Saturday, October 27, 2007

De re Belief

De dicto belief ascription: (1) Jones believes that the tallest spy is a spy.

De re belief ascription: (2) Jones believes, of the tallest spy, that he is a spy.

I find the latter puzzling. Does it commit us to haecceities, or deep facts about the identities of things? Hopefully we may account for the phenomenon without such metaphysical cost. Here's the vital question: what must the world be like in order to satisfy the ascription made in #2?

I don't think it requires any mysterious identity facts. One platitude about de re belief is that its object must actually exist. (I can have de dicto beliefs about unicorns, but there is no unicorn such that I have beliefs about it.) Further, de re belief requires some kind of causal acquaintance (however indirect) with its object. So perhaps #2 is just a convenient way to describe the conjunctive fact that Jones believes (de dicto) that the person he bears such-and-such relation to is a spy, and Jones does indeed bear such-and-such relation to someone (who is now counted as the object of the belief when we ascribe it in de re terms).

(3) Jones believes that Smith is the tallest spy.

Unlike in the previous case, this seems to imply the corresponding de re ascription:
(4) Smith is believed by Jones to be the tallest spy.

Jackson accounts for this via his descriptivist analysis of proper names. The representational content of 'Smith' (for Jones) is just some unique associated descriptive property that builds in the aforementioned acquaintance relation, and thus licenses "exportation" to de re ascription.

Does that do the trick, or is there more to de re belief than captured above? (Do some philosophers deny that de re belief is reducible in this way to the de dicto?)


  1. Richard,

    Just a quick question...

    You asked us to consider:
    (3) Jones believes that Smith is the tallest spy.
    You said that this seems to imply:
    (4) Smith is believed by Jones to be the tallest spy.

    One thing worries me.

    First, I don't 'feel' the pull of saying that (4) is a consequence of (3). Consider:
    (3+) Jones believes that Smith is the tallest spy, but that's because he does not realize that Smith is a fictional character.
    It seems that I'd not want to assert (4) and (3+), but if (3+) is consistent with (3), (4) doesn't follow from (3).

  2. Hmm, nice point. (3) is consistent with the failure of the additional requirements for de re belief ascription, e.g. that the target actually exists. (This probably supports my more general point, I think.) So (4) only follows when those additional conditions obtain.

  3. I don't think it can be as simple as you suggest for familiar Kripkean reasons. Jones may not have any belief about the relation he bears to such-and-such or his belief about that relation may be wrong. I've heard of Godel and have some non-causal beliefs about him but I do not know when I first learnt about him or who from.

  4. Causal descriptions are still descriptions. (E.g. perhaps you identify Godel as whoever stands at the end of the causal chain from your tokening of 'Godel', back through whoever you picked the term up from, etc. That would do.)

  5. Richard,

    Why think that some descriptions are exportable? For any description, it seems to me, there is a difference between:

    (i) believing that the D is F,
    (ii) believing of the D that it is F.

    One cannot in general tell by introspection whether there belief is exportable. Whether their de dicto beliefs ground de re beliefs depends on whether some EXTERNAL conditions are satisfied.

    But if you are right, it seems to me, there would be kinds of descriptions for which it would be an immediate consequence of having the de dicto belief that you also have the de re belief. I don't think there are. Do you have any examples of these kind of exportable descriptions?

  6. Jack, right, I mean to reduce de re belief to the conjunction of:
    (1) the right kind of de dicto belief
    (2) an external condition of causal acquaintance.

    I don't think the belief alone suffices.

  7. You miss the point. Maybe I don't have any causal beliefs about Godel or maybe they are wrong. Maybe I think Godel is Schmidt and so Godel refers to whoever is at the end of the Schmidt chain or maybe I've never thought about the causal theory of reference, or maybe I first picked up Godel from someone who misused the name and yet I still managed to pick up info about the real Godel.

    I can be related to Godel in some way without that relation being transparent.

  8. How is this a problem for my view?

  9. Its a problem because you want to reduce a de re belief to a de dicto

    "perhaps #2 is just a convenient way to describe the conjunctive fact that Jones believes (de dicto) that the person he bears such-and-such relation to is a spy, and Jones does indeed bear such-and-such relation to someone"

    But if the descriptive information Jones associates with X actually picks out someone else your conjunction is not satisfied. Similarly if he has no descriptive belief about X. But do you want to say Jones has no belief about X in such cases?

    Just as we refer to Godel even if we falsely believe he rather than Schmidt was the originator of incompleteness theorems we can have beliefs about Godel even if we have false beliefs about him.

  10. Right, so it looks like there are a couple of issues here:

    (1) We may relax the de dicto belief requirement. I then reduce de re belief to (a) a de dicto belief of the right kind (leaving it somewhat vague for now precisely what, if anything, is required here); and (b) the external condition of acquaintance. That'll work, right?

    (2) General objections to descriptivism. These strike me as misguided. What our response to the Schmidt scenario tells us is simply that 'the originator of the incompleteness theorems' is not the description we associate with the name 'Godel'. The actual associated description is whatever tracks our intuitive judgments across all such scenarios one might describe. Plausibly, it's some kind of causal description. Anyhow, there's obviously some description or other in the vicinity, since the very methodology we use here is to present us with a description (of a scenario) and invite us to say which of the described people we thereby associate with which names.

    Now, if someone really had no (correct) descriptive beliefs about X (even, e.g., "The person everyone else calls 'X'"), they would presumably be unable to say who 'X' refers to in any given scenario. Such a situation seems (a) rare; and (b) one in which the speaker doesn't succeed in referring at all.

    To sum up: the Kripkean scenario method cannot possibly refute descriptivism. At most, it can remind us of some specific beliefs that we don't essentially associate with a name after all (as with Godel and his theorem). But there must be some associated description or other, since we are able to pick Godel out when given nothing but a description of the scenario.

  11. Richard -- first, thanks for sponsoring this interesting site. Concerning de re belief, I generally agree with your (October 31) conclusions, although I have a question about your (1)(b) ("external condition of acquaintance") that I'll get back to in a minute. However, I've been wondering if your characterization of the example (3) "Jones believes that Smith is the tallest spy" is completely correct. (3) could be interpreted as (a1) "Jones believes that someone named 'Smith' is the tallest spy" or, perhaps (a2) "Jones believes: 'Smith is the tallest spy'. While we can use (3) to mean (a1) or (a2), we generally don't -- getting (3) across in the sense of either (a1) or (a2) requires some extra contextual cues in most cases. Leaving (a1) and (a2) aside, then, the natural interpretation for (3) is de re, with the proper name 'Smith' being used by the speaker to make a reference. The reason is that that's how proper names are normally used -- to make make a reference and not to convey descriptive content (which is not to say that there is absolutely no descriptive content in a name). Now, I agree that we allow speaker's reference to 'Smith' in (3) to describe Jones' belief only if we think that Jones has some belief about an individual in (descriptive) terms that uniquely apply to Smith. However, I don't think that Jones has to associate these descriptive terms with 'Smith', as you seem to have suggested in your initial posting. We may correctly use a sentence like (3) in cases where Jones does not know that Smith is named 'Smith', or in which we have no idea whether or not he does. Indeed, this is one of the primary reasons why we use de re, because we think Jones has a belief about a specific individual but we don't know the exact descriptive terms in which Jones thinks of Smith. The usual reason for this is that we think that Jones has a rich mental 'story' about Smith, but one that he has not articulated to us in way that allows us to pick a uniquely identifying term to use in attributing a de dicto belief (or, perhaps, in some cases, because we know too much about the 'story' and have no reasonable way to single out one among the many available descriptive terms that Jones has in mind). This leads to a broader point about proper names, which is that we normally bestow them on individuals (or objects) for which we have -- or expect to have -- a rich 'story', to save us the trouble of having to decide and agree each time we refer to the individual on exactly which term of reference to use. In the case of existing individuals, the 'story' -- and the denotation of the proper name -- of course has a causal link to the individual named, a link which we can exploit if necessary to clear up confusions about denotation. However, as an earlier commenter noted (Clayton, I believe), proper names can also denote fictional or mythical characters -- Rhet Butler, Santa Claus, etc. -- and we can use them to refer to these characters in expressing de re beliefs as well. We use proper names in these fictional cases because -- at least in the right context -- they are also associated with a 'story' that will be familiar to th speaker and audience, and can be presumed to be familiar to a person whose beliefs we're describing (Jones) as well. However, when we use a proper name like 'Santa Claus' to express a de re belief of Jones -- as we certainly may -- we are not committing ourselves either to the existence of this individual or to a causal link between Jones (the believer) and Santa Claus. So, going back to the issue of (1)(b), while an "external condition of acquaintance" may be a necessary condition for de re beliefs about existing entities, it may not be fundamental to de re belief per se.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)