Abstract

This talk examines the relationship between two alternative approaches to the evolution of social behaviour: inclusive fitness theory, and multi-level selection. A growing consensus in evolutionary biology maintains that these theories are not really competitors, despite what was once thought, but are in fact ‘equivalent’. I argue that this is correct in a sense, in that it is usually possible to find a correct expression for gene frequency change using either approach. However this only shows that the approaches are predictively equivalent, not that they are causally equivalent. In general in science, predictive equivalent is not usually taken to imply equivalence tout court; and I argue that this general moral applies to the case at hand. I examine a number of examples where either inclusive fitness or multi-level selection seems more ‘causally appropriate’ than the other. I end with a discussion of the suggestion that inclusive fitness is preferable on the grounds that it preserves the ‘individual as maximizing agent’ analogy.