Sunday, October 19, 2008

Explaining Higher-Order Desires

The unwilling addict has an overwhelming first-order desire for heroin, but wishes that this were otherwise. How are we to explain this second-order desire? Presumably it is not an intrinsic, brute desire. We might think it instrumental: the addict desires good health, etc., and ridding himself of his desire for heroin would help bring about this desired end. But here's a puzzle: in that case his first-order desire for heroin, being stronger than his desire for health, should give rise to an equally overwhelming second-order instrumental desire to be rid of his desire for health (for the sake of the more-desired end of using heroin). As pointed out in class last week, the first-order weightings should replicate themselves all the way up the hierarchy, if higher-order desires are instrumental. After all, instrumental desires are not really additional desires at all; they are simply the pairing of an underlying (intrinsic) desire with a means-end belief. They are in this sense 'transparent' -- mere placeholders for an underlying desire, which provides the entire motivational force (filtered by one's degree of credence in the means-end belief). So understood, instrumentalism cannot explain the emergence of substantive desire hierarchies, where our higher-order desires govern (rather than merely reflect) our first-order ones.

Yet even if higher-order desires cannot be purely instrumental, it seems clear that they depend in some sense on first-order desires. Plausibly, his desire for health is the reason why the unwilling addict is unwilling. There is some explanatory relation here. But what? We might opt for a mere causal-historical explanation: the desire for health happened to give rise (through some non-rational process or psychological quirk?) to a new (intrinsic) 2nd-order desire to be rid of the addiction. But this implies that the desires are now independent, so even if he ceased to care about his health and all that, the addict would retain his brute desire to be rid of the addiction. That doesn't sound right.

Is there a third option? Perhaps the desire for health not only created, but also sustains, the second-order desire. Can we make sense of this proposal without collapsing it into the first failed account (i.e. transparent instrumentalism)? I think we can, on the model of conditional desires. The higher-order desire to be rid of addiction is implicitly conditional on one's desiring of other goods in life (e.g. health). If one loses those other desires, then the higher-order desire is neither thwarted nor satisfied, but simply 'cancelled' -- it no longer applies. Nonetheless, so long as the condition (of one having those other desires) is met, this does count as a further desire in addition, with weight of its own that is not simply inherited from others.

I should note that this still doesn't answer the comparative question why one came to have this higher-order desire rather than some other; but perhaps that is not a fact in need of explanation.* What this account succeeds in explaining is how higher-order desires can in some sense depend upon first-order desires, without merely replicating their weightings (as instrumental desires would, given their 'transparent' nature).

* = Having said all that, I think we would do better yet to abandon strict Humeanism and allow that value judgments can give rise to desires. The stronger higher-order desire to be rid of the addiction thus reflects our cognitive judgment that health (etc.) is better than heroin.

Actually, I'm inclined to deny that the addict really desires heroin (non-instrumentally). It seems more plausible to say he desires to be rid of the unpleasant craving sensation, and either taking heroin or somehow losing the addiction would equally satisfy this negative desire. The difference between 'willing' and 'unwilling' addicts is then not a matter of second-order desire at all, but simply a refinement of the negative first-order desire, by further specifying which of the two possible means is preferred. (Maybe I'm working with an overly intellectualized sense of 'desire' here, but I guess that depends on what you want to do with the concept.)


  1. This seems to be a feature of all cases of akrasia -- when I make an all-things considered judgment that I want to go on a diet, but I still eat the pizza, we can attribute a 2nd-order desire to me to not have the desire to eat the pizza...

    But if we see these as cases of akrasia, then it seems like our behavior need no longer match our strongest desire.

    Even on something close to a straight-up Humean account*, we can make sense of the fact that I think I have more reason to diet than to eat the pizza by saying that either I don't really desire the pizza, or whatever desire I have for the pizza is outweighed by the desires for, e.g., health, attractiveness, etc. that counsel not eating the pizza.

    Then we can treat the higher-order desires as purely instrumental as we please, since the weighing will replicate appropriately: we'll say that I have second-order desires to lose the desire for health as well as to lose the desire for pizza, yet the latter is stronger than the former, just because my desire for health is stronger than my desire for pizza.

    The fact that I eat the pizza, on that story, does not necessitate the conclusion that my pizza-desire is stronger. Just that something else (irrationality, a slipped marble, some non-desire source of motivation) took over my will.

    * I say "close to" because I'm not sure how the Humean account deals with akrasia. Would a proper Humean deny it even exists?

  2. "if we see these as cases of akrasia, then it seems like our behavior need no longer match our strongest desire."

    But akrasia is not when you fail to act on your strongest desire. (The Humean might consider this impossible, as a matter of definition.) It's when you act contrary to your better judgment.

    "whatever desire I have for the pizza is outweighed by the desires for, e.g., health, attractiveness, etc. that counsel not eating the pizza."

    That just doesn't sound like an accurate description of the unwilling pizza-lover. Their problem is precisely that their desire for tasting pizza is stronger than their contrary desires (and they wish it wasn't). [That's what explains their unfortunate behaviour: their giving in to the desire to eat pizza is not some inexplicable freak occurrence.]

    But even if we accept your story here, the higher-order desire structure doesn't do any work, whereas Humeans like Frankfurt think it should, so I'm interested in whether we can make sense of this...

  3. I'm thinking that a Humean account of akrasia might look something like this:
    - If my desire for X is strongest, I have most reason for X. (The claim that makes this account Humean.)
    - Akrasia is acting against what I (judge that I -- let's assume I'm judging accurately here) have most reason to do.
    - Therefore, akrasia is also acting against what I have most desire to do.

    I'm not sure whether this matches the pizza or the heroin cases. Is serving an addiction more like a compulsion? We often think we're doing things that we don't *really* want to do because we can't help ourselves. (Does the person with OCD desire to wash his hands 5 times an hour?)

  4. Ah, I think we're talking about difference senses of 'Humeanism' here. You mention the normative claim that desires are reasons. But I was talking about the claim that action is caused by our beliefs and desires.

    I assume the pizza-lover does act in eating the pizza -- though perhaps you are questioning this when you ask whether it might be "more like a compulsion". That might have more plausibility in strong addiction cases, but I take it nobody is really "addicted" to pizza. This is just ordinary weakness of will: a transitory desire overwhelming one's better judgment.

    "We often think we're doing things that we don't *really* want to do because we can't help ourselves."

    I doubt the pizza-lover would deny that he does actually desire the pizza. The sense in which he "doesn't really want it" is arguably just that he has a higher-order wish not to have this desire in the first place. (At least I take it that's the kind of story Frankfurt would tell.) So the desire - though real enough - is not 'wholehearted'.

    In a case like OCD I'm more inclined towards the line mentioned at the end of my post, i.e. there's an unpleasant sensation they wish to be rid of (and giving in to the impulse is one way to get rid of it).

  5. I don't think a Humean is committed to the claim that all action is caused by our beliefs and desires, just action for reasons -- that is, the explanatory flip side of the normative claim. A Humean who takes on the stronger claim that all actions are caused by desires has to deal with cases like reflexes, and that seems impossible. (I don't have any desire, the satisfaction of which is advanced by kicking when the doctor hits me on the knee.)

    I've written a post with an expanded argument from the Humean perspective here. It's a bit hasty, but I think it points to a way for the Humean to say that pizza isn't a desire. Though ultimately it might come down to the battle of intuitions -- as a frequent akratic pizza eater, I'm not sure I'd describe akratic pizza eating as pursuant to a desire...

  6. Not all behaviours are actions -- I take it the latter must be intentional. Even so, this doesn't entail your normative claim, since one may act intentionally even when there are no good reasons that could justify one's so acting. (People do dumb things all the time.) So I guess you were instead talking about 'motivating' or 'explanatory' reasons.

    But anyway, thanks for the link to your expanded argument. One quick thought: even if you think some cases of apparent akrasia are really compulsions, don't you think there are other cases where we do (voluntarily) act on some strong desire which doesn't match our all-things-considered judgment of what we should do? One may imagine any number of everyday examples of non-compulsive procrastination, indulgence, etc.

  7. True, but are those sorts of cases usually the object of second-order desires? I suppose I could solve my procrastination problem by no longer liking internet chess (or blogs, etc. etc.), but I don't have a desire to get rid of my desire for internet chess -- just a desire to have stronger will to play it less often.

    I suppose it's conceivable that we could have second-order desires in such cases, but I'm skeptical that we actually do so with any frequency.

  8. Do Humeans believe in the 'will'? I'd expect them to insist that to alter your actions we must alter (the strengths of) your desires.

    Fwiw, I'm pretty sure I do desire that I had a stronger desire to write philosophy papers. (That's not to say I wish I had no desire at all to do other things. But I wouldn't mind adjusting the weightings a bit. This more moderate version still qualifies as a higher-order desire, which cannot be explained purely instrumentally.)

  9. I just remembered that Frankfurt talks about higher-order 'volitions' -- what I want is not just that I have a stronger desire to write philosophy, but that this desire is effective, i.e. it leads all the way to action.

  10. Hmm... I'll have to read some more Frankfurt then -- that sounds like exactly the sort of thing I might be interested in incorporating into my own version of the Humean view.


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.)