On the appointed day the unarmed crowd of the Gothic youth was carefully collected in the square or forum; the streets and avenues were occupied by the Roman troops, and the roofs of the houses were covered with archers and slingers. At the same hour, in all the cities of the East, the signal was given of indiscriminate slaughter; and the provinces of Asia were delivered, by the cruel prudence of Julius, from a domestic enemy, who in a few months might have carried fire and sword from the Hellespont to the Euphrates. The urgent consideration of the public safety may undoubtedly authorize the violation of every positive law. How far that or any other consideration may operate to dissolve the natural obligations of humanity and justice is a doctrine of which I still desire to remain ignorant.

—Edward Gibbon, The Decline and Fall of the Roman Empire (1776), 2.36

I

What lesser evils may a society commit when it believes it faces the greater evil of its own destruction? This is one of the oldest questions in politics and one of the hardest to answer. The old Roman adage—the safety of the people is the first law—set few limits to the claims of security over liberty. In the name of the people's safety, the Roman republic was prepared to sacrifice all other laws. For what laws would survive if Rome itself perished? The suspension of civil liberties, the detention of aliens, the secret assassination of enemies: all this might be allowed, as a last resort, if the life of the state were in danger. But if law must sometimes compromise with necessity, must ethics surrender too? Is there no moral limit to what a republic can do when its existence is threatened? As Edward Gibbon retold the story of how the Romans slaughtered defenseless aliens in their eastern cities in 395 C.E. as a preemptive warning to the barbarians massing at the gates of their empire, he declined to consider whether actions that political necessity might require could still remain anathema to moral principle. But the question must not only be asked. It must be answered.

If the society attacked on September 11, 2001, had been a tyranny, these ancient questions might not be relevant. For a tyranny will allow itself anything. But the nation attacked on that bright morning was a liberal democracy, a constitutional order that sets limits to any government's use of force. Democratic constitutions do allow some suspension of rights in states of emergency. Thus rights are not always trumps. But neither is necessity. Even in times of real danger, political authorities have to prove the case that abridgments of rights are justified. Justifying them requires a government to submit them to the test of adversarial review by the legislature, the courts, and a free media. A government seeking to respond to an attack or an expected danger is required to present the case for extraordinary measures to a legislature, to argue for them with reasons that might convince a reasonable person, and to alter the measures in the face of criticism. Even after extraordinary measures receive legislative approval, they will still come under review by the courts.

The first challenge that a terrorist emergency poses to democracy is to this system of adversarial justification. The machinery of legislative deliberation and judicial review grinds slowly. Emergencies demand rapid action. Hence they require the exercise of prerogative. Presidents and prime ministers have to take action first and submit to questions later. But too much prerogative can be bad for democracy itself.

In emergencies, we have no alternative but to trust our leaders to act quickly, when our lives may be in danger, but it would be wrong to trust them to decide the larger question of how to balance liberty and security over the long term. For these larger questions, we ought to trust to democratic deliberation through our institutions. Adversarial justification is an institutional response, developed over centuries, to the inherent difficulty of making appropriate public judgments about just these types of conflicts of values.1 Citizens are bound to disagree about how far the government is entitled to go in any given emergency. Because we disagree deeply about these matters, democracy's institutions provide a resolution, through a system of checks and balances, to ensure that no government's answer has the power to lead us either straight to anarchy or to tyranny.

In a terrorist emergency, we disagree, first of all, about the facts: chiefly, what type and degree of risk the threat of terrorism actually presents. It would make life easy if these facts were clear, but they rarely are. Public safety requires extrapolations about future threats on the basis of disputable facts about present ones. Worse, the facts are never presented to the public simply as neutral propositions available for dispassionate review. They come to us packaged with evaluation. They are usually stretched to justify whatever case for action is being made. Those who want coercive measures construe the risk to be great; those who oppose them usually minimize the threat. The disagreements don't end there. Even when we agree about the facts, we may still disagree whether the risks justify abridgments of liberty.

What prevents such a system from falling prey to the tyranny of the majority is the system of checks and balances and, more broadly, the democratic process of adversarial justification itself. While injustice can always be justified if you have to justify it only to yourself, it is less easy when you have to justify it to other democratic institutions, like courts and legislatures or a free press. Thus presidents or prime ministers may not see anything wrong in a stringent measure, but if they know that this measure will have to get by the courts and the legislature, they may think twice.

Besides these constitutional checks and balances, there would also be the democratic check of competing social, religious, and political interests in the nation at large. One of the most lucid versions of this argument is to be found in Federalist No. 51, where in discussing the federal system's balance of federal and state power, the authors go on to say that while all authority in the United States will be derived from the power of the majority,

the society itself will be broken into so many parts, interests, and classes of citizens, that the rights of individuals, or of the minority, will be in little danger from interested combinations of the majority. In a free government, the security for civil rights must be the same as that for religious rights. It consists in the one case in the multiplicity of interests and sects; and this may be presumed to depend on the extent of country and the number comprehended under the same government.4

Against this pragmatic view there is a moral view of democracy which maintains that it is something more than majority rule disciplined by checks and balances. It is also an order of rights that puts limits to the power of the community over individuals. These limits are not there just for prudential reasons, to prevent governments from riding roughshod over individuals. The rights are also there to express the idea that individuals matter intrinsically. Democracies don't just serve majority interests, they accord individuals intrinsic respect. This respect is expressed in the form of rights that guarantee certain freedoms. Freedom matters, in turn, because it is a precondition for living in dignity. Dignity here means simply the right to shape your life as best you can, within the limits of the law, and to have a voice, however small, in the shaping of public affairs. Government for the people, in other words, is something more than government for the happiness and security of the greatest number. The essential constraint of democratic government is that it must serve majority interests without sacrificing the freedom and dignity of the individuals who comprise the political community to begin with and who on occasion may oppose how it is governed. Rights certainly owe their origin to the sovereignty of the people, but the people—and their representatives—must steer majority interests through the constraints of rights.

Aharon Barak, president of Israel's Supreme Court, describes these two conceptions of democracy as “formal” and “substantive.”5 Other scholars have contrasted a “pragmatic” reading of the U.S. Constitution with a “moral” reading.6 In normal times, these two meanings of democracy—one stressing popular sovereignty, the other stressing rights; one privileging collective interests, the other privileging individual dignity—are interdependent. You can't have a democracy without rights, and rights cannot be secure unless you have democracy. But in terrorist emergencies, their relation breaks apart. What makes security appear to trump liberty in terrorist emergencies is the idea—certainly true—that the liberty of the majority is utterly dependent upon their security. A people living in fear are not free. Hence the safety of the majority makes an imperative claim. On this view, rights are political conveniences a majority institutes for its defense and is therefore at liberty to abridge when necessity demands it. Those who defend a rights-based definition of democracy will then argue that rights lose all effect, not just for the individuals at risk, but for the majority as well if they are revocable in situations of necessity.

Civil libertarians think civil liberties define what a democracy is. But the recurrently weak and shallow public support for civil liberties positions suggests that many Americans disagree. They believe that the majority interest should trump the civil liberties of terrorist suspects.8 For these democrats, rights are prudential limits on government action, revocable in times of danger; for civil libertarians, they are foundational commitments to individual dignity that ought to limit government action in times of safety and danger alike. For one side, what matters fundamentally is that democracies prevail. For the other, what matters more is that democracies prevail without betraying what they stand for.

When citizens of a democracy insist that what matters most in a terrorist emergency is the safety of the majority, they are usually saying that rights are at best a side constraint, at worst a pesky impediment to robust and decisive action. Those who think this are also likely to believe that international agreements, like the Geneva Conventions or the Torture Convention, should not limit what the United States can do in a war on terror. Since the threat is primarily directed at the United States, it must respond according to its own system of law, not according to anyone else's standards. To take this position, however, is also to assume that the lives of your own citizens matter more than the lives of people in other countries. It is, as Ronald Dworkin has pointed out, to base policy on the premise that Americans come first.11 Those who disagree will usually be committed to the idea that a democracy's ethical commitments are universal and apply both to its own citizens and to its enemies.

These debates are also about whether some measures are just plain wrong. Consequentialists argue that measures which aim to save lives and preserve the security of the citizens cannot be wrong if they actually succeed in doing so. They are wrong only if they don't work—that is, if they produce a chain of further harms, like more terrorist attacks. Most civil libertarians take the view that some actions remain wrong even if they work. So torturing someone to divulge terrorist actions is wrong, no matter what useful information is extracted, and hence no democracy should ever have anything to do with torture. A third position lies between the two. It maintains that consequences can matter so much, for example, saving thousands of people from terrorist attack, that it might be worth subjecting an individual to relentless—though nonphysical—interrogation to elicit critical information. But this style of interrogation, which would push suspects to the limits of their psychological endurance, would remain a violation of their dignity. It would be a lesser evil than allowing thousands of people to die, but its necessity would not prevent it from remaining wrong.

This third position—which gives this book its title—maintains that necessity may require us to take actions in defense of democracy which will stray from democracy's own foundational commitments to dignity. While we cannot avoid this, the best way to minimize harms is to maintain a clear distinction in our minds between what necessity can justify and what the morality of dignity can justify, and never to allow the justifications of necessity—risk, threat, imminent danger—to dissolve the morally problematic character of necessary measures. Because the measures are morally problematic, they must be strictly targeted, applied to the smallest possible number of people, used as a last resort, and kept under the adversarial scrutiny of an open democratic system.

A lesser evil position holds that in a terrorist emergency, neither rights nor necessity should trump. A democracy is committed to both the security of the majority and the rights of the individual. Neither a morality of consequences nor a morality of dignity can be allowed exclusive domain in public policy decisions. If each of these ethical principles has legitimate claims, the resulting framework is going to be complex, to say the least. In it, there are no trump cards, no table-clearing justifications or claims. What works is not always right. What is right doesn't always work. Rights may have to bow to security in some instances, but there had better be good reasons, and there had better be clear limitations to rights abridgments; otherwise, rights will soon lose all their value. At the same time, a constitution is not a suicide pact: rights cannot so limit the exercise of authority as to make decisive action impossible. Finally, international standards matter. Nations are not moral islands: they should conform to international standards, both to comply with the treaties and conventions that nations have signed and to pay what Thomas Jefferson called “decent respect for the opinions of mankind.”

A lesser evil morality is designed for skeptics, for people who accept that leaders will have to take decisive action on the basis of less than accurate information; who think that some sacrifice of liberty in times of danger may be necessary; who want a policy that works but are not prepared to make what works the sole criterion for deciding what to do. Such an ethics is a balancing act: seeking to adjudicate among the claims of risk, dignity, and security in a way that actually addresses particular cases of threat. An ethics of balance cannot privilege rights above all, or dignity above all, or public safety above all. This is the move—privileging one to the exclusion of the other—that produces moral error. They are all important principles—all must be weighed in the balance equally—and nothing trumps.

This is an ethics of prudence rather than first principle, one that assesses what to do in an emergency with a conservative bias against infringements of established standards of due process, equal protection, and basic dignity. A conservative bias assumes that in terrorist emergencies, the first response is usually wrong. Tried and tested standards of due process should not be hastily discarded. These standards are more than procedures, anchored in legal tradition. They reflect important commitments to individual dignity. Protection of the law means, concretely, that no one should be held indefinitely, without charge, without access to counsel or judicial review. Moreover, persons can be detained only for what they have done, not for who they are or for what they think, profess, or believe. A key conservative principle would be that blanket detentions and broad roundups of suspects are always a mistake, because they violate the law's principle of the individuality of guilt. It is invariably wrong to arrest or detain on the principle of guilt by association, based on race, ethnicity, or religious affiliation. Any detention policy must be targeted to individuals against whom probable cause can eventually be demonstrated. By these standards, the United States failed the test in its detention of nearly five thousand aliens, mostly single males of Muslim or Arab origin, after September 11. None have been found to merit charging with terrorist offenses. In retrospect, the whole exercise seems to have been as unnecessary as it was unjust.12

While a conservative bias will enable us to see through most of the overhasty reactions to terrorist emergencies, it may not be adequate when we have to face terrorists who control weapons of mass destruction. If the threat is sufficiently great, preemptive detention of suspects, together with military or police action to disarm, disable, or neutralize the threat, may be necessary. It is unrealistic to think that commitments to dignity, coupled with a conservative bias against departing from tried legal standards, will be sufficient to cope with any eventuality in the future. In the wake of another mass casualty terrorist attack, on or above the scale of September 11, most bets—and gloves—would be off. Even extreme necessity, however, cannot override democratic processes and the obligation to balance strong measures with basic commitments to full public justification.

If a war on terror may require lesser evils, what will keep them from slowly becoming the greater evil? The only answer is democracy itself. Liberal democracy has endured because its institutions are designed for handling morally hazardous forms of coercive power. It puts the question of how far government should go to the cross fire of adversarial review. Adversarial review procedures do not just pit one branch of government against another. Within each branch, there are, or should be, checks and balances, fire walls that guarantee the independence of institutions that perform intra-agency review. The General Accounting Office, for example, keeps the spending propensities of other federal agencies of the U.S. government in check. One branch of the Justice Department has recently criticized another branch's handling of administrative detainees after September 11, and that branch has changed its practices.13

In this process of adversarial review the test of reason is not the test of perfection. Citizens usually accept the decisions that result, not because they are right but because they are reasonable, and because democratic review affords a genuinely adversarial and open contest of opinions. Of course, even the most open process may produce perverse results. Senator Joseph McCarthy harassed and defamed individuals suspected of Communist sympathies in the full glare of publicity, and with the support, for a time, of majority opinion. While open proceedings are fallible, they at least create the possibility for correcting error. If McCarthy persecuted innocent people in open proceedings, he was also brought down by open proceedings.14 Ultimately, if open proceedings fail to produce answers that command the assent of citizens, it is up to citizens themselves to force the institutions—through public criticism and the electoral process—to come up with better answers. What is striking about democracy is the role of distrust in keeping the system honest. The system of checks and balances and the division of powers assume the possibility of venality or incapacity in one institution or the other. The ultimate safety in a democracy is that decisions filtered down through this long process stand less of a chance of being wrong than ones decided, once and for all, at the top.

The war waged against terror since September 11 puts a strain on democracy itself, because it is mostly waged in secret, using means that are at the edge of both law and morality. Yet democracies have shown themselves capable of keeping the secret exercise of power under control. So long as “a decision for secrecy should not itself be secret,” secrecy can be controlled.15 Legislatures can take hearings on sensitive intelligence matters in camera; judges can demand that state prosecutors justify secret hearings or the withholding of information from the defense. The redlines should be clear: it is never justified to confine or deport an alien or citizen in secret proceedings. Openness in any process where human liberty is at stake is simply definitional of what a democracy is. The problem is not defining where the red-line lies, but enforcing it. A democracy in which most people don't vote, in which many judges accord undue deference to executive decisions, and in which government refuses open adversarial review of its measures is not likely to keep the right balance between security and liberty. A war on terror is not just a challenge to democracy; it is an interrogation of the vitality of its capacity for adversarial review.

II

Having laid out the rudiments of a lesser evil approach to a war on terror, I need to say something about the word evil. First of all, not all evil is done by evil people or with evil design. Some of the worst things done to human beings are done with the best will in the world. The evil that is characteristic of democracies usually results from the blindness of good intentions. The evil I have in mind is committed by the officials of liberal democratic states who know they are not supposed to do wrong and who serve institutions that are created to guard against it. They may do wrong, nonetheless, believing these actions are justified because they forestall still greater harms, or, given the scale of modern bureaucracy, because they hardly know the consequences of their actions at all.

But why should democracies have anything to do with evil? Why expose their servants to such moral hazard? Why not stay safely on the side of pure legality? The answer is that we are faced with evil people, and stopping them may require us to reply in kind. If so, how do we keep lesser evils from slipping into greater ones?

Let me admit that the very process of justifying an act as a lesser of two evils is an exercise in moral risk. We can legitimately do so only if we actually know what we are doing and do not try to pretend that the necessary character of an evil act excuses its morally dubious character. Thus killing an innocent person to save the lives of hundreds of others might be a lesser evil, but the act would still be wrong. The law might accept a plea in mitigation but it would not excuse the act's criminal character. The Israeli Supreme Court has ruled that an agent of the state may make a defense of necessity if accused of torturing someone: this excuse might mitigate the penalty for violating the law, but it would not excuse the torture itself, which remains criminal.16

I do not want to minimize the moral hazard of resort to evil means. Sometimes we can accurately predict this hazard but more often we cannot. Choosing a lesser evil to ward off a greater one may bring the greater one to pass nonetheless. When escalating a military conflict, for example, a commander may choose the lowest possible increase in the gradient of force out of a desire to minimize harms and to secure a military objective at the lowest possible cost. But this may only redouble an opponent's willingness to resist, with the unintended result that the conflict costs more lives, possibly to both sides, than a short, sharp escalation would have done.17 Bad consequences are not always predictable, and so in choosing the lesser evil course, we may have to take a shot in the dark, knowing, unfortunately, that good intentions cannot exempt us from blame when bad consequences result.

A further reason why any recourse to a lesser evil is bound to be morally hazardous is that human beings are so adept at inventing good intentions, coming up with plausible excuses for atrocious consequences. The ancient Greek playwright Euripides gives us a particularly unsparing account of this in Medea. It is a play about a woman who kills her two children, in order, so she claims, to spare them the horror of being killed by strangers:

Women, my task is fixed as quickly as I may

To kill my children, and start away from this land,

And not, by wasting time, to suffer my children

To be slain by another hand less kindly to them.

Force every way will have it they must die, and since

This must be so, then I, their mother, shall kill them.18

But since she also wants revenge on the father of the children, who has abandoned her, it is impossible to see Medea in an unqualified moral light. She may be sparing her children, but she may also be sacrificing them to her own fury. She even admits this:

I know indeed what evil I intend to do,

But stronger than all my afterthoughts is my fury,

Fury that brings upon mortals the greatest of evils.

Moreover, from the standpoint of her children, who is to say that being murdered at the hands of your own mother is a mercy compared to being murdered at the hands of strangers? Only if we can be certain that they would be tortured, abused, and then killed by strangers can we justify her act as a lesser evil. Because Euripides is a great playwright, he leaves us—the audience—uncertain about this issue, which is why two thousand years later, we still leave the theater wondering whether she is a self-justifying monster or a tragic angel of mercy.

As Euripides shows us, human beings can justify anything as a lesser evil if they have to justify it only to themselves. In Medea's case, the audience sees more clearly than Medea ever does, even if they cannot disentangle her motives. In this play, evil appears as the incapacity to take any distance—through reason—from the primal force of feeling, so that all strong emotion becomes automatically self-justifying.

Evil can also appear in rational form, in the careful and deliberate choice to do harm, motivated by a rational but mistaken calculation of anticipated good. Either way, in hysterical grief or in cold calculation, the course of the lesser evil can lead to tragedy or to crime. But as Euripides insists, these choices are unavoidable elements of human experience. A war on terror presents leaders with such choices: harming some to save others, deceiving some to entrap others, killing some to preserve the freedom of others. Democracy is designed to cope with tragic choice, and it does so by understanding that if anyone can justify anything, provided they only justify it to themselves, they are less likely to be able to carry it out if they are forced to do so in adversarial proceedings before their fellow citizens.

But this does not mean that democrats do not sometimes commit evil. No system of politics is capable of saving us from moral hazard. Indeed, as Machiavelli understood a long time ago, dubious decisions are not just accidental incidents in political life; they are intrinsic to political action. Machiavelli famously insisted that the moral qualities we admire in private life—probity, honesty, forbearance—can be liabilities in public life, and that to apply private scruples to the decisions that have to be taken when a republic's security is at stake might be to condemn the republic to disaster.19 Thanks to Machiavelli, we are familiar with the irony that a politician who, in private life, would condemn acts of killing, must not hesitate to order his armed forces to kill the republic's enemies en masse. The same political leader who would be ashamed to lie to his own family must not hesitate to dissemble in the legislature when the public safety requires that a secret mission to protect the republic be concealed from scrutiny.

If Machiavelli says that all political life necessarily entails lesser evils, he does not ask whether democracy imposes particular limits on the kinds of evils that democrats can contemplate. While international law sets standards for states regardless of whether they are free or tyrannical, well or ill governed, it seems evident that democratic states will hold themselves to higher standards of dignity and due process.

They do so because liberal states seek both to create a free space for democratic deliberation and to set strict limits to the coercive and compulsory powers of government.20 This is the double sense in which democracies stand against violence: positively, they seek to create free institutions where public policy is decided freely, rather than by fear and coercion; negatively, they seek to reduce, to a minimum, the coercion and violence necessary to the maintenance of order among free peoples.

This is not true of democracy's other twentieth-century competitors. In Hitler's Germany or Stalin's Russia, law, politics, and culture were all ordered so as to eliminate the very idea that government violence was problematic. Far from being evils, Hitler's and Stalin's acts of extermination were heralded as necessary to the creation of a utopia: a world of class unity and social justice, or, in the case of the Thousand Year Reich, a nation purified of the racial enemy.21 If such were the Utopias that class warfare and racial extermination were supposed to serve, violence in their service could hardly be a crime. Hence the very idea that violence can be a lesser evil has meaning only in societies very different from these.

Thanks to the rights they entrench, the due process rules they observe, the separation of powers they seek to enforce, and the requirement of democratic consent, liberal democracies are all guided by a constitutional commitment to minimize the use of dubious means—violence, force, coercion, and deception—in the government of citizens. It is because they do so in normal times that they feel constrained to do so in times of emergency. Otherwise, these societies will not be true to who they are. When citizens consent to be ruled, they do so on the condition that the abridgment of their freedom, necessary to maintain a free and secure public realm, must be kept to a minimum. This implies that in a liberal democracy even government based on consent remains coercive. The coercions in question range from collection of taxes and the imposition of fines to punishment for criminal or civil liability. Coercion may be necessary to maintain social order, but in a democratic theory of government it is an evil, and it must be kept to a strict minimum.22 Why else would a liberal society put such store in rights, if it did not seek to protect individuals from the abusive exercise of coercive power?

This account of a liberal democracy may sound strange to some, because it stresses the coercive powers of government and fails to emphasize its enabling role in creating public goods—schools, roads, public security, hospitals, and welfare services—that allow individuals to exercise their freedom. They are positive goods, created by the consent of the governed. Yet majority consent does not eliminate the problem of minority constraint. These positive goods are paid for by a coercive measure—taxation—which most, but not all, citizens accept for the sake of the greater good. Not all citizens will agree about how much of their private income should be taxed to support this public infrastructure, nor how extensive this infrastructure should be. Disputes about this constitute the largest part of public politics, and the arbitration of these disputes, by legislation and by elections, inevitably leaves some citizens convinced that their freedom has been unduly constrained. There is simply no consensus about the proper extent of public goods or about the proper extent of government's power. At the margins, the constraint intrinsic to government will be experienced, at least by some citizens, as a lesser evil, to be submitted to as a condition of public life.

It might be asked whether coercive yet necessary uses of government power deserve to be called an evil at all. Taxation may be unpopular but hardly counts as an evil. Yet other acts of government, like punishment, which inflict direct harm on individuals, do raise the specter of evil. Or at least they do in our type of society. Only liberal democracies have a guilty conscience about punishment. Totalitarian societies have enthusiastically embraced coercion as a positive social instrument to create desired social types, ideal workers, obedient citizens, enthusiastic party apparatchiks. Only in liberal societies have people believed that the pain and suffering involved in depriving people of their liberty must make us think twice about imposing this constraint even on those who justly deserve it.23 The fact that it is necessary and the fact that it is just do not make it any less painful. It is necessary that criminals be punished, but the suffering that punishment causes remains an evil nonetheless.

It might be said that this example fails to distinguish necessary actions that cause harm from unnecessary actions caused by malice or gross negligence. Another way of saying the same thing is that a necessary evil cannot really be an evil at all, since it is a characteristic of evil that it is not necessary but gratuitous. I still want to hold on to the idea of the lesser evil, because it captures the idea, central to liberal theory, that necessary coercion remains morally problematic.

To insist that justified exercises of coercion can be defined as a lesser evil is to say that evil can be qualified. If two acts are evil, how can we say that one is the lesser, the other the greater? Qualifying evil in this way would seem to excuse it. Yet it is essential to the idea of a lesser evil that one can justify resort to it without denying that it is evil, justifiable only because other means would be insufficient or unavailable. Using the word evil rather than the word harm is intended to highlight the elements of moral risk that a liberal theory of government believes are intrinsic to the maintenance of order in any society premised upon the dignity of individuals.

Thus even in times of safety, liberal democracies seek to limit the use of force necessary to their maintenance. These limits seek to balance the conflict between the commitments to individual dignity incarnated in rights and the commitments to majority interest incarnated in popular sovereignty. In times of danger, this conflict of values becomes intense. The suppression of civil liberties, surveillance of individuals, targeted assassination, torture, and preemptive war put liberal commitments to dignity under such obvious strain, and the harms they entail are so serious, that, even if mandated by peremptory majority interest, they should be spoken of only in the language of evil.

In a war on terror, I would argue, the issue is not whether we can avoid evil acts altogether, but whether we can succeed in choosing lesser evils and keep them from becoming greater ones. We should do so, I would argue, by making some starting commitments—to the conservative principle (maintaining the free institutions we have), to the dignity principle (preserving individuals from gross harms)—and then reasoning out the consequences of various courses of action, anticipating harms and coming to a rational judgment of which course of action is likely to inflict the least damage on the two principles. When we are satisfied that a coercive measure is a genuine last resort, justified by the facts as we can understand them, we have chosen the lesser evil, and we are entitled to stick to it even if the price proves higher than we anticipated. But not indefinitely so. At some point—when we “have to destroy the village in order to save it”—we may conclude that we have slipped from the lesser to the greater. Then we have no choice but to admit our error and reverse course. In the situation of factual uncertainty in which most decisions about terrorism have to be taken, error is probably unavoidable.

It is tempting to suppose that moral life can avoid this slope simply by avoiding evil means altogether. But no such angelic option may exist. Either we fight evil with evil or we succumb. So if we resort to the lesser evil, we should do so, first, in full awareness that evil is involved. Second, we should act under a demonstrable state of necessity. Third, we should chose evil means only as a last resort, having tried everything else. Finally, we must satisfy a fourth obligation: we must justify our actions publicly to our fellow citizens and submit to their judgment as to their correctness.

III

The challenge in assessing which measures might be permissible is to find a viable position between cynicism and perfectionism. Cynicism would maintain that ethical reflection is irrelevant: the agents of the state will do what they will do, and the terrorists will do what they will do, and force and power alone will decide the outcome. The only question to ask about these means is whether they work. The cynics are wrong. All battles between terrorists and the state are battles for opinion, and in this struggle ethical justifications are critical, to maintain the morale of one's own side, to hold the loyalty of populations who might otherwise align with the terrorists, and to maintain political support among allies. A counterterror campaign probably can be run only by cynics, by professionals schooled in the management of moral appearances, but even cynics know that some moral promises have to be kept if they are to be believed at all. Preventive detention to withdraw suspicious aliens from the general population might disrupt terrorist networks, but it might so enrage innocent groups that they would cease to cooperate with the police. Torture might break apart a network of terrorist cells, but it would also engender hatred and resentment among the survivors of the torture and further increase their support among disaffected populations. There is simply no way to disentangle the technical question of what works from the political question of what impact such methods will have on the struggle for opinion that is the essence of any campaign against terror. Extreme measures, like torture, preventive detention, and arbitrary arrest, typically win the battle but lose the larger war. Even cynics know that Pyrrhic victories are worse than useless.

As for moral perfectionism, this would be the doctrine that a liberal state should never have truck with dubious moral means and should spare its officials the hazard of having to decide between lesser and greater evils. A moral perfectionist position also holds that states can spare their officials this hazard simply by adhering to the universal moral standards set out in human rights conventions and the laws of war.

There are two problems with a perfectionist stance, leaving aside the question of whether it is realistic. The first is that articulating nonrevocable, nonderogable moral standards is relatively easy. The problem is deciding how to apply them in specific cases. What is the line between interrogation and torture, between targeted killing and unlawful assassination, between preemption and aggression? Even when legal and moral distinctions between these are clear in the abstract, abstractions are less than helpful when political leaders have to choose between them in practice. Furthermore, the problem with perfectionist standards is that they contradict each other. The same person who shudders, rightly, at the prospect of torturing a suspect might be prepared to kill the same suspect in a preemptive attack on a terrorist base. Equally, the perfectionist commitment to the right to life might preclude such attacks altogether and restrict our response to judicial pursuit of offenders through process of law. Judicial responses to the problem of terror have their place, but they are no substitute for military operations when terrorists possess bases, training camps, and heavy weapons. To stick to a perfectionist commitment to the right to life when under terrorist attack might achieve moral consistency at the price of leaving us defenseless in the face of evildoers. Security, moreover, is a human right, and thus respect for one right might lead us to betray another.

A lesser evil morality is antiperfectionist in its assumptions. It accepts as inevitable that it is not always possible to save human beings from harm without killing other human beings; not always possible to preserve full democratic disclosure and transparency in counterterrorist operations; not always desirable for democratic leaders to avoid deception and perfidy; not always possible to preserve the liberty of the majority without suspending the liberties of a minority; not always possible to anticipate terrible consequences of well-meant acts, and so on. Far from making ethical reflection irrelevant, these dilemmas make ethical realism all the more essential to democratic reflection and good public policy. The fact that liberal democratic leaders may order the surreptitious killing of terrorists, may withhold information from their voters, may order the suspension of civil liberties need not mean that “anything goes.” Even if liberties must be suspended, their suspension can be made temporary; if executives must withhold information from a legislature in public, they can be obliged to disclose it in private session or at a later date. Public disinformation whose sole purpose is to deceive the enemy might be justified, but deliberately misleading a democratic electorate with a view to exaggerating risk or minimizing hazard can never be. The same balancing act needs to be observed in other cases. If the targeted killing of terrorists proves necessary, it can be constrained by strict rules of engagement and subjected to legislative oversight and review. The interrogation of terrorist suspects can be kept free of torture. Drawing these lines means keeping in clear sight the question of whether these means reinforce or betray the democratic identity they are supposed to defend.

Keeping lesser evils from becoming greater ones is more than a matter of democratic accountability. It is also a matter of individual conscience. Hannah Arendt once argued that being able to think for yourself is a precondition for avoiding evil, especially in large bureaucracies where there is a premium against independent thought. She said that the one common denominator uniting opponents of Nazi rule in Germany was a capacity to ask, at all times, what kind of person one was or wished to be. Those who refused to kill others, she said, “refused to murder, not so much because they still held fast to the command ‘Thou Shalt Not Kill’, but because they were unwilling to live together with a murderer—themselves.”24

No society can avoid official crimes and brutality unless this sense of responsibility is widely shared among public officials. Rules and procedures are not enough. Character is decisive, and there is some reason to think that democracies encourage the right kind of character. People who grow up in societies with constitutional rights are taught to believe that their opinions matter, that they are entitled to a certain fairness and due process in official dealings, and that they have a responsibility to the rights of others. But we cannot be sure that democracy teaches us all to do the right thing.

Moreover, no matter how good our moral learning we all stand in need of the scrutiny of good institutions. A war on terror puts these institutions under strain. It is not always possible to subject intelligence agents and Special Forces to full democratic scrutiny and control. Yet the agents themselves remain citizens, and their responsibility to the constitutional order they defend remains the tribunal of last resort to save them and us from a descent into barbarism. We depend for much of what we know about abuses of power on whistle-blowers, honest people who could not stand what they were being asked to do.25 Any democracy that wants to fight a clean war on terror needs to safeguard the rights of whistle-blowers, in the most secret agencies of government, to tell the truth to elected officials and the media. The only way to prevent zones of impunity from opening up in our government is for legislatures to insist on their rights of oversight, for the media to continue to demand access, and for the law to sustain the rights of whistle-blowers to tell the truth.

But these are not the only moral checks on a war on terror. Internationally ratified human rights instruments, together with the UN Charter and the Geneva Conventions, widen the audience of justification beyond the electorates of democratic states under direct attack, to a broader network of states and international bodies, whose views must be taken into account. Their views matter because liberal democracies constitute a community of values as well as a community of interests, and successful joint action against terrorism will soon become impossible if states disregard their allies, ignore their objections to national policies, and seek unilateral advantage or exemption from international commitments.26

International standards matter but we must not assume that nations always agree about what they mean. European countries disagree with the United States about the legitimacy of the death penalty, and they have refused to extradite terrorist suspects to the United States, where capital punishment may be the penalty. International conventions prohibit torture, but the exact point at which intensive interrogation passes over the redline into torture is a matter of dispute. The Geneva Conventions protect the idea of civilian immunity, but who counts as a civilian remains controversial. International conventions set standards, but each country may interpret them differently. How political leaders do so depends on what their domestic electorates appear to allow. But a political standard is not necessarily an ethically normless or relativist one. Public opinion will not accept simply anything. The norms that govern a war on terror are not the monopoly of government. They are susceptible to influence by moral entrepreneurship. Human rights activists and members of civil liberties NGOs will seek to raise the barrier of the morally permissible, while groups representing the military and the police may want to lower it. In any liberal democracy, standards for a war on terror will be set by adversarial moral competition.

As a contribution to this process of standard setting, I would propose the following tests for policy makers. First, a democratic war on terror needs to subject all coercive measures to the dignity test—do they violate individual dignity? Foundational commitments to human rights should always preclude cruel and unusual punishment, torture, penal servitude, and extrajudicial execution, as well as rendition of suspects to rights-abusing countries. Second, coercive measures need to pass the conservative test—are departures from existing due process standards really necessary? Do they damage our institutional inheritance? Such a standard would bar indefinite suspension of habeas corpus and require all detention, whether by civil or military authorities, to be subject to judicial review. Those deprived of rights—citizens and noncitizens—must never lose access to counsel. A third assessment of counterterror measures should be consequentialist. Will they make citizens more or less secure in the long run? This effectiveness test needs to focus not just on the short term, but on the long-term political implications of measures. Will they strengthen or weaken political support for the state undertaking such measures? A further consideration is the last resort test: have less coercive measures been tried and failed? Another important issue is whether measures have passed the test of open adversarial review by legislative and judicial bodies, either at the time, or as soon as necessity allows. Finally, “decent respect for the opinions of mankind,” together with the more pragmatic necessity of securing the support of other nations in a global war on terror, requires any state fighting terrorism to respect its international obligations as well as the considered opinions of its allies and friends. If all of this adds up to a series of constraints that tie the hands of our governments, so be it. It is the very nature of a democracy that it not only does, but should, fight with one hand tied behind its back. It is also in the nature of democracy that it prevails against its enemies precisely because it does.

• 1.

  Cass R. Sunstein, Designing Democracy: What Constitutions Do (New York: Oxford University Press, 2001), 6–8, 13–47. Dennis Thompson and Amy Gutmann, Democracy and Disagreement (Cambridge: Harvard University Press, Belknap Press, 1996), 41–49.

• 2.

  John Hart Ely, Democracy and Distrust: A Theory of judicial Review (Cambridge: Harvard University Press, 1980), 4: “thus the central function, and it is at the same time the central problem, of judicial review: a body that is not elected or otherwise politically responsible in any significant way is telling the people's elected representatives that they cannot govern as they'd like… This in America is a charge that matters.”

• 3.

  William H. Rehnquist, All the Laws but One: Civil Liberties in Wartime (New York: Knopf, 1998), 222.

• 4.

  The Federalist No. 51, http://memory.loc.gov/const/fed/fedpapers.html (accessed December 4, 2003).

• 5.

  For an account of how a Supreme Court seeks to balance the two meanings in times of terrorism, see A. Barak, “A Judge on Judging: The Role of a Supreme Court in a Democracy,” Harvard Law Review 116, no. 1 (November 2002): 16–162, and especially 36–46 and 148–60.

• 6.

  Ronald Dworkin, “Philosophy and Monica Lewinsky,” New York Review of Books, March 9, 2000; Ronald Dworkin, “Posner's Charges: What I Actually Said” (2000), http://www.nyu.edu/gsas/dept/philo/faculty/dworkin/ (accessed December 4, 2003). Ronald Dworkin, Freedom's Law: The Moral Reading of the American Constitution (Cambridge: Harvard University Press, 1996), vs. Richard Posner, The Problematics of Moral and Legal Theory (Cambridge: Harvard University Press, Belknap Press, 1999).

• 7.

  Peter Irons, Justice at War (New York: Oxford University Press, 1983), 9–13, 57–64. David Cole, “An Ounce of Detention,” American Prospect, September 9, 2003.

• 8.

  USA Today/CNN Gallup Poll Results (August 2003), http://www.lifeandliberty.gov/subs/s_people.htm (accessed December 4, 2003). To the question “Do you think the Bush Administration has gone too far, has been about right or has not gone far enough in restricting civil liberties in order to fight terrorism?” between 55 and 60 percent of respondents say “about right,” the numbers remaining stable since June 2002. Forty-eight percent of the sample believe that the Patriot Act gets the balance between liberty and security “about right.”

• 9.

  David Cole, Enemy Aliens (New York: New Press, 2003).

• 10.

  For the idea of an enemy of the human race, see the discussion in Hannah Arendt, Eichmann in Jerusalem: A Report on the Banality of Evil (New York: Viking, 1963).

• 11.

  Ronald Dworkin, “Terror and the Attack on Civil Liberties,” New York Review of Books, November 6, 2003.

• 12.

  Cole. Enemy Aliens.

• 13.

  U.S. Justice Department Inspector General Report on Administrative Detention (June 2003), http://www.usdoj.gov/oig/special/03–06/index.htm (accessed December 4, 2003).

• 14.

  Robert D. Marcus and Anthony Marcus, eds., “The Army-McCarthy Hearings, 1954,” in American History through Court Proceedings and Hearings, vol. 2 (St. James, N.Y.: Brandywine Press, 1998), 136–51.

• 15.

  Dennis F. Thompson, Political Ethics and Public Office (Cambridge: Harvard University Press, 1987), 118; also Dennis Thompson, “Democratic Secrecy,” Political Science Quarterly 114, no. 2 (Summer 1999): 181–93. See also U.S. Senate Final Report of the Select Committee to Study Governmental Operations with Respect to Intelligence Activities (Washington, D.C.: U.S. Government Printing Office, 1976), 11–14.

• 16.

  Israeli Supreme Court Judgment on the Interrogation Methods Applied by the GSS (September 6, 1999), http://www.us-israel.org/jsource/Society_&_Culture/GSS.html (accessed December 4, 2003):

  Moreover, the “necessity” defense has the effect of allowing one who acts under the circumstances of “necessity” to escape criminal liability. The “necessity” defense does not possess any additional normative value. In addition, it does not authorize the use of physical means for the purposes of allowing investigators to execute their duties in circumstances of necessity. The very fact that a particular act does not constitute a criminal act (due to the “necessity” defense) does not in itself authorize the administration to carry out this deed, and in doing so infringe upon human rights.

• 17.

  Michael Ignatieff, Virtual War: Kosovo and Beyond (New York: Metropolitan, 2000), conclusion.

• 18.

  Euripides, Medea, trans. Rex Warner (New York: Dover Publications, 1993), 40.

• 19.

  Niccolò Machiavelli, The Prince, trans. and ed. Robert M. Adams (New York: Norton, 1977), 44–45, “On the Reasons Why Men Are Praised or Blamed—Especially Princes”: “And again, he need not make himself uneasy at incurring a reproach for those vices without which the state can only be saved with difficulty, for if everything is considered carefully, it will be found that something which looks like virtue, if followed, would be his ruin; whilst something else, which looks like vice, yet followed brings him security and prosperity.” Isaiah Berlin, “The Originality of Machiavelli,” in Against the Current (New York: Viking, 1980).

• 20.

  Isaiah Berlin, “Two Concepts of Liberty” in Four Essays on Liberty (Oxford: Oxford University Press, 1969), 165:

  If I wish to preserve my liberty, it is not enough to say that it must not be violated unless someone or other—the absolute ruler, or the popular assembly, or the King in Parliament, or the judges, or some combination of authorities, or the laws themselves—for the laws may be oppressive—authorizes its violation. I must establish a society in which there must be some frontiers of freedom which nobody should be permitted to cross. Different names or natures may be given to the rules that determine these frontiers: they may be called natural rights, or the word of God, or Natural Law, or the demands of utility or of the “permanent interests of man”; I may believe them to be valid a priori, or assert them to be my own ultimate ends, or the ends of my society or culture. What these rules or commandments will have in common is that they are accepted so widely, and are grounded so deeply in the actual nature of men as they have developed through history, as to be, by now, an essential part of what we mean by being a normal human being. Genuine belief in the inviolability of a minimum extent of individual liberty entails some such absolute stand. For it is clear that it has little to hope for from the rule of majorities; democracy as such is logically uncommitted to it, and historically has at times failed to protect it, while remaining faithful to its own principles. Few governments, it has been observed, have found much difficulty in causing their subjects to generate any will that the government wanted. “The triumph of despotism is to force the slaves to declare themselves free.” It may need no force; the slaves may proclaim their freedom quite sincerely: but they are none the less slaves. Perhaps the chief value for liberals of political—“positive”—rights, of participating in the government, is as a means for protecting what they hold to be an ultimate value, namely individual—“negative”—Liberty.

• 21.

  Michael Ignatieff, “Genocide: An Essay,” in Simon Norfolk, For Most of It I Have No Words (London: Dewi Lewis, 1998).

• 22.

  John Stuart Mill, “On Liberty” (1869), in On Liberty, ed. Edward Alexander (Peterborough, Ont: Broadview Press, 1999), pp. 51–52:

  The object of this Essay is to assert one very simple principle, as entitled to govern absolutely the dealings of society with the individual in the way of compulsion and control, whether the means used be physical force in the form of legal penalties, or the moral coercion of public opinion. That principle is that the sole end for which mankind are warranted, individually or collectively, in interfering with the liberty of action of any of their number, is self-protection. That the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others. His own good, either physical or moral, is not a sufficient warrant. He cannot rightfully be compelled to do or forbear because it will be better for him to do so, because it will make him happier, because, in the opinions of others, to do so would be wise, or even right. These are good reasons for remonstrating with him, or reasoning with him, or persuading him, or entreating him, but not for compelling him, or visiting him with any evil in case he do otherwise. To justify that, the conduct from which it is desired to deter him, must be calculated to produce evil to some one else. The only part of the conduct of any one, for which he is amenable to society, is that which concerns others. In the part which merely concerns him, his independence is, of right, absolute. Over himself, over his own body and mind, the individual is sovereign.

• 23.

  See my A Just Measure of Pain: Penitentiaries in the Industrial Revolution (New York: Pantheon, 1978) for a discussion of Beccarian and post-Beccarian theories of punishment in the European Enlightenment.

• 24.

  Hannah Arendt, “Personal Responsibility under Dictatorship,” in Responsibility and Judgment, ed. Jerome Kohn (New York: Schocken Books, 2003), 44.

• 25.

  Daniel Ellsberg, Secrets: A Memoir of Vietnam and the Pentagon Papers (New York: Viking, 2002). For a whistle-blower's career in the CIA, see Robert Baer, See No Evil: The True Story of a Ground Soldier in the CIA's War on Terrorism (New York: Crown Publishers, 2002).

• 26.

  See Michael Ignatieff, ed., American Exceptionalism and Human Rights (Princeton: Princeton University Press, 2004).