Trusting In the Word of Madmen

EDITED rather dramatically to delete an intro that obscured more than it illuminated.

Resolved: Even if extraordinary rendition weren't immoral, it would be stupid.

I don't expect to convince people who don't already agree with me that torture isn't effective--I'm no expert, and I clearly want very badly to believe that it's not effective. But here are three pieces of evidence:

1. Terry Karney's September 29, 4:30 pm comments to this post:

Allow me, as an interrogator, not so long returned from Iraq, with friends who were/are still there and in/were in Afghanistan (among them the author of, "The Interrogators, which book I now commend to you, in light of the questions you are asking) to answer the quesetion:

"IF you knew that a few hours of excruciating nail-pulling..." [that is a quote from another poster]

Not only no, but hell no. First, it doesn't work (and I've gone on, at length on the subject at Washington Monthly, Electrolite, in my blog Pecunium and anywhere else I encounter the pernicious doctrine that torture is a useful tool.

Even, in the rare case, a victim gives up truthful information, the validity is suspect, and as much time as it would have taken to get it from him otherwise is need to verify it. If torture is used as a standard tool then we get a cycle of positive feedback where the torturer leads the victim to the answers he's trying to confirm.

2. Praktile's post, "The Law of Unintended Consequences," about "one instance in which torture achieved its intended purpose, but set in motion a chain of events eventually leading to the fall of the World Trade Center on the morning of September 11th, 2001."

3. A Reuters article entitled "Spies 'lap up' info from torture", which is what prompted this post.

It had occurred to me that even if one accepted that torture was good policy, it did not make much sense to rely on countries like Syria and Egypt to interrogate suspects under torture for us and faithfully describe their confessions. At a minimum, they were likely to exploit it to harm domestic opponents as well as dangerous terrorists.

Great Britain's ambassador to Uzbekistan agrees with me. The Reuters article describes a memo he wrote in July, condemning his country's reliance on evidence gained under torture in Uzbekistan, given to the CIA, and shared with Britain:

"We receive intelligence obtained under torture from the Uzbek Security Services, via the U.S. We should stop," Murray wrote. "This is morally, legally and practically wrong.

The practice "fatally undermines our moral standing. It obviates my efforts to get the Uzbek government to stop torture; they are fully aware our intelligence community laps up the results."....

Murray said the material was disinformation designed to trick the United States and Britain into giving aid.

"TORTURED DUPES"

"Tortured dupes are forced to sign up to confessions showing what the Uzbek government wants the U.S. and the UK to believe: that they and we are fighting the same war against terror.

"I repeat that this material is useless -- we are selling our souls for dross. It is in fact positively harmful," he wrote.

"The aim is to convince the West that the Uzbeks are a vital cog against a common foe, that they should keep the assistance, especially military assistance, coming, and they should mute the international criticism on human rights and economic reform."

Murray said Britain's own spy agency lacked the knowledge to evaluate the material, which it received from the American CIA.

"MI6 have no operative within a thousand miles of me and certainly no expertise that can come close to my own in making this assessment," he wrote.

He described meeting an old man who was forced to watch his sons being tortured until he signed a confession admitting links to Osama bin Laden. "Tears were streaming down his face. I have no doubt they had as much connection with bin Laden as I do."

Comments

[Jesurgislac]

I am not naive enough to argue that there are no possible benefits--that suspects interrogated under torture never reveal useful intelligence that we may use to stop an attack.

I don't think the word for that argument is "naive": I think the word is "realistic".

Fourth reason why torturing suspects is a bad idea: it doesn't work.

Remember Abu Ghraib, and the reports that some of the humiliations visited upon the prisoners there are in fact techniques taught to US and UK special forces? (That’s “taught” as in “direct hands-on experience, giving and receiving.”) It’s supposed to prepare them to withstand such treatment if they’re captured.

What I said at the time was that I happened to know that those reports were correct. I also said I wasn’t going to explain how I knew it. That’s still my position.

What I will say is that there’s an aspect of that training that the media didn’t fully understand. As my source explained it, the exercise is a double-blind setup. Having the designated prisoners learn to resist severe interrogation methods is only half of what’s going on. The other half is that the designated interrogators are told in advance that the prisoners have been given certain pieces of target information. Their task is to elicit that information during interrogation.

The trick is that the designated prisoners haven’t been given that info. However, once the exercise is in progress—I gather it’s unpleasantly realistic—they make it up anyway. These are tough guys. They know they’re supposed to be resisting, but they break anyway, because that’s how torture works; and when they do, they start talking.

At the end of the day, the interrogators will have accumulated a mass of realistic-sounding intelligence on the subjects they were told to investigate. At that point they get to find out that the prisoners had no such information, and that none of the intelligence they’ve gathered is valid.

That is, the designated prisoners are learning what torture can do. The designated interrogators are learning what it can’t do, which is elicit reliable information. (cite)

[Katherine]

I'm arguing less than you may think I am arguing. I believe:
1) in general, torture works badly and worse worse than other interrogation methods,
2) but there are specific occasions when suspects have revealed accurate information under torture.
3) there's no way to be certain of it, but there are probably specific occasions when suspects have revealed accurate information under torture that they otherwise would not have revealed.

On Uzbekistan, note that I don't know whether or not the tortured suspects were "rendered" there by the U.S. Uzbekistan is mentioned as a country we've sent suspects to, but none of the dozen or so suspects whose names I know where sent there. I would guess that most people Murray's memo describes were Uzbek citizens arrested in their own country, but it also speaks to the reliability of intelligence from renditions.

[Katherine]

And the title refers to this quotation from Bush's RNC acceptance speech:

"I faced the kind of decision that comes only to the Oval Office -- a decision no president would ask for, but must be prepared to make. Do I forget the lessons of September the 11th and take the word of a madman, or do I take action to defend our country?"

[Jesurgislac]

there's no way to be certain of it, but there are probably specific occasions when suspects have revealed accurate information under torture that they otherwise would not have revealed.

But it still doesn't work.

If I tell you ten lies and one truth, and you know only that out of the 11 statements one of them is true, you're still no better off than you were before, because you don't know which. In fact, with information obtained under torture, you don't even know that: it's more as if I told you a hundred things, and you think that some of them may be true, but you don't know which, and you don't know how many, and you don't know to what degree any of the hypothetically-true statements are true.

[Terry Karney]

Jesurgislac has the right of it.

Information is only useful if it's provable, and torture doesn't get reliably provable information. It get information the subject thinks the torturer wants to know.

If I have the time to pursue the leads the victim gives up, and verify them, I had the time to 1: break him and get it without the problems torture brings into the picture, or 2: time to get it from someone else.

It is possible, that a person who knows one thing, might be compelled to truthfully tell that one thing? Yes. That's the Dershowitz arguement. We know the guy has the info (how we came to this certainty is never revealed, the a priori assumption is just made, without which this line of reasoning fails) and we know we have a limited time.

Under those circumstances it is possible he will give it up... it's also possible we have to chase down a couple of false leads and come back to torture him some more before he gives it up.

But, if we can't be certain he knows... then what you have is garbage to work with.

And GIGO applies.

TK

[Katherine]

Argh. But I don't even really disagree with him. It's a semantic argument. I only conceded "that suspects interrogated under torture" may conceivably "reveal useful intelligence that we may use to stop an attack."

I did this, quite honestly, because people inevitably bring up a suspect who was tortured and revealed useful intelligence. (Abdul Hakim Murad is the usual example, though the same articles that cite his case as proof that torture works also say that it was not the severe torture but the threats of being sent to Israel that broke him.) The point is that an example like that doesn't even get torture close to passing a cost benefit analysis, let alone providing moral justification.

I hesitated to use the word "benefit" but I thought the rest of the post made it clear that I was using it in a very narrow sense. I take it they don't?

[Katherine]

actually, I bet that the source of the confusion is "intelligence that we may use to stop an attack". I didn't mean an imminent attack or the ticking time bomb hypo. I meant intelligence that leads us to other suspects & leads to us breaking up terrorist cells and imprisoning other members. And as I said, there's no way to know in a given case that a suspect only confessed because he was tortured. In most cases that's probably not true. I assume there are cases where it has been true only because there is unfortunately no shortage of suspects and I assume different people break and different points for different reasons.

[Nathan Hamm]

Katherine, the Uzbek suspects, I can assure you, absolutely were not "rendered" by the US. In fact, after each series of terrorist attacks in the country this year, the US has been instrumental in reducing the Uzbek impulse to round up tons of people. Somewhere in the neighborhood of 10,000 were arrested after the 1999 bombings. 300 were arrested after the first '04 bombings. After the second attack this year, in which our embassy was hit, the US provided investigative assistance.

Anyway, Murray's story strikes me as having a lot of problems, not the least of which has to do with his shaky character. He is close to losing his job and perhaps looking for the save that the UK press gave him before.

[Katherine]

We may have sent a few people there, but yeah--it would be in the realm of two or three, not hundreds. I'll update to make that clear.

[Jesurgislac]

Katherine - thanks for your comment (October 11, 2004 05:59 PM) - light dawns. I think you're right: we may be arguing semantics.

[Terry Karney]

Some of it is my hair-trigger response to anything which can be used to justify torture.

I don't think that sentence was clear, it made it seem that, with the right safeguards, a system might be implemented where the, "benefits," of torture can be made to outweigh the detriments.

I do have to say, that I am flattered to have gotten such prominent play (it's happened once or twice before, on this same issue), I just wish we weren't having the need for such a conversation. It out to be beyond the pale, from the get go.

Keep fighting the good fight.

TK