© Koninklijke Brill NV, Leiden, 2010 DOI: 10.1163/157180610X488209
International Negotiation 15 (2010) 107–131 brill.nl/iner
Nagorno-Karabakh Negotiations:
Though the Prism of a Multi-Issue Bargaining Model
Taleh Ziyadov*
Azerbaijan Diplomatic Academy (ADA) and University of Cambridge
88 Shamil Azizbeyov Street, Baku, Azerbaijan AZ 1009
(E-mail: tziyadov@ada.edu.az)
Received 10 April 2009; accepted 15 June 2009
Abstract
This article examines various phases in the Nagorno-Karabakh (NK) negotiations between Armenia and
Azerbaijan using a Multi-Issue Bargaining Model – a modified version of the traditional bargaining
model. It offers micro-level and phase-by-phase analysis of the negotiation process, mediation efforts
and proposed mechanisms for the settlement of the NK conflict. Issues on the negotiation table and the
evolution of the Azerbaijani and Armenian positions over time constitute a central focus of the article.
The multi-issue model is applied to each negotiation phase in the NK conflict from 1994 until 2009.
Keywords
Azerbaijan; Armenia; Nagorno-Karabakh; bargaining model; bargaining space; Aliyev; Ter-Petrosian;
Kocharian; Sargsyan; negotiation; separatism
The Nagorno-Karabakh (NK) conflict between Armenia and Azerbaijan is an
interesting case study for those who deal with international negotiations. Since
the signing of the ceasefire agreement in 1994, the parties have witnessed interesting
turns and shifts in their negotiation positions. The process itself evolved
each time the negotiations were halted or collapsed as a result of domestic pressure,
resignation of the head of the state, coup d’état attempts, interference of a
third party or a shooting in the Parliament. Its gradual evolution provided
important lessons for the mediators and researchers of the NK conflict. As there
is no single formula to resolve all conflicts around the world, each conflict resolution
and its negotiation process leaves its own particular mark on the study of
international negotiation – so it is with the NK negotiations.
*)
Taleh Ziyadov is Assistant Dean of Academic Affairs at the Azerbaijan Diplomatic Academy (ADA)
in Baku, Azerbaijan. He holds a Master’s degree in Eurasian, Russian, and East European Studies from
the Edmund A. Walsh School of Foreign Service, Georgetown University. He is currently pursuing his
PhD degree in International Studies at the University of Cambridge (UK).
108 T. Ziyadov / International Negotiation 15 (2010) 107–131
Despite the fact that the actual hostilities ceased in 1994, the parties have yet
to find a mutually acceptable solution to the conflict. Armenian forces have
taken control over roughly 16 percent of Azerbaijan’s territory. The conflict
resulted in the ethnic cleansing of over a million civilians from their homes and
villages on both sides. It continues to be in a stage of unstable peace. The ceasefire
is maintained by Armenian and Azerbaijani military forces without the presence
of international peacekeepers. However, every year, as result of sniper
shootings or mines, several dozen soldiers and civilians die on both sides of the
front line. Numerous attempts by the NK separatists to legitimize their claims of
“unification with Armenia” or self-proclaimed “independence” have failed and
the province remains de jure a part of Azerbaijan, albeit de facto under the control
of the Armenian forces. Seven additional occupied regions of Azerbaijan
adjacent to NK are kept “hostage” as a “security” or a “buffer zone.”
Many articles on the NK conflict and its negotiation process have been policy
oriented and focused primarily on the empirical side of the conflict. There has
been very limited work done using theoretical models to examine the NK negotiation
process more systematically. This article will attempt to analyze various
phases in the NK negotiations using a version of the traditional bargaining
model. It will start with a short description of the mediation efforts in the NK
conflict and proposed mechanisms for settlement. It will state the known positions
of the two “principal parties” to the conflict, the Republic of Armenia and
the Republic of Azerbaijan, and the two “interested parties,” NK Armenians and
NK Azerbaijanis, on the disputed issues and provide descriptive analysis of various
negotiation phases in the NK peace process through to 2009. It will then
introduce the Multi-Issue Bargaining Model – a modified version of the traditional
bargaining model – and apply this model to each negotiation phase from
1994 until 2009.
Mediation Efforts and Mechanisms for Settlement
Since 1992, the conflict has been mediated under the auspices of what is today
the Organization for Security and Co-operation in Europe (OSCE) and its
Minsk Group, which consists of eleven member states and is currently co-chaired
by representatives from three countries: France, Russia, and the United States.
The mediation efforts began in March 1992 by the OSCE, then the Conference
on Security and Cooperation in Europe (CSCE) in the midst of armed clashes
that were taking place in the NK region of Azerbaijan. But the process itself was
chaotic at best. Each time the mediators tried to initiate a meeting between the
Armenian and Azerbaijani officials, one or the other region of Azerbaijan was
occupied by Armenian forces. By mid-1993, the Armenian forces had gained
important advantages and had taken over several regions of Azerbaijan that are
T. Ziyadov / International Negotiation 15 (2010) 107–131 109
adjacent to the NK region. Between mid-1993 and mid-1994, several counteroffensive
campaigns were launched by the Azerbaijani side to retake some of the
lost territories, but they resulted only in marginal territorial gains and came at a
heavy cost.1
By mid-1994, the current line of contact was established, which
resulted in the Armenian forces occupying Nagorno-Karabakh proper and an
additional seven adjacent regions of Azerbaijan (De Waal 2003: 292–295).
Over the years that followed, the OSCE Minsk Group has circulated a number
of proposals in efforts to resolve the conflict. Two of these proposals or methods,
commonly discussed during the NK negotiations, are presented below.
The first method is called a “step-by-step approach,” which comprises several
stages. The first stage consists of Armenia gradually pulling its troops from the
surrounding Azerbaijani areas outside NK (except for the Lachin corridor that
connects Armenia and NK). After this stage, certain security measures would be
taken, including the deployment of peacekeeping forces, to allow Azerbaijani
IDPs to return to the liberated lands and the two states and two communities to
re-establish communication links. The proposal leaves the discussions on the
final status of NK and the Lachin corridor to a later stage, after the implementation
of the above steps are completed. The main idea behind the step-by-step
proposal is to restore trust between Armenians and Azerbaijanis in hopes that
this will lead to a commonly agreeable compromise on the question of the status
of NK.
The “package approach,” on the other hand, is a proposal whereby all outstanding
issues pertinent to the resolution of the conflict are addressed at once,
which also includes determining the final status of NK before the Armenian
forces are pulled out of the occupied territories.
Parties to the Conflict and Their Positions
There are two “principal” and two “interested” parties in the NK negotiations.
The two “principal parties” are Armenia and Azerbaijan, while the two “interested
parties” are the Armenian and Azerbaijani communities of the NK region.
Also referred to as the “Baker Rules,” this view of the parties to the conflict,
which is shared by the OSCE Minsk Group today, was first proposed by U.S.
Secretary of State James Baker in 1992 as a compromise solution between Armenia
and Azerbaijan to the question of NK representation in the negotiations
(Baguirov 2008: 19–20). It is important to note that up to 1998, when the former
separatist leader of NK Armenians Robert Kocharian became President of
1)
From Fall 1993 until Spring 1994, Azerbaijan was able to retake parts of the Agdam, Fizuli and
Agdere (formerly Mardakert) regions, which were insignificant compared to the seven lost regions and
NK. During the 1994 winter campaign, Azerbaijan suffered heavy casualties, losing almost 1,500 soldiers
in an attempt to take back the Kalbajar region (De Waal 2003: 236–237).
110 T. Ziyadov / International Negotiation 15 (2010) 107–131
Armenia, the views of the “interested parties,” particularly NK Armenians, had
played a critical role in the negotiation process. Too often the views of NK
Armenians, who held a rather uncompromising stance, diverged from the position
of Armenia.
Armenia (Also Represents the Interests of NK Armenians)
Prior to 1998, Armenia had mostly favored a step-by-step approach to the NK
conflict resolution. But because this approach was delaying the decision on the
final status of the NK region, it ran counter to the position of the NK Armenians,
who demanded either unification of NK with Armenia or full independence
from Azerbaijan. Levon Ter-Petrossian, the former President of Armenia
(1991–1998) and the initiator of this approach, was unable to convince his
opponents and gather enough political support for the step-by-step resolution
method. As a result of growing criticism and pressure from the NK “clan” within
the Armenian leadership, he was ousted from his office in 1998. Since 1998, the
positions of Armenia and NK Armenians have more or less overlapped due to
the fact that the two subsequent Armenian presidents, Kocharian (1998–2008)
and Serge Sargsyan (2008–present), were originally from the NK region; they
have been representing the interests of NK Armenians in the negotiations under
the auspices of the Republic of Armenia.
NK Armenians vs. Armenia
NK Armenians have primarily insisted on the “package approach,” which since
1998 has also been adopted by Armenia. Today, neither Armenia nor NK Armenians
openly demand unification of NK with Armenia as they once did, and
instead seek international recognition for the self-proclaimed and unrecognized
NK republic. Armenia’s position on the package and step-by-step proposals has
evolved as well. In 2005, the Armenian Foreign Minister Vardan Oskanian and
the then Armenian Defense Minister and current President Serge Sargsyan,
addressed the Armenian National Parliament about the possibility of moving to
a step-by-step approach (Abasov and Khachatrian 2005).2
However, the Armenian
side still insists that NK cannot be vertically subordinated to official Baku.
In other words, Yerevan and NK Armenians seem to consider both approaches
as long as they both lead to de jure recognition of NK as an independent state
outside of the jurisdiction of Azerbaijan.
2)
Also see speeches of the former Armenian Foreign Minister Vardan Oskanian and the former Armenian
Defense Minister and current President Serge Sargsyan, at the parliamentary hearings in March in
Yerevan, Armenpress, March 29, 2005.
T. Ziyadov / International Negotiation 15 (2010) 107–131 111
Azerbaijan (Also Represents the Interests of NK Azerbaijanis)
Azerbaijan’s main position is based on the restoration of its territorial integrity,
though this argument seems to have been in question between 1998 and 2001,
when Baku and Yerevan were seriously discussing the possibility of territorial
swaps. At various phases of negotiation, Azerbaijan favored either a step-by-step
or a package approach depending on the implications of the proposed agreement.
For a brief period (1998–2001), the conflicting parties considered territorial
swaps, but later dropped this idea altogether. Azerbaijan’s statements on
granting NK the “highest level of autonomy” have not changed much. As long
as the parties can find a commonly acceptable solution within the territorial
integrity of Azerbaijan, Baku can agree to give as much self-rule to NK within
Azerbaijan as the Republic of Tatarstan has within the Russian Federation.3
As
for the question of the Lachin corridor, which connects Armenia to NK, Baku
believes that this issue could be resolved by obtaining similar access to a road
connecting Azerbaijan proper with its enclave, Nakhichevan, via Armenia.
NK Azerbaijanis vs. Azerbaijan
The views of the NK Azerbaijanis are essentially identical to those of official
Baku. The majority of the internally displaced population of Azerbaijan (570,000
in total) is from the seven occupied regions around NK and between 42,000 and
46,0004
(Yunusov 1998) Azerbaijani IDPs are from the NK region itself
(UNHCR 2005 and IDMC 2007). The primary demand of the Azerbaijani
IDPs, including NK Azerbaijanis, is the liberation of all occupied territories
3)
The idea of granting the NK region of Azerbaijan the equivalent status of Russia’s Republic of Tatarstan
was first voiced by the late President of Azerbaijan, Heydar Aliyev, during the first trilateral meeting
between Aliyev, Ter-Petrosian and then Russian President Boris Yeltsin in 1994 in Moscow. When Aliyev
said that Baku would consider granting NK the status equivalent to that of the Republic of Tatarstan,
Ter-Petrosian was amused and asked in disbelief “Can you really grant the status of Tatarstan to
Nagorno-Karabakh?” From a personal interview by the author with Vafa Quluzade, former senior advisor
to Heydar Aliyev (March 2009).
4)
The 1979 Soviet census, which is considered to be the most reliable, puts the population of the NK
region at 162,000 (123,000 Armenians (76%) and 37,000 Azerbaijanis (23%)). The last official 1989
Soviet census, which was disputed by Azerbaijan, puts the number of Azerbaijanis living in the NK
region at about 42,000, constituting 23% of the total population of the region. The Armenian population
was shown to account for 77% of the region. Azerbaijan disputed the result of this census based on
the previous census statistics of 1939, 1959, 1970 and 1979, which showed a steady increase in Azerbaijani
population and a gradual decline in Armenian population, from 88% in 1939 to 76% in 1979. In
October 1990, Azerbaijan conducted another census which included all 51 major Azeri towns and villages
inside the NK region. The new results declared the number of NK Azerbaijanis at about 46,000,
representing 24% of the total population of the NK region (Yunusov 1998). Today, all 46,000 NK
Azerbaijanis live as IDPs outside of the NK region and the more realistic number of NK Armenians living
in the region is estimated between 60,000 and 80,000 (De Waal 2003: 285). For more information
about the number of refugees and Internally Displaced Persons (IDPs) in Azerbaijan also see UNHCR
(2005), available online at http://www.unhcr.org/refworld/docid/442d369d4.html, and IDMC (2007),
available at http://www.unhcr.org/refworld/docid/460bd4512.html.
112 T. Ziyadov / International Negotiation 15 (2010) 107–131
including the NK region. NK Azerbaijanis state that the final status of NK
should take into account their views and, therefore, their return to the NK
region should precede any talks about the final status of the region. This view
also represents the official Azerbaijani position. Since 2003, Baku has firmly
argued that all outstanding issues, including the determination of the final status
of NK, can only be resolved gradually. In other words, today Baku considers the
package proposal unattainable and favors the step-by-step approach so long as
the solution of the conflict is within the territorial integrity of Azerbaijan, which
is to say no de jure separation of NK from Azerbaijan.
Phases in the NK Negotiation Process
Both step-by-step and package methods, as well as other proposals by the OSCE
Minsk Group mediators, have been developed and evolved over time at various
phases of negotiations. In general, the stages of the NK negotiations can be
divided into a Pre-Negotiation Phase (1992–1994); First Phase (1994–1998);
Second Phase (1998–1999); Third Phase (1999–2004); and Fourth Phase
(2004–2009).
Pre-Negotiation Phase: 1992–1994 (May) – Imitation of Negotiations
The period from 1992 to 1994 was marked with an escalation of the conflict
between Armenia and Azerbaijan. By the time the ceasefire was signed in May
1994, the Armenian forces were in control of some 16 percent of Azerbaijan,
including the additional seven regions outside of NK. Nonetheless, the meetings
between the Armenian and Azeri officials continued to take place even in the
midst of full scale military operations.
In 1992, as clashes between Armenian and Azeri forces started to spread inside
the NK region, the then Acting President of Azerbaijan, Yagub Mammadov, and
the then President of Armenia, Levon Ter-Petrosian, met in Tehran in an attempt
to sign a framework document for a peace deal. On May 9, 1992, the two presidents
inked a document on the general principles of the peace accord. That very
day, the major Azeri city inside of the NK region, Shusha, was occupied by the
Armenian forces. This was a huge embarrassment for Ter-Petrosian, who seemed
to want to end this conflict peacefully and as soon as possible (De Waal 2003:
180). There are divergent views as to why the meeting in Tehran ended in fiasco.
One argument is that it was done to undermine Ter-Petrosian, who arguably was
not in control of the situation inside of the NK region and allegedly parts of
Armenia’s military and intelligence apparatus.5
Other views focus on the com-
5)
Quluzade claims that Ter-Petrosian did not have control over the NK Armenians, and some of his
own military and intelligence units inside Armenia, which still received directions from Moscow, rather
than Yerevan. From the interview with Vafa Quluzade.
T. Ziyadov / International Negotiation 15 (2010) 107–131 113
peting mediator agendas and their conflicting interests, particularly of Russia,
which tried “to establish a monopoly on the right to lead, mediate and control
the peace process” (Zulfuqarov 2005:39). Hence, capturing the Azeri town of
Shusha on the same day as signing the peace accord in Tehran may have been an
attempt to undermine Ter-Petrosian’s position inside Armenia and the Iranian
mediation efforts in general, by “teaching” the latter a lesson to stay out of the
mediation process (De Waal 2003: 180).
Similar incidents took place during other meetings as well. The two sides
would be discussing a ceasefire or a timetable for the withdrawal of the Armenian
forces from recently occupied Azeri regions, and the news of fresh Armenian
advancement would shock the attendees. This was the period of “imitation
of negotiations.” While Baku and Yerevan may have had genuine interest in finding
a settlement to the conflict, they nonetheless also took advantage to
strengthen their negotiation positions with new territorial advancements on the
ground (Interview with Vafa Quluzade). In addition, the OSCE mediation
efforts were inconsistent and internal competition between the mediating parties
made the process ineffective. As Ter-Petrosian later acknowledged, prior to 1996
the OSCE mediation “was simply a bluff [and] there was absolutely no peace
process... [the mediators] competed among themselves more than they thought
about the Karabakh issue” (De Waal 2003: 229–230).
Phase 1: 1994 (May) – 1998 (February) – A Missed Opportunity
Serious negotiations started only after the Moscow-brokered ceasefire agreement
entered into force at midnight on May 11–12, 1994, shortly after the parties
agreed on the Bishkek Protocol, drafted on May 4–5 in Bishkek, Kyrgyzstan,
which “called on all conflicting sides to heed again the voice of reason: to ceasefire
at midnight on 8 to 9 May” (De Waal 2003: 238). The period from the signing
of the ceasefire accord until the resignation of Ter-Petrosian in February
1998 was the time when two leaders, Ter-Petrosian and Heydar Aliyev, who
was elected president of Azerbaijan in October 1993, started “a joint diagnosis
of the common problem,” in an “effort to identify a formula to solve that problem”
(Hopmann 1995: 41; Zartman and Berman 1983). An analysis of their
behavior and statements shows that the two presidents adopted a problem-solving
approach (Hopmann 1995) to the settlement of the Nagorno-Karabakh
conflict.
Ter-Petrosian was in an advantageous position at the negotiation table. Armenia
was in control of NK and the additional seven adjacent regions of Azerbaijan.
Yerevan also managed to close the military gap with Azerbaijan, owing to
Russia’s generous one billion dollars worth of arms shipments to Armenia
between 1992 and 1994 (De Waal 2003: 199). The urgency for settlement and
socio-economic pressure from some 800,000 internally displaced persons and
refugees in Azerbaijan also played into the hands of Armenia.
114 T. Ziyadov / International Negotiation 15 (2010) 107–131
Yet, two of Armenia’s four international borders, with Turkey and Azerbaijan,
were closed as a result of the war (the border with Iran and Georgia remained
fully open). The economic situation inside the country was dire and Armenia
was left out of ongoing energy projects in the region. In addition, Azerbaijan’s
moves on the diplomatic front started to pay off by gaining the support of international
organizations, which complicated Armenia’s position. Both sides had
reached what Touval and Zartman call a “hurting stalemate” (Zartman 2008: 232).
During the OSCE Lisbon Summit in December 1996, the OSCE Chairmanin-Office
outlined three main principles, later known as “the Lisbon Principles,”
which were to form part of the conflict settlement. These were:
1) Territorial integrity of the Republic of Armenia and the Republic of Azerbaijan;
2) Legal status of Nagorno-Karabakh defined in an agreement based on the principle of selfdetermination,
which confers on Nagorno-Karabakh the highest degree of self-rule within Azerbaijan
(emphasis mine);
3) Guaranteed security for Nagorno-Karabakh and its whole population, including mutual obligations
to ensure compliance by all the parties with the provisions of the settlement (Abasov and
Khacatrian 2005: Appendix 15).
In 1997, following the Lisbon Summit, the OSCE Minsk Group was re-organized
and began to mediate with three co-chairs from France, Russia, and the
United States. It was after that when the co-chairs came up with two proposals:
the so-called “package deal” and a “step-by-step”6
or “stage-by-stage” proposal.
The former included a peace agreement between Armenia and Azerbaijan and a
separate agreement on the status of NK, but based on the Lisbon Principles.7
The latter envisaged a gradual settlement whereby, in the beginning, Armenia
would pull out its troops from the six occupied Azerbaijani regions (except from
the Lachin corridor that connects Armenia to NK), peacekeepers would be
deployed in the conflict zone, IDPs would return to their homes, communications
would be restored and confidence building measures would take place.
Only after these steps were taken, would the parties move to the second stage
and discuss issues related to the Lachin corridor, the return of Azeri IDPs to NK,
the fate of the Azeri town of Shusha and the final status of NK.8
6)
In fact, the idea of a step-by-step or gradual settlement was first proposed before 1997 by Armenian
President Levon Ter-Petrosian. Azeri President Heydar Aliyev responded positively to this call, though
Baku may have played its oil card and at times opted for a package deal in hope that oil revenues would
strengthen its negotiation position in the future.
7)
According to former Armenian Foreign Minister Vardan Oskanian, the package proposal in 1997 was
based on the Lisbon principles and that is why the Armenian side rejected it. See the speech of former
Armenian Foreign Minister Vardan Oskanian at the parliamentary hearings in Yerevan, Armenpress,
March 29, 2005.
8)
See “OSCE Minsk Process on the Settlement of the Nagorno-Karabakh Conflict between Armenia
and Azerbaijan” brief, available on the website of the Azerbaijan Embassy to Spain at www.azembajada.es/.
T. Ziyadov / International Negotiation 15 (2010) 107–131 115
There are different opinions among Armenian and Azerbaijani officials on the
then positions of the two sides with respect to a package deal. Since the package
proposal presented in May and July 1997 was based on the Lisbon Principles,
Baku considered it an “acceptable starting-point” (Zulfuqarov 2005: 40). Later,
it would also prefer a step-by-step proposal. Yerevan, on the other hand, mainly
favored a step-by-step approach but had serious disagreements with NK Armenians
on this issue, who favored a package deal where the status of NK must be
an “independent state” and nothing less. Yerevan’s negotiation strategy was based
on the return of the occupied territories in exchange for peace, which would
provide strong security guarantees and self-rule for NK Armenians. This
approach did not overrule a possible agreement within the territorial integrity of
Azerbaijan. NK Armenians, on the other hand, based their position on a “territories
for status” formula, which is to say the return of the occupied territories in
exchange for de jure separation of NK from Azerbaijan (Libaridian 2005: 36).
The NK Armenians rejected the package proposal presented in May and July
1997 because it was not the package deal for which they were hoping, although
both Baku and Yerevan thought it was acceptable with some reservations and
could be improved with further negotiations (Zulfuqarov 2005; Libaridian
2005). In the absence of support from NK Armenians, the proposal failed to
make any progress.
Following the package proposal, the OSCE Minsk Group came up with a
step-by-step proposal in September 1997, which was acceptable to both Baku
and Yerevan, but again not to NK Armenians who insisted on a package deal
that would guarantee the separation of NK from Azerbaijan. Due to the controversial
nature of the topic and as a compromise, Ter-Petrosian wanted to postpone
the discussion on the final status of NK to a later stage, and the step-by-step
proposal seemed to be the right mechanism to allow this to happen. By October
1997, he was ready to sign a peace deal envisaging a gradual settlement of the
conflict. But Ter-Petrosian’s decision resulted in internal divisions within the
Armenian political spectrum. Even his close allies, such as Yerevan Mayor Vano
Siradeghian and the then Foreign Minister Alexander Arzumanian, resigned and
some 40 out of 96 delegates from the ruling party defected to the opposition
(Croissant 1998: 123). Some members of his own cabinet started to demand his
resignation, which he submitted on February 3, 1998.
In general, during the 1994–1998 period, the final status of NK was still
viewed as a matter of self-determination and territorial integrity of Azerbaijan.9
As Girard Libaridian, the former senior adviser to Ter-Petrosian, stated:
9)
For example, the former Ambassador of Armenia to the United States often talked about the concept
of territorial integrity and once stated that the solution to the NK conflict “should be based on the
“right to self-determination and territorial integrity and should be an outcome of peaceful negotiations.”
The Washington Times, December 9, 1995.
116 T. Ziyadov / International Negotiation 15 (2010) 107–131
Ter-Petrosian sought a compromise where the Armenian side would concede that [NK] would be
legally part of Azerbaijan; in return Azerbaijan would agree to a status above the nominal autonomy
that the NKAO [Nagorno-Karabakh Autonomous Oblast/Province] had enjoyed until 1988,
but a notch below independence (Libaridian 2005: 36).
Because the parties desired to find a compromise solution quickly, there was real
bargaining space that could have led to a long-term peace in the region. However,
the concept of territorial integrity was buried soon after Ter-Petrosian was
forced to resign in February 1998. This was the end of an era of missed opportu-
nities.
Phase 2: 1998 (March)–1999 (October) – “Karabakhization” of Armenia’s Position
The period after the resignation of Ter-Petrosian witnessed some major shifts in
the negotiation positions of Armenia and Azerbaijan. In March 1998, Robert
Kocharian, native of the NK region and the former secessionist leader, was
elected President of Armenia. Kocharian’s coming to power was an indication
that from there on the interests of NK Armenians at the negotiation table would
be represented by Armenia directly.10
Unlike Ter-Petrosian, Kocharian refused to delay the decision on the final status
of NK and insisted on adopting a package approach. Likewise, Kocharian
changed the strategy in the negotiations, shifting it from a “problem-solving
approach” to a “bargaining approach,” thus adhering to a more hard-line, maximalist
position using distributive bargaining (Walton and McKersie 1965; Hopmann
1995: 27).
The major shift was related to the status of NK, which Kocharian insisted
could not be left within Azerbaijan. As former Foreign Minister of Azerbaijan,
Tofik Zulfuqarov stated:
The basic premises of [Kocharian’s] standpoint [were] that: the liberation of part of the territories
around [NK was] possible only after Azerbaijan’s agreement to the independence of [NK]; and the
territory situated between... [NKAO-Nagorno-Karabakh Autonomous Oblast/ Province] and
Armenia must be given to Armenia. This strategy [was] clearly intended to reduce Azerbaijan to
capitulation, an outcome seen by the Armenian side as the logical consequence of its own military
victory (Zulfuqarov 2005: 41).
10)
Prior to the Lisbon Summit in December 1996, the representatives of both Azeri and Armenian
communities of NK participated in the plenary Minsk Group discussions as “interested parties.” From
thereon, Azerbaijan insisted that all negotiations should be based on the Lisbon principles and should
be conducted directly between Baku and Yerevan without participation of interested parties. The broader
format of talks, including the Azeri and Armenian communities, is envisaged at a later stage in the negotiations
conditioned to the Armenian troop pullout from the occupied territories of Azerbaijan. From
the personal interview by the author with Deputy Foreign Minister of Azerbaijan, Araz Azimov (March
2009).
T. Ziyadov / International Negotiation 15 (2010) 107–131 117
Therefore, any proposals or options considering a solution within the territorial
integrity of Azerbaijan became unacceptable to the Armenian side. It was during
this time that Russia revealed the so-called “Primakov Plan” on a “common
state” solution, which was proposed not only to Armenia and Azerbaijan but
also to Georgia and Moldova (Zulfuqarov 2005: 41). The basic idea was that the
secessionist units within Azerbaijan, Georgia (i.e. the proposal applied to Abkhazia
only) and Moldova would form together with their respective host states a
“single state” with internationally recognized borders. But this idea lived only for
five months (March–December 1998) and was rejected by Baku (Zulfuqarov
2005: 41).
Yet, in 1999, Heydar Aliyev seemed to soften his position and made a controversial
move that would be considered a radical shift in Azerbaijan’s negotiation
strategy. The shift in Heydar Aliyev’s position became known to the public only
a few years after the collapse of the 1999 negotiations. It was the time when the
parties started to discuss the territorial swaps, also known as the Goble Plan,
named after U.S. State Department regional expert, Paul Goble (De Waal 2003:
263). The plan suggested an exchange of corridors (Lachin and Meghri) that
would link Armenia to NK and Azerbaijan to its Autonomous Republic of
Nakhichevan.
While internally there had been some discussions about possible exchanges of
equal sized territories between Armenia and Azerbaijan, this was by no means an
exchange of that sort. The content and direction of negotiations envisaged the
eventual political separation of NK from Azerbaijan – a painful concession that
Aliyev had to make if he wanted to sign any potential agreement with Kocharian.
In exchange, Armenia would liberate all occupied territories around NK
(except for the Lachin corridor) and Baku would get a strip of territory in Armenia
which would connect Nakhichevan with Azerbaijan proper (Interview with
Vafa Quluzade). This was indeed a radical shift in Baku’s original position, which
was to resolve the dispute within the territorial integrity of Azerbaijan.
The negotiations were taking place in strict secrecy and the two presidents
were meeting regularly, often tet-a-tet, on their own initiatives. The plan was to
sign a peace accord during the OCSE Summit in Istanbul in November 1999.
But in October 1999, three senior Azerbaijani officials, including Aliyev’s longtime
foreign affairs advisor Vafa Quluzade, the Foreign Minister Tofik Zulfuqarov,
and the head of Aliyev’s secretariat, Eldar Namazov, resigned from their
posts in apparent protest and disagreement with this move (De Waal 2003: 264).
Aliyev was shaken but unmoved and still thought that a breakthrough would be
possible during the OSCE Istanbul Summit in November. Quluzade did not
share his optimism.11
Soon everyone would understand why.
11)
Vafa Quluzade was the first to resign followed by the resignations of Zulfugarov and Namazov three
weeks after. In his conversations with Heydar Aliyev, before and after his resignation, Quluzade stated
118 T. Ziyadov / International Negotiation 15 (2010) 107–131
On October 27, 1999, a group of armed men entered the Armenian Parliament
and shot dead eight people, including Prime Minister Vazgen Sargsyan and
the Speaker of the Armenian Parliament, Karen Damirchyan. These two individuals
were the most influential Armenian politicians at the time. Two of Kocharian’s
political opponents and potential challengers died, but so did the idea of
signing a peace accord during the OSCE Istanbul Summit in November 1999.
As Defense and Prime Minister, Sargsyan’s support was crucial for Kocharian
if Baku and Yerevan were to sign any peace agreement on NK (De Waal 2003:
265–266). Sargsyan seemed to have agreed to support the agreement that was to
be signed in November 1999. U.S. Deputy Secretary of State Strobe Talbott,
who held discussions with Sargsyan in Yerevan just before the latter was assassinated,
also confirms that “the deal was very much within reach.”12
While some
may have considered the idea of territorial swaps unworkable from the start, and
no one knows what would have been the fate of an agreement proposing territorial
exchanges, it is nonetheless important to acknowledge that the two parties
were ready to sign such an agreement in November 1999.
Some Armenian and Azeri sources allege that the shooting in the Armenian
parliament was orchestrated by the Russian security services to prevent the signing
of the peace accord during the Istanbul OSCE Summit in November (Safaryan
2005).13
While this assertion may lack clear evidence, it does show, however,
how Russia is often perceived as a third party mediator in Armenia and Azerbaijan.
As Hopmann notes “powerful third parties may attempt to influence the
outcome of the negotiation to serve their own national interests rather than the
interests of the disputed parties” (Hopmann 1996: 227). Losing the land connection
to Iran via Armenia, as a result of territorial swaps between Armenia and
Azerbaijan and fears of possible integration of both states into the Euro-Atlantic
organizations may have appeared to be threatening Russia’s regional and security
interests at the time.14
But as De Wall describes Moscow’s involvement in the
NK peace process, “The crux of the problem was that Russia was both involved
in the conflict and also its only serious mediator” (De Waal 2003: 233).
that some elements in the peace deal ran against the national interests of Azerbaijan. But more importantly,
he believed that resolving a conflict between Armenia and Azerbaijan at the time was not in the
interest of Moscow, hence, he always thought that the signing of this agreement would not take place.
From the interview with Vafa Quluzade.
12)
On October 27, 1999, U.S. Deputy Secretary of State Strobe Talbott was in Yerevan meeting with
Armenian officials, including Vazgen Sargsyan. Talbott states that he believed “the deal was very much
within reach, and that Sargsyan’s apparent willingness to support it, indicated in that meeting, was crucial.”
From the author’s personal communication with Strobe Talbott (May 2009).
13)
In his paper, Safaryan refers to Col. Aleksandr Litvinenko who claimed that the shooting was allegedly
organized by the Russian security services to prevent the signing of the peace deal. See also (Zyberk
2007). From the interviews with Vafa Quluzade and a senior Armenian official who held a key government
post in Armenia at the time.
14)
Security of Russia’s southern flank or its perceived interests in the South Caucasus may have
demanded to keep the status quo and both Armenia and Azerbaijan insecure.
T. Ziyadov / International Negotiation 15 (2010) 107–131 119
In general, this was the period when both sides seemed to have made some
fundamental shifts in their previous positions and came close to signing an
agreement, the implementation and success of which was very much in question.
Phase 3: 1999–2004 – Moving to Stalemate
The consequences of the October 1999 shooting in the Armenian parliament
had serious implications for the subsequent negotiations. Kocharian was less
cooperative referring to the fragile security situation inside Armenia.15
He also
retreated from the initial idea of exchanging the land corridors, downgrading it
only to the right of access. Meanwhile, Heydar Aliyev started to experience serious
health problems and began to think about succession issues.
The most serious talks between Kocharian and Aliyev were held in Paris and
Key West, Florida in 2001. With the help of the US Secretary of State Colin
Powell, the two presidents had the “most high-profile and intensive negotiations
ever on the dispute” in Key West in April 2001 (De Waal 2003: 267). Aliyev
seemed to compromise on some of the issues regarding NK’s status, but probably
understood that it would be impossible to sell this deal domestically. Not
long after his return from the Key West, Aliyev denounced any compromises.
All hopes for a possible agreement were demolished when he died at the age of
80 in December 2003.
Phase 4: 2004–2009 – Back to Territorial Integrity: Prague Process and Basic Prin-
ciples
The last or present phase of the NK negotiations began with Ilham Aliyev, son
of the late President Heydar Aliyev, who became the President of Azerbaijan in
October 2003. Contrary to his father the younger Aliyev embarked from a more
traditional position based on the supremacy of international law and inviolability
of Azerbaijan’s territorial integrity. He stated many times that Azerbaijan was
willing to wait and restore its territorial integrity by any means, including using
force, if necessary. He denounced any Key West documents that had been discussed
between his father and Kocharian, promised only the highest level of
autonomy to the NK region, and tried to internationalize the conflict, moving
the discussion into the UN General Assembly format.16
The new phase in negotiations started in April 2004 when the Foreign Ministers
of Armenia and Azerbaijan met in Prague to re-start the talks. Following the
initial meeting in Prague, the two ministers as well as the two presidents had
15)
Interview with Araz Azimov.
16)
In the 59th UN General Assembly, Azerbaijan introduced a resolution on the issue of illegal resettlement
of Armenians in the occupied territories controlled by the Armenian forces. Later, as a compromise,
the parties agreed to send the OSCE fact-finding mission to the conflict zone to determine the
number of illegal settlements and resolution was withdrawn.
120 T. Ziyadov / International Negotiation 15 (2010) 107–131
conducted a series of major talks in 2004 and 2005. This process became known
as the Prague Process (Mehdiyev 2005). In August 2005, the Minsk Group cochairs
introduced another concept of “interim status,” or what Azimov calls an
“interim regime.”17
Finally, in 2007, a new proposal called the “Basic Principles
for the Peaceful Settlement of the Nagorno-Karabakh Conflict” was introduced
(OSCE 2007; OSCE 2008).
The most recent phase in negotiations was structured to tackle more general
issues first and then move to discuss more complex topics and specific details,
what Zartman calls moving from “formula to detail” (Zartman and Berman
1983: 135–138, 146–149). This approach is a combination of a package
approach and a step-by-step approach. In other words, if Baku and Yerevan can
agree on “Basic Principles” of the settlement, which includes agreeing on all contentious
issues, including how to determine the final status of NK, the agreement
could be implemented in phases.
As in the previous talks, the major issues on the table now are the same except
for the two new concepts of an “interim regime” and a “popular vote” or “referendum.”
Other issues include a gradual withdrawal of Armenian forces from
seven occupied regions of Azerbaijan (except for a transit corridor/road in
Lachin), the demilitarization and deployment of peacekeeping forces to the conflict
zone, the return of IDPs to liberated territories, the restoration of communications
and transportation links between the parties, and other confidence
building measures (Interview with Araz Azimov).
The concept of an “interim regime” or an “interim status” refers to a temporary
status that could be granted to NK until its final status is determined. The
concept of a “popular vote” or “referendum” implies a public vote inside NK at
some future point in time, where both Armenian and Azeri communities would
vote to determine the final status of NK (Interview with Araz Azimov). The
positions of Baku and Yerevan vary significantly on these two concepts. Yerevan
wants to make sure that the “interim status” for NK would be “international,”
similar to the one granted to Kosovo before the province declared its independence.
Baku, on the other hand, views this timeframe as an “interim” or “temporary”
regime rather than an “international status” (Interview with Araz
Azimov). Similarly, Armenia demands a “referendum” inside NK whereby only
Armenian residents of NK could vote on NK’s final “political” status, which may
include an independence option as well. Azerbaijan, however, demands the
return (not “the right of return”) of Azeri IDPs to NK prior to any popular vote,
17)
This concept is based on the idea that Azerbaijan could recognize the de facto present-day situation in
NK until the final status of the region is determined in exchange for Armenia liberating several occupied
Azeri territories outside of NK. Azimov states that this “interim regime” or “interim status” by no means
refers to “interim international status” and that by the final status Baku means “legal” and not “political”
status of NK. Yerevan has its own interpretation of this interim regime, which might be different from
the one above. From personal interview with Deputy Foreign Minister Araz Azimov.
T. Ziyadov / International Negotiation 15 (2010) 107–131 121
which can only determine the “legal status” of NK within Azerbaijan and not a
“political” one. In short, Yerevan seeks independence for NK, even if a delayed
one, while Baku insists on a resolution of the conflict within the territorial integrity
of Azerbaijan.
Traditional Bargaining Model vs. Multi-Issue Bargaining Model
To visualize the above descriptive analysis of the NK negotiation process since
1994 and better understand the positions of the parties and the issues on which
they agreed and disagreed, a modified version of the traditional bargaining
model will be used – a Multi-Issue Bargaining Model. A Traditional Bargaining
Model (Hopmann 1996: 55) (see Fig. 1) is used to illustrate the positions of the
two negotiating parties on a particular issue with their preference curves, preferred
and minimum acceptable outcome points, and gains and losses relative to
nonagreement. If the range of mutually acceptable agreements overlaps, it creates
a “bargaining space” for mutually acceptable agreement. If not, there is no
“bargaining space,” and hence no agreement can be achieved.
The traditional bargaining model is helpful in demonstrating the maximum
and minimum positions of the parties, their gains and losses, and whether or not
there is a bargaining space which could result in an agreement. But many times
the negotiations are conducted around multiple issues which together form a
single agreement. If there is no bargaining space and the parties cannot come to
a mutually acceptable agreement, one cannot really examine what were the particular
issues that led to the collapse of the potential agreement, if any. In other
words, the traditional bargaining model is limited in representing the disputed
issues on the negotiation table and the level of difficulty or contention between
the parties with regards to a particular issue. Hence, a revised model is used to
Figure 1. A traditional two-party bargaining model.
1. Horizontal Axis = Issue Dimension
2. Vertical Axis = Gains (+) and Losses (-)
Relative to Nonagreement (0)
3. A-A’ = A’s Preference Curve
4. B-B’ = B’s Preference Curve
5. a = A’s preferred outcome
a’ = A’s minimum acceptable outcome
b = B’s preferred outcome
b’ = B’s minimum acceptable outcome
6. E = point of “equitable” solutions,
where gains of both parties relative to
nonagreement equal.
7. S = Bargaining Space
122 T. Ziyadov / International Negotiation 15 (2010) 107–131
illustrate all outstanding issues in the negotiations, the level of difficulty/contention
of each issue, and on which issues the parties have had a mutually acceptable
agreement (bargaining space) during various phases of negotiations (see Fig.
2). The bargaining space in this multi-issue bargaining model does not imply
that the parties have already agreed on the issues within the space, but rather
that these issues are not seen as non-agreeable by both sides and both parties
have had mutually acceptable agreements concerning these issues during various
stages of negotiations. It is also possible to show a preference curve of more than
two negotiating parties in this model, which will be demonstrated below.
1. Horizontal Axis = Issues Dimension /
Gains (+) & Losses (-)
2. Vertical Axis = Difficulty/Contention
Level (Left Vertical Axis represents A’s
approach to a particular issue, Right
Vertical Axis represents B’s approach to a
particular issue. The model assumes that
the approach of A and B is equal with
respect to any issue).
3. 0b– A’ = A’s Preference Curve
4. 0a – B’ = B’s Preference Curve
5. 0a = A’s preferred outcome
0b = B’s preferred outcome
a1,a2,a3... an
= A’s minimum acceptable
outcome on issues A1,2,3... n
b1,b2,b3... bn
= B’s minimum
acceptable outcome on issues B1,2,3... n
6. A1–B1= Issue #1
7. A2–B2= Issue #2
8. A3–B3= Issue #3
9. An-Bn= Issue #n (The most contentious
issue)
10. A1,2,3... n = A’s preferred outcome on issue
A1,2,3... n
11. B1,2,3... n = B’s preferred outcome on issue
B1,2,3... n
12. S = Bargaining Space for issues #1,2,3
13. There is no mutually acceptable
agreement on issue #n, since A (An- an
)
and B (Bn- bn
) do not overlap at any
point.
Figure 2. A simple multi-issue bargaining model.
T. Ziyadov / International Negotiation 15 (2010) 107–131 123
A Multi-Issue Bargaining Model
Applying the Model to the NK Negotiations
The multi-issue model will be applied to the NK negotiations in phases and it
will reflect the changes in the positions of the parties and their gradual evolution.
But prior to using this model for the NK negotiations, it is important to
define the issues that have been the subject of negotiations since 1994.
For the purpose of better demonstrating the availability or lack of bargaining
space, the model assumes that “the level of difficulty or contention” on any issue
is equal for both parties. However, it is possible to illustrate that particular issues
are more critical or contentious for one party than for the other. The difficulty
level for an issue listed below will be shown by a number assigned to each letter.
For instance, the parties have more chances to come to a mutually acceptable
agreement on the issue (A1–B1) than on the issue (A7–B7). Because it is difficult
to assign the exact weight for each issue and some issues may be interlinked and
have similar difficulty or ease, these representations are only relative and approximate.
If an issue is assigned an additional small letter to the number (i.e. A3bB3b),
it means this particular issue was not on the negotiation table during a
certain phase and was introduced later on as a change in negotiation tactics. It is
also essential to acknowledge that some issues like security guarantees, the
make-up of peacekeeping forces, blockades and severed communication links,
demilitarization of the conflict zone, and others, while important, have been
nonetheless discussed in conjunction with other more central issues and have
not been viewed as very contentious. Thus, they are either shown under a single
category or presumed to be a part of another issue listed in Table 1 below.
The list of selective issues used in the model for various phases of NK negotiations
is as shown in Table 1.
Phase 1: 1994 (May) – 1998 (February)
As previously discussed, Phase 1 of the NK negotiations was probably the most
encouraging in terms of the opportunities and the problem-solving approach
that both sides had with respect to urgent resolution of the conflict. Fig. 4 shows
that both Armenia and Azerbaijan had a common understanding on all issues on
the negotiation table.
Even on the most contentious issues like the status of NK (A6–B6), the return
of Azeri IDPs to NK(A5–B5) and the status of the Lachin corridor (A5–B5), the
two sides held common ground whereby they were willing to consider a step-bystep
approach to rebuild the trust between Armenians and Azerbaijanis and
gradually pave the way for a long-term peace in the region. The preference curve
of the Armenian separatists in NK (Line C (NK)), however, differed significantly.
The NK Armenians were only willing to compromise on issues of the six
124 T. Ziyadov / International Negotiation 15 (2010) 107–131
Table 1. Selective Issues in the NK Negotiations
A1–B1 – Issue #1: Security Guarantees (including peacekeeping force) and
Demilitarization of the Conflict Zone
A2–B2 – Issue #2: The return of Azerbaijani IDPs to six regions outside of
NK (except for Lachin)
A3–B3 – Issue #3: The pullout of Armenian forces from six regions of
Azerbaijan outside of NK (except for Lachin)
A3a–B3a – Issue #3a: The pullout of Armenian forces from five regions
of Azerbaijan outside of NK (except for Lachin and
Kalbajar)
A3b–B3b – Issue #3b: The status of Kalbajar region
A4–B4 – Issue #4: The status of the Lachin corridor and the return of
Azerbaijani IDPs to Lachin
A4a–B4a – Issue #4a: The status of the Meghri corridor
A5–B5 – Issue #5: The return of Azerbaijani IDPs to Shusha and NK
A6–B6 – Issue #6: The final status of NK
* This list may not include all of the issues in the NK negotiations, but it does list most of
the outstanding issues discussed by the parties during various phases of the NK negotiations.
It could be expanded if needed.
Figure 3. Map of Azerbaijan.
MAP OF AZERBAIJAN
The former boundaries of
Nagorno-Karabakh
Autonomous Oblast (NKAO)
in Azerbaijan SSR
Azerbaijani Territories under
the control of Armenian Forces
ARMENIA
Ganja
Nagorno-Karabakh
7
6
5
3
2
4
1 The Lachin
Corridor
The Meghri
Corridor
Shusha
Baku
1 - Agdam 5 - Qubadli
2 - Fizuli 6 - Lachin
3 - Jabrayil 7 - Kalbajar
4 - Zengilan
The Meghri Corridor
The Lachin Corridor
Line of Control
Occupied Territories outside of NK
Nakhichevan
Nakhichevan
(Azerbaljan)
(Azerbaljan)
Nakhichevan
(Azerbaijan)
T. Ziyadov / International Negotiation 15 (2010) 107–131 125
occupied Azerbaijani territories around NK (A3–B3), the return of Azeri IDPs to
these regions (A2–B2) and agreeing on stationing peacekeeping forces in the conflict
zone (A1–B1). But all these issues were linked to the final status of NK
(A6–B6). And because NK Armenians insisted on a direct link between Armenia
and NK, the question of returning Lachin (A4–B4) was non-negotiable. The NK
Armenians were only willing to offer a minority status to Azeri IDPs from NK
(A5–B5), as the former were in the majority in the region. In other words, the
approach of NK Armenians was independence in exchange for the occupied
Azerbaijani territories. As Libaridian states while differentiating between the
positions of Yerevan and NK Armenians:
Ter-Petrosian’s focus was the continued secure and free existence of the Karabakh Armenian population
in their historic land. The basic formula for a negotiating strategy would be the return of territories
for peace. [NK], on the other hand, by and large opted for a ‘territories for status’
formula... (Libaridian2005: 36)
Thus, while Azerbaijan had a common agreement on all issues (S) with Armenia,
Baku could have agreed only on three issues (s) with NK Armenians. The divergence
in views of Armenia and NK Armenians was the most crucial element of
the collapse of this phase of negotiations. It was primarily NK Armenians,
1. Line A (Az) = Azerbaijan’s Preference
Curve
2. Line B (Ar) = Armenia’s Preference Curve
3. Line C (NK) = Preference Curve of NK
Armenians *
4. Horizontal Lines = Issues (i.e. A1–B1,
A2–B2 etc.) See Table 1
5. A1,2,3... 6 = Azerbaijan’s preferred outcome
on issues #1,2,3... 6
6. B1,2,3... 6 = Armenia’s preferred outcome
on issue# 1,2,3... 6
7. a1,a2,a3,a4,a5,a6 = Azerbaijan’s minimum
acceptable outcome on issues #1, 2,3... 6
8. b1,b2,b3,b4.b5.b6 = Armenia’s
minimum acceptable outcome on issues
#1,2,3,... 6
9. S = Bargaining Space (convergence in
positions of Baku and Yerevan)
10. s = Bargaining space (convergence in
positions of Baku and NK Armenians)
* Because the interest of NK Azerbaijanis were represented by official Baku, their preference curve is
presumed to be the same as Azerbaijan’s (Line A (Az))
Figure 4. A multi-issue bargaining model in Phase 1.
126 T. Ziyadov / International Negotiation 15 (2010) 107–131
brought to the Armenian government by Ter-Petrosian himself, including Kocharian,
who forced Ter-Petrosian to resign in February 1998.
Phase 2: 1998 (February)–1999 (October) – “Karabakhization” of Armenia’s Position
Kocharian’s rise to power in March 1998 meant that the position of Armenia
would differ and be closer to that of NK Armenians. Fig. 5 illustrates the gradual
shifts in the positions of the two parties. Armenia would toughen its position
and move its preference curve (Line B (AR)) closer to the preference curve of
NK Armenians (Line C (NK)) (see Fig. 6). From this time on, the positions of
NK Armenians were represented by Armenia.
Toughening Armenia’s position would have resulted in a stalemate if Azerbaijan
was to stick to its original position. A commonly acceptable agreement on all
issues (S – bargaining space) could only be created if Baku was willing to satisfy
some of Yerevan’s tough demands and soften its initial position. To ease this process,
the parties introduced another issue of territorial swaps – exchanging the
Meghri corridor in Armenia (A4a–B4a) for the Lachin corridor in Azerbaijan. This,
however, meant that the final status of NK would suit the demands of the Armenian
side (gains for (B) Armenia and losses for (A) Azerbaijan) (see Figure 7).
By October 1999, despite several setbacks among his own supporters, President
Aliyev of Azerbaijan and President Kocharian of Armenia seemed ready to
sign a peace agreement at the OSCE Istanbul Summit in November 1999. But
the agreement was never signed and the deal was called off shortly after the
shooting in the Armenian Parliament in October 1999.
Phase 3: 1999–2004 – Moving to Stalemate
After the October 1999 killing in the Armenian Parliament, Yerevan backtracked
on the idea of giving Azerbaijan the Meghri corridor in exchange for the Lachin
corridor and downgrading it to a “right of access” (see Fig. 8).
Figure 5. A multi-issue bargaining model
in Phase 2
Figure 6: Armenia’s position toughens
T. Ziyadov / International Negotiation 15 (2010) 107–131 127
Figure 7. Azerbaijan softens its position.
Figure 8. Backtracking: towards a stalemate.
128 T. Ziyadov / International Negotiation 15 (2010) 107–131
The Meghri corridor was the major carrot that encouraged the Azeri president
to consider accepting the status of NK that would satisfy the Armenian side.
Without the Meghri option there was no incentive for a compromise. Moreover,
understanding that such a peace deal would be political suicide, especially during
a period of political succession, Heydar Aliyev backed down (see Fig. 9).
This led to a stalemate in the NK negotiations that lasted until 2004.
Phase 4: 2004–2009 – Back to Territorial Integrity
Since 2004, the two parties and the OSCE Minsk Group mediators have opted
for a “formula-detail” approach, trying to agree on sets of issues from the “Basic
Principles” framework. In our model, this would be a “maximization of bargaining
space,” trying to work with positions of each party and bringing them to a
commonly acceptable agreement on as many issues as possible. In general, this
period has witnessed some shifts in the positions of the two parties and also
changes in their negotiation tactics. The first shift was in Baku’s position, which
reverted to its initial formula of resolving the NK conflict only within the territorial
integrity of Azerbaijan. There was a brief stalemate after the newly elected
president of Azerbaijan met with his Armenian counterpart and proposed to
restart the negotiations, as he was not going to follow the same negotiation strategy
as his father. Soon the parties restarted the negotiation process, first on the
level of foreign ministers and then on a presidential level. The idea of agreeing
on a framework of the settlement was appealing to both sides because this document
would only be complete if both parties found a mutually acceptable agreement
to all of the outstanding issues in the document.
During this process, Yerevan softened its position, but also added a new issue
to the negotiation table, which was the status of Kalbajar (A3b–B3b). Kalbajar is
another Azerbaijani region outside of NK, which, like Lachin, is located between
Armenia and NK. Yerevan insisted that it is a strategic region that should be left
to Armenia along with Lachin. This also meant that the discussions on the pullout
of Armenian troops from the seven Azerbaijani regions adjacent to NK
would be downgraded to five regions instead of six (A3a–B3a). Nonetheless, over
the course of the negotiations, the parties seem to have come to a mutually
acceptable understanding on the timetable of the pullout of Armenian forces
from all seven regions, including Kalbajar. The issue of the Lachin corridor
(A4–B4) is still being formulated and there are encouraging signs that the parties
are likely to find a common solution to this issue as well. The two remaining
issues that have not yet been agreed upon, the return of Azeri IDPs to NK
(A5–B5) and the final status of NK (A6–B6), are in fact interlinked. The Azeri
position is that all Azeri IDPs from NK should return to NK prior to any popular
vote on the final status of the province, which is viewed by Baku only as a
“legal status” within the territorial integrity of Azerbaijan. Armenia and NK
T. Ziyadov / International Negotiation 15 (2010) 107–131 129
Armenians resist the idea of the return of Azeri IDPs and just offer the right of
return. They also call on a referendum that could allow NK Armenians to separate
de jure from Azerbaijan. These are the two most contentious issues that
remain outside of the range of the bargaining space for now.
Conclusion
The multi-issue bargaining model is a useful tool to illustrate the issues on the
negotiation table, the positions of the parties, the level of agreement on particular
issues and shifts in the positions that occurred during various phases of the
NK negotiations. In analyzing all four phases of the NK peace process, the first
phase (1994–1998) stands out as a particularly important period when the two
sides used a problem-solving approach to come to a mutually acceptable resolution.
Here, the position of NK Armenians played a critical role, which resulted
in the collapse of the negotiations when Ter-Petrosian was forced to resign in
1998. While there was a bargaining space between Baku and Yerevan, the position
of NK Armenians differed significantly, which prevented the parties from
signing any agreements.
When Kocharian came to power in March 1998 and started the second phase
of the NK negotiations (1998–1999), he effectively erased the existing bargaining
range and turned the conflict into a zero-sum situation. Yet, despite his
uncompromising position on the question of the final status of NK, Kocharian
considered the option of territorial swaps that created another bargaining space.
Figure 9. Maximizing the bargaining space.
130 T. Ziyadov / International Negotiation 15 (2010) 107–131
During this period, the parties seem to have reached a mutual agreement that
they planned to sign in November 1999, but one doubts whether it would have
succeeded in the end. The subsequent phase (1999–2004) involved backtracking
from initial compromises and the hardening of positions.
In the final phase (2004–2009), the parties adopted a formula-detail approach,
which is a combination of the step-by-step and package proposals. While most
of the issues on the negotiation table have been agreed upon, the two most contentious
issues – the final status of NK and the return of NK Azerbaijanis to the
region – remain unresolved.
Although useful in adding a multi-issue dimension to the analysis of the negotiations,
the multi-issue model has some constraints. The model does not explain
why exactly the parties shifted their positions or how and under what conditions
this shift occurred. It is also limited in illustrating third party interventions that
have been crucial in the NK peace process. In some cases, third party intervention
did not result in successful resolution of a hurting stalemate, but led to
one.18
Nonetheless, the model does help to visualize the bargaining range in the
negotiation and the issues to which the parties have found or could find mutually
acceptable agreement. It is useful in drawing some lessons from past negotiation
phases and focusing on the most contentious issues for future negotiations,
trying to identify the intractable components of the conflict and find a better
formula acknowledging areas of mutual agreement.
Compared to the traditional bargaining model, the multi-issue model could
be used to analyze issues individually and in conjunction with other issues. This
would allow a better understanding of the importance of each issue to one or the
other party and could explain the occurrence of stalemates and deadlocks in
negotiations, albeit retrospectively, which are sometimes hidden in the issue
dynamics. Overall, this multi-issue model could enrich our understanding of the
negotiation process as a whole.
Acknowledgements
I would like to thank Professors I. William Zartman and P. Terrence Hopmann
at the Johns Hopkins University for their helpful advice and encouragement.
Thanks are also due to the anonymous peer reviewer whose comments and suggestions
improved this paper.
18)
For example, Russia too often played a role of spoiler in the negotiations between Armenia and Azerbaijan,
particularly in the 1990s. It suggested proposals that did not necessarily aim to break through
the stalemate, but rather reflected the interests of Moscow. The so-called “Primakov Plan” of a “common
state” solution in 1999 was one of such proposals, which aimed to secure Russia’s influence in Azerbaijan,
Georgia and Moldova.
T. Ziyadov / International Negotiation 15 (2010) 107–131 131
References
Abasov, Ali and Khachatrian, Haroutiun (2005). The Karabakh Conflict – Variants of Settlement:
Concepts and Reality. Baku-Yerevan: Areat/Noyan Tapan.
Baguirov, Adil (2008). “Nagorno-Karabakh: Basis and Reality of Soviet-era Legal and Economic
Claims Used to Justify the Armenia-Azerbaijan War” Caucasian Review of International Affairs
2, 1.
Croissant, Michael P. (1998). The Armenia-Azerbaijan Conflict: Causes and Implications. Westport,
CT: Praeger Publishers.
De Waal, Thomas (2003). Black Garden: Armenia and Azerbaijan through Peace and War. New
York and London: New York University Press.
Hopmann, Terrance P. (1995). “Two Paradigms of Negotiation: Bargaining and Problem Solving,”
ANNALS, AAPSS, 542 (November).
—— (1996). The Negotiation Process and the Resolution of International Conflicts. Columbia, SC:
University of South Carolina Press.
Internal Displacement Monitoring Centre (IDMC) (2007). Azerbaijan: IDP Living Conditions
Improve, but Peace Deal Remains Elusive. Geneva: IDMC.
Libaridian, Gerard (2005). “The Elusive ‘Right Formula’ at the ‘Right Time’: A Historical Analysis
of the Official Peace Process,” in Laurence Broers, editor, Elites and Societies in the Nagorny
Karabakh Peace Process. London: Conciliation Resources.
Mehdiyev, Elkhan (2005). “Armenia-Azerbaijan Prague Process: Road Map to Peace or Stalemate
for Uncertainty?” Caucasus Series, May.
OCSE (2007). “OCSE Minsk Group Co-chairs Issue Statement on Nagorno-Karabakh,” Press
Release, November, 29.
—— (2008). “OCSE Minsk Group Co-chairs Issue Statement on Nagorno-Karabakh, Press
Release, March 19.
Safaryan, Stephan (2005). “Conflicts in the South Caucasus: Political, Security and Development
Challenges,” GFSIS Roundtable Discussion, July 5–6.
UN High Commissioner for Refugees (UNHCR) (2005). Assessment of the IDP situation in Azerbaijan
and Cooperation Mechanisms in Place to Address their Needs, March 24.
Walton, Richard E. and McKersie, Robert B. (1965). A Behavior Theory of Labor Negotiations.
New York: McGraw-Hill.
Yunusov, Arif (1998). “The Armenia-Azerbaijan Conflict: Demographic and Migration Aspects,”
Central Asia and the Caucasus Journal 16.
Zartman, I. William (2008). Negotiation and Conflict Management: Essays on Theory and Practice.
London and New York: Routledge.
Zartman, I. William and Berman, Maureen (1983). The Practical Negotiator. New Haven, Conn.:
Yale University Press.
Zulfuqarov, Tofik (2005). “Obstacles to Resolution: An Azerbaijani Perspective,” in Laurence Broers,
editor, Elites and Societies in the Nagorny Karabakh Peace Process. London: Conciliation
Resources.
Zyberk, Henry Plater (2007). “Misunderstanding Russia: Alexander Litvinenko,” in Andrew Monaghan,
editor, The UK& Russia -A Troubled Relationship – Part 1. UK: Defense Academy of the
United Kingdom, Conflict Studies Research Centre: Russian Series.
Copyright of International Negotiation is the property of Martinus Nijhoff and its content may not be copied or
emailed to multiple sites or posted to a listserv without the copyright holder's express written permission.
However, users may print, download, or email articles for individual use.