The Humean Addict

Must a Humean accept higher-order desires?

Let’s fix some terminology. Say a second-order desire is a desire about a desire. The standard examples of second-order desires are cases where one is dissatisfied with some aspect of one’s character — for example, one is a heroin addict, and wishes not to be one — that is, desires heroin, and desires the elimination of the desire for heroin. Or one could be a glutton, and desire the elimination of one’s desire for food. One could, I suppose, also have positive second-order desires: I could desire to desire to give more to charity. But it’s a little harder to make sense of positive second-order desires (if I desire to desire to give more to charity, isn’t that the same as desiring to give more to charity?). The most convincing examples are about compulsive first-order desires that one would be better off rid of — like Ulysses’s desire to listen to the Sirens, which he avoided by plugging his ears.

A Humean is a theorist who accepts certain strong views about the relationship between practical reason and desires. It’s not totally clear exactly how to formulate those views, but I’ll give it a stab as follows: a Humean needs to take on at least two claims:

1. If I have reason to X, I must have some desire, D, such that Xing contributes to the satisfaction of D.

2. If I think I have reason to X (consciously or unconsciously), I must have some desire, D, such that I think Xing contributes to the satisfaction of D.

Note that the Humean must limit this demand to action for reasons. The Humean can’t claim that all actions are done pursuant to desires, on pain of immediate refutation by things like reflexes (I have no desire to flinch when the optometrist blows the air into my eye).

As Richard points out, certain features of second-order desires require explanation — they seem to give rise to some funny consequences.* This seems to be a particular problem for Humeans, because non-Humeans can always throw out the notion that desires must be based on other desires, all the way down.

But I think everyone, including the Humean, has a simpler answer: to deny that the objects of these “second-order desires” are really desires at all (and thus deny the existence of second-order desires).

We might think that this move is unavailable to the Humean, because the Humean is committed to propositions 1) and 2), and thus to the claim that when the heroin addict takes the heroin, he is doing so pursuant to some desire. That thought is the target of this post: I think the Humean needs accept no such thing.

Here’s the argument in outline. Note that I’m taking advantage of the fact that the most convincing examples of second-order desires are cases of compulsion or addiction.

1) All actions for reasons are intentional actions.

2) For X to be an intentional action, the agent must believe, consciously or unconsciously, that (s)he is Xing, and for X to be an intentional action done for reason R, the agent must believe, consciously or unconsciously, that (s)he is Xing for reason R.

3) When a person is engaging in compulsive behavior, (s)he need not, and often does not, believe that (s)he is doing so for any reason.

4) Compulsive behaviors are not intentional actions for any reason. (from 2, 3)

5) Compulsive behaviors are not actions for reasons (from 4, 1)

6) Even on the Humean account, compulsive behaviors need not be (though they may be) rooted in desires (from 6, Humean claims).

Notes on these claims:

1) is, I think, pretty conventional. I’m a little leery of how such a claim could treat unconscious motivation, but if we accept a notion of unconscious intention as well, then it seems plausible enough.

2) is ripped off loosely from the account in Kieran Setiya’s Reasons Without Rationalism, and he, in turn, took it from Anscombe. There are some problems with this claim (Davidson is sometimes thought to have blown it to bits with a famous example involving carbon paper), but Setiya offers a reasonably plausible defense of it in a modified form (the modifications are not relevant here — they basically amount to weakening it such that the agent has to believe something), again with some reservations that I have about unconscious motivation.

3) is where someone might leave the train. But I think 3) is true: the phenomenology of compulsive behavior isn’t necessarily “I’m taking the heroin because I desire it,” it might well be “why am I taking the heroin? I wish I could stop… why am I doing this?”

I’ve never been any kind of a drug addict, but I have been on diets, and I’m generalizing this claim about the phenomenology of compulsive behavior from the experience of breaking a diet. I’ve sat down at a pizza joint, ordered and consumed pizza, all the while thinking “why am I doing this?” I’ve also believed, while eating the pizza, that I was doing so because of a compulsion to do so. I imagine that drug addicts are even more alienated from their behavior, and that behavior even more compulsive than drug addiction (like the handwashing compulsions of OCD sufferers) are even more so, though this is something that would require empirical research to actually demonstrate.

(Edit: as I was writing this post, Mike put up a comment to a previous post, which offers some confirmation of my intuition about the phenomenology of compulsive behavior.

One might say that in these cases there’s an unconscious belief that one is eating the pizza, shooting the heroin, etc., because one desires it, but that’s incredibly ad hoc and would make the belief condition on intentional action pretty much vacuous.

One might also say that the heroin addict believes that he’s consuming the heroin because he’s addicted to it. But that’s not a belief about a reason. It’s a belief about a brute fact of biology, just as when the optometrist blows air into my eye to test the pupil dilation, I believe I’m flinching because I have a flinch reflex.

* See Richard’s post for details. Briefly, it seems like the desire to get rid of the heroin addiction is an instrumental desire rooted in the desire for health, but, then, how come, if the first-order desire for heroin is stronger than the first-order desire for health, there isn’t a second-order desire to get rid of the desire for health?


7 Responses to “The Humean Addict”

  1. Daniel Goldberg Says:

    One might also say that the heroin addict believes that he’s consuming the heroin because he’s addicted to it. But that’s not a belief about a reason. It’s a belief about a brute fact of biology . . .

    There are strong reasons to doubt that addiction is a “brute fact of biology,” at least in part because it is question-begging to rely on a conception of biology that reduces all “biological facts” to reflexes. That’s a mechanical, automatic view which does not accord with a great deal of things we might take to be biological facts, as Lewontin argues nicely. I talk about some of this briefly here.

    There’s a large literature on the socialization of addiction, which is a part of my dissertation. Caroline Jean Acker has good material on this, as does Marcus Boon and Joe Gabriel.

  2. Paul Gowder Says:

    Thanks Daniel. I’d like to know more about the phenomenology of addiction (though without taking the most direct route, viz., becoming addicted to something myself).

    The issue isn’t really whether we call it a brute fact of biology or not, though, it’s whether a person suffering from an addiction experiences themselves as acting for reasons, in the same sense that I experience myself acting for a reason when I, say, write this comment. I guess a good heuristic (because it’s hard to say exactly what it might mean to believe one is acting for a reason) is whether someone who is servicing an addiction experiences it as something (s)he is doing, or something that is happening to him/her.

  3. Daniel Says:


    The question of reasons is a good one. I guess I’m just not sure whether the addict experiences themselves as acting for reasons — I would tend to say “yes” in most cases, but that’s at least partly a function of my desire to preserve some concept of agency (which is harder than you might think when you are as big a believer in the power of social determinants as I).

  4. Paul Gowder Says:

    Hmm… I guess my sense is that agency is exactly what addicts lack — the whole idea of addiction is that one is driven to do things that one doesn’t really want to do?

  5. Daniel Goldberg Says:

    But they can’t entirely lack agency, otherwise how could the addict ever alter his/her behavior?

  6. Paul Gowder Says:

    Good point. Tough question, too. Perhaps we need sharper tools? This might not be a question that can be resolved by talking about having or not having agency.

  7. Daniel Says:

    Are you questioning the power of my kung fu????

Leave a Comment