Engaging with the world

Eriksen's site

  • facebook
  • wordpress
  • yu
  • Academia
  • Nflag






africaA non-ethnic state for Africa?

On integration, disintegration, difference and similarity in contemporary Africa

Thomas Hylland Eriksen

Lecture delivered at ASEN conference, London, 11 May 1996; subsequently published in Ethnicity and Nationalism in Africa (1999), edited by Paris Yeros.


The following discussion on social and political identification, with particular reference to that vast continent conveniently summarised under the label "sub-Saharan Africa", is partly epistemological, partly analytic and partly plainly policy-oriented. I will set out by outlining some standard social science models of nationhood and ethnicity, then proceed to a discussion of social and political identification and integration with reference to a few examples, before considering a perspective which, in my view, has been neglected in too many scholarly analyses of collective identity formations.

How does a sense of group membership develop, and under which circumstances do groups behave as inverted refrigerators - emanating coldness outwards for every degree of heat painstakingly generated inwards? What is it about collective identification that makes it so susceptible to being exploited by warlords and Machiavellan power brokers? And - granted that ethnic group sentiment cannot be done away with by scholars and politicians fuelled by Enlightenment sentiment - how can all the energy invested into ethnic politics be harnessed for the progress of humanity?

Such are some of the typical questions asked by academics studying ethnicity and nationalism. Being one of those academics myself, I should add that something important is often missing from our scholarly diagnoses of contemporary ethnicity and nationalism: The main paradigm represented by academics is not only at odds with, but to a large extent incommensurable with, the experienced life-worlds of the people we ostensibly study and, perhaps, carry out policy recommendations about. It also, with a few exceptions, tends to caricature those life-worlds, often creating contrasts between a (benign, liberal) cosmopolitan attitude and a (totalitarian, irrational) localist or ethnicist attitude; a kind of contrast which is less marked in ongoing social life than in social science models, and which is both inaccurate and potentially politically harmful. Let us keep this in mind as we move on to the world of models, explanations and empirical examples. Towards the end of the lecture, I shall return to the discrepancy.


It is widely held that the members of human groups have an "innate" propensity to distinguish between insiders and outsiders, to delineate social boundaries and to develop stereotypes about "the other" in order to sustain and justify those boundaries. If this is indeed the case, ethnicity can be conceived of as being nearly as universal a characteristic of humanity as gender and age - unlike phenomena like nationhood and nationalism, which have been so conceptualised in the academic community as to concern the modern world only (Anderson 1991, Gellner 1983). Marx and Engels held, probably correctly, that gender, age and the insider--outsider distinction were universal criteria of social differentiation. If, on the other hand, ethnicity as we conceptualise it can be shown to be a product of a particular kind of society, it can of course not be regarded as an ahistorical and universal phenomenon. I have argued elsewhere that this discussion is a non sequitur (Eriksen 1995); today, I would instead like to draw attention to the very process of collective identification, which is currently represented as a negotiable commodity in most of the world and which may just as well end in nihilism, individualism, postmodernism and creolism as in vehement and stubborn identity politics with all the conventional trappings of firm boundary-markers and negative stereotyping of others - to denote some of the more extreme options.

By isolating "ethnicity" as a focus for research, one easily loses everything else from sight. This is perhaps the cardinal sin committed by many students of ethnicity; and although one should not overestimate the importance of academic research, it can have very noticeable effects on the outside world through its potential as a self-fulfilling prophecy. Concepts can serve as both intellectual tools of liberation and straitjackets. Their only claim to legitimacy lies in their ability to help us conceptualise the outside world more accurately; when they cease to do that job, they are ready for replacement by others.


It has become a ritual exercise in social science theses and most theoretically ambitious papers on ethnicity and nationalism to interrogate the theories of nationalism developed by the late Ernest Gellner (1983), Eric Hobsbawm (1992) and Anthony D. Smith (1986), respectively, concerning whether nationalism and ethnicity are "old" or "new" phenomena and to which extent they are "invented". While everyone seems to agree that nationalism is a child of that fusion between Enlightenment and Romanticism we are accustomed to label Modernity, not everyone is convinced that it is this historically recent. For didn't already the Vikings distinguish between Dane, Swede and Norwegian? And didn't the ghostwriters of the Bible attribute a saying about Jews and Greeks to Jesus Christ? And weren't the ancient Greeks pathologically xenophobic? To this, Modernists would reply that although the collective self--other distinction harks back to the mists of prehistory, the peculiar reflexive character of the modern individual was missing and besides, the nation-states places peculiar demands on its citizens and an abstract solidarity from them.


At this point, enter Benedict Anderson, famously quoted for the title of a book few seem to have read (Anderson 1991); with all due respect, if they had, they would have realised that already on page 6, Anderson notes that in a certain sense, all communities beyond the family are imagined; in other words, to state that the nation is an "imagined community" is pretty vacuous as a distinguishing mark. Rather than focusing on ethnic identity as the paradigmatic prerequisite for nationhood, as so many do, Anderson emphasises the impact of print technology and a capitalist system of distribution in his explanation of the development of imagined communities of an unprecedented scale, involving, in shared commitment, solitarity and ritual communion, very large numbers of people who will never meet.


Although Anderson notes the similarity between ethnicity and nationalism as modes of belongingness, he does not explicitly relate nationalism to ethnicity; he does not talk of the "ethnic origins of nations" (Smith 1986). An historian of South-East Asia, Anderson writes extensively about nationalist movements in the Philipphines and Indonesia; countries which, if they are to be considered nations at all, definitely are non-ethnic ones.

Nations are conceptualised and defined in crucially different ways in different parts of the academic community. In the popular European press, incidentally, nationalism has virtually become a synonym for xenophobia. In my own discipline, social anthropology, nationalism tends to be seen as identical with ethnic nationalism (cf. e.g. Banks 1996), whereas political scientists often regard it as a chiefly civic kind of ideology, for example, what Habermas has spoken of, in the German context, as Verfassungspatriotismus ("Constitutional Patriotism"). The two "kinds" of nationalism, sometimes described as "German" and "French" or even "East European" and "West European" are not mutually exclusive in practice. Even the civic British nation, if it exists as an imagined community, has an easily identifiable dominant ethnic group, namely the English; just as the WASPs can be seen, perhaps with slightly greater difficulty, as the hegemonic ethnic group of the USA. This does not, however, mean that nations are by default built around the shared collective memories, territorial attachments, customs and values of ethnic groups. I will shortly invoke an example which shows the opposite. Besides, any sane discussion of nationhood in Africa must find its point of departure in a model of nationhood which does not equate national identity either with ethnic identity or with a subservient ethnic identity. In the African context, bickering about ethnic diversity as being somehow irreconcilable with nationhood, describing the African states as being "unnatural", with "artificial boundaries" etc., would lead us nowhere, and would also reveal a rather weak understanding of actual European nations, which, as it has been pointed out repeatedly since Walker Connor's seminal article on nationhood and ethnicity (Connor 1974), are much less mono-ethnic than it has been customary to believe.


The issue at this stage concerns to what extent a sense of common identity can be developed in the poly-ethnic African countries, and which models of the nation can be reconciled with the facts on the ground.

In order to discuss this question properly, it is necessary to dwell a little on the problem of cultural integration and its relationship to ethnicity and nationhood. Most theorists and commentators on nationalism describe it as an ideology which promotes cultural homogeneity and a subjective feeling of we-hood. The first clause is evidently true, but it needs qualifying. How much do the citizens of a nation need to have in common in order to be regarded as culturally homogeneous? The answer cannot be of an either-or type, since shared culture is a matter of degree. Smith (1995) and others have stressed that a shared public culture is sufficient, but even if that much is conceded, there are some very real difficulties associated with the delineation of such a public culture. Even societies regarded as homogeneous, such as mono-lingual countries, have greatly differentiated public spheres. The social philosopher Jon Elster once suggested that a society might be defined as a place where people stop at red traffic lights. This will certainly not be sufficient as a general definition of a nation, but it is difficult to tell exactly what is. This, obviously, is a core issue.


The second clause, regarding the subjective feeling of we-hood, is no less difficult to define in an unambiguous way, if not only because the compass and composition of the we-group shift situationally. In general, there is a widespread tendency, also present among academics, to conflate cultural similarity and subjectively defined nationhood. One may have a lot of the latter without much of the former, as when conservative Protestant fundamentalists in Western Norway and post-marxist babyboomers in Eastern Norway take it for granted that they share the same culture, despite their very considerable differences at the level of objective culture. Conversely, the middle class of Milan and the middle class of Lyon may have a lot in common culturally, but not at the level of collective self-identity. The imagined community does not have an existence unless it is being imagined actively by its members. This does not mean that it is any more "imaginary" than other communities, but that it can only exist at the intersubjective level. This is perhaps nowhere more evident than in the "soft" African states, which only rarely come into being as relevant aspects of interpersonal life-worlds.


The common conflation of shared culture with shared ethnic identity makes the task of crafting accurate concepts even more difficult. Just as cultural similarity does not by itself lead to collective identification, nor does ethnic similarity vouch for cultural similarity, even if there is often a strong correlation. As it has been pointed out many times, the most protracted and bitter ethnic conflicts are often played out between ethnic groups who, at the level of culture, are very similar, such as Hutus and Tutsis in Rwanda, or Orthodox, Catholics and Muslims in Bosnia.


The role of history is a related question in discussions on nationalism and nationhood. To what extent do nations need a shared past as a foundation for their present collective identity? Although it is difficult to imagine a national ideology totally devoid of invocations of shared memories, their importance varies, and some theorists assert that the shared past is always a more or less arbitrary invention. In some nations, the future seems to play a more important role in popular national imagery than the past; Sweden and the USA are perhaps good examples of this. Post-colonial African nation-states have also often stressed ideas of progress and modernity rather than the mythical past in official national symbolism.


Let me sum up this part of the discussion before moving on to the main example. Some of the moot points in the discourse over nationalism with immediate relevance to the discussion about African nationhood, are the following:


(i) Whether or not ethnic groups are "modern" or not is a contested issue. In the classic collection edited by Fredrik Barth (1969), most of the contributions dealt with traditional societies. For my purposes here, however, I shall accept that ethnic corporations in Africa are recent developments associated with the rise of individualism, capitalism and the state: Earlier corporations and collective identities, although they might have carried a label later seen as ethnic, were constructed according to a different logic.


(ii) The relationship between conceptualisations of ethnicity and of nationalism is complicated. Some analysts see nations simply as ethnic groups writ large, with leaderships of state-building ambitions; others stress the civic, non-ethnic aspects of nations. Scarcely a single state is ethnically homogeneous, and many do not even have a dominant ethnic group. In the following, I shall see the African nation as a supra-ethnic or poly-ethnic phenomenon, which may nevertheless be appropriated by ethnic groups in a number of ways, including power monopoly and secessionism.


(iii) The relationship between nationhood and culture has been insufficiently explored. Although a theorist like Ernest Gellner has brazenly, and probably correctly, stated that nationalism and successful nation-building inevitably imply cultural homogenisation, the degree of cultural similarity varies both inside and beyond the boundaries of the state in question.


(iv) Cultural similarity, ethnic incorporation and collective self-identity are granted unequal weight in rivalling conceptualisations of nationhood. In Eric Hobsbawm's work, for example, the collective self-identification takes on a paramount importance, while A.D. Smith stresses the sharing of customs and memories, frequently through an idiom of ethnicity.

(v) The role of historical continuity, whether real or imagined or both, is also a difficult aspect of nationhood and one faced with deep gravity by the African states, many of them manifestly lacking a shared history.
Commenting on the genre of biography, Virginia Woolf lets the protagonist of her novel Orlando (1928) assume an enormous and seemingly unrealistic variety of social roles and identities, right down to the point of changing gender. Yet, the narrator says towards the end of the book, she has written only a small fraction of the possible biographies that could be written about Orlando; in other words, people have many more facets than a single book might reveal.


The people of Mauritius, I think it would be fair to say, collectively exploit a fair proportion of Orlando's vast role repertoire. Mauritius, a multi-ethnic island-state in the south-western Indian Ocean, has for historical reasons an ethnically very diverse population of about one million, four major religions, a large but uncertain number of languages, and no indigenous population (Eriksen 1988, Bowman 1990). Widely considered an economic miracle in the 1990s, Mauritius is also a stable multi-party democracy which has experienced several changes of government since independence in 1968.


Arriving on the island to carry out fieldwork in early 1986, I half expected to find a society where postmodern relativism was as deeply ingrained as the faith in technological progress had been in Europe in the 1950s; instead, I encountered a very wide range of perfectly solid and confident personal identities, often based on qualitatively different premises. The eclectic approach to identification which can be observed in a society such as Mauritius is evident already in the now obsolete colonial grid for dividing the population into ethnic categories. The last of several such classificatory exercises, abolished in 1982, divided the Mauritian population into four mutually exclusive categories: Hindus, Muslims, Chinese and "General Population". Two of them are religious categories, one refers to an ancestral country, and the final one is residual and contains most but not all of Mauritius' Catholics, with origins as diverse as France and Madagascar, and with no collective sense of ethnic community. However, if a common myth of origin is an important defining mark for an ethnic group, then the island has at least eight ethnic groups, if one chooses to stress the endogamy rule, the number may rise to around twenty, whereas if ancestral language is to be invoked as a differentiating criterion, fifteen ethnic groups might be counted. This ought to make it clear, if there should be any doubt, that ethnicity is and remains a relational and situational kind of social phenomenon. However, not just the ethnic identities are important in Mauritius although they have justly or unjustly formed the focus of most research on identification in the island. Let me, by way of illustration, provide a list of the over-arching identities I have recorded during fieldwork. By over-arching, I mean identities which may in certain situations overrule all other identities and appear as imperative in the sense that they induce action.


(i) Ethnic identity. Although criteria for ethnic differentiation are not consistent with each other, there is always a close link between ethnic identity and kinship. Politics, jobs and marriage are often, but far from always, regulated through an ethnic idiom.


(ii) Class identity. Trade unions frequently cut across ethnic lines, and two general strikes in independent Mauritius testify to their occasional efficacy. Also, the sense of belonging to a certain class rather than to an ethnic group is strong in parts of the urban middle class, which often intermarries and tends to regard itself as cosmopolitan. Shared professional identity, also not negligible under certain circumstances, can also be included here.

(iii) Gender identity. There are several feminist organisations, which have a certain impact on public debate and which explicitly try to bridge gaps between Indo-Mauritian and Creole women. Of course, gender identity is also highly relevant in a great number of everyday situations, probably most.


(iv) Age. Youth increasingly tend to share social idioms, networks and activities irrespective of ethnic background, in addition to in most cases going to multiethnic state schools.


(v) Religious identity. Although the correlation between religion and ethnic identity is high, Chinese Catholics, Indo-Christians and Muslim Creoles, among others, worship with members of ethnic groups other than their own.


(vi) Local identity. Although the ethnic element is rarely entirely absent from local politics in Mauritius, villages are often united on single issues, and sometimes divided along non-ethnic lines (cf. Eriksen 1988).

(vii) Political identity. Several political groups are explicitly based on non-ethnic premises, although national politics in Mauritius remains thoroughly ethnified.


(viii) Linguistic identity. Although most Mauritians speak a French-lexicon Creole (Kreol) as their first language, other languages are also spoken in the island. Notably, French-speakers, regardless of ethnic affiliation, militate for the continued strong position of French in public life. Some of them would be, for example, Tamils by ethnic origin.


(ix) Supra-ethnic national identity. Although this form of identification is most evidently present in parts of the growing urban middle class, among intellectuals, academics and the like, there have been historical situations where very large numbers of Mauritians have visibly associated themselves with the multi-ethnic nation rather than with their ethnic group; for example, during and after the ritual following the death of Mauritius' first Prime Minister Sir Seewoosagur Ramgoolam, and in connection with large international sports tournaments (Eriksen 1994). Significantly, during the controversy concerning whether Mauritius should change its Commonwealth status to that of a Republic (which happened on March 12, 1992), views did not diverge systematically along ethnic lines.


Mauritius now seems to be in the process of developing a common set of supra-ethnic, national myths and symbols which is invested with meaning and relevance by the bulk of its population, although ethnic identification still remains strong. Some of this mythical material harks back to the mists of colonialism, but some of it relates to the turmoil and social unrest in the years around Independence (Eriksen 1993b). It is impossible to state whether this symbolic framework will in the long run prove too feeble and fragile to sustain a sense of unity among its ethnic and non-ethnic groups and networks; it is nonetheless clear that any viable Mauritian supra-ethnic nationalism will have to reconcile itself not only with ethnicity, but also with the emerging non-ethnic constituent parts of society. Ethnic, non-ethnic and post-ethnic elements, in other words, coexist side by side in a precarious, but nonetheless stable equilibrium.


Two complementary, and sometimes competing, models of the Mauritian nation coexist, and each has its symbols and rituals. One of them is commonly spoken of as le pluriculturalisme mauricien and sees the nation as being made up by the very "cultural mosaic" that it embodies, and locates nationhood to the interface between the constituent ethnic or cultural groups and their mutual respect. The other model lifts nationhood to a supra-ethnic level and depicts Mauritian nationhood as the universal values and institutions that all Mauritians share; the political, legislative and educational systems, the territory, and recently, the successful export economy. To the extent that these formal, or officially constructed, visions of nationalism are successful at the informal or popular level (cf. Eriksen 1993a), they appeal both to ethnic sentiment and to supra-ethnic nationhood; both to particularistic yearning for community and historical embeddedness, and to the universalistic ideals of the bureaucratic state. Although it is by no means unproblematic to develop, and although it requires a great deal of compromise and improvisation as one goes along, a combination of these two nationalisms seems to solve many of the potential contradictions inherent in multi-ethnic states.

If this had been a conference on xenophobia and exclusion in Europe, I would at this point have raised the question of what Europe can learn from Mauritius. Here, it is more appropriate to ask what Africa could learn from Mauritius (a question already asked by post-apartheid policy makers in South Africa). The answer is simple. The model can be learnt, quickly and freely. On the other hand, it is unlikely that it can be implemented wholesale through the application of mere political will, since several of the institutional and objective underpinnings of the Mauritian model are missing in most African countries:


(i) The boundaries of the Mauritian nation are not questioned, and no European Besserwisser or local identity politician can plausibly argue that its borders are "artificial". There can be no secessionism, no irredentism in such a small, isolated island, unlike in many postcolonial African states, from Nigeria's Biafra war to the successful Eritrean secession and the idea of an Afrikander Volkstaat.


(ii) The Mauritian population has a high level of education and a de facto high level of cultural integration, which makes a national public sphere possible. The media infrastructure - radio, TV and to some extent newspapers - reaches most households. As a result, the political system is governed by a system of common denominators.

(iii) Mauritius is a country of such small scale that the gap between elite and masses, or between centre and periphery, is much less than in most African countries. Although many African countries, like Mauritius, practice power sharing between major ethnic groups, there is comparatively little social integration between elites and masses within each ethnic group. Thus, power sharing at the ethnic group level is not tantamount to national integration, if the state bureaucratic elite is more or less cut off from ordinary people. Admittedly, nepotism remains a well established practice in parts of the Mauritian public sector, but unlike in several African countries, it cannot be said that the state is run by kinship corporations.


I have mentioned unquestioned sovereignty, high levels of education and cultural integration, and small scale as factors which may partly account for the relative success of Mauritian nation-building. As regards the second point, I might add that the dialect variations in Mauritian Creole tend to follow regional and not ethnic lines, and that although cultural differences based on ethnicity invariably crop up during conversations and are sometimes observed as well, they rarely interfere with the functioning of the shared institutions of society. The message is, in a word, that if such shared institutions work reasonably impartially and according to universalistic principles, ethnic diversity is no obstacle to nation-building. Somalia could in its tragic way provide the mirror-image of this argument: The Somalis speak the same language, share the same culture and ethnic identity in any reasonable meaning of the term, and yet their country has fallen into pieces because of a lack of functioning shared institutions.


In this brief discussion of Mauritian nationhood, I have mentioned ritual a few times. Rituals are crucial confirmers and producers of collective identification in any society; they simultaneously create a sense of identity and justify a power structure. When rituals fail to engage people, as is the case with many state-sponsored Independence Day celebrations in postcolonial societies all over the world, the power structure they implicitly symbolise is not justified, therefore, and people go elsewhere for their collective identification. In a recent book on ritual in contemporary Africa (Comaroff & Comaroff 1993), symptomatically, none of the rituals analysed, with the possible exception of Nigerian witchcraft accusations as mediated by the popular press, seem to be credible candidates as nationally cohesive forces. In creating communitas, to use Turner's (1969) celebrated term, among the members of single ethnic groups, they contribute to strengthening rather than weakening ethnic incorporation.


From the notion that ethnic identity is a threat to national cohesion in multi-ethnic societies, it would follow that ethnic rituals are dangerous centrifugal forces. The loyalty to the state would then be inversely proportional to the degree of loyalty to the ethnic group. This, indeed, is the view held by many politicians and theorists, and not just in Africa. This is nevertheless a position which, through overestimating the political dimension of ethnicity and underestimating its dimension of identification, may paradoxically inspire a politicisation of ethnicity. A suppressed kind of cultural practice may easily re-emerge as a resentful political one. State intolerance towards ethnic rituals does not usually lead to popular support for state rituals, but rather to the creation of ethnic countercultures. This is one problem which has been avoided in the dual Mauritian construal of the nation, where cultural expressions of ethnicity are positively encouraged and not seen as a threat to nationhood, but instead as complementary to it.


The paucity of rituals which are actually, and not merely officially, cohesive at the national level in many African countries, is a symptom, not a cause. The causes for the lack of national cohesion have to be found elsewhere - in the usually oligarchic and often kleptocratic political structure, in the widespread lack of lingua francas spoken by the bulk of the population, and in labour markets which fail to offer jobs to individuals who can no longer return to functioning clans and who therefore remain suspended in limbo between a lost traditional society and a mock modernity which reveals itself as little more than a showroom. The question is why it is that divisive and conflictual ethnicity emerges, or - as some would have it - re-emerges, in this kind of situation. In order to fully understand this, it is necessary to look more closely at the phenomenon of identification than I have done so far.

What is the self; what does the word "I" mean, let alone the word "we"? As A P Cohen has noted in an important recent book (Cohen 1994), mainstream social science has tended to avoid the question, taking individual agency for granted and regarding groups in society as "social facts", even if they pay lip service to the fact that identification is situational and relational: given appropriate structural conditions, they need not be politicised. Further, social identity cannot be taken for granted by the analyst, but must always be investigated empirically, and it varies which identity is construed as "the most primordial" - whether, say, gender, ethnicity, kinship or class is considered a person's most crucial criterion of belongingness in a particular society or a particular context. Therefore, an investigation of identification must begin with the individual.


People are loyal to ethnic, national or other imagined communities not because they were born into them, but because such foci of loyalty promise to offer something deemed meaningful, valuable or useful. This kind of perspective is sometimes denounced as "utilitarian" or "individualist", or even "reductionist". I would like to argue that it is not. First, as I have argued with reference to Mauritius, ethnic identification is but one of several identities any individual engages in. It is made relevant under particular circumstances which the individual may not exert much control over. Secondly, what is deemed valuable is culturally determined; it is defined from within, so to speak. This means that there may be, and indeed often are, discrepancies between definitions of the good life, especially in societies where cultural diversity is considerable. Thirdly, the "instance" that finally perceives alternatives and chooses between them, is a human being, an individual. Fourthly, it may be that say, kinship and natal villages will always command nostalgia and warm sentiment, but unless they are socially activated through some kind of resource flow perceived as relevant by the actor, they remain at the level of representations and do not emerge as social and political corporations. It is within this kind of framework we may ask under which circumstances classes and ethnic groups emerge, respectively.


In other words, individuals choose their allegiances, if not under circumstances of their own choice. For them to invest symbolically, politically or economically into a corporation or an imagined community, it must offer something in return. Which collectivities are at any point the most important both to individuals and to the functioning of society, is therefore an empirical question. This is not utilitarianism; it is tautology.

Now to the question of why specifically ethnic corporations emerge, and why they tend to be poised against the state: Commenting on the traumas of post-colonial Africa, Basil Davidson recently wrote: "The jubilant crowds celebrating independence were not inspired by a 'national consciousness' any more than were the Romanian peasants and their coevals in the nation-states crystallized some decades earlier from Europe's old internal empires. They were inspired by the hope of more and better food and shelter" (Davidson 1992, p. 185). When this failed to materialise, they oriented themselves to new - or old - foci for social allegiance. For the great Pan-Africanists, the nation-state may have been too small; for very many Africans, it was way too big, unless they happened to live in a ministate such as Lesotho or Mauritius, or if they they had the kind of Western education and middle-class experiences which gave sense to the African nation-state as an imagined community - or, again, if the state had something substantial to offer by way of education, employment, security etc. For the vast majority of Africans, a community of this scale did not tally with their personal experiences, which were strictly local.


Generally, as every European sceptical of the Maastricht treaty would agree, when a large-scale community fails to deliver the goods, structures at the medium and low levels of scale emerge. Although these structures have always been present in African countries, they become increasingly invested with potential as the higher level (the state bureaucracy) is weakened or simply severed from the majority of its citizens (who thereby become citizens in name only). We should keep in mind that a substantial part of what is conventionally described as African ethnicity is simply kinship; in other words, the failed nations are replaced by pre-existing structures, whose functioning has been transformed by historical changes, and whose political importance is inversely correlated with the strength and legitimacy of the state. That an individual in, say, Congo should place his bets on his clan rather than on the state, should be no surprise, as ethnographic studies show (Ekholm Friedman 1993). In some parts of the world, notably North Africa and the Middle East, political Islam plays the same part: it stands for resentment to the corrupted and inefficient state and an alternative path to authentic belongingness, prosperity and integrity. In general, politicised countercultures are liable to take on an ethnic character in societies where kinship is the most important local principle of social organisation.


Corporate kinship, in Africa and elsewhere, ideally works as a segmentary system. The segmentary model of identification, made famous by Evans-Pritchard (1940), but familiar to a vast number of people including Jean-Marie Le Pen, depicts each individual's modes of belonging as a set of concentric circles where loyalty, in absolute terms, becomes weaker as one moves outwards from the centre. It is illustrated in the famous formula "It is I against my brother, my brother and I against our cousins, our cousins, my brother and I against our remote relatives, etc." Each level of allegiance is activated when circumstances make it relevant. The contraction of such a system of concentric circles indicates that structures at a high level of scale are weakening and breaking down. This seems to be happening in many African societies in the nineties, where an increasing amount of capital - political, economic and symbolic - seems to be flowing through the inner circles. The severing of links between elites and masses is one important indication, which is serious not just from the peasant's point of view, but also from a nation-building perspective. As noted by Lonsdale (1996:9) in a recent discussion of Kikuyu ethnicity and Kenyan nationhood, "the experience of state power sems necessary to the growth of nationhood".


An alternative view of social identification could represent a person's identities as a set of partly overlapping group allegiances. Such multiple identities cannot be placed in concentric circles in orderly ways; they can scarcely be represented graphically at all. They cut across each other: every person has a shared identity with different people at different times, according to the situation; one belongs to a profession, a political interest group, a neighbourhood, a kin group and so on. The most fundamental human form of identification is arguably gender. In this kind of context, the status sets of individuals are not clustered about multiplex relationships to a limited number of people; they are diverse and flexible. The stiff and inflexible concept of culture typical of 20th Century academia, incidentally, has made it difficult to understand these complexities fully, since it has been difficult to conceptualise group membership as a relative and relational thing. As noted by Cohen (1994), group boundedness has simply been taken for granted in virtually all anthropological studies of culture and in all social science studies of ethnicity - and, let me add, in modern identity politics worlwide. A simple distinction between personal identity and group identity would nevertheless be sufficient to show that boundaries are relational and that groups, even if their existence is not necessarily negotiable, are situational at least with regards to their relative importance.


The policy implications are obvious: Cross-cutting ties and conflicting loyalties may contribute to reducing tension and conflict potential. A world of many small differences is safer, all other things being equal, than a world of a few major ones - such as the ones promoted by ethnic nationalisms. Here it may be noted that even a theorist sympathetic to nationalism such as Anthony D. Smith (1991; cf. Smith 1995), stresses the need for states to encourage multiple and conflicting loyalties.


It would seem, then, that if it is correct that the focus of social organisation in large parts of Africa is shifting from the grand, abstract imagined community to more manageable, tangible ones, tension may be reduced. I do not think so, for the simple reason that the weakening of the state entails a weakening of that over-arching set of rules governing intergroup relations; that is, the universalistic structure, or common denominators, referred to in the Mauritian example. The division of labour between the different levels of scale and principles of organisation characteristic of Mauritian society, which is, incidentally, necessary to any society worth more than a few traffic lights, is skewed and incomplete in many African countries, where the highest level of scale is, to varying degrees, disengaged from the other levels (cf. Hydén 1983).


If a sweeping statement can be allowed, let me offer the following: What is at stake for the majority of Africans is not primarily a vindication of their "roots", popular culture and so on - in this, African ethnicity is quite different from West European ethnicity. While meaning and communitas may certainly be scarce, more urgent concerns would in most cases be the satisfaction of basic needs as well as compelling reasons for believing that one's personal ambitions will lead somewhere. When the state institutions cease to deliver, kinship and, by extension, ethnicity, is often the only alternative. It is chiefly in this context that we must understand the emergence of modern, sometimes militant African ethnicity as mass movements. The state would in this case be a Trojan horse concealing not identity politics but kinship organisation.


I began this lecture as a comparison of different models of nationhood, with explicit reference to the multi-ethnic, post-colonial states of Sub-Saharan Africa. Obviously, most African states did not easily fit the standard model of nations as culturally more or less homogeneous, historically constituted entities. Neither, of course, did most European, American and Asian states. The crisis in the African state, I have subsequently argued, is no more a result of ethnic heterogeneity than the breakup of Yugoslavia was caused by centuries old ethnic hatred (see Bringa 1995 if in doubt). Ethnic homogeneity can contribute to national cohesion, but as the contrasting examples of Mauritius and Somalia show, it is neither a necessary nor a sufficient condition. I have argued that in so far as presumedly shared state-level institutions - education, politics, labour market - do not function according to more or less universalistic principles and are perceived as such by the non-elite members of society, the state not only is illegitimate, but it also fails to form a sustained focus of loyalty and identification for the majority of its citizens. This may seem a banal insight, but it is meant as an urge for us, the academic community specialising in the study of other people's modes of identification, not to neglect the study of the subjective level of experience.

The social science project is, like the African nation-building one, essentially an Enlightenment one, but its subject-matter is largely constituted as a Romantic world in the wide sense; one drawing on unifying metaphors rather than analytic dissecting tools for its cognitive power. In this divergence lies a great potential for misrepresentation and, as I think the African post-independence experience shows, political futility and disruption.


Theorists of culture have in recent years developed a critical attitude towards "native" appropriations of the concept of culture for ideological purposes. The classic anthropological concept of culture, originating in German romanticism (where Herder is the main reference) and cultural relativism (from Franz Boas onwards), depicts the world as an archipelago of more or less isolated cultures. Within this kind of model, culture becomes reified; it becomes like a fixed object, or a bounded vessel containing "a people". Against this model, theorists have posited a view stressing the flow, ambiguity and unbounded character of systems of meaning. One has tended to view kin groups and ethnic corporations as mere constructions, which they are in a certain sense, but not to their members, for whom they are resources which channel security, hopes and dreams.


Native theories tend to be, in some way or other, essentialistic and argue the objective, thinglike character of social and cultural identity. Criticism from an academic point of view does not change this. It may indeed seem that we, as theorists, represent perspectives which may be analytically valid but which are politically hopeless, which part of the post-independence African experience, whereby politicians and theorists eagerly applied European models to their societies, may show. It may of course be the case that the theory is avant-garde, but there can be no doubt that there are profound and systematic differences between the experiences and structures of relevance (Relevanzstrukturen, Schütz and Luckmann 1979) theorists and "natives", respectively, draw on in their constructions of identity. Academics are by default "cosmopolitans", and it is no accident that some of the most interesting writings on identity have been produced by Jews, social "anomalies" in European societies. Extrapolating from our own experiences of the world, we run the risk of generalising from our own, highly specialised experiences. As a result, large parts of the population we theorise about would feel alienated from the theoretical models and question their validity, since their own experiences and models of the world are radically different - not only from ours, but also from our representations of theirs!


The academic, or cosmopolitan, mode of engaging with the world could be described as detached, logical, disinterested and discursive. (It is not irrelevant that the largest group of Internet users by far are academics.) The academic discursive field is supra-spatial, or disengaged from place. By contrast, local modes of engaging with the world are experience-based, sensual, engaged and practical. These world views are intrinsically connected with concrete places. They are often stigmatised (by cosmopolitans) as fascist, racist, reactionary and so on, after having been caricatured as much more solid and absolute than they may actually be.


Or perhaps it would be more accurate to state that both modes of engaging with the world are practical, but derive from qualitatively different kinds of practices, different embedded experiences and thus different interests in a wide sense. The city-dweller walks on tarmac and lives in a flat; the rural dweller walks on soil and lives in the family's farm. (It is, in my colleague Eduardo Archetti's words [personal communication], impossible to develop roots in an environment where one walks on tarmac and not the earth itself.) The urban social network is based on the public sphere of anonymous individuals, while the rural one is based on kinship and neighbourhood. It is perfectly understandable that different groups, with radically different experiences, do not develop the same ways of relating to kinship, resources, belonging and identity. It is therefore the source of some worry that recent criticisms of ethnic and nationalist ideologies have not incorporated a sympathetic understanding of such variations in our frequently one-sided dismissals of "indigenous essentialisms". Blut und Boden may be a perfectly appropriate metaphor for the collective interests of rurals, just as human rights are more important to urban dwellers. Further, the sheer force of argument is scarcely sufficient to make people change their views on identity, belonging and loyalty. The rationale behind subjective identification with a collective entity is simply, as I have discussed at length, that it has something to offer which is deemed valuable, meaningful or useful. Mauritians are not by birth more civilized, more tolerant, more industrious or more democratic than say, Togolese or Angolans, but for structural reasons they are able to relate to identity options which make such virtues seem sensible and beneficial.
Tailbiting occurs in two very different ways when ethnicity is considered. While ethnopoliticians and their supporters bite the tails of others, scholars increasingly tend to bite their own. They encounter risks of self-referential inconsistency, circular argument and infinite regress, and ought to find no easy consolation in the alternatives of stringent positivism and closed objectivism. It remains indubitable that ethnic groups are created from within - subjectively and intersubjectively - and also that a mere examination of the objective conditions for their genesis does not provide a full explanation for their existence. Many of us pay lip service to rather fuzzy ideas of "interplay" between objective and subjective, or historical and structural, factors; but it is another matter altogether to demonstrate the complexity of ethnic phenomena.


What is crucial at this point is that we academics become able, not necessarily to sympathise with, but at least to understand what it is that make people tick: in order to carry out an analysis of an entire society, we must first be able to understand the subjective experiences upon which people act.


A largely implicit theme in my lecture, announced at the beginning, has been the contrast between social science models of ethnicity and nationalism on the one hand, and ongoing group identification in actual societies on the other hand. Social scientists, from political scientists to anthropologists, have excelled in developing sophisticated formal models of societal formations and cultural systems of classification, while paying much less attention to modelling and understanding the ongoing flow of social life, which relate to those societal formations and draw on those systems of classification, but which cannot be reduced to them, because persons and not systems act.


In many of our societies, including in this case both European and African ones, it seems that the academic, political and economic élites are moving in one direction and the rest of the population in another, and simplistic academic models of the world do not exactly mitigate this problem. An important task for analysts must therefore consist in relaying an understanding of life-worlds not just the way they are described by locals, but to come as close as possible to the way in which they are experienced. Only then can we claim to have understood African ethnicity, and only then should our policy recommendations carry any weight.



Anderson, Benedict (1991) Imagined Communities, 2nd edition. London: Verso.
Banks, Marcus (1996) Ethnicity: Anthropological Constructions. London: Routledge.
Barth, Fredrik, ed. (1969) Ethnic Groups and Boundaries. The Social Organization of Cultural Difference. Oslo: Scandinavian University Press.
Bowman, Larry (1990) Mauritius: Democracy and Development in the Indian Ocean. Boulder: Westview.
Bringa, Tone (1995) Being Muslim the Bosnian Way: Identity and Community in a Central Bosnian Village. Princeton: Princeton University Press.
Comaroff, Jean and John Comaroff, eds. (1993) Modernity and its Malcontents. Ritual and Power in Postcolonial Africa. Chicago: University of Chicago Press.
Cohen, Anthony P. (1994) Self Consciousness. London: Routledge.
Connor, Walker (1974) A nation is a nation, is a state, is an ethnic group, is a, Ethnic and Racial Studies, 1 (4), pp. 378--400.
Davidson, Basil (1992) The Black Man's Burden. Africa and the Curse of the Nation-State. London: James Currey.
Ekholm Friedman, Kajsa (1993) Den magiska världsbilden. Om statens frigörelse från folket i Folkrepublikken Kongo. Stockholm: Carlssons.
Eriksen, Thomas Hylland (1988) Communicating Cultural Difference and Identity. Ethnicity and Nationalism in Mauritius. Oslo: Occasional Papers in Social Anthropology, 16.
- (1993a) Formal and informal nationalism. Ethnic and Racial Studies, 16 (1), pp. 1-25.
- (1993b) Ethnicity and nationalism: Anthropological perspectives. London: Pluto.
- (1994) Nationalism, Mauritian Style: Cultural unity and ethnic diversity. Comparative Studies in Society and History , 36 (3), pp. 549--574
- (1995) What is happening to the concept of ethnicity? Razprave in Gradivo (Ljubljana), 29--30, pp. 165-177
Evans-Pritchard, E.E. (1940) The Nuer. Oxford: Clarendon Press.
Gellner, Ernest (1983) Nations and Nationalism. Oxford: Blackwell.
Hobsbawm, Eric (1992) Nations and Nationalism since 1780, 2nd edition. Cambridge: Cambridge University Press.
Hydén, Göran (1983) No Shortcuts to Progress: African Development Management in Perspective. London: Heinemann.
Lonsdale, John (1996) Moral ethnicity, ethnic nationalism and political tribalism: the case of the Kikuyu. Unpublished ms., University of Cambridge
Schütz, Alfred and Thomas Luckmann (1979) Strukturen der Lebenswelt. Band I. Frankfurt-am-Main: Suhrkamp.
Smith, Anthony D. (1986) The Ethnic Origins of Nations. Oxford: Blackwell.
- (1991) National Identity. Harmondsworth: Penguin.
- (1995) Nations and Nationalism in a Global Era. Cambridge: Polity.
Woolf, Virginia (1928) Orlando. London: Hogarth Press.