Nov 17, 2005

the efficacy of torture

Regarding torture (once more), Mark Olson claims,
I haven’t been convinced by Mr Kuznicki’s arguments that they never work, and I don’t think I’ve seen data to that effect from anyone who has access to the data and has done a careful study to verify that assumption or not. This is a matter which can surely be settled. The question is, why hasn’t it? There is the distinct possibility that the reason this hasn’t settled for the public is that the answer is the uncomfortable one, not the one that nice people in warm rooms wish it might be, not the one that voters would cheer.
The research is out there. It's just that Olson hasn't taken the time to find it. Consider a document that demands a much wider audience, "A Consequentialist Argument against Torture Interrogation of Terrorists. In language both banal and horrifying, Jean Arrigo addresses theories behind torture's efficacy to justify a particular course of action. If torture works, we don't yet have solid evidence to prove it.
Even under the Nazis, torture interrogation failed to break dozens of high state officials and military commanders involved in late-war plots to assassinate Hitler. According to Peter Hoffman’s History of the German Resistance: 1933-1945[27]:
Six months from the start of their investigations the Gestapo still had nothing like precise knowledge of the resistance movement.......This lack of information and knowledge is all the more astounding in that Himmler's men employed every means to extract confessions.... Moreover all forms of torture were used without hesitation....
Hoffman attributes the failure of the Gestapo to the “fortitude of their victims....”

A criminological analysis of 500 British court cases found that police interrogation of defendants contributed little to discovery and conviction. Rather, the study concluded that interrogation fulfilled certain psychological and administrative needs and that “police perceptions of reality dominate the criminal process.”[122]
Torture, Arrigo notes, simply isn't as effective as other methods.
Here I pass over a considerable literature pointing to the greater efficacy of noncoercive interrogation based on social skills: subtlety and finesse of interrogation,[34] sympathy with the subject,[35] appeal to the subject’s self interest,[36] and outright deceit and trickery.[37]...
Arrigo considers the obvious statistical fact that many, if not most, of the subjects will be useless, or worse--innocent.
What proportion of ignorant or innocent suspects are likely to be interrogated under torture? Modern crime statistics indicate that among suspects arrested and charged with serious crimes, one-half to three-quarters are not convicted, depending on the state of jurisdiction.[78] Under the proposed torture interrogation program, a detainee firmly believed to be involved in serious acts of terrorism will likely be tortured. The secrecy and urgency of terrorist cases certainly cannot improve the rate of accuracy over serious criminal convictions, for the counterterrorist program rejects normal judicial safeguards—the right not to testify against oneself, the right to legal counsel, habeas corpus, bail setting, public hearings, and so on. Moreover, in tracking terrorist operations it is customary to interrogate individuals just because they are acquainted with a person who has been detained, not because they are suspected of crimes.[79] So an error rate of one-half to three-quarters of torture interrogees would be a low estimate. For a long historical comparison, examination of court records for 625 cases of torture interrogation in France, from the 1500s through the mid-1700s, showed approximate rates of error—that is, no confession on the rack, under repeated drowning, crushing of joints, and the like—in 67% to 95% of cases, depending on the province.[80]
Arrigo considers the purported urgency that demands torture, and finds that the threat of societal backlash or instability caused by programs of torture is as destructive as the threat of civilian casualties; this, in Arrigo's final estimation, is reason to reconsider torture as an effective means of interrogation.
The moral error in reasoning from in the ticking bomb scenario arises from weighing the harm to the guilty terrorist against the harm to the prospective innocent victims. Instead, the harm to innocent terrorist victims should be weighed against the breakdown of key social institutions and the state-sponsored torture of many innocents. Stated most starkly, the damaging social consequences of a program of torture interrogation evolve from institutional dynamics that are independent of the original moral rationale.
Now, Olson may respond by attacking this evidence, as it is secondary, and perhaps oversimplifies the matter. But the larger point is that the information is accessible, and that these judgments can and must be made in its light. The burden of proof, as it always has, lands squarely on the shoulders of those who would promote, excuse, or rationalize immoral means to a moral end.

4 comments:

MT said...

Did you read the New Yorker (Atlantic?) article awhile back (by Mark Danner?) about the Israeli experience? There's kinds of playing dirty that some might call torture and which sources credible-sounding-to-me say work and have saved many lives in Israel (i.e. counting Jewish lives only).

Mark said...

I won't attack the evidence, but just that Mr Arrigo isn't quite asking (and answering) the same question..

My question was: Given the data amassed from the variety of techniques how does interrogation stack up? Mr Arrigo argues that coercion doesn't work well. So? Neither does any other method of intelligence gathering.

We get data from very many sources, all are mostly bad and the reliability varies widely but for the most part intelligence from any of the sources is usually (I understand) quite poor.

Where does interrogation fit in this spectrum? Given that there are different styles of interrogation, you might "bin" the "interrogation" data differently (separating subtle, sympathetic, trickery, self interest, and coercion). But certainly from the current conflict the data is hard to tease out without the will to know and security clearance to dig. Congress and the WH have that clearance. You and I do not.

Mr Arrigo argues that interrogations are often unsuccessful. He ignores the fact that every other method of intelligence gathering is also mostly unsuccessful as well.

If you google the question, you see a vast number of discussions by people who argue in the absence of data. For a post Enlightenment culture is that the right methodology?

Perhaps the best approach might be to demand that our President and/or Intelligence oversight committees look into this. Like I said, they may have done so already and if they got an answer that implied that was effective but because of the political ramifcations do you think they would trumpet it? Or do you think things might play out pretty much like they have been.

Jim Anderson said...

Mark,

I guess you didn't notice the part where Arrigo mentions the relative lack of efficacy of torture compared to other forms of interrogation.

McCain's Newsweek piece brings up a poignant example: when tortured to name his fellow soldiers, McCain listed the starting line of the Green Bay Packers. There is simply no guarantee that torture will produce actionable intelligence, since in a broken psychological state people will say anything to comply, even what they know to be false.


Murky Thoughts, it was The Atlantic, an article by Mark Bowden. His latest perspective is rather weakly drawn in an Opinion Journal article.

Anonymous said...

Just in case anybody's still reading, my thoughts:

First, it's up to Mark to prove that torture works, not up to us to prove that it doesn't.

Second, as to the administration or it's appointees examining this, with their oh-so-secret 'insider knowledge' - riiiight. That's up there with 'vast stockpiles of WMD's', smoking gun mushroom clouds, and Saddam-Al Qaida ties.