Punishment versus restitution: A formulation

Kinds of coercion

How should criminals be treated in a libertarian polity? Is it permissible to punish them? Why or why not? In what follows I’d like to outline the answers I personally have reached to these questions, stressing that I speak only for myself, and would be happy to receive comments and criticism.

Let’s define coercion as the forcible subjection, actual or threatened, of the person or property of another to one’s own uses, without that other’s consent. In light of this definition, it is possible to distinguish three kinds of coercion:

• Defensive coercion: I use coercion against you, but only to the extent necessary to end your aggression against me (or someone I legitimately represent).
• Retaliatory coercion: I use coercion against you, but while you are aggressing against me (or someone I legitimately represent), my coercion exceeds the extent necessary to end such aggression on your part.
• Initiatory coercion (or aggression): I use coercion against you, although you are not using coercion against me (or anyone I legitimately represent).

The justification of coercion

Given these definitions, there are four possible positions one might take on the justification of coercion:

1. Coercion is never justified.
2. Defensive coercion is justified, but retaliatory and initiatory coercion are not.
3. Defensive and retaliatory coercion are justified, but initiatory coercion is not.
4. Defensive, retaliatory, and initiatory coercion are all justified.

The libertarian principle of non-aggression clearly rules out option (4): The initiation of coercion is not permissible. But each of the remaining three options appears to be compatible with libertarianism’s ban on aggression; and in fact each option has some libertarian defenders.

But while options (1) through (3) may all be compatible with the letter of libertarianism, it does not follow that they are also all equally compatible with its spirit. Indeed, I wish to argue that (2), and only (2), expresses the spirit of libertarianism, and consequently that libertarian principles, properly understood, permit self-defence against the criminal, but prohibit his or her punishment (at least in any ordinary sense of punishment).

Why not pacifism?

Option (1) might initially seem the most attractive. If coercion is evil, why not forswear it altogether? Doesn’t the use of violence, even in self-defence, reduce the victim to the moral level of the aggressor? For many people, radical pacifism resonates with such deep religious and cultural values as Christ’s advice to turn the other cheek, the doctrine of ahimsa (non-violence) in many Indian religions, and even the familiar maxim that “two wrongs don’t make a right.” Further, we admire such preachers and practitioners of non-violence as William Penn, Tolstoy, Thoreau, Garrison, Gandhi, and Martin Luther King. Within the libertarian movement itself, many have been inspired by the writings and personal example of anarcho-pacifist Robert LeFevre.

Yet from a libertarian point of view, there seem to be drawbacks to the radical pacifist position. Libertarians see themselves as defenders of rights; but the difference between rights and other sorts of moral claims lies in the fact that rights are legitimately enforceable. Suppose I gratuitously insult one of my neighbours, and steal the car of another. Each of my neighbours thereby acquires a moral claim against me: The first neighbour has a claim to an apology, the second a claim to his car. But only the second claim can properly be described as a right. My second neighbour has a right to have his car returned, and may legitimately use coercion to enforce his claim. But my first neighbour has no right to an apology; I ought to apologise, but no one may legitimately force me to do so. Not all sins are crimes.

But if coercion is never justified, even against aggressors, then the distinction between the two sorts of moral claim vanishes: Neither of my unfortunate neighbours may legitimately use coercion to enforce his claim against me. But what makes a moral claim a right rather than something else is precisely the fact that coercion may be used to enforce it. Whoever endorses radical pacifism, then, is committed to denying that anyone has any rights — a rather odd position for a libertarian to be in!

Spheres of authority

Libertarians generally see rights as setting boundaries around people. Each person has a sphere of authority within which they may do as they please, without external interference; but they may not cross beyond their own boundary and engage in actions within someone else’s sphere of authority (except with that person’s permission). The classic expression of this idea is: “Your right to swing your fist ends where my nose begins”.

The libertarian pacifist, however, can no longer consistently endorse this picture. Suppose I grab onto your nose, thus invading your sphere of authority. Before I did so, you were free to scratch your nose whenever you liked. But now that I have a firm grip on your nose, you cannot scratch your nose without first knocking my hand away. Yet if the radical pacifists are right, it would be immoral for you to knock my hand away. It follows that, under those circumstances, it is no longer morally permissible for you to scratch your nose. Through no fault of your own, because of my immoral action, your own nose is now no longer within your sphere of authority.

But this seems unfair. Why should my aggression be allowed to constrict your domain of legitimate activity? Why should my grabbing your nose make you lose your rights over it? It seems more in accordance with the libertarian conception of justice to say that by grabbing your nose I have put myself into your sphere of authority, rather than taking your nose out of it; and as a result, you can now coerce me without exceeding your just authority. What’s wrong with initiatory coercion is that it exceeds the bounds of the coercer’s sphere of authority, but defensive coercion does not exceed those bounds, and so is legitimate.

In the case of third-party intervention, defensive coercion is justified to the extent that the intervener is acting as the victim’s agent. Presumably, this involves acting with the victim’s actual authorisation, when the victim is able to give or withhold consent, or else acting as the victim would authorise (so far as can be determined), when the victim is for one reason or another unable to give or withhold consent-perhaps because of unconsciousness, infancy, mental illness, or simply pressure of time. For example, suppose I attack you, and Martina, acting as your agent, intervenes to defend you. By invading your boundary, I have put myself under your authority. You may exercise this authority directly; but you are equally within your rights in exercising it vicariously, through your agent Martina..

But the argument that justifies defensive coercion does not justify retaliatory coercion. If I use more coercion against you than is necessary to end your aggression against me, then in effect I am going beyond merely exercising my legitimate authority within my own sphere. If each person’s freedom may be justly limited only by the equal freedom of others, what could justify me in limiting your freedom by more than is necessary to restore my own?

What counts as defence?

I have argued that the principle most consistent with the spirit of libertarianism endorses defensive coercion, but prohibits not only initiatory but also retaliatory coercion. But how much latitude does this restriction allow us? What does this principle allow us to do to criminals, beyond fighting them off at the moment of the aggression?

First of all, remember that we defined coercion as the forcible subjection, actual or threatened. of the person or property of another without that other’s consent. If I come running. toward you brandishing a sword, you need not wait until I actually cut you before taking defensive measures. By manifesting a murderous intent toward you, I have already placed myself under your authority. Hence it is permissible to imprison or exile criminals, to the extent that they pose a continuing danger to the innocent.

Requiring the criminal to pay compensation to the victim can also be justified on the defensive grounds. Consider the following three cases.

1. I break into your house. Here I am clearly trespassing on your property, and you have the right to use coercion to get me to leave, since your home falls within your sphere of authority.
2. I break into your house, and slip your radio into my knapsack. In this case, you may do more against me than simply kicking me out of your house, because I, by retaining an item of your property on my person, have failed to vacate your sphere of authority. Hence you may use coercion to get the radio back. I remain under your authority until you recover your property.
3. I break into your house, and smash your radio with a hammer. The fact that your radio no longer exists does not alter the fact that I remain under your authority until the radio (or its equivalent in value) is restored to you. Thus I may legitimately be coerced into compensating you for your loss.

Note that this justification of defensive coercion has nothing to do with the aggressor’s responsibility for his or her actions. If I have been hypnotised into attacking you, you still have the right to fight me off. If a wind blew me onto your property against my will, you still have the right to remove me. And likewise, if I accidentally destroy your property, I still owe you compensation. What matters is that I have entered your sphere of authority and so may be coerced into leaving it; whether I got into your sphere voluntarily or involuntarily is irrelevant. Thus it seems to me that a libertarian concept of rights favours a strict-liability approach: That is, people are liable for the damage they cause, regardless of whether they caused that damage deliberately or accidentally.

Revenge or restitution?

Critics of the position I’m defending often ask the following question: “What if someone you loved were murdered? Would you be content with seeing that the murderer paid you back and was locked up, or would you want to see the murderer dead?” In my case, at least, the answer is: Yes, I’d want to kill the murderer. And I might be justified in doing so if the government released the murderer while he still posed a danger to others — for then my action might count as defensive rather than retaliatory (whatever my motivations might be). But suppose I know the government isn’t going to release the murderer. Even then, the desire to take revenge by slaying the slayer is a perfectly natural and forgivable reaction. But should the vengeful emotions of victims and their loved ones replace rational analysis as the foundation of jurisprudence? Important questions of rights and justice should not be decided in the heat of anger.

I should point out, incidentally, that under a restitution-based system, victims who did take revenge would not be treated harshly. After all, retaliatory coercion is not permissible against them either. If you kill me because I killed your loved one, then you would be required to pay compensation to my next of kin (though as Randy Barnett, another critic of retaliatory coercion, has pointed out, you could discharge this obligation by handing back to my next of kin, as compensation for my death, the very same money I originally gave you as compensation for your loved one’s death); but you would probably not be locked up, because you do not appear to pose a threat to others. Hence the danger of revenge by the next of kin would pose a serious deterrent to murder. As Louis XV of France said to one of his courtiers who had committed a murder: “I grant you pardon — but I also pardon whoever will kill you”. (In the case of murder victims without a next of kin, the right to compensation might be granted to whoever “homesteaded” it by pursuing the matter in court.)

Might such a system encourage a cycle of revenge, with each side in a dispute claiming a life in return for the life taken previously by the other side, as the two sides pass the same compensation money back and forth? Such does not seem to be the lesson of history. In the stateless societies of pre-Christian Northern Europe, the blood-feud was originally a pervasive feature of social life; but this began to change with the institution of a restitutive system. Wergeld (mangold; that is, monetary compensation for a human life taken) gave feuding parties a powerful financial incentive to accept compensation and end the cycle of revenge,.rather than taking another life. If such economic motives could tame the blood-feud in a society that glorified revenge as a matter of honour, a restitutive system should be even more successful in a society like ours, which at least pays lip service to the condemnation of revenge.

The limits of defence

Are there limits to what one may do to an aggressor in order to enforce one’s rights? For example, if you swallow something belonging to me, does that give me the right to cut you open in order to retrieve my property?

As I have written elsewhere: “The dilemma here is analogous to that in Shakespeare’s The Merchant of Venice. Antonio owes Shylock a pound of his flesh, but he does not owe Shylock any blood; yet of course Shylock cannot claim his pound of Antonio’s flesh without taking some of Antonio’s blood as well. Portia reasons that since Antonio does not owe Shylock any blood, Shylock has no right to take Antonio’s blood, and so has no right to take the pound of flesh either (since, although he has a right to the pound of flesh, he cannot exercise that right without doing something he has no right to do — namely, take Antonio’s blood). Shylock, on the other hand, reasons that since Antonio does owe Shylock the flesh, Shylock has the right to take it, and so has the right to take the blood too (since that follows from something Shylock has a right to).” (Roderick T Long, “Abortion, Abandonment, and Positive Rights: The Limits of Compulsory Altruism,” Social Philosophy and Policy, vol. 10, no. 1 (1993), p. 174.)

By analogy, we can distinguish two possible principles to govern defensive coercion:

1. Shylock’s Principle: If I aggress against you, you have the right to coerce me in whatever way is necessary to remove me from your sphere of authority.
2. Portia’s Principle: You have no right to coerce me, even if doing so is necessary to remove me from your sphere of authority.

Portia’s Principle is simply pacifism again, and I’ve already argued that pacifism is at odds with the spirit of libertarianism; so we can reject Portia’s Principle. But what about Shylock’s Principle? Although strictly speaking it licenses only defensive coercion, not retaliatory coercion, Shylock’s principle nevertheless strikes me as defective in much the same way as retaliatory coercion. If Shylock’s Principle is correct, then I would be justified in shooting a toddler if that were the only way to prevent the toddler from treading on my toe. But such a response would surely be disproportionate to the seriousness of the aggression.

To my mind, then, the most libertarian approach to the legitimation of defensive coercion is the following one:

3. Principle of Proportion: If I aggress against you, you have the right to coerce me in whatever way is necessary to remove me from your sphere of authority, so long as your coercion is not disproportionate to the seriousness of my aggression.

Thus not even all defensive coercion is automatically justified. Coercion, to be legitimate, must pass three tests: First, it must be a response to aggression on the part of someone else; second, it must be necessary to end or prevent that aggression; and third, it must be proportionate to the seriousness of the aggression.

Let me guard against a possible misinterpretation of this principle. It might seem that if the defensive response must be proportionate to the threat, then we can never be justified in using greater force than our aggressor (for example, killing someone to prevent them from inflicting serious but not fatal harm on us). I think that would be a mistaken inference. An aggressive killing is worse than a defensive killing. Hence aggression need not be fatal in order for deadly force to be a proportionate defensive response to it.

Crime without punishment

I’ve argued, in effect, that we may legitimately impose fines or prison sentences on criminals. So haven’t I justified punishment? I don’t think so. The point of punishment is to impose some sort of suffering on the criminal — either for retributive reasons (the criminal deserves to suffer) or for reasons of deterrence (other criminals will be discouraged when they see how we make criminals suffer). On the defensive model, while fines and incarceration may in fact cause suffering to the criminal, that is not their aim; rather, the aim is to restore the victim’s rights. After all, we do not think that those who violate others’ rights accidentally should be made to suffer; but the only difference between a willing aggressor and an accidental aggressor lies in the contents of their thoughts — a matter over which the law has no legitimate jurisdiction. Hence, I conclude, we may not legitimately treat willing criminals any differently from accidental criminals (except to the extent that they differ in likelihood of posing a continuing threat). Mandatory restitution to the victim is justified on libertarian grounds as an expression of defensive coercion; but punishment, I believe, constitutes not defensive but retaliatory coercion, and so is not permissible.