Proof-conditional theories of meaning are set off from essentially and avowedly truth-based ones by the insistence that truth-conditions unverifiable in principle or practice are not a sound basis for a theory of meaning.4 Was there a penguin within five feet of the south pole thirty days ago? Questions of this kind, and others involving the distant future, have been offered as counterexamples to truth-conditional foundations for a theory of meaning. But the range of application of a proof-theoretic semantics will be circumscribed no less severely than that of its truth-conditional foil. Everyday means for getting an answer to the question concerning the penguin are as close to hand as those for knowing whether or not someone's software did give us a proof of the Four Colour Theorem. To argue, then, that proof positive informs everyday linguistic practice --- `grasp' rather than proto-juridical imputability --- involves a leap of faith comparable to that which antirealists impute to their opposite numbers.
There are, of course, other objections to the viability of a verificationist theory of meaning, not least because it is far from clear what it would amount to once expressions of propositional attitude are being considered. But then, attitude contexts (Kim believing that Sandy talks) are problematic for any would-be compositional theory of meaning, as Schiffer (1987) has argued.5 In what follows I shall try to make a substantive case for a theory of meaning based on the specific objection levelled at the proof-theoretic approach.