This article responds to objections to the account of permissible harming developed in Defensive Killing, as raised by Christian Barry, Jeff McMahan, Kimberly Ferzan, Massimo Renzo and Adil Ahmad Haque. Each paper deserves much more attention than I can give it here. I focus on Barry’s important observations regarding the liability to defensive harm of those who fail to rescue. In response to McMahan, I grant some of McMahan’s objections to my rejection of the moral equivalence of threats and bystanders, but reject his analysis of my Shield cases. I welcome much of Ferzan’s development of my account of ‘futile’ defence, but offer some concerns regarding her own view of when honour can be appropriately defended. I argue that Renzo’s objections to my account of bloodless invasions are unpersuasive, and identify some problems with Renzo’s own view. Finally, I defend my account of civilian liability against Adil Haque’s critique.
1 Defensive Killing
Defensive Killing defends an account of permissible defensive harming. Part One explores when individuals may inflict harm in defense of their own lives. It primarily focuses on the conditions under which a person becomes morally liable to defensive harm – that is, when she forfeits her right against such harm, such that harming her does not wrong her. I pay particular attention to the liability of those who indirectly threaten, for example by blocking an escape route of which a threatened person needs to avail herself, or by contributing to a threat directly posed by someone else, such as by hiring a hitman. The account of liability I defend shares its roots with Jeff McMahan’s view that one is liable to defensive harm only if one is morally responsible for a threat of unjust harm, although McMahan and I disagree on various theoretical and substantive points concerning liability and permissible defense.
Part Two defends the claim that the principles that govern harming in war are the same as those that govern harming in ‘ordinary life’ (the so-called ‘reductivist’ approach to the ethics of war). I then consider the implications of my account of defensive harming for specific aspects of harming in war. I argue, amongst other things, that if liability to defensive harm is underpinned by moral responsibility for an unjust threat, many civilians who contribute to unjust wars will be liable to defensive harm. Despite its counter-intuitiveness, I argue that this result is correct. We may have other moral reasons to refrain from harming even liable civilians (such as the likelihood of causing disproportionate collateral harm). But many civilians would not be wronged by being killed by just combatants, if such killings were a means of saving the lives of innocent people whom those civilians have helped to endanger.
I’m very grateful to the contributors to this symposium for their careful and insightful engagement with Defensive Killing. Each paper deserves much more consideration than I can offer here. Since each discusses a different part of the book, I’ll consider each paper in turn.
2 Enablers, Omitters, and Liability
Case 2: Ella removes a rock that would otherwise stop a trolley that is hurtling downhill toward Vicky.
Boulders: Ella sees Vicky walking along the beach. Ella is on the rocky cliffs overhead. A large boulder is held securely in place by a smaller boulder. Ella manages to push the smaller boulder aside. The large boulder rolls off the cliff, crushing Vicky to death.
Ella clearly kills Vicky in Boulders – indeed, it’s hard to see any significant difference between Boulders and a case in which Ella simply pushes the large boulder off the cliff onto Vicky. And there doesn’t seem to be any significant difference between clearing the path for the boulder in Boulders, and clearing the path for the trolley in Case 2. In both cases, Ella initiates a threat to Vicky: as things stand before Ella interferes, Vicky is not threatened by either the trolley or the large boulder. (Note, then, that neither is a case of preventing Vicky from being saved. She doesn’t need saving until Ella acts.)
Contra Barry, I also think that if Ella intentionally removes the rock in order that the trolley strikes Vicky, she uses the trolley as a tool to kill Vicky. Her action seems akin to intentionally leaving a person lying on a trolley track because their body will prevent the trolley’s hitting five. Leaving the person there makes use of her, even though one does not manipulate her (one can, I think, make use of a person by allowing her to be killed). Similarly, Ella does not manipulate the trolley by removing the rock, but she does manipulate the environment to ensure that the trolley hits Vicky. Thus, she uses the trolley as a tool.
Of course, according to the view defended in dk, this suggests that Vicky may kill Ella in self-defense not only if Ella removes the rock, or intentionally rolls away the small boulder, in order to unjustly kill Vicky, but also if she does so innocently, unaware that she thereby threatens Vicky. As I’ll discuss in Section Three, my view of killing innocent threats has shifted somewhat since I wrote the book (or, more accurately, my intuition that such killings are permissible persists – and extends to killing innocent Ella – but I am now less persuaded by the arguments I gave in dk).
The second gap that Barry identifies concerns people who fail to save. As Barry points out, it’s very plausible that those who responsibly fail to rescue (such as Abel, who fails to costlessly trigger a fence that will protect Vicky from the trolley) are liable to harm that seeks to avert the relevant threat. Abel may not only be forced to rescue Vicky, but also to do so at a greater cost than he was initially required to bear to save Vicky.1 I’ll call people who (knowingly) fail to save omitters.
Barry suggests that the plausibility of the view that omitters are liable to preventive harm forces us to conclude that either bystanders are not always materially innocent with respect to threats, or that material innocence does not preclude liability. A counterfactual analysis of causation, for example, will deem Abel a cause of Vicky’s death, making him a straightforward candidate for liability to defensive harm.2 However, even if we think omissions are not causes, Abel nevertheless omits to do something that would have prevented Vicky’s death. Barry suggests that this ‘quasi-causal’ connection plausibly suffices to ground liability.
2.2 Liability for Failing to Save
One question, then, is whether omitters might be liable to defensive killing in order to prevent proportionate harm to those whom they fail to save. Barry suggests that only culpable omitters are liable to defensive harm, and only highly culpable omitters – those who know all the relevant facts, can prevent very serious harm at little cost to themselves and, perhaps, who want the victim(s) dead – are liable to be killed, and perhaps then only to prevent many deaths.
This different treatment of omitters compared to other threateners is justified by the fact that omitters do not “make people worse off than were they to be absent from the scene altogether.” I agree with Barry that our duties to rescue are less stringent than our duties not to threaten. One way to understand this claim is as asserting a lower cost threshold for when one may fail to save compared to when one may cause harm. For example, Abel may fail to rescue Vicky if saving her will cost him his hand, but Dirk may not push the trolley towards Vicky to avoid the loss of his hand.
But if the fact that omitters do not make people worse off explains these different justificatory thresholds, it’s not clear what then underpins the separate claim that only highly culpable unjustified omitters are liable to be killed, and perhaps then only in defense of many lives. For Barry, it is not enough that Abel’s failure is unjustified – that is, that Abel knows that he could save Vicky at a trivial cost. Rather, he is liable to be killed only if he also “welcomes Vicky’s death and would be disposed to bring it about himself.” Barry also suggests that even if we doubt that even very culpable bystanders are liable to be killed to save one person, they must be so liable to prevent 20 deaths. But if a threatener like Dirk is liable to be killed if he is morally responsible – but not culpable – for an unjustified threat to a single person’s life, it’s not clear why an omitter is not so liable for an unjustified failure to rescue a single person’s life. Iterating that the omitter doesn’t make the victim worse off seems to double-count this fact, which has already been taken into account in setting the cost threshold at which the failure becomes unjustified. Moreover, even though Abel does not make Vicky worse off than she would have been in his absence, he does make her worse off than she has a claim against being, since he unjustifiably fails to rescue her. We might think that this is the appropriate baseline, rather than the counterfactual baseline Barry employs. So, whilst the liability of omitters certainly demands more thought, I don’t think the view that Barry proposes here can be correct.
3 Killing Innocent Threats
Jeff McMahan’s paper argues against my long-held view that it’s permissible for Victim to kill in self-defense even an innocent person who will otherwise kill him. I think McMahan is right that reframing Michael Otsuka’s Moral Equivalence Thesis so that it appeals to the impermissibility of killing innocent indirect threateners – rather than of using bystanders as human shields – avoids the objections that I leveled against that thesis.
However, Otsuka also claims that Victim may not continue to hold a flagpole, upon which Falling Person will be lethally impaled, because this too constitutes killing an innocent threat. I argued that, if so, it is hard to avoid the conclusion that Victim ought also to drop a shield he is holding and so on (letting Falling Person kill him). I argued that this seems implausible. In response, McMahan argues that we can distinguish Ray Gun from my Shield cases, because the Shield cases, along with Flagpole, are instances of (intentionally) allowing Falling Person to die, rather than of killing her.
McMahan might be right that Otsuka was simply mistaken to bite the flagpole bullet: if holding the flagpole is to (intentionally) allow Falling Person to die, Otsuka can permit Victim to hold the flagpole whilst denying that he may vaporize Falling Person. But presumably McMahan would not allow Victim to pick the flagpole up foreseeing that Falling Person will be impaled upon it – if he does that, McMahan can hardly prohibit using the ray gun. But in the section McMahan quotes, I ask whether it is plausible, “to grant Victim a permission to move the shield over himself [that is, pick up a shield] but deny him a permission to continue to hold the flagpole.” I continue: “I don’t see how it can be plausible, because picking up a lethal object—which is what the shield is—looks more like a killing than does merely continuing to hold the flagpole.” McMahan dismisses my objection by saying that “the four additional cases” that I discuss (Spacious Well, Stationary Shield, Existing Shield, and Flagpole) are instances of allowing Falling Person to die. But what I describe in this passage is a fifth case – Victim’s picking up, rather than continuing to hold, a shield, the hitting of which will kill Falling Person. As I say, this looks more like a killing than Otsuka’s case of continuing to hold the flagpole. It strikes me as implausible to grant (as McMahan does) that Victim may move himself under a shield, and continue to hold a shield, but deny that he may pick up a shield. But I doubt that in picking up the shield, the hitting of which will kill Falling Person, Victim is plausibly described as allowing Falling Person to die. Thus, there still seems to be a pretty big bullet for opponents of a permission to kill innocent threats to bite.
That said, I do have doubts about the arguments in favor of a permission to kill innocent threats that I advanced in the book. Like Gerald Lang, I’m increasingly drawn to the idea that the root of permissible defense may lie in the non-liability of the victim, rather than the liability of the attacker.3 The (rather simple) idea is that my right not to be harmed permits me to use proportionate defense against those who will otherwise harm me. As Lang argues, McMahan focuses on the fact that Falling Person will not kill Victim impermissibly (since she fails in no duty by killing him). But nor does she kill permissibly: she lacks a right to kill Victim.
Whilst I think this approach is promising, and has various advantages over its rivals, it needs further development. In particular, it’s not clear how this approach can avoid the implication that McMahan identifies at the end of his paper – namely that Victim may kill Falling Person after she has landed on him if doing so is necessary to save his life. It’s hard to see why this difference in timing affects what may be done to Falling Person – she is still what threatens Victim’s life, and she will have killed him if he is not saved – and yet it seems impermissible to kill her after she has landed on him. Again, and unsatisfactorily, I’m not sure what we should say about such cases, but I’m also not ready to give up on killing innocent threats.
4 Honor and ‘Futile’ Defense
4.1 Expressive not Defensive
Defensive Killing considers what we might say about cases of seemingly futile self-defense. For example, I argued that Fran is plausibly permitted to break Eric’s wrist in Rape even if she knows that it will not even mitigate Eric’s attack. It is hard to believe such force would wrong Eric. And yet, since Fran knows it cannot avert the rape, it’s also hard to explain how Eric could be liable to such force. I suggested that if punitive reasons underpinned a permission to harm Eric, they also permit Fran to harm him a week later, which gives us reason to be skeptical that punitive reasons are the right explanation of Fran’s seeming permission. I proposed that a plausible alternative explanation of ‘futile’ defense holds that force can serve to defend a victim’s honor (although I now think that ‘moral standing’ is a better way to capture the value at stake).
Kimberly Ferzan points out that whereas I frame my proposal as permitting the victim to defend her honor, my exploration of this idea implicitly draws upon both defensive and expressive goals. And, Ferzan argues, it’s the expressive value of such force that is the most promising way to explain its permissibility. This model both reflects a “deeply entrenched” idea of the value of resistance, and also solves some problems that arise with the defensive model.4
I am sympathetic to much of what Ferzan says in her criticism and development of my account. I’ll return the favor by criticizing her analysis of Lucky Escape, in which Murderer is culpably trying to kill Victim, and Victim can save his life by either easily parachuting to safety, or killing Murderer. My view holds that Murderer is liable to be killed by Victim, since he is morally responsible for an unjust threat to Victim’s life, and killing him is a proportionate means of averting that threat. I suggested that it’s nonetheless all-things-considered impermissible for Victim to kill Murderer – there is a general moral prohibition on causing gratuitous harm – but Murderer would not be wronged by being killed, and may not kill Victim in counter-defense. I argued that alternative accounts of liability, which hold that Murderer is liable to only necessary harm, and thus wronged by any defensive harm that Victim inflicts, struggle to explain why Murderer lacks a permission to kill Victim if Victim employs defensive force instead of jumping.
However, I also noted that if, as Ferzan and I agree, defense (or expression) of honor warrants inflicting some harm on Eric in Rape, it should also warrant inflicting some harm on Murderer. Ferzan is skeptical of this move. She argues that the value at stake in Lucky Escape is that of standing one’s ground, which manifests honor only as a “testosterone-filled, chest-beating contorted conception of masculinity”, quite different from the sort of honor at stake in Rape, and with which our enquiry is properly concerned. The question in Lucky Escape is, Ferzan claims, whether Victim has a duty to retreat – “Victim is clearly defending his rightful space, and this is an interest that is properly taken into account in what Victim may defend in Lucky Escape.” But “the expressive value in ‘not giving up’ in the face of defeat has no purchase in Lucky Escape… the value of expression in the face of certain loss simply does not obtain when Victim will lose nothing (or what little he will lose is better explained by a different conception).”5
I disagree with Ferzan’s characterization of what’s at stake in Lucky Escape. If certain threats make it appropriate to express one’s moral standing, as Ferzan and I agree is the case in Rape, this suggests that what matters for the appropriateness of honor-expressing harm is what Murderer is trying to do to Victim, not what Victim needs to do to save his life. As Ferzan notes, my account is underpinned by an objection to the disrespectful attitude that Eric and Murderer display towards their victims. Ferzan’s analysis implies that since Victim can save his life by jumping, all that is at stake in Lucky Escape is Victim’s right not to move from a place where he is entitled to be. Her approach thus fails to distinguish a case in which Victim must step aside to avoid colliding with a fellow pedestrian who refuses to alter his path from someone who is trying to murder Victim. But the wrong Murderer intends is not that of occupying Victim’s rightful space, nor of forcing Victim to move from his rightful space: rather, it is that of unjustly killing Victim because he hates him. And objecting to the fact that Murderer is trying to unjustly kill him does not seem to be a particularly masculine, testosterone-fueled trait, utterly “analytically distinct” from the objection Fran has to being attacked by Eric. By making Lucky Escape a case about retreat, Ferzan misses the point about threats to honor – namely that if expressing honor is appropriate in the face of sexual assault, it is plausibly appropriate in the face of attempted murder, and this affects the harm to which Murderer is liable.
Given this, it’s not clear that assertions of moral standing make sense, or are permissible, only in the face of the certain defeat of some other interest, as Ferzan suggests. Her account of expressing honor draws on the value of ‘fighting to the end’ when one knows that one will lose; Victim may not use expressive harms in Lucky Escape because (unlike Fran) he does not face a “certain loss.” But why must the loss be certain, rather than threatened, in order for assertions of moral standing to be appropriate? And does it matter if the loss arises only because the victim does not avail herself of permissible means of defense? Imagine that Fran could, in fact, avert the rape by breaking both of Eric’s legs, but doesn’t want to do this. However, she does want to break his wrist as a way of communicating her moral standing. Does she still satisfy Ferzan’s condition of expressing her honor in the face of certain defeat? Hers is no longer a struggle plausibly characterized as “fighting to the end,” putting up “a tenacious fight,” or “not giving up.” But this doesn’t seem to alter the permissibility of breaking Eric’s wrist in order to express her moral standing. Of course, Ferzan correctly stresses that there is no requirement on victims to resist – my point here is rather that Ferzan’s characterization of what is expressed by breaking the wrist isn’t obviously applicable to this sort of case, and yet expressions of honor still seem permissible and appropriate even if Fran also has the means of averting the rape. And if that’s true, there’s reason to think that Victim can appropriately express his honor by harming Murderer in Lucky Escape.
5 Bloodless Invasions
5.1 Two Objections
Massimo Renzo’s paper critically examines my arguments concerning bloodless invasions. Some writers, most notably David Rodin, have argued that reductive individualist views cannot accommodate a right of national-defense against these invasions, since individual political rights do not warrant lethal defense.6 I may not kill even a fully culpable person who will otherwise prevent me from voting, even though this person violates my political rights.
I argued that reductive individualism allows that one may do more to prevent multiple rights violations than to prevent single rights violations. And, I argued, some interests that might not individually warrant lethal force against even a culpable aggressor can, once aggregated, justify the infliction of lethal defense. For example, even though I may not kill you to prevent your breaking my arm, I may kill you to prevent your breaking, say, 20 people’s arms (or my arm 20 times). Thus, contra Rodin, the fact that one may not kill to protect a single person’s political rights does not preclude reductivists’ permitting lethal defense of many people’s political rights, as are threatened by bloodless invasions.
Renzo objects that the aggregative strategy misleadingly aggregates the goods at stake without also aggregating the harms to the soldiers who will be killed in a defensive war. Whilst it might be true that aggregated political rights warrant lethal force, it doesn’t follow that they warrant force on the scale of war.
But this part of my argument was not intended to support the proportionality of defensive war. It was intended to rebut Rodin’s inference from the impermissibility of lethally defending an individual’s right to the impermissibility of lethally defending many people’s rights. Moreover, I explicitly acknowledge that the defense of even many people’s aggregated political rights, in the absence of threats to other goods, might not sanction force on a scale that would count as war.7 Since this claim relies on aggregating the interests of those who will be killed, I don’t think my discussion of aggregation is misleading in the way Renzo suggests.
Renzo’s second objection – that the aggregative strategy makes the possession of a right of national-defense sensitive to the size of a country’s population – doesn’t strike me as an objection at all. On the contrary, it seems to me very plausible that the number of people whose rights are at stake determines the permissible degree of defensive force. Renzo notes that “both ordinary morality and international law attribute the right to wage wars of national defence to political communities, regardless of their population size.”8 But this isn’t very illuminating, given that ‘war’ is not a fixed amount of force. This simply leaves open the question of whether more may be done in defense of India’s 1.3bn members’ rights than in defense of the Vatican’s 500 members’ rights. And certainly international law, and, one would hope, ordinary morality, grants states the right to wage only proportionate wars of national defense (Renzo himself iterates that any war of national defense must satisfy a proportionality constraint). As I argue below, it’s hard to see how proportionality could be wholly insensitive to the number of citizens whose rights are being defended. Moreover, since international law grants states the right to wage wars of national defense regardless of whether the state is valued by its citizens, Renzo’s own account conflicts with the law. This rather undermines his invoking of the law as a reason to doubt my view.
The idea that there is a moral objection to regarding the right of national-defense as numbers-sensitive – that it is “unfair to grant the right to wage wars of national defence to large political communities, but not to small ones” – seems unpersuasive. By way of comparison, consider the widely-held and plausible view that ten people, but not two, may divert a runaway trolley to where it will kill one. Even if there is some sense in which it is unfair that the two may not divert the trolley – if only more people’s interests had been at stake, the two could have defended themselves! – this hardly challenges the view that killing one to save two is disproportionate and thus impermissible. Likewise, it does not seem to me to be importantly unfair that the Vatican may kill far fewer people in defense of its 500 citizens’ rights than India would be permitted to kill in defense of its 1.3bn citizens’ political rights, even if this might mean that the Vatican is not allowed to defend itself on a scale that counts as war.
5.2 The Step-Change Argument
Renzo and I agree that part of the significance of threatening the political rights of a large number of people, compared to frustrating a single person’s right to vote, is that doing so produces a change in the type of harm being threatened. If we prevent, say, all non-Londoners from voting in a uk general election, we do not simply iterate the harm of denying individuals the right to vote, but rather fundamentally change the uk’s political system from a legitimate democracy to an illegitimate dictatorship. I suggested that something similar happens in the case of a bloodless invasion: even if the new rulers do not change the laws, the fact that these laws are now illegitimately imposed, backed by threats of unjust violence should citizens resist, deprives uk citizens of important goods in a way that is quite different from what happens when a single person is wrongly prevented from getting to the polls. Call this the Step-Change Argument.
Renzo agrees with the Step-Change Argument, but also says that it is “besides the point,” since we are interested in “aggressions in which the only interest set back is the interest in maintaining control over political goods such as political independence or territory, even assuming that no further interests are set back.” Renzo insists that it’s at least theoretically important to know what we can do in defense of purely political interests, and, moreover, that bloodless invasions might in fact be very common, since “bloodless aggressions are not wars in which no blood is spilled, but rather wars in which no blood is spilled unless the victims of the attack fail to surrender.”
I agree that we can distinguish ‘pure’ political invasions from invasions that effect a fundamental unjust political change of the sort described above. Indeed, I do distinguish these: as mentioned above, I grant that in the pure cases, defensive force on a scale that would count as war might be impermissible even if political interests can warrant lethal defense. But the Step-Change Argument is intended to support the permissibility of more extensive defensive force in most (if not all) actual cases of bloodless invasion, because most if not all such cases will also threaten these other interests.
The Step-Change Argument isn’t besides the point, then, but rather making a different point – namely that the claim that one may use only limited force in the face of purely political invasions might look overly restrictive, but this turns out to be an illusion once we recognize that hardly any – perhaps no – bloodless invasions will be purely political invasions, if by this we mean that they threaten only political interests without also affecting citizens’ broader interests. One cannot coercively take control of the uk’s political institutions, imposing new political and cultural values on uk citizens and threatening to kill those citizens if they resist, without thereby changing the uk from a legitimate democracy to an illegitimate dictatorship. And thus the harms at stake in these invasions include – plausibly as a conceptual matter – the other serious harms that I identified. The important point is not that invasions can be bloody and yet ‘bloodless,’ but rather that, given that these invasions are always conditionally violent, they always threaten the broader interests that are the focus of the Step-Change Argument.
Renzo’s own proposal focuses on the value of political self-determination, asking whether one may kill merely to protect self-determination (in isolation of the other harms that I consider). In response to Allen Buchanan, who argues that democracy is in fact in tension with individual self-determination, Renzo argues that respecting individual self-determination requires respecting the self-determination of the community, irrespective of whether its decisions align with our personal preferences. By way of analogy, Renzo claims it would wrong him to interfere with his ability to make choices with his partner about how to raise their child, even if that interference brought about his preferred outcome.
But I take it that Buchanan’s concern is that ‘my’ political community might never reflect my preferences: political life is, for many people, like being part of a couple where, after a process of discussion and deliberation, the more dominant partner always gets their way, and the outcome is never what the weaker partner would have chosen. Whilst Renzo considers what we should say about people who do not participate in their political communities – he thinks that these non-self-determining citizens have no reason to prefer the continuation of V to A – he doesn’t consider the case of citizens who participate in deliberative processes, but whose preferences never align with the outcomes of that deliberation. The point is not that these citizens might not value V, but rather that, given that the community consistently frustrates their attempts to satisfy their preferences, it seems a bit implausible to say that they value it as an expression of their own self-determining agency, which is the foundation of Renzo’s account of national-defense.
Renzo acknowledges that his own proposal has an aggregative element, requiring us to compare the interests of V’s self-determining citizens against the interests of its non-self-determining citizens who will be exposed to risks of serious harm by war. However, he claims that this is not problematic aggregation, since it does not commit him to the view that V may employ lethal defense only if V is sufficiently populous. Renzo argues that individuals may kill in defense of sufficiently central aspects of their own interests in self-determination. Since political self-determination is, for Renzo, a manifestation of individual self-determination, it follows that “no matter how small V is, insofar as a sufficiently large share of its population values its membership in the political community, V may permissibly resort to lethal force to defence itself (provided that the jus ad bellum principles are met).”9 A political community of just two people may kill in defense of their collective self-determination.
But my claim is simply that, quite generally, how much force – lethal or otherwise – one may use is determined by how many people’s interests are at stake. Say that a community of 500 people can prevent a coup by killing the leader of a rebel group, and assume that (for the sake of argument) this coup threatens only the self-determination of the group and no other goods. My view permits the use of lethal force in this case. In contrast, if the group would need to kill not only the leader, but also his 1,000-strong army, and perhaps a hundred innocent civilians, my view would prohibit such killings.
But if we increase the size of the threatened community – imagine it’s a small state of two million people –, killing the combatants and the civilians could be proportionate. I find it hard to make sense of how Renzo thinks defensive force could satisfy the ad bellum proportionality constraint, which he mentions several times as a constraint on the right of national-defense, without paying attention to the size of the political community – that is, to the degree of harm threatened by the invasion. He surely does not believe that the community of two people may kill 1,000 people and several hundred civilians in order to protect their right to self-determination, even if this were necessary. But once we grant this, insisting that no matter how small the community, it may always wage a defensive war – if we are understanding war as force above a certain threshold – must be a mistake. And if his claim is rather that the community may use proportionate force, and that this can include lethal force (but perhaps not on the scale of war), depending on how many people are threatened and how many people they would need to kill, his view seems to collapse into mine.
6 Civilian Liability
6.1 Contributing to Harm, and Liability
Drive-By: Terrorist holds a gun to Driver’s head, and orders him to drive the car for a drive-by shooting, in which Terrorist will kill Victim.
Uber: The same as Drive-By except that if Driver refuses, then Uber will happily drive Terrorist to kill Victim in exchange for $20.
Haque thinks it’s still possible that Driver is liable to defensive killing if he drives in Uber (although he claims that “any intuition regarding Driver’s liability seems far weaker than common intuitions about civilian immunity.”) But, he argues, if Driver is so liable, it’s because he jointly perpetrates the killing of Victim with Terrorist. Since, according to Haque, civilians do not jointly perpetrate harms in war, Driver’s liability on grounds of joint perpetration in Uber does not support the view that civilians are liable to defensive harm.
Haque says that he agrees with me that Driver is liable to defensive killing only if he is morally required to allow himself to be killed rather than help Terrorist.11 But we are not, in fact, in agreement on this point. It is not only when I ought to bear a lethal cost rather than contribute to a threat that I can be liable to defensive killing if I contribute. As I argue, Pedestrian need not bear a lethal cost to get out of Victim’s way as he flees Murderer.12 But if she stays put to avoid the trivial cost of spoiling her shoes, Victim may lethally knock her out of the way to save his own life.
Consider Haque’s Uber case again. Haque does not comment on the potential liability of the Uber driver, who will “happily” drive Terrorist to kill Victim in exchange for $20. Imagine that another Uber driver would have driven Terrorist had the first Uber driver refused, so that, according to Haque, the first driver merely superfluously contributes to, rather than enables, the killing.13 My intuition that such a person is liable to defensive killing is not weak at all. On the contrary, it strikes me as very implausible that the driver would retain his usual rights against harm, or that Victim would lack a right of self-defence against him. And the Uber driver’s position is more directly analogous to the position of many contributing civilians than Driver in my original Drive-By case, since, as Haque concedes, most contributing civilans do not contribute to unjust wars in order to avoid serious harms to themselves.14
The problem for Haque is that even if it is true that Driver need not bear a lethal cost rather than drive the car in Drive-By – that is, even if Driver is justified in driving to avoid being killed – this doesn’t settle the question of what to say about people who unjustifiably contribute to unjust lethal threats, such as someone who contributes to a murder in exchange for $20. This is the crucial point, because many civilians voluntarily and unjustifiably contribute to unjust wars. Some do so in exchange for money. Some do so because they support their country’s war effort. Some contribute to avoid comparatively trivial costs, of the sort that, by Haque’s own lights, do not justify contributing.15 Thus, in Uber, Haque unwittingly gives us a much closer analogy to many civilian contributions than I offered in Drive-By. And, I suggest, the intuition that a prospective innocent victim may defensively kill a person who knowingly contributes to an attempt to murder her in exchange for $20 is at least as strong as our intuitions about civilian immunity.
What of Haque’s suggestion that if Driver is liable in Uber, this is only because he is a joint perpetrator of the killing? Haque claims that since civilians jointly perpetrate unjust killings with combatants only if they directly participate in hostilities, and very few civilians do so, Driver’s liability on grounds of joint perpetration does not threaten a general principle of civilian immunity.
Informant: Albert realizes that his neighbors are concealing Beth, a Jew, in their attic. He goes to tells the Nazi authorities, who will raid the house in order to send Beth to a concentration camp, where she will be killed. Beth can kill Albert to prevent his informing on her. Had Albert not informed on Beth, Carl would have informed on Beth.16
I think that Albert is clearly liable to be killed by Beth both to prevent his informing, and after he has informed on her if killing him will enable Beth to escape the Nazis. But Albert does not jointly perpetrate harm: he will not be part of the raid, and to inform is not to directly participate in hostilities. Nor is Albert’s contribution necessary for Beth’s being killed. Thus, it doesn’t look as if civilians can be liable only if they jointly perpetrate harms.
Again, we don’t need to settle the question of whether Albert may inform on Beth to save his own life in order to see that Informant undermines Haque’s claims about civilian immunity. As above, civilians do not typically contribute to wars because they will otherwise be killed. Many contribute, as Haque grants, unjustifiably. Albert voluntarily informs on Beth. Answering the question of when one may justifiably threaten others does not give us a verdict on the liability of unjustified threateners.
6.2 Mediated Harms and Intervening Agency
Pinch: Bully wants to pinch Victim’s arm, which will hurt. Victim knows that if (and only if) he tries to prevent this, Bully will become so enraged that he will kill five people.
Kill: Bully wants to kill Victim. Victim knows that if (and only if) he tries to prevent this, Bully will become so enraged that he will kill five people.
Victim would not be required to bear a lethal cost to rescue five people, and so need not suffer a lethal cost for their sake in Kill.
And yet, Haque points out, Driver in Drive-By would not be required to rescue Victim at the cost of his life either. So why it is impermissible for Driver to drive the car? As Haque puts it, my view holds that Driver has a duty to allow himself to be killed rather than facilitate the harming of Victim, but Victim need not allow himself to be killed rather than trigger the killing of the five. Triggered mediated harms are morally comparable to allowing harm; facilitated harms are morally comparable to doing harm. Haque points out that we need an explanation of this difference.
Video: Anthony is trying to unjustly kill Victim. Bryan wants Victoria dead. Bryan is filming Anthony’s attack on Victim. If Victim kills Anthony in self-defense, Bryan will send the video to Anthony’s brother, Chris. Chris will be so enraged that he will kill Victoria. If Victim does not defend himself, Chris will not kill Victoria.
In Video, Victim’s defense provides Bryan with the means of causing Victoria’s death – namely a video of Anthony’s death that will cause Chris to kill Victoria. However, his doing so is a side-effect of Victim’s defending himself. Neither the video, nor Victoria’s resultant death, is the means by which Victim saves his own life.
In contrast, when Driver helps Terrorist kill Victim in order to save himself, his helping to harm Victim is the means by which Driver saves himself. Since harming people as a means is, in general, harder to justify than harming them as a side-effect, this plausibly explains the different permissions in Kill and Drive-By.
Helen Frowe is Professor of Practical Philosophy at Stockholm University, where she directs the Stockholm Centre for the Ethics of War and Peace.
I think this is generally true of people who breach their duties. See Frowe ‘Civilian Liability’ (unpublished manuscript) and Frowe ‘Intervening Agency, Indirect Contributions and Civilian Liability’ (unpublished manuscript).
‘Preventive harm’ might be a more neutral label between the views that Abel is or isn’t a cause of the threat to Vicky, but I’ll stick with defensive here.
Gerald Lang, ‘What Follows From Defensive Non-Liability?’ Proceedings of the Aristotelian Society (2017) 117/3 231–252. I suggest this in ‘Agent-relative Prerogatives and Harming’ (draft manuscript).
For another discussion of some of these problems, see Joseph Bowen, ‘Necessity and Liability: On an Honour-Based Justification for Defensive Harming,’ Journal of Practical Ethics (2016) 4/2: 79–93.
Ferzan, p. 15.
Rodin, ‘The Myth of National Self-Defence’, in Cécile Fabre and Seth Lazar (eds.) The Morality of Defensive War (Oxford: oup, 2014): 64–84.
Note too that the share must be sufficiently large not only to outweigh non-consensually exposing the non-self-determining citizens to risks of serious harm, but also to warrant the harms that V’s combatants will inflict on A’s combatants and civilians. And, following Jonathan Parry, we might think that the problem is not only that one exposes the non-self-determining citizens to risks of harm, but that one cannot invoke the defense of those citizens as part of the justification for a defensive war. This seems to me to clearly affect the degree of defensive harm that V may inflict. See Parry, ‘Defensive Harm, Consent and Intervention,’ Philosophy and Public Affairs, (2017) 45/4: 356–396.
Contra Haque, the claim that those who responsibly pose indirect threats can be liable to defensive harm is defended at some length in the book.
Haque, p. 6.
I criticize Haque’s notion of superfluity in Frowe, ‘Civilian Liability’.
Adil Ahmad Haque, Law and Morality at War (oup, 2017), p. 71.
Haque, p. 3; Haque, Law and Morality at War (oup, 2017), pp. 70–71.
My discussion here draws on that in Frowe, ‘Civilian Liability’.
Haque raises this as a tension between what I say about mediated harms and enabling harms, but enabling is basically to be an indirect contributor (perhaps of a necessary means).