Strategic trust

Arguably the most prominent notion of trust in political science and international relations is what Uslaner (Reference Uslaner2002) calls ‘strategic trust’. In this understanding of the concept, actors trust on the basis of beliefs about others’ interests, cooperating when they believe that others’ interests ‘encapsulate’ their own (Hardin, Reference Hardin2006). Trust is a belief that potential partners have a self-interest in cooperation, generally an incentive in building or sustaining a long-term, mutually beneficial relationship. Tyler (Reference Tyler2001: 287) calls this approach ‘calculativeness’ as it envisions a cold, sober evaluation of the advantages and disadvantages of trust. Whether to trust, argues Hardin (Reference Hardin2006: 22), is a ‘matter of prudential assessment, not moral choice’. Actors make such judgments on the basis of information. Hardin writes that the two central elements of the rationalist account of trust are the incentives of the trusted to honor the trust, and knowledge to justify the truster’s trust (Yamagishi and Yamagishi, 1994: 139; Hardin, Reference Hardin2006: 44). There is of course also strategic distrust, a belief that specific others are trustworthy based on information, albeit incomplete, about their interests.

This is a highly structural account of trust. Strategic trust develops in situations in which actors have an incentive to honor their agreements. It has nothing to do with the attributes of the individuals cooperating. The implication of strategic trust is that even if there is an ongoing relationship of mutual self-interest in one area, those involved draw no general lesson about the moral character of the other into other domains. The relationship might produce incentives for potential partners to honor their commitments in other fields but primarily through the mechanism of a linkage with previously existing areas of cooperation. There is ‘no moral force’ in strategic trust (Uslaner, Reference Uslaner2002: 22). As Uslaner writes, the etymology of strategic trust is ‘A trusts B to do X’ (2002: 21).

International relations scholars have recently begun utilizing this conception of trust explicitly (Kydd, Reference Kydd2005). However, it has a much longer pedigree. This is the kind of trust utilized, often only implicitly, in rationalist work in international relations. It provided the foundation of what was known as ‘neoliberal institutionalism’, an approach pioneered by Robert Keohane and others but which has now reached beyond studies of international cooperation into all elements of state interaction. States are unsure as to whether others will maintain their end of the bargain, uncertain whether others will keep their word. Cheating, which in collaboration games yields a ‘sucker’ payoff, is a key concern (Stein, 1982; Grieco, Reference Grieco1988; Mearsheimer, 1994). Incomplete information stymies potentially rewarding collaboration.

In keeping with this conception, rationalist work on trust focuses almost exclusively on changing the structure of the situation to generate more cooperation, such as by iterating the game or by reducing the number of players in the game to prevent free-riding. By institutionalizing cooperation, international organizations can facilitate this process. Through the provision of information, trust of particular others becomes possible. As shadows of future are lengthened, linkages are created, and reputations put at a premium, the level of trust between egoists increases as their interests come to encapsulate one another. All of these mechanisms focus almost exclusively on creating incentives for trustworthy behavior through adjustments to the strategic environment (Axelrod, 1984; Keohane, Reference Keohane1984; Oye, 1985).

Moralistic trust: generalized and particularized

Trust, however, can also be based on beliefs about the honesty and integrity of potential partners. This is the other basis of trust, what is called moralistic trust. Messick and Kramer (Reference Messick and Kramer2001: 91) are describing moralistic trust when they define trust as making decisions as if others will abide by ordinary ethical rules, most important of which are truth-telling and not harming others. This is a ‘social conception of trust’ in which trusters believe that intentions and behavior reflect traits of the trustee, rather than the situation (Tyler and Degoey, Reference Tyler and Degoey2004: 332). Hoffman (Reference Hoffman2002: 20–22) calls it ‘fiduciary trust’. It is based on ‘an implicit theory of personality’, a belief that others have consistent personalities and traits that do not vary by situation (Sztompka, Reference Sztompka1999: 75; Mercer, Reference Mercer2005: 95). When we trust, we are making judgments about the character and integrity of partners (Yamagishi and Yamagishi, 1994: 132; Larson, Reference Larson1997: 22; Booth and Wheeler, Reference Booth and Wheeler2008: 243–245). When moralistic trust is present, the truster does not need to restrict that trust of another to transactions or cooperation of a particular kind, because while interests vary across situations, character does not (Kramer et al., Reference Kramer, Brewer and Hanna2004: 367). Moral attributions allow us to draw broader conclusions about the trustworthiness of others without the need to constantly collect information in every new situation or insist on specific reciprocity. It provides a deeper and more durable foundation for cooperation.

Moralistic trust varies in scope. It might apply to a specific other or most others. If moralistic trust extends more broadly so as to constitute a belief that not all, but most others can be trusted, we can speak of generalized trust. The etymology of generalized trust is ‘A trusts’ or ‘A is trusting’ (Uslaner, Reference Uslaner2002: 4). Generalized trust rests on a general belief in the honesty and benevolent character of others. When Rotter defines trust as the ‘generalized expectancy held by an individual that the word, promise, oral or written statement of another individual or group can be relied on’, he is talking about generalized trust (Rotter, Reference Rotter1980: 1).Footnote 2 Generalized trust is a belief in human benevolence (Cook and Cooper, 2003: 215).

Generalized trust must, by definition, be moralistic, as we cannot collect information about the interest of all prospective interaction partners. As Sztompka (Reference Sztompka1999: 70) writes, ‘Trustfulness may incline people to grant trust, and suspiciousness to withhold trust, quite independently of any estimate of trustworthiness. The origins of trustfulness or suspiciousness are not epistemological. They have nothing to do with knowledge about the partners of future engagements… .’ Mercer (Reference Mercer2005: 95) writes that it involves ‘certainty beyond observable evidence’. Hardin (Reference Hardin2001: 5), somewhat pejoratively, calls it a ‘non-cognitive’ approach. Better stated, it is ‘non-calculative’ in the utilitarian sense. It is most certainly cognitive in that it based on a set of beliefs about others.

The presence and therefore the relevance of generalized trust might not be immediately obvious or prevalent in international relations, but its absence is at the heart of Waltz’s version of structural realism. The opposite of generalized trust is fear, the presumption that most others cannot be trusted.Footnote 3 Fear has a long association with classical realism, but structural realism is generally understood as relocating realism’s grounds for pessimism about cooperation in international politics from the ‘first image’, what classical realists regarded as human nature, to the systemic level. However, Waltz needs fear, and sneaks it through the back door. As a number of analysts have noted, all major security dilemma theorists, going back to Butterfield and Herz, rely on fear to make their arguments go (Booth and Wheeler, Reference Booth and Wheeler2008: Ch. 2; Tang, Reference Tang2008). Not all self-described realists rely on the assumption of fear, but many do (Rathbun, 2007b; Booth and Wheeler, Reference Booth and Wheeler2008; Tang, Reference Tang2008).Footnote 4

In Theory of International Politics, Waltz (Reference Waltz1979) maintains that the possibility of opportunism given uncertainty about intentions, combined with the lack of any external protector, is enough to induce recurrent patterns of balancing power, the security dilemma, and war. Anarchy and uncertainty preclude the creation of trust and international cooperation. States engage in self-help, manifested in the accumulation of power against potential adversaries, real or presumed, through military build-ups and alliance formation. Waltz (Reference Waltz1979: 105) maintains that ‘the condition of insecurity – at least, the uncertainty of each about the other’s future intentions and actions – works against their cooperation’.

Yet anarchy and uncertainty alone themselves do not generate what Glaser (Reference Glaser1994) calls the ‘competition bias’ in Waltz’s neorealism. Only if actors react to uncertainty in a particular way, by assuming the worst, do the dynamics that Waltz predicts ensue (Booth and Wheeler, Reference Booth and Wheeler2008: 22–23; Tang, Reference Tang2008). There is a presumption that states believe others have malign aims, or must be treated as if they do. Brooks (Reference Brooks1997) calls this a ‘possibilistic’ logic. In Waltz’s realism, states distrust not specifically, but generally. They have a general inclination not to trust. This working assumption is insensitive, perhaps impervious, to revision through accumulation of information, and precludes objective, case-by-case assessments of others’ trustworthiness. The presence of fear is evident in the work of those neorealists who argue that states respond to the problem of uncertain intentions, that is whether others should be trusted, not by thoroughly and carefully parsing the evidence case-by-case in search of information about their objectives, but by balancing power (Grieco, Reference Grieco1988: 500). According to many neorealists, states do not carefully evaluate another state’s intentions because a state can never have enough certainty to fully let down its guard, and intentions can change (Glaser, Reference Glaser1994; Taliaferro, Reference Taliaferro2000: 144–145; Kydd, Reference Kydd2005: 15). The implicit mechanism of fear and the consequent focus on power over information explains why many structural realist scholars have dedicated little attention to issues of signaling and screening, as noted by Glaser (Reference Glaser1994) and Booth and Wheeler (Reference Booth and Wheeler2008: 27). Power, not past action, is how states gauge intentions and judge credibility (Press, Reference Press2005).Footnote 5

It was the implicit replacement of fear with strategic distrust that led Keohane (Reference Keohane1984, Reference Keohane1989), in particular, to such different conclusions regarding the prospects for international cooperation than Waltz. By removing the presumption of a general fear of others’ intentions that is not responsive to information (and generates the concern for ‘relative gains’), states are able to collaborate and might reach mutually beneficial outcomes even in an anarchic environment. Nevertheless, it would be incorrect to label neoliberal institutionalists optimistic or idealistic about the character of international relations; rather they are agnostic. They believe in the contingent possibilities for international relations in particular situations when states have common interests. Trust is specific to particular situations and based on information. Neoliberal institutionalists envision states proceeding skeptically, careful not to be burned, but not to the degree that they miss opportunities for potentially rewarding cooperation. In keeping with their strategic conception of trust, they maintain that whether to trust or fear others is a function of the situation, not a general predisposition on the part of states.Footnote 6

Generalized trust (or fear) is a belief held about the overall trustworthiness of others as a whole. It is a quality of the truster, not the relationship or even the characteristic of the specific target of distrust or trust (Sztompka, Reference Sztompka1999: 97). In contrast, particularized trust characterizes a relationship in which one individual or group believes that a specific other or group is inherently trustworthy in a moralistic sense. The etymology of what Uslaner (Reference Uslaner2002: 28) calls particularized trust is ‘A trusts B’. What we could call particularized fear would be the opposite. Unlike generalized trust, particularized trust is relational in nature, rather than dispositional, even though both are moralistic varieties of trust. Particularized trust is dyadic, perhaps created and certainly sustained by opposing identities. We trust others like us and fear those who are different. Nevertheless, these two types of trust are likely related. It is likely that relationships characterized by particularized fear (trust), for instance, are more likely to develop among individuals with a more fearful (trusting) disposition.

Particularized trust is essential for Alexander Wendt’s (Reference Wendt1992, Reference Wendt1999) version of constructivism. Wendt’s constructivism does not see conflict as an inevitable or general feature of international relations but does maintain that a ‘culture’ of anarchy marked by fear might arise. States are not doomed to repeat forever the cycles of conflict and acrimony as they are in realism. They can embark on a more cooperative path and change the dynamics of their interaction. Trust is the mechanism. It is what Wendt (Reference Wendt1999: 358) calls the ‘fundamental problem of collective identity formation’. While the core variable in constructivism is generally thought to be identity, trust is logically prior to the formation of a common identity (Adler and Barnett, Reference Adler and Barnett1998: 46). One cannot feel part of a community composed of others whom he or she believes might mean him harm. He or she must believe that his or her needs will be respected.

Trust for Wendt is created through interaction. Acts of concession or generosity can create a reinforcing spiral of cooperation. Trust-building might begin through a rationalist process of signaling and conveying information, but the process is more transformative, allowing for a redefinition of self and other from adversary to partner to friend. However, for this process to take hold, states must overcome their fear of ‘being engulfed… by those with whom they would identify’ (Wendt, Reference Wendt1999: 357). One way to provide this assurance is through some sort of institutional constraint, as would be the case in both structural realism and rationalism. But Wendt argues this is not sufficient. States must ‘believe that others have no desire to engulf them, nor would do so out of self-interested opportunism… [T]hen it will be easier to trust that in identifying with them, their own needs will be respected, even in the absence of external constraints’ (1999: 359) (emphasis added). If states trust other states in an environment in which there is no institutional sanction, even if the latter might have a self-interest in defection, trust rests on a judgment about the other’s integrity and honesty rather than the situational incentives of the time. Trust is not a function of the situational exigencies of the moment, but rather predicated on a belief that others have a particular character, that they are inherently trustworthy. The kind of trust operating in this instance must be based on a belief that others feel morally compelled to reciprocate cooperation – that is, moralistic trust.

Of course, Wendt’s constructivism does not deterministically assume that this process will occur and is equally comfortable describing a fearful dynamic overtaking state interactions. It does, however, stand somewhat alone in offering an account for the potential development of deeper cooperation based on moralistic trust. Wendt’s use of this moral variety of trust leads it to the conclusion that the possibilities for cooperation in international relations are far more profound than those foreseen by rationalism, with its focus on strategic trust.

Monadic effects of social orientation: projecting fear and trust

There is evidence that dispositions to trust serve as a core feature underlying political ideology and link domestic and foreign policy positions. Jost et al. (Reference Jost, Glaser, Kruglanski and Sulloway2003), on the basis of a remarkable review and synthesis of studies on the source of differences between left and right, conclude that a significant share, if not most, of the variation in political ideology is explained by the right’s greater level of fear. The left, in contrast, is more trusting. The logical response to a belief that others are untrustworthy is of course to accumulate power to use against those who would do harm (Jost et al., Reference Jost, Napier, Thorisdottir, Gosling, Palfai and Ostafin2007). In domestic affairs, these tendencies exhibit themselves in a tendency for the right to adopt more authoritarian positions on civil liberties. In foreign policy, the right endorses greater military spending and less cooperation in international organizations where others might take advantage. The left puts greater faith in diplomacy and multilateralism (Hurwitz and Peffley, 1987; Chittick et al., 1995; Rathbun, 2007a). Differences on trust appear to structure ideological cleavages and link domestic and foreign policy beliefs, suggesting that individuals’ social orientations serve as core values anchoring their political program in general.

If domestic political ideology bleeds into foreign policy because both are manifestations of social orientation and propensities to trust, social psychology expects and can explain variation and internal contestation over foreign policy based on differences in core beliefs about others. The connection also allows us to observe and measure social orientation at the domestic level independently of its behavioral manifestation at the foreign policy level and avoid tautological reasoning. As certain individuals with particular social orientations or groups of like-minded individuals enter or even dominate a state apparatus, their views are translated into policy.

Even while this is only a monadic effect of social orientation on foreign policy, it still has implications for structural international relations paradigms, such as rationalism and structural realism. Theorists in the these traditions might mistake the foreign policy behavior of a particular type of truster as evidence for the nature of the international system, when in fact what they are capturing is dispositions predicated on certain beliefs about that system emerging from a social orientation that resembles the logic of the paradigm in which they work. A competitor might claim the system made him do it, but that does not mean it did. Only careful analysis will tell.

Of course, in these instances, realists are drawing conclusions about the nature of the system as a consequence of the actions of individuals who represent the state. Because individuals do not have tremendous choice in their eyes, they are able to talk meaningfully of the state as something more than the individuals who occupy its key posts. Given its focus on disposition, social psychology is less inclined to think of states in anthropomorphic terms, and more as a shorthand for those directly and heavily involved in the policy-making process. The presence of variation in social orientation among foreign policy makers in the same structural setting would be evidence for the social-psychological view.

Domestic political contestation based on different social orientations also provides leverage to evaluate a constructivist critique of a social-psychological approach, that the phenomenon I am describing is based on culture rather than beliefs. Wendt argues that there are three potential ‘cultures of anarchy’ marked by dynamics that are roughly similar to the beliefs of the three different social orientations. Wendt distinguishes between ‘role-identities’ and ‘roles’, the former being ‘subjective self-understandings’, the latter being ‘objective, collectively constituted positions that give meanings to those understandings’. Roles are macro-level phenomena that transcend the individual level of analysis and are critical for creating culture. Beliefs ‘come and go as individuals take on or discard beliefs’, whereas roles ‘persist as long as someone fills them’ (Wendt, Reference Wendt1999: 259). Roles take on a life of their own, independent of individuals. Intersubjectivity trumps subjectivity. It is this stress on roles that allows some constructivists, like realists, albeit in a different way, to think of states as ontological units in and of themselves. Again, given its stress on variation in type, a social orientation approach expects more contestation than consensus. To the extent that we witness significant domestic contestation over foreign policy, we are more inclined to think in terms of ideology rather than culture.

However, domestic ideological cleavages are not the only place to observe different types of trust and trusters. Particular societies might be composed primarily of a particular type of truster. In that sense it would be possible to say that ‘states’ have a social orientation. For instance, liberal democratic societies are premised on a more benign view of human nature (Doyle, 1986, 1997; Owen, Reference Owen1997; Rousseau, Reference Rousseau2006). In other words, we might say that mature liberal societies have a predominance of individuals with cooperative social orientations. Uslaner (Reference Uslaner2002: Ch. 8) cites evidence that democratic societies are marked by higher degrees of generalized trust than authoritarian societies.Footnote 10 While some argue that distrust drives liberal democracy, as evident in the importance placed on accountability of elected officials and the necessity of a free press (Levi, Reference Levi1998), recall that generalized trust is the belief that most others can be trusted, but not everyone. Liberalism is distrust of power centralized in the hands of a few individuals, but trust in the broader general public. The king does not get the same benefit of the doubt.

Thinking of liberal democratic states as more cooperative and trusting in social orientation provides a micro-logic to some findings about liberal societies and conflict. Democracies are found to be less inclined to initiate armed hostilities, even against non-democracies (Rousseau et al., Reference Rousseau, Gelpi, Reiter and Huth1996). In other words, there is a monadic effect of democracy in the instigation of conflict. If democracies are indeed marked by generalized trust, this behavior makes sense, as they are less likely to feel that they will be taken advantage of if they enter into diplomatic negotiations. Ceteris paribus, democracies are better able and more likely to make the first conciliatory gesture in an effort to elicit trust. However, unlike a constructivist account that might consider this democratic culture, a social-psychological approach is also comfortable explaining the presence of non-liberals in liberal societies, who are always present (Owen, 1997).

A focus on dispositions to trust also leads to a potentially unique social-psychological approach to the study of international cooperation and the design of international organizations.

As reviewed above, rationalists understand trust in situational terms. It has nothing to do with the attributes or the character of the individuals cooperating. Therefore, rationalist work on cooperation has focused almost exclusively on how changing the structure of the situation can create incentives for trustworthy behavior. Institutions and organizations can help.

The rationalist literature on cooperation, because it relies on strategic trust, specifies a particular cause and effect relationship between international organizations and trust. Distrust drives the creation of international organizations, which are the producers of strategic trust. And where distrust is greater, states create stronger institutions. The argument that states pool or delegate sovereignty or make firmer commitments to circumvent defection is a common one in the rationalist literature (Snyder, 1984; Garrett, 1992; Pollack, 1997; Abbott and Snidal, 1998; Moravcsik, 1998, 2000; Lake, 1999; Weber, 2000).

This is a plausible argument, but social psychologists would suggest a different mechanism. Significant international cooperation, in which states delegate or pool sovereignty, involves placing one’s fate in the hands of others. Particularly in highly uncertain situations, in which numerous issues are potentially at stake among a large number of partners, it becomes difficult, if not impossible, to predict one’s interests and the likely alignments of others. Surrendering some degree of control over policy in these instances would be an indication of generalized trust, not distrust.

Social psychology inverts the relationship seen by rationalism. Trust (better stated, generalized trust) precedes and is necessary for the creation of international organizations. Generalized trust is a ‘springboard’ in uncertain situations to leap into the ‘outside world’, ‘emancipating’ us from the secure confines of stable relationships and allowing us to seek other cooperative partners with whom there might be greater gains (Messick and Kramer, Reference Messick and Kramer2001: 108; Yamagishi, Reference Yamagishi2001: 141). Experimental research shows that in prisoner’s dilemma games in which subjects are given the option to enter or leave the game, it is trusters who tend to play and keep playing (Orbell et al., Reference Orbell, Schwartz-Shea and Simmons1984; Orbell and Dawes, Reference Orbell and Dawes1993). Non-trusters, in contrast, tend to leave games or form committed relationships among specific others with whom they are familiar (Yamagishi et al., Reference Yamagishi, Cook and Watabe1998).

By this logic, individuals with different social orientations will demonstrate different propensities to build international organizations and have different preferences over institutional design. Cooperators will be less isolationist and more inclined to pool or delegate sovereignty. As compared to cooperators, competitors are likely to insist on a veto in institutions or flexibility clauses that allow members to withdraw or limit the term of an agreement. They will want to restrict membership so as to lower the number of members and allow for easier monitoring and enforcement. These dispositions to trust are best seen by looking for variation amongst the preferences of similarly situated actors, for instance in domestic political settings, something that rational design research cannot explain (Olson, 1971; Snidal, 1985; Kahler, 1992; Abbott and Snidal, 2000; Koremenos et al., 2004: 24). Interestingly, even where social-psychological type affects foreign policy in a monadic way, in terms of approaches to international cooperation, it might produce an outcome – an international organization – that helps govern the international system.

Dyadic effects of particularized trust and fear: rivalries and the democratic peace

Trust and fear, more specifically particularized trust and particularized fear, are implicit in accounts of arguably the two most important dyadic phenomena in international relations – rivalries and the democratic peace. The former concerns itself with spirals of fear, the latter with spirals of trust. The logics of both apparently rest on a social-psychological foundation.

The international relations literature on rivalries proceeds from a startling observation – that the majority, perhaps even the predominance of international conflicts, whether they be crises, or wars, take place between a small number of dyads (Goertz and Diehl, Reference Goertz and Diehl1993; Thompson, Reference Thompson1995). These rivals have relations that at first glance appear to be the type expected by structural realism and security dilemma theorists. Defensive actions on the part of state A are understood as potentially offensive by state B and are met in kind. This causes concern on the part of state B and the process can begin a spiral of escalating hostility (Thompson, Reference Thompson2001: 561–562). However, the central premise of rivalry theory is that conflicts between rivals taken own a ‘life of their own’ (Thompson, Reference Thompson1995: 196) due to the ‘psychological baggage’ (Colaresi and Thompson, Reference Colaresi and Thompson2002Reference Colaresi and Thompsonb: 1182) that comes with repeated crisis and conflict. Vasquez (Reference Vasquez1996: 532) defines rivalries as relationships ‘characterized by extreme competition and… psychological hostility…’. Thompson (Reference Thompson2001: 561) understands it as ‘very much a social-psychological process’.

Although not always explicit, the psychological baggage that accumulates in rivalries seems to be what I would call particularized fear, an expectation on the part of participants that a specific other is inherently untrustworthy. Colaresi and Thompson (Reference Colaresi and Thompson2002a: 271) write, ‘In a rivalry context, as opposed to more moderate crisis situations, uncertainty is likely to take on a different meaning…Within rivalry, this may increase the fear of a dangerous environment where escalation is inevitable and thus encourage hyper-vigilance. Thus, under increased insecurity, rivals are likely to prepare for the worst.’ This is not strategic distrust, in which states are simply objectively updating their conception of others’ interests based on prior experiences. Rivals have categorized others as enemies who are intent on harm. They have made attributions about others character, not their interests. Rivals have ‘had time to develop images of their adversaries as threatening opponents with persistent aims to thwart their own objectives’ (Colaresi and Thompson, Reference Colaresi and Thompson2002a: 263–264).

The psychological nature of distrust and fear in rivalries is evident in how it affects both rivals’ information processing and their conception of their interests. The non-rationalist premise on which rivalry theory is built is evident in the fact that rivalry theorists understand their subjects as prone to cognitive biases based on prior experience, biases that distract them from their objective interests. Colaresi and Thompson (Reference Colaresi and Thompson2002a: 269) write that ‘a rivalry context with its increased mistrust and expectation of conflict will affect the way objective information is processed, all else being equal… . Information passed between rivals must be processed within the conflictual setting of that relationship.’ The difficulty of transcending a relationship marked by an untrustworthy image of an adversary has been demonstrated by attribution theorists. Social psychology tells us that cooperative and peaceful demonstrations of intent on the part of an adversary are attributed to situational imperatives, whereas hostile actions are attributed to an untrustworthy disposition (Jervis, Reference Jervis1976; Mercer, Reference Mercer1996). Once in place, these images are difficult to break free from. ‘The expectations become more rigid, less sensitive to changes in adversary behavior, and less in need of continued reinforcement’, writes Thompson (Reference Thompson2001: 561–562).

Rivalry research is also non-rationalist in that the nature of conflict in rivalry is different than in other types of conflictual relationships. Vasquez writes that rivals are ‘governed primarily by their attitude toward each other rather than by the stakes at hand’. Vasquez (Reference Vasquez1996: 532) calls this ‘negative affect’. This would include not only fear – the belief that others are untrustworthy and intent on harm. It would generate positive utility for any harm to the other regardless of the costs to the self. We might call this hate, or what social psychologists have labeled as an ‘aggressive’ social orientation. Whereas ‘normal conflict’ is marked by cost–benefit analysis, rivalries ‘get out of hand’ and can lead to what would normally be considered self-defeating behavior, according to Vasquez. He writes, ‘The relationship is difficult to change because each side is involved in a vicious circle in which hostility makes actors define issues in ways that are intractable and threatening and actors become hostile, in part, because of the way they have defined the issues that divide them’ (1996: 532–533). In other words, rivals transform the game objectively given by the strategic circumstances.

Rivalries have a beginning and grow. They are not a product of objective, structural features of the strategic setting. Rivalry research is based on the premise that the past crises and wars between a particular set of states have an effect on future ones, independent of the structural features of the interaction such as relative capabilities or the persistence of the issues at stake. Crises and wars between the same pair of states cannot be considered independent of one another. In the language of trust, rivalries are marked by particularized fear that arises from the history of conflict. This kind of fear is moralistic in that participants have made a judgment about inherent traits of the other side. However, it is not dispositional in that it is not applied beyond that particular other. It is a quality of the relationship not the objective strategic situation or the actors involved. Continuing conflict reinforces and buttresses fear of the other side.

Given that particularized trust is a quality of a relationship and not reflective of the disposition of those in the relationship, constructivism is of considerable help in explaining the persistence of rivalries once constituted. Rivalries are, in Wendt’s terms, roles that require an ‘other’ in order to be taken up. Once constituted, a rivalry is an identity relationship based on an image of the other that is hard to break regardless of the different preferences of state leaders. Colaresi (Reference Colaresi2004) finds that efforts to initiate rapprochement in rivalries fail at greater rates than in other protracted conflicts, and that unreciprocated efforts result in higher likelihood of losing office. The same is not true in non-rivalry settings, where there is no effect of failure on leadership tenure. Rivalries create an intersubjective cultural context that is irresponsive to the subjective beliefs of individuals.

However, given that rivalries are created through a process of interaction, the features that sustain rivalries cannot explain their origins or their end. The latter must be exogenous. On this basis, rationalists have concluded that there is no reason to believe that rivalries have psychological dynamics. As rivalries only develop over time, after somewhat persistent conflict, the beginning of conflicts between would-be rivals must have other sources. As claimed by Gartzke and Simon (Reference Gartzke and Simon1999), those same factors likely explain their persistence. They also argue that given the relatively small number of rivalries, the higher level of conflict among rivals can just as easily be explained stochastically as a statistical aberration.

Rivalry theorists have met this theoretical challenge empirically. Colaresi and Thompson (Reference Colaresi and Thompson2002b) show that while crises tend to beget a higher likelihood of future conflicts among the same pair of states, self-identified rivals have a much higher likelihood still of another crisis. Rivalry increases the odds of war by a factor of almost three, even while controlling for other factors. Crises between rivals, as opposed to non-rivals, tend to escalate more quickly, to higher stages in the escalation ladder, and often over non-violent, even symbolic cues. While power disparity between non-rivals decreases the likelihood of conflict, there is no such effect on rivalry dyads. The psychological baggage trumps rational calculation.

Nevertheless, rationalists do raise an important theoretical question, even if they do not have a satisfactory empirical answer. To the extent that rivalries are a function of particularized fear, which is a relational factor that transcends the attributes and dispositions of those involved, something exogenous to the relationship must explain changes in the direction of this process. Rivalries have been found to have an arc (Colaresi, Reference Colaresi2005). They escalate and de-escalate. Types and social orientation might play an important role.

Different types are very present in the work of those who have tried to specify a micro-logic to the creation of rivalries, even if they are underconceptualized. Absent any preexisting image of the other as enemy, it is likely that the instigators of rivalries and subsequently their greatest defenders are competitors. Theorists of rivalries generally claim that there is a distribution of hawks and doves (in social orientation parlance, competitors and cooperators). Hawks are more inclined to initiate conflict in general, and the conflicts that their policies produce can produce shifts in the overall level of particularized fear vis-à-vis their adversary in their societies, which adds to their stature and puts them in position to sustain and escalate the rivalry (Vasquez, Reference Vasquez1993; Senese and Vasquez, Reference Senese and Vasquez2008). Conversely, de-escalation might begin with gestures of conciliation more likely to come from dispositionally more trusting individuals and groups. Particularized and generalized trusts are likely linked empirically.

Social psychology also helps explain interactive aspects of the democratic peace. The democratic peace is a dyadic phenomenon in which democracies are thought to behave peacefully towards other democracies, but not towards autocratic regimes. The incidence of large-scale war and even crisis between democracies has been found to be exceedingly rare, whereas democracies are no less likely to become engaged in warfare with autocratic regimes than the latter are amongst one another. The peaceful nature of democratic dyad relations does not require significant elaboration. I discussed the monadic effect of the democratic peace above. Trusting gestures are likely to be matched by regimes marked by a similar cooperative social orientation. Hence, democracies rarely fight wars with one another (Maoz and Russett, Reference Maoz and Russett1993).

More puzzling is why democracies, if they in fact embrace non-violent methods of conflict resolution, do not exhibit a generically more peaceful foreign and defense policy as a whole toward all other states. A social-psychological focus on trust helps explain this puzzle. Kuhlman and Wimberley (Reference Kuhlman and Wimberley1976) and Kuhlman and Marshello (Reference Kuhlman and Marshello1975) find that while cooperators tend to begin with cooperation in social dilemmas, they do not continue to cooperate against a 100% competitive strategy, opting to defect against defection (although individualists resort to defection much more quickly). Cooperative social orientations are based on an expection of reciprocity, not altruism. Generalized trust is a default tendency that adjusts in light of the circumstances. In the face of consistent information revealing a breach of trust, cooperation declines sharply on the part of these generalized trusters.

The combination of ‘behavioral assimilation’ by cooperators and competitors’ consistent resistance to cooperation leads to what Kelley and Stahleski (Reference Kelley and Stahleski1970) call the ‘triangle effect’, in which the behavior of individuals with higher predispositions to cooperate based on a greater disposition to trust eventually converges on non-cooperation over time in the face of defection. It might be that even if liberal democracies are predominantly composed of individuals with a cooperative social orientation, when they find reciprocity unforthcoming, they are forced to behave like competitors. Therefore they are no less likely to fight non-democracies than non-democracies are to fight each other, even if they are less inclined to initiate conflict (see above).

Liberal democracy is based both on shared beliefs and a set of institutions. It rests on both a benign and trusting view of human nature and a set of procedures to select political leaders. The advantage of a social-psychological approach is that it can explain both the short-term variation in foreign policy in democracies that is the natural result of electing different types into positions of leadership, as well as the long-term tendency of democracies to behave differently in foreign affairs as a result of the culture of generalized trust that marks most but not all of its citizens. Owen (Reference Owen1997) notes that even liberal democracies have illiberal leaders from time to time, but that the general inclination of democracies in foreign affairs is nevertheless distinct. The dynamics of illiberal democracies, those with democratic institutions but without a trusting democratic culture, would be, indeed have been found to be, different (Mansfield and Snyder, Reference Mansfield and Snyder2002).

Systemic effects of social orientation: selection mechanisms allowing cooperation

A social-psychological approach to international relations based on social orientation is on toughest terrain at the systemic level, where the levels of analysis, the external validity and the ecological fallacy critiques are most applicable. Structural realists might assert that the nature of international relations, the lack of any institutional enforcement of cooperation, means that the applicability of social-psychological findings to foreign policy and international relations is highly questionable. Structure is much more powerful when the stakes are high and protections are low. Cooperating in an experiment with a couple of dollars is one thing, playing the high stakes game of international politics, where one’s very survival is in question, is another. Vulnerability is simply too high to trust and will overwhelm any variation in dispositions to trust. In this view, the material nature of international politics acts as a selection mechanism for only competitive orientations. In this vein, Tang (Reference Tang2008) makes the case for the evolutionary superiority of a fearful human nature. This view is not necessarily antithetical to an emphasis on types. Individual states and statesmen are free to trust. They will simply be weeded out of the system. However, it would lead to doubts that there are significant numbers of individualists, much less cooperators.

Neorealists miss that moralistic trust, or even strategic trust, can yield greater payoffs for cooperators and individualists that makes them stronger and better able to take the risk of trusting again in the future. Obviously the risks of cooperation in international politics are higher relative to other arenas, but so might be the rewards. Moralistic trust, both particularized and generalized, serves as a form of ‘social capital’, allowing for higher levels of cooperation without the need for costly enforcement and the collection of information (Putnam, 1993; Sztompka, Reference Sztompka1999: 62; Tyler, Reference Tyler2001: 285; Cook and Cooper, Reference Cook and Cooper2003: 209). It allows states to seek potentially greater gains with unfamiliar cooperative partners. It frees its holders from isolation or the familiar but restrictive confines of those with whom one regularly interacts. As stated above, moralistic trust ‘emancipates’ individuals (Yamagishi, Reference Yamagishi2001). It might facilitate cooperation even in the most dangerous, anarchical environments. In prisoner’s dilemma experiments in which subjects can enter and exit the game, those willing to trust generally fare better than more fearful individuals do (Orbell and Dawes, Reference Orbell and Dawes1993). Cooperators do better than competitors.

The observation that cooperators might do better than cooperators also suggests a mechanism by which the distribution of cooperators and competitors might change over time, changing with it the overall character of the system in which they operate. A social-psychological approach to international relations predicated on the importance of types would explain systemic features of the environment in this way, as a consequence of the aggregation of the traits of the units that make it up. Systemic qualities are ‘emergent macro-phenomena’, to use Cederman’s term, created by a multitude of micro-level interactions. Ahn and Esarey (Reference Ahn and Esarey2008) suggest such a game theoretic model consistent with this notion of what biologists call ‘reciprocal altruism’ (Trivers, Reference Trivers1971). As mentioned above, individualists or cooperators, if they can successfully identify one another, might fare better based on their cooperation than competitors. If this cooperation yields material benefits, more pro-social individuals (or states) are in a better position to defend themselves. They can carve out a space for themselves, replicate, and thrive. Their relative distribution in the overall population might even grow over time. We can imagine this process in biological terms, if dispositions to trust are passed genetically, or in social terms, as a cultural phenomenon. There is evidence for both (Uslaner, Reference Uslaner2002; Oxley et al., 2008). This process of selection could provide the link between the individual level of analysis and systemic features of the environment and helps address and redress the ecological fallacy critique leveled at social psychology.

Lars-Erik Cederman (Reference Cederman2001b) provides a very similar, if not identical, selection argument to that of Ahn and Esarey (Reference Ahn and Esarey2008) to explain the democratic peace. Cederman frames his argument similarly, asking how more benign democratic polities manage to attempt cooperation if it exposes them to potential extinction. Implied is the danger of trust. However, for the reasons identified above, non-democratic, competitive orientations are not necessarily dominant. Cooperators can thrive, even amongst competitors, if they are able to identify other cooperators. While Ahn and Esarey write of ‘credentialing mechanisms’, Cederman calls them ‘strategic tags’. The most obvious is democracy itself. Democracy might serve to provide evidence of trustworthiness through a strategic trust mechanism, its very transparency, or through a moralistic trust mechanism, democracies’ culture of non-violent conflict resolution (Maoz and Russett, Reference Maoz and Russett1993; Owen, Reference Owen1997; Schultz, 1999; Ikenberry, 2001). As a result, democracies thrive on the basis of mutual exchange and their more peaceful relations. If the benefits gained by cooperation among democracies leads to the proliferation of democracy itself, as Kant predicted it could, the very nature of international politics changes. Cederman (Reference Cederman2001a) offers some empirical evidence that the more trusting democracies, over time, created such a transformation.

This description evokes Wendt’s stylized account of the origins of a different culture of anarchy. As more and more members of a collectivity come to believe others are enemies, a tipping point is reached at which these representations take over (1999: 264). An individual level phenomenon becomes a social structural one. This is the movement from agency to structure in ‘structuration’.

Wendt’s understanding of this process is cultural. But the beliefs of individuals come first and they likely vary. The process is driven at the beginning by psychological processes. Fear comes before identity. The motor is at the individual level. The social-psychological hunch would be that there would always be different types, even if the overall distribution varies over time and place. A social orientation approach to international relations would be skeptical that any particular type predominates so much so that there is a systemic ‘culture of anarchy’. Indeed the existence of different types is necessary to begin any change in a culture of anarchy. Anarchy can only be what states make of it if there are some types who think a little differently. The interaction process foreseen by Wendt’s constructivism leading to the creation of moralistic trust is most likely confined to particular relationships, and not broad enough to create an overall nature of the system. It is mostly a dyadic approach. For a culture of trust or fear to form, states (or individuals) would have to draw general conclusions from their more local experiences and these are unlikely to be homogenous.

Regardless of which mechanism is correct, demonstrating the systemic manifestations of micro-level processes is extremely difficult, something which might be thought to give a certain edge to structural realism. However, a social-psychological approach based on type alerts us to the possibility that even neorealism’s systemic qualifications are dubious. Structural realism and realism in general expect international relations to be largely conflictual and premised on self-help, lacking in trust and cooperation except in situations of mutual threat. Yet rivalry theory notes that conflict in the international system is relatively rare empirically and confined to a small number of dyads. As discussed, hostility amongst them is driven less by their objective structural and material circumstances than the accumulated psychological baggage of fear. Rivalry theory would tell us that realism oversamples on conflict, drawing general conclusions about systemic qualities of the international environment that are in fact psychological and confined in space and time to dyads. Rivals ‘create and sustain a relationship of atypical hostility’, writes Thompson (Reference Thompson2001: 558) (emphasis added). In other words, neorealism is reductionist.