I would like to thank members of the John Howard Society and the International Association for Philosophy of Law and Social Philosophy (IVR) for comments in Toronto, June 1990. The comments of Roger Shiner and Kadri Vihvelin were the source of many second thoughts and some revisions, though I did not, in the end, modify my view to accommodate their skepticism. I also must thank colleagues at Dalhousie for their comments on an earlier draft.

1. Murphy, Jeffrie C., “Mercy and Legal Justice,” in Coleman, J. & Paul, E. Frankel, eds, Philosophy and Law (Oxford: Basil Blackwell, 1987) 1.Google Scholar

2. Much of Murphy’s discussion is based on earlier discussions by Alwynne Smart, Claudia Card, and P. Twambley. See Smart, A., “Mercy,” (1968) 43 Philosophy 345 Google Scholar, reprinted in Acton, H., The Philosophy of Punishment, (New York: St. Martin’s Press, 1969);Google Scholar Card, C., “On Mercy” (1972) 43 Philosophical Rev. 182;Google Scholar Twambley, P., “Mercy and Forgiveness” (1976) 36 Analysis 84.Google Scholar The references to Smart below are to the reprint in Acton

3. Twambley and Murphy both hold that mercy makes good sense in relation to civil suits, where individuals are free to waive their right to compensation. Murphy’s paper, for example, attempts to provide a contractarian basis for the recognition of a rule which would allow private individuals room for mercy in relation to defendants against whom they have a claim.

4. Reference is to the 1888 Selby-Bigge version of the Treatise as revised by Nidditch, P.H. (Oxford: Clarendon Press, 1978).Google Scholar In this second edition, the pagination of the original Selby-Bigge edition is preserved.

5. I should make clear that the recent skeptical position is very different from Hume’s except in its conclusion. The strategy of the former, as we will see, is to show that whatever might plausibly be thought of as a manifestation of the independent and opposed virtue of mercy is either a part of justice, properly understood, or not a virtue. Hume’s position, on the other hand, clearly involves the possibility of conflict, at least between justice and humanity. But, on Hume’s view, to allow an appeal to humanity in order to constrain the pursuit of justice is to allow one’s judgment to be swayed by considerations which are less important. Thus, if mercy necessarily involves deferring to what should be overridden, it cannot be a virtue. Moreover, two quite different conceptions of justice are involved. On Hume’s view, justice is the disposition to comply with established conventions which confer rights on individuals and to act in ways that maintain these conventions. The concept of justice which the contemporary paradoxes presuppose is the virtue of retributive impartiality. This complicates the business of comparison. But it is interesting that one can proceed from quite different theoretical bases in generating skepticism about mercy. I am preparing a discussion of the Humean opposition between justice and humanity (“Hume: Conflict of Virtues”) I argue that there is not a good basis in Hume’s own theory of the virtues for his assumption that justice must take precedence over humanity.

6. There are papers which deal with the general issue in a way that is intended to resolve the supposed conflict between mercy and justice, viz., Hestevold, H., “Disjunctive Dessert,” (1983) 20 A.P.Q. 357;Google Scholar Roberts, H., “Mercy” (1971) 46 Philosophy, 352;Google Scholar Stebra, J., “Can a Person Deserve Mercy?” (1979) 10 J. Soc. Phil. 11;Google Scholar Sverdlik, S., “Justice and Mercy” (1985) 15 J. Soc. Phil. 36.Google Scholar But these solutions do not re-establish mercy as a prerogative of a judge in criminal cases. Hestevold says at 358, for example, I will not claim that mercy, as I will analyze it, is manifested by our present systems of institutional justice, nor will I argue that it should be. In this paper I defend both the claim that mercy is manifested by our legal system and the claim that it should be.

7. Some of Alwynne Smart’s examples of mercy are of this type, supra, note 2.

8. Claudia Card, supra note 2.

9. See Fauteux, G. et al., Principles and Procedures Followed in the Remission Service of the Department of Justice of Canada (Ottawa: Queen’s Printer, 1956)Google Scholar c. IV “The Prerogative of Mercy” at 32: “Both kinds of pardon [free pardon and ordinary pardon] proceed from the same source as an act of grace, but the first is an act of grace to which the recipient is morally entitled, while the second is a pure act of grace.” Obviously, the present discussion is not concerned with pardons of the first sort, based on the innocence of the accused. (These are “acts of mercy” relative to the power, as opposed to the authority, of the state.) The Fauteux report deals with procedures through which the judgments of judges can be altered (in the direction of leniency) by political authorities outside of the courts. Though the present discussion focuses on the discretion of judges in discussing mercy, precisely the same problems can be raised in considering “ordinary pardons” based on compassion. The Report, like this paper, defends remissions of this type, see at 36.

10. The fact that B’s act would be described as a “mercy killing” adds an unintended complication. It should suffice to say that the skeptics addressed in this paper are not denying the possibility of genuine acts of mercy outside of the context of legal punishment. Although this is not mercy in the sense under discussion here (since there is no blameworthy act) it is interesting to note that much of the forgoing account of mercy does fit this case. In the case described (i) B is not just doing what (moral) justice requires, (ii) It would be a grotesque joke to suppose that compassion for some other person (e.g. some family member for whom the wife had become a burden) could be the motive for an action properly described as “mercy killing”, (iii) There is a sense in which B’s action is optional. Few people have argued for a positive right to die, which would entail that B (or someone) had an obligation to kill. Of course, one could hold this position. My analysis would imply that the defenders of such a position would not be speaking with strict accuracy if they continued to speak of this as mercy killing.

11. Murphy, supra, note I at 6–7.

12. Here I ignore the very real possibility of interpretive differences based on case law.

13. The case derives from Smart’s discussion and is reconsidered by Murphy.

14. This is a part of what Claudia Card argues: “One sometimes gets the impression that the latter phrase [‘tempering justice with mercy’] is taken to mean something like “being a little unjust in order to make room for a little charity.” (“On Mercy”, supra, note 2 at 191.) She argues for the “coherence” of justice with mercy. The “mercy principle” allows the court to temper legal justice with considerations which we generally relegate to the category of “cosmic” or “poetic” justice.

15. Or, where the term had lost much of its present significance.

16. Twambley, supra, note 1 at 85–86; Murphy, supra, note 1 at 9–12.

17. There are some important exceptions, e.g., Canada’s Narcotics Control Act which makes the minimum penalty for importing narcotics a seven year sentence. This, of course, excludes any considerations of mercy which would move the judge below that minimum.

18. I shall ignore here the complications generated by the fact that crimes do not always involve harms (consider attempts, victimless crimes, etc.).

19. One could argue, of course, that the willingness of the victim to press charges is a relevant difference which can be relied on to distinguish these assault cases. This just ignores the point that the same crime will get different treatment; the assault does not cease to be a crime simply because its victim will not press charges.

20. The skeptics might attempt to support their position by reference to the specific legal sense which ‘mercy’ has taken on in connection with executive clemency. In a passage cited (with approval) by the Supreme Court of Canada, Dicey speaks of this as an “arbitrary” power implicit in the Royal prerogative to grant pardons, which connects historically with supremacy of the sovereign: The ‘prerogative’ appears to be historically and as a matter of actual fact nothing else than the residue of discretionary or arbitrary authority, which at any given time is legally left in the hands of the Crown. [Cited from Dicey, , Law of the Constitution, in Clayton Ruby, Sentencing (Toronto: Buttersworth, 1976) at 421.]Google Scholar It should be clear, however, that this paper is not about this specific legal sense of ‘mercy,’ but about propriety of mercy, as it is ordinarily understood, as a factor in legal decisions concerning punishment.

21. Boyle, C.K. & Allen, M.J., Sentencing Law and Practice, (London: Sweet & Maxwell, 1985) at 272–73.Google Scholar

22. Ronald Dworkin has alerted us to the difficulties one can get into in thinking about the exercise of discretion. See, e.g., Taking Rights Seriously (Cambridge, Mass.: Harvard University Press, 1977)Google ScholarPubMed chs. 2.5 and 3.4. From the fact that a legal decision-maker has discretion it does not follow that the decision is unconstrained by legal principles or that it is in any sense arbitrary. It need not even be true that the decision is not subject to appeal. In the passage, from Boyle & Allen above, mitigation is said to be “very much in the discretion of the judge,” but the authors proceed to articulate a nest of rules and principles which constrain these decisions—and, most of these constraints are themselves gleaned from cases in which sentence were appealed. The judge must do her best to arrive at a reasonable judgment based on a whole set of factors which may compete with each other. The judge is constrained by a set of principles which may also compete and which cannot be (or at least have not been) lexically ordered.

23. See, for example, the report of the Canadian Sentencing Commission summary, Sentencing Reform: A Canadian Approach (Ottawa: Supply and Services Canada, 1987).Google Scholar “The Paramount principle governing determination of a sentence is that the sentence be proportionate to the gravity of an offense and the degree of responsibility of the offender for the offense” (5).

24. Many are not matters of mercy. That a defendant is pregnant may be a basis for lessening a sentence, though it is not necessarily a reason that connects with the welfare of the offender. That a certain type of crime is becoming prevalent may be a reason for a “deterrent sentence” which sets an example.

25. Barbery (1975) 62 Cr. App. R. 248. This and the following cases are sorted from a discussion of mitigation in English cases, supra, note 21 at 272–79. See supra, note 20.

26. Tierney (1982) Crim. L.R. 53. See Boyle & Allen, supra, note 20 at 278.

27. Supra, note 21 at 273.

28. Fell (1975) Crim. L.R. at 349.

29. Fauteux, supra, note 9 at 36.

30. I am assuming here the mixed position on the justification of punishment defended by, e.g., Hart, H., Punishment and Responsibility (Oxford: Oxford University Press, 1968)Google Scholar and Ten, C., Crime, Guilt, and Punishment (Oxford: Oxford University Press, 1987).Google Scholar