There’s a running conversation over at the Corner about the parallels between opposition to torture and pacifism, which is really just a thinly-veiled version of one of those tendentious hypotheticals about the nuclear bomb in Manhattan whose location (you know with apodictic certainty) can only be uncovered through the judicious application of thumbscrews. As such, the analogy suffers from the same two defects: One empirical, one conceptual.
The empirical problem is this popular but stunningly stupid version of cost-benefit analysis that fails to seriously consider costs. Which, I guess, makes it “benefit analysis.” In other words, if it seems as though torture ever yields important and actionable intelligence more quickly than alternative methods, we’re supposed to take it for granted that this completes the necessary utilitarian analysis. And this is just absurd. How does torture affect the willingness of enemy combatants to surrender? How much does it complicate our relations with allies? How many people does it help to radicalize against the United States? How many non-radicals does it leave sufficiently disgusted that they’re less motivated to assist the U.S. in fighting radicalism in their communities? You’ll notice that torture-fans never really attempt to deal remotely seriously with any of these questions; they just babble inanities about how Fanatics Will Hate Us No Matter What. Which, of course, some will—but that’s hardly to the point, is it?
The conceptual mistake is to suppose that we’re faced with a binary choice between a pure consequentialism that just mechanically adds up all the yums and ouches or a kind of absolutist deontology that hews to a principled rule, and damn the consequences. The point of invoking pacifism is to imply that if you want to consider any non-consequentialist moral properties of certain kinds of acts, you’re compelled by relentless logic to the most extreme possible position. The thing is, pretty much nobody really thinks this way. Most people—the vast majority—will say it’s immoral to secretly chop up a healthy vagrant for organs to save five other people. We’re not just interchangeable tokens in some great social calculus, but individuals with individual rights that must be respected—rights that trump maximization of social welfare. Except that if suddenly we’re sure we could save a thousand or ten thousand or ten million people by killing one innocent, most of us will at some point say, reluctantly, that it ought to be done after all.
This is supposed to be theoretically unsatisfying because it seems inconsistent—if the weight of the numbers can outweigh those supposed rights after all, what’s so special about ten thousand or whatever other number? Why not five? But the demand for consistency here is misplaced. Our seemingly incoherent intuition here is grounded in the sense, which I think is correct, that consequentialist and deontic approaches both capture some sort of important moral truth, but that the two frameworks are (or have thus far seemed) incommensurable or incompatible in some deep way. The best we can do is to say that sometimes one will seem more appropriate, sometimes another. But it’s pure question-begging to insist that the internal standards of one framework or the other be used to determine where that transition point between approaches lies. That’s just tantamount to saying that the standard-setting framework is really paramount after all. I prefer the position that’s less theoretically tidy if it has the advantage of not being obviously wrong: At some point the consequences are so extreme that you ditch rights talk and just count noses—but it’s not a point we often encounter, even in counterterror policy.