1 Some philosophers have argued that we should apply the deterrence justification to the institution of punishment and the retributive justification to particular acts within this institution; for classic statements of this “mixed” view see Rawls, John, “Two Concepts of Rules”, Philosophical Review 64 (1955), 3–32CrossRefGoogle Scholar, and Hart, H. L. A., “Prolegomenon to the Principles of Punishment”, in his Punishment and Responsibility (Oxford: The Clarendon Press, 1968), 1–13Google Scholar. But these arguments seem to me to rely on a dubious distinction between an institution and the acts of which it is composed. The rights-based justification I will defend has many of the same attractive consequences as this mixed view without relying on its dubious assumptions about institutions.

2 Locke, John, Two Treatises of Government, ed. Laslett, Peter (2nd ed.; New York: Cambridge University Press, 1967).Google Scholar

3 Nozick, Robert, Anarchy, State, and Utopia (New York: Basic Books. 1974). 10, 137–142.Google Scholar

4 On this see Hart, H. L. A., “Are There Any Natural Rights?”, Philosophical Review 64 (1955), 175–191.CrossRefGoogle Scholar

5 Enforcement rights are often said to include not only a right of self- (and other-) defence and a right to punish but also a right to exact compensation, where this right to exact compensation is sometimes exercised alongside the right to punish and sometimes exercised when it would be wrong to punish. The right to exact compensation, however, need not be regarded as a separate enforcement right. If we say that alongside their ordinary rights persons have more complex rights not to be harmed without compensation being paid them afterwards, we can say that exacting compensation prevents the violation of these rights in exactly the same way that self-defence prevents the violation of simpler rights. Why compensation sometimes can and sometimes cannot be accompanied by punishment—and what distinguishes the two cases—is an extremely difficult question, which lies outside the scope of this paper. For a discussion see Nozick, , Anarchy, State and Utopia, 57–73.Google Scholar

6 A libertarian rights theory of this kind is presented in Kant, Immanuel, The Metaphysical Elements of Justice: Part I of the Metaphysics of Morals, trans. Ladd, John (Indianapolis: The Liberal Arts Press, 1965)Google Scholar, and discussed by Hart in “Are There Any Natural Rights?”.

7 Those who do not believe there is a right not to be tickled as such can imagine that X is trying to violate all the (fairly unimportant) rights that he would violate if he tickled Y without his consent.

8 Anarchy, State, and Utopia, 62–63.Google Scholar

9 In Anarchy, State, and Utopia, 62Google Scholar, Nozick gives a different version of this upper limit qualification, using the notion of harm, and saying that the upper limit on the harm we can inflict on an attacker is some function f(H) of the harm H which he is threatening to visit on us, where f(H)>H, or at least f(H)≥ H. If the notion of harm which Nozick is using here is the ordinary utilitarian notion (as it certainly seems to be—see especially the remarks on 58 and 75) then I am not sure that it is the appropriate one to be using at this point in a rights theory, and I am not sure that, if it is used, it will always yield the right results. I will argue below that the right not to be physically assaulted is not less important than the property right which a person has who owns a weapon, and that as a result my version of the upper limit qualification always allows the victim of an assault to destroy his attacker's weapon if that is the only way he can prevent himself from being physically beaten. Nozick's version, however, does not always allow this. Whether I can destroy my attacker's weapon or not depends at least in part on how much he will be harmed by its destruction, if my attacker is not going to do me a very great harm (I already have one broken arm so another will not be much of an added inconvenience), and if he is very fond of his weapon (it is a family heirloom and its destruction will cause him untold misery), then Nozick's version of the upper limit qualification says I cannot destroy his weapon to prevent him from breaking my arm. This is surely not what a libertarian rights theory ought to say.

10 Nozick discusses a (different) version of this minimum necessary qualfiication in his “Moral Complications and Moral Structures”, Natural Law Forum 13 (1968), 1–50Google Scholar. See especially the discussion of Principle VII.

11 Analogues of these two qualifications as they apply to self-defence are recognized in the legal systems of Canada, Britain and the United States, though they are not recognized in Germany or the Soviet Union. For a helpful survey see Fletcher, George P., Rethinking Criminal Law (Boston: Little, Brown, 1978), 855–875.Google Scholar

12 In “The Paradox of Punishment”, Philosophy and Public Affairs 9 (1979–1980), 42–58Google Scholar, Alan H. Goldman worries that something like the upper limit qualification forbids us to imprison persons for crimes against property, even though imprisoning these persons is necessary if our laws protecting property are to be effective. This worry only arises because Goldman's version of the upper limit qualification is incorrect. “If we ask which rights are forfeited in violating the rights of others”, he says, “it is plausible to answer just those rights that one violates (or an equivalent set)”, where equivalence “is to be measured in terms of some average or normal preference scale, much like the one used by the utilitarian when comparing and equating utilities and disutilities” (45). Since most people would prefer losing several thousand dollars to spending five years in prison, Goldman concludes that the upper limit qualification forbids us to give five-year prison terms for thefts of several thousand dollars. Goldman's problem here is similar to Nozick's: he is try ing to generate a ranking of rights using concepts which belong properly in a utilitarian rather than a rights-based theory. If he used genuine rights concepts he would find that property rights and rights not to be imprisoned are not ranked with respect to each other, and that the upper limit qualification permits imprisonment for crimes against property.

13 Nozick does not, curiously enough, appeal to his version of the minimum necessary qualification when he discusses punishment in Anarchy, State, and Utopia, 59–63Google Scholar, but instead presents a version of the “retributive calculus”. Not only does this mark a (to my mind unattractive) departure from Locke, it also weakens the parallel which Nozick draws later on in his discussion between issues about punishment and issues about self-defence.

14 As is often pointed out, the studies have not produced evidence that capital punishment is not a deterrent to murder either. But the onus of proof in this question is surely on the defenders of capital punishment to show that it is.