Less Blame, Less Crime? The Practical Implications of Moral Responsibility Skepticism
Most philosophers believe that wrongdoers sometimes deserve to be punished by long prison sentences. They also believe that such punishments are justified by their consequences: they deter crime and incapacitate potential offenders. In this article, I argue that both these claims are false. No one deserves to be punished, I argue, because our actions are shot through with direct or indirect luck. I also argue that there are good reasons to think that punishing fewer people and much less harshly will have better social consequences, at a reduced overall cost, then the long prison sentences that are usually seen as required for social protection.
In every nation, a great deal of state money is spent on the punishment of criminals. The bulk of this expenditure is on prisons. In the United States, federal, state and local governments spent an estimated U$75 billion on corrections in 2008 (Schmitt, Warner & Gupta 2010). There were more than one and half million people imprisoned in the United States at the end of 2011, the great majority in state facilities (Carson and Sabol 2012). This is enormously costly: not only directly, in dollar terms, but also (perhaps especially) in its indirect effects. Around 40% of the prison population is black (though blacks make up less than 14% of the total population). One in three black men can be expected to go to prison in their lifetime; a higher percentage than go into higher education (Lyons & Pettit 2011). Upon release into the community, former inmates have significantly worse employment prospects than those who have not been imprisoned. These facts have effects on majority black communities, in particular: they are implicated in childhood poverty, behavior problems and mental illnesses, and may help to constitute a criminogenic environment, perpetuating the problems (Lyons & Pettit 2011).
Of course the US criminal justice system is widely acknowledged to have distinctive problems all of its own. In other countries, a far smaller percentage of people are imprisoned, and imprisonment does not seem to be as racially biased as it appears to be in the US. Whereas the United States imprisons 756 people per 100,000 of its population, the median rate for Southern and Western European countries is a dramatically lower 95 (Walmsley 2009). Costs, direct and indirect, are correspondingly lower. Nevertheless, even in these cases, the costs—to the taxpayer (including the cost of foregone tax incomes), to families, communities, and the prisoners themselves—are very significant.
Imprisonment does not only have costs, of course: it has benefits too. These benefits accrue to many different groups of people. Punishment is typically justified, in very important part, by its role in deterring crime, and by the fact that it incapacitates individuals some of whom would go on to commit more crime were they not imprisoned. To the extent that deterrence and incapacitation are genuine, imprisonment therefore brings benefits to potential victims of crime. Imprisonment can also—in theory at any rate—benefit the person imprisoned him or herself. It might provide an opportunity for the treatment of drug addictions or mental illness, the acquisition of skills, and so on. In practice, actual prisons tend to make the kinds of problems that lead to crime worse, not better (to increase drug use and mental illness, and worsen life prospects, making crime relatively more attractive, and to cause the breakdown of communities, thereby indirectly contributing to high crimes rates), but in principle punishment could benefit the person punished; sometimes, prisons actually fulfill this function. Prisons also provide employment; the growth of the penal system in the US has made it an important contributor to economic activity.
The costs and benefits of punishment are directly relevant to whether it is justified. But most philosophers think that these questions are secondary when it comes to justifying punishment. What matters most, they claim, is whether the incarcerated deserve to be punished. If they do not, then the question whether punishment is justified becomes a question not merely of weighing up its costs and benefits, but also of comparing its costs and benefits to alternative methods of responding to crime. In this article, I will argue that punishment is not justified on the basis of desert, and that therefore we should be assessing its costs and benefits relative to alternatives. I will suggest that less punitive responses to crime may in fact have benefits: to criminals, potential victims of crime, and society more generally. Rejecting the notion that people deserve to be punished opens our eyes to possibilities of responding to crime that will be more effective, cheaper and more humane.
People deserve to be treated better or worse because of the way they have acted (alone) only if they are morally responsible for their actions. Desert, at least with regard to actions and omissions, is backwards looking: someone deserves something (reward, praise, punishment, or what have you) just because of what they have done or failed to do in the past. Other kinds of justifications, that do not invoke moral responsibility, are forwards looking: they ask about the future consequences of distributing benefits and burdens. In asking whether people deserve to be punished for breaking the law, we ask whether they are morally responsible; whether they deserve to be treated better or worse on backwards-looking grounds alone.
Debates about moral responsibility have usually focused on whether the freedom required for responsibility is compatible with causal determinism. Some philosophers have argued that if the universe is deterministic—that is, roughly, if the physical facts plus the laws of nature determine a unique future, for every instant of the universe—than no one can be morally responsible. The classical argument for this conclusion turned on access to alternative possibilities: allegedly, if our actions are determined, we lack the power to act in more than one way, and we are therefore not responsible for our actions. More recent versions of the argument have turned on whether we are the source of our actions, in a certain sense, independent of worries about alternative possibilities. Compatibilists—philosophers who hold that causal determinism is compatible with freedom—say that we do not need access to alternatives in the ‘all in’ sense to be morally responsible: we need only to be responsive to good reasons and not to be coerced or compelled to act against our better judgment.
These debates, in the metaphysics of free will, deserve to be taken seriously. However I will not address them here. Instead, I will present independent arguments for the conclusion that agents are not morally responsible for their actions.
Luck and Moral Responsibility
Libertarians in the philosophy of free will are philosophers who hold that free will is incompatible with causal determination of action—that if conditions sufficient for my behavior are always in place well before I act, I do not act freely—but who also hold that we have free will (note that this view is entirely independent of libertarianism as a view in political philosophy, which holds that the just state is one in which there is minimum interference with individual choice; it is quite possible to be a compatibilist about free will and a libertarian in political philosophy). According to most libertarians, we can choose between genuinely open alternative possibilities. Compatibilists have long accused libertarianism of being susceptible to an argument from luck. Absence of determinism does not increase our control over our actions, they argue: rather, it makes how we act subject to responsibility-undermining luck. If the universe is indeterministic, such that when I contemplate whether to perform a good action or a bad, there is some chance that an indeterministic event might cause me to choose the first and some chance that an indeterministic event might make me choose the second, then I do not control which I do (everything about me—my beliefs, desires, values, and so on—is consistent with me doing either). And that entails that how I choose is a matter of luck. Luck is incompatible with moral responsibility: if the only difference between two agents is the product of luck, then they can’t deserve differential treatment (Mele 2006).
This argument against libertarianism is a powerful one. However, the problem of luck is by no means confined to libertarianism. Luck is ubiquitous in our lives: people are unlucky to fall ill or lose their jobs (which isn’t to say, of course, that everyone who falls ill or loses their jobs is unlucky: if I am sacked for stealing office supplies, that’s not just bad luck) and lucky to win the lottery or to narrowly escape an accident. Most physicists believe that the universe is indeterministic, but it is not because determinism is false that we are often lucky and unlucky (it remains an open question how frequently the kind of indeterminism the physicists maintain is true affects events of the kind that matter to us). An account of luck should therefore not suppose that indeterminism is required for luck.
In earlier work, I have defended a detailed account of luck (building on the work of Pritchard 2005 and Coffman 2007). Here I shall just give the outlines of the account. An event or state of affairs is lucky if it is chancy, significant and out of the control of the person who it is lucky for. To say it is chancy is to say something about the probability of its occurrence. Roughly, the less probable an event, the luckier it is if it occurs (so winning the lottery is very lucky, but correctly guessing ‘heads’ when a coin is tossed is much less lucky). Just how improbable an event must be to count as lucky is sensitive to how significant it is to the person. If it doesn’t matter at all, then it is not lucky no matter how unlikely it is. If it matters a great deal, then it need not be all that unlikely to be lucky or unlucky (compare winning 50 cents by guessing heads, with surviving a single round of Russian roulette using a six-chambered revolver and one bullet: clearly the person who survives at Russian roulette is luckier than the person who wins 50c, even though surviving Russian roulette was rather more probable than winning 50c). Finally, the event must be out of the agent’s control to be lucky for her: if I am able to control whether an event occurs, its occurrence is not lucky for me.
This account gives us a way of judging when an event is lucky or not. If I am hit by a car in a situation in which being hit is very unlikely (say on the sidewalk of a quiet road), I may be very unlucky. But if I am hit by a car in a situation in which being hit by a car is rather likely (say I have fallen asleep in the middle of the highway) I am not unlucky. When Tiger Woods sinks a straightforward putt, he is not lucky, because his skill and control ensures that he sinks putts like that 99 times out of 100. If I were to sink that putt, I would be lucky because I rarely sink putts like that. In order to understand the significance of luck for human life, though, we need an account of luck that is able to explain effects beyond those it has (directly) on the events within a life.
In particular, we need to be able to account for what Thomas Nagel (1979) called constitutive luck: luck that helps to form people into the kinds of people they are. Some people seem to be the victim of terrible constitutive luck: they are born with disabilities that are incompatible with flourishing lives, for instance. Other people seem to be the beneficiaries of wonderful constitutive luck: they seem to be born talented, healthy and resilient. It is hard to model this kind of luck using the resources of the account of lucky events, because there are difficulties in making sense of how lucky an individual is to be born with a particular trait. However, it is easy enough to use the resources of the account of lucky events to construct an account of constitutive luck. Something is constitutively lucky for a person if it is significant for her, out of her control and it is relatively unlikely in the circumstances into which she is born. This account gives us a context-relative account of constitutive luck. For instance, if a person is born at a time in history when most people die at 40, she might be lucky to be so endowed as to live to 50, whereas if she were born into a developed country today with an expected lifespan of 50, she would be unlucky.
The principle on which critics of libertarians rely when they argue that libertarianism makes actions unacceptably subject to luck is this: no one can be responsible for anything that is for them just a matter of luck. The best way to bring this home is to compare two agents. If the only difference between them is a matter of luck, then they can’t deserve different treatment. Suppose that two marksmen are each shooting at a target. Each is equally skilled and has practiced equally hard. Each focuses all their attention on the target and fires. Marksman A’s shot hits the centre of the target. Marksman B’s bullet flies straight and true, but fails to hit the target because a bird flies into its path, deflecting it. In that case, we would not think that A deserves praise that is not due to B, because luck and luck alone explained the difference between them. I will argue that luck and luck alone explains the differences between agents who find themselves performing wrongful acts and those who don’t.
First, consider the influence of lucky events on how people may find themselves acting badly or well. When we deliberate about what to do, it is often true that small things can make a decisive difference. For instance, it might be true that I would fail to let the person walking in front of me know that a $20 bill has fallen out of his pocket if it occurred to me, at the right time, that he looks well-dressed enough not to need the money (or if it occurred to me that his shirt reminds me of someone I don’t like, or if my recent stress and therefore sense that I deserve a break crossed my mind, or what have you; of course, none of these thoughts need be conscious for them to influence my behavior). Now, the occurrence or nonoccurrence of thoughts like these may be a matter of luck for me: they may be significant (since they help settle what I end up deciding), out of my control and relatively unlikely (in the right circumstances, odds of 50% may be sufficient for a thought to be lucky, so I may be lucky if it occurs to me and if it does not). In cases like this, which I claim are quite common, how I end up deciding and therefore acting may be a matter of responsibility-undermining luck.
To see how this kind of luck may undermine moral responsibility, let’s work through an example. I see the $20 bill fall from someone’s pocket and quickly deliberate whether to alert him to his loss, before it is too late and he runs across the road. If an appropriate thought occurs to me, I alert him; if it does not, I do not. Suppose I do not: am I to blame? The only difference between me in the case in which I alert him and me in the case in which I do not is a difference due to luck. But as we have seen, two agents cannot deserve different treatment if the only difference between them is due to luck. I can’t deserve different treatment in one case than in the other, it seems; therefore I am not to blame if I keep the money (unless, bizarrely, I am also to blame in the case in which I return the money).
Cases like this one—cases in which an action has the moral character it has due to luck—are common. We often perform one kind of action rather than another because a chance thought happens to strike us at the right moment. The cause of that thought may often be a chance factor in our environment, often one of which we are not consciously aware. These features—a billboard, a snatch of conversation, or even the way the light glances off the window, triggering a (possibly unconscious) memory—serve to prime our behavior, in much the same kinds of ways in which the primes used in psychological experiments may cause us to act one way or another (priming occurs when information of which we are not conscious, or which has effects on us of which we are not conscious, makes other ideas more accessible to us and thereby alters our behavior). Sometimes, there will be no illuminating explanation of why a thought happened to occur (though there will be causal explanation, such an explanation may not really explain the occurrence in psychological terms). Given that there are some circumstances in which most of us are capable of performing laudable as well as somewhat shoddy actions, we are all sometimes vulnerable to this kind of responsibility-undermining luck.
However, there are also a wide variety of circumstances in which we are typically not vulnerable to this kind of luck. There may be no circumstances in which, constituted as I am, I am capable of murder were the right thought happen to strike me. In any case, with regard to very many murders it would be bizarre to excuse the murderer on the grounds that such a thought happened to strike him. A would-be murderer may find himself contemplating such a course of action because the right thought happens to strike him, but many murders require some degree of planning and therefore take time. A chance thought may set us down a path, but when we have time to reflect on what we are doing, the influence of such chance thoughts may be reduced, as we bring our stable beliefs and values to bear to its assessment, and as new, sometimes conflicting, thoughts occur to us. So while there are grounds for excusing some actions on the grounds that they are caused by responsibility-undermining luck of the kind outlines above, there are many other actions in which such grounds are lacking.
In such cases, agents are not vulnerable to luck in the kinds of considerations that happen to strike them because the necessary conditions for such vulnerability are not satisfied. To be vulnerable to this kind of luck, a stray thought must be capable of making an important difference to how we act (perhaps by making an important difference to how we reason) and that occurs only when things are relatively finely balanced. That is, it is only when we are so constituted that actions with quite different and conflicting moral characters are compatible with our attitudes and values that luck makes a difference in this kind of way (note that it is not enough that things be so balanced that chance factors can make a difference to our behavior for us to be excused on these grounds—no one should be excused for murder because it is a matter of luck that they bludgeoned their victim to death, rather than stabbing them—the luck has to make a significant difference to the moral character of the action to serve as an excuse). Now, things are not finely balanced, in the right kind of way, when, and only when, agents are so constituted that their attitudes and values are resistant to this kind of chance influence. And that happens as a result of constitutive luck. In other words, the explanation for why agents are often resistant to one kind of luck is that they are subject to another kind of luck instead.
Consider the normally virtuous agent, going about her daily business. Perhaps someone is rude to her, or cuts her off in the car park. Perhaps, as a result, she experiences a flicker of irritation that causes her (perhaps unconsciously) to entertain the thought of punching the person in the nose. But she doesn’t do it; she doesn’t even allow herself to entertain the thought for more than a fleeting instant. The explanation of why she does not punch the offender, in circumstances in which someone else might, is that it is inconsistent with her values to do so; unless she was subject to extremely severe provocation, there is no coherent story in which those values could explain that kind of behavior. So she is resistant to this kind of responsibility-undermining luck. But why is she resistant? In other words, why does she have the values and temperament she has?
The explanation will cite her constitutive luck. Her genes may explain her relatively even temperament, which makes her resistant to provocation. Her upbringing was relatively stress free, thereby ensuring that she does not experience the kind of anxiety from which those with stressed infancies suffer, which dispose them to react more rapidly and less reflectfully to perceived threats. She has received a good enough education, including a moral education, through which her responses to people and situations were trained in such a manner that aggression is not an immediately accessible strategy for her. All these things are not things for which she is responsible. They are her good luck. Had she been born differently constituted, she would not have had her even temperament. Had her environment been more stressful, her role models less virtuous, or had violence been rewarded in her social setting, she would have been disposed to react quite differently. She would have been vulnerable to passing thoughts: chance thoughts might have led to her behaving badly, or well. Or she might even have been resistant to such thoughts because she is too bad for such vulnerability. Her actual lack of vulnerability to luck in what thoughts occur to her is explained by luck in her constitution.
Philosophers have sometimes argued that constitutive luck does not undermine moral responsibility because normal agents take responsibility for their constitutive luck. Normal agents play an active role in shaping who they are and what their values are: they modify their initial values (those they have due to constitutive luck) or they endorse them. This claim—that agents play an active role in self-shaping—is certainly true. It is, however, false, that in so doing these agents escape the responsibility-undermining effects of constitutive luck.
When we shape ourselves, we do so in ways that either simply express the values and attitudes we have as a result of constitutive luck, or in ways that are vulnerable to luck in what thoughts happen to strike us and other chance events. We shape ourselves, but our self-shaping is no less subject to luck than is our initial constitution.
Consider our even-tempered, moderately virtuous, agent again. She may not have always been either even-tempered or moderately virtuous; she may be partially responsible for how she is now, in the causal sense of ‘responsible’ (that is, it may be true that how she is is partially the result of things she did). Perhaps she started out quite different and deliberately shaped her responses. But however she did it, the self-shaping was the product of luck, in one way or another. Suppose, first, that in shaping herself she acted as she took herself to have most reason to act. This might occur because she was already disposed to usually (though not always) respond in the kinds of ways in which she is (pretty much) always disposed to respond now, or because she is disposed to value the way in which she now responds. In that case, in shaping herself she was expressing her constitutive luck. She made choices that seemed (often enough) natural and obvious to her to make; as we saw above, when things are not evenly balanced—when choices seem natural and obvious—constitutive luck explains why our reasons for action have this kind of character. In this kind of case, she does not escape luck; rather, she expresses it.
Suppose, however, that shaping herself in this way did not seem so natural and obvious to her. Suppose that things were more finely balanced, so that she could just as easily have ended up hardening herself to the needs of others. These are precisely the kinds of conditions in which luck in the considerations that occur to her may have a decisive effect (and luck in her environment, too: for instance, in who she encounters and befriends, and the moral outlooks they happen to have). In cases like this, to the extent to which her self-shaping does not express her constitutive luck, it is the product of luck around the time of choice.
Most cases of self-shaping will be mixed: we will express aspects of our constitutive luck but also respond in ways affected by lucky events around the time of choice in shaping ourselves. And the relative influence of the different kinds of luck may change over time: as we develop a more stable personality and outlook, our choices will likely come to seem more obvious and natural to us, and will be correspondingly less vulnerable to luck around the time of choice. But self-shaping does not allow us to escape the effects of luck; at best, it allows us to escape the effects of one kind of luck by ensuring that another kind plays a more prominent role. This is not surprising: if we are subject to luck in our choices prior to self-shaping, we are subject to luck in self-shaping, because the choices involved in self-shaping are, after all, just choices like any other.
If luck undermines moral responsibility—as I claim—self-shaping does not enable us to become morally responsible, because it is itself the product of (responsibility-undermining) luck. Luck is not a problem that can be solved. It must be avoided or dissolved if we are ever to hold anyone morally responsible.
There are a number of philosophers who might shrug their shoulders at the arguments presented above. They would deny that luck is a problem we need to solve. They accept that someone might be lucky or unlucky to have a good, or bad temperament, a good or bad education, or to find themselves in challenging circumstances—the kinds of circumstances in which ordinarily good people may easily find themselves condoning or even engaging in terrible actions; think of ordinary Germans in the 1930s, or ordinary Southerners in the United States prior to the civil rights movements. Nevertheless, they hold, these kinds of luck do not undermine responsibility. Someone might be unlucky to come from a deprived background, to have had bad role models and to have the kind of life prospects that make crime more tempting than it is for me (say), and yet be morally responsible for their crimes. Some philosophers are unperturbed even by the fact that how we choose is often subject to chancy events occurring around the time of choice: they maintain we can be responsible for these lucky choices.
These philosophers defend a variety of views. For some, what matters is whether our actions genuinely express our attitudes (Smith 2005). For others, what matters is whether we respond to moral reasons in acting (Arpaly 2002). I do not intend to argue against these and other views that deny that luck undermines responsibility. Rather, I aim to establish something less demanding: that whatever plausibility they have is not sufficient to justify the more severe kinds of blame that responsibility is usually taken to underwrite. That is, none of these theories suffice to establish that agents are responsible enough, or responsible sufficiently beyond reasonable doubt, to justify sanctioning them; in particular, they can’t justify the kinds of hard treatments typically involved in punishment.
First, I want to highlight some of the moral costs associated with accepting the view that we can justifiably sanction agents even when they are lucky, limiting myself to the kinds of cases which are commonly encountered (rather than being concerned, as philosophers often are, with all possible cases). In the actual world, I claim, many of the people we blame are the victims of a double dose of bad luck. First, they are unlucky to be badly off in various ways, and unlucky that being badly off in these ways disposes them toward crime. They may then be unlucky to actually commit a crime. This double dose of luck should give us pause before we conclude that these ordinary criminals deserve punishment.
We might illustrate this kind of claim by considering the kinds of criminals that routinely pass before our courts: car thieves, burglars, drug addicts, and so on. These people typically come from relatively deprived backgrounds, which explains several facts about them. It helps to explain why they find crime tempting (alternative strategies of securing income and respect are much less available to them than to us). It helps to explain why they are less likely to think crime is morally unacceptable (one is likely to be less horrified by crime if one’s father or other respected older figures are criminals, or if going to prison is so common in your neighborhood that it carries little stigma). Deprivation can also cause cognitive and psychological problems that make crime more likely, either by reducing the capacity for self-control (Hackman & Farah 2009) or by reducing IQ and therefore reducing the capacity to understand potential consequences. In ordinary cases, criminals are disposed toward crime because they are—unluckily—badly off. Then they are blamed by those who are—luckily—well off, for giving in to temptations we do not face, with meager resources to withstand them. It may be false that there but for the grace of God we go, but we have our luck, and not our merit, to thank for that.
Still, not every criminal is from a deprived background. White collar criminals are sometimes from privileged backgrounds; even drug addicts and burglars sometimes come from such social strata. These people do not seem to be the victims of a double dose of bad luck; rather, they are the beneficiaries of good luck, who have gone astray. Even in cases like this, I claim, a double dose of bad luck is to be found: there is some explanation of why they turned to crime, and that explanation will cite luck (bad company, losing a job, or a genetic predisposition, say). In any case, I do not mean these thoughts to be decisive. Some people will maintain that we can hold some or all of these people responsible despite their bad luck, and I have given them no argument that must force them to change their mind.
However, once we recognize what is at stake, these philosophers are no longer entitled to their complacency. Given that the argument from luck (and, indeed, other arguments designed to show that no one is morally responsible for their actions; Pereboom 2001; Waller 2011; Zimmerman 2011) is powerful, and given that the stakes are high, we ought to err on the side of refraining from sanctioning.
The stakes are high for an obvious reason: because sanctioning involves the imposition of suffering on those sanctioned. Deprivation of liberty is, for almost all people, a great burden. Add to this the kind of regimentation and disrespect that is normal in prisons and the costs of imprisonment to those imprisoned can be seen to be high. These features are intended to be part and parcel of punishment: it is to these conditions that judges sentence criminals, to deter them and others and (according to some theorists) to redress the moral balance by having them suffer in some kind of correspondence to the benefit they took in committing a crime. Actual prisons tend to be awful places, because they have features that are not part and parcel of the punishment but which often accompany punishment. Abuse of all kinds, both by other prisoners and sometimes by those in authority, is very common. Mental illness is often exacerbated by imprisonment. Drugs are widely available and prisoners seeking relief from boredom may find themselves addicted. Even after release the punishment may, in some sense, continue: the stigma of imprisonment dogs former inmates, ensuring that legal means of supporting themselves are more difficult to obtain.
Of course, the stakes are not always this high. Sometimes they are higher still (consider supermax prisons or capital punishment); sometimes they are lower (not every prison is a hell hole, and there are a variety of alternatives to imprisonment, from community service to fines). The higher the stakes, the more powerful arguments for the compatibility of luck and responsibility must be, if sanctions are to be justified.
We should err on the side of thinking that sanctions are not justified, because sanctioning involves the imposition of harm, and there is a strong moral presumption against harming. That is, everyone agrees that we ought to refrain from harming other agents unless there are very good reasons not to refrain; thus, the imposition of harm is permissible only if a substantive burden of justification is successfully shouldered. Given the presumption against harm, sanctioning requires that the burden of proof be decisively on the side of harm; the bigger and the more irrevocable the harm, the better the justification must be (Vilhauer 2009). It is therefore not enough to show that it is not irrational to dismiss the luck argument; we can justify sanctions only if we can show that the argument really lacks force (all things considered). And no one, I claim, has been able to do that.
It might be objected that the presumption against harm is not as hard to overcome as suggested; it does not require, in particular, that we show that the luck argument lacks genuine force. We can overcome the presumption much more simply, by showing that the consequences of not harming are likely worse than the consequences of harming: we can invoke deterrence, incapacitation, and rehabilitation to justify sanctions. But that’s the wrong sort of justification. What’s at issue is whether people can deserve to be sanctioned, independent of the consequences. If we must resort to consequentialist grounds to justify sanctions, then it becomes an empirical issue whether punishment (say) can rightfully be imposed. It is fully compatible with the consequentialist approach, for instance, that the correct response to wrongdoers is to reward them, because that’s what has the best consequences. Indeed, in certain cases that is not merely theoretically possible but even likely: it may be that the best response to some lawbreakers is to offer them opportunities for education, and even food and (decent) lodgings, not to punish them.
What kinds of sanctions remain justified?
Still, the availability of some kind of consequentialist defense of sanctions leaves open the possibility that showing that sanctions are unjustified on the basis of desert makes no real difference to how we should act or to public policy. No matter whether criminals are typically, even always, bad as a result of bad luck, some of them are indeed bad, and left to themselves they will make other people suffer through their actions. These people must be prevented from committing theft, assault, and worse crimes. Other criminals may not be really bad; nevertheless without the threat of sanctions, or simple incapacitation, they will make other people’s lives worse, by stealing their car or their television, or by losing control and taking a swing at them. Showing that sanctions are unjustified does not begin to show that we can afford to close the prisons.
One response might be to make prisons much less harsh than they currently are. Indeed, given that prisons are breeding grounds for further crime and that violence and abuse are often rife in them, we have good reasons to do this. We should ensure that mental illness and addiction is treated in prison, not worsened. We ought to ensure that prisoners get educational opportunities to give them a chance of lawful employment when they get out. But it is natural to think that we can’t go too far in the direction of making prisons secure, safe environments in which prisoners get the opportunity to make up for some of their past bad luck. The less harsh the prisons, the less the deterrent effect of the threat of imprisonment, so there are consequentialist grounds for thinking that prisons must be at least somewhat harsh. Indeed, some have worried that the logical conclusion of skepticism about moral responsibility is that prisons are made so pleasant that they come to be attractive, and people begin to commit crimes in order to get incarcerated (Smilansky 2011). They have also pointed out that making prisons less harsh might be costly, in pure financial terms: these costs weigh against that course of action.
Once we give up on the thought that sanctions can be deserved (or, more cautiously, that we are justified in imposing harsh sanctions), we can begin to weigh up the costs and benefits of imprisonment and rival responses to wrongdoing in consequentialist terms. It is becomes an empirical matter how harsh prisons should be. The worries mentioned above notwithstanding, there are very good reasons to think that the optimal harshness is very much less than is typical today. In this final section, I will tentatively suggest that the optimal sanctions may, in some ways, be extremely light.
Though there is good evidence that deterrence works, at least for some kinds of crimes, there is also evidence that human beings are typically more responsive to positive incentives than to negative (Levy 2013). The threat of punishment does indeed have an effect on behavior, but that effect is smaller than the promise of reward (and the two combined are more effective still). For this reason, we do better to offer criminals and potential criminals opportunities if they refrain from crime, perhaps coupled with the threat of punishment if they do not, then to focus on punishment alone. Genuine opportunities—training, jobs, treatment for mental illness and addiction—are likely to prove far more effective than mere threats. Moreover, such opportunities are much cheaper to provide, in the long run, since they have the potential to make the potential criminal a productive member of society, ensuring ongoing tax revenues and economic activity.
By coupling threats with opportunities, we can reasonably hope to reduce the number of people imprisoned, and therefore reduce the suffering endured by offenders (suffering which, we are assuming, is undeserved). We do so at the same time as we reduce economic costs and the number of crimes committed. There is also reason to think that when we must punish, we may be able to punish much less than we do now. Though threats have a genuine deterrent effect, this effect diminishes rapidly. Two years imprisonment is much less than twice as effective a deterrent as one year. The optimal amount of imprisonment time, measured by its deterrent value alone, may be startlingly low. Swiftness and certainty of punishment correlate far better with deterrence than does sentence length (Kleiman 2009). Further, shorter sentences actually increase certainty and therefore have a multiplying effect on deterrence: prisoners do not tend to appeal short sentences and courts have little patience for them when they do. As Kleiman (2009: 3) notes, in criminal justice systems like those prevailing in most developed countries, “severity is incompatible with swiftness and certainty”.
The criminological evidence is mixed and its interpretation difficult. However, there is no doubt at all that we can punish much less than we currently do, and achieve better results for all concerned (potential victims of crimes, tax payers, and offenders themselves). It may be that we can achieve these results with sentences that are months, and not years, long for most offenders. We should not pretend these solutions will reform all criminals. Some people will remain threats to innocent others, no matter how much we encourage them to change. Some psychopaths probably fall into this category, though some people believe that even psychopaths will refrain from harming others if other strategies prove rewarding. We may need to incapacitate some offenders for years or even decades. However, we may not need to punish them: though they are deprived of their liberty (at some cost to them, of course) they may be compensated in various ways. A lot depends on how many people fall into this class of offender (in a more just society, in which resources are better distributed, there will probably be fewer people who experience the kinds of environments that lead to being irredeemably bad; there is evidence that many people genetically disposed to violence will not develop aggressive tendencies without exposure to abuse (Caspi et al. 2002)). If numbers are low, then we can afford to ensure that conditions for those who must be incapacitated compensate them for deprivation of their liberty.
In a more just society, worries about making imprisonment so lenient it becomes positively attractive also diminish. How attractive decent food and lodgings are depend on available alternatives: if one is sleeping rough, then the disincentive represented by the deprivation of liberty dwindles in importance. This worry can also be significantly mitigated by the fact that for non-psychopathic offenders, sanctions may be rather unpleasant, though also quite brief (Kleiman (2009: 94) suggests that a 40 hour stint in solitary confinement without radio or television might be aversive enough to have a high deterrent value for a juvenile offender; of course, one would want to be sure that these benefits did not come at the cost of offenders’ mental health). It might be that a diagnosis of psychopathy—for which there are good tests—is required to avoid the harsher sanctions, ensuring that imprisonment retains its deterrent value for those (and only for those) that can be deterred.
If no one deserves punishment, the way is open for us to explore which sanctions can best promote the goals we aim to achieve. Though the evidence remains inconclusive, it may be that if we give up on thinking that people must be punished in proportion to their offenses, we can greatly reduce prison populations and sentence lengths, thus saving a large proportion of the money we currently expend on the corrections system. We can do this while reducing crime and its impacts on victims, and reducing the suffering of offenders and potential offenders. We may not be able to achieve all these goals except in a society that is already just, but moving toward these goals may constitute a step toward such a society.
Acknowledgements: I would like to thank two anonymous referees for the Journal of Practical Ethics. This work is supported by a generous grant from the Australian Research Council.
Arpaly, N. 2002. Unprincipled Virtue. New York: Oxford University Press.
Carson, E.A. & Sabol, W. J. 2012. Prisoners in 2011. Bureau of Justice Statistics. http://bjs.ojp.usdoj.gov/content/pub/pdf/p11.pdf
Caspi A., McClay J., Moffitt T.E., Mill J., Martin J., Craig I.W., Taylor A. & Poulton R. 2002. Role of genotype in the cycle of violence in maltreated children. Science 297 (5582): 851–4.
Coffman, E.J. 2007. Thinking about Luck. Synthese 158: 385-398.
Hackman, D. A. and Farah, M. J. 2009. Socioeconomic Status and the Developing Brain. Trends in Cognitive Science 13: 65–73.
Kleiman, M. 2009. When Brute Force Fails. Princeton: Princeton University Press.
Levy, N. 2013. Punishing the Addict: Reflections on Gene Heyman. In Thomas Nadelhoffer (ed). The Future of Punishment. Oxford University Press, pp. 233-245.
Lyons, C.J. & Pettit, B. 2011. Compounded Disadvantage: Race, Incarceration, and Wage Growth. Social Problems 58: 257-280.
Mele, A. 2006. Free Will and Luck. Oxford: Oxford University Press.
Nagel, T. 1979. Moral Luck. In Mortal Questions, New York: Cambridge University Press, pp. 24-38.
Pereboom, Derk. 2001. Living Without Free Will. Cambridge: Cambridge University Press.
Pritchard, D. 2005. Epistemic Luck. Oxford: Oxford University Press.
Schmitt, J., Warner, K. & Gupta, S. 2010. The High Budgetary Cost of Incarceration. Centre for Economic and Policy Research. http://www.cepr.net/documents/publications/incarceration-2010-06.pdf.
Smilansky, S. 2011. Hard Determinism and Punishment: A Practical Reductio. Law and Philosophy 30: 353-367.
Smith, A. M. 2005. Responsibility for Attitudes: Activity and Passivity in Mental Life. Ethics 115: 236–271.
Vilhauer, B. 2009. Free Will and Reasonable Doubt. American Philosophical Quarterly 46:131-140.
Waller, B.N. 2011. Against Moral Responsibility. Cambridge, Mass.: The MIT Press.
Walmsley, R. 2009. World Prison Population List (eighth edition). International Centre for Prison Studies. http://www.prisonstudies.org/info/downloads/wppl-8th_41.pdf.
Zimmerman, M. J. 2011. The Immorality of Punishment. Peterborough, Ontario: Broadview Press