...plus a preview of forthcoming analyses of a priority and analyticity.
This is a correspondence-inspired followup to 'Two Concepts of Metaphysical Modality'.
What I there called 'metaphysical modality in the broad sense' should probably be thought of as including what David Chalmers calls the logical or epistemic modality (also sometimes the 'indicative' modality), tied to what is a priori consistent. This despite my characterizing it in part with the term 'non-epistemic'.
What's going on? It may look like there is some disagreement between me and Chalmers about whether some modality has some status which we both call 'epistemic'.
I don't think this is so, however. It seems there are just different uses of 'epistemic' at work here. Chalmers doesn't think that what he calls the logical or epistemic modality depends upon our knowledge or conventions (any more than logic does). Whereas I was using 'epistemic' in 'non-epistemic' (and tend to use it elsewhere) to mean something like 'depending on or having to do with knowledge'.
This also connects with a bit in Chalmers's short piece on the tyranny of the subjunctive, where he fields the objection 'But indicative necessity is epistemic!' and says 'So?'.
For me that always seemed like a bullet that really didn't have to be bitten - in fact, on that ought not to have been. But this is probably because I was reading 'epistemic' to mean something like 'depending on, or having to do with, knowledge'. I now suspect I was reading it contrary to Chalmers's intention. One moral: take care using and interpreting overworked philosophical words!
Now for the preview...
For my part, I think the core notion behind the a priori/a posteriori distinction in propositional typology - which distinction is tied to what Chalmers calls the logical, epistemic or indicative modality - isn't really 'epistemic' in the sense that I tend to give that word. 'A priori' could of course be stipulated to be about knowability without experience or whatever, but what underlies and explains that knowability has to do with the nature of the proposition itself.
I'm working on cashing that out in terms of the internal meaning determining the truth value. And that sounds like analyticity, but to that I reply that it's the full internal meaning in question here, which need not be fully grasped in order to pass as understanding the proposition in question. To define analyticity, I want to use a notion of a 'meaning-radical' - something like a bit of conceptual structure such that having it on board suffices for you to count as understanding the proposition in question - and say that a proposition is analytic iff its meaning-radical determines its truth value. Or one of its meaning-radicals, if there can be more than one. (Meaning-radical/analyticity are pretty vague notions, I think.)
This also provides a nice pithy answer to Kant's big question; the synthetic a priori is possible because full internal meaning can outrun meaning-radicals. This connects to the idea of the linguistic division of labour, and Wittgensteinian points about the widely-scatteredness of the determinants of meaning.