Abstract

A core Darwinian idea is that evolution will lead to well-adapted organisms, with phenotypes that maximize their fitness relative to the available alternatives. Grafen’s ‘formal Darwinism project’ attempts to make this idea precise, by explicitly linking the process of natural selection and the optimality of individuals’ phenotypes. Grafen’s analysis ties in closely with the unity-of-purpose constraint on agency, but does not amount to a general vindication of adaptationist assumptions. Under frequencydependence, the theory of adaptive dynamics shows that natural selection does not necessarily lead to phenotypes which maximize fitness conditional on their being fixed in the population. These results suggest that there is no theoretical principle to the effect that natural selection will tend to produce adaptation. The justification for agential thinking in biology must thus be empirical, not theoretical.