Complicitous liability in war

Abstract

Jeff McMahan has argued against the moral equivalence of combatants (MEC) by developing a liability-based account of killing in warfare. On this account, a combatant is morally liable to be killed only if doing so is an effective means of reducing or eliminating an unjust threat to which that combatant is contributing. Since combatants fighting for a just cause generally do not contribute to unjust threats, they are not morally liable to be killed; thus MEC is mistaken. The problem, however, is that many unjust combatants contribute very little to the war in which they participate—often no more than the typical civilian. Thus either the typical civilian is morally liable to be killed, or many unjust combatants are not morally liable to be killed. That is, the liability based account seems to force us to choose between a version of pacifism, and total war. Seth Lazar has called this “The Responsibility Dilemma”. But I will argue that we can salvage a liability-based account of war—one which rejects MEC—by grounding the moral liability of unjust combatants not only in their individual contributions but also in their complicit participation in that war. On this view, all enlistees, regardless of the degree to which they contribute to an unjust war, are complicitously liable to be killed if it is necessary to avert an unjust threat posed by their side. This collectivized liability based account I develop avoids the Responsibility Dilemma unlike individualized liability-based accounts of the sort developed by McMahan.

1 Introduction

For the past 15 years the moral equivalence of combatants (MEC) has been subjected to sustained and withering attacks. Jeff McMahan, in particular, argues that, for the most part, unjust combatants (those fighting in furtherance of an unjust cause) are not morally permitted to kill just combatants (those fighting in furtherance of a just cause).¹ Just combatants, however, for the most part, are permitted to kill unjust combatants. This is so, not because unjust combatants have consented to such an attack, nor because attacking them constitutes a lesser evil, but because unjust combatants have made themselves morally liable to be attacked—they have done something to lose their right not to be attacked. This account is thus known as a “liability-based” account of just war. Put roughly, unjust combatants have made themselves liable to be attacked by contributing to the unjust harms in virtue of which the cause for which they are fighting is unjust. Just combatants do not make themselves liable to be attacked by targeting unjust combatants, because engaging in necessary and proportionate defense against an unjustified threat does not make one morally liable to be attacked by that threat. One cannot lose one’s right not to be attacked by acting with justification (McMahan 2009, pp. 14, 38ff). Thus, even when unjust combatants confine their attacks to military targets, they kill individuals who are not liable to be killed; unjust combatants—unlike just combatants—cannot satisfy the requirement that they attack only legitimate targets. MEC, by imposing the same principles governing combat on both sides in a war, wrongly permits unjust combatant to attack illegitimate targets. MEC is, then, mistaken as a moral doctrine. The moral principles governing combat must distinguish between just and unjust combatants.

But this sort of account faces a dilemma posed by Seth Lazar (2010). He points out that many unjust combatants contribute very little to the war in which they participate. Despite the name “combatant”, a significant portion of combatants do not see combat. Many occupy support roles that are in aggregate indispensible to the war but which individually do not contribute substantially. And those who do participate in combat often do so ineffectively. Yet, according to McMahan, it is their contributions to the unjust war that make them liable to be attacked. Lazar argues that if such combatants are liable to be attacked in virtue of their contributions, then the threshold for liability to be attacked must be so low that many civilians on the unjust side must be liable as well, for their individually marginal contributions to the war. It seems, then, that the liability-based account fails to insulate civilians from liability to attack. This is a problem for the account—it seems to countenance terrorism. Lazar notes that one might ‘raise’ the threshold of liability to attack, thereby insulating civilians—but this would also insulate unjust combatants who do not pose direct threats and who contribute no more than a typical civilian does. The discrimination required to attack solely those combatants who contribute substantially, would pragmatically preclude waging a war with a just cause. We are then left with a version of contingent pacifism according to which modern wars, as they are actually fought, are unjust.²

McMahan’s account thus faces a dilemma—his account forces him to choose between countenancing total war, and requiring contingent pacifism. Avoiding both total war and contingent pacifism requires setting the bar for liability high enough to insulate civilians from attack, but low enough to avoid insulating a significant number of combatants from attack. Lazar argues that McMahan’s account rules out this middle-of-the-road view—he is stuck between the horns of contingent pacifism and total warfare. Calling this “the responsibility dilemma”, Lazar argues that it “seems sufficiently serious either to prompt reevaluation of McMahan’s critique of MEC, or to suggest that we need a quite different approach to the ethics of war than that offered either by McMahan or by Walzer” (Lazar 2010, p. 189). I hope to develop such an approach—one that vindicates the denial of MEC without succumbing to the responsibility dilemma.

I will argue in favor of a liability-based account of just war in which unjust combatants are collectively liable for the threats posed by other combatants on their side. The account I develop fundamentally differs from McMahan’s in what grounds the liability of unjust combatants. On this account, combatants (except for those whose participate in a war under duress) can be morally liable to be killed in virtue of participating as a combatant in an unjust war, even if that participation ultimately results neither in posing threats nor in contributing substantially to the war. To this end, I will develop a preliminary account of collective liability in which participating in certain sorts of cooperative projects makes each participant, regardless of how much she contributes, partially liable for what others foreseeably do in furtherance of the group’s goals. By grounding the liability of unjust combatants not only in their individual contributions to an unjust war but also in their complicit participation in that war, the collectivized liability-based account I develop avoids the responsibility dilemma—unlike McMahan’s individualized liability-based account. A more general strength of the proposed collectivized liability-based account is that it treats wars the way they are actually fought: as a fundamentally cooperative activity.³

In the next section I further explicate and defend the responsibility dilemma. In the sections following that, I develop a preliminary collectivized liability-based account of just war, after which I explicate how it vindicates MEC while avoiding the responsibility dilemma.

2 The responsibility dilemma

On McMahan’s individualized liability-based account, an agent is liable to be attacked just in case that agent is responsible, if only minimally, for significantly contributing to an unjust threat severe enough to justify the degree of preventive harm intentionally imposed on the agent as a necessary means of reducing or eliminating that unjust threat (McMahan 2008, p. 22; 2009, pp. 32–37). An agent is minimally responsible for a threat if she has made relevantly voluntary choices that foreseeably contribute to the resulting threat. Consequently, the vast majority of unjust combatants—including conscripts—are liable to be attacked. But, as mentioned in the introduction, it seems that this individualistic account cannot reconcile the claim that civilians on the unjust side are generally not liable to be attacked, with the claim that unjust combatants are generally liable to be attacked, because, (1) civilians and combatants are in general responsible for their contributions to an unjust war, and (2) a substantial portion of combatants individually contribute no more than many civilians contribute individually.

What evidence do we have for (2)? Based on post-combat interviews of soldiers, Brigadier-General S.L.A. Marshall concluded that only 15–25 percent of Allied riflemen in World War II who were in a position to fire their weapons at an exposed enemy soldier, did so. Though Marshall’s methodology has since been criticized (Spiller 1988, pp. 63–71), its general findings have been defended by Lt. Col. Dave Grossman in his book On Killing (which is on the Unites States Marine Corp. Commandant’s Required Reading List, and is required reading at the FBI Academy, the DEA Academy, West Point, the USAF Academy, and numerous Peace Studies programs). Specifically, Grossman notes that Marshall’s claims have been confirmed by numerous other studies (Grossman 2009, p. xvi).⁴

Call those combatants who do not contribute substantially to the war’s aims, “ineffective combatants”. The property of effectiveness is dispositional—a combatant need not be occurrently contributing at t in order for the combatant to be an effective combatant at t. For example, a sleeping combatant can be an effective combatant. (Combatants who are hors de combat—such as soldiers incapacitated by injury—are technically non-combatants and are thus not ineffective combatants). Whether a combatant is ineffective comes apart from whether she is competent—a competent combatant might be ineffective through no fault of her own. And an incompetent combatant might be effective despite her incompetence.

Can ineffective combatants be permissibly targeted? It seems that on McMahan’s account, ineffective combatants are not liable to be attacked since they (by definition) neither pose nor significantly promote any threat; they have done nothing to lose their right not to be attacked. On McMahan’s view, intentionally killing those who are not liable to be killed is permissible only if doing so is necessary to avert a substantially greater evil—otherwise the killing violates the constraint of proportionality.⁵ This is because, on McMahan’s account, killing non-liable persons intentionally is a much greater wrong than killing them merely foreseeably (McMahan 2009, pp. 20, 218ff). Thus, if ineffective combatants are not liable to be killed, and if both ineffective and effective combatants are intentionally targeted, then the scope of wars that can be morally fought is severely limited, since the vast majority of wars will violate the proportionality constraint.⁶

Because, on an individualized liability-based account, whether a civilian or a combatant is liable to be attacked depends on whether and how much she contributes to an unjust war, the gulf between combatants and civilians is fixed not by necessity, but by contingent factors. McMahan believes that these factors are constant enough to establish strong though defeasible injunctions against intentionally targeting civilians. But such a theory is confounded by the prevalence of ineffective combatants. As a result, an individualized liability-based account of war radically shrinks the gulf between civilians and combatants. We are forced to choose between an overly inclusive and an overly exclusive picture of who is liable to be killed in war. This is the responsibility dilemma.

But ultimately, the question of whether an individualized liability-based account can cope with the prevalence of ineffective combatants is merely revealing of what is fundamentally mistaken in such account. Even if combatants are rarely ineffective, it is mistaken, I think, to claim that a necessary basis of an unjust combatant’s liability is her contribution to the unjust war. In my view, any attempt to cast the net of liability by adverting to the largely contingent contributions of individual combatants will only approximate, at best, a more fundamental basis of their liability—namely, their complicitous participation in an unjust war.

3 The bank robbery and murder case

Many philosophers over the past century have been largely skeptical of collective liability. Perhaps the most widely-cited in the twentieth century is H. D. Lewis. He argued vehemently that an individual can only be morally responsible for what she causes, or for what she fails to prevent (Lewis 1948, p. 3). Lewis’s view seems to be shared by some philosophers writing on the ethics of war, including McMahan: “all doctrines of collective liability,” he writes, “are false” (McMahan 2010, p. 61).

I will argue that Lewis and McMahan are mistaken. An individual’s participation with others in a cooperative project can serve as a basis of complicitous liability for what the other participants do in furtherance of the projects aims, even if the complicitously liable individual’s participation contributes nothing to the achievement of the project’s aims. On this view, an ineffective participant in a cooperative project can be complicitously liable to be killed, provided that doing so averts substantially wrongful threats posed by her effective co-members in furtherance of the project’s unjust aims. Consequently, ineffective combatants are liable to be opportunistically killed.⁷ By joining and participating (however ineffectively) in the military, combatants can become liable for the foreseeable contributions made by other combatants in furtherance of the military’s aims.⁸ This account, I will argue, dissolves the responsibility dilemma, while vindicating McMahan’s reasons for rejecting MEC.

I will begin with an analogy—an example of a cooperative project consisting of a bank robbery—in which I argue that ineffective participants can, intuitively, be permissibly killed in order to prevent the threats posed by their effective cohorts. I will then ground this intuition by developing a preliminary account of complicitous liability, which is sensitive to circumstances endemic to warfare.

Suppose a criminal mastermind puts together a plan for robbing a bank. She hires five individuals, each of whom agrees to participate in the robbery. The recruits are made aware that part of the plan is to kill the witnesses in the bank. The mastermind does not physically participate in the robbery—instead, she provides the plan, the layout of the bank, the equipment, etc. One of the recruits, J, is stationed on a second floor balcony above the bank, as a look-out. Her role is the least important. The mastermind would have commenced with the plan even without a look-out. Suppose that J is not a very effective look-out—in fact, she falls asleep on the job. Fortunately for the robbers, J’s incompetence has no negative effect on the robbery, though her participation does not causally contribute to the robbery or murders either. The plan succeeds, and two witnesses are killed.

Though J causally contributed nothing, she bears some liability for the murder of the witnesses and the theft of the money. The claim is not simply that J is liable to retributive punishment, which isn’t subject to a condition of effectiveness. Rather, J is liable to be killed if it is necessary to avert the harms posed by the other participants. That is, J is liable to be opportunistically killed. It is intuitively permissible to kill J if it is necessary to stop the murders, even though J neither contributes to the murders, nor does anything to prevent others from stopping the murders. Suppose that, several blocks from the bank, a firearm and ham radio enthusiast learns the details of the impending robbery and murders when she accidentally overhears their walkie-talkie conversations on her radio. She hears on the radio that the robbers are only moments away from entering the bank, and that they plan on shooting as soon as they enter. She tries calling the bank, but receives only an automated answer. From her vantage point at the window, using binoculars, she sees J on the second floor, whom she confirms on the radio as one of the robbers. With no less violent means of stopping the impending murders, and with no other robbers to target, the enthusiast takes out her sniper rifle, and shoots J who has now dozed off against the balcony’s railing. She reasonably expects that the impact of the shot will send J over the railing, onto the street below, where the getaway driver is waiting, thereby forcing them to abort the robbery and murders. The attempt is successful and saves the lives of the would-be witnesses in the bank.

What the enthusiast did was intuitively permissible, despite that J neither causally contributed to, nor prevented anyone from stopping the robbery and murders. Clearly, the permission to kill J cannot be grounded in utilitarian reasoning—if J were an uninvolved bystander, killing her would be intuitively impermissible. Rather, it is because J is complicitously liable that she can be permissibly killed, if necessary to stop the murders.

The fact that J can be liable despite that she contributed nothing is reflected in the treatment of accomplice liability in Anglo-American criminal law. Ordinarily, an agent is criminally liable only if she has culpably committed or attempted to commit a criminal act. Waiting above a bank that one expects to be robbed is not itself criminally wrong, in that there are few jurisdictions in which this would support a conviction of attempted robbery. But an agent who has individually committed no criminal act can be at least partially liable for the criminal acts of others with whom she participates in furtherance of a joint criminal goal. On these grounds, J can be held criminally liable for murder and armed robbery, despite that she neither committed nor successfully contributed to either act. J’s criminal liability is grounded not in what she individually does, but in her intentional (albeit ineffective) participation in an armed robbery and murder.

It should be obvious that I take J’s role in the robbery and killings to be analogous to the role of an ineffective combatant in an unjust war. Virtually all combatants (effective or not) who agree to participate in the military by acting in accordance with roles designed to facilitate the achievement of the military’s aims, can be complicitously liable for what their fellow combatants foreseeably do in furtherance of the military’s aims. Ineffective unjust combatants are complicitously liable to be attacked if necessary to prevent the harms imposed by effective unjust combatants, for the same reason that J is liable to be attacked if necessary to stop the other robbers from killing. Insofar as killing ineffective combatants is unavoidable if just combatants are to engage in defense against effective unjust combatants, doing so is permissible, since ineffective combatants are complicitously liable to be opportunistically killed in furtherance of just aims.

McMahan has argued⁹ that an individualistic liability-based account might be able to accommodate J’s liability to be opportunistically killed. Such a view might ground her liability not in her complicitous (though ineffective) participation, but in her culpability. On this view, a culpable individual can be liable to be opportunistically killed, even if she makes no causal contribution to the harm which is averted by killing her.¹⁰ Thus, on this view, J can be permissible killed—though non-culpable ineffective unjust combatants will not be complicitously liable.

But if McMahan’s argument is correct, we can opportunistically kill an individual who is culpably but unsuccessfully attempting to commit a murder if it is necessary to avert a completely unrelated murder that is about to be committed by a completely unrelated individual somewhere else in the world. This seems strange. But I recognize that intuitions are not decisive on this point (or even with respect to whether J is liable to be opportunistically killed in the first place). For this reason, I will outline an account of complicitous liability which will provide grounds for thinking that ineffective participants in a cooperative project with unjust ends can be complicitously liable. The account I present will be synoptic—I cannot here develop a comprehensive account of individual responsibility for cooperatively committed harms. But the outline will demonstrate how an account of this type can salvage the argument against MEC by underwriting the complicitous liability of ineffective combatants.

On the proposed account, I am liable for the wrongs committed by other members of a cooperative project in which I intentionally participate, even if my participation fails to contribute causally to those wrongs. Call this the “Complicity Principle”. Clarifying and defending it will take two steps. First, I will explicate the concept of a ‘cooperative project’. Second, I will provide preliminary reasons for thinking that participants in a cooperative project can bear complicitous liability for the acts foreseeably committed by the other participants. In doing so, I will make the Complicity Principle more precise by saying more about the conditions that determine the degree to which participants bear complicitous liability.¹¹

4 The structure of cooperative projects

A cooperative project consists of individuals who share participatory intentions. A participatory intention is an intention to act according to a role, the function of which is to contribute to a cooperative act. Thus a participatory intention has an individual agent as its subject, and has a role as its object—and this role, in turn, has a cooperative act as its object. When individuals have participatory intentions with roles that have one and the same cooperative act as their objects, these individuals share participatory intentions. And a cooperative act is, in turn, an outcome consisting in or caused by the individual contributions of those who act according to shared participatory intentions.¹²

Acting on participatory intentions by acting in accordance with a role is to “do one’s part” in a cooperative project. Whether participatory intentions are shared—that is, whether we have roles promoting one and the same cooperative act—will depend on how the cooperative acts referenced in the objects of our roles are described. Suppose a squad of soldiers is participating in the rescue of a wounded POW. Two soldiers in the squad are carrying the wounded POW while six others are laying down suppressive fire. Each of the two soldiers assisting in carrying the wounded POW shares participatory intentions in that they intend to do their part in furtherance of one and the same cooperative act: carrying the wounded POW. Under this description, their participatory intentions are not shared with the six soldiers laying down suppressive fire, who, in turn, share participatory intentions of their own. As a result, the six laying down suppressive fire are participating in one cooperative project, and the two carrying the wounded soldier are participating in another. But all eight soldiers in the squad share the broader participatory intention of doing their parts in furtherance of rescuing the wounded POW. The token act of helping lay down suppressive fire manifests the broader shared participatory intention of doing one’s part in furtherance of the cooperative act of successfully rescuing the POW; mutatis mutandis for the token act of assisting in carrying the wounded POW. Thus the narrower cooperative project of laying down suppressive fire, and the narrower cooperative project of carrying the wounded POW, are both nested within the broader cooperative project of rescuing the POW. The entire squad is itself nested within the even broader cooperative project the members of which compose a platoon, in virtue of the more general participatory intentions that they all share. As we move outward concentrically, generalizing shared participatory intentions along the way, we can characterize each soldier in the squad as part of a cooperative project encompassing the nation’s army, and eventually the armed forces as a whole (and perhaps the state and the nation’s citizenry itself—a possibility which I will consider shortly).

To each soldier in the squad we can attribute partial authorship of the cooperative act in which rescuing the POW consists. That is, each soldier is an inclusive author of this cooperative act.¹³ Each soldier in the squad is also an inclusive author of the cooperative acts which the rescue of the POW consists in. Thus the soldiers laying down suppressive fire are inclusive authors of the cooperative act consisting in physically carrying the POW to safety; and the soldiers physically carrying the POW are inclusive authors of the cooperative act consisting in laying down suppressive fire. This is reflected in the fact that any soldier can appropriately say, without equivocation: “we laid down suppressive fire and carried the POW to safety”. We can ascribe inclusive authorship even to those soldiers whose attempted contributions were ineffective. This is because each soldier’s decision and subsequent attempt to act according to her role—which has the cooperative act of saving the POW as its object—relates her teleologically to that cooperative act. This is just to say that she has made it her function to help bring about this act; this functional relation, which survives a failure to contribute effectively, grounds the attribution of inclusive authorship. (I will say more about how and why this functional relation grounds inclusive authorship when I discuss complicitous liability among members of a cooperative project).

We can attribute inclusive authorship for rescuing the POW not only to the soldiers in the squad, but to any other members of any other cooperative project which shares with the squad the participatory intention of furthering the cooperative act of winning the war being fought. This is because rescuing the POW is part of the broader cooperative act of winning the war. More generally, when individuals cooperate by intentionally acting according to roles, and each role has one and the same object (O), each cooperator bears inclusive authorship for O. Each cooperator bears inclusive authorship in virtue of the fact that they have willingly taken on the function of assisting one another in furtherance of O. This means we can attribute inclusive authorship for rescuing the POW to all the combatants in the country’s armed forces, since rescuing the POW is partly constitutive of the shared object of their respective functional roles—i.e., the shared object of accomplishing the war’s aims. Even though the vast majority of the individuals who bear inclusive authorship for this cooperative act did not contribute to the rescue, any combatant can appropriately say “we rescued the soldiers” (or “we massacred civilians in rescuing the POW” if that is what the squad did).

Against the account of inclusive authorship of combatants, one might point out that combatants are often alienated from the aims of the war in which they fight. Put in the terms of the account I have outlined, combatants are often alienated from the objects of their roles—especially when the description of these objects is generalized. They might regard their contributions to the cooperative act merely as foreseen side-effects of some other goal they have—a goal for which acting according to their roles is a mere means. For example, a combatant might not care whether the ultimate aims of the war in which she is participating are achieved—she cares only about getting paid, or earning the respect of her peers, or accomplishing her mission, or simply ending her tour of duty. Or, to return to the previous example, a member of the squad rescuing the POW might identify strongly with the participatory intentions she shares with those in the squad, while failing to identify with the broader participatory intentions of those in her regiment, or with the navy, or with the country’s armed forces in general. But despite her alienation, such a combatant counts as bearing the participatory intention of achieving the war’s aims; to have this participatory intention, it is not necessary have an intention to contribute to war’s aim. It is only necessary to have an intention to act according to a role, where the function of this role is to contribute to the war’s aims.¹⁴ Thus a combatant can be largely alienated from the cooperative act to which she has the function of contributing, and yet still have the participatory intention of achieving the war’s aims, so long as she intends to do her part in the cooperative project—i.e., to act according to her role.

But what if, as is often the case, combatants have varying conceptions of what the aims of a war are? In a paradigmatic cooperative project, each participant is fully aware of who the other participants are, what their roles are in the cooperative project, and what cooperative act they are all contributing to. But large and complex cooperative activities are rarely like this. According to Michael Bratman, Raimo Tuomela, and Kaarlo Miller, individuals jointly commit an act only if each individual intends that the group commit that act. In small-scale, closely-knit cooperative projects in which plans are formulated jointly and in which each individual has a substantial influence on the cooperative act (which is precisely the sort of cooperative projects that these theorists tend to focus on) it might indeed be plausibly said that each participant intends the joint achievement of the cooperative act. But it is a mistake to analyze larger, more complex cooperative projects in this way. In projects where many participants have only marginal influence on the outcome, each participant plans and acts in ways directed toward the aim of fulfilling his or her own role, rather than towards the aim of achieving the cooperative act. This is because such a participant is not in a position to substantially influence whether the cooperative act is achieved. Indeed, such participants might have varying conceptions of the cooperative act in furtherance of which they are participating. Combatants, specifically, might be unaware of all (or even any) of the aims of the war in which they are participating. But individuals can share participatory intentions even given an indeterminate or mistaken description of the cooperative act serving as the object of their roles. Sharing participating intentions requires only that a candidate participant regard herself as doing her part in contributing to a cooperative act consisting in or caused by the individual contributions of the others participants—even if she is uncertain what is required of her, what the cooperative act is, and who the other participants are.¹⁵

Having outlined the basic structure of cooperative projects, I now turn to the Complicity Principle.

5 Complicitous liability in cooperative projects

According to the Complicity Principle, a participant in a cooperative project who does her part in furtherance of a cooperative act but whose participation fails to contribute causally to that cooperative act can nonetheless bear complicitous liability. The basis of her liability lies partially in her decision to do her part in the cooperative project by enacting a role in that project. As I noted, enacting a role in the cooperative project relates the participant teleologically to the cooperative act—the decision makes it her function to help bring about this act. Because she willingly took on this function, we can ascribe to her inclusive authorship of the function’s end—i.e., the cooperative act that the participants in the cooperative project commit together. And this ascription of inclusive authorship provides a basis for holding her accountable for the cooperative act, if the cooperative act is wrongful.

The account of cooperative projects I have outlined, then, provides a basis for the Complicity Principle. We have grounds for holding combatants complicitously liable for the military aims which they are tasked with promoting. Combatants who do their part in a war—that is, combatants who (effectively or not) act in accordance with roles designed to further the cooperative act of achieving the war’s aims—can be complicitously liable for that cooperative act, since such combatants are inclusive authors of that cooperative act. Combatants can bear complicitous liability even if they are ineffective since what grounds their complicitous liability is not their causal contribution to the cooperative act, but the functional relation of their role to that act.

But not all participants in a cooperative project will bear the same degree of complicitous liability. One factor determining how much complicitous liability a participant bears is the degree of inclusive authorship that can be properly attributed to the participant. And one factor determining the degree of inclusive authorship a participant bears is the type of role that the participant has in the cooperative project. All roles have the function, broadly conceived, of contributing to the cooperative act serving as its object. But some roles have the function of contributing far more than others. All things being equal, the greater the degree to which one is supposed to contribute to a cooperative act, the more prominent one’s role in that collective act. The more prominent one’s role in a cooperative act, the greater the degree of inclusive authorship that the participant bears. And since inclusive authorship is a basis of complicitous liability, the stronger the attribution of inclusive authorship, the greater the degree of complicitous liability that an individual will bear for the cooperative act. Thus, returning to the previous example, ineffective soldiers in the squad rescuing the POW will bear greater inclusive authorship for the rescue of the POW than will, for example, sailors aboard a minesweeping vessel—even if their respective contributions to the rescue of the POW are on a par. The soldiers in the squad bear greater inclusive authorship since they have roles that feature more prominently vis-à-vis the end of rescuing the POW.

This factor determining the degree to which a participant is complicitously liable helps dissolve a potential worry: that the account of complicitous liability that I have outlined might over-generalize by illicitly implicating civilians.¹⁶ McMahan, raising this sort of worry, writes that if combatants bear complicitous liability, then civilians will as well, since “the combatant is not more a part of the collective than the noncombatant” (McMahan 1994, p. 212). More specifically, one might worry that, on the analysis of cooperative projects that I have presented, “the armed forces” is part of a larger cooperative project, encompassing all the civilians engaged in war-planning, war-provisioning, and war-making. Moreover, one might point out that since the military is an element of the state, it is natural to think of the government as a cooperative project which includes the citizens as its members, if the citizens vote government officials into office, or if they pay taxes part of the purpose of which is to fund the military. This, again, suggest, that my account of complicitous liability over-generalizes by implicating a significant proportion of civilians.

But this worry is unfounded. The function of civilians, such as tax-payers, is to contribute in ways that have only a marginal impact, at best, on the war. So even if a typical tax-paying civilian, for example, ultimately contributes no more to an unjust war than an ineffective combatant does, the ineffective combatant can still bear substantially greater complicitous liability than the civilian. A combatant’s role, in a straight-forward sense, is designed to contribute to a degree far greater than the typical civilian’s. After all, a combatant’s successful contribution to an unjust war is enough to ground individual liability to be opportunistically killed—but not so for a typical civilian’s successful contribution to an unjust war. What underwrites the fact that an ineffective combatant will be more liable is not simply the fact that she would have contributed substantially if she had been an effective combatant. Rather, the degree to which she is supposed to contribute serves as a metric determining the relative prominence of the role she bears, which, in turn, partly determines her degree of inclusive authorship. The upshot is that even if both ineffective combatants and marginally contributing civilians are complicitously liable for the cooperative acts in which the pursuit of the war’s aims consist, we can still claim that the ineffective combatant but not the civilian is complicitously liable to be killed.

The distinction between those with roles designed to contribute substantially and those with roles designed to contribute marginally will not perfectly track the distinction between civilian and combatants. It is at least possible for there to be a class of combatants whose roles are designed to contribute so little to the war, that combatants of this sort would bear no more complicitous liability than the typical voting and tax-paying citizen. Certainly there are military personnel who do not pose direct threats—such as combat engineers engaged in explosive ordinance disposal, convoy soldiers driving military supply trucks, boom operators responsible for refueling aircraft mid-flight, technicians aboard navy combat ships, Special Operations Forces engaged in reconnaissance, combat division military police charged with manning traffic control checkpoints, etc. But a combatant occupying a subsidiary role of this type contributes to a degree far greater than the average tax-paying civilian (as is evidenced by the fact that the finances required to train and fund a combat engineer, a boom operator, a navy technician, etc. far outstrips the funds garnered by the government from the average tax-paying citizen).

If the account I have presented re-draws the line between liable and non-liable targets of attack in an unjust war, it is likely to do so by implicating certain types of civilians. Military scientists and high-ranking government officials, for example, bear roles designed to contribute more than that of the typical combatant. As a result, they can indeed be complicitously liable to be opportunistically killed, if the war being fought is unjust. But this hardly entails total war.

The upshot is this: ineffective combatants contribute to an unjust war no more than the typical civilian—but unlike such civilians, ineffective combatants have roles which, when enacted effectively, contribute far more than the typical civilian. By taking on a prominent role of this type, a combatant willingly makes it her function to contribute substantially to the war—and this function underwrites a degree of inclusive authorship and thus a degree of complicitous liability significantly greater than that of the typical civilian.

But what if a combatant is non-culpably mistaken about what the war’s aims are? In many cooperative projects, a non-culpable mistake regarding the nature of the cooperative act will mitigate or eliminate the participant’s complicitous liability for the cooperative act. In such cases they are not to blame for taking on the function of contributing to the wrongful cooperative act. But the complicitous liability of combatants who are non-culpably mistaken about the unjust aims of the war is not mitigated—at least not the extent that would make them non-liable to opportunistic killing. This is because any minimally responsible combatant is in a position to know that participating in any war is a morally risky activity, in the sense that war always involves imposing prima facie wrongs—e.g., killing, destruction, etc. They are in a position to recognize that even if the available evidence suggests the war is just, there is still a chance that the war is unjust, in which case the acts committed in furtherance of the war’s aims constitute egregious moral wrongs. By participating in a war—even one that seems just—she is participating in a morally risky activity. If the risk actualizes, then she is morally liable to be attacked, in virtue of having courted this risk.

A consequence of this view is that unjust combatants are liable to be killed even if they are in a situation epistemically equivalent to that of the just combatant on the other side. This might seem problematic, since there must be a morally relevant difference grounding the difference in liability between the combatants on both sides. But there is a morally relevance difference: those on the unjust side are actually complicitous in imposing unjust threats. A difference in moral luck can constitute a morally relevant difference by serving as a basis of liability for a wrongful threat when those posing the wrongful threat willingly risked doing so. The decision to gamble with the lives of innocents can serve as a basis of liability if the gamble is lost. The upshot is that ineffective unjust combatants can be liable to be opportunistically attacked even if they are non-culpably mistaken in their belief that the war is just, since such combatants are in a position to recognize that participating in a war is to participate in a morally risky cooperative project.

The argument for the Complicity Principle in general—and for the specific claim that ineffective combatants are complicitously liable to be opportunistically attacked—can be summarized as follows. Ineffective unjust combatants can be complicitously liable to be opportunistically killed despite that their causal contributions are minimal since they bear substantial inclusive authorship for the unjust aims serving as the shared object of their respective functional roles. Their shared participatory intentions—i.e., their intentions to act according to roles that have as their respective objects one and the same cooperative act—serve as a basis for their inclusive authorship. This is because each combatant, by taking on the role of a combatant, makes it her function to contribute to the war’s aims—by taking on this function, we can attribute to her inclusive authorship of the function’s end. The inclusive authorship and thus the complicitous liability that each combatant potentially bears (unlike the complicitously liability that civilians potentially bear) is substantial—enough to ground a permission to kill the combatant if it is necessary to prevent her cohorts from imposing unjust threats—since it is her function to contribute substantially to the war’s aims.

6 Wars with just and unjust aims

So far, I have been arguing as if all wars have either only just aims or only unjust aims. But in many wars each side will have at least some unjust aims (even if they are subsidiary to the achievement of a larger, just aim). Call this a “morally heterogeneous war”. If combatants are complicitously liable for the unjust aims pursued by their effective cohorts, then it seems the presence of any unjust aim is enough to make all the combatants on that side complicitously liable to be opportunistically killed. Thus it might seem that MEC resurfaces in any conflict in which both sides pursue at least one unjust aim. But this is mistaken.

A combatant liable to be killed can be permissibly killed only if doing so satisfies the constraint of proportionality—that is, only if doing so is not disproportionate in the harm it directly or indirectly inflicts on those who are not at all liable to be harmed. If killing a combatant helps forestall the achievement of both unjust and just aims, then we bear some responsibility for having indirectly contributed to forestalling the war’s just aims. If these just aims are significantly more numerous or morally important than the unjust aims, then killing the combatant can be impermissible, despite that she is liable to be killed.

Still, one might argue that if the permissibility of killing a combatant depends crucially on comparing the effects that killing the combatant will have on forestalling the just aims versus the unjust aims of the war in which she fights, then we are again left with an account in which the basis of permissible killing is highly contingent. The worry is that, on this account, discriminating between permissible and impermissible targets in morally heterogeneous wars is pragmatically impossible.

But this is mistaken—on the account I have outlined, discriminating between permissible and impermissible targets is not pragmatically unrealistic. After all, one way to discriminate between permissible and impermissible targets in a morally heterogeneous war is to abandon or eliminate the war’s unjust aims. Consider the following example. Suppose our adversary is waging a war with both just and unjust aims. We are, in turn, resisting our adversary’s aims. Thus we too are waging a morally heterogeneous war: resisting our adversary’s unjust aims is just, and resisting our adversary’s just aims is unjust. Because our adversary is pursuing both just and unjust aims, it is pragmatically very difficult for us to discriminate between enemy combatants who can and cannot be permissibly killed. But suppose that we cease resisting our adversary’s just aims—which, presumably, we well should. That is, we allow them to achieve their just aims. Since our adversary is now pursuing only unjust aims by means of war, we no longer have the problem of discriminating between enemy combatants who can and cannot be permissibly killed; neither we nor our adversary is waging a morally heterogeneous war. By ceasing our pursuit of the unjust aims, we affect the moral status of the war our adversary is waging, by shifting it from a morally heterogeneous war to a purely unjust war, thus dissolving the problem of discrimination.

The upshot is that we do have a way to discriminate between permissible and impermissible targets in certain morally heterogeneous wars—namely, by ceasing pursuit of the war’s unjust aims. This comports with a pre-existing obligation: in general, we already have a very strong reason to eliminate or abandon the pursuit of unjust aims in morally heterogeneous wars. Now we have even greater reason: in waging a war with an unjust aim, we will have difficulty discriminating between permissible and impermissible targets.

Still, one might argue that even the best wars will inevitably include some unjust subsidiary aims (i.e., means to the achievement of larger aims). They crop up, despite the best efforts of military and civilian leaders. But if we are fighting in an overwhelmingly just war with only a few minor unjust aims, it is very unlikely that a substantial portion of our attacks in that war will violate the proportionality constraint. Hence, it is very unlikely that a substantial portion of the enemy’s attacks are made just by the fact that the war we are waging includes minor unjust aims. Consequently, the violations of proportionality resulting from pursuing a minor unjust aim in an overwhelmingly just war are very unlikely to either a) require the party waging the overwhelmingly just war to cease waging the war altogether or b) undergird an equal permission for both side of the conflict sides to kill.

But what of a conflict in which both sides wage wars rife with just and unjust aims—and which refuse to abandon their respective unjust aims? On my account, will combatants on both sides have an equal permission to kill? Depending on the details (which I cannot address here), MEC will indeed be applicable to wars of this type. The upshot is that though all enlistees in a morally heterogeneous war are liable for that war’s unjust aims, killing such combatants will still be impermissible if doing so violates the proportionality constraint by indirectly forestalling the achievement of sufficiently important just aims. Thus the account I have developed does not force us to conclude that if a war is morally heterogeneous then MEC applies in that war—though there will be some morally heterogeneous wars in which MEC applies.

7 The problem of conscription

A combatant conscripted in an unjust war on pain of prolonged imprisonment or worse will not be complicitously liable to be killed. Since her participatory intention to act in accordance with her role as a combatant is coerced, she is not to blame for her inclusive authorship of the cooperative act in which achieving the war’s aims consists. Ineffective conscripts of this sort are thus not liable to be killed, since they neither contribute substantially to the war, nor do they bear complicitous liability. (Unjust effective conscripts who participate under duress would still liable to be attacked, since their liability is grounded not merely in the decision to participate, but in their minimal responsibility for the unjust threats they pose).

Whether a coercive threat undermines complicitous liability depends on the severity of that threat and the wrongfulness of participation in the cooperative project. Since participation in furtherance of an unjust war is typically morally egregious, a mild threat will not defeat complicitous liability. A fine, or a short prison sentence, is presumably not enough. However, given the harsh treatment to which some governments subject those who attempt to evade conscription during wartime, many unjust, ineffective conscripts will not be complicitously liable.

To admit that many unjust ineffective conscripts are not liable to be attacked does not resurrect the responsibility dilemma—we are not forced to choose between contingent pacifism or total war. This is because all actual wars include many ineffective combatants, but not all actual wars include many ineffective combatants whose participation is compelled under threats of prolonged imprisonment or death. Still, the participation of these sorts of combatants is all too common. As a result, the scope of wars that can be permissibly fought—if a collectivized liability-based account of war is correct—is narrower than what McMahan thinks it is, but not as narrow as Lazar thinks it is. A collectivized liability-based account of war carves out a middle path.

An upshot is that whether a war violates proportionality by resulting in the deaths of too many non-liable individuals—including many unjust, ineffective, conscripted combatants who participate under duress—will, to some extent, be determined by ‘facts on the ground’, such as how many conscripts there are, what coercive threats are leveled against them, and whether those threats are avoidable (by paying bribes, for example). But the best account of the morality of war ought to be sensitive to such facts. Though the object of complicitous liability—the cooperative act—might be an event out of the control of the liable agent, the basis of complicitous liability—her participation in the cooperative project—must lie in an event that is relevantly in the agent’s control: her voluntary entry into (or exit from) the cooperative project.

8 Conclusion and remaining issues

The account I have outlined of complicitous liability in war is, I believe, an improvement upon McMahan’s account in the following way. On McMahan’s account an individual’s liability in an unjust war is determined largely by the contributions she makes to that war. The trouble with such an account is that we cannot expect this highly contingent basis of liability to warrant categorizing liable and non-liable targets in such a way that unjust combatants systematically fall in the liable category and civilians systematically fall in the non-liable category. The account I develop does a better job, I believe, of grounding liability in a way that tracks the combatant/noncombatant distinction. Specifically, I have argued that one of the criteria grounding the permissibility of killing unjust combatants—the criterion of liability—can be satisfied systematically by all voluntarily participating combatants in an unjust war.

But as I argued in Sect. 6, this is not to say that all voluntarily participating combatants in an unjust war can be permissibly killed. Specifically, in a morally heterogeneous war, killing liable combatants might violate the proportionality constraint by indirectly forestalling the achievement of just aims. In such wars, a basis of permissibly killing enemy combatants is indeed contingent in a way that prohibits systematically categorizing all combatants in a morally heterogeneous war as a permissible target of attack. But at one remove, this is precisely the sort of upshot we ought to expect—it would be strange if we could systematically categorize all combatants in a morally heterogeneous war as permissible targets of attack. An upshot is that there will likely be some morally heterogeneous wars in which MEC applies.

Many questions remain. To what extent can a combatant be liable for the war-crimes committed by other combatants who acted without the authorization of military leaders? Will an unjust combatant who joins with the intention to sabotage her own side bear complicitous liability? What are the conditions under which a combatant participating in a war rife with just and unjust aims can be permissibly attacked? These are only some relevant questions which I have not addressed here—and which a collectivized account of liability in war must ultimately address. But doing so is beyond the scope of this paper.

My intention, rather, has been to salvage a liability-based account of war—one which rejects the MEC—by developing a collectivized liability-based account of war. More generally, I hope to have taken a first step in demonstrating how a collectivized liability-based account of just war can accommodate the ethical implications of both the cooperative nature of warfare and the highly attenuated nature of the contributions that constitute participation in warfare.