Consequentialist Theories of Punishment (Part Three)

(Part One, Part Two)

Welcome to this, the third and final part in my series of posts on consequentialist theories of punishment. The series is working off some of the material contained in Michael Zimmerman’s book The Immorality of Punishment. The book argues for two conclusions. The first is that there are very serious moral reasons that count against the legal imposition of punishment (defined as “intentional harm that is fitting”). The second is that, all things considered, the weight of moral reasons counts against punishment.

I’m not concerned with Zimmerman’s total defence of those claims in this series. My interests are much narrower than that. I’m only concerned with how he defends those claims in relation to consequentialist theories of punishment. These are theories of punishment which seek to justify the imposition of punishment by reference to its desirable consequences, most often that of harm reduction.

There are three such theories dealt with by Zimmerman, each of which highlights a different mechanism through which punishment can achieve its harm-reductive effects. The first is the rehabilitative theory; the second is the incapacitative theory; and the third is the deterrence theory. We’ve already looked at the arguments for and against the rehabilitative theory. In today’s post we’ll look at the arguments for and against the incapacitative and deterrence theories.

Recall at the outset, as was discussed in part one, that the key to evaluating these theories is to ask a contrastive question: Is a punitive form of incapacitation/deterrence more likely to achieve the end of harm-reduction than a non-punitive form? Only if the answer is “yes” will the theories succeed in justifying punishment. This is for a simple moral reason: punishment involves the intentional harming of another human being and it is, all things considered, better not to harm than to harm.


1. The Incapacitative Theory of Punishment
According to the incapacitative theory of punishment, punitive acts, practices and institutions have a harm-reductive effect insofar as they incapacitate harm-producing people. The most common way in which this is achieved is through incarceration, which removes the harm-producing person from society. But other types of punishment can have this effect too. Most obviously, the death penalty could have an incapacitative effect.

The argument might look like this:

• (1) If punitive practices are more likely to reduce harm by incapacitating harm-producing people than non-punitive practices, then there is (at least one) moral reason counting in favour of punitive practices. 

• (2) Punitive practices are more likely to reduce harm through incapacitation than non-punitive ones. 

• (3) Therefore, there is (at least one) moral reason counting in favour of punitive practices.

The principle stated in premise (1) can be accepted for the sake of argument. Premise (2) is where most of the action is. The first problem with it is non-contrastive in nature: can punishment actually achieve the end of incapacitation? Maybe some forms can (e.g. permanent incarceration), but not all (e.g. temporary incarceration, fining or lashing). Further, it seems like there would be certain people for whom the incapacitative effect would be redundant. For example, someone who was never going to commit a crime again, either due to some general disposition or due to the exceptional nature of their crime, would not need to be incapacitated. For such people, punishment could not be justified as it would not achieve the harm-reductive end that is needed to justify it. In practice, it might be very hard to identify such people, and consequently there might be a tendency to err on the side of caution and punish most criminals as a result — Zimmerman suggests this is borne out by the practice on parole boards — but such precaution would still lead to unjustifiable instances of error.

I’m somewhat unsure about this argument. It seems to me that a certain risk of error is both tolerable and justifiable. Indeed, in law it seems like achieving the correct balance of errors (false positives to false negatives) is a constant goal of the system. The trade-off would be the all-important question: are we achieving a sufficient harm-reductive effect to offset the damage done by unjustifiable imprisonment? But this might be a very difficult question to answer in the absence of any ability to identify those for whom punishment has the desired effect and those for whom it does not. Maybe that’s what Zimmerman was getting.

The second problem with premise (2) is probably more telling, though once again I’m somewhat uncertain about it. This problem is contrastive nature: what possible advantage could accrue from punitive incapacitation compared to non-punitive forms? Zimmerman thinks there is none: punishment adds nothing here. It’s difficult to evaluate this argument since Zimmerman doesn’t sketch what he thinks a non-punitive form of incapacitation would look like. To me, it seems like any form of incapacitation would be somewhat punitive since it would involve the elimination of certain choices and the undermining of autonomy. This would be true even in cases of incarceration in a very pleasant environment. And even there problems arise. If the environment is made too pleasant — e.g. if the incarcerated person can interact with a whole group of other prisoners, work in rewarding jobs and so forth — it could open up opportunities for further harmful conduct and thereby undermine the incapacitative effect. Indeed, this happens in contemporary prisons where conditions are far from non-punitive. So we have to wonder about this part of Zimmerman’s argument.


2. The Deterrence Theory of Punishment
According to the deterrence theory of punishment, punitive acts, practices and institutions can be justified in terms of their deterrent effect. In other words, they can be justified because they discourage would-be criminals from engaging in harmful conduct. This satisfies the goal of harm-reduction. The argument runs as follows:

• (4) If punitive practices are more likely to deter would be criminals than non-punitive ones, then there is (at least one) moral reason counting in favour of punitive practices. 

• (5) Punitive practices are more likely to deter would be criminals than non-punitive ones. 

• (6) Therefore, there is (at least one) moral reason counting in favour of punitive practices.

Again, we can skip over premise (1) here and focus our attention on premise (2). In a non-contrastive sense, it seems to be undeniably true that punishment has a deterrent effect. Zimmerman concedes as much. There is a common sense argument for this. To take just one example, would you break the speed limit if you knew that getting caught would result in an automatic five year prison sentence? No doubt the probability of getting caught will factor into the decision too, but I have to say that, from my perspective, an automatic five-year sentence for speeding would make me far less likely to break the limit. In addition to common sense, there seem to be empirical cases backing up the deterrent effect of punishment. Zimmerman cites the case of Singapore, which imposes the death penalty for a wide range of offences, and has a very low crime rate (presumably as a result of punitive practices). Unfortunately, no proper social-scientific analysis of the data is conducted, but the claim is certainly not implausible.

So there is good reason to think that punitive practices have a deterrent effect. But are there non-punitive practices that could be equally as effective? One could argue that punishment is really not needed in order to achieve the deterrent effect. Rather, it is the threat of punishment that is needed. To return to the speeding example, my decision not to speed is guided by the threat of five year imprisonment, not by the actual fact that I have been imprisoned for five years.

The problem with this line of reasoning is that it is very difficult to see how the threat of punishment could have the deterrent effect without there being, at least some, occasional cases of actual punishment. In order to be effective, the threat needs to be credible, and if punishment is never imposed it couldn’t really be credible. One possible source of salvation is fake punishment. In other words, the threat could be made credible by giving people the illusion that they will be punished. However, this would seem to require a massive, and probably unsustainable, conspiracy in order to work. Thus, it seems more likely that an actual system of punishment would be more effective as a deterrent than a fake one.


3. Punishing the Innocent
Perhaps surprisingly then, Zimmerman finds the deterrent argument to be successful. Obviously, this is not a welcome conclusion given the overall project of his book. So what does he do to avoid it? Predictably enough, he falls back on the old problem of punishing the innocent. Ostensibly, this is a problem that works against all consequentialist theories of punishment, but since deterrence is the only remaining candidate we’ll focus on that.

I’ve discussed the problem at greater length before. In essence, it boils down to this: if punishment is to be justified in terms of its deterrent effect, then one could justifiably punish the innocent on this basis. Why? Because the goal is achieving the greatest level deterrence and punishing the innocent as well as the guilty would have a greater deterrent effect than punishing the guilty alone. But there are serious moral reasons against punishing the innocent, so if a deterrence-based justification allows for this, there must be something wrong with that justification.

In effect, this is a slippery slope argument. It claims that embracing the deterrence principle will lead, either logically or causally, to the unpalatable consequence of punishing the innocent. As follows:

• (7) If we accept the deterrence principle of punishment, then we must also accept that it is okay to punish the innocent. 

• (8) It is not okay to punish the innocent. 

• (9) Therefore, we ought not to accept the deterrence principle.

There are a variety of responses open to the proponent of deterrence. The first is to bite the bullet and accept that punishing the innocent is okay. I looked at that possibility in my earlier post on this topic. The other is to argue that there is some logical or causal blockage that prevents the slide from deterrence to punishing the innocent. I’ll look at those possibilities here.

From the causal perspective, one could argue that punishing the innocent is unlikely to have a deterrent effect. Indeed, punishing the innocent may serve to reduce the deterrent effect of punishment, particularly if it is done in a heavy-handed fashion. After all, if our likelihood of being punished was the same irrespective of whether we had engaged in harmful conduct or not, we would lose the incentive to forbear from harmful conduct. Zimmerman acknowledges this point, but thinks it would only imply that governments should be careful in how they punish the innocent, not that they wouldn’t be justified in doing so.

From a logical perspective, one could argue that there’s no obvious inconsistency created by introducing a special clause in the deterrence principle that excludes punishment of the innocent. We would achieve a sufficient deterrent effect from punishing the guilty anyway, so why we would need to take the unwelcome step of punishing the innocent? For Zimmerman, the problem with this argument is that it assumes it is easy to justify punishing the guilty. But he thinks this isn’t the case and he dedicates subsequent chapters of his book to showing why. Unfortunately, those chapters lie beyond the scope of this series, so I’ll have to leave it there. Read the book if you want more.

To sum up, Zimmerman thinks that neither the incapacitative nor the deterrent theory of punishment provides a satisfactory justification for punitive practices. The incapacitative theory fails because non-punitive incapacitative practices are possible and they would be more justified than punitive ones. In my analysis above, I expressed some doubts about this by wondering whether non-punitive incapacitation could really exist. As regards the deterrence theory, this seems much more promising since there do not appear to be viable non-punitive alternatives. Nevertheless, Zimmerman thinks that even the deterrence theory has its problems because it justifies punishing the innocent. But, as we just saw, there may be some reasons for doubting that too. So, I’m a little more sanguine about the deterrence theory than Zimmerman is.