It is a delicate matter to defend deterrence, the doctrine that it is the very lethality of nuclear weapons that lessens the likelihood of their use sufficiently to make us safe. That the consciousness of that lethality in the corridors of power in Washington and Moscow has played an important role in the keeping of the peace since the advent of the nuclear age is beyond doubting, as is the unwisdom of tampering with that consciousness, of accepting theories or technologies that will diminish the terror with which the prospect of nuclear war has been traditionally regarded and make nuclear weapons in any way less inhibiting to use. Still, if it is possible to underestimate the contribution that nuclear weapons make to the prevention of nuclear war, it is possible to overestimate it, too.

The essential fragility of deterrence must never be forgotten. Its central irony, that what may destroy us may be relied upon to deliver us, really is intolerable. That is the basis in reality for the criticism of deterrence from both the right and the left, for Ronald Reagan’s dream of ballistic missile defenses and for the peace movement’s dream of nuclear disarmament. Many of their arguments, to be sure, originate in moral, ideological and strategic assumptions that may be refuted. Against a great many alternatives, deterrence may be defended without apology. But there is at least one criticism of deterrence that may not be refuted: it may fail.


The failure of deterrence is a plain possibility. It does not require any form of scientific or strategic expertise to see it. Deterrence is a wager upon a broad variety of technological, military, political and diplomatic arrangements, any of which may collapse in a crisis; but at bottom it is a wager upon the human heart. Who would dare make do with such a wager? History gives no grounds for such a quantity of trust. Just a reading of Thucydides should suffice to rattle the most devout defender of deterrence. There is evil in man, and folly. There has never been a weapon that was never used; and many more weapons were used unjustly than justly. The tragic dimensions of human experience, then, must haunt the nuclear debate.

The idealists of the nuclear debate who plead for the abolition of nuclear weapons, or the abolition of war, or the abolition of power politics, or the abolition of the system of sovereign states, certainly appreciate the magnitude of the nuclear danger; but they do not correctly represent either human history or human nature with their perfectibilian proposals. Their idealism is the same as a counsel of despair. The magnitude of the nuclear danger means, rather, that realism—which is to say, deterrence—is the better way. A time in which there are 50,000 nuclear weapons in the arsenals of two great powers caught in political and philosophical competition is not a time to reform man. It is a time to manage him; that is, to deter him.

Yet deterrence may be complacently or uncomplacently constructed. Too many of its defenders construct it complacently. They avert their gaze from its flawed human origins, from the extent to which it is itself an expression for a state of emergency, from the luckless chance that it may fail. Consider some of the most influential American deterrers. Robert McNamara, for example, has stated categorically "that nuclear weapons serve no military purpose whatsoever. They are totally useless—except only to deter one’s opponent from using them." Mr. McNamara is correct, of course, that there is no meaningful military mission that can be accomplished by weapons of mass destruction, that the military means will always exceed the political ends. Perhaps there are no circumstances of any sort in which the use of nuclear weapons will be justified, though the revolution that such a belief wreaks with Western strategy should be appreciated.

What if the Russians, however, use them? Of what use will the truth about the military inutility of nuclear weapons be to the president who will be faced with the responsibility of nuclear retaliation? Presumably Mr. McNamara is not suggesting that the president not act at all. But he has not provided the president with any guidance about the courses of action that may be allowed after the deterrent use of nuclear weapons has disappointed.

Consider also a school of liberal strategists whose enthusiasm for deterrence has resulted in an enthusiasm for nuclear weapons. The most sophisticated member of this school, and an important influence upon the academic study of deterrence theory, is Kenneth Waltz. Mr. Waltz has written a close analysis of the many rewards of nuclear deterrence that concludes in a plea for nuclear proliferation:

The world has enjoyed more years of peace since 1945 than had been known in this century. . . . Nuclear weapons do not make nuclear war a likely prospect, as history has shown. . . . Much of the writing about the spread of nuclear weapons has this unusual trait: It tells us that what did not happen in the past is likely to happen in the future. . . . A happy nuclear past leads many to expect an unhappy nuclear future. . . . With more nuclear states the world will have a promising future. . . . Nuclear weapons make it possible to approach the deterrent ideal.

This is a little backward. Nuclear weapons are not there to create deterrence. Deterrence is there to cope with nuclear weapons. And the nuclear past has been peaceful, but it has not been exactly happy.

More examples could be cited of such complacency. There is a touch of liberal meliorism here, of liberalism’s familiar disinclination to think the darker thoughts about man. There is also a hidden apocalyptic premise. The rejection of the idea of operational thinking about nuclear weapons implies that the end of deterrence will be the same as the end of history. More specifically, it implies that once any nuclear weapons are used, all nuclear weapons will be used. The possibility that the arsenals may not be emptied, that a nuclear exchange may consist of an infernally small number rather than an infernally large number of launchings, is not considered. It implies, too, that immediately after deterrence fails, from the moment that a nuclear weapon is fired, there will be nothing left to save.

To be sure, the remorseless advance to annihilation that is euphemistically known as "escalation" is a very plausible way to consider the progress of a nuclear war; and much of the operational thinking about nuclear weapons is flawed precisely by an egregious refusal to entertain the possibility of escalation with sufficient seriousness. It makes sense to plan for the worst case. But it does not make sense to plan only for the worst case. In their planning for the post-deterrence predicament of the United States, many defenders of deterrence, well, they do not plan at all. Instead, as the British philosopher W.B. Gallie has remarked in a brilliant discussion of the apocalyptic approach toward nuclear weapons, "all less bad cases—all possibilities of forestalling or moderating the danger before it reaches its climax—are dismissed as delusive, as if on the principle that only the worst is bad enough for us!"

Thinking about the use of nuclear weapons, runs the distinguished liberal argument, makes the use of nuclear weapons more thinkable. You cannot consider the possibility that deterrence may fail without contributing to the likelihood of its failure. The argument is valid, as will be seen, against a good deal of official and unofficial wisdom about nuclear use. Still, the liberal evasiveness must end. Defenders of deterrence must have an answer to this most difficult question of all, for three reasons.

First, the absence of an answer weakens the argument for deterrence itself. As is well known, the argument has been challenged on the grounds of "credibility"; that is, it has been said that for the threat of retaliation to remain plausible, weapons systems must be developed and deployed that can make less destructive strikes against more discrete targets. It is not uncontrolled nuclear violence but controlled nuclear violence that the Soviet Union will really fear; a suicide threat, which is an integral part of the doctrine known as mutual assured destruction, will not be taken seriously. Thus the "flexible response" of the McNamara strategy, which culminated in the "limited nuclear options" of the Schlesinger strategy. To this rather exceptionable argument, however, the defenders of deterrence have replied robustly that if deterrence is "working," it is not because of "options," but because of the open-endedness of "options"; not because either the United States or the Soviet Union is persuaded that it can control the level of nuclear violence, but because neither the United States nor the Soviet Union is persuaded that anybody can control the level of nuclear violence. The point is not that escalation is always automatic, but that escalation is always possible; and until somebody proves that it is impossible, neither side will have a rational reason to start a war.

Still, the argument against deterrence is not thereby exhausted. The deterrer’s work is never done. There is another ground on which deterrence has been challenged, to wit, that it may fail, and that it is therefore morally, strategically and intellectually irresponsible not to foresee its failure. So it is. It is also a disservice to the theory of deterrence, and to the millions of people whom that theory is supposed to free from anxiety.

Second, the refusal to think beyond the success of deterrence makes such thinking of no consequence for policy. The position of the president is not like the position of the political scientist or the pundit. The president needs more than an analysis; he needs a policy. But too much of the theory of deterrence is conceived at a level of conceptual complexity that makes it utterly useless for the purposes of policy. It will not help to present the president, in the middle of a crisis, with game theory, or with murmurings about "mirrorings."

Third, it is essential that those who fully appreciate the enormity of nuclear war do not surrender the subject of nuclear use to those who do not fully appreciate it. While the failure of deterrence will certainly constitute the greatest mistake in human history, it may not constitute the last mistake. There are more wrong and less wrong, more catastrophic and less catastrophic things that can be done with nuclear weapons. It happens, in part due to the indifference of liberal strategists to the dirty details of nuclear strategy, that most of the operational thinking about the post-deterrence period has been of the more wrong and more catastrophic kind.

The insistence that nuclear weapons must never be used is of no help in a debate about the proper course of action once they have been used. An idea about the period of war cannot be refuted by an idea about the period of peace. Once the United States has suffered a nuclear strike, the president will have to make many decisions that will be not only about the automaticity of retaliation, about which the theory of deterrence teaches, but about the structure of retaliation, about which the theory of deterrence teaches not at all. What remains of nature and of culture after a war between the superpowers will depend upon the intelligence of these decisions.

The next stage in liberal thinking about nuclear weapons, then, must be the confrontation with the meaning of the fragility of deterrence—with its military meaning, and not merely its moral meaning. Is there a way to think about the failure of deterrence that cannot be accused of contributing to the failure of deterrence? That is the question that its defenders must try to answer. Obviously nuclear disarmament is an answer; deterrers, too, dream of disarmament. But until nuclear disarmament is achieved, in the interim that feels like an infinity, the question remains. As long as the power to destroy the world rests in the imperfect and infirm hands of men, provision must be made. And not just any provision; and probably not the provision that has already been made.


The need for policy has usually been interpreted as a need for strategy. Mutual assured destruction, it is argued, is not a strategy; for strategy, it is argued, means the purposeful application of military means to political ends. Of MAD this is certainly true. Indeed, MAD is nothing other than the negation of strategy, insofar as it denies that these military means can have a purposeful application to any political ends. The doctrine of mutual assured destruction, according to which nuclear use is essentially indiscriminate, can furnish no instruction to the military planner, whose job is precisely to discriminate. The planner must make choices about targets and weapons, and the order in which the latter will be dispatched against the former. He requires criteria for these choices. MAD obviously cannot furnish such criteria, since it considers all the choices to come to the same thing, which is the extermination of societies. Moreover, to the extent to which there is a prescription for policy made by MAD, it is to make the worst strike first, to begin a nuclear war with an attack on cities. This makes the situation of the planner morally intolerable, in addition to militarily intolerable; it puts the planner in the business of the slaughter of innocents.

The moral objection to MAD has been accompanied throughout by a military objection. Indeed, the United States has probably never had a pure version of MAD as an actual plan. Recent studies of American targeting policy deduce from the declassified documentation that, with the exception of a short-lived and bureaucratically inspired adoption of the doctrine of mutual assured destruction by the McNamara Pentagon, it has never been the American intention to strike first against cities. During the period that bombers were the means of nuclear delivery, the lessons of the Strategic Bombing Survey were not lost upon nuclear planners; and the notions of "flexible response" and "nuclear options" made their appearance shortly after missiles became the means. Thus there is no point in clinging to MAD as a strategy, because it never was one.

The planner, then, has a need for strategy. The failure of MAD to satisfy the need seems to signify a failure of the destructive difference of nuclear weapons to provide a different mode of military thought and military action. There is a new weapon, but there is not a new strategic idea. There is, however, an old strategic idea. That is, there is the tradition of classical strategy, to which the planner promptly repaired. Again, the poor planner’s predicament is not hard to understand.

He is frequently told that he must honor the difference between conventional weapons and nuclear weapons, but he is not told how. He is frequently told that nuclear weapons must never be used, but he is the man who must plan for the possibility of their use. And so nuclear strategy has developed under the sign of conventional strategy, and committed, from its beginnings, what Hans Morgenthau called "the conventional fallacy." From General Curtis LeMay’s comment that a nuclear bomb is just another bomb to the recent comment of one strategist that the targeting of control, command and communication centers by small intercontinental ballistic missiles would constitute "a form of maneuver warfare," there runs through the operational thinking about nuclear weapons a fateful consistency between the categories of conventional warfare and the categories of nuclear warfare. (The consistency is more pronounced, and less accompanied by ambivalence and argument to the contrary, in the Soviet strategic literature.) The most famous official formulation of "the conventional fallacy" was certainly Secretary of Defense Robert McNamara’s speech at the University of Michigan at Ann Arbor in June 1962:

The United States has come to the conclusion that to the extent feasible, basic military strategy in nuclear war should be approached in much the same way that more conventional military operations have been regarded in the past. . . . That is to say, our principal military objective in the event of nuclear war . . . should be the destruction of the enemy’s military forces. . . . Specifically, our studies indicate that a strategy which targets nuclear forces only against cities or a mixture of civil and military targets has serious limitations for the purpose of deterrence and for the conduct of general nuclear war.

This is the strategic doctrine that came to be called counterforce; and it is in the name of counterforce that the distinction between conventional weaponry and nuclear weaponry has been blurred more or less constantly from the Kennedy Administration through the Reagan Administration. The doctrine of counterforce targeting received its most rigorous treatment, and was adopted most unapologetically, in the "limited nuclear options" of the Schlesinger Pentagon. It persists in the "countervailing strategy" of Harold Brown and the "prevailing strategy" of Caspar Weinberger, which are the logical consequences of the reliance of nuclear strategy upon conventional notions by default, of the idea that the need for strategy may be met by strategy as it has been classically conceived. Both represent what Colin Gray tellingly calls a "war-fighting, or classical strategy, approach to nuclear deterrence [italics added]."

The "countervailing strategy" consisted, briefly, in the plan to make deterrence more "credible" by designing American nuclear forces of sufficient flexibility that they would successfully "put at risk" targets of critical significance for the Soviet Union, most notably their military forces and their structure of command, control and communication, and thereby successfully "deny" the Soviets any prospect of meaningful military success. Not the possibility of American victory, but the impossibility of Soviet victory, was the official rationale for the strategy.

The intellectual pathos of the doctrine is profound. It is a kind of halfway house between mutual assured destruction and the complete conventionalization of nuclear war. This accounts for its many confusions. Thus one official of the Brown years wrote in the same breath that the "countervailing" strategy "does not assume that a nuclear war is likely to be protracted" and that "the United States must act to improve its ability to conduct a sustained exchange." No such pathos may be discovered, however, in the "prevailing strategy" of Caspar Weinberger. It is less troubled than its predecessor by the problem of conventionalization, and speaks candidly about "prevailing" (in another formulation, of "prevailing with pride") in a nuclear war that would be fought "over a protracted period," with a distinct emphasis upon "decapitation," or the destruction of the Soviet military and political authority.

Both "countervailing" and "prevailing" have been sternly criticized. These strategies are indeed caught in many of the contradictions observed by their critics. There is the problem of escalation, of Clausewitz’s familiar "friction." The certainty that a nuclear war can be controlled is no greater than the certainty that a conventional war can be controlled, which was never very great; indeed, in the case of nuclear war there can be no certainty at all. And there is the problem of co-location. Too many of the enemy’s "counterforce" targets are too close to too many of the enemy’s "countervalue" targets; there is no way to destroy the Soviet military and its centers of command without destroying Soviet cities as well, which means that the Soviet Union will have little reason to interpret our "limited" strikes as a sign of restraint.

But there is a larger historical irony about the evolution of nuclear strategy that has rarely been remarked, and that accounts in large measure for its present problem. It is that the theory of counterforce preceded the technology of counterforce. The moral and military flaws in the destruction of cities were discovered many years before the weapons were discovered that could destroy anything less. Recent improvements in the technology of terminal guidance appear to have made the instruments of a limited nuclear strike finally imaginable. But a limited nuclear strike, not a limited nuclear war. The development and deployment of nuclear warheads of higher accuracy and lower yield will not put an end to the peril of escalation. Only the dismantling of the warheads of lower accuracy and higher yield will accomplish that; and it is a distant dream. To be sure, when the "ladder of escalation" concludes in the possibility of holocaust, it is not a small thing to add some more rungs to the ladder. But it is important to recognize, too, that the fundamental facts of nuclear life have not been altered.

So it is that we have arrived at the present moment in nuclear planning. Deterrence may fail. Therefore, nuclear policy. Therefore, nuclear strategy. Therefore, conventional strategy. Therefore, the fighting of a nuclear war.

But a nuclear war must never be fought, as Ronald Reagan likes to say. Is there a way, then, to break this melismatic movement from the fragility of deterrence to the fatality of a nuclear war? Is there a form of operational thinking about nuclear weapons that honors the difference between conventional weapons and nuclear weapons? Is there a strategy for the use of nuclear weapons in the aftermath of deterrence that seeks neither to "countervail" nor to "prevail"? Is there, for the planner and the policymaker, a lesser evil?

Perhaps there is.


There is one thing that would constitute a moral and rational course of action in a nuclear war. That is, to end it. The objective of nuclear strategy should be what strategists refer to as "war termination." Bernard Brodie made the point succinctly not long before his death: "The main war goal upon the beginning of a strategic nuclear exchange should surely be to terminate it as quickly as possible and with the least amount of damage possible on both sides."

Brodie’s proposal for the objective of operational thinking about nuclear weapons would seem to be utterly banal. Who could be opposed to the ending of a nuclear war? Yet a close examination of the pronouncements of the secretaries of defense of the past decade reveals some rather critical complications. James Schlesinger, be it said at once, seems to have been the policymaker most alive to the question of war termination. Indeed, the "limited nuclear options" that he forthrightly installed at the center of American nuclear strategy appear to have been designed partly for the purpose of allowing for some kind of armistice. In his annual report for 1975, which introduced the new construction of "flexibility," Schlesinger observed that: "What we need is a series of measured responses to aggression which bear some relation to the provocation, have prospects of terminating hostilities before general nuclear war breaks out, and leave some possibility of restoring deterrence. . . ." A year later, he stated with good sense: "Surely if there is only a small probability that limited response options would deter an attack or bring a nuclear war to a rapid conclusion, without large-scale damage to cities, it is a probability which . . . we should not foreclose."

And yet Schlesinger proceeded to vitiate substantially his stated objective for strategy. In his important briefings to the House Foreign Relations Committee in 1974, he uttered the words that have come to haunt the official position on war termination. The aim of American "flexibility," he said, was to stop the war "at the lowest possible level consistent with U.S. objectives." Those are not many words, but they take a lot back. This locution, which may be found in the official and unofficial statements of Schlesinger’s successors, is not as platitudinous as it may seem; indeed, as will become clear, it achieves nothing less than a restoration of the conventional fallacy to nuclear strategy.

Donald Rumsfeld presided over such a restoration rather unabashedly. He called for "a system of multiple . . . capabilities combined with a survivable control, command and communications network to permit the president to direct the strategic nuclear forces in a deliberate and controlled pursuit of national objectives." These objectives he defined grandiosely as the capacity "to retard significantly the ability of the U.S.S.R. to recover from a nuclear exchange and regain the stature of a twentieth-century military and industrial power more rapidly than the United States." "More rapidly" means, presumably, that the arms race will be succeeded by a recovery race.

Harold Brown had a less ambitious agenda for American strategic forces after the collapse of deterrence. While he confessed candidly, in his annual report for 1979, that "we cannot afford to make a complete distinction between deterrent forces and what are so awkwardly called war-fighting forces," he added that the objective of the latter would be "to ensure that no adversary would see himself better off after a limited nuclear exchange than before it. We cannot permit an enemy to believe that he could create any kind of military or psychological asymmetry that he could then exploit to his advantage." There is no consideration here of American advantage as an aim for operational thinking about nuclear weapons; only of an absence of Soviet advantage. The line is fine.

It is not fine at all in the post-deterrence plans of Caspar Weinberger, where the restoration of the conventional fallacy is pretty well perfected. In his annual report for 1984, for example, he numbers among the objectives of our nuclear forces "in the event of an attack, to deny the enemy his objectives and bring a rapid end to the conflict on terms favorable to our interests; and to maintain the political and territorial integrity of the United States and its allies." Later in the text he remarks that "should deterrence fail, our strategy is to restore peace on favorable terms." In his report for 1985, he is more ringing about the conditions he sets for nuclear war termination. "We must plan for flexibility in our forces and in our options for response, so that we might terminate the conflict on terms favorable to the forces of freedom, and reestablish deterrence at the lowest possible level of violence, thus avoiding further destruction."

These subclauses about "favorable terms," about advantage and disadvantage, smuggle into American plans for nuclear use in general, and into American plans for nuclear war termination in particular, the most characteristic principle of conventional thinking about war, which is the distinction between victory and defeat. I say smuggle, because no American administration has so far adopted a doctrine of victory in nuclear war. While the notion of victory has been renounced, however, the rhetoric of conquest has been retained. This has the consequence of diluting considerably the priority of war termination after the failure of deterrence.

War termination is an objective of American nuclear strategy only when it is the immediate objective of American nuclear strategy. No military or political goal of any kind may be allowed to obstruct the effort to end a nuclear exchange. When the annihilation of American society and Soviet society may have already begun, the capture of West Germany by the East or of East Germany by the West should be of little consequence. Nuclear weapons are supremely unsuited to the recapture of territory, which they can merely destroy. More important, the notion that the termination of a nuclear war should await the analysis of the results on the battlefield, that its timing should be determined by the prospects for the recovery of a military or political position, or more generally for a return to the status quo ante bellum, is a pretext for the fighting of a nuclear war, and nothing else.

The immediacy of the imperative to stop a nuclear war, the absolute nature of the notion of nuclear war termination, has been bitterly criticized: "the best prospect of all for minimizing (prompt) damage lies in surrendering preemptively." This is not merely demagoguery; it is also a dismal collapse into conventional categories of strategic thinking. True, the sort of war termination offer that must be made in a nuclear war may be seen as surrender in a conventional war. Still, there is a critical difference. In a conventional war there are winners and losers. A war termination offer made by a loser to a winner can be called surrender. In a nuclear war there are only losers. A war termination offer made by a loser to a loser cannot be called surrender. It can be called rational. Dispensing with another round of fighting is all that should be desired after deterrence fails.

The notion of war termination usefully sharpens, then, the distinction between conventional warfare and nuclear warfare. The distinction, indeed, may be defined by it. We may say that in a nuclear war war termination is a strategic objective. Of course this defies the Clausewitzian canons, according to which war termination is the result of calculations of profit and loss. If deterrence disappears, however, such canons and such calculations must disappear with it.


It is not enough to want to end a nuclear war. You must be able to. War termination must not only be posited; it must be planned. The mighty military, strategic and technological tendencies of America’s nuclear arsenal cannot be reversed or redirected in a time of crisis or a time of war. By then it will be too late. Thomas Schelling, one of the lonely few who have worked on the problem of nuclear war termination, has called it "the crucial challenge" and observed that "in earlier times, one could plan the opening moves of war in detail and hope to improvise plans for its closure; for thermonuclear war, any preparations for closure would have to be made before the war starts." There is no evidence, however, that the Pentagon, even during the tenure of officials who understood the immense importance of nuclear war termination, has developed plans to end the war with anything like the devotion that it has developed plans to fight the war. Moreover, the university, which frequently is the government’s quarry for precepts of policy, has likewise allowed the study of war termination to languish. The literature is scant; it is almost entirely about the termination of conventional warfare; and it is owed in large part to the difficult duration of the Vietnam War, which inspired a bout of serious academic interest in the late 1960s and early 1970s in the methods of ending a war. The state of the subject, then, inside and outside government, is something of a scandal.

It is not a subject to which this writer can contribute. The theory and practice of war termination must be worked out by those who have the scientific and strategic training. Still, I wish to propose that it be placed on their agenda, and that it be given a pride of place. And I will make bold to list a number of conditions for the swift and successful termination of a nuclear war that rather obviously suggest themselves:

The Principles of Strategy. A nuclear war will be terminated by those who have been schooled in the idea and the instrumentalities of war termination.

This factor—the mentality of the officials on whose watch deterrence fails—should not be overlooked. When deterrence fails, it will almost certainly fail with a limited nuclear strike. But a limited nuclear strike can mean many things. The technology of countervailing, prevailing and terminating is the same. What will matter, therefore, is what the employers of the technology think that they are doing with it. The history of warfare shows again and again the extraordinary influence of intellectual and physical drill upon military behavior. If the planners (and the politicians) with the authority to use nuclear weapons come into a crisis with the conviction that they can use them to countervail, to prevail, or to win, the conviction will carry the day. There will be no time for a discussion of the first principles of the nuclear world, not in a crisis, and certainly not in a war. But there is time now.

Targeting policy. The termination of a war will not be negotiated by a party that believes it has nothing to lose. A strike against cities, or more generally against a population, will almost certainly preclude the possibility of preventing any further strikes.

This is the good news, and all the good news, about counterforce. The theory of war termination suggests something important about (horibile dictu) the character of a first strike, which is that it must be seriously limited. A massive first strike will make war termination impossible (not to speak of all the other things it will make impossible). But in the initial stage of a nuclear war, at least, counterforce is not yet war-fighting. In fact, as Kenneth Waltz has correctly observed, it is perfectly compatible with what he calls, oxymoronically but on the mark, "a deterrent strategy," a strategy for nuclear use that commences with "a few judiciously delivered warheads" that hold in reserve the threat of the society’s final destruction and "are likely to induce sobriety in the leaders of all the countries involved and thus bring rapid de-escalation." (To this extent, moreover, a "countervailing strategy" may be tolerated; that is, to the extent to which the negotiation of an end to the war will require that the Soviets, too, perceive no meaningful "advantage" from nuclear use; but that is all the confidence in countervailing that seems justified.)

Nor will the termination of a war be negotiated by a party that believes it has no bargaining power. A strike against strategic forces, the "first strike" that has featured prominently and passionately in the recent debates about the "window of vulnerability" and the MX, will certainly not completely succeed, if only because the accuracy of an ICBM should not be confused with the reliability of an ICBM; but even a partial success could leave the other side with the feeling that it has been robbed of the strength it would need to negotiate a halt and not a surrender. The scale of a counterforce attack, then, will make a difference, too. A very large counterforce strike would probably induce a feeling of despair that would tempt the merely human command of the side that was struck into an orgy of retaliation. It is unused weapons, and not destroyed weapons, that will provide the assurance necessary for the conclusion of an armistice.

Force Design. The scale of the counterforce attacks that may be made must not be prejudiced by the design of our weapons.

The fewer warheads on a missile, the better. The arming of missiles with multiple warheads (MIRVs) made them more attractive targets. The development of ICBMs in the direction of ten warheads to a missile (and more, according to some excessively excited planners) makes large counterforce attacks more attractive, and thus more likely.

Similarly, the higher the accuracy and the lower the yield of strategic warheads, the better. Not for the purpose of making deterrence more "credible," and not for the purpose of making a nuclear war "winnable," but for the purpose of keeping the apocalypse at bay for as long as possible. To be sure, the "low" yield that is sometimes spoken of by strategists is not quite low enough to nullify our nightmares. Different types of nuclear weapons mean different types of horrors and nothing else. Still, the chances of concluding the war go up as the dimensions of the catastrophe with which it began go down.

More important than not deploying a first-strike weapon is not deploying a first-strike force. Such a force would furnish the technological foundation for the misguided notions that the nuclear forces of the other side can be completely removed, and that they can be removed without murdering the number of innocent people that would transform a counterforce motive into countervalue action. A first-strike force, a force of vulnerable land-based MIRVed missiles, like the MX, deployed beyond a certain number, would tempt the Soviet Union (and the United States, since the Soviet ICBM force is equally vulnerable and more than equally deadly) into the kind of attack that will invite, no, implore, escalation.

Control and Command. The termination of a war can be negotiated only with those who have the authority to negotiate it.

For the purpose of preventing escalation and putting an end to hostilities, nothing could be more stupid than the destruction of the political and military authority of the enemy. In 1945 the United States chose to bomb Hiroshima and Nagasaki, and not Tokyo. But the destruction of such Soviet authority has been American policy for years. Despite Harold Brown’s thoughtful observation, in his annual report for 1982, that "we recognize the role that a surviving supreme command could and would play into the termination of hostilities and can envisage many scenarios in which the destruction of them would be inadvisable and contrary to our best interests," the Carter Administration finally adopted Presidential Directive 59, which placed a high priority on the targeting of the Soviet leadership. Top priority accorded to the destruction of Soviet command and control was codified by the Reagan Administration, where it is called "decapitation."

The initial stage of a nuclear negotiation will be tacit; it will consist in the correct identification of a counterforce attack, of the restraint that it implies. The attack itself will be a message. (For this purpose, the instruments of attack assessment must be secure.) Thus, the telephone must be spared. James Schlesinger has reflected rather straightforwardly that:

. . . if we were to maintain continued communications with the Soviet leaders during the war, and if we were to describe precisely and meticulously the limited nature of our actions, including the desire to avoid attacking their urban-industrial base, that in spite of whatever one says historically in advance that everything must go all out, when the existential circumstances arise, political leaders on both sides will be under powerful pressure to continue to be sensible.

The telephone, however, must work. There is probably no more pressing need for American research and development than the creation of redundant lines of communication that will remain absolutely secure. We must be able to communicate not only with our enemy, however, but also with ourselves. A nuclear war must not be the work of the submarine captain who believes that deterrence has failed and has lost contact with his superiors.


Having made the argument for war termination as the proper objective of nuclear strategy, I must make the apology for the argument. There are practical problems. The most formidable is that the Soviet Union may not cooperate. While Soviet strategists have devoted much more energy to the planning for the period after deterrence, there does not seem to be much in their planning about war termination. Quite the contrary; the Soviets seem to think (judging by what has been declassified and translated into English) that a nuclear war should be fought exactly like a conventional war. They are not against deterrence, as the evidence of our continued existence shows. Yet they are not particularly troubled by the conventional fallacy, either.

The alarming asymmetry between the logic of counterforce and the logic of countervalue must be grasped in its full force. It takes two to restrain a war, but only one to go all the way. If the Soviet Union chooses to treat the failure of deterrence like the outbreak of a conventional war, if it strikes first and strikes massively, our own provisions for something different will not avail us. To be sure, we must do what we can to influence Soviet strategy; but beyond the influence of our ability to annihilate Soviet society, there may be little that we can do. We can persuade the Soviets to keep the peace much more easily than we can persuade them to fight a certain kind of war. This is one of the most sobering facts of strategic life.

There is another risk that must not be overlooked. It is that the adoption of war termination as a policy may weaken deterrence slightly, in the sense that the Soviets will fear the outcome of an armistice less than the outcome of mutual destruction or fighting to "victory." Still, as Pierre Hassner has rightly observed, a certain loss in deterrence may be accepted for the sake of a certain gain in the possibility of terminating a war. The risk is probably justified, since (in Soviet eyes, too) the prospect of escalation continues, all plans of any kind notwithstanding, to do the work of deterring.

But the primary problem with the idea of war termination is not practical. It is philosophical. In the end, a faith in American rationality may be no more justified than a faith in Soviet rationality, since it is a faith in human rationality. Moreover, it is probably the most groundless faith in human rationality that may be imagined, since it follows upon the most irrational act in human history, which is the first use of a ballistic missile. Moreover, the second use of a ballistic missile, which the doctrine of deterrence demands, may be the second most irrational act in human history. For this reason I have assumed that the second use will occur (though an argument against it may be made); while it may have no rational reason, it may be counted upon to have an irrational reason, to wit, revenge. How we get into a nuclear war is a subject about which much has been written and thought. How we get out of a nuclear war, if indeed we do, is a different subject. It is against the third use, then, that I would want us to plan; and, if there is indeed a difference between the failure of deterrence and the end of the world, even more desperately than we have planned against the first.

It must seem like reliance upon a broken reed to recommend that the failure of deterrence will culminate in deterrence. But that is just the recommendation. "Intra-war deterrence," as we have seen, is the official idiom of war termination; and the title is apt. Certainly this constitutes an even wilder wager than deterrence itself. This wager assumes that there will appear an atmosphere of trust in the wake of a nuclear exchange. The credibility of the threat to escalate, to carry out the punishment of society that a more sensible strategy had put off, it is said, will persuade both sides to negotiate a cease-fire. But a cease-fire requires a promise. What about the credibility of the promise? Who will take the word of a nuclear warrior?

War termination, then, is finally an argument for the formulation of rules for nuclear combat: rules of nuclear engagement and rules of nuclear diplomacy. It is an expression of what Hans Morgenthau called "a humanitarian impulse within the framework of an utterly inhuman enterprise." That is asking a lot. Still, the same may be said of the countervailing strategy, and of the prevailing strategy. They, too, are a call for rules, for some kind of mutual understanding that the worst strikes will not be made first, that retaliation will be proportionate to aggression, and so on. Except that they would take the outbreak of nuclear hostilities all the way into a nuclear war; and they would make no provision for a prompt attempt to limit the damage, not by quitting while we are ahead but simply by quitting. War termination, to be sure, is a version of restraint; but it turns out to be much less Panglossian than the other versions of restraint, since it abjures any notion of advantage. For that reason the idea of war termination casts an important light upon the idea of mutual assured destruction. It accepts calmly the strategic futility of MAD. But it does not reject MAD altogether. For, if MAD is a bad description of nuclear strategy, MAD is a good description of nuclear reality. Indeed, it is the accurate name of our condition. As long as we continue to possess this terrible power, we continue to live at the risk of mutual assured destruction. History may be periodized from the moment that the atom was split; and no refinement of counterforce technology can do anything about it. It is precisely because the world is MAD that war termination is the only morally and practically sound objective of nuclear strategy. As long as the possibility of escalation to extinction is present, a nuclear exchange must instantly be ended. The theory of war termination is not lulled by MAD, then, into ignoring the chance that deterrence may fail; but it insists, too, upon MAD’s profound truth.

The theory of war termination has an interesting analytic role as well. It helps to show the sense in which the distinction between deterrence and war-fighting is not empty, as some have maintained. War-fighting means more than merely the use of nuclear weapons; war-fighting means the continued and purposeful use of nuclear weapons. There is a difference between a nuclear exchange and a nuclear war: a difference in the degree of damnation. The continued use of nuclear weapons constitutes a nuclear war. It represents a rejection of the fear that is implicit in both deterrence and war termination, in favor of a false ambition for military and political profit. The countervailing strategy and the prevailing strategy are not the dark side of deterrence, as their apologists claim, or its corollaries; nor are they historically inevitable when deterrence fails. They are conscious strategic choices based upon specific strategic assumptions. They may be challenged in the name of other strategic choices based upon other strategic assumptions. The theory of war termination makes such a challenge. It provides an opportunity to break the monopoly of the war-fighters on the operational dimension of the nuclear question. Not all the strategists who think about the failure of deterrence are nuclear war fighters. Only most of them are.

War termination may provide a solution to the intellectual perplexity of liberal strategists and to the strategic perplexity of nuclear use. It is indeed that rare species of nuclear speculation, a way of thinking about the failure of deterrence that cannot be accused of contributing to its failure. Unlike some of the fantasies that have settled comfortably in the seats of American and Soviet power, war termination is strenuously sober about the technological and strategic circumstances in which we live. It denies neither the reality of counterforce nor the reality of escalation.

To be sure, it may turn out to be a fantasy, too. If only because of the speed with which a nuclear war will begin—the speed that may prove to be the most disastrous dogma of strategy in the nuclear age—war termination may prove to be physically and psychologically impossible. The worst may not happen first, but it may happen fast. Perhaps the apocalyptics among us are right. But perhaps they are not. In any event, there is no reason to ensure that they are right. The requirements of war termination are worth a stupendous intellectual, technological and financial effort. The chance of putting an end to a nuclear war between the United States and the Soviet Union may be slim, but it is all the chance that the rest of history may have.

You are reading a free article.

Subscribe to Foreign Affairs to get unlimited access.

  • Paywall-free reading of new articles and a century of archives
  • Unlock access to iOS/Android apps to save editions for offline reading
  • Six issues a year in print, online, and audio editions
Subscribe Now
  • Leon Wieseltier is the Literary Editor of The New Republic and the author of Nuclear War, Nuclear Peace. The author wishes to thank Pierre Hassner, Robert W. Tucker and Michael Walzer for their comments.
  • More By Leon Wieseltier