« Escher Sentences | Main | Linguistic Just-So Stories »



It's no part of my project to eliminate the role of intuition (I think that would be hopeless) or to lend support to empiricism (I'm arguing in favor of modal rationalism, after all). Of course the notion of ideal conceivability in particular is full of normative epistemic notions which I make no effort to eliminate. I don't think the gap between a "conceivability" view and a "modal intuition" view is as big as you suggest: one might see positive conceivability itself as involving a sort of modal intuition. There are other sorts of modal intuition, though, so I think the conceivability account is somewhat more specific about what's going on in the relevant sort of modal reasoning.



Point taken. I intended only to say that the conceivability approach is (at least prima facie) /consistent/ with a form of neo-empiricism. The idea being that empiricists can take on board positive conceivability (understood as a form of imagination) without swinging all the way over to a full-blown modal rationalism. The main point, then, was just to undercut any empiricist hopes of co-opting your proposal for their own dark ends, not to attribute those dark ends to you directly.

But on the critical side. One danger for your view is that, unless you eliminate rational intuition as a distinct capacity, it threatens to push conceivability considerations out altogether. Specifically, if rational intuition can do what positive conceivability is designed to do (plus what it can't do), then parsimony suggests we not bother with conceivability at all (or treat it as merely derivative in some way). This might happen for example in cases where the "imagined" situation outstrips our imaginative capacities, but where we nevertheless have rational intuitions about its possibility. That is not an argument, just a red flag. What extra work is conceivability doing? For my part, I suspect it is doing some ineliminable work, but I don't think I can say what it is with any degree of confidence.


Well, "rational intuition" just seems like a very broad and blunt instrument here. For example, each of the eight varieties of conceivability that I discuss in the paper presumably counts as a sort of rational modal intuition, and there are lots of other sorts of rational modal intuitions. If one's interested in an entailment from epistemic to modal notions (which is the subject of my paper), it seems to me that a blunt appeal to modal intuition won't work, as there are too many misleading modal intuitions. So one needs something more specific -- hence ideal primary positive conceivability, and so on.

I suppose an alternative would be to appeal to "idealized modal intuition", i.e. intuitions about what's possible that stand up to ideal rational scrutiny, but here there's a serious danger of circularity. For example, it seems that differences in philosophical views about the epistemology of modality will ramify into differences in what one takes to an idealized modal intuition. The advantage of doing things in terms of conceivability is that modal notions and modal judgments don't play this direct role.


In the first post you suggest that positive conceivability "involves" modal intuition; but in the second post, you suggest that positive conceivability is a species of modal intuition. These are importantly different proposals and, to my mind at least, the first is far more plausible.

I agree that modal intuition simpliciter is a blunt instrument, as blunt as perception simpliciter. That I saw that p suggests I have good evidence that p unless it is further revealed that my perception occured under unfavorable perceptual circumstances. Similarly, that I intuit that p suggests that I have good evidence that p unless it is further revealed that my intuition occurs in unfavorable cognitive circumstances. However, while we have a pretty good handle on what unfavorable cognitive circumstances come to in the case of perception, it is far less clear what they come to in the case of intuition. Bealer has noted some pretty obvious ones (e.g., lack of focus, too much beer, and so on). As I see it, the value of the work on conceivability lies in giving us a clearer understanding of the conditions under which our intuitions are to be trusted, not a clearer understanding of the intuitions themselves.


I working on a paper right now that is loosly related to this topic. Dave already knows the argument but I'm interested to see what you (Mark) think of it, especially given the passage you cite from Dave. You cite

[T]o positively conceive of a situation is to imagine (in some sense) a specific configuration of objects and properties. ... . When one imagines a situation and reasons about it, the object of one's imagination is often revealed as a situation in which S is the case, for some S. When this is so, we can say that the imagined situation verifies S, and that one has imagined that S. Overall, we can say that S is positively conceivable when one can imagine that S: that is, when one can imagine a situation that verifies S. (Chalmers, "Does conceivability entail possibility?", p. 150.

I offer the following counterexample to thinking of concievability in this way.

An author of fiction knows apriori that the names he creates fail to refer. So, take any fictional sentence n is F. The author of that sentence knows apiori that n is F is not true. So there will be no scenario that verifies n is F, since each scenario will be centered by marking the author. However, it seems pretty clear that n is F is concievable for the author. While Salinger knows apriori that there is no Holden Caulfield. He can still concieve of a concrete Holden having such-and-such properties. In fact, this seems like a natural way to describe what fiction writers do.
Hence, we need to give up defining concieveablility in terms verification. But, the problem this poses for Dave is that verification is, as far as I understand, just a way to model epistemic space in terms of apriority. If my counterexample is sound, then we need to reject defining epistemic space in terms of apriority as well.


Hi Nick,

There are a couple of things to say about your case. First, it is not clear to me that an arbitrary sentence involving a fictional name is false simply because the name fails to refer. For instance, it seems true that Holmes lived on Baker St. Although I don't have a considered opinion on fictional discourse, these sorts of observations suggest that the semantics for fictional names is nonstandard--descriptivist or something like it. (I assume Dave would here want to invoke the primary/secondary distinction.)

Second, unless we take the name "Holden Caulfield" to be descriptive, I doubt that Salinger can conceive of a situation in which Holden Caulfield exists. At most, Salinger can conceive of a situation like the one described in the book in which an individual named "Holden Caulfield" does such and such. On the other hand, if we do take the name to be descriptive, then it is false that we can know a priori that Holden had such and such properties is false (and, even if we can, we can't know a priori that it is necessarily false--which would be the relevant claim).


Hi Mark

Sorry I haven't responded till now but I've been travelling. About your comments. I argee that not every sentence involving a fictional name is false (or, to be more careful, not true). Following Kripke and van Inwagen, one can make a distinction between sentences within a fiction versus sentences outside of a fiction that contian fictional names. However, my point is only concerned with sentneces that occur within a fiction. These fictional sentences are not true, otherwise it wouldn't be a work of fiction. Now consider any arbitrary sentnece within a fiction (n is F). Now do a modal profile of the sentence using scenarios centered on the author of n is F. Since the author knows a priori that n is non-referring, or at least knows a priori that n is F is not true, n is F will not be verified at any scenario.

It should follow, given Chalmers's framework, that n is F is inconceivable at all scenarios centered on our author. However, without giving an argument here why it is in fact conceivable. I'll just state that if it is concievable (for our author), then Chalmers is in trouble. We have good reason to suppose it is concievable. Therefore, we have good reason to suppose that Chalmers's framework is in trouble.

Lastly, fictional names cannot be descripitive in an epistemic 2-D framework, because in both the primary and secondary dimensions of meaning a descriptive intension would determine actual concrete objects. Following Kripke though there cannot be a concrete Sherlock. And if I am correct about Chalmers view, if it is a priori that n is non-referring, then n cannot refer in any scenario.

Thanks for the comments.

The comments to this entry are closed.

August 2011

Sun Mon Tue Wed Thu Fri Sat
  1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30 31      
Blog powered by Typepad