1 For a discussion of proportionality, see Thomas Hurka ‘Proportionality in the Morality of War,’ Philosophy & Public Affairs 33 (2005): 34-66.

2 See Daniel Statman ‘On the Success Condition for Legitimate Self-Defense,’ Ethics 118 (2008), 659. I have substituted his use of ‘immoral’ with ‘impermissible.’ I don't think anything hinges on this substitution.

3 Banbaji, Yitzhak ‘Culpable Bystanders, Innocent Threats and the Ethics of Self-Defense,’ Canadian Journal of Philosophy 35 (2005), 586Google Scholar

4 Ibid. 608

5 Ibid. 610

6 It could be that Benbaji means that what really does most to explain the right to use force against culpable aggressors is their reduced claim on protection, and that the moral interest of the victim in avoiding disrespectful treatment is only a minor addition. If so, the paper is not very good at conveying this idea. Anyway, my aim is to demonstrate that the interest in avoiding disrespectful treatment will only play a very minor role, if any role at all, and not to show that Benbaji is in the wrong.

7 There is no option to punish the wrongdoer later, or to compensate for the lost fingers.

8 Some people may not think it permissible to proceed in White and Grey because protecting fingers is not enough to warrant killing even very culpable aggressors. They may envisage that limbs rather than fingers are at stake, and should now see a difference between these two cases and a corresponding Blue case.

9 Statman ‘On the Success Condition for Legitimate Self-Defense,’ 670.

10 Ibid. 669

11 Ibid. 672

12 Ibid. 678

13 I have no idea why ‘evil’ is used by Statman. And he doesn't always use it: ‘However, when the threat to Victim's honor is parasitical, so to say, on other threats — to her property, bodily integrity, or life (‘parasitical offenses’), the measures she is allowed to take should be proportionate to the primary objects of harm rather than to the offense to honor implied by being subject to these threats’ (ibid. 678).

14 See Statman ‘On the Success Condition for Legitimate Self-Defense,’ 663, where he refers to Jeff McMahan who held the view that the proportionality restriction governing self-defense against an innocent attacker is stronger than that governing self-defense against a culpable aggressor. See McMahan, ‘Self-Defense and the Problem of the Innocent Attacker,’ Ethics 104 (1994), 265-266. And while Statman does seem to think that culpability matters, he relies on Benbaji's account of it, see Statman ‘On the Success Condition for Legitimate Self-Defense,’ 682-3.

15 Again, one might replace fingers with limbs here without distorting the argument.

16 See Statman ‘On the Success Condition for Legitimate Self-Defense,’ 664.

17 Ibid. 664

18 However, see Statman ‘On the Success Condition for Legitimate Self-Defense,’ 686, where Statman seems to indicate that it is impossible to protect one's honour if one doesn't fight. Interestingly, a referee also questioned ‘how ‘firing his gun in the air’ might seriously contribute to saving Wayne's honor. Actually that would seem to make things worse for Wayne and expose the extent of his helplessness.’ I really don't understand this view. What matters for saving one's honour, I would think, is the attitude you have toward what's facing you, not whether you are able to get in a few futile shots.

19 I know that has been a historical practice in certain situations where some chivalrous behavior has been explained in terms of defending another's honour. And indeed, if that is what one believes, I would guess that is what one achieves.

20 I owe this proposal to an anonymous referee of this journal.

21 See ‘On the Success Condition for Legitimate Self-Defense,’ 678, quoted earlier in this paper.

22 ‘Whenever victims of aggression are overwhelmed by an aggressor but, nonetheless, find the courage to rise against him through some form of determined resistance, however hopeless, they are thereby reaffirming their honor, or — to use a rather out-of-date expression — they show themselves to be men (or women) of honor’ (ibid. 679).

23 I owe this proposal to Jeff McMahan.

24 And here I side with Statman, who admits that he feels more certain about the existence of the puzzle than about his ability to offer a convincing solution to it (see Statman ‘On the Success Condition for Legitimate Self-Defense,’ 666).

25 In fairness to Benbaji, it should be observed that he is aware of this problem and tries to meet it with considerations that are supposed to further beef up the interests of the victim. These attempts do not seem very promising though.

26 Statman ‘On the Success Condition for Legitimate Self-Defense,’ 667

27 It could in addition be that while moral culpability reduces an aggressor's claim on protection, the victim's interest in protecting herself against disrespectful treatment, to protect her honour, or perhaps otherwise to promote one of her interests, creates further latitude for actions against the culpable aggressor. While this may be true, I have been unable to find examples that clearly indicate any such additional increase. I therefore leave this open.

28 There might be further reasons for distinguishing morally between two types of innocent agents and aggressors. One might, for instance, distinguish between innocent aggressors that have engaged in risky activity and those that have not (see McMahan, ‘The Ethics of Killing in War,’ Ethics 114 (2004): 693–733CrossRefGoogle Scholar and ‘The Basis of Moral Liability to defensive Killing,’ Philosophical Issues 15 (2005): 386-405). For instance, the driver of a car that goes out of control through no fault of her own is innocent but not non-responsible; she has freely taken up the risky activity of driving. McMahan's contention that there are distinctions to be made here seems plausible. Against McMahan, however, I would suggest that responsible agents who have become innocent threats by choosing to participate in risky activities have to bear more of the cost than ordinary innocent aggressors and threats, and not simply more than innocent bystanders. Being responsible for engaging in risky activities widens the range of actions we can take against those engaging in them, though the baseline is not what can be done towards innocent bystanders, but towards innocent non-responsible contributors.

29 For the view that it is impermissible to kill innocent aggressors in self-defence, see N. Zohar, ‘Collective War and Individualistic Ethics: Against the Conscription of ‘‘Self-Defense,’’’ Political Theory 21 (1993): 606-22; Otsuka, Michael ‘Killing the Innocent in Self-Defence,’ Philosophy and Public Affairs 23 (1994), 74–94;CrossRefGoogle Scholar McMahan, Jeff ‘Self-Defense and the Problem of the Innocent Attacker,’ Ethics 104 (1994);CrossRefGoogle Scholar Rodin, David War and Self-Defense (Oxford: Clarendon Press 2002), 70–99,CrossRefGoogle Scholar and McMahan, Jeff ‘The Basis of Moral Liability to defensive Killing,’ Philosophical Issues 15 (2005): 386–405.CrossRefGoogle Scholar

30 See, for instance, Thomson, Judith Jarvis ‘Self-Defence,’ Philosophy and Public Affairs 20 (1991): 283—310Google Scholar and Mapel, D. R. ‘Innocent Attackers and Rights of Self-Defense,’ Ethics and International Affairs 18 (2004).CrossRefGoogle Scholar

31 Benbaji and Statman may agree that we need a position between the two listed positions on the use of defensive force against innocent aggressors. The disagreement between us is in the explanation of why moral culpability matters. While they primarily think it has to do with the increased interests on the part of the victims, I think it has to do with the aggressor's reduced claim on protection.

32 I do that in Øverland, Gerhard ‘Forced Assistance,’ Law and Philosophy 28:2 (2009): 203–32,CrossRefGoogle Scholar and at length in ‘Dividing Harm’ forthcoming in the Journal of Moral Philosophy.

33 The asymmetry between innocent aggressors and their victims could be justified on contractual grounds. Central in such an argument would be the claim that being an initiator would often serve as a proxy for culpability. And, since being an initiator is a discernable external feature, as opposed to moral culpability, sacrificing the initiator, or at least shuffling cost in his or her direction, is likely to save more innocent people from harm in the long run as we are less likely to get it wrong. Epistemic uncertainty gives rise to a presumption in favour of having a rule that gives priority to the defending party since such a rule will give innocent contracting parties a higher chance of survival and of avoiding harm. See Øverland ‘Forced Assistance’ and ‘Dividing Harm.’

34 Choosing (3) would obviously be permissible, but it would be supererogatory.

35 Note that there could be a reduction in permissible level of imposed cost as compared with maximum required duty to shoulder cost in the sense that it would be permissible to impose less than what the person would have a duty to bear in the first place. Reasons for this could be respect for a person's free choice, or perhaps the mere psychological cost of being forced to do something. So that if 100 were the cost one was obliged to shoulder, the cost others were permitted to impose could be only 100 — q, where the constant q denotes any appropriate number reducing the level of enforcement cost. I shall assume there is no need for a q-moderator.

36 See Thomson, Judith Jarvis ‘Self-Defense and Rights,’ in Rights, Restitution and Risk, Patent, William ed. (Cambridge, MA: Harvard University Press 1986).Google Scholar For Thomson's later position, see her article ‘Self-Defense,’ Philosophy and Public Affairs 20 (1991): 283-310.

37 See, for instance, Uniacke, Suzanne ‘In Defense of Permissible Killing: A Response to two Critics,’ Law and Philosophy 19 (2000): 627–33.Google Scholar

38 See Thomson ‘Self-Defense,’ 290. Her example goes like this: ‘A villain is shooting at you, and your only defense is to run. But your only path to safety lies across a bridge that will hold only one person, and there is already a man on it; if you rush onto the bridge, he will be toppled off it into the valley below.’

39 That you may impose considerably more harm on the incapacitated aggressor than an innocent bystander does not imply that you may impose the same level of harm on him now that he is incapacitated as when he still posed a threat to you. Since he didn't cause you any harm there is no contribution aspect that transfers only the culpability factor.

40 Interestingly, it seems to me that it might even have been permissible to save your pet, rather than the incapacitated aggressor.

41 This was suggested by an anonymous referee of this journal.

42 Matters would be different, of course, if A made a conditional threat and said: ‘Your napkin or your life.’ See Øverland, Gerhard ‘Conditional Threats,’ Journal of Moral Philosophy 7 (2010): 334–45.CrossRefGoogle Scholar

43 For a discussion of many such cases see Øverland, Gerhard ‘Moral Taint: On the Transfer of the Implications of Moral Culpability,’ Journal of Applied Philosophy 28 (2011): 122–136.Google Scholar

44 Repentance will typically come with a feeling of sadness for one's past behaviour. The essential element of repenting is nevertheless not this feeling-sorry-for feeling, nor any kind of suffering on the part of the wrongdoer, but his or her improved moral character.

45 See Derek, Parfit ‘Personal Identity and Rationality,’ Synthese 53 (1982): 227–41,Google Scholar and Parfit, Derek Reasons and Persons (Oxford University Press 1984), 313Google Scholar and 341.

46 Hurka considers the number of soldiers that may be killed to defend one of our own, and doesn't think there is a limit. I disagree, and my disagreement will depend on the level of moral culpability on the part of the soldiers on the other side. See Hurka ‘Proportionality in the Morality of War,’ 58.

47 Some would probably say that even in the case of full culpability, one couldn't kill all the remaining people in the world. And this would be the case in both Luke and Wayne. Anyway, this might be enough of a special case that it can be set aside.

48 Note that if Luke's killing were in vain, it would have been just like the case of Wayne.

49 The observations in the last section explain why the success condition is more important with regard to the question of defensive force in war than private selfdefence. Taking up arms will normally include the death of innocents and not-soculpable individuals, civilians and soldiers alike, and that can normally be justified only if there is a likelihood of success. See Statman ‘On the Success Condition for Legitimate Self-Defense,’ 682.

50 I would like to thank Christian Barry, Jeff McMahan and an anonymous referee of this journal for helpful comments.