The War That Must Not Occur
Jean-Pierre Dupuy, Translated by Malcolm DeBevoise



This book is about nuclear warfare. As a philosopher, I have pondered the implications of this nightmarish scenario for twenty years, without claiming to be either an expert or a specialist. More generally, I have thought about the history and philosophy of violence since encountering the work of René Girard in the late 1970s. This inevitably led me to consider catastrophes and the problem of evil, particularly in relation to the threats to the future of humanity posed by climate change, the risk that advanced technologies—na-nobiotechnologies, synthetic biology, human genome editing, and the like—may escape the control of their inventors, and, not least, nuclear war. Nuclear war has furnished me with a template for a form of rational doomsaying that describes a relationship to the future that I call projected time. This conception of time implies that once what is at risk is monumental, beyond all human measure, it is legitimate to hold that catastrophic events, once they become possible, are bound to occur. Here possibility implies necessity. The difficult thing to understand is that this necessity in no way amounts to fatalism. It may be within our power, as I show in this book, to postpone catastrophe ad vitam aeternam. Nevertheless, we must regard catastrophe as necessary; otherwise our wanting to delay its occurrence would not be sufficiently motivated. If catastrophe is merely possible, then its non-occurrence is equally possible. It is not a contradiction, in other words, to believe in both the necessity of the future and its indeterminacy.

The problem of nuclear warfare leads us to reformulate some of the most important and most difficult questions of metaphysics.1 A reliance on abstraction and a priori reasoning becomes unavoidable in the light of a very simple and quite stunning fact. Thomas Schelling, whose writings on the mathematical theory of games have had a great influence on nuclear doctrines, memorably referred to it at the beginning of his lecture in Stockholm in December 2005 on being awarded the Nobel Prize in Economic Sciences: “The most spectacular event of the past half century is one that did not occur. We have enjoyed sixty years without nuclear weapons being exploded in anger.” More than fifteen years later, this observation still holds. Schelling’s opening words are often recalled, but what he said next has for the most part been forgotten: “In 1960 the British novelist C. P. Snow said on the first page of the New York Times that unless the nuclear powers drastically reduced their nuclear armaments thermonuclear war within the decade was a ‘mathematical certainty.’2 Nobody seemed to think Snow’s statement extravagant.”3 The coupling here of necessity and indeterminacy is a paradox well worth reflecting upon.

The situations and events that I review here and in the first two chapters of this book were selected mainly for the purpose of illustrating the concepts analyzed in the final two chapters. Most of what has been written and said about the Ukrainian crisis, for instance, comes almost exclusively under the head of geopolitics. This dimension is essential, but it is by no means the only one that needs to be taken into account. Nuclear war has its own syntax, which is superimposed on the intentions and decisions of world leaders. Putin’s psychology no doubt plays a role in the present instance, just as Trump’s did during the North Korean crisis. Nor can the importance of Ukraine in Russian history and culture be neglected, any more than the role played by the United States in the military command structure of NATO can be. But when a potentially nuclear confrontation begins to escalate, these so-called actors look more like marionettes, driven this way and that by forces beyond their control, even if these are forces of their own making. They still believe that they are in control, of course, that violence obeys their will, but the truth is otherwise: violence manipulates them according to its own laws.

It is this aspect of the matter I am concerned with in the present book. The Ukrainian crisis is only a particular case study. Nevertheless, I hasten to take advantage of the opportunity to introduce the book to an American audience in order to sketch the outlines of an analysis of the events taking place today in accordance with the method I develop in the pages that follow.

Most people, except French experts and others who claim that the current crisis cannot possibly lead to a nuclear conflict amounting to a third world war, anxiously wonder what the chances4 are of this scenario coming to pass. Will Putin drop an atomic bomb on a Ukrainian city in order to make Zelensky surrender? Scott Sagan, a prominent authority, thinks it is by no means an implausible outcome. The United States did just this, he reminds us, in order to force Japan’s surrender in 1945.5 Will the Russian president go so far as to target a European capital in order to punish NATO for providing Ukraine with increasingly powerful and sophisticated weapons? Given Russia’s quantitative superiority, is he prepared to launch intercontinental ballistic missiles against the only other great nuclear power, the United States?

At this juncture we need to consider a prior question, the one that Schelling raised in his Nobel lecture: how are we to explain the fact that since 9 August 1945, the date of Nagasaki’s decimation, no atomic bomb has been dropped on civilian populations with the aim of exterminating them? If we knew the answer to this question, we might be in a position to estimate the chance that this mysterious blessing will continue to favor us—as though there were a good fairy who watches over humanity and prevents it from destroying itself.

Why should the unprecedented power of the atomic bomb not be a sufficient reason to dissuade anyone from even thinking of using it? Isn’t the principle of deterrence entailed, as a matter of practical reason, by its immeasurable destructiveness? Who could possibly have an interest in escalating a conflict to the point that there are no winners, only losers? These questions have been with us since 1945, and they remain no less perplexing today. Attempts were made during the Cold War to reduce both the power of atomic weapons and the range of the missiles that carry them in the hope of bringing the devastation produced by a nuclear conflict nearer to that which a traditional war is capable of producing. Eventually it became clear, however, that these so-called tactical weapons and missiles must be banned. Their relatively small explosive force6 encourages military planners to use them on the battlefield, as in the case of conventional weapons, with the risk of getting caught up in a nuclear spiral whose inevitable tendency, as can be shown by a priori reasoning, is to spin out of control and lead to mutual annihilation. Just as the explosion of an atomic bomb triggers the thermonuclear reaction within a hydrogen bomb, so too the use of tactical nuclear weapons on the battlefield is the surest way for intercontinental ballistic missiles to be brought out from their silos, notwithstanding that these missiles are supposed, by virtue of their passive existence alone, to assure nuclear peace.

This explanation in terms of interests—the interests of each person and the interest of all—nonetheless runs up against the sobering lessons of our own time. The tragedy of human history is that very often it destroys the very people who make it, even though each one of them tries only to satisfy his or her own interests. Tactical nuclear weapons were in fact banned for a time, at least partially, as we will see. Today, however, they are more numerous than ever.

An entirely different explanation for the absence of nuclear war is that we narrowly escaped it; chance and purely chance—that is to say, luck—has spared us the worst. Historians of the nuclear era record many incidents that could have set off a fatal escalation but did not: miscommunication among key figures, errors of interpretation, rash calculations of risk, fits of rage, and so on. In each case horror only just missed becoming reality. I examine several such episodes. The relative weakness of the explanation from luck has to do with the fact that we do not know whether chance was responsible for the incident that threatened disaster or whether it was because of chance that disaster was averted. For want of a common preternatural cause for this series of near catastrophes, it is reasonable to suppose that one day luck will grow tired of the coin always coming up heads and that the moment will inevitably come when it will come up tails—not least because this should have happened long ago.

The simplest, most obvious, and most common explanation is that it is because deterrence has succeeded that nuclear war was able to be prevented. On this view, the possession of an atomic arsenal has only one aim: to dissuade other nuclear powers from attacking first, by threatening them with disproportionate retaliation if they pay no heed and, if need be, by extending this threat to a nonnuclear attack that would imperil the nation’s vital interests. A good part of the present book is devoted to discussing this claim. Once again, what makes it a genuine philosophical puzzle is the lack of empirical evidence and the corresponding need to resort to a priori reasoning.

Many philosophers and strategists have concluded that deterrence can work only if each of the leaders of two rival nuclear powers has reason to believe that the other is irrational.7 The chief obstacle to nuclear deterrence is that the threat of retaliation on which it depends is not credible. If deterrence fails, will the nation that has been attacked really carry out its threat and unleash a suicidal escalation? Does one have to be mad, or pretend to be mad, in order to be credible? The soundness of deterrence turns on the answer to this question.

Whatever the answer may be, a major reason to doubt that deterrence is mainly responsible for the absence of nuclear war for almost eighty years now is that it has seldom actually been practiced. In its pure form, deterrence requires giving up the very thing that gives armed forces their legitimacy: the power to defend. By making it clear to an adversary that you will do nothing to stop any missiles that it may launch against you—for example, by means of an antiballistic missile shield—your adversary can be assured that you will not attack first. For if you were to launch a first strike, that would not prevent your adversary from carrying out the threat of a ruinous reprisal by virtue of its capacity to launch a second strike. In that case no one attacks first, and what is called a balance of terror (or, a remarkable oxymoron, nuclear peace) is in principle realized. But the abandonment by the armed forces of their primary mission, to defend the nation from attack, is not a price they are prepared to pay. I analyze several striking cases where the principle of deterrence has been disregarded altogether.

What purpose have nuclear weapons served, then, if they are only remotely associated with the absence of nuclear war? Paradoxically, they have made the possibility of a first strike more likely. In making the first move (“preemptively,” in the jargon of nuclear strategists), one side responds to a potential attack as if it had already taken place. The answer comes before the question—thus the temporal inversion that forms the leitmotif of this book: retaliation in advance. Whatever their official nuclear doctrines may say to the contrary, a first strike has never been ruled out by either Soviet (later Russian) or American leaders. Nevertheless, convincing an adversary that one is prepared to strike first is no less problematic than convincing an adversary that one will forcefully respond to its attack. Here again the problem of credibility arises. A first strike will not be sufficient to neutralize an enemy who retains the capacity to retaliate. It is therefore necessary to make the enemy believe that you will be able to absorb a retaliatory strike while limiting the damage from it—which is to say that you will remain fully capable of retaliating against retaliation. That may be very difficult to do.

The United States and Russia have had, and continue to have, an ambivalent attitude toward an element of nuclear doctrine known, misleadingly, as “escalate to de-escalate.” Their vacillation in this regard illustrates a dilemma facing both of them in respect of deterrence and preemption that has a direct bearing on the Ukrainian crisis. The idea of escalating in order to de-escalate, introduced by Schelling in The Strategy of Conflict (1960), has influenced several generations of strategists. The doctrine of flexible response formulated by Robert McNamara shortly afterward and the concepts of limited nuclear war, escalation control, and the like are so many variations on the same theme. The simplest way to think about it is to compare it to the logic of an auction, where one keeps pushing the price up until the other bidders drop out. Similarly, one increases the intensity of combat with nonnuclear (conventional) forces until resort to a nuclear strike seems inevitable in order to end the conflict by forcing the enemy to yield. This is what is called de-escalation.

I advance a number of arguments showing not only the vacuousness of this idea but also the dangers that are bound to follow if it is put into effect. Carl von Clausewitz was perhaps the first to point out, in On War (1832), that as a theoretical matter there is no decisive move that will put a halt to an increasingly violent sequence of events. More often than not, Clausewitz observed, logistical and other obstacles—what he called the fog of war—combine to prevent mutual annihilation from coming to pass. In the case of nuclear war, by contrast, these same obstacles accelerate an escalation to extremes.

Both American and Russian nuclear strategists recite the credo of nuclear deterrence: a limited attack cannot be deterred by making a threat of limited retaliation credible; it can be deterred only by sustaining a moderate probability of mutual annihilation. In practice, however, the fact remains that the option of escalating to de-escalate continues to tempt military planners, particularly on the Russian side, so far as we can judge from unofficial public statements. According to Alexei Arbatov, a senior national security advisor, “Conventional precision weapons should be capable of inflicting sufficient losses on attacking NATO forces and bases to induce NATO either to stop its aggression, or to escalate it to the level of massive conventional warfare, including a ground offensive. This would then justify Russia’s first use of tactical nuclear weapons.”8

From the point of view of deterrence, this can be seen only as an admission of failure. The fact that we have escaped a third world war involving strategic nuclear weapons, even that no atomic weapon of limited power has yet been used under battlefield conditions, seems nothing short of a miracle. In order to provide a satisfactory explanation for this extraordinary state of affairs—what Schelling called an event that did not occur—it will be necessary to rely on a negative form of metaphysical argument, after the example of negative (or apophatic) theology. For the moment, however, we know enough to be able to venture an answer to the question that concerns us here, namely, whether it is possible that the Ukrainian crisis will lead to nuclear war and, if so, what is the likelihood of this happening.

On 1 February and 2 February 2019 two things happened that went largely unnoticed by world public opinion and that in large measure account for the present state of affairs. First Trump and then Putin the next day announced that their countries intended to withdraw from the treaty signed more than thirty years earlier, in 1987, by Ronald Reagan and Mikhail Gorbachev, who thereby agreed to eliminate from their respective nuclear arsenals all ground-launched cruise missiles and ballistic missiles with ranges of between 500 and 5,500 kilometers (310 and 3,420 miles). The Intermediate-Range Nuclear Forces (INF) Treaty was a misnomer, for the accord did not limit nuclear forces per se. It banned a certain class of missiles, whether they were equipped with nuclear warheads or not. The American withdrawal took effect on 2 August 2019.

It will be instructive to recall the historical background to this joint decision. For more than ten years, between 1976 and 1987, the Euromissile crisis aroused enormous fear and controversy in European nations. In March 1976, the Soviet Union deployed in its western region SS20 missiles having a range of about 5,000 kilometers, capable therefore of reaching not only Western Europe but also China and Japan. The American president, Jimmy Carter, felt confident, in accordance with the accepted principles of deterrence, that America’s long-range strategic nuclear weapons sufficed to discourage the Soviet Union from launching a surprise attack on Europe. But the West German chancellor, Helmut Schmidt, urged the United States to intervene directly. In December 1979, at a summit of NATO leaders, it was decided to enter into negotiations aimed at forcing the Soviet Union to withdraw its SS20s and, if these talks failed, to install Pershing II medium-range missiles in West Germany within four years.

There followed a confused period that survives in popular memory today mainly in the form of slogans. The German pacifists, supported by the French Communist Party, declared that they would rather be “red than dead,” to which François Mitterrand replied that “the pacifists are in the West and the missiles are in the East.” As it turns out, the Pershing IIs were deployed in West Germany on schedule, in November 1983. Shortly thereafter the geopolitical situation completely changed with the coming to power of Mikhail Gorbachev in March 1985. He and Reagan met in Reykjavik a year and a half later, in October 1986, and came close to reaching an agreement in principle on a general disarmament. I discuss the reasons for the summit’s failure in chapter 2. Détente had nonetheless been established and led in the following year to the INF Treaty, signed on 8 December 1987. It is this treaty that Trump and Putin repudiated in February 2019.

To no one’s surprise, the two leaders accused each other of bad faith. Both could plausibly argue that the other nation had long been in violation of the INF Treaty. Trump, it has been said, had no policy in any area that did not consist merely in undoing what Barack Obama had done before him, but in this regard at least he was a worthy successor. In 2014, the Obama administration had become alarmed by Russian deployment of a cruise missile system that in every respect resembled the kind banned by the INF Treaty. The Russians had begun testing this system long before, however, in 2008, while making no attempt to hide what they were doing; indeed, Putin openly complained five years later that whereas Russia was constrained by the treaty, it was surrounded in Asia by countries that were free to equip themselves with medium-range nuclear weapons, most notably China. The United States, after much indecision, concluded that it had no satisfactory response under the existing regime and declared the treaty dead.

Meanwhile, Russia accused America of cheating—for example, by taking the liberty of installing ballistic missile defense systems in Eastern Europe. Apart from the fact that they violated the ABM Treaty, these shields could easily be converted into offensive weapons. Furthermore, armed drones, which can be adapted to the same purpose as missiles, did not exist in 1987.

NATO, for its part, declared that Russia was wholly to blame for violating the treaty and, more than this, that the treaty was not worth saving—a strange position, endorsed by France, since it was in large measure thanks to this treaty that peace in Europe had been guaranteed and the security of NATO preserved for more than thirty years. In a nuclear world, however, where rationality is indistinguishable from madness, it is not necessary to separate the good from the wicked.

It is against this background that the most recent events are to be seen, including Putin’s decision to invade Ukraine while threatening anyone who might stand in his way with nuclear annihilation. The relative military strength of Russia and the United States needs to be taken into consideration as well. The end of the Cold War brought about a spectacular reversal in the balance of power between Washington and Moscow in terms of their respective shares of nuclear and conventional weapons. Before 1989, the Soviet Union’s superiority in conventional weapons was manifest, and the United States had sought to compensate for this disadvantage by building up its nuclear arsenal. After the collapse of the USSR, the Pentagon turned its attention elsewhere, particularly to regional conflicts for which conventional weapons were better suited than atomic bombs. During this same period, Putin built up Russia’s nuclear arsenal.

It was not nuclear warfare in general that the United States neglected, but tactical nuclear warfare. American doctrine called for conventional weapons to be used in regional conflicts and, in the event that escalation for the purpose of de-escalation became necessary, reliance on strategic nuclear weapons delivered by ICBMs. Today the United States has only a hundred or so tactical nuclear warheads in Europe, distributed between Germany, the Netherlands, Belgium, Italy, and Turkey. Russia has perhaps twenty times this many, a substantial number of them based in the exclave wedged between Poland and Lithuania (a strategic place if ever there was one) known as the Kaliningrad Oblast—a tragic historical irony, since Immanuel Kant wrote Perpetual Peace (1795) in the capital of this region, known as Königsberg when it was part of Prussia; the parents of Hannah Arendt, who wrote The Origins of Totalitarianism (1951), were from here as well.

Putin is evidently proud of the superiority of his tactical nuclear arsenal, which is what matters in the present instance. More disturbing still, he believes that Russia is better prepared than the United States to ride out a nuclear exchange between the two countries. We have seen that this is the condition of a successful preemptive strike—that is, of successfully attacking first.

Given this much, how did the two nuclear superpowers react to their mutual rejection of the INF Treaty in 2019? Recall that this agreement imposed limits on missiles, whether they carry nuclear weapons or not. The United States and NATO immediately seized the opportunity that now presented itself of stationing nonnuclear intermediate-range and shorter-range missiles in Europe. But they failed to reckon with the Russian response, repeated several times, calling upon the United States and NATO to impose a moratorium on the deployment in Europe of missiles equipped with nuclear warheads. This demand has remained a dead letter. Emmanuel Macron, while categorically refusing to accept it, raised a pertinent question: “Has the absence of dialogue with Russia made the European continent any safer? I don’t think so.”9

Here, one point of technical detail has considerable importance. It is impossible to determine whether or not a ballistic missile is carrying a nuclear warhead until it reaches its target. Faced with this indeterminacy, Russia has elected to treat any missile approaching its territory as evidence of a nuclear attack. According to official doctrine, this is reason enough for Russia to launch its own nuclear missiles even before enemy missiles land on its soil—forcing the United States, which had thought it had a free hand once more to deploy both conventional and nuclear missiles in Europe, to think again. All this, it needs to be kept in mind, took place in the months leading up to Putin’s invasion of Ukraine.

In the first chapter of this book I consider a similar case, the upshot of which is that, in the nuclear age, it is necessary to treat all alerts produced by early warning systems, whether true or false, as though they were true. This is an essential property of the theory of apocalyptic prophecy that I argue for here: once a major catastrophe appears to be possible, one must assume that it is going to occur. In the case of the war in Ukraine, as in all others of its kind, the rule is this: given several possible outcomes, it is necessary to concentrate on the worst one, however great or small the degree of its future necessity10—and this in order to prevent it from occurring.

I am aware that I have not answered the question that is uppermost in everyone’s mind. Will Putin launch nuclear missiles against a Ukrainian or a European city? As philosophers often do, I have reformulated it. This apocalyptic scenario is evidently a possibility, since nothing stands in the way of an error or some unlucky accident precipitating catastrophe, whether due to “noise” in the system, a miscommunication, or the diabolical cycle of humiliation that creates resentment and then causes it to be acted on. It needs to be assumed, as I say, that the worst is going to occur, and, for just this reason, everything that can be done must be done to ensure that it will not occur.

The French president was certainly imprudent in saying, as he did on 3 June 2022, that Russia must not be humiliated.11 This was interpreted in psychological and moral terms, as if to say, shockingly, that due attention needed to be paid to the feelings of the aggressor. Instead Macron should have said that the world is at the mercy of the whims of one of its leaders, meaning that the international system is structurally unstable—an academic way of saying extremely fragile.

My analysis almost completely neglects the geopolitical aspect of the question. I do not by any means wish to minimize its importance. My purpose in what follows is simply to demonstrate the decisive power of the instrument of destruction employed: an atomic weapon. The instrument is not neutral; whether it brings about good or evil does not depend on the intentions of those who make use of it. Not long ago America mourned for the victims of a mass shooting that took the lives of nineteen children who were scarcely ten years old. The killer was eighteen years old. From this latest episode, a good many Americans concluded that access to firearms should be expanded rather than limited. Violence is needed to cast out violence. Violence alone, it is believed, can protect against the mental illness and radical malignancy of others. Whoever makes this argument is blind to the quasi-autonomy of decision that firearms have acquired. It is as though possessing them means giving up one’s free will. Likewise, as I trust this book will demonstrate, the mere possession of nuclear weapons is a moral abomination.

24 January 2023


1. I have been led in particular, as we will see in chapter 4, to propose a new solution to one of the oldest metaphysical aporia, the Master Argument of Diodorus Cronus.

2. Snow should have written “mathematical necessity.” A physical chemist as well as a novelist, known for his theory of the “two cultures,” Snow was a powerful mind, but pardonably unfamiliar with the philosophical distinction between necessity--an ontological category--and certainty--an epistemological category.

3. Thomas C. Schelling, “An Astonishing Sixty Years: The Legacy of Hiroshima,” Nobel Prize Lecture delivered at Beijersalen, The Royal Swedish Academy of Sciences, Stockholm, 8 December 2005.

4. The word chance, if only through its etymology, refers to the concepts of contingency and probability, which, as I go on to show, are not adequate for our purposes. For the moment, however, it will do.

5. See Scott D. Sagan, “The World’s Most Dangeous Man: Putin’s Unconstrained Power Over Russia’s Nuclear Arsenal,” Foreign Affairs, 16 March 2022,

6. The yield of such weapons may nonetheless be as much as twenty times that of Little Boy, the bomb that destroyed Hiroshima.

7. Critics of nuclear deterrence typically assert the opposite. They hold that nuclear deterrence is an abstract theory assuming perfect rationality on the part of actors. Since this condition cannot be satisfied in practice, they conclude that the very concept of nuclear deterrence is a nonstarter.

8. Alexei Arbatov, “Reducing the Role of Nuclear Weapons,” paper presented to the International Conference on Nuclear Disarmament, Oslo, Norway, 26–27 February 2008, pp. 5–6; the emphasis is mine.

9. Reuters, Aerospace and Defense News, 28 November 2019, quoted by Brennan Deveraux, “Why Intermediate-Range Missiles Are a Focal Point in the Ukraine Crisis,” War on the Rocks, 28 January 2002,

10. I employ this rather unusual formulation deliberately, for it saves me from having to rely on the notion of probability, which, as I say, has no validity here.

11. Reuters, “Russia Must Not Be Humiliated Despite Putin’s ‘Historic’ Mistake, Macron Says,” 4 June 2022,