Language and Ethnic Identity

cover title: author: publisher: isbn10 | asin: print isbn13: ebook isbn13: language: subject publication date: lcc: ddc...

Author: William B. Gudykunst (Editor)

103 downloads 1980 Views 1MB Size Report

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

Report copyright / DMCA form

DOWNLOAD PDF

cover

title: author: publisher: isbn10 | asin: print isbn13: ebook isbn13: language: subject publication date: lcc: ddc: subject:

next page >

Language and Ethnic Identity Gudykunst, William B. Multilingual Matters 1853590215 9781853590214 9780585240787 English Anthropological linguistics, Sociolinguistics. 1988 P35.L267 1988eb 401/.9 Anthropological linguistics, Sociolinguistics.

cover

next page >

< previous page

page_i

next page > Page i

Language and Ethnic Identity Edited by William B. Gudykunst MULTILINGUAL MATTERS LTD Clevedon. Philadelphia

< previous page

page_i

next page >

< previous page

page_ii

next page > Page ii

Library of Congress Cataloging in Publication Data Language and ethnic identity / edited by William B Gudykunst p. cm Bibliography: p. Includes index ISBN 1853590215 1. Anthropological linguistics. 2. Sociolinguistics I Gudykunst, William B. P35 L267 1988 8818702 401'.9dc19 CIP British Library Cataloguing in Publication Data Language and ethnic identity. 1. Ethnic groups. Identity. Role of language I Gudykunst William B. 305.8 ISBN 1-85359-021-5 Multilingual Matters Ltd Bank House, 8a Hill Road Clevedon, Avon BS21 7HH England

242 Cherry Street Philadelphia, Pa 19106-1906 & USA

Copyright © 1988 William B. Gudykunst and the authors of individual chapters. All rights reserved No part of this work may be reproduced in any form or by any means without permission in writing from the publisher Typeset by Wayside Books, Clevedon, Avon. Printed in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire

< previous page

page_ii

next page >

< previous page

next page >

page_iii

Page iii

Contents Language and Ethnic Identity: An Overview and Prologue William B. Gudykunst and Karen L. Schmidt

1

Achieving 'Unmarkedness' in Organisational Discourse: A Praxis Perspective on Ethnolinguistic Identity Stephen P. Banks

15

Social Mobility and Language Use in an American Company in Japan Patricia M. SanAntonio

35

Language Attitudes, Frames of References and Social Identity: A Scottish Dimension Dominic Abrams and Michael A. Hogg

45

Language and Social Identity: Israelis Abroad T. F. McNamara

59

Ethnicity and Immigrant Values: Religion and Language Choice Deepa Punetha, Howard Giles and Louis Young

73

Language and Intergroup Perception in Spain Maria Ros, J. Ignacio Cano and Carmen Huici

87

Imposed Identity and Linguistic Strategies Joha Louw-Potgieter and Howard Giles

105

Language Attitudes and Vitality Perceptions: Intergenerational Effects among Chinese Canadian Communities Itesh Sachdev, Richard Bourhis, Sue-wen Phang and John D'Eye

131

Social Psychological and Linguistic Impediments to the Acquisition of a Second Nigerian Language among Yoruba and Ibo C. N. Mgbo-Elue

153

Afro-American Identity Labels and Communication Effectiveness Michael L. Hecht and Sidney Ribeau

163

< previous page

page_iii

next page >

< previous page

page_1

next page > Page 1

Language and Ethnic Identity: An Overview and Prologue * William B. Gudykunst and Karen L. Schmidt Department of Communication, Arizona State University Abstract The purpose of this essay is to overview the study of language and ethnic identity. The general role of language in social categorisation and the influence of social categorisations on language attitudes, as well as the specific influence of ethnic identity on the microsociolinguistic, macrosociolinguistic, and social psychological aspects of language use, language attitudes, sociolinguistic stereotypes, ethnolinguistic vitality, and speech accommodation are reviewed. The contributions to the present volume also are outlined in the context of these areas of research. Individuals use social categories to order their social environment and reduce the complexities of the world. Language is one of the major factors used to categories others (Giles & Johnson, 1981). Language use also plays a major role in the development of social identity in general (Eastman, 1985) and ethnic identity in particular (Giles & Johnson, 1981). Social identity is 'that part of an individual's self-concept which derives from his [or her] knowledge of his [or her] membership in a social group (or groups) together with the values and emotional significance attached to that membership' (Tajfel, 1978: 63). The language or dialect speakers use provides cues that allow others to determine if speakers are members of an ingroup or an outgroup. Research indicates that ethnic group members identify more closely with those who share their language than with those who share their cultural background (Giles, Taylor & Bourhis, 1973). Individuals also evaluate speakers perceived as ingroup members more favourably than those perceived as outgroup members (Hogg, Joyce & Abrams, 1984). Language and ethnic identity are related reciprocally, i.e. language usage influences the formation of ethnic identity, but ethnic identity also influences language attitudes and language usage. The interrelationship between language and ethnic identity is studied from sociolinguistic (e.g. Gumperz, 1982a,b) and social psychology of language (e.g. Giles, 1977; Giles & Saint-Jacques, 1979) perspectives. Extensive research on various aspects of language and ethnic identity (e.g. language attitudes, second language acquisition, code-switching) has been conducted in recent years using a variety of theoretical and methodological perspec-

< previous page

page_1

next page >

< previous page

page_2

next page > Page 2

tives. The purpose of this volume is to present representative and innovative research on the topics studied under the rubric 'language and ethnic identity' conducted in several different areas of the world. The research presented uses a variety of methodological approaches, including matchedguise experiments, intensive interviews, surveys, ethnographic observations, and discourse analysis. They also examine different levels of analysis regarding the relationship between language and ethnic identity. All studies, however, draw upon the social psychology of language for their theoretical rationale. The remainder of this introductory essay is devoted to overviewing the study of language and ethnic identity and outlining the contributions to the present volume. The overview is organised around Bourhis' (1979) three categories of factors that influence code-switching in interethnic encounters: microsociolinguistic, macrosociolinguistic, and social psychological. While Bourhis used these categories to organise research on code-switching, they also are applicable to the research on language and ethnic identity more generally. It should be noted at the outset that micro- and macrosociolinguistic factors related to language and ethnic identity are studied from sociolinguistic and social psychological perspectives (conversely, the social psychological factors also are studied from a sociolinguistic perspective). The focus of the present overview and the contributions to this volume is the social psychological perspective. Microsociolinguistic Factors Microsociolinguistic factors associated with language and ethnic identity in general code-switching in particular 'include the topic, setting, and purpose of conversation as well as characteristics of the interlocutors' (Bourhis, 1979: 119). These factors, therefore, encompass Hymes' (1972) taxonomy of situational determinants of speech. 'Situational' and 'metaphorical' code-switching are the most common speech strategies influenced by microsociolinguistic factors. Situational switching occurs when the choice of the code used is guided by the rules and social norms of the situation. Metaphorical switching, in contrast, occurs when ingroup speakers intersperse their conversations with utterances from outgroup languages. There is extensive evidence that microsociolinguistic factors influence speech strategies in encounters with members of the same and other groups (see Dittmar, 1976; Giles & Powesland, 1975; Gumperz, 1982a,b; Gumperz & Hymes, 1972, for reviews). Studies of immigrants in the United States, for example, indicate that English is used in public formal settings, while native language is used in informal, nonpublic settings (e.g. Edelman, Cooper & Fishman, 1966; Kimble, Cooper & Fishman, 1969; Ryan & Carranza, 1977). Data from the Philippines (Sechrest, Flores & Arellano, 1968), Paraguay (Rubin, 1968), Israel (Herman, 1961), and Nigeria (Mgbo-Elue, this volume) support this conclusion. Banks (this volume) extends traditional discourse analytic work to examine the influence of language-in-use on changing ethnolinguistic identities. He argues that the boundary between marked and unmarked ethnic discourse (cf. Scherer &

< previous page

page_2

next page >

< previous page

page_3

next page > Page 3

Giles, 1979) is soft and permeable, while the boundary between low and high power positions is hard and less permeable. Banks presents evidence from the discourse of an ethnolinguistic minority sales manager to support four general propositions: (1) members of ethnolinguistic minority groups must cross the soft boundary from marked to unmarked ethnic discourse before crossing the boundary from low to high power position; (2) crossing the boundary from marked to unmarked ethnic discourse is a function of the individual's strategies to maximising rewards, as well as the norms, values, and discourse routines in the organisation; (3) there is an implicit promise from the organisation that if members of ethnolinguistic minorities cross the boundary from marked to unmarked ethnic discourse, they will have the opportunity to cross the low to high power boundary; and (4) individuals who cross the boundary from marked to unmarked ethnic discourse 'subtract from' their ethnolinguistic identity (cf. Giles, 1979). San Antonio (this volume); also examined language-in-use as an ethnolinguistic identity marker in her study of intergroup communication in a United States corporation in Japan. She argues that speaking Japanese is linked closely with Japanese ethnolinguistic identity. In the company which she observed, the explicit language policy, however, is that only English is to be spoken. The English ability of the Japanese employee varied from almost no ability to speak the language to almost native fluency. San Antonio's ethnographic observation indicated that the use of English by Japanese employees in a range of situations (e.g. meetings, interacting with visitors) is a marker for the claimed identity of being a Japanese with whom the Americans could work. American managers interpreted the use of English and following American communication norms (e.g. first names, joking) positively and not using English, negatively (e.g. Japanese were evasive). English directly influences Japanese employees' status and role within the organisation. The employees who are fluent in English 'protect' employees who do not speak English well in meetings by speaking-up and answering Americans' questions. Employees who do not speak English well are grateful for the help of those who speak it better. The use of English in this particular setting, therefore, does not necessarily 'subtract from' the Japanese ethnolinguistic identity. San Antonio's research clearly demonstrates that communicators' second language competence is a major microsociolinguistic factor that must be considered in the study of language and ethnic identity. There are numerous second language learning models (see Gardner, 1985, for reviews of the major models), but only the intergroup model (Giles & Byrne, 1982) deals in-depth with the effects of social categories on second language acquisition. Giles & Byrne (1982) argue that when there is weak ingroup identification, quiescent interethnic comparisons, low perceived ingroup vitality, soft/ open perceived ingroup boundaries, and strong identification with other groups, members of subordinate groups will attain proficiency in the dominant group's language. When the two 'primary' determinants of second language acquisition motivation, perceived vitality and perceived hardness of boundaries between groups, are inversely correlated with motivation an 'intermediate' situation exists (Beebe & Giles, 1984). Ball, Giles & Hewstone (1984) extend the theory to account for these situations using a cusp catastrophe. This extension allows the

< previous page

page_3

next page >

< previous page

page_4

next page > Page 4

theory to account for conflicting motives that are based on speakers' perceptions of the relations between their own and other linguistic groups. One recent study (Hall & Gudykunst, 1986) tested Giles & Byrne's (1982) original intergroup model of second language acquisition. Hall & Gudykunst's data generally support the model for international students studying in the United States. Their results, however, indicated that the stronger the ingroup identification, the greater the perceived competence in the outgroup language. While incompatible with the original statement of the intergroup theory, this finding is compatible with Lambert, Mermigis & Taylor's (1986) study which suggests that the more secure and positive members of a group feel about their identity, the more tolerant they are of members of other groups. Similar observations emerge from other studies (e.g. Bond & King, 1985; Pak, Dion & Dion, 1985). While incompatiable with Giles & Byrne's original formulation, Hall & Gudykunst's (1986) findings are compatible with Giles, Garrett & Coupland's (1987) suggested revision of the intergroup theory of second language acquisition. Hall & Gudykunst (1986) also found the influence of ingroup identification, interethnic comparisons, perceived vitality, perceived group boundaries, and identification with other groups on second language competence is not mediated by integrativeness as suggested by Clément & Kruidenier (1983) or Gardner (1985). Support for the intergroup theory also emerges from Giles & Johnson's (1987) study of Welsh speakers in Britain. Macrosociolinguistic Factors Macrosociolinguistic influences on language and ethnic identity include those sociostructural factors that influence speech strategies. Bourhis (1979) argues these include the relations between the groups involved vis-á-vis bilingualism and diglossia (see e.g. Fishman, 1967), language planning and/or legislation (see e.g. Bourhis, 1984b), as well as the individuals' group memberships and the ethnolinguistic vitality of these groups. Several of the studies presented in this volume support and/or extend Tajfel's (1978) social identity theory to language usage or draw upon Giles & Johnson's (1981) analysis of ingroupoutgroup membership and language. Abrams & Hogg's (this volume) research, for example, revealed that Tajfel's research on ingroup favouritism can be extended to evaluations of speakers' status, likely employment, and solidarity. Their study also indicated that while ingroup favouritism correlated positively with social identity, language attitudes vary considerably depending on the degree of self-categorisation (Turner, 1987) salient to the speaker. One of the major implications of this study is that degree of self-categorisation as a member of an ethnic group must be considered in future research on language and social identity. McNamara's (this volume) study of Israelis living in Australia also supports social identity theory. His research demonstrated that social identity influences language attitudes, as well as language maintenance and shift of the immigrant group. Specifically, he found that changes in social identity are accompanied by

< previous page

page_4

next page >

< previous page

page_5

next page > Page 5

changes in language attitudes favouring English over Hebrew and a shift to the use of English among the immigrants' children. (Note: McNamara, in press, also has reinterpreted numerous Australian studies of language and ethnic identity from a social identity theory perspective.) Punetha, Giles & Young's (this volume) research revealed that language choice affects values of Asian (Hindus, Muslims, and Sikhs) immigrants in England. More specifically, they found that immigrants choosing to answer a questionnaire in English (outgroup language) held values intermediate to indigenous respondents and Asian immigrants choosing to answer the questionnaire in their ingroup language. This study is the first empirical exploration of Asians in the United Kingdom regarding values and language preference. It has clear methodological implications for conducting research on language and social identity. Forcing respondents to answer questionnaires or to be interviewed in a specific language will influence their responses. Since individuals' choice of language is a function of their ethnic identity and influences their responses, care must be taken in interpreting results of studies where they are forced to use a specific language. Ros, Cano & Huici's (this volume) research also is consistent with social identity theory. Their study revealed that social identity and perceived ethnolinguistic vitality affect intergroup relations and attitudes toward the major languages (Castilian, Catalan, Basque, Valencian, and Galacian) in Spain. Subtractive identity (i.e. the difference between identification with a specific linguistic ingroup and identification with Spain), however, was a better predictor of the specific intergroup relations to emerge than was identification with ingroup alone. The subtractive identity measure used in this study should prove valuable in future studies of language and ethnic identity in multiethnic environments. Louw-Potgieter & Giles (this volume) also examined language and social categorisation, but from a different angle. They studied the linguistic strategies used to deal with imposed identities of high status group members in South Africa. Their research revealed that when there is an incongruity between members of one group's self-definition and the identity imposed on them by members of another group, they attempt to escape the imposed identity by changing the criteria for group membership and by differentiating themselves from the group with which they are associated. Both of these processes involved specific language usage strategies, but the specific strategies varied depending upon the groups and the relations between them. This study is the first extended study of language and identity in South Africa from a social psychological perspective. It also is the first to examine dissidents and the intragroup dynamics that exist in South Africa. The concept of 'group vitality' was introduced by Giles, Bourhis & Taylor (1977). They argued that ethnolinguistic vitality influences the degree to which group members will act as a group when interacting with individuals from outgroups. Two types of group vitality, objective and subjective, were distinguished in their model. Subjective group vitality refers to a group's societal position as perceived by its members, while objective group vitality refers to the group's position as indicated by available data on group membership and activities. They proposed that the more vitality an ethnolinguistic group has, the more likely the group will survive as a distinctive linguistic group in a multilingual setting.

< previous page

page_5

next page >

< previous page

page_6

next page > Page 6

Giles, Bourhis & Taylor (1977) isolated three categories of variables influencing ethnolinguistic vitality: status, demographic, and institutional support. The status variables are those which influence the prestige of a language group (e.g. economic, social, and language status). Giles and his associates contend that the greater a linguistic group's status, the more vitality it has. Demographic factors involve the group size and its distribution within regions and a country as a whole. The more favourable the demographic situation, the more vitality the group has. Finally, institutional support is related to the representation of the language use in societal institutions (e.g. churches, businesses, government agencies) and the greater the institutional support for a language group, the greater its vitality. In order to measure subjective perceptions of ethnolingusitic vitality, Bourhis, Giles & Rosenthal (1981) devised the Subjective Vitality Questionnaire (SVQ). The three dimensions of vitality are included in the questionnaire and each questionnaire item is designed to measure group members' subjective view of their own group in comparison to a salient outgroup. The SVQ appears to have high internal reliability (Labrie & Clement, 1986). Several studies have been conducted which compare different ethnic groups' perceptions. Giles, Rosenthal & Young (1985), for example, studied ethnolinguistic vitality perceptions of Greek- and Anglo-Australians and found that the two groups agreed that Anglos' vitality was higher on certain status and institutional support items, but disagreed about each other's sociostructural positions. Bourhis & Sachdev (1984) studied Italian and English Canadians in both an equal setting, and a majority setting. Their research indicated that both groups had more realistic perceptions in a majority setting than in an equal setting. English Canadians, however, were biased more against the Italian language in the equal setting than in the majority setting. In a study of ethnic Chinese and non-Chinese in Hong Kong, Young, Giles & Pierson (1986) found that ingroup vitality perceptions are associated with the amount of exposure to the outgroup language. Sachdev, Bourhis, Phang & D'Eye's (this volume) extended research on ethnolinguistic vitality to Chinese immigrants in Canada. Their study revealed that perceived vitality is consistent with objective vitality estimates of high English Canadian and low Chinese Canadian vitality on status, demographic and institutional support factors. Sachdev and his associates also discovered Canadian born Chinese exaggerated the perceived vitality of Cantonese more than first generation Chinese Canadians; Canadian born Chinese further reported that a Canadian born Chinese perceived Cantonese to be used less in the home and church than first generation Chinese Canadians. This study is the first attempt to examine the vitality perceptions of different generations of a visible minority in Canada. Hewstone & Giles (1986) contend that stereotypes provide the content of social categories. Based on their analysis, at least four generalisations are warranted: stereotyping is the result of cognitive biases stemming from illusory correlations between group membership and psychological attributes; stereotypes influence the way information is processed (i.e. more favourable information is remembered about ingroups and more unfavourable information is remembered about outgroups); stereotypes create expectancies (hypotheses) about others and individuals try to confirm those expectancies; and stereotypes constrain others' patterns

< previous page

page_6

next page >

< previous page

page_7

next page > Page 7

of communication and engender stereotype confirming communication (i.e. they create self-fulfilling prophecies). Language cues the activation of stereotypes. Bond (1983), for example, examined the influence of language choice on autostereotypes (Chinese) and heterostereotypes (Westerners) of Chinese bilinguals in Hong Kong. His research indicated that language has a significant effect on activating different auto- and heterostereotypes. This finding is consistent with studies which demonstrate that the language of questionnaires used in research can affect the responses to items dealing with ethnicity (Bond & Yang, 1982; Marine, Triandis, Betancourt, & Kashima, 1983). Other research demonstrates that language's impact on stereotypes is different from that of ethnicity. Specifically, Bond (1985) found that language influences Chinese bilinguals' judgements of the likeability, Westernisation, and benevolence of Chinese and British speakers, while ethnicity influences the speakers' ethnic-group preference. Stereotypes also influence language choice. Beebe (1981), for example, found that when Chinese-Thai bilingual children are interviewed by a person who looks Chinese, but speaks standard Thai, they use Chinese phonological variants. Similarly, Bell (1982) discovered that stereotypes of listeners influence how New Zealand broadcasters read scripts. Language and stereotypes are the focus of two articles in this volume. Mgbo-Elue (this volume) examined the stereotypes Yorubas and Ibos in Nigeria hold regarding the other group and the affect these stereotypes have on individuals' desire to learn the outgroup language. She found a positive association between attitudes toward the outgroup and its language and desire to learn the outgroup language. Her research also revealed that positive stereotypes of the outgroup increase the likelihood that individuals will learn the outgroup language. This paper presents data from the first matched guise technique study in this part of Africa. Ribeau & Hecht (this volume) examined the labels Afro-Americans use to define themselves and their perceptions of interethnic communication effectiveness. They argue that ethnic identity is a function, at least in part, of semantic labels individuals use to define themselves. The semantic labels reflect a group identification and a consciousness of group membership. Ribeau & Hecht's research revealed that individuals who label themselves 'Black' perceive less negative stereotyping in their conversations with whites than individuals who label themselves 'Black-American' or 'Afro-American.' They contend that the use of the label 'Black' reflects the conservatism and assimilation motivation of the individuals selecting the label. Hewstone & Giles (1986) developed a stereotype-based model of intergroup communication breakdown which focuses upon sociolinguistic stereotypes and language production/reception strategies. They argue that in intergroup situations sociolinguistic stereotypes are activated. The stereotypes activated and the speech pattern used are a function of the individuals' ethnolinguistic identities that are relevant in the particular context, as well as the relations between the groups involved in the context (cf. Taylor & McKirnan, 1984). the relations between the groups influences the communicative distance established between the

< previous page

page_7

next page >

< previous page

page_8

next page > Page 8

groups (e.g. Lukens', 1979, distances of 'indifference,' 'avoidance,' and 'disparagement' or Gudykunst & Kim's, 1984, distances of 'sensitivity' and 'equality'). The ethnolinguistic stereotypes are used to explain difficulties in intergroup communication and make attributions about members of the outgroup. Language production and reception strategies feedback and reinforce ethnolinguistic stereotypes and, in combination with the context, lead to breakdowns in or dissolutions of intergroup communication. Given this general overview of language and social categorisation, we now turn our attention to the specific area of code-switching. Social Psychological Factors The major social psychological factor influencing code-switching is speech accommodation (for a recent review of research and a statement of speech accommodation theory in proposition form, see Street & Giles, 1982, or Thakerar, Giles & Cheshire, 1982; for an expansion of the theory to cover accommodation more generally, see Giles, Mulac, Bradac & Johnson, 1986; also see Gallois, Franklyn-Stokes, Giles & Coupland, in press, for an integration of communication accommodation theory and ethnolinguistic identity theory). There is often a tendency for members of ingroups to react favourably to outgroup members who linguistically converge toward them (Bourhis & Giles, 1976; Giles & Smith, 1979). This, however, is not always the case. Giles & Byrne (1982) point out that as an outgroup's members begin to learn the speech style of the ingroup, ingroup members will diverge in some way so as to maintain linguistic distinctiveness. Jargon is often a derivative of this effort to diverge from a linguistically close group, thus helping to provide control over group membership. Reaction to speech convergence by outgroup members thus depends upon the intent attributed to the speaker (Bourhis, 1984a; Genesee & Bourhis, 1982; Simard, Taylor & Giles, 1976). The use of speech divergence and/or maintenance to assert positive group identity has been demonstrated in reference to interethnic encounters (Bourhis, 1984a; Bourhis & Giles, 1977; Giles, Bourhis & Taylor, 1977). Values influence the nature of the accommodation that occurs. Bond & Yang's (1982) research, for example, revealed that the more importance Chinese in Hong Kong place on Chinese values, the more they display ethnic affirmation and the less they display interethnic accommodation. Context also affects accommodation. Research in Taiwan (van den Berg, 1986) indicated that sales people accommodate to the language customers use, while customers in banks accommodate to the clerk's language usage. The language itself also influences accommodation. John, Young, Giles & Hofman (1984) found that Arabs and Jews in Israel differentiate ingroup values from outgroup values more when tested in Hebrew than when tested in English. In related research, Genesee & Bourhis (1982) studied the influence of situational norms, sociocultural status, ingroup favouritism, and interpersonal accommodation on evaluations of code-switching. Their findings indicated that social

< previous page

page_8

next page >

< previous page

page_9

next page > Page 9

categories and situational norms interact to influence evaluations of language choice and that this interaction also impacts upon the role of ingroup favouritism. They argue that one group's evaluations of code-switching by members of another group is not consistent throughout a conversation. Evaluation of language choice initially is based on situational norms, but it is based on interpersonal language accommodation later in conversations. One plausible explanation is that individuals begin to view the other person less in terms of a member of a social category and more as a distinct individual. Bourhis (1985) extended Genesee & Bourhis' (1982) analysis in an attempt to integrate sociolinguistic and speech accommodation approaches to code-switching. He examined customer-clerk communication between Quebec Anglophones and Francophones. His research revealed that listeners' evaluations of dialogues depends 'on a dynamic interaction of factors including situational norms, language status, interpersonal accommodation and in-group favoritism' (p. 130). Bourhis' data indicated that following situation norms is the safest strategy in initial stages of hostile intergroup encounters, but normative strategies are adopted in later stages depending upon the communicators' goals, desire to assert group identity, and their affective responses to each other. The initial choices, however, influence the choices made later in conversations. Bourhis (1987) outlined initial research to extend this line of work to superior-subordinate language-usage within the Canadian government. His linguistic Work Environment Survey developed for this research should prove useful in other studies designed to examine language use, speech accommodation, and ethnolinguistic vitality in bilingual work settings. While all of the contributions to this volume were summarised in the previous two sections, several also examined accommodation processes. Banks' study, for example, revealed that minorities must accommodate to the majority regarding their discourse before access to power positions becomes available to them. Similarly, San Antonio's study of the use of English in an American corporation in Japan reveals that Japanese accommodation to the use of English and following American communication norms is a prerequisite to improving their status in the organisation. Other articles in this volume (e.g. Ros, Cano & Huici; McNamara; Louw-Potgieter & Giles) also indirectly examined language accommodation processes. Ball, Giles & Hewstone (1985) extended accommodation theory to take into consideration situational construals. Similar to their earlier extension of the intergroup model of second language acquisition (Ball, Giles & Hewstone, 1984), they argue that a catastrophe model is necessary to account for the joint influence of individual predispositions and social pressures (e.g. situational norms) on speech accommodation. In group-salient encounters, both convergence and divergence occur, but there is no intermediate accommodative speech used except under conditions of extreme hostility or friendliness. A cusp catastrophe, therefore, is necessary to explain accommodation under these conditions. The dynamic which accounts for convergence or divergence is the maximisation of self-esteem in terms of the intergroup and interindividual comparisons being made.

< previous page

page_9

next page >

< previous page

page_10

next page > Page 10

Finale The purpose of this essay was to overview research on language and ethnic identity and place the contributions to the present volume within the context of major research traditions. Space limitations, however, did not allow for a complete review of previous research (see Gudykunst & TingToomey, in press, for a more complete review). The research reviewed and the articles included in this Special Issue clearly demonstrate that the study of language and ethnic identity is a dynamic and rigorous research area within the study of the social psychology of language. The contributions to the present volume were selected to represent (1) the different levels of analysis (e.g. micro- and macrosociolinguistic and social psychological), (2) the different topics of research (e.g. language attitudes, speech markers, ethnolinguistic vitality, sociolinguistic stereotypes), (3) the different methodological approaches (i.e. self-report questionnaires, interviews, matched guise experiments, discourse analysis, and participant observation), and (4) the research conducted in different parts of the world (i.e. Australia, Canada, England, Japan, Nigeria, Spain, South Africa, and the United States). Taken together, the articles in this Special Issue emphasise the importance of social identity theory (e.g. Tajfel, 1981; Turner, 1987) and ethnolinguistic identity theory (e.g. Giles & Johnson, 1987) for the study of language and ethnic identity. All studies contained in this volume (including those using discourse analysis and ethnographic observation) draw on these theories directly or indirectly. The research presented herein, therefore, demonstrates that social and ethnolinguistic identity theories have the scope necessary to explain the influence of ethnic identity on such diverse processes as language attitudes, second language competence, sociolinguistic stereotypes, perceived ethnolinguistic vitality, and speech accommodation. These related theoretical perspectives provide a powerful heuristic for the study of language and ethnic identity. Several important theoretical and methodological issues also emerge from the papers in this volume. Theoretically, the degree to which individuals self-categorise themselves as members of ethnic groups, for example, is critical to understanding their language attitudes (Abrams & Hogg). In multiethnic environments, however, self-categorisation alone cannot adequately explain language attitudes. Rather, it is necessary to take into consideration the difference between self-categorisation as a member of the larger culture and self-categorisation as a member of a specific ethnolinguistic group (i.e. Ros, Cano & Huici's subtractive identity). Further, individuals not only use specific linguistic strategies to define themselves, they use linguistic strategies to deal with identities imposed by members of outgroups (Louw-Potgieter & Giles). In addition, subordinate group members' convergence toward dominant group language subtracts from their ethnolinguistic identity in some situations (e.g. Banks' study of Mexican-Americans in a predominantly white organisation), but not in others (e.g. San Antonio's study of Japanese in a United States corporation in Japan). Future research, however, is necessary to define the types of situations under which convergence is subtractive. Finally, the propositions Louw-Potgieter & Giles present regarding

< previous page

page_10

next page >

< previous page

page_11

next page > Page 11

imposed identity and dissent offer new theoretical insights into intra- and inter-group processes and set an agenda for future research. The major methodological issue to emerge from the contributions to this volume has to do with the language in which respondents answer questions. Punetha, Giles & Young found that the language in which Asian immigrant respondents choose to answer questionnaires influenced the degree to which their values approached those of the indigenous population; those choosing to answer in English (outgroup language) held values closer to the indigenous population than those choosing to answer in their ingroup language. This clearly suggests that language choice is linked closely to ethnic identity. Most studies of language and ethnic identity, however, do not allow respondents to choose the language in which they want to be interviewed or answer questionnaires. Future research and reviews of past research, therefore, must take care in drawing conclusions regarding ethnic identity and other variables when respondents are not forced to answer questions in a specific language and given a choice of the language in which they wish to respond. The contributions which follow not only reflect the recent developments in the study of language and ethnic identity, but also suggest new directions for future research. Given the diversity of the research presented in this volume, it provides a good overview of where the field is and where it is going in the future. Acknowledgements * Howie Giles suggested this Special Issue, provided assistance throughout its preparation, and commented on an earlier version of this essay. His constant support is acknowledged gratefully. Steve Banks also provided valuable suggestions on an earlier version of this essay. References Ball, P., Giles, H., and Hewstone, M. (1984) The intergroup theory of second language acquisition with catastrophic dimensions. In H. Tajfel (ed.). The Social Dimension (Vol. 2). Cambridge: Cambridge University Press. (1985) Interpersonal accommodation and situational construals: An integrative formulation. In H. Giles & R. St. Clair (eds), Recent Advances in Language, Communication, and Social Psychology. London: Lawrence Erlbaum. Beebe, L. M. (1981) Social and situational factors affecting the strategy of dialect code-switching. International Journal of the Sociology of Language, 32, 139-49. Beebe, L. M. and Giles, H. (1984) Speech accommodation theories: A discussion in terms of second-language acquisition. International Journal of the Sociology of Language, 46, 5-32. Bell, A. (1982) Radio: The style of news language. Journal of Communication, 32. 150-64. Bond, M. H. (1983) How language variation affects inter-cultural differentiation of values by Hong Kong bilinguals. Journal of Language and Social Psychology, 2, 57-66. (1985) Language as a carrier of ethnic stereotypes in Hong Kong. The Journal of Social Psychology, 125, 53-62. Bond, M. H. and King, A. Y. C. (1985) Coping with the threat of Westernization in Hong Kong. International Journal of Intercultural Relations, 9, 351-64.

< previous page

page_11

next page >

< previous page

page_12

next page > Page 12

Bond, M. H. and Yang, K. S. (1982) Ethnic affirmation versus cross-cultural adaptation: The variable impact of questionnaire language on Chinese bilinguals in Hong Kong. Journal of Cross-Cultural Psychology, 13, 169-85. Bourhis, R. Y. (1979) Language in ethnic interaction. In H. Giles & R. Saint-Jacques (eds), Language and Ethnic Relations. Elmsford, NY: Pergamon. (ed.) (1984a) Conflict and Language Planning in Quebec. Clevedon, England: Multilingual Matters. (1984b) Cross-cultural communication in Montreal: Two field studies since the Charter of the French language. International Journal of the Sociology of Language, 46, 33-47. (1985) The sequential nature of language choice. In R. Street & J. Capella (eds), Sequence and Pattern in Communicative Behaviour. London: Edward Arnold. (1987) Linguistic Work Environments and Language Usage in Bilingual Settings. Paper presented at the Second World Basque Congress, Basque Language Conference, San Sebastian, Spain. Bourhis, R. Y, and Giles, H. (1976) The language of co-operation in Wales. Language Sciences, 42, 13-16. (1977) The language of intergroup distinctiveness. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Bourhis, R. Y., Giles, H. and Rosenthal, D. (1981) Notes on the construction of a 'subjective vitality questionnaire' for ethnolinguistic groups. Journal of Multilingual and Multicultural Development, 2, 145-55. Bourhis, R. Y. and Sachdev, I. (1984) Vitality perceptions and language attitudes: Some Canadian data. Journal of Language and Social Psychology, 3, 97-126. Clément, R. and Kruidenier, B. G. (1983) Orientations in second language acquisition. Language Learning, 33, 273-91. Dittmar, N. (1976) Sociolinguistics. London: Edward Arnold. Eastman, C. M. (1985) Establishing social identity through language use. Journal of Language and Social Psychology, 4, 1-26. Edelman, M., Cooper, R. L. and Fishman, J. A. (1966) The contextualization of school-children's bilingualism. Irish Journal of Education, 2, 10611. Fishman, J. A. (1967) Bilingualism with and without diglossia: Diglossia with and without bilingualism. The Journal of Social Issues, 23, 29-38. Gallois, C., Franklyn-Stokes, A., Giles, H. and Coupland, N. (in press) Communication accommodation theory and intercultural encounters: Intergroup and interpersonal considerations. In Y. Y. Kim & W. B. Gudykunst (eds), Theory in Intercultural Communication. Newbury Park, CA: Sage. Gardner, R. C. (1985) Social Psychology and Second Language Learning. London: Edward Arnold. Genesee, F. and Bourhis, R. Y. (1982) The social psychological significance of code-switching in cross-cultural communication. Journal of Language and Social Psychology, 1, 1-27. Giles, H. (ed.). (1977) Language, Ethnicity, and Intergroup Relations. London: Academic Press. (1979) Ethnicity markers in speech. In K. R. Scherer & H. Giles (eds), Social Markers in Speech. Cambridge: Cambridge University Press. Giles, H., Bourhis, R. and Taylor, D. M. (1977)Towards a theory of language in ethnic group relations. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Giles, H. and Byrne, J. (1982) An intergroup approach to second language acquisition. Journal of Multilingual and Multicultural Development, 3, 17-40. Giles, H., Garrett, P. and Coupland, N. (1987) Language Acquisition in the Basque Country: Invoking and Extending the Intergroup Model. Paper presented at the Second World Basque Conference, Basque Language Conference, San Sebastian, Spain. Giles, H. and Johnson, P. (1981) The role of language in ethnic group relations. In J. Turner & H. Giles (eds), Intergroup Behavior. Chicago: University of Chicago Press.

< previous page

page_12

next page >

< previous page

page_13

next page > Page 13

(1987) Ethnolinguistic identity theory: A social psychological approach to language maintenance. International Journal of the Sociology of Language, 68, 69-99. Giles, H., Mulac, A., Bradac, J. J. and Johnson, P. (1986) Speech accommodation theory: The next decade and beyond. In M. McLaughlin (ed.), Communication Yearbook 10. Beverly Hills, CA: Sage. Giles, H. and Powesland, P. F. (eds) (1975) Speech Style and Social Evaluation. London: Academic Press. Giles, H., Rosenthal, D. and Young, L. (1985) Perceived ethnolinguistic vitality: the Anglo- and Greek-Australian setting. Journal of Multilingual and Multicultural Development, 6, 253-69. Giles, H. and Saint-Jacques, B. (eds) (1979) Language and Ethnic Relations. Oxford: Pergamon Press. Giles, H. and Smith, P. M. (1979)Accommodation theory: Optimal levels of convergence. In H. Giles & R. St. Clair (eds), Language and Social Psychology. Oxford: Basil Blackwell. Giles, H., Taylor, D. M. and Bourhis, R. Y. (1973) Toward a theory of interpersonal accommodation through speech: Some Canadian data. Language in Society, 2, 177-92. Gudykunst, W. B. and Kim, Y. Y. (1984) Communicating with Strangers: An Approach to Intercultural Communication. New York: Random House. Gudykunst, W. B. and Ting-Toomey, S. (in press) Language and social categories. In H. Giles & P. Robinson (eds), Handbook of Social Psychology and Language. London: Wiley. Gumperz, J. (1982a) Discourse Strategies. London: Cambridge University Press. (ed.) (1982b) Language and Social Identity. London: Cambridge University Press. Gumperz, J. and Hymes, D. (eds) (1972) Directions in Sociolinguistics. New York: Holt, Rinehart, & Winston. Hall, B. J. and Gudykunst, W. B. (1986) The intergroup theory of second language ability. Journal of Language and Social Psychology, 5, 291-302. Herman, S. (1961) Explorations in the social psychology of language choice. Human Relations, 14, 149-64. Hewstone, M. and Giles, H. (1986) Social groups and social stereotypes in intergroup communication. In W. B. Gudykunst (ed.), Intergroup Communication. London: Edward Arnold. Hogg, M. A., Joyce, N. and Abrams, D. (1984) Diglossia in Switzerland? A social identity analysis of speaker evaluations. Journal of Language and Social Psychology, 3,185-96. Hymes, D. (1972) Models of the interaction of language and social life. In J. Gumperz & D. Hymes (eds), Directions of Sociolinguistics. New York: Holt, Rinehart and Winston. John, C., Young, L., Giles, H. and Hofman, J. E. (1984) Language, values, and intercultural differentiation in Israel. The Journal of Social Psychology, 125, 527-29. Kimble, J., Jr., Cooper, R. L. and Fishman, J. A. (1969) Language switching and the interpretation of conversations. Lingua, 23, 127-34. Labrie, N. and Clément, R. (1986) Ethnolinguistic vitality, self-confidence and second language proficiency: An investigation. Journal of Multilingual and Multicultural Development, 7, 269-82. Lambert, W. E., Mermigis, L. and Taylor, D. M. (1986) Greek Canadian's attitudes toward own group and other Canadian ethnic groups: A test of the multiculturalism hypothesis. Canadian Journal of Behavioural Sciences, 18, 35-51. Lukens, J. (1979) Interethnic conflict and communicative distance. In H. Giles & R. Saint-Jacques (eds), Language and Ethnic Relations. Elmsford, NY: Pergamon. Marin, G., Triandis, H. C., Betancourt, H. and Kashima, Y. (1983) Ethnic affirmation versus social desirability. Journal of Cross-Cultural Psychology, 14, 173-86. McNamara, T. F. (in press) Language and social identity: Some Australian studies. Australian Review of Applied Linguistics. Pak, A., Dion, K. L. and Dion, K. K. (1985)Correlates of self-confidence with English among Chinese students in Toronto. Canadian Journal of Behavioural Sciences, 17, 369-78.

< previous page

page_13

next page >

< previous page

page_14

next page > Page 14

Rubin, J. (1968) National Bilingualism in Paraguay. The Hague: Mouton. Ryan, B. E. and Carranza, M. A. (1977) Ingroup and outgroup reactions to Mexican-American language varieties. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Scherer, K. R. and Giles, H. (eds) (1979) Social Markers in Speech. Cambridge: Cambridge University Press. Sechrest, L., Flores, L. and Arellano, L. (1968) Language and social interaction in a bilingual culture. Journal of Social Psychology, 76, 155-61. Simard, L. M., Taylor, D. M. and Giles, H. (1976) Attributional process and interpersonal accommodation in a bilingual setting. Language and Speech, 19, 374-387. Street, R. and Giles, H. (1982) Speech accommodation theory. In M. Roloff & C. Berger (eds), Social Cognition and Communication. Beverly Hills, CA: Sage. Tajfel, H. (1978) Social categorization, social identity, and social comparison. In H. Tajfel (ed.). Differentiation between Social Groups. London: Academic Press. (1981) Human Categories and Social Groups. Cambridge. MA: Cambridge University Press. Taylor, D. M. and McKirnan, D. (1984) A five-stage model of intergroup relations. British Journal of Social Psychology, 23, 291-300. Thakerar, J. N., Giles, H. and Cheshire, J. (1982) Psychological and linguistic parameters of speech accommodation theory. In C. Fraser & K. Scherer (eds), Advances in the Social Psychology of Language. Cambridge: Cambridge University Press. Turner, J. (1987) Rediscovering the Social Group. London: Basil Blackwell. van den Berg, M. E. (1986) Language planning and language use in Taiwan: Social identity, language accommodation, and language choice behavior. International Journal of the Sociology of Language, 59, 97-115. Young, L., Giles, H. and Pierson, H. (1986) Sociopolitical change and perceived vitality, International Journal of Intercultural Relations. 10, 459-69.

< previous page

page_14

next page >

< previous page

page_15

next page > Page 15

Achieving 'Unmarkedness' in Organisational Discourse: A Praxis Perspective on Ethnolinguistic Identity Stephen P. Banks Department of Communication, Arizona State University Abstract An ethnomethodological approach to ethnolinguistic identity is presented using the linguistic concept of 'markedness' as a heuristic innovation to explore the dynamics of ethnolinguistic minorities in work settings. A matrix is proposed that consists of a marked/unmarked boundary and a high/low power boundary. Four propositions relating to crossing the boundaries are advanced, and illustrative evidence supporting each proposition is developed in a discourse analysis of talk in a large hotel work site. The paper concludes that a more praxisoriented perspective on ethnolinguistic identity can enrich understanding of the three concepts embedded in that term ethnicity, language, and identity. In the last analysis, any sociopsychological image of the self, in fact the very possibility of a self concept, is inextricably dependent on the linguistic practices used in everyday life to make sense of our own and others' actions. (Potter & Wetherall, 1987) Before the turn of the century, workers in U.S. factories who considered themselves to be ethnic minorities could retreat to the linguistic and cultural familiarity of 'ethnic enclaves' within their workplaces (Norkunas, 1987). Such insulation was possible because minimal interaction with others was permitted under the forms of command and control exercised by early factory managers. The availability of such enclaves was crucial to the preservation of group identities, however, because the workplace was the site of interethnic contact par excellence, and contact often implied conflict that was unanticipated and unmanaged by employers. Today the situation for ethnic minorities is vastly different from the early factory experience and at the same time exaggeratedly like it. It is different in virtue of the decades-long infusion of human relations theory into management practices and employers' recognition of the production gains associated with positive, informal social relationships in the workplace. Now interaction is widely encouraged,

< previous page

page_15

next page >

< previous page

page_16

next page > Page 16

and isolates are ferreted out so they can be integrated into communication networks (Rogers & Rogers, 1976). Ethnic minorities' situations today also are different from those of the last century because of technological and occupational changes that require ever-greater degrees of worker interaction (Forester, 1981; Simon, 1981). Much about the worklife situations of ethnic minorities, however, has not changed, except to grow more extreme. The ethnic and linguistic composition of workplaces generally is becoming more polymorphous in the U.S. and Canada (Fishman, Gertner, Lowy & Milan, 1986; Glazer, 1983; Lachapelle & Henripin, 1986). Moreover, Jupp, Roberts & Cook-Gumperz (1982) have noted that in Great Britain over the past decade there has emerged a new multiethnic and multilingual workplace. As the planet shrinks in terms of communication and transportation, it also is possible that newcomers, for a variety of reasons related to employment conditions, will be less able than their predecessors to balance assimilation and ethnic identity (Glazer, 1983). Thus the workplace remains the site par excellence where intergroup contact and tensions are experienced. It follows that work institutions are places where ethnolinguistic identities likewise experience crucial challenges and tensions. The term, ethnolinguistic identity, presents several major problems that all too commonly are overlooked in research and theory development. Ambiguities emerge from analysis of the term's three component concepts ethnicity, language, and identity. It would be trivial to point out that various researchers apply different definitions to some or all of these components; a more fundamentally conceptual examination of each is in order so that 'we can . . . check the insights [already] claimed against other evidence' (Edwards, 1985: 155). The following section of this paper examines each of the three component concepts as a way of developing an understanding of ethnolinguistic identity that is grounded in ethnomethodology and discourse. A brief characterisation of some problematic aspects of each concept is followed by a revised, ethnomethodological view that solves some definitional or operational problem. I then present an argument for analysing interethnic discourse in terms of markedness theory (Comrie, 1976; Levinson, 1983), and, relying on 'insights already claimed' and the implications of markedness, I advance four propositions about the relationships between the power structures in work organisations and ethnolinguistic identities of members who are minorities (in the sense of ethnic nonpowerholders). To provide 'other evidence' against which to check claims generally and to illustrate the propositions, I then analyse transcribed segments of discourse of managers and other workers in a large hotel in the southwestern part of the United States. Refocusing Ethnicity, Language, and Identity Ethnicity as Phenomenon Despite decades of theory and research, little agreement exists among scholars on how to conceptualise ethnicity (Edwards, 1985; Gans, 1979; Harmann, 1986;

< previous page

page_16

next page >

< previous page

page_17

next page > Page 17

Martinelli, 1986; Ross, 1979). Perhaps the most popular view of ethnicity among researchers in ethnolinguistic identity is that adopted by Edwards (1985), following his detailed review of prominent perspectives on ethnicity. To Edwards, ethnicity is a form of historical allegiance to a group, invoking both a felt group boundary and some attachment to an observably real past (p. 10). An elaboration of this basically static, subjective approach, which is inspired by Tajfel's (1978, 1981) work on the social psychology of groups and Barth's (1969) work on boundaries, is the multifaceted view of Giles and his associates (e.g. Bourhis, Giles & Rosenthal, 1981; Edwards & Giles, 1984; Giles & Byrne, 1982; Giles & Johnson, 1981). Their by now familiar 'ethnolinguistic vitality' theory, including the more recent excursions into group members' subjective assessments of group vitality, focuses on the group's social status, demographic profile, and institutional support. A clear definition of ethnicity in this tradition is not apparent; it is assumed to be generally coterminous with the boundaries of the group identified by the researcher as worthy of comment and study. More anthropological views of ethnicity debate whether objective features such as race, nationality, linguistic group, tribe, or religion or subjective aspects such as a member's perception of belonging to any such social or descriptive category determine any individual's ethnicity (see, e.g. Burgess, 1978; Martinelli, 1986). Both of these traditional views and the social psychological approaches just mentioned conceive of ethnicity as a relatively enduring quality a member has, in virtue of either self-identification or external ascription. None accounts well for situational changes in perceived ethnicity that result from conditions of interaction (Smith, 1984) or for longer-term changes based on evolution of bilingual-diglossic circumstances (Martinet, 1986). From a praxis oriented view, however, ethnicity is more fruitfully understood as evidence in a specific interaction at a given historical moment which is inferentially adequate for judging an individual's degree of attachment to some identifiable cultural group. Insofar as ethnicity denotes difference (Barth, 1969), the cultural group to which an individual may be inferred to be attached typically is the nonstandard, nonlocal, or minority group, and in all instances would be the less normative group for the given circumstances of interaction. Thus, ethnicity is not an inherent quality of dress, cuisine, gesture or language but is an inference about a member's group attachment that might be based on the presence of certain attributes like dress, cuisine, gesture or language. Moreover, such judgements are made, often simultaneously, by both the ethnic person and others present, and the inferences are arrived at by a process of social negotiation in virtual time and space (Smith, 1984). One's sense of one's own ethnicity waxes and wanes, emphasising some group features and then others, depending on the current social situation one is engaged in (Gans, 1979). The ethnomethodological approach to ethnicity, then, accounts for changes in perceptions of a member's ethnicity situationally and over time, both as perceived by self or by others. When ethnicity is linked to concrete interactional settings, it is a 'formal structure of practical action' (Garfinkel & Sacks, 1970), which is taken not as epiphenomenon but as phenomenon in its own right. As such ethnicity is not a representational term that indexes a more abstract quality of the individual;

< previous page

page_17

next page >

< previous page

page_18

next page > Page 18

rather it is the basis for inferences about the individual within a specific social circumstance. Language as Discourse Many treatments of ethnolinguistic identity embrace 'language' as an indexical expression that refers to a system of symbols and rules for their use that is shared by some unique group (e.g. Bourhis, 1979; Fishman, 1986; Giles, Bourhis & Taylor, 1977; Giles & Johnson, 1987; Haarmann, 1986; Khubchandani, 1983; Shukla, 1985). Similarly, social identity theory encompasses language choice (e.g. Bond, 1983; Liebkin, 1982) and language attitudes (e.g. Ryan & Giles. 1982) within the same systematic, macrosociolinguistic view of language. Speech accommodation theory (Giles & Powesland, 1975; Giles, Mulac, Bradac & Johnson, 1986) and studies of speech stereotyping (e.g. Hopper & Williams, 1973; Kalin & Rayko, 1980) have contributed much to ethnolinguistic identity research. What is not given priority in these views of language is its situated instantiation as discourse the record of communicative interaction by persons in virtual time and space. Viewing the language component of ethnolinguistic identity as discourse takes account of situational factors, such as the nature, role and moods of participants, the physical settings, social histories, topics of discussion, etc. It also situates social meaning as it emerges in ongoing discourse in the centre of the analysis of ethnolinguistic identity (Chock & Wyman, 1986). An important modification of the established approach to language, then, is to include contexts of language use and to focus on particular social meanings in talk. Identity as Produced in Talk Turner (1982: 19) discriminates both self-concept, a relatively enduring cognitive structure, and a subset labelled self-images, which are identities produced independently in any given situation. Situationally, locus of control over identity is switched from personal to social (i.e. group) identity, depending on conditions of the moment. While most discussions of ethnolinguistic identity give nodding recognition to self-images, the predominant assumption about identity is that it is a transituational cognitive structure of individuals that reflects beliefs about group membership (Tajfel, 1981). Lost in this view of identity is the aspect of identity that responds to the individuals's present circumstances. Czyzewski (1982) points out, however, that powerful concepts of social identity have been developed in the symbolic interactionist tradition. He critiques the contributions of Mead, Goffman, Sacks and others and concludes that communicative experience is ineluctably social and 'does not exist beyond embodied practices' (p. 7). Hadden & Lester (1978) demonstrate how self-identity is produced in talk. They reframe identity in interactionist terms as the emergence of self through interactionist work:

< previous page

page_18

next page >

< previous page

page_19

next page > Page 19

As interaction proceeds temporally, others' responses or assumptions about others' responses can reinform both the form and content of a person's identifying work. Thus, members' stock of knowledge is not a static stockpile of typifications. Rather, it is changed and reinforced as particulars are invoked and assembled into patterns in order to display any aspect of the social world. In short, accounts of any 'thing', and of identity in particular, must be conceptualized as a continually evolving process. (p. 333) Taking identity to be a feature of participants that is created and displayed in interaction facilitates inferences about possible relationship between language and ethnicity. It moves analysis one step closer to authentic speakers whose authentic attitudes about their language is evident in their specific instances of talk (Crystal, 1986). An ethnomethodological approach to ethnolinguistic identity thus provides two significant modifications of the currently popular research postures. First, it allows ethnicity to 'float' upon the instant of interaction, requiring it to be neither objective and historical nor hidden and psychological. Ethnicity becomes what ethnicity does interactionally. Second, a praxis view focuses analytic attention on language as both subject and object; the work that discourse does must be available in the discourse itself. Neither of these points negates human memory for past interaction and the cognitive structure of Turner's self-concept; they do, however, add to the analysis of identity a pertinent focus on contributions of the here-and-now. Certainly, interpretation of any discourse must call on analysis of the manifold social circumstances, as well as analysis of the microlinguistic and immediate interactional events (Brown & Yule, 1983). A contextual dimension that must be considered when one seeks to understand ethnolinguistic identity in work organisations is that of institutional and individual power (Deschamps, 1982). The next section discusses power as part of a matrix with social markedness. Marked and Unmarked Discourse The term 'marked' or 'marker' occasionally has been used to refer to features of discourse that identify the speaker as belonging to some category or other (e.g. Scherer & Giles, 1979; Schiffrin, 1987; Scotton, 1983). This view of markedness entails the concept of unmarked speech, that is, speech which does not exhibit cues of personal categories the analyst is interested in. Hymes (1974) refers to unmarked speech as 'natural' and uses the concept of markedness to describe situations in which speech style is incongruous with the social circumstances (p. 111). A more broadly applicable sense of markedness, and one that is more conventionally used among sociolinguists, is the case where one element 'is more usual, more normal, less specific than the other (in markedness terminology it is marked, the others unmarked)' (Comrie, 1976). Levinson (1983) builds on this notion by asserting that marked cases have more morphological material and more likelihood of irregularity than unmarked cases. Moreover, unmarkedness is related to

< previous page

page_19

next page >

< previous page

page_20

next page > Page 20

preference: markedness generally signals nonstandardness in a negatively valued sense, such that an individual who exhibits markedness is seen as different from a standard, normal, and natural reference category. I apply the concept of markedness in this broad sense to describe the interaction of ethnolinguistic minorities and non-minorities in work settings. The most extremely marked case of linguistic differences is that in which one member does not speak the language that constitutes the standard language of the enterprise. The literature on language standardisation is much too broad to review here; I will assume that in any work institution one system of linguistic and pragmatic rules generally is preferred by powerholders over other systems (cf. Banks, 1987). A somewhat less marked case located toward the other end of the difference continuum is the ethnolinguistic minority who has not mastered or wishes not to display the prescribed phonological practices of the standard. Although there is a continuum of discourse practices ranging from vastly different to identical, at some point of difference between an intuitively known standard and the discourse of the moment the difference becomes noticeable to the participants; at that point markedness of speech begins. The dispreference for marked discourse, however, is grounded in more than linguistic differences (Dittmar & yon Stutterheim, 1985; Edwards, 1979; Gumperz, 1982). Linguistic practices inevitably are bound up with socioeconomic status and interpersonal perceptions of speaker traits. Thus, marked speech is marked in virtue of both linguistic differences and status differences of the participants in interaction. The reciprocal influence of speech and social variables on markedness has been pointed out repeatedly (e.g. Brown & Levinson, 1979; Edwards, 1979; Trudgill, 1983). This is not to say that the discourse of high status persons who use language that deviates noticeably from the standard will be unmarked. It is just that, as Gwyn Williams (1979) notes, they have fewer boundaries to cross to achieve institutional successes. In a rough analogy to Giles' (1979) model of ethnic boundaries, the boundary between marked and unmarked discourse is 'soft'. By this I mean that individuals find the social milieux in work institutions conducive to their movement from producing marked to producing unmarked discourse. Employers encourage workers to learn effective expression, and 'effective' entails the sense of standardness. This is so because in modern, complex organisations, sharing discourse practices not only simplifies and makes more efficient task-oriented talk but also promotes cohesiveness and a common sense of direction among members. In addition, unmarkedness draws markedness toward standardisation because of the non-normative nature of markedness and because of the association of standardness with power and prestige (Kramarae, Schulz & O'Barr, 1984). Deschamps (1982) recognises this relationship and describes the institutional consequence of power and linguistic differences on identity: 'The dominant discourse assigns their place to individuals through defining, locating and ordering each of them in relation to others' (p. 88). In this compact statement, he includes three elements that are key to a fuller understanding of language and ethnic identity: (a) it is discourse and not language qua language that affects the ordering of social relations in workplaces; (b) linguistic and institutional power differences

< previous page

page_20

next page >

< previous page

page_21

next page > Page 21

may be analysed as affecting individuals, regardless of their group membership; and (c) dominance is massively involved in ordering group membership and members' identity. Deschamps posits two aspects to identity, 'de facto' identity and 'imaginary' identity. The former refers to identity that individuals accrue by attributions through social relationships, i.e. the 'me'; imaginary identity is an application of situational norms describing a proper subject, a normative 'I'. Deschamps argues that dominant group members collapse these two aspects of identity as members define themselves, since they have power to 'make normal' their sense of self. Dominated groups, on the other hand, define themselves as de facto belonging to some group, but must 'share . . . the imaginary I represented by the dominant' (p. 89). In this way, identity is parallel to discourse the marked case is the identity of the dominated, and the unmarked case is the identity of the dominant. This sense of markedness is pivoted on the allocation of institutional power. Powerful members establish norms of all kinds for others, both explicitly and indirectly through their routine discourse practices (Fowler, 1985; Fowler, Hodge, Kress & Trew, 1979). Power represents a second dimension in an (admittedly incomplete) description of ethnolinguistic minorities' fates in work institutions. The boundary between low and high power in organisations, marked at the threshold of managerial positions, is relatively impermeable. At the first level of management, members' community of interests and goals are assumed generally to be identical with those of the organisation; it is here that one becomes identified with 'master' and not with 'employee' (Perrow, 1986). Often this special boundary crossing is signalled by institutional changes a separate compensation system, loss of union protection, access to managerial privileges, acquisition of authorities to direct the behaviour of others. The two factors I have explored that influence ethnolinguistic identity in work institutions can be laid out in a simple matrix.

< previous page

page_21

next page >

< previous page

page_22

next page > Page 22

Based on the preceding discussion and the assumption that relatively permeable boundaries will be crossed more readily than relatively less permeable ones, several generalisations can be offered about crossing these particular boundaries. First, movement from marked to unmarked discourse will precede an individual's movement from low to high power. Not only does a member in quadrant 4 who attempts to standardise his or her discourse become more favourably perceived by members in quadrants 2 and 1, but standardisation also might be a prerequisite for such boundary-crossers to articulate their special needs, talents and potential contributions (Dittmar & von Stutterheim, 1985). In addition, movement from low to high power is a consequence of high power members specifically authorising any boundary-crossing; the movement from marked member to unmarked member typically occurs at the initiative of the boundary-crosser. Second, there is more to achieving unmarkedness than the quest for material rewards in the organisation, although advancement still might be a powerful motivation (cf. Kanter & Stein, 1979, for various perspectives on upward mobility). The workplace is an encompassing, socialising environment; membership in a work organisation submerges individuals in a pool of pre-established norms and values that transcend those of extraorganisational reference groups. Similarly, the workplace setting is constituted by a set of routine interactional practices, particularly discourse routines, that members must learn so as to be effective. All these influences pressure members with marked uses of language to seek unmarked-ness. Third, since standardness is invoked by high power members and is inherent in the identity of high power members, they express an implicit invitation to low power members to 'be like' them, not only in terms of behaviour, appearance and rewards, but also like them in power. Aspiration to climb the corporate ladder is inculcated in part by example and demonstration; no respectable organisational theorist would omit from an exegesis on worker ambition and motivation a discussion of financial rewards and perquisites. Thus, ethnolinguistic minorities are likely to anticipate a subsequent boundary-crossing from low to high power as being facilitated by a prior move from marked to unmarked discourse. The expected movement in the matrix, then, is from 4 to 2 and subsequently from 2 to 1; high power marked speakers generally would not be expected to move to unmarkedness, so movement from 3 to either 2 or 1 would not be observed. The fourth proposition involves effects of movement on the marked speaker. Weak forms of linguistic relativity theory and recent research findings (cf. Giles. 1979: 256-57) suggest that linguistically dominated groups who regularly adopt the language of dominant groups experience a degradation of their subordinate group identity. Thus it is likely that the 4 to 2 boundary-crossing would result in a sense of loss of earlier identity and a substitution of affiliation with the dominant group. Illustrations in Discourse Evidence that the foregoing propositions have more than intuitive appeal is found in the talk of members of a large hotel in the U.S. Southwest. The data were

< previous page

page_22

next page >

< previous page

page_23

next page > Page 23

recorded and transcribed by the author as part of an ongoing study of two hotels in a worldwide chain, which I have given the pseudonym 'Major Hotel corporation.' The first set of comments are provided by the Director of Sales (identified as SD) in an informal interview with two investigators (R1 and R2). SD is a third generation Japanese-American in her late twenties who was born in Hawaii and lived in a mixed Hawaiian/Japanese/White community until she was 21 years old. In this excerpt she responds to R1's question about speech characteristics and organisational advancement. 1 4 R1:

Is it likely or possible that a person with a heavy

5

accent uhm would be advanced to at least upper or

6

middle levels of management like the executive

7

committee or above?

8 SD:

A:hm, let's see if we have anybody like that. Ahm

9

(.) It's probably possible, ya:h but=ah: I don't

10

know anybody like tha=oh, chefs/. They all have

11

accents.

12 R1:

that's . . .

13

[

14 SD:

But that's acceptable.

15

[

16 R2:

Right.

17

[

18 R1:

Yeah.

19 SD:I mean all chefs. Ah=th=they don't deal really with 20

the public so much, they're behind doors. Ahm (.)

21

st (.) Not that I know of . . .

SD cannot identify any middle or upper level manager in the organisation who has a 'heavy accent'; then she notes that all chefs have accents. Her rising intonation on 'chefs' in line 10 implies that she is unsure if chefs qualify as members of the group asked about. In the hotels where SD has worked chefs are members of the executive committee; consequently, they dearly are members of upper level management. Her equivocation could be because of the degree to which chefs' accented speech is accepted as normal: most chefs in this hotel corporation are Europeans who enter the firm at high level positions, and their accented speech is part of 'chefness'. It also is possible that she sees chefs' accentedness as differing in some salient quality from that of Asian-Americans, particularly those from Hawaii who grew up speaking a Hawaiian creole. In either case, her reply to the question is in the negative. Notice that her account for why chefs can have accented speech is function-based (lines 19 and 20) chefs have an ethnolinguistic enclave within which they work. Yet later in the discourse, she accounts for why an aspiring Japanese-American manager would want to change his speech: 94 SD: They'll = they'll probably go, ''wadja sa:y?" or they 95

make fun of you so you don't (.) you won't wanna

96

talk like that anymore.

< previous page

page_23

next page >

< previous page

page_24

next page > Page 24

SD's view of chefs and accentedness evidently does not envision members of the organisation ridiculing chefs' accents or challenging how or what they say. Thus, SD's talk supports the contention that movement from quadrant 4 of the matrix to quadrant 2 is more likely than movement from 3 to 1 or 2. Also, the fact that she can not identify any manager who has moved from quadrant 4 to 3 indicates that in her experience, ethnolinguistic minorities in management must have moved from 4 to 2 or from 4 to l. Moving from low power, marked speech directly to high power, unmarked speech i.e. from 4 to 1 is the only movement pattern besides 4 to 3 that allows for crossing the power boundary before crossing the markedness boundary. The unlikelihood of a quadrant 4 to 1 movement is expressed in SD's description of her own motives: 28 SD: . . .And when I 29

came to //name of state// and worked here, I

30

decided that that wasn't gonna (.) help me in my

31

career. 'N' so then I, I really practiced at s=at

32

really enunciating 'n' saying the instead of duh,

33

and=ahm (.) a:nd not using the Hawaiian words.

Changing her speech was for SD an instrumental act, an effort that was aimed at helping her career advancement. The conscious, intentional aspect of her movement from 4 to 2 is expressed here; what is not highlighted is the perhaps less noticeable influence of challenges and ridicule that she referred to (lines 94-98, above) when analysing the speech of another, similarly situated Hawaiian-born manager. SD noted that he, too, would have to adjust his speech in order to avoid censure and to rise in the corporate hierarchy. Insofar as SD changed her discourse practices in anticipation of improved career advancement possibilities, she can be said to support the third proposition movement from marked to unmarked discourse anticipates movement from low to high power. The fourth proposition holds that crossing the soft boundary from marked to unmarked discourse degrades the member's attachment to the linguistically dominated group. Here SD's account is most dramatic: 33 SD: and=ahm (.) a:nd not using the Hawaiian words. = It 34

kinda pulls you away from your culture a bit. And

35

now that I go ba:ck, ahm and I speak like this,

36

this is to me: normal/and I can't (.) go back and

37

talk like that 'cause if I do and I try to, people

38

think, "oh you=you're t=you're trying to fit in:

39

an: . . .

40 R2: It's not ah: //inaudible// 41

[

42 SD: You're kinda in the middle. 43 R2: Yeah. 44 SD: 'Cause when [»] I go home 'n' I talk like this 45

they, they think that I'm trying t' be better than

46

them, an' if I try ta have (.) the Hawaiian accent

< previous page

page_24

next page >

< previous page

page_25

next page > Page 25

47 or the local accent there ahm and it=it obviously 48 doesn't sound right anymore so that they=they think 49 that . . . The first noteworthy feature of this account is the bind the boundary-crosser faces: if she speaks in her ethnolinguistic group with her dominant group speech she is rejected as 'trying to be better', and if she tries to speak as the linguistically subordinate group does she is labelled as an ingratiator. Having crossed the soft boundary, SD's new discourse patterns are 'normal' to her, and she sees her former linguistic group as marked with an accent. The bind is all the more trenchant when it is recognised that SD describes her rejoining the subordinate group in terms of 'going home'. It is unclear from the transcript shown here what SD means by 'it obviously doesn't sound right anymore' at lines 47 and 48. Conceivably 'it' refers to her creole, which doesn't sound right to her Hawaiian group members. Or perhaps her Hawaiian speech does not sound right to her. An alternative hypothesis is that her corporate speech does not sound right, but the inclusion of 'anymore' argues against that interpretation. The ambiguity in 'obviously doesn't sound right' is somewhat resolved in subsequent talk: 51 R1: You've lost some skill in speaking the di=in the 52

Hawaiian . . .

53 SD: Yeah. 54 R1: Pidgin. 55 SD: I know the words an' all that, but . . . 56 R1: There is an accent. 57 SD: I f=I feel kinda foolish to say, y'know (.) just 58

to not ta enunciate anymore.

The interviewer attempts here to provide his own account for what has happened to SD. When he suggests that she has lost her language skill, she begins a counter-explanation at 55; as she grapples with the subtlety of her situation, he interjects for her a possible predicate: 'There is an accent.' On the contrary, SD then says that she feels 'foolish' to speak in the ways of her ethnolinguistic group. The recurrence of 'anymore' harks back to line 48 and implies that to not enunciate anymore is to not sound right anymore. Hence weight is given to the explanation that the local speech doesn't sound right anymore means doesn't sound right to her; and her utterance at 47 to 49 could speculatively be completed with 'they think that I am trying to be better than they are because I don't approve of the sound of their speech anymore'. It is not so much her lost skill, then, as her lost identification with the discourse practices of her Hawaiian group that causes discomfort when she interacts with them. It is clear in this case that, even if her sense of ethnic group identity is not degraded by her boundary-crossing in the workplace, her linguistic identity is deeply affected toward the standards of social appropriateness held by the dominant group. Moreover, the degree to which ethnicity and language can be partitioned in circumstances such as this remains problematic.

< previous page

page_25

next page >

< previous page

page_26

next page > Page 26

Evidence from SD's talk presented so far in support of the four propositions is explicit. That is, the evidence is derived from the descriptions and accounts told by SD. A deeper level of evidence can be found in the patterns of discourse practices themselves among high power, unmarked speakers and low power, marked speakers. The following excerpt is taken from a new employee orientation session at the same hotel in which a personnel manager (MGR) has asked newcomers each to read aloud sections from the employee handbook as a way of familiarising them with company policies. The second speaker here (EMP) is a middle-aged female immigrant from Central America whose first language is Spanish. WKR is another newcomer, not an immigrant to the U.S. 15 MGR:

Ahm (.) Maria. Why don't you

16

take the next one \? Workers' compensation.

17 EMP:

Ow::, like reading I don' know the wor:ds.

18 MGR:

Oh:. Okay.

19 EMP:

Ha=heh=heh=heh.

20 MGR:

You could take=d'you=do you wanna try?

21 EMP:

M:. //inaudible// this one on insurance?

22 MGR:

Ah: this one right here.

23 EMP:

Okay. The ploymen' insurance is provided uhm I

24

don't say that right=hhh . . .

25 WKR:

You on workers' compensation?

26 MGR:

Uh huh/.

27 WKR:

//points to page// well it's on this. (.) This one.

28 EMP:

Ah working compention?

29 WKR:

M hm/.

30 EMP:

Okay. (.) This insurance is paid full by Mahor for

31

you medical 'spenses and continue (.) ah=I don't

32

know what what is.

33

[]

34 WKR:

Sources.

35

[

36 MGR:

You're doing very well.

37 EMP:

Sources of income:: under certain' cir=circum=

38

=s=s=cir . . .

39 [ 40 MGR:

Circumstances, um hm.

41 EMP:

Cumstances if you s::oon be: (.) o:: ort on the

42

yob.

43 MGR:

O:ka:y, good. Ahm, see you did fi::ne. (.) Ah:m

44

workers' compensation is provided by the state . . .

45

[]

46 EMP:

I was scare=hhh=heh.

The explicit evidence that markedness is drawn toward unmarked speech is the institutional exercise of having newcomers read aloud the corporate policies on employee benefits. MGR already is familiar with the language competencies of

< previous page

page_26

next page >

< previous page

page_27

next page > Page 27

each newcomer because she has had them introduce themselves and talk briefly about their previous employment and hobbies; consequently, she must have anticipated Maria's difficulties with reading the section on workers' compensation. By performing this exercise, however, newcomers are situated inside both the employer's policy perspective and the employer's way of articulating it. They speak in the vocabulary of the workplace, and they do so publicly. Even linguistically marginal members are placed in this relation to the employer's philosophy and language, and the challenging aspects of the task (see line 46) may cause them to interpret their movement toward unmarkedness as an accomplishment. Fowler (1985) calls such explicitly manipulative interaction 'directive linguistic practices.' Those more indirect, subtle influences of discourse which inculcate values and attitudes below members' level of self-reflexive awareness are labelled 'constitutive linguistic practices.' In the preceding excerpt, Maria is involved as recipient of at least two constitutive linguistic practices. First, she is shown the importance of trying. Career ambition and worker innovation are championed at the Major Hotel corporation, and to dare to accomplish something difficult, embarrassing or unusual is a valued trait. In this case, it is significant that what Maria is encouraged to try is to speak publicly in the employer's language and specific vocabulary. Thus the paradigm case of trying for Maria is to move toward the soft boundary of markedness. The second constitutive practice is evaluation. When she tries to speak the standard English of the workplace, Maria hears, 'you're doing very well.' Not only is a positive sanction given for her attempts to achieve unmarkedness, but the sanctioning act establishes who evaluates social performances and what constitutes acceptability. A high power member with unmarked discourse says that trying to speak the standard English of the institution is laudable; consequently, the paradigm case of doing well also is to move toward unmarked discourse. Conclusion This paper argues that it is speakers in virtual settings of interaction, in addition to languages per se, that reflect, identify, maintain, promote, inhibit, or discriminate against ethnolinguistic minority status. The praxis approach to identity demands that interaction be examined as research object as well as resource; as long as the topic of interest is the intersection of language, ethnicity, and identity, discourse must be seen as 'bespeaking itself' while it 'bespeaks self'. In the transcribed talk of members of the Major Hotel, it is apparent that nonstandard discourse is marked and there is institutional pressure for members to move from marked practices toward unmarked ones. While the evidence produced here is anecdotal and begs for widespread confirmation in other sources of member talk, it nevertheless captures the authentic, in situ responses of members to their social circumstances. Most importantly, a praxis approach intimately focuses attention on the personal histories in which the effects of achieving unmarkedness in discourse are played out. This sensitising aspect of discourse analytic work makes it an essential component of the effort to better understand the dynamics of ethnolinguistic identity.

< previous page

page_27

next page >

< previous page

page_28

next page > Page 28

Appendix 1 TR:MAJ: 12887 SD = Sales Director R1 = Researcher #1 R2 = Researcher #2 1 R1: We were just talkin' about the problems for a 2

person with a heavy accent in Sales.

3 SD: M hm/. 4 R1: Is it likely or possible that a person with a heavy 5

accent uhm would be advanced to at least upper or

6

middle levels of management like the executive

7

committee level or above?

8 SD: A:hm, let's see if we have anybody like that. Ahm 9

(.) It's probably possible, ya:h but=ah: I don't

10

know anybody like tha=oh, chefs/. They all have

11

accents.

12 R1: That's . . . 13

[

14 SD: But that's acceptable. 15

[

16 R2: Right. 17

[

18 R1: Yeah. 19 SD: I mean all chefs. Ah=th=they don't deal really with 20

the public so much, they're behind doors. Ahm (.)

21

st (.) Not that I know of, tha=I mean I, (.)oh

22

that's one thing when I ca:me ahm I used to live in

23

Hawaii, was born 'n' raised there. A:nd everybody

24

there speaks what we call pidgin English. 'N'

25

that's: (.) y'know you just don't enunciate/and =

26

=ahm y'know you throw some Hawaiian words in there

27

an' you throw some Japanese words in there and hh

28

only people there can understand that. And when I

29

came to //name of state// and worked here, I

30

decided that that wasn't gonna (.) help me in my

31

career. 'N' so then I, I really practiced at s=at

32

really enunciating 'n' saying the instead of duh,

33

and=ahm (.) a:nd not using the Hawaiian words =it

34

kinda pulls you away from your culture a bit. And

35

now that I go ba:ck, ahm and I speak like this,

36

this is to me: normal/and I can't (.) go back and

37

talk like that 'cause if I do and I try to, people

38

think, 'oh you=you're t=you're trying to fit in:

39

an: . . .

< previous page

page_28

next page >

< previous page

page_29

next page > Page 29

40 R2: It's not ah: //inaudible// 41 [ 42 SD: You're kinda in the middle. 43 R2: Yeah. 44 SD: 'Cause when [»] I go home 'n' I talk like this 45

they, they think that I'm trying t' be be better than

46

them, an' if I try ta have (.) the Hawaiian accent

47

or the local accent there ahm and it=it obviously

48

doesn't sound right anymore so that they=they think

49

that . . .

50 R2: Yeah. 51 R1: You've lost some skill in speaking the di=in the 52

Hawaiian . . .

53 SD: Yeah. 54 R1: Pidgin. 55 SD: I know the words an' all that, but . . . 56 R1: There is an accent. 57 SD: I f=I feel kinda foolish to say, y'know, (.)just 58

to not ta enunciate anymore.

59 R1: Yeah. (.) How long did it take you to make that 60

transition, so that you were speaking pretty,

61

y'know, standard \?

62

[

63 R2: Normal. 64 SD: Heh = heh hhh heh = hh. Hate t' say normal. Hhh ahm 65

(..) ahm, well=I was brought up in=in=ah a

66

community where there were predominantly white

67

people, so white people really don't talk (.) so

68

much the local dialect.

69

[

70 R1: Haole talk. 71 SD: Right. 72 R1: Heh = heh. 73 SD: So=ahm, I didn't have as much of a problem as maybe 74

someone else would, ahm, probably abou:t maybe a

75

year. = But I know someone that just started with

76

Major as Director of Sales at City Center, and she

77

has ahm=he has an accent. Ahm (.) And I think a=for

78

him it's gonna take a much longer ti:me. Because . . .

9

[

80 R1: He's 81

Hawaiian?

82 SD: Ahm he's: (.) h=he was born an'=I think and rai:sed 83

in Hawaii but he's Japane:se. But nevertheless he

84

has the accent. So=ahm, it might take him a little

85

bit longer because he's spent more of his life

< previous page

page_30

next page > Page 30

86

there than I = I left when I was: twenty-one years

87

old.

88 R2: Do you know if he was aware of that (.) something 89

//inaudible// you did change?

90 SD: I don't think so. He probably will learn that as he 91

gets into: working at the hotel and getting

92

exposed ta (.) Major people even, y'know.

93 R2: Yeah. 94 SD: They'll=they'll probably go, "whadja sa:y?" or they 95

make fun of you so you don't (.) you won't wanna

96

talk like that anymore.

97 R2: Yeah. Note 1. Notational conventions used in the transcripts are as follows: (a) Double slash marks, // //, contain transcriber comments; (b) single parentheses, (..), indicate speaker pauses; each period enclosed within the parentheses indicates a pause length of approximately one second; (c) brackets, [ ], indicate overlapping talk; (d) ellipsis at the end of a speaker turn, . . ., indicates end of speaker turn that does not contain an apparent end of the utterance; (e) colon, :, shows elongated sound; (f) underscoring, ______, indicates vocal emphasis; (g) falling slash, \, is used for falling vocal tone; (h) rising slash,/, is used for rising vocal tone; (i) equal sign, =, shows run-on speech with no pause; (j) right arrows, », indicate increased speech rate; (k) left arrows, «, indicate slowed speech rate; and (l) hhh indicates audible intake of breath. Sequential line numbers are included on the left margin of the transcript. Line numbers shown here refer to the transcript segment that is shown in its entirety in Appendix 1. The latter segment of transcript, concerning new employee orientation, is shown in its entirety within the body of the paper. References Banks. S. P. (1987) A critical analysis of ideology and discourse in two hotels. Unpublished Ph.D. dissertation. University of Southern California, Los Angeles, CA. Barth, F. (ed.) (1969) Ethnic Groups and Boundaries. Boston: Little, Brown and Company. Bond, M. H. (1983) How language variation affects inter-cultural differentiation of values by Hong Kong bilinguals, Journal of Language and Social Psychology, 2, 57-66. Bourhis, R. Y. (1979) Language and ethnic interaction. In H. Giles & B. Saint-Jacques (eds), Language and Ethnic Relations (pp. 117-42). Oxford: Pergamon Press. Bourhis, R. Y., Giles, H. and Rosenthal, D. (1981) Notes on the construction of a 'subjec-

< previous page

page_30

next page >

< previous page

page_31

next page > Page 31

tive vitality questionnaire' for ethnolinguistic groups, Journal of Multilingual and Multicultural Development, 2, 145-55. Brown, G. and Yule, G. (1983) Discourse Analysis. Cambridge: Cambridge University Press. Brown, P. and Levinson, S. (1979) Social structure, groups, and interaction. In K. R. Scherer & H. Giles (eds), Social Markers in Speech (pp. 291342). Cambridge: Cambridge University Press. Burgess, M. E. (1978) The resurgence of ethnicity: Myth or reality? Ethnic and Racial Studies, I, 265-85. Chock, P. P. and Wyman, J. R. (eds) (1986) Discourse and the Social Life of Meaning. Washington, D.C.: Smithsonian Institute Press. Comrie, B. (1976) Aspect: An Introduction to the Study of Verbal Aspect and Related Problems. Cambridge: Cambridge University Press. Crystal, D. (1986) Authentic what, who(m), and when? In J. A. Fishman, A. Tabouret-Keller, M. Clyne, Bh. Krishnamurti & M. Abdulaziz (eds), The Fergusonian Impact, Volume 2: Sociolinguistics and the Sociology of Language (pp. 411-18). New York: Mouton de Gruyter. Czyzewski, M. (1982) The problem of identity in modern sociology of interaction, Sociolinguistics Newsletter XIII, 2, 4-7. Deschamps, J-C. (1982) Social identity and relations of power between groups. In H. Tajfel (ed.), Social Identity and Intergroup Relations (pp. 8598). Cambridge: Cambridge University Press. Dittmar, N. and von Stutterheim, C. (1985) On the discourse of immigrant workers: Interethnic communication and communication strategies. In T. van Dijk (ed.). Handbook of Discourse Analysis, Volume 4: Discourse Analysis in Society. London: Academic Press. Edwards, J. (1979) Language and Disadvantage. New York: Elsevier. (1985) Language, Society, and Identity. Oxford: Basil Blackwell. Edwards, J. and Giles, H. (1984) Applications of the social psychology of language: Sociolinguistics and education. In P. Trudgill (ed.), Applied Sociolinguistics. London: Academic Press. Fishman, J. A. (ed.) (1986) The Rise and Fall of the Ethnic Revival: Perspectives on Language and Ethnicity. New York: Mouton. Fishman, J. A., Gertner, M. H., Lowy, E. G. and Milan, W. G. (1986). Ethnicity in action: The community resources of ethnic languages in the United States. In J. A. Fishman (ed.), The Rise and Fall of the Ethnic Revival: Perspectives on Language and Ethnicity, (pp. 195-282). New York: Mouton. Forester, T. (1981) The jelly bean people of Silicon Valley. In T. Forester (ed.), The Microelectronics Revolution (pp. 65-71). Cambridge, Mass: MIT Press. Fowler, R. (1985) Power. In T. van Dijk (ed.), Handbook of Discourse Analysis, Volume 4: Discourse Analysis in Society (pp. 61-82). London: Academic Press. Fowler, R., Hodge, B., Kress, G. and Trew, T. (1979) Language and Control. London: Routledge & Kegan Paul. Gans, H. J. (1979) Symbolic ethnicity: The future of ethnic groups and cultures in America, Ethnic and Racial Studies, 2, 1-20. Garfinkel, H. and Sacks, H. (1970) On formal structures of practical actions. In J. C. McKinney & E. A. Tiryakian (eds), Theoretical Sociology: Perspectives and Developments. New York: Appleton-Century-Crofts. Giles, H. (1979) Ethnicity markers in speech. In K. R. Scherer & H. Giles (eds), Social Markers in Speech (pp. 252-90). Cambridge: Cambridge University Press. Giles, H., Bourhis, R. Y. and Taylor, D. M. (1977) Towards a theory of language in ethnic group relations In H. Giles (ed.), Language, Ethnicity and Intergroup Relations (pp. 307-48). London: Academic Press. Giles, H. and Byrne, J. (1982) An intergroup approach to second language acquisition, Journal of Multilingual and Multicultural Development, 3, 17-40.

< previous page

page_31

next page >

< previous page

page_32

next page > Page 32

Giles, H. and Johnson, P. (1981) The role of language in ethnic group relations. In J. Turner & H. Giles (eds), Intergroup Behaviour (pp. 199-243). Chicago: University of Chicago Press. (1987) Ethnolinguistic identity theory: A social psychological approach to language maintenance, International Journal of the Sociology of Language, 68, 69-99. Giles, H., Mulac, A., Bradac, J. J. and Johnson, P. (1986) Speech acommodation theory: The next decade and beyond. In M. McLaughlin (ed.), Communication Yearbook 10 (pp. 13-48). Beverly Hills: Sage. Giles, H. and Powesland, P. (eds) (1975) Speech Style and Social Evaluation. London: Academic Press. Glazer, N. (1983) Ethnic Dilemmas 1964-1982. Cambridge, Mass: Harvard University Press. Gumperz, J. J. (ed.) (1982) Language and Social Identity. Cambridge: Cambridge University Press. Hadden, S. C. and Lester, M. (1978) Talking identity: The production of 'self' in interaction, Human Studies, I, 331-56. Haarmann, H. (1986) Language in Ethnicity: A View of Basic Ecological Relations. New York: Mouton de Gruyter. Hopper, R. and Williams, F. (1973) Speech characteristics and employability, Speech Monographs, 40, 296-302. Hymes, D. (1974) Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia: University of Pennsylvania Press. Jupp, T. C., Roberts, C. and Cook-Gumperz, J. (1982) Language and disadvantage: The hidden process. In J. J. Gumperz (ed.), Language and Social Identity (pp. 232-56). Cambridge: Cambridge University Press. Kalin, R. and Rayko, D. (1980) The social significance of speech in the job interview. In R. N. St. Clair & H. Giles (eds), The Social and Psychological Contexts of Language (pp. 39-50). Hillsdale, N J: Lawrence Erlbaum. Kantor, R. M. and Stein, B. A. (eds) (1979) Life in Organizations: Workplaces as People Experience them. New York: Basic Books. Khubchandani, L. M. (1983) Plural Languages, Plural Cultures. The East-West Center: University of Hawaii Press. Kramarae, C., Schulz, M. and O'Barr, W. M. (1984) Language and Power. Beverly Hills: Sage. Lachapelle, R. and Henripin, J. (1986) The Demolinguistic Situation in Canada: Past Trends and Future Prospects (trans. D. A. Mark). Montreal: The Institute for Research on Public Policy/L'Institut de recherches politiques. Levinson, S. (1983) Pragmatics. Cambridge: Cambridge University Press. Liebkin, K. (1982) The Swedish-speaking Finns: A case study of ethnolinguistic identity. In H. Tajfel (ed.), Social Identity and Intergroup Relations (pp. 322-367). Cambridge: Cambridge University Press. Martinelli, P. (1986) A test of the McKayand Lewins ethnic typology, Ethnic and Racial Studies, 9, 196-210. Martinet, A. (1986) The dynamics of plurilingual situations. In J. A. Fishman, A. Tabouret-Keller, M. Clyne, Bh. Krishnamurti & M. Abdulaziz (eds), The Fergusonian Impact, Volume 2: Sociolinguistics and the Sociology of Language (pp. 245-52). New York: Mouton de Gruyter. Norkunas, M. K. (1987) Women, work and ethnic identity: Personal narratives and the ethnic enclave in the textile city of Lowell, Massachusetts, The Journal of Ethnic Studies, 15, 27-48. Perrow, C. (1986) Complex Organizations: A Critical Essay (third edition). New York: Random House. Potter, J. and Wetherall, M. (1987) Discourse and Social Psychology. London: Sage. Rogers, E. M. and Rogers, R. (1976) Communicating in Organizations. New York: The Free Press.

< previous page

page_32

next page >

< previous page

page_33

next page > Page 33

Ross, J. A. (1979) Language and the mobilization of ethnic identity. In H. Giles & B. Saint-Jacques (eds), Language and Ethnic Relations (pp. 1-13). Oxford: Pergamon Press. Ryan, E. B. and Giles, H. (eds.) (1983) Attitudes toward Language Variation. London: Edward Arnold. 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, 115-36. Scherer, K. & Giles, H. (eds) (1979) Social Markers in Speech. Cambridge: Cambridge University Press. Schiffrin, D. (1987) Discourse Markers. Cambridge: Cambridge University Press. Shukla, H. L. (1985) Language, Ethnicity and History. Delhi: B. R. Publishing Corporation. Simon, H. A. (1981) What computers mean for man and society. In T. Forester (ed.), The Microelectronics Revolution (pp. 419-33). Cambridge, MA.: The MIT Press. Smith, S. J. (1984) Negotiating ethnicity in an uncertain environment, Ethnic and Racial Studies, 7, 360-72. Tajfel, H. (1978) Social categorization, social identity, and social comparison. In H. Tajfel (ed.), Differentiation between Social Groups. London: Academic Press. Tajfel, H. (ed.) (1981) Human Groups and Social Categories. Cambridge: Cambridge University Press. Trudgill, P. (1983) On Dialect. New York: New York University Press. Turner, J. C. (1982) Towards a cognitive redefinition of the social group. In H. Tajfel (ed.), Social Identity and Intergroup Relations (pp. 15-40). Cambridge: Cambridge University Press. Williams, G. (1979) Language group allegiance and ethnic interaction. In H. Giles & B. Saint-Jacques (eds), Language and Ethnic Relations (pp. 5765). New York: Pergamon Press.

< previous page

page_33

next page >

< previous page

page_34

next page > Page 34

< previous page

page_34

next page >

< previous page

page_35

next page > Page 35

Social Mobility and Language Use in an American Company in Japan 1 Patricia M. SanAntonio Department of Anthropology, Arizona State University Abstract This paper is based on a year of ethnographic research in an American company in Japan. The company, a high-tech computer firm, has a policy of requiring English to be used in all overseas offices. Like most other foreign companies in Japan, this company has difficulty attracting high quality Japanese employees. They need to hire Japanese managers who have both business experience and English language skills. At the same time the company attempts to compete with the Japanese business community by hiring new college graduates, overlooking the English language requirement. As a result, there is considerable variation in the English language abilities of Japanese employees. This causes problems because the American managers insist on the use of English for meetings, classes, and any interactions involving foreigners. English facility and the ability to deal with Americans socially becomes a source of power for English proficient Japanese employees. The use of English and Japanese in the company becomes an important, means of restricting access to information as well as a source of power and advancement to a subset of employees. Ethnographic examples of language use in the company as it relates to the issue of power brokerage and employee advancement in the company are presented in the paper. American companies are faced with a problem when establishing a branch in Japan of deciding what to do about language use in the firm. There are three options available: to use the indigenous language only, to require all employees to use the foreign language, or to have a policy that both languages are spoken. There are problems in communication with any of these three choices (Holden, 1985). This is especially true in Japan, where the Japanese language and its use play a vital role in Japanese identity both within Japanese society and in the relationship of Japanese to foreigners (Miller, 1977). This paper focuses on the implications of one such policy choice for the Japanese employees of an American firm in Japan. The use of Japanese/English as a salient cultural marker and language choice as a strategy for personal advancement and control is discussed. The analysis is based on ethnographic research carried out in one American company in Japan.

< previous page

page_35

next page >

< previous page

page_36

next page > Page 36

Background Japanese and American Attitudes to Language Use Language is an important aspect of ethnic identity for both Japanese and Americans (Miller, 1977). In Japan there is a strong relationship between being Japanese and speaking Japanese (Loveday, 1982, 1986). The importance of the Japanese language as an identity marker is illustrated both by the hostility and problems met by the foreigner who speaks Japanese with some fluency. Since speaking Japanese is linked closely with ethnic identity, when a foreigner speaks Japanese well, the ethnic difference between foreigner and Japanese is reduced to the 'Asian-ness' of the Japanese (Miller, 1977). This creates problems because the Japanese do not base their identity on a notion of fellowship with other Asians, but consider themselves unique (Miller, 1977). The Japanese cultural hierarchy and identity vis à vis other Asian ethnic groups is threatened. Ira foreigner speaks Japanese well, then the cultural attributes which enable an individual to claim Japanese identity are emphasised making it difficult for Japanese to retain the view of themselves as unique. Americans too, believe that there is a close relationship between the use of English and identity as an American. The current controversy in the Southwest over English as the official language and proposed legislation to forbid the official use of Spanish, is one example of this attitude. Americans associate bilingual or multilingual individuals with low status groups because of their immigrant past (Wardhaugh, 1986). As in Japan, Americans who speak a foreign language well in certain situations may be suspect as to their loyalty or motives. In one Harvard Business School case, an American manager who had 'gone native', learning to speak Japanese well, was fired because his superior felt that Japanese language ability kept the employee from acting in the best interests of the American company. When the American spoke Japanese he was adopting the speech markers of the outgroup, threatening his colleagues' American ethnic identity and the sense of loyalty to the dominant ethnic group (Taylor, Meynard & Rheault, 1977). There has been much research concerning the cultural differences between Japanese and Americans (Goldstein & Tamura, 1975; Omens, Jenner & Beatty, 1987; Ramsey & Birk, 1983). In spite of the many differences between their cultures, Japanese and Americans share similar attitudes to language use and the relationship of language to ethnic identity. Ethnic Identity and Language Use The study of ethnicity has proven invaluable in understanding how people organise themselves in a heterogeneous, complex environment. Studies have dealt with the dynamic, interactional nature of ethnicity and have shown that behaviour across ethnic group boundaries are theoretically the most interesting (Barth, 1969; Cohen, 1974; Leach, 1954; LeVine & Campbell, 1972).

< previous page

page_36

next page >

< previous page

page_37

next page > Page 37

The focus of this paper is the instrumental use of ethnicity in competition for access to resources. Language use is a tool in the manipulation of identity on the individual level, important as a political strategy (Cohen, 1974). The use of English in the Japanese context becomes a bid for access to scarce resources, power for Japanese employees in an American company. Resources include access to American managers, information, and promotion and approval in the American corporation. In the Japanese context, where language and identity are inextricably linked, the willingness of Japanese employees to use English in a range of situations becomes a marker of their desire for inclusion, or integration into the American corporate group. Second language learning is difficult in Japan because of the strong identification between Japanese language and identity (Hall & Gudykunst, 1986; Nisugi, 1974). Speaking English in various situations in the company is tantamount to making a claim to the identity of a Japanese with whom Americans can work. The Company's Language Policy The American company studied has a very strong corporate identity. Founded in the late 1960s during the explosion of firms in Silicon Valley, the company culture was formulated by engineers, and designed with the needs of a high tech industry in mind. The company culture has American values and this 'American-ness' is considered by its employees to be the foundation of the company's continued technological success. When expanding to overseas locations, the corporation made a decision to preserve the 'American' culture orientation on a worldwide basis because company values were considered necessary to successfully carrying out their work. Accordingly, the company chose an explicit English-only language policy. It was required for Japanese employees to use English when speaking with foreigners in Japan, the United States based corporate headquarters, and the offshore factories. American managers would often state 'Company people speak English.' Given the English-only language policy, the American company needs to hire Japanese employees who are bilingual in Japanese and English. However, this policy adds to already existing constraints on the company in hiring Japanese employees. First, the supply of bilingual Japanese is small, and the number of Japanese willing to work for a foreign company is even smaller. Second, the American company, in trying to compete with Japanese companies in the Japanese marketplace, has been attempting to hire the kind of new college graduates whom Japanese companies also recruit. These constraints on hiring result in a mix of Japanese employees actually being hired by the American company in Japan. There is a considerable variation in English language competency, experience with the American culture, and interest in working with foreigners among the Japanese employees. The American expatriates chosen to work in the corporation's Japan office also reflect the company's American value orientation and English only language

< previous page

page_37

next page >

< previous page

page_38

next page > Page 38

policy. Paralleling the situation among Japanese employees, there are two kinds of foreigners sent to Japan from the United States. The first, and most common type of American employee has no particular interest or experience in Japan. Such employees come to Japan out of curiosity, or interest in the extra pay given to expatriates. The second type of employee is more rare and is not preferred by the American company. These are Americans who have had experience in Japan as students or missionaries, for example, or who are married to Japanese wives. These employees are often able to speak Japanese or are very interested in learning Japanese. Therefore both American expatriates in the company in Japan, and Japanese employees vary considerably in their backgrounds, interest, and commitment to working in the Japanese/American environment, as well as in their Japanese-English language ability. Officially there is no variation; all employees speak English. Unofficially, language aptitude varies with several foreigners being able to communicate in Japanese and many Japanese employees unable to communicate in English. In this environment, the gap between the ideal and the real situation makes language ability a salient feature of interaction in the company. In this context, language use is a marker which can be manipulated situationally by an employee. Language Use in the Company The American company in Japan affords many possibilities for interaction between Japanese and Americans. I will concentrate on two general situations meetings and interaction with visiting foreignersthe types of interactions which first drew my attention to the language problem. Meetings Meetings provide an interesting example of language behaviour. If a foreigner is present, English must be spoken. Several American managers told me that English was always spoken at meetings, even when only Japanese were present. This assumption reflects the American belief that there is a natural relationship between the corporation's business and the use of English. During meetings attended by only Japanese, however, Japanese is spoken and employees are animated and talkative. Meetings with foreigners in attendance are noticeably quieter. Compared with the Japanese managers, foreigners in general tend to conduct meetings in a very formal, controlling style with an emphasis on turn taking and concise answers. The foreign managers are of high status within the company and clearly are deferred to by the Japanese. In upper level staff meetings all of the Japanese managers speak English well and meetings tend to be quiet with orderly turn taking and short clear answers to the questions posed by the foreign manager. If all of the Japanese attending have approximately the same English language skills then language is not at issue. The

< previous page

page_38

next page >

< previous page

page_39

next page > Page 39

situation changes and language becomes a salient marker when Japanese employees with varying skills are present at the same meeting. Japanese often are hesitant to speak English in front of other Japanese, are conscious of differences in English speaking ability, and hesitant to volunteer information or opinions (Hildebrandt & Giles, 1980). Usually one employee dominates the meeting, speaking directly to the foreign manager, volunteering information or restating what another Japanese employee has said. The Japanese who deal aggressively with foreigners in meetings must have more than English language ability, they must also have the cultural knowledge to interact with the foreign managers in an appropriate way. Japanese employees who have been trained abroad have access to the culturally correct use of language such as joking, use of first names, and the use of colloquial American speech which American managers interpret as understanding, aggressiveness, and interest. When English speaking Japanese are in a meeting with Japanese who speak very poor English, the language situation is complicated by the language policy. American managers expect all Japanese attending the meeting to speak English and become angry if questions are not answered quickly and clearly. Managers interpret silences as ignorance or evasiveness on the part of the employees. An English speaking, culturally competent Japanese employee can control such a meeting. The non-English speaking Japanese are grateful to English speaking Japanese for dealing with the American and keeping attention focused away from them. The American manager is grateful to the English speaking employee for facilitating at the meeting and clarifying the situation. The Japanese often fall quiet and allow one Japanese employee to do most, if not all the talking, speaking only when forced to by a direct question from the American manager. As a result of the language policy there is no translating at meetings. If a Japanese employee has difficulty expressing an idea, another Japanese will not help him overtly because this would draw attention to his lack of English ability. Either the idea is never clarified or, more likely, an English speaking Japanese will speak for the other employee in order to keep the meeting going smoothly. A Japanese employee who speaks English well in such a situation becomes a mediator. He interprets the business situation for the American, expressing his idea of the concerns and position of other Japanese employees. He may be one of the few to understand directives and the opinion of the Americans and he interprets American speech and behaviour for other Japanese employees. A culturally competent English speaking Japanese employee claims an identity as a Japanese employee who strives to be part of the American corporate group. He does this by his aggressive, mediating use of English in a meeting. His convergence to the American speaking style is a speech marker signalling a desire to be integrated into the American corporate group (Giles, Bourhis & Taylor, 1977). Other Japanese are constrained by the language policy not to challenge his claim to business knowledge or interest. To avoid censure for their lack of language skills it is in the interest of many Japanese employees to allow mediation between themselves and foreign managers. The foreign manager accepts the claim because

< previous page

page_39

next page >

< previous page

page_40

next page > Page 40

it facilitates the conduct of the meeting and gives him feedback from the Japanese employee couched in a language and style he can understand and evaluates positively. The claim is accepted and the Japanese employee who acts as a mediator is judged by the American manager to be intelligent, ambitious, and interested. When opportunities arise for promotion or training, American managers often recommend the mediating type of employee because of their experience in interacting with him and their judgement of his abilities. Other Japanese employees are not resentful, but are grateful for the mediation and admire his expertise in dealing with foreigners. This lack of resentment comes from the fact that other Japanese employees consider the English-fluent Japanese to have a foreign identity. This shift in identity is illustrated clearly when Japanese employees use first names when referring to the English-fluent Japanese, an exclusively foreign style of address. This is an example of how language use influences the judgement of ethnic identity (Fitch & Hopper, 1983). Occasionally, Japanese employees will compete for the attention of American managers. At meetings this may take the form of one Japanese employee translating from English to Japanese or correcting vocabulary or grammar for the supposed benefit of another Japanese employee. This only occurs between Japanese with approximately the same English ability. When two competing Japanese employees are equal in English ability, American cultural knowledge assumes importance. English fluent Japanese who have no foreign experience and cannot successfully compete with Japanese employees who do, often reject making a claim to membership in the American group and turn instead to a different strategy. They reject comparison with the other type of Japanese by rejecting American values openly and emphasising their Japanese-ness. This nationalistic strategy often took the form of stating explicitly to both foreigners and Japanese in the company that they did not like to speak English, socialise with foreigners, or have any interest in going to the United States. This was one way of maintaining their ethnic distinctiveness, gaining the respect of aggressive, individualistic American managers, and presenting themselves as 'authentic' Japanese. Both types of Japanese are mutually critical of each other. Interaction with Visiting Foreigners The American corporate headquarters sends a number of foreigners to Japan on business on a regular basis. During an important marketing project for a new product, an American technical expert came to Japan from California. The American, a well-known corporation manager, was scheduled to speak on technical matters in the Japan office and with the representatives of Japanese companies and distributors. This senior manager with high level technical expertise and a knowledge of the current situation in the American office spoke no Japanese. As an upper level manager, the American would usually interact with only the most senior Japanese managers, but in this instance he needed to visit outside the company and upper

< previous page

page_40

next page >

< previous page

page_41

next page > Page 41

level Japanese managers could not be spared from their jobs. A young English-fluent Japanese customer engineer was assigned to escort the American. The Japanese engineer had been trained in the United States and spoke colloquial English. Through his role as interpreter, the Japanese employee was involved in higher level meetings than he would ordinarily attend. The Japanese employee also was able to spend time with the upper level American and in subsequent meetings and interactions, repeated company and product gossip he had heard from the American referring to people, as the American had, by first names. The Japanese and American also used first names with each other, an indication on the part of the Japanese of his acceptance of American interactional style. While escorting the American in the company the Japanese employee had a freer schedule and more autonomy, a sign of high status in an American work situation (Jackall, 1978). The American technical expert was scheduled to speak at a Marketing Staff meeting. The meeting was run by a foreigner, held in English, and was for the upper level business staff. The Japanese customer engineer was not, as a rule invited to attend this meeting. During the American's visit, one of his jobs was to get the American to various meetings. During the Marketing Staff meeting, the Japanese engineer interrupted the meeting and in English, using the American's first name asked the foreign manager if 'you're ready for him yet'. The foreign manager was annoyed and said no curtly, but the Japanese persisted in asking whether he should 'send the guy in' he made it clear that he was in control of the American's schedule and should be informed of the American's commitments. The young Japanese marketing engineer was also in an enviable position vis à vis other Japanese employees in the firm. The American manager was scheduled to give a general presentation to Japanese employees of the firm. The Japanese customer engineer escorted him into the meeting room. They had obviously eaten together at MacDonald's and the Japanese employee was acting very casual, lounging in a chair at the front of the meeting room finishing his coke. As it became time for the lecture to begin the Japanese engineer stood up and introduced the American in Japanese. Occasionally during the presentation he would translate into Japanese, controlling the flow of the presentation. The translation was not necessary because the Japanese engineers were chosen to attend the lecture because they could speak English. In front of the other Japanese employees, the Japanese marketing engineer assumed a mediator position between them and an American manager who controlled valuable technical information. After the American returned to the United States, the Japanese engineer was invited to give insights into California policy to the Americans at staff meetings. The cultural and English ability of this employee gave him access to valuable technical and corporate information. During the American's visit, the Japanese employee stressed his interpreter position through language use as a way of making his claim to American corporate identity. Although this was a visit of unusual importance, there are enough American visitors to make this type of interaction an important arena of language choice and opportunity for Japanese employees. Each department received foreign visitors on a regular basis. In one department, the Japanese employees were reluctant to socialise with foreigners and

< previous page

page_41

next page >

< previous page

page_42

next page > Page 42

unable to speak English easily. The Japanese employees took the foreign manager out to dinner, but lapsed quickly into Japanese which infuriated the American manager. Angry at what she considered to be purposeful exclusion, the American spent subsequent visits socialising with the Japanese employees in another department, all of whom were English fluent and socially comfortable in dealing with foreigners. As a result of this interaction, one of the English fluent Japanese was promoted on the recommendation of the American manager. Some of the foreign educated Japanese may seek out one on one interaction with Americans and it is these Japanese whose abilities and interests are known by the Americans. In group interactions, one Japanese who is more aggressive in English dominates the conversation. Lunch groups in the company cafeteria are segregated sexually, as well as by nationality and status. In order to eat with a foreigner, a Japanese employee had to be willing to speak English, another barrier to casual personal contact with Americans. It is interesting to note that English-fluent Japanese generally do not associate with Japanese-fluent Americans. There are several reasons for this. First, English-fluent Japanese prefer to control an interaction with an American by speaking English. If they speak Japanese with an American their special status arising from their bilingualism is neutralised in the face of the American's competence in Japanese. Secondly, English-fluent Japanese are sensitive to American cultural attitudes and realise that Japanese language ability is not valued by their American superiors. Speaking Japanese is a speech marker of the outgroup. Japanese who are making a claim to integrativeness in the American corporate group avoid the use of a speech marker which contradicts this claim. With the limited opportunity available to Japanese in an American company, language use becomes a salient marker and a tool in the competition for resources. The official English policy reduces the conflict between Japanese employees of differing English abilities. Conclusion There is disagreement within the company about whether or not the language policy works. Certainly the policy creates many problems that it could be argued, negatively affect the overall efficiency of the company. It can be argued, however, that the corporation's 'English-only' policy maintains the boundary between Japanese and American identity. This policy maintains a boundary and a clear ethnic identity between Japanese and American employees with authority clearly kept in the hands of Americans. American expatriates are faced with a confusing, complicated environment. The Japanese environment is threatening to the foreigner because in Japan, unlike in other foreign countries, Americans do not have the upper hand. It is important then, that the foreign employee feel in control of a business situation especially in this company where control and aggressiveness are major components of their corporate culture. The English only language policy restricts and controls information. Official information must be disseminated in English all foreigners have access, and certain Japanese have access. American managers often refuse to allow memos to

< previous page

page_42

next page >

< previous page

page_43

next page > Page 43

be translated into Japanese. The issue is not one of efficiency and language comprehension, but one of control. Because only select Japanese within any given department can communicate easily in English, the number of Japanese whom foreigners deal with closely is greatly reduced. Those Japanese who communicate easily and well with Americans deal with the Americans. Such Japanese may or may not be in the direct hierarchical line of communication, but they communicate with Americans by virtue of their language and social skills. The English policy circumvents the indigenous Japanese hierarchy and creates its own American communication-based hierarchy. As a result of legislating language use, social structural complexity is reduced. The input from the Japanese is reduced, and when it occurs it is funnelled through the English proficient Japanese. We can see the benefits for Americans in adhering to an English only language policy. It is in the interest of Americans to retain the English only language policy as a way of maintaining control. Given this situation, there is pressure from foreigners to maintain the policy and to adhere to the English only rule. It is possible for English-fluent Japanese in a situation where many employees cannot communicate easily in English, to use their supposedly universal ability to further their own ends. Input is reduced because a Japanese must communicate in English with a foreigner, and if they cannot do this they do not communicate at all. The status and role of a Japanese employee in the American company evolves from language ability in interaction. Besides the full complement of roles in the company, this interethnic situation creates a number of additional roles, which a Japanese employee can fill, using as a stepping stone to higher positions in the company. Language based roles are not formally designated, but arise through the need of Americans to communicate with the Japanese employees. Language based roles include that of interpreter and mediator. Aggressive Japanese employees may fill these roles in special situations and use this as a way to get noticed and positively evaluated by their foreign superiors. According to Giles & Byrne's (1982) intergroup theory of second language ability, Japanese would have difficulty in second language acquisition. Japanese difficulties in learning, and embarrassment in using a foreign language have been noted in the literature (Hildebrandt & Giles, 1980). There is a strong relationship between Japanese ethnic identity and language, strict ingroup boundaries, and a strong sense of Japanese uniqueness and cultural superiority. Speaking English in Japan threatens Japanese ethnic identity. A Japanese employee is making a choice in favour of a foreign identity by using English and converging to American speech and interaction style. For a subset of Japanese employees the company's English only language policy provides an arena in which language use becomes a salient marker of ethnic identity, and the manipulation of language abilities becomes an important tool in a bid for access to scarce resources. Note 1. The research reported in this paper was conducted in 1986-87 on a Japan Foundation dissertation grant.

< previous page

page_43

next page >

< previous page

page_44

next page > Page 44

References Barth, F. (1969) Ethnic Groups and Boundaries. Oslo: Universitetsforlaget. Cohen, A. (1974) Urban Ethnicity. London: Tavistock Publications. Fitch, K. and Hopper, R. (1983) If you speak Spanish they'll think you're a German: Attitudes toward language choice in multilingual environments, Journal of Multilingual and Multicultural Development, 4, 115-27. Giles, H., Bourhis R. Y. and Taylor, D. M. (1977) Towards a theory of language in ethnic group relations. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Giles, H. and Byrne, J. (1982) An intergroup approach to second language acquisition, Journal of Multilingual and Multicultural Development, 3, 17-40. Goldstein, B. and Tamura, K. (1975) Japan and America. Rutland: Charles E. Tuttle Company. Hall, B. J. and Gudykunst, W. B. (1986) The intergroup theory of second language ability. Journal of Language and Social Psychology, 5, 291-301. Hildebrandt, N. and Giles, H. (1980) The English language in Japan: A social psychological perspective, JALT Journal, 2, 63-87. Holden. N. (1985) The communication gap: How Western engineers experience and react to Japanese language and society, Multilingua, 4-1, 19-25. Jackall, R. (1978) Workers in a Labyrinth: Jobs and Survival in a Bank Bureaucracy. New York: Universe Books. Leach, E. R. (1954) Political Systems of Highland Burma. Cambridge, MA: Harvard University Press. LeVine, R. A. and Campbell. B. (1972) The Problem of Ethnic Boundaries in Ethno-Centrism. New York: John Wiley & Sons. Loveday, L. (1982) The Sociolinguistics of Learning and Using a Non-native Language. Oxford: Pergamon Press. (1986) Explorations in Japanese Sociolinguistics. Amsterdam: John Benjamins Publishing Company. Miller, R. A. (1977) The Japanese Language in Contemporary Japan. Washington D.C.: American Enterprise Institute for Public Policy Research. Nisugi, M. (1974) Images of spoken Japanese and spoken English. In J. Condon & M. Sato (eds), Intercultural Encounters with Japan. Tokyo: Simul Press. Omens, A. E., Jenner, S. R. and Beatty, J. R. (1987) Intercultural perceptions in United States subsidiaries of Japanese companies, International Journal of Intercultural Relations, 11, 249-64. Ramsey, S. and Birk, J. (1983) Preparation of North Americans for interaction with Japanese: Considerations of language and communication style. In D. Landis & R. W. Brislin (eds), Handbook of Intercultural Training, Vol. III. New York: Pergamon Press. Taylor, D. M., Meynard, R. and Rheault, E. (1977) Threat to ethnic identity and second language learning. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Wardhaugh, R. (1986) An Introduction to Sociolinguistics. Oxford: Basil Blackwell, Inc.

< previous page

page_44

next page >

< previous page

page_45

next page > Page 45

Language Attitudes, Frames of Reference, and Social Identity: A Scottish Dimension 1 Dominic Abrams University of Dundee and Michael A. Hogg University of Melbourne Abstract It was predicted from social identity and self-categorisation theories that ingroup speakers should attract more positive evaluations than outgroup speakers, especially when both are of equal status. However, the ingroup may be redefined by changes in the levels of categorisation salient to the perceiver. In Scotland, both intra-national and international levels are relevant at various times. Using the matched guise technique, in three conditions each possible pairing of middle class versions of Dundee, Glasgow (both Scottish) and RP English accents were presented to teenagers from Dundee. On ratings of speaker status, likely employment, and solidarity there was a clear pattern of ingroup favouritism. As predicted, the ingroup accents were evaluated positively and out-group negatively. Thus, Glasgow accents were evaluated negatively when contrasted with Dundee accents, but positively when contrasted with RP. Moreover, levels of ingroup favouritism correlated positively with measures of identification with Scotland. The data suggest language attitudes are susceptible to considerable variation, depending on the level of self-categorisation salient to the perceiver. Implications for issues of ethnolinguistic vitality and for the similarity attraction hypothesis are discussed. Hogg, Joyce & Abrams' (1984) study of language attitudes towards High German and Swiss German in Switzerland revealed that where language varieties do not differ in prestige, attitudes can be predicted to result from a social identity dynamic. Specifically, speakers will attract more favourable evaluations if they are perceived as ingroup members than as outgroup members. Thus, for example, regional accented speakers are perceived as more likeable and attractive (high solidarity) than are standard speakers (see Giles & Powesland, 1975). However, this pattern of evaluations is affected by a number of other variables, notably the context in which speech occurs, and the perceived socio-economic status of the speakers. Using English and Mexican-accented American, and RP and Welsh-accented English, Ryan & Carranza (1975) and Creber & Giles (1983), respec-

< previous page

page_45

next page >

< previous page

page_46

next page > Page 46

tively, found that a standard speaker is upgraded vis à vis a non-standard speaker on status variables, especially in a formal context. However, as Ryan & Sebastian (1980) have observed, the use of standard and non-standard accents often is correlated subjectively with social class. When listeners are informed that a non-standard speaker is in fact middle class, their evaluations on status measures are markedly more favourable. This does raise the question of whether Creber & Giles' (1983) and Ryan & Carranza's (1975) findings are due to class/formality or to accent/formality interactions. Giles & Sassoon (1983) orthogonally manipulated message formality, speaker's accent (Standard versus Cockney English), and speaker's social class label (middle vs. working). They found only two main effects on status ratings. Middle class and RP speakers were rated as having higher status than working class and Cockney speakers, respectively. One problem with both the Ryan & Sebastian (1980) and the Giles & Sassoon (1983) studies is that the class manipulation was simply based on descriptive background information about the speaker. This manipulation does not remove the possibility that the 'middle class' non-standard speaker is perceived as being peculiar, or deviant from middle class norms (cf. McKirnan & Hamayan, 1984), and somehow in-genuine. An alternative method of eliminating the confounding with perceived class is to present different middle class accents, one of which is RP. In this situation, when a non-RP accented listener hears his or her own middle-class variety, it should receive favourable evaluations relative to RP, given that it denotes equivalent real status, and also ingroup membership of the speaker (but see Edwards, 1982 for conflicting evidence). Thus, there should be ample scope for a social identity dynamic to come into play. This hypothesis is drawn from Tajfel & Turner's (1979) analysis of the macro-social aspects of social identity which states that while legitimate status differences inhibit ingroup bias, in the absence of such differences (and of immediate power imbalances; cf. Sachdev & Bourhis, 1984), individuals will display ingroup biases as these serve to maintain positive social identity by positively differentiating the ingroup from outgroups. Scotland may be considered high on objective vitality according to the criteria set out by Giles, Bourhis & Taylor (1977) who conceive of ethnolinguistic vitality as being determined by the institutional support, status, and demography of an ethnolinguistic group. Scotland, although part of the United Kingdom, is separated from England geographically, historically, and institutionally. It has its own legal and education systems, its own church, independent television stations (with Scottish accented news presenters), and formerly its own monarch. In recent years it has become the major UK base for oil exploration and drilling in the North Sea, the fruits of which have increased the wealth of Britain as a whole. In the 1970s the Scottish Nationalist Party enjoyed great surges in popularity, again partly as a result of Scotland's new found status and prestige as the home of Britain's oil riches. The high objective vitality of Scotland could plausibly be reflected in high subjective vitality (people's perceptions of the objective vitality of their group; see Johnson, Giles & Bourhis, 1973), and may partly be responsible for discontinuities in accent evaluations, both within and outside of Scotland. While English

< previous page

page_46

next page >

< previous page

page_47

next page > Page 47

people's perceptions of dialect prestige within England appears to reduce the further the variety is based from London, Scottish accents are accorded greater prestige than their northern English neighbours (Brook, 1963). A study by Cheyne (1970), carried out before the oil production was substantial, revealed that Scottish subjects tended to rate Scottish regional accents lower on status, but higher on solidarity than English regional accents. Unfortunately, Cheyne did not say which accents were presented, and only noted a possible confounding of perceived nationality and perceived social class. It is likely, however, that Scottish subjective vitality is currently somewhat higher than it was in the mid 1960s, when Cheyne conducted his study. This was manifested behaviourally during the 1987 general election, in which the Labour party claimed a 'mandate for Scotland', after a landslide defeat of the Conservatives in Scotland (but the reverse in England). Moreover, a recent large scale survey of 16-19 year olds from Fife (adjacent to Dundee) revealed a very high level of identification with Scotland, and a perception that advantages held by people living in England were unfair or illegitimate (Abrams, Emler, Jamieson & Martens, 1987). We therefore predict that when Scottish listeners evaluate Scottish and English standard accents, the former will be rated more highly on both solidarity and status measures. Moreover, since the psychological dynamic is a desire for positive social identity, we predict that ingroup favouring ratings will correlate positively with feelings of identification with Scotland, and with separatist, vitality preserving attitudes (cf. Bourhis, Giles & Rosenthal, 1981; Bourhis & Sachdev, 1984). In addition to the subjective vitality of Scottish people, and our prediction that Scottish subjects will favour Scottish over English accents in an equal status comparison (i.e. all accent varieties are middle-class versions), is a further issue pertinent to our social identity predictions. Although not addressed in Cheyne's study, it is likely that 'regional' Scottish accents each attract different evaluations within Scotland, just as different regional English accents do in England (see Giles & Powesland, 1975). This raises the interesting question of how a listener from Scottish area 'XA' will regard a speaker from Scottish area XA, first in comparison with an 'XB' speaker and second in comparison to a speaker from England, 'Y'. There are at least two possibilities. A simple similarity-attraction prediction is that XA will be rated more favourably than XB, and XB than Y (see Byrne, 1971). Similar predictions would be derived from an assumption that language attitudes are developed historically, relatively stable and enduring aspects of a person's ethnicity (cf. Giles et al., 1981). Both analyses would assume that, for any particular evaluative dimension, people have in mind a mental scale, with different language varieties placed at different points on the scale. Social identity theory, with its emphasis on the active process of differentiation between categories, predicts simply that people strive to evaluate ingroups more positively than outgroups (see Hogg & Abrams, 1988). Furthermore, the definition of which groups are ingroups or outgroups is not only determined by objective intergroup relations, but is also a subjectively conceived relationship based on the contrasts that are most salient and meaningful in any particular situation (the same assumptions form the basis of Turner, Hogg, Oakes, Reicher & Wetherell's, 1987, self-categorisation theory). That is, a sub-

< previous page

page_47

next page >

< previous page

page_48

next page > Page 48

jective frame of reference is imposed on person stimuli, within which others are divided into ingroup and outgroup category members. To return to our example, when groups XA and XB are salient, the ingroup is XA, and the listener will subjectively focus on the A/B distinction and exaggerate the differences between XA and XB, usually favouring XA. When the contrast is between XB and Y, the ingroup becomes XB (this is now a category which includes the self more than does the other category), with a particular focus on the X/Y distinction, and hence its position shifts towards the positive end of the scale, while Y is at the negative end. Thus, as the subjective frame of reference shifts, so attitudes to any particular object or group shift also. Apart from providing a more parsimonious explanation of attitude shifts, and a more plausible model than the absolute judgement model (implying the possibility of an infinite number of positions on either a huge scale or one with tiny increments), this prediction has been borne out in numerous experiments on the effects of social categorisation on conformity, attitude polarisation, and norm formation (e.g. Abrams, Wetherell, Cochrane, Hogg & Turner, in submission; Eiser & Stroebe, 1972; Hogg & Turner, 1987). The present study compares the absolute and relative judgement models using a matched guise procedure (Bourhis, Giles & Tajfel, 1963; Lambert, Hodgson, Gardner & Fillenbaum, 1960) in which school pupils from Dundee, Scotland, listen to middle class accents of two varieties. There are three contrast conditions. In the first, subjects hear two Scottish accents (Dundee and Glasgow), in the second, Dundee is paired with English RP, while in the third, Glasgow is paired with English RP. Our general prediction is that intergroup differentiation will occur in all conditions, so that Dundee is always rated most, and RP least favourably, but that ratings of Glasgow will be more favourable when compared to RP than when compared to Dundee. In order to investigate the effects of class and formality (as partial replication of studies reported above), subjects were drawn from two schools, one a middle class private school, and the other a working class state school. As pointed out by Giles & Ryan (1982), the majority of studies conducted in the UK have used southern middle class undergraduate students as subjects. Younger people are less likely to accord higher status to RP (Giles, 1970; Giles, Harrison, Smith, Creber & Freeman, 1983). When the experimenter does not speak with an RP accent, and the context is less formal, status evaluations of RP are even less likely to exceed those of local varieties (Creber & Giles, 1983; Price, Fluck & Giles, 1983). In the present study, the speakers read both a formal and informal passage in each guise (see Ryan & Carranza, 1975; Hogg et al., 1984). It was predicted that, as the accents were middle class, pupils at the middle class school should more easily identify with ingroup speakers, and hence display more ingroup favouritism than those at the working class school. No specific predictions were made concerning the formality of context, although we note that previous studies have found reduced solidarity with standard speakers in formal settings (Hogg et al., 1984), and heightened solidarity with regional speakers in informal settings (Bourhis, Giles & Tajfel, 1973; Sebastian & Ryan, 1985). These findings may not be replicated with the equal status varieties employed in the present study.

< previous page

page_48

next page >

< previous page

page_49

next page > Page 49

Method Design Social class, and the experimental variables of comparison condition, accent designation and situation formality were orthogonally manipulated in a 2 (middle class subjects vs. working class subjects) × 3 (Dundee/Glasgow vs. Dundee/RP vs. Glasgow/RP) between subjects and 2 (ingroup vs. outgroup) × 2 (formal vs. informal) within subjects design. The matched guise technique was adopted, using two speakers for each accent and context. Participants and Sessions One hundred and twenty male and female 14-15 year old school pupils from Dundee, of which 60 were attending a state school in a working class area, and 60 were attending private school in a middle class area, were assigned randomly to conditions, and the entire design was run within each school. The study was conducted by a female experimenter who was born and raised in Dundee, and who spoke with a middle class Dundee accent. Procedure and Materials The experiment was introduced as being concerned with how people rate personalities when only limited information is available (as in the Hogg et al., 1984 study). Subjects listened to a tape recording of eight stimulus voices (two accent types x formal/informal × speaker 1/speaker 2) in randomised order. Having listened to all 8 guises the tape was replayed, and subjects were asked to rate each one on a separate page of a special answer booklet. Each page required the subject to rate the speaker on eight 5-point bipolar adjective scales, four representing the status dimensions (educated/uneducated, intelligent/unintelligent, successful/ unsuccessful, rich/poor), and four the solidarity dimensions (friendly/unfriendly, kind/cruel, trustworthy/untrustworthy, good/bad). These dimensions were selected as they were found to be useful both by Carranza & Ryan (1975) and Hogg et al. (1984). A ninth item in the same format represented the likely occupation of the speaker (office worker/factory worker). This was included as a more direct measure of the speaker's perceived social class. All scales were scored so that the higher number represents the more positive evaluation. The stimulus tape involved two female speakers from Dundee, who were trained by native speakers from Glasgow and London prior to recording both of two passages in each of the three guises. Female speakers were used so as to make the stimulus voices and experimenter the same sex, thus avoiding confoundings due to differential sex-identification with speakers and the experimenter. Sex of speaker was not in itself considered likely to affect the relative impact of the independent variables (cf. Giles, Smith, Ford, Condor & Thakerar, 1980). A separate

< previous page

page_49

next page >

< previous page

page_50

next page > Page 50

tape was then produced, containing (in random order) only the relevant two guises for each condition. The speakers were instructed to assume the same personality and to be paralinguistically consistent for all readings. The two passages were those used by Hogg et al. (1984). Each was approximately 160 words long, and dealt with a different topic to manipulate situational formality. The informal situation involved a mother and son at home, discussing playing football with friends, while the formal situation involved a classroom discussion concerning the development of Jazz. After rating the speakers, subjects returned their booklets to an envelope before receiving a further booklet of questions, all presented as 5-point Lickert scales. These were two questions concerning feelings of identification with each of Dundee, Scotland, England, and Britain; three questions about perceptions of Scottish self-government and the value and persistence of Scottish accents, respectively; and questions about the importance of learning French, Latin, Gaelic, German, and Spanish as second languages; and a question about each parent's occupation. After completing this booklet subjects were informed that the study was over, and then participated in a discussion about the study. Debriefing revealed that no subjects were aware of the experimental hypotheses, or that the eight passages had not involved eight different speakers. Results Precautionary analyses revealed no significant effects of sex of any consistency or theoretical relevance. This factor has therefore been dropped from the presentation of the analyses which follows. For the class manipulation, the measure of subjects' parental occupations (coded in terms of Registrar General's criteria for socio-economic status) revealed a highly significant difference between the two schools in the predicted direction (F(1,111) = 105.85, p<0.001), confirming that the two schools were indeed composed of pupils from different social class backgrounds. Speaker Ratings The items determined a priori as measuring solidarity were averaged, as were the four measuring status (cf. Hogg et al., 1984). Speaker ratings are therefore represented by three measures: solidarity, status, and employment rating. These were analysed using a 3 (Contrast) × 2 (Class) betweensubjects × 2 (group rated) × 2 (context formality) between subjects repeated measures ANOVA. Although the findings reported below were somewhat stronger for one speaker, there were no significant interactions with speaker. This factor was therefore excluded from the analyses reported below, in which ratings of the two speakers for each guise were averaged.

< previous page

page_50

next page >

< previous page

page_51

next page > Page 51

Table 1 Cell means for effects of Condition and Group on ratings of speakers' solidarity, status and job Comparison Condition Dundee vs. Glasgow Dundee vs. RP Glasgow vs. RP Group In Out In Out In Out n=40 n=40 n=40 Solidarity 2.70b 2.44a 2.84b 2.52a 3.10c 2.50a Status 2.94c 2.60b 3.20d 2.44b 3.40d 2.19a Job 3.34c 2.71b 3.51c 2.37a 3.80d 2.15a Note: Scales run from 1 (low) to 5 (high). Means with different superscripts within each row differ significantly (Newman-Keuls, p<0.05). On the solidarity ratings there was a highly significant main effect of group (F(1,97) = 35.24, p<0.001) in which ingroups (mean = 2.88) were rated more highly than outgroups (mean = 2.48) within all conditions. A significant effect of condition (F(2,97) = 4.80, p<0.01) was qualified by a condition × group interaction (F(2,97) = 3.45, p<0.05). Specifically, while all outgroup ratings were equivalent, the rating of the ingroup (Glasgow) in the Glasgow/RP condition was higher than the rating of the ingroup (Dundee) in the Dundee/Glasgow and the Dundee/RP conditions (NewmanKeuls, p<0.05), as can be seen in Table 1) 2 On the status ratings we obtained no main effect of condition, but an even more powerful main effect of group (ingroup mean = 3.17, outgroup mean = 2.41: F(1,99) = 122.21, p<0.001), which was again qualified by an interaction with condition (F(2,99) = 14.72, p<0.001). The same pattern was obtained as on the solidarity measure, only this time the ratings of both outgroup and ingroup were significantly more extreme in opposite directions in the Glasgow/RP condition than in all others (see Table 1). There were also some interesting interactions involving class. The main effect of class (F(1,99) = 8.77, p<0.01) in which the middle class subjects gave higher status ratings overall, was qualified by an interaction with group (F(1,99) = 7.87, p<0.01). Middle class pupils regarded ingroup speakers as higher status than did working class pupils (means = 3.37, 2.98, respectively). There were no differences in ratings of outgroups. There were also three interactions, but no main effect involving context. Most important theoretically is the group × context interaction (F(1,99) = 6.54, p<0.05). The difference between ratings of ingroup and outgroup in the formal context was less than that in the informal context (see Table 2). The significant condition × context interaction (F(2,99) = 4.51, p<0.05) revealed that formal speakers were rated higher than informal in the Glasgow/RP condition (means = 2.88 vs. 2.74), but lower than informal speakers in the other two conditions (means = 2.77, 2.74 vs. 2.88, 2.82). Finally, the class × context interaction (F(1,99) = 13.94, p<0.001) indicated that middle class pupils rated formal context speakers (mean = 2.97) more highly than did working class pupils (mean = 2.62) (Newman-Keuls

< previous page

page_51

next page >

< previous page

page_52

next page > Page 52

Table 2 Cell means for effects of Group and Context on ratings of speakers' solidarity status and job Group Ingroup Outgroup Context Formal Informal Formal Informal Solidarity 2.53 2.88x 2.53 2.45y Status 3.13b 3.22b 2.47a 2.37a Job 3.45b 3.63c 2.47a 2.36a Note: Scales run from 1 (low) to 5 (high). Means with different superscripts a, b, c differ significantly (Newman-Keuls, p<0.05). For solidarity, the interaction is nearly significant (F(1,97) = 2.71, p< 0.07), and means with superscripts x and y differ (p<0.10). p<0.05), whereas their evaluations of informal context speakers did not differ (means = 2.84 and 2.77, respectively). For the job evaluations the pattern was highly similar to that for status ratings (see Tables 1 and 2). There were no significant main effects of condition, class or context, but there was a highly significant main effect of group (ingroup mean = 3.55, outgroup mean = 2.41; F(1,108) = 156.43; p<0.001). This was qualified again by an interaction with condition (F(2,108) = 11.47, p<0.001) in which once again, least differentiation arose in the Dundee/Glasgow condition, and most in the Glasgow/RP condition. As with the status and solidarity measures, the difference between the two ratings of Glasgow between these conditions was highly significant (p<0.001, using t-test), There was also a class × group interaction (F(1,108) = 4.31, p<0.05). The middle class pupils rated the ingroup as having a better job (mean = 3.64) and the outgroup as having a worse job (mean = 2.33) than did the working class pupils (means = 3.44, 2.49, respectively, p<0.05 by Newman-Keuls test). The interaction between group and context (F(1,108) = 7.69, p<0.01) took the same form as that on the status measure (see Table 2). There were no other significant interactions. In summary, a remarkably clear pattern of results has emerged across the three measures. There is a highly significant favouring of ingroup over outgroup in every condition, which is heightened, especially on status measures, in the Glasgow/RP condition. Ingroups are preferred in informal contexts, while this is not true of outgroups. Middle class pupils express stronger ingroup favouritism than do working class pupils. Identification and Nationalism The remaining measures were entered into a factor analysis with oblique rotation. Five factors emerged with eigenvalues >1.0, accounting for 65% of the variance. Factor loadings of >0.30 were regarded as significant and are reported in

< previous page

page_52

next page >

< previous page

next page >

page_53

Page 53 Table 3 Correlations between post-experiment identification with Scotland, separatist attitudes, pro-British sentiment, and ratings of speakers' solidarity, status and job Post-experiment Variable Accent Rating Scottish Identity Separatism ProBritish Dundee Solidarity 0.10 -0.13 -0.16 Status 0.22* -0.06 -0.11 Job 0.20* -0.04 -0.18 Glasgow Solidarity 0.35*** 0.14 -0.24* Status 0.37*** 0.17 -0.17 Job 0.25* 0.15 -0.20* RP Solidarity 0.13 -0.21* 0.12 Status 0.19 -0.31** 0.28** Job 0.12 -0.26* 0.19* Note: Pearson product moment correlations *p<0.05, **p<0.01, ***p<0.001

n 72 73 78 77 79 79 68 70 74

brackets below. The first factor tapped pride in being from Dundee (0.67), a sense of belonging to Dundee (0.41), feeling strong ties with Scotland (0.76) and being glad to be seen as Scottish (0.74). This factor seems to measure Scottish identity. A second factor comprised responses on the importance of learning the three modern languages (French (0.68), Spanish (0.71), German (0.90)). The third seemed to tap the need for separatism (or perhaps vitality), as it included measures of being unpatriotic to England (0.41), favouring self-government for Scotland (0.32) and believing that accents will not become uniform across Britain (0.86). The fourth factor involves anti-Britishness, as it includes the two measures of not thinking of oneself as British (0.64) and not feeling loyal to Britain (0.61). The fifth factor includes the two measures of the importance of learning the ancient languages of Latin (0.61) and Gaelic (0.54). For each factor a summed score was derived by adding those items which loaded significantly. These scores were then correlated with the ratings of the three different accents (thus, for any accent, subjects from only two conditions are involved). As can be seen in Table 3, the identity and separatism factors correlate in different ways with each accent. Scottish identity is positively related to the status and job ratings of the Scottish speakers, and of the solidarity rating of the Glasgow speaker, but is unrelated to the ratings of the RP speaker. In contrast, the separatism factor is negatively related to ratings of the RP speaker, but unrelated to rating of the Scottish speakers. This suggests that when Scottish identity is salient it encourages both positive ingroup evaluations in absolute terms, and a desire for intergroup distinctiveness, or separatism, but that the latter only occurs when a non-Scottish outgroup is invoked (i.e. when the differentiation can be made).

< previous page

page_53

next page >

< previous page

page_54

next page > Page 54

Discussion The results provide clear support for a social identity analysis of variation in language attitudes. As predicted, evaluations of the Glasgow accent shifted from negative to positive when the comparison accent changed from being Dundee to being RP. This shift illustrates the differentiation process in which meaningful contrasts are made within a particular frame of reference. It also reveals how social psychological processes can have a substantial influence even within enduring sociolinguistic contexts. Specifically, when real status and power differences are less salient, or attenuated, accent loyalties may shift in the direction of whichever accent is most easily adopted as a criterion for ingroup membership. Thus, when a Dundonian makes a comparison between Glasgow and RP, it is not the 'Glaswegian-ness' or 'Dundonian-ness' of the accent which is salient, but the features which distinguish it from RP, as being Scottish. The fact that Dundee and Glasgow speakers have quite different accents (mid-east and midwest Scottish, respectively (Murison, 1981)) becomes irrelevant. However, when the comparison is between Dundee and Glasgow, the east-west differences become more salient, and a Dundonian is likely to categorise accents in terms of 'Dundonian-ness'. Pursuing this line of reasoning, one would predict also that there would be distortions in perceptions of the extremity of the local accents according to the comparison accent. Ryan, Carranza & Moffie (1977) found that naive students were sensitive to varying degrees of Spanish accentedness when rating Spanish-English speakers. Perhaps they would have made fewer distinctions between speakers if an additional comparison group (say German accented English speakers) had been introduced. In this situation the frame of reference is shifted, and listeners would be predicted to attend to different features of speech. If the present study were to be repeated including measures of perceived accentedness, the Dundee and Glasgow accents might be perceived as less extreme or typical of their areas, and more typical of 'Scottish' in the conditions where they are contrasted with RP, than where they are contrasted with each other. Clearly there are fruitful avenues for further research here, including the possibility of introducing yet higher level outgroups (e.g. Australian, American), to see if Scottish subjects will make favourable evaluations of RP speakers. The effects of context obtained in the present study are interesting in that they contrast with those of Bourhis et al. (1973), Ryan & Carranza (1975), and Creber & Giles (1983). In the Ryan & Carranza, and Creber & Giles studies, a standard speaker was upgraded on status ratings in a formal context. However, this was in the context of a contrast in which the standard speaker was already in a high status group, and lower status regional accented speakers attracted relatively greater solidarity. As Bourhis, Giles & Tajfel (1973) found, the sense of solidarity when rating a regional accented speaker, increases if that speaker is a high status group member in an informal context (i.e. downward divergence is displayed in terms of speech accommodation theory (Giles, 1984)). The present results show that when speakers' socioeconomic status is high in all guises the use of the ingroup variety in an informal context can actually elevate the speaker's perceived status. This

< previous page

page_54

next page >

< previous page

page_55

next page > Page 55

seems sensible in that, irrespective of the particular accent, speakers would be expected to use a more standard, or middle class, form in a formal context (and hence one can learn little about the true status of the person), whereas only speakers who really are middle class will use the middle class variety in informal contexts. Given the desire to render ingroup members positively distinct, this explains why only ingroup speakers attract higher ratings in informal contexts, while outgroup speakers do not. A second interesting finding is that the middle class subjects were more confident in differentiating between groups on status dimensions. This may have been because, coming from a powerful group within Scotland, they had a greater sense of identity with Scottish middle class speakers, and felt more confident about expressing favourable evaluations of them. Certainly, the data fit with the idea that both subjective vitality and objective power and status are likely to encourage accent loyalty (cf. Bourhis & Sachdev, 1984). Notes 1. Some of the data from this article were presented at the 3rd International Conference on Social Psychology and Language, Bristol, July 1987. We wish to thank Gillian Christie for help with data collection and Diane Houston for comments and advice on an earlier version. Correspondence concerning this article should bc directed to Dominic Abrams, Department of Psychology, The University, Dundee DD1 4HN, Scotland, U.K. 2. There was also a significant class x condition interaction (F(2,97) = 3.81, p<0.05) qualified by a class × condition × groups × context interaction (F = 3.77, p<0.05). Ingroup bias increased from the Dundee vs. Glasgow, to Dundee vs. RP, to Glasgow vs. RP conditions for the formal context. For the informal context this pattern also emerged among the middle class subjects but working class subjects displayed less ingroup bias in the Dundee vs. RP condition than in the others. We do not have a ready explanation for this interaction. References Abrams, D., Wetherell, M., Cochrane, S., Hogg, M. A. and Turner, J. C. (In submission) Knowing what to think by knowing who you are: Selfcategorization and the nature of norm formation, conformity and group polarization. University of Dundee. Abrams, D., Emler, N. P., Jamieson, L. and Martens, J. (1987) First analyses of the ESRC 16-19 initiative survey data from Scotland. Working paper, University of Dundee. Bourhis, R. Y., Giles, H. and Rosenthal, D. (1981) Notes on the construction of a 'Subjective Vitality Questionnaire' for ethnolinguistic groups, Journal of Multilingual and Multicultural Development, 2, 145-55. Bourhis, R. Y., Giles, H. and Tajfel, H. (1973) Language as a determinant of ethnic identity, European Journal of Social Psychology, 3, 447-60. Bourhis, R. Y. and Sachdev, I. (1984) Vitality perceptions and language attitudes: Some Canadian data, Journal of Language and Social Psychology, 3, 97-126. Brook, G. L. (1963) English Dialects. London: Deutsch. Byrne, D. (1971) The Attraction Paradigm. New York: Academic Press.

< previous page

page_55

next page >

< previous page

page_56

next page > Page 56

Carranza, M. A. and Ryan, E. B. (1975) Evaluative reactions of bilingual Anglo and Mexican American adolescents towards speakers of English and Spanish. International Journal of the Sociology of Language, 6, 83-104. Cheyne, W. M. (1970) Stereotyped reactions to speakers with Scottish and English regional accents, British Journal of Social and Clinical Psychology, 9, 77-79. Creber, C. and Giles, H. (1983) Social context and language attitudes: The role of formality-informality of the setting. Language Sciences, 5, 155-62. Edwards, J. (1982) Language attitudes and their implications among English speakers. In E. B. Ryan & H. Giles (eds), Attitudes Toward Language Variation: Social and Applied Contexts. London: Edward Arnold. Eiser, J. R. and Stroebe, W. (1972) Categorisation and Social Judgement. London: Academic Press. Giles, H. (1970) Evaluative reactions to accents, Educational Review, 22, 211-27. (ed.) (1984) The dynamics of speech accommodation, International Journal of the Sociology of Language, Special Issue, 46. Giles, H., Bourhis, R. Y. and Taylor, D. M. (1977) Towards a theory of language in ethnic group relations. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Giles, H., Harrison, C., Smith, P. M., Creber, C. and Freeman, N. H. (1983) Developmental and contextual aspects of children's language attitudes, Language and Communication, 3, 141-46. Giles, H. and Powesland, P. F. (1975) Speech Style and Social Evaluation. New York: Academic Press. Giles, H. and Ryan, E. B. (1982) Prolegomena for developing a social psychological theory of language attitudes. In E. B. Ryan & H. Giles (eds), Attitudes Toward Language Variation: Social and Applied Contexts. London: Edward Arnold. Giles, H. and Sassoon, C. (1983) The effects of speaker's accent, social class background, and message style on British listeners' social judgements, Language and Communication, 3, 305-13. Giles, H., Smith, P. M., Ford, B., Condor, S. G. and Thakerar, J. N. (1980) Speech style and the fluctuating salience of sex, Language Sciences, 2, 260-82. Hogg, M. A. and Abrams, D. (1988) Social Identifications: A Social Psychology of Group Processes and Intergroup Relations. London: Routledge. Hogg, M. A., Joyce, N. and Abrams, D. (1984) Diglossia in Switzerland? A social identity analysis of speaker evaluations, Journal of Language and Social Psychology, 3, 185-96. Hogg, M. A. and Turner, J. C. (1987) Social identity and conformity: A theory of referent informational influence. In W. Doise & S. Moscovici (eds), Current Issues in European Social Psychology (Vol. 2). Cambridge: Cambridge University Press. Johnson, P., Giles, H. and Bourhis, R. Y. (1973) The viability of ethnolinguistic vitality: A reply, Journal of Multilingual and Multicultural Development, 4, 255-69. Lambert, W. E., Hodgson, R. C., Gardner, R. C. and Fillenbaum, S. (1960) Evaluational reactions to spoken language, Journal of Abnormal and Social Psychology, 60, 44-51. McKirnan, D. J. and Hamayan, E. (1984) Speech norms and attitudes toward outgroup members, Journal of Language and Social Psychology, 3, 2130. Murison, D. (1981) The Scottish language. In D. Daiches (ed.), A Companion to Scottish Culture. London: Edward Arnold. Price, S., Fluck, M. and Giles, H. (1983) The effects of language of testing on bilingual pre-adolescents' attitudes towards Welsh and varieties of English, Journal of Multilingual and Multicultural Development, 4, 149-61. Ryan, E. B. and Carranza, M. A. (1975) Evaluative reactions towards speakers of standard English and Mexican-American accented English, Journal of Personality and Social Psychology, 31, 855-63. Ryan, E. B., Carranza, M. A. and Moffie, R. W. (1977) Reactions towards varying degrees of accentedness in the speech of Spanish-English bilinguals, Language and Speech, 20, 24-26.

< previous page

page_56

next page >

< previous page

page_57

next page > Page 57

Ryan, E. B. and Sebastian, R. J. (1980) The effects of speech style and social class background on social judgements of speakers, British Journal of Social and Clinical Psychology, 19, 229-33. Sachdev, I. and Bourhis, R. Y. (1984) Minimal majorities and minorities, European Journal of Social Psychology, 14, 35-52. Sebastian, R. J. and Ryan, E. B. (1985) Speech uses and social evaluation: Markers of ethnicity, social class and age. In H. Giles & R. N. St. Clair (eds), Recent Advances in Language, Communication, and Social Psychology. London: Erlbaum. Tajfel, H. and Turner, J. C. (1979) An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (eds), The Social Psychology of Intergroup Relations. Monteray, Ca.: Brooks-Cole. Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D. and Wetherell, M. (1987) Rediscovering the Social Group: A Self-Categorization Theory. Oxford: Blackwell.

< previous page

page_57

next page >

< previous page

page_58

next page > Page 58

< previous page

page_58

next page >

< previous page

page_59

next page > Page 59

Language and Social Identity: Israelis Abroad 1 T. F. McNamara Department of Russian and Language Studies, University of Melbourne Abstract This paper explores the potential of social identity theory for explaining language attitudes and language maintenance and shift in an immigrant group. It reports a study of Israeli native speakers of Hebrew and their children in Melbourne from an intergroup perspective. The study investigates the Israelis' perceptions of the new intergroup setting in which as immigrants they find themselves and the social categories salient within it. As yordim (a Hebrew word with negative connotations referring to Israelis living permanently abroad) the subjects in the study report that they are stigmatised by Israelis remaining in Israel and by the wider Jewish community in Melbourne; this stigma is to some extent internalised. With reference to the Gentile community, subjects are stigmatised both as immigrants and more particularly as Jews, an experience which the inter-group situation in Israel precludes. The multiple group membership of subjects (yordim, Jews, immigrants) is a key to understanding the intergroup relations involved. It was hypothesised that the transformation of the social identity of the subjects is accompanied by language attitudes favouring English over Hebrew among the immigrants and a rapid shift to English among their children. This was confirmed in the data. Social identity theory (Tajfel & Turner, 1979; Tajfel, 1981; Turner, 1982; Hogg Abrams, 1987) provides an explicit theory of the intergroup setting in which language attitudes are formed. Most studies of language attitudes and the associated phenomena of language maintenance and shift among Australian immigrant groups have not related their findings to the dynamics of the context in which such phenomena occur (McNamara, 1987). This paper explores the way in which the altered social context in which members of an immigrant group find themselves affects the social identity, and hence the language attitudes, of members of the group. Each intergroup context creates its own set of salient categories for intergroup comparison and the establishment of social identity. Social identity is thus dynamic, interactive and context-dependent. In the act of immigration, immigrants are entering a new social context, in which their former social identity may have to be substantially redefined. In this paper, I will attempt to trace this process among Israeli immigrants to Australia and their children. This numerically insignificant group reveals the process and its linguistic implications particular clearly.

< previous page

page_59

next page >

< previous page

page_60

next page > Page 60

The Subjects: Israelis in Australia The data in this paper come from a partial analysis of material from an ongoing interview study of Israelis in Melbourne. So far, 60 informants have been interviewed, representing 50 families and a further five single individuals. I estimate that there are over 3,000 Israelis and their children in Melbourne, forming part of a Jewish population of some 35,000 in a city of almost 3 million. The Israelis are a minority within a minority. 'Israelis' are defined in this paper as those who left Israel after the age of 14 and who spent at least 13 years of their lives before the age of 25 in Israel. Of my sample, 30% were born outside Israel, but of these all had arrived in Israel by age 12; average age on arrival was 6 years. All claimed to be native speakers of Hebrew, and to feel 'most themselves' in Hebrew. Average length of residence in Australia was 12 years, the range being from 5 to 32 years. The average age of those interviewed was 37.43 years, with a range of 24 to 49 years. The Israeli Diaspora Emigration from Israel (Hebrew yerida) is a controversial matter, as Israel, the country which exists to bring an end to Jewish exile, is experiencing a new Diaspora. Figures on emigration from Israel are hard to establish, and controversial; figures for the number of Israelis permanently in the United States, for example, are an issue of dispute, with estimates ranging from fewer than 100,000 to over half a million. Evidence for something like the latter figure is given in a recent study by Fishman (1984) on mother-tongue claiming in the United States, where Hebrew has over half a million mother tongue claimants (a rapid increase in a matter of a very few years); Fishman (personal communication) believes that these claimants are Israelis settled in the United States rather than, for example, American Jews claiming Hebrew as a 'heritage' language. Ritterband (1986), among others, has argued for the lower figure. A figure of half a million represents a substantial proportion of Israel's Jewish population of 3.5 million, a serious situation for a beleaguered society where population is a matter of survival. Israel has for several years been experiencing a net loss of population through processes of immigration and emigration. Emigration traditionally has been stigmatised, as departure of citizens in wartime has been in other contexts: witness the British attitude towards Auden and Isherwood's decision to leave England for America just prior to the outbreak of World War II, or the current criticism of E1 Salvadoreans in Australia by their fellowcountrymen at home, who accuse them of 'abandoning the struggle'. Such an attitude finds expression in Hebrew, which has two terms for emigrant: m'hager, meaning one who leaves any society, and yored (literally, 'one who goes down'; plural yordim), a special term for emigrants from Israel; the latter term has strong negative connotations. In a famous remark some years ago, the then Israeli Prime Minister, Yitzhak Rabin, on a visit to New York, described the yordim of New York as the 'dregs' of Israeli society. (There is recent evidence that popular attitudes to this question are changing in Israel, as

< previous page

page_60

next page >

< previous page

page_61

next page > Page 61

people accommodate to the reality of the extent of emigration; as one informant put it, 'There is now a yored in almost every family in Israel'.) The Decision to Leave Israel The Israelis in this study seem to have internalised the stigma of yerida. Seventy-six per cent said that they had experienced guilt and conflict over the realisation that they were leaving or had left Israel for good (N = 38). Of those who had lived in Australia for five years or more, 71% expressed the intention of returning to live permanently in Israel in the future (N = 43); this was associated with the syndrome of denial that they were yordim at all, and were here only temporarily. There are in fact jokes about this among Israelis: the stereotype of the Israeli abroad is of a person living with his or her bags permanently packed, waiting for the 'right moment' to return, which will never come. This 'myth of return' is of course not exclusive to Israeli emigrants, but it is particularly strong among them. Given that emigration (often a decision to live permanently in Australia once they are there, rather than a decision made prior to leaving Israel) is a difficult decision, how are people led to it? Commonly stated reasons in my sample were: economic opportunities; marriage to a non-Israeli (38% were married to Australians); frustration or conflict in a job; educational opportunities; and desire to escape a tense military situation, particularly where male children, who would be required to serve in the army, were involved. It is worth noting that no informants left Israel because of opposition to the goals of the state; all informants expressed support for Israel and the goals of Zionism. Until recently, it was comparatively unusual for Israelis to apply as family groups to leave Israel, despite the presence in Israel of Australian immigration officials. Often, individuals or couples came to Australia as part of a tour, or to visit relatives, and stayed as it were inadvertently. Deliberate emigration was not the preferred pattern, as the following data on emigration patterns show (N = 40): marriage to Australian Jews or Gentiles (38%); heads of households, in family group, with children (20%); young married couples, without children (18%); as teenagers, with nonIsraeli born parents (18%); and on temporary contracts with the Israeli government or Israeli firms, or with Jewish schools (6%). Social Categories Through the process of immigration the Israelis enter an intergroup setting in which three ethnolinguistic groupings are salient: native born AngloAustralians; migrants from non-English speaking backgrounds; and Australian Jews. It was hypothesised that the following intergroup categorisations would be salient for the Israelis in the study: 1. Jewish vs Gentile (in interaction with Gentiles). In this interaction, a specifically Israeli identity is less significant. Gentiles are on the whole not very suc-

< previous page

page_61

next page >

< previous page

next page >

page_62

Page 62 cessful in making a distinction between 'Jewish' and 'Israeli' (see below, p. 219). 2. Yordim vs Australian Jews (in interaction with Australian Jews). In contrast to Gentiles, Australian Jews are able to recognise cues to categorisation as 'Israeli' (and hence 'yored'), a salient category in the Jewish community. 3. Yordim vs Israelis in Israel, or temporarily resident in Australia (in interaction with Australian Jews, or with other Israelis). This bears on the stigma associated with yerida, discussed above. 4. Migrant vs native-born Anglo-Australian (in interaction with native-born Anglo-Australians). The migrant presence is very obvious in Melbourne; for example, 20.7% of the population of Melbourne were reported as using a community language other than English in the 1976 census (Clyne, 1982: 6). Immigrant groups are low status groups, and immigrant languages have lower status than English (see below). Note that these categories differ from those salient in Israel, where important categories (cf. Smooha, 1978) include: 1. Israeli vs non-Israeli. 2. Religious (the minority) vs non-religious (the majority) a distinction expressed (confusingly for Gentiles) by Israelis as 'Jewish' vs 'Israeli'. Thus Israelis are sometimes to be heard saying 'I'm not Jewish, I'm Israeli', which may simply mean that they are not religiously observant; or it may mean that they do not share the traditional values of Jewish communities in the Diaspora, but instead identify with the values of a secular Zionist society in which Jews enjoy majority, not minority status. 3. Ashkenazi (belonging to the European Jewish tradition) vs Oriental (belonging to the eastern Jewish tradition). These categories may no longer be salient for Israelis in Australia. The argument of this paper is that the transformation in social context involves a transformation in the possibilities for the development and maintenance of a particular social identity, which in turn will alter language attitudes. To investigate the salience of the above hypothesised categories for Israelis in Australia, the Israelis in the sample were asked to estimate intergroup perceptions and attitudes as follows: a) To what extent do you think there is any negative feeling towards Israel and to Israelis specifically as Israelis (not as Jews) among Australians as a whole? ( , N = 24); b) To what extent do you think there is anti-Semitism in Australia at present? (

, N = 27);

c) To what extent do you think there is any negative feeling towards Israelis living in Melbourne among the Jewish community as a whole? ( N = 25); and d) To what extent do you think there is any negative feeling towards migrants among Australians as a whole? (

< previous page

page_62

,

, N = 28).

next page >

< previous page

page_63

next page > Page 63

The scale used was: (1) none, (2) a little, (3) some, (4) a fair amount, (5) a lot, and (6) don't know. In interaction with Gentiles, Israelis felt they were perceived negatively in terms of their general migrant origin, or, if specifically, in terms of their Jewishness. This is confirmed by significant differences between the scores for questions (a) and (d) (Wilcoxon's T = 10, N = 15, p<0.01 (twotailed)), and for questions (a) and (b) (Wilcoxon's T = 0, N = 14, p<0.01 (two-tailed)). The appropriate interpretation of the low score for 'Israeliness' is unclear. Does it suggest positive attitudes towards Israelis, or indifference.'? Responses to this question generally focussed on perceptions of media representation of Israel or of popular attitudes to Israel as a state, rather than on personal encounters with Gentiles in which 'Israeliness' had been salient. Comments about negative media reports of Israel in recent years were balanced by comments about earlier more positive attitudes, for example at the time of the Six Day War or the raid on Entebbe Airport. Several informants commented on the irrelevance of Israel as a political entity to ordinary Australians. As far as the lack of salience of 'Israeliness' in personal encounters is concerned, it is not really surprising that 'Israeliness' is not a particularly salient category for Gentiles. The cues to such a categorisation (e.g. recognition of a Hebrew accent, or of specifically Hebrew names) are not generally available to most Gentiles, whereas cues to categorisation in terms of Jewish identity are well-known, and reinforced by conditioning which draws on a centuries-old tradition of Christian antagonism to Jews. The experience of being perceived by the outgroup in terms of 'Jewishness' rather than 'Israeliness' is something new for the Israelis, most of whom have themselves had no direct experience of Jewish minority existence in a Christian society. In fact, non-religious Israelis (the majority) are not necessarily used to seeing themselves as 'Jewish' at all, a term which, as has been mentioned above, they often use to refer to Orthodox or Ultra-Orthodox Jews in Israel, or to Diaspora Jews. Several informants articulated their awareness of the changed categories in the new social context and the implications of outgroup perceptions: The minute an Israeli leaves Israel, I believe he is a Jew . . . as far as the non-Jewish community is concerned. In Israel he can think as an Israeli but in the Diaspora he becomes a Jew whether he likes it or not. (Informant #42) I don't like to be called Jewish, I'm Israeli. In Israel, there were no Jews and Christians. There were Israelis and non-Israelis. (Informant #11) Israeliness is not the same as being Jewish. Here, being Jewish takes over because the main point is to distinguish yourself from Christians. (Informant #1) In Israel, I used to [make a distinction between 'Jewish' and 'Israeli'] - not here. In Israel l used to say I'm Israeli, I'm not Jewish because I quite resent all those Ultra-Orthodox . . . but here . . . because I live in the Exile . . . I don't see much difference. You have to have both of them, to strengthen the ties. (Informant #31)

< previous page

page_63

next page >

< previous page

page_64

next page > Page 64

The Israelis' reported perceptions of anti-Semitism confirm that categorisation as 'Jewish' is evaluated negatively by the Gentile outgroup. This is confirmed by explicit awareness of specific anti-Semitic incidents, mostly verbal, occasionally violent, in all but 5 of 38 interviews examined on this point; of these, 23 involved reports of incidents that were directed against the informants or their children personally. One factory worker had been presented with a wedding present by the supervisor on behalf of her co-workers with the words 'To be honest with you they didn't want to buy you a present because you're Jewish'. Another woman, driving with her Orthodox husband, responded to a signal from another driver to wind down her window at traffic lights, only to be greeted with a volley of anti-Semitic abuse. Such incidents, while infrequent, serve to draw a boundary and to create an unmistakable awareness of membership groups. However, it is not only, or perhaps even most importantly, incidents such as these which make Israelis aware of the salient categories in the new social context. The pervasiveness of Christian symbolism in mainstream Australian culture acts as a constant reminder to the Israelis of their Jewish 'otherness'. It is possible to trace the growth of an awareness of the relevant categories in the development of social identity in the child. One mother discussed the process in the following way: It took her quite a long time to understand the fact that we're Jewish and it comes up usually during Christmas. Well she sees all the ads on television and it looks beautiful and she says 'Can we have a Christmas tree?' so I said 'Look we can't have a Christmas tree because we're Jewish and Jews don't have Christmas trees' but sometimes Hanukkah [Jewish Festival of Lights, usually falling in December] comes but I said to her 'I don't have to justify because the Christians have Christmas we have Hanukkah, this is how it goes and this is how it is'. So now every time we go she says 'Is it true that we're Jewish, nachon she anachnu Yehudim?' and things like this. I said 'Yes, we're Jewish' . . . I remember going with my son through the same stages because we came here he was three and a half and by the time he was four and a half we went through a Christmas and . . . we went actually a year later we went on a trip and we got to Canberra and went into a motel and one of the maids there asked the kids 'Did you have nice presents for Christmas?' (it was Christmas time) and my son said to her 'We don't celebrate Christmas' (he was four and a half) 'we're Jewish but we had presents for Hanukkah' . . . Often this process of self- and other-categorisation involves linguistic cues that may at first be rather confusing to the child, as this nice example shows: My children when they came here they didn't know that they were Jewish . . . my boy asked me 'I . . . I'm Jewish?' or in the synagogue when they were taken my little boy asked the Rabbi at the end . . . he put up his hand 'Are you Jewish?' . . . the Rabbi! . . . and then I asked 'Why did you ask the Rabbi that question?' He said 'He couldn't speak Hebrew so I wasn't sure if he was Jewish or not . . . he spoke in English so how can I know?' He didn't quite connect the Rabbi with Judaism with everything . . . so they hadn't realised,

< previous page

page_64

next page >

< previous page

page_65

next page > Page 65

being always in a Jewish community, taking for granted, you never think of it unless you see someone else that is not, so you know what you are . . . like looking in a mirror . . . In interaction with Australian Jews, Israelis report that they encounter negative attitudes to themselves as Israeli yordim. This contrasts with Israeliness being a non-issue for the Gentile majority; there is a significant difference between the scores for questions (a) and (c) above (Wilcoxon's T = 0, N = 16, p<0.01 (two-tailed)). Israeli immigrants are a low-status group within the Melbourne Jewish community as a whole. Partly this is a question of attitudes to yerida (emigration). Australian Jews express their loyalty to Israel by reflecting official Israeli attitudes on this issue, and also perhaps their guilt over their own decision not to live in Israel themselves. These attitudes naturally enough breed a certain resentment among the Israelis. One informant reported replying to Australian Jews who asked him aggressively what he was doing in Australia: 'I came here to make room for you in Israel.' Negative stereotypes of yordim are further reinforced by the idea that it is only the 'bad' type of Israeli who tends to leave Israel. Majority Jewish attitudes, however, can also be understood in terms of relations between the Gentile majority and Australian Jews, who are anxious about anyone rocking the boat or drawing negative attention to the Jewish community. There is anxiety, for example, about recent police publicity given to Israelis wanted in connection with drug dealing. More generally, unassimilated newcomers are traditionally a threat to the position of more assimilated members of minority communities. One informant put it, perhaps too strongly, in the following terms: Australian Jews would like to hide, would like to blend nicely; if we Israelis are loud, talk Hebrew in public, are obviously Israeli and Jewish, it embarrasses them. In this view, the low status of Israelis can be linked to an awareness of pressure from majority attitudes and values. The Salience of Group Membership Giles & Johnson (1981) propose a number of factors influencing the salience of ethnolinguistic group membership, including: (a) perceived ethnolinguistic vitality; (b) perceived group boundaries; and (c) multiple group membership. Each of these factors will be examined in relation to the identity of the Israelis in the study. Perceived Ethnolinguistic Vitality Perceived ethnolinguistic vitality refers to a group's perception of itself as an (enduringly) distinctive and active collective entity in intergroup situations (Giles, Bourhis & Taylor, 1977; Bourhis, Giles & Rosenthal, 1981). Israelis in my study had a low perceived ethnolinguistic vitality. This has to do with the issue of the stigma of yerida, discussed above. In answer to the question, 'Is there a possibility

< previous page

page_65

next page >

< previous page

page_66

next page > Page 66

of an Israeli ethnic or cultural identity surviving outside Israel?', responses were (N = 36): Yes (17%), Yes (qualified) (14%), and No (67%). In answer to the question, 'Is there a future for Hebrew as a spoken, vernacular language outside Israel?', responses were (N = 18): Yes (17%), Yes (qualified) (17%) and No (61%). Israelis are reluctant to pass on the stigmatised identity of the yored to their children. Some informants commented as follows: Personally I don't want to see that future abroad. I prefer all the Jews to be in Israel. (Informant #44) There is no second generation of Israelis; the second generation, they are Jews. They do not consider themselves any more Israelis, are not by any means Israelis. Israelis by definition are people living in Israel, or were from Israel and speak Hebrew. I would like the second generation to live in Israel; if they are going to live in Australia, should be Australian Jewish way . . . there is nothing spectacular about being Israeli in Australia; if you're going to be an Israeli, be in Israel. (Informant #58) It is also striking that Israelis, unlike other ethnolinguistic groups in Melbourne, have very few of their own institutions (i.e. separately from the Jewish community). For example, they do not have a separate social club which is exclusively theirs, where Hebrew is the natural language to use. This is not accidental; partly, Israel has not been prepared to institutionalise, and hence to be seen to be encouraging, the existence of separate communities of yordim; partly it is that the Israelis themselves have mixed feelings about such a club. Some welcomed the idea, but others felt that there was something not quite right about Israelis formally getting together just because they were Israeli. There was also resistance to the idea of publicly and obviously mixing with other Israelis. For some informants, it was or had been important to distance themselves from the category yordim; as stated above, many went to great lengths to rationalise their position in Australia as being temporary, and were at pains not to be seen to be associating with Israelis who could be seen as falling into the stigmatised category. Perceived Group Boundaries There is some evidence that intergroup boundaries in relation to the Gentile outgroup are perceived somewhat differently by Israelis as compared with other members of the Jewish community. Israelis are less likely to perceive a boundary than Australian Jews. Informants were asked, 'Is there any greater chance that the children of Israeli couples here will assimilate than the children of other Jewish couples?' Responses were as follows (N = 35): Yes (40%), No (46%) and Don't know (14%). Several of those who said 'No' commented on the fact that as the Australian Jews were Englishspeaking and were more assimilated to mainstream norms than the more recently arrived Israelis, the Israelis were in fact less likely to assimilate. Those who answered 'Yes' explained their responses in rather different terms:

< previous page

page_66

next page >

< previous page

page_67

next page > Page 67

Yes . . . but only when the Israelis continue living like they lived in Israel . . . in other words like Israelis. If they live like Jews . . . consciously . . . then the child might be insulated from assimilation . . . but otherwise if there isn't that conscious effort to make the child Jewish then they would assimilate more easily. (Informant #42) Yes, definitely yes, from my own experience . . . because again when you are born in a free country like Israel, you don't . . . your main issue in life is not being Jewish . . . right . . . people that live outside Israel, the main issue is Judaism . . . I was never asked or thought of me as Jewish . . . I always thought first Israeli because . . . Judaism just wasn't an issue, you learn it in history or Bible Studies or something like this . . . (Informant #57) Another informant commented: Being Israeli, you are not as aware of the outside environment as the Jews are. (Informant #58) This same informant commented that it was his Australian Jewish wife who insisted that the children go to a Jewish day school; left to himself, he would have sent them to the local high school, because that was the natural thing to do in the Israeli context. Others commented that Israelis and Australians shared a 'majority mentality', in contrast to the 'minority mentality' of Diaspora Jews. Israel had come into existence specifically to render obsolete the traditional minority status and what were seen as the accompanying complexes and conflicts of Jews. Despite this we have seen above how the pervasiveness of Christian cultural symbols and the experience of persistent anti-Semitism may constitute a boundary which discourages or prevents assimilation into the Gentile outgroup. It is important to remember that this is a Gentile-Jewish boundary, that is, one reflecting the Jewish rather than the specifically Israeli identity of the informants. One would expect a reflex of perception of such a boundary to be heightened awareness among the informants of the Jewish, as against the Israeli, dimension of their identities. The reaction of some informants has already been noted (pp. 219 above), and provides some evidence on this point. Further evidence is provided in the responses of informants to questions about religious beliefs, which may have the function of expressing awareness of traditional values and groupings as much as of expressing personal religious values. The Israelis in my sample remained on the whole not very involved with religion: 55% described themselves as 'not religious at all' or 'opposed to religion'; 21% said they were 'traditional' or 'somewhat religious'; only 24% said they were 'very' or 'moderately' religious (N = 42). However, there was some evidence that attitudes had changed as a result of living in Australia. In answer to the question, 'Have you become more or less religious since you came to Australia, or has your attitude remained the same?', 40% said they had become more religious, 5% said less, and 55% said there had been no change (N = 42). Some comments illustrate the point. Referring to the secular/ religious divide in Israeli society, one informant said: I am less anti-religious here . . . In Israel, I felt like bashing heads . . . (Informant #58)

< previous page

page_67

next page >

< previous page

page_68

next page > Page 68

Another expressed similar attitudes: When I was younger, I used to be very anti-religious; you can afford it in Israel, but not here. (Informant #62) Other comments were: Religion keeps Jews together in the Diaspora. (Informant #56) I do go to synagogue much more than I would do in Israel . . . there I considered myself only Israeli . . . because I didn't have to prove I am Jewish . . . here I go every festival. (Informant #31; cf. p. 219 above). A boundary also exists for the Israelis between themselves and the Australian Jewish majority. Perceptions of negative attitudes have already been reported. What is the Israelis' response to this? Adapting a procedure used to measure the extent of social relations between Australian Jews and Gentiles developed by Taft (1973), an Index of Social Relations with each outgroup (non-Jews and non-Israeli Jews) was developed to measure social distance between the Israelis in the sample and non-Israelis, both Jewish and Gentile. Responses to the following questions were used in developing the indices: (1) What percentage of your personal friends are (a) Israeli (b) Jewish, non-Israeli (c) non-Jewish? (2) Do you feel more at ease among Israelis or non-Israelis? (3) Do you feel more at ease among Jews or non-Jews? (4) Do you think Jewish people accept you socially as fully as Israelis do here? and (5) Do you think non-Jewish people accept you socially as fully as Jewish and Israeli people do? For each Index, a range of -5 (extremely outgroup-oriented) to +7 (extremely ingroup-oriented) was possible. Including only those interviews where there are complete data, the following results were found: Index of Social Relations with non-Jews: Mean score = +3.76 (N = 25); and Index of Social Relations with non-Israelis: Mean score = +3.31 (N = 35). There was complete data for both indices for 23 informants. There was no significant difference between the scores on each index for this group; in other words, informants did not see themselves as significantly closer to either outgroup (Wilcoxon's T = 32.5, N = 12, p<0.05 (two-tailed)). There was also a significant correlation between scores on each index: in other words, informants who tended to identify with the (Israeli) ingroup in relation to one outgroup also tended to identify with the (Israeli) ingroup in relation to the other outgroup; informants who tended not to identify with the ingroup in relation to one outgroup also tended not to identify with it in relation to the other (Kendall's Tau = +0.32, N = 23, p<0.05 (two-tailed)). In answer to the question on friendships with ingroup and outgroup members, the mean percentages were as follows (N = 38): Other Israelis (52.5%), Jewish, non-Israeli (32%), and Non-Jews (15.5%). There is evidence, then, that the Israelis prefer to keep themselves separate from both the Gentile and the Australian Jewish outgroups. This might be expected to lead to maintenance of Hebrew at least among the immigrants themselves; the case of their children is a different matter, as we shall see. Informants also were asked about their affiliation to institutions within the Jewish community. Thirty-four per cent had some affiliation with at least one

< previous page

page_68

next page >

< previous page

page_69

next page > Page 69

institution; 66% had no affiliation with any institution. This figure is low when compared with the Melbourne Jewish community as a whole, where the rate of affiliation is over 70% (Lippmann, 1973). However, the figure for Israeli affiliation in Melbourne is comparable to the rate of affiliation of Israelis in the New York area (31%), which in turn was similar to the rate of affiliation of other New York Jews (32%); the rates for Orthodox Israelis and other New York Jews were higher, and again similar (Ritterband, 1986: 123). A finding which contrasts with the New York findings is the figure for the percentage of the children of Israelis who have had a formal Jewish education: 84% of the children of the Israelis in the study had attended a Jewish school in Melbourne (N = 57); this compares with only 25% of children of Israelis in New York, and 40% of other New York Jewish children (Levi, 1986: 173). Partly this is a question of the unique position of the Jewish day school movement in Melbourne compared with other Jewish communities in the Diaspora (cf. Klarberg, 1980, 1983; Lubin, 1985), and it may also reflect a bias in the sample of Israelis, which was obtained initially through contacts within the Jewish community and subsequently through a 'snowballing' technique: The choice for Israelis in interaction with Australian Jews, then, is either to remain entirely separate or to assimilate to Australian Jewish norms. The solution seems to be for the first generation of immigrants to do the former, while encouraging their children to do the latter. In terms of the children, the problem of passing on a specifically Israeli identity outside Israel is that the Israelis' conflict over their identity as yordim precludes any desire for such a course; cf. the discussion of perceived ethnolinguistic vitality above, and the comments of the informants quoted there. Multiple Group Membership The discussion so far has already illustrated the importance of this concept for Israelis in Melbourne. In interaction with Gentiles, it is their Jewish identity which is salient, or their identity as 'migrants'; in interaction with Australian Jews, it is their identity as yordim. It is also worth noting that the Israelis may have little control over how they are identified. It would be advantageous to be seen by Gentiles as Israelis rather than as Jews, and by Australian Jews as Jews rather than as Israelis; but in fact the reverse happens. The notion of a hierarchy of ethnic identities is important in studies of social identity in ethnically complex settings (McNamara, 1987). An Egyptian in Australia, for example, may be 'Arab', 'Muslim' and 'Egyptian', to go no further than that; an Anglo-Australian may be 'Anglo-Saxon', 'Christian' and 'Australian'. Which identity comes into play will depend on the nature of the intergroup interaction involved. Implications for Language Attitudes and Language Shift The linguistic consequences of the new situation the Israelis find themselves in are complex, but are likely to point in one direction: language shift to English.

< previous page

page_69

next page >

< previous page

page_70

next page > Page 70

Hebrew as the language of an immigrant group in Australia shares the low status accorded to all migrant languages in Australia (cf. Mackiewicz & Kee, 1986 and the discussion of their findings in McNamara, in press). Majority group attitudes on this point are rapidly internalised by migrants themselves (Mackiewicz & Kee, 1986; Callan & Gallois, 1982). The language attitudes of the mainstream Jewish community are in turn an internalisation of majority group (i.e. Gentile) attitudes. Hebrew has a lower status than English as a vernacular language in the Jewish community; being Jewish in Australia does not involve being able to communicate in Hebrew. Hebrew has an important and traditional symbolic function for the Australian Jewish community (it symbolises commitment to Israel, and it is the language of traditional Jewish ceremonial) but its actual communicative function is limited. This important distinction and its consequences for language maintenance in other immigrant communities in Australia is discussed in Clyne (1988). Informants were asked the following questions on the use of modern Hebrew by non-Israelis in Australia: 'What do you think of the Hebrew taught in Australia?' and 'What do you think of the Hebrew used by non-native speakers in Australia?'. Two further questions were asked about attitudes to code-switching between Hebrew and English, both intra- and inter-sententially. Responses to the first two questions usually involved comments (including from several teachers of Hebrew themselves) about the limited success of attempts to teach Hebrew in the Jewish day schools, and the attitudes of parents and students: Hebrew was often seen, particularly in the non-Orthodox schools, as a low-status subject. Code-switching norms were invariably in favour of English, even among Hebrew teachers in the staffroom of at least one major Jewish school. Children mocked the accents (in English) of their Israeli teachers of Hebrew. In fact, the problematical status of modern Hebrew in the curriculum of the non-Orthodox schools may be associated with the low status of the Israeli teachers of the subject (cf. Klarberg, 1985, in support of the views expressed here). To the extent that Israelis are led to redefine themselves in the long term as Jews in the new intergroup context in the ways discussed above, we may expect them to show a commitment to the enduring symbolic function of Hebrew but not necessarily to its communicative function, in line with mainstream Jewish norms. This is one alternative. We have already seen that the alternative of passing on a separate Hebrew-speaking Israeli identity is problematic, as it involves a commitment to the stigmatised identity of the yored (cf. p. 222 above). The third possibility, of assimilation to the Gentile outgroup, will of course result in a shift to English. The above analysis would suggest that a rapid language shift to English is likely in the second generation, and this is borne out by the evidence. Informants were asked to assess their children's Hebrew for each of the skills of speaking, listening, reading and writing on the following scale: (1) like a native speaker, (2) very well, (3) quite well, (4) can get by, (5) a little, and (6) not at all. Results for 28 families studied were as follows: (a) Australian-born children, or Israeli-born with no education in Israel Speaking ( , N = 50), Listening ( , N = 50). Reading ( ,N= 43), and Writing ( N = 43), and (b) Israeli-born

< previous page

page_70

next page >

< previous page

next page >

page_71

Page 71 children, with at least one year's schooling in Israel Speaking ( , N = 9).

, N = 9), Listening (

, N = 9), Reading (

, N = 9), and Writing (

A further measure of language shift was linguistic interaction in the home. Informants were asked about the language used by each participant in dyadic interactions between each parent and each child. The following 5-point scale was used: (1) All Hebrew, (2) Mostly Hebrew, (3) 50-50 Hebrew-English, (4) Mostly English, and (5) All English. The results were as follows: (a) Language used by each parent to each child (Total number of dyads = 120) All or mostly Hebrew (22%), 50-50 Hebrew-English (10%), and All or mostly English (68%); and (b) Language used by each child to each parent (Total number of dyads = 118) All or mostly Hebrew (12%), 50-50 Hebrew-English (2%), and All or mostly English (86%). The data are evidence of a fairly massive shift to English in the second generation, an outcome predicted by the analysis above, Conclusion Social identity theory provides a powerful framework for the analysis of language attitudes and language maintenance and shift in intergroup settings. This study has dealt with a detailed study of one small community in Australia which illustrates the potential of the theory particularly clearly. In another paper (McNamara, 1987) the discussion is extended to reinterpret the findings of other recent studies in the Australian context. The prognosis for enduring bilingualism in Australia is not encouraging if the analysis supported by the theory is correct. Note 1. This is a revised version of a paper given at the 3rd International Conference on Social Psychology and Language at the University of Bristol, July 20-24, 1987. Requests for reprints should be sent to: Tim McNamara, Department of Russian and Language Studies, University of Melbourne, Parkville, Victoria 3052, Australia. References Bourhis, R. Y., Giles, H. anti Rosenthal, D. (1981) Notes on the construction of a 'subjective vitality questionnaire' for ethnolinguistic groups, Journal of Multilingual and Multi-cultural Development, 2, 2, 145-55. Callan, V. J. and Gallois, C. (1982) Language attitudes of Italo-Australian and Greek-Australian bilinguals, International Journal of Psychology, 17, 345-58. Clyne, M. G. (1982) Multilingual Australia. Melbourne: River Seine. (1988) The German-Australian speech community: ethnic core values and language maintenance, International Journal of the Sociology of Language, 72, 67-83. Fishman, J. A. (1984) Mother tongue claiming in the United States since 1960, International Journal of the Sociology of Language, 50, 21-100.

< previous page

page_71

next page >

< previous page

page_72

next page > Page 72

Giles, H., Bourhis. R. Y. and Taylor D. M. (1977) Towards a theory of language in ethnic group relations. In H. Giles (ed.). Language, Ethnicity and Intergroup Relations. London: Academic Press. Giles, H. and Johnson, P. (1981) The role of language in ethnic group relations. In J. C. Turner & H. Giles (eds), Intergroup Behaviour. Oxford: Basil Blackwell. Hogg, M. A. and Abrams, D. (1987) Social Identity, Intergroup Relations and Group Processes. London: Methuen. Klarberg, M. (1980) Ideology and language planning: Melbourne's Jewish schools, ITL Review of Applied Linguistics, 49-50, 67-90. (1983) Diglossic education: The Jewish tradition, some Australian manifestations and their implications, Journal of Intercultural Studies, 4, 2, 55-65. (1985) Hebrew and Yiddish in Melbourne. In M. Clyne (ed.), Australia, Meeting Place of Languages. Pacific Linguistics C-92. Canberra: Australian National University. Levi, L. G (1986) Israelis in New York and the Federation of Jewish Philanthropies: A study of anomie and reconnection. In A. Dashefsky (ed.), Contemporary Jewry, Volume 7 (First Annual). New Brunswick, NJ: Transaction Books. Lippmann, W. (1973) Melbourne Jewry: A profile. In P. Y. Medding (ed.), Jews in Australian Society. South Melbourne: Macmillan/Monash University. Lubin, C. (1985) Of two worlds: A comparative case study on the social settings of Jewish day schools in Melbourne, Journal of Intercultural Studies, 6, 1, 18-40. Mackiewicz, G. & Kee, P. K. (1986) Intergroup attitudes and behaviour in Adelaide. Paper presented at National Research Conference on Ethnicity and Multiculturalism. University of Melbourne, May 14-16. McNamara, T. F. (1987) Language and social identity: Some Australian studies, Australian Review of Applied Linguistics, 10, 2, 33-58. (in press) Language and relations between ethnic groups. Social psychological perspectives. In T. J. Quinn & T. F. McNamara, Issues in Second Language Learning: General and Particular. Geelong, Victoria: Deakin University. Ritterband, P. (1986) Israelis in New York. In A. Dashefsky (ed.), Contemporary Jewry, Volume 7 (First Annual). New Brunswick, NJ: Transaction Books. Smooha, S. (1978) Israel, Pluralism and Conflict. London: Routledge and Kegan Paul. Taft, R. (1973) Jewish identification of Melbourne Jewry. In P. Y. Medding (ed.). Jews in Australian Society. South Melbourne: Macmillan/Monash University. Tajfel, H. (1981) Human Groups and Social Categories. Cambridge: Cambridge University Press. Tajfel, H. and Turner, J. (1979) An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (eds), The Social Psychology of Intergroup Relations. Monterey, CA: Brooks/Cole. Turner, J. C. (1982) Towards a cognitive redefinition of the social group. In H. Tajfel (ed.), Social Identity and Intergroup Relations. Cambridge: Cambridge University Press.

< previous page

page_72

next page >

< previous page

page_73

next page > Page 73

Ethnicity and Immigrant Values: Religion and Language Choice Deepa Punetha Department of Psychology, University of Allahabad, India Howard Giles Department of Psychology, University of Bristol, U.K. and Louis Young Department of Psychology, National Taiwan University, Taipei Abstract While the British literature on multilingualism has invoked value differentials not only between Asian immigrant groups and the host culture but also amongst the former, no empirical evidence actually exists. Hence, an extended Rokeach Value Survey was administered to three Asian groups (Hindus, Muslims and Sikhs) and to a British indigenous sample; the bilingual immigrants were offered a choice of the questionnaire in their own ethnic tongue or in English. Canonical variate analyses showed large intergroup value differentials with language choice having an effect for all Asian groups (greatest for Sikhs and least for Hindus). The specific differentials, however, depended on the dimension examined. Immigrants choosing English had a value position intermediate to the indigenous group and the Asians choosing their ingroup tongue. South Asian immigrants to Britain have had problems adapting to their new culture in part because of large-scale differences between their home and the host country with respect to climate, culture, language and religion (Anwar, 1979; Kannan, 1978). Many of these groups have tried to maintain their religious and ethnic identities in Britain (Allen, 1971; Ballard, 1979; James, 1974) oftentimes intentionally remaining separate from mainstream Anglo society. Those who resist intercultural mixing do so for several reasons: they are unable to speak good English, are less educated, are from a rural background and have a 'dream' of returning to their home country (Brooks & Singh, 1979). Indeed, many of those who have this 'myth of return' perceive their own culture in static terms and change is viewed as being a possible or inevitable threat to the retention of their distinctive value systems. Yet at the same time, many Asian immigrants who are educated professionals become relatively westernised in their behavioural styles

< previous page

page_73

next page >

< previous page

page_74

next page > Page 74

and outlook (Nowikowski & Ward, 1979; Smith, 1976) and acculturated in their value systems (cf. Eisenstadt, 1954). While the extant literature highlights the notion of value differentials among South Asian immigrants and the host culture, as well as among the immigrant groups themselves (e.g. Bagley & Verma, 1983; Ballard, 1982; Swarm, 1985; Tomlinson, 1984), no comparative empirical research has explored the structure of these value systems nor therefore the extent to, or conditions under, which value acculturation is expressed, if at all. Hence, by utilising a modified version of the Rokeach Value Survey (see Rokeach, 1973) which has been adopted in several countries (e.g. Feather, 1975; Moore, 1976; Ng et al, 1982; Penner & Ank, 1977), the first objective of the present study was to investigate the values of three South Asian groups (Hindus, Muslims, and Sikhs) in Britain in comparison with a British indigenous sample. Each of the Asian groups has, of course, its own particular history, religious practices and beliefs, habits, lifestyle and so forth, most of which are retained in the British context (see Kanitkar, 1981; Khera, 1981; King, 1981; Kuppuswamy, 1975 for Hindus; Anwar, 1979; Ballard, 1982; Khan, 1979; Tomlinson, 1984 for Muslims; Ballard & Ballard, 1979; Ghuman, 1980; Helweg, 1979; Rex & Moore, 1967 for Sikhs). Furthermore, all of these groups have organised themselves by forming religious and cultural committees, and religious festivals and ceremonies are collectively celebrated. They attempt to maintain their cultural values and traditions and to transmit them to succeeding generations. Since they consider language as a vehicle for cultural and value maintenance (cf. Taft, 1977), children are taught their mother tongue in temples, mosques and guradwaras. The leaders and parents are constantly demanding recognition of their mother tongues in the mainstream schools. Indeed, such a perspective is consonant with much theory concerning the relationships between language and ethnicity (Berry, 1984; Giles, 1977; Giles & Johnson, 1987). The second aim of this study was to determine the role of language of testing in the elicitation of values from the three ethno-religious groups, as previous research has in certain (albeit not all) cultural contexts shown it to be an important variable (e.g. John, Young, Giles & Hofman, 1985; Johnson & Giles, 1982; Marin, Triandis, Betancourt & Kashima, 1983). However, researchers invariably have administered bilingual respondents with a value instrument in one language or the other; in other words, subjects have had no option in which language they have had to express their values. When bilinguals undertake the task in an outgroup language, sometimes they have accommodated towards the values inherent in the other culture (e.g. Ervin, 1964). In contrast, when forced to use the ingroup language, some apparently have compensated by affirming their ingroup values more strongly (see Bond, 1983; Bond & Cheung, 1984). In the study to be reported below, Asian bilingual subjects were afforded a choice as to the language in which they would prefer to answer to the questionnaire, English or their ethnic tongue. If, as many scholars have argued (Bourhis, 1979; Giles, Scherer & Taylor, 1979), language choice is at least in part a barometer of individuals' cultural dispositions, then it is predicted that those Asians choosing English as their preferred medium in this context would be more Anglo in their value expressions than those bilinguals preferring to use their own ethnic tongue.

< previous page

page_74

next page >

< previous page

page_75

next page > Page 75

Method Participants A total of 255 subjects (Ss) from Hindu, Muslim, Sikh (Asian) and white British (indigenous) groups from a cross-section of cities (11) in mainland Britain (viz., Southern and Northern England, Scotland and Wales) participated in the study and were predominantly middle class (as evident from their self-reports of educational attainment, occupation and income). In an attempt to explore general value systems, an adult (predominantly nonstudent) population was tested (mean age = 32.6 years, S.D. = 14.2). All 131 Asian Ss (56 male, 74 female & 1 unknown) were bilingual in their own language and English and noted to be so when interacting with the female Asian multilingual experimenter who talked to them in both their tongues (usually in a, not uncharacteristic, code-switching between them). Mean duration of stay in Britain for this first generation Asian sample was 14.4 years (S.D. = 6.7) with mean figures for Hindus, Muslims and Sikhs being 14.3, 15.4, and 17.9 years respectively; mean ages of these same groups were 31.2, 28.3, and 29.3 years respectively. The indigenous sample of 124 (42 males and 82 females) was monolingual English with a mean age of 35.7 years, and heterogeneous in terms of religious background (64% Anglican/Christian, 27% agnostic/atheist, and 9% Nonconformist/Catholic). The language spoken at home for the majority of these Asian Ss was English and the ethnic tongue in a mixed form. The Asians were asked for their choice of language (Hindi/Gujrati/Urdu/Punjabi/English) in which they would like to fill out the questionnaire. Seventeen Hindus, 27 Muslims and 23 Sikhs gave a preference for their ethnic tongue (Gujrati or Hindi, Urdu, and Punjabi, respectively) while 35 Hindus, 18 Muslims and 11 Sikhs opted for the English language. The mean ages and duration of British stay for those Asians choosing the ingroup language were 31.4 and 14.8 years respectively and those choosing English were 27.9 and 13.8 years respectively. (The differences were not statistically significant.) All the selfselected Asian groups were matched reasonably in terms of duration of stay in the U.K. and all samples were similar in terms of age and sex composition. Questionnaire and Procedure An extended version (40 items) of the Rokeach Value Survey (Ng et al., 1982) formed the English version of the questionnaire and through back translation procedures provided versions in Gujrati, Hindi, Urdu and Punjabi. In the questionnaire, Ss were requested to rate each of the 40 values (see Table 1) on 9-point scales (Ng, 1982). The Ss contacted were those who would be likely to co-operate in academic research as suggested by indigenous and Asian associates and community leaders in the different cities. The volunteer samples represent over 97% of those approached in this manner. The Ss were told that the purpose of the study was to discover what issues were important to people.

< previous page

page_75

next page >

< previous page

page_76

next page > Page 76

Results The data were computed first so as to provide rank orderings of values and as such were based on the comparative total score given to a particular value by that subject group. As can be seen from Table 1, 'family security' and 'honesty' were ranked highly by all groups (that is except for Sikhs regarding the former value in their own language and the latter one in English). 'Social justice' and 'happiness' were ranked highly by the indigenous group and by Asian Ss in English, and 'salvation', 'power' and 'imagination' were ranked lowly by all groups. The three highest rank values for the indigenous sample were 'happiness', 'honesty' and 'freedom'. 'Family security' was ranked first by Sikhs in English, and by Hindus irrespective of language choice. The highest ranked value for Muslims in English was 'social justice' with 'self-respect' and 'freedom' ranked second and third respectively. For Muslims Who chose their ethnic tongue, 'self-control' and 'clean' were ranked joint first, and 'responsible' third. The two highest ranked values for the Sikhs who chose their ingroup language were 'honesty' and 'responsible'. The data were then examined using canonical variate analysis (CVA) which provides insight into those values which would maximally differentiate the various groups as well as indicate the relative 'distance' between them. Of the six possible dimensions obtained, the first accounted for 46.99% of the total variance of the model and the remainder accounted for 13.96%, 13.61%, 9.58%, 9.24% and 6.62% respectively. Since the first three dimensions accounted for a high proportion of the total variance (74.56%), the results focus on these three. The relative similarity of the subject groups in this three-dimensional CVA space is shown in Figure 1 below. A well-defined pattern of results was obtained on the first dimension. The indigenous sample occupies a position at one end of this dimension, Asians who chose their own ingroup languages were at the other end with their counterparts choosing English in an intermediate position. The most important values which this dimension represents were 'cheerful' (0.218), 'forgiving' (0.379), 'helpful' (-0.318), 'obedient' (-0.193), and 'social recognition' (-0.189). Hence, the indigenous Britons regarded 'cheerful' and 'forgiving' more importantly than the Asian immigrants who when choosing their own language (and Muslims and Sikhs especially so) valued 'helpful', 'obedient' and 'social recognition' in high regard. Two other features of this spatial arrangement are also worthy of note as shown in Figure 1. First, it was the Sikhs who were the most aligned to the indigenous sample when they chose English as the language of testing. Second, while language choice affected values in all three groups, the effect was greatest for Sikhs and least for Hindus. The values which were most important on the second dimension were 'freedom' (-0.329), 'independent' (-0.295), 'loving' (-0.302), 'polite' (-0.353), and 'self-control' (0.382). Sikhs, irrespective of their language choice, regarded the first four values as more important than any other group, with Muslims choosing English holding 'self-control' in highest regard. Language choice had the maximum differentiating effect for Muslims, with Hindus minimally affected by this variable.

< previous page

page_76

next page >

< previous page

next page >

page_77

Page 77 Table 1 The hierarchical structure of values of the seven subject groups Groups** Values* Hn Mn Sn He Me 1.A comfortable life 39 24 36 22 11 2.An exciting life 27 38 40 37 30 3.A sense of accomplishment 18 28 35 18 36 4.A world of peace 23 25 3 7 13 5.A world of beauty 27 35 32 36 39 6.Equality 25 11 23 8 19 7.Family security 1 4 24 1 4 8.Freedom 12 29 3 4 3 9.Happiness 5 14 22 2 6 10.Inner harmony 2 14 29 6 19 11.Mature love 23 38 34 34 22 12.National security 21 17 27 33 27 13.Pleasure 16 28 24 30 28 14.Salvation 40 27 38 40 38 15.Self-respect 2 7 31 17 2 16.Social recognition 30 7 29 30 17 17.True friendship 5 18 11 10 9 18.Wisdom 11 29 18 15 4 19.Equity 22 4 15 22 33 20.Power 38 34 39 39 40 21.Social justice 12 21 20 4 1 22.Ambitious 16 21 3 21 10 23.Broadminded 32 11 27 11 33 24.Capable 27 4 16 26 25 25.Cheerful 15 18 24 19 25 26.Clean 5 1 9 12 13 27.Courageous 9 7 14 30 21 28.Forgiving 35 31 33 22 28 29.Helpful 9 11 7 12 17 30.Honesty 2 10 1 3 6 31.Imagination 37 40 37 38 37 32.Independent 20 35 6 20 22 33.Intellectual 25 31 16 22 32 34.Logical 30 37 19 28 30 35.Loving 33 33 8 12 6 36.Obedient 35 18 9 34 33 37.Polite 12 14 12 27 15 38.Responsible 8 3 2 9 11 39.Self-control 33 1 20 15 15 40.Self-determination 18 23 12 29 22 *The values were presented to Ss with their connotative meanings in parenthesis also. **Hn Hindus choosing the ingroup language Mn Muslims choosing the ingroup language Sn Sikhs choosing the ingroup language He Hindus choosing English Me Muslims choosing English Se Sikhs choosing English Bri (white) British

< previous page

page_77

Se 11 31 25 2 11 3 1 9 3 11 11 16 32 38 20 36 15 29 36 40 6 32 29 25 23 3 34 27 27 16 39 16 20 34 7 16 7 9 23 22

Bri 33 28 20 4 17 13 6 3 1 11 14 31 32 39 19 37 8 24 35 40 7 36 21 22 15 16 18 12 10 2 28 23 26 34 5 38 28 9 27 25

next page >

< previous page

page_78

next page > Page 78

Figure 1 The relative similarity of the subject groups in three-dimensional graphical represenation

< previous page

page_78

next page >

< previous page

page_79

next page > Page 79

On the third dimension, the most important values were 'courageous' (-0.469) and 'helpful' (-0.415). Sikhs who chose their own language regarded these values as more important than all other groups, but less important when they chose English for testing. Two other features of this spatial arrangement are worthy of note here. First, the indigenous sample was aligned more closely to the values espoused here when the Asian choice was their ingroup tongue than when it was English. Second, while Sikhs again find language maximally differentiating for value expression, Hindus are affected to a great extent too. A further CVA on just the three groups of Asian Ss without taking into account religion showed that the first dimension of the two dimensions possible accounted for 77.47% of the variance. The spatial arrangement thus provided (see Figure 2) remained very similar to the first dimension discussed above in that 'forgiving' (-0.390) and 'helpful' (0.302) were the most important values. It is evident here that the relative positions of Asians who chose English for testing in this pooled analysis emerged in a position that approaches the centre between the other two groups.

Figure 2 The relative similarity of the subject groups without 'religion' as a mediating factor in one-dimensional graphic representation Discussion The comparative analysis of the groups revealed that the indigenous white British group values 'forgiving', 'cheerful', 'independent' and 'freedom' whereas the Asian groups construe 'helpful', 'self-control', 'obedient' and 'social recognition' as relatively important. This difference seems broadly in line with the social structures to which the groups belong. For instance, the indigenous emphasis on 'forgiving' seems to be related to the Christian value and interestingly this item was ranked highly by white American (Christian) Ss in Rokeach's (1973) early study. As far as 'cheerful', 'independent' and 'freedom' are concerned, these are arguably individualistic values 'to decide for oneself' and are emphasised in egalitarian contexts of Western culture and education (Feather, 1975). The values espoused by the Asian groups are implicit in the emphasis placed in their cultures on group loyalty, family affiliation, strong kinship relations, interdependence, and conformity to social and traditional norms (e.g. Ballard, 1982; Ballard & Ballard, 1979; Kannan, 1978; Khan, 1979). Indeed, the sizeable and

< previous page

page_79

next page >

< previous page

page_80

next page > Page 80

continuous migration and settlement of Asian groups into Britian was possible only because of strong ingroup ties and societal values. Every member expects help from other members, and especially the family which is part of the backbone of Asian identity (Swann, 1985). Indeed, contact with the home family in their home countries provides Asians not only with a constant social support network but also a vital reference group which engenders difficulties for them in establishing and keeping contact with the host society. The culture itself inculcates a feeling of guilt in members if they do not help another member of their family, caste or creed (Kuppuswamy, 1975; McClelland, 1971). Moreover, help-giving and help-taking are considered legitimate acts in the religious ethics of these Asian groups (e.g. Anwar, 1979; Ballard & Ballard, 1979). The caste system, the role of women, respect for elders and arranged marriages are all hierarchical features which still prevail in Britain, albeit in a flexible manner (Bowen, 1981; Brooks & Singh, 1979) even amongst Sikhs (Dhanjal, 1976). 'Obedient', 'social recognition' and 'self-control' are thereby quite explicable in terms of Asian value systems. Turning now to the role of language choice in the findings, it is interesting to note that the societal values were ascribed more to Asians who undertook the task in their ethnic tongue than in English. It was the Sikhs who evinced the closest profile to the indigenous group's values. This pattern might exist for several (complementary) reasons. First, some components of the Khalsa culture and Western society are similar with respect to an emphasis on self-sufficiency and Sikhism's avowed stance on social equality (James, 1974; Mukherjee, 1982; Shackle, 1985). Second, Ballard & Ballard (1979) have found that Sikhs have had a good and continuing relationship with the British since 1857 in the British Punjab. Among other Asian migrants, Sikhs were the first to settle in Britain. The long exposure to British traditions, particularly the educational system (Helweg, 1979; Taylor, 1976), might be reflected in their value structure. Third, it also has been concluded that Sikhs adjust well to new and different situations (Ballard & Ballard, 1979). Brooks & Singh (1979) point out that Sikhs are changing their views on cultural customs such as arranged marriages and dowries in an appreciation of some British habits they consider worthwhile. Sikhism emerged some 500 years ago out of a quest for freedom and independence from the dogmatic and orthodox influences of Hinduism and Islam (Shackle, 1985). To establish an independent ethno-religious-linguistic identity out of political, economic and social hazards then requires a great deal of 'courage' and 'helpfulness' apart from a sense of 'freedom' and 'independence'. Mukherjee (1982) points out that Khalsa is an attempt to unify Hinduism and Islam into an aggressive, militaristic brotherhood. Still today in Britain, particularly if children are attracted to the British culture and life style (e.g. hair-cut), and start to devalue their own customs, parents try and immerse them in Sikh values through tales of heroic deeds and the chivalry of Sikh martyrs (Helweg, 1979). The emergent concept of soldier-saint suggests a simultaneous presence of bravery, valour and aggression on the one hand, and a working for welfare of others on the other. A combination of these values has indeed been reflected in the present study by Sikh values expressed in English and in Punjabi. This group's cultural flexibility (to which we have previously attested) might well foster the Sikh

< previous page

page_80

next page >

< previous page

page_81

next page > Page 81

bilingual's ability both to accommodate British values via English and traditional orientations via Punjabi. In addition, this analysis, together with the fact that they are arguably the most physically distinctive (e.g. turbans; see also, Bidwell, 1975) and cohesive of Asian immigrants in Britain (Helweg, 1979), provides some clues as to why they might be the most value-differentiating of the own-language-choosing Asian groups (cf. Tajfel, 1978). Language choice also had an influential effect on Muslim value expression. They too valued 'helpfulness' and 'self-control' when rating in their ethnic tongue. Maintaining kinship ties with relatives is a religious act for this group with self-control being necessary to lead a life according to the Quran. Indeed, Muslims segregate themselves from mainstream culture because of negative experiences of discrimination from the host society and the intermixing is likely to be detrimental to the maintenance of their religious and traditional practices (Jeffrey, 1976; Now-ikowski & Ward, 1979). The difference in their value structures in the ethnic and English language seem to be the outgrowth of existing in a strong ingroup environment of a minority culture which is embedded in the majority world. Results showed that the Hindus occupied a medium position in comparison with the Sikhs and Muslims to the extent that they afforded values such as 'helpfulness' and 'self-control' only moderate importance when choosing their ingroup language whilst affording those of 'freedom' and 'independence' a similar status when choosing the English language. This kind of ambivalence in value structure is supported by the basic characteristics of Hinduism. In comparison with the other two Asian religions, Hinduism is not, according to King (1981), a centralised religious tradition in the sense that an overall focus can not be pointed to. In other words, there is no one founder, no one central creed, nor one supreme religious leader. It has a long history of over 2,000 years with many gods and goddesses. Elasticity and tolerance is part of the ethos as illustrated by a picture of Jesus in the Leeds Hindu temple. Moreover, unlike Islam and Sikhism, more than one language is spoken by Hindu adherents (e.g. Punjabi, Bengali, Urdu, and many Southern Indian languages), even in Britain. Although its ethnic boundaries are therefore permeable (Banton, 1983), and doubtless perceived so to be (Giles & Johnson, 1986), this does not imply that Hindus court wholesale assimilation or value acculturation. As do the Muslims, Hindus in Britain feel that the host society rejects them and threatens their cultural existence (Bowen, 1981). Besides extending samples so as to investigate the roles of gender (Ahmed, 1984; Bhatti, 1976; Khan, 1976), age and generation (Kannan, 1978; Thompson, 1974), ethno-religious and national subjective identifications (Triandis, Kilty, Shanmugam, Tanaka & Vassiliou, 1972), and socioeconomic status (Young, Louw-Potgieter & Giles, 1986), future research might attend to elucidating the possible reasons for language choice in this and other contexts as different underlying motives are likely to yield consequentially quite different value structures. In this regard, an interesting empirical question would be to determine whether the same effects reported herein would emerge had these same Ss been forced to undertake the task in one or other of their languages. Our guess after engaging many of the Asian Ss in post-investigation discussion is that they would persist as there is doubtless a dialectical relationship between values and language choice.

< previous page

page_81

next page >

< previous page

page_82

next page > Page 82

Stable values, aspirations and modes of self-presentation may well have mediated language choice in our situation but doubtless continued preferences for certain languages in certain settings will also reshape values as well (Williams, 1979). Other interesting issues in more naturalistic settings would be to determine (i) the conditions under which Asian language (as well as dialect/accent) choice is identified as a marker of cultural values in intragroup situations by other ingroup speakers (Simard, Taylor & Giles, 1976), and more generally, (ii) not only the structure of ethnic values but to assess the social functions of their expression in different situations (Johnson & Giles, 1982; Tajfel, 1981), and (iii) the perception of in-/outgroup value differentials as a function of language choice and historical change amongst Asian, other immigrant groups, and the indigenous population. In sum, the results showed that Asian immigrants not only have quite a different value structure from the indigenous group but that important (and just as large) differences emerged between Sikh, Muslim and Hindu samples. Moreover, language choice had quite a profound effect in shifting value expression, the magnitude of which was arguably in the order of the ethno-religious groups just mentioned. Interestingly also, the nature of the differentiating effect of language choice for these groups depended upon the dimension examined. For instance, on the first CVA analysis, it had its only appreciable effect for Hindus on the third dimension and its largest effect for Muslims on the second dimension; indeed, this latter group was the only one for whom language choice had a really differentiating effect on all three dimensions. For Sikhs, language choice had no effect on the second dimension but its largest effect on the third. In any case, in this context with these Ss, an individual's language choice was, as predicted, a good indicator of cultural values expressed. Finally, it could well be that previous research which has demonstrated the role of outgroup language on value elicitation was not so much a language of testing effect per se as one in which Ss either fulfilled the demand characteristics of a forced linguistic choice or, under other intergroup circumstances, experienced reaction against a lack of any (naturalistic) sociolinguistic option. Acknowledgements Gratitude is expressed to the Association of Commonwealth Universities in the United Kingdom for funding the first author for a postdoctoral scholarship at the University of Bristol (1984-85), to W. Hickson and J. N. Singh for assistance in data collection, to R. Barot, N. F. Kapadia, B. Sandhu, and A. Singh, for their services in the back translation process, and to M. H. Bond, J. H. Crook and G. Jahoda for comments on an earlier version of this paper. References Ahmed, S. (1984) Cultural racism in work with women and girls, Conference Report on Women: Cultural Perspectives. London: Bedford College.

< previous page

page_82

next page >

< previous page

page_83

next page > Page 83

Allen, S. (1971) New Minorities, Old Conflicts: Asian and West Indian Migrants in Britain. New York: Random House. Anwar, M. (1979) The Myth of Return: Pakistanis in Britain. London: Heinemann. Bagley, C. and Verma, G. (1983) Introduction: Issues in multicultural education. In C. Bagley & G. Verma (eds), Multicultural Childhood. Hampshire: Gower. Ballard, C. (1979) Conflict, continuity and change: Second generation South Asians. In V. S. Khan (ed.), Minority Families in Britain: Support and Stress. London: Macmillan. (1982) South Asian families. In R. N. Rapoport, M. P. Fogarty & R. Rapoport (eds), Families in Britain. London: Routledge and Kegan Paul. Ballard, R. and Ballard, C. (1979) The Sikhs: The development of South Asian settlements in Britain. In J. L. Watson (ed.), Between Two Cultures: Migrants and Minorities in Britain. Oxford: Blackwell. Banton, M. (1983) Racial and Ethnic Competition. Cambridge: Cambridge University Press. Berry, J. W (1984) Multicultural policy in Canada: A social psychological analysis. Canadian Journal of Behavioural Science, 16, 354-70. Bhatti, F. M. (1976) Language difficulties and social isolation: The case of South Asian women in Britain. New Community, 5, 115-17. Bidwell, S. (1975) The Turban Victory. Southall: Sri Guru Singh Sabha. Bond, M. H. (1983) How language variation affects inter-cultural differentiation of values by Hong Kong bilinguals, Journal of Language and Social Psychology, 2, 57-66. Bond, M. H. and Cheung, M-K. (1984) Experimenter language choice and ethnic affirmation by Chinese trilinguals in Hong Kong. International Journal of Intercultural Relations, 8, 347-56. Bourhis, R. Y. (1979) Language and ethnic interaction. In H. Giles & B. Saint-Jacques (eds), Language and Ethnic Relations. Oxford: Pergamon. Bowen, D. G. (1981) The Hindu community in Bradford. In D. G. Bowen (ed.), Hinduism in England. Bradford: Bradford College. Brooks, D. and Singh, K. (1979) Ethnic commitment versus structural realities: South Asian immigrant workers in Britain. New Community, 7, 1920. Dhanjal, B. (1976) Sikh women in Southall: Some impressions. New Community, 5, 109-14. Eisenstadt, S. N. (1954) The Absorption of Immigrants. London: Routledge and Kegan Paul. Ervin, S. M. (1964) Language and TAT content in bilinguals, Journal of Abnormal and Social Psychology, 68, 500-507. Feather, N. (1975) Values and Education in Society. New York: Free Press. Ghuman, P. A. S. (1980) Bhattra Sikhs in Cardiff: Family kinship and organization. New Community, 8, 309-16. Giles, H. (ed.) (1977) Language, Ethnicity and Intergroup Relations. London: Academic. Giles, H. and Johnson. P. (1986) Perceived threat, ethnic commitment and inter-ethnic language behaviour. In Y. Y. Kim (ed.), Interethnic Communication: Recent Research. Newbury Park, CA: Sage. (1987) Ethnolinguistic identity theory: A social psychological approach to language maintenance, International Journal of the Sociology of Language, 68, 69-99 Giles, H., Scherer, K. R. and Taylor, D. M. (1979) Speech markers in social interaction. In K. R. Scherer & H. Giles (eds), Social Markers in Speech. Cambridge: Cambridge University Press. Helweg, A. W. (1979) Sikhs in England: The Development of a Migrant Community. Delhi: Oxford University Press. James, A. G. (1974) Sikh Children in Britain. London: Oxford University Press. Jeffrey, P. (1976) Migrants and Refugees: Muslim and Christian Pakistani Families in Bristol. Cambridge: Cambridge University Press. John, C., Young, L., Giles, H. and Hofman, J. E. (1985) Language, values, and intercultural differentiation in Israel, Journal of Social Psychology, 125, 527-29.

< previous page

page_83

next page >

< previous page

page_84

next page > Page 84

Johnson, P. and Giles, H. (1982) Values, language and intercultural differentiation: The Welsh-English context. Journal of Multilingual and Multicultural Development, 3, 103-16. Kanitkar, H. K. (1981) Caste in contemporary Hindu society. In D. G. Bowen (ed.), Hinduism in England. Bradford: Bradford College. Kannan, C. T. (1978) Cultural Adaptation of Asian Immigrants: First and Second Generations. Bombay: India Printing Works. Khan, V. S. (1976) Pakistani women in Britain, New Community, 5, 99-108. (1979) Migration and social class: Mirpuris in Bradford. In V. S. Khan (ed.), Minority Families in Britain: Support and Stress. London: Macmillan. Khera, A. K. (1981) The status of women in Hindu society. In D. G. Bowen (ed.), Hinduism in England. Bradford: Bradford College. King, U. (1981) Hinduism in a Western context. In D. G. Bowen (ed.), Hinduism in England. Bradford: Bradford College. Kuppuswamy, B. (1975) Social Change in India (2nd ed.). Delhi: Vikas. McClelland, D.C. (1971) Some themes of Hindu culture in India. In A. R. Desai (ed.), Essays on Modernization of Underdeveloped Societies (Vol. II). Bombay: Thacker. Marin, G., Triandis, H. C., Betancourt, H. and Kashima, Y. (1983) Ethnic affirmation versus social desirability: Explaining discrepancies in bilinguals' responses to a questionnaire, Journal of Cross-Cultural Psychology, 14, 173-86. Moore, M. (1976) A cross-cultural comparison of value systems, European Journal of Social Psychology, 6,249-54. Mukherjee, T. (1982) Sri Guru Singh Sabha: Southall. In A. Ohri, B. Manning & P. Cyrno (eds), Community Work and Racism. London: Routledge and Kegan Paul. Ng, S. H. (1982) Choosing between ranking and rating procedures for the comparison of values across cultures, European Journal of Social Psychology, 12, 169-72. Ng, S. H., Akhtar-Hossain, A. B. M., Ball, P., Bond, M. H., Hayashi, K., Lim, S. P., O'Driscoll, M. P., Sinha, D. and Yang, K. S. (1982) Values in nine countries. In R. Rath, H. S. Asthana, D. Sinha & J. B. P. Sinha (eds), Diversity and Unity in Cross-Cultural Psychology. Lisse: Swets & Zeitlinger. Nowikowski, S. and Ward, R. (1979) Middle class and British? An analysis of South Asians in suburbia. New Community, 7, 1-10. Penner, L. A. and Ank, T. (1977) A comparison of American and Vietnamese value systems, Journal of Social Psychology, 101, 187-204. Rex, J. and Moore, R. (1967) Race, Community and Conflict: A Study of Sparkbrook. London: Oxford University Press. Rokeach, M. (1973) The Nature of Human Values. New York: Free. Shackle, C. (1985) The Sikhs before and after Indian Independence, Asian Affairs, 14, 183-93. Simard, L., Taylor, D. M. and Giles, H. (1976) Attribution processes and interpersonal accommodation in a bilingual setting, Language & Speech, 19, 374-87. Smith, D. J. (1976) The Facts of Racial Disadvantage. London: Political & Economic Planning (PEP), No. 650. Swarm Report: Education for all (1985) A Report of the Community of Inquiry into the Education of Children from Ethnic Minority Groups. London: Her Majesty's Stationery Office. Taft, R. (1977) Coping with unfamiliar cultures. In N. Warren (ed.), Studies in Cross-Cultural Psychology, vol. 1. Tajfel, H. (ed.) (1978) Differentiation Between Social Groups. London: Academic. (1981) Social groups and social stereotypes. In J. C. Turner & H. Giles (eds), Intergroup Behaviour. Oxford: Blackwell. Taylor, J. H. (1976) The Half-way Generation: A Study of Asian Youths in Newcastle-upon-Tyne. Berks: NFER Publishing Co. Thompson, M. (1974) The second generation: Punjabi or English, New Community, 3, 242-48.

< previous page

page_84

next page >

< previous page

page_85

next page > Page 85

Tomlinson, S. (1984) Home and School in Multicultural Britain. London: Batsford. Triandis, H. C., Kilty, K. M., Shanmugam, A. V., Tanaka, Y. and Vassiliou, V. (1972) Cognitive structures and analysis of values. In H. C. Triandis (ed.), The Analysis of Subjective Culture. New York: Wiley. Williams, R. M. (1979) Change and stability in values and value systems: A sociological perspective. In M. Rokeach (ed.), Understanding Human Values: Individual and Societal. New York: Free Press. Young, L., Louw-Potgieter, J. and Giles, H. (1986) Values as functions of ethnicity and socioeconomic background: A cross-national study, Journal of Multilingual and Multicultural Development, 7, 253-67.

< previous page

page_85

next page >

< previous page

page_86

next page > Page 86

< previous page

page_86

next page >

< previous page

page_87

next page > Page 87

Language and Intergroup Perception in Spain Maria Ros and J. Ignacio Cano Universidad Complutense, Madrid and Carmen Huici U.N.E.D. Abstract With the background of the multilingual situation in Spain, which has undergone a very rapid change in the past decade, this study centres in the inter-group perception of social groups who share Castilian (Spanish) with their own ethnic languages, Catalan, Basque, Galician or Valencian. A sample of 165 university students answered several questions related to language competence, attitudes, social status and ingroup social identity. Using the matched-guise technique they also evaluated speakers representative of the five linguistic communities. Results show that the content of social categories (stereotypes) and language attitudes vary according to ingroup social identity and subjective vitality of their languages. Introduction The multilingual character of the Spanish state is defined by the coexistence of four languages: Castilian, Catalan, Basque and Galician. Valencian is often considered a geographical variety of Catalan though there is some disagreement on its status in relation to Catalan. The 1978 Constitution acknowledges Castilian as the national language of the state and grants the rest of the languages, formerly dialects, the coofficiality in their respective 'Comunidades Autónomas' (from now on C.A.). 1 The official status that C.A. languages enjoy today was not the norm in the recent history of Spain. In contrast, the people who speak these languages have been constantly beset by obstacles which frequently negated their right to use them. In fact, it is difficult to understand the revitalisation of their ethnolinguistic identities without a reference to the historical relations of these languages with Castilian. This relation has been and is still today, especially in the case of Galician and Basque, a diglossic relationship (Ferguson, 1959). Castilian was the only language of educa-

< previous page

page_87

next page >

< previous page

page_88

next page > Page 88

tion, administration and legislation, while use of the other languages was restricted to informal contexts like the family. Diglossia reached its peak after 1940. With Franco's victory, all languages except Castilian ceased to be official and were lowered to dialects. Education, administration and mass media were Castilianised. Furthermore, the use of minority languages in public was interpreted as a sign of separatism and treated accordingly. The policy of linguistic assimilation increased illiteracy in the minority languages. Socialisation patterns converged to the dominant language, particularly for those attempting upward mobility. The reduction of these languages to more rural, informal and uneducated contexts did not completely deter their use. Catalans were the most successful group in keeping their mother tongue as a language of certain prestige. This diglossic situation began to change slowly after the restoration of democracy in 1977. Political autonomy for 'Comunidades Autónomas' and the rising status of their languages started a new period of normalisation at all institutional levels. Their use has been increasing slowly but systematically in administration and mass media (Catalan, Basque and Galician have their own community television channels). Nationalistic movements have regained vitality and are expressing it through the linguistic dimension. However, this trend of normal language use still shows a certain imbalance in favour of Castilian at present. According to the model of sociolinguistic vitality (Giles, Bourhis & Taylor, 1977) we can draft the different vitality that these languages enjoy today by presenting an approximate outline of the situation. We think that their vitality has to be considered interdependently and in relation to Castilian (see Table 1). Table 1 Sociolinguistic vitality of C.A. languages of Spain Status Demography Institutional Support Castilian High High High Catalan High Med-High Med-High Basque Medium Low Medium Valencian Med-Low Medium Low Galician Med-Low Medium Low

Overall High Med-High Medium Med-Low Med-Low

Castilian is the national language of Spaniards. Widely used in most countries of Central and South America, it enjoys a remarkable position in status, demography and institutional support. Catalan is the native language of over six million people. It is spoken in all levels of society, including the cultural and economic realms. Understanding Catalan has become necessary for upward mobility. In Catalonia, Castilian is predominant among industrial immigrants who amount to over one and a half million people. However, most people who live in Catalonia identify themselves as Catalans and see this language as the basic dimension of their social identity (Strubell, 1981).

< previous page

page_88

next page >

< previous page

page_89

next page > Page 89

Valencian does not have strong vitality (Ros & Giles, 1979). It is predominantly used in rural areas; people in urban areas, especially Valencia city, speak mostly Castilian (Mira, 1981). Furthermore, as people ascend in their educational level or social status Castilian becomes predominant (Ros, 1984). Institutional support for the language is much weaker than in Catalonia. Some Valencians do not feel that their language is an important feature of their social identity. Others strive to define their 'psychological distinctiveness' (Giles, 1973) by enlarging the difference with Catalan and regard theirs as a unique variety. Basque, recently standardised around the 'Batua' variety, is spoken by 27% of the population of the Basque C.A. Native speakers are called 'euskaldunes' and they are concentrated mainly in rural areas. Industrialisation brought a mass of Castilian-speaking immigrants to the Basque C.A., which today constitute about 40% of the population. They are called 'eraldunes', that is, Basque people who can only speak Castilian. Although in the past Basque native speakers seeking social mobility spoke Castilian to their children, the changing status of the language has altered this trendthe young and more educated are now learning and speaking the language (Ruiz Olanbuenag et al., 1983). Institutional support for the language is growing, encouraging its use in primary and secondary education as well as in the mass media. Despite this rather diglossic situation in favour of Castilian, the Basque language seems to be the best symbol of their social identity (Ugalde, 1979). This positive 'psycholinguistic distinctiveness' is based on two aspects: their language is not shared by any other social group and it has unique historical roots as a very antique variety. Galician is spoken by approximately 83% of the population, who live mostly on agricultural and fishing resources. The majority of Galician speakers live in rural areas. Cities, especially the capitals of the four provinces, speak mainly Castilian. Vilariño (1981) points out a direct relationship between social class and language preference, Castilian having the higher status. Even though institutional support is slowly increasing the use of Galician in education, Castilian continues to be more instrumental. Favourable attitudes towards the community language use are not as strong as in other C.A. Research on the relation between language and secondary education reveals that teachers have better expectations concerning students' achievement when they come from urban, Castilian-speaking contexts than from rural Galician-speaking ones (Rojo, 1981). Moreover teachers consider that speaking Galician is a hindrance for learning. In short, we can see that the overall vitality of these languages is quite different; Castilian holds the dominant position, followed by Catalan, Basque, Valencian and finally Galician. We suspect that these various patterns of sociolinguistic vitality will differentially affect the role of language in their ethnic identity. Theoretical Framework Many social groups can be categorised readily by their distinct language varieties. Some ethnic and national groups also choose language as the most salient dimension of their social identities (Fishman, 1977; Ryan & Carranza, 1977; Ryan

< previous page

page_89

next page >

< previous page

page_90

next page > Page 90

& Giles, 1982). When we consider language as a salient dimension of intergroup relations in Spain, we acknowledge the existence of several ethnic groups that have their own ingroup language as well as Castilian, the national language of the state. These groups are redefining their statuses in a more favourable direction and expressing it by means of a wider use of their languages. Two theoretical frameworks seem most adequate for the explanation of language behaviour in our intergroup context: Tajfel's (1978, 1981) theory of inter-group relations and social change and Giles et al. (1977) theory of language in ethnic group relations. Stated in its simplest form, Tajfel's theory suggests that members of a group, in search of a positive social identity, compare themselves in a number of valued dimensions with members of outgroups. The aim of these intergroup social comparisons is to obtain ingroup distinctiveness through positive differentiation from the outgroup. Tajfel devotes much of his theoretical attention to groups which possess a negative social identity and to the means to restore it. Ingroup identity in Tajfel's theory is the result of a comparative process of the ingroup vis à vis outgroups in search of psychological distinctiveness. Even though the theory has proposed ingroup identity in relational terms, most measures of this concept have been done in an absolute way. Therefore ingroup identity has been often measured on its own, without considering the outgroup in relation to which ingroup identity has meaning. This strategy does not pose much of a problem when there is a single possible outgroup preselected in a binary way (e.g. men/women). However, when the concept is applied in an intergroup setting, where multiple comparisons can be made, it becomes more problematic and needs to be considered in a more relational way. When applied to the role of language in ethnic group relations, Giles et al. (1977) propose an integration of Tajfel's theory and Giles' theory of interpersonal accommodation through speech (Giles, 1973). The theory analyses the sociostructural factors which can influence whether an ethnolinguistic minority will or will not seek to establish its own ingroup language as a viable means of communication. These factors are status, demography and institutional support, all of which form the ethnolinguistic vitality of social groups. The theory is also concerned with the speech strategies of convergence and divergence when referring to an interethnic context. When members consider their inferior status as fair they will attempt to leave their ingroups converging socially and psychologically. They might do this by adopting the outgroup language, a strategy called 'linguistic convergence' (Giles, 1973). However, when groups in a subordinate position feel their status to be illegitimate, they will tend to dissociate themselves from the outgroup by stressing their social identity via language behaviour, a strategy called 'psycholinguistic distinctiveness'. By diverging or emphasising their own ethnic accent, dialect or language, ingroup members accentuate the differences between themselves and the outgroups on a valued dimension of their group identity (Bourhis & Giles, 1977; Bourhis et al, 1979). Research done on the social categorisation of groups whose language is the best symbol of their ethnicity has revealed that the content of social categorisation (e.g. stereotypes) depends on the social status that their speakers have as a group

< previous page

page_90

next page >

< previous page

page_91

next page > Page 91

in larger society (Giles & Hewstone, 1986). Therefore, speakers of standard varieties are ascended over speakers of non-standard varieties on status and prestige and devalued on interpersonal attraction (Giles & Powesland, 1975; Ryan & Car-ranza 1977; Ros, 1984). Moreover, when social categorisation is made salient between groups of different ethnolinguistic vitalityQuebecois versus English speakers in Canada (Lambert et al., 1960); Mexican-Americans versus English-Americans (Carranza, 1982) or Valencian speakers versus Castilian speakers in Valencia (Ros, 1984) there is a tendency among speakers of a language with low vitality to upgrade out-groups on status dimensions and ingroups on solidarity dimensions. Ingroup members compensate their inferior position by enhancing interpersonal attraction in this way. Past intergroup research has not studied the conjoint relation of subjective ethnolinguistic vitality of groups, social identity and intergroup social categorisation. Moreover, research on social categorisation in intergroup terms has been concerned mainly with one ingroup and one outgroup. Nevertheless, adopting Tajfel's concepts in a multiple group comparison, we may find that the notions of one ingroup and one outgroup are restrictive. It would seem quite reasonable, given the historical roots of the present inter-group situation, that Castilians should be the most prominent outgroup. However, we think that this would not hold true when ingroups have low ethnolinguistic vitality or do not perceive their language as the central dimension of their social identity. This would imply further consequences for the social evaluation of ingroup and outgroup members. Evaluations have been made via semantic differential scales. Our study was designed to test whether the central dimensions obtained in previous studies converge with trait dimensions obtained when subjects are allowed to respond freely. The Empirical Study Our research focused on the subjective ethnolinguistic vitality of five different 'Comunidades Autónomas' in Spain: Castile, Galicia, the Basque Country, Catalonia and Valencia. Through indirect measurement, we explored their language competence and their attitudes and values towards the two languages (Castilian, that is, Spanish, and their respective languages). Identification with the different social groups, objective and subjective social status and the saliency of language in their definition of their social identity were also considered. Using the matched-guise technique we elicited categorisation of the ingroups and outgroups, attempting to assess the content of these social categories through free responses. We also examined the affective reactions aroused by each language. A sample of 500 students divided by age (three subsamples of twelve, fifteen and eighteen-year-old subjects) was selected for the four bilingual C.A. and for Madrid. 2 Data will only be shown from the subsample of eighteen-year-old students. The sample consisted of a total of 165 university undergraduates (around

< previous page

page_91

next page >

< previous page

page_92

next page > Page 92

30 for each community) in the first course of History. Thus, they were matched reasonably regarding age, social and economic status. Subjects were asked to complete a questionnaire related to their language usage, attitudes towards this use, their linguistic skills in both languages, and their level of national identification, apart from various items about their perception of their linguistic group situation. Next, following the matched-guise technique (Lambert et al., 1960), a tape was played for the subjects and they heard the same weather forecast in each of the five languages. The order of the messages was randomised for every community. However, all subjects listened to their ingroup language in the last position. In addition, Basque never appeared first for it is the only language that is not understood outside its C.A. and, therefore, the meaning of the message could have been misinterpreted. The voices, though belonging to different speakers for the sake of native accents, had been matched in pitch and pace as far as possible. A small test among native judges confirmed all the voices as native. After hearing each of the five speakers, students filled in another short questionnaire where they were asked to identify the speaker's language and to describe in their own manner the people who speak that language. They also were provided with seven-point scales to rate the feelings evoked by each message (the adjectives were joy, pleasantness, fear, contempt, shame and laughter). These scales were placed in opposite evaluative directions to avoid automatic sets of responses. Subjects were not aware of the nature of the task before they came into the classroom. Results and Discussion Language and Social Identity: Basic Findings The matters covered in the first questionnaire included the following basic dimensions: (1) Use of Castilian and the C.A. language, (2) Linguistic competence in both languages, (3) Attitudes towards the use of language: language preference, (4) Identification with the ingroup and various outgroups, (5) Instrumental and integrative value accorded to each language, (6) Feelings aroused by the C.A. language, and (7) Perceived status of the ingroup, both from the perspective of the respondent himself and from the outgroups'. In some cases single questions were used to measure the dimension, while in some others several questions were averaged to compute global indexes. An index to account for the level of C.A. ingroup identification was constructed. The item directly asking ingroup identification, however, turned out to have a high mean for practically all cases and a small standard deviation, which did not enable us to correlate it with other variables. Therefore, we constructed an index (Idausus) by subtracting the identification with Spain from the original identification with the C.A. ingroup. 'Idausus' was used as a measure of identification, always allowing for the fact that it was not a direct standard but a subtractive one. The way the scales were presented, one after the other, leads us to believe that the number

< previous page

page_92

next page >

< previous page

page_93

next page > Page 93

marked on each scale acquires its full meaning in comparison with the others. The use of Spanish was predominant over the C.A. language, while attitudes point to a clear preference for the community language. This apparent contradiction can be explained partly by the significantly better knowledge people claim to have of Spanish (F = 41.88; df = 1,144; p<0.001). Competence in the community language is perhaps not good enough to allow further use (which respondents seem to wish). Subjects identified more with their community than with Spain (F = 70.5; df = 1,166; p<0.001). They reported a greater integrative value for their language than for Spanish (F = 105.38; df = 1,141; p<0.001). The feelings evoked by the community language were systematically positive. The valuation of one's own community was also very high, significantly higher than the perceived outgroup evaluation of one's own group (F = 111.89; df = 1,165; p<0.001). This is confirmed by the rating that subjects expressed about the fairness or unfairness of this valuation (mean = 1.63; 1 = fair, 2 = unfair). This suggests that communities feel underestimated as a whole. The change in this underestimation, if it is ever to take place, should be undertaken in a collective way according to the subjects (mean = 1.86; I = individually; 2 = collectively), 85% of them expressing that this change will come or at least may come in the future. A factor analysis of the contexts of language use and the contexts of attitudes towards the use was carried out and the results showed three factors: (1) formal contexts of language use; (2) formal contexts of attitudes; and (3) familiar contexts both of use and of attitudes. Two conclusions can be drawn from this. The first is that familiar and formal contexts form two quite different blocks where linguistic use and preference differ considerably. The second is that while in formal contexts, the use and the attitudes toward language seem to follow different paths, both dimensions (usage and attitude) are apparently very close when in familiar contexts. In simpler words, people seem to speak the language they want to when they are at home, but not in other situations. Another factor analysis was executed using responses to questions on identification with various groups to determine the extent they were considered ingroups. We also included the subtractive index we had created (idausus). Two factors were obtained, one, again defined rather by familiar and informal indexes (family, friends) and the other directed towards wider political and social groups. The next step was to examine the structure of the seven main dimensions to see how they covaried with one another. The results of the principal factors confirmed our expectation and revealed a close relation between most of the dimensions. Mastery of C.A. language, use of it, preference for its use compared to Castilian, positive attitudes aroused by this language and identification with the community all saturated very high in the first factor. The nearest correspondence is found indeed between the linguistic preference of using C.A. language and the subtractive identification (r = 0.74), between the feelings evoked by the C.A. language and the subtractive identity (r = 0.58) and between the feelings and the attitudes of preference (r = 0.73). These three variables form what we might call an affective-attitudinal dimension regarding the ingroup and its language. The other two variables, mastery and use, lead us to think of a more objective behavioural

< previous page

page_93

next page >

< previous page

page_94

next page > Page 94

dimension whose relation with the first one is not as close because it also depends on external factors. We can then picture an image of members strongly identified with their ingroup who know and use their language, preferring it to Spanish, and who accordingly experience positive feelings when listening to it. At the other extreme we would have members not very identified with the ingroup, who neither master nor use their language, and who would rather speak Spanish. It should be noted that the correspondence between the evaluation of one's own group (valorsub) and the level of identification with it is not very high (r idausus-valorsub = 0.17; r idau 3-valorsub = 0.40). Thus, it is not uncommon to find individuals who, despite their clear identification with the ingroup, acknowledge its inferiority (at least on some dimensions). The second factor represents the values accorded to the language, both instrumental and integrative. While the integrative value had a moderate correlation with the first factor, the instrumental one seems to rely on other external causes encountered in reality. There is a significant relationship (F = 4.59, df = 1,161; p<0.05) between the degree of ingroup identification (Idau) and whether respondents think the other groups fairly assess theirs. The greater the identification with the ingroup, the more underestimation perceived. Finally, a regression analysis was carried out, where the attitudes or preference for either language was used as the dependent variable. The independent variables were linguistic mastery, use of the language, feelings when hearing it and subtractive identification. The preference for the language to use depended first on the level of identification (subtractive) with the ingroup (Adjusted R2 = 0.56), second on the feelings aroused by the local language (Adjusted R2 = 0.64 for both variables) and third by the actual use of the language (Adjusted R2 = 0.69 for the three variables). We find here again a relative discrepancy between the preference for a language and the actual use of it. The level of identification determines the preference for the language, but not the actual use. With the use of language as the dependent variable, the identification was not included in the regression equation due to its lack of significance (the stepwise method used rejects variables when their F is smaller than 3.84 or its p higher than 0.05). Language and Social Identity: Profile of Each Community From the point of view of linguistic competence in C.A. languages, Catalans had a significantly higher degree of competence than Valencians or Galicians with the Basques in the middle of these two extremes (F = 8.81; df = 3,141, p<0.001). The specific means were: Galicians = 4.06, Catalans = 5.40, Basques = 4.73, and Valencians = 3.60 (1 = no idea, 7 = perfect mastery). This description of linguistic skills is in accordance with the frequency of language use they claim to have. Catalans seem to use their language significantly more often than all the other communities, be it written or oral usage. Basques appear once again between the pole of Catalans in one direction and the pole of

< previous page

page_94

next page >

< previous page

page_95

next page > Page 95

Valencians and Galicians in the other (F = 15.44; df = 3,141; p<0.001). The specific means were: Valencians = 1.49, Galicians = 1.55, Basques = 1.84, and Catalans = 2.37 (1 = only Spanish, 3 = only the C.A. language). The difference between Catalans and the other groups increases when we focus on written use, not because Catalans write more often in their language than they speak it, but because writing Basque, Galician and Valencian seems to be clearly less generalised than speaking them (F = 31.09; df = 3,141, p<0.001). The specific means were: Galicians = 1.13, Valencians = 1.17, Basques = 1.28 and Catalans = 2.21. Catalans are also the only group which uses its language orally more frequently than Spanish, the others using Castilian in far many more contexts than their own language (F = 5.87; df = 3,141, p<0.001). The specific means were: Galicians = 1.57, Valencians = 1.60, Basques = 1.84 and Catalans = 2.21. As for the attitudes towards the usage (preference), the situation changes. It is now Basques who favour most the use of their language compared to Spanish, Catalans in second place. These two groups together with Galicians advocate the use of their language rather than Spanish, while Valencians are not clearly inclined to one or the other. The differences in this dimension are significant between the two extreme groups, Basques and Valencians (F = 6.55; df = 3,141; p<0.001). As far as the feelings aroused by the C.A. languages are concerned all bilingual communities expressed positive ones towards their ingroup language. Catalans showed the least positive feelings this time while the Basques marked the most positive ones almost in every occasion. The next dimension where groups differ is the integrative and instrumental value accorded to the languages. All groups except Valencians assessed their language as significantly more integrative than Spanish (F = 105.38; df = 1,141; p<0.001). However, only Catalans and Basques expressed that their respective language has more instrumental value than Spanish, and the difference is significant only for the first case (F = 14.10; df = 1,39; p<0.001). Thus, Catalans were the only community to express that their language is significantly more integrative and more instrumental than Castilian. The other groups conferred Spanish a higher instrumental value though not significantly. The five communities displayed roughly the same distribution when they were asked the degree to which their language was integrative or instrumental. On the other hand, there was a wide dispersion when they were asked the same questions for Castilian. Catalans, for instance, were of the opinion that Spanish is neither really integrative nor truly instrumental. The various groups do not differ in the degree of identification with the ingroup, which is extremely high in most cases, so a new index of subtractive identification (identification with C.A. minus identification with Spain) was developed, as explained before. This subtractive variable shows a dramatic difference between Basques and the remaining groups (F = 16.28; df = 4,162; p<0.001). The specific means were: Castilians = 0.34, Valencians = 0.35, Galicians = 1.23, Catalans = 2.65, and Basques = 4.24, (minimum = -6, maximum = +6). A similar but inverse discrepancy is found in the identification with Spain. The Catalan group

< previous page

page_95

next page >

< previous page

page_96

next page > Page 96

follows Basques and both of them are then significantly more identified with their community (subtractive identification) than some other groups below like Valencians and Castilians. Being Catalan or Basque seems to mean being less of a Spaniard. Castilians are the only ones to be more identified with Spain than with their C.A. (F = 23.70; df = 4,162; <0.001). The specific means were: Galicians = 4.63, Catalans = 3.31, Basques = 2.22, Valencians = 6.04, and Castilians = 6.00 (minimum = 1, maximum = 7.0). Speaking the C.A. language was mentioned as the main feature of the ingroup self-definition by Catalans and Basques. Living in the community was also cited as important by these two groups, but neither of these differences with the other groups was significant. Having been born in the region was the crucial characteristic for Galicians and Valencians and it was also important though not predominant for Catalans. On the other hand, Basques conferred a significantly smaller relevance to this factor as definition of their identity (F = 4.44; df = 3,136; p<0.01). Speaking Spanish was considered an irrelevant component of their identity by all the bilingual groups. Speaking both languages was judged a more important constitutent of their belonging to the group though not a prominent one. Valencians grant this characteristic a significantly higher position than the other groups (F = 5.42; df = 3,133; p=0.01), thus confirming their neutral position as regards the use of Castilian or their local language. It must be noted that speaking the C.A. language was a more relevant factor than speaking Castilian for all four communities; the differences ranging from the case of Basques (the highest difference between both languages) to that of Valencians (the lowest). We can then draw the conclusion that the linguistic identity therefore was based on the local language rather than on Castilian. The analysis of the last dimension, the perception of the ingroup social status indicated that Catalans and Castilians share the belief that they are reasonably well valued within the Spanish state, in a degree which is significantly higher than that of the other groups. Galicians at the other extreme are convinced of their general devaluation by the outgroups (F = 15.32; df = 4,162; p<0.001). The specific means were: Galicians = 2.13, Basques = 3.51, Valencians = 3.57, Catalans = 4.47, and Castilians = 4.93 (minimum = 1, maximum = 7). Parallel to this, Galicians disagreed most with this perceived status, while Castilians and Catalans had self-evaluations which were nearest to the external evaluation they claim to receive. The difference between these latter and Galicians or Valencians is statistically significant (F = 7.75; df = 4,161; p<0.001). The specific means were: Castilians = 0.65, Catalans = 1.00, Basques = 1.91, Valencians = 2.46, and Galicians = 3.31, (-6 = highest perceived social evaluation compared to one's own, +6 = lowest perceived social evaluation compared to one's own). Summarising, the profile of each linguistic group would come out as follows. Catalans have a definite mastery over their language and they use it very often, in fact more than Spanish. Their attitude towards this usage is consistent with it, preferring Catalan to Castilian. The feelings evoked by their language are not particularly positive, probably because they are accustomed to using it every day and it is, therefore, nothing extraordinary for them. Catalan seems to be integrative and also instrumental in their opinion, while Castilian is not. This confirms that

< previous page

page_96

next page >

< previous page

page_97

next page > Page 97

Catalan is established firmly in most social contexts. They show quite a high level of identification with Catalonia and they affirm that speaking Catalan is important for the Catalan identity. They do not feel underestimated by the other groups. Basques do not have as high linguistic skills as Catalans and they do not use their language as often, especially for writing. Spanish is still predominant, partly due to its instrumental value. On the other hand, their attitude towards the use of languages is the most radical one and they see the Basque identity as based on speaking the language, which they do not always do since they have not mastered it. This 'unsatisfied militant' attitude is accompanied by a great distance from Spain (and accordingly from Spanish) and by very positive feelings aroused by their language. A language with which they are not as familiar as the Catalans are with theirs. Galicians have a narrower knowledge of their language than the other two groups and they use it much less often (very seldom when writing). Their preference for their language is also milder than in the case of Catalans and Basques and they see Spanish as slightly more instrumental. Their level of subtractive identification is in the middle, but they still have a higher identification with their community than with Spain. Their most distinctive feature is their conviction of being severely underestimated by the other groups. Valencians are the group that has the least knowledge of their language. Its use is not very widespread, at a level similar to Galician. They show no inclination for Valencian or Castilian and their identity is not significantly related to their language, which they judge slightly less instrumental than Spanish. Their identification with Valencia is practically the same as their identification with Spain. In short, they do not have a strong linguistic identity for they do not have a strong Valencian identity. They also feel that they are underestimated. Castilians are the most peculiar group in the study because they are not bilingual and they share their only language with the other groups. Their identity is understood best in a Spanish, rather than a Castilian sense. They favour the use of Spanish, even in other C.A. They have no special feelings when hearing Spanish, possibly because they have no other language to compare it with. They do not appear to have a linguistic conflict to produce a differential identification. They believe they are well considered in other parts of the state. C.A. Groups Stereotypes The descriptions of C.A. groups generated from the free answers of the subjects after listening to the five recorded messages were content analysed according to a system of 70 categories generated for this purpose. This category system was exhaustive and all categories were mutually exclusive. The unit of analysis was each phrase which expressed an independent idea. Every unit was also rated in terms of its evaluative aspect as positive, negative or neutral. Once the content analysis was performed, only categories that had been used several times in the description of one or more groups were retained. Others were collapsed into larger categories according to similarity in theoretical terms. Thus the number of

< previous page

page_97

next page >

< previous page

page_98

next page > Page 98

categories was reduced to build a summarised system, which varied slightly for each particular analysis. In the case of the analysis we are going to present now the categories were as follows: 1. Competence and achievement. Intellectual (e.g. intelligent-stupid), cultural and professional. 2. Sociability. Relational traits (e.g. rough-gentle; nice-unpleasant; warm-cold). 3. Personality. Personality traits not included in the previous category. 4. Ambitious/Active Passive. 5. Morality. Traits that involve a clear moral evaluation (e.g. mean-generous; selfish-altruist; honest-deceitful). 6. Sociopolitical attitudes. For example: conservative, centralist, progressive. 7. Sociodemographic characteristics. For instance, young, old, rural, urban. 8. Subjective evaluation. This category reflects: a) the subject's evaluation (acknowledged by the subject himself as his personal opinion) towards C.A. groups and their language. b) explicit support or criticism of the group in its attempt of self-assertion. c) emotional reaction towards the group. d) references to personal experiences with the group. 9. Characteristics of the language. Language evaluation and description. Language diffusion. Comparison to other languages. 10. Identity. National identity or identification. Cultural identity. Active defence of language or culture. Language as a sign of identity. 11. Lack of identity. Explicit mentioning of the absence of the aspects referred to in the previous category. 12. Melting pot. Description of the group as a mixture of different types of people. 13. Evaluation by society. Evaluation that the group receives by society at large. 14. Ethnocentrism. Reference to outgroup rejection. 15. Group cohesion. Reference to intragroup solidarity. 16. Others. As it can be seen in Table 2 the three categories most frequently used to describe C.A. groups are 'Sociability', 'Personality' and 'Identity'. These three seem to be the most salient for group description. Other dimensions found are 'Competence', 'Morality', 'Characteristics of the Language' and 'Sociopolitical Attitudes'. The images of the five linguistic groups obtained in this study, which should be considered as tentative, are as follows. Galicians are viewed as politically conservative, rural, lacking in competence and very sociable. They are positively assessed in moral traits and perceived as having cultural identity. Catalans are described as competent, but unsociable. They are seen as unkind and they are thought to feel superior. They are also viewed as stingy. Other dimensions used to define Catalans are a strong group identity and a tendency to reject outgroups.

< previous page

page_98

next page >

< previous page

next page >

page_99

Page 99 Table 2 C.A. groups descriptions Categories

Galician

Catalan

Groups Basque % E 13 (1.62) 50 (2.39) 46 (1.84) 2 (1.33) 10 (1.08) 14 4 13 (2.78) 14 (2.12) 36 (1.91) 1 2 2 (2.33) 5 8 (2.05)

Valencian

%1 E2 % E % Competence 23 (2.19) 29 (1.25) 11 Sociability 57 (1.92) 48 (2.5) 45 Personality 52 (1.99) 49 (1.84) 47 Ambitious 14 (1.87) 9 (1.76) 2 Morality 18 (1.34) 33 (2.41) 15 Sociopolitical 29 13 6 Sociodemographic 24 5 5 Subject evaluation 14 (2.19) 12 (1.75) 20 Characteristics of language 13 (1.71) 15 (2.12) 18 Identity 31 (1.85) 40 (1.64) 28 Lack of identity 6 2 4 Melting pot 1 2 1 Evaluation by society 3 (2) 7 (2.33) 2 Ethnocentrism 1 16 6 Group cohesion 1 (2) 5 (1.83) 2 1. % of subjects using the category. 2. Evaluation (in parentheses) ranges from 1 = positive through 2 = neutral to 3 = negative.

< previous page

page_99

E (1.5) (1.77) (1.57) (1.5) (1.83) (2.11) (2) (1.85) (2) (2)

Castilian % 14 39 41 1 9 14 5 14 20 14 3 17 3 8 1

E (1.6) (1.99) (1.99) (1.6) (1.73) (1.88) (2.08) (1.90) (1.65) (1)

next page >

< previous page

next page >

page_100

Page 100 Table 3 Dimensions of intergroup categorisation Dimensions Castilians Catalans Group Galicians Basques Sociability = +* Personality = + + =* Identity +* + + Competence + +* + Morality = -* + + High ethnolinguistic Low ethnolinguistic vitality groups vitality groups + Positive evaluation = Neutral evaluation - Negative evaluation * Group with the highest frequency of occurrence in the dimension

Valencians + + = = +

Basques are described negatively in sociability traits, probably because of the frequent attribution of the trait 'aggressive'. On the other hand, they receive a better evaluation in other personality features. The other common dimension used to picture the Basques is their identity. The image of Valencian speakers consists mainly of sociability and other positive personality traits. In fact, this group receives the best evaluation in these two salient dimensions. Castilian speakers have a rather neutral image both as regards personality and sociability. Accordingly their outlook is that of a group conceived as a mixture of various kinds of people, i.e. a melting pot. A synthetic picture of the five C.A. groups is presented in Table 3. We show the main dimensions combined with the evaluation that accompanies them when used in relation to the different groups. In this table the status of groups in ethnolinguistic vitality is also taken into account. When we consider the status dimension in ethnolinguistic terms, we see that our results are in accordance with other studies in the area of intergroup relations: the superiority of high status groups in competence, and that of lower status groups in sociability and moral characteristics. Additionally, two of the high status groups in our study are viewed as superior in a highly valued dimension: identity. At present, within the Spanish context of transference of power from the central state to the 'Comunidades Autónomas', national identity is often held as an important dimension for the negotiations between the central state and the different C.A. Conclusions When we consider language as the salient dimension of intergroup relations in Spain, we acknowledge the existence of several social groups that speak their

< previous page

page_100

next page >

< previous page

page_101

next page > Page 101

ingroup language, as well as Castilian. Two factors, ingroup identity and ethnolinguistic vitality, seem to account for the intergroup relations that exist. Social identity theory has focused attention on polarised social situations characterised by one ingroup and one outgroup, while many social settings allow for multiple comparisons between groups, and the same outgroups may serve very different functions. In our study all subjects seem to identify strongly with their ingroup. However, subtractive identity seems a better relational measure to reflect ingroup identification in this context. Through the use of this index different patterns of relation to the Spanish group seem to emerge: Basques and Catalans dissociate more extremely, while Valencians and Galicians appear less distant. Furthermore, for this latter group Spaniards do not appear as an outgroup at all. Finally, Castilians are the only group to identify more with Spain than with their ingroup. Therefore, the same group can have different meanings in a multiple intergroup context: for some groups Spain becomes a salient outgroup, while for other groups it is a supracategory where the ingroup itself feels included. Interestingly, this pattern clearly correlates with the importance groups attach to their ingroup languages as part of their social identities. As for subjective vitality, the results show that the use of Castilian (Spanish) is predominant over the other standard languages. However, the attitudes and integrative value attached to the use of Castilian point to a clear preference for the C.A. language. Three variables seem to be highly intercorrelated, subtractive identification, feelings towards the language and attitudes of preference for ingroup language (an affective-attitudinal dimension referred to identification and language). One end of this dimension consists of members strongly identified with their ingroups, who know and use their language rather than Castilian, and experience positive feelings when listening to it. At the other extreme, there are members who do not identify strongly with their ingroup, who neither master nor use their language and who would rather use Castilian. Subjective status of the ingroups and identification are not correlated. Thus, it is not uncommon to find ingroups perceiving their status as inferior. When it comes to group evaluation by society, the more subjects identify with the ingroup, the more they feel underestimated by the rest of society. There seems to be no direct relation between this perception and the choice for individual or social strategies of change. When we consider the language profiles regarding subjective vitality a clear picture seems to emerge. Castilian has the highest overall vitality regarding competence, actual use and instrumental values. Catalan is quite close to the former on these dimensions. Basque comes next since it is not as widespread as Castilian in the Basque Country. However, Basques have the most radical favourable attitude towards the use of the ingroup language. Finally, Galician and Valencian varieties have a more restrictive use and less instrumental value in relation to Castilian. Taking into account both factors, social identity and language vitality, the groups seem to cluster in the following way: Catalans are simultaneously high on both dimensions, Basques have a medium vitality and high identity while Gali-

< previous page

page_101

next page >

< previous page

page_102

next page > Page 102

cians and Valencians are low on both dimensions. When intergroup categorisation is carried out on the bases of the language dimension, the subjects seem to give a comprehensive yet rather accurate view of the situation. Castilians, the only non-bilingual group, have a low identity and a high language vitality. However, this vitality should not be attributed directly to the Castilian group since their language is not restricted to one C.A., but used nationwide. They tend to upgrade those groups with high vitality and/or high identity (Castilians, Catalans and Basques) in competence and to some extent downgrade the same groups in sociability. While the opposite trend seems to happen with respect to groups low in vitality and identity (Galicians and Valencians), which are conversely highly valued in moral traits. As it was mentioned before, general results in the studies concerning relations between groups of different status show that high status groups are characterised in terms of competence and economic success and lower status groups in terms of warmth and interpersonal attraction. One could interpret that reinforcing this mutual image works in favour of maintaining the status quo of the intergroup structure, while allowing some dimensions of positive distinctiveness for the inferior groups involved. The basic dimensions (competence, personality, sociability and morality) found in other studies (Giles, 1973; Ryan, 1979) also tend to emerge when content analysis of free responses is used. Acknowledgement This research was conducted with funding from the C.A.I.C.Y.T. (Ministry of Science and Education). Notes 1. 'Comunidades Autónomas' are the present semi-federal divisions of the Spanish state, some of which also have a language of their own as we have seen. 2. Madrid was taken as the representative of Castile for practical reasons. Although it constitutes at present a 'Comunidad Autónoma' by itself, it has been historically the capital and the geographical centre of Castile. 3. 'Idau' is the direct index of identification, not subtractive. References Bourhis, R. Y. and Giles H. (1977) The language of intergroup distinctiveness. In H. Giles (ed.), Language, Ethnicity, and Intergroup Relations. London: Academic Press. Bourhis, R. Y., Giles, H., Leyens, R. and Tajfel, H. (1979) Psycholinguistic distinctiveness: Language divergence in Belgium. In H. Giles & R. St. Clair (eds), Language and Social Psychology. London: Blackwell. Carranza, M. A. (1982) Attitudinal research on Hispanic language varieties. In E. Ryan & H. Giles (eds), Attitudes towards Language Variation. London: Edward Arnold.

< previous page

page_102

next page >

< previous page

page_103

next page > Page 103

Ferguson, J. (1959) Diglossia. Word, 325-40. Fishman, J. (1977) Language and ethnicity. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Giles, H. (1973) Accent mobility: A model and some data, Anthropological Linguistics, 15, 87-105. Giles, H., Bourhis, R. and Taylor, D. (1977) Towards a theory of language in ethnic group relations. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Giles, H. and Hewstone, M. (1986) Social groups and social stereotypes in intergroup communication: A review and model of intergroup communication breakdown. In W. B. Gudykunst (ed.), Intergroup Communication. London: Edward Arnold. Giles, H. and Powesland, P. (1975) Speech Style and Social Evaluation. London: Academic Press. Lambert, W. E., Hodgson, R., Gardner, R. C. and Fillenbaum, D. (1960) Evaluational reactions to spoken languages, Journal of Abnormal and Social Psychology, 60, 44-51. Mira, J. (1981) Població i llengua al Pais Valenciá. Valencia: Institutió Alfons el Magnanim. Rojo, J. (1981) En torno alas actitudes linguisticas de los profesores de E.G.B. en Galicia, Revista de Educacion, 268, 131-56. Ros, M. (1984) Speech attitudes to speakers of language varieties in a bilingual situation. In M. Ros & M. Strubell (eds), International Journal of the Sociology of Language, 47, 73-90. Ros, M. and Giles, H. (1979) The Valencian language situation: An accommodation perspective, I. T. L.: Review of Applied Linguistics, 44. Ruiz Olabuenaga, J. I. et al. (1983) La lucha del Euskera: conocimiento, uso y actitudes. Zarautz: Gazteiz. Ryan, E. (1979) Why do low-prestige varieties persist? In H. Giles and R. St. Clair (eds), Language and Social Psychology. Oxford: Blackwell. Ryan, E. and Carranza, M. (1977) Ingroup and outgroups reactions to Mexican-American language varieties. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Ryan, E. and Giles, H. (eds) (1982) Attitudes towards language variation. Social and Applied Contexts. London: Edward Arnold. Strubell, M. (1981) Llengua i població a Catalunya. Brna: La Magrana. Tajfel, H. (ed) (1978) Differentiation between Social Groups. Studies in the Social Psychology of Intergroup Behaviour. London: Academic Press. (1981) Human Groups and Social Categories: Studies in Social Psychology. London: Cambridge University Press. Ugalde, M. (1979) Conflicto lingüistico en Euskadi. Bilbao: Ediciones Vascas. Vilariño, P. (1981) Las Reivindicaceones Autonomicas en España. Analisis por Regiones. Sada: Edicios do Castros. Vilariño, P. and Sequerias, J. L. (1982) Parametros y procesos básicos de la sociedad gallega, Revista Internacional de Sociologia, 44, 497-512.

< previous page

page_103

next page >

< previous page

page_104

next page > Page 104

< previous page

page_104

next page >

< previous page

page_105

next page > Page 105

Imposed Identity and Linguistic Strategies Joha Louw-Potgieter Psychology Dept., University of Natal, Durban, South Africa and Howard Giles Psychology Dept., University of Bristol, U.K. Abstract The aim of this paper is to investigate intragroup variability and its effect on social identity. Based on assumptions of power differentials within groups, and the attempts of the relatively more powerful to impose their notion of group membership on the relatively less powerful, a set of theoretical propositions regarding identity and dissent is formulated. These propositions deal with instances of congruence between the notions of identity held by the powerful and the powerless, instances of incongruence between the notions of these two subgroups, and possible reactions of the powerless in the latter instances. The role of language in these attempts at identity imposition, and reactions to it, is examined in detail. As identification with a group precedes dissent within it, any study of intragroup variability and (by implication) dissent should also deal with group identity (Louw-Potgieter, 1988). Following Tajfel and his co-workers on identification (Tajfel, 1978; Tajfel & Turner, 1979; Turner, 1982), we agree that subjective feelings of group belongingness are essential for identification with a group. Social identity is 'that part of an individual's self-concept which derives from his knowledge of his membership of a social group (or groups) together with the value and emotional significance attached to that membership' (Tajfel, 1981: 255). Apart from self-definition as a group member, authors acknowledged that the perception of others (especially other group members) might play a part in a person's experience of his/her group membership, and the issue of group boundaries in terms of internal or external criteria of group belongingness was introduced (Breakwell, 1979; Tajfel, 1978). External criteria refer to 'objective' standards (e.g. skin colour, fluency in the group's language, etc.), while internal criteria could be regarded as the concepts members hold of what constitutes legitimate membership (cf. Fishman, 1977). Despite acknowledging the roles of self and other in the identification process, the impression might still be that a person has relative freedom of choice regarding identification with a group. This ignores the fact that some groups, or some group

< previous page

page_105

next page >

< previous page

page_106

next page > Page 106

members, possess more power than others and, by virtue of this power, can impose their notion of identity upon the less powerful (Billig, 1976; Deschamps, 1982; Guillaumin, 1972). As .identity has been described as a multidimensional, dynamic process, it could also be that the powerful's definition of identity might fluctuate or change over time. When this happens, group members adhering to the old definition might discover that they are no longer regarded as part of the ingroup, while former outgroup members might find that they are unexpectedly perceived as new ingroup members. Following from these notions of imposition of and change in identity, one could envisage instances where the relatively less powerful might: (1) agree with and embrace the identity imposed on them (congruence); or (2) disagree with and reject the identity imposed on them (incongruence). This imposition and acceptance/rejection of identity might occur on an inter- or intragroup level. As our concern is with dissent within a high status group, this paper will focus on intragroup processes. In instances of incongruence, a relatively less powerful group member might regard self as a group member, but might not be accepted as such by the dominant or powerful within the group. In cases such as these, where group members experience conditions conducive to staying in the group, Tajfel (1978) predicted an intensification of actions to ensure the future superiority of the group: enhancement of existing distinctions, creation of new distinctions, etc. Empirical evidence indicated that a high status group, perceiving its position as both illegitimate and unstable, adopted strategies of creativity rather than direct competition (Turner & Brown, 1978). Condor & Abrams (1984), in their study on women's identity, indicated that different sets of beliefs regarding the position of the group in society might exist in relatively less powerful groups. Identification might, therefore, be based on a specific set of beliefs, or might take the form of complete individualisation from the group. On the other hand, incongruence might result from the dominant defining the dominated as group members, while the dominated perceive self as non-group members. In such cases where members experience conditions conducive to leaving the group, Tajfel (1978) predicted strategies of individual mobility from the group. Individual mobility, social creativity and social competition have also been suggested as possible strategies for group members experiencing negative or threatened identity. In addition, non-conforming group members might opt for social creativity strategies, like avoidance of painful comparisons with the out-group, attempts to change the value of ingroup characteristics in a more positive direction, and comparing the ingroup with the outgroup on some new dimensions (Giles & Johnson, 1981; Tajfel & Turner, 1979). It should be noted, however, that most of these suggested strategies are based on a conception of intragroup solidarity and might, therefore, not be applicable in cases of intragroup variability. We are also of the opinion that any dissident group member, whether deciding to remain within the ingroup or to leave it, might experience complex feelings of group belongingness which might be manifested in intricate avoidance attraction strategies towards the ingroup (see Merton, 1965, 1976). In either one of the incongruence instances above, a dissident group member might be perceived (by self and/or others) as member and non-member

< previous page

page_106

next page >

< previous page

page_107

next page > Page 107

simultaneously, and, therefore, as sharing and not sharing the group's position in society (see Okamura, 1981). Within the structural constraints of society, the dissident group member thus has to make complex decisions regarding the variable significance of his/her social identity: e.g. whether to affirm group membership when rejected by the ingroup; whether to deny group membership in the face of unwanted, ascribed membership; and how to deal with unexpected, residual feelings of group belongingness towards the repudiated, formerly significant group, etc. It has been asserted that the salience of group membership might be heightened in certain situations of, for example, confrontation, crisis and ritual (Vincent, 1974) which highlight the differential rights and privileges associated with status disparities between groups (Despres, 1975) rather than in others. On the basis of these empirical data and theoretical suggestions, the following assumptions, on which our model is based, can be formulated: (1) whatever the particular terms used to describe the subjective experience of identification or the 'objective' judgement of others regarding conformity to group norms, both these factors might influence social identity; (2) power differentials might exist within as well as between groups; (3) the relatively more powerful (whether constituting a group, or a sub-group within a group) might attempt to impose their notion of identity upon the relatively less powerful; (4) this notion of identity might change over time; (5) the powerless might react to the imposition of identity by the powerful in a variety of ways; and (6) these reactive identity management strategies might be more complex than hitherto suggested. A Model of Identity and Dissent We propose the following model (see Table 1) to deal with identification and dissent within a powerful group with an insecure social identity: Table 1 The dominated as defined by self and the dominant Other-Definition Self-Definition (Provided by the (Provided by the Dominant) Dominated) Defined in Defined out Defined in A B Defined out C D Cells A and D represent the congruent instances, while Cells B and C represent the incongruent instances. We are hypothesising that: 1. in Cells A and D, where there is little difference between self- and other-definition, the dominated would not engage in attempts to escape or reject the dominant's strategy, respectively, to draw them into or expel them from the ingroup;

< previous page

page_107

next page >

< previous page

page_108

next page > Page 108

2. in Cells B and C, the 'misfit' between self- and other-definition would motivate the dominated to engage in strategies to escape the imposed identity conferred upon them by the dominant; 3. those individuals who are defined out of the ingroup against their will (Cell B) would attempt to confirm their group membership by using strategies which are aimed at changing the dominant's criteria for group membership; 4. those individuals who are defined into the ingroup against their will (Cell C) would attempt to repudiate this imposed membership by using strategies which are aimed at differentiating self from the ingroup; 5. as language has been identified as a particularly important dimension of ethnic identity (e.g. Christian, Gadfield, Giles & Taylor, 1976; Leclézio, Louw-Potgieter & Souchon, 1986), strategies of identity imposition, affirmation, or denial, would centre on the language of the specific group; 6. however, as language is not acontextual, it should combine with other, important identity determinants (e.g. self-definition and economic wealth (Giles, Llado, McKirnan & Taylor, 1979), political orientation and/or skin colour (Louw-Potgieter & Giles, 1987), etc.) in this process of identity management. An Application of the Model The model presented here should be regarded as a tentative framework for the study of identity and dissent. It is, therefore, not presented as a formal theory, but rather applied to a single ethnolinguistic group, Afrikaans speakers, in order to test it and to gather more information regarding the specific identity management strategies used by dissident group members. As our conceptualisation of 'Afrikaner' differs from commonly held stereotypes, a few remarks regarding this group are required. In terms of political power, it could be regarded as a high status group, albeit illegitimate and insecure high status (status defined here as the outcome of a comparison on a specific dimension (Tajfel & Turner, 1979)). In the eyes of casual observers, intragroup cohesion and loyalty have become stereotypical attributes of this group and the notion of dissidence from it is usually met with disbelief or amusement (Van den Berghe, 1970). We are, therefore, rejecting the notion of volkseenheid (unity of the nation propagated by the National Party [NP] in order to mobilise all Afrikaans speakers as 'Afrikaners' during and after the 1948 election) and the subsequent impression of ingroup solidarity created by it. Historical evidence of internal strife and shifting intragroup alliances (Davenport, 1977; Wilson & Thompson, 1975) support the notion of the existence of group members critical of their group's privileged position within South African society who were and are trying to distance themselves from this undesirable group membership. In order to go beyond the idea of monolithic Afrikanerdom, we are conceptualising this group as consisting of various sub-groups of Afrikaans speakers; e.g. Afrikaners, traditional Afrikaners, ware (true) Afrikaners, dissident Afrikaners, verligtes (the enlightened), verkramptes (reactionaries), angry Afrikaners, philo-

< previous page

page_108

next page >

< previous page

page_109

next page > Page 109

sophically analytical Afrikaners, etc. (W. de Klerk, 1973; W. A. de Klerk, 1975; Louw-Potgieter, 1986). In terms of the model proposed, a powerful sub-group of Afrikaans speakers, constituted by those who have the political power to enforce their definition of group identity upon other, less powerful sub-groupings can be isolated. This powerful group has been identified (see Louw-Potgieter, 1986) as the supporters of P. W. Botha within the NP, the political party currently (and since 1948) in power in South Africa. This sub-group is able to create social categories and to transform these categories into social reality (for instance, the NP's Population Registration Act of 1950 in which South Africans were classified as belonging to specific 'racial' groups). By virtue of their power (see Khleif, 1979), they control the media and it is, therefore, easy for them to disseminate their current definition of what an Afrikaner is (see Giliomee, 1979). For this sub-group, an in-depth interview study indicated that a group member should speak Afrikaans, share a common culture, fate, background, and history with other Afrikaners, and be concerned with the maintenance and preservation of his/her unique Afrikaner identity (Louw-Potgieter, 1986). However, Louw-Potgieter & Giles (1987) found that this sub-group perceived those who were white, voted for the NP and spoke Afrikaans to be highly similar to themselves, and regarded those who spoke Afrikaans, but did not support the NP; or those who spoke Afrikaans, but did not share the same skin colour, to be highly dissimilar to themselves. In consequence, the attributes of being white and voting for the NP were added to the above definition of an Afrikaner, thereby constituting the norm, the reference point, or 'élite' definition of a group member (Deschamps, 1982; Guillaumin, 1972). We shall label this sub-group the Afrikaners and henceforth all other sub-groups of Afrikaans speakers will be categorised, labelled and described from the perspective of the Afrikaners. The Dominated from the Perspective of the Dominant Four different empirical studies will be referred to henceforth. The first three studies (Studies 1, 2 and 3) are described fully in Louw-Potgieter (1986), while the multidimensional scaling study of Louw-Potgieter & Giles (1987) constitutes the fourth (for further details on the samples and procedures see Note 1). In this section, the discussion on identity imposition will centre on the bases for ascribed identity, and linguistic strategies and labels used to designate belongingness to, or distance from the ingroup. Cell A This cell of the model represents a dominated or less powerful sub-group of Afrikaans speakers whose self-definitions are congruent with the way in which they have been defined by the Afrikaners. They differ from the Afrikaners only in that they are not opinion leaders or norm creators and could be conceived of as loyal camp followers and supporters of the leading, dominant Afrikaners.

< previous page

page_109

next page >

< previous page

page_110

next page > Page 110

Cells B and C Represent a 'misfit' between the self-definition of specific sub-groups and the way in which they have been defined by the Afrikaners. This 'misfit' is often represented in the differential, ascriptive labels used by the dominant and the dominated, reflecting their respective perceptions regarding location within the social system (Lever, 1978). Cell B In this cell, we have Afrikaans speakers who regard themselves as group members, but who are defined out of the group by Afrikaners. Firstly, the basis of exclusion might be political, as in the following examples. Liberale/Liberaliste Afrikaners regard this sub-group as too left-wing, because they are critical of the policies of the NP; thus, despite the fact that this sub-group conforms to most of the membership criteria regarded as essential by Afrikaners, they are defined out of the group on the basis of political orientation. Afrikaners described them as 'people who . . . do not want to maintain our white identity'; or 'a person, born Afrikaans, who repudiates his descent and culture'; and 'Afrikaans speakers who arc embracing those who are alien to the volk' (Study 3, p. 524). Currently there is no commonly known and used label for this sub-group and the more emotional labels of the past (e.g. volksverraaier [renegade of the nation], kommunis [communist], etc.) have fallen into disuse. It could be argued that, by not naming those individuals who could be considered both as Afrikaners and non-Afrikaners, the Afrikaners are attempting to remove such confusing instances from everyday life (Leach, 1966). On the other hand, the lack of a label for left-wing Afrikaans speakers could indicate that the status quo itself is changing and that such Afrikaans speakers are gradually becoming more acceptable as group members (Louw-Potgieter, 1986). In order to provide a label for this sub-group, we have chosen the most recent of the rather outdated labels used by Afrikaners to designate this group, namely liberale or liberaliste (liberals or liberalists). It is interesting to note that Afrikaners rarely use this appellation in conjunction with the adjective denoting group membership (e.g. 'liberale Afrikaners' [liberal Afrikaners], or 'Afrikaner liberaliste' [Afrikaner liberalists]), thereby indicating that Afrikaans speakers of this persuasion are not regarded as group members. It was noted by an Afrikaner that 'amongst us Afrikaners, the word liberaal usually has a negative meaning . . . The most commonly accepted definition holds that a liberalis is a proponent of integration between White and Black in our country' (P. S. Dreyer, 1977). Afrikaners have acknowledged that they use a defining out strategy to rid their group of these undesirable members: '. . . If one then finds that there arc Afrikaners of whom one cannot be proud, then one would like to cut them from the body of the volk,'

< previous page

page_110

next page >

< previous page

page_111

next page > Page 111

but suggested simultaneously that this should be resisted for the sake of group unity (Coetzee, 1971: 49-50). These ambivalent attitudes regarding inclusion/ exclusion of this specific sub-group were confirmed by Study 3, in which Afrikaners indicated that they were willing to accept specific Liberaliste into their group. It should be stressed again that, at times when the status quo itself is changing, some group members might still retain the original group ideology (an exclusivity based on support for strict apartheid), while others are exhibiting more recent ideas (the ideology of 'reform') about their group's position in society. Ware Afrikaners Afrikaners regard this sub-group as too right-wing, because they have left the NP and joined ultra right-wing parties, like the Conservative Party (CP) (founded after the 1982 breakaway from the NP) and the Herstigte Nasionale Party (HNP) (founded after the 1969 breakaway), or belong to organisations like the Afrikaner Weerstandsbeweging (AWB) (Afrikaner Resistance Movement), or the Afrikaner Volkswag (AV) (Afrikaner Guard of the Nation). Although they conform to most of the membership criteria regarded as essential by Afrikaners, and in addition seek to preserve the 'true spirit' of Afrikanerdom as represented by former leaders, like Hendrik Verwoerd (known as 'the architect of apartheid'), they are excluded from the Afrikaner group and are perceived by Afrikaners as highly dissimilar to themselves on an integrationist-segregationist dimension (LouwPotgieter & Giles, 1987). It was decided to label this group Ware Afrikaners (true Afrikaners). The origin of this term is unknown, but it could be speculated that, at its inception, it simply denoted a 'true' or prototypical member of this group, as opposed to more peripheral members. After the 1969 breakaway from the NP, the image of the true Afrikaner was altered significantly, as it became unclear whether such a person belonged to the NP or the HNP. In 1982, a more serious breakaway occurred when Andries Treurnicht, a powerful figure on the right, left the NP to constitute the CP. While the NP was making propaganda for a broader South Africanism, especially in the light of the forthcoming constitutional changes, and fought an election and a referendum on the issue of state security (Giliomee, 1981), the CP stuck to the idea of an exclusive Afrikaner nationalism and claimed that they represented the true Afrikaners. Those dubbed the 'reformers' within the NP soon labelled the CP the 'ultra right-wing', and the term Ware Afrikaners acquired a rather derogatory meaning and was used often in conjunction with 'Superafrikaner', a term denoting a member of the secret organisation, the Broederbond, or an arch-conservative. In Study 3, it became clear that the label 'Ware Afrikaner' caused some problems for Afrikaners: the majority reported that they did not use the term any more, mainly because they were unsure of the exact meaning of it. Those who reported using the term either used it to denote a prototypical group member, or in a derogatory, jocular fashion to describe 'people who make a ''religion" out of their so-called Afrikanerhood, who see themselves as superior to others' (Study 3, p. 520). It would seem that, while this sub-group had been contained within the NP, this label was used by Afrikaners in a normative

< previous page

page_111

next page >

< previous page

page_112

next page > Page 112

sense, but once it had defected from the NP Afrikaners started using the term in a negative, derogatory, or ironical sense. It is evident that Afrikaners regard these Ware Afrikaners no longer as group members. The dilemma, however, is to reject from their ranks those who consider themselves essential and prototypical group members. The general strategy seems to be the propagation of a new ideology of 'reform', while simultaneously accentuating how outdated the Ware Afrikaners' conception of 'Afrikaner' is. Afrikaners have been of the opinion that: 'it is our task to handle the great changes in our country . . . [but] the Conservative Party sows the suspicion that National Party is about to sell the white man down the river . . .' (W. de Klerk, 1982: 30). and that: '. . . what these people [the CP] do not seem to realise, however, is that the world of 1984 is not the world of 1936' (Gerber, 1984: 25). Secondly, still in Cell B, the Afrikaners might use skin colour, or 'race', as the basis of exclusion from the group. Coloureds Afrikaners have officially classified this sub-group as Coloureds. 2 Despite the fact that 83% of them are Afrikaans-speaking (Republic of South Africa Population Census, 1980) and, in addition, conform to most of the membership criteria regarded as essential by Afrikaners, Afrikaners seek to exclude them on the basis of skin colour (see Beukes, 1974; Viljoen, 1971). It is clear that Afrikaners are of the opinion that 'language is essential for Afrikanerhood, but is not sufficent in itself. Afrikanerhood entails more than that, it demands a common race and culture. It would, therefore, be more correct to use the term "Brown Afrikaans speakers" regarding the Coloureds' (Viljoen, 1971: 16). It has been asserted that 'Afrikaner nationalism has been of incredible importance for the establishment of Afrikaans as a cultural and official language, but from Afrikaner nationalism negative implications for Afrikaans have also resulted: the division of the speech community by means of a colour barrier and the expulsion of brown Afrikaans speakers. . .' (Ponelis, 1984: 30). Cell C The idea that the powerful's conceptualisation of identity might change over time is one of the basic assumptions of our model. At times, Afrikaners have pursued an exclusive ideology of apartheid, which resulted in firm boundaries between perceived ingroup and outgroup. At other times, however, pragmatic considerations (e.g. broadening of the power base) have forced Afrikaners to co-opt those perceived as non-Afrikaners into the ingroup. This inclusive strategy has resurfaced recently in the form of an ideology of 'reform'. It has been noted that lately Afrikaners have exhibited more tolerance for 'heretics' (i.e. white, left-

< previous page

page_112

next page >

< previous page

page_113

next page > Page 113

wing Afrikaners speakers) (W. de Klerk, 1982) and that intergroup alliances, especially with 'our Afrikaans speaking Brown people' (Editorial, Rapport, 17/7/1982) are mentioned with more frequency. Attempts at inclusion of Afrikaans speakers, irrespective of party political affiliation or skin colour, have usually been motivated by threat, whether it be cultural assimilation by the English-speaking group or political domination by the Black majority. It should be kept in mind, however, that, just as 'reform' remains on the ideological level for Afrikaners, so does the strategy of inclusion of some Afrikaans-speaking sub-groups into the Afrikaner group. A complex approach-avoidance pattern could thus be expected from Afrikaners towards those Afrikaans speakers perceived to be different from them. It should also be noted that, despite the fact that Afrikaans is their home language, specific sub-groups of dominated Afrikaans speakers might not define themselves as Afrikaners, and might, therefore, experience a 'misfit' between self- and other-definitions in these situations of attempted inclusion. Examples are as follows: Politically Left-Wing, White Afrikaans Speakers 3 Usually, this sub-group, which does not desire ingroup membership themselves, would be regarded by the Afrikaners as too left-wing to be included in the Afrikaner group. However, under specific circumstances, Afrikaners might seek to include them (albeit reluctantly), in order to swell the ranks of the Afrikaner group. In such cases, the common language and culture are stressed, while differences in political outlook are played down. Part of Afrikaners' incursive strategy is to admit that the group is heterogeneous and could, therefore, accommodate them. It has been asserted by Afrikaners that 'we should not disqualify Afrikaners who do not conform to our definition of what constitutes an Afrikaner. Today, being an Afrikaner has a much more heterogeneous meaning. There are many kinds of Afrikaners but they are still Afrikaners' (W. de Klerk, 1973). The lack of a label for this sub-group is noticeable. It could be argued that, in this case, Afrikaners are employing a strategy which is aimed at drawing these Afrikaans speakers into the group, and would, therefore, not use any labels which could differentiate them from the group. If labels are coined (e.g. 'the dissident Afrikaner, the angry Afrikaner, the philosophically analytical Afrikaner' (W. A. de Klerk 1975: 341)), they usually consist of a distinguishing adjective used with the noun, Afrikaner thus denoting the heterogeneity while simultaneously stressing the group membership. Bruin Afrikaners (Literally, Brown Afrikaners) As in the case of Politically left-wing, white Afrikaans speakers, this sub-group does not regard itself as ingroup members and, usually, is not regarded as such by Afrikaners. For reasons of group survival, and in times of threat, Afrikaners might also consider incorporating them into the ingroup. In what has been termed 'power sharing without losing control' (Giliomee, 1984: 29), Afrikaners have co-opted Coloured (and Indian) representatives as junior partners into the political system in 1983. Within this context (and especially while making propaganda for these limited constitutional reforms), the label 'Coloured' has been changed by

< previous page

page_113

next page >

< previous page

page_114

next page > Page 114

the Afrikaners to Bruin Afrikaners and the language and culture which this sub-group shares with them were accentuated. The Director of Language and Arts Research of the Human Sciences Research Council has recently made the assumption that '. . . the Coloureds, of whom the overwhelming majority speak Afrikaans as their home language, regard themselves as part of the Afrikaans community' (Prinsloo, 1984: 102). In order to illustrate the historically ambivalent attitudes Afrikaners have had towards the Coloured people for much of the twentieth century (Trapido & Marks, 1987), and the approach-avoidance strategies resulting from this ambivalence, one need only examine the actions and decisions of the Federasie vir Afrikaanse Kultuurverenigings (FAK: Federation of Afrikaans Cultural Organisations), an umbrella body of all Afrikaans cultural organisations. At the 1983 meeting of this organisation, the CP succeeded in winning control of the executive and getting a number of ultraconservative proposals accepted. In reaction to this coup, an editorial in an NP newspaper stressed that the cultural activities of ordinary people, and specifically 'our Afrikaans speaking Brown people', were more important than those of official cultural organisations (Rapport, 17/7/1983: 14). However, in 1985, when the NP regained control of the FAK executive, it was reaffirmed that this superordinate cultural organisation would not consider Coloureds as members. Also, in Study 3, approximately half of the sample of Afrikaners indicated that they never used the term 'Bruin Afrikaners' because they defined group membership in a more exclusive way, or in a more inclusive way, or regarded subjective identification with a group as more important than ascription to it. The other half of the sample of Afrikaners who reported that they used the term either used it to distinguish between groups of Afrikaans speakers, or to emphasise the similarity of all Afrikaans speakers. The complexity of the meaning of this label for Afrikaners is reflected in the following explanation provided by a respondent regarding his specific usage of it: 'To identify with a fellow language user who is not White' (Study 3, p. 522). Cell D Represents those sub-groups of Afrikaans speakers who do not define themselves as Afrikaners and who are, despite the common language and/or other shared attributes, considered 'too far gone' by the Afrikaners to be included in the group. These sub-groups would, for instance, not be considered, in specific situations, as possible ingroup members (like the sub-groups in Cell C), because they are regarded by Afrikaners as much more extreme. The following sub-groups fitting this cell could be identified: White Afrikaans Speakers Who have Rejected Their First Language and Become Anglicised In order to escape undesirable Afrikaner group membership, this sub-group has rejected the Afrikaans language for English (see Ponelis, 1984). In a study on the White élite groups in South Africa, Van der Merwe, Ashley, Charton &

< previous page

page_114

next page >

< previous page

page_115

next page > Page 115

Huber (1974), found that 5% of the Afrikaans speakers showed a language shift towards English, while only 0.3% of the English speakers had accepted Afrikaans as their first language. The authors noted that all language groups have certain institutions (e.g. schools, clubs, etc.) which reflect the general ethos of the group. Any shift in language would also mean a shift towards those institutions where the group's values are inculcated. It could, therefore, be predicted that Afrikaans speakers who have adopted English as their first language would become part of the English speech community and be regarded as members of an outgroup by Afrikaners (cf. linguistic strategy of individual mobility, Giles & Johnson, 1981). White Afrikaans Speakers Who are Regarded as Left-Wing Extremists Members of this sub-group have attempted to escape Afrikaner group membership by adopting a radical political orientation and/or joining what Afrikaners would regard as 'extreme' political groups. They are usually labelled 'Communists' by Afrikaners, whether they subscribe to this ideology or not. Harrison (1981:211) noted that for Afrikaners: '. . . Communism undermined traditional South African attitudes towards race. Communists advocated equal civil and political rights for all and it required no great leap of Nationalist imagination to believe that those who advocated equal civil rights for all South Africans must therefore be Communists' (see also P. Dreyer, 1980; Vermaak, 1966). Afrikaans Speakers Who have Been Classified Officially as 'Bantu' (Meaning Black) In 1968, 55,880 black people in South Africa had Afrikaans as mother tongue (Beeld, 14/7/1968), while in 1970 1% of the black people in Namibia had Afrikaans as mother tongue (Republic of South Africa Population Census, 1970). Despite the shared language, this sub-group is perceived as 'too different' from Afrikaners to be included in the group. Louw-Potgieter & Giles (1987) also found that Afrikaners perceived 'Africans' (a term generally used by Afrikaners to denote black inhabitants of South Africa and/or Africa) as being the most dissimilar group from themselves on an Afrikaner/non-Afrikaner dimension. Schuring (1981) commented on a language phenomenon in Afrikaans called Tsotsitaal (language of the rogues). This type of Afrikaans is used as lingua franca in black residential areas like Soweto and Mamelodi in Transvaal. Ponelis (1984: 34) remarked that 'unfortunately we know very little about this commonly used Afrikaans dialect'. The fact that Afrikaans was still being used in the 1980s in areas which had been the focus of the 1976 riotssparked off by the enforcement of Afrikaans in black schools makes this phenomenon and its neglect by Afrikaners all the more extraordinary. Reactions of the Dominated to Imposed Group Identity Viewing imposed identity from the perspective of the dominated might provide insight into the strategies used by them to escape this imposition. The reaction of each sub-group in Cells B and C (in Cells A and D, the congruence instances,

< previous page

page_115

next page >

< previous page

page_116

next page > Page 116

these strategies are not expected) to Afrikaners' imposition of identity will be discussed. Liberale/Liberaliste The awareness of being defined out of the Afrikaner group by the Afrikaners is clearly reflected in the following quotations from interviews with Liberaliste: 'I think in a certain sense one is so marginalised that in the end . . . I have landed on the sidelines without consciously intending to be there'; and '[a dissident Afrikaner] is a man who starts saying no to a model other people have created about themselves, about their ethnic group, and, therefore, the cultural framework in which they move, and then wanting to prescribe it to this person' (Study 2, p. 471). The reactions of Liberaliste to the defining out strategy of the Afrikaners can be classified according to Tajfel's (1978) three-component view of social identity; namely strategies concerning categorisation, comparison and identification. Each type of strategy wil be discussed below. (a) Categorisation Strategies Firstly, the results of Studies 2 and 3 indicated that Liberaliste stressed the importance of self-categorisation (as opposed to categorisation by others) for identification with a group. On being categorised by Afrikaners as non-group members, respondents noted: 'It is their business how they regard me; it is not going to change the way in which I see myself'; or'. . . they cannot diminish my ethnic identity by what they think of me and how they treat me: I myself decide about that' (Study 3, p. 537). The active rejection of other-categorisation again illustrates our ideas about the strategies involving linguistic categories; clearly this is no passive accceptance of labelling attempts by others, but a highly creative and dynamic response on the part of the recipient of the label. Secondly, Liberaliste attempted to change the content of the category by indicating that they defined 'Afrikaner' in a way which differed from the definition supplied by Afrikaners. In Study 2, language emerged as the most common membership attribute: a third of the respondents regarded the single attribute of language as sufficient for group membership (e.g. '. . . the Afrikaner is someone whose mother tongue is Afrikaans'; '. . . an Afrikaner would be any person who is Afrikaans-speaking'; or 'Afrikaner one would describe as a person who speaks Afrikaans' (p. 473)), while the overwhelming majority included this attribute in their definition of a group member. Not one respondent mentioned that voting for the NP constituted a criterial attribute and very few regarded skin colour as essential. In Study 3, this sub-group defined an Afrikaner as somebody who speaks Afrikaans, defines self as an Afrikaner and shares a common culture with other Afrikaners. Not a single respondent endorsed political party affiliation as a neces-

< previous page

page_116

next page >

< previous page

page_117

next page > Page 117

sary attribute for group membership and very few endorsed skin colour. These definitions have specific implications for the perception of group boundaries and clearly reflect the inclusive nature of the Liberale's perception of a group member: any Afrikaans speaker, irrespective of skin colour or politics (or any of the other criteria deemed necessary by Afrikaners), who perceives self as a group member and shares the group's culture. Thirdly, it could be said that Liberale tried to change the definition of 'Afrikaner' in a specific way: they tended to over-emphasise inclusive attributes (e.g. language, self-definition, culture, etc.) and tended to under-emphasise exclusive attributes (e.g. skin colour, political orientation, etc.). For instance, Liberale said: 'I myself try to emphasise the language and free association with the language and language groups and a person who tries to define me out on the basis of political reasons is directly in opposition to the things I try to do in the political and cultural area'; and '[it is laughable for] a political group to claim the right for themselves to define a cultural group in an exclusive manner' (Study 3, p. 535). In Study 3, when asked about their perceptions of Afrikaners' criteria for group membership, Liberale over-estimated the importance Afrikaners attached to exclusive criteria (being white, being from (white)West European descent), while under-estimating the importance of more inclusive criteria (self-identification, language, and culture). Fourthly, evidence exists that Liberale attempted to change the labels, or the meanings of the labels applied to them and other Afrikaans speakers by Afrikaners, to more positive descriptions. In Study 3, the majority of Liberale said that they did not use the term 'linkse Afrikaners' (left-wing Afrikaners) for a variety of reasons. The following suggestions were offered as alternatives to this term: 'liberal Afrikaners', 'critical or free Afrikaners', and 'liberal-minded Afrikaners'. Fifthly, when being categorised by self or others (i.e. in situations where ethnolinguistic identity was made salient), Liberaliste firstly affirmed their group membership, while, almost immediately thereafter, differentiating themselves from Afrikaners. Aware of being regarded as members and nonmembers simultaneously, this sub-group employed the following self-presentation strategies: selective disclosure of membership; self-description in terms of related, but more 'innocuous' group memberships (e.g. 'of Africa', 'South African', etc.); enhancing own difference from the ingroup by overemphasising group membership; self-mockery (e.g. 'I am one of those racists, called Afrikaners', etc.); and using others' incorrect categorisations to enhance own difference from Afrikaners (for instance, 'I am often perceived as being an English-speaking person . . . and then it is really enjoyable afterwards to come back with a vengeance and . . . to change into broad Afrikaans') (Study 2, pp. 348-52). Data from Study 2 (see pp. 332-60) also revealed that Liberale experienced shared pride, shame and/or guilt as group members, even though they attempted to differentiate themselves from Afrikaners. Moreover, it became clear that Liberale were of the opinion that rational argument was the most effective way of transmitting their alternative definitions

< previous page

page_117

next page >

< previous page

page_118

next page > Page 118

or changed labels to Afrikaners and other interested parties. The following quotation illustrates this: 'I will try, to the best of my ability, to explain my motivation for my viewpoint and the way in which I understand who is an Afrikaner, in the hope that they would understand better and perhaps be willing to accept a broader understanding of Afrikanerhood' (Study 3, p. 534). (b) Comparison Strategies The strategies identified under this heading were mostly on the intragroup level and were concerned with differentiation from Afrikaners. Some of these comparison strategies might also, to an extent, overlap with the categorisation strategies already mentioned. Firstly, the main strategy to emerge under this heading can be identified as differentiation from fellow group members (Condor, 1984; Turner, 1982), or the primus inter pares (PIP) effect (Codol, 1984). In order to achieve positive identity, individuals might assign positive attributes of group membership to themselves and any negative characteristics to other group members (Turner, 1982) or, in assigning positive attributes of group membership to all members, they might nevertheless assume that they are still closer to the normative ideal (Codol, 1984). In Study 2, three respondents reported that they perceived 'dissident Afrikaners' to be the Ware Afrikaners, thus appropriating this term for self-description and simultaneously changing the value of the term in a more positive direction. Study 3 provided further support for this strategy, as Liberale indicated that they were not affected by Afrikaners' exclusive strategies, because they regarded themselves as the prototypical group members and the true Afrikaners. This is illustrated by the following quotation: 'I would say that Afrikaners who are future-oriented like me are opening the way for our volk. I would state that our kind are really and truly the true Afrikaners' (Study 3, p. 538). Secondly, Liberale used a strategy whereby they extracted positive kernels from decidedly negative stereotypes held by others about 'Afrikaners'. For instance, it was alleged that '[w]hat black people say is that they regard Afrikaners more highly than the English, because the Afrikaners hate them more honestly' (Study 2, p. 381). (c) Identification Strategies The two strategies classified under this heading involved, firstly, not identifying with all those attributes described by Afrikaners as essential group attributes, and, secondly, not identifying with all members of the ingroup. As these strategies also involved re-categorisation or changing the content of a category, a certain overlap with (a) might be expected. The first strategy is clearly reflected in the attempts of Liberale to change the definition of 'Afrikaner' (see [a] for examples and details). Studies 2 and 3 both revealed that often the Afrikaner group as a whole was overlooked for identification purposes, but that similar others or specific sub-groups of Afrikaans speakers provided sources of identification. For instance, on

< previous page

page_118

next page >

< previous page

page_119

next page > Page 119

describing his reaction to the exclusive strategy of Afrikaners, a Liberalis commented that '. . . there are, in any case, enough other (Ware) Afrikaners with whom I would be able to associate' (Study 3, p. 538). Apart from these popular intragroup strategies, Liberale infrequently engaged in intergroup strategies in reaction to Afrikaners' exclusive strategy. In Study 2, it was found that some respondents categorised themselves as South African, African, academic, liberal, sociologist, intellectual, etc., and acted in terms of these superordinate memberships rather than in terms of their Afrikaner group membership. In Study 3, further evidence of movement towards alternative groups was found. Eight respondents replied that exclusion by the Afrikaners did not affect them any more, as they have formed alliances with other groups; for instance, English speakers, South African blacks, foreigners, or dissident Afrikaners. The following quotation reflects this strategy: 'I have already built up such a wide and intimate reference group outside the Afrikaners that I don't exactly feel ostracised if they do not want me' (Study 3, p. 540). The general impression obtained from the data discussed was that Liberale seemed to be quite confident of the fact that they were the true group members and were, therefore, not very much affected by the defining out strategies of the Afrikaners. Ware Afrikaners This sub-group's claim of prototypicality and its ultra-conservative stance as regards the content of 'Afrikaner' provide the Afrikaners with an embarrassing dilemma as far as exclusion is concerned. Afrikaners have attempted to exclude Ware Afrikaners from the ingroup mainly by means of ridicule, while Ware Afrikaners themselves have resorted to the following strategies to these attempts: (a) Categorisation Strategies Firstly, Ware Afrikaners attempted to change the content of 'Afrikaner' to differentiate themselves from Afrikaners. In Study 3, Ware Afrikaners defined 'Afrikaner' as a person who is concerned with the maintenance and preservation of a unique Afrikaner identity, who shares a common culture with other Afrikaners, who is white, who shares a common history, background and language with other Afrikaners, who speaks Afrikaans and who is a member of one of the Afrikaans churches. Secondly, it could be alleged that Ware Afrikaners tried to change the definition of 'Afrikaner' in a specific way: when describing how they perceived Afrikaners' conceptualisation of a group member, it became clear that they under-estimated the importance Afrikaners attached to language, culture, and identity, and overestimated the importance of voting for the NP. It could be concluded from these results that Ware Afrikaners regarded themselves as custodians of 'true Afrikanerhood' and were trying to differentiate themselves from Afrikaners by emphasising more exclusive group attributes (e.g. skin colour, church membership, etc.) and minimising Afrikaners' endorsement of criterial attributes (e.g.

< previous page

page_119

next page >

< previous page

page_120

next page > Page 120

language, culture, identity, etc.). This was further supported by Louw-Potgieter & Giles (1987), who found that on a separatistintegrationist dimension Ware Afrikaners perceived themselves at the separatist end and Afrikaners at the integrationist end of the dimension, and, therefore, as highly dissimilar to themselves. This exclusive definition of a group member has inevitable implications for Ware Afrikaners' perceptions of group boundaries and membership: by requiring members to conform to a host of exclusive criteria, the group boundaries could be hardened and 'undesirable' members could be kept out. In view of this subgroup's conservative and racist ideology, it could be asserted that such 'undesirability' would centre mainly on political orientation and skin colour (see Louw-Potgieter & Giles, 1987). In Study 3, for instance, it was found that no Ware Afrikaner used the label 'Bruin Afrikaner'. The reasons given for this were based on the assumption that group members could only be white: 'Coloureds are a separate volk, even though they speak Afrikaans'; or 'the Coloureds speak Afrikaans, but I do not regard them as Afrikaners' (Study 3, p. 529). Thirdly, there exists evidence that Ware Afrikaners changed the labels, or the meanings of the labels, applied to them by Afrikaners to more positive descriptions. In Study 3, it was found that they used the term 'Ware Afrikaner' in its original, 'positive' sense: '[A] collection of White, Afrikaans-speaking people who have unified a common feeling of nationalism under the name Boer'; and 'A Boer . . . a White who is part of our population and who recognises and respects the history and covenants of our forefathers '(Study 3, p. 521). By way of explanation, it should be pointed out that, recently, ultra right-wing Afrikaans speakers have started propagating a 'white homeland' option an ideology which stresses that all Afrikaners who strive for 'racial purity' should secede from South Africa to form an independent, white Boer nation. These two quotations might also provide evidence that the term 'Boer' is used by Ware Afrikaners for self-descriptive purposes and more positively than by other groups of South Africans. Black South Africans, for instance, invariably use this label in a derogatory manner to refer to the police or the white, Afrikaans-speaking régime. One noticeable aspect of Ware Afrikaners' membership claims as described above is the biblical justifications they employed to bolster their group belongingness. Upon becoming aware of Afrikaners' strategy to exclude them from the ingroup, they reported that such people could only be perceived as 'victims of an anti-Christian, international propaganda campaign that started in Babel, long before Christ', and that one could only pity them, 'pray for them', or ask: 'Father, forgive them, because they don't know what they are doing' (Study 3, pp. 542-46). (b) Comparison Strategies As was the case for Liberale, the strategies for Ware Afrikaners identified under this heading were mostly on the intragroup level and were concerned with differentiation from Afrikaners. As Ware Afrikaners were adamant about the fact that they were the prototypical group members, it could be predicted with some

< previous page

page_120

next page >

< previous page

page_121

next page > Page 121

certainty that the PIP effect (Codol, 1984) or, in this specific case, the mother-in-law strategy (Condor, 1984) would be the main comparison strategies used by this sub-group. The results from Study 3 supported this prediction, as this strategy, often accompanied by biblical justifications, proved to be the most popular. It became clear that Ware Afrikaners regarded Afrikaners as the 'deviants' whose definition of 'Afrikaner' constituted an 'aberration'. The following quotations illustrate these ideas and this specific intragroup comparison strategy: '. . . how can one try to rob the identity of someone who is a true nationalist, while oneself is busy making more and more concessions to other ethnic groups at the expense of one's own group?'; or 'I regard them as South Africans and not as Afrikaners. A person who does not want to preserve his own volk's identity and who does not want to associate with the whites, but with the masses of South Africa, cannot lay claim to Afrikaner identity' (Study 3, p. 544) 4 (c) Identification Strategies Three different strategies could be identified under this heading. Firstly, while they identified with all the attributes described by Afrikaners as essential, it became clear that Ware Afrikaners attempted to add to these. These additions (a white skin, membership of an Afrikaans church) reflect their conservative and exclusive notion of a group member (see [a] above for details). Secondly, Ware Afrikaners did not identity with the group as a whole, and especially not with Afrikaners whom they regarded as 'impure' group members. The following quotations illustrate this: Asked how he would feel if Afrikaners said that they did not regard him as a group member, a Ware Afrikaner respondent in Study 3 replied: 'Proud, because I do not want to be regarded as one of them' (p. 542). Another concluded that '. .. I do not want to be associated with volksverraaiers' (p. 544). Thirdly, Ware Afrikaners suggested that they might secede from the Afrikaners by establishing a rival Afrikaner group. As mentioned in (a), and reflected in the following quotations, this group was to include only Ware Afrikaners: 'I would continue to campaign for an Afrikaner-volkstaat'; or q would have to examine whether I should insist on acceptance or whether I would form my own group' (Study 3, p. 547). As with the Liberale, the same general impression of confidence in their group membership was obtained from the data of Study 3 of the sample of Ware Afrikaners (see, for instance, [b] on prototypicality). The reactions of this sub-group to the defining out strategy of the Afrikaners reflected their attempts at differentiation and ranged from '[pitying] them for their ignorance and uprooted-ness' (Study 3, p. 542) to pride in being different from them. Coloureds Recent events in South Africa have diminished dramatically the likelihood of this sub-group aspiring to Afrikaner group membership. While individual mobility of light-skinned Coloureds into the white Afrikaans-speaking or English-speaking groups was a relatively common occurrence in the 1950s-1960s, social

< previous page

page_121

next page >

< previous page

page_122

next page > Page 122

pressure from more politicised sub-group members precludes this to a large extent today. It was, therefore, not surprising that our data did not yield a single respondent who could be classified into this sub-group. It could, however, be theorised that a minority of Afrikaans-speaking Coloured people of a certain ideological persuasion, who might covertly desire Afrikaner group membership, might still exist (e.g. some Coloureds who were willing to take part in the NP's new, three-tiered parliament). On the basis of this information, strategies mentioned for this sub-group have to remain on the level of speculation. (a) Categorisation Strategies Firstly, Coloureds might try to change the content of 'Afrikaner' to indicate that they do not regard the Afrikaners' definition of group membership as acceptable. Secondly, the specific way in which this sub-group might try to change the definition might include under-emphasis of those criteria which they do not share (e.g. white skin, NP membership, etc.), and over-emphasis of more inclusive attributes which they do possess (e.g. the Afrikaans language, a common culture and history, the fact that they are descended directly from white, Afrikaans speakers, etc.). This might have implications for perceived group boundaries and membership, as they might perceive all Afrikaans speakers who share a common culture to be eligible for group membership. Thirdly, it could be predicted that Coloureds might reject the label 'Coloured', perceived by them as crude (Senekal, 1984), and possibly also as a reminder of the basis of their exclusion. They might, therefore, adopt more 'positive' labels, like 'Bruin Afrikaners' or 'Bruinmense' (literally, brown people), or create a new label, devoid of all skin colour connotations. (b) Comparison Strategies As there are no indications that Coloureds regard themselves as the true group members, we cannot predict a PIP effect for this sub-group. The only possible prediction could be that Coloureds might perceive some similarity between themselves and Liberale (for instance, in terms of their definitions of group membership), and might co-operate with them in order to establish a common, satisfactory social identity. (c) Identification Strategies Following from (a) and the strategies used by Liberale and Ware Afrikaners, it could be predicted that Coloureds might not identify with all the attributes regarded as essential for group membership by Afrikaners. Secondly, Coloureds might not identify with the group as a whole, but might select only perceived similar others (i.e. Coloureds and Liberale) for identification purposes. It has been hypothesised that those who are defined into the Afrikaner group against their will (Cell C) might attempt to repudiate this imposed identity by engaging in strategies of differentiating themselves from the Afrikaners. The identity management strategies of the following two subgroups will be discussed in the same terms as before:

< previous page

page_122

next page >

< previous page

page_123

next page > Page 123

Politically Left-wing, White Afrikaans Speakers In times of threat, 'renegades' and 'heretics' such as these might suddenly become acceptable as group members in order to broaden the power base of Afrikaners. As Afrikaners change and relax their criteria for group membership (for instance, the current move towards 'reform'), this sub-group might unexpectedly find itself being perceived as possible group members and react in the following manner to attempts to include them in the ingroup: (a) Categorisation Strategies Firstly, this sub-group regarded self-categorisation as the basis of group identity, thereby implying that an individual should have the freedom of choice as regards this issue and not be dragooned into a group on the basis of inescapable or immutable attributes (Studies 2 and 3). This is reflected indirectly in the following excerpt from Study 2: a Liberalis said that his inclusive definition of 'Afrikaner' would encompass a politically left-wing, white Afrikaans speaker, 'provided that he regards himself as an Afrikaner, because I know many so-called dissident Afrikaners who don't want to have anything to do with the concept, Afrikaner' (p. 482). Secondly, this sub-group used categories outside the Afrikaner group for self-categorisation purposes. In Study 2, these categories emerged as South African, African, academic, intellectual, world citizen or stateless citizen. Thirdly, a strategy of complete individualisation might be opted for. In Study 2, there is evidence that there were respondents who did not think of themselves in Afrikaner terms or, for that matter, in any other group terms. The following quotation illustrates this: 'I regard myself . . . as someone who has broken away completely, who has no affinity for the group, or the nationality, or the ethnicity' (p. 342), and he added that he did not just reject the Afrikaner group, but 'broke away from this whole group thing' (p. 326). Fourthly, when asked to define 'Afrikaner', language emerged as the common attribute in the definitions supplied by this sub-group. However, any similarity in definition ended there: some respondents supplied their own definition and indicated how it differed from that of the Afrikaners, while others supplied only the definition they perceived to be used by Afrikaners, and a single respondent said that it was too difficult to define 'Afrikaner'. All this could be seen as support for the idea that this sub-group have rejected any notion of affiliation with Afrikaners and, therefore, were not really interested in semantic games involving the image of this undesirable group. Data from Study 2 indicating a lack of shared pride in the group supported this idea. However, evidence of shared shame and guilt indicated that this sub-group might still harbour residual feelings of group belonging-ness (see pp. 352-57). Fifthly, for this sub-group, the label 'dissident Afrikaner' meant either those Afrikaans speakers who have broken completely with the group, or two different

< previous page

page_123

next page >

< previous page

page_124

next page > Page 124

categories; namely one which has distanced itself completely and another which has retained links, but is rebellious against the definition of the group as provided by the Afrikaners (in other words, differentiating between Liberale and Politically left-wing, white Afrikaans speakers) (Study 2). From the results of Studies 2 and 3, it could be concluded that this sub-group showed a lack of interest in the labels used by Afrikaners to denote other sub-groups of Afrikaans speakers and did not attempt to change these labels in any way. (b) Comparison Strategies No comparison strategies emerged for this sub-group, which could be construed as further evidence that they viewed group membership as undesirable or irrelevant (e.g. intragroup comparisons would no longer be relevant, as Afrikaners did not constitute a reference group any more, while intergroup comparisons would be made on dimensions other than identity power, for instance, etc.). (c) Identification Strategies Following from (a), and illustrated by the results of Study 2, attempts to seek out broader group memberships, or opting for a purely individualistic self-identification, were the strategies used most often by this sub-group. It could be theorised that the ultimate form of this latter strategy of differentiation from the ascribed ingroup would be to reject the ingroup's language, thus destroying the most important basis of ascription (in other words, a shift to Cell D). One respondent in this sub-group replied to our letters in English and, although he spoke Afrikaans during the interview, he stressed the fact that he thought in English and German when busy with academic matters. Bruin Afrikaners Unlike the Coloureds, who aspired to Afrikaner group membership, this sub-group has rejected any attempt of Afrikaners to include them in the ingroup. In a television interview, Jakes Gerwel (7/7/1985), a renowned Coloured academic, stressed that an Afrikaner is a white person who speaks Afrikaans and that Coloureds are not part of the group called Afrikaners. He said that Coloureds form part of a clearly delineated group of their own, which shares their language with the Afrikaner, just as the Irish are a separate group which incidentally speak the same language as the English. The following strategies were employed by this sub-group in order to escape the undesirable identity Afrikaners attempted to impose on them: (a) Categorisation Strategies Firstly, this sub-group attempted to escape the imposed identity by defining 'Afrikaner' in such a way that it did not include them. From the results of Study 3, it is clear that this sub-group did not perceive themselves as members of the Afrikaner group, as they stressed that 'Afrikaner' meant being white, belonging to one of the Afrikaans churches, speaking Afrikaans, sharing a common

< previous page

page_124

next page >

< previous page

page_125

next page > Page 125

background with other Afrikaners, voting for the NP, and sharing a common history with other Afrikaners. Secondly, the specific way in which they used the definition to exclude themselves involved simultaneously admitting that they spoke the same language as Afrikaners, but stressing that their skin colour precluded them from being group members (see Gerwel, and definition, above). Thirdly, Bruin Afrikaners used categories other than Afrikaner for self-categorisation. The category used most often appeared to be English-speaking South Africans. To acquire membership in this category, this sub-group might attempt to reject the Afrikaans language and become anglicised. There is no doubt that large numbers of Afrikaans-speaking Coloureds are turning their backs on what they perceive as the language of the oppressor (see Gerwel, 1975, on the preference of English as communication medium between parents and children in this sub-group). Often combined with this anglicisation strategy is the tendency to categorise self as black and seek membership in black consciousness groups. Such a strategy would indicate effectively the ultimate attempt at distancing self from the political power of the group whose language one shares or has shared. The following illustrates this strategy: the (Coloured) Labour Party's decision in 1983 to support the proposed three chamber parliament has been described as more than a simple political decision: 'It was, for many Coloureds, a statement about their identity. For those who hated apartheid because it kept them apart from the whites, the door to power has at last opened. For those who see their future alongside the country's black majority, the decision is a massive betrayal' (Zille, 1983: 30). Fourthly, it was confirmed by Study 3 that this sub-group rejected the label 'Bruin Afrikaner', and were not concerned with labels indicating the inclusion or exclusion of other sub-groups of Afrikaans speakers by Afrikaners. (b) Comparison Strategies A single intragroup comparison strategy, aimed at differentiation from the Afrikaners and centring on the Afrikaans language, was identified. Instead of rejecting Afrikaans, this sub-group stressed the fact that the Afrikaans they used differed from the 'Boers' language'. In this context, Gerwel (quoted in Senekal, 1984) noted that even during the riots of 1976 the political language at the University of the Western Cape, where the majority of students were Coloured, remained Afrikaans and concluded that the issue was not the language per se, but an attitude associated with the language. Zille described the 'new' language used by this sub-group as 'a lively patois of the [Cape] Flats, called Gamtaal [the language of Ham], a vibrant mixture of Afrikaans and English, fast-flowing and high-pitched' (Zille, 1983: 32). Giles & Johnson (1981) noted that a social creativity strategy where 'inferior', 'sub-standard' or 'minority' language, dialect or slang are no longer stigmatised, but are proudly displayed as a group symbol of cultural pride, might be used by members of a dominated group to differentiate themselves from the dominant group.

< previous page

page_125

next page >

< previous page

page_126

next page > Page 126

(c) Identification Strategies Based on information in (a) and (b), two identification strategies were identified for this sub-group. Firstly, respondents sought out other groups (English-speaking South Africans or black South Africans) with which to identify. Secondly, this sub-group identified with the Afrikaans language, but stressed that the Afrikaans they valued differed significantly from the Afrikaans spoken by Afrikaans. The general impression of this sub-group's reaction to the Afrikaners' inclusion strategy was one of intense suspicion and surprise, coupled with a cynical realisation that possible inclusion was based on considerations of self-interest on the part of Afrikaners. Summary and Conclusion Hypothesis 1 predicted that in the congruence instances (Cells A and D) there would be an absence of strategies to escape imposed identity. Although not discussed in this paper, the results of Study 3 indicated that there was no significant difference between the responses of Afrikaners and their loyal camp followers (see pp. 488-89 for more details). Representatives of Cell D were not investigated, as the focus of this paper and the related studies was on incongruence instances. Hypothesis 2 predicted that in (incongruous) Cells B and C respondents would be motivated to escape the imposed identity. Abundant support for this hypothesis is evident in the data. Hypothesis 3 predicted that individuals who are defined out of the ingroup against their will (Cell B) would engage in strategies attempting to change the Afrikaners' criteria for group membership. Results reported herein supported this and elaborated on the specific ways in which different sub-groups might attempt to do this; that is, using mainly creative strategies aimed at changing the 'élite' definition of the group. Hypothesis 4 predicted that those individuals who are defined into the ingroup against their will (Cell C) would attempt to repudiate this imposed identity by attempting to differentiate themselves from the ingroup. This hypothesis received ample support and elaboration from the results reported here. Hypothesis 5 predicted that, as language is central to identity, strategies to impose, affirm or deny identity would centre on the group's language. This was firmly supported with language emerging as the most common membership criterion mentioned by all sub-groups. Indeed, the identity management strategies centred mainly on the content of a linguistic category; namely the name of the ethnolinguistic group. Hypothesis 6, assuming that language is not acontextual, predicted that it would combine with other, important determinants in identity management strategies. This was supported by the fact that skin colour and political orientation, together with language, formed the three main criteria for group membership. Other important criteria mentioned were: a shared culture, history, fate, background, self-identification, and specific church membership. On the basis of these results, we propose the following tentative theoretical propositions regarding intragroup identity and dissent:

< previous page

page_126

next page >

< previous page

page_127

next page > Page 127

Proposition 1: Power differentials within groups might create various powerful and powerless sub-groups. Proposition 2: Within such a heterogeneous group, the powerful might attempt to impose their conception of identity on the less powerful. Proposition 3: In instances of congruence, the powerless agree with this imposed identity and do not engage in attempts to escape it. Proposition 4: In instances of incongruence, the powerless disagree with this imposed identity and might engage in attempts to escape it. Proposition 5: In the specific instance of incongruence where non-member-ship is imposed on the powerless who might desire membership, they might engage in creative studies on an intragroup level aimed at changing the powerful's definition of group membership. Proposition 6: In the specific instance of incongruence where membership is imposed on the powerless who might desire non-membership, they might engage in inter- or intragroup strategies aimed at differentiating self from the undesirable group. Proposition 7: As language forms a very important attribute of ethnolinguistic group membership, strategies of identity management might centre on the group's language. Proposition 8: As language is not acontextual, it might be combined with other important membership attributes in these strategies of identity management. At the very basis of these propositions, and clearly reflected in the data, is the central idea that linguistic categories might mean different things to different people; so-called intragroup relations and communication may be just as complex and potentially conflictual as intergroup relations. The existence of a specific linguistic category might well imply that it is used, but it does not yet tell us by whom, in which situations, and with which connotations. Should then linguistic categories be accepted unquestioningly by researchers, the end result might be extensive documentation on how the powerful use these labels. In order to find out how all group members perceive, use, and avoid the linguistic labels available to them, empirical investigation, not blind acceptance of 'obvious' usage, is essential. A case in point and a fruitful area in which to apply the theoretical propositions advanced here might be how 'feminist' men engage with the linguistic categories created and used by 'traditional' men. Finally, how individuals representative of our four quadrants speak to each other and to the dominant élite may have important implications for communication accommodation theory (Gallois, Franklyn-Stokes, Giles & Coupland, in press) as well as Gudykunst's (1986) axiomatic model of intergroup communication. Notes 1. Study 1: This required respondents to write an autobiography, according to specific guidelines, regarding their change from a dissident Afrikaner to a traditional Afrikaner. Twenty two respondents, 21 men and one woman, who were well-educated, middle class, professional, Afrikaans speakers whose parents were also Afrikaans-speaking,

< previous page

page_127

next page >

< previous page

page_128

next page > Page 128

took part in the study. (See Louw-Potgieter, 1986:195 for geographical, age and occupational details, if needed). Study 2: This was a follow-up to Study 1. Twenty one of the 22 respondents in Study 1, plus an additional four respondents, took part in it. These additional respondents were well-educated, middle class, professional, Afrikaans-speaking men whose parents were also Afrikaans speaking. A focused interview, based on social identity theory principles, was used to interview them. (See Louw-Potgieter, 1986:302 for age, geographical and occupational details, if needed.) Study 3: In this study, six different sub-groups of Afrikaans speakers were asked to complete a questionnaire concerning their perceptions of the content and boundaries of the Afrikaner category. A sub-group representing Afrikaners consisted of 23 white, middle class men who were members of the National Party in Natal. Another sub-group representing Liberale consisted of the same respondents who took part in Study 2. A third sub-group representing Ware Afrikaners, consisted of 15 white, middle class respondents, 12 men and three women, who were members of the ultra right-wing AWB. (The Afrikaners and the Ware Afrikaners for this study, were the same samples used in Louw-Potgieter & Giles, 1987.) The fourth sub-group representing Coloureds was not included in this study, as no respondents fitting this category answered the questionnaire. A fifth sub-group consisted of Politically left wing, white Afrikaans speakers. There were only two respondents in this category and they were well-educated, white, middle class men. Both of them took part in Study 1 and 2. The sixth sub-group, called Bruin Afrikaners, consisted of 5 respondents, three men and two women. They were well-educated, middle class people who had been classified as 'Coloured' by the Afrikaners. The small samples in this study need to be explained: data were gathered during 1985, a time of escalating intergroup conflict in South Africa. A state of national emergency had been declared, giving the state wide ranging power to arrest any person suspected of 'subversive' acts. In the light of these circumstances, it is therefore understandable that people who were known for their opposition to the state, were very relucant to take part in studies with a political slant. (See Louw-Potgieter, 1986:488-94 for any further information. There are actually small discrepancies in the numbers of respondents mentioned in that source and here: questionnaires received too late to consider in the dissertation were nonetheless included in this paper.) 2. In this chapter, the labels 'Coloured', 'Indian', etc., are used without inverted commas or the adjective 'so-called'. This usage should in no way be construed as approval of the South African government's 'racial' categorisation system or seeking to perpetuate it. These labels are used simply because they (unfortunately) have a certain reality within South African society, and are used in this specific way because any additions might complicate matters further, especially in the latter sections of this paper, where labels are discussed. 3. Politically right-wing, white Afrikaans speakers could not fall in this cell, because they conceive of themselves as the true group members and would, therefore, not define themselves out of the Afrikaner group. 4. It should be noted that, while the PIP effect emerged for both Liberale and Ware Afrikaners, the basis of differentiation is dissimilar. Liberale are trying to differentiate themselves from Afrikaners in terms of their radical, inclusive definition of 'Afrikaner', while Ware Afrikaner are using their traditional, conservative and exclusive definition as a basis. These processes can, therefore, not be understood in isolation from the sub-group's ideas about the group. References Beukes, D. (1974) Afrikaner se plig om sy identiteit te handhaaf, Die Burger, 12 July. Billig, M. (1976) Social Psychology and Intergroup Relations. London: Academic Press.

< previous page

page_128

next page >

< previous page

page_129

next page > Page 129

Breakwell, G. (1979) Woman: Group and identity, Women's Studies International Quarterly, 2, 9-17. Christian, J., Gadfield, N. J., Giles, H. and Taylor, D. M. (1976) The multidimensional and dynamic nature of ethnic identity, International Journal of Psychology, 11, 281-91. Coetzee, P. C. (1971) Dan bly hy nog 'n Afrikaner, Buurman, 1, 49-50. Codol, J-P. (1984) Social differentiation and non-differentiation. In H. Tajfel (ed.), The Social Dimension, Vol. 1. Cambridge: Cambridge University Press. Condor, S. (1984) Sex group membership and womanhood as an aspect of 'social identity'. Unpublished manuscript, Psychology Dept., University of Lancaster, England. Condor, S., and Abrams, D. (1984) Womanhood as an aspect of social identity. Group identification and ideology. Paper presented at the International Conference on Self and Identity, Cardiff, July. Davenport, T. R. H. (1977) South Africa: A Modern History. London: Macmillan. De Klerk, W. (1973) Die Transvaler, 21 December. (1982) The delicate balance that could decide our fate, Sunday Times, 7 November. De Klerk, W. A. (1975) The Puritans in Africa: A Story of Afrikanerdom. London: Rex Collins. Deschamps, J-C. (1982) Social identity and relations of power between groups. In H. Tajfel (ed.), Social Identity and Intergroup Relations. Cambridge: Cambridge University Press. Despres, L. A. (1975) Ethnicity and resource competition in Guyanese society. In L. A. Despres (ed.), Ethnicity and Resource Competition in Plural Societies. The Hague: Mouton Publishers. Dreyer, P. (1980) Martyrs and Fanatics: South Africa and Human Destinies. London: Secker and Warburg. Dreyer, P.S. (1977) Afrikaner-Liberalisme. Pretoria: Boekenhout. Fishman, J. A. (1977) Language and ethnicity. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Gallois, C., Franklyn-Stokes, A., Giles, H. and Coupland, N. (in press). Communication accommodation theory and intercultural encounters: Intergroup and interpersonal considerations. In Y. Kim & W. B. Gudykunst (eds), Theoretical Perspectives in Intercultural Communication. Newbury Park, CA: Sage. Gerber, C. (1984) The final test, Sunday Times, 22 July. Gerwel, J. (1975) Die band tussen Afrikaans en die Bruinman is versteur. Deurbraak, July. Giles, H. and Johnson, P. (1981) The role of language in ethnic group relations. In J. C. Turner & H. Giles (eds), Intergroup Behaviour. Oxford: Blackwell. Giles, H., Llado, N., McKirnan, D. J. and Taylor, D. M. (1979) Social identity in Puerto Rico, International Journal of Psychology, 14, 185-201. Giliomee, H. (1979) The growth of Afrikaner identity. In H. Adam & H. Giliomee (eds), The Rise and Crisis of Afrikaner Power. Cape Town: David Philip. (1981) Afrikaners soek al hoe meer na 'n nuwe identiteit, Rapport, 4 October. (1984) Can the Afrikaner share power and not lose control? Sunday Times, 24 June. Gudykunst, W. B. (1986) Toward a theory of intergroup communication. In W. B. Gudykunst (ed.), Intergroup Communication. London: Edward Arnold. Guillaumin, C. (1972) L'Idéologie Raciste: Genèse et langage Actuel. Paris: Mouton. Harrison, D. (1981) The White Tribe of Africa. London: BBC. Khleif, B. B. (1979) Insiders, outsiders and renegades. In H. Giles & B. Saint-Jacques (eds), Language and Ethnic Relations. Oxford: Pergamon Press. Leach, E. (1966) Anthropological aspects of language: Animal categories and verbal abuse. In E. H. Lenneberg (ed.), New Directions in the Study of Language. Cambridge, Massachusetts: MIT Press. Leclezio, M. K., Louw-Potgieter, J. and Souchon, M. B. S. (1986) The social identity of Mauritian immigrants in South Africa, Journal of Social Psychology, 126, 61-9. Lever, H. (1978) South African Society. Johannesburg: Jonathan Ball. Louw-Potgieter, J. (1986) The social identity of dissident Afrikaners. Unpublished doc-

< previous page

page_129

next page >

< previous page

page_130

next page > Page 130

toral dissertation, Bristol University. (1988) Afrikaner Dissidents: A Social Psychological Study of Identity and Dissent. Clevedon: Multilingual Matters. Louw-Potgieter, J. and Giles, H. (1987) Afrikaner identity: diversity amongst the right, Journal of Multilingual and Multicultural Development, 8, 283-92. Merton, R. K. (1965) Social Theory and Social Structure. New York: The Free Press. (1976) Sociological Ambivalence and Other Essays. New York: The Free Press. Okamura, J. Y. (1981) Situational ethnicity, Ethnic and Racial Studies, 4, 452-65. Ponelis, F. A. (1984) Afrikaans binne meertalige verband. In K. P. Prinsloo & M .C. J. van Rensburg (eds), Afrikaans: Stand, Taak en Toekoms. Pretoria: HAUM. Prinsloo, K. P. (1984) Stand van die Afrikaanse taal: Tans en in die Toekoms. In K. P. Prinsloo & M. C. J. van Rensburg (eds), Afrikaans: Stand, Taak en Toekoms. Pretoria: HAUM. Republic of South Africa Population Census. (1970) Pretoria: Government Printer. (1980) Pretoria: Government Printer. Schuring, G. K. (1981) Die basilek van flaaitaal, Tydskrit vir Geesteswetenskappe. 21. Senekal, J. H. (1984) Afrikaans: 'n Toekomsblik. In K. P. Prinsloo & M. C. J. van Rensburg (eds), Afrikaans: Stand, Taak en Toekoms. Pretoria: HAUM. Tajfel, H. (ed.) (1978) Differentiation Between Groups. London: Academic Press. (1981) Human Groups and Social Categories: Studies in Social Psychology. Cambridge: Cambridge University Press. Tajfel, H. and Turner, J. C. (1979) An integrative theory of intergroup conflict. In W. G. Austin and S. Worchel (eds), The Social Psychology of Intergroup Relations. Monterey, California: Brooks/Cole. Trapido, S. and Marks, S. (1987) South Africa since 1976. An historical perspective. Paper for the David Davies Memorial Institute for International Relations. Turner, J. C. (1982) Towards a cognitive redefinition of the social group. In H. Tajfel (ed.), Social Identity and Intergroup Relations. Cambridge: Cambridge University Press. Turner, J. C. and Brown, R. J. (1978) Social status, cognitive alternatives and intergroup relations. In H. Tajfel (ed.), Differentiation Between Social Groups. London: Academic Press. Turner, J. C. and Giles, H. (eds) (1981) Intergroup Behaviour. Oxford: Blackwell. Van den Berghe, P. (1970) Race and Ethnicity. New York: Basic Books. Van der Merwe, H. W., Ashley, M. J., Charton, C. J. and Huber, B. J. (1974) White South African Elites: A Study of Incumbents of Top Positions in the Republic of South Africa. Cape Town: Juta. Vermaak, C. (1966) Braam Fischer, the Man with Two Faces. Johannesburg: APB Publishers. Viljoen, G. v N. (1971) Behoud en uitbouing van die Afrikaner se identiteit, Sarp, 7 (10) 43-4:(11) 16-7. Vincent, J. (1974) The structuring of ethnicity, Human Organisation, 33, 375-79. Wilson, M. and Thompson, L. (1975) The Oxford History of South Africa. Vol. 2. Oxford: Clarendon Press. Zille, H. (1983) Between Amandla and the PeeCee, Frontline, 3, 29-35.

< previous page

page_130

next page >

< previous page

page_131

next page > Page 131

Language Attitudes and Vitality Perceptions: Intergenerational Effects Amongst Chinese Canadian Communities Itesh Sachdev 1 University of London, U.K. Richard Bourhis McMaster University, Canada Sue-wen Phang University of Manchester, U.K. and John D'Eye University of London, U. K. Abstract First (FGCC) generation and Canadian-born (CBC) subjects of Chinese origin in Canada completed the Subjective Vitality Questionnaire and a sociolinguistic questionnaire concerning self-reported usage and evaluations of Cantonese (CC) and English (EC) in different domains. Perceived vitality reflected the objective estimates of high EC and low CC vitality on demographic, institutional support and status factors. However, contrary to expectations, CBCs exaggerated perceived Cantonese vitality on some dimensions relative to FGCCs. Self-reports and evaluation of language use reflected the dominance of the English Language in public settings, but CBCs agreed less than FGCCs about Cantonese usage in the home and church. Overall, vitality perceptions of succeeding generations of Canonese-speakers were most useful in understanding language usage and evaluation patterns when used in conjunction with group identification and language proficiency information. Introduction In an attempt to place social psychological processes mediating interethnic communication in their appropriate sociostructural contexts, Giles, Bourhis & Taylor (1977) developed a taxonomy of factors which were hypothesised to affect groups' ethnolinguistic vitalities, i.e. their ability to survive as distinctive and active collective entities in intergroup settings. The aim of the present study was

< previous page

page_131

next page >

< previous page

page_132

next page > Page 132

to explore the relationship between group vitality, identification, language use and attitudes across two generations of an immigrant minority community in Canada. According to Giles et al. (1977) status, demographic strength, institutional support and control factors combine to make up the vitality of ethnolinguistic groups. Groups' strengths and weaknesses in each of these domains may be assessed to provide a rough classification of ethnolinguistic groups as having low, medium or high vitality. Low vitality group members are most likely to assimilate linguistically or cease to exist as a distinctive collectivity. High vitality group members are most likely to survive as a distinctive collectivity in multilingual settings. Since its introduction, the construct of ethnolinguistic vitality has been incorporated into a variety of models in cross-cultural communication (e.g. Bourhis, 1979), second language acquisition (e.g. Giles & Byrne, 1982; Giles & Johnson, 1987), language attitudes (e.g. Ryan, Giles & Sebastian, 1982). ethnic identification (e.g. Giles & Johnson, 1981) and relations between the sexes (e.g. Kramarae, 1981). The original formulations of Giles et al. (1977) focussed largely on an 'objective' analysis of ethnolinguistic vitality gathered from available sociological and demographic information. The Subjective Vitality Questionnaire (SVQ, Bourhis, Giles & Rosenthal, 1981) was developed to take into account individuals' representations of vitality which mediate ethnolinguistic behaviour (Moscovici, 1981; Johnson, Giles & Bourhis, 1983; cf. Husband & Saifullah Khan, 1982). Interestingly, Giles, Rosenthal & Young (1985) obtained impressive factor analytic support from SVQ responses for the subjective reality of ethnolinguistic vitality along the dimensions of demography, status and institutional support and control. The relationship between perceived vitality, language proficiency, usage, attitudes and group identifications has recently received some empirical attention. For instance, Bourhis & Sachdev (1984) obtained supportive evidence for a hypothesised positive relationship between perceived ingroup vitality and self-reported language attitudes and behaviour in a variety of public and private settings. Interestingly, their results also suggested that the demography of the immediate school settings within which their studies were carried out mediated perceptions of vitality and attitudes towards language usage. Other research has suggested that group members may systematically minimise or exaggerate the vitality of their own or salient outgroups depending on their levels of social identification (Johnson, 1984), learning of, and proficiency in own- and out-group languages (Sweeting, 1982; Young, Pierson & Giles, 1984). Unlike an objective vitality analysis, the subjective vitality approach adopted by recent studies has demonstrated the flexibility of the SVQ to take into account specific intergroup relations situations as well as assessing the degree of intragroup heterogeneity in perceptions of vitality. Undoubtedly, an important source of intragroup heterogeneity having serious implications for language and group survival may be found in the perceptions and behaviour of succeeding generations of ethnolinguistic groups. However, systematic empirical investigation on the impact of intergenerational differences on the survival of groups is still in its infancy.

< previous page

page_132

next page >

< previous page

page_133

next page > Page 133

Reitz (1980) in a review of studies and social surveys conducted in Canada concluded that members of east and west European minority groups (e.g. Italians, Germans, Ukrainians and Poles) tend to loosen their ethnic group 'ties' and social/ethnic group networks relatively quickly after immigration and particularly in the second and third generations (see also O'Bryan, Reitz & Kuplowska, 1976; Bourhis, 1987). Measures of group ties have included levels of ingroup interaction, ethnic identifications and language retention. After immigration, ethnic identification appears to weaken more than ingroup interaction and immigrants quickly begin to think of themselves as 'Canadians'. The most susceptible tie seems to be the ingroup language. Similar to findings obtained in U.S.A., Reitz (1980) obtained evidence that owngroup language knowledge and use, though widespread amongst adult immigrants, is much less prevalent in the second generation and rarely to be found in the third generation in Canada. Though the vitality of minorities from European backgrounds has yet to be systematically investigated, it appears that their vitality is declining rapidly and seriously undermines their chances of survival as distinct ethnolinguistic groups in multicultural Canada. The paucity of research on intergenerational differences and ethnolinguistic vitality is most apparent in the case of 'visible', non-European minorities (e.g. Chinese, Japanese, East and West Indians) who have long been the targets of discrimination in Canada (e.g. see Blanc, 1987). Unlike east and west European immigrants, visible minority groups may not 'pass' easily as Anglo (or Franco) Canadians even after linguistic assimilation. Indeed where physiognomy and skin-colour are criterial for category membership, visible minority group members may neither rely nor be allowed to rely a lot on language retention to preserve their sense of ethnic identification and group cohesion. Interestingly, Pak, Dion & Dion (1985) found amongst Chinese students in Toronto that self-rated confidence in English did not necessarily constitute any corresponding loss to their own cultural identity. The present study represents a first attempt to systematically monitor the vitality perceptions of succeeding generations of a visible minority in Canada. The groups chosen for this study were first generation Canadian Chinese (FGCC) and second generation Canadian-born Chinese (CBC) of Hong Kong descent living in the city of Toronto (Ontario). Although Chinese Canadians (CCs) living in Toronto form the second largest CC community in Canada, this group is definitely in a minority relative to the English Canadians (ECs) who form the single largest ethnic group in Toronto. The first aim of the present study was to assess the degree to which subjects' perceptions accurately reflected 'objective' assessments of the groups' vitalities. The SVQ, modified for tile Toronto setting and comparing the relative vitality of ECs and CCs on each vitality dimension (see Bourhis et al., 1981), was completed by FGCC and CBC subjects attending Christian churches in Toronto. Assessment of groups' 'objective' vitality was carried out using census data and other sociological information. Previous research in Canada (e.g. O'Bryan et al., 1976; Bourhis, 1987) has clearly identified strong attitudinal support for heritage language retention

< previous page

page_133

next page >

< previous page

page_134

next page > Page 134

amongst Chinese Canadians, although specific intergenerational differences amongst Chinese Canadians have yet to be systematically documented. The second aim was thus to investigate how vitality perceptions relate to the frequency of, and attitudes towards, the use of ingroup and outgroup languages. To investigate this issue all subjects completed a sociolinguistic survey comprising two parts: (i) self-reported usage of Cantonese and English language in different domains such as home, church, school, work, etc.; (ii) Evaluation of the use of these languages in the different sociolinguistic settings. Our simplest general expectations, based on Bourhis (1979, 1987) and Bourhis & Sachdev (1984) were formulated as follows: (a) Cantonese would be used more frequently and in a wider range of settings by group members who perceive Chinese vitality to be higher than those who perceive it to be lower. (b) Evaluations of Cantonese usage will be more positive in a wider range of settings by group members who perceive Chinese vitality to be higher than those who perceive it to be lower. The above general expectations were assessed in the context of two other hypotheses concerning intergenerational differences, vitality perceptions, language usage and language-use evaluations: (c) FGCCs' perceptions of ingroup vitality were expected to be higher than those of CBCs. This prediction was based on 'objective' assessments suggesting that the knowledge and use of Cantonese is more prevalent in the FGCCs than CBCs. (d) FGCCs were expected to have more favourable attitudes in a wider range of settings about the use of Cantonese than SGCCs. 'Subjective' vitality data may only be optimally used in conjunction with 'objective' information about the group's ethnolinguistic vitality (Bourhis et al., 1981). The following section is a brief assessment of the 'objective' vitality of Chinese (CC) and English Canadians (EC) in the city of Toronto using the taxonomy provided by Giles et al. (1977). The 'objective' analysis of vitality was carried out using Canadian Government Census statistics and other sociological and historical information. CC and EC Demography In the 1800s gold mines and the building of railways employed as many as 17,000 lowly paid Chinese labourers (mainly from Guangdong province). Exhaustion of the gold mines and completion of the transcontinental railways at the turn of the century led to large-scale unemployment amongst Chinese immigrants. While many returned to China, others moved to the Canadian Prairies and further eastward to cities such as Toronto. By the Second World War urban Chinese communities were well established in Vancouver, Montreal and Toronto.

< previous page

page_134

next page >

< previous page

page_135

next page > Page 135

Table 1 Census1 profile of English Canadians (ECs), Chinese Canadians (CCs) and others in Toronto by stated ethnic origin Census Total Population No. of ECs2 No. of CCs3 No. of Others'4 of Toronto absolute no. & % absolute no. & % absolute no. & % Year 1941 667, 457 523, 588 2,326 141,543 (100%) (78.4%) (0.4%) (21.2%) 1951 675,754 465,457 2,879 207,418 (100%) (68.8%) (0.4%) (30.8%) 1961 1,824,481 1,107,203 20,534 696,744 (100%) (60.7%) (1.1%) (38.2%) 1971 2,089,220 1,474,335 20,765 594,120 (100%) (70.6%) (1% ) (28.4%) 1981 2,998,947 2,136,975 67,910 794,062 (100%) (71.3%) (2.3%) (26.4%) 1. This table was compiled from Canadian Census data published in 1942, 1952, 1962. 1972 and 1982. 2. ECs include all those who reported British descent: English, Scottish, Welsh and Irish. 3. CCs include all those who reported Chinese descent. 4. Others exclude ECs and CCs, includes in order of importance German origin, French, Polish, Ukrainian, Dutch and others but missing Italian, Greeks and Portuguese. Table I indicates that the proportion of CCs in Toronto increased two and a half times from 0.4% in 1941 to 1.1% in 1961. Relative to the ECs who have been in a majority, the CC proportion in Toronto remained at about 1% in Toronto according to the 1971 census. The 1981 census showed that the ECs and CCs had stabilised to their majority and minority proportions of 71.3% and 2.3%, respectively. Census data indicate that by the early 1970s the distribution of non-Anglo immigrant populations in Toronto was residentially concentrated in a single central region of the city. Neighbourhoods, where clusters of families of similar ethnic origin resided, included the Chinese to the east, in what is distinctively recognised as Toronto's Chinatown. Such areas bordering on the central business district were preferred by Chinese immigrants because they provided low-cost housing and also were close to the sources of unskilled employment, e.g. restaurant catering, laundry, etc. Recent changes in immigration policies based on a 'points' system tended to favour the wealthy, skilled and educated Chinese. Most recent Chinese immigrants have come from Hong Kong, where they were educated in the British school system. Recently settled Chinese Canadian business people and professionals have sought careers in the Anglo-Saxon/Celtic community using the highly visible Chinatown mainly as a cultural buffer zone during the post immigration

< previous page

page_135

next page >

< previous page

page_136

next page > Page 136

period. Most recent Chinese immigrants tend to live in the more affluent suburbs of Toronto, away from the older Chinatown area of the city (O'Bryan et al., 1976). Little information is available concerning the endogamy and birthrates of CCs and ECs in Toronto. A survey by Reitz (1980) suggested that CCs in Toronto maintained higher rates of endogamy relative to all European groups except for the Southern Europeans. Marriage registers indicate that interethnic marriages are less frequent amongst CCs than amongst ECs in Toronto. The 30 fold increase in the population of Chinese in Toronto from 1941 to 1981 suggests that immigration rather than high birthrate is responsible for the increase of CC population in Toronto. Overall, the demographic trends have stabilised such that relative to the EC majority, Chinese Canadians remain a small though highly visible minority in Toronto. CC and EC Institutional Support and Control Whereas ECs and the English language are extremely well represented in all formal and informal institutions of Toronto (i.e. high vitality), this has not been the case for Chinese Canadians and their language. Relative to the Northern and East European immigrant groups, CCs have long been the target of racism and job segregation in Canada (Reitz, 1980). Indeed before the Second World War, anti-Chinese sentiments were actually enshrined in several pieces of specific government legislation, including 'head' and 'landing' taxes for each person of Chinese origin and denial of federal voting rights in 1929. These practices denied Chinese representation and control over formal institutions such as education, business, the mass media and civil administration. The history of discrimination contributed to the formation of professional services and organisations whose aims were specifically to satisfy the distinctive needs of the CC community in Toronto. Parents associations backed up by the Mon Sheong Foundation sought changes within the school system and asked for schools to teach the Chinese language, culture and philosophy. Beginning in the late 1970s, Chinese language schools have been established in Toronto as well as in other cities such as Vancouver, Victoria and Calgary. The objective of such schools is to maintain the Chinese language and traditional culture in the lives of the younger generation. Surveys (e.g. Reitz, 1980) suggest widerspread support for language retention amongst the Chinese in Toronto (68%), although intergenerational differences have yet to be documented. Cantonese (Hong Kong) is the most common language used amongst people of Chinese origin, although Mandarin and a variety of other dialects are also spoken. At present, the CC community in Toronto is served by four Chinese language newspapers. A multicultural radio station in the Toronto area provides Chinese language programming which averages about 9-10 hours weekly. The federal government through its Multilcultural Directorate provides limited services and helps subsidise annual CC cultural activities in Toronto. There are six churches in Toronto offering religious services in Cantonese.

< previous page

page_136

next page >

< previous page

page_137

next page > Page 137

At present, the CCs are a fairly active collectivity as regards informal instituitional representation and control in Toronto. Reitz (1980) suggested that unlike European minorities, second and third generation CCs were likely to maintain their distinctiveness as they remain targets of discrimination from the host community. Overall, ECs may be characterised as having high vitality, CCs as having medium vitality and European minorities such as the Germans and Dutch as having low vitality on institutional support factors. Status of CCs and ECs in Toronto Though the vitality of groups on status factors may be the most difficult to assess, it is clear that the dominant ECs, being the founders of the modern city of Toronto, enjoy the highest amount of prestige. CCs, on the other hand, having long been the target of discrimination are accorded much lower prestige, even in comparison to European minorities. Job segregation, low literacy levels and educational qualifications of the early Chinese immigrants helped maintain their low job status. Although the present Chinese community have higher job status than the earlier immigrants, there are some indications that members of the present Chinese community still suffer from restricted job opportunities (e.g. Reitz, 1980). Moreover, results of various studies have shown that negative stereotypes of CCs abound (e.g. Berry, Kalin & Taylor, 1977; Phang, 1983). Although the actual prevalence of negative stereotypes of Chinese amongst CCs themselves remains to be systematically investigated there are indications that Chinese Canadians may sometimes internalise low status ascriptions accorded by majority group members (e.g. Phang, 1983; also see Aboud & Christian, 1979). Today CCs have reason to feel more proud of their cultural and economic achievements because of their recent successes in improving their economic and residential status. Many of the new Chinese immigrants are professionals, technicians and business people who have attained high social status. Recent gains for the Cantonese language in education and the mass media has increased the status of the Cantonese language in Toronto. Method Subjects Two groups of 31 Chinese Canadians (CCs) with Cantonese as their mother tongue and whose parents were born in Hong Kong were recruited from three Christian churches in Toronto where services were conducted in Cantonese. The first generation Chinese Canadian group (FGCCs) had resided in Canada for at least ten years and their mean age was 30.5 years (16 females, 15 males). The second generation Chinese Canadians were all born in Canada (CBC) and had a mean age of 20.5 years (17 females, 14 males).

< previous page

page_137

next page >

< previous page

page_138

next page > Page 138

Procedure Subjects were tested in their Sunday school classrooms after the religious service. Each subject completed a three-part questionnaire. The first part consisted of the Subjective Vitality Questionnaire (SVQ, Bourhis et al., 1981) adapted for CCs in the Toronto setting. The SVQ involved rating CC and EC groups on 21 items measuring group vitality along the three dimensions of status, demography and institutional support (using 7-point Likert scales). This first part also included three items to directly measure subjects' perceptions of 'overall' ingroup-out-group vitality. These items assessed 'how strong and active' the ingroup and out-group had been in the past, are in the present, and would be in the future, in Toronto (see Bourhis & Sachdev, 1984). An additional item attempted to assess specific perceptions of 'how strong and active' the ingroup and outgroups would be when Hong Kong becomes part of mainland China in the 1990s. The use of 'strong and active' originated from Osgood, Suci & Tannenbaum's (1957) 'potency' and 'activity' dimensions which emerged from their semantic differential analyses. The second part of the questionnaire consisted of self-reported percentage use of Cantonese and English in nine domains of use including the home, church, stores, work setting, etc. Respondents then rated their degree of agreement with the use of Cantonese and English in the above language use domains on 7-point Likert scales. The third part of the questionnaire comprised items dealing with biographical, proficiency, identification and intergroup contact information (in the main, responses were on 7-point Likert scales). Results and Discussion General Background Information on Respondents Background information concerning the two groups of respondents is presented in Table 2. Items on the third part of the questionnaire were compared using t-tests for independent samples. Results indicated that FGCCs rated themselves as being able to understand and speak English slightly less well than CBCs. FGCCs also reported a stronger Cantonese and weaker Canadian accent in their English than CBCs. These results were expected since FGCCs grew up as native speakers of Cantonese in Hong Kong and emigrated to Canada in their late teens. Results in Table 2 also show that intergenerational differences in self-reported ability to understand and speak Cantonese were clearly in favour of the FGCCs. These results concur with previous findings amongst a variety of European and non-European minorities in Canada showing sharp declines in the knowledge and use of heritage languages by second generation group members (e.g. O'Bryan et al., 1976; Bourhis, 1987). Results in Table 2 show that on identification items, CBCs felt more Canadian, more proud to be Canadian, less Hong Kong Chinese and less proud to be Hong

< previous page

page_138

next page >

< previous page

page_139

next page > Page 139

Table 2 Information on First Generation Chinese Canadians (FGCC) and Chinese-born Canadian (CBC) respondents* Item FGCC t p CBC (df = 60) (n = 31) (n = 31) 1. How well speak English 5.81 6.77 5.04 0.001 2. How well understand English 6.16 6.74 4.16 0.001 3. How well speak Cantonese 6.48 2.61 10.73 0.001 4. How well understand Cantonese 6.51 3.22 9.27 0.001 5. Canadian accent in English 3.97 5.87 4.93 0.001 6. Cantonese accent in English 4.29 1.48 8.47 0.001 7. Feel Canadian 4.74 5.71 3.04 0.004 8. Feel Hong Kong Chinese 5.19 1.90 8.88 0.001 9. Proud to be Canadian 5.26 5.97 2.02 0.04 10. Proud to be Hong Kong Chinese 5.81 3.64 5.92 0.001 11. Proud to be English Canadian 3.90 5.13 2.62 0.01 *All scores are on a 7-point scale; higher scores indicate higher ratings Kong Chinese than FGCCs. Surprisingly, CBCs also felt more proud to be English Canadians than FGCCs, even though as oriental 'visible' minority group members they were unlikely to 'pass' as typical Anglo-Canadians. It appears that CBCs may well be assimilating to the Anglo-Canadian mainstream culture. The trend of CBC assimilation to the dominant EC group was further confirmed by analyses of contact and liking items (see Table 3). ANOVA results shown in Table 3 suggest that although both groups reported that they had more contact with ECs than CCs, CBCs liked contact with ECs more than contact with CCs. CBCs also reported they liked ECs more than CCs. No significant differences were obtained in the ratings made by FGCCs towards ECs and CCs, suggesting no differentiation in the liking for both ingroup and outgroup others. Overall, as with European Canadian immigrant groups (see Reitz, 1980), these results suggest that linguistic and cultural assimilation can be felt to be achieved by the second generation even amongst visible minority group members such as the Chinese Canadians in Toronto. Subjective Vitality Ratings Ss' ratings for vitality items were analysed using a 2 × 2 MANOVA procedure. The first factor (G) consisted of the two groups of CCs (FGCC & CBC), while the second factor (V), treated as a repeated measures factor, represented the actual

< previous page

page_139

next page >

< previous page

page_140

next page > Page 140

Table 3 FGCC and CBC reports of the amount of social contact and liking they have toward Chinese Canadians (CCs) and English Canadians (ECs)* Interaction F Main F Item Evaluation toward CCs Evaluation toward ECs Main F by FGCC Ss by CBC Ss by FGCC Ss by CBC Ss Group (G) Ratings G×R (df = 1, 60) (df = 1,60) (df = 1,60) Q1. Daily social contact you have 3.87 3.45 5.13 5.97 0.89 24.73 2.75 NS p<0.001 NS Q2. Social contact you like to have 5.36 4.10 5.13 5.65 1.56 5.96 10.72 NS p<0.02 p<0.002 Q3. Liking 5.39 4.42 5.16 5.55 0.86 8.17 18.38 NS p<0.006 p<0.001 *All scores are on a 7-point scale (strongly disagree = 1, strongly agree = 7) Table 4 Overall vitality rating. Perception of CC and EC vitality in the past, present, future and 1990s by FGCC and CBC subjects Overall Vitality Ratings of CC Ratings of EC Main F Interaction Main F Vitality Vitality Vitality F Group by by by Item by G×V (V) FGCC CBC FGCC CBC (G) (df = (df = 1,60) (df = Ss Ss Ss Ss 1,60) 1,60) Q24. Perceived Vitality of CC/EC groups 20-30 years ago in Toronto 1.65 2.00 5.52 5.19 0.00 172.98 1.59 NS p<001 NS Q22. Perceived CC/EC groups in Toronto today 2.90 4.54 5.36 5.16 9.78 56.34 20.28 p<0.003 p<0.001 p<0.001 Q23. Perceived Vitality of CC/EC groups 20-30 years from now in 4.00 4.74 5.41 4.81 0.06 17.15 14.29 Toronto NS p<0.001 p<0.001 Q25. Perceived Vitality of CC/EC groups in the 1990s in Toronto when 3.77 4.42 5.16 5.03 0.64 18.35 2.75 Hong Kong becomes part of mainland China NS p<0.001 NS

< previous page

page_140

next page >

< previous page

page_141

next page > Page 141

ratings of EC and CC vitality in Toronto. The overall MANOVA revealed a significant G × V interaction F(25,36) = 2.52, p<0.006; a main effect for repeated measures factor V, F(25,36) = 92.22, p<0.001; and a main effect for G, F(25,36) = 2.92, p<0.002. The univariates contributing to this main effect are discussed below. Results of questions drafted to directly measure group members' perceptions of overall vitality (i.e. 'how strong and active') are presented in Table 4. They show that both groups rated overall EC/CC subjective vitality in the past to be in favour of the ECs, thus accurately reflecting the 'objective' analyses presented in the introduction. Moreover, concurring with 'objective' estimates, results in Table 4 also suggest that both groups perceive a moderate gain in vitality today relative to the past. Vitality estimates after Hong Kong becomes part of mainland China suggest that both groups of subjects expected ECs to have more vitality than CCs in Toronto in the 1990s (cf. Young, Giles & Pierson, 1986; Pierson, Giles & Young, 1987). Inspection of the significant interactions on perceived vitality today and in the future suggested a systematic intergenerational divergence with the 'objective' estimates. Contrary to expectation, the second generation subjects (CBCs) perceived the differential in favour of EC vitality to be much smaller than first generation subjects (FGCCs), both today and in the future. Ratings by FGCCs concerning the vitality of CCs in Toronto today and in the future are relatively closer to the objective estimates than are the ratings of CBCs. The CBCs' tendency to overestimate the vitality of CCs was also reflected on one demography and three institutional support items of the Subjective Vitality Questionnaire (see Table 5). Perceived Demography Results in Table 5 show that both FGCC and CBC subjects perceived ECs to be proportionately greater in number, were emigrating at higher rates and being less endogamous than CCs in Toronto. Though both groups accurately acknowledged that ECs were in a majority relative to the CCs in Toronto, second generation subjects perceived the difference in favour of ECs to be smaller than first generation subjects. Perceived Institutional Support On all institutional support items, ECs and English Were consensually perceived to be better represented and have more institutional control than CCs and Cantonese (Table 5). However, on three items viz. political power, representation in government services and business/industry, there were some systematic intergenerational differences indicated by the significant interactions. Inspection of the means indicates that second generation subjects tended to perceive a smaller difference in favour of EC vitality over CC vitality on these three items than first generation subjects.

< previous page

page_141

next page >

< previous page

next page >

page_142

Page 142 Table 5 Subjective vitality ratings of CCs and ECs in Toronto today: 2 × 2 ANOVA results SVQ Items Ratings of CC Ratings of EC Main F Vitality Vitality Group by FGCC by CBC by FGCC by CBC (G) Ss Ss Ss Ss (df = C = Cantonese language 1,60) E = English language Q4. C/E Representation in Government Services 1.48 2.10 6.94 6.58 1.18 NS Q7. C/E Representation in Mass Media 2.36 3.32 6.45 6.26 3.52 NS Q 10. Teaching of C/E in Schools 1.61 2.03 6.81 6.61 0.87 NS Q14. C/E Representation in Business and Industry 2.26 3.13 6.45 6.26 4.28 p<0.04 Q17. C/E Representation in Churches 3.52 4.07 6.13 5.81 0.23 NS Q6. CC/EC Group Control of Business and 2.71 3.65 5.29 5.61 9.20 Economy p<0.004 Q13. CC/EC Group Political Power 1.77 2.89 5.90 5.55 2.96 NS Q18. CC/EC Group Cultural Representation 2.77 3.97 5.52 5.81 8.40 p<0.005 Q19. CC/EC Group Control of Mass Media 2.13 2.89 5.84 5.90 3.05 NS Q20. CC/EC Group Control of Schoolboards and 1.77 2.42 5.77 5.81 2.14 Schools NS Q3. Proportion of CC/EC Population 1.58 2.42 4.39 4.61 3.96 p<0.05

Main F Interaction F G × V Vitality (df= 1,60) (V) (df = 1, 60) 1656.53 p<0.001 127.55 p<0.001 941.31 p<0.001 299.82 p<0.001 71.90 p<.0.001 89.69 p<0.001 262.22 p<0.001 80.59 p<0.001 188.62 p<0.001 254.57 p<0.001 97.82 p<0.001

15.72 p<0.001 3.48 NS 3.70 NS 6.34 p<0.01 2.88 NS 1.63 NS 11.30 p<0.001 3.13 NS 1.71 NS 1.75 NS 1.47 NS

(table continued on next page)

< previous page

page_142

next page >

< previous page

page_143

next page > Page 143

Table 5 continued SVQ Items C = Cantonese language E = English language Q9. CC/EC Minority/Majority Q5. CC/EC Birthrates Q11. CC/EC Immigration to Toronto Q15. CC/EC Emigration from Toronto Q12. Extent of CC/EC Exclusive Ingroup Marriages Q1. Language Status of C/E in Toronto Q20. Language Status of C/E Internationally Q21. Group Wealth of CC/EC Q8. Group Prestige of CC/EC Q16. CC/EC Pride in Cultural Achievements

< previous page

Ratings of CC Vitality Ratings of EC Vitality' Main F by FGCC by CBC by FGCC by CBC Group Ss Ss Ss Ss) (G) (df = I, 60) 2.58 3.23 5.84 5.29 0.03 NS 4.19 4.39 4.55 4.29 0.01 NS 4.10 4.48 3.71 4.00 1.73 NS 2.10 3.52 3.42 4.45 24.66 p<0.001 5.55 5.68 4.61 4.13 0.56 NS 3.29 3.45 6.71 6.61 0.02 NS 2.74 3.03 6.23 6.19 0.42 NS 4.68 4.77 4.52 4.74 0.61 NS 3.48 3.77 5.48 5.45 0.32 NS 4.48 5.36 4.32 4.81 6.85 p<00.01

page_143

Main F Interaction F G × V Vitality (df = 1,60) (V) (df =1,60) 130.55 6.56 p<0.001 p<0.01 0.23 0.71 NS NS 2.01 0.02 NS NS 36.86 1.08 p<0.001 NS 21.29 1.30 p<0.001 NS 201.63 0.31 p<0.001 NS 109.82 0.26 p<0.001 NS 0.23 0.10 NS NS 89.34 0.69 p<0.001 NS 1.58 0.47 NS NS

next page >

< previous page

page_144

next page > Page 144

Perceived Status On three out of five status items ECs were consensually perceived to have higher vitality than CCs. First and second generation CCs perceived ECs to have higher prestige than CCs in Toronto and English to have more status than Cantonese both in Toronto and internationally (Table 5). Overall, in terms of the first aim of this study, results suggested that responses on the majority of SVQ items reflected the 'objective' estimates of vitality in that ECs were perceived to have higher vitality than CCs. However, contrary to expectation, second generation subjects tended to overestimate perceived ingroup vitality, relative to first generation subjects on some items. There was greater consensus on perceived status than on perceived demography and perceived institutional support dimensions. The Sociolinguistic Questionnaire Subjects' self-reported language use and evaluations were analysed using a 2 × 2 MANOVA procedure. The first factor (G) consisted of the two groups of CCs, while the second factor (L), treated as a repeated measures factor, represented the actual language (English or Cantonese) being rated. The overall MANOVA revealed a significant G x L interaction F(9,52) = 10.68, p<0.001, and a repeated measures main effect L, F(9,52) = 717.10, p<0.001. Univariate analyses of each usage and evaluation item were computed. Results of these analyses were statistically clear cut as shown in Table 6 and Table 7. Results shown in Table 6 clearly support the hypothesis that as the low vitality group, both groups of CCs reported speaking Cantonese less frequently than English in most public domains of language use including stores, at social events, at work, at school, in hospitals and while using government services. However, some interesting intergenerational differences also emerged. First generation Chinese Canadians self-reported higher levels of Cantonese usage than members of the second generation in almost all settings. The two informal but public settings where this was most apparent were in stores/restaurants and at social events, where FGCCs reported greater use of Cantonese than CBCs. In the most private setting, the home, which O'Bryan et al. (1976) considered the last bastion of minority language use, FGCCs reported much greater use of Cantonese than English. In contrast, the CBCs self-reported much higher levels of English than Cantonese usage in the home. A similar, but less pronounced pattern emerged in amounts of Cantonese and English usage in church settings (see Table 6). Evaluations of language use in these different settings reflected the actual self-reported patterns in all settings by both groups. Results in Table 7 suggest that though both FGCCs and CBCs agreed more with English than Cantonese usage in most settings, there were some systematic intergenerational differences. Specifically, first generation subjects were generally more supportive of Cantonese usage than second generation subjects, especially in informal and private settings.

< previous page

page_144

next page >

< previous page

page_145

next page > Page 145

Table 6 FGCC and CBC reports of percent Cantonese (c) and English (E) language usage in various Toronto settings* Sociolinguistic Percent Cantonese Percent English Main F Main F Language Use (L) Setting usage usage Group by FGCC by CBC by FGCC by CBC (G) (df = 1,60) Ss Ss Ss Ss (df = 1,60) Q1. Percent use of Cantonese/English (C/E) at 8.42 3.00 3.84 9.00 1.94 1.34 home NS NS Q2. Percent use of C/E in stores and restaurants 4.90 1.68 7.61 10.52 0.73 97.16 NS p<0.0001 Q3. Percent use of C/E in church 6.94 1.74 5.58 10.26 4.29 29.28 p<0.04 p<0.001 Q4. Percent use of C/E during community and 4.10 1.39 8.36 10.61 4.78 143.13 social events p<0.03 p<0.001 Q5. Percent use of C/E in classroom 2.16 1.32 9.84 10.68 0.00 317.45 NS p<0.001 Q6. Percent use of C/E in schoolyard 2.55 1.26 9.45 10.74 0.00 295.74 NS p<0.001 Q7. Percent use of C/E in hospital 2.16 1.26 9.87 10.74 0.11 403.04 NS p<0.001 Q8. Percent use of C/E at work 2.19 2.00 10.06 10.00 3.11 240.64 NS p<0.001 Q9. Percent use of C/E when using public 1.10 1.19 10.94 10.81 1.00 6826.2 services NS p<0.001 *All scores are on an 11-point scale (1 = 0%, 11 = 100%)

< previous page

page_145

Interaction F G ×L (df = 1,60) 74.39 p<0.001 27.37 p<0.001 55.63 p<0.001 19.43 p<0.001 3.08 NS 7.33 p<0.009 4.29 p 0.02 NS 0.92 NS

next page >

< previous page

page_146

next page > Page 146

Table 7 FGCC/CBC evaluations of Cantonese (C) and English (E) language usage in various Toronto settings* Main F Main F Evaluation Agreement with Sociolinguistic Agreement with (E) Group English setting Cantonese (df = 1,60) (G) usage usage by FGCC Ss by CBC Ss by FGCC by CBC (df = Ss Ss 1,60) Q1. At home 5.81 4.94 3.81 5.74 4.66 3.98 p0.03 p0.05 Q2. In stores and restaurants 3.84 4.00 5.77 6.29 2.20 85.49 NS p0.001 Q3. In church 5.22 5.23 4.94 6.10 4.28 1.89 p0.04 NS Q4. During community and social 4.32 4.16 5.81 6.39 0.78 64.91 events NS p0.001 Q5. In the classroom 3.39 3.48 6.03 6.32 0.52 76.98 NS p0.001 Q6. In the schoolyard 3.94 3.87 5.77 5.87 0.00 43.90 NS p0.001 Q7. In hospital 4.10 3.90 5.81 6.10 0.04 55.08 NS p0.001 Q8. At work 3.00 3.68 6.29 6.29 2.12 110.33 NS p0.001 Q9. When using public services 3.65 3.71 6.07 6.39 0.51 77.35 NS p0.001 * All scores are on a 7-point scale (1 = strongly disagree, 7 = strongly agree)

< previous page

page_146

Interaction F G × E (df = 1,60) 22.01 p0.001 0.60 NS 7.56 p0.008 2.60 NS 0.10 NS 0.08 NS 0.85 NS 1.45 NS 0.20 NS

next page >

< previous page

page_147

next page > Page 147

Indeed, there were clear indications of language shift within the home, and to a lesser extent in the church. Whereas FGCCs showed a clear preference for Cantonese use, CBCs preferred English to be the language of the home and the church. Language use and evaluation data in the present study are clearly related to the social identification and proficiency data. Cantonese was used and evaluated most favourably by bilingual FGCCs who strongly identified themselves as Hong Kong Chinese and spoke Cantonese (and English) fluently. Conversely, English use and evaluation was most favourable amongst CBC subjects who identified themselves most as English Canadian (but not Hong Kong Chinese) and spoke English (but not Cantonese) fluently. These results support literature positing a link between social identity, language proficiency, language usage and evaluation (e.g. Bourhis, 1979; Giles & Johnson, 1981, 1987). The hypothesised link between perceived vitality, language use and evaluation was supported in that both groups reported greater use of English (the high vitality language) than Cantonese (the low vitality language) in a wide range of public domains in Toronto. Contrary to expectation, first generation immigrants who perceived ingroup vitality to be lower (accurately) than their second generation counterparts, used Cantonese more, and were more favourable about its use in a wider range of settings than second generation subjects. Clearly the relationship between subjective vitality, language use and language evaluations needs to be reexamined. One promising avenue of explanation may lie in assessing the relationship between social identification, vitality and language use in the light of findings obtained by Johnson (1984) and Sweeting (1982). Their results (in Wales and Scotland respectively) suggested that when group members identify strongly with their ingroups, perceptions of low vitality may be viewed as threatening and spur efforts to maintain and even increase use of the ingroup language. It is thus suggested that FGCCs' favourable Cantonese usage and evaluation patterns may partially represent a strategy for enhancement of self and social identity due to low perceived ingroup vitality. There were some indications from previous studies (e.g. Liu, 1975) and from subjects' comments in the present study that first generation group members were quite concerned about the linguistic and cultural assimilation of second generation Chinese Canadians in spite of the sustained efforts to preserve their Chinese culture and language within the Canadian context. It is noteworthy that widespread support for heritage language retention has been obtained in a variety of European and non-European minority groups in Canada, even amongst ethnolinguistic group members who report total loss of their heritage language (e.g. O'Bryan et al., 1976; Bourhis, 1987). The relationship between CBCs' vitality perceptions, language usage/evaluation and social identification patterns was more puzzling. Findings from the present study showed that second generation Chinese Canadian subjects identified more with ECs and preferred contact with ECs more than with CCs. These results suggest that second generation Chinese Canadians are fast assimilating into the Anglo-Canadian mainstream culture. To the extent that CBCs have internalised

< previous page

page_147

next page >

< previous page

page_148

next page > Page 148

the dominant, negative views of CCs prevalent in Canadian society (e.g. Berry et al., 1977) and attempt to 'pass' as ECs they may view other CCs as 'outgroups' and even derogate them (also see Aboud & Christian, 1979). If CBCs consider CCs to be an outgroup and do not wish to identify with CCs, then small increases in 'objective' CC vitality may actually be perceived as large and threatening. Under these circumstances one would expect lesser agreement in favour of Cantonese than English usage in all domains, including the home, as was the case with the CBCs. Preliminary data collected amongst Cantonese speakers of Hong Kong origin in London (U.K.) provides some cross-cultural support for the above identity based analyses (Sachdev, Bourhis, D'Eye & Phang, in preparation). Undoubtedly, the discrimination suffered by the early immigrants contributed in large measure to the linguistic and cultural assimilation portrayed by their progeny, i.e. CBCs. Interestingly, Blanc (1987) suggested that though intense discrimination against Japanese immigrants to Canada led the second generation ('Nisei') to assimilate linguistically, there are some indications that the failure of assimilative strategies in deterring discrimination has led a significant minority amongst the third generation ('Sansei') to search for their cultural roots (cf. Reitz, 1980). In this vein it is noteworthy that McLeod & Wright (1971) have suggested that job status of earlier Chinese immigrants and their descendants is still decidedly low, and many remain victims of discrimination. More recent studies (e.g. Reitz, 1980) suggest that although Chinese Canadians today have fairly high job status, income and education, they are segregated in particular jobs, and their job status and income are not commensurate with their high levels of education. In the present study, first generation subjects (who also reported having experienced discrimination), are likely to have undergone quite a different life experience than the second generation who, as recent immigrants, were admitted under a 'points for professional and educational qualifications' rule. Biographical information from the subjects indicated that FGCC subjects (unlike CBCs) enjoy high status and income due to their high educational qualifications and membership in a variety of professional and corporate ranks. In contrast, CBCs are likely to have faced pressures from the home to maintain their ethnolinguistic distinctiveness while being powerfully 'drawn' in the opposite direction by the dominant majority AngloCanadian culture and the English language throughout their lives in Canada (e.g. Chan, 1976). Such differing life experiences of first and second generation subjects undoubtedly contribute to the variance in the results obtained in studies such as the present one. Clearly the relationships between the 'stages of an ethnic group's life-cycle' (Reitz, 1980) exemplified in succeeding generations of minority groups, and their perceived vitality, identity and language use/evaluations constitute important avenues for further research. Concluding Notes Results of the present study of intergenerational differences in vitality perceptions amongst Chinese Canadians allow us to make the following five concluding points:

< previous page

page_148

next page >

< previous page

page_149

next page > Page 149

First, though the 'objectively' high vitality of English Canadians was consensual, the Subjective Vitality Questionnaire uncovered systematic differences in perceptions between two generations of Chinese Canadians. Second, that Chinese Canadian vitality perceptions reported by first generation members accorded more with 'objective' estimates than the relatively inflated estimates of their second generation counterparts. Third, results suggested that subjective vitality estimates do not account for the observed patterns of language usage and language evaluation amongst intergenerational groups. Language usage patterns obtained in the present study reflect language retention ratios obtained with other first and second generation immigrants across Canada (see Bourhis, 1987). Fourth, that the relationship between perceived vitality, language use and evaluations is, in all likelihood, mediated by factors such as ethnolinguistic identification, intergroup discrimination and the 'drawing power' of the host community language and culture. Fifth, the present study clearly suggests that the optimal use of the Subjective Vitality Questionnaire in predicting language behaviour and attitudes is achieved only if it is used in conjunction with objective assessments of vitality and indices of group identification and language retention. Acknowledgements We are grateful to Michel Blanc, Paul Crook, Malcolm Edwards, Bill Gudykunst and Rob Kennedy for comments on earlier versions, technical assistance and translation. Notes 1. Requests for reprints and comments to: Itesh Sachdev, Dept. of Applied Linguistics, Birkbeck College, University of London, 43 Gordon Square, London WC1H 0PD. References Aboud, F. E. and Christian, J. D. (1979) Development of ethnic identity. In L. Eckensberger, Y. Poortinga & W. J. Lonner (eds), Cross-Cultural Contributions in Psychology. Netherlands: Swets and Zeitlinger. Berry, J. W., Kalin, R. and Taylor, D. H. (1977) Multiculturalism and Ethnic Attitudes in Canada. Ottawa: Ministry of Supply and Services. Blanc, M. (1987) Changing ethnolinguistic group relations in Canada, 1967-1986, London Journal of Canadian Studies, 4 (in press). Bourhis, R. Y. (1979) Language in ethnic interaction: A social psychological approach. In H. Giles & B. Saint-Jacques (eds), Language and Ethnic Relations. Oxford: Pergamon Press.

< previous page

page_149

next page >

< previous page

page_150

next page > Page 150

(1987) Social psychology and heritage language research: A retrospective and future trends for Canada. In J. Cummins (ed.), Heritage Language Research in Canada. Ottawa: Ministry of Supply and Services. Bourhis, R. Y., Giles, H. and Rosenthal, D. (1981) Notes on the construction of a 'Subjective Vitality Questionnaire' for ethnolinguistic groups, Journal of Multilingual and Multicultural Development, 2, 144-55. Bourhis, R. Y. and Sachdev, I. (1984) Vitality perceptions and language attitudes: Some Canadian data, Journal of Language and Social Psychology, 3, 97-126. Chart, I. (1976) The problems of language learning by Chinese immigrants. In Paper on the Chinese Community, Ministry of Culture and Recreation, Multicultural Development Branch (Ottawa), 9-30. Giles, H. and Byrne, J. (1982) The intergroup model of second language acquisition, Journal of Multilingual and Multicultural Development, 3, 1740. Giles, H. and Johnson, P. (1981) The role of language in ethnic group relations. In J. C. Turner & H. Giles (eds), Intergroup Behaviour. Oxford: Blackwell. (1987) Ethnolinguistic identity theory: A social psychological approach to language maintenance, International Journal of the Sociology of Language, 68, 256-69. Giles, H., Bourhis, R. Y. and Taylor, D. (1977) Towards a theory of language in ethnic group relations. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Giles, H., Rosenthal, D. and Young, L. (1985) Perceived ethnolinguistic vitality: The Anglo- and Greek-Australian setting, Journal of Multilingual and Multicultural Development, 3,253-69. Husband, C. and Saifullah Khan, V. (1982) The viability of ethnolinguistic vitality: Some creative doubts, Journal of Multilingual and Multicultural Development, 3, 193-205. Johnson, P. (1984) Language and ethnicity: The Welsh-English context. Unpublished Ph.D Thesis, University of Bristol, U.K. Johnson, P., Giles, H. and Bourhis, R. Y. (1983) The viability of ethnolinguistic vitality: A reply, Journal of Multilingual and Multicultural Development, 4,255-69. Kramarae, C. (1981) Women and Men Speaking. Rowley, Mass.: Newbury House. Liu, F. (1975) The elderly Chinese. In 'Papers on the Chinese community', Ministry of Culture and Recreation, Multicultural Development Branch, 31-36. McLeod, D. B. and Wright, E. N. (1971) Parents' Occupations, Students' Mother Tongues, and Immigrant Status: Further Analyses of the Ever)' Student Survey. Toronto: Board of Education for the city of Toronto, Research Department. Moscovici, S. (1981) On social representations. In J. Forgas (ed.), Social Cognition. London: Academic Press. O'Bryan, K. G., Reitz, J. G. and Kuplowska, O. M. (1976) Non Official Languages. Ottawa: Ministry of Supplies and Services. Osgood, C. E., Suci, G. and Tannenbaum, P. H. (1957) The Measurement of Meaning. Urbana, Ill.: University of Illinois Press. Pak, A. W-P., Dion, K. L. and Dion, K. K. (1985) Correlates of self-confidence with English among Chinese students in Toronto, Canadian Journal of Behavioural Science, 17,369-78. Phang. S-W. (1983) Evaluative reactions to Anglo and visible minority speakers in Canada. Unpublished Manuscript, McMaster University. Pierson, H., Giles, H. & Young, L. (1987) Intergroup vitality perceptions during a period of political uncertainty: The case of Hong Kong, Journal of Multilingual and Multicultural Development, 8, 451-60. Reitz, J. G. (1980) The Survival of Ethnic Groups. Toronto: McGraw Hill. Ryan, E. B., Giles, H. and Sebastian, R. J. (1982) An integrative perspective for the study of attitudes towards language variation. In E. B. Ryan & H. Giles (eds), Attitudes towards Language Variation: Social and Applied Contexts. London: Edward Arnold. Sachdev, I., Bourhis, R. Y., D'Eye, J. and Phang, S. (in preparation) Ethnolinguistic identity, vitality and language attitudes: Some U.K. data. London, Birkbeck College.

< previous page

page_150

next page >

< previous page

page_151

next page > Page 151

Sweeting, H. (1982) Aspects of psycholinguistic distinctiveness: A Scottish example. Unpublished B.A. thesis: University of Edinburgh. Young, L., Giles, H. and Pierson, H. (1986) Sociopolitical change and perceived vitality. International Journal of Intercultural Relations, 10, 459-69. Young, L., Pierson, H. and Giles, H. (1984) Perceived group vitalities: Effects of academic specialisation in Hong Kong, Linguistische Berichte, 101, 66-73.

< previous page

page_151

next page >

< previous page

page_152

next page > Page 152

< previous page

page_152

next page >

< previous page

page_153

next page > Page 153

Research Note Social Psychological and Linguistic Impediments to the Acquisition of a Second Nigerian Language among Yoruba and Ibo C. N. Mgbo-Elue University of Lagos The purpose of the research presented in this paper is to examine the language attitudes and stereotypes of the Yorubas and Ibos in Nigeria. The Yoruba tribe geographically inhabit the Western Region of Nigeria and live in a number of monarchical type clans clustered around the kingdom of Ire. Among the largest groups are the Oyo, Ijebu, Ondo and Ekiti. The Ibos occupy the greater part of the Eastern Region of Nigeria. Based on oral traditions, they are said to have originated from Egypt and moved southwards around the fourteenth century to their present location. With an increase in population the Ibos spread into five groups the Onitshas, the Nri-Awka, the Owerri, the Ogoja and the Western Ibo on the West bank of the River Niger. Expansion and isolation among the various groups resulted in deep dialectical differences in languages. Religion also is a barrier between the two tribes. The Yorubas are mostly Muslims, with a small minority of Protestant Christians, while the Ibos are mainly Roman Catholic. Very few Ibos are Muslim. Two studies were conducted to examine the language attitudes and stereotypes of the Yoruba and Ibos. The first study used an interview technique to assess language attitudes directly. The second study employed the matched guise technique to test language attitudes indirectly. Study I Subjects were chosen using a stratified random sampling technique. Subjects were selected at random in each of the three settings professional offices, business premises and educational institutions in Lagos and Enugu. At the secretariat, for example, respondents on every other floor were interviewed. When the person selected did not satisfy the sampling criteria (e.g. Leorghe was a native speaker of Yoruba or Ibo), others were approached until someone who satisfied the sampling criteria was found. The same procedure was applied at business premises and educational institutions. This procedure is similar to that used by Young & Willmott (1962).

< previous page

page_153

next page >

< previous page

page_154

next page > Page 154

Inspection of the civil service salaried structure in Nigeria suggested that it would be possible to obtain a good spread of respondents according to the characteristics of education, occupation and income simply by stratifying the sample in terms of mean income figures. Listing the seventeen grade levels by average annual wage and salary income of Nigerians in the labour force population, the list was cut into three levels: N8,000 and over (grade level 12 upwards), N4,000-N7,999 (grade levels 08-11), N1,200-N3,999 (grade levels 01-07). Since it is crucial to adequately represent all income levels and associated characteristics within the population, equal number of respondents were selected from each of the three levels. The final sample consisted of 288 subjects. The subjects were stratified by income (three levels), ethnic group (Yorubas, Ibos), and workers 20-40 years old, workers 41-60 years old, students 16-19 years old, and students 20-40 years old (there were 12 subjects for each of the 24 levels). There were no significant differences (based on chi-square tests) between Yorubas and Ibos with respect to (1) when they first learned the outgroup language, (2) from whom they learned the outgroup language, (3) where they learned the outgroup language, (4) how they learned the outgroup language, (5) why they learned the outgroup language, (6) to whom they speak the outgroup language, (7) how often they speak the outgroup language, (8) where they speak the outgroup language, (9) whether or not they think the outgroup language should be used on radio and television broadcasts, (10) whether or not there are times when they would rather speak the outgroup language than the ingroup language, (11) whether or not there are times they pretend not to understand the out-group language, (12) whether or not there are times they do not like to speak the outgroup language, (13) whether or not they speak the outgroup language to their children, (14) their attitudes toward other Nigerian languages, and (15) general attitude toward bilingualism. Significant differences emerged on several issues. First, a larger percentage of Ibos (78.4%) than Yorubas (51.6%) wanted their children to learn the outgroup language (X2 = 17.87, df = 1, p=0.001). Second, a larger percentage of Ibos (65.9%) than Yorubas (44.7%) liked the outgroup language (X2 = 11.74, df = 10, p<0.001). Third, a larger percentage of Ibos (81.8%) than Yorubas (65.3%) thought the outgroup language should be taught in schools (X2 = 9.25, df = 1, p<0.01). Fourth, a larger percentage of Ibos (62.0%) than Yorubas (46.5%) thought people generally liked the outgroup language (X2 = 6.20, df = 1, p<0.01). Fifth, a larger percentage of Ibos (79.3%) than Yorubas (49.1%) had positive attitudes toward other Nigerian languages (X2 = 21.39, df = 1, p <0.001), even when only monolinguals were questioned (Ibos = 69.2%, N = 71; Yorubas = 33.3%, N = 105; X2 = 15.20, df = 1, p<0.001). Sixth, a larger percentage of Ibos monolinguals (95.6%) than Yoruba monolinguals (82.5%) had positive attitudes toward bilingualism (X2 = 5.33, df = 1, p<0.05). Seventh, a larger percentage of Yorubas (66.4%) than Ibos (60.7%) selected their ingroup language as the best for Nigeria's national language (X2 = 142.8, df = 4, p<0.001; other options were Hausa, Edo, or other). The results of the present research revealed that the Nigerian respondents, with bilingual skills in the outgroup's language, learned the second language in infor-

< previous page

page_154

next page >

< previous page

page_155

next page > Page 155

mal situations from friends and peer groups. Second language learning took place, in most cases, in childhood and the reason for learning the second language was to be able to communicate. These results are consistent with Gardner's (1979) model which proposes that informal learning contexts facilitate better oral proficiency and more favourable non-linguistic outcomes than formal learning contexts. The bilinguals (112) interviewed displayed oral proficiency as revealed by standardised oral proficiency tests. Respondents' claim that they learned the second language because they want to communicate and to understand outgroup members is consistent with Macnamara (1973). He argued that children in the streets learn a second language faster than children in the classroom because the need to communicate in the streets is so great that children will improvise. In the classroom, however, children seldom have anything so important to say to the teacher that requires them to improvise. Another important category, according to Gardner (1979), is the social milieu which can be either monolingual or bilingual. The cultural beliefs prevalent in the social milieu influence the extent to which individual differences affect achievement in second language acquisition. The majority of the Nigerian respondents learned the second language in the capital city, Lagos. According to Gardner's model, their second language competence may be a function of situational anxiety. The results suggested that Yorubas and Ibos sometimes pretend not to understand the second language; often prefer to use the second language rather than their native language; sometimes do not like to speak the second language; and tend not to speak the second language to their children. The Yoruba and Ibo bilinguals used the second language occasionally with friends in such places as at work, in school and church, and around the neighbourhood. The respondents also claimed there are times when they prefer to speak in their native language, and other times when they chose to speak in the second language. The majority of members of both ethnic groups further indicated there are times when they pretend not to understand the second language. If these bilinguals associated with the outgroup through common residence, religion or other social category membership, they identify most with the group associated with their language at those times when they pretend not to understand the second language. The present results, therefore, support Turner's (1987) self categorisation theory which suggests that when individuals' subjective frame of reference and selfcategorisations shift, their language attitudes also shift. (Editor's Note: For a more complete discussion of this issue, see Abrams & Hogg's attitude in this volume.) Results of the present study indicated that parents who have favourable attitudes toward the Yoruba/Ibo community and who feel that learning the out-group language is valuable encourage their children to study the other language and may also speak the language to their children. The data suggested that the Ibo hold a significantly more positive attitude toward their children learning Yoruba than the Yorubas hold toward their children learning Ibo. The possible discrepancy between what respondents say they do or would do and what they actually do should be pointed out. While the Ibo say they would like their children to learn Yoruba, it should not necessarily be inferred that they actually do so when

< previous page

page_155

next page >

< previous page

page_156

next page > Page 156

confronted with actual situations. Studies of self-evaluation tests (Labov, 1966; Trudgill, 1972) reveal that respondents 'over-report'. Labov found that respondents tended to over-report the use of high-status pronunciations, while Trudgill reported that both working-class and middle-class males in Norwich claimed to use non-standard forms when in actual fact they did not. The data revealed that the Ibo monolinguals have a positive attitude toward Yoruba language while the Yoruba monolinguals have a negative attitude toward Ibo language, but there is no significant difference between the bilinguals and the monolinguals. These results are consistent with the data on attitudes toward other Nigerian languages; the Ibo bilinguals and monolinguals expressed a positive attitude toward other Nigerian languages, while the Yoruba bilinguals and monolinguals expressed a negative attitude toward other Nigerian languages. One possible explanation for the negative attitude toward second language learning expressed by the Yoruba is that the linguistic environment affects them differently than it affects the Ibo (Shockey, 1984). The structural differences between Yoruba and Ibo could explain the negative attitude toward language learning expressed by the Yoruba group. The linguistic background of the Yorubas and the language distance between Ibo and Yoruba (see Giles, 1979, discussion of hard and soft linguistic boundaries) makes it more difficult for Yorubas to learn Ibo. Beebe & Giles (1984), however, claim that it is possible for a native speaker of French to successfully learn Chinese with the right attitude and motivation. Structural differences, therefore, cannot fully account for the Yoruba not learning Ibo. The results indicated that Yorubas and Ibos strongly identify with their own ethnic group. There was a strong relationship between ethnic group and choice of a national language, with each group naming its own language as the best choice for a national language. Just under a third of the respondents, however, did not want any of the three major languages as a national language. This has implications for Nigerian language planning. As Tollefson (1981) proposed in his model of centralised and decentralised language planning, opinions of local people have to be sought and incorporated into language policies, otherwise the result will be inefficiency, the withdrawal of plans (e.g. the Philippines), or serious conflict (e.g. India). Study II The second study tested whether there is a relationship between the ethnic group of the listener and the ethnic language used by the speaker and stereotypes of the speaker. Drawing on Hogg, Joyce & Abrams (1984) research, it was hypothesised that judges evaluate speakers of their ingroup language more positively than they evaluate speakers of outgroup language. Following Lambert (1967) and Ryan & Sebastian's (1980) research, it was further hypothesised that sex and age of the judge, and whether the judge is a monolingual or bilingual affect their evaluations of outgroup speakers. Initially, eight bilingual speakers were recruited for the Study. There were five female and three male speakers. A short paragraph (10 lines) of a text on the high

< previous page

page_156

next page >

< previous page

page_157

next page > Page 157

cost of commodities in the markets was read by each speaker to familiarise him or her with the text. They were then asked to read the passage in the two languages, while their speech was taped. The Yoruba voices were recorded on one tape with short spaces between each voice and the Ibo voices on another tape. Two university language specialists in Yoruba and Ibo were asked to judge whether the voices were native speakers of the languages. Only one female speaker was judged to be perfect in native-like qualities of both languages. Finally, two groups of university undergraduates Yorubas and Ibos (20 in each group) were asked to listen to the tapes. The students were asked to listen to one voice at a time and afterwards to give the ethnic group of the speaker and the reasons for their answers. The eight voices were judged for native-like qualities. The results corresponded with the previous assessment of the linguists. Only one speaker was judged to be a native speaker of both languages, by all 40 judges. Ideally, it was important to have more than one speaker, but it would have been unwise to use bilingual speakers who would be assessed on the one hand as a Yoruba and on the other as a non-native speaker of Ibo. Ideally, too, both sexes should have been represented. Given the results of the pretest, only one speaker was used in the Study. Subjects were asked at the time of the first study whether they would like to participate in an experiment which required them to be at the university for a period of not more than one hour. Those who agreed to take part were recorded by their identification numbers. A list of 60 judges was obtained and they included individuals from the three socio-economic groups, the two ethnic groups, adults and students and both sexes. The Yoruba judges were to assess the speakers in Lagos university language laboratory and the Ibo judges at the Enugu Campus. The passage had to contain a text common to both native speakers and also it had to be interesting to sustain the attention of the judges. The content of the passage was about the disadvantages to family life of the increased materialistic tendency in Nigeria. The problem was a current one, as the country is going through a period of rapid affluence and economic competitiveness. The judges were told that this was an experiment to find out if they could assess a person's personality by listening to her voice. They were told that they should imagine themselves listening to a speaker who is in another room, to ignore the language being used, and the content of the speech (see Gardner & Lambert, 1972). They were instructed to imagine the type of a person the speaker was, whether tall or short, intelligent or not, dependable or not. The judges also were instructed to listen to each voice and then when the tape stops to assess the personality of the speaker using the rating scales provided. An example was displayed on the board. The personality traits were discussed with the judges. The status/prestige traits included competence, intelligence, ambitiousness, self-confidence, courage and good leadership. Personal integrity/character traits included sincerity, kindness and dependability. Altogether 22 personality traits were used. A factor analysis of the 22 traits yielded four factors: (1) competence included competence, intelligence, ambitiousness, self-confidence, courage and good leadership; (2) personal integrity included dependability, sincerity, kindness, conscientiousness and good character; (3) social attractiveness included sociable, likeable, good sense of

< previous page

page_157

next page >

< previous page

page_158

next page > Page 158

humour, affectionate, entertaining, tall, and handsome/beautiful; and (4) negative attributes included shy, cold, wicked, and arrogant. Separate analyses were conducted for each of the three indices that were constructed. 1 There were significant main effects for ethnicity [F(1,54) = 12.16, p<0.001] and speaker's language [F(1,54) = 6.10, p<0.01]. The main effects indicate that the Ibo judges evaluated the voices more negatively than the Yoruba judges. The Ibo voice was evaluated more positively on competence traits than the Yoruba voice. The main effects were moderated by a significant interaction effect [F(1,54) = 226.40, p<0.001]. The interaction suggests that Yoruba judges rate the Yoruba speaking voice with qualities of competence, prestige and leadership, but do not attribute the same qualities to the Ibo speaking voice. Conversely, the Ibo judges reacted more favourably to the Ibo voice on competence traits than they did to the Yoruba voice. The effects of mono- vs. bilingual and speaker language also were examined for competence. The main effect for mono-bilingual was not significant [F(1,54) = 2.52, p = ns], but there was a significant interaction effect [F(1,54) = 16.01, p<0.001]. The bilingual judges rated the Ibo voice as competent, but the monolingual judges were undecided whether the speakers were competent or not. Further, age did not have a significant main effect [F(1,54) = 1, p = ns], but it also interacted with speaker language [F(1,54) = 4.30, p<0.05]. The adults (41-60 years old) rate the Ibo speaker as incompetent; while the students (20-40 years old) rate both speakers as borderly competent. Ethnicity did not have a significant main effect [F(1,54) = 3.25, p = ns] on personal integrity, but there was a significant main effect for speaker's language [F(1,54) = 5.98, p<0.05]. There also was a significant interaction effect [F(1,54) = 223.90, p<0.001]. The combined effects mean that the Yoruba judge rated the voice speaking in his native language more favourably on personal integrity trait than the Ibo voice, and vice versa. There was no significant difference in personal integrity based on ethnicity [F(1,54) = 2.54, p = ns], but there was a significant interaction effect with speaker's language [F(1,54) = 12.49, p<0.001]. The bilingual judges rated the Ibo speaker higher on personal integrity and the Yoruba speaker lower. The monolingual judges rated both voices similarly, they were undecided whether the speakers had personal integrity or not. The analysis of social attractiveness by ethnicity [F(1,54) = 1, p = ns] and speaker's voice [F(1,54) = 1, p = ns] did not yield significant main effects. There was, however, a significant interaction effect [F(1,54) = 54.51, p<0.001]. The judges rated the speaker favourably when she used the ingroup language, but unfavourably when she used the outgroup language. Thus, Yoruba judges rate the Yoruba speaker favourably on social attractiveness scale, but not the Ibo speaker. Likewise the Ibo judges rated the Ibo speaker favourably, but not the Yoruba speaker. Mono-bilingualism did not have a significant effect on social attractiveness [F(1,54) = 1, p = ns], nor was there a significant interaction with speaker's voice [F(1,54) = 1.64, p = ns]. Ethnicity [F(1,54) = 9.99, p<0.01] and speaker's voice [F(1,54) = 9.98, p<0.01] had a significant effect on the negative attributes. The Ibo judges were more negative in their evaluation of the voices than the

< previous page

page_158

next page >

< previous page

page_159

next page > Page 159

Yoruba judges. The Ibo voice was evaluated more positively than the Yoruba voice. The main effects were tempered by a significant interaction effect [F(1,54) = 71.73, p<0.001]. The language of the speaker determined whether she was rated favourably or not. Yorubas refused to affirm the Yoruba speaker's negative attributes, even though they were affirmative about the Ibo speaker's negative attributes. The Ibo judges affirmed the Yoruba speaker's negative attributes, but denied that Ibo speaker had such attributes. The Ibo speaker was downgraded in evaluation by the Yoruba judges on all traits. Ibo judges, however, evaluated the Ibo speaker positively and favourably on all traits. In the same vein, the Ibo judges evaluated the Yoruba speaker very negatively on all the personality traits. This tendency to downgrade a speaker from the opposite group is similar to the findings of Lambert et al. (1960) and Hogg, Joyce & Abrams (1984). The only difference between the Nigerian results and the Lambert et al. results, is that neither the Yoruba nor the Ibo judges tended to downgrade the speaker in their native language, whereas, the French judges actually downgraded the French guises. Lambert et al. interpreted this phenomenon as evidence of a 'minority group reaction'. There was a tendency for the Ibo judges to evaluate the Yoruba guise slightly more favourably on the social attractiveness scales entertaining, showing affection and good sense of humour but not on the basic competence or personal integrity scales intelligence, dependability. These results bear some similarity to the findings of Carranza & Ryan (1975) who found that there was a tendency for the low-prestige language (Spanish) to receive a more positive evaluation on 'solidarity' traits such as friendliness and trustworthiness than on status traits. These studies and others reveal that a person's speech serves as a convenient identifier and this is readily used to evoke a social stereotype, based on ethnic group, race or social class. The Nigerian data reveal that ethnic group is the variable which has the greatest impact on listeners' evaluations. Thus, the judges tended to associate the speaker in their native language with qualities of leadership, competence, and integrity but to downgrade the guise in the second language. The variable age also was significant in the assessment of competence. The adults (41-60) rated the Ibo guise as more competent than the Yoruba guise, while the students (20-40) rated both guises as fairly competent. Also, the bilingual judges evaluated the Ibo speaker more positively on personal integrity and social attractiveness scales than they evaluated the Yoruba guise. The monolingual judges were in all cases more moderate in their evaluations of both guises. One possible explanation for this is that there were more Ibo bilinguals (17) than Yoruba bilinguals (5) in the sample. No significant findings emerged based on sex differences between the judges, nor on socio-economic differences. Conclusions The opinion that every Nigerian should know a second 'major' Nigerian language (Hausa, Yoruba and Ibo) prompted this research. The research is the first

< previous page

page_159

next page >

< previous page

page_160

next page > Page 160

to ascertain the state of bilingualism among Yorubas and Ibos in each others' language, attitudes toward the learning of these two languages, attitudes toward the learning of other Nigerian languages, learning domain, process and use of the second language. The purpose of the research also was to attempt to examine ethnic stereotypes based upon language and attitudinal constraints on second language learning. The first study, using the direct method of eliciting attitudes, revealed that the Ibo sample expressed a more positive attitude toward the teaching and learning of Yoruba and also toward other Nigerian languages. The Yoruba sample expressed negative attitudes toward learning Ibo language and other Nigerian languages. Furthermore, there were more Ibo bilinguals in Yoruba language than Yoruba bilinguals in Ibo. Study I, therefore, suggests that if individuals' attitudes toward other languages are positive and favourable, they will learn the other group's language. A passive attitude has a negative influence on the acquisition of the target language. This can be taken a step further and it can be inferred that a speaker of several Nigerian languages has a more tolerant attitude toward other Nigerian languages and their speakers. The second study, using the indirect method of eliciting attitude, indicated that the Yoruba and Ibo hold biased and stereotyped impressions about speakers of the other group's language. The Ibos, however, viewed the Yoruba speaker as sociable, with good sense of humour, quite affectionate and entertaining. In terms of basic personality traits such as dependability, intelligence, competence and ambitiousness, both groups view the other group as clearly inferior. These stereotypes may explain why many of the bilinguals never actually use the second language. Respondents from the two ethnic groups claim that there are times when they choose not to use the second language and times when they pretend not to understand the other language. The hypothesis that polyglots have more tolerant attitudes toward the languages they are able to speak and to its speakers should be modified slightly in this instance. The Ibo sample contained more bilinguals (over half) and the group expressed a more positive attitude to Yoruba language, but the group does not hold a positive attitude of Yoruba speakers. The Yoruba group, on the other hand, did not express a positive attitude toward Ibo or toward Ibo speakers. The discrepancy between attitude and behaviour (Parkin, 1977; Trudgill, 1972; Labov, 1977) should be borne in mind when discussing respondents' claims of when, how and why they use a particular language. This limits the extent to which generalisations can be made from the findings of the present research. Note 1. A more complex factorial design could have been used. The manipulation of too many factors simultaneously, however, would have resulted in empty cells. The simplest factorial design (2 × 2) was sufficient to study interactions and test hypotheses. Since this study is primarily concerned with the study of ethnic groups, the full result of the two-way analysis of ethnic group of the judges by the voice of the speaker are reported. Only the significant findings on all other tests are discussed.

< previous page

page_160

next page >

< previous page

page_161

next page > Page 161

Bibliography Agheyisi, R. and Fishman, J. A. (1970) Language attitude studies: A brief survey of methodological approaches, Anthropological Linguistics, 12, 131-57. Beebe, L. and Giles, H. (1984) Speech accommodation theories: A discussion in terms of second language acquisition, International Journal of Sociology of Language, 46, 5-32. de Vos, G. and Romanucci-Ross, K. (1975) Ethnicity: Vessel of meaning and emblem of contrast. In G. de Vos & K. Romanucci-Ross (eds), Ethnic identity: Cultural Continuity and Change. Palo Alto, CA: Mayfield. Fishman, J. A. and Spolsky, B. (1976) The Whorfian hypothesis in 1975: A sociolinguistic evaluation. In H. Fisher & R. Dias-Guerrero (eds), Language and Logic in Personality and Society. New York: Academic Press. Giles, H. (1970) Evaluative reactions to accents, Educational Review, 22, 211-27. (1979) Ethnicity markers in speech. In K. Scherer & H. Giles (eds), Social Markers in Speech. Cambridge: Cambridge University Press. Giles H. and Bourhis, R. (1973) Dialect perception revisited, Quarterly Journal of Speech, 59, 337-42. Giles, H. and Powesland, P. F. (eds) (1975) Speech Style and Social Evaluation. London: Academic Press. Hogg, M. A., Joyce, N. and Abrams, D. (1984) Diglossia in Switzerland? A social identity analysis of speaker evaluations, Journal of Language and Social Psychology, 3, 185-96. Labor, W. (1977) The social motivation of a sound change. In W. Labor (ed.), Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. Lambert, W. E. (1967) The social psychology of bilingualism, Journal of Social Issues, 23, 91-109. Lambert, W. E., Hodgson, R. C., Garner, R. C. and Fillenbaum, S. (1960) Evaluational reactions to spoken language, Journal of Abnormal and Social Psychology, 6, 44-51. Parkin, D. J. (1977) Emergent and stabilized multilingualism: Polyethnic peer groups in urban Kenya. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Ryan, E. B. and Sebastian, R. J. (1980) The effects of speech style and social class background on social judgements of speakers, British Journal of Social and Clinical Psychology, 19, 229-33. Shockey, L. (1984) All in a flap: Long term accommodation in phonology, International Journal of Sociology of Language, 46, 87-96. Trudgill, P. (1972) Sex, covert prestige and linguistic change in urban British English of Norwich, Language and Society, 1, 179-96. (1974) Sociolinguistics. Harmondsworth, England: Middlesex. Young, M. and Wilmott, P. (1962) Family and Kinship in East London. London: Penguin.

< previous page

page_161

next page >

< previous page

page_162

next page > Page 162

< previous page

page_162

next page >

< previous page

page_163

next page > Page 163

Research Note Afro-American Identity Labels and Communication Effectiveness 1 Michael L. Hecht Communication Research Center, Department of Communication, Arizona State University and Sidney Ribeau Dean of Undergraduate Studies, California State University, San Bernadino The ethnic experience in modern society provides insight into the rich diversity of human experience. Through the eyes of ethnic communities it is possible for one to gain access to the unique relationship existing between cultural experiences and communication codes by means of which these ethnic groups eventually must interact with the dominant social order. The Afro-American2 community provides a useful model to study these phenomena. The range of the 'black experience' in America seems to have been regulated by chance, political activism, and the emergence of an aggressive group consciousness. Beginning with the enslavement of Africans by the Portuguese toward the end of the fourteenth century and extending through the Black Consciousness movement of the 1960s, Afro-American social reality clearly indicates a desire for self definition and social control. One means of investigating this process is by the study of the semantic labels used by this group. 'An individual's language is intricately bound up with her or his sense of identity and group consciousness' (Smitherman, 1977). Victimisation, integration, and assertive self-identification characterise the evolution of Afro-American semantic labelling. Discussing the importance of semantic labelling for this group, Smitherman (1977) comments: What's in a name, then? Everything, as we acknowledge that names are not merely words but concepts which suggest implications, values, history and consequences beyond the word or 'mere' name itself. Words fit into a total symbolic and cultural system and can only be decoded within the context of that system. This labelling approach to the study of ethnic communication is consistent with Berger & Luckman (1966) and others (Eastman, 1985; Giles, Bourhis & Taylor, 1977; Lambert, 1967; Tajfel & Turner, 1979) who believe that social reality is constructed largely through linguistic processes such as self-identity.

< previous page

page_163

next page >

< previous page

page_164

next page > Page 164

To date the studies of the Afro-American community have been largely descriptive. Most research has focused on communicator style, seeking to articulate the typical indicants of this ethnic group (Asante & Noor Al-Deen, 1984; Bachman & O'Malley, 1984; Crosby, Bromley & Saxe, 1980; Hecht & Ribeau, 1984; Ickes, 1984; Jarmon, 1980; Kochman, 1970; 1982; LaFrance & Mayo, 1978). This research documents a unique AfroAmerican style for language (Kochman, 1972, 1982; Smitherman, 1977), nonverbal behaviour (Baxter, 1970; Johnson, 1971; Jones & Aiello, 1973; LaFrance & Mayo, 1976; Rosegrant & McCroskey, 1975), and interaction patterns (Asante & Noor Al-Deen, 1984; Gudykunst & Hammer, in press; Ickes, 1984; Kochman, 1982; LaFrance & Mayo, 1978). While describing typical conversational behaviour, these studies do not address the question of communication effectiveness. Since research in mainstream culture indicates that typical and effective communication are not isomorphic (Hecht, 1984), the effectiveness orientation is theoretically and pragmatically important. Without an understanding of effective communication, theory is based on normative or dysfunctional accounts of interaction. The study of effectiveness provides a comprehensive view of the range of human interaction while infusing theory with the ability to impact and improve pragmatic, everyday interaction. Recent research indicates that Afro-Americans differ from other ethnic groups in the issues they see as most salient to their intra-ethnic communication satisfaction (Hecht & Ribeau, 1984) and rules (Collier, 1982; Collier, Ribeau & Hecht, 1984). In addition, Hecht & Ribeau (1987) have described an Afro-American perspective on interethnic communication with whites. In that study seven primary issues were identified for AfroAmericans' satisfaction with communication with whites: acceptance, emotional expressiveness, authenticity, negative stereotyping, understanding, goal attainment, and powerlessness. Acceptance refers to a feeling that the white person accepts, confirms, and respects the Afro-American interactant. Emotional expressiveness is the communication of feelings. Authenticity in communication is expressed by genuineness and open disclosure on the positive side, and evasiveness on the negative. Rigid racial categorisation constitutes Negative stereotyping, which also involves distortion of the Afro-American person's individuality. Feelings of Understanding emerge when interactants feel their meanings are shared. Goal attainment is manifested when interactional objectives are met. Finally, Powerlessness is invoked when Afro-Americans feel controlled, manipulated, and trapped. These issues underlie Afro-Americans' satisfying communication with whites. These studies provide useful descriptions of Afro-American social behaviour. However, to group the studies and claim the designation of a single style may be to assume that all Afro-Americans share a common ethnic identity. It may, in fact, be true that some of the descriptions are associated with one identity, while others are associated with a second identity. Identification with a particular group membership influences the way individuals perceive their environment and communicate with those around them (Tajfel, 1978; Turner, 1982). In fact, Banks (1981) has suggested that 'there is no one ethnic identity among Afro-Americans that we can delineate, as social scientists have sometimes suggested, but many complex

< previous page

page_164

next page >

< previous page

page_165

next page > Page 165

and changing identities among them.' His explanation emphasises the 'emerging,' 'changing' and 'complex' nature of ethnic identity and is consistent with Smitherman's discussion of Afro-American semantic evolution. A careful reading of Afro-American history supports this claim. Survival has required Afro-American flexibility. One test of a culture is its ability to dynamically adapt to a changing world. When this test is applied for AfroAmericans, the results might be found in multiple ethnic designations. Fairchild (1981) investigated what he called 'race names' to determine how the designations Black, Negro, and Afro-American were stereotyped by whites. Results indicated that Blacks were more likely to be viewed as lazy, loud, and rude (Afro-Americans least so) and Afro-Americans were more frequently viewed as talkative (Blacks least so). This approach shifts the focus from perceived self-identification to attributed identification, but points out the differences in behaviour attributed to semantic designations. Not only do Afro-Americans attribute different meaning to their semantic labels, whites also view Afro-Americans through a semantic filter. In a recent study Hecht & Ribeau (in press) identified four ethnic labels denoting four separate identities. The labels, Black, Black American, BlackAfro-American, and Afro-American, differentiated self-perception, communication style, and relationship to mainstream culture. Upon reanalysis it appeared that only three distinct groups exist: Blacks, Black-Americans, and Afro-Americans. Blacks choose their label because it is acceptable and is based on social consensus. They describe their communication style in terms of slang, verbal aggressiveness and willingness to talk. They are mildly patriotic and implicitly accepting of the status quo. They are the most conservative group. Black-Americans derive their identity from being both Black and American, utilising a label which reflects their background, place of origin and heritage. They see their communication as characterised by dialect and code switching and they are aware of assimilation pressures but maintain a developing sense of ethnic consciousness. Afro-Americans, the most political group, emphasise ancestry and birth as the key elements of their label. They see their communication style as distinguished by nonverbal and dialectic differences. They are pragmatic about their relationship to the mainstream group, realising that their roots are not American, but they must succeed in the existing system if they are to get ahead. Ethnic identity labels reflect different attitudes toward self and mainstream America. The identity labels align themselves along a continuum from most to least assimilationist. People who label themselves Afro-American are the less assimilated, while those labelling themselves Black or BlackAmerican are more assimilated. In addition, both Afro-Americans and Black-Americans recognise a duality in their identity; one that reflects both cultural roots and mainstream culture. Accordingly, the identity labels filter and shape the ways group members interact with white America. Given the existing descriptive research it is important to now focus on how people who apply different ethnic identity labels interact effectively with white Americans. It is our thesis that as a result of their semantic labelling, Blacks,

< previous page

page_165

next page >

< previous page

page_166

next page > Page 166

Black-Americans, and Afro-Americans differ in their relationships with whites. More specifically, we theorise that differences should be manifested in the issues perceived salient in interethnic conversations. Our research question was: Interactants who label their ethnic identity Black, Black-American, and Afro-American differ in perceptions of Acceptance, Emotional Expressiveness, Negative Stereotyping, Authenticity, Understanding, Goal Attainment, and Powerlessness as issues in conversations with whites. There were 61 respondents in this study, 37 from California, 13 from New York, and 11 from Arizona. Respondents were recruited from college classes as well as community groups, reflecting a diverse sample. The average age was 31.3 and the average family income was $28,362. The sample was predominantly female (43). Respondents were asked to recall satisfying and dissatisfying conversations they had recently experienced with someone of a different ethnic group and respond to a questionnaire. Respondents were chosen on the basis of self-identification as a United States-born Black, Black-American, or Afro-American who had conversed with a white American. Not all respondents could recall both satisfying and dissatisfying conversations. As a result, a total of 105 conversations were described. After recalling the conversation, respondents answered a series of questions guiding their recall and description: 1. describe the location and topic of conversation, 2. describe and explain what the respondent did or said that was satisfying or dissatisfying, 3. describe and explain what the other did or said that was satisfying or dissatisfying, 4. describe and explain anything else in the conversation that was satisfying or dissatisfying, and 5. indicate the ethnicity of their conversational partner. The next part of the questionnaire asked respondents to provide the label they use to describe their ethnic identity. They also were asked to supply demographic information regarding their gender, age, and family income. As a preliminary analysis a series of Chi Square tests were performed to insure that the ethnic identity labels were independent of demographic factors. Yates' correction factor was used when expected frequencies were less than 5. The analyses supported this assumption. Nonsignificant Chi Square values were observed for the relationship between the labels and age (3.46, df = 2), gender (0.57, df = 2), family income (4.01, df = 6), and geographic location (0.74, df = 4). These analyses indicate that the labels are truly reflective of ethnic identity and not the other demographic variables. Two research assistants were provided with definitions and examples of the seven conversational issues: acceptance, authenticity, understanding, goal attainment, powerlessness, negative stereotyping, and emotional expressiveness. They discussed the categories and independently coded 20 of the questionnaires. Inter-rater reliability was calculated to be 0.67. This reliability is similar to that reported

< previous page

page_166

next page >

< previous page

page_167

next page > Page 167

in a separate study for two different coders (Hecht & Ribeau, 1987) and is adequate considering the difficulty of the coding task and the goal of describing group rather than individual differences. Thereafter, a single assistant coded the remaining questionnaires. A series of Chi Squares were computed to test the relationship between ethnic identity labels and the issues. Separate Chi Squares were computed for each issue. Since powerlessness and negative stereotyping were used almost exclusively in descriptions of dissatisfying conversations and goal attainment was used almost exclusively for satisfying conversations, these analyses were applied to 3 × 2 tables (identity labels by not used/used). Ail other analyses were applied to 3 × 3 tables (identity labels by not used/used for satisfying conversation/used for dissatisfying conversation). The Chi Square values for acceptance (3.45), emotional expressiveness (2.98), understanding (1.08), goal attainment (0.25), authenticity (2.16), and powerlessness (1.90) were not significant at the 0.05 level. Only the value for negative stereotyping (6.13) was significant at the 0.05 level. BlackAmericans and Afro-Americans perceived stereotyping in their dissatisfying conversations more often than Blacks. The results of this study pose important questions regarding the form and content of Afro-American/white communication. Our earlier research clearly identified both themes in Afro-American/white conversations (Hecht & Ribeau, 1987) and a differentiation in the selection of semantic labels and communication behaviours by Afro-Americans (Ribeau & Hecht, in press). Why then was negative stereotyping the only theme that differentiated groups? First, one must consider the context of American race relations as the framework of Afro-American/ white dialogue. Historically Afro-Americans have separated the nuances of their culture from personal interaction with whites. Differences that were in-group remained in-group for purposes of survival. Variations were secondary to confrontation with the major threat: racial discrimination and prejudice. Thus an Afro-American interpretation of our five themes should elicit the strongest response when considering discrimination. Negative stereotyping, as a variant of discrimination, is an issue which is central to all Afro-American/white relations. A position on this issue becomes a starting point for group cohesion, separating nationalist from assimilationist, separatist from integrationist and liberals from conservatives. The topic saliency of negative stereotyping demands a response which emerges from the group consciousness. Thus in our study the group labelling themselves 'Black' report experiencing the least amount of negative stereotyping, while those labelling themselves 'Afro-American' report the most. This is consistent with previous analyses (Ribeau & Hecht, in press) in which Blacks were found to be the most conservative and assimilationist group. Conversely, Afro-Americans were found to be the most nationalistic, and are highly aware of the duality of their African and American heritages. Blacks, then, are either less aware or sensitive to stereotyping, or behave in a way that results in less actual stereotyping by whites. A second critical factor in interpreting our results is the growing sense of socio-ethnic identity among Afro-Americans. The semantic labels selected all focus on unique aspects of cultural experience, providing insight into an ethnic world.

< previous page

page_167

next page >

< previous page

page_168

next page > Page 168

However, when this in-group experience confronts the 'other' (in this case whites), it may dissolve for the sake of survival for the entire group. AfroAmerican, Black-American, and Black speak to one's ethnicity, yet they all emerge from the experience of American racism. It is perhaps more important to speak with one rather than many voices when confronting a historical nemesis. Therefore, the themes of acceptance, emotional expressiveness, authenticity, understanding, goal attainment, and powerlessness, which address more conversational style issues, do not differentiate among the ethnic identities. Instead, Afro-Americans, as a group, share these issues in their conversations with whites. Negative stereotyping, unlike the other themes, calls for a response that is conditioned by the shared screenings of the semantic label. Just as many Afro-Americans converse with whites in the terms of mainstream culture by code switching, they might also unconsciously submerge variations of ethnic identity in the context of social discourse. Finally, ethnic identity is determined by many dimensions which are perhaps best measured in relation to context and issue saliency. Our themes differentiate communication between Afro-Americans and whites, but might not speak to the subtleties of in-group variations that are influenced by topic and socio-racial sensitivity. Specifically, this study addressed the types of Black ethnic identity. Other researchers have addressed the degree of identity, labelling this construct 'ethno-linguistic vitality' (Bourhis, Giles & Rosenthal, 1981; Giles, Bourhis & Taylor, 1977). It may be that the conversational issues addressed above are related to the degree of identification or the combined effect of type and degree. Future research should address this issue. Notes 1. An earlier version of this paper was presented at the 3rd International Conference on Language and Social Psychology, Bristol, July 1987. 2. The label Afro-American is used here. However, other labels (e.g. Black, Black-American) could also be used to denote this ethnic group. It is not the authors' intent to impose a label. The label, Afro-American, will be used for convenience and stylistic consistency. References Asante, M. K. and Noor Al-Deen, H. S. (1984) Social interaction of black and white college students: A research report, Journal of Black Studies, 14, 507-16. Bachman, J. G. and O'Malley, P. M. (1984) Yea-saying, nay-saying, and going to extremes: Black-white differences in response styles, Public Opinion Quarterly, 48, 491-509. Banks, J. A. (1981). Multi-ethnic Education: Theory and Practice. Boston, MA: Allyn and Bacon. Baxter, J. C. (1970) Interpersonal spacing in natural settings, Sociometry, 33,444-56. Berger, P. L. and Luckman, T. (1966) The Social Construction of Reality. Garden City, NY: Doubleday & Co., Anchor Books.

< previous page

page_168

next page >

< previous page

page_169

next page > Page 169

Bourhis, R. Y., Giles, H., & Rosenthal, D. (1981). Notes on the construction of a 'Subjective Vitality Questionnaire' for ethnolinguistic groups. Journal of Multilingual and Multicultural Development, 2, 145-155. Collier, M. J. (1982) Communication Rules and Perceived Outcomes within and between Domestic Cultural Groups: A Comparison of Intracultural and Intercultural Communication among Acquaintances. Unpublished doctoral dissertation, University of Southern California. Collier, M. J., Ribeau, S. and Hecht, M. L. (1986) Intracultural communication rules and outcomes within three domestic cultural groups, International Journal of Intercultural Relations, 10, 439-57. Crosby, F., Bromley, S. and Saxe, L. (1980) Recent unobtrusive studies of black and white discrimination and prejudice: A literature review, Psychological Bulletin, 87, 546-63. Eastman, C. J. (1985) Establishing social identity through language use, Journal of Language and Social Psychology, 4, 1-26. Fairchild, H. H. (1981) Black, Negro, or Afro-American? The Differences are Crucial! Paper presented at the National Association of Black Psychologists meeting, Denver, CO. Giles, H., Bourhis, R. and Taylor, D. M. (1977). Towards a theory of language in ethnic group relations. In H. Giles (ed.), Language, Ethnicity and Intergroup Relations. London: Academic Press. Gudykunst, W. B. and Hammer, M. R. (in press) The influence of ethnicity, gender, and dyadic composition on uncertainty reduction in initial interaction, Journal of Black Studies. Hecht, M. L. (1984) Satisfying communication and relational labels: Intimacy and length of relationship as perceptual frames of naturalistic conversations, Western Journal of Speech Communication, 48, 201-16. Hecht, M. L. and Ribeau, S. (1984) Ethnic communication: A comparative analysis of satisfying communication, International Journal of Intercultural Relations, 8, 135-51. (1987) A Black Perspective on Interethnic Communication Effectiveness. Paper presented at the meeting of the Speech Communication Association, Boston, MA. Hecht, M. L. and Ribeau, S. (1987) Socio-cultural roots of ethnic identity: A look at Black America, Journal of Black Studies. Ickes, W. (1984) Compositions in black and white: Determinants of interaction in interracial dyads, Journal of Personality and Social Psychology, 47, 330-41. Jarmon, C. (1980) Racial beliefs among blacks and whites: An evaluation of perspectives, Journal of Black Studies, 11, 235-47. Johnson, K. (1971). Black kinesics: Some nonverbal communication patterns in Black culture. In L. Samovar & R. Porter (eds), Intercultural Communication: A Reader (1st edition). Belmont, CA: Wadsworth. Jones, S. E. and Aiello, J. R. (1973) Proxemic behavior of black and white first-, third-, and fifth-grade children, Journal of Personality and Social Psychology, 25, 21-27. Kochman, T. (1970) Toward an ethnography of black American speech behavior. In N. E. Whitten, Jr. & J. F. Szwed (eds), Afro-American Anthropology. New York: Free Press. (ed.). (1972) Rappin' and Stylin' out: Communication in Urban America. Urbana: University of Illinois Press. (1982) Black and White: Styles in Conflict. Chicago, IL: University of Chicago Press. LaFrance, M. and Mayo, C. (1976) Racial differences in gaze behavior during conversations: Two systematic observational studies, Journal of Personality and Social Psychology, 33, 547-52. LaFrance, M. and Mayo, C. (1978). Cultural aspects of nonverbal communication. International Journal of Intercultural Relations, 2, 71-89. Lambert, W. (1967) The social psychology of bilingualism, Journal of Social Issues, 23, 91-109. Rosegrant, T. J. and McCroskey, J. C. (1975) The effects of race and sex on proxemic behavior in the interview setting, Southern Speech Communication Journal, 40, 408-20.

< previous page

page_169

next page >

< previous page

page_170

next page > Page 170

Smitherman, G. (1977) Talkin' and Testifyin': The Language of Black America. Boston, MA: Houghton Mifflin Company. Tajfel, H. (1978) Social categorization, social identity, and social comparison. In H. Tajfel (ed.), Differentiation between Social Groups. London: Academic Press. Tajfel, H. and Turner, J. (1979) An integrative theory of intergroup conflict. In W. Austin & S. Worchel (eds), The Social Psychology of Intergroup Relations. Monterey, CA: Brooks/Cole. Turner, J. (1982) Towards a cognitive redefinition of the social group. In H. Tajfel (ed.), Social Identity and Intergroup Relations. Cambridge: Cambridge University Press.

< previous page

page_170

next page >

< previous page

page_171

next page > Page 171

Courses and Conferences Fulbright (USA/UK) Commission International Colloquium 1988 'Communication, Health and the Elderly' University of Wales Conference Centre, Newtown, Mid Wales, U.K. 12-15 December, 1988 Social support, health and the elderly The social construction of health and ageing The elderly in caring/medical settings Sociolinguistics of intergenerational communication Social cognition and health communciation Convenors: Nikolas Coupland, University of Wales, Cardiff Howard Giles, University of Bristol John Wiemann, University of California, Santa Barbara Offers of papers (before 1st July, 1988) and requests for further details to: Dr N. Coupland, Fulbright 1988 Centre for Applied English Language Studies, University of Wales, Cardiff, Colum Drive, Cardiff CF1 3EU, Wales, UK. Partial financial subsidy towards accommodation/travel may be available to some succesful paper-proposers through Fulbright sponsorship.

< previous page

page_171

next page >

< previous page

page_172 Page 172

Index A Accent 45-58 Afrikaans 105-130 Asian languages 73-86 B Bilingualism 35-44 British-Asians 73-86 C Canada 131-152 Cantonese 131-152 Categorisation 105-130 Code-switching 1-14 Communication intercultural 35-44 intergroup 163-170 problems 35-44 Communicative competence 163-170 Culture 163-170 D Discourse 15-34 E Ethnic boundaries 15-34 identity 1-14, 15-34, 35-44, 87-96, 131-152, 153-162, 163-170 Ethnicity 105-130, 163-170 Ethnolinguistic vitality 1-14, 87-96, 131-152 Ethnomethodology 15-34 H Hindus 73-86 I Ibo 153-162 Identity 59-72, 105-130 Immigrants 59-72 Intergroup 87-96, 163-170 J Japanese 35-44 L Language 1-14 attitudes 45-58, 59-72, 87-96, 131-152, 153-162 choice 73-86 intergenerational 131-152 shift 59-72 strategies 105-130 M

Matched guise 45-58 Markedness 15-34 Minority groups 59-72 Muslims 73-86 N Nigeria 153-162 S Scotland 45-58 Second language acquisition 153-162 Self-categorisation 45-58 Sikhs 73-86 South African 59-72 Social categorisation 59-72, 87-96 comparison 105-130 identity 1-14, 45-58 Sociolinguistics 35-44 Speech accommodation 1-14 Stereotypes 87-96, 163-170 V Values 73-86 W Work settings 15-34 Y Yoruba 153-162

< previous page

page_172