¹ This article is based on an assessment of events through early June 1994.

² The term ‘civil conflict’, as used in this paper, refers to a violent conflict between two or more parties within the boundaries of the same state, in which the goals of the conflicting parties are stated in the context of that particular state's polity. The 1983 US intervention in Grenada, for instance, would not be an instance of civil conflict. This definition, however, does not preclude consideration of those cases where civil conflict has transboundary ramifications, or where one or more of the conflicting parties is performing proxy roles for external agents.

³ Conflicts which are ‘resolved’ or terminated through one party's military victory are not considered here.

⁴ This discussion draws upon the work of Stuart Nagel, Policy Evaluation Methods New York, (1992); Pruit, D. G. and Rubin, J. Z., Social Conflict: Escalation, Stalemate and Settlement (New York, 1986)Google Scholar; and Touval, S. and Zartman, I. W. (eds), International Mediation in Theory and Practice Washington, (1985)Google Scholar. This argument also parallels the emphasis on pacts in the comparative literature on democratization. As O'Donnell and Schmitter state, pacts are likely when two parties ‘can neither do without each other nor unilaterally impose their preferred solution on each other if they are to satisfy their respective divergent interests’. O'Donnell, Guillermo and Schmitter, Philippe, Transitions from Authoritarian Rule. Tentative Conclusions about Uncertain Democracies Baltimore, (1986), p. 38Google Scholar. Di Palma adds to this, showing how political actors with conflicting interests have an incentive to accept a pact when otherwise ‘their immediate prospect is a reciprocal stalemate, fed by recalcitrance and polarization, and without visible exit’. Di Palma, Giuseppe, To Craft Democracies. An Essay on Democratic Transitions Berkeley, (1990), p. 56Google Scholar.

⁵ This situation could be represented in terms of game theory as a Prisoners’ Dilemma game, which consists of the following pay–off matrix.

In such a game, both players clearly have a collective incentive to cooperate, that is, to negotiate, rather than continue fighting. But, at the same time, there remains a tendency for each individual actor to defect from the negotiations and continue to fight.

While game theory helps to conceptualize the choices faced by the warring actors in a situation of civil war, game theory itself has nothing to say concerning the origin of actors' preferences and changes in actors' preferences. Thus, it is necessary to consider contextual factors that account for the overcoming of the collective action problem embedded in the Prisoners' Dilemma game. It is in this sense that this article considers the role of the UN and a series of other factors that are, properly speaking, external to the game. The usefulness of game theoretic language is thus derived from its ability to pinpoint with a high degree of simplicity and elegance the key obstacles to progress in the peace process.

⁶ Whitehead, Laurence, ‘Democracy by Convergence and Southern Europe: A Comparative Politics Perspective’, in Pridham, Geoffrey (ed.), Encouraging Democracy: The International Context of Regime Transition in Southern Europe London, (1991), pp. 54–5Google Scholar. The role of third–party mediation, of course, has long been discussed in the conflict resolution literature. See, among others, Hoffman, Mark, ‘Third–Party Mediation and Conflict Resolution in the Post–Cold War’, in Baylis, John and Rengger, N. J. (eds.), Dilemmas of World Politics: International Issues in a Changing World Oxford, (1992), pp. 260–86Google Scholar; and Princen, Thomas, Intermediaries in International Conflict Princeton, (1992)Google Scholar.

⁷ The term ‘international intervention’ refers to a multilateral intervention, or an intervention by a regional or global international organization, in a civil conflict. The intervention's expressed purpose should be that of obtaining a negotiated resolution of the conflict, and not the forceful imposition of the goals of any one of the intervening parties. Good examples of the kinds of intervention encompassed in the definition are the UN interventions in Cambodia, El Salvador, and Namibia, as well as the multilateral West African intervention in Liberia. This definition would thus preclude, for instance, the 1989 US intervention in Panama, or the 1991 Indian intervention in the Maldives, as well as all other unilateral interventions. However, an exception would be made in those instances where the purpose of unilateral intervention is the restoration of peaceful conditions, and where the intervention is carried out in collaboration with, or with the support of, various international bodies. A good example of such intervention is Italy's recent role in brokering the peace accords in Mozambique.

⁸ Diehl, Paul F., ‘When Peacekeeping Does Not Lead to Peace: Some Notes on Conflict Resolution’, Bulletin of Peace Proposals, 18 (1987), pp. 47–53CrossRefGoogle Scholar.

⁹ The in–depth case studies of El Salvador and Cambodia were used to pinpoint which factors were likely to affect the disposition of the conflicting parties to continue cooperating, and to formulate some initial hypotheses concerning the role of these factors, through a logic Lijphart outlines in his discussion of ‘hypothesis–generating’ case studies. These factors have also been highlighted in the literature on conflict resolution, as well as in other complementary bodies of literature. In a second stage of the research, these hypotheses were tested comparatively. On the use of case studies in the development of theory, see Lijphart, Arend, ‘Comparative Politics and the Comparative Method’, American Political Science Review, 65 (1971), pp. 682–93CrossRefGoogle Scholar, pp. 691–3; Eckstein, Harry, ‘Case Study and Theory in Political Science’, in Greenstein, Fred I. and Polsby, Nelson W. (eds.), Handbook of Political Science, vol. 7, Strategies of Inquiry Reading, MA, (1975), pp. 79–137Google Scholar; and George, Alexander L., ‘Case Studies and Theory Development: The Method of Structured, Focused Comparison’, in Lauren, Paul Gordon (ed.), Diplomacy: New Approaches in History, Theory and Policy New York, (1979)Google Scholar.

¹⁰ As stressed in bargaining theory, particularly as applied to corporatist pacts, the existence of welldefined collective actors with a leadership able both to speak for the rank–and–file membership and to enforce conditions agreed to with other parties on this membership, is a precondition for any negotiation to take place and for the resulting agreements to be enforced. In this article we are concerned with the characteristics of the actors that affect the likelihood of the signed accord being implemented. The reason for dyadic disputes being more amenable to resolution is that an increase in the number of parties to a dispute leads to an increase in the number of conflicting interests that need to be coordinated in the negotiation and implementation phases. The destabilizing tendencies are, accordingly, increased. Lange, Peter, ‘Unions, Workers, and Wage Regulation: The Rational Bases of Consent’, in Goldthorpe, John (ed.), Order and Conflict in Contemporary Capitalism: Studies in the Political Economy of Western European Nations Oxford, (1984), pp. 98–123Google Scholar, pp. 106–8. This type of argument has been made in the context of international intervention and conflict resolution by Diehl, Paul F. and Kumar, Chetan, ‘Mutual Benefits from International Intervention: New Roles for United Nations Peacekeeping Forces’, Bulletin of Peace Proposals, 22 (1991)CrossRefGoogle Scholar.

¹¹ The term ‘external party’ is used to refer to the parties that were initially involved in the conflict, and not to the mediating party. For example, while in Cambodia the external mediator is the UN, the external parties include China, Thailand, and Vietnam.

¹² The democratization literature has also argued that the process of pact–making sometimes calls for guarantees being extended to certain parties in return for their willingness to negotiate with their opponents and back political solutions to conflictive situations. See O'Donnell and Schmitter, Transitions from Authoritarian Rule, pp. 28–32, 69. Of course, electoral laws and constitutional structures will also determine whether power–sharing and respect for minority rights are likely to result, and whether the losing party will have a realistic chance of coming to power at a future date. See Shugart, Matthew S., ‘Guerrillas and Elections: An Institutionalist Perspective on the Costs of Conflict and Competition’, International Studies Quarterly, 36 (1992), pp. 121–52CrossRefGoogle Scholar.

¹³ Ruggie, John Gerard, ‘Wandering in the Void: Charting the UN's New Strategic Role’, Foreign Affairs, 72 (1993), pp. 26–31CrossRefGoogle Scholar.

¹⁴ Rosenau, James, The United Nations in a Turbulent World New York, (1991)Google Scholar.

¹⁵ Rosenau, James, Turbulence in World Politics: A Theory of Continuity and Change Princeton, (1991)Google Scholar.

¹⁶ This study, thus, primarily employs John Mill's method of difference, as discussed by comparativists such as Lijphart, ‘Comparative Politics and the Comparative Method’.

¹⁷ The following discussion draws upon Vickery, Michael, Kampuchea, Politics, Economics and Society London, (1986)Google Scholar; McAuliff, John and McDonnell, Mary Byrne, ‘Ending the Cambodian Stalemate’, World Policy Journal, 7 (1989–1990), pp. 71–105Google Scholar; Haas, Michael, Genocide by Proxy. Cambodian Pawn on a Superpower Chessboard (New York, 1991)Google Scholar; Chandler, David P., A History of Cambodia Boulder, 1992, 2nd edn)Google Scholar; and Papp, Daniel S., ‘Soviet and American Peacemaking Efforts in Regional Conflicts in Asia, Afghanistan and Cambodia’, in Goodman, Melvin A. (ed.), The End of Superpower Conflict in the Third World (Boulder, (1992), pp. 93–137Google Scholar. For background information covering the 1945–79 period, see Chandler, David P., The Tragedy of Cambodian History: Politics, War, and Revolution since 1945 New Haven, (1991)Google Scholar.

¹⁸ Vietnam and Laos also participated in these meetings.

¹⁹ The rulers in Phnom Penh wanted to retain control of Cambodia in the transitional period, with any interim authority being limited to organizing the elections. The three remaining factions, which included the Khmer Rouge, wanted a coalition government for Cambodia.

²⁰ The spectre of a new Khmer Rouge government created an uproar in the West. In particular, the US, no longer burdened by Cold War concerns, took strong action against the Khmer Rouge. It pressured China to stop all aid to the Khmer Rouge and China complied, in large part due to its effort to restore its tarnished image in the wake of the Tiananmen Square massacre. The US also put pressure on Thailand, which began to clamp down on Khmer Rouge bases in its territory. Furthermore, the US withdrew recognition from the CGDK in the UN and even promised humanitarian aid to the government in Phnom Penh.

²¹ For the entire text of the final peace accord, see the Agreements on a Comprehensive Political Settlement of the Cambodia Conflict, United Nations, DPI/1180–92077–January 1992–10M (New York, 1992).

²² On the surface of things, the assertion that UNTAC would administer Cambodia appears to contradict the UN Charter, which prohibits UN trusteeship of a member state (Article 78) and bans any UN intervention in matters that are considered as being within the domestic jurisdiction of any state (Article 2/7). The UN, however, got around these limitations by creating the SNC, defined in the Paris accords as ‘the unique legitimate body and source of authority in which, throughout the transitional period, the sovereignty, independence, and unity of Cambodia are enshrined’. The SNC, to which Prince Sihanouk was elected as head, then ‘delegate[d] to the United Nations all powers necessary to ensure the implementation of this Agreement’. (Alternatively, the Security Council could have ordered an enforcement action under chapter VII to restore international security, as happened in the Gulf war; this was not considered practicable for Cambodia.)

The powers that would be delegated to UNTAC were clearly defined. First, all government offices relating to foreign affairs, finance, defence, internal security and information were to be placed under the direct control of UNTAC, since the neutral functioning of this bureaucracy was deemed essential to ensure fair elections. Second, UNTAC maintained supervisory powers over any government agency which could, through less–than–neutral functioning, adversely affect the fairness of the elections. These agencies included those handling commerce, communications, and education. Third, the special representative in charge of UNTAC could conduct investigations into agencies such as those dealing with cultural or religious affairs if s/he felt that their activities impinged upon the election's outcomes. Jurisdictional disputes arising from these guidelines were to be settled as follows: UNTAC had plenary authority in all matters pertaining to the elections; in other matters, the UN special representative would listen to the SNC if the latter proposed policy through Prince Sihanouk on the basis of unanimity among the SNC members. Even in the latter scenario, the Paris Agreement suggested that the special representative accept the SNC's advice if it was consistent with the objectives of the Agreement as determined by the special representative. In other words, UNTAC was being handed almost complete authority for the administration of the country, even though the language of the Paris Accords was such as to suggest that the UN would act only on the basis of delegation from the local actors. For more details, see Ratner, Steven R., ‘The United Nations in Cambodia: A Model for Resolution of Internal Conflicts?’, in Damrosch, Lori Fischer (ed.), Enforcing Restraint: Collective Intervention in International Conflicts New York, (1993)Google Scholar.

²³ The following discussion draws upon LeoGrande, William M., ‘After the Battle of San Salvador’, in Coleman, Kenneth M. and Herring, George C. (eds.), Understanding the Central American Crisis Wilmington, (1992)Google Scholar; Karl, Terry Lynn, ‘El Salvador's Negotiated Revolution’, Foreign Affairs, 71 (1992), pp. 147–64CrossRefGoogle Scholar; Baloyra, Enrique, ‘Salvaging El Salvador’, Journal of Democracy, 3 (1992), pp. 70–80CrossRefGoogle Scholar; Tulchin, Joseph and Bland, Gary (eds.), Is There a Transition to Democracy in El Salvador? Boulder, (1992)Google Scholar; and Munck, Gerardo L., ‘Beyond Electoralism in El Salvador: Conflict Resolution Through Negotiated Compromise’, Third World Quarterly, 14 (1993), pp. 75–93CrossRefGoogle Scholar. For background information, see Baloyra, Enrique, El Salvador in Transition Chapel Hill, (1982)Google Scholar.

²⁴ Following Christian Democrat Napoleon Duarte's victory in the May 1984 presidential race certain prospects for peace did emerge. As a reformist, Duarte proposed a meeting between the government and the FDR–FMLN. In October 1984, thus, a dialogue began in the town of La Palma. But the US, especially after Reagan's 1984 reelection, pushed for an escalation of the war and clearly sought to undermine the peace process.

²⁵ As in the case of ASEAN with regard to Cambodia, the first tentative peace initiatives were regional. The formation of the Democratic Convergence, indeed, was made possible by the Central American Peace Plan or Esquipulas II accord, signed by the elected presidents of Guatemala, El Salvador, Honduras, Nicaragua and Costa Rica in August 1987. On the regional initiatives, see Latin American Studies Association, ‘Extraordinary Opportunities … and New Risks. Final Report of the LASA Commission on Compliance with the Central American Peace Accord’, LAS A Forum, 19 (1988)Google Scholar; and Child, Jack, The Central American Peace Process, 1983–1991 Boulder, (1992)Google Scholar.

²⁶ Adams, Jan S., A Foreign Policy in Transition: Moscow's Retreat from Central America and the Caribbean, 1985–1992 Durham, (1992)CrossRefGoogle Scholar.

²⁷ Negotiations between the Cristiani government and the FMLN started in September 1989, but rapidly the army attempted to disrupt the talks, a move immediately responded to, with a new offensive, by the FMLN. At this point, the UN stepped in. In early 1990, it organized secret meetings between the government and the FMLN, which then led to the beginning of formal talks in April 1990. From the beginning, the UN Secretary General Javier Perez de Cuellar oversaw the negotiation process. The role of the UN became more important after the October 1990 agreement by the negotiating parties to allow the UN Secretary General's representative, Alvaro de Soto, to put forward his own proposals. Karl, ‘El Salvador's Negotiated Revolution’, pp. 154–5.

²⁸ On the negotiations, see Karl, ‘El Salvador's Negotiated Revolution’, pp. 154–61. For the entire text of the final peace accord, see ‘Acuerdos de Paz de Chapultepec’, Revista del Pensatnienlo Cenlroamericano, 214 (1992), pp. 32–82Google Scholar.

²⁹ There were also differences in terms of how the negotiations were carried out in El Salvador and Cambodia. In El Salvador the parties to the conflict played an important role in bargaining over key issues. Another difference concerns the regional peace efforts, which were clearly more productive in Central America, mainly due to the less complicated pattern of external involvements.

³⁰ This positive assessment of progress in El Salvador does not seek to brush aside its more problematic aspects. There were some clear violations of the peace accord during 1992. On the part of the government, President Cristiani delayed the purge of those military officers identified as having violated human rights beyond what he had agreed to, in effect violating the peace accords. On the part of the guerrillas, evidence emerged that they had stockpiled weapons, also in violation of the peace accords. A series of problems persisted well into 1994. Indeed, UN Secretary–General Boutros Boutros–Ghali informed the UN Security Council in May 1994 that aspects of the peace accords pertaining to the areas of public security, land reform, the recommendations of the UN–appointed Truth Commission, and programmes for the re–insertion of former combatants into civilian life, remained to be fully implemented. Thus, the peace process continued with the extension of the UN mission to El Salvador (ONUSAL) for six months beyond its original 1 June expiration. What is most important, however, is the attitude of key actors. As with the well–documented irregularities during the electoral process in 1994, a key sign of the progress toward peace is that the leftist parties did not seek to make too much out of them, fearing that the possible electoral gains they could make in this manner would overshadow a broader gain: that they had finally become an accepted actor in El Salvador's electoral process and that the new government was seen as legitimate by all sectors of Salvadoran society.

³¹ Watch, Asia, ‘Cambodia: Human Rights Before and After the Elections’, Asia Watch 5 (May 1993), pp. 9–12Google Scholar.

³² An associated problem is that the overall ineptitude of the government, and infighting between the FUNCINPEC and their Hun Sen partners, had paralyzed the new government. Corruption and administrative chaos are endemic among both the bureaucrats and the military, and little is left of the UN's legacy in terms of creating some accountability for human rights abuses, particularly by the army and the security forces. This latter factor has been primarily responsible for generating grassroots support for the Khmer Rouge in some rural areas. On the recent advances of the Khmer Rouge and the response of the government, see Thayer, Nate and Chanda, Nayan, ‘Things Fall Apart…’, Far Eastern Economic Review, 19 May 1994, pp. 16–18Google Scholar; and Shenon, Philip, ‘U.S. Considers Supplying Arms to Help Sihanouk in Cambodia’, New York Times, 15 May 1994, pp. 1, 8Google Scholar.

³³ On the May 1993 election and its aftermath, see Jeldres, Julio A., ‘The UN and the Cambodian Transition’, Journal of Democracy, 4 (1993), pp. 104–16CrossRefGoogle Scholar, pp. 109–14. For a detailed critique of UNTAC's strategy of focusing primarily on holding elections, to the detriment of other peacekeeping goals, see ‘Cambodia: Elections in the Killing Fields’, in Jean, Francois (ed.), Life, Death, and Aid: The Medecins Sans Frontieres Report on World Crisis Intervention New York, (1993), pp. 79–86Google Scholar.

³⁴ The unification of the FMLN came about in October 1980, more or less at the insistence of Fidel Castro. A condition for his offering any assistance to the rebel forces was that they have a clearly identifiable head. Hence, the five main guerrilla factions at the time came together in the FMLN. Recently, in May 1994, when the new Legislative Assembly was sworn in, a dispute between two wings of the FMLN broke out. But the potential split in the FMLN comes at a point in time when the major obstacles on the way to peace appear to have been overcome.

³⁵ Even though the individual members of the coalition signed the accord separately, the CGDK had acted as a unified body.

¹⁶ Diehl and Kumar, ‘Mutual Benefits from International Intervention’.

³⁷ For example, one of the biggest problems in El Salvador was the existence of a big rural refugee population. Part of the Salvadoran peace process involved the resettlement of these refugees. In initiating and promoting this process, the UN's Observer Mission in El Salvador operation followed the lead of the various Salvadoran organizations that already existed for the purpose. Significant in this regard were the National Coordinator of Repopulation, a non–governmental organization representing the displaced, the Salvadoran government's commission on the displaced (CONADES), and the Christian Committee for the Displaced (CRIPDES), a church–sponsored NGO. The UN's cooperation with these agencies helped restore and strengthen civil society after years of warfare, thereby allowing the peace process to gain a momentum that is likely to continue once the UN begins to withdraw. See Central America Report, 13 March 1992Google Scholar.

³⁸ On the UN role in Cambodia, see Jeldres, ‘The UN and the Cambodian Transition’, pp. 106–9.

³⁹ ‘Easy Scapegoat: People Blame the UN for all their Woes’, Far Eastern Economic Review, 22 October 1992Google Scholar.

⁴⁰ Jeldres, ‘The UN and the Cambodian Transition’, p. 107.

⁴¹ As mentioned above, the Democratic Convergence was made up of important politicians aligned with the FMLN, which came together as a new coalition representing a leftist agenda, following the signing of the Esquipulas II accord in August 1987. In the March 1991 elections they gained 12 per cent of the vote and 8 of 84 seats in the unicameral National Legislature.

⁴² ‘FMLN Close on ARENA'S Heel’, Central America Report, 14 August 1992Google Scholar.

⁴³ On the concept of ‘ripeness’, see Zartman, I. W., Ripe for Resolution: Conflict and Intervention in Africa Oxford, (1985)Google Scholar.

⁴⁴ This role of external actors also affects the contribution of international intervention in the negotiation phase. In this regard, the end of the Cold War and the increasing willingness and interest of major powers to allow for greater UN initiative has made it possible for the UN to take advantage of these opportunities and to offer its services as a broker in Cambodia and El Salvador. Where the degree of interest on the part of both regional as well as global actors to see something done is much lower, as in Sudan or Tajikistan, this lack of attention impedes a peace process from even starting.