Robert Bates, in turn, explains the persistence of ethnic groups in Africa by "their capacity to extract goods and services from the modern sector and thereby satisfy the demands of their members" (1983, 161). With resources such as land, state allocations, and high governmental positions in scarce supply and highly valued by all communal interests, ethnic membership is viewed as a means of maximizing the ability of individuals and groups to compete. Social interactions reinforce ethnic identities, carrying them beyond the purely material realm and giving them meaning in a wider range of relations. In particular, ethnic groups tend to possess strong norms of exclusion that override more diffuse universalistic norms, thus reinforcing group solidarity and promoting extremism (Hardin 1995, 101 and 140-1). As individuals interact with others in their social environment, ethnic groups thus have a strong tendency to form and become politically salient.

The centripetal forces that drive groups together, however, do not necessarily lead to the polarization of the larger society. Ethnic identities and even vibrant ethnic organizations can coexist with a wide range of other, potentially cross-cutting identities and organizations. Two catalysts--ethnic activists and political entrepreneurs are necessary to produce polarization.

Timur Kuran (1995) examines the role of ethnic activists in the process of what he calls ethnic dissimilation. Recognizing that all individuals desire to belong to groups, but that the strength of this desire can differ, Kuran argues that individuals with especially strong needs to identify with ethnic kin can persuade directly or induce indirectly through their own behavior others to increase their public ethnic activity. Over time, ethnic activists can lead individuals to increase their preferences for ethnic activity or their manifest ethnic behaviors in order to maintain standing within the group. For example, as associates make more frequent and visible displays of consuming ethnic foods, individuals must likewise increase their consumption (even if they do not especially like the foods) to remain part of the group. Such increases, moreover, can be self-reinforcing, thereby creating further increases in identifiably ethnic behaviors. In this way, ethnic activists and the social pressures they foster can cause previously integrated communities to separate along ethnic lines. As individuals can be driven by social pressures to represent falsely their true preferences, and there may be multiple equilibria, Kuran demonstrates that the process of dissimilation can be rapid, apparently spontaneous, and essentially unpredictable.

Political entrepreneurs both reflect the polarization of societies and, through their actions, propel this process further. As Stephen Saideman (1995) notes, ethnicity often provides a key marker for self-aggrandizing politicians seeking to build constituencies for attaining or keeping political power. As an identifiable (if not "fixed") characteristic, ethnicity allows for selective benefits to be targeted to specific communities--and for politicians representing those communities to claim credit for delivering the goods; at the same time, ethnic cleavages allow political entrepreneurs to mobilize grievances against distributions of benefits that are perceived to be unfavorable to the group. Thus, while ethnicity is certainly not the only political marker, it is a highly visible and easily used vehicle for political mobilization.

Politicians in the middle of the political spectrum or those who court ethnically heterogeneous constituencies are often vulnerable to extremists and others seeking to draw support from only a narrower but more ethnically homogenous constituency. When faced with extremist challenges or the threat of such challenges, even centrist politicians can be driven to embrace a more "ethnic" position and defend more vigorously communal interests a phenomenon often referred to as ethnic outbidding (Rothschild 1981; Horowitz 1985). The smaller the constituency willing to support a universalistic program, the more likely politicians will be drawn toward to the extremes.

Political entrepreneurs seeking power based on ethnic appeals also reinforce social polarization. Political entrepreneurs are often indistinguishable from ethnic activists. Like activists, they can highlight and legitimate ethnic associations and affinities. By doing so, they raise the political saliency of ethnicity and increase the likelihood that members of their ethnic group will support them rather than more centrist politicians. In framing issues for the public, moreover, extremists will exaggerate the hostility of others and magnify the likelihood of conflict--thereby distorting public debate and images of other groups and driving coethnics toward them for power and support. Milosevic's control over the media in Serbia, for instance, allowed him to present a one-sided view of Croat violence toward Croatian Serbs (Weingast 1995, 20). In short, political entrepreneurs stimulate ethnic fears for their own aggrandizement.

The polarization of society, driven by ethnic activists and political entrepreneurs, is magnified by political memories and myths, on the one hand, and emotions, on the other. Political memories and myths can also lead groups to form distorted images of others and, especially, to see others as more hostile and aggressive than they really are. Such memories and myths are often rooted in actual events--and probably could not be long sustained absent such historical roots. At the same time, events can be distorted, evolving into legends that justify the superiority of one group over another, desires for retribution for perceived grievances, or group hatreds.

Following decolonization in Africa, for instance, political memories of past conflict directly contributed to violent encounters, even instances of "selective genocide" (Lemarchand and Martin 1974). Imperial repression created communications gaps between rulers and ruled; it also allowed imperial officials great latitude in allocating fiscal resources and recruiting imperial adjuncts among the local population. Over time, however, these intentional and unintentional acts of imperial ethnic preference spawned hurts and angers toward minority identity groups perceived as having close working relationships with the colonizers. With independence, the resulting perceptions of comparative disadvantage contributed to a spiral of fear and aggressive behavior, which grew precipitously whenever the stereotypic images of other groups were supported by actual events. Thus, substantive competition over land and other resources combined with symbolic hurts from past humiliations and denials of group status (for example, among the Hutu in Rwanda) to contribute to highly destructive outcomes. With the rough hand of the imperial buffer removed, centralized bureaucratic and military state power no longer kept ethnic adversaries at bay and violent encounters ensued.

In Eastern Europe, political memories and myths have both defined the groups themselves and stimulated acute fears of mutual exploitation. The Croats and Serbs, for instance, formerly conationals and now enemies, have both used history and religion to lump each other into tight ethnic blocs determined on a destructive course and therefore deserving of pitiless retaliation (Glenny 1992, 85). In cultivating the enemy image, leaders in the former Yugoslavia and elsewhere not only stereotype and express their hostility toward their opponents, but they also force the appearance of conformity among their group members. Such an insulation of an organized body of people from complex reality can be a harbinger of impending chaos.

Emotions may also cause individuals and groups to act in exaggerated or potentially "irrational" ways that magnify the chances of conflict. We are extremely suspicious of emotions as explanations of conflict, at least at a first stage. Many analysts leap prematurely to the conclusion that ethnic conflict--because it appears so counterproductive and so vicious--must be irrational by nature (see Connor 1994). In our view, many aspects of ethnic conflict can be understood as the perhaps unfortunate but nonetheless rational outcomes of group interactions. However, we would be remiss to ignore such emotions as hostility and alienation as possible sources and catalysts of ethnic conflict.

Many analysts point to a deep psychological--perhaps even physiological--need for humans to belong to a group (Horowitz 1985). Part of this is a need to distinguish between "us" and "them" as individuals search for belonging and security. This need underlies Kuran's model of ethnic dissimilation. In the process of drawing distinctions, however, individuals often overstate the goodness of their own group while simultaneously vilifying others. Where such emotional biases exist, groups are likely to interpret the demands of others as outrageous, while seeing their own as moderate and reasonable; to view the other as inherently untrustworthy and likely to defect from any ethnic contract, while believing themselves to be reliable; to insist upon adequate safeguards against the possible defection of the other, but interpreting the efforts of others to impose similar restrictions on them as a sign of "bad faith;" to believe that the other is withholding information or being purposively deceptive, while they are being open and honest, and so on. Emotions magnify both group solidarity and intergroup tensions (Van Evera 1994).

Politics under conditions of extreme scarcity also contribute to a win/lose mentality where ethnic representatives seek favorable inclusion, even domination, in order to avoid the risks of marginalization. In such a context, political competition and conflict act as magnifiers of a people's uncertainty about its future. Individuals understandably fear the consequences of modernization and the application of programs of structural adjustment, anticipating the loss of jobs and status, and the need for massive readjustments in terms of new values, outlooks and orientations (Rothchild and Groth 1995, 74-75). Under such circumstances, ethnic identities are more likely to become suffused with belligerent stereotypes, as hostility toward ethnic adversaries, fanned by the mass media, provide an outlet for exaggerated fears and suspicions. ₁₀

The emotional power of ethnic attachments is typically increased by the unifying effects of what are perceived to be external threats. People who have little in common with others may unite when they feel threatened by external enemies. Thus, the shared identity of the Hutu in Burundi emerged only recently with the Tutsi repressions of 1972. An external threat of aggression led the organizationally distinct Hutu of the north-center to join forces with those in the south-Imbo, something that Warren Weinstein has described as "enforced ethnicity" (Weinstein 1972, 27). Similarly, in Chechnya, when very disparate interests felt threatened by Russian power, they overcame their difference and made common cause in the face of Russian intervention. This emotional pull may be cultivated by elites. After examining elementary school textbooks in former Yugoslavia, one analyst concludes that "Not a single act of heroism, or personal valor and death is mentioned for the sake of achieving freedom within the community. Only harmony is required within so as to facilitate defending the community from the enemy" (Pesic 1994, 77). Much like the "rally round the flag" effect that takes place within states threatened by external aggression, ethnic leaders can mobilize their members against threats posed by other ethnic groups. Such mobilization creates cohesion against internal group "traitors," national minorities (such as Russians living in the Ukraine), and external state and ethnic enemies, and results in greatly strengthened collective capacities for good or evil (Brubaker 1995).

Where strategic interactions between groups call for external actors to mediate and provide information to the groups and, possibly, create credible guarantees of new ethnic contracts, strategic interactions within groups require changing the incentives of the groups themselves and, especially, the ethnic activists and political entrepreneurs who lead them. As discussed below, targeted interventions are necessary to decrease the social and political salience of ethnicity and prevent the polarization of society. Such interventions are best taken early. Once the society has become polarized, there is little--at least in the short term--that outsides can do to reintegrate the polity.

Because conflict can escalate dangerously when leaders and constituents become entrapped in a situation where political memories and myths and emotions create menacing intergroup perceptions, it is also important that third parties take initiatives to clear up misperceptions and correct (or at least offset) emotionally generated fears and biases. Preferably, this should also occur at an early stage. At times, as in the 1971--72 negotiations in the Sudan (Rothchild and Hartzell 1993), external intermediaries can be instrumental in encouraging rival parties to understand and empathize with each other's feelings and predicaments. They can help influence groups to see themselves as others see them, and to view others as they would want to be viewed. Provided such third party actors are perceived as fairminded and detached providers of information, they can often assist groups to recognize the distortions in the information they receive and correct for these distortions in evaluating the attitudes and emotions of other groups.

Together, strategic interactions between and within groups can produce environments of fear in which ethnic tensions and conflicts can grow. As Pesic recognizes, it is the future that threatens, but it is often interpreted through the past. While each strategic dilemma discussed above is sufficient to produce and explain the outbreak of ethnic conflict, they almost always occur simultaneously.

Ethnic activists and political entrepreneurs polarize societies, exacerbating strategic dilemmas. The tendency toward polarization, in turn, is magnified by political memories and myths and emotions. Combined, these forces create a devastating brew of ethnic rivalry and violence.

II. The International Spread of Ethnic Conflict

As noted in the Introduction, the fear that disputes might spread across state borders accounts for much of the increase in scholarly and policy interest in ethnic conflict today. Our discussion so far has focused on the origins of ethnic conflict within states. These origins, however, provide essential building blocks for understanding the international spread of ethnic conflict. Unfortunately, the magnitude of problem and the processes through which ethnic conflict spreads remain poorly understood. Our analysis here is more tentative than that above, but provides an approach to conceptualizing and studying the problem. Understanding how and why ethnic conflict spreads remains an important research frontier.

Ethnic conflict spreads across state borders in two ways. Diffusion occurs when an ethnic conflict in one state increases the probability of conflict in a second. In other words, if conflict in Bosnia incites similar violence directly or indirectly in the Soviet successor states, the conflict will have diffused. Escalation occurs when a conflict in one country brings in new, foreign belligerents--whether neighbors or great powers with global reach. If Greece or Turkey were to become embroiled in the current Balkan wars, for example, the conflict will have escalated. ₁₁

Our focus on the international spread of ethnic conflict is not meant to imply that every such episode has an international dimension. Both scholars and policy makers need to recognize and respect the autonomy of particular ethnic conflicts from international pressures (Smith 1993, 28). It is also important to recognize that ethnic conflict is inherently self-limiting (see Fearon 1995b). Ethnic conflicts differ from ideological and, possibly, religious conflicts in that ethnic groups are by definition limited, while the latter principles are more nearly universalistic. Ethnic conflicts may still spread beyond the original kin groups, but they are unlikely to produce global conflagrations unless they become linked with other issues (Halperin 1995). Nonetheless, the international spread of ethnic conflict--even within limited, regional contexts--is a growing source of concern.

Diffusion

The papers in this project reach different conclusions on the extent to which ethnic conflicts diffuse between states. Focusing on the aggregate level, Kuran (1995) and Stuart Hill, Donald Rothchild, and Colin Cameron (1995) develop a model and provide evidence, respectively, of how politicized ethnicity and the tactics of mass conflict spread to group members and their leaders in other countries who face similar political conditions. In studying specific cases, on the other hand, Fearon (1995b) and Saideman (1995) locate the sources of contemporary ethnic conflicts primarily within states and question whether conflicts abroad play a significant role in precipitating violence. This disparity in views underlines not a contradiction but an important insight. The seeds of ethnic conflict, while possibly blown in from abroad, germinate and take root only in fertile soil. Unless the local conditions are right--or, perhaps more appropriately, wrong--diffusion is unlikely; but when circumstances are receptive, ethnic conflict can take root and become devastating.

Building upon the causes of ethnic conflict discussed above, diffusion can occur in four ways. These four processes are not necessarily exclusive and all may occur simultaneously. First, events abroad may change directly the ethnic balance of power at home, disrupting the existing ethnic contract and precipitating conflict. Through this first route, ethnic conflict may actually be contagious in the full sense of this overused term. For instance, refugee flows from a neighboring state may substantially alter the state's own ethnic composition (Newland 1993). Armed insurgents from one state may seek refuge in a second and stir up local conflicts in their wake. These are constant concerns for many African states (Keller 1995). Similar changes in the ethnic balance of power can occur in the breakup of federal states, even without the actual migration of peoples across recognized international borders. Once central political authority in Yugoslavia began to unravel, the relevant ethnic balance of power shifted from the federal level to the now independent republics. As this shift occurred, minority groups--previously protected by their kin in other regions--were left exposed and vulnerable. This emboldened the new majority and threatened the new minority in each state, undermining the ethnic contract and leading the groups into a spiral of violence (see Fearon 1993, Hardin 1995, 156 63, and Djilas 1995). In this way, Slovenia's relatively minor conflict with Serbia diffused to the other republics, and became more virulent with each additional succession. ₁₂

Second, ethnic conflict in one country may prompt groups in another to make more extreme demands. Groups in one state, witnessing ethnic mobilization or, more importantly, political success by ethnic groups in another, may increase their own political agitation and demand a significantly greater share of the resource pie--increasing the probability of conflict. Kuran (1995, 24) develops a strong argument on the importance of this "demonstration effect" in stimulating ethnic dissimilation abroad. Similarly, ethnic conflict abroad may cause groups to update their beliefs about the likely demands of other groups in their own country. Even in the absence of any change in the underlying political power of groups or in the claims made, if the groups believe others are now more likely to challenge the existing ethnic contract and issue greater demands, their best response may be to strike preemptively before the others have actually increased their levels of mobilization. Thus, changes in beliefs about the likely behaviors of others can precipitate conflict even in the absence of any manifest demands or actions. As groups update their beliefs about one another by observing events elsewhere, ethnic conflict can literally materialize out of thin air. Islamic fundamentalism appears to have stimulated greater concerns in France about its large Algerian minority and, especially, fears that the latter is likely to make appeals for greater autonomy and a more favorable distribution of resources; while such concerns have stimulated sporadic violence by French rightists, the conflict has not reached to date the level of widespread violence.

Third, and in ways similar to those just discussed, ethnic conflict abroad may lead groups to update their beliefs about the efficacy of the political safeguards contained in their existing ethnic contracts. For example, if events abroad suggest that the economic leverage wielded by wealthy minority groups is less effective than previously believed, the poorer majority may become emboldened and the minority threatened--again, precipitating conflict without any manifest changes in the underlying conditions at home. In this context, Serbia's suppression of agitation for elevation to republic status in Kosovo in 1981 provided an important signal to other groups in Yugoslavia about the efficacy of existing federal safeguards. Hardin (1995, 157) dates the unraveling of Yugoslavia to this event.

Finally, ethnic conflict abroad may lead groups to update their beliefs about the costs of protest or, ultimately, violence and their probability of success. Effective protest or violence abroad may lead groups at home to believe that they too may be able to obtain valued ends through coercion. Hill, Rothchild, and Cameron (1995) provide strong evidence for the diffusion of political tactics from one country to another. Slovenia's relatively easy break from Yugoslavia, precipitating a ten-day war between the Yugoslav National Army (JNA) and separatist forces with fewer than 70 casualties, gave "the impression that the dissolution of a country was not so difficult after all" (Woodward 1995, 146). Similarly, if groups believe that violence will provoke the international community to insist on the punishment of ethnic aggressors, but events suggest that the international community is unlikely to impose such punishments, groups will then lower their estimated costs of using violence and become more likely to use force. For instance, Haiti's military leaders inferred from America's precipitant disengagement from Somalia, as well as its earlier pullout from Lebanon, that the United States lacked the will to absorb significant costs in an internal conflict and could be forced to back down--an inference that was confirmed when a small group of demonstrators on the docks prompted the first American landing party to withdraw. The various ethnic groups in Bosnia drew similar inferences, and reasonably doubted early attempts by the United States and NATO to intervene in the conflict; this may have led to the eventual deployment in Bosnia of a larger than otherwise necessary force to demonstrate the commitment of the United States.

In these four ways, then, ethnic conflict in one country may precipitate similar conflict in another. Whether events have this effect, however, depends upon local conditions, the initial beliefs of groups on the scene, and the lessons drawn by these groups. For instance, adversaries may believe that resorting to violence is so costly that even substantial changes in these beliefs will still not produce manifest conflict. Conversely, the distribution of ethnic power or the beliefs of the groups about this distribution may be such that violence is inevitable; if so, events abroad may appear to cause the outbreak of conflict--and may in fact be a contributing factor--but conditions at home are the real driving forces. Identifying how conflicts diffuse requires a model of ethnic relations (such as the one posed in the previous section), estimates of the variables in this model, and close attention to how events abroad change these estimates and, especially, the beliefs of the groups.

It is important to note that all of the processes discussed above can both increase and decrease the likelihood of conflict. Successful conflict management abroad can reduce the direct spillovers and lead groups to temper their own demands, reduce their expectations of the likely demands of others, have greater confidence in the safeguards in their existing ethnic contracts, and recognize the high costs of violence. For example, South African whites drew positive lessons from the earlier experiences of the minority communities in Kenya, Namibia, and Zimbabwe after the transfer of power to African-led majority regimes, thereby facilitating the recent transition to one person, one vote elections and a modified form of majority rule in that country. Even in the conflict-prone 1990s, the lessons drawn need not be one-sided. The peaceful transition in South Africa, the emerging peace between Israel and its neighbors in the Middle East, and the peaceful divorce between the Czech Republic and Slovakia are shining examples of progress toward stable ethnic relations that may offset, in part, the harsh events that have occurred in Bosnia, Rwanda, and other recent tragedies. At the very least, some of the warning signs from the 1990s are more ambiguous than a focus on the overt conflicts might suggest.

There is, however, reason to expect that conflict may diffuse more readily than peace. The beliefs of groups are central to the outbreak of ethnic conflict within countries and the diffusion of ethnic conflict across countries. Information shapes these beliefs, and today flows mostly to groups from the international media. The media, in turn, contains within it an important selection bias. Conflict occurs in those countries in which the underlying conditions are most ripe--the balance of ethnic power is precarious, the demands made by each side are large, and the costs of conflict are small. Almost a truism, we--and other peoples around the globe--observe conflict where it is most likely. International news reports, which provide the raw material for the conclusions drawn by ethnic groups everywhere, are heavily biased toward conflict. The evening news does not feature balanced reports of deadly conflicts, on the one hand, and conflicts that did not happen or that were successfully nipped in the bud, on the other. This selection bias thus distorts the information received by individuals and groups and may cause them to see other groups at home as more threatening or prone to violence than they really are. The media sends one-sided messages and receivers may draw one-sided lessons. This selection bias is likely to be even more extreme in the "partisan" press that is associated with one or the other side in a particular conflict. ₁₃ Although diffusion can, in theory, work both ways as a damper and spur to ethnic conflict--in practice the selection bias of the media will tend to heighten ethnic fears and provoke ethnic conflict abroad.

Coupled with this biased information flow are the emotions discussed above that often produce deep insecurities in individuals--especially when placed in potentially threatening environments. An individual's desire to belong to and identify with an ethnic group can prove an emotionally satisfying experience when peace and regularized patterns of intergroup relations prevail. However, when ethnic leaders exaggerate trends from abroad to stimulate fear and mobilize their supporters for competition and conflict, the result may be to entrap groups in deadly encounters from which there is no escape. Efforts to promote the security of one group then leave all groups with a heightened sense of insecurity. Intergroup linkages become gravely weakened, leading to societal incoherence and, at times, to state collapse. As the actors retreat to their safe, ethnic containers, fewer and fewer of them are willing to risk contacts and cooperative initiatives with members of other communities, leaving them increasingly isolated and enmeshed in a Hobbesian world of group against group. Where polarization becomes complete and the state is overpowered by a single ethnic group, it is only a short step to the communal killings of colonial Algeria, Bosnia-Herzegovina, Burundi, Rwanda, Sri Lanka, and Nagomo Karabakh.

Escalation

Ethnic conflicts also escalate to include additional, foreign belligerents. ₁₄ Where diffusion occurs in part through information flows that condition the beliefs of ethnic actors elsewhere, escalation occurs through the more "traditional" routes of other interstate conflicts--alliances, spillovers, irredentism, diversions, and internal weakness.

Ethnic ties and antagonisms frequently motivate countries to become involved in ethnic conflicts elsewhere. In this form of "ethnopolitik," coethnics in one state are propelled by feelings of solidarity with their ethnic kin in a second. This largely occurs between neighbors where ethnic groups span national borders. India's intervention in Sri Lanka (Cooper and Berdal 1993, 186) and Hungary's attention to the treatment of its ethnic brethren elsewhere in Eastern Europe (Woodward 1995, 219) are prime examples.

In their study of the escalation of interstate conflicts, Randolph Siverson and Harvey Starr (1991) find that states join ongoing wars when they possess opportunity, defined by shared borders, and willingness, represented by a preexisting alliance. Most states lack the ability to project force over long distances, and thus contiguity conditions the ability of states to become involved in a conflict. Alliances reflect a self-defined interest in the security of another belligerent. In a similar vein, Will Moore and David Davis (1995) reason that ethnic alliances--cases where a majority group in one state is a minority group in a second--should have similar effects. Examining the behavior of all international dyads that are either contiguous or contain at least one great power (presumably possessing global reach), Moore and Davis find that ethnic alliances are an important source of interstate conflict. This effect is particularly prominent when the minority kin group is politically mobilized, indicating a higher level of ethnic conflict within that state. While there are obvious exceptions, such as the alliance between Catholics in Ireland and Northern Ireland which has not resulted in heightened interstate conflict, the strength of the overall pattern is noteworthy. In short, ethnic ties are a primary source of escalation.

Ethnic conflict within a state can also act as a trigger for interstate conflict in four other ways. First, in ways similar to the first and most direct path of diffusion above, ethnic warfare may spill over into neighboring territories and draw states into conflict. Ethnic combatants in one state may use the territory of a second for staging areas, retreats, and so forth--with or without the latter's consent. This spillover can lead to recriminations between the two affected states and, in cases of "hot pursuit," direct border clashes that contain the potential to spiral out of control. Such spillovers have been a frequent worry in Africa and Southeast Asia (Keller 1995). In March 1991, in but one possible example, Charles Taylor's forces in Liberia joined with Sierra Leonean dissidents and invaded Sierra Leone (Wippman 1993, 170). There have also been problems in the former Yugoslavia, with minor incidents occurring at the Austrian border, Serbian/JNA overflights of Hungary and the bombing of a border village, and JNA forces withdrawing through Italian territory (Steinberg 1993, 52). While potentially dangerous, and often used as an excuse for involvement by neighbors looking for a greater role, such spillovers can be resolved amicably. However, as borders become more fluid in areas of ethnic instability, violent encounters may become more frequent.

Second, ethnic mobilization often contains within it an irredentist dimension, as ethnic leaders demand the reunification of an often mythical but nonetheless politically salient ethnic homeland--typically defined as the largest area of territory ever controlled or believed to have been controlled by the group (Carment 1993, Carment and James 1995). Examples include Somalia's invasion of the Ogaden region of Ethiopia, Nazi Germany's threat to incorporate the Sudetenland, and Pakistan and India's continuing conflict over Kashmir.

Third, ethnicity provides a strong basis for "diversionary wars" stimulated by political leaders beset by opposition at home and seeking to rally support for their continued rule by inciting conflict abroad. ₁₅ Ethnicity and its emotional appeal to an "us versus them" outlook provides a particularly salient principle of organization and support. This was precisely the strategy used by Serbian President Slobodan Milosevic; faced with growing opposition to his presidency and a majority that favored far-reaching economic reforms, which he opposed, the embattled president "played the ethnic card" and precipitated the collapse in Yugoslavia (Bates and Weingast 1995; Djilas 1995, 85 and 105).

Fourth, predatory states within the region may consider states with significant internal conflicts to be easy targets. With the other weakened by internal dissent, aggressor states may calculate that their prospects for an easy, cheap victory are now greater than before; challenging the target is thus more attractive. Ethiopia's internal weakness, for example, appears to have contributed to Somalia's 1977 challenge in the Ogaden (Carment and James 1995, 94). Such strategies may also backfire, however, as predatory states often appear to miscalculate the rally effect that their aggression provokes. This may have been the case in the Iran-Iraq war, when Saddam Hussein sought to take advantage of the revolution in Iran to settle outstanding territorial issues to his advantage; contrary to his expectations, Saddam actually mobilized support for the new regime in Iran and locked the two countries into a long and bloody war.

Although still poorly understood, diffusion and escalation are real and, correctly, important concerns to policy makers worldwide. The studies commissioned for this project and summarized where appropriate in this paper shed some light on this complex process of diffusion, but our understanding remains at a rudimentary level. For scholars and analysts interested in the international spread of ethnic conflict, no question is more important than how and why groups learn from conflicts abroad. Even at this stage, however, it is clear that both diffusion and escalation can result in devastating ethnic conflicts not only for the groups involved but potentially for other states as well. Not only must such conflicts be managed to end the violence, but they must also be controlled to inhibit their spread.

III. The Management of Transnational Ethnic Conflict

Managing ethnic conflicts is a complex and ongoing process. Success is both difficult to achieve and transitory. There are no permanent resolutions, only temporary "fixes." The analysis above, however, clearly suggests that the key to successful management lies in reassuring minority groups of their physical safety. To foster stability and constructive ethnic relations, the rights and position of the minority must be secured. Confidence-building measures undertaken at the national level are the preferred instrument to this end. In light of group fears and elite ambitions, however, international intervention may be necessary and appropriate. Even so, confidence-building measures and international interventions are imperfect. In the end, ethnic groups are left without reliable safety nets and some measure of conflict may be inevitable.

Conflict Management as Middle-Level Theory

International relations theory and conflict management theory and practice are often viewed as distinct forms of knowledge. The former, which we have focused on in our discussion of the origins and spread of ethnic conflict, is concerned with developing broad, systematic insights into the interactions of states and groups irrespective of their applicability to specific problems. The latter focuses on the use of various types of knowledge and experience to achieve certain desired outcomes. International relations theorists seek abstract but generalizable knowledge to understand better the forces at work in international life. Conflict managers make use of middle-level theories, case specific knowledge, and intuition to respond effectively to the many, unique challenges in the political environment.

Although both types of knowledge make distinct contributions to our understanding of international politics and ethnic relations, it is also important to note that links between these approaches exist and can prove beneficial for both. Those concerned with general theory may well gain new insights as they take account of the political and social factors that are a part of the real-world decision process; here, such variables as past state/ethnic relations, political memory, the availability of mass media information, diffusion experience, the political culture of accommodation and violence, and spatial configurations come immediately to mind. And those charged with the task of formulating policy, decisions that often leave little time for research and reflection, can nonetheless benefit from using the general knowledge of international relations theory to aid in the analysis of specific problems. The discussion above, for instance, diverts the attention of practitioners away from ancient hatreds as an explanation of ethnic conflict and focuses it, instead, on the strategic interactions between and within groups that can produce violence. Conflict managers can, we believe, profit from these general insights.

Although some common ground does exist between these two forms of knowledge, in the end it is those prescribing policies who must choose from international relations theory what will help them to understand and deal with pressing foreign policy issues. As Alexander George observes: "General knowledge of international relations produced by scholars can be only an input to, not a substitute for, the policy analysis of a specific problem conducted within the government. Policy analysts, not academic scholars, have the difficult task of adapting the available general knowledge of a given strategy or foreign policy undertaking to the particular case at hand" (George 1993, 21 [italics deleted]; see also Lepgold 1995).

Constructive and Destructive Incentive Structures

The variety of ethnic experiences makes the development of generic guidelines for conflict management extremely difficult. Where ethnic groups possess effective safeguards, share pacific expectations, and feel secure in their relationship with the state and each other, inter-group competition tends to be constructive. Ethnic leaders are not fearful for their group's future and can operate within existing political institutions to maximize group interests. These elites abide by the rules of the game because they perceive the possibility of achieving beneficial outcomes for themselves and their ethnic constituents. The result over time is a growing sense of confidence on all sides about the intentions of ethnic rivals. As rivals demonstrate their commitment to deliver on bargains, confidence evolves and elites develop pragmatic, even positive, perceptions about each other. The possibilities for such constructive inter-ethnic relations should not be underestimated. Ethnic groups have lived side by side in amity for centuries in many areas of the world. To focus exclusively on the destructive side of ethnic relationships perpetuates dangerous political myths.

But what about situations where safeguards, shared norms, and pragmatic perceptions are absent and the prevailing incentive structure encourages ethnic leaders to adopt damaging courses of action? Destructive relations are not the norm, but they can surface where the incentive structure makes violence seem expeditious. Facing the dilemmas above, political leaders play the ethnic card in a calculated effort to benefit themselves and their constituents, even though the cost appears high in terms of the common good of the society as a whole. Threatening language and action can lead to societal polarization, precipitating a situation of grave suspicion and uncertainty for all. As David Rieff notes, "the Bosnian Serbs won because they knew how to take old fears and old complaints, repackage them, and cause otherwise decent Serbs, people from a national community with no more of an innate predilection for murder than any other national community, to commit genocide . . .What began as a tactic of pure massacre and terror in villages," he continues, "had evolved within six months into a sophisticated system for the destruction of a people" (Rieff 1995, 112 [italics added]). In light of such menacing possibilities, and especially its planned, genocidal variant, can ethnic brutality be constrained? Can confidence-building mechanisms within the state and international pressures and guarantees reassure uncertain peoples as to their future and make cooperation possible? There are no reliable safety nets in our anarchic world able to secure the cultural or physical survival of beleaguered ethnic peoples. Nonetheless, we believe confidence-building measures and international interventions can promote cooperative interethnic encounters (Stein 1990, 111).

Coping with Uncertainty and Fear

How can state and international leaders create an environment that avoids the rhythm of destructive relations culminating in ethnic violence? How can they promote a sense of security and moderate politics? In discussing the management of ethnic conflict, Kwame Anthony Appiah points to a seeming paradox when he observes that the attempt to manage identities brings reason to a subject where the identities themselves are based largely on imagined origins (Appiah 1992, 178). Appiah is right that ethnic identities lack a fixed, centuries-old, primordial basis. In Bosnia and Rwanda, as John Steinbruner has remarked, there are no readily ascribed bases for defining distinct ethnic identities; rather, elites evoke these perceived differences to rally their members for the maximization of individual and group interests; in doing so, however, ethnicity becomes "the organizing basis for conflict" (Steinbruner 1995, 7). Clearly, a focus on management must distance itself from the debate between the primordialists and instrumentalists, for group consciousness can result in murderous actions irrespective of whether this awareness is of long standing or not.

As shown above, it is difficult to transcend the dilemmas that produce collective fears. As a consequence, we have little alternative but to recognize that there is no form of insurance sufficient to protect against dilemmas of this sort, only the possibility of limiting their impact. Croatia's ruthless treatment of its Serbian population in the early 1 990s went virtually unrecognized in the Western press. In Bosnia, international appeals by the contact group and the United Nations had little influence on Serbian treatment of the Muslim population. The fall in eastern Bosnia of such safe-areas as Srebrenica and Gorazde, and the ruthless ethnic cleansing and presumed murders that followed, are stark reminders of the limited influence of the international community. Only the NATO bombings of September 1995 appear to have penetrated the cycle of conflict in Bosnia, but it is still too soon to tell whether this intervention, the Dayton Peace Agreement that followed, and now the deployment of the NATO-led Implementation Force will produce a long-term peace. The aborted restoration of hope in Somalia provides a second telling reminder of the limited ability of the international community to mitigate collective fears.

In theory, preventive diplomacy appears to be the most appropriate form of action that domestic and international leaders can take before the frenzied ethnic killings begin and before such confrontations create vested interests in prolonging the crisis. Fearing another downward spiral into widespread violence in Burundi, the Clinton administration and the international community have taken a series of preventive measures to bolster moderate forces and to "deter extremists from fomenting violence or overturning the current fragile power-sharing arrangement," including the facilitation of local dialogues, the deployment of military observers, radio appeals, the strengthening of the judicial system, high-level visits to reassure moderate politicians that the world community remains concerned, and the provision of developmental assistance (Friedman, 1995). As Bruce Jentleson notes, acting to deter conflicts from escalating into crises is "unassailable" in principle, but difficult to formulate and implement in practice. As Jentleson and others conclude, preventive diplomacy involves complicated political and military decisions regarding the domestic jurisdiction of states, the timing of interventions, and problems of political will and capacity on the part of intervening states and organizations (Jentleson 1995, 1, Stedman 1995a, 14-20, Lund 1995, 161-163). Clearly, as illustrated by the crisis in Rwanda, international actors are not prepared at this juncture to develop effective strategies for dealing with ethnic disputes in advance of the manifest conflict.

It must also be seen that the scope for third-party mediation between contending ethnic-based rivals at both the negotiation and implementation stages is limited in effectiveness. Internal wars are particularly difficult to negotiate, because ethnic enmities tend to be so deep and the stakes so high. At times, even despite the depth of these animosities, negotiators have managed to bring about an end to the fighting, as in Zimbabwe and Angola. Yet, it is important not to expect too much from mediation and negotiation. As the data on negotiations indicate, settlements are difficult to achieve and at least as difficult to maintain, even where a third party is prepared to step between the adversaries. Roy Licklider, largely reconfirming earlier studies by Stephen Stedman and Paul Pillar, finds that only 14 out of 57 civil wars between 1945 and 1993 were settled through negotiations (Licklider 1995, 684, Stedman 1991, 5-7, Pillar 1993, 25). Barbara Walter (1995) suggests that interstate wars are actually easier to bring to a negotiated conclusion than civil wars; in the former, the two parties remain on opposite sides of their border, but in the latter the disputants must re-merge themselves into a single unit and face larger problems of credible commitment as a result. Particularly in cases where insurgencies have an ethnic or nationality dimension, a mediated agreement may be difficult to arrange because highly sensitive issues of legitimacy are involved (Frei 1976, 70).

The difficulties encountered in negotiated solutions are also reflected in the implementation process. Unlike the ending of wars by means of military victory or capitulation, where the power of identity groups to resist central authority is largely eliminated, negotiated settlements leave ethnic entities with sufficient space to frustrate the ambitions of state elites (Licklider 1995, 685). The concessions necessary to bring about agreement often result in a complex and indeterminate process at the consolidation stage. Because deep distrust of an opponent remains in place at the time an agreement is set in motion, commitments made at the bargaining table may not be credible. Within-group rivalry may come into full view after the agreement, the provisions of the agreement may be vague and cause new tensions, and the world community may not be prepared to give sustained support to peacekeeping and peacemaking. The presence of such dilemmas continues to threaten the renewal of violence.

Accordingly, it was prudent for chief American negotiator Richard Holbrooke to state at the 1995 talks on Bosnia that, with several of the main issues still unresolved, there could be no 30 guarantee that any settlement would hold (Sciolino 1995, A6). Even now, after laboriously working out an agreement, the Dayton peace accord may still unravel. After all, the Bosnian Serbs rejected a power-sharing arrangement in 1992 and, and during the current implementation phase, may fail the commitment test again. Indeed, efforts by the Bosnia Serbs to prolong their control of the suburbs of Sarajevo suggests that their commitment to the peace accords is quite shallow.

In Rwanda, extremist Hutus feared the consequences of the power-sharing provisions of the 1993 Arusha accords. These provisions gave the Tutsi-led Rwandese Patriotic Front responsibility for five of the 20 ministries (including the Ministry of the Interior) as well as 40 percent of the enlisted ranks and 50 percent of the officer positions in the new army. In the case of Rwanda, then, the power-sharing agreement itself "contributed to [a] polarization of political tensions," because the Hutu hard-liners viewed it as jeopardizing their position of political influence in the country (Newbury 1995,15). Fearing for their future under this agreement, the extremists launched a highly destructive preemptive strike.

While it is important to recognize how difficult it is to overcome ethnic fears by means of negotiated agreements, it is also necessary to appreciate that there are few alternatives to negotiations if both sides are to be brought into the solution. For a mutually satisfactory peace to take place, a two-step negotiating process is essential: first, among the key elements within the group, and then between the groups themselves. Operating rules must be hammered out in these talks regarding inclusive coalitions, proportionality in recruitment and allocations, autonomy, provisions on electoral competition, and so forth. The ensuing negotiations are likely to be protracted and difficult, largely because the various factions and groups lack a clear chain of command (making commitments difficult to produce) and because they understand fully that the terms they accept will cast a long shadow over their future. But if each of the parties concludes that its alternatives are limited, its present course unduly costly, and its stake in its rival's willingness to cooperate with an agreement significant, they may then begin to negotiate in good faith.

Given possibilities for converging interests, it is important to discuss the instruments available to adversaries and mediators for mitigating fears and reducing the possibilities of conflict (Oye 1985, 3). We turn first to the negotiation of confidence-building measures that hold out some promise of promoting interethnic cooperation. The acceptance and implementation of these measures can enhance prospects for interdependence through iterated encounters, possibly culminating in a growth of confidence in each other's goodwill. However, because some conflicts cannot be expected to end by this process, the situation may require external intervention to enforce stability and promote a consensus among the contending state and ethnic actors. In the post-Cold War world, interventions in intrastate ethnic conflicts are likely to encounter strong public resistance in the advanced industrial states--even though "the greatest need for U.S. forces is peace-keeping and peace enforcement" in precisely such confrontations (Maynes 1995, 109). The very process of mobilizing ethnic constituencies for the bitter intrastate wars that can follow inhibits Western democracies from becoming involved militarily in what promise to be protracted engagements. Consequently, the safety nets for ethnic minorities remain fragile.

Confidence-Building Measures

Confidence-building measures seek to reassure ethnic peoples about their future. As Saadia Touval puts it, "actors may resort to insurance and other forms of risk management in order to reduce risks" (Touval 1982, 19). Through packages of coercive and noncoercive incentives, the state attempts to assure ethnic minorities about their place in society. By means of these concessions, it seeks to get recalcitrant elites to rethink their belligerent practices and cooperate in joint problem-solving initiatives. To overcome minority fears, confidence-building measures must be appropriate to the needs of those who feel vulnerable to the majority-backed state. The challenge, as I. William Zartman observes, "is to keep the minority/ies from losing" (Zartman 1996). Such safeguards, if handled sensitively over the years, may be able to cope with the central questions of sharing private information and making credible commitments. There are four major trust-building mechanisms for helping ethnic minorities deal with their feelings of insecurity.

Respect

The security of ethnic peoples is in no small way based on a reciprocity of respect. Unless each side views its opponent as honorable and having legitimate interests, relations are likely to be marred by a history of intended or unintended affronts. "The more invidious the intergroup comparison and the larger the area of unacknowledged claims to group legitimacy," writes Donald Horowitz, "the more intense the conflict, all else being equal" (Horowitz 1985, 215). Ethnic affronts can be highly injurious to a group's pride and self-esteem, widening social distance between groups and exacerbating fears among ethnic minorities that their children will be relegated to second-class status for an indefinite time.

Relations in Bosnia, worsened by polarization and increasingly hostile perceptions, have been further aggravated by the contempt Serbs have reportedly shown their Muslim adversaries. Describing themselves as the only people in former Yugoslavia "who have the talent, energy, experience, and tradition to form a state," they characterize their adversaries as representing "all that is base, undesirable, and naturally and other forms of risk management in ordersubordinate" (Cigar 1995, 775). And in the Sudan, southerners, with strong memories of slavery and perceptions of low status, bridle at any new evidence of disrespect. Thus, they viewed the Sudanese government's decision to apply Islamic (Sharia) law to them (as well as the Muslims living in the country's north) as a confirmation of their second-class status (Amnesty International 1995, 57). Their resentment boiled over in 1994, when the minister of state in the president's office, at the mediation talks in Nairobi held by the Inter-Governmental Authority on Drought and Development, allegedly treated both southerners and the IGADD mediators with contempt when rejecting the southerners' call for self-determination and a secular state (Sudan Democratic Gazette (1994,3).

To be sure, the fears of ethnic minorities may be overstated. Minorities in Eastern Europe are described as having "an exaggerated fear of the loss of identity"--a legacy of distrust of majority authorities that causes them to make broad demands for legal guarantees. The majorities, fearful that this will start them down the slippery slope toward the breakup of their states, refuse to consent to these demands (Watts 1995, 92-93). But to build confidence it is imperative that dominant state elites take minority ethnic resentments and anxieties into account. Unless old psychological hurts are taken seriously, regimes will be unable to avoid the problem of "wounded tigers" in years ahead. In this regard, those involved in the management of ethnic disputes can learn much from C. E. Osgood's Graduated and Reciprocated Initiatives in Tension Reduction (GRIT) strategy for easing conflict between the superpowers during the Cold War (Osgood 1962). Osgood's suggested approach of repeated overtures (in this case by a dominant majority controlling the state) without expectations of an immediate tit-for-tat response could stimulate full negotiations between equals. Unless past wrongs are redressed and the sting of disparagement is removed from current ethnic interactions, internal negotiations will remain clouded by an overhang of bitterness and suspicion and minority uncertainty regarding adversary intentions will contribute to serious conflicts.

Power Sharing

When ethnic minorities fear that their exclusion from the decision-making process will leave them exposed and vulnerable to majority preferences, conflict management requires an effort by the state to build representative ruling coalitions. In conceding to ethnic minority members a proportionate share of cabinet, civil service, military, and high party positions, the state voluntarily reaches out to include these minority representatives in public affairs, thereby offering them an important incentive for cooperation. In South Africa, for example, President Nelson Mandela agreed to include power-sharing provisions in the interim constitution in an effort to reconcile the economically-dominant local white community as well as to build confidence among largely white investors abroad.

Power sharing can be informal (e.g., Kenya, 1960s) or formal (e.g., Nigeria, 1979), and can take place in authoritarian (e.g., Zambia, 1980s) or democratic (e.g., South Africa, mid-1 990s) settings. In both Eastern Europe and Africa, there has been a mixed pattern of "hegemonic exchange" regimes--centrally-controlled one or no-party regimes that allow a limited amount of bargaining to take place between state, ethnic, and other elites. Under the authoritarian administrations of Josip Broz Tito in Yugoslavia or Felix Houphouet-Boigny in Cote d'Ivoire, nationality or ethnic representatives met with the president in cabinet sessions, where strong differences were sometimes aired by group spokespersons behind closed doors. The resulting power-sharing systems are quite diverse; yet, they have in common a form of coordination in which a somewhat autonomous state and a number of less autonomous ethnic-based and other interests engage in a process of mutual accommodation in accordance with commonly accepted procedural norms, rules, or understandings (Rothchild 1986a, 72). These elite-power-sharing arrangements are inevitably temporary, but while they last they provide some security for political and ethnic minorities.

Pacted democracy, with its rough reflection of the configurations of elite power, can prove relatively easy to organize in an interim constitutional situation. Elite pacts provide a relatively stable form of governance for a transitional period, but if they remain unresponsive to public demands for change over too long a period, then the pactmakers risk becoming isolated from their supporters, allowing new uncertainties to surface (Karl 1986, 217--218). In principle, there is no logical reason why such structural arrangements cannot lead to a more open system of sharing-even full democracy, as happened in Colombia. In reality, however, Colombia possesses a rather unique form of party (not ethnic) relations, creating grave doubts about power-sharing's ability to prove an effective confidence-building mechanism over the long term and in other places.

Because of the fragility and temporary nature of these state-inspired inclusive coalitions, such mechanisms are likely to provide only minimal assurances to ethnic minorities. First, as already indicated, with ethnic balances of power nearly always in flux and information limited, these arrangements are necessarily transitional ones. Second, if poorly negotiated and implemented, the incomplete ethnic contracts may be rejected eventually by the groups they are designed to protect. The number of people appointed to the cabinet or civil service is not in and of itself a guarantee of proportional group influence (Mattes 1993, 76). Minority representatives can, as in the Sudan at various times, be assigned insignificant portfolios and therefore wield only minor influence. Majority and minority parties may also both pull back from power-sharing arrangements, regarding these arrangements as co-opting them into a system they view as still potentially threatening. In Sri Lanka, hard-line elements within the minority Tamil community rejected President Chandrika Kumaratunga's 1995 proposal giving them control over a semiautonomous region in the north, a proposal also opposed by some of the more nationalist Sinhalese in her own cabinet (Sisk 1996); at the same time, the insurgent Liberation Tigers of Tamil Eelam reportedly charged that the government's tactic of peace negotiations was in fact a pretext to cover its planned military offensive. Similarly, hard-liners among the majority Hutu in Rwanda, resentful of the power-sharing provisions of the 1993 Arusha accords, launched a preemptive strike that included a genocidal assault against both the Tutsis and some Hutu moderates.

When not applied with great care, then, power-sharing arrangements can backfire. Ethnic elites must be prepared to interact with other elite representatives they find personally repugnant, something difficult to do under normal circumstances but especially so where the norms of collaborative politics are not in place. In order to achieve inclusiveness, should those crafting power-sharing structures bring ethnic extremists into decision-making bodies to build greater confidence? Again, local circumstances are critically important in assessing the appropriate course to follow. For instance, in situations where a majority African government has perceived its local white community to be useful in achieving its developmental objectives as well as in reassuring others abroad about the safety of their investments in the country--as in Kenya, Zimbabwe or South Africa--government authorities have restrained their anger over the humiliations of colonial times and taken care to include white representatives in important cabinet positions. However, in situations where Africa's ethnic groups regard each other's ambitions for control of the state and its hold on publicly-controlled resources suspiciously, as in Burundi and Rwanda, their essentialist perceptions of their rival's intentions frequently lead to an inflexible stance in favor of appointing their group members or conciliatory outsiders to high government positions. It may be possible to justify an exclusion of radical adversaries in such circumstances. Thus, including Hutu extremists in a post-Arusha government in Rwanda, for example, would not likely have led to an easing of interethnic tensions.

Moreover, where the majority-dominated state remains unprepared to respond to legitimate minority demands for full participation in decision-making activities, power-sharing schemes are likely to unravel and become themselves a source of grave insecurity. Power-sharing by itself is not a solution to ethnic fears. While it offers some safeguards against exploitation, it cannot prevent extremist elites and their supporters from polarizing society and pulling on the social fabric.

Elections

Although elections represent only a brief episode in a larger political process, they can have enormous influence on intergroup collaboration and conflict. Where favorable circumstances prevail (i.e., an agreement on the rules of the political game, broad participation in the voting process, and a promising economic environment), elections can promote stability. In democratic regimes, where institutionalized uncertainty provides many players with an incentive to participate, the election process can legitimate the outcome (Przeworski 1991, 26). All groups have a reason to organize and, through coalitions, they are given an opportunity to gain power in the future. This prospect of competing in accordance with the procedural norms of the system can be reassuring to minority interests; not only do they have a chance to advance their individual and collective interests, but they have reason to be encouraged by the majority's commitment to the electoral contract. The effect is to preempt conflict.

The implications of elections, however, can also be troubling in multiethnic settings. With opportunities limited and competition for positions and resources intense, some leaders can choose to further their individual and collective interests--even though at a high cost in terms of the society's overall well-being. As noted above, where ethnic leaders seek to promote their parochial interests by attempting to outflank their centrist rivals through militant appeals to their ethnic kinsmen, the result may be to increase strife and undermine the frail, cross-cutting linkages that buttress democratic regimes. Ethnic outbidding heightens minority insecurity. Groups make greater demands on the state and one another. They strain against their own commitments to the existing ethnic contract--and worry about the commitment of the other. Values of restraint and civility are weakened and suppressed emotions of dominance come to the fore in majority circles. In some circumstances leaders can repackage and play upon latent grievances in such a way as to foster a collective response highly damaging to their stereotypic enemies. As a result, elections in certain circumstances can prove very destabilizing, threatening minorities with the possibility of discrimination, exclusion, and even victimization.

In practice, those crafting constitutions have organized elections in two ways to promote inclusive coalitions. First, electoral rules can be set so that candidates are forced to appeal to more than one ethnic group. In an effort to give presidential candidates an incentive to appeal to a broad, cross-section of communal groups to gain the necessary support, for example, the aborted 1993 constitution in Nigeria provided that a winning candidate would be deemed to be elected when that person secured a simple majority of the total number of votes cast as well as one-third of the votes cast in each of at least two-thirds of the states. The intent here was evident. In securing a majority of votes in a multiethnic society, moderate appeals, with their overarching themes, were expected to win out over parochial ones. The effect of adopting such an electoral system would likely be to build a measure of confidence among ethnic minorities regarding their future political status.

Second, electoral rules can also be crafted to ensure some minimal representation of all ethnic groups in the society. Those seeking to encourage minority representation in party lists and in ruling coalitions have looked favorably on systems of proportional representation (PR). For example, in structuring the elections for the Russian State Duma (or lower chamber of parliament) in 1993, the legal drafters provided for a chamber of 450 members, half on the basis of single-member constituencies and half on the basis of PR. Constituencies vary enormously in size, ranging from as few as 13,800 in the Evenki autonomous region to over 700,000 in Astrakhan province. Several observers have concluded that such a system ensures the representation of small ethnic peoples in the State Duma (van den Berg and Simons forthcoming). Similarly, in South Africa, the African National Congress agreed, somewhat reluctantly, to use PR during the transition period to give racial and ethnic minorities a sense of security at a difficult time of transition (Sisk 1993, 87). However, because the present system seems cumbersome and fails to generate close links between a member of parliament and his or her constituents, ANC leaders are currently thinking of modifying the electoral system to be more like those in Germany or Russia, with their mixtures of PR and plurality voting systems.

The way that state elites structure electoral arrangements is likely to prove critical in building confidence in minority circles. Nigeria's broadbased electoral formula and PR are two possible ways of encouraging minority ethnic participation and inclusion; yet, they are likely to endure only as long as they retain support among key groups and state elites. At such time as the majority shifts its concern away from the values of representativeness, a change in electoral rules can take place. Unless this change is handled fairly and with extreme sensitivity, it can be perceived by minority groups as inimical to their interests. As a consequence, considerable time is required before minorities come to see electoral laws as reliable foundations for their security.

Regional Autonomy and Federalism

In the recent scholarly literature, much attention has been paid to the administrative benefits of decentralized political systems and strengthened local and regional initiatives. Frequently, however, writers have shown little concern for the role that decentralization schemes can play in managing political conflict. By enabling local and regional authorities to wield a degree of autonomous power, elites at the political center can promote confidence among local leaders who come to exercise a limited but important set of administrative responsibilities.

Measures on decentralization, regional autonomy, and federalism featured in peace negotiations in Bosnia, Sri Lanka, Cyprus, Sudan, Angola, Mozambique, and South Africa. In each, they provided insurgent militias with an important incentive for responding positively to the government or third-party mediator's proposals for settling the conflict. The American-brokered peace initiative in Bosnia achieved a key breakthrough in the September 1995 negotiations, for example, when the Bosnian government agreed to recognize an autonomous Bosnian Serb entity, called Republika Srpska. In exchange, Serbia and Croatia accepted the legal existence of Bosnia and Herzegovina with its present borders and endorsed the division of the country--51 percent of the territory to the Bosnian government and Bosnian Croats, and 49 percent to the Bosnian Serbs. All three parties perceived control of Bosnia's space to be critically important for their survival once peace came into effect. Given the deep suspicions that continue to prevail among these identity groups, however, it remains to be seen whether such a compromise, as enshrined in the Dayton Peace Agreement, can endure and contain the hostilities.

In attempting to create a new balance between state and society, groups turn to decentralization as a means of placing institutional limitations on unbridled central authority. Politically marginalized groups have vivid memories of excessive state penetration and a continuing fear of majority domination. Decentralization and the authority these schemes allow local elites can, therefore, become confidence-building mechanisms that safeguard the place of minorities in the larger society. In Ethiopia, for example, President Meles Zenawi looks to a scheme of ethnic federalism as a means of reversing the repressive, hegemonic practices of previous governments that have led to internal wars (McWhirter and Melamede 1992, 33). The 1994 Constitution gives the nations making up Ethiopia wide powers, including an unconditional right of self-determination and secession.

Nevertheless, experiments with decentralized systems in India, Pakistan, Cyprus, Sri Lanka, Kenya, Uganda, South Africa, Sudan, and Ethiopia reveal serious practical difficulties in securing majority-backed state acceptance for these attempts to insulate minority interests from central authority. Determined to prevent the division of the state, public officials have taken firm action to avert a weakening of control. In extreme cases, they have revoked previous concessions. Thus, as Yugoslavia began to disintegrate in 1989, President Slobodan Milosevic rescinded the autonomous provincial status within Serbia given to largely Albanian-populated Kosovo by former President Tito. Similarly, Sudan's President Gaafar el-Nimeiry, the main advocate of political accommodation with the Southern Sudan Liberation Movement insurgents in 1972, backtracked on his commitments formalized in the Addis Ababa accords and began to dismantle the quasi-federal compromise. In a series of moves intended to placate hard-line, Muslim elements within his government, Nimeiry intervened in southern regional elections, changed regional boundaries, redivided the southern region, applied Sharia law to non-Muslims, and ultimately, abrogated the agreement itself. In these and other cases, the voiding of concessions on autonomy heightened tensions, leading to new or renewed violence. Where central governments, jealous to guard their prerogatives, have changed their mind and nullified concessions already made, the result has been extreme suspicion, tension, and broken commitments that have led in some cases to a renewal of heavy fighting.

Conceived as a type of safeguard, regional autonomy and federalism have had, in some instances, unintended consequences that have actually increased conflict. Despite efforts to decentralize power in South Africa and Ethiopia, the fiscal dominance of the political center has tended to undercut the significance of regional authorities. Moreover, efforts to delineate boundaries have increased conflict between ethnoregional identity groups. In contemporary Russia, the arbitrary way in which internal boundaries divide ethnic peoples has been a major source of tension (Lapidus and de Nevers 1995, 3). The regional boundaries set up by Ethiopia's government appear to favor Tigray and the Afars, at the expense of the formerly-dominant Amhara and the Somali Isaks in the Awash Valleyland. Unless carefully crafted, decentralization schemes may worsen rather than improve inter-ethnic relations.

In sum, confidence-building measures are potentially creative instruments in the hands of state elites intent on reassuring ethnic minorities. They indicate a sympathetic concern on the part of those in power to the fears and uncertainties of minorities. By acknowledging and showing respect for difference and by agreeing to share resources, state positions, and political power with exposed and vulnerable groups, these measures reduce the perceived risks of association and provide incentives for cooperation. They can also become the basis for an iterated process that can culminate over time in a shared sense of common fate among diverse communities.

However, such concessions represent conflict management, not conflict resolution. They can reduce some of the surface factors giving rise to ethnic fears, but they do not alter the basic relationships that cause these fears in the first place. They represent only partial safeguards. The risks in ethnic encounters remain in place, even if covered over by concessions from powerful state actors. And because there is always the possibility that groups will adopt more threatening forms of interaction, these confidence-building measures never eliminate the information failures, problems of credible commitment, and security dilemmas that are embedded in ethnic encounters. As Adam Przeworski astutely observed, "if sovereignty resides with the people, the people can decide to undermine all the guarantees reached by politicians around a negotiating table. Even the most institutionalized guarantees give at best a high degree of assurance, never certainty" (Przeworski 1991, 79).

External Intervention

If confidence-building mechanisms within the state are not sufficient to overcome the incentives for violence rooted in the strategic interactions of groups, it is necessary to turn to the international environment and ask whether external intervention can safeguard minorities against their worst fears. For many observers, sovereignty is linked to responsibility: state elites are expected to guarantee minority rights and to provide the means for establishing and maintaining regularized patterns of state-society and interethnic relations. The state, with its monopoly of force, is often in a position, as one South African mediator described it in 1995, to "enforce stability" between local warring parties (in this case, in the East Rand townships in his country). But who will intercede if the state is unable or unwilling to secure the safety of its minority peoples? What forms will this intervention take? And which of the interventions, if any, are likely to have a significant impact on intrastate conflicts?

At the outset, it is important to note that some of the states in Eastern Europe, the former Soviet Union, and Africa are notably "soft" and unable to enforce their regulations throughout the territory under their influence. As noted above, the decline of the state contributes to an environment in which intergroup violence can take place. Other states, lacking effective control and unwilling to live with the uncertainties of ongoing negotiations, opt for heavy-handed repression in an effort to compensate for their weaknesses (Lapidus and de Nevers l995, 35). In such abusive contexts, such as Siad Barre's Somalia, the state itself can become the source of intense conflicts with ethnoregional opponents.

Where international action is sanctioned, external actors, concluding that sovereignty is not being exercised in a responsible manner, can decide to intervene in intrastate conflicts to protect minority interests and to insulate the international community against spreading violence. As Stephen Krasner (l995) has shown, states have a long history of intervention in the ethnic (and religious) affairs of others. The principle of sovereignty has never been articulated or respected in the clear-cut manner often assumed by scholars of international relations. Many of the treaties settling European affairs in the aftermath of World War I contained provisions obligating states to protect the political and religious rights of minorities within their borders. More recently, the United Nations Charter affirmed an international commitment to basic human rights and fundamental freedoms, and Secretary General Boutros Boutros-Ghali now believes that "the time of absolute and exclusive sovereignty . . . has passed" (Boutros-Ghali 1992, 9). The Conference on Security and Cooperation in Europe (now the Organization on Security and Cooperation in Europe), has always sought to promote human rights within states, with a meeting of national experts on minority problems stating in early 1991 that "issues concerning national minorities, as well as compliance with international obligations and commitments concerning the rights of persons belonging to them, are matters of legitimate international concern and consequently do not constitute exclusively an internal affair of the respective state" (quoted in Kampelman l993, ix). Lori Fisler Damrosch, in turn, states that "large segments of the international community have been willing to endorse strong collective action in a wide range of situations," including genocide, interference with the delivery of humanitarian relief, violation of cease-fire agreements, collapse of civil order, and irregular interruption of democratic governance (1993, 12).

Nonetheless, since 1945 there has been a strong insistence by many countries on the protection of national autonomy afforded by the juridical principle of sovereignty. This emphasis on internal autonomy has often been strongest where states themselves were weakest, and has been particularly noteworthy in Africa (Jackson and Rosberg 1982). Yet, today ethnic conflicts and their possible spread have thrust issues of "humanitarian" intervention onto the policy agendas of the United States and many other countries. As Keller (1995) indicates, even in Africa there is a greater willingness on the part of state leaders to entertain limitations on the notion of sovereignty. This without doubt reflects their shock over the extreme brutality of ethnic wars and the inability of states with limited legitimacy and low capabilities to surmount these challenges (Obasanjo forthcoming). Despite this change in attitude, it remains an open question whether these leaders will be prepared to sanction international interventions directed against their own countries.

External intervention takes three broad forms: noncoercive intervention, coercive intervention, and third-party mediation during both the negotiation and implementation stages. We look briefly at each of these forms, drawing conclusions in each case about their anticipated effects on intrastate conflicts.

Noncoercive Intervention

In our shrinking global environment, international actors are increasingly distressed over the violation of minority rights taking place in other countries. They are also concerned with the possibility of diffusion, with demands for self-determination spreading and civil wars spilling over into neighboring territories--as seen in Liberia and Bosnia. This sense of alarm has, intermittently, brought action with outside states or multilateral organizations protesting infractions or exerting pressure on the transgressors. Western governments, encouraged by their domestic publics to denounce breaches of human rights in Bosnia, Chechnya, Rwanda, and Sudan, have criticized these abuses through quiet, behind-the-scenes diplomacy and at public fora.

Assertions of international norms are important in raising the costs of unacceptable behavior, especially when their advocates offer an alternative set of interests which defectors can mobilize around and challenge the ensconced ethnic leaders (Gagnon 1994/1995, 139). States are also in a strong position to use inclusion or exclusion from the international community to reward or punish regimes and ethnic leaders who deviate from internationally accepted norms. The promises of inclusion or the pains of exclusion can at times create strong incentives to behave in a more responsible fashion. Thus Serbian President Milosevic's desire to be accepted by Europeans and North Americans enabled Western diplomats to influence his behavior at the bargaining table, even causing him to make concessions on the emotionally charged issue of Bosnian government control over a unified Sarajevo. This concession brought him into contention with the Bosnian Serb representatives at Dayton, and with some hard-liners in the capital city, exposing deep within-group differences to the world community. Similarly, conditions on membership in international organizations appear to be mitigating ethnic conflicts in Hungary and Romania and Turkey's desire for acceptance in Europe may be limiting its actions against its Kurdish minority.

In South Africa, external protest and sanctions raised the cost of doing business, gaining access to technology and raw materials, and travel. Above all, international condemnation challenged state and governmental legitimacy. Sanctions physically punished the regime, something that became painfully evident in South Africa's loss of dominance in the air war over Angola--brought on in part by the air force's inability to secure spare parts. Yet, the symbolic impact of sanctions was also important, because it represented a clear statement of sympathy for black hardship and moral disapproval of apartheid policies by the international community (Strack 1978, 12). Yet, while South Africans found the costs of sanctions to be discomforting and burdensome, they did not hurt the main body of the white constituency sufficiently to alter priorities that is, until President F. W. de Klerk's remarkable change of heart on negotiating with the anti-apartheid opposition in the early l990s (Sisk 1995).

Given the extreme emotionalism over security issues that brings aggressive ethnic leaders to the fore in the first place, external appeals, exhortations, and pressures are not likely in and of themselves to dissuade determined elites from their abusive courses. A Ratko Mladic, the Bosnian Serb military commander, or a Saddam Hussein remains sufficiently insulated from world pressures that what transpires at diplomatic meetings or in the global press may have little immediate impact on them or their militant followers. Pointing to the crimes of "ethnic cleansing" or to the violation of safe areas in Bosnia, for instance, did little to deter such extremists as Dr. Radovan Karadzic, the Bosnian Serb leader, or Mladic. In fact, the crimes that were committed actually eliminated dissenters and waverers within the Bosnian Serb community and bound the ethnic community together tighter than ever before.

Frustrated by this willingness to endure external censure, international actors have sought to influence local preferences through the use of mandatory sanctions, the denial or threatened denial of recognition, the provision of intelligence information, and the provision of food and economic assistance to one or more parties. The West has also been actively involved in easing Eastern European insecurities by engaging majority governments in a discussion of their ethnic problems, facilitating communications between majorities and minorities, and pushing the two sides toward accommodation wherever possible. The effect, notes Larry Watts (1995, 95), has been to ease the siege mentality and the sense of vulnerability of local actors; nevertheless, it remains unrealistic to hope that these pressures will compel regional governments to change their basic policies.

Noncoercive interventions, as well as other forms of intervention discussed below, place the initiating party or parties in a dilemma. On the one hand, seeking to overcome the problems of between group strategic interactions, initiating countries need to engage in confidence-building in order to establish trust and to reassure majority and minority groups. On the other, targeting within-group interactions, the efforts of third parties to alter the incentives of leaders and punish particular groups (the Hutu in Rwanda or the Serbs in Bosnia) complicate the process of building trust across groups. In dealing with this dilemma, the initiating states can benefit from precise information on such subjects as within-group points of tension and capabilities and from the adoption of a mix of policies gauged to promote conciliatory behavior on the part of target groups.

In brief, noncoercive interventions of the type discussed here can be helpful in raising the costs of purely ethnic appeals and in structuring the incentives of group leaders prepared to accept international norms for the purposes of recognition and acceptance. Where conflicts are intense, however, exhortations and international warnings may not deter or end violence. The most noncoercive intervention can do in such situations is create a climate in which ethnic appeals and violence are perceived by all as illegitimate and, therefore, marginally less likely to be used.

Coercive Intervention

Third parties intervene militarily in intrastate conflicts in a peace-keeping or peacemaking role for a variety of reasons: to ensure food deliveries to the starving (e.g., Somalia, Bosnia), protect designated safe areas (e.g., Iraq, Bosnia), defend threatened peoples (e.g., Liberia, Rwanda), and establish a new regime (e.g., Uganda). States may be motivated by hegemonic ambitions, concerns for regional stability, sympathy for oppressed groups, a sense of international responsibility, or simple humanitarianism (Cooper and Berdal 1993, 197). They may also intervene to maintain their own moral values, deter the possible use of weapons of mass destruction, or forestall further diffusion (Brown 1993,120). Whatever the motive of states, the rise in ethnic conflict today creates new demands and opportunities for coercive intervention.

Although the classic notion of sovereignty is being questioned today, such military actions, which involve the penetration of another country's space, inevitably elicit challenges on the grounds of legitimacy. They can also involve high costs in building political coalitions and marshaling financial and human resources. In the end, the outcomes of these initiatives are less than inspiring. For every successful military intervention, as in the Gongo or Persian Gulf crisis, there are the failed efforts in Sri Ianka, Somalia, and Lebanon. As new intrastate emergencies arise, it can be anticipated that diplomats, analysts, and increasingly publics will raise doubts about the prudence of such undertakings.

External interventions have two primary effects. First, intervention can alter the internal balance of ethnic power and may lead groups to moderate their demands. Except perhaps where the two (or more) sides have reached a "hurting stalemate" (Zartman 1985) and the purpose of the intervention is exclusively to separate the forces and keep the peace, interventions always have political implications (Carr 1993). Even in Somalia, where negotiations on establishing a transitional national council held out hopes for a settlement in 1993, the initial humanitarian mission eventually favored one claimant to power (Ali Mahdi Mohamed) over the other (Mohamed Farah Aideed), ultimately causing the politicization of the mission (Hirsch and Oakley 1995). By typically favoring by design or default the weaker side in any internal conflict, external powers reduce the stronger side's chances for success. This, in turn, restrains the stronger party's demands. To the extent that such restraint takes hold, intervention can improve the prospects for agreement.

This positive effect of intervention, however, is offset by an opposite and possibly equal reaction. To the extent that its prospects of failure decline, and its prospects of success improve, the weaker side is likely to increase its demands and ask for more at the bargaining table (Wittman 1979). With both effects occurring simultaneously in any intervention, the "bargaining gap" between the parties is likely to remain the same--or close to it. The outcome of the negotiations may shift in favor of the weaker side, but the distance between the groups may remain as wide as ever.

These opposing tendencies have been played out fully in recent events in Bosnia. Once the NATO countries intervened decisively in September 1995 on behalf of the Bosnian government, and against the Bosnian Serb forces, the latter--pressured by Serbian President Slobodan Milosevic--quickly moderated their demands and moved towards accepting the 49 to 51 percent partition of the country that they had earlier rejected. ₁₆ At the same time, however, the Croats saw new opportunities on the battlefield and at the negotiating table, and the United States and its allies had to step up pressure on the Bosnian government and Croatia not to exploit their increased leverage.

As the Bosnian case demonstrates, unless pressure is exerted on both sides to moderate their demands, intervention by itself will not necessarily enhance the prospects for agreement. In this case, however, the United States did place pressure on all the combatants and brought the parties to an agreement in Dayton. Simultaneous pressures were also critical in the Zimbabwe independence negotiations, where a coalition of mediators brought substantial influence to bear on both delegations at the 1979 Lancaster House conference. The partisan effects of all interventions must be recognized and incorporated into any plan for bringing the disputants to a successful agreement. Nor does intervention, by itself, solve the information failures that may have thwarted agreement in the first place. On this score, intervention is an instrument of very limited effectiveness.

The second primary effect of intervention is to guarantee new ethnic contracts between the warring parties--at least during an interim period. As discussed above, problems of credible commitment hinder the efforts of groups to resolve their differences peacefully. Creating situations in which each side believes that the other will live up to the agreement may prove more difficult than actually bridging the competing interests of the groups. The primary attraction of external intervention is that an outside state can enforce an agreement, thereby providing the necessary credibility that is other-wise lacking between the parties to the dispute. Thus, in Namibia in 1989, the third-party enforcer was in a position to raise the costs of breaking agreements by monitoring the implementation process, highlighting violations of the peace agreement, and focusing an inter-

national spotlight on any breaches that occurred (Fortna 1995); the lack of any equally effective third-party enforcer in neighboring Angola following the signing of the Bicesse accords and the first round of the 1992 elections increased the incentives for Jonas Savimbi, President of the National Union for the Total Independence of Angola, to defect from the agreement and resume the civil war.

The promise of the post-Cold War world is that the great powers, freed from the shackles of superpower competition, can now intervene to mitigate ethnic conflict by providing external guarantees of social order. If the warring parties themselves cannot commit to uphold their pacts, external powers can lead the groups to peaceful solutions by enforcing any agreement they might reach. The paradox of the post-Cold War world, however, is that absent the bipolar competition that drove them into the far reaches of the globe, the United States and other powers now lack the political will necessary to make a sustained commitment to the role of external guarantors of new ethnic contracts. This lack of resolve partly reflects the vagaries of public opinion. For example, during the 1994 Rwanda crisis only 28 percent of the American public surveyed in June 1994 favored sending U.S. troops to stop the killing (CBS News, 1994); two months later, however, when the policy objective shifted to humanitarian aid, 69 percent of respondents favored such a strategy (Yankelovich, 1994). As these data indicate, intervention is a weak reed upon which to rest hopes for the successful management of transnational ethnic conflict.

The key issue in determining the success of any external guarantee is the commitment of the external powers. External guarantees are highly intrusive and, to the extent they seek to change relations among local groups, potentially expensive. They work only when the local parties to the conflict believe that the outside powers are resolved to enforce the ethnic contract in a fair manner into the indefinite future. Thus the Clinton administration's plan to level the military playing field in Bosnia by helping the Muslims to rearm and to retrain their soldiers will not only prove financially costly but, if not administered carefully, may jeopardize the impartiality of the peace-keepers and possibly lead to escalating violence. Even so, the behavior of the external powers today is not the crucial factor; rather, a more fundamental question is whether the warring parties or potential combatants believe the external powers will be there to protect them tomorrow--and in the days and years after that. Absent a belief in the fair-mindedness and stamina of the external powers, intervention in any form will fail to mitigate the conflict.

Unfortunately, both countries with strong and weak interests in intervening often find themselves unable to offer credible guarantees. Countries with strong interests in the conflict--those who are vitally affected by the fighting or the outcome either tend to be partisan or are perceived by the combatants as partisan--as was the case with France's intervention in Rwanda in 1994. One or both sides to the conflict, therefore, will doubt the willingness or ability of the outside power to enforce the new ethnic contract in an evenhanded and fair manner. To choose an extreme hypothetical example that makes the point clearly, Serbia would not be regarded by the Croats and Muslims as an impartial enforcer of any new ethnic contract between themselves and the Bosnia Serbs--regardless of any promises by Serbia to remain neutral. Precisely because it has close ties to its co-ethnics across the border and strong interests in the outcome of the conflict in Bosnia, Serbia would be biased, or expected to be biased, in carrying out its guarantees. Similarly, having essentially taken sides in the same conflict, the United States and its NATO allies may have forfeited their status as credible guarantors; although many in the West worry about Russian involvement in peace-keeping operations in Bosnia, in part because they favor the Serbs, this may be the only way of checking the bias or feared bias of NATO and ensuring that the warring parties accept external oversight. The more biased the external guarantor, the less likely the sides will be to reach an effective and enforceable agreement. However, when outside powers have interests in a stable outcome, rather than in the victory or loss of either side, they may be perceived by all as a fair-minded facilitator. Britain's role in Zimbabwe is a positive example of an interested party able to work with a coalition of external mediators to push negotiations ahead to a successful outcome.

Countries with weak interests in the conflict, on the other hand, will tend to lack or will be perceived as lacking the political stamina to enforce any new ethnic contract into the indefinite future. The Eastern African countries making up the Inter-Governmental Authority on Drought and Development have launched mediatory initiatives in recent years in the Sudan, but they lack the economic and military capabilities to enforce an agreement between the combatants at this juncture. Likewise, the United States was unwilling to bear any substantial cost in human lives to guarantee the peace in Somalia--although some casualties are probably inevitable in all peacemaking operations. When the interests of the outside power in the conflict is weak, each of the warring parties is more likely to believe that the external guarantor will default on its commitment--and will therefore continue fighting. An external guarantee that the parties fear will soon evaporate is no guarantee at all.

There are many reasons why states might possess only weak interests in guaranteeing a new ethnic conflict even when appearances suggest the opposite. First and foremost, political instability abroad is typically broad but shallow in its effects; conflicts that may diffuse across borders or instability that threatens to breed centers of international terrorism and crime affect all countries, but none intensely. It is a collective bad, if you will, subject to the usual tendency to free ride on the efforts of other states (Olson 1965). While the international community might benefit significantly from ending the conflict, it is in no country's interests to pay unilaterally the substantial costs of resolving the dispute by acting as the world's policeman. This is one plausible interpretation of the hesitancy of the United States in taking a leadership role in Bosnia. In this view, Presidents Bush and Clinton held back hoping that the Europeans would step forward and carry the financial and military burden; only when the Europeans proved themselves unprepared to assume the costs did the United States take the lead. Second, false understandings of the nature of ethnic conflict can also increase the reluctance of states to become involved. Analysis or political rhetoric that portrays ethnic conflicts as primordial contests of centuries long duration subtly raise the expected costs of any possible intervention and reduce the expected benefits; natural or inevitable conflicts appear to be harder to resolve than other, more "man made" conflicts. In this case, the widespread acceptance of the view that the present Balkan conflict is based upon ancient hatreds is both an impediment to action and an excuse for inaction.

Weak interests, however, need not preclude states from becoming involved in ethnic conflicts. Especially after the Cold War, states are freer to respond to international events according to their own internal, domestic political whims and fancies (Lake forthcoming). Humanitarian intervention is driven by many of the same information flows that drive the process of diffusion. As John Chipman writes

The flow of information and the effect of television mean, especially for the great powers who retain power projection capabilities, that public opinion might support the deployment of force for preventive diplomacy humanitarian aid peace-keeping, or pacification, even without the national interests of the 'expeditionary state being remotely engaged. This impels states to become involved in the parochial quarrels of others. Leaders may know that their citizens will not support heavy casualties where no great strategic interest is at stake, but often the demand 'to do something' cannot be refuted by reasoned argument (Chipman 1993, 239).

Public exposure to information about human tragedies, thus, may induce an emotional response sufficient to draw an outside power into a conflict. These same emotions, however, are likely to prove insufficient to sustain interest once the full costs of the mission become apparent.

Weak commitments produce ambiguous policies that may, in the end, exacerbate rather than resolve conflicts. Public commitments encourage the weaker party to believe that the external power supports it, thereby prompting the group to fight on and hold out for a better deal than its position on the battlefield warrants (Djilas 1995, 102). Ambiguity and vacillation, however, may simultaneously persuade the stronger party that the external power does not possess sufficient stamina, and that this local actor too may improve its position by continuing to fight. Indeed, the strong party may even target the external power in an attempt to raise the latter's costs of intervention and force its withdrawal from the conflict. This ambivalent commitment is the true tragedy of the current United States policy in the Balkans. One of the most important lessons from this analysis is that if external powers are going to intervene in ethnic conflicts, they must do so in a way that is credible to the groups involved.

Although easier to organize, unilateral external action represents an intermittent and somewhat unreliable means of responding to ethnic violence in the contemporary period. Even if the French did enter Rwanda in the later stages of the 1994 crisis, this was a belated and atypical initiative. More common is the case of the Sudan where a decade of brutal encounters has engendered mediatory maneuvers but no decisive military intercession between the warring parties. As noted above, with the end of the Cold War, elites in the developed, Western countries are looking inward, reluctant to take on the challenge of risky humanitarian interventions in distant countries. Even in areas adjacent to the NATO powers, like Bosnia and Macedonia, Western Europeans and Americans are not prepared to act as peace enforcers, only as mediators and peace-keepers.

Some of the problems of political legitimacy and burden-sharing are less difficult when the intervention is mounted by multilateral regional or global organizations. The problems of coalition building, financing, and domestic public opinion remain, but the collective nature of the intervention tends to blunt opposition at home and abroad. To some extent, the United Nations' action to insulate the Kurdish area of Iraq from Saddam Hussein's retaliatory measures represents a pioneer effort to protect a vulnerable nationality group. Another unique initiative was the entrance of the Economic Community of West African States (ECOWAS) onto the Liberian scene in 1990, where it prevented the National Patriotic Front of Liberia's takeover of Monrovia, and with that city's fall, a possible threat to the Krahn and Mandingo peoples living there.

The difficulties of funding the ECOWAS effort and building up a larger force to ensure a sustained commitment, however, point up some of the larger problems associated with multilateral peace enforcement. International organizations, after all, are state-based institutions that rely upon their members to support their activities. They cannot cope with the burgeoning problems of intrastate conflict unless their members pay their dues promptly and provide the necessary military equipment, logistical support, and trained manpower (Boutros-Ghali 1995, 11). Smaller states such as Ghana have expressed their uneasiness about continuing what seems an indefinite commitment of scarce resources in Liberia (Ofosuhene and Osafo-Mensah, 1994, 1); large and wealthy states, such as the United States, are slow to meet their financial responsibilities and appear intent on reducing their contributions to United Nations peace-keeping and peace-enforcing activities in the years ahead. To achieve "acceptable costs," international organizations will, without doubt, have to make some difficult choices on which humanitarian interventions to undertake (Haass 1994, 27). Only in special cases, then, can multilateral coercive interventions be expected to have anything but a limited impact.

Finally, external interventions are not likely to solve the underlying problems associated with ethnic insecurity and violence or to change the local balance of power between the parties. Interveners can attempt to reinforce international norms and enforce agreements, but in the final analysis, conflict management requires an effort by the local parties to work out acceptable rules of interaction. This is not to say that containing conflict is not an important achievement, only that containment by itself is not a permanent solution. External intervention does not solve the dilemmas above or create a desire among the parties to restore normal relations.

Third-party Mediation

External mediators can encourage adversaries to reconsider their alternatives and to opt for peaceful, negotiated solutions to their differences. As noted above, only one-fourth of intrastate wars in the post-World War II period have been settled through mediation and negotiation, a percentage that reflects the strength of state and ethnic identifications and the difficulty that intrastate groups have in compromising with their adversaries. As Charles William Maynes (1995, 111) contends: "There is . . . a desperate quality to civil wars that makes them particularly hard to control once they start." Even so, as wars reach a mutually hurting stalemate and leaders on both sides perceive an "intolerable situation" with little expectation of military victory, there is a chance that the fatigued parties will come to the table and bargain in earnest (Zartman 1985, 232). Despite the emotionalism and organizational imperatives surrounding civil wars, a number of them--including Cambodia, Nicaragua, Zimbabwe, Angola, Namibia, Mozambique, and possibly now Chechnya, Bosnia, Palestine, and Israel--have been or are close to being settled by means of negotiations. One must not anticipate too much from mediatory efforts, but a grim, self-fulfilling outlook is also not appropriate.

The difficulties normally associated with mediation are compounded by the obstacles to implementation. In the 1980s and 1990s, several laboriously negotiated agreements have been signed only to see them fall apart at the implementation stage--for example, in Ethiopia and Eritrea (1962), Sudan (1982), Uganda (1985), Angola (1975, 1992), and Rwanda (1994). A large part of the responsibility for these failures lies with adversary parties and their inability to make credible and reliable commitments. Their distrust of one another's intentions was so deep that the peace agreement crumbled when ambiguity opened the way to renewed confrontation. However, part of the explanation for the failure of agreements is attributable to the international community and its unwillingness to provide the mediators with the needed economic, logistical, police, and military support to oversee the processes of disarmament, integration of the armed forces, repatriation of refugees, and holding of general elections. In addition, the guarantees made to one or more rivals by foreign governments and multilateral organizations may come to lack credibility in the eyes of local actors if domestic publics lose interest in far-off conflicts and retreat from commitments made at the high point of the struggle.

International interventions, then, may prove necessary but not sufficient to ensure long term commitments to peaceful relations. Where confidence-building measures fail to overcome _ ethnic fears and iterated bargaining processes stall, the international community becomes by default the guarantor of last resort for desperate peoples caught up in intense civil wars. Yet even here, the safety net is frail, dependent on the goodwill of local adversaries and the sustained concern of foreign governments and publics. In such an environment, all may be born free, but find themselves chained by fear.

Thwarting the International Spread of Ethnic Conflict

Both confidence-building measures and external interventions are imperfect instruments for ensuring ethnic peace. Some measure of ethnic conflict may be inevitable. Confidence-building measures and external interventions are also targeted at particular ethnic conflicts--both actual and possible. They work primarily by influencing internal conditions and, thus, the probabilities of intrastate conflicts. As ethnic conflict requires fertile soil in which to grow, this emphasis on local conditions is necessary and appropriate even in a project such as ours that focuses on the international spread of ethnic conflict. By reducing the chances of violence in particular locations, confidence-building measures and interventions affect indirectly the international processes of contagion and expansion discussed above. Nonetheless, they do not address these processes directly. Even if individual ethnic conflicts can be only imperfectly managed, is it possible to influence the extent to which they are likely to spread across national borders? What can be done to limit the processes of diffusion and escalation?

With the exception of the first and most direct route, where conflict abroad changes the existing balance of ethnic power at home, diffusion occurs largely by changing the political strategies and, more fundamentally, the information and beliefs of ethnic groups. Ethnic conflict abroad may embolden groups to make more extreme demands, prompt groups to expect others to make more extreme demands, and alter the evaluations groups make about the distribution of ethnic power within their societies, the efficacy of their existing safeguards, and the costs of using violence. Because of the selection bias in the media and the self-aggrandizing strategies of egoistic politicians, conflict is more likely than peace to diffuse through these channels.

Recognizing the importance of information flows and the beliefs leaders and groups hold about the intentions and abilities of others, the most direct and potentially most effective policy instruments available to the international community as it attempts to limit the diffusion of ethnic conflict is to ensure that objective, unbiased and balanced information is widely available to all ethnic groups at risk of being engulfed by conflict. In short, states and the international community in general can seek to offset the selection bias in the media and pronouncements of national political entrepreneurs and correct the distortions in beliefs that follow from it. By providing objective and balanced information, the international community can help to reduce the severity of both the between group and within-group strategic problems discussed above. In turn, this directly limits the likelihood that ethnic conflicts will diffuse around the globe. Mediation in specific conflicts often serves this same purpose. But wider communication of accurate information can also create a general climate that inhibits the spread of conflict, for it mitigates the sources of ethnic fear and insecurity.

This need for readily available, objective information suggests a new mission for the information arms of the United States government, built up during the Cold War but now coming under the budgetary knife, and possibly the United Nations. Agencies like Radio Free Europe, which provided information on economic and political trends that was simply unavailable in the communist countries of Eastern Europe and the Soviet Union, could now serve equally well in disseminating balanced reports on ethnic conflicts and, more important, successful cohabitations. Such objective information might have made a difference in Bosnia and Rwanda, although, by itself, better information is unlikely to have prevented either conflict in its entirety.

The international community in general, and its leading states in particular, can also act to limit the escalation of ethnic conflicts to new belligerents. States can aid in limiting the potential for escalation by reaffirming credibly their commitment to the rights of regional autonomy within states and, in certain circumstances, national self-determination outside existing states, but only if pursued and obtained through peaceful means (Kampelman 1993, xi). The divorce between the two halves of Czechoslovakia or between Ethiopia and Eritrea, widely accepted by other states, are exemplary in this regard.

Similarly, the international community must recognize and legitimate the need to protect minority rights within other countries; at the same time, it should reserve to its regional and international bodies the sole right to sanction coercive interventions, lest the pursuit of this principle expand rather than mitigate interstate conflicts. Protecting minority rights is a double edged sword. By limiting the rights of coethnics in neighboring states to intervene unilaterally, the potential for transforming local conflicts into regional conflicts is greatly reduced. The pull of ethnic alliances can be diminished. At the same time, this principle obligates the broader international community to assume responsibility for protecting these same minorities--as it has done for the Kurds in northern Iraq.

Neither of these principles for containing the escalation of ethnic conflict will, of course, work without a credible commitment by the community of states to punish transgressors and forcibly protect minorities if necessary. The problems of building support for the defense of principles are no different than those found in intervening in specific conflicts--although they may be even harder to solve in the abstract than in immediate cases of ethnic violence. This highlights again the problems and limits of forcible interventions.

Conclusion

Most of the time, most ethnic groups live side-by-side in harmony with one another. Even in cases where ethnic minorities might otherwise be at risk, some states have been determined to promote stable ethnic relations on their own, and have made concessions on minority group inclusion, participation, autonomy, and access to resources. The effects of these adjustments to the grievances and insecurities of smaller or less powerful ethnic groups is, for the most part, to ease intergroup tensions. However, an awareness that regimes can always change their preferences and retract these concessions leaves minorities fearful of the future. Information failures, problems of credible commitment, the security dilemma, and the risk that ethnic activists and political entrepreneurs will polarize society lurk in the background of all ethnically divided polities. Conflict always remains a possibility.

When politicians lack legitimacy and a solid base of supporters, they may decide to outbid moderate competitors within their own ethnic group by making militant appeals to group interests. In situations of perceived group disadvantage, economic decline, hostile political memories, or widespread anomie, such appeals may awaken long dormant "malignant nationalisms" and lead to escalating violence (Van Evera 1994, 8). Irrespective of whether ethnicity is an invented category or not, a leader's call for action to a people having some form of shared identity can awaken a consciousness of common grievances and a desire to rectify these perceived wrongs. The group can rapidly attract supporters not wanting to be left out, enabling its leaders to use their new influence and power to force centrist politicians to placate them on key issues. Provided that the aggressive nationalists continue to broaden their appeal and keep their movement unified, they can, in a multiethnic context, manipulate emotions, expand their demands, and cause violence to surface. In worst case scenarios, as in Bosnia, Burundi, and Rwanda, a progression from intermittent to systematic violence can occur, creating new destruction and enmities that complicate a return to normal relations.

Particularly where an element of local anarchy is present and the state is weak, a spiral of negative encounters that leads to violence remains a very real possibility. Information failures occur as the state loses its ability to arbitrate between factions and groups hold back information and suspect others of doing the same. Problems of credible commitment arise as ethnic contracts collapse and groups fear that others will not uphold their promises. Incentives to preempt drive groups to fight first and seek the basis for compromise later. In multi-ethnic polities with past histories of conflict and distrust, the social fabric can be very weak and easily tom apart.

In their fear, political minorities, recognizing the state's limited capacity to maintain order and its potential to act repressively in some cases, look outward to the international community for protection. They hope the international community will restore a balance of power and hence make systematic, state-sanctioned ethnic killing too costly for the hard-line majority leadership to condone. The international response, however, has all too often been feeble and unconvincing. Western countries condemned ethnic cleansing in Bosnia or Sudan and appealed for negotiated settlements; they have also mediated peace agreements between the adversaries with some success in recent years. However, as these situations continue to degenerate, these same external powers are inhibited by calculations of prudence and fail to intervene more forcefully. The effect is to leave the aggressive nationalists with an incentive to stay their murderous course. In a world community in which domestic publics oppose external interventions and many states are inclined to free ride on the efforts of others, only fragile safety nets are held out for the vulnerable.

What can be done? It is important to start by recognizing that Western and other countries have strong interests in a stable world, one in which their people can safely travel and conduct business and population movements and flows of refugees do not cause new tensions and instabilities. It is also important to acknowledge the limits of international action. Yet the growing potential for ethnic conflicts around the globe requires new ideas and concerted efforts by regional and global actors to prevent future disasters.

The increase in ethnic and national confrontations requires new thinking about the diffusion of destructive ideas on interethnic relations across international borders. New norms on dealing with dangerous radio broadcasts, diaspora communities living in neighboring countries, national and state boundaries, or secession attempts might go a long way toward making some of the worst cases of group tension more manageable. Also, the acceptance of decentralization schemes within the state, including subregional autonomy and confederation, might provide incentives for pragmatic cooperative relationships. To be sure, because these schemes are a response to ethnic fears and build a certain amount of social tension into their structures, they are likely to provoke intense anxieties in majority circles. But with a concerted effort by international donors to link economic assistance to the acceptance of these institutional accommodations, leaders in majority communities could come to see their wider interests served by striking a deal with local adversaries.

References

Adam, Heribert and Kogila Moodley. 1993. "South Africa: The Opening of the Apartheid Mind," in The Politics of Ethnic Conflict Regulation, ed. by John McGarry and Brendan O'Leary (New York: Routledge).

Amnesty International. 1995. 'The Tears of Orphans': No Future Without Hunan Rights (New York: Amnesty International).

Anderson, Benedict. 1983. Imagined Communities: Reflections on the Origins and Spread of Nationalism (London: Verso).

Appiah, Kwame A. 1992. In My Father's House: Africa in the Philosophy of Culture (New York: Oxford University Press).

Arutiunov, Sergei. 1995. "Ethnicity and Conflict in the Caucasus," in Ethnic Conflict and Russian Intervention in the Caucasus, ed. by Fred Wehling, IGCC Policy Paper No. 16. (IGCC Project on The International Spread and Management of Ethnic Conflict).

Bates, Robert H. 1983. "Modernization, Ethnic Competition, and the Rationality of Politics in Contemporary Africa," in States Versus Ethnic Claims, ed. by Donald Rothchild and Victor A. Olorunsola (Boulder: Westview).

Bates, Robert H. and Barry R. Weingast. 1995. "Rationality and Interpretation: The Politics of Transition. Paper presented to the Annual Meeting of the American Political Science Association, Chicago, IL, Aug. 31--Sept. 3.

Boutros-Ghali, Boutros. 1992. An Agenda for Peace (New York: United Nations).

Boutros-Ghali, Boutros. 1995. "Supplement to an Agenda for Peace: Position Paper of the Secretary-General on the Occasion of the Fiftieth Anniversary of the United Nations," United Nations, General Assembly, 50 sess., A/50/60 (January 3).

Brass, Paul, ed. 1985. Ethnic Groups and the State (London: Croom-Helm).

Brown, Michael E., ed. 1993. Ethnic Conflict and International Security (Princeton: Princeton University Press).

Brubaker, Rogers. 1995. "National Minorities, Nationalizing States, and External National Homelands in the New Europe." Daedalus (Spring). Also IGCC Project on The International Spread and Management of Ethnic Conflict.

Carment, David. 1993. The International Politics of Ethnic Conflict: The Interstate Dimensions of Secession and lrredenta in the Twentieth Century, A Crisis-Based Approach. Ph.D. dissertation. McGill University, Montreal, Canada.

Carment, David. 1994. "The Ethnic Dimension in World Politics: Theory, Policy and Early Warning." Third World Quarrerly, 15, 4, 5512.

Carment, David and Patrick James. 1995. "Internal Constraints and Interstate Ethnic Conflict: Toward a Crisis-Based Assessment of Irredentism." Journal of Conflict Resolution, 39, 1, 82-109.

Carr, Caleb. 1993. "The Consequences of Somalia," World Policy Journal 10, 3 (Fall), 1-4.

CBS News, 1994. (June 23).

Chipman, John. 1993. "Managing the Politics of Parochialism," in Brown, pp. 237-63.

Cigar, Norman. 1995. Genocide in Bosnia: The Policy of 'Ethnic Cleansing,' (College Station: Texas A&M University Press).

Connor, Walker. 1994. Ethnonationalism: The Quest for Understanding (Princeton: Princeton University Press).

Cooper, Robert and Mats Berdal. 1993. "Outside Intervention in Ethnic Conflicts," in Brown, 181-205.

Damrosch, Lori Fisler, ed. 1993. Enforcing Restraint: Collective Intervention in Internal Conflicts (New York: Council on Foreign Relations).

Djilas, Aleksa. 1995. "Fear thy Neighbor: The Breakup of Yugoslavia," in Nationalism and Nationalities in the New Europe, ed. by Charles A. Kupchan (Ithaca: Cornell University Press), 85-106.

Dominguez, Virginia R. 1989. People as Subject, People as Object: Selfhood and Peoplehood in Contemporary Israel (Madison: University of Wisconsin Press).

Esman, Milton J. 1994. Ethnic Politics (Ithaca: Cornell University Press).

Fearon, James D. 1993. "Ethnic War as a Commitment Problem." Unpublished ms., University of Chicago, June.

Fearon, James D. 1995a. "Rationalist Explanations for War." International Organization 49, 3 (Summer 1995), 379-414.

Fearon, James D. 1995b. "Commitment Problems and the Spread of Ethnic Conflict." IGCC Project on The International Spread and Management of Ethnic Conflict.

Federal Republic of Nigeria. 1987. Report of the Political Bureau (Lagos: Federal Government Printer).

Frei, Daniel. 1976. "Conditions Affecting the Effectiveness of International Mediation," Papers of the Peace Science Society (International), vol. 26.

Friedman, Townsend. 1995. "Statement ... Before the Joint Hearing of the Senate Foreign Relations Subcommittee on Africa and the House International Relations Subcommittee on Africa on Rwanda and Burundi," April 5.

Fortna, Virginia P. 1995. "Success and Failure in Southern Africa: Peacekeeping in Namibia and Angola," in Beyond Traditional Peacekeeping, ed. by Donald C. F. Daniel and Bradd C. Hayes (Houndsville: McMillan).

Gagnon, V. P., Jr. 1994/95. "Ethnic Nationalism and International Conflict: The Case of Serbia," International Security 19, 3 (Winter), 1366.

George, Alexander L. 1993. Bridging the Gap: Theory and Practice in Foreign Policy (Washington, D.C.: The United States Institute of Peace Press).

Glazer, Nathan and Daniel P. Moynihan. 1975. Ethnicity: Theo7y and Experience Cambridge: Harvard University Press).

Glenny, Misha. 1992. The Fall of Yugoslavia (New York: Penguin Books).

Haass, Richard N. 1994. "Military Force: A User's Guide," Foreign Policy, no. 96 (Fall), 21--37.

Halperin, Sandra. 1995. "The Spread of Ethnic Conflict in Europe: Some Comparative-Historical Reflections." IGCC Project on The International Spread and Management of Ethnic Conflict.

Hardin, Russell. 1995 . One for All: The Logic of Group Conflict (Princeton: Princeton University Press).

Hill, Stuart, Donald Rothchild, and Colin Cameron. 1995. "Tactical Information and the Diffusion of Protests." IGCC Project on The International Spread and Management of Ethnic Conflict.

Hirsch, John L. and Robert B. Oakley. 1995. Somalia and Operation Restore Hope: Reflections on Peacemaking and Peacekeeping (Washington, D.C.: United States Institute of Peace Press).

Horowitz, Donald L. 1985. Ethnic Groups in Conflict (Berkeley: University of California Press).

Issacs, Harold. 1975. Idols of the Tribe: Group Identity and Political Chang (New York: Harper and Row).

Jackson, Robert H. and Carl G. Rosberg. 1982. "Why Africa's Weak States Persist," World Politics 35, 1, 1-24.

Jentleson, Bruce W. 1995. "Preventive Diplomacy and Ethnic Conflict: An Idea in Search of a Strategy." IGCC Project on The International Spread and Management of Ethnic Conflict.

Jervis, Robert. 1978. "Cooperation Under the Security Dilemma," World Politics 30, 2 (January).

Kampelman, Max M. 1993. "Foreword," in Damrosch, vii--xii.

Kaplan, Robert. 1993. Balkan Ghosts A Journey through History (New York: St. Martin's Press).

Karl, Terry L. 1986. "Petroleum and Political Pacts: The Transition to Democracy in Venezuela." In Transitions from Authoritarian Rule: Latin America, ed. by Guillermo O'Donnell, Philippe C. Schmitter, and Laurence Whitehead (Baltimore: Johns Hopkins University Press).

Keller, Edmond J. 1995. "Transnational Ethnic Conflict in Africa." IGCC Project on The International Spread and Management of Ethnic Conflict.

Norman Kempster and Art Pine. 1995. "Bosnia Peace Treaty Approved; Use of U.S. Troops is Sought," Los Angeles Times, November 22, A1 and A16.

Krasner, Stephen D. 1993. "Westphalia and All That." In Ideas and Foreign Policy: Beliefs, Institutions, and Political Change, ed. by Judith Goldstein and Robert O. Keohane (Ithaca: Cornell University Press).

Krasner, Stephen D. 1995. "Minority Rights and the Westphalian Model." IGCC Project on The International Spread and Management of Ethnic Conflict.

Kuran, Timur. 1995. "Ethnic Dissimilation and its Global Transmission." IGCC Project on The International Spread and Management of Ethnic Conflict.

Lake, David A. Forthcoming 1996. "Toward a Democratic Foreign Policy." IGCC Policy Paper no. 23.

Lake, David A. 1996. "Anarchy, Hierarchy, and the Variety of International Relations," International Organization 50 (Winter): 1-33.

Lapidus, Gail W. and Renee de Nevers, ed. 1995. Nationalism, Ethnic Identity and Conflict Management in Russia Today (Stanford: Stanford University Center for International Security and Arms Control).

Lemarchand, Rene and David Martin. 1974. Selective Genocide in Burundi, no. 20 (London: Minority Rights Group).

Lepgold, Joseph. 1995. "Is Anyone Listening? International Relations Theory and the Problem of Policy Relevance," paper presented at the 1995 Annual Meeting of the American Political Science Association, Chicago, August 31-September 3.

Licklider, Roy. 1995. "The Consequences of Negotiated Settlements in Civil Wars, 1945-1993," American Political Science Review 89, 3 (September).

Lijphart, Arend. 1967. The Politics of Accommodation: Pluralism and Democracy in the Netherlands (Berkeley: University of California Press).

Lipschutz, Ronnie and Beverly Crawford. 1995. "Ethnic Conflict Isn't," IGCC Policy Brief no. 2 (March).

Lund, Michael S. 1995. "Underrating 'Preventive Diplomacy'," Foreign Affairs 74, 4 (July/August).

Mattes, Robert B. 1993. "Beyond 'Government and Opposition': An Independent South African Legislature," Poltikon 20, 2 (December).

Maynes, Charles W. 1995. "Relearning Intervention," Foreign Policy no. 98 (Spring), 96-113.

McWhirter, Cameron and Gur Melamede. 1992. "Ethiopia: The Ethnicity Factor," Africa Report 37, 5 (September/October).

Meadwell, Hudson. 1989. "Cultural and Instrumental Approaches to Ethnic Nationalism," Ethnic and Racial Studies 12, 309-28.

Moore, Will H. and David R. Davis. 1995. "Does Ethnicity Matter? Ethnic Alliances and International Interactions." IGCC Project on The International Spread and Management of Ethnic Conflict.

Moynihan, Daniel Patrick. 1993. Pandaemonium: Ethnicity in International Politics (New York: Oxford University Press).

Newbury, Catherine. 1995. "Background to Genocide: Rwanda," Issue 23, 2.

Newland, Kathleen. 1993. "Ethniconflict and Refugees," in Brown, 143-63.

Obasanjo, Olusegun. Forthcoming. "The African Region and the Cold War: A Balance Sheet," in Africa and the New International Orde1: Studies of State Sovereignty and Regional Security ed. by Edmond J. Keller and Donald Rothchild (Boulder: Lynne Rienner).

Ofosuhene, Kwabena and Amma Osafo-Mensah. 1994. "Lay Down Your Arms," Daily Graphic (Accra), (November 22).

Olson, Mancur. 1965. The Logic of Collective Action (Cambridge: Harvard University Press).

Olson, Mancur. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities (New Haven: Yale University Press).

Oye, Kenneth A. 1985. "Explaining Cooperation Under Anarchy: Hypotheses and Strategies," World Politics 38,1 (October), 1--24.

Osgood, C. E. (1962). An Alternative to War or Surrender (Urbana: University of Illinois Press).

Percival, Valerie and Thomas Homer-Dixon. 1995. Environn7ental Scarcity and Violent Conflict: The Case of Rwanda. American Association for the Advancement of Science, Washington, D.C.

Pesic, Vesna. 1994. "Bellicose Virtues in Elementary School Readers," in Warfare, Patriotism, Patriarchy: The Analysis of Elementary School Textbooks, ed. by Ruzica Rosandic and Vesna Pesic (Belgrade: Centre for Anti-War Action).

Pillar, Paul R 1993. Negotiating Peace: War Termination as a Bargaining Process (Princeton: Princeton University Press).

Posen, Barry R. 1993. "The Security Dilemma and Ethnic Conflict," in Brown, 103--24.

Przeworski, Adam. 1991. Democracy and the Market (Cambridge: Cambridge University Press).

Rapoport, David C. 1995. "Observations on the Importance of Space in Violent Ethno-Religious Strife." IGCC Project on The International Spread and Management of Ethnic Conflict.

Rieff, David. 1995. Slaughterhouse: Bosnia and the Failure of the West (New York: Simon & Schuster).

Rogowski, Ronald. 1988. Commerce and Coalitions: How Trade Affects Domestic Political Alignnments (Princeton: Princeton University Press).

Rothchild, Donald. 1986a. "Hegemonial Exchange: An Alternative Model for Managing Conflict in Middle Africa," in Ethnicity, Politics, and Development, ed. by Dennis L. Thompson and Dov Ronen (Boulder: Lynne Rienner).

Rothchild, Donald. 1986b. "Interethnic Conflict and Policy Analysis in Africa," Ethnic and Racial Studies 9, 1 (January).

Rothchild, Donald and Alexander J. Groth. 1995. "Pathological Dimensions of Domestic and Intemational Ethnicity," Political Science Quarterly 110,1 (Spring).

Rothchild, Donald and Caroline Hartzell. 1993. "The Peace Process in the Sudan, 1971--1972," in Stopping the Killing: How Civil Wars End, ed. by Roy Licklider (New York: NYU Press).

Ruggie, John G. 1994. "The U.N.: Between Peacekeeping and Enforcement," Foreign Affairs: Agenda 1994 (New York: Foreign Affairs).

Saideman, Stephen M. 1995. "Is Pandora's Box Half-Empty or Half-Full? The Limited Virulence of Secessionism and the Domestic Sources of Disintegration." IGCC Policy Paper no. 18.

Sciolino, Elaine. 1995. "U.S. Envoy Advises Caution on Bosnia Peace Agreement," New York Times, September 13.

Simons, Anna. 1995. "Democratization and Ethnic Conflict: The Kinship Connection." IGCC Project on The International Spread and Management of Ethnic Conflict.

Sisk, Timothy D. 1996. Living Together? International Intervention and Power Sharing in Ethnic Conflicts (Washington, D.C.: United States Institute of Peace Press).

Sisk, Timothy D. 1995. Democratization in South Africa: The Elusive Social Contract (Princeton: Princeton University Press).

Sisk, Timothy D. 1993. "South Africa Seeks New Ground Rules," Journal of Democracy 4, 1 (January).

Siverson, Randolph and Harvey Starr. 1991. The Diffusion of War: A Study of Opportunity and Willingness (Ann Arbor: University of Michigan Press).

Smith, Anthony D. 1986. The Ethnic Origins of Nations (New York: Basil Blackwell).

Smith, Anthony D. 1993. "The Ethnic Sources of Nationalism," in Brown, 27-41.

Snyder, Jack. 1993. "Nationalism and the Crisis of the Post-Soviet State," in Brown, 79-101.

Stedman, Stephen J. 1995a. "Alchemy for a New World Order: Overselling 'Preventive Diplomacy'," Foreign Affairs 74, 3 (May/June), 14-20.

Stedman, Stephen J. 1995b. "Mediation in Civil War," The International Dimensions of Internal Conflict, ed. by Michael E. Brown (Cambridge: M.I.T. Press).

Stedman, Stephen J. 1991. Peacemaking in Civil Wars: International Mediation in Zimbabwe, 1971980 (Boulder: Lynne Rienner).

Stein, Arthur A. 1990. Why Nations Cooperate: Circumstance and Choice in International Relations (Ithaca: Cornell University Press).

Steinberg, James B. 1993. "International Involvement in the Yugoslavia Conflict," in Damrosch, 27--75.

Steinberg, Stephen. 1981. The Ethnic Myth: Race, Ethnicity, and Class in America (New York: Atheneum).

Steinbruner, John. 1995. "Remarks," in U.S. Intervention in Ethnic Conflict, ed. by Fred Wehling, ICCC Policy Paper no. 12.

Sudan Democratic Gazette. 1994. No. 53 (October).

Touval, Saadia. 1982. "Managing the Risks of Accommodation," in Termination of Wars, ed. by Nissan Oren (Jerusalem: Magnes Press).

van den Berg, Ger P. and William B. Simons. Forthcoming. "The New Russian Political System: Its Features and How It Is Created," Review of Central and East European Law.

van den Berghe, Pierre. 1981. The Ethnic Phenomenon (New York: Elsevier).

Van Evera, Stephen. 1994. "Hypotheses on Nationalism and War." International Security, 18, 4, 5--39.

Wagner, Robert H. 1993. "The Causes of Peace," in Stopping the Killing: How Civil Wars End, ed. by Roy Licklider (New York: New York University Press).

Walter, Barbara F. 1985. "The Resolution of Civil Wars: Why Negotiations Fail." Unpublished paper.

Watts, Larry L. 1 99S. "Ethnic Tensions: How the West Can Help," World Policy Journal 12, 1 (Spring).

Weingast, Barry R. 1995. "Constructing Trust: The Political and Economic Roots of Ethnic and Regional Conflict." Unpublished ms., Stanford University.

Weinstein, Warren. 1972. "Conflict and Confrontation in Central Africa: The Revolt in Burundi, 1972," Africa Today 19, 4 (Fall).

Welsh, David. 1993. "Domestic Politics and Ethnic Conflict," in Brown, 430.

West Africa. May 25, 1981.

Williamson, Oliver. 1985. The Economic Institutions of Capitalism: Firms, Markets, and Relational Contracting (New York: Free Press).

Wippman, David. 1993. "Enforcing the Peace: ECOWAS and the Liberian Civil War," in Damrosch, 157-203.

Wittman, Donald. 1979. "How a War Ends: A Rational Model Approach," Journal of ConJ7ict Resolution 23 (December): 7433.

Woodward, Susan. 1995. Balkan Tragedy: Chaos and Dissolution After the Cold War (Washington, D.C.: Brookings).

Yankelovich Partners, 1994. (August 8).

Young, Crawford. 1993. The Rising Tide of Cultural Pluralism: The Nation-State at Bay? (Madison: University of Wisconsin Press).

Zartman, I. William. 1995. "Putting Humpty-Dumpty Together Again." IGCC Project on The International Spread and Management of Ethnic Conflict.

Zartman, I. William. 1985. Ripe for Resolution: Conflict and Intervention in Africa (New York: Oxford University Press).

*: Acknowledgments. The IGCC Project on the International Spread and Management of Ethnic Conflict was made possible by the generous financial support of the Pew Charitable Trusts. We thank Barbara Bunerton of IGCC for her administrative and conference support. We are indebted to Fred Wehling, IGCC Coordinator of Policy Research, for his numerous contributions to the project. Finally, we thank the members of the IGCC Working Group, and especially the authors of the papers commissioned for this project, who participated in six conferences over 18 months; as reflected in the following document, we have learned much from the papers and discussions. Back.

Note 1: Remarks to the IGCC Working Group on the International Spread and Management of Ethnic Conflict, October I, 1994. Back.

Note 2: For a discussion of the distributional implications of international economic integration, see Rogowski 1988. Back.

Note 3: This sub-section and the next draw heavily upon recent essays by James Fearon (1993 and 1995a), two of the best theoretical works on conflict between organized groups. Back.

Note 4: This is similar to market failures in technology. For firm A to assess accurately the technology offered by firm B, the latter must reveal fully the process for sale; after B reveals the technology to A, however, A no longer has any incentive to purchase it. In any arms length sale, then, B will not reveal fully the technology and A will only be willing to pay less than what it is "truly" worth. Back.

Note 5: The term "ethnic contract" was, we believe, coined by Leonard Binder at the first meeting of the IGCC Working Group on the International Spread and Management of Ethnic Conflict, May 114, 1994. On relational contracting more generally, see Williamson (1985); for an application to interstate relations, see Lake (1996). Back.

Note 6: Aleksa Djilas (1995, 99) argues that the communist party served as the primary safeguard in Yugoslavia, largely through coercion and repression, and that the defeat of the party in the I990 elections left a political vacuum. He faults the party for not developing stable institutions that could have regulated relations among the republics national groups and protected their political, cultural, and territorial rights. Since Bosnia's Parliament, courts, press, and policy, had no authority as impartial institutions, affiliation with one's national group emerged as the only source of protection, whether of one's human rights or physical security." Weingast (1995, 16 (fn. 17) and 17), on the other hand, credits Marshall Tito for constructing a set of veto mechanisms institutionalizing trust among the groups. Back.

Note 7: Beliefs are used here in their game theoretic sense to refer to the conditional probably of an actor holding one set of preferences (intentions, in the text, payoffs from a game, more formally) rather than another. Beliefs are formed subjectively by actors, largely on the basis of past interactions. Back.

Note 8: Simons (1995) argues that democracy is unworkable in an environment where genealogy rules. Back.

Note 9: The security dilemma has a long pedigree in international relations. Jervis (1978) gave the concept its modern form. Posen (1993) was the first to apply it to ethnic relations. Back.

Note 10: In both Serbia and Rwanda, the radio proved a powerful weapon for broadcasting elite messages of hostility. Thus, Serbia's President, Slobodan Milosevic, used the state radio to mobilize his supporters for war, while in Rwanda, Hutu ideologues employed the privately-owned Radio Mille Collines to arouse their followers. Back.

Note 11: Please note, spread, diffusion, escalation and even contagion are often used as synonyms. We restrict the terms diffusion and escalation to the particular processes defined here and use the term spread for the more general tendency. Some analysts distinguish between positive diffusion, where an event increases the probability of a similar event occurring elsewhere. and negative diffusion, which reduces the probability. Similarly, some analysts differentiate between horizontal escalation, which increases the number of actors involved in the conflict, and vertical escalation, which increases the intensity or level of violence in the conflict. We are concerned here only with horizontal escalation and use the term to refer to an increase in the number of actors. Our analysis suggests little about the level of violence in the conflict. Back.

Note 12: Both Fearon (1995b) and Saidernan (199S) consider this domestic rather than international contagion. Admittedly, where one should draw the line between domestic and international is ambiguous. Refugee endows have also been important in the former Yugoslavia (Steinberg 1993, 53 ). Back.

Note 13: Rational individuals will, of course, recognize the possibility of selection bias in the media. In turn, they will discount information on conflict when they update their prior beliefs. Correctly discerning the degree of bias, however, can be extremely hard, particularly for poorly informed individuals in a highly partisan environment. The likelihood of drawing the correct conclusion from the information most readily available is very low. If individuals are not fully rational, biased information flows may have much more extreme effects. Back.

Note 14: Countries also become involved in foreign ethnic conflicts, occasionally but not necessarily through international organizations, as peacekeepers or peace enforcers. As discussed in more detail below, the distinction between belligerents and peacekeepers in one form or another is often hard to sustain in practice (Ruggie 1994 99). Peacekeepers do not always remain impartial and can easily blend over into belligerents. For every Cyprus, where the United Nations forces have remained neutral there are the Somalias and Liberias, where UNOSOM Il and ECOWAS, respectively, stepped over the border into belligerency. Nonetheless, the initial intent of the countries, at least, and the processes of involvement differ between these two primary routes. The escalation of ethnic conflicts to include additional belligerents is discussed here. Peace-keeping and peace enforcement are discussed below under the question of management. Back.

Note 15: Diversionary wars are also called "conflict transformations;" Carment (1994) and Carment and James (1995). Back.

Note 16: Milosevic's role in the October 1995 negotiations was important, because it carried with it an implied threat: if the Bosnian Serbs refused to be more accommodating at the bargaining table, their Serb kinsmen across the border could further reduce their military support. Back.