On Torture

“Death: Suppose a terrorist has hidden an atomic bomb on Manhattan Island which will detonate at noon on July 4 unless… (here follow the usual demands for money and release of his friends from jail.) Suppose, further, that he is caught at 10 a.m. of the fateful day, but—preferring death to failure— won’t disclose where the bomb is. What do we do? If we follow due process— wait for his lawyer, arraign him— millions of people will die. If the only way to save those lives is to subject the terrorist to the most excruciating possible pain, what grounds can there be for not doing so? I suggest there are none. In any case, I ask you to face the question with an open mind.” -Michael Levin

I recently read this argument in a Norton collection of readings for logical analysis, and I found it to be brutally repugnant. Obviously, Levin’s thought experiment is less of an exercise in thinking than it is a crafted argument in his favor. How is it that the terrorist in one breath is “preferring death to failure” and then in the next can be physically convinced into failure? Clearly, there are several assumptions here:

1. Torturing will be successful in breaking the will of the terrorist.
2. Based on such a success, future success is likely.
3. The terrorist you have has the information needed to find the bomb.

Each of these assumptions triggers new sets of entirely different arguments that require evidence and further deduction. Beyond this thought experiment he clarifies that torture is “justified only to save lives, it is justifiably administered only to those known to hold innocent lives in their hands.” He further remarks that since terrorists often proclaim themselves as perpetrators or televise their actions, questions that arise from people concerned about identification errors are disingenuous. But let us imagine the first hypothetical again with added strength. Let’s imagine that the terrorist is 100% confirmed as the initial bomb placer, and let’s further assume that the torture methods were successful in getting the terrorist to confess the true original location of the bomb. Even this strengthened version of the hypothetical doesn’t account for the possible externalities that could have altered the position of the bomb. For example, what would have stopped other terrorists, upon receiving the knowledge that their comrade was captured, from moving the bomb and covering their tracks? Up until that point, all of the known data was true, and the terrorist was known to be the one with innocent lives in their hands. For the sake of time and clarity, the point here is that a confession, even from the one who is known to be directly responsible, does NOT necessarily provide the required information to avert a crisis event. But if Micheal Levin’s point is that we should torture someone if:

1. We know they are 100% guilty.
2. We know, once heard, that the information they provided is 100% accurate.
3. The information they could provide is still materially beneficial to stop the crisis.
4. There is literally nothing else to do except torture the perpetrator. (Not even the most obvious courses of action like looking for the bomb supposedly hidden are available)

then this argument proceeds much like a salesman who has the perfect mattress for someone not interested in buying a mattress. Ultimately, I view torture as a course of action with gambling odds so bad that to engage with it would be to knowingly sink the moral bankroll of the nation and the prospects of the investigation. I argue that even when there is pressure from the context, we should leverage our cutting-edge technology, investigative skills, and law enforcement bureaus to produce results.