The manipulation of ethnicity: from ethnic cooperation to violence and war in Yugoslavia Anthony Oberschall Abstract The story of the failure of bargaining among ethnic elites and of international diplomacy is well known. What has not been well explained is the spread and support for xenophobic nationalism and ethnic violence among people who had lived cooperatively for thirty-five years. I draw on key ideas of four views on ethnicity and ethnic conflict, and add the concept of cognitive frame in ethnic relations. Yugoslavs possessed two ethnic frames in their minds, an ethnic cooperation and peace frame for normal times, and a crisis frame anchored in World War II memories. Elite contention and mass media propaganda awakened the dormant crisis frame, suppressed the normal frame, and spread insecurity and fear. I explain why ethnic manipulation succeeded, people believed falsehoods, voted for nationalists, how moderates were purged and why men in militias killed innocent civilians. Keywords: Yugoslavia; nationalism; ethnic violence; cognitive frame; conflict; polarization. Four views on ethnicity and ethnic violence are common.1 In the 'primordial' view, ethnic attachments and identities are a cultural given and a natural affinity, like kinship sentiments. They have an overpowering emotional and non-rational quality. Applied to the former Yugoslavia, the primordialist (Kaplan 1993) believes that despite seemingly cooperative relations between nationalities in Yugoslavia, mistrust, enmity, even hatred were just below the surface, as had long been true in the Balkans. Triggered by fierce competition for political power during the breakup of Yugoslavia and driven by the uncertainties over state boundaries and minority status, these enmities and hatreds, fuelled by fear and retribution, turned neighbour against neighbour, and district against district, in an expanding spiral of aggression and reprisals. Although the primordial account sounds plausible, and it is true that politicians activated and manipulated latent nationalism and ethnic fears, some evidence contradicts it. Ethnic cleansing was more commonly militias and military Ethnic and Racial Studies Volume 23 Number 6 November 2000 pp. 982—1001 © 2000 Routledge Journals, Taylor & Francis Ltd ISSN 0141-9870 print/1466-4356 online The manipulation of ethnicity 983 against civilians than neighbour against neighbour. In seventeen assaults against villages during the ethnic cleansing of Prijedor district in Bosnia in May/June 1992, we found that the aggressors wore military and paramilitary uniforms and insignia. In fourteen assaults, the survivors did not recognize any of the aggressors, who did not bother to wear masks or disguises. These 'weekend warriors' from central Serbia openly bivouacked at the Prijedor police station (Vulliamy 1994, p. 94). The primordial theory omits the fact that ethnic hatreds can subside as a consequence of statecraft and living together. DeGaulle and Adenauer managed to reconcile the French and German people. Why no lasting conciliation in Yugoslavia after forty years of ethnic peace? In the second, 'instrumentalist' view, ethnic sentiments and loyalties are manipulated by political leaders and intellectuals for political ends, such as state creation (Rosens 1989). For Yugoslavia, the instrumentalist explanation highlights Serb nationalists' goal of a Greater Serbia (Cigar 1995), and a similar Croat nationalism (Djilas 1995). Ethnic cleansing resulted from a historical longing by Serbs for a Greater Serbia, with deep cultural roots. Milosevic and Serb nationalists tried to implement it when the opportunity arose in the late 1980s and early 1990s. Greater Serbia required ethnic cleansing of non-Serbs from areas inhabited by a majority of Serbs and the corridors linking Serb population clusters. Although there is evidence that ethnic cleansing was a state policy, orchestrated by the highest authorities in Serbia and the Bosnian Serb leadership, this explanation ignores that many Bosnian Serbs did not want secession, that many Serbs in Croatia at first backed moderate nationalists, and that many Serbs evaded the draft (Milosevic 1997, p. 109). The instrumentalist view assumes an ethnic consensus that initially does not exist. But if many were reluctant to wage war and to participate in ethnic cleansing, how did ethnic extremists prevail over these moderates? The third 'constructionist' view of ethnicity and ethnic conflict was originally formulated by Kuper (1977). It supplements the insights of the primordial and of the instrumentalist views. Religion or ethnicity are very real social facts, but in ordinary times they are only one of several roles and identities that matter. There is a great deal of variance in a population on ethnic attachments and identities. In the words of Linz and Stepán (1996, p. 366) 'political identities are less primordial and fixed than contingent and changing. They are amenable to being constructed or eroded by political institutions and political choices'. The constructionist view offers insights but is incomplete. How are nationality and ethnicity constructed and eroded by political mobilization and mass media propaganda? A fourth model of ethnic violence (Posen 1993; Gagnon 1997) centres on state breakdown, anarchy, and the security dilemma that such conditions pose to ethnic groups who engage in defensive arming to protect 984 Anthony Oberschall their lives and property against ethnic rivals, which then stimulates arming by other ethnic groups like an arms race between states. The driving motivations are not ethnic hatreds but fear and insecurity. In the Yugoslav crisis Michael Ignatieff (1993a, p. 42) puts it thus: Once the Yugoslav communist state began to split into its constituent national particles the key question soon became: will the local Croat policeman protect me if I am a Serb? Will I keep my job in the soap factory if my new boss is a Serb or a Muslim? The answer to this question was no, because no state remained to enforce the old ethnic bargain. There is a security dilemma in ethnic conflict, but why so much ethnic violence without state breakdown? Can insecurity and fear be spread by propaganda even when daily experience contradicts the allegations of ethnic hostility and threat? Can the powerful fear the weak? Building on the four views and mindful of Brubaker and Laitin's (1998) criteria for a satisfactory theory of ethnic violence, I use the idea of latent nationalism at the grass roots, and show how it was activated; I highlight ethnic manipulation by political leaders, and explain why manipulation was successful; I take into account the variance in ethnic identities and analyse why extremists prevailed over moderates; I focus on the security dilemma and ethnic fears and insecurity, and show how fears and insecurity grew from lies and propaganda. To this arsenal of concepts and models for generating the dynamics of ethnicization and collective violence, I add 'cognitive frames'. Combining all, I seek to explain how forty years of cooperative ethnic relations ended with collective violence and war. Prijedor: a case-study To get a sense of what is to be explained about ethnic conflict and violence at the grass roots, consider the Prijedor district in Northwest Bosnia where major ethnic violence took place in the spring of 1992. In the 1991 Census, Prijedor district was 42.5 per cent Serb and 44 per cent Muslim. It was surrounded by districts that had either a slight Serb majority or were close to even, as Prijedor was. Prijedor Serbs were not an isolated Serb minority island surrounded by a sea of Muslims and Croats. There had been no Serb complaints of mistreatment, discrimination, or intimidation in Prijedor by non-Serbs, or vice versa. On the contrary, as a bewildered Muslim refugee from Prijedor stated, In Prijedor there were no conflicts between nationalities. We didn't make the distinctions. My colleague at work was an Orthodox Serb, The manipulation of ethnicity 985 we worked together. When we were children we went to the Orthodox church or the mosque together ... I don't understand. Before there were never any problems between us. We lived together. My sister is married to a Serb, and a brother of my wife is married to a Croat (Mazowiecki 1993, §24). According to the Bassiouni Report (UN Security Council 1994, Annex 5: Prijedor), Serbs held the leading positions in Prijedor in 1991, as they had done for decades. Among the leading Serbs were the director of posts and telecommunications, the managers of the radio station and the newspaper, the managing directors of the largest employers and state enterprises such as the mining company, top administrative jobs in social and medical services, and important posts in local government and the police. In the 1991 elections, the predominantly Muslim SDA won thirty seats; the Serb SDS twenty-eight, and thirty-two went to other parties.2 The Muslims refrained from taking over a number of leading posts to which their electoral victory entitled them because they believed in power-sharing. Even so, the SDS blocked the work of the Prijodor Assembly and organized a parallel governance for Serbs, in alliance with the SDS leaders in nearby Banja Luka. In Bosnia as a whole, the Serbs shared political power and controlled the most important military forces. As in other towns and cities in Bosnia, the SDS in Prijedor organized a successful Serb plebiscite for Greater Serbia. A parallel Serb governance, called the 'Crisis Committee', secretly created an armed force of Serbs with weapons obtained from Serbia and the JNA (Malcolm 1996, p. 225). Serb crisis committees were also formed among Serbs in some of Prijedor district's towns and villages. On the night of 29 April 1992, without any provocation or a shot being fired, 1,775 well-armed Serbs seized the city of Prijedor in a coup d'etat. By this time the Prijedor local government had completely lost power to various Serb groups. Paramilitaries had seized the radio and television transmitters and cut off all but Serb transmissions. The Serb coup d'etat in Prijedor is similar to what happened elsewhere in Northern Bosnia. Non-Serb leaders were arrested and shortly afterwards disappeared, presumed executed. The Muslim police and other officials were fired from their posts. Schools closed; the newspaper ceased publication, and a Serb paper was started. Non-Serbs were harassed, intimidated, fired from their jobs (Hukanovic 1996, p. 65). Amid incessant house searches, weapons, mostly hunting guns, belonging to non-Serbs, were rounded up. After the attempt on 30 May by the Patriotic League of Croats and Muslims — an armed formation of 150 fighters — to retake the old city, many non-Serb inhabitants were arrested and sent to the infamous Omarska camp (Human Rights Watch 1992, vol. 2, pp. 405ff). At Omarska, prisoners were tortured, brutalized, starved and killed. The guards were rural Serbs from nearby villages; the interrogators were 986 Anthony Oberschall Prijedor police inspectors (UN Security Council 1994, §148). Nearby Kozarak got attacked on 26—28 May by the army and paramilitaries, and 600 townspeople were killed. The attackers looted and burnt down the houses. People were rounded up and some were executed: those shot were Muslim leaders whose names appeared on a list. Atrocities took place elsewhere in the district. Several observations should be made about the events in Prijedor. Muslims and Serbs had lived in peace before the conflict erupted. The Serbs were neither a numerical minority, nor discriminated against. They not only had a share of power, but they had the biggest share, and they were well armed. Why, then, did Serbs fear their fellow citizens in Prijedor? A cartoon from this period expresses the puzzle well. It shows a bearded Serb paramilitary, armed to the teeth, with guns, hand-granades, ammunition belt, knives, waving a machine gun, looking worried, and yelling at the top of his voice, T am being threatened!' (Monnesland 1997, p. 460). There was no anarchy, no state breakdown in Prijedor. The Serbs used the police and military of a functioning government to subdue the non-Serbs. Serbs may have been apprehensive about their future in an independent Bosnia, but even in Bosnia they had a big presence —numerical, military, political, economic. There was no spontaneous violence initiated by Serb civilians against non-Serbs, nor vice-versa. Instead, there was a highly organized, secretly prepared coup d'etat, like the 1917 Bolshevik seizure of power in Russia. The arming of the Serbs in Prijedor was straightforward. Seselj, the head of the largest paramilitary and an extremist rival to Milosevic, boasted in a BBC documentary (Silber and Little 1995, p. 230): 'Milosevic organized everything. We gathered volunteers and he gave us special barracks ... all our uniforms, arms, military technology, and buses'. As in the Russian revolution with the Soviets, the Serb parallel government was not only an instrument for seizing power from non-Serbs but of stripping the moderate Serbs of any influence and authority. What was the reaction of ordinary Serbs to these events? Though there is no information on Prijedor itself, one can learn from what observers recorded in nearby Banja Luka. Peter Maas (1995) reports that a Serb lawyer there estimated that 30 per cent of Serbs oppose such things [ethnic cleansing], 60 per cent agree or are confused and go along with the 10 per cent who 'have the guns and control the television tower'. Such variance in response and the Banja Luka proportions struck the informants I interviewed as a good estimate for their locality as well. An armed, organized 10 per cent who control mass communications can have its way when the majority supports it overtly or tacitly or is confused, and when the opposition is unorganized, divided, and scared. One has to explain how it was that 60 per cent were supportive of or confused on ethnic cleansing, since their support and quiescence were necessary for the success of the extremist 10 per cent. The manipulation of ethnicity 987 Was violent conflict inevitable? In a multinational state such as Yugoslavia, nationality will be a salient dimension of political contention, and there will be leaders and intellectuals with a nationalist ideology and agenda. The Yugoslav constitution and its political institutions were delicately balanced and crafted to deal with nationality. A nationalist challenge would inevitably zero in on stateness, minority rights and power-sharing: if accepted boundaries of political units are renegotiated or remade, who decides which peoples and territories belong to new and old political entities? Will all peoples in the new units be equal citizens for governance, or will majority ethnonational affiliation become the admission ticket for full citizenship? Once unleashed, nationalism in Yugoslavia set on a collision course the two largest nationalities, the Serbs and the Croats. With a quarter of Serbs living outside Serbia, a centralized Yugoslav state was a guarantor of Serb security. For Croats and their history of opposition to Hapsburg rule, a decentralized state and weak federation meant control of their own destinies, unencumbered by inefficient state agencies and enterprises staffed and controlled by Serbs. Nevertheless, nationality issues could have been sorted out with democratic institutions in a confederation, with collective rights for minorities, and with systems of political representation in elections and collective decision rules in assemblies that would protect minority voice and favour coalitions rather than majority domination (Sisk 1996). With these reforms, nationalist leaders would have found it difficult to rally the citizenry to their cause. In a country with great differences in economic development and standards of living between the Republics, there will be disagreements over economic policies, taxation, transfers, subsidies across regions, and abandoning socialism for a market economy. All Republics had experienced dramatic economic gains since World War II. Yugoslavia was not beyond economic repair. As in other communist states in the late 1980s, the Yugoslav communist leaders wanted to remain in power. Some reprogrammed as reform communists, and hoped to move into European-style social democracy. Others chose ethnonationalism as the issue that would carry them to power and create a new principle of legitimacy for the post-communist regime. Moderate nationalists stood for conciliation among nationalities; extremists were willing to pursue their goals with force and violence. The defeat of the moderates was not inevitable. Why did xenophobic nationalism resonate with the citizenry? How is it that when the media unleashed the war of words and symbols before the war of bullets, so many believed the exaggerations, distortions and fabrications that belied their personal experiences? 988 Anthony Oberschall Ethnic relations before the crisis Survey research on ethnic relations in mid-1990 found that in a national sample of 4,232 Yugoslavs, only 7 per cent believed that the country would break up into separate states, and 62 per cent reported that the 'Yugoslav' affiliation was very or quite important for them (Cohen 1993, p. 173). On ethnonational relations, in workplaces, 36 per cent characterized them as 'good', 28 per cent as 'satisfactory', and only 6 per cent said 'bad' and 'very bad'. For ethnonational relations in neighbourhoods, 57 per cent answered 'good', 28 per cent 'satisfactory', and only 12 per cent chose 'bad' and 'very bad' {Yugoslav Survey 1990, p. 25). For the majority of Yugoslavs, on the eve of the Yugoslav wars, nationalist contention in the public arena did not translate into hostile interpersonal ethnic relations. In May/June 1998 I interviewed several former managers of large state enterprises about workplace cooperation among their employees in 1990 and 1991: a chemical plant in Tulsa, a metallurgical factory in Banja Luka, a construction company in Zadar, a national park where hundreds camped, a law department of a large state enterprise, a music school with several hundred pupils in Sarajevo, a high school in Knin, a university faculty in Sarajevo. None reported any hostile incidents, fights, non-cooperation between workers, employees, clients, pupils, campers of different nationalities, right up to the time that war broke out in Croatia and Bosnia. Others report confirming evidence on good relations among nationalities before the wars started. Glenny (1992, p. 19) writes about Knin, later a hotbed of extremist militia activity: 'Before May 1991 Croats and Serbs lived together in relative contentment ... nobody in their wildest fantasy would have predicted that within 12 months ... Croat soldiers would massacre innocent Serbs while Serb fighters would mutilate innocent Croats'. Most ordinary people were stunned by the violence that descended upon them, unexpectedly. Jeri Laber (1993), the human rights activist, talks to refugees in a camp: "How was it before the war?" "Before the war it was super ... my neighbours were Muslims, Croats. We celebrated our holidays together. A few months before war broke out, people started separating. It was after Bosnia's independence was recognized. Our neighbours avoided us." Ignatieff (1993b) is puzzled, 'What is difficult to understand about the Balkan tragedy is how ... nationalist lies ever managed to take root in the soil of shared village existence ... In order for war to occur, nationalists had to convince neighbors and friends that in reality they had been massacring each other since time immemorial.' The manipulation of ethnicity 989 The manipulation of ethnicity For explaining ethnic manipulation one needs the concept of a cognitive frame. A cognitive frame is a mental structure which situates and connects events, people and groups into a meaningful narrative in which the social world that one inhabits makes sense and can be communicated and shared with others (Snow etal. 1986). Yugoslavs experienced ethnic relations through two frames: a normal frame and a crisis frame. People possessed both frames in their minds: in peaceful times the crisis frame was dormant, and in crisis and war the normal frame was suppressed. Both frames were anchored in private and family experiences, in culture and in public life. In the normal frame, which prevailed in Tito's Yugoslavia, ethnic relations were cooperative and neighbourly. Colleagues and workers, schoolmates and teammates transacted routinely across nationality. Some did not even know or bother to know another's nationality. Intermarriage was accepted. Holidays were spent in each others' Republics. Except in Kosovo, the normal frame prevailed for most Yugoslavs throughout the 1980s. The crisis frame was grounded in the experiences and memories of the Balkan wars, the first and second world wars — and other wars before that. In these crises, civilians were not distinguished from combatants. Old people, children, women, priests were not spared. Atrocities, massacres, torture, ethnic cleansing, a scorched-earth policy were the rule. Everyone was held collectively responsible for their nationality and religion, and became a target of revenge and reprisals. A soldier reported on his orders in the Balkan war of 1912, 'We are to burn the villages, massacre the young, and spare none...'; another writes in a letter, 'It is horrible ... there is nothing but corpses, dust and ashes ... Pillage is going on everywhere' (Carnegie Endowment 1914). Change the date to 1943, to 1992, and not a word need be altered. Tito had wanted to eradicate the crisis frame, but it simmered in the memories of older people, the families of victims, intellectuals and religious leaders. Milosevic, Tudjman and other nationalists did not invent the crisis frame; they activated and amplified it. A year before war the Bosnian Serb leader Radovan Karadzic stated, 'The Serbs are endangered again ... this nation well remembers genocide [in World War II]. Those events are still a terrible living memory. The terror has survived 50 years (Sudetic 1998, p. 84)'. The deputy mayor of Prijedor justified the Omarska camp where Croats and Muslims were put to death in 1992 as a payback for the World War II Jasenovac camp where thousands of Serbs, died: 'There is a direct connection ... During world war two the Croats killed us; this time it was the other way round, we killed them. Perhaps in fifty years, it will happen again to us' (Cohen 1998, p. 479). If the normal frame prevailed in the 1980s as shown by the survey 990 Anthony Oberschall findings, how did nationalists activate and amplify the crisis frame after decades of dormancy? The emotion that poisons ethnic relations is fear (Lake and Rothchild 1996): fear of extinction as a group, fear of assimilation, fear of domination by another group, fear for one's life and property, fear of being a victim once more. After fear comes hate. The threatening others are demonized and dehumanized. The means of awakening and spreading such fears in Yugoslavia were through the newsmedia, politics, education, popular culture, literature, history and the arts. The crisis frame in Yugoslavia was first resurrected by Serb intellectuals over the plight of the Kosovo Serbs. Because of a higher Albanian birthrate and higher Serb out-migration, Kosovo changed from 23 per cent Serb in 1971 to 10 per cent in 1989. Serb nationalists alleged that Albanians were threatening Serbs into leaving, and that the police and judiciary were not protecting Serbs against Albanian violence. Charges in the newsmedia of sexual assault and rape by Albanians against Serbs were widely believed by Serbs. The public was fed exaggerations and total fabrications that fit the crisis frame. A news story reported In Pec, on the weekend, three young Serb women were raped by Albanians. One rapist was later arrested, the [Albanian ] judge sentenced him to 15 days in jail ... the same day the same judge sentenced a Serb man to 40 days in jail for singing a patriotic Serb song (Z. Markovic 1998, p. 323). Yet an analysis of violent crime statistics in Kosovo in the 1980s by Serb social scientists found that rates of rape and attempted rape in Kosovo were lower than in Central Serbia, in the Vojvodina, and in all Yugoslavia, and rapes tended to be within, not across, nationality (Popovic et al. 1990). Social science findings were known to few. Mira Markovic (1996), Milosevic's wife and an important politician, called the victimization of Serbs in Kosovo 'feudal terror'; The Serb Academy of Sciences and Arts called it 'genocide'. Fear of extinction was spread with highly inflated figures on the ethnic killings in World War II. Serbs claimed 700,000 casualties in the Jaseno-vac camp alone; a more accurate estimate is 100,000 dead, of whom half were Serbs (Anzulovic 1999, pp. 100—104). Croats made similarly inflated claims for the Bleiburg massacres. Everyone engaged in a war of numbers sought to prove victimhood and justification for a pre-emptive strike. Fears of extinction polarized ethnic relations. Bogdan Denitch (1996, p. 81) recalls: Everyone was traumatized by all the talk of world war two atrocities ... even those who had seemed immune to nationalism. Old personal ties and friendships crumbled as many intellectuals I knew, as well as The manipulation of ethnicity 991 friends and family members, rallied to the defense of their own nation. The pressure to do so was immense. In my interview with a Serb refugee one can trace how the atrocities discourse switched on the crisis frame: 'We were afraid because nationalists revived the memory of World War II atrocities ... nationalist graffiti on walls awakened fears of past memories; it was a sign that minorities [Serbs in Croatia] would not be respected and safe'. Fears of domination, oppression and demographic shrinkage were roused by the incessant rape and genocide discourse. Biljana Plavsic, a professor of biology at Sarajevo University before she became a Bosnian Serb leader, wrote in the newspaper Borba (Cohen 1998, p. 222) '... rape is the war strategy of Muslims and Croats against Serbs. Islam considers this something normal...'. A Belgrade university student I interviewed recalls a Serb writer asserting before a student club in 1990 'Serbia has three enemies: transnational capital, Islam and the Vatican. Their goal is to replace the population of Yugoslavia with Muslims from Arab countries'. Such statements by the political and cultural elites had become commonplace. Dubrana Ugresic (1998) rightly accuses such intellectuals as being the 'great manipulators' who spread a 'culture of lies'. Ordinary people echo the intellectuals' and the media crisis discourse. A young Serb soldier under a map of Europe in his barracks that has huge areas coloured in green [for Islam] tells a reporter, 'the Muslims expelled us from Kosovo with their sexual organs ... they want to do the same here [in Bosnia] ... the way they reproduce they need room. You will soon feel that elsewhere in Europe' (Cohen 1998, p. 222). At the high end of the military hierarchy, General Panic, whose artillery flattened Vukovar, draws a huge green line on the map of Europe, and explains, 'This is the Green Transversal, the attempt to cut Europe with a green line, a Muslim line, down from Bosnia to Kosovo to Albania to Turkey ... it cuts the Christian world in half (Cohen 1998, p. 434). The Prijedor police chief claimed that Muslims would kill all Serbs over three and put all Serb women aged fifteen to twenty-five into harems to breed janissaries (Gutman 1993, p. 113). Peter Maas (1995, p. 113) asks a Serb refugee couple why they had fled their village. Their answer: Muslims planned to take over, a list of names had been drawn up, Serb women were to be assigned to Muslim harems after the men had been killed. They had heard about it on the radio; the Serb military had uncovered the plan. The journalist probes: 'Did any Muslims in the village ever harm you?' They reply, 'Oh no, our relations with the Muslims in the village were always good, they were decent people'. In the minds of the Serb couple, the crisis frame had eclipsed the normal frame. What under peaceful circumstances were totally implausible events — young women become sexual slaves in harems for breeding janissaries; a fifteenth- and 992 Anthony Oberschall sixteenth-century style Turkish/Islamic invasion of Europe — become credible narratives of ethnic annihilation and domination within the crisis frame. Fear and the crisis frame provided opportunities for nationalists to mobilize a huge ethnic constituency, get themselves elected to office, and organize aggressive actions against moderates and other ethnics. With a 'Save the Serb nation' platform, Milosevic ousted moderate reform communists within the Serb Communist Party. Milosevic seized power, gained control of the state broadcasting organizations, and created a broad-based nationalist movement that gave him a populist, post-communist legitimacy. In the summer of 1988, high on the agenda was the 'retaking of Kosovo'. In a dozen cities crowds of 10,000 to 200,000 demanded the resignation of incumbents and their replacement with Milosevic nationalists. At the core of these assemblies were professional demonstrators dressed in folk costumes, carrying placards and banners rich with Serb national and Orthodox religious symbols, for example, the names of Serb kings and saints, and, of course, 'Slobo'. The coverage in the mass media became a vast nationalist learning experience and legitimated the manifestations of Serb nationalism suppressed under Tito (Monnesland 1997, p. 319). Populist nationalism worked. The Vojvodina and Montenegro party leaderships resigned and were replaced by Milosevic loyalists. Abolishing the autonomous provinces of Kosovo and Vojvodina precipitated a constitutional crisis. According to Djilas (1995, p. 82), Kosovo and Vojvodina had their own representatives in the federal, state and party bodies where most of the time they had voted against Serbia. The two provinces also had the right to veto any change in the Serbian constitution. The nationality balance in Yugoslav politics was thus disturbed. Serbia gained control of over half the votes in all federal bodies and institutions. Slovenes and Croats reacted with their own nationalism. There was grass-roots resistance to nationalism and to activation of the crisis frame. A content analysis of news stories in Oslobodjenje for 1990 indicates that municipalities, youth and veterans' organizations, and trade unions repeatedly protested against ethnic polarization and hatreds. The municipal council of Vojnic organized a rally of brotherhood and unity; in Mostar, the Youth Association and a trade union stated that 'there is no place for the ghosts of the past ... we condemn the spread of nationalist hatred'; in Glamoc and Kladanj, the Socialist Association of Working People sent a message to nationalists Tudjman, Draskovic and Seselj that it would not participate in civil war. Veterans' organizations in Novi Pazar, Kladanj, and Banja Luka condemned 'attempts to break up Yugoslavia'. Important as this opposition was, it The manipulation of ethnicity 993 was countered by the spread of populist nationalism. Oslobodjenje in 1990 is full of affirmations of national symbols and identities: the renaming of localities; the reburial of bones of atrocity victims from World War II; nationalist graffiti on churches, mosques, monuments and in cemeteries; fights over flags, ethnic insults, nationalist songs, ethnic vandalism. To many, these were signs that normal times were sliding into crisis, and the authorities had lost control. Mass communications and propaganda research help to explain why ethnic manipulation worked and why the crisis frame eclipsed the normal frame. First, in a classic study on communication and persuasion, Hovland, Janis and Kelley (1963) found that fear arousing appeals, originating in a threat, were powerful and effective in changing opinion and belief. Furthermore, the most important reaction to fear is removing the source of threat, precisely what nationalists were promising to do in Yugoslavia. Second, studies of propaganda routinely find that repetition is the single most effective technique of persuasion. It does not matter how big the lie is, so long as it keeps being repeated (Brown 1963). Third, much of what we know is vicarious knowledge and not based on personal experience. We accept the truths of authorities and experts whom we respect and who have socially recognized positions and titles. Who could really tell or check how many Serbs had been massacred by Ustasha? Fourth, outright falsehoods were common and intentional. According to a media analyst, 'In Serbia and Croatia, TV fabricated and shamelessly circulated war crime stories ... the same victims would be identified on Zagreb screens as Croat, on Belgrade screens as Serb' (Milosevic 1997, p. 119). A broadcast manager in Sarajevo explained to me how 'patriotic journalism' falsified news: when some began slanting the news in favour of their own nationality, others followed because they did not want their nationality to be disadvantaged. With no one to stop unprofessional practices, news reporting and analysis became a competitive spiral of propaganda. Fifth, mass communications studies of the two-step flow of communication (Wright 1959) show that in ordinary circumstances crude propaganda from 'patriotic journalism' is discounted because people are exposed to a variety of broadcast messages and because they check media messages against the beliefs and opinions in their social milieus in interpersonal relationships and conversations. Ethnic crisis politics breaks down the two-step flow. According to several informants, when politics became contentious, it strained friendships across nationality. Either one avoided discussing public affairs and politics with a friend in order to remain friends, or one stopped being friends, and turned for discussion of such matters to a fellow ethnic with whom agreement was likely. In either case, exchange of political views across ethnic boundaries is impoverished. Each group becomes encapsulated; dialogue and understanding cease. 994 Anthony Oberschall In these circumstances lies fitted the crisis frame, and mass media propaganda polarized ethnic relations at the grass roots. For Knin, Glenny (1992) reports: Without a doubt one of the most important actors on the Knin stage which transformed the consciousness of this dozy town was ... Serbia Radio Knin ... the people of Knin were extremely dependent on this radio station ... Radio Knin is an accomplice in the dissemination of falsehood and the perpetuation of divisive myth which has turned one hapless narod [people] against another equally innocent one (Glenny 1992). A Croat-American youth whose family originated from the small island of Olib in Dalmatia and who regularly spent summer holidays there observed the growth of Croat nationalism (Pulsic 1998). Among new manifestations of Croat nationalism in the 1990s were: use of 'Croat' words in preference to local Olibski dialect; the singing of Croat and Ustasha songs in taverns and on roads; a new 'Croatian' way to salute drinking pals; the removal of Tito portraits from bars, restaurants and stores, and their replacement by Croat crest and flag; vandalization of local World War II partisan hero's grave and memorial; the looting and occupation by squatters of Serb summer homes; replacement of village oriented leaders in the Olib government by HDZ members and sympathizers; anonymous threats sent in letters to ex-communists and moderates; rudeness to non-Croats, especially Serb visitors and summer people, who have since stopped coming. Pulsic believes the nationalist surge was media driven, since there were no island events that could account for change. Croat TV, which everyone watched, had audience monopoly; the non-Croat YUTEL programmes were broadcast at 1 a.m. only. 'During three weeks watching television news, I did not observe a single critical opinion of President Tudjman or the HDZ.' Even sports coverage was patriotic. When a Zagreb soccer team lost to a Belgrade team, none of the goals scored by the Serbs was shown in the sports summary. Nationalists win the 1990 elections Second only to the mass media wars for the revival of the crisis frame were the 1990 elections. Every town and city experienced the founding of political parties, often at a huge rally in a public building or a sports stadium, during which speaker after speaker gave vent to exaggerated nationalist rhetoric and hostile pronouncements and attacks against other nationalities. Drawing on the Oslobodjenje content analysis, one finds that in Bosanska Gradiska the local HDZ handed out flags that read 'Serbs are swine, Serbs should leave', whereas three days later in The manipulation of ethnicity 995 Novi Pazar, Vuk Draskovic, the Serb novelist-politician, proclaimed at an election rally 'All those who like Turkey [Muslims] should go to Turkey'. Such discourse and accompanying election brawls created insecurity and fears about social stability. In the election campaigns, bigotry, hatred and misinformation about nationality circulated freely and reached millions of people nightly on television. The introduction of a multiparty system in communist Yugoslavia was flawed (International Commission on the Balkans 1996, pp. 26—33). In the elections of 1990, citizens of the various republics had no chance to vote for maintaining the federation, despite the fact that most wanted it to continue. There was a proliferation of political parties, seventeen in Slovenia, thirty-three in Croatia, forty-one in Bosnia. The voters lacked information; political parties and leaders had limited time and resources to organize and campaign. According to Woodward (1995), campaigns centred on 'symbols and personalities': In a world of competing symbols and personalities, at a point of political transition, nationalism has a particular advantage. The message is simple, relies on the familiar, takes little resources, does not have to develop a new political language and explain the complexities of democratic institutions and market economy ... nationalist appeals thus provide the easiest route to politics for politicians without established constituencies and party organization (Woodward 1995, p. 124). Nationalists persuaded voters not to 'split the ethnic vote' but to vote as a bloc for the nationalists because the other nationalities would bloc-vote and gain power. Bloc-voting became a self-fulfilling prophecy. Strategic voting produces perverse effects (Schelling 1978). The politicians elected were more nationalist than their voters. Even so, the most nationalist parties did not win large majorities of the vote. In Croatia, the HDZ won 41.5 per cent of the popular vote, but that translated into 58 per cent of seats in the legislature under the winner-take-all system of representation. In Serbia, Milosevic and the SPS got 47 per cent of the eligible voters, 65 per cent of those voting, and 78 per cent of the seats (Woodward 1995). Once elected, the nationalists moved rapidly to consolidate their power. Repression of minorities and moderates The demise of the moderates was due to a combination of electoral defeats, loss of credibility about being effective in a crisis, and intimidation and threats from extremists. Milar Pupovac, a Serb party leader in Croatia who stood for Serb-Croat dialogue and conciliation, was vilified and threatened by both Croat and Serb nationalists (Stitkovac 1997, p. 159). In Osijek, a Serb economist who was also the vice-mayor and 996 Anthony Oberschall had been elected as an independent moderate to the Zagreb legislature in 1990, described to me the rise of extremism and ethnic violence. 'It was safer to take sides than being for peace in the middle', as her own case testifies. Former students avoided her. One later told her, 'It was dangerous to be seen talking to you in public'. While she was attending the legislature in Zagreb, her house was bombed. When she asked the mayor for police protection, he told her she would get none. When she herself was targeted by a sniper, she joined other refugees who left Osijek. The nationalist winners purged their ethnic opponents and moderates of their own nationality from party and state positions. The targets were sent anonymous threat letters, were fired from their jobs, forced into military service, charged with treason, subversion and plotting armed rebellion, and subject to office and house searches for weapons, radio transmitters and 'subversive' literature. In Banja Luka a 45-year old Croat I interviewed was dismissed from a state position, detained and immediately 'assigned' to military labour service, which meant digging trenches, building fortifications and removing mines for Serbs while being exposed to cross-fire from both armies. I asked how moderate Serbs had reacted when non-Serbs were fired from their positions, evicted from their homes, and the mosques were dynamited. 'They [moderate Serbs] were scared; even death threats were not uncommon,' he replied. A moderate Serb in Banja Luka who refused to join the SDS and to serve in the army confirmed it: he was dismissed from his faculty position, and was totally marginalized. Former friends crossed the street when they saw him approaching. It was the same everywhere. In a Bosnian example reported by Mazowiecki (1992, §8—11), 'According to a witness [from Bosanska Dubica], the elected authorities who were moderates and who tried to prevent acts of violence were dismissed or replaced by Serbian extremists'. Other methods were cruder, such as the murder of Josip Rheil-Kir, the regional Croat police chief of the Slavonian part of Croatia, a moderate who had negotiated cease-fires between Serbs and Croats, and who was gunned down by an HDZ extremist official. Ordinary people could not escape ethnic polarization. In an interview a Serb taxi driver explained: 'No one wanted the coming war, but if I don't fight, someone from my side [Serb] will kill me, and if my Muslim friends don't fight, other Muslims will kill them'. The overthrow of moderates by extremists or radicals is well known in the great revolutions: Girondins were overthrown by the Jacobins in the French revolution and all groups were overthrown by the Bolsheviks in the Russian revolution.3 The means of seizing power are similar. The radicals create parallel governance to the state and come to exercise de facto authority in many institutions, and militias and mutineers execute a coup d'etat. Then the remaining moderates are purged (Brinton 1957). The manipulation of ethnicity 997 It happens in ethnic violence as well. It did so in the mixed ethnic districts of Croatia and Bosnia, and it happened in Prijedor. Militias take over Militias and paramilitaries roamed far and wide and perpetrated ethnic cleansing, massacres, atrocities and other war crimes, as in the Prijedor district. One recruitment pool for militias proved to be soccer hooligans. The largest and most criminal militia, Arkan's Tigers, was founded in May 1990 from the Red Star fan club in Belgrade. The fans popularized Serb symbols, flags, colours and chants during games and rallies, and engaged in violent clashes with other fan clubs. One of their chants parallels the political programme of the extreme Serb nationalists: 'We are the fans from proud Serbia, Slobo [Milosevic] you Serb, Serbia stands behind you; come out on the benches, greet the Serbian race; from Kosovo to Knin, Serb is united with Serb' (Colovic 1998). The Tigers set up a military training camp visited by Red Star soccer stars. Other recruitment occurred in prisons, jails and secondary schools. Militiamen were not necessarily fanatics filled with hatred to start with. Judah (1997) described how a Serb militiaman got recruited by his peers from the local SDS who pressured him for weeks: 'We've all got to take up arms, or we'll disappear from here'. He had Muslim and Croat friends. Would they protect him against extremists of all nationalities? Not likely, if it got violent. So he 'took out a gun'. Peer pressure, fear, not only of Muslims but of extremist Serbs who might finger him as a 'traitor', were the major reasons for joining a militia. Some of these men were unemployed and expected a job in the coming Serb government as militia or police. Once the young man 'took out a gun' he became encapsulated in a quasi-military unit subject to peer solidarity and ethnic loyalty. He was trained in weapons and indoctrinated with the beliefs and norms of the crisis frame about other ethnics (Block 1993): a. Collective guilt: 'They' act in unison; children grow into adults; women give birth to future warriors; even old people stab you from behind; 'they' will never change. b. Revenge and retaliation: 'They' massacred 'us' in the past, and are about to do it again, in fact they have already started. A settling of scores is justified; an eye for an eye. c. Deterrence/first strike: Disable them before they strike, which is what they are about to do, despite appearances, because they are secretive and treacherous. d. Danger/survival: These are extraordinary times, one's entire nationality is threatened, and extreme measures are justified. e. Legitimacy: Ordinary people and militias are justified in taking 998 Anthony Oberschall extreme measures because the constituted authorities have not come to the defence of our people. These are the rationalizations and the justifying norms for unrestrained, collective, ethnic violence. Other motives for collective violence were economic gain, peer pressure and lack of accountability. From being an ordinary man in normal times the militiaman changed into being a killer at crisis times. The Bassiouni report (UN Security Council 1994) counted eighty-three paramilitaries in Bosnia alone operating between June 1991 and late 1993, fifty-three for Serbs, with an estimated 20,000^0,000 members, thirteen for Croats, with 12,000—20,000, and fourteen for Bosniac, with 4,000-6,000 men. In view of 700,000 Bosnian Serb men aged fifteen to thirty-five, militiamen were 10—20 per cent of the Serb men of military age in Bosnia. Ten to 20 per cent of adult males in militias, added to the military and police, are more than enough for death and destruction against civilians on a massive scale. Conclusion My account is not a narrative of events but an analytic explanation for the breakup of Yugoslavia amid collective violence. No explanation of such a complex event is complete and comprehensive; mine is no exception. Key events have been narrated as failed negotiations among political leaders and failed diplomacy. Yet what gave leaders credibility and power is that they were elected and had a visible following. How did they accomplish that? On the eve of the wars, Yugoslavs reported cooperative interpersonal ethnic relations and opposed a breakup of the state. Nationalist leaders succeeded in manipulating ethnicity by spreading fear, insecurity and hatred, which advanced their political agenda of separate national states. To explain their success I draw on elements from the primordialist, instrumentalist and constructionist views on ethnicity and on the theory of ethnic violence originating in fear and insecurity. To these I add the concept of a cognitive frame which clarifies élite-grass-roots linkage and ethnic manipulation. Nationalism, ethnic identity and attachment alone, however intense, do not explain grass-roots ethnic actions. Yugoslavs possessed two frames on ethnic relations: a cooperative frame for normal, peaceful times, as in the decades of the fifties to the eighties. They also possessed a dormant crisis frame anchored in family history and collective memory of wars, ethnic atrocities and brutality. Threats and lies that were implausible and dismissed in the normal frame could resonate when the crisis frame was switched on: they became persuasive, were believed, and inspired fear. In the waning days of Communism, nationalists activated the crisis frame on ethnicity by playing on fears of ethnic annihilation and The manipulation of ethnicity 999 oppression in the mass media, in popular culture, in social movements, and in election campaigns. Elite crisis discourse resonated at the grass roots, made for ethnic polarization, and got nationalists elected. Once in office, nationalists suppressed and purged both moderates in their own ethnic group and other ethnics. They organized militias who perpetrated acts of extreme violence against innocent civilians. They conducted war according to the crisis script. Without the tacit, overt or confused support of the majority, the nationalist leaders could not have escalated ethnic rivalry and conflict into massive collective violence. Acknowledgements This research was supported by a grant from the United States Institute of Peace. In addition to published sources and documents, the findings are based on a content analysis of news stories in Oslobodjenje for 1990, and on some fifty interviews with informants interviewed in Serbia, Bosnia and Croatia in May/June 1998. I thank Dr. Ivana Vuletic and students in my collective action course for research assistance with the content analysis. I also thank the scholars who commented on my presentations on this topic at conferences, and the interviewees in the former Yugoslavia. The views expressed in this work are my responsibility alone. Notes 1. I use the words and concepts ethnic, ethnonational and nationality interchangeably to refer to the Serbs, the Croats, the Albanians, the Muslims, and other similar large peoples in Yugoslavia, selecting one or another term to conform to English usage, for example, one writes 'ethnic cleansing' and not 'nationality cleansing'. 2. The following acronyms are used:HDZ — Croat Democratic Union, Croat nationalist party led by Tudjman JNA —Yugoslav National Army SDA —Party of Democratic Union in Bosnia led by Izetbegovic SDS —Serbian Democratic Party, Serb nationalists in Bosnia and Croatia SPS —Serbian Socialist Party, Serb nationalist party led by Milosevic 3. The overthrow of moderates by extremists has also been observed in ethno-national conflicts. It happened in Rwanda before the genocide when Hutu extremists targeted and attacked Hutu moderates (Prunier 1995). Kuper (1977) provides examples from other ethno-national conflicts. References ANZULOVIC, BRANIMIR 1999 Heavenly Serbia, New York: New York University Press BLOCK, ROBERT 1993 'Killers', The New York Review of Books, 21 October BRINTON, CRANE 1957 The Anatomy of Revolution, New York: Doubleday BROWN, J. A. L. 1963 Techniques of Persuasion, London: Penguin Books BRUBAKER, ROGERS and LAITIN, DAVID 1998 'Ethnic and nationalist violence' Annual Review of Sociology, vol. 24, pp. 423—52 1000 Anthony Oberschall CARNEGIE ENDOWMENT FOR INTERNATIONAL PEACE 1914 Report of the Commission to Inquire into the Cause and Conduct of the Balkan Wars, Washington DC CIGAR, NORMAN 1995 Genocide in Bosnia, College Station, TX: A&M University Press COHEN, LEONARD 1993 Broken Bonds: the Disintegration of Yugoslavia, Boulder, CO: Westview COHEN, ROGER 1998 Hearts Grown Brutal, New York: Norton COLOVIC, IVAN 1998 'Fussbal, Hooligans und Krieg', in Thomas Brenner et al., Serbiens Weg in den Krieg, Berlin: Berlin Verlag DENITCH, BOGDAN 1996 Ethnic Nationalism, Minneapolis, MN: University of Minnesota Press DJILAS, ALEKSA 1995 'Fear thy neighbor: the breakup of Yugoslavia', in Charles Kupchan (ed.), Nationalism and Nationalities in the New Europe, Ithaca, NY: Cornell GAGNON, V P. 1997 'Ethnic nationalism and international conflict: the case of Serbia', in Michael Brown et al., Nationalism and Ethnic Conflict, Cambridge, MA: MIT Press GLENNY, MISHA 1992 The Fall of Yugoslavia: The Third Balkan War, London: Penguin Books GUTMAN, ROY 1993 A Witness to Genocide, London: Macmillan HUKANOVIC, REZAK 1996 The Tenth Circle of Hell, New York: Basic Books HUMAN RIGHTS WATCH 1992 War Crimes in Bosnia-Hercegovina, New York IGNATIEFF, MICHAEL 1993a Blood and Belonging: Journey into the New Nationalism, New York: Farrar Strauss ------1993b 'The Balkan tragedy' New York Review of Books, 13 May JUDAH, TIM 1997 The Serbs, New Haven, CT Yale University Press INTERNATIONAL COMMISSION ON THE BALKANS 1996 Unfinished Peace, Washington, DC: Carnegie Endowment for International Peace KAPLAN, ROBERT 1993 Balkan Ghosts, New York: St Martin's Press KUPER, LEO 1977 The Pity of it All: Polarization in Racial and Ethnic Relations, Minneapolis, MN: University of Minnesota Press LABER, JERI 1993 'Bosnia, questions about rape', The New York Review of Books, March LAKE, DAVID and ROTHCHILD, DONALD 1996 'Containing fear: the origin and management of ethnic conflict', International Security, vol. 21, pp. 41—75 LINZ, JUAN J. and STEPÁN, ALFRED 1996 Problems of Democratic Consolidation, Baltimore, MD: Johns Hopkins University Press MAAS, PETER 1995 Love the Neighbor: A Story of War, New York: Knopf MALCOM, NOEL 1996 Bosnia: A Short History, New York: New York University Press MARKO VIC, MIRA 1996 Answers, Kingston, Canada: Quarry Press MARKOVIC, ZORAN 1998 'Die Nation, Opfer und Rache', in Thomas Bremer et al, Serbiens Weg in den Krieg, Berlin: Berlin Verlag MAZOWIECKI, TADEUSZ 1992 Report on the Situation of Human Rights in the Territory of the Former Yugoslavia, United Nations: Economic and Social Council ------1993 Seventh Periodic Report, 10 June MILOSEVIC, MILAN 1997 'The media wars: 1987-1997', in Jasmina Udovicki and James Ridgeway (eds), Burn This House Down: the Making and Unmaking of Yugoslavia, Durham: Duke University Press MONNESLAND, SVENN 1997 Land Ohne Wiederkehr: Ex Jugoslavien, die Wurzeln des Krieges, Klagenfurt: Wieser POPOVIC, SRDJA et al. 1990 Kosovski Cvor: Dresiti ili Seci [Kosovo Knot: Untie it or Cut it] Belgrade: Bibliotéka Kronos POSEN, BARRY 1993 'The security dilemma and ethnic conflict', in Michael Brown (ed.), Ethnic Conflict and Internal Security, Princeton, NJ: Princeton University Press PRUNIER, G. 1995 The Rwanda Crisis, New York: Columbia University Press PULSIC, WESLEY 1998 'Olib Nationalism', Honours thesis, Sociology Department, University of North Carolina, Chapel Hill The manipulation of ethnicity 1001 ROSENS, E. 1989 Creating Ethnicity, London: Sage SCHELLING, THOMAS 1978 Micromotives and Macrobehavior, New York: Norton SILBER, LAURA and LITTLE, ALLAN 1995 Yugoslavia: Death of a Nation, New York: TV Books SISK, TIMOTHY 1996 Power Sharing and International Mediation in Ethnic Conflicts, Washington, DC: US Institute of Peace SNOW, DAVID et al. 1986 'Frame alignment processes, micromobilization and movement participation', American Sociological Review, vol. 51, no. 4, pp. 464-81 STITLOVAC, EJUB 1997 'Croatia: the first war', in Jasmina Udovicki and James Ridgeway, (eds), Burn This House Down: the Making and Unmaking of Yugoslavia, Durham: Duke University Press SUDETIC, CHUCK 1998 Blood and Vengeance, New York: Norton UGRESIC, DUBRANKA 1998 The Culture of Lies, University Park, PA: Pennsylvania State University Press UN SECURITY COUNCIL 1994 Final Report of the Commission of Experts [Bassiouni Report], New York, UN, S/1994/674 VULLIAMY, ED 1994 Seasons in Hell, London: St Martin's Press WOODWARD, SUSAN 1995 Balkan Tragedy, Washington DC: Brookings WRIGHT, W. 1959 Mass Communication, New York: Random House YUGOSLAV SURVEY 1990 'Public Opinion Survey on the Federal Executive Council's Social and Economic Reform', 31 March, pp. 3—26 ANTHONY OBERSCHALL is Professor of Sociology at the University of North Carolina at Chapel Hill. ADDRESS: Department of Sociology, University of North Carolina at Chapel Hill, 212 Hamilton Hall, #3210, Chapel Hill, NC 27599-3210, USA. email: tonob@email.unc.edu