The genetical evolution of social preferences: where the categorical imperatives of Hamilton, Kant, and Nash meet

This paper models the genetical evolution of individual behaviour rules that guide the choice of strategies in pairwise assortative interactions under incomplete information. Building on results at the cross-roads of evolutionary theory and game theory, it is first shown that in an uninvadable population state of behaviour rule evolution, individuals are compelled to use strategies that are Nash equilibria of a lineage fitness game. Thus, choice behaviour evolves to be representable as the maximization of a utility function, as if each individual holds a personal rational preference. Second, the paper contrasts two representations of personal utility that are found to be uninvadable. The first is semi-Kantian in form. This preference averages a fitness self-interest with a Kantian interest, where the average involves the genetic relatedness between interacting individuals. The Kantian interest evaluates the consequence of own behaviour for own fitness, assuming the interaction partner adopts the same behaviour as self. The second preference is a personal inclusive fitness. This preference combines a self-regarding interest with a relatedness weighted other-regarding interest. Each of these interests takes the form of an average effect, which evaluates the consequence of expressing own behaviour, instead of average population behaviour, on a statistical average fitness to self and the interaction partner, respectively.