M. Hechter and D. Okamoto
Department of Sociology, University of Washington,
Department of Sociology, University of Arizona, Tucson, Arizona
Given the global trend of increasing ethnocultural diversity and the outbreak of nationalist movements based on cultural, linguistic, and territorial identities, this review focuses on social and political mechanisms that lead to the emergence of minority group collective action. This kind of collective action is seen as a function of three necessary conditions: the formation of distinctive social identities, the overcoming of free riding, and the development of institutional structures promoting the demand for greater autonomy. The article examines the debates, theories, and empirical evidence concerning these three conditions. We conclude by noting that the most important impediment to progress in this field is the relative paucity of historical and cross-national databases that are required to test many of the theories in the literature.
minority groups; ethnicity; nationalism; conflict; social identity; collective action.
This article is one of 13 articles in the Domestic Political Violence and Civil War compilation.
The boundaries of states rarely coincide with those of national or ethnic groups (Ra’anan 1990). Most states contain several culturally distinct groups whose language, religion, tradition, and historical experiences are not shared and often are at odds with one another. In 1972, Connor (1972) noted that only 12 of the 132 states then extant were culturally homogeneous. Since then, the disjuncture between state and ethnonational boundaries has grown larger. Moreover, the United Nations estimated that in 1993, as many as 100 million individuals lived outside their country of birth or citizenship, and that at least 18 million were refugees (Williams 1994). Migration, fragmentation, and annexation have been contributing to ethnocultural diversity on a global scale. These trends are of interest to political scientists because multiethnic states are at some risk of experiencing violent ethnic conflict, secessionist movements, and government-sponsored ethnic cleansing.
The current prevalence of secessionist and nationalist movements raises deep questions about how states foster unity among people with diverse cultures and historical experiences, how individuals construct social identities, and how minority groups manage to engage in collective action. In this essay, we review the principal social and political factors that lead to the emergence of nationalist identities and movements among minorities—that is, groups whose members share a distinctive social identity based on their culture, language, territory, or individual ascriptive traits.
Readers are urged to consult prior reviews on closely related topics: ethnicity (Yinger 1985), ethnic conflict in comparative perspective (Williams 1994), ethnic mobilization (Olzak 1983), and ethnic and nationalist violence (Brubaker & Laitin 1998). This review, by contrast, is organized on the premise that minorities affect statewide political outcomes only under three conditions: when (a) they have distinctive social identities, (b) they have the potential to engage in collective action, and (c) their political demands are not likely to be met by the existing institutional arrangements. We discuss each of these conditions for the mobilization of minority groups by examining the debates and theories in the literature and corresponding empirical evidence.
A word about this evidence. The bulk of the literature on minorities focuses on particular movements and on the unique historical, economic, political, and cultural circumstances that led to them. These case studies provide a rich narrative that can clarify concepts, apply—and sometimes test—theoretical perspectives, and suggest new hypotheses. Comparative case studies, in particular, seek to extract patterns from sets of broadly comparable cases. Fruitful typologies and conceptual schemes have also developed out of this methodological tradition. For example, typologies of nationalism (Mughan 1979, Meadwell 1983, Rogowski 1985) and a taxonomy of macropolitical forms of ethnic conflict regulation (McGarry & O’Leary 1993) have been useful in describing different movements.
Large data sets permit more rigorous testing of rival theoretical propositions both within and across polities. These data sets typically consist of survey data gathered from individual respondents (Sekulic et al 1994, Diez Medrano 1994, Bollen & Diez Medrano 1998) or are compiled from official statistics about the sociodemographic characteristics of minority groups and reports of collective-action events, generally derived from samples of newspapers (Olzak 1992, Gurr 1993). Case studies provide the historical background from which statistical models can be sensibly interpreted and understood (see Ganguly 1996). Most analyses, whether qualitative or quantitative, focus on culturally distinct groups with strong nationalist movements (Belanger & Pinard 1991, Hechter 1999, Leifer 1981). However, some analyses focus on entire states (Bollen & Diez Medrano 1998).
CONDITIONS PROMOTING THE FORMATION OF MINORITY SOCIAL IDENTITY
The Micro Foundations of Minority Social Identity
Although there are a number of alternative microscopic approaches to social identity [including those based on evolutionary biology (van den Berghe 1981, Whitmeyer 1997)], social psychological theories have received the greatest empirical validation. The first prominent social psychological theory of identity formation is known as realistic group conflict theory. Borrowing its inspiration from Marx and Simmel, early social psychologists (Lewin 1948, Sherif 1966) argued that the principal cause of identification with groups is mutual dependence (see Brown 1986:Ch. 6). This mutual dependence, in turn, arises from peoples’ experience of a common situation or predicament (Rabbie & Horwitz 1988). Experimental research confirms that people arbitrarily selected into groups and subjected to differential treatment on that account indeed develop strong identification with their groups. However, the effect of differential treatment goes beyond mere identification: Given decisions about the allocation of resources, individuals tend to provide more of the resources at their disposal to in-group rather than out-group members. The key implication of realistic group conflict theory is that social identities are likely to be a by-product of intergroup stratification.
More recent research, however, challenges the idea that dependence and conflict over resources are necessary conditions for the development of group identity. Currently, the most popular explanation for the formation and salience of group identification is social identity theory (Tajfel 1974, 1981, Tajfel & Turner 1979, Abrams & Hogg 1990, Hogg et al 1995). According to this theory, individuals have a fundamental desire to attain positive self-esteem. This desire, in turn, motivates two kinds of sociocognitive processes: categorization, in which individuals are led to distinguish between social groups; and self-enhancement, in which people come to emphasize norms and stereotypes that favor the in-group.
The novel finding in this research is that in order to produce group identification and in-group bias in a given population, all a third party has to do is categorize individuals (studies of this sort are part of the “minimal group” research paradigm). In one well-known experiment, for example, Tajfel (1970) arbitrarily divided in two a group of British university students based on their preference for an abstract image painted by either Klee or Kandinsky. Once categorized, the members of each group began to develop a social identity and made biased allocations of resources favoring the in-group. There is some evidence that given categorization, such discrimination in favor of one’s group leads to an increase in members’ self-esteem (Oakes & Turner 1980, Lemyre & Smith 1985).
At the same time, social identity theory also implies that identification patterns vary with the stratification of groups. According to the theory, individuals identify with high-prestige groups because doing so contributes positively to their self-esteem. By the same token, individuals avoid identifying with low-prestige groups because doing so lowers their self-esteem. This hypothesis too has largely been supported by the empirical literature. However, some research on the link between self-esteem and a negative social identity is inconsistent with social identity theory. Sachdev & Bourhis (1984) found that group members engaged in comparable levels of discrimination whether they were in high- or low-prestige groups. Other experimental research indicates that social identification with a group may actually be increased when the group is threatened or stigmatized. For example, Turner and his colleagues (1984) reported that in-group defeat produces even higher levels of in-group preference than does success among committed group members. Many studies also indicate that identification with disadvantaged or stigmatized groups is associated with high rather than low individual self-esteem (Rosenberg 1979, Cross 1985, Crocker & Major 1989). There is more evidence that in-group favoritism enhances self-esteem than that low self-esteem motivates intergroup discrimination (Hogg & Abrams 1990, Crocker et al 1993).
If everyone desires positive self-esteem, then how do people come to terms with membership in low-prestige groups? They may (a) attempt to gain entry into the dominant group (social mobility), (b) interpret their group’s traits as a badge of pride rather than disparagement (social creativity), or (c) engage in collective action designed to raise their group’s status (social change) (Tajfel & Turner 1979). These strategies have been explored at some length (van Kippenberg 1984, van Kippenberg & van Oers 1984, Taylor et al 1987, Jackson et al 1996). For example, Ellemers and her colleagues (1988) found that group members engaged in collective strategies when intergroup boundaries were highly impermeable. However, the evidence is mixed: Jackson et al (1996) found that permeability of group boundaries had the predicted effects on social creativity strategies but not on social mobility strategies.
The evidence lends partial support to social identity theory. As predicted, categorization does lead to in-group identification, and identification, in turn, leads to higher self-esteem among group members. However, a negative social identity does not always lead to low self-esteem, and—a finding at variance with the theory—the strategies designed to combat such an identity are not necessarily carried out. In laboratory-created groups, experiments suggest that group members engage in social mobility and social creativity strategies to combat a negative social identity, but the results are mixed for real social groups (Ellemers et al 1988, Jackson et al 1996). Moreover, social identity theory does not predict which coping strategy will be selected in a given condition.
If members perceive that their group has shared interests, the salience of their social identity independent of third-party categorization is raised. Because social identity theory assumes that there is no mutual dependence between group members (Turner 1987), this finding is important. Flippen et al (1996:883) argue that “people form social categories on the basis of similarity/dissimilarity with others, but these categories do not become in-groups/out-groups until some kind of perceived mutual dependence creates the belief that members of different categories will act for or against self-interest.” When subjects perceive that they have interest interdependence with other group members in minimal group experimental settings, they will favor these members, but categorization is not sufficient to produce in-group bias in the absence of mutual dependence (Locksley et al 1980, Rabbie et al 1987, 1989, Flippen at al 1996). All told, then, social psychological research highlights the effects of two independent mechanisms for the formation of all types of social identities—interest interdependence and categorization.
Each of these mechanisms is implicated in two important macrosocial processes associated with the formation of durable and distinct group identities: state building, which often results in the categorization of minority groups, and the development of labor markets, which often produces mutual dependence.
The Macro Foundations of Minority Social Identity
Distinctive social identities are nurtured by a collective sense of a common ancestry, a common religion, and, in the most general sense, a common history. Early theorists of ethnicity and nationalism claimed ethnic communities were natural, primordial, and given (Shils 1957, Geertz 1963, Isaacs 1975). These theorists focused on the content of culture to explain the intensity and meaning of ethnic attachments and viewed racial and ethnic boundaries as fundamental, ascriptive, and immutable. Many contemporary scholars agree that ethnicity is not rigidly ascribed but is socially constructed (Barth 1969, Horowitz 1975, Anderson 1983, Gellner 1983, Calhoun 1998, Hechter 1999). For example, several studies have documented movement across racial and ethnic boundaries, where new groups were created from previously diverse cultural groups: Cornell (1988) on pan-Indian consciousness; Padilla (1985) on Latino identity; Espiritu (1992) on Asian- American panethnicity; Young (1976), Nnoli (1989) on pan-Igbo ethnicity in Nigeria; and Young (1976), Nagata (1981) on pan-Malay ethnicity in Malaysia. Some scholars even emphasize the notion that ethnic communities are created by the interests of ethnic entrepreneurs and state elites (Wallerstein 1980, Gellner 1983, Brown 1998, Hechter 1999). Brass (1991) is a forceful proponent of this position: “Ethnicity and nationalism are … political constructions. They are creations of elites who draw upon, distort, and sometimes fabricate material from the cultures of the groups they wish to represent in order to protect their well-being or existence or to gain political and economic advantage for their groups as well as for themselves.” The claim that ethnic group boundaries are not primordial, but socially constructed, is now the dominant view.
That said, there remain dissenters who argue that the content of culture cannot be discounted as a condition for the construction of ethnic boundaries. Conversi (1995:82) argues that the construction of ethnic identities relies on the “pre-existing diffusion of shared symbols and cultural elements as well as on memories of a shared past and myths of a common destiny.” Likewise, Smith (1996) suggests that economic and political circumstances are important in explaining why nationalisms emerge, but just as vital are “deep” ethnosymbolic resources (Armstrong 1982, Hastings 1997). In particular, the durability and character of a given nationalism can be largely explained by analyzing the ethnohistorical, religious, and territorial heritages that its proponents draw on. These ethnoheritages set the limits and provide the pattern within which modern elites must operate if they are to be successful in mobilizing their conationals. Likewise, Ollapally & Cooley (1996) argue that in order to understand why certain identity affiliations associated with identity movements are ultimately selected over other competing identities within the same space, the content of the identities must be examined.
Whereas some writers see the roots of national identity as extending far into premodern history, others (Anderson 1983, Hobsbawm & Ranger 1983) view them as invented traditions emanating from the rise of market society (“print capitalism”) or state-building activities (Hobsbawm 1992). Smith (1998) offers a useful summary of the controversy about the modernity of nationalism. Perhaps the most judicious conclusion is that national identity is a (relatively) modern construction that is sometimes built on prior cultural foundations.
Direct Rule, State Building, and National Mobilization
Prior to industrialization, central rulers were unable to exert direct control over geographically distant territories, many of which were culturally distinct from the core areas of their states (Hechter 2000). As a result, they were compelled to employ some form of indirect rule to govern these peripheries (Tilly 1990). Under indirect rule, central rulers designated either local authorities or their own nominees as their agents. In return for guaranteeing the security of their territories and revenue for the center, these agents were given sweeping governance powers. They were responsible for maintaining the peace and protecting against invaders, they adjudicated disputes, and they provided the bulk of the public goods enjoyed by their subjects.
Industrialization and modern communications technology made it possible for central rulers to govern geographically remote territories directly. In attempting to seize a greater proportion of governance and public goods provision, central rulers were thwarted at every turn by indirect rulers—those local political and ecclesiastical authorities whose powers they sought to usurp (de Swaan 1988). In their struggles to maintain power, local elites sometimes mobilized nationalist movements based on the cultural distinctiveness of their populations. These movements challenged the political stability as well as the legitimacy of the modernizing state. Prior to direct rule, local leaders demanded, and tended to receive, the loyalty of their subjects. With the growth of direct rule, central rulers demanded an increasingly great share of that loyalty, and to win it they engaged in a variety of state-building activities designed to promote the ideology of the national state (Hobsbawm 1992, Breuilly 1993). State building invariably led to a categorization of populations into two classes: those who adhered to the dominant (and legitimate) culture, and those who did not. As a result, central rulers made strenuous efforts to stamp out peripheral languages, religions, and social mores (Weber 1976). As social identity theory implies, these typically heavy-handed state-building policies—which constitute a blunt form of categorization—tend to produce a reactive nationalism in peripheral areas. State-building efforts often erupt in conflict against the state in order to gain or restore control over a homeland, or to change the balance of rights and resources in a particular region (Esman 1985, Horowitz 1985). Thus, the timing of state-building activities in the Ottoman Empire is closely correlated with the timing of nationalist resistance movements among its culturally distinct subject territories (Hechter 2000).
Labor Markets and Intergroup Stratification
Once it was widely believed that the rise of market society would slowly but surely reduce the political salience of cultural differences. According to this view, cultural differences would cease to matter because market forces would ensure that people from different cultures (and ethnicities) would live, go to school, and work together. Because of these forces, places in educational institutions and jobs in the economy would be awarded on the basis of individual skills rather than cultural (or ascriptive) characteristics (Parsons 1951). Because there was no reason to expect that culture had any implications for one’s skills, in a market society one’s cultural background should have little influence on one’s earning power. In consequence, there should be little residential or occupational segregation. As neighborhoods, schools, and firms became increasingly culturally heterogeneous, cultural intermarriage naturally would follow and the social basis of cultural distinctions would melt away.
Based as it is on neoclassical economic principles, the logic of this diffusionist expectation is crystal clear. However, it is inconsistent with much empirical reality. It is a truism that cultural distinctions have important political implications in nearly every society in the world—not least in the most advanced societies. Intergroup conflict persists because development, and migration, tend to have differential effects on cultural groups (Gellner 1964). One of the principal sources of these effects emanates from labor market processes.
Two different labor market mechanisms have been proposed in the literature. The first suggests that distinct social identities emerge as a by-product of mutual interaction and shared experience. In particular, a market society tends to create a cultural division of labor in which individuals with distinctive cultural markers cluster hierarchically or segmentally. In a hierarchical cultural division of labor, high-status occupations tend to be reserved for individuals of the dominant culture while jobs at the bottom of the stratification system are for others (Hughes 1943, Hughes & Hughes 1952, Bonacich 1972, Hechter 1999, Boswell & Brueggemann 2000). The Indian caste system offers an extreme example. Cultural markers that confine individuals to separate social classes take on high salience. In a segmental cultural division of labor (Hechter 1978), distinctive cultural groups are clustered into highly specialized occupations. In the 1970 census, Greek immigrants to the United States, for example, were disproportionately found in restaurant employment. Dutch immigrants, however, had no comparable occupational specialization. Because of their similar work experience and propinquity, individuals who are concentrated in the same occupations tend to develop distinctive identities. On this account, Greek immigrants had a stronger social identity (as indicated by rates of endogamy, for instance) than their Dutch counterparts.
The second mechanism suggests that distinctive social identities tend to be triggered only when out-group members begin to compete for jobs effectively controlled by insiders. Competition theorists (Deutsch 1953, van den Berghe 1967, Barth 1969, Hannan 1979, Ragin 1979, Nielsen 1980, Olzak 1982, 1992, Banton 1983) argue that economic and political modernization erode the social bases for smaller-scale ethnic identities (such as villages, tribes, or dialects) while encouraging collective action based upon larger-scale ethnic boundaries (such as regional or party lines). According to this view, urbanization, the expansion of industrial and service sectors, the increasing scale and complexity of production organization, the development of peripheral regions, and state building create the potential for ethnonational movements and political parties because these factors initiate contact and competition between culturally distinct populations. At the heart of competition theory lies a biological metaphor: Social boundaries are formed when an out-group invades an in-group’s established ecological niche. For competition theorists, therefore, national and ethnic boundaries are activated when socioeconomic development decreases the barriers between different ethnic populations, spurring intergroup competition over scarce resources.
Cultural division of labor and competition mechanisms have often been framed as alternatives. As a consequence of this framing, a large literature has attempted to assess their relative merits. In his study on the Celtic fringe in Britain, Hechter (1999) provides historical evidence that a cultural division of labor existed in preindustrial Ireland and Wales. Using electoral and census data from British counties, he found that variations in conservative voting patterns from 1885 to 1966 were explained more by cultural variables than by class-related variables, which supports predictions from the cultural division of labor theory over those of the developmental model. Using the same data as Hechter (1999), Leifer (1981) also tested the cultural division of labor and developmental perspectives to further isolate the effect of ethnic ties on mobilization. Using regression analyses, he found that dual subordination (interaction of economic disadvantage and ethnic subordination) and economic disadvantage had no significant effect on voting patterns, which supports neither perspective. However, Leifer’s results do reveal that ethnic subordinate status is significant, which indicates that ethnic ties are important in explaining mobilization, but not in ways anticipated by the cultural division of labor model.
Ragin (1979), Nielsen (1980) analyzed ethnic separatist voting patterns by examining the sources of support for the Welsh nationalist party in Britain and the Flemish nationalist party in Belgium. Both researchers claim that the highest levels of support for the ethnic nationalist party came from economically advanced areas. Employment in the tertiary and advanced sectors increased nationalist support, which suggests that competition within these sectors due to recent incorporation of the peripheral population increased ethnic solidarity. Similarly, Olzak (1982) found that participation in more advanced sectors of the Quebec economy was not associated with ethnic separatism, but that the proportion of bilinguals—which she takes as an indicator of language competition—increased ethnic separatism. Olzak concluded that collective behavior and separatist voting patterns in Quebec supported the ethnic competition perspective. Diez Medrano (1994) tested these mechanisms by analyzing the voting patterns in the Basque country in northern Spain. He found that the concentration of immigrants in blue-collar jobs was positively associated with immigrants voting for Spanish parties. In addition, ethnic segmentation in the labor market was not associated with the voting behaviors of immigrants or natives. He concluded that the cultural division of labor mechanism better explained the ethnic voting patterns in Basque country. Data from Mettam & Williams (1998) on employment patterns in Estonia indicate that at the end of the Soviet era, the segmental cultural divisions of labor that operated in Estonia contributed to Estonian nationalist identity. Finally, cultural division of labor has been used to account for political conflicts between Israeli Jews (Peled 1990, 1998).
All told, both mechanisms are likely to come into play. Each seems to account for nationalist behavior in different countries during different periods of time. For example, Hechter (1999) shows that during one time period, the cultural division of labor is most useful in explaining nationalism in Britain’s Celtic fringe, whereas Ragin (1979) suggests that during a subsequent period, the competition model is more effective. Similarly, Nielsen (1980) found that the rise of the Flemish movement, prior to World War II, supports the cultural division of labor analysis, but after 1945 ethnic competition seems to dominate.
Although these two mechanisms have often been treated as mutually incompatible, this conclusion is probably hasty. Nielsen (1985:147) argues that “…it is even possible that one [mechanism] is more correct in the case of one country as compared with another, or even for one country in the case of one historical period and not another, depending on which trends have major causal effects in the situation.” Moreover, the two mechanisms may be responsible for explaining related but distinct phenomena (Hechter 1994). It may be that the cultural division of labor explains how and why ethnic and racial identities become salient relative to other social identities, whereas the competition mechanism explains how ethnic collective action is triggered once these salient identities are formed.
Institutional Factors in the Formation of Group Boundaries
Recent scholarship has emphasized a number of institutional factors that lead to the construction of group boundaries. Nagel (1994) claims that political policies and institutions can affect the strength and even the existence of ethnic group boundaries. For example, immigration policies can influence the composition, location, and class position of immigrants, which ultimately affects their process of group formation and assimilation in the host country (Horowitz 1985, Pedraza-Bailey 1985, Light & Bonacich 1988, Espiritu 1996, Reitz 1998). Ethnically linked resource policies can also activate ethnic boundaries. For example, in 1960, the division of the Nigerian state into three regions resulted in the formation of three regionally based, ethnically linked political parties (Nagel 1986). The competition between these political parties led not only to a heightening of ethnic boundaries but also to the eventual disintegration of the Nigerian political system and the secession of the Eastern Region (Biafra). Additionally, official categories, such as census racial classifications, can create a sense of group membership and help to constitute groups (Starr 1978, Brubaker 1996). Political recognition of a particular ethnic group, whether positive or negative, can raise the group’s self-awareness and encourage organization. For example, ethnic and cultural identities are often heightened by “racial profiling,” purposeful action taken by the state to associate certain crimes and characteristics with an ethnic or cultural group. Political recognition of an ethnic group can also increase identification and mobilization between officially unrecognized groups facing the prospect of exclusion from an ethnically defined political arena (Nagel 1986). Darroch (1981) argues that a variety of Canadian federal programs implemented in the 1980s, including those recommended by the Royal Commission on Bilingualism and Biculturalism, may have encouraged Quebeckers to engage in a more vocal separatist movement. In Nigeria, the ethnic boundary that designated Yoruba peoples from non-Yoruba peoples expanded only after colonization and contact with groups who did not share the same descent myth and language (Nnoli 1989). Similarly, a politicized pan-Malay ethnicity began to emerge only after British colonization (Young 1976, Nagata 1981, Ongkili 1985).
Determinants of Assimilation in Countries of Immigration
Assimilation is a process in which the boundaries of culturally distinct groups gradually become attenuated. Assimilation has been intensively studied in countries of immigration, particularly in the United States (Park 1950, Gordon 1964, Hirschman 1983, Yinger 1985). In these countries, the government has actively pursued policies promoting assimilation. Researchers in the United States have attempted to gauge the extent to which certain racial and ethnic groups have assimilated into the host society by using such indicators as educational and occupational attainment, wage inequality, language maintenance, neighborhood segregation, and interethnic and interracial friendships and marriages. Assimilation has been relatively rapid in countries of immigration because the cultural minorities in these countries tend to be spatially dispersed. Even so, research on ethnic and racial groups in the United States helps us to understand some of the conditions that promote assimilation in the multicultural context.
Educational and occupational assimilation clearly is affected by generation and cohort status. Using census data, Lieberson & Waters (1988) observed that American men of European ancestry increased their educational and occupational attainment with each new generation. Niedert & Farley (1985) found that the occupational returns for educational attainment improved when comparing first and third generations for European groups, Asians, and Mexicans. In addition, Alba’s (1988) evidence suggests that there is convergence across birth cohorts of European Americans: Rates of college attendance and graduation among European men and women in later cohorts increased to match the rates of their American-born British counterparts. However, several studies have documented the low levels of achievement among certain groups, such as African-Americans, which may increase with each new generation but which do not converge with the core Anglo group with respect to education, income, and occupation (Chiswick 1978, Featherman & Hauser 1978, Niedert & Farley 1985).
The mode of incorporation also influences the subsequent development of immigrant communities (Portes & Rumbaut 1990, Reitz 1998). In their analysis of the educational progress of second generations, Portes & MacLeod (1996) found that the relative advantages or disadvantages associated with Cuban and Vietnamese immigrant communities, due to different modes of incorporation, remained significant even after controlling for family socioeconomic status. In a comparative study of the United States, Canada, and Australia, Reitz (1998) found that social institutions, such as education, labor markets, and social welfare, shaped the extent to which immigrants will assimilate in terms of socioeconomic status. For example, in the United States, the laissez faire immigration policy, less-regulated labor markets, and weak welfare state combine to produce lower entry-level earnings for immigrants, which retards the assimilation process. More generally, immigrants who are granted legal status, the prospect of citizenship, and resettlement assistance by the host community will assimilate more rapidly, both economically and in their social and psychological integration (Light 1984, Bailey & Waldinger 1991, Zhou 1992, Portes & Stepick 1993).
Studies of spatial assimilation (Massey 1985)—the integration of ethnic minorities into neighborhoods where the dominant group resides—indicate that for Asians and Latinos, the most powerful determinant of the racial and ethnic composition of their neighborhoods is their own socioeconomic status. The more income and education one has, the larger the percentage of Caucasians in the neighborhood (Massey & Denton 1987, Alba & Logan 1993, Logan et al 1996, Alba et al 1997). For African-Americans, this relationship does not hold. Even when African-Americans do move to the suburbs, they experience relatively high levels of segregation compared with Asians and Latinos (Massey & Denton 1988). Overall, African-Americans still face a spatial assimilation process hampered by racial stereotypes (Massey & Denton 1988, 1993).
For Latinos but not Asians, linguistic assimilation is another predictor of moving into suburban neighborhoods largely populated by Caucasians (Alba & Logan 1991, Alba et al 1997). Alba and his colleagues (1999) compared the determinants of suburbanization in 1980 and 1990. In the 1990s, the effect of linguistic assimilation in addition to recent immigrant status for Asians, Afro-Caribbeans, and Cubans was much weaker. Such findings indicate that the achievement of spatial assimilation is not necessarily associated with other forms of assimiliation.
More than any other intergroup indicator, intermarriage represents the final outcome of assimilation (Gordon 1964). Intermarriage rates are also affected by generation: With each new generation of Asian-Americans (Montero 1980, Wong 1989, Lee & Yamanaka 1990, Hwang et al 1994) and Latinos (Murguia & Frisbie 1977, Schoen et al 1978, Fitzpatrick & Gurak 1979, Gurak & Fitzpatrick 1982), the rates increase. In addition, research indicates that members of ethnic or racial minority groups with higher levels of education tend to marry outside their groups more often than do their less-educated counterparts (Sandefur & McKinnell 1986, Lieberson & Waters 1988, Schoen & Wooldredge 1989, Kalmijn 1993).
Intermarriage is also affected by the relative size and distribution of racial and ethnic populations. Several studies indicate that group size is negatively related to intermarriage rates because members of small populations have less chance of meeting potential spouses from within their group than from other groups (Blau et al 1982, 1984, Blau & Schwartz 1984, Alba & Golden 1986, Schoen 1986, Gilbertson et al 1996). The geographic concentration of an ethnic population also affects intermarriage. For example, Asian-Americans living in California, where they are concentrated, tend to marry outside their group less often than do those living elsewhere in the United States (Wong 1989). Similarly, the rate of endogamy for African-Americans is higher in states where the proportion of African-Americans in the population is higher (Kalmijn 1993). Members of geographically concentrated ethnic groups have a greater probability of meeting potential spouses from within their own groups. In addition, community sanctions against intermarriage are likely to be greater in geographically concentrated groups (Spickard 1989). By the same token, spatial dispersion favors intergroup contact and sociability, and this increases intermarriage rates. There is some empirical support for a positive relationship between heterogeneity and intergroup association (Blau et al 1982, 1984). Although some recent studies produced inconsistent results (South & Messner 1986, Anderson & Saenz 1994, Hwang et al 1994), these may be due to differences in the measurement of group heterogeneity. Other studies found that income inequality, opportunity for contact, status diversity within the ethnic group, and ethnic language maintenance are also significant predictors of intermarriage (South & Messner 1986, Anderson & Saenz 1994). All told, the empirical literature suggests that the size and distribution of the group shape the rate of intermarriage, the final stage of assimilation.
Determinants of Assimilation Elsewhere
There is less research on assimilation in countries with spatially concentrated minority groups. Several studies show that high levels of education and income, residence in an urban area, and youthfulness promote identification with the dominant culture. In his study of the determinants of nationalistic identification with Spain, Herranz de Rafael (1998) found that those who reside in urban areas and perceive themselves to be in higher class positions were more likely to identify with the nation-state. Other variables, such as religious participation, blue-collar occupation, white-collar occupation, political identification, and education, had significant effects, but not consistently. Similarly, Bollen & Diez Medrano (1998), using a national representative survey, found that a high level of education was the most important individual-level variable for explaining attachment to Spain, as measured by feelings of morale and sense of belonging. In the context of the former Yugoslavia, using survey data from Croatia, Bosnia, and Serbia, Sekulic et al (1994) examined the factors that influenced people to identify themselves as members of a multinational state rather than as members of a specific nationality. Their statistical analyses indicate that urban residence, political participation in the Communist Party, having parents of different nationalities, being young, and being Croat led to Yugoslav identification. Similarly, Laitin’s (1998) study of Russians in the newly independent republics of the former Soviet Union indicates that education and the number of years a respondent has resided in the republic are both positively and significantly related to assimilation. Thus, the research on the formation of an assimilated national identity indicates that other forms of assimilation—education, socioeconomic status—tend to lead to an assimilated identity within a multinational state.
Laitin (1998) found that the most significant predictor of linguistic and cultural assimilation among Russian-speaking peoples in Kazakhstan, Estonia, Latvia, and the Ukraine was the size of the titular population—nationality groups for which the republics were named—in the respondents’ city of residence. Russians in these republics are likely to assimilate when they reside in cities where the percentage of the titular population is high. However, Laitin also found strong interrepublic differences in assimilation. He claims that the different modes of incorporation of elites from peripheral republics resulted in different incentives for assimilation and, therefore, varied outcomes for each republic. To complement the macrostructural elite-incorporation model, he also used a microlevel tipping model to explain the mechanisms behind assimilation. Russians assimilated faster in the Baltic states than in Ukraine or Kazakhstan because of high expected economic returns for linguistic assimilation and acceptance of language by in- and out-group members.
Different forms of assimilation—socioeconomic, spatial, marital, identity—are facilitated by common individual-level factors, such as generation, age, socioeconomic status, and urban residence. However, new studies indicate that some types of assimilation are not necessarily associated with others for all immigrant groups, which suggests that assimilation may not be a unitary process. The evidence reviewed here suggests that structural factors, such as the mode of incorporation and the size and concentration of the ethnic group, contribute to whether ethnic group members integrate into the host society and, if so, how fast. Most of the literature focuses on outcomes; too often, discussions and tests of the causal mechanisms responsible for assimilation are missing (Laitin’s research is exceptional in this regard).
CONDITIONS PROMOTING THE COLLECTIVE-ACTION POTENTIAL OF MINORITY GROUPS
The mere existence of distinctive social identities among minorities has no necessary consequences for political mobilization. Nationalist collective action is unlikely to occur in the absence of (a) preexisting social groups formed to provide members with insurance, welfare, and other kinds of private goods, (b) a widespread demand for autonomy or outright independence, and (c) the opportunity to act collectively on behalf of one’s ethnic group.
The Capacity to Overcome Free Riding
Scholars have long recognized the importance of mutual dependence as a condition of group formation (Hechter 1987, Rabbie & Horowitz 1988, Belanger & Pinard 1991). Individual group members will be dependent on the group to the degree that they do not have other alternatives by which to pursue their goals. Several other factors, such as altruism, proximity, shared territory, similar preferences, and shared labels (Rabbie & Wilkens 1971, Chai 1996), all contribute to lowering the costs of organizing a group and encourage the formation of minority groups. But even where shared interests and dependence exist, collective action will not occur if group members free ride on the cooperative behavior of other members. Coercing participation or offering selective rewards to those who participate in collective action were originally identified as the only solutions to the free-rider problem (Olson 1965). Because selective incentive solutions are themselves subject to the second-order free-rider problem (Frolich & Oppenheimer 1970), the popularity of this solution has waned in recent years. Theorists recognize that other kinds of incentives and case-specific conditions are likely to be important in promoting mobilization (Lichbach 1995). These include the group’s monitoring capacity (Hechter 1987), the magnitude and character of its social rewards (Chong 1991), members’ fear of incurring social penalties (Hardin 1995), conformity to internalized norms of political action (Muller et al 1991), and self-conscious cooperation to promote collective rationality [for documentation of the increasing influence of strategic views of collective action, see Hardin (1982), Moore (1995)].
These conditions are likely to flourish in groups once organized to provide their members with private goods (Hechter 2000). For example, the origins of many Western European nationalist movements lay in groups established by interested parties, such as teachers and clergy, to promote or protect peripheral cultures and religions in which they had a significant personal stake (Hroch 1985). Likewise, the American civil rights movement depended heavily on African-American churches to overcome free riding (Morris 1984:807). Two studies employing the Minorities at Risk data set suggest that a group’s prior mobilization for political action has a strong effect on its current political behavior (Gurr 1993, Olzak & Tsutsui 1998). A handful of studies find that a group’s organizational capacity for mobilization also leads to communal protest and rebellion (Gurr 1993, Lindstrom & Moore 1995, Gurr & Moore 1997). Additionally, the historical loss of group autonomy, grievances, and group coherence have a positive effect on the organizational capacity for rebellion, whereas the loss of group autonomy and institutionalization of democracy increase the ability of an ethnic group to get its members to engage in, or at least condone, protest behavior (Gurr 1993, Gurr & Moore 1997).
If, as these findings suggest, support for nationalist movements is primarily instrumental, then an important conclusion follows. Nationalist dynamics ought to be related to the central state’s actual treatment of minorities. The more responsive the state is to the political demands of its minorities, the less support there should be for antistate—and especially violent—forms of nationalism. This instrumental view, however, implies nothing about the content of minority political demands. There is no reason to believe that the minority’s direct material interests (e.g. in jobs or economic growth) must trump its cultural interests (e.g. in a given language or religion). The ultimate rationale of any state is the provision of public goods. The key political question is this: To what extent does the state provide an optimal bundle of the public goods demanded by the minority? For example, if Corsicans want their children to be able to speak the vernacular in school, will Paris accommodate their demand? To the degree that Paris can accommodate the Corsicans, support for Corsican separatism ought to decline.
In turn, the state’s response to challenging political activity is important in shaping the extent to which an outbreak of violence will occur. If nationalist political parties are not recognized by the state, and if their activities are repressed, violent collective action is likely to occur. Low levels of repression tend to increase political protest and violence, but high levels of repression decrease the likelihood of these outcomes (Muller 1985, Muller & Seligson 1987, London & Robinson 1989, Boswell & Dixon 1990, Muller & Weede 1990, Gurr & Harff 1994). A semirepressive regime, which may exact sanctions and punishment at a middle level, tends to increase political violence. More studies are needed to better understand the dynamic interactions between states and political dissidents. Current research indicates that dissidents respond to state repression by substituting nonviolent behavior for violent behavior and vice versa (Lichbach 1987, Moore 1998). An important advance in the literature is the finding that there are short- and long-term effects of repression on actors (Opp & Ruehl 1990, Rasler 1996).
Although the instrumental logic that undergirds much of this literature may seem to be unassailable, its implications are far from clear. In fact, there is little consensus about the kinds of political institutions that are most likely to contain nationalism.
Institutional Factors Affecting the Demand for Minority Collective Action
Electoral arrangements have long been implicated in the rise of nationalism. As John Stuart Mill recognized, majoritarian democratic regimes with single-member districts may result in a tyranny of the majority. For this reason, a variety of alternative institutions have been proposed—from proportional representation (Grofman & Lijphart 1986, Lijphart & Aitkin 1994, Sartori 1997), to consociationalism (Lijphart 1984), to various forms of federation (Riker 1964).
Whereas vast literatures are devoted to each of these institutions, it is the role federation plays in nationalism that has garnered the lion’s share of recent attention. There are three different views of the matter. One is that federation undercuts nationalist conflict by providing minorities with a more optimal mix of government-provided goods (Brass 1991, Gurr 2000, Hechter 2000). A second view is that federation exacerbates nationalism by providing minority elites with resources (principally patronage and tax revenues) they can use for mobilizing nationalist movements against central authorities (Roeder 1991, Bunce 1999). The third view is that federation has no determinate effects on nationalism because it is subject to a host of noninstitutional contingencies, such as “reputational cascades” (Kuran 1998). Because these rival arguments are based on the same instrumental premises, their merits can only be assessed against the empirical record.
The empirical record, however, is remarkably murky. In fact, the literature lends credence to all three propositions. Although propositions of this kind are best tested with quantitative data, these data tend to be in short supply. Even so, it is possible to make one interesting generalization. Scholars who have attempted quantitative cross-national analyses tend to support the first view (Gurr 2000, Hechter 2000). By contrast, students of the former Soviet empire and its former dependencies (Roeder 1991, Brubaker 1996, Treisman 1997, Laitin 1998, Bunce 1999, Hale 2000, Snyder 2000) seem convinced of the second. Most historians, we suspect, are convinced of the third proposition.
One possible interpretation of the conflicting evidence comes to mind. This interpretation rests on Riker’s (1964) view of federation as an exchange relation, or bargain, between relevant agents in the center and in the periphery. The utility of any such bargain depends on the resources each party brings to the relationship. If the center in a federation loses resources—or is perceived as having lost them because of failures in war and an inability to maintain social order—peripheral leaders are likely to be emboldened to strike out on their own (Hechter 2000). Because federation grants them more resources than their counterparts in centralized regimes, nationalism is likely to arise. Moreover, according to this logic, nationalism should arise first in those parts of a federation that have the greatest resources (Hale 2000). In centralized regimes, however, nationalist or secessionist movements may be more likely to break out in less developed, rather than more developed, regions (Horowitz 1985, Hechter 1992). Resolving this issue is an important task for future research. Studies aiming to disentangle the relationship between regime structure and nationalist movements must be not be limited to one particular political context but should be broadly cross-national.
Political Opportunity Structures
The most expansive view of nationalist movements suggests that they depend on the structure of political opportunities (Jenkins & Perrow 1977, Tilly 1978, McAdam 1982, Tarrow 1983, McAdam et al 1996). The concept of political opportunity encompasses far more than regime structure. McAdam (1996) specifies four dimensions of political opportunity: the relative openness or closure of the political system (presumably, federation would figure here), the stability of elite alignments, the presence or absence of elite allies, and the state’s capacity for repression.
Political mobilization of all kinds is facilitated when the central state is perceived to be in crisis. When economic and social institutions collapsed in Iran, Algeria, Kazakhstan, Uzbekistan, and Kyrgyztan, Islamic movements offered goods that were previously provided by the state (Ollapally & Cooley 1996). In each case, the collapse of the central state encouraged challenging movements. This account dovetails neatly with our discussion of the importance of prior organization for collective action. The increasingly perceptible weakness of the Soviet state, which could not manage to defeat an Afghan insurgency, led to its fragmentation (Hechter 2000) and to a welter of nationalist movements in the titular republics. Students of revolution likewise point to the importance of national crises that render the powerful vulnerable (Skocpol 1979, Arjomand 1988, Goldstone 1991). In his study of 227 communal groups—“cultural groups that do not have recognized states or institutionalized political status”—in 90 countries, Gurr (1993:160) found that democratization provided opportunities that spurred the mobilization of these groups. Gurr’s results show that expansion of state power during the 1960s and 1970s intensified communal rebellion in the 1980s but diminished protest. These effects were especially pronounced in the Third World among ethnonationalists and indigenous peoples, groups who faced the greatest losses of resources and autonomy due to the expansion of direct rule.
Despite its intuitive appeal, the concept of political opportunity is often criticized for its lack of specificity. For example, Gamson & Meyer (1996:275) state that political opportunity is “in danger of becoming a sponge that soaks up virtually every aspect of the social movement environment…it threatens to become an all-encompassing fudge factor for all the conditions and circumstances that form the context for collective action.” Political opportunity may be heuristically useful for interpreting the emergence of social movements and political activities, but measurement and specification of the concept need to be improved (see Kriesi et al 1995, Soule et al 1999).
The International Context
To this point in the review we have concentrated on causes of minority collective action that are endogenous to existing state borders. Yet it is evident that the international context has important effects on political outcomes in states. In their study of ethnic mobilization at the world system level, Olzak & Tsutsui (1998) found that during the 1970s, a state’s membership in international organizations decreased ethnic violence, whereas during the 1980s, ethnic diversity and membership in international organizations increased ethnic protest. The authors argue that these findings support the claim that the diffusion of global norms about human rights via ties to international organizations constrains violent ethnic activity while encouraging ethnic mobilization and nationalism in dependent, peripheral states.
But how does the political behavior of groups in other countries affect political mobilization and conflict in general? Kuran (1998) claims that the international context affects the process of “ethnic dissimilation” via three distinct mechanisms that work together: (a) a demonstration effect, where the ethnic discrimination that is a dominant part of the political discourse in one country heightens ethnic categories among the citizens of another country; (b) an expectation effect, where news about the rising ethnic activity in another country alerts people to the possibility that dissimilation can occur in their own countries; and (c) a reputation effect, where global norms affect whether dissimilation is a project to be undertaken. However, most studies, including Kuran’s (1998), do not actually test causal mechanisms but instead focus on outcomes. Lindstrom & Moore (1995) found that protest and rebellion in neighboring countries influenced mobilization and rebellion by ethnic groups (contagion) but did not find that ethnic groups were influenced by kinship groups in other countries (diffusion). Research suggests that links to kinship in other countries affect conflict occurring between states. Davis et al (1997), Moore & Davis (1998) examined the effect of transnational ethnic ties on the conflict and cooperation behavior of states in the international system and found that the level of conflict between two states was higher if both states contain members from the same ethnic group and if ethnic group members are politically privileged in one state but not in the other state.
There is also empirical support for the diffusion of tactics across group and state boundaries. In a study examining patterns of peaceful protest in 17 Western industrialized nations (1950–1982), Hill and his colleagues (1998) found that after 1960, when a state had substantial ethnic divisions and widely available access to television, and when the people had seen nonviolent tactics used effectively in the United States, a greater number of peaceful protests occurred. The authors suggest that diffusion of tactics occurs when groups are experiencing the same political conditions. Halperin (1998) argues that prior to 1945, social, economic, and political conditions in Europe did diffuse ethnic conflicts to other states, but that because after World War II minorities and lower classes were integrated into the political process, circumstances today do not encourage such trends.
However, other researchers locate the sources of ethnic conflicts primarily within states and question whether the effects of demonstration and contagion facilitate the outbreak of violence and secession. Saideman (1998) argues that after the fall of communism, the Yugoslav, Czech, and Soviet federations faced similar problems; these nation-states had to rebuild politics and maintain power. Such changing conditions led to insecurity on the part of minorities in the region. Fearon (1998) calls this the commitment problem, where ethnic minorities find themselves without a third party to guarantee that majority leaders will not exploit them in the new state. In this circumstance, secession is seen as a superior alternative. Both Saideman and Fearon criticize the logic of contagion via demonstration effects; proponents of demonstration effects assume that followers will engage in further political action when there is the distinct possibility that secessionist movements or political conflict might actually discourage such behaviors in other nation states. Saideman (1998) points out that the secessionist efforts of Croatia and Bosnia were costly in terms of lives lost and damage done—a situation unlikely to inspire similar activities elsewhere.
Overall, the empirical evidence indicates that links to kinship in other countries do not necessarily affect protest or rebellion activity by ethnic groups unless both groups face similar political circumstances, but that such links affect the level of conflict or cooperation between two neighboring states. There is some evidence that political action in neighboring countries positively affects mobilization and rebellion. However, there is more evidence that political and social circumstances (i.e. disruptive social change, such as the fall of communism) affect whether minorities will engage in secessionist movements. In addition, links to international organizations tend to increase the level of mobilization in the name of sovereignty. Once again, there is a need for identifying and testing the mechanisms behind processes of diffusion and contagion in an international context as the world continues to globalize.
Given the variety of minority groups and the different social and political circumstances they face, constructing an overarching theory of minority group collective action is a daunting challenge. As this review has attempted to show, there is much to gain by focusing on causal mechanisms responsible for the political mobilization of minority groups. To enrich our knowledge about these processes and mechanisms, however, it is necessary to explicitly define the “nationalist movements” so often taken to be hallmarks of minority groups’ political mobilization. In most of the accounts of nationalist movements in the literature, there is an alarming lack of specificity about what forms this kind of behavior entails. Such accounts could be describing the formation of ethnic political parties or nationalist organizations, protest or rebellion, military activities carried out by nationalist organizations, or secessionist votes. Of course, these categories can be disaggregated ad infinitum. Progress will be slow until some consensus on the types of nationalist movements found in the historical record can be realized. Several scholars have attempted to construct more effective categories of ethnic conflict (Heraclides 1990, Chazan 1991, Carment 1993) and of types of participants in political protest and rebellion (Horowitz 1985, Gurr 1993). We urge others to follow their lead. The most important impediment in this field remains the paucity of historical and cross-national databases adequate to test the many theories that have been advanced to account for the fate of minority groups in an increasingly interdependent world.