The methodology implicit in empirically grounded social scientific studies of religion naturally allies with forms of semantic holism. However, a well known argument which questions whether holism in general is consistent with the fact that languages are learnable can be extended into an epistemological one which questions whether holism is consistent with an empirical methodology. In other words, there is question whether holism, in fact, makes social science possible. I diagnose the assumptions on which that objection rests, pointing out that (...) they are not necessary for semantic holism per se. I argue that a form of semantic holism along the lines advocated by W. V. O. Quine and Donald Davidson can avoid the objection. Moreover, when seen in this light, it has the resources to withstand an even deeper methodological problem. In so doing, though, several important but overlooked implications for the study of religion emerge. (shrink)