What the Dog Saw: And Other Adventures

could not see it.

None of these postmortems, however, answer the question raised by the Yom Kippur War: was this pattern obvious before the attack? This question – whether we revise our judgment of events after the fact – is something that psychologists have paid a great deal of attention to. For example, on the eve of Richard Nixon’s historic visit to China, the psychologist Baruch Fischhoff asked a group of people to estimate the probability of a series of possible outcomes of the trip. What were the chances that the trip would lead to permanent diplomatic relations between China and the United States? That Nixon would meet with the leader of China, Mao Tse-tung, at least once? That Nixon would call the trip a success? As it turned out, the trip was a diplomatic triumph, and Fischhoff then went back to the same people and asked them to recall what their estimates of the different outcomes of the visit had been. He found that the subjects now, overwhelmingly, “remembered” being more optimistic than they had actually been. If you originally thought that it was unlikely that Nixon would meet with Mao, afterward, when the press was full of accounts of Nixon’s meeting with Mao, you’d “remember” that you had thought the chances of a meeting were pretty good. Fischhoff calls this phenomenon “creeping determinism” – the sense that grows on us, in retrospect, that what has happened was actually inevitable – and the chief effect of creeping determinism, he points out, is that it turns unexpected events into expected events. As he writes, “The occurrence of an event increases its reconstructed probability and makes it less surprising than it would have been had the original probability been remembered.”

To read the Shelby report, or the seamless narrative from Nosair to bin Laden in The Cell, is to be convinced that if the CIA and the FBI had simply been able to connect the dots, what happened on September 11 should not have been a surprise at all. Is this a fair criticism or is it just a case of creeping determinism?

3.

On August 7, 1998, two Al Qaeda terrorists detonated a cargo truck filled with explosives outside the US embassy in Nairobi, killing 213 people and injuring more than four thousand. Miller, Stone, and Mitchell see the Kenyan embassy bombing as a textbook example of intelligence failure. The CIA, they tell us, had identified an Al Qaeda cell in Kenya well before the attack, and its members were under surveillance. They had an eight-page letter, written by an Al Qaeda operative, speaking of the imminent arrival of “engineers” – the code word for bomb makers – in Nairobi. The US ambassador to Kenya, Prudence Bushnell, had begged Washington for more security. A prominent Kenyan lawyer and legislator says that the Kenyan intelligence service warned US intelligence about the plot several months before August 7, and in November of 1997 a man named Mustafa Mahmoud Said Ahmed, who worked for one of Osama bin Laden’s companies, walked into the US embassy in Nairobi and told American intelligence of a plot to blow up the building. What did our officials do? They forced the leader of the Kenyan cell – a US citizen – to return home, and then abruptly halted their surveillance of the group. They ignored the eight-page letter. They allegedly showed the Kenyan intelligence service’s warning to the Mossad, which dismissed it, and after questioning Ahmed, they decided that he wasn’t credible. After the bombing, The Cell tells us, a senior State Department official phoned Bushnell and asked, “How could this have happened?”

“For the first time since the blast,” Miller, Stone, and Mitchell write, “Bushnell’s horror turned to anger. There was too much history. ‘I wrote you a letter,’ she said.”

This is all very damning, but doesn’t it fall into the creeping-determinism trap? It is not at all clear that it passes the creeping-determinism test. It’s an edited version of the past. What we don’t hear about is all the other people whom American intelligence had under surveillance, how many other warnings they received, and how many other tips came in that seemed promising at the time but led nowhere. The central challenge of intelligence gathering has always been the problem of “noise”: the fact that useless information is vastly more plentiful than useful information. Shelby’s report mentions that the FBI’s counterterrorism division has sixty-eight thousand outstanding and unassigned leads dating back to 1995. And, of those, probably no more than a few hundred are useful. Analysts, in short, must be selective, and the decisions made in Kenya, by that standard, do not seem unreasonable. Surveillance on the cell was shut down, but, then, its leader had left the country. Bushnell warned Washington – but, as The Cell admits, there were bomb warnings in Africa all the time. Officials at the Mossad thought the Kenyan intelligence was dubious, and the Mossad ought to know. Ahmed may have worked for bin Laden but he failed a polygraph test, and it was also learned that he had previously given similar – groundless – warnings to other embassies in Africa. When a man comes into your office, fails a lie-detector test, and is found to have shopped the same unsubstantiated story all over town, can you be blamed for turning him out?

Miller, Stone, and Mitchell make the same mistake when they quote from a transcript of a conversation that was recorded by Italian intelligence in August of 2001 between two Al Qaeda operatives, Abdel Kader Es Sayed and a man known as al Hilal. This, they say, is yet another piece of intelligence that “seemed to forecast the September 11 attacks.”

“I’ve been studying airplanes,” al Hilal tells Es Sayed. “If God wills, I hope to be able to bring you a window or a piece of a plane the next time I see you.”

“What, is there a jihad planned?” Es Sayed asks.

“In the future, listen to the news and remember these words: ‘Up above,’” al Hilal replies. Es Sayed thinks that al Hilal is referring to an operation in his native Yemen, but al Hilal corrects him: “But the surprise attack will come from the other country, one of those attacks you will never forget.”

A moment later al Hilal says about the plan, “It is something terrifying that goes from south to north, east to west. The person who devised this plan is a madman, but a genius. He will leave them frozen [in shock].”

This is a tantalizing exchange. It would now seem that it refers to September 11. But in what sense was it a “forecast”? It gave neither time nor place nor method nor target. It suggested only that there were terrorists out there who liked to talk about doing something dramatic with an airplane – which did not, it must be remembered, reliably distinguish them from any other terrorists of the past thirty years.

In the real world, intelligence is invariably ambiguous. Information about enemy intentions tends to be short on detail. And information that’s rich in detail tends to be short on intentions. In April of 1941, for instance, the Allies learned that Germany had moved a huge army up to the Russian front. The intelligence was beyond dispute: the troops could be seen and counted. But what did it mean? Churchill concluded that Hitler wanted to attack Russia. Stalin concluded that Hitler was serious about attacking, but only if the Soviet Union didn’t meet the terms of the German ultimatum. The British foreign secretary, Anthony Eden, thought that Hitler was bluffing, in the hope of winning further Russian concessions. British intelligence thought – at least, in the beginning – that Hitler simply wanted to reinforce his eastern frontier against a possible Soviet attack. The only way for this piece of intelligence to have been definitive would have been if the Allies had had a second piece of intelligence – like the phone call between al Hilal and Es Sayed – that demonstrated Germany’s true purpose. Similarly, the only way the al Hilal phone call would have been definitive is if we’d also had intelligence as detailed as the Allied knowledge of German troop movements. But rarely do intelligence services have the luxury of both kinds of information. Nor are their analysts mind readers. It is only with hindsight that human beings acquire that skill.

The Cell tells us that, in the final months before September 11, Washington was frantic with worry:

A spike in phone traffic among suspected Al Qaeda members in the early part of the summer [of 2001], as well as debriefings of [an Al Qaeda operative in custody] who had begun cooperating with the government, convinced investigators that bin Laden was planning a significant operation – one intercepted Al Qaeda message spoke of a “Hiroshima-type” event – and that he was planning it soon. Through the summer, the CIA repeatedly warned the White House that attacks were imminent.

The fact that these worries did not protect us is not evidence of the limitations of the intelligence community. It is evidence of the limitations of intelligence.