I had a chat about my favorite topic in the pub the other day. That's right, beer. Shortly after that, I had another chat about modality. What started off (as I recall) as chat about my research soon drifted off to a strange (but wicked cool) mid-point between my research and that of my friend, let's call him Schmomas. Schmomas' research is into the nature of logical laws and structure, whether it is something worldly a la Sider, or something unworldly, a la Hirsch. After apologising for what is probably a gross oversimplification of Schmomas' research, I'll get to the point.
Schmomas, I'm sorry for grossly oversimplifying your research.
Having discussed the potential for essentialist accounts of modality, we started talking about how one might account for logical truth in particular. Fine reckons that we can ground such things in the essences of logical concepts (though I'm guessing he doesn't really mean to say 'concepts'), but let's face it, that's kinda weird. If we're going to be asking people to accept essences, let's at least try to slow down and not reify absolutely everything we can think of. So Schmomas and I started to talk about the potential for an account that split modality in two. What would happen if you accounted for all the de re necessity in terms of the essences of things, but then turned to linguistic conventions when it came to things like logical truth and de dicto necessity? This is a leap from a traditional essentialist picture which, as far as I can tell, takes the unconventional (excuse the pun) step of basically treating all de dicto necessity as just de re necessity about propositions or whatnot.
So how about it? It looks pretty weird at first. Especially as it portrays a serious theoretical divide where folk don't often put one, and what's more it places necessity de re as the more real of the two, which bucks against the historical trend. Now that's enough to get some hats spinning, but I think there's something to be said for it. For a while now I've been wondering if modality de re and de dicto are just too different to be covered realistically by the same theory. I remember making some slightly drunken midnight tweets about this a while ago (oh yeah, that's how I roll). Imagine my joy to find that someone else actually had thoughts along the same lines! It seems to me that modality de dicto and de re are really different, more so than most people acknowledge (at least publicly). I'm beginning to wonder if it isn't a significant advantage of a theory if it accounts for the two in radically different ways. Think about it. Necessity de re is about stuff. Could that thing be more like this other thing? Did that thing have to be like that? etc. Necessity de dicto just isn't like that. It's about the way that we reason, the way that we think about the world. The apparent overlap between the two doesn't really seem any more significant than the use of certain bits of vocabulary. Normative claims involve the same kind of vocabulary, and we don't expect to cover those with the same account. Of course there are de dicto claims that seem to overlap a lot more. "It could have been the case that there was a dragon in my bathtub" seems to be talking about stuff (in a way) just as much as "My dragon can't get out of the bathtub!" But let's look at that. Whilst the latter is a modal claim about what's possible of the abilities of my dragon, the former just isn't. It's a case of us reasoning about the world, and about the compatibility of properties. The reason why we think there could have been a dragon in my bathtub is that we think the properties we associate with dragons are consistent, and that they are consistent with the mechanics of bathroom fixtures. The only kind of stuff that the de dicto claims could be about are propositions, logical notions, properties, or the way we think about them.
Oh course, if you want a more unified account, but still don't want to go reifying things willy-nilly, then you could be a conventional essentialist. You ground all modal facts in the essences of things, and you give essences to anything that you want there to be modal facts about. This means just about everything. Numbers, logical forms, the lot. This can go along the lines of the Finean essentialist account of modality. Then you just go and be a conventionalist about neo-Aristotelean essence and Bob's your uncle, you don't have to worry about reifying all those things you just gave essences to because those essences are conventional and don't commit you to much at all. The advantage of this is that you get to talk about modality and think in modal terms without having to worry about any of the spooky stuff that us metaphysicians get bullied for not thinking of as being all that spooky. A potential downside is this doesn't really seem to be the way we think. It hangs on a fair few 'ifs', like the viability of a complete essentialist account of modality and a conventionalist account of essence (both of which I plan to write about in some depth), and also the slightly cringe worthy claim that all de dicto necessities are really just de re ones in disguise.
So there you go. Some ramblings from the pub. Feel free to digest and reject at your leisure.