I think it comes down to a question of guilt or innocence. A criminal has committed a major sin, and hence deserves a major punishment: an eye for an eye, a tooth for a tooth, and a life for a life. Even just an ordinary adult will have some track-record of sin behind them -- none of us are perfect. They might not quite be evil enough to deserve to be targetted directly, but nevertheless it wouldn't be such a terrible thing if they were to become the victims of collateral damage in war. But an unborn baby, having had no opportunity to sin, is completely and utterly innocent, an unblemished soul, and consequently of greater moral worth.

As far as I can discern, that's roughly the idea that those fundamentalists have. Speaking for myself, I regard this attitude as wholly abhorrent, both antithetical to the teachings of Jesus and morally repugnant in itself. But, hey, that's just my opinion, and what do I know?

I agree that standard utilitarians would find it hard to justify retribution as such, that is, without appeal to such further effects as you mention: deterrence and satisfaction to injured parties. But I do not agree that this leaves many punishments without a utilitarian justification. The problem you see is that "it is easy to conceive of scenarios in which punishment does not act (1) as a deterrent to crime or (2) to relieve the suffering of any injured parties."

In fact, I believe, this is not so easy. Let's start with (2). Suppose someone murders a man who was a bit of a loner and not much liked by the few people who knew him. He won't be missed. So it may seem that punishing the murderer won't give satisfaction to any injured parties because no one is really injured by the man's demise. But this overlooks that there are others who have lost a loved one to murder and others on whose life an attempt has been made. They, too, are injured parties -- not injured, to be sure, by the murderer of our man, but injured by others with murderous intent. And they are likely to get satisfaction from the fact that murderers get caught and punished appropriately. Or so the utilitarian might say in response to you.

Somewhat similar thoughts apply to your (1). Here you may have in mind, perhaps, very rare crimes that virtually no one has any inclination to commit. It may seem that punishing such a crime would not deter anyone. But now imagine that we accepted your argument and stopped punishing crimes that virtually no one has any inclination to commit. Then those who have such very rare inclinations would take notice and feel much freer to follow them. The utilitarian response then is that, by punishing a crime that virtually no one would want to commit we are deterring not merely crimes of this kind, but also all the many other kinds of crimes that almost no one has any inclination to commit.

Perhaps you had another thought in the deterrence case. Perhaps you thought of crimes whose incidence would not be affected by the prospect of punishment. The law does recognize such cases, and does in fact offer leniency when the accused is shown to have been temporarily insane, for example. Utilitarianism would endorse such leniency, but with caution: there is a genuine danger that such leniency can have deleterious effects on deterrence by encouraging prospective criminals to believe that, with some advance planning and a good lawyer, they'll probably be able later in court to create some reasonable doubt about their sanity at the time of their crime.

In conclusion, I do not think that there really are many scenarios of pure retribution where punishment neither strengthens deterrence of crime nor gives satisfaction to any injured parties. With respect to the problem you pose, utilitarianism can account pretty well for our punitive practices. Utilitarianism does poorly, I think, in accounting for the severity of the punishments we mete out. Utilitarians would impose severe punishments when the extra harm to the criminal brings a larger reduction in the harm done by crime. Utilitarians might then impose a very severe punishment for a very slight crime, so long as this punishment has a dramatic impact on the incidence of this crime. An extreme example: imposing the death penalty for not wearing seat belts might cause a good number of executions annually. But it would also terrify millions of current scofflaws into reliably wearing their seat belts, thus saving perhaps many more persons than get executed from dying in accidents. If this were so, utilitarians might approve the death penalty in this case, while it would strike the rest of us as insanely disproportionate.

My impulse is to say that we're mixing apples and eggnog. It's true that retribution is part of the way we think about punishment. But however we understand retribution, it's hard to see how the State's wrong against you would make it okay for you to rob me or beat me up. After all, even though I'm part of the body politic, qua private citizen, I don't represent the state. And in any case, robbery, embezzlement, assault and various sorts of mayhem are the kinds of things we should shy away from. Giving someone a free pass on future misdeeds because of past mistreatment by the State seems to miss that point.

If Mr. Tankleff is indeed found not guilty, then it's not unreasonable to think that he should receive some sort of compensation. But 17 years (or even 17 days) of punishment-free crime doesn't seem like the right way to go. Things being what they are (no time machines, no guaranteed life-extending potions...) monetary compensation sounds like a much better idea, even though it can't make up for 17 lost years.

The word "right" is generally used more broadly than this, so that rights may give way when a lot is at stake. Some philosophical literature may suggest otherwise -- people talk of rights as "side constraints" or "trumps" -- but when you look more closely, they too agree that most rights should be understood as giving way at some point (though for Nozick this point comes rather late: when there is the threat of "catastrophic moral horror").

How much needs to be at stake, and for whom, are matters that get built into the content of the right. Thus, in some jurisdictions property rights are quite strong (property can be expropriated only for an overwhelmingly important purpose) whereas in other jurisdictions property rights are much weaker. Even in the latter jurisdictions, the word "right" is not out of place, so long as an expropriation is based on more than just a showing that it would be better on the whole for this property right to be infringed. You do not have a property right in your car if you must compete on equal terms with others for its use (e.g., show that you need it today more than they do).

Most rights, both moral and legal, are then of this form: When you have a right to X, then no one is permitted to deprive you of X unless this is done for a purpose of such-and-such kind and importance, and unless compensation of such-and-such kind is offered.

Now let's apply all this to the right that interests you, the right not to be punished unless one is guilty. Here deliberate infringement is rarely at issue (though it may be for a judge who considers sending an innocent person to jail to scare off would-be law-breakers). Rather, the problem is probabilistic infringement, running a risk that we'll be infringing the right in question. A general prohibition on ever running any such risk would be very constraining -- you could never drive a car, for instance, because you're thereby subjecting others to a tiny risk of a right-infringing physical injury. Must we minimize the risk, by reducing the speed limit "as low as possible", for instance? I don't think so. But we must keep the probability below certain thresholds that are determined in part by the cost of lowering these thresholds even further.

In the punishment case, one cost is financial, the cost of sophisticated evidence collection, trials, and so on. But another very important cost is crime: As standards of evidence are raised, more guilty go free, possibly to commit another crime, and more people are insufficiently deterred from choosing a career in crime. Now in accordance with the above analysis, one enjoys a right not to be punished unless one is guilty only if one's interest in not being so punished is given greater weight than the interests of potential crime victims. This is suggested in the famous formulation by William Blackstone: “It is better that ten guilty persons escape than one innocent suffer.” Put more bluntly, it is better that ten innocent persons be victimized by criminals than that one innocent person be officially victimized by us. Now what the correct ratio is here, the ratio at which no strengthening of the protections of the accused are thought to be required by their right, this is a matter of dispute, and philosophy can hardly give a precise answer (such as "8.39:1"). What philosophers can say is that people enjoy something worthy of the name "right" only when this ratio is a good bit above 1:1. What empirical studies can add is that some groups in our society have not enjoyed even the right you speak of in even the weakest sense of the word. In many states, criminal justice has been practiced in ways that foreseeably led to huge numbers of African Americans being innocently convicted and very severely punished, and these wrongs could have been very largely avoided through reforms of these practices that would have had a negligible impact on crime rates or none at all.

A good question! As it turns out, not everyone agrees that the purpose of punishment is to reform people. In fact, some philosophers (Kant is perhaps the foremost) held that the only justification for punishment is that the person deserves it, and if we punish for the sake of making someone better, we fail to show proper respect for them -- we manipulate them for our own ends.

Setting the question of Hell aside for the moment, we can see that the argument you raise, if correct, would also count against capital punishment. But whatever one's views on the rightness of capital punishment, the widespread support it has in some places makes clear that many people see punishment as a matter of giving people what they deserve rather than reforming them. This idea, fleshed oout and elaborated, is often called the retributive theory of punishment, though it's important not to confuse retribution with revenge. Retributive theorists would maintain that the punishment must always be proportional to the crime, and so excessive punishment -- say, sentencing a petty thief to 20 years in prison -- would be wrong. Kant went further. He insisted that even in executing a murderer, we must respect his humanity. "His death... must be kept free from all maltreatment that would make the humanity suffering in his person loathsome or abominable," Kant wrote. Whether executing someone is really consistent with respecting his humanity is open to debate, but Kant insisted that justice demands that the murderer's punishment must fit his crime, and hence nothing short of execution will do. Whether we agree or not, what's clear is that the aim of the punishment here is not to make someone a "better person" but to serve a particular conception of justice.

But now we can return to your worry about hell, and what we see is that no matter what our view of punishment, it's hard to understand how anything a finite human being does could earn them a never-ending stay in Hell. In fact, many religious people agree. Not all believers think that everyone outside of those who confess the proper doctrine are destined for an eternity of torment. Those who do have a conception of punishment that reformers and retributivists alike are bound to find puzzling and repugnant.

That parental discipline may constitute a form of abuse depends entirely on what you accept under the label of "discipline". Consider for example a family in which following some religious practices - like preying before supper, or not eating certain kinds of food - is considered as part of a discipline that children are obliged to follow, and a 10 years old child (that is, someone who is cognitively able to take at least partly autonomous decisions on her moral preferences, even is she still doesn't have reached the institutionally established "age of reason", usually 16, 18 or 21 according to the countries) who refuses to comply. In this case, I would consider a sanction of her behaviour as a form of abuse. Punishing her for not complying to a rule she doesn't want to endorse because she finds it incompatible with her ideas and moral feelings is a form of abuse.

Abusing children means prescribing them a system of rules of disciplines without taking their stance and thinking about what is reasonable to accept now and in their future and what is questionable from their point of view now and in their future. Of course it is very difficult to evaluate what are the cases in which we have the right to act in an authoritative way with our children and cases in which we haven't this right, because given the asymmetry in experience, cognition and moral development between our children and us, very often we must force them to comply with practices and norms whose interest and value they cannot immediately understand.

One may argue that paternalism - that is, restricting the freedoms of dependants in what it is claimed to be their own interest - could serve as an appropriate moral test for parental discipline. But I don't think it suffices. In many cases we have the right to decide on behalf of our children in our own interest: we may for example proscribe some of children's actions or behaviours by appealing to the respect they owe us, that is, to the regard they should have to our way of living. So paternalism doesn't suffice to discriminate between legitimate and illegitimate coercion that we exerce on them.

A more plausible moral test for discipline would be the following: we can legitimately impose children to comply with some rules and practices if we consider our attitude towards the values these practices help to promote and the reasons why we endorse them. If the values these practices help to promote (hygiene, respect, instruction etc...) are values that we share and accept on a reasonable basis and not simply on authority, and furthermore we think we will be able in the future to articulate a reasonable explanation to our grown up children on why we decided to force them in such a way when we had control on their lives, then we can legitimately impose a discipline.

A different question is when punishment for not complying with parental discipline can be considered an abuse because, for example, it is too severe. Here I think we should take a long term perspective and balance the corrective effects of the punishment on education with long term effects of the punishment on the emotional and moral development of the child. For example, public humiliation, a form of punishment that was very common in schools and families, should be avoided, because its long term consequences may be extremely negative, whereas any form of punishment that involves taking the responsibility of an action (like for example learning to admit a fault), can have positive consequences on the way the child will conduct herself in the future. It will be useful also in this case to take the children's stance and try to "simulate" their understanding of the causal path which goes from their action to the punishment.

It is true that, once a person has been executed, she is no longer around to suffer the loss of years she might otherwise have lived. But the point of an execution is not to punish the person after she's dead, but before. She is subjected to the experience of living on death row and later to the experience of being killed in the execution chamber; and she must expect all along that many things she cared for are less likely to thrive or to come to fruition.

You might respond that this answer works only for people who know about their impending execution. What about someone who is killed painlessly in her sleep? Could this ever be construed as a punishment?

We can give an affirmative answer if we think of punishment in a somewhat extended sense as the setting back of a person's interests. Suppose you have given offense to someone and, in order to punish you, he has been embezzling money from your account. Being an affluent entrepreneur, you never notice the losses (you rather take your business to be less successful than it really is). But, nonetheless, there is a sense in which he really succeeded in punishing you. Your wealth is something you really care a great deal about, and it has really been substantially dimished over time below what it would be if he had not siphoned off money from your business account. Similarly, you might be punished by someone who is spreading false rumors about you that damage your reputation among the people who know you -- even if no one confronts you with these rumors and you thus remain ignorant of how your reputation has been gravely damaged.

Once we allow that there can be plausible cases of unexperienced punishments, then it may also be plausible to say that someone is punished after her death by setting back interests that were important to her. (After all, if the punishment is not experienced anyway, why should it matter whether the punished person is capable of such an experience at the time of the punishment?) By destroying the reputation of someone after her death, one can reduce the value of her life as she thought of it -- and likewise by destroying her best artwork (which she had ardently hoped would be admired by posterity).

Of course, this is stretching the ordinary notion of punishment a bit, but not, I believe, beyond recognition.

Of course, murder is not a victimless crime! But how can that be, Alex asks, if the victim no longer exists in order to suffer the harm that has been done to him? If you must exist in order to have interests, then how can a dead person’s interests suffer as a result of his death?

To see the harm that is suffered by a murder victim, let’s think first about what it means to be harmed. If I were to harm Harry, what sort of thing would I have to do him? Intuitively, when I harm Harry, my actions make him worse off than he would have been had I not acted as I did. So when I spread vicious gossip about Harry, I have harmed him because, had I not spread the vicious gossip, his reputation would have been intact, and he would have been well-respected in his community, loved by his family, and able to complete more easily certain projects about which he cares deeply, projects that require the good will and cooperation of others. Because of my vicious gossip, Harry is now a social outcast, unloved and unaided.

So let’s try out this definition of harm:

But now, it seems, we have a problem. If I kill Harry, how can we compare the state that Harry would have been in, had I not killed him, to the state that he’s now in, namely, dead? Since he is dead (and we’ll suppose, non-existent), he’s in no state at all. How can we compare this "non-state" to his state he would have been in had he been alive?

The answer to this puzzle, I think, is this. If Harry had survived, he would have attained all of the goods that generally come with living– pleasure, deep relationships with others, philosophical knowledge . . . (complete this list with whatever you count as genuine goods). Of course, had he lived, it’s likely that he would have had some hard times, too– some pain, frustration, heart-break, and so forth. But so long as his life would have been worth living for him, the goods that he would have had, had he survived, would have outweighed the bads that he would have had, had he survived. When I kill Harry, I prevent him from attaining these goods.

When we attempt to figure out the harm that Harry has suffered when I kill him, we should not compare Harry’s state after his death to the state that Harry would have been in, had I not killed him: for the reasons that I give above, such a comparison is impossible. Instead, when we attempt to figure out the harm that Harry has suffered when I kill him, we should compare the totality of goods that Harry would have had over the entirety of his life, had I not killed him, to the totality of goods that he had actually attained in the life that I cut off. If his life would have been worth living, then I did indeed harm Harry when I killed him: I deprived him of all of the goods that he would otherwise have had, had I not killed him.