An earlier version of this paper was presented at the Oxford Legal Philosophy Colloquium in March 2002. I am very grateful to Jeremy Horder for his response to my presentation, to John Gardner and other participants in the colloquium for their comments, and to Joseph Raz for an encouraging remark. I also thank Richard Bronaugh, Arthur Ripstein, Malcolm Thorburn, and Bruce Chapman for their comments on a draft. Erica Bussey provided excellent research assistance, and the Social Sciences and Humanities Research Council of Canada provided financial support.

1. Raz, Joseph, Practical Reasons and Norms, 2nd ed. (Oxford: Oxford University Press, 1999)CrossRefGoogle Scholar.

2. Ibid. at 15-16.

3. Ibid. at 17-19.

4. Ibid. at 17.

5. Ibid. at 18.

6. Gardner, John, “The Mysterious Case of the Reasonable Person” (2001) 51 U. Toronto L.J. 273 at 275–76.Google Scholar

7. Gardner, John, “Justifications and Reasons” in Simester, A.P. & Smith, A.T.H., eds., Harm and Culpability (Oxford: Clarendon Press, 1996) 103 at 105.Google Scholar

8. R. v. Dadson (1850), 4 Cox C.C. 358. Dadson, a police constable, was attempting to arrest a man stealing growing wood. The man fled, and Dadson shot and wounded him. Under the law of England at the time, the use of this degree of force was justified to effect an arrest of a “fleeing felon.” Theft of growing wood was not a felony unless the accused had been previously twice convicted of the same offence. In fact, the man had previously been twice convicted and so was a “fleeing felon.” But Dadson did not know of the man’s prior convictions and so could not have been firing at him for that reason.

9. Gardner, “Justifications,” supra note 7 at 106.

10. Gardner, John, “The Gist of Excuses” (1998) 1 Buff. Crim.L. Rev. 575 at 578CrossRefGoogle Scholar.

11. Ibid.

12. Fletcher reaches the same result for different reasons: see Fletcher, George, Rethinking Criminal Law (Boston: Little, Brown 1978) at 762–69 Google Scholar. Uniacke’s position is somewhat more complex, but she too ultimately concludes that the law should regard reasonably mistaken self-defence as an excuse rather than a justification: Uniacke, Suzanne, Permissible Killing (Cambridge: Cambridge University Press, 1994) at 39–41.CrossRefGoogle Scholar Robinson reaches the same result more directly. Since, in his view, justifications have to do with the objective qualities of the act and depend in no way on the actors” beliefs, mistaken use of defensive force could not be justified, but at best excused. See Robinson, Paul, “A Theory of Justification: Societal Harm as a Prerequisite for Criminal Liability” (1975) 23 UCLA L. Rev. 266 at 283–84 Google Scholar. I do not consider Robinson’s approach in this paper; I suspect it would not avoid all my criticisms of Gardner and Fletcher.

13. Gardner, “Justifications,” supra note 7 at 124.

14. This is no doubt the feature of Logan’s situation that Robinson would emphasize: see Robinson, Paul, “Competing Theories of Justification: Deeds v. Reasons” in Simester, A.P. & Smith, A.T.H., eds., Harm and Culpability (Oxford: Clarendon Press, 1996) 45.CrossRefGoogle Scholar

15. There are obvious analogies between this situation and Dadson (which Gardner and I both think was rightly decided) and the problem of the impossible attempt on which see Fletcher, Rethinking, supra note 12 at 146-84; Duff, R. Antony, Criminal Attempts (Oxford: Clarendon Press, 1996) at 76–115 and at 378-84Google Scholar; Robinson, “Competing Theories”, supra note 14 at 60-61; Stewart, Hamish, “The Centrality of the Act Requirement for Criminal Attempts” (2001) 51 U. Toronto L.J. 399.Google Scholar

16. If Logan’s situation doesn’t seem paradigmatic enough, imagine again a situation in which he was protecting his child or his spouse.

17. Some of this impatience is visible in Smith, J.C., Justification and Excuse in the Criminal Law (London: Stevens, 1989) at 12 and at 126 Google Scholar, and in Stuart, Don, Canadian Criminal Law: A Treatise, 3d ed. (Toronto: Carswell, 1995) at 421-23Google Scholar. Smith is more concerned with whether a particular type of situation should amount to a defence than with whether the defence in question is an excuse or a justification. I will nonetheless suggest below that some of Smith’s reasoning supports the idea that reasonably mistaken defensive force should be a justification.

18. Gardner, “Gist of Excuses”, supra note 10 at 590.

19. Gardner, “Reasonable Person”, supra note 6 at 289-90.

20. “One cannot act for a reason unless one believes in its validity.” Raz, supra note 1 at 48 Google Scholar. Just so; and a fortiori, one cannot believe in the validity of a reason, indeed one cannot have a view about its validity, if one does not know of its presence.

21. Compare Fletcher, George, “The Right and the Reasonable” (1985) 98 Harv. L. Rev. 949 at 978-79CrossRefGoogle Scholar; Dressler, Joshua, “New Thoughts about the Concept of Justification in the Criminal Law: a Critique of Fletcher’s Thinking and Rethinking” (1984) 32 UCLA L. Rev. 61 at 80–81Google Scholar and Gardner, “Justifications”, supra note 7 at 108.

22. Compare Greenawalt, Kent, “The Perplexing Borders of Justification and Excuse” (1984) 84 Colum. L. Rev.1897 at 1922-23Google Scholar.

23. This analysis raises the following question: what, precisely, is Logan’s guiding reason in Case 1? Is it “the suspect had a loaded gun” or “the suspect was about to kill or wound Briscoe or Logan”? It seems that, on Gardner’s and Raz’s analysis, it must be the latter. The fact that an event would have occurred had another event not intervened must count as a fact that can in turn be a reason for action; otherwise, one could never say, “Avoiding a is a good reason for x to φ.”

24. Horder, Jeremy, “Self-Defence, Necessity and Duress: Understanding the Relationship” (1998) 11 Can. J. L. & Juris. 143 at 147-48Google Scholar.

25. The Supreme Court of Canada has held that a correct but unreasonable hunch did not justify the exercise of an intrusive common law search and seizure power: R. v. Golden [2001] 3 S.C.R. 769 at paras. 109-112

26. For one of many contributions to the economic analysis of insurance, See Arrow, Kenneth J., “Optimal Insurance and Generalized Deductibles” in Individual Choice under Certainty and Uncertainty (Cambridge, MA: Harvard University Press, 1984) 212.Google Scholar

27. Gardner, Compare, “Reasonable Person”, supra note 6 at 279 Google Scholar. If I understand his discussion of probability and practical reason correctly, Gardner would draw this conclusion only for the risk-neutral individual. Risk-aversion by itself would be a perfectly proper reason to buy insurance, whether or not the risk materializes, because having insurance would be a way of counteracting the aversion to risk. “Insuring to avoid risk aversion” would be a different action from “insuring against a loss”. Compare ibid at 283. But if people are expected utility maximizers, as is usually assumed in economic analysis of insurance issues, risk-aversion arises indirectly if an individual has diminishing marginal utility of income or wealth, not directly from the individual’s feelings about risk. Consequently, the distinction Gardner requires seems tenuous at best; that is, for an expected-utility maximizer, avoiding a risk just is insuring against a loss.

28. Compare Perry, Stephen R., “Second-Order Reasons, Uncertainty, and Legal Theory” (1989) 62 S. Cal. L. Rev. 913 at 922–27.Google Scholar

29. Raz, , supra note 1 at 19.Google Scholar

30. Ibid. at 17 (beliefs), at 58-59 (norms), at 65-72 (decisions).

31. Horder, , supra note 24 at 148.Google Scholar

32. Rather than treating a probability as a reason, Raz treats it as a reason to believe in a fact which, if established, would itself be a reason; but he does not spell out what is meant by a “reason to believe.” Supra note 1 at 22 and at 198; see also Raz, Joseph, “Facing Up: A Reply” (1989) 62 S. Cal. L. Rev. 1153 at 1154 n. 2.Google Scholar Raz would, I expect, treat an apparent reason as a reason to believe in a fact which, if true, would itself be a reason.

33. See Savage, Leonard J., The Foundations of Statistics, 2nd ed. (New York: Dover, 1972) at 27–68 Google Scholar; Perry, Stephen R., “Risk, Harm, and Responsibility” in Owen, David G., ed., Philosophical Foundations of Tort Law (Oxford: Clarendon Press, 1995) 320 at 322-30Google Scholar.

34. It may seem strange to criticize Fletcher, of all people, for making an argument that is insufficiently normative, given that Fletcher has consistently emphasized the normative over the descriptive in his account of basic concepts of criminal law. I see Fletcher’s argument concerning reasonably mistaken self-defense as an unfortunate lapse from his usual method.

35. Fletcher, , Rethinking, supra note 12 at 761.Google Scholar

36. Ibid. at 762.

37. Ibid. at 763.

38. Ibid. at 764-67.

39. Curtis might make a lucky guess as to who was in the right here: he might have a correct “hunch” that Green was acting lawfully or an incorrect “hunch” that the shopkeeper was acting lawfully. Gardner’s view might be that Curtis would be excused if he helped the shopkeeper and vindicated if he helped Green. See Gardner, , “Justifications”, supra note 7 at 125 n. 39 Google Scholar. Note that a “hunch” need not be reasonable to vindicate. It is unclear what effect Gardner thinks vindication ought to have on criminal liability: if the usual requirement that a belief about a justifying or excusing condition being reasonable were imposed, a person who was vindicated in this sense would not be justified.

40. The scenario in this paragraph was inspired by the puzzle presented in Christopher, Russell, “Self-Defense and Defense of Others” (1998) 27 Phil. & Pub. Affairs 123 Google Scholar; see also Ripstein, Arthur, Equality, Responsibility and the Law (Cambridge: Cambridge University Press, 1999) at 201 Google Scholar n. 50.

41. Parents who have tried to adjudicate disputes between young children may be familiar with this sensation.

42. Horder, supra note 24 at 145.

43. Cf. Ripstein, supra note 40 at 199.

44. I do not mean to suggest that all behavior that is not criminally punishable is reasonable, rather that all behavior that is reasonable is not criminally punishable. Thus, a reasonable mistake of fact should always be a good defence to a criminal charge; but for some offences, even an unreasonable mistake of fact might, for other reasons, be a good defence.

45. Something like this seems to underlie the argument in Gardner, “Justifications”, supra note 7 at 117-22.

46. Gardiner, , “Reasonable Person”, supra at 274 Google Scholar, original emphasis.

47. Gardner, , “Gist of Excuses”, supra note 10 at 578.Google Scholar

48. Ibid. at 579, original emphasis.

49. Gardner, John, “On the General Part of the Criminal Law” in Duff, R. Antony, ed., Philosophy and the Criminal Law (Cambridge: Cambridge University Press, 1998) 205 at 237-38Google Scholar.

50. Brudner, Alan, The Unity of the Common Law (Berkeley: The University of California Press, 1995) 229-40Google Scholar.

51. I have discussed this possibility with Brudner, but he has not, as far as I know, adopted it in print. See also the English cases mentioned infra in note 59.

52. Brudner favours fault-differentiation of offences rather than outright acquittal for some unreasonable mistakes of fact: see Brudner, Alan, “Proportionality, Stigma and Discretion” (1996) 38 Crim. L.Q. 302.Google Scholar

53. Dressler, supra note 21 at 92-93 (characterizing but not adopting a rule-utilitarian account); see also Yeo, Stanley, Compulsion in the Criminal Law (Sydney, AU: The Law Book Company, 1990) at 32 Google Scholar (treating reasonably mistaken self-defence as a justification, based on a “balance of harms” test).

54. For Ripstein, , supra note 40 at 180-82Google Scholar, this is the source of the familiar rule that a mistake of law is not a defence.

55. Ibid. at 191-93.

56. Ibid. at 193.

57. Fletcher has rightly made this idea a central theme of his work on defences; see, for instance, Fletcher, , Rethinking, supra note 12 at 798–810 Google Scholar. See also Chapman, Bruce, “A Theory of Criminal Law Excuses” (1988) 1 Can. J. L. & Juris. 75.Google Scholar

58. This type of excuse is obviously rather different from those outlined in Gardner, “Gist of Excuses”, supra note 10. The requirements Gardner imposes on an excuse seem more likely to generate justifications than excuses.

59. Self-defence claims that are mistaken in other ways could, in principle, support excuses rather than justifications. See, for instance, Re S. Christian 872 P.2d 574 (Cal. 1994) (honest but unreasonable belief in the need to act in self-defence reduced murder to manslaughter). This might be one way of understanding the anomalous English rule that a mistaken belief in the need for self-defence need only be honestly held, not reasonable, to result in an acquittal. See R. v. Williams, [1987] 3 All E.R. 411 (C.A.); Beckford v. Queen [1987] 3 All E.R. 425, [1987] 3 W.L.R. 611 (PC); and the discussion in Smith, , supra note 17 at 103-7Google Scholar. Or one might, in principle, treat a disproportionate response to a reasonably perceived threat as a partial excuse, along the lines of provocation. In R. v. Faid (1983) 2 C.C.C. (3d) 513, the Supreme Court of Canada considered this possibility but rejected it, not primarily on grounds of principle but rather because there was no statutory basis for it.

60. A classic discussion is found in Dan-Cohen, Meir, “Decision Rules and Conduct Rules: On Acoustic Separation in Criminal Law” (1984) 97 Harv. L. Rev. 625.Google Scholar

61. See Raz, Joseph, The Authority of Law (Oxford: Clarendon Press, 1979) at 210–29 Google Scholar. I refer to the “instrumental demands” of the rule of law because I do not wish to rule out the possibility that the rule of law makes more substantive demands as well, which might well apply equally to decision rules and to conduct rules.

62. To take another example, ex post facto changes in the law of evidence are generally not thought to present problems for the instrumental value of the rule of law because, except perhaps in the law of privilege, people do not rely on those rules in governing their affairs. It is hard to see, for example, that anyone (apart from the legal profession) has a certainty interest in the rule against hearsay: no-one can complain when a court creates a new exception for a reliable form of hearsay. See, for example, R. v. B. (K.G.), [1993] 1 S.C.R. 740 at 782-83.

63. Curiously, most criminal statutes do not ex Pressly direct citizens not to engage in the conduct they prohibit; rather, they define offences and set out punishments.

64. Fletcher, George, “The Nature of Justification” in Shute, Stephen, Gardner, John & Horder, Jeremy, eds., Action and Value in Criminal Law (Oxford: Clarendon Press, 1993) 175 at 177CrossRefGoogle Scholar.

65. Ibid. at 180.

66. Thorburn, Malcolm, “The Gist of Justifications” (2003) [unpublished, Columbia Law School] at 32–34.Google Scholar

67. Fletcher, George, “The Individualization of Excusing Conditions” (1974) 47 S. Cal. L. Rev. 1269 at 1304.Google Scholar

68. Fletcher has argued that the principles governing justifications cannot be, and should not be, as precise as offence definitions. Fletcher, “The Nature of Justification”, supra note 65 at 180-81. It is one thing to say that the principles governing justifications may be looser than offence definitions; it is another thing to say that the principles governing justification are so loose that their application can depend on considerations that are not reasonably available to the actor. My argument requires rejection of the latter, but not of the former.