The Sense of `A Name of a Truth-Value'

Abstract

1. No one needs to be persuaded that Frege went wrong in merging sentences with singular terms in a single logical type. Even so, those who on occasion find themselves expounding Frege may well feel the need for a relatively compact argument that shows how he went wrong. There are arguments available, both long and short. The definitive short arguments are those presented by Wittgenstein, at Tractatus 4.063 and 6.111. Among longer arguments, none is more compelling than Dummett's contention in The Interpretation of Frege's Philosophy (e.g., ch. 19) that the 'merger theory' (as I shall call it) perversely undermines the very formulability of Frege's own 'context principle'. None of these exactly fits the bill I have in mind. Though succinctly expressed, Wittgenstein's points can take a while to explain; also, there is an obvious risk that the explanation will import themes foreign to Frege, in the face of which he would remain unmoved. Dummett's argument has the merit of resting on premises Frege clearly held, indeed held dear. Yet it feels as if it suffers from an indirectness inappropriate to the issue. It seems we can see that Frege went wrong. An argument that turns on so broadly theoretical a matter as the priority of sentences in semantic explanation, and thus the context principle's indispensability in a philosophical appreciation of thought and its expression, may confirm that impression; one doubts that it is apt to illuminate it. What I aim to offer here is a short(-ish) argument, with premises drawn from regions of Frege's thought local to the issue. Perhaps those who have felt the need I mentioned will find it useful; but as I also mentioned, I do not intend to persuade readers of anything they did not already believe.