Monday, November 14, 2005

Novemberfest and Conventionalism

A busy couple of weeks at Arche, with two workshops (on semantic paradox and vagueness respectively), and talks by (among others) Greg Restall, JC Beall, Hartry Field, Diana Raffman, Achille Varzi and Dominic Hyde (with many others in town - e.g. Steve Yablo, Richard Heck, Graham Priest). I'll try and post some paper reports here as the fortnight progresses.

In the meantime, does anyone think the following is a way to rescue conventionalism about necessary/a priori/analytic truth from one obvious type of objection to it?

Here's the objection (as expressed by BonJour):

[w]hat convention might be adopted that would make it possible for something to be red and green all over at the same time? It is, of course, obvious that new conventions could change the meaning ... of the words ‘red’ and ‘green’, but there is no plausibility at all to the idea that such changes would result in the falsity of ... the proposition that nothing can be red and green all over at the same time, as opposed to merely altering the way in which [that proposition is] expressed.
(From In Defence of Pure Reason, p. 53.)

Mightn’t the conventionalist try and distinguish two ways of understanding the claim that had our conventions been different it would have been possible for something to be red and green all over? On one of these, the relevant counterfactual worlds are being assessed by us, and therefore our own conventions are in play. So we deny that these worlds where our conventions are different are worlds and something can be red and green all over (because our actual-world conventions fix that nothing can be red and green all over in any world, including these ones). On the other approach we consider, not what is true at those worlds considered as worlds governed by our actual conventions, but what is true at those worlds considered as worlds governed by the conventions we have at the worlds in question. So on this second approach we accept that there are some worlds where it is possible for something to be red and green all over at the same time.

Maybe the conventionalist could argue that our intuition that changing our conventions wouldn’t change the facts (the intuition driving BonJour's objection) is well-enough preserved by the result we get on the first approach. But on the second approach there are worlds where the proposition is made false by the fact that we have different conventions at those worlds. And this (she might say) is enough to rescue the thought that we could have had different conventions which would have made it false that nothing is red and green all over - i.e. enough to rescue conventionalism from the objection.


Ross Cameron said...

Right, bloody computer just ate my last attempted post, so let's try again. (Every day, being a luddite seems more and more attractive.)

I talked about a similar issue with Ted Sider a while back. He was toying with the idea of a two-dimensionalist treatment of 'necessarily'. When we consider a world w, in which different conventions are in play, as counter-factual, the actual necessary truths are necessary with respect to that world; but when we consider w as counter-actual the necessary truths are those that would be necessary were the conventions of w the actual conventions adopted.

This seems close to your idea. As a conventionalist about modality (on tuesdays, thursdays and every second sunday) I think this is the only sense in which the conventionalist should say that were the conventions different, different modal truths would be true. It is a weak agreement with that claim, since for all that has been said S5 is the correct modal logic.

Incidentally, I think Bonjour is confusing the issue by focusing on the conventions governing 'red' and 'green' rather than the conventions governing modal language. It is, according to me, nothing to do with how we use the colour terms that nothing could be red and green all over at the same time, but rather down to how we use the term 'could'.

Carrie Jenkins said...

Cool - Sider's idea does seem close to mine. (Does he have anything written on this, do you know?)

So on Tuesdays and Thursdays do you feel like this is a good-enough and/or the best available response to criticisms like BonJour's?

(PS Just to be clear, Boo conventionalism; I'm not trying to defend it, just interested in seeing whether - as I suspect - the best objections will need to be framed in non-modal terms ...)

Anonymous said...

I agree: boo conventionalism. It is a wednesday after all.

Sider hasn't said that in print, as far as I know. In fact, he was somewhat tentative about it in conversation - he convinced me more than himself I think.

In my conventionalist mode I think it is a good response to the Bonjour-type criticism. Bonjour's criticisms are directed at a weak sort of conventionalism - the kind of theory that is subject to Bonjour's criticisms we have known for a while isn't a runner, due to Quine. The sophisticated conventionalist should not accept the move from 'p is true by convention' to 'had the conventions been different, p could have been false'. And the Sider/Jenkins point shows exactly how to resist this move.

And in a few hours: hooray conventionalism!