Just Warfare Theory and Noncombatant Immunity

Richard J. Arneson^†

Introduction

According to just war theory, a just war is a war against military aggression or the serious intentional threat of military aggression or a war of intervention to protect fundamental human rights. A just war must also satisfy a proportionality norm: the reasonably expected moral gains of commencing and sustaining military intervention must exceed the reasonably expected moral costs.¹ In this tradition, the justice of the war is regarded as a separate issue from the justice of the conduct of the war. Justice in warfare requires, above all, respect for noncombatant immunity.² Those engaged in war are prohibited from deliberately attacking those who are not soldiers, those who are not political leaders of soldiers, and those who are not supplying soldiers with the necessities to carry out warfare. Combatants are those whose activities materially assist the war effort (or in a more narrow construction, those engaged in the war effort).

The right of noncombatant immunity forbids inflicting harm on noncombatants as either an end in itself or as a means to an end. In other words, noncombatants have the right not to be deliberate targets of attack.³ The right of noncombatant immunity, however, condones unintended harm to noncombatants, provided the proportionality norm described earlier is observed. The proportionality norm means that the good effect that one aims to achieve must be greater than the collateral damage to noncombatants that one foresees, but does not intend. Additionally, the proportionality norm requires that there must not be another option available that realizes the same expected benefit but with less expected collateral damage. Noncombatant immunity also extends to combatants who have ceased to be contributors to the war effort, either by surrendering or by becoming incapacitated.⁴

This essay examines the justice in warfare component of the just war theory.⁵ How should we regard the right of noncombatant immunity as just characterized? Common-sense rhetoric tends to regard respect for noncombatant immunity as a litmus test for moral rectitude. Contemporary statements of just war theory reflect this view, as in this succinct formulation: “Terrorism strikes at the defenseless, not at the combatant forces of a social unit, and is thus by nature a crime against humanity.”⁶ Politicians appeal to this sentiment, drawing sharp lines between “us” and “them.” The first Secretary of the Department of Homeland Security, Tom Ridge, observed, “We face a hate-filled, remorseless enemy that takes many forms, hides in many places, and doesn’t distinguish between innocent civilians and military combatants.”⁷ My question is the following: what is the nature and moral force of this distinction?⁸ I shall conclude that we should reject the idea that just warfare theory prescribes an absolute and exception-free condemnation of what we are prone to call terrorist acts.

Although it is not universally accepted, the norm that warring nations should not deliberately attack civilians under any circumstances is deeply engrained in popular moral opinion and embodied in current international law and treaties.⁹ Despite the criticisms developed below, the wide acceptance of these norms may be doing significant good. So why criticize them? My suspicion is that the position that the duty to refrain from attacking civilians in war and military operations is well supported by acceptable, fundamental nonconsequentialist¹⁰ moral principles is unfounded. And like any unfounded moral position, it probably does more harm than most like to think.

The following examination of just war theory represents an internal rather than external critique of this philosophical position. The key question here is not what we should regard as morally acceptable, all things considered. Instead, the question is what is morally acceptable according to the deontological tradition of moral thought, in particular what is the morality of self-defense against aggressive attack? I must give notice at the outset that this account of the morality of self-defense is, to some degree, revisionary. I conclude that just warfare theory, as recently elaborated by Michael Walzer, Paul Ramsey, and other distinguished thinkers, rests on fundamental errors.¹¹

Are individuals who do not threaten violence entitled to immunity from deliberate attack? The proposed line between combatants and noncombatants locates the moral boundary between people who are, and those who are not, materially contributing to the war effort.

In her essay, War and Murder, G. E. M. Anscombe perhaps sharpens the line in a way that narrows the class of permissible targets. Her moral rule is that one must never attack the innocent, and “[w]hat is required, for the people attacked to be non-innocent in the relevant sense, is that they should themselves be engaged in an objectively unjust proceeding which the attacker has the right to make his concern; or—the commonest case—should be unjustly attacking him.”¹² In this sense, innocence and lack of innocence are clearly distinct from innocence and culpability: one can be engaged in an objectively unjust proceeding while being blamelessly ignorant of its unjust character. An individual, therefore, can be morally innocent but not innocent in a sense entitling her to immunity from justified attack.

There remains a question as to what counts as being an engaged participant in a war effort. Must one be (a) doing something that one believes, with good reason, will advance the war effort and also (b) doing something that actually advances it? Or must one also (c) intend to advance the war effort by one’s actions? Or would any one of (a) through (c) by itself be sufficient? Because (a), (b), and (c) may exist to varying degrees, noncombatant status is a sliding rather than binary classification.

Some scholars narrowly draw the distinction, defining combatants as individuals “who are engaged in fighting.”¹³ Emphasizing the perspective of a soldier struggling to discern permissible targets in a just war, C. A. J. Coady suggests that noncombatants are those “prosecuting the harms that are believed to legitimate resort to responsive violence . . . .”¹⁴ This definition implies that individuals who merely provide soldiers with goods required in the course of ordinary life, rather than what they need to perform as soldiers, are noncombatants.¹⁵

Soldiers, of course, are unable to fight without daily nourishment, just as they are unable to fight without bullets. Those supplying the resources enabling soldiers to fight materially contribute to the war effort. One might take the position that the relevant demarcation is a causal notion: when one’s activities, if successful, would significantly increase the probability that group X’s war effort will prevail, one materially contributes to X’s war effort. An individual performing these activities is, therefore, a combatant for purposes of delimiting the just war immunity. This formulation draws the circle broadly. If a nation is mobilized for an exhaustive war effort, almost any productive activity may indirectly aid the war effort. Consider the following hypotheticals. If previously unemployed civilians join the labor force producing goods for home consumption, other workers are freed to make supplies for troops, and if some civilians make propaganda films boosting the resolve of the home work force, again production for military use grows.

A narrow understanding of the idea of the combatant corresponds with a moral line that could be drawn between ways people assist evildoers in ordinary life. A restaurant owner who serves a meal to a known bandit does not materially assist the bandit’s crimes, we suppose. A gun dealer who sells a gun to the same bandit, however, is implicated in the latter’s subsequent crimes, we might think. In short, an individual who provides a bank robber with what she needs in the ordinary course of life is not engaged in the enterprise of bank robbing; however, one who supplies the same bank robber with guns and a map indicating the exact location of a bank vault is so engaged. Similarly, an individual who supplies soldiers the means of warfare is engaged in the enterprise of war, while one who supplies the very same soldiers with food, clothing, and shelter during peacetime is not.

The common-sense way of drawing these lines, nevertheless, collapses under scrutiny. There might be a morally relevant distinction between the intention of the meal provider and the gun provider, but then again, there might not be. The meal provider might intend to facilitate the bandit’s crimes, while the gun owner might not. The act of the meal provider might do more than the act of the gun provider to increase the probability that a crime will be committed, or the amount of wrongful damage done if a crime is committed, or both. The line between what is needed in the ordinary course of life and what is needed specifically to prosecute the problematic crime or military act does not necessarily correspond to any significant moral distinction. Of course, the just war theorist is a deontologist who distinguishes between doing and permitting harm, but both the meal provider and the gun provider are plainly on the doing side of that line. Each might try to excuse his assistance of the bandit by saying, “If I don’t do it, someone else will,” which would mean that his activity is not a “but for” cause of the bandit’s crimes. In either case, the particular circumstances might vindicate or defeat the proffered excuse. Consider a case in which the meal provider and the gun provider are both equally positioned to know that their particular assistance to the bandit will enable him to commit a crime. In this case, both are materially assisting a crime and equally wrong for doing so. The fact that, psychologically, the provision of a gun has a more vivid and salient connection in our minds to the ultimate wrongful acts is not, per se, morally relevant.

What holds in ordinary life also holds in war. We should draw the line broadly, regard as combatants all who materially assist the war, and acknowledge that status as a combatant is a matter of degree. It should be noted that I endorse the view that medical personnel who treat wounded soldiers behind the lines, thereby facilitating their return to battle, materially assist a war effort and hence qualify as combatants. Good consequences may flow from establishing and upholding conventions that stipulate that medical personnel and farmers supplying aid to the troops and other indirect providers of aid should be deemed illegitimate targets of military attack. Considerations of expedience, however, should be sharply distinguished from the considerations that should figure in a proper development of just warfare theory.

II. The Moral Shield Protecting Noncombatants

Thus far I have tried to mark the line between combatants and noncombatants. But the combatant/noncombatant distinction does not coincide with the line that divides those who are morally legitimate targets of violence according to natural law and just war theory, on the one hand, from those who are not legitimate targets, on the other. So at least I shall argue. To see this point, it helps to consider self-defense scenarios.

In a just war, some people perpetrate lethal violence against other people in order to advance the just war cause. The question then arises, on whom may such violence be legitimately perpetrated? By way of example, take the scenario in which one or a few people, in order to save their own lives, perpetrate lethal violence against one or a few people in circumstances where such violence qualifies as self-defense. Who are morally appropriate targets in self-defense scenarios? This self-defense scenario offers a simple case of violence that many will consider permissible, so the judgments we make after reflection about self-defense offer some guidance for how to decide who may be killed in the course of prosecuting a just war.¹⁶

Consider the fault forfeits first principle. In a situation in which there is a wrongful threat to an innocent person’s life, and only killing another person can avert that threat, it is morally better that among those who might die, a person who is significantly and culpably at fault with respect to this situation should be the one who dies. The following examples illustrate the principle.

A. Accommodation

You are being chased by a villainous aggressor who is trying to kill you and will succeed, unless you jump up to a niche where the aggressor cannot follow. There is a bystander standing in this niche who has a perfect right to be there. The bystander sees that if he accommodates you by moving to the back of the niche, you can jump to it and both of you will be safe. If the bystander does not move in this way, you could still save yourself, but only by jumping to the niche and jostling the bystander, causing a fatal fall. The bystander also sees that this is the case. The bystander declines to move to the back of the niche.

The example described does not specify the mental state of the niche occupier, whose failure to accommodate you creates a predicament in which either you or she must die. Perhaps this bystander believes she is entitled to remain there and expects you to respect her right even in your desperate plight. I deny that the bystander plausibly possesses any such right. On the facts described, she is at least grossly negligent, and hence significantly and culpably at fault, by virtue of failing to help you by moving to the back of the niche. The “fault forfeits first” principle dictates that it is morally preferable for the unaccommodating bystander to die rather than you. To save your life, it is then morally permissible for you to jump to the niche, causing the bystander to fall. Would you be seriously culpable for causing the death of the bystander? The fault forfeits first principle responds in the negative: you are entitled to privilege an innocent life over a culpable one. Acting to produce this morally better state of affairs is not acting wrongly and does not render you culpably at fault.

B. Guilty Past

Suppose that Smith is an innocent aggressor currently engaged in attacking you. He is wrongfully trying to inflict lethal violence against you, but he is not culpable in this respect. Perhaps he is acting on the basis of false beliefs, and that if these beliefs were true, the attack would be justified. Moreover, he is not culpable for having these false beliefs. The culpable agent here is Jones, who used deception to induce Smith to form these beliefs. Suppose Jones’ evil plan was to trick Smith into killing you. But now, Jones is no longer doing anything that menaces your life, and we can suppose there is nothing Jones can do that will now annul his deception and remove the threat to your life that Smith’s aggression represents.

You have a right not to be killed in these circumstances—a right that Smith is violating. Two courses of action, either one of which would save your life, are available. You could kill Smith, the innocent aggressor, and thus prevent him from killing you. Or you could kill Jones, which would sufficiently unnerve Smith to incapacitate him. Perhaps Smith, positioned so that he sees Jones, will witness your act of killing Jones and, as Smith is very emotionally attached to Jones, Smith will be prevented from acting effectively, thereby eliminating the threat against your life. These are the sole life-saving responses available.

Any plausible theory of justifiable homicide would approve killing Jones to save your life. The pertinent facts are that Jones has contributed to bringing about the situation in which your life is in peril, Jones is morally culpable for doing so, and your killing Jones would remove the peril against your life. Even though Jones is not currently acting in a way that menaces your life and Smith is, Jones’ moral guilt and his causal responsibility for your peril jointly negate his moral immunity from harm. This claim does not settle the interesting question of whether it would be morally acceptable to kill Smith in self-defense if the option of killing Jones were not available. The principle of fault forfeits first, applied to this sort of case, justifies the judgment endorsed above.

The lesson of the self-defense case is readily applicable to the case of just warfare. Consider a spy working for the Allies in Germany during World War II. It turns out that to fulfill his mission the spy needs to kill someone. There are two possibly useful homicides, either of which would contribute equally to the spy’s mission. One is to kill a young soldier raised in a culture that trained young men of ordinary sensibilities not to question the civil authority. The young soldier believes he is doing the right thing by serving in the German military. He is not blameworthy for having this belief, so he is morally not culpable for his soldiering activity. The other possibility would target a civilian, a noncombatant not engaged in any war-enabling activity. But this civilian is a fervent Nazi of evil mind. This noncombatant worked ardently to facilitate Hitler’s rise to power and the consolidation of Nazi rule. I submit that you are morally obligated to kill the culpable civilian rather than the non-culpable combatant.

The case of a justified military strike against noncombatants, as described above, involves three factors: (i) the moral culpability of the noncombatant target; (ii) the impact of killing the noncombatant on prosecution of the just war cause; and (iii) the noncombatant’s historic contribution to initiating and sustaining the unjust war effort.¹⁷ Would the presence of only two of these factors justify attacking noncombatants? Imagine a potential noncombatant target who culpably endorses the unjust war effort. The noncombatant is placed so that killing her would advance the just war effort; however, she neither contributed to the initiation of the unjust war nor assisted in sustaining it. Again we can consider how we should respond to analogues of this decision problem that arise in simple self-defense scenarios.

C. Guilty Bystander Trying to Inflict Harm

You are driving up a narrow mountain road and see Evil Aggressor approaching in a large armored vehicle. Evil Aggressor intends to murder you in a head-on collision. Your only recourse is to drive onto the shoulder of the road. The shoulder happens to be occupied by an admirer of Evil Aggressor, Guilty Bystander, who is cheering and gloating at the prospect of your demise. Let us suppose it is clear that Guilty Bystander just happens to be there, and that he has not deliberately chosen to occupy the site you now need for your survival. Furthermore, he cannot maneuver to enable both of you to share the shoulder safely. You must either drive onto the shoulder, killing Guilty Bystander, or be killed by Evil Aggressor. Moreover, although the admirer is powerless concretely to threaten your life, he is doing his best to harm you. He is throwing snowballs to distract you and hasten your demise. Assume his efforts are entirely futile. Still, we might think that the combination of wrongfully taking pleasure at your anticipated demise and acting with evil intent, though entirely impotently, to facilitate your demise suffices to render this guilty bystander significantly and culpably at fault with respect to your mortal plight. Thus, the fault forfeits first doctrine would justify driving onto the shoulder, killing the Guilty Bystander, to save your own life.

D. Guilty Bystander Disposed to Inflict Harm

We might imagine a variant of this case involving an even more attenuated connection between the bystander’s conduct and your injuries. First, suppose that the case is as described above, except that the Guilty Bystander is not trying actively to harm you now but is disposed to harm you right now if he could. I suppose that the fault forfeits first doctrine also applies here because of Guilty Bystander’s evil intent. It is morally preferable that the guilty bystander dies rather than you and killing Guilty Bystander to save your own life seems justified. I would add that in both of the preceding hypotheticals, Guilty Bystander’s evil attempt, or evil disposition, would suffice to render the bystander significantly and culpably at fault in a way that involves forfeiture, in this context, of his right not to be killed.

E. Guilty Bystander Exulting in Anticipated Evil

The next case for consideration eliminates the element of wrongful intent to cause harm. The guilty bystander is merely guilty of possessing a wrongfully positive attitude toward your imminent demise. He may cheer the Evil Aggressor, take sadistic satisfaction in contemplating your wrongful death, or exult in the triumph of evil. In this case, the guilty bystander is purely a bystander. He has no opportunity or intent to cause harm. Presented with this sort of example, many deontologists would insist that the bystander has not forfeited his right to be free from harm, and that you, the innocent driver, are forbidden to harm the bystander, even to save your life.

I disagree. I should add that the fault forfeits first principle commits its proponents to the claim that in a situation in which someone must die, it is morally better that one who is significantly and culpably at fault with respect to this situation should die rather than any non-culpable person who is available to die instead. This holds even when the significantly culpable individual neither causes, threatens, attempts nor is disposed to cause any concrete harm. At least this is so if you can be seriously at fault regarding a situation even though your fault involves neither the violation of anyone’s rights nor any wrongful agency aimed at harm. A wrongful attitude toward the evil that others are perpetrating or threatening can be sufficient to negate protection. This position, while controversial, strikes me as correct. Merely taking malicious pleasure in the misfortunes of others does not establish sufficient culpability to dissolve the shield that the status of mere bystander confers. But if one varies the case by making the gloating of the guilty bystander increasingly malign, the judgment that it is wrong to harm him in order to save the innocent becomes attenuated and eventually dissipates altogether.¹⁸ (Imagine a person whose life activity entirely revolves around celebratory rehearsal in thought of horrific immoral acts, such as torture-murder of the innocent.)

F. Fault Forfeits First Doctrine in Just Warfare

The principle of fault forfeits first carries over to the issue of determining the morally preferred targets of violence in the course of prosecuting a just war. If one has a sufficiently important moral cause to justify waging war against the forces opposing that cause, and if killing someone would also sufficiently advance that cause to justify a killing, it would be morally preferable to kill a target significantly culpable with respect to the war rather than an innocent person.

In many cases, one can only advance the just war by killing enemy combatants, whether or not they are culpable. In many cases, when one could kill either combatants or noncombatants to advance the just cause, the combatants will be more culpable, or at least no less culpable, than the noncombatants. But if one is fighting combatants who are not culpable, and if the noncombatants are guilty bystanders in any of the ways detailed above, it will be morally preferable to kill noncombatants rather than combatants to gain a comparative advantage for the just cause.