![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
Access to resources through ethnic group membership is often presumed to affect the intensity of ethnic identification. We examine this premise using survey data on three ethnic groups in Mauritius: Creoles, Hindus, and Muslims. Two key findings emerge from our research. First, access to material resources explains only a modest proportion of total variation in ethnic identification within each group. Second, the resources that affect ethnic identification differ significantly across groups. Access to political goods through group membership affects Hindu identification but is unrelated to ethnic identification among Creoles or Muslims. Conversely, access to economic goods affects Creole and Muslim identification but has no effect on Hindu identification. Explaining these group differences leads us beyond a basic means–ends instrumentalist model to identify conditions that likely mediate the relationship between individual interests and collective identification including the divisibility of economic goods relative to political goods in Mauritius.
Acknowledgements
An earlier version of this article was presented at the 2004 annual meeting of the American Sociological Association, Philadelphia. The research presented here was supported by grants from the National Science Foundation (NSF 92-114) and the Social Science Research Council. I would like to thank all those who participated in the study and assisted in data collection. I would also like to thank Humphrey Costello, the Editors of this journal and the anonymous reviewers for their insightful and helpful comments.
Notes
1. The terms ‘circumstantialism’ or ‘situationalism’ are often used in place of ‘instrumentalism’ to reflect newer approaches to ethnicity that focus on the conditions under which ethnic-based interests and identities emerge. These approaches maintain an instrumentalist foundation by suggesting that ethnic identification is strengthened when interests are perceived to be furthered by ethnic membership (see Hempel Citation2004). The term ‘instrumentalism’ is used throughout the paper because this term focuses analytic attention on this core foundation which, we argue, has not been adequately tested.
2. Thus, for example, Banton (Citation1983) writes of the presence of ‘group consciousness’, Hechter (Citation1975) of ‘status solidarity’ and later (2000) ‘group solidarity’, Olzak (Citation1992) of ‘ethnic solidarity’, Gurr (Citation1993) of ‘strong sense of group identity’, and Hardin (1995) of ‘group identification’.
3. We would like to express our appreciation to an anonymous reviewer for emphasizing these points.
4. In an effort to discourage ethnic communalism, the Central Statistics Office of Mauritius, which conducts the national census, has not collected data on ethnicity since 1983. The percentages reported are based on estimates cited in a number of sources (see Eriksen Citation1991; Hollup Citation1993; Carroll and Caroll 2000).
5. The area described refers to the main island of Mauritius where more than 99 per cent of Mauritians live. The Republic of Mauritius is also composed of Rodrigues Island, the Cargados Carajos Shoals, and Agalega. These areas were not included in the study. On the main island, neighbourhoods tend to be more class than ethnically based, although there are some ethnic-based residential patterns in Mauritius: most members of the Chinese community live in urban areas; a majority of whites live in the more affluent sections of Curepipe and Floréal; Muslims constitute a strong majority in Plain Verte, an area on the outskirts of the capital of Port Louis; many rural villages are predominantly Hindu; and Creoles living in rural areas tend to be concentrated along the coastline. These patterns aside, there is no ethnic group that is distinctively removed from other ethnic groups in Mauritius.
6. Many actually live on top of one another. Construction of homes in Mauritius often takes place vertically for all but the most advantaged as families add stories to their home to accommodate the families of grown children.
7. Hollup (1993, p.110) further argues that although caste identities are less important that those based on ethnic affiliation, castes or caste associations have emerged for political purposes as pressure groups along the same lines as other ethnic communities. As such, caste association is also likely to influence voting patterns in Mauritius.
8. The Central Statistics Office of Mauritius reports that between 1990 and 2000, the number of Mauritians who reported that Creole was the ‘usual or most often spoken language at home’ rose from 62 per cent to 70 per cent while the use of Asian languages, including Hindi and Bhojpuri, declined (Mauritius CSO Citation2000).
9. There is some suggestion (Laville Citation2000; Boswell Citation2005) that the label ‘Creole’ may be rejected by some who adopt the label ‘Catholique’ to bring themselves semantically on par with Hindus and Muslims. We found little evidence of this in our interviews, however. This may be due to our asking respondents their religion prior to asking their ethnicity.
10. Official statistics on income, education, and occupation by ethnicity are not available (see note 4). Descriptive statistics from the survey are provided in .
11. Following each election, the Electoral Supervisory Commission looks at the self-reported ethnic background of the sixty elected members and determines how many come from the four recognized communities in the Constitution: Hindus, Muslims, Chinese, and the ‘General Population’, a residual category mainly including Creoles and Franco-Mauritians. If the proportion of elected members is less than that community's proportion of the Mauritian population, ‘best losers’ – candidates from the underrepresented ethnic communities who stood for election but who were defeated – make up the difference (Srebrnik Citation2000; see also Mathur Citation1991). To determine which communities are underrepresented, authorities use the 1972 official population census, as this was the last census that required Mauritians to report their ethnicity.
12. Because no single party can get the majority alone or by relying on mobilizing support from one ethnic group, political parties have to forge strategic alliances and coalitions in order to obtain and maintain political power (Hollup Citation1993, p. 263). For an extensive historical account of coalitions in Mauritius see Bowman (Citation1991) and Srebrnik (Citation2002).
13. A striking illustration of this occurred recently when the Government redesigned the nation's paper currency. Previously, Mauritian currency included denominations written in English, Arabic, Hindi, and Tamil along with an illustration of Mauritius’ first Prime Minister, Sir Seewoosagur Ramgoolam. In 1998, a newly designed currency was released which featured drawings of leaders from each of Mauritius's major ethnic communities and a small but significant change altering the positions of the languages so that the Hindi demarcation now appeared above the Tamil demarcation. This alteration was seen as a sign of Hindu usurpation by many in Mauritius, both Tamil and non-Tamil alike (Bunwaree Citation2002). Demonstrations held in protest of the new currency shut down the capital of Port Louis for two days and resulted in the forced resignation of the governor and director general of the Central Bank as well as the removal of the bank notes from circulation at an estimated cost of over $2 million (Rs 58 million) to the Mauritian economy.
14. Prior to its implementation, the questionnaire was translated into French and Creole – the two most widely spoken languages in Mauritius. This gave respondents the option of conducting the interview in either language. Two experienced translators worked independently to translate the questionnaire into both languages. These translations were then compared, edited, and synthesized in direct consultation with the translators. A language expert from the University of Mauritius then reviewed each of the final translations to ensure consistency in meaning.
15. Of those interviews not completed, 181 were non-responses and 58 were refusals or a non-response/refusal rate of 23 per cent. The remaining 4 interviews were terminated early by the respondent or the interviewer.
16. Latent variables are variables that are not directly observable and, therefore, cannot be measured directly. Multiple manifest or observable indicators are used instead as approximate indirect measures. Common examples of latent variables include social status, intelligence, and economic development (Bollen Citation1989).
17. Given the overlap between religion and ethnicity in Mauritius, respondents were first asked their religion prior to being asked their ethnicity.
18. In addition to political and economic resources, psychological motivations including self-esteem, social status, optimal distinctiveness, or uncertainty reduction are also emphasized in the literature (see, for example, Tajfel and Turner Citation1979; Horowitz Citation1985; Brewer Citation1991; Hogg Citation2005). Indeed, there is very suggestive evidence in this literature that collective identification is motivated primarily by the desire for ‘psychic satisfactions’ that come from positive social comparison potentially involving, but not limited to, the relative economic and/or political status of an ethnic group.
19. A number of criteria were used for defining the number of factors underlying these indicators: first, each factor had an eigenvalue over one and fell above the ‘break’ in scree test plots; second, each factor accounted for over 10 per cent of the common variance; and, third, the factor could be interpreted, i.e. at least three indicators loaded significantly on the factor, the factor has substantive meaning that is shared by the indicators, and indicators loading on other factors were distinct.
20. Ancillary analyses indicated the need for an additional parameter correlating the errors between access to homes and loans. The rationale for this specification is that access to housing is predicted, in part, by access to loans.
21. Tests of configural equivalence examine whether the groups being compared employ a similar conceptual framework. Tests of metric equivalence examine whether the slopes relating the observed variables to the corresponding latent construct differ significantly across groups. Detailed tables with results of these tests for the measures are available upon request.
22. Although specifying additional factors that contribute to a more complete explanation of variation in ethnic identification is beyond the scope of this paper, we are attracted by arguments that address the ‘assumed givens’ (Geertz Citation1963) of ethnic membership, most notably those that examine the relationship between social categorization and cognition (e.g. Jenkins Citation1994, 2000; Gil-White Citation1999; Brubaker, Loveman and Stamatov Citation2004; Horowitz Citation2002).
23. Additional factors include lack of information about alternatives, costs of moving (transfer costs), and the strength of personal ties (see Hechter Citation1987, pp. 46–7).
24. Alternatively, economic opportunities could be so available they require no signalling.
25. These perceptions are likely to be exacerbated by the relative lack of national coordination among Creole associations and organizations and by internal cleavages based on class and skin-tone within the Creole community (see Laville Citation2000; Boswell Citation2005). Hindus also have strong corporate groups with extensive ties of mutual commitments while Creole kin groups are less extensive and based on a stronger individualism (see Hollup Citation1993). These factors may contribute to the continuing popular support among many Creoles for the Mouvement Militant Mauricien (MMM), a multiethnic political party which mobilizes on class rather than ethnic-based commonalities.
26. Where a collective good is defined by Olson (Citation1965, p.14) as ‘any good such that if any person … in a group … consumes it, it cannot feasibly be withheld from others in that group’.