Psychological Explanations of International Conflict

Janice Gross Stein. Handbook of International Relations. Editor: Walter Carlsnaes, Thomas Risse, Beth A Simmons. Sage Publication. 2002.

International conflict can be conceived as a set of interactive and interdependent decision problems that arise within the context of world politics. The context shapes, but does not determine, the choices leaders make and, indeed, through their choices, leaders change the context and structures of world politics. Individual and collective theories of choice, the focus of systematic research in cognitive and social psychology, become relevant to international politics insofar as we accept the proposition that leaders of collectivities–states or groups–acting alone or as part of a group, make decisions that have consequences not only for themselves but for others. The choices of one leader can have a powerful effect on the structure of the decision-problem itself and on the choices available to others.

Psychological approaches are useful in establishing boundary conditions for on-going arguments within structural explanations of systemic patterns. They help as well to specify theories built on identity formation and change and on system-wide norm creation and observance. Psychological theories are also useful in explaining the bounded choices political leaders make to escalate–and to de-escalate–conflict, and in the analysis of patterns of interaction that are the result of linked series of choices.

In international conflict, important decision-problems are typically ill-structured. The definition of the problem is often contested among the parties and there is little agreement on the nature of the stakes or their value. Even when the representation of the problem is not contested, the environment often does not provide timely and accurate feedback nor do leaders have the opportunity to engage in repeated trials over time to generate robust probability distributions (Newell and Simon, 1972; Simon, 1973;

Voss and Post, 1988; Voss et al., 1983; Voss et al., 1991). Leaders are consequently often uncertain about the problem, stakes, the values and intentions of others, and the constraints that define the problem. Abstract values must be constructed, interpreted and reformulated as precise objectives in the light of specific contingent circumstances. Often, the options themselves are not known and leaders must identify the options as well as their consequences. At times, the structure of a problem emerges only through the process–individual or collective–of constructing the representation of the problem. Leaders’ representations of decision problems have a significant impact on their construction of preferences and identification of options, as well as on their choices.

Psychological theories are especially useful in the explanation of international conflict when representations of problems are contested, when these problems are not routine, the stakes are high to the choosers, and when the environment offers sufficient degrees of freedom to permit a range of choice. Under these conditions, institutional routines are often not considered adequate and the role of leaders, acting alone or collectively, is critical. In the first part of the chapter, I begin by reviewing the contribution of cognitive psychology to the explanation of processes of problem representation in ill-structured problems typical of non-routine conflict environments in international politics. I examine next the contribution of prospect theory, the leading alternative to theories that model choice as subjectively expected maximizing decisions. I then move from the individual level to explore the contribution of social psychology to the explanation of small group decisions. Many foreign policy decisions are made in the context of a small group, functioning either as a collective chooser, or as advisers to a leader with final executive responsibility. In the second part of the chapter, I move beyond choice as the focus of explanation to examine the contribution of social psychology to the analysis of mass ethnic conflict, where identity and group dynamics are important drivers of conflict in the international system.

Psychological Explanations of Choice

Cognitive Approaches to Individual Choice

Cognitive psychology has identified a series of systematic deviations from the norms of rational choice. Deviations from rational-actor assumptions about judgement, estimation and choice are explained by the need for simple rules that permit timely responses to the complex, often ill-structured problems and the uncertain environments of the kind political leaders typically face in international relations. These responses are adaptive in routine, well-structured environments but can produce significant distortions in complex poorly structured environments. There is no single cognitive theory of choice, and cognitive psychologists have identified no dominant decision rule (Mintz and Geva, 1995). Instead, they have specified and categorized the filters through which people process information, and the simplifying mechanisms they employ to help them make sense of the world. Political psychologists, drawing on research done in cognitive psychology, see leaders of collectivities facing choices about conflict as cognitive managers struggling to manage inherent uncertainties and complexities through typical cognitive ‘short-cuts.’ Although cognitive psychology provides no unified theory of choice, it explains why people deviate from ideally rational choice and alerts the analyst of international politics both to the importance of identifying leaders’ representation of their problems, and to a menu of systematic strategies of simplification leaders are likely to use.

Cognitive Stability: Beliefs, Schema, Heuristics, Biases and Information Processing

Cognitive psychology has demonstrated that people’s prior beliefs strongly affect information processing (Grayson and Schwartz, 1999; Larson, 1997; Sanbonmatsu et al., 1997; Wegener and Petty, 1998). Theories of cognitive consistency postulate that individuals seek to maintain the consistency of their ‘belief systems’ against discrepant information in ways that lead them to depart from norms of rational inference and choice. Indeed, exposure to contradictory information frequently results in the strengthening of beliefs. The strengthening of beliefs after exposure to contradictory information results from the processes of reasoning people use to explain the apparent inconsistency. The discount rate of information that is inconsistent with organizing beliefs is systematically higher than rational norms would dictate and people tend to choose options whose anticipated outcomes are consistent with established beliefs.

In international politics, leaders can be expected to discount systematically new information and resist change in fundamental beliefs (Little and Smith, 1988). President George Bush, for example, required a consistent stream of evidence over a protracted period of time before he changed his belief about Mikhail Gorbachev. Indeed, even a consistent stream of evidence was insufficient; it took the destruction of the Berlin Wall to overcome his resistance. Discounting has also been used to explain the success of strategies of deception and the consequent strategic surprise experienced by intelligence officials. The failure by American intelligence to detect Japanese intentions before the attack on Pearl Harbor, the failure of Israel’s intelligence to predict the Egyptian and Syrian attack in 1973, and the failure to predict Iraq’s invasion of Kuwait in 1990 have all been explained not by the absence of good evidence but rather by the tendency of officials to discount systematically evidence that was inconsistent with prevailing beliefs. Analyses of a wide range of political leaders, working in divergent political systems, suggest very similar processes of discounting information is incompatible with belief systems are at work.

Attribution theorists have emphasized the importance of ‘schema,’ or individuals’ concepts and their defining attributes. A schema is a working hypothesis about some aspect of the environment, and may be a concept of the self (self-schema), other individuals (person schema), groups (role-schema), or sequences of events in the environment (scripts) (Fiske, 1986; Fiske and Taylor, 1984: 140; Lau and Sears, 1986; Walker, 1988). People use schema to organize their interpretation of their environment and develop scripts–a working hypothesis about the environment–to prepare for action. Unlike theories of cognitive consistency, attribution theories do not assume that an individual’s collection of schema form a coherent system. But like theories of cognitive consistency, they too assume that schema, once formed, are resistant to change.

The well-established tendency to discount information that is discrepant with existing schema contributes significantly to cognitive stability. The postulate that schema are resistant to change can be interpreted as consistent with statistical logic if people assign a low variance estimate to their expectations. Psychological research contradicts this interpretation through repeated observations that exposure to discrepant information strengthens rather than undermines existing schema.

Common heuristics and biases can impair processes of rational revision and judgement as well (Kahneman et al., 1982; Nisbett and Ross, 1980; Suedfeld and Tetlock, 1997; Von Winterfeldt and Edwards, 1986). Heuristics are rules people use to test the propositions embedded in their schema, and may be thought of as convenient short-cuts or rules-of-thumb for processing information. Three of the best documented heuristics are availability, representativeness and anchoring. The availability heuristic refers to people’s tendency to interpret ambiguous information in terms of what is most easily remembered (Ross and Sicoly, 1979; Taylor, 1982; Tversky and Kahneman, 1973). The heuristic of representativeness refers to people’s proclivity to exaggerate similarities between one event and a prior class of events, typically leading to significant errors in probability judgements or estimates of frequency (Kahneman and Tversky, 1972, 1973; Tversky and Kahneman, 1982). The heuristic of anchoring refers to an estimation of magnitude or degree by comparing it with an ‘available’ initial value (often an inaccurate one) as a reference point and making a comparison (Fiske and Taylor, 1984: 250–6, 268–75).

All three heuristics help explain crucial errors in estimation and judgement by mapping the effects of prior mental states. Availability and representativeness provide a convincing account for the tendency of Israel’s leaders to relate Arab threats to the Nazi holocaust, despite differences in capacity and context. These heuristics affect not only estimates of probability but also the representation of the problem Israel’s leaders use. Anchoring accounts in part for Arafat’s failure to recognize the significant differences between the offer made by Prime Minister Barak in the autumn of 2000–one which included a return of approximately 90 per cent of the West Bank to Palestinian sovereignty as well as all the Palestinian neighborhoods in Jerusalem–and the far more limited offers made by earlier leaders.

Cognitive biases also lead to errors in attribution. The egocentric bias, which leads people to exaggerate the likelihood that the actions of others are the result of their own prior behavior and to overestimate the extent to which they are the target of those actions, contributes to an escalation spiral in a conflict environment. When people exaggerate the importance of dispositional over situational factors in explaining the behavior of others, and attribute greater coherence and meaning to others’ behavior, and to stimuli in general, than reality warrants, they commit the fundamental attribution error (Fiske and Taylor, 1984: 72–9; Kahneman et al., 1982; Nisbett and Ross, 1980). The fundamental attribution error makes it more likely that the leaders will attribute hostile intentions to others and that they will discount the situational constraints other leaders face. This kind of reasoning can fuel mutual perception of intentional hostility that is unwarranted by behavior, exacerbate security dilemmas, and spur arms races and escalation to violence.

Research suggests that the fundamental attribution error is most likely to occur when the observed behavior is consistent with prior beliefs about the actor (Kulik, 1983). Consistent behavior is especially likely to be attributed to dispositional factors even when there are compelling situational explanations. Dispositional attributions are likely to be highly accessible for consistent behavior and the readily available belief that the actor is ‘just the kind of person’ who would engage in such behavior may make people insensitive to situational factors that might have evoked consistency. The representative bias may therefore help to explain the occurrence of fundamental attribution errors when observed behavior is consistent with prior beliefs. This kind of reasoning can sustain and deepen hostile representations of others’ intentions, and make conflict escalation more likely and conflict resolution more difficult.

Impact of Embedded Enemy Images on Conflict

Once schema are embedded, they are extraordinarily difficult to change. This is particularly true of hostile imagery. An image refers to a set of beliefs, or the hypotheses and theories that an individual or group is convinced are valid. An image includes both experience-based knowledge and values, or beliefs about desirable behavior (Vertzberger, 1990: 114–27.) Insofar as enemy images contain an emotional dimension of strong dislike, there is little incentive to seek new information (Druckman, 1994: 50, 63). Stereotyped images generate behavior that is hostile and confrontational, and increase the likelihood that an adversary will respond with hostile action. A cycle of reciprocal behavior then reinforces adversary images by providing allegedly confirming evidence of hostile intentions. Adversarial images tend to become self-fulfilling and self-reinforcing and can fuel spirals of international conflict (Chen and Bargh, 1997: 541–60; Pruitt and Rubin, 1986: 117–18).

Research has established at least three different schemas of enemies: imperials, barbarians and degenerates (Herrmann and Fischerkeller, 1995). Throughout the Cold War, the Soviet leadership saw the United States as an ‘imperial’ enemy, Chinese leaders have at times stereotyped others as ‘barbarians,’ and the Ayatollah Khomeini in Iran described Western leaders as ‘degenerates.’ These schemas informed the representation of problems that leaders in the Soviet Union, China and Iran constructed, and influenced the way they processed information and estimated probabilities, and the choices they made.

Cognitive processes tend to support stereotypical images and biases once they are established. People make heavy use of social stereotypes in predicting one another’s personal characteristics and behavior. Drawing on stereotypes, they also tend to make social predictions with greater subjective certainty or confidence than can be justified by their objective accuracy (Brodt and Ross, 1998). Behavioral memory retrieval processes operate differently when judging ingroup and outgroup members, contributing to a perpetuation of social biases. Implicit theories of others were linked to individual differences in evaluative processing (Hong et al., 1997; Sherman et al., 1998). Attribution processes also influence judgements of discriminatory behavior (Burgess and Borgida, 1999). All these processes work to reinforce rather than mitigate international conflict (Little and Smith, 1988).

Mitigating Factors

Although these are systematic errors and biases, their impact can be mitigated in part by the institutional and group setting in which decision-making takes place. The group context of the decision may mitigate or accentuate individual biases. Group processes can reinforce and strengthen the tendency to discount inconsistent information, by appealing to group solidarity, or deliberately structure processes that allow and indeed encourage early challenges to prevailing beliefs (t’Hart et al., 1997). I return to this argument when I examine the dynamics of collective choice. At a more general level, transparent systems, which allow for scrutiny and accountability, can help to reduce the impact of some of these biases by forcing inconsistent information into the system in a timely manner (Lerner and Tetlock, 1999; Tetlock and Lerner, 1999). A political system where powers are shared helps to reduce these kinds of biases by leaders while the kind of isolation characteristic of leaders of authoritarian states–Hafez al-Asad in Syria and Saddam Hussein in Iraq–reinforces biases. Ceteris paribus, democratic and pluralist systems can better mitigate these kinds of biases than closed authoritarian political systems.

Revision of Schemas, Inference and Estimates

Stability in enemy images is the default and change the exception. Yet conservatism does not hold unconditionally. Schema do change, although they generally tend to change gradually over time, rather than undergo quick and far-reaching conversion. Schema theory has not yet developed an integrated set of propositions about why schema change. In large part because schema theories focus on whole schemas, they are relatively static (Kuklinski et al., 1991). The centrality of schema, their refutability, the diagnosticity of discrepant information, the pattern of attribution and cognitive complexity have all been identified as predictors of the likelihood of revision and, by extension, of change in judgement.

Change is in part a function of the rate at which discrepant information occurs, and how diagnostic leaders and officials consider the information. Contradictory evidence dispersed across many instances should have a greater impact on schema than a few isolated examples (Crocker et al., 1983). As people consider information inconsistent with previous knowledge, they incorporate into their schema the conditions under which the schema does not hold, permitting gradual change and adjustment (Higgins and Bargh, 1987: 386). Important schema are challenged only when there is no other way to account for contradictory data that people consider diagnostic. Greater change will occur when information arrives in large batches, rather than bit by bit. President George Bush, as I noted, did not change his image of Gorbachev even though the Soviet leader made a series of unilateral gestures to the United States. Only when information about large changes arrived in a rush, did he finally alter his well-established image. Even the strongest schema cannot withstand the challenge of strongly incongruent information or a competing schema that fits the data better (Markus and Zajonc, 1985).

Significant change in schema about another also occurs when subjects are exposed to incongruent information and are persuaded that the behavior is not arbitrary, but reflects the nature of the target. Croatian and Muslim leaders, for example, did not change their image of Serbians because they attributed the change in Serbian policy to the military setback they suffered at the hands of the Croatian forces in Krajina. Change occurs when inconsistent information is attributed to dispositional rather than situational factors. The general tendency to prefer situational rather than dispositional attributions for incongruent behavior explains why change occurs so infrequently.

Change is also a function of cognitive complexity, or the intricacy of the cognitive rules used to process information about objects and situations. Cognitive complexity refers to the structure or the organization of cognition rather than to the content of thought. Complexity has somewhat contradictory effects on schema change. The more complex the cognitive system, the more capable is the decision-maker of making new or subtle distinctions when confronted with new information (Tetlock, 1985). Experts with highly complex cognitive schema are more sensitive to new information than novices with low cognitive complexity, whose schema are likely to be fixed and rigid (Conover and Feldman, 1984). Those who possess multiple judgement dimensions also tend to possess rules of abstraction that facilitate the integration and comparison of information. They tend to produce alternative interpretations of new information, but by using their capacity for abstraction and integration, are able to resolve these ambiguities (Vertzberger, 1990: 134–7). Experts who have more relevant information can more easily incorporate inconsistent information as exceptions and special cases. Incongruent data therefore have less impact on their schema than they would have on those of novices (Higgins and Bargh, 1987). Paradoxically, experienced leaders are more likely to be conservative in their judgements than are novice leaders with little experience in foreign policy.

Framing Effects and Prospect Theory

Cognitive psychology has identified a series of systematic biases that are likely to distort the judgements leaders make when they construct representations of high-stakes international conflict. An important stream of research relates framing not to the individual but to the situation, and predicts strong effects on choice. Prospect theory–the most important corrective to the subjective expected utility variant of rational choice theory–maintains that choice is influenced by how a decision-problem is framed (Kahneman and Tversky, 1979). Framing describes the way in which a choice can be influenced simply by the order in which options are presented or the language that is used to describe the options. Research suggests strongly counter-intuitive results: choice can be manipulated by the order and presentation of the options available, without changing the substance of the problem. That simple frame changes can elicit changes in preferences violates one of the most fundamental axioms of rational choice. The policy implications for the management of international conflict are obvious and strong. When, for example, Secretary of State James Baker had tried, with no success, to persuade Prime Minister Yitzhak Shamir of Israel to join the Madrid Process in 1991, he finally cautioned him to think not about what he would gain if he came to the table, but what he would lose if he stayed away. Shamir agreed to participate.

Baker intuitively recognized an important proposition put forward by prospect theorists. People frame problems around a reference point and consider options from its vantage. When the options they identify are in the domain of gain, people tend to be risk-averse, and when the options they identify are in the domain of loss, people tend to choose the risky option. Risk is consequently not a function of individual predisposition, but of the framing of problems (Bazerman, 1986; Farnham, 1990, 1992; Levy, 1992; Stein and Pauly, 1989; Tversky and Kahneman, 1981).

Prospect theory identifies situational conditions that promote risk-aversion and risk-acceptance. Leaders are especially likely to be more risk-prone than rational choice would suggest when they have not normalized for a (recent) loss, and when they treat small probabilities as the equivalent to certainty. They are likely to be more risk-averse when they are in the domain of gain or have normalized for gain, when they overweigh small probabilities of failure by focusing on salient examples of failure, and when they systematically undervalue large probabilities of success.

Evidence in support of prospect theory is robust across a wide variety of experiments in a broad range of cultures. Prospect theory has been extensively applied to the analysis of foreign policy decision-making (Levy, 1992; Farnham, 1997; McDermott, 1998), and provides a convincing explanation, for example, of President Sadat’s decision to go to war in 1973; he had not normalized for the loss of the Sinai seven years earlier. A threat-based strategy of deterrence by Israel only pushed him further into the domain of loss. It also explains Arafat’s decision to escalate the violence in October 2000; he was in the domain of loss and under-weighted the probability of loss from a return to violence in comparison to the certain loss of full sovereignty over East Jerusalem. Here too threat-based strategies appear to have reinforced risk propensity and contributed to escalation.

Prospect theory provides a decision rule embedded within the cognitive frames leaders construct and suggests propensities to risk-acceptance that can have serious consequences for escalation and dampening effects on bargaining and negotiation. When leaders view their own concessions as losses, and those of their opponent as gains, they will tend to overvalue their own concessions and underestimate those of their opponents (Jervis, 1989). This kind of dynamic makes negotiated agreements more rather than less difficult and is one candidate explanation of why leaders engaged in international conflict so often leave value on the table when they are bargaining. Similarly, when leaders fear loss, they are more likely to choose options that risk escalation (Levy, 1992).

The predisposition to loss aversion under specified conditions is a powerful explanation of the escalation of international conflict and can be usefully integrated into the systematic design of strategies of conflict prevention. It holds promise especially because it locates risk-taking propensity in attributes of the situation as well as in characteristics of the individual leader. An emphasis on leaders’ framing of the problem complements the expected consequences of the fundamental attribution error where leaders tend to overweight personality and undervalue situational determinants of choice (McDermott, 1998: 165–86).

Mitigating Factors

Cognitive psychology developed in the laboratory. Biases and heuristics were discovered in the course of experiments in a highly controlled environment, but the laboratory differs from the real world in a number of significant ways. This raises important questions about whether concepts tested in the laboratory can be transposed to and made operational in naturalistic settings.

Attribution studies typically ask subjects explicitly to determine why an event occurred, whereas most events occur in the absence of specific causal questions. The elicitation procedure of structuring the problem and defining the tradeoffs may systematically bias the results of the studies (Enzle and Shopflocher, 1978; Pyszczynki and Greenberg, 1981: 31). Moreover, the tasks subjects are asked to perform in most prospect theory experiments are generally trivial and highly structured. They are also unrelated to other tasks and judgements the subjects are likely to perform outside the laboratory. Judgements made by political leaders on foreign policy issues, by contrast, are often interrelated and deal with issues that are significant and highly valued. People’s recognition of the importance of a decision may influence their thoroughness in collecting and evaluating information and choice of decision-making rules (Jervis, 1986). Some cognitive psychologists, however, dispute the proposition that important judgements will encourage the adoption of more explicit and articulated cognitive processes. They argue that laboratory procedures often result in an underestimation of the magnitude of inferential failings (Nisbett and Ross, 1980: 220–2, 250–4).

Biases may matter less in foreign policy decision-making than they do in the laboratory because the continuous environment of foreign policy decision-makers gives them some feedback that often permits them to approach decisions incrementally, repeatedly correcting for past mistakes. Change is most likely when decision-makers receive timely feedback from competitive markets (Tetlock, 1998: 880). This kind of structured environment and timely feedback is, as I have argued, unlikely in the high stakes, complex and uncertain environments characteristic of high value international conflict. My arguments suggests that, other things being equal, systematic psychological biases are likely to be more rather than less important in an explanation of high value international conflict.

Social and political conditions may also work to minimize biases. When controlled political systems become more open, for example, leaders may learn more about the intentions and the range of choices available to others and, consequently, shift their reference points and give greater weight to attributes of others in explaining change. Again, ceteris paribus, democratic pluralist political systems are most likely to mitigate these kinds of biases.

Psychological Explanations of Collective Choice

Group Decision-Making

Social psychology, through its analysis of small group dynamics, helps, under specified conditions, to explain important decisions about international conflict. These kinds of explanations are useful when problems are ill-structured, when a single small group is at the apex of the policy-making process, or when policies develop out of the interplay among a number of groups (t’Hart et al., 1997).

Of particular interest is how groups arrive at a collective representation of a decision problem. Groups may tend toward ‘simple conformity,’ where collective discussion works to minimize differences in order to construct a shared representation of a problem and undifferentiated analysis of the advantages and disadvantages of alternatives. They may do so, for example, through a process of ‘anticipatory compliance,’ the tendency of those lower down in the political or administrative hierachy to adopt problem representations that conform to the real or perceived predispositions of senior decision-makers (Stern and Sundelius, 1997). They may also do so to preserve group solidarity and maximize group cohesiveness.

This kind of collective decision-making is especially likely to occur when policy-makers feel threatened and are under stress. It produces pathologies in policy-making: inadequate attention to alternative problem representations, unduly short search for information and discounting of inconsistent information. Pressures to conform within groups reduce their capacity to develop differentiated representations of problems, debate values and resolve conflict. Although key explanatory variables are frequently difficult to operationalize and measure in real-world policy-making systems, the expected ‘flawed’ processes have been widely documented across a variety of crisis decision-making contexts, in diverse political systems, across cultures.

Mitigating Factors

Not all groups develop these kinds of tendencies toward conformity. A central differentiator is whether or not leaders intervene actively to establish norms and processes that actively promote debate about alternative representations of a problem and, more generally, signal their tolerance of dissent, especially at the early stages of decision-making. A culture of deliberation and argument flourishes only when leaders actively promote debate and differentiation. Generally, balanced critical deliberation tends to be associated with more open, pluralistic, and facilitative leadership styles that are more likely to be found in democratic cultures that emphasize norms of accountability. These kinds of deliberations are especially likely to occur when there is a rough balance of power and policy-making resources among group members. When access to and control of information is roughly evenly distributed, group membership is heterogeneous rather than homogeneous, and expertise is not the monopoly of one or two members, pressures to conformity are likely to be less (t’Hart et al., 1997).

Social Psychological Explanations of Conflict

Group Identity and Conflict

International conflict grows not only out of the interaction among states and their leaders, but also increasingly out of the violence among ethnic groups that spills across international borders. In the past ten years far more people have been killed in civil wars than in inter-state wars, and it is civil wars that have provided the greatest challenge to international institutions struggling to manage conflict. Social psychology addresses the dynamics of conflict among groups and processes of conflict management, reduction and resolution. It pays particular attention to incompatible group identities as a permissive context of conflict.

Two important bodies of scholarship in international relations challenge the importance of inter-group differences and incompatible group identities as significant contributors to violent conflict. Structural explanations of conflict generally give little attention to the processes that mediate between attributes of the environment and behavior. Realist explanations that focus on competition for scarce resources or changes in patterns of alignment assume that conflict can be explained independently of the collective identities of contending groups. They treat collective identities as epiphenomenal.

A second body of scholarship uses rational choice models to explain the resort to violence as an optimal response to collective fears of the future (Fearon and Laitin, 2000; Lake and Rothchild, 1996; Posen, 1993). As groups begin to fear for their safety, strategic dilemmas arise that are exacerbated by information failures and problems of credible commitment, and, fueled by political entrepreneurs, conflict explodes into violence. Violence becomes a rational response to strategic dilemmas fueled by fear. Here, rational choice explanations are compatible with psychological explanations insofar as they develop the intervening mechanism that transforms fear into violence. Lake and Rothchild argue, for example, that ethnic activists deliberately play on fears of collective insecurity, which are in turn magnified by political memories and anxieties.

Social psychology addresses the origins and triggers to the collective fears that prepare the ground for violence. Converging streams of evidence from social psychology, cultural anthropology, international relations and comparative politics suggest that individuals and groups are motivated to form and maintain images of an enemy as part of a collective identity even in the absence of solid, confirming evidence of hostile intentions.

Enemy images can be a product of the need for identity and the dynamics of group behavior. Social psychologists have identified a fundamental human need for identity–the way in which a person is, or wishes to be known by others; it is a conception of self in relation to others. One important component of individual identity is social identity, or the part of an individual’s self-concept which derives from knowledge of his or her membership in a social group or groups, together with the value attached to that membership (Tajfel, 1981: 255). Social psychologists suggest that people satisfy their need for positive self-identity, status or reduction of uncertainty by identifying with a group (Hogg and Abrams, 1993: 173). These needs lead to bolstering and favorable comparison of the ‘in-group’ with ‘out-groups’ (Brewer and Schneider, 1990: 169–84; Hogg, 1992; Messick and Mackie, 1989; Tajfel, 1982; Tajfel and Turner, 1986: 7–24). Membership in a group leads to systematic comparison and differentiation, and often to derogation of other groups.

The most striking finding of social psychologists is that social differentiation occurs even in the absence of material bases for conflict. This need for collective as well as individual identity leads people to differentiate between ‘we’ and ‘they,’ to distinguish between ‘insiders’ and ‘outsiders,’ even when scarcity or gain is not at issue. In an effort to establish or defend group identity, groups and their leaders identify their distinctive attributes as virtues and label the distinctiveness of others as vices. This kind of ‘labeling’ responds to deep social-psychological needs and can lead to the creation of enemy stereotypes. An examination of massive state repression leading to group extinction, for example, concluded that genocides and politicides are extreme attempts to maintain the security of one’s ‘identity group’ at the expense of other groups (Harff and Gurr, 1988).

Ethnocentrism draws on myths that are central to group culture and breeds stereotyping and a misplaced suspicion of others’ intentions (Booth, 1979; Eberhardt and Fiske, 1996; Fiske, 1998). Strong feelings of self-group centrality and superiority, however, do not necessarily culminate in extreme or violent behavior. The critical variables are the kinds of environments in which groups seek to satisfy their needs and the norms that they generate and accept. Certain kinds of international and domestic conditions facilitate the collective formation of enemy images (Taylor and Moghaddam, 1987).

Mediating Factors

Social identity and differentiation do not lead inevitably to violent conflict (cf. Mercer, 1995). If they did, conflict would occur at all times, under all conditions. First, personal and social identities are often in tension with one another. By identifying strongly with a group, people inevitably de-emphasize their individual identity, and those with a strong sense of individual identity give less weight to their group identities. Human rights activists, for example, characteristically identify less with a particular group and more with norms of individual responsibility. Second, people also generally identify with several groups and typically identify with a group whose importance is salient in a given situation (Turner et al., 1987). Which group identity is activated is situationally specific.

The critical question is under what conditions identity and violent conflict are related. Why are relationships among some groups so much more competitive–and violent–than among others? Hutus and Tutsis have engaged in violent conflict six time since 1962 while Quebecois and Anglophones in Canada, despite important and deep differences between the two groups, have not fought for over two hundred years. Moreover, substantial numbers of Quebecois also share multiple identities, including strong and positive identification with Canada. What explains why strong group identity precipitates violent conflict only in some situations?

The answer may lie partly in the stability of democratic pluralist systems where citizens have internalized norms of conflict resolution. Canada is a stable democratic system whereas the successor states to the former Yugoslavia were not. Even stable democratic systems, however, have experienced violent conflict. Much depends on the variability of identity. Social identity is not given; social learning theorists and constructivists argue that it is constructed through membership in a group and through interaction with others (Bandura, 1973; Harre, 1986). The patterns of identity formation and mapping are critical. Conflict does not develop when the sources of identities or the identities themselves are compatible. When the identity an individual chooses is incompatible with the identity imposed by others or the social context in which identity is constantly being recreated, conflict can develop. Muslims living in Bosnia–Herzegovina, for example, defined themselves as Serbs or Croats until the 1970s, when the Serb and Croat identities began to be recreated to exclude Muslims. Only then did they begin to define themselves as Bosnian Muslims with a distinct political identity. Even then, however, incompatible political identities may not be sufficient to create violent conflict. To return again to the Canadian example, some Quebecois see fundamental incompatibilities between being Quebecois and Canadian, but do not consider a resort to force. They do not because they are committed to norms of fairness and due process, and they expect that these commitments will be reciprocated by their counterparts in English Canada (Stern, 1995).

Several important conditions have been identified that sharpen identity and prepare the terrain for violent conflict. The first set of factors operate between groups within incompatible identities, while the second set is internal to the groups. Ethnic or national identity intensifies during periods of social, economic, or political crisis, when uncertainty grows and the mechanisms in place to protect one group from another lose their credibility (Lake and Rothchild, 1996: 43). As central authority declines in the context of socio-economic or political crisis, fears about physical security grow, and groups invest in measures to protect themselves, making the violence they fear more likely (Posen, 1993). State weakness, its perceived incapacity to protect one group from the anticipated violence of another, is an important trigger of violence among groups with incompatible identities.

Identity conflict is often a competition for ownership of the state and control of its resources. States can stand above and attempt to mediate conflict–by, for example, giving representation to different groups as in Belgium–or be the creature and the instrument of one exclusive group, as in Nigeria where the Hausa Fulani dominate the military regime (Brass, 1995; Gurr, 1993). The expropriation of the identity, symbols and resources of the state by one group to the exclusion of others is a strong predictor of the likelihood of violence.

Conflict can trigger violence among groups under conditions of scarcity. Some evidence suggests that culturally and physically similar groups can generate hostility and aggression toward one another due to competition for scarce resources (Sherif, 1966). Some analyses of civil violence similarly conclude that relative deprivation is the most important condition for participants in collective violence (Gurr, 1970: 12–13). As the gap grows between material expectations and assets, aggression toward those perceived as the cause of relative deprivation grows and intensifies. The competition for scarce resources is exacerbated when the state actively controls the distribution of important resources. In the former Yugoslavia, for example, Slovenians and Croatians actively resented federal redistribution of resources to poorer regions of the country. Loss aversion is likely to intensify when groups compete for scarce resources in a context of decline: when expectations remain stable, but capabilities decline, prospect theory expects that people who are experiencing a decline in their assets or ‘loss,’ are especially likely to make risky choices (Gurr, 1970: 46–50; Stein and Pauly, 1989). Yet, the Czechs and Slovaks competed for scarce resources and divided assets without a resort to violence. Competition for resources and relative deprivation, a sharpened version of competition, cannot satisfactorily account for violence among groups with differentiated and competing identities.

Conflicts of identity are likely to escalate to violence when group members consider that recognition of another’s identity can compromise their own, when they perceive the granting of rights to the other as an abdication of their own identity, and when they fear that the other group may move pre-emptively to make gains at their expense. Throughout much of its history, the Israeli–Palestinian conflict has been this kind of existential conflict; because both identities are tied to the same territory, leaders on both sides long felt that acknowledgement of the other’s identity would fundamentally compromise their own (Kelman, 1982: 61). When one or the other group has attempted, for example, to seize territory and establish a presence on contested ground, violence has resulted. When the state is too weak or unwilling to constrain pre-emptive action by one group, the other becomes more fearful, loses confidence in institutional arrangements, deepens the perception of the hostility of their ethnic rival, and prepares for violence. The intense violence between Palestinians and Israelis in 2000 did not erupt after the failure of bargaining, but when the Palestinian leadership perceived that Israel was tangibly asserting sovereignty over holy places in East Jerusalem. It had never normalized for the loss of sovereignty, and Israel’s position accentuated the Palestinian sense of loss.

Leaders of ethnic groups manipulate group fears to solidify their positions within their own ethnic community. Ethnic activists, with a strong need to identify with their ethnic group, manipulate identities and fears to produce a rapid and spontaneous process of social polarization that magnifies hostility and fear among groups (Fearon and Laitin, 2000). As polarization proceeds, members of an ethnic group are pressured by their leaders–and by the reciprocal intensification of hostility in the other group–to identify only with their ethnic group and to break any cross-cutting ties. In the former Yugoslavia, for example, despite a high degree of social integration among Croats and Serbs, ethnic activists were able to initiate a process that broke apart families and forced members to self-identify with a single group. In a related process, ‘political entrepreneurs,’ who see opportunities for political gain, may take advantage of a process of social polarization to achieve political ends. They deliberately reinterpret histories and traditions to sharpen ethnic differentiation, heighten grievance and increase fear (Ranger, 1983). Slobodan Milosevic was both an ethnic activist and a political entrepreneur: he exaggerated Croatian violence against Croatian Serbs and the Muslim threat to Serbia in Kosovo as a pretext to consolidate and expand the political power of the Serbs when the state structure of Yugoslavia weakened following Tito’s death.

‘Spoilers,’ or militant ethnic activists have also fomented social polarization when new political arrangements that would cut across ethnic cleavages seem likely. After the moderate Hutu and Tutsi reached a painful compromise on new arrangements for political leadership in Rwanda, the militant Hutus, anticipating their exclusion from political power and marginalization, deliberately planned the assassination of the moderate Hutu leadership and a genocidal campaign of violence against Tutsis. The Rwandan genocide is often mistakenly explained as the result of competition for scarce resources, or the weakening of the state structures, or a primordial rivalry between the two dominant ethnic groups. None of these is a sufficient explanation of the outburst of genocidal violence. Militant leaders who feared marginalization and loss from institutional arrangements that would have dampened polarization chose to execute others rather than to accept a diminished political status. They were able, however, to mobilize support for genocidal action because they expertly played on long-standing ethnic fears.

Entrepreneurial leaders or elites whose domestic support is uncertain or threatened can manipulate identities to bolster political loyalty. A leading non-governmental observer of human rights concluded that ‘time after time, a proximate cause of violence is governmental exploitation of communal differences … The “communal card” is frequently played, for example, when a government is losing popularity or legitimacy, and finds it convenient to wrap itself in the cloak of ethnic, racial, or religious rhetoric’ (Human Rights Watch, 1995: viii).

To gain public support, parochial interest groups that benefit from militarist or imperialist policies create strategic rationalizations or ‘myths.’ Over time, some elites come to believe the myths that they have learned, making these images extraordinarily resistant to change. A process of myth-making that perpetuates hostile imagery is most likely when concentrated interest groups trade and log-roll (Snyder, 1991: 2–6, 31–49). The salience and intensity of identity myths are closely tied to the perceived stakes of ethnic relations (Esman, 1986, 1994). The greater the gap between expectations and capabilities, the more important the values that are endangered by declining capabilities, and the smaller the range of other satisfactions that can compensate for the loss in assets, the more receptive populations are to elite attempts to manipulate identities (Gurr, 1970: 59).

Differences in domestic political conditions make some kinds of populations more receptive to elite manipulation than others. In controlled political regimes, leaders and elites who dominate the instruments of communication can more easily manipulate identities and mass images. Not only the kind of regime but also the organization of society has an impact on the creation of hostile imagery. The hallmark of a deeply divided society, that is likely to sustain significant hostile imagery and experience violent conflict, is the presence of separate structures, organized on the basis of identity, that infuse every aspect of society. In Lebanon, for example, political office from the center to local levels traditionally has been allocated on the basis of religious identity. In these kinds of societies, creation and maintenance of ethnic stereotypes and enemy images is easily done.

This analysis suggests that differentiated identities are not themselves a cause of violent conflict. Even when incompatible identities are present, violence is likely only when it is triggered by the exclusionary acts of leaders, either by monopolizing the resources of the state against groups within their own societies, or to press claims against those within others. Leaders and elites evoke threats to political identity that then provoke stereotyping and contribute to violence.

Psychological Explanations and Conflict Resolution

Hostile imagery must change if enduring conflict is to be reduced and resolved. Inter-state conflict has been managed and routinized without modification in elite, much less public images, but recurrent civil violence as well as bitter inter-state conflict cannot be resolved unless images change and leaders and publics learn. The process must also be reciprocated. Once leaders or groups begin to change their image of their adversary and are interested in attempting to resolve their conflict, they must change the image their adversary has of them if conflict reduction is to make any progress.

Strategies of conflict resolution that focus only on competing interests will likely not be sufficient to provoke the learning that is fundamental to the change of hostile imagery and identity conflict. In both enduring inter-state rivalries and bitter ethnic conflict, interests are shaped by images and beliefs that in turn are partially shaped by identity. What we see as a threat is a function in large part of the way we see the world and who we think we are.

If threatened identities facilitate the creation of hostile imagery and contribute to violent conflict, then securing these identities must be a fundamental component of conflict resolution. If they are to be effective, peace-makers who confront bitter civil wars or enduring state rivalries must address interests in the broader context of images and identity. In the former Yugoslavia, the conflict could at best be managed temporarily by territorial partition and safe havens. The conflict can be resolved only if the parties recognize the legitimacy and the permanence of the others’ identities. President Sadat’s recognition of Israel’s legitimacy was the critical key that unlocked the long and difficult peace process between Israel and its neighbors.

In conflict between states, reciprocal recognition of legitimacy and renunciation of the use of force can most directly secure threatened identities and reshape images. Civil conflicts may be more difficult to resolve, in part because of the proximity of clashing identities and the intensity of fear and emotion (Crawford, 2000: 150; Stedman, 1988). Fractured states can be reconstructed through political separation and mutual recognition of competing identities, through a ‘consociational’ or group building-block approach, where elite leaders accommodate and groups remain distinct with constitutional guarantees, or through an integrative approach, which seeks to forge multi-ethnic coalitions with cross-cutting ties (Sisk, 1995).

Mutual recognition and political separation is the most farreaching strategy of conflict reduction. In 1989, after a brutal civil war that lasted over a decade, leaders of Lebanese religious groups modified the fundamentals of their pre—war consociational bargain. Instead of privileging the Maronite Christian community, Muslims and Christians now share power equally. The bargain still provides for a Maronite Christian president, a Sunni Muslim prime minister, and a Shi’a president of the National Assembly. Political decisions are still made by leaders at the top while their communities remain distinct.

The forging of multi-ethnic coalitions with cross-cutting ties is yet another strategy. This was the principal demand of the Muslim leadership of Bosnia–Herzegovina. The agreement reached in Dayton honors a multi-ethnic Bosnia in principle, but in its political arrangements provides for de facto separation of Bosnian Serbs from Muslims and Croats. Kosovo simmers unresolved as Kosovars press for formal independence and Serbia insists on integration. In all these cases, conflict reduction required more than reciprocation of small concessions in a gradually building process. The core of the solution lies in the often difficult decision by senior leaders to acknowledge, respect and accommodate different identities and share political power.

All these strategies assume that identities and images are fixed and that they must be accommodated as they are. Such a pessimistic assumption is unwarranted. Research in social psychology suggests that individual stereotyping can be overcome, but at times educational and social processes can inadvertently reinforce bias (Fiske, 1998; Lopez et al., 1998; Malo and Olson, 1998; Petty et al., 1998; Slomczynski and Shabad, 1998). Others argue that identity is not given, but that it is socially reconstructed as interactions develop and contexts evolve (Teske, 1997). Benedict Anderson (1991) observed that nations, unlike families and clans where individuals can know the others, are ‘imagined communities,’ whose past, tradition and connections are interpreted and reinterpreted through time. Political identities similarly depend on imagined communities whose traditions are constructed and reinterpreted. Identities can consequently be reshaped and reconfigured as leaders and communities restructure their relationships.

Identities are complex structures, with components that emphasize shared communitarian traditions and norms that usually include emphasis on protection of the weak, social responsibility, generosity, fairness and reciprocity as well as honor, reputation and vengeance. Emphasis given to these different norms varies with the situation. Skilled mediators can emphasize the positive values of responsibility, fairness and compassion as important elements of honor and reputation. Appeal to the ‘best’ in the tradition of an identity may shift the emphasis within an ‘imagined community’ to create the space for fairness and reciprocity which can ultimately change images, reshape interests and culminate in tolerance and recognition of others’ identities.

Psychological Explanations and International Relations Theories: The Scope of Analysis

The evidence of systematic patterns of deviation from norms of rational choice is robust. Although much of the evidence grows out of controlled laboratory experiments, political psychologists have documented these patterns across a wide range of institutional settings and political cultures in the analysis of decisions about high value international conflict. Yet, the dominant theories in international relations use much more limited assumptions about choice and motivation: realists treat states as unitary security maximizers and liberals and neoliberal institutionalists conceive of states as wealth maximizers. Both accept the representation of a problem as given and establish the contours of the decision problem from attributes of the payoff matrix; in this sense, they reason backward. Constructivists do problematize the representation of choice but focus on states as creators of norms. The theoretical–and empirical–gaps among these theoretical constructions and between psychological analyses seem very wide. The gaps may be less than they seem, however, when the boundary conditions of different theoretical approaches are examined more carefully. A central challenge is to identify the scope conditions of analysis. Under what conditions can psychological tools be used in combination with leading theoretical approaches to increase explanatory power and enrich the analysis? Are psychological analyses competitors to the prevailing theories, or can they be useful complements under specified conditions? How can psychological approaches best be embedded within the structural analyses of the dominant approaches?

Rational and Psychological Explanations of Choice: The Scope of Analysis

Realist and liberal approaches embed rational choice at their core. Psychological approaches and rational choice have different comparative advantages. Rational choice is theoretically elegant while psychological approaches provide descriptive accuracy of processes of decision-making. Each has different disadvantages: the evidence from psychological studies is now robust that people are not ‘rational actors,’ except in the most trivial and uninteresting situations, yet psychologists have not yet developed powerful general theories that explain choice. Theorists of rational choice who accept concepts of ‘bounded rationality’ find it easier to engage in conversation with cognitive psychologists.

Rational choice provides a transposable deductive apparatus for the formal analysis of interstate interactions. Rational choice theorists forego descriptive accuracy in specific cases and insist that treating states as if they were unitary rational actors yields dividends in explanatory and predictive power if predictions of the model match behavior. Through simplified representations of strategic interactions, rational choice theories claim to provide powerful explanation and prediction of state behavior that would otherwise be obscured by empirical detail (Morrow, 1995). Rational choice analysis is most successful when decision-problems typically concern examples of well understood classes of events that are formally identical to large numbers of other events. In the analysis of competitive markets, or large organizational systems, rational choice can provide elegant, powerful and at times counter-intuitive explanations.

International conflict occurs in a relatively small universe, however, with small numbers. It is therefore difficult for analysts and decision-makers alike to identify among historical antecedents a significant number of ‘like’ cases on crucial dimensions from which to tease out generic problem representations, probability distributions and metrics of value across dimensions of options. Much rational-choice scholarship on international conflict deals with the apparent complexity and indeterminacy of decision problems in international conflict through a strategy of simplified representation. Such simplified representations are often analytically tractable, in the sense that equilibrium solutions to games, when they exist, yield determinate predictions of behavior (Snidal, 1986). A strategy of simplified representation assumes that the abstraction captures the essence of a decision problem for states or group leaders, notwithstanding their confusion and uncertainty about stakes, options, costs, benefits, likelihoods and the interests, goals and intentions of the adversary. The claim is prima facie plausible when reality and representation are isomorphic, and becomes less plausible as the isomorphisms break down (Green and Shapiro, 1994; Moe, 1979). While simplified representation might be a useful analytical strategy in well-structured decision problems, it is unlikely to help in the analysis of international conflict where problems are almost always ill-structured. In ill-structured decision problems, the choice of game to use as a representation of a strategic interaction (and the specification of players’ preference orderings) represents an arbitrary stipulation that undermines the analytical utility of the representation as an explanatory tool.

Cognitive and social psychology explain how leaders cope with this uncertainty and complexity. They commonly interpret ambiguous situations in the light of personally salient historical experiences, or employ behavioral rules of thumb that reflect idiosyncratic ‘lessons of history’ (Khong, 1992; Lebow, 1981; Neustadt and May, 1986). Since these are idiosyncratic variables that none the less strongly influence problem representation, judgement and estimation, they greatly complicate rational choice analysis. Problem representation is a significant variable in any international conflict and, consequently, it is difficult to justify formal abstractions from real-world decision-problems of the kind rational choice theorists seek. The concepts and methods of cognitive psychology are comparatively well suited to understanding problem representation (George and George, 1956; Hermann, 1974, 1980; Holsti, 1989; Holsti and George, 1975; Janis, 1982; Jervis, 1976; Jervis et al., 1985; Lebow, 1981; Mandel, 1986; Stein, 1992; Stein, 1994).

Analyses of the debates between modelers of deterrence and spirals of escalation illustrate the complementarities between the two approaches. The debate can be better formulated as a problem of scope conditions. Rational choice theorists model deterrence as a set of interdependent rational choices, and have identified counter-intuitive strategic choices to resolve the dilemmas they design. Psychological explanations have focused on problem representation, judgement, motivation and fear, have specified how leaders as cognitive managers cope with ambiguous information in threatening environments, and have argued that the strategies deduced from rational models can be counterproductive and culminate in dangerous escalatory spirals (Achen and Snidal, 1989; Downs, 1989; Fearon, 1993; Huth and Russett, 1990, 1984; Jervis, 1989; Jervis et al., 1985; Lebow and Stein, 1989, 1990a, 1990b, 1994).

Prospect theory helps to resolve these seemingly contradictory findings. When states are fundamentally satisfied with the status quo, they can be considered in the domain of gain, and as appropriate targets for deterrence threats, since they are likely to be risk-averse with respect to gains. Under these conditions, except at the extremes of the probability distribution, rational choice is likely to provide a parsimonious explanation. When, however, states have experienced significant loss, and have not normalized for the loss, deterrence threats are likely to provoke escalation by leaders who are risk-acceptant with respect to losses.

Cognitive psychology and rational choice can be complementary, once the critical scope conditions are established. After the representation of the problem is identified, rather than assumed, scholars can examine the complexity of the environment, the scope of uncertainties, and the attributes of the situation. They can then establish whether leaders are in the domain of gain or loss and assess the likelihood that the calculus will be utility-maximizing, satisficing, or approximate the systematic deviations from rational norms that psychologists expect.

Constructivism and Psychological Explanations

There is a much more natural fit between constructivist explanations of international conflict and psychological analyses. While the conversation with rational choice centers on cognitive psychology, the dialogue of constructivists is with social psychology. Constructivists develop a concept of choice that is deeply informed by leaders’ identity–how they define their state or group, who they are, and how they see themselves in relation to others. Constructivists have expanded the repertoire of psychological explanations of international relations–that traditionally focused on beliefs, images, and judgement of leaders–to include the collective or shared beliefs that constitute a common identity, and processes of norm creation and norm observance. Unlike realists and liberals, constructivists do not take identities and interests as given, but rather as created largely through interaction with others. They build into the concept of identity not only interests but attention to norms as a constitutive element (Finnemore and Sikkink, 1998; Price, 1997; Ruggie, 1998; Tannenwald, 1999).

This emphasis on constructed identities as the explanation of choice makes possible a useful dialogue with social psychological theories that examine the conditions under which norms become the criteria for choice. Social psychology explores the conditions under which criteria of equity, fairness and justice override the maximization of interest as the decision rule. Social identity theory in psychology emphasizes how identity changes a critical component of constructivist arguments.

Social identity theory can enrich and broaden arguments about identity change. Social identity is created not only through interaction with others, as constructivists suggest, but as social psychologists argue, through processes of identification with a group. Whether people identify with a particular social group is a matter of choice and, in choosing an identity, people struggle between the contradictory imperatives of inclusion and differentiation (Brewer, 1993; Brewer and Schneider, 1990). Canada, for example, is part of the larger North American trading system while seeking to protect its cultural distinctiveness. The cognitive mechanisms that produce group identification–or social identity–are categorization and social comparison (Abrams and Hogg, 1990). Categorization sharpens inter-group boundaries and produces stereotyping (Hogg and Abrams, 1988: 19–20).

As I have argued earlier, however, stereotyping does not always lead to conflict. Groups and states have a choice of strategies to deal with identity conflict: they may seek to assimilate to a more favorable identity, they may choose to redefine the value of their identity, or finally, they may choose to compete (Tajfel and Turner, 1979). Some states currently excluded from the European Union emphasize their identity as democratic states observing norms of human rights and respect for legal processes. Their leaders have chosen to assimilate to an identity that they regard as more favorable; identity change becomes a strategy of entry and opportunity. ASEAN members have chosen to emphasize the value of their distinctiveness by creating a group that is defined by the ‘Asian way.’ At the extreme, when groups feel that their identity is threatened, and barriers to inclusion are insuperable, they choose to compete. Social identity theory helps to explain why leaders select from a repertoire of available social representations, and by examining the thickness of barriers to inclusion, helps to identify the conditions under which states and groups are likely to choose one or another strategy of identity change.

Psychology and Theories of International Conflict

Critics of psychological approaches to international politics contend that these approaches may be relevant to individual choice but they are both too limited, too under-determined, too messy, and situated at an inappropriate level of analysis to explain the large systemic patterns of world politics. Proponents argue that evidence of systematic patterns of thinking and choice, both at the individual and the group level, are robust, and that broad deductive arguments, premised on flawed assumptions that violate much of what we now know about how individuals and collectivities define themselves and choose, are unlikely to provide valid foundations for powerful explanations of international and inter-group conflict. This has largely been a dialogue of the deaf.

Psychological approaches are useful in establishing boundary conditions for on-going arguments within structural explanations of systemic patterns. They help to refine and inform explanations that rely implicitly on theories of choice that are unlimited by varying conditions. The ‘messiness’ of psychological theories and evidence can give specificity–and rigor–to big debates about the dynamics of international conflict and its prevention. Psychological approaches also help to specify theories built on identity formation and change and on norm creation and observance.

Psychological theories are also useful in explaining the bounded choices political leaders make to escalate–and to de-escalate–conflict, and in the analysis of patterns of interaction that are the result of linked series of choices. The recurrent puzzle of why bargaining processes, for example, leave value on the table, or fail to produce the agreements that are transparently obvious to the disinterested omniscient observer, is often powerfully explained by psychological theories of choice. A related puzzle–the repeated failure of one set of leaders to read and interpret accurately the signals of another set–is also well explained by psychological theories. Theoretical and empirical progress in the explanation of escalation and de-escalation of conflict can benefit by embedding psychological theories within the dominant rational and constructivist traditions and by carefully examining their complementarities and their relative purchase.