Abstract

Multiple realizability seems to be empirically justified and provides the conceptual basis for the autonomy of the special sciences. But it is mysterious. In this talk I propose a new reductionist approach to the special sciences that removes the mystery: I explain why the special sciences kinds appear to be multiply realized although they are identical with physical kinds and in what sense the special sciences kinds and laws are autonomous although they are physical laws. This approach is based on the approach by Hemmo & Shenker to the reduction of thermodynamics to statistical mechanics.