1. J.L. Austin: “A Plea for Excuses,” in LVH Proceedings of the Arisi. Soc ‘y (1956–1957) at 1.

2. Many commentators have regarded the distinction between justification and excuse to be crucial to an understanding of criminal law. See Michael Corrado, ed., Justification and Excuse in the Criminal Law (New York: Garland Publishing, 1994).

3. Austin was not concerned to develop a comprehensive taxonomy of defense categories. At a minimum, however, at least one other kind of desert-based defense must be countenanced. See Douglas Husak, Philosophy of Criminal Law (Totowa, NJ: Rowman & Littlefield, 1987) at Ch. 7.

4. Supra note 1 at 3 (emphasis in original).

5. Austin refers to “partial justification” as well as to “partial excuse.” Ibid, at 3. In explaining his reasons for focusing on excuses, Austin writes that “I shall not be so anxious to talk about [justifications] because they have enjoyed more than their fair share of philosophical attention.” Ibid. at 2. However true this observation may have been in 1957, moral and legal theorists have done far more work on excuses than on justifications in the interim.

6. Many commentators define “defense” so that the concept of a complete defense is redundant and the concept of a partial defense is incoherent. According to one commentator, “a true defense is one that, if proved, results in the acquittal of a defendant, although the prosecutor has proved beyond a reasonable doubt every element in the definition of the crime.” Joshua Dressier, Understanding Criminal Law, 2d ed. (New York: Matthew Bender, 1995) at 182. The point of Dressler’s definition is to disqualify what I call denials (in supra note 3) as “true defenses.” An additional implication of his definition, however, is that a partial defense is not a “true defense.”

7. Several factors may help to explain the relative lack of interest among commentators in partial defenses. One such factor is that partial defenses are clearly less important than complete defenses. A circumstance that precludes liability altogether is more valuable to a defendant than one that merely reduces the severity of his punishment, and thus can be expected to have received more attention from theorists.

8. Black’s Law Dictionary (St. Paul, MN: West Publishing Company, 1990).

9. U.S. Sentencing Commission, Sentencing Guidelines and Policy Statements (1987) Sec. 5K1.1.

10. Ibid., Sec. 5Kl.1.

11. Presumably, the rationale for recognizing such circumstances as mitigating is that they “sav[e] the time of prosecutors and courts and thereby contribut[e] to the smooth running of the criminal justice system.” See Andrew Ashworth, Sentencing & Criminal Justice, 2d ed. (London: Butterworths, 1995) at 137. Even alternative rationales—for example, that “a guilty plea is evidence of remorse” (Ibid, at 137)—does not provide a desert base for mitigation.

12. See text, infra Section III.

13. See D.A. Thomas, Principles of Sentencing, 2d ed. (London: Heinemann Library of Criminology and Penal Reform, 1979) at 47.

14. Some commentators who endorse the claim that defendants are not entitled to a consideration of their mitigating circumstances purport to compromise rather than to reject the principle of proportionality. Immediately after alleging that “credit for mitigating factors is not an entitlement of the offender,” Thomas proceeds to argue that “the sentencer is permitted to refrain from making an allowance for mitigating factors in order to achieve a recognized penal objective, such as general deterrence [or] the prevention of further offences … [TJhese considerations are not normally held to justify a sentence which is disproportionate to the immediate offence, but they may justify the sentencer in ignoring mitigating factors and giving no credit to them.” Ibid, at 47.

Sometimes it is difficult to decide whether these claims are designed merely to describe positive law or to provide a normative defense of it. If Thomas is offering a descriptive generalization about the practice of sentencing authorities, I have little doubt that his claim is true. But it is hard to see how he could defend the position that sentencing authorities may disregard a partial defense without also violating the principle of proportionality. I see no principled reason why the discretionary status of partial defenses should be any different from that of complete justifications and excuses.

15. For further clarification of the relation of justifications to excuses, see Douglas Husak, “The Serial View of Criminal Law Defenses” (1992) 3 Crim. L. Forum 369.

16. See, for example, Kent Greenawalt, “The Perplexing Borders of Justification and Excuse” (1984) 84 Colum. L. Rev. 1897.

17. See text, infra Section V.

18. For a development of the contrast between concepts and conceptions, see Ronald Dworkin, Law’s Empire (Cambridge, MA: Harvard University Press, 1986) at 70–72.

19. See text, infra Section IV.

20. This baseline is described as a “‘heartland,’ a set of typical cases” in the U.S. Sentencing Guidelines. Supra note 9, Sec. 1.6.

21. For a discussion of such issues, see Andrew von Hirsch & Nils Jareborg, “Gauging Criminal Harm: A Living-Standard Analysis” (1991) 11 Oxford J. of Legal Stud. 1 at 6.

22. I am hesitant to say that these paradigm or standard cases should be constructed solely by statistical data about the typical offender and offense. Such a proposal would be especially problematic in formulating a paradigm or standard case of an inchoate offense. The statistically average drunk driver, for example, is not involved in a crash. Only the statistically atypical case provides any insight into whether and to what extent such conduct should be punished.

23. Some defenses, to be sure, are all-or-nothing. Consider a kind of defense I call denials, which differ conceptually from both justifications and excuses. Denials have the form “No, I did not [commit the crime of which I am charged]” rather than “Perhaps I did, but…” This kind of defense cannot admit of degrees or be partial. For a further discussion of kinds of defenses, see supra note 3 at ch. 7.

24. Some commentators deny the existence of partial justifications or excuses. I do not survey the grounds on which such claims have been made. For an argument in favor of the view that “a particular act or offence cannot be partially justified,” see Suzanne Uniacke, Permissible Killing (Cambridge: Cambridge University Press, 1994) at 13. On some accounts, excuses cannot admit of degrees. If excuses are assimilated to invalidating or nullifying conditions, as in the law of contracts, it is difficult to understand how they could be partial. For a defense of this conception of excuses, see H.L.A. Hart, Punishment and Responsibility (New York: Oxford University Press, 1968) at 29.

25. For an interesting empirical study alleging that “the trouble with blame is [its] all or nothingness,” see Sharon Lamb, The Trouble With Blame (Cambridge, MA: Harvard University Press, 1996) at 12.

26. In questioning whether complete justifications, for example, preclude wrongfulness altogether, I am not simply making the familiar observation that the wrongfulness of justified conduct is only outweighed rather than eliminated. This feature of justifications is an important ground on which some theorists have argued that victims of justified acts may be entitled to compensation. See Judith Thomson, The Realm of Rights (Cambridge, MA: Harvard University Press, 1990) at 93–96.

27. I do not deny that some complete justifications and excuses may preclude wrongfulness or blame altogether—but only that not all complete defenses do so.

28. I suspect that the view that a completely justified act must involve no wrongdoing helps to account for the reluctance of many theorists to believe that killings by battered women can be justified. Yet many of these same theorists sympathize with battered women, and do not approve of their conviction and punishment. Thus they are led to conceptualize these defendants as excused. See, for example, Mira Mihajlovich, “Does Plight Make Right: The Battered Woman Syndrome, Expert Testimony and the Law of Self-Defense” (1987) 62 Indiana L. J. 1253.

29. The supposition that only a tiny quantum of wrongfulness is required to justify the imposition of criminal liability is incompatible with the existence of a de minimis defense in most jurisdictions. This defense is available when the conduct of the defendant is “too trivial to warrant the condemnation of conviction.” Model Penal Code, 2. 12(2).

30. Consider a second example. Suppose that a jurisdiction does not require defendants to retreat before they are permitted to kill in self-defense. One can hold that the act of retreat is preferable to the act of killing in self-defense without undermining the conclusion that killings by defendants who fail to retreat are completely justified.

31. Sometimes the location of this line seems nearly arbitrary. According to one commentator, “the basis [for assuming a clear distinction between excusing conditions and mitigating excuses (i.e., those excuses taken into account by way of sentencing discretion)] is rarely articulated, and when it is, it seems unconvincing.” Martin Wasik, “Partial Excuses in the Criminal Law” (1982) 45 Modern L. Rev. 516 at 516.

32. By a theory of mitigation, I mean nothing more ambitious than one or more principles to govern the issue of whether mitigation should be extended or withheld.

33. An exception must be made for the context of capital punishment. Commentators have paid much greater attention to the circumstances that mitigate what might otherwise be capital murder. See, for example, James Acker & Charles Lanier, “In Fairness and Mercy: Statutory Mitigating Factors in Capital Punishment Cases” (1994) 30 Crim. L. Bulletin 299.

34. Andrew Ashworth and D.A. Thomas are the two most notable exceptions to this generalization. See supra note 13; supra note 11; text, infra note 61.

35. I suspect that the absence of a theory of mitigation helps to explain why many commentators believe that sentencing authorities have die discretion to disregard a partial defense. If a defendant is entitled to have his mitigating circumstances taken into consideration, but there is uncertainty about whether a given circumstance mitigates, then the justifiability of all sentences is called into doubt. Theorists endeavor to block this skeptical conclusion by contending that sentencing authorities need not take partial defenses into account.

36. “A sentencing system tailored to fit every conceivable wrinkle of each case can become unworkable.” Supra note 9, Sec. 1.2.

37. Ibid., Intro, to ch. 3.

38. Ibid., Sec. 3B1.2 authorizes a downward departure if a defendant was a “minimal” or a “minor participant in any criminal activity.”

39. Ibid., Sec. 3E1.1 audiorizes a downward departure if a defendant “clearly demonstrates acceptance of responsibility for his offense.”

40. See supra note 9.

41. 18 U.S.C. Sec. 3553(b). Such a catch-all would be inconceivable (as a violation of the principle of legality) if it pertained to complete justifications.

42. Supra note 9, Sec. 5H1.10. In addition, the Guidelines Manual lists a number of circumstances that “are not ordinarily relevant” in departing from the Guidelines. These include education and vocational skills, mental and emotional conditions, physical condition, employment record, and family ties and responsibilities.

43. Annual Reports of the U.S. Sentencing Commission attempt to provide guidance to sentencing authorities by listing the circumstances that have been affirmed on appeal to allow (or not to allow) a departure from a guideline sentence.

44. For a discussion of some of the considerations that should lead authors of sentencing guidelines to provide a general statement of the principles that can justify a departure from a presumptive sentence, see Andrew von Hirsch, “Structure and Rationale: Minnesota’s Critical Choices” in Andrew von Hirsch, Kay Knapp & Michael Tonry, eds., The Sentencing Commission and its Guidelines (Boston: Northeastern University Press, 1987) 84 at 103–04.

45. I will continue to refer to “code-relativity,” even though the same point applies to common law jurisdictions in which criminal offenses are created by the judiciary rather than by a legislature.

46. This new defense is neither a justification nor an excuse, but an instance of what I call a denial in supra note 3, ch. 7.

47. I say “in principle” because, as a practical matter, there is surely some upper limit to the complexity that should be allowed in a criminal code. No legislator should strive to create thousands of offenses to correspond to the many varieties of mitigating circumstances that should affect sentencing.

48. For an account of this distinction, see H.L.A. Hart, supra note 24 at 15.

49. Much has been written about formal mitigation in general and about provocation in particular. The best book-length treatment is Jeremy Horder, Provocation and Responsibility (Oxford: Oxford University Press, 1992).

50. Many theorists believe that the concept of attempted manslaughter is incoherent. Qua attempt, this offense requires an intention with respect to its result; qua manslaughter, however, this offense lacks an intention to kill. According to one commentator, attempted voluntary manslaughter, but not attempted involuntary manslaughter, is a possible offense. See R.A. Duff, Criminal Attempts (Oxford: Clarendon Press, 1996) at 23.

51. See text, infra Section V.

52. See ibid., Section II.

53. Model Penal Code, Sec. 2.13.

54. See supra note 11 at 134–35.

55. I do not discuss whether it is possible to include aggravating circumstances along with mitigating circumstances in a proposed unification of partial and complete defenses.

56. At least, this generalization applies to “true defenses.” See supra note 6. It does not apply to what I call denials. See supra note 3. Some commentators purport to identify defense-types other than justifications and excuses. According to Paul Robinson,_/ive distinct types of defenses exist. In addition to justifications and excuses, Robinson identifies “failure of proof defenses, “offense modifications,” and “public policy” defenses. The first of these correspond to what I call denials, which do not admit of degrees. The last of these lacks a basis in the offender’s desert. Robinson describes the remaining category—offense modifications—as follows: “While the actor has apparently satisfied all the elements of the offense charged, he has not in fact caused the harm or evil sought to be prevented by the statute defining the offense.” Paul Robinson, Criminal Law Defenses, vol. 1, (St. Paul, MN: West Publishing Company, 1984) at 77.

57. But see the examples of “quasi-desert mitigation” provided in Andrew von Hirsch, “Sentencing Guidelines and Penal Aims in Minnesota” (1994) 13 Crim. Justice Ethics 39.

58. See text, infra Section V.

59. See text and citation, supra at note 104.

60. I assume that nothing that happens to the defendant after the commission of his offense can affect his desert for committing the offense. Thus such factors as the defendant’s extraordinary degree of post-offense restitution, given mitigating significance in U.S. v. Lieberman, 971 F.2d 989 (3d Cir. 1992) can have no basis in his desert. For further discussion, see Douglas Husak, “Already Punished Enough” (1990) 18 Philosophical Topics 79.

61. Some of these issues are discussed in Andrew Ashworth, “Justifying the Grounds of Mitigation” (1994) 13 Crim. Justice Ethics 5. Curiously, Ashworth argues in favor of (or against) recognizing a partial defense without invoking his earlier view that mitigating circumstances are “almost defenses.” See Andrew Ashworth, Principles of Criminal Law, 2d ed. (Oxford: Clarendon Press, 1995) at 18–19 [hereinafter ‘Principles’] and supra note 11 at 134.

62. Neither complete nor partial defenses need be thought to have any sort of logical priority to one another; both are derived from the same substantive theories.

63. Principles, supra note 61 at 18–19. Elsewhere, Ashworth writes that many defenses “are restricted tightly in the expectation that courts will award substantial mitigation of sentence where the circumstances fall just outside the legal requirements for a defence.” Supra note 11 at 134.

64. This train of thought might be called the “consolation theory” of partial defenses. An argument that comes close but fails to win a complete acquittal establishes a claim in mitigation as a kind of a consolation prize.

65. In Section V, I will argue that analogues of partial defenses in complete defenses cannot always be understood in these terms.

66. See text, infra Section I.

67. In other words, theory and example must be brought into “reflective equilibrium.” For an account of this methodology, see John Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1972) at 48–51.

68. See the discussion of “victim derogation” in C.R. Snyder, Raymond Higgins & Rita Stucky, Excuses (New York: John Wiley & Sons, 1983) at 85–86.

69. Of course, convicted criminals may be treated in ways that otherwise would be offenses. Whether this fact presents complications for the unifying hypothesis depends on how one understands analogues of partial defenses in complete defenses.

70. In making this claim about principles and doctrines, I do not mean to suggest that nothing about the criminal justice system favors wealthy defendants. See, for example, Jeffrey Reiman, And the Poor Get Prison (Boston: Allyn and Bacon, 1996).

71. See supra note 42.

72. The socio-economic status of the defendant, however, is somewhat more controversial. See text, infra Section V.

73. Principles, supra note 61.

74. Supra note 13 at 195.

75. I make no effort to canvass each of the possible theories of excuse. See supra note 6 at 189–93.

76. See Michael Moore, “Choice, Character, and Excuse,” in Ellen Paul, Fred Miller & Jeffrey Paul, eds., Crime, Culpability, and Remedy (Oxford: Basil Blackwell, 1990) at 29.

77. See the description of a paradigm of a “perfectly voluntary choice” in Joel Feinberg, Harm to Self (New York: Oxford University Press, 1986) at 113–17.

78. For a defense of this theory, see Michael Bayles, “Character, Purpose, and Criminal Responsibility” (1982) 1 L. and Phil. 5; and Peter Arenella, “Character, Choice and Moral Agency: The Relevance of Character to Our Moral Culpability Judgments” in Paul, Miller & Paul, eds., supra note 76, 59.

79. See supra note 6 at 273. The Model Penal Code, Sec.2.09, incorporates this requirement by allowing the defense only when “a person of reasonable firmness” would have been unable to resist the threat.

80. The U.S. Sentencing Guidelines, supra note 9, Sec. 5K2.12 provide for a reduction in a guideline sentence “if the defendant committed the offense because of serious coercion, blackmail or duress not amounting to a complete defense.”

81. See Model Penal Code, Sec. 3.04(2)(b).

82. As with excuses, I will make no attempt to canvass each of the possible theories of justification. See supra note 6 at 186–89.

83. Paul Robinson, Criminal Law (New York: Aspen Publishers, 1997) at 402.

84. Ibid, at 401.

85. The U.S. Sentencing Guidelines, supra note 9, Sec. 5K2.11, call for reduced punishment for “lesser harms.” The example provided is a mercy-killing.

86. See Doughty (1986), 83 Crim.App. 319.

87. See supra note 20.

88. Perhaps this same response can be made to long-standing uncertainty about whether intoxication should function as a mitigating circumstance. The paradigm offender of many crimes may well be somewhat intoxicated, so the intoxication of a particular defendant cannot serve to differentiate him from this paradigm. For a discussion of the association between crime and intoxication, see Mitchell Keiter, “Just Say No: The Rise and Fall of the Intoxication Defense” (1997) 87 J. of Criminal Law & Criminology 482.

89. Perhaps the amount of mitigation to which this offender is entitled is so tiny that policy reasons militate against including it at all. I am unclear to what extent the disregard of considerations with a desert base involves a rejection of the principle of proportionality. But desert need not be the only factor to play a role in what qualifies as a desert-based theory of punishment and sentencing.

90. See supra note 36.

91. Ultimately, the question whether circumstances differ at all—in degree or in kind—cannot be resolved without criteria to decide how properties (or circumstances) are individuated—a difficult topic I do not address here.

92. Jurisdictions differ about the circumstances under which ignorance of law provides a complete defense. See Model Penal Code, Sec. 2.04(3).

93. After maintaining that ignorance of law does not ordinarily affect the liability of the defendant, commentators frequently add that “it may affect his punishment.” See, for example, J.C. Smith & Brian Hogan, Criminal Law, 5th ed. (London: Butterworths, 1983) at 68.

94. See Michael Travers, “Mistake of Law in Mala Prohibita Crimes” (1995) 62 U. Chi. L. Rev. 1301.

95. See Douglas Husak & Andrew von Hirsch, “Culpability and Mistake of Law” in Stephen Shute, John Gardner & Jeremy Horder, eds., Action and Value in Criminal Law (Oxford: Clarendon Press, 1993) 157 at 157.

96. Model Penal Code, Sec. 2.09.

97. In this respect, the debate resembles what might be said about entrapment in jurisdictions that do not allow a complete defense. See text, infra Section III.

98. See the dissenting opinion of Lord Simon in D. P.P. for Northern Ireland v. Lynch (1975), A.C. 653 at 686.

99. Commentators who are skeptical that the substantive criminal law can be placed on a rational foundation are likely to respond that the question of whether a particular circumstance has an analogue in a complete defense cannot be answered. For a defense of this kind of position, see Alan Norrie, Crime, Reason and History (London: Weidenfeld and Nicolson, 1993).

100. Model Penal Code, Sec. 2.09(1).

101. Ibid., Commentaries to 2.09, at 375.

102. See supra note 42.

103. Model Penal Code, Sec. 3.02(l)(a).

104. One commentator speculates that “the true basis…for the intuition that deprivation should excuse is sympathy for those who have suffered.” Stephen Morse, “Diminished Capacity” in Shute, Gardner & Horder, eds., supra note 95, 239 at 266.

105. See Douglas Husak, “Motive and Criminal Liability” (1989) 8 Crim. Justice Ethics 3.

106. I do not pretend that these brief remarks about the connection between poverty and necessity are definitive. If theorists steadfastly deny that good motive can ever be relevant to criminal liability—or that good motive can ever reduce the wrongfulness of a criminal act—then we will be hard-pressed to use the sufficiency hypothesis to defend the mitigating significance of good motive.

107. This boundary is equally elusive in the context of complete defenses. See supra note 16.

108. One commentator describes such circumstances as “quasi-desert mitigation”—as “circumstances that do not directly affect the harmfulness of the conduct or the actor’s personal culpability in committing it, but which nevertheless seem somewhat related to retributive concerns.” Supra note 57 at 46.

I would think that the desert base of this circumstance is relatively clear. Von Hirsch’s categorization of this circumstance as “quasi-desert” is best understood as reflecting his uncertainty that it amounts to either a partial justification or a partial excuse. Perhaps some desert-based mitigating circumstances are neither partial justifications nor excuses. If so, the unifying hypothesis would be undermined. I believe, however, that this circumstance does “affect the harmfulness of the conduct.” Uncertainty about whether this circumstance reduces harmfulness may derive from the supposition that the loss of property is the only harm caused by such conduct. This supposition, however, seems overly narrow and restrictive. In any event, a discussion of this example shows how sentencing theory can be used to better understand the nature of the harms sought to be prevented by the substantive criminal law.

109. Perhaps an analogue of this mitigating circumstances need not be sought in a wholly different offense. Arguably, robbery can be analyzed as a composite crime, involving elements of both assault and theft. The mitigating circumstance of the reassuring robber can be found in that component of robbery that involves assault.

110. See the discussion of abandonment in Duff, supra note 50 at 66–75.

111. Supra note 1.