Monday, October 03, 2005

Imagination and modality at FBC

For philosophers who can pull themselves away from gawking at the distraught Republicans, Jonathan has a nice post on connections between modality and imagination. I've wondered for a while whether one can use imagination to explain modal thought and discourse in the way that philosophers like Gibbard use (supposed) mental states of norm-acceptance to explain normative thought and discourse. I haven't thought enough about this to say anything particularly good, but I like Jonathan's observation that imagining and belief often interact as follows: we imagine a particular situation where (perhaps among other things) p is the case, and our belief that p -> q causes us to imagine q subsequently happening.


Aidan said...

I'm worried that the kind of story Ichikawa tells can't be what underwrites a large portion of our modal knowledge. To take a limiting case, it doesn't seem plausible that I learn that 'P and not-P' entails the truth of the counterfactual conditional '(P and not P) so Q' by imagining a particular situation where 'P and not P' is true, then work my way somehow to the conditional.

I have a bunch of other similarly picky points I wanted to make, and perhaps they're not that damaging, but I think I have a stronger point to make, so I'll cut to the chase. The proposal seems to at best explain how we could come to be justified in believing a subjective conditional, and extending it to account for our modal knowledge concerning the truths of counterfactual conditions seems problematic (glossing over lots of details, the difference is that a subjective p so q is true if q is true at the closest p world(s), while the counterfactual is true if q is true in all p worlds).

The point should be obvious; imagining a particular situation where p is the case isn't going to help determining the truth of the counterfactual p so q, where we need to establish that q holds in all p worlds.

I hope nobody thinks that appealing to a further belief that p -> q will help here (I take it that's meant to be the material conditional, Neil?). It's just not true that our knowledge of the truth of counterfactual conditionals always (ever?) rests on a justified belief in some corresponding material conditional, for obvious reasons:

Take the counterfactual p so q. Assume p is in fact false. p -> q is true, but clearly knowing that won't help someone who is imagining a p situation to justifiably imagine it being or becoming a q situation. To put the point another way, it can't in general be the case that 'our belief that p -> q causes us to imagine q subsequently happening' (from Neil, not FBC post), since that conditional can be true without *any* causal connection between p and q.

Of course, Ichikawa doesn't claim otherwise, but the problem is that it seems he must if he is to offer even a flawed epistemology for counterfactual conditionals. So I stand by my opening comment, this is at best a partial account of what justifies our modal talk and thought; if we want a remotely unified account we should look elsewhere.

Aidan said...

Actually, I'm starting to think my use of subjunctive and counterfactual isn't standard (I'm wondering where I picked it up). But nevermind, the point can clearly be made in more standard terminology; it's not clear how we could be come to be justified in believing a strict conditional on this picture, since we haven't (and presumably often couldn't have) imagined all the relevant p situations in the manner prescribed by the account.

Jonathan said...

I guess I should clarify the dialectic. Williamson offered a substantial argument, including several pages of modal logic, for the interesting substantive philosophical thesis that modal reasoning is just a special case of counterfactual reasoning. I'm not committing as to whether that's right, and I certainly didn't mean to have compellingly argued that it was right. I'm just talking about counterfactuals; let's leave modality to Williamson for now.

Now, how do I get from my imagining that p to an imagining that q, given that p is false but does counterfactually imply q? Well, it's an uninteresting case if I just already know that p implies q. Sure. But there are lots of ways I might do it. I don't really have a particular answer in mind, nor do I think there's a standard way. Here is my suggestion: it's however it is that I would go from a belief that p to a belief that q. There might be inference in there, or innate scientific theory, or whatever. I don't mean to have solved the problem; I've just tried to assimilate it to another problem.