Xi Jinping Is Not Stalin
How a Lazy Historical Analogy Derailed Washington’s China Strategy
To recall the atmosphere of September and October 1962 now seems almost as difficult as to recreate the weeks, more than two decades earlier, before the attack on Pearl Harbor. But if we are to understand the onset of the Cuban missile crisis, it is worth the effort. Indeed we may learn something about the problems of foreseeing and forestalling or, at any rate, diminishing the severity of such crises by examining side by side the preludes to both these major turning points in American history. In juxtaposing these temporally separate events, our interest is in understanding rather than in drama. We would like to know not only how we felt, but what we did and what we might have done, and in particular what we knew or what we could have known before each crisis.
Afterthoughts come naturally following the first wave of relief and jubilation at having weathered the missile crisis and forced the withdrawal of the missiles. But it is good to keep in mind the obvious contrast with Pearl Harbor. At the least, Pearl Harbor was a catastrophe, a great failure of warning and decision. At the very worst, the missile crisis was a narrow escape. Taken as a whole, however, its outcome must be counted as a success both for the intelligence community and the decision-makers. But a comparison of the failure at Pearl Harbor and the Cuban success reveals a good deal about the basic uncertainties affecting the success and failure of intelligence.
It is true for both Pearl Harbor and Cuba that we had lots of information about the approaching crisis. In discussing this information it will perhaps be useful to distinguish again between signals and noise. By the "signal" of an action is meant a sign, a clue, a piece of evidence that points to the action or to an adversary's intention to undertake it, and by "noise" is meant the background of irrelevant or inconsistent signals, signs pointing in the wrong directions, that tend always to obscure the signs pointing the right way. Pearl Harbor, looked at closely and objectively, shows how hard it is to hear a signal against the prevailing noise, in particular when you are listening for the wrong signal, and even when you have a wealth of information. (Or perhaps especially then. There are clearly cases when riches can be embarrassing.)
After the event, of course, we know: like the detective-story reader who turns to the last page first, we find it easy to pick out the clues. And a close look at the historiography of Pearl Harbor suggests that in most accounts, memories of the noise and background confusion have faded quickly, leaving the actual signals of the crisis standing out in bold relief, stark and preternaturally clear.
After the crisis, memories fade and recriminations take their place. For a time the Cuban missile crisis figured as an outstanding triumph for the United States-in the swift discovery of "hard evidence," in the retention of American initiative, in the strict security maintained and in the taut control of power by the Executive Committee. Today, some of these aspects of the Cuban crisis have been thrown into doubt, and in particular, critics talk of a significant intelligence failure in anticipating the crisis. In both Pearl Harbor and Cuba the notion of a conspiracy of silence has been raised, the suggestion that we knew all along and failed to act, that Kennedy, like Roosevelt, had some special information which he withheld, or that information was so obvious that even a layman could have interpreted it correctly.
New York's Senator Keating, for example, was explicit and articulate in insisting that he believed long-range or medium-range missiles and Soviet combat troops were in Cuba as early as August. On August 31 he said in the Senate that he had reliable information on landings between August 3 and August 15 at the Cuban port of Mariel of 1200 troops wearing Soviet fatigue uniforms. He also reported that "other observers" had noted "Soviet motor convoys moving on Cuban roads in military formation," the presence of landing craft, and of suspicious cylindrical objects that had to be transported on two flatcars, and so on. He claimed that his statements had been verified by official sources within the U.S. Government. Between August 31 and October 12 he made ten Senate speeches warning of the Soviet military build-up.
After the crisis, Congressmen naturally wondered why we had not listened to Senator Keating, why it was possible to have had these warnings and many others and still be surprised on October 15. But failures to foresee and to forestall catastrophes are by no means abnormal. Military men and statesmen have no monopoly on being taken by surprise. The example of the Dallas police department springs to mind, and the murder of Oswald which gave rise, like Pearl Harbor, to rumors of conspiracy in high places and in local governments. Nor are American businessmen and financiers immune. Witness the $150 million De Angelis vegetable-oil scandal, where normally cautious bankers suddenly found they were holding empty storage tanks as security for their loans.
Conspiracy with the culprit, however, is hardly a universal line of explanation, as is suggested by a recent natural catastrophe-the earthquake in and near Alaska that sent a tidal wave to shatter the northern shore of California and caught some towns unprepared in spite of timely warnings. For the warnings sounded just like many others in the past that had not been followed by tidal waves. These are all American examples, but Singapore, "Barbarossa" (the German attack on Russia) and many others suggest that we are not dealing with a purely national susceptibility to surprise.
Defense departments and intelligence agencies, of course, continually estimate what an opponent can do, may do, intends to do. They try to gauge the technical limits within which he is operating, to determine his usual ways of behavior, under what conditions he will probe, push or withdraw. They try to measure what risks he will take, and how he might estimate the risks to us of countering him. Much of this work by American analysts is sound, thorough, intelligent, frequently ingenious and sometimes brilliant- but not infallible. Unhappily, any of these estimates may be partly, but critically, wrong. A wealth of information is never enough.
To get a rapid idea of the mass of data available for predicting the Cuban crisis and the Pearl Harbor attack, let us run through the main intelligence sources. In the case of Cuba, there was first of all magnificent photographic coverage as well as visual reconnaissance. The Navy ran air reconnaissance of all ships going in and out of Cuba, especially ships originating in Soviet or satellite ports during the summer of 1962, and intensified this sort of coverage during September. High-level photographic reconnaissance by U-2s over the island of Cuba was taking place at the rate of one flight every two weeks until the month of September, when it increased to once a week.[i] Low-level photographic reconnaissance began only after the President's speech of October 22-the first being on October 23. In addition to photography, we had voluminous accounts from Cuban refugees who were leaving the island in a steady stream. We had agents stationed on the island who were reporting, and we were listening to radio broadcasts from Cuba. The Cuban press, while carefully controlled, was making some announcements which are interesting in retrospect. A number of European correspondents stationed on the island were reporting to their newspapers, though the American press was not welcome.
Finally, but by no means least, we had Castro's pronouncements. His casual interviews with reporters, debates with students, interrogations of prisoners, and nearly interminable television speeches offer a rich fount of information. If you wait long enough, it seems, Castro will tell you everything. The only problem in a crisis is that you may not be able to wait that long. Castro is noted for his slyness, and he is perhaps better able than most Cubans to keep a secret. But sometimes he cannot resist hints that may reveal a trap before his victim falls into it. And often in real rather than calculated anger he will show his hand.
For predicting the Pearl Harbor attack, the United States Government had an equally impressive array of intelligence sources. Though aerial surveillance of the Japanese fleet was limited, the Navy had developed a system of pinpointing the location of ships and deducing their types by radio-traffic analysis. This was accomplished by analyzing the call signs of various ships, even though we could not read the content of the messages. Any change in call signs was in itself a cause for alarm, and it took usually several weeks of close listening to an enormous amount of traffic to re-identify the call signs. Call signs were changed on November 1, 1941, and again on December 1. We had not identified the new ones by December 7.
While we had not broken any military codes, we did have one superlative source that is perhaps comparable to the evidence provided by U-2 photography. That was the breaking of the top-priority Japanese diplomatic code, known as MAGIC, as well as some less complicated codes used by Japanese consular observers. We were listening in on diplomatic messages on all the major Tokyo circuits-to Rome, Berlin, London, Washington and so on. Colonel Friedman, an Army cryptographer, had devised a machine for rapidly decoding these messages, so that, in general, we knew what a message said before its intended Japanese recipients. Our ground observers, stationed in key ports along the coast of China and Southeast Asia, were reporting in by radio.
Ambassador Grew and his Embassy staff in Tokyo were experienced observers of local economic and political activities. Grew himself had a very sound estimate of Japanese character and diplomacy, but as Japanese censorship closed in during the last few weeks before the attack, Grew had to warn Washington that he was unable to report accurately on any military preparations then under way. American newspaper correspondents in Japan were also quite well informed and shrewd in their reporting. In addition to our own sources, we exchanged information with British intelligence. At that date, our own intelligence officers did not trust British intelligence fully. They expressed a certain amount of unease over British methods of picking up information, which they regarded as sophisticated but underhanded. As General Sherman Miles put it, U. S. intelligence preferred to be "above board." However, the British provided us with some good leads and lots of corroborative information. And there was, of course, the Japanese press, which proclaimed Japan's undying hostility to the American presence in Asia, and announced with increasing violence the Japanese intention to expand to the south.
In sum, for each of the two crises there was plenty of information suggesting its advent. Even though Cuba is a closed society, and even though Japan, in the last weeks, was under heavy censorship and tight security, the data provided by U. S. intelligence agencies were excellent. Once more, then, we come to the question, what went wrong? With all these data, why didn't we know that Japan would attack Pearl Harbor on December 7? Why, when it seems so clear in retrospect, didn't we anticipate that Khrushchev might put medium-range missiles into Cuba? Why didn't we seize the first indications that such installations were on the way? Weren't these early signs clear enough?
Unfortunately, they were not, and almost never are. Even with hindsight, we are not able to reconstruct the exact sequence of events that led to the Cuban missile crisis. Most of our sources are alive, and some of them are talking. But what can we say with certainty about Cuban and Soviet motives? Castro, for example, has spoken on many occasions about why missiles were put into Cuba. But he swings between the view that he requested them and the view that Khrushchev suggested the idea and that he, Castro, felt so indebted economically he had to accept. He has mentioned two motives-one, defense against an American invasion that he believed was imminent, and the other, the need to advance the international cause of socialism, which implied that the missiles were for offense as well as defense. Khrushchev's story is more consistent, but also more "official": he cites only the need to help Cuba prepare against an American invasion. But of course for active Cuban defense, long-range missiles are not necessary. Speculation on Soviet and Cuban motives still continues.
With hindsight, we can look back now and see that during the crisis there were naturally many confusions embedded in the mass of intelligence reports. A report of a "missile" might refer to a surface-to-air missile which is approximately 30 feet long, to the nose cone of a surface-to- surface missile which is about 14 feet long, to its body which is almost 60 feet long, or to a fuel storage tank. Or perhaps it might just represent the imagination of an excited Cuban refugee. Most of these objects were seen at night through closed shutters and in motion. Visual observation, except by a highly trained observer, was not likely to be accurate even as to the length of the object. And Senator Keating did not act altogether responsibly in perpetuating this confusion centering around the word "missile." He was right when he described the total build-up as alarming, but he was proceeding beyond the evidence in suggesting, as he did, that he had positive proof of the presence of medium-range missiles,[ii] and of the capability for rapid transformation of surface-to-air missiles into medium- range surface-to-surface missiles.
Or take the presence of Soviet combat troops. President Kennedy's critics noted after the crisis that in his October 22 speech he made no mention of combat troops in Cuba, although the American public was later informed of their presence. Actually, Soviet troops, organized into four regimental units, totaled approximately 5,000 men. They were located at four different spots, two near Havana, one in Central Cuba and one in Eastern Cuba. They were equipped with modern Soviet ground-force fighting equipment, including battlefield rocket launchers similar to the American "Honest John." This equipment, along with the accompanying barracks and tent installations, was not identifiable, or at least was not identified, until we started photographing at low level. For this reason, President Kennedy made no demand about removal of troops on October 22, but kept to the colorless term, "Soviet technicians." While U-2 photography is almost as magical as the MAGIC code at the time of Pearl Harbor, like the code, it is limited; it cannot reveal all.
For the layman, the feeling persists that there must be some marvelous source that will provide a single signal, a clear tip-off that will alert the American forces and tell them exactly what to do. Unfortunately, there is no instance where such a tip-off arrived in time, except perhaps in the Philippines in 1941, when General MacArthur had a minimum of nine hours' warning between his knowledge of the Pearl Harbor attack and the initial Japanese assault on his own forces. The news of the attack on Pearl Harbor clearly did not tell him what alert posture to take, since his planes were found by the Japanese attackers in formation, wing-tip to wing-tip on their bases.
Instead we must wait for a number of signals to converge in the formation of a single hypothesis about the intentions and actions of an opponent. This is a necessary but slow process. In 1962, for example, General Carroll, head of the Defense Intelligence Agency, became suspicious of Soviet activities on the basis of several pieces of data from different sources. According to Secretary McNamara's testimony,
. . . [Carroll] had had thousands of reports like this. What gradually formed in his mind was a hypothesis based on the integration of three or four pieces of evidence, one of which was not a report at all, one of which was a recognition through photographic analysis that a SAM (surface-to-air missile) site appeared to be in a rather unusual place. . . . Gradually over a period of time-I do not know over what period of time-but sometime between the 18th of September and the 14th of October, there was formulated in his mind a hypothesis specifically that there was the possibility of a Soviet ballistic missile installation in a particular area, a hypothesis that had been formulated previously and had been tested previously and found to be in error with respect to other locations.
His only action here-I think quite properly his only action here-was to test that hypothesis, to submit it to the targeting group that targets the reconnaissance missions, and place that target on the track for the next reconnaissance mission, which was the October 14 mission.[iii]
This period of time from September 18 to October 14 is not long for the crystallization of a hypothesis.[iv] It is long only in relation to the speed of the missile installation. This sort of time difference is a perpetually agonizing aspect of intelligence interpretation. Collection, checking of sources and interpreting all take time. There is always delay between the intelligence source and the evaluation center, and between the center and the final report to the decision-maker. Even then, the decision- maker may merely request more information before taking action. In the meantime, the opponent moves forward.
In the Cuban missile crisis, for example, there were delays in the identification of surface-to-air missiles. From July 29 to August 5, Cuban refugees reported that "an unusual number of ships" unloaded cargo and passengers at the ports of Havana and Mariel. All Cubans were excluded from the dock. By August 14 these reports reached U. S. intelligence agencies, which the next day requested U-2 photo coverage of the suspect areas. On August 29 the flight was made. From the first visual observation on July 29 to the overflight on August 29 a full month passed.
This August 29 flight turned up the first hard evidence of surface-to-air missiles in Cuba. During September, surveillance flights seem to have been stepped up: the U-2 flew on September 5, 17, 26, 29, and on October 5, 7 and 14. On the September 5 flight, which took in the San Cristobal area a hundred miles east of Havana, the photographs showed no evidence of medium- range missiles. A flight scheduled for September 10 was canceled, perhaps because a U-2 had been shot down over Red China the previous day. According to the American press, all U-2 flights stopped while the United States waited for the world reaction.
Secretary McNamara testified that available evidence indicated the first landing of mobile M.R.B.M.s occurred on September 8, and that construction of the sites did not begin before September 15 to 20. It is possible that September 10 photography might have shown some activity at the San Cristobal site. The September 17 flight was of little use because cloud cover obscured the areas photographed. However, between September 18 and 21 further Cuban reports came to U. S. intelligence, and these were evaluated on September 27. They eventually led to the flight on October 14, again over San Cristobal. This flight produced the first reliable evidence of medium-range missiles on the island.
In spite of the frequency of the U-2 flights, there is a lag of 33 days from the first visual observation made by a Cuban exile on September 8, and reported on September 9, to October 14, the day that hard evidence was obtained. There is a lag of 39 days between September 5 and October 14, during which no flights covered the San Cristobal area. This gap in coverage was not apparent until some inquiring Congressmen pressed their cross-examination. When William Minshall of Ohio asserted that the U-2 flights had been covering the wrong end of the island, General Carroll pointed out that it was necessary to cover the eastern and central portions also. Secretary McNamara supported him by pointing out that the September 5 flight over San Cristobal "showed absolutely no activity whatsoever." He also recalled that this was the hurricane season, "and the weather in that part of the Caribbean is very bad. We had a number of flights canceled during that period." Mr. Minshall then produced the official weather report showing clear days in the vicinity of Havana, and said that "the weather from September 25 to October 2, at least at 7:00 in the morning, was generally clear." No one pointed out at that time that weather forecasts, not actual weather, determined the schedule of U-2 flights.
Photographic coverage, then, was apparently being scheduled on the assumption that any Soviet construction would proceed at a pace which might be considered rapid according to our own experience in installing similar equipment. Secretary McNamara repeated several times that there was no missile construction activity in the Havana area on September 5, as if this, coupled with the pressing need to get clear pictures of other parts of the island, were sufficient reason for not covering the area again until October 14. This judgment, with hindsight, may have been correct, but in the absence of the full intelligence picture the layman can only wonder why it was not possible to cover more than one section of the island on a single U-2 sortie, or why it was not possible to make several simultaneous sorties when good weather prevailed. Perhaps Secretary McNamara's statement, made under pressure of Mr. Minshall's criticism, to the effect that "we were facing surface-to-air missile systems that might be coming into operation," indicates that the flight schedule was sensitive to the political atmosphere. The fact is that there were increasing dangers to our pilots as the SAM sites became operational. With the Republicans now in opposition, it was easy for some of them to forget the extreme embarrassment of the Eisenhower régime at the shooting down of the U-2 over the Soviet Union in 1960 and the collapse of the Paris summit that followed. Certainly after the publicity given to the U-2 shot down over Red China on September 9, the United States would not want to lose such a plane over Cuba. U-2 planes are never armed; and the August 29 flight had showed surface-to-air missile installations in western Cuba.
Naval photography shows a somewhat similar gap. Photographs of the crates containing IL-28 bombers were taken on September 28 but not evaluated until October 9, and not disseminated until October 10. This identification of bombers capable of carrying a nuclear or non-nuclear payload of 6,000 pounds and with a combat radius of about 700 nautical miles[v] came together with a report of October 15 evaluating the U-2 photographs of M.R.B.M.s.
This sort of delay can easily be paralleled in the Pearl Harbor intelligence picture. In the handling of the coded messages, there was inevitably a delay-from interception of the message at the intercept station through transmission to the decoding center in Washington, determination of priority in handling, assignment for full decoding, assignment for translation and the actual translation, to final delivery to the approved list of recipients. The longest delay recorded in the Congressional hearings is 54 days between interception and translation. Part of the delay is a function of the time necessary for transmission. Part of the delay comes from checking the accuracy of the reports, which is necessary for responsible decision. But these delays in response must all be seen against the forward march of events.
In Cuba, the rapidity of the Russians' installation was in effect a logistical surprise comparable to the technological surprise at the time of Pearl Harbor. Before September 1962 we were scheduling U-2 flights approximately two weeks apart, because we couldn't believe that capabilities could change significantly within a shorter period. But Secretary McNamara testified in his first background briefing (October 22) that the medium-range mobile missiles were planned to have a capability to be de-activated, moved, reactivated on a new site and ready for operation within a period of about six days. The Stennis Report, which reviewed the entire intelligence operation, refers to "a matter of hours."[vi] In one instance, between two sets of photographs separated by less than 24 hours, there was an increase of 50 percent in the amount of equipment visible. On the date of withdrawal, October 28, the medium-range missiles were fully operational. Intelligence estimates set December 15 as the outside date for the non-mobile I.R.B.M.s to be operational.
This kind of technological or logistical surprise may be either a secret so carefully guarded that it doesn't reach our intelligence agencies until after the event; or it may happen too swiftly, too near the outbreak of the crisis, to be transmitted and evaluated in time. In the case of Pearl Harbor, there were two technological changes that failed to reach either the intelligence agencies or the commanding officers who needed the information: (1) that the Japanese had fitted fins to their torpedoes which would permit bombing in the shallow waters of Pearl Harbor; and (2) that the combat radius of the Zero fighter plane had been stretched to 500 statute miles, making possible aerial attack on the Philippines from Formosa. Both of these developments came to fruition only a few weeks before Pearl Harbor.
Besides technological surprise and the inevitable physical delays involved in transmission and checking, there are more subtle obstacles to accurate perception of signals. First, there is the "cry-wolf" phenomenon. Admiral Stark actually used this phrase in deciding not to send Admiral Kimmel any further warnings about the Japanese. An excess of warnings which turn out to be false alarms always induces a kind of fatigue, a lessening of sensitivity. Admiral Kimmel and his staff were tired of checking out Japanese submarine reports in the vicinity of Pearl Harbor. In the week preceding the attack they had checked out seven, all of which were false.
General Carroll had the same problem with missiles in Cuba. Refugee reports of missiles had been coming in for a year and a half and the first San Cristobal report of September 9 describing that suspect area, later confirmed as harboring medium-range missiles, was "comparable to many other reports . . . similarly received and checked out," and found to reveal not surface-to-surface missiles, but surface-to-air or nothing at all. This history of mistaken observations by the refugees tended to reinforce the feelings of fatigue and disbelief. There was also a justifiable reaction to the fact that refugee exaggerations of anti-Castro ferment in Cuba had not been properly discounted at the time of the Bay of Pigs, and that their self-interest in wanting to return to Cuba had not been properly weighed. This background increased the reluctance of the intelligence agencies to credit their reports without careful verification. Besides the refugees, members of the Congressional opposition were also using exaggeration and pressure, because they had an interest in overstating provocation in order to indicate laxness on the Administration's part. Senator Keating claimed to have hard evidence at a time when, it seems, such evidence did not exist. Opposition pressure tended to evoke a natural counter-pressure from the Administration, which responded by charging irresponsibility in its critics, and which insisted on caution and the necessity for special evidence before entering on such serious action. In this way the opposition served in some respects as rein rather than simply as spur.
Another obstacle to objective evaluation is the human tendency to see what we want to see or expect to see. The Administration did not want open conflict with the Soviet Union. It was working on a program of trying to relax tensions, of which a test-ban agreement was one important though distant goal. It most definitely did not want an offensive Soviet base in Cuba, in the same way that Zermatt, the famous Swiss ski resort, did not want typhoid fever and refused to acknowledge its existence until epidemic proportions had been reached. Just as President Roosevelt wanted no war in the Far East-no war on two fronts-and didn't want to believe that it could happen, so we didn't want to believe that the Soviets were doing what they were doing.
When this is the background of expectation, it is only natural to ignore small clues that might, in a review of the whole or on a simple count, add up to something significant. For example, the large ships that turned out to be the villains in the Cuban case had especially large covered hatches. They were unloaded at night by Soviet personnel, and all Cubans were excluded from the docks. The contents, whatever they were, were moved at night. The decks were loaded with 2½- and 5-ton trucks and cars. But these ships, in transit, had been noted to be riding high in the water. If intelligence analysts in the American community had been more ready to suspect the introduction of strategic missiles, would this information have led them to surmise, before as well as after October 14, that these ships carried "space-consuming [i.e. large volume, low density] cargo such as an M.R.B.M." [vii] rather than a bulk cargo? Roger Hilsman points out that these vessels had been specially designed for carrying lumber, and "our shipping intelligence experts presumably deduced that lumbering ships could be more easily spared than others." "We knew," Hilsman writes, "that the Soviets had had some trouble finding the ships they needed to send their aid to Cuba."[viii] This is a good illustration of the way we can adjust (without doing violence to the facts) a disturbing or unusual observation to "save" a theory-in this case that the Soviets would not send strategic missiles to Cuba.
Our estimate of Soviet behavior included, of course, some expectation of how the Russians would react to what we were telling them, to our warnings in words and acts. However, we overestimated the clarity of our signals. General Maxwell Taylor had visited Florida bases on August 25 with a great deal of publicity. Naval reconnaissance of ships approaching Cuba had been stepped up to the point where U. S. planes were shot at by nervous Cubans on September 2. Castro reacted with great restraint in commenting on this incident-a fact which might in itself have been thought suspicious. But above all, on September 4, President Kennedy announced the installation of surface-to-air missiles in Cuba which had been confirmed by the photographs of August 29. He said with the greatest care that we would not tolerate an offensive base or the installation of missiles capable of reaching U. S. territory. He made the distinction between offensive and defensive weapons, and he did this publicly in a way that put him on the spot. To anyone familiar with the workings of the American political system, this should have indicated that we were "contracting-in." The President was deliberately engaging his own prestige and that of the country. He was reacting to the Republicans as well as to Castro. He was justifying not acting up to a certain point, but making it more likely that he would act beyond that point. In other words, he was drawing a line, and he was making it extremely unlikely that we would back down if that line were crossed. Again on September 13, the President called attention to the firmness of his commitment.
To the official Administration statements, we must add the formal announcements by the opposition party. Senator Everett Dirksen of Illinois and Charles Halleck of Indiana, the Republican Congressional leaders, both issued statements on Cuba on September 7. Halleck warned that the increases in armaments and numbers of military technicians supplied by the Soviet Union to Cuba made the situation there "worse from the point of view of our own vital interests and the security of this country." Senator Dirksen invoked the Monroe Doctrine and defined current Soviet military aid to Cuba as a violation of that doctrine. He pointed out that, in view of our treaty commitments, either the Organization of American States should immediately agree on a course of action or, quoting President Kennedy's speech of April 20, 1961, the United States should act on its own, "if the nations of this hemisphere should fail to meet their commitments against outside Communist penetration."
American elections and their accompanying distractions have been the subject of world-wide speculation and concern. Yet they are not always easy for an outsider to understand. These protests from the opposition were taking place in a setting of pre-election debate, and Khrushchev may have hoped to exploit that fact. He may not have been aware that the alarm expressed by the Republicans was something President Kennedy could not ignore. In addition to explicit proposals and resolutions about the Monroe Doctrine, there was the President's request for Congressional authorization to call up 150,000 reserves. This action too should have been a warning signal; it did trigger a Soviet reassurance that Moscow had no need for an offensive base in Cuba. However, the Soviets did not find these warnings weighty enough to reverse their plans for installation.
Another major barrier to an objective U. S. evaluation of the data was our own estimate of Soviet behavior. The Stennis Report isolated as one "substantial" error in evaluation "the predisposition of the intelligence community to the philosophical conviction that it would be incompatible with Soviet policy to introduce strategic missiles into Cuba."[ix] Khrushchev had never put medium- or long-range missiles in any satellite country and therefore, it was reasoned, he certainly would not put them on an island 9,000 miles away from the Soviet Union, and only 90 miles away from the United States, when this was bound to provoke a sharp American reaction.
In considering this estimate of Soviet behavior, let us remember that the intelligence community was not alone. It had plenty of support from Soviet experts, inside and outside the Government. At any rate, no articulate expert now claims the role of Cassandra. Once a predisposition about the opponent's behavior becomes settled, it is very hard to shake. In this case, it was reinforced not only by expert authority but also by the knowledge both conscious and unconscious that the White House had set down a policy for relaxation of tension with the East. This policy background was much more subtle in its influence than documents or diplomatic experience. For when an official policy or hypothesis is laid down, it tends to obscure alternative hypotheses, and to lead to overemphasis of the data that support it, particularly in a situation of increasing tension, when it is important not to "rock the boat."
In the case of Pearl Harbor, there was a concentration on Atlantic and European affairs, which led to a kind of neglect of, or tendency to ignore, Far Eastern signals, and to a policy of staving off the outbreak of a Pacific war as long as possible. In the last months especially, this tendency was combined with a desire to avoid incidents. The wording of the final warning messages to the Army and Navy reflected this concern:
If hostilities cannot repeat not be avoided the United States desires that Japan commit the first overt act. This policy should not repeat not be construed as restricting you to a course of action that might jeopardize your defense. Prior to hostile Japanese action you are directed to undertake such reconnaissance and other measures as you deem necessary but these measures should be carried out so as not repeat not to alarm civil population or disclose intent. . . . Undertake no offensive action until Japan has committed an overt act.[x]
These directives have been frequently characterized as "do-don't."
Another attempt to avoid incidents was the Navy order of October 17 to re- route all trans-Pacific shipping to and from the Far East through the Torres Straits (between New Guinea and Australia), thus clearing the sea lanes to the north and northwest of the Hawaiian Islands. This order followed a warning of possible hostile action by Japan against U.S. merchant shipping. We avoided any incidents in these sea lanes, and at the same time we cut off the possibility of visual observation of the Japanese task force bound for Pearl Harbor.
In the autumn of 1962, pursuing a policy of reducing tension, the Kennedy Administration made very little allowance for deception in Soviet statements, for false reassurances that would quiet justifiable American fears. On September 2, TASS published a joint communiqué on Soviet military aid to Cuba, referring to the August 27 visit to Moscow of Che Guevara and Emilio Aragones. The Soviet Government announced assistance in metallurgical work and the sending of technical specialists in agriculture to Cuba. They added that
views were also exchanged in connection with threats of aggressive imperialist quarters with regard to Cuba. In view of these threats the government of the Cuban Republic addressed the Soviet government with a request for help by delivering armaments and sending technical specialists for training Cuban servicemen.
The Soviet government tentatively considered this request of the government of Cuba. An agreement was reached on this question. As long as the above- mentioned quarters continue threatening Cuba, the Cuban Republic has every justification for taking necessary measures to insure its security and safeguard its sovereignty and independence, while all Cuba's true friends have every right to respond to this legitimate request.[xi]
This was reassuring in a negative understated way: it limited military aid to vague "armaments" and "technical specialists." On September 11, in response to the President's request to call up reserves, a higher-keyed, if not hysterical, pronouncement was issued by TASS. This started with an attack on "bellicose-minded reactionary elements" and "the provocations the United States Government is now staging, provocations which might plunge the world into disaster of a universal world war with the use of thermonuclear weapons." In the U.S. Congress and in the American press, the Soviet Government claimed, an unbridled propaganda campaign was calling for an attack on Cuba and on Soviet ships "carrying the necessary commodities and food to the Cuban people." "Little heroic Cuba" was pictured as at the mercy of American imperialists, who were alarmed by the failure of their economic blockade and calling for measures to strangle her. Particularly serious was the President's action in asking Congress' permission to call up 150,000 reservists. The statement then embarked on a series of jeers at the ridiculous fears of the American imperialists. The peace-loving Soviet Union was sending agronomists, machine-operators, tractor-drivers and livestock experts to Cuba to share their experience and knowledge and to help the Cubans master Soviet farm machinery.
What could have alarmed the American leaders? What is the reason for this Devil's Sabbath? . . . Gentlemen, you are evidently so frightened you're afraid of your own shadow. . . . It seems to you some hordes are moving to Cuba when potatoes or oil, tractors, harvesters, combines, and other farming industrial machinery are carried to Cuba to maintain the Cuban economy. We can say to these people that these are our ships and that what we carry in them is no business of theirs. . . . We can say, quoting a popular saying: "Don't butt your noses where you oughtn't." But we do not hide from the world public that we really are supplying Cuba with industrial equipment and goods which are helping to strengthen her economy.[xii]
A bit farther on, having had its fun, TASS recalled that "a certain amount of armaments is also being shipped from the Soviet Union to Cuba" and that Soviet military specialists had also been requested by the Government of Cuba. However, the number of Soviet military specialists sent to Cuba "can in no way be compared to the number of workers in agriculture and industry sent there. The armaments and military equipment sent to Cuba are designed exclusively for defensive purposes and the President of the United States and the American military just [like] the military of any country know what means of defense are." The statement went on to imply that any threat to the United States was a figment of the American imagination. The major reassurance then followed:
The Government of the Soviet Union also authorized TASS to state that there is no need for the Soviet Union to shift its weapons for the repulsion of aggression, for a retaliatory blow, to any other country, for instance Cuba. Our nuclear weapons are so powerful in their explosive force and the Soviet Union has so powerful rockets to carry these nuclear warheads, that there is no need to search for sites for them beyond the boundaries of the Soviet Union. We have said and we do repeat that if war is unleashed, if the aggressor makes an attack on one state or another and this state asks for assistance, the Soviet Union has the possibility from its own territory to render assistance to any peace-loving state and not only to Cuba. And let no one doubt that the Soviet Union will render such assistance just as it was ready in 1956 to render military assistance to Egypt at the time of the Anglo-French-Israeli aggression in the Suez Canal region.
This sort of reassurance had also been privately delivered to the President, and the misuse of the private channel apparently shocked President Kennedy as much as the creation of the strategic base in Cuba.
President Kennedy and his staff had believed the Soviet reassurances. Their reaction to what they regarded as deception was one of genuine outrage, for one of the President's basic tenets had been that a state of mutual trust between the great powers was an important part of the problem of relaxing tension. And there is a considerable body of literature which goes farther and isolates the attitude of mutual suspicion itself as the central danger today in international relations.
It is a permanent problem of diplomacy to know where to draw the line in extending trust to unfriendly states. A certain amount of healthy suspicion of the opponent's public statements is in order. The President deliberately tested the willingness of Gromyko to lie, after the President knew the truth, but before the Russians knew that he knew. The trap set by the President aroused the indignation of some of those very Americans who urge mutual trust. But the President of the United States would be simple indeed if he did not build his trust cautiously on the basis of many such probings. The Russian performance in the fall and winter of 1962 made it perfectly clear that we cannot take at face value Russian statements-even those made only to the top American leadership in privacy and without those constraints that might be imposed by having the Chinese or other Communist powers or the non-aligned or our own allies listening.
In periods of high tension it is commonly accepted that deception will be an enemy tactic. Before the Pearl Harbor attack Japanese deception was very refined and ingenious. It involved, among other things, giving shore leave to large numbers of Japanese sailors, reinforcing garrisons on the northern border of Manchuria to give an impression of a thrust to the north, issuing false war plans to Japanese commanders and substituting true ones only days before the attack, and on the diplomatic side continuing the appearance of negotiation. For deception is not confined to statements, but must also be translated into actions.
It is important for the enemy's security that he keep his signals quiet. On the Soviet side this meant that all movement on the island of Cuba must take place at night. The Cubans were excluded from the docks and from many of the missile construction areas. Troops were kept below decks, and unloaded equipment was camouflaged or hidden under the trees. On our own side, in the period before October 22, tight security was important to preserve the initiative. And this tight security was maintained through the next few weeks. The members of the group close to the President, known as the Executive Committee or EXCOM, were directly supervising decisions normally left to lower command levels and were doing paper work normally handled by their staffs. This sort of procedure is fine for a couple of weeks, but it means the neglect of other areas of government and, in particular, other areas of foreign policy.[xiii] Richard Neustadt, a keen observer, reminds us that the Sino-Indian conflict was in progress at the same time, and offers a "lay impression" that "at least one side effect of Cuba" was to tighten the time and narrow the frame of reference of the decision-then in the making-on Skybolt.[xiv] Under conditions of tight security, there is also a danger that we may keep signals not only from the enemy but also from ourselves. There are a good many who feel that careful study by a wider range of experts might have been useful at the time and would be useful now, particularly with regard to the Kennedy-Khrushchev communications. These, like MAGIC, were very closely held during the crisis and had to be read and interpreted swiftly at the time.
Another set of signs we may have misread or missed were those appearing in official Cuban statements. Castro is so verbose and temperamental that we tend not to listen carefully to his speeches. And his controlled press is so dull that we are equally careless about that. In addition, the policy of embargo and explicit isolation of the island tends to carry over in a curious way to ignoring the voice of Cuban officialdom.
It is interesting now to review the Cuban press of 1962 for clues we might have picked up. After Raul Castro's July visit to Moscow, the warmth of the references to the Soviet Union increased noticeably. Thanks and praise became the order of the day. On September 11, the day of the falsely reassuring TASS statement, the Cuban newspaper Revolución underlined the threat of thermonuclear war invoked by TASS. The front page was printed with a single white headline on a black background, and it said: "Rockets Over the United States if Cuba is Invaded." Forcing the Soviet Union's hand in this way had been Cuban policy for some time, so that it was natural for our experts to take this as another instance of Cuban wishful thinking.
Finally, in intelligence work the role of chance, accident and bad luck is always with us. It was bad luck that September-October is the hurricane season in the Caribbean, so that some reconnaissance photography was unclear and certain flights were canceled. It was bad luck that the Red Chinese shot down a U-2 on September 9. In 1941 it was bad luck that we had cut all traffic on the Northwest Passage to Russia, and thereby made visual observation of the Pearl Harbor task force impossible. It was bad luck that there was a radio blackout in the Hawaiian Islands on the morning of December 7, and that Colonel French of the Communications Room then decided to use commercial wire instead of recommending the scrambler telephone for the last alert message.
To sum up then, in both the Pearl Harbor and Cuban crises there was lots of information. But in both cases, regardless of what the Monday morning quarterbacks have to say, the data were ambiguous and incomplete. There was never a single, definitive signal that said, "Get ready, get set, go!" but rather a number of signals which, when put together, tended to crystallize suspicion. The true signals were always embedded in the noise or irrelevance of false ones. Some of this noise was created deliberately by our adversaries, some by chance and some we made ourselves. In addition, our adversary was interested in suppressing the signs of his intent and did what he could to keep his movements quiet. In both cases the element of time also played against us. There were delays between the time information came in, was checked for accuracy, evaluated for its meaning, and made the basis for appropriate action. Many of these delays were only prudent, given the ambiguities and risks of response.
The interpretation of data depends on a lot of things, including our estimate of the adversary and of his willingness to take risks. To make our lives more complicated, this depends on what he thinks the risks are, which in turn depends on his interpretation of us. We underestimated the risks that the Japanese were willing to take in 1941, and the risks that Khrushchev was willing to take in the summer and fall of 1962. Both the Japanese and the Russians, in turn, underestimated our ultimate willingness to respond.
It is important to understand that the difficulties described are intrinsic. By focusing on misestimated capabilities, dispositions and intentions, we obscure the fact that, without a very large and complex body of assumptions and estimates, the data collected would not speak to us at all. If there were no technological constraints whatsoever-if, for example, a large missile installation could be put in place in an instant-no reconnaissance, no matter how frequent, could provide assurance that we would not at any moment face a massive new adversary. The complex inferences involved in the act of interpreting photographs are made possible only by a large body of assumptions of varying degrees of uncertainty, ranging from principles of optics and Euclidean geometry through technological, economic and political judgments. The inferences from the interpretations themselves in turn are based on an even wider range of uncertain beliefs. But just because a very large body of partially confirmed beliefs and guesses is involved in interpreting a reconnaissance photograph or the observations of a Cuban refugee or intelligence agent, it is possible to interpret the photograph or observations in many differing ways. Our beliefs, as Willard Van Orman Quine has put it, are "underdetermined" by our experience, and they do not face experience separately, statement by statement, but always in mass, as a collection. We have a good deal of freedom as to what statements to adjust in the light of any new and seemingly disturbing report.
An observation or its report does not seize us, then, and force any specific interpretation. This relatively free situation of hypotheses in intelligence is no different in kind from that of hypotheses in the more exact sciences such as physics. A more naïve empiricism once suggested that statements in physics could be refuted definitively by observation, by the result of a crucial experiment. But a great many physicists and students of the logic of science, at least since Pierre Duhem, have shown that even the interpretation of the simplest experiment depends implicitly on comprehensive theories about the measuring instruments and a great deal else. It is always possible therefore to "save" a theory or hypothesis by altering some other one of the large set of our beliefs that connects it with any given observation.
If this is true in the more exact sciences it is most obviously true for the role of observations and their interpretation in such spheres of practical activity as the operation of an intelligence agency, and the inferences and decisions of an executive. Here the assumptions that shape interpretation are likely to be more multifarious and also less explicit and therefore often less tentatively held. This puts it mildly. Some of the relevant assumptions may be held passionately. They are likely to include wishful or self-flattering beliefs, items of national pride or claims at issue in partisan debate. In the case of Japan, some of the critical assumptions concerned technology-the range, speed and man?uvrability of the Zero plane, the supposed inability of the Japanese to do any better than the Americans in launching torpedoes in shallow water. In the case of Cuba again some critical assumptions were technological; for example, the minimum time required to put into place and make operational a medium-range ballistic missile. Others concerned the politics and character of the Soviet, Cuban and American leadership and their estimates of each other's willingness to take a chance. Our expectations and prior hypotheses guide our observations and affect their interpretation. It is this prior frame of mind, now changed, that we forget most easily in retrospect. And it is this above all that makes every past surprise nearly unintelligible-and inexplicable except perhaps as criminal folly or conspiracy.
The genuine analogies between Pearl Harbor and Cuba should not obscure the important differences. A study of the Pearl Harbor case makes clear that the problem of getting warning of an impending nuclear raid today is much harder than the problem of detecting the Japanese attack some 20 years ago. It is against this increased difficulty that we must balance improvements in intelligence techniques and organization. But the missile crisis illustrates something else, namely that there are other acts very much short of nuclear war of which we want to be apprised, and here our improved techniques and organization can put us ahead of the game. Action was taken during the missile crisis and taken in time to forestall Soviet plans. For while we can never ensure the complete elimination of ambiguity in the signals that come our way, we can energetically take action to reduce their ambiguity, by acquiring information as we did with the U-2. And we can tailor our response to the uncertainties and dangers that remain.
In the Cuban missile crisis action could be taken on ambiguous warning because the action was sliced very thin. After reconnaissance reduced the ambiguity, the response chosen kept to a minimum the actual contact with Russian forces, but a minimum compatible with assuring Khrushchev that we meant business: quarantine, the threat of boarding, the actual boarding of one Lebanese vessel chartered to the Soviet Union. Further, it was a response planned in great detail as the first in a sequence of graded actions that ranged from a build-up of U.S. Army, Marine and Tactical Air Forces in Florida and our southeastern bases to a world-wide alert of the Strategic Air Command. We had been partially prepared for such sequences of action short of nuclear war by the Berlin contingency planning, and this put us in a position to use the warning we had accumulated. If we had had to choose only among much more drastic actions, our hesitation would have been greater.
The problem of warning, then, is inseparable from the problem of decision. We cannot guarantee foresight. But we can improve the chance of acting on signals in time to avert or moderate a disaster. We can do this by a more thorough and sophisticated analysis of observers' reports, by making more explicit and tentative the framework of assumptions into which we must fit any new observations, and by refining, subdividing and making more selective the range of responses we prepare, so that our response may fit the ambiguities of our information and minimize the risks both of error and of inaction. Since the future doubtless holds many more shocks and attempts at surprise, it is comforting to know that we do learn from one crisis to the next.
[i] Flights over the island took place on September 5, 17, 36, 29, October 5, 7 and 14. The irregularity is attributed to bad weather.
[ii] See testimony, September 17, 1962: United States Senate, Committee on Foreign Relations and Committee on Armed Services, Situation in Cuba, 87th Cong., 2d Sess., 1962, p. 7, 12; U. S. News and World Report, November 19, 1962 (distributed week of November 12), p. 87; and speech to the Senate, October 12, 1962.
[iii] U. S., Congress, House of Representatives, Subcommittee on Department of Defense Appropriations, Department of Defense Appropriations for 1964, 88th Cong., 1st Sess., 1963, p. 45-46. These hearings contain most of the intelligence data cited in this article.
[iv] According to Roger Hilsman, the request for a U-2 flight covering the western end of the island was made on October 4-ten days before the flight was actually made. "The Cuban Crisis: How Close We Were to War," Look, August 25, 1964, p. 18.
[v] According to W. W. Kaufmann, The McNamara Strategy, Harper & Row, 1964, p. 270. According to John Hughes, Special Assistant to General Carroll, "about 600 nautical miles," Hearings, p. 15.
[vi] U. S., Congress, Committee on Armed Services, Preparedness Investigating Subcommittee, Investigations of the Preparedness Program, Interim Report on Cuban Military Build-Up, 88th Cong., 1st Sess., 1963, p. 3.
[vii] "Department of Defense, Special Cuba Briefing by the Honorable Robert S. McNamara, Secretary of Defense, State Department Auditorium, 5:00 p.m., February 6, 1963." A verbatim transcript of a presentation actually made by General Carroll's assistant, John Hughes.
[viii] Op. cit. p. 18
[ix] Op. cit., p. 3.
[x] U. S., Congress, Joint Committee on the Investigation of the Pearl Harbor Attack, Pearl Harbor Attack, 79th Cong., 2d Sess., 1946, Part 14, p. 1407.
[xi] The New York Times, September 3, 1962.
[xii] Text of Soviet statement, The New York Times, September 12, 1962.
[xiii] According to Secretary Rusk, "Senior officers did their own typing; some of my own basic papers were done in my own handwriting, in order to limit the possibility of further spread. . . ." C.B.S. Reports, televised interview of Secretary Rusk by David Schoenbrun, November 28, 1962.
[xiv] U. S. Congress, Senate Subcommittee on National Security Staffing and Operations of the Committee on Government Operations, Administration of National Security, 88th Congress, 1st Session, 1963, Part 1, p. 97, testimony of March 25, 1963.