* This publication was prepared in part under a grant from the Kennan Institute for Advanced Russian Studies of the Woodrow Wilson International Center for Scholars, Washington, D.C. The statements expressed herein are that of the authors and are not necessarily those of the Woodrow Wilson Center.Google Scholar

The author would like to thank Paul Cummings for assistance in assembling materials for this article. The author is also grateful to the organizers of the State Department Title VII research presentations and the Kennan Institute for Advanced Russian Studies discussion series for providing opportunities to hear valuable comments on my research from scholars and practitioners specializing in ethnopolitical conflict and international organizations. Many thanks to the staff of the United Nations’ Department of Political Affairs, Europe Division and Asia Division, and the Department of Humanitarian Affairs for answering the author's questions and facilitating the author's access to the United Nations. Special thanks to Dr Walter Dorn for his insights on the work of the United Nations and the Department of Peacekeeping Operations.Google Scholar

1. Nations, United, Supplement to an Agenda for Peace, position paper of the Secretary-General on the occasion of the fiftieth anniversary of the United Nations, A/50/60-S/1995/1, para. 10–11.Google Scholar

2. Gates, Robert, From the Shadows: the Ultimate Insider's Story of Five Presidents and How They Won the Cold War (New York: Simon and Schuster, 1996), pp. 511, 527.Google Scholar

3. Trueheart, Charles, “U.N. Alerted to Plans for Rwanda Bloodbath,” Washington Post (25 September 1997), A1, A28.Google Scholar

4. For example, Dorn, Walter, “Early Warning of Armed Conflict by the U.N. Secretary General,” paper presented at the Synergy in Early Warning Conference, Toronto, 17 March 1997, p. 3; Gordenker, Leon, “Early Warning: Conceptual and Practical Issues,” in Rupesinghe, Kumar and Kuroda, Michiko, eds, Early Warning and Conflict Resolution (New York: St. Martin's Press, 1992), pp. 1–14.Google Scholar

5. George, Alexander L. and Holl, Jane E., “The Warning-Response Problem and Missed Opportunities in Preventive Diplomacy,” discussion paper, Carnegie Commission on Preventing Deadly Conflict, March 1997, pp. 3–5. Whereas the warning-action continuum is a necessary part of defining how early is “early” in EW, one must take care not to imply that only warnings that are acted upon qualify as early warnings. In that case, measures of early warning success will be tautological. Successful conflict prevention due to absence of premature early action (because of good warning) would not qualify as EW success.Google Scholar

6. For example, a corridor chat or a classified note to the Secretary-General do not count (U.N. can act without such documents, or use some of them and ignore others). However, Secretary-General's reports to the Security Council do count, because they are required for the U.N. to pass resolutions and act. The reports also represent consolidated inputs from many U.N. agencies.Google Scholar

7. These are the likes of Bosnia, Chechnya, Moldova, Rwanda, or East Timor, as distinct from the 1918–1920 civil war in Russia (fought predominantly over the nature of the political system) and also from L.A. riots (where ethnic and economic issues, rather than changing the U.S. political system or overthrowing a U.S. president, were at stake).Google Scholar

8. The number of casualties and refugees/IDPs captures most of these measures of violence.Google Scholar

9. These conditions fulfill a simple but often-overlooked research design rule: the dependent variable must be allowed to vary (in our case—predicted conflict magnitude and estimate probability). See King, Gary, Keohane, Robert O. and Verba, Sidney, Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton: Princeton University Press, 1994), pp. 129–130.CrossRefGoogle Scholar

10. Singer, David, in Ted Gurr, Robert, ed., Handbook of Political Conflict (New York: Simon Schuster, 1990), pp. 494–495; Duffy, Gavan, “An Early Warning System for the United Nations: Internet or Not?” Mershon International Studies Review, No. 39, 1995, pp. 315–318.Google Scholar

11. For a post-Soviet case study illustrating this neo-realist view, see Ignatieff, Michael, Blood and Belonging: Journeys into the New Nationalism (New York: Farra, Straus and Giroux, 1993). On domestic anarchy theory, see Jackson, Robert, Quasi-States: Sovereignty, International Relations and the Third World (Cambridge: Cambridge University Press, 1990).Google Scholar

12. Fein, Helen, Lives at Risk: A Study of Violations of Life Integrity in 50 States in 1987 based on Amnesty International 1988 Report (New York: Institute for the Study of Genocide, 1988); Mansfield, Edward and Snyder, Jack, “Democratization and the Danger of War,” International Security, No. 20, Summer 1995, pp. 5–38; Rousseau, David L., “Regime Change and International Conflict: Is Democratization Really So Dangerous,” paper presented at the annual meeting of the American Political Science Association, Washington, D.C., 28–31 August 1997.Google Scholar

13. Rummel, R. J., “National Attributes and Foreign Conflict Behavior,” in Singer, J. David, ed., Quantitative International Politics: Insights and Evidence (New York: Free Press, 1968); Treisman, Daniel S., “Russia's ‘Ethnic Revival': The Separatist Activism of Regional Leaders in a Postcommunist Order,” World Politics, No. 49, January 1997, pp. 212–249. A classic application of this analysis to the Soviet Union suggested that the Central Asian republics, rather than the Baltics or Ukraine, had the greatest secession potential: d'Encausse, Helen Carrere, Decline of an Empire: The Soviet Socialist Republics in Revolt (New York: Harper Colophon Books, 1989).Google Scholar

14. Given the ambiguity of the essentialist variables it is hardly surprising that ethnological monitoring of Russia's ethnic tensions by the Russian Academy of Sciences’ Institute of Ethnology and Anthropology, in conjunction with the U.S.-based Conflict Management Group, failed to identify the potential for violent conflict in Chechnya until as late as October 1994, when Grozny was already under air attacks by Moscow-backed Chechen opposition to Dudayev. In a bulletin issued at that time one gets a good picture of ethnic tensions in Omsk or Orenburg, but no analysis of an ethnopolitical conflict rapidly degenerating into war in Chechnya. See, Conflict Management Group, “Ethnic Conflict Management in the Former Soviet Union,” Bulletin, October 1994; for reports about escalation of violence in Chechnya see BBC, Summary of World Broadcasts, 12 October 1995, SU/2124 B/8.Google Scholar

15. Horowitz, Donald L., Ethnic Groups in Conflict (Berkeley: University of California Press, 1985); Gurr, Ted Robert, Minorities at Risk: A Global View of Ethnopolitical Conflicts (Washington, D.C.: United States Institute of Peace Press, 1993), p. 82; see also Roeder, Philip G., “Soviet Federalism and Ethnic Mobilization,” World Politics, No. 43, January 1991. The slow, reluctant secession of predominantly Muslim and resource rich Soviet Central Asian republics from Moscow also works against the more sophisticated hypothesis advanced by Horowitz (ibid., pp. 229–290) that backward non-titular ethnic groups within economically less developed regions will have the greatest propensity for separatism, while advanced ethnic groups in backward regions will have the lowest separatist propensity.Google Scholar

16. Brown, Michael, ed., The International Dimensions of Internal Conflict (Cambridge, MA: Center for Science and International Affairs, 1996).Google Scholar

17. Gurr, Ted Robert, “A Conceptual Framework for Screening and Interpreting Information is Required,” Mershon International Studies Review, No. 39, 1995, pp. 318–319.Google Scholar

18. A problem known as endogeneity. See, King, , Keohane, and Verba, (1994), pp. 107–108.Google Scholar

19. King, , Keohane, and Verba, (1994), p. 120. Thus, broadly inclusive data-processing networks—such as the one employed by the Russian Academy of Sciences’ Institute for Ethnology and Anthropology and Harvard's Conflict Management Group—make the observed effects explainable in terms of too many variables.Google Scholar

20. For example, a dynamic data model of ethnopolitical genocide, captures a complex balance of indicators, classified as “accelerators for communal challengers” (communal group clashes, regime policies threatening group status, factional strife within groups, formation of communal paramilitary units, and kindred group action in nearby countries) and “accelerators of regime insecurity” (rise of opposition coalitions, clashes between regime supporters and communal groups, increased external support for communal groups, and threats of external involvement). The study of eight ethnopolitical conflicts confirms that genocide is more likely when these indicators are more pronounced and when the ratio of regime instability to communal challenger accelerators increases. However, the same study implies that Slovenia and Croatia should have seen the same levels of ethnic violence, that Bosnia should have been a more violent case than Rwanda, and that a largely non violent outcome in both Kosovo and Macedonia should not have occurred. See Davies, John L. and Harff, Barbara, with Speca, Anne L., “Dynamic Data for Conflict Early Warning,” Conference Proceedings, Synergy in Early Warning, Toronto, 15–18 March 1997, pp. 91–112.Google Scholar

21. Friedman, Thomas L., “The Physics of Mideast Peace,” The New York Times, 15 September 1997, A23.Google Scholar

22. An Agenda for Peace, Preventive Diplomacy and Related Matters, A/RES/47/120.Google Scholar

23. Mission of the Department of Political Affairs (revised 1 January 1996), http://www.un.org/…a/docs/mission.html. Designed along the lines of a national foreign ministry with departments responsible for geographic areas and desk officers covering specific countries, the DPA incorporated five former departments dealing with political affairs.Google Scholar

24. The five departments and offices were the Department for Disarmament Affairs, the Department for Special Political Questions, Regional Cooperation, Decolonization and Trusteeship, the Department of Political and Security Council Affairs, and the Office for Research and Collection of Information. For 1993 changes see report of the Secretary-General (A/C 5/48/9).Google Scholar

25. Interview with John Renninger, New York, 13 August 1997.Google Scholar

26. For these reasons, having anything like a standard operating definition of EW is considered unnecessary, as is the development of assessment blueprints or models. The DPA desk officers, interviewed between 12–15 August and 8–10 September 1997, were Boventer, Gregor (Georgia), Horst Heitmann (Russia), and Vladimir Goryayev (Tajikistan and Central Asia).Google Scholar

27. Interview with John Renninger, U.N. Headquarters, New York, 13 August 1997. The model was also confirmed by responses of desk officers (note 21) to questions on what shapes their EW priorities. By target choice I mean cases that would be assigned to a specific desk officer, regularly monitored, and reported on via formal channels to the Secretary-General.Google Scholar

28. de Jonge Oudraat, Chantai, “The United Nations and Internal Conflict,” in Brown, Michael E., ed., The International Dimensions of Internal Conflict (Cambridge: The MIT Press, 1996), pp. 489–536. The U.N.'s resources—contingent on member states contributions—are infinitely smaller than those of the major powers. For example, the U.N.'s entire budget (which includes peacekeeping expenses) amounts to $7–8 billion a year, equal to 3% of the U.S. Department of Defense budget. See, ibid., p. 497.Google Scholar

29. Involvement here means direct participation in hostilities (unlike in Nagorno Karabakh, regular Russian troops have taken part in military operations in Chechnya, Tajikistan and Abkhazia).Google Scholar

30. According to Horst Heitmann, during Russia's massive attack on Chechnya in late 1994—early 1995, the Secretary-General blocked a proposal to write reports on human rights violations by both sides to the Security Council, anticipating a hostile Russian reaction.Google Scholar

31. Interview with Derek Boothby, Director of the Europe Division, at the Department of Political Affairs, New York, 13 August 1997. In EW, as in taking political action, senior practitioners have internalized the view that the U.N. “plays a marginal role” and, with a few exceptions, “there is no potential conflict situation that the United Nations would be better placed to detect or alert than major powers.” Interview with John Renninger, August 1997.Google Scholar

** Here and throughout the article, the terms “separatist” and “separatists” do not imply any judgement as to the validity of claims for Abkhazian independence. These terms simply describe the proponents of Abkhazia breaking away from Georgia, such as the government of Vladislav Ardzinba, as opposed to Abkhazians who advance no independence claims or to the provisional government of Abkhazia loyal to Tbilisi which was in place in Sukhumi in 1992–1993.Google Scholar

32. According to the Georgian government, the violence resulted from Abkhazian separatists’ overreacting to Shevardnadze sending about 3,000 guards to pursue pro-Gamsakhurdia militants into Western Georgia. The Abkhazian side has since asserted, however, that chasing these “Zviadists” was used by Tbilisi as a pretext to crush the Abkhazian drive for independence. Aklayev, Airat, Interethnic Conflict and Political Change in the Former USSR (Dundas: Peace Research Institute, 1995), p. 222. Based on Nezavisimaya Gazeta, 15 August 1992.Google Scholar

33. Despite a cease-fire agreement reached on 3 September 1992 in Moscow by the Republic of Georgia, the separatist leaders of Abkhazia and the Russian Federation, fighting continued, claiming 447 dead and nearly 2,000 wounded in the first 100 days. During 2–4 September, approximately 50,000 people (mostly Abkhazians) fled Abkhazia. Aklayev, p. 225, based on Nezavisimaya Gazeta, 25, 28 November 1992.Google Scholar

34. Low-level hostilities, marked by frequent exchanges of artillery and rocket fire, armed skirmishes and air raids, continued for nearly a year, mainly over control of Sukhumi and Ochamchira. With the end seemingly in sight, after the Moscow 1992 cease-fire agreement was renewed in Sochi on 27 July 1993 and following the dispatch of U.N. military observers, the Abkhazian forces launched a massive attack on Sukhumi on 16 September 1993. By 24 September, 246 Georgians were reported killed and 2,500 wounded in intense fighting (with indeterminate Abkhaz losses). By 1 October, over 100,000 people—mostly ethnic Georgians—were on the run; 80,000 refugees were caught between the Abkhaz and Georgian forces along the Kodori River. BBC World Service, Summary of World Broadcasts, 17 September–6 October 1993.Google Scholar

35. Serebryannikov, V. V., Voina v Chechne: porozheniye smuty [War in Chechnya: The Origins of Turmoil], in Aydayev, Yu A., ed., Chechentsi: istoriya I sovremennost' (Moscow: Mir domu tvoyemu, 1996), p. 16 reports that in May 1992, Pavel Grachev, then newly appointed as Russia's Defense Minister ordered the commander of the North Caucasus Military District to transfer control of over 50% of weapons in Chechnya to the Chechen leader, Dzhokar Dudayev.Google Scholar

36. Letter Dated 7 October 1992 from the Secretary-General Addressed to the President of the Security Council, S/24633, 8 October 1992, p. 1. Whereas warnings and reports from various agencies and officials within the U.N. must have reached the Secretary-General prior to 10 September 1992, they cannot be considered legitimate instances of effective EW. Firstly, they were outside the stream of official documents upon which the U.N. action could be based. Secondly, even if made they were not strong enough to induce the Security Council to request regular reports on Abkhazia earlier (to say nothing of taking any preventive action).Google Scholar

37. According to Gregor Boventer, the DPA desk officer who has covered Georgia since 1992, these confidential reports often differ from public documents, particularly when estimating the role of major powers. However, in John Renninger's view, with which the desk officers agreed, the reports and letters of the Secretary-General to the Security Council—drafted and finalized at the DPA in conjunction with the DHA and DPKO—are representative of DPA's and U.N.'s overall tendencies in situation assessment. Interviews, August, September 1997.Google Scholar

38. Various letters and notes verbales by member states, international treaties, agreements, and statements are routinely circulated as Security Council documents. This part of the information stream reaching the Security Council may be analyzed in terms of institutional noise-to-signals ratio within the U.N., but these documents are not reflective of U.N.'s internal collection and analysis process. More importantly, perhaps, these documents do not reflect what the Security Council wants to learn about the situation so that it can take action—a critical factor in organizational threat assessment and EW.Google Scholar

39. One plausible explanation for the “territorial escalation” assessment by the U.N. would be that the Secretariat took the cue from Eduard Shevardnadze's assertions in a number of communications to the Secretary-General in autumn and winter of 1992 to the effect that the disintegration of Georgia will lead to extremist separatism engulfing “the entire Caucasus region” and ruining Georgia's neighbors and “the whole mankind.” (Letter Dated 25 December 1992 from the Chairman of the Parliament and Head of State of the Republic of Georgia Addressed to the Secretary General, S/25026, 30 December 1992, p. 2). Also, as an organization of member states, the U.N. is mandated with maintaining (and, hence, monitoring) international peace and security. Therefore, without a projection that conflict may spread beyond a single member state, the U.N. technically has little legal basis for preventive action and EW. In the Abkhazian case, assessments that violence will spread to other regions proved wrong.Google Scholar

40. According to Gregor Boventer, a desk officer charged with Georgian affairs, the U.N. takes into account both the macro factors such as the economic situation, and such micro conditions as the number of cigarettes the Abkhazian leader, Vladislav Ardzinba, smokes during meetings, the locations where U.N. missions are received, or the number of times Ardzinba consults with his aides. Interview, 15 August 1997.Google Scholar

41. United Nations Security Council, Report of the Secretary-General Concerning the Situation in Abkhazia, Georgia, S/26551, 7 October 1993, pp. 2–5; United Nations Security Council, Resolution 858 (1993), S/RES/858, 24 August 1993.Google Scholar

42. Ibid.Google Scholar

43. Ibid., quote on p. 6.Google Scholar

44. United Nations Security Council, Report of the Secretary-General on the Situation in Abkhazia, Republic of Georgia, S/25188, 28 January 1993, p. 5.Google Scholar