The Liability of Justified Attackers

Abstract

McMahan argues that justification defeats liability to defensive attack (which would have far-reaching consequences for the ethics of war, in particular for the thesis of the moral equality of combatants). In response, I argue, first, that McMahan’s attempt to burden the contrary claim with counter-intuitive implications fails; second, that McMahan’s own position implies that the innocent civilians do not have a right of self-defense against justified attackers, which neither coheres with his description of the case (the justified bombers infringe the rights of the civilians) nor with his views about rights forfeiture, is unsupported by independent argument, and, in any case, extremely implausible and counter-intuitive; and third, that his interpretation of the insulin case confuses the normative relations between an agent’s justification and non-liability (or lack thereof) on the one hand and permissible or impermissible interference with the agent’s act on the other. Similar confusions, fourth, affect his discussion of liability to compensation.

1 Introduction

Jeff McMahan claims that there is a moral inequality of combatants, in the sense that the combatants on the unjustified side in a war are liable to attack while the combatants on the justified side in a war are not.¹ There are different definitions of liability, but they all imply that someone who is liable to attack has no right not to be attacked, and thus would not be wronged by an attack. However, this moral inequality claim has been challenged: since many combatants on the justified side infringe the rights of innocent bystanders, for example by “collaterally” killing them, they would, on certain accounts of liability to attack, become liable to attack themselves (Steinhoff 2007, 95–97; and 2012a).

Jeff McMahan rejects such accounts of liability: he claims that “justification exempts one from liability to defensive attack” (2014a, 118). In the following I will argue that he is mistaken, referring to his most recent and sustained defense of his claim.

McMahan pursues two strategies in arguing against the view that justified attackers (or “threateners,” as he calls them) are liable to attack. His first strategy seeks to show that such a view has counter-intuitive implications. Thus, this is basically a reductio ad absurdum strategy. His second strategy, in contrast, is to try to undermine his opponents’ positive arguments for their view to give positive support for his own view.

Before considering his arguments, we have to set the scene by describing his main example, the Tactical Bombers. McMahan’s own description of this case is very long and detailed, but for the purposes of the present discussion not all of these details are relevant.² We can summarize his example as follows:

There is a bomber crew of 5 people who have, in the context of a humanitarian intervention, the mission to bomb and destroy a military target. If they are successful, 100 innocent civilians in the state in which the intervention takes place will be saved, and there are no alternative means of saving them. However, 5 innocent villagers will be killed in this attack as a side-effect.³

McMahan assumes that this attack will be proportionate and justified. Thus, the tactical bombers are justified attackers. We will grant this assumption here.

Let us now turn to McMahan’s first strategy. McMahan’s prime target is the view that “the criterion of liability to defensive attack is moral responsibility for a threat of wrongful [he means unjust, that is, rights-violating or rights-infringing] harm to another” (2014a, 114). He calls this view “the responsibility account.” However, as he himself subscribes to a responsibility account, although one that is qualified by the doctrine that “justification exempts one from liability to defensive action” (2014a, 118), it is useful to distinguish between an unqualified responsibility account and a qualified one.

For McMahan, the problem of the unqualified responsibility account lies precisely in its rejection of the doctrine in question (“justification defeats liability to defensive attack”), for this rejection allegedly has certain “implausible implications,” namely “that the tactical bombers are liable to be killed in defense of the villagers, that neutral third parties therefore have a liability justification for killing them, and that the bombers have no right of self-defense either against the villagers or against third parties (though they might have a different justification for defensive action derived from the importance of achieving their mission)” (2014a, 118).

From McMahan’s discussion and his references to my work it would appear that he thinks that I endorse the unqualified responsibility account (2014a, 115, n. 6). However, I actually reject both the unqualified and the qualified responsibility accounts, because both of them lead to counter-intuitive, namely rather draconian, implications about who is liable to attack (Steinhoff 2012a, section 4.2). I do, however, indeed subscribe to an account that implies that justified attackers are liable to attack. Yet, to simply assume that this is an “implausible” implication would obviously be question-begging in the context of McMahan’s attempt to defend his doctrine that “justification defeats liability to defensive attack”. McMahan would have to show that it is an implausible implication.

And indeed, he does try to show this. His argument is that the first implication (the bombers are liable) of the unqualified responsibility account implies the other two implications (“neutral third parties therefore [my emphasis] have a liability justification for killing them, and … the bombers have no right of self-defense either against the villagers or against third parties”).

2 Who has a Right to Self-Defense – the Justified Bombers or the Innocent Civilians?

Let us start with the last implication, namely that “the bombers have no right of self-defense either against the villagers or against third parties.” McMahan states:

Many people, of course, will not find it counterintuitive to suppose that the villagers are permitted to shoot down the bombers in self-defense. But what is counterintuitive is the claim that, while the five villagers are permitted to kill the five bombers in self-defense, the bombers are not permitted to kill the villagers in self-defense. For that to be true, it seems that there must be some significant moral asymmetry between the villagers and the bombers. Yet on the assumptions most favorable to the villagers, both groups act with moral justification in threatening to harm the other. The only difference is that the bombers have attacked first. But that is of course precisely what they were morally justified in doing. (2014a, 114)

In reply, first, a morally significant difference between the bombers and the villagers is not only that the bombers attack first, but that the bombers, by attacking first, are infringing the rights of the villagers. The villagers, in contrast, are not infringing anybody’s rights by living in the village. Thus, the villagers defend themselves only after their rights have been threatened, while the bombers bomb the villagers without the villagers having threatened their lives (or anybody else’s, for that matter). This moral asymmetry is entirely unaffected by the fact of the bombers’ justification.

Second, the claim that the bombers are liable to attack does not have the counter-intuitive implication that the bombers cannot justifiably defend themselves against the villagers. As McMahan himself acknowledges, the bombers can – and will under the circumstances – have a necessity justification for defending themselves, that is, they have a justification based on the importance of their mission (2014a, 113 and 118).

One might perhaps be tempted to argue that the fact that the bombers have a necessity justification to bomb the military target and to thereby foreseeably kill the villagers does not imply that they then also have a necessity justification for defending themselves against the villagers, that is, for killing them intentionally.⁴ After all, many philosophers claim that there is an important difference between killing a person foreseeably and killing a person intentionally. Yet, as many other authors have pointed out,⁵ this is very hard to believe. To wit, a person with a hammer who intentionally smashes a fly on an innocent person’s forehead (someone offered him money for killing the fly), foreseeing that the person’s skull will then be crushed too, and a person who intentionally crushes an innocent person’s skull (someone offered him money for doing so), foreseeing that the fly on the forehead will then be crushed too, are, in all Western jurisdictions, simply murderers. Law does not look more favorably on the first skull-crusher than on the second. Nor should it: the idea that there is a morally significant difference between the two cases seems to be entirely counter-intuitive. Things do not change if we assume that there was actually a justification for wielding the hammer. Suppose, for instance, that the fly is the carrier of the doomsday virus which would exterminate humanity, and we are faced with the last chance (there are no other means than the ones described here) to keep it from escaping the lab. In the first scenario the person’s skull is crushed as a side-effect, since the fly is sitting on it, in the second scenario, the innocent person’s skull is crushed because the fly is sitting behind it and one wants to give the anti-fly sniper a clear shot. The innocent person in the second case will certainly not think that his situation has dramatically improved if he swaps places with the innocent person in the first case, and it is unclear why there should be something morally more dubious about the first kind of attack than about the second one. If there was a difference, this would mean that there has to be a number x of people prevented from dying through the virus that would make smashing the skull justifiable in the first scenario while in the second it would remain unjustifiable. Again, this seems to be extremely hard to believe.

Moreover, and even more importantly, we are currently discussing McMahan’s attempt to burden the account defended here with “implausible implications.” However, one can only derive implications from an account on the grounds of its own assumptions, not on the grounds of assumptions the account rejects. Thus, while it is true that McMahan assumes that there is a morally significant difference between intentional and foreseen killing, I, for one, emphatically reject this assumption. Accordingly, the account defended here most certainly does not have the implausible implication that the tactical bombers may not defend themselves against the civilians. They may defend themselves on the basis of a necessity justification.

Yet, this is, according to McMahan, “not what is generally meant by a right of self-defense,” and he states that some people (including him, apparently) may be “troubled by the idea that the bombers lack a right of self-defense” (2014a, 117). He then considers the attempt to escape this allegedly troubling idea by claiming that the bombers’ liability (the forfeiture of their right to life) “does not entail the loss of their right of defense as well” and affirms that this claim would be “doubtfully coherent” (2014a, 117). I do not want to take a stance on whether the claim in question (a claim I nowhere make) is coherent or not; rather, I would like to shed some doubts on the coherence of McMahan’s own claims.

To wit, McMahan insists that the bombers do have a right of self-defense against the villagers. But how is that possible within the framework of his theory? After all, he emphasizes that on his account of liability people can only become liable to attack by forfeiting rights through their own responsible action (2009, 7–37; 2012, 296). Moreover, he explicitly defines “justified threateners” as “people who act with moral justification but whose justified action will wrong or infringe the rights of others – in this case, the villagers’ right not to be killed.” Why? Because “the villagers’ right against attack has been neither waived nor forfeited, it is overridden” (2014a, 105). In other words, the villagers have done nothing to forfeit their right to life.

Yet, McMahan overlooks the fact that the villagers have also done nothing to forfeit their right of self-defense. But if the villagers have a right of self-defense against the bombers, then it is conceptually and logically impossible that the bombers have a right of self-defense against the villagers.

This is at least the case if we are talking about so-called claim rights. As McMahan notes, “a claim right is … a right against intervention” (2009, 62).⁶ The bomber’s shooting the villagers when the villagers try to defend themselves with their anti-aircraft gun obviously is a way of interfering with the villagers’ self-defense. Thus, it is simply impossible that the bombers have a claim right to defend themselves against the villagers if the villagers have a claim right to defend themselves against the bombers. But while McMahan deems it “doubtfully coherent” that the bombers lose their right to life but retain their right to self-defense, he apparently and interestingly does not deem it “doubtfully coherent” to imply that, conversely, the villagers retain their right to life but lose their right to defend their lives.

Perhaps one might be tempted to speculate that the villagers forfeit their claim right to defend themselves by defending themselves. However, that is as absurd as claiming that someone forfeits his right to life by living. Another possibility would be to deny that people have a claim right to self-defense. Maybe they only have a liberty right: a liberty right of self-defense would merely imply that I do not wrong the aggressor if I defend myself, but it would not imply that the aggressor wrongs me by defending himself against my defense. Yet, first, this possibility does not sit well with McMahan’s insistence that aggressors have no right of self-defense against defenders (2009, 14). Second, if people only have a liberty right of self-defense, then this would also be true of the bombers. But if it is, as McMahan claims, “doubtfully coherent” to claim that someone loses his right to life (which in McMahan’s account is certainly a claim right) but keeps his right to self-defense, then it would also appear to be doubtfully coherent to claim that the bombers retain their right to life but would nevertheless not be wronged by villagers who try to kill them in order to defensively preempt the bomber’s self-defense against the villagers. Third, if the villagers indeed had no claim right to defend themselves against the justified bombers, then the justification of the latter would in effect not only defeat their own liability but also the villagers’ claim right to self-defense. But why does it then not also defeat the villagers’ right to life? Again one would like to have an argument for this curious asymmetry between the right to life and the right to self-defense, but McMahan offers none.⁷ Fourth, that people have a basic claim right to self-defense (and not just a liberty right) is an extremely plausible assumption. Without it, it would, for instance, be impossible to explain why the state would wrong us (as we certainly intuitively and quite rightly think it would) if it prohibited us from defending ourselves (see Kadish 1976, 884).

Another possibility one might speculate about is to apply so-called rights specification theory to the right of self-defense and claim that this right includes a right to defend oneself against culpable attackers, but not a right to defend oneself against justified attackers. Yet, this would be obviously entirely ad hoc and question-begging. It would also again raise the question of why the same cannot be said about the right to life: an innocent person’s right to life might not include the right not to be killed in a necessary and proportionate military attack. A further question is why one should not specify the right to life in such a way that it does not include a right not to be attacked by the innocent victims of one’s justified but rights-infringing acts. In this case, “justification defeats liability” would not help the justified bombers – McMahan’s account of liability involves, as already noted, forfeiture through responsible action, but a right one does not have in the first place cannot be forfeited, which means that the “justification defeats liability” protection against forfeiture would be useless for the bombers. They would not be “liable” to attack, but they would simply have no right not to be attacked. Either way, stipulatively “specifying” rights such that they conform to one’s philosophical predilections does not amount to an argument. While a certain rights specification could be a legitimate result of an argument about the respective rights of the villagers and the bombers, one cannot simply assume it.

Another suggestion I have come across is that McMahan could claim that the villager’s right not to be killed supports a claim right to do some things in self-defense, but not a claim right to kill in self-defense.⁸ Thus, if the civilians tried to kill the bombers, they would engage in excessive self-defense, and hence the bombers could defend themselves against this by killing the civilians. There are two problems with this suggestion. The first is that it seems objectionably ad hoc and arbitrary. To wit, if bombers who justifiably initiate an attack on entirely non-liable people are not liable to be killed (although they are liable to some force, according to the suggestion), why then should civilians be liable to be killed although they certainly have at least a partial excuse to defend themselves against others and although, moreover, these others are not entirely non-liable and, furthermore, initiated a lethal attack against them? Of course, one can simply claim that there is such a curious asymmetry here – but then this is, indeed, just a mere claim. The second problem is that this suggestion simply contradicts what McMahan says. He nowhere states that justification merely diminishes liability; rather, he says that justification defeats “liability to defensive harm” (2008a, 231), that it “exempts one from liability to defensive attack,” or that it “excludes” one from it (2014a, 118 and 119). In other words, on his account the bombers are not liable at all. But this, as already argued, makes McMahan’s position incoherent since it surreptitiously and inconsistently strips the civilians of a right that, by McMahan’s own lights, they have done nothing to forfeit.

Finally, it is indeed intuitive to suppose that the villagers are permitted to shoot down the bombers. As McMahan himself admits, he is “sympathetic to this intuition and once sought to defend it” (2014a, 108). Yet, if we grant (McMahan – counter-intuitively – does not grant this anymore) that the villagers may shoot down the bombers, then the question arises how that can be explained without supposing that the bombers are liable to attack. If we suppose that the bombers are not liable and retain their right to life, then the only possible explanation is that the villagers are permitted to override the rights of the bombers, probably on the basis of an agent-relative permission to give their own interests greater weight than the interests of the bombers. However, McMahan – rightly, in my view (Steinhoff 2015) – rejects this idea as implausible (McMahan 2014a, section 7.5). But if this idea is implausible, then the liability-based explanation – being the only one remaining – is plausible by default, as it were – unless McMahan had independent arguments to reject either the idea that the villagers are permitted to shoot down the bombers or independent arguments to support the idea that justification indeed does defeat liability. Yet, McMahan does not have independent arguments against the first idea, and his arguments for the second idea are unconvincing, as I will argue in a moment in sections 4 and 5.

Moreover, McMahan’s new position, namely that the villagers are not justified in defending themselves or permitted to defend themselves (2014a, 108 and 136), implies, in conjunction with his claim that “the criterion of liability to attack in war is moral responsibility for an objectively unjustified threat of harm” (2009, 35), that the villagers become liable to attack if they defend themselves. In fact, since the threat they pose is lethal, they would become liable to be killed (if otherwise the threat cannot be averted); and according to McMahan, it is permissible to kill any number of people who are liable to be killed (2014b, 137). All this taken together leads to certain unpalatable conclusions, however.

Consider this variation of McMahan’s tactical bombers. The general wants to bomb an ammunitions factory. He knows that thereby he will save 100 innocent lives, and kill 5 innocent bystanders. Yet, he finds out that 500 other villagers, when they see the bombers approaching, will try to defend the 5. Each of the 500 has an anti-aircraft gun, and it would be necessary to kill each of them to reach the target. (Thus, it is not the case that the number of the villagers decreases each villager’s contribution to the lethal threat that the bombers face: each of the villagers is a lethal threat in his or her own right.) On McMahan’s account, all these 500 (but not the 5) villagers would be liable to attack and could be killed in the course of achieving the mission. Maybe I am being squeamish and overestimate the worth of human life, but to me this seems to be a morally entirely unacceptable consequence of his account.

But, one might object, do the villagers, by taking part in the hostilities, not become combatants themselves, and thereby legitimate targets under international law? If one takes into account a feature of McMahan’s original example that I have only mentioned in a footnote, namely that the villagers are citizens of a neutral country (their village is across the border in that country), this would be wrong, since the threat they are facing is not lawful,⁹ and self-defense against unlawful attack does not qualify as taking part in the hostilities (Schmitt 2010, 34). If we are talking about villagers in the country where the actual war is taking place, however, the legal situation might well be different. Yet, we are concerned with the moral question. And while the 500 villagers might in one sense take part in the hostilities, it should be noted that there is a difference between strict self- or other-defense against an ongoing or imminent attack on the one hand and on the other actively, and again and again, seeking out the enemy to kill him (search and destroy) (see on this distinction Cochran 1996). The villagers engage in the former, not in the latter. It seems, therefore (at least to me), that the implication of McMahan’s account that all 500 villagers would become liable to attack and that hence the bombers’ mission would remain proportionate, although they will now kill five times more people than they save – people, moreover, who did not initially pose a threat but only reacted to an unjust threat posed to their fellow villagers by the bombers – is not an implication that speaks in favor of McMahan’s theory.

Thus, our discussion so far has shown that the implication of the unqualified responsibility account (and of other accounts that hold justified attackers liable to attack) that the tactical bombers of McMahan’s example have no right to self-defense is not implausible at all. Rather, it is McMahan’s contrary claim that is implausible. In fact, it seems to contradict some of McMahan’s very own basic premises about rights-forfeiture and about what justified bombers actually do: attack people who have not forfeited their rights.

3 Do Neutral Third Parties have a “Liability-Justification” to Kill the Justified Bombers?

Let us now turn to the implication that “neutral third parties have a liability-justification” for killing the tactical bombers (2014a, 118, see also 114–117). Is this really implausible? That is not so clear, but be that as it may: as far as I am concerned, I happen to reject the idea that liability alone can justify anything (Steinhoff 2016). On my definition of liability a person is liable to be killed if she has no right not to be killed. However, how can a person’s mere lack of a right not to be killed provide by itself a justification to kill her? I do not think that a 500-year-old oak tree holds a right against me not to be cut down, but I fail to see how that could possibly justify me in cutting it down (compare also Uniacke 1996, 191). That same logic applies to persons.¹⁰

Thus, contrary to McMahan’s (2014a, 115) explicit attribution of this claim to me, I have most certainly not claimed that “there is an agent-neutral liability justification for killing the bombers.”¹¹ First, I have not made the claim that the justified bombers become liable to attack by just anyone: I made one exception (2012a, 357–358). This need not concern us here, though. More important is that I have explicitly distinguished between liability on the one hand and justification or even permissibility on the other (2012a, section 5), which means that agent-neutral liability does not imply agent-neutral justification. For the “unjust combatants” to be justified in shooting down the justified bombers it is simply not sufficient that the latter are liable to attack; rather, the fact, for instance, that the unjust combatants are defending people they have special relations to, like family members, friends, or their compatriots more generally, can provide them (sometimes) with a justification.¹² As McMahan once admitted himself: “If the civilians have a right of self-defense, it’s hard to believe that those they’ve paid to protect them aren’t permitted to help them do what they’re permitted to do.”¹³ On the other hand, completely neutral third parties simply do not have such a justification. Thus, I submit that the account I endorse does not produce counter-intuitive results, but in fact precisely the intuitive ones.¹⁴ Likewise, while, in my view, the unqualified responsibility account does in fact have certain counter-intuitive implications (which it shares with McMahan’s qualified account), it does not have the implausible implications McMahan ascribes to it.

Moreover, at one point McMahan actually acknowledges that there is this possibility of sufficiently disconnecting liability from justification: “One might agree that the unjust combatants are not permitted to kill the bombers but also claim that this is not because the bombers are not liable to be killed.” (2014a, 116) He then presents an example that is somehow (I admit: I do not know how) supposed to undermine this strategy. He envisions a situation where the “unjust combatants” change their mind, so that the civilians who the justified bombers want to save from the allegedly unjust ones by engaging in the attack that will kill the (smaller number of) other innocent people (the villagers) “are no longer threatened by the unjust combatants” (2014a, 116). He then asks us to consider whether the reformed, formerly unjust combatants, in order to do what is right, should attack the justified bombers to save the villagers or instead allow the bombers to kill the villagers and save the other civilians themselves. This thought experiment, however, is incoherent: if the civilians are no longer threatened (either because the formerly unjust combatants refrain from killing them or, if necessary, will protect them themselves against their non-reformed former comrades), then the attack the allegedly justified bombers are about to engage in is not justified but in fact entirely unnecessary – and therefore there are good reasons to save the innocent villagers from the unjustified, liable bombers.¹⁵

I conclude that McMahan’s first strategy of defending the “justification defeats liability to defensive attack doctrine,” namely by trying to reduce the negation of this doctrine to absurdity, is unsuccessful.

4 What Really Matters: On Permissibly Resisting Justified Bombers and Impermissibly Resisting (Certain) Unjustified and Liable Thieves

McMahan’s second strategy is to try to undermine arguments against the view that justification defeats liability (to defensive attack), which in turn is supposed to positively support his own view that justification does defeat liability to defensive attack. McMahan’s central example in the context of this strategy is the following:

... a passerby finds a person in a diabetic coma. If the diabetic does not receive a shot of insulin within minutes, he will die. The passerby knows that the house immediately across the street from where the diabetic lies belongs to someone who has a bountiful supply of insulin. That person not being home, the passerby breaks in, takes some insulin, and saves the life of the diabetic. (2014a, 119)

While in the past McMahan made the more general claim that justification defeats liability (not only liability to defensive attack) and asserted that this was also true in law (2008a, 233–234),¹⁶ he now admits that tort law would require the passerby – although he has acted justifiably – to compensate the owner of the insulin. And he adds that many people think that this also reflects the demands of morality (2014a, 118).

In “response to this challenge” McMahan argues.

that while justification does not exclude liability to compensate those one has harmed, it does exclude liability to defensive harm. One reason why this might be true is that to hold a justified threatener liable to compensate his victim is not to permit anyone to prevent his justified action, whereas to hold him liable to defensive action is to permit others to prevent the justified action. (2014a, 119)

Yet, this is simply wrong. To hold a justified actor liable to defensive force definitely does not amount to allowing interference with his justified act. McMahan should know this, since he himself provides an example of a case where it would be wrong to prevent even the unjustified action of a liable agent (2014a, 107, the police sniper example).¹⁷Consider this easier case: Karl is about to slap me out of spite. I know that if I defend myself against this attack, Karl’s father, the dictator, will kill 100 innocent people. Karl is liable to defensive measures by me – he cannot complain if I defend myself – but it seems that I ought not to defend myself all things considered. My defense against liable Karl would be impermissible given what is at stake.

In the same vein, it is not the passerby’s justification that rules out interference, but what is at stake, namely an innocent life. Consider a variation of McMahan’s insulin example. The passerby, out of spite and in order to harm the owner of the insulin, intentionally takes the much more expensive one of two insulin charges (it is so expensive because it helps not only against the common diabetes of the unconscious diabetic, but also against two incredibly rare kinds of diabetes), although the other one would be as effective. Since necessity justifications are subject to constraints of proportionality and, well, necessity, this act of taking the more expensive insulin would be unjustified. But suppose interfering with this act would not leave enough time to then provide the diabetic with the other charge. Is the owner of the insulin justified in stopping the unjustified, liable passerby? It does not seem so. But this means that liability and justification has not much to do with the reason why the passerby must not be stopped.

Of course, McMahan affirms that “the killing of the bombers is not ruled out solely because it would prevent the saving of the 100 civilians” (2014a, 114). I agree, of course, but I agree because the killing of the bombers is not ruled out at all. In the modified insulin case, in contrast, stopping (let alone killing) is ruled out because it would prevent the saving of the diabetic, a saving which is possible without imposing excessive costs (like severe bodily injury or even death) on the owner of the insulin. Justification and non-liability have nothing to do with it: it is simply a question of weighing moral benefits and moral costs against each other.

It is also interesting that McMahan sees the following difference between the case of the diabetic and the case of the tactical bombers:

Whereas the owner would have a duty to provide the needed insulin were he at home, I have stipulated that the villagers (who are the cost-bearers in this case, as the owner is in the diabetic case) would not be required to act in a way that would sacrifice their own lives to save the 100 civilians. (2014a, 112)

Yet, sacrificing their lives is exactly what McMahan requires from the civilians. He does not require them to commit suicide, yes, but of course one can sacrifice one’s life without committing suicide. If, for instance, someone asked me not to shoot the tiger that is attacking me because if I did, the tiger could not go on to then also kill the villain who otherwise would kill 20 innocent people, then, yes indeed, it seems that what I am being asked is to sacrifice my life for the benefit of these other people (Steinhoff 2014).

Be that as it may, McMahan then considers a variation of the insulin case where the costs to the owner are so high that he is not under a duty to surrender the insulin. Surrendering it would be supererogatory. And McMahan claims:

It also seems plausible to suppose that, even if the owner had no duty to provide the life-saving resource, the passerby’s justification in taking it would shield her from liability to harmful defensive action by the owner. The owner would, it seems, be permitted to thwart the passerby’s efforts by other means, but not by means that would involve the infliction of serious or substantial harm on the passerby. (2014a, 122)

Note the swift shift from merely “harmful defensive action” to “the infliction of serious or substantial harm.” It might well be plausible that the passerby is not liable to be killed or mutilated under these new circumstances, where the costs to the owner are significant. It is plausible because killing or mutilating him would be disproportionate. In contrast, that the passerby would not be liable to any form of harmful defensive measures (pushing, holding, or a punch to the solar plexus) although he would infringe the owner’s rights seems to be entirely implausible.

This brings us to the actually decisive difference between the case of the bombers and the insulin case. This decisive difference is also an obvious one: the passerby is only taking some insulin; the bombers, however, are about to kill or mutilate the villagers. Given, therefore, that the costs imposed on the villagers and the costs imposed on the owner of the insulin differ not only quantitatively but qualitatively – there mere property rights are infringed, here, however, rights to bodily integrity and life¹⁸ – it is puzzling that McMahan takes the impermissibility (in the original case) of preventing the passerby from taking the insulin as justifying a statement like this:

So even if the general claim that justification excludes all forms of liability is false, it may still be true that justification excludes liability to defensive harm, and that is all that is necessary to rule out the claim that the bombers are liable to defensive action either by the villagers or by third parties. (2014a, 119)

It “may” be true that justification excludes liability to defensive harm, but that cannot be established by the insulin example. That example might (or might not) show that sometimes justification defeats liability to defensive harm, but it does not show that it always does. In other words, while for an argument in support of the liability of the bombers it is quite sufficient to demonstrate that a justification to kill and maim innocent people does not exclude liability to defensive harm,¹⁹ it is certainly not sufficient for McMahan’s argument in support of the non-liability of the bombers to show that a non-killing, non-maiming thief’s justification to steal insulin defeats liability to defensive harm.

Not to ignore the qualitative differences is of the utmost importance. As far as a justified attacker’s liability to defensive harm is concerned, there seems to be no sliding scale, as it were, between minor harms and harms to life and bodily integrity. Rather, there seems to be a moral threshold which is based on the “separateness of persons,” to use Rawls’s term (1999, 167). German law, for example, holds that innocent people cannot reasonably be expected to sacrifice their own lives or bodily integrity for the survival of strangers, and therefore denies a necessity defense when it comes to killing or maiming innocent people (Erb 2003, 1350–1351 and 1387). However, while I think that the first part is absolutely correct, the conclusion does not follow. If the stakes are high enough, killing innocent people can be justified, but given that they cannot reasonably be expected to sacrifice themselves for strangers, they cannot reasonably be expected not to fight back (Steinhoff 2014). This, incidentally, seems to be precisely the position taken by those US states that have adopted the Model Penal Code’s account of the necessity justification (Steinhoff 2012a, 360–362). It is, I submit, a quite reasonable position, which can certainly not be undermined by McMahan’s insulin example.

5 Who “Should” Compensate? – and why that does not Matter

McMahan contemplates a second response to the challenge posed by the intuition that the passerby must compensate the owner of the insulin, namely to deny that this intuition is correct and to insist that the “stronger claim that moral justification excludes all forms of moral liability is true” (2014a, 119).

Why, however, should that claim be true (it is most certainly not a very intuitive claim)? One of McMahan’s explanations is:

The sacrifice of the insulin may simply have been what morality required of [the owner]. But given that he was not at home at the time the insulin was needed, the passerby acted in his absence to fulfill the duty he would have had if he had been at home. Thus no one is liable to compensate the owner for the loss of what he was morally required to sacrifice. (2014a, 121)

Yet, this “Thus” is a clear non-sequitur. Suppose I go into a bakery and say: “One cheesecake, please.” The baker answers: “Of course.” He has thereby entered into a contract with me and is now morally required to indeed give me the cheesecake. But that he has a Hohfeldian duty towards me to give me the cheesecake does not absolve me from my Hohfeldian duty towards him to pay him for it. Maybe, however, contractual duties are different.²⁰ So let us look at a duty that is not contractual: the duty of gratitude or reciprocation. To wit, if a neighbor loans me flour when I ask him for it, then this seems to ground a duty on my part to lend him flour (all else being equal) if he asks for it. But that hardly implies that he can just walk into my apartment in my absence, take the flour, and be absolved of any duty of compensation.

Likewise, that the owner has a duty to let the passerby take the insulin does not imply that the passerby has no duty to compensate the owner. (In others words, that the owner has to “sacrifice” the insulin does not mean that he also has to sacrifice its financial value.) As regards the claim that the owner would not have had a claim to compensation if he had provided the insulin himself – this is neither here nor there. The observation that justified rights infringements give rise to duties to compensate cannot be refuted by pointing to cases that do not involve rights infringements in the first place (the owner does not infringe his own rights by providing the insulin).

McMahan makes a number of further claims that are, in my view, simply irrelevant for the issue in question. For example, he states that “the burden of the rescue should ideally be shared by everyone in the society (or indeed by everyone in the world)” (2014a, 120).²¹ But just as McMahan confused above an agent’s liability to interference with a second agent’s permission to interfere, he now confuses what people should do to some other people (for example, share the burdens with them) with what those other people are liable to. In that context he also considers another option (potentially more relevant to the non-ideal world since “there is no such [burden-sharing] scheme in place”) (2014a, 120), namely to impose the costs of the rescue on its beneficiary: “The diabetic can … have no reasonable objection to being required to restore the owner to the position he would have been in had he not made (or been forced to make) the sacrifice, when the outcome of this would still be much better for the diabetic than what would have happened to him in the absence of the sacrifice.” (2014a, 120) Setting aside the question whether this is true or not, it is, again, irrelevant: that the diabetic cannot have a reasonable objection (which, incidentally, is not so clear) if he is required to compensate, does not imply that the passerby can have a reasonable objection if the owner first asks him to compensate. Indeed, it would appear that the passerby, whose act directly infringed the rights of the owner, owes the owner compensation; while the diabetic, whom the passerby directly benefited, perhaps owes compensation to the passerby.

McMahan further considers what should happen if the beneficiary does not have the means to compensate and states: “Indeed, it seems implausible to me, regardless of what the law says, to suppose that third parties would be morally permitted to coerce the passerby to provide that compensation, given that she has acted with full moral justification.” (2014a, 121) Yet, provided the passerby has enough money, this seems entirely plausible to me, regardless of what McMahan says. Judith Jarvis Thomson notes: “[I]f A wants to do a certain good deed, and can pay what doing it would cost, then—other things being equal—A may do that good deed only if A pays the cost himself.” (Thomson 2008, 365) While I think that she is overstating her case here, it seems that the owner has not only a complaint against the passerby for infringing on his property rights in the first place, but also a further complaint if the passerby wants to impose the costs of his good deed on the owner.²² Besides, McMahan’s point is irrelevant again: even if it were true that third parties would not be morally permitted to coerce the passerby to provide compensation, this does not imply that the passerby is not under a duty towards the owner to provide compensation (after all, there is the famous “right to do wrong” – others are not necessarily permitted to force one to do what one has an obligation or a duty to do [Waldron 1981]).

In that context, let me come back to the proviso: what if the passerby does not have enough money, what if he is really poor, and taking his money for the insulin would impose an unreasonable hardship on him? In that case, I submit, perhaps neither the state nor the owner should insist that the passerby compensate the owner. But, again, this does not mean that he has no duty to compensate, that he is not liable to pay compensation. Out of beneficence or compassion the owner should not insist on his right (German law calls inappropriate insistence on one’s rights abuse of rights, but for there to be an abuse of a right, there obviously has to be that right in the first place), but, conversely, the passerby should at least apologize to the owner: “Look, I am really sorry, but I can’t pay for the damages I unjustly inflicted on you.” This confirms that he is not released of his duty towards the owner (unless the owner releases him himself), even if, all things considered, one should not make him pay out of compassion and mercy.

Thus, McMahan’s discussion of compensation in no way warrants his conclusion that “the justified agent is not liable to compensate the victim,” let alone his further conclusion that “[t]his, in turn, supports the more general claim that justification excludes liability” (2014a, 122). Nor does it warrant his claim that “the case of the insulin shows that the intuitions about compensation to which Steinhoff appeals are not so robust as he supposes” (2014a, 123). First, while showing that justification does not defeat liability to pay compensation shows that justification does not always defeat liability, conversely, showing that justification defeats liability to pay compensation would still not show that it does always defeat liability – including a justified killer’s or mutilator’s liability to defensive attack by his innocent victims. Second, McMahan’s discussion shows nothing about the robustness of our intuitions, not least since his discussion of how we should distribute burdens or what we are or are not permitted to do misses the mark – it has little to do with the question of who is liable to what.

6 Conclusion

McMahan argues that justification defeats liability to defensive attack. In response, I argued, first, that McMahan’s attempt to burden the contrary claim with counter-intuitive implications fails; second, that McMahan’s own position implies that the innocent civilians do not have a right of self-defense against justified attackers, which neither coheres with his description of the case (the justified bombers infringe the rights of the civilians) nor with his views about rights forfeiture, is unsupported by independent argument, and, in any case, is extremely implausible and counter-intuitive; and third, that his interpretation of the insulin case confuses the normative relations between an agent’s justification and non-liability (or lack thereof) on the one hand and permissible or impermissible interference with the agent’s act on the other. Similar confusions, fourth, affect his discussion of liability to compensation. I conclude that McMahan has failed to demonstrate that justification defeats liability to defensive attack.