Ever come across an argument that you just know can’t be right but you can’t pin down its flaw? I had that experience recently while reading an older paper on punishment, Alan Wertheimer’s “Deterrence and retribution.” (Ethics 86 (1976): 181-199) Quick background: Retributivists about criminal punishment often criticize consequentialist views on punishment, especially those that hold that deterrence is the aim of punishing an individual, on the grounds that such views would permit the “punishment” of those who are not responsible for criminal wrongdoing. If. e.g, if children, the insane, and others lacking in mens rea could be deterred from crime by the credible threat of punishment, why not punish these individuals, despite their lack of responsibility for their criminal acts? Indeed, retributivists ask, if we could deter individuals by occasionally punishing those who lack actus reus (those whose actions do not even meet the behavioral standards for criminal conduct), why not punish those who haven’t even engaged in criminal conduct? Wertheimer’s argument aims to defang this criticism a bit:
“…to say that an agent deserves to be punished entails that the agent is responsible for his behavior. And to say that an agent is responsible is to say that he is, in principle, deterrable. Consider the case of Jones — a pure kleptomaniac. When placed in certain situations, Jones will steal, even if the legendary policeman is standing next to him. No punishment which has been or could be administered to Jones will have any effect on his behavior. I take it that although we might want to isolate or quarantine Jones, because his behavior violated our rights, we would not say that Jones deserves to be punished. … Now assume that all stealing is performed by persons like Jones. There are some persons who do not steal and who would not steal even if there were no punishments for stealing. There are other persons who do steal and who would steal even if it were certain that they would be punished for stealing. Under such conditions, those who do steal do not deserve to be punished for the same reason that Jones did not deserve to be punished — they lack the responsibility or deterrability necessary to the justification of their punishment. …With regard to deterrence, [retributivism] and [utilitarianism] are extensionally equivalent if not intensionally equivalent.” (183-84)
In effect, Wertheimer’s argument replies to the retributivist criticism by claiming that there are no individuals as imagined in the criticism: individuals who are not responsible for their wrongdoing, but who can be deterred, or vice versa. So the class of the deterrable and the responsible is the same class of individuals.
For reasons I can’t identify, I find this argument troubling. Retributivists can of course reply to Wertheimer’s argument on conceptual grounds, by pointing out that this move leaves the link between wrongdoing and deserved punishment a contingent matter. And Wertheimer doesn’t elaborate on why the classes would be co-extensive. I’d suggest something like reasons-responsivness (or the lack thereof). Presumably, the threat of punishment gives an agent a reason not to engage in crime, so an agent who is not deterrable in principle is not responsive to the reasons that punishment provides. Likewise, perhaps being responsive to reasons is a necessary condition for responsibility as well.
In any event, I’m not sure what I find misguided about this argument. I suspect it has something to do with how we might distinguish those who are simply not deterred by punishment from those who are not deterrable in principle. Any ideas?
Wertheimer assumes that “we would not say that Jones [a pure kleptomaniac] deserves to be punished,” for Jones isn’t deterrable. This assumption is, I believe, false. I think that whether we think that he deserves to be punished depends on whether or not he is a willing kleptomaniac. If he’s someone who wishes he could stop stealing but just can’t control himself, then he doesn’t deserve to be punished — that is, provided he’s not responsible for having acquired his compulsion. But if he endorses his thieving ways, then we would, I believe, think that he deserves to be punished even if it’s true that he couldn’t stop himself if he wanted to. So the class of the deterrable and the class of responsible are not coextensive. Jones, the willing kleptomaniac, is morally responsible for his act of thievery even though he is not deterrable.
Let me add a thought to Doug’s comment. One conventional way of thinking about punishment is that it has an expressive function (this point was initially stressed by Joel Feinberg, I think). So, as, e.g., CL Ten points out in his “Fantastic Counterexamples and the Utilitarian Theory,” we might distinguish between punishment, on the one hand, and quarantine and other non-punitive forms of imposed unpleasantness, on the other, by holding that cases of punishment are those where we want to express disapproval towards the commission of a wrong (or illegal) action. This, of course, is wholly independent of the question of the responsibility or deterrability of the person convicted (and, to the consternation of consequentialists about punishment, it is independent of the guilt of the person convicted, which might generate some of those fantastic counterexamples if our performing such expressive acts is good enough to outweigh the bad generated by occasionally convicting the innocent in service of such expression).
Thanks folks. Now that I think it over, the problem that strikes me is this: The best enthymemic sense I can make of Wertheimer’s argument is that it locates a common necessary condition for criminal responsibility and deterrability in principle, namely, something like responsiveness to reasons. But I think there might be interesting examples of the criminally non-responsible who are not deterrable (I’ll leave out ‘in principle’ for the moment) because while their actions are responsive to reasons, they respond only to an extremely limited set of such reasons.
Let’s suppose Hitler was undeterrable in that nothing that could have been credibly threatened against him by the U.S., its allies, etc. would have dissuaded him from attempting to achieve Nazi domination of Europe. This need not entail, however, that he is not deterrable *in principle*, if that is taken to imply that his actions are not rationally grounded or that he is unresponsive to reasons altogether. The strongest reason Hitler recognized, let’s say, was the aim of European domination, and he understood all reasons that spoke against his efforts to dominate Europe as miniscule in normative comparison. Now Hitler was not unresponsive to reasons, but he was not actually deterrable. His incapacity to be persuaded not to try and dominate Europe was not due to his arationality, but due to his rather peculiar economy of practical reasons. Indeed, this seems to me what characterizes the fanatic: being under the grip of one all-encompassing reason. But in any event, his non-deterrability is not a conceptual truth about his rationality; there is just no actual reason that could deter him.
Now this raises an interesting question about sanity and responsibility. For it seems to me that we often want to say that the insane and irresponsible are unresponsive to reasons. But might not some cases of irresponsibility be like what I have attributed (whether correctly or not, I’m not in a position to say) to Hitler, cases where a person is guided by a small set of alarming or bizarre considerations that they count as reasons. Imagine, for example, someone who compulsively feels the need to immerse her elbow in a bucket of mud and professes deep conviction that this is a great and wonderful thing to do, and no threats of shame, embarassment, punishment, etc., can convinve her not to do this. This strikes me as a behavior for which we should not hold her responsible. What that might suggest is that some instances of non-responsibility involve substantive defects in rationality (i.e., taking certain non-reasons as reasons) rather than systemic defects of arationality.