take nothing for granted
March 25, 2008An entirely new draft of an old paper, here. Comments would be nice. Here’s the introduction:
This paper is a brief plea for the relevance of linguistic phenomena (specifically, the phenomenon of semantic presupposition) to epistemological theorizing about the a priori. In particular, I propose, explain, defend, and apply the following constraint on knowability a priori (with ‘KA’ abbreviating ‘it is knowable a priori that’).
(AK) For all φ, ψ such that φ semantically presupposes that ψ: if KAφ, KAψ.
Roughly, (AK) claims a sentence’s content is knowable a priori only if its semantic presuppositions are too. §2 defines the notion of ‘semantic presupposition’ invoked by (AK). §3 makes use of this definition (and some plausible assumptions about the closure of knowability a priori under a priori knowable entailment) to argue in favor of (AK). The rest of the paper is devoted to exploring the (mostly negative) implications of
(AK) for the a priori. Well-known arguments for the contingent a priori and a priori knowledge of logical truth founder when the semantic presuppositions of the putative items of knowledge are made explicit. Likewise, certain kinds of analytic truth turn out to carry semantic presuppositions that make them ineligible to be items of a priori knowledge. On a happier note, I argue that (AK) offers an appealing, theory-neutral explanation of the a posteriori character of certain necessary identities, as well as an interesting rationalization for a commonplace linguistic maneuver in philosophical work on the a priori.
P.S. This is a very drafty draft. I haven’t put in the acknowledgements yet, but I will when I’ve had a chance to make the paper a bit stronger.
March 26, 2008 at 3:21 pm
this is a particularly obnoxious and unhelpful comment (sorry), but when i scrolled through this last night i must have loled like 5 times at your notational conventions
congrats on the draft. i’ll read it soon (soon as you read my thesis).
March 26, 2008 at 4:19 pm
which conventions?
May 4, 2008 at 3:49 pm
Nice paper. I had one little quibble. At the very end, you say that Boghossian’s account of our knowledge of logical truth runs into trouble because he presupposes that there exists a logical operator that will make the intro and elim rules for (e.g.) ∧ valid. You are certainly right that he sometimes puts things that way.
However, he is sometimes more careful:
“a particular constant means that logical object, if any, which makes valid a specified set of sentences and/or inferences involving it.” (”Analyticity Reconsidered,” p.376, my emphasis)
So then Boghossian’s claim is analogous to
‘If there’s a King of France, then he’s bald,’
which does not presuppose the (unique) existence of the KoF.
Also, I was wondering if you could help me out with something I’ve been wondering about. Is there a standard (set of) argument(s) against the view that sentences exhibiting presupposition failure are literally false (but pragmatically unacceptable), instead of neither true nor false? I would equally appreciate a direct, positive argument that such sentences are are gappy instead of false. If you don’t have time to write up a whole answer, a (specific) reference would be greatly appreciated.
Thanks,
Greg
May 14, 2008 at 2:37 pm
Hi Greg,
Thanks for your comment!
Point taken about Boghossian. But it would be wrong for Boghossian to insist on this, since retreating to this weaker claim would only get us knowledge of the following claim:
(1) If ‘&’ designates a logical object making the introduction and elimination inferences for ‘&’ valid, then ‘P&Q -> P’ is a logical truth.
It would be misleading to say that knowing (1) amounts to having knowledge of logic.
Excepting Strawsonian intuition-mongering, I don’t know what the standard arguments are “against the view that sentences exhibiting presupposition failure are literally false (but pragmatically unacceptable), instead of neither true nor false.” The notion that lots of presuppositions should be modeled as lexically encoded definedness constraints on functional semantic values (from which the truth-value-gap view follows pretty much immediately) seems to have been a fruitful hypothesis in formal model-theoretic semantics. Because of this, many semanticists seem to take it for granted that this is the right way to go (see, e.g., Heim and Kratzer’s remarks on the definite determiner). I’m not inclined to worry about this — the view is an elegant, useful one, and that’s enough for me. All this by way of saying that I’m afraid that I’m the wrong person to defend the view for a skeptic.