IDENTITY, INTEREST AND EMERGENT RATIONALITY:

AN EVOLUTIONARY SYNTHESIS*
Michael W. Macy ABSTRACT Identity and interest paradigms provide rival explanatory frameworks: One portrays group solidarity as the affirmation of cognitive categories, while the other looks for the incentives that motivate collective action among rational egoists. Neither is fully adequate. Identity theory emphasizes the cohesive effects of similarity but overlooks the causal importance of self-interest in a collective good. Interest theory emphasizes the solidary effects of interdependence but struggles to explain enthusiastic self-sacrifice and moral righteousness. I propose a synthetic reformulation of identity and interest, based on two corresponding strategies in the evolution of cooperation: kin and reciprocal altruism. This evolutionary approach, however, entails a fundamental rethinking of basic concepts in rational choice theory a shift from purposive to emergent rationality. KEY WORDS: altruism, evolution, identity, learning, rationality

Introduction During the American Revolution, General Washington pondered the motives that might lead a struggling band of poorly equipped farmers to take up arms against the most powerful empire in human history:
I do not mean to exclude altogether the Idea of Patriotism. I know it exists, and I know it has done much in the present Contest. But I will venture to assert, that a great and lasting War can never be supported on this principle alone. It must be aided by a prospect of Interest or some reward.1

Since the time of the Sophists, social and political theorists have debated whether public spirit or self-interest is the more secure foundation for

428

Michael W. Macy

a viable community. During the great philosophical and religious debates of the seventeenth and eighteenth centuries, conservatives saw community as a divine representation of the Universal, while liberals countered that enlightened self-interest was more reliable and less ill-tempered than were appeals to the passions.2 Pragmatic eclectics, like Washington, believed that the strongest mortar contains some mix of the two ingredients. Among contemporary social scientists, the debate has been stripped of its philosophical wrappings and reduced to two contrasting approaches to the social psychology of group solidarity, one based on shared identity, the other on mutual self-interest. The identity paradigm is all about the cohesive effects of similarity, while the interest paradigm sees interdependence as the motive for cooperation and exchange. This paper will briefly recapitulate the arguments, then point out complementary limitations in each approach that suggest the need to transcend the old identity-interest divide. Finally, I will propose an evolutionary synthesis centered on a theory of emergent rationality.

Shared Identity and Expressive Collective Action Oberschall and Kim define identity as the answer to
the basic existential questions, Who am I? and With whom do I belong? The categorization of a we and a they is a fundamental social psychological manifestation of human sociality. We and they are experienced and socially validated with we-feeling, approval seeking, attachment and conformity to group attitudes and norms, and a corresponding tendency to distancing, negative affect, and stereotyping of other groups and social categories, the they (1996: 64).

These ideas were first enunciated by Heider (1958) and Sherif et al. (1961), refined in Tajfals (1982) social identity theory, and elaborated further by Hogg and Turner (1987). The basic behavioral assumption is that individuals order the social world by carving out cognitive categories through interaction with others, an idea that is grounded in the pragmatist epistemology of James (1981), Peirce (1955), and Mead (1934).3 In order to navigate these interactions, we seek coherence, predictability, and the coordination of expectations. Without the ability to predict how others will respond to a situation, and to know what responses others expect, social life would be unbearably capricious, random, and incoherent. To avoid this, individuals anchor themselves socially by exaggerating

IDENTITY, INTEREST AND EMERGENT RATIONALITY

429

similarities among those they categorize together with themselves, and by exaggerating differences with those in groups that provide a clear contrast. Heiders (1958) balance theory, for example, posits a need for cognitive consistency in our orientations toward others. If Adam likes Bonnie and Bonnie likes Charlie, then upon meeting Charlie, Adam will be predisposed to accept Charlie into his circle. Homophily (Homans, 1951; McPherson and Smith-Lovin, 1987) also contributes to structural balance. If Adam and Charlie share many salient attributes (religion, class, friendship with Bonnie, etc.), then they will be more likely to interact with one another and to develop positive interpersonal attraction. These tendencies toward categorization, stereotyping, in-group favoritism, and out-group prejudice can enhance the capacity of interactants to act collectively in pursuit of common interests. However, most identity theorists follow Heider in explicitly rejecting an instrumental basis for group cohesion (interpersonal attraction) and solidarity (compliance with group norms). For identity theorists, interests are only the surface of things. What is beneath the surface is a strong emotion, a feeling of a group of people that they are alike and belong together (Collins, 1992:28). This shared identity transforms interests into the objects of expressive behavior. The interests become moral rights and become surrounded with a kind of symbolic halo of righteousness (1992:29). Others see both interests and identity playing a part, but not in the same groups. Instrumental models of collective action may be applicable to interest groups but cannot explain the expressive content of enthusiastic social movement mobilizations. Turner (1996), for example, acknowledges the contribution of self-interest as...a strong undercurrent in all human motivation, but warns against the reduction of expressive collective action to the instrumental pursuit of shared interests:
Mere interest groups may devote tremendous energy and resources and follow sophisticated rational strategies in promoting their collective interest, but without the sense of mission characteristic of social movements. Cognitively, what is required is a conviction that others in society, or even the whole society, will benefit from the reforms being promoted, that those who stand to gain are more noble than self-interested, and that those who lose deserve to lose because of their evil motives and schemes. The latter conviction is enshrined in the ubiquitous identification of villains in the ideologies of social movements (1996:8).

The polarization of the noble self and the evil other is the signature of identity-based symbolic collective action. Dependence on others for

430

Michael W. Macy

affirmation of self means that the social interactions that confer a sense of reality and identity (Berger, 1966:21) are highly charged with emotion. These affirmations are a shield against terror in the face of the innate precariousness of all social worlds (1966: 23). The emotional energy invested in identity is most evident when members of the group are confronted with deviance. Transgression and trespass of social boundaries can trigger virulent expressions of indignation that illuminate the borders of the group. This is the basic idea in Durkheims contention that crime is normal. For Durkheim, crime is not simply the disruption even of serious interests, it is an expiation for an act that offends strong and defined states of the collective conscience (1960:80). It is this indignation that makes an act criminal. We must not say that an action shocks the common conscience because it is criminal, but rather, that it is criminal because it shocks the common conscience (1960:81). This offensiveness, or repugnance, induces a passionate reaction that is instinctive and unreflected (1960:85). Hence, for Durkheim, public punishment is not a deterrent but an expiation, an emotionally charged collective affirmation of identification with and commitment to the group. Following Durkheim, identity theorists stress internalized compliance with collective obligations and expressive demarcation of group membership. Social control, from this viewpoint, does not promote order by creating incentives that motivate instrumental compliance. Rather, it mobilizes the emotional energy that solidifies attachment to the group. The effect is not to suppress deviance but to reveal it, and thereby to patrol, clarify, and affirm the normative boundaries that define the group. Eriksons wayward Puritans (1966) and Gusfields Temperance activists (1963) are two classic case studies of the mobilization of solidarity in an enthusiastic social movement. Erikson showed how a theory of repression as a deterrent to deviance misspecifies the causal relationship between punishment and crime. Despite the terrible penalties inflicted on hapless victims, the campaigns against witches and Quakers caused the rate of deviance to rapidly escalate. These crime waves were the consequence of repressive campaigns, not the cause. According to Gusfield (1963), the Temperance struggle was not really about the evils of alcohol, but rather, the evils of ethnic immigration that threatened to redefine what it means to be an American. Following Gusfield, Luker (1984) has argued that the contemporary abortion controversy is not about the point at which life begins, it is a referendum on motherhood. Participation in these zealous movements can be explained as an expressive response to challenges to the social status of

IDENTITY, INTEREST AND EMERGENT RATIONALITY

431

middle-class Protestants in the Gusfield case and traditional homemakers in Lukers. This strong aversive response parallels the curiously violent indignation of Puritans to the refusal of Quakers to doff their caps to the authorities. To sum up the identity paradigm, identities determine who we care about, and who cares about us. They emerge as actors sort themselves into categories that affirm members expectations of others, and through that, their sense of a shared reality. Identities are not acquired passively but are actively negotiated through the demarcation of group boundaries. Social and moral boundaries are defined by identifying and punishing deviants. Compliance with conventions is motivated not by instrumental calculation of the risk of punishment but by an emotional investment in the identity these conventions delineate. Violent punishment is not calibrated to deter deviance; rather, it is unleashed as an expression of indignation at the violation of normative boundaries. Social movements are mobilized as expressive responses to real or imagined incursions across the borders of the group. The collective action in turn affirms the shared identities that animate participation.

Mutual Interests and Purposive Collective Action In striking contrast to the identity approach, the point of departure for the interest paradigm is not similarity but complementarity. Groups are held together not by shared definitions but by shared interests in cooperation among interdependent individuals (Kelley and Thibaut, 1978). Interdependence means that the optimal outcome for a given choice is contingent on the choices of other decision-makers, and vice versa. Game theorists have formalized the problem using mixed-motive games, most notably, the Prisoners Dilemma (Rapoport and Chammah, 1966). In these games, the optimal choice for each individual is defection (or freeriding). However, if everyone defects, all suffer the collectively worst possible outcome. This tension between individual and collective interests is the defining problem in collective action theory (Macy 1993). Collective action obtains not through subordination of individual wills to a shared consciousness, but through overlapping interests that motivate purposive actors. The emphasis is not on emotional attachment to the group but on instrumental compliance with collective obligations and social control centered on deterrence of free-riding. The social exchange theories of Homans (1961), Emerson (1962), and Blau (1964) explain cohesion and cooperation in small groups

432

Michael W. Macy

characterized by face-to-face interaction. Exchange theories posit selfinterested actors who cooperate because the gains outweigh the costs. The rewards provided by the group, especially social approval, reinforce members commitment and generate affective ties (liking). Emotional attachments are thus the consequence of group solidarity and cohesion, not the cause. Contemporary elaborations focus on network structure, status, power, and justice in social exchange. These include Cook and Yamagishis (1983) equidependence theory, Markovskys (1992) network exchange theory (Markovsky, et al., 1993), Bienenstock and Bonacichs (1992) core theory, Willers elementary theory (Willer and Anderson, 1981), Friedkins (1992) expected value theory, Bergers expectation states theory (Berger et al., 1974), Gray and Tallmans (1984) satisfaction balance theory, and Jassos (1980) theory of distributive justice. Lawler and Yoon (1996) revisited the problem of cohesiveness that concerned early exchange theorists, but with a new twist. They incorporated emotional attachment as both cause and consequence of repeated exchange. Nee and Ingram (forthcoming) extend the exchange-theoretic approach from networks to institutions. The core idea is that institutions are permeated with informal sanctioning mechanisms, embedded in local networks and interpersonal relationships. They also see Ellickson (1991) and Ostrom (1990) as important contributors to theories of informal social control based on face-to-face interaction in small groups, and argue that these ideas may also be usefully applied to institutional norms. Ellickson showed how informal arrangements, beyond the shadow of the law, are not only less costly but also more flexible in allowing small transgressions, so long as the books are balanced in the long run. Ostrom pointed to ongoing interactions within a community as an alternative to reliance on central enforcement of environmental regulations. Nee and Ingram see these informal controls as the missing link between networks and institutions as structures that constrain opportunism in social exchange. The interest paradigm also informs studies of large-scale collective action. Hechters (1987) theory of solidarity points to the capacity of the group to monitor and sanction individual compliance with the obligations of membership. Formal sanctions, including expulsion, deter deviant behavior that compromises the collective good. This neo-Hobbesian solution to the problem of social order poses what Oliver (1980) has called a second-order free rider problem. How can

IDENTITY, INTEREST AND EMERGENT RATIONALITY

433

formal social control explain individual contributions to collective action if the provision of inducements is itself a public good? Members of an interest group know they will benefit from enforcement of member obligations even if they let others bear the cost of the Leviathan. From this angle, neo-Hobbesian theories would appear to have it backwards: Collective action does not depend on the institutional capacity for social control; rather, it is the other way around. Heckathorn (1989) has advanced a novel rational-choice solution to the second-order free-rider problem. He argues that enforcement can be much more cost-effective than compliance. Hence, rational deviants may be willing to enforce the very norms they violate. He points to the example of corrupt sheriffs in the Old West who nevertheless increased the level of order (1989:97). Other interest theories emphasize mobilization rather than sanctioning. Fireman and Gamson (1979), as well as Marwell and Oliver (1993), downplay the free-rider problem and focus instead on the problem of efficacy, the ability to make a difference. More formally, the collective action problem is a game of Stag Hunt, not Prisoners Dilemma. My best payoff may be to cooperate, but only if you (or enough others) choose to cooperate as well. Mobilization of participation in collective action thus requires a critical mass of vanguard contributors who can trigger a chain-reaction as the prospects for success begin to snowball. Chongs (1991) case study of the civil rights movement is an important empirical application of the Stag Hunt (or Assurance) game.

Fawning Zealots and Rotten Parents The interest and identity paradigms have complementary strengths and weaknesses. Identity theorists show how solidarity can emerge from interpersonal attraction and cohesion, regardless of instrumental interest in the production or exchange of valued resources. This emotional attachment provides a handle on the effects of righteousness and moral outrage in many groups, but it misses what should be the starting point of an explanation, the power of collective action to satisfy individual needs. The assumption that behavior is largely expressive, and not directed toward the efficient attainment of a goal, precludes the attribution of purpose as an explanation for action. Without an account of the source of the compulsion, expressive behavior appears capricious and arbitrary. Hence, the literature on enthusiastic social movements tends toward case-

434

Michael W. Macy

specific (and often ad hoc) interpretation and description over systematic theoretical integration. In contrast, a key strength of the interest approach is the theoretical rigor and abstraction of rational choice theory, as noted by Oberschall and Kim in proposing a purposive model of identity. We are using rational choice for no other purpose than making testable deductions from assumptions and propositions that are explicitly stated, and which are more difficult to make systematically and unambiguously without formal theory (1996:66). Oberschall and Kim propose an identity-action model of ethnic conflict that reduces identity to purposive choice. According to the authors, actors have multiple and competing sources of identity, such as ethnicity or community. This generates social pressures to conform to contradictory expectations (1996:70). Actors selectively conform so as to maximize approval, and their choices in turn produce positive and negative externalities that strengthen or weaken the pressures on each individual. In short, the relative salience of competing identities shapes the interests that guide action, and collective action, in turn, defines the salient boundaries for the enforcement of identity. Identity is both subject and object of the interests of an actor. The impetus for this approach comes from Coleman. Identification, he argued, is a process in which one actor has adopted, or taken up, the others interest (1990:158). He offered the examples of parental bonding, patriotism, and communal euphoria which he attributed to the instrumental pursuit of a shared interest in emotional gratification (1990:161). The search for incentives that motivate expressive behavior confronts a Zen-like antilogy. Euphoric gratification is likely to elude those who pursue it deliberately. The feeling is more readily attained as the unintended byproduct of participation in righteous or enthusiastic collective action. Or consider the paradox of the fawning zealot. Groups unified by a shared political or spiritual mission may withhold approval from those whose sacrifices appear to be motivated by an obsequious desire to gain personal status and approbation, rather than the advancement of a higher cause. Hence, aspiring martyrs must be careful not to appear overly eager. Similar problems arise on the enforcement side. Expiating punishment is often meted out even when its effects on compliance are counterproductive, as when an angry and abusive parent reacts inappropriately to a temper tantrum by a rotten kid, to borrow a character from Beckers (1991) theorem. The rotten parent wants to end the childs intemperate violation of familial norms but cannot suppress the

IDENTITY, INTEREST AND EMERGENT RATIONALITY

435

urge to vent excessive frustration, even though this only aggravates the source of the parents displeasure. When confronted, the parent might justify the behavior as the need to teach the child a lesson. This squares with a deterrence theory of social control, but only as an account of rationalization, not rational action. In sum, it is obviously prudent to comply with the rules of the group, and compliance is often instrumentally motivated (and thus notoriously unreliable). Nevertheless, there remains a large residual that defies reduction to purposive self-interest. This includes unthinking, compulsive, and fervid compliance with internalized, self-enforced norms, as well as angry and vengeful enforcement. In addition, there is the problem of participation in enthusiastic groups where social approval and emotional gratification are attained more efficiently when not directly pursued. In these groups, it becomes difficult to know just what advice to give a perfectly rational actor, except perhaps, to stop acting rationally. Such are the logical traps that await the effort to discover ulterior motives for the affirmation of identity. This, then, is our dilemma: On the one hand, rational choice theory obtains its analytical power from the use of expected outcomes (or utilities) to explain the actions that produce them. On the other hand, this consequentialist mode of explanation is difficult to apply to expressive behavior and enthusiastic collective action without badly distorting the motivational state of the actors.

Evolution and Emergent Rationality The way out of the dilemma is the explanatory power of the unintended consequences of action. Consequentialist explanations pose a teleological quandary: how can the future influence the past? In rational choice theory, the link to the future is intention: Action is explained as the effort to maximize the expected return on competing investments. By rational choice, I refer to theories that posit the analytical ability of purposive actors to accurately predict the outcomes of alternative choices based on knowledge of causal processes. In these models, unintended outcomes may constrain future choices, and even influence future preferences,4 but they have no explanatory power as incentives for action. Evolutionary and learning-theoretic models, in contrast, posit iteration, not intention, as the link to the future (Macy, 1993, 1996b). Outcomes attract action via gradient search, without the need for a map of the evolutionary landscape. The outcomes that matter are those that have

436

Michael W. Macy

already occurred, not those that an analytical actor might expect to obtain in the future. The consequences of recurrent decisions can select for efficient action by adaptive actors, whether or not anticipated benefits motivate the associated choices. By extending the interest paradigm to include both intended and unintended rationality, the evolutionary approach promises to provide an integrated account of both identity and interest as motives of collective action. Iterative search also relaxes the highly restrictive cognitive assumptions in rational choice theory. Behaviorists charge that the calculus of expected utility is at best a prescriptive model of how choices ought to be made, one that bears little resemblance to actual decision-making (Simon, 1992). Indeed, many economists now concede that utility-maximizing behavior describes the end result of market exchange but not the process by which it obtains (Winter, 1986). Results from two decades of laboratory research led Orbell and Dawes (1991) to reject subjective expected utility as a descriptive model of human behavior, preferring instead a set of heuristics, or rules of thumb, that can be explained (and perhaps justified) insofar as they economize on cognitive effort (1991:517). Economists, including Hayek (1967), Simon (1992), and Vanberg (1994), have also concluded that economic behavior is based on rules, not choices. Routine decisions, especially those by institutional actors, are more often rule-governed than analytical. Rules are embedded in heuristic conventions and habits, symbolic and ceremonial rituals, moral codes, social customs and protocols, and institutional norms and routines. In game-theoretic evolutionary models, rules encode strategies, like the widely acclaimed Tit for Tat popularized by Axelrod (1984). These strategies are formalized as inputoutput functions, where the input is a set of conditions of varying complexity and the output is an action. Although behavioral economists contend that rule-based behavior is mandated by cognitive limitations and the costs of information, social psychologists point to a much broader and more basic requirement: the possibility of social life (Berger, 1966). Among social species, we are unique in our plasticity. Unlike social insects, we have very little genetic programming for collective behavior. Were we not creatures of habit, routine, and heuristic devices, effective coordination would be impossible a cacophony of inappropriate responses to unexpected reactions from others. The argument thus builds on and extends the cognitive categories model in social identity theory. Rules make behavior predictable, or at least, enough so that interdependent individuals can influence one another in response to the influence they receive, thereby

IDENTITY, INTEREST AND EMERGENT RATIONALITY

437

carving out locally stable patterns of interaction. In short, rules are not simply analytic shortcuts that lower the cognitive costs of decisionmaking. They are the language of social life. Of course, our species is also unique in its analytical capabilities; hence, an obvious objection to the behavioral rational action model is the portrayal of intelligent thinkers as robotic simpletons. Clearly, rule-based behavior would be hopelessly inefficient were it not for the fact that stimulus-response patterns can improve (and become more sophisticated) over time. The key behavioral assumption is that the consequences of an action, which may or may not have been consciously anticipated, modify the probability that the action will be repeated the next time the same input conditions are met. Repeated exposure to a problem recycles the lessons of the past, allowing the consequences of alternative courses of action to be iteratively explored. Human intelligence is not the issue. Our cognitive capacity for language does not obviate the evolution of words and signs, or our need to learn their meaning through practice. The same pragmatist epistemology also applies to problem-solving behavior and implies that rationality is grounded in action, not calculation. Rationality does not spring from the heads of the actors; rather, it emerges from their hands and feet from problem-solving, trialand-error, and other processes of gradient search. This emergent rationality obtains through two types of experiential feedback, reproduction, which alters the frequency distribution of individuals with various rules in a given population, and reinforcement, which alters the probability distribution of rules competing for attention in a given individual. In biological reproduction, rules are genetically encoded and transmitted. In cultural evolution, reproduction occurs through role modeling, occupational training, social influence, imitation, and vicarious reinforcement. Most expected-utility models assume instantaneous calculation and global optimization. In contrast, reproduction and reinforcement are dynamic processes that generate emergent solutions through adaptive responses that improve over time. Because these improvements are piecemeal, emergent rationality is meliorizing, not optimizing. Melioration is highly path dependent and not very good at backing out of cul de sacs on the evolutionary landscape. Both reinforcement and reproduction are biased toward better strategies, but they carry no guarantee of finding the best of all possible solutions. Hill-climbing thus violates the optimality assumption in rational choice theory -- the actors cannot be assumed to attain the highest peak, however relentless the search. Even so, the highest achievements of natural selection far surpass

438

Michael W. Macy

anything created by the intellect of an engineer, causing skeptics to suspect the unseen intervention of some higher intelligence. Consider, for example, an economic model of emergent rationality. A firms problem-solving strategies improve over time through exposure to recurrent choices, under the relentless selection pressure of market competition. Sub-optimal routines are removed from the repertoires of actors by learning and imitation, and any residuals are removed from the population by bankruptcy and takeover. The outcomes may not be optimal, but we are often left with well-crafted routines that make their bearers look much smarter than they really are (or need to be), like a veteran outfielder who catches a fly ball as if she had calculated its trajectory. We thus arrive at the possibility that rationality can emerge from behind the backs of the actors and can target both the intended and unintended consequences of action. If the link to the future is repetition, not calculation, then actions can be attracted by unintended outcomes that alter the probability that the associated behavior will be repeated. Unintended consequences may account for the persistence of expressive and righteous behavior that lacks any instrumental (or ulterior) motive. For example, Franks (1988) evolutionary model of trust and commitment formalizes the emergent rationality of emotions like vengeance and sympathy. An angry or frightened actor may not be capable of deliberate pursuit of self-interest, yet the response to the stimulus has consequences for the individual, and these in turn can modify the probability that the behavior will be repeated, through reinforcement, imitation, or some combination of learning and reproduction.

Kin and Reciprocal Altruism Still, the rationality of emotion is not the most radical departure from the conventional portrayal of self-interested optimizing. Not only may actors be unaware that their actions are self-serving, it is not even necessary, in an evolutionary model, to assume any benefit whatsoever to the individual. This is admittedly a tentative argument, one that requires more rigorous specification and empirical testing. However, if valid, it means that emergent rationality is applicable not only to expressive behavior but also to behavior that is genuinely altruistic. Altruism entails more than a prudent detour in the pursuit of self-interest (e.g., a well-publicized charitable donation). By definition, altruistic behavior requires sacrifice not just an investment that pays back with a compensating benefit, but a

IDENTITY, INTEREST AND EMERGENT RATIONALITY

439

net loss, an outcome that is inferior to an available alternative. How can this be rational? With purposive rationality, the outcomes that motivate action correspond, necessarily, to the interests of the actor, thereby precluding genuine altruism. In contrast, the outcomes that attract emergent rationality need not always correspond to the interests of the individual. An earlier generation of evolutionists posited the interests of the group as the target of selection pressures that favored altruistic behavior (WynneEdwards, 1962). This group-selectionist theory has now been largely discredited. Many ethologists and theoretical biologists now believe that the proper specification of the unit of selection is neither the individual organism nor the group. The evolutionary game theorist Maynard Smith (1976) has proposed a provocative alternative: the underlying strategy (whether a genetic or heuristic rule). This rule-centered approach can be applied to an explanation of kin altruism in nature, the strategy to sacrifice for the benefit of close genetic relatives (Ruse and Wilson, 1986; Alexander, 1987; Hamilton, 1964). From a rules-eye view of the problem, the altruist serves the interests not of the beneficiary but of the selfish gene that controls its behavior (Dawkins, 1976). A gene for kin altruism can improve its viability by directing a transfer of vital resources from its agent to another organism that carries the same gene. Allison (1992) has used formal models to extend the kin altruism model to benevolence based on cultural relatedness, such as geographical proximity or a shared cultural marker. The social equivalent of Dawkins selfish gene is a cultural rule, encoded in an aphorism, norm, heuristic, routine, protocol, ritual, or strategy. Dawkins (1976) call these memes, which he defines as ideas that can jump from one brain to another via imitation. However, ideas are poor candidates for evolution, due to blending and mutation. In contrast, rules are patterned behavioral responses to stimuli, carved out through repetition, stored in neural pathways, and continually checked against error in local interaction. For convenience, we can define the self-interest of a rule as its evolutionary stake in the outcomes of the actions it instructs, insofar as these outcomes influence the odds that the rule will flourish in the face of competitive selection pressures.5 However, as illustrated with kin altruism, a self-interested rule need not imply a program that instructs egoistic behavior by its agent. Suppose a rule could propagate faster by ordering its carrier to export life-chances, as might happen when the beneficiary carries the same rule. Then the rule could be said to have a self-interest in the altruistic behavior of its agent.

440

Michael W. Macy

The logic of kin altruism might explain why cultural evolution seems to favor a tendency to associate with those who are similar, to differentiate from others, and to defend the identified group against social trespass with the emotional ferocity of parents defending their offspring. Solidary expressions of identity can thus be reduced to rational self-interest, but it is the evolutionary interest of the rule, not the interest of its agent, that is the effective cause. We thus arrive at a key difference between purposive and emergent rationality. Purposive rationality applies to the individual as the unit of analysis, and therefore entails the axiom of self-interested behavior. Emergent rationality applies to rules (or strategies) as the units of analysis, as well as to the individuals who act on these rules. This means that the behavior the rules instruct need not necessarily advance the interests of the individual agent. Self-interested behavior can then be modeled as empirically variable, not axiomatic. It must be stressed, however, that the self-interested individual is clearly not precluded by emergent rationality, quite the contrary. Kin altruism is only one of two highly robust helping strategies that have been identified by theoretical biologists and evolutionary game theorists, the other being reciprocal altruism. The latter refers to the trading of favors without an enforceable quid pro quo. Reciprocal altruism is therefore fully compatible with individual self-interest and not a form of genuine altruism. Rather, it is a straightforward illustration of prudent behavior, Frank concludes, enlightened prudence, to be sure, but self-interested behavior all the same (1988:34-5). An enlightened egoist may transfer resources to another under the expectation that this behavior is likely to trigger sufficient compensation. Thus, Beckers (1991) theory of parental benevolence (including the rotten kid theorem referenced above) is actually an example of reciprocal, not kin altruism the exchange of support between care-givers and their dependents at different ends of the life-cycle. Although reciprocal altruism can be modeled using the individual as the unit of analysis, it may nevertheless be instructive to consider a ruleselectionist formulation. Again taking a rules-eye view of the problem, the manifestations of self-interested rules can be altruistic, so long as the benefits ultimately convey relief from the relentless pressure of selection. When these benefits to the rule obtain immediately from the behavior, we classify the behavior as self-interested. When the programmatic benefits are separated from the behavior in time and space, we classify the behavior as altruistic. With reciprocal altruism, the benefit is temporally removed in that the rule has learned how to export life-chances as a way

IDENTITY, INTEREST AND EMERGENT RATIONALITY

441

to trigger a later return on the investment.6 With kin altruism, the benefit is spatially removed, in that the rule has learned how to export lifechances from one of its agents to another, such that the inclusive fitness (Hamilton, 1964) of all its carriers is greater than before. Given that space and time are alternative measures of distance, it seems reasonable to classify both reciprocity and kin-benevolence as altruistic behaviors (or phenotypes) that provide indirect benefits to the self-interested rules (or genotypes) in which these behaviors are encoded. The only difference is whether the benefits must leap over time or space to make their way back to the rules that generated them. It is noteworthy how closely these two cooperative mechanisms parallel the sociological distinction between identity and interest as motives for collective action. Kin altruism is triggered by similarity, while reciprocal altruism explains cooperation among interdependent egoists. Reciprocal altruism requires on-going relations, as might be found in instrumentally motivated interest groups. Kin altruism requires densely clustered social ties among tightly knit in-groups and explains the importance of group boundaries around close cultural relatives, as well as the need for cultural markers that signify membership. To sum up, emergent and purposive rationality share the consequentialist assumption that the outcomes of action satisfy a set of relevant interests. However, emergent rationality differs on several important points: 1. 2. 3. 4. Outcomes attract efficient behavior via iteration, not intention; hence models are dynamic rather than static. Emergent rationality is meliorizing, not optimizing. Behavior is based on rules, not choices. Intended and unintended consequences have comparable explanatory power; hence rational self-interest can account for expressive solidarity in highly enthusiastic groups. The self-interests that select behaviors are not necessarily those of individuals but can also be those of the rules that control action. Hence, emergent rationality can account for individual self-sacrifice in groups that provide no compensation. Two rules for cooperation, kin and reciprocal altruism, illustrate the evolutionary logic that underlies both identity and interdependence as sources of group solidarity.

5.

6.

442

Michael W. Macy

Formal Models of Emergent Rationality Emergent rationality lends itself to formalisms that can be implemented analytically and by computer simulation. The latter is often mandated by the nonlinearity of gradient climbing on an evolutionary landscape. Recent contributions include Axelrod (1984), Allison (1992), Frank (1988), Kollock (1993), Heckathorn (1996), and Lomborg (1996). (For an overview, see Bainbridge et al., 1994 and Macy and Flache, 1995.) These studies show how formal models of kin and reciprocal altruism might be used to study the solidary effects of cultural markers and strategic interdependence. This approach is relatively new, and much work remains to be done, especially on direct applications of kin altruism to the evolution of collective identity. In recent papers, I have used simulation experiments to explore problems that relate to the emergent rationality of identity-driven collective action. Following Frank (1988), Macy and Skvoretz (1995) modeled the evolution of trust among agents who must learn to read telltale signs of character. In this experiment, cooperation in a Prisoners Dilemma required the coordination of shared meanings of symbolic cues. Conventions were created by an adaptive process tied to the outcomes of compliance, while the consequences of compliance depended on the emergence of a consensus about the meaning of the telltale signs. In another experiment, Flache and Macy (1996) modeled performance of group-rewarded work teams whose members used social approval to discourage free-riding in a multiplex Prisoners Dilemma. In the absence of outside sources of social support, dependence on the team for approval might be expected to improve compliance with production norms, as predicted by Homans (1951). However, the simulations showed how dependence can instead cause performance to suffer. With informal controls, peer pressure can trap work teams in badly sub-optimal equilibria. Although dependence makes team members more responsive to social sanctions, it also makes them less likely to risk friendships by sanctioning those who give them approval. Social approval then flows into the maintenance of dyadic relations rather than enforcement of compliance with collective obligations. In a follow-up experiment (Macy, Kitts, and Flache, 1997), we used a Hopfield device (a type of artificial neural network) to model the effects of homophily on cooperation in these same work teams. Homophily refers to the tendency for likes to attract. Attraction in turn determines the value placed on approval and the propensity to imitate. Conversely, we

IDENTITY, INTEREST AND EMERGENT RATIONALITY

443

assume that unlikes repel and that repulsion, in turn, motivates efforts at differentiation. Simulations showed how groups then tend to split into cliques of workers and shirkers. As in the earlier experiment with dyadic friendships, we found that dependence on the group for social approval did not prevent the formation of an oppositional clique. Another recent experiment used genetic algorithms to study the evolution of reciprocal altruism (Macy, 1996a; see also Nowak and Sigmund, 1993). The simplest and best-known implementation of reciprocal altruism is Tit for Tat or TFT. This rule never takes advantage of others but retaliates swiftly if provoked. However, TFT has two serious weaknesses. First, because TFT refuses to exploit, a population of reciprocal altruists can be invaded by naive (unconditional) cooperators who then invite the return of more aggressive and predatory rules (Axelrod, 1984). Second, TFT is vulnerable to random miscues that trigger cycles of recrimination when interacting with other reciprocators (Bendor, 1993; Kollock, 1995). The earlier example of the abusive parent of a rotten kid illustrates this echo effect, or self-reinforcing sub-optimal equilibrium. Simulated evolution showed that reciprocal altruism is more robust when based on a rule for conditional reciprocity. The program, known as PAVLOV (or Win-Stay, Lose-Shift), is as simple as TFT: Favor actions whose outcomes are satisfactory and avoid those that are not. Conceptually, TFT and PAVLOV are complements: TFT teaches, PAVLOV learns. TFT rewards cooperation and punishes defection, as if trying to modify the partners behavior. Conversely, PAVLOV modifies its own behavior by repeating choices whose outcomes are satisfactory and avoiding those that are not. PAVLOV avoids the pitfalls of TFT by occasionally testing to see if the partner uses conditional or unconditional rules. To avoid blood feuds, PAVLOV stops retaliating if it appears that a noncooperative partner is acting defensively rather than aggressively. Unlike TFT, PAVLOV periodically extends an olive branch, but then quickly retracts it if the peace gesture is rebuffed. Although this makes PAVLOV vulnerable to unconditional aggressors, the risk is mitigated by a complementary trait: PAVLOV has no remorse about exploiting naive altruists who unwittingly invite invasion by predators. Hence, PAVLOV stops reciprocating cooperation if it appears that a cooperative partner can be exploited. This aggressiveness is clearly not altruistic. Unlike TFT, which never seeks to exploit, PAVLOV reciprocates cooperation only when crime appears not to pay.

444

Michael W. Macy

Conclusion The interest-identity distinction turns on the relative explanatory power of the consequences and meaning of participation in collective action. Consequences include the costs of contribution, value of the public goods, and access to selective incentives. The meaning of participation refers to the symbolic affirmation of shared social classifications and normative protocols that regulate interaction and structure social life. Yet symbolic behavior entails consequences, even if these are unintended by the actors. Evolutionary theory can tap the explanatory power of the unintended consequences of symbolically meaningful and expressive collective action. Evolution is used here to refer to an iterative and stochastic learning process that alters the probability distribution of rules in an ecology characterized by finite resources that constrain propagation. Evolutionary models make interest explanations applicable to expressive as well as purposive behavior, but the interests that matter are those of the underlying rules, and not of the agents whose behavior the rules control. This points to the possibility for a more general theory of group solidarity and social order, encompassing both identity and interdependence, and grounded in self-interested rationality, properly understood. Nevertheless, a final note of caution is in order. Unlike genes, behavioral rules lack any physiological equivalent of a transferable nucleic acid that might preclude blending, replication error, or the inheritance of acquired traits. There is no social equivalent of the central dogma in molecular genetics that makes the gene plausible as the unit of selection. Hence, much more theoretical research is needed before we can know if rule-selection can be rigorously applied to social and cultural evolution. The point of this paper is only to suggest that the effort may be warranted by the potential for important breakthroughs in empirical knowledge if rule-selectionist theoretical models can be made to work.
NOTES This research was supported by a grant from the National Science Foundation, #SBR 95-11461. I wish to thank Victor Nee and the editors for their comments and suggestions. 1. Washington, in a 1778 letter to John Bannister, cited in Mansbridge, 1990: 26. 2. For an insightful survey of the history of self-interest in political theory, see Mansbridge, 1990. 3. In a related paper (Macy, 1996b), I compare pragmatism and rationalism as the philosophical foundations for evolutionary and analytical game theory, respectively.

IDENTITY, INTEREST AND EMERGENT RATIONALITY

445

That paper also advances a distinction between forward-looking theories of rational choice and a pragmatist alternative I call rational action. 4. For example, A major use of the concept of social capital depends on its being a byproduct of activities engaged in for other purposes (Coleman 1990, p. 312). 5. Care must be taken not to read purpose or intent into this characterization. By definition, rules instruct actions by individuals (and organizations) that have adopted the rule, and these actions have consequences that may alter the viability of the rule. Hence, like organizations, rules have interests, even though they are not purposive and do not pursue interests in the way that a purposive individual might. A rulecentered approach is thus incompatible with assumptions of purposive action and is strictly limited to applications where iteration, not intention, is the link between outcome and action. 6. Some readers may be bothered by the claims that can learn how to spread more effectively, as if they were intelligent viral entities that infect and manipulate humans for their own nefarious purposes. That might make interesting science fiction, but the point here is only to introduce a rules-eye view of social life, so that our attention is directed to the evolutionary consequences of behavior, not for the agent, but for the rule that instructs the agent. Evolutionary game theorists often write about the fortunes of competing strategies using similar animating language. REFERENCES Alexander, R. 1987. The Biology of Moral Systems. New York: Aldine de Gruyter. Allison, P. 1992. The Cultural Evolution of Beneficent Norms. Social Forces 71(2):279-301. Axelrod, R. 1984. The Evolution of Cooperation. New York: Basic Books. Bainbridge, W., E. Brent, K. Carley, D. Heise, M. Macy, B. Markovsky, and J. Skvoretz. 1994. "Artificial Social Intelligence." Annual Review of Sociology 20:407-436. Becker, G. 1991. A Treatise on the Family. Cambridge, MA: Harvard University Press. Bendor, J. 1993. Uncertainty and the Evolution of Cooperation. Journal of Conflict Resolution 37:709-34. Berger, J., T. Conner, and M. Fisek. 1974. Expectations States Theory: A Theoretical Research Program. Cambridge, MA: Winthrop. Berger, P. 1966. The Sacred Canopy. New York: Anchor. Bienenstock, E. and P. Bonacich. 1992. The Core as a Solution to Exclusionary Networks. Social Networks 14:231-44. Blau, P. 1964. Exchange and Power in Social Life. New York: Wiley. Chong, D. 1991. Collective Action and the Civil Rights Movement. Chicago: U. Chicago Press. Coleman J. 1990. Foundations of Social Theory. Cambridge, MA: Harvard U. Press. Collins, R. 1992. Sociological Insight. Oxford: Oxford University Press. Cook, K. S. and T. Yamagishi. 1992. Power in Exchange Networks: A PowerDependence Formulation. Social Networks 14:213-30. Dawkins, R. The Selfish Gene. 1976. Oxford: Oxford University Press. Durkheim, E. 1960. The Division of Labor in Society. New York: Free Press. Ellickson, R. 1991. Order Without Law. Cambridge: Harvard University Press. Elster J. 1989. The Cement of Society. Cambridge, England: Cambridge U Press. Emerson, R. 1962. Power Dependence Relations. American Sociological Review. 27:31-41.

446

Michael W. Macy

Erickson, K. 1966. Wayward Puritans. New York: Wiley. Fireman, B. and W. Gamson. 1979. Utilitarian Logic in the Resource Mobilization Perspective. In M. Zald and J. McCarthy, eds., The Dynamics of Social Movements, pp. 8-44. Cambridge, MA: Winthrop Press. Flache, A. and M. Macy. 1996. The Weakness of Strong Ties: Collective Action Failure in a Highly Cohesive Group. Journal of Mathematical Sociology 21:3-28. Frank, R. 1988. Passions Within Reason: The Strategic Role of the Emotions. New York: Norton. Friedkin N. 1992. An Expected Value Model of Social Power: Predictions for Selected Exchange Networks. Social Networks 14:213-30. Gray, L. and I. Tallman. 1984. A Satisfaction-Balance Model of Decision Making and Choice Behavior. Social Psychology Quarterly 42:146-159. Gusfield, J. 1963. Symbolic Crusade. Chicago: U. Illinois Press. Hamilton, W. 1964. "The Genetical Evolution of Social Behaviour." Journal of Theoretical Biology 7:1-52. Hayek, F. 1967. Studies in Philosophy, Politics and Economics. Chicago: University of Chicago Press. Hechter M. 1987. Principles of Group Solidarity. Berkeley: U. of California Press. Heckathorn, D. 1989. Collective Action and the Second-Order Free-Rider Problem. Rationality and Society 1:78-100. Heckathorn, D. 1996. The Dynamics and Dilemmas of Collective Action. American Sociological Review 51: 250-277. Heider, F. 1958. The Psychology of Intergroup Relations. New York: Wiley. Hogg, M. A. and J. C. Turner. 1987. Social Identity and Conformity: A Theory of Referent Information Influence. In W. Doise and S. Moscovici, eds. Current Issues in European Social Psychology, vol. 2, pp. 139-182. New York: Cambridge University Press. Homans, G. 1951. The Human Group. New York: Harcourt Brace. Homans, G. 1961. Social Behavior. Its Elementary Forms. New York: Harcourt Brace & World. James, W. 1981. Principles of Psychology. Cambridge: Harvard University Press. Jasso, G. 1980. A New Theory of Distributive Justice. American Sociological Review 45:3-32. Kelley, H. and J. Thibaut. 1978. Interpersonal Relations: A Theory of Interdependence. New York: Wiley. Kollock, P. 1993. An Eye For An Eye Leaves Everyone Blind: Cooperation And Accounting Systems. American Sociological Review 58:768-86. Kollock, P. 1995. Transforming Social Dilemmas: Group Identity and Cooperation. In P. Danielson, ed., Modeling Rational and Moral Agents. Oxford: Oxford University Press. Lawler, E. and J. Yoon. 1996. Commitment in Exchange Relations. American Sociological Review 61:89-108. Lomborg, B. 1996. Nucleus and Shield: The Evolution of Social Structure in the Iterated Prisoner's Dilemma. American Sociological Review 51:278-307. Luker, K. 1984. Abortion: the Politics of Motherhood. Berkeley: University of California Press. McPherson, J. M. and L. Smith-Lovin. 1987. Homophily in Voluntary Organizations: Status Distance and the Composition of Face-to-Face Groups. American Sociological Review 52:370-9.

IDENTITY, INTEREST AND EMERGENT RATIONALITY

447

Macy, M. 1993. Social Learning and the Structure of Collective Action. In E. Lawler et al. eds., Advances in Group Processes, 10:1-36. Greenwich, CT: JAI Press. Macy, M. and A. Flache. 1995. "Beyond Rationality in Theories of Choice." Annual Review of Sociology 21: 73-91. Macy, M. and J. Skvoretz. 1995. "The Secret of Success: A Telltale Sign for Trusting Strangers." Paper presented at the Annual Meeting of the American Sociological Association. Washington, D.C. Macy, M. 1996a. Natural Selection and Social Learning in Prisoner's Dilemma: Coadaptation with Genetic Algorithms and Artificial Neural Networks. Sociological Methods and Research 25: 103-137. Macy, M. 1996b. Pragmatism, Emergent Rationality, and Evolutionary Game Theory. Presented at New York Meeting of the American Sociological Association. Macy, M., J. Kitts, and A. Flache. 1997. The Weakness of Strong Ties II: Collective Action Failure in a Self-organizing Social Network. Presented at Toronto Meeting of the American Sociological Association. Mansbridge, J. 1990. Beyond Self-Interest. Chicago: University of Chicago Press. Markovsky, B. 1992. Network Exchange Outcomes: Limits of Predictablility. Social Networks 14:267-286. Markovsky, B., J. Skvoretz, D. Willer, M. Lovaglia and J. Erger. 1993. The Seeds of Weak Power: An Extension of Network Exchange Theory. American Sociological Review 58:197-209. Marwell, G. and P. Oliver. 1993. The Critical Mass in Collective Action: A MicroSocial Theory. Cambridge, England: Cambridge University Press. Maynard Smith, J. 1976. Group Selection. Quaraterly Review of Biology 51:277-289. Mead, G. H. 1934. Mind, Self, and Society. Chicago: University of Chicago Press. Nee, V. and P. Ingram. Forthcoming. Embeddedness and Beyond: Institutions, Exchange, and Social Structure. In M. Brinton and V. Nee, eds. The New Institutionalism in Sociology. New York: Russell Sage. Nowak, M. and K. Sigmund. 1993. A Strategy of Win-Stay, Lose-Shift that Outperforms Tit-for-Tat in the Prisoners Dilemma Game. Nature 364:56-8. Oberschall, A. and H. Kim. 1996. Identity and Action. Mobilization 1:63-86. Oliver, P. 1980. Rewards and Punishments as Selective Incentives for Collective Action: Theoretical Investigations. American Journal of Sociology 85:1356-75. Orbell, J. and R. Dawes. 1991. A Cognitive Miser Theory of Cooperators Advantage. American Political Science Review 85:515-28. Ostrom, E. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge, England: Cambridge University Press. Peirce, C. 1955. In J. Buchler, ed., The Philosophical Writings of Peirce. New York: Dover Press. Rapoport, A. and A. Chammah. 1966. Prisoner's Dilemma: A Study in Conflict and Cooperation. Ann Arbor: University of Michigan Press. Ruse, M. and E. O. Wilson. 1986. Moral Philosophy as Applied Science. Philosophy 61:173-92. Sherif, M., O. J. Harvey, B. J. White, W. R. Hood, and C. W. Sherif. 1961. Intergroup Conflict and Cooperation: The Robber's Cave Experiment. Norman: The University of Oklahoma Press. Simon, H. 1982. Models of Bounded Rationality. 2 vols. Cambridge, MA: MIT Press. Simon, H. 1992. "Decision-Making and Problem-Solving." In M. Zey, ed. Decision Making: Alternatives to Rational Choice Models. Newbury Park: Sage Publications.

448

Michael W. Macy

Tajfal, H. 1982. Social Identity and Intergroup Relations. Cambridge, England: Cambridge University Press. Turner, R. 1996. The Moral Issue in Collective Behavior and Collective Action. Mobilization 1:1-16. Vanberg, V. 1994. Rules and Choice in Economics. London: Routledge, 1994. Willer, D. and B. Anderson. 1981. Networks, Exchange, and Coercion: The Elementary Theory and Its Applications. New York: Elsevier. Winter, S. 1986. Comments on Arrow and on Lucas. In R. Hogarth and M. Reder, eds. Rational Choice: The Contrast Between Economics and Psychology, pp. 24350. Chicago: University of Chicago Press. Wynne-Edwards, V. 1962. Animal Dispersion in Relation to Social Behaviour. Edinburgh: Oliver and Boyd. ________________________ MICHAEL MACY is Professor of Sociology at Cornell University. He studies cooperation in social dilemmas using computer simulation and laboratory experiments with human subjects. His recent papers are available on his website at www. people.cornell.edu/pages/mwm14.