Is Lewis's ‘Genuine Modal Realism’ Magical Too?

Abstract

David Lewis' has described the view that possible worlds are states of affairs or properties or propositions as ersatz modal realism, and has criticized it as 'magical' because the theory requires us to have special magical powers in order to understand it. According to Lewis, possible worlds are not 'abstract objects' like these. They are maximal spatiotemporal regions: every part of a possible world is spatiotemporally related to every other part (except in the limiting case of a possible world with no parts, one which is, say, a mathematical point). Among the possible worlds the actual world is but one, and its actuality consists in the fact that it is identified indexically, as the one we are in. Thus, the actual world is that spatiotemporal object all of whose parts are spatiotemporally related to us. Other possible worlds exist in the one and only sense in which anything exists (thus Lewis is no Meinongian there are no other modes of existence). What makes all possible worlds but ours merely possible, as opposed to actual, is that we are not in any of them: none of their parts are spatiotemporally related to me or you, or here, etc., though all of each possible world's parts are spatiotemporally related to each of its other parts. The chief attractions of this system are that (a) it provides a unified theory that can explain much about modality, causality, intentionality, and intensionality, and (b) it can do so 'reductively'. For it makes possibility and necessity intelligible even to those with the gravest doubts about modality. It does so by defining modal claims as extensional ones about spatiotemporal regions: P is merely possible if P is true in some maximal spatiotemporal region other than the spatiotemporal region constituted by the actual world. P is necessary if its denial is false in the actual maximal spatiotemporal region and in every other maximal spatiotemporal region. So, all we need in order to understand modality is the notion of spatiotemporal region. Or rather all we need is the notion of 'spatiotemporal region unconnected to any other spatiotemporal region'. Since this is apparently a non-modal notion, modal facts will turn out to be explained in terms of non-modal ones. I shall suggest, however, that the notion of spatiotemporal region Lewis