The Undue Moralization of Individual Conduct in War

In this section we argue that two of the most basic, evidently unchanging characteristics of war make impossible the full moral assessment of individual conduct in war and therefore block the moral individualization of the rules that should guide the fighting. First, war is a confrontation ultimately between states or other collective entities. As a result, few individuals will be liable entirely on the basis of their own behavior to the treatment they are subjected to as adversaries during this confrontation. Adversaries deal with each other primarily as representatives of their respective collectives. As Rousseau famously phrased it, “War, then, is not a relation between men, but between states; in war individuals are enemies wholly by chance, not as men, not even as citizens, but only as soldiers.”Footnote ⁶ Second, a significant number of individuals fighting in war will be genuinely unsure or mistaken about their own justification for fighting and will be unavoidably ignorant of the moral justifications of the various individuals they encounter on the battlefield. It follows that the moral calibration of individual conduct during combat operations on the basis of the rights of individual adversaries is not an achievable goal. Any set of rules that pretends to possess the virtue of conformity with an individual rights–based morality in fact permits the harming of individuals that are not (fully) liable to that fate. Individualized rules for the conduct of war are hypocritical and/or unworkable.

Why would a rights-based morality be unable to anchor rules for the conduct of individuals in war? After all, morality is perfectly capable of regulating ordinary physical confrontations between individuals. Though physical confrontations are no longer conceived of as providing a legitimate mechanism for generating principled distributions of goods and ills, morality provides rules for just or at least justifiable conduct within them. These rules generally rest on the differentiation between aggressor and defender. Morality devises rules for self-defense based on the notion that we can in most cases tell who initially threatened another's rights, or alternatively who is responsible for creating a forced choice situation that threatens individual rights. Morality, then, works with the resulting distribution of liability among the individuals involved when prescribing rules for conduct.

War does not treat the human being as an end in itself, even when it is geared toward the overall best preservation of individual rights, such as during the rescue of a people subjected to genocide. The individual is at the disposition of the state, traditionally the only legitimate belligerent and the subject of international legal regulation. Nowadays, other collective entities, such as national liberation movements, ethnic factions, and rival political regimes, engage in war at least as often as states. The point here is not to establish what kind of entity can be or tends to be a belligerent, and our thesis is not that a belligerent ought to be a certain type of collective entity. We make the purely descriptive observation that war is a confrontation between collective entities that is manifest partly in many physical confrontations between and among individuals, and point out that these latter confrontations take on their meaning, and at least some of their moral status, as part of the physical confrontation between the collective entities.

If we seek to regulate war by analogy to the approach of morality to conflict between individuals, so that the rules for permissible conduct attach to the moral status of each collective entity, and then assess which entity is responsible for the emergence of the confrontation, can we not still extrapolate to what each individual combatant ought to do during combat operations? McMahan's challenge to the conventional view has fallen on such fertile ground precisely because he suggested making just this analogy. He proposed, first, to differentiate between belligerent collective entities in the way we differentiate between individuals in an ordinary physical confrontation, according to which side is responsible for the confrontation having arisen in the first place; and, second, to treat individuals in accordance with their specific connection to the endeavor of their own belligerent entity—namely, the contribution to the threat posed by the collective entity for which they are responsible.

McMahan acknowledges that it is impossible to ensure a precise fit between the liability of individuals for contributing to an unjustified collective threat and the harm that is inflicted on them during war.Footnote ⁷ He maintains that the best available rules for conduct in war, hence, permit the killing and wounding only of those individuals whose liability is above a certain threshold and who fight for the unjustified collective. In this reading, what individuals are allowed to do in war depends at least to a certain extent on their individual moral choices, and the harm inflicted on individuals in war is partially morally accounted for by their own conduct.

We suggest, however, that McMahan's connections between behavior and the resultant wounds or death are loose. First, significant obstacles exist to merely identifying those individual adversaries liable above a certain threshold. These obstacles are extensively discussed elsewhere and would, according to several commentators, prevent law from systematically implementing McMahan's morally prescribed rules.Footnote ⁸ Second, even if the distribution of intentional harm in battle could somehow be restricted to only those who are liable above a certain threshold, the loss of life or the particular wounds actually sustained would often be excessive to any wrong committed by that specific individual. Which fault in the decision to participate in a war makes it morally appropriate that one returns home a quadriplegic? How many combatants behave during war in a way that calls for their bleeding to death from a grenade?Footnote ⁹ While much of the appeal of McMahan's individual liability approach rests on the claim that it would give the individual her moral due, it largely fails in exactly that endeavor if applied in an undifferentiated way to everyone belonging to one of the two adversaries whose liability is above a single threshold. The assessment of liability is binary, but the types of wounds and kinds of deaths suffered by the liable (and the non-liable) range indefinitely from the minor to the horrific. As a result, the diverse fates met by combatants in war cannot be shown to be individually morally justified.Footnote ¹⁰

Walzer portrays fighters on both sides primarily as equally victims.Footnote ¹¹ It is their shared victimhood, above all, that makes it appropriate that we should have the same expectations for them all. They have an equal right to kill because they have all been victimized by being placed in the circumstances they are in, and so opponents are entitled to defend themselves in those circumstances by (if necessary) killing or wounding each other. In addition, Walzer claims that the combatant consents to being “made into a dangerous man” and thus loses his immunity from attack.Footnote ¹² If we accept Walzer's thesis of the moral equality of combatants on both sides, the question remains what justifies the radically different treatment of combatants compared to civilians. Moreover, even if, given their victimization, one granted combatants a general basis for defending themselves, it would not follow that they were free specifically to kill or severely wound the equally victimized combatants on the opposing side.

How is further reciprocal victimization the morally appropriate response to the initial parallel victimizations? By the same token, even if we thought that the reduced vulnerability resulting from their becoming “dangerous men” justified combatants bearing considerably more of the harm committed during war than their generally more vulnerable civilian compatriots, how is it morally reasonable that some individuals, but not others, suffer death or grievous wounds when all consented to increase their threat potential?Footnote ¹³ After all, the decision to be made into a dangerous man may often have been coerced, excusably ill-informed, or sincerely misguided.

In sum, the kinds of harms regularly suffered by individuals in war are morally inappropriate for many of the kinds of faults typically committed by individuals in choosing to participate in war. Claims to the contrary amount to an undue moralization of individual conduct in war. Neither conventional just war theorists nor their revisionist counterparts have provided adequate grounds for thinking that either all participation (Walzer) or wrongful participation (McMahan) should render a combatant liable specifically to the risks of death or wounds in battle. To a certain extent, the argument advanced here echoes the revisionist literature that explains the disproportionate harming of many individuals in war as a distribution problem: epistemic obstacles to knowing who responsibly contributes what to the collective threat prevent us from attuning harm to liability. This distribution problem is real and significant. Human individuals are morally responsible for the genuine choices they make. As a result, they vary greatly in how they ought to be treated, beyond equal respect for their basic rights. One of the true horrors of war is precisely that individuals encounter fates that are utterly disconnected from their individual moral status. We have no reason to believe that what the vagaries of combat inflict on an individual is even approximately the harm that is her appropriate share.

Beyond this familiar point, we also suggest that this distribution problem is incurable because the moral rules in war cannot work with individuals' choices as the sole reference point for determining their treatment and permitted conduct. The initial wrong, from which flows the moral assessment of individual acts in war, is the act of a collective entity. It is not the individual's moral status as one natural person and the merit of her acts that fully and definitively determine her fate in war, but also her moral status as a member of a collective entity and her acts as part of a collective endeavor. It is not merely the extent of her liability for threatening another individual, but the extent of her liability for partaking in a much larger threat to the opposing collective that a normative framework for the just allocation of harms would have to grapple with. Contra McMahan, judgments of liability in war are prima facie collective. That is partially why in war many individuals necessarily encounter harm that is disproportionate to their own morally wrong choices.

That war is a confrontation between collective entities not only means that morality needs to navigate two levels of agency and, in the process, loses any exclusive distributive focus at the individual level. The confrontation between collective entities, we submit, creates evil that is unaccounted for at the level of individual combatants. If, per impossible, all individual combatants bore exactly the harm that is due to them given their responsible contribution to the threat of their collective entity, we hypothesize that not all harm inflicted in war would be divvied up. The harm generated in a physical confrontation between collective entities is larger than the sum total of individual moral transgressions by combatants. In this reading, the observation that many individuals in war are harmed far beyond what they appear to be liable for individually is explained by the fact that individuals must bear the residual harm that stems from being bound up with a collective entity.Footnote ¹⁴

The impossibility of giving each individual her moral due and, hence, the large-scale infringement of individual rights is a noncontingent feature of combat operations. In turn, useable guidance for conduct in war could not be derived directly from knowing what exactly combatants have rights to or would somehow be morally liable to endure. Rules that attempt to hinge the permission to attack on the moral status of the attacked individual inevitably either provide a permission to infringe another's rights or would need to be so complex that they do not guide the combatant's actions at all. For most combatants in most wars, following a strict individual rights–based morality would require them to stop fighting. That is what we mean by saying that the character of war is ultimately a physical confrontation between collective entities, which prevents the moral individualization of the rules that guide it while it is under way.

If rules for conduct in war cannot allocate life and death, physical soundness and maiming wounds, systematically among individuals according to any normative principle that attaches to them qua individuals, should the rules not at least be able to treat individuals differentially according to whether they form part of a justified or an unjustified collective enterprise? This is the other major aspect of McMahan's challenge to the just war orthodoxy. Besides claiming to tailor the moral guidance of the rules to individuals, he proposes to make them asymmetrical between the embattled collectives. We have explained why the individualization seems noncontingently impossible. For us, asymmetry would have to mean that we allow all individuals on one side to fight and prescribe that all individuals on the other side hold still, as morality would analogously require in an ordinary physical confrontation.

However, in an ordinary physical confrontation, even though the facts might in the moment be contested, we do not generally doubt that a third-person perspective is available from which it is possible to tell how liability for attack and rights forfeited or retained are distributed between the two individuals fighting. In certain cases national courts retroactively adopt this third-person perspective. Regarding war, it seems, no third-person perspective is readily to be had. First and foremost, this is because no court has competence to interpret the facts authoritatively. If, for instance, the International Court of Justice (ICJ) were consulted, its ruling would be unlikely to be rendered until several years after the conflict began. Is this a contingent difficulty? Is all we would need, then, in order to have asymmetrical rules for conduct in war an effective, reliable, and prompt mechanism of adjudication about just resort?

Even if the competences of the ICJ could be expanded and its workings reformed, under the current international order, premised now on the sovereign equality of states, the individual is, for better or for worse, much more immediately subject to the authority of the state and the socialization inculcated by national institutions and media. He or she is therefore much less likely to consider an international court's ruling the valid third-person perspective to adopt.Footnote ¹⁵ We suspect that even if we had an international court for the adjudication of just resort, the likely recipients of rules about conduct in war and, so to speak, the law's target audience would include many combatants who are unclear about the justice of their cause; who mistakenly believe that they are fighting on the basis of a just cause; or who, even if they are able to see the lack of a just cause from a valid third-person perspective, such as an international court, fight anyway, because the anticipated consequences of disobeying the state are much more immediate than those of disobeying an international court.

It is crucial to stress that we do not claim that all such combatants are morally justified or even excused for being unclear, mistaken, or persistent in fighting. But the fact that wars are being fought in part by combatants who belong in one of these three categories creates an imperative for rules that do not depend on recognition of where objective justice, even at the level of collective entities, lies. While a normative framework for guiding conduct in war cannot be morally individualized because of the noncontingent features of war as a physical confrontation between collective entities that individuals only represent, it contingently cannot be asymmetrical, given the current institutional setup of the international system. These are the constraints, contingent and noncontingent, within which a workable normative framework for conduct in war has to operate.

While conventional just war theory puts forward rules that are individualized to only a small degree and completely symmetrical, the proposal for distributing the harm that is intentionally and inevitably inflicted in war—put forward, for instance, by Walzer—draws on differences in individual moral status: the equal consent to increased threat through becoming dangerous and the equal victimization that both distinguish combatants from noncombatants. Though the underlying attempt to track individual rights is rather unpretentious, this explanation and justification for the permissibility of killing combatants in war presupposes that moral individualization is what a normative framework for conduct in war should ultimately strive for. Like McMahan's liability threshold, noncombatant immunity justified by a combatant's choice to be made into a dangerous man significantly fails to give the individual his specific moral due, effectively treating all combatants collectively, all the while upholding the pretense of echoing (if faintly) the terms of individual rights–based morality.

Military Necessity as a Restraining Rule: The St. Petersburg Assumption

If systematically protecting individual rights is not an achievable purpose for rules during the conduct of war, what can we aim for instead? What about general “moral damage limitation”: not straightforwardly doing what is right and avoiding what is wrong (killing only individuals liable to that fate), but avoiding as much wrong as possible in the circumstances? Crucially, that cannot in practice mean avoiding all unjustified killing, given the distribution problem discussed and the consequent impossibility of morally individualizing the rules. In any case, even if an act of killing can ultimately be justified, it is prima facie morally problematic. It follows that when matters have reached the stage of a physical confrontation between collective entities, where rules can no longer systematically guide participants in distinguishing between justified and unjustified killing, the remaining moral achievement available is to limit all killing as much as possible.

Rules for conduct in war “permit” killing only in the minimal sense that they do not prohibit some killing. Many of the killings that will not be prohibited will be morally wrong. Yet, these rules need not be in the business of blessing what they do not prohibit: they simply do not prohibit, as such, morally wrongful killing and wounding. By not prohibiting some wrongful killing they can effectively prohibit even greater evils—namely, unlimited (including much wrongful) killing. To what degree can rules effectively limit killing in war? Only to the extent that they do not prohibit war as such. This alternative pragmatic goal for rules in war—limiting killing in war as much as possible—hence means allowing only killing that is militarily necessary and, indeed, only some of it.

The spirit of the international laws regulating conduct in war, the laws of armed conflict (LOAC), has long been informed by this pragmatism. The law states its purpose to be “alleviating as much as possible the calamities of war,”Footnote ¹⁶ given that the purpose of going to war is to prevail militarily. Hence, as a matter of law nothing is permitted that is not necessary for military success. Being militarily necessary is a logically necessary condition of legal permissibility. However, authoritative legal commentators write about these matters in ways that are unhelpful. For instance, the 1987 Commentary on the Additional Protocols of 1977 posits necessity to be “the limit of legality” as “any violence which exceeds the minimum that is necessary is unlawful.” The commentary then goes on to render matters circular by claiming that “it is on this principle [necessity] that all law relating to the conduct of hostilities is ultimately founded.”Footnote ¹⁷ This suggests that all prohibitions under the laws of war stem from a lack of necessity. But necessity is not the only restriction. The principle of proportionality, for instance, prohibits conduct that, even though it might be necessary, can be expected to cause unintended but foreseeable civilian damage that is excessive.

By the same token, we reject the position that everything that is necessary is also legal.Footnote ¹⁸ We propose to uphold necessity and legality as logically independent requirements that are nevertheless connected. Military action is fully permissible legally only if it is both militarily necessary and not otherwise in violation of international law. Military necessity cannot override international law and is itself bounded by independent considerations of legality, as has recently been reaffirmed in the authoritative British Manual of the Law of Armed Conflict: “Since military necessity permits the use of force only if it is ‘not otherwise prohibited by the law of armed conflict’, necessity cannot excuse a departure from that law.”Footnote ¹⁹ Historically, “military necessity” has been interpreted in profoundly pernicious ways.Footnote ²⁰ The second cumulative criterion of “not otherwise prohibited by the law of armed conflict” provides a bulwark against such interpretations. In modern LOAC, military necessity serves only as a limit on the use of force, not a justifying ground for exceptions.

Can military necessity ground rules for combat operations that guide individuals' conduct and, in fact, limit killing in war? After all, even if we exclude otherwise explicitly prohibited acts, such as killing protected medical personnel or inflicting excessive collateral damage, what attacks belligerents consider necessary seems to be open to interpretation, so that the criterion permits attacks on a wide range of objects and persons. Everything depends on knowing more about how “military necessity” is to be interpreted, specifically with regard to which ultimate goal the necessity of means is to be determined.

We submit that in international law “military necessity” has one corollary that provides it with moral significance even though (and perhaps partly because) it is not formulated in distinctively moral terminology. It does so by specifying the frame of reference for the determination of what is militarily necessary. We shall refer to it as the St. Petersburg assumption. The St. Petersburg Declaration of 1868 was the first formal treaty of international law banning a specific category of weapons (explosive projectiles weighing under 400 grams) because it was considered to inflict unnecessary suffering on combatants. Drafted at the behest of the Russian government by an international military commission consisting exclusively of military men,Footnote ²¹ the declaration provided the precedent for many treaties to follow, and it preceded by more than thirty years the attempts by the international community to codify general rules in the Hague for the conduct of hostilities.

The operative core of the LOAC is specified in the preamble to the declaration: “The only legitimate object which States should endeavour to accomplish during war is to weaken the military forces of the enemy.”Footnote ²² The preamble thus contains the assumption that often an “enemy can be overcome by sufficiently weakening its military forces. Once its military forces are neutralized, even the politically, psychologically or economically strongest enemy can no longer resist.”Footnote ²³ The criterion of military necessity hence appeals only to the goal of overcoming the enemy militarily. Since military success can often be attained by weakening the military forces on the other side, military necessity does not require allowing any other means—even other means, if any existed, that would in fact also be effective or even more efficient.Footnote ²⁴ It is necessary to allow only one kind of means that can be sufficient.

To “weaken the military forces of the enemy” contains an implicit “directly.” Interpreted in accord with the St. Petersburg assumption, military necessity never requires weakening the military forces of the enemy indirectly, because it can possibly instead be done directly.Footnote ²⁵ If indirect measures were allowed, such as attacking civilians in order to break their morale and thereby undermine their support for their military forces, the principle of military necessity would be largely meaningless, even permitting, for instance, World War II–style city busting. Clearly, the point is that the military forces on one side are to engage the military forces on the other—directly.

Since the rules are intended to limit the conduct of both sides to a war, they have to avoid prescribing behavior based on judgments about the justifiability of the resort to particular wars by particular parties. The normative framework must be completely agnostic concerning what force is being used for, morally or politically. As a result, the necessity to which it appeals can only extend to actions necessary to achieve a generic military victory. The difference between doing what is necessary to achieve military victory and doing what it takes, for instance, to achieve the political goals of a war is that the former restricts the range of possible actions to weakening the opposing military forces with the goal of overcoming them, even if that endeavor will not ultimately lead to achievement of the political goal of the war (for example, leverage at the negotiation table or regime change).

If one could do whatever is necessary to achieve any goal of a war, then military necessity would have relatively little constraining bite. Accordingly, in order to exercise any morally significant restraint the principle of military necessity needs the St. Petersburg assumption that accompanied it in 1868 to limit acceptable goals. The restriction to a single kind of action—namely, weakening the opposing military forces—has profound moral significance in spite of not being phrased in explicitly moral categories.

The fundamental principle of discrimination currently in force in international law can be interpreted as hinging on the principle of necessity combined with the St. Petersburg assumption. The principle of discrimination creates a class of people whom it is permissible to target because they are functionally connected to achieving a military victory: combatants. It is necessary to engage them for a possibility of military victory to exist; to that end it is not necessary, by contrast, to engage civilians. After all, civilians are not allowed to directly participate in war, the competition between two militaries. Their engagement can hence be presumed unnecessary for military victory. Each belligerent designates a class of people that is allowed or forced to carry out the collective armed confrontation. For this confrontation to go forward and ultimately be decided, it is necessary to engage those permitted to directly participate in it, but no one else.

If we ground and interpret discrimination between combatants and civilians by drawing on necessity, the principle that secures immunity from direct attack for a large segment of society by putting another segment in harm's way does not even have the pretense of providing a moral justification for wounds and deaths among combatants on the individual level. In turn, it provides little moral solace for the loved ones of a combatant killed or maimed. But if military necessity interpreted on the St. Petersburg assumption underlies discrimination, the latter rules out many of the most calamitous military practices of the twentieth century, such as terror bombing both by air forces and suicide bombers. It draws a definitive line and avoids the danger of harmful consequences inherent in criteria that ground discrimination in threat and/or liability of combatants and often end up suggesting the targeting also of civilians, some of whom can be construed as liable or threatening on an equal plane with combatants.Footnote ²⁶

The assertion about the targeting of combatants is not implicitly comparative. No claim is made that the targeting of combatants is the best or the fastest or the most efficient means of military success, nor even a claim that it is better, faster, or more efficient than any alternative means of pursuing military success. The virtue of conduct guided by necessity is that it can succeed militarily while effectively protecting many civilian lives. These are modest assertions. The St. Petersburg assumption builds a simple and plausible case for actions that will de facto protect the moral rights of many civilians to life and physical security. The moral price is that it explicitly refrains from protecting the moral rights of combatants.

It is important, however, not to appear to overstate the military utility of killing combatants. It is far from the case militarily that the more killed, the better. The brutal attrition of forces during the trench warfare in World War I and the body-count “strategy” of the United States in Vietnam show the absurdity of the notion that military success lies in killing as many combatants as possible. Although McMahan, for instance, proposes that killing in war be restricted to those who are “morally liable” (in some sense) as individuals to be killed, if we are honest we should acknowledge that the rules for the conduct of war cannot in general restrict the killing of combatants. This is not, however, to express enthusiasm for killing as many combatants as possible, which is often not only militarily unnecessary but is also distinctly militarily counterproductive. The killing of which and how many combatants is in fact necessary is an empirical question and requires a practical judgment.

Obviously, any given party may not be able to weaken the military forces of their particular adversary and thereby succeed militarily. The targeting of combatants according to military necessity provides an opportunity, not a guarantee or even an equal chance, to succeed in a military undertaking. One may not in fact be able to defeat the military forces of the enemy, in which case one will lose the war. At least one side loses every war, except in the purest stalemates. Military necessity requires only that not all means of success be ruled out on moral grounds: one means of succeeding must not be ruled out. Just as it was crucial to notice above that military necessity is claimed to be only logically necessary for the permissibility of military actions, it is now essential to understand that the St. Petersburg assumption claims only that the targeting of combatants can sometimes be sufficient for military success.

Pragmatism, Morality, and the Laws of War

This pragmatic normative framework for the conduct in war acknowledges two realities on which the law of armed conflict is already premised. First, agents to whom rules for the conduct of war are addressed have a prior commitment to the pursuit of military success, having already decided to resort to war. Consequently, rules that are incompatible with all effective military action risk being ignored and, thereby, not preventing any harm from occurring. Second, rules of conduct designed specifically to make combat difficult for combatants fighting for parties who are not justified in fighting, or otherwise to reverse the decision to resort to war, might be largely redundant of the laws for the resort to war and miss the morally valuable alternative of limiting the harm done by all sides in all the wars that are in fact fought.

From the point of view of an individual rights–based morality it is obviously deeply troubling that the calamities of war are not simply to be minimized without qualification, but that instead the minimization is subject to the strong condition that enough kinds of otherwise morally objectionable practices are to be allowed to make it possible for some warring parties to succeed militarily. But we submit that the clear-eyed acknowledgment of the absence of justification for the distribution of the casualties among individual combatants afforded by necessity is morally preferable to rules that, in fact, likewise permit wrongful conduct, but operate under false pretenses, such as the liability approach or the justification of combatant non-immunity because of the combatants' consent to having been rendered threatening.Footnote ²⁷ The proposed normative framework's ambition in moral terms is limited. But we believe that more modest limits that are honest and workable are preferable to more ambitious limits that are hypocritical in claiming a conformity with a morality based on individual rights that they do not achieve (McMahan's and Walzer's) and/or are unworkable (McMahan's).

Specifically, compared to McMahan's liability approach, the necessity principle complemented by the St. Petersburg assumption has two more advantages. The first is that it serves a morally important purpose that is neglected by rules of conduct oriented entirely around the rightness or wrongness of the decision to participate in the war: that is, limiting the morally unjustified killing and wounding of persons (and destruction of property) that occurs within all the wars that are not prevented by the refusals of individuals to participate (whether or not they ought to have refused), which is to say, all the wars that do in fact happen. The rules of conduct that pivot on military necessity can seize this crucial opportunity to protect the moral rights of all civilians. The many wars that cannot be prevented, plus the few, if any, that ought not to be prevented, are limited to actions that are both required by the successful prosecution of a war and not otherwise contrary to international law. Such an approach constrains war to the degree that war can be constrained without being prohibited, by limiting the violence to the otherwise legal violence in fact needed to attempt to win a war.Footnote ²⁸

The other advantage of the proposed framework relates to the ultimate aim, which we embrace, of translating rules for the conduct in war into enforceable law. One normative aim of law is to afford the individual protection from the arbitrary power of the state. Law does this by telling the individual exactly what she needs to do in order to avoid incurring prosecution or punishment; it does so by being determinately action-guiding. Laws that require combatants to make impossible judgments about the individual moral status of the random persons they suddenly face on the battlefield would fail in this task. Necessity interpreted on the basis of the St. Petersburg assumption, by proposing the clear and simple rule only ever to engage combatants, effectively “relieves the individual of the cognitive burden of forming her own judgements.”Footnote ²⁹ Despite its prescinding from individual rights as the touchstone for permissible conduct in war, the necessity framework safeguards one important individual right of combatants: security under the law.Footnote ³⁰

Over the remaining paragraphs we address three concerns that we anticipate will be raised with regard to the normative framework centered on necessity and the St. Petersburg assumption. The first is the question whether the applicability of the same rules for all sides in all wars does not mean that the content of the rules has somehow been weakened specifically in order to accommodate those whose resort to war is unjustified. This concern is a mistake and rests on confusion. A party whose resort to war is justified is not entitled to conduct its war any differently from any other party. It is simply entitled to conduct a war that could attain military success. All the rules for the conduct of war have the form “even if, only if”: that is, even if one is justified in resorting to war, one may conduct the war only if one abides by the necessary restraints.

What is militarily necessary is a matter of empirical, largely technical judgment, consisting of the selection of means that are in fact necessary to ends. To incorrectly judge that a particular means is necessary to a particular end is to make an empirical error, to get the facts wrong. Such judgments are independent of whether the party pursuing the end is justified in doing so. In other words, it is not the case that if pursuing an end were justified, means A would be necessary to reach this end, but if pursuing the same end is unjustified, means B becomes necessary to reach it.Footnote ³¹ The limits are the limits; and the fundamental limit is: is this means in fact necessary for reaching military victory? This empirical judgment is completely independent from the moral judgment about the justifiability of this party's pursuing a particular goal by resorting to violence.Footnote ³² Limiting the further harms and wrongs issuing from activity that ought not to be occurring at all is a morally distasteful, yet morally vital, enterprise at the heart of the laws for the conduct of war.

Some may ask, “Are you then approving of a combatant on a side that is not justified in going to war killing combatants on the other side, which is justified in going to war?” But the question has no answer because it is deeply ambiguous. One unambiguous question is: “Are the actions of a combatant fighting for a side that is not justified in going to war, including his killing or wounding of combatants on the opposing side, justified, all things considered?” The answer is: no, because no one fighting for a side that is not justified in going to war is fully justified; such a person has made a moral error in choosing to go to war. A different unambiguous question is: “If someone sincerely but mistakenly believes, after considering the matter as conscientiously as he is capable of doing, that his side is justified in going to war, is the best way for him to conduct himself from that point forward to abide by the laws for the conduct of war?” The answer is: yes—and this is the question that occupies us here.

The answer to this question does not entail that whoever acts sincerely acts well; that is subjectivist nonsense. It does imply, however, that those who make genuine moral mistakes about going to war are not necessarily morally liable to be maimed or killed. Having first made the wrong decision about joining in the war, they may from that point have conducted themselves as well as possible. If they are subsequently killed or wounded, we have no good reason for joy, satisfaction, or even complacency. In war many combatants on both sides suffer or die tragically and without full justification. Sadly, no development in the laws for the conduct of war will change that.

While some commentators might be uncomfortable with the wide range of actors that are granted a chance of military success (regardless of the justice of their cause), the opposite objection will also be raised. Thus, the second question we anticipate is: “Does the imperative of military necessity interpreted in accord with the St. Petersburg assumption deny a chance of success in situations of a justified resort to force?” We see no basis for the complaint that too little opportunity for military success is allowed if no route is provided by which the weak can defeat the strong. It is not only possible but not unusual for justice to be on the side of the weak, not the strong, so that if the strong succeed militarily against the weak, justice will be defeated. We agree that something ought to be done about this in as many cases as possible. However, we strongly doubt that the solution to this problem is to try to create means by which the weak can defeat the strong by the use of force—by, for instance, allowing the weak to attack civilians in hope of breaking the will of the strong. A better path to ensure that justice prevails in international relations would be to provide the weak with nonmilitary means of successfully pursuing their just goals.

Most important, there is no natural right that a party ought to have an equal opportunity, or any opportunity at all, to succeed in a military action. After all, the current international system is premised on a blanket prohibition of the use of force. If one has a political goal that one cannot achieve by means of a military victory gained in compliance with the St. Petersburg assumption, one simply has no permissible way of pursuing this political goal with the use of military force. Yet, it is implausible to claim that to wish to defend oneself effectively against, say, a threat of annihilation is to have a goal that does not justify a resort to military action. After all, self-defense is the recognized exception to the prohibition of the use of force. Specifically for cases of self-defense, then, is the St. Petersburg assumption too demanding?Footnote ³³ If attacks against civilians were indispensable to one's defense, then they would be necessary. This would create a direct conflict between the implications of the principle of military necessity and the requirements of the St. Petersburg assumption. This is not a problem distinctive to our approach. It arises whenever an exceptionless restriction is placed on what one may do to win. It is precisely this concern that led Michael Walzer to formulate the doctrine of supreme emergency, which allows attacks on civilians when such attacks are believed to be necessary to self-defense against a threat to the continued existence of one's community and expected to be effective against and proportional to that threat.Footnote ³⁴

At its most abstract, the question is whether there is any goal such that any means necessary to attaining it may be employed—is there an end that justifies whatever means it takes? It is perfectly possible to answer: no—some actions are unreservedly evil and are not to be performed, no matter what the cost. Yet, we have not asserted that the killing of civilians is a supreme evil, but only that the most effective possible way to limit the mayhem of war is to prohibit direct attacks on civilians. We have simply said that the St. Petersburg assumption is a generally effective limit that it is possible to maintain during military combat. This means that any exception would need to be formulated with great care and in great detail. Parties to wars routinely claim that they are fighting in self-defense. If an exception took the form of saying, “except if you believe you are fighting in self-defense,” the entire game would have been given away—the St. Petersburg assumption would simply have been abandoned. What are proposed as exceptions are often holes in the dike that undermine the entire structure.

If we managed to limit access to the exception, by formulating an inescapable test as to when a party is fighting in self-defense, the question remains what the normative framework congenial to attacks on civilians would say. “One may target anyone whose death one believes will contribute to military success”? That would not be a restraint, but the absence of any restraint. This would be too much like saying that the best rule to constrain individual behavior is that “everyone objectively ought to do as she subjectively thinks best.” One of the great merits of our framework is that it provides a clear workable rule that protects large numbers of people. The first challenge for the advocate of attacks on civilians in cases of self-defense is to formulate a proposal that has the essential features of a general principle and provides clear, workable protection for large numbers. When that is forthcoming, we can compare it with the suggestion here.

A better way to safeguard the right to effective self-defense than to grant an exception to the constraints on conduct is to make sure that the assault on civilians is never genuinely necessary. Such necessity depends on the absence of any superior alternatives. One obvious alternative is that sufficient numbers of other parties come to the assistance of the attacked party so that the threat can be defeated by collective military action in accord with the St. Petersburg assumption. Article 51 of the Charter of the United Nations acknowledges an “inherent right” to self-defense (without mentioning any restrictions on the conduct of that defense), but allows the right to be exercised only until the Security Council authorizes an effective collective defense.Footnote ³⁵ At the very least, one would not need to abandon the St. Petersburg assumption until the Security Council had failed to authorize effective collective action. Arguably, after a Security Council veto of a collective military rescue, responsibility for any civilian casualties caused by the threatened society would partly lie with the party casting the veto.

The third concern we anticipate is that we have slipped into some disreputable form of consequentialist reasoning. We advocate a type of rules that appear effective in protecting rights in the situation for which the rules are intended: a collective contest of arms. We argue that they will produce better consequences than the stricter-appearing rule that one must not kill anyone, including combatants, who is not morally liable to be killed. Thus, we have an indirect argument for a framework that turns on what the rules accomplish. But what they accomplish is in fact the protection of rights—namely, the rights of civilians. What they do not accomplish is to protect the rights of those combatants who are in no sense morally liable to be killed. The argument, however, is not that this is somehow the maximum net good—that is, that the rights or lives of the civilians protected are worth more than the rights or lives of the combatants not protected. We have no idea what such an assertion about a net good from a trade-off among lives could possibly mean. All we are saying is that protecting the rights of civilians is an achievable aim, while the rights of combatants are beyond the reach of a workable normative framework.

Those who embrace truly net-consequentialist reasoning will likely find fault with our framework. The claim that often the direct killing of civilians will minimize human death and suffering overall because it is more efficient than allowing combatants to kill each other and thus means that a war ends sooner has been made from time immemorial.Footnote ³⁶ It was made famous by the Lieber Code's dictum: “The more vigorously wars are pursued, the better it is for humanity. Sharp wars are brief.”Footnote ³⁷ Apart from the difficulties of settling the factual correctness of such grand speculative generalizations, a defender of our framework would want to raise the difficulties of formulating a generally constraining law that would embody any proposed efficiency rule: “Kill the persons whose deaths will most immediately and directly lead to military success.” Again this seems to be the absence of a constraint rather than a constraint.

Not surprisingly, true rule consequentialists who attempt to devise rules that minimize killing in war overall by allowing the attack on those individuals whose death would end war quickest would be hard to find among philosophers. It is in military doctrine, whose normativity is primarily directed toward winning wars and only in the second instance toward meaningfully regulating them, that the efficiency argument is encountered. Space does not permit us to identify the manifold guises in which it appears in military discourse and the ways in which it underlies many strategies in modern warfare.Footnote ³⁸ The targeted killing of nonmilitary regime leaders is only one major current example of the weakening of a legal category (combatants) and the expansion of permissible killing in war based on an interpretation of discrimination that rests on an efficiency argument rather than on a necessity argument. A reaffirmation of the St. Petersburg assumption and the principle of military necessity would provide an effective bulwark against the efficiency-based interpretation of discrimination, which is the single most significant challenge to limiting killing in war in contemporary armed conflict.Footnote ³⁹

The efficiency argument shows that it is really not only a commitment to noncombatant immunity that we need in order to effectively limit killing in war. For different reasons neither Walzer nor McMahan, whom we have used as foils in this analysis to demonstrate the merits of the proposed normative framework, calls for an abandonment of the legal distinction between combatant and civilian. For Walzer the distinction between combatants and civilians adequately tracks individual rights in war, and McMahan believes that since law cannot at the moment properly track morality, in discriminating “in a way that attracts general agreement, we can best succeed in insulating significant areas of human life from the destructive effects of war.”Footnote ⁴⁰ Our argument in this article is that even if we all agree on noncombatant immunity, it matters what underlies the application of discrimination.

We have demonstrated that the underlying criterion matters because it determines whether the rules that permit killing in war are hypocritical and whether they can actually be consistently applied on the battlefield. We close with the observation that, in addition, what we think explains why we are allowed to kill combatants also defines who we think is a combatant. If combatant killing is “justified” by individual liability, threat potential, or efficiency, it is ultimately a matter of tenuous perception, interpreted amid fear and stress, about the attacker who dies. If, on the other hand, we attach discrimination to necessity and the St. Petersburg assumption, it simply tracks who among its own individuals each collective allows to take up arms and carry out the confrontation. We advocate a morally thin, but honest, workable, and—most important—definitive limit on killing in war.