language community’s unnecessary attachment to the language. dekhiye jab [...] aapni maatri bhaashaa se pyar ho jaataa hai, aur wo jivan paryant rehta hai… …mujhe laagtaa hai ki mere native language community aaj bhi aapni bhaashaa ko pakad ke kuch is kadar baithi hui hai aur usme kuch haad tak jadtaa itni zyadaa hai ki wo baahaar nahi nikaalnaa chaahte[… ]unhe aapni bhaashaa se prem karne ke alawaa aur tamaam bhaashaao ko sikhnaa caahiye aur us jadtaa se baahaar nikaalnaa caahiye…6 ‘when (one) […] falls in love with his mother tongue, it stays for life. … I think that people of my language community are so caught up with

Language Maintenance and Shift in Migration

149

their language that they have some uneasiness from which they do not want to get out […] apart from loving their own language they should learn all the languages and come out of that uneasiness ’.

He used the often-repeated ‘sweet’ metaphor for Bengali language. Bengali mujhe bahut meethi zabaan laagti hai, bahut meethi bhaashaa laagti hai isiliye bahut chesthaa karke sikhi hai. ‘I feel Bengali is a very sweet language so I learned (it) with some effort’

He echoes the resentment of many other respondents’ towards English. English se mujhe […] chadma-superiority […] dikhaanaa chaahte hai ki main bahut jaankaar hu. […]to mujhe thodi si vyaktigat problem is tarike par hai. … ‘English, it seems, has a ‘fake superiority’ (associated with it).(they) want to show that “I know a lot” […] so I have a personal problem with this way.’

For Hindi, he discusses its value for everyday communication and, like many more respondents, its importance in national integration. daraasal Hindi vyavahaar ki bhaashaa hai […] dinchaarya ki bhaashaa hai. roj ki.… mujhe laagtaa hai ke Hindi me ek shakti hai[…] pure desh ko ek saath baandhaa jaa saktaa hai. raashtriya bhaashaa kaa ye kartavya hotaa hai ki wo pure desh ko ek sutra me baandh karke raakkhe, aur iske liye mujhe Hindi zyadaa samarth laagti hai. ‘Actually Hindi is a language of use [..] everyday language …I feel that Hindi has this power to tie a nation together. It is the duty of a national language that it ties the entire nation together, and I feel Hindi is best suited for this.’

Respondent no 2 is a 21 year old student who was born and brought up in Kolkata. He said he considers West Bengal as his homeland and wants to settle here in future if he had a choice but also pointed out that he faces language based discriminatory behaviour in the city. English nahi jaante hai […] ek alag tarike se dekhte hai ‘if you don’t know English, they look at you differently’

150

Chapter Eight

For Bhojpuri, he shows a lot of emotional attachment and regrets that he cannot use it as much as he wants to. aachaa hai, bahut aachaa laagtaa hai bolna me. aapnaapan laagtaa hai. aapno se baat karte hai to bahut aachaa laagtaa hai. aagar mujhe chale to main Bhojpuri me har kaam karu lekin […] [smiles] chaltaa hi nahi hai. ‘Very good. Feels good to speak it. Feels like it is your own. When one talks with one’s own, it feels very good. If it were up to me, I will do everything in Bhojpuri but […] it does not work’

He did not demonstrate any negative views against English and emphasised its usefulness. … English to international language hai. ye to hameshaa har jagaah kaam degaa. matlab ki India me bhi aur out of country me bhi. to jaannaa zaruri hai ye. ‘English is an international language. It works all the time everywhere. I mean in India and out of the country. So it is important that you know this.’

Like many others, he linked Hindi with India. (he used the word ‘Hindustan’ instead of ‘Bharat’ for India here). Hindi to aapni bhaashaa hai. […]Hindustan me rehne ke liye to Hindi zaruri hai.. ‘Hindi is our language[…] if you are staying in India, then Hindi is important’

Respondent number 3 is a 26-year-old man who has lived in Kolkata for ten years. His first language in school was Hindi. He undertook his higher studies in Kolkata and he is pursuing research work in Hindi literature. He is one of the few respondents that reported Bhojpuri as their mother tongue. He spoke enthusiastically about different languages, especially about Bengali. He also demonstrated a lot of emotional attachment to Bhojpuri. bahati madhur bhaashaa hai bahati pyari isko baanglaa me bolte hai “ mishTi bhaashaa” baanglaa khubi misti bhaasaa hai […]baanglaa ke baad jiskaa sthan aataa hai wo bhojpuri kaa aataa hai madhurtaa me hai baanglaa ke baad hindi kaa nahi aataa hai

Language Maintenance and Shift in Migration

151

‘(Bhojpuri is a) very sweet language, extremely lovely, what, in Bengali is called “sweet language”. Bengali is a very sweet language….after Bengali, Bhojpuri comes next in sweetness, not Hindi.’

While speaking about English, he acknowledged the importance of the language, but mentions at the same time that it should be graded lower than a ‘national language’, making clear his dissatisfaction with the importance attached to it. dekhiye english as a language jo hai aacchi bhaashaa hai aur bhaashaa ke taur par sikhnaa caahiye kyunki ye national language ki baat nai hai international language hai English aur ye aapke jibikaa se jur gayi […] iskaa education se hai utnaahi sambandh iska jibkaa se ho gaya hai […] lekin as a national language as a mother language and as a our language iske rup me sabse pehele aagar baangaal me reheti hai to baanglaako aur dusre roop me Hindi ko mahat denaa caahiye[…]English ko as a state language as … mane prasaashan bhaashaa baanaa denaa usko thik hai naa ye thik nai hai. aur english bolnewaale ko ye maanaa jaa raahaa hai ki aap jo hai ye bahat bidwaan ho gaye – nahi…” ‘look, English as a language is a very good language, and as a language one must learn it because it is not a case of national language but international language. English is connected with your job and it is equally connected with education […] but as a national language, as a mother language and as our language we must give importance to Bengali, if you are in Bengal or else to Hindi…English as a state language… I mean to make it language of administration, is not right. And it is assumed the English speaking people are very educated—No…’

It may be noticed here that he mentioned Hindi but not Bhojpuri in the list of mother languages or ‘our languages’. But while talking about Hindi, he started by saying that it is nobody’s mother language, but at the same time he goes on to say that it is “national” language and then argued a case of Hindi as a lingua franca with an integrative function. Hindi as a language raastra bhaashaa hai aur hindi kisiki maatri bhaashaa nahi hai. is desh me hindi kisiki maatribhaasaa nahi hai. ye bhram phelayaa gayaa hai ki Hindi kisiki maatribhaashaa hai. jese meri, main hindi partaa hoon, likhtaa hoon, paraa hoon, lekin meri maatribhaashaa Bhojpuri hai Hindi nahi hai […]ek aisi bhaasaa hai jo puri bhaartiya bhaashaayo ki jurnewaali kadi ho saakti hai […]to Hindi bhaashaako as a national language communication language ke roop me waise langua franca bolte hai, uske roopme jo hai isko aapnaanaa caahiye … ‘Hindi as a language is the national language and Hindi is nobody’s mother language. In this country Hindi is nobody’s mother language. This is a

152

Chapter Eight misconception that has been circulated that Hindi is somebody’s mother tongue. I, for example read in Hindi, write in Hindi, have studied in Hindi but my mother tongue is Bhojpuri, not Hindi […] (Hindi is) one language that has the power to connect the entire country […]. So Hindi, as a national language, a communication language—what you call lingua franca, should be accepted by all…’

Respondent No 4 is a 25 year old guest lecturer, who was born and brought up in Kolkata. He first denied the status of Bhojpuri as a language and called it a boli instead, which can be roughly translated as either speech or a dialect. But having called it a boli, he went on to talk about the vast area in which Bhojpuri is spoken and about the different varieties available there. actually bhojpuri language nahi hai ye …boli hai […] UP Bihaar generally bhojpuri pradesh bahat baraa maanaa jaataa hai aur wahaa pe bahat vary kartaa hai bhojpuri . ‘actually Bhojpuri is not a language…it is a boli (speech variety) […] UP (Uttar Pradesh) and Bihar are (the two states) which are considered to be big Bhojpuri states, Bhojpuri has many varieties there.’

When talking about English, he echoed the view of many other respondents about its foreignness. Its ok but I don’t prefer English kyunki ye humari desh ki bhaashaa nahi hai. ‘It is ok but I don’t prefer English because it is not a language of our nation’

While commenting on Hindi and the issue of national language, he expressed his disappointment with the states of southern India for not accepting Hindi whole-heartedly. haamaare desh ke national language hai to uske baareme kuch kahenaa naa kahene se kuch badalnewaalaa nahi hai […]main south ke baareme thoraasaa..bahat disappoint hoon kyunki south me Hindi ko lekarke dusre tarike ki dhaarnaa hai […]wo log India ka hissa bhi nahi maantaa hai culturally nahi maantaa hai, mentally bhi nahi maantaa. to wahaan pe Hindi ka birodh bahat pahele bhi huaa hai ‘(Hindi is the) national language of our country, whatever you say that is not going to change […] I am a little… a lot disappointed with the south because in south they have a different opinion about Hindi […] they do not

Language Maintenance and Shift in Migration

153

consider (themselves) to be a part of India – culturally or mentally. So opposition to Hindi has also happened there quite sometimes back’

Respondent no 5 is a 37 year old third generation resident of Kolkata. He is an Assistant Professor in Hindi and comes from a highly educated family and his first language in school was English. He was also born and brought up in Kolkata, but is in regular contact with his place of origin, Varanasi (formerly Benaras), where he goes every weekend to visit his parents, who have settled there after their retirement. He spoke fluently in Hindi, Bengali and English and the only respondent in the sample, who is aware that Hindi is not India’s national language. When asked about his mother tongue he replied Hindi, but when asked to talk about his native language, he brought up the issue of Bhojpuri and talked at length about language politics involving the possibility of formation of a new state where Bhojpuri is likely to be the official language expresseing his lack of enthusiasm about the prospect. …that question very much revolves around politics because native language means, if I say Bhojpuri, which is the original language of Benaras now the case history is a little bit different. […] they are demanding a new state—Purvanchal. Aur India kaa politics kaa jo haal hai, khub taaraataari raajyoTaa hoeo jaabe jaar capital hobe Benaras. takhon ki hobe bolaa different, ekhon mushkil. […]ektaa tourist city Benaras nie kikore ekta rajyer economy cholbe bojhaa jaacche naa. kintu etaa hobe. […]. lekin problem ye hai ki, Bhojpuri, for a log time, it has always been the colloquial medium. But for a long time it has been under the shadow of Awadhi which is very neighbouring. […] lekin sheer population er strength jodi dekho, it lies with Bhojpuri. And a political stamp o jokhon peye jaabe ektaa stateer official language jokhon hoe jaabe tokhon poristhiti khothaaye giye daaraabe ekhon bolaa jaacche naa. ‘… that question very much revolves around politics because by native language, if I say Bhojpuri, which is the original language of Benaras, now the case history is a little bit different. […] they are demanding a new state – Purvanchal. And the way Indian politics works, they will get it very soon and the capital of the state will be Benaras. It is difficult to say what will happen then […]. How the economy of the state will run based on one tourist city Benaras, no one knows. But it will happen. […] but the problem is Bhojpuri, for a log time, has been the colloquial medium. But for a long time it has been under the shadow of Awadhi, (another mother tongue under Hindi) which is a neighbouring language.[…] Sheer strength of population lies with Bhojpuri and, when it will get a political stamp, when it will become the official language of a state, then one cannot say what the situation will be like.’

154

Chapter Eight

He expressed strong negative feeling about the importance that English is enjoying in the country and said that he himself is educated in English and he is giving his son an English education, but it is out of compulsion, not by choice. Ideally he would have preferred a school where the medium of instruction is Hindi, but because of ‘ambitious’ politicians, English is a necessity for a good living in India, so he cannot pursue his ideals. Well, it’s a compulsion which we are bearing and we will be bearing. That’s it. It’s a Frankenstein we have given birth and we will have to bear. […] aami Hindi’r lecturer aamaar chele ICSE te pore. aami experiment korte paarbonaa. Benaras e onek bhaalo bhaalo Hindi medium school aache[…]oke antoto shei jaaygaaye aamaake daar korie die jete hobe jaate o kichu kore khete paare. […]When my father didn’t try to experiment with me. Why should I go and experiment with him? Khudiram bhalo kintu paasher bari te hole aaro bhaalo. taai na?, nijer chele na hoaai bhaalo. […] Nobody tries to mess with it nowadays. When India got independent […] humko koi zarurat hi nahi thi English beech me laane ki. aapko official language Hindi baithaa denaa caahiye thaa, national language aapke paas itni saari vernaculars thi. aap har state ke liye compulsory kar do aapnaa mother tongue pado aur Hindi pado. Baat khatam! […]. aapke aapne ambitions itne bade bade the we have made many kind of historical blunders. raashtriya khel hai, raashtriya phul hai, raashtriya gaan hai, raashtriya pashu hai, ek kaayde ki raashtria bhaashaa aapke paas nahi hai. har saal do chaar tho us 8th amendment me baadhaate chale jaate hai, […]. to English we have to bear. iska faydaa US ko ho raaha hai, America ko ho raahaa hai, England ko ho raaha hai. […]It’s very regretting. ‘Well it’s a compulsion, which we are bearing with and we will be bearing. That’s it. It’s a Frankenstein we have given birth to and we will have to bear. […] I am lecturer of Hindi, my son studies in ICSE (English medium) school. I cannot experiment. Benaras has many good Hindi medium schools.[…] but I must put him in a position where he can earn his living in future.[..] When my father didn’t try to experiment with me, why should I go and experiment with him? […] Ksudiram (a well known freedom fighter who died young) is good, but it’s better if he is my neighbour’s son, not mine. […] Nobody tries to mess with it (English) nowadays. When India became independent […] we did not need to bring English in the middle. You needed an official language; take Hindi. For national language you have so many vernaculars. For every state you should have made the mother tongue and Hindi compulsory. End of discussion. […] your own ambition is so big that we have made many historic blunders. There is a national game, a national flower, a national song, a national bird, a national animal but you do not have national language worth saying. Each year two three languages are added to the 8th schedule of the constitution […] English – we have to bear (with it), US, England have benefited from it. It is regrettable.’

Language Maintenance and Shift in Migration

155

5. Conclusion In conclusion, it can be said that among the Bhojpuri speakers studied in this paper, there is a visible inter-generational shift in the use of Bhojpuri in the home domain. Among the different dominant languages that are available in this community, Hindi is the most influential language in the given context. The shift in the home domain indicates a shift from Bhojpuri to Hindi. English, though prestigious, is not influential in this shift as it is considered alien, distant and generally viewed negatively. Bengali, which is numerically a dominant language in the city, also seems to have little influence in the language shift process, though the general attitude towards the language is positive. But Hindi exerts more influence in everyday usage and language preference as it is accepted or assumed to be the national language and enthusiastically adopted as the standard version of the mother tongue. It is the ideological positioning of the community towards Hindi and the other languages that they face that seems to have made this language the most dominant language for the migrated community.

Notes 1

The paper is based on some of the findings of a survey research being carried out at the Linguistics Department, Calcutta University with financial support from University Grants Commission extended under the University with Potential for Excellence scheme. I also acknowledge the help of my colleague Dr Jayati Chatterji, NEHU, for her comments and suggestions on an earlier version of this paper. 2 As per report published in The Statesman 05/03/2005 (page 9) the Bengali speaking population is only 37% of the total population of Kolkata. 3 In this paper, I would use the terms ‘speech’ and ‘speech variety’ as a neutral term as opposed to language or dialect. 4 Ethnologue (2007) reports a total of 38,546,000 speakers with 36,500,000 speakers in India. 5 I have elsewhere explored the possibility of considering Bhojpuri and Standard Hindi (Khariboli) as pluricentric varieties. The paper entitled ‘Bhojpuri as a nondominant variety of Hindi’ is under consideration for publication In: Rudolf Muhr (ed.) (forthcoming): Non-Dominant Varieties of Pluricentric Languages: Getting the Picture. In memory of Michael Clyne. Wien et. al., Peter Lang Verlag. 6 Transcription convention used: Italics: conversation in Hindi Italics and underlined: conversation in Bengali

156

Chapter Eight

References and Further Reading Chand, V. 2011. Elite positionings towards Hindi: Language politics, political stances and language competence in India. Journal of Sociolinguistics. 15/1, pp. 6-35. Chatterji, S. K. 1963. Languages and literatures in modern India. Calcutta: Bengali Publishers. —. 1970. The origin and development of Bengali language, 3 Vols London, George Allen and Unwin (First published in 1926 by Calcutta University Press). Dasgupta, J., and Fishman, J. A. 1971. Inter-state migration and subsidiary-language claming: an analysis of selected Indian census data. International Migration Review, 5/ 2. The Impact of Migration on Language Maintenance and Language Shift. pp. 227-249. Grierson, G. A.1883-7. Seven grammars of the dialects and subdialects of the Bihari language, Parts I-VII. Reprinted 1980, Delhi, Bharatiya Publishing House. Haugen, E. 1966. Dialect, language, nation. American Anthropologist, 68, pp. 922-935. Kerswill, P. 2006. Migration and language. In: Sociolinguistics/Soziolinguistik. An international handbook of the science of language and society. Berlin: De Gruyter, pp. 2271-2285 Khubchandani, L. M. 1978. Toward a selection grammar: Fluidity in modes of address and reference in Hindi-Urdu. Indian Linguistics. 39, pp. 1-24. —.1985. Diaglossia Revisited. Oceanic Linguistics Special Publication. No 20. For Gordon H Fairbanks. pp. 199-211. Kloss, H. 1967. Abstand languages and Ausbau languages. Anthropological Linguistics. 9, pp. 29-41. Lippi-Green, R. 1994. Accent, Standard Language Ideology, and Discriminatory Pretext in the courts. Language in Society. 23/2. pp. 163-198. Milroy, J., and Milroy, L. 1985. Authority in Language: Investigating language prescription and standardisation. London: Boston and Henley, Routledge and Kegan Paul. Milroy, J. 2001. Language ideologies and the consequences of standardization. Journal of Sociolinguistics. 5/4. pp. 530-555. —. 2007. The ideology of the standard language. In: C. Llamas, L. Mullany and P. Stockwell (eds.), The Routledge Companion to Sociolinguistics. New York: Routledge, pp. 133–139.

Language Maintenance and Shift in Migration

157

Parasher, S. V. 1980. Mother tongue-English diglossia: A case study of educated Indian bilinguals’ language use. Anthropological Linguistics 22/4, pp. 151-162. Tiwari, U. N. 2001. The origin and development of Bhojpuri. Kolkata: The Asiatic Society. (First published in 1960) Trudgill, P. 2004. Glocalisation and the Ausbau sociolinguistics of modern Europe. In: A. Duszak and U. Okulska (eds.), Speaking from the Margin: Global English from a European Perspective. Frankfurt am Main: Peter Lang:, pp. 35-49. Tsitsipis, L. D. 1995. The coding of linguistic ideology in Arvanitika (Albanian) language shift: congruent and contradictory discourse. Anthropological Linguistics 37/4, pp. 541-577. Wiley, T. G., and Lukes, M. 1996. English only and Standard English ideologies in the U.S. TESOL Quarterly 30/3, Language Planning and Policy. pp. 511-535. Woolard, K. A., and Schieffelin, B. B. 1994. Language Ideology. Annual Reviews in Anthropology, 23: pp. 55-82.

Exercises 1. What role does standard language ideology play in forming language attitude? Do you see a similar pattern (as discussed in this paper) in other languages familiar to you? 2. What factors may affect language maintenance or shift in instances of migration? 3. What are the most pertinent problems for a linguistically diverse nation like India? Are there any advantages of this diversity? What should be the ideal attitude for policy makers to handle linguistic diversity?

CHAPTER NINE LANGUAGES IN CONTACT IN MINORITY COMMUNITIES: THE CASE OF FRANCOPROVENÇAL IN ITALY CARMELA PERTA

1. Introduction Italian repertoire is formed by the official language, Italo-Romance dialects and in some cases it is constituted by the coexistence of them with a minority language.1 Starting from the latter reality, this paper has a twofold aim: firstly, it will analyse the vitality of a minority language Francoprovençal - through describing the speakers’ competence and use of the language; secondly, it will examine the distribution of the languages in the social repertoire of two communities where Francoprovençal is spoken. As is known, with Francoprovençal is meant a set of Gallic-Romance varieties spoken in France, Switzerland and Italy, differing on a phonetic basis from French on the one hand and from Provençal on the other (Ascoli 1878). Though such a dialect group is effectively Abstandsprache (in Kloss’s terms ‘languages of distance’, that is languages which are structurally distant enough from Italian to be considered different languages) within the larger group of Romance languages, it does not have a reference standard which covers and unifies the whole range of related dialects, and which functions as an identification symbol. The fact that the identity of Francoprovençal as a language is mainly scholarly has important consequences for its official status and for its maintenance today. In Italy, Francoprovençal varieties are spoken in Valle d’Aosta, in six valleys in Piemonte and in two small villages in the Apulia region. Even though they present common traits, each variety differs from the other, to the extent that a common ethnic-linguistic identity among the communities does not exist, even throughout Italy.

Languages in Contact in Minority Communities

159

Figure 9-1: Faeto and Celle San Vito

2. Historical Sketch The two villages in Apulia in which I conducted fieldwork - Faeto and Celle San Vito - are particularly interesting in that their sociolinguistic situation is very different, notwithstanding their closeness (they are 3 kilometres from each other) and their isolation from other Francoprovençal communities in Italy. There are conflicting accounts as concerns the origin of the Francoprovençal communities in Apulia: it is not certain whether these

Chapter Nine

160

groups were Angevins or Waldenses. Instead, the time period of their settlement has been established to be between the late of 1200s and 1500s (Sobrero 1974, pp. 41-42; Telmon 1994, p. x). The lack of consensus on their origin has created confusion of identity, since most of the community members declare they speak Provençal rather than Francoprovençal (Perta 2008; 2010). In fact, it was only in 1888 that Suchier, on the basis of the language used in a brief novel2, established that Faeto and Celle varieties belong to the Francoprovençal group, rather than to the Provençal one.

3. The Fieldwork As stated in the introduction, the aim of the study is to analyze Francoprovençal vitality; it was ‘measured’ on the basis of the speakers’ competence and use of the minority language. Methodologically, a representative stratified sample formed by 52 informants in Faeto and 63 for Celle San Vito3 was used; subjects, who were interviewed on the basis of a semi-structural questionnaire4, were segmented according to the following variables: x demographic variables (‘age’ and ‘gender’); x social variables (‘education’ and ‘job’). On the basis of speakers’ age, the samples were stratified as follows: 1. 2. 3. 4.

layer 1: layer 2: layer 3: layer 4:

informants from 3 to 19 years old; informants from 20 to 39 years old; informants from 40 to 69 years old; informants from 70 or more.

As concerns the variable ‘education’, these criteria were observed: 1. 2. 3. 4. 5.

subjects without any education; subjects with primary education; subjects with secondary education; subjects with high education; subjects with higher education.

As concerns the variable ‘job’, informants were segmented on the basis of: 1. 2.

agriculture related jobs; industry related jobs;

Languages in Contact in Minority Communities

3. 4.

161

services related jobs; no job (unemployed, housewife, student)5.

Quantitative methods were used to analyse the data, particularly the package SPSS was used to estimate the statistical association between the variables through a linear equation6. The social repertoire of the two communities comprises Italian, local Italo-Romance dialect, and Francoprovençal, all of them distributed differently in the individual repertoire. On the basis of a scale from 1 to 10, subjects were asked to say how much they used each language7 in conversations with different interlocutors, particularly in family (with parents, grandparents, siblings, partner, children, grandchildren), with agegrading friends, with colleagues, in school, in church, in topic-bound conversations, when dreaming and mentally counting8. Through a factorial technique the whole set of independent variables was reduced to two dimensions; on the basis of the dimensions it was possible to interpret speakers’ language behaviour, considering them as the ratio of the their language choices.

4. Faeto On the basis of the results, in Faeto the minority language is largely used by the population (92% of the informants declared to be active speakers of Francoprovençal). Only a low rate of subjects (8%) belonging to the age layers 1 and 2 state they know Francoprovençal passively9. Moreover, none of the social variables – ‘age’, ‘gender’, ‘education’ and ‘job’ - make a statistically significant contribution to the variation in the speakers’ competence; in other words, such variables do not appear to be correlated to the speakers’ degree of competence in Francoprovençal. Excluding any Italo-Romance dialect, two languages are used by the community: Italian and Francoprovençal; their functional distribution is balanced, since Francoprovençal is used more than Italian in all informal settings; while in formal settings, Italian is commonly used more than Francoprovençal, above all in schools and in church10.

4. 1 Language Choices in Faeto: A Factorial Interpretation The two dimensions which represent the whole set of independent variables11 have positive and negative values. As concerns dimension 1, it is positively correlated to the use of Italian (with ‘mother’, ‘father’, ‘teacher’, ‘grandparents’, ‘schoolmates’, ‘older siblings’, for ‘mentally

162

Chapter Nine

counting’, with ‘younger siblings’, and for conversations about ‘everyday life’ and with ‘partner’).

Figure 9-2: Dimension 1, positive variables

Figure 9-3: Dimension 1, negative variables

The same variables are negatively correlated to the use of Francoprovençal, as is clear form Figure 9-3. Variables which are represented by dimension 2 are positevely correlated to the use of Italian and negatively correlated to the use of Francoprovençal; they are conversations with ‘colleagues’ and with ‘one’s own children’ as is represented in Figure 9-4.

Languages in Contact in Minority Communities

163

Figure 9-4: Dimension 2, positive and negative variables

In other words, we move from a core of interlocutors with whom the speakers’ language choices are stable (dimension 1) to an area where there is more variability (an area represented by both dimensions), and finally we reach a border (represented by variables of dimension 2) where variability related to the language to be used rapidly increases.

Figure 9-5: Dimensions and ‘age’

Thus, dimension 1 could be labelled ‘inner’, since it is represented by interlocutors with whom speakers would have spontaneous/unconscious language interaction; the result is stable uses of language. While, dimension 2 could be labelled ‘outside’ since it is represented by interlocutors

164

Chapter Nine

with whom speakers would have a conscious language behaviour based on a precise decision of the language to be used. Considering the correlation of the two dimensions with speakers’ age, it is clear that most of the speakers belonging to layer 4 and some to layer 3 are negatively correlated with both dimensions, in other words there is not any variability in their language uses, since Francoprovençal is the preferred language in all settings. Moreover, for some speakers of layer 4, 3 and 2 there is a higher variability in their language choices in ‘inner’ situations, this means that they use Italian more than the previous group. Instead, most of the speakers of layer 2 and all the speakers of layer 1, for whom there is also a negative correlation with dimension 1, are located in the positive sector of dimension 2; this means that they use Francoprovençal only in ‘inner’ situations, and more Italian than Francoprovençal in ‘outside’ situations.

Figure 9-6: Dimensions and subjects

This is clear from Figure 9-6, which shows speakers’ language behaviour divided on the basis of the two dimensions. In dimension 1, starting from positive values related to speakers of layer 1 and 2 (which means variability in their language uses) more and more negative values are recorded as the speakers’ age increases. This means that speakers use more Francoprovençal than Italian with interlocutors with whom they have spontaneous language behaviour. In dimension 2, the curve is less shaped since the values are negative in almost all the cases; this means that Francoprovençal is used more frequently than Italian. However, there are high values towards the positive sector in correspondence of speakers of layer 2, 3 and 4; this means that Italian is used more with interlocutors belonging to the ‘outside’ world rather than those belonging to the ‘inner’ world. It is worth

Languages in Contact in Minority Communities

165

mentioning that some speakers belonging to layer 2 display a deep sense of loyalty towards Francoprovençal rather than some speakers of layer 3 do.

5. Celle The results show that 56% of the informants are actively competent in Francoprovençal, 22% state that they know Francoprovençal passively and 22% do not know the minority language at all. ‘Age’, ‘education’ and ‘job’ are statistically significant variables, since they are the factors which determine the degree of speakers’ competence in Francoprovençal; this means that loyalty towards the minority language is associated with elderly, rural areas, and primary sector employment12. Francoprovençal is rarely used in the considered settings, also in the family, a fundamental domain for the maintenance of a minority language. Moreover, the repertoire of speakers which are 11 to 39 years old is based only on Italian and its diatopic variety, both used according to the formality of the situation, hence excluding any use of Francoprovençal. Instead, for speakers who are 39 and over, the use of Francoprovençal increases and the use of the Italo-Romance dialect decreases as the speakers’ age.

5. 1 Language Choices in Celle: A Factorial Interpretation Through factorial analysis two dimensions were obtained13; dimension 1 is positevely correlated with the variables related to the use of Italian (with ‘teacher’, ‘schoolmates’, ‘friends of layer 1’, ‘priest’, conversations about ‘economics/politics’, with ‘friends of layer 3’) and of the local Romance dialect (with ‘partner’, ‘older siblings’, ‘father’, ‘friends of layer 4’), whereas it is inversely correlated with variables which represent the use of Francoprovençal (with ‘teacher’, ‘friends of layer 4’, ‘friends of layer 3’ ‘friends of layer 2’, ‘schoolmates’). On the other hand, dimension 2 is represented by variables related to the use of the Romance dialect (in conversations about ‘everyday life’ and with ‘mother’) and of Italian (with ‘older siblings’, ‘one’s own children’, ‘father’), while it is inversely correlated to the variables related to the use of Francoprovençal (with ‘mother’, in conversations about ‘everyday life’ and with ‘younger siblings’). The following figure shows a comparison among the variables related to the use of Italian and Francoprovençal according to both dimensions.

166

Chapter Nine

Figure 9-7: Francoprovençal vs. Italian

Figure 9-8 shows a comparison between the use of Francoprovençal and the use of the local Romance dialect in both dimensions.

Figure 9-8: Francoprovençal vs. dialect

Languages in Contact in Minority Communities

167

Figure 9-9: Dimensions and ‘age’

From Figure 9-9 which pictures the correlation between the dimensions and the speakers’ age, it is clear that only some old speakers are negatively correlated to both dimensions; this means that they use Francoprovençal more than Italian and the local Romance dialect. Moreover, some speakers of layer 4 and 3 are negatively correlated with dimension 1 and positevily correlated with dimension 2, whereas for other speakers belonging to these age layers the opposite correlations are recorded; finally, for the rest of layer 4 and 3, and all layer 2 and 1 there is a positive correlation with both dimensions, this means that these speakers prefer Italian and the local Italo-Romane dialect to Francoprovençal. Therefore, dimension 1 could be labelled ‘past’, while dimension 2 ‘present’: in the ‘past’ Francoprovençal was preferred to Italian and above all to the local dialect, while in the ‘present’ the community use both Italian and the local Romance dialect, with the latter replacing Francoprovençal. This is clear by analysing the language choices of the speakers, divided according to their age, together with the two dimensions.

Chapter Nine

168

Figure 9-10: Dimensions and speakers

Considering both dimensions, the positive values recorded for speakers of layer 1 and 2 show that their language space is divided between Italian and the local Romance dialect, therefore Francoprovençal is completely excluded from their choices. This panorama changes as the speakers’ age increases, since negative values are recorded for the two dimensions, this means that Francoprovençal was and is preferred to Italian and to local dialect in the past and in the present. It is worth noting that for some speakers of layer 3 there are negative values for dimension 1 that is for the ‘past’, this means that they use more Francoprovençal than some speakers of layer 4 do in both the dimensions of time.

6. Discussion Speakers’ language use and the attitude they have towards Francoprovençal are the key factors for sketching the social repertoire14. H LANGUAGE

Italian

L LANGUAGE

Italo-Romance dialect

Francoprovençal

Table 9-1: Repertoire in Celle In Celle the languages used by the community are Italian (used as the high language (HL) also in formal settings as to determine a situation of

Languages in Contact in Minority Communities

169

dilalia (Berruto 1995)), Italo-Romance dialect and Francoprovençal, both in competition in the functional area of the law language (LL). The local variety of Francoprovençal is considered as ‘the old dialect’, as is clear from the words of this speaker belonging to layer 415. \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\

1. ?quando dici di parlare il dialetto a cosa ti riferisci? al cellese ah: ?ma è come quello che si parla a Faeto? no, quello è il dialetto vecchio ?come vecchio? prima qua parlavamo lu cigliàje + uguale + un poco a Faeto, ma prima

[Translation] \Res\ ?when you say that you speak dialect what do you refer to? \Sub\ to cellese \Res\ ah: ?is it the same as the one spoken in Faeto? \Sub\ no, that one is the old dialect \Res\ ?in which sense is it old? \Sub\ once we spoke lu cigliàje + the same + a bit to Faeto dialect, but once

According to the speakers’ opinion, this ‘old’ dialect is different from the new one which is similar to the one spoken in nearby towns and is used by the whole community, as is clear from this part of interview with a speaker of layer 316. \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\

2. ?ma il dialetto lo conosci? quello vecchio un poco + abbastanza, ma qua parliamo il dialetto di qui scusa non ho capito + ?cosa è quello vecchio? lu cigliàje ?ma è quello vecchio o il dialetto di qui? mò c’è un altro ah: uno vecchio e uno nuovo ?e quello vecchio come hai detto che si chiama? cellese, mò si dice provenzale + ah: no, francoprovenzale ?è non è pure di qui? sì, ma non lo parliamo più molto, parliamo cellese ?cellese? ?ma non è il francoprovenzale? no, il cellese di ora, come quello a Castelluccio, un poco come Lucera Pure

[Translation] \Res\ ?do you know dialect? \Sub\ the old one a little bit + enough, but we speak the local dialect now

Chapter Nine

170 \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\

sorry I don’t understand + ?what is the old one? lu cigliàje ?but is this the old one or the local dialect? now there is another one ah: an old one and a new one ?and how is the old one called? cellese, now they say Provençal + ah: no, Francoprovençal ?is it a local one as well? yes, but we don’t speak it a lot, now we speak cellese ?cellese? ?it is Francoprovençal, isn’t it? no, new cellese, as the dialect spoken in Castelluccio, similar to the Lucera one as well

In other words, the minority language is less used than the ItaloRomance dialect, which is named cellese the label once used by the community to refer to Francoprovençal; in this type of repertoire a dangerous situation for the vitality of the minority language could be seen, because the Italo-Romance dialect competes with the minority language in informal settings (Berruto 2009: 181). This is what I found in Celle where the local dialect gains space in informal settings, to the point that the minority language is completely cancelled from the repertoire of the young speakers while keeps leaving in that of the old speakers.

- functions +

ĸ Italian ĸ Local Romance dialect ĸ Francoprovençal - society +

Figure 9-11: Representation of the repertoire in Celle

Instead, the threat that comes from the high language (HL) towards the minority language is obvious, and a type of repertoire with the standard language (or better its regional variety) used as the HL and the minority language used as the LL is balanced because of their good functional distribution. This is the situation found in Faeto.

Languages in Contact in Minority Communities

Italian ĺ

171

ĸ Francoprovençal - functions +

- society + Figure 9-12: Representation of the repertoire in Faeto

This is clear from this part of the interview with a speaker of layer 3. \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\

3. ?parli altre lingue o dialetti ùtre a lu taliàn? ah: ?ti interessa lu fajetàre? ?il faetano? ?e cos’è? ?lu dialètte de Faìte? come il foggiano, il napoletano !no! ma: ?e lu dialètte? ?stì gli étte lu vóte dialètte + de Faìte? è lu dialètte de Faìte + ma non è un poco italiano + si chiama provenzale

[Translation] \Res\ ?do you know other languages apart from Italian? \Sub\ ah: ?are you interested in faetano? \Res\ ?faetano? ?and what is this? ?Faeto dialect? As Foggia dialect, Naples dialect \Sub\ !no! \Res\ but: ?is it a dialect? ?is it the local dialect + Faeto dialect? \Sub\ it is Faeto dialect + but it’s not like Italian + it’s called Provençal

In other words, it seems that in the community the language market includes Italian and Francoprovençal excluding the presence of any ItaloRomance dialect. H LANGUAGE

Italian

L LANGUAGE

Francoprovençal

Table 9-1: Repertoire in Faeto

172

Chapter Nine

This is is proved by the following interview with a speaker belonging to layer 3. \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\ \Res\ \Sub\

4. ?acchì vù ve parlà pa ‘ndùtte lu taliàn, vero? ++ ge parle + nùsse ne parlùn lu provenzale ah: ?lu provenzale? !eh il provenzale! mò lu chiammanùnt francoprovenzale ?é lu ++ il vostro dialetto? gli i étte lu dialètte de lu nóte paìje + de Faìte ah: ?quindi è come il lucerese, il foggiano? nu: ?e che cos’è? è un dialetto diverso…

[Translation] \Res\ ?you speak Italian only here, don’t you? \Sub\ ++ I speak, we speak Provençal \Res\ ah: ? Provençal? \Sub\ !eh Provençal! Now they call it Francoprovençal \Res\ ?is it ++ the local dialect? \Sub\ it’s the local dialect, Faeto dialect \Res \ ah: ?so is it like Lucera, the Foggia dialect? \Sub\ no: \Res\ ?and what is this? Sub\ it’s a different dialect…

In conclusion, as was demontrated by analysing the relationship between the minority language and the other languages used by the community, it was possible to delineate the social repertoire and the functional distribution of the languages which seem to be key factors in the maintenance of the minority language.

Notes 1

Italian law 482/1999 Regulations regarding the protection of historic language minorities was approved on the 15th December 1999, thus giving a national framework to minority language preservation and renaissance. The legislation made it possible to use the historic minority languages in education, in public offices, in local government, in the judicial system, in the mass media, and allowed for the reinstatement of place and personal names. The status and prestige of minority languages thus appeared to be strengthened, even though only given minority languages are protected by the national law (Albanian, Catalan, German, Greek, Slovene, Croatian, French, Franco-Provençal, Friulan, Ladin, Occitan and

Languages in Contact in Minority Communities

173

Sard) notwithstanding more than twelve minority languages are spread throughout Italy. 2 The novel appeared in Papanti 1875. 3 The main study was preceded by a pilot study whose results were used as variables to calculate the sample (see Perta 2008). 4 In most of the cases the language used in the interviews was Italian but with switchings either in Francoprovençal or in the local Romance dialect. 5 These factors are considered to be independent variables used to analyse their statistical contribution to the variation in speakers’competence. 6 This equation predicted the values of a dependent variable from knowledge of specified values of same independent variables – i.e. predictors. The type of the regression used – categorical regression – was chosen since it allows the analysis of both quantitative and qualitative data. 7 Speakers were explained that they should also consider the situations in which more than a language was used in a single speech act. 8 These factors are considered to be independent variables to explain the variation in the speakers’ language choices. 9 For the complete results see Perta 2008. 10 It is worth mentioning that in Faeto local authorities and planners, in compliance with national regulations, are implementing various measures to strengthen their language position and to promote it in new settings, particularly the formal ones. 11 Dimension 1 could explain 59.329% of the cumulative variation, dimension 2 explains 7.785%. 12 For the complete results see Perta ivi. 13 Dimension 1 represent 27.873% of the cumulative variation, while dimension 2 14.544%. 14 See Dal Negro and Iannàccaro 2003. For the different types of repertoire in Italy see Berruto 1995 and Mioni 1988. 15 In the parts of interviews \Res\ stands for researcher, \Sub\ stands for subject; the norms used for the transcription of Francoprovençal are taken from Minichelli (1994) and Nagy (2000). Other criteria are: 1. 2. 3. 4. 5. 6. 16

word word + word: ?word? !word!

Francoprovençal; Italian; stop; emphasis; question; esclamation.

It is worth noting that the informants in (1) and (2), who declared they know Francoprovençal actively, did use few Francoprovençal, notwithstanding my efforts in trying to insert it in the conversation.

174

Chapter Nine

References and Further Reading Ascoli, G. 1878. “Schizzi Francoprovenzali”. Archivio Glottologico Italiano: pp. 61-120. Berruto, G. 1995. Fondamenti di Sociolinguistica, Bari: Laterza. Berruto, G. 2009. “Repertori delle Comunità Alloglotte e ‘Vitalità delle Varietà Minoritarie”. In: C. Consani, P. Desideri, F. Guazzelli and C. Perta (eds.), Alloglossie e Comunità Alloglotte nell'Italia Contemporanea. Teorie, Applicazioni e Descrizioni, Prospettive, Roma: Bulzoni: pp. 173-198. Dal, N. S., and Iannàccaro, G. 2003. “«Qui parliamo tutti uguale, ma diverso». Repertori Complessi e Interventi sulle Lingue”. In: A. Valentini, P. Molinelli, P. Cuzzolin. and G. Bernini (eds.), Ecologia Linguistica. Roma: Bulzoni, pp. 431-450. Minichelli, V. 1994. Dizionario Francoprovenzale di Celle San Vito e Faeto, Alessandria: Edizioni dell’Orso. Mioni, A. 1998. “Standardization Processes and Linguistic Repertoires in Africa and Europe. In: P. Auer and A. Di Luzio (eds.), Variation andConvergence. Studies in Social Dialectology. Berlin/New York: de Gruyter, pp. 294-320. Nagy, N. 2000. Faetar. München: Lincom Europa. Papanti, G. 1875. I Parlari Italiani in Certaldo alla Festa del V Centenario di Messer Giovanni Boccacci. Livorno: Forni, pp. 62-65. Perta, C. 2010. “La Varietà Francoprovenzale della Puglia fra Mantenimento e Perdita linguistica”. In: M. Iliescu, H.M. SillerRungaldier and P. Danler (eds.), Actes du XXV Congrès International de Linguisique et de Philologie Romanes. Tome I. De Gruyter: pp. 215222. Perta, C. 2008. Repertori e Scelte Linguistiche nelle Comunità Francoprovenzali della Puglia. Roma: Aracne. Sobrero, A. A. 1974. “Il Franco-provenzale in Capitanata: Storia Interna e Storia Esterna di una Parlata Alloglotta”. In: A. A. Sobrero (ed.), Dialetti Diversi. Proposte per lo Studio delle Parlate Alloglotte in Italia. Lecce: Milella, pp. 33-64. Telmon, T. 1994. Presentazione in V. Minichelli 1994: I-XIII.

Languages in Contact in Minority Communities

175

Exercises 1. How could Francoprovençal be considered? Where is it spoken? 2. Describe the different vitality of the minorità language in Faeto and Celle San Vito. 3. Outline the social repertoire in Faeto. 4. Outline the social repertorie in Celle San Vito.

CHAPTER TEN LANGUAGE MAINTENANCE AND LANGUAGE SHIFT AMONG AFRO-COSTA RICAN WOMEN LISBETH PHILIP

1. Introduction Immigrant contact settings where immigrants speak a language different from the host country create interesting case studies; especially in the areas of language maintenance and language shift (see Fishman 1972 and 1978; Clyne 2005, and others). With the group this study focuses on, the Afro-Costa Rican women in the Province of Limón, Costa Rica, the bilingual situation came about when their ascendants - West Indian immigrants, the great majority from Jamaica- settled in the province of Limón beginning in the 1870s. The West Indians lived in enclaves where English was the language used in the home as well as in the workplace. After a period of initial resistance, a gradual shift from the exclusive use of English to the incorporation of Spanish took place when social and educational reforms were implemented to integrate the West Indian Blacks to the national Costa Rican identity. This marked the passing from monolingualism to bilingualism and the beginning of the transformation from West Indian Black to Afro-Costa Rican. The Republic of Costa Rica is located in the Central American isthmus, bordered by Nicaragua on the north and Panama on the south. The country is divided into seven provinces, which are further subdivided into cantons. One of the provinces is Limón, the geographical setting of this study. The province of Limón is often referred to by Costa Ricans as a ‘mosaic of cultures’, since it is the only province of Costa Rica that is populated by a variety of ethnic groups that constitute minority groups in the country, including Blacks, Chinese, and a number of indigenous Amerindians. The population to the north of the old railroad line (Línea Vieja) is mainly composed of white or mestizo Hispanics. In the area of Colorado, close to the Nicaraguan border, the largest communities are

Language Maintenance and Shift among Afro-Costa Rican Women

177

populated by migrants from Nicaragua who are of different ethnic backgrounds (White, Mesquito Indian and Black). The Afro-Costa Rican population in the province is concentrated in the communities along the southern part of the railroad line up to the Panama border (this includes Siquirres, Matina, Limón and Talamanca). The Chinese, who are numerically small but play an important role in the economic life of the region, are largely concentrated in the city of Puerto Limón. It is also important to note that the indigenous population – the Bribris, Cabécares and others - is concentrated in the canton of Talamanca, south of Limón. However, of the three groups mentioned, the Blacks constitute the dominant minority group. In the census data of the year 2000 Afro-Costa Ricans constituted 1.9%, that is 72,784 of the country’s population. This information was acquired through the incorporation of a racial variable in the national census (Instituto Nacional de Estadística y Censos 2001). The census demonstrated that the AfroCosta Rican group is still overwhelmingly concentrated in the province of Limón, which had a total of 74.4% Afro-Costa Ricans (see Putnam 2002b), concentrated in the cantons of Siquirres, Matina, Talamanca and Limón. This last canton contains Puerto Limón, the province’s largest city and the major Atlantic port in the country.

2. West Indian Black Migration to the Province of Limón The mid-nineteenth century is marked by the presence of North American companies that had vested economic interests in the development of the infrastructure of the Central American Caribbean coast. As Holm (1983) points out, North American companies maintained a prominent position first in the construction of railroad systems and later, in the establishment of banana plantations in the Atlantic coast in Central America. These companies recruited labourers from Jamaica and other West Indian islands for both enterprises and created enclaves for their operations in which English was working language (Holm 1983). The Black West Indian labour force was turned to after many failed attempts in the search for labourers. This occurred not only in Costa Rica, but also in the construction of the Panama railroad, which employed West Indian workers on a large scale (Bourgois 1989; Putnam 2002a). The first West Indian immigrants to Costa Rica came from various parts of the Caribbean, arriving in the decades after the emancipation of slaves in the various islands. Jamaicans at first constituted a small number of this group (Meléndez 2005).

178

Chapter Ten

In the years following the first migration of Jamaican labourers, immigration of Black West Indian workers to Costa Rica for the construction of a railroad system was constant, and Jamaicans soon began to greatly outnumber the other Black West Indian groups. In 1883, the census data indicated that approximately 47% of the population in Limón was of Jamaican origin (Casey Gaspar 1979). In the meantime, banana production had been spreading along the railroad line. By the 1890s, large successful banana plantations, owned in whole or in part by Minor Keith, had been formed, resulting in the displacement of small landowners (Putnam 2002a), and the formation by Keith of the United Fruit Company (UFCO) in 1899. Jamaicans became the preferred workers on the plantations due to their physical resistance and knowledge of banana production, which had been initiated in Jamaica before it became an industry in Costa Rica (EcheverriGent 1992; Meléndez 2005). The number of Black West Indian workers grew because employment opportunities in the West Indies did not provide the same opportunities and treatment that the UFCO had to offer (see Chomsky 1996). Railroad workers were contracted temporarily to work in Costa Rica until the duration of the assignment for which they were hired. From the perspective of the Black immigrant, who, according to Duncan (2005), had developed a sense of loyalty to the British Crown, Costa Rica was, from the onset of their arrival, a ‘provisional’ or a ‘temporary’ environment. In 1891, the completion of railroad construction gave rise to the second phase of the financial venture of the United Fruit Company in Costa Rica: The formation of banana enclaves in the region (Municipalidad de Limón 1992; Herzfeld 2002). Having spent twenty years working on the railroad, many workers extended their stay to work in the banana plantations. In the 1927 census, under the classification of ‘foreigner’ (Stewart 1999), the total number of Jamaican Blacks in the province of Limón had reached 18,000. By 1930, however, the banana economy in the Caribbean area experienced an economic crisis, as the disease ‘sigatoka negra’ struck (Municipalidad de Limón 1992). The United Fruit Company, which had exercised great control in the region, departed, and the Costa Rican government, gained economic and social control of the region. By 1939 all ties with Jamaica were broken; Costa Rica was no longer thought of as a ‘provisional’ or ‘temporary’ environment by the immigrants. Yet, these immigrant workers were confined to the Caribbean region due to an immigration law that prohibited them from travelling outside the

Language Maintenance and Shift among Afro-Costa Rican Women

179

region. During this period, Blacks were neither Costa Rican nor Jamaicans (Municipalidad de Limón 1992). After the Revolution of 1948 in Costa Rica, by decree of law N.836, on November 4, 1949, the descendants of the Jamaican immigrants were granted the right to Costa Rican citizenship1. The province of Limón began to experience a major demographic change, as many more Hispanics migrated into the region. Communities that were inhabited exclusively by the Afro-Caribbean worker became more ethnically diverse neighbourhoods where Spanish and Limón English were both spoken. This began the transformation of the West Indians into an Afro-Costa Rican society. Gradual shifts from English to Spanish took place with education reforms, and the children of the West Indian immigrant were slowly integrated into the national identity.

3. The English Spoken in Limón The linguistic scenario of the Afro-Costa Rican population is a rather complex situation. The ambiguous status of the English spoken in Limón was noted in the opening lines of a Costa Rican newspaper article entitled “Limón: Lengua, legado y pérdida” which stated: ‘Por muchos años las variedades lingüísticas de Limón han sido un tema de debate sobre el que no ha habido acuerdo’ [For many years the linguistic varieties of Limón have been a subject of debate from which no agreement has emerged.]2 In the article, the author examined the linguistic peculiarities of Standard Limón English (a prestigious form used in formal situations) and Limón English Creole spoken in everyday situations by a significant number of the Black population in the Province of Limón in Costa Rica. Researchers (Herzfeld 1978, 2002a; Wright 1982; Spence 1997) have classified the languages spoken by members of the Black community into Spanish, the official language of Costa Rica; Standard Limón English, which is used in formal settings, and the Limón English Creole, also labelled as Mekatelyu or Patois, which is used in informal conversation among its speakers. Winkler (2000), on the other hand, although recognizing that there are varieties of English spoken in the area, concluded that ‘... en la ciudad de Limón casi no quedan dominios sociales del habla que pertenezcan exclusivamente al inglés o al inglés criollo’ (190) […in the city of Limón, there are hardly any social domains left in which the language spoken belongs exclusively to English or to Creole English.] Based on this, for the purposes of her study, she chose to label the varieties of English spoken in the area simply as English ‘(Para simplificar, voy a usar la

180

Chapter Ten

palabra inglés para referirme tanto al inglés estándar como al criollo limonense)’ [To simplify, I am going to use the word English to refer to both standard English and Limonense Creole.] Given the many linguistic characteristics that both varieties share, it is generally agreed that the English spoken in the area is a descendent of Jamaican English/Creole (Herzfeld 1978, 2002a, 2002b; Wright 1982; Holm 1983; Masís Morales and Mora Lobo 1985; Winkler 2000). However, like much of the English spoken on the eastern coast of Central America, the English spoken in Limón can also be characterized as part of a continuum in which many varieties overlap, providing features that oscillate from the most basilectal variety to a more standardized West Indian English variety (see Holm 1983 and Herzfeld 2002b for an analysis of Central American English/Creole). Notwithstanding the different views concerning the proper classification of the varieties of English spoken in the region, Spence (1998) points out that for many speakers Limón English Creole represents a handicap rather than an advantage for social-economic climbing. Hill Hudson (2001:199) cites John R. Edwards, as follows: ‘… language varieties which diverge from standard English are likely to be viewed, even by speakers of those varieties, less favorably than Standard.’ The holding of such negative views is illustrated by the observation of writers such as the one whose quote is cited at the beginning of this section concerning the reluctance of people to admit that they speak ‘Mek-a-tel-yu’: Lingüísticamente ninguna de las dos es mejor, pues ambas cumplen con un solo propósito: establecer y mantener los canales de comunicación abiertos entre los miembros de una comunidad. Sin embargo, y esta es otra de sus características, pocos admiten que hablan mek-a-tel-yu (October 1997). [Linguistically neither of the two is better, since both serve one purpose: to establish and maintain open communication channels among members of the community. However, and this is another of its characteristics, very few admit that they speak Mek-a-tel-yu.]

Today, the urban life of the Jamaican descendents is concentrated mainly in the cities of Puerto Limón and of Siquirres (Bozzoli de Wille 2005), and although according to Taylor (1968) size of a speech community does not guarantee linguistic longevity, it is estimated that the number of speakers of Limón English is significant: around 55,000 (Herzfeld 2002a:10)3. For purpose of clarity in light of this continuum, I will refer to the English spoken in Limón, including standard Limón English, Limón English Creole and its varieties as Limón English, following a variation of Winkler’s (2000) approach, as noted previously.

Language Maintenance and Shift among Afro-Costa Rican Women

181

4. Significance of the Study In the analysis of language maintenance and language shift, one of the most common traits affecting language behaviour involves attitudes. Attitudes can be defined as the amount of positive or negative feeling one holds toward a specific social object or class of social objects. Affective variables have traditionally been studied in second language learning, following the work of Gardner and Lambert (Gearoid 1984; Genesse, Rogers and Holobow 1983). The effects of affect in second language learning can be attributed to attitudes and motivation (Gardner 1988; Gearoid 1984; Massey 1986). According to Stone and Rubenfeld (1989, p. 230) attitudes represent a response by the individual to a social object, and for Gearoid (1984) they are characterized by three components: a) the affective component; b) the cognitive component; and c) the behaviour component. Gardner and Lambert’s social approach to attitudes (cited in Gearoid 1984) is influenced by people’s perception and beliefs about the ethno-linguistic group who speak the language in question. Some researchers take the position that ‘Identification with a community language is probably the key factor in group strengthening and stabilization’ (Dabène and Moore 1995, p. 23), whereas others maintain that ‘Identification with a language and positive attitudes towards it do not guarantee its maintenance’ (Romaine 1995, p. 43) and that ‘degree of proficiency has little to do with language attitudes’ (Paulston 1994, p. 21). For some cultures, the value attached to language tends to be more important than social, historical and cultural factors (Clyne 2005). According to Søndergaard and Norrby (2006, p. 106) when language is a core value, language maintenance can be expected to be high, and it can be expected to be low if the group in question does not include language among its core values. However, social and historical pressures to integrate members of a linguistic minority, as well as stigmatized perceptions from the outside, tends to reinforce negative attitudes and encourage shift among speakers of less prestigious languages or language varieties. In the case of Limón English, the factors mentioned above and the increased prestige of Spanish since the adoption of the Costa Rican nationality by many Limón English speaking Blacks, together with their gradual assimilation, especially within the third and fourth generations, to the national Costa Rican identity, make it appropriate to ask what the future holds for Limón English. It is particularly important to examine this question from the Afro-Costa Rican women’s perspective regarding the role that attitudes and identity play in

182

Chapter Ten

language maintenance and language shift (hereon LMLS), since women are traditionally viewed as key players in the transmission of language.

5. The Sites of the Study The sample studied consisted of 127 female respondents who participated in a sociolinguistic interview. The selection of interview sites was of great relevance to this study since the major area of the Afro-Costa Rican cultural and linguistic elements is located from Puerto Limón and the neighbouring areas along the Línea Vieja (the old railroad line) up to Siquirres. Puerto Limón and the town of Siquirres are the two main economic areas, and the largest Afro-Costa Rican population resides there. Seventy-one were interviewed in Puerto Limón and fifty-six were interviewed in Siquirres.

6. The Profile of the Participants Of the 127 participants interviewed, fifteen were born in Costa Rica (outside the Province of Limón); 104 participants were born in the province of Limón and eight were born abroad. In this sample the participants’ age group is distributed as follows: sixty-one are in the 18-45 age group, forty five belong to the 46-65 age group and twenty-one are in the 66 and above age group. One hundred and twenty five (n=125) participants reported their educational level. Seventeen completed elementary school whereas one hundred and eight attended or completed high school, technical/vocational training or university. The occupations/professions varied for 125 participants. Fifteen did not have an occupation, twelve held jobs in the personal services category, twentytwo participants were involved in sales and office occupations and seventy-six occupied the professional category. One hundred and twentyseven (n=127) participants reported their marital status. Forty-six participants were single, fifty are married and thirty-one are either separated, divorced or widowed.

7. Methodology The sociolinguistic questionnaire was adapted from Rubin’s (1974) Questionnaire on Linguistic Use in Paraguay and supplemented with information from published articles such as Fridland (2003) and Milroy (2002). The questionnaire was designed in both English and Spanish. All the questionnaires were administered individually by the author to elicit

Language Maintenance and Shift among Afro-Costa Rican Women

183

information from participants about themselves. To protect the participants’ identity, all questionnaires were numbered and coded; the labels PLnn were assigned to the informants in Puerto Limón; SIQnn for the informants in Siquirres and PL/SIQnn to label those informants who were from Puerto Limón but were interviewed in Siquirres.

7. 1 Language of Interview As Figure 10-1 indicates, the interviews were conducted in the language chosen by the participant. Thirty-four participants chose to be interviewed in English, thirty-five chose Spanish and two participants chose to be interviewed in both languages.

Figure 10-1: Place and language of interview

7. 2 The Variables Tested Apart from asking the standard demographic questions (age group, place of birth, education level, marital status and occupation/profession), the questionnaire was designed to obtain information about the participants’ overall perceptions of their social and linguistic interaction with other members of the community. Though the sociolinguistic questionnaire comprised six parts, this article only focuses on two parts: one which included attitudes towards Spanish and the varieties of English spoken in the area. In this part the questionnaire covered five areas:

184

Chapter Ten

a. Language(s) participants identify with the most b. Language(s) participants feel more comfortable speaking c. Importance of speaking the ancestor’s language d. Participant’s effort in preserving Limón English e. Participant’s opinion of the teaching of Limón English in schools Another part included the participant’s inter ethnic and intra ethnic contacts and the nature and source of most of her contact with other ethnic groups, in order to determine the participants’ degree of interaction and integration with members of other ethnic groups (White/Mestizo Hispanics, Chinese, etc.). In this part the questionnaire covered three areas: a. Ethnic Group of participants b. Ethnicity of the participants’ social circle c. Ethnicity of the majority of participants’ contacts These variables were tested to determine any association between attitudes and identity in LMLS.

8. Analysis of Data Analyses of variance were conducted to determine the effects of attitude and identity in LMLS among the Afro-Costa Rican women in this sample. The level of significance was set at p 0.05. Of the many challenges I encountered in the analysis of these data, one was sorting out the labels participants used to designate the varieties of English they speak. These included: English, Limón Creole, Creole English, Patois, Mekatelyu, Broken and Jamaican English. For statistical purposes in the analysis of LMLS, it was necessary to lump all the varieties into one category labelled English. For this reason, table 1 is presented only to illustrate the broad range of responses that participants gave when I asked the question ‘what language(s) do you speak?’ Nonetheless, the various labels were important to provide us with knowledge of the participants’ general attitudes towards the varieties spoken in the two communities. Interestingly, of the 127 participants, two monolingual (English) participants as well as fifty-six English-Spanish bilinguals emphasized the word ‘English’ to express the language they speak. When I asked them what variety of English they speak, PL81 stated the following: ‘English is English especially if you read it and write it’, and SIQ47: ‘As long as people read and write in English, they speak English.’ Similar comments and reactions were shared by the other participants who view the language they speak as English and nothing else.

Language Maintenance and Shift among Afro-Costa Rican Women

185

(N= 127) Language(s) participants speak

Frequency

Spanish

1

Valid percent .8%

English

2

1.6%

Spanish and English

56

44.1%

Spanish and Limón English

24

18.9%

Spanish and Limón Creole

8

6.3%

Spanish and Patois

8

6.3%

Spanish, Limón English and Mekatelyu

1

0.8%

Spanish, Limón Creole and English

9

7.1%

Spanish and Creole English# #According to this speaker her Creole is a Nicaraguan Creole which she claims is different from the one spoken in Limón.

1

.8%

Spanish, Mekatelyu and English

3

2.4%

9

7.1%

Spanish and Broken English

1

0.8%

Spanish, Patois and English

2

1.6%

Spanish, Broken English and English

1

0.8%

Spanish and Jamaican English

1

0.8%

Spanish, Limón English and English

Table 10-1: Language(s) participants speak The term Mekatelyu has been used in publications (Wright 1982; Herzfeld 2002a) to designate Limón English Creole. There is a tendency to use the label ‘Broken’ as a folk name for Limón English Creole among some Blacks, whereas the label ‘Patois’ is also employed by both Blacks and Hispanics for the English spoken in the province. In order to

186

Chapter Ten

determine what these terms evoke for the participants, I asked them to explain the difference between Limón English, Limón English Creole, Mekatelyu and Patois and which varieties they considered ‘bad’ or ‘broken.’ I arranged their responses to the labels Mekatelyu and Patois based on shared judgments and views: from unawareness, through dissent to conformity. To highlight their views, I kept their statements in the original language (see notes for my translations of the comments made in Spanish).

9. Attitudes toward Mekatelyu and Patois Not all Limón English speakers recognized the term Mekatelyu as a language. Such was the case of PL11, PL18, PL31, PL45, PL48 and PL83 whose answers varied from ‘what is Mekatelyu?’ and ‘I don’t know who invented the term Mekatelyu’ to ‘No one speaks that here’. For those who recognized the term Mekatelyu, they considered it offensive because it is often perceived as an invention of the White/Mestizo population, especially from San José, to denigrate the English spoken by the Blacks in the province: SIQ12: ‘Según las personas de raza blanca, la gente de Siquirres habla Mekatelyu’4. PL/SIQ20: ‘Los blancos dicen que los negros hablan el Mekatelyu’5. PL40 ‘Mekatelyu is just an invention from the Spaniard’. PL87: ‘Mekatelyu es una forma despectiva del josefino’6.

The label Mekatelyu comes from the expression make I tell you [OGMCV'NLW] used in Limón English as an idiomatic expression and at times a support word. As expressed by PL07: No entiendo el concepto del inglés Mekatelyu. Mekatelyu es una expresión que han designado con el nombre del idioma. Mekatelyu es una ofensa, es como decir ‘Ud. no sabe hablar.’ No estoy de acuerdo con el Mekatelyu7.

This view is also supported by many participants in this study, like PL63 who stated ‘[I] don’t know what Mekatelyu is. It is just a word used in English, not the name of a language’ and PL79: ‘Mekatelyu is a word, is not a language’. PL31: ‘Mekatelyu and Patois have nothing to do with the way people speak here’.

Language Maintenance and Shift among Afro-Costa Rican Women

187

Patois, on the other hand, is regarded by some as the French spoken by the people of St. Lucia who settled in the province: PL18: ‘Patois is different from Mekatelyu. People from St. Lucia speak a different language than we’. PL81: ‘Patois speakers were a group of people from a French Island living in Cieneguita. Children of Patois speakers assimilated with English speakers. Mekatelyu is an expression; people who call English language Mekatelyu are wrong. They need to learn’.

Following a similar trend, some participants stated that Patois is broken French, a mixture of French and English or the Patwa spoken in Jamaica: SIQ01: ‘Patois mixed with French, here we speak English’. Referring to the language spoken in the province, she says: ‘I don’t accept it [the language] like Mekatelyu, White people say that Limonenses speak Patois/Mekatelyu; we’s American’. SIQ03: ‘Limón English is more better; [we] try to speak more better. Mekatelyu is bad language, Patois is a French dialect’. SIQ40: ‘Patois is French. There is no language here called Mekatelyu’. SIQ43: ‘In Cahuita they speak Mekatelyu - broken English-. Patois is not correct French. Westfalia people [in Limón province] speak that French’. PL/SIQ16: ‘Patois mix with French and English. Mekatelyu is an idiomatic expression. No language named Mekatelyu here’. PL60: ‘El patois es inglés y francés. Aquí en Limón ya no se habla el patois’8. PL61: ‘Patois es el inglés de Jamaica; mezcla de inglés y francés. El inglés limonense es un retazo de lo que quedó del patois, combinándolo con el español. Mekatelyu, también llaman así el inglés criollo. Viene de una expresión idiomática’9.

188

Chapter Ten

9. 1 Bad or Broken? SIQ49 claims that people often label broken forms of speech as Mekatelyu and Patois. She claims that her English is not entirely correct but it is not Patois as people often refer to it. She does not know what Patois is but claims that the Spanish gave it that name. Similar views came from SIQ35: ‘Creole is the broken language. There is no Patois here and Mekatelyu does not exist’; and from SIQ26: ‘I don’t know where people get Mekatelyu from. At home we speak a more relaxed English’. Contrary to the above statements, for SIQ44: Limón English is not broken English, but some people still consider it as broken. Patois means broken in any language; it means that your language is not correct. In Limón we have a Jamaican accent; both parents and teachers had it. People generalize by the accent and call it broken.

For the following participants, there seems to be an esthetic dimension of what the varieties represent. Although they claim that Limón English is different from Mekatelyu and Patois, Mekatelyu has a negative value attached to it since it represents a lower form, even more ‘imperfect’ from the one they speak: SIQ32: ‘El inglés de Siquirres es el criollo; aquí no se habla mekatelyu. El mekatelyu tiene variantes, por ejemplo: en mekatelyu se dice [C YG LW C IQ] o [MQOLC] y en criollo [YG LW IYC+0] o [MQO iGˆ] para decir ‘adónde va’ o ‘venga’ (respectivamente). El Patois es francés’10. SIQ48: ‘Mekatelyu es un inglés muy feo. Aquí no se habla eso’11. PL26: ‘Mekatelyu es más ordinario, rudo, rústico. Nuestro inglés es más refinadillo. Todos nos entendemos’12. PL53: ‘Inglés mekatelyu es como pachuco’13. PL73: ‘Patois es más bajo, flat, broken. El criollo es más elevado, más estándar. Mekatelyu es como patois pero no lo tengo muy claro lo que es mekatelyu’14. PL90:‘Broken English es mekatelyu, patois es más flat’15.

9. 2 Separate but Equal There are participants who did not make a distinction between Limón English, Mekatelyu and Patois, and for these participants, are all the same:

Language Maintenance and Shift among Afro-Costa Rican Women

189

SIQ41: ‘Mekatelyu y patois es lo mismo, lo único que es más broken [que el inglés limonense]. Todos los que hablan inglés, hablan mekatelyu’16. SIQ54: ‘Es lo mismo; lo importante es no perder la visión de la cultura’17. PL09: ‘Limón English is Mekatelyu’. PL34: ‘Here in Limón, we speak Mekatelyu’. PL36: ‘Is the same’. PL37: ‘Limón English is Mekatelyu; flat English. Ten percent [of the people] speak good English, all the rest is flat’. PL52: ‘Mekatelyu es el inglés limonense, todo el mundo le da diferente título, pero es lo mismo’18. PL62: ‘Mekatelyu is same as Creole’. PL70: ‘El inglés de Limón es patois’19. PL71: ‘Limón English [is] Mekatelyu; broken’.

And for other participants, if there is a distinction, it lies in the pronunciation: SIQ02: ‘The pronunciation is different’. SIQ13: ‘The difference between Limón English and Patois is in the pronounce, [the] way people spell words, plenty difference’. PL/SIQ11: ‘…en la pronunciación hay formas diferentes de hablar’20. PL15: ‘Certain words are different; pronounced differently’. SIQ36: ‘There is a difference in the pronunciation’. PL55: ‘Mekatelyu pronunciation is different. Rules are not well applied’. PL56: ‘Mekatelyu and Patois are the same thing; is a matter of pronunciation’.

190

Chapter Ten

9. 3 English and Some Spanish = Mekatelyu Studies have argued that code switching and code mixing are some of the features present in Limón English. For the following participants, Mekatelyu represents a combination of English and Spanish. As stated by these three participants: PL02: ‘Mekatelyu mix of Spanish and English; everybody mix English and Spanish’. PL21: ‘Mekatelyu es una mezcla con un poquito de español y el otro [patois], no’21. PL44: ‘Mezclan palabras en inglés y usos de muletillas en español. Para describir una persona arrogante y pesada: es que she fresh’22.

10. Factors Related to Attitudes To further examine their attitudes, participants were asked to respond to the following questions: Which language do you identify yourself with the most? Which language are you most comfortable with (speaking)? How important is it to speak your ancestor’s language? Do you make an effort in preserving Limón English? Do you think that Limón English should be taught in schools? (The results are displayed in Table 2, next page). Of the 127 participants, thirty-seven (29.1%) reported that they identified themselves with Spanish; forty-four (34.6%) identify with both languages whereas forty-six (36.2%) identify only with English. For SIQ25, the English language is one of the reasons she lives in Siquirres, and she identifies with speakers of it. On the other hand, for SIQ40 it is not with the language but the group that she identifies with the most. Among friends she speaks ‘broken’ but if she is introduced to someone from abroad she will try to modify her English. For PL84, Limón English is part of her roots. She claims that she is not interested in British or American English. SIQ47 stated that she tries not to speak Spanish except if necessary. Of the 126 participants who responded to the question ‘which language are you most comfortable with speaking?’ forty-six (36.5%) reported Spanish; forty-four (34.9%) feel equally comfortable speaking both languages and thirty-six (28.6%) felt more comfortable speaking English. It appears that is sometimes difficult to handle certain concepts in English, especially when it comes to the professional setting.

Language Maintenance and Shift among Afro-Costa Rican Women

191

(N= 127) Attitudes toward Spanish, Limón English and its varieties

Frequency

Valid percent

37

29.1%

Spanish/English

44

34.6%

English

46

36.2%

Spanish

46

36.5%

Spanish/English

44

34.9%

English

36

28.6%

Not important

2

1.6%

Somewhat

4

3.3%

117

95.1%

NO

18

15.3%

YES

100

84.7%

NO

59

49.6%

YES

60

50.4%

Language(s) participant identifies with the most Spanish

Language(s) participant feels more comfortable speaking

Importance of speaking ancestor’s language

N

127

126

123

Very Effort in preserving Limón English

Should LE be taught in schools

118

119

Table 10-2: Participants’ attitudes towards Spanish, Limón English and its varieties

192

Chapter Ten

Of the 123 participants who answered the question: ‘how important is it to speak your ancestor’s language?’ only two (1.6%) stated that it was not important whereas four (3.3%) stated that it was somewhat important. However, for the overwhelming majority, 117 (95.1%), speaking the ancestor’s language is very important. Of the 118 participants who answered the question ‘do you make an effort in preserving Limón English?’ only eighteen (15.3%) stated that they did not make an effort to preserve Limón English. Though the majority, 100 participants (84.7%) appeared to be supporters of the maintenance of their language, the reality for some like SIQ22 is that the environment does not lend itself to speaking the heritage language. In Siquirres, there seem to be a consensus among the participants that many of the traditions left by the Black West Indians are dying out, and consequently this is affecting the fate of their language. For SIQ09, maintenance of the language is important for the maintenance of the culture and ethnicity. As SIQ18 explains: ‘Ya se está perdiendo mucho la identidad, los orígenes. Necesitan saber de dónde viene su lengua materna’23. In Siquirres, this may be attributed to the larger part of the population being White/Hispanic or mixed, who are no longer exposed to the traditions of the Blacks. SIQ15 is an English teacher who said that she does not have children and her family prefers to speak Spanish. SIQ21 explained that the majority of her friends do not speak Limón English, and she does not have the strength to maintain it. For her as well for PL/SIQ07, when there is mixture of different ethnic groups, the language disappears. SIQ46, who happens to be the daughter of SIQ41, stated ‘No practico el idioma, todo mi entorno es hispano. Mi mundo es hispano’24. Similar comments were made by speakers in Puerto Limón. PL20 stated that English is temporary; her environment is Spanish since her husband is White/Hispanic. PL43 mentioned that her friends are ‘White’ and she does not pay attention to the people when they tell her to speak English: ‘No le ago caso a la gente cuando me dice que hable inglés. Mis amistades son blancas’25. Another problem that participants face is the feeling that that the younger generatibon is not interested in maintaining Limón English. SIQ01 stated that only the aged people speak Limón English; young people mostly want to speak Spanish. As SIQ53 puts it, ‘the youth don’t want to speak the English, they are favouring Spanish’. For SIQ54 there is a resistance among her children to speaking the language. She adds to that ‘a los hijos les molesta la mezcla del inglés y español ya que ellos están aprendiendo el inglés americano en la escuela y en el colegio. Por eso prefieren hablar conmigo en español’26. .

Language Maintenance and Shift among Afro-Costa Rican Women

193

Participants complained that even among the Blacks there is embarrassment and shame in speaking the language: SIQ38: ‘Hay muchos negros de granja27 que no quieren hablar inglés’28. PL85: ‘En Limón la gente rechaza el inglés real’29.

As PL/SIQ16 states: People still feel embarrass talking their language; they don’t know the value of being different. People are ashamed. For my family is clear that their language has to continue, to maintain the heritage.

The following two participants stated that they did not make an effort to maintain Limón English because as PL48 stated, she waits for people to make the first step to talk to her in English and for SIQ48 ‘solo lo hablo cuando lo debo hablar’30. In addition to the examples above, there are other reasons why speakers reject maintaining the language:

1. Raised by ‘Spanish’ Speaking Mother In Limón there is a common expression: ‘si la mamá es blanca y el papá es negro, los hijos crecen hablando español’31. This was the case for PL14, PL19 and PL89. PL14, whose father is Black, was raised by her mother who is Hispanic. Both her father and paternal grand-parents are speakers of English but the language was not transmitted to her. PL19 was also raised by a Spanish speaking mother and a Spanish/English speaking father. Although she claims to speak a standard form of English, according to her it presents many difficulties. She does not identify herself with any of the varieties of English spoken in the area, but prefers to speak a standard form of the language, the one she is learning at the university. PL89’s mother only spoke Spanish, but her maternal grandmother spoke Spanish and English. Spanish was always spoken at home. She speaks English even though she was not raised with the language.

2. To Avoid being Ridiculed by the Spanish People PL80 says that she speaks Limón English to communicate with others but she prefers Spanish. Her experiences and training have been in Spanish.

194

Chapter Ten

In the university she decided to improve her Spanish because her classmates made fun of her pronunciation. As she puts it: En mi generación, el inglés me ocasionaba problemas en la pronunciación del español, debido a eso les hablo español a los niños, para evitar burlas32.

That is why she had to define herself as a Costa Rican. Speaking English has only helped her at the international level.

3. A Matter of Trust The following participants said that speaking Limón English is a matter of trust, even among Blacks. PL87: ‘el español tiene un impacto mayor, el inglés representa confianza y cercanía’33. PL/SIQ 27 only speaks English at home and with people very close to her. For her, speaking English at work means that she has acquired a level of familiarity in the workplace; if she feels that she has not reached that level of trust, she will not speak to her co-workers in English: Si hay negros de los cuales no les tengo confianza, no les hablo en inglés. Hablar inglés es adquirir un grado de confianza, sobretodo en el trabajo. El español es para todo lo administrativo. Mis padres se ríen cuando hablo correcto. La gente me mira con mala cara cuando uso un inglés más estándar, piensan que es pretensión o aires de grandeza34.

4. Preference for ‘Standard’ English Participants who are English teachers are the main advocates of ‘Standard’ English. PL55, who is an English teacher, mentioned that she does not know how to speak Limón Creole. The language in her environment both at work and at school is in Standard English and she watches television in American English. Another English teacher PL/SIQ29 stated that she speaks a combination of American and British English. For her Standard English is American or British English and she claims that in Limón most of the people speak a standard English which is British English, though some people speak Creole: ‘I could not speak Creole at home. I had to speak Standard English’. For her, English is more dominant, and she claims she learned Spanish watching television. Even though PL73 is a speaker of Limón English, she is an English teacher and makes an effort to improve her American English by communicating in a more Standard English. SIQ28 is an English teacher and according to her, ‘I try to practice to talk the standard way’. SIQ19

Language Maintenance and Shift among Afro-Costa Rican Women

195

who is also a teacher, only speaks Limón English with friends and in the province of Limón but in all other situations she speaks American English. Other participants who lived in the States also displayed admiration for American English. SIQ37 lived in New York for 33 years and prefers to speak American English, for her: ‘no me interesa ese inglés, prefiero el inglés americano’35. Participants were divided concerning the teaching of Limón English in schools. Of the 119 participants who responded to the question: ‘do you think that Limón English should be taught in schools? Fifty-nine (49.6%) stated that Limón English should not be taught in schools whereas sixty (50.4%) stated that it should. Some believe that Limón English should be taught only in the province (SIQ09, SIQ32, SIQ42) or at least it should be taught as a subject (PL20) in order that Costa Ricans know the origins of the Black people of the province (SIQ18) and to maintain the culture (PL52). Participants (PL07 and PL11) stated that not only Limón English but also the indigenous languages should be taught at schools. SIQ25 mentioned that people already learn the oral tradition at home; therefore they should learn the standard at school. SIQ24 believe that it is their native language and is part of their environment, so she favours the teaching of Limón English at schools since it is very important for the maintenance of their language. For the participants who stated that Limón English should not be taught in schools, many believed that it is the standard form (either American or British) that should be taught. For these and many more participants, schools should teach ‘correct’ (proper) English: SIQ02: SIQ05: SIQ 10: PL/SIQ11: PL83: PL82: PL80:

‘El inglés correcto; necesitamos un cambio’36. ‘Es mejor enseñarles el verdadero inglés’37. ‘They should learn the right way’. ‘El inglés perfecto sí’38. ‘English is the key for everything; a better English’. ‘Standard American English’. ‘Correct British English’.

Most participants are fully aware that Limón English is not recognized internationally, and the market only caters for what it is considered proper English (mainly American English). In the words of SIQ55, ‘si no aprenden el inglés americano, no le dan trabajo a ninguno’39. Most bilingual schools in the province praise the teaching of Standard English. In some bilingual schools, English teachers are not even allowed to use the vernacular outside the classroom, not even among their

Chapter Ten

196

colleagues. According to the Director of the Colegio Experimental Bilingüe in Siquirres, English classes are conducted in American English. If the English taught is considered ‘negro caribeño’40 the parents of the students complain, even those who cannot speak the language.

11. Factors related to Ethnicity and Ethnic Relations Of the 126 participants who reported their ethnic background, an overwhelming majority of 115 (91.3%) claimed to be Blacks as opposed to eleven (8.7%) who claimed to be of a mixed ethnic background (Table 103). (N =126) Frequency

Valid Percent

Mixed

11

8.7%

Black

115

91.3%

Total

126

100.0%

Ethnicity

Table 10-3: Ethnicity of participants The ethnic composition of the social groups of ninety-five participants (78.4%) consists of an equal mixture of two or more ethnic groups, as Table 10-4 indicates.

Social Group

(N =127) Frequency

Exclusively from the same background as you

Valid Percent 7

5.5%

Mainly from the same background as you

19

15.0%

An equal mixture of two or more ethnic groups

95

74.8%

6

4.7%

127

100.0%

Mainly from a different background as you Total

Table 10-4: Social group ethnic composition

Language Maintenance and Shift among Afro-Costa Rican Women

197

(N=126) Frequency

Valid Percent

Yes

63

50.0%

No

45

35.4%

Equal Mixture

18

14.3%

126

100.0%

Contacts outside ethnic group

Total

Table 10-5: Do you consider the majority of your contacts to be with people outside your ethnic group? (N= 127) N Language(s) participant identifies with the most Spanish Spanish/English English Language(s) participant feels more comfortable speaking. Spanish Spanish/English English Importance of speaking ancestor’s language. Not important Somewhat Very Effort in preserving Limón English NO YES Should LE be taught in schools NO YES

X (S.D.)

p-value

0.64 (.11) 0.72 (.11) 0.81(.09)

0.001

0.64 (.11) 0.74 (.09) 0.84 (.07)

0.001

0.62 (.009) 0.65 (.07) 0.74 (.11)

0.08

0.62 (.13) 0.75 (.10)

0.001

0.71 (.12) 0.75 (.12)

0.12

127

126

123

118

119

Table 10-6: Language attitudes associated with LMLS (The mean (X) range is 0 – 1. A mean of 1 indicates most likely to maintain English)

198

Chapter Ten (N= 127) N

Location of people with whom participant spends most of her time

X (S.D.)

126

0.463

Inside the community

0.74 (.12)

Inside and outside the community

0.75 (.13)

Outside the community

0.71 (.11)

Ethnic background

126

0.013

Mixed

0.65 (.17)

Black

0.74 (.11)

The majority of participants contacts are with people outside their ethnic group Yes

126

0.037 0.71 (.11)

No

0.77 (.12)

Equal Mixture

0.72 (.14)

Social group ethnic composition Exclusively from the same background Mainly from the same background An equal mixture of two or more ethnic groups Mainly from a different ethnic group

p-value

127

0.001 0.86 (.07) 0.84 (.08) 0.70 (.12) 0.72 (.10)

Table 10-7: Ethnicity and inter/intra ethnic relations associated with LMLS (The mean (X) range is 0 – 1. A mean of 1 indicates most likely to maintain English)

Language Maintenance and Shift among Afro-Costa Rican Women

199

Sixty-three participants (50.0%) stated that the majority of their contacts take place outside their ethnic group, in contrast to forty-five (35.4%), whose majority of contacts are with people inside their ethnic group (Table 10-5, page 196). For most participants, the majority of their contacts with other ethnic groups take place almost everywhere in the street and in every aspect of their life, more specifically at work.

12. Findings 12. 1 Language Attitudes High level of significance of LMLS was found with the language participants identify themselves with most, the language participants felt more comfortable speaking and how much effort they exert in preserving Limón English. Even though the importance placed in speaking the ancestor’s language was higher for most participants, the association between it and LMLS was of borderline significance. The teaching of Limón English in schools was not associated with LMLS (see Table 10-6, page 196).

12. 2 Factors Associated with Ethnicity As Table 10-7 (see page 197) shows, the results shown in the Ethnicity and inter/intra ethnic relations are surprising. Only the ethnic composition of the participants’ social group ethnic composition was highly significant in its association with LMLS (Table 10-7). Participants’ exclusive interaction with members of their same ethnic background was much higher that those who interacted with an equal mixture of two or more ethnic groups. Both ethnicity and contacts with members outside their ethnic groups were significant but not to the extent of the ethnic composition of the participants’ social group. The location of people with whom the participants spend most of their time did not prove to be statistically significant. Participants’ mobility inside or outside the community was not a factor associated with LMLS.

13. Discussion In terms of language attitudes, the results indicate that English maintainers are located within the subgroup of women who identify themselves most with English, who feel more comfortable speaking it, and

200

Chapter Ten

who make efforts to preserve Limón English. Despite the positive impact that identification with English may have in its maintenance, the reality for most participants is that Spanish only and Spanish in combination with English are the languages in which they feel more comfortable. It is also worth mentioning that, even though the majority of the participants believe that it is very important to speak their ancestral language, its impact in maintenance of English is uncertain as it was demonstrated to have a borderline significance. As for the teaching of Limón English in schools, this was not an influential factor in the maintenance of English. Although the ethnic background of the participants was not an influential factor in maintenance of English, the ethnic composition of their social group was notably influential. Higher English maintenance was found among women whose social group is exclusively or mainly of the same ethnic group of the participants, that is, Black. In addition, English maintainers are located within the subgroup of participants whose contacts are within their ethnic group regardless whether these contacts are inside and outside their community. The majority of the participants who claimed to socialize with members of different ethnic groups exceeded by far those women who socialize exclusively or mainly with members of the Afro-Costa Rican community. This brings us to the next question: what’s in a name? The problem of demarcation of language varieties as studied by Klingler (2005) in the varieties of French spoken in Louisiana reveals that stigmatized Creole languages, like Louisiana Creole, are often considered corrupted versions of their lexifying language. Klingler (Ibid) explains that the problem becomes more complex when labels are attached to designate the language used by those who use it. As noted, the participants had different opinions about the varieties of the English spoken and the value and need to maintain it. Their perceptions concerning maintenance of English seem to be motivated more by their individual choices, which are often generated by external pressures and perceptions mainly of the Josefino, the people of the capital who, according to the participants, view their language with disdain. Will the shift of Limón English be toward Spanish or to Standard English? In Spence’s (1998) study Language Attitudes of Limon Creole Speakers, respondents, who appeared to be female in the majority, displayed both a general admiration for ‘Standard English’ and a strong support for the maintenance of Limón Creole (LC) in the community. However, for many, passing on their language (LC) to younger generations represents a conflict. Spanish, as the author puts it, ‘is thought of in terms of status, academic and social advancement’ (111). Despite

Language Maintenance and Shift among Afro-Costa Rican Women

201

Spence’s conclusions, for the Afro-Costa Rican women in this study, the ability to speak English has facilitated employment in many sectors of their professional lives. Some participants in this study are English teachers; others work in commercial and tourist areas where they are in constant interaction with English speakers. The strong supporters of preserving English are aware that English is the key for success. For these women English is as highly valued as Spanish.

14. Concluding Remarks It is interesting to speculate on what the future holds for the English linguistic practices of the Afro-Costa Rican population, as their country has entered a new economic phase by opening its doors to the economic power of the United States. Lara, Barry and Simonson (1995, p. 143) describe the effects of the American in Costa Rica as follows: To the first-time visitor, the extent of the U.S, influence in Costa Rica is startling. Menus in the classier restaurants are commonly bilingual, U.S. brand-name products are readily available and there is a large U.S. community. Sale signs for choice houses and property are frequent only in English. To an extent not seen elsewhere in Central America, Costa Rica has been inundated with American Consumer culture - in the food people eat, the style of dress they wear and even the brand of political ideology they endorse. (Lara et al. 1995, p. 143)

Costa Rica has for some time aggressively tried to attract American retirees, most of whom speak only English. In addition, Costa Rica is noted for its ecotourism and attracts many foreign tourists. The large number of English speaking American tourists, as well as the increasing use of English as a lingua franca for non-native English speaking tourists, adds another element possibly affecting the vitality of Limón English, which becomes less useful in this context. However, it should be noted that the increasing economic opportunities generated by tourism and the presence of a significant number of American retirees may well become another factor reinforcing change in the English spoken by the Afro-Costa Rican population in and outside the province of Limón. The degree of importance that the vernacular receives from its speakers will eventually determine its fate. Maintenance of Limón English will require a strong implementation of educational policy planning as well as educating its people about the importance of speaking it, in order to revitalize the linguistic practices of the language. If not, like many linguistic minorities, Limón English will find its refuge at homes and

202

Chapter Ten

among those speakers who feel that it should be preserved as a symbol of their West Indian Black heritage.

Notes 1

Source: Colección de Leyes y decretos 1949 -1950. No 836 p. 677 (4 de noviembre de 1949) Artículo 1º 2 Source: Barrantes, V. La República, October 1997. 3 According to Herzfeld, who obtained the data from the publication Ethnologue: Languages of the World (Grimes 1997 online) and supported by Norval Smith (1995:341), the number of Limón Creole English speakers is 55.100. 4 According to the White people, the people in Siquirres speak Mekatelyu. 5 Whites say that Blacks speak Mekatelyu. 6 Mekatelyu is a pejorative form used by the people of San José. 7 I don’t understand the concept of Mekatelyu English. Mekatelyu is an expression that is used to label the language. Mekatelyu is offensive and it is like saying ‘you don’t know how to speak.’ I do not favor Mekatelyu. 8 Patois is English and French. Here in Limón, Patois is no longer spoken. 9 Patois is the English of Jamaica; [a] mixture of English and French. Limonense English is a remnant of what was left of Patois, combined with Spanish. Creole English is also called Mekatelyu. It comes from an idiomatic expression. 10 Siquirres English is Creole. Here Mekatelyu is not spoken. Mekatelyu has its variants, for example, in Mekatelyu one says [C YG LW C IQ] o [MQOLC] and in Creole [YG LW IYC+0] o [MQO KGˆ] to say ‘where are you going’ or ‘come here’. Patois is French. 11 Mekatelyu is ugly English; it is not spoken here. 12 Mekatelyu is more ordinary, tough and rough. Our English is more refined. We all understand one another. 13 Mekatelyu English is like an uneducated form of speech. 14 Patois is lower and Creole is more elevated and more standard. Mekatelyu is like a Patois but it is not quite clear to me what Mekatelyu is. 15 Broken English is Mekatelyu, Patois is more flat. 16 Mekatelyu and Patois are the same only that it is more broken [than Limón English]. Everyone who speaks English, speak Mekatelyu. 17 It’s all the same; the important aspect is that we should not lose sight of the culture. 18 Mekatelyu is Limón English, everyone gives it a different title but it is the same. 19 Limón English is Patois. 20 ...in the pronunciation there are different ways of expressing it. 21 Mekatelyu is a mixture with a little bit of Spanish and the other one [Patois] is not. 22 [They] mix words in English with support words in Spanish. To describe an arrogant and unpleasant person: es que she fresh. 23 We are losing our identity and origins. They need to know where their maternal language comes from.

Language Maintenance and Shift among Afro-Costa Rican Women

203

24

I don’t practice the language. My environment is Hispanic. My world is Hispanic. 25 I don’t pay attention to people when they tell me to speak English. My friends are White. 26 My children are annoyed when Spanish and English are mixed, since they are learning American English in Elementary and High School. For that reason, they prefer to speak with me in Spanish. 27 The expression ‘negros de granja’ in this context does not mean ‘field Blacks.’ It is an expression used for Blacks who pretend to be White. 28 There are many Blacks who pretend to be to be White who do not want to speak English. 29 In Limón, people reject the real English. 30 I only speak it when I have to. 31 If the mother is White and the father is Black, the children are raised speaking Spanish. 32 In my generation, English interfered with my pronunciation in Spanish, for that reason, I speak Spanish to the children, to avoid being ridiculed. 33 Spanish has a larger impact, English represents trust and closeness. 34 If there are Blacks whom I don’t trust, I don’t speak to them in English. Speaking English means acquiring a level of trust, especially at work. Spanish is for all administrative matters. My parents laugh at me when I speak properly. People look at me negatively when I use a more Standard English; they think that it is snobbery or airs of grandeur. 35 I don’t care about that English [referring to Limón English] I prefer American English. 36 We need proper English; we need a change. 37 It is better to teach them the real English. 38 Perfect English, yes. 39 If they don’t learn American English, [people] are not considered for jobs. 40 Caribbean Black English.

References and Further Reading Bourgois, P. 1989. Ethnicity at Work. Divided Labor on a Central American Banana Plantation. Baltimore and London: The Johns Hopkins University Press. de Wille, B., and Eugenia, M. 2005. Bryce Laporte, Roy Simon. 1962 – Social Relations and Cultural Persistence (Or Change) Among Jamaicans in a Rural Area of Costa Rica. In: Q. Duncan and C. Meléndez (eds.), El negro en Costa Rica. Costa Rica: Editorial Costa Rica. pp. 265-276. Casey, G. J. 1979. Limón 1880 -1940. Un estudio de la industria banarera en Costa Rica. San José: Editorial Costa Rica.

204

Chapter Ten

Chomsky, A. 1996. West Indian Workers and the United Fruits Company in Costa Rica, 1870- 1940. Baton Rouge: Louisiana State University Press. Clyne, M. 2005. Australia’s Language Potential. Sidney, Australia: University of New South Wales Press Ltd. Dabène, L., and Moore, D. 1995. Bilingual Speech of Migrant people. In: L. Milroy and P. Myusken (eds.), One speaker, two languages. Crossdisciplinary perspective on code-switching. Cambridge: Cambridge University Press. pp. 17- 44 Duncan, Q. 2005. El negro antillano: inmigración y presencia. In: Q. Duncan and C. Meléndez (eds.), El negro en Costa Rica. Costa Rica: Editorial Costa Rica, pp. 115- 175. Echeverri-Gent, E. 1992. Forgotten workers: British West Indians and the early days of the banana industry in Costa Rica and Honduras. Journal of Latin American Studies, 24(2): pp. 275-308. Fishman, J. 1972. Language in Sociocultural Change. Stanford, California: Stanford University Press. Fishman, J. 1978. Advances in the Study of Societal Multilingualism. The Hague, Paris, New York: Mouton Publishers. Fridland, V. 2003. Network strength and the realization of the southern vowel shift among African Americans in Memphis, Tennessee. American Speech, 78(1): pp. 3-30. Gardner, R. C. 1988. Attitudes and motivation. Annual Review of Applied Linguistics, 9: pp. 135-148. Gearoid, C. 1984. The affective dimension in second/foreign language learning: An interactional perspective. CLCS Occasional Paper No. 11. Dublin, Ireland: centre for language and communication studies. Genesee, F., Rogers, P., and Holobow, N. 1983. The social psychology of second language learning: Another point of view. Language Learning, 33(2): pp. 209-224. Herzfeld, A. 1978. Vida o muerte del criollo limonense. Revista de Filología y Lingüística 4(2): pp. 17-24. —. 1998. Cohesion in Limonese Creole. In: C. Pauline, B. Lalla, V. Pollard and L. Carrington (eds.), Studies in Caribbean Language II. Papers from the Ninth Biennal Conference of the Society for Caribbean Linguistics, 1992. St. Augustine, Trinidad, West Indies: The Multimedia Production Centre School of Education. Faculty of Humanities and Education. The University of the West Indies, pp. 163 – 177. Herzfeld, A. 2002a. Mekaytelyuw – La lengua criolla. Costa Rica: Editorial de la universidad de Costa Rica.

Language Maintenance and Shift among Afro-Costa Rican Women

205

—. 2002b. Los criollos de base inglesa en Centro América: Investigaciones recientes. In: N. Díaz, R. Ludwig and S. Pfander (eds.), La Romania Americana. Vervuert, pp. 265-227. Hill Hudson, B. 2001. African American Female Speech Communities – Varieties of Talk. Wesport, Connecticut; London: Bergin and Garvey. Holm, J. (ed.) 1983. Varieties of English Around the World - Central American English. Julius Groos Verlag Heidelberg Druck und buchbinderische Verarbeitung: Beltz Offsetdruck Hemsbach. Instituto Nacional de Estadísticas y Censos. 2001. I-59n IX Censo Nacional de Población y V de Vivienda del 2000: Resultados Generales/Instituto Nacional de Estadística y Censos. San José, C.R: INEC. http://www.conedsa.com/biblioteca/INEC,%202001%20%20Resultados%20Censo%202000.pdf Klingler, T. 2005. Le problème de la démarcation des variétés de langues en Louisiane: étiquettes et usages linguistiques. In: A. Valdman, J. Auger, and D. Piston-Hatlen (eds.), Le français en Amérique du Nord: Etat-présent. Quebec: Les Presses de l’Université Laval, pp. 349-367. Lara, S., Barry, T., and Simonson, P. 1995. Inside Costa Rica. The Essential Guide to its Politics, Economy, Society, and Environment. Albuquerque, New Mexico: Resource Center Press. Masís Morales, O., and Mora Lobo, E. 1985. Las oraciones completivas en el criollo de Limón: Análisis sintáctico de un idiolecto. Revista de Filología y Lingüística, XI (2): pp. 111-121. Massey, D. A. 1986. Variations in attitudes and motivation of adolescent learners of French as a second language. Canadian Modern Language Review, 42(3): pp. 607-618. Meléndez, C. 2005. Aspectos sobre la inmigración jamaicana. In: Q. Duncan and C. Meléndez (eds.), El negro en Costa Rica. Costa Rica: Editorial Costa Rica., pp. 77 -111. Milroy, L. 2002. Social Networks. In: J. K. Chambers, P. Trudgill, and N. Schilling-Estes (eds.), The handbook of language variation and change. Massachusetts, United States of America., Oxford, United Kingdom: Blackwell. Municipalidad de Limón. 1992. Luchas y esperanzas.100 años de historia doble e inconclusa del cantón de Limón. Uruk Editores, S.A. Paulston, C. B. 1994. Linguistic Minorities in Multilingual Settings. Implications for Language Policies. Amsterdam/Philadelphia: John Benjamins Publishing Company. Putnam, L. 2002a. The Company They Kept. Migrants and the Politics of Gender in Caribbean Costa Rica, 1870 – 1960. Chapel Hill and London: The University of North Carolina Press.

Chapter Ten

206

Putnam, L. E. 2002b. La población afrocostarricense según los datos del Censo 2000. Presentado en el simposio “Costa Rica a la luz del censo 2000”. Costa Rica 5 y 6 de agosto del 2002. Romaine, S. 1995. Bilingualism. Oxford, United Kingdom and Cambridge, United States of America: Blackwell. Rubin, J. 1974. Bilingüismo nacional en el Paraguay. México, D.F: Instituto Indigenista Interamericano. Søndergaard, B., and Norrby, C. 2006. Language maintenance and shift in the Danish community in Melbourne. International Journal of the Sociology of Language, 180: pp. 105-121. Spence Sharpe, M. 1997. A case study of language shift in progress in Port Limon, Costa Rica. Revista de filología y lingüística de la Universidad de Costa Rica. XXIII(1): pp. 225-234. Spence Sharpe, M. 1998. Language attitudes of Limon Creole speakers. Revista de Filología y Lingüística. XXIV(1): pp. 101-112. Stewart, R. 1999. Limón Real. San José: Litografía e Imprenta LIL, S.A. Stone, G. B., and Rubenfeld, S. A. 1989. Foreign languages and the business curriculum: What do the students think? Modern Language Journal. 73(4): pp. 229-239. Taylor, D. 1968. New Languages for Old in the West Indies. In: J. Fishman, J. (ed.), Readings in the Sociology of Language. The HagueParis: Mouton, pp. 607-619. Winkler, E. G. 2000. Cambio de códigos en el criollo limonense. Revista de filología y Lingüística de la Universidad de Costa Rica. XXVI(1): pp. 189-196. Wright M. F. 1982. Problemas y métodos para la enseñanza del inglés como segunda lengua a los hablantes del mek-a-tel-yu en la provincia de Limón. Filología y Lingüística, 8 (1 v 2): pp. 129-135.

Exercises 1. Explain which of the following are most likely to be the key factors associated with attitudes and language maintenance and shift among the speakers of Limón English: x x x x x

Personal identification with the languages spoken Comfort in speaking the languages Importance of speaking their ancestors’ language Efforts in preserving Limón English The teaching of Limón English in schools

Language Maintenance and Shift among Afro-Costa Rican Women

207

2. In this paper, does ethnicity play an influential role in the maintenance of the English spoken among the participants? Explain why you think it does or does not. 3. Despite the fact that the overwhelming majority of the participants describe themselves as supporters of their language, they were divided about the teaching of Limón English in schools. Do you believe that their views can be attributed to the stigmatizing of Limón English by the “White” Costa Rican, or that their views are related to the economic pressures of the country, that is, speaking Standard English will provide more employment opportunities for them in Costa Rica? Explain. 4. Given the fact that there is a shift in status, what conclusions can you draw about the fate of Limón English based on the participants’ attitudes concerning their language? In your opinion, in which direction will the shift take place in future generations? Spanish? Spanish and Limón English? Or Spanish and Standard English? Explain. 5. Do you think that the shifts in status of Limón English as described are unique to that country and its particular circumstances, or can they be generalized to language choices made in other countries? Do you know any parallel cases?

CHAPTER ELEVEN SURVIVAL AGAINST ALL ODDS: LONGEVITY OF SRI LANKA PORTUGUESE CREOLE SHIHAN DE SILVA JAYASURIYA1

1. Introduction The Portuguese encounter with Sri Lanka in the early 16th century was driven by trade and proselytisation in the East. An accidental visit to Galle in 1505, whilst on a voyage to the Maldives, was followed by planned visits to Sri Lanka. Establishing trading posts, constructing fortresses and settlements began twelve years later, from 1517 onwards. This was followed by colonisation of parts of the Island, and atypically in Asia, not merely in the coastal areas but also inland. The Portuguese were dragged into local politics and they acquired a desire to colonise the entire island. Various plans to achieve this goal failed and the Dutch routed the Portuguese. But neither the Dutch nor the British who followed them could wipe away the Portuguese linguistic imprint on the island. Though the period of domination by the three colonisers was almost the same – 1505 to 1658 (Portuguese), 1658 to 1796 (Dutch) and 1796 to 1948 (British) – the extent of land controlled by each coloniser varied. Similarly, the imprints made by the Europeans varied and the intangible heritage are strong. By 1638, the Portuguese controlled a large area of the island. The Kandyan kingdom remained independent excepting for Batticaloa and Trincomalee which came under Portuguese control only for a few years. A fortress was built by the Portuguese in Trincomalee in 1623 and five years later, in 1658, a fortress was built in Batticaloa. These two areas have become strongholds of the Portuguese creole language – Sri Lanka Portuguese Creole (SLPC) – and will be discussed in this paper. Being related to Indo-Portuguese, this contact language was known as Indo-Portuguese of Ceylon in the 19th century (Ceylon being the name given to the island by the British). Hugo Schuchardt, the ‘Father of Creole

Survival Against All Odds

209

Linguistics’ considered the Indo-Portuguese of Ceylon to be the most important language in this group (de Silva Jayasuriya 1999). IndoPortuguese served as a lingua franca and was a bridging tongue between the three colonisers and the locals. The Dutch also spoke Indo-Portuguese due to intermarriage with mestiças (people of Portuguese and Sri Lankan descent). Although Dutch was used in administration, Dutchmen spoke creole at home with their families. The British learnt the creole when they administered the island, particularly during the initial phase, until English became the new lingua franca.

2. Contemporary Sri Lanka Portuguese Creole The Portuguese linguistic imprint in Sri Lanka is two-fold. On the one hand, it is apparent in a contact language which resulted from the Portuguese – Sri Lankan encounter - Sri Lanka Portuguese Creole (SLPC), which is still spoken. On the other hand, it is apparent in other Sri Lankan languages – Sinhala, Tamil and Sri Lanka Malay – through Portuguese words which have lasted for more than 350 years and have now become part of the vocabulary of these languages (de Silva Jayasuriya 2010a). Firstly, linguistic features of SLPC which deviate from European Portuguese will be illustrated through data obtained in a booklet of ‘Portuguese Cantignas’ (Cantigas or songs) (1976) from the Catholic Burgher Union of Batticaloa. The Catholic Burgher Union was established in 1927 to preserve the language and culture of the Burghers. Naturalistic data that I obtained from Batticaloa Burghers in 2008 and in 2011 are also included. Whilst considering linguistic features that deviate from European Portuguese, this paper theorises on the survival of SLPC for half a millennium. When the Goan Jesuit, Sebastião Rodolfo Dalgado (1900) was Vicar-General in Sri Lanka (Ceilão to the Portuguese), at the end of the nineteenth century, Indo-Portuguese of Ceylon (SLPC) was predicted to become moribund. SLPC is the mother-tongue of people now called Portuguese Burghers (meaning citizen) which now encompasses the term mestiço/mestiça (meaning of mixed parentage). Central to the activities of the Burghers is the Catholic Burgher Union, which provides inspiration for cultural maintenance. On the bright side, several young Burghers are involved with the activities of the Union. Although religion is a binding force for the Burghers, it does not differentiate them from other ethnic groups. Some Sinhalese and Tamils are also Catholics. But language, music and dance differentiate the Portuguese Burghers from other groups.

210

Chapter Eleven

Free education for the masses was introduced in the 1940s and all Sri Lankans are entitled to entirely government-subsidised, ‘free’ schooling and university education. Places for tertiary education are competitive because the number of places at universities is limited. But education is available in Sinhala, Tamil or English and these three languages are now official languages. There is no pressure for small minorities such as the Portuguese Burghers or Malays to study in their mother-tongue. According to Achinthya Bandara (2010), an undergraduate at the University of Colombo (Sri Lanka), who carried out field work in the Eastern Province, Burghers live in several areas within the districts of Batticaloa, Trincomalee and Ampara as follows: 1. Batticaloa District: Valachchenei, Vinayagapuram, Iruthayapuram, Dutch Bar, Eravur, Jayanthipuram, Thannamunei, Kallady, Koolawady, Karopakany, Thiyavathuwan, Mamunagam, Nochchiyamun, Kalkuda 2. Trincomalee District: Uppuweli, China Bay, Niyathapurei, Oris Hill, Trincomalee Town, Palauthu, Lingenagar, Selvanayapuram 3. Ampara District: Akkaraeipaththu, Karathivu, Manalchenei, Nappitavenmuni

Kurunchiadi,

Kalmunei,

Panthirippu,

SLPC speakers are ethnically diverse. Africans were brought to Sri Lanka by the colonial powers and as SLPC was the lingua franca it is not surprising that they learnt creole (de Silva Jayasuriya 2010b). African migration to Sri Lanka is multi-layered and the various waves of migrants were heterogeneous, making it difficult to maintain African languages. In the circumstances, SLPC became the mother-tongue of Africans born in Sri Lanka but as the creole has been replaced by English, and the day-today linguistic demands for this small minority comes now from Sinhala and Tamil, they have shifted away from the creole (de Silva Jayasuriya 2011). In Puttalama (northwestern province), SLPC is now spoken only by the elderly Afro-Sri Lankans. There are Portuguese-speakers, not only in Trincomalee and Batticaloa, but also in other parts of the island particularly in the northwestern province (Puttalama district) mentioned above and in the western province (Wattala and Gampaha). There may be vestiges of the creole in the central province (Wahakotte), southern province (Galle) and Northern Province (Jaffna). These are areas for future field work and research. Although the Portuguese-speaking Burghers are conspicuous in

Survival Against All Odds

211

Batticaloa and Trincomalee, some have moved out of their communities, employment and marriage being the main reasons, and a few isolated Portuguese-speakers live in various parts of the island.

2. 1 Word Order European Portuguese or Standard Portuguese is a prepositional, verbmedial language. The two major substratum languages that influenced SLPC were Sinhala and Tamil, both of which are postpositional verb final languages. Therefore, it is not surprising that given the centuries of language contact, SLPC should now be mainly a verb final language. But interestingly, SLPC still shows variation between SVO and SOV word order. 1. (Puttalama – de Silva Jayasuriya 2005) Yo I S

nasci born V

Eu nasci I born S V ‘I was born here’

aki here O

(SLPC)

aqui here O

(SP)

2. (Puttalama – de Silva Jayasuriya 2005) Yo S

arus O

Eu come S V ‘I eat rice’

kume V

(SLPC)

arroz O

(SP)

3. (Gampaha – de Silva Jayasuriya 2008) Para For

mi me S

dos two

Eu tenho duas I have two S V ‘I have two children’

crianças tem children have O V crianças children O

(SLPC)

(SP)

Chapter Eleven

212

4. (Batticaloa – de Silva Jayasuriya 2011) Eu I S

cum with

Dorothy ja Dorothy TNS O

Eu fui com I went with S V ‘I went with Dorothy’

a DEF ART

anda go V

(SLPC)

Dorothy (SP) Dorothy O

5. (Batticaloa – de Silva Jayasuriya 2011) Etus They S

ja TNS

vi come

aki here V

oito eight O

Eles virão aqui as oito e meia da They came here at eight and half of S V O ‘They came here at half past eight’

ores hours

mevu half

noite night

(SLPC)

(SP)

2. 2 Noun-adjective Order Noun-Adjective order shows variation from being the typical form where the adjective precedes the noun: 6. (Puttalama – de Silva Jayasuriya 2005) Yo I

badi brother N

Eu I ‘I have one brother’

unha one ADJ

te have

tenho have

(SLPC)

um one ADJ

irmão brother N

(SP)

Survival Against All Odds

213

7. (Gampaha – de Silva Jayasuriya 2008) Para mi quatre lingueras te pappiya, Sinhala, Tamil, English, Portuguese (SLPC) (As) for me four languages ASP speak, Sinhala, Tamil, English, Portuguese

ADJ

N

Eu falo quarto linguas, Cingala, Tamil, Inglês e Português (SP) I speak four languages, Sinhala, Tamil, English and Portuguese ADJ N ‘I speak four languages, Sinhala, Tamil, English and Portuguese’ 8. Or the form where the adjective followed the noun: Vossa Your

die day N

O vosso Your

grande great ADJ grande great ADJ

(SLPC)

dia day N

(SP)

‘Your great day’

2. 3 Noun-genitive Order In SLPC, the genitive or possessive, which normally denotes ownership, could have been influenced by several languages – English, Sinhala or Tamil. 9. (Puttalama – de Silva Jayasuriya 2005) Minha My

O The

mamma - s mother’s GEN

nome name N

nome da minha name of my N GEN ‘My mother’s name is Miseliya’.

Miseliya Miseliya

mae mother

(SLPC)

é is

Miseliya Miseliya

(SP)

Chapter Eleven

214

10. (Gampaha – de Silva Jayasuriya 2008) Minha Mama-s GEN Grant Outschoon

nome Ann Mary Mace Outschoon Pai-s nome N GEN N (SLPC)

O nome da minha mama é Ann Mary Mace Outschoon e o nome do meu N GEN N GEN pai é Grant Outschoon (SP) ‘My mother’s name is Ann Mary Mace Outschoon and my father’s name is Grant Outschoon’. 11. (Gampaha – de Silva Jayasuriya 2008) Minha fiya-s nome Augustina Jacqueline Dias, fiyo-s nome GEN N GEN N Julian Percival Dias

(SLPC)

O nome de minha filha é Augustina Jacqueline Dias e o nome de meu N GEN N GEN filho é Julian Percival Dias (SP) ‘My daughter’s name is Augustina Jacqueline Dias and my son’s name is Julian Percival Dias’

2. 4 Adposition Order 12. (Batticaloa Cantigas 1976) Bunitu lumara branku janala diante POST

(SLPC)

Bonito luar branco diante da janela (SP) PREP ‘Beautiful white moonlight in front of the window’

For several centuries, creole-speakers have been in touch with Sinhala (an Indic language) and Tamil (a Dravidian language), both verb final languages. Hence the shift is not surprising.

2. 5 Tense-mood-aspect Markers In the creole, there is a loss of inflection and verbs are reduced to a single form. Free morphemes – ja, te, lo – indicate the past, present and future. Ja and lo are adverbials (ja – already; lo(go) – soon). The etymon

Survival Against All Odds

215

of te is uncertain as there are two possibilities - ter (‘to have’) or the present progressive estar (‘to be’). For example: 13. (Batticaloa 1976) Te canta o’galu ja cunsa nov dia Canta o galo começou novo dia ‘The cockerel sings, a new day began’

(SLPC) (SP)

14. (Batticaloa 1976) Ano par ano nos lo festiamos Ano após ano nós festejamos ‘Year by year we will feast’

(SLPC) (SP)

15. (Batticaloa de Silva Jayasuriya 2011) Eu I

cum with

Dorothy ja Dorothy TNS

Eu fui com I went with ‘I went with Dorothy’

anda go

a DEF ART

(SLPC)

Dorothy (SP) Dorothy

2. 6 Gender Agreement and Number Agreement Gender and number agreement between the possessive pronouns and the nouns does not exist as in Standard Portuguese. The possessive pronoun has been reduced to the singular feminine form.

2. 7 Possessive Pronouns The possessive has been reduced to the feminine singular form – minha (my), vossa (your), nossa (our) or sua (her) in SLPC. 16. (Batticaloa 1976) Vi mingha amor par baila Vem bailar comigo meu amor ‘Come to dance with me my love’

(SLPC) (SP)

Chapter Eleven

216 17. (Batticaloa 1976) Vossa corasam O vosso coração ‘your heart’

(SLPC) (SP)

18. (Batticaloa 1976) Abri vossa porte nona dantru para vi Abri a vossa porta, senhora, para ver por dentro ‘Open your door lady to come inside’

(SLPC) (SP)

19. (Batticaloa 1976) Nossa Santo poderoso Nosso Santo poderoso ‘Our powerful Saint’

(SLPC) (SP)

20. (Batticaloa 1976) Folgadu tem nossa tare Polca há na nossa terra ‘We have polka in our land’

(SLPC) (SP)

21. (Batticaloa 1976) Lancers Kafferigna nossa nonas lo baila Lancers e cafrinhs bailarão as nossas donas ‘Our ladies will dance Lancers and Kaffrinha’

(SLPC) (SP)

22. (Batticaloa 1976) Evuya thoma praserthama nona sua casa (SLPC) Eu tive o prazer de tocar na casa da senhora (SP) ‘I had the pleasure of playing in the lady’s house’

2. 8 Plurals Loss of Inflection is observed in the plurals also. 23. (Batticaloa 1976) Sathi padera Sete pedras ‘Seven stones’

(SLPC) (SP)

Survival Against All Odds

217

24. (Batticaloa 1976) Mingha sappatu Os meus sapatos ‘My shoes’

(SLPC) (SP)

25. (Batticaloa 1976) Fuula de rosa foya tem verdi Rosa flor tem folhas verdes ‘Rose flower has green leaves’

(SLPC) (SP)

The final ‘s’ indicating plurality is some times observed but with a loss of vowel indicated by apostrophe: 26. (Batticaloa 1976) Flor’s em botao Flores em botão ‘Flower buds’

(SLPC) (SP)

2. 9 Prepositions The preposition par also functions as an infinitive marker, an object marker and preposition. 27. (Batticaloa 1976) Jada par beva par me vignu de Natal Dá-me para beber o vinho de Natal ‘Gave me to drink Christmas wine’

(SLPC) (SP)

28. (Batticaloa 1976) Rogai por nos sempre Roga sempre por nos ‘Ask for us always’

(SLPC) (SP)

2. 10 Interrogative Pronouns The interrogative pronoun, como in Portuguese is kilai (ki lai – ‘what way’) in creole.

Chapter Eleven

218 29. (Batticaloa 1976) Kilai evu lo iskise Como posso esquecer ‘How will I forget?’

(SLPC) (SP)

30. (Batticaloa 1976) Kiyora preham A que horas oram ‘At what time is the prayer’

(SLPC) (SP)

2. 11 Demonstrative Pronouns Standard Portuguese ‘e’ has a value of ‘i’ in creole. 31. (Batticaloa 1976) Tem iste dia nossa alagria Este dia é a nossa alegria ‘This day is our happiness’

(SLPC) (SP)

2. 12 Indefinite Pronoun Standard Portuguese ‘o’ is pronounced ‘u’. 32. (Batticaloa 1976) Tudus ki faye parmi, corsam te duve Todos me dizem, que o meu curação está doer ‘Everyone tells me that my heart is hurting’

(SLPC) (SP)

33. (Batticaloa 1976) Toda a amargura Toda a amagura ‘All the bitterness’

(SLPC) (SP)

2. 13 Negatives 34. (Batticaloa 1976) Evu nanda large basu yesta pega ratu Eu nunca te deixei pegar num rato ‘I never left you to catch a rat’

(SLPC) (SP)

Survival Against All Odds

219

2. 14 Quantifiers Muito is only found in the masculine form and seems to bbe an exception to the general pattern. 35. (Batticaloa 1976) Ista die tem muito alegria Neste dia de muita alegria ‘This day has much happines’

(SLPC) (SP)

2. 15 Orthography In the 1970s, the Minutes of the Annual General Meetings of the Catholic Burgher Union in Batticaloa were recorded in SLPC. This indicates that creole was written down, at least by some Burghers, and not only spoken and in lyrics as today. The orthography of the Cantigas is interesting for several reasons. There has been an attempt to maintain Portuguese orthography though the accents, nasalisation and cedilha have been omitted possibly due to the lack of technology. But there are variations from the Standard Portuguese orthography which can be attributed to the multilinguality scribes. For example: Consonants k, y, w which are not in the Portuguese alphabet are used in SLPC. 36. (Batticaloa 1976) Kavesa Foya Wide

(SLPC): cabeça (SLPC): folha (SLPC): vida

(SP) (SP) (SP):

‘head’ ‘leaf’ ‘life’

An apostrophe indicates the loss of a vowel denoting elision of the interior unstressed syllable. 37. (Batticaloa 1976) Flor’s (SLPC): flores ‘flowers’

(SP):

Final unstressed vowels in Standard Portuguese – a, e, o – have the values of e, i. u. In SLPC, the orthography is compatible with the pronunciation For example:

Chapter Eleven

220 38. (Batticaloa 1976) Barcus Ratu

(SLPC): barcos (SLPC): rato

(SP): (SP)

‘boats’ ‘rat’

(SP): (SP):

‘This’ ‘day’

39. (Batticaloa 1976) (SLPC): Este (SLPC): dia

Iste Die

Nasalised vowels represent Portuguese diphthongs: 40. (Batticaloa 1976) Limam Savan Pavam

(SLPC): limão: (SLPC): sabão: (SLPC): Pavão

(SP) (SP) (SP)

‘lemon’ ‘soap’ ‘peacock’

In SLPC, ‘k’ instead of ‘qu’ and ‘v’ instead of ‘b’ are used: 41. (Batticaloa 1976) Kevrado (SLPC): quebrado

(SP):

‘broken’

In SLPC ‘s’ is written instead of ‘ç’: 42. (Batticaloa 1976) Lansu (SLPC): lenço (SP): Descansa (SLPC): descança (SP):

‘handkerchief’ ‘tired’

I have also included some naturalistic data that I recorded in Batticaloa Town in 2011 – a conversation between Mr Stanislaus Ockersz (President of the Catholic Burgher Union of Batticaloa, now 80 years old and his 73 year old wife, Mrs Irene Ockersz (née Rosairo). They are multilingual and also speak English, Tamil and Sinhala. Mrs Ockersz was educated to GCE (General Certificate of Education) Ordinary Level and Mr Ockersz to GCE Advanced Level. He also passed the Sri Lanka Administrative Service examinations, a requirement for those entering the Service. Though they speak SLPC, nowadays their home language is English. Their parents spoke SLPC at home. They are Roman Catholics. Mass is held in English on Sundays only and in Tamil, every other day.

Survival Against All Odds

221

43. (Batticaloa – de Silva Jayasuriya 2011) Onte Yesterday

griya church

pa INF MKR

Ontem eu fui Yesterday I went ‘Yesterday I went to church’

a to the

anda go

(SLPC)

igreja church

(SP)

44. (Batticaloa – de Silva Jayasuriya 2011) Eu cum Dorothy I with Dorothy Eu fui com I went with ‘I went with Dorothy’

ja anda TNS go a DEF ART

(SLPC) Dorothy Dorothy

(SP)

45. (Batticaloa – de Silva Jayasuriya 2011) Oi Today

ki what

kere want

kusina? cook

(SLPC)

O que queres que cozinhe, hoje? What you want that cook today ‘What do you want to cook today?’ 46. (Batticaloa – de Silva Jayasuriya 2011)

(SP)

Oi Today

kusina cook

bom good

dia day

vide galina lo because chicken MOD

(SLPC)

Porque o dia esta bom, hoje, eu cozinharei galinha (SP) Because the day is good today I will cook chicken ‘Because today is a good day, I will cook chicken’. 47. (Batticaloa – de Silva Jayasuriya 2011) Minha dos irmam, grand irmam cum pikinin irmam te vi kumera (SLPC) My two sisters big sister with little sister ASP come food As minhas duas irmas, a irma mais velha e a mais nova, elas virao My two sisters sister older and younger they are coming comer (SP) to eat ‘My two sisters, elder sister and little sister are coming for a meal’.

Chapter Eleven

222

48. (Batticaloa – de Silva Jayasuriya 2011) De kusina galina cum verduru, cambram pesci par friya To cook chicken with vegetables prawns fish for frying

(SLPC)

Eu vou cozinhar galinha com vegetais, camarões, e peixe frito (SP) I am going to cook chicken with vegetables prawns and fish fried ‘I am cooking chicken with vegetables, prawns, and frying fish’. 49. (Batticaloa – de Silva Jayasuriya 2011) Nos pe cinquante gentes lo vi vinte cinqu machis, Us DAT fifty people will come twenty five boys,

(SLPC)

A nossa casa virão cinquenta pessoas, vinte cinco rapazes, (SP) Our house will come fifty people twenty five boys, 50. (Batticaloa – de Silva Jayasuriya 2011) Kinsi famias cum das crianças nossa gerasamas cum savinhos (SLPC) Fifteen women with ten children our relatives with known people Quinze mulhers e dez crianças, os nossos familias e pessoas conheçidas (SP) Fifteen women with ten children our relatives with people known ‘To our house, will come twenty five boys, fifteen women and ten children, our relatives with known people’. 51. (Batticaloa – de Silva Jayasuriya 2011) Fora tare em Sony su irma crianças ja vi te Outside land in Sony GEN sister children TNS come PF

(SLPC)

As crianças da irma da Sony vieram for a da area Children of sister of Sony have come outside of the area ‘Sony’s sister’s children from outside the area have come’.

(SP)

52. (Batticaloa – de Silva Jayasuriya 2011) Etus They

lo MOD

vi come

aki here

Eles virão aqui as oito e They will come here eight and ‘They came here at half past eight’

oito eight

ores hours

meia da noite half at night

mevu half

(SLPC)

(SP)

Survival Against All Odds

223

53. (Batticaloa – de Silva Jayasuriya 2011) Daz ani primer olya per avere me ja vi te Ten year before saw to now only TNS come PF

(SLPC)

Nós vimo-los ha dez anos atras We saw them PF ten years ago ‘We had seen them only ten years ago’.

(SP)

54. (Batticaloa – de Silva Jayasuriya 2011) Muite alegre nos pe etis pe olya ores Very happy we OBJ MKR they INF MKR see times

(SLPC)

Quando os virmos nos ficaremos muito felizes When they saw us we were very happy ‘When we saw them, we were very happy’

(SP)

55. (Batticaloa – de Silva Jayasuriya 2011) Etus pe nos tudu luara mostra primer kum avere tem defuransa They DAT our all place show earlier with now has difference

(SLPC)

Nós mostramos-lhe de tudo, a diferença entre o anteriormente e o We showed them all the difference between the formerly and agora (SP) now ‘We showed them everywhere, the difference between then and now’ 56. (Batticaloa – de Silva Jayasuriya 2011) Pikining crianses oras nos macha me iskola ka vi tone (SLPC) Little children times we walk only school PF go later Quando nos eramos crianças andavamos para ir a escola, mais tarde When we were children we walked to school but later iremos em (SP) we went in 57. (Batticaloa – de Silva Jayasuriya 2011) Karata kum Buggy cart with

cars cars

(SLPC)

Carroça com carros (SP) Cart and cars ‘When we were children, we walked to school, but later we went by cart and cars’.

Chapter Eleven

224

58. (Batticaloa – de Silva Jayasuriya 2011) Teng katri crianses, dos famias, dos machas I have four children two girls two boys Eu tenho quarto filhos, duas meninas dois rapazes I have four children, two girls, two boys ‘I have four children, two girls and two boys’.

(SLPC)

(SP)

59. (Batticaloa – de Silva Jayasuriya 2011) Tudus All

ya TNS

ka PF

casa married

(SLPC)

Todos ficam casados All have married ‘All have married’

(SP)

60. (Batticaloa – de Silva Jayasuriya 2011) Ung One

famia girl

me only

Somente uma menina Only one girl ‘Only one girl is here’

aki here ça here

(SLPC)

esta is

(SP)

61. (Batticaloa – de Silva Jayasuriya 2011) Otros tudus servi vide casa ja mara te ala Colombo tu (SLPC) Others all work because house TNS built PERF there Colombo LOC Todos os outros trabalham em Colombo onde eles construiram as casas (SP) All the other work in Colombo where they constructed the houses ‘All others work in Colombo and have built houses there’

SLPC illustrates contact-induced language change. Even though it was typologically similar to European Portuguese by the 19C, SLPC was tending towards being a verb-final language. If this could be attributed to contact-induced language change through bilingualism or multilingualism among creole-speakers, the question then arises as to when the variation began. To answer this question, we need to find data from the centuries prior to the nineteenth. Until then, we cannot pinpoint the time when the process of change started.

Survival Against All Odds

225

Peter Bakker (2006, p. 156) states that the Sri Lankan Sprachbund is unusual in that it involved only pairs of languages, rather than multilingual situations. I am not sure we could exclude the scenario where creolespeakers were multilingual, though their proficiency in the languages would vary depending on where they lived and the languages spoken by the other Sri Lankan communities with whom they came into contact regularly. Some scholars consider the term Sprachbund difficult to define but Campbell (2006, p. 2) points out that this term is more or less synonymous with linguistic area, diffusion area or convergence area. Ian Smith (1978) discusses the problems of defining and recognising processes associated with convergence and the difficulty of looking at the changes and inferring whether they occurred as part of pidginisation, creolisation or realignment. The Eastern Province has been affected by the ethnic conflicts that lasted for almost thirty years and accurate population statistics have not been obtained. The Burgher communities themselves have started to collect statistics and the numbers supplied are their estimates. It is not known yet how many Burghers actually speak the creole, nor the level of proficiency. In Trincomalee, Mrs Stefani Dominicus (née Barnes), Secretary of the Burgher Foundation established in 1948, told me that there are about 300 Burgher families in the Trincomalee district now. Many have moved from Trincomalee town to a village called Palayuttu where the largest number of families (158) is now based. There are 48 families in Selvanayapuram, 37 in Trincomalee Town, 17 in Sambaltivu, 7 in Oris Hill and 5 in China Bay. Mrs Dominicus speaks Creole with her father, Felix Barnes. A few families such as the Lappens, for example, in Palayuttu still speak SLPC at home. Boris Lappen, the second child in the family, now a 17 year old, studying for his GCE Advanced Levels, was able to converse in Creole with Mr Emmanuel Paul (69 year old) and Mrs Linda Paul (née Rosairo) (58 year old) in Palayuttu. Mrs Paul pointed out that several families in Palayuttu were able to speak Creole. Mr Paul had worked for the Royal Navy in Trincomalee and Mrs Paul was originally from Batticaloa but had moved to Trincomalee. Boris’s elder brother, Gerard, was offered a place to read for a medical degree at the University of Peradeniya, a few miles from Kandy. Gerard is the first student from Palayuttu to enter this prestigious university. Although education is free up to the tertiary level in Sri Lanka, places at university are limited and medical courses are particularly competitive. Gerard studied in Tamil and there had not been an option for him to study in English, but from next year, students will be able to choose whether they wish to follow lessons in Tamil or English. A

226

Chapter Eleven

Portuguese volunteer who had arrived from Portugal was teaching European/Standard Portuguese for a few months after which the students, including Boris, would sit an exam and be awarded a certificate of proficiency. Boris was confused between SLPC and Standard Portuguese. Given that teaching materials and learning resources are available for Standard Portuguese, decreolisation would become a reality if the Burghers learnt Standard Portuguese. The volunteer was preparing a word list in three languages – SLPC, Standard Portuguese and English. Nevertheless, Bros’s linguistic repertoire has expanded now and he speaks four languages – Indo-Portuguese of Ceylon, Portuugese, English and Tamil. There is a core group of young Burghers who are working to uphold their ancestral traditions. Burgher marriage ceremonies are not complete without the Kaffrinha, danced by four couples with the bride and groom opening the dance, followed by the bridesmaid and bestman, who are then followed by two other couples. On my recent visit to the Batticaloa Burgher Union, as is typical, I was greeted with lively music and singing in creole which was typically played for weddings. Traditional food broeder (a cake associated with the Dutch) and fafferoti (a sweet made from rice flour to which some times raisings and cadjunuts are added) was served. Intergenerational transfer of their cultural values is strong. Given that many of them have stuck to the trades of their ancestors and are artisans - carpenters, masons - and are not professionals, their chances of migrating out of their home towns for employment are low. Though they have admirable skills as master carpenters, for example, their marketability in the labour market is limited. Of the Batticaloa Burgher community, about 90% are carpenters and mechanics whose earning capacity does not enable them to make any savings. A few are businessmen (motor cycle agents), school teachers (teaching English in Tamil schools) and government servants (clerks, for example). A member of the Batticaloa community is a lecturer at the Southwestern University (Akaraipattu). He obtained a Doctorate in Mathematics from a university in South Africa, and his wife has a Masters degree from the Eastern University (Batticaloa). The community has an Accountant, a few Engineers, a lady Lawyer, and a lady Dentist. The professionals are young (less than 35 years old) and have migrated from Batticaloa to advance their career prospects. Many Burghers lived in Dutch Bar, very close to the beach. On 26th December 2004, when a Tsunami washed over the shores of Sri Lanka, 290 Burgher families became homeless; 120 were lost to the Indian Ocean (Leonard 2005). According to Sonny Ockersz, there are about 2,000 Burgher families in the Batticaloa district now.

Survival Against All Odds

227

The Roman Catholic religion has been a binding force for the Portuguese Burghers and they have maintained their identity over the centuries. There has been out-marriage with other ethnic groups. Not surprisingly, the Portuguese Burghers do not have Portuguese surnames only. There has been intermarriage with other Europeans, though surnames or family names do not always indicate ancestry. Whilst some surnames of the Burghers such as Andrado, Dias, Rosairo, signal a Portuguese connection, others such as Outschoon and Ockersz remind us of this community’s contact with other Europeans. Ethnicity, language and identity are discrete though interconnected. The Burghers are a cohesive group and music, song and dance play an important part in binding them. Also, being Roman Catholics, the church and associated festivities play a significant part in maintaining group cohesion. Endogamy, and following the trades of their forebears, not only limits their socio-economic status but also affects their social mobility. There are a sufficient number of Burghers to create a community and maintain the spoken language. But as the younger generations are educated in the local non-fee paying schools and where the medium of instruction caters mainly for the major ethnic groups in the area, the tension to maintain the Creole will be exacerbated. Introduction of the English medium as an option would enhance the marketability of the Burghers who opted to be educated in English, particularly in the international labour market. The Burghers are clear about the importance and also the limitations of Sinhala and Tamil. They acknowledged that English is important in order to be internationally focused in a globalised world. For the Burghers, pressure for survival, gaining employment and economic enhancement, tipped the balance in favour of learning other languages. This has increased competency in second and third languages but puts SLPC at further risk. The language itself has been handed down orally and this may reflect the level of literacy or education amongst the community. But as this changes, and if resources are made available through the Burgher Union (Batticaloa) and Burgher Foundation (Trincomalee), creole would be empowered. Newsletters, church music and pamphlets, for example, could be produced to held the community in reading and writing a creole. Since the mid-17th century, SLPC has developed without contact with European Portuguese because the Portuguese officials were allowed to leave when they were ousted by the Dutch (de Silva Jayasuriya 2008a). There has been no opportunity for decreolisation or the loss of creole features when a creole language comes into contact with the lexifier language. SLPC still contains creole features as illustrated but it is now an

228

Chapter Eleven

endangered language. However, SLPC is deeply embedded in the songs of the Burghers which ring out their Lusitanian identity. Many Burghers desire to keep the language alive and uphold it as an important part of their Portuguese ancestry. The community are attempting to preserve SLPC by teaching the younger generations, the Creole, but it remains to be seen if these efforts will be strong enough not to snap under the weight of the pressures of learning the local languages needed at school and in the street. Even when there are out-marriages, Burgher grandparents speak to their grandchildren in Portuguese, in order to pass on their language to the younger generation. This is in contrast to the Afro-Sri Lankans to whom the Portuguese language is not a key factor in identity. In Sirambiyadiya, creole speakers are shifting towards Sinhala, the language that they need on a day-to-day basis and the language which the younger generations learn at school. But even in Sirambiyadiya, as the Afro-Sri Lankans are assimilating, their culture, albeit creolised, is preserved through music (de Silva Jayasuriya 2008b). In Sri Lanka, nobody is forced to assimilate and it has remained pluricultural but as communities get smaller and outmarriages increase, there is a tension between holding on to one’s cultural elements and allowing them to change. However, as a market has been created for creolised Portuguese songs by the Sirambiyadiya Afro-Sri Lankans, there seem to be knock-on effects for the Burghers. The BBC World Routes programme met these communities and recorded a few of their songs in September 2011. As media interest and coverage internationally increases, on both ethnic groups identified with the Portuguese, the communities would have a further incentive to hold on to their centuries-old traditions which, in turn, would lead to the preservation of creole.

3. Concluding remarks There is an unavoidable tension between integration and in holding on to one’s ancestral language which impinge on identity. Bilingualism or multilingualism is inevitable and as long as the Burghers uphold their Portuguese identity, creole will be alive, not necessarily in the spoken form but more embedded in lyrics of Indo-Portuguese songs. As creole becomes more endangered, Portuguese identity would be a construct of new elements – music, song and dance. Through the inseparable artistic traditions as stated in the Indian texts – nacca (dance), geeta (song) and vƗdita (music) which form a troika, a Portuguese identity resurfaces and re-establishes itself.

Survival Against All Odds

229

Notes 1

My thanks go to Mr and Mrs Sonny Ockersz (Batticaloa), Boris and Gerard Leppen, Mrs Stefani Dominicus, Mr Felix Barnes and Mr and Mrs Emmanual Paul.

References and Further Reading Bakker, P. 2006. The Sri Lanka Sprachbund: The Newcomers Portuguese and Malay. In: Y. Matras, A. McMahon and N. Vincent (eds.), Linguistics Areas: Convergence in Historical and Typological Perspectives Basingstoke: Palgrave MacMillan. Bandara, A. 2010. Behavioural Changes in Sri Lanka Portuguese Creole. A paper presented to the Royal Asiatic Society Colombo, Sri Lanka. Campbell, L. 2006. Areal Linguistics: A Closer Scrutiny. In: Y. Matras, A. McMahon and N. Vincent (eds.), Linguistics Areas: Convergence in Historical and Typological Perspectives Basingstoke: Palgrave MacMillan. Dalgado, S. R. 1900. Dialecto Indo-Português de Ceylão. Lisboa: Imprensa Nacional. De Silva Jayasuriya, S. 1999. 'On the Indo-Portuguese of Ceylon': A Translation of a Hugo Schuchardt Manuscript. Portuguese Studies, 15, pp. 52-69. —. 2008a. The Portuguese in the East: A Cultural History of a Maritime Trading Empire. London: I B Tauris Academic Publishers. —. 2008b. African Identity in Asia: Cultural Effects of Forced Migration. New Jersey: Markus Wiener. —. 2010a. Persisting Portuguese Linguistic Impressions in India and Sri Lanka. In: Portuguese in the Orient:The Portuguese in Sri Lanka and India. Kandy, Sri Lanka: International Centre for Ethnic Studies. De Silva Jayasuriya, S. 2010b. African diaspora in Asian Trade Routes and Cultural Memories. UK: Edwin Mellen Press. —. 2011. Language Maintenance and Loss among Afro-Asians in South Asia. In: K. Ihemere (ed.), Language Contact and Language Shift: Grammatical and Sociolinguistic Perspectives. Munich: LINCOM Studies in Language Typology 17. Leonard, E. 2005. As related by M Rozario. Caught up in the tsunami: Forgotten Batticaloa Burghers. http://www.sundayobserver.lk/2005/01/23/fea17.html Smith, I. R. 1978. Realignment and Other Convergence Phenomena. University of Melbourne Working Papers in Linguistics 4, pp. 67-76.

Chapter Eleven

230

Abbreviations ADJ ART ASP COMP CONJ DAT DEF ART GEN INF MKR LOC MOD N O OBJ MKR PF POST PREP S SLPC SP TNS V

adjective article aspect completive conjunction dative definite article genitive infitive market locative modality noun object object marker perfective postposition preposition subject Sri Lanka Portuguese Creole Standard Portuguese tense verb

Exercises 1. 2. 3. 4.

Is language more resistant to change than other cultural factors? What factors impinge on the survival of language? Should creole speakers also learn the base language? Should education be available in the mother-tongue?

PART III: LANGUAGE ATTITUDES AND IDENTITY

EDITOR’S INTRODUCTION TO PART THREE

Appel and Muysken (1987, p. 16) correctly observe, the fact that languages are not only objective, socially neutral instruments for conveying meaning, but are linked up with the identities of social or ethnic groups has consequences for the social evaluation of, and the attitudes towards languages. Put simply, language attitudes are the feelings people have about their own language variety or the languages/ language varieties of others. Language can play a very important role in how we are identified. It fits with other indicators of social identity and group membership such as style of clothes, type of haircut and taste in music. Identity, whether it is an individual, social or institutional level, is something which we are constantly building and negotiating all our lives through our interaction with others. For instance, Ronþeviü, Blaževiü, and Šepiü in their paper investigate multilingualism and national identity in mixed-marriage families. More precisely, they attempt to confirm the claim that children in mixed marriages have shared identity. According to Bourdieu (1990), language is a constituent of all forms of social interaction and it influences the shaping of one's identity. For Barth (1969), language is not only a means of communication, but also a symbol of identity within society, a mark of belonging. The attitudes towards a language reflect the attitudes towards its users. The general notion is that families represent the foundation of society, and they play a crucial role in shaping and preserving one's identity. There are no shifts in regularity patterns in cases where family members belong to the same language community and share the same national identity – the patterns of belonging can be logically outlined. However, the case is quite different in mixed-marriage families where partners belong to different language communities (i.e. have different national identities). Thus, the authors assess and determine to which extent the usage of mother's minority language and the perception of domination by one parent influence the shaping of one's identity, whether it becomes a clear identity or a shared one. For Hawes and Mirvahedi in their paper, they examine a possible link between the process of language attrition and Azeris’ increasingly common codeswitching to Farsi, employing Farsi structures within Azeri sentences. Code-switching in Tabriz may be an attempt to be fair towards the various

Editor’s Introduction to Part Three

233

language communities or perhaps to show ‘loyalty’ to the central government. It is particularly interesting as performed by high profile individuals, in that it does not necessarily highlight one’s own individual identity (e.g. Myers-Scotton 2006), nor suggest a convergence reaction to a change in audience (e.g. Bell, 2010). Rather, it seems to assert one’s membership of mainstream society and that, by implication, the identity of one’s audience ought to mirror this. The final paper in this section by Mann uses the “voices” of questionnaire respondents to reflect attitudes toward Anglo-Nigeria Pidgin (ANP) in twelve urban centres in Nigeria in a context of non-formal, nonstandard, socially-marked varieties of language. The paper also draws on the data and observations of Mann (1993b; 2000b; 2001; 2004; 2011), Ihemere (2006), Igboanusi (2008), Akande and Salami (2010), and others, for its discussions. Attitudes toward ANP are re-analyzed from the perspective that, in spite of several negative and apparent formal weaknesses attributed to this language, but on account of its everincreasing sociolinguistic vitality, it may have become, notwithstanding, more recently perceived as a threat, not only to ethnic minority languages in urban centres, but also to English – the prestigious official language of politico-administration, education and the mass media in Nigeria.

References and Further Reading Akande, A. T. and Salami, L. O. 2010. Use and attitudes towards Nigerian Pidgin English among Nigerian University students. In: Millar, R.M. (ed.), Marginal Dialects: Scotland, Ireland and Beyond, 70-89. Aberdeen: Forum for Research on the Languages of Scotland and Ireland. Barth, F. 1969. Ethnic Groups and Boundaries: The Social Organization of Culture Difference. Oslo: Universitetsvorlageet. Bell, A. 2010. Back in Style: Reworking audience design. In: M. Meyerhoff and E. Schleef (eds.), The Routledge Sociolinguistics Reader. Abingdon: Routledge. pp. 32-52. Bourdieu, P. 1990. The Logic of Practice. Cambridge: Polity. Igboanusi, H. 2008. Empowering Nigerian Pidgin: A challenge for status planning? World Englishes, Vol. 27, No. 1, pp. 68-82. Ihemere, K. U. 2006. An integrated approach to the study of language attitudes and change in Nigeria: The case of the Ikwerre of Port Harcourt City. In: Arasanyin, O. F and M. A. Pemberton (eds), Selected Proceedings of the 36th Annual Conference on African

234

Editor’s Introduction to Part Three

Linguistics, Somerville, MA, Cascadilla Proceedings Project, 2006, pp. 194-207. Mann, C. C. 1993b. The sociolinguistic status of Anglo-Nigerian Pidgin: An overview. International Journal of the Sociology of Language, 100/101, pp. 167-178. —. 2000b. Reviewing ethnolinguistic vitality: The case of Anglo-Nigerian Pidgin. Journal of Sociolinguistics, 4/3, pp. 458-74. —. 2001. Anglo-Nigerian Pidgin: A Sociopsychological Survey of Urban, Southern Nigeria. Unpublished Doctoral thesis, Edinburgh,University of Edinburgh. —. 2004. Attitudes toward Anglo-Nigerian Pidgin in urban, Northern Nigeria: A preliminary report. Paper presented at the Society for Pidgin and Creole Linguistics Meeting (11-15 August, 2004), World Trade Center, Curaçao. —. 2011. Socially-marked varieties of language: The Sociocommunicational Hypothesis and Anglo-Nigerian Pidgin. Tshwane University of Technology Professorial Inaugural Lecture (18 March, 2011), Pretoria. Myers-Scotton, C. 2006. Multiple Voices: An introduction to bilingualism. Blackwells: Malden, USA and Oxford.

CHAPTER TWELVE MULTILINGUALISM AND NATIONAL IDENTITY IN MIXED-MARRIAGE FAMILIES MARINA RONýEVIû, BRANKA BLAŽEVIû AND TATJANA ŠEPIû

1. Introduction The concept of identity refers to individuals, but it is very often explained by determining the social groups these individuals belong to – family, friends, education and other institutions, historical context, social class. All of these factors influence the individual on various levels: language, attitudes, education, belonging to a certain “place” within a society. Identity, as an integral part of an individual, can be viewed from various perspectives. Nowadays, new modern conceptions of identity and individuality give the right to express dissimilarity and diversity as to one's individual, ethnic, national, linguistic or sexual identity (Bucholz and Hall 2004). It can therefore be said that everyone has a multilayered identity and that different constituents i.e. aspects of identity will either emerge or will be highlighted under certain conditions, and/or will be hidden and repressed in others. Identity theory treats society as a ‘result of repeated, patterned behaviours of individuals. The theory examines the self and how actors attach meanings to the multiple roles they play’ (Carter 2007); researchers within this theoretical framework examine ‘how identities are embedded within social structures, and how social structures influence the notion of what it means to be an individual. Identity theory also examines how micro-level processes create and maintain different meanings actors have for themselves and others, and how these meanings perpetuate themselves to maintain social order. This perspective sees the self as emerging from social interaction and portraying it to others through the identities that are appropriate in specific situations’ (Carter 2007). The term identity is used in various ways; one can talk about personal or group identity. In fact,

236

Chapter Twelve

someone's identity is shaped by looking for similarities within the group he/she belongs to, and at the same time, however, he/she accepts and rejects certain identity traits. While personal identity highlights the differences between an individual and the group, social identity constitutes the ways individuals and groups bond by means of recognizing differences between them as a group and other individuals and groups and emphasizing similarities within their own group (Howard 2000). Ethnicity is a very important constituent of identity, and refers to a series of situations in which groups of people act and live together, i.e. it refers to a nation (Ostergard 1992a, p. 32 in Jenkins 1997, p. 19). National identity is a common list of other identities put together. On one hand the state has the most prominent role in the creation of national identity, as it is achieved through its institutions, citizenship, national loyalty and pride and other forms (Weber 1976). On the other hand, ethnic groups are founded on the belief of having a common ancestry. The realization of common interests encourages ethnic identification. Ethnic identification, moreover, arises from and through the interaction with other groups. Ethnic identity presupposes a cultural bondage to a given group which transmits its culture by means of language, religion, laws, traditions and institutions (Liebkind in Fishman 1999). The role of language in determining one's identity is extremely important as it is an instrument to express the self and the group (Liebkind in Fishman 1999); it is a means of interaction when raising a child, it demarks an ethnic group; it influences the creation of group identity, yet at the same time the very same group identity influences the attitude towards a language and its usage. In the general opinion the family plays a determinant role in shaping and preserving a person’s identity. When family members belong to the same language community, the identity problem does not exist. On the other hand, in families where partners belong to different language communities it is not easy to guess which language or ethnic identity will the children feel more as their own (Ronþeviü 2010). What the authors propose to examine in this paper is the identity of children in mixedmarriage families and the role played by the language – especially in bilinguals – in shaping their identity. This, in turn, will entail another question, whether such persons have a clear or a shared identity. More precisely, the authors will try to assess the presence of shared identity in children coming from bilingual mixed-marriage families, and the extent to which the usage of mother’s minority language and the dominant figure of one of the parents influence the clear or shared identity of a person.

Multilingualism and National Identity in Mixed-Marriage Families

237

2. Research Methodology and Sample The survey was conducted on a sample of a hundred respondents of both genders (49% male and 51% female), aged 18 to 72 and with a similar educational background. All of the respondents are Croatian citizens coming from mixed-marriage families – one of the parents a Croat and the other of foreign ancestry/origin (Italy, Slovenia, Bosnia and Herzegovina, Czech Republic, Poland, Germany, Kosovo, Serbia, and Japan). The research was carried out by means of a quantitative analysis (closed questions, self-evaluation scale) and a statistical analysis of the questionnaire: arithmetic middle, t-test for the significance of a small sample number, Ȥ2 test and a one-way analysis of variance. A closed questions questionnaire was prepared for the purpose of this survey. Along with personal details, it gathered data on parents’ origins; the sense of dominance by one of the parents during the respondent’s childhood, the usage of the minority language within and outside the family, data about recognizing/acknowledging and expressing one’s own identity. A self-evaluation scale was prepared for the purpose of this survey. The scale consists of five statements distributed on twenty-two Likert-type scales scaling 1 (not at all), 2 (a little), 3 (partially), 4 (fairly) and 5 (fully). These rating scales refer to variables that affect the respondent’s identity, his/her minority language, the importance of the minority language and/or the Croatian language, some emotions evoked by the very usage of the minority language and the frequency of minority language usage. The task of the respondents was to evaluate a Likert item, i.e. a statement, by rating its level of frequency expressed by the scale. The questionnaire was anonymous. In a general introduction a brief explanation of the purpose of the survey was given to the examinees. It must be said that the survey did encounter some difficulties. Indeed, even though it was stated out clearly that the questionnaire was anonymous; it was not easy to find respondents willing to cooperate. Respondents did not feel at ease when being asked about their nationality, as it was and still is indeed a very delicate issue. Even though no one explicitly expressed the reasons for this uneasiness and reluctance, it is possible to assume it was due to fear of negative labelling. Moreover, part of the respondents did not answer the question about having a clear or a shared identity. Questionnaires were submitted for completion individually or in groups, but respondents took their own time in filling it out. Everything was supervised by an examiner.

Chapter Twelve

238

3. Research Within the framework of this research, the authors wanted to assess the following goals: ƒ the tendency of the overall sample as to the self-evaluation of one's own identity with respect to various factors (minority language, place of birth, language of education); ƒ the presence of a statistically significant difference with respect to one's gender; ƒ the presence of a statistically significant difference in using the mother's minority language and the father's minority language in individuals who consider both parents to be dominant; ƒ the presence of a statistically significant difference in the results as to self-evaluating the minority language in individuals who consider themselves to have a clear identity and those who have a shared identity.

3. 1 Analysis of the Tendency of the Overall Sample as to the Self-evaluation of One's Own Identity In accordance with the first postulated goal the authors established that: -

82% of the respondents can clearly define their own identity; 24% of the respondents consider their nationality extremely important (50% partially important and 26% not important at all); 19% of the respondents believe they cannot freely express their identity to others; 65% of the respondents declared to have shared identity; 61% of the respondents consider their inner/intimate sense of nationality to coincide with what they have officially declared; 68% of the respondents are bilingual.

On the basis of the obtained and statistically analysed results it was possible to draw a few general conclusions with respect to various factors. The place of birth and the language of the respondent have a slightly greater influence on the respondents’ identity, whereas religion has little or almost no influence (Table 12-1).

Multilingualism and National Identity in Mixed-Marriage Families YOUR IDENTITY IS AFFECTED BY:

N

Your place of birth

99

3.141414

Place of birth of your mother

99

2.545455

100

2.630000

Minority language

99

3.101010

Language of education

98

3.448980

Official language

99

3.212121

Religion

98

1.979592

Something else

27

1.444444

Place of birth of other family members

239

M

Table 12-1: Tendency of the overall sample as to the self-evaluation of one's own identity with respect to various factors The knowledge of the minority language is more passive than active. Respondents can understand and speak the minority language better than they can read or write in it, as greater importance is given to the official language (Table 12-2). The minority language:

N

M

I understand

97

4.422680

I speak

98

4.010204

I read

98

3.989796

I write

98

3.642857

Table 12-2: Knowledge of the minority language Respondents feel at ease, satisfied and proud when they use the minority language and thus feel little or no fear, shame and uneasiness (Table 12-3).

Chapter Twelve

240

Using the minority language it evokes:

N

M

Pride

97

3.556701

Fear

97

1.319588

Shame

97

1.206186

Satisfaction

98

3.867347

Dissatisfaction

97

1.247423

Ease

97

3.711340

Uneasiness

96

1.145833

Table 12-3: Emotions evoked by using the minority language

3. 2 Analysis of the Knowledge of the Minority Language with Respect to the Gender of the Respondents In accordance with the second goal the authors assessed a statistically significant difference with respect to the gender of the respondent when asked about their ability to write in the minority language. Women, according to their self-evaluation, as can be seen in Table 12-4, seem to write better than men. This can be explained by the women’s predisposition to communicate with family members in the minority language, for example, for correspondence. The minority language:

M men

M women

t-value

Df

p

N men

N women

I understand I speak I read I write

4.382979 3.958333 3.833333 3.312500

4.460000 4.060000 4.140000 3.960000

-.35307 -.39375 -1.10620 -2.11557

95 96 96 96

.724818 .694638 .271406 .036970

47 48 48 48

50 50 50 50

Table 12-4: Knowledge of the minority language with respect to the gender of the respondents Women consider the minority language more important than men (table 5), proven by the statistically significant difference (t=3.088, p