ABSTRACT
Social science explanation does not fit the traditional model of scientific explanation in terms of "laws": there simply are none. So "reductionism" in the classical philosophy of science sense is not an option. The developing field of social neuroscience approaches the relationship from the bottom up: from the identification of cognitive mechanisms grounded in neural processes with observable, measurable correlates, to known social phenomena such as empathy or free-rider punishment. The core computationalist model – consisting of a pre-given architecture in which complex modules have evolved from more rudimentary ones – sets the contours for how the standard model applies to social phenomena. Computationalism comes in different forms with respect to representations: one view is minimalist and treats information and its processing as a causal process. The strong form of this argument, however, treats representations as complex inferential objects best modeled, even for animal cognition, as parts of syntactic systems.
