Water is Not H

In defending semantic externalism, philosophers of language have often assumed that there is a straightforward connection between scientific kinds and the natural kinds recognized by ordinary language users.1 For example, the claim that water is H2O assumes that the ordinary language kind water corresponds to a chemical kind, which contains all the molecules with molecular formula H2O as its members. This assumption about the coordination between ordinary language kinds and scientific kinds is important for the externalist program, because it is what allows us to discover empirically the extensions of ordinary language kind terms. While I am sympathetic to the semantic externalist project, I think that the discussion of chemical kinds by philosophers of language has been rather badly oversimplified, hiding difficulties that arise when we try to coordinate scientific kinds with the natural kinds recognized by ordinary language users.2 In this paper, I will examine these difficulties by looking more closely at the chemist’s notion of water. To help with this examination, I will begin by making explicit a principle on which I believe semantic externalists rely. The coordination principle is the thesis that scientific kinds and the natural kinds recognized by natural language users line up or can be mapped onto one another one-to-one. A brief examination of an externalist picture of kind reference will show how the coordination principle is relied on.