Natural Kinds and Nominal Kinds

Abstract

There is an important and often overlooked distinction between natural kind terms and nominal kind terms, as I have argued elsewhere.' Some familiar examples of natural kind terms are 'gold', 'water', and 'tiger'. According to Putnam and Kripke natural kind terms are rigid designators. Putnam holds that this is due to their indexicality. In the ideal case, the extension of a natural kind term is not determined by descriptive concepts or lists of non-trivial necessary and sufficient conditions. A natural kind term is meant to refer to whatever has a common underlying trait or essence, even though this underlying trait may be unknown to us. When we speak of water, we believe that there is a chemical nature that makes some stuff water, independently of its superficial characteristics. We believe that we know what this nature is in the case of water-namely, being H20. This is an empirical matter, however, and thus we might be wrong. There is no (even loose) analytical specification of what it is to be water. It is my view that nominal kind terms differ from natural kind terms in that the extension of a nominal kind term is not gathered by an underlying trait. Perhaps the best examples of nominal kind terms are the names of common artifacts such as 'pencil', 'bottle', and 'chair'. Underlying traits play no role in the semantics of such terms. The essence of a nominal kind is not a natural essence, rather nominal kinds, at least in the ideal case, do have a linguistic essence. This linguistic essence is the definition of the nominal kind term. The extension of a nominal kind term is determined by an analytical specification of superficial features such as phenomenal properties, and/or form, function, or origin. In this paper I will suggest a useful criterion for distinguishing between natural kind terms and nominal kind terms. A general term is indexical, according to Putnam, if and only if its extension is determined by a similarity relation pegged to