Interesting Links for 17-12-2014

[andrewducker]

• The good horror movies of 2014
  (tags: movies horror recommendation )
• The supposed dilemma over torture doesn't exist
  (tags: torture )
• Skeptics, deniers, and contrarians: The climate science label game
  (tags: science globalwarming )
• Etsy is now selling pills that make you poop glitter. (Merry Christmas!)
  (tags: faeces awesome funny )
• Idiot fans try

Comments

[steer]

So you may have spotted I was less than 100% convinced about that "The supposed dilemma of torture" article. Here's my thinking. I should say at the outset I am 100% opposed to torture in any real-life situation

[andrewducker]

I can see where you're coming from now.

But I think there are two prongs here:
1) The "Torture is immoral no matter what" argument.
2) The "Torture does not work" argument.

(1) fails if dealing with a pragmatist who can imagine situations where the alternative is even more immoral to them.
(2) fails if there are circumstances where torture does work.

Therefore, in order to be as effective as possible, you need both prongs.

(Plus, probably, also the "Even if it works in the immediate sense, it alienates people and creates more terrorists in the long-term." prong.)

[steer]

(2) is far too strong an argument because it only really takes one case to break it. The CIA's William Francis Buckley might be one such case

[andrewducker]

You can always fall back from an absolutel form of (2) to "Torture works extremely badly, and produces incorrect information more often than it produces useful data, and as an investment of time/energy is of very low value".

(But yes, I prefer my approach too)

[naath]

I think to say something "works" it has to work more often than "once, ever".

If you want to act on the torture evidence you need to be convinced that such evidence is mostly right.

If such evidence is ofter/usually a pack of lies made up to make you stop hitting them then you might as well have used a random system for picking names (or whatever other details it is you are allegedly extracting).

So, sure, sometimes you get names and go kill a bunch of CIA people. But sometimes you get names and go kill a bunch of totally innocent (of CIA involvement) people and spend lots of resources and piss people off with you *to no good end* (assuming "fewer CIA people" is a "good end").

[steer]

I think to say something "works" it has to work more often than "once, ever".That's because your mindframe is that of someone who thinks torture is abhorrent and shouldn't happen (as is mine). If your mindframe is someone who is OK with it then they will think "OK, it worked then, will it work now." So "works once" is probably good enough for that person. It's certainly good enough for them to dismiss someone saying "torture never works" as wrong. And, of course, one thing in that report was that the CIA was lying about effectiveness. If you're that way minded then, of course, you believe it does work because people have told you it does work and you've likely spoken to people who've told you it does work and that the report in question is a pack of lies

[naath]

But to know that it even *actually worked* that once you'd have to find some confirmation that those dudes were CIA dudes.

Obviously if some liar has told you "sure they were" then maybe you believe that (or rather, maybe some people would believe that)

The point about the pack of lies is not because "torture is awful" (it is) but because that is one of the ways torture *doesn't work*. Sure, if the torturee just stays silent (or swallows a cyanide pill, or whatever) then that's a really *obvious* failure of your torture - but when you get what looks like an answer, that doesn't mean you *got* an answer.

It's like if you have a "diagnostic test" that says that *everyone* has breast cancer, well, sure - you diagnosed some breast cancer (some people have it) but it didn't WORK because of all those people you said have breast cancer when they don't.

[steer]

Have a look at what I say in the other thread about the situation of getting what looks like an answer...

[andrewducker]

Oh - also, you probably missed this:
http://andrewducker.livejournal.com/3202829.html?thread=31429645#t31429645
which led into this post by Nancylebov:
http://nancylebov.livejournal.com/628049.html

Where I hopefully made it clear that even if torture worked I would still absolutely not be in favour of it.

[steer]

I presumed that this would, in fact, be your attitude but it's good to have it confirmed.

[kalimac]

I don't think anybody except those engaging in extremely sloppy wording ever said that torture never gets you accurate information. What I've always seen said is that it's not reliable. A high percentage is bad, and you can't tell what's bad or good. That means even the good information is useless. Unless you don't care what happens when you act on your unreliable information, and the same people who don't care whether who they're torturing is innocent probably don't care about that, either.

A more mundane parallel would be getting information from Wikipedia. A lot of what's on Wikipedia is thoroughly accurate. But woe betide any scholar who relies on it.

The sex education parallel doesn't work because the dynamic of stuffing information into brains is completely different from extracting it. Also, there's no torture. The moral question of getting information out of prisoners becomes entirely different if nobody's torturing them for it.

[steer]

The comparison with sex education was not intended as a parallel for extracting information my point was, pick some idea X you like, really any idea that requires practitioners and subjects. If you read a report saying
"1) Idea X was implemented by unskilled and unqualified practitioners on this occasion."
"2) Idea X was implemented often on people where it could never succeed on this occasion."
"3) Idea X failed on this occasion."

Now, for any idea whatsoever that you're keen on would you ever say as a basis of that report "Oh dear, I was wrong, idea X doesn't work, what a shame, I liked it so much"? No, you'd say "How ludicrous, idea X can still work perfectly fine with well trained people and the right subjects. What kind of person would think this is evidence that idea X doesn't work?"

It's sloppy thinking, confirmation bias. You only accept it because you came to the report already believing torture doesn't work in practical circumstances (as did I).

A high percentage is bad, and you can't tell what's bad or good.

[kalimac]

No, it's not that arguing from principles leads to problems. It's that you're using the wrong principles. Bad information exists in all areas of life. Most of the time you can take that into account. But the whole purpose of the "ticking bomb" torture scenario is to force you to rely on the accuracy of the information given. If it's not accurate, you're wasting your exceedingly valuable time.

[steer]

Yes. But the point is, that you make some statement and say "torture doesn't work because of X"... that's easily followed with "Great, we'll torture when X is not true".

So, "ticking bomb" scenario where it's hard to verify information, no go, don't use torture. Other situation where information once obtained is easy to check then do use torture. I suspect this is not what you actually think.

[kalimac]

It's true that the unreliability of torture-acquired information is less critical in other scenarios. But that's irrelevant, because there's more than one principled argument against torture. I was just focusing on this one.

The whole reason the "ticking bomb" scenario was invented was as a way to bypass or short-circuit all the other well-known arguments against torture. But it is the most susceptible to this argument.

[steer]

But that's irrelevant, because there's more than one principled argument against torture. Yes, that's my real point. The principled argument is the one that should be stuck to because the pragmatic "it doesn't work" is susceptible to people finding circumstances where it does