Examining ethnic violence and partition in Bosnia-Herzegovina

Abstract

A number of scholars have argued that the separation of ethnic groups is the best way to resolve ‘ethnic wars’, or wars in which the primary cleavage is coded as ethnic. This partitionist view identifies ethnic mixing as the root of conflict, and partitionists thus prescribe policies that will ‘unmix’ ethnic groups—the drawing of new borders and potential population transfers. This paper argues that the partitionist position misidentifies ethnicity and demographics as the cause of conflicts. Using examples from the war in Bosnia from 1992 to 1995, I demonstrate how a more dynamic approach to examining conflict can better illustrate the complex and contingent emergence of the boundary activation mechanism that produces violence. I conclude with an argument that policy prescriptions should incorporate dynamism into their structures. The Dayton Agreement's durability nearly 10 years after the war's end demonstrates the merits of flexible policies, as opposed to the draconian measures prescribed by the partitionists.

Acknowledgements

I am grateful to Fiona Adamson, Stephen Deets, Paul Dixon, Stacie Goddard and anonymous reviewers for insightful comments on an earlier draft of this paper, as well as to Fotini Christia and Lara Nettelfield for helpful points.

Notes

1. Scholars disagree on whether ethnic civil wars differ significantly from other civil wars. Some argue that civil conflicts defined in terms of ethnicity can make them more difficult to resolve than other types of civil conflict. See Licklider (1995), Sambanis (2001) and Mason & Fett (1996).

2. Sambanis (2000). This is particularly the case for ‘ethnic wars’, pp. 469–471.

3. Downs (2001) outlines a variant that does not explicitly invoke the security dilemma, but which manifests these symptoms.

4. See Kumar (1997a) for a good overview of both moral and practical critiques.

5. A nation is a community of individuals who intersubjectively share this identity. See Anderson (1983) and Connor (1994).

6. The simplifying assumption in international relations that states are unitary actors perpetuates this illusion.

7. Some partitionists argue that demonstration effects do not exist. This assertion goes against vast evidence of copycat strategies by minority groups, particularly when the same ethnic community is located across different states. It is no accident that Hungarians in a number of countries have outlined autonomy proposals, and that extremist Albanians in Macedonia instigated some attacks in 2001 after events in Kosovo.

8. In Petersen's terms, different communities have different equilibria for mobilization.

9. I am grateful to Jack Snyder for this point.

10. Nor are heterogeneous communities significantly more likely to engage in general violence; regime type and socioeconomic factors are more correlated with violence than heterogeneity. See Ellingsen (2000).

11. Whether Tito himself was of mixed Slovene–Croat heritage (Jelavich) or whether he had a singular Croat identity (Kaufmann) is a discrepancy that illustrates how identities may be selectively identified for different purposes. See Jelavich (1995, p. 269), Kaufmann (1996a, p. 142).

12. Stroschein (in progress) examines this dynamic in detail.

13. See, for example, King, Keohane & Verba (1994).

14. For a discussion of time and duration, see Pierson (2004; 2003).

15. While subsequent surveys may reveal similar responses, the act of taking an initial survey establishes the conditions under which those responses are given according to certain categories, making them seem more durable. Anderson (1983, ch. 10).

16. Paraphrased from Tilly (2003).

17. The ability or desire of individuals to make ‘rational’ choices thus varies according to these conditions.

18. Although Roger Petersen, Charles Tilly, and Ashutosh Varshney have taken on such studies.

19. Following common usage, the term ‘Bosniac’ is used to refer to the Bosnian Muslims.

20. Kaufmann's emphasis on mixing leads him to interpret even Israeli–Palestinian conflict over Jewish settlements to be the result of ‘remixing’ rather than of attempts to mark and claim territory. Kaufmann (1998, p. 147).

21. The international community attempted to derail these efforts.

22. There is some distinction between Croat and Muslim sections of the Federation, but this is far less visible than the distinction between the RS and the Federation.

23. At the republic level, ethnic tolerance ran high in heterogeneous republics such as Bosnia and Vojvodina, with low values in homogeneous Kosovo.

24. Overall, minority groups tended to be more tolerant than majorities, consistent with sociological theory on demographic interactions. See Blau (1977). Kaufmann (1996a, pp. 148–149), in contrast, argues that enclaves are preferable to more interspersed populations.

25. H. D. Forbes (1997) describes this contrast as an aggregation problem, the result of different views of power and competition at individual and group levels.

26. In some cases this mechanism may have been more informal—Susan Woodward (1995, esp. pp. 41–45) describes a tendency for traditional loyalties to be used as a criterion for allocating jobs in economically deprived regions.

27. The Yugoslav People's Army became more homogeneously Serb and Montenegrin throughout the conflict, as a number of Slovenes, Croats, Bosniacs and Macedonians began to leave the forces.

28. A similar dynamic was observed with relation to violence against Jews in Europe during WWII.

29. Mystery witness faces Milosevic at Hague trial, New York Times, 24 April 2003.

30. See also the OSCE map representing population distribution in 1996, p. 121.

31. Each entity has its own legislature, but the RS legislature is unicameral and the Federation's is bicameral. There are also seats for the other groups in the Republic parliaments; a Croat and a Muslim hold seats in the RS parliament. Bieber (2002, pp. 212–213).

32. But the SRS was banned from the 2000 elections by the OHR. Keane (2002, p. 98).

33. There are interesting variations in local voting for non-nationalist parties. Individuals in mixed areas have been shown to be more likely to cast votes for non-nationalist parties, particularly when the demographic mix includes Bosniacs. In addition, areas with strong population density and the western zone of the RS have a higher propensity to vote for non-nationalist parties. Pugh & Cobble (2001). Tuzla and Sarajevo demonstrated significant amounts of non-nationalist voting.