Why Torture Doesn't "Work"

This point in Philip Zelikow's op-ed yesterday in the NYT is essential:

What the committee may well find, after all the sifting, is that the reports were a critical part of the intelligence flow, but rarely if ever affected a “ticking bomb” situation. Yet the main rationale for using extreme methods is to save time. To the extent that the methods are more than just a way of debasing an enemy, their added value is in breaking people quickly, with the downsides including unreliability.

That is one reason the methods of torment do not stack up well against proved alternatives that rely on patience and skill.

As is Daphne Aviatar's rejoinder:

After you’ve tortured and humiliated someone, how can you possibly know what he would have told you if you’d gained his confidence instead?