CIAO

email icon Email this citation


Civil War and the Security Dilemma

Institute on War and Peace Studies

February 1997

Introduction

Jack Snyder and Robert Jervis

Since the end of the Cold War, a series of costly civil wars, many of them ethnic conflicts, have dominated the international security agenda. The international community, often acting through the United Nations or regional organizations like NATO, has felt compelled to intervene with military forces in many of these conflicts, including the four cases that comprise the heart of this study: the former Yugoslavia, Somalia, Cambodia, and Rwanda. The mixed record of partial successes, failures, and in some cases counterproductive interventions suggests an urgent need to extract lessons from these experiences with a view toward developing a better conceptual framework to guide future policy choices. 1

Seeking a tool to help assess these lessons, we examine whether the concept of the security dilemma, widely used to explain conflict among states in the anarchical international system, provides analytical insight into civil wars as well. 2 A security dilemma is a situation in which each party's efforts to increase its own security reduce the security of the others. This situation occurs when geographical, technological, or other strategic conditions render aggression the most advantageous form of self-defense. Thomas Hobbes himself pointed out that the state provides a buffer between the individual and the dangers of anarchy: because individuals are more vulnerable than are states, the security dilemma is likely to be more severe in civil than in international anarchy. When state authority disintegrates, patterns more familiar in international than in domestic politics move to the fore.

A full understanding of the role of security fears in causing and perpetuating civil war requires more than just a knowledge of the threatening, anarchical situation facing the combatants. Behavior under the security dilemma is shaped not simply by the strategic situation, but also by the participants' perceptions of that situation and their expectations of each others' likely behavior in that situation. Or putting it in V.P. Gagnon's somewhat different terminology, the security dilemma is a "social construction." 3 Thus, interveners must confront not only the circumstances that constitute the security dilemma--namely, anarchy and offensive advantages--but also the ideas and social forces that produced the dilemma in the first place and that may reproduce it unless the interveners can neutralize them.

Another complicating factor is that, in virtually every case that the contributors to this volume examined, the security fears of the parties to civil conflict were intertwined with their predatory goals--that is, with exploitative desires that would not necessarily diminish if their security problems were solved. Moreover, it is often difficult to separate security-driven and predatory motivations, since long-term fears may drive security-seekers to take every opportunity to exploit others in an effort to build up their reserve of strategic resources, even when they face no immediate security threat. For both of these reasons, the strategies of interveners must consequently address both kinds of motives simultaneously. As most of the contributors to this volume stress, few contemporary civil conflicts are driven purely by security fears. Understood more broadly, however, the concept of the security dilemma can be a fruitful diagnostic tool for both scholars and practitioners who wish to understand and influence the dynamics of civil conflict.

The security dilemma and strategies of intervention in collapsed states

A number of the civil wars of the 1990s have followed in the wake of the collapse of the Soviet empire, the unraveling of other Communist states, or the governmental breakdown of so-called "failed states" in the Third World. Some states, such as Yugoslavia, broke up along ethnic lines, whereas others, such as Somalia or Tajikistan, collapsed into a more general disorder, where clans, regions, patronage networks, or other groups that could command some guns contended for the control of economic resources and territory. These contemporary civil conflicts seem to replicate the well-known pattern of Hobbesian competition for security in the "state of nature," where no sovereign power protects fearful individuals from each other. In this anarchical setting, prudent self-help may require preventive attacks to hedge against possible threats, even in the limiting case where everyone seeks only security. 4

Policy-makers, members of the informed public, and scholars commonly focus on the ways in which domestic conflict, often between different ethnic groups, leads to breakdown of authority. But causation can run in the opposite direction. The breakdown of internal authority and legitimacy may generate conflict among groups because each now fears that others will move against it. In Tajikistan, for example, conflict among ethnic groups and other kinds of factions was a consequence of the breakdown of state institutions rather than a cause of it. 5 It is not only that a strong state may be needed to keep ethnic hostility in check, but also that the absence of effective law enforcement tends to produce defensively-motivated behavior that magnifies or even creates conflict and violence. Thus, it is not surprising that David Laitin and James Fearon find that violent ethnic conflict within states is rare; 6 it often requires the disintegration of the state to set in motion the forces that lead to widespread ethnic violence.

When asked about the difference between the peaceful transition in South Africa and the bloodshed in the former Yugoslavia, Nelson Mandela replied: "In countries where innocent people are dying, the leaders are following their blood rather than their brains." 7 This parallels Sigmund Freud's famous response to Einstein's question as to the causes of war: "The Ideal condition of things would of course be a community of men who had subordinated their instinctual life to the dictatorship of reason." 8 Both these claims underestimate the extent to which violence can be the considered response to a situation that offers both danger and opportunity, especially when actors can be badly harmed if they are not able to protect themselves.

Three general types of prescription follow from the security-dilemma diagnosis. The first is the one recommended by Hobbes: establish a sovereign authority capable of enforcing a hegemonic peace upon all the fearfully contending parties. As Richard Betts has argued so persuasively, the question that drives civil conflict is "who rules?" Consequently, one solution to the security dilemma is to answer that question quickly and decisively. 9

The second solution is to devise a situation in which the parties can provide for their own security through strictly defensive measures. The security dilemma evaporates if the best defense is not a good offense, but simply a good defense: if so, everyone can be secure simultaneously. This can be achieved by making the contending groups more compact geographically, by evening out imbalances of power and dramatic shifts in relative power, by deploying weapons that are most useful in positional defense and least useful in attack (assuming that these can be distinguished), and by providing reliable monitoring of the military preparations of the contending groups. 10 As we will discuss further below, however, the cost of such policies is to reinforce the divisions within a society. At its most extreme, this can lead to splitting a country into separate parts; in other cases, it can solidify ethnic divisions and loyalties, making a state figuratively if not literally a federation.

A third solution is for the contending parties to lock themselves into an institutional framework that guarantees their mutual self-restraint, once they lay their weapons down. Of course, surrendering all power to a sovereign authority is one form of this solution, but in many cases this is unpalatable, because the only candidates for the role of hegemon are the feared and hated partisans of one of the warring groups. Consequently, the groups have an incentive to devise institutional arrangements that commit themselves to mutual cooperation without risking exploitation by the other.

Such arrangements may be based on either of two principles: delegation to neutral authorities, or balanced powersharing among the interested parties. 11 Pursuing a strategy of delegation, the contending groups may commit themselves to support the creation of powerful independent bodies that are neutral among the warring parties: e.g., a professionalized army, neutral courts, depoliticized police, and a technocratic economic bureaucracy. Thus, the strategy of delegation to neutrals is a more benign, consensual variant of the Hobbesian solution, creating a sovereign power that stands above the parties. Alternatively, the groups may commit themselves to a powersharing scheme, in which the warring groups retain their organizational coherence and manage many of their affairs internally, while agreeing to share legislative and bureaucratic power proportionally. In this scheme, the groups retain the corporate ability to look out for their own interests, yet institutionally commit themselves to a set of rules governing cooperation on matters that cannot be conveniently left to the separate parties. Powersharing has some characteristics of the defensive stalemate solution, in that the groups retain some of their capacity for self-help, but instead of physically separating the parties, powersharing seeks to establish a stable balance based on reciprocity in a partially integrated polity.

Both of these institutional solutions present conundrums from the standpoint of the security dilemma. The delegation strategy must overcome the warring groups' fear that they could be exploited by their enemies as they disarm themselves during the transfer of security functions to a neutral authority. The groups will also fear that the new authority might not be neutral, benign, and successful. Thus the groups may fear both that the new institutions will be too powerful and that they will not be powerful enough. 12 In the latter case, they will not be able to protect the group against its rivals; in the former case, those who dominate the institutions may put their own interests ahead of the community good, enriching and entrenching themselves at the expense of a well-ordered society.

Furthermore, institutions do not bind by magic; they normally do so by creating a pattern of behavior around which expectations converge. Yet these new institutions with no track record of shaping expectations must overcome the warring parties' powerful situational incentives to hedge against the risks of delegation. Such hedging, which might take the form of retaining significant armed force at the group's disposal, is likely to destroy the delegation strategy or turn it into a wary variant of the powersharing strategy.

Insofar as powersharing rests on the residual ability of the groups to act in their own self-defense, it is as likely to recreate the security dilemma as solve it. Powersharing reifies the contending groups and insures that all political mobilization must take place within the framework of the rival segments. Moreover, since powersharing eschews the full partitioning of the polity in favor of continued political and economic integration, it perpetuates the mutual interdependencies and vulnerabilities that heighten the security dilemma. 13

It is precisely for these reasons that civil wars are so rarely resolved by mutual agreement among the warring factions, unless powerful outsiders intervene to guarantee the settlement. 14 We will take up this question of prescriptions for interveners in the final section. But first we will take up two issues on which these prescriptions hinge in part: the relationship between predatory and security motives, and the perceptual and social dynamics that create or heighten the security dilemma.

The interaction of predatory and security motives

The purest type of security dilemma is a situation in which security is the overriding objective of all of the protagonists, yet attempts by one party to increase its security reduce the security of the others. At the opposite end of the spectrum, some conflicts may be driven entirely by the desire of one or both parties to exploit or dominate the other for reasons that would not diminish even if security were not in jeopardy. In between are a variety of situations in which security and non-security motives are both present. 15

It is important to keep the concepts of the security dilemma and predation separate and to understand that elements of each are present in almost every spefici situation, albeit in different ways and different proportions. In a pure security dilemma, even if all actors are not perfectly satisfied with the status quo (i.e., even if they would choose another outcome were there no costs and risks attached to doing so), all prefer a world of the status quo to one in which any actor can use force to seek change. In other words, all would prefer to end anarchy if effective security guarantees were possible. This is not the case when actors are predators. Here the desire to expand and the willingness to run risks to do so is great enough so that a freezing of the status quo would not be acceptable.

No individual case is ever entirely of one type or the other. Actors often feel they need to expand in order to be secure. Sometimes such beliefs are rationalizations for more purely predatory drives; at other times they are not, and it is extremely difficult for later analysts, let alone contemporary observers, to tell which is which. Furthermore, unless such observers can provide the actor with alternative means to security, they must treat the actor as a predator in this situation, irrespective of its "true"motives. It is also the case that predators are hardly immune from security worries. In Rwanda, for example, extremists in the Hutu-dominated government were fearful not only that a Tutsi victory or powersharing agreement would end their ability to exploit the state for economic gain, but also that they would be powerless to fend off demands for accountability for their past genocidal crimes. 16 This case, like almost all real cases, falls somewhere between the two ideal types of security-driven and predatory aggressors.

We are mindful of the risk of concept stretching: we acknowledge that the added value of the security dilemma concept decreases as situations move away from the idealtype of purely security-driven behavior. Nonetheless, we contend that understanding the interaction between predatory and security motives is valuable for designing effective strategies of intervention.

Such mixed situations might be analyzed in a number of ways. It might be that predatory and security motives are both present, and can be treated as equal and independent. If so, policies to mitigate conflict would have to deter the parties' predatory inclinations at the same time as they reassured their security fears. An example is Otto von Bismarck's defensive alliances with Russia and Austria, promising to fight alongside whichever one attacked the other.

However, reassuring fearful predators may be distasteful; it could be seen as rewarding vile behavior and encouraging it in others. Moreover, a settlement that holds predators harmless gives them a chance to start up the game of predation at the next opportunity, as soon as international interveners withdraw. Thus, when both predatory and security motives seem to be present, it may make sense to try to dig deeper to understand the underlying roots of these aims. In some cases, predatory motives may the primary cause, and security fears derivative. Because aggression in an anarchical balance-of-power system provokes resistance and hostility, predators' security fears are likely to be especially acute. 17 Conversely, in some cases, security fears may be primary, and predatory behavior derivative. As we will argue below, the security dilemma often tends to turn even security-driven actors into predators, defined as actors who prefer exploiting others to cooperating with them, even when short-run security threats are small. Thus, the security dilemma gives rise to predators, and predation intensifies the security dilemma. 18

In a very few cases, the situation facing the contending parties in anarchy may resemble Rousseau's Stag Hunt in that each side would prefer mutual cooperation, but is driven to defect solely by the fear that the other side has or will develop a different preference. Even in Stag Hunt, however, the curse of foresight may cause predatory motives to develop. As many realist scholars have noted about international politics, the actor's concern is often not that the other side is currently aggressive or that the current situation is threatening, but that it may become so in the future as others change their capabilities and intentions. Thus, it is incorrect to argue that security concerns and the security dilemma would disappear if the actors could be certain of others' benign intentions. The fact that others or their successors can change, and cannot commit themselves not to, are at the heart of the security dilemma. In these circumstances, unilateral gains from defection may lead to the accumulation of relative advantages which serve as a hedge against future defection by the opponent. 19 Consequently, even if none of the parties starts with predatory motives, the desire to protect one's future position under conditions of the security dilemma can transform a situation from Stag Hunt into a Prisoners' Dilemma, in which exploiting the other side is preferred to mutual cooperation. 20

Even without this dynamic, Prisoners' Dilemma is more common than Stag Hunt, because there is almost always some underlying conflict of interest between the parties. Even if cooperation would make both parties better off, they may have conflicting preferences over the division of those joint gains. Although fear and vulnerability are not the only forces at work, they may still be important in exacerbating conflicts and interacting with other disturbing factors. In many cases, the reaching of an agreement is inhibited both by fears that the other side will cheat and by hopes to gain a better distribution of the values in dispute. On other occasions, the security dilemma may make a situation or settlement particularly fragile while the immediate cause of its destruction can be any number of particular shocks and frictions.

Even the existence of a severe conflict of interest does not mean that the security dilemma is irrelevant. Predatory preferences and behavior at one stage of a conflict, which cannot be immediately explained by the security dilemma, may have been generated by security-related preferences and behavior at an earlier stage. Just as Stag Hunt easily degenerates into Prisoners' Dilemma in the absence of a sovereign authority, so too PD can easily degenerate into the game of Deadlock, in which conflict is preferred to mutual cooperation. Most obviously, this can be the case with spirals of conflict. To understand the dynamics at work, we have to ask why the actors prefer to defect and determine whether at least a part of the reason is the drive for security. In some cases, people may prefer conflict and Deadlock because they think it is the best or the only route to their security: offensive military technology or high subjective requirements for security may place mutual security beyond reach, at least unless there are significant changes in the situation, including outside intervention. In other cases, actors may come to believe that the other is such a menace that they can be secure only if it is crippled, if not destroyed. Deadlock may result, for example, when one party sees a preventive advantage in fighting now, rather than later. Thus the security dilemma, and actions and perceptions that flow from it, can transform a difficult and tense situation, but one with possibilities for cooperation, into one of complete conflict of interest and physical combat.

In a parallel manner, the disintegration of state authority may not only create or awaken security fears, but generate behavior that makes later situations much more intractable. In these circumstances, group identities may become increasingly important, not only intrinsically, but because the security of individuals becomes implicated with the fates of the contending groups. Group identity can then be a consequence of conflict as much as a cause of it and can be fueled by security concerns. Before the start of the fighting in the former Yugoslavia, most Muslims and Serbs did not believe that their security was menaced because of their ethnicity. In fact, in the normal competition in Yugoslav political and social life, these categories were not the most salient. But as the conflict developed, personal security and personal well-being more broadly became defined in terms of the security and well-being of one's ethnic group. While this was partly a matter of psychological identification, everyone else's psychological identification made the identification an objective reality for any given person: individual Serbs, Croats, and Muslims were menaced by people of the other ethnic groups whether or not the person's own sense of self was defined by ethnicity. Others were doing the defining in a way that put it beyond any one person's control. The situation is even sharper in situations when the relevant group identity is readily observable by externals--e.g., skin color.

Thus, the security dilemma can give rise to predatory behavior, zero-sum conflicts of interest, and inimical, mutually exclusive identities. Conversely, predatory behavior intensifies the security dilemma. Hobbes understood this perfectly well: in his view, the reason for generalized fear in the state of nature was not simply the reciprocal fear of surprise attack, but also the knowledge that men had unlimited appetites, which would drive them to aggression out of greed. If men are greedy, or if resource scarcity makes them predatory, this intensifies security concerns, because it increases the expectation that others will defect. Once this expectation becomes entrenched, the security dilemma can take on a life of its own, trapping both predators and prey in rivalries whose costs outweigh any possible gains from exploitation. For example, Germany was not a status quo state in the years before World War I; nonetheless, its aggressive policy in 1914 can be understood only in terms of the security dilemma provoked by its own earlier belligerence, not as the result of a simple calculation that the predatory benefits of conquest would be worth the costs and risks. 21 This is one way to understand the paradox of the predatory "tinpot dictators," described in this volume by David Laitin, who fight harder and harder for the control of ever diminishing state resources.

Several of the contributors to this study argue that predatory strategies can create, exacerbate, or take advantage of the security dilemma in yet another sense. That is, leaders often try to manipulate security concerns in order to solidify their positions and extract additional resources from society. To strengthen their hold on their followers, they may incite conflicts that make the latter more vulnerable and entice or force them to join in the enterprise of collective "self-defense." This is calculated to intensify group solidarity and the need for a continued role for the predatory leaders. Viewed from the standpoint of the leaders' motivations, this may be a purely predatory manipulation, not a security dilemma, but viewed from the standpoint of their followers, who are trapped in the conflict spiral unleashed by the leaders' actions or misrepresentations, the security dilemma may become very real indeed. This is a central theme of the contributions by Susan Woodward on the former Yugoslavia, Bruce Jones on Rwanda, and Rui deFiguerdo and Barry Weingast, who draw on both of those cases. 22 In this sense, a security dilemma may be viewed as an outcome of social processes, not simply a strategic situation.

The security dilemma as a social situation

Even when we conceive of the security dilemma as a situation, it is a social situation with social and perceptual causes, not simply a fact of nature. Thus, the security dilemma is a cause of behavior, yet it is also an outcome to be explained. None of the elements that fuel the security dilemma--neither anarchy, nor offensive advantages, nor expectations that others will defect--can be taken for granted as unproblematic givens, which follow unmediated from technology or geography. If interveners are to undo the security dilemma, they need to understand the social and perceptual factors that shape it.

Arguably, international anarchy might be taken for granted as a starting point for analysis on the grounds that, at least since the Treaty of Westphalia, states have not been subject to any overarching sovereign power. 23 The anarchy that prevails in collapsed states and empires, however, is much more problematic. In those latter cases, the anarchy is not long-standing, but is caused by whatever caused the collapse of the state. Barry Posen claims that the collapse of the Yugoslav state triggered security fears that led to the competitive mobilization of ethnic self-help groups; that is, anarchy caused an ethnic security dilemma. However, if it was ethnic nationalism that tore apart the Yugoslav state and thereby caused the situation of anarchy, then anarchy is an effect, not a cause, of ethnonationalism. Of course, once put in motion by nationalist ideas or politicians, the security dilemma may have unfolded largely by its own logic, just as Posen describes it. As Chaim Kaufmann argues, once inimical identities are hardened as a result of intense ethnic rivalry, it may not matter whether the strategic competition between them was triggered by a "mere" social construction. 24

If the fact of anarchy has often been taken for granted by students of the security dilemma, they have long understood the role of perception and ideology in strategists' calculations of offensive and defensive advantages. In both world wars, indeed throughout much of modern history, soldiers and political leaders misread the implications of prevailing military technologies, in part because of the inherent difficulty of guessing the consequences of untested innovations, but mainly because of organizational biases coloring such assessments. 25

Similarly, estimates of the likelihood that the opponent will defect cannot be inferred simply from the logic of the strategic situation. Few anarchical settings are as uniquely compelling as a burning house, in which everyone behaves the same regardless of the social, cultural, or intellectual understandings that they bring to the situation. 26 Theoreticians of Realpolitik disagree about the relative merits of aggressive versus defensive strategies of self-help, about which circumstances require balancing alliances and which allow passing the buck to other balancers, and about whether the other group is largely driven by fear or whether other is (or has become) so deeply hostile that a just reconciliation is simply impossible--and so do real world leaders. Which path is chosen depends on one's theories of how the balance of power operates, one's guesses about the advantages of attacking, and about the extent to which predatory aims are conjoined with security goals on both sides. 27 Thus, estimating the likelihood that an opponent will defect requires understanding how the opponent subjectively sizes up the situation, as well as assessing the opponents' mix of motivations.

Such estimates are ripe for social and psychological misconstruction. On the cognitive psychological plane, it is well established that people tend to exaggerate the extent to which the actions of others are determined by their innate disposition, while underrating the role of situational constraints, including the observers' own behavior. This tendency operates with special strength when the other's behavior harms the actor. Furthermore, not only are people who expect others to defect in Prisoners' Dilemma driven to defect themselves, but the other's defection will be taken as evidence that cooperation is not possible. In conflictual relationships, this fuels the security dilemma, since both sides are prone to see the other as innately predisposed to defect. 28

On the cultural plane, ethnic groups often hold mythic views of the perfidy of out-group adversaries, which establish a biased baseline for judging the likelihood of defection. As Posen acknowledges, it is not only the anarchical situation, but also elites' purposeful misconstrual of historical conflicts that convinces the ethnic rank-and-file that the opponent's defection is likely. To trigger a security-driven conflict, all that is required is that people believe that such assessments might be true, deFigueredo and Weingast contend. The rest can be accomplished by rational calculations about the high cost of misplaced trust and by belligerent actions by one's own side that elicit responses consistent with worst-case assessments of the foe. Thus if actors have reason to believe that the costs of being conciliatory will be very high if the other side does not reciprocate, they may behave aggressively even though they think that the other probably is benign. Here too, psychology can compound the problem; people are loath to see themselves moving against their neighbors simply because there is only a chance that these people may be a menace to them. When people adopt a hostile stance toward others, they are likely to convince themselves that these neighbors are proven aggressors, not just a hypothetical danger. 29 Furthermore, Posen notes, the actions that trigger a group's fears may be undertaken not only by manipulative leaders, but also by uncontrolled extremist factions or thugs who seek to profit from the turmoil they create. 30

Insofar as the security dilemma is the outcome of a social process and not simply a strategic situation, interventions to manage the security dilemma may require more than simply creating military defensive advantages and monitoring troop deployments. They may also require attention to the content of history textbooks, the introduction of professionalized media, the policing of undisciplined mafia groups, and the creation of alternative incentives for elite groups who might otherwise be tempted to "play the security dilemma card." 31

At the same time, it would be wrong to infer that the security dilemma is just a figment of people's imaginations. People and nations know from hard experience that the real balance of power and real vulnerability to outsiders, not just imaginings about them, are crucial to their fate. Although estimates of the strategic situation may be subjective and erroneous before the fact, once the fighting starts, there is an objective strategic reality that asserts itself. The Schlieffen plan fails, and the trenches are dug; panzers penetrate the impenetrable Ardennes, and Paris falls, no matter what myths were believed beforehand. Social mythmaking is preoccupied with purported strategic facts not because they are all made up anyway, but precisely because everyone realizes how important the assessment of those strategic facts is to everyone's choices and the resulting outcomes. People try to understand their real strategic situation, because they know their fates depend on it, but under uncertainty their ability to analyze these social facts is clouded by cognitive biases and by the manipulations of strategic ideologists, who have their own parochial agendas. From this perspective, the job of interveners in civil conflict is both to create a strategically stable situation and also to a create social setting that is conducive to the accurate assessment of that situation by the contending parties.

Implications for strategies of intervention

Intervention by the international community, if done astutely, can promote the resolution of civil wars by facilitating any of the three mechanisms for solving the security dilemma. Intervention can establish a hegemon by helping one side to win or by imposing direct rule by outsiders. Intervention can also end a conflict by balancing the power of the contending groups, by creating defensive military advantages, or by helping the parties to arrange for a strategically defensible partition of the territory in dispute. Indeed, irrespective of the technologies involved, outsiders can in effect create a great defensive advantage by promising to come to the aid of whichever side is attacked. 32 If both sides know that resuming the fighting, or even behaving provocatively, will call down the intervener's wrath, then each gains a significant measure of security that does not come at the other's expense. Finally, intervention can provide the capability to monitor and enforce new institutional arrangements during a transitional period until locals' expectations converge on the new pattern and vested interests in the new status quo emerge.

However, intervention can also undermine these potential solutions. Inept intervention can prolong wars by preventing one side from prevailing, by transferring offensive weaponry to the parties to the dispute, or by preventing locals from locking in a de facto partition. Interveners can also exacerbate conflict by pursuing internally inconsistent strategies. Just as one cannot back a hegemon and partition power at the same time, so too one cannot delegate power to neutral bodies, yet also provide for powersharing arrangements which give the contending groups the wherewithal to hamstring these new institutions. As Susan Woodward's contribution to this volume shows, the Dayton Accords on Bosnia are riddled with precisely these kinds of internal contradictions, yielding predictably counterproductive results. 33

Regardless of which of these approaches is adopted, a focus on security concerns and the security dilemma can lead to a better understanding of the situation and to fruitful prescriptions for both participants and outsiders. Analysis can go astray if it is insufficiently attentive to these topics. For example, many discussions about the importance of free elections as a way of helping to solve civil conflict neglect the participants' fear that there will be no subsequent free elections after the first one, or that the majority will win and dictate to the minority. As a result, holding elections often intensifies civil and ethnic conflicts. 34 Similarly, much of the stress on the importance of disarming the participants makes the same error as the 1950s literature on international disarmament which failed to realize that a reduction in arms that makes the participant less secure will be part of the problem rather than the solution. 35 Of course the ultimate goal is for the participants to no longer fear each other, but arrangements that lead one side or the other to fear that they are about to be stripped of all protection are not likely to produce even temporary peace.

Using the framework of the security dilemma and the related Prisoners' Dilemma allows analysts, participants, and outsiders to tap into knowledge of how they can increase cooperation in a situation in which both sides prefer to defect while the other side cooperates, yet both sides also prefer mutual cooperation to mutual defection. Parties to the conflict as well as interveners can try to make offense distinguishable from defense, give advantages to the defender, decrease the benefits of exploiting the other on any single play, make the payoff for cooperation much better than bearing the costs of conflict, make the sucker's payoff less than disastrous, and develop contingent strategies, such as tit-for-tat. 36 An agreement that could benefit both sides in the long run can be undermined if it is the case--or if it is believed to be the case--that one side can gain a major advantage from breaking it at a particular point. Outsiders can often function effectively here as guarantors. Even when they cannot or will not impose a settlement, they may be able to provide the forces that can protect each side against the other's defection and symmetrically remove from each the temptation to defect. At least as important, they may be able to provide each with information about what the other is doing that can reduce unfounded fears.

However, adopting a security-dilemma perspective offers no panacea for interveners. One obvious difficulty is that outsiders may lack the will or resources to transform the structure of the game so radically. In addition, interveners who want to defuse the security dilemma face hard choices in deciding how they will try to accomplish this. Some strategies are antithetical to others.

For example, attempts to stabilize an existing security dilemma are likely to work at cross purposes with attempts to "deconstruct" it. Reducing the vulnerabilities of the parties and creating defensive advantages means working with groups as they are, even reifying and further entrenching them. In contrast, trying to reverse the social construction of a security dilemma is likely to mean undercutting the identity myths, institutions, and leadership that hold a group together. Thus, attempts to deconstruct a group increase its vulnerability, threaten its leadership, and thus intensify the security dilemma, at least in the short run. This tradeoff between short run stabilization and long run transformation is a constant theme in our case studies: working with the Somali warring factions versus disarming them, partitioning Bosnia versus integrating it, sharing power with genocidal Rwandan extremists versus excluding them from power. At least in the short run, the interests of peace and justice may be in conflict with one another; to guarantee the peace may be to maintain an unjust settlement. Kaufmann's chapter argues that once ethnic conflicts are intense, the right answer is to stabilize the strategic situation by moving people and partitioning the state. Other authors favor strategies that transform the underlying social relationships that led to the security dilemma in the first place. Woodward argues that, at the very least, interveners should avoid the mistakes of the Dayton Agreement, which pursued both of these antithetical strategies at the same time.

In the long run, then, reducing the security dilemma may lead to the maintenance of a situation that outsiders and many of the participants regard as only a second-best solution, if that. Throughout this paper we have talked about armed "sides"; these should not exist in an undivided country and making arrangements posited on their existence may go a long way to ensuring that they do not disappear. Providing security for identifiable groups is one thing that outside intervention, especially military intervention, can do, however. The knowledge and resources required, although not trivial, are often within reach. But it is far from clear that outsiders either know what to do to encourage a united country or would have the required abilities to do so.

Similar tradeoffs are likely to arise in choosing between deterring predators and reassuring fearful parties. If the international community intervenes with overwhelming force, no tradeoff arises. Predators lack the ability to fight back effectively; they are packed off to jail or to retirement villas in the South of France. If predators cannot be so easily overpowered, however, difficulties arise. Since predators are also fearful, they need to be simultaneously deterred from predation and reassured about their survival. If predators have a secure regional base, like the Khmer Rouge, they can simply be contained there. But often predators cannot survive unless they maintain a monopoly of exploitative state power, in which case offense and defense are indistinguishable. The Tutsi minority in Burundi, for example, counted for its survival on maintaining a dictatorship and a monopoly over military power. To them, democratization and proportional representation in the military ranks and officer corps, measures pressed on them by international donors, were indistinguishable from a death sentence, since they felt that maintaining their control was the only way to guarantee against victimization by the Hutu majority they had brutalized. 37 The conceptual apparatus of the security dilemma may not provide a neat answer to such dilemmas, but it at least highlights the tradeoffs that have so often been ignored in interventions by the international community. Compromises and settlements that appear to outsiders to be feasible and even to contain a modicum of justice may be out of reach because of the participants' beliefs and construals of reality, no matter how unreasonable they may appear to be.

Tradeoffs are also involved in choosing among the three types of solution to the security dilemma: establishing hegemony, creating a defensive stalemate, and institutionalizing cooperation, either through delegation to neutrals or through powersharing. For example, one conceivable solution to Laitin's problem of incessant conflict among hard-pressed "tinpot" Somali warlords is to return to the Cold War practice of elevating one of them through massive military and economic aid to the status of a secure leader. 38 Yet enthroning an asset-stripping faction risks perpetuating the underlying problem.

Likewise, creating stalemates and transferring populations would make policing a cease-fire easier, yet even if it works, this solution reifies the warring groups and rewards those who have played the security dilemma card. Moreover, in many cases, it may be impossible to devise a lasting stalemate dividing a country whose political economy has been deeply integrated. Thus, the problem of obtaining security or of reaching any agreement in a civil war may be greater than it is in wars between well-established sovereign states. Not only are states harder to kill than are most domestic factions, but there can be more rapid change as each group's power rises and falls in the normal ebb and flow of domestic politics and social change. Freezing the status quo--or freezing any settlement--is not likely to be feasible and, indeed, many strains are placed on the settlement as power relations change. Lebanon and Nigeria cannot even conduct a census for this reason. Furthermore, there often are only two main factions, which heightens the relative gains problem.

Consequently, institutionalized cooperation normally seems to be the most attractive solution, though problems and tradeoffs abound here as well. We have already discussed the dangers of institutional delegation to others who have different incentives, as well as the dangers of reifying contending groups through powersharing. In light of these problems, how can institutions succeed in binding parties to a formula for resolving their conflict and making their promises credible?

As deFigueredo and Weingast show in their contribution to the volume, gauging the opponent's character type is often crucial to assessing the likelihood of defection in a security dilemma. Forming an institution that restrains oneself, especially when one has a temporary advantage, can be seen as an index to one's type. Moreover, adhering to an institution can lock in one's type by creating vested interests in perpetuating a profitable arrangement. 39 This is true, for example, of the process of European unification. Not only could observers reason that a Germany that was willing to join in a significant degree of integration would not be an aggressive type, but also the process of integration reinforced German incentives to cooperate and thus locked in the benign German identity and interests. Changes in institutions and changes in behavior can sometimes do even more, changing actors' definitions of their interests, and consequently changing the actor's type, not just revealing it or locking it in. For example, decades of participation in the institutions of NATO and the European Community have contributed to redefining German identity. Germans who once might have wanted to regain the "lost territories" in the East or gain political and economic hegemony over Western and Central Europe now may have different preferences because of a decreased identity with Germany, an increased identity with Europe as a whole, and an associated set of changes in values and means-ends beliefs about what will make Germany as a country and themselves as individuals prosperous. 40 In principle, integrative institutions may be able to engender similar transformations in the wake of civil wars, especially if the strongest local actors use these institutions to reveal their type as restrained cooperators at the outset.

Interveners can provide incentives to undertake such initiatives as well as the resources to make them safe and effective. One drawback of intervention-induced cooperation, however, is that it reveals less about the local actors' true character. Thus local observers might conclude that the most favored tinpot would remain a powersharing democrat only as long as the foreign aid kept flowing. If so, the tinpot could prove himself to be a cooperator only by taking irrevocable steps to lock in the commitment to cooperate, such as the full integration of the dominant faction's army into a professionalized force commanded by neutral officers or ones drawn from all factions. For actors who are unwilling to go this far, then, external intervention may decrease their ability to demonstrate that they will remain committed to peacefully sharing power once they are no longer compelled to do so.

Credibility is also a problem for external interveners, who as Walter shows, will not be able to transform local actors' expectations if their commitment is seen to be limited in time or scope. Although intervening to some extent ties the actor's reputation to finishing the job, the most credible commitments perhaps can only be made by states who have a real interest at stake from the beginning. This was a major problem in the intervention in Rwanda, where the Hutu hard-liners were correct to calculate that no real enforcement would be forthcoming. Indeed, an intervener's lack of will in one case may undercut the credibility of a different intervening power in another case: Bruce Jones' account reveals that the Hutu extremists killed a contingent of Belgian UN peacekeepers at the start of the genocide because they inferred from the American retreat in Somalia that the Belgians would respond to such violence by withdrawing.

Credibility problems are exacerbated, moreover, when the intervener is not a unitary actor. For example, the Nigerian government may have had a strong interest in maintaining order in Liberia (in part to show that it was a true regional power), but it lacked the state strength to see that this policy was implemented by the armed forces on the scene, which instead concentrated on looting. In somewhat the same way, the US military has successfully resisted what little pressure there was from the administration to interpret its mission in Bosnia not as guaranteeing the cease-fire, but as enforcing the Dayton Accords. As significant as these problems are, the basic point remains that external intervention is likely to be cheaper--and hence more credible--when the security dilemma drives the conflict than when substantive disagreements do, because less has to be changed: each side has "merely" to come to believe that the other cannot cheat and threaten it. This may make intervention easier to sustain, because of the perception that it is not only impartial but moral in preventing unnecessary conflict.

In short, outside intervention can help overcome the security dilemma in many ways, from increasing transparency and timely warning by giving each side accurate information about what the other is doing, to providing for the safety of the participants at local "summit meetings," to overseeing local arms control (which could include giving weapons and training, especially of a types that are particularly appropriate for defense), to threatening to cut off aid to any faction that resumed the fighting, to keeping large numbers of troops on the ground to "guarantee" that agreements will be kept (the quotation marks indicating the ambiguity of this term). Yet poorly designed interventions can inadvertently institutionalize the security dilemma, as Woodward argues it has in Bosnia, through a failure to understand tactical tradeoffs. The contributors to this volume do not all agree on what is the right mix of tactics to defuse the security dilemma in conditions of civil war, nor do they agree on its relative weight in comparison with other ways of analyzing the factors that perpetuate conflict. Nonetheless, their sharp analyses and vivid case material go a long way toward laying out those tradeoffs and suggesting an array of strategies that interveners can use to manage them.

Notes:

Note 1: Two valuable efforts along these lines are Michael E. Brown, ed., The International Dimensions of Internal Conflict (Cambridge: MIT Press, 1996), and Ariel Levite, Bruce Jentleson, and Larry Berman, eds., Foreign Military Intervention (New York: Columbia University Press, 1992). Back.

Note 2: Robert Jervis, "Cooperation under the Security Dilemma,"World Politics 30, no. 2 (January 1978), 167-213; Barry Posen, "The Security Dilemma and Ethnic Conflict,"Survival 35, no. 1 (Spring 1993), 27-47; Barbara Walter, "The Critical Barrier to Civil War Settlement,"International Organization 51, no. 3 (Summer 1997). Back.

Note 3: V. P. Gagnon's remarks as a discussant for Susan Woodward's contribution to this volume, conference on International Intervention in Civil Wars, Columbia University, February 14, 1997. Back.

Note 4: In addition to Posen and Walter, see Chaim Kaufmann, "Possible and Impossible Solutions to Ethnic Civil Wars,"International Security 20, no. 4 (Spring 1996), 136-175; James Fearon, "Ethnic War as a Commitment Problem,"unpubl. ms., University of Chicago, 1993; David Lake and Donald Rothchild, "Containing Fear: The Origins and Management of Ethnic Conflict,"International Security 21, no. 2 (Fall 1996), 41-75. Back.

Note 5: Barnett Rubin, "Russian Hegemony and State Breakdown in the Periphery,"in Barnett Rubin and Jack Snyder, eds., Organizing Post-Soviet Political Space (London: Routledge, forthcoming). Back.

Note 6: James Fearon and David Laitin, "Explaining Interethnic Cooperation,"American Political Science Review, 90, no. 4 (December 1996), 715-735. Back.

Note 7: Quoted in Anthony Lewis, "Mandela the Pol," New York Times Magazine, March 23, 1997, p. 43. Back.

Note 8: Freud correspondence. Back.

Note 9: Richard Betts, "The Delusion of Impartial Intervention,"Foreign Affairs 73, no. 6 (November/December 1994), 20-33. Back.

Note 10: See Jervis, Posen, and Kaufmann. Back.

Note 11: An excellent overview of such approaches is Timothy Sisk, Power Sharing and International Mediation in Ethnic Conflicts (Washington, DC: United States Institute of Peace, 1996). Concise statements of these two viewpoints are Donald Horowitz, "Making Moderation Pay: The Comparative Politics of Ethnic Conflict Management,"451-476, and Arend Lijphart, "The Power-Sharing Approach,"491-510, both in Joseph Montville, ed., Conflict and Peacemaking in Multiethnic Societies (New York: Lexington Books, 1991). Back.

Note 12: Avner Greif, "The Institutional Foundations of Genoa's Economic Growth: Self-Enforcing Political Relations, Organizational Innovations, and Economic Growth During the Commercial Revolution,"Stanford University, Department of Economics, manuscript, November 1996. Back.

Note 13: On the dangers of economic interdependence, see Kenneth Waltz, Theory of International Politics (Reading, MA: Addison-Wesley, 1979), chapter 7 . Back.

Note 14: Walter, "Critical Barrier." Back.

Note 15: Charles Glaser, "The Security Dilemma Revisited,"World Politics 50:1 (October 1997), esp. 189-198. Back.

Note 16: Arguing the former is Bruce Jones in this volume; for the latter view, see Jack Snyder and Karen Ballentine, "Nationalism and the Marketplace of Ideas,"International Security 21, no. 2 (Fall 1996), 5-40, citing African Rights, Rwanda: Death, Despair and Defiance (London: Africa Rights, September 1994), by Rakiya Omaar and Alex de Waal, 30-34, 44; Alan J. Kuperman, "The Other Lesson of Rwanda: Mediators Sometimes Do More Damage Than Good,"SAIS Review 16, no. 1 (Winter-Spring 1996), 221-240 Back.

Note 17: Jack Snyder, Myths of Empire (Ithaca: Cornell University Press, 1991), 11-2, 22. Back.

Note 18: On the problem of dealing with predatory "spoilers"of a potential settlement, see Stephen Stedman, "Spoiler Problems in Peace Processes,"International Security 22:2 (Fall 1997), 5-53. Back.

Note 19: Robert Powell, "Absolute and Relative Gains in International Relations Theory,"American Political Science Review 85 (December 1991), 1303-1320. Back.

Note 20: Jervis, "Cooperation Under the Security Dilemma." Back.

Note 21: Marc Trachtenberg's failure to take this distinction into account underlies his dispute with Van Evera over World War I as a conflict of interest as "opposed to"a security dilemma. See Marc Trachtenberg, "The Meaning of Mobilization in 1914,"International Security 15, no. 3 (Winter 1990-91), 120-150. Back.

Note 22: See also V. P. Gagnon, "Ethnic Nationalism and International Conflict: The Case of Serbia,"International Security 19, no. 3 (Winter 1994/95), 130-166 Back.

Note 23: We put aside here the question of the factors that determine how states react to, modify, and even create this condition. Back.

Note 24: Kaufmann, "Possible and Impossible Solutions." Back.

Note 25: On the chronic misreading of the offense-defense balance, see Jervis, "Cooperation under the Security Dilemma,"and Jack Levy, "The Offensive/Defensive Balance of Military Technology,"International Studies Quarterly, June 1984. Back.

Note 26: For the burning house analogy, see Arnold Wolfers, Discord and Collaboration (Baltimore: Johns Hopkins University Press, 1962), 13. Back.

Note 27: Thomas Christensen and Jack Snyder, "Chain Gangs and Passed Bucks,"International Organization 44, no. 2 (Spring 1990), 137-168; Randall Schweller, "Bandwagoning for Profit,"International Security 19, no. 1 (Summer 1994), 72-107. Back.

Note 28: Lee Ross and Richard Nisbett, The Person and the Situation: Perspectives of Social Psychology (New York: McGraw-Hill, 1991). Back.

Note 29: The attribution processes involved are explicated in Harold Kelley and Anthony Stahelski, "Social Interaction Basis of Cooperators' and Competitors' Beliefs about Others,"Journal of Personality and Social Psychology 16 (September 1970), 66-91. Back.

Note 30: Posen, "Security Dilemma,"33. Back.

Note 31: The latter is Gagnon's phrase from his comments on Woodward's paper; on the other points, see Stephen Van Evera, "Hypotheses on Nationalism and War,"International Security 18, no. 4, (Spring 1994), 5-39; and Snyder and Ballentine, "Nationalism." Back.

Note 32: Stephen Van Evera, "The Cult of the Offensive and the Origins of the First World War,"International Security 9, no. 1, (Summer 1984), 58-107, reprinted in Steven Miller, ed., Military Strategy and the Origins of the First World War (Princeton: Princeton University Press, 1985), esp. 96-101. Back.

Note 33: Susan Woodward in this volume Back.

Note 34: Snyder and Ballentine, "Nationalism"; Crawford Young, The Politics of Cultural Pluralism (Madison: University of Wisconsin Press, 1976), 158; Mahmoud Mamdani, Citizen and Subject: Contemporary Africa and the Legacy of Late Colonialism (Princeton: Princeton University Press, 1996), 300. Back.

Note 35: Virginia Gamba's multivolume UNIDIR study pays insufficient attention to this problem. Back.

Note 36: Jervis, "Cooperation under the Security Dilemma"; on the last point, Robert Axelrod, Evolution of Cooperation. Many of the more specific ways of doings this are summarized in Kenneth Oye, ed., Cooperation under Anarchy. Back.

Note 37: Snyder and Ballentine, "Nationalism,"33-34; Réné Lemarchand, Burundi: Ethnocide as Discourse and Practice (Cambridge: Cambridge University Press, 1994), 129, 176, 185, 187. Back.

Note 38: Note that Laitin denies that warlords' short-run security motives were decisive in the Somali conflict, and thus is skeptical that the security dilemma applies to his case. However, in light of our analysis of the interrelationship between predatory and security motives, especially in long-run strategic competitions, it seems to us that the Hirshleifer "war of attrition"model employed by Laitin overlaps heavily with the security dilemma. In any case, the point argued here applies equally to both. Back.

Note 39: Anna Eliasson, dissertation-in-progress, Columbia University. Back.

Note 40: Daniela Engelmann, Hans-Joachim Knopf, Klaus Roscher, and Thomas Risse-Kappen, "Identity Politics in the European Union: The Case of Economic and Monetary Union (EMU),"prepared for Politics of Economic and Monetary Union (Helsinki: Finnish Institute for International Affairs, 1997). Back.

 

CIAO home page