Monday, May 13, 2013

Locating the benefit of future-directed desires

Inspired by Princeton's recent (and very interesting) "Workshop on Well-being", I've been thinking about the debate between Bradley and Dorsey over the question: Where in time should Desire Theorists "locate" the benefit of a satisfied future-directed desire? At the time (t1) of the person's having the desire, or the later time (t2) where the object of the desire, i.e. the desired "good", is located?

I'm not sold on the idea that harms and benefits must be temporally located at all.  But supposing we do want to locate the benefit here, the former option -- locating it at the time of the state of desire -- seems to me to make most sense.  This may be illuminated by translating welfare talk into 'for the sake of' talk.

Let's say that (i) S at t1 desires that P obtains at t2, and (ii) P does indeed obtain at t2.  According to the desire theory in question, P's obtaining at t2 is good for S, in virtue of S's desire at t1.  The question is: is P's obtaining at t2 good for S at t1, or at t2?  Or, better: is P's obtaining at t2 desirable for the sake of S-at-t1, or for S-at-t2?

This re-statement helps to bring out that, in assigning "temporal locations" to harms and benefits,  we are not thereby treating them as some kind of ghostly substance that itself has a kind of physical presence in space-time.  Rather, it is just to pin down more precisely the recipient of the harm or benefit.  It is not just S-as-a-whole that is harmed or benefited (we are supposing), but specifically the momentary timeslice that is S-at-t1, or S-at-t2.  And when seen in these terms, it's clear that the relevant recipient is the timeslice doing the desiring, not the timeslice (if there even is one) that happens to be temporally co-existent with the previously (but no longer) desired object.

So this suggests that Desire Theorists who want to count future-directed desires should reject Bradley's Internalism ("How well things are going for a person at a time is wholly determined by the things happening at that time that are intrinsically good or bad for that person.") in favour of Dorsey's Time of Desire View ("If S desires at t1 that P, S benefits at t1 iff P obtains at any time").


  1. Richard,

    I'm not sure if this is as clear as you were suggesting (but perhaps I misunderstood your argument!). It seems that you are saying that the desiring at S-at-t1 is the intended recipient of the benefit of the satisfied desire, rather than S-at-t2, when the satisfied object actually is concurrent with a now previously held desire . But if this is true, then desires that become fulfilled at a t2 where there is also an opposite desire than the one at t1 possess benefit and harm for two different timeslices. While possible, this outcome does not seem "desirable" (see what I did there?), due to the complications this causes when deciding which S deserves concern (Do we weigh the desires of S-at-t1 or the desires S-at-t2?. It seems this could be avoided (and also avoid the problem of S-at-t2's benefit not referring to any held desire) by only ascribing a benefit to the time when the desire to be satisfied is present at the timeslice before the actual satisfaction. Is this a bad interpretation of your argument?

    1. Hi Korey, right, on my preferred understanding of the Desire View, if we add that S at t0 desires that not-P, then P's obtaining at t2 is good for S-at-t1 but bad for S-at-t0, so there's a real trade-off there. That seems to me the right thing to say about the case (at least from the perspective of a desire theorist who wants to count future-directed desires at all).

  2. [Dale Dorsey writes:]

    Hi Richard,

    I think I generally agree with you that it's fine to say that welfare is not temporally located. But I think your argument here is a good way to reject the time-of-object view in favor of the time-of-desire view, if one's a DS theorist. The worry, I think, is an alternative, viz., concurrentism. Concurrentism holds (actually, this is the view I call in my paper "concurrentism lite", but ignore this) that x is good for A at t1 if and only if A desires x at t1 and x occurs at t1. Basically it's a restriction on temporal welfare: you're only better-off at a time when you desire something at that time and the thing occurs at the time. And so this view could accept Internalism and the "for the sake of" point you make, plausibly in my view. I suppose I think concurrentism lite fails just for the sort of space-time asymmetry reasons I was interested in pressing, but not all find that plausible.


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.)

Note: only a member of this blog may post a comment.