Does torture work? As Bill Clinton might have said, it depends on what you mean by “work.” If your goal is to extract a confession for use at one of Stalin’s show trials, then torture works well enough. If, on the other hand, your goal is to get reliable intelligence out of a person, to get him for example, to tell you the details of a potential terror plot or name his confederates, then just as clearly torture does not work. A person being tortured is liable to say just about anything to get the pain to stop and isn’t apt to be much of a stickler for whether or not his statements are true. Since there is no way to separate the true from the false screams, answers given by a suspect under torture are worse than worthless and should never form a part of our interrogation policy.
It is true that torture could sometimes produce accurate information. So can a magic eight ball, but no one would think to build our counter-terrorism strategy around the answers it gives. If I ask the magic eight ball whether it will rain tomorrow the answer it gives might be accurate, but it won’t be reliable. Magic eight balls are sometimes right, but are very often wrong, and since there is no way to tell if an answer it gives is accurate or not its answers are never reliable and only a fool would rely on them. So also with torture. If one scours the Internet, one might find a few cases in which the use of torture produced accurate intelligence (though not nearly as many as you might think). But so what? Police sometimes employ psychics to help them solve crimes, and I’m sure there have been a few cases where information provided by the psychics have led police to a breakthrough in the case. This doesn’t mean that we ought to be employing psychics in the war on terror. You might as well argue that the lottery is a good financial investment because people do occasionally win it. We can only get the benefit of the information torture produces if we are willing to rely on it without knowing whether or not it is true. Since this is self-evidently a bad policy, the only logical choice is to forgo the use of torture altogether in favor of more reliable interrogation techniques.
People have sometimes tried to get around the reliability problem by demanding that any information gained via torture be checked out and corroborated before it is considered at all reliable. If a suspect says that the bomb is in a certain location, you go and look in that location and check to see if what he’s saying is right. If he says ‘the enemy will attack at dawn from the east’ and you have no way of checking whether what he’s said is right, then you disregard it completely. It’s a nice theory, but in practice things don’t seem to work that way. Invariably people are unwilling to ignore the intelligence gained under torture even when it is unverified, and this often leads them to do many stupid things. Part of the pre-war intelligence connecting Iraq and al-Qaeda came from information produced under torture, for example. Reason may tell us to ignore unverifiable information gleaned under torture, but experience shows that the temptation to act on that information is just too strong to resist.
Even where information produced under torture can be checked out, it often isn’t worth it to do so. The “trust-but-verify” approach also leaves itself open to manipulation. In Vietnam solders sent to check out information gained under torture often found themselves walking into ambushes, and in Algeria captured members of the FLN were told that if tortured they should give up the names of members of rival (and more moderate) political groups.
Non-torturous techniques are not only more in keeping with our ordinary notions of human decency, they are also more effective. The most effective interrogator at Guantanamo (according to the interrogators themselves) is “an older woman who adopts a nurturing attitude.” During World War II, the British were able to identify all but three of the hundreds of German spies operating in the country without the use of torture. Many of these spies became double agents, giving false information to the Germans and likely saving the lives of tens of thousands of civilians. Traditional interrogation methods can work even under extreme circumstances. The NYPD secured actionable intelligence from a suspect in the millennium-bombing plot in just six hours on December 30, 1999 – all using standard interrogation methods.
Not only is the use of harsh methods unnecessary, but it can be counter-productive. One of the things that upset the FBI about Guantanamo is that it seemed like whenever they were making headway with a particular detainee, developing a rapport and getting him to divulge information, the CIA would step in and start using more “forward leaning” methods, instantly eliminating whatever progress they had made.
Defenses of torture typically rely on highly improbable hypothetical situations. We are to assume that we know (somehow) that a plot involving massive loss of human life is about to occur and that we know (somehow) that a certain person has information that could stop the plot, and that we know (somehow) that nothing short of torture will get the information out of him. This is only a couple of steps away from asking us to suppose that we could know whether a particular statement made under torture is accurate. It is, in fact, a recipe for error, since the more extreme and unfamiliar a situation is, the more likely that we will fail to take key elements into account in our moral analysis.
But even if torture would be justifiable in such an unlikely scenario, this wouldn’t mean that we should carve out an exception to the law in such a case. Most of us would say that civil disobedience can sometimes be justified. This does not mean, however, that a formal exception should be written into every single law permitting people to break the law when (they believe) they are justified in doing so. Such an exception would have to be written in general enough terms that it would end up swallowing, if not the rule, then at least so many cases where it oughtn’t apply as to make the rule itself unworkable.
So also with torture. We may wish to restrict torture to cases where there is a ticking time bomb, but it will soon come to be used where there is no bomb at all. We may want to limit it to case where we know that the person has the information we need; we will end up using it in cases where we only suspect it – all will end up torturing the innocent and the guilty alike. (We have released hundreds of people from Guantanamo Bay, some of whom claim to have been tortured. Either we have put a lot of terrorists back on the streets, or we have mistreated a lot of innocent people). We may mean to use torture only as a last resort; we will find we are resorting to it at the first sign of frustration. This is not idle speculation. When Israel legalized torture in the 1990s this is exactly what happened. What was intended as an extraordinary method became widespread and routine, and all attempts to put real limits on it failed.