Hacker News Reader

Comment by anvandare | Hacker News Reader

An example, I'm afraid, of situations that happen all too often around the world. When we look at the justice system as any other abstract system (with positive/negative feedback loops) we find that the actors (the judges, prosecutors, investigators, etc.) are never rewarded for finding the truth. After all, how can you ever know that you found the truth? They are rewarded for finding a truth. For making the picture fit.

And that is why you should never collaborate with (or trust) the system. They aren't interested in what's good for you, they are interested in what's good for themselves: finding someone to convict and thereby pushing their yearly evaluation into the positives.

[ed.: typo]

View on HN

> ...They aren't interested in what's good for you, they are interested in what's good for themselves:...

Very true, but then that's why the aversarial justice system is so important. It's ok for prosecutors and police to use every means legally at their disposal to try and prove someone guilty, as long as that person has an independent advocate with sufficient authority to protect the accused and challenge the prosecution case.

The problem here is that the way these people were treated was perfectly legal and even morally acceptable at the time. Clearly lengthy unecessary periods of solitary confinement, drugs and agressive interrogation without an advocate present should be completely unacceptable. They even used simulated drowning. The paraelells to Abu Ghraib, Guantanamo, the treatment of Bradley Manning and CIA waterboarding are stark. It's pretty clear that these kind of abuses simply do not work.

People subjected to these treatments will say, and even come to believe absolutely anything, so any information that comes out of such processes is worthless. That's completely aside from the deep immorality of using such techniques in the first place. ALl it does is provides moral cover and justification for oppressive regimes to use such techniques completely indiscriminately.

Is the adversarial system really any better or worse than the inquisitorial system (I believe they use the latter in Iceland):

http://en.wikipedia.org/wiki/Adversarial_system

http://en.wikipedia.org/wiki/Inquisitorial_system

As the inquisitorial system came out of the inquisition NO is the obvious answer.

<snip> agressive interrogation without an advocate present should be completely unacceptable <snip>

LOL, that is the standard in much of Europe today. That is of course not incongruent with what you said - it just nuances your 'at the time'.

It doesn't make sense to me to say that torture is ineffective simply because subjects have incentive in the face of torture to say or believe anything.

For example, it doesn't matter if a person keeps on outputting contradictory claims of reality. What matters is if this person can produce evidence of novel knowledge. And of course even if someone delivered information in a non-tortuous interrogation, any government would fact-check.

I also don't believe it's very easy to confuse someone's existing memories. I don't think I can simply torture somebody, with professionals and equipment at my disposal, and cause discrete changes in beliefs ("My real name is Reek!").

These world governments no doubt have data on varying techniques of torture, and I have no doubt there has been a very long race, begun awhile ago, to map out the theory of torture, and I think it is a bit preliminary to say that torture doesn't work until we get a hold of some systematically collected data.

> What matters is if this person can produce evidence of novel knowledge.

Except that in most cases you can't distinguish novel knowledge from made-up bullshit with real certainty. If you could, you wouldn't have to torture them. And don't bother dredging up contrived ticking-bomb scenarios where it's different - yes, they exist, but they are anything but common and even then often not clear cut.

And that's not even touching on the fact that you almost never have 100% certainty that the suspect actually knows what you want them to tell you at all.

> And of course even if someone delivered information in a non-tortuous interrogation, any government would fact-check.

No. Most of them wouldn't, when they've just been told what they wanted to hear. Or (worst but common case) they'll fact-check by arresting the people the victim just named and torturing them as well.

> I also don't believe it's very easy to confuse someone's existing memories.

Your belief is utterly wrong. Read the goddamn article. And that wasn't even using physical torture. Heck, people's memory often enough get confused all of their own. People suck at remembering events clearly.

> These world governments no doubt have data on varying techniques of torture, and I have no doubt there has been a very long race, begun awhile ago, to map out the theory of torture, and I think it is a bit preliminary to say that torture doesn't work until we get a hold of some systematically collected data.

Hopefully that will never happen (it pretty definitely hasn't happend so far; some of the Nazi medical experiments may have come close but had a different focus), the consensus among people with actual experience seems to be clear: intel gained through torture is of worse quality than that gained through other interrogation techniques, and employing torture has many negative effects on your own side.

http://www.ritholtz.com/blog/2011/05/interrogation-experts-f...

What I had in mind when I was thinking of novel data, would be someone who can authenticate the value of their potentially false claims by stating something that nobody else could know. Like the serial killer who states special information. An alternative situation could be those who are looking for cryptography keys. I still believe that the presence of novel data could help a torturous interrogator sieve desperate lies from actionable data.

Confirmation of novel data that nobody else could've known authenticates the value of your knowledge. Government may have been reckless before by acting before confirming, but if they just include a confirmation step, then they could have a means of differentiating actionable data from junk.

After looking over the site, and doing some Wikipedia searching, it does appear that there are ways to reliably damage or confuse memory, even in innocuous settings. Such as ads meant to manipulate memory.

The last point is the one I wanted to make, because I felt that arguments about the efficacy of torture is similar to arguments about the financial costs of the death penalty -- it's a highly ephemeral and gambling argument that hinges on the state of science or technology.

So I was wondering what happens to the discussion on ethics once we begin to accept the efficacy of torture?

It's trivially easy to implant false memories. Here's Aza Raskin doing it to an audience just using a single question: http://vimeo.com/15886853#t=21m27s Also, you should look up Beth Rutherford's story. Edit: Oh, her story is actually earlier in the video. I guess it's pretty famous by now.

"It's ok for prosecutors and police to use every means legally at their disposal to try and prove someone guilty, as long as that person has an independent advocate with sufficient authority to protect the accused and challenge the prosecution case."

I don't necessarily agree with that. The defense should be using every means available, the prosecution should restrict themselves to things that get at the truth. As extreme examples, presenting fabricated evidence or withholding exculpatory evidence would clearly be unacceptable even if laws were changed to allow it. I am not remotely confident that laws proscribe everything they should proscribe.

The problem isn't so much centered on the "shoulds"/ideals it's centered on reality and a more game theoretic approach. Instead of everyone crossing their heart to be good... how do you make the desired outcome have the least friction in being realized?

I don't think the game theoretic approach is wrong, but you're ignoring that threat of censure (explicit or implicit, formal or informal) is an important part of the system, and in that dimension the informal ideals and "shoulds" are highly relevant, and need to be factored into the game theoretic analysis. If a prosecutor knows that they'll get accolades for exploiting a new way to screw defendants, that's a different system than if they know they'll face condemnation for it, even if there's nothing formal.

This is an interesting insight into the dangers of ill-chosen metrics. It could only happen in a system in which the police were expected to produce (and were in various ways rewarded for) convictions instead of e.g. a maximally accurate presentation of events that can be used to decide how society should respond.

Convictions are far easier to measure (and achieve, of course). How would you go about constructing an efficient, corruption-resistant system that is measured by its accuracy? It seems like a hard problem, but not necessarily an unsolvable one.

Debugging support issues sometimes feels like this. If you go into a complex system looking for bugs then you'll find plenty - but not necessarily the one that's causing the current issue. What you're really looking for is to fully understand the reason the system does what it does. And then you can change it.

Also: further evidence that information produced under torture cannot be relied upon, yet our states still do this. Perhaps sometimes it works? Or is it just very successful in producing "information" that the system wants and expects to receive, which is then interpreted as an objective success that validates the method?

Kind of difficult to study that.

Perhaps an "ends cannot be used to justify means" principle is simply a necessary precondition for benevolent civilization.

A nice formulation of the first part of yout post is Goodhart's Law: If you use a correlated metric to measure something qualitative, people will optimize toward the metric, and it'll stop correlating.

Some reading:

http://en.wikipedia.org/wiki/Goodhart's_law

http://lesswrong.com/lw/1ws/the_importance_of_goodharts_law/

See also: the breakdown of the link structure of the web because of the rise of google.

> Also: further evidence that information produced under torture cannot be relied upon, yet our states still do this. Perhaps sometimes it works? Or is it just very successful in producing "information" that the system wants and expects to receive, which is then interpreted as an objective success that validates the method?

Both, really. There is certainly a small class of scenarios where it does work, but you cannot know in advance where your case is in that class. And is great at making it look like you're "getting results" if you don't look to closely.

Another factor: torture is definitely good for maintaining a "reign of terror" style of regime. Until it isn't. I think there's even a passage in Macchiavelli's "The Prince" (i.e. the opus that basically coined the "end justify means" meme) that says ruling people by making them love you is most effective but impossible to do all the time, so ruling through fear is the next best thing - yet one should carefully avoid being hated.

> Perhaps an "ends cannot be used to justify means" principle is simply a necessary precondition for benevolent civilization.

As a basic principle, definitely. Though no simple rule can solve all moral dilemmas.

Reply on news.ycombinator.com