Материал: Rethinking ethnicity, 2nd Edition by Richard Jenkins

22 Rethinking Ethnicity

questions – about the nature of collective social forms – in which Barth is interested. In their enthusiasm for processes of boundary maintenance, negotiation and strategizing, many of those whose analyses owe much – if not everything – to Barth have taken for granted, and continue to take for granted, what he did not, the ‘group’ that is the object of their research.

Two other major criticisms are worth brief consideration. First, there is the apparent difficulty experienced by many anthropologists in handling the distinction between the ethnic and the ‘racial’. Banks suggests (1996: 97–100), that the most typical anthropological response to this problem is either to ignore it or just to mention it in passing. It is certainly not sufficient to adopt Wallman’s solution and simply dismiss the distinction as ‘a quibble’. John Rex has argued, for example, that we must pay attention to the importance and significance of folk ‘racial’ categorizations, because:

a far wider set of situations are based upon cultural differentiation of groups than those which are commonly called racial and … few of them have anything like the same conflictual consequences that racial situations do. (Rex 1973: 184)

Since Rex wrote that, events have somewhat caught up with the second part of his argument: Northern Ireland, the Balkans, southern Sudan, Iraq, one could doubtless cite other ethnic conflicts which seem to give it the lie. But it is really only somewhat and only seem. Ethnicity is still a wider classificatory or organizational principle than ‘race’, and it remains true that few ethnic conflicts are as bloody as ‘racial’ ones (and the genocide in Rwanda is depressingly instructive here).4 Furthermore, the boundary between ‘race’ and ethnicity is historically variable; what was ‘racial’ before 1945 may be more publicly acceptable as ‘ethnic’ today.

The second critical point is that anthropologists are reluctant to acknowledge and ‘deal with’ power. This apparent lack of enthusiasm is the basis for what sometimes appears a disciplinary predilection to treat ethnicity as a positive social resource. This criticism is related to the first, in that ‘racial’ categories and racism are characteristic of situations of domination. It is also bound up with the relative lack of theoretical attention which has been given by anthropologists to relationship between ethnicity, on the one hand, and class and stratification, on the other (Eriksen 2002: 49–51). Despite the important contributions made by political anthropologists to our understanding of how humans dominate each other (e.g. Gledhill 2000), and recent heterodox anthropological contributions to debates about power (e.g. Rapport 2003), there is still, at least, a case to be answered here. However, if there is a disciplinary ‘blind spot’ about power, it is not an inherent theoretical incapacity. The same is true of the post-Barthian approach to ethnicity: one of the papers in the Ethnic Groups and Boundaries collection was, for example, concerned with ethnic identity as a social stigma (Eidheim 1969).

These criticisms do not reflect fatal, unavoidable shortcomings in the anthropological model of ethnicity. They reflect aspects of social anthropology – its preoccupation with social integration, and an ontology of the human world framed in terms of corporate groups – that have long been contested within the discipline itself5 and were precisely the problems that led Barth to formulate his arguments in the first place. As I shall discuss below, these shortcomings, if such they can be called, are not peculiar to anthropology, either: they have re-emerged recently, in

From Tribes to Ethnic Groups 23

Brubaker’s critique of contemporary conceptualizations and studies of ethnicity (Brubaker 2002; 2004: 7–27), as general sociological problems. An answer to both criticisms can, I suggest, be found in the distinction between groups and categories, a distinction which is elided in bald assertions, by Barth and Wallman, that ethnicity depends on ascription from both sides of the group boundary. It is important to distinguish in principle between two analytically distinct processes of ascription: group identification and social categorization. The first occurs inside and across the ethnic boundary, the second outside and across it.

Social categorization, in particular, is intimately bound up with power relations and relates to the capacity of one group successfully to impose its categories of ascription upon another set of people, and to the resources which the categorized collectivity can draw upon to resist, if need be, that imposition. To acknowledge the significance of the distinction between group identification and categorization is to place relationships of domination and subordination on the theoretical centre-stage.

Similarly, the distinction between groups and categories may help to clarify the contrast between ethnicity and ‘race’. Banton has argued that ethnicity is generally more concerned with the identification of ‘us’, while racism is more oriented to the categorization of ‘them’ (1983: 106). From this point of view ethnicity depends, of necessity, upon group identification, while ‘race’ or racism is most typically a matter of social categorization. Such an understanding is, however, only helpful if it is heavily qualified. First, group identification and social categorization are inextricably linked, logically, inasmuch as similarity entails difference (and inclusion exclusion) and processually, with respect to individual as well as collective identification (Jenkins 2008). Furthermore, on the one hand, hostile categorization with reference to putative ethnic or cultural criteria is common, and, on the other, groups may come to identify themselves positively in ‘racial’ terms, as in the more or less supremacist ideologies of societies such as Imperial Britain, National Socialist Germany, Imperial Japan, or the South Africa of apartheid.

So although group identification and social categorization are analytically distinct, each is bound up with, and makes no sense without, the other. Ethnicity may, for example, be strengthened or generated as a response to categorization; similarly, an aspect of a group’s ethnic identification may be – indeed is likely to be – the categories with which it labels other groups or collectivities. In each of these cases the centrality of power relationships – a centrality that many ethnicity theorists have been reluctant to include in their analyses – must be acknowledged.

To return briefly to the quotation from John Rex, above, ethnicity is a more general social phenomenon than racism or ‘racial’ categorization. It is equally clear that ethnicity, although its emphasis may conventionally be thought to fall upon group identification, is routinely implicated, through the signification of cultural or ethnic markers, in processes of categorization. Bearing these considerations in mind, racism may be understood as a historically specific facet of the more general social phenomenon of ethnicity. As such, it characterizes situations in which an ethnic group (1) dominates, or attempts to dominate, another set of people and (2) seeks to impose upon those people a categorical identity that is primarily defined by reference to their purported inherent and immutable differences from, and/or inferiority to, the dominating group. Viewed from this perspective, the distinction between ethnicity and ‘race’ is more than a quibble, and demands the most serious attention (Miles 1982: 44–71; Rex 1986: 18–37; M.G. Smith 1986: 188–94). This is

24 Rethinking Ethnicity

not least the case because ‘race’, as a generic concept and in its local versions, is nothing but socially constructed (Wade 1993, 2002).

Groups and boundaries revisited …

In its generic emphasis upon social construction, its recognition of fluidity as well as solidity, and its refusal to naturalize ethnic identification in the face of powerful essentialist discourses, the post-Barth ‘basic anthropological model of ethnicity’ remains arguably the best approach to the topic. It has survived a fair amount of criticism to become the increasingly comfortable intellectual furniture of several generations of anthropologists interested in ethnicity and analogous collective identifications. From Cohen’s analysis of the ‘symbolic construction of community’ (1985), to Baumann’s multi-layered exploration of ‘discourses of identity’ in west London (1996), to Balzer’s study of the historical and contemporary twists and turns of ‘indigenous’ identification in Siberia (1999) – and these are necessarily invidious personal choices – its influence is wide-ranging.

This influence is also obvious in research, across disciplines, which does not explicitly mention Barth. Good examples are Lentz’s historical anthropology of ethnic identification in northern Ghana since initial colonization (1994, 2000), a loosely related set of studies of the ethnic complexity of contemporary Miami (Grenier and Pérez 2003; Grenier and Stepick 1992; Portes and Stepick 1993; Stepick et al. 2003) and Lamont’s sociology of the ‘boundary work’ of class and ‘race’ in France and the United States (1992, 1995, 1997, 1999; Lamont and Molnár 2001). In these cases, and many others that I could have selected, the apparent oversight – if, indeed, fairness permits me to describe it as such – is not a matter of ignorance, conceit or carelessness. I have taken care to choose examples where these are not conceivable. Nor are these authors making the claims to radical novelty that I criticized in Chapter 1. The explanation is, I think, more simple: once a set of ideas becomes sufficiently accepted and influential, it simply passes into the axiomatic common sense of the field and becomes common property.

Although the intellectual genealogy of the ideas about ethnic identification that I am exploring in this book reaches back to the sociology of Weber and Hughes, there seems little doubt that Barth’s Ethnic Groups and Boundaries was ‘particularly germinal’ (Lamont and Molnár 2002: 174) in providing the stimulus for their eventual widespread acceptance. As Malesevic has put it:

Barth’s Ethnic Groups and Boundaries provided nothing short of a Copernican revolution in the study of ethnicity. Barth turned the traditional understanding of cultural difference on its head. He defined and explained ethnicity from the outside in: it is not the possession of cultural characteristics that makes social groups distinct but rather it is the social interaction with other groups that makes the difference possible, visible and socially meaningful. (2004: 2–3)

Accepting this, the previous sociological history of these ideas reminds us that the ‘basic anthropological model of ethnicity’, as I have called it, is but one version of a long-standing and wide-ranging interactionist sociology of identification (Jenkins 2008; Malesˇevic´, 2004: 61–77).

From Tribes to Ethnic Groups 25

Very recently, that interactionist sociology has provided the context for Rogers Brubaker’s influential, and sceptical, revisiting of the conceptualization of ‘ethnic groups’ in particular, and ‘identity’ more generally. Inspired not least by Barth, and by the perspective that sees groups as somewhat indeterminate, in that their boundaries and membership are changeable and uncertain, Brubaker insists that ethnic groups, as he believes they are most generally understood within social science and elsewhere – as definitely bounded, internally more or less homogeneous, and clearly differentiated from other groups of the same basic kind – don’t actually exist. They are not, in fact, real. It is only a shared sense or image of ‘groupness’ that is real; the participants in ethnic conflicts are thus actually individuals and organizations, not ethnic groups (Brubaker 2002; 2004: 7–27). Ethnicity, for Brubaker, is a point of view of individuals, a way of being in the world (2004: 64–87; Brubaker et al. 2004). Employing similar logic, Brubaker further argues that identity more generally is not real, either, in the sense that it is not a ‘thing’ that people can be said to have or to be (2004: 28–63; Brubaker and Cooper 2000). Instead we should talk about ongoing and open-ended processes of identification. By this logic, identity does not impel people to do anything; it is, rather, people who engage in identification.

Up to a point, Brubaker is right, in both of these respects. It is certainly true, as I argue in this book, for example, that whatever reality can be attributed to groups depends on people thinking that groups exist and that they belong to them. It is also certainly true that identity depends on processes of identification and does not determine, in any mechanistic or causal sense, what individuals do (Jenkins 2008). Individual behaviour is a complicated outcome of conscious decision-making, habit, emotion, health and well-being, access to resources, knowledge and worldviews, the impact on them of what others do, and probably several other factors besides. Neither (ethnic) group membership nor (ethnic) identity can be said to determine anything.

Brubaker is, however, right only up to a point (and for a further supportive critique of his position, which is complementary to that which I offer here, see Ruane and Todd 2004). In contrast to the definition of groups that Brubaker presents as wrong-headed, conventional social science wisdom – that is, as clearly demarcated and bounded, homogeneous collectivities – there is another, more minimal definition, which simply says that a group is a human collectivity the members of which recognize its existence and their membership of it (Jenkins 2008: Chapter 9). There are no implications of homogeneity, definite boundaries, or, crucially, coordination of collective action. This is the definition that informs my arguments in this book, and it commands considerable support across a broad social science spectrum. Viewed from this perspective, which is also Brubaker’s own position, his distinction between apparently non-existent groups and real ‘group-ness’ doesn’t make much sense, in that groups are constituted in and by their ‘group-ness’ (see also the debate between Brubaker [2003] and Calhoun [2003a, 2003b]).

In a hard-nosed, almost puritanical, search for unambiguous analytical categories – which, in the search for the really ‘real’, have more than a whiff of both materialism and positivism about them – Brubaker’s is a sensible argument that, driven to a logical extremity, ends up somewhere less sensible. He is attempting to impose conceptual order on a human world that is not a straightforward place, in which fuzziness, ambiguity and paradox are part of everyday life. While social scientists

26 Rethinking Ethnicity

must strive for maximum conceptual clarity, our concepts must also be grounded in the observable realities of the human stuff with which we deal. If we attempt to impose concepts that are too straight-edged on this messy reality we risk divorcing ourselves from it, and, in Bourdieu’s words, substituting the ‘reality of the model’ for a ‘model of reality’ (1977: 29).

These critical comments notwithstanding, it is important to acknowledge that the social theoretical model of unitary, homogeneous groups at which Brubaker aims his darts is not a mirage. John Urry, for example, argues that the impacts of globalization mean that it no longer makes sense for the basic unit of sociological analysis to be the spatially and socially bounded set of locally focused institutions that we call ‘a society’ (Urry 2000a, 2000b). He argues, instead, that, entering a new epoch in which the old rules no longer apply, we must shift our focus away from boundaries and towards individual and collective patterns of spatial and social ‘mobilities’. He is, in effect, arguing that there are no more ‘societies’.

Despite his own argument, however, Urry remains a prisoner of the social theoretical conventional wisdom that is the focus of Brubaker’s critique, which overstates the unitary solidity and boundedness of ‘societies’ – and of groups in general – and, in the process, misunderstands them. Urry’s argument about social change only makes sense if societies as he understands them – as largely self-regulating, possessing definite boundaries, culturally unified if not uniform, different in important respects from other societies, and self-reproducing – have, historically, been the norm of human collective organization. However, viewed in the light of the historical and ethnographic records, this ‘formal organization’ or ‘nation-state’ model of human groups does not correspond particularly well to the long-run observable realities of the human world. It has probably never been the only, or even the dominant, form of collective human organization. That the last two or three centuries or so have witnessed increasingly determined efforts, in some regions of the world – and, increasingly, everywhere – to formalize and exert control over the boundaries of collective identification and territory and the ‘cultural stuff’ within them, does not mean that fixed, definite boundaries or cultural homogeneity are universal human collective patterns. If they were, less determination and effort might be required to maintain and police them, and they might not, as in Urry’s account of the present state of globalization, be threatened today. His argument is an exemplary illustration of social theory’s need to reflect upon and transcend its axiomatic modernism.

Urry’s work, in raising the issue of the territoriality of modern nation-states, offers a bridge into the last topic that I wish to consider here. Where Barth emphasized that interaction across group boundaries produces and reproduces group identifications, a more recent body of literature, taking its lead from Sahlins’s pioneering historical research on the Pyrenees (1989), has begun to explore border regions as distinctive locations, worthy of research in their own right (Donnan and Wilson 1999; Lamont and Molnár 2002: 183–5; Wilson and Donnan 1998). In addition to their role in dramatizing and underlining the national state identities that meet on their ground, territorial borders – and I suspect that this is an argument with wider applications – are revealed as imprecise zones that may, to at least some degree, be independent of, or different from, the states in question. They offer possibilities for the emergence of distinctive ‘border identities’ and ways of doing