Thursday, August 13, 2009

Is addiction voluntary?

I'm pretty fascinated by this new book on addiction. The author (Gene Hyman), a psychologist and professor at BC (I think), argues that addiction is voluntary. I don't have anything profound to say about the matter, but I think that his framing of the issue is interesting.
He begins by pointing out that most addicts are ex-addicts, meaning that roughly 75% of addicts have stopped being addicts, and that among these, most have done so without clinical treatment. So it seems that most people who become addicts stop on their own. And why have they stopped? Heyman's answer: because continuing the addiction became too costly. Whether for financial, health or family reasons, addicts tend to quit when, to put it bluntly, it just ain't worth it anymore.

And yet, Heyman admits that addictive behavior is compulsive. So why call addiction voluntary? Because addictive behavior seems to be for the most part sensitive to the sorts of incentives that guide everyday choices. And this is just what Heyman means by 'voluntary': an action is voluntary not by virtue of its cause (whether by an unfettered will or by some antecedent brain state), but by virtue of its sensitivity to incentives. Schizophrenia, Alzheimer's, bipolar--these are utterly insensitive to costs and incentives, and are thus diseases. But addiction, Heyman argues, for the most part is sensitive to incentives and so is voluntary. Thus, it is not a disease. This conclusion seems to follow even if we acknowledge the truth that tendencies for addiction are heritable. (Many critics of Heyman seem to ignore this point, not realizing that his definition of 'voluntary' is consistent with voluntary behavior being both heritable and a function of the brain).

So my first question: Why can't diseases sometimes be voluntary? Here's something a philosopher might say: addiction is voluntary, but not free. Addictive behavior is the effect of first-order decisions, which, like all voluntary decisions (which is to say, decisions per se), are sensitive to costs and benefits and follow a preference structure (utility function). But the addict, usually, is not very happy with the choices s/he makes, and to this extent, is not free. Can we allow voluntary but unfree diseases? If so, would this neutralize the annoyance many people feel towards Heyman's thesis? (Incidentally, we might already have a name for such 'voluntary diseases,' viz., character flaw.)

A second question: Why does the label matter so much? There has been a lot of mean ink directed towards Heyman, often from recovering addicts. One obvious reason is that we as a community have decided that we don't want to blame addicts for their actions. It's important that we continue to refer to addictive behavior as a disease because it's important that we inoculate the addict against moral opprobrium and judgment. But why is that important? Because it's not effective. Expressing moral disapproval is not very likely change an addicts behavior.

And so a third question: What kind of costly, self-harming practices wouldn't be better influenced by treating them more like addictions than like moral choices? Deceit, thievery, infidelity, cowardice, exploitation--these sorts of behaviors are almost always in the end self-destructive, and so why not treat them also like diseases? The answer surely has something to do with the fact that by labeling something as a disease we undercut our right to moral indignation, which is usually a peculiarly pleasurable feeling, and therefore one which we will usually protect.

UPDATE: This is amusing. From Sunday's NYT Magazine:
"Our national weight problem brings huge costs, both medical and economic. Yet our anti-obesity efforts have none of the urgency of our antismoking efforts. “We should declare obesity a disease and say we’re going to help you get over it,” Cosgrove said.
Indeed. This article tells a story about the efforts to label alcoholism as a disease. Apparently, after the repeal of prohibition, a predictable surge in alcoholism followed. But alcoholics were told that this was a moral failing, and so was accompanied by a surge in just the same sorts of moralizing forces that had been responsible for prohibition in the first place. The doctors on the Research Council on the Problem of Alcohol needed to get convince the people to send alcoholics their way, and realized that calling it a disease was a good way to accomplish that. Notice that Dr. Cosgrove (from the Cleveland Clinic) in the quote above is advancing the same sort of reasoning: it's not that we care particularly whether or not alcoholism/obesity is a disease, we just observe that by calling it a disease we help to effect better results.


  1. I haven't read the book, but I will. Three points just on your reflections:

    1. Is it possible that not all addicts are the same? Specifically, that ex-addicts differ in some neurochemical way from continuing addicts? If so, this would also help to distinguish the recovery rates for different drugs, i.e. alcohol v. meth.

    2. I've read the claim repeatedly that 'blame' is not effective in treatment. I find it baffling. I mean, I know what you're getting at: it's senseless to 'morally blame' someone for something they have done if they couldn't have done otherwise. That's like blaming a rock for tripping you. This is also why we see abusive childhoods and mental retardation as mitigating circumstances: in a sense, such people have been robbed of agency by forces beyond their control. Like the rock, we see them from more as objects than subjects, and indeed their actions make more sense as a set of chemical reactions than as the choices of a rational and autonomous being: the lack simple predictive capacities and the impulse control that we associate with maturity.

    However, I do think a species of 'amoral' blame applies and is helpful for thinking about these issues. Specifically, that nobody else can be held accountable for the alcoholic's addicition, and the only solution is to take this irresponsible and unblameable subject and turn him into a responsible and blameable one. It's not about whether the act was optional, nor about a metaphysical account of freedom and agency. Rather, it's simply a matter of attribution. That's what the 12-step programs do, forcing the addict to think of herself as 'the one who did these things.' The way alcoholics do this is by 'taking responsibility' for their past actions, by willingly submitting themselves to blame even though they couldn't have, in the past, done otherwise. Then, they take this newfound responsibility and apply it to present decisions and future choices. I think of it as an 'educative' rather than 'merit-based' account of praise and blame. It seems that addiction recovery is the process by which blameless subjects come to regard themselves as blameworthy.

    3. Thinking of character flaws and vices as diseases seems to be most helpful when it's paired with an Aristotelian psychology in which we recognize a distinction between knowledge and willpower. The akratic subject needs more than just general knowledge about her lack of self-control, she needs to develop strategies in the face of rashness and incontinence. So this would be possible for cowardice or infidelity, but not for thievery or deceit unless it was literally compulsive kleptomania or pathological lying.

  2. Yo Josh. So, I had written up pretty longish replies, and then my computer shut down without warning and I lost it all. So here are shorter replies:

    re: (1) What I like about Heyman's framework is the meaning he has given to 'voluntary.' He uses a concept of voluntary action that has no need to evoke an agent. Anything that is sensitive to incentives is voluntary, whether that be brain states or noumenal selves or identity-poles and so on. So yes, he grants that of course there are neural differences most likely among addicts, but however widely those differences vary, so long as on either end of the spectrum we are still talking about cost-sensitive behavior, we are talking a voluntary rather than involuntary disease.

    re: (2) Yea, I was neither rejecting nor endorsing the reasoning, just saying what I take that reasoning to be. And I like the way you're trying to square this circle. The attitude that comes across in some accounts I've read is that addicts accept responsibility for whatever past harms they've done, not by identifying with them in the normal way, but in the way that an older brother might accept responsibility for something a younger sibling has done, or in the way an officer accepts responsibility for the costly behavior of a subordinate.

    re: (3) Point taken, but I did have something like the klepto and pathological liar in mind. These, too, I imagine are spectrum disorders, implying that most liars and thieves lie and rob not as random, one-off acts, but through some feature of their character.

  3. 1) Clearly, voluntary should not be defined dichotomously. Some things are more voluntary than others. Through exercise, drugs, and/or medication, I can control my heart rate, but it's very costly for me. Most people would agree that the word "voluntary" shouldn't apply.

    Interpreting Heyman's definition as allowing this, the obvious empirical measure of Heyman's concept is elasticity: the ratio of % change in quantity consumed to % change in price.

    For drugs, this ratio is very low (in absoulte value)--Large increases in the price don't change quantity consumed very much--but it's not 0--there's still some effect.

    2) "The answer surely has something to do with the fact that by labeling something as a disease we undercut or right to moral indignation..."

    Carrying this logic forward, a likely reason that drugs rather than, say, thievery or infidelity have gotten the disease makeover is that there's more demand for rehabilitation with drugs. After alienating your friends and family
    on a coke binge, you'll pay good money to get back in good with them, and they probably don't value being indignant toward you as much. With, say, thievery or infidelity, though, it's less likely for this to be the case. Thieves rarely have any incentive to continue a long-term relationship with those from whome they've stolen, and people who cheat on their spouses usually want to leave them for their new lovers, so there's not as much benefit to anyone in changing social norms toward these behaviors (changing norms is a costly enterprise). Niether ther victims nor the perpetrators of these acts benefit as much from the disease framework.


  4. re: (1) yea, elasticity is exactly the right way to go here. Basically, Heyman is arguing that drug consumption is much more elastic than has been assumed. Make the opportunity costs of drug use higher--and the benefits of avoiding that next hit higher--and people have a tendency to kick the addiction.

    As to wider issue of what to count as voluntary, I also agree that there are lot of blurred lines remaining even if we adopt an elasticity approach and stop worrying about a possible agent. My hunch right now is that we might want to distinguish voluntary from nonvoluntary action according to whether such action has a direct or indirect effect on its intended object. So, it's true that I can have an effect on my heart rate, but only indirectly, but practicing meditation, repetition, etc. By contrast, I can directly affect the position of my arm, by deciding to move it. Of course this distinction itself can be problematized....

    re: (2) Sure, that's probably part of it. But people with no dog in the fight (people who don't know any addicts directly, don't work with addicts, etc) are often as indignantly wed to definition of addiction as a disease as anyone else. I think that they perceive a gross moral mistake: to blame someone for heroin addiction, for example, is like blaming someone for cancer.

  5. You guys need to institute an e-mail subscription to further comments.

    Anyway, vis-a-vis #1, I still suspect that some addicts are so neuro-abnormal that they have massively inelastic demand, at best only accepting close substitutions. But that's an empirical question. Certainly this framework does a lot of interesting work problematizing common sense frames of reference in addiction.

    re #3: Even if most liars and cheaters have character flaws (i.e. neuro-abnormalities) it still seems like they'd only benefit from having those flaws described as a disease if there's a treatment available. For drug use, some group therapeutic treatments (12 steps, etc.) seem to work only for one class of addicts, but not for others. It seems like the people who suffer worst from the disease-model are those for whom available treatments aren't curative. With lying et al, I'm not sure if there are *any* treatment models, so as a class they may be much worse off under a disease model. Contrast this with obesity, for which perhaps the only effective treatments are surgical and pharmacological, and I think a pattern starts to emerge. Anything that can be treated by doctors is a disease, regardless of its causation.