Does Kant Have a Theory of Punishment?

  • Jeffrie G. Murphy
Part of the Philosophical Studies Series book series (PSSP, volume 54)


Several years ago I thought that I understood Kant’s theory of criminal punishment—an understanding aided by a Rawlsian reconstruction of certain aspects of that theory.1 I thought that Kant’s theory was profound, inspiring, and—although subject to certain problems—the only morally acceptable theory of punishment.2 I was confident that all philosophical work on problems of crime and punishment could be done within a generally Kantian perspective (which I took to be an unambiguously retributive perspective), and some of my own work exhibited that confidence to an uncritical degree.3


Supra Note State Coercion Criminal Punishment Moral Justification Moral Evil 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    See J. Murphy, Kant: The Philosophy of Right, 109–49 (1970) [hereinafter Murphy, Kant];Google Scholar
  2. 1.
    J. Murphy, Kant’s Theory of Criminal Punishment, in Retribution, Justice, and Therapy: Essays in the Philosophy of Law 82–92 (1979) [hereinafter Retribution, Justice, and Therapy] .Google Scholar
  3. 2.
    See Murphy, Marxism andRetribution, 2 J. Phil. & Pub. Aff. 217–43 (1973), reprinted in Retribution, Justice, and Therapy, supra note 1, at 93–115.Google Scholar
  4. 3.
    For a more expanded treatment of the theory of crime and punishment, see J. Murphy & J. Coleman. The Philosophy of Law: An Introduction to Jurisprudence 69–112 (1984).Google Scholar
  5. 4.
    SeeMurphy, Retributivism, Moral Education, and the Liberal State, 4 Crim. Just. Ethics 3–11 (1985).CrossRefGoogle Scholar
  6. 5.
    I. Kant, Universal Practical Philosophy, in Lectures on Ethics 55 (L. Infield trans. 1978) [hereinafter Universal Practical Philosophy] .Google Scholar
  7. 6.
    Id . at 62–63.Google Scholar
  8. 7.
    Id. at 67–68.Google Scholar
  9. 8.
    I. Kant, Critique of Pure Reason *A317/B373 (N. Smith trans. 1965).Google Scholar
  10. 9.
    Id. at *A551/B579.Google Scholar
  11. 10.
    I Kant, Foundations of the Metaphysics of Morals *406–07 (L. Beck trans. 1969)[hereinafter Foundations of the Metaphysics of Morals]. All references to Kant’s Groundwork are cited to Beck’s translation entitled Foundations. Google Scholar
  12. 11.
    Reprinted in Philosophical Correspondence, 1759–99, at 199 (A. Zweig trans. 1967) translating 11 I. Kant, Gesammelte Schriften *398–99 (Prussian Academy ed. 1900).Google Scholar
  13. 12.
    I. Kant, Religion Within the Limits of Reason Alone 16 (T. Greene & H. Hudson trans. 1960).Google Scholar
  14. 13.
    Id. at 33–34.Google Scholar
  15. 14.
    Id . at 60–61.Google Scholar
  16. 15.
    Id. at 87–90.Google Scholar
  17. 16.
    Id. at 158 n.*Google Scholar
  18. 17.
    I. Kant, To Perpetual Peace: A Philosophical Sketch, in Perpetual Peace and Other Essays *366 (T. Humphrey trans. 1983).Google Scholar
  19. 18.
    I. Kant, The Doctrine of Virtue *459–60 (M. Gregor trans. 1964) [hereinafter Doctrine of Virtue] .Google Scholar
  20. 19.
    To facilitate reference to the passages from Kant quoted supra text accompanying notes 5–18, the passages have been assigned numbers and are referred to in the text by those numbers.Google Scholar
  21. 20.
    Note that this justification, though teleological, is not utilitarian. The goal aimed at is a world in which rights are protected and respected, not where happiness or preference satisfaction is maximized.Google Scholar
  22. 21.
    A hypothetical imperative is a command of reason that binds us or represents the rational thing to do contingent upon our having a certain desire. A categorical imperative binds us or represents the rational thing to do independently of any desire we might have. See Foundations of the Metaphysics of Morals, supra note 10, at *414.Google Scholar
  23. 22.
    I. Kant, The Metaphysical Elements ofJustice *331(J. Ladd trans. 1965) (Rechtslehre) [hereinafter Metaphysical Elements of Justice] .Google Scholar
  24. 23.
    Id. at *333.Google Scholar
  25. 24.
  26. 25.
    Id. at *363.Google Scholar
  27. 26.
    The apportioning of happiness to virtue is, for Kant, the highest good (summum bonum). Kant’s belief that this is unattainable on earth is one of the reasons that Kant postulates the existence of God, and it is part of his “moral proof” for the existence of God. See J. Murphy, The Highest Good as Content for Kant’s Ethical Formalism, Kant-Studien, Band 56, Heft 1, 102–10 (1965).Google Scholar
  28. 27.
    Kant puts it this way: “the universal law of justice is: act externally in such a way that the free use of your will is compatible with the freedom of everyone according to a universal law.”Metaphysical Elements of Justice, supra note 22, at *231.Google Scholar
  29. 28.
  30. 29.
    For an argument that this defense of the legitimacy of coercion does not work, see R. Nozick, Anarchy, State, and Utopia 88–119 (1974).Google Scholar
  31. 30.
    Metaphysical Elements of Justice, supra note 22, at *331 (brackets in translation).Google Scholar
  32. 31.
    Kant is typically regarded as a paradigm retributivist, but one point of this paper is to bring into question the extent to which he is actually a retributivist in his theory of punishment.Google Scholar
  33. 32.
    For an attempt to develop a theory of criminality in these terms, see Why Have the Criminal Law at All?, in J. Murphy & J. Coleman, supra note 3, at 114–23.Google Scholar
  34. 33.
    For an expansion of this Rawlsian interpretation of Kant’s political philosophy, see Murphy, Kant, supra note 1.Google Scholar
  35. 34.
    See Metaphysical Elements of Justice, supra note 22, at *230–33.Google Scholar
  36. 35.
    Id. at *331; see supra text accompanying note 30.Google Scholar
  37. 36.
    See id. at *331–37.Google Scholar
  38. 37.
    Rawls, Legal Obligation and the Duty of Fair Play, in Law and Philosophy 3–18 (S. Hook ed. 1964);Google Scholar
  39. 37.
    Morris, Persons and Punishment, 52 Monist 475–501 (1968).CrossRefGoogle Scholar
  40. 38.
    Murphy, Kant, supra note 1, at 142–43.Google Scholar
  41. 39.
    See supra note 4.Google Scholar
  42. 40.
    See especially A. R. Duff, Trials and Punishments 205–32 (1986).Google Scholar
  43. 41.
    See id. at 211–12.Google Scholar
  44. 42.
    It would still be an important concern, of course, for it attempts to deal with the problem of free riding. But it must be a secondary concern. One will seek to prevent free riding in a practice that provides for a common benefit, but the common benefit (which gives the practice its primary point or justification) cannot itself be the prevention of free riding.Google Scholar
  45. 43.
    B. Aune, Kant’s Theory of Morals 167–68 (1979).Google Scholar
  46. 44.
    It may, for example, be possible to give good rule-utilitarian arguments for limiting punishment to those who deserve it. See J. Murphy & J. Coleman, supra note 3, at 131–36.Google Scholar
  47. 45.
    Metaphysical Elements of Justice, supra note 22, at *331 (emphasis added).Google Scholar
  48. 46.
    P. Riley, Kant’s Political Philosophy 109 (1983).Google Scholar
  49. 47.
    SeeMurphy, Kant’s Concept of a Right Action, 51 Monist 574–98 (1967).CrossRefGoogle Scholar
  50. 48.
    R. Wolff, The Autonomy of Reason 15 (1973) (emphasis added).Google Scholar
  51. 49.
    See Universal Practical Philosophy, supra note 5, at 57–70. Brief remarks on responsibility may be found in the Introduction to the Metaphysics of Morals. See Metaphysical elements of Justice, supra note 22, at *224, *228.Google Scholar
  52. 50.
    See, e.g., I. Kant, On the Proverb: That may be True in Theory but is of no Practical Use, in Perpetual Peace & Other Essays *299–300(T. Humphrey trans. 1983) [hereinafter Theory & Practice] . Kant’s example (from Cicero) of the two men at sea fighting over a plank is perhaps not—because of the competitive element involved—the best case in terms of which to discuss necessity. For a splendid discussion of necessity cases (mainly of murder and cannibalism to avoid starvation after shipwreck),Google Scholar
  53. 50a.
    see A. Simpson, Cannibalism and the Common Law (1984). The main focus of Simpson’s book is the famous 1884 English case, familiar to all law students, Regina v. Dudley & Stephens, 14 Q.B. 273 (1884). With respect to this case, Simpson claims that various parties involved for the government felt a conflict between two values: (i) that the law should assert an absolute prohibition against the killing of the innocent and (ii) that the defendants in this particular case, because of the terrible pressure they had been under, should not be made to suffer very much in the way of punishment. A way to satisfy both demands, proposed by various government parties prior to the trial, was this: convict Dudley and Stephens of murder, sentence them to death, and then have the Crown pardon them or commute their sentences. A. Simpson, supra, at 198–255. The latter was in fact the outcome. Id. at 247–48. The scenario was to be an application of strict legal justice conjoined with sovereign mercy and compassion. Given that Kant sometimes seems to care about issues of excuse and is sometimes sensitive to human frailty and vulnerability to powerful temptation, but given also that his theory of the strictly external nature of legal justice seems to preclude his working a strong theory of excuse into the law itself, we might have expected him to follow a similar course—namely,to suggest that when the law as interpreted and administered by the court fails (perhaps necessarily) to take due account of such “inner” factors, then perhaps it is proper for the sovereign to use the power of pardon to show mercy or compassion (or at least administer highly individuated justice) in these cases. Kant does not, however, take this line. Indeed, he explicitly argues that the sovereign should not exercise the power of pardon in such cases. His reason is not one we might expect (e.g.,that the sovereign is no better at reading the inner heart than anyone else) but is rather the following: The right to pardon a criminal (jus aggratiandi), either by mitigating or by entirely remitting the punishment, is certainly the most slippery of all the rights of the sovereign. By exercising it he can demonstrate the splendor of his majesty and yet thereby wreak injustice [unrecht] to a high degree. With respect to a crime of one subject against another, he absolutely cannot exercise this right, for in such cases exemption from punishment (impunitas criminis) constitutes the greatest injustice toward his subjects. Consequently, he can make use of this right of pardon only in connection with an injury committed against himself (crimen laesae majestatis). But, even in these cases, he cannot allow a crime to go unpunished if the safety of the people might be endangered thereby. Metaphysical Elements of Justice, supra note 22, at *337.Kant does suggest, however, that the sovereign should exercise his power of pardon when social utility demands it: Anyone who is a murderer—that is, has committed a murder, commanded one, or taken part in one—must suffer death. This is what [legal] justice as the Idea of the judicial authority wills in accordance with universal laws that are grounded a priori. The number of such a deed might, however, be so large that the state would soon approach the condition of having no more subjects if it were to rid itself of these criminals, and this would lead to its dissolution and a return to the state of nature, which is much worse, because it would be a state of affairs without any external legal justice whatsoever. Since a sovereign will want to avoid such consequences and, above all, will want to avoid adversely affecting the feelings of the people by the spectacle of such butchery, he must have it within his power in case of necessity (casus necesitatis) to assume the role of judge and to pronounce a judgment that, instead of imposing the death penalty on the criminals, assigns some other punishment that will make the preservation of the mass of the people possible, such as, for example, deportation. Id. at *334. This is puzzling. According to Kant’s theory of public crimes (crimes for which punishments are appropriate), all crimes are in some sense crimes against sovereignty. And thus all crimes, not simply ones directed against the person who happens to hold sovereign office, should be equally eligible for pardon. And if social utility (of all things!—given Kant’s frequent sermons on the priority of justice over utility) serves to justify pardon, then it is unclear why mercy, compassion, and a desire to attain greater precision and individuation with respect to justice (than is possible for the courts) might not also serve as justification. The issues of forgiveness, mercy, and individuated justice (and the relation of these virtues to the power of pardon) are simply too rich for the superficial treatment Kant here accords them. For further discussion of these matters, seeGoogle Scholar
  54. 50b.
    J. Murphy, Forgiveness and Resentment, in Midwest Studies in Philosophy VII: Social and Political Philosophy 503–16(P. French, T. Uehling, Jr. & H. Wettstein eds. 1982);Google Scholar
  55. 50c.
    J. Murphy, Mercy and Legal Justice, Soc. Phil. & Pol’y, Autumn 1986, at 1–14.Google Scholar
  56. 50d.
    See also J. Murphy and J. Hampton, Forgiveness and Mercy (1988).CrossRefGoogle Scholar
  57. 51.
    Metaphysical Elements of Justice, supra note 22, at *235–36.Google Scholar
  58. 52.
    There are also, of course, utilitarian grounds that might prompt one to refuse to recognize a right of necessity—e.g.the belief that such a right might be abused, encouraging people to see necessity before it is actually present.Google Scholar
  59. 53.
    Metaphysical Elements of Justice, supra note 22, *230–31.Google Scholar
  60. 54.
    Theory & Practice, supra note 50, at *297.Google Scholar
  61. 55.
    What would it mean to call the law coercive if it did not at least thwart wants and desires?Google Scholar
  62. 56.
    J. Rawls, A Theory of Justice 118–92 (1971).Google Scholar
  63. 57.
    For an argument that hypothetical consent theories are of little value in moral, political, and legal theory, see R. Dworkin, Taking Rights Seriously 150–83 (1977).Google Scholar
  64. 58.
    In Regina v. Dudley & Stephens, 14 Q.B. 273 (1884), Lord Coleridge, C.J., expressed such a concern: “Who is to be the judge of this sort of necessity?... It is plain that the principle leaves to him who is to profit by it to determine the necessity which will justify him in deliberately taking another’s life to save his own.” Id. at 287. Of course one must not be too hasty in applying such “who is to judge?” arguments, since the question can be addressed to any principle at all (e.g., self-defense, provocation, etc.). One answer to the question “who is to judge?” is simply this: the jury after hearing all the evidence.Google Scholar
  65. 59.
    According to Simpson this was perhaps an accepted practice among sailors in the 19th century and earlier. See A. Simpson, supra note 50, at 122–32.Google Scholar
  66. 60.
    Let me again remind the reader that part of my argument in this essay is that we must not simply assume the correctness of the common interpretation of Kant as an unambiguous retributivist.Google Scholar
  67. 61.
    For example: Kant claims that children born out of wedlock—and thus in the state of nature rather than in civil society—may be thought to have less value and fewer rights than legitimate children. The child has, after all, “crept surreptitiously into the commonwealth ( prohibited wares), so that its existence as well as its destruction can be ignored.” Metaphysical Elements of Justice, supra note 22, at *336.Google Scholar
  68. 62.
    Id. at *336–37.Google Scholar
  69. 63.
    See J. Murphy, Cruel and Unusual Punishments, in Retribution, Justice, and Therapy, supra note 1, at 223–49.Google Scholar
  70. 64.
    Metaphysical Elements of Justice, supra note 22, at *332.Google Scholar
  71. 65.
    Id. at *363.Google Scholar
  72. 66.
  73. 67.
    For a sensitive discussion by Kant of the importance of inner sensibility, see id. at *332 (discussion of the possible relevance of class distinctions to criminal punishment); id. at *333 (discussions of the different motives that might lie behind treason or rebellion). There are times when Kant clearly wants to consider such matters. But most of the time he does not—perhaps because at some level he realizes that such concerns do not sit well with his insistence that the law must deal only with external actions.Google Scholar
  74. 68.
    Id. at *363.Google Scholar
  75. 69.
    There is another problem here for the “fair play” interpretation of Kantian retributivism. If the wrongness of criminal conduct simply consists in the unfairness of failing to exercise the self-restraint of obeying the law in cases where one expects others to do this, it is hard to see how—on this theory of wrongness—criminaloffenses could be graded on a scale of seriousness so that any form of a proportionality principle could be applied. Murder seems worse than robbery, but it is not in any obvious way more unfair in the relevant sense.Google Scholar
  76. 70.
    If there is an argument, it is this: free riding is an obvious example of injustice. Taking from an individual something equivalent to what his free riding has gained him is thus a clear example of doing justice.Google Scholar

Copyright information

© Springer Science+Business Media Dordrecht 1992

Authors and Affiliations

  • Jeffrie G. Murphy
    • 1
  1. 1.College of LawArizona State UniversityTempeUSA

Personalised recommendations