Abstract
The literature on the evolution of preferences of individuals in strategic interactions is vast and diverse. We organize the discussion around the following question: Supposing that material outcomes drive evolutionary success, under what circumstances does evolution promote Homo economicus, defined as material self-interest, and when does it instead lead to other preferences? The literature suggests that Homo economicus is favored by evolution only when individuals' preferences are their private information and the population is large and well-mixed, so that individuals with rare mutant preferences almost never get to interact with each other. If rare mutants instead interact more often (say, due to local dispersion), then evolution instead favors a certain generalization of Homo economicus including a Kantian concern. If individuals interact under complete information about preferences, then evolution destabilizes Homo economicus in virtually all games.