Are Chemical Kinds Natural Kinds?

Abstract

Chemical categories fail to satisfy three widely accepted conditions for being a natural kind: hierarchy, discreteness and independence from interests. This failure presents a choice. Should we reject the chemical categories as natural kinds, reject the conditions themselves, or dispense with the whole idea of a natural kind? I argue that we should reject the conditions, for three reasons. Firstly, the independent justifications for those conditions are weak. Secondly, some chemical categories are so central to philosophers’ discussions of natural kinds that it hardly makes sense to deny that they are natural kinds, yet they fail to be hierarchical, discrete or interest-independent. Thirdly, their failure to meet these three conditions has no tendency to undermine their usefulness in a realist understanding of chemical classification and its role in prediction and explanation. As this is just the kind of thing that the notion of a natural kind was always supposed to do, realist philosophers of science are also justified in resisting calls for the elimination of that notion.