Can Punishment Be Justified?
Thomas Ash

Our practice of punishment seems to require justification. It involves society deliberately inflicting suffering on some of its members. When private individuals steal, or kidnap, or kill, we generally say this is wrong; but when a criminal is fined, imprisoned, or executed we deem it quite alright.

Though most people would endorse the punishment of criminals if asked, it is likely that they have not seriously considered the possibility that the amount of suffering inflicted counts so heavily against the practice that it is in fact wrong. A similar phenomenon can be observed in people's attitudes to war and other state undertakings: perhaps because no one individual has to 'do the dirty work' we fail to apply our normal strict standards. Or perhaps we are felt to be pushed into both these practices out of necessity, leaving us no choice. they cannot often express their reasons. Normally, you find several different threads of thinking tied together. Of course, it is perfectly sensible to feel that several of the traditionally self-sufficient philosophical justifications of punishment may be necessary. But we should be wary of the common mistake of assuming that since punishment 'must' be justified, any rationalisation that can be offered must be part of the truth. Our prisons, fines, and (in some places) electric chairs may not be justified at all, and we should subject all proposed rationales to critical scrutiny.

A traditional notion, revived in the 1970s, is that criminals should be punished because they deserve it. The central concept here is the familiar one of desert, central to notions of justice and 'just deserts.' But what does it mean to say that prison sentences and the like are deserved, and why are they? As Ted Honderich, a critic, points out, it cannot mean simply that they are right, as this begs the question. A hint as to the answer can be found in the name usually given to this tradition: retributivism.

The term 'retribution' implies that offences should be balanced out with equivalent penalties. The most famous, or perhaps notorious, example of this is the Babylonian lex talionis, adopted in the Old Testament: "An eye for an eye, and a tooth for a tooth."[2] Nowadays, we tend to think such direct equivalence barbaric, and it was even condemned by Jesus in the New Testament. Nonetheless, the principle is followed more or less completely in Saudi Arabia and other states with sharia law, and even Western justice embodies the similar principle, put so well by the Lord High Executioner in Gilbert and Sullivan's The Mikado: "Let the punishment fit the crime." Here is what seems like an intuitively correct standard: we should hand out sentences proportionate to the severity of the original crime. There is a clear notion of desert here, though as we shall see this might be justified externally.

However, a critic could still ask what it means for an offence to be "balanced out", and what an "equivalent" penalty is. As the familiar phrase goes, "two wrongs don't make a right", and if someone has been killed, no sentence can really alter that. What must be meant is that from society's viewpoint, crimes have to be punished to prevent criminals from enjoying their ill-gotten gains. This is a meaningful sense of "balancing out", as it may be assumed that criminals have benefited from breaking the law (or at least expected to, though this is - problematically - not quite the same thing.) People who choose to obey the law in the same situation presumably forego some satisfactions. An unorthodox critic could still claim that there's no use crying over spilt milk. Retributivists seem compelled to say that a burglar should be locked up, even if he is the last man on Earth. This seems odd: in whose eyes is anything being "balanced out?" Isn't this just to inflict pointless suffering.

There are some problems with applying retributivism in this mild, equalising sense. Firstly, it is often hard to see what an "equivalent" punishment is. If we have rejected the lex talionis's suggestion that a rapist should be raped, we are left with the impossible (and perhaps unhealthy?) determination of what period in jail is equivalent. In a clear-cut case, that of the pickpocket who steals $20 and causes no further distress, asking him to pay back $20 would be a wholly ineffective punishment, as he would have lost nothing. This suggests some practical, forward-looking factor may need to be considered in deciding sentences, and thus in justifying punishment.

The question of appropriate penalties aside, retributivist theories - unlike many others - tend to fit quite well with the law as it stands. Breaking the law can be seen as taking advantage where others choose not to, and as this is wrong in itself, quite aside from any external consequences, it deserves to be punished. The law is often seen as a contract which all of society signs. When someone makes a contract, we say they agree to incur some penalty if they go back on it. This is cited as another strength of retributivism: criminals accept their own punishment, and are not just treated as means to others' ends. But criminals obviously do not explicitly accept their punishment when it comes, nor can they really be said to accept it beforehand. The analogy given is that of our implicit agreement to pay when we ask a cabbie to go somewhere, but this is misleading in two ways. Firstly, career criminals often explicitly reject all laws, and see those who obey them as saps. Secondly, we do not have a choice of whether to live in a country with laws or not. Many laws in the past have been unjust, and it would be ludicrous to argue that slaves gave their tacit consent to slavery by living in societies where it was a central part.

A conclusion we can draw with some certainty is that a desire for retribution is at the heart of our notions of justice. As we shall see, other theories may play a necessary part, but none give so suitable a description. However, we may feel some discomfort at this. Honderich correctly identifies the desire of victims in particular and indirectly hurt society general to have grievances satisfied as central to any retributivism. This need not be unbridled vengeance, but it clearly flows from the same problematic source. It is interesting that Nietzsche, a critic of morality, explained punishment in terms of "feelings of anger and vengeance, in the desire to hurt those who have hurt us."[3] No wonder some modern retributivists deny that this is the source of their theories[4]. Yet the force they have does often derive from it. It is a recognisable fact of human nature that when we see someone getting ahead through unjust (but unspotted) means, we have a strong desire to see them "get what's coming to them" just for the sake of it.

Determinism is another potential problem for retributivism. This is a long-standing area of contention, with great philosophers on both sides, but it is fair to say that there are strong arguments for it. These include the fact that every event must either be random, or have a chain of causes stretching back to before the individual's life, and also neuroscientific observations that thoughts correspond to (theoretically predictable) neural events. Compatibilists like Hume and Hobbes have argued that we can still have free will in a determinist universe, so long as our actions are our own, and not forced upon us by anyone or anything. But, as Honderich has pointed out in several books[5], compatibilism does not let us keep our 'common sense' notions of free will unaltered. In particular, it makes it harder to judge someone as guilty, and thus deserving of punishment, if given his exact genes and environment (both of which were outside his control) we would have made the same choice.

Since the 19th century in particular, people have put forward what they felt were more enlightened justifications of punishment than simple retribution. Retributivism was seen (rightly) as a backwards-looking theory, hurting criminals back for what they'd done, and they felt (rightly) that this was not in itself different from the original hurt. So the theories of punishment they offered were characteristically forward-looking, justifying it in terms of its external deterrence value or reformative character. These are not vulnerable to determinism, and rest on no perplexing metaphysical notion of 'desert.'

Deterrence is a popular, and distinctly consequentialist, theory. Though we can't undo the original crime, through punishing it we can hope to deter future ones, and at least do some good, as well as of course keeping criminals off the streets. Now, if the assumption that punishment does deter is false, this falls at the first hurdle. The claim that people, especially those who go to prison, learn that "crime doesn't pay" does seem to have some evidential backing, with criminals committing fewer crimes (or at least getting caught less!) as time goes on. Bank robbers used to search each other for lethal weapons before each job - a prudent measure, given the hefty penalties they could have received. But many crimes are committed in moments of passion, when it is crazy to commit them from a purely prudential perspective. Deterrence cannot justify punishments here, or indeed for low-profile crimes no one ever hears about: as such, it is far from justifying ordinary criminal justice. However, even if deterrence did make all crimes less attractive, this might not be the best way in which to do so. It appeals purely to people's self-interest, and gives criminals no reason not to commit a crime if they can be sure they won't get caught, as a reform-based system might.

But this looks even less like a justification of punishment, as such. Why not just give a con man a persuasive sermon, or a lecture, or some therapy, or some drugs? If someone claims that a jail sentence (obviously not a fine - or execution!) would be more effective, they are just inviting incredulity. Our prisons seem to do a fine job promoting recidivism, although most informed people are aware that they expend scandalously little effort on rehabilitation. It is implausible that the infliction of suffering is effective at reform, unless it is taken to communicate disapproval. And most criminals are not very susceptible to tut-tutting. Recent evidence also suggests that psychiatrists' claims to be able to expertly mould people's personalities other than through persuasion were overblow. And punishment is not persuasion.

Both deterrence and reform theories have a distinctly paternalistic ring, justifying punishment because it is either for society's good or the criminal's own good. As such, they are sometimes criticised as treating criminals as means in some technocratic scheme, and not respecting them as rational individuals. This echoes Kant's famous injunction to treat human beings always as ends and never as means. A.C. Ewing's reformative defence of punishment[6] violates this by recommending what looks like a king of brainwashing: using punishment to habitually reinforce the wrongness of crime in the minds of both criminals and ordinary people. Reform is often praised as respecting offenders, rather than simply locking them away, but the way to do this would be through rational argument or treatment, not punishment. Brainwashing hardly respects criminals as autonomous agents, as Kant demanded that we should.

Both reformative and deterrence theories are open to a still more powerful objection: since they treat punishment as a way of reducing crime, without any reference to desert, they justify victimising the innocent, if this will do the job. It is no use pointing out that these are theories of punishment, which by definition applies only to the guilty. Since they make no reference to desert, by their own logic they provide no reason why victimisation is intrinsically wrong. There are plenty of reasons for thinking it is extrinsically wrong, though[7]. Those who need to be deterred, and reformed, are those who committed deliberate crimes in the first place. In practice, victimisation would undermine deterrence by detaching punishment from crime. It would set a dangerous precedent for judges, and would be likely to leak out. Also, the reason that we, as (presumably!) largely innocent people react negatively to the prospect of punishing an innocent man is that it would make all of us fearful of punishment, despite having chosen law-abiding lives, and would especially distress us as we would feel we had nothing to make ourselves targets.

But this cannot be the only reason. We can always imagine a case, unlikely ever to happen, in which either an innocent or a guilty man were executed, causing equal distress and the exact same deterrent effect (presumably as everyone thought they were both guilty.) In this case, we would still prefer that the innocent man go free, and this can only because some notion of retribution and 'desert' is part of our common notion of justice. Only this is sufficient to explain our revulsion at victimisation. Some element of retribution, especially desert, is then a part of our common notion of justice - this explains our revulsion at victimisation.

But is this one of those 'common sense' beliefs that we should treat as irrational, like past beliefs in nationalism and hereditary rule? I tend to think that victimisation is wrong because of the harmful consequences for the victim, not because of any metaphysical notion of desert. Suffering is always an evil, whether someone is a burglar or a saint or an insurance salesman. We do in fact accept punishment of the innocent, sometimes in quite serious cases, with our systems of strict and vicarious liability, and exemplary punishment. A shopkeeper whose clerk puts adulterated milk on the shelves, in complete ignorance and with all reasonable precautions in place, is just as much an innocent man as the victim of a mistaken lynching. A Danish judge who gives someone an exceptionally tough sentence "to send a message" is acting in just the same way as the Saudi Arabian mutawa'een (religious police) chopping off a thief's hand. The reason we particularly object to the latter is that the punishment goes so far beyond usual notions of desert that it causes too much distress. So the general wrongness of victimization is explicable by its bad consequences.

If you feel that punishment is unjustified in these circumstances, another set of crimes pose an even greater difficulty for all three traditional justifications. These are those crimes, like pedophilia, that we feel the perpetrators cannot help, because they are in the grip of some mental disease. Nonetheless, these offences are treated as crimes as abhorrent and worthy of punishment as murder, and certainly unleash vengeful feelings like no others. Unfortunately, it is widely accepted that you cannot deter or reform people suffering from conditions such as pedophilia. Retribution, on the face of it, might seem to be as applicable (or inapplicable) in this case as any other. But P.F. Strawson pointed out that, though feelings of resentment and a desire for revenge are natural in any situation where we are wronged, we learn to moderate them in cases where the perpetrator could not have helped what he or she did. None of this is to argue that we should not lock up pedophiles to protect children from them - and them from vigilantes - but this cannot be a case where the infliction of “hard treatment” is justified.

The last thing to consider is the ambiguous idea of justice. It is defined by the Cambridge Dictionary of Philosophy as "each getting what he or she is due", which suggests a notion of desert. But it could also mean that, in Bentham's phrase, "everybody to count for one, and nobody for more than one."[8] This is itself an ambiguous notion. It could simply mean that everyone's interests are equal in the hedonic calculus. Or it could be seen as pointing to something like Ted Honderich's "Principle of Equality." This is that our punishments, while preventing offences, should also restore the equality of welfare and distress they disrupted. Equality is taken by many as a good thing, although it is unclear whether this is because it is the most efficient distribution or because it is simply taken to be good in itself (both could be doubted.) Of course, as Honderich recognizes, our societies have nothing like this kind of equality, and as criminals often come from the least privileged groups, the legal system can reinforce this. This leads him to suggest that punishment can only be justified in the context of a broader political effort to ensure equality.

In its current from, justice isn't justified. This shouldn't be surprising: most attempts to explain pre-existing instincts and institutions - as opposed to working out with an open mind what can be justified - will lead to a similar conclusion. The ideas of retribution and desert are inextricable from the kind of punishment we have today, but they are fatally undermined by determinism and the fact that they boil down to inflicting suffering to satisfy the instinct of revenge. But a more enlightened version of justice and punishment, involving deterrence, reform and a concern for what kind of society we want, could be justified.


1. Crime and punishment, R.A. Duff, Routledge Encyclopedia of Philosophy

2. The lex talionis was originally found in Hammurabi's Code

3. The Moral Philosophers, Richard Norman, p136

4. Walker and Murphy have claimed it reduces retributivism to "[sentencing] because the people outside in the street are baying for blood" and "just a primitive bit of intuitive vindictiveness" respectively. Where they do have a point is that Honderich's account of retributivism as aiming to satisfy the victim's grievances is fundamentally consequentialist, and not really a desert-based theory at all.

5. Including a section in Punishment: The Supposed Justifications, A Theory of Determinism Vols 1 & 2 and a popular exposition of that theory in How Free Are You?

6. Expressed in The Morality of Punishment (London, 1929)

7. As Honderich points out, again in Punishment: The Supposed Justifications.

8. Attributed to Jeremy Bentham by JS Mill in Utilitarianism, chapter 5 [near end]

Sources of further study on this topic

Criminology programs teach you more about crime and punishment.

© 2003

About the author...
Thomas Ash is the webmaster of Big Issue Ground and Atheist Ground. He is a graduate student studying philosophy at Merton College in the University of Oxford. He received his BA in the subject from the University of Cambridge, where he was president of the University Atheist and Agnostic Society. As you may have guessed, he is an atheist - a positive one in fact. More information can be found here.

Feedback on the site, or on this essay, is always welcome. If it's complimentary, you may also want to put it in the Guestbook, for all to see!


Essays on related topics...

Title Author
Can Punishment Be Justified? Thomas Ash
For A Pro-Life Death Penalty Jim Blair
Do Murderers Deserve Death? Steve Kangas [off site]

Number of unique visits to Big Issue Ground so far...

visits so far...