Abstract

Indirect reciprocity explains exchange within large groups where a favour is returned by someone other than the beneficiary. Evolutionary games used to analyze this phenomenon are of two types: some authors show that simple strategies based on information about the past deeds of a partner are sufficient for indirect reciprocity to emerge; others indicate that complex strategies based on motivations have better dynamic properties. These complex strategies require knowledge of a considerable number of past interactions of partners and their former partners and use complicated reasoning. I present another approach to improve the performance of simple strategies using the ecological rationality concept. According to this approach, information can be useful in solving a problem even if it is not logically related to it. A probabilistic connection between information and an environmental criterion is sufficient. I present two simple strategies based on such connections. They use reference and experience. Neither needs information extending beyond the most recent interactions of chosen individuals. Using computer simulation I show that they can support indirect reciprocity. The strategies are particularly useful when conditions are unfavourable e.g. when the cost of helping is high and/or when information is difficult to obtain.

Keywords