This October is the 50th anniversary of the Cuban missile crisis, when the United States and Soviet Union stared each other down after Soviet nuclear missiles were discovered on the island of Cuba.

Much has emerged about these events as successive troves of information have been uncovered — including records from the Soviet secret service and ruling Presidium1, and from the US Department of State and Navy2. Perhaps the most remarkable are the hours of audio recordings that US President John F. Kennedy made secretly of the deliberations of the Executive Committee of the National Security Council (ExComm), with which he met almost daily during the height of the crisis, between 16 and 28 October 1962.

The existence of the ExComm recordings — more than 20 hours' worth from these 13 days alone — was known only to Kennedy's close associates until they came to light during the Watergate investigations in 1973. The tapes were eventually declassified and, with a few excisions, released to the public in the mid-1990s. Historians have looked to them and the transcripts3 for insight into the positions taken by the dozen or so ExComm members, including Robert Kennedy, the attorney general and president's brother, and Robert McNamara, the Secretary of Defense at the time4.

President John F. Kennedy meets the US Executive Committee of the National Security Council in October 1962 to discuss Soviet nuclear missiles in Cuba. Credit: CORBIS

I am the first to subject the recordings to scientific scrutiny, using the tools of conversation analysis. This subfield of sociology starts with the meticulous transcription of conversation — with all its interruptions, slips and hesitations. Assuming that everything that happens does so for a reason, it asks why, when and with what consequences people do things such as interrupt, restart sentences and repeat others' words. It also helps to shed light on why particular discussions unfold as they do, from conversations about mundane matters, such as about what to have for dinner, to talk of greater moment.

The Cuban missile crisis is often held up as a model of rational decision-making. In many accounts, the only decision maker that mattered was Kennedy, and the ExComm's contribution is minimized. My study of these recordings, however, suggests that Kennedy was susceptible to persuasion, which was shaped by conversational vicissitudes and exigencies, yet also motivated and constrained by the pace of events in the outside world. My analysis challenges familiar ideas: that Kennedy steadfastly resisted calls for military action; that once he expressed a preference his advisers promptly fell in line; that his decisions were the obvious ones given the objective logic of the geopolitical situation; and that the ExComm's consideration of the options was thorough and balanced. In fact, the recordings reveal a president clearly swayed by the advice he received — and the manner in which he received it.

Time to talk

The president learned of the missiles on 16 October (see 'Cuban missile crisis timeline'). His first major decision was to impose a naval blockade on Cuba, which he announced to the world on 22 October after revealing the discovery of the missiles. According to most accounts, President Kennedy was drawn to the idea of a blockade because it seemed like a strong action that did not involve an immediate military attack. Certainly the president was not eager to begin dropping bombs, but the blockade option itself carried serious risks. No one thought it would be sufficient to force Soviet Premier Nikita Khrushchev to withdraw the missiles already on the island, which meant that an air strike might eventually be needed to destroy them. In the interim, the Soviets would have time to finish work on the missiles, some of which could be fired at US cities — perhaps by accident or without authorization — in the event that the United States attacked.

How, then, did Kennedy come to choose a blockade? I identified and examined every instance of talk about a blockade and subsequent air strike in the tapes to see whether the danger of a missile launch was reiterated, omitted or anticipated but unsuccessfully articulated. I found a marked shift, early on 18 October, from frequent warnings, particularly from McNamara, about the danger of bombing operational missiles on the island, to a string of statements (the first from Llewellyn Thompson, former US ambassador to the Soviet Union) supporting the blockade and air strike but omitting any reference to this danger.

Credit: BETTMAN/CORBIS; HULTON-DEUTSCH COLLECTION/CORBIS

After this shift, whenever someone tries to reintroduce the risk of a nuclear response, that person is interrupted, talked over or ignored. Twice, for instance, Robert Kennedy tries to warn about the danger of allowing the Soviets time to finish work on the missiles before a US attack, but both times he is thwarted. The first time, Maxwell Taylor, Chairman of the Joint Chiefs of Staff, talks over him. The second time, McNamara repeatedly interrupts him before he can finish his thought.

I believe such efforts at 'suppression' reflected a growing determination to reach consensus on some course of action when the best course remained elusive, and to combat attempts to threaten this consensus. This in turn allowed President Kennedy to make a choice, which he did on the 20 October, that he could justify in light of recent talk — even if it flew in the face of much of what had been said earlier.

This consensus was possible because, for the first week of the crisis, the United States kept its knowledge of the missiles a closely guarded secret, so the ExComm could revisit the blockade option again and again until objections were squelched. When all options look bad, a decision may hinge on the ability of groups to talk in circles until such a time as the more frightening consequences of one course of action disappear from view.

Overtaken by events

Kennedy's second main decision concerned the implementation of the blockade. Soon after the president made his public announcement on 22 October, Khrushchev ordered Soviet ships carrying weapons to be diverted away from Cuba2 (although a warhead-laden ship called the Alexsandrovsk slipped into port in Cuba before the blockade was put into effect). Other ships were allowed to continue on their course. Although Kennedy believed that these ships were not carrying missiles, some felt that an interception would be needed so that Khrushchev would not doubt US resolve.

Four Soviet-bloc ships were put forward as candidates. Two (the Kimovsk and Grozny) were ordered by Khrushchev to turn back before they could be intercepted. Another (the East German Völkerfreundschaft) was a passenger ship that no one had the stomach to challenge for fear of loss of life. The only serious candidate for interception was the Bucharest, a Soviet tanker that, lacking deck cargo, was almost certainly carrying only oil. The ship turned up at the blockade line early on the morning of 25 October and, before it could be discussed by the ExComm, was permitted to pass after it declared its cargo benign. Thus, the ExComm meeting later that morning was, in part, about whether to chase it down and board it belatedly.

The standard story is that the president let the Bucharest go because he wanted to give Khrushchev more time to back down, or to respond to diplomatic overtures at the United Nations5.

In fact, the tapes reveal a Kennedy who is very worried about appearing weak if he does not order the ship to be boarded. His hand is stayed by opposition from his more-vocal advisers. They assure him that the failure to intercept could easily be justified to domestic and international audiences, even as they thwart attempts to give sustained consideration to the way that the Soviets might interpret the move. Kennedy remains unconvinced. By the end of the meeting he still seems inclined to board the ship, but postpones a decision until later in the day.

The same process could have had a different ending.

Before the ExComm reconvened, however, a state-department briefing for legislators leaked the fact that the Bucharest had already crossed the blockade. Rather than admit indecision, the Pentagon announced that the navy had let it pass after ascertaining that it carried no contraband. When the ExComm reconvened later in the afternoon, it took the passage of the tanker as irreversible — showing how indecision can be transformed into decision by the pace of events.

Kennedy's third big decision concerned the deal that resolved the crisis. Late on 26 October, he received a private letter in which Premier Khrushchev offered to withdraw his missiles in return for a US pledge not to invade the island. Before this could be discussed by the ExComm the next morning, the group received another offer, this one made publicly, demanding that President Kennedy remove the North Atlantic Treaty Organization's nuclear (Jupiter) missiles from Turkey. The ExComm was dumbfounded, but Kennedy's advisers quickly converged to recommend that the president accept the first offer and ignore the second. The president resisted, however, insisting that Khrushchev would never be content with such a deal, having publicly proposed a missile trade.

The group reached an impasse on the afternoon of 27 October. Neither side could muster an argument that the other side would accept, even though everyone felt pressured to respond to Khrushchev's offer before the end of the day, partly for fear that if there was any delay he would issue additional demands.

Aerial shots of Soviet missile launch sites in Cuba provoked a naval blockade by the United States. Credit: BETTMANN/CORBIS

On the tapes, President Kennedy's advisers try, in various ways, to work around their boss. Twice, for instance, the president is allowed to talk on at length about the futility of accepting Khrushchev's first offer, receiving no challenge despite providing openings (such as pauses) for disagreement.

Several times people even suggest that President Kennedy remove himself from the letter-drafting process. Robert Kennedy does so most forthrightly. This results in an explosion of laughter at his audacity, and at the way he is pointing out the elephant in the room — that the president has become the primary impediment to consensus as to what the president should do.

Finding little support for the position he has taken, President Kennedy eventually relents and agrees to the wording of a letter that accepts Khrushchev's first offer, even though it hints at later negotiations over “other armaments”.

Yet Kennedy remained unconvinced that this would be enough. Following a more intimate meeting off-tape on that night, he deployed his brother to secretly deliver a promise that the Jupiter missiles would be removed from Turkey within a few months of the end of the crisis, on the condition that Khrushchev did not publicize that part of the deal.

Khrushchev accepted, having been about to agree to the terms of Kennedy's letter when he received word of the further concession2. This marked the beginning of the end of the crisis. Although there followed several months of tense negotiations about inspections and Soviet nuclear bombers1, by Christmas all the offending weapons had been removed from Cuba.

Not so rational

The Cuban missile crisis had a good ending, and the ExComm deliberations deserve some credit. But this was not a rational decision-making process in any traditional sense. The discussions were undoubtedly beneficial in bringing to light the various options and their associated risks, and in forcing the different factions (the 'hawks' and 'doves') to hone their arguments. But they were also structured by conversational machinery — based, for instance, on the rules of turn-taking and the expectation that one say something relevant to whatever was said last — that is not easily put to the service of comparing the consequences of competing courses of action.

In fact, the need to reach consensus at each stage required the ExComm to avoid, or cease, consideration of some of the risks: the risk of having to bomb operational missiles if the blockade failed; the risk that letting the Bucharest past would leave Khrushchev with the impression that Kennedy was weak; and the risk that by accepting Khrushchev's first offer and ignoring his second, the first real path out of the crisis (like-for-like missile withdrawal) would be sacrificed.

The same process could have had a different ending. Had a consensus in favour of the blockade been forged even a day sooner, for instance, it might have gone into effect in time for the Navy to intercept the Aleksandrovsk. Given that this ship was under orders to sink itself rather than be boarded2, that would have sent history careering down a different path entirely.

The lesson for scientists, advisers and policy-makers is that the details and mechanics of conversation matter. Talk is useful for decision-making, but its conventions do not ensure that sustained attention is given to all the things that could go wrong. Given enough time, all branches of the decision tree might receive their due, but during a crisis time is in short supply, and the hardest decisions might require that some branches are neglected and even wilfully abandoned.