At the height of the siege of Sarajevo, when hundreds of shells were hitting the Bosnian capital every day, many there believed they would not survive. The United Nations airlift, which provided food for many who would otherwise have starved, was constantly being interrupted as Serb forces repeatedly shelled the airport runway. As the months turned into years and the siege was not lifted, it seemed, both to Sarejevans and to many sympathetic foreigners, that either the city would be destroyed or some kind of outside intervention would have to take place.

"How can you in the West allow this to go on?" an actress at Sarajevo’s National Theater asked me in January of 1993. "My teeth are falling out, I’m covered in eczema, I haven’t had a bath in months, and I’m a privileged person by the standards of this town. I know no one outside really understands what is going on here; you all think we’re savage Balkan people pursuing our ancient ethnic bloodlusts. It’s utter nonsense, of course. But even if that’s what you think, even if you won’t come help us because we have right on our side, why can’t you help us for humanitarian reasons? Why can’t you just stop the siege?"

Overhead, as if it had been choreographed, a jet screamed through the afternoon sky. It was a NATO fighter on a routine reconnaissance. The actress smiled. "Good timing, wasn’t it?" she said. "If it would dive down and drop some bombs, I could have a bath."

She was right about the befuddlement afflicting most people in the U.S. and in western Europe when they thought about Bosnia. Those who believed that the Bosnian government should be aided because it was in the right were always in a small minority, even among those who thought force needed to be used to bring the ethnic cleansing and the siege of Sarajevo to an end. It was more common to hear the argument that what was happening was so horrible that it simply had to be stopped, and that if military means were needed then military means had to be employed. In other words, what was most credible to the vast majority of those who paid attention to Bosnia—even at the height of the slaughter, a small percentage of the Western public—was not a political intervention but an intervention that needed to be undertaken solely on humanitarian grounds.

In the end, when the mass murder of more than seven thousand Muslim men and boys at Srebrenica finally forced the Western powers to act, they did so not out of a political determination to restore a unified Bosnian state but instead out of a version of a humanitarian impulse. The bloodshed and the slaughter, Western leaders finally concluded, could not be allowed to go on.

Humanitarian intervention is at once an immensely powerful and a terribly imprecise idea. No formal legal definition of it exists, but its fundamental premise is that outside powers have the right and, perhaps, under some circumstances, the duty to intervene to protect people in other countries who are being victimized, even if what is taking place is a conflict within a State. Whereas classical interventions are political in character and involve one State either imposing its will by force on another or coming to the aid of another (and thus in no sense challenging the long-standing notion that State sovereignty should be for all practical purposes inviolable), humanitarian interventions offer a direct challenge to such notions of sovereignty. This is especially true for those interventions directly into the internal affairs of a single State. In a deep sense, they also sidestep considerations of the political rights and wrongs of a given conflict. What matters, from the perspective of the State or group of States contemplating a humanitarian intervention, is the effect a conflict has on civilians.

An example of this kind of thinking was the debate in 1996 over a Canadian proposal for a humanitarian intervention in what was then eastern Zaire to protect the millions of Hutu refugees who were at risk from both the attacks of Tutsi-led Rwandan forces and from the ebb and flow of the Zairean civil war. The Canadians argued that the rights of a civilian population at risk outweighed any other consideration, including the effect that such a humanitarian deployment might have on the political struggle then taking place in Zaire. Those who argued against the deployment in effect were saying that humanitarian needs alone could not justify such outside interference. Many also made the cautionary argument that forecasting the long-term effect of humanitarian military intervention was itself fraught with uncertainty.

As a matter of international law, humanitarian intervention remains purely a matter of the political preferences of the person making the argument. It seems generally accepted that the Security Council can declare anything it likes to be a "threat to international peace and security," subject not to any genuinely objective constraints of law but only to the political vetoes of its permanent members. Its determinations are shaped by public opinion, international activists, CNN, and the political considerations of Security Council members, as well as by broader, and more principled policies and laws. Within international humanitarian law (IHL), the leap from provisions providing for the delivery of humanitarian relief to military intervention is a long one, but not too long for those politically motivated to do so.

In practice, humanitarian intervention has often served as a justification for States to act in conflicts where there is no domestic support for more straightforward political interventions. The public in North America and western Europe has, for all the talk of compassion fatigue, proved remarkably sympathetic to the use of force to avert or bring to an end a humanitarian disaster. On the other hand, humanitarian intervention has also been a justification for other political motivations. In Rwanda, in 1994, it was commonly assumed that the French humanitarian intervention dubbed Operation Turquoise used the humanitarian imperative as a cover for France’s decision to continue to try to influence events in the Great Lakes region of Africa with military force and, more specifically, to save the French-supported, but genocidal, government. And historically, many of the imperial campaigns launched by the European colonial powers in the nineteenth century were justified on humanitarian grounds.

Today, humanitarian interventions have been largely the brainchildren of UN bureaucrats and of humanitarian relief organizations who are unable to operate safely in conflict zones. These groups have become some of the most fervent interventionists. A French humanitarian official, Bernard Kouchner, even popularized the legal theory of the French scholar Mario Bettati of the "right of intervention." Of course, whether people would react in the same way if the price in the lives of soldiers and in money grew too high is another question entirely. This is in part due to the fact that Western politicians routinely, but often improbably, describe humanitarian interventions as safe. When, as with the U.S.-led United Nations Operation in Somalia (UNOSOM II), this turns out to be false, pressures to abort humanitarian interventions mount quickly and usually become politically irresistible.

In international law, the tug of war between State sovereignty and the obligation of other countries to maintain international peace and security continues to rage. In practice, the idea, at least in the case of so-called failed States, that the principle of State sovereignty is simply irrelevant and the needs of persons preeminent has been gaining favor and is becoming a common starting point for discussions of what to do about situations where there is immense human suffering and little prospect of immediate relief. That was the way the U.S. intervention in Somalia was presented to the American public, and how the Canadians made the case for their proposed deployment in eastern Zaire. As the costs of humanitarian interventions mount, the willingness of outside States to intervene to protect people from systematic violations of internationally recognized human rights or provide them with relief is an open question. But humanitarian intervention remains an immensely attractive idea to many, and in the absence, after the end of the Cold War, of either any real system of international security or any real transfer of authority to supranational institutions like the UN, it is likely to remain an enduring, if half-hearted, one.