Wednesday, January 13, 2010

Connecting the Dots

Stephen Colbert put his own hilarious spin on the “underwear bomber” and the failure of the national security system to “connect the dots” and thus predict and interdict the would-be bomber.
Colbert offers up a montage of pundits (and the president) talking about the failure to connect the dots, and then notes that connecting the dots is the easiest puzzle of them all.
Connecting the dots, it seems, is not rocket science.
Actually, it's way more difficult than rocket science. Rocket scientists depend upon immutable laws of physics and gravity. They can test and retest on various scales and project from those tests what will happen on a full scale with near certainty.
Try that with human behavior.
I've been reading William Poundstone's Priceless, a fascinating book on the theory of pricing and decision making. Poundstone relates a story from Nobel-prize winning psychologist Daniel Kahneman's formative experience of sitting in on psychoanalysts' patient conferences. One such conference, Poundstone reports, stuck in Kahneman's memory because the patient did not attend -- having committed suicide the night before. Poundstone writes:
"It was a remarkably honest and open discussion," Kahneman mordantly observed, "marked by the contradiction between the powerful retrospective sense of the inevitability of the event and the obvious fact that the event had not been foreseen."

The analysts, who presumably knew the patient well and were in quite regular conversation with him, failed to connect the dots.
The same was quite obviously the case with the Ft. Hood shooter last year. Maj. Nidal Malik Hasan's behavior had, as was widely reported, raised concerns among his colleagues but others close to him were shocked at the shootings. The imam at his mosque said to CNN, "The quiet, very peaceful person coming in and out of the mosque, I couldn't believe he could have done this."
On the other hand, a Pentagon report acknowledged a failure on the part of a terrorism task force "to connect the dots" which included the major's e-mail correspondence with a radical Yemeni cleric who has also been connected with the alleged underwear bomber.
Moreover, Major Hasan's former colleagues at Walter Reed had wondered aloud about his mental stability and capacity to serve.
All of which leads to the same contradiction of a powerful retrospective sense of inevitability and the obvious fact that no one saw it coming.
We had a young friend visiting over the weekend. She is a Virginia Tech graduate who was in school when Seung-Hui Cho killed 32 students during a rampage shooting in April, 2007. As CBSNews noted, "A special state panel convened after the shooting concluded the school had misinterpreted privacy laws and had failed to connect the dots."
Our friend noted that she had several classes with Cho. "He was a little odd and very quiet," she recalled, but said she would never have thought he was likely to commit such a crime. Her recollection sounded remarkably like the ones heard so often following such shocking events. The response, "he was so quiet; I never would have thought he'd do such a thing," has become a cliche in the aftermath of these spasms of violence.
The dots, in other words, are damnably difficult to see in advance.
In a technologically advanced society such as ours it comes as no surprise that in the aftermath of an attempted terror attack we would turn to technology to protect us. In this case, the security experts are turning to the full body scanning technology to find the dots that we do not seem able to connect.
The problem is that even when all the dots are on full display connecting them -- when it involves predicting human behavior -- is incredibly difficult and inexact. Poundstone spends an entire book demonstrating that the rational actor school of economics fails miserably to explain something a fundamental to economic activity as pricing and purchasing decisions. At least when it comes to those basic decisions the rational animal is unreasonable. And if such relatively simple transaction are not subject to the predictions of reason why should we expect that far more complex human interactions would be?
On the other hand, what Poundstone ultimately demonstrates is not that it is impossible to predict human decision making with regard to price and purchase, but rather the decisions that we make are contrary to the expectations of prevailing economic theory. Perhaps what we need with respect to security is a better theory.
There are plenty of quiet, troubled, angry young men who will never shoot up their college campuses or stuff explosives into their underwear or join a terror cell. And there is the occasional seemingly well-adjusted young woman who turns out to be a serial killer.
All of which suggests that, while we can hope that the national security apparatus improves its own internal communications and ability to cut through its own red tape, we probably cannot expect that the human dots will get any easier to connect until and unless we begin to look at them differently -- not through a different scanner but through different eyes.


jamie said...

Thanks for this David. I'm writing my thesis on a French sociologist, Pierre Bourdieu, who spent his career trying to show why both personalism (Sarte, Rational Action Theory, most phenomenology) and deterministic structuralism fail to explain anything about human practice. His late book Practical Reason is a good short introduction to the arguments.

Christian Wright said...

There is a pair scenes in the recent Star Trek movie that show the same thing. In the first, Spock's father explains that, due to his position as ambassador to earth, it was "logical" for him to marry Spock's mother that he might learn more about human culture. In the second, immediately after her death, he says to Spock, "I married her because I loved her." We certainly have the capacity for rational thought be we are not merely the 'rational animal' that rational action theory would want us to believe.