John Oliver discusses torture

[theprestige]

Who here has argued that we should use torture?

I haven't argued it here, but in general I think that we "should" use torture in the same way that we "should" kill civilians; i.e., according to principles similar to those applied by international law on the question of collateral damage in warfare. But even this leaves difficult questions unanswered. Reasonable people can disagree over the military distinction of non-state militia fighters, the military advantage of stopping an attack whose probability is only guessed at, and the proportionality of any given harsh interrogation method, for example.

Even if I think it has a principled place in the toolbox, I would understand if it were prohibited due to lack of answers to those difficult questions.

 

[marplots]

Torture is used on a daily basis across the US to encourage certain behaviors. The police call it "pain compliance techniques."

Is it OK to suggest that torture works well for that purpose?

 

[Belz...]

It all rounds in one's expectation of the result of acquiescence to the interrogator's demands. In your scenario I assume the expectation is that once they get what they want the captive will eventually be released.

Yes, the possibility that you wouldn't expect to be released has crossed my mind. However I'd think that, were I in that scenario, the hope that this will happen would facilitate my compliance.

But you're right: the mindset of an actual terrorist might very well be vastly different, and this will alter the torturer's expectations of truthfulness, perhaps to the point that, in that person's case, torture will be entirely inefficient.

 

[Upchurch]

Torture is used on a daily basis across the US to encourage certain behaviors. The police call it "pain compliance techniques."

I may be wrong, but I was assuming we were talking about using pain techniques on captive prisoners. Using painful techniques to subdue people who are not in custody doesn't really fit into that definition.

 

[theprestige]

Torture is used on a daily basis across the US to encourage certain behaviors. The police call it "pain compliance techniques."

Is it OK to suggest that torture works well for that purpose?

Seems like a textbook demonstration of the efficacy of torture, and the proper method for effective torture: Apply the technique. Test the response. Stop applying the technique when the test result is satisfactory.

Everything else is just figuring out what results justify the method, what methods are proportional, and whether an effective and timely test of the response is practical.

 

[Beelzebuddy]

You are demanding too much of the method. First of all, as Zuggurat pointed out, the same criticism applies to any interrogation method used. It even applies when someone is entirely willing to tell you the truth - they can still be mistaken or ill-informed or biased.

And that's why the second part - consilience - rules the day. We do it in areas completely divorced from torture - we look for converging lines of evidence to give us confidence that can't be obtained with one line of evidence alone.

I reject the picture where "tell me anything and the torture stops" is the rule. You'd instead use torture as a punishment or a threat for lying, not in a "make it stop now" way. I'd wager that the most effective torture is the torture that never actually happens at all.

What I hoped to accomplish with the post was to shift the conversation from "torture obviously works because it is capable of returning a single true positive along with a sea of type 1 errors" to a question of accuracy. I don't really care how you or Ziggurat thinks it ranks against other methods of information gathering, nor what you think the "right" way to torture someone should be, so long as you continue to consider the signal to noise ratio while doing so.

 

[p0lka]

torture is a tribal trait,
the ability to switch to us and them,
I hope we all grow up soon.

 

[Upchurch]

torture is a tribal trait,
the ability to switch to us and them,
I hope we all grow up soon.

We used to be, on this issue. At some point, we un-grew up.

 

[Chuck Guiteau]

I think that there are actually two separate questions here:
1.Does torture provide an effective means of intelligence gathering?
2.Should torture be used as a tool of intelligence gathering?

Look at it this way:
1. Does the A-bomb work?
2. Should the A-bomb be used?

 

[marplots]

I may be wrong, but I was assuming we were talking about using pain techniques on captive prisoners. Using painful techniques to subdue people who are not in custody doesn't really fit into that definition.

Don't they use the same techniques to gain compliance after the arrest as before? For example, getting a prisoner to submit to commands.

 

[Allen773]

This topic is torture.


(not that I disagree with John Oliver).

 

[Upchurch]

1.Does torture provide an effective means of intelligence gathering?

I would suggest rather:

1. Does torture provide a sufficiently more effective means of intelligence gather over more conventional means?​

It isn't enough that torture works or even that it works as well as conventional interrogation techniques. I'd argue that it isn't even enough if torture worked slightly more efficiently.

Most of us seem to agree that using torture is a bad thing, so torture would have to be significantly superior in gathering accurate intelligence in order to justify its use.

 

[jaydeehess]

Seems like a textbook demonstration of the efficacy of torture, and the proper method for effective torture: Apply the technique. Test the response. Stop applying the technique when the test result is satisfactory.

Everything else is just figuring out what results justify the method, what methods are proportional, and whether an effective and timely test of the response is practical.

Is there backing to the claim?
Again the expectation of the captive would come into play, including suing the police.

 

[sunmaster14]

Monketi Ghost and sunmaster14 (there are many other versions of his pro-torture argument in other threads)

I have never argued that we should use torture.

 

[marplots]

What I hoped to accomplish with the post was to shift the conversation from "torture obviously works because it is capable of returning a single true positive along with a sea of type 1 errors" to a question of accuracy. I don't really care how you or Ziggurat thinks it ranks against other methods of information gathering, nor what you think the "right" way to torture someone should be, so long as you continue to consider the signal to noise ratio while doing so.

I don't actually know what the signal to noise ratio is. I suspect it varies by individual and the torturer could decide whether the technique is helping or not. This would mirror other policies where the relevant expert gets a say in the specific application.

But there's more than just raw accuracy involved. With consilience, you use every method you can think of - statements from other prisoners, documents, intelligence gathering, statements made without torture and with, circumstantial stuff. It all gets pooled together to try and draw a coherent picture.

Ideally, the entirety of the process yields the best signal to noise ratio, not any one technique standing alone. My only objection is evaluating torture only as a standalone test for veracity, when I don't think any testimony can meet that standard.

The better question is then whether or not having torture in the stew is worthwhile. That's much harder to answer. You need a control condition and hopefully blinding.

It's a mess the other way round too. If I torture someone and they tell me what I want to know, I still can't say torture "works," at least not exclusively - I might very well have gotten the same information with a bribe or a threat instead.

 

[jaydeehess]

It's a mess the other way round too. If I torture someone and they tell me what I want to know, I still can't say torture "works," at least not exclusively - I might very well have gotten the same information with a bribe or a threat instead.

Or faster.

 

[Monketi Ghost]

Or faster.

I'm quite sure bribes are attempted first.

 

[marplots]

I'm quite sure bribes are attempted first.

For $20, I'll agree.

 

[Fudbucker]

I just think of myself: no matter how friendly an interrogator acts, he's not getting my PIN number. But as soon as he starts pulling my fingernails out (and threatens to cut fingers off if he catches me lying), I'll tell him what he wants to know.

 

[Fudbucker]

Again, I don't think anyone is saying that you can't get information from people using torture. What people are saying is that there are more reliable and faster ways to get the information than torture. And that there is both scientific and logical evidence to support that.

But suppose you're an agent in Nazi Germany, and you're trying to find out the plans for a coming offensive, and you get caught. No matter how friendly your Nazi interrogator is, would you ever reveal the names of other operatives you know that are working undercover in Germany? I don't think you would because you hate the Nazi regime (and by extension, it's interrogators).

But what if they did as Zigg was suggesting, and started asking you some questions they do and don't know the answer to, and torture you for every wrong answer? You might give up those names, if the pain is bad enough, and you think it will stop the torture.