1. This terminology originated with Bentham. See An Introduction to the Principles of Morals and Legislation (Burns, J.H. & Hart, H.L.A. eds., 1970), ch. 8.Google Scholar

2. From here on I shall generally drop the “merely.” It should he kept in mind, however, that I assume throughout that intended consequences are themselves always to some extent foreseen, so that the real issue is what has to be added to foresight in order to produce intention.

3. For a recent, sympathetic discussion of the DDE, see Boyle, Joseph M. Jr., Toward Understanding the IMnciple of Double Effect, 90 Ethics 527–38 (1980).CrossRefGoogle Scholar Among its more sophisticated critics are Bennett, Jonathan, Morality and Consequences, in The Tanner Lectures on Human Values, vol. 2Google Scholar (McMurrin, Sterling ed., 1981Google Scholar); Davis, Nancy, The Doctrine of Double Effect: Problems of Interpretation, 65 Pacific Phil. Q. 107–23 (1984)CrossRefGoogle ScholarPubMed; and Kagan, Shelly, The Limits Morality, ch. 4 (1989).Google Scholar

4. The distinction also appears to play a role in the civil law. For example, certain forms of governmental discrimination against a suspect class may be constitutional if the government acts knowingly but not if it acts purposefully; see discussion in Simons, Kenneth W., Rethinking Mental States, 72 B. U. L. Rev. sec. I (1992).Google Scholar

5. For a concise summary of this debate and the confusion it has engendered in British criminal law, see White, Alan, Misleading Cases, ch. 4 (1991).Google Scholar

6. Hyam, [1974] 41Google Scholar All E.R. In addition, physicians and other medical personnel (such as Dr. Jack Kervorkian) who have been accused of mercy-killing or assisted suicide often attempt to defend themselves by arguing that their actions were not motivated by an intention to cause death (even though that was certainly a foreseen consequence).

7. For a very useful discussion, see Adams, Frederick, Intention and Intentional Action: The Simple View, 1 Mind & Lang. 285–93 (1986).CrossRefGoogle Scholar RA. Duff has suggested that ascriptions of intentionality may also depend on normative factors such as “the seriousness of the expected effect, the character of the action which causes it, and the adequacy of the precautions taken [against bringing about the expected effect] (Intention, Agency & Criminal Liability 98 (1990)Google Scholar). This view is criticized in Laing, J.A., The Prospects of a Theory of Culpability: Mens Rea and Methodological Doubt, 14 Oxford J. Legal Studies 74–80 (1994).CrossRefGoogle Scholar

8. Although I shall follow the usual practice of using these terms more or less interchangeably, it is worth noting that in fact most intrinsic ends are not ultimate ends, i.e., ends valued only for their own sake, but arc also valued for their instrumental importance in facilitating the occurrence of further ends. I value playing golf, for example, not only because of the intrinsic pleasure it affords (on those rare occasions when it is not an exercise in utter futility) but also because of its health benefits and the opportunities it affords for socializing.

9. It is worth distinguishing between a means for achieving some objective and an agent's means for doing this. The former is the fundamental, objective notion in terms of which the latter is defined. My means for achieving some goal is what I believe would be a means for so doing (and, obviously, I am sometimes mistaken about this). Hence it may turn out that what I take to be a means to some goal—and thereby intend—is, in reality, but a side effect of my act because it does not, contrary to my expectations, causally contribute to any of my ends.

10. This formulation is not entirely accurate since it could be interpreted to classify as a means anything that is on the causal path to the agent's goal. Suppose, for example, that in order to achieve some morally important objective, I must run over (or shoot through) some innocent person who is physically in my way (tins sort of case is discussed by Quinn, Warren S., Actions, Intentions, and Consequences: The Doctrine of Double Effect, 18 Phil. & Pub. Aff. 345 (1989)).Google ScholarPubMed My own intuition is that any harm done to such a person would be only a side effect of my act, not a means to my goal (Kagan thinks otherwise; see his discussion of the difference between “strong” and “weak” means in Kagan, , supra note 3 at 138–44Google Scholar). However, the proposal in the text would apparently classify it as a means because its occurrence would causally contribute to the realization of the goal in the sense that the precise manner in which the goal was actually realized would have been causally affected by whatever wns on the relevant causal path. What seems needed here is an appeal to the distinction between something's being required for the occurrence of a particular concrete state of affairs that serves to realize whatever goal is in question and something's being causally required in order for the goal to be realized somehow. Something which is merely on the causal path to my goal need be causally required only in the first of these two ways.

11. Hence a side effect can be temporally prior to, simultaneous with, or subsequent to either the means or the end. I shall suppose that, all else equal, it makes no moral difference which of these slots a side effect occupies; for a contrasting view see Kamm, F. M., Nonconsequentialism, the Person as an End-in-Itself, and the Significance of Status, 21 Phil. & Pub. Aff. 378–79 (1992).Google Scholar

12. Actually, it is by no means clear that even the distinction between effects which are intended as ends and those which are only foreseen has any real moral significance. Jeff McMahan (in private correspondence) has raised the following challenge:

Consider two cases: (1) One intends some bad effect, B, as one's intended intrinsic end but foresees that one's action will have some good effect, G, that is not intended either as a means or as an end. (2) One intends G as one's end but foresees that one's action will have B as an unintended side effect. In both cases G outweighs B. Can we give reasons for thinking act 1 is worse than act 2 (or, more strongly, that 2 is permissible while 1 is not)? Certain consequentialists would be unable to find any difference between them.

Indeed, I would be prepared to go further and suggest that the very notion of an “intrinsic” end masks serious philosophical problems and that we might do better to eschew reference to this concept entirely and to speak, instead of something like the agent's overall plan of action. However, my reason for not pursuing these complexities here is that it is difficult to see how one could plausibly defend the significance of the intended/foreseen contrast without assuming both that the notion of an intrinsic end is (relatively) unproblematic and that there is a moral difference between intending something as an end and merely foreseeing it. My purpose in this paper is to argue that, even if these assumptions are granted, there are still no persuasive grounds for distinguishing morally between effects that are instmmentnlly intended effects and those that are merely foreseen.

13. See Boyle, , supraGoogle Scholar note 3, passim, for an extensive discussion of all of the relevant conditions.

14. Bratman, Michael, Intention. Plans and Practical Reason 155–60 (1987).Google Scholar

15. Id. at 156.

16. See, e.g., Beardsley, Monroe C., Intending, in Values and MoralsGoogle Scholar (Goldman, Alvin I. & Kim, Jaegwon eds., 1978Google Scholar), and Audi, Robert, Intending, Intentional Action, and Desire, in The Ways of DesireGoogle Scholar (Marks, Joel ed., 1986Google Scholar). Such a view has also enjoyed considerable support in legal circles. See Dressier, Joshua, Understanding Criminal Law 98–99 (1987).Google Scholar

17. Additional criticisms of the desire model are found in Moore, Michael, Intentions and Mens Rea, in Issues in Contemporary Legal Philosophy 246Google Scholar (Gavison, Ruth ed., 1987Google Scholar) and Duff, , supra note 7, at 52–73.Google Scholar

18. A good example of this sort of strategy is Duffs “test of failure”; see Duff, , supra note 7, at 61–63.Google Scholar Although I cannot argue for it here, I believe this proposal is also valnerable to the criticism mentioned in the text.

19. This line of criticism is developed more fully in Bennett, , supra note 3, at 100–05.Google Scholar Further problems are noted in Norrie, Alan W., Oblique Intention and Legal Politics, Crim. L. R. 795–96 (1989).Google Scholar

20. Bratman, . supra note 14, at 140–43.Google Scholar

21. Thomas Nagel grounds his account of the significance of the intended/foreseen distinction in terms of a similar understanding of its nature; see The View From Nowhere 180–185 (1986).Google Scholar For some important objections, see Kagan, , supra note 3, at 166–70.Google Scholar

22. A somewhat similar case is discussed by Bratman, but his example is less extreme since he docs not suppose that the pilot is aiming at the civilians. See Bratman, , supra note 14, at 191–92 (n. 6).Google Scholar

23. One might argue that there is some doubt as to whether TAB would fully exhibit the second (or even the third) of these dispositions with respect to the civilian casualties. For suppose that he has knowledge of a bomb shelter that the Civilians are unaware of and forms the intention of informing them of this (out of a humanitarian desire to minimize their losses). Assuming that what is needed is only their being visible just prior to the dropping of the bombs, it would seem that he need not be disposed to constrain this further intention, which seems to imply that he does not possess everything involved in the second practical reasoning disposition with respect to the prospect of causing the Civilian casualties. A similar sort of case could be used to argue that he lacks the third disposition as well.

Friends of the intended/foreseen distinction, however, should be reluctant to interpret intentions in such a fine-grained manner since doing so provides even TEB with an escape hatch. After all, as Bennett has pointed out, TEB could maintain that he too lacked this disposition with respect to the deaths of the Civilians (and hence did not really intend them) since his objective would have been fully achieved as long as “the people's bodies should be inoperative for long enough to cause a general belief that they were dead, this belief lasting long enough to speed the end of the war” (Bennett, , supra note 3 at 111Google Scholar). For an ingenious and sophisticated attempt to deal with the problem of “shaving” one's intentions, see Quinn, , supraGoogle Scholar note 10; for important criticisms, see Fischer, john Martin, Ravizza, Mark, & Copp, David, Quinn on Double Effect: The Pmblem of “Closeness” 103 Ethics 707–25 (1993)CrossRefGoogle Scholar and McMahan, Jeff, Revising the Doctrine of Double Effect, 11 J. App Phil. 201–12 (1991).CrossRefGoogle Scholar

24. Dworkin, Gerald, Intention, Foreseeability, and Responsibility, in Responsibilty, Character, and The Emotions 348Google Scholar (Schoeman, Ferdinand ed., 1987).Google Scholar

25. Id. at 351.

26. Id. at 352.

27. Id. at 352 (emphasis added).

28. Moreover, some of Dworkin's suggestions seem premised upon a misunderstanding of what is required to defend the direct/indirect distinction. For example, the claim that, in cases of indirectly intended harm, the agent is more able to alter his plans than in cases of directly intended harm, would seem to be true only if everything is not equal, i.e., only in cases where the indirectly intended consequence is not required for the achievement of the agent's goal.

29. Glanville Wlliams, Oblique Intention, 46 Cambridge L. J. 436 (1987).

30. See, e.g., J.F. Lever, Means, Motives, and Interests in the Law of Torts, in Oxford Essays in Jurisprudence, First Series (A.G. Guest ed., 1961).

31. “An intentional invasion of another's interest in the use and enjoyment of land is unreasonable if the harm is significant and the actor's conduct is (a) for the sole purpose of causing harm to the other …”(Second Restatement of Torts sec. 829 (emphasis added).

32. In fact, the criminal law does evince a concern with a person's intrinsic objectives, as is evidenced (for example) by the general willingness to classify reckless homicide as murder where the conduct involved has no redeeming social value; see George Fletcher, Rethinking Criminal Law 265 (1978).

33. [1915] 1 K. B.616.

34. Williams, supra note 29, at 437.

35. Id. at 437.

36. Other important issues that need to be addressed in cases of this sort are discussed by Duff, Intentions Legal and Philosophical, 9 Oxford J. Legal Studies 84–85 (1989).

37. Williams, supra note 29, at 425–30.

38. Id. at 426.

39. Duff, supra note 7, at 176.

40. It is interesting to note that, according to Scottish law, “wicked recklessness,” i.e., a complete indifference as to the saferty of another person, is also sufficient for possessing the mens rea murder. For a provocative defense of this view (together with the appropriate legal citations), see Robert Goff, The Mental Element in Murder, 104 LAW Q. R. 30–59 (1988). For a spirited rejoinder, see Glanville Williams, The Mem Rea for Munler Leave It Alone, 106 Law Q. R. 387–97 (1989).

41. Duff, supra note 7, at 113 (emphasis in original).

42. Charles Fried, Right and Wrong 40–41 (1978).

43. Boyle, supra note 3, at 536–37. Similar sentiments arc expressed by Robert Nozick, Philosophical Explanations 497 (1981), and John Finnis, Intention and Side-Effects, in Liability: New Essays in Legal Philosophy 61–64 (R.G. Frey & C.W. Morris eds., 1991).

44. This sort of closeness is to be sharply distinguished from another sort that is often discussed in connection with the Doctrine of Double Effect, viz., that which is sometimes alleged to hold between the intended and foreseen effecLs themselves; for a discussion of the latter, see Fischer, Ravizza and Copp, supra note 23.

45. Sanford Levy, The Principle Double Effect, 20 J. Value Inquiry 37–39 (1986). Bennett seems to be making a similar point when he asserts that “[t]hc concept of intention has a role [only] in second-order morality, i.e., in guiding judgments on people in respect of past actions,” and that “it is a mistake to think of first-order morality–morality for the guidance of deliberating agents–as making any use of the concept of the delibcrator's future intentions” (Bennett, supra note 3, at 97).

46. Boyle, for example, argues that the DDE is best construed as a principle at justification rather than a principle of excuse. See Boyle, supra note 3, at 529–33.

47. Duff supra note 7 at 112. He seems to assume, however, that consequentialists cannot recognize such harms; for a criticism of this assumption, see Gorr, Motives and Rightness (forthcoming in Philosophia).

48. Gorr, supra note 47.

49. I should note that while I am an “objectivist” with respect to questions of liability (since I hold that the latter is not determined exclusively by the agent's psychological states) I am a “subjectivist” with respect to culpability (which I maintain is so determined). Duff is an objectivist with respect to both since he has argued that “What relates me to a criminal harm as a responsible agent is partly the subjective connection of my action to that harm, and partly its objective connection.” (Duff, , supra note 7, at 204ffGoogle Scholar). For a rigorous defense of a thorough going subjectivist account of culpability, see Michael Zimmerman, An Essay on Moral Responsibility (1988).

50. I hasten to point out that a satisfactory formula will almost surely be more complicated than this. Many who are not extreme noncomequcntialists and who subscribe to a mixed theory of punishment would still insist that liability should at least be Umitedhy culpability. On some such accounts, dangerousness would serve only to help determine the precise amount of punishment within that limit. (I am grateful to one of the referees for reminding me of this.)

51. I am assuming, of course, that there is no plausibility in the supposition that the former is more culpable than the latter.

52. Goff, , supra note 40, at 57.Google Scholar

53. A referee for this journal suggested that the motivation for equating intended means to intended ends rather than to foreseen side effects may be “the Kantian version of deontology, according to which the objects of morality's agent-relative prohibitions are intentions, not actions.” There are, however, at least two reasons why, even if this is the case, it is unlikely to provide much support for those who think the law should recognize the intended/foreseen distinction. First, the law is primarily concerned with actions, at least in the sense that intentions in the absence of actions do not come within its purview. Second, even if psychological states are taken to be the primary locus of moral/legal evaluation, the fundamental question at issue is simply transposed. Instead of talking about the alleged moral difference between consequences intended as means and consequences that are only foreseen, we would simply concern ourselves with the moral difference between the psychological state of intending as a means and the psychological state of merely foreseeing. All of my criticisms would apply, mutatis mutandis, as much to the latter distinction as they do to the former.

54. As Gilbert Harman has observed, side effects will normally not be as salient as means or ends, a fact that is of considerable importance in a theory of rational decision making. See Harman, , Change in View, ch. 9 (1986).Google Scholar

55. Duff, Antony, Intention, Responsibility and Double Effect, 32 Phil. Q. 4 (1982).CrossRefGoogle Scholar I disagree with Duff's contention that ascriptions of intentionality are literally senseless in such contexts: it is enough to note that they are at least highly misleading by virtue of what they conversationally imply.

56. It is revealing that Duff, while attempting to sketch the contours of a nonconsequentialist theory of responsibility in which the intended/foreseen distinction would make sense, insists that merely foreseen effects, though unintended, are nevertheless sometimes brought about intentionally. See Duff, supra note 7, at 76–82, 109–15.Google Scholar For the very fact that Duff feels the need to make such a concession suggests that even he finds it difficult to maintain the rather sharp separation between the intended and the foreseen that his view commits him to. In any event, there are good reasons to doubt that any such distinction exists; see Adams, supraGoogle Scholar note 7. And even if there were good reasons for acknowledging such a distinction, it would not appear to strengthen the position of those who ascribe normative significance to the intended/merely foreseen contrast. For in that case we would simply have to ask all the same questions about the significance of the distinction between intended effects and effects which are (merely) intentional.

57. Collingwood, R.G., An Essay on Metaphysics 285–327 (1940).Google Scholar