The following is a revised and enlarged version of the Walter E. Edge lecture at Princeton University on November 15, 1983.


There are so many aspects of the nuclear war problem that anyone who talks about it, especially in a relatively short span of time, must choose a particular aspect to deal with. It is all too easy to get tangled up in terminology, technicalities, or the controversies of the moment. I am going to discuss what I consider to be the chief danger or threat of nuclear war. It is what I call “nuclear temptations,” by which I mean nothing more than the temptations to use nuclear weapons. These temptations have taken various forms, some of which are still with us. But temptations in this field, as in life generally, come with inhibitions, and so one will naturally lead to the other as we go along.

But first I wish to ask a potentially embarrassing question: What business do people like ourselves have to discuss the problem at all? We are not nuclear experts or military professionals. Why, then, should we think that we have a right to speak or even to think about it?

The answer to this question was given almost exactly a century and a half ago by Karl von Clausewitz, whose book is still the most profound analysis of the nature of war. One of Clausewitz’s best-known rules of war goes as follows: “Everything in strategy is very simple, but that does not mean that everything is very easy.”1 By strategy, Clausewitz meant the grand, the fundamental, lines of a military plan or conception, not the manner in which it was executed.

In the Clausewitzian sense, the grand strategy of nuclear war is also simple, though that does not mean that everything about it is very easy. Its strategic simplicity is what permits us—non-nuclear experts and nonmilitary professionals—to think seriously about the problem of nuclear war. One such simple conception is that of nuclear “deterrence,” and it did not need nuclear experts or military professionals to think of it. In fact, it was first conceived by people very much like ourselves. The full story of how and when the idea of nuclear deterrence was first conceived remains to be told. I am merely going to touch on it to bear out my point about nonexperts and the nuclear war problem.

The first atomic bomb was dropped on Hiroshima on August 6, 1945. Only a month later, President Robert Hutchins of the University of Chicago invited a remarkable group of scientists, economists, sociologists, and government officials—forty-six in all—to an “Atomic Energy Control Conference.” It was held two weeks later, between September 19 and September 22. Two of the participants are of the greatest interest to us.

One was Professor Jacob Viner, then Professor of Economics at the University of Chicago. Viner was, of course, no atomic expert or military professional. But he already knew one thing, as he put it in a talk two months later: “A single atomic bomb can reduce a city and its population to dust.”2 From this simple enough premise, Viner deduced a remarkable number of consequences, only one of which I need mention here.

In his talk at the Chicago conference, which was supposed to be on “Economic Implications for Strategic Foreign Policy,” Viner made a striking allusion to the strategic military implication of atomic warfare. Though the United States was then the only country to have the bomb, Viner already foresaw that the monopoly could not last and, moreover, that a stage of parity, or equal destructiveness, was bound to come about.

Here is a key sentence in a memorandum he wrote for the conference: “Retaliation in equal terms is unavoidable and in this sense the atomic bomb is a war deterrent, a peace-making force.”3 The term “deterrent” was thus used for the first time in this connection. In his later talk, in November 1945, Viner developed the thought behind his original insight. That thought might be summed up in this way: If one atomic power is capable of retaliating in kind against another atomic power, each is capable of deterring the other from using atomic weapons.

The second pioneer in this field was Bernard Brodie, then at Yale University. Brodie had specialized in the history of naval warfare, but he too was no nuclear expert or military professional. He also attended the Chicago conference, where he spoke directly on “Strategic Consequences of the Atomic Bomb.” Brodie had been a student of Viner’s and years later paid back his debt by declaring that “whatever is good in my work owes something to him.”4 Brodie’s talk contained a mention of “possible deterrent value” in connection with the atomic bomb.5 The word “deterrent,” therefore, had crept into Brodie’s thinking at this time, but the summary of his talk on record is too cryptic to reveal fully what he may have meant.


But whereas Viner, as far as I can tell, dropped the subject after 1945, Brodie stayed with it for the rest of his life. In a book published in 1946, Brodie produced the classic formulation of the military consequence of atomic weapons. Three of his sentences have been quoted innumerable times, but I may perhaps be excused for quoting them once again, because they sum up the essence of the matter better than anything else:

Thus far the chief purpose of our military establishment has been to win wars. From now on its chief purpose must be to avert them. It can have almost no other useful purpose.6

Brodie thus hit on the main point—the strategic essence—that this weapon was capable of such mutual destruction that it could have no useful purpose except to make war between atomic or thermonuclear powers irrational and suicidal.

Previous wars had been rational to the extent that they had served some political purpose. In fact, this had been Clausewitz’s main point. Here, again, he had something basic to say to us today:

No one starts a war—or rather, no one in his senses ought to do so—without first being clear in his mind what he intends to achieve by that war and how he intends to conduct it.7

Let us go over Clausewitz’s proposition more slowly, with special relevance to nuclear war:

No one starts a war…without first being clear in his mind what he intends to achieve by that war….

What would a nuclear power intend to achieve by waging a nuclear war against another nuclear power? Win the war? But what does victory mean if the enemy has a nuclear force that can retaliate in kind? What is victory worth if the nuclear aggressor must take the risk—at the very least the risk—of mutual devastation?

Then the second part of Clausewitz’s proposition:

No one starts a war…without first being clear in his mind…how he intends to conduct it.

But how is a nuclear war to be conducted? This question immediately raises another: Do nuclear weapons represent a qualitative rather than a quantitative change in warfare? If the answer is qualitative, as most theorists agree, there is no experience or precedent for conducting it. Armed forces learn how to use new weapons by trial and error. The function of the machine gun, for example, was misconceived at the outset. It took time and failure to learn how to use the machine gun to best advantage. But no one in his senses—to use Clausewitz’s language—is going to be able to learn how best to use nuclear weapons, especially those of the greatest destructiveness, by a process of trial and error. The risk and the cost make such experimentation in real combat prohibitive.

Some such reasoning flowed from the fundamental insight first stated by Viner and Brodie. Their analysis has been refined; various distinctions have been introduced; but the central idea of deterrence is still very much with us. It was recently restated by former Secretary of Defense Robert McNamara, who was not always such a firm believer. In an article the newspapers played up as if he had said something new and original, McNamara wrote that nuclear weapons “are totally useless—except only to deter one’s opponent from using them.”8 Viner and Brodie had been there thirty-seven years earlier. For this they should be properly honored.


But the idea of deterrence has peculiar flaws in it. It is by its very nature not a strategy for waging war; it is rather a nonstrategy or an antistrategy. In all previous human history, weapons have been invented to be used; the more effective they were, the more they were used. But nuclear weapons are too effective to be used. This paradox is almost too much for the military mind, and its civilian counterparts, to bear.

This nuclear nonstrategy is also very expensive. It devours untold sums of the national budget, incredible quantities of rare and costly materials, prodigious amounts of human knowledge and ingenuity. And all this for something that cannot or should not be used? All this for something that is merely intended to deter someone else from using it?

Another problem with deterrence is linguistic. In ordinary usage, deterrence is treated as if it were a thing, a doer, an active agent. The question is asked, “If deterrence fails?”—as if deterrence by itself can fail or succeed. This reification of deterrence is natural enough, but it can be grossly misleading and distorting. The concept of deterrence is nothing more than a mode of analysis, a shorthand for a relationship between nuclear weapons and political decision making. The weapons do not make decisions to be used or not to be used; political leaders make the decisions. If nothing can be gained from using these weapons, the political leaders will or could be deterred from using them. If “deterrence” fails, the political leaders—or those who have elected or tolerated them—will have failed. As long as hostility and rivalry persist, weapons will be their instruments, not their causes. The real success or failure of deterrence depends on political thinking, not on unthinking objects that obey the will of men.


Deterrence, then, is not a stockpile of weapons. Just here is the source of the greatest misunderstanding of and discomfort with the whole doctrine of deterrence. Because a stable balance of nuclear weaponry is necessary for the deterrent effect, it is all too common to transfer the hatefulness of those weapons to the balance that inhibits their use. Those weapons will not disappear however much we may fear and detest them; the idea of deterrence cannot be blamed for bringing them into existence or making them continue to exist. To reject deterrence because the weapons are rejected is to give up one of the few—perhaps the only convincing rationale why the weapons cannot be used for any sane, credible purpose. It is the link with politics that is the best hope of deterring the employment of nuclear weapons.

Even here, however, there is a hitch. The same nuclear weapon that can be used for deterring can also be used for fighting. This ambiguity is inherent in nuclear weapons, and we may have to live with it as we live with many lesser ambiguities in our lives. Some people seem to think that they have to be against deterrence because they are against nuclear weapons, as if one has to approve of nuclear weapons because one may favor a policy of deterrence. The real dividing line should be between those who wish to give nuclear weapons a war-deterring and those who wish to give them a war-fighting role. The distinction cannot be determined in practice solely by whether nuclear weapons are involved; they are inevitably involved in both cases. We must draw the line by determining whether the level of nuclear weaponry far exceeds the requirements for deterrence, whether the types of weapons are far more suitable for war fighting than for war deterring, and whether the official strategic doctrine encourages or requires the use of nuclear weapons. In this case as in so many others, policy-makers are far more apt to give themselves away by their actions than by their words.

In some sense, then, deterrence is psychological or, better yet, political. It depends on the political calculation of political leaders who might want to use nuclear weapons for political gain. If there could be no gain, there would be no point in their use. Nevertheless, the equation is always going to be made by human beings, not by the weapons themselves. There is something scary about the responsibility of mere human beings for such ultimate decisions. It is for this very reason that rationality is our best hope and guide in this awful predicament. By rationality, I do not mean to suggest the need for any great feat of wisdom or foresight; nothing more is needed than the rational will to survive—and the leaders of even the most aggressive and detestable regimes have that. The worst way of dealing with the problem is to build up an atmosphere of mindless terror, which almost surely leads to fatalism or abdication.

We would be in an even worse position than we are in already if we though that deterrence worked its wonders automatically and inevitably. Nothing could be more dangerous than taking deterrence for granted. We would then be really vulnerable to accidents and misjudgments; that they can happen is what must make us guard against them or at least take the necessary precautions to minimize their damage. The subtlety of Bernard Brodie’s mind was never better displayed than in this little-noted observation in his late, great work, War and Politics:

It is the curious paradox of our time that one of the foremost factors making deterrence really work well is the lurking fear that in some massive confrontation crisis it might fail. Under these circumstances one does not tempt fate.9

The temptation is to say: Since no conceivable use of nuclear weapons makes sense, let’s get rid of them. We don’t go on building cars or computers or anything else in order not to use them. Why go on building nuclear weapons in order not to use them?

The logic is impeccable; the reality is something else. In the first place, all nations having nuclear weapons would have to get rid of them altogether, simultaneously. There is no serious chance of that happening.10 More than that, it would be necessary to get rid of the knowledge of how to make nuclear weapons. That is not even worth dreaming about; it is out of the question. There are a great many things in life easier to get rid of than nuclear weapons—cigarettes, for example. No one expects us to get rid of cigarettes—or cigarette smokers—all at once, universally and simultaneously. Probably no one expects us to get rid of them ever.

So we have been left with the nonstrategy of deterrence. It amounts to this: We have nuclear weapons; we are going to have them; but the weapons themselves are of such a nature that we—and other nuclear powers—dare not use them.

It is not the most elegant or satisfying solution—if it can be called a solution. As long as these weapons exist, there is always the danger that they may be used. At best, deterrence belongs to the lesser-evil, or faute-de-mieux, variety of human conduct. No one likes to choose the lesser evil—but isn’t that what we are doing most of the time in our lives? We rarely get the chance to choose between the perfect good and the ultimate evil. The problem of the lesser evil is not that it is less but just how much less it is. If a lesser evil is infinitely preferable to the greater evil, we are lucky to have it. So let us not scoff at lesser evils, even in the case of nuclear arms.

Nevertheless, the lesser-evil nature of nuclear deterrence makes it vulnerable to pulls from two directions. One is the utopian—simply to do away with these weapons altogether, unilaterally or universally. The utopian program is not my subject, so I do not have to discuss it. The other pull, however, is my subject—it is “nuclear temptations.”


It is best to start with a brief history of these temptations.

The original temptation came in with the defense strategy of NATO. The temptation arose from the problem of finding a means of defending Western Europe against what was conceived to be an overwhelming Soviet advantage in conventional, or non-nuclear, forces. NATO adopted a military policy of using atomic weapons to deter or to defeat a conventional attack.

That is still official NATO doctrine. General Bernard Rogers, the Supreme Allied Commander in Europe, recently declared that the alliance, if attacked conventionally, would have to resort to nuclear weapons “fairly quickly,” unless its conventional strength were considerably increased, which seems to be unlikely.11 I do not wish to linger on this—the original—nuclear temptation, because it is an old story. At this point I merely wish to point out that this policy was adopted when the Soviets did not have or were far behind in nuclear weapons; the same policy, however, prevails in a condition of nuclear parity, that is, the “retaliation in equal terms” that Jacob Viner predicted in 1945.

This form of deterrence also has a serious flaw. The most controversial of nuclear weapons are those that can reach the Soviet Union from the United States and the United States from the Soviet Union—the ICBMs, or intercontinental ballistic missiles. They are known as “strategic deterrents”—a misuse of the term “strategic,” which, unfortunately, has by now become customary. A recent official US statement affirms: “Ultimately, the most important link is that between forces in Europe—both conventional and nuclear—and the US strategic deterrent.”12

It has long been difficult to believe in the efficacy—or, as the current jargon has it, “credibility”—of this deterrent. No president, it has been argued, would risk the devastation of the United States in defense of someone else’s territory, even that of Western Europe. The first obligation of an American president is to the safety of the United States, whatever other obligations there may be. The argument goes that the use of the strategic deterrent against the Soviet Union would lead to the use of the Soviets’ strategic deterrent against the United States, with the result of mutual devastation. So one of the best students of the nuclear war problem—English, not American—has flatly stated: “The United States would be irrational to commit suicide on behalf of Western Europe….” 13 Charles de Gaulle had said more or less the same thing at least two decades ago.14 It is not a new problem, but no one has yet thought of a good answer.


So now we have at least two serious problems with traditional deterrence—it depends on a huge investment in weapons that will never be used; and it assumes the willingness of an American president to risk the devastation of the United States in behalf of other countries, albeit allies. There are other problems, no doubt, but these two are enough for our present purpose.

For at least three decades now, efforts have been made to get around such problems. It is these efforts that have brought about the most tempting inducements to use nuclear weapons.

First, there were the so-called tactical nuclear weapons, which are still with us. Tactical nuclear weapons, developed in the early 1950s, were small enough to be used on a battlefield. NATO officially decided in 1954 to use tactical nuclear weapons in defense of Europe. The Soviets then introduced tactical weapons on their side. Thus a distinction was created between “strategic” and “tactical” nuclear weapons. In effect, there were now “bad” nuclear weapons—the strategic—because they were most useful against cities and civilian populations, and there were “good” nuclear weapons, because they could be used within a much smaller area by enemy units in combat.

The development of tactical weapons brought with it another temptation—that of “limited nuclear war.” A limited war would have to be “controlled,” so now we had another tempting proposition—the “controlled” and “limited” nuclear war. If, in effect, the only weapons used were the smaller, less destructive ones, a nuclear exchange could be said to be “controlled” and “limited.” The two terms were really interchangeable because a nuclear war would have to be controlled to be limited and limited to be controlled.

Now, in order for a nuclear war to be limited and controlled, it was necessary to fit the technology to the strategy. The new technology was supposed to produce nuclear weapons that were precise enough to be “discriminating.” That was the favorite word—“discriminating.” The Hiroshima type of weapon had been hopelessly imprecise and undiscriminating. It was effective only against big targets, such as cities. But the smaller, tactical weapons were now touted as precise and discriminating, so that they could avoid mass destructiveness.

For example, Henry Kissinger first attracted widespread attention in 1957 with a book called Nuclear Weapons and Foreign Policy. It advocated a policy of waging limited nuclear war, largely based on the new tactical weapons. Kissinger then sought “to break down the atmosphere of special horror which now surrounds the use of nuclear weapons” and “to overcome the trauma which attaches to the use of nuclear weapons.” And what kind of weapons were needed to accomplish these goals? They had to be, Kissinger then thought, both “destructive” and “discriminating.”15

A great fuss was made about the limited-war doctrine in the late 1950s. Its nuclear version was so shaky that Kissinger—to his credit—repudiated his earlier position in another book in 1961, only four years later. In 1957, he had come out in favor of limited nuclear war and against conventional war;16 in 1961, he came out against limited nuclear war and in favor of conventional war, with the use of nuclear weapons only as a last resort.17

Yet the technological solution to the nuclear dilemma continued to captivate. In 1973, Fred Charles Iklé, the present under secretary of defense for policy, advocated “taking advantage of modern technology” by exploiting “the potential accuracy of ‘smart’ bombs and missiles.” He wanted “assured destruction” of the enemy’s “military, industrial and transportation assets” instead of “the killing of vast millions,” as if one were the antithesis of the other. This strategy, Iklé admitted, was not an alternative to deterrence; it was nothing more than a change of form, not of substance.18 Thus the technological nostrum for conducting nuclear war in a way to avoid the destruction of cities and the massacre of civilians is hardly a new idea; it goes back at least a quarter of a century in one form or another, as if we could return to a nuclear version of the premodern art of warfare.

What was wrong with those “discriminating” and “accurate” nuclear weapons? In the first place, they were not discriminating or accurate enough. In June 1955, NATO held an exercise to find out what casualties might result in tactical nuclear warfare. In less than three days, it was found, 1.5 to 1.7 million people would expect to be killed and 3.5 million wounded if only 268 bombs fell on German soil. The rate of German casualties would be five times that suffered in World War II as a whole. In 1960, NATO maneuvers in Schleswig-Holstein showed that between 300,000 and 400,000 civilian deaths were to be expected within forty-eight hours of the initiation of tactical nuclear warfare. These figures did not take into account the effects of radiation and follow-up diseases. At that time the future German Chancellor Helmut Schmidt was moved to protest: That the concept of tactical nuclear warfare “should remain in force is inconceivable.”19

Something else was even more troublesome and embarrassing. Where was this limited war with tactical nuclear weapons going to be fought? The obvious answer was: in Europe and, primarily, in Germany. On second thought, the Europeans in general and Germans in particular were not enamored of this prospect. A limited nuclear war is a war limited to Europe. That circumstance has always made a limited war more attractive to Americans than to Europeans.

The problem of tactical nuclear weapons was soon compounded by that of so-called intermediate nuclear weapons. Intermediate weapons may be defined as those with a range great enough to reach the Soviet Union from Western Europe and to reach Western Europe from the Soviet Union. They are thus classified somewhere between the tactical nuclear weapons, with a range short enough to be used on a battlefield, and the strategic nuclear weapons, with an intercontinental, or Soviet-American, range. The Soviet SS-20s and the American Pershing IIs, much in dispute today, are typical of the intermediate weapons.

The Pershing IIs illustrate a point that I have previously tried to make about the ambiguity of nuclear weapons. One reason the European members of NATO originally wanted Pershing IIs was that they were supposed to “couple” European and American nuclear weaponry. It was reasoned that a threat to the Pershing IIs would be regarded in the United States as great enough to bring into play the “strategic,” or intercontinental, weapons in the United States. But then doubts arose. It was also charged that the Pershing IIs could just as well “decouple,” or dissociate, the nuclear defense of Europe from that of the United States. An alarm was raised that an American president was more likely to accept Soviet retaliation against weapons in Europe than against those in the United States.

Which is right? It seems to me that either one may be right—or wrong. The weapons themselves will not couple or decouple; they can be used for either purpose. The decision will be made politically in circumstances we cannot now foresee. Ambiguity and uncertainty hover over almost every aspect of the nuclear question.20 Anyone who can believe that the United States would passively permit the domination or destruction of Western Europe could be convinced of the decoupling theory: those who find it hard to imagine that the United States would not regard such an attack on Western Europe as a mortal threat to itself will lean over to the coupling side. The irony is, however, that the decision will not be made because there are a few hundred more intermediate-range weapons in Europe; the same decision would be made without them, because the United States is coupled with Western Europe by interest, culture, and geopolitical imperatives, not by any particular weaponry.


In any case, temptations soon came in new guises. Four terms characterize the variations on the old theme—options, escalation, flexible response, and counterforce.

The idea of “options” was brought forward to get rid of the nightmare of all-out nuclear war. It held that the president had to have something between all or nothing to defend against a Soviet attack. The concept of “options” seemingly left everything wide open, from any kind of conventional war to any kind of nuclear war or a combination of both.

The temptation to use nuclear weapons, if necessary, was therefore inherent in the concept of “options.” But when might it be necessary? It was hoped that a war could be waged, at least at the outset, at the lowest level of violence, in a conventional manner. But the losing side, it was also realized, was bound to try to overcome some disadvantage by a process of escalation, that is, by bringing new forces or weapons into play.

Thus “options” and “escalation” were intimately related: the process of escalation was bound to result in the optional use of nuclear weapons. Abstractly, one could envisage the process of escalation going from conventional warfare to the use of tactical nuclear weapons, then to intermediate nuclear weapons, and finally to strategic nuclear weapons.

But “options” and “escalation” were really parts or aspects of a more basic doctrine—that of “flexible response.” It was officially adopted by NATO, in 1967, at the urging of the United States and after much resistance by the Europeans. The Europeans had always preferred to put their trust in the American strategic nuclear umbrella to deter the Soviets from any kind of war, conventional as well as nuclear. Flexible response, or, as it was also called, graduated deterrence, implied that a war could be fought conventionally or with lesser nuclear weapons in Europe before the United States might be called on to use its strategic nuclear weapons, that is, the very weapons that risked American self-destruction. Nevertheless, flexible response remains official NATO doctrine to this day.

Flexible response has one irresistible attraction. It can be all things to all countries and all people. It does not, in principle, promise to use nuclear weapons; it does not promise not to use them. It is an accordion-like policy; you can stretch it out or pull it in as much as you like.

The reality about flexible response is something else. It has one fatal military defect: it does not tell what kind of war to prepare for. It tells military planners to prepare for any kind of war, which is exactly the same as telling them that they cannot prepare for any particular war. Choices always have to be made, and choices cannot be made if the response is so flexible that it must cover all possible eventualities.

In fact, a choice has been made. That choice has boiled down to one between a conventional war and a nuclear war. In 1982, General Rogers explained: “If flexible response is to be credible, it must be supported by an adequate military capability for each leg of the NATO triad of forces—strategic nuclear, theater nuclear and conventional.” In order to keep the nuclear threshold in Europe as high as possible, he advocated an adequate conventional deterrent, which would require an average annual real increase in defense spending of about 4 percent for the six years between 1983 and 1988. 21 The NATO members had previously agreed on an annual 3-percent increase, which had not been met. No one seems to expect a 4-percent increase, and even if it were met, there is much skepticism that it would really be enough, because the putative enemy need only increase his conventional forces to match.

More recently, General Rogers has told a rather more somber story: “The record shows that nations in the Alliance have never fully met their commitments to conventional force improvements. As a result NATO, while continuing to proclaim its faith in the declaratory policy of Flexible Response, has in fact mortgaged its defense to the nuclear response.”22 So we are back to the nuclear temptation in the name of Flexible Response.

Of all the nuclear temptations, however, the most seductive and most menacing is still to come; it is the “counterforce” doctrine. It is different in kind from the other three I have just mentioned—options, escalation, and flexible response. These three theoretically leave open the possible use of nuclear weapons and to that extent do not foreclose the issue. The counterforce doctrine requires the use of nuclear weapons, but in a certain way.

For those not accustomed to nuclear jargon, two terms should be briefly explained. They are countervalue and counterforce. Countervalue means nuclear attacks against cities or civilian populations and industries in highly populated areas. Counterforce refers to nuclear attacks against military targets, such as the enemy’s own nuclear weapons, military units, or facilities.

The counterforce doctrine first appeared, in the late 1950s and early 1960s, for two main reasons. It was argued that attacks against cities were not militarily productive; they killed the wrong people, namely civilians. And attacks against enemy cities were bound to bring similar attacks against our own cities, with the result that no American president was likely to adopt a countervalue, or city-oriented, strategy.

Counterforce presents a very strong temptation to use nuclear weapons, because it promises to take much of the horror out of nuclear war. No one can contemplate with equanimity a devastating attack on the entire social fabric of any country, let alone our own, but an attack on its nuclear weapons or even military establishment does not arouse quite the same repulsion. In one way or another, all present efforts to make nuclear war more feasible go back to the counterforce doctrine.

Here again, the doctrine needed a technological foundation. It was necessary to make two main assumptions. One was that military targets were physically or geographically separate and distinct from civilian targets. The other was that it was possible to develop nuclear weapons capable of distinguishing between the two. Again, as in the case of tactical weapons, the favorite terms are “precise” and “discriminating,” but now they seem to be applied to all kinds of nuclear weapons, even those of greatest range and most destructive power.

The counterforce strategy was actually adopted by Secretary McNamara in 1962 during the Kennedy administration. But he then quickly backed away from it and became a convert to a form of the classical deterrence doctrine, which he called “mutual assured destruction,” meaning that if both sides were assured of destruction, they would stay away from it as serving no conceivable political purpose.23 In effect, the goal again became how not to fight a nuclear war rather than how to fight one.

The reasons for this shift are just as valid today as they were then. In the first place, for counterforce to work, both sides have to adopt the same policy. No one, however, expects the United States and the Soviet Union to agree on how to fight a nuclear war or to guarantee their adherence to the same strategy in advance.

It was also realized that a counterforce policy was likely to cause such great civilian casualties that the line between counterforce—military—and countervalue—civilian—was purely theoretical and largely illusory. A Department of Defense study in the 1960s estimated that between 30 and 150 million Americans and a comparable number of Russians were likely to die in a nuclear war, even if efforts were made to stay away from highly populated areas. 24 In 1981, a group of UN experts found that a minimum of five to six million immediate civilian casualties and 400,000 military casualties would result if 1,500 nuclear artillery shells and 200 nuclear bombs were used by both sides against each other’s military targets.25 In effect, counterforce targeting was no panacea for what ailed countervalue.


Another reason for the shift away from counterforce strategy was somewhat more complicated and takes us some way into the darker recesses of nuclear war theory. It was the threat of a first strike. This threat has always hovered over nuclear war strategy, but it became particularly acute in the case of counterforce planning, which implies in the first place an attack against the enemy’s nuclear forces.

The trauma of the first strike comes about in the following way: Basically, there are two ways of conceiving a possible nuclear war. One is that it will resemble a conventional, or non-nuclear, war—only more so. There will be one or more fronts; a development of hostilities with some degree of gradualness; a mixture of weaponry; in general, a protracted, more or less controlled escalation. The other conception is peculiar to nuclear war. It can be conceived as an almost immediately catastrophic exchange, with millions of casualties suffered in one, two, or three days.

Neither of these alternatives is particularly appealing—to put it mildly. The protracted nuclear war would be, at best, a protracted agony. As for the nuclear cataclysm, nothing more need be said about it. So the problem presents itself: how to get around both of these unpleasant alternatives?

The logic of the situation points to a way out—to knock out the enemy’s nuclear weapons before they can be fired. If they—or most of them—could be knocked out at the very outset, the enemy would be prevented from waging a cataclysmic or protracted nuclear war. In short, a first strike is, logically, the most effective way to wage a nuclear war. That is what makes it so tempting and dangerous.

But to be successful a first strike must benefit from two preconditions: it must be thorough, and it must come as a surprise. If it is not thorough, it invites retaliation, which would begin a cycle of mutual devastation. If it does not come as a surprise, it would invite a preemptive first strike by the other side.

Of course, no nuclear nation will admit that it has ever contemplated or is even capable of contemplating a surprise attack and an unprovoked first strike. What cannot be denied is that they are inherent in the logic of the nuclear dilemma. That is why both sides fear them so much and charge the other with preparing for them.

But logic does not exhaust reality. The full reality is that a surprise attack and a first strike would be an infinitely risky business. They would have to be totally successful or they would open the aggressor to devastating retaliation and retribution. Without any experience of nuclear warfare, no one knows, and no one can know, what a surprise attack would achieve. It would have to be a go-for-broke operation. In the abstract, the first strike would seem like an attractive proposition. In the real world, it is an almost senseless gamble.

In any case, for these reasons and others, McNamara gave up the counterforce temptation after 1962. But now we are getting it again, and in a worse form than ever before.


The present phase began in 1974, during the Nixon administration. It was sponsored by then Secretary of Defense James Schlesinger. His new policy was basically no more than a variation on an already old theme—that of “options.” The argument, still in vogue today, maintained that the president should not be limited to choosing between no nuclear war and all-out nuclear war. He should instead be able to engage in all forms and degrees of conventional and nuclear war. Schlesinger’s National Security Study Memorandum of 1974 brought forward the option of threatening Soviet military targets.

After Schlesinger came Harold Brown, secretary of defense in the Carter administration. In 1980, President Carter issued Presidential Directive 59, which played more variations on the theme of “options.” This directive has never been made public, so we are dependent on what Mr. Brown and other insiders have said about it.

According to Mr. Brown: “There is a good chance that any US-Soviet nuclear exchange would escalate out of control.” Nevertheless, the United States must prepare for just such a nuclear exchange, that is, a limited nuclear exchange that would probably escalate out of control. Why? Because the United States must have a “victory-denying” response—“victory-denying” is typical of the fudging language customary in this field—to a Soviet effort to obtain victory in a limited nuclear war.26

Notice: the whole idea is predicated on the assumption that the Soviets may seek some sort of nuclear superiority to obtain victory in a limited nuclear war—the same sort of war that is unlikely to stop short of an all-out exchange. After all this, Mr. Brown also tells us that “superiority is an idle goal.”27 Yet without superiority, victory could not be obtained in a limited nuclear war and it would almost certainly escalate out of control. If Presidential Directive 59 follows Mr. Brown’s exposition, it is a mishmash of contradictory premises and prescriptions.

After Brown came Caspar W. Weinberger, the present secretary of defense in the Reagan administration. Mr. Weinberger is another wholesale options merchant. He has offered the president one of the most treacherous options of all, though the idea may not have originated with him or his advisers.28 This option is the conduct of a protracted nuclear war in which the United States “must prevail” or out of which it must “emerge” with “terms favorable” to us. This policy was enshrined in a document entitled “Fiscal Year 1984–1988 Defense Guidance,” issued in the spring of 1982, which had to be leaked in order for ordinary citizens to know about it.

Ironically, Mr. Weinberger has indirectly criticized his own policy. The idea of a protracted or prolonged nuclear war is so indefensible that he tried to repudiate it in a letter sent to a number of US and foreign publications in August 1982. He also tried to repudiate the concept of a nuclear victory in a letter sent to me in July 1983. What seems to have happened in this: The policy of waging a protracted nuclear war and of prevailing in such a war has been adopted officially but disavowed publicly. The least that can be said of this two-tracked or two-faced policy is that it is a strange way of conducting serious business in a democracy.

Here is Mr. Weinberger on both sides of these issues:

For protracted nuclear war: US forces must be able to maintain “through a protracted conflict period and afterward, the capability to inflict very high levels of damage” on Soviet industry. Should a Soviet attack “nevertheless occur, United States nuclear capabilities must prevail under the condition of a prolonged war.”29

Against protracted nuclear war: “I am increasingly concerned with news accounts that portray this Administration as planning to wage protracted nuclear war, or seeking to acquire a nuclear ‘war-fighting’ capability. This is completely inaccurate….”30

For winning: “…United States nuclear capabilities must prevail…. earliest termination of hostilities on terms favorable to the United States”.31 “…to achieve political objectives and secure early war termination on terms favorable to the United States and its allies.”32 “You show me a Secretary of Defense who’s planning not to prevail and I’ll show you a Secretary of Defense who ought to be impeached.”33

Against winning: “…we do not believe there could be any winners in a nuclear war”;34 “…our belief that there could be no winners in a nuclear war.”35

Finally, a still greater temptation has recently been put forward by influential nuclear war theorists. Like most temptations, nuclear or otherwise, there is nothing new or original about it; it merely pushes the temptation further than anyone has dared to do in the past. These tempters advocate the development of nuclear weapons that could attack targets so “precisely and discriminately” that they could safely be used against the enemy’s weapons “without mass destruction.”36 Their nuclear war would be something like a ping-pong game in which each side would “precisely and discriminately” drop its nuclear warheads on the other side’s weapons. Since this scheme holds out the prospect of avoiding mass destruction, it is more tempting than the prospect of repeating the heavy civilian casualties and widespread destruction of the two conventional world wars in this century.

As I have tried to show, we have been through all this before. Targeting the enemy’s weapons or military facilities may reduce civilian casualties at the outset, but these will still be so high—somewhere in the millions—that it is irresponsible and heartless to play around with the likelihood of avoiding mass destruction. The Soviets, at least, have clustered many of their nuclear weapons and installations in proximity to their cities, especially Moscow. There would be no way of adequately testing our precise and discriminating weapons, even if—someday—we should have them. There is no reason to believe that both sides would agree to use precise and discriminating weapons only, especially if one side should be put at a disadvantage in the development of such weapons. There is no reason to believe that either side would trust the other, even if they both agreed to use such weapons only. If those precise and discriminating weapons did not knock out all or most of the other side’s nuclear weapons at once, retaliation could only take the form of a more indiscriminate counterattack. For one thing, the same type of weapon would no longer be available to both sides; for another, one side’s precise and discriminating weapons would already have been shot off, thus no longer offering a useful target to the other.

Thus the technological cure is a form of the disease. It is actually a prescription for a potential first strike, the most dangerous of all nuclear temptations. It is interesting to note that an analysis was made in 1968 to determine the relative number of casualties in the event of a Soviet or an American first strike. The paradoxical result was a finding that there would be more American casualties in the case of an American first strike than in the case of a Soviet first strike. The paradox arose because it was figured that the side striking first would go after military targets, whereas the retaliating side would mainly hit cities.37 All of which suggests that this is not a subject for weak nerves or soft heads.

The ultimate temptation, in my view, is the type of indoctrination we are now getting. It is indoctrination in the feasibility of some kind of nuclear war or at least the use of some kind of nuclear weapons in some kind of war. Feasibility is almost always expressed in terms of the counterforce strategy and the technology necessary to carry it out. As you may suspect, I consider them to be a snare and a delusion. I see no way of getting rid of nuclear weapons; and I see no political purpose served by using them. All the theorizing about limited, controlled nuclear war is just that—theorizing. We may be fortunate that the decisions will be made not by theorists but by practical politicians, who have to ask themselves what conceivable political purpose would be served by running the risk of a devastation that will know no politics.

But there is politics and politics. There are the political rules by which nations decide what weapons they need and political games which they play in negotiations with other nations. In the first case, the governing question is—or should be—“What is enough?” In the second case, it almost invariably becomes “Who’s ahead?” Negotiations between nuclear nations, even in the name of reductions, seem to bring out the worst in them. Since they are overstuffed with nuclear weapons, they can afford to negotiate in the realm of redundancy, so that it makes little practical difference what partial reductions they decide on, if any. The real target is not so much the other side’s nuclear weapons as its political will and unity. It makes little military difference whether the SS-20s are stationed on the western borders of the Soviet Union, from which they can reach all or most of Western Europe, or whether, as the Soviets have recently threatened, they are moved to East Germany and Czechoslovakia. The threat is made to be psychologically intimidating, because it cannot be anything else.

The redundancy factor is one of the most tantalizing. It is difficult to establish just where it begins. Yet it is not difficult to recognize that we have long since passed that point, wherever it is. Former Secretary of Defense Brown tells us: “A single 50-kiloton to 1-megaton nuclear explosion over a city will kill several hundred thousand people. The United States and the Soviet Union each have more than 6,000 such warheads, and the vehicles to deliver them, in their stockpiles. Thus each has much more than enough capability to destroy all the large and medium-sized cities in the other’s country.”38 At the General Assembly of the United Nations in September 1983, President Mitterrand said that the US and USSR both have “a nuclear system of 2,000 to 3,000 launchers, carrying 8,000 to 9,000 nuclear warheads.” They are enough to “reach and destroy each other seven or eight times over.” The French alone are said to have the ability to wipe out thirty Soviet cities.39 Yet they too are adding to their nuclear armory.

Marshal Nikolai V. Ogarkov, chief of the Soviet General Staff, has asserted that both the United States and the Soviet Union have the ability to fire at a single launching on the order of 12,000 warheads with a total yield of 3,400 megatons from strategic nuclear weapons, or 170,000 times the yield of the first Hiroshima bomb. Yet it is generally agreed that one thousand warheads would be more than enough for deterrent purposes.40 Presumably Brown, Mitterrand, and Ogarkov have access to the most trustworthy information, denied to ordinary mortals.

Another and more fearsome measure of redundancy has been offered by Professor Carl Sagan on behalf of a distinguished group of scientists. It is usually accepted, according to this study, that both American and Soviet arsenals contain about 18,000 strategic thermonuclear warheads, with an aggregate yield of 10,000 megatons, in addition to about 35,000 tactical and intermediate weapons. Yet, no more than 500 to 2,000 strategic warheads would be enough to trigger a climatic catastrophe threatening the survival of the human race. “In summary,” Professor Sagan writes, “cold, dark, radioactivity, pyrotoxins and ultraviolet light following a nuclear war—including some scenarios involving only a small fraction of the world strategic arsenals—would imperil every survivor on the planet.” Thus the redundancy factor here amounts to a minimum of 16,000 strategic thermonuclear warheads. In addition, a first strike is estimated to require from 2,200 to 4,500 attacking warheads to be effective. This number would clearly approach and possibly overreach the threshold of a climatic catastrophe. The study concludes, “a major first strike may be an act of national suicide, even if no retaliation occurs.” 41

The enormous redundancy of nuclear weapons on both sides has a direct bearing on the efficacy of a nuclear “freeze.” I am not opposed to it; I am simply not enthusiastic about it. The present stocks of nuclear arms would be frozen at such high levels that it would (a) make little difference in the ability of both sides to destroy each other, and (b) operate mainly to keep down the sphere of redundancy. A freeze cannot at best take the place of a mutually deterrent nuclear balance, whatever the level of the freeze might be. The great utility of a freeze or a comprehensive test ban is not so much in itself but in what it might lead to; it might conceivably be a first step in an agreement to stabilize the nuclear confrontation and eventually to reduce the level of forces. There are all sorts of gimmicks and placebos being offered in the nuclear marketplace, but on close examination they all depend ultimately on deterrence to give them a semblance of plausibility.42

The lesson I draw from the entire course of Soviet-American nuclear negotiations is that they have done more harm than good; they increase tensions and mistrust; at best they end up by letting both sides keep what they really want to keep and give up what they really do not need. The greatest nuclear reductions would come about if each side simply decided by and for itself to maintain whatever stable force level was necessary to deter the other side from taking incalculable risks. Neither side will ever negotiate itself below this level anyway, so negotiations ultimately turn on gaining political rather than military advantage.

But “what if deterrence fails?” This question is often asked, with an air of triumph, as if the possible failure of deterrence were a reason for rejecting it. Such an attitude is comparable to that of rejecting a life-support system in a hospital because it may fail or be inadequate to keep a mortally ill patient alive. Yet the possibility of failure does confront us with the fearful problem of what to do if some sort of nuclear war should break out. Toward the end of his life, Bernard Brodie gave the answer that the main goal should be “to terminate it as quickly as possible and with the least amount of damage possible—on both sides.”43 That was the attitude of one who had thought that almost anything was better for mankind than total nuclear war.44 I have also been driven ineluctably to this conclusion, without, however, pretending to know how it will be possible to terminate a nuclear war with the least possible damage. Yet Brodie’s view is infinitely preferable to one which calls for a “Nuclear Strategy: The Case for a Theory of Victory”45 or of another which claims that the Soviet Union has found a theory of victory by way of nuclear “superiority.”46

To my mind, the obvious answer to the question “What if deterrence fails?” is that we do not know what will happen. We have no experience with the failure of nuclear deterrence, and without experience we have little or nothing to go by. Wars have been notoriously unpredictable, and nuclear wars must surely be the most unpredictable of all. We do not know how, where, by whom, or to what extent deterrence would fail. It would seem to be the most ordinary prudence and elementary common sense to make sure that we are not responsible for its failure, that we do whatever we must to limit the damage to ourselves and our allies, and to induce the other side to terminate the conflict as quickly as possible in its own interest. But all this is so far in the realm of the contingent and unpredictable that no one can be sure what such a war would be like or how the antagonists and the world at large could even survive it in recognizable condition. This very uncertainty is an element of deterrence. If it is any comfort, we know more about how to deter a nuclear war, judging from almost half a century of some sort of deterrence, than we know how to fight one.

I am, therefore, a believer in the lesser evil of nuclear deterrence. I believe in it because it is by far the lesser evil, not because it is good. The main enemy at present is not a nuclear balance that results in mutual deterrence; it is the propaganda about the feasibility of nuclear war by way of precise and discriminating weapons allegedly capable of avoiding mass destruction. The main reason nuclear weapons have not been used thus far is precisely the belief that they cannot be launched for any useful political purpose and that mutual mass destruction can be of no conceivable benefit to either side. But now the Pied Pipers of a protracted nuclear war and of precise and discriminating nuclear weapons are trying to lure us to break through the psychological and political barriers to nuclear war.

Lord Henry Wotton told Dorian Gray that the only way to get rid of a temptation is to yield to it. Clearly that would not do in this case. The only way for us to get rid of this temptation is to know it for what it is and to reject it precisely and discriminatingly.

“To use or not to use”—that is the “to be or not to be” of our time and for as long as we can now foresee.

A Postscript

In my exchange with Secretary of Defense Caspar Weinberger, in The New York Review of August 18, 1983, I devoted some lines to an article by Albert Wohlstetter in Commentary of June 1983:

This vision of a controlled nuclear war, capable of hitting only military targets “precisely and discriminately,” was recently proclaimed as the new nuclear gospel by Albert Wohlstetter, no stranger to the Pentagon, in the June 1983 issue of Commentary, which has become an organ of the nuclear warriors. Wohlstetter in effect conceives of nuclear war as if it could be a kind of ping-pong game in which nuclear warheads would drop “precisely and discriminately” on each side’s military installations “without mass destruction.” This is the most perverse and dangerous nuclear temptation that has been dangled before us in a long time. It is the apotheosis of the limited, controlled nuclear war that is even less likely to hurt people than the bad, old-fashioned artillery and aerial bombardments which were responsible for vast mass destruction.

I also commented on Wohlstetter’s pose of moral superiority:

Apparently it is moral to bring nuclear war closer by making it more like any other kind of war and even more precise and discriminating. But it is immoral to remove nuclear war from rational calculation for the reason that it is likely to be so mutually devastating that it could serve no useful political end. It is insidious to urge that one type of nuclear war is immoral, but that another, more seductive kind—neatly, cleanly “military”—is by implication moral or at least more morally acceptable. To pretend that moral distinctions can be made between allegedly different types of nuclear wars is already taking a most slippery and menacing step toward breaking the nuclear barrier. If we should bite off our tongues before uttering one word in this discussion, that word is “morality.”

Mr. Wohlstetter has chosen to answer me in the letter columns of the December 1983 issue of Commentary. Wohlstetter also takes issue with others who wrote in about his article, “Bishops, Statesmen, and Other Strategists on the Bombing of Innocents,” but I need only comment on the matters that directly concern me. The points in question are fundamental to those raised by the “nuclear temptations,” and, therefore, a confrontation with his views may contribute to a better understanding of what is at stake.

Wohlstetter says that “it’s hard to resist taking just a few well-aimed smacks at such fierce but hollow defenders of this establishment faith [mutual assured destruction] as Theodore Draper, who nowadays appears regularly in The New York Review of Books.” It’s not clear to me why he found those smacks hard to resist, inasmuch as he returns to my alleged transgressions again and again, or what sinister significance there is in appearing regularly in The New York Review. Still, let us see how well-aimed his “smacks” are.

Curiously, one of his best-aimed smacks is by Wohlstetter against Wohlstetter. The Wohlstetter who wrote the reply took just such a smack at the Wohlstetter who wrote the article. In the article of June 1983, the first Wohlstetter came out for nuclear weapons with “the capability to attack targets precisely and discriminately,” and he deplored opposition to “research and engineering on ways to destroy military targets without mass destruction” (italics added). In the reply of December 1983, the second Wohlstetter completely dropped the term “precisely” and fell back on “discriminate and proportionate.” The latter are obviously so vague and general—how discriminate? Proportionate to what?—that they may mean anything or nothing. “Precisely” was evidently too precise to be repeated.

There is also no longer any suggestion of getting nuclear weapons “without mass destruction.” Instead, Wohlstetter now tells us that he has “never been in the slightest bit unclear about the slaughter involved in war, even a nonnuclear war” and that “war would be terrible, even if we were able to confine the destruction effectively to military targets.” Once again Wohlstetter, has silently corrected Wohlstetter, not me.

As for that metaphor about a nuclear ping-pong game, Wohlstetter writes: “Theodore Draper, at his surliest, suggests that I think a nuclear war would be like ping-pong” and “cannot conceive of any use of nuclear weapons in between a neat well-mannered game of ping-pong and total mutual annihilation.” Once more it escapes me why I had to be “at my surliest” to suggest that Wohlstetter’s original version of his preferred nuclear war would resemble “a kind of ping-pong game.” I am sure that I would have written the same thing even if I had been at my sunniest.

It should be recalled that what I wrote was based on Wohlstetter’s original article. If the nuclear war is going to be fought with weapons that drop precisely on other nuclear weapons, and could be developed to do so without mass destruction, the effect would be very much like a nuclear ping-pong game. If one read only Wohlstetter’s reply, one would never know that he had originally espoused nuclear weapons so precise that he could envisage dropping them without causing mass destruction.

Wohlstetter knows better than to pretend that I cannot conceive of a nuclear exchange that is somewhere between a neat well-mannered ping-pong game and total mutual annihilation. But that is not the point. What was originally in question was what Wohlstetter, not I, could conceive. He initially conceived of a neat, if not particularly well-mannered, nuclear ping-pong game, a war of dropping nuclear weapons precisely on nuclear weapons or their installations. It is not particularly hard to conceive, at least in the abstract, of a nuclear war coming somewhere between those extremes, but I do not think that my conception—or his—is all that important. We may conceive of almost any kind of nuclear war, but that is all it would be—a conception. The reality is another matter.

This brings me to the main point. Basically, Wohlstetter has resurrected the counterforce doctrine of fighting a limited, controlled nuclear war in which both sides concentrate on “enemy military power.” He contrasts favorably small nuclear weapons with conventional weapons, on the grounds that the former would do less direct damage and would be better able “to keep the chain of violence under political control.” If both sides limited themselves to such nuclear weapons, he reasons, they could do most damage to each other’s military targets and minimum damage to “non-combatants.” Wohlstetter also contends that anyone who threatens to do much more than that, especially if the policy were based on “mutual assured destruction” or MAD, is bluffing or has no intention of carrying out the threat.

One oddity is that he feels it necessary to launch a lengthy attack on MAD, as if it were the “establishment faith,” and as if I represented the “establishment.” He himself acknowledges that MAD “has never been operational policy in the United States” and “could not be.” The declared American policy since the early 1970s has been based on “flexible response,” which has been increasingly interpreted in the past decade in the light of the counterforce doctrine, such as Wohlstetter advocates. Whose “establishment” is he fighting? He is really defending the current official orthodoxy, but one would never know it from his article or reply.

Wohlstetter’s method is entirely conceptual. All that actually comes out of it is the proposition that a limited, controlled nuclear war would be better than a total, uncontrolled nuclear war. He never touches the real question of how feasible or likely a limited, controlled nuclear war may be. It never seems to occur to him that one should be cautious about celebrating the comparative virtues of a limited, controlled nuclear war if it is not likely to remain limited and controlled or even if the risk of its becoming unlimited and uncontrolled is too great to take. Almost no one who has written on the subject believes that a nuclear war can be safely controlled. The most recent skeptic is an old soldier with impeccable military credentials. Lieutenant General Arthur S. Collins, Jr. (US Army, ret.) has contributed an extraordinarily realistic analysis of just the kind of war that Wohlstetter conceives of. Here are some of his reflections:

The scenarios for how a tactical nuclear war will be kept limited are just not credible—at least not to me. Most of those who postulate scenarios for tactical nuclear war have a faculty of describing situations in which understanding, cooperation, and compliance by the enemy make those scenarios credible. There is an unwarranted assumption that compliance will be forthcoming….

But nothing in human history compares to the tactical nuclear battlefield where several hundred—or several thousand—nuclear weapons will be used by both sides. The physical devastation and psychological trauma of the nuclear battlefield will be without parallel….

Many war games and studies have shown that where the first use of nuclear weapons is careful and discrete, the side initiating the nuclear attack would be overwhelmed by a sudden and massive enemy nuclear response….

The lack of an acceptable doctrine for tactical nuclear war contributes to the distortion, but it is not due to neglect. There have been studies and wargames by the hundreds, but still no credible doctrine.47

General Collins’s judgments—there are many more, equally pungent and instructive—particularly apply to Wohlstetter’s type of nuclear scholasticism.

The nub of Wohlstetter’s argument is that the Soviets will follow our example if we assure them that we will use nuclear arms only against their weaponry, not against civilian concentrations. That is what he calls a “non-suicidal response.” An attack against their civilians, who “in any case, do us no harm,” would be suicidal, because the Soviets would return fire in kind. Thus the success of Wohlstetter’s policy hinges on a mutual agreement to conduct a limited, controlled nuclear war only—one limited to military targets and politically controlled from beginning to end.

All depends, then, on these two conditions. How likely are they to be fulfilled? Soviet doctrine has always called for a massive, overwhelming nuclear response. In March 1983, Marshal Nikolai V. Ogarkov, chief of the Soviet General Staff, gave an interview to Leslie H. Gelb of The New York Times (March 17, 1983). Ogarkov said that “to keep such a [nuclear] war limited will not be possible” logically, and that “inevitably, such a war will extend to allout war.” Most American and almost all European authorities agree with him or consider that no reliance can be put on a barrier between limited nuclear war and all-out nuclear war.

In any case, it matters little what anyone says about intending to wage a strictly limited nuclear war. The dynamic character of war, past and future, makes such self-denial a peacetime luxury or an academic exercise. Wohlstetter has a curious way of venting his spleen on those who, allegedly like myself, “threaten mutual annihilation in order to deter.” This is a strange aberration. I believe that any nuclear war will bring with it the risk of mutual annihilation—hardly an original thought—and, therefore, that the risk itself acts as a deterrent to engaging in such a war. The risk of mutual annihilation inheres in the nature of this beast. It is there whether anyone who is misguided enough to disagree with Wohlstetter threatens it or not. Even if mutual annihilation is too strong a term, the “terrible” destruction to be expected even in a counterforce nuclear exchange, as Wohlstetter admits is the case, would be great enough to amount to a near-suicide pact. No matter how we conceive a nuclear war might begin, we cannot be sure—we can hardly conceive—how it will end.

The real issue is whether the temptation of a nuclear war limited to the assured mutual destruction of nuclear weapons or military targets only makes a nuclear war appear to be more feasible and, therefore, more likely. It is not, as Wohlstetter would like to have it, whether anyone threatens the assured mutual destruction of entire societies. If, as Wohlstetter believes, a policy of MAD cannot be taken seriously anyway, and it has not been American policy for almost two decades, it is hard to understand why Wohlstetter should have worked himself into such an unseemly fit of ill-aimed “smacks.” If MAD should ever come to pass, it will most likely come about as the final stage of a nuclear war that is supposed to be limited and controlled à la Wohlstetter.

There is no necessary moral choice between the admittedly terrible slaughter of a limited, controlled nuclear war and the possible mutual annihilation of an all-out nuclear war, and even less if the terrible slaughter may well end up in mutual annihilation. If morality enters at all, it is in how to avoid such a choice, not in how to make it.

This Issue

January 19, 1984