Many have felt that it is impossible to defend autonomous laws of social sc
ience: where the regularities upheld are law-like it is argued that they ar
e not at base social scientific, and where the phenomena to be explained wo
uld seem to require social descriptions, it is argued that laws governing t
he phenomena are unavailable at that level. But is it possible to develop a
n ontology that supports the dependence of the social on the physical, whil
e nonetheless supporting the explanatory power of genuinely autonomous soci
al scientific laws? The aim of this paper is to show that reductive explana
tion is not a requirement of a 'naturalist' ontology, thereby defending an
account of supervenience as a suitable framework within which to recognize
a metaphysical relationship between the natural and the social that is cons
istent with the pursuit of autonomous nomological social scientific explana
tions.