Trust and Reciprocity: Interdisciplinary Lessons for Experimental Research

7. Conflict, Interpersonal Assessment, and the Evolution of Cooperation: Simulation Results

170 Chapter 7 Conflict, Interpersonal Assessment, and the Evolution of Cooperation: Simulation Results James Hanley, John Orbell, and Tomonori Morikawa H ow cooperative dispositions might have evolved among social animals has, for many years, been productively addressed within the prisoner’s dilemma paradigm. That game captures the intuition that by cooperating, individuals can often produce more than is possible by their separate efforts and also that selfinterest can lead individuals to undermine their cooperative efforts.1 The structure is robust, with only minor elaborations necessary to show how populations can realize their cooperative opportunities. The best-known such elaboration is iteration of the game; simply requiring players to interact in a sequence of prisoner’s dilemmas can lead them to adopt cooperation-inducing strategies such as tit-for-tat (see, for example, Axelrod 1984; Nowak and Sigmund 1992) and winstay , lose-shift2 (Nowak and Sigmund 1993, see also Schuessler 1989 and Vanberg and Congleton 1992). Nevertheless, studying the prisoner’s dilemma in isolation from the full complexity of a subject’s social life might have led researchers to miss important processes that bear on the evolution of cooperation. In particular, natural societies make it possible for individuals to engage in conflictual games with one another as well as cooperative ones;3 and the fact that choices must be made among these two types of games as well as within them raises the possibility that the evolu- Conflict, Assessment, and Cooperation 171 tion of cooperativeness has been critically influenced by the fact that those “prior” between-game choices are necessary. Notably, individuals intending cooperation and defection might differ in their willingness to play cooperative games, moving disproportionately into conflictual games and, thus, altering the proportion of cooperators and defectors in the cooperative games that are consummated. As Robert Trivers (1971) classically points out, anything that alters the probability that cooperators will enter into play with one another also alters the probability that cooperation will survive (see also Orbell and Dawes 1991), and adding the option to play a conflictual game might have just that effect. We address that possibility in the present chapter. To do so, we developed an evolutionary simulation—the Anticipatory Interactive Planning Simulation—that gives agents the option of playing both cooperative and conflictual games with one another but in which choices must be made among such games. To make those choices, individuals are equipped with capacities that give them some control over the information that others receive about their intentions and also over the accuracy of the information that they receive about others’ intentions. These mechanisms are subject to evolution by natural selection, as is the disposition to cooperate, although not cooperating or defecting behaviors as such. In these terms, we arrive at two conclusions: First, among populations in which individuals have only such cooperative games available—in which the option of fighting does not exist—there is no selection on the capacity to accurately recognize others’ intentions, and cooperativeness itself is negatively selected , under some parameters such that the society itself dies. Second , when individuals within populations do have the option of fighting among themselves, positive selection on the capacity accurately to recognize others’ intentions does happen, and it becomes possible also for cooperative dispositions to evolve in a positive direction . Theoretical Background Our study is developed from the “Machiavellian” or “social-intelligence ” hypothesis. In essence, that hypothesis proposes that the human brain has been, in substantial part, designed by natural selection for negotiating a complex social space comprising conspecifics. Historically , group living provided improved defense against predators, better success as predators, more-ready access to mates, and (sometimes ) assistance in raising children. Group living also brought members into competition with one another, however, with some being more successful in that competition than others, and whatever genes 172 Trust and Reciprocity contributed to that success were disproportionately passed on to successive generations. By this hypothesis, three critical cognitive mechanisms were subjected to such selection: those facilitating quick and accurate assessment of others’ capacities and intentions; those facilitating successful manipulation of information received (and believed) by other such individuals; and those that would translate information received (and believed) into adaptively appropriate actions. These ideas underlie the notable paradigmatic convergence that has been happening in recent years between developmental and comparative psychology, cognitive ethology, theory of mind, and human intentionality. They have also proved fertile for developing hypotheses —some of them now well supported by laboratory...