The Efficacy of Torture

Not long ago I wrote a post arguing, among other things, against the use of torture as an instrument of public policy. I stand by this conviction for three reasons. First, though I recognize that while torture will probably always occur here and there as an inevitable consequence of war, I believe that it is fundamentally an immoral form of behavior. For torture to be moral one must give up the notion that people ought to have certain rights simply because they are people. Having given up such a core concept, the only realm in which one can discuss morality is within the confines of some particular group. Being moral only on an ad hoc basis (that is, only toward “good” people) leaves the whole concept rather superfluous. A promise not to beat one’s friends is not particularly constraining. My second objection to torture comes from a belief that the value of any information gained by exercising the practice as an instrument of public policy is probably far outweighed by the level of hatred it inspires in an enemy, aiding in his recruiting efforts, hardening his resolve, and justifying his own resort to the use of similar forms of atrocity. I know that such an assertion, like most others that involve the psychology of groups, is all but impossible to prove. That does not render it untrue, but I admit it weakens it as an argument. Third, I believe that a public policy which condones torture is fundamentally incompatible with the stability of a Republic. If the government can exercise such a policy as a legal option, no citizen can long feel free to dissent. A nation that condones the torture of foreigners will, soon enough, employ torture against its own citizens. As an instrument of repression, torture has few peers.

These things said, there is one argument against torture that is not only ludicrous, but dangerous and tragic in its own right. That is the argument that torture simply doesn’t yield any useful information. You can read a synopsis of this position in the article at the link below:

http://harpers.org/archive/2009/09/hbc-90005768

In essence, the argument is that neurobiological research shows that the extreme stress of torture has a negative effect on memory, producing statements that are uncertain or susceptible to the interrogator’s suggestion. Further, while torture increases one’s willingness to talk, it does not guarantee that one will speak the truth. I don’t have any particular doubts about the science itself, but the conclusions drawn, at least by Harper's, The Daily Beast, et al, are oversimplifications and erroneous.

The presumption is that stress not only impairs memory, but that it impairs it to such a degree as to render it wholly dysfunctional. Further, the Harpers article, at least, fails to make any distinctions between different methods of questioning, corroboration between different subjects, corroboration with outside evidence, etc. The implicit scenario is one in which the subject of torture is simply abused until he discerns from sloppy interrogation technique more-or-less what it is that his torturers expect to hear, whereupon he delivers up that narrative and his torturers go away with a presumption they have uncovered a fact. As much as we may hate the practice of torture, we are not justified in presuming that its practitioners are uniformly that credulous and amateurish. The CIA, for example, has been torturing people since its origins as the OSS in Second World War. While they are far from infallible, there is no reason to believe that they are any less methodical about torture than they are about any other intelligence gathering activity.

Deriving information from any sort of interrogation is a signal-to-noise ratio problem – a matter of sorting what is true from the noise of both impaired memory and deception on the part of the subject, and from poor questioning technique on the part of the interrogator. Methods and circumstances matter.

Imagine six people have been physically separated since capture and are asked, under torture, to give up the name of their immediate superior. One successfully resists, three give unique names, and two give the same name. While not utterly conclusive, the result of getting the same name from two separated individuals certainly increases the probability that the two are telling the truth. If the neurobiological objection implied by Harper’s were correct, all six subjects would be so impaired that their answers would be essentially random, thus any consensus that occurred between them would also be either random or the product of leading questions, and, therefore, irrelevant. This, frankly, is absurd. If our cognitive abilities failed utterly under severe stress, no one could be tortured effectively for the purpose of extracting information – but we probably could not have survived the rigors of our evolutionary history either.

The kind of scenario described above has not been unusual in wartime. During the Second World War, for example, the Soviets, the Germans, the Japanese and others all conducted, with some regularity, battlefield interrogation under torture along these general lines. Military interrogations are an entirely different matter from, say, the interrogations of the Spanish Inquisition. The Inquisition’s goal was, in some sense, to get at the truth – but since whether one was a heretic or not was entirely dependent on the Inquisition’s verdict, torture served largely to make the words of the accused conform to the Inquisitor’s prior conclusions. Having been broken by torture one confessed to one’s sins in open court. In a case like this, accurate memory of past events was not particularly important. Military interrogations are, on the other hand, wholly disinterested with the guilt, innocence, or immortal souls of the interrogation subjects. Their goal is simply to extract specific, factual information about the status of enemy forces. The military interrogator does not record his (or her) findings in a history book, but disseminates them to military commanders who in turn use them as the basis for their battle plans. If the information deviates very greatly from actual facts, plans will fail and soldiers will die in greater numbers. As a rule, battlefield commanders tend to notice such things. If we accept the Harper’s headline that “Torture Doesn’t Work,” we have to conclude that thousands of military leaders just never caught on. Somehow, they never noticed that the reports of their intelligence officers, derived from tortured prisoners’ statements, were never any more reliable than random guesses. Military leaders have often been dogmatic and thick headed – but they are not completely blind. They do eventually discard methods that don’t produce results.

Again, I am not advocating torture. The mere fact that something may be effective does not make it justifiable. However, it must be admitted that the empirical record trumps neurobiological theory and educational credentials. The mere fact that an expert does a study, and a group of people append that study to their cultural narrative, does not put the matter beyond the realm of further inquiry and rational doubt. To believe that torture never yields useful information is to believe that those who employ it are either too stupid to notice or too sadistic to care. While I am no admirer of the KGB, the Gestapo, the Kempeitai, or the CIA, one must have a decidedly peculiar view of history to imagine the depredations on their unfortunate victims never yielded at least enough correct information to justify the trouble of asking questions. If one’s goal is simply to produce terror or to vent sadistic impulses, asking questions serves no purpose.

That much of the American left believes the neurobiological argument without question should not be surprising. It fits their view that military and intelligence people are by nature simply stupid and sadistic. The neurobiological argument is dangerous, however, because it abandons the difficult but worthwhile moral and political arguments for a rhetorically powerful but empirically vapid claim. And even if that claim were true, as a basis for judgment, it would still have very dangerous implications. If efficacy were the real problem, would torture become acceptable if a drug could be administered which improved the victim’s memory? Is the problem with “enhanced interrogation” that it is an affront to basic decency – or merely that it isn’t enhanced scientifically by someone with the proper credentials?

It seems that plain empirical truth does not have many friends these days. Fortunately, it needs none.