基于Evolution,Trust,andReciprocity.pptVIP

Evolution, Trust, and Reciprocity Robert Kurzban The Evolution of Cooperation: Constraints on Models No model of the evolution of cooperation should predict promiscuous cooperation Cue-based models (“green beard”) have profound difficulties, in particular the problem of mimics. Plausible models of the evolution of cooperation must provide a basis by which exploitation is minimized, or otherwise eliminate the problem The Evolution of Cooperation: Minimizing Exploitation Exploitation can be minimized through: Punishment (but note the 2nd order problem) Boyd et al., 2003 Withdrawing cooperation, changing groups… Aktipis, in prep. Indirect reciprocity models: limiting transactions? Panchanathan Boyd, in press (JTB) Exploitation irrelevant in: kin selection models, honest signaling models The Evolution of Cooperation: Cognitive Mechanisms A well-designed cognitive system should be fashioned to reap benefits of cooperation while minimizing exposure to exploitation: Process cues that inform computations regarding cooperation (costs/benefits, features of others, etc) Decide whether/how much to trust (endure cost with expectation of future benefit) Models of the evolution of cooperation Kin selection doesn’t generalize because r falls off fast independent assortment of alleles makes “similarity” irrelevant. Reciprocal altruism doesn’t generalize easily because (e.g., Axelrod, 1986) When common, “tolerant” reciprocators are replaced by defectors. When rare, “intolerant” reciprocators find insufficient conspirators (unless there is a lot of assortment). Punishment presents a 2nd order problem (e.g., Boyd Richerson, 1988) Models of the evolution of cooperation Some models suggest that striving for uniqueness/irreplaceability will lead to mechanisms that cause individuals to deliver benefits to others. Note: KS and RA are very different, and suggest that it’s not the number of interactants that matters, but the relevant selection pressures (pace Caporael et a