APPENDIX

Early Theoretical Models That Pertain to Conflict De-escalation

In this appendix, we first focus briefly on listing a number of factors that have been mentioned as important to conflict de-escalation processes. Afterward, we quickly move on to the consideration of fourteen models that are relevant to rivalry de-escalation and termination. These models include Rosati 1994 and Welch 2005 on foreign policy restructuring; Checkel 1998 and Legro 2005 on ideational shifts in foreign policy; Diehl and Goertz 2000, Maoz and Mor 2002, and Cox 2010 on rivalry onset and termination; Rock 1989, T. Armstrong 1993, and Kupchan 2010 on rapprochement; Lebow 1995, Pervin 1997, and Orme 2004 on adversarial de-escalation and accommodation; and Womack 2006 on asymmetrical politics.¹ Since these models overlap considerably with our expectancy revision model, we wish to highlight their efforts in order to make the case for a partial synthesis.² We find that all of these models appear to be looking at similar explanations, but they privilege different combinations of factors more than others. At the same time, these models stress similar sets of variables that produce a surprising amount of overlap in the existing literature.

Students of conflict resolution maintain that the de-escalation of protracted conflicts ultimately depends on conducive background conditions that encourage adversaries to believe in settlement. Shifts in these background conditions create moments of opportunities or “ripeness” when peacemaking efforts are likely to have favorable outcomes. Table A.1 provides a list of some of the many factors that scholars have linked to the de-escalation process. It shows that de-escalation may arise from domestic conditions, international contexts, and the relations between the adversaries (Kriesberg 1991).

Past research also indicates, however, that there is no simple direct relationship between any of these factors and the termination of protracted conflict across many domains (Licklider 1993: 303). All of these variables have different sorts of impacts under different circumstances. Hence, we have no single pattern or theory of conflict resolution. In addition, Licklider (1993: 303–304) argues that conflict resolution more likely reflects a set of processes in which certain critical choice points emerge. Decisions at these points lead to alternative strategies of conflict termination. But, which situations (factors or combination of factors) present critical choice points and why is left unclear.

Table A.1. Conditions That Bring about De-escalation in Protracted Conflicts

Processes		References
Hurting stalemates		Zartman and Aurik 1991; Zartman 1993, 2001; Grieg and Diehl 2005, 2006; Weible and Sabatier 2006
Promise of reciprocity		Ward 1982; Touval and Zartman 1985; Goldstein and Freeman 1990; Zartman 1993; Lebow 1995; Rajmaira 1997; Goldstein et al. 1998, 2001; Bogumil 2002; Gintis et al. 2005
Extending peace initiatives		Kriesberg 1992; Hutchison 2001; Zartman 2001; Garb and Allen Nan 2006
Leadership changes and shifts in the internal distribution of power		Evangelista 1991; Licklider 1993; Bennett 1997; Mendelson 1998; Diehl and Goertz 2000; Colaresi 2004
Leadership autonomy		Stein 1993a; Coleman 1997
Third-party mediation		Touval 1975; Pruitt 1981; Becovitch, 1995; Hartzell 1999; Regan and Stam 2000; Schrodt and Gerner 2004
Mutual trust and credible commitments		Kelman 1995; Walter 1999; Bar-Tal 2000; Kydd 2005
Organizational failure and learning		Hermann 1990; Levy 1994; Stein 1995
Shifts in the external distribution of power (e.g., military defeats, loss of patronage)		Bueno de Mesquita, Siverson, and Woller 1992; Mitchell 1995; Huth 1996; Kriesberg 1998; Chiozza and Goemans 2004
Internal economic crises		Stein 1993b; Waterman 1993; Bennett and Nordstrom 2000
New perceived threats to security		Rock 1989; Lebow 1995; Thompson 2001b; Kupchan 2010

In an effort to pare down every argument that is potentially relevant to our understanding of rivalry termination, we focus on fourteen previous models of conflict de-escalation that overlap with features in our expectancy revision model. We begin with Rosati's (1994) more general model for foreign policy restructuring.

Rosati and Foreign Policy Restructuring

Rosati assumes that societies at any given point in time contain competing domestic forces that work toward and against reinforcing the status quo (Boulding 1970). Which groups gain the upper hand and in what circumstances accounts for any changes that may occur. Rosati's model concentrates on the interactions among state institutions, society, and their environment, which have an impact on the development of strategies and policies. State bureaucracies are viewed as inert institutions that move only when pushed hard to do so (Allison 1971; Halperin 1974). Meanwhile, decision makers, who preside over state policy formulation, are disinclined to alter the belief systems that underpin their choice of strategies (Holsti 1967; Jervis 1976). Mass publics also tend to be conservative because they are more comfortable with the status quo than with the uncertainties of change. Thus, vested interests and tradition pose obstacles to potential changes and leaders who advocate for such change. Concomitantly, the international environment also presents a number of built-in constraints on domestic policy shifts, including alliances, commercial relationships, legal commitments, and norms. Continuity rather than discontinuity more often characterizes the development of foreign policies and strategies (Rosenau 1981; Holsti 1982; Goldmann 1988; Hermann 1990; Volgy and Schwarz 1994).

Nonetheless, societal and environmental alterations occur and the question then becomes whether states and leaders modify their foreign policies to adapt to these new circumstances. If they choose to avoid altering existing policies, then leaders and their governments risk the increasing probability that their policies will fail, which in turn will expand the criticism from groups that advocate for foreign policy change. Domestic or international crises also open up the opportunity for such change as defenders of the status quo are overwhelmed by pressures both at home and abroad for a reorientation. In sum, foreign policy failures expose the liabilities and inappropriateness of existing strategies, shake conventional belief systems, upend government personnel supporting the status quo, and bring forth new personnel committed to different foreign policy approaches.

Rosati quotes Nisbet (1969: 282) to summarize the nature of this crisis-induced dynamic: “A given way of behaving tends to persist as long as circumstances permit. Then…the way of behaving ceases to be possible as the result of some intrusion, some difficulty which is the consequence of event or impact, and a period of crisis ensues.”

Hence, crises create the necessary space for policy innovators and their innovations. However, there is no guarantee that such crises will lead to change. If catalysts produce positive outcomes, then leaders and governments are likely to reinforce policy continuities. If, however, crises are perceived to have negative consequences, then they are more likely to generate the quest for new policies—and the more severe the impact of negative events, the greater is the potential for foreign policy reorientation.

Based on this perspective, foreign policy restructuring tends to be fairly abrupt and episodic (Krasner 1984). Without crises, incremental change is the likely consequence. Only when the intersection of ongoing societal and environmental shifts and traditional policies combine with critical events that lead to the defeat of status quo defenders do we expect to observe a significant redirection of foreign policy. If, however, status quo defenders maintain their political control, then they are likely to support only minor or moderate alterations in foreign policy.

In sum, Rosati's model reemphasizes that true policy changes represent punctuations in an otherwise steady and stabilized course.³ And, crises are the catalysts that overcome the normal resistance of policymakers to undertake new policy directions.

Welch and Painful Choices

While Rosati stresses the role of crises, Welch (2006) emphasizes the pain associated with the high costs of making foreign policy shifts. Welch starts with the presumption that foreign policy situations are structured to avoid major changes. Inertia is more probable than innovation. The main reasons for this bias are rooted in uncertainty, interpretation problems, and the costs of change. Welch claims that foreign policy analysts are not very good at anticipating change and that when change does emerge they are unlikely to recognize it. In addition, foreign policy organizations are slow to adapt to new environments since they are reluctant to abandon well-established routines.

Stressing the role of cognitive processes, Welch argues that policymakers are reluctant to blaze new trails. They are more likely to distort incoming stimuli than to alter their fundamental beliefs about external adversaries and threats. Leaders are also unlikely to perceive when opportunities emerge for change. And, in those cases where some policymakers observe the payoffs in trying something novel, they will assess the potential costs of pursuing them as too high. Policymakers realize that admitting their earlier policies were flawed while pursuing new policies is very risky politically. They calculate that they will be blamed personally if their new policies are unsuccessful.

Consequently, decision makers are more likely to pass up possible gains associated with potentially chancy new policies in order to avoid losses even if continuing old strategies is seen to be equally or more risky. Only when decision makers perceive that the losses associated with new policies exceed the risks of doing something different are they likely to support change. Thus, there are individual and organizational biases toward “staying the course,” which ensure that foreign policy shifts will be rare and unlikely unless a disaster is looming. In this respect, leaders must be forced to change. In other words, “Foreign policy change is most likely when decision-makers perceive that their current policies are incurring painful costs; that a failure to change policy is virtually certain to result in further painful costs; and that at least one option open to them holds forth the possibility of an acceptable outcome, even if that acceptable outcome is not highly likely” (Welch 2006: 46).

Welch's argument about the inertial tendencies of foreign policy actors is compelling. Welch cites a variety of factors that work against foreign policy change: the cognitive propensities of individuals to preserve their beliefs in spite of contrary evidence; the tendencies of individuals to prefer avoiding losses over achieving potential gains; and the interests of bureaucracies to avoid innovation. Hence, the impediments to change are high and only when decision makers have their backs against a wall are they willing to pursue a new strategic tack.

Despite its appeal, Welch's argument raises questions. When do decision makers perceive they have little choice but to experiment? Certainly, there have been leaders who have preferred to go down with their ships of state—the extreme example of Adolf Hitler in a bunker in Berlin in 1945 comes to mind—than to consider more realistic (and perhaps less painful) policy alternatives. Just what does it take to compel decision makers to contemplate altering their tactics or strategies? In fact, how often do decision makers stay the course before they are replaced by different decision makers? Welch's approach seems to be tied to the likelihood that incumbents will change their minds. But, what about the possibility that turnovers of the elite are more likely to be associated with foreign policy change? How frequently do incumbents actually execute foreign policy shifts as opposed to new policymakers who are less tied to older strategies? Ideally, we would like a theory that takes into account the broader circumstances that bring about the probability of foreign policy change—as opposed to Welch's more restrictive approach on how policymakers change their own courses of action.

In a rivalry context, what prevents decision makers from escalating hostility as opposed to de-escalating it? Either course of action could constitute a significant foreign policy shift. To what extent can we generalize about whether foreign policy shifts will be more cooperative or more belligerent? Can our theoretical understanding of foreign policy change be oriented to explaining more than just a generic shift?

Checkel and New Ideas in Foreign Policy

Crises and events that can either reinforce the status quo or lead to major deviations are useful clues. But we need greater specificity in our search for ways to account for rivalry de-escalation and termination. One model closely approximates our own approach, albeit with differences. Checkel (1997) models the behavioral changes in Soviet foreign policy at the end of the Cold War on the following assertions:

1. The greater the external uncertainty or foreign policy crisis, the greater are decision makers' searches for new ideas. Thus, external uncertainty or a foreign policy crisis creates policy windows that encourage advocates of new ideas to promote them.
2. Policy entrepreneurs are critical to the empowerment of new ideas, especially in more centralized states that lack alternatives for affecting change.
3. Because centralized states lack alternatives for affecting change in comparison to decentralized states, new ideas are less likely to reach elites in centralized states than they are in decentralized states. But once decision makers in centralized states embrace new ideas, they are more likely to put them into practice and change state behavior than is the case in decentralized states.

Hence, external change and uncertainty create space for policy entrepreneurs to promote new ideas. Those states that are more authoritarian are more resistant to these cues because their systems are more closed, but they are also more readily manipulated once the new ideas are adopted. In contrast, states that are less authoritarian are thought to be more open to new ideas but less likely to implement them. Since Checkel's study focuses exclusively on an authoritarian setting, he has yet to test his ideas across centralized and decentralized states. This issue distinguishes Checkel's approach from our own.

Although we find the institutional argument to be interesting, we are unwilling to assert that institutional differences play a significant role in cases of rivalry de-escalation. That is not to say that all governments operate similarly in holding on to rivalry perceptions or that states, for instance, with electoral contests may be less susceptible to rivalry outbidding as leaders pursue votes. Rather, we prefer to remain agnostic about the role of institutions on the origins and implementation of foreign policy change at this time.

Another difference between our approaches is Checkel's focus on generic ideational change. For instance, our expectancy revision model emphasizes the shift in ideas about strategic enemies. Checkel, in contrast, is interested in ideational shifts more broadly. For instance, ideas associated with strategies of economic development may fit Checkel's emphasis on regime differences more readily than rivalry strategies. So, while our approaches appear to overlap considerably, Checkel's is couched at a more general level.

Checkel's attention to external uncertainty is a third difference. Our expectancy revision model stresses specific types of internal and external shocks while Checkel's model applies more generically to any foreign policy crisis that generates uncertainty for decision makers. We are unwilling to say that all crises are likely to have the same impact. Nor do we believe that general foreign policy crises will be the most likely settings for rivalry de-escalations, especially if the crises involve the two rivals as antagonists. In these cases, the level of intense threat is apt to discourage rather than encourage the reassessment of a rival's intentions.

Finally, we agree with Checkel's argument that changes in external environments can produce opportunities for policy entrepreneurs to lobby successfully for different approaches to solving problems. However, we maintain that the process by which policy entrepreneurs are able to gain access to or control of their governmental bureaucracies that are wedded to old strategic ideas takes time and usually occurs when foreign or domestic policy problems have become acute. Policy entrepreneurs are frequently bystanders waiting for openings to press their ideas more aggressively. Alterations in the external environment can bring about new challenges or problems for decision makers. If these alterations (either in the external or internal setting) discredit incumbents sufficiently, then policy entrepreneurs may be positioned to replace them and in that event they might be able to pursue strategic innovations.

Legro and Changing Orthodoxies in Foreign Policy

Along lines similar to Checkel's approach, Legro's approach (2005) focuses on changes in dominant foreign policy orthodoxies.⁴ Although actors have interests and preferences, Legro is interested mostly in how collective beliefs or ideas guide the selection of strategies that deliver the behavior that actors want to play in world politics: integrationist, separatist, or revisionist roles.

At the heart of Legro's model, which is also critical to our own perspective on rivalry termination, is his view on external shocks. Similar to other scholars' work on foreign policy shifts, Legro maintains that shocks create windows of opportunity for change in both institutions and dominant orthodoxies. However, Legro quite rightly notes that shocks can be indeterminate. He argues that we do not know which shocks are most important in bringing about change and whether these shocks will have the same impacts across different societies. Moreover, scholars tend to look at only those cases where shocks have produced change and ignore cases where shocks have failed to produce change. Without this comparison, Legro maintains that it is difficult to discern how or whether shocks really mattered. Thus, the standard approach of “old thinking > shock > new thinking” is unsatisfactory.

Legro modifies the standard model with a two-stage argument. An old dominant orthodoxy must first collapse. In most situations, the collapse of collective beliefs is unlikely to occur. That is, dominant orthodoxies are likely to remain dominant orthodoxies for some time. How then are dominant orthodoxies sustained or swept away? Legro advances an auxiliary theory of ideational collapse that is based on four possible outcomes. If actors choose a strategy that corresponds to the prevailing foreign policy ideational framework and the results of the strategy are evaluated positively, then dominant ideas will remain. In contrast, if actors choose to pursue strategies that fail to correspond to ideational guidelines, then the results of such strategies will matter little. For instance, if the results are perceived to be negative, then the outcome will reinforce traditional ideas; and if the results are conceived to be positive, then actors will still have few incentives to challenge dominant orthodoxy. In short, “happy” outcomes are unlikely to lead to a revision in traditional ideas. However, the one situation that is most likely to create turmoil involves cases where strategies correspond to orthodox views and yet produce failed outcomes. In this context, the greater the failure, the more likely it is that actors will overthrow traditional orthodoxies.

Collective ideas do not exist without controversy since they are fought over by partisans of different perspectives. Major failures bring critics out in the open as well as shake the convictions of those who once subscribed to the dominant orthodoxy. At this point, the potential for moving to a new set of collective beliefs is greatest. But this potential must be realized by a shock that overwhelms the defenses built around entrenched ideas. The presence of a shock is not enough; the severity and the circumstances in which they occur are important as well.

In the second stage of Legro's model, new ideas must be consolidated before they become the dominant way of thinking. Consolidation, in turn, depends on primarily two factors. One is the number of alternative interpretations that exist at the time of collapse. Too many alternatives increase the likelihood that factional fighting will prevent the adoption of a new orthodoxy. In the absence of any obvious alternatives, actors will fall back on their old ways of thinking. Thus, the most favorable setting for consolidation is a situation where only one strong alternative approach emerges. The absolute failure of the old orthodoxy thereby opens up the possibility for actors to switch to an opposing viewpoint.

Consolidation also depends on how well the new way of thinking seems to be working. If new ideas appear to be failing as badly as the old ones, then they are less likely to remain dominant for long. If new ideas appear to be related to more successful outcomes, then they are likely to be retained.

Thus, there are several components in the Legro model. Shocks are prominent, but not just any shocks will do. The most crucial ones are linked to major policy failures that were expected to succeed. Then, too, these shocks only “open the door” to the possibility of a shift in prevailing orthodoxies. Partisans of alternative ideas must struggle and compete to have their perspectives adopted. New orthodoxies that emerge must be reinforced by success before they become consolidated or institutionalized in foreign policy bureaucracies. Legro's model is very similar to the one that we developed. Yet, we maintain that shocks do not always have to be linked to major failures before strategic revisions must occur. Rivals, as we saw in the chapter on India and Pakistan, are quite capable of resisting the implications of major shocks.

Rivalry Approaches to Onset and Termination

Diehl and Goertz on the Necessity of Shocks

Whereas shocks are prominent in the Legro model, they are the primary focus in the framework of Diehl and Goertz (2000). Their basic argument is that environments and interactions tend to move toward stability. In order for rivalries to begin or end, something must disrupt the stability. Hence, internal or external shocks introduce this necessary but insufficient condition. External shocks create new opportunities and expectations for rivals to initiate or terminate their conflicts, especially if the power distribution between them has been affected in a substantial way. Meanwhile, internal shocks change leaderships, regimes, and the amount of resources available for engaging in foreign policy activities.

Examples of suitable external shocks include world wars, quick changes in relative capability standings, and changes in territorial control. Primary types of internal shocks are independence, civil wars, and regime changes, especially shifts to democracy. Empirically, Diehl and Goertz find that over 90 percent of rivalry births and deaths have been associated with some kind of shock. However, their evidence also indicates that shocks are more critical for the commencement of rivalries and less so for the termination of rivalries. Hence, they conclude that shocks are a weak necessary condition for ending rivalries.

The strong emphasis on shocks is understandable. They appear to be quite common and readily associated with foreign policy changes. However, they do require careful interpretation. Should independence be viewed as an equivalent shock in kind to world wars or democratization? Often, independence does not come as a shock but is fully anticipated. It may or may not alter expectations all that much in the sense that pre-independence feuds may be continued in the post-independence era. Still, independence can significantly alter the environment in which states operate. One can no longer rely on the colonial power to police older feuds or disputes over boundaries that had been relatively suppressed in the colonial era. Similarly, democratization may reflect gradual or abrupt processes of change. It can alter fundamentally the decision-making personnel and foreign policy goals at play. So, if shocks play critical roles in rivalry behavior, then we need to ensure that their impacts on intervening events are also abrupt and unexpected and tied closely to rivalry termination processes.

Maoz and Mor

Another approach with considerable utility is advanced by Maoz and Mor (2002), who believe that they can explain rivalry onset, escalation, and termination with a relatively straightforward theoretical premise.⁵ Peaceful dyads encompass mutually satisfied states, or, if there is dissatisfaction in the dyad, the dissatisfied state is too weak to do anything about it. Conflictual dyads, therefore, involve states that are not satisfied with the dyadic status quo and believe that they can do something about it.

In other words, rivals perceive that they share fundamentally incompatible goals—a perception that may be predicated on relatively accurate or inaccurate premises. Consequently, rivals are likely to engage in cycles of repeated conflict unless something changes along the lines already identified. Eventually, rivalry termination occurs when dissatisfied states become satisfied via a perceptual shift about the level of incompatibility with a rival or through the recognition that their initial perceptions were wrong. Alternatively, dissatisfied states can also transform into satisfied states when they become convinced that they lack the capability to continue contesting the status quo.

Of course, these types of circumstances may be only temporary. Maoz and Mor draw particular attention to warring rivals in which each war leaves one side triumphant and the other side in defeat. The outcome can lead to some stabilization only as long as domestic changes do not bring into power a decision-making group that is either more dissatisfied or more impressed by its capabilities than the group that lost the last war. When one of these conditions reemerges, another round of overt hostilities can be anticipated.

We also note that the outcomes of these warring rivalries since 1945 have not been all that decisive in altering the perceptions of the participants.⁶ That is, winners in war are less likely to eliminate their international adversaries. One side is often defeated but its territory is neither occupied nor absorbed into the winner's expanded boundaries. Recurring conflicts between rivals may have eliminated one of the parties in earlier historical periods, but this result may be less likely in a contemporary period in which there are international restraints on the winners of war. Moreover, during the Cold War, rivals found their military capabilities rebuilt fairly quickly as a result of superpower assistance.

Although the argument of Maoz and Mor is both parsimonious and theoretically compelling, we have some problems with their approach. First, Maoz and Mor equate rivalry with conflict density. In short, rivals are those states that clash frequently and after these clashes cease, they are no longer considered to be rivals. We contend that this line of thinking indicates that Moaz and Mor are really explaining outbreaks of repeated conflict but not necessarily rivalry. Our argument is that while rivals are much more likely to become involved in repeated, overt hostilities than nonrivals, rivalries themselves can persist in the absence of explicit changes in conflict. Once we assume that rivalry can exist without overt conflict, Maoz and Mor's emphasis on satisfaction and capabilities becomes less compelling as an explanation of rivalry termination. While satisfaction and capabilities continue to be important for understanding the onset and escalation of rivalry, they are less applicable if termination is not associated with the end of overt hostilities as we contend. Hence, we believe that other variables are required to understand rivalry terminations.

Another problem with the approach of Maoz and Mor is that it is very dyadic-centric.⁷ Whether rivals are satisfied or dissatisfied is an issue that is determined by the status quo that exists between members of the dyad. This may be the case for some rivalries, because dyadic members can be so totally focused on each other that little else seems to matter. However, this may not always be the case. Rivalries can be influenced by other processes that exist outside of the dyadic context. For instance, other external rivalries can influence the internal behavior of rival dyads. We maintain that an overly dyadic focus is likely to overlook the role that regional and global influences play in rivalry termination as our later case studies show.

Yet even within the scope of Maoz and Mor's theory, it is quite conceivable that satisfaction and capabilities may be contingent on events and processes occurring outside of a dyadic rivalry. For instance, the emergence of a new, mutual threat from a third party may lead dyadic rivals to reevaluate their traditional perceptions of threat. Another possibility is a shift in exogenous processes that diminish perceived threats—as in the case of the Chinese, who reassessed their threat perceptions about the United States as U.S. involvement in Vietnam declined in the late 1960s. Capabilities, too, can be directly affected by the loss of external patronage as illustrated by the Vietnamese-Chinese rivalry. Although Vietnamese decision makers did not have to deescalate their conflict with China in the aftermath of the Soviet Union's disintegration, they did have to reassess their relative capability to continue the rivalry at a fairly intense level.

Although Maoz and Mor (2002: 32–42) recognize that perceptual changes between rivals can be brought about by exogenous events, they tend to consider such factors outside the scope of their theory. Hence, a crucial difference between their approach and ours is that we believe factors often exogenous to a rivalry dyad are important to termination processes. Therefore, if external shocks are critical as other theorists argue, we want them inside a model of termination as opposed to remaining outside of it.

Cox's Rivalry Termination Model

Cox's (2010) model, unlike many other rivalry approaches, is specifically designed to account for rivalry terminations. Rivalries are perpetuated by three foreign policy orientations, which are present on both sides of the relationship. Hawks want a rivalry to continue at all costs; doves prefer to end the rivalry; and in between are the moderates who maintain the rivalry until they are no longer convinced that it is feasible to do so. Thus, the trick to bringing about rivalry termination is to reduce the influence of the hawks, increase the influence of doves, and convince moderates that the rivalry is more costly than beneficial.

These three “objectives” are more likely to be accomplished in the context of policy failures. If failures are acute and occur in both the domestic and foreign environments, then they will increase the probability of a major change in the ruling coalition. This is especially so if domestic policy failures are associated with economic setbacks. Otherwise, a leadership change is less likely if the failures are limited to foreign policy alone. Democratic institutions help to facilitate the turnover of leadership in these contexts, but they are not required. Nor is it guaranteed that hawks will fall out of favor while doves replace them; but nevertheless, policy failures create the opportunity for this to occur. In that instance, the opening for rivalry termination also emerges.

At this point, the focus shifts to the other member of the rivalry. If this adversary has also experienced domestic and foreign policy failures, then the opportunities for a change in leadership also occur. When changes in leadership occur on both sides of the rivalry more or less at the same time, the prospects of rivalry termination improve considerably. However, in the event that this does not happen, the outcome for rivalry termination will then hinge on whether the side that is attempting to change its approach offers something that the adversary is seeking. If not, then rivalry termination is unlikely. Another obstacle emerges when an adversary delays in responding to the other side's initiatives.

Cox's model is certainly testable. It is also attractive because it specifies a temporal sequence of events when bargaining could lead to rivalry termination. For instance, leadership change occurs first and is followed by a policy change; depending on the nature of these changes, they could lead to the end of a rivalry. However, the hawk-dove dichotomy in the model is awkward since decision makers may not always be easily identified as fitting into one of these categories. Presumably, the most important condition is whether decision makers who are committed to an ongoing set of policies are removed in favor of decision makers who are committed, or open, to new alternative policies. So, whether one group is more or less ideologically conservative than another is beside the point. There is also the problem of identifying policy failures, which ultimately depends on whether actors perceive them as such. One way to avoid this problem is to focus on how shocks create openings for new leadership without the requirement that they must lead to the replacement of one set of leaders for another.

Rapprochement

Rock on Peace Breaking Out

Perhaps, the earliest explicit argument for rivalry termination is Rock's (1989) interpretation of “why peace breaks out” between hostile great powers. In the first stage of his model, Rock maintains that decision makers decide that they need to abandon the war option before resolution can be achieved. Decision makers are likely to be motivated to do so when they face three conditions: 1) the emergence of a new threat; 2) the necessity of preserving economic connections; and 3) and the imperative of avoiding civil war. Once decision makers on both sides decide to reconcile, they form suitable policies in the second stage of the model. Rock argues that reconciliation is facilitated in this stage by three other factors. First, the adversaries have minimal overlap in, or commitment to, geopolitical objectives. Second, the adversaries have little economic rivalry that stems from producing and/or exporting similar commodities or lending capital in the same areas. Third, they have similar sociopolitical attributes that are likely to reduce their ideological incompatibilities.

In addition to these motivating and facilitating factors, Rock argues that a catalyst of some sort is usually needed to overwhelm the inertia associated with one side's negative perceptions of the other. Borrowing from Lebow (1981), he views acute crises as the most probable catalyst. If one side is defeated in the crisis, then that defeat may encourage decision makers to think anew about their rival. Another scenario is that defeat may weaken the influence of hard-liners or it may resolve the disputes at stake between the rivals. In short, crises and/or their outcomes encourage one or both sides to look for alternatives, and settlements that occur in the context of these crises can even generate some trust and good will on both sides of the bargaining table.

Rock downplays the role of human agents while putting greater stress on the presence or absence of key external factors that facilitate reconciliation. If these facilitating factors are present, then rapprochement is likely to occur regardless of the competence or incompetence of diplomats.⁸ Thus, Rock's model has three critical components: decision-maker motivations, catalysts, and facilitating factors. Reconciliation begins when decision makers' motivations and catalysts break through the policy inertia; it is achieved mostly likely if the obstacles (or facilitating factors) line up favorably.

We agree that policy inertia must be overcome by either catalysts or shocks. An interstate crisis can serve this role, but we are not sure that defeats in crises are guaranteed to produce a reevaluation of prevailing strategies. Such defeats may lead to greater future hostility and even war if decision makers resolve to pursue a more hard-line stance in the next confrontation.⁹ We believe that other types of shocks and motivations need to be considered for rivalries to end. As for Rock's facilitating factors, we are not sure whether they should be viewed as sources of rivalry or obstacles to rivalry termination. For instance, ideological incompatibilities, geopolitical conflicts, and economic rivalry are reasons for increased perceptions of threat and hostility between states. If these factors are absent initially, then the emergence of rivalries is likely to be low. If they are present, then there is no reason to presume that rivalry termination is likely to appear as an option that decision makers would consider.

We believe that it is not the presence or absence of these factors that are crucial, but whether and how the nature of these factors change over time. We argue against viewing these factors as static obstacles that must be removed for rivalries to end. Moreover, we do not agree with the notion that rivalries will persist as long as the reasons for rivalry persist. The annals of rivalry termination suggest that ideological incompatibilities, economic rivalry, and geopolitical contestation can be overcome given the appropriate circumstances—as illustrated by one of Rock's motivating factors (the emergence of a new and more serious threat) for instance.¹⁰ Consequently, we are not convinced that “facilitating” factors necessarily trump “motivational” or catalytic factors. We also maintain that reciprocity and reinforcement are critical facilitating factors in the process of rivalry transformation. Despite the shortcomings in Rock's model, we draw on a number of its arguments to inform our own expectancy revision model.

Armstrong

Tony Armstrong (1993) presents another set of pertinent clues for rivalry termination.¹¹ Although he (1993:134) concentrates on negotiation tactics that two rivals employ once they explore the prospects of rapprochement, he offers two generalizations about the “international circumstances” that will encourage both sides to come to the table in the first place. He maintains that improved relations are more likely to succeed when the following conditions are present for both governments: 1) the general strategic and/or diplomatic situation threatens to deteriorate and continued hostility obstructs the governments' ability to address the problem; and 2) decision-maker expectations about the nonviability of their adversary's regime are replaced by expectations of the viability of the regime.

The first condition relates to environmental deterioration that implies an implicit invitation to contemplate a change in strategies. Armstrong focuses on perceptions of impending deterioration as opposed to Rosati's realized crises. Yet both of Armstrong's conditions introduce the idea of expectations—albeit not very explicitly. His first condition speaks to expectations of strategic deterioration. His second addresses expectations about an adversary's viability in a context of abandoning earlier hopes that the other side might solve one side's problems by conveniently collapsing. While this factor does appear in some cases, it is not clear that it shows up in enough cases to make it a compelling explanatory element. Still, one could argue that Armstrong's conditions are anchored clearly in decision-maker expectations that encompass perceptions concerning more generally the state of the environment in which they operate and more specifically the status of the enemy rival.

Kupchan's Stable Peace Model

Kupchan (2010) embeds rivalry termination within a more elaborate stage theory designed to explain stable peace outcomes. Since stable peace goes far beyond rivalry termination, our current interest is focused primarily on the initial stages. The first phase is centered on one side's unilateral accommodation. When a state is beset with multiple threats, it is logical for that state to attempt to reduce the number of threats with which it is coping by selectively signaling its benign intentions. If concessions can work to eliminate one adversary in a crowded field, then the accommodating side is likely to have improved its general security position, at least if it is successful.

The first phase dampens the selected rivalry. A second phase furthers it to the extent that both sides become more accommodative. Gradually, the rivalry disappears and opens up the possibilities of the third and fourth phases, which allow for increased interaction and the formation of communal identities between the two states. Whether all four phases will come about depends in turn on institutional restraints and compatibilities in terms of social orders and culture. Domestic restraints on the exercise of power facilitate the communication and imagery associated with relatively benign intentions. More essential, however, are compatible social structures in which the distributions of power among various types of domestic groups are not threatened by more peaceful interactions with the former adversary. To generate new narratives of common identity, it is also critical to be able to draw on a common cultural heritage.

Thus, Kupchan begins with limited rapprochement (phase 1), which can broaden into something like a security community if the appropriate ingredients are in place. For our purposes, however, this model resembles most closely the British-U.S. transition from rivalry to a special relationship in the twentieth century. In that respect, it narrows its broader rivalry applicability by specifying a single path—unilateral threat reduction—as the termination trigger. The desire to reduce the number of existing threats does arise from time to time, but it is not sufficiently widespread to explain a large class of rivalry terminations. Something more general is needed to encompass multiple triggers for rivalry termination.

Adversarial De-escalation and Accommodation

Lebow on Adversarial Accommodation

A different approach, although one that certainly agrees with Orme's emphasis on leadership skills and choices (discussed below), is taken by Lebow (1995:181), who hypothesizes that conciliatory accommodation between two adversarial states is most likely when leaders 1) are committed to internal reforms that require or are expected to benefit from improved relations with an adversary; 2) recognize that confrontation has been counterproductive in the past; and 3 expect their adversaries to respond positively to conciliatory overtures.

Lebow refers to these conditions as propositions about alternative pathways to accommodation. He believes (1995: 182) that other pathways also exist and suggests that mutual fear of a third party and economic incentives may also play some role. In a subsequent article (Lebow 1997: 165) he adds a new condition relating to the domestic feasibility of mobilizing support for conciliatory foreign policy. Moreover, implicit to both of his treatments is a very strong focus on the leaders involved. The ending of the Cold War was very much a Gorbachev story. The de-escalation of Egyptian-Israeli tensions was very much a Sadat story, and so on. Yet in both analyses, Lebow (1997: 175) also makes sure to stress that his hypotheses “represent a tentative step toward a more comprehensive theoretical explanation…and need to be incorporated into a broader theory that more fully specifies the…other conditions.”

The expectancy theory that we have put forward is intended to approximate that broader theory to which Lebow refers, even though it may not be exactly what he had in mind. A variant on his first condition (commitment to domestic reforms) is incorporated as one of several sources of shock (Rosati's crises)—as are his suggestions of third-party threat and economic incentives. His second condition does not appear in the theory in any explicit way (essentially, it is taken for granted that innovative policy entrepreneurs are dissatisfied with the old ways of doing things), but his third condition (expected reciprocity) is one of the main components. The idiosyncratic element of leadership is translated into a strong version of policy entrepreneurship. Finally, his fourth condition (domestic support) has a role to play as one source of continued reinforcement for what is seen as a gradual process of de-escalation once such a course is set in motion.

Pervin on the Transformation of Antagonism

David Pervin's (1997) perspective on the transformations of antagonisms in the Middle East presents a different type of challenge. He begins with a set of assumptions that merge into something that approximates a compact and widely applicable theory. Decision makers are rational actors who pursue the primary goal of staying in power. But to achieve this goal, decision makers must have access to the resources that make success possible. Whatever affects access to the necessary resources will force decision makers to reconsider their strategies. Whether new strategies are likely to be forthcoming, in turn, depends on a variety of other factors. Policy alterations require that inertia be overcome, and the greater the previous investment in a conflict situation, the greater may be the sunk costs that must be overcome or ignored. Whether and how policies can be altered also depends on how difficult it is to obtain resources from domestic and international environments. The greater the ease of access to resources, other things being equal, the less likely is policy alteration.

Another factor is the symbolic value of the conflict. Minor conflicts are more easy to overcome than are conflicts that can have a large number of significant repercussions. For instance, there may be an explicit dispute over who is allowed to control adjacent territory. Territorial issues can be sufficiently intractable in their own right, but they become even more awkward when surrender might have implications for a state's regional leadership position. Implications such as these compound the costs of concessions. Then, too, concessions can have variable impacts on domestic actors. Internal political and economic changes may be deemed too expensive to salient constituencies. Finally, still another consideration concerns how long prevailing circumstances are likely to persist and what might follow in their wake. If the current international environment is thought to be relatively long-lived, or if the future alternatives are even less desirable, then decision makers may be more likely to adjust their strategies.

What one should look for, then, are actors adapting their strategies to changing environments, subject to a number of caveats. It is not just any change that matters but, rather, ones pertaining to the survival of decision makers and the resources essential to that goal. Pervin illustrates his perspective by a brief examination of foreign policy changes in the relations between Jordan, Syria, the PLO, and Israel in the mid-1990s. The environmental changes were multiple and sequential. The Cold War between the United States and the Soviet Union had ended, leaving the United States the primary surviving superpower in the Middle East. Soviet allies could no longer count on being rearmed after losing a war with Israel. Arab states could no longer play the two superpowers against each other. The general effect was to make interstate war less likely and Arab actors less independent of external patrons. Another change was the second intifada, which made the perceived costs of Israel holding on to territory in the West Bank indefinitely more costly.

A third major change was the Iraqi invasion of Kuwait and the eventual defeat of Iraq by a large coalition led by the United States. The salience of the United States was reinforced while Jordan and the PLO found themselves more or less on the losing side of the Persian Gulf War. Palestinians were expelled from the Gulf area, putting pressure on Jordan to absorb them despite major domestic economic problems while the PLO's funding was cut severely. In addition, the PLO was facing an increasingly serious challenge from Hamas over the leadership of Palestinian resistance to Israel. Jordan was being pressed by the United States to improve its relationship with Israel in return for debt relief and arms.

Syria, in contrast, had lost its Soviet patronage but gained from being part of the winning coalition in the Persian Gulf War. Its domestic economic problems were much less severe than Jordan's and, unlike the PLO, the Syrian government had no domestic challenger. Its main territorial issue with Israel was difficult to approach incrementally. Either Israel would return the Golan Heights or it would not. To concede on this issue would not only be extremely unpopular at home, it would also reverberate throughout the region, negatively affecting Syria's claims to Arab leadership.

Which of the three Arab actors was most likely to reduce its conflictual relationships with Israel? The theory predicts that Jordan and the PLO, as the most vulnerable actors, would be more likely to increase cooperative interactions with Israel than would Syria, the least vulnerable of the three. This is precisely the outcome that developed in the mid-1990s. Israel and the PLO negotiated the Oslo Accords. Jordan and Israel negotiated an end to their interstate rivalry. Little change took place in the Israeli-Syrian relationship. Of course, Israel need not have been receptive to cooperative overtures, but the overthrow of Likud's rule by Labor made receptiveness far more likely. Israel's economic future was also seen as benefiting from a substantial decrease in conflict. Opportunities for investment and technological production could be expected to expand in a less risky setting.

Pervin greatly (over)simplifies the increased cooperation in the Middle East by contrasting three preference schedules.¹² One “Arab” preference schedule is compared to two Israeli (Likud and Labor) sets of preferred futures.¹³ Pervin's basic point, however, is that the Arab and Likud preferences had been vitiated by environmental changes. Israel was too strong militarily and the Arab actors too weak militarily to destroy Israel. Without the fallback support of Soviet patronage, Arab-initiated warfare was both unlikely and too dangerous. The intifada had made holding on to the occupied territories gained in 1967 less attractive than had been the case earlier. As a consequence, the Arabs were reduced to their third and fourth preferences (peacefully regaining or conceding the territories). The defeat of Likud by Labor meant a preference shift within Israel, brought on by the Labor schedule appearing more attractive than the views held by the incumbent regime. The ascendancy of Labor meant that Israel was prepared to negotiate peacefully over the status of the occupied territories at the same time that at least some of its Arab opponents were also prepared to be more cooperative than in the past.

Pervin goes on to note that the longer these environmental conditions persisted, the more likely it would be that Middle Eastern relationships would be transformed. The key is for the international system to create incentives for cooperation. The longer these incentives remain in effect, the more solid the ground for gains through cooperation—especially if important domestic actors view their interests as being best served by the maintenance of pacific behaviors and environments.

We have no problem with arguing that environments, prevailing preferences, and inter-actor relationships evolve or change over time. When environments change, actors are likely to need to reconsider whether their leading preferences and strategies need to change as well. The end of the Cold War, the intifada, and the Persian Gulf War certainly constitute or reflect environmental changes in the Middle East. We emphasize the same types of factors but call them “shocks” instead of changes for they are not simply gradual changes but rather abrupt ones that often catch the actors involved by surprise.¹⁴ As shocks, they represent opportunities for reconsidering strategies toward adversaries, among others. To the extent that the responses to the shocks lead to behavior that is more in tune with the altered international environment, actors can be said to have adjusted or adapted their behavior. In this respect, we are not really in disagreement with Pervin's approach. Rather, we are emphasizing different aspects of the same environmental changes.

We do not feel the need to assume that decision makers are rational actors engaged in careful cost-benefit calculations while also primarily seeking to remain in power. Real world examples of behavior that conform to these assumptions are not difficult to find, but they are not critical to the building of our theory. We do accept the significance of changes in decision makers' access to resources encouraging revised calculations. Shocks are frequently shocking because they impede or cut off access to resources that decision makers had come to rely on in earlier times.

We certainly have no reservations about accepting the utility of preference schedules. Political actors rarely fixate on one possible eventuality as the sole acceptable outcome. Instead, they tend to identify more and less favored outcomes. They may work harder to attain the more favored outcomes, but sometimes they must accept the less favored ones. Alternatively, actors who are fixated on prevailing preferences that are no longer seen as viable may find themselves ultimately removed from office.¹⁵

One of the more interesting aspects of Pervin's argument is the implication that images of adversaries never change. Preferences may have to give way to altered realities from time to time, but the idea that decision makers' perceptions of their adversaries and adversarial intentions might change and then generate new slates of preferences and strategies is entirely absent. We think this assumption is too strong, but it does highlight an important dimension of analyzing rivalry terminations. If one emphasizes the perceptual/psychological foundations of rivalry, then it is not difficult to construct a model of rivalry termination that would be predicated on decision makers realizing that their opponents were not quite as malign as they had previously thought and, therefore, more suitable for cooperative exchanges. Pervin, no doubt, would reject such a model as naïve and implausible. In contrast, we do not rule out such a possibility, but we do not find it all that probable. We think policy about rivals is based on expectations about what strategies are most likely to work, given the adversary's capabilities and intentions and one's own capabilities and preferences. A more benign view of an adversary could lead to altered expectations. But so too could perceptions of an adversary's weakness or strength, an adversary's apparently altered strategies, and one's own relative weaknesses or strengths that would also encompass changes in opportunities and threat perceptions.

Accordingly, decision makers do not have to learn to love their adversaries in order to ratchet down interstate tensions and rivalry relationships. Threat perceptions, for instance, are malleable. They can be exaggerated for conscious effect and they can be downplayed when convenient. One illustration is the U.S. reaction to the importation of Soviet missiles in Cuba in 1962 and its much more low-keyed reaction to information on the development of Soviet submarine bases in Cuba less than a decade later. Neither of the reactions necessarily implied that U.S. decision makers had drastically altered their image of Soviet decision makers and their intentions. Similarly, the negotiations involving Nixon, Kissinger, Mao, and Zhou in the early 1970s did not hinge on abrupt U.S. or Chinese reappraisals of U.S. or Chinese benignness or intentions. The beginning of U.S.-Chinese rivalry termination was instead predicated primarily on calculations of the utility of reduced U.S.-Chinese hostility vis-à-vis U.S.-Soviet and Sino-Soviet relations.

We see such changes in interstate relations as something more than merely moving down one's collective preference schedule. But we also think that agents are often necessary to the process of changing preferences. That is, preference adjustments do not happen automatically. They require policy entrepreneurs who may or may not be able to manipulate their bureaucratic resources toward new ways of doing business with adversaries.¹⁶ Thus, Hubert Humphrey, if he had defeated Nixon, might not have been able to, or even likely to, seize the opportunity that Nixon and Kissinger exploited in 1972. That does not mean Nixon was necessary to the U.S.-Chinese rivalry termination process, but he was certainly a part of it. Much the same statement could be applied to Rabin in the Israeli-Palestinian negotiations of the mid-1990s. His ascendancy to the Israeli prime ministership and his subsequent removal from that office were not marginal to the success or failure of the Oslo Accords. Policy entrepreneurs can be an important variable in altering preference schedules.

Finally, we agree that time is critical to rivalry transformation. Decisions to de-escalate antagonisms are not like light switches that can be turned on and off. Rivalry de-escalation processes are begun at some point and then take time to gel into more concrete relationships. But more than time is required. The initial decisions to alter an adversarial relationship need considerable reinforcement. Both sides need to feel that they have not made a fundamental error in judgment. The identity of the policymakers involved may change, but their behavior toward the former adversary must remain roughly consistent with expectations of reduced threat. Otherwise, it is only too easy to slip back into the conventional groove of hostility between rivals. This reinforcement process is also a critical part of the rivalry termination process.

Orme on Rivalry De-escalation

Orme's (2004) interpretation of rivalry de-escalation is similar to Armstrong's. Orme believes that analysts give too little credit to statesmen. It is their task to monitor an adversary's intentions and capabilities and to respond as best they can within the constraints imposed by levels of domestic support for various foreign policies. If the issues that separate rivals are intractable, then there may be little recourse but to fight for one side's interests.

Where issues are tractable, however, some possibility of negotiating a noncoercive outcome is possible. A variety of factors may encourage or discourage successfully negotiated cooperation. Leaders may feel, for instance, that their political survival hinges on the continuation of external conflict. If so, then genuine cooperation with external adversaries would seem to be unlikely. Proliferating external threats and economic stagnation, however, can create an atmosphere of pessimism that may facilitate negotiation.

It is pessimism that Orme anoints as the key to peace between rivals. As long as rivals believe that they are either stronger than their adversaries or that there is some good chance that their own capabilities will improve in the near future, negotiated accommodation is improbable. Intransigence, resolve, and/or outwaiting the enemy seem most advisable. If one rival thinks its own capabilities are weaker or unlikely to improve, then some type of accommodation may become attractive—both as a leadership goal and a strategy that can be sold to domestic audiences. In this respect, Orme believes that pessimism is more important than distrust between rivals. Given sufficient pessimism and effective leadership, the distrust can be overcome. In the absence of substantial pessimism, few incentives exist to make any concessions.

Orme may be right about pessimism. It is hard to imagine optimists making substantial compromises. One or both sides need to feel a need to try something different and break out of an international relationship that appears to be thwarting the attainment of other goals. It seems quite possible that deteriorating military and economic capabilities, as Orme suggests, can encourage some decision makers to reassess their foreign policies toward traditional foes. But, as in the case of Maoz and Mor's (2002) approach, Orme's focus is primarily on the psychological mood within one or more rivals. Moods can be rather tricky to operationalize. Whose moods count? How ephemeral are decision makers' moods? Are they up one week and down the next? One way to look at what we are doing is to say that our question involves what triggers more pessimistic outlooks and, once triggered, what else is necessary to bring about rivalry de-escalation.

Asymmetrical Politics - Womack

The final model to consider is Womack's (2006) asymmetrical politics model. Womack begins with the assumption that asymmetrical relationships between states work differently from symmetrical relationships. Symmetry allows both sides to contemplate equivalent moves. Asymmetry means that some significant moves are foreclosed to the weaker side. For instance, the stronger side might be able to invade and conquer a weaker opponent but it is most unlikely that the weaker opponent could do the same. As a consequence, asymmetrical relationships will tend to be susceptible to relative inattentiveness by the stronger side and overattentiveness by the weaker side. The stronger side will tend to focus on relationships with states that are more competitive while the weaker side may become overly sensitive to interpreting the stronger side's actions and intentions.

A second theoretical characteristic of asymmetrical relationships is differential expectations. The stronger side expects the weaker side to show deference to the stronger side's greater capabilities. The weaker side expects the stronger side to respect its own autonomy. Misperceptions can arise on one or both sides, but the chances for conflict are especially magnified if the weaker side ceases to be deferential after perceiving the stronger side to be threatening its autonomy. It is also hardly unknown for the weaker side to exaggerate its own capability to withstand attack from stronger parties. What may thus be viewed as a policy nuisance initially by the stronger side can be seen as a matter of national survival by the weaker side.

Should the stronger side attempt to bring the weaker party to heel coercively, it is unlikely to mobilize its full capabilities. The degree of explicit threat simply is not there for an all-out attack. The stronger side probably also has stronger enemies to worry about. Should the weaker side muster sufficient resistance to thwart the stronger side's punitive attempt, a stalemate may ensue in which neither side is likely to win. The weaker side has too much to lose to surrender and can utilize asymmetric warfare techniques to its own advantage. The stronger side can keep the stalemate going indefinitely but risks losing prestige, resources, and domestic support for the effort if the stalemate is allowed to persist indefinitely.

When both sides conclude that the stalemate is unlikely to work in their favor and that a less hostile relationship is preferable to the status quo, negotiations over normalization are conceivable. Then the question becomes whether some new framework of deference for the stronger side and autonomy for the weaker side can be constructed as negotiators consider the terms of resolving outstanding policy differences. The movement toward a “normal” relationship is not likely to be immediate because suspicions on both sides must be first overcome. But if the normalization process appears to be working, decision makers on both sides may become convinced that the new relationship can work. Until that point is reached, though, some probability of intermittent threats and crisis lingers on. Should the status of normalcy be attained, however, the new relationship is apt to be more stable than was the case prior to the development of active hostility because both sides are more likely to avoid some of the misunderstandings that led to the earlier escalation of conflict.

In Womack's worst-case scenario for asymmetrical relationships, something leads to conflict escalation and an attempt to resolve the issue by the stronger side through the use of military force. If the application of military force fails to achieve the desired outcome and the weaker party is able to resist successfully the stronger side's efforts, then a state of stalemate sets in. Of course, nothing guarantees that the stalemate will not persist indefinitely. At least one side, and preferably both, has to realize that the stalemate is less desirable than yielding on policy differences. At this point, some potential for negotiation and normalization of the relationship is possible. Normalcy is attained when both sides reach the point that they can assume that policy differences are likely to be negotiable.

One of the definite pluses of Womack's model is that it was developed to conceptualize processes in what we view as an asymmetrical rivalry between China and Vietnam. It offers, therefore, a rivalry model with some attractiveness even though Womack does not use the term rivalry. The twin theoretical features of over/under attentiveness and defense/autonomy deserve further application to other asymmetrical conflicts, such as that between Russia and Finland or the United States and Cuba. Womack's perspective also emphasizes the expectations of the antagonists—a perspective that we showed some appreciation for in Chapter 2. The stress on the slowness of normalization processes is also highly appreciated. Yet the basic nature of a model explicitly developed to deal with asymmetrical situations limits its applicability for rivalry purposes. Asymmetrical rivalries have existed and some have continued much longer than one would expect. Still, they constitute a definite minority in the rivalry population.

The question remains whether we need to focus solely on models for specific types of situations or whether it is possible to develop more general conceptualizations that can be applied to both symmetrical and asymmetrical rivalries. At the same time, we are not suggesting that researchers must choose one or the other approach. Limited models can have great utility for the populations that they address, especially if they are fairly consonant with more general models. We think the model developed in Chapter 3 is quite compatible with Womack's useful interpretation.

Summary Observations on the Other Approaches

Thus, there are a few points that can be made based on this very selective sampling of the literature on, or pertinent to, rivalry de-escalation. One is that earlier efforts have been content, for the most part but certainly not entirely, to advance isolated hypotheses (see, for instance, Table A.1) or very general models. The need for a broader and more comprehensive treatment has not gone unrecognized. Another is that the substantive arguments relied upon in earlier efforts need not be abandoned. No effort need be made to incorporate every possible proposition, but it is quite likely that a more comprehensive yet explicit theory will overlap substantially with earlier hypotheses and models. The difference is that the theory needs to tie the components together in some coherent fashion. It also needs to generalize as much as possible the conditions that are thought to influence efforts at de-escalation. That may mean using different language to convey (and to cover) types of behavior similar to those discussed in earlier propositions. Finally, the sampled literature by Rosati, Checkel, Legro, Maoz and Mor, Armstrong, Orme, Welch, Lebow, and Womack highlights, without exception, the standard mode of analysis on rivalry de-escalation. Things change in decision makers' environments that lead them to revise either their view of the adversary or the wisdom of their existing strategy for coping with the adversary. This theme constitutes the core of the expectancy theory developed in this book. Changing decision makers' minds is critical and necessary to creating situations in which de-escalatory negotiations between rivals can begin—a point made by Armstrong, Orme, and Womack. But, we have argued, these changes are not sufficient to bring about rivalry de-escalation.

We have proposed that an expectancy framework is best positioned to frame the types of processes upon which we are focusing. It offers a viable means of weaving together a wide range of seemingly disconnected variables into a coherent explanation of de-escalation. We maintain that protracted conflicts are processes that are subject to change in the context of environmental challenges. In other words, protracted conflicts are sustained by the strategies and policies of adversaries, their expectations, and their institutions. Over time, these strategies are reproduced in a routine fashion that produces inertia.

Protracted conflicts can be seen as partial equilibrium systems. They exhibit stability over time, but during certain periods the possibilities emerge for rapid, dramatic, and nonincremental change in the level and intensity of the conflict. A key property of protracted conflicts, therefore, is their inertia—the point at which the adversaries' expectations and behavior are said to be in equilibrium. In other words, the continuity (or stability) of protracted conflicts reflects the entrenched expectations that each side has of the other in terms of their tactics, strategies, and past actions.

Something has to give for fundamental change to emerge in the relations between two adversaries. We think protracted conflicts de-escalate when adversaries adopt new interpretations, understandings, and expectations of their opponents (see Adler 1991). When is such innovation most likely to occur? We think that crises or shocks are the critical factor in bringing about the realization of alternative expectations. Periods of stasis or equilibrium are eventually interrupted by shocks or incidents that tip the level and intensity of the protracted conflict into a new direction that can lead either to escalation or de-escalation (Young 1998: 12-15).

Shocks consist of an event or a short-term chain of events that are potentially major turning points in political behavior. They can be either exogenous or endogenous, although neither is assumed to have a more important effect than the other. Exogenous shocks emerge from the environment that is external to the protracted conflict (as in the cases of changes in international distributions of power or war). Endogenous shocks occur within the domestic contexts of the adversaries (as in economic crises).¹⁷

Shocks have both subjective and objective dimensions (Knight 1998). The objective dimensions are the event or series of events that occur abruptly and have the capacity to bring about a change in behavior. The subjective dimensions deal with the orientations, attitudes, or perceptions of the adversaries involved. If this interpretation is correct, then adversaries are unlikely to change their policies unless shocks induce the realization that existing strategies (based on outmoded expectations) are infeasible or potentially disastrous if they continue. Only then does cooperation and compromise become possible.

Political shocks, as a consequence, are necessary conditions for the initiation and consolidation of de-escalation. They are transforming events that cause adversaries—or, more likely, some elites within the antagonistic rivals—to adopt new ways of thinking. When adversaries realize that the strategy they had been pursuing cannot triumph or they cannot gain more by continuing it, a more accommodative strategy may offer a better alternative (Kriesberg 1998: 217). Therefore, shocks can change the expectations and strategies of at least some of the critical participants who then attempt to undertake new policies that move intractable conflicts into seemingly less intractable ones.

Yet shocks are at best necessary but not sufficient triggers for de-escalation. They are not sufficient because shocks do not always result in initial accommodations, especially if they fail to elicit new ways of thinking in one or both of the adversaries. Moreover, shocks may have long-term impacts as result of their intermediary effects on other internal or external processes that more directly bring about changed expectations. What these shock-induced changes are likely to be are hard to predict given the variety of events and circumstances surrounding protracted conflicts. Hence, it is impossible to guarantee when and where shocks will lead to political breakthroughs.

What events qualify as shocks with the potential to de-escalate hostilities? Unfortunately, the list of possible candidates is long and diverse. However, recent empirical research by political scientists and sociologists on social movements, revolutions, regime transitions, and elite settlements point to the role of certain types of recurring events: defeat in warfare, economic crises, withdrawal of foreign support, and the loss of societal support for existing policies.

While military defeats and/or the withdrawal of foreign support have typically been associated with revolutionary situations, state breakdowns, and crises within the elite (Tilly 1978; Skocpol 1979; Leng 1983; Reiter 1994; Goertz and Diehl 1995, 1997; Goldstone 1991; Oberschall 1996; Dogan and Higley 1998), they also play a critical role in de-escalating protracted conflicts. These events affect the capacity of adversaries to mobilize the necessary resources to sustain protracted conflict. Diminished resources (in terms of economic and/or political support) may force parties to reevaluate the conditions under which they are willing to seek an initial agreement (Lichbach 1995: 167–209).

A sharp deterioration in a country's aggregate economic performance (including the perception that economic conditions will continue to worsen) can threaten the political survival of ruling elites unless they can reverse the situation. Ruling elites who face challenges from their opposition groups and mobilized publics are likely to reverse policies that are construed as obstacles to economic recovery (Haggard and Kaufman 1995; Dogan and Higley 1998). Consequently, leaders will reevaluate their willingness to prolong conflicts that drain important societal resources or block needed reforms for long-term economic growth (Bunce 1981; Evangelista 1991; Licklider 1993; Lebow 1995; Bennett 1997). Stein's (1993b) analysis of the bargaining environment regarding the Camp David Accords of 1979 provides a good example. She argues that acute political and economic problems in both Israel and Egypt (although more intense in Egypt) changed the bargaining strategies of Sadat and Rabin, both of whom sought a political agreement that would reduce not just their security threats abroad but would pave the way for more foreign aid and economic liberalization at home.

One reason why protracted conflicts last as long as they do is because they are sustained by what Baumgartner and Jones (1993) call “policy monopolies” within one or both of the adversaries. Policy monopolies reflect dominant political understandings about a policy issue—akin to Legro's dominant orthodoxies—in this case the conditions surrounding interactions and negotiation with an adversary. They are reinforced by existing institutional structures (which limit access to the policy process) and powerful supporting ideologies that reside within these institutions. These buttressing ideologies are usually connected to core political values, which are communicated directly through images and rhetoric to the public. Policy monopolies last until they are supplanted by new political and/or social movements that are predicated on new ideas that become popular and diffuse throughout the population (Baumgartner and Jones 1993: 6-16)—at least in more democratic settings. In less democratic settings, it may suffice to supplant policy monopolies exclusively within elite circles.

Consequently, internal uprisings, civil strife, or protest cycles (Burton and Higley 1998; McAdam, Tarrow, and Tilly 1997) may reverse policy monopolies by either replacing old institutions and/or elites with new ones or encouraging old elites to reevaluate their modes of thinking. Defeat in war and economic crises may also lead to these sorts of changes as well. In any case, new expectations can emerge, which may lead to policy changes vis-à-vis external adversaries.

To the extent that shocks or crises hasten the re-evaluation process, some actors (certainly not necessarily all) move from one set of collective understandings to another. If the direction of the shift produces negotiations, then these collective understandings will be communicated through diplomacy. That is, leaders will express their new understandings to their opponents, which should facilitate the process of negotiation and conciliation (Adler 1991). Cooperation and de-escalation, therefore, are rooted in the strategies of political survival. How, why, and which environmental crises are likely to bring about escalation or de-escalation was discussed in more detail in Chapter 2.

Shocks can induce indirect changes as well. Disasters and crises reinforce preexisting perceptions of a problem among elites and the public alike. If people perceive that existing policies are not meeting certain goals, that the costs of these policies are escalating, or that they are generating unanticipated negative consequences, then public support for ruling elites and their policy monopolies or orthodoxies may erode. Shifts in public opinion, for instance, can lead to administrative or legislative turnovers from which policy entrepreneurs are likely to emerge. By their very definition, policy entrepreneurs—actors seeking political gain (and improved policy) by innovating new ways of treating public problems—bring about the destruction of policy monopolies by advocating new solutions to old problems and selling new proposals to elites and political and/or social movements. Most of all, they exploit political events and opportunities to differentiate their positions on pressing policy problems (Kingdon 1995: 182). They are political pioneers, willing to venture into new terrain by pursuing new solutions to old problems.

Yet, policy entrepreneurs are hardly passive actors waiting in the wings for the appropriate constellation of factors to appear. They are apt to be proactive. They seize opportunities to carve out more political space that makes policy change possible. Shocks can help generate these opportunities, especially if they have produced major structural breakdown, fragmentation, and dissension. Entrepreneurs are then in a position to restructure their political environments in ways that provide them with enough political power to overcome internal commitments to older, more entrenched policy strategies. In short, policy entrepreneurs have to have more than new ideas. They must have sufficient political resources to support and sustain new strategic overtures toward cooperating with their adversaries.

Finally, they need reciprocal gestures from their adversaries lest their efforts to promote greater cooperation resemble the act of one hand clapping. Decision makers may also need pressure and encouragement from third parties. Third parties can also sometimes smooth the way toward greater cooperation. Moreover, the cooperative sequences, once instituted, must continue to be reinforced by the actions of both sides (Burton and Higley 1998). If not, then there is always some likelihood of one or both parties reverting to their default position of protracted conflict. Thus, post-settlement shocks may either derail the process or reinforce the commitment by elites to abide by the initial accommodation. Conciliation and cooperation, therefore, can be as protracted as the initial conflict seemed.

We contend, in sum, that shocks, expectation revisions, policy entrepreneurs, reciprocity, third-party pressure, and reinforcement can be critical to the de-escalation and termination of rivalry relationships. We do not expect each and every factor to be equally important to changing strategies. The questions are whether the theory is helpful and whether the various factors specified by the theory are of equal importance.