評判生成規範の寛容性が社会的交換の協力率に及ぼす効果: 社会的ネットワーク上での進化シミュレーション

We investigated the effect of tolerance of reputation-making norms on cooperation by conducting an evolutionary simulation with more realistic constraints than those in previous studies. Specifically, agents could make social exchanges only within their neighborhoods on the network, and could refer to the reputation of others only within two steps on the network when exchanging and rewiring. Under these constraints, we compared the effect of three reputation-making norms which varied across the levels of tolerance: i.e., image scoring, standing, and strict discriminator. The results were as follows. Under the standing norm, while social exchanges increased in number because the network became dense, cooperation was highly likely to collapse because defection tended to be more adaptive than cooperation. On the other hand, under the strict discriminator norm, while cooperation was stable, the social exchanges decreased in number because the network became sparse. These results indicate that the tolerances of reputation-making norms would affect not only the robustness of cooperation but also the structure of the social network. This implies that the intolerant reputation-making norm focusing only on level of cooperation may have negative effects on social capital by shrinking the social network as a by-product. In the second simulation, we added perceptual errors that caused agents to misinterpret another's reputation. The results showed that the cooperation was stable even under tolerant standing norm. This implies that when social exchanges are done under the constraints of the network structure and perceptual errors, even the tolerant reputation-making norm can maintain cooperation.