Abstract
In this paper, two recent interpretations of current work in behavioral genetics are rejected. Genetic reductionism, the view according to which genetic properties are causally sufficient for phenotypic traits, is dismissed because it ignores the fact that there are not only causal pathways from DNA to phenotype but also pathways that feed back from, for instance, the cytoplasm or the outside world to the genetic material. The complexity of development is acknowledged in the view known as developmentalism, which claims that a complex causal network of genetic and non-genetic factors is responsible for phenotypic outcomes. However, genetic explanations, at least in some contexts, do seem to have a privileged status. Heuristic identity theory appears to offer a more adequate interpretation for behavioral genetics. It is in many cases very illuminating to hypothetically identify phenotypic traits with genetic properties. We claim that McCauley and Bechtel's proposal calls for at least two constraints to avoid a wild proliferation of implausible identities. First, more emphasis should be placed on explanatory failures. Second, the most interesting identifications are those that make `qualitative leaps', that is, they must apply across distinct levels of analysis.