Jussi Haukioja

Semantic externalism about a class of expressions is often thought to make conceptual analysis about members of that class impossible. In particular, since externalism about natural kind terms makes the essences of natural kinds empirically discoverable, it seems that mere reflection on one's natural kind concept will not be able to tell one anything substantial about what it is for something to fall under one's natural kind concepts. Many hold the further view that one cannot even know anything substantial about the reference-fixers of one's natural kind concepts by armchair reflection. In this paper I want to question this latter view and claim that, because of the way our standard methodology of doing theories of reference relies on semantic intuitions, typical externalists in fact presuppose that one can know the reference-fixers of one's natural kind concepts by mere armchair reflection. The more interesting question is how substantial such knowledge can be. I also take some steps toward answering this question.

A recent strategy for defending physicalism about the mind against the zombie argument relies on the so-called conditional analysis of phenomenal concepts. According to this analysis, what kinds of states our phenomenal concepts refer to depends crucially on whether the actual world is merely physical or not. John Hawthorne, David Braddon-Mitchell and Robert Stalnaker have claimed, independently, that this analysis explains the conceivability of zombies in a way consistent with physicalism, thus blocking the zombie argument. Torin Alter has recently presented three arguments against the conditional analysis strategy. This paper defends the conditional analysis strategy against Alter’s objections

We argue that the term “intuition”, as it is used in metaphilosophy, is ambiguous between at least four different senses. In philosophy of language, the relevant “intuitions” are either the outputs of our competence to interpret and produce linguistic expressions, or the speakers’ or hearers’ own reports of these outputs. The semantic facts that philosophers of language are interested in are determined by the outputs of our competence. Hence, philosophers of language should be interested in investigating these, and they do this by testing what we would say or understand in hypothetical communication situations. In the final section of the paper we suggest some methods for investigating these outputs which are independent of whether subjects report them, and hence which might be used as an alternative to the standard use of hypothetical cases