scorecardresearch Skip to main content
Alan M. Dershowitz | Part 5

A choice of evils

Should democracies use torture to protect against terrorism?

edel rodriguez for the boston globe

Preventing terrorism often requires a choice of evils: to target terrorists and their leaders, knowing civilians will be killed; to collect massive amounts of private data in an effort to track terrorists; to detain potential terrorists without trial. But perhaps no choice of evils is more controversial than the use of torture to secure real-time intelligence against imminent acts of terrorism.

Is there ever a justification for the use of torture?

What if a captured terrorist bragged that he knew the location of a ticking nuclear weapon in a large city and refused to disclose it?

What if a child were kidnapped, as occurred in Germany several years ago, and the kidnapper was apprehended and refused to disclose where he had supposedly buried the child with only a few hours of air?

These sorts of cases — some real, others hypothetical — have been at the center of the debate that began in this country following the terrorist attacks on Sept. 11, 2001, over whether it is ever justified to use torture in an effort to prevent significant harms, including mass casualty terrorist attacks.


Several basic positions have emerged. The first is that regardless of the feared harm, torture should never be used. In support of this absolutist view it is sometimes argued that torture never works, because a tortured person will tell the torturer anything to stop the pain. This is a factual assertion that is difficult to prove. Although torture has produced many false confessions, it has also played a role in uncovering some self-proving truthful statements, such as the locations of bombs, plans, or people. The reality is that torture may sometimes work, but that no statement made under torture should ever be believed unless it is self-proving or corroborated by hard evidence.

A related argument acknowledges that torture may sometimes produce self-proving, valuable, real-time information, but that there are better methods for securing such information than by torture. This may well be true, but it doesn’t provide a complete rebuttal to the counter-argument that in a given case, torture may be the only or best tactic for saving lives.

If there are, in fact, such cases, then we have to face up to the conflict between the moral and the factual: Should torture, as a moral matter, be permitted in a case where, as a factual matter, it may save lives?


Proponents of the use of torture in such cases argue that saving lives has a higher value than avoiding infliction of pain, especially where the lives are innocent and where the pain would be inflicted on guilty terrorists. Jeremy Bentham constructed a trenchant “law school hypothetical case” in support of his view that based on a cost/benefit analysis, torture should sometimes be permitted. Bentham imagined a gang of torturers who, if they remained at liberty, would torture 100 innocent victims. He then asked whether it would be moral to torture one guilty member of that gang “to make known the place” where the other torturers could be found and apprehended, and thus save 100 innocent victims from torture. His answer was yes, based on the greatest good for the greatest number.

President Bill Clinton gave a similar answer: “Every one of us can imagine the following scenario. We get lucky and we get the number three guy in Al Qaeda, and we knew there’s a big bomb going off in America in three days and this guy knows where it is. Don’t we have the right and the responsibility to beat it out of him?”

He then pulled back when his wife disagreed with him. Obama agreed with Hillary Clinton that there should be an absolute prohibition against torture, even if there is a a ticking bomb — a situation we have, fortunately, not faced.


My own view is the same as Obama’s and Hillary Clinton’s. As a moral matter, I am categorically opposed to torture, without exception. But as a factual matter, I think every president would at least consider the option of torture if confronted with an actual ticking bomb. President Clinton said he would consider torture and President George W. Bush actually authorized the use of waterboarding, even in non-ticking bomb situations. Bush denied that waterboarding and other forms of extreme interrogation measures he approved were torture, but they would seem to fit any reasonable definition of that term.

If I’m right, and if every president would, in fact, consider opting for the torture of one terrorist rather than permitting thousands of innocent Americans to be blown up, then the following question must be asked: Would it be better or worse for a law to be passed requiring the president to secure a warrant before (or, in a real emergency, during or right after) he could employ this drastic measure? Such a law would implicitly legitimate torture in extreme situations, and that’s a bad thing, but it would also create visibility and accountability, which is a good thing.

Once again, we are faced with a terrible choice of evils.

My own view — which is controversial among liberal and conservatives alike — is that, on balance, visibility and accountability are essential to democracy, even if it means lending some legitimacy to an immoral and despicable tactic such as torture. I wish no one would ever torture, but I’m sure some will if the ticking bomb situation were ever to arise. That’s why I favor torture warrants.


Such is the nature and complexity of principled decision-making when confronting the evils of terrorism within the rule of law. There are no perfect answers, but some are worse than others. In such situations, democratic accountability suggests that we should generally opt for the least-worst approach that is most compatible with both the rule of law and the realities of terrorism.

Alan M. Dershowitz is an emeritus professor of law at Harvard University. His newest e-book, “Terror Tunnels: The Case For Israel’s Just War Against Hamas,’’ has just been published.