I’ve been revising a paper on essentialism about natural kinds, which focuses on the case of chemical kinds. A central motivation for the paper is what I take to be a confusion regarding Putnam’s classic Twin Earth scenarios: could the chemical properties of water be reproduced by some molecular structure other than H2O — say, XYZ — and if so, should we consider this substance to be water? The usual answer to the second part of the question is widely accepted: XYZ is not water. However, the empirical part of this concern is rarely discussed, at least by metaphysicians. This is no doubt partly because most metaphysicians lack the necessary knowledge of chemistry to be able to say much about it, but often it appears that the metaphysical possibility of XYZ reproducing the chemical properties of water is simply assumed. As philosophers of chemistry have been arguing for a while now, the story is far from being this simple — Paul Needham for instance argues that there is no plausibility in the claim that the essence of the natural kind ‘water’, if it is a natural kind, reduces to its microstructure (see his ‘Microessentialism: What Is The Argument?‘, 2011). There is also some discussion about this issue in the Stanford Encyclopedia of Philosophy entry ‘Philosophy of Chemistry‘, by Needham and Robin Hendry.
Regardless of the lack of scientific detail in the usual discussion regarding natural kinds and chemical kinds in particular, it is of course controversial what the role of empirical knowledge is regarding the existence of metaphysically possible molecular structures that could replicate the chemical properties of H2O. Chemistry will presumably be of some help here, but I do contend that we also need metaphysical a priori work to determine what is metaphysically possible. Putnam distances himself from the whole debate with his 1990 paper ‘Is Water Necessarily H2O?’, where he expresses his doubts about the notion of metaphysical possibility in general — he also speculates about Kripke’s interpretation of the issue. But regardless of what Putnam and Kripke themselves may think about the issue, there is an established interpretation of the Twin Earth scenarios which, for want of a better notion, I call the ‘Kripke-Putnam framework’. Philosophers like Scott Soames, who identify themselves as ‘Kripkean’, continue to work with this framework.
Now, the lesson of the Twin Earth scenarios could be understood in two radically different ways:
- INST: We know a priori that chemical substances have only one metaphysically possible instantiation which produces the chemical properties (i.e. macrophysical water-like properties, such as boiling point) of that substance.
- IDENT: We know a priori that chemical substance A is identical with chemical substance B iff they share their molecular composition.
INST appears to be an open question: we are going to need some substantial, essentialist knowledge about chemical substances as well as empirical knowledge of chemistry to be able to establish its truth. However, the Twin Earth scenarios seem to concern IDENT rather than INST, since what is at stake are our intuitions in cases where we do encounter substances such as XYZ, which replicate the chemical properties of water. Yet, in the Twin Earth scenarios it is assumed that a substance such as XYZ is metaphysically possible — that is, it is metaphysically possible that XYZ produces the same chemical properties that H2O does — and the question is simply whether we would call this substance water.
In contrast, in the paper that I’m revising I argue that the metaphysical, essentialist content of presumed a posteriori necessities concerning natural kinds, such as ‘Water is H2O’ (or perhaps better: ‘Water is composed of H2O molecules’) boils down exactly to a priori content in the sense of INST rather than IDENT. The claim is supported by a survey of recent work in the philosophy of chemistry.
Drop me a message if you’re interested in reading the draft, although it’ll probably soon find it’s way to my website as well…