Abstract

Modal conventionalism is the view that two crucial forms of sameness are mind-dependent. There is no phenomenon of sameness in kind, on this view, except in virtue of our conventions for individuating nature’s kinds; there is no phenomenon of numerical sameness across time, for an individual member of some natural kind, except in virtue of our conventions for individuating such members.1 Modal conventionalism has its realist opponents. These opponents have argued, following Kripke’s lead more than thirty years ago (Kripke 1972), that the boundaries of at least many of nature’s kinds are carved out by nature itself, and not by our classifi catory practices (e.g., Millikan 2000: passim, e.g., pp. 25 and 72). But they have not generally argued, with anything approaching the same vigor, that modal conventionalism is wrong in its other main claim. They have not in general argued that the world-given connections among properties, that make those properties mind-independently be membership conditions for some natural kind, at the same time make those properties mind-independently be persistence conditions for the members of the kind—properties the departure of which constitutes a ceasing-to-exist for the object that formerly had them. On the contrary, opponents of modal conventionalism have..