The murderous ‘ethnic cleansing’ of civilian populations remains one of the unexplained scandals of world history, although such events seem to have occurred almost as frequently as social revolutions. Over the past 150 years alone, mass killings of indigenous groups by colonial or settler states, of Armenians by Turkish forces and their allies, of Jews by the Nazis, of Tutsis by Hutus, have far exceeded any rational military or economic calculation. But historical and comparative sociology has had relatively little to say about these deeds. Debate about the causes of ethnic cleansing is instead dominated by ahistorical and individualistic models. Michael Mann’s impressive The Dark Side of Democracy makes a giant step toward specifying the concrete social structures and circumstances that produce such results. Its scale is vast—over 500 pages of dense theorization and historical narrative, encompassing a temporal arc that stretches from ancient Assyria to the Rwandan genocide—while its unforgettable analyses of perpetrators and their actions display an almost ethno-methodological sensibility to the micro-foundations of social life, a new dimension for this master of the grand narrative. It is a major achievement.

The Dark Side of Democracy’s mass of historical evidence is marshalled to test a strikingly bold central thesis: that ethnic cleansing is the dark side of democracy, in the sense that the latter is premised on the creation of an ethnic community that ‘trumps’ or ‘displaces’ class divisions. It is worth unpacking his usage of these terms a little more at the outset. First, democracy for Mann is primarily understood not as a set of institutions but as an ideology of equality, one that legitimates itself through a claim to represent the people and aims at a popular redistribution of social power. Second, ‘ethnic cleansing’ is defined as the attempt to create mono-ethnic populations for a given political unit; this is not necessarily murderous, and may more often involve assimilation, whether coercive or not. It is in this sense that Mann sees ethnic cleansing and democracy as having an elective affinity to one another, in two respects: first, most democracies develop on the basis of relatively mono-ethnic populations, and second, democracy carries ‘the possibility that the majority might tyrannize minorities.’ Ethnic cleansing is, then, ‘the dark side of democracy’ both in the sense of being a precondition for its emergence, and because it is generally perpetrated by democratic, or democratizing, regimes.

Mann initially presents his argument in terms of eight bold theses, following this with a chapter on ethnic cleansing prior to the nation-state that argues for the fundamental modernity of the process. A series of minutely researched case studies comprises the empirical core of the work: the New World, Armenia, Nazi Germany, Yugoslavia and Rwanda. Interleaved with these are three ‘intermezzos’—on the non-German Axis countries; on ‘Communist Cleansing’; and on ‘Counterfactual Cases’, where murderous cleansing did not occur. Broad discussion of the historical background to each case is followed by quasi-ethnographic accounts of the actual process of mass killing. Yet despite, or perhaps because of, its comparative range and density of historical detail, few reactions to the book have done justice to it. Responses have so far been of two main types: focused, respectful assessments of specific theses, and broader attacks on the notion that ethnic cleansing is the ‘dark side’ of democracy. The latter argument has generally been made from the perspective of a bland, right-thinking liberalism that instinctively reacts against any attempt to besmirch the good name of the procedurally regulated circulation of elites that contemporary social science terms ‘democracy’. A more serious, critical engagement with the book requires a fuller reconstruction of Mann’s central argument, and an examination of the theoretical cogency and empirical adequacy of his key claims—as well as a sense of how the present work should be understood within the context of Mann’s wider intellectual trajectory.

The eight theses at the core of Mann’s argument can be read as a set of increasingly specific preconditions for genocide as the most total and violent form of ethnic cleansing. The first two set out the overall parameters for his theory: one, that ‘murderous cleansing is modern, because it is the dark side of democracy’; second, ‘ethnic hostility arises where ethnicity trumps class as the main form of social stratification’. The next three refer primarily to a set of geopolitical factors: murderous cleansing occurs where two ethnic groups make a claim to the same territory; where one ethnic group feels threatened but also capable of eliminating the other; and where sovereignty breaks down ‘amid an unstable geopolitical environment that usually leads to war’. The final three theses concern the perpetrators: murderous cleansing is not generally their initial intent; there are three levels of perpetrator—party elites, militants and civilian constituencies; and, lastly, ordinary people are ‘brought by normal social structures into committing murderous ethnic cleansing’.

In sum: ethnic conflict becomes murderous when key social forces, in multi-ethnic and geopolitically unstable environments, conceive of democracy as the rule of an ‘indivisible, united, integral’ people. In contrast, where the people is conceived of as ‘diverse and stratified’, and class differences are politically institutionalized, the potential for mass ethnic killings is blocked by countervailing, non-democratic features of these societies. It is for this reason that neither pre-capitalist agrarian societies nor established liberal democracies tend to engage in ethnic cleansing. The former, according to Mann, tend to have cosmopolitan upper classes and locally oriented producing classes. In such societies ethnicity, a sense of cross-class solidarity, and ipso facto ethnic cleansing, are rare. Established democracies, meanwhile, are unlikely to commit ethnic cleansing not because they are democratic, but because ‘the politics of class, region and gender’ dominate and implicitly moderate the tendency of democracy to undermine such differences. Northwestern Europe, then, has been relatively immune, because here, democratic rights were gradually extended down the social structure; the fact that these polities did ‘not try to eliminate exploitation’ meant that the national community remained divided along class lines. Indeed, for Mann, the ‘institutionalization of class conflict has been the main political accomplishment of the modern West.’ Outside of the northwestern core states—particularly in Eastern Europe—democracy meant the rule of the whole people, and was associated with an attempt to ‘repress’ class conflict, rather than institutionalize or entrench it; in these cases, Mann contends, ethnic groups could emerge as social actors undivided by class.

The theoretical crux of Mann’s argument thus seems to be that class conflict, especially when institutionalized, tends to undermine ethnic conflict. The association between democracy and ethnic cleansing stems from the threat that the former poses to class stratification. The probability of ethnic cleansing for Mann thus follows a parabola as democratization increases: first rising, then declining. Ethnic cleansing is typical of democratizing states emerging from old regimes, where the class structure of agrarian bureaucracies has collapsed, but fully developed industrial class conflict has not yet emerged. In these conditions, an organic conception of the people can arise, unconstrained by class antagonism—and in some cases, permit ethnicity to ‘trump’ class. For Mann, then, class and ethnicity are not just independent, but to a large extent alternative, forms of social stratification.

How theoretically cogent and empirically adequate are these claims? Mann’s own evidence imposes an obvious objection, raised by many of the critical responses to the book. None of the classic cases of murderous ethnic cleansing occurred under the aegis of a democratic regime: the Armenians were massacred under the Ottoman Empire; one of the most authoritarian states in history carried out the Final Solution; in Rwanda, mass killing of Tutsis took place under an authoritarian Hutu party-state. The only real support for democratization as a basis for lethal ethnic cleansing comes, firstly, from 1990s Yugoslavia, where ethnic consolidation under elected nationalist governments often became murderous—although not, as Mann correctly points out, genocidal—and, secondly, from instances in democratic colonial or settler states. Here, Mann presents some striking and iconoclastic material to support his thesis that some of the worst genocides occurred in the most democratic environments. For instance, he records that California’s 1850 Constitution enshrined universal white male suffrage, ‘the most advanced form of democracy of the age’; and that in little over a decade, the Californian Indian population had been reduced by 80 per cent—exceeding the rate at which the Third Reich exterminated Europe’s Jews. In Mexico, by contrast, the conquistadors had faced a highly articulated society, and needed local allies in order to establish control over its resources. Although colonial rule was brutal and murderous, it did not amount to genocide: cross-elite cooperation created a ‘mestizo class/caste ruling over the indios’, within the framework of Habsburg absolutism. Spanish rule was fundamentally less exterminist than the ‘settler democracies’ of Australia or the United States. Yet there are many exceptions; to name but one, the mass deportations of Circassians and Chechens from the Caucasus during the 1860s were conducted by Tsarist armies as part of the standard arsenal of expansionism. Thus it may seem initially that the book’s central theoretical claims stand in a skewed relationship to the empirical work.