Formalization Does Not Entail Neutrality: Evolutionary Game Theory and the Sociobiology Controversy

Abstract Model makers often find themselves in the following situation: in order for the model to be able to represent the dynamics of the target system, they have to make certain assumptions that pertain to the system. The problem with these assumptions is not that they are wrong, but that they cannot be derived from observations and experiments. Thus, alternative modeling assumptions can also be used, and the data does not adjudicate between the sets of assumptions. Since a decision must be made, the decision derives from considerations such as the theoretical commitments of the modeler. Focusing on assumptions of this sort, the article examines the assumptions of evolutionary game theory (EGT). Following previous authors who noticed that there were in fact two strands of EGT, the current work compares the set of modeling assumptions made in each. The attempt here is not to adjudicate between the two sets, but to demonstrate that for each strand of EGT, an alternative set of assumptions was available. My hope is that the examination of the chosen assumptions will reveal the theoretical commitments that motivated the choice. Focusing on the latter strand of EGT, the article addresses the following question: What assumptions about evolution did EGT embody that align this modeling tradition with sociobiology? By demonstrating these similarities, I hope to support a broader argument, namely, that models are not neutral structures with respect to their target system.