African Traditions Online Encyclopedia Wiki

A review of four decades of anglophone research. BY CAROLA LENTZ*. web address of pdf source :












Ethno-nationalist movements in post-Communist Eastem Europe, brutal ‘ethnie cleansing’ in the former Yugoslavia, radical right-wing violence against foreigners in Western Europe and the growing attraction

of old-new racist ideologies, sometimes in the guise of seemingly liberal ‘multiculturalism’...: these recent developments, like the racial conflicts of the 1960s and 1970s in the U.S.A., are a brutal reminder

that ethnicity cannot simply be explained away,neither with modemization theories about stubbom but dying relies of pre-modem mentalities nor neo-Marxist concepts of ‘false consciousness’. The global ethnicization of social identities and conflicts may at least reassure Africans and Africa scholars that ethnie or tribal articularism is not the specifically African problem it once appeared to be. In the years to corne, ethnicity, in whatever oncrete form and under whatever name,Will be so important a political resource and an idiom for creating communitythat today’s social scientists and anthropologists have no choice but to confront it. A review of the past four decades of research on ethnicity and tribalism in Africa may perhaps aid in a better under-standing of ethnicization processes outside Africa as well.

The literature on ethnicity and tribalism in Africa is so voluminous that this essay cari only survey the most important lines of research and refer to some relevant case-studies. It Will concentrate on Sub-Saharan Africa (particularly west and southem Africa) and on the Englishlanguage literature which, 1 suspect, is not always well-known to the Francophone social science reading public. After some introductory remarks on the controversy over primordialist versus constructionist concepts of ethnicity I Will outline three major currents in research on ethnicity in Africa, each of which has been dominated by a different discipline and is deeply entwined with the actual history of ethnicity. These currents, in chronological order (although with temporal overlap),are studies of tribalism in the context of labour migration and urbanization undertaken by British social anthropologists particularly in the 1950s and early 196Os, discussions centred on politicized ethnicity and nation-state integration in the 1960s and 1970s, particularly among political scientists, and finally the study of the colonial ‘invention of tradition’ (RANGER, 1983) and ‘creation of tribalism’ (VML, 1989), which has been carried on by historians of Africa since the 1980s. Some general thoughts on the genesis and development of ethnie communities and discourses in colonial and post-colonial Africa round out the discussion.


‘Ethnicity’ is a dazzling, ambiguous category, at once descriptive and evaluative-normative. It has long since ceased to be the exclusive domain of social scientists, having entered the practical vocabulary of politicians and social movements2. In both spheres, the terms ‘ethnicity’ and ‘ethnie group’ frequently absorb, overlap or replace other concepts such as ‘race’ or ‘tribe’ which have become problematic. In the early 197Os, for example, SOUTHALL (1970: 47-48) called for the replacement of ‘tribe’, then current among Africanists, by ‘ethnie group’. His argument was not that the latter term offered greater analytical clarity, but rather that the primitive connotations of the former affronted the sensibilities of African colleagues3. In South Africa. referentes to culture and ethnicity allow liberal as well as Christian-nationalist Afrikaaners to emphasize difference without resorting to the biologistic and discredited concept of race (D~OW, 1987). ‘Ethnicity’ functions like the joker in a tard-game: it cari be introduced into various play sequences, taking on the characteristics - in this case, connotations and conceptual vagueness - of the tard it replaces.

At the same time, the ubiquitous use of ‘ethnicity’ has contributed to its reification and naturalization. Classifying the most diverse historical forms of social identity as ‘ethnie’ creates the scientifically questionable but politically useful impression that a11 ethnicities are basically the same and that ethnie identity is a natural trait of persons and social groups. If, following SIVIITH( 1991: 52), we use the ter-m ethnies to refer equally to Old Testament Canaanites, early medieval Normans and modem-day Basques and Sikhs, it is no great feat to claim ‘a greater continuity between pre-modem ethnies and ethnocentrism and more modem nations and nationalism than modemists of a11 kinds have been prepared to concede’. This is not an argument which bears up to historical scrutiny. Rather, it is a nominalist operation intended to provide scholarly legitimation for ethno- ational& ideologies.And here we find ourselves at the centre of the controversy between (neo)primordialists like Smith and constructionists who take ethnicity to be an historically specific and sociahy generated pattem of identity.

Both concepts cari with some justification invoke the chequered history of the Greek term ethnos, which, as in the case of ethnicity today, was above all a political category. In Homer, ethnos, still free of connotations of a common culture, language or history, mainly referred to large, undifferentiated groups of either animals or warriors (in the sense of ‘swarm’ or ‘throng’). Later, Aristotle used the word as a term for both Greek and non-Greek segmentary societies (or the ‘segmentary state’, as Ehrenberg renders it), as opposed to polis, the Greek urban polity. In New Testament Greek, ethnos stands for ‘heathen’, and the adjective derived from it, ethnikos, for ‘barbarian’ and ‘uncivilized.’

Thus ethnos is embedded in a context-specific welthey dichotomy and was, to a certain extent, originally associated with others’ and a lower stage of civilization or political development. Apparently only in the context of the consolidation of the Ottoman Empire in the fifteenth century did ethnos become a term of self-identification for Greek Orthodox Christians and, finally, in conjunction with nineteenth-Century Greek nationalist efforts, a term connoting a ‘we group’ with a common culture and history”.

The opposition between ‘individualism’ and ‘romantic collectivism’ which has marked European intellectual history since the eighteenth century (GELLNER, 1993) also t-uns through the debate about ethnicity, and it has become common practice to distinguish between constructionist (or formalist) and primordialist (or essentialist) approaches”.

Constructionists emphasize that ethnicity is not suprahistorical and quasi-natural membership in a group. but rather a social identity constructed under specific historical-political circumstances. They insist that researchers must not naively adopt the actors’ own discourses of ethnie identity, which typically claim ‘hereditary’ membership in an ethnie group as a group ‘overlapping and including the family’, a common history and cultural similarity (ELWERT, 1989). Ethnie groups, So the basic assumption of the constructionists goes, exist only in the plural, in the relationship between ‘us’ and ‘others’. BARTH ( 1969: 14-15), in particular, has criticized the equation of ethnicity with a common culture, insisting that ethnie groups are only constituted through the construction of social boundaries - as self-ascription and ascription by others”. Constructionists emphasize the subjective manipulability, flexibility and strategic quality of ethnicity, but arguments and positions vary widely in detail. Some authors, like BANTON (1983) apply rational choice theories to ethnicity or otherwise look for the ‘objective’ interests upon which ethnie identity is allegedly based. Still others study the cultural construction of social identity, often ignoring questions of power (SOLLORS, 1989). Yet others stress the political instrumentalization of ethnicity by social movements (ARONSON, 1976) or competing élites (BRAS~, 1991).

Essentialist concepts of ethnicity emphasize the significance of ‘primordial tics” and a ‘given’ common history (ancestry), culture and language. VAN DEN BERGHE'S ( 198 1: Il- 12) sociobiological model, which conceives of ethnicity and race as ‘expansions of kinship’ and of ethnocentrism and racism as ‘biologically evolved mechanisms of pursuing self-interests’, has not been embraced in this extreme form by any other social scientisP. ISAAC~’ (1975: 34) scarcely less problematic metaphor-laden definition of ethnicity as a ‘basic group identity’ which a11 members inherit at birth and which satisfies the human need for ‘belongingness and self-esteem’ much better than the ‘secondary group identities’ acquired later in life has been and continues to be more widely accepted. Geertz’s analysis of the role of ethnie ties in the new African and Asian states, which combines primordialist and historicalpolitical arguments has been particularly influential, however.

According to Geertz, it was only in the context of economic and political ‘modemization’ that ethnicity became a virulent idiom for defending particularist interests. But the ethnicity thus mobilized is itself presented as a traditional ‘primordial attachment’: “that stems from the ‘givens’ - or, more precisely, as culture is inevitably involved in such matters, the assumed ‘givens’ - of social existence: immediate contiguity and kin connection mainly, but beyond them the givenness that stems from being born into a particular reIigious commufiity, speaking a particular dialect of a language, and following particular social practices” (GEERTZ, 1973: 259).

Although these and similar postulates of the undoubted and a priori givenness of ethnie identity have been refuted by countless empirical studies, primordialist approaches have proved extraordinarily tenacious. This may be due in part to the deficiencies of simplistic constructionist theories, which have had a difficult time explaining such phenomena as why people are prepared to die for ideologies of identity which supposedly arose out of rational political interests. It is nonetheless extraordinary, as Comaroff has pointed out, how stubbomly the controversy between (neo)primordialists and constructionists is constantly being recast. The only explanation he cari find for this phenomenon is that the questionable theories themselves belong to the ideological arsenal of contemporary identity politics: “Ethno-nationalisms see their own roots in prima1 attachments: it is by virtue of these attachments - and by effacing the traces of their historical construction - that claims to ethnie self-determination are typically conceived and justified. As a result, primordialism appears to account for, and to valorize, this kind of identity.

By contrast, Eure-nationalism [that envisages a secular state founded on universalist principles of citizenship and a social contract] locates its origins in narratives of human agency and heroic achievements. It is, alike for those who hold it as worldview and for those who seek to analyze it, an historical çreation; not surprisingly, it seems most persuasively illuminated by one or the other form of constructionism. And hetero-nationalism [that seeks to absorb ethno-national identity politics within a Eure-nationalist conception of political community] tends to be rationalized and explained by recourse to neoprimordial instrumentalism. Both the former and the latter hold that cultural identity has a primai basis; an immanent, enduring essence that is bound to express itself as soon as its bearers find cause and/or occasion to assert common interest. And both agree, explicitly or implicitly, that - inasmuch as such assertions are founded on ‘natural’ affiliations - they are undeniably right and proper” (COMAROFF, 1993: 33-34). 

Such political connotations and the peculiar mixture of instrumentalist and primordialist arguments cari also be located in many of the studies of ethnicity in Africa outlined below.


In the 1940s and 195Os, issues of social change, particularly the phenomena of labour migration and urbanization, gained in significance in British social anthropological research on Africa. A number of methodologically and theoretically innovative studies of ‘African urban systems’ (MAYER, 1962: 576) appeared, most of them focusing on the Rhodesian copper belt, which exercised a strong influence on the field more generally”. In the mining towns, but also in other larger towns in southern Africa, the majority of the urban population was composed of migrant worlters who returned, whether voluntarily or not, to their rural places of origin after shorter or longer stays in the City. Early studies which focused chiefly on the cultural implications of this mobility interpreted the adaptation of rural migrants to urban conditions as a process of ‘detribalization’, with clearly negative connotations (e.g. WILSON, 194142). Gluckman, Mitchell, Epstein and other scholars from the Rhodes-Livingstone Institute insisted, in contrast, that the town and rural tribal homes represented different social fields, in which the migrants developed different forms of behaviour and organization appropriate to their respective situations. Or, as Gluckman expressed it in his famous pronouncement:

“An African townsman is a townsman, an African miner is a miner: he is only secondarily a trîbesman [...] the moment an African crossed his tribal bnundary. he was ‘detribalised’, outside the tribe, though not outside the influence of the tribe. Correspondingly. when a man retums from the towns into the politica