Why torture doesn't work: A definitive guide
In the wake of the Senate report cataloging a whole lot of torture committed by the CIA, Dick Cheney has been reduced to arguing that torturing people — even innocent ones — is worth doing if you eventually get good results. The ends justify the means.
I can see why he makes this argument — he's simply got no other option. It is now obvious that what the CIA did was illegal, brutal torture. Claims that it kept the nation safe are all that Cheney has left.
But Cheney is wrong: Torture doesn't work and never has.
I have referenced the work Torture and Democracy, by Darius Rejali of Reed College, many times in the past. It is widely agreed to be a benchmark work on torture — perhaps the most thorough investigation and analysis of the subject available. Here's what Rejali says, to put this question to rest for all time.
Over 12 years of research, Rejali examined the use of torture in the U.S., Great Britain, Nazi Germany, Soviet Russia, South Vietnam, and Korea. He looked at torture inflicted during the French-Algerian War, as well as at the Abu Ghraib prison in Iraq and at Guantánamo Bay. His research found that there is no record of any successful use of torture to gather intelligence, not even in totalitarian states.
The reasons for this are more complex than is popularly supposed.
The first has to do with the nature of pain itself. Causing someone pain is not like turning a dial on a stove. Greater damage to the body often translates as less pain, since the body, in shock, shuts down the pain system (as victims of car accidents or shootings can often attest). Going too far, too fast with torture can simply desensitize people or cause them to black out. Furthermore, different people have different thresholds for pain, and they use certain types of pain to mask other ones. As a result, even with technological assistance, it is simply impossible to torture in any scientific, reproducible way.
Torturers understand this, and so are drawn to two blunt techniques: (1) apply maximum allowable pain, so as to push past all limits and (2) vary the torture methods widely to exploit as many phobias and specific weaknesses as possible. One perverse result of this is that there will be constant pressure to ignore limits set by the law in favor of a maximum diversity of pain.
Second, torture badly corrodes organizations that practice it. One of the most striking aspects revealed by the Senate report is the incompetence of the CIA, the kind of stuff that would bankrupt a lemonade stand, like losing track of who it had in custody. This is not a coincidence. As Rejali writes, in agencies that torture, "behavioral and organizational indicators show a rapid decay in professionalism."
In addition to a temptation to break the rules, torture regimes encourage competition between interrogators to break prisoners first. All of this rots traditional investigative skills, as interrogators turn to torture as a quick and easy road to success. Why do complicated forensics if you've got a cattle prod? As a result, arguments like the ones Charles Krauthammer once made — that torture works but is a monstrous evil that should be used only rarely and under strict limitations — are not only wrong, but also unrealistic. Once used, torture inevitably spreads like kudzu.
Third, torture directly undermines traditional intelligence gathering. Whether solving traditional crimes or penetrating the enemy in wartime, the cooperation of the public is the most valuable intelligence resource. Even in the communist dictatorships of Eastern Europe, working relationships were critical to keeping the secret police informed. Torture of sources, by contrast, destroys trust and makes normal interrogation dramatically more difficult, which is why those dictatorships kept their torture confined to political dissidents.
Furthermore, what little information is produced under torture is extremely unreliable. Detainees with a score to settle may falsely rat out old enemies, hoping they will be tortured instead. Detainees with no information will sometimes try to appease their torturers with lies, making interrogators waste time and effort chasing false leads. The CIA did just this, in fact. The Senate report documents at least one instance in which the CIA tortured a detainee, who gave them bad information, which led to more innocent people being detained.
Even when prisoners say true things, the interrogators very often do not believe them. This happened to John McCain when he was tortured in North Vietnam. Formal studies show that torturers cannot reliably distinguish truth from falsehood.
Indeed, inflicting pain on someone can directly damage memory. Extreme trauma often short-circuits the process of recent memory consolidation. When a woman called Sheila Cassidy was tortured by the Pinochet regime, she was so broken mentally that she could not remember the address the interrogators wanted. Sleep-deprived victims also become disoriented and confused, and can accidentally convince themselves of things interrogators are suggesting to them, producing more false leads.
A study of the Phoenix Program, a CIA program of torture and assassination of Viet Cong members during the Vietnam War, found that it "victimized at least 38 innocents for every one actual Viet Cong agent"; the ratio is likely closer to 78 to 1. The reason is that the CIA's database was filled with crappy intelligence produced by torture, a result that fits with studies of torture in Northern Ireland and British Cyprus. Rejali concludes:
For harvesting information, torture is the clumsiest method available to organizations, even clumsier in some cases than flipping coins or shooting randomly into crowds. The sources of error are systematic and ineradicable. [Torture and Democracy]
That brings us to the ticking time bomb thought experiment, where someone is known to have information about an imminent attack but will not talk. This is the centerpiece of the pro-torture case. Setting aside the fact that this sort of situation is extraordinarily rare, there is no reason to think time-limited, high-pressure torture would be any more successful than in other circumstances. On the contrary, all the problems with torture identified above are made worse by a time constraint: the techniques are limited, as slow ones must be ruled out; pain must be applied more quickly, thus increasing the risk of blackouts, desensitization, or memory damage; and time wasted chasing false leads becomes an even greater loss.
Of course, the overwhelming amount of evidence will not convince many conservatives. Torture is how the Republican Party projects toughness. The fact that it is worthless for intelligence will be denied, just like with climate change. Instead, the bluster of Dick Cheney and other people deeply implicated in the Bush torture regime and the ensuing cover-up (like Jose Rodriguez, who destroyed CIA interrogation tapes) will be cited as unquestionable proof.
But for anyone with even a scrap of intellectual integrity, the evidence is beyond question. Torture doesn't work. Full stop.