David Chalmers

David Chalmers (Australian National University), “Probability and Propositions,” with commentary by David Braun (University of Rochester).  The paper and commentary can be found here.  Chalmers' reply can be found here.

Comments

[uriah]

Very cool paper, David. I wonder if someone couldn't object that, at the end of the day, it a bit odd to say that when I believe that the weather is nice, the object of my belief is an ordered pair of an enriched intension and an extension. Such an ordered pair may be a good way to *model*, or *represent* (in the mathematical sense), the object of my belief, but it cannot literally be *that which is being believed*.

[David Chalmers]

Hi, I take it the same worry arises for any view where the objects of belief are formally defined entities involving sets, sequences, and the like -- e.g. it arises equally for Russellian complexes and for sets of possible worlds. Likewise, for views where numbers and sets are identified with formal constructions. Around here one can easily get into the frame of mind where one wants to say "everything is what it is, and not another thing". On the other hand, it doesn't seem so awful to bite the bullet and say that people can believe structured complexes of enriched intensions (n.b. not ordered pairs of enriched intensions and extensions). If such a theory came to be accepted, this might come to sound no odder than the claim that one is drinking H2O. Speaking for myself, I don't much mind making the marginally weaker claim that such entities model the objects of belief (or that one believes* these entities rather than believing them, where belief is in some strong way by belief*). I'm not certain how to settle the issue, though, or how much turns on it at the end of the day.

[Benj]

Nice paper, Dave. Where in your taxonomy does the following view appear: the hall of mirrors guy assigns credence 1 to the proposition that this ball is that ball, but an intermediate credence to the proposition that, say, the object presented in this experience is the object presented in that experience (similarly, 1 is assigned to the proposition that Tinasky is Hawkins, while intermediate credence is assigned to the person I call 'Tinasky' is the person I call 'Hawkins')?

(A concern about this view would be that the referentialist propositions do no explanatory work for the Bayesian, while all the work is done by the descriptive propositions.)

[David Chalmers]

That strategy is discussed under reply (4), section 2, on propositional surrogates. I say (i) that it's pretty awkward to for a referentialist to say that such descriptive surrogates will always be available, as they often argue for their view via Kripke's epistemological argument against the availability of associated descriptions, and an adaptation of those arguments seems to work equally well in this context, (ii) metalinguistic surrogates aren't always available, and in any case (iii) if one applies this strategy in general, it looks like a view on which the objects of credence are nonreferential.

Incidentally I've just noticed that the version of the paper on the OPC website is an older draft. The current version is on my website at http://consc.net/papers/probability.pdf (this is the version that David Braun commented on). The changes aren't huge, but they make a minor difference in a few places. I'll ask the organizers to update the version on this website.