April 02, 2005

Column 2005-4-02 Commentary

In this column, Brooks takes an idiosyncratic -- one might even say idiotic -- look at intelligence reform. Brooks knows what's wrong with the CIA: it tries to hard to be scientific. Apparently, the American intelligence community is insufficiently touchy-feely. It needs to try harder to empathize with our enemies, to imagine how we appear to them and how they will interpret our actions. That's interesting, as I had been led to believe that people who did that were liberal softies, and that our enemies have only one motivation: they hate freedom. Regardless, Brooks argues for an "intuitive generalist" approach to intelligence, claiming that the "scientific method" used by the CIA, though seemingly authoritative, is actually counterproductive. Instead, analysts should be sent off to study "Thucydides, Tolstoy and Churchill to get a broad understanding of the full range of human behavior." Brooks doesn't mention whether or not they should also study something directly relevant to their specialty, but we'll be generous and assume that he simply forgot. And Brooks goes so far as to back up his argument with a source, so you know that he's serious (though in true Brooksian fashion, he can't even cite his source correctly: for Donald Zagoria, read Alan Whiting throughout). Strangely, though, his source is a paper by a Yale undergraduate discussing the CIA's failure to correctly predict China's willingness to negotiate with the United States in 1970. And that's really all you need to know about this column: Brooks constructs his entire argument on the basis of one 35-year-old data point. Well, nobody will ever accuse him of being overly scientific.

Brooks's first problem is that the CIA is not using the scientific method. In fact, if the CIA was performing rigorous experiments to test its hypotheses, they would probably have a considerably better record than they do. Sadly, it's difficult to concieve of an experiment that would test the hypothesis, that, for example, North Korea has eight nuclear weapons, that doesn't involve the fiery destruction of Seoul and Tokyo. Therefore, to claim that the CIA is using the scientific method to come to its conclusions is completely incorrect. What Brooks says the CIA is really using is the pseudo-scientific method, as perfected by Samuel Huntington. This method involves quantifying as many of the things you are talking about as you can so that your work looks scientific, even if the things you attempt to turn into numbers cannot really be quantified and your attempts to do so give nonsensical results. The prime example of this (or at least the one that springs readily to mind) is Huntington's claim that South Africa was a satisfied society in the early sixties under apartheid, based on a calculated "satisfaction index". (Huntington backed up his claim by stating, in flat contradiction of the facts, that there had been "no major riots, strikes, or disturbances" in South Africa.) However, it should be noted that Brooks offers no real facts to back up his claim that the CIA has fallen into the Huntington trap. There is a lot of rhetoric about "bloodless compilations of data by anonymous technicians" and "systematic, codified and bureaucratic" processes, but no actual facts. Which is too bad, because this is the strongest part of Brooks's argument: social science's attempt to be scientific are inherently doomed to failure because of the inability of social scientists to do controlled experiments. Sadly, he presents no non-anecdotal evidence to show that American intelligence analysts have gone down this blind alley.

This is not too surprising, though: after all, Brooks barely provides any evidence to back his claim that a more intuitive method is necessary. Aside from his 35-year-old story about China, he "justifies" this intuitive generalist b.s. almost exclusively with this sentence: "Individuals are good at using intuition and imagination to understand other humans." Apparently, Brooks believes that if Malcom Gladwell writes a best-selling book about something, that makes it right (and Gladwell's book, Blink, is actually about the value of first impressions; I don't think even Brooks proposes that intelligence analysis should be based entirely on the first impressions of the analysts). This sentence is so general as to be meaningless, and ignores the fact that individuals can also be abysmally bad at using intuition and imagination to understand other humans. For evidence of this, we need to go no further than our own President, who said, back in June 2001, that "I found [Putin] to be very straightforward and trustworthy. . . . I was able to get a sense of his soul." Not only did Bush's intuition fail him, his first impression was also wrong. Maybe we shouldn't base all of our intelligence analysis on this intuition stuff after all. Incidentally, Brooks's claim that his "intuitive generalists" are more "modern and rigorous" than the pseudoscientists is ludicrous, unless he's using some new definition of rigorous (or of intuition). Imagination and intuition involve going outside of the rules and making leaps that cannot be justified by strict logic: i.e., the exact opposite of rigor. But perhaps Brooks is simply trying to lead by example. There is no certainly no rigor and a whole lot of imagination in this piece of social analysis.

Brooks finishes his column by saying that he'll know that reform has happened "when there's a big sign in front of C.I.A. headquarters that reads: Individuals think better than groups." This may confuse those of you who believe that the point of this column is to agitate for a more intuitive approach to intelligence analysis. Apparently, you haven't been paying attention to the column. Remember, intuition and imagination are key: Brooks practices what he preaches and makes a "novelistic judgement" not from any logic or evidence but from his "broad understanding of the full range of human behavior" gained from reading Tolstoy and Churchill. Personally, I think that a good judgement of whether or not reform has worked would be performance-based, but I guess this kind of naive, surface-level thinking is why I don't have a column at the New York Times.

There are also very good reasons for Brooks to disdain a performance-based evaluation of the CIA. For one thing, he might have to talk about mistakes more recent than 1970, such as the intelligence community's confident assertion that Iraq had WMDs. There was, in fact, a big report published about that quite recently: Brooks even mentions it in his column (he refers to the "presidential panel on intelligence"), though he doesn't say what exactly the report was investigating. In fact, Brooks somehow manages to write a column about intelligence reform without mentioning Iraq's nonexistent WMD's or 9/11, even though most calls for reform of the CIA were based on its bogus claims about the former and failure to prevent the latter. The reason that Brooks chooses to cite a 35-year-old mistaken analysis that didn't even have serious consequences is because neither of these two modern intelligence failures fit into Brooks's argument. The question of whether or not Iraq had WMD's did not require reading Thucydides and Tolstoy, and to answer it did not require imagination or intuition. All it required was an objective analysis of the evidence. All of Brooks's intuitive generalist methods -- gaining a deep understanding of Iraqi society, for instance -- were irrelevant. In fact, it seems all too likely that the problem was not that analysists lacked imagination and intuition, but that they had too much (especially of the former). It is instructive to note that Brooks's column on the Duelfer report (laughably entitled "The Report That Nails Saddam") dismissed the finding that there had been no WMD's as a small part of the report that had been blown out of proportion by partisan Democrats. Instead, Brooks focused on parts of the report that revealed that Saddam really, really wanted WMD's and that he was gaming the oil-for-food program and the sanctions in an attempt to restart his program. All true, and exactly the kind of things that an intuitive generalist would focus on as proof positive that Saddam had a weapons program, but the fact remains that there was no weapons program, and Saddam was not an immediate threat and would not have been a threat for some years.

So, if a lack of intuition was not at fault, why was the intelligence so wrong? Why, for instance, did the U.S. ignore the work of U.N. inspectors that refuted most of the claims they made about Saddam's WMD programs prior to the war? Why did the intelligence community base its assertions about Saddam's extensive biological weapons program on a single defector, codenamed Curveball, whose reports were never verified and whose credibility was cast into doubt as long as ago as May 2000 when his very first interview by an American revealed that, contrary to the information the Department of Defense had been given, he spoke excellent English, and that he was likely an alcoholic? Why was it that the CIA did not meet Curveball until March 2004, and that the CIA did not figure out that Curveball had been out of Iraq during the time he was supposedly working on bioweapons until January 2004? It's not as if Curveball was a bit player: according to the recently released report,
"Virtually all of the Intelligence Community’s information on Iraq’s alleged mobile biological weapons facilities was supplied by a source, codenamed 'Curveball.'" But somehow the CIA neglected to make any basic checks on their source, despite doubts about his authenticity, until long after the war. What could have caused such incompetence? Might it have been administration pressure?

Similar questions arise when one considers the other intelligence failure Brooks ignores, the fact that the CIA and FBI were caught flatfooted by September 11. Intuitive generalism doesn't help here, either, since even the old-fashioned way of doing things figured out that Al Qaeda would try to attack in the U.S., or so one would gather from the title of the August 6 PDB "Bin Laden determined to strike in U.S.". In fact, warnings that Bin Laden was planning an attack involving one or more airplanes, quite possibly as weapons, dated back to 1998. The FAA issued several warnings of imminent hijackings in 2001, and the NSA reported a large number of communications suggesting the possibility of an attack. In fact, in July 2001 the FAA, FBI, and INS were warned that terrorist attack was coming soon, and this warning was taken seriously enough that non-essential travel by countererrorism personnel was suspended. And yet the administration took no other action. Perhaps 9/11 could not have been foreseen or prevented, but the administration could at least have made an effort, rather than largely ignoring counterterrorism.

Thus, it appears that Brooks's mystifying omission of the WMD and 9/11 questions is actually susceptible of a perfectly reasonable explanation: Brooks is, as usual, acting as an administration shill. He places all the blame for intelligence failures on "epistemological" problems in the intelligence community dating back at least 35 years, thus neatly deflecting attention from the fact that the main fault in both cases lies with an administration that manipulated or ignored intelligence to suit its purposes. The problem is not that the CIA occasionally fails to read the minds of foreign leaders correctly: mind reading is difficult, and my guess is that the CIA is just about as good at it as anybody else in the game. The problem is that the CIA's intelligence analysts were not allowed to do their job, or ignored when they did if their conclusions were not what the administration wanted to hear. Brooks is correct when he says that "the problem is not bureaucratic," but it's not epistemological either: it's political, and won't go away until the administration does too.

Since it's become traditional, I'll ask: how much longer does the Times plan on allowing Brooks to use its editorial page to lay down smokescreens for the administration? And surely I'm not the only one who feels that Brooks attacking the CIA's intelligence failures is a serious case of the pot calling the kettle black?