Thursday, December 11, 2014
An NPR interview this morning made two interesting points about the recent report on torture by the CIA, the first of which I had thought of myself. This was that it is possible that some authors of the report, or perhaps just those who summarize it, are engaging in belief overkill. They are saying that the extreme methods of interrogation never did any good. They don't need to say this. They could just say, more cautiously, that the amount of good done was so small as not to justify the use of these methods. Belief overkill is the distortion of beliefs so that all conflict disappears and no arguments oppose the favored conclusion, even if the conflict would be easy to resolve.
The second point, made in passing, was that it may in fact be impossible to find a case in which the extreme methods led the CIA from ignorance to certainty all at once. But intelligence does not work like this. It is all about probabilities. (The interview did not use this term, but it was implied.) If a confession extracted under extreme pressure changed the probability of some conclusion, such as an immanent attack or the presence of a terrorist at some location, the probability may change enough to warrant a change in plans. Such changes could happen if a confession agreed, or disagreed, with other evidence already available.
Given the possibility that probabilities were affected, the difficulty of claiming that these methods "never mattered at all" seems to increase. Again, this claim is not needed. The methods are bad in themselves and in the precedents they set for others.