I suppose we are at an impasse. I don't understand how you can say, on the one hand, that torture may be effective at preventing a greater tragedy (under very limited circumstances), yet not simply acknowledge that if it isn't an option even in those very limited and extreme cases, the greater tragedy might occur. That seems willfully obtuse. Of course none of that should imply that torture should be taken lightly or is somehow benign. It may simply be the best of a set of terrible options at any given moment. Or maybe it should be banned altogether. Either way, there is a cost.

In your repeated sojourns through history's hall of horrors (including, of course, the Nazis), you fail to demonstrate how controlled use of an extreme tactic like torture leads inexorably to the Holocaust or the Gulag. While I am no fan of capital punishment, one might as well say that its employment in cases of particularly gruesome and premeditated homicide leads inexorably to mass murder. Such "slippery slope" arguments are nonsense.

As for the remark about my handle, I would counter that FDR was the wiliest of chief executives in weighing competing moral claims. His machinations to involve the U.S. in WWII (including such acts as cutting off the Japanese oil supply) were anything but forthright, and certainly got a lot of Americans killed in the process, but he understood the greater need of defeating fascism. It's that messy world of competing values that most of us have to live in.

I think it's a bit of reach to state that, just because unscrupulous politicians or news propagandists could concoct, in retrospect, a scenario that would attribute the success of some terrorist act to the lack of torture, such a narrative of events could never be correct regardless of the specific facts involved or the motives of those conducting the analysis.

The events of Sept. 11 are ultimately not a good example of the limited scenario we're arguing about because, for a number of different reasons, including bureaucratic lethargy and executive indifference, the Moussaoui lead was never exploited as it should have been. There's at least a chance that Sept. 11 could have been prevented through ordinary interrogation and research had all the cylinders of the intelligence apparatus been firing together.

But suppose we did have Atta in the basement at 4:00 a.m. that day, and suppose all available intelligence - wiretapping, examination of his associates' travel patterns, etc. - revealed that something big was likely to happen on that day. Suppose he is completely uncooperative or is failing a polygraph. Suppose agents' attention had been appropriately focused on this group for a while now and they already possessed a great deal of knowledge about the group's motives and goals. What they lack is specific knowledge about what will happen today.

If the answer to "What do we do?" is always, "We don't torture," okay. That's the end of the debate. But my point is that there will always be Monday morning quarterbacking of the worst sort in situations like this, much of it pernicious and dissembling and desperate to rid itself of any responsibility. For example, it would be equally disingenuous to expect perfection from normal intelligence or interrogative activities. In a situation where the staff involved was moving with all deliberate speed and operating within normal guidelines, there's obviously a strong chance that any planned terrorist act could be executed before all the dots were connected.

Low probability does not equal no probability. If we're going to take a morally pure approach ("We never torture"), and I'm not necessarily arguing that we shouldn't, then I maintain we have to be prepared to cleave to that high standard of morality even in the face of a scenario that might make it very uncomfortable.

That was an excellent response to a sincere question on my part. I agree with all the conditions you enumerated, and acknowledge that it would be an exceedingly rare circumstance when those variables were accounted for.

What your analysis implies, however, is that everyone who claims to be utterly against torture must continue to hold that position even on the occasion when it has a likelihood of being effective, and even when a hundred or a thousand or a hundred thousand people die as a result of an official's refusal to use any means necessary. That's logically consistent and morally admirable, but I frankly don't have enough faith in human nature, and definitely not in U.S. political culture (Exhibit A - our government's reaction to 9/11), to believe that people will ever be so sober in the face of great tragedy.

Speaking hypothetically, if it were ever determined that a significant terrorist attack might have been prevented by torture of a key individual under controlled conditions, I would fear for the life of the person in charge of the proceedings who refused to do so. The individual would likely be ruined professionally if not financially, and would have to defend himself/herself against every kind of scurrilous charge, even from the government whose laws he/she was obeying. That's why I wonder if some rarefied exceptions couldn't be made in the law, though I of course recognize all the dangers that entails.