Introduction
On 25 November 2016, Fidel Castro died. He was the last surviving elite participant in the 1962 nuclear crisis, widely regarded as the closest humanity ever came to nuclear war. With his passing we have lost a direct link with the experience of very intense fear of imminent nuclear war, and with the learning of the crucial role of luck in preserving the world from nuclear devastation. From now on, our interpretation of the danger of the most dangerous crisis in the history of the nuclear age is radically detached from direct experience at the highest levels of decision-making.
At the same time, all nuclear weapons states are developing vast programmes to enhance their nuclear weapons capabilities, tensions between Russia and the West remain high, and the current US president is suspected by some to be more prone to using nuclear weapons in anger than any of his predecessors.Footnote 1 In such a context, this article offers a broader investigation of our beliefs about the ability to control nuclear weapons and manage nuclear crises based on a study of the so-called ‘Cuban Missile Crisis’ (hereafter referred to as the Crisis).
This crisis, which today is widely considered to be the closest mankind has ever come to nuclear war, is an essential case study for assessing the fear-inducing effects of nuclear weapons and understanding the possibility of nuclear learning.Footnote 2
‘Learning’ in this context means learning from history, and it assumes that the interpretation of key events in the nuclear age plays a decisive role in the behaviour of policymakers in crisis situations.Footnote 3 My understanding of learning is based on the three following premises. First, I accept the assumption that national experience is a major source of learning.Footnote 4 Second, I accept the finding of the literature portraying overconfidence as a source of increased risks.Footnote 5 Third, I assume that shared international learning about the limits of controllability of nuclear weapons is an important precondition for more informed nuclear decision-making and public deliberations on the subject.Footnote 6 As a consequence of these three premises, I regard the absence of learning, or forms of memory that systematically deny the role of luck and promote overconfidence, as contributing to nuclear danger.Footnote 7 The fact that learning is rare does not make its absence any less problematic or puzzling.
The core puzzle of this article is the following: the latest assessment of the Crisis emphasises the underestimation of the danger at the time, the limits of control over nuclear weapons, and the role of luck in the peaceful outcome of the Crisis. However, not all policy and scholarly communities have taken these insights seriously. The unbearable lightness of luck seems constantly to escape the learning process.Footnote 8 This article explores three reasons for this failure to learn, focusing on ideational factors. I do not deny the role of institutional and bureaucratic dynamics in the entrenchment of representations related to nuclear weapons, which will be explored in a subsequent essay, but I do not analyse them here. Staying at the level of ideas emphasises the responsibility of scholars and analysts to work in the name of avoiding overconfidence without waiting for structural or institutional change.Footnote 9
Forms of learning that promote overconfidence have been documented across national cases. In India and Pakistan, as Russel Leng and S. Paul Kapur have noted: ‘the learning that has occurred has been largely dysfunctional and dangerously hawkish’.Footnote 10 In the US, Pulitzer Prize winner and longtime analyst of US nuclear weapons policy Richard Rhodes similarly observed: ‘despite several close calls, … no one in authority believes the damned things will go off, and so everyone wants to play with them, like treasure hunters wallowing in a vault of golden coins laced with guardian scorpions, like children discovering the loaded gun their parents thoughtlessly neglected to lock away.’Footnote 11 By contrast with the US and the UK, the other two nuclear-armed NATO states, France has received relatively little attention. Drawing on unexploited primary sources to illuminate this less-known case, the following article identifies and explains the limits to French nuclear learning.
In this context, France is a relevant case for two reasons. First, as Beatrice Heuser eloquently showed: ‘while nuclear weapons also have a metaphysical dimension for other nuclear powers, it is most developed in France’.Footnote 12 France displays in particularly acute form some of the sources of overconfidence in the controllability of nuclear crises that can been found in other nuclear armed states. If official representatives of all nuclear weapons states tend to portray themselves as responsible custodians of their nuclear arsenals, the French stand out for their public display of confidence in the perfect safety record of their nuclear arsenal. On 1 May 2015, French Ambassador Jean-Yves Simon-Michel stated before the Main Committee I of the Nonproliferation Treaty Review Conference that: ‘There is no risk of the [French nuclear] weapons being used non-intentionally.’ Two weeks before, he had given a tour of the Luxeuil Air Force base and its former weapons storage facilities. He added that the visitors ‘will have seen [this] for themselves’.Footnote 13 This article treats such public confidence in control over nuclear weapons as part of a puzzle, and analyses how such a statement has become possible and acceptable. Second, there is only very limited scholarship on the French experience and memory of the Crisis. In terms of historiography, most of the scholarship on the Crisis still tells the story of a bipolar confrontation and ignores the perspective and agency of actors other than the US and USSR, Cuba included.Footnote 14 What limited literature does exist on France is not available in English and ignores most of the primary and secondary sources analysed in this article: recently published memoirs, declassified archival material, testimonies of former policymakers and high-level military personnel, representations of the Crisis in the French media, specialised military and intellectual publications, and high school textbooks.Footnote 15 Overall, little has been written on the memory of key episodes in the nuclear age in France and their impact on French thinking about nuclear safety, security and possible nuclear futures.Footnote 16
In addressing the puzzle of how the discoveries about the role of luck in the outcome of the Crisis, never adequately refuted, have not been integrated in the accepted body of knowledge about nuclear history and policy, this article makes three contributions. First, the focus on new French primary sources is relevant to security studies as well as policymaking circles given frequent misconceptions about nuclear dynamics. Nuclear security studies scholars too often make unwarranted theory-driven extrapolations based on few or no primary sources, and latent assumptions about a common narrative of the danger, causes and consequences of the ‘most dangerous crisis of the nuclear age’ are widespread in Western nuclear policy circles.Footnote 17 Second, and consequently, the article tackles a widespread assumption of automaticity linking the presence of nuclear weapons to a fear-induced deterrent effect. Vipin Narang labelled this as ‘an existential bias’ in the literature on deterrence, focused on the acquisition of nuclear weapons as the only meaningful threshold, a bias that Daniel Deudney locates in the tradition of ‘automatic deterrence statism’ in the literature on nuclear weapons policy.Footnote 18 Indeed, this assumption crosses the policy divide over the desirability of nuclear disarmament. On the anti-disarmament side, Susan Martin recently summarised the issue in the following terms: ‘nuclear weapons cannot be stripped of their values as a strategic deterrent’ since ‘[their] strategic deterrent capability … stems from their material characteristics’.Footnote 19 As James Lebovic aptly summarises, in this tradition, the deterrent effect does not come from rational calculation but from fear: ‘the existential deterrent acquires its power from the non rational world of fear’.Footnote 20 On the pro-disarmament side, proponents of ‘virtual nuclear deterrence’ or ‘weaponless deterrence’ have to subscribe to an even more radical version of this idea for their concept to make sense. In their case, the weapons do not even need to be assembled in order to deter, their existence and the fear that they trigger need only exist in the mind of the deterred.Footnote 21 In other words, whether it is for deterrence or disarmament, the mere presence of nuclear weapons (physical or virtual) is expected to inspire fear on account of their destructive capacity; such a dynamic, if granted, would clearly constrain the scope for re-evaluation of these weapons. In other words, the value of nuclear weapons would be irreducibly determined by the physics of nuclear devastation and unaffected by social and political factors. Given the position of France in the middle of the expected European battlefield for the coming nuclear war of the 1960s, if it turns out that the French experience and memory of the Crisis were largely unmarked by fear, this widely-held assumption of the automaticity of the effect of the presence of nuclear weapons will be opened to further challenge. As a consequence, the valuation of nuclear weapons will be reassessed as a socially and historically constructed process, irreducible to destructive capacity alone.Footnote 22 Third, empirically, this study of the French case responds to Nick Ritchie’s invitation to study ‘regimes of value’ of nuclear weapons beyond his foundational case study of the UK.Footnote 23 It also contributes to a growing literature on nuclear fear and atomic anxiety that has not documented the French experience in detail or consulted French-language primary sources.Footnote 24
The rest of this article is divided in five parts. I first review the most recent literature on the Crisis, with evidence showing that the situation was indeed extremely dangerous and that its peaceful outcome cannot be reduced to successful, fully informed crisis-management. Nuclear weapons use was avoided in the autumn of 1962 not through restraint on the part of President Kennedy and the Soviet leadership only, but as a result of decisions made by individual nuclear operators, under conditions of incomplete or incorrect information. I then identify three reasons for why this interpretation, strengthened over the past three decades and never persuasively refuted, has still not been fully incorporated in scholarship and public discourse: rhetorical practices of epistemic and practical inconsistencies that acknowledge the role of luck but do not treat it adequately and bring it back within the realm of controllability; a disciplinary rejection of counterfactual thinking as a legitimate scholarly practice and the memory of an idiosyncratic experience of the Crisis exempt of fear. It is historically inaccurate and politically unwise, I argue, to expect that analysts’ post-facto confirmation of the danger confronted at the time will entail a corresponding learning process on the part of policymakers and the public, and I conclude by outlining the implications of this point for the scholarship in history and security studies.
The latest assessment of the possibility of nuclear use during the Cuban Missile Crisis
Even though discussion over the dangers of the Crisis continues, scholarship from the last three decades has reached compelling conclusions about the role of luck in its outcome, contingent on limited information, misperceptions, the limits of safety and presidential command and control over nuclear weapons, as well as the potential of accidents. As a 2015 literature review concluded, ‘had everyone stopped researching … the crisis in the 1980s our awareness of the risks of nuclear war would be greatly diminished. If we had chosen to draw the lessons on the basis of our knowledge, we would almost certainly have drawn the wrong lessons.’Footnote 25
First, the limits of political leaders’ knowledge about nuclear weapons during the Crisis have been exposed. For example, the Kennedy administration massively underestimated the number of Soviet troops in Cuba, ignored that they arguably could fire tactical weapons based in Cuba, and that Castro was actively pushing the Soviets to be more aggressive.Footnote 26 Excomm members, we now know, were not informed when the head of Strategic Air Command (SAC), General Thomas Power, took the unprecedented decision to place American nuclear forces on Defense Condition 2. One of those missiles was even launched for a test flight from Vandenberg Air Force base, in California, on 26 October 1962 at 4 am, as planned before the Crisis, without the president knowing about it.Footnote 27
Second, and consequently, command and control problems also have to be taken into account, so that complete presidential control cannot be taken for granted. As Scott Sagan wrote twenty years ago, in a classic account:
Many serious safety problems, which could have resulted in an accidental or unauthorized detonation or a serious provocation to the Soviet government, occurred during the crisis. None of these incidents led to inadvertent escalation or an accidental war. All of them, however, had the potential to do so. President Kennedy may well have been prudent. He did not, however, have unchallenged final control over U.S. nuclear weapons.Footnote 28
For example, during the night of 26–7 October 1962, at the height of the Crisis, an American U-2 strayed into Soviet air space over the Arctic. Soviet fighter jets scrambled to intercept the U-2 while F-102 interceptors were sent to escort it home and prevent the Soviet MIGs from freely entering US airspace. Given the circumstances, the F-102s’ conventional air-to-air missiles had been replaced with nuclear-tipped ones and their pilots could decide to use nuclear weapons: ‘the interceptors at Galena were armed with the nuclear Falcon air-to-air missiles and, under existing safety rules, were authorized to carry the weapons in full readiness condition in any “active air defense” mission.’Footnote 29 Fortunately, the spy plane turned back and the Soviet jets held their fire.Footnote 30 On the Soviet side, a similar problem of the delegation of the authority to use nuclear weapons occurred.Footnote 31 Moreover, it is now well established that President Kennedy was often in a minority in opposing military action against Cuba during the 13 days of the Crisis (to follow the American periodisation).Footnote 32 A usual interpretation of the Crisis assuming complete control from the two leaders might find this comforting: his ability to resist Excomm pressure shows that he would never have used force, let alone nuclear weapons, during the Crisis.Footnote 33 However, finding confidence in the president’s consistency is in itself problematic. It neglects, for instance, Kennedy’s severe illness, dependence on steroids, and reliance on competing medical opinions and treatments from doctors with whom he did not share what the others were prescribing him, which might have affected his judgement.Footnote 34
Third, the most recent research shows that weapons safety was very rudimentary at the time. For instance, in the early 1960s NATO weapon handlers pulled the arming wires out of a Mark 7 nuclear warhead while they were unloading it from a plane. ‘When the wires were pulled, the arming sequence began – and if the X-Unit charged, a Mark 7 could be detonated by its radar, by its barometric switches, by its timer or by falling just a few feet from a plane and landing on a runway.’Footnote 35 In Italy, a few months before the Crisis, Jupiter missiles were hit by lightning four times, ‘some of their thermal batteries fired, and in two of the warheads, tritium gas was released into their cores’ and may have boosted a nuclear detonation.Footnote 36 One of the missiles rolled down a hill.Footnote 37 So, beyond the command and control problem outlined above, the safety of the weapons was very problematic at the time and accidental detonation of NATO weapons could have happened during the Crisis.
Fourth, those possibilities of accidents and miscalculations could have increased the danger of the Crisis in an indirect fashion, by degrading the common knowledge of the unacceptability of nuclear war that arguably kept the two leaders away from the brink: they could have led one of the two to think that nuclear war was imminent. As a result, one can reasonably claim that the danger of war was higher than leaders thought at the time.Footnote 38 In 2015, Len Scott reviewed the latest developments in the scholarship about the Crisis on this point and concluded: ‘the emphasis on contingency and unacknowledged risk has accelerated with more evidence. Better understanding of the role of misperceptions, miscalculations and mistakes … suggests that the risk of nuclear war was greater than thought by decision-makers at the time and by commentators subsequently.’Footnote 39 Similarly, Réachbha FitzGerald concluded a 2007 review of the literature with the idea that: ‘recent research into the operational aspects of the crisis has revealed that both the US and the Soviet leaderships had less knowledge and control over their military than they or anybody else realized.’Footnote 40 The consensus even goes beyond the literature in English and extends to the award-winning 2014 volume in Italian by Leonardo Campus, I Sei Giorni Che Sconvolsero Il Mondo (The Six Days that Shook the World).Footnote 41
Taken together, all those elements about the limits of controllability and predictability of the Crisis point to a role played by luck in its outcome. In 1994, Len Scott and Steve Smith already observed,
there is now ample evidence that the fact that the crisis did not lead to nuclear war was due, in large part, to good luck. In our view this is a most important finding since it undermines the claims of those who think that nuclear crises can be safely managed and that command and control systems will work as they are meant to work.Footnote 42
Dean Acheson proposed this interpretation as early as 1969. He attributed the outcome of the Crisis to ‘plain dumb luck’, and other key participants would come to share this view after a handful of oral history conferences conducted by James Blight and his colleagues.Footnote 43 Among them were former US Secretary of Defense Robert McNamara and Nikolai S. Leonov, KGB head of Cuban Affairs at the time. This view is now supported by the scholarly evidence presented above and explicitly endorsed by Scott Sagan, who writes that: ‘there was an element of good luck involved in avoiding accidental war in October 1962’.Footnote 44 This cannot be reduced to good management. Twenty years later, Campbell Craig raises important objections against Sagan’s work, but still fundamentally agrees with this point.Footnote 45
In a nutshell, then, even if one cannot know for sure what would have followed the first nuclear strike, the scholarship of the last three decades shows that the Crisis was much more dangerous than leaders thought at the time. Problems of limited or false information, misperceptions, limits of safety and presidential command and control over nuclear weapons as well as accidents susceptible to create the impression that war was imminent, could all have led to inadvertent escalation. A narrative of control cannot fully account for the peaceful outcome of the Crisis and the part played by luck.
Reasserting control by practical and epistemic inconsistencies
After those discoveries, the few voices who still want to argue that the dangers of the Crisis were manageable can only do so via an inconsistency: they at first claim to acknowledge the limits of safety and controllability of nuclear weapons and the unpredictability of all pathways to nuclear war, but they then either reduce the realm of what was possible to what is measurable, or reintroduce the idea of manageability as the inevitable outcome of policy-relevant research. In the first case, which I characterise as an epistemic inconsistency, the risk frame makes it impossible to account for the limits of control and safety identified above or for the role of luck, in particular if luck is defined as an effect of unquantifiable uncertainty.Footnote 46 In the second case, which I characterise as a practical inconsistency, the role of luck and the limits of controllability will be denied in practice so that acknowledging them will not lead to a questioning of managerialism as an approach of nuclear crises on epistemological, ethical, or political grounds.Footnote 47
John Lewis Gaddis is perhaps the most explicit practitioner of the epistemic inconsistency view in English-language literature. He writes:
Just how close [we came to nuclear war], though, remains at issue. … The tendency afterwards was to lower the odds. Hawks found it inconceivable that Khrushchev would have used nuclear weapons to defend Cuba in the face of such overwhelming American superiority. Doves insisted that whatever the numerical balance Kennedy would never have authorized an invasion of the island. … Calculating risks retrospectively is almost as difficult as trying to anticipate them: in any complex system so many things can go wrong that it is difficult to know what might – or might not have. A reasonable place to start, though, is with those in command.Footnote 48
As this quote suggests, this is not only the place where Gaddis’s analysis starts but also where it ends, even though he claims otherwise. In focusing on the two leaders, Gaddis renders the findings about the limits of leaders’ controls about the weapons, the possibility of pre-delegation of use as well as accidental explosions due to the limits of safety, invisible. By emphasising calculable risks and probabilities, he cannot account for unquantifiable pathways to disaster, also called possibilities. Lee Clarke perfectly identified the problem when he wrote: ‘We need to think in terms of chances and odds and likelihoods. But we shouldn’t concentrate so much on probabilities that we forget the possibilities.’Footnote 49
A similar reduction of the unquantifiable/unknowable to the impossible as a result of an approach in terms of risk assessment can be found in Vincent Touze’s otherwise remarkable study of the Crisis. Commenting on Sagan’s scenarios for potential accidents, he confesses: ‘it is true that the developments are very technical and one is not able to forge an opinion’.Footnote 50 As a result, he neglects them entirely as inconsequential. Instead of suspending judgement or acknowledging the limits of what can be known, he even goes so far as to treat possible accidents as only having a confirming effect on previously observed trends: ‘all the incidents would only have had one single … effect anyway: to confirm US resolve and lead the USSR to give up.’Footnote 51 The game-theoretic literature on brinkmanship based on risk thinking that has taken the Crisis as a textbook case is a typical example of this inability to grasp the unbearable lightness of luck; it also treats the unquantifiable as impossible and therefore negligible.Footnote 52 This intellectual move illustrates that, as Mary Douglas aptly wrote, ‘risk is not a thing, it’s a way of thinking’.Footnote 53 Most important, this way of thinking is oriented towards a desire for control and faith in that control.Footnote 54 This implication of risk thinking is most obvious in a research memorandum for RAND, published in 1960 in which the founding father of risk-based nuclear strategy, Thomas Schelling, writes about ‘a controlled loss of control’.Footnote 55 As I have argued elsewhere, Schelling’s legacy of conflating uncertainty and risk in Strategy of Conflict has produced the illusion that risk thinking was able to capture uncertainty and luck. The effect of this illusion has been unduly to reassert control.
Such epistemic inconsistency is frequently displayed in the limited literature in French on the subject. In most other cases, the role of luck is acknowledged in theory but its implications and the questions it would raise are immediately denied at the same time, via its reincorporation in a narrative of control. For instance, Marie-Hélène Labbé, who writes that ‘for a week, the world came close to apocalypse’Footnote 56 in her two-page treatment of the Crisis, notes that ‘later revelations nuanced the idealized image of a perfectly managed crisis’ and recognises the pressure Castro put on Khrushchev to be more aggressive towards the US. However, she ends up underplaying the role of luck in the outcome of the Crisis: the ‘risk [of use of nuclear weapons] … was omnipresent but controlled during the Cold War’.Footnote 57 Similarly, the 2014 Livre noir du nucléaire militaire accepts that the 1962 crisis is the closest humanity came to nuclear war and that it could have been triggered without political approval, but concludes by looking at the controllable side of the problem, suggesting that the two leaders should have learned from it that small nuclear arsenals are enough to deter.Footnote 58 This conclusion obviously ignores the episodes mentioned above in which the peaceful outcome of the Crisis is not reducible to a success of deterrence, notably because it resulted from a decision taken out of false information or because the exact cause of the behaviour is not well established.
Beyond the epistemic inconsistency, scholarship can escape the unbearable lightness of luck without actually refuting it by a form of practical inconsistency. Graham Allison and Philip Zelikow’s 1999 revised edition of the classic but highly criticised Essence of Decision exemplifies this tendency. They ostensibly accept normal accident theory, that is, the idea that tightly coupled and interactively complex systems like nuclear weapons will inevitably lead to accidents, yet they conclude with a recommendation in terms of manageability which is at odds with the theory or would at best assume that another disaster will happen before the nuclear one, which would make it less significant.Footnote 59 They conclude that ‘the potential for dysfunctionality exists and must be managed’,Footnote 60 when the key insight of normal accident theory is that systemic accidents in complex and tightly coupled systems cannot be managed away. This is clear in Scott Sagan’s summary in 2004: ‘scholars in the “normal accidents theory” school have argued that organizations that exhibit both high degrees of interactive complexity and tightly coupled operations will suffer serious accidents despite their efforts to maintain high reliability and safety.’Footnote 61
Former member of the latest two commissions on the French Livre blanc sur la défense et la sécurité nationale (white book on defense and national security) Bruno Tertrais displays a similar form of practical inconsistency. Nuclear crises have to remain manageable, so he starts his short analysis of the risk of nuclear escalation in history by deliberately setting aside the case of the Crisis. He can therefore conclude that: ‘there does not seem to be any example when nuclear weapons have been really “close” to being employed’.Footnote 62 A similarly inconsistent reluctance to acknowledge the role of luck is observable in the writings of the late French nuclear strategist and policy adviser Thérèse Delpech. Given that she recognises Cuba as ‘the most dangerous nuclear crisis so far’ and accepts the widest variety of dangers, it is very revealing that she ends up hedging her acknowledgement of the role of luck in the outcome of the Crisis. She accepts that: ‘during the second peak [of the Crisis, on 27 October], a nuclear war was almost launched on more than one occasion’Footnote 63 and recognises that escalation was possible during the blockade.Footnote 64 She also recognises the lack of information and existence of misperceptions on both sides, and acknowledges that the short-range atomic weapons in Cuba could have been fired.Footnote 65 Finally, she agrees with the highest assessment of the risk of nuclear retaliation and writes that: ‘if those weapons had been fired at U.S. troops, the United States would have retaliated with nuclear weapons’.Footnote 66 Building on Nikolai Leonov’s testimony mentioned above, she concludes, reluctantly and with a caveat that: ‘Luck may have played a part as well.’Footnote 67 This is a foregone conclusion: for practical purposes, nuclear crises have to be managed, so they have to be manageable even if this conclusion depends on an inconsistent double operation of obfuscating the role of luck without being able to disprove it.Footnote 68
Reasserting control by rejecting counterfactual thinking
In this section, I focus on rejection of counterfactual thinking as an epistemological hurdle that has led French scholarship away from a full acknowledgement of the danger of the Crisis. This rejection of counterfactual thinking seems to be a disciplinary limitation of diplomatic history, which has had a particularly strong influence on scholars of the Crisis in France.
A recent debate in Security Studies between a diplomatic historian and his political scientist colleague illustrates the reluctance of diplomatic historians vis-à-vis counterfactual thinking. Frank Gavin, a trained diplomatic historian turned security studies scholar, exposes a frequent concern among diplomatic historians: ‘No matter how plausible, “what-ifs” are not part of our mission. This has led some historians to take a rather dim view of counterfactual exercises, believing they do violence to the craft of historical reconstruction.’Footnote 69 His scepticism appears clearly in his persisting sense that a historical argument without counterfactuals is possible and that some of the historians’ critiques of counterfactual analysis remain valid. He writes: ‘It is close to impossible to develop theories or frameworks for understanding a complex world without imagining alternative causal paths, shifting variables, and different outcomes.’Footnote 70 He concludes: “these essays convincingly remove many of the arguments a historian might make about dismissing counterfactual analysis entirely’.Footnote 71 So even the moderate diplomatic historian ‘is not without scepticism’ vis-à-vis the use of counterfactual reasoning in a US context where it is more accepted than in France. This hostility is one of the reasons why French analysts, strongly influenced by the tradition of diplomatic history, do not adequately engage with the dangers of the Crisis. As Ned Lebow notes, the rejection of counterfactual thinking leads analysts to a historical teleology that retrospectively neglects the role of luck and the validity of other possible worlds that come with it. ‘If the Cuban missile crisis had led to war – conventional or nuclear – historians would have constructed a causal chain leading ineluctably to this outcome.’Footnote 72
Interestingly, in 1964 the French philosopher and strategist Pierre Hassner may have been the first to identify retrospective illusions of control in the few French accounts of the Crisis. He observed that prevailing accounts treated the actual outcome of the Crisis as necessary and were unable to grasp the Crisis itself as an event in the philosophical sense of the term. His contrast between French accounts that assume that the Crisis has settled the matter and the American accounts focusing on surprise, accidents, and pathways to escalation is still valid five decades later.Footnote 73
The lack of legitimacy of counterfactual reasoning and the study of other possible worlds in French IR circles is such that it is not a matter of debate.Footnote 74 This can be explained by disciplinary origins. IR in France borrowed its theoretical foundations from international law and diplomatic history, or what the founding figure Pierre Renouvin would call histoire des relations internationales (history of international relations) which remains based on the need for empirical and archival evidence.Footnote 75 Six decades later, the heir of Renouvin as Professor of the History of International Relations at the Sorbonne wrote: ‘the historian … keeps a specificity. A part of his or her approach is empirical. … To that effect, the work on sources, and in particular on archives, remains at the heart of his or her craft.’Footnote 76
As suggested earlier in the debate between Gavin and Lebow, such empiricism – by treating counterfactual reasoning as an unacceptable scholarly practice – allows for a denial of the need to study the danger of the Crisis. Quite tellingly, Renouvin’s direct successor interprets the outcome of the Crisis as resulting from ‘tacit bargaining’ between the two actors.Footnote 77 As a result, political scientist Jean-Yves Haine revealed the general French attitude towards the missile crisis when, in 2004, he wrote that: ‘One has difficulty grasping the intensity of the showdown.’Footnote 78 While anglophone IR scholars have paid sustained attention to counterfactual thinking and even established ‘what-if history’ as a subfield since the 1990s, this has not at all been the case in France.Footnote 79
Even Vincent Touze, who wrote the most serious study of the Crisis available in French, ends up failing to entertain the possibility that it might have ended differently than it actually did and quickly dismisses attempts at considering possible outcomes as based on ‘hypotheticals’.Footnote 80 He does so despite being well aware of the existing scholarship in English reviewed above. His denial is justified on the grounds of disciplinary boundaries and the impossibility of testing alternative outcomes; to him counterfactual thinking is illegitimate in political science. He writes: ‘Sagan’s goal is not to contribute to the study of foreign policy. His purpose is to contribute to accidents theory. As far as our subject is concerned, his examples are all hypothetical.’Footnote 81 Further, he continues: ‘One can very well conceive of nuclear accidents during the crisis in the US … The effects they could have had belongs so much to speculation that the subject does not seem legitimate.’Footnote 82 If done rigorously, such ‘speculation’ is called counterfactual thinking.
Reasserting control by remembering an idiosyncratic experience of the Crisis exempt of fear
Contrary to the latent assumptions of a common narrative of the danger, causes and consequences of the ‘most dangerous crisis of the nuclear age’, the French experience of the Crisis is idiosyncratically deprived of fear, and this interpretation has been relatively stable over time. The contrast with other nuclear-armed NATO allies and permanent members of Security Council is striking.
Among policymakers, President Kennedy was particularly worried about inadvertent escalation and being forced to use nuclear weaponsFootnote 83 and British Prime Minister Harold MacMillan lost sleep over the dangers of the Crisis; on 4 November, he assessed that ‘everything was at stake…’ in the Crisis.Footnote 84 On the contrary, French President Charles De Gaulle did not voice any particular fear at the time and the French military and intelligentsia do not seem to have felt such a threat, except in press reports for a few days after Kennedy’s speech on 22 October.Footnote 85 A biographer of de Gaulle, for instance, assumes that ‘because no American missiles were based in France, … she would not be an early target, even if it came to a nuclear exchange’,Footnote 86 which is evidence for the ‘great calmness’ of France during the Crisis.Footnote 87 Macmillan confirms this impression when he writes that, faced with the threat of imminent nuclear war, ‘the French were … contemptuous’Footnote 88 while other British Cabinet ministers worried that war was imminent. On the evening of 28 October 1962, Quintin Hogg, Baron Hailsham of St Marylebone, Lord President of the Council, whose wife had just given birth, considered whether to baptise the child himself.Footnote 89 On the contrary, Etienne Burin des Roziers, general secretary of the Elysee Palace, makes fun of the fear of disaster in his diary, which was not meant to be made public before decades.Footnote 90
In the press and the population, a high level of fear was felt for at least a week in the US and the UK. On 2 November 1962, the New Yorker could record: ‘We waited for something to happen, gauging, minute by minute, in something like pain, our ignorance of what the next minute would bring, and feeling the dead weight of the conviction that no one on earth – not the president, not the Russians-knew what it would bring.’Footnote 91 As Alice George summarised it, ‘for many Americans, it represented one week of emprisonment in fear … For many Americans, the clock was ticking loudly, and like men before a firing squad, they waited and wondered whether they would feel anything before the end came.’Footnote 92 The French press, relying mostly on American information, generally only insisted on the risk of war following President Kennedy’s speech, around 24–8 October.Footnote 93 Joelle Beurier’s study of the year 1962 as represented in the French popular photo-journalistic outlet Paris-Match shows that in its pages the violence of the world was dramatised during the first half of the year but euphemised during the second, when the events this article engages with took place. Indeed, the only French source suggesting serious fears of nuclear war came from the French Embassy in Washington, which expressed in a telegram deep worries associated with not having an anti-nuclear shelter.Footnote 94
A few elements of the Crisis appeared on French television: the 8 pm news programme, for example, showed footage of the meeting between President Kennedy and Andreii Gromyko, as well as the confrontation between Adlai Stevenson and Valerian Zorin at the 26 October session of the UN Security Council.Footnote 95 On balance, however, none of this conveyed a significant level of fear or awareness of the danger in the context of the time. A study of the world news segment broadcast in French movie theaters at the time, Regards sur le monde, suggests this absence of fear very strongly. The Crisis does not appear until 31 October 1962 (ten days after the Kennedy speech, and more than two weeks after the beginning of American suspicions about missiles in Cuba). This three-and-a-half-minute segment emphasised control in the voiceover: ‘anything was possible, then the miracle of wisdom and firmness defeated the fatality of unfolding dangers’.Footnote 96 Interestingly, the order in which news items were presented that week did not change to reflect a particularly grave danger: the sport section followed, as usual.
Similarly, the main outlet on military issues in France, the Revue de Défense Nationale, carried a few columns worrying about the increasing arms trade between Cuba and the Soviet Union, but prior to 1963 it published nothing in particular on the Crisis itself. This is all the more surprising as two articles published earlier in 1962 in this same journal discuss the risk of inadvertent escalation.Footnote 97 Prominent French strategist General Pierre-Marie Gallois was already convinced at the time that the situation was safe due to information he received from his contacts in NATO, who confirmed that the NATO military did not feel ‘on the brink’.Footnote 98 Pierre Hassner, then a close collaborator of major French strategist Raymond Aron, emphasised the contrast between the French indifference and the British alarm and anticipation of the end of the world.Footnote 99
While many French intellectuals, in particular Albert Camus and Jean-Paul Sartre, had taken an early stance on nuclear weapons after Hiroshima, the French intelligentsia was conspicuously silent during the Crisis.Footnote 100 Of course, one needs to keep in mind that Camus passed away before the Crisis, in January 1960. But Sartre, André Malraux, Simone de Beauvoir, and prominent Catholic intellectual Francois Mauriac seem to have been completely silent about the Crisis.Footnote 101 None of the journals published by the French intellectual elite in the fall of 1962 touched upon the problem of the Crisis: there is no mention of it in Les temps modernes, Esprit, or La revue des deux mondes.Footnote 102 The only significant discussion appeared in L’Humanité, the daily newspaper of the French Communist Party, in which Sartre and Aragon published a message of support to the Cuban people in the name of association France-Cuba on 25 October. But this statement made no mention of the nuclear threat.Footnote 103
Given the intimate connection with Charles de Gaulle cultivated by a part of the French population, which wrote to the president to share their hopes and fears, it is also quite telling that a search through the correspondence received by the Presidency of the French Republic in October 1962 from private citizens, mayors of French cities, and French clergymen does not reveal a single letter expressing fear or asking for clarification regarding the nuclear situation in Cuba.Footnote 104 Similarly, none of the cards of best wishes for the year 1963 which have been kept mentioned the favourable conclusion of a very risky and scary crisis.Footnote 105 In short, an in-depth study of French sources suggests that the French experience of the Crisis ignored the risk of nuclear war or nuclear weapons accidents and did not display the level of fear expected by security studies scholars who deduce a deterrent effect from the destructive potential of nuclear weapons.
This absence of fear persists in more recent statements by the few French public figures who have expressed an interpretation of the Crisis. In 2014, Ambassador Gabriel Robin, a French official working in the French delegation to the European Economic Community (EEC) in Brussels during the crisis, remembered words of danger but no experience of it: ‘We had the feeling to be on the verge, possibly, of a nuclear danger.’ However, he continued by saying that this was ‘not his business’ since he was in charge of economic relations and it did not disturb his daily work. He concluded: ‘We were saying “this is frightening” but we did not really think about it.’ He similarly remembered that Minister of Foreign Affairs Maurice Couve de Murville came to visit during the Crisis and that he ‘was not distraught at all’.Footnote 106 This is consistent with the telegram Couve sent to all the representatives of France abroad on 18 September 1962, a month before the tensions rose, in which he wrote: ‘even if the idea is constantly suggested, nowhere is it said expressly that were the USA to intervene in Cuba, the Soviet Union would retaliate by triggering a nuclear war.”Footnote 107
The same feeling of indifference and lack of awareness of the danger appears in the two chapters on France and NATO in the 1993 French volume on L’Europe et la crise de Cuba (Europe and the Cuban Crisis) edited by diplomatic historian Maurice Vaïsse. Pierre Gallois writes about ‘American hysteria’Footnote 108 adding that ‘except for pacifist movements, public opinion worried only moderately’.Footnote 109 ‘While housewives hoarded, governments only manifested a limited apprehension.’Footnote 110 Future President Valery Giscard d’Estaing, who as a then 36-year-old finance minister used to meet with General de Gaulle at least once or twice a week – or so he says in an interview in 2000 – goes so far as to say that it was a time of ‘insouciance’ and ‘the happiest moment in his career’.Footnote 111 Apparently, then, he had no memory of any experience of existential fear.
Most interestingly for the argument of this article, no major French intellectual, scholar or strategist has changed his mind concerning the meaning of the Crisis over the following fifty years in spite of the compelling discoveries that have been made: the presence of nuclear weapons in Cuba that was ignored by the Kennedy administration, the debate about whether the Cubans had the authority to fire them, the close calls revealed by the oral history conferences of the 1980s and 1990s, Sagan in 1993 and Dobbs in 2008. Gabriel Robin offered a revisionist interpretation of the causes of the Crisis and its resolution but his book was ignored by scholars as well as policymakers.Footnote 112
Raymond Aron did not live to see the major revelations of the 1990s, but he lent his authority to a persisting narrative of control and underestimation of the danger of the Crisis.Footnote 113 In 1976, he wrote that: ‘the most spectacular episode of the Cold War, the direct showdown of two nuclear-weapons powers … excludes by no means an interpretation consistent with the concepts of classical diplomacy.’Footnote 114 In a 1980 television interview with Dominique Wolton and Jean-Louis Missika, Aron considered that: ‘Even the Cuban crisis of 1962 was not truly a nuclear emergency.’Footnote 115
For most of his life, General Gallois – Aron’s intellectual adversary in the so-called ‘Great Debate’ over French nuclear strategy in the early 1960s – did not think the Crisis worth writing about. For him, the evolution and outcome of the Crisis were just additional evidence that nuclear weapons created a shared rationality – even when dealing with enemies like the Soviet Union. Gallois felt that the Crisis was not threatening because its progressive evolution led to the official alerting of strategic nuclear forces that in turn forced the two great powers to negotiate. ‘Surprise was impossible.’Footnote 116 This has been a consistent interpretation over the years on the part of Gallois, who never considered the Crisis to be a ‘nuclear crisis’.Footnote 117 This is the case even if his archives show that he had been exposed to literature about accidental use of nuclear weapons and inadvertent nuclear war.Footnote 118
A third essential French nuclear strategist, André Beaufre, gave a revealing interview in 1971, in which he explicitly treats nuclear war as impossible because nuclear dynamics are perfectly predictable and controllable. When asked whether he believed in the risk of a total nuclear war, he responded:
at the nuclear level, we are able to … exactly calculate what the war will look like. We were not able to do so with much smaller means such as the pistol, the machine gun, or even the canon. Errors and dreams were allowed. Today, this [sic] is no longer possible. We are working with something which is infinitely calculable and we know ahead of time what the results will be. Men, because they are not crazy, will not do certain things. … Therefore, I do not believe in the great catastrophe nor in atomic death.Footnote 119
Surprisingly, and very tellingly for the argument of this article, this opinion is shared by a longtime opponent of nuclear weapons, General Claude Le Borgne, who described the Crisis as ‘sort of a psychodrama that the two great heads of state, then nuclear beginners, played for their own education’,Footnote 120 adding that ‘the fact that public opinion saw the world on the brink of war is another matter, a moot point’.Footnote 121 In a way, this startling absence of fear during the Crisis has led both the conservative proponents of rational deterrence theory and radical anti-colonialists on the left to embrace a narrative based on the retrospective illusion of safety and control.Footnote 122
Even the official directives from the French Ministry of Education regarding the teaching of this episode in middle and high school perpetuate this retrospective illusion of control compatible with an experience of the event devoid of fear.Footnote 123 Teaching the Crisis is no longer mentioned in the official guidelines the years for which school is still mandatory. It has disappeared from the guidelines for ages 11 to 15 since 2013 and the Crisis does not appear in the list of 43 landmark dates from the third millennium BC to today that French students have to master at the end of mandatory schooling.Footnote 124 For the rest of high school, the recommendations suggest that the Crisis allows the teacher to: ‘insist … on the weight of deterrence and on the will of the players to avoid a direct confrontation’.Footnote 125
Overall, a study of the French experience of the Crisis and discourse about it among scholars, military officers and education officials shows an absence of fear that has been translated over time into a narrative of control. This relative fearlessness contrasts both with implicit assumptions in policy circles of a common understanding of the Crisis and with security studies scholars’ conviction in the dread induced by the destructive capability of nuclear weapons.
Conclusion
Full awareness of the limits of knowledge of and control over nuclear weapons is crucial for historical accuracy, for nuclear learning, and as a starting for a fruitful nuclear weapons policy debate that would include strategic, ethical, and political concerns.
This awareness is all the more important as overconfidence has been shown to be a cause of increased danger. Therefore, learning from the Cuban Missile Crisis is essential given that, over the last thirty years, analysts have discovered and confirmed that it was one of the most dangerous events in the history of the nuclear age, the peaceful outcome of which was partly due to luck, and that in the preceding three decades they had been overconfident in the ability of good management to explain its peaceful outcome. In order to understand the construction of such overconfidence, I have accepted the idea that learning occurred mostly at the national level and focused on the French case, in comparative perspective. I have first shown how those scholarly findings have not been adequately refuted and argued that those who do not take these elements seriously have done so via epistemological or practical inconsistencies. The rejection of counterfactual thinking as a legitimate scholarly practice is another way of rendering these findings invisible without having to refute them. Finally, I have used the French example to show that an experience and official memory of the Crisis that is not based on fear does over time fuel overconfidence in the safety, controllability, and predictability of nuclear crises. American and British elites and populations did not grasp the full extent of the dangers at the time but, unlike the French, they certainly experienced fear. These ideational and disciplinary factors would be sufficient to prevent the problem from emerging as an issue of public concern in France; they also call for responsible nuclear scholarship to address them without waiting for structural or policy change. This opens up three avenues for research.
First, social scientists cannot let Fidel Castro take the unbearable lightness of luck with him to the grave. Following from the efforts of cognitive psychologists to uncover our tendencies to deny luck retrospectively, further exploration of the politics of luck and how the distinction between risk and uncertainty (as uncontrollability and unknowability even of the boundaries of the possible) has been blurred would be a first critical step towards a reconceptualisation of nuclear controllability, a reconceptualisation that would place luck at the heart of political and ethical action, power and responsibility over time.Footnote 126 At the empirical level, this approach would involve treating the question of ‘how close did we come to nuclear disaster’ as a starting point, and focusing on pathways towards disasters rather than patterns identified in the past and expected to bind possible futures.
Second, cases of near-nuclear use need to be requalified as events worth investigating.Footnote 127 Diplomatic historians and security studies scholars could fruitfully join forces independently of their views on the value of counterfactuals. Indeed, we need further investigations into the history of nuclear-armed states both to unearth primary documentation about the past security and safety record of nuclear arsenals and to allow for rigorous counterfactual thinking, going beyond risk thinking.Footnote 128 Comparative critical oral histories of nuclear close calls would likewise help recover the limits of control over nuclear weapons as a legitimate object of scholarly interest. Such an approach would also tackle directly the scholarly problem of uncritical reliance on accounts from former officials, while at the same time addressing the policy problems that result from misguided assumptions of a shared experience and interpretation of events like the Crisis.Footnote 129 The empirical dimension of this article is only a first step in this direction.Footnote 130 This research programme allows analysts to start working against overconfidence without waiting for structural or institutional reforms and suggests that they have a responsibility to do so.
Third, in security studies, it is crucial to reassert the socially and historically constructed process of valuation of nuclear weapons instead of perpetuating the widespread assumption that the destructive capability of nuclear weapons triggers adequate fear, which in turn initiates a learning process that is sufficient for existential deterrence to work everywhere. If the relevant French decision-makers were indeed adequately afraid at the time but did not leave any evidence of it, the contentions of this article would remain valid: that absence of evidence would only be additional evidence of the scholarly blinders entailed by a rejection of counterfactual thinking; in any case, and whatever the unspoken thoughts of decision-makers at the time, their public display of fearlessness would be consequential for future generations of leaders. Evidence of private doubts on the part of statesmen would only bring to the fore the need to reconnect nuclear weapons scholarship with democratic theorising and the issue of citizens’ right to know. Identifying the effects of rejecting counterfactual thinking and documenting the diverse experiences and memorialisations of nuclear danger as components of a process of valuation of nuclear weapons are ways to understand and counter overconfidence in their controllability. This is crucial for policy-relevant learning because the coexistence of this diversity of memories with the retrospective illusions of unanimity and control gives an unsettling resonance to Peter Sloterdijk’s claim: ‘the only catastrophe that seems clear to all is the catastrophe which no one survives’.Footnote 131
Biographical information
Benoît Pelopidas holds the junior chair of excellence in security studies at Sciences Po. He is also an affiliate of the Center for International Security and Cooperation (CISAC) at Stanford University and a visiting fellow at Princeton University’s Programme on Science and Global Security.
Acknowledgements
I want to thank Philippe Gallois, Pierre Hassner, General Claude Le Borgne, Ambassador Gabriel Robin and Maurice Vaïsse for the time they gave me early on in this research. M. Gallois deserves particular thanks for sharing documents with me. I also want to thank James Cameron, Lynn Eden, Nicolas Guilhot, David Holloway, Jacques Hymans, Venance Journé, Garret Martin, Leopoldo Nuti, Laura Stanley and the participants in seminars at the Woodrow Wilson Center for Scholars, CISAC, Stanford University, La Sorbonne Nouvelle and the Nuclear Insecurities Working Group at the University of Bristol. Grey Anderson, Barton Bernstein, Sebastien Philippe, and Jutta Weldes deserve special thanks for their scrupulous and precious feedback over several drafts … over several years. I am grateful to Georges Le Guelte who has been supportive of my work since 2009. Funding for this research was provided by a British Academy’s Newton Advanced Fellowship for a project with Professor Carlo Patti on ‘Global Nuclear Vulnerability: the effects of the Cuban Missile Crisis on British, French and Brazilian nuclear policies’, the Nuclear Proliferation International History Project at the Woodrow Wilson Centre for Scholars, and small travel grants from the Global Insecurities Centre at the University of Bristol. A University Research Fellowship from the Institute of Advanced Studies from the University of Bristol gave me time to bring this research to completion; the global systemic risk research community and the programme on science and global security (SGS) at Princeton University provided a home to complete it. Finally, the three anonymous reviewers of EJIS deserve special thanks for helping me improve the article significantly.
