« Our student veterans | Main | A Good Week for Civil Liberties and (Maybe) Old-Fashioned Liberalism »

Wednesday, November 11, 2015

The Fungibility of Intentional and Unintentional Punishment

In my prior post, I argued that punishment theorists often speak of punishment in a narrow sense that only applies to intentional inflictions while people more generally tend to think of punishment in a broader sense that includes not only intentional inflictions but others that are foreseen (and maybe even just foreseeable). Much is at stake here because if retributivists only attempts to justify intentional inflictions, they will fail to justify anything like our actual punishment practices which include lots of harm that are foreseen but are arguably not intended as punishment (e.g., the harms to offenders and their families from being deprived of each other; the reduction of First Amendment rights when imprisoned, the emotional distress of confinement, etc.).

Alec Walen, in his helpful and interesting entry on retributivism in the Stanford Encyclopedia of Philosophy, tries to fill the gap. He offers what I think of as a non-punishment shadow theory to justify aspects of our punishment practices not directly addressed by the retributivist justification of punishment. To my claim that retributivists fail to justify punishment to the extent that they fail to justify the varied emotional suffering prisoners experience, Walen writes:

[E]ven unintended differences in suffering are morally significant. But they can justifiably be caused if (a) the punishment that leads to them is itself deserved, (b) the importance of giving wrongdoers what they deserve is sufficiently high, and (c) the problems with eliminating the unintended differences in experienced suffering are too great to be overcome.

Kudos to Walen for acknowledging that (non-accidental) unintended inflictions of harm require justification. That's a point I've been emphasizing for a while. Some retributivists (see p. 24 here) would like to say that they simply need not justify side effects of punishment because these side effects are not punishment. But all of our actual punishment practices involve both intentional and unintentional harms. Carving off intentional inflictions of punishment from the broad notion of punishment leaves retributivist discussion cut-off from real-world punishment practices.

I'm afraid, though, that Walen says too little to defend his three-part test. He simply asserts it (perhaps confident in its double-effect style reasoning). His (a) and (b) essentially say that if the value of retributive punishment is high enough, then it justifies punishment side effects. But this is exactly the claim I've been calling on retributivists to justify and explain in more detail. Here it's just an assertion and not clear that the condition is ever satisfied. Moreover, it provides no affirmative reason to inflict side effect harms. On Walen's view, retributivism offers no justification for side-effect harms except to the extent that they are needed to inflict intentional harms. And these are serious harms. Imagine if we put school children in an environment with a high risk of sexual assault. Outrageous! Yet that's what we do with prisoners. So the justification of the side-effect harm has to be quite strong. It can't be "used up" by the fact that we've already relied on desert to justify delivering proportional punishment. Moreover, can we not imagine non-incarcerative methods of punishment with fewer side effect harms? (To the extent retributivists support incarceration, it's awfully convenient for them that this method of giving people what they deserve also happens to incapacitate the dangerous.) 

Two further points: First, Walen's view seems to accord with my own claims that we need to measure the subjective experience of punishment, at least in some respects. How can we be confident that the value of retributive punishment exceeds the side effect harms if we don't measure those harms?

Second, Walen is saying that we have affirmative reasons to impose the intentional inflictions of punishment and permission to impose side effects. I wonder, however, why we don't have to adjust the purposeful inflictions to accommodate the side effects. If A and B are equally blameworthy but A will experience his confinement much more severely, why incarcerate A and B for the same period of time if there is an easy method of making their total harm more equal?

My point is easiest to understand when the units of intentional infliction of harm are the same as the units of side-effect harms. Imagine a futuristic method of punishment. Rather than incarcerating offenders, we spray them with "gravitons" that limit their liberty by slowing them down. Future retributivists have solved problems of proportionality and simply look up an offense, say 100 units of crime seriousness, and then set their guns to 100 gravitons so that the intentional infliction of punishment precisely matches offense seriousness.

There is a catch, however. Graviton guns fire 15 extra units 98% of the time. So setting the gun for 100 units will typically spray 115 units. If the value of retribution is significantly high in some case, Walen seems committed to the view that you can set the gun to 100 and fire away, almost certainly leading to someone receiving 115 gravitons total. I think most of us would say, and perhaps Walen would agree, that you have to set the gun to 85 units to achieve the ultimate 100 units. But notice that doing so falls short of the goal of intentionally inflicting 100 units of punishment. (One might quibble about what your intentions really are if you set the gun to 100, given that it fires in excess so frequently. But note this is not an unrealistic assumption. We sentence people to deprivations of liberty in prison knowing that they will suffer side effect harms with probability greater than 98%.)

In any event, if you agree that it would be better to set the gun to 85 units, then don't we have to shorten prison sentences to accommodate harms that we inflict as side effects? This is where standard double-effect reasoning may break down. The intentional portion of sentences can be titrated up and down to make up for foreseen side effects. So the crux of the debate may turn on how fungible the intentional and unintentional harms of punishment are and how strong the obligation is to avoid side-effect harms. Walen offers some comments in his piece suggesting that side-effect harms are not fungible with intentional inflictions, a topic I'll discuss in an upcoming post.

I should add that one cannot fault Walen for his brief discussion of how he would justify the side-effect harms of punishment. He's writing, after all, in an encyclopedia entry. But to the extent he claims to cite a flaw in my reasoning, I see insufficient discussion to backup his claim. (Adapted from work in progress.)

Posted by Adam Kolber on November 11, 2015 at 12:56 PM in Criminal Law, Legal Theory | Permalink


I am grateful to Adam Kolber for taking the time to engage me, and in a generous spirit, on this blog. My primary response is that he is attacking on solid ground here. But I have one thing to say that may help the retributivists, and another to say "yes, they are in trouble"--maintaining my own diffident [bad son at Passover] relationship to retributivism.

In defense of the retributivists, one might say that criminals not only *deserve* punishment, but they partially forfeit their right not to suffer certain side effects. Or, putting it more subtly, they forfeit their equal standing to complain about side-effects. They can still complain, and the side-effects still matter. They still have to be justified. But they may matter *less* than they would for those who do not also deserve punishment. This responds to Kolber's argument: "Imagine if we put school children in an environment with a high risk of sexual assault. Outrageous! Yet that's what we do with prisoners." My point: prisoners are not innocent children. Prisoners' interest in not suffering sexual assault is meaningful, but nothing like the interest children have in not suffering the same. And it might be worth adding that we send children to institutions where they *do* get sexually assaulted, and we do that because the alternatives are worse--too expensive, too much distrust, too restrictive of children's lives, etc.

Saying this requires me to make a modification to my three prong response to Kolber. I should modify prong (b) to read: "the importance of giving wrongdoers what they deserve is sufficiently high given the strength of their claim not to suffer (which may be lessened as a matter of partial rights forfeiture).

Now the concessionary point: as I said in my SOP entry: "Adam Kolber (2013) raises the most difficult proportionality-based criticism of retributivism." What I was referring to was his discussion of pre-trial detention. It is not meant as punishment, yet it is universally subtracted from the intended sentence. This corresponds to his point about the extra unintended gravitons... indeed, it would seem quite odd not to take them into account. I won't rehearse here everything I wrote about how retributivists might try to take them into account, but I will repeat the last sentence: "This is not to say that retributivists could not come up with an adequate response to Kolber, but they have work to do."

I won't end in *quite* so concessionary a tone, however. There is one more thing to say that may explain why sentences should not always be shorter, as Kolber's overfiring retributive machine suggests. It is reasonable to think that in general punishment is meant to include almost all of the normally negative features of it: the bad food, the separation from loved ones (though not their separation from the prisoner; that is just an unavoidable logical other-side-of-the coin), the crappy bed and other loss of creature comforts, the loss of liberty, the loss of career progress, etc. There will still be variation around the mean in how these losses are experienced, but it is not so easy to eradicate that. Individualized tailoring of sentences would quickly lead to scamming and be *very* difficult to administer. The contrast here is with pre-trial detention, which always falls on one side of the ledger, as harming not intended as punishment. That would explain why credit for that is given, while credit for most other excess hardship is not. That still leaves the question of what to do about prison assaults. Those are presumably *not* intended as punishment, as guards *should* prevent those. But the latest data I have seen shows that the rate of prison rape is actually "only" about 4% per year, and the murder rate is lower than on the outside. That is not to say that the rate of sexual and other assault is not too high, but it is arguably not *so* high that it should displace prison time the way that pre-trial detention does.

Lastly, let me be clear, saying this does not address Kolber's deeper objection that retributivists are still committed to giving credit for harms that are not inflicted as punishment, and this is in tension with the rest of the theory. Kolber is right to press on that, and retributivists need to think harder about how to respond.

Posted by: Alec Walen | Nov 11, 2015 9:53:11 PM

The comments to this entry are closed.