When the Taliban took control of Kabul, Afghanistan, in late 1996, they soon launched a sustained military offensive to the north, an area they did not control. The following May, however, Abdul Malik Pahlawan, an Uzbek leader of the so-called Northern Alliance, which had been defending the region, struck a power-sharing agreement with the Taliban — who marched right into Mazar-i-Sharif, a key northern city.
All of two days later, Malik changed his mind, recognizing that his group would not have as much power as he had hoped. Quickly joining forces with two other ethnic groups in the area, Malik and his Uzbek followers repelled the Taliban in a bloody battle, eventually regaining control of the northern provinces.
This episode contains a larger lesson: Contrary to the common perception, political alliances during civil wars are not formed along immutable religious, ethnic or linguistic lines, according to the research of MIT political scientist Fotini Christia. As she explains in a new book, “Alliance Formation in Civil Wars,” published this month by Cambridge University Press, such alliances are often created for balance-of-power reasons, and stretch across religious or ethnic boundaries. Moreover, factions can develop within homogenous groups — leading seemingly solid allies, representing the same identity groups, to oppose each other.
“We see a civil war as black-and-white, a two-sided conflict between a government and rebels,” Christia says. “But usually it is a more dynamic situation.” In these more fluid circumstances, she adds, “Two groups can be friends one day and bitter enemies the next.”
The practical upshot of Christia’s findings is that many civil wars, though often described as manifestations of ancient sectarian conflicts, are often fought between factions whose leaders are more pragmatic — possibly suggesting that these wars can be resolved if the right incentives are in place.
“We have to be a lot more nuanced in the way we understand the dynamics of civil wars,” Christia says. “The process of alliance formation and alliance breakdown can take on a life of its own as a conflict unfolds.” Or, as she writes in the book, “policy makers should not be looking to race, language or religion to predict or preclude civil wars’ allies.”
On the ground in Afghanistan and Bosnia
Christia’s book is based on two years of field research in Afghanistan and Bosnia, combined with an analysis of 53 multiparty civil wars around the globe. She interviewed dozens of political leaders and warlords in the two countries, including some former Taliban leaders, and drew upon a variety of written materials, from alliance agreements to memoirs.
One of the characteristics of these conflicts is that they often include a multitude of groups, making pragmatic alliances a necessary condition of resolution.
“Unless a group is powerful enough to win a conflict on its own, what we’re going to see is this constant conflict and realignment among groups,” Christia explains.
In Christia’s view, there are two specific mechanisms through which alliances form. The first, she writes, is the “evolution of the relative power balance between groups” — such as in the frequently shifting alliances in Afghanistan. The second is what she calls “group fractionalization,” the splitting of seemingly natural allies. During the war in Bosnia, she notes, groups of Bosnian Muslims fragmented from each other. Thus, while much of the violence against them originated with Serbs and Croats, a certain amount of violent conflict in the war pitted members of this formerly solid bloc against each other.
To be sure, Christia acknowledges, religious, ethnic and linguistic identities are important to political leaders and their followers, and are often invoked, in historical terms, during civil wars.
“My argument does not suggest that ethnicity doesn’t matter,” Christia explains. “It just tries to show why and when and how it becomes important. Identity factors may not determine alliance choices per se, but the fact that leaders need to use them to justify their choices shows that they’re clearly useful for public consumption.”
Nonetheless, Christia adds, it is significant that even in times of armed conflict, group identity along traditional lines is not the be-all and end-all determining who takes which side in a war.
“In the context of war, that’s when we presume that identities become sticky,” Christia says. But in reality, she adds, even though “a group may not change the way it views itself, it will reinvent how it relates to other groups.”
For further study: Iraq, Lebanon and Syria
The new book has gained praise from other political scientists in the field. Stathis Kalyvas, a political scientist at Yale University, has called it “a rigorous and compelling explanation for the puzzle of alliance formation and group fragmentation in multiparty civil wars.”
In the book, Christia herself calls for more empirical research on the topic, “to test the theory more comprehensively on the full universe of cases.” She suggests fieldwork and historical examination of Iraq and Lebanon, and adds that the ongoing conflict on Syria, involving several different religious and ethnic groups, should be examined along these lines as well.
Christia’s own current research involves field experiments about religion, identity formation and social action in Yemen — a subject potentially related to her book, though looking at issues of identity in somewhat more microscopic fashion. She has also recently published research on the same topic based on studies she helped conduct in Bosnia.
Ultimately, when it comes to civil wars, Christia says, there is no template for resolution; instead, understanding the flexibility of local leaders may help bring such episodes to a close. “The knowledge we have at the beginning of the conflict may not really inform what the conflict is about as it unravels,” she says. “There may be a reason those wars started, but they can be reshaped and recast as the conflict goes on.”