the boundaries of language

Apologies for the lack of posts recently. I spent most of last week up in St Andrews for the Arche Vagueness Conference, which unsurprisingly turned out to be a very fun trip.

Let me take the chance to thank Andreas, Ole, Ralf and Andri, who very kindly put me up, and enabled me to be part of an effort to break the world record for the number of male philosophy graduate student bloggers simultaneously living under one roof. Cosmic. (Dan was also in town for the conference, and I got the chance to meet Avery, so, to adapt Nick Asher's joke, St Andrews seemed blogger-dense; between any two philosophy bloggers there was another one.)

The talks themselves were a slightly mixed bag, as always. I don't have much to add to the exchange between Andreas and Dan on Nathan Salmon's paper. I found myself very dissatisfied with Scott Soames' attempt to respond to the Dummett-Glanzberg challenge to the possibility of truth-value gaps (see here for links and discussion). Basically (and no doubt wildly oversimplified) the challenge is this. The intrinsic aim of assertion is truth, and either an assertion achieves that aim or it doesn't. So there are only two possible verdicts concerning whether an assertion achieved its aim; the idea that there might be assertions which fit into some third bracket is just incoherent. The suggested upshot is that a third truth-value is, at best, radically unmotivated.

I'm not really sure what to make of the Dummett-Glanzberg line of reasoning. Soames, however, attributed to them the thesis that truth is the norm of assertion, in the sense that Williamson has given that phrase; the thesis that it's constitutive of the speech act of assertion that one must only assert the true. He then recycled Williamson's arguments for favoring the knowledge rule of assertion over the truth rule, concluding that all the Dummett-Glanzberg argument showed is that his position is inconsistent with an incorrect account of the norm of assertion. But there wasn't even so much as a comment on the switch from talking of the aim of assertion to talking of the constitutive norm of assertion, and the two are quite different. It might well be true that the intrinsic aim of an assertion is to hit the truth, and the argument that truth is a binary notion can then proceed in the envisaged manner from the observation that whether that aim has been achieved is itself a binary matter. But that is, so far as I can tell, consistent with holding that the norm governing the making of assertions is more demanding than truth (for instance, the knowledge rule) or even less demanding (like the reasonably- or rationally-believed rules). Certainly there's some work to be done getting clear on the idea of an 'intrinsic aim' of assertion. But Soames' take seemed utterly uncharitable - aside from any other issues, the attribution of the thesis that only truth warrants assertion to the Dummett of 'Truth' should have raised alarm bells concerning Soames' interpretation of the challenge. And even putting aside these doubts about whether Soames had correctly identified his target, it was difficult to feel he'd really engaged the issues; there was simply no discussion whatsoever of the many powerful criticisms that have been made of the Williamsonian account of assertion in the literature (including Matt's excellent defense of the truth account over the knowledge account). It just seemed to be taken for granted that Williamson's arguments must be cogent, something I think noone has the right to assume in the current climate.

We had a characteristically high-altitude discussion from John MacFarlane and Dorothy Edgington, acting as his commentator, on how to motivate degree-theoretic accounts of vagueness. The main sticking point between them was whether we should endorse Stephen Schiffer's suggestion that confronted with a borderline case of baldness, tallness and thinness, we should have just as much (or nearly as much) credence in the truth of 'x is bald and x is tall and x is thin' as we should have in each of the conjuncts, contrary to what would be the case if credences about borderline cases obeyed the axioms of classical probability theory. The case against Schiffer and MacFarlane, led by Edgington and Roy Sorensen, turned on the observation that our credence in the truth of 'y is bald and y is tall and y is thin' should be higher than the analogous claim about x, when y is a borderline case of thinness and tallness but a absolutely clear case of baldness. While the standard picture can accommodate this point easily, the alternative pictures proposed by Schiffer and MacFarlane both (more or less) assigned a conjunction the min of the three conjuncts, and so both claims would be assigned the same value.

Roy Sorensen's talk probably supplied the most entertainment value, and gave what, for me, was the quote of the conference:

'Epistemicism is a pie-eating competition, and the prize is more pie.'

The highlight, however, was the final session, with Crispin Wright giving his paper 'The Illusion of Higher-Order Vagueness', and Mark Sainsbury responding. I'd heard Crispin's paper last year in Austin, so there weren't any real surprises. But as Mark pointed out in his closing remarks on the conference, Crispin has been thinking about this issue for over 40 years, and we got a little bit of that distilled into 40 minutes in his talk. It left me looking forward even more to his collected papers on the topic, The Riddle of Vagueness.

Anyway, it was a real pleasure to be back in St Andrews; it was great to catch up with everyone, and I'm very glad I got to see the last official moments of the Vagueness project.

Labels: Conferences, Crispin Wright, Mark Sainsbury, Vagueness