• Home

OSCE Mediation of Nagorno-Karabakh conflict

When the Nagorno-Karabakh conflict erupted at the end of 80s and beginning of 90s no one could presage it would remain immune to a final solution over decades. Armenia and Azerbaijan have locked horns over the destiny of the region that was part of Azerbaijan throughout decades, despite the opposing narratives of Armenia suggesting the idea of Karabakh’s Armenian origins. The conflict was the offspring of Tsarist Russia’s wanton “divide and rule” policy that allowed it to maintain its lever over large swathes of territories by exploiting ethnic fault lines. It was due to this logic that large numbers of ethnic Armenian population were resettled in Azerbaijani territories, including Nagorno-Karabakh, in the beginning-mid of 19th century.


This purposeful manipulation with ethnic card created fertile grounds for future ethnic conflict that exploded when the ramshackle Soviet Union was biding goodbye to its existence. Hence the inherited Nagorno-Karabakh conflict that the Organization on Security and Co-operation in Europe (OSCE) is desperately trying to solve for two decades. This piece is not about the origins and history of Nagorno-Karabakh conflict and therefore does not aim to delve into long description of conflict background. The purpose is to highlight the OSCE efforts to mediate the conflict and analyze the conduct of its work from the prisms of mediation theories. How once can rate the level of success by the OSCE in its rather protracted meditation work? What are the interlocking determinants that define the boundaries of possible success? Is there a factor of bias, and if yes, how does it blend with the idea of impartial mediation? Those are the critical questions that this essay will seek to address.  To be able to do it covering short history of the mediation process presents to be necessary. 

Dr. Esmira Jafarova is a Visiting Scholar at Columbia University, Harriman Institute in the city of New York. She received her Ph.D. from the University of Vienna, Austria. She has held diplomatic positions at the Permanent Missions to the OSCE and the UN. The views expressed in the current work are purely her own and do not represent any organization.

A little history

International involvement in Nagorno-Karabakh conflict between Armenia and Azerbaijan was not possible at the outbreak of hostilities, since the conflict started before the dissolution of the USSR. This situation excluded any initiatives by international actors, since it could have been interpreted as the interference into the internal affairs of a sovereign state. Things come to head after the disintegration of the Soviet Union and subsequent internalization of the conflict. In 1991-1992 there were several mediation attempts by Russia, Kazakhstan and Iran. The United Nations did not shoulder the responsibility for conflict mediation due to already being overburdened with other priorities around the world.[1] Upon the membership of both states in the Conference on Security and Co-operation in Europe (CSCE) in 1992, the latter as a regional arrangement under Chapter VII of the UN Charter undertook a major role of mediating the conflict resolution.

At the beginning, the CSCE sent a special mission of rapporteurs under the leadership of Czech diplomat, Karel Schwarzenberg to collect the information about the conflict. Another mission following the first one was led by Mr. Dienstdier, the CSCE Chairman in Office.[2] After the Khojaly massacre[3] on February 26, 1992, the CSCE Ministerial Council took a decision to convene a conference in Minsk on the Nagorno-Karabakh conflict, to work out a final settlement to the conflict. However, due to dissonances on the idea of conference the Minsk Conference was quickly replaced with “Minsk Group”, which comprised of  eleven participating states (Belarus, Czech Republic, Slovakia, France, Germany, Italy, Russia, Sweden, the US, Armenia and Azerbaijan)[4]. In March of 1995, the mandate of the Minsk Group co-chairs was adopted. This gradual retrenchment from a conference to co-chairmanship institution in fact marked the start of intensive shuttle diplomacy between the parties in quest of a solution.

From its inception, CSCE’s mediation efforts were negatively affected by the string of inclement events. Occupation by Armenian forces of Azerbaijan cities Shusha (8 May) and Lachin (18 May) further escalated the conflict and complicated peace efforts. In November 1992, during Prague meeting of the CSCE Conference of Senior Officials (CSO) a decision was taken to establish a special planning group in Vienna, which had to prepare the Advance Monitoring Group to be deployed in the region. Furthermore, at the Rome meeting of the Minsk group in February 1993 a preliminary agreement was signed to send a group of special observers to monitor the situation. However, new Armenian attacks in Kelbajar district of Azerbaijan at the end of March, and new surge in the level of violence in the front line played havoc with peace plans.[5]

Political tensions in Azerbaijan and change of leadership therein coupled with Armenia’s further military advancements[6] further conditioned a bleak setting for CSCE’s rather timid initial mediation efforts. In these circumstances, growing skepticism in Minsk Group’s capacity to deliver effective solutions to cease the hostilities on the ground contributed to Russia’s increasing profile as a mediator. In an apparent disregard to the CSCE, Russian government headed by Victor Chernomyrdin was trying to seize control over the situation by mediating between the parties in order to reach a ceasefire deal. Instead of the international “monitoring force” she wanted to organize and control the CIS “separation force”, though under the auspices of the CSCE, which would enable Russia to permanently station itself in the conflict zone.[7] However, in 1993 in view of internal political turbulence in the country Russia had to shelve its plans to dominate conflict resolution, which enabled the CSCE to reenergize its activities. With the efforts of the Swedish co-chair of the Minsk Group meetings between the conflicting parties were resumed. Nevertheless, this window of opportunity quickly closed due to Armenian attacks on two Azerbaijani regions, Zangelan and Goradiz. Control over the mediation process was again assumed by Russia, who brokered the cease-fire on May 12, 1994.[8]

After the cease-fire agreement Russia was back to “dominate the process” mood. If Armenia, as a traditional Russia ally was not only averse to, but also highly content with Russia’s increasing share in the process, Azerbaijan, on the contrary, was not happy about Russia’s attempts to take the lead in the process. It was suspicious of Russia’s intentions to repeat Georgian scenario, and under the pretext of peace efforts and peacekeeping forces to root itself indefinitely in Azerbaijani territories. After the signing of the so-called “Contract of the Century” in September, 1994 between the State Oil Company of the Azerbaijan Republic (SOCAR) and foreign oil companies on the exploration of the oil resources of the Caspian basin, the West started to perceive Azerbaijan in more strategic terms. Necessity of ensuring security and stability in this part of the world now correlated with the urgency of securing the critical infrastructure projects that were in the process of becoming.

“Second breath” or new wave of activism

The above developments brought a new wave of activism in Minsk Group’s work. Following years, especially 1996-1997 marked the increased intensity in mediation activity. At the OSCE Lisbon Summit of 1996, the Member States laid out very important three principles as a legal basis for the peaceful settlement process. The principles were as follows: 1) territorial integrity of the Republic of Armenia and the Republic of Azerbaijan Republic; 2) legal status of Nagorno-Karabakh, defined in an agreement based on self-determination, which confers on Nagorno-Karabakh the highest degree of self-rule within Azerbaijan; 3) guaranteed security for Nagorno-Karabakh and its population, including mutual obligations to ensure the compliance by other parties with the provisions of the settlement.[9]

However, the OSCE consensus rule became the part of the problem that prevented the inclusion of all three principles in the final Lisbon document. Although 53 out of 54 participating states, including Russia, accepted the principles, Armenia’s rejection to join consensus rendered impossible the reflection of the above principles in the main Lisbon document, and achievement of any mutually agreeable arrangement based on Lisbon principles. Thus, this protracted inter-state conflict was left in anticipation of the next steps for the peaceful resolution.

In 1997 the composition of the Co-Chairmanship of the Minsk Group was altered[10] in accordance with the Chairman-in-Office’s decision to include France, the Russian Federation and the United States, which remains unchanged till present time. Change in composition of the Minsk Group had almost immediately ushered in new initiatives. In 1997 the co-chairs initiated famous three proposals, –  the so-called “package” variant, “stage by stage” or “phased” and “common state” proposals – that envisaged negotiated solutions to the conflict. The “package” approach aimed at achieving the solution on cessation of hostilities/withdrawal of armed forces by Armenia and the agreement on final status of Nagorno-Karabakh region in one stage. “Phased” solution implied the cessation of hostilities/withdrawal of armed forces by Armenia in the first place to be followed by the negotiations on the status of the Nagorno-Karabakh region. The third and the most contentious “common state” proposal promoted the idea of Nagorno-Karabakh region becoming a state-territorial formation in the form of Republic and constituting a common state with Azerbaijan within its internationally recognized borders.[11] All three proposals failed to forge a base for a negotiated solution, due to parties’ different security considerations and preferences for one or the other model of conflict resolution.

Engaging face-to-face

After the failure of all three proposals, the Minsk Group stopped preparing new proposals, and the next strategy became the organization of tête-à-tête negotiations between Armenian and Azerbaijani leaders to discuss again all three proposals individually. In 1999-2001 the presidents met several times in Washington, Istanbul, Geneva, Davos, Moscow, Yalta, Paris and Key-West. The negotiations between the Presidents were highly confidential, but the basic idea of the talks followed the logic that the peace deal could comprise of the elements of two or more proposals if the Presidents could succeed to achieve any progress. However, mutually exclusive demands of the parties coupled with Armenia’s rigid negotiating position to perpetuate its gains from the war rendered these talks ineffective. Suspicions in the ability of the Minsk Group to perform effective mediation went on the rise again. Nevertheless, the tête-à-tête strategy was not given up.

The mediators employed a new approach towards the negotiations by establishing contacts mainly on ministerial level in order to build proper environment for further negotiations by the presidents. In 2004 the Prague Process, which envisaged direct bilateral negotiations between the Azerbaijani and Armenian Foreign Ministers was initiated. In 2004-2006 there was optimism about the window of opportunity to achieve a settlement between the election cycles in both countries (parliamentary elections in 2005 and the presidential elections in 2008 in Armenia and Azerbaijan). During the presidential meeting in Kazan in 2005 the Co-chairs proposed the so-called “basic principles” for further negotiation.[12]

These “basic principles” were later refined and presented to the parties at the OSCE Madrid Ministerial in 2007 as the so-called “Madrid Principles”. The principles in fact combined the “package” and “phased” solution approaches, and envisaged the withdrawal of Armenian forces from the occupied territories that are adjacent to Nagorno-Karabakh region with special modalities for Lachin and Kelbajar districts. Restoration of social and transport infrastructure, resumption of trade relations should be the part of the plan, while establishment of interim international security arrangements for Nagorno-Karabakh region was considered until the voting on status takes place.

The modalities of a referendum or a population vote for defining the future legal status of the region are to be agreed and conducted as a result of a negotiated agreement between Armenia and Azerbaijan, and this vote should be conducted in a non-coercive environment in which all the citizens have opportunity to envisage their positions after public debate. The parties’ opinions, however widely differ on constituencies, modalities and the scope of the potential vote on the region’s status with Azerbaijan making reservations that 1) IDPs from Nagorno-Karabakh should equally participate in the vote and therefore, should previously be returned to the occupied areas; 2) according to Article 3 of the Azerbaijani Constitution any vote which could result in independence of Nagorno-Karabakh should be a nationwide referendum.[13] Azerbaijan is ready to give the Nagorno-Karabakh region the highest degree of autonomy within its internationally recognized borders.

In 2009 the above Madrid Principles underwent another revision, but the points of contention nevertheless remained unchanged. Since then, until January 2012 Sochi meeting of the presidents, the negotiations although snail-paced, still were more or less on track. However, Armenia’s position during the negotiations has weathered visible fluctuations. Despite giving its assent to the concept and constituting elements of the Madrid principles, and participating in the talks on their basis, she apparently was mostly unhappy about the clause concerning the withdrawal of its forces from the occupied territories. In 2011 Kazan meeting of the presidents Armenia went on asking to revise previously agreed provisions of Madrid principles that dealt with the withdrawal of its forces from the occupied territories. This of course was met with Azerbaijan’s backlash, who retorted against any revision in the long negotiated document. For Azerbaijan, liberation of its occupied territories remains as an utmost priority and agreeing to anything less than that is a non-starter. Persisting disagreements between the parties fortified by this new tide of intransigence mothballed the entire negotiation process. Negotiations were largely stalled up until the Vienna summit in November 19, 2013 between the presidents.

When the talks finally resumed in Vienna expectations were cautiously optimistic about the continuity this meeting could provide. Although nothing substantial emerged from that meeting, the presidents nevertheless committed themselves to re-launch of the direct negotiations. This in itself was already a positive sign after a long break. The next meeting of the presidents was said to take place sometime in the beginning of February 2014 after the preparatory meeting of the foreign ministers on January 24, 2014 in Paris. However, in consistence with the unlucky pattern, the cease-fire violations involving causalities started to surge in the frontline. The parties again got ensnared in mutual recriminations and blame game. The expected next meeting of the presidents did not take place.

Evaluation of OSCE mediation via theoretical lenses

Based on the description above, and using the theoretical assumptions of mediation theories, the evaluation of OSCE mediation efforts could be conducted. Scholars theorizing conflict mediation lay out quite a comprehensive approach that could be employed while analyzing this particular case. They define meditation as being successful when it has made a considerable and positive difference to conflict resolution and interaction between the parties. Mediation is considered to be partly successful when it has initiated dialogue and negotiations between the conflicting states, and it is deemed to be of limited success when it has achieved a cease-fire.[14]

Against the backdrop of the aforementioned assumptions and considering the overall picture of peace negotiations between the conflicting sides, the OSCE mediation could perhaps be labeled as being partly successful, since it has indeed succeeded in initiating a dialogue between the parties.[15] In terms of reaching a final solution though, the OSCE mediation efforts could not be seen as being successful. However, it goes without saying that possible success of the OSCE mediation activities hinges upon a number of different factors that could be explained using the conjectures of the mediation theory.

There are several factors that influence the success of mediation. For the purpose of analysis the most relevant ones will be focused on: characteristics of the parties (previous relations between the conflicting parties) [16], nature of the conflict and the strategies of mediation. Firstly, it is assumed that the previous relations between the conflicting parties matter in defining the success of mediation. If the previous relations between the warring sides were friendly, then there are more chances for the mediation to be successful than in the case if the parties had a troubled past.[17] While sketching the success of OSCE efforts from this prism, it is not difficult to imagine how the co-chairs are challenged by mediating between the parties that share a history filled with rancor and antagonistic mutual perceptions. The fact the Armenians and Azerbaijanis have a history of ethnic clashes complicates the achievement of rapprochement on the disputed issues thus diminishing the chances for successful mediation.

Secondly, the nature of the dispute is defined as one of the important factors impacting the outcome of the mediation. When vital interests such as sovereignty and territorial integrity are affected, it is much more difficult for the mediators to succeed. Ideology disputes, disputes over the issues of resources and ethnicity have greater possibility to be resolved through successful mediation than security and sovereignty disputes.[18] Indeed, looking through the prisms of this statement, it could be argued that success of mediation by the OSCE in the Nagorno-Karabakh conflict is to a large extent impeded by as many like to refer “intractable” nature of the conflict shaped by the acrimonious debate on the so-called conflict of two principles of international law: territorial integrity and the rights of peoples to self-determination. The Nagorno-Karabakh conflict is about hard security issue and involves vital interests such as sovereignty and territorial integrity. This for sure has an overwhelming impact on the OSCE potential to succeed in its mediation efforts.

Third element is the strategies of mediation. According to the authors, there are three main strategies that lead the mediation behavior: communication-facilitation strategies, procedural strategies and directive strategies. In communication-facilitation strategies mediators mainly execute the function of communication between the parties with a little control over the substance of the issues. In procedural strategies mediators have a more formal control and may define the technical factors such as mediation environment, number and type of meetings, agenda of the meetings, etc.. Directive strategies though, are more advanced in a sense that the mediators influence the content, substance as well as the process of the mediation. In directive strategies mediators provide incentives, issue ultimatums, offer rewards and punishments, and introduce new proposals. It is argued that the directive strategies are the most successful among others.[19]

Based on these assumptions, the OSCE mediation could be assessed positively as throughout its work the Minsk Group has used directive strategies by presenting various proposals to the parties. However, the group was only presenting the proposals to the parties’ perusal without any incentives, rewards, punishments, ultimatum, etc. as required by directive strategies, thus diminishing the effectiveness of this strategy. This actually evinces the gist of the problem with the OSCE mediation. The Minsk group has self-confined itself only to working out new deals and proposals without undertaking a responsibility for their acceptance by the parties. As the above description suggested, it has become a tradition that in case if any of the conflicting parties is not satisfied with the proposed deal, the OSCE would still leave it at their discretion, this time preferring communication-facilitation strategy, would wait for the parties themselves to find zones of agreement.

Therefore, in terms of providing the parties with food-for-thought and being involved in the substance of the negotiations, the OSCE could be considered as doing a successful job, which, however, is not a full success. As it was laid out above, the OSCE mediation efforts could be named as being partly successful, because of the absence of a final solution to the conflict. Years of mediation have been remarkable with directive strategies – the initiation of frequent and continuous bilateral contacts, preparation of blueprints for a future peace deal, –  which however have not yielded a final lasting solution to the conflict, because the OSCE lacks the capacity to influence conflicting parties’ behaviors and is fully dependent on their opinions.

There is also a factor of bias in international mediation, which cannot be overlooked while dealing with Minsk Group’s mediation activity. It is assumed that mediators in general have certain interests that motivate their involvement in conflict resolution. This also indicates that they have something at stake in the given conflict that derives from political, economic context, or close relationship with any of the conflicting sides. A biased mediator is thought to be the one who has something at stake or have preferential political, economic and other kinds of relations with one of the conflict parties. On the contrary, an impartial mediator is seemingly balanced or “neutral” in a sense that he or she has no opinion or preference in the conflict and its outcome. A biased mediator may not be the only available party to mediate, but also the one with greatest lever on the conflict side that needs to change the most.[20] There are suggestions that even the biased mediators can achieve a breakthrough despite of their biases, but different forms of mediator bias can influence the parties’ perception about the whole mediation process. The partial mediator can sometimes unleash a political process that possesses leverage, weight and counterweight, or carrots and sticks.[21]

Three Minsk Group co-chair states, Russia in particular, have not always been perceived as being non-tendentious. This perception was the strongest in Azerbaijan, who is suspicious of Russia’s extensive strategic and political alliances with Armenia, and believes that this special relationship undermines Russia’s image as an impartial mediator and translates into certain policy outcomes. However, Russia is not the only country in the Minsk Group Troika, who in one or other ways has strong pro-Armenian connections. Both the United States and France have very active and omnipresent Armenian Diaspora and lobby groups that extensively promote Armenia’s interests within these two states. Foreign policies of United States and France in regard to Nagorno-Karabakh conflict apparently do not happen in isolation from their domestic politics.[22] All three Minsk Group co-chairs have closer political, cultural relationship with Armenia than Azerbaijan, although Azerbaijan increasingly bears more appeal in economic terms. However, obviously it is hard to break the entrenched specificities of domestic politics and societies in all three countries from being translated into their foreign policies. Ironically though, despite the presence of bias on the part of the mediators, they are also the ones with the greatest lever on the conflict side that needs to change the most. This statement argues that with little more efforts the United States, Russia and France could have convinced Armenia to withdraw from the occupied Azerbaijani territories, thus creating the basic, but the most important condition for the advancement of peace process. However, in this concrete case, the mediator bias unfortunately works not for the sake, but against the peace process.

Summing up

In concluding the above analysis, it should once again be reiterated that the OSCE mediation efforts could be seen as partly successful. Although, the OSCE has the right mechanism to make difference in the conflict resolution process and its co-chairs have been highly engaged with mediation and facilitation work, the present format as such does not have the capacity and necessary tools to support attaining final and lasting peaceful solution to the conflict. Despite all the good work done so far, by extending the same treatment to both states – one being a victim of occupation, while other seeking to maintain its territorial acquisitions, – the OSCE runs the risk of losing its face as an impartial mediator. With matters remaining static in their present form, the likelihood that the parties can achieve negotiated conflict resolution based on norms and principles of international law as well as relevant UNSC resolutions unfortunately seems rather meager. In the given situation there are two options that would further define the fate of the peace process. In the first option, the OSCE continues to preserve the current discourse agitating a balanced and equal approach to both conflicting parties, thus maintaining dangerous and unproductive status quo for some more years to come. This of course would be the worst-case, however, the most likely scenario.

In the second option, however, the OSCE as a regional arrangement under Chapter VII of the UN Charter could rethink its long entrenched equal treatment of both parties and make appropriate adjustments to its position by placing primacy on the implementation of relevant resolutions of the UN Security Council on Nagorno-Karabakh conflict. Only then the OSCE could preserve its image as an impartial negotiator, and insure itself from reoccurring accusations blaming the organization for ineffectiveness. Peace and reconciliation are long overdue in the region torn by conflict and mistrust. Security and confidence building is not delivered automatically to those states in the most affected parts of Europe upon their membership in the OSCE. It takes hard work, dedication and well-thought strategies. The OSCE has a real chance to demonstrate that it is capable of continuously fostering those assets that its name stands for.


[1] However, the UN did speak up in the most active phases of the conflict in 1993 by adopting UN SC resolutions 822, 853, 874 and 884. All of those resolutions recognized territorial integrity of Azerbaijan and called for withdrawal of Armenian forces from the occupied Azerbaijani territories.

[2] Erjan Kurbanov, “The Armenian-Azerbaijani Conflict over Nagorno- Karabakh and Prospects for OSCE  Mediation,” in  S. Neil MacFarlane and Oliver Thranert, (eds.) Balancing Hegemony: The OSCE in the CIS, (Queen’s University Kingston, Ont. Center for International Relations, 1997):87

[3] At night from February 25 to 26 the Armenian armed forces occupied the Azeri town of Khojaly. The occupation was carried out with active support of several units of the Russian Army’s 366th regiment. In a few hours 613 innocent and unarmed were killed. Among them were 106 women, 83 children. 8 families were totally exterminated. 25 children were totally, and 130 children were partly orphaned. 476 people became disabled persons (of them 76 were minors). 1275 people were taken into hostage and even though afterwards most of the hostages were released from captivity, the fates of 150 of them are still unknown. (http://www.khojaly.org/)

[4] At the beginning the Minsk Group consisted only of the nine participating states, excluding Armenia and Azerbaijan, however, later in the year the two parties to the conflict joined the Group.

[5] Oliver Paye and Eric Remacle, “UN and CSCE policies in Transcaucasia” in Bruno Coppieters (ed.), Contested Borders in the Caucasus, (VUB University Press, 1996):119-120.

[6] In July 1993 Armenian forces captured further Azerbaijani territories, also the city of Aghdam, which was not the part of the Nagorno- Karabakh autonomous region.

[7]  David D. Latin and Robert Grigor Suny, “Armenia and Azerbaijan: Thinking a way out of Karabakh”, Middle East Policy, vol. VII, no.1(October 1999):159.

[8] Ibid., 160-161

[9] Statement of the OSCE Chairman-in-Office, Annex 1 to the Lisbon Document, December 2-3, 1996. From http://www.osce.org.

[10] Between 1992 and 1997 Italy, Sweden, Finland and Russia assumed the chairmanship.

[11] “Key texts and agreements from the Nagorno-Karabakh conflict and peace process” Conciliation Resources, available from (http://www.c-r.org/accord-article/key-texts-and-agreements-accord-nagorny-karabakh). Accessed on March 1, 2014.

[12] “Nagorno-Karabakh: Risking War,” International Crisis Group Europe Report, N 187, (November 14, 2007):2

[13] “Ibid.,5.

[14] Jacob Bertcovitch and Allison Houston, “The study of international mediation: theoretical issues and empirical evidence” in Jacob Bertcovitch (ed.) Resolving International Conflicts: The Theory and Practice of Mediation, (USA, Lynne Rienner Publishers, inc.1996):19

[15] I do not define it under the category of limited success. Since the cease-fire between the parties were negotiated mainly by individual actor, Russia, not by the OSCE

[16] In “Characteristics of the conflict” I will focus only on the previous relations between the conflicting parties, leaving aside parties’ political context and parties’ power, since they will require more in depth analyses, which is beyond the scope of this paper.

[17]  Jacob Bertcovitch and Allison Houston, 22

[18] Ibid., 24,25

[19] Ibid. 30

[20] Peter J. Carnevale and Sharon Arad, “Bias and impartiality in international mediation”, in Jacob Bercovitch (ed.), Resolving International Conflicts, The Theory and Practice of Mediation, (Lynne Rienner Publishers, USA, 1996):40-42.


[22] To bring a recent example, the resolution “Situation in the occupied territories of Azerbaijan” initiated by the Azerbaijan in 2008 during 62th session of the UN General Assembly about the fires in the occupied territories was voted against only by Armenia and three co-chair states: the US, Russia and France. The resolution was adopted; however the co-chairs substantiated their stance with their disagreement to bring the matters related to Nagorno-Karabakh conflict to the UN framework. But the mere act of their siding with Armenia in the voting for the resolution in fact testifies to the presence of a bias in their approaches to the conflict resolution.

  • January 31, 2021