Were George Orwell to thumb his way through Fotini Christia's meticulously researched Alliance Formation in Civil Wars, he would nod knowingly. Christia's thesis about how participants are sold on particular alliances during multiparty civil wars centers on the near-Orwellian malleability of those narratives. One year, fighters are told they are dying for their ethnicity or religion; the next year, they are told to risk their lives for minority rights or civic nationalism, and are expected to get on with it. So it goes. “Central to the thesis of this book,” Christia writes, is the idea that “notions of shared identity are . . . employed instrumentally to justify the power-driven alliance decisions that are actually made by elites” (p. 46). The language connoting the puttylike quality of nonelites, who are swayed from one identity to another as a result of the strategic choices of elites, can be troubling, but given Christia's audience, this is not especially problematic. The field of political science is comfortable with the idea that elites find accepting audiences for self-serving narratives concerning alliance choices, from arenas as vastly different as ethnic party politics in Zambia (Dan Posner, Institutions and Ethnic Politics in Africa [Cambridge: Cambridge University Press, 2005]) to conflict among major global powers (John J. Mearsheimer, The Tragedy of Great Power Politics [New York: W. W. Norton and Company, 2001]).

Christia's main theoretical contribution, however, is in explaining when and why we can expect alliances to form and fall apart, rather than the messaging that accompanies such transitions. She argues that groups are chiefly motivated by their share of power in a postwar order. They therefore not only want to be on the winning side in any civil war, but also do not want to share victory with too many others. In other words, groups want to be part of a “minimum winning coalition,” an idea first popularized by William Riker and adopted since in vastly different substantive contexts, now including multiparty civil wars. Consequently, alliances in such wars tend to be constantly shifting: as one side gains ascendency, the groups constituting the alliance begin to distrust one another, emphasizing the “minimum” more than the “winning.” Commitment problems between weak and strong ensue, and lead to the weaker actor abandoning the alliance. An alliance growing stronger on the ground, in terms of territory under administrative control, is an alliance more likely to fall apart—this is the bottom line for Christia.

Christia's theoretical framework is transported from analyses of alliances carried out by neorealists in International Relations a generation ago (Stephen Walt, The Origins of Alliances [Ithaca, N.Y.: Cornell University Press, 1987]). As she notes, however, it is more suited to the study of civil wars than international politics (p. 52). After all, what defines anarchy more than a civil war? The book's empirical research brings to light her intuitively appealing theoretical argument, reducing the Byzantine-like complexity of two civil wars in each of Afghanistan and Bosnia-Herzegovina, as well as their alphabet soup of militant groups, to something digestible. Christia's cross-time evidence is consistent with her theory: alliances broke up when fortunes were on the rise, and groups suffered splits or takeovers when disproportionately suffering battlefield losses. The empirical chapters (3–7), which are focused on making the case for the argument at hand as opposed to a general history of those conflicts, are fine-grained and impressively multimethod: Christia employs personal interviews with those involved in the wars under focus—including warlords and former fighters—primary documents, secondary sources, and Geographic Information System technology to map changes in territorial control, rendering an encyclopedic account of how groups allied with one another, split, experienced takeovers, and gained or lost territory. Students of the two major civil wars in Afghanistan's history—the almost-decade long war between the communists and mujahidin in the 1980s, and the intra-mujahidin war in the mid-1990s—in particular will learn a great deal from this book. They will, if they share this reader's reaction to Christia's work as it relates to Afghanistan, also brace themselves for an escalation of violence in the country after the U.S. pulls out combat troops in 2014.

While valuable for students of civil war and alliances, Christia's work suffers from inattention to a number of concerns. First, on a theoretical level, the logic of her argument suggests that multiparty civil wars are extremely difficult to end because any time a coalition makes headway in the conflict, it is ripe for divorce; sustained gains are difficult. Absent an overwhelmingly powerful actor, or massive external intervention—such as that of Pakistan on the side of the Taliban in the 1990s—such multiparty civil wars would likely be caught in an endless loop of gains made and alliances broken. One recalls the joke about a student responding to a prompt on the spelling of “banana” by retorting that she knew how to start but did not know when to stop (“B, A, N, A, N, A, N, A, N. . .”).

Is Christia's argument similar? When does the loop of battlefield ascendency and alliance fissure stop? Aside from her discussion of external intervention or the exogenously given prospect of a decisive victory by one side, she is silent on the subject. Christia claims that civil war termination “is beyond the scope of this work” (p. 50), arguing that hers is a theory of civil war alliances only. But this misses the point that her theory of alliances makes the termination of civil wars a puzzle. If she is right, then multiparty civil wars should almost never end, absent external intervention.

Second, Christia is too eager to challenge identity-based explanations, both in the theoretical and empirical chapters. This is a strange preference, given that she readily and repeatedly concedes that identity has something to do with who allies with whom, even if she shows it cannot explain short-term variation in alliance behavior (p. 45). If identity can exist alongside, rather than instead of, power as a key variable in explaining different aspects of alliance behavior, then Christia need not seek to diminish the role and power of identity. Indeed, the nature of her inquiry privileges levers that are more likely to display short-term sensitivity, and here power undoubtedly wins over identity. After all, one's power can rise or fall appreciably within weeks, but one's identity is necessarily more permanent.

Third, there is a methodological issue. For Christia, the intra-mujahidin civil war in Afghanistan (1992–98) and the ethnic war in Bosnia-Herzegovina (1992–95) serve as theory-building cases, while Afghanistan a decade prior (1978–89) and Bosnia-Herzegovina during World War II (1941–45) serve as tests of the theory. This sits uncomfortably for the reader, who is doubtless aware that one cannot test a theory with the same case that gave rise to it. Christia's explanation is that the earlier wars in each country were primarily nonethnic while the latter two were ethnic (p. 27). The implication is that each of the wars is an independent event. I am not so sure, especially in the Bosnian context, where Christia herself writes that the 1992–95 civil war “directly referenced symbols and events from a past war that had traumatized the psyche of the different constituent nations residing in Bosnia” (p. 197), showing that people in the area may themselves not accept that the two were purely independent.

These issues notwithstanding, Christia's work is recommended reading for students of groups in alliance behavior, conflict processes generally, and Afghanistan's two civil wars. The book's arguments are sharpened by her impressive marshaling of various data sources and methods, and represent an important step forward in the study of civil wars and insurgencies.