The Just War Framework

Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) The Just War Framework1 Abstract Much work in the ethics of war is structured around the distinction between jus ad bellum and jus in bello. This distinction has two key roles. It distinguishes two evaluative objects— the war ‘as a whole’, and the conduct of combatants during the war—and identifies different moral principles as relevant to each. I argue that we should be sceptical of this framework. I suggest that a single set of principles determines the justness of actions that cause nonconsensual harm. If so, there are no distinctive ad bellum or in bello principles. I also reject the view that whilst the justness of, for example, ad bellum proportionality rests on all the goods and harms produced by the war, the justness of combatants’ conduct in war is determined by a comparatively limited set of goods and harms in a way that supports the ad bellum–in bello distinction. Keywords: jus ad bellum, jus in bello, proportionality, necessity, evaluating war 1. Jus ad bellum and jus in bello The last decade or so has witnessed a surge of interest in the ethics of war amongst philosophers. Prior to this revival, both philosophical and public debates on the ethics of war adhered to a fairly broad consensus on two central points. First, that war is to be understood as a relation between collectives—most obviously states—rather than as a relationship between individuals. Second, that making judgments about war involves judging whether the war satisfies two sets of established criteria: the principles of jus ad bellum and the principles of jus in bello. Jus ad bellum is commonly interpreted as 1 Thanks to Daniel Statman for helpful suggestions that inspired the paper, and to Seth Lazar, Victor Tadros, Jonathan Parry, and Cécile Fabre for helpful written comments. Thanks also to Tom Dougherty, Massimo Renzo and J. P. Smit for helpful discussions. Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) determining the justice of resorting to war, although, as I’ll argue here and has been observed elsewhere, this is misleading, since on any plausible view these principles also apply across the duration of a war. Jus in bello is commonly understood as determining the justice of combatants’ conduct during war. As Michael Walzer puts it in his influential Just and Unjust Wars, ‘The moral reality of war is divided into two parts. War is always judged twice, first with reference to the reasons states have for fighting, second with reference to the means they adopt.’1 This framework has lately been extended to include jus ex bello (the justice of ending war) and the development of jus post bellum (justice after war).2 More recent work on the ethics of war has challenged the claim that war is a relationship between states. For example, one prominent school of thought known as reductive individualism holds that war is to be understood in terms of the rights and duties of individuals and can be judged by the rules governing harming between individuals outside of war.3 This approach has revisionary implications for many substantive issues in the ethics of war. For example, it undermines the idea that noncombatants are not legitimate targets in war. In ordinary life, people making indirect contributions to unjust harms can forfeit their rights against being intentionally harmed. Reductive individualism thus implies that noncombatants can forfeit their rights against being intentionally harmed by contributing to unjust wars.4 It also challenges the widespread belief that combatants fighting in unjust wars can be the moral equals of those fighting in a just war. When we think about conflicts between individuals, we draw a sharp moral asymmetry between those engaged in unjust aggression and those engaged in justified defence against that aggression. If this asymmetry obtains in war, we must reject the thesis of moral equality between combatants.5 This entails a rejection of one aspect of the bifurcation of war: if just cause matters for the status of combatants, it cannot be true that ad bellum and in bello judgments are, as Walzer claims, ‘logically independent’ of each other. 6 But, surprisingly, Walzer’s model of the structure of just war theory, with the ad bellum–in bello distinction at its core, has survived these revisionist critiques largely intact. 7 Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) While few now endorse the view that issues of jus in bello can be settled independently of jus ad bellum, just war theorists still routinely distinguish between the principles that determine the justness of resorting to war and the principles that determine the justness of combatants’ conduct during the war. Questions about just cause, for example, are described as ad bellum questions, even though the justness of a cause clearly matters in bello. Even those who reject traditional, collectivist approaches to war develop their work around this familiar framework. For example, in Cosmopolitan War, Cécile Fabre cites the way in which just war theory is ‘standardly divided’ into principles governing resort to war and principles governing conduct within war, adopting this distinction for the structure of her book.8 Jeff McMahan distinguishes between, for example, ad bellum proportionality and in bello proportionality.9 David Rodin suggests that the ad bellum–in bello distinction articulates ‘moral reasons appropriate to distinctive forms of moral problem generated by different aspects of conflict’. 10 Books on the ethics of war—such as this one—are typically structured around the ad bellum–in bello distinction, reinforcing the idea that these categories raise distinct moral problems or address different questions. And, of course, international law sharply distinguishes between the laws governing recourse to war and the laws governing conduct in war: as Carsten Stahn puts it, jus ad bellum and jus in bello are as regarded as ‘distinct normative universes’.11 Given the prevalence of the ad bellum–in bello distinction and its apparently ingrained place in our just war thinking, one could be forgiven for assuming that it must have an ancient lineage in the just war tradition. But, as Robert Kolb puts it, while ‘the august solemnity of Latin confers … the misleading appearance of being centuries old’, the structural use of the distinction that Walzer describes is a comparatively recent phenomenon.12 Oliver O’Donovan describes the ‘modern (not traditional) distinction between just resort to war (ius ad bellum) and just conduct in war (ius in bello)’ as a ‘secondary casuistic distinction, not a load-bearing one’.13 Nicholas Rengger argues that whilst the distinction is a useful heuristic, ‘[t]he problem is that the modern revival of the Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) [just war] tradition has elevated it to an architectonic’. 14 This doesn’t mean, of course, that early just war scholars were not writing about the justness of resorting to war or about the just conduct of war. But it does mean that we should be cautious about suggesting, as McMahan does, that just war theory ‘has traditionally been divided into two sets of principles’.15 Given what McMahan says immediately afterwards—that classical theorists such as Francisco de Vitoria took the justness of actions within war to rest on the justness of resorting to war—it seems neither helpful nor accurate to regard the ad bellum–in bello distinction as a historical division between two sets of principles if that division is meant to reflect mutual independence. (Indeed, it’s a bit hard to make sense of the idea of dividing sets of principles at all.16 The view McMahan has in mind is more accurately described as the claim that war itself is divided into resort and conduct and that some moral principles can be sensibly applied to only one, but not both, of these aspects of war.) The view that I will defend here—that the categories of jus ad bellum and jus in bello are not useful—perhaps looks less heretical against this historical background. In Section 2, I suggest that there are no distinctive ad bellum or in bello principles. Any principle that matters for the justness of declaring war matters for the justness of a specific offensive in war. We might grant that this claim but argue that these principles are sensitive to different goods and harms depending on whether we are talking about jus ad bellum or jus in bello. When we make ad bellum proportionality judgments, for example, we take into account all the (predicted) harms and benefits of the war. But when we make in bello proportionality judgments, we take into account only a comparatively small range of goods and harms pertaining to a specific offensive. In Section 3, I argue that this view is mistaken: the justness of in bello actions is not limited in this way. The justness of an action depends on all its interactions with other actions, and this is no less true for the justness of individual offensives than for the justness of the decision to declare war. Given this, there’s no Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) principled demarcation of relevant effects and harms that is reflected by the ad bellum–in bello distinction. Seth Lazar has suggested that anyone who endorses some kind of collectivism in ethics will need some analogue of the ad bellum/in bello distinction. I agree that if one endorses collectivism, one will need to do some demarcating of relevant actions (e.g., if one believes that an individual can be ‘on the hook’ for the wrongful actions of other members of her collective, one will be picking out a subset of actions for which she can be thus responsible). I’m not a collectivist, although I don’t defend anti-collectivism here. However, it’s not clear to me why we would call this an analogue of the ad bellum–in bello distinction: the relevant collectives and subset of actions aren’t going to map onto that distinction. So I think collectivists can share the reasons I give here for scepticism about this pervasive bifurcation of war. Section 4 addresses the thought that jus ad bellum gives us valuable summative assessments of war. I suggest that summative assessments of war—understood as a single verdict on the war as a whole—are not useful and that it is a mistake to focus our moral evaluations in this way. In any war, some individual actions will be just and some will be unjust. Even wars that we typically describe as just—such as the Allies’ war against Nazi Germany—contain many instances of wrongdoing, ranging from mildly disproportionate or overly risky offensives to appalling atrocities. Given this, it is unsurprising that people often baulk at sweeping evaluations that purport to give us an overall assessment of a war (commonly expressed as the thought that ‘there’s no such thing as a just war’). One response to this kind of scepticism is to embrace pacifism, concluding that war is always impermissible. But those who think that war can be permissible typically respond by disaggregating the actions of the war. We grant that, of course, some of the offensives were wrong, and some atrocities were committed, and that these actions were unjust. But the decision to resort to war can still be justified in light of the importance of securing the just cause despite foreseeing the concomitant wrongdoing. I suggest that if we’re interested in Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) the morality of a war, we should be primarily interested in these more nuanced, individualized evaluations, rather than summative assessments of wars. 2. The Principles Governing Nonconsensual Harming In this section, I argue that there is no difference in content between jus ad bellum and jus in bello. There is, rather, a single set of principles that governs both the justness of resorting to war and the justness of particular acts within war. This is a formal claim that holds irrespective of the substantive content of this set of principles: it holds simply that whichever principles determine the justness of harming as part of a specific offensive within war also determine the justness of declaring or continuing war. Contrary to Rodin’s suggestion, there are no distinctive moral reasons that matter only, for example, in bello and not ad bellum. I think this formal claim is a natural upshot of various arguments that have been put forward in support of reductive individualism as well as of the more general view, reductivism (which might or might not be individualist). Defences of reductivism in just war theory typically focus on the relationship between harming in war and harming in selfdefence. But the more general reductivist claim is simply that there is a single set of principles governing all actions that cause nonconsensual harm, whether those actions are part of self-defence, policing, measures short of war, or war. If this is correct, there cannot be different principles that govern the resort to war compared to the conduct of war, and so reductivists should find the view defended here appealing. However, the claim that actions that cause nonconsensual harm are always judged by the same principles is also open to those who take a nonreductivist or ‘exceptionalist’ approach to war (although exceptionalists will think that some of the justifications for harming generated by these principles are satisfied only in war). I’ll defend the formal claim by considering the principles traditionally identified as governing jus ad bellum and jus in bello. I’ll argue that each of these principles either Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) matters both ad bellum and in bello or is subsumed under one of the other principles. Thus, there are no distinctive ad bellum or in bello principles. However, the focus on these particular principles is primarily illustrative. It’s compatible with the formal claim that a completely different set of substantive principles govern the inflicting of nonconsensual harm. On traditional accounts of jus ad bellum, a war is just if and only if it has a just cause, is a last resort for securing that just cause, is proportionate to that just cause, is waged in order to secure that just cause, is waged by a legitimate authority, fought for the right intention, and has a reasonable prospect of success. The standard account of the content of jus in bello is a requirement to discriminate between legitimate and illegitimate targets and to cause only harm that is necessary for securing and proportionate to a military advantage. It is uncontroversial, then, that the constraints of necessity and proportionality apply to the resort to and continuation of war and to specific offensives within war. And, I think it is now fairly widely accepted that the criterion of just cause plays a central role in our judgments about specific actions and offensives in war.17 As Jeff McMahan and Thomas Hurka have argued, having a just cause is a prerequisite of any individual offensive’s satisfying proportionality. Proportionality weighs morally relevant harms against morally relevant goods. If one lacks a relevant moral good to secure—if one’s aim is simply to contribute to the unjust expansion of territory, for example—there’s nothing to outweigh the harms that one causes.18 A similar argument applies with respect to necessity. The fact that force is the least harmful means of achieving an end can form part of a justification for using force only when the end is morally good. The fact that I need to kill you to steal your wallet does not provide some partial justification for killing you. When the end is morally wrong, there’s no role for necessity to play.19 Notice that this does not mean that the only just cause for using force is the just cause that initially gave rise to war. As McMahan has pointed out, there could be other just causes for using force, such as defending noncombatants whose country lacks a just cause Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) from unjust attacks by combatants whose country is fighting for a just cause. 20 A combatant might have a just cause for some of her actions and lack a just cause for others. I’m somewhat sceptical about the importance of right intention, where this is understood as requiring that a war be fought for the reasons that in fact justify it and not as a pretext for furthering some other national interest, even if the war would in fact secure the just cause via (otherwise) legitimate means.21 Similarly, I’m not sure how much intentions matter for the permissibility of individual actions. But whilst this debate will affect the substantive content of the set of principles governing harming, it poses no threat to the formal claim I’m defending here since it’s hard to see how intentions could matter for the justness of declaring or continuing of a war but not for the justness of specific offensives that make up the war. If one’s intentions partly determine the permissibility of one’s actions, this will be equally true for combatants engaged in fighting and politicians engaged in declaring war. That is, if there is a right intention constraint on the use of force, it will equally constrain the declaring, continuing, and fighting of war. 22 I’m also sceptical, as are many other just war theorists, that the traditional notion of legitimate authority plays any role in determining the justness of war. 23 The idea that only heads of state are eligible to declare war, or that only their wars can be just, is shown to be implausible by civil wars and just revolutions. The more recent appeals to consent or authorization as a way of capturing what seems morally relevant about authority strike me as attractive,24 but I also think that these considerations will come under the remit of the proportionality constraint. Roughly, whether, for example, the intended beneficiaries of a war consent to being exposed to the (risk of the) harms of war affects the goods that can be included in the proportionality calculation.25 But again, if one thinks legitimate authority important for the justness of resorting to war, one should extend this to jus in bello: it will also matter whether a combatant is killing on behalf of a state or other relevant group. Finally, I’m also unsure whether the requirement that wars should have a reasonable prospect of success is best conceived of as an independent constraint on the use Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) of force or as part of the proportionality constraint. But, whichever model we prefer, it will matter for the declaring, continuing, and fighting of war—all must have a reasonable prospect of achieving their ends if they are to justify the inflicting of collateral harm. Even if a combatant may kill twenty civilians as a side effect of securing a given military advantage, it does not follow that she may kill twenty civilians as a side effect of an offensive that has only a very low chance of securing the military advantage. The foregoing supports the view that those principles typically identified as governing jus ad bellum also apply in bello. But to secure the conclusion that there are no distinctive ad bellum or in bello principles, we also need to show either that we can omit the requirement of discrimination from our principles governing the use of force or that discrimination also matters for the resort to and continuation of war. This requirement enjoins combatants to distinguish between legitimate and illegitimate targets—traditionally interpreted as the distinction between military personnel, equipment, and installations, on the one hand, and civilians and civilian infrastructure on the other. More recently, some writers have urged that discrimination should be interpreted as a requirement to discriminate between people who are liable to be harmed and people who are not so liable.26 There are two ways in which the requirement of discrimination might be employed (whatever the relevant criterion for discrimination turns out to be). The first is as an absolute prohibition on intentionally harming the members of the relevant group (e.g., on intentionally harming civilians or intentionally harming nonliable people). The second is as a claim that intentionally harming the members of the relevant group is harder to justify than harming other people. Whichever model we choose, though, it will matter for the justness of declaring war as well as the justness of specific offensives. If one predicts that the war will be fought in a way that involves intentionally harming members of the relevant group, this will either make it impermissible to declare war (on the absolutist model) or harder to satisfy proportionality (on the higher justificatory threshold model). Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) 3. Relevant Goods and Harms Say we grant that whatever moral principles determine the justness of conduct within war also determine the justness of resorting to war or of continuing a war. Nevertheless, we might insist that there are still important differences between our ad bellum and in bello judgments. Specifically, we might think that, even once just cause is satisfied, the range of goods and harms relevant to in bello judgments is much narrower than the range of goods and harms relevant to ad bellum judgments. I’ll address this objection in the context of our judgments concerning proportionality and necessity. 3.1. Proportionality Henry Shue argues that there are two proportionality principles: a macro-level test concerning the resort to war, as well as a micro-level test concerning the conduct of war … thought each to apply independently. One must ask prior to going to war, would the evil to be prevented by military action in this case be worth engaging in a war overall, and ask throughout any war engaged in, would this particular military engagement make a sufficiently great contribution to potential victory to be worth the death and destruction likely to result? (Shue, 2005, p. 748) David Rodin similarly suggests that, with respect to jus in bello, ‘proportionality requires balancing the harm of a particular action in the course of a war against the requirements of military necessity. It prohibits military actions which are “excessive in relation to the concrete and direct military advantage anticipated”’. But, in contrast, ‘proportionality in the context of the jus ad bellum … requires a fundamentally different kind of balancing— between the harms brought about by the pursuit of a defensive war and the nature of the threat to which the war is a response’.27 The language of the ‘harm of a particular action’ and the ‘direct military advantage’ reflects the narrow range of outcomes that is typically thought relevant to combatants’ conduct in war compared to the broad array of goods and harms that determine ad bellum proportionality. Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) According to the view I will defend here, it’s true that the justness of declaring and continuing war is grounded in a very wide range of goods and harms. But, I will argue, all these goods and harms are also relevant to the justness of specific offensives. Nothing that is relevant to the justness of a political leader’s decision to continue a war is irrelevant to the justness of a combatant’s decision to continue fighting in that war, and vice versa. There’s thus no difference between what I’ll call the sensitivity of moral principles that varies depending on whether we are talking about an ‘ad bellum’ action or an ‘in bello’ action. Here’s a simple case that demonstrates why the relevant goods and harms for proportionality are not limited to those that one directly causes. Prediction: A runaway trolley is heading to where it will break innocent Engineer’s leg. Engineer can divert the trolley down a sidetrack to where it will break Workman’ finger. But Engineer knows that if she diverts the trolley, Workman will then divert it again down a second side-track, killing innocent Pedestrian in order to spare himself the cost of a broken finger. If Engineer’s proportionality calculation were sensitive only to the goods and harms that she herself causes, it would not be disproportionate for her to divert the trolley. She would be permitted to break Workman’s finger as a side effect of defending her leg, and she does not directly cause the death of Pedestrian since that death depends on Workman’s intervening agency. But the fact that Workman will predictably kill an innocent person to avoid the broken finger renders Engineer’s defence disproportionate. She may not divert the trolley away from her leg if she knows that Workman will then lethally divert the trolley towards Pedestrian. The range of harms that are relevant to the permissibility of her defence extend beyond those that she will directly cause. This doesn’t mean, of course, that actions that give rise to wrongdoing are always impermissible. The interactions between our actions are more complex than that. For Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) example, one can sometimes be justified in acting even if one foresees that others will then engage in wrongdoing as a consequence, as in Prediction Two: Prediction Two: A runaway trolley is heading to where it will kill fifty people. Bystander can divert the trolley down a sidetrack to where it will break Workman’s legs. But Bystander knows that if she diverts the trolley, Workman will then divert it again down a second side-track, killing an innocent person in order to spare himself the cost of broken legs. In this case, Bystander is justified in diverting the trolley despite knowing that, if she does so, Workman will then engage in the wrongful killing of an innocent person. It would have been proportionate for Bystander to directly divert the trolley to where it would kill one innocent person as a side effect of saving the fifty, and so her prediction that Workman will wrongly inflict this harm does not make Bystander’s action disproportionate. But Workman’s actions are still relevant to the proportionality of her Bystander’s actions: if she knows that Workman will divert the trolley to where it kills a hundred people, she may not divert it away from the fifty. There can also be cases in the permissibility of our actions that can depend upon wrongdoing, such as Murderous Prediction: Murderous Prediction: A runaway trolley is heading to where it will kill Walker. Passer-By can divert the trolley away from Walker down a side-track. Jogger is on the side track, and will be killed if the trolley hits her. However, Murderer is just about to shoot Jogger in the head, so she will already be dead when the trolley hits her. It would be disproportionate for Passer-By to divert the trolley if she would thereby kill Jogger. But when she sees that Murderer will unavoidably kill Jogger anyway, diverting the trolley becomes proportionate and permissible in virtue Murderer’s wrongdoing. Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) These are just some of the myriad ways in which our actions can interact and that will affect their permissibility. Notice that none of these claims depends on any underlying commitment to collective responsibility or notions of acting together. There’s no sense in which Bystander and Workman are acting together, or Passer-By and Murderer are acting together. And yet what Workman does can determine the proportionality of Bystander’s actions, just as what Murderer does can determine the proportionality of Passer-By’s saving Walker. How our actions interact with other actions is equally important in war: whether a specific offensive is justified is not, as Rodin suggests, determined only by the goods and harms that it directly produces. Consider Torture: Torture: Captain A is leading Unit A on a mission to break into the enemy’s military base. Captain B is going to provide Captain A with essential information for getting through the base’s security. Once inside, Captain A’s unit will kill a proportionate number of combatants to secure the base. There is no less harmful way to secure the base, which is important for winning a just war. However, the information that will be provided by Captain B will be extracted by Unit B from a captured enemy combatant by torturing his children in front of him. Assume that torturing the children is so harmful as to make Unit B’s actions disproportionate. Considered independently of Unit B’s actions, nothing that Unit A does is impermissible. They will use proportionate force to secure an advantage that is necessary for winning their just war. But the harms that Unit B will inflict are surely relevant to whether Unit A acts permissibly. If refusing to carry out their offensive might prevent the torture of the children, for example, it’s plausible that Unit A should refuse to capture the base, even though the harm of torture is not a direct result of their offensive. It is false, then, that the range of goods and harms relevant to in bello judgments is limited to those directly caused one’s own actions or by the actions of a very small subset of individuals tied Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) to a specific offensive. It also matters how one’s actions causally influence, or depend upon, the actions of others. Moreover, there’s no reason to think that the range of relevant goods and harms is limited to those brought about by one’s own unit, or platoon, or regiment. Those are not morally significant demarcations. Even if one is more likely to causally influence the actions of one’s own unit, this is a merely contingent fact, not indicative of some genuine moral boundary. Any goods and harms—whether they are inflicted by one’s own side or even by the opposing side—can affect the permissibility of one’s own actions. Once we recognize this, it looks like the range of goods and harms that determines the justness of any individual combatant’s decisions could in principle be as extensive as the range of goods and harms that determines the justness of a political leader’s ad bellum decisions. We might object that it is surely implausible that a combatant’s actions could be rendered impermissible by the actions of some other individual or individuals who are perhaps thousands of miles away. But closer reflection suggests otherwise. We can see this most clearly with respect to just cause: if a political leader is offered reasonable peace terms that she declines, continuing to fight may be unjust because war is no longer the least harmful means of securing the just cause. But fighting can also be rendered disproportionate by the excessive destruction of troops elsewhere. If a combatant’s evidence is that the war will continue to be fought in a disproportionate way, this, too, could make it impermissible for her to continue to fight. The fact that those troops are not physically close to her or part of her own unit does not make their excesses irrelevant to the justness of her participation in the war. Note that this claim is not vulnerable, on either a fact-relative or an evidencerelative account of justification, to the objection that no combatant could possibly know such an extensive range of facts as a political leader.28 On the fact-relative account, the justness of action depends on objective truths. It does not matter whether the combatant (or the political leader) knows these truths. If her action is fact-relative unjustified, she will Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) act wrongly in performing it. What she believes or has evidence of might furnish her with an excuse, but it won’t provide her with a justification for acting. On an evidence-relative account, the justness of an action depends on whether that action is the right thing to do in light of the agent’s available evidence. It is possible that, if she has different evidence available to her, an ordinary combatant’s continuing to fight might be justified even if her leader’s continuing the war is unjustified (and vice versa). But that doesn’t undermine the claim that there is no limited range of goods and harms that determines the permissibility of specific offensives compared to the wide range of goods and harms that determines permissibility of declaring or continuing war. Both should act according to their evidence—which, for a combatant, includes her evidence about the goods and harms arising from other people’s actions even if they are causally or physically remote from her own. All that matters for the view being defended here is that there is no evidence that might be significant for the justness of a leader’s actions but irrelevant to the justness of a combatant’s actions. It looks implausible, then, to hold that the ad bellum–in bello distinction does useful work by directing our attention to distinct sets of goods and harms that determine the answers to different sorts of questions we can ask about war. The justness of specific offensives is not sensitive to only a narrow range of goods and harms that are a direct upshot of that offensive. Anything that might be relevant to the justness of declaring or continuing a war is relevant to our evaluation of specific offensives not only because combatants’ just cause for using force will typically be the cause justified the leader’s decision to declare war, but also because the justness of the means by which that cause is pursued depends on how those means interact with other actions. 3.2. War-Dependent Goods We might think that the sensitivity of ad bellum proportionality differs from the sensitivity of in bello proportionality in the following way. Ad bellum proportionality weighs the harm Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) we expect to cause in war against the harm we hope to prevent by waging war. But this cannot exhaust what is weighed in bello. Consider the following case: Rescue: Ten members of Unit A, whose country is fighting a just war, are captured by a group of enemy combatants, who will kill them at dawn. Unit B have a plan to rescue them. In doing so, they will collaterally kill one civilian. It looks like Unit B’s proportionality calculation should weigh the good of rescuing the ten combatants against the harm of killing of the civilian. But the good of saving the captured combatants cannot be a factor in our ad bellum proportionality calculation—not because rescuing people can’t be a just cause for resort to force, but because (in this case) the existence of the good of the rescue is conditional on whether we resort to force. One cannot count in one’s decision about whether to go to war the good of rescuing a person from a harm that he will face only if one decides to go to war. These are, in a sense, goods that don’t yet exist since their existence turns on the decision to fight. Thus, it looks like there will be goods that matter in bello that do not count ad bellum.29 (This is compatible with thinking that the harms of cases like Rescue count ad bellum. We can predict that our combatants will be captured and that rescuing them will cause harm, and we should include these harms in our assessments of whether to go to war.) I think that this objection plays upon the fairly pervasive, but mistaken, idea that the principle of jus ad bellum apply to a one-off judgment made prior to resort to war. Some people explicitly endorse this idea. Brian Orend writes that ‘jus ad bellum … concerns the justice of resorting to war in the first place’. 30 Megan Braun and Daniel Brunstetter claim that ‘in war, principles such as just cause and last resort need only be satisfied at the outset of a conflict’. 31 Others slip into this terminology when speaking rather more loosely: Fabre recently suggested that jus ad bellum, jus in bello, and jus ex bello are ‘a convenient way to demarcate various phases in the initiation, conduct and termination of a war’, even though she also claims that the categories lack ‘deep conceptual or normative significance’. 32 Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) But we can see that this understanding of jus ad bellum is mistaken by thinking about ways in which a war might satisfy the conditions for the just use of force at its outset but cease to satisfy one or more of those conditions as the war progresses. For example, it might become apparent some months into a war that it cannot be won without a disproportionate loss of life. Or, a less harmful alternative to war might suddenly present itself. Most obviously, a group may not continue to fight after it has secured its just cause (unless some other just cause arises, such as ensuring stability in the defeated state or defending civilians against revenge attacks). We can ask of a war at any time whether it is just, and our answer might vary between different stages of the same war. 33 How does this bear on the objection concerning the role of war-dependent goods in proportionality? Well, we can grant that when a leader is thinking about declaring war, she cannot factor in savings that don’t yet need enacting and that will need enacting only because of the war. If jus ad bellum judgments were restricted to the declaration of war, it would seem that there are indeed some in bello goods that don’t count ad bellum. But once the war is being fought and the combatants need saving, the good of saving them becomes relevant to whether the leader may continue to use force—that is, to the ad bellum question of whether she may continue the war. Preventing unjust imprisonment and killings are legitimate aims of force. Thus, rescuing combatants who have been captured in the course of their legitimate defence is a good of the war that can be factored into her proportionality calculation about whether the war should continue to be fought. We can see this clearly if we imagine that peace terms are offered that are satisfactory in every respect except that the enemy proposes to execute all of its prisoners of war. These people would not have been endangered were it not for the war. But the good of saving them certainly matters for our assessment of whether continuing to fight the war is proportionate. The fact that these goods aren’t included in the ad bellum proportionality calculation prior to war is irrelevant—they can’t be included in any in bello calculations prior to war either since there is no in bello at that point. Cases such as Rescue suggest that there are some Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) goods that might matter when we’re considering the continuation of war that don’t count when we’re considering the resort to war. But that doesn’t support a difference in sensitivity between jus ad bellum proportionality and jus in bello proportionality. All that matters for the purposes of my argument is that, at any time at which there is an in bello proportionality calculation that is sensitive to these kinds of goods, ad bellum proportionality will be similarly sensitive to them. 3.3. Necessity On the standard view, ad bellum necessity governs the whole war, such that war is justified only if it is the least harmful means available of securing the just cause. For example, a political leader who is wondering whether to continue with a war should compare the costs and harms of war with, for example, those of diplomacy or surrender. In bello necessity, in contrast, governs specific offensives, such that an offensive is justified only if it is the least harmful means available of securing some specific end that will contribute to winning the war. A combatant who is deciding whether to carry out a particular offensive has no option of ending the war or accepting peace terms. Thus it seems that her in bello deliberations about necessity are very different from the ad bellum deliberations of a political leader. But the idea that necessity is indexed to a particular agent and her available options is a familiar one.34 Consider Attack: Attack: Attacker is trying to simultaneously kill Amy and Beth. Amy has a shotgun that she can use to defend herself by killing Attacker. Beth has a taser with which she can nonlethally disable Attacker, but she refuses to use it. Beth acts impermissibly in refusing to use the taser. But her refusal means that the least harmful means of defence available to Amy is killing Attacker with the shotgun. Amy’s action of killing Attacker therefore satisfies the necessity condition. Similarly, from the perspective of a political leader, it might be that an unnecessarily violent (but still proportionate) war is the least harmful means of securing a just cause available to her. She may predict that her troops will cause more harm than is necessary Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) and that a less bloodthirsty army would secure the just cause less harmfully. And yet, since those are the only troops she has and their fighting will still be proportionate, it could still be permissible for her to order the war. But the observation that what is necessary depends on the set of actions available to an agent doesn’t show the distinction between jus ad bellum and jus in bello to be a useful one. It merely shows that it’s still true in war that necessity is indexed to an agent’s options. At any point in the chain of command, combatants will have different option sets available to them: some will direct entire campaigns, whilst others will lead particular missions or participate in specific offensives. Nothing helpful follows from singling out the perspective of the political leader and calling this ad bellum necessity, as if this is somehow tells us that the war overall is the least harmful means of achieving the just cause. In the case described, the war as it will be fought is not the least harmful means—many of the offensives will be gratuitously violent. It can be important, of course, that we recognize that it was still the least harmful means available to the political leader. But privileging her option set as determining the overall necessity of the war seems to me wrongheaded. As I’ll suggest next, our moral judgments of war should primarily involve a more nuanced assessment of individual actions rather than summative assessments of the perspective of a political leader. 4. Summative Assessments of War Surely, we might insist, there is value in asking whether, for example, the Kosovo intervention was just or ought to have been fought—that, as Seth Lazar claims, ‘we both do and need to be able to evaluate wars as a whole.’35 Lazar argues that if we focus on the justification of individual actions, we can no longer engage in this valuable practice. Rather, we can ask only about the justification of individual actions that make up the war and then offer the thought that a war is justified only if each of those individual actions is justified. He argues that this sets the bar for justification too high, making ‘justified wars practically impossible’.36 Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) I think we should be cautious about how valuable summative assessments of wars really are. By ‘summative assessments’, I mean assessments that purport to aggregate all the actions of the war to produce a single verdict of ‘just’ or ‘unjust’—as Lazar puts it, to ‘evaluate the war as a whole’. O’Donovan suggests that such assessments are not, in fact, the purpose of just war theory: [I]t is very often supposed that just war theory undertakes to validate or invalidate particular wars. That would be an impossible undertaking. History knows of no just war, as it knows of no just peoples. Major historical events cannot be justified or criticised in one mouthful; they are concatenations and agglomerations of many separate actions and many varied results. One may justify or criticise acts of statesmen, acts of generals, acts of common soldiers or of civilians … but wars as such, like most large-scale historical phenomena, present only a great question mark, a continual invitation to reflect further on which decisions were, and which were not, justified.37 A more nuanced approach of this sort, one that does not try to aggregate all the component actions and harms of a war into a single verdict, strikes me as a much more useful way to evaluate wars. For example, imagine that we want to know whether a given war was proportionate. One way to assess this is to simply take the total harm inflicted—lives lost, injuries inflicted, cities destroyed—and ask if this total harm was warranted by (trying to) avert the threat that was the cause for war. If it was, we might think that we don’t need to know anything more to know whether the war was proportionate—that we don’t need to know how many specific offensives were proportionate and how many were not. But this strikes me as morally misleading: it’s unclear why we should be interested in that kind of overall judgment. Imagine a war in which some very important targets were captured or destroyed with very little collateral harm, even though their capture or destruction would have warranted significant collateral harm, and many comparatively Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) minor targets were captured or destroyed with vastly disproportionate force. It might be that the total amount of harm inflicted in this war was proportionate to the just cause. But pronouncing the war ad bellum proportionate simply fails to capture the fact that many of the composite actions of the war were disproportionate. And if we’re concerned with the morality of a war, then the more nuanced, individualized approach strikes me as much more interesting. In the example just given, the overall proportionality of the war would not undercut our duties to compensate those harmed in disproportionate offensives or to punish those involved in carrying out those offensives. Nor would the fact that a combatant fought in this war tell us whether she acted permissibly. Asking whether the war as a whole was proportionate seems akin to asking whether Bystander and Workman’s actions in the Prediction cases are, taken together, proportionate. This doesn’t mean that we cannot answer questions about whether, for example, war should be declared. It can be true that a politician ought to pursue a war that could be fought less harmfully if fighting it less harmfully is not an option for her, and that it is better that she declare war than not, and so on. But even if we determine that the action of declaring war is justified, this should not be viewed as a verdict that the actions of the ensuing war are justified. It’s merely an assessment of the leader’s action, given the options available to her. We can similarly assess whether Bystander in Prediction is justified in diverting the trolley without taking our answer to provide an overall assessment of whether Bystander and Workman’s actions taken together are justified. In some sense, Lazar is correct that this approach entails that a war can be properly described as justified only if each composite action is justified. But this is not an objection since this approach holds that the overall description of a war as justified is not terribly interesting. What we should be interested in is whether a political leader was justified in declaring or continuing a war, and she can be so justified even foreseeing that others might engage in wrongdoing as a result, as in Prediction.38 Indeed, one can be justified in acting even if the success of one’s action depends upon wrongdoing, as in Murderous Prediction. Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) The upshot of focusing on individual actions is not that we set the bar for justified resort to war impossibly high. It’s rather a recognition that, just as it’s unhelpful to ask in the Prediction cases whether the agents’ actions ‘taken together’ or ‘as a whole’ were justified, it’s unhelpful to ask whether all the actions of a war, taken together, were justified. There can be no meaningful single verdict on those actions taken together. 5. Conclusion I have suggested that there is no difference in content between the principles of jus ad bellum and the principles of jus in bello. Rather, a single set of principles governs the declaring, continuing, and fighting of war. I have also argued that we should reject the idea that judgments about the justness of particular offensives in war are limited to the evaluation of goods and harms directly caused by that offensive. The justness of any action depends on how it interacts with other actions, which can include actions performed by other units, even if they are physically or temporally remote, or by members of allied or opposing forces. Once we widen the scope of relevant effects in this way, the idea that the ad bellum–in bello distinction helpfully demarcates the range of relevant evaluanda for moral principles looks false. I also rejected the idea that the category of jus ad bellum enables us to make valuable summative assessments of war, understood as a single verdict on the justness of the war as a whole. There is no helpful evaluation to be made about whether the set of individual actions that compose a side of a conflict are, taken together, ‘overall’ just or justified. This is compatible with saying that a political leader was justified in declaring or continuing a war. But it denies that we should take the perspective of the political leader— whether a war was necessary from her perspective, for example—and take this to determine whether the war ‘taken as a whole’ was necessary. A war that was necessary from the leader’s perspective could include many gratuitously harmful offensives. If we know this, it is misleading to nevertheless declare the war just with respect to necessity. Our moral Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) evaluations of war should be more nuanced, focusing on the justification of individual actions rather than on attempts to pronounce on the war as a whole. Notes 1 Michael Walzer, Just and Unjust Wars: A Moral Argument with Historical Illustrations (New York: Basic Book, 1977). 2 See, e.g., Darrel Moellendorf, ‘Two Doctrines of Jus Ex Bello’, Ethics 125 (3: April 2015), 653–673. David Rodin prefers the term jus termination to jus ex bello. See, e.g., Rodin, ‘The War Trap: Dilemmas of Jus Terminatio’, Ethics 125 (3: April 2015), 674–695. 3 See, for example, Jeff McMahan, Killing in War (Oxford: Oxford University Press, 2009); Cécile Fabre, Cosmopolitan War (Oxford: Oxford University Press, 2011); Helen Frowe, Defensive Killing (Oxford: Oxford University Press, 2014); Victor Tadros, ‘Orwell’s Battle with Brittain: Vicarious Liability for Unjust Aggression’, Philosophy and Public Affairs 42 (1: 2014), 42–77. Of course, there are other theoretical approaches to war—see Seth Lazar’s contribution to this volume for an excellent taxonomy. 4 See, e.g., Frowe, Defensive Killing; Kai Draper, War and Individual Rights (New York: Oxford University Press, 2015) 5 E.g. McMahan, Killing in War. 6 Walzer, Just and Unjust Wars, 21 7 In ‘War’s Endings, and the Structure of Just War Theory’, Seth Lazar argues that we should replace the ad bellum–in bello distinction with what he claims is a more useful a distinction between Command Ethics and Combatant Ethics. I think that Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) proposal is vulnerable to the same objections that I’ll level here against the ad bellum–in bello distinction. 8 Fabre, Cosmopolitan War, 4. 9 McMahan, Killing in War, 19. 10 David Rodin, ‘The War Trap: Dilemmas of jus terminatio’, Ethics 125 (3: April 2015), 674–695, at 676. 11 Carsten Stahn, ‘”Jus ad bellum”, “jus in bello” … “jus post bellum”? Rethinking the Conception of the Law of Armed Force’, The European Journal of International Law 17 (5: 2007), 921–943, at 925. 12 Robert Kolb, ‘Origins of the Twin Terms Jus Ad Bellum and Jus in Bello’, International Review of the Red Cross 37 (Special Issue 320: 1997), 553–562. Thanks to Greg Reichberg and Rory Cox for helpful discussion of this. 13 Oliver O’Donovan, The Just War Revisited (Cambridge: Cambridge University Press, 2003), 15. 14 Nicholas Rengger, ‘Jus in bello’, in War: Essays in Political Philosophy, edited by Larry May and Emily Crookstone (Cambridge: Cambridge University Press, 2008), 32. 15 McMahan, ‘War’, in The Oxford Handbook of Political Philosophy, edited by David Estlund (New York: Oxford University Press, 2012). 298. 16 Thanks to Victor Tadros for pointing this out. 17 See, e.g., Estlund, ‘On Following Orders in an Unjust War’, Journal of Political Philosophy 15 (2: 2007), 213–234; Lazar, this volume. Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) 18 Thomas Hurka, ‘Proportionality in the Morality of War’, Philosophy and Public Affairs 33 (1: 2005), 34–66, at 37; and Jeff McMahan, ‘Just Cause for War’, Ethics and International Affairs 19 (3: 2005),1–21, at 5. 19 Seth Lazar suggests that gratuitously harming someone is worse than harming a person because it’s necessary to achieve an unjust end (see Sparing Civilians, chapter 3) I disagree: I don’t think that an unjust harm becomes less bad if it also contributes to achieving some further unjust end. 20 McMahan, ‘Just Cause for War’, 2. See also Saba Bazargan, ‘Aiding and Abetting Unjust Wars’, Journal of Moral Philosophy 8 (2011), 513–529; Uwe Steinhoff, ‘Debate: Jeff McMahan on the Moral Equality of Combatants’, Journal of Political Philosophy 16 (2: 2008), 220–228; and James Pattison, ‘When Is It Right to Fight? Just War Theory and the Individual-Centric Approach’, Ethical Theory and Moral Practice 16 (1: 2013), 35–54, at 40. 21 See Frowe, ‘Judging Armed Humanitarian Intervention’, in The Ethics of Armed Humanitarian Intervention edited by Don E. Scheid (Cambridge: Cambridge University Press, 2011). For the view that intentions do matter, see Tadros, ‘ Journal of Practical Ethics, and his contribution to this volume. 22 Rengger argues that, e.g., Augustine also took there to be a right intention constraint on ordinary combatants that ‘cut across the “dividing line” of jus ad bellum and jus in bello that is in fact absent in Augustine …’. 23 See, e.g., Cecile Fabre, ‘Cosmopolitanism, Just War Theory and Legitimate Authority’. 24 See, e.g., Jonathan Parry, this volume; Seth Lazar, ‘Authorisation and the Morality of War’, Australasian Journal of Moral Philosophy 94 (2: 2016), 211–226. Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) 25 See Jonathan Parry, this volume. 26 McMahan, Killing in War; Frowe, Defensive Killing. 27 David Rodin, War and Self-Defense (New York: Oxford, 2002), 114. 28 Fact-relative, evidence-relative, and belief-relative permissibility are distinguished by Derek Parfit in On What Matters: Volume One (Oxford: Oxford University Press, 2011), 143. But see also Frank Jackson, ‘Decision-Theoretic Consequentialism and the Nearest and Dearest Objection’, Ethics 101 (3: 1991), 461–482. 29 See, e.g., Seth Lazar, ‘War’, The Stanford Encyclopedia of Philosophy (Summer 2016 Edition), edited by Edward N. Zalta, http://plato.stanford.edu/archives/sum2016/entries/war/, §4.5. 30 Ibid. 31 Braun and Brunstetter Megan Braun and Daniel Brunstetter, “Rethinking the Criterion for Assessing CIA-Targeted Killings: Drones, Proportionality and Jus ad Vim,” Journal of Military Ethics, 12, no. 4 (2013), pp. 304 – 324, 317. 32 Fabre, ‘War Exit’, p. 632. In an exchange between myself and Fabre on Pea Soup, she concedes that it may be a mistake to talk of the categories as useful. See http://peasoup.typepad.com/peasoup/2015/05/ethics-discussions-at-pea-soupcecile-fabres-war-exit-with-critical-precis-by-helen-frowe.html 33 See, e.g., McMahan, ‘Just Cause for War’, 2; Frowe, The Ethics of War and Peace: An Introduction, 2nd edn. (New Abington: Routledge, 2015), 54, 57; Cecile Fabre, ‘Guns, Food and Liability to Attack in War’, Ethics 120 (1: 2009), 36–63, at 49 and 57; Seth Lazar, ‘War’s Ending and the Structure of Just War Theory’ Forthcoming in Lazar and Frowe (eds.) Oxford Handbook of Ethics of War (New York: OUP) (forthcoming); Darrel Moelloendorf’s work on jus ex bello and David Rodin’s work on jus terminatio, e.g., ‘War Termination and the Liability of Soldiers for the Crime of Aggression’, in Jus Post Bellum: Towards a Law of Transition from Conflict to Peace, edited by Carsten Stahn and Jann Kleffner (T. M. C. Asser, 2008). 34 See, e.g., Lazar, ‘Necessity in War and Self-Defence’, Philosophy and Public Affairs, 40, 1 (2012), pp. 3 - 44 35 Lazar, this volume. 36 Ibid. 37 Oliver O’Donovan, The Just War Revisited, 13 38 For discussion of what I have elsewhere called mediated harms that are triggered (but not enabled) by one’s otherwise permissible behavior, see Frowe, Defensive Killing, chapter 5, and Rodin, ‘The Myth of National Defense’, in The Morality of Defensive War, edited by Cécile Fabre and Seth Lazar (Oxford: Oxford University Press, 2014).