READ IN French
Political Science / Book Review

Vol. 4, NO. 4 / July 2019

The Doomsday Machine: Confessions of a Nuclear War Planner
by Daniel Ellsberg
Bloomsbury, 425 pp., $27.00.

Daniel Ellsberg is well known for releasing the Pentagon Papers to the New York Times and the Washington Post. Their publication led indirectly to the Watergate scandal and the downfall of Richard Nixon. Less well known is that in 1961, while working for the RAND Corporation as an economist specializing in rational choice theory, Ellsberg was seconded to the White House and the Pentagon to work alongside Robert McNamara drawing up plans for a nuclear war. This period is the subject of Ellsberg’s new book.

When he left RAND in January 1970, Ellsberg made copies of all the documents in his safe, not only those pertaining to the Vietnam War.1 It seemed obvious to him that making known the horrifying risks involved in American nuclear strategy was much more important than denouncing the Vietnam scandal. Upon reflection, Ellsberg elected to release the Vietnam papers first. At the beginning of the book, he explains this decision in pragmatic terms, namely the urgency of ending an absurd war. But the remainder of the book suggests another, deeper reason: a feeling of helplessness. Despite the inherent dangers of nuclear weapons, it seemed to Ellsberg that the world remained “blind to apocalypse,” as Günther Anders has put it.2

The Doomsday Machine is a fantastic book, albeit one with an extraordinary omission: the index contains no entry for “deterrence,” a concept that is usually integral to any discussion about nuclear warfare. The word itself appears here and there throughout the book, but deterrence is never discussed as a key concept. Within the extensive literature on atomic weapons, such an oversight would be unprecedented. As it turns out, the omission was intentional.

A Weapon of Use

Nuclear weapons are widely believed to be weapons not meant for use. In seeking to acquire them, a nation has only one goal: to prevent others from using them. This is the notion at the heart of deterrence theory. The most basic form of deterrence theory is known as MAD, short for Mutually Assured Destruction. Ellsberg also has little to say on this topic. Above and beyond its formulation as a doctrine or strategy, MAD is a situation, and therefore also a state of affairs. This is certainly the case if two opposing nuclear powers exhibit both vulnerability and invulnerability. On the one hand, they are both vulnerable because they can be destroyed by their opponent. On the other hand, they are also both invulnerable, because they will not be destroyed without also destroying their opponent. The latter implies that, regardless of any attacks launched by their opponent, a second-strike capability is still retained. Contrary to all other military doctrines, it is essential that neither protagonist tries to protect against itself attacks from their opponent.3

In principle, MAD ensures that neither side attacks first—to do so would be suicidal. It is for this reason that the deterrent threat must be credible, even when it seems completely senseless to put it into effect. In 1969, Nixon was informed that the only practical means of retaliation in the event of a Soviet nuclear attack would kill tens of millions of people within a matter of hours. Nixon was appalled. His national security advisor, Henry Kissinger, is reported to have exclaimed, “How can one rationally … make a decision to kill eighty million people?”4 Of course, it made no sense whatsoever for Kissinger to appeal to any ethical considerations. Acts with such monstrous consequences defy judgment on those terms. This disregard for rationality is, in fact, fundamental to the internal logic of deterrence theory.

In the summer of 2017, tensions between the North Korean leader Kim Jong-un and US President Donald Trump escalated dramatically. Despite the fact that a nuclear confrontation on the Korean peninsula would inevitably result in millions of deaths, the two leaders displayed little restraint, exchanging threats and insults with reckless abandon. North Korea spoke of exacting “thousands-fold” revenge against the US, while Trump boasted that the US would respond with “fire and fury like the world has never seen.”5 In his dealings with North Korea, Trump’s behavior has been seen by some commentators as imitating tactics previously used by Nixon.6 During the Vietnam War, Nixon had theorized that if he pretended to be exasperated to the point of madness by trying to end the war, the North Vietnamese would beg him to make peace immediately. Far from being the sole invention of Nixon, the madman theory is, in fact, integral to the MAD doctrine, a point made by Thomas Schelling in The Strategy of Conflict.7 It can sometimes be rational, he suggested, to mimic irrationality.8

While MAD ensures that neither side will attack first, the ensuing state of peace is so singular as to have acquired a distinct, albeit oxymoronic, description: nuclear peace. In Ellsberg’s view, it would be a mistake to assume that attaining this paradoxical state of peace has been the objective of a nuclear-armed United States. This is the main thesis advanced in The Doomsday Machine. Since the administration of Harry Truman, there has not been a single US president, Ellsberg notes, who has pledged not to use nuclear weapons first. Many prepared for a preemptive strike in response to an imminent attack, whether conventional or nuclear. Preemption is retaliation by anticipation—striking second first. Deterring a surprise nuclear attack by the USSR, according to Ellsberg, has never been a priority for the United States. Instead, the goal has been to minimize the damage caused by the response to a first strike made by the United States.

Finding evidence in support of Ellsberg’s ideas is not difficult. Prior to becoming Trump’s new national security advisor, John Bolton wrote in a February 2018 op-ed for the Wall Street Journal that he considered it “perfectly legitimate for the United States to respond to the current ‘necessity’ posed by North Korea’s nuclear weapons by striking first.”9 In a television interview that took place two years earlier during the presidential campaign, MSNBC’s Chris Matthews attempted to obtain an undertaking from Trump that he would never initiate a nuclear attack against Europe. “I’m not going to use nuclear [sic],” the candidate remarked, “but I’m not taking any cards off the table.”10 Trump’s response was widely criticized and relentlessly mocked, as if it were a blunder. It has also been suggested, incorrectly, that the sentence he uttered contains a contradiction. “Trump was taking the same position of every president since Truman,” Ellsberg notes, “and of every major candidate in that long period, definitely including his rival Hillary Clinton.”11

Contrary to popular opinion, Ellsberg believes that the US military has never regarded the use of nuclear weapons as a taboo—whether for threats or actual strikes. Nuclear weapons are far from the weapons of nonuse suggested by the theory of deterrence. With an overwhelming array of examples, Ellsberg demonstrates that when it comes to deciding between preemption and retaliation, the former has always been preferred, and there is no reason to think that it will not triumph in the future. It turns out that the world is much more dangerous than we had imagined. Ellsberg is, of course, obliged to acknowledge that since 1945, nuclear confrontation has not progressed beyond being a threat. He does not consider the possibility that deterrence—in the sense that it can be seen as a situation or state of affairs, rather than a doctrine—might not have been entirely for nothing.

Ellsberg’s account of the Cuban missile crisis, in which he was a major player, is perhaps the strongest example he presents. He recounts the events of October 1962 with the skill of a thriller writer, the suspense building as the crisis unfolds. John F. Kennedy and Nikita Khrushchev were fully aware that the survival of humanity was at stake, and neither sought a nuclear conflict. Nonetheless, despite the obvious danger, the pair proceeded to threaten each other for entirely pragmatic reasons. In order to wring concessions from their opposite number, such as the withdrawal of American missiles from Turkey in exchange for the Soviets leaving Cuba, the leaders were prepared to delay reaching an agreement for a few hours—at a time when it had been estimated that there was a ten percent chance of a nuclear apocalypse.12 Clearly, this scenario has little connection to MAD.

Orders of Magnitude

A  recurring theme throughout the book is the relentlessly extreme character of the phenomena, events, and magnitudes that Ellsberg contemplated during his time as a nuclear strategist. He describes the shock he felt in the spring of 1961 when he learned that, if implemented, the plans for nuclear war that he had helped develop would, in all probability, result in six hundred million deaths—“[a] hundred Holocausts.”13 These figures are so excessive that they fail to provoke feelings of fear, revulsion, sorrow, compassion, or even pity. At this scale, strategic calculations are so far removed from any sense of humanity as to be completely meaningless.

At one point in the book, Ellsberg describes a truly horrific scene:

People fleeing suffocation in the shelters took to the streets to escape and became blazing torches unable to move in the melting asphalt. [The city], like Venice, was covered with canals, to which mothers raced with their children to get away from the heat. The smaller canals began to boil, and families boiled to death by the thousands.14

Although these events took place in Japan toward the end of the war, the location was not Hiroshima or Nagasaki. On the night of March 9, 1945, five months before the atomic bombs were dropped, Tokyo was destroyed in a firebombing raid by the US Air Force. Around 100,000 people were killed. Achieving this level of death and destruction required over 300 bombers and more than 1,500 tons of bombs, most of which were incendiary devices.15 A single atomic bomb achieved a similar result in both Hiroshima and Nagasaki. The officer who ordered the use of incendiary bombs, General Curtis LeMay, later the head of the US Strategic Air Command between 1948 and 1957, saw only one significant difference between conventional and atomic weapons: increased efficiency.

Along with Ellsberg, the historians Gar Alperovitz and Barton Bernstein are insistent that Truman’s decision to resort to nuclear weapons was not preceded by any ruminations on morality.16 The threshold of moral horror had been crossed long before Truman made his decision in the days following the successful Trinity test in New Mexico on July 16, 1945. “Annihilation of an urban civilian population by fire,” Ellsberg remarks, “had already become the American way of war from the air, as it had been the British way since late 1940.”17 All cats are gray in the dark.

After the war, American strategists were seemingly untroubled by the transition from the atomic to the hydrogen bomb, much as they had been for the transition from incendiaries to the atomic bomb—despite expected casualty rates in the tens of millions. In 1961, Ellsberg was dismayed to discover that the estimated number of Soviet casualties resulting from an American first strike had increased tenfold during a single twelve-month period in the 1950s; it now stood at “more than two hundred million in the Soviet bloc alone.”18 Initially, he was at a loss to explain the increased estimates. “Had someone concluded,” Ellsberg pondered, “that ‘killing a nation’ with four hundred atomic bombs that would kill tens of million Russians was not [emphasis original] enough devastation for deterrence?” As it turned out, the plans of attack had not changed, nor had the strategy, nor the targets. Fission bombs had simply been replaced with fusion bombs that were ten times more lethal. Arithmetic had won out again.

The extreme character of these figures is matched in counterpoint by similarly horrifying calculations involving vanishingly small probabilities. In 1942, a group of physicists working on the Manhattan project debated a question raised by Edward Teller: whether detonating a nuclear device would ignite the atmosphere, ending all life on earth. One of the physicists, Hans Bethe, looked into the problem and “found that it was just incredibly unlikely.”19 Although concerns about such an event were dismissed by many of the senior scientists, some of the project’s leaders, such as Arthur Compton, remained fearful. In a 1959 interview quoted by Ellsberg, Compton described how the risk of catastrophe was assessed.

If, after calculation, [Compton] said, it were proved that the chances were more than approximately three in one million that the earth would be vaporized by the atomic explosion, he would not proceed with the project. Calculations proved the figures slightly less—and the project continued.20

In retelling a breathtaking story, Ellsberg, despite his expertise in rational choice theory, chooses to focus on the leading figures in the lead up to the Trinity test, asking whether they really believed that atmospheric ignition was a realistic possibility.21 The absence of a theoretical perspective on these deliberations and calculations is one of the book’s limitations. Still, we can hazard a guess at the reasoning employed by those involved. They likely had neither the time nor the background to ponder the issue from a metaphysical point of view. Suppose that epsilon represents the probability that the catastrophe will occur. Between a positive, albeit vanishingly small, value for epsilon and zero epsilon there is a yawning chasm. A calculation of mathematical expectation has no meaning. There is no positive value of epsilon for which the odds would justify risking the apocalypse.22

Doomsday Machines

The best way to ensure the success of a nuclear first strike, one might assume, is to target the opposing president or leader, the individual with sole authority to press the button. In an effort to counter this vulnerability, suppose that an enemy nation has delegated the power to launch a nuclear attack to multiple people, each of whom can, under certain circumstances, also delegate that power. Suppose, moreover, that alongside this cascading structure, reprisals are launched automatically and that this task has been assigned to a computerized system. The deterrent power of this arrangement seems assured. But suppose, finally, that the enemy nation fails to inform its rival about the existence of the automatic reprisal system. If the first nation then seeks to decapitate its adversary, it will be confronted with a thermonuclear hydra—mutual assured destruction would be inevitable.

This absurd scenario evokes the dark humor of Dr. Strangelove. In the film, the Soviets invent a doomsday machine that responds automatically to an American nuclear attack by detonating a network of buried nuclear bombs and rendering the earth uninhabitable for a century, but they neglect to warn the US leadership of its existence. As Martin Hellman has noted, far from being simply satire, the film could almost be considered a documentary.23 “Everything in that film,” Ellsberg has remarked, “existed as an operational reality at the time.”24

During the Cold War, neither the Americans nor the Soviets thought to let each other know that they had each developed doomsday machines by delegating the power to launch attacks beyond their leadership. At the height of the Cuban missile crisis, the American leaders were unaware that Soviet submarines patrolling the waters around Cuba were equipped with nuclear warheads. As a result, the Americans almost provoked nuclear attacks on several occasions by putting Soviet submarine commanders in situations in which they could have chosen to launch reprisals. Ellsberg acknowledges the lunacy of this situation. To admit to the world that a doomsday machine had been developed, he explains, was to risk terrifying one’s countrymen, allies, and, of course, the rest of the planet.25 This secretiveness reveals just how little importance deterrence had for the leaders of the United States and the Soviet Union. And this is Ellsberg’s point. The fact that each nation did not inform the other about the existence of a deterrent is proof that deterrence was not their goal.

Ellsberg broadly defines the doomsday machine as

a very expensive system of men, machines, electronics, communications, institutions, plans, training, discipline, practices, and doctrine—which, under conditions of electronic warning, external conflict, or expectations of attack, would with unknowable but possibly high probability bring about the global destruction of civilization and of nearly all human life on earth.26

Rather than focusing on the dangers posed by nuclear proliferation and terrorism, widely considered the most pressing threats in the current era, Ellsberg believes that it is the American and Russian doomsday machines that constitute the greatest danger to humanity. These two doomsday machines, he writes, must be dismantled: “And that, at minimum, is what we must hasten to do.”27 Each is “susceptible to being triggered on a false alarm, a terrorist action, unauthorized launch, or a desperate decision to escalate.”28

At 8:10 a.m. on January 13, 2018, the inhabitants of the Hawaiian archipelago received an alert on their mobile phones: “BALLISTIC MISSILE THREAT INBOUND TO HAWAII. SEEK IMMEDIATE SHELTER. THIS IS NOT A DRILL.” Subsequent investigations determined that a government official had mistaken a message sent during a civil defense training exercise for a real alert, and then relayed it to the population using an emergency warning system. It took thirty-eight minutes for a false-alarm notification to be sent. Three days later, a similar situation unfolded in Japan. An erroneous alert warning of an incoming North Korean missile was broadcast, urging people to take shelter. Mercifully, a correction was sent five minutes later.

Thirty-eight minutes, let alone five, is more than enough time for an alert to reach military commanders and political leaders. After an incoming attack has been detected, the president of the United States has between five and ten minutes to decide whether to launch a battery of intercontinental ballistic missiles. Waiting any longer would risk the missiles being destroyed in their silos. It is not inconceivable that a false alarm could trigger a global nuclear conflict.

“I do not know with what weapons World War III will be fought,” Albert Einstein remarked in 1949, “but World War IV will be fought with sticks and stones.”29

Translated and adapted from the French by the editors.


  1. The copied documents concerning nuclear warfare were subsequently lost in unusual circumstances and have never been found. Ellsberg notes that “a good deal of what was lost has since been declassified, in particular over the last thirty-two years by FOIA requests.” Daniel Ellsberg, The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017), 10. 
  2. This is a reference to the words of the German philosopher and critic of the atomic bomb, Günther Anders, who described us as living in a state of Apokalypse-Blindheit (apocalypse blindness). Günther Anders, “Über die Bombe und die Wurzeln Unserer Apokalypse-Blindheit” (On the Bomb and the Roots of our Apocalypse-Blindness), in Die Antiquiertheit des Menschen, vol. 1 (Munich: C. H. Beck Verlag, 2002). 
  3. The incarnation of this condition is the Anti-Ballistic Missile Treaty signed by Leonid Brezhnev and Richard Nixon in 1972. 
  4. Daniel Ellsberg, The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017), 271. 
  5. Meghan Keneally, “From ‘Fire and Fury’ to ‘Rocket Man,’ the Various Barbs Traded Between Trump and Kim Jong Un,” ABC News, June 12, 2018. 
  6. Jerome Slater, “Trump Is Using the Madman Theory in North Korea Policy,” The National Interest, June 3, 2018; Steve Coll, “The Madman Theory of North Korea,” The New Yorker, September 24, 2017; Tim Naftali, “The Problem with Trump’s Madman Theory,” The Atlantic, October 4, 2017; David Sanger, “Trump on North Korea: Tactic? ‘Madman Theory’? Or Just Mixed Messages?The New York Times, April 28, 2017. 
  7. Thomas Schelling, The Strategy of Conflict (Cambridge, MA: Harvard University Press, 1960). 
  8. Ellsberg, who was Schelling’s student, does not quote his mentor’s book, but agrees with this idea. He recalls the lecture he gave in 1959 titled “The Political Uses of Madness,” given at Henry Kissinger’s seminar at Harvard. This is the same Kissinger who would later call Ellsberg “the most dangerous man in America.” It should not be forgotten that Ellsberg is an expert in the theory of rational choice and game theory. He has contributed much to the theory of decision-making while facing the uncertain future. He is the author of the well-known Ellsberg paradox, questioning Leonard Savage’s theory of expected utility. Daniel Ellsberg, The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017), 310–11. 
  9. John Bolton, “The Legal Case for Striking North Korea First,” The Wall Street Journal, February 28, 2018. 
  10. Daniel Ellsberg, The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017), 330. 
  11. Ibid., 330. 
  12. Although it goes unmentioned, Ellsberg is aware that this was a zero-sum game with low stakes being played out on the precipice of an abyss. 
  13. Ibid., 2–3. 
  14. Ibid., 258. 
  15. Wikipedia, “Bombing of Tokyo.” 
  16. Gar Alperovitz, The Decision to Use the Atomic Bomb and the Architecture of an American Myth (New York: Vintage Books, 1996); Barton Bernstein, “The Atomic Bombings Reconsidered,” Foreign Affairs 74, no. 1 (1995): 135. 
  17. Daniel Ellsberg, The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017), 261–62. 
  18. Ibid., 270. 
  19. John Horgan, “Bethe, Teller, Trinity and the End of Earth,” Scientific American, August 4, 2015. 
  20. Pearl S. Buck, “The Bomb—The End of the World?The American Weekly, March 8, 1959. 
  21. Daniel Ellsberg, Chapter 17: “Risking Doomsday I,” in The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017). 
  22. Ultimately, the choice of protagonists was implicitly reduced to the choice of a temporal metaphysics. For further developments on the temporal metaphysics applied to the nuclear threat, see my essay “MAD-Made World,” Inference: International Review of Science 3, no. 3 (2017), and especially the much lengthier French version, which was published in the same issue under the title “La dissuasion nucléaire et la métaphysique du temps.” Interested readers might also consult my recent book, La Guerre qui ne peut pas avoir lieu : Essai de métaphysique nucléaire (Paris: Desclée de Brouwer, 2019). 
  23. Martin Hellman, lecture given at the Center for International Security and Cooperation (CISAC), Stanford University, February 5, 2018. 
  24. Quoted in “‘Dr. Strangelove’ is Basically a Documentary,” Wired, March 3, 2018. 
  25. Daniel Ellsberg, The Doomsday Machine: Confessions of a Nuclear War Planner (New York: Bloomsbury, 2017), 306. 
  26. Ibid., 339. 
  27. Ibid., 340. 
  28. Ibid., 339. 
  29. Albert Einstein, in an interview with Alfred Werner, Liberal Judaism 16 (April–May 1949), Einstein Archive 30–1104, quoted in Alice Calaprice, The New Quotable Albert Einstein (Princeton: Princeton Universty Press, 2005), 173. 

Jean-Pierre Dupuy is Professor Emeritus of Social and Political Philosophy at the École Polytechnique in Paris and Professor of Political Science at Stanford University.

More from this Contributor

More on Political Science


Copyright © Inference 2024

ISSN #2576–4403