Multicultural Citizenship

Christian Joppke. Handbook of Citizenship Studies. Editor: Engin F Isin & Bryan S Turner. Sage Publication. 2002.

On the list of hyphenated citizenships the ‘multicultural’ one certainly takes a prominent place. It signals a general concern for reconciling the universalism of rights and membership in liberal nation-states with the challenge of ethnic diversity and other ascriptive ‘identity’ claims. In this chapter, I first discuss multicultural citizenship in theory, pointing to the shortcomings of both ‘radical’ and ‘liberal’ approaches to justifying minority rights under the generic, and paradoxical, notion of multicultural citizenship. Secondly, I look at the ways in which multicultural citizenship has been practiced in liberal nation-states. This investigation reveals a gap between the theory and the practice of multicultural citizenship: a mechanism to accommodate ethnic, national, and other minorities in theory, multicultural citizenship in practice has been a variant of nation-building in a few new settler societies without independent founding myths. In addition, I argue that the state-centered notion of multicultural citizenship deflects from the decentered accommodation of multicultural minority claims in functionally differentiated societies, which remains short of official state recognition. Considering the difficulties of grounding multicultural citizenship in theory, and considering the diverse and often contested practices that it seeks to encompass, I conclude that the notion of multicultural citizenship is too vague and multifaceted to be a useful tool of sociological analysis. It also runs counter to a trend toward de-ethnicization in liberal states, in which the cultural impositions of the majority on minority groups are growing thin, thus removing the case for minority rights.

Multicultural Citizenship in Theory

A good way of approaching multicultural citizenship is to explicate its critiqued or modified opposite. One influential author has identified the latter as ‘universal citizenship’ (Young, 1989). Universal citizenship refers to the equality of rights and status that have come to define membership in liberal-democratic nation-states: ‘citizenship for everyone, and everyone the same qua citizen’ (Young, 1989: 250). Universal citizenship, product of the French and American revolutions, has its own historical opposite: the tiered and multiple subject statuses and the particularistic rights and duties attached to them under feudalism. In fact, ‘universal citizenship’ is a pleonasm, because universalism as the rejection of particularistic rights and statuses is constitutive of citizenship as such. Accordingly, universal citizenship is meaningful only if viewed as the opposite of multicultural citizenship. The latter is thus paradoxical, because it seeks to (re)particularize a form of membership that is inherently universalistic.

From a different point of view, the notion of universal citizenship is not a pleonasm, but an oxymoron. As long as there is no world state, citizenship means membership of a particular state. Citizenship as state membership is ‘inherently group-differentiated’ (Kymlicka, 1995: 124), and thus the exact opposite of ‘universalist.’ This ambivalence of ‘universal citizenship,’ to be either pleonasm or oxymoron, reflects the dual nature of citizenship as both ‘internally inclusive’ and ‘externally exclusive’ (Brubaker, 1992: Ch. 1). The distinction between internal and external aspects of citizenship points to an important limitation of the meaning of multicultural citizenship: to the degree that the latter is a commentary on the shortcomings of universal citizenship, it focuses only on the internal rights dimension, and takes no account of the external state-membership dimension of citizenship.

The abstraction from the external dimension of citizenship in most discussions of multicultural citizenship is puzzling. After all, the multiplication of cultures and ethnic groups within contemporary nation-states is largely the result of cross-border migrations, which is partially shaped by (and, in turn, impacts on) a state’s formal citizenship laws and policies. In multicultural citizenship’s most concise formulation (Kymlicka, 1995), special group rights (for immigrants, for example) are compensation for axiomatically assumed strong nationalizing practices even in contemporary liberal states, which are said to have a discretionary ‘right… to determine who has citizenship’ (Kymlicka, 1995: 124). As Kymlicka provocatively argues, this right of states ‘rests on the same principles which justify group-differentiated citizenship within states,’ and ‘accepting the former leads logically to the latter’ (1995:).

Following the same logic, a relaxing of the axiomatically assumed state discretion on citizenship would remove the main justification for group rights. In light of a liberalization of citizenship law across immigrant-receiving Western states (see Weil, 2001), which is part of a larger trend toward de-ethnicization in such states (see Joppke and Morawska, forthcoming: Ch. 1), there is evidence that this is actually happening. However, by focusing only on the internal rights and ignoring as an invariable parameter the external state-membership dimension of citizenship, Kymlicka has ruled out by design the possibility of changes in the external aspect of citizenship impacting on its internal rights dimension. In sum, a proper assessment of multicultural citizenship has to take account of both the internal and the external dimensions of citizenship.

Before elaborating on this, it is important to note that current versions of multicultural citizenship differ in their relationship to universal citizenship. For feminist and (post)Marxist radicals, the relationship is one of critique and substitution (Young, 1989, 1990); for liberals, it is one of complementarity and linear addition (Kymlicka, 1995; Carens, 2000). The thrust of Iris Marion Young’s radical formulation is to denounce the ‘universal’ in universal citizenship as the disguised particularism of the dominant group(s). ‘Oppression’ is key to her scenario: society is seen as composed of ‘social groups,’ which are either dominant or oppressed. Not much is said about the dominant group(s) (is it one or several?), despite occasional reference to ‘white middle-class men’ (Young, 1989: 268). This omission is perhaps not accidental, because the dominant can hide their groupness under the cloth of universalism. ‘Differentiated citizenship,’ which for Young is mostly about special representation rights in the polity, is reserved for ‘oppressed’ groups, whereby oppression is defined rather broadly as including anything from economic exploitation to cultural discrimination. From this broad definition of oppression follows a long list of groups entitled to differentiated citizenship: ‘Women, blacks, Native Americans, Chicanos, Puerto Ricans and other Spanish-speaking Americans, Asian Americans, gay men, lesbians, working-class people, poor people, old people, and mentally and physically disabled people’ (Young, 1989: 261). Without any commentary, a second list adds ‘young people,’ while dropping ‘Asian Americans’ (1989: 265). The underlying reasoning is apparently ad hoc; and ‘differentiated citizenship’ for what turns out to be the vast majority of the US population seems a rather impracticable idea.

Young’s failure to come up with a more concise definition and elaboration of what constitutes an ‘oppressed group’ is instructive. It shows the difficulty of building a theory of multicultural citizenship around the notion of ‘oppression.’ This notion is too vague and simplistic to account for the asymmetries of power and resources in complex societies. As an inherently polemical (or ‘critical’) concept, oppression thrives on its (utopian) opposite, the absence of oppression. Has it ever existed? Can it exist at all, particularly if group differentiation is not only an inevitable but a ‘desirable process’ in modern societies (Young, 1989: 261)? Why should oppression stop when The Others are in charge? Finally, there is a systematic ambivalence about the inclusive or exclusive thrust of differentiated citizenship, and thus about its relationship to its critiqued opposite, ‘universal citizenship.’ If ‘universal’ is just a smokescreen for dominant group interests, the purpose cannot be inclusion into this false universal (as it had been in the—negatively evaluated—‘emancipatory momentum of modern political life’ 1989: 250). Accordingly, the quest for differentiated citizenship is presented as a ‘politics of difference’ that rejects traditional ‘inclusion’ (Young, 1990: Ch. 6). However, in other places the whole point of differentiated citizenship is still seen as ‘mak[ing] participation and inclusion possible’ (Young, 1989: 273). Inclusion into what, one is inclined to ask, if existing institutions and representations (such as ‘universal citizenship’) are just instruments of dominant groups.

The notion of oppression does not figure centrally, in fact, hardly appears at all, in Kymlicka’s (1995) liberal alternative of multicultural citizenship. Furthermore, for Kymlicka the relationship between universal and multicultural citizenship is not one of critique and substitution, but of simple addition. Rather than being subjected to a radical critique, universal rights are fine; the problem is that they are not enough for certain groups: ‘A comprehensive theory of justice in a multicultural state will include both universal rights, assigned to individuals regardless of group membership, and certain group-differentiated rights or “special status” for minority cultures’ (Kymlicka, 1995: 6).

Whereas Young’s key concept was oppression, Kymlicka’s is ‘societal culture.’ Individuals need societal culture as a context of meaningful choices: without it there is no freedom (Kymlicka, 1995: Ch. 5). In addition, access to a societal culture can become an issue of equality and justice under certain circumstances (1995: Ch. 6). No state, not even liberal states, can be culturally neutral; for example, in its selection of an official language a state inevitably promotes the majority culture, at the cost of the culture of minority groups that may reside in the same territory. Given the inevitable nexus between state and majority culture, the traditional liberal answer to ethnic and cultural difference, ‘benign neglect,’ is not enough: liberal justice requires special rights recognizing and protecting the cultures of minority groups.

Kymlicka’s distinctive contribution has been the liberal mainstreaming of minority rights. After Kymlicka, the earlier confrontation between liberal defenders of universal citizenship and radical proponents of multiculturalism and group rights has lost its basis: it is not a radical critique of existing institutions, but those liberal principles on which existing institutions are built that require multicultural citizenship.

It is therefore worth scrutinizing this theory in more detail. A crucial difference to Young is the drastic narrowing of the minority groups entitled to special rights: only ethnic and national minority groups qualify. This is due to a narrow definition of the conditioning factor of group rights, ‘societal culture.’ Kymlicka defines societal culture as shared history, language, and territory, making it ‘synonymous with “a nation” or “a people”—that is,… an intergenerational community, more or less institutionally complete, occupying a given territory or homeland, sharing a distinct language and history’ (Kymlicka, 1995: 18). This definition excludes non-ethnic groups, such as gays and lesbians, the disabled, or lifestyle groups, as multicultural claimants (see Kymlicka, 1998: Ch. 6).

However, the claims of the only two legitimate multicultural groups in Kymlicka’s liberal scenario, national minorities and immigrants, differ in significant ways. And, as I would like to add critically, they differ in ways that ultimately militate against the very notion of multicultural citizenship. National minorities, such as the Catalans in Spain, the Quebecois in Canada, or the Aborigines in Australia, have the strongest claims within Kymlicka’s scheme. All of them have ‘institutionally complete’ cultures, that is, cultures that cover the full range of human activities, needs, and functions (1995: 78). Moreover, as the victims of nation-state building, they are forced to reside in states that do not carry the marks of their culture (most notably, their language). To accommodate the always potentially secessionist national minorities within multinational states, strong ‘self-government rights’ are required, and also justifiable from a liberal point of view. Kymlicka does not hide the fact that these rights pose a serious threat to the integrative function of citizenship, because their thrust is separation, not integration (1995: 188).

By the same token, the nationalist, state-seeking ambition of national minorities is imperfectly captured, even trivialized, by the notion of multicultural citizenship. The very case of Quebec, which partially motivated Kymlicka’s theory, demonstrates this. The Quebecois have always fiercely rejected Canada’s multiculturalist policies, because Canada’s binational founding structure is insufficiently visible in them. In fact, Canadian official multiculturalist policy was introduced just two years after the Official Languages Act of 1969 had made French the second official language of Canada, and it was an obvious attempt to make this concession to the francophone community acceptable to the country’s other minorities, the immigrants and the Aborigines. The Quebecois understood this symmetrizing, levelling function of official multiculturalism, and refused to be considered just a minority among other minorities in a multicultural Canada. It is astonishing that Kymlicka, who is perfectly aware of the stern aspirations of (some) national minorities, and who has argued in particular that only an ‘asymmetrical’ understanding of Canadian federalism could accommodate Quebec (Kymlicka, 1998: Ch. 10), has distorted the asymmetrical, monocultural pretensions of national minorities in the symmetrical and pluralist notion of multicultural citizenship.

The pluralist thrust of multicultural citizenship is more plausible in the case of immigrants, the second legitimate claimant in Kymlicka’s scheme. In fact, all official multiculturalist policies, starting with Canada and Australia in the early 1970s, have prominently (though not exclusively) targeted immigrants. However, immigrants pose their own difficulties for Kymlicka’s scheme. In contrast to strong self-government rights for national minorities, immigrants are due only more moderate ‘polyethnic rights’ examples being exemptions from some general laws that discriminate against minority beliefs and practices, or special benefits (like support for ethnic organizations or mother-tongue instruction and services) that accrue to the majority population automatically. However, qualifying these measures as ‘rights’ is misleading, and it would be more appropriate to call them contingent policies even within Kymlicka’s scheme. Why? The purpose of minority rights is to secure access for minority groups to their own societal culture. However, immigrants, in voluntarily leaving their country of origin, have ‘waived’ the right to their culture (Kymlicka, 1995: 96). Accordingly, the thrust of ‘polyethnic rights’ is integration into the majority culture. Kymlicka’s low-key stance on immigrant rights is healthily realistic: no state would continue admitting immigrants if they arrived with the right to recreate their homelands; and immigrant groups are usually ‘too small and dispersed’ to form viable societal cultures (Kymlicka and Rubio Marin, 1999: 146). However, this realism can be turned against the theory itself. If immigrants have ‘waived’ the right to their societal culture, there is no ground within this theory to endow them with any special ‘right’ at all. To call those immigrant integration policies of states that are more contingent and public order-oriented than rights-based an instance of ‘multicultural citizenship’ seems to be overstated, even misleading.

Most critics of Kymlicka’s theory of multicultural citizenship have zeroed in on its key concept of societal culture (e.g. Benhabib, 1999: 53-6). Joseph Carens rightly detects in its monolithic contours the ‘old logic of the nation-state’ (2000: 66), making it ‘much better suited to a monocultural conception of citizenship than to a multicultural one’ (2000: 65). Most national minorities, particularly the decimated and beaten ones, could never venture on the building of an institutionally complete nation-like culture, from schools to media and hospitals, leaving the basis of their rights claims unclear; and for immigrants ‘it is not clear why (they) are entitled to any special rights to maintain their distinctive cultural commitments’ (2000: 57). There is indeed a tension in Kymlicka’s concept of culture between being either too thick or too thin: ‘too thick’ to give a realistic account of the relationship between liberal states and culture; ‘too thin’ to justify any minority rights at all, particularly for immigrants. Let me develop both lines of criticism in turn.

On the one hand, states are axiomatically presented in Gellnerian terms as strong and tireless nation-builders, as guardians of a thickly conceived majority culture, now as in the nineteenth century, the high point of industrialism and nation-building in the West (Kymlicka, 1995: 76f). This is not a realistic picture for contemporary liberal states. Consider their treatment of immigrants. If one takes the nationalizing practices of states as variable rather than parameter, one sees that in contemporary liberal states there is very little that these states expect of and impose on their newcomers, even at the point of acquiring citizenship. If this is the case, it is not clear why these states should concede minority rights in return for their (very minimal) cultural impositions. Immigrant integration policies are everywhere clothed in multicultural rhetoric, shunning the ‘assimilation’ of immigrants (see Joppke and Morawska, forthcoming). Even in an extreme case of nation-building, like Quebec, the only nationally distinct imposition is the requirement to adopt French language in public life. The other integration requirement in Quebec’s immigrant policy is a dual commitment to democracy and pluralism, which is not specific to Quebec but generic to all liberal democracies (Carens, 2000: 113).

Language, in fact, boils down to the one substantive, and not just procedural, imposition on immigrants. Partially in response to immigration, liberal states have gone a long way toward tilting all (however implicit) ethnic preferencing in their policies and institutions—the shrill Foulard affair in France has been the exception to the generally smooth and noiseless adaptation of European states to the Islamic religion imported by some of their immigrants (see Bauböck, forthcoming). Language is different, because the state has to rely on it in its very functioning—the state can distance itself from religion (and it has actually done so), but not from language. However, language differs from religion in that a person can speak several languages, but can adhere to only one religion. This suggests that the identitarian implications of language use are less than those for religious practice. At least it is not clear why the adoption of another language would deprive a person of a meaningful context of choice.

The relaxing of liberal states’ nationalizing practices is equally visible in the attribution of citizenship. While in international law states have the sovereignty to determine their nationality laws, a creeping rights-logic has rendered this a sovereignty on paper only. This is especially visible in Europe, whose jus sanguinis tradition had at first erected high hurdles to citizenship for immigrants. To better integrate their later-generation labor and postcolonial immigrants, most European states have in the meantime added jus soli elements to their blood-centered nationality laws. With the exception of Luxembourg, Greece, and Austria, all member states of the European Union now provide a right to citizenship to their second-and third-generation immigrants (see the overview in Weil and Hansen, 1999). In addition, most European states have significantly lowered the requirements for naturalization. Germany, for instance, which was until its recent citizenship reform the proverbially ethnic state, in the early 1990s introduced non-discretionary as-of-right naturalization for later-generation immigrants of legal residence and in effect no longer required these citizenship applicants to be culturally assimilated (Joppke, 2000).

These recent changes of immigrant integration policies and nationality laws in liberal states have important implications for multicultural citizenship: if minority rights are compensation for states’ strong nationalizing practices, the weakening of these nationalizing practices removes the case for (this type of) minority rights.

On the other hand, in response to Waldron’s ‘cosmopolitan’ alternative multiculturalism (Waldron, 1992), Kymlicka has admitted to a rather ‘thin’ picture of societal cultures in modernized societies. Citing the case of modern Quebec, Kymlicka finds that all have a place in it, ‘e.g. atheists and Catholics, gays and heterosexuals, urban yuppies and rural farmers, socialists and conservatives, etc’ (Kymlicka, 1995: 87). In fact, to be Quebecois today ‘simply means being a participant in the francophone society of Quebec’ (ibid)—this indicates again the unique position of language in the contemporary liberal state’s cultural impositions. If this is the case, it is not clear why the state’s inherent alignment with this thin and pluralistic culture, which excludes virtually no one, should necessitate compensatory minority rights.

Multicultural Citizenship in Practice

Whatever the difficulties of justifying multicultural citizenship at the theoretical level, does it exist anywhere in the real world? Here it is important to distinguish between explicit multicultural citizenship, in which the latter is an official state program, and implicit multicultural citizenship, in which diversity claims have widely diffused without being written on the forehead of the state concerned. Explicit multicultural citizenship can be found in less than a handful of Western states. The most prominent examples are Canada and Australia, where the very notion of multiculturalism originated in the early 1970s. Canada made a start in 1971 with its policy of ‘multiculturalism within a bilingual framework’ (quoted in Kymlicka, 1998: 55). As indicated above, this multiculturalism is compensation for accommodating the francophone national minority of Quebec, and thus rather separate from the concerns of the latter. Interestingly, its underlying concern is not so much minority recognition as state neutrality, or, in Kymlicka’s terms (1998: 57), ‘(to) separate the… dominance of… common languages… from the historical privileging of the interests or lifestyles of the people descended from the historically dominant groups.’ In line with this de-ethnicizing function, Canadian multiculturalism is an integrative offer for the whole society, not just for minorities. This is explicit in the Canadian Multiculturalism Act of 1988, which ‘acknowledges the freedom of all members of Canadian society to preserve, enhance and share their cultural heritage’ (quoted in Kymlicka, 1998: 185; emphasis added).

The nation-building function of multiculturalism is even more visible in Australia. One of its central documents, the National Agenda for a Multicultural Australia, passed by a Labour government in 1989, stresses that multiculturalism is a ‘policy for managing the consequences of cultural diversity in the interests of the individual and society as a whole’ (quoted in Castles and Davidson, 2000: 166; emphasis added). More than its Canadian precursor, Australian multiculturalism stresses the limits of diversity: ‘Multicultural policies are based on the premise that all Australians should have an overriding and unifying commitment to Australia, to its interests and future first and foremost’ (quoted from the National Agenda, in Castles and Davidson, 2000).

Canadian and Australian multiculturalist policies have gone along with a liberalization of citizenship laws, which had previously been tainted by racial selectivity. The Australian Citizenship Act of 1973, for instance, considerably lowered the residence and language requirements for naturalization, and no longer asks for a ‘transfer of cultural attachments,’ only for a procedural commitment to liberal democratic values (Castles and Davidson, 2000: 168). Castles and Davidson therefore conclude that ‘Australia’s citizenship rules are now multicultural rather than national’ (2000: 169). More correct would be to say that multicultural citizenship in Australia (as well as in Canada) is a distinct way of conceiving of national citizenship.

The few explicitly multiculturalist policies in Europe, notably those of Sweden and the Netherlands, look rather different. They are not identity options for society as a whole, but target immigrants only. In this sense, they are closer to ‘multicultural citizenship’ in Kymlicka’s sense. However, it is also misleading to couch European multiculturalism policies in the language of citizenship, because one of their initial purposes was to protect the status of immigrants qua aliens and not to impose on them the citizenship of the receiving state.

It is worth referring here to the rather curious Swedish experience. A key purpose of Sweden’s multicultural Immigrant and Minority Policy launched in 1975 was to give immigrants the ‘freedom of choice’ between maintaining their ethnic identity or adopting a (obviously ethnically conceived) Swedish identity (see Wieviorka, 1998: 686). This implied that immigrants would not be forced into Swedish citizenship (though citizenship was easier to acquire in Sweden than in most other countries of Europe). Expressions of the upgrading of alien status were the introduction of local and regional voting rights for immigrants in 1975, and—most important—the inclusion of a clause on ethnic and linguistic minority protection in the Swedish constitution in 1976. However, a parliamentary inquiry in the early 1980s correctly noted that in international law only autochthonous minorities that is, long-settled, territorially concentrated minorities with citizenship status—were entitled to minority protection. In addition, the inquiry suggested that the proper meaning of ‘freedom of choice’ could not be the state’s active furthering of ethnic minority identities, but its adopting a ‘neutral position’ and desisting from forced cultural assimilation (Soininen, 1999: 690). The government heeded this advice, renaming its ‘immigrant and minority policy’ ‘immigrant policy.’ This was but a step in Sweden’s gradual withdrawal from its explicitly multiculturalist policy. In the 1990s, escalating unemployment among immigrants moved the attention from cultural to economic issues. The multicultural society was no longer a desirable project for the future but an unavoidable reality that had to be mastered by a centrist rather than difference-oriented state policy. Now the Swedish government even tilted the ‘immigrant’ reference from its immigrant-related policy, calling the latter simply ‘integration policy.’ Its stress is no longer to protect immigrants as ethnic groups, but to enable them as individuals to ‘acquire the Swedish tools which can be needed to manage on one’s own in Swedish society’ (a government statement of 1995, quoted in Soininen, 1999: 692).

The Dutch withdrawal from explicit multiculturalism has been even more extreme (see Entzinger, 1999). The Dutch Minderhedennota (ethnic minorities policy) of 1983 earmarked eight official immigrant minorities for ‘emancipation,’ not within Dutch society, but within state-supported ethnic parallel societies, reminiscent of the ‘pillar’ tradition in this religiously and ideologically divided country (see Lijphart, 1968). This most multicultural of all European immigrant policies soon ran into problems. First, the ethnic diversification of migrant streams in the age of asylum-seeking made it simply impracticable to provide each ethnic group with its own infrastructure, including ethnic schools, media, and social services. Second, the focus on cultural autonomy proved inadequate for the most pressing problem facing immigrants: unemployment and socioeconomic marginalization. In response, much as in Sweden, there has been a reorientation toward ‘open[ing] up the existing institutional arrangements to immigrants, rather than aiming at the development of new, parallel institutions’ (Entzinger, 1999: 10). All references to ethnic minorities, even the very notion of immigrant, were erased from the state’s new ‘integration policy,’ and the new emphasis was on integrating newcomers as individuals rather than as members of groups. The Dutch withdrawal from explicit multiculturalism culminated in the 1998 Law on the Civic Integration of Newcomers, which requires new (non-EU) immigrants to take 600 hours of civics and Dutch language classes.

The demise of official multiculturalist policies in Europe is not the result of a right-wing backlash. As in the Netherlands, impeccable liberals have driven the change. The insight took hold that it was counterproductive to fuel the centrifugal thrust of ethnically diversifying societies with explicitly multiculturalist policies. This implies a fundamental re-evaluation of multiculturalism, which is no longer seen as a normative goal but as an empirical exit-condition of state policy.

In contrast to the precarious nature of explicitly multiculturalist policies, implicit multiculturalism is deeply entrenched throughout immigrant-receiving Western states. It reflects the simple fact that liberal states cannot but sit on top of pluralizing societies. A good example of this is the United States. Unlike the other new settler nations, (Canada and Australia), the United States does not officially consider itself a multicultural society. Nathan Glazer, who captured the pervasive reality of American multiculturalism in the happy notion that ‘we are all multiculturalists now’ (1997), also pointed to the fact that there was ‘nothing multicultural yet’ about US naturalization law, which requires citizenship applicants to swear an oath of allegiance to their new country (Glazer, 1999: 196). Unlike Canada and Australia, which are still today part of the British Commonwealth, the United States has its own founding myth, one that invites ethnic pluralism through its exclusively political content, but also checks such pluralism through its very existence. The recent ‘Americanization’ campaign of the federal government invokes this distinct founding myth of ‘liberty, democracy and equal opportunity’ (US Commission on Immigration Reform, 1997: 26), 5 which is not available in Canada or Australia, and substituted there by their post-British nation-building commitment to multiculturalism.

American multiculturalism rests on the dual pillars of affirmative action and public education (see Joppke, 1999: Ch. 5). Affirmative action is an example of the ‘special representation rights’ identified by Kymlicka (1995: 31-3) as an intermediate, third category of minority rights. Its purpose (though not its reality) is to be temporary only, and to redress discrimination on the basis of race (as well as sex and other ascriptive markers, such as, more recently, physical handicaps). However the state has not carved out official racial categories in order to give them public recognition; rather, racial categorizing is an unintended consequence of antidiscrimination laws and policies, originally color-blind, that were driven towards color-consciousness by concerns of administrative efficiency (see Skrentny, 1996). Accordingly, classifying affirmative action as an instance of multicultural citizenship may appear to be stretching the meaning of the term, if the latter is meant to be recognition and protection of minority cultures. However, affirmative action is a policy with many faces (see Skrentny, 1998). One of them is the quintessentially multicultural idea of mirror representation, which sees society as a composite of ‘groups’ and calls for their proportional representation in key sectors such as the polity, higher education, and the workplace (on the idea of mirror representation, see Phillips 1995).

The main site of American multiculturalism is the ‘world of education’ (Taylor, 1992: 65), where it has appeared as the claim for non-Western-centered public school and college curricula. If minority cultures have found public recognition in the USA, it is mostly through their successful entry into the curriculum—up to a point, as Charles Taylor critically remarks, where the ‘presumption’ of the equal value of the creative expressions of minority cultures is replaced by the ‘peremptory demand for favorable judgments of worth’ (Taylor, 1992: 71). This has been exhaustively discussed under the rubrics of ‘culture wars,’ ‘political correctness,’ etc. It is more interesting for our purposes that this form of multicultural citizenship is thoroughly entrenched yet has remained short of official state recognition also because in the federal US polity the responsibility for public education is multiply divided, horizontally between public and private actors and organizations, and vertically between the federal, state, and substate levels.

The American case points to a central shortcoming of the idea of multicultural citizenship: its fixation on the state. This fixation is perhaps unavoidable because citizenship refers to a relationship between the individual and the state. However, it has obscured the multiple entry points of multicultural claims in the fabric of functionally differentiated societies. To catch the pervasive reality of implicit multiculturalism, we have to change the root image of modern society: not (exclusively) bounded and steered from the top or by a state, but (also) composed of a multiplicity of autonomous subsystems. Among the latter the political system is only one, and not one that could claim to be more central than the others—in Luhmann’s diction, modern societies have ‘neither peak nor center’ (1986: 167-182).

One important sphere in which implicit multiculturalism has quickly taken hold is markets. Because of the ‘Hispanic market,’ Spanish has established itself as the unofficial second language in certain (southwestern and southeastern) parts of the United States automatic bank tellers in California give customers the choice of English or Spanish; large billboards on Los Angeles’ glamorous Wilshire Boulevard advertise their products in Spanish; the leading newspapers in Los Angeles and Miami now publish Spanish-language editions. An advertisement in the business section of the New York Times has the obvious answer to the question, ‘Why Hispanic?’: ‘Because in the next 15 years Hispanic buying power in New York will double to $89.9 billion dollars!’ (quoted in Zolberg and Woon, 1999: 37, fn.74). As Zolberg and Woon (1999: 26) conclude their important observations, there is now a ‘market-driven multiculturalism’ in the USA, ‘quite independently of any public policy choices.’

A second sphere in which implicit multiculturalism has found entry is the legal system. A staple in multicultural-citizenship reasoning is that the individual-rights principle of non-discrimination is not enough to protect minorities. This underestimates the teeth of this legal principle. In Europe, for example, general constitutional provisions on family rights and religious freedoms have been sufficient to exempt a particularly vulnerable group, Muslim girls, from parts of the public-school curriculum that their parents deem incompatible with Islamic norms (see Albers, 1994). In Germany, a landmark decision by the Federal Administrative Court in early 2000 forced the Senate of Berlin to recognize an Islamic sect (Milli Görüs) as an official religious organization, with the right to teach Islam in Berlin’s public schools. Not explicit minority rights, but universal legal principles seem to be sufficient to put Islam on the path to establishing itself as the fourth official religion in Germany.

Sometimes multicultural recognition claims are not only indirectly satisfied by the law’s individual non-discrimination principle, but have come to shape legal strategies and principles directly. An example of the latter is the recent ‘cultural defense’ strategy in American criminal law (see Coleman, 1996). It builds on a longstanding trend in American criminal law toward ‘individualized justice.’ Departing from the previous principle that every offense in a like legal category calls for identical punishment, individualized justice takes into account mitigating circumstances and the individual character and propensities of the offender in the assessment of guilt and punishment. Cultural defense injects the defendant’s culture as one such mitigating and individualizing circumstance into the criminal process. It argues that someone raised in a foreign culture should not be held fully accountable for conduct that violates domestic law if that conduct would be acceptable under his or her native law. Successfully invoked by immigrant defendants in California, New York, Georgia and Minnesota, the cultural defense strategy has led some courts to reinterprete rape among Hmong refugees as part of their traditional courtship customs; to consider wife-beating and killing among Chinese as conditioned by ‘traditional Chinese values about adultery and loss of manhood’ (quoted in Coleman, 1996: 1109); and to exonerate from manslaughter charges a Japanese mother who had drowned her three children because in Japanese culture mother-child suicide is an accepted method for betrayed wives to escape shame. As Coleman points out, there is a tension between cultural defense and individualized justice, because in the former the question of moral culpability is not answered by resort to the particular individual’s beliefs, but by summarily invoking his or her subgroup’s cultural standards (1996: 1126f).

More disturbingly, this ‘clearest example of how multiculturalism has influenced the law’ (Coleman, 1996: 1100) violates the equal protection doctrine and antidiscrimination principle, the cornerstones of US civil rights law, because it denies justice to the immigrant women and children harmed by immigrant offenders. Multicultural criminal law thus poses a ‘Liberals’ Dilemma’ (1996: 1096): the liberal impulse in criminal law to protect the offender (see Dworkin, 1978: 135f) leaves unprotected the victims of immigrant crimes. Moreover, the liberal’s multicultural defense of the immigrant offender ‘re-institutes a bifurcated criminal code that is frighteningly similar to the old slave codes and to the black codes that briefly existed after the Civil War’ (Coleman, 1996: 1144)—to undo what had been whole point of the 14th Amendment’s equal protection clause.

The example of multicultural criminal law raises a larger problem for advocates multicultural citizenship: the question of how to deal with illiberal minority cultures. Feminist authors in particular have pointed to the fact that endorsing (very often chauvinist and authoritarian) minority cultures may amount to the suppression of women and internal dissidents (e.g. Shachar, 1999). This is a very serious charge that, in my view, has not been convincingly rebutted by defenders of multicultural citizenship.8 Kymlicka (1995: Ch. 8) distinguishes in this context between minority rights as ‘external protections’ (which secure equality between minority and majority groups in society and are therefore legitimate from a liberal point of view) and minority rights as ‘internal restrictions’ (which suppress the autonomy of the members of minority groups and therefore cannot be endorsed by a liberal). Building his theory of minority rights on the principle of individual autonomy (rather than toleration) allows Kymlicka to be more critical of illiberal minority groups than some liberals who reject group rights but allow for internal restrictions in the name of toleration (e.g. Kukathas 1992). However, the liberal theorist’s rejection of minority rights that restrict the autonomy of minority individuals does not mean, according to Kymlicka (1995: 171), that liberals can impose their principles on illiberal groups: ‘Liberals have no automatic right to impose their views on non-liberal national minorities.’ Much as in the world of interstate relations, where the principle of nonintervention is firmly established, all that liberals can hope for in their dealings with illiberal national minorities is the soft power of ‘dialogue’—in a word, much as we might despise illiberal minority practices, we have to let them go. Note, however, that this hands-off approach may be relaxed in the case of voluntary immigrants, where ‘it is more legitimate to compel respect for liberal principles’ (Kymlicka, 1995: 170)—only, ‘how’ legitimate this is and where the margin of the tolerable ends even this most succinct account of the limits of toleration does not say.

Conclusion

Where does this discussion of the theory and practice of multicultural citizenship leave us? At the level of theory, attempts to ground minority rights in ‘oppression’ (Young, 1989) and ‘societal culture’ (Kymlicka, 1995) have run into difficulties. With the vague concept of ‘oppression,’ all of society is turned into a composite of minorities, in a kind of apartheid in reverse. The more concise concept of ‘societal culture’ prudently narrows the range of legitimate multicultural claimants, one of which, however, sees its monocultural, nationalist ambitions trivialized and distorted by the pluralist notion of multicultural citizenship (national minorities), while the other’s rights claims are left without a basis (immigrants).

Kymlicka’s liberal theory has the advantage over Young’s radical theory of being more closely aligned with actual state practices regarding minorities—no abstract principles are held against states from the outside, but liberal states are confronted with the normative implications of some of their own time-tested practices regarding minorities. However, the central shortcoming of Kymlicka’s theory is its exclusive focus on the internal rights dimension, disregarding as an unchangeable parameter the external state-membership dimension of citizenship. If the argument is that axiomatically assumed strong nationalizing practices on the external citizenship dimension justify minority rights, this nexus is empirically rendered obsolete by the trend toward de-ethnicization in liberal states. Particularly in response to immigration, liberal states have excised most ethnic references from their citizenship laws and integration policies—it is de rigueur in all of them not to ‘assimilate’ immigrants, even at the point of citizenship acquisition. As liberal states, in response to ethnically diversifying societies, are busily tilting all ethnic majority preferencing and referencing, it would be strange to demand that they reverse this trend for ethnic minorities (except in the rare and serious cases of state-seeking national minorities).

Michael Walzer (1992: 100f), in a stridently liberal rejection of multiculturalism, has drawn a distinction between the ethnically neutral American ‘nation of nationalities,’ where ‘there is no privileged majority and there are no exceptional minorities,’ and the ethnic ‘nation-states’ of Europe, whose ‘governments take an interest in the cultural survival of the majority nation.’ Since state neutrality is realized in the United States, there is no point in granting minority rights here; in the ethnic states of Europe, according to Walzer, minority rights are more appropriate, even though these states may find such rights impracticable. Walzer is both right and wrong: right in his intuition that liberal states can (and do) live up to the ideal of neutrality, thus rendering the idea of minority rights pointless; but wrong in his belief that striving for public neutrality is a privilege of the United States. The United States is not different in kind from European states. All European immigrant-receiving states are moving in the same, American direction of politically constituted nationhood and territorial citizenship.

The trend towards state neutrality also in Europe is perhaps best documented in the recent controversy in Germany over a ‘deutsche Leitkultur’ (German dominant culture) that immigrants should adopt. The notion was introduced by the conservative opposition party in parliament (CDU), as an antidote to Germany’s current opening toward new labor migration. Interestingly, when pressed to define it, its proponents could not say what exactly the deutsche Leitkultur was. For the CDU parliamentary leader, it consisted of the ‘constitutional tradition of our Basic Law,’ the ‘European idea,’ equality of women, and the German language—in that order, with the only specifically ‘German’ marker appearing last, and overlapping with the functional language requirement. This was also precisely how the SPD Chancellor, like most in the political élite an opponent of the notion of Leitkultur, defined the criteria of immigrant integration. A CDU position paper on immigration, notable also for the party’s retreat from its long-held mantra that Germany was ‘not a country of immigration,’ finally included the contested notion of Leitkultur. It identifies as ‘Christian-occidental culture’ the value added by Leitkultur to the constitutional and language obligations that are agreed by all. However, this ‘culture’ is already circularly interwoven with Germany’s laws and constitution, and most disturbingly—it does not contain anything that is particularly ‘German’: every country in Europe, and many countries beyond Europe too, share this ‘culture.’

David Miller, philosophical proponent of liberal nationalism, defined national identity, among other things, as a ‘distinct public culture’ (1995: 27), which is meant to be more than the ‘common currency of liberal democracies,’ because it provides an answer to the question ‘why the boundaries of the political community should fall here rather than there’ (1995: 163). Germany’s inconclusive wrangling overLeitkultur shows that a liberal state cannot formally commit its immigrants to anything that exceeds the procedural canon of liberal democratic rules. This creates the paradox that the political community that immigrants are to be socialized into has to remain unnamed. And, for our purposes, it leaves unclear what exactly ‘multicultural citizenship’ is supposed to remedy.

At the level of practice, multicultural citizenship as written on the forehead of the state has remained exceedingly rare. As the few European states that once practiced multicultural policies (though not: multicultural citizenship policies) are moving away from them, perhaps only Canada and Australia qualify—though their multicultural citizenship differs from that of the theorist by being a citizenship for all, not just for minorities. At the same time, the state-centered ‘top-down’ notion of multicultural citizenship has deflected from the multiple ‘bottom-up’ successes of multicultural claims in the decentered subsystems of differentiated societies, which have remained short of official state recognition. Considering that multiculturalism is de facto everywhere in liberal societies, whereas it is explicit policy only in some countries (and for some groups therein), it may be better to use a diverse vocabulary to capture a diverse reality—and not to swallow the latter under the general and in important respects misleading rubric of ‘multicultural citizenship.’