Review Essay

Violence and the State in the Twentieth Century

MARK MAZOWER

Not long ago, modernization was thought to lead to prosperity, stability, and social welfare. When historical sociologists in particular sought to explain episodes of political violence along the path (or paths) to the modern era, they tended to see these as temporary. Both Barrington Moore and Charles Tilly, for instance, stressed the role of coercion and social conflict in modernization, but only as elements in a process of transition. Of late, however, violence has moved center stage, and the twentieth century is increasingly characterized by scholars in terms of its historically unprecedented levels of bloodshed. "More human beings had been killed or allowed to die by human decision than ever before in human history," Eric Hobsbawm has written. For Isaiah Berlin, the twentieth century was "the worst century there has ever been." Genocide, ethnic cleansing, and the killing of unprecedented numbers of civilians both in wartime and through acts of massive political repression have all contributed to what Charles Maier has described as an epoch of "moral atrocity." 1	1
For some, the causes are to be located in the innate violence of nationalism and the nation-state, for others in the rise of an impersonal bureaucracy and new forms of government; still others blame the Enlightenment and its various ideological offspring. Modernization, according to Stanley Tambiah, has brought the world not only mass literacy, urbanization, and rising living standards but also "massive civil war and gruesome interracial and interethnic bloodshed." "Ethnic cleansing," states Norman Naimark, in his study of the phenomenon, "is a product of the most 'advanced' stage in the development of the modern state." According to Omer Bartov, the Holocaust was "the culmination . . . of a process begun in the late eighteenth century and still continuing." The all-seeing bureaucracy, its repressive, all-controlling panoptic impulses to the fore, often more or less tightly tied to the paranoiac character of the individual despot, is now blamed for many of the woes of the twentieth century. 2	2
Behind this radical shift in perspective lies a combination of cultural, intellectual, and political developments. The belated reckoning of Western intellectuals with communism, especially after 1989, renewed an interest in the theories of totalitarianism. Then, in a desire to counter popular stereotypes of "ancient ethnic hatreds," which were commonly invoked in the 1990s to explain the recrudescence of violence globally after the end of the Cold War—especially in the Balkans, the Caucasus, and central Africa—many scholars began to emphasize the modern and state-derived character of mass violence. The strength and causal importance of more or less spontaneous crowd behavior, riot, and popular violence was downplayed by focusing attention on political elites and their proxies. Above all, the rise of Holocaust studies has seen the Final Solution—genocide at the hands of a highly organized state apparatus—turned into a paradigm for understanding modern violence, if not modern life altogether. Although many Holocaust scholars remain reluctant to contextualize Nazi violence in a broader, comparative historical framework, others are less inhibited, and even many of those who insist on the uniqueness of the genocide of the Jews want to claim it as a or perhaps the defining event of the twentieth century. 3	3
The cumulative effect of these developments has been to highlight the central role played by the violent state, and to see modern mass violence in terms derived from the experience of a small number of historiographically dominant European paradigms. It is, however, questionable how far these paradigms allow us to understand the origins of such diverse events as the massacres that accompanied the partition of India in 1947, la Violencia in Cold War Colombia, or the expulsion of ethnic Germans from Eastern Europe after 1945. Before this can be done, we will need a more finely grained analysis of what is meant by the state and what role different agencies and different forms of the state may have played at different times. We will need to test hypotheses derived chiefly from historical sociology or political philosophy against the facts, scrutinize the power of ideology in its historical context, and reintroduce the role of historic contingency both in time—the catalytic impact of wars, civil wars, and other upheavals—and space—geopolitical location, the proximity of disputed borders—in understanding why such large numbers of non-combatants have been killed by official or semi-official agencies in the twentieth century.	4
Above all, it must be open to question how far we can understand numerous other episodes of modern mass violence if we insist on viewing them in the shadow of the Final Solution, the event that continues historiographically to loom even over the numerous mass murders carried out by communist regimes. 4 The issue here is not modernity: there can be little doubt that the genocide of European Jewry cannot be understood in terms of atavistic throwbacks to medieval hatreds: Nazi anti-Semitism did not merely represent a revival or continuation of earlier Christian attitudes but drew extensively on contemporary racial science for its authority and legitimacy. Sociologically, leading Nazi cadres included highly educated individuals. The technology employed, and the state that deployed it, had a fair claim to be among the most advanced in the world at that time. Yet most other states that have perpetrated acts of mass violence over the past century were less efficient, differently organized, and motivated by different sets of beliefs and strategies. Perhaps the time has come, as I shall seek to suggest in this essay, to reconsider the usefulness of the Holocaust as a historical benchmark for modern mass violence, and to ask how useful the categories most recently associated with it—namely, genocide and ethnic cleansing—are as instruments of historical analysis.	5
Take, to start with , the two cases that most closely approximate—in scale and intention—the Nazi genocide. In 1915–1916, at least 800,000 Armenian civilians were killed in cold blood by Ottoman forces. There can be little doubt that the killings were deliberately planned and carried out at the highest reaches of the Ottoman state. The fact that the Armenian populations of Istanbul and Izmir were largely untouched simply means that the goal was not, as in the Nazi case vis-à-vis the Jews, complete extermination. But this is not the only difference between the two cases. The use of Ottoman rather than Nazi ethnic categories meant that some Armenians could escape death through conversion. Moreover, the structure of the Ottoman state differed sharply from that of the Third Reich: there was a cabal at the top rather than a single leader: we know far less about the Teskilât-i Mahsusa than about the SS, and we still lack an analysis of who really held power in Istanbul in 1914–1915. Above all, while the Third Reich plotted the extermination of the Jews at the height of German supremacy, in the spring of 1915 the very existence of the Ottoman Empire had been thrown into question by Russian victory at Sarikamis, British victory at Suez, and the threat of seaborne invasion of the Dardanelles. Stretched to its limits, facing imminent extinction itself, the Ottoman state possessed nothing comparable to the industrialized killing machinery of the Reich. 5	6
More murderously efficient, in terms of victims over time, than either of the above was the Hutu Power regime in Rwanda during the spring of 1994. Once again, it is not clear that the Nazi model of the state or of its geopolitical predicament is very helpful in understanding what occurred. Hutu Power did, of course, espouse an extremist ideology that depicted the Tutsi as a racial threat; many Hutus—like many Germans between the wars—saw themselves as victims of history and thus found it easier to turn their enemies into victims, too. But beyond this, it is hard to see what purpose the German comparison can serve. Sociologically, the key feature of Rwandan society was the pressure on land created by its extremely high population density and the vulnerability of a largely rural society to fluctuations in international markets. The regional context was all-important—with events in Rwanda closely linked both to political change in neighboring Burundi (where in 1972 there had been massacres of Hutus, memories awakened by the Tutsi army's killing of Burundi's first Hutu president in October 1993) and to the successful invasion of the country by the Tutsi-led Rwandan Patriotic Front from across the Ugandan border. There was thus a background of acute Hutu-Tutsi violence at the political and military levels already. Far from ruling the world, the government was about to be pushed out of power following the Arusha Accords. Faced with having to make the transition from one-party to coalition government, some Hutu Power extremists mobilized in defense of their privileges in a fashion not dissimilar to—though far more lethal than—Slobodan Miloevi's reaction to the breakdown of one-party rule in Yugoslavia. Here is a case where genocide was a sign not so much of the extremists' strength as of their weakness—both at home and abroad—and an instrument for consolidating themselves in power. Nazi extermination took place in quasi-secrecy, at least so far as the German populace was concerned; in Rwanda—as in Anatolia earlier—it was ubiquitous and inescapably public. And international powers played a much greater determining role than in the Nazi case, both in the past as colonial governors and at the time—France and the United States between them effectively prevented UN action at a time when it might have stopped the killings. 6	7
All three cases—the Holocaust, the mass killing of the Ottoman Armenians, and Rwanda—count as episodes of genocide, if the term is to have any meaning at all. But genocide is a slippery term for historians, for several reasons. As its significance in international law becomes greater, its legal connotations start to complicate its historical usefulness. The definition embodied in the UN Genocide Convention is both too limited—it acknowledges ethnic, racial, and religious but not political or economic repression—and astonishingly open-ended. It does not confine itself to episodes of mass murder. Cultural repression also counts in certain circumstances as genocide according to the Convention, even when no one is killed as a result. What matters for the lawyers are not the numbers of victims but the proportion of an ethnic group that is affected: in 1996, for instance, five miners in Brazil were convicted of genocide after the killing of sixteen Yanomami Indians. The lawyer may focus on the similarities with what happened in Rwanda; the historian is struck by the differences. But above all, the historian must surely take into account the comparative rarity of the phenomenon, at least in the commonsense usage of the term. Very few regimes have tried to wipe out entire ethnic groups by killing them, and characterizations of the twentieth century as "the century of genocide" exaggerate the significance of what is in fact a rather rare occurrence. If we are going to explore modern mass violence, we must cast the net wider. 7	8
More common throughout the late nineteenth and twentieth centuries as empires collapsed and nationalism gained ground have been those intermediate-range policies of violence known since the 1990s as "ethnic cleansing"—policies characterized by a combination of massacre and expulsion, deliberate acts of terror and looting, social humiliation and mass rape. Forced population movements themselves are no more a symptom of modernity than massacre; they have formed part of the repertoire of imperial rule at least since the early modern era. But just like mass killing, deportations also acquired new connotations in the twentieth century. And "ethnic cleansing" has come in the past decade to refer to a huge range of events, including the flight of Muslims from the Black Sea littoral into the shrinking Ottoman Empire, the Balkan Wars of 1912–1913, the expulsion of ethnic Germans from Eastern Europe, Palestinians from Israel in 1948, and members of the "captive nations" in the USSR under Joseph Stalin, right up to events during the 1990s in the Balkans and the Caucasus. 8	9
It was the wars in the former Yugoslavia from 1991 to 1999 that gained the term currency. In Bosnia, in particular, ethnic cleansing took the form of targeted assaults on isolated towns and villages by an overwhelming military force of regular troops, often backed by heavy artillery, from the Yugoslav People's Army, and associated paramilitary and irregular units. The latter were known to terrorize civilian populations by random shootings of non-Serbs (initially, later the practice spread on the Croat side as well), thereby prompting the rapid flight of the rest of the inhabitants. Often, the men were incarcerated in temporary camps, and women were raped. The result of this policy—and it clearly was a policy—was that hundreds of thousands of non-Serbs fled their homes within weeks. Ethnic cleansing, in other words, in this case, emptied the land of a very large proportion of the ethnic undesirables, who were turned into refugees. At least from the Bosnian Serb perspective, this was, or quickly became, the rationale for the war. Violence was needed, to force people to leave their homes. The goal was not their total extermination but their flight across the border into neighboring states. From the Serbian point of view, ethnic cleansing was an integral part of nation-building, or to be more precise nation-enlarging.	10
Like genocide, and indeed the term "Holocaust" itself, the label "ethnic cleansing" has since turned into a means of attracting attention to and claiming significance for various more or less neglected episodes in the past. Yet the parallels with the Bosnian case are frequently less striking than the differences, and reveal the difficulty of making a hard and fast connection between organized violence, the homogenization of populations, and nation or state-building. First, we need to establish intentionality, and to distinguish forced deportations, refugee movements fueled by deliberate public acts of terror, from panics and more voluntary migrations. It is often very difficult to distinguish between expulsion and panic, and the two may be intertwined, as the case of the Germans of East Prussia in 1944–1945 suggests. Although some historians refer to the flight of thousands of Crimean Tatars into Ottoman lands in the face of Russian advances during the nineteenth century as "ethnic cleansing," recent research suggests that this often took place despite the wishes of the imperial Russian authorities, not because of it: their flight meant the loss of a valuable agrarian labor force and the abandonment of their lands. In fact, many Tatars and Balkan Muslims fled once political control passed from Muslim into Christian hands because they did not wish to remain in a non-Muslim state, especially one where their sons might face conscription into an army that could be used against the Ottomans. On the other hand, the label ethnic cleansing provides a closer fit for Russian treatment of tribesmen in the Caucasus, where the imperial army decided to expel thousands of its former opponents once the Shamil uprising had been crushed. It is no coincidence that whereas the Tatar colonists who resettled in Ottoman Bulgaria got on peaceably enough with their neighbors, the Circassians became a mainstay of the irregulars deployed across the Balkans to suppress Christian nationalist uprisings. 9	11
More crucially, in cases of coercion, we need to ask who organized the violence and how they fitted into what passed for a state apparatus. In the Bosnian case, for instance, there was an intricate set of relations between the government of Serbia, the Yugoslav People's Army, and the makeshift paramilitary groups clustered around Radovan Karadi and other notables. Violence was as much a means of making a new state, and laying claim to a prominent place within it, as of securing or expanding the power of an existing one. The anthropologist Cornelia Sorabji has suggested that "rather than an organisation of specific violent techniques, [the war in Bosnia] suggests the organisation of a context in which people are enabled to inflict whatever disorganised torture they may dream up. The context is organised, however." It would be worth relating this account of an over-stretched or incipient state, encouraging and inciting the collaboration and complicity of private individuals and groups, to Jan Gross's analysis of how totalitarian regimes of occupation in the Polish-Lithuanian borderlands similarly organized a context in which violence could be generated locally between 1939 and 1941. The process continued as Soviet power was reestablished amid the expulsion of ethnic Germans from Poland's western borderlands a few years later (and Ukrainians from its eastern borderlands), again offering an example of the way ethnic violence formed part of the establishment of a new system of state power. An analogous case is that of the Albanian Muslim Cham minority, expelled from northwestern Greece on charges of collaboration at the end of the war—an act initiated largely by local military powerbrokers and only afterwards ratified, as it were, by the beleaguered Greek state far away in Athens. 10	12
One common feature of all the above cases is that they took place in Europe or on its fringes, a part of the world where state structures have been relatively well organized since the early modern era. Forcing peoples from their lands has happened, of course, on a much wider scale, and forms part of the history of global colonialism: prioritizing the state's role in the perpetration of large-scale violence has had the effect of marginalizing this kind of settler violence, whose victims have tended to be indigenous peoples leaving few historical records behind them. In areas such as Australia, Russia, Africa, and the Americas, violence against natives was perpetrated by mostly European colonists, sometimes backed from afar by the metropolitan power but often the result of local initiatives in these frontier societies, motivated by the colonists' desire to control land, water, and other resources. Whether this should count as ethnic cleansing is a moot point, an issue that hinges largely on what the political goals of the perpetrators were. But settler violence may well be connected with mainstream ethnic cleansing more closely than is commonly admitted, since it often established attitudes and practices for the policing of the indigenous inhabitants of colonial societies, which came in the early twentieth century to influence police and military behavior by European states vis-à-vis their own populations and those of adjacent conquered territories, too. 11	13
The issue of how ethnic cleansing is organized is closely linked to the prior question of how and at what point the relevant policies were formulated and arrived at. The expulsion of 10–12 million ethnic Germans from Eastern Europe needs to be seen in the context of the previous six years of total war, initiated by a German regime that itself had brought unprecedented bloodshed and forced upheaval of populations to the region. There was unanimity on the part of most of the major actors—the Soviet Union, its Allies and satellites, and indeed popular opinion in the region—that short-term justice and long-term regional peace could not be gained without the expulsion of the Germans. Violence and mass rape at the hands of the Red Army accelerated the panic that had set in even before the Reich's defeat. But whereas in Eastern Europe, the ground had been laid for the expulsion of the Germans—an expulsion that took place largely after the fighting was over—by prior discussion of national and international leaders, in other cases of mass expulsions happening at the same time, the picture looks different.	14
In the 1948 war for Palestine, recent research has confirmed that Israel's war of independence was accompanied by several massacres and by the deliberate expulsions of civilians. Zionist statesmen such as David Ben-Gurion and Chaim Weizmann had speculated earlier about getting rid of Palestine's Arabs after the creation of the Jewish state. Nevertheless, the linkage between these statements and the actions that took place on the ground subsequently is difficult to prove. It seems likely that the idea of expelling Arab civilians in as large numbers as possible came in the course of the fighting itself and was, perhaps for that reason, realized in more haphazard and partial a fashion than in the German case. In the Israeli case, the state itself came into being in a rapidly changing environment in which military commanders found themselves able to make far-reaching decisions. 12	15
All these ambiguities around the state's role as agent of mass expulsion are redoubled in the case of the 1947 partition of India, where ethnic cleansing—if that is indeed what occurred—took place in something close to a political vacuum. In terms of sheer numbers, the scale of displacement was second only to the expulsion of ethnic Germans, going on at the same time. In terms of those killed, it was far more bloody and murderous. Yet the killings were triggered by the manifest incapacity of the colonial Indian state at all levels to control the situation in vital border areas, notably the Punjab. Once the new postcolonial states came into existence, on either side of the new border, they proved able to bring the killing to a halt relatively quickly, leaving a large Muslim population in India and a smaller Hindu minority across the border. The violence here is thus scarcely to be attributed to the all-powerful modern state—whether colonial or postcolonial—although its prime cause, of course, had been the partition policies shaped in New Delhi and London. 13	16
Here, it is obvious that the logic of state interest is insufficient to account for violence. Must we turn to more individualistic and subjective factors? Not necessarily: between the level of the state and that of the individual perpetrator lies that of local and regional powerbrokers, whose importance is increasingly emphasized by scholars of communalism in India as well as by those of its Russian near-equivalent, the tsarist pogrom. According to Joya Chatterji's study of Bengal, for instance, the 1946 "Great Calcutta Killing," which left at least 5,000 dead, was "not a riot but a civil war," involving political imperatives and the organization of volunteer groups by local politicians and elites. The following year, virtual private armies and paramilitary organizations like the Sikh jathas, often including men with military expertise acquired in British service, outwitted the hard-pressed Punjab Boundary Force. But even such a smaller-scale instrumentalism needs to take into account the more subjective and emotional aspects of mass violence of this kind. For if violence sometimes served to terrorize populations and force them to flee, at others it betrayed a less rational motivation, as frequently massacres took place along roads and rail lines in which the victims were murdered precisely while they attempted to leave the area. "We simply went mad," recalls one Sikh who participated in a killing spree in a Muslim village during the Indian partitions. "We were swept away by this wild wave of hatred," were the words of a Muslim perpetrator in the same events. In like manner, Tambiah refers to "jubilant violence," and sees the 1983 anti-Tamil riots in Colombo and the 1984 anti-Sikh riots in Delhi as epicenters of mass emotion. Less jubilant perhaps, but the Nazi death marches displayed a similar irrationality. It is at this point perhaps, that Europeanists—whose state-centered approach has often hidden a reluctance to consider the idea that occasionally ordinary people enjoy or take pride in killing—could learn from scholars of South Asian communal violence. In the work of Veena Das and Sudhir Kakar, in particular, we see a subtly modulated psychology at work that avoids the conceptual rigidities found in some well-known discussions of Nazi perpetrators. 14	17
No doubt because there can be little question of the Soviet state's ambition to organize society as effectively as possible, the terms "ethnic cleansing"—and indeed "genocide"—have been applied there, too. For some scholars, these categories provide a way of locating the experience of the USSR in a much broader context, linked to the Holocaust and other aspects of the darker side of modernity. Peter Holquist presents interwar Bolshevik political surveillance as "a subfunction of the modern form of politics," while Amir Weiner has sought to show that, in Ukraine in the 1940s, the Soviet leadership was carrying to an extreme the purifying and violently exclusionary elements inherent in the "aesthetic enterprise" of building a "better, purer and more beautiful community." In such arguments, one may discern the influence of Michel Foucault's thinking about governmentality as well as a vein of late twentieth-century anti-utopianism that points to the dangers inherent in all schemes for human improvement and social engineering. 15	18
But although the state-directed character of political violence under the Bolsheviks cannot easily be questioned, the degree to which communist violence should be "ethnicized" may. The USSR did draw on ethnic categories in its political policing and began deportations of members of specific national groups in the 1930s. No other state had ever moved—or, for that matter, killed—people on a comparable scale. But most Soviet deportations were not directed against national groups in their totality: initially, they were designed to secure borders and thus possessed a quasi-military function of a kind that could be traced back to imperial Russian policy in World War I, and even earlier. 16	19
Although the huge deportations of Polish and Baltic people in 1940–1941 did not aim at the displacement of anything close to entire nations, the treatment of Chechens, the Crimean Tatars, and others does take us much closer to the Yugoslav paradigm of ethnic cleansing. Nevertheless, there remains an important difference. The Soviet authorities envisaged deportation as collective punishment: however, it was not deportation out of the country but farther into the interior. Thus the border—so crucial in both cases—had a very different function each time: point of no return, in the Yugoslav, German, or Palestinian cases, but for the Soviets, rather a neuralgic zone, whose security might demand the removal of suspect groups inland away from possible contact with enemy states: on the one hand, expulsion from the political community, but not on the other.	20
Does this count as ethnic cleansing? Yes, according to J. Otto Pohl, who sees Joseph Stalin drawing on older imperial Russian traditions of population displacement; yes, too, for different reasons, for Terry Martin, who argues that the mid-1930s saw a transition from class-based deportations to ethnic targeting as part of the state's modernization drive. Yet the extent to which the deportations reveal an underlying racial dimension to communist ideology is altogether harder to prove. Of the switch from an autonomist to a more repressive line toward nationalities in the 1930s there can be little doubt. Even during the Great Purge, nationality was not among the key target categories: but between the liquidation of "Polish spies" and the operations directed against entire peoples after 1937, when both the Koreans and the Black Sea Greeks were targeted en masse, there was clearly an important shift in official thinking. Yet unlike "ethnic cleansing" Yugoslav-style, this was not about destroying nations so much as targeting "counter-revolutionaries" in the context of Moscow's fear of imminent war. The need felt by some contemporary scholars to demonstrate a racial dimension to Soviet communist policies appears to derive from, or at least to reflect, the continued power of the Holocaust paradigm in discussions of European mass violence. Yet to the outside observer, it scarcely seems to make much difference from the ethical point of view whether Stalin connived in, or tolerated, the deaths of millions of Ukrainians during the famine years of the early 1930s because they were Ukrainians or whether, as seems more likely, because Ukrainian farmers held the key to the regime's control of the food supply: to that extent, the issue of genocide is a red herring, although it matters a good deal of course to Ukrainian nationalists anxious to ascertain their own relationship to their communist past. 17	21
Whether or not the USSR went in for ethnic cleansing and genocide, it was certainly more murderous toward its own citizens—at least in peacetime—than any other country had been until that point. Was this perhaps because of its totalitarian character? Such a question suggests comparing its forms and levels of violence with those of the other great totalitarian regime of mid-century Europe: Adolf Hitler's Germany. Curiously, though, in view of the political salience of the issue (which fueled heated intellectual debates in both France and West Germany in the late twentieth century over the comparative criminality of the two regimes), there has been little in the way of sustained historical comparison. In fact, from the viewpoint of levels of political repression, the two systems before 1940 were rather different: the Gulag's interwar population soared far above 1 million, while the Nazi camps were generally well below 100,000 inmates; nor could the scale of mass shooting in and around the Great Terror be said to bear much resemblance to such relatively small beer as the Night of the Long Knives. Nothing in the pre-war experience of the Third Reich compared with such episodes as the shooting of 9,000 people by the NKVD in Vinnytsia alone in 1937–1938. In purely quantitative terms, it is clear which political system required a higher degree of coercion to maintain itself in power. The war, however, saw the numbers gap narrow quickly: for although the population of the Gulag rose to around 2 million, the Nazi camp system expanded much faster between 1939 and 1944, eventually attaining a population of around a million inmates. Like the OGPU/NKVD earlier, the SS was now transformed into a major industrial producer. One difference was that the Germans established several industrialized extermination camps for racial mass murder that had no Soviet equivalents. Another was that the Nazi system—largely a product of the war—was only closed down by military defeat. The fate of the Gulag was more complex: given a new lease on life, and indeed extended across Eastern Europe in Stalin's last years (in 1953, it held more than 2.7 million prisoners), the camp system shrank, and prisoners were given amnesty in the thaw that followed. 18	22
Does the concept of totalitarianism gives us much purchase on the rise and fall of these two systems? Probably not: there are too many fluctuations, too many discrepancies: indeed, if we were to count Benito Mussolini's Italy—a country that passed fewer than fifty death sentences for political crimes before 1939—as a totalitarian state (a characterization espoused by the regime itself at various times and recently adopted by some of its historians), the heuristic value of the concept would seem even more doubtful. Moreover, totalitarianism itself suggests a kind of structural approach to a phenomenon that was clearly much influenced by highly conjunctural factors, above all, by war. The same conclusion emerges from a study of the even more highly charged terrain of mass deaths. Let us try to set aside an intellectually redundant if highly charged debate over whether Hitler or Stalin was responsible for more victims. (For what it is worth, a recent estimate suggests that while the Stalinist regime "may have caused the premature death of more people than Hitler's regime . . . [the evidence] does not show that it purposely killed more people.") The totalitarianism thesis offers no obvious explanation for the scale and chronology of killings in either case. For both the Nazi and the Soviet regimes, there were moments of escalation when the killing toll mounted very rapidly. For the Nazis, this was, again, the war, first in autumn 1939 and then, a new threshold, in June 1941, with the invasion of the USSR and the onset of a self-proclaimed "war of annihilation." For the Soviets, the invasion of the Baltic states and Poland in 1939–1940 was a similar such moment, though not the first: as Nicholas Werth has argued, there were at least four cycles of violence—starting with 1917 (thereby reinforcing the point that it was not only the Second but also the First World War that played a key role in escalating norms of violence). In particular, Werth alludes to the 1919–1920 "massive extermination" of the Cossacks as a precedent for future mass killings, and argues that what the Bolsheviks did after 1920 was to extend the principle of civil war to their own society. This strikes me as a key insight that helps us understand why some states killed so readily. As the Italian anti-fascist Carlo Rosselli first pointed out, it was a feature of mid-twentieth-century ideological states that they rather readily blurred the boundary between internal and external enemies, and thus redrew the political dividing line within their own societies between those deemed loyal and thus regarded as in practice or potentially beyond the pale. In this respect, they differed sharply from their nineteenth-century predecessors, for whom disloyalty and treachery were two separate concepts. 19	23
Yet it is worth noting that not all communist—or indeed totalitarian—regimes practice unremitting violence, or experience the same levels of violence as one another, as a more comparative survey of communism indicates. The sheer numbers involved in the Chinese state's tyranny over its own people dwarf even the Soviet case: a prison camp population of nearly 5 million by 1949, rising to nearly 10 million by the early 1980s; a huge death toll—estimates range from 20 to 43 million—in the 1959–1961 famine, a direct result of the Great Leap Forward. As in the Soviet case, alongside the unmistakable role of ideology (here intensified by Mao Zedong's rivalry with the USSR) must be set the impact of nearly two decades of war, colonial occupation (by the Japanese), and civil war. And just as Stalin's death saw the repression scaled down, so Mao's death led to a new moderation, as the army disbanded the Red Guards. Any explanation of communist violence must surely be able to encompass the termination as well as the genesis of episodes such as the Terror or the Cultural Revolution, and this is likely to require explanations that see ideology not so much as a causative factor in its own right but as an element in the political struggle within the state apparatus between different groups and factions, whose interests ride on the promotion or termination of violence. 20	24
The bloodbath carried out by another communist movement, the Khmer Rouge, and brought to an end by the invasion of a communist neighbor, Vietnam, may be viewed in similar fashion. While China was moving away from revolutionary radicalism, perhaps the most radical experiment of all was being undertaken to the south in Cambodia, under the leadership of Pol Pot. His brief but almost unbelievably murderous reign involved a kind of competition to demonstrate the profundity of his achievement. "We are making a unique revolution," boasted one cadre. "We are much better than the Chinese who look up to us. They are trying to imitate us but they haven't managed it yet. We are a good model for the whole world." Jean-Louis Margolin sums up his achievements as follows: "Money was abolished in a week; total collectivisation was achieved in less than two years; social distinctions were suppressed by the elimination of entire classes of property owners, intellectuals and businessmen; and the ancient antagonism between urban and rural areas was solved by emptying the cities in a single week." In just four years, perhaps a million people were executed and another 700,000 or more died from hunger and disease: some professions and some minorities were almost entirely wiped out. Nearly half the population of the capital may have perished. 21	25
If death on this scale is hard to explain, a few factors seem obviously relevant: in the first place, the Khmer Rouge came to power in the midst of a war next door that had spilled over into Cambodia with catastrophic social and political effects, including pogroms against the Vietnamese, massive American bombing raids, and huge movements of population. (It should also be borne in mind that the Khmer Rouge fell from power once the war in Vietnam was over.) This tremendous wartime flux did not end with the arrival of the Khmer Rouge: on the contrary, they began their reign by massacres, "reeducation" campaigns, and, eventually, the total evacuation of the capital, Phnom Penh. This was a huge upheaval, which threw the traditional institutions of society into turmoil and facilitated the extreme social engineering of the country's new rulers. However, it was surely the combination of ideology and internal party politics that was chiefly responsible for the ensuing bloodbath. The very weakness of the Kampuchea Communist Party resulted in purges, massacres, and witch hunts; it was, moreover, locked in a bitter struggle to assert its autonomy from the Vietnamese party—the eastern part of the country, close to Vietnam, was where the worst massacres took place, in 1978. At the same time, famine, which had first attained serious dimensions in 1976, continued to plague the country. Impossible economic targets, poor planning, the rejection of technology, and a hostile attitude toward any criticism were the ingredients of disaster. The dehumanization created first by war and social destabilization in general, and by the famine in particular, helped erode the moral inhibitions against widespread killing. The regime aimed, for example, to weaken the family, which it saw as a threatening institution, and in fact it insisted on usurping family roles. It also deliberately disregarded traditional burial norms in the official treatment of the dead. Meanwhile, the death sentence was meted out for a wide range of crimes, and an atmosphere of paranoia and suspicion ensured its frequent use. Perhaps that paranoia had been there from the start: the national anthem started with heavy emphasis on the "bright red blood that covers towns and plains." But it was exaggerated both by the character of its mysterious leader and by the fact that his movement was rather weakly based and insecure and used terror in order to shore up its position. 22	26
Much of the discussion of communist ideology has focused on the degree to which socialism could ever have been made the subject of a political program without giving rise to massive bloodshed. But there was a geopolitical dimension to ideology too, which is less frequently highlighted in the literature. The Cold War in particular played an enormous role in creating a climate of mutual hostility and paranoiac suspicion between and within the two great power blocs that divided the world. Anti-communism identified enemies at home as allies of those abroad, and this, at the very least, provided a language that made mass killing permissible to the leaders of the Free World as well.	27
In the early postwar era, as America's international role grew rapidly, mass killings and repression increased under Washington's gaze. This began on a relatively small scale in Greece during its civil war (1947–1949) but moved to a different level in Central America and Southeast Asia. In Cuba, Venezuela, Peru, and elsewhere, anti-guerrilla campaigns were pursued through the 1950s and 1960s. The impact of state terror, aerial bombing of peasant areas, mass rape, execution, and massacre varied from country to country. In Bolivia, there were relatively few casualties; but in Colombia, la Violencia was responsible for roughly 200,000 deaths. In Peru, the bombing of guerrilla zones in the 1960s was the most intensive in South America until the Nicaraguan war of the late 1970s. In Guatemala, consistently among the most violent states in the region, the army killed roughly 100,000 people between 1980 and 1985 alone. 23	28
Anti-communism offered a legitimizing language shared by South American elites and their patrons in Washington. But more important was the evolution of Western military thinking about counter-insurgency, developed on the basis of experience during and after World War II and then spread from the U.S. Army via instructors and advisers to senior army and police chiefs in client states. This is a complex and tangled piece of intellectual history that has not received as much attention from theorists of mass violence as it perhaps deserves. Its study has been pioneered by scholars of "Western state terror," and their work points the way to a clearer understanding of how far the counter-insurgency doctrines espoused by U.S. Army thinkers were actually taken up and followed by their clients and how far the latter were following their own inclinations. 24	29
The role of the army as a key instrument of political violence in the apparatus of the Cold War state, its power bolstered by Washington's acquiescence, is illuminated by events in Indonesia in 1965. Nowhere is the oddly skewed character of the academic debate on mass violence more evident than in the extraordinary neglect of this case by historians outside the small number of area specialists concerned. The story, briefly told, is that a botched PKI coup was crushed by the army and civilian vigilantes, who went on to murder perhaps half a million people in a campaign to wipe out communism in the country. The army's role was both organizing and permissive. It inflamed public emotions by displaying publicly the bodies of six army generals assassinated by the coup plotters, and when senior army officers indicated their approval of the initial revenge killings, in mid-October, only then did massacres spread across the country. Many people died under the guise of an ideological crusade for reasons that had little or nothing to do with ideology. But this was perhaps inevitable when, as one survivor described it, "the atmosphere of vengeance spread everywhere" and when a new figure—the local executioner—emerged, who needed to display his prowess and necessity. The same army—give or take a few purges—was responsible for further bloodletting in East Timor after the Portuguese pulled out a decade later, and it is tempting to see the first massacre as a laboratory for later violence. A crucial element in Indonesia was the stance of the United States (and later of Australia, too); the unapologetic memoir of former U.S. Ambassador Marshall Green makes clear the kind of mentality that guaranteed the Indonesian military a more or less free hand. 25	30
Armies had not enjoyed such power in the totalitarian one-party state, which built on the bureaucratic traditions of its imperial Hohenzollern or Romanov predecessors, precisely by establishing civilian control through the party and a powerful secret police. Similarly, in Miloevi's Serbia, the power of the Yugoslav People's Army was curbed by frequent purges and, even more important, by the creation of alternative private channels for Miloevi through the Interior Ministry to special police and paramilitary units, whose relatively unrestrained use of violence against civilians both alarmed and challenged the army itself. From this point of view, postcommunist Yugoslavia, the Third Reich, and the USSR form a striking contrast with the far more common type of praetorian state found in many postcolonial countries outside Europe, where civilian politics remains weak and power is concentrated in the hands of a military apparatus that sees itself as the guardian of national values, especially where borders are fragile and societies are multi-confessional and multi-lingual. 26	31
There were, of course, pre-war instances of such states even inside Europe: Poland and Bulgaria in the 1930s, or countries where generals established dictatorships, as in Greece, Romania, and Franco's Spain, which emerged with high levels of repression after a bitter civil war. (Hungary and Finland, both of whom experienced bitter civil wars in the aftermath of World War I, would make useful comparators.) In the postwar period, military dictatorships—or guided democracies—became far more common and were involved in some of the worst episodes of mass killing. When East Pakistan broke away from Pakistan to form Bangladesh in 1971, the Pakistani army killed hundreds of thousands of people. The Nigerian army in secessionist Biafra, the Indonesian army in East Timor, and the Turkish military in southeastern Anatolia all saw themselves as defending the unity of the new state from the forces of fragmentation and disintegration. In the Philippines, thirty years of fighting between government forces and the New People's Army and National Liberation Front led to at least 70,000 deaths. In Algeria, some 80,000 to 100,000 people are estimated to have died in the 1990s alone, as the army struggles to keep Islamicists from power. In these cases, it is surely the weakness of the states concerned, not their strength—in particular, the weakness of their traditions of civilian politics—that helps explain the prominence of the military and their apparent impunity.	32
But not everything should be blamed on the military: as events in Algeria, southeastern Turkey, Sri Lanka, and South America demonstrate, terror also forms part of the repertoire of insurgent groups. States are often weak because others challenge their monopoly of the use of force in pursuit of religious, regional, or political ends. In the many insurgencies that proliferated in the postwar era, there emerged guerrilla organizations and insurgent armies that also deliberately deployed violence against civilians. Stathis Kalyvas has suggested, in a kind of rational-choice model of insurgent violence, that guerrillas seek to use terror against civilians in order to maximize their support, especially when their hold over the latter is weak, when peasants show signs of defecting to the other side, and when the general level of violence in the wider military confrontation escalates. He points out that cruelty and public displays of power may form part of the same overall strategy to secure loyalties and deter defections, as they do for the army. And, although Kalyvas does not make much of it, revolutionary ideology may be an important factor, too. In Peru, for instance, the violence of the state was matched and eventually exceeded by that of the Maoist insurgent group, Sendero Luminoso (Shining Path). As the army became more sophisticated and less crudely violent in its approach to the peasant highlanders, the insurgents became more uncompromising, talking of a revolutionary "river of blood" and a "radical intensification of violence." By the late 1980s, their leader, Abimael Guzmán, was promising to "exterminate whole communities" in the effort to stop them cooperating with the state authorities. Indeed, Guzmán's movement seemed to go beyond Mao in according violence an absolute value as a form of revolutionary purification. But he was also engaged in a competition with the army to show the peasantry who was tougher. The decision to concentrate on the urban areas, a decision that led eventually to Guzmán's arrest, was an indication that the army had won that particular competition. 27	33
Finally, just as the Russian and Chinese civil wars may be seen as the seedbeds of a propensity for those running the state to wage war on their own society, so elsewhere the violence inherent in imperialism itself may have played a part in establishing norms of state violence after independence in post-imperial states. In post-imperial Eastern Europe, this process had already started with the Balkan Wars of 1912–1913, ramifying after 1918 into complex intra-ethnic struggles in Eastern Europe, and Central Asia, as well as into new major military conflicts such as the Russian civil war, the Russo-Polish war, the Greco-Turkish war, and the struggle for control of Lebanon and Syria. It was this decade—lasting until 1922–1923—that was the catalyst for genocide, ethnic cleansing, and massive forced population movements for the first time in history. As Aviel Roshwald argues in a recent, very valuable work of synthesis entitled Ethnic Nationalism and the Fall of Empires, the whole area from the Baltic states to the Middle East constituted one vast arc caught up in this dynamic of imperial collapse and nationalist struggles. What expanded the scale of the violence was that war had also led to a substantial enlargement of the state's powers over its citizens. Propaganda, welfare, and martial law were among the manifestations of this; so, too, was the rise of special services and paramilitary units allied informally to political centers of power. Detention camps, internal exile, and forced labor units for suspected subversives were employed from France to tsarist Russia. 28	34
These policies reflected the shifting balance of administrative power from civilian to military authorities. Entire populations were deported on the grounds that they were not trustworthy: tsarist forces deported communities of Jews from the Pale, fearing their pro-German sentiments; Habsburg, especially Hungarian troops, targeted Serb villages in Habsburg lands adjacent to the border with Serbia, burned houses and killed some civilians, and imprisoned others. Greek males were sent inland to labor details by Ottoman troops in the summer of 1914. The militarization of state bureaucracies also meant tougher treatment of suspected spies, the incarceration of aliens in special camps and the extension of censorship. Karl Kraus's satires on the Habsburg war machine and its lethal inefficiency should perhaps be read less metaphorically than is customary and more as warnings of the potential for bureaucratic violence that had been exposed in 1914. Where civilians did resist military occupation forces, as in Belgium, Serbia, and eastern Anatolia, for example, exemplary executions and reprisals took place and were often deliberately photographed and publicized to increase their deterrent effect. A thread of continuity of policy and attitude runs from the Habsburg reprisals against Slav civilians in Serbia in 1915–1916 and the policies of the Wehrmacht field commanders there twenty-five years later. Out of this turmoil emerged not only the Bolshevik state but also the Kemalist state in republican Turkey and the new heavily militarized state structures of the post-mandatory Middle East. 29	35
Outside Europe, the death agony of empire was more protracted and the struggle to hang on more violent: as resistance and protest escalated, imperial policemen in defensive mode developed counter-insurgency doctrines of their own. The British, influenced by fiscal constraints and by their own distinctive policing traditions, defined a doctrine of "minimum force," although they did not always live up to it in Ireland, Palestine, Iraq, or Kenya. In Algeria and Madagascar, the mass killings of thousands of civilians accompanied the French effort to retain power. In Angola, Congo, and Mozambique, long-running anticolonial wars turned into ongoing civil wars after Belgian and Portuguese colonialists pulled out. It would be absurd to attribute the power of armies in politics across the Third World entirely to the violence of colonialism; the weakness of civilian party politics, the impact of the Cold War, and indeed the growth of the modern arms trade must all have played a crucial role. Nevertheless, it would be wrong for historians to take imperialism's historical legacy of violence for granted: it was, after all, the Europeans who had exported modern warfare across the globe to build those empires in the first place, although the Japanese and others learned fast. 30	36
And that legacy of violence may also have contributed in its way to the militarization of Europe itself. Drawing on Hannah Arendt and others, we can discern continuities in the practice of organized violence between European policing of the colonies in the late nineteenth century and the later shift to total war on the continent itself. Military and policing doctrines, more or less imbued with a racist contempt for the enemy, emerged in a colonial context that would help shape the new brutality displayed by European armies toward non-combatants on the Old Continent after 1914: thus precursors for the Wehrmacht's brutality toward civilians in World War II have been found in the expeditionary force sent to crush the Boxer Rebellion, as well as in the 1904 war against the Herero in South West Africa (now Namibia), where the military commander, Adolf von Trotha, insisted unequivocally that "the nation must be annihilated as such." But it would be premature to see such behavior as a peculiarly German issue before far more research has been undertaken into the comparative development of colonial policing tactics among the various European Great Powers. One advantage of such research is that it might open up a perspective that allows us to link the violence unleashed by major states in Europe itself between 1930 and 1950 with both an earlier history of imperial violence and a subsequent history of violent decolonization and postcolonization globally during the Cold War. 31	37
In conclusion , it is striking how much the debate about mass violence has been dominated by a small number of decontextualized European exemplars—notably, the Holocaust and Stalin's USSR. I have tried to suggest in this essay that generalizations based on the handful of emblematic mid-twentieth-century cases of mass killing may be of only limited usefulness in understanding other episodes in which large numbers of people have been murdered, imprisoned, or forced to leave their homes. Structural or systemic forms of explanation need to be able to encompass the contingency of geopolitical location and the impact of wars. And both the Holocaust and the experience of Soviet communism may be better understood in a historical context that stretches back to the age of empire and forward to encompass the spread of independent, more or less violent, states across the globe.	38
In this story, war is evidently a crucial catalyst. In cases such as the Third Reich, total war led directly to an exponential increase in the murderousness of its policies. This is one important difference with the Soviet Union, which did not need an actual war to kill large numbers of its own civilians. On the other hand, the experience of total war—whether in 1914–1918 or in the subsequent civil war—did clearly have an enormous impact on the way the Soviet state regarded its own citizens. And "dirty wars" more generally, which blurred the distinction between combatant and civilians—whether civil wars, colonial policing operations, or later "liberation struggles"—helped generate new military doctrines and practices that increased the state's readiness to use violence against its own subjects, whether that state was being run by communist party cadres or military men.	39
And who exactly controlled the state turns out to be the crucial issue: Hitler's Germany and Stalin's Russia—for all that historians and political scientists have sought to demonstrate the apparently anarchic character of their bureaucracies—differed from most other countries in the past precisely because of the relative orderliness, ambit, and coherence of their state machinery. There are, at the other end of the spectrum, instances of mass violence where the state is either entirely absent or an onlooker. The state, for instance, was often virtually absent in settler colonies, leaving settlers to battle it out with indigenous peoples for control of land and resources. As we can see in the case of North America and Australia, the great technological imbalance between European colonists and their opponents often resulted in massacres or forced population movements, which the colonial state then had to respond to, after the event, through its judicial or administrative machinery. Nor is the situation entirely different today in some of the more remote and inaccessible parts of South America and South Asia, where indigenous peoples continue to find themselves faced with violence at the hands of settlers and the agents of economic interest groups. Indeed, the neglect by most historians of the whole issue of political violence directed against indigenous peoples underscores the need to move away from an overtly state-dominated understanding of mass violence. 32	40
Rather more common, and more lethal, is the situation in which a weak state vies to preserve its existence in the face of the threat of organized violence from armed insurgents. There, both sides utilize violence and may generate their own doctrines or rationales for it. During the Cold War, such struggles were often described in terms of the broader cosmic conflict then under way between communism and anticommunism. But the problem of weak states has not disappeared with the collapse of European or Soviet communism, and the high levels of violence against civilians that result remain a fact of life in eastern Africa, the Philippines, and elsewhere.	41
Although events since September 11 have reinforced elite fears in the West—already visible since the end of the Cold War—that future political challenges are indeed going to emanate not so much from established states as from fissiporous and unstable states that offer refuge to, or may indeed be taken over by, groups of terrorists, writing the state's obituary, as some are already doing, is surely also premature. The war against Afghanistan has been justified largely in such terms. Is it possible, then, that we stand at the beginning of a new era, in which the capacity to commit violence on a large scale passes from states to terrorists? Amid all the hype, it may be prudent to recall the difference between violence and terror: it may well be that more civilians have been killed in Afghanistan by military forces allied to the United States in the war against the Taliban than were killed on September 11 in the USA; more Palestinians have certainly been killed by the Israeli Defense Forces in the past two years than Israelis have been killed by suicide bombers. Terrorists rely on terror, such as that spread—more effectively than death—by the anthrax attacks in the USA, but the ability to use overwhelming force still remains in the hands of technologically advanced states. We need not to write off the violent state but to understand better what it does and how it behaves. And to do this, we will need to abandon the very partial and highly Eurocentric version that still dominates the agenda in contemporary history of what counted in the century that has just passed. 33	42
Mark Mazower is Anniversary Professor of History at Birkbeck College, University of London, and the author of several books including Inside Hitler's Greece: The Experience of Occupation, 1941–1944 (1993), Dark Continent: Europe's Twentieth Century (1998), and The Balkans: A Short History (2000). He studied at Oxford and Johns Hopkins, Bologna Center, and has also taught at Princeton and Sussex. He is currently writing a history of Salonica from the fifteenth century to the present. Notes Thanks are due to Marwa Elshakry, Laura Engelstein, Mark Levene, Stephen Kotkin, Jonathan Mazower, and the anonymous readers of earlier drafts of this essay. Without their encouragement and help, this piece would not have been written. 1 Barrington Moore, Jr., Social Origins of Dictatorship and Democracy (London, 1966); Charles Tilly, Coercion, Capital, and European States, AD 900–1990 (London, 1990). Compare Anthony Giddens, The Nation-State and Violence (Cambridge, 1985); Michael Mann, The Sources of Social Power (London, 1986); J. L. Talmon, The Origins of Totalitarian Democracy (New York, 1970); Yves Ternon, L'état criminel: Les génocides au XXe siècle (Paris, 1995); E. J. Hobsbawm, The Age of Extremes: The Short Twentieth Century (London, 1994), 12; Berlin cited by Michael Ignatieff, Isaiah Berlin: A Life (New York, 1998), 301; Charles S. Maier, "Consigning the Twentieth Century to History: Alternative Narratives for the Modern Era," AHR 105 (June 2000): 807–31, quote 812. The issue of the total number of victims of mass violence will not concern me here. The statistical problems appear to be overwhelming. Those interested in the available figures will want to start with the pioneering but highly unreliable Gil Elliot, Twentieth Century Book of the Dead (London, 1972); as well as Rudolph J. Rummel, Death by Government (New Brunswick, N.J., 1994); and Barbara Harff and T. R. Gurr, "Toward an Empirical Theory of Genocides and Politicides: Identification and Measurement of Cases since 1945," International Studies Quarterly 32 (1988): 359–71. 2 Omer Bartov, Murder in Our Midst: The Holocaust, Industrial Killing, and Representation (New York, 1996), 67–70; Stanley J. Tambiah, Leveling Crowds: Ethnonationalist Conflicts and Collective Violence in South Asia (Berkeley, Calif., 1996), 3–4; Norman M. Naimark, Fires of Hatred: Ethnic Cleansing in Twentieth-Century Europe (Cambridge, Mass., 2001), 8. 3 The rise of an interest in the Holocaust is charted by Peter Novick, The Holocaust in American Life (Boston, 1999). Its paradigmatic quality is asserted by Zygmunt Bauman, Modernity and the Holocaust (Cambridge, 1989); and Omer Bartov, Mirrors of Destruction: War, Genocide, and Modern Identity (New York, 2000). On the issue of uniqueness, see from among a huge literature G. Rosenfeld, "The Politics of Uniqueness: Reflections on the Recent Polemical Turn in Holocaust and Genocide Scholarship," Holocaust and Genocide Studies 13 (Spring 1999): 28–62; and Alan S. Rosenbaum, ed., Is the Holocaust Unique? Perspectives in Comparative Genocide (Boulder, Colo., 1996). Compare the judicious comments by Maier, "Consigning the Twentieth Century to History," 812, 826–29. It is striking, to take one example of the reluctance of historians of the Third Reich to contextualize their subject through comparative work, that a recent collection of essays on Nazi violence ("La violence nazie," special issue of Revue d'histoire moderne et contemporaine 47 [April–June 2000]) fails to make a single comparison with other regimes or episodes. Similarly exclusive is a volume to which I am a contributor: Hannes Heer and Klaus Naumann, eds., The German Army and Genocide: Crimes against War Prisoners, Jews and other Civilians in the East, 1939–1944 (New York, 1999). For a very recent example of a work on atrocities committed by the German army in World War I, showing the path to a more contextualized approach to this subject, see John N. Horne, German Atrocities, 1914: A History of Denial (New Haven, Conn., 2001). The pioneering dissertation by Richard Cavell Fattig, on the historical evolution of German military attitudes toward civilians, remains unpublished: Fattig, "Reprisal: The German Army and the Execution of Hostages during the Second World War" (PhD dissertation, University of California, San Diego, 1980). 4 Symptomatic of the continuing power of the Holocaust as a historical category is the polemical extension of the term in works such as Horst Möller, Der rote Holocaust und die Deutschen: Die Debatte um das "Schwarzbuch des Kommunismus" (Munich, 1999), and the tendency of new scholarship on the USSR (as is discussed below) to "ethnicize" if not "racialize" the character of the Stalinist repression in the 1930s and 1940s. Compare too Patrick Raszelenberg, "The Khmers Rouges and the Final Solution," History and Memory 11 (Fall–Winter 1999): 62–93. 5 Yves Ternon, Les Arméniens: Histoire d'un génocide (Paris, 1977); Vahakn N. Dadrian, The History of the Armenian Genocide: Ethnic Conflict from the Balkans to Anatolia to the Caucasus (Providence, R.I., 1995); on the Teskilât, there is still only Philip Stoddard, "The Ottoman Government and the Arabs, 1911 to 1918: A Preliminary Study of the Teskilât-i Mahsusa" (PhD dissertation, Princeton University, 1963); and two articles by Dadrian, "The Role of the Special Organisation in the Armenian Genocide during the First World War," in Panikos Panayi, ed., Minorities in Wartime: National and Racial Groupings in Europe, North America, and Australia during the Two World Wars (London, 1993), 50–83; and "The Secret Young Turk-Ittihadist Conference and the Decision for the World War One Genocide of the Armenians," Holocaust and Genocide Studies 7 (Fall 1993). Richard G. Hovannisian, ed., Remembrance and Denial: The Case of the Armenian Genocide (Detroit, Mich., 1999), gives little attention to the broader Ottoman context. Useful from this point of view is Manoug Joseph Somakian, Empires in Conflict: Armenia and the Great Powers, 1895–1920 (London, 1995); as well as the still unsurpassed W. E. D. Allen and Paul Muratoff, Caucasian Battlefields (Cambridge, 1953); for background, see Feroz Ahmed, "Unionist Relations with the Greek, Armenian and Jewish Communities of the Ottoman Empire, 1908–1914," in Benjamin Braude and Bernard Lewis, eds., Christians and Jews in the Ottoman Empire, vol. 1 (New York, 1982), 387–434. 6 Howard Adelman and Astri Suhrke, eds., The Path of a Genocide: The Rwanda Crisis from Uganda to Zaire (New Brunswick, N.J., 1999); Philip Gourevitch, We Wish to Inform You That Tomorrow We Will Be Killed with Our Families: Stories from Rwanda (New York, 1998); Gérard Prunier, The Rwanda Crisis: History of a Genocide (1995; New York, 1997); Mahmood Mamdani, When Victims Become Killers: Colonialism, Nativism, and the Genocide in Rwanda (Princeton, N.J., 2001). For insight into French policy toward Rwanda during the genocide, see Gérard Prunier, "Operation Turquoise: A Humanitarian Escape from a Political Deadend," in Adelman and Suhrke, Path of a Genocide, 281–307. 7 On the Yanomami, see Survival International, Disinherited: Indians in Brazil (London, 2000), 49–53. Mark Levene, "Why Is the Twentieth Century the Century of Genocide?" Journal of World History 11 (Fall 2000): 305–36. 8 Andrew Bell-Fialkoff, "A Brief History of Ethnic Cleansing," Foreign Affairs 72 (Summer 1993); and Bell-Fialkoff, Ethnic Cleansing (New York, 1996); Naimark, Fires of Hatred; J. Otto Pohl, Ethnic Cleansing in the USSR, 1937–1949 (New York, 1999). Robert M. Hayden, "Schindler's Fate: Genocide, Ethnic Cleansing, and Population Transfers," Slavic Review 55 (Winter 1996): 727–48, and subsequent replies, is as much about the hypocrisy of international attitudes toward the expulsion of the ethnic Germans as it is about the expulsions themselves. For useful cautionary remarks, see Rogers Brubaker and David D. Laitin, "Ethnic and Nationalist Violence," Annual Review of Sociology 24 (1998): 423–52. 9 Justin McCarthy, Death and Exile: The Ethnic Cleansing of Ottoman Muslims, 1821–1922 (Princeton, N.J., 1995); less polemically, Alexandre Toumarkine, Les migrations des populations musulmanes balkaniques en Anatolie (1876–1913) (Istanbul, 1995); B. Glyn Williams, "Hijra and Forced Migration from 19th Century Russia to the Ottoman Empire: A Critical Analysis of the Great Crimean Tatar Emigration of 1860–1861," Cahiers du monde russe 41 (January–March 2000): 79–108; W. Brooks, "Russia's Conquest and Pacification of the Caucasus: Relocation Becomes a Pogrom in the Post-Crimean War Period," Nationalities Papers 23 (1995): 675–86; Stephen D. Shenfield, "The Circassians: A Forgotten Genocide?" in Mark Levene and Penny Roberts, eds., The Massacre in History (New York, 1999), 149–63; M. Pinson, "Ottoman Colonisation of the Circassians in Rumili," Etudes balkaniques 9 (1973). Compare A. Derslid, "Imperial Russification," in John Morison, ed., Ethnic and National Issues in Russian and East European History (London, 2000), chap. 3. 10 Cornelia Sorabji, "A Very Modern War: War and Territory in Bosnia-Hercegovina," in Robert A. Hinde and Helen Watson, eds., War: A Cruel Necessity? The Bases of Institutionalised Violence (London, 1995), 80–99; Jan Tomasz Gross, Neighbors: The Destruction of the Jewish Community in Jedwabne, Poland (Princeton, N.J., 2001); Tim Snyder, "'To Resolve the Ukrainian Problem Once and for All': The Ethnic Cleansing of Ukrainians in Poland, 1943–1947," Journal of Cold War Studies 1 (Spring 1999); and T. Piotrowski, "Akcja 'Wisla': Operation 'Vistula,' 1947: Background and Assessment," Polish Review 43, no. 2 (1998): 219–38. 11 I return below to the relationship between colonial empire and Europe in this context. Dirk Moses of the University of Sydney is working on frontier violence in settler societies. For the Americas, see David E. Stannard, American Holocaust: The Conquest of the New World (New York, 1992). For an example from Africa, see Arthur Keppel-Jones, Rhodes and Rhodesia: The White Conquest of Zimbabwe, 1884–1902 (Kingston, Ont., 1983), my thanks to Mark Levene for this reference. 12 On the idea of expulsion, see Nur Masalha, Expulsion of the Palestinians: The Concept of "Transfer" in Zionist Political Thought, 1882–1948 (Washington, D.C., 1992); Anita Shapira, Land and Power: The Zionist Resort to Force, 1881–1948 (Stanford, Calif., 1992); Eugene L. Rogan and Avi Shlaim, eds., The War for Palestine: Rewriting the History of 1948 (Cambridge, 2001); Benny Morris, The Birth of the Palestinian Refugee Problem (Cambridge, 1987); and Morris, "Operation Dani and the Palestinian Exodus from Lydda and Ramle in 1948," Middle Eastern Journal 40, no. 1 (1986): 82–109. 13 S. Aiyar, "'August Anarchy': The Partition Massacres in Punjab, 1947," South Asia 28 (1995): 13–37; Ian Copland, "The Further Shores of Partition: Ethnic Cleansing in Rajasthan, 1947," Past and Present, no. 160 (1998): 203–39; David Gilmartin, "Partition, Pakistan and South Asian History: In Search of a Narrative," Journal of Asian Studies 57, no. 4 (1998): 1068–95. 14 Joya Chatterji, Bengal Divided: Hindu Communalism and Partition, 1932–1947 (Cambridge, 1994), 232; Veena Das, ed., Mirrors of Violence: Communities, Riots and Survivors in South Asia (Delhi, 1990), 25; Urvashi Butalia, The Other Side of Silence: Voices from the Partition of India (London, 2000), 58–59; Tambiah, Leveling Crowds; see also Sudhir Kakar, The Colors of Violence: Cultural Identities, Religion, and Conflict (Chicago, 1996), for an insightful study of the psychology of the "strong men" at the heart of these bouts of violence. 15 Peter Holquist, "'Information Is the Alpha and Omega of Our Work': Bolshevik Surveillance in Its Pan-European Context," Journal of Modern History 69 (September 1997): 415–50; Amir Weiner, Making Sense of War: The Second World War and the Fate of the Bolshevik Revolution (Princeton, N.J., 2001). More generally, see Martin E. Malia, The Soviet Tragedy: A History of Socialism in Russia, 1917–1991 (New York, 1994); and James C. Scott, Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed (New Haven, Conn., 1998). 16 See Eric Lohr, "Enemy Alien Policies within the Russian Empire during World War One" (PhD dissertation, Harvard University, 1999); Terry Martin, "The Origins of Soviet Ethnic Cleansing," Journal of Modern History 70 (December 1998): 813–61; Michael Gelb, "The Western Finnic Minorities and the Origins of the Stalinist Nationalities Deportations," Nationalities Papers 24 (June 1996): 237–68; Gelb, "An Early Soviet Ethnic Deportation: The Far-Eastern Koreans," Russian Review 54 (July 1995): 389–412; Gelb, "Ethnicity during the Ezhovshchina: A Historiography," in Morison, Ethnic and National Issues in Russian and East European History, 190–99. 17 Pohl, Ethnic Cleansing; Keith Sword, ed., Deportation and Exile: Poles in the Soviet Union, 1939–48 (London, 1994); and Z. Siemeszko, "The Mass Deportations of the Polish Population to the USSR, 1940–1941," in Keith Sword, ed., The Soviet Takeover of the Polish Eastern Provinces, 1939–41 (London, 1991), 217–32. On the Ukrainian famine, and in particular the issue of whether or not it counts as "genocide," see David R. Marples, Stalinism in the Ukraine in the 1940s (London, 1992), 22–23. 18 David Rousset, L'univers concentrationnaire (Paris, 1946). The best comparative treatment of the political prisoner in the twentieth century—a badly neglected topic—is now Polymeris Voglis, Becoming a Subject: Political Prisoners during the Greek Civil War, 1945–1950 (Oxford, 2001), chaps. 1–2. 19 On Italian totalitarianism, see Alexander de Grand, "Cracks in the Facade: The Failure of Fascist Totalitarianism in Italy 1935–9," European History Quarterly 21 (1991); Stephen Wheatcroft, "The Scale and Nature of German and Soviet Repression and Mass Killings, 1940–1945," Europe-Asia Studies 48, no. 8 (1996): 1319–53; Nicholas Werth in Stéphane Courtois, et al., The Black Book of Communism: Crimes, Terror, Repression (Cambridge, Mass., 1999). On a comparison of nineteenth and twentieth-century concepts of treachery, see Voglis, Becoming a Subject, chap.1; and Barton L. Ingraham, Political Crime in Europe: A Comparative Study of France, Germany, and England (Berkeley, Calif., 1979). 20 Jean-Louis Margolin, "China: A Long March into Night," in Courtois, Black Book of Communism, 463–547; Lynn T. White, Policies of Chaos: The Organizational Causes of Violence in China's Cultural Revolution (Princeton, N.J., 1989); Jasper Becker, Hungry Ghosts: China's Secret Famine (London, 1996). 21 Jean-Louis Margolin, "Cambodia: The Country of Disconcerting Crimes," in Courtois, Black Book of Communism, 577–95. 22 See in addition to Margolin, David P. Chandler, The Tragedy of Cambodian History: Politics, War, and Society since 1945 (New Haven, Conn., 1995); Marie Alexandrine Martin, Cambodia, a Shattered Society (Berkeley, Calif., 1994); and Ben Kiernan, The Pol Pot Regime: Race, Power, and Genocide in Cambodia under the Khmer Rouge, 1975–79 (New Haven, 1996). 23 Dirk Krujit, "Exercises in State Terrorism: The Counter-Insurgency Campaigns in Guatemala and Peru," in Kees Koonings and Dirk Krujit, eds., Societies of Fear: The Legacy of Civil War, Violence and Terror in Latin America (London, 1999), 33–63; Timothy P. Wickham-Crowley, "Terror and Guerrilla Warfare in Latin America, 1956–1970," Comparative Studies in Society and History 32, no. 2 (1990): 201–37. 24 Michael McClintock, The American Connection, Vol. 2: State Terror and Popular Resistance in Guatemala (London, 1985); and McClintock, "American Doctrine and Counter-Insurgency State Terror," in Alexander George, ed., Western State Terrorism (Cambridge, 1991). 25 See the very useful essay by Robert Cribb, "The Indonesian Massacres," in Samuel Totten, William S. Parsons, and Israel W. Charny, eds., Century of Genocide: Eyewitness Accounts and Critical Views (New York, 1997), 236–47; and Robert Cribb, ed., The Indonesian Killings of 1965–1966: Studies from Java and Bali (Clayton, Victoria, 1990); Harold A. Crouch, The Army and Politics in Indonesia (Ithaca, N.Y., 1978); Marshall Green, Indonesia: Crisis and Transformation, 1965–1968 (Washington, D.C., 1990), 68. 26 Amos Perlmutter, The Military and Politics in Modern Times: On Professionals, Praetorians, and Revolutionary Soldiers (Ne