Hiding in her house in Kigali, hearing the cries of her neighbors as they were being slaughtered, Monica Uwimana did not know where to turn for help. Gangs of youths were moving from house to house with machetes and nail-studded clubs killing Tutsis like her. She called UN employees. They said they could not help. She called a foreign reporter and asked for advice. The journalist had none.

Many Rwandans believe that UN peacekeepers had an obligation to help them. But the UN Security Council withdrew most of the troops at the height of the Rwanda genocide in April 1994 and did not order a new force in until it was too late.

In theory, civilians can turn to a neutral country and try to make it to a foreign embassy. But even if Monica Uwimana could have gotten past the roadblocks, she might have been turned away, for most embassies take in and protect only victims who have a proven link to their country, or whose own country has formally requested that its citizens be protected.

Rwanda was the extreme example of a humanitarian crisis in which the international community and its representatives all but disappeared from the scene when they were most needed. And it encapsulates the predicament for victims of armed conflict, of crimes against humanity, or, as in this instance, of genocide. States have the obligation to pay compensation for violating international conventions and are responsible for the misdeeds of their troops. The Hague Convention of 1907 states: “A belligerent party which violates the provisions of the said Convention shall, if the case demands, be liable to pay compensation. It shall be responsible for all acts committed by persons forming part of its armed forces.” Similar language is found in the first Additional Protocol of 1977. Individuals responsible for war crimes, crimes against humanity or genocide may in time face criminal charges before institutions like the war crimes tribunals for the former Yugoslavia and Rwanda, or the recently formed International Criminal Court. Ultimately in these cases there will be accountability. But at the moment of maximum violence, protection is almost nonexistent.

The International Committee of the Red Cross (ICRC) is one of the first places victims turn. Its mandate is based on the Geneva Conventions: to protect the most vulnerable individuals, be they prisoners of war or civilians who come under attack; to trace the missing and reunite them with their families; to supervise repatriation of prisoners; and to remind all sides in a conflict that they are obliged to uphold the conventions. During the Rwanda genocide, its expatriate and local staff expanded their mandate and sheltered as many as nine thousand Rwandans in the ICRC compound. But the ICRC was overwhelmed. Its representatives collected thousands of wounded civilians and brought them to hospital. That did not spell protection, for on one occasion, soldiers entered the hospital after the foreigners had departed and massacred the patients. On another, Tutsis were dragged out of ICRC vehicles and killed. In parts of Rwanda, local staff of the national Red Cross participated in the killing.

The UN High Commissioner for Refugees (UNHCR), by its mandate, attempts to provide protection for those who flee their country because of conflict. The massive growth in the number of internally displaced persons (IDPs) has led UNHCR to set up operations in an increasing number of countries to handle them. And an ever-growing number of non-governmental organizations (NGOs) are attempting to help the victims of armed conflict. But few national or international NGOs operate in the midst of the conflict, unless an outside power guarantees their security. And the conflict in Bosnia shows the limits of such protection, for outside military force was deployed to protect convoys of food and medicine attempting to reach UN-declared “safe areas,” but did almost nothing to safeguard the civilians or, when they came under attack, the safe areas as well.

In the post-Cold War era, aid organizations increasingly are targeted along with the civilians they are protecting. The head of the ICRC mission in Bosnia was targeted and killed in June of 1992, leading the ICRC to withdraw its entire presence from that country temporarily. It was then that Bosnian Serbs began their full-scale ethnic cleansing campaign. In Burundi, three ICRC delegates were ambushed on a main road and murdered in 1996. The same year, in Chechnya, gunmen broke into an ICRC compound and killed six delegates. And in 2003, a suicide bomber attacked the ICRC building in Baghdad, killing two officials and ten bystanders. Attacks on those who try to help victims are the most dramatic evidence of the total disregard for the laws of war that marks contemporary conflict.

Under international law, States are obliged to prevent and, that failing, punish grave breaches of the Geneva Conventions and the first Additional Protocol. For reasons of domestic politics, States often prefer to stay out of each others’ conflicts; and the judicial instruments to justify intervention have at best weak enforcement mechanisms. The 1948 Genocide Convention, which requires States parties to “prevent and to punish” genocide, contains no mechanism for determining if a genocide is under way. Article 90 of Additional Protocol I of 1977 set up an International Humanitarian Fact-Finding Commission to investigate charges of war crimes, but it requires the consent of both parties to function and has been in general disuse. The newest attempt is the UN’s recent creation of an International Criminal Court, whose prosecutor is empowered to launch investigations into war crimes and crimes against humanity while crimes are underway.

As for compensation, victims have a right under international and domestic law to sue, but after a genocide such as Rwanda’s, there is rarely anyone rich enough left to pay the damages. The victims of the Nazi Holocaust received reparations from successor German governments. But “comfort women,” used as sex slaves by Japanese troops during World War II, have yet to receive State compensation fifty years later, even though the government admitted in 1993 that Japan’s military was responsible for the crime. On April 27, 1998, the Tokyo district court ordered compensation paid to three Korean women, but in March 2006, the Hiroshima High Court overturned this landmark decision. The UN Commission on Human Rights has advocated a set of principles to assist victims of human rights violations. The principles include a family’s right to know the fate of missing people, the right to reparations, the right to criminal justice, and the principle that governments take steps to prevent recurrences. These principles are only recommendations, however.

By luck, chance, and the grace of God, Monica Uwimana survived. Her five children, who had been staying with grandparents in the countryside, were murdered. She has little faith in international law.

Two emblematic figures, the exile and the refugee, loom large in our consciousness at the beginning of this new century as at the end of the last. Vast refugee flows have become a feature of the contemporary world. Part of this can be attributed to the fact that it is easier to cross borders now than it was, say, 150 years ago. People have always wanted to move away from danger zones; today, whatever the risks, they have the means to do so. As a result, when Hurricane Mitch devastated Honduras in 1998, many predicted that within a matter of months Honduran refugees would pour into the United States.

The Honduran refugees were fleeing a natural disaster and, as such, were something of an anomaly, since refugees from natural disasters constitute only a small percentage of the 34 million refugees and displaced people in the world. When most people think of refugees they usually think of the victims of political repression, as in East Timor, religious persecution, as in Tibet, or of civilians fleeing a war zone in which they have become the targets, as in a dozen conflicts from the Democratic Republic of Congo to Kosovo.

Of course, there is nothing new about repression, and the question of whether war has grown more barbarous in the twentieth century remains a controversial one. Those who, in the aftermath of World War II, devised the international humanitarian law (IHL) on refugees seemed to have believed that it had. So do those who have spent time in such killing zones as Bosnia, South Sudan, and the Great Lakes region of Africa. There, civilians are more often than not the preferred targets of the belligerents, and every villager caught in these maelstroms is, when viewed in a certain light, a potential refugee.

Perhaps this is why Sadako Ogata, the United Nations High Commissioner for Refugees, once observed that “refugees are the symptoms of the ills of an age”—our age. In almost every crisis that has plagued or baffled us since the end of the Cold War, from Tajikistan to Burundi, the refugee issue has been at the center. There has been no escaping it, and, it seems, no resolving it.

This is not because of an absence of laws, but, rather, a lack of implementation. Indeed, if the political will of the powerful nations of the world matched the legal protections that already exist for refugees, many of the cruelest tragedies of the last part of the twentieth century might have been greatly diminished. A raft of treaties passed in the aftermath of World War II, and, half a century later, the accumulating weight of customary law guarantees refugee rights. Unlike in the case, say, of internally displaced persons (the distinction is that a refugee is a person who crosses an international border), refugees enjoy a wide array of rights and protections including the right to certain kinds of legal aid and material assistance. Compared to what should be, the situation of refugees in the world today is appalling; compared to what it would be without IHL, it is at least not hopeless. The most important laws concerning refugees are the Fourth Geneva Convention of 1949, the 1977 First Additional Protocol to the Geneva Conventions, and the 1951 UN Convention Relating to the Status of Refugees. The Geneva Conventions required a certain humane standard of treatment for civilians who do not enjoy diplomatic status. The Fourth Convention granted refugees the right not to be returned to the country where they faced danger or could legitimately claim that they would be subject to religious or political persecution. Additional Protocol I extended the standard for civilian protection set in the 1949 Geneva Conventions to include all civilians regardless of their nationality.

The UN Convention Relating to the Status of Refugees defines who refugees are and how they are to be treated. For the most part, the guarantees the convention grants are the basic human rights outlined in other international legal instruments. Refugees are not to be returned to the place where they face persecution, nor, except on grounds of national security, are they to be expelled without due process. They are not to be treated as illegal aliens (a key right, particularly in Western Europe and North America where the authorities routinely try to claim that people claiming refugee status are really economic migrants). Their right to move about in their country of asylum is not to be unnecessarily restricted, and they are to be given identity papers if they do not have them.

In a world awash in refugees, where people in rich countries feel overwhelmed by the press of economic migrants, legal and illegal, and where people in poor countries in areas adjacent to conflict zones have neither the resources nor the expertise to deal with vast refugee flows (2 million Rwandan refugees crossed the border into Zaire in less than a week in the summer of 1994), this has meant that where refugees have been concerned, the situation has grown more and more difficult. In particular, the burdens on the main international organization charged with protecting refugees, the Office of the UN High Commissioner for Refugees, have become excruciating. More generally, the gap between the law and realities on the ground is greater in the area of refugee rights than almost any other. Perhaps that gap is an emblem of failure. But many refugee advocates say that it gives us the means actually to make the ideals of refugee protection a reality, whereas if no such body of law existed, and the idea of protection had not been enshrined within it, the situation would be even more dire than it is.

Military necessity is a legal concept used in international humanitarian law (IHL) as part of the legal justification for attacks on legitimate military targets that may have adverse, even terrible, consequences for civilians and civilian objects. It means that military forces in planning military actions are permitted to take into account the practical requirements of a military situation at any given moment and the imperatives of winning. The concept of military necessity acknowledges that even under the laws of war, winning the war or battle is a legitimate consideration, though it must be put alongside other considerations of IHL.

It would be overly simplistic to say that military necessity gives armed forces a free hand to take action that would otherwise be impermissible, for it is always balanced against other humanitarian requirements of IHL. There are three constraints upon the free exercise of military necessity. First, any attack must be intended and tend toward the military defeat of the enemy; attacks not so intended cannot be justified by military necessity because they would have no military purpose. Second, even an attack aimed at the military weakening of the enemy must not cause harm to civilians or civilian objects that is excessive in relation to the concrete and direct military advantage anticipated. Third, military necessity cannot justify violation of the other rules of IHL.

Moreover, the action in question has to be in furtherance of a military, not a political, goal. This poses obvious problems of characterization. Is persuading the enemy to surrender a military or political goal? Is “persuading” the enemy to surrender by aerial bombardment a military or political goal.

What constitutes a military objective will change during the course of a conflict. As some military objectives are destroyed, the enemy will use other installations for the same purpose, thereby making them military objectives and their attack justifiable under military necessity. There is a similarly variable effect on the determination of proportionality. The greater the military advantage anticipated, the larger the amount of collateral damage—often civilian casualties—which will be “justified” or “necessary.” This flexibility also appears with regard to the prohibition of the use of weapons that cause “superfluous injury or unnecessary suffering.” The greater the necessity, the more suffering appears to be justified. Thus, even in the Advisory Opinion on the Legality of the Use of Nuclear Weapons the majority of judges in the International Court of Justice in The Hague left open the possibility that a State might be able to justify its use of nuclear weapons where the very survival of the State was under serious threat.

State practice recognizes that judgments about military necessity often require subjective evaluations with incomplete information on the battlefield and imperfect knowledge of where the failure to take action might lead. For this reason, great discretion has always been attached to commanders’ judgments, especially those made under battlefield conditions. Rarely, if ever, is the judgement of a field commander in battle—balancing military necessity and advantage—subject to legal challenge, let alone criminal sanction. An exception would be when the method of warfare used by the commander was illegal per se, and therefore not covered by the claim of military necessity.

In some cases, there is a presumption that certain actions are unlawful; it was not possible to prohibit them in absolute terms but they are unlawful unless justified by “imperative military necessity.” This qualification of “absolutely necessary” or “for reasons of imperative military necessity” puts a significant burden of proof on those invoking the exception. Examples include the Fourth Geneva Convention, which restricts the internment of protected persons and the transfer or deportation from an area of occupied territory; Additional Protocol I, which would normally prohibit a scorched-earth policy but which allows it in exceptional circumstances in national territory; and Additional Protocol II, which normally prohibits the internal displacement of the civilian population.

In the course of hostilities, these rules impose significant restraints on the conduct of law-abiding forces, but those forces may be able lawfully to invoke military necessity where their very survival or the requirements of winning the conflict are at stake.

From a distance, the line of cars and buses that snaked along the rural stretch of northwest Bosnian roadway looked like a traffic jam on that languid August afternoon, as girls in summer white bicycled past fields of cornflower blue. Up close, the terror in the eyes of the people of Sanski Most, the belongings stuffed into satchels and plastic bags, and the police riding shotgun told the real story.

More than fifteen hundred Slavic Muslims were being forced from their homes that day in mid-August 1992 by ethnic Serbs trying to purge northern Bosnia of their neighbor-turned-enemy. The Serbs had even provided city buses to transport those without cars, though the generosity was soon to degrade into a terrifying all-night trek over blood, body parts, and across the front lines. Four months into the war, the UN High Commissioner for Refugees had begun to realize that the mass movement of civilians in Bosnia was not the chaotic happenstance of war, but rather a calculated, orchestrated transfer of populations aimed at creating ethnically pure areas.

Europe had not seen this kind of mass expulsion of civilians since World War II. Hitler’s atrocities across the continent had given rise to the 1949 Geneva Conventions, which include specific protections for civilians. Article 49 of the Fourth Geneva Convention declares that “individual or mass forcible transfers . . . are prohibited, regardless of their motive.”

Those in the Sanski Most convoy, by virtue of being moved to another location in the same country, are regarded under international humanitarian law as internally displaced persons (IDPs). If sent across an international border, it would be a deportation and they would be treated as refugees.

Additional Protocol II of 1977, which applies in internal conflicts, provides that forced civilian displacement may be undertaken legally only when civilians’ very safety or “imperative military reasons” require it. In addition, Article 17 says that civilians cannot be forced from their “whole territory” for reasons connected with the conflict. The article does not say unambiguously what is meant by “territory.” The International Committee of the Red Cross (ICRC) Commentary to the Additional Protocols states that the intent here is to minimize civilian displacement that is politically motivated.

The standard is the same for international or internal conflicts: if civilians have to be moved for either of those two reasons—safety or military imperatives—their evacuations are to be under protected, hygienic, and humane conditions, and as short-lived as possible. None of that applied in the case of Sanski Most, which fell under Bosnian Serb control in the earliest days of the war and saw no fighting for three years until autumn 1995.

International law, therefore, was unambiguous. Article 49 of the Fourth Geneva Convention prohibits the mass transfer of Bosnia’s civilians and Article 17 of Additional Protocol II prohibits the expulsions.

Despite IHL, here they were. The daylong journey had descended into the heart of darkness as the convoy turned from the main road onto isolated country paths and made its way through an increasingly hostile gauntlet of Serb soldiers and civilians shouting, “Butcher them, butcher them!” As the sun set, the worst was yet to come. With guns stuck in their faces, the people of Sanski Most were ordered from the buses and their cars, most of which were seized. With what belongings they could carry, they were sent on foot into the darkness through a no-man’s land separating two armies. I went with them.

Old and young, fit and feeble, we trekked along a mountain road that was cratered by mortar impacts, mined on one side, and covered by snipers. In places, the blood was so thick our shoes stuck momentarily to the road, and we stumbled onto chunks of human flesh and other remnants—teddy bears, backpacks, slippers—of those who had gone before. At the front line—a high wall of boulders dynamited from the mountainside—a crippled man was carried over, his wheelchair passed after him. Babies were handed to their mothers. Old men and women, stooped with age, struggled over the rocks to the government-controlled side. The foot journey lasted six hours and covered twelve miles.

By December 1995, just after the Dayton peace accord, at least 1.2 million Bosnians had been internally displaced. Three years later, only a fraction had returned home. “Civilians, not soldiers, were the principal and often intentional victims in the Bosnian conflict. Forced displacement was not just a by-product but an objective of military action and persecution,” Sadako Ogata, the UN High Commissioner for Refugees, said at the time.

But three years later, it happened again in the Serbian province of Kosovo, where Serb troops spent much of 1998 crushing an uprising by the 90 percent ethnic Albanian population. As many as 350,000 people had been displaced by September 1998. The majority were women and children.

Not all internal displacements are ethnic cleansing. In Colombia, both sides in the government-rebel conflict have forcibly relocated civilian populations to gain political or economic advantage. Worldwide, there were an estimated 20 to 25 million IDPs in 1998, compared to about 13 million refugees, that is, those forced across international borders.

Legal and humanitarian experts concede existing law is insufficient and weakened by a lack of political will. Interpretations of “imperative military reasons” vary widely, and governments also worry about getting involved in what may be considered the internal affairs of another State. According to Francis Deng, the UN’s top representative on IDPs, “Lack of political will is ultimately the issue. Even if you had fine principles, fine laws, but you don’t have the will to enforce them, then it’s as good as a dead letter.”