Breaking Up (a Country) Is Hard to Do

By Gary Bass
Sunday, August 27, 2006

Now that the dreaded words "civil war" have been officially dropped into the Iraq debate, the next word the White House should brace itself for is "partition." As Iraq spirals out of control, arguments for dividing the country along ethnic lines have begun to surface with increasing frequency among scholars, diplomats and others.

As early as November 2003, Leslie H. Gelb, a Carter administration assistant secretary of state, wrote an op-ed in the New York Times that suggested breaking up the British-imposed "historical defect" of a unified Iraq and replacing it with "a three-state solution: Kurds in the north, Sunnis in the center and Shiites in the south." Earlier this summer, in a confidential memo to Prime Minister Tony Blair that was leaked to the BBC, William Patey, the outgoing British ambassador to Iraq, warned that "a de facto division" of the country is more likely than a transition to a stable democracy. And in his new book, "The End of Iraq," Peter Galbraith, a former U.S. ambassador to Croatia and a longtime advocate for the Kurdish cause, argues that Americans must recognize that "Iraq has broken up in all but name."

A partitioned Iraq would join a long line of chopped-up countries. After World War I, President Woodrow Wilson invoked national self-determination to justify carving up the multiethnic Ottoman and Austro-Hungarian empires. Partition was later tried as a last-ditch solution to ethnic violence in places such as Cyprus, India, Ireland, Palestine and, to some extent, Bosnia. Kosovo, now a province of Serbia with a restive population that is 90 percent ethnic Albanian, is likely to be next.

At first glance, the idea of separation seems appealing: If they can't live together, let them live apart. But what do we really understand about such a radical step? Surveying the literature of past experiments in partition reveals that the closer one looks, the messier the outcome gets.

One core problem with partition is that it's excruciatingly hard to draw a neat line to divide groups. Some partition advocates have therefore called for moving the people first. In a 1996 article, "Possible and Impossible Solutions to Ethnic Civil Wars," published in the journal International Security, Lehigh University political scientist Chaim Kaufmann argued that because ethnic warfare hardens group identities, ethnic civil wars can only really end with total victory by one group or by physically separating all of them. Rather than send in troops to save multiethnic nations, he wrote, the international community should "facilitate and protect population movements to create true national homelands." As a last resort, it should intervene decisively by drawing a separation line; the remaining civilians of the enemy ethnic group on the wrong side of the line should be interned, to become part of a population exchange after the war. (He cheerily calls this "conquer and divide.")

Kaufmann praised the role of the League of Nations in helping to organize ethnic population exchanges between Greece and Turkey following World War I, and declared that a successful mission in Bosnia would have required that the international community "overcome its squeamishness about large-scale population transfers." And he argued that Hutu refugees not be allowed to return to Rwanda after the 1994 genocide, proposing instead an orderly resettling of more than a quarter of the populations of Rwanda and Burundi to prevent "another genocide" one day. In a follow-up article in the same publication two years later, titled "When All Else Fails," he suggested that India's 1947 partition didn't go far enough -- it should have carved up Kashmir and created a Sikh homeland, too.

In a New York Times op-ed in 1993, during the war in Bosnia, the controversial University of Chicago political scientist John Mearsheimer similarly called for moving borders and populations to create "a Bosnian state peopled almost exclusively by Muslims, a Croatian state for Croatians and a Serbian state made up mainly of Serbians." In 1995, he and MIT's Stephen Van Evera published "When Peace Means War" in the New Republic, arguing that the United States should never have signed the Dayton peace accords, because they allowed for an insufficient partition. While Bosnia was now made up of two entities -- a Serb republic and a Muslim-Croat federation -- the latter had to be quickly partitioned, too, or it would bloodily collapse, reigniting the wider war. (More than a decade later, this still hasn't happened.) They also wanted the United States to help with population transfers such as moving Serb civilians out of Sarajevo.

The critics of partition, on the other hand, see separation as a kind of ethnic cleansing with a human face. In a 2000 statistical study published in World Politics, "Partition as a Solution to Ethnic War," Yale political science professor Nicholas Sambanis found that partitions did not significantly reduce the risk of wars breaking out again. He points to the 1998-2000 war between Ethiopia and Eritrea after a 1993 partition; the violent 1992 collapse of the partition of Somaliland; and the recurring wars between India and Pakistan since partition, including the 1971 partition that sliced Bangladesh from Pakistan -- not to mention the post-partition Arab-Israeli wars of 1956, 1967, 1973 and 1982. He warns against carving up warring African countries into many little monoethnic states, which would only replace civil war with international war.

In his 2000 book, "Ethnic Groups in Conflict," Duke University professor Donald Horowitz writes that "the only thing secession and partition are unlikely to produce is ethnically homogenous or harmonious states." India today, for instance, has more than 100 million Muslims.

Still, partition theorists think that groups can never be made to live together again after ethnic war. This is surely true in some cases, but how many? After generations of warfare, it seems impossible to create a common sense of Indian and Pakistani identity, or Israeli and Palestinian identity. But ethnic cooperation is common worldwide. Not that long ago in Bosnia, Muslims, Croats and Serbs were all peacefully intermixed as both Bosnians and Yugoslavs.

Even if partition could be imposed, the creation of ethnic statelets gives an international seal of approval to the ethnic nationalists. As Stanford professor James D. Fearon points out in "Separatist Wars, Partition, and World Order," an article in the summer 2004 issue of Security Studies, if the world promises statehood as a reward for a particularly bloody ethnic war, that gives a perverse incentive to rebels such as the Kosovo Liberation Army: The worse the violence, the better the outcome. This is also an alarming signal to send to multiethnic countries such as Belgium, Canada, India, Indonesia, Nigeria -- and even the United States.

With all this to consider, it's no wonder that even partition's stoutest supporters see it as a last-ditch option. If Iraq is partitioned, it probably will be only after the United States experiences the same kind of panicky desperation that helped prompt Britain's mid-century partitions in its crumbling imperial possessions.

Gary J. Bass is an associate professor at the Woodrow Wilson School of Public and International Affairs at Princeton.

© 2006 The Washington Post Company