If you are a deflationist of sorts, you might reasonably toy with the thought that, when ‘that’s true’ is said of an attempted assertion which misfires so badly that no proposition is expressed, then the purported endorsement of a saying misfires too. The suggestion naturally leads to the view that the Liar sentence doesn’t express a proposition. Field briefly touches on the view at p. 132 of Saving Truth from Paradox, and brusquely rejects it (without mentioning any of its defenders):
I would like to point out (what is widely known but still deserves emphasis) that it simply won’t work on any ordinary notion of proposition. Consider my remark that what the least intelligent man in the room was saying isn’t true [when sadly, I myself fitted that description]. When I said that, I firmly believed it, and had good evidence for so believing it: I firmly believed that Joe Schmoe was the least intelligent man in the room, that he had uttered ‘Maine is bigger than Colorado’, and that Maine is not bigger than Colorado; and I had good evidence for all these claims. What I said wasn’t nonsense. Indeed, had the facts been different — if I hadn’t overestimated my own intelligence or underestimated his –what I said would have been clearly true. So if we’re going to talk of propositions at all, and talk of sentences and belief-states as expressing propositions, then this would seem to be a clear example: I stood in the belief relation to a perfectly coherent proposition which (like many propositions that are believed) is “conditionally paradoxical” in classical logic. Unfortunately, the conditions for classical paradoxicality were satisfied despite my evidence to the contrary.
But this is very surprising — or at least, surprising when offered without considerably further elaboration in 2010 (as if the notion of proposition is not contested). Imagine a parallel debate with that kind of externalist about content who says that in certain cases of reference-misfiring, we fail to express a thought. Thus, Field hallucinates Gwyneth Paltrow sitting across the restaurant, and gesturing to empty space, purports to express a singular proposition, saying “that woman was brilliant in Shakespeare in Love“. The externalist of a certain stripe (Gareth Evans, for example) will hold that here — due to referential misfiring — no proposition is expressed. Now imagine Field similarly protesting
I would like to point out that this too simply won’t work on any ordinary notion of proposition. When I said, at the restaurant, that that woman was brilliant in Shakespeare in Love, I firmly believed it, and had good evidence for so believing it. I firmly believed that Gwyenth Paltrow was brilliant in Shakespeare in Love, and that I could see that Gwyneth was sitting across the restaurant; and I had good evidence for both those claims. What I said wasn’t nonsense. Indeed, had the facts been different — if I hadn’t been deceived by candlelight –what I said would have been true. So if we’re going to talk of propositions at all, and talk of sentences and belief-states as expressing propositions, then this would seem to be a clear example: I stood in the belief relation to a perfectly coherent proposition which (like many propositions that are believed) misrepresented the world despite my evidence to the contrary.
But of course the externalist has by now a familiar patter with which aims to fend off this kind of (he would claim) question-begging attack. No one would think that the imaginary paragraph I’ve just foisted on Field would pass muster without a lot more elaboration and a lot more engagement with the externalist debate (I’m not taking a view here about who is right — just emphasizing the need for more than the imaginary Field says). No one would think that in, that context, we can just assume that the same proposition is expressed whether the world goes luckily or unluckily as to whether Gwyneth was actually present when the claim was made. So I wonder why the real Field thinks it is so obvious that the same proposition is expressed whether things go luckily or unluckily when the world threatens a “contingent liar”? The defender of a no-proposition diagnosis of the Liar will clearly balk at Field’s too-quick and unargued assumption.
There’s good reason to think propositions are *not* hyperintensional entities (shared content arguments), and there’s good reason to think the object of beliefs *are* hyperintensional. Hence there’s good reason to think the object of beliefs are not propositions. So the guy in the example stood in a relation to some (“perfectly coherent”) hyperintensional thing even though he stood in no such relation to a proposition.
Moreover, there’s good reason to think that meanings are *not* propositions (or that characters are contents, etc.) So gappy sentences can be meaningful (rather than “nonsense”) even when they express no proposition (e.g. because they have truth conditions (Davidson)).