Published in The Monist, Vol. 57, No. 4 (1973): 527-550.
THE KILLING OF THE INNOCENT
JEFFRIE G. MURPHY
After war has been begun, and during the whole period thereof up to the attainment of victory, it is just to visit upon the enemy all losses which may seem necessary either for obtaining satisfaction or for securing victory, provided that these losses do not involve intrinsic injury to innocent persons, which would be in itself an evil.
Suarez, The Three Theological Virtues
Fight on land and on sea
All men want to be free
If they don't, never mind
We'll abolish all mankind
Peter Weiss, Marat/Sade
Murder, some may suggest, is to be defined as the intentional and uncoerced killing of the innocent; and it is true by definition that murder is wrong. Yet wars, particularly modern wars, seem to require the killing of the innocent, e.g., through antimorale terror bombing. Therefore war (at least modern war) must be wrong.
The above line of argument has a certain plausibility and seems to lie behind much philosophical and theological discussion of such problems as the Just War and the nature of war crimes.1 If accepted in full, it seems to entail the immorality of war (i.e., the position of pacifism) and the moral blameworthiness of those who participate in war (i.e., warmakers and uncoerced soldiers are all murderers). To avoid these consequences, some writers will challenge some part of the argument by maintaining
- that there are no innocents in war or
- that modern war does not in fact require the killing the innocent or
- that war involves the suspension of moral considerations and thus stands outside the domain of moral criticism entirely or
- that contributing to the death of innocents is morally blameless so long as it is only foreseen but not intended by those involved in bringing it about (the Catholic principle of the Double Effect) or
- that the prohibition against killing the innocent is only prima facie2 and can be overridden by even more important moral requirements, e.g., the defense of freedom.
In this paper I want to come to terms with at least some of the important issues raised by the killing of innocents in time of war. The issues I shall focus on are the following:
- What does it mean, in a context of war, to describe an individual as innocent?
- Why is it morally wrong to kill individuals so described?
- Is the moral wrongness merely prima facie (i.e., subject to being overridden by other, more weighty, moral considerations) or is it absolute?
I shall avoid making any final judgments on the morality of war (modern wars or otherwise) since such judgments would involve not just philosophical claims but also empirical claims, e.g., the claim that in fact modern wars cannot be waged so as to avoid the killing of the innocent. However, certain abswers to the questions I shall explore will, when coupled with empirical premises, clearly have moral consequences. For example: If the killing of' innocents is absolutely and not just prima facie immoral (a philosophical claim), and if modern wars necessarily involve the killing of innocents (an empirical claim), then war for modern times must be absolutely condemned, i.e., pacifism is the required moral posture for the twentieth century. I shall, leave it to sociologists and military historians, however, to supply the correct factual premises.
I shall, then, mainly be concerned to seek answers to the above three questions. However, before launching into my discussion, I want to make one thing clear. Just because I am focusing upon the killing of innocents, it should not be thought that I am confident that it is right to kill the guilty and thus that it is only the killing of the innocent that stands in need of analysis and justification. I have no such confidence. I am starting with what any man who is not a moral imbecile must admit to be, at the very least, an obvious prima facie objection to indiscriminate war. In so doing, I do not mean to suggest that there may not be strong nonobvious objections to be made against discriminate war.
THE CONCEPT OF INNOCENCE
The notions of innocence and guilt seem most at home in a legal context and, somewhat less comfortably, in a moral context. Legally, a man is innocent if he is not guilty, i.e. if he has not engaged in conduct explicitly prohibited by rules of the criminal law. A man may be regarded as morally innocent if his actions do not result from a mental state (e.g., malice) or a character defect (e.g., negligence) which we regard as morally blameworthy. In any civilized system of criminal law, of course, there will be a close connection between legal guilt and innocence and moral guilt and innocence, e.g., murder in the criminal law has as one of its material or defining elements the blameworthy mental state (mens rea) of "malice aforethought." But this close connection does not show that the legal and moral concepts are not different. The existence of strict liability criminal statutes is sufficient to show that they are different. Under a strict liability statute, a man can be guilty of a criminal offense without having, at the time of his action, any blameworthy mental state or character defect, not even negligence.3 However, the notion of strict moral responsibility makes little sense; for an inquiry into moral responsibility for the most part just is an inquiry into such matters as the agent's motives, intentions, beliefs, etc.4 Also, the issue of legal responsibility is much more easily determinable than that of moral responsibility. For example: It is noncontroversial that negligence can make one legally responsible. Anyone who doubts this may simply be given a reading assignment in any number of penal codes.5 But whether or not negligence is a mental state or a character defect for which one is morally responsible is a matter about which reasonable men can disagree. No reading assignment or simple inquiry into "the facts" will lay this worry to rest.6
Now our reasonably comfortable ability to operate with these concepts of guilt and innocence leaves us when we attempt to apply them to the context of war. Of course, the legal notions will have application in a limited number of cases, i.e., with respect to those who are legally war criminals under international law. But this will by no means illuminate the majority of cases. For example: Those who have written on the topic of protecting innocents in war would not want to regard the killing of an enemy soldier engaged in an attack against a fortified position as a case of killing the innocent. He is surely, in the right sense (whatever that is), among the guilty (or, at least, among the noninnocent) and is thus a fitting object for violent death. But he is in no sense legally guilty. There are no rules of international law prohibiting what he is doing; and, even if such rules were created, they would surely not involve the setting up of a random collection of soldiers from the other side to act as judges and executioners of this law. Thus the legal notions of guilt and innocence do not serve us well here.
What, then, about moral guilt or innocence? Even to make this suggestion slausible in the context of war, we surely have to attempt to narrow it down to moral innocence or guilt of the war or of something within the war -- not just moral innocence of guilt simpliciter. That is, we surely do not want to say that if a bomb falls (say) on a man with a self-deceiving morally impure leart who is a civilian behind the lines that this is not, in the relevant sense, a case of killing an innocent. Similarly, I think it would be odd for us to want to say that if a soldier with a morally admirable character is killed in action that this is a case of killing an innocent and is to be condemned on those grounds. If we take this line, it would seem that national leaders should attempt to make some investigation of the motives and characters of both soldiers and civilians and kill the unjust among both classes and spare the
just. (Only babes in arms would be clearly protected.) Now this sort of judgment, typically thought to be reserved for God if for anyone, is surely a very disquieting thing if advocated for generals and other war leaders. Thus the notions of moral innocence and guilt simpliciter must be dropped in this context.
Suppose, then, we try to make use of the notions of moral innocence of the war or moral guilt of the war (or of something within the war). Even I here we find serious problems. Consider the octogenarian civilian in Dresden who is an avid supporter of Hitler's war effort (pays taxes gladly, supports warmongering political rallies, etc.) and contrast his case with that of the poor, frightened, pacifist frontline soldier who is only where he is because of duress and who intends always to fire over the heads of the enemy. It seems reasonable to say that the former is much more morally guilty of the war than the latter; and yet most writers on the topic would regard killing the former, but not the latter, as a case of killing an innocent.
What all this suggests is that the classical worry about protecting the innocent is really a worry about protecting noncombatants. And thus the distinction between combatants and noncombatants is what needs to be illucidated. Frontline soldiers are clearly combatants; babes in arms clearly are not. And we know this without judging their respective moral and legal guilt or innocence. And thus the worry, then, is the following: Under what circumstances is an individual truly a combatant? Wars may be viewed as games (terrible ones of course) between enemies or opponents. Who, then, is an enemy or opponent?
One suggestion for defining a combatant might be the following: Only soldiers engaged in fighting are combatants. But this does not seem adequate. For if killing an enemy soldier is right, then it would also seem to be right to kill the man who orders him to the frontline. If anything, the case for killing (say) a general seems better, since the soldier is presumably simply acting in some sense as his agent, i.e., the general kills through him. Perhaps the way to put the point, then, is as follows: The enemy is represented by those who are engaged in an attempt to destroy you.7 And thus all frontline combat soldiers (though not prisoners, or soldiers on leave, or wounded soldiers, or chaplains, or medics) are enemies and all who issue orders for destruction are enemies. Thus we might try the following: Combatants are those anywhere within the chain of command or responsibility -- from bottom to top. If this is correct, then a carefully planned attack on the seat of government, intended to destroy those civilians (and only those) directing the war effort, would not be a case of killing noncombatants or, in the relevant sense, innocents.
But what is a chain of command or responsibility? It would be wrong to regard it solely as a causal chain, though it is at least that. That is, the notion of responsibility has to be stronger than that expressed in the sentence "The slippery pavement was responsible for the accident." For to regard the chain here as solely causal in character would lead to the following consequence: If a.combatant is understood solely as one who performs an action which is a causally necessary condition for the waging of war, then the following are going to be combatants: farmers, employees at a city water works, and anyone who pays taxes. Obviously a country cannot wage war if there is no food, no management of the basic affairs of its cities, and no money to pay for it. And of course the list of persons "responsible" for the war in this sense could be greatly extended. But if all these persons are in the class of combatants, then the rule "protect noncombatants" is going to amount to little more than "protect babies and the senile." But one would, I think, have more ambition for it than that, e.g., one would hope that such a rule would protect housewives even if it is true that they "help" the war effort by writing consoling letters to their soldier husbands and by feeding them and providing them with emotional and sexual relief when they are home on leave. Thus I think that it is wrong to regard the notion of chain here as merely causal in character.
What kind of chain, then, is it? Let us call it a chain of agency. What I mean by this is that the links of the chain (like the links between motives and actions) are held together logically and not merely causally, i.e., all held together, in this case, under the notion of who it is that is engaged in an attempt to destroy you. The farmer qua farmer is, like the general, performing actions which are causally necessary conditions for your destruction; but, unlike the general, he is not necessarily engaged in an attempt to destroy you. Perhaps the point can better be put in this way: The farmer's role bears a contingent connection to the war effort whereas the general's role bears a necessary connection to the war effort, i.e., his function, unlike the farmer's, is not logically separable from the waging of war. Or, following Thomas Nagel,8 the point can perhaps be put in yet another way: The farmer is aiding the soldier qua human being whereas the general is aiding the soldier qua soldier or fighting man. And since your enemy is the soldier qua soldier, and not qua human being, we have grounds for letting the farmer off. If we think of a justified war as one of self-defense,9 then we must ask the question "Who can be said to be attacking us such that we need to defend ourselves against him? Viewed in this way, the farmer seems an unlikely candidate for combat status.
This analysis does, of course, leave us with borderline cases. But, since there are borderline cases, this is a virtue of the analysis so long as it captures just the right ones. Consider workers in a munitions factory. Are they or are they not combatants? At least with certain munitions factories (making only bombs, say) it is certainly going to be odd to claim that their activities bear only a contingent connection to the war effort. What they make, unlike food, certainly supports the fighting man qua fighting man and not qua human being. Thus I should be inclined to say that they are properly to be regarded as combatants and thus properly subject to attack. But what about workers in munitions factories that only in part supply the war effort, e.g., they make rifles both for soldiers and for hunters? Or workers in nonmunitions factories that do make some war products, e.g., workers in companies, like Dow. Chemical, which make both Saran Wrap and Napalm? Or workers in ball bearing factories or oil refineries, some of their product going to war machines and some not? Here, I submit, we do have genuine borderline cases. And with respect to these, what should we do? I should hope that reasonable men would accept that the burden of proof lies on those claiming that a particular group of persons are combatants and properly vulnerable. I should hope that men would accept, along with the famous principle in the criminal law, the principle "noncombatant until proven otherwise" and would attempt to look at the particular facts of each case as carefully and disinterestedly as possible. I say that I hope this, not that I expect it.
Who, then, is a combatant? I shall answer this question from the point of view of one who believes that the only legitimate defense for war is self-defense.10 It is, in this context, important to remember that one may legitimately plead self-defense even if one's belief that one's life is being threatened is false. The only requirement is that the belief be reasonable given the evidence that is available. If a man comes to my door with a toy pistol and says, pointing the pistol at me, "Prepare to meet your Maker for your time has come," I act in my self-defense if I kill him even if he was joking so long as my belief was reasonable, i.e., I had no way of knowing that the gun was a toy or that he was joking. Thus: combatants may be viewed as all those in the territory or allied territory of the enemy of whom it is reasonable to believe that they are engaged in an attempt to destroy you.
What about our Dresden octogenarian? Is he a combatant on this analysis? Since he does not act on authority, it is at least prima facie odd to regard him as part of a chain of command literally construed -- the concept of command being most at home in a context of authority. He does not, of course, have much to do with the war effort; and so we might find his claim that he is "helping to defeat the Americans" quaint on purely factual grounds. And yet none of this prevents its being true that he can properly be said to be engaged in an attempt to destroy the enemy. For people can attempt even the impossible so long as they do not know it is impossible. Thus I am prepared to say of him that he is, in fact, engaged in an attempt to destroy the enemy. But I would still say that killing him would count as a case of killing a noncombatant for the following reason: that the concept of attempt here is to be applied, not from the agent's point of view, but from the point of view of the spectator who proposes to plead self-defense in defense of his acts of killing. Combatants are all those who may reasonably be regarded as engaged in an attempt to destroy you. This belief is reasonable (though false) in the case of the frontline soldier who plans always to shoot over the heads of the enemy and unreasonable (even if true) in the case of our octogenarian. It would be quite unreasonable to plan a bombing raid on a nonmilitary and nonindustrial city like Dresden and say, in defense of the raid, that you are just protecting yourself or your country from all those warmongering civilians who are attempting to destroy you. For making such a judgment imposes upon you a burden of proof which, given the circumstances of war, you could not satisfy. You probably could not get any evidence for your claim. You certainly could not get what the law calls a "preponderance of the evidence" -- much less "proof beyond a reasonable doubt."
Combatants, then, are all those of whom it is reasonable to believe that they are engaged in an attempt at your destruction. Noncombatants are all those of whom it is not reasonable to believe this. Having the distinction, we must now inquire into its moral importance.
WHY IS IT WRONG TO KILL THE INNOCENT?
From what I have said so far, it should be obvious that this question cannot adequately be answered unless one specifies the sense of innocence that one has in mind. First, with respect to moral innocence, we can take babes in arms as paradigms. Here I should argue that no reasons can be given for why it is wrong to kill babies: neither are any reasons needed. If anything can be taken as a brute datum for moral philosophy, surely the principle "Do not kill innocent babies" is a very good candidate -- much more plausible for an ethical primitive than, say, "promote your self-interest" or "maximize the general utility" (other candidates that have been offered for ethical primitives). The person who cannot just see that there is something evil about killing babies could not, I suspect, be made to see anything else about morality and thus could not understand any reasons that one might attempt to give. And any "ethical" theory which entailed that there is nothing wrong at all with killing babies would surely deserve to be rejected on the basis of this counterexample alone. Miss G. E. M. Anscombe puts the point in the following way:If someone really thinks, in advance, that it is open to question whether such an action as procuring the judicial execution of the innocent should be quite excluded from consideration -- I do not want to argue with him; he shows a corrupt mind.11
Consider an example from literature: If a reader of Melville's Billy Budd cannot see why Captain Vere is deeply troubled because he has a legal obligation to execute Billy, a morally innocent (thought legally guilty) man, what would one say to him? What could one do except repeat, in a louder tone of voice perhaps, "But, don't you see, Billy is morally innocent"! What I am saying is that, though it is reasonable to expect ethical theories to correct some of our moral intuitions or pretheoretical convictions, we cannot reasonably expect these theories to correct all of them -- since there is a class of such intuitions or convictions which makes ethical theory itself possible and testable.12 Kant, I think, was perhaps one of the first to begin to see this when he argued that utilitarianism cannot be correct because it cannot account for the obvious wrongness of making slaves of people, punishing the innocent, etc.
Now with respect to legal innocence (when this does not, as it often does, overlap moral innocence) a set of reasons for not killing (or punishing) the innocent can be given. These reasons will be in terms of utility (it makes people insecure to live in a society which punishes them whether they obey the rules or not), justice (it is not fair to punish people who conform their conduct to the rules), and general observations about the purposes and techniques of the criminal law as a system of control through rules. As John Rawls has argued,13 a system of criminal law can be viewed minimally as a price system for controlling conduct (i.e., the rules state a price that has to be paid for certain bits of behavior in the hope that most people will regard the price as too high to pay). And obviously this sort of system is going to work only if the price is generally charged when and only when the "commodity" (i.e., the behavior) is "purchased."
When we move to war, however, difficulties begin to present themselves. We have seen that what people have had in mind when they have argued for the protection of innocents in war was neither clearly moral innocence nor clearly legal innocence. Rather it was noncombatant protection. But when we put the principle as "Do not kill noncombatants" we lose the background of moral and legal thinking which makes the principle seem plausible when formulated in terms of innocence.
Why, then, should we worry about killing noncombatants and think it wrong to do so -- especially when we realize that among the noncombatants there will be some, at any rate, who are morally and/or legally guilty of various things and that among the combatants there will be those who are morally and/or legally innocent?
What I suggest is the following: If one believes (as I do) that the only even remotely plausible justification for war is self-defense, then one must in waging war confine one's hostility to those against whom one is defending oneself, i.e., those in the (both causal and logical) chain of command or responsibility or agency, all those who can reasonably be regarded as engaged in an attempt to destroy you. If one does not do this, then one cannot be said merely to be defending oneself. And insofar as one is not defending oneself, then one acts immorally in killing one's fellow human beings. The enemy can plausibly be expanded to include all those who are "criminal" accomplices -- those who, in Judge Learned Hand's phrase, have a "stake in the venture."14 But it cannot be expanded to include all those who, like farmers, merely perform actions causally necessary for the attack -- just as in domestic law I cannot plead self-defense if I kill the one (e.g., the wife or mother) who feeds the man who is engaged in an attempt to kill me.15
In passing, I should note why I described my position as weakly as I did -- namely, that self-defense is the only remotely plausible defense of war. I do this because, with respect to nations, the whole idea of self-defense is strongly in need of analysis. What, for example, is it for a state to die or to be threatened with death? Can nations, like individuals, fear death and act compulsively on the basis of that fear? And, insofar as the death of a state is not identical with the deaths of individual human beings, why is the death of a state a morally bad thing? Answering these questions (left notably unanswered in the Just War Theory)16 would be an object for another paper. I mention them here to point up latent problems and to further explain my unwillingness, expressed earlier, to affirm that even the killing of the "guilty" (combatants) is morally justified. I am certainly inclined to suspect the Just War Theory's willingness to regard war as an activity giving rise to autonomous moral problems (i.e., problems solely about war and not reducible to ordinary moral problems) and thus to make moral capital out of suspect notions like "national honor" and "national interests" and "national self-defense." Though I am obviously inclined to regard the concept of self-defense as having an important application in the context of war, I am sceptical that the "self" to be legitimately defended must always be the nation or state. It is at least worth considering the possibility that the only moral problems arising in war are the oldest and most common and most important -- namely, are human beings being hurt and killed, who are they, and why are they?
IS KILLING THE INNOCENT ABSOLUTELY IMMORAL?
That is, is killing the innocent a prima facie immorality that can be overridden by other, more weighty, moral requirements or is it absolutely immoral, i.e., incapable of being overridden by any other moral requirements? Miss Anscombe17 holds that the prohibition against killing the innocent is absolute in this sense. And one would suspect that she would echo Kant's sentiments that we should do no injustice though the heavens fall. Unfortunately, she tends simply to assert her position rather than argue for it and so fails to come to terms with the worry that might bother anti-Kantians -- namely, does it not matter upon whom the heavens fall? I agree with Miss Anscombe that no argument can be given to demonstrate that there is something wrong with killing babies, but it does not follow from this alone that this "something" is not capable of being overridden by another "something" -- saving the lives of even more babies, perhaps.
Now in trying to come to grips with this issue, I propose to start with the case of killing babies as a clear example of killing creatures innocent in every possible sense. If a case can be made out that it is sometimes right to kill them, then I assume it will follow a fortiori that it is sometimes right to kill those who may be innocent in a less rich sense, e.g., merely noncombatant. Of course if it is not ever right to kill babies, this will not in itself show that it is not ever right to kill noncombatants; for it may be the special kind of innocence found in babies (but not necessarily in noncombatants) which protects them.
First of all, we need to ask ourselves the question "What is it deliberately to kill a baby?" I am certain that Miss Anscombe does not mean the following: that one deliberately kills a baby whenever one pursues a policy that one knows will result in the deaths of some babies. If anyone meant this, then that person would have to regard the construction of highways as absolutely immoral. For it is a statistical fact that on every completed highway a certain number of babies meet their deaths in accidents. Yet normally we do not regard this as a moral case against highways or as a moral proof that highway engineers are murderers. What is done is to weigh the social value of a highway against the knowledge that some deaths will occur and judge that the former outweighs the latter. (If we did not make this judgment, and if this judgment was not reasonable, then we would be acting immorally. For example: Suppose we let people blast with dynamite whenever and wherever they felt like it -- conduct with little or no social value -- in spite of our knowledge that this would result in many children dying.) Of course, situations comparable to the highway example arise in war. For example, consider the pinpoint bombing of a military installation in a war reasonably believed to be necessary and just coupled with the knowledge that bombs will occasionally go off target and that occasionally a wife, with baby in arms, may be visiting her husband on the base.
The natural way to interpret Miss Anscombe's view is as follows: One kills a baby deliberately either when one (a) brings about the death of a baby as one's final purpose or (b) brings about the death of a baby as a means to one's final purpose.18 Since war is hardly to be regarded as motivated by fetishistic infanticide, it is (b) which is crucial. And (b) lets off the highway engineer. For the highway engineer, whom we do not want to regard as a murderer, will not be a murderer on (b). The deaths of babies in highway accidents are in no sense means to the socially useful goal of good transportation but are rather accidental byproducts. That is, highway transportation is not furthered by these deaths but is, if anything, hindered by them. And so it is unreasonable to suppose that highway engineers desire the deaths of babies because they want to use their deaths as a means to their goal of transportation. Similarly with the accidental deaths of babies resulting from a pinpoint bombing raid on a military installation.
But consider the following kind of case: One knows that one is fighting a war in defense of civilization itself. (It was not unreasonable to regard the war against Hitler in such terms.)19 Suppose also that one knows that the only way (causally) to bring a power like Hitler's to a collapse is to undercut his support among the German people by achieving their total demoralization. Further suppose that one knows that the only way to do this is by the obliteration terror bombing of civilian centers (e.g., Dresden) so that, by killing many German babies among others, one can create a desire on the part of the German people to abandon the venture. I am not suggesting that we could ever in fact know these things (they might be false) or that we should ever even let ourselves believe such things.20 But my worry here is one of principle -- namely, assuming the factual situation is as described, would it be wrong on principle to initiate a campaign of antimorale obliteration bombing? Here I take it that Miss Anscombe will say that, even in such circumstances, those making a decision to initiate such a campaign are murderers. For they will be deliberately killing the innocent as a means to their goal and that (no matter how good the goal) is absolutely wrong.
But why? Miss Anscombe fails to bring fully into the open a latent issue that is absolutely crucial here -- namely, are we as morally responsible for our omissions (e.g., failing to save lives) as for our commissions (e.g., killing people)? Of course there are differences between the two expressions, but are they morally relevant differences? If our basic value here is the sanctity of life, or the sanctity of innocents, or the sanctity of babies, then -- as Jonathan Bennett has pointed out21 -- it is hard to see their moral difference. For consider the following kind of case: Suppose we know that a victory by Hitler would mean the extermination of all or a great many non-Aryan babies. And further suppose that these babies far outnumber the German babies to be killed in an obliteration bombing campaign.22 Now, given this knowledge, what would a man who really values the lives of babies do? Is the moral case to rest upon the different descriptions "killing babies" and "letting babies die"? If so, why? If the argument is that by not positively killing we will at least be preserving our own moral purity, then it is important to note that this argument, in addition to being rather selfish, is question-begging. For to assume that one remains morally pure if one does nothing is to beg the question of whether we are as responsible for omissive as for commissive conduct. If moral purity means never choosing anything which one will have to regard as in some sense wrong and regret for all one's days, then moral purity may be impossible in a complex world. Albert Camus based this theory of rebellion on this kind of claim -- a theory which Howard Zinn summarizes as follows:Camus spoke in The Rebel of the absurdities in which we are trapped, where the very acts with which we seek to do good cannot escape the imperfections of the world we are trying to change. And so the rebel's "only virtue will lie in never yielding to the impulse to be engulfed in the shadows that surround him, and in obstinately dragging the chains of evil, with which he is bound, toward the good."The issue here is not over whether we should ever allow ourselves to be persuaded by any argument that killing the innocent is in fact necessary; and I am certainly not suggesting that I find plausible such arguments as have actually been given in the past. For a practical maxim I am much in favor of the slogan. "Never trade a certain evil for a possible good." However, this does not solve the issue of principle. If the good (e.g., saving the lives of scores of babies) is certain and not just possible, is it anything more than dogmatism to assert that it would never be right to bring about this good through evil means? The maxim "Never trade a certain evil for a certain good" is by no means self-evidently true and, indeed, does not even seem plausible to many people. Thus I do not think that it has yet been shown that it is always absolutely wrong, whatever the consequences, to kill innocent babies. And thus it has not yet been shown that it is absolutely wrong to kill those innocent in a less rich sense of the term, i.e., noncombatants. Of course we may feel differently about actually killing innocents and simply letting innocents die; but I do not think that this phenomenological evidence in itself proves anything. Bomber pilots no doubt feel differently about dropping bombs on babies from thirty thousand feet than they would about shooting a baby face to face, but surely this does not show that the acts differ in moral quality.
Feelings do not prove anything in morality; but they sometimes point to something. And it is at least possible that we have not yet captured the worry which motivates the responses of people like Kant, Miss Anscombe, and those ho want to defend some version of the doctrine of the Double Effect. If what they really value is the lives of babies and other innocents simpliciter, then -- as Bennett argues -- it does not seem that the distinction between "killing babies" and "letting babies die" will help them to save their principle, But perhaps this is to conceive their position too teleologically. That is, we have so far (with Bennett) been assuming that the person who says "never kill babies" says this because he sees the maxim as instrumental to something else that he values -- namely, the lives of babies. But it is at least possible that he does not hold this principle to be instrumentally right (in which case he would be subject to Bennett's refutation) but intrinsically right, i.e. right in itself or from its very description. The problem is to explicate this notion of intrinsic rightness in such a way that it does not involve either of the two following pitfalls noted by Bennett:
- Authoritarianism, e.g., "God commands not killing babies."
- Dogmatism, e.g., "It is just absolutely wrong to kill babies; and, if you do not see this, you are just too corrupt to talk to."
Now the reason why (a) is a bad move is obvious -- namely, it is an appeal authority rather than reason and thus has no place in a philosophical discussion of moral questions. But (b) is more problematical. For I have said that I am willing to accept such a move if made in the name of there being something wrong with killing babies. Why will I not accept such a move in favor of its being absolutely wrong to kill babies?
Roughly, I should argue as follows: That there is something wrong with killing babies (i.e., that "A is a case of killing a baby" must count as a moral reason against A) explicates, in part, what may be called "the moral point of view." To be worried about moral issues just is, among other things, to be worried about killing innocents. But the judgment "Never kill babies under any circumstances" does not explicate the moral point of view but is, rather, a controversial moral judgment -- or, if you prefer, explicates a moral point of view rather than the moral point of view. And so, to build it into the moral point of view is to beg a controversial question of moral substance -- something which presumably metaethics should not do. Someone who said "I see nothing at all wrong with killing babies so let's bomb Dresden" would be an amoral monster -- one with whom it would be senseless to conduct a moral argument. But one who sincerely said "Of course it is terrible to kill babies but I believe, to save more lives, we must regretfully do it in this case" is not such a monster. He is one with whom, if we think him wrong (immoral), we should hope to be able to argue.24 Since, among war supporters, there are (one would hope) few of the former sort but many of the latter sort, being able to argue with and persuade the latter has some practical importance. It will hardly do simply to say to them "But I can just see the absolute wrongness of what you contemplate and, if you do not, I refuse to discuss the matter with you." We know, alas, what will then happen.
Now I am not going to pretend that I can give anything resembling a proof that it is absolutely wrong to kill the innocent (though not to allow them to die), but I hope at least to be able to elaborate a way of thinking which (a) does give some sense to such a prohibition and (b) cannot be condemned as simple dogmatism or authoritarianism, i.e., is a way of thinking.25
The way of thinking I want to elaborate is one in which the notion of people's having rights plays a predominate role. (Kant's ethical theory, in broad outline, is one such way of thinking.) And an ethical outlook in which the notion of rights looms large will want to draw a distinction between the following two claims:
- Doing A to Jones would be to violate one of his rights.
- It would be bad to do A to Jones.
To use Kant's own examples:26 If I have made a promise to Jones, he has a right to expect me to keep it, can properly regard me as having wronged him if I do not keep it, and could properly expect that others (the state) should coerce me into keeping it. Kant's opaque way of putting this is to say that keeping a promise is a perfect duty. A quite different situation is the following: If Jones is in distress (assuming I have not put him there) and I could help him out without extraordinary sacrifice, he would certainly want me to help him and I would be doing something bad if I did not help him. But he does not have a right to expect my help, is not wronged if I fail to help him, and it would be unreasonable to expect the state or anyone else to coerce me into helping him. Helping him is beneficence and (unlike justice) is a comparatively weak moral demand. Kant's opaque way of putting this is to say that the duty to help others in distress is imperfect.
When'a man has a right, he has a claim against interference. Simply to refuse to be beneficent to him is not an invasion of his rights because it is not to interfere with him at all. When a person uses his freedom to invade the rights of others, he forfeits certain of his own rights and renders interference by others legitimate. (Kant calls this a moral title or authorization -- Befugnis -- to place "obstacles to obstacles to freedom.")27 Thus if I have an imperfect duty to help others, I may interfere with those trying to harm those others because, by such an attempt, they have forfeited their right against interference. Here I have the imperfect duty; and, since those attacking have by the attack forfeited certain of their rights, I violate no perfect duty in interfering with them. Thus there is no conflict here. However, if the only way I could save someone from harm would be by interfering with an innocent person (i.e., one who has not forfeited his rights by initiating attack against others) then I must not save the person, for this would be to violate a perfect duty. And, in cases of conflict, perfect duties override imperfect duties.
Suppose that Jones is being attacked by Smith. In such a case it is certainly true to say that Jones's rights to liberty, security, etc. are being threatened and that Smith, therefore, is acting wrongly and thereby forfeits his right to be left free from interference. Thus I would not be acting wrongly (i.e., against Smith's rights) if I attacked him to prevent his attack on Jones. Similarly, Jones would not be acting wrongly if he defended himself. However, it does not follow from any of this that I have a duty to help Jones or even that Jones has a duty to defend himself. Defense, though permissible, is not obligatory. This being so, it does not follow that Jones has a right to be saved by me. Thus, since it is far from obvious that Jones has a right to be saved even from an attack by the guilty, it is even more implausible to assert that he has a right to be saved if so doing would involve killing the innocent. (Consider the following: We are all, at this very moment, sitting and talking philosophy and are thus omitting to save the lives of countless people we might save throughout the world. Are we acting wrongly in so doing? If we are, is this because all these people have a right to be saved by us?)
Now what sort of a moral view could one hold that would make one accept the principle that perfect duties, resting on rights, override imperfect duties, not resting on rights? I think it is this: a view which makes primary the status of persons as free or choosing beings who, out of respect for that status, are to be regarded as having the right to be left alone to work out their own lives -- for better or worse. This is a basic right that one has just because one is a person. Respecting it is what Kant calls respecting the dignity of humanity by not treating people as means only. Part of respecting them in this sense is not to use them as a means in one's calculations of what would be good for others. It is fine (indeed admirable) for a person to sacrifice himself for others by his own choice; but it is presumptuous (because lacking in respect for his choices) if I choose to sacrifice him. This is his business and not mine. I may only interfere with the person who, by his own evil actions, has forfeited his right against interference. Innocent persons by definition have not done this. And therefore it is absolutely wrong to sacrifice the innocent, though not to kill aggressors. On this view there is something terribly perverse in arguing, as many do, that a defense of freedom requires a sacrifice of those who in no way give their free consent to the sacrifice.28
Of course babies are not yet, in the full sense, free or choosing beings who clearly have rights. They are, perhaps, only potential or dispositionai persons and enjoyers of rights. But if one accepts the maxim "Innocent until proven otherwise" they may be regarded as equally protected in the above way of thinking. For they certainly cannot be described in the only way which, on this view, makes harmful interference permissible -- namely, described as having, through their own deliberate acts of aggression, forfeited their right to be left in peace.
Now this view that what is central in morality involves notions like rights, dignity, freedom, and choice (rather than notions like maximizing the general utility) cannot be proven. But it is a plausible view which may lie behind the maxim "Never kill the innocent" and is a view which would be sacrificed
I (at least greatly compromised) by the maxim "Kill the innocent to save the innocent." I am myself deeply sympathetic to this way of thinking and would make neither the compromise nor the sacrifice. But I cannot prove that one ought not make it. Neither, of course, can my teleological opponent prove his case either. For we lie here at the boundaries of moral discourse where candidates for ultimate principles conflict; and it is part of the logical character of an ultimate principle that it cannot be assessed by some yet higher ("more ultimate"?) principle.29 You pays your money and you takes your choice. It is simply my hope that many people, if they could see clearly what price they have to pay (i.e. the kind of moral outlook they have to give up and what they have to put in its place) would make the choice against killing the innocent.
Consider the following example: Suppose that thousands of babies could be saved from a fatal infant disease if some few babies were taken by the state and given over to a team of medical researchers for a series of experiments which, though killing the babies, would yield a cure for the disease. In what way except degree (i.e., numbers of babies killed) does this situation differ from the rationale behind antimorale obliteration bombing raids, i.e., is there not a disturbing parallel between Allied raids on Dresden and Tokyo and
Nazi "medicine"? With respect to either suggestion, when we really think about it, do we not want to say with the poet James DickeyHolding onto another man's wallsHow can any such thing be in the interest of humanity when its practice would change the very meaning of "humanity" and prevent us from unpacking from it, as we now do, notions like rights, dignity, and respect? No matter how good the consequences, is there not some point in saying that we simply do not have the right to do it? For there is, I think, an insight of secular value in the religious observation that men are the "children of God." For this means, among other things, that other people do not belong to me. They are not mine to be manipulated as resources in my projects. It is hard to imagine all that we might lose if we abandoned this way of thinking about ourselves and others.
My hat should crawl on my head
In streetcars, thinking of it,
The fat on my body should pale.30
My appeal here, of course, is in a sense emotive. But this in my judgment is not an objection. Emotive appeals may rightly be condemned if they are masquerading as proofs. But here I am attempting to prove nothing but only to say -- "Here, look, see what you are doing and what way of thinking your doing it involves you in." If one sees all this and still goes forth to do it anyway, we have transcended the bounds of what can be said in the matter.
What about noncombatants? Though they are not necessarily innocent in all the senses in which babies are, they clearly are innocent in the sense I have elaborated above -- namely, they have not performed actions which forfeit their right to be free from execution (or, better: it is not reasonable for the enemy to believe this of them). Thus, in a very tentative conclusion, I suggest the following: I have not been able to prove that we should never kill noncombatants or innocents (I do not think this could be proven in any ordinary sense of proof): but I do think that I have elaborated a way of thinking which gives sense to the acceptance of such an absolute prohibition. Thus, against Bennett, I have at least shown that one can accept the principle "Never kill the innocent" without thereby necessarily being an authoritarian or a dogmatic moral fanatic.
AUTHOR'S NOTE: Many people were kind enough to comment on this manuscript in various stages of its preparation, and I should like to take this opportunity to express my sincere appreciation for the help that they provided. In particular, I should like to thank the following: Lewis White Beck, Robert L. Holmes, Gareth Matthews, Ronald Milo, Richard Wasserstrom, Donald Wells, Peter Winch, and Anthony D. Woozley. I should also like to thank the members of my graduate seminar on war at the University of Arizona for stimulating discussion. Since I have perversely gone my own way on several points in spite of advice to the contrary, I alone am to be held responsible for any errors that remain. In setting a philosophical framework for a discussion of the problem of killing the innocent in war, I have been greatly influenced by Richard Wasserstrom's "On the Morality of War: A Preliminary Inquiry" in his useful collection of essays War and Morality (Belmont. Calif: Wadsworth Publishing Co., 1970). I am presupposing that the reader is familiar with Miss G. E. M. Anscombe's two articles "Modern Moral Philosophy," Philosophy, 33, No. 124 (January 1958), and "War and Murder" (in the Waserstrom collection). I am also presupposing a familiarity with Jonathan Bennetts' "Whatever the Consequences," Analysis, 26, No. 3 (January 1966).
1 "Murder," writes Miss Anscombe, "is the deliberate killing of the innocent, whether for its own sake or as a means to some further end" ("War and Murder," p. 45). Deliberate killing of the innocent (or noncombatants) is prohibited by the Just War Theory and is a crime in national law. A traditional account of the Catholic Just War Theory may be found in Chapter 35 of Austin Fagothey's Right and Reason: Ethics in Theory and Practice (St. Louis: C. V. Mosby Co., 1963). A useful sourcebook for inquiry into the nature ot war crimes is the anthology Crimes of War, ed. by Richard A. Falk, Gabriel Kilko, and Robert Jay Lifton (New York: Random House, 1971).
2 By "prima facie wrong" I mean "can be overridden by other moral requirements"—not, as a literal translation might suggest, "only apparently wrong."
3 For example: In the criminal offense of statutory rape, the defendant is strictly liable with respect to his knowledge of the age of a girl with whom "he has had sexual relations, i.e., no matter how carefully he inquired into her age, no matter how reasonable (i.e., nonnegligent) his belief that she was of legal age of consent, he is liable if his belief is in fact mistaken. For a general discussion of such offenses, see Richard Wasserstrom's "Strict Liability in the Criminal Law," Stanford Law Review, 12 (July 1960).
4 In discussion, Richard Wasserstrom has expressed scepticism concerning my claim that there is something unintelligible about the concept of strict moral responsibility. One could regard the Old Testament and Oedipus Rex as containing a strict liability conception of morality. Now I should be inclined to argue that the primitiveness of the Old Testament and of Oedipus Rex consists in these peoples not yet being able to draw a distinction between legality and morality. However, I am prepared to admit that it might be better to weaken my claim by maintaining simply that no civilized or enlightened morality would involve strict liability.
5 In California criminal law, for example, vehicular manslaughter is defined as vehicular homicide "in the commission of an unlawful act, not amounting to felony, with gross negligence; or in the commission of a lawful act which might produce death, in an unlawful manner, and with gross negligence . . ." (California Penal Code, 192, 3, a).
6 For an excellent discussion of moral and legal responsibility for negligence, see H. L. A. Hart's "Negligence, Mens Rea and Criminal Responsibility," in his Punishment and Responsility: Essays in the Philosophy of Law (Oxford: Oxford University Press, 1963).
7 I say "engaged in an attempt" rather than "attempting" for the following reason: A mortar attack on an encampment of combat soldiers who happen to be sleeping is surely not a case of killing noncombatants even though persons who are asleep cannot be attempting anything. Sleeping persons can, however, be engaged in an attempt -- just as sleeping persons can be accomplices in crime and parties to a criminal conspiracy. Being engaged in an attempt, unlike attempting, is not necessarily a full time job. I am grateful to Anthony Woozley for pointing this out to me.
8 Thomas Nagel, "War and Massacre." Philosophy and Public Affairs, 2 (Winter 1972). In the same issue, Richard Brandt replies to Nagei in his "Utilitarianism and the Rules of War." I am grateful to Professors Nagel and Brandt for allowing me to read their articles prior to publication.
9 For reasons of simplicity in later drawing upon important and instructive principles from the criminal law, I shall use the phrase "self-defense." (I shall later want to draw on the notion of reasonable belief in the law of self-defense.) However, what I really want to focus on is the concept of "defense" and not the concept of "self." For it seems to me that war can be justified, not just to defend oneself or one's nation, but also to defend others from threats that transcend nationality, e.g., genocide. If one wants to speak of self-defense even here, then it must be regarded as self-defense for the human, not just national, community. The phrase "self-defense" as it occurs in what follows should always be understood as carrying this qualification. And, of course, even clear cases of self-defense are not always necessarily justified. Given the morally debased character of Nazi Germany, it is by no means obvious that it acted rightly in trying to defend itself near the end of Worlrl War II (i.e., after it had ceased to be an aggressor).
10 Remember that this carries the qualification stated in note 9. For a survey of the law of self-defense, the reader may consult any reliable treatise on the criminal law, e.g., pp. 883 ff. of Rollin M. Perkins's Criminal Law (Brooklyn, N.Y.: Foundation Press, 1957). The criminal law is a highly moralized institution, and it is useful (though by no means always definitive) for the moral philosopher in that it provides an accumulated and systematized body of reflection on vital moral matters of our culture. For my purposes, I shall in what follows focus upon the reasonable belief condition in the law of self-defense. Other aspects of the law of self-defense (e.g., the so-called "retreat requirement"), have, I think, interesting implications for war that I cannot pursue here.
11 "Modern Moral Philosophy," p. 17. For reasons I shall note later, Miss Anscombe weakens her point by adding the word judicial.
12 See William H. Gass, "The Case of the Obliging Stranger," Philosophical Review, 66 (April 1957).
13 John Rawls, "Two Concepts of Rules," Philosophical Review, 64 (January 1955).
14 See, fo example, United States v. Falcone, United States Court of Appeals, Second Circuit, 1940, 109 F:2d 579.
15 Consider the case of the homicidal diabetic: He is chasing you through the woods of an enclosed preserve, attempting to kill you for sport with a pistol. However, because of his medical condition, he must return to a cabin in the middle of the preserve every hour in order that his aged mother can give him an insulin shot. Without it, he will take ill or will die and will thus be forced to abandon his attempt to kill you. Even if blocking that insulin shot seems your only hope, killing the mother in order to do it would be a very doubtful case of selt-defence.
16 Austin Fagothey, in his well-known statement of the Just War Theory cited previously, allows injury to national honor to count as a ground for a war of self-defense. I cite Fagothey's book because of its influence. It was, for example, appealed to by Justice Douglas in the 1971 Gillette case where the defendant refused induction on the grounds of his belief that the Vietnam War is unjust (Gillette v. United States, 401 U.S. 437.
17 "Moswern Moral Philosophy."
18 Another factor, relevant both to war and to the highway example, is the following: acting with the knowledge that deaths could be prevented by taking reasonable precautions and yet not taking those precautions. Such grossly negligent or reckless behavior, while perhaps not "deliberate" in the strict sense, is surely immoral in either context.
19 See Michael Walzer, "World War II: Why Was This War Different?," Philosophy and Public Affairs, 1, No. 1 (Fall 1971).
20 As the citizens of London and Hanoi have illustrated, for example, terror bombing has a tendency to backfire. Rather than demoralizing the enemy, it sometimes strengthens their courage and will to resist.
21 Jonathan Bennett, "Whatever the Consequences," Analysis, 26, No. 3 (January 1966).
22 I take it that, from a moral point of view, their being German babies is irrelevant. As Howard Zinn has argued, we should accept the principle that "all victims are created equal" (Disobedience and Democracy: Nine Fallacies on Law and Order [New York: Random House, 1968], p. 50).
23 Ibid., p. 40.
24 One important metaethical inquiry that has been conducted in recent years concerns the nature of moral judgments and the moral point of view or "language game." The task here is to distinguish the moral from the nonmoral or amoral. Normative ethics, on the other hand, is concerned to distinguish the moral from the immoral. This is too large a dispute to enter here, but I can at least make my own commitments clear. Unlike the so-called "formalists" (e.g., R. M. Hare, The Language of Morals [Oxford: Oxford University Press, 1952]), I am inclined to believe that the moral point of view must be defined, in part, in terms of the content of the judgments it contains. Here I am siding, if I understand them correctly, with such writers as H. L. A. Hart (The Concept of Law [Oxford: Oxford University Press, 1961]), and G. J. Warnock (The Object of Morality [London: Methuen, 1971]). Someone who does not see that "A is a case of killing an innocent" is a relevant reason against doing A, does not understand what moral discourse is all about, what it is necessarily concerned with. And this is so no matter how much he is prepared to universalize and regard as overriding his own idiosyncratic imperatives. If this is true, then the gulf between metaethics and normative ethics is not quite as wide as many have supposed, since the relevance of certain substantive judgments is now going to be regarded as part of the meaning of morality as a point of view, language game, or form of life. There still is some gulf between metaethics and normative ethics, however, since I should argue that only the relevance, and not the decisiveness, of certain substantive judgments (e.g., do not kill the innocent) can be regarded as a defining feature of the moral point of view. Normative ethics, however, is primarily interested in which of the relevant moral considerations are, in certain circumstances, decisive. (I am grateful to Ronald Milo for discussing these matters with me. He is totally responsible for whatever clarity is to be found in my views.)
25 Though I shall not explore it here, I think that the religious acceptance of the principle could be elaborated as a way of thinking and so distinguished from the authoritarianism and dogmatism that Bennett so rightly condemns. The view I shall develop also has implications (though I shall not draw them out here) for the abortion issue. For more on this, see Philippa Foot's "Abortion and the Doctrine of the Double Effect," Oxford Review (1967). This has been reprinted in Moral Problems: A Collection of Philosophical Essays, ed. by James Rachels (New York: Harper and Row, 1971), and in my An Introduction to Moral and Social Philosophy: Basic Readings in Theory and Practice (Belmont, Calif.: Wadsworth Publishing Co., 1973). She sees the doctrine of the Double Effect as groping in a confused way toward an insight of moral importance: "There is worked into our moral system a distinction between what we owe people in the form of aid and what we owe them in the way of non-interference."
26 These well-known examples are drawn from Kant's Foundations of the Metaphysics of Morals, trans, by Lewis White Beck (Indianapolis: Bobbs-Merrill Co., 1959), pp. 39 ff.; pp. 421 ff. of the Academy edition of Kant's works. Kant distinguishes perfect duties (duties of respect) and imperfect duties (duties of love) in many places. See, for example, The Doctrine of Virtue, trans, by Mary J. Gregor (New York: Harper and Row, 1964), pp. 134-35; p. 463 of the Academy edition.
27 Kant, Metaphysical Elements of Justice, trans, by John Ladd (Indianapolis: Bobbs-Merrill Co., 1965), pp. 35 ff.; pp. 230 ff. of the Academy edition. See also pp. 94 ff. and 107-9 of my Kant: The Philosophy of Right (New York: St. Martin's Press, 1970).
28 Someone (e.g., Jan Narveson in his "Pacifism: A Philosophical Analysis," Ethics, 75 ) might say the following: If what you value are the rights of people, does this not entail (as a part of what it means to say that one values something) that you would recognize an obligation to take steps to secure those rights against interference? Perhaps. But "take steps" does not have to mean "do whatever is causally necessary." Such an obligation will presumably be limited by (a) what I can reasonably be expected to do or sacrifice and (b) my moral judgment about the permissibility of the means employed. It does seem perverse for a person to say "I value the rights of people above all else but I do not propose to do anything to secure those rights." However, it by no means seems to me equally perverse to say the following: "I deplore interferences with human rights above all else and I will do everything in my power to prevent such interferences -- everything, of course, short of being guilty of such interferences myself."
29 See Brian Barry, "Justice and the Common Good," Analysis, 21-22 (1960-61). Though the existentialists tend to overdo this sort of claim, there is truth in their claim that there are certain moral problems that are in principle undecidable by any rational decision procedure. Such cases are not perhaps as numerous as the existentialists would have us believe, but they do arise, e.g. when candidates for ultimate moral principles conflict. In such cases, we find it impossible to nail down with solid arguments those principles which matter to us the most. Perhaps this even deserves to be called "absurd." In his article "The Absurd" (Journal of Philosophy, 68, No. 20 [October 21, 19711), Thomas Nagel suggests that absurdity in human life is found in "the collision between the seriousness with which we take our lives and the perpetual possibility of regarding everything about which we are serious as arbitrary, or open to doubt." It is here that talk about faith or commitment seems to have some life.
30 James Dickey, "The Firebombing," in Poems 1957-1967 (Middletown, Conn.: Wesleyan University Press, 1967), p. 185. Reprinted with the permission of the publisher.