Five Errors in the Reasoning of Alan Dershowitz
by
Elaine Scarry*

In Torture: A Collection

Editor Sanford Levinson

Oxford University Press 2004
pp.281-290

At the center of Alan Dershowitz's recent account of torture is the argument that a hypothetical case can be imagined in which saving a city from a nuclear, chemical, or biological bomb might depend on torturing the terrorist who placed it there or knew where it was hidden.  His chapter "Should the Ticking Bomb Terrorist Be Tortured?" is the centerpiece of his book Why Terrorism Works, and his essay herein is again structured around the dramatic instance of the ticking bomb, which occurs at the beginning, middle, and end of his argument.  He believes that in such a situation it would be permissible to torture if one first obtained a judicial or executive warrant; the prohibition against torture, dissolved by means of the warrant, would continue in place for any act of torture that had not been warranted.

The first error in Alan Dershowitz's argument is that he wrongly addresses us as a population whose members are morally impaired. (It is of course the case that if we disagree with him we are perceived to be deficient:

p.281

the charge of "hypocrisy" to any who believe the prohibition against torture should remain firmly in place recurs at intervals throughout his two essays; but the problem I wish to identify is more grave and applies to us whether we agree or disagree with him; and I believe it distorts his reasoning about the key questions.)  Let us see precisely how this is so.

Introducing an "imaginable" occasion for torture that has no correspondence with the thousands of cases that actually occur has the effect of seeming to change torture to a sanctionable act.  As Henry Shu urges in his essay herein, the unwavering prohibition against torture must be kept in place; and should the unlikely "imaginable" instance actually ever occur, the torturer would have to rely on convincing a jury of peers that the context for the act was exceptional.

But exposing the defect of the ticking bomb argument requires that we go further.  Anyone, we are told, who had the choice between on the one hand torturing and saving-the-city and on the other hand not torturing and not saving-the-city would be likely to choose the first.  That may be.  But so, too, anyone confronted with the choice between on the one hand saving-the-city and being herself imprisoned, or on the other hand not saving-the-city and not being imprisoned, would almost certainly also choose the first.  That is, torturing should be perceived with the same acute aversion with which one's own legal culpability and one's own death are perceived; and while it is possible that a jury would exonerate someone in this situation, it does not follow that any such guarantee should be provided before the fact.  Nor should someone enter into the act expecting exoneration after the fact.  That one might have to do something someday that is wrong does not mean the act has ceased to be "wrong" and "punishable."  It is unlikely that any savior of the city would actually be inhibited by the lack of preexisting moral and legal assurances of immunity.

It is a peculiar characteristic of such hypothetical arguments on behalf of torture that the arguer can always "imagine" someone large-spirited enough to overcome (on behalf of a city's population) his aversion to torture, but not so large-spirited that he or she can also accept his or her own legal culpability and punishment.

The first major error in Alan Dershowitz's argument, then, is that he severely midjudges the compatriots to whom, and about whom, he is speaking. He rules out, at the outset, the possibility that if one of us had the chance to save the Earth from the scourge of a nuclear weapon, the person

p. 282

would forfeit his or her liberty or even life to carry out that act. The entire argument is premised on the idea that the population lacks the simple attribute of courage. The act of inflicting torture requires no courage (the aversiveness is wholly borne by someone else), whereas the forfeit of one's future liberty requires that some portion of the severe adversity be endured by the actor himself.

Dershowitz tells us--both in his book and in the present essay--that he repeatedly asks his students and lecture audiences to raise their hands if they believe someone who could stop a nuclear bomb by torturing would do so; he reports that invariably many hands rise in the air. What if he followed that simple experiment with another: "Raise your hand if you believe that someone who saw she had it within her power to save hundreds of thousands of lives would forfeit her own liberty or give her own life?"  Will not as many hands go up when this question is posed as when the imaginary opportunity to torture is posed?  In fact, many more people (such as soldiers) have shown themselves willing to give their lives to save other human beings than have ever shown themselves willing to perform an act of torture, so it is unclear why any legal impunity needs to offered to cover the unlikely ticking bomb situation.

The way a person's legal culpability for torture enables him to test the situation in front of him can be seen by noticing not just the final action that is taken but the stations along the way.  If I believe I am in the presence of someone who knows where an armed nuclear bomb is ticking away, I must ask myself how certain I am that this person actually knows that information.  Now the testing ratio comes into play.  If I say "I am confident enough that he holds this knowledge that I am willing to torture him;' then I ought also to be able to say "I am confident enough that he holds this knowledge that I am willing to forfeit my liberty and possibly my life in order to procure that knowledge."  If I instead find myself saying "Come to think of it, I'm not quite sure enough that I can give up my liberty to it;" it is the signal to revise my assessment of confidence that torturing him would produce any knowledge worth having.  Performing this test is more accurate (and certainly more rapid) than finding a judge who can issue a warrant, unless we design the warrant situation as one in which any judge who generates the warrant also agrees to go to jail.

In addition to the two defects of the argument on behalf of issuing warrants to permit torture in the ticking bomb situation--first, that it assumes

p. 284

a cowardly population incapable of acting without prior guarantees of immunity, and second, that it eliminates the procedure for testing one's level of confidence--there is a third problem. The ticking bomb scenario is often described as highly improbable.  What makes it improbable is not the existence of a ticking bomb (it is entirely possible that a terrorist or a deranged state leader will one day try to use a nuclear bomb, or a chemical or biological weapon capable of killing hundreds of thousands).  What instead makes the ticking bomb scenario improbable is the notion that in a world where knowledge is ordinarily so imperfect, we are suddenly granted the omniscience to know that the person in front of us holds this crucial information about the bomb's whereabouts.  (Why not just grant us the omniscience to know where the bomb is?)

In the two and a half years since September 11, 2001, five thousand foreign nationals suspected of being terrorists have been detained without access to counsel, only three of whom have ever eventually been charged with terrorism--related acts; two of those three have been acquitted.¹  When we imagine the ticking bomb situation, does our imaginary omniscience enable us to get the information by torturing one person?  Or will the numbers more closely resemble the situation of the detainees: we will be certain, and incorrect, 4,999 times that we stand in the presence of someone with the crucial data, and only get it right with the five thousandth prisoner?  Will the ticking bomb still be ticking?

Almost all aspects of our post--September 11 world bring us face to face with our lack of omniscience. We have failed, in two and a half years, to find the anthrax murderer, despite the fact that the precisely identified strain of anthrax limits the pool of eligible candidates to a tiny handful of people; this is not like finding a needle in a haystack, it's like finding a needle in a bright red pincushion that contains one needle and nineteen straight pins.  We have gone to war against a country that was "known" to have weapons of mass destruction, only to find it had none.  We knew our troops would be welcomed as liberators, at least by the Shi'ites, who are now killing our soldiers as the Sunnis hang our civilians from bridges.  And yet, despite our overwhelming miscalculation, mismeasurement, and inability to solve mysteries both before and after they happen--or to put it in the fairest light, despite the excruciating difficulty of ever being right--we are asked to entertain the possibility of lifting the unconditional prohibition against torture, and to do so by imagining that one of us will recognize the ticking bomb accomplice the moment we see him.

p. 285

Oddly, and conversely, torturing a person in order to get information about a ticking bomb is sometimes introduced in situations where the information is already available through means that require inflicting no cruelty.  Although Alan Dershowitz usually formulates the ticking bomb license-to­torture as a case in which hundreds of thousands of people stand to die from a nuclear, chemical, or biological weapon, he at one point asks us to imagine that we are back on September 11; and that by capturing and torturing the hijacker on one of the planes we can learn the target of another plane and enable the people in that targeted building to evacuate.  But a great deal of information about the plane that eventually hit the Pentagon was known from FAA air controllers, radar images, and a passenger cell phone call placed directly to the Justice Department.  For fifty-five minutes before the plane hit the Pentagon, it was clear that American Airlines Flight 77 was highly likely to be one of the hijacked planes, since it was off course, was not answering air controllers, and had turned off its secondary radar; for twenty minutes before it hit the Pentagon, it was certain (because passenger Barbara Olson phoned her husband, Theodore Olson, in the Justice Department) that the plane was under the control of hijackers; for twelve minutes before it hit the Pentagon, an air controller saw it on primary radar headed for Washington; for nine minutes before it hit the Pentagon, a C-130 watched it flying fast and low.²  We need to wonder why we were not able to use this available information to get people out of the Pentagon and other Washington buildings, rather than supposing that it would have been helpful to torture someone to learn its target.

To summarize, then, the ticking bomb scenario, with warrant as a license to torture, presents us with three major problems: (1) it assumes a population that is (against robust evidence) cowardly and self-regardingable and willing to torture but unable and unwilling to themselves suffer harm; (2) it assumes a population that is (against robust evidence) omniscient; and (3) by providing legal immunity, it eliminates the felt-aversiveness to cruelty that acts as a way to test one's level of conviction that thousands of lives are at risk and that one is uniquely positioned to act as their savior.

Alan Dershowitz has asked us to put aside our commitment to an unwavering prohibition on torture, to enter into an open debate with him, and to step into that debate by passing through the threshold of the ticking bomb case, a case whose framing assumptions are erroneous. But it may be useful to proceed forward and examine some of his other arguments, for it is apparent that here, too, he presents us with errors both in the substance

p. 286

and the style of the reasoning; and perhaps if the errors are articulated, he will be persuaded to return to his own original position, which (he several times tells us) was until recently a blanket condemnation of torture.

The proposal we are asked to contemplate is one in which a judge or an executive branch officer will issue a warrant licensing the holder of the warrant to torture, and the claim is that the existence of the warrants will, by introducing judicial scrutiny and by providing a documentary record, reduce the number of incidents of torture that take place and increase the accountability of those carrying out such acts.

This is a puzzling claim.  We are asked to assume that a judge or executive branch officer, acting under the pressure of a ticking bomb, will be able to discriminate between acceptable and unacceptable cases.  Are acceptable cases those that involve weapons of mass destruction (and therefore tens of thousands of deaths) and unacceptable cases those that involve smaller num­bers of injuries?  Or is some factor other than number of persons the key, as at many moments appears to be the case in Dershowitz's own examples?  Dershowitz might fairly complain that we only lack an answer to that question because so far the debate has not really gotten underway, and therefore the practical details of the arrangement have not been worked out.

But there does already exist a solid basis for our skepticism that a warrants system will produce coherent discrimination.  The court set up to issue warrants under the Foreign Intelligence Surveillance Act (FISA) has declined only one requested warrant in twenty-five years: the estimated number of warrant requests is twenty-five thousand.³  If a torture warrant court were based on this modeI,⁴ the incidence of torture would not be likely to decline, nor would the level of accountability increase.

But let us assume that the torture warrant court would, unlike the FISA court, operate with a high level of resistance and would grant only a small number of warrant requests (those providing strong evidence that extreme injury is about to take place very soon, and also providing strong evidence that one specified prisoner holds the key that will enable us to prevent the injury).  In other words, let us assume that a coherent principle of discrimination is at work, and now let us see if we can decipher how this will decrease the incidence of torture and increase the level of accountability.  Nothing about the results appears to let us reach this conclusion.

Under this new system, the prohibition against torture will dissolve in those cases where the torturer obtains a warrant but will continue in place for any act that has not first been warranted.  Of these two groups of per-

p.287

sons, the permitted and the prohibited, is it the first group, the second group, or the two groups together that enable us to achieve a level of accountability that surpasses the level available under our longstanding blanket prohibition of torture?  Clearly, it cannot be the second group. All acts of torture that are carried out without a warrant (either because the torturers refuse to consult the court, or because they surreptitiously carry out their acts of torture even after their application for warrants have been turned down) will be undocumented--or, more accurately, they will have only the level of documentation that we have today under the blanket prohibition (a point that will be returned to).  Is it, instead, the first group that will be accountable?  We will, by virtue of the warrant, have a record of the person's actions, the reasons for those actions, the outcome (assuming the warrant holder does not stray from what he or she requested and what the judge granted) and can now request to see concrete evidence of the ticking bomb that was dismantled.  But since the torturer has, by means of the warrant, already been released from the usual constraints against torture, in what does his or her accountability consist?  Didn't the judicial review, by taking account of his or her proposed actions, release him or her from further accountability?  It may be that we can review his or her actions: would this mean we should understand the warrant as a temporary grant of permission that is, upon review, subject to retrospective revocation, at which point the torturer's exemption from punishment would dissolve?  Long experience with search warrants suggests the opposite: search warrants, far from facilitating review, historically have tended to close the door on review.  In The Bill of Rights: Creation and Reconstruction, Akhil Amar describes the way the search warrant has often acted as a shield against the charge of trespass.⁵

The torture warrant system, then, appears to leave us with an unknowable number of illegal instances that cannot be reviewed (that is, cannot be reviewed in any way that is not already available to us under our current blanket prohibition) and a knowable number of legal instances that because they have been warranted are unlikely to be reviewed--even in the way that is currently available to us under the blanket prohibition system.  Under the proposed system, our ability to review acts of torture has gotten no better, and actually appears to have gotten worse, than under our present blanket prohibition.

Alan Dershowitz, then, credits the warrant system with a power of documentation and accountability it does not appear to have.  Conversely, he

p. 287

undercredits the forms of documentation and accountability that already exist under the present across-the-boards prohibition.  He seems to believe that if someone wants the ban on torture to be absolute yet acknowledges that torture occurs, the person must be a hypocrite who pretends to denounce brutality while letting it take place "under the radar." The most baffling moment in his essay comes when he accuses William Schulz, the executive director of Amnesty International USA, of having an insufficient understanding of the purpose of accountability. Because William Schulz opposes warrants for torture, as well as warrants for brutality, testilying, and prisoner rape, Alan Dershowitz asks Schulz: "Do you prefer the current situation in which brutality, testilying and prisoner rape are rampant, but we close our eyes to those evils?"

Does Alan Dershowitz not know, or has he somehow forgotten, that Amnesty International's major work in the world is relentlessly to document instances of torture that have taken place, to make a public record, and through that record, to bring public pressure to bear on stopping the acts of torture even as they are taking place?  Torture is itself a ticking bomb (it inflicts grave and widespread physical injury); Amnesty works to stop it, not only before it goes off but often in the very midst of its explosion.  The suggestion that Amnesty, because it opposes warrants, prefers that "we close our eyes" is astonishing, given that no group has so steadily required us to keep our eyes on torture. This does not mean that the record is close to complete.  Amnesty International continually reminds us that its own records are incomplete, that it can document instances of torture only in countries where Amnesty members are permitted to speak with prisoners.  Many other research bodies--newspapers, human rights groups, congressional or UN investigative groups--also contribute to the widespread commitment, and ability, to document torture.⁶

We encounter, then--in addition to the three major problems earlier summarized in the ticking bomb frame--a fourth and fifth major problem in the proposal for warranting torture. The fourth is that Alan Dershowitz credits the warranting system with a power to provide documentation and accountability that it does not appear to have.  The fifth is that his proposal greatly undercredits the forms of documentation and accountability already available to us.  These allow us continually to strive for some measure of accountability, while keeping national and international prohibitions on torture fully in place.

p.289

Although I have focused here on the framing assumptions and substance of his argument, problems also occur in the form and style of that argument.  He often exposes the flaw of a particular idea (e.g., the necessity defense), form of sequencing (e.g. slippery slope), or phrase (e.g., "torture lite")--seeming in each case to repudiate it in unequivocal terms--only to bring that idea, form of sequencing, or phrase back into the service of his own proposal, usually without identifying it by its earlier name or label.  The phrase "nonlethal torture" though it literally designates a horrifying set of practices, is in the context of his essay used as though it meant "moderation," without announcing that cruel linguistic trick in the way "torture lite" openly does.  He critiques the "necessity" defense, accurately identifying it as "the most lawless of legal doctrines" and warning us that it is so elastic it can accommodate any person and any position; yet his warrant system gives center stage to the necessity defense, bestows on it a material form, and turns it into a formal procedure.  Through this method of repudiating, then using, phrases, forms of sequencing, and ideas, he protects his arguments by giving them deniability.  Were we to fault him for relying on the "necessity defense," he would look startled, indignant, and quote back to us the four pages in which he has discredited that defense.  Thus we arrive at the climactic moment in his essay where he quotes those who fault him for countenancing the legitimation of torture--quotes them with astonishment, as though he cannot comprehend where on earth such descriptions (straightforward summaries of his view) could possibly have come from.

No one should take Alan Dershowitz lightly (even when face to face with his light, bright spirit).  He means business.  He intends to open a debate.  He intends that debate, in turn, to reopen the law, to alter it, to replace the blanket prohibition on torture with partial legitimacy.  He assumes--rightly-­that he and those to whom he addresses himself have the power to change law and legal practice.  He dedicates his book on the ticking bomb to the "nearly ten thousand students" he has taught over thirty-eight years at Harvard Law School (many of whom now hold legal positions around the country and the world). "You are our future," he tells them. "Preserve it from our enemies."

Let us hope that his former students and all other readers will see the errors in his reasoning and conclude that the best way to preserve the future from "our enemies" is to reaffirm each day the blanket prohibition on tor-

p.290

ture, and to work with newspapers, human rights groups, and investigative bodies to document and hold those who torture accountable for their acts.

Notes

1. David Cole, lecture, Harvard Law School, September 24, 2003.

2. The inability to act on available information is described in more detail in Elaine Scarry, Who Defended the Country? ed. Joshua Cohen and Joel Rogers (Boston: Beacon Press, 2003).

3. David Cole, Enemy Aliens: Double Standards and Constitutional Freedoms in the War on Terrorism (New York: New Press, 2003), 68; Dan Eggen and Susan Schmidt, "Secret Court Rebuffs Ashcroft," Washington Post, August 23, 2002, 1.

4. Alan Dershowitz does sometime appear to be assuming the warrant court will be, like the FISA court, secret.  He at one point speaks of a "special cabinet committee or judicial panel authorized to approve special measures under extraordinary circumstances" (n. 3).

5. Akhil Reed Amar, The Bill of Rights (New Haven, Conn.: Yale University Press, 1998), 71-73. Amar writes: "In the end, the Fourth Amendment framers accepted some warrants as necessary but imposed strict limits on these dangerous devices.  Warrantless searches did not pose the same threat because those searches would be subject to full and open after-the-fact review in civil trespass cases featuring civil juries" (73).

6. For an example of the documentation of acts of torture (and those who have trained torturers) gathered by multiple research groups, see Timothy Kepner, "Torture 101: The Case against the United States for Atrocities Committed by School of the Americas Alumni," Dickinson Journal of International Law (2001). Kepner assesses the potential liability of the School of Americas in a domestic court, using the Alien Tort Act and the Torture Victim Protection Act of 1991; he takes into consideration complications that arise from the Federal Tort Claims Act, the Combatant Activity Exception, the Foreign Country Exception, and the Political Questions Doctrine.