Abstraction and Explanatory Relevance; or, Why Do the Special Sciences Exist?

Abstract

Nonreductive physicalists have long used multiple realizability to argue for the explanatory “autonomy” of the special sciences. Recently, in the face of the local reduction and disjunctive property responses to multiple realizability, some defenders of nonreductive physicalism have suggested that autonomy can be grounded merely in human cognitive limitations. In this article, I argue that this is mistaken. By distinguishing between two kinds of abstraction I show that the greater explanatory relevance of some special-science predicates (to certain explananda) is both nonanthropocentric and not solely based on considerations of multiple realizability.