The Problem of Immoral Threats
Truman used the atomic bomb to end a war that seemed to him limitless in its horrors. And then, for a few minutes or hours in August 1945, the people of Hiroshima endured a war that actually was limitless in its horrors. “In this last great action of the Second War War,” wrote Stimson, “we were given final proof that war is death.”1 Final proof is exactly the wrong phrase, for war had never been like that before. A new kind of war was born at Hiroshima, and what we were given was a first glimpse of its deadliness. Though fewer people were killed than in the fire-bombing of Tokyo, they were killed with monstrous ease. One plane, one bomb: with such a weapon the 350 planes that raided Tokyo would virtually have wiped out human life on the Japanese islands. Atomic war was death indeed, indiscriminate and total, and after Hiroshima, the first task of political leaders everywhere was to prevent its recurrence.
The means they adopted is the promise of reprisal in kind. Against the threat of an immoral attack, they have put the threat of an immoral response. This is the basic form of nuclear deterrence. In international as in domestic society, deterrence works by calling up dramatic images of human pain. “In the groves of their academy,” wrote Edmund Burke of the liberal theorists of crime and punishment, “at the end of every vista, you see nothing but the gallows.”2 The description is uncomplimentary, for Burke believed that domestic peace must rest upon some other foundation. But there is this much to be said for the gallows: in principle, at least, only guilty men need fear the death it brings. About the theorists of deterrence, however, it must be said, “In the groves of their academy, at the end of every vista, you see nothing but the mushroom cloud”—and the cloud symbolizes indiscriminate slaughter, the killing of the innocent (as in Hiroshima) on a massive scale. No doubt, the threat of such slaughter, if it is believed, makes nuclear attack a radically undesirable policy. Doubled by a potential enemy, the threat produces a “balance of terror.” Both sides are so terrified that no further terrorism is necessary. But is the threat itself morally permissible?
The question is a difficult one. It has generated in the years since Hiroshima a significant body of literature exploring the relation between nuclear deterrence and just war.3 This has been the work mostly of theologians and philosophers, but some of the strategists of deterrence have also been involved; they worry about the act of terrorizing much as conventional soldiers worry about the act of killing. I cannot review this literature here, though I shall draw upon it freely. The argument against deterrence is familiar enough. Anyone committed to the distinction between combatants and noncombatants is bound to be appalled by the specter of destruction evoked, and purposely evoked, in deterrence theory. “How can a nation live with its conscience,” John Bennett has asked, “and know that it is preparing to kill twenty million children in another nation if the worst should come to the worst?”4 And yet, we have lived with that knowledge, and with our consciences too, for several decades now. How have we managed? The reason for our acceptance of deterrent strategy, most people would say, is that preparing to kill, even threatening to kill, is not at all the same thing as killing. Indeed it is not, but it is frighteningly close—else deterrence wouldn’t “work”—and it is in the nature of that closeness that the moral problem lies.
The problem is often misdescribed—as in the following analogy for nuclear deterrence first suggested by Paul Ramsey and frequently repeated since:5
Suppose that one Labor Day weekend no one was killed or maimed on the highways; and that the reason for the remarkable restraint placed on the recklessness of automobile drivers was that suddenly everyone of them discovered he was driving with a baby tied to his front bumper! That would be no way to regulate traffic even if it succeeds in regulating it perfectly, since such a system makes innocent human lives the direct object of attack and uses them as a mere means for restraining the drivers of automobiles.
No one, of course, has ever proposed regulating traffic in this ingenious way, while the strategy of deterrence was adopted with virtually no opposition at all. That contrast should alert us to what is wrong with Ramsey’s analogy. Though deterrence turns American and Russian civilians into mere means for the prevention of war, it does so without restraining any of us in any way. Ramsey reproduces the strategy of the German officers during the Franco-Prussian War who forced civilians to ride on military trains in order to deter saboteurs. By contrast with those civilians, however, we are hostages who lead normal lives. It is in the nature of the new technology that we can be threatened without being held captive. That is why deterrence, while in principle so frightening, is so easy to live with. It cannot be condemned for anything it does to its hostages. It is so far from killing them that it does not even injure or confine them; it involves no direct or physical violation of their rights. Those critics of deterrence who are also committed consequentialists have had to imagine psychic injuries. Thus Eric Fromm, writing in 1960: “To live for any length of time under the constant threat of destruction creates certain psychological effects in most human beings—fright, hostility, callousness . . . and a resulting indifference to all the values we cherish. Such conditions will transform us into barbarians. . . .”6 But I don’t know of any evidence that bears out either the assertion or the prediction; surely we are no more barbarians now than we were in 1945. In fact, for most people, the threat of destruction, though constant, is invisible and unnoticed. We have come to live with it casually—as Ramsey’s babies, traumatized for life in all probability, could never do, and as hostages in conventional wars have never done.
If deterrence were more painful, we might have found other means of avoiding nuclear war—or we might not have avoided it. If we had to keep millions of people under restraint in order to maintain the balance of terror, or if we had to kill millions of people (periodically) in order to convince our adversaries of our credibility, deterrence would not be accepted for long.7 The strategy works because it is easy. Indeed, it is easy in a double sense: not only don’t we do anything to other people, we also don’t believe that we will ever have to do anything. The secret of nuclear deterrence is that it is a kind of bluff. Perhaps we are only bluffing ourselves, refusing to acknowledge the real terrors of a precarious and temporary balance. But no account of our experience is accurate which fails to recognize that, for all its ghastly potential, deterrence has so far been a bloodless strategy.
So far as consequences go, then, deterrence and mass murder are very far apart. Their closeness is a matter of moral posture and intention. Once again, Ramsey’s analogy misses the point. His babies are not really the “direct object of attack,” for whatever happens on that Labor Day weekend, no one will deliberately set out to kill them. But deterrence depends upon a readiness to do exactly that. It is as if the state should seek to prevent murder by threatening to kill the family and friends of every murderer—a domestic version of the policy of “massive retaliation.” Surely that would be a repugnant policy. We would not admire the police officials who designed it or those pledged to carry it out, even if they never actually killed anybody. I don’t want to say that such people would necessarily be transformed into barbarians; they might well have a heightened sense of how awful murder is and a heightened desire to avoid it; they might loathe the work they were pledged to do and fervently hope that they never had to do it. Nevertheless, the enterprise is immoral. The immorality lies in the threat itself, not in its present or even its likely consequences. Similarly with nuclear deterrence: it is our own intentions that we have to worry about and the potential (since there are no actual) victims of those intensions. Here Ramsey has put the case very well: “Whatever is wrong to do is wrong to threaten, if the latter means ‘mean to do.’ . . . If counter-population warfare is murder, then counter-population deterrent threats are murderous.”8 No doubt, killing millions of innocent people is worse than threatening to kill them. It is also true that no one wants to kill them, and it may well be true that no one expects to do so. Nevertheless, we intend the killings under certain circumstances. That is the stated policy of our government; and thousands of men, trained in the techniques of mass destruction and drilled in instant obedience, stand ready to carry it out. And from the perspective of morality, the readiness is all. We can translate it into degrees of danger, high and low, and worry about the risks we are imposing on innocent people, but the risks depend on the readiness. What we condemn in our own government, as in the police in my domestic analogy, is the commitment to murder.a
But this analogy, too, can be questioned. We don’t prevent murder any more than we control traffic in these bizarre and inhuman ways. But we do deter or seek to deter our nuclear adversaries. Perhaps deterrence is different because of the danger its advocates claim to avoid. Traffic deaths and occasional murders, however much we deplore them, do not threaten our common liberties or our collective survival. Deterrence, so we have been told, guards us against a double danger: first, of atomic blackmail and foreign domination; and second, of nuclear destruction. The two go together, since if we did not fear the blackmail, we might adopt a policy of appeasement or surrender and so avoid the destruction. Deterrence theory was worked out at the height of the cold war between the United States and the Soviet Union, and those who worked it out were concerned above all with the political uses of violence—which are not relevant in either the traffic or police analogies. Underlying the American doctrine, there seemed to lurk some version of the slogan “Better dead than Red” (I don’t know the Russian parallel). Now that is not really a believable slogan; it is hard to imagine that a nuclear holocaust was really thought preferable to the expansion of Soviet power. What made deterrence attractive was that it seemed capable of avoiding both.
We need not dwell on the nature of the Soviet regime in order to understand the virtues of this argument. Deterrence theory doesn’t depend upon a view of Stalinism as a great evil (though that is a highly plausible view) in the same way that my argument about terror bombing depended upon an assertion about the evils of Nazism. It requires only that we see appeasement or surrender to involve a loss of values central to our existence as an independent nation-state. For it is not tolerable that advances in technology should put our nation, or any nation, at the mercy of a great power willing to menace the world or to press its authority outwards in the shadow of an implicit threat. The case here is very different from that which arises commonly in war, where our adherence to the war convention puts us, or would put us, at a disadvantage vis-à-vis them. For disadvantages of that sort are partial and relative; various counter-measures and compensating steps are always available. But in the nuclear case, the disadvantage is absolute. Against an enemy actually willing to use the bomb, self-defense is impossible, and it makes sense to say that the only compensating step is the (immoral) threat to respond in kind. No country capable of making such a threat is likely to refuse to make it. What is not tolerable won’t be tolerated. Hence any state confronted by a nuclear adversary (it makes little difference what the adversary relationship is like or what ideological forms it assumes), and capable of developing its own bomb, is likely to do so, seeking safety in a balance of terror.b Mutual disarmament would clearly be a preferable alternative, but it is an alternative available only to the two countries working closely together, whereas deterrence is the likely choice of either one of them alone. They will worry about one another’s readiness to attack; they will each assume their own commitment to resist; and they will realize that the greatest danger of such a confrontation would not be the defeat of one side or the other but the total destruction of both—and possibly of everyone else too. This in fact is the danger that has faced mankind since 1945, and our understanding of nuclear deterrence must be worked out with reference to its scope and imminence. Supreme emergency has become a permanent condition. Deterrence is a way of coping with that condition, and though it is a bad way, there may well be no other that is practical in a world of sovereign and suspicious states. We threaten evil in order not to do it, and the doing of it would be so terrible that the threat seems in comparison to be morally defensible.
If the bomb were ever used, deterrence would have failed. It is a feature of massive retaliation that while there is or may be some rational purpose in threatening it, there could be none in carrying it out. Were our “bluff” ever to be called and our population centers suddenly attacked, the resulting war could not (in any usual sense of the word) be won. We could only drag our enemies after us into the abyss. The use of our deterrent capacity would be an act of pure destructiveness. For this reason, massive retaliation, if not literally unthinkable, has always seemed undo-able, and this is a source of considerable anxiety for military strategists. Deterrence only works, they argue, if each side believes that the other might actually carry out its threat. But would we carry it out? George Kennan has recently given what must be the moral response:9
Let us suppose there were to be a nuclear attack of some sort on this country and millions of people were killed and injured. Let us further suppose that we had the ability to retaliate against the urban centers of the country that had attacked us. Would you want to do that? I wouldn’t . . . I have no sympathy with the man who demands an eye for an eye in a nuclear attack.
A humane position—though one that should probably be whispered, rather than published, if the balance of terror is to be sustained. But the argument might look very different if the original attack or the planned response avoided cities and people. If a limited nuclear war were possible, wouldn’t it also be do-able? And might not the balance of terror then be re-established on the basis of threats that were neither immoral nor unconvincing?
Over a brief timespan, in the late 1950s and the early 1960s, these questions were answered with an extraordinary outpouring of strategic arguments and speculations, overlapping in important ways with the moralizing literature I described earlier.10 For the debate among the strategists focused on the attempt (though this was rarely made explicit) to fit nuclear war into the structure of the war convention, to apply the argument for justice as if this sort of conflict were like any other sort. The attempt involved, first, a defense of the use of tactical nuclear weapons in deterring and, if that failed, in resisting conventional or small-scale nuclear attacks; and it involved, secondly, the development of a “counter-force” strategy directed at the enemy’s military installations and also at major economic targets (but not at entire cities). These two had a similar purpose. By holding out the promise of a limited nuclear war, they made it possible to imagine actually fighting such a war—they made it possible to imagine winning it—and so they strengthened the intention that lay behind the deterrent threat. They transformed the “bluff” into a plausible option.
Until the late 1950s, the tendency of most people was to regard the atomic bomb and its thermonuclear successors as forbidden weapons. They were treated on analogy with poison gas, though the prohibition on their use was never legally established. “Ban the bomb” was everyone’s policy, and deterrence was simply a practical way of enforcing the ban. But now the strategists suggested (rightly) that the crucial distinction in the theory and practice of war was not between prohibited and acceptable weapons but between prohibited and acceptable targets. Massive retaliation was painful and difficult to contemplate because it was modeled on Hiroshima; the people we were planning to kill were innocent, militarily uninvolved, as removed from and ignorant of the weapons with which their leaders threatened us as we were of the weapons with which our leaders threatened them. But this objection would disappear if we could deter our adversaries by threatening a limited and morally acceptable destruction. Indeed, it might disappear so entirely that we would be tempted to give up deterrence and initiate the destruction ourselves whenever it seemed to our advantage to do so. This was certainly the tendency of much strategic argument, and several writers painted rather attractive pictures of limited nuclear war. Henry Kissinger likened it to war at sea—the very best kind of war, since no one lives in the sea. “The proper analogy . . . is not traditional land warfare, but naval strategy, in which self-contained, [highly mobile] units with great fire power gradually gain the upper hand by destroying their enemy counterparts without physically occupying territory or establishing a front line.”11 The only difficulty is that Kissinger imagined fighting a war like that in Europe.c
Tactical and counter-force warfare meets the formal requirements of jus in bello, and it was seized upon eagerly by certain moral theorists. That is not to say, however, that it makes moral sense. There remains the possibility that the new technology of war simply doesn’t fit and cannot be made to fit within the old limits. This proposition can be defended in two different ways. The first is to argue that the collateral damage likely to be caused even by a “legitimate” use of nuclear weapons is so great that it would violate both of the proportionality limits fixed by the theory of war: the number of people killed in the war as a whole would not be warranted by the goals of the war—particularly since the dead would include many if not most of the people for whose defense the war was being fought; and the number of people killed in individual actions would be disproportionate (under the doctrine of double effect) to the value of the military targets directly attacked. “The disproportion between the cost of such hostilities and the results they could achieve,” wrote Raymond Aron, thinking of a limited nuclear war in Europe, “would be colossal.”12 It would be colossal even if the formal limits on targeting were in fact observed. But the second argument against limited nuclear war is that these limits would almost certainly not be observed.
At this point, of course, one can only guess at the possible shape and course of the battles; there is no history to study. Neither moralists nor strategists can refer to cases; instead they design scenarios. The scene is empty; one can fill it in very different ways, and it is not impossible to imagine that limits might be maintained even after nuclear weapons had been used in battle. The prospect that they would be maintained and the war extended over time is so frightening to those countries on whose soil such wars are likely to be fought that they have generally opposed the new strategies and insisted upon the threat of massive retaliation. Thus, as André Beaufre has written, “Europeans would prefer to risk general war in an attempt to avoid war altogether rather than have Europe become the theater of operations for limited war.”13 In fact, however, the risks of escalation will be great whatever limits are adopted, simply because of the immense destructive power of the weapons involved. Or rather, there are two possibilities: either nuclear weapons will be held at such low levels that they won’t be significantly different from or of greater military utility than conventional explosives, in which case there is no reason to use them at all; or their very use will obliterate the distinction between targets. Once a bomb has been aimed at a military target but has, as a side effect, destroyed a city, the logic of deterrence will require the other side to aim at a city (for the sake of its seriousness and credibility). It is not necessarily the case that every war would become a total war, but the danger of escalation is so great as to preclude the first use of nuclear weapons—except by someone willing to face their final use. “Who would even launch such hostilities,” Aron has asked, “unless he was determined to persist to the bitter end?”14 But such a determination is not imaginable in a sane human being, let alone in a political leader responsible for the safety of his own people; it would involve nothing less than national suicide.
These two factors, the extent even of limited destruction and the dangers of escalation, seem to rule out any sort of nuclear war between the great powers. They probably rule out large-scale conventional war, too, including the particular conventional war about which the strategists of the 1950s and 1960s were most concerned: a Russian invasion of western Europe. “The spectacle of a large Soviet field army crashing across the line into western Europe in the hope and expectation that nuclear weapons would not be used against it—thereby putting itself and the USSR totally at risk while leaving the choice of weapons to us—would seem to be hardly worth a second thought. . . .”15 It is important to stress that the bar lies in the totality of the risk: not in the possibility of what the strategists called a “flexible response,” finely adjusted to the scope of the attack, but in the stark reality of ultimate horror should the adjustments fail. It may well be that “flexible response” enhanced the value of a counter-population deterrent by making it possible to reach that final point in “easy” stages, but it is also and more importantly true that we have never begun the staged escalation and are never likely to begin it, because of what lies at the end. Hence the persistence of counter-population deterrence, and hence also the virtual end of the strategic debate, which petered out in the middle 1960s. At that point, I think, it became clear that given the existence of large numbers of nuclear weapons and their relative invulnerability, and barring major technological breakthroughs, any imaginable strategy is likely to deter a “central war” between the great powers. The strategists helped us to understand this, but once it was understood it became unnecessary to adopt any of their strategies—or at least, any particular one of them. We continue to live, then, with the paradox that pre-existed the debate: nuclear weapons are politically and militarily unusable only because and insofar as we can plausibly threaten to use them in some ultimate way. And it is immoral to make threats of that kind.
Before deciding (or refusing) to live with this paradox, I want to consider in some detail the work of the Protestant theologian Paul Ramsey, who has over a period of years argued that there exists a justifiable deterrent strategy. From the beginning of the moral and strategic debates, Ramsey has been a sharp opponent of the advocates of counter-city deterrence and also of those of its critics who think that it is the only form of deterrence and therefore opt for nuclear disarmament. He has condemned both these groups for the all-or-nothing character of their thinking: either total and immoral destruction or a kind of “pacifistic” inertia. He argues that these twin perspectives conform to the traditional American view of war as an all-out conflict, which must therefore be avoided whenever possible. Ramsey himself, I think, is a Protestant soldier in a different tradition; he would have Americans gird themselves for a long, continuous struggle with the forces of evil.16
Now if there is to be a justified deterrent strategy, there must be a justified form of nuclear war, and Ramsey has conscientiously argued “the case for making just war possible” in the modern age. He takes a lively and well-informed interest in the strategic debates and has at various times defended the use of tactical nuclear weapons against invading armies and of strategic weapons against nuclear installations, conventional military bases, and isolated economic objectives. Even these targets are only “conditionally” permissible, since the proportionality rule would have to be applied in each case, and Ramsey does not believe that its standards will always be met. Like everyone (or almost everyone) who writes about these matters, he has no zest for nuclear combat; his main interest is in deterrence. But he needs at least the possibility of legitimate warfare if he is to maintain a deterrent posture without making immoral threats. That is his central purpose, and the effort to achieve it involves him in a highly sophisticated application of just war theory to the problems of nuclear strategy. In the best sense of the word, Ramsey is engaged with the realities of his world. But the realities in this case are intractable, and his way around them is finally too complex and too devious to provide a plausible account of our moral judgments. He multiplies distinctions like a Ptolemaic astronomer with his epicycles and comes very close at the end to what G. E. M. Anscombe has called “double-think about double effect.”17 But his work is important; it suggests the outer limits of the just war and the dangers of trying to extend those limits.
Ramsey’s central claim is that it is possible to prevent nuclear attack without threatening to bomb cities in response. He believes that “the collateral civilian damage that would result from counter-force warfare in its maximum form” would be sufficient to deter potential aggressors.18 Since the civilians likely to die in such a war would be the incidental victims of legitimate military strikes, the threat of counter-force warfare plus collateral damage is also morally superior to deterrence in its present form. These are not hostages whom we intend to murder (under certain circumstances). Nor are we planning their deaths; we are only pointing out to our possible enemies the unavoidable consequences even of a war justly fought—which is, we could honestly say were we to adopt Ramsey’s proposal, the only sort of war we were preparing to fight. Collateral damage is simply a fortunate feature of nuclear warfare; it serves no military purpose, and we would avoid it if we could, though it is clearly a good thing that we cannot. And since the damage is justifiable in prospect, it is also justifiable here and now to call that prospect to mind for the sake of its deterrent effects.
But there are two problems with this argument. First, the danger of collateral damage is unlikely to work as a deterrent unless the damage expected is radically disproportionate to the ends of the war or the value of this or that military target. Hence Ramsey is driven to argue that “the threat of something disproportionate is not always a disproportionate threat.”19 What that means is this: proportionality in combat is measured, let’s say, against the value of a particular missile base, while proportionality in deterrence is measured against the value of world peace. So the damage may not be justifiable in prospect (under the doctrine of double effect), and yet the threat of such damage may still be morally permitted. Perhaps that argument is right, but I should stress that its result is to void the proportionality rule. Now there is no limit on the number of people whose deaths we can threaten, so long as those deaths are to be caused “collaterally” and not by taking direct aim. As we have seen before, the idea of proportionality, once it is worked on a bit, tends to fade away. And then the entire burden of Ramsey’s argument falls on the idea of death by indirection. That is indeed an important idea, central to the permissions and restraints of conventional war. But its standing is undermined here by the fact that Ramsey relies so heavily on the deaths he supposedly doesn’t intend. He wants, like other deterrent theorists, to prevent nuclear attack by threatening to kill very large numbers of innocent civilians, but unlike other deterrent theorists, he expects to kill these people without aiming at them. That may be a matter of some moral significance, but it does not seem significant enough to serve as the cornerstone of a justified deterrent. If counter-force warfare had no collateral effects, or had minor and controllable effects, then it could play no part in Ramsey’s strategy. Given the effects it does have and the central part it is assigned, the word “collateral” seems to have lost much of its meaning. Surely anyone designing such a strategy must accept moral responsibility for the effects on which he is so radically dependent.
But we have not yet seen the whole of Ramsey’s design, for he doesn’t pull back from the hardest questions. What if the likely collateral damage of a just nuclear war isn’t great enough to deter a would-be aggressor? What if the aggressor threatens a counter-city strike? Surrender would be intolerable, and yet we cannot ourselves threaten mass murder in response. Fortunately (again), we don’t have to. “We do not need . . . to threaten that we will use [nuclear weapons] in case of attack,” Bernard Brodie has written. “We do not need to threaten anything. Their being there is quite enough.”20 So it is, too, according to Ramsey, with counter-city strikes: the mere possession of nuclear weapons constitutes an implicit threat which no one actually has to make. If the immorality lies in uttering the threat, then it may in practice be avoided—though one may wonder at the ease of this solution. Nuclear weapons, Ramsey writes, have a certain inherent ambiguity: “they may be used either against strategic forces or against centers of population,” and that means that “apart from intention, their capacity to deter cannot be removed from them. . . . No matter how often we declare, and quite sincerely declare, that our targets are an enemy’s forces, he can never be quite certain that in the fury or the fog of war his cities may not be destroyed.”21 Now, the possession of conventional weapons is both innocent and ambiguous in exactly the way Ramsey suggests. The fact that I am holding a sword or a rifle doesn’t mean that I am going to use it against innocent people, though it is quite effective against them; it has the same “dual use” that Ramsey has discovered in nuclear weapons. But the bomb is different. In a sense, as Beaufre has said, it isn’t designed for war at all.22 It is designed to kill whole populations, and its deterrent value depends upon that fact (whether the killing is direct or indirect). It serves the purpose of preventing war only by virtue of the implicit threat it poses, and we possess it for the sake of that purpose. And men and women are responsible for the threats they live by, even if they don’t speak them out loud.
Ramsey presses on. Perhaps the mere possession of nuclear weapons won’t be enough to deter some reckless aggressor. Then, he suggests, we must distinguish “between the appearance and the actuality of being . . . committed to go to city exchanges. . . . In that case, only the appearance should be cultivated.”23 I am not sure exactly what that means, and Ramsey (for once) seems reluctant to say, but presumably it would allow us to hint at the possibility of massive retaliation without actually planning for it or intending to carry it out. Thus we are offered a continuum of increasing moral danger along which four points are marked out: the articulated prospect of collateral (and disproportionate) civilian deaths; the implicit threat of counter-city strikes; the “cultivated” appearance of a commitment to counter-city strikes; and the actual commitment. These may well be distinct points, in the sense that one can imagine policies focused around each of them, and these would be different policies. But I am inclined to doubt that the differences make a difference. To rule out the last for moral reasons, while permitting the first three, can only make people cynical about one’s moral reasons. Ramsey aims to clear our intentions without prohibiting those policies that he believes necessary (and that probably are necessary under present conditions) for the dual prevention of war and conquest. But the unavoidable truth is that all these policies rest ultimately on immoral threats. Unless we give up nuclear deterrence, we cannot give up such threats, and it is best if we straightforwardly acknowledge what it is we are doing.
The real ambiguity of nuclear deterrence lies in the fact that no one, including ourselves, can be sure that we will ever carry out the threats we make. In a sense, all we ever do is “cultivate the appearance.” We strain for credibility, but what we are putatively planning and intending remains incredible. As I have already suggested, that helps make deterrence psychologically bearable, and perhaps also it makes a deterrent posture marginally better from a moral standpoint. But at the same time, the reason for our hesitancy and self-doubt is the monstrous immorality that our policy contemplates, an immorality we can never hope to square with our understanding of justice in war. Nuclear weapons explode the theory of just war. They are the first of mankind’s technological innovations that are simply not encompassable within the familiar moral world. Or rather, our familiar notions about jus in bello require us to condemn even the threat to use them. And yet there are other notions, also familiar, having to do with aggression and the right of self-defense, that seem to require exactly that threat. So we move uneasily beyond the limits of justice for the sake of justice (and of peace).
According to Ramsey, this is a dangerous move. For if we “become convinced,” he writes, “that in the matter of deterrence a number of things are wicked which are not,” then, seeing no way of avoiding wickedness, we will “set no limits on it.”24 Once again, this argument is precisely right with reference to conventional warfare; it catches the central error of what I have called the “war is hell” doctrine. But it is persuasive in the case of nuclear warfare only if one can describe plausible and morally significant limits, and that Ramsey has not done; nor have the strategists of “flexible response” been able to do it. All their arguments depend upon the ultimate wickedness of counter-city strikes. The pretense that this is not so carries with it dangers of its own. To draw insignificant lines, to maintain the formal categories of double effect, collateral damage, noncombatant immunity, and so on, when so little moral content remains is to corrupt the argument for justice as a whole and to render it suspect even in those areas of military life to which it properly pertains. And those areas are wide. Nuclear deterrence marks their outer limits, forcing us to contemplate wars that can never be fought. Within those limits there are wars that can and will and perhaps even should be fought, and to which the old rules apply with all their force. The specter of a nuclear holocaust does not invite us to act wickedly in conventional wars. Indeed, it probably is a deterrent there, too; it is hard to imagine a repetition of Dresden or Tokyo in a conventional war between nuclear powers. For destruction on such a scale would invite a nuclear response and a drastic and unacceptable escalation of the struggle.
Nuclear war is and will remain morally unacceptable, and there is no case for its rehabilitation. Because it is unacceptable, we must seek out ways to prevent it, and because deterrence is a bad way, we must seek out others. It is not my purpose here to suggest what the alternatives might look like. I have been more concerned to acknowledge that deterrence itself, for all its criminality, falls or may fall for the moment under the standard of necessity. But as with terror bombing, so here with the threat of terrorism: supreme emergency is never a stable position. The realm of necessity is subject to historical change. And, what is more important, we are under an obligation to seize upon opportunities of escape, even to take risks for the sake of such opportunities. So the readiness to murder is balanced, or should be, by the readiness not to murder, not to threaten murder, as soon as alternative ways to peace can be found.