The taking of hostages poses an acute moral and practical dilemma for a democratic society. Justice and compassion prompt us to seek the release of hostages at all costs, including negotiating with terrorists and — what negotiation normally entails — making at least some concessions. But the harsh reality is that while a deal may free innocent victims, by rewarding terrorism it encourages repetition. Hence, a choice between two evils.

Yet we need not remain at that impasse. Each hostage crisis is different, and all are fraught with human and diplomatic pressures and constraints. But like any problem of social choice, they also reflect an underlying moral question of interests and ends. In fact, the hostage dilemma can be seen to reflect a conflict between two great ethical traditions in Western thought.

One of these derives mainly from Kant, and emphasizes the transcendent value of the individual. Kant argues that people should never be treated as means, only as ends. Individual freedom is the central value in Kantian ethics, and in Western society, because freedom is instrumental to all other values. Kant’s theory captures, with exquisite analytic subtlety, much of what the moral enterprise is all about; that is why, despite the complexity of much of his writing, his legacy as a thinker and humanist is so profound in the liberal tradition.

The other theory is utilitarianism. Expounded by the 19th-century British philosophers Jeremy Bentham and John Stuart Mill, it sacrifices the individual to the collective, basing morality on the idea of achieving “the greatest good for the greatest number.” While in some ways intuitively appealing, utilitarianism has never succeeded in defining or quantifying the “greatest good”, reconciling competing goods, or weighing intensity of satisfaction against number of people satisfied. Today it is a largely discredited doctrine, and for good reason.

Yet despite the flaws of utility, and the virtues of Kantian ethics, as broad moral theories, an interesting reversal occurs when one considers a particular class of concrete moral problems. This class might be called “dire cases”: cases where the stakes are not just competing rights and interests, or the ordinary moral decencies, but human life itself, on a terrifying scale. Hostages constitute just such a case.

Dire cases are unique because, whereas interests may be weighed and compared, lives can only be counted; if, as Kant suggests, every life is of equal and infinite worth, then in dire cases all we can do is try to save the most innocent lives overall. This means we must not make real concessions to terrorists if doing so will at best save some lives in the short run. When it comes to respecting the value of individuals (or distinguishing between those who use innocent people for political ends and true “freedom fighters”) we are Kantians; but when it comes to saving human life, we must be more like utilitarians. This is not a corruption of the liberal-individualist ethic. Kantian principles are not invalidated or embarrassed by dire cases; they simply don’t apply. Likewise, I doubt that Bentham and Mill had dire cases in mind, yet it’s in precisely those cases — characteristically modern, tragic, and extreme, with moral calculation reduced to a crude arithmetic — that their theory makes sense.

Not surprisingly, dire cases abound in wartime. A striking example was Winston Churchill’s decision not to evacuate Coventry in World War II, despite foreknowledge of a German air raid, to guard the secret that the Allies had cracked the Enigma codes. Disclosure would have cost an incalculable strategic advantage and many lives, if not the war itself. On a smaller scale, terrorism poses a similarly dreadful choice between saving some innocent lives now and uncounted others in the future. The apparently humane choice isn’t necessarily the most humane. However, there are some consolations. For one thing, it is not in principle a political decision. The question is one of means, not of ends. Terrorism, like Nazism, is a challenge that all democratic nations must face in concert. Moreover, the choice is seldom so stark until intermediate strategies have been exhausted. Authorities must first learn who the terrorists are, what they want, and how flexible they may be, while also gaining time to prepare other diplomatic or tactical moves. It is the stated policy of the U.S. Government not to negotiate with terrorists. But the real problem isn’t talking; it’s making concessions, such as the shipment of arms to Iran.

A state that is the target of terrorist demands must do everything possible to secure the safe release of hostages, except capitulate. Here Kant and Mill would doubtless agree: terrorism is an act of war that must be discouraged, even at the risk of some innocent life — preferably by expert diplomacy, but if all else fails, by force.  ■