The Myth of Global Ethnic Conflict

John R. Bowen

Much recent discussion of international affairs has been based on the misleading assumption that the world is fraught with primordial ethnic conflict. According to this notion, ethnic groups lie in wait for one another, nourishing age-old hatreds and restrained only by powerful states. Remove the lid, and the cauldron boils over. Analysts who advance this idea differ in their predictions for the future: some see the fragmentation of the world into small tribal groups; others, a face-off among several vast civilizational coalitions. They all share, however, the idea that the world's current conflicts are fueled by age-old ethnic loyalties and cultural differences. ¹

This notion misrepresents the genesis of conflict and ignores the ability of diverse people to coexist. The very phrase "ethnic conflict" misguides us. It has become a shorthand way to speak about any and all violent confrontations between groups of people living in the same country. Some of these conflicts involve ethnic or cultural identity, but most are about getting more power, land, or other resources. They do not result from ethnic diversity; thinking that they do sends us off in pursuit of the wrong policies, tolerating rulers who incite riots and suppress ethnic differences.

In speaking about local group conflicts we tend to make three assumptions: first, that ethnic identities are ancient and unchanging; second, that these identities motivate people to persecute and kill; and third, that ethnic diversity itself inevitably leads to violence. All three are mistaken. [End Page 3]

Contrary to the first assumption, ethnicity is a product of modern politics. Although people have had identities--deriving from religion, birthplace, language, and so on--for as long as humans have had culture, they have begun to see themselves as members of vast ethnic groups, opposed to other such groups, only during the modern period of colonization and state-building.

The view that ethnicity is ancient and unchanging emerges these days in the potent images of the cauldron and the tribe. Out of the violence in Eastern Europe came images of the region as a bubbling cauldron of ethnonationalist sentiments that were sure to boil over unless suppressed by strong states. The cauldron image contrasts with the American "melting pot," suggesting that Western ethnicities may melt, but Eastern ones must be suppressed by the region's unlikable, but perhaps necessary, Titos and Stalins.

Nowhere does this notion seem more apt than in the former Yugoslavia. Surely the Serbs, Croats, and Bosnians are distinct ethnic groups destined to clash throughout history, are they not? Yet it is often forgotten how small the differences are among the currently warring factions in the Balkans. Serbs, Croats, and Bosnians all speak the same language (Italy has greater linguistic diversity) and have lived side by side, most often in peace, for centuries. Although it is common to say that they are separated by religion--Croats being Roman Catholic, Serbs Orthodox Christian, and Bosnians Muslim--in fact each population includes sizeable numbers of the other two religions. The three religions have indeed become symbols of group differences, but religious differences have not, by themselves, caused intergroup conflict. Rising rates of intermarriage (as high as 30 percent in Bosnia) would have led to the gradual blurring of contrasts across these lines.

As knowledgeable long-term observers such as Misha Glenny have pointed out, the roots of the current Balkan violence lie not in primordial ethnic and religious differences but rather in modern attempts to rally people around nationalist ideas. "Ethnicity" becomes "nationalism" when it includes aspirations to gain a monopoly of land, resources, and power. But nationalism, too, is a learned and frequently manipulated set of ideas, and not a primordial sentiment. In the nineteenth century, Serb and Croat intellectuals joined other Europeans in championing the rights of peoples to rule themselves in "nation-states": states to be composed of one nationality. For their part, Serbs drew on memories of short-lived Serb national states to claim their right to expand outward to encompass other peoples, just as other countries in Europe (most [End Page 4] notably France) had done earlier. That Balkan peoples spoke the same language made these expansionist claims all the more plausible to many Serbs. ²

At the same time, Croats were developing their own nationalist ideology, with a twist: rather than claiming the right to overrun non-Croats, it promised to exclude them. Nationalism among the Croats naturally was directed against their strong Serb neighbors. When Serbs dominated the state of Yugoslavia that was created after the First World War, Croat resentment of Serbs grew. The most militant of Croat nationalists formed an underground organization called Ustashe ("Uprising"), and it was this group, to which the Nazis gave control of Croatia, that carried out the forced conversions, expulsions, and massacres of Serbs during the Second World War. The later calls to war of the Serb leader Slobodan Milosevic worked upon the still fresh memories of these tragedies.

But the events of the Second World War did not automatically lead to the slaughters of the 1990s; wartime memories could have been overcome had Yugoslavia's new leaders set out to create the social basis for a multiethnic society. But Marshal Tito chose to preserve his rule by forbidding Yugoslavs from forming independent civic groups and developing a sense of shared political values. Political opposition, whether in Croatia, Serbia, or Slovenia, coalesced instead around the only available symbols, the nationalisms of each region. Tito further fanned nationalist flames by giving Serbs and Croats privileges in each other's territories--Serbs held positions of power in Croatia, and Croats in Belgrade. In the countryside these minority presences added to nationalist resentments. Tito's short-term political cleverness--nostalgically remembered by some in the West--in fact set the stage for later slaughter. Resentments and fears generated by modern state warfare and the absence of a civil society--not ethnic differences--made possible the success of the nationalist politicians Milosevic and Franjo Tudjman.

The Legacy of Colonialism

But what about Africa? Surely there we find raw ethnic conflict, do we not? Our understandings of African violence have been clouded by visions, not of boiling cauldrons, but of ancient tribal warfare. I recall a National Public Radio reporter interviewing an African UN official about Rwanda. Throughout the discussion the reporter pressed the official to discuss the "ancient tribal hatreds" that were fueling the slaughter. The official ever so politely demurred, repeatedly reminding the reporter that mass conflict began when Belgian colonial rulers gave Tutsis a monopoly of state power. But, as happens so often, the image of ancient tribalism was too deeply ingrained in the reporter's mind for him to hear the UN official's message. [End Page 5]

What the African official had to say was right: ethnic thinking in political life is a product of modern conflicts over power and resources, and not an ancient impediment to political modernity. True, before the modern era some Africans did consider themselves Hutu or Tutsi, Nuer or Zande, but these labels were not the main sources of everyday identity. A woman living in central Africa drew her identity from where she was born, from her lineage and in-laws, and from her wealth. Tribal or ethnic identity was rarely important in everyday life and could change as people moved over vast areas in pursuit of trade or new lands. Conflicts were more often within tribal categories than between them, as people fought over sources of water, farmland, or grazing rights.

It was the colonial powers, and the independent states succeeding them, which declared that each and every person had an "ethnic identity" that determined his or her place within the colony or the postcolonial system. Even such a seemingly small event as the taking of a census created the idea of a colony-wide ethnic category to which one belonged and had loyalties. (And this was not the case just in Africa: some historians of India attribute the birth of Hindu nationalism to the first British census, when people began to think of themselves as members of Hindu, Muslim, or Sikh populations.) The colonial powers--Belgians, Germans, French, British, and Dutch--also realized that, given their small numbers in their dominions, they could effectively govern and exploit only by seeking out "partners" from among local people, sometimes from minority or Christianized groups. But then the state had to separate its partners from all others, thereby creating firmly bounded "ethnic groups."

In Rwanda and Burundi, German and Belgian colonizers admired the taller people called Tutsis, who formed a small minority in both colonies. The Belgians gave the Tutsis privileged access to education and jobs, and even instituted a minimum height requirement for entrance to college. So that colonial officials could tell who was Tutsi, they required everyone to carry identity cards with tribal labels.

But people cannot be forced into the neat compartments that this requirement suggests. Many Tutsis are tall and many Hutus short, but Hutus and Tutsis had intermarried to a such an extent that they were not easily distinguished physically (nor are they today). They spoke the same language and carried out the same religious practices. In most regions of the colonies the categories became economic labels: poor Tutsis became Hutus, and economically successful Hutus became Tutsis. Where the labels "Hutu" and "Tutsi" had not been much used, lineages with lots of cattle were simply labeled Tutsi; poorer lineages, Hutu. Colonial discrimination against Hutus created what had not existed before: a sense of collective Hutu identity, a Hutu cause. In the late 1950s Hutus began to rebel against Tutsi rule (encouraged by Europeans on their way out) and then created an independent and Hutu-dominated [End Page 6] state in Rwanda; this state then gave rise to Tutsi resentments and to the creation of a Tutsi rebel army, the Rwandan Patriotic Front.

The logic of rule through ethnic division worked elsewhere, too. The case of Sri Lanka (formerly Ceylon) shows how, even when colonizers did not favor a single group, colonial rule could foster interethnic violence. The Sinhalese and Tamils of Sri Lanka have a common origin and, contrary to stereotypes of dark Tamils and light-skinned Sinhalese, they cannot easily be distinguished by their physical characteristics. The distinction between them is based mainly on the language spoken. Before this century there was little conflict between them; indeed, they did not think of themselves as two distinct kinds of people. Then came British rule. As they did throughout their empire, the British ruled Ceylon by creating an English-speaking elite, and, here as elsewhere, their favoritism engendered an opposition. In Ceylon this opposition took on racial and religious overtones. The majority of those who had been left out of the elite spoke Sinhalese and were Buddhists, and they began to promote a racist notion of Sinhalese superiority as an "Aryan race." After independence it was this Sinhalese-speaking group that gained control of the new state of Sri Lanka, and began to exclude Tamils from the best schools and jobs, mainly by requiring competence in Sinhalese. Not surprisingly, Tamils resented this discrimination, and some--initially only a few--launched violent protests in the 1970s. These riots led to massive state repression and, by a logic similar to that shaping Tutsi rebellions in Rwanda, to the creation of the Tamil Tigers (the Liberation Tigers of Tamil Eelam) and their demands for an autonomous Tamil region. As the anthropologist Stanley Tambiah has argued, the island's violence is a late-twentieth-century response to colonial and postcolonial policies that relied on a hardened and artificial notion of ethnic boundaries. ³

In these cases and many others--Sikhs in India, Maronites in Lebanon, Copts in Egypt, Moluccans in the Dutch East Indies, Karens in Burma--colonial and postcolonial states created new social groups and identified them by ethnic, religious, or regional categories. Only in living memory have the people who were sorted into these categories begun to act in concert, as political groups with common interests. Moreover, their shared interests have been those of political autonomy, access to education and jobs, and control of local resources. Far from reflecting ancient ethnic or tribal loyalties, their cohesion and action are products of the modern state's demand that people make themselves heard as powerful groups, or else risk suffering severe disadvantages.

Fear from the Top

A reader might say at this point: Fine, ethnic identities are modern and created, but today people surely do target members of other ethnic [End Page 7] groups for violence, do they not? The answer is: Less than we usually think, and when they do, it is only after a long period of being prepared, pushed, and threatened by leaders who control the army and the airwaves. It is fear and hate generated from the top, and not ethnic differences, that finally push people to commit acts of violence. People may come to fear or resent another group for a variety of reasons, especially when social and economic change seems to favor the other group. And yet such competition and resentment "at the ground level" usually does not lead to intergroup violence without an intervening push from the top.

Let us return to those two most unsettling cases, Rwanda and the Balkans. In Rwanda the continuing slaughter of the past few years stemmed from efforts by the dictator-president Juvenal Habyarimana to wipe out his political opposition, Hutu as well as Tutsi. In 1990-91 Habyarimana began to assemble armed gangs into a militia called Interahamwe. This militia carried out its first massacre of a village in March 1992, and in 1993 began systematically to kill Hutu moderates and Tutsis. Throughout 1993 the country's three major radio stations were broadcasting messages of hate against Tutsis, against the opposition parties, and against specific politicians, setting the stage for what followed. Immediately after the still unexplained plane crash that killed President Habyarimana in April 1994, the presidential guard began killing Hutu opposition leaders, human rights activists, journalists, and others critical of the state, most of them Hutus. Only then, after the first wave of killings, were the militia and soldiers sent out to organize mass killings in the countryside, focusing on Tutsis.

Why did people obey the orders to kill? Incessant radio broadcasts over the previous year had surely prepared them for it; the broadcasts portrayed the Tutsi-led Rwandan Patriotic Front as bloodthirsty killers. During the massacres, radio broadcasts promised the land of the dead to the killers. Town mayors, the militia, the regular army, and the police organized Hutus into killing squads, and killed those Hutus who would not join in. The acting president toured the country to thank those villagers who had taken part in the massacres. Some people settled personal scores under cover of the massacre, and many were carried away with what observers have described as a "killing frenzy." The killings of 1994 were not random mob violence, although they were influenced by mob psychology. ⁴

In reading accounts of the Rwanda killings, I was struck by how closely they matched, point by point, the ways Indonesians have described to me their participation in the mass slaughters of 1965-66. [End Page 8] In Indonesia the supposed target was "communists," but there, too, it was a desire to settle personal scores, greed, willingness to follow the army's orders, and fear of retaliation that drove people to do things they can even now barely admit to themselves, even though many of them, like many Hutus, were convinced that the killings stopped the takeover of the country by an evil power. In both countries, people were told to kill the children and not to spare pregnant women, lest children grow up to take revenge on their killers. Americans continue to refer to those massacres in Indonesia as an instance of "ethnic violence" and to assume that Chinese residents were major targets, but they were not: the killings by and large pitted Javanese against Javanese, Acehnese against Acehnese, and so forth.

The two massacres have their differences: Rwanda in 1993-94 was a one-party state that had carried out mass indoctrination through absolute control of the mass media; Indonesia in 1965-66 was a politically fragmented state in which certain factions of the armed forces only gradually took control. But in both cases leaders were able to carry out a plan, conceived at the top, to wipe out an opposition group. They succeeded because they persuaded people that they could survive only by killing those who were, or could become, their killers.

The same task of persuasion faced Serb and Croat nationalist politicians, in particular Croatia's Franjo Tudjman and Serbia's Slobodan Milosevic, who warned their ethnic brethren elsewhere--Serbs in Croatia, Croats in Bosnia--that their rights were about to be trampled unless they rebelled. Milosevic played on the modern Serb nationalist rhetoric of expansion, claiming the right of Serbs everywhere to be united. Tudjman, for his part, used modern Croat rhetoric of exclusionary nationalism to build his following. Once in power in Croatia, he moved quickly to define Serbs as second-class citizens, fired Serbs from the police and military, and placed the red-and-white "checkerboard" of the Nazi-era Ustashe flag in the new Croatian banner.

Both leaders used historical memories for their own purposes, but they also had to erase recent memories of new Yugoslav identities, tentatively forged by men and women who married across ethnic boundaries or who lived in the cosmopolitan cities. The new constitutions recognized only ethnic identity, not civil identity, and people were forced, sometimes at gunpoint, to choose who they "really" were. ⁵

Contrary to the "explanations" of the war frequently offered by Western journalists, ordinary Serbs do not live in the fourteenth century, fuming over the Battle of Kosovo; nor is the current fighting merely a playing out of some kind of inevitable logic of the past, as some have written. It took hard work by unscrupulous politicians to convince ordinary people that the other side consisted not of the friends and neighbors they had known for years but of genocidal people who would kill them if they were not killed first. For Milosevic this meant [End Page 9] persuading Serbs that Croats were all crypto-Nazi Ustashe; for Tudjman it meant convincing Croats that Serbs were all Chetnik assassins. Both, but particularly Milosevic, declared Bosnian Muslims to be the front wave of a new Islamic threat. Each government indirectly helped the other: Milosevic's expansionist talk confirmed Croat fears that Serbs intended to control the Balkans; Tudjman's politics revived Serbs' still remembered fears of the Ustashe. Serb media played up these fears, giving extensive coverage in 1990-91 to the exhumation of mass graves from the Second World War and to stories of Ustashe terror. This "nationalism from the top down," as Warren Zimmerman, the last U.S. ambassador to Yugoslavia, has characterized it, was also a battle of nationalisms, with each side's actions confirming the other's fears.

If Rwanda and the Balkans do not conform to the images of bubbling cauldrons and ancient tribal hatreds, even less do other ongoing local-level conflicts. Most are drives for political autonomy, most spectacularly in the former Soviet Union, where the collapse of Soviet power allowed long-suppressed peoples to reassert their claims to practice their own languages and religions, and to control their own territory and resources--a rejection of foreign rule much like anti-imperial rebellions in the Americas, Europe, Asia, and Africa. Elsewhere various rebellions, each with its own history and motivations, have typically--and erroneously--been lumped together as "ethnic conflict." Resistance in East Timor to Indonesian control is a 20-year struggle against invasion by a foreign power, not an expression of ethnic or cultural identity. People fighting in the southern Philippines under the banner of a "Moro nation" by and large joined up to regain control of their homelands from Manila-appointed politicians. Zapatista rebels in Chiapas demand jobs, political reform, and, above all, land. They do not mention issues of ethnic or cultural identity in their statements--indeed, their leader is from northern Mexico and until recently spoke no Mayan. Other current conflicts are raw struggles for power among rival factions, particularly in several African countries (Liberia, Somalia, Angola) where rival forces often recruit heavily from one region or clan (giving rise to the notion that these are "ethnic conflicts") in order to make use of local leaders and loyalties to control their followers. ⁶

Ethnic Diversity and Social Conflict

This brings us to the third mistaken assumption: that ethnic diversity brings with it political instability and the likelihood of violence. To the contrary, greater ethnic diversity is not associated with greater interethnic conflict. Some of the world's most ethnically diverse states, such as Indonesia, Malaysia, and Pakistan, though not without internal conflict and political repression, have suffered little interethnic violence, while countries with very slight differences in language or culture (the former [End Page 10] Yugoslavia, Somalia, Rwanda) have had the bloodiest such conflicts. It is the number of ethnic groups and their relationships to power, not diversity per se, that strongly affect political stability. As shown in recent studies by political scientist Ted Gurr, and contrary to popular thinking, local conflicts have not sharply increased in frequency or severity during the last ten years. The greatest increase in local conflicts occurred during the Cold War, and resulted from the superpowers' efforts to arm their client states. (The sense that everything exploded after 1989, Gurr argues, comes from the reassertions of national identity in Eastern Europe and the former Soviet Union.) ⁷

By and large, the news media focus on countries racked by violence and ignore the many more cases of peaceful relations among different peoples. Take Indonesia, where I have carried out fieldwork since the late 1970s. If people know of Indonesia, it is probably because of its occupation of East Timor and its suppression of political freedoms. But these are not matters of ethnic conflict, of which there is remarkably little in a country composed of more than three hundred peoples, each with its own distinct language and culture. Although throughout the 1950s and 1960s there were rebellions against Jakarta in many parts of the country, these concerned control over local resources, schooling, and religion. An on-again, off-again rebellion where I work, on the northern tip of Sumatra, has been about control over the region's vast oil and gas resources (although the Western press continues to stereotype it as "ethnic conflict").

Cultural diversity does, of course, present challenges to national integration and social peace. Why do some countries succeed at meeting those challenges while others fail? Two sets of reasons seem most important, and they swamp the mere fact of ethnic and cultural diversity.

First there are the "raw materials" for social peace that countries possess at the time of independence. Countries in which one group has been exploiting all others (such as Rwanda and Burundi) start off with scores to settle, while countries with no such clearly dominating group (such as Indonesia) have an initial advantage in building political consensus. So-called centralized polities, with two or three large groups that continually polarize national politics, are less stable than "dispersed" systems, in which each of many smaller groups is forced to seek out allies to achieve its goals. And if the major ethnic groups share a language or religion, or if they have worked together in a revolutionary struggle, they have a bridge already in place that they can use to build political cooperation. ⁸

Take, again, the case of Indonesia. In colonial Indonesia (the Dutch East Indies) the Javanese were, as they are today, the most numerous people. But they were concentrated on Java and held positions of power only there. Peoples of Java, Sumatra, and the eastern islands, along with Malays and many in the southern Philippines, had used Malay as a [End Page 11] lingua franca for centuries, and Malay became the basis for the language of independent Indonesia. Islam also cut across regions and ethnicities, uniting people on Sumatra, Java, and Sulawesi. Dominance was "dispersed," in that prominent figures in literature, religion, and the nationalist movement tended as often as not to be from someplace other than Java, notably Sumatra. Moreover, people from throughout the country had spent five years fighting Dutch efforts to regain control after the Second World War, and could draw on the shared experience of that common struggle. ⁹

One can see the difference each of these features makes by looking next door at culturally similar Malaysia. Malays and Chinese, the largest ethnic groups, shared neither language nor religion, and had no shared memory of struggle to draw on. Malays had held all political power during British rule. On the eve of independence there was a clear fault line running between the Malay and Chinese communities.

The Importance of Political Choices

But these initial conditions do not tell the whole story, and here enters the second set of reasons for social peace or social conflict. States do make choices, particularly about political processes, that ease or exacerbate intergroup tensions. As political scientist Donald Horowitz has pointed out, if we consider only their starting conditions, Malaysia ought to have experienced considerable interethnic violence (for the reasons given above), whereas Sri Lanka, where Tamils and Sinhalese had mingled in the British-trained elite, should have been spared such violence. And yet Malaysia has largely managed to avoid it while Sri Lanka has not. The crucial difference, writes Horowitz, was in the emerging political systems in the two countries. Malaysian politicians constructed a multiethnic political coalition, which fostered ties between Chinese and Malay leaders and forced political candidates to seek the large middle electoral ground. In Sri Lanka, as we saw earlier, Sinhalese-speakers formed a chauvinist nationalist movement, and after early cooperation Tamils and Sinhalese split apart to form ethnically based political parties. Extreme factions appeared on the wings of each party, forcing party leaders to drift in their directions.

But political systems can be changed. Nigeria is a good example. Prior to 1967 it consisted of three regions--North, South, and East--each with its own party supported by ethnic allegiances. The intensity of this three-way division drove the southeast region of Biafra to attempt to break away from Nigeria in 1967, and the trauma of the civil war that followed led politicians to try a new system. They carved the country into 19 states, the boundaries of which cut through the territories of the three largest ethnic groups (Hausa, Yoruba, and Igbo), encouraging a new federalist politics based on multiethnic coalitions. The [End Page 12] new system, for all its other problems, prevented another Biafra. Subsequent leaders, however, continued to add to the number of states for their own political reasons. The current leader, General Sani Abacha, is now adding to an already expanded list of 30 states; this excessive fragmentation has broken up the multiethnic coalitions and encouraged ethnic politics anew. A similar direction has been pursued by Kenya's Daniel arap Moi, who has created an ethnic electoral base that excludes most Kikuyus, increasing the relevance of ethnicity in politics and therefore the level of intergroup tensions.

What the myth of ethnic conflict would say are ever-present tensions are in fact the products of political choices. Negative stereotyping, fear of another group, killing lest one be killed--these are the doings of so-called leaders, and can be undone by them as well. Believing otherwise, and assuming that such conflicts are the natural consequences of human depravity in some quarters of the world, leads to perverse thinking and perverse policy. It makes violence seem characteristic of a people or region, rather than the consequence of specific political acts. Thinking this way excuses inaction, as when U.S. president Bill Clinton, seeking to retreat from the hard-line Balkan policy of candidate Clinton, began to claim that Bosnians and Serbs were killing each other because of their ethnic and religious differences. Because it paints all sides as less rational and less modern (more tribal, more ethnic) than "we" are, it makes it easier to tolerate their suffering. Because it assumes that "those people" would naturally follow their leaders' call to kill, it distracts us from the central and difficult question of just how and why people are sometimes led to commit such horrifying deeds.

John R. Bowen is professor of anthropology and chair of the Program in Social Thought and Analysis at Washington University in St. Louis. His most recent books are Muslims Through Discourse: Religion and Ritual in Gayo Society (1993) and a forthcoming coedited book on comparisons in political science and anthropology. He is now completing a study of changes in religious and civil law in Indonesia.

Notes

1. Two of the most widely read proponents of the view I am contesting areRobert Kaplan, in his dispatches for The Atlantic and in his Balkan Ghosts: A Journey Through History (New York: St. Martin's, 1993), and (writing mainly on large-scale conflict) Samuel P. Huntington, in "The Clash of Civilizations?" Foreign Affairs 72 (Summer 1993): 22-49. My concern is less with the particular difficulties of these writers' arguments, about which others have written, than with the general notion, which, as with all myths, survives the death of any one of its versions.

2. See Misha Glenny, The Fall of Yugoslavia (New York: Penguin, 1992), for a balanced and ethnographically rich account of the Balkan wars. On recent tendencies in European nationalisms, see especially Rogers Brubaker, Nationalism Reframed: Nationhood and the National Question in the New Europe (Cambridge: Cambridge University Press, 1996). Brubaker makes the important point that "nationalism" should be treated as a category of social and political ideology, and not a pre-ideological "thing."

3. S.J. Tambiah, Sri Lanka: Ethnic Fratricide and the Dismantling of Democracy (Chicago: University of Chicago Press, 1986). For a different view on the culture of violence in Sri Lanka, see Bruce Kapferer, Legends of People, Myths of State (Washington, D.C.: Smithsonian Institution Press, 1988).

4. Among recent overviews of massacres in Rwanda and Burundi, see Philip Gourevitch, "The Poisoned Country," New York Review of Books, 6 June 1996, 58-64, and René Lemarchand, Burundi: Ethnic Conflict and Genocide (Cambridge: Cambridge University Press, 1995).

5. That there were memories, fears, and hatreds to exploit is important to bear in mind, lest we go to the other extreme and argue that these conflicts are entirely produced from the top, an extreme toward which an overreliance on rational-choice models may lead some analysts. Russell Hardin's otherwise excellent One for All: The Logic of Group Conflict (Princeton: Princeton University Press, 1995) errs, I believe, in attributing nothing but rational, self-aggrandizing motives to those leaders who stir up ethnic passions, ignoring that they, too, can be caught up in these passions. The cold rationality of leaders is itself a variable: probably Milo sevi c fits Hardin's rational-actor model better than Tudjman, and Suharto better than Sukarno. In each case, it is an empirical question.

6. The same points could be made concerning the religious version of"ancient hatreds," such as Muslim-Hindu relations in India. However peaceful or conflictual "ancient" relations may have been (and on this issue there continues to be a great deal of controversy among historians of South Asia), the often bloody conflicts of the past ten years in India have been fueled by ambitious politicians who have seen boundless electoral opportunity in middle-class Hindu resentment toward 1) lower castes' claims that they deserve employment and education quotas, and 2) the recent prosperity of some middle-class Muslims. See the penetrating political analyses by Susanne Hoeber Rudolph and Lloyd I. Rudolph in the New Republic, 22 March 1993 and 14 February 1994, and a historical and ethnographic study by Peter van der Veer, Religious Nationalism: Hindus and Muslims in India (Berkeley: University of California Press, 1994).

7. See Ted Gurr, Ethnic Conflict in World Politics (Boulder, Colo.: Westview, 1994).

8. See Donald L. Horowitz, Ethnic Groups in Conflict (Berkeley: University of California Press, 1985), 291-364.

9. I would propose "dispersed dominance" (a situation in which each of several groups considers itself to dominate on some social or political dimension) as a second important mechanism for reducing intergroup conflict alongside the well-known "cross-cutting cleavages" (a situation in which one or more important dimensions of diversity cut across others, as religion cuts across ethnicity in many countries). "Dispersed dominance" takes into account social and cultural dimensions, such as literary preeminence or a sense of social worth stemming from putative indigenous status. It is thus broader than, but similar to, political mechanisms such as federalism, when these mechanisms are aimed at (in Donald Horowitz's phrase) "proliferating the points of power." It is the empirical correlate to the normative position articulated by Michael Walzer in Spheres of Justice (New York: Basic Books, 1983) that dominance in one sphere (or dimension) ought not to automatically confer dominance in others.

[an error occurred while processing this directive].