Abstract
David Papineau, Jerry Fodor and many others wonder how the conjunction of the following three positions can be true: 1) Special science laws: There are lawlike generalizations in the special sciences. These sciences trade in kinds that are such that statements about salient, reliable correlations that are projectible and that support counterfactuals apply to the tokens coming under these kinds. 2) Non-reductionism: The laws of some of the special sciences cannot be reduced to physical laws. 3) Physicalism: Everything there is in the world supervenes on the physical, that is, is fixed by the distribution of the physical properties in the world. The obvious problem is that (3) implies that the similarities among tokens in the world, accounting for the kinds in which the special sciences trade, and the correlations among such tokens, accounting for the laws of the special sciences, are fixed by the distribution of the physical properties. By contrast, (2) implies that some of the laws seizing such correlations are not reducible to physical laws. By using the term “token”, I mean a particular instantiating a property. Papineau’s proposal to reconcile these three positions is to account for (2) in terms of selection (pp. 6-9): There can be laws in the special sciences that are not reducible to physical laws if and only if these laws focus on effects that are selected for in a given context independently of the mechanisms by which they are brought about. Thus, the fact of there being such laws and their non-reducibility to physics do not contradict physicalism (3). The drawback is that the kinds that figure in such laws cannot enter into a rich network of laws 199 and that nothing can be causally efficacious insofar as it is a member of such a kind. In these comments, I shall try to push Papineau further in the direction of a reductive physicalism, thus solving the problem by simply abandoning (2)..