Review Forum: Michael Mann, The Dark Side of Democracy

Michael Mann.
The Dark Side of Democracy: Explaining Ethnic Cleansing.
Cambridge University Press, 2005, 548 pp.
$US 24.99 paper (0-5221-53854-8), $US 70.00 hardcover (0-521-83130-X)

When I accepted the challenge of reviewing Michael Mann's colossal book, The Dark Side of Democracy: Explaining Ethnic Cleansing, I was aware that I was giving up the home field advantage. I am not a sociologist but rather a historian. Like most historians, I was trained to specialize, not to generalize. My area of expertise is eighteenth-century Europe. Familiarity with what Robert R. Palmer called “The Age of the Democratic Revolution” [1] is certainly useful in this context, considering that democracy figures in the very title of Mann's book. But I don't read much contemporary sociology, and I have never studied ethnic cleansing systematically. So, I presumed that Mann's book would overawe me with theoretical vigor. And I anticipated that when it came time to make critical suggestions, I would probably just propose a longer timeframe and a few factual corrections.

My overall reaction turned out to be just the opposite. The book is impressive in its historical dimensions. Mann has done a great deal of original research. He even provides a major empirical contribution in an area of history — Nazi Germany — that looked like it was already saturated with data. The temporal perspective of the book is long. Finally, Mann has a talent for narration — for sketching biographies and thickly describing episodes of murderous conduct.

Paradoxically, the book was disappointing only on the theoretical level.

This review will focus almost exclusively on the book's theoretical structure. I am taking this approach for three reasons. First, the book is simply too extensive in scope; its empirical content cannot be summarized and evaluated in a way that would be brief and fair at the same time. Secondly, in a sociological journal, primacy ought to be given to the book's overarching claims, its general hypotheses — the sociological categories that underlie its history. Finally, after reading the book, I found that as much as I admired its breadth of information, I remained confused by its basic intellectual goals. What does “the dark side of democracy” mean? What constitutes Mann's “explaining” of ethnic violence? My personal judgment of the book crystallized within the framework of these intellectual perplexities.

In this critical review, it will become clear that my standards for what constitutes coherent sociological explanation are very high. My intuition, in fact, is that I have approached Mann's book with a standard more exacting than that operating within the field of sociology today. If that is the case, then this review can perhaps give rise to discussion not just about Mann's book but also about the disciplinary standards that have made this book possible. The issue, to put it bluntly, may not be that I need to lower my standards — it may be that you, dear reader, need to be stimulated to raise yours.

I do want to emphasize, before launching the theoretical discussion, that the book is extremely informative. There are very substantial chapters on the cleansings that occurred in North America, Armenia, the Nazi regime and its allied states, three communist regimes (Stalin, Mao, and Pol Pot), Yugoslavia, and Rwanda. Students and scholars unfamiliar with these episodes of ethnic cleansing will learn a great deal about what happened. I especially recommend the chapters on Armenia, the Nazis, Yugoslavia, and Rwanda because they vividly narrate the series of events by which mundane political aspirations escalated into extreme violence. The following analysis, however, deals mainly with the argumentative structure of the book as a whole. “This book presents a new theory of ethnic cleansing,” says the descriptive blurb on the book's back cover. What is this theory?

Thesis 1: The Dark Side of Democracy

Mann outlines the theoretical apparatus in an opening chapter called “The Argument.” The chapter is 33 pages and enumerates a long series of heterogeneous theses. Each thesis appears, disappears, and reappears throughout the book in unpredictable ways. Entire chapters seem to contradict particular theses. But the theses are so diverse that, at every moment of the book's unfolding, Mann can reach into his collection of claims and suggest that at least one of them is being confirmed. One could call such an approach “multi-causal” or “kaleidoscopic.” But I found myself thinking instead of “sleight of hand.”

More precisely, Mann enumerates eight primary theses, but some of them are broken down into various sub-theses indicated by letters. We thus have Thesis 1, Thesis 1a, 1b, and in the end, about fifteen distinct theses. With this nomenclature, there is no mistaking how each thesis is designated. But how do the theses inter-relate? What is the architecture that unifies them, either by making them compatible with each other or by establishing a hierarchy of truthfulness among them? The plethora of theses invites these questions. By focusing selectively on some of the main theses, I wish to highlight the chaotic relationship among the arguments.

Mann's Thesis 1 is that murderous ethnic cleansing is the dark side of democracy. The ideal of rule by the people tends to convert demos into ethnos, generating organic nationalism and encouraging the cleansing of minorities. Mann refines Thesis 1 with Theses 1b and 1c by stating that democracies are most likely to produce ethnic violence in the early phases of their development — at the time of founding, or if the democratic project is carried by a group of colonists, at the time when this democratizing group takes over and settles a foreign territory. So far, so good. The reader gets the impression that the book will look closely at the institutionalizing of democracies and how these formative moments yield ethnic violence.

Thesis 1d, however, states, “Stably institutionalized democracies are less likely than either democratizing or authoritarian regimes to commit murderous cleansing” (italics added). At this point, I detect two difficulties. First, declaring that “stably institutionalized” democracies do not produce “murderous cleansing” sounds like a tautology. It also sounds redundant in relation to the previous theses about the founding of democracies. For if ethnic violence is concentrated in the founding period, then it goes without saying that it will diminish after the founding. I would add that tautologies about the relationship between democracy and ethnic cleansing are not infrequent in this book. For example, “democracy can encourage ethnic tensions and violence if political parties organize by ethnicity” (440). Of course, once we have ethnic parties, we are on the road to ethnic conflict. But the more important question is whether democratic ideals, by their very nature, inevitably produce political parties organized by ethnicity.

Secondly, Thesis 1d admits that ethnic cleansing occurs in authoritarian regimes, not just democratic ones. Does this mean that ethnic cleansing is the dark side of all democratic regimes — or of some democratic and some undemocratic regimes? To frame this duality differently: Is the book about how ethnic violence stems from democratic political processes? Or is the book about ethnic violence in a multitude of contexts not reducible to any particular political type? The answer to these questions will very much affect the ontological status of ethnic cleansing. On the one hand, the very title of the book, The Dark Side of Democracy, and Thesis 1 encourage us to see ethnic cleansing as derivative of modern democracy. On the other, once Mann concedes that non-democratic regimes have ethnic cleansing too, it sounds like ethnic cleansing has its own being, it own undetermined history — so we end up following its itinerary into whatever regime-type it leads us. The book thus ends up being too long for its main thesis. It would be more elegant scientifically, the thesis would be more concentrated and compelling, if Mann had focused only on democratic regimes, or offered a concept of democracy that was large enough to include apparently undemocratic ones.

Large stretches of the book are not about democratic regimes at all. Even the cover of the book belies the title printed upon it. The cover shows the burning of the Reichstag in 1933. The reader at first wonders if Mann intends to characterize the Nazis as democratic (a contention I would not reject out of hand, given some of the egalitarian elements of Nazi ideology — all members of the Volk were considered honorable regardless of class), or if he sees the Nazis as an undemocratic exception to thesis 1. He ultimately treats it as an exception; the Nazis were not democrats. Thesis 1 thus starts out looking like a global sociological explanation of ethnic cleansing, but it becomes a partially confirmed, largely refuted, speculation about democracy.

I would add that the four chapters of the book that deal with the Nazis and their auxiliaries (Italy, Bulgaria, Romania, etc.) occupy well over 100 pages of the book. Then we have a chapter on communist regimes adding about another 40 pages. So a good chunk of the book demonstrates that ethnic cleansing is not merely the dark side of democracy. Moreover, it's also questionable whether some of the other regimes described as democratic really were. For example, Mann suggests that the Young Turks who came to commit genocide against the Armenians began with aspirations to establish representative government. There's some truth in this, but establishing democratic institutions was not the Young Turks' main goal. Influenced heavily by French positivism, the Young Turks identified less with 1789 and 1848 and more with the science-oriented reaction against popular government. In positivistic philosophy, the form of government matters much less than the scientific grounds on which policy is based. One specialist on the ideology of the Young Turks goes to far as to say that the Young Turks were not democratic at all. Instead, central in their ideology were pseudo-biological theories of race. On this ideological base, which combined Comte's cult of science, especially biology, with an intensely racial view of the world, the Young Turks began to envisage a homogeneous Turkish state.[2] Thus, I would suggest that far from being the dark side of democracy, ethnic cleansing was often grounded in a counter-democratic impulse: the ambition to transcend the politics of popular will by means of a politics of scientific reason.

Finally, there are certain regimes that both Mann and I would consider democratic. But it's debatable whether they perpetrated ethnic cleansing. The key cases here are France and the United States. France comes to mind as the quintessential democratic nation. The Revolution of 1789 unleashed the Terror and the massacres of the Vendée, but it's not clear that this violence occurred on the terrain of ethnic ideology. Though I suggested at the beginning that I would not pick apart the book on the basis of factual omissions or errors, I do think it's worth underscoring that the book does not treat France at all. Can any analysis of modern democracy be compelling without some treatment of the country that gave us the Declaration of the Rights of Man, that abolished slavery in 1794, and that inspired much of the world to move along the path of democratic universalism?

For the United States, another country whose democratic destiny we can agree upon, Mann does provide a thorough chapter. Here he compares Lincoln to the Nazis, and argues that the dwindling of the native-American population in the 1800s was a form of murderous ethnic cleansing. He is not the first scholar to make these claims, but I found this chapter, above all others, to be factually biased. There is a considerable amount of misinformation about the Founding Fathers and the American Constitution. For example, Mann at one point says that the framers of the Constitution met in Philadelphia for only “two weeks” (56). In reality, they debated intensively for four months, from May to September, 1787. He makes inaccurate remarks about the Framers' age and social background that are designed to portray the Framers as elitists. His analysis of the place of “property” in the Founders' thinking also struck me as crudely belittling.[3] His overall message is that the founding of the United States was a cursory affair in which no consideration at all was given to anyone but middle-aged, white, property owners.

Mann proceeds to implicate these solipsistic whites in the native-American disaster by saying that they deliberately refrained from inoculating Indians against smallpox. “Slaves were inoculated against smallpox. Indians were not” (89). But according to Steven T. Katz, after the discovery of the smallpox vaccination in 1797, all Indians were encouraged to be vaccinated, and following the epidemic of 1831-1832, U.S., government policy required the vaccination of Indians. The resistance of the western tribes to this statute contributed directly to their decimation by epidemics in the late 1830s and early 1840s.[4] I do no not mean to suggest that Mann's views are entirely without merit. But I would never put this chapter in the hands of students without presenting them with the counter-arguments of scholars, such as Katz, who consider it fruitless to liken the American case to European genocides.

Thesis 3: The Bi-Ethnic Thesis

The ambiguities of Thesis 1 are compounded, in my opinion, by at least three other theses. According to Thesis 3, murderous ethnic cleansing is imminent when two strong ethnic groups (as opposed to multiple ethnic groups) compose the society. For example, Mann observes that in most of Africa, “a major restraining factor has been the sheer multiethnicity of African states. Hence, there has been little ethnic cleansing in most of Africa” (428). I will refer to Thesis 3 as the “bi-ethnic” thesis. It's plausible, and also has a good pedigree. Voltaire wrote: “If you have two religions in your midst they will cut each other's throats; if you have thirty, they will live in peace.” [5]

Yet, one difficulty with the bi-ethnic thesis is that too many of the regimes that perpetrated ethnic cleansing were multiethnic. The Ottoman empire had Greeks, Kurds, Arabs, Mongols, Nestorian Christians, Jews, nomads, and so forth. Yet, for the Young Turks, the Armenians became enemy number one. Similarly, in the areas where Jews were victimized (Germany, Poland, the Ukraine, etc.), there was no pre-given conflict between Jews and some other group for control of the state. Mann tries to finesse this issue by observing that since Jews were associated with the communist party, the Nazis could “plausibly” regard the Jews as part of the communist threat. About the cooperation of Lithuania in the Holocaust, Mann writes: “Jews were 7 percent of the general Lithuanian population. They were 15-16 percent of Communist Party and Komsomol members ... So a Judeo-Bolshevik threat seemed minimally plausible” (282). About the Ukraine, Mann writes: “During the 1930s they [Jews] were 5 percent of the population but constituted over 10 percent of the Communist Party ... the Judeo-Bolshevik enemy thesis might again attain minimal plausibility” (288). For the Armenian genocide, Mann finesses the issue in a similar manner: “Though the Turks possessed overwhelming force within the empire vis-à-vis the Armenians, the two sides could be equalized by foreign support that produced a real threat of political extinction among Turks” (138).

The bi-ethnic thesis is profoundly undermined when Mann inflects it to mean that at certain times, one group can perceive another group as a threat to its control of the state. Saying that a bi-ethnic state tends to produce militant ethnic ideology and aggression is one thing. Saying that a variety of factors can lead people in a multi-ethnic state to adopt a bi-ethnic mentality, a Manichean paranoia, is another. Moreover, the treatment of the Jews in passages like those quoted above is theoretically undeveloped. Why does 10 percent involvement in the communist party make the Judeo-Bolshevik thesis “plausible”? Why shouldn't we consider 20 percent the threshold? Or 50 percent? And what does it mean to say that these persons were Jews at all? Many were atheists and would not have identified themselves as Jewish. Some were surely of mixed religious descent. It would have been equally plausible for Nazis to assume that anyone of Jewish or partly Jewish descent who joined a communist party was no longer Jewish. In short, Mann's concept of Judaism as an ethnicity is too similar to the ethnic ideology he ought to be analyzing. (Not to imply that he harbors any prejudices against Jews or any other group. His humanitarian standpoint is evident throughout the book.)

It appears to me that Mann is struggling to take a purely ideological fact (fear of a Jewish conspiracy) and harness it for his bi-ethnic thesis. But the bi-ethnic thesis is originally not about ideology; it is about the real (not imaginary) ethnic composition of a state. The bi-ethnic thesis may well work for Africa, where a multiplicity of ethnic groups helps to impede ethnic cleansing. But it doesn't work for Europe, where a multiplicity of ethnic groups has not impeded ideological polarization on the basis of ethnic categories of thought.

Even if the bi-ethnic thesis were confirmable, we would still wonder how it relates to thesis 1, the dark side of democracy. Thesis 1 is oriented toward the political content of the regime in which ethnic violence takes place. The bi-ethnic thesis, in contrast, is about the sheer mechanics of political confrontation. This thesis draws attention away from the particular logic of democratic culture and emphasizes instead the more generic logic of bipolar political competition. It could apply to any kind of state. The bi-ethnic thesis can't be a corollary of Thesis 1 because it is too different in kind. The bi-ethnic thesis, instead, looks like a whole thesis in its own right. Again, we are forced to confront an ontological ambiguity. We are left wondering which of the two theses is meant to get at a fundamental reality? What is contingent, what is necessary, in the history of ethnic cleansing? How do these seemingly incongruent theses fit together to form a general explanation of ethnic cleansing?

Thesis 5: Warfare

Thesis 5 states: “Going over the brink into the perpetration of murderous cleansing occurs where the state exercising sovereignty over the contested territory has been factionalized and radicalized amid an unstable geopolitical environment that usually leads to war.” Here one could focus again on tautological elements (“Going over the brink ... radicalized ... unstable geopolitical environment”). But the discussion that follows the passage makes it clear that war is the key term here. Ethnic cleansing occurs in the midst of war. Inter-state violence tends to unleash violence within the state. The thesis is sensible in itself; but again, it's hard to relate the point cumulatively to the previous theses. Is modern warfare by itself enough to generate ethnic violence? Or must war combine with other causes? Does war accelerate the trend toward democracy, or vice versa? Do democracies tend to create a kind of international rivalry, which in turn intensifies internal ethnic rivalry? If so, how does this claim relate to the previous claim that democracies tend to breed ethnic cleansing at the time of their founding? How many of these factors are critical in themselves? How many merely correlate to other causes? We have lots of interesting theses, systematized by means of numbers and letters; but what is the logic — what is the hierarchy of explanation?

Thesis 7: The Perpetrators

Thesis 7 focuses on the individual perpetrators of murderous ethnic cleansing. Mann has done a great deal of research into the social backgrounds and personal motives of people who kill others on the basis of their ethnicity. What distinguished the perpetrator of ethnic violence from other members of the populations? Some excellent parts of the book stem from Mann's systematic and thoughtful examination of this issue. His analysis of the Nazis is a veritable tour de force. I predict that Mann's treatment of the Nazis will quickly become one of the most widely cited interpretations of the Nazi perpetrators.

With a sample of 1581 Nazi killers, Mann observes that a disproportionate number of them were Catholics. He also observes that high number came from borderlands areas of the German empire. This second finding is particularly exciting. Individuals from lost or borderlands areas tended to become militant about ethnic cleansing. Many were forced from their homeland after World War I. Having been housed in refugee camps, having lost their homes and their way of life, these dispossessed Germans later formed a core of militants who aggressively advocated the revision of Germany's borders and the purging of non-Germans from the population. “The origins of mass murder,” Mann, writes, “lay substantially in embittered ethnic imperial revisionism” (227).

Mann deserves credit for coding his sample of perpetrators for lost territories or threatened border areas. No one had done this before with precision. The argument is especially important because it encourages us to see how the perpetrators considered themselves to be victims. Instead of thinking only of “Germans” and “Jews” we must think more subtly of “Germans who feared being French” (or “Germans who did not wish to be Poles”) and the escalation of their resentments into radical programs. Mann also applies the borderlands thesis productively to the genocides in Yugoslavia and Rwanda. He does not apply the thesis to Armenian or communist ethnic cleansings. But this inconsistency is not the real difficulty. The real problem, as I see it, is that the important borderlands insight never registers an impact on the global lessons of the book that concern the perpetrators of violence. For Mann's overarching view of the perpetrators is that they are not distinguished from the rest of humanity in any significant ways.

In spite of the specificity of his empirical findings, Mann is heavily invested in the “ordinary men” argument — an argument that readers familiar with Christopher Browning's work will recognize. In contrast to Daniel Goldhagen, who argued that nearly all Nazi perpetrators had intense hatred for Jews, Browning argued that preexisting anti-Semitism played only a minor part. [6] Browning drew on the experiments of the Yale psychologist, Stanley Milgram. The basic idea is that apart from a few militants, killers are not ideologically motivated but are drawn into killing by social pressures, such as fear of ridicule or of punishment, that are sufficient to induce human beings in general to join in the slaughter.

This approach implies that the perpetrators of Nazi violence did not have a distinctive sociological or ideological profile. Though Mann's research on the Nazis seems to lean in a different direction, his Thesis 7, overall, is very much in the spirit of Browning, whom he cites with approval contra Goldhagen. In the book's conclusion, Mann writes:

The entire group of perpetrators is driven by the many motives that are normally found among ordinary people participating in more mundane social movements. Ordinary people are brought by normal social processes behind the ethnic barricades and then into committing murderous ethnic cleansing ... Radicals at all levels are helped to kill by their sense of righteousness. But even you or I could do it, for reasons of career, comradeship, patriotism, work routines, and other mundane human motives. We are humans, capable of evil. (504)

Thus, Mann ultimately opts for an unhistorical explanation. It is simply human nature, it is “normal,” to go along with violence. In this context, it seems to me, his acute observations on the borderlands phenomenon cease to have general sociological significance. Moreover, the “ordinary man” thesis is so generic, so universal, that it provides no support for the more sociological theses in the book about democracy.

Some Moral Considerations

My intuition is that Mann gravitates toward the “ordinary men” thesis because its universalism is morally appealing for him. It permits him to avoid being judgmental about specific national groups. Some scholars have criticized Goldhagen (incorrectly, in my opinion) as harboring a hatred of Germans that is akin to the German hatred of Jews. Mann clearly seeks to avoid blaming specific groups of people for ethnic cleansing. He seeks a more self-critical sociology. The “ordinary men” thesis allows him to suggest to his readers, no matter where they are from, that they have no right imagine the perpetrators as other. None of us is immune to hatred and violence, when the sociological conditions for ethnic cleansing are ripe.

But there's a problem here. I wouldn't do it. I refuse to confirm Mann's intuition that I myself would commit murderous ethnic cleansing, even if I were placed in the same circumstances as those who have perpetrated violence. Assertions to the effect that “either you or I would do it” occur several times in Mann's book, and I found them alienating, not edifying. I believe my moral and religious education has conditioned me to refrain from murdering innocent people on account of their ethnicity. I realize that my assertion cannot be confirmed — but neither can Mann's assertion that I would become a killer. Moreover, by making a public affirmation, I am surely increasing the chances that if the moment of truth did occur, I would do the right thing. Our public pronouncements play a role in conditioning our conduct. They force us to live up to the image we have fashioned for others. None of this is conclusive, of course. But my main point is that the “ordinary men” approach just doesn't move me as an ethical exhortation. In fact, it strikes me as morally counter-productive. It is more inspiring to hear people declare firmly that there are certain lines they will not cross, no matter what the circumstances. The moral fabric of the human race depends on such principles.

In his last chapter, “Combating Ethnic Cleansing in the World Today,” Mann takes a different view. He suggests that the key to preventing ethnic cleansing is avoiding the social conditions that explain ethnic cleansing. But as my critique of his explanatory theses suggests, I do not emerge with a clear vision of these social preconditions. Providing no appeal to the faculty of human courage, offering instead only a convoluted series of highly academic theses, Mann's book cannot compete effectively against ideologies of ethnic hatred. For these ideologies, in spite of their gross distortions, do at least recognize the human need for meaning — for examples of uncompromising commitment and paradigms of explanatory coherence. This leads me to some reflections on the purpose of sociology.

Is it the purpose of sociology today to avert discussions of moral responsibility? Does it aim to avoid responsibility even for its own intellectual claims by multiplying them to the point where nothing of a general nature can be refuted because everything of a general nature has been asserted in one way or another? Does contemporary sociology contribute to the dehumanization of man?

Daniel Gordon
University of Massachusetts

Daniel Gordon (Ph.D. in History, University of Chicago; M.S.L. Yale Law School) is professor of history at the University of Massachusetts and coeditor of the journal Historical Reflections.

The Darkside of Democracy: Explaining Ethnic Cleansing.
A Reply to Daniel Gordon

Daniel Gordon claims that he has “very high standards of explanation,” “more exacting” than those of sociologists. I doubt this, since in the case of my book he has not even understood the nature of the explanation it offers. I would be very sorry if your readers would deduce from the incoherence of his review that my book is also incoherent. Gordon flits around among some of my arguments and chapters, but in a rather desultory way. The book does have an overall explanation, but he fails to detect it. So I will first undertake what should be the duty of the reviewer: present the book’s argument. Then I will turn to factual inaccuracies and disagreements.

My explanation of ethnic cleansing is not of the single-factor or covering-law type which he seems to expect. As I say on page 9, it concerns not laws at all, but empirical tendencies concerning most cases of murderous ethnic cleansing. It is a multi-factor explanation, yet the factors are inter-related, not “heterogeneous” and “disconnected,” as Gordon says. Its eight main theses enter cumulatively into my explanation. Each, if present, escalates ethnic tensions further, into conflict, then violence, then mass murder. Escalation is contingent on the added pressures contained in each successive thesis. There will not tend to be further escalation of ethnic conflict if the additional pressure is not present. That is the first “hierarchy of explanation” he calls for.

This is all stated in the first pages of the book. What could be clearer? But Gordon seems to expect that each of my theses will in itself and independently explain a given proportion of ethnic violence. He lists many questions about the general relations between ethnic cleansing on the one hand, and three of my theses (concerning democracy, bi-ethnicity and war) on the other. Yet his questions are beside the point. On democracy, he says he would prefer I ask “whether democratic ideals, by their very nature, inevitably produce political parties organized by ethnicity.” That would be a nonsense. Of course they do not, unless a number of other conditions are present. His questions about the supposed general “logic” of bi-ethnic situations and of war are similarly problematic. The book deals not with static relations between variables, but with a process of escalation. Remember also that these are all empirical generalizations, which stand or fall not by any general “logic” but by whether they are empirically true – which is what 90% of the book tries to discover through its series of case-studies.

My theses also form a second type of hierarchy. They proceed from the general to the particular – from thesis 1 (a) which specifies the historical era in which murderous cleansing occurs, to the individuals who perpetrate it, discussed in theses 7 and 8. My model is one of process, sequence, narrative, interaction, not of one of universal relations between static variables. I would have assumed this would be self-evident to someone who describes himself as a historian.

The model begins with my most general proposition about democracy, thesis 1a. Murderous ethnic cleansing is the product of the modern era of democracy, the aspiration to “rule by the people.” Here two meanings of “the people,” demos and ethnos, may get confused, and thus danger looms in the process of the search for the demos. So it occurs more amid processes of democratization than in institutionalized democratic or authoritarian states. My analysis here is rather concrete. With the exception of the Nazis, most of the leading perpetrators in my case studies (including most of the foreign collaborators of the Nazis, who occupy many of the pages he says the Nazis occupy) started off with democratic goals. But under the pressures described in my other theses, they perverted those initial ideals, and confused the demos with the ethnos, first in an exclusionary “organic nationalism” and then in violence to “cleanse” the ethnos. Thus the general process is revealed most precisely in individual perpetrators’ own political careers.

Naturally, since murderous ethnic cleansing is fairly uncommon even in the modern period, we need more explanatory factors than merely democratic aspirations. My second thesis adds that sentiments of ethnicity must predominate over those of class. This is not a mere tautology, since in these cases class consciousness is rechanneled in consciousness of ethnic exploitation by one side and of a claim to civilizational hierarchy by the other (Hutus claimed they were exploited, Tutsis that they stood for civilization and progress). In fact, I say that murderous conflict occurs between hierarchically-arranged ethnic groups. This re-channeling intensifies ethnic hostility. In a counter-factual case, India since 1950, severe ethnic/ religious conflict intermittently reaches as far as murderous riots but then tends to fade. Here I argue that the strength of caste differentiation blended with the Congress and leftist parties’ emphasis on class, have helped restrain ethnic escalation. This thesis led me in another chapter to explore the distinctive features of class-based murders (“classicide”) in the communist countries.

My third proposition concerns the politics of certain bi-ethnic situations. Here Gordon objects that one of my cases, the late Ottoman Empire was multi-ethnic. Indeed. So was Yugoslavia, South-West Africa, and other cases. But murderous cleaning came as one main bi-ethnic conflict came to dominate and polarize (of course, Yugoslavia contained a series of bi-ethnic confrontations, and in Bosnia there were also brief understandings between two out of three of them). Polarization in the Ottoman case happened when the eventual perpetrators switched from considering themselves “Ottomans” (a non-ethnic identity) to “Turks,” and when the eventual victims came to think of themselves as an Armenian nation. Greeks, Arabs, Kurds etc were more marginal to the process of escalation to genocide, though all became caught up in it in the ways I describe. But I also specify three further pre-conditions for the escalation of bi-ethnic conflict, which Gordon does not mention. First, the rival ethnicities must be “old” but not “ancient” – my stance is a qualified “modernist” view of ethnicity. Second, there must be a political dispute: each must claim its own state over some or all of the same territory – which only some Armenian nationalists were doing at this time among the minorities of the Ottoman Empire. Third, both sides must have a plausible, achievable vision of statehood – in particular the domestically weaker side will tend to believe it will get help from outside, either from co-ethnics abroad or from some great Power with presumed influence on local events. Help would supposedly come from Armenians within Russia, backed up by the Tsarist state itself. Otherwise the weaker side will not resist, but submit to discrimination and/or low-level ethnic violence – which I say remains the commonest outcome of all the prior pressures I have just mentioned. Very little of this refers to any “generic logic of bipolar political competition.” And it is also a product, not of Gordon’s abstract “democratic culture” (whatever that is), but of the ethnos replacing the demos.

However, even so, we find that moderates/ realists in the rival communities usually retain enough power to draw back from murderous confrontation. That is why I add the fifth thesis of geopolitical instability, normally involving a difficult war, to which this ethnic rivalry has some relationship. Yet again, it is not war in general, but wars generating a political crisis to which ethnic escalation might be considered a solution. As factionalism intensifies in the crisis, “radicals” emerge suggesting that added strength will come from ethnic “purification,” and “retaliation” against the other community.

Even so, these radicals do not initially conceive of mass murder. In a sixth thesis ignored by Gordon, their initial “Plan A,” perhaps to achieve cleansing by, for example, forced emigration, is unsuccessful. They escalate to a more violent “Plan B,” and maybe (if again frustrated), to a “Plan C,” genocide. I was surprised to discover such sequences, and this thesis will be especially controversial. Those who speak in the name of the victims prefer their perpetrator to have been genocidal and “evil” from the beginning. Tragedy resulting from purposive action seems more meaningful than that ensuing more contingently. However, I do distinguish between perpetrators like Hitler and his circle for whom escalation seems almost “logical” given their initial views, and those who could not have conceived beforehand that this is what they might end up doing.

My last two propositions deal with the actual perpetrators. Gordon does identify them both, yet says that my “overarching view” of them is that they are “ordinary people” – ie that my eighth thesis dominates my seventh thesis. I did not men to convey this impression, since the seventh is a necessary precondition for the eighth. Number seven states that there are three levels of core perpetrator, the elite of the “party-state” (in the end, all the perpetrating states have this form) exercising power top-down, “popular” paramilitaries exercising military power “sideways,” and mass (though never majority) pressures from below, which center on “core constituencies” of radicalism – some of which Gordon does praise. Only when these three levels jointly generate radical ethnic movements, does my eighth thesis become relevant: their combined pressures can indeed persuade “ordinary people” to join in murderous cleansing.

I admit that adding to “ordinary people” the words “like you and me” is an attempt to provoke the reader, but it is a provocation to the kind of moral reflection Gordon himself calls for. I discuss two levels of “ordinariness,” those located within core constituencies (who are likely to be more “ideological killers”) and those outside them (whose motivations are more likely to be more mundane). I talk about “me” in relation to a particular core constituency (pp 29-30). I note that most German professors in the early 1930s were probably Nazi sympathizers (most students certainly were), and I instance one in particular. Otto Ohlendorf was, like me, a Professor of Sociology (though also of Law) who had studied fascism (as I have). He later led one of the terrible SS Einsatzgruppen. Gordon says proudly, almost boastfully, that he would never be tempted into evil. Has he ever been tested by such circumstances? Thank God, I have not.

As for “you,” I say that if you are a present or past soldier or policeman, or a refugee, or living near threatened borders, or a young male who likes violent sports or an occasional bar brawl, or if you conceive of your religion as being “the soul” of your nation, then you would be more at risk of being a perpetrator. But, of course, only some people with such identities did become perpetrators. I suggest (mainly from the very rich data that I collected on Nazi perpetrators – which is why I spend most time on them) that these more ideological perpetrators did so not usually in one single “choice,” but through a “career “ in which they accepted one-by-one escalating doses of radicalism-cum-violence. But the more “ordinary” the perpetrator, the more suddenly and unexpectedly they were confronted with a single devastating moral choice – to kill or not. A few bravely refused, most gulped and did it, mainly because of mundane social pressures or careerism, a few did it with enthusiasm. But almost all the perpetrators then constructed a “rational” and even a “moral” tale” of why they did it – normally that this was “self-defence” or merely “retaliation.” This is the most disturbing empirical reality that I describe in this book.

None of this evades issues of moral responsibility – a charge which I find a little offensive. My analysis only makes it a more complex issue. People do not generally think in advance about whether they would commit murder (it is so far away from their everyday life experience). They confront it suddenly, yet in recognizable social situations in which authority, peer pressure and careerism all entwine with morality. In the end I do advance a theory of evil. Evil is not located primarily in completely alien “others,” whether these are individual sadists or whole ethnic groups like supposedly “genocidal Germans” or “primitive Hutus” (though there were sadists, and many Germans and Hutus did become consumed by hatreds). Most of my perpetrators became evil through choices made through and amid recognizable social processes and pressures occurring within our own civilization . In fact, they included not just Germans or Serbs, but also our British and North American ancestors.

Now for empirical matters. I say that my theses are only empirical tendencies, lacking universal applicability. None of them apply unequivocally to all my cases – since this is history, which is messy, not a neat laboratory for hypothesis-testing. Murderous ethnic cleansing can skip one or even two of them.

I also say repeatedly that my model does worst in the Nazi case, and so we should not take the Holocaust as being the typical case – though this is tempting, since it is easily the best-documented one. It is extraordinary, the only case with a seven-hundred year history of intermittent violence against the victim, where the victims did virtually nothing to provoke mass killing, and where Jews constituted only 0.7% (and declining) of the Weimar German population. It is bizarre that Gordon would cite Katz against me on the uniqueness of the Holocaust, since Katz and I mostly agree. I am uncomfortable mainly with Katz’s tendency to confuse “different” with “lesser” suffering in other cases, for he sometimes minimize their horrors. I do defend the comparative analysis of genocide – and of the other types of murderous cleansing which I distinguish in an elaborate typology (on page 12) of which Gordon makes no mention. I also produce more evidence than Gordon mentions for the Jewish link to Bolshevism, which I only claim gave “minimal plausibility” (among Germans and especially their Eastern allies) to the notion of a Judeo-Bolshevik enemy. I myself say that Nazi ideology has to do a lot of work in this case. It is gratifying that he finds my data on the over-representation of Catholics and those living in border areas “exciting.” But this is a key part of my overall explanation of why the Nazis did feel so threatened and why they committed genocide. He even quotes my sentence “The origins of mass murder lay substantially in embittered ethnic imperial revisionism” – again this is a key part of my explanation as to why the Nazis sought to annihilate Jews.

I disagree with what Gordon says about the Young Turks, and so do some other scholars, though these issues remain controversial. Though Hanioglu does emphasize positivism, racial theory and elitism in Young Turk ideology, he makes them too homogeneous by removing thinkers who do not fit – like Gökalp, the most influential of all, a Durkheimian (Hanioglu says none of them was influenced by Durkheim!). I think that Gordon rather exaggerates the role of great intellectuals (he is a historian of ideas) at the expense of practical politics. In 1908-09 the Young Turks were still political allies of the Armenian nationalists. They were trying to restore a constitutional order, now based on elections. The rupture between the two parties, and of the Young Turks from democratic forms, happened over the next few years. It concerned politics, both domestic (especially education) and geopolitical, not positivism or racialism. The Young Turks remained divided into factions espousing more multi-cultural Ottoman and narrower Turkish identity politics until just before World War I, with Ottoman liberalism also still hanging on in parts of the movement. Most scholars put the decisive shift away from Ottomanism within the party core as occurring within the congresses of 1912-1913. In 1913 also began the large-scale resettlement program, from which the institutions of perpetrating large-scale violence on Armenians (and, on a lesser scale, on Greeks) later emerged. I term this process as the emergence of “organic nationalism” in the period 1908-1915 (I trace similar processes in almost all my case-studies).

Furthermore, many Armenians did hope that Russia would help them achieve their own nation-state, and some Armenian nationalists did assist the invading Russian army. In early 1915 the Young Turk regime did indeed fear “political extinction.” The Russians had invaded from the east, the British and French from the south, and the British had also landed at Gallipoli, close to the capital. The Young Turks expected they would be shortly retreating into what they had recently begun to think of as their “heartland,” Anatolia, where unfortunately most of the Armenians also lived. The radicals among them argued Anatolia should be cleansed of people of dubious loyalties in order to defend it. This rapidly escalated to genocide. [7]

It is understandable that a historian of the ideas of the French Revolution would chide me for not including France. Again he seems to assume my book is seeking to explain democracy rather than ethnic cleansing. But since France was not seriously riven by rival ethnic groups, there could be little confusion there of demos and ethnos. As for the Americans, he is correct that I mis-state the duration of the Constitutional Convention, but the delegates to the Convention were all white males, and virtually all were at least middle-aged and of substantial property – only two (one was Hamilton) might be considered self-made men, though there were also half-a-dozen poorish relations of prominent families, whose influence helped them ascend. Most of the founders had originally wanted to restrict the franchise more, but the war had forced them to be more inclusive. That is what I say and it is correct.

But since this is a book about ethnic cleansing, not democracy, my main argument about the founding fathers concerns their policies toward native Americans. I quote “exterminist” exhortations made by Washington, Jefferson, Jackson and the first three governors of California. Gordon does not challenge these. He falsely says that I compare Lincoln to the Nazis. In fact, I call him “a relative moderate among presidents” (p94). I also say explicitly that the extermination of the native Americans was not like the Nazi Holocaust, since far more of the Indians died through ethnocide (in this case death through disease) than genocide, while the genocidal episodes came as largely unconnected bursts rolling westward across the continent over a long period of time. Thus there was no overall organization. Katz refers in his article to the most benign phase and aspect of American disease policy. I say that callousness was the predominant American response to the massive death-rate produced by the ethnocidal combination of disease and malnutrition among people who had been expelled from the most productive land. The result was an approximately 97% elimination of the native population, intermittently urged on by elected politicians. This was over twice as high as the overall death-rate in areas colonized by the (less democratic) Spanish.

Only his empirical argument about the Young Turks, if correct, would damage my overall explanation. Of course, my theses lock together in an overall explanation of ethnic cleansing which more extensive research might prove false. I make many empirical generalizations in my book. No doubt some will turn out to be incorrect, and more will remain controversial. I expect to generate debate and disagreement. But I hope this will be based on what my book actually says.

Michael Mann, M.A., DPhil., Hon.DLitt.
Professor of Sociology, UCLA.
2004-2005 Pitt Professor of American History and Institutions,
University of Cambridge

Review Forum Explaining Ethnic Cleansing

Review Forum
Explaining Ethnic Cleansing