Does torture work?

[The Minsky Moment]

It does if your goal is getting someone to tell you something you want to hear.

Which is a huge problem in the real world where there is always temptation to use intelligence info to support pre-existing conclusions and policies.  I.e. we have this huge national security problem with witches curses, we know this because when we torture the witches, they confess their diabolical curses. 

[Josquius]

I better general question would be something like "Ignoring the moral implications, can torture be an effective means under some circumstances of extracting useful information?"

I think the answer to that question is "Yes".

I mean I guess that's better as it has an answer choice that makes sense (unlike current poll) but a little too akin to Der's silly poll that had only one real answer.

Yes.  That question only has one answer. Too obvious.
In this one we are asking does it get a accurate result in more than half of cases.

[The Brain]

It's effective at extracting confessions, and torturers often gain a good deal of enjoyment from it. It presumably also discourages people from opposing the torturing authority.

Indeed. My impression is the answer is yes.

[Razgovory]

Question is vague.  Work doing what?  It certainly hurts, so in that sense torturing works.

[CountDeMoney]

Nine out of 10 ex-girlfriends say Yes.
One was unavailable for comment.

[grumbler]

Of course torture works.  If you try hard enough to hurt someone, you will succeed.

[Admiral Yi]

I suggest anyone interested in this question dig up Mark Bowden's excellent article from a few years back in the Atlantic.

[Tamas]

Well if you use enough torture SOMEBODY will tell you what you want to hear, that's for sure. I guess that makes it highly efficient.

It does if your goal is getting someone to tell you something you want to hear.

If the goal is to actually extract actionable intelligence "what you want to hear" is both not useful, and usually not even known to the person being interrogated.

Thanks for explaining my point 

[dps]

Of course torture works.  If you try hard enough to hurt someone, you will succeed.

I dunno.  If a lot of posters here aren't any better at torture than they are at setting up polls, attempts at torture by them might actually make their intended targets feel better.

[Berkut]

I better general question would be something like "Ignoring the moral implications, can torture be an effective means under some circumstances of extracting useful information?"

That's a better general question but still not a good one.  It's a very low bar.  Just about any technique can be effective under *some* circumstances, e.g. the broken clock tells the time correctly twice every day.  There are whole categories of game theoretic problems where the optimal strategy is to randomly vary responses, but that doesn't mean that a magic 8-ball or a quarter is a useful decision making tool. 

It's true that torture is likely to yield information - the question is how useful or reliable that information and what cost is to be assigned to the receipt of unreliable information.  If it turns out ex post that you can point to individual instances where the information turned out to be useful, that doesn't mean it makes sense as a technique to use a priori. 

No argument from me.

I think torture is a pretty terrible idea*, but I think the attempt to pretend like it is cannot be effective is both morally suspect and intellectually dishonest.

It is a terrible thing to do. Full stop. It is not terrible because it doesn't work. It's efficacy is irrelevant to it's repugnance.

[Berkut]

It does if your goal is getting someone to tell you something you want to hear.

Which is a huge problem in the real world where there is always temptation to use intelligence info to support pre-existing conclusions and policies.  I.e. we have this huge national security problem with witches curses, we know this because when we torture the witches, they confess their diabolical curses. 

Indeed. One of many reasons why we should not torture people.

Or put people in interrogation rooms for 16 hours until they "confess" to a crime just to get out of the damn room.

[Berkut]

Of course torture works.  If you try hard enough to hurt someone, you will succeed.

I dunno.  If a lot of posters here aren't any better at torture than they are at setting up polls, attempts at torture by them might actually make their intended targets feel better.

Why drag poor spicey into this?

[derspiess]

Hey man, that ended up being a fun discussion that went on for pages.  SUCCESS

[Berkut]

Hey man, that ended up being a fun discussion that went on for pages.  SUCCESS

That is OUR success, not yours!

[Barrister]

It does if your goal is getting someone to tell you something you want to hear.

Which is a huge problem in the real world where there is always temptation to use intelligence info to support pre-existing conclusions and policies.  I.e. we have this huge national security problem with witches curses, we know this because when we torture the witches, they confess their diabolical curses. 

Indeed. One of many reasons why we should not torture people.

Or put people in interrogation rooms for 16 hours until they "confess" to a crime just to get out of the damn room.

Lengthy interviews are not torture, and really should not be equated.

But yes, even confessions should generally be confirmed with some kind of hard evidence, or with the confession including details that only the guilty party would know.