Torture Sometimes Works
Or:
Judo'ed by Krauthammer

I've been meaning to write a post on this for awhile now, and this Krauthammer post prompts me to at least say something brief about it.

There can't be any real doubt that torture sometimes works. That is, that there have been, in mankind's torturous history, at least a few cases in which torture has elicited the desired information. How many acts of torture have there been in the history of humans? (Just stick with H. sapiens sapiens here...) Millions? Hundreds of thousands? It would be foolish to deny that hundreds or thousands of such acts have worked--that is, elicited the relevant information.

If we limit our attention to people who actually know the relevant information, and diligent torturers, is it at all plausible to suggest that torture works less than, say, one percent of the time? Such a suggestion sounds absurd to my ear, at least.

(Note: I'm intentionally not looking for any scientific evidence on this right now, in part to make a point about what we can safely assume in cases like this.)

Now, much of the current liberal chatter about torture seems to come dangerously close to suggesting that torture just flat-out never works.

And that's almost certainly false.

There really can't be any significant doubt that that's false. It would be a kind of miracle if it never worked.

The relevant point seems to be that it cannot be relied upon to work.

(Or, rather: that's one point. We're ignoring the straight-forward moral questions here.)

As we now have very good reason to believe, torture is not only an unreliable means of extracting truth, but it is actually a fairly reliable method of generating falsehoods. And that makes it unreliable. Even if the victim of torture does utter some truths, he is likely to have uttered many falsehoods, and we are not in a position to tell which is which. And even if some of the tortured utter mostly truths, we aren't in a position to distinguish them from the tortured who utter mostly falsehoods.

(And, of course, we should feel disgusted that we're even thinking and talking like this, IMHO...)

Now, in the chaos of the internet, people who want fast and clean routes to their preferred conclusions have a tendency to complain about being corrected in such ways. But it is very, very important that we be clear about points like this.

Why?

Well, first because it's just important to be precise about such things in general. But, second, because being precise here helps to avoid being judo'ed by cherry picking fallacies. (Which fallacies, one might reasonably argue, have characterized the Bush era: we can't count the ballots because some of them are unreadable; we have to attack Iraq because some evidence indicates that they have WMD and some evidence indicates links with al Qaeda.) And now, of course--and explicitly with Krauthammer:

It was o.k. to torture because it sometimes works.

Now, many who have already virtually committed themselves to the proposition that torture never works will be inclined to deny Krauthammer's premise, when what they should do is question the validity of his inference: from it sometimes works (i.e. it sometimes produces true information) it does not follow that it was permissible that we did it. Even ignoring all other factors and considerations here, we'd at least need something more like: it usually works/usually produces true information.

Now, taking this route is going to make the argument a lot messier in more ways than the obvious ones, but that's the price of attending to the facts.

Currently, I worry that liberals are going overboard in their anti-torturism. It never works...it's always wrong... Witness the anger it provokes to even discuss ticking-time-bomb cases.

The most important point is that I think liberals are moving toward a false moral position. But in case you're not worried about that: you're about to get judo'ed by conservatives. They're going to use the momentum of your own exaggerated arguments against you. Like Krauthammer just did.

I myself allowed myself to be shamed into shutting up about ticking-time-bomb cases. In particular, in response to disapproval, I allowed myself to say that such cases "never happen," when I knew that the truth was that they were probably approximated to a greater or lesser degree by some actual cases. And then Krauthammer, in effect, Judo'ed me by pointing to the Waxman case.

Now, I don't think that the Waxman case actually shows much about our own policy of torture...it wins Krauthammer the point only against dialectically over-extended and off-balance liberals...but that seems to be most of us these days. I won't try to discuss the Waxman case at length here, but I've thought a bit about such cases of late. My vague thought thus far is that there is an important difference between a case in which a prisoner might know when we might be attacked and a case in which a prisoner is definitely complicitous in an on-going crime. Roughly my current, gut-reaction position is: don't torture for me just because you think there might be an attack against us somewhere, some time; but if I'm buried in a box somewhere and one of the kidnappers is sitting in front of you, you better damn well make with the tough stuff. (And, needless to say, that goes for anyone in the box.) I'm not sure this will turn out to be a defensible position, but it fuels my hunch that the Waxman case doesn't tell us as much as Krauthammer thinks it does.