I-Languages and T-sentences

This paper is about the relevance of Liar Paradoxes for truth-theoretic accounts of linguistic meaning. In " Framing Event Variables, " I develop an independent objection to such accounts. But in both essays, the first numbered sentence is the one displayed below. (1) The first numbered sentence in " Framing Event Variables " is false. Let 'Lari' be a name for the first numbered sentence in that other paper, Pietroski (2015). Since the definite description in (1) can be used to describe Lari, the grammatical subject of (1) can be used to talk about that very sentence. So (1) can be used to say that (1) is false. Likewise, (2) (2) The second numbered sentence in " I-Languages and T-sentences " is not true. can be used to say that it isn't true. In my view, this isn't paradoxical: it's true that (2) isn't true, and it's false that (1) is false. But such examples do tell against the Davidsonian thesis (D); (D) for each Human Language H, some Tarski-style theory of truth for H is the core of a correct theory of meaning for H where a Human Language is a spoken or signed language that a biologically normal human child can acquire, given an ordinary course of experience. In section one, I say more about thesis (D). But details aside, many Human Languages permit sentences like (1) and (2), in which a word like 'true' is combined with devices that can be used to describe a linguistic expression and express negation. These sentences are grammatically akin to (3) and (4). (3) The largest frog in Walden Pond is green. (4) Kermit is not red. Though as discussed below, it seems that any truth theory that accommodates (1) and (2) will be too sophisticated to serve as the core of a plausible meaning theory that applies to (3) and (4); and this concern is amplified if Human Languages are I-languages in Chomsky's (1986) sense. In short, sentences like Lari highlight a deep tension between (D) and an independently plausible conception of Human Languages. I think the best response is to deny that (1) and (2) have truth conditions, and then admit that (3) and (4) also fail to have truth conditions. While sentences of a Human Language are often used to express truth-evaluable thoughts, it remains a hypothesis that these sentences are themselves truth-evaluable. And this hypothesis has implausible …