The twenty-first century begins, much as the twentieth ended, with estimates as high as hundreds of thousands of civilians being slaughtered in international and civil wars.2 Reports of recent international conflicts tell of civilian deaths exceeding combatant deaths, sometimes by multiples.5 For example, During Israel's summer 2006 conflict with Hezbollah in Lebanon the Israeli Ministry of Foreign Affairs claimed the deaths of 119 Israeli combatants4 compared to independent reports of Lebanese civilian casualties amounting to at least 700.5 Still more recently, the Iraqi Health Minister shocked the world by reporting that the U.S.-led war to overthrow the Iraqi government from 2003 to 2006 had caused approximately 150,000 civilian deaths (previously reported at less than 50,00O).6 That number seems still more unsettling when compared to the estimated number of Iraqi military personnel killed in the war-approximately 12,000/ Deaths in the coalition forc003 to 2006 numbered fewer than 3000.8 Meanwhile, the civilian body count in the Sudanese civil war in Darfur reportedly continues to approach 250,000.9 These figures have raised sufficient concern with the U.N. Security Council that it has recurrently, since 1999, adopted resolutions deploring civilian casualties in armed conflicts and urging states to comply with their obligations under international humanitarian law.10

The disparities between combatant deaths and civilian deaths merely represent a long-term trend in modern armed conflicts. According to the International Committee of the Red Cross and various U.N. reports, the ratio of civilian to combatant casualties was between 5% and 10% in the First World War and then dramatically leapt to 50% during the second World War." By the 1990s, 75% of all casualties resulting from armed conflicts were civilian, and in some cases the rate has allegedly reached as high as 90%.'2 While the ICRC and U.N. trend figures are likely exaggerated by the inclusion of post-conflict casualties-for example, death by disease and starvation resulting from lack of access to normal food, clean water, and medical facilities-in more recent calculations not factored into older figures,13 there is little doubt that civilian casualty rates have increased dramatically.14

All this is, of course, precisely the kind of tragedy that international humanitarian law is designed to prevent. The discrimination principle, long and widely recognized as a rule of customary international law and codified in several core humanitarian law treaties, prohibits combatants from directly attacking civilians and civilian property.15 Discrimination does not, however, offer any protection from the consequences of overzealous or careless attacks on military targets that endanger civilians. To shield civilians from the worst effects of attacks on military targets, the customary norms of necessity and proportionality restrict the modalities of attack. The modern necessity doctrine permits otherwise unlawful attacks on military targets when such attacks are necessary to prevail. By the same token, however, it prohibits combatants from using force endangering civilians or their property except to the extent that such force reasonably appears essential to attaining military victory.16 The proportionality principle, first codified in 1977 in Additional Protocol I to the 1949 Geneva Conventions, prohibits attacks against civilian targets and attacks likely to cause "incidental loss of civilian life, injury to civilians, damage to civilian objects, or a combination thereof, which would be excessive in relation to the concrete and direct military advantage anticipated."17 Attacks violating the principles of necessity and proportionality are considered to breach what the ICJ has called "intransgressible principles of international customary law"18 and may be considered grave breaches of the Geneva Conventions.19

There are many reasons to believe that discrimination and its related principles of necessity and proportionality should effectively protect civilians from the worst effects of armed conflicts. Among these are the longevity and widespread acceptance of these principles by states,20 the greater degree of detail in modern conventional law to guide the decisions of military commanders,21 and the adoption of new technologies for avoiding civilian casualties, such as GPS-guided bombs and missiles. Given these developments, one might suppose that armed conflicts have evolved to present fewer and milder dangers to civilians than in the past. But as the discussion of civilian casualty rates implies, such a supposition could hardly be more wrong. If any evidence can prove that international humanitarian law has failed to protect civilians from military operations, it is the disproportionate ratio of civilian to military casualties despite the presence of technological and operational factors that should have enabled belligerents to reduce such casualties. Consider that, at most, 18% of the deaths caused by the 2003 war in Iraq were combatant casualties in spite of the coalition forces' technological supremacy.22 Meanwhile, states continue to trumpet their commitment to protecting civilians from intentional attacks, as the Security Council unanimously did in April 2006, when its members reaffirmed the "responsibility to protect" civilian populations from large-scale attack.23 The divergence between the professed goals of states in adopting international humanitarian law and civilian suffering during modern armed conflicts raises troubling questions about the efficacy of the legal norms protecting civilians.

Perhaps it should not then be surprising that civilian immunity as implemented in international humanitarian law is not universally acclaimed. The implementation of international humanitarian law's civilian protections clearly leaves much to be desired. But even beyond the law's pragmatic failure to reduce casualties, recent discussions have attacked the very concept of civilian immunity, questioning its moral coherence, its independence from the ius ad bellum, and its impartiality toward underrepresented classes such as women or ethnic minorities seeking independence or secession. In this Article, I explore both theoretical and operational challenges to civilian immunity to determine whether the concept merits preservation, alteration, or replacement.

On the theoretical score, two fundamental challenges to civilian immunity have become popular in recent years. The first posits that the distinction between civilians and combatants is morally arbitrary and that the relevant ethical standards for subjecting categories of persons to attack in the ius in bello should be some other criteria, such as moral innocence or capacity to harm the belligerent state or its combatants. Some variants of this argument also attack the dualistic nature of the laws of war (ius ad bellum versus ius in bello) and argue for their moral inseparability.

The second challenge has emerged from increasingly common feminist critiques of the ius in bello. Several scholars have criticized the principle of discrimination as reflecting an inherent gender bias. Equating men with "protector" combatant and women with "protected" civilian classes, these critiques conclude that the principle of distinction is both a product and a perpetuator of gender stereotyping. Civilian immunity is rejected by these authors as a fundamentally, and in some views unalterably, flawed concept.

Operational challenges to civilian immunity attack not the concept itself, but its implementation in the law of war. The rules codified in the 1949 Geneva Conventions and their 1977 Additional Protocols-especially those embodying the principles of necessity and proportionality-have long been derided as too vague and inefficacious to justify their inclusion in a major convention.24 Because necessity and proportionality are uncertain in application, impose only tepid duties, and are rarely enforced in any event, their inclusion in international humanitarian law could be considered counterproductive. What is especially interesting about these criticisms is that many states that initially opposed the inclusion of proportionality in Additional Protocol I to the 1949 Geneva Conventions did so not because it constrains the effectiveness of their armed forces, but rather because it does not restrain them well enough.25 This family of challenges invokes jurisprudential weaknesses of humanitarian law, as well as its de facto failure to protect civilians as a basis for questioning whether some aspects of civilian immunity actually inflict harms on civilians that outweigh the benefits provided. Operational criticisms of civilian immunity, then, contrary to theoretical ones, are based on the idea that there is too little civilian immunity rather than too much or the wrong kind. The accelerating rate of civilian casualties lends particular force and urgency to these operational critiques.

Each of these critiques of civilian immunity and its underlying doctrines will be considered and analyzed in the following pages. Ultimately, I do not find any of them sufficiently trenchant to justify rejecting altogether civilian immunity as currently embodied in international law. But each critique makes important observations about the weak points in the implementation of civilian immunity in international humanitarian law, which in turn suggest useful avenues of reform to strengthen civilian immunity in pursuit of a more meaningful and enforceable ius in hello.

II. ETHICAL CHALLENGES TO CIVILIAN IMMUNITY

War necessarily involves violence, and so long as we agree that killing people can at least sometimes be justified, the task of an ethically defensible international humanitarian law must be to identify who may be killed, under what circumstances, and how such killing should take place within a morally acceptable paradigm. Currently, international humanitarian law embodies a fairly simple modelgrounded in the conviction that it is not morally acceptable to kill civilians if such killing can be avoided-that applies upon commencement of an armed conflict: Civilians and civilian property are immune from direct attack.26

Based on the same conviction, when a military objective cannot be achieved without unintended civilian casualties, the attack may only go forward if the attacker complies with the relevant rules of international humanitarian law-in this case, necessity, proportionality, and related rules codified in Additional Protocol I-to minimize unnecessary civilian casualties.

Active members of the armed forces (i.e., those who have not surrendered, become prisoners of war, or been rendered hors de combat by illness or wounds) are always subject to direct attack, apparently based on the conclusion that it is always morally acceptable to kill such combatants, even when not factually necessary.

While this is the current ethical model of international humanitarian law, it is certainly not the only one possible. As the model makes clear, the beliefs underlying civilian immunity appear to be quite simple-in moral terms, combatants' lives have no value requiring legal protection by belligerents, while civilians' lives have a great deal of it. Civilian immunity is supported by chivalric military values, epitomized by MacArthur's quotation in the prolegomena attributing to combatants the role of a protector class and to civilians the role of a protected class. To those trained in the European school of human rights philosophy, with its paradoxical Kantian foundation of equal and inviolable human dignity27 and Hegelian superstructure attributing to the reified state an inherent right to claim the loyalty and even lives of its citizens,28 the concept of discrimination may seem unproblematic, or at least somehow reconcilable. Because slaughtering the unarmed is especially the repugnant to human rights values, it may seem superfluous to back confidence in the rectitude of the principle of civilian immunity29 with a careful analysis of the ethics underlying international humanitarian law. But philosophers make it their business to raise such questions, and they have done so vigorously.

Diplomats, politicians, nongovernmental organizations, the media, and other international elites commonly label civilians with adjectives like "innocent" and "defenseless"30 (in MacArthur's terms, "weak and unarmed"), implying through these terms that killing such persons would be morally reprehensible. By definition, civilians are indeed unarmed and relatively defenseless, with only a few exceptions unimportant here. These labels are in effect assertions of objective fact-whether a specific civilian has taken up arms in pursuit of military activity can be ascertained through the production of evidence.31 The term "innocent" differs from these others; it connotes immunity to moral condemnation, an assertion of status capable of verification only pursuant to an accepted ethical theory. This difference between the kinds of terms applied to civilians is highly significant for purposes of ethical analysis. One can be defenseless, weak, and unarmed without being innocent, just as one can arm and defend oneself while maintaining moral innocence (i.e., without necessarily violating any rules of right moral conduct). Ethical critiques of civilian immunity can be divided into two overlapping classes based on this distinction. Both classes share the assumption that killing can only be justified if the target is morally culpable in the sense of intentionally directing his or her activities to the attainment of an immoral objective.

The first class of critiques focuses on the claim of "innocence" to determine the justifiable subjects of attack. In most of these views, the justice of the underlying armed conflict and each individual's relation to it is inseparable from the question of moral innocence. Morally innocent persons are those who do not actively seek to attain an unjust end. Such persons should be immune from attack, whether they are civilians or combatants. Morally culpable persons, conversely, should be subject to attack regardless of their status as civilians or combatants.

The second class shifts the focus away from the justice of the armed conflict and evaluates instead the combatant's decisionmaking within the conflict as a fixed, morally neutral context. These critiques may accept international law's dichotomization of ius ad bellum and ius in bello without concluding that this classification morally justifies subjecting combatants as a class to lawful attack or civilians as a class to immunity from attack. Critiques of this kind compare the situation of the combatant choosing targets to that of an individual defending himself from an external attack and thus engage in a process of moral reasoning by analogy. The conclusion is that the justification for immunity from attack should turn neither on a person's membership in the class of combatants or civilians nor on the person's moral innocence or culpability respecting the armed conflict as a whole, but rather on the direct threat of harm posed by the person (civilian or combatant) to a combatant or his armed forces. In other words, this class of critique seems to correspond less to the claims that persons should be immune from attack because they are "innocent" than to claims that they should be immune because they are "unarmed" and "defenseless" (or, more accurately, offenseless). Each class of critiques will be evaluated in turn.

A. Moral Innocence and the Just War

The assertion that civilians as a class are "innocent" cannot be more than a figure of speech. Nobody could reasonably believe that mere membership in the class of civilians constitutes automatic exoneration from moral blame for the state's wartime conduct. Civilian industrialists and politicians who have engaged in morally reprehensible conduct such as enslavement and genocide during armed conflicts have been convicted of and punished for war crimes and human rights violations by international criminal tribunals for this reason.32 In an ethical analysis, moral innocence is a question of conduct and intentions. The context in which the conduct and intentions occurs, such as membership in the class of civilians or combatants, is relevant but merely one factor in the analysis rather than the sole determinant of the conclusion. Some authors have accordingly argued that discrimination as a rule of international humanitarian law is morally indefensible because, notwithstanding the qualification of civilians as "innocent," civilian immunity fails to correspond with any coherent conception of moral innocence. Because such arguments typically account for the justice or injustice of the underlying war effort, it is necessary to take a brief detour from the ius in bello to the ius ad bellum.

Ius ad bellum relates solely to the legality of initiating the use of armed force. That subject is, like international human rights law, peculiarly susceptible to ethical analysis. The attainment of a just world order under law presupposes constraints on organized violence between and within states. A primary function of international law is to adopt and enforce legal rules defining when the use of military coercion is legitimate and limiting its legitimacy to situations where it is morally justified. The ius ad bellum contributes to this objective by defining the circumstances under which states may resort to armed force. lus ad bellum is located in both customary international law and the U.N. Charter, but has little place in either the Hague law or the Geneva law, which define most of the ius in bello rights and obligations of states after hostilities have commenced.33

The ius in bello governs the specific conduct of military operations. It presupposes the existence of an armed conflict, and it applies regardless of the legality (or illegality) of any belligerent's position under the ius ad bellum. lus in bello contains rules governing the conduct of hostilities by belligerents and their combat forces in both legal and illegal wars without distinction. Its goal, like the goal of ius ad bellum, is to limit violence to morally justifiable purposes and amounts. But the ius in bello applies in much more varied factual circumstances and must, therefore, be much more detailed and complex. lus in bello does not merely determine the legality of a single decision of a state; it governs innumerable decisions by military commanders at multiple levels of the chain of command in every kind of pre-combat, combat, and occupation situation.

A rejection of international law's position, in which ius ad bellum is divorced from ius in bello, underlies much criticism of civilian immunity. The temptation to reject any specific killing unless morally justified in a long view-in this case, in pursuit of a just casus belli-is understandably strong. The weighing of military advantage against the loss of civilians lives or property must be undertaken in the context of the worthiness of the objectives of the war. Professor Hurka argues:

If "military advantage" justifies killing civilians, it does so only because of the further goods such advantage will lead to, and how much it justifies depends on what those goods are. Compelling though it is, this view has the radical implication that no act by soldiers on a side without a just cause can satisfy proportionality: if their acts produce no relevant goods, they can never be just. ... [I]f we consider the morality of war rather than its legality, the independence of its two branches cannot be maintained. Whether an act in war is in bello proportionate depends on the relevant good it does, which in turn depends on its ad bellum just causes.34

One ethical objection to civilian immunity in international law, then, is its basis in a strict division between ius ad bellum and ius in bello.

At the very least, in this view, killing a combatant who is pursuing a just objective in a just manner should be considered immoral and, therefore, illegal. This is the first step to breaking down a flat rule of civilian immunity into a more complex rule conferring immunity from attack on combatants pursuing a just war as well. Indeed, the chief British prosecutor at Nuremberg tried to make precisely this argument: "The killing of combatants in war is justifiable . . . only where the war itself is legal. But where a war is illegal... there is nothing to justify the killing, and these murders are not to be distinguished from those of any other lawless robber bands."15 The corollary of this principle is that combatants recruited to fight unjustly an attacker waging a just war can defend themselves, if at all, with only minimal force. Moreover, combatants involuntarily waging an unjust war should receive greater protection than their more willing fellow combatants. As Jeff McMahan has argued, the view that the justification for self defense against Innocent Attacker (i.e., one who fights unwillingly) is the same as the justification for self defense against a Culpable Attacker (i.e., one who fights willingly) is implausible. If it were true, then, "whatever restrictions there are on the defensive use of violence against an Innocent Attacker would also apply to self-defense against a Culpable Attacker. But intuitively the restrictions on self-defense against an Innocent Attacker are considerably more stringent."36

The second step depends on the assertion that it is more morally justifiable to attack civilians contributing to an unjust war than to attack combatants contributing to a just war. Civilians voluntarily contributing to an unjust war, either morally or materially, are not "innocent" in the ethical sense, but are rather offenders against international peace and, consequently, more appropriate targets for attack than combatants prosecuting a just war. The logic of this view can be extended still further, justifying the incidental killing of some morally innocent civilians in pursuit of the just war to the extent that such killing does not outweigh the morally justifiable goals of the armed conflict itself. In other words, unless one is a very strict Kantian, the bellum iustum may justify killing a certain number of morally innocent civilians to prevent another, more morally reprehensible result from occurring, such as the deaths of many more innocent persons, combatant, or civilian.37

The consequences of an individual combatant's actions under the ius in bello should partly depend, then, on the moral justifiability of his or her state's decision to use armed force under the ius ad bellum. If that decision is a just one, killing morally culpable civilians may arguably be as justifiable as killing morally culpable combatants. Conversely, intentionally killing civilians or combatants contributing to a just war cannot be morally permissible. If the state's resort to armed force is not morally justified, no attack by that state's combatants can be morally justified.

It does not follow from these observations that every citizen of a state waging unjust war should be subject to indiscriminate slaughter. What does follow is that some persons in a state waging unjust war may be innocent of intent to further that war, while others may have culpable intent, and innocence does not necessarily depend on one's status as a civilian or combatant. Contrary to popular perceptions of combatants as aggressive and mercenary and of civilians as innocent and inoffensive, specific members of the military may engage in an unjust armed conflict unwillingly, while civilians may actively and intentionally support or contribute to a conflict.38 For example, an attacking combatant may have been conscripted into service against his will, just as a civilian may politically support an unjust war effort. Indeed, a civilian journalist, clergy, or celebrity may encourage or prolong the conflict by rallying public support for it and encouraging voluntary enlistment, thereby presenting a more tangible (though indirect) threat to the opposing belligerent than an individual soldier who participates involuntarily could ever pose. In this sense, some combatants may be morally innocent of any intention to commit a harmful or unethical act, while the opposite may be said of some civilians. The distinction between combatants and civilians as lawful targets of attack, in this perspective, is little better than morally arbitrary.

These critiques of civilian immunity have strong theoretical appeal. The concept of binding the ius ad bellum to just war theory has long been a project promoted by international lawyers and diplomats.39 States and their predecessors, empires, have recurrently sought to justify their resort to arms on moral grounds throughout history, beginning at least with the Roman ius fetiale and advanced most concretely through the U.N. Charter's prohibition on the use of force for territorial aggrandizement and other aggressive purposes. Indeed, the very purpose of IMJ ad bellum is to deter war and sustain a just world public order under law. Whether the world public order is "just" presupposes an ethical judgment, inviting precisely the kinds of moral analysis that just war theories provide.

As for the insistence that within this just war context, the morally relevant criteria that should guide legal rules specifying appropriate targets for attack is moral innocence rather than membership in the civilian or combatant class, this theory also has its charms. It is merely tautological to observe that moral culpability and innocence are characteristics not of entire classes of persons but of the individuals within those classes. Except in a theory of collective responsibility for military organizations or states, a theory long rejected by the ius in bello,40 the treatment of individuals should correspond to their personal moral characteristics. From the perspective of a single combatant deciding whether a single "enemy" is an ethically acceptable target of attack, then, both the context in which killing occurs (just or unjust war) and the intentions of the individual target are morally relevant.

As is often the case with theoretically appealing arguments, however, this one does not work out in practice. Attempting to translate strict ethical analysis into concrete rules of international law would fail to eliminate the moral arbitrariness of the current rules and would create entirely new ethical problems. To understand why, it is first necessary to consider how the conditions under which judgments about the legality of an act under the ius ad bellum differ from similar judgments under the ius in bello.

The decision to deploy armed force is generally a one-time decision-either the potential belligerent attacks or not, invades or not, defends or abdicates. International politics often provide ample opportunity for moral deliberation before a decision is made to resort to armed force, making the grounding of ius ad bellum in just war theory not merely desirable but feasible. The decision to use armed force is, moreover, subject to accepted legal principles limiting the use of force mainly to situations of individual or collective self-defense, current or anticipatory, and to humanitarian intervention in limited circumstances. In cases of borderline legality, states often have the time and resources to research the facts that would enable a reasonably certain judgment of the proposed use of force's compatibility with the ius ad bellum.

The arguments for integrating moral theory into the decision of a state to employ armed force apply less readily in factually rich combat situations. Combatants, whose behavior is currently governed only by ius in bello, typically lack the luxuries of careful and objective legal analysis and time to consider and weigh options when making tactical decisions. One implication or stipulation of moral innocence critiques is that they require each combatant to objectively evaluate the morality of a specific individual act of attack, including consideration of the justice of the cause for which the combatant and his target fights. Aside from the unavailability of objective information to combatants in the field, the ius in bello cannot provide a precise moral compass to combatants for judging the justice of harming individual targets. Such a project would be hopelessly Utopian. First, the justice of a resort to armed conflict is rarely black and white and never uncontested. The world community lacks an impartial general arbiter of which wars are just and which unjust,41 and the states and their populations cannot always be trusted to judge objectively the merits of their own case. As Francis Bacon observed, "[t]here is that justice imprinted in the nature of men, that they enter not upon wars (whereof so many calamities do ensue) but upon some, at least specious, grounds and quarrels."42 Political and military elites never admit when they are engaging in an unjust war, or even in a war of outright aggression and territorial aggrandizement. They invent some nontrivial argument to justify the most naked aggression, as Iraq did when invading Kuwait in 199043 and as Argentina did when invading the Falkland Islands in 1982.44 In both cases, historical and moral claims to the territory were used to justify the armed aggression. As Edmund Spenser observed four hundred years ago:

For never wight so evill did or thought,

But would some rightfull cause pretend, though rightly nought.45

Potential combatants themselves cannot reasonably be expected to possess the legal expertise, knowledge of political facts, and psychological neutrality to evaluate the justice or legality of these claims fully and impartially, particularly where the belligerent has not made full information about the conflict publicly available. International humanitarian law merely accounts for the normally subjective and subheroic human nature in refusing to punish a belligerent's citizens for not coming to their own objective, individual assessment about the justice and legality of their state's resort to armed attack.

Even if potential combatants could objectively identify unjust wars, it would be unrealistic to require them to resist participating when domestic law mandates their enlistment or cooperation. Armed conflicts sweep up the populations of the belligerent states regardless of the most strident individual moral objections of members of the population, even in democratic states. The threat of enforcement of domestic law-which among its penalties for refusal to register for a draft may include imprisonment or capital punishment-is far more immediate and threatening than the latent possibility of enforcement of international law at some future point. Requiring combatants facing imminent imprisonment or execution to engage in such delicate speculations defies credibility. If their activities are to be governed by regulations, the legal rules must be more definite than just war theory currently makes them.

If demanding that combatants objectively and accurately assess the justice of the decision to resort to armed conflict is unrealistic, how much more unrealistic is the expectation that they could identify which civilian and combatant adversaries are morally culpable and which are morally innocent? A soldier or bomber pilot has virtually no means to judge the intentions of his target. The principle of distinction, in contrast, can claim the advantage of simplicity and clarity. Lawful combatants by definition wear distinct uniforms and emblems and carry their arms openly regardless of their personal misgivings about the conflict in which they play a part.46 Except by surrender, they have no way of communicating clearly to the opposing combatants their moral innocence. As Professor McMahan concludes, it is impossible for pragmatic reasons to distinguish between those willingly posing a threat of harm and those doing so unwillingly.47 Similarly, a combatant may not be able to distinguish readily between a civilian intentionally contributing to an unjust war and a civilian objecting to the war. And in any case there are degrees of moral culpability (e.g., ambivalence or apathy about, or qualified support of, an unjust war), few of which can be judged at a glance, and some of which presumably should not subject one to direct military attack.

All this does not mean, however, that the ius in bello is doomed to moral arbitrariness. A flat rule of civilian immunity unquestionably results in some morally culpable civilians avoiding attack and some morally innocent combatants dying, but it creates offsetting benefits. Given the reality that combatants may find themselves agents in an armed conflict of indeterminate justice and facing opponents of uncertain innocence, the challenge is to craft rules of international law that preclude indiscriminate killing. International humanitarian law sacrifices the unattainable standard of universal individual justice in pursuit of the more feasible goal of minimizing the total number of unnecessary casualties given the realities that combatants necessarily face.48

A potent ethical argument for a blanket rule of civilian immunity and combatant nonimmunity, then, is that in the political, psychological, and sociological context in which international law must operate, the rule is both practical and progressive in serving the worthy goal of diminishing the total human suffering caused by armed conflicts.49 In the St. Petersburg Declaration of 1868, the great powers declared:

Considering:

That the progress of civilization should have the effect of alleviating as much as possible the calamities of war;

That the only legitimate object which States should endeavour to accomplish during war is to weaken the military forces of the enemy;

That for this purpose it is sufficient to disable the greatest possible number of men ... .50

In effect, the St. Petersburg Declaration set a policy, continuing to the present, designed to keep the devastation caused by armed conflicts within the bounds of weakening enemy military forces. Any system of morality that values human life and happiness and consequently seeks to minimize unnecessary suffering deplores legal norms not narrowly tailored to these ends. The discrimination, necessity, and proportionality principles are designed to serve this end by limiting de lege the objects of attack to a subset of all persons and property that are de facto susceptible to attack. The discrimination principle does so by permitting only attacks that, as a general rule, will not contribute significantly to the resolution of the conflict. Even if international law achieves that goal imperfectly, it does so through rules that are (or at least could be made to be) applicable to the battlefield.

The justification for a universal license to harm combatants, then, is that their intentions can pragmatically be judged only as a class (i.e., by the hostile intentions of the military force to which they belong). The judgment whether some individuals in a class identified as a legitimate target of attack (combatants) may not "deserve" death, while others in a different class identified as immune from attack (civilians) do "deserve" death necessarily transcends the scope and competence of the ius in bello.^ Moral desert simply does not and cannot enter into its calculations. The role of humanitarian law, given the inevitable occurrence of some armed conflicts of whatever nature, is to constrain belligerents from spreading unnecessary devastation without unduly impinging on their perceived self-interest in achieving military victory. Although this approach ignores important theoretical ethical issues, it is not itself ethically neutral or arbitrary.

B. Capacity to Harm

There is an alternative approach to amending the discrimination principle for greater ethical coherence that does not require importing just war concepts into the ius in bello. Perhaps, given the necessary independence of the ius in hello from the ius ad bellum, the relevant ethical litmus test is neither combatant status nor moral innocence, but individual capacity to harm the belligerent. International humanitarian law already reflects the notion that capacity to harm may alter the strict principles of discrimination. Civilians who take up arms against a belligerent are lawful subjects of attack,52 and members of the military no longer capable of threatening violence (hors de combat) are immune from attack.53 The rules of humanitarian law do not, however, directly correlate legality of attack to threat of harm. For example, direct attacks on any combatant or noncombatant member of the adverse military forces (except religious and medical personnel) are permitted regardless of the lack of immediate threat posed by such individuals, while direct attacks on civilians are impermissible even if their activities are necessary for sustaining the war effort.

Legitimate questions have been raised as to why the line between lawful objects of attack should be drawn in this manner. Lawrence Alexander has claimed that self-defense could ethically justify attacks on civilians if they play a "necessary or sufficient" role in causing the threat posed by the combatants.54 Robert Fullinwider contends to the contrary that killing of combatants alone in armed conflicts is an independent moral rule justified by the right of states to self-defense.55 Because enemy combatants pose an immediate threat of harm to the belligerent, the belligerent is justified in attacking them. In contrast, there is no moral justification for killing enemy civilians, in contrast, because they are incapable of putting the belligerent in immediate jeopardy. Alexander and Fullinwider agree in assigning an ethical justification to an attack based on the target's capacity to harm the attacker. They further agree in basing this conclusion on models of self-defense.

But as their different opinions illustrate, if the litmus test for morally justified killing is based on the capacity of combatants to harm the enemy, this raises the more general question of how to define the "harm" that justifies drawing the line between legitimate targets of attack in one place rather than another. If one defines harm, as Alexander does, to mean any causal and material contribution to the prosecution of the armed conflict in general, then a norm uniformly permitting intentional attacks on combatants and prohibiting such attacks on civilians is difficult to justify. If one defines harm, as Fullinwider does, to include only direct threats to the belligerent, then only those combatants and civilians who are in the position to commit violence against the belligerent should be lawful targets of attack. Fullinwider's view more nearly reflects the current rules of international humanitarian law,56 but why not define harm more broadly?

To the extent that direct attacks on civilians would damage an opposing belligerent's military effort, such attacks seem more justifiable than attacks on military forces not hors de combat but presently in no position to harm the attacker or impede its military victory. After all, combatants generally do not operate as independent agents. They are enmeshed in a societal fabric that includes a support structure encompassing most classes of working civilians, from the most basic producers of staple goods even to spouses or relatives who free combatants for military duty by caring for their children. Civilians and civilian infrastructure perform an indispensable role in the success of the armed forces, providing funds, power generation, food, clothing, supplies, armaments, and ammunition to the armed forces." Civilians such as miners, farmers, transportation providers, or software engineers may encourage or support the armed conflict actively through the provision of goods, technology, or services. Journalists, novelists, and television news anchors may discourage peace by actively fomenting jingoism or merely by whitewashing the horrors of the conflict. Even apparently innocuous entertainers, who visit war camps to raise military morale, may thereby prolong or intensify the conflict. The harm-causing distinction between civilians and military persons and objects blurs yet more as traditional military functions, such as aircraft maintenance, troop transportation, security, and combat training are contracted out to civilian corporations, as the United States has increasingly done.58

Finally, it should be noted that the distinction between civilians and combatants is also a temporal one to some extent. Every society has civilians who are permanently ineligible for combatant status due to physical, mental, or legal59 impairment of one kind of another. But many of today's civilians are capable of becoming tomorrow's combatants. International law does not forbid a belligerent to augment its military forces by drafting or recruiting civilians in the course of a protracted armed conflict.

The negotiators of the 1977 Additional Protocols recognized these concerns, and the relevance of capacity to harm made its way into the principles relating to attacks on property. Additional Protocol I classifies as "military objectives" any "objects [that] by their nature, location, purpose or use make an effective contribution to [the] military action and whose total or partial destruction, capture or neutralization, in the circumstances ruling at the time, offers a definite military advantage."60 Such objects are lawful targets of attack, even though they may have a predominantly civilian use,61 so long as they are not "indispensable to the survival of the civilian population."62 Accordingly, the United States military currently takes the position that civilian economic facilities such as power plants, telecommunications systems, and rail yards that "indirectly but effectively support and sustain the enemy's war-fighting capability" are not entitled to immunity from attack.63 It is not difficult to find historical instances of automobile assembly plants being converted into military truck and tank factories, or civilian seagoing vessels being commandeered for military use.64

However, the rules apply only to property. Neither conventional nor customary international law permits direct attacks on civilian persons who, "by their nature, location, purpose or use make an effective contribution to the military action and whose . . . destruction [or] capture . . . , in the circumstances ruling at the time, offers a definite military advantage." Some have argued that the effective contributions of civilians to the "war-fighting" capabilities of states dilute the ethical force of the discrimination norm.65 In a sense, these critiques of the discrimination principle reincarnate the pre-1945 amoral concept of total war,66 conceiving of the state as a unitary actor-a kind of societal war machine-in which civilians are as integral to the war-making function as combatants.

The force of these arguments derives primarily from the analogy between private, personal, self-defense situations, and armed conflicts. Killing someone who poses a direct threat to oneself may as a general rule be ethically justified if necessary for self-defense. Killing someone who poses no direct threat may be as a general rule ethically unjustified. But consider the example of someone who poses no direct threat to an individual person, but enables another to pose such a threat. A gun shop owner (Civ) is supplying a dangerous person (Com) with bullets, which Com uses to threaten the person's life from time to time by taking potshots from his house as the person walks down the street. The police refuse to be involved. Civ is aware of the threat and, while he perhaps has nothing against the person, refuses to stop supplying the bullets to Com. If not for Civ, nobody would sell bullets to Com. The person's life is in danger, and it can only be made safe by killing Civ or Com or both. It may be possible for the person to attempt to kill Com at great risk to himself, or he could attempt to kill Civ at a significantly lesser risk to his own life. Is it ethical to reduce the risk to the person by killing Civ, whose threat is definite but indirect and unintended, instead of Com, whose threat is direct and intended? Reasonable persons could differ on the conclusion, but in these circumstances, killing Civ could be rationally justified. This argument seems at first glance to undermine the discrimination principle on grounds of self-defense rather than desert.

However, translating this kind of hypothetical to international armed conflicts proves problematic. The role of civilians in supplying logistical support, raising morale, and otherwise indirectly encouraging the armed forces is more attenuated and varied than the example suggests. Under current law, the fact that a civilian or group of civilians plays a definite role in supporting armed forces does not in itself convert them to legitimate targets of attack under humanitarian law. There is a sound reason for this limitation. Reverting to pragmatics again, it is difficult to distinguish between civilians who provide such support and civilians who have no facilitating role whatsoever in the war effort. It is the role and duty of combatants to use their best efforts to harm the opposing belligerent. In addition, combatants have the means and the presumed intention of harming opposing combatants. They are a legitimate target of attack because, by their reified role of "combatant," they are required and expected to focus their primary efforts on harming the enemy or assisting those who do, and they are capable of doing so immediately and effectively.

The role of civilian carries with it no such duty, and civilians typically perform functions of infinite variation. These roles contribute to military operations on a sliding scale of indistinguishable gradations, from those persons whose activities could make no discernible impact on military ventures to those working full time in munitions factories. In their role of "civilian," they do not necessarily pose any threat to combatants, either direct or indirect. A rule permitting combatants to attack civilians who make a significant contribution to the armed conflict would lend itself easily to abuse, potentially weakening the rules of armed conflicts. This is one of those rare issues on which the slippery slope argument is fairly cogent. A bright line between lawful and unlawful targets of attack prevents the increasingly aggressive use of force and counter-force that has historically tended to lead to unrestrained warfare.

Of course, this does not mean that the line should be drawn between combatants and military targets on one hand and all civilians and civilian targets on the other. Humanitarian law carves out exceptions for certain designated classes of civilians that clearly make a material contribution to armed conflicts. Thus, humanitarian law permits some forms of attack directed at civilians working to advance the war effort directly." But given the typically limited information about targets available to combatants and their commanders in a combat situation, international law permits them to rely on heuristics in determining legitimate targets of attack.68

If legitimate attacks on specific targets contribute to limiting armed conflicts, then the current discrimination principle justifies itself. Combatants are expected to wear uniforms and indicia of membership in the armed forces and to operate under responsible command, thereby facilitating their distinction from civilians. While Additional Protocol I diluted this requirement in situations of guerilla warfare,69 distinguishing uniformed members of a military organization from civilians is easier than distinguishing civilians promoting a war effort from civilians undermining or neutral to it.

There is an additional consideration justifying this bright-line rule of discrimination. Allowing attacks on civilians would set a precedent that legitimizes the use of force against the attacker's civilians by unconventional modalities such as terrorism. In attacking enemy civilians, belligerents undermine the ethical distinction between attacking those who pose a direct threat and those who do not. Indeed, some commentators have criticized international humanitarian law as "ultimately a system designed to protect the self-interests of the more powerful states"70 because they can afford to hurl long-distance attacks on less powerful states, with their attendant high rate of civilian casualties, while remaining legally protected from "terrorist" attack on their civilian populations at home.71 While this view is mistaken in asserting that humanitarian law is "designed" or intended for that purpose, the moral arguments against terrorism buckle when belligerents attack civilians intentionally or are indifferent to their safety.

In the final analysis, although the philosophical observations discussed in this section do not significantly undermine the policy rationale for civilian immunity, they do lead to an important conclusion. In public international law, helium iustum depends conceptually on a judgment of the legality and morality of the initiation of armed conflict by a state that has been provoked by a casus belli. Because the reality of armed conflict does not, however, permit belligerents distinguish between morally innocent and morally guilty targets (combatant or civilian), there will always be a certain amount of morally unjustified killing in any armed conflict (1) no matter how morally and legally justified its initiation was (because a valid casus belli does not justify or excuse every act committed in the course of an armed conflict permissible under the ius ad bellum), and (2) no matter how effectively the belligerents manage to distinguish between combatants and civilians in directing their firepower. The second observation holds true because, although civilian immunity can be justified as preventing unnecessary killing on the whole, it cannot be tailored to require belligerents to attack only (or even primarily) morally justifiable targets. Indeed, this conclusion calls into question the very possibility of bellum iustum except in circumstances where the resort to force is fully morally justified, conservative in scope, and strictly proportional to legitimate goals pursued.72

III. CIVILIAN IMMUNITY AS GENDER BIAS

If the discrimination principle serves an ethically justified end, it does not necessarily follow that the principle operates fairly in all circumstances. Feminist political and legal theorists have increasingly mounted challenges to the gender neutrality of ius in bello in general and to civilian immunity specifically. The first such critiques came in the 1980s and 1990s, when Judith Gardam argued that the vague duties toward civilians in international humanitarian law compare unfavorably with the more explicit and protective legal rules applying to combatants.73 From this inequality of protection and concern, Gardam deduced a gender bias in international humanitarian law. Observing that all civilians are "protected persons," she interpreted combatants to assume the role of "protectors" (a role, as the introductory quote from General MacArthur illustrates, that combatants have sometimes willingly claimed for themselves) and invoked feminist theory to equate all "protected persons" with persons of female gender, regardless of their sex, leaving all combatants identified with the male gender.74 While Gardam never specifically identifies who attributed these genders to civilians and combatants as classes, she concludes that whatever forces cause these attributions render the principle of discrimination a gendered one.75

More recently, Helen Kinsella has agreed that discourse of gender is vital to identifying and defining combatants and targets,76 and, extending Gardam's theory, Kinsella has claimed that civilian immunity is "dependent upon, not merely described by," gender discourses, and indeed "relies upon" gender "for its very possibility."77 Kinsella seizes on language in the International Committee of the Red Cross's official commentaries to Geneva Convention IV regarding the need for protection of vulnerable and nonthreatening civilians, which she equates to women.78 From this, Kinsella argues that because innocent civilians by definition do not pose a threat to belligerents, and because some feminist theorists insist that women are socially conceived as posing "no potential harm" to enemy combatants, women (as opposed to children, the elderly, and male civilians) are the only quintessential civilians.79 Kinsella concludes that the distinction between combatants and civilians is founded upon gender stereotypes-indeed, she infers that the very concept of civilian immunity presupposes gender stereotyping of women and women alone as innocent and in need of protection.80

Similarly, trying to bridge the ethical theories discussed in Part II above and the feminist theories of Gardam, Judith Hicks Stiehm, and others, Laura Sjoberg has recently argued that "the very foundations of the non-combatant immunity principle are flawed and problematic" because the distinction between combatants and civilians is a gendered one and because the moral responsibility of each "enemy" combatant or civilian for perpetrating an unjust war cannot be gauged without attempting "to understand the composition and political commitments of the people in the opposing society" in making their targeting choices.81

These arguments all share as their common basis the assumption that social stereotypes relegating women (and primarily women) to the status of "protected" persons exist and have legal relevance, so that international law equates combatants with males and civilians with females. They conclude, in each case, that all who do not take up arms, whether male or female, are socially assigned a female gender.82 Male combatants, the "protector" class, are contrasted in each case with female civilians, the "protected" class.83

Although Kinsella adduces no evidence to support her assertion that only women are the quintessential civilians, there is reason to believe that women and children are grouped together and equated with civilians in the international discourse on the protection of civilians from the effects of armed conflicts. Charli Carpenter has published a persuasive account of how intergovernmental and nongovernmental humanitarian organizations and political elites overwhelmingly tend to describe civilians as comprised of women and children, or to characterize these groups as especially "vulnerable," with little or no reference to civilian men.84 For their own political and economic reasons, and pursuant to (or in exploitation of) pre-existing cultural tropes, international elites commonly perpetuate the myth that women and children are more likely than men to be direct targets of attack and are more innocent and vulnerable than men during armed conflicts.85

Recognizing the ubiquity of the equation of women and children with civilians in international discourse does not, however, mean accepting that the discrimination principle itself is a gendered concept. The equation of women and children with civilians may color public perceptions of what kinds of persons suffer most from armed conflict-indeed, the equation may be made with that very purpose in mind86-but it has no more effect on the operational law of war than if the same elites asserted that all combatants were women or children. For purposes of international law, the concern is not that public perceptions of civilians may be inaccurate, but how civilians experience armed conflict and specifically whether the discrimination principle causes female civilians to experience it differently from male civilians.

It is a fact that all but a tiny percentage of combatants in most conflicts are men,87 but international law imposes no limitations on the sex of combatants or civilians.88 If the assignment of female gender and minority age to civilians is evident in political and social discourse (and, tautologically, in some feminist writings)89, it is nowhere to be found in the principle of discrimination itself.

To understand why, it helps to start with the simple fact that most of the ius in bello is doctrinally gender-neutral. With only a few exceptions, international humanitarian law provides the same rights to, and imposes the same obligations on, combatants and civilians regardless of whether they are respectively male or female. There is no exception to the principle of discrimination for female combatants or male civilians, and international law does not limit either men or women to combatant or civilian categories. There are, of course, principles of humanitarian law other than discrimination that apply to women and not to men, but these are either neutral to or enhance, rather than degrade, the level of protection afforded to women.90

Nor does the misleading rhetoric identified by Carpenter reflect the reality of the composition of civilian populations during armed conflict; civilian populations even during armed conflict are usually composed of approximately equal numbers of men and women. Because armed conflict usually involves only a modest percentage of a state's population, a substantial number of civilians are boys and men. Age, disability, need for civilian services, the limited scope of the armed conflict, and other factors almost always keep many men out of combat." Accordingly, the U.N. High Commissioner for Refugees reports that only about half of international refugees are female (either adult or minority).92 The equation of "protector" and "protected" with "male" and "female" respectively may be an international cultural trope, but all relevant elites realize that civilians are composed of approximately equal numbers of males and females. They may speak about the urgent necessity of protecting "women and children" from the effects of armed conflicts, but they are aware that the consequence of doing so is also to protect male civilians from those same effects.

That does not, of course, complete the analysis because the absence of gender bias in the black letter cannot guarantee that the application of the discrimination principle protects women's interests de facto in the same manner or to the same extent that it protects men's interests. A norm can be formally gender-neutral but gendered in conception and gender-biased in practice. One central question respecting international law, then, is whether the rules relating to civilian immunity somehow disadvantage women by subjecting them to disproportionate dangers or deprivations during armed conflict because of their sex. The evidentiary foundation for any assessment of the gender neutrality or bias of civilian immunity in practice must be the evaluation of how female civilians experience war differently from male civilians in terms of the effects of attacks upon them.

The following sections will engage in that assessment by examining three critical ways in which female civilian suffering during armed conflict may be disproportionate to the suffering of male civilians. First, female civilians may experience higher rates of casualties than male civilians from direct or indirect attacks. Second, they may suffer greater deprivations of food, water, shelter, clothing, and medical services than men, and they may be subjected to exploitation by those providing these services on such a scale that the consequences are comparable to an attack. Third, they may be more likely to be subjected to specific forms of attack disproportionately to men, such as sexual violence. If international humanitarian law fails to account for and rectify disproportionate suffering by female civilians in any of these areas, the next question becomes whether that failure is attributable in some way to the international humanitarian law principle of civilian immunity.

A. Intentional Attack and "Collateral Damage "

The incidence rate of civilian deaths relative to combatant deaths may have grown in the last century, but the available evidence indicates that female civilians experience fewer, rather than more, deaths from either direct attack or collateral damage than do their male counterparts. Beginning the analysis with the most severe form of violation of the discrimination principle-intentional attacks on civilians-the evidence strongly contradicts the conclusion that women civilians are subjected to unfavorable gender bias. The U.N.'s Inter-Agency Standing Committee and the United Nations International Children's Emergency Fund (UNICEF) jointly report that adult and adolescent male civilians are much more likely to be intentionally executed, detained, or tortured in armed conflict situations than are female civilians.93 At Sbrenica and in Rwanda, for example, adult male civilians suffered by far the highest rate of execution by armed bands.94 Similarly, the ICRC reports that in some recent conflicts, 96% of detained and 90% of missing civilians have been male.95 During the Cambodian Civil War from 1975 to 1978, for example, while the life expectancies of both genders became extremely low, the life expectancy of females exceeded the life expectancy of males by 50%.%

The pattern of greater harm to male civilians is replicated in a milder form with regard to unintended civilian casualties, also known as collateral damage. For example, the 1990 Gulf War resulted in the deaths of approximately 40,000 Iraqi soldiers (all male), 46,000 male civilians, and 39,000 female civilians.97 During the recent Intifada in Palestine and Israel, 40% of Israeli civilian casualties were female and only 8% of Palestinian civilian casualties were female.98 More generally, the ICRC has reported, since 1999, 65% of the weapons injuries to civilians it has treated have been to adolescent and adult males (excluding males over fifty years of age).99

From these statistics, the view that female civilians suffer disproportionately from the worst forms of harm in armed conflicts-death and weapon injuries-has little to support it. There may be some armed conflicts in which female civilians suffered significantly higher casualty rates than male civilians, but the trend seems rather the opposite. Thus, the claim that it is the "female" gender of civilians that results in civilian suffering is diametrically opposed to the facts. If male civilians are exposed to a significantly greater risk of serious harm than female civilians, the logical conclusion is that if gender plays any role in civilian immunity, it is intentional or unintentional discrimination against civilian males by opposing male combatants. Male civilians appear to be more vulnerable and in need of protection than female civilians. This observation reveals a curious paradox arising from the reasoning of the feminist authors identified earlier. If the traits of vulnerability and need for protection qualify one as "female," then male civilians are in reality closer to having a "female" gender than are female civilians themselves.

B. Deprivation and Exploitation Among Civilians

Casualty figures do not tell all, however. Immunity from attack does not guarantee immunity from disproportionate suffering or increased mortality from the long-term effects of armed conflicts. Except in cases of genocide or ethnic cleansing, most of the suffering caused by armed conflicts takes the less extreme form of deprivations inflicted on civilian populations by their own armed forces or by opposing forces. Here, the case for gender bias is considerably stronger. In the last five years, the effects of armed conflicts on women have been the subject of study by the United Nations and nongovernmental organizations. In 2001, the ICRC published a detailed report on the status of women in armed conflicts.100 The following year, two independent experts under the auspices of the United Nations Development Fund for Women (UNIFEM) published a study of the impact of armed conflicts on women based on interviews with a variety of (exclusively female) refugees and civilians in conflict areas.101 At the same time, the U.N. Secretary General published a report102 on the effect of armed conflicts on women and girls pursuant to a request by the Security Council.103

The ICRC and U.N. Secretary-General's reports document gender discrimination in several areas that are either unaddressed by international humanitarian law or that are formally addressed but under-enforced.104 For example, in some regions, men may receive preferential access to food, clean water, medicine, and other supplies based on cultural, economic, or political factors.105 In at least one recent conflict, discriminatory access to necessities led to highly disproportional death rates among female children.106 A more recent and systematic study has shown that although in almost all states women live longer than men on average both in peace and war, the gender gap slightly decreases on average during armed conflicts.107 The effect, though extremely small,108 is statistically significant, suggesting a marginal sex-discriminatory effect in average mortality during most armed conflicts.

The cause of this discriminatory effect on life expectancy is not, of course, international humanitarian law itself, but rather entrenched cultural and economic gender bias. For example, a much greater decrease in the gender gap in life expectancy may be observed during natural disasters (on average, about 1.5%).109 Most sex-discriminatory practices are prohibited by the 1949 Geneva Conventions and their Additional Protocols110 or international human rights instruments, but international humanitarian law is far from comprehensive in addressing gender bias during armed conflict. Geneva Convention IV requires belligerents to allow the passage of food and medical supplies to civilians through a blockade,111 and Additional Protocol I requires that priority in the distribution of food and medicine be given to expectant and nursing mothers,112 but neither instrument requires egalitarian distribution to females generally (or any other demographic group).113 The silence of international humanitarian law on these matters is regrettable and indicates some lack of attention to women's needs in armed conflicts. The protection of civilians as a class from direct, indiscriminate, or disproportional attacks does not itself guarantee perfect gender equality in the effects of armed conflict.

Nonetheless, such biases are not sufficient evidence that civilian immunity is per se a gendered concept. They may indicate some of the unfortunate consequences of the under-representation of women in negotiating the Geneva Conventions and their Additional Protocols, and they certainly evidence a lamentable lack of attention by states until very recently to the effects of armed conflicts on women as a class. However, the claims addressed here are not that international humanitarian law fails to minister to the needs of female and male civilians completely equally in all instances (a criticism with much evidence to support it), but rather that the law of civilian immunity is fundamentally gender-biased. To say that women's needs as civilians are not fully met by provisions of international humanitarian law relating to relief supplies and refugees does not tend to prove that female civilians would be better off if civilian immunity were abolished and civilians were suddenly made the lawful object of attack. For that proposition, there is no evidence in any of the recent reports on women's experiences in armed conflicts or in the current empirical literature.

C. Sexual Violence as an Attack on Civilians

Sexual violence is by far the most pernicious form of harm suffered by women civilians in armed conflicts. During the recent civil strife in Rwanda, Yugoslavia, Sierra Leone, Uganda, and Sudan, for example, there were mass rapes of female civilians, sometimes deliberately calculated to impregnate the victims.114 Sexual slavery, mass rape, and trafficking of women and girls remain far from unusual in armed conflicts even today.115 The uses of sexual violence are manifold and shocking in their diversity and brutality. Because sexual violence usually takes the form of an attack on civilians, it is relevant to the humanitarian law of civilian immunity.

Some feminist authors have criticized ius in bello for failing to protect women against sexual violence,116 but this is more a criticism of the law's enforcement mechanisms than of the doctrine itself. Rape by combatants has violated the law of war since the nineteenth century at the latest,117 and possibly well before then.118 It is true that none of the early international humanitarian law conventions clearly forbade sexual violence until the 1907 Hague Regulations outlawed rape under the dubious rubric of "violation of family honor."119 But the fourth Geneva Convention of 1949 and the 1977 Additional Protocols make "rape, enforced prostitution, or any form of indecent assault" of women a war crime,120 and these crimes are commonly considered a grave breach of the Conventions.121 More recently, the Statutes and the Rules of Evidence and Procedure of both the ICTY and ICTR provided for prosecution of rape and other sexual assaults as violations of humanitarian law,122 and Article 5(g) of the ICTY's Statute labels rape a crime against humanity.123 Both tribunals have undertaken groundbreaking prosecutions of sexual assaults, including the incitement of sexual violence, under the heading of crimes against humanity.124 Similar decisions have been reached by the Inter-American Court on Human Rights and the European Court of Human Rights.125 The Statute of the Special Court for Sierra Leone establishes jurisdiction over rape, forced prostitution, and other sexual assaults.126 The Special Court has indicted several defendants for rape as a crime against humanity, as well.127 Finally, the Rome Statute of the International Criminal Court includes rape, sexual assault, and other gender crimes among prosecutable crimes against humanity and war crimes.128 It is, therefore, inaccurate to say that ius in hello does not formally protect female civilians against sexual violence.

The other criticism leveled against positive international law respecting sexual violence is that such violence is not clearly a grave breach of the Geneva Conventions and a crime against humanity. As Gardam observes, "it would have been preferable if rape of women had been designated as a grave breach rather than leaving it as a question of interpretation."129 Although this is undoubtedly true, the neglect of clarity, while unfortunate, does not seem to represent radical gender bias in contemporary ius in bello, much less in the concept of civilian immunity. It represents instead a gender bias in the priorities assigned in international law to harm inflicted on women relative to harm inflicted on men. That lacuna has been rectified at least partially by the international statutes and prosecutions described above.

Of course, doctrine may not reflect practice. International humanitarian law is potentially vulnerable to criticism for the failure of the international community systematically to detect and punish sexual violence in armed conflicts at all command levels and to provide assistance and compensation to the victims. That states have historically taken relatively little decisive action to prevent and punish sexual violence in armed conflicts evidences precisely the kind of gender discrimination that has resulted from the historical and continued dearth of women in international diplomacy.130 The fact that sexual violence has long been prohibited as a wartime practice has not prevented the law from suffering neglect and under-enforcement because of the historical worldwide political repression of women. Unlike norms relating to the treatment of prisoners of war or wounded combatants, violations of the prohibition on rape and other forms of sexual violence have not regularly resulted in reprisals or punishment of the offenders and their commanding officers. Prosecutions for rape as a war crime have been sporadic in the past,111 and there has until very recently been no institution for awarding reparations to the victims.132 While ad hoc international and mixed tribunals have prosecuted and will undoubtedly continue to prosecute combatants for some incidents of sexual violence, the large scale of sexual violence in recent armed conflicts suggests that the international community has failed to inculcate in combatants a repulsion to, and fear of certain and severe punishment for, sexual violence. Nor does international humanitarian law provide an effective system for detection of sexual violence through individual complaint procedures.133

These are important criticisms of humanitarian law. In some ways, they overstate the gender bias problem, as many other violations of humanitarian law, including the pillage, torture, mutilation, and outright slaughter of civilians by military forces, have not consistently been punished. In any case, it does not follow from a conclusion that gender bias exists in the under-enforcement of the prohibition on sexual violence that the discrimination principle is a gendered concept, embodies gender bias, or operates to the disadvantage of women. Although sexual violence is a form of intentional attack on civilians that is directed in practice primarily at women, international humanitarian law does prohibit such attacks regardless of the gender of the victim or perpetrator. To the extent that the international community has neglected these norms relative to other norms protecting civilians by hypothesis, a de facto gender bias may exist in a de lege gender-neutral world public order. This observation does not call for the abolition or radical reform of the principles of discrimination, necessity, or proportionality, but rather an adjustment in enforcement priorities.

D. Conclusions

The available evidence indicates that the discrimination principle is not significantly gendered or gender-biased in conception or substance, although there is cause for more research into the question of whether there is gender-bias in terms of priorities and enforcement of the kinds of attacks to which civilians may be subjected (i.e., sexual versus non-sexual) and the distribution of resources among civilians by relief agencies. On the whole, however, the gendered construction of civilian immunity espoused by many international elites appears to operate more to the disadvantage of men than to women."4 The equation of civilian status with the female gender in the social or political spheres does not translate to gender bias against women in the law; international humanitarian law is formally adequate to protect women from direct attack and the major war-related abuses.135

Beyond the lack of evidence adduced by those who would equate gender with combatant status under humanitarian law, there is a more fundamental objection-it leads to no useful policy recommendations. One might conclude-as may be inferred from Gardam's argument-that because international humanitarian law is supposed to protect the "female-gendered" civilians, and because such civilians nonetheless suffer greatly in armed conflicts, that women's interests are best served by strengthening laws that protect civilians.136 But greater protection of civilians is only one possible solution to the alleged gender discrimination, and it is one that does not solve the underlying problem that Gardam and others purport to identify. To eliminate the supposedly gendered nature of humanitarian law, it would be more effective to adopt new ius in bello requiring equal representation of men and women among combatants, thereby achieving gender equality among both combatants and civilians and rendering an equation of gender to combatant status even less plausible than it already is.137 Or, alternatively, the distinction between combatants and civilians might be abolished altogether, with both being equally subject to attack. Either of these approaches would more rigorously address the alleged gender bias in international humanitarian law than Gardam's proposal. But the "women questions" of feminist theory remain unanswered: Will women be better off? Will the horrors of war fall less heavily upon them if civilian immunity is abolished or women are conscripted equally with men? Will they find themselves closer to sharing equal political and economic power with men? No radical reform of civilian immunity, much less its abolition, can achieve these objectives.

IV. DISCRIMINATION, NECESSITY, AND PROPORTIONALITY AS OPERATIONAL NORMS

The third class of critiques of civilian immunity is the most venerable, but also the least metaphysical. These approaches do not question the desirability of distinguishing between combatants and civilians for purposes of determining lawful targets of attack, but rather criticize the rules the international community has settled upon to protect civilians from attack-necessity, proportionality, and the guiding rules to achieve them set forth in Additional Protocol I to the 1949 Geneva Conventions-as ineffectual or counterproductive.

Such critiques are based upon the observation that if modern armed conflicts expose civilians to similar or greater risks than combatants, as is often the case, then the natural conclusion is that the necessity and proportionality principles are either consistently disregarded or are too vague to be useful in making responsible decisions about methods and targets of attack. If the former proposition is accurate, then improved enforcement mechanisms could address the problem-a subject beyond the scope of this Article-but in any case the soundness of the legal concepts remains undisturbed. If the latter proposition is more accurate, the solution is a revision of the legal norms themselves to more efficiently effectuate the purposes of humanitarian law.

In this part, I will consider whether international humanitarian law provides insufficient guidance on the legality of attacks that may harm civilians or civilian property and conclude that the norms of necessity and proportionality, although of limited assistance to military commanders and criminal tribunals, probably offer enough helpful guidance to justify their continued respect. Accepting that necessity and proportionality are useful components of international humanitarian law does not, however, justify their continuation if they create more problems than they resolve. I will, therefore, also consider some ways in which necessity and proportionality may be criticized as hindering rather than helping the protection of civilians during armed conflicts.

A. Ineffectual or Flexible by Design?

Necessity and proportionality, it will be recalled, are the principles defining the lawfulness of attacks that may be expected to cause incidental civilian deaths. As a conventional and customary standard of humanitarian protection, military necessity itself sets no definite limits on the method or targets of attack; the rule merely ensures that the use of force against civilian objects is not entirely gratuitous. The limited protection provided by necessity reflects its origins. Historically, military necessity was invoked to justify acts that otherwise violated customary international law on a theory that the need for military victory trumped all other values, including the safety of large numbers of civilians.138 Only since the 1907 Hague Conventions has military necessity evolved into a constraint rather than an expansive principle."9 Necessity no longer justifies a violation of other rules of ius in bello, such as the prohibition on hostage-taking or proportionality.140 Its protective value is limited to forbidding attacks that endanger civilians and that cannot be justified for any military purpose, such as attacks on undefended localities or the random destruction of civilian property.141

The proportionality principle offers more definite protections to civilians, but it also leaves much to the discretion of belligerents. The most specific proportionalityrelated guidance relates to prohibitions or limitations on the use of designated inhumane and highly indiscriminate weapons142 and attacks on dams, dykes, nuclear power facilities, and other potential sources of uncontrollable destruction.143 Beyond these rules, the parameters of proportionality analysis grow vague and uncertain. The general rule requires military commanders to weigh the "concrete and direct" military advantage gained by an objective or tactic against the expected incidental harm to civilians and destruction of their property.144 More specific rules in Additional Protocol I and various protocols to the 1980 Convention on Conventional Weapons flesh out the standard.145 Article 51(4)(c) of Additional Protocol I prohibits attacks that strike civilian and military targets indiscriminately, while Article 51(5)(a) prohibits attacks that treat "as a single military objective a number of clearly separated and distinct military objectives located in a city, town, village or other area containing a similar concentration of civilians or civilian objects."146 Additional Protocol HI to the Convention on Conventional Weapons similarly provides:

2. It is prohibited in all circumstances to make any military objective located within a concentration of civilians the object of attack by air-delivered incendiary weapons.

3. It is further prohibited to make any military objective located within a concentration of civilians the object of attack by means of incendiary weapons other than air-delivered incendiary weapons, except when such military objective is clearly separated from the concentration of civilians and all feasible precautions are taken with a view to limiting the incendiary effects to the military objective and to avoiding and in any event to minimizing, incidental loss of civilian life, injury to civilians and damage to civilian objects.147

Finally, Article 57 of Additional Protocol I requires military commanders to choose "means and methods of attack" that minimize civilian casualties and dictates that when confronting multiple military objectives conferring a "similar military advantage," they must choose the objective posing the least danger to civilian lives and objects.148

These rules extend substantially more protection to civilians than the naked obligation to refrain from attacking civilians directly. Yet the inclusion of detailed rules of proportionality was opposed by several delegations to the negotiations on the Additional Protocols, not because proportionality limited military options excessively, but rather the opposite. The Hungarian delegate, for example, saw little evidence that the principle of proportionality had been consistently adopted in state practice. In his view, "either the rule was not well established and hence not binding; or it existed and could not be applied in armed conflicts; or it existed and was applied, but the results of its application provided the best argument against it."149 Several other delegations from geographically, culturally, politically, and legally diverse countries, including those of Czechoslovakia, Egypt, Germany, Italy, North Korea, Poland, Romania, and Uganda, similarly thought the principle too subjective to offer any meaningful protections to civilians against perceived military exigencies and proposed excising it from the draft.150 Their argument was well supported by international practice. In 1977 there was not a single well-known instance of prosecution for a violation of proportionality.151 Even today, publicized international or domestic prosecutions for violation of proportionality are exceedingly rare, due at least in part to the difficulty of applying the rule. The Prosecutor's Office of the ICTY, for example, has established a policy of prosecuting violations of proportionality only "in cases where the excessiveness of the incidental damage was obvious."152 To trigger such a prosecution, the relative number of civilian deaths must be so clearly disproportional that no reasonable military commander could argue its legitimacy.153 Conceding so much discretion to belligerents is unlikely to be conducive to a conscientious application of the rule.

The arguments of the delegations opposing the inclusion of proportionality in Additional Protocol I have proved oracular. Today, the observation that necessity and proportionality standards are so nebulous as to render them incapable of providing guidance in most actual combat situations is commonly accepted by international lawyers.154 One common explanation for the lackluster performance of this legal rule in protecting civilians is that proportionality calculations are unavoidably subjective. It has sometimes been observed that, being based on heterogeneous species-military objectives and civilian lives-for which there is no accepted value ratio or relation, such calculations are objectively impossible.155 The calculus becomes still less feasible when the military values are abstract, such as positional superiority or advance occupation. If it is difficult to assess the proportionality of the loss of ? number of civilian lives to a combat tank, assessing the number of lives proportional to occupation of a commanding view of a battlefield or major thoroughfare, or even a decisive military victory, seems doubly abstract.

The subjectivity of the value of a military objective relative to the cost of human lives is compounded by its dependence on the decisionmaker's position in the command chain. Operational officers may view numbers of civilian losses from a different perspective than that of commanding officers high in the chain of command.156 To a platoon commander, the order to obliterate a well-positioned village of one thousand civilians defended by a platoon of fifteen combatants may seem disproportional, while to a division commander, one thousand civilian deaths may seem trivial to the importance of assuming positional superiority before the enemy reinforces the village. Or the reverse may be true-the deaths of even a dozen soldiers under the officer's command to protect a hundred enemy civilians may seem unjustified. The proportionality standard offers no clear basis for impugning any of these judgments, however contradictory to each other in their calculations. Indeed, the travaux préparatoires to Additional Protocol I indicate that although many delegates insisted that the relevant decisionmaker's level should be at high levels of command,157 that matter was intentionally left undefined in the instrument to allow belligerents the option of determining command responsibility,158 increasing the variability and subjectivity of the principle's application.

It is also clear from the travaux of Additional Protocol I that several of the negotiators deliberately sought to keep proportionality standards flexible to avoid constraining their abilities to weigh civilian lives according to their own standards.159 A handful of states went so far as to insist that the rules of proportionality could not constrain a state's right to self-defense at all when the survival of the state's independence was at stake, implying that military victory could legitimately be valued above a potentially unlimited number of civilian lives.160 At other times during the negotiations, proposals were offered to make the rule of proportionality more definite, as when Sweden suggested that proportionality could only justify civilian deaths where the civilians were located in the "immediate" vicinity of the military objective.161 These proposals were typically dismissed without published discussion. It is clear that the negotiators of Additional Protocol I purposefully declined opportunities to increase the potential of the necessity and proportionality principles to protect civilians at the cost of imposing additional constraints on methods of warfare.

The resulting vagueness of the standards renders them susceptible to evasion or marginal observance. Neither necessity nor proportionality imposes a significant obstacle to military commanders acting in a manner that evidences a decided overvaluation of military victory or the lives of their own combatants over civilian lives-even many more civilian lives per combatant lost. The norms might be criticized as "vague" or praised as "flexible," but in any case they represent a final consensus to accept known and substantial risks to civilians as an acceptable price to pay to avoid diminishing the potential to achieve military objectives. Necessity and proportionality were designed, in short, to offer distinctly limited protection to civilians. The fact remains, however, that Additional Protocol I's protections improve substantially on those available to civilians prior to 1977. To criticize proportionality for its vagueness and subjectivity is not, then, equivalent to denying its potential contribution to ameliorating some of the worst effects of armed conflicts on civilians. Proportionality as a legal principle has merits that weigh against letting the best become the enemy of the better. At the very least, the rule improves the prospects of civilians in three ways.

The first and most obvious is that Additional Protocol I contains specific rules and guidance that narrow the range of offensive and defensive options available to military commanders. Indiscriminate terror attacks, carpet bombing, or shelling near civilian population centers, and similar tactics sometimes in use before 1945 are clearly prohibited. More generally, as mentioned earlier, if multiple military options and targets present themselves, the option least harmful to civilians must be selected.

Second, even vague and subjective rules require military commanders to weigh in some manner the effects of their tactical decisions on civilians and civilian objects and to consider alternative means and targets of attack rather than automatically proceeding with the most desirable plan from a purely martial view. Although military commanders will vary in how they strike the balance between risks to civilians and the importance of various military objectives, the necessity and proportionality doctrines at least require them to engage in that analysis rather than assuming the supremacy of military objectives. Valuing the lives of foreign civilians has become an integral part of military planning for all responsible state military organizations.

Finally, the necessity and proportionality doctrines may assist combatants to develop a culture or mentality in which the military role as protector of civilians and civilian objects figures into strategic, tactical, and operational thinking at all levels of command. A military culture that respects civilians and views its own role as encompassing not just military victory but humanitarian ideals may take greater precautions for the safety of civilians and weigh military objectives against dangers to enemy civilians from a more balanced perspective. Proportionality, in effect, creates a value to counterbalance the tendency to value military victory for its own sake.

Given these benefits, the criticisms of the necessity and proportionality principles as unnecessary or harmful operational norms seem misdirected. This is not to say that, given a preference for greater protection of civilians relative to military effectiveness (assuming that trade-off is a necessary one), the necessity and proportionality norms could not be improved in some ways. On the contrary, there is plenty of room for more specific norms strengthening civilian protections. But it does provide good cause to doubt that international humanitarian law would achieve its objectives more efficaciously in the absence of these norms, however vague and difficult to apply they may seem.

B. Are the Norms Counterproductive?

If necessity and proportionality justify their inclusion in the canon of humanitarian law by the three contributions discussed above, the implication that these norms are actually counterproductive seems even less credible. Perhaps the opponents of including proportionality in Additional Protocol I considered that whatever protections it would guarantee to civilians would be offset by some disadvantage. One possibility is that the posited disadvantages of necessity and proportionality have less to do with the content of the law itself than with the indirect effect the rules might have on relevant actors.

It could be argued that international humanitarian law performs a signaling function that conveys harmfully deceptive information to combatants, civilians, and international elites. This critique references the information-disseminating function of the law,162 but argues that necessity and proportionality convey inaccurate rather than accurate information, thus skewing the calculus of rational action. The United States made a very similar argument in seeking to table a draft treaty to end the production of weapons-grade plutonium and uranium as lacking an effective verification mechanism:

Mechanisms and provisions that provide the appearance of effective verification without supplying its reality could be more dangerous than having no explicit provisions for verification. Such mechanisms and provisions could provide a false sense of security, encouraging countries to assume that, because such mechanisms and provisions existed, there would be no need for governments themselves-individually or collectively-to be wary and vigilant against possible violations.163

In the context of proportionality, if the rules of warfare might, for example, convey inaccurate information to civilians about their chances of injury or death based on their decision to stay put or flee approaching armies or air strikes, at least some civilians can be expected to react accordingly. Civilians concerned they will be endangered by military movements of belligerent forces are presumably more likely to evade the approaching attack than civilians convinced that the chance of being subjected to violence is remote. Acting on that belief, they may seek shelter, flee to neutral states or refugee camps, or avoid proximity to military objectives, thereby decreasing their chances of becoming casualties. If, on the other hand, civilian immunity principles deceptively reassure civilians that such self-preservative action is unnecessary because rules of international humanitarian law shield them from danger, it could be argued that the doctrines of necessity and proportionality distort the rational decisionmaking of civilians by leading them to believe they are less likely to be victimized by enemy forces than they are in fact. Acting on these beliefs, they may be less likely to evacuate current battle spaces or regions likely to become imminent battle spaces.164

State officials furnish a second example of actors whose behavior may be influenced by the information-distorting function of the law. There are many reasons why states commonly refrain from intervening (or using intergovernmental organizations like NATO or the United Nations to intervene) to prevent or halt armed conflicts. To the extent that state representatives perceive that an armed conflict is creating or will create a humanitarian crisis, they have some additional incentive to intervene to prevent or promptly terminate the conflict. In contrast, they may perceive little advantage in interfering in armed conflicts fought primarily between foreign combatants and having no substantial effect on civilian populations unless their own national interests are implicated. The political appeal of intervention to prevent the slaughter of "helpless" and "innocent" civilians is more likely to motivate states to act decisively than intervention to stop soldiers from killing other soldiers.165 The available evidence appears to show that, all else being equal, conflicts involving atrocities committed against civilians, as in BosniaHerzegovina, Kosovo, and the Congo in 1998-2003,166 are slightly more likely to elicit a decisive international intervention than what are perceived by political elites (or, often more important, their electors through information reported by the media)167 as primarily military conflicts, such as the 1980-1988 Iran-Iraq War or the 1998-2000 Eritrean-Ethiopian War. Political studies of the causes of intervention in foreign states support this hypothesis.168 This is not to say that humanitarian intervention is generally probable in armed conflicts even of disastrous human consequences, and especially ain non-international conflicts. But in deciding whether to intervene in a foreign armed conflict, conflicts widely perceived as humanitarian catastrophes are somewhat more likely to provoke intervention. To the extent that civilian immunity principles persuade the international community that armed conflicts in general may not be expected to cause widespread death and suffering to defenseless civilians, then, such immunity may be thought to deter states from intervening to prevent or terminate the conflicts.

A final example is supplied by the combatants themselves and their supporting populations. One commentator has praised the proportionality principle as "useful as an acknowledgement of the unfortunate inevitability of incidental civilian casualties in war."169 Alternatively, proportionality may be precisely the opposite. Far from reminding the armed forces and supporting civilian population that civilian casualties are unfortunate or inevitable, humanitarian law may create a mistaken impression that wars can be fought "humanely" and that any resulting civilian casualties were unforeseeable at the time of the decision to resort to armed force. The illusion of relative humanity fostered by civilian immunity and its accompanying principles may in some cases justify in the minds of combatants and their supporting populations a use of armed force that, in its total effects, causes death, destruction and suffering disproportional to the good achieved by the use of force. That is to say, prior to considering the effects of humanitarian law once the decision to use armed force has been made, civilian immunity may actually result in more a frequent resort to armed force for less than compelling purposes-in brief, for recreational warfare. Humanitarian law may in this way work against the peaceful resolution of international disputes.170

Although each of these hypothesized information-distorting effects has interesting implications, none of them seems particularly credible; they are all based on questionable assumptions. One key shared assumption is that the relevant persons are aware of the humanitarian law norms. Although state officials and combatants themselves are expected to be familiar with humanitarian law, this assumption is probably unjustified with respect to civilian populations. There is no automatic mechanism for educating civilians of their rights under the laws of war before or during an armed conflict. If civilians do not flee at the onset of an armed conflict, it is likely to be due to some reason other than a belief that they are protected adequately by international humanitarian law.

Another assumption shared by the arguments is that, by accepting that the information is available, the relevant persons have the ability and the will to rely upon it in some way in their decisionmaking. Civilians often have few self-protective options even if they believe the dangers of nearby armed conflict are grave. Mobile mechanized and aerial warfare have enormously decreased the notice period before a civilian space becomes a battle space. In some cases, civilians will not have time to remove themselves from the scene of combat; in others, exit routes (such as roads and bridges) may be blocked or destroyed before they can evacuate.171 In other cases, civilians simply have no access to safe facilities like refugee camps or bunkers or lack means of transportation away from the combat zone.172 Even when civilians have the ability to evacuate, they face considerable dangers of displacement-the lot of a refugee is rarely free of risk and suffering. In such cases, civilians' beliefs (if any) in the protection offered by international humanitarian law are irrelevant because they have no means of acting on those beliefs. As for combatants, they may be inducted into military service, either by prior contract or by mandatory draft, regardless of their views on whether armed conflict is justified .

Even those who know the rules of warfare and possess the ability to act upon whatever information they convey may have sound reasons for disregarding civilian casualties. What deters international intervention in armed conflicts, for example, is probably less an optimistic belief in the efficacy of humanitarian law than it is the presence or absence of specific political characteristics and consequences of the conflict unrelated to the dangers the conflict poses to civilians.173 The fact that the widespread characterization of an armed conflict as a humanitarian catastrophe makes intervention more likely does not mean that its catastrophic dimensions are enough by themselves to incite international intervention, as should be clear from the numerous disastrous slaughters of civilians in which no intervention or extremely tardy intervention occurred.174

In summary, arguments that the discrimination, necessity, and proportionality principles may create risks to civilians through their information distorting effects remain mostly hypothetical. The force of this critique is especially diminished by the dubious assumption that the relevant actors rely on humanitarian law as a source of information about the consequences of armed conflicts. Even if humanitarian law did exercise some information distorting function, it is far from clear that it would outweigh the benefits the principle provides to civilians even on the most critical view of the civilian immunity principles.

V. BALANCING MILITARY EFFECTIVENESS AND HUMANITARIAN IDEALS

The discrimination principle admittedly does not embody the ideal ethical basis for the regulation of international violence. There are sound arguments for tying individual criminal responsibility to participation in the international use of force conducted immorally or to immoral ends. After all, when a state deploys its military for clearly unjust and illegal ends, such as territorial aggrandizement or naked aggression, a standard of individual responsibility that exonerates combatants from punishment for participating in such a use of armed force seems to divorce international humanitarian law from any theory of ethics that attributes free will and individual moral responsibility to all persons.

The assumptions of free will and individual moral responsibility are, however, merely theories, and encoding them in international law presupposes a veritable babushka of other nested assumptions. Individuals must be assumed (1) to have all the relevant facts about the international situation; (2) to be capable of seeing through misleading statements or lies made by political leaders and the media; (3) to be able to control their feelings of nationalism; and (4) to resist the social and possibly legal pressures to participate in the armed conflict. Given that requiring all potential participants in an armed conflict to judge accurately and objectively its relative justice and the relative moral culpability of others involved in the conflict is not a feasible option for structuring international humanitarian law, the current rules are an appropriate and morally defensible compromise. The discrimination, necessity, and proportionality principles pursue the worthy goal of limiting the destructive effects of armed conflicts to the extent politically feasible. Without dismissing the importance of ethical analysis of the ius ad bellum; this Article rejects the practicality of tying it to ius in bello.

Accepting that the limitation of suffering caused by unavoidable armed conflicts is a morally worthy goal, the germane question is not how international law can hold all individuals responsible for all of their immoral, or questionably moral, decisions in the course of an armed conflict. Rather, it is how to limit the suffering caused by armed conflicts to the minimum necessary given the political constraints on further abridgment of military discretion on the battlefield. Incremental changes to humanitarian law are more likely to achieve the political critical mass necessary to amend the operative treaties than paradigm-shifting proposals to abolish civilian immunity or make participation in any unjust war a crime against humanity. Advocating reform of humanitarian law rather than the abolition of civilian immunity is a rather conventional proposal, and in any case examining the avenues of reform is beyond the scope of this Article. It is not, however, inapt to mention a few promising options with a view to stimulating future discussion.175 The alternatives may be divided provisionally into three categories: (1) making the content of the relevant norms more definite and protective; (2) improving enforcement mechanisms and institutions; and (3) supplementing necessity and proportionality with alternative concepts and rules.

First, improving the definiteness and clarity of the rules underlying military necessity and proportionality (for example by adding specific constraints on attack or defense methods and permissible targets) has at least two important advantages. Greater definiteness increases the enforceability and deterrence value of the norms, as it becomes more difficult for belligerents who recklessly inflict civilian casualties to argue that unduly dangerous tactics fall within a broad margin of appreciation created by the vague language of international norms. After all, if almost any attack can be credibly argued to comply with the rules as necessary and proportional, even when the number of civilian casualties greatly exceeds the number of combatant deaths or the realistic military value of the target attacked, the goal of minimizing the unnecessary suffering caused by armed conflicts remains very distant.

Another advantage to improving definiteness is the option most likely to gather political consensus, as international lawyers can focus on advancing rules likely to achieve maximal prevention of civilian casualties at minimal military disadvantage. It is important to remember that not every state sought more definite rules of proportionality during the 1977 Additional Protocols.176 Proposals to tighten the restrictions on modes of attack for the greater protection of civilians either failed or were kept off the agenda due to the opposition of the United States and other powerful states.177 The most likely avenue of opposition is that detailed rules may be inapposite to the unpredictable constellation of circumstances present in any military encounter. lieutenant Colonel William Fenrick argued in 1982 that more detailed and objective rules elaborating a proportionality standard in the ius in bello "might well be too inflexible to be applied in the myriad circumstances of combat."178 Whether this will create an insuperable problem depends of course on how the rule is crafted, but we can expect the states most likely to resort to force in international relations, including for humanitarian purposes, to oppose proposals for significantly greater definiteness on these grounds.

Keeping in mind these political constraints, a few examples of potential reforms might include (1) a categorical prohibition on scorched earth tactics in foreign territory designed to deny foreign civilian facilities to advancing enemy forces; (2) a dynamic redefinition of military objectives so that the scope of permissible objectives narrows dramatically when a belligerent achieves decisive air superiority or positional superiority;179 (3) a specific rule prohibiting the sacrifice of any significant number of civilian lives in hopes of achieving a highly improbable chance of military success; (4) a legal presumption that a certain ratio of civilian-to-combatant deaths resulted from a disproportional attack; or, (5) as Marco Roscini has suggested, a requirement that civilian casualties and suffering to be factored into the calculus of proportionality must include not only immediate deaths, injuries and destruction, but likely long-term consequences, including starvation, homelessness, disease, and poverty.180 Some of these proposals, if implemented, are likely to reduce civilian casualties more than others, just as some will be more politically and militarily objectionable than others. Because there are a great variety of potentially more definite rules, the proposals can be tailored to balance these considerations as needed.

Greater definiteness does not, however, inevitably address the problem of enforcement. Creating clear rules is one thing; finding the political will in the international community to enforce them in dangerous combat situations is quite another. Without enforcement, more definite rules promise only to make clearer than before that violations of international humanitarian law usually go unpunished-quite the opposite of deterring the disregard of civilian safety. As things stand, the extreme rarity of prosecution and punishment for violations of necessity and proportionality181 attests not only to the vagueness of the norms but to the unreliability of the current enforcement options. These formerly might have included belligerent reprisals, but these have now been severely constrained by Additional Protocol I.182 This leaves the threat of domestic prosecution, prosecution before an international tribunal such as the International Criminal Court or an ad hoc tribunal, or court-martial by the enemy belligerent as the primary means of enforcement. The proliferation of civilian casualties proves at least that the threat of judicial enforcement is insufficiently communicated or credible to military commanders and combatants in many armed conflicts.

Second, as with proposals to increase definiteness, proposals to improve enforcement are likely to encounter political opposition. The U.N. security Council permanent members' highly divergent geopolitical interests and ideologies limit its ability to commit to ensuring the punishment of war criminals. Also deterring reform, the very states whose intervention in armed conflicts is necessary and desirable to prevent civilian catastrophes object to the possibility of prosecutions.183 Systematizing intervention during armed conflicts is even less politically palatable, as evidenced by the refusal of the U.N. security Council thus far to implement the recommendations of the Panel on U.N. Peacekeeping Operations to create a standing rapid-deployment force of peacekeepers.184 While a more incremental improvement in enforcement mechanisms is certainly possible, it seems optimistic to hope that a rise in enforcement of necessity and proportionality norms sufficient to deter many military excesses will result.

As discouraging as it may seem to abandon the hope of more definite and better enforced rules for the protection of civilians in armed conflicts, as Bismarck observed, "politics is the art of the possible." While reformation in this direction may transcend the possible, it is the task of international lawyers to find solutions in the realm of the possible for advancing desirable policy goals such as the diminution of the suffering caused by armed conflicts.

Third, the universe of effective and politically acceptable reform proposals is large indeed, but one especially promising proposal is worth revisiting briefly here. International law already binds states to provide compensation for victims of violations of the laws of war.185 W. Michael Reisman's 1997 proposal, directed toward expanding and modifying this obligation to ameliorate the suffering of civilians posits:

Compensation in humanitarian law should be conceived on two levels, with two measures of damages. First, and without regard to the question of violation of the law of war, belligerents must compensate injured noncombatants or their survivors promptly, in proportion to the degree to which each caused the injuries suffered. Measure of damages here may be determined by general principles of compensation for civil liability on which there is ample agreement and for which there are detailed models. Thus, compensation will be a humanitarian instrument for repair of an injury suffered by an innocent party. The issue is not absolute liability, for a state may substantially reduce, if not eliminate, liability by using more discriminating (and hence more operator-vulnerable) weapons, thereby "internalizing" what would otherwise be collateral damage....

Second, compensation should also be conceived of as a sanction for violations of treaty terms-in short, an international expiation for criminal responsibility. Measure of damages here will be determined in sanction terms by reference to such factors as the gravity of the offense, intentionality, etc.186

This proposal has the substantial advantage of fulfilling the long recognized but rarely respected obligation to compensate for violations of the laws of war "if the case demands."187

Another strength of the proposed approach is that unlike the typical international humanitarian law analysis, it does not force military commanders to make immediate and legally consequential military judgments based on the current nebulous standards of necessity and proportionality. If it is unclear whether an attack meets the standards of necessity and proportionality, the attack may proceed if the belligerent can afford to compensate the civilians, or their heirs, who will suffer from it. Except in cases of grave breaches of humanitarian law, this turns an obscure legal analysis into a fairly concrete economic analysis.

A second advantage of compensation is its focus on remedying the damage caused by disproportional attacks rather than merely deterring the attacks-an area where humanitarian law remains ineffectually. Compensation of such harm is desirable both from the civilian point of view and because it may mitigate any information-distorting function of humanitarian law with respect to belligerent states and their supporting populations. In short, this scheme wtmld increase the costs of armed conflicts in proportion to the suffering inflicted on the opposing civilian population. This would create a financial disincentive to engage in recreational warfare and would provide ammunition, as it were, to opponents of questionably justified armed conflicts.

A third strength of the proposal is that it forces political elites, who are often responsible for pressuring military commanders to conduct hostilities with less respect for the rules of warfare than military culture might prefer, to share in the opprobrium for disproportional attacks. Because political elites must usually answer for the economic consequences of warfare, large civilian losses are more likely to impede their ability to fund armed conflicts and, once the costs become clear, may redound to their political discredit. Under Reisman's scheme, it will be less feasible for the civilian political leadership to demand aggressive military tactics and expect the blame for disproportional civilian losses to fall solely on military commanders.

The attractions of a compensation regime have led some to question whether replacing the concept of war crimes altogether with a compensation regime might vindicate civilian interests more effectively.188 Yet, with all the advantages of compensation, it is preferable whenever possible to deter the excessive loss of civilian lives rather than pay for them. Despite some exuberant economic theories, the monetization of human life and suffering, while often unavoidable, is never the ideal policy option. A duty to compensate will have some preventative value, especially insofar as motivates states to systematize disincentives for violating the laws of war, but the problem of individual action remains. An individual is far more likely to respond to threats of personal punishment and stigmatization than to the possibility of the state bearing some economic expense because of his actions.

The second part of Reisman's proposal addresses punitive deterrence directly. This part presupposes violations of ius in bello are sufficient to give rise to criminal responsibility. In the cases of necessity and proportionality, a finding that states have violated these principles is likely only under current law in the most extreme circumstances, if ever. Thus, the advantages of Reisman's approach do not supplant the need for more definite guidance on the rules of necessity and proportionality if those norms are to be usefully retained.

The basic conclusions defended here are threefold. First, the concept of civilian immunity is impervious to criticisms that it fails to account for the justice of the entire armed conflict as a whole and that it does not consider the ethical position of individual combatants and civilians. The quest for such highly nuanced rules of warfare is quixotic and undermines the goal of minimizing the destruction and suffering caused by armed conflicts that cannot realistically be prevented.

Second, increasingly popular objections to civilian immunity concerning its fundamentally gendered concept have little to recommend them in terms of the actual substance or the implementation of the discrimination doctrine. Apart from social and political rhetoric and the domestic practices of states, civilian immunity as a legal principle is gender neutral. There may be other provisions of international humanitarian law that reflect gender bias, but this observation is hardly equivalent to claiming that, in a gender-free world, civilians would be lawful targets of attack or that the concept of civilian immunity would assume a fundamentally different form.

Third and finally, the legal rules of military necessity and proportionality may be criticized as insufficiently definite and detailed to shield civilians and civilian objects against all the effects of armed conflicts, but they clearly represent an improvement over a basic duty not to attack civilians directly and intentionally. The argument that these norms cause harm by misrepresenting the risks of armed conflicts to civilians is unpersuasive on the facts.

Civilian immunity represents a useful and progressive norm of international law that reflects not the minimization of harm to civilians as a supreme value, but rather a compromise between that goal and the demands of most states to be allowed to pursue military victory without unduly objectionable constraints. As humanitarian values gradually displace militaristic values in the world public order, the balance may shift further toward increased protection of civilians through legal reforms defining necessity and proportionality in greater detail, more systematic enforcement of these norms, alternative approaches to protecting civilians and their property, or some combination of these. In the meantime, the advantages of civilian immunity and its related doctrines more than justify their continued observance as rules of both customary and conventional international law.