In a recent book and in a number of interviews, essays, and op-eds, Marc Thiessen has revived discussion of the tactics used by the Bush administration to obtain actionable intelligence to prevent terrorist attacks. The Christmas day arrest of Farouk Abdulmutallab confirms that this is still a live issue: ought Mr. Abdulmuttalab to have been waterboarded immediately in order to determine whether any other attacks were in the offing?
Thiessen’s defense of “enhanced interrogation tactics,” including waterboarding, has been especially striking for his reliance on the conceptual apparatus of the Catholic moral tradition. That tradition has held that the state has a duty to protect its citizens against aggression, and that in prosecuting that duty the state may use force, even lethal force.
Thiessen has further identified the principle of double effect as essential to this justification: the harm that is done to an assailant is licit, by this principle, only if it lies outside the intention of the agent who causes the harm. According to Thiessen, the intent of the interrogator “is not to cause harm to the detainee; rather it is to render the aggressor unable to cause harm to society. The act of coercive interrogation can have a double effect (to protect society and to cause harm to the terrorist) but one is intended, the other is not.”
One might ask how the subject of a coercive interrogation is to be understood as a “threat.” In most examples of defense against an aggressor where double effect is implicated, the aggressor physically threatens, and force is used to repel that threat, even if it is foreseen that harm, and even lethal harm, will be done as a consequence of the repelling—if, for example, it is done with a gun, or a bomb. Thiessen’s claim is interesting in this regard: the suspect is a threat in virtue of his knowledge of an impending attack, and protecting against this threat requires inducing the suspect to give up the information that he is concealing.
In this essay, I do not address the question of Catholic teaching on torture or “torture lite,” as enhanced interrogation is sometimes called. I think that Thiessen is correct in thinking that the teaching of the Catechism is not obviously definitive in ruling out torture for interrogatory reasons, since the Catechism does not specifically mention interrogatory torture by name. However, the Catechism’s repudiation of a number of other kinds of torture as “contrary to respect for the person and for human dignity” is quite strong, and its mention of torture “to extract confessions” could perhaps be read broadly to include interrogatory torture. It does not seem unlikely that stronger statements regarding this form of torture will be forthcoming from the magisterium.
I do not question Thiessen’s motives, honesty, or good will. He is correct that defense of society is a foundational obligation of public servants, and he himself performs an important public service by encouraging serious reflection on what is morally permissible and practically necessary in pursuit of this end. Additionally, as I mention below, the infliction of some pain should probably not be considered torture, and is not intrinsically impermissible.
Nevertheless, there are difficulties in relying of the principle of double effect in this case. And it is those difficulties, when considered in light of the limits of the permissible infliction of pain, which cast doubt on the legitimacy of waterboarding a suspect for the sake of obtaining information he might possess.
The essential difficulty in Thiessen’s argument is his application of double-effect reasoning. Double effect has its primary application in the domain of what one is doing, where what one does, including the consequences of what one does, will encompass both good and evil effects. Thus, I consider the possibility of taking medicine to cure a disease, and realize that taking that medicine will not only cure me, but also cause me to lose my vision. I may be said both to have cured my disease and to have brought on blindness, but, insofar as it was only the cure that I intended, and blindness was only foreseen, but not chosen, then bringing on blindness was not part of my act in the fullest sense—the sense that encompasses all and only what I intended.
This analysis has application where an attacker is physically prevented from fulfilling what he has set out to do: I stop the attacker by hitting, or shooting, or stabbing him. Stopping, or repelling, is what I intend, and the harm that I thereby cause is, when I act rightly, outside my intention. But in the case of interrogation, what is envisaged—the end pursued by the interrogator—is, in fact, something that is to be done by the suspect. This is a difficulty, because what I do ends, under most circumstances, with the choice of another agent to do something. I can, of course, establish conditions under which others will be motivated to do something, but its being done is still their doing, not mine.
Now this raises a problem for the double effect defense of harsh interrogation for the following reason. If there are two effects, as on Thiessen’s analysis, and the good one—the suspect revealing critical information and thus ceasing to be a threat—is not strictly speaking a part or consequence of my action, but is rather the action of another, then the double-effect analysis seems stalled: what I am doing, it seems, is causing the harmful effect, with a view to establishing conditions that will induce the suspect to bring about the beneficial effects.
If this is so, then the question must return to whether some particular tactic—waterboarding, for example, or mutilation, or sensory deprivation, to give three very different examples—is in fact a form of harm that is intended by the interrogator. For the double effect defense identifies the intention to harm as one that is impermissible. The agent who physically repels an attacker need not, to reiterate, intend any harm at all: his intention is only to use force to stop the aggression. Is there, then, something that the interrogator intends that is other than harm, which has the consequence of creating conditions that make it more likely than not that the suspect will confess?
Since pain is not itself a form of harm, there might well be. On this view, pain, because it is not a harm but rather is a sign of proper physiological functioning of an organism, is not something that it is intrinsically impermissible to intend. While this view seems implausible to some, it is not without merit: A dentist may try to create pain in your mouth in order to diagnose the location of a cavity. And, if spanking children is permissible it must be because the infliction of pain just as such is not morally impermissible.
So the infliction of some pain, ranging from various forms of minor discomfort to something presumably quite unpleasant might be permissible as a way of inducing another to render himself no longer a threat. But here is the location of what I will call the Torturer’s Dilemma.
The morally upright interrogational torturer wishes to create conditions in which the suspect will talk. One way that this could be done would be to create conditions in which it is impossible for the suspect not to talk: to create conditions of such agony that the suspect is no longer in control of his actions and literally has no choice but to speak. The intention is to push the agent past his breaking point, to break him down so that he is less of an agent and more of a tool of the interrogator.
I take it that this intention will be rejected by anyone who accepts the basic framework of double effect for the analysis of interrogation: the intention to break someone down is an intention to harm them, and that is morally impermissible. We should note, in passing, that if this is how interrogational torture is defined—as the attempt to break down a suspect so as to force him to talk—then there is an absolute moral norm against torture.
I have argued previously on Public Discourse that it is very hard to see repeated application of the techniques of enhanced interrogation as anything other than an attempt to break down the suspect, and nowhere is this more true than in the case of waterboarding. To waterboard someone repeatedly, multiple times a day for many days, seems like an attempt to grievously damage another human being.
Thus one side of the Torturer’s Dilemma: the infliction of pain to this point is a form of harm (and there are various other ways of inflicting harm as well that are more direct than the infliction of unbearable pain—mutilations, for example).
But, on the other side, the interrogational torturer, by my account, can inflict some pain for the sake of inducing a suspect to talk. And perhaps, given this, someone like Abdulmutallab, unlike Khalid Sheik Mohammed, ought to have been subjected to somewhat more inducement to provide information. He ought, perhaps, to have been made somewhat less, rather than more comfortable. Yet the limits of such an approach seem clear: a level of pain that falls short of harming a suspect by beginning to disintegrate his psycho-somatic functioning is hardly likely to induce a serious terrorist to answer questions voluntarily. When compared against other forms of interrogation, such as those that attempt to gain the suspect’s trust and confidence, a form of interrogation using only this degree of pain—a degree that would not begin to break the suspect down—might in fact have little to recommend itself.
I suspect that a single application of waterboarding does not constitute torture, and can be done without an intention to harm. Those who have assisted journalists, such as Christopher Hitchens, or military personnel, in being waterboarded, certainly did not intend them harm. But is there any reason to think that such one-off applications of this technique would provide serious intelligence? And how could one be sure, save by further applications, to the point at which one was convinced that the suspect was not holding out?
In any event, the upshot of my discussion is this: if, as the double effect defense presupposes, waterboarding or some other interrogation technique is done in a way that is expected to cause harm to the suspect, then that harm is most likely intended as a means by the interrogator and double effect will not justify it. And if such techniques are performed with the intention to cause pain, but not either direct physical harm, or psychological disintegration, then they are likely to be ineffective. Either way, it is, in my view, a good thing that United States’ policy has moved (as it did in the second Bush term) beyond the grim, if understandable, policies of the first few years after 9/11.