The number of planets is not a number

Abstract

In a number of places, Quine deploys variations on a certain argument to show that modal contexts are referentially opaque and therefore suspect.1 Here is a streamlined version: ... Since (3) is obviously false (we do not think that the number of planets must be nine), we must conclude that modal contexts are not well behaved in that they do not allow the substitution salva veritate of co-referring expressions. They are referentially opaque (and a symptom of a suspect underlying metaphysics). Quine takes it as evident that the inference in (Q) has a false conclusion.2 Since it appears to be valid, the trouble must lie in one of the premisses. If by ‘necessarily’ we mean ‘it is analytic that’ (as Quine does in this argument), (1) appears to be unexceptionable. But perhaps (2) does not licence substitution in the way Quine assumes (Cartwright 1971, Marti 1989). Or perhaps there is a scope ambiguity in (2) (Smullyan 1948, Neale 1990). In either case, the inference would be invalid and (Q) would show nothing about modal contexts. I shall not discuss these well-known and controversial lines of resistance.3 Instead, I shall suggest that there is a different reason to worry about (Q). Contrary to what Quine says and most commentators accept, (2) is not an identity statement, hence it cannot license the problematic inference. Before showing this, I briefly discuss another reply that leaves the usual interpretation of (2) unchallenged. The fact that it, too, leaves things unsettled will, I hope, add to the incentive to take a closer look at the construal of (2) shared by critics and defenders of (Q) alike.