9 'Land for Peace' 'Can you think of a conflict where the winning party has given up land for a promissory note?' I was in Washington, in the , hubbub following the close of an event at a downtown thinktank in 2014. My interlocutor was an activist in one of the Armenian lobbying groups active in the US capital. Her question conceded too much to a victor's outlook, but captured the elusiveness of the 'land for peace' formula widely seen as the key to an Armenian-Azerbaijani peace. Nearly thirty years of diplomacy mediated by the Organization for Security and Co-operation in Europe (OSCE) has yielded only two documents signed by Armenian and Azerbaijani leaders, both brokered by Russia. These are the 12 May 1994 ceasefire, known as the Bishkek Protocol and the Moscow Declaration of 20 November 2008,, a symbolic commitment to a peaceful resolution that subsequent events proved meaningless. Despite continuous dialogue, plentiful peace proposals and intermittent high-level attention from global leaders, the Armenian-Azerbaijani conflict still awaits its peace conference. Why has mediation failed? The deep dynamics ?f rivalry explored in previous chapters offer a range of explanatlOns, from overlapping and indivisible geopolitical visions, to the dynamICS of hybrid regimes, the strategic parity of truncated asym~etry, the diffusion of international leverage, and the congeahng of de facto realities in the territory at the heart of the rivalry. This chapter considers three sets of factors that can loosely be considered intrinsic to the process of mediation itself. They trace a broadly chronological arc from the earliest mediation efforts while hostilities were still ongoing in the 1990s to the 276 'Land for Peace' resent day. The first is the impact of mediator rivalries. The ~omination of the OSCE's mediation body, the Minsk Group, by three global powers, France, Russia ,and the 1!~ited Sta~es, which also represent three poles in today s competItIve EuraSIan eopolitics, accounts for the popularity of geopolitical perspec~ves in explaining mediation failure, especially in Azerbaijan. As Azerbaijani Deputy Foreign Minister Araz Azimov contends, 'the Minsk Group is an institution that has been used and abused for geopolitical purposes'.1 Geopolitical perspectives also account for one of the most frequently asked questions about the peace process: whether othe~ medi~tors could ac~ieve more. There is ample evidence of medIator nvalnes hampen.ng early ~fforts. ~et after the mid-1990s the nexus of external Interests In avertIng another Armenian-Azerbaijani war has resulted in what one former mediator calls 'shockingly good' cooperation between Russian and Western counterparts.2 Superseding mediator rivalries from the late 1990s were problems relating to the structure and sequencing of the negotiating agenda. This agenda, already confronting the contradiction between territorial integrity and self-determination, was overlaid by the dramatic outcomes of the 1992-4 war. Most obviously, the occupation by Armenian forces of seven districts surrounding Nagorny Karabakh gave rise to a working distinction between the 'consequences' and 'causes' of conflict. This distinction yields the 'land for peace' formula. Simply put, this proposes that the occupied territories return to Azerbaijani jurisdiction in exchange for an agreed determination of status for Nagorny Karabakh. Between 1997 and 2004 a succession of proposals attempted to manage this equation in a variety of different ways. None succeeded. On the Armenian side a more demanding alternative, 'land for status', vies with 'land for peace'. This ties the return of occupied territories to an explicit recognition of Karabakh's secession from Azerbaijan, casting the territories as bargaining chips in a game of geopolitical extortion. This approach never overcame Azerbaijani resistance to territorial fragmentation, and increasingly confronts deepening Armenian attachments over time to 'augmented Armenia' and perceptions of the territories as more existential than collateral in significance. 277 Armenia and Azerbaijan poor relation to Richard Holbrooke, US envoy to the conflict in Bosnia in the 1990s, Maresca recalls: I couldn't get the high-level attention from the Russians I wanted and there was a lot of confusion because there was no single Russi~ policy. There were two or three policies, depending on whether you were. dealing with the Ministry of Defence or the Ministry of Foreign Affairs. I would have meetings with Ministry of Defence people and the Russian co-Chair, my negotiating partner, wasn't allowed in the room because he represented a different policy.8 Russia's policies oscillated between pushing for a Russianled peacekeeping operation under the formal aegis of the Commonwealth of Independent States (CIS) - similar to that introduced into Abkhazia, Georgia - and support for a multilateral CSCE operation. Yet although the Bishkek Protocol, brokered by veteran Russian diplomat Vladimir Kazimirov, commits the parties to 'suggest Parliaments of the CIS member-states to discuss the initiative ... on creating a CIS peacekeeping force' this was never acted upon, largely due to Azerbaijani resistance: The Armenian-Azerbaijani ceasefire of 12 May 1994 emerged as self-regulating, overseen by neither a CSCE-Ied multilateral nor a Russian-led CIS peacekeeping force. It consequently owed its durability to belligerent exhaustion, not international oversight. The CSCE's Budapest summit of 5-6 December 1994 solidified the mediation structure. At the summit, the CSCE transformed itself into the Organization for Security and Co-operation in Europe, an ambitious jump in scope and purpose. With regard to the Armenian-Azerbaijani conflict, the summit mandated 'a single coordinated effort' to consist of newly appointed co-Chairs of the still pending Minsk Conference, who would chair the Minsk Group, the collective of states originally designated as participants to the conference, and report to the OSCE's Chairman-in-Office (CiO).9 The summit also mandated the establishment of a HighLevel Planning Group (HLPG), 'to make recommendations on, inter alia, the size and characteristics of the [peacekeeping] force, command and control, logistics, allocation of units and resources, rules of engagement and arrangements with contributing States,.l0 280 'Land for Peace' The co-Chair system replaced the earlier rotating single chair initially held by Italy then Sweden. A dual system was introduced whereby Russia had a permanent chair and a rotating chair was held by a 'neutral' state (Sweden in 1994, then Finland in 1995-6).11 This arrangement acknowledged Russia's role while embedding it within a multilateral format. Russian acquiescence reflecte~ a newfound commitment to the OSCE as a regional mechamsm balancing the perceived encroachment of Westernled struc:ures, such as NATO's Partnership for Peace, into the post-Soviet space.12 The addition in 1995 of the onerously titled Personal Representative of the Chairperson-in-Office on the Conflict Dealt with by the OSCE Minsk Conference (PRCiO) completed the OSCE's mediation structure. Since July 1996 Ambassador Andrzej Kasprzyk, of Polish origin, has filled the post of PRCiO; together with five field assistants, he is the OSCE's only field presence.13 The PRCiO is responsible for monitoring the ceasefire with pre-arranged, bi-monthly inspections, and has played a critical role in, inter alia, crisis communication, prisoner exchanges and facilitating the exchange of human remains in the aftermath of April 2016's 'four-day war'. The consolidation of the Minsk Group relieved but did not resolve the problem of fractured mediations. Under the FinnishRussian co-Chairmanship in 1995-6, the Minsk Group held regular meetings. Concurrently a backchannel was opened between presidential envoys Gerard Libaridian and Vafa Guluzade articulate negotiators with a good personal rapport. At the sa~e time in. February 1996 the new Russian Foreign Minister Yevgeny Pnmakov also circulated a blueprint proposing loose confederal relations monitored by Russian-led peacekeeping operations.14 Parallel international initiatives were finally reined in through the mtroduction in January 1997 of a permanent troika of Minsk Group co-Chairs, consisting of Russia, as before, in addition to France and the United States. IS Carey Cavanaugh, former US co-Chair of the Minsk Group in 1999-2001, explains the virtues of this approach: Many countries would dream of having such a negotiating structure. You've got all the major powers at the table ... If you have a solution, 281 Armenia and Azerbaijan -no single party could implement it on their own: Russia has to be there, it is the only power with the local presence to provide security. But Russia doesn't have the finances nor the international influence with financial organisations like the International Monetary Fund international NGOs, the donors ... that's where the US comes in' we could provide that. And you need Europe too, that's a geographi~ and political inevitability, that's where these countries are ultimately headed. So you have the political, military and economic backing for a settlement, and what's more, the endorsement of three members of the United Nations Security Council. You don't get much better than that.16 Not all actors see the Minsk Group troika in such a positive light. John Maresca observes that the arrangement 'brought the interests of the co-Chair nations into the mechanism. There were built-in incentives to bring geopolitical interests to the table and it created disincentives to change the mechanism.>17 Yet th~ finalisation of the troika also preceded the most productive era of the Armenian-Azerbaijani peace process. Mediator rivalries faded into the background of an equally complex set of challenges posed by the nature of the negotiations agenda. Structure and Sequence (1997-2004) Between 1997 and 2004 Armenian and Azerbaijani leaders discussed a succession of concepts for peace within the framework of the Minsk Group. In variable - sometimes diametrically contrasting - ways, all grappled with a set of problems associated with the structure of the conflict, the issues at stake, and possible sequences of their resolution. None found viable solutions. A first problem was the structure of the conflict. The ArmenianAzerbaijani conflict confronts all mediation efforts with the contested primacy of territorial integrity and self-determination, and the presence of a non-state actor in the form of the unrecognised Nagorno-Karabakh Republic (NKR). Unlike the four resolutions issued by the United Nations Security Council in 1993, all of which had explicitly highlighted territorial integrity, in its early 282 'Land for Peace' deliberations the CSCE, then OSCE, had fudged this issue.18 The 1992 mandate for the Minsk Conference stipulated the participation of 'elected and other representatives' of Nagorny Karabakh as 'interested parties'.1 9 This ostensibly covered both the Armenians of Karabakh, and the Karabakh Azerbaijani minority displaced from the territory. However, by 1996 the proliferation of secessionist claims in the former Yugoslavia, heightened international perceptions of the brutality accompanying secessionism, and the transformation of the CSCE into an organisation composed of ethnically diverse states, several of which confronted separatism or outright secessionism, resulted - in combination with an Azerbaijani diplomatic offensive - in a more statist vision of European security at the OSCE's 2-3 December 1996 Lisbon summit. The principle of territorial integrity was explicitly affirmed for Georgia and Moldova.20 Owing to Armenia's exercise of its veto, a similar commitment to Azerbaijan was relegated to a separate statement by the chairman affirming the territorial integrity of both Armenia and Azerbaijan, self-rule for Karabakh within Azerbaijani borders, and security guarantees for the 'whole population' of Karabakh as the parameters for a solution to the Armenian-Azerbaijani conflict.21 Azerbaijan considered the Lisbon summit a diplomatic triumph and validation of its position. Its statist emphasis confronted Karabakh Armenians with the prospect that military victory might result in less than independence.22 But this also flagged the problem of a state-based negotiation of a conflict in which a non-state actor - the NKR was central. A second problem was the structure of the issues comprising the negotiation agenda. Negotiators in the mid-1990s defined two agendas: 'military-technical issues', interpreted as addressing the consequences of conflict, and the 'status issue', interpreted as its original cause and concerned with determining the status and rights of Karabakh Armenians.23 In addition to humanitarian issues shared by all parties, for Azerbaijan the salient consequences of the conflict included the occupation, in whole or in part, of the Azerbaijani regions of Agdam, Fizuli, Jebrayil, Zangelan, Qubatly, Lachin and Kelbajar surrounding the former autonomous oblast. For Armenians, the salient consequences 283 Armenia and Azerbaijan -were the. b.lock.ades ~aintained by Azerbaijan and Turkey, and the persIstIng Insecunty confronting the NKR as an unre . d " . cog?ISe ~ntlty In an International environment privileging territo . I 1 t A' 1" nan egnty.. ~ Imp IClt assumption accompanied this division of the negotIatIon agenda. This was that since Azerbaijan'S los ~ad been greater in terms of territory and internal displaceme ses Its fundamental interests were driven more by consequences thnt, c . h . an auses, suggestmg t at It would be tractable on causes in order to undo the c~nsequences of conflict. This assumption underpinned ;he alternatIve formula of 'land for status', a hardline variation on land for peace'. !his division of the negotiation agenda generated in turn a thIrd. problem, namely whether consequences and causes of conflIct should be dealt with simultaneously or sequentially. In t~e .parlance of the Armenian-Azerbaijani negotiations, these dIstInct approaches are popularly labelled 'package' and 'ste _ by-step' res~ectively. The packag~ approach implies a one-ste~, comprehenSIve agreement on all Issues.24 In theory, this enables trade-o~fs in. which parties 'win' or 'concede' according to the vana~le Importance they accord to individual issues in an over-~rc~Ing grand bargain. This assumes, however, that the ?egotIatIng parties do not privilege the same issue as the most Important. The package approach also introduces the element ?f con?itionality, since agreement on one negotiation agenda IS contmgent on agree~ent ~f all. Conversely, the step-by-step ~pproach reduces condItIonalIty by delinking agendas and allowIng pro?ress on one independently of agreement on the other. :~e lOgIC of this approach is that the most intractable issues are kIcked down the road', allowing trust and confidence to build up t~rough the resolution of lesser issues. The problem here is that Implementation of the earlier steps has a direct causal impact o~ the shape and form that resolution of the more critical issues wIll e:e~tually take. In the context of the Armenian-Azerbaijani ?egotlanons the critical distinction between these approaches IS whether to 'frontload' or 'backload' the issue of Nagorny Karabakh's status. 284 'Land for Peace' Three Package Proposals and a Step-by-Step Alternative Between 1997 and 2001 Armenian and Azerbaijani leaders discussed what were essentially three package plans that - even if vaguely - specified the contours of a final status solution, and one step-by-step concept that only committed the parties to a deferred negotiation of status at a future, unspecified date (see Table 9.1). The three package proposals are the 'package' proposal of May-June 1997, the Russian-inspired 'common state' proposal of November 1998, and the 'territorial swap' concept discussed by Presidents Heydar Aliyev and Robert Kocharian in 1999-2001. These three concepts can loosely be interpreted as resolving the status issue through autonomy, confederalism and secession respectively. Each had a distinct political genealogy. The Minsk Group co-Chairs elaborated the first package proposal through the early months of 1997, and presented it to the parties in May-June. There is some ambiguity over the labelling of this proposal as a package deal. US diplomat Philip Remler clarifies that the Armenian and Azerbaijani leaderships always understood the proposal as a phased negotiation, in which Agreement I on security issues would be adopted before the negotiation of Agreement II on status issues began.25 Nevertheless, even if understood as a phased negotiation, the proposal followed closely in the wake of the Lisbon summit and reflected its emphasis on territorial integrity in prescribing a final status for Karabakh. The published text of Agreement II defined Nagorny Karabakh as 'a statal and territorial formation, within the borders of Azerbaijan'.26 The proposal listed an extensive set of rights consistent with self-government for Nagorny Karabakh and free mobility and migration to Armenia, yet the envisaged status was less than independence. Karabakh Armenians would elect representatives to the Azerbaijani parliament, and hold Azerbaijani passports - albeit specially annotated - and their holders would not be considered foreigners in Armenia. This was, in effect, 'land for a liberal peace', additionally qualified by the fact that as per Agreement II, Armenia would only recognise Azerbaijan'S territorial integrity once Karabakh's status had been mutually agreed.27 Within its 1988 borders, Karabakh would be linked to Armenia 285 Armenia and Azerbaijan bringin.g ,about a. s.ituation where a Karabakh native occupied Armema s top polItIcal post and negotiated on behalf of Armenia an~ ~arabakh. President Heydar Aliyev, his health in decline and polItIcal succession on his mind, sought to simplify the conflict's legacy to his son. Over sixteen meetings in a two-year period workin? 'often alone', the presidents elaborated the Aliyev~ K?chanan plan.30 The evolution of their plan and its.details are stIll shrouded in secrecy and controversy. Unlike previous and ~ater peace plans, no document exists today in the public domain: it was ~ 'two-man solution' never submitted to wider scrutiny. The idea of a territorial exchange as a solution had circulated since the beginning of the Armenian-Azerbaijani conflict in 1988. Several external observers, including US State Department analyst Paul Goble, Nobel Prize winner Andrey Sakharov and Turkish President Turgut 0zal, had all ventured ideas of this kind.31 The idea's attractiveness lay in its potential to simultaneously resolve the territorial problems posed by Nagorny Karabakh's enclave and Nakhichevan's exclave status, two geopolitical 'islands' seeking access to their 'mainlands'. Aliyev and Kocharian developed a concept that accepted the post-war status quo, but also saw in it a historic opportunity to permanently disentangle Armenian and Azerbaijani geo-bodies. They discussed the transfer of Nagorny Karabak~ to Armenian jurisdiction in exchange for the ceding of a corndor across Armenia's southernmost Meghri region to Azerbaijan. For Azerbaijan, a coveted corridor to Nakhichevan was the quid pro quo for losing Karabakh. No longer divided into mainland and exclave, Azerbaijan would have lost one form of territorial integrity, but gained another. For Armenia, a lost border with Iran was the corresponding quid pro quo for finally legalising possession of Karabakh. The underlying logic was 'land ~or status', but with a radical twist that included the ceding of de Jure Armenian territory as part of the deal. It was the ultimate geopolitical solution, taking a transactional attitude to territorial integrity in quest of a cartographic fix. An initial iteration of this plan failed due to the shocked reactions of elites when the presidents revealed their plan, and to the devastating impact of political assassinations whose connection to the proposal remains the subject of febrile speculation. 288 'Land for Peace' On :he A.zerbaij~ni side, three senior Azerbaijani officials, firstly preSidential adVisor Vafa Guluzade and then Foreign Minister Tofik Zulfuqarov and Presidential Secretary Eldar Namazov resigned in October 1999 in protest at the ideas under discussion: Guluzade's response, on learning of the plan, was to tell Aliyev: 'You are n~t a ~andlord to give away our lands.>32 Three days later, t~rrons~ mterceded when former journalist and political extremist Nam Hunanyan, his brother and three others broke into the Armenian parliament on 27 October 1999 and shot dead Armenian Prime Minister Vazgen Sargsyan, Parliamentary Speaker Karen Demirchian and six other senior officials. The link between the assassinations and the peace process is much conjectured but unproven. It hinges on the assumption that the assent of war hero and networked strongman Vazgen Sargsyan was crucial to overcoming likely domestic resistance to the plan. Yet there is no evidence that the assassins could have known of his disposition towards the plan, on which accounts in any case vary.33 Domestic political motives, rather than bringing down a peace plan, may account better for the assassinations, yet the impact was the same: the peace process halted for a year. The territorial swap plan was revived in 2001 in a second iteration that offered Azerbaijan less. Former Foreign Minister of Armenia Vartan Oskanian explains what was under discussion: Karabakh with Lachin was being given to Armenian sovereignty. ~aragraph 2 was clear on this point ... and what we were giving In return was sovereign use of the corridor through Armenian territory to link Azerbaijan and Nakhichevan. [There) was the difference b.e~ween the two sovereignties: sovereign use meant that [Azerbaijani citizens) will cross the Armenian border without any border control, no one could stop them on that road to Nakhichevan ... But you cannot put a gas station on that road because it is not your territory, you Just have sovereign use.34 These ideas were discussed at talks in Key West Florida in April 2?01. But no breakthrough ensued, and the ;eace pro~ess lapse.d mto desuetude. Even without the destabilising effects of political murder, domestic resistance to territorial swap concept 289 Armenia and Azerbaijan UlillPltli en SUIi qntnll t, ~U'I••• Figure 9.1 Poster-board, Meghri, Armenia, 2015. The legend reads: 'Armenians, Meghri is the door to your home!' Author photo. was critical. From an Azerbaijani perspective, what was discussed at Key West 'was not even a territorial swap, because we were getting only a corridor'.35 Even Oskanian concedes that the formula seemed uneven, and expectations that Azerbaijan would withdraw were high. On the Armenian side the conceding of the border with Iran was widely seen as prohibitively costly. Seventeen years later poster-boards in the town of Meghri still declared the region to be the 'door to Armenia' (see Figure 9.1). The three package proposals are striking in their coverage of the full spectrum of possible outcomes. The logic of the first package proposal, in essence, was a solution consistent with Azerbaijani territorial integrity that compensated Armenians with wide-ranging rights and veto-points precluding - in theory - the imposition of a status not to their satisfaction. Conversely, the territorial swap proposal essentially accorded with the Armenian position and compensated Azerbaijanis with access to Nakhichevan. The 290 'Land for Peace' common state proposal sought an ambiguous third path of horizontal relations. That none of these three plans were accepted is an indication of the limitations of the package approach, and its promise of trade-offs, when the conflict parties privilege the same issue - the status of Nagorny Karabakh - as the most important. Each package plan demanded of one party or another concessions on this issue that they were not willing to make, which even generously conceived compensation could not indemnify, and which outside powers could not compel them to agree to. What, then, of the phased approach? After the rejection of the first package plan in August 1997, the Minsk Group held further consultations and presented an updated, 'step-by-step' proposal in September. The essential difference from the first package proposal was the omission of Agreement II issues. Rather than a specified commitment to a status solution, Paragraph XI only committed the parties to further negotiation of the status of Nagorny Karabakh. Several sensitive issues previously explicit in Agreement I of the package proposal, such as the Lachin corridor and displaced community return to Shusha and Shahumyan, were also included in this basket of deferred issues. The plan consequently provided for the return of occupied territories and demilitarisation without an explicit determination of the crucial status issue. This was 'land for a peace subject to further negotiation'. With reservations, Baku accepted the concept. Armenian President Ter-Petrossian's advocacy of this plan, articulated in his article 'War or Peace? Time for Reflection', was rejected by Prime Minister Kocharian and his supporters, who then forced Ter-Petrossian's resignation in February 1998.36 The step-by-step proposal is remembered today as a significant moment when Armenian and Azerbaijani leaderships aligned around the same plan as a basis for negotiations, but did little to prepare or advocate for it. It failed for several reasons. First, it still had the shadow of the Lisbon summit hanging over it. The plan was closely associated with - indeed an outgrowth of - the package plan of only a few months before that had offered Karabakh Armenians a status less than independence. Second, the plan effectively entrusted the fate of the Karabakh Armenians to Yerevan, and consequently to President Levon 291 Armenia and Azerbaijan T.e~-Petrossian. His and the Pan-Armenian National Movement's VlSlOn of a 'compliant Armenia' was ambivalent on the role of Nagorny Karabakh in a sovereign Armenian state. Moreover his hold on power had been weakened by allegations of fraud and post-electoral violence at his re-election as president in 1996 a~cumulating discontent over the economy and corruption, and ?IS own dependence on Kocharian's nationalist legitimacy to stay In . power. But perhaps most significantly, the step-by-step plan failed because of the 'promissory note' problem identified at the beginning of this chapter. The NKR leadership saw no reason to concede to a vague 'land for peace' formula - which they saw as leading back to autonomy - when 'land for status' was within reachY The plan contradicted a strategic calculus that Azerbaijan had more to gain from reversing the consequences of the conflict than it had to lose from being tractable on its cause: the status of Karabakh. As Heydar Aliyev's apparent pliability in 1999-2001 subsequently showed, this was not an outlandish calculation. But it proved ultimately untenable because of the under-estimation of t~~ power behind the narratives constituting Azerbaijani geopolmcal culture examined in Chapter 2, and an Azerbaijani calculus t~at the coming power asymmetry with Armenia made concesSlOns on status unnecessary. Towards a Hybrid Approach: From Proposals to Principles Following the discarding of the territorial swap concept, talks resumed in the 'Prague Process' from 2002 between presidential envoys, Deputy Foreign Ministers Araz Azimov and Tatul Markaryan. In 2004 Foreign Ministers Vartan Oskanian and Elmar Mammadyarov took up the process and engaged in new ideas. First, they sought to reconcile the package and step-by-step conundrum through a hybrid approach. This envisaged agreement on the mechanism for deciding the final status of Karabakh, but the deferral of its implementation until after the return of the occupied territories and settling of security issues.38 In other words, the strategy was to frontload agreement on the status determination mechanism, but backload its deployment in the process. Second, rather than setting out elaborated proposals, the 292 'Land for Peace' foreign ministers sought to agree a framework agreement of basic principles on the basis of which a comprehensive peace agreement would then be developed. The result of these negotiations was the Basic Principles (inform~lly referred to as th~ 'Madrid ~rinciples'), presented by the Minsk Group co-ChaIrS In the Spamsh capital in November 2007. The document presented in Madrid constituted a set of fourteen bullet points over two sides of A4, which remain the basis for negotiations to this day.39 They provide for withdrawals of Armenian forces from the occupied territories (with a distinct ti~et~ble for the region of Kelbajar), a corridor 'of an agreed Width between Karabakh and Armenia, a right of return for all displ~ced persons and the deployment of an international peacekeeping force. Two of the Principles provide for a two-step solution to the question of status. Nagorny Karabakh would receive an i.n~erim status, effectiv~ly codifying its de facto status of today, until ItS final legal status IS determined in a plebiscite at a time to be further negotiated. Vartan Oskanian explains: Interim status plus the prospect of a referendum sometime down the road, for us, was almost tantamount to independence. You have interim [status] until the referendum, and you know the result of the referendum because it is clearly stated that the proportion [of Armellian and Azerbaijani voters respectively] cannot be more than 80:20, so that was guaranteed. We know how the Armenians would vote, so you have interim status which is almost independence and the prospect for self-determination expression through the referendum, the combination was for us independence. That was our face-saving.4o In ~his VISion, the Madrid Principles essentially offer an alternative route to con.sens~al secession that the territorial swap proposa~ had made Imagmable, but with a 'softer landing'. The f~ce~savmg element for Azerbaijan was that at the moment of slgnmg a fra~ework agreement, the plebiscite would only be a futur~ c?mmltment; territorial integrity, for the moment, would r~m~m mta~t ~s other territories returned to Azerbaijani jurisdlctlOn. ThiS IS not, however, a vision shared in Azerbaijan, 293 Armenia and Azerbaijan -where offi.cials .ar~ less sanguine on the face-saving potential of th~ Madnd Pnnciples. They highlight the problem of signin tWlCe: the Madrid document commits leaders to signing off on th: framework, and the subsequent negotiation of a comprehensive agr.eement within six months. This has significant political implicatlOns, as noted by Araz Azimov: The adoption of the Madrid Principles as a text would be lauded as a breakthrough, but there is no agreement in the Principles, it would be in a spiral coming out of them. The same problems would extend into the negotiation of a comprehensive peace agreement, they wouldn't end with a framework agreement. The practical effect of accepting the Basic Principles would thus be zero, you wouldn't get territories coming back or people returning to them straight away. But the political impact would be devastating.41 Given that the Madrid document stipulates the boundaries, electorate and unlimited nature of status options to be offered in the future plebiscite, its outcome can indeed be seen as a foregone conclusion. Unsurprisingly, there have been multiple working versions of the Principles circulating at different times, or even concurrently, with language suiting one side or the other. What are sometimes referred to as 'updated' Principles refer not to a plebiscite, for example, but to a 'mutually agreed and legally binding expression of will'. The Azerbaijani vision of this vote contrasts sharply with that elucidated above by Vartan Oskanian. As Azimov explains: Yes, there will be a vote on status at the end of the process ... We see interim status as the recognition of [Karabakh Armenians'] status until the determination of their final status within the framework of territorial integrity. It means the legitimation of a local authority, of economic relations, of tourism and so on. Karabakh Azerbaijanis would have the same legal rights, so they too would have some kind of interim status .. . [Karabakh Armenians] have to agree to interim status with police forces, demobilisation of the army they have there, legal security forces. They have to become legalised within our system, not Armenia's.42 294 'Land for Peace' These perspectives indicate a growing divergence between an Armenian reading of the Madrid Principles in which the referendum is the first principle on which all the others hang, and an Azerbaijani reading in which the referendum is the last principle, the need for which is obviated by the successful enactment of the others.43 The hybridity of the Madrid Principles thus did not overcome the essential problems of the structure and sequencing of the negotiating agenda: We ended up in a situation where there is no document on the table, and what we are arguing today is not the substance of one particular document, but we are arguing about which document should be the basis of our talks.44 A determined effort by President Dmitri Medvedev in 2009-12 generated the last occasion when a breakthrough was plausibly anticipated, at!l summit in the Russian city of Kazan inJune 2011, h P · . 1 45 but was insufficient to secure an agreement on t e nnClp es. Among mediators there is a sense of inevitability to the ideas contained in the Madrid Principles. As US co-Chair James Warlick noted in a speech at the Carnegie Endowment for International Peace in 2014, 'after each failed round [of negotiations], the building blocks of the next "big idea" were similar to the last time'.46 This is underscored by the sheer longevity of the Madrid Principles. While other proposals have come and gone in a matter of months, they have lain on the negotiating table for twelve years. Over that extended period, however, the conceptual refinement of the Madrid Principles was gradually eclipsed by a new political logic undermining the very assumptions on which they were based. Liberal Assumptions, Illiberal Practices The Madrid Principles, in their provisions for rights, electoral mechanisms, inclusivity and participation, represent a liberal model of conflict resolution. They reflect the core principles of the Helsinki Final Act, and the OSCE's foundational purpose 295 Armenia and Azerbaijan as a regional security organisation in post-Cold War Eurasia. The measures they envisage are increasingly at odds with a normative context evolving over their lifespan, both regionally and globally, challenging liberal norms and practices. Globally, the management of internal conflicts has become a key area of COntestation as the United Nations Security Council has been repeatedly deadlocked over appropriate responses to conflicts in Sri Lanka, Sudan, Kosovo, Libya, Syria and Ukraine. The hegemony of the liberal peace as a model 'exporting' conflict resolution norms through democratisation, human rights and liberal governance, has been challenged both theoretically and empiricallyY Regional hegemons, including Russia and China, openly reject international liberal norms and manage their own internal conflicts through authoritarian alternatives. While the threshold of recognition for new states remains high, the fracturing of global opinion on the secessions of Kosovo and Crimea in particular undermined prior assumptions of a unified, law-bound approach to internal conflicts in Eurasia. These factors have constrained the OSCE's peacebuilding impact.48 Several of its mediation structures have never in fact been activated, while consensual decision-making has held the organisation hostage. In 2017 the OSCE conceded the closure of its field office in Armenia, its last ground presence in the South Caucasus, due to Azerbaijani objections to its support of demining activities in Armenia.49 Beyond these issues, however, the OSCE's commitments to liberal norms of conflict resolution are increasingly at odds with the challenge of non-liberal approaches to conflict. In the 19905 the liberal peace was conceived, like democratisation, as diffusing across a global normative periphery. As was the case for democratic transition, this vision was rapidly understood to be overambitious. Alternative conceptions of 'post-liberal' or 'hybrid' peace subsequently emerged that allowed for combinations of liberal international and local political norms in addressing conflict.5o Over time, however, it became clear that the liberal peace confronted more than the residual resistance of non-liberal actors. Rather, the problem was the development of a coherent, illiberal alternative that did not seek to adapt, hybridise or cohabit with the norms and practices of a liberal peace, but to manage 296 'Land for Peace' conflict in ways consistent with the preserva.tion of authoritarian Ie. This alternative entails the suppreSSlOn of armed rebel~i~n through an array of political, social and eco~o~ic policies, and has been termed 'illiberal peace' or 'authontanan conflIct management'.51 . . . . In the context of the Armenian-AzerbaiJalll nvalry thIS alternative might be called the 'authoritarian conflict strategies' ?f ~he parties. These are premi~ed on the idea t~at liberal ~eaceb~lldmg cannot arrive at resolutlOns of the conflIct compatIble WIth the continued rule of networked regimes. Authoritarian conflict strategies do not seek to de-escalate or terminate rivalr~, but r~ther to exploit it as a domain for the dev~lopm~n~, expenmen:atlon ~nd deployment of practices embeddmg .e.xlstmg power hl~r~rch~es. They entail the homogenising of polltlcal space, the legltlmat~on of the state to the detriment of other actors, and the strengthenmg of illiberal political and security controls dimini~hing ac~ountability. Their impact on mediation, and .the sol~tlons envlsage.d in the Madrid Principles in particular, IS to dIssolve the baSIS for liberal norms to govern measures such as interim status, the return of displaced communities, inclusive governance an? el.ectoral mechanisms, or credible security guarantees. Authontan~n conflict strategies are Janus-faced in that they seek both to c.o~ta~n conflict without recourse to liberal norms, and to explOIt It m order to embed illiberal norms in the wider governance of the state. They channel the generalised insecurity of enduring rivalry into a resource for political domination. . This argument comes with important caveats.. FIrst, a~th.oritarian conflict strategies do not imply a normatIve assoClat.lOn only with authoritarian regimes. They define a set o~ practlces enacted within specific policy domains related to conflIct, and as such can also be enacted by ostensibly democratic states.52 Israel and Sri Lanka in 2006-8 offer examples of formally democratic states enacting authoritarian practices in specific, conflict-relate.d spaces or issue areas. Such zones of exception beyond derr:ocratlc oversight are typically unsustainable, however, as they ultl~ately undermine the wider democratic order. A second caveat IS that the underlying logic of authoritarian conflict strategi~s may v~ry. In Azerbaijan's case, such strategies are congruent WIth the WIder 297 Armenia and Azerbaijan practices of a hegemonic authoritarian regime. In Armenia's case they can also be seen as instruments aimed at the consolidation of military victory as an alternative to a peace agreement. The nature of the victory, involving substantial territorial overspill and the growing power asymmetry with Azerbaijan, builds in a dependence on authoritarian conflict strategies in order to sustain a tenuous victor's peace. Finally, I am not arguing that those negotiating on behalf of Armenia and Azerbaijan since the mid-2000s, often with great skill and sophistication, have done so cynically. Indeed, most of t.heir labour has been undone by the impacts of these strategies, whiCh generally serve the interests of the networked regimes to which foreign ministers are not traditionally close. I examine here three aspects to authoritarian conflict strategies. The first concerns a strategy of control that disables the voice and representation of significant constituencies affected by conflict, prevents dialogue, and produces a singular, hegemonic discourse about the conflict. The second defines a process that I call the communalisation of the narratives, issues at stake, and essence of the conflict, binding people into homogenised, ascriptive identities and silencing other political agendas and conceptions of identity. The third involves the deployment of coercion that transforms the political arena through violence and justifies the strengthening of authoritarian political and security controls. Control Liberal models of conflict resolution seek to open up peace processes to diverse stakeholders and, by acknowledging and reconciling opposed views, to legitimate peace through inclusivity. These are the principles on which the OSCE defines its approach to mediation.53 Authoritarian conflict strategies conversely seek to limit the expression of differing views and the agency of other actors, and to promote a single hegemonic narrative that exclusively legitimates the state.54 The strategy of control has been visible first in the narrowing of the negotiating table in the Minsk Group itself. The original mandate for the Minsk Conference, as already noted, specified the participation of 'elected and other representatives' from 298 'Land for Peace' Nagorny Karabakh as interested parties. Early Minsk Group talks in 1992-6 took place in this wider format. This not only created concerns for Azerbaijan regarding the tacit recognition of the NKR, but admitted an official platform where Karabakh Armenian grievances could be legitimately raised and expressed. Azerbaijan sought to preclude this through the elaboration of a discursive and political equivalence between the 'two communities of Karabakh', with equal claims to self-determination: Our vision is of two communities equally footed, equally provided for and engaged in self-rule, whether bi-communally or separately. I would prefer it to be bi-communally in shared institutions, but if that is not possible then separately. But both should have equal status.55 Rancorous argument regarding the status of delegations from Karabakh ensued. Maresca recalls of the Minsk Group's first meeting in 1992: 'after a day of wrangling we found a solution by agreeing that the representatives of the two ethnic population groups from Nagorno Karabakh would be associated with the delegations of Armenia or Azerbaijan'.56 This wider format ended in 1997 as Azerbaijan successfully leveraged the more statist emphasis of the Lisbon summit, and as former NKR leader Robert Kocharian acceded to the Armenian presidency and negotiated for both Armenia and Karabakh from 1998. Since then the Minsk Group has narrowed to become the near-exclusive preserve of presidents and foreign ministers: 'the presidents refuse any translation by their own staff. We work in English with Aliyev and in Russian with Sargsyan and there are no local interpreters involved. Overall, there are not more than five people involved from both sides.'5?This format affirms the exclusive legitimacy of the heads of state, emphasises the interstate dimension of the conflict and denies agency to other actors. One Azerbaijani official likens dialogue between Armenians and Azerbaijanis of Karabakh on status and security to talks between Presidents Aliyev and Sargsyan on nuclear disarmament: 'They simply cannot decide on such matters because it is beyond their remit.'58 According to another Azerbaijani official: 'of course the two communities in Karabakh do not decide anything'.59 Articulated in this way, 299 Armenia and Azerbaijan the 'communities approach' unravels the initial inclusivity of the Minsk Conference mandate, and deprives constituencies on both sides of a voice in matters relating directly to their own status and security. Azerbaijan has also sought to delegitimate Karabakh Armenian claims by portraying the conflict solely through the framework of occupation. Official Azerbaijani discourses reject interpretations of the conflict in terms of 'civil war', 'inter-communal violence' or 'self-determination'. Emphasis on Armenia as an occupying power excludes the idea that there are legitimate local grievances in Nagorny Karabakh. By this reading, as several Azerbaijani policy-makers have impressed upon me over the years, there is no conflict in Nagorny Karabakh. There are only interfering geopolitical forces, to which only securitised responses are appropriate. Azerbaijan has also sought to map the conflict onto the discourse of the global 'war on terror'. Addressing the United Nations Security Council on 4 May 2012 during Azerbaijan's non-permanent accession to that body, President Ilham Aliyev argued that 'Areas affected by armed conflict - especially territories under foreign military occupation - create conditions conducive to networking between terrorists and those acting in such territories.'6o Securitising external powers as the source of conflict justifies colossal military spending, the secrecy that enshrouds these flows, and precludes attempts to initiate more liberal policies that could acknowledge local dynamics. Armenians similarly depict Karabakh Azerbaijanis as illegitimate interlocutors, with whom dialogue is 'a waste of time. We maybe could have discussed this in the 1990s but not now.'61 In a symmetry of exclusion, de facto officials depict Karabakh Azerbaijanis in much the same way as they are themselves depicted in official Azerbaijani discourse, as an instrument of hostile state power rather than a community with legitimate concerns, grievances and rights: 'we exercise statehood, but Karabakh Azerbaijanis are merely appointees'.62 When posturing for international audiences, official Karabakh Armenian discourses admit the possibility of Azerbaijani displaced community return. In reality, a strong taboo on any interactions in a 'bicommunal' format reinforces the gradual effacing of the historical 300 'Land for Peace' presence of Azerbaijanis in Nagorny Karabakh. The liberalism of the,Karabakh Armenian self-determination claim thus gazes exclUSIvely outwards. This also entails a denial of the claim of the larger Azerbaijani population internally displaced from the occupied territories surrounding the former oblast. Return to these areas, according to this logic, must be symmetric with the return of Arrr:enian refugees, to ot?er parts of Azerbaijan.63 Through such dls~urslve deflectIOn, dIsplaced populations and the spaces from whIch they were displaced are homogenised in the service of a chilling equation dictating that all displaced persons must return, or none will. Authoritarian actors also seek to control the media environment surrounding a conflict, and to suppress alternative sources o~ information and interpretation contradicting official lines. Smce the late 2000s, Azerbaijan has compiled a 'black list' of foreign citizens deemed to have visited Nagorny Karabakh illegally.64Since 2010 between 50 and 100 people have been added to t,he list ev~ry year; by February 2018, 707 people appeared ?n ,It. AnalYSIS of ,th~ pro~essions of those visiting the territory mdlcates that the lIst IS a hIghly targeted instrument driven by an awareness o~ the Importance of narrative. Journalists, including bloggers, WrIters and other media professionals accounted for just over 30 per cent (215 people). Foreign parlia'mentarians and those visiting the territory in the capacity of observers of its de facto elections are also a particular focus. The denial of access to seceded territories is a characteristic element of counter-secession strategies, allowing for the ongoing expression of the parent state's claim to the territory. But in 2017 the blacklist hit the headlines when Azerbaijan prosecuted blacklisted Israeli-Russian blogger Aleksandr Lapshin for having visited Nagorny Karabakh after securing his extradition from Belarus.65In what was the firs~ ?ros~cution of its kind, Lapshin was sentenced to three years' ImprIsonment (he was released three months later). , The blacklist polarises the transmission of alternative narratIves about life in Nagorny Karabakh, by leaving it in the hands of those alrea~y committe~ to the territory's Armenian identity or who ,have lIttle to lose m their relationship with Azerbaijan. The mam casualty of this situation is the field of independent 301 Armenia and Azerbaijan knowledge about the territory, and the possibility of triangulating the hegemonic narratives disseminated by the parties with observations from the field. Censorship, moreover, has not been limited to contested territory in Nagorny Karabakh. In 2011 US Ambassador Matthew Bryza to Azerbaijan was denied access to the site of an ancient Armenian cemetery deliberately destroyed at Julfa (Jugha) in Nakhichevan in 2005.66 Communalisation Nationalism has of course been present in different forms since on the onset of the Armenian-Azerbaijani conflict. Communalisation describes a more specific process aimed at homogenising political identities through ahistorical story-telling centred on ethnic identities unqualified by time, place or circumstance: The enemy here is mythical, it's not about real people or real agendas. If you look back at the original demands, they were to do with textbooks and TV towers, not this zero sum fight to the death. It's a fight with an imagined enemy, sustained by historical memories.67 As we saw in Chapter 5, communalisation is central to demobilisation by framing other kinds of identity positioning as illegitimate. In Armenia, a communal narrative structured around eternal images of the self and 'the Turk' has vied with a legal-political discourse of self-determination since the onset of the conflict. Militarism, considered by many necessary in order to mobilise sufficient resources in the context of an asymmetric conflict, competed with liberal discourses focused on rights, emancipation and harmonisation with global democratic norms. Growing Line of Contact violence, and April 2016's 'four-day war' in particular, strengthened the hand of the former. In the aftermath of the 'four-day war' figures in Karabakh condemned premature institution-building efforts, calling for an exclusive emphasis on the military.68 In Armenia, this emphasis took the form of the 'Nation-Army' concept announced by Minister of Defence Vigen Sargsyan in October 2016. According to Sargsyan, the concept envisaged that the 'entire population, not just those who serve in 302 'Land for Peace' the armed forces, should have many scientific, economic, industrial or other projects related to the army'.69 Initially manifested as a 1,000-dram levy (around $2.30) on monthly salaries, the concept extended to new recruitment programmes, amendments to arrangements for draft deferral among students, and directing financial and human resources towards the domestic defence industry. Public demand for the eradication of corruption in the army, heightened by Armenian losses in April 2016, was genuine. However, the Nation-Army concept spread into the sphere of 'military-patriotic education', identifying, inter alia, 'the existence of citizens reluctant to protect the country' and the 'tendency to communicate foreign cultural values' as internal threats. This signified for Armenian liberals a wider project in communalised militarism, taking Armenia in the direction of a 'garrison state' peopled by soldiers bound by duty, not citizens endowed with rights.?o That most Armenian citizens did not share this vision was vividly demonstrated less than two years later in April 2018's Velvet Revolution. In Azerbaijan, the shift to communalisation was starkly illustrated by the case of Ramil Safarov, an Azerbaijani military officer convicted of the gruesome murder of an Armenian counterpart, Gurgen Markaryan, with an axe at a NATO training seminar in Budapest on 19 February 2004. Sentenced to life imprisonment by a Hungarian court, he was extradited to Azerbaijan in August 2012 where he received a hero's welcome, a presidential pardon and promotion, and financial reward.?l The Safarov case caused a furore in Armenia; in Azerbaijan, it marked a significant transition from the late 1980s, when those killing Armenians in communal violence were depicted as circumstantial hooligans. In the aftermath of Safarov's return, in January 2013 former parliamentarian and celebrated novelist Akram Aylisli released a draft novella, entitled Stone Dreams, set against the backdrop of the anti-Armenian pogrom in Baku in January 1990.72 Depicting Armenians in an empathetic light, Aylisli was publicly condemned, his books burnt in the street, and a bounty placed on his ear by a member of parliament.?3 Azerbaijani arguments highlighted Safarov's origins in the occupied city of Jebrayil, and the fact that Armenian political culture has similarly lionised terrorists operating in groups such 303 Armenia and Azerbaijan as the Armenian Secret Army for the Liberation of Armenia Whataboutism is ubiquitous in Armenian-Azerbaijani rhetorics' but in ,this case ignores the fact that it is Azerbaijanis and no~ Armemans who seek a future cohabitation. The tragic irony of the Safaro:,c~se is the convergence it signified with a much-quoted an~ c~lt1Clsed comment by Armenian President Robert Kocharian clalmmg a fundamental 'ethnic incompatibility' between the two nations,?4 ,Sinc,e the late .2000s Azerbaijan has effectively implemented,~hls aXlOm as polley through a near-total ban on entry into Aze~balJan by any ethnic Armenian - whatever their citizenship -:- wIth extremely rare and choreographed exceptions usually relatmg to political or sporting events. Others, such as Turkish citizen and professional pianist Burak Bedikyan, scheduled to perform at the tenth anniversary party of mobile phone company Azercell in 2006, or Estonian citizen Karina Oganesyan, a local government official in Tallinn and delegation member to a conference in Baku in March 2018, are deported on arrivaJ.75 Coercion A third aspect to Armenian and Azerbaijani authoritarian conflict strategies is coercion. This is a long-term trend with roots in the dynamism of evolving power asymmetry, accelerating in 2014-15 and culminating in the 'four-day war' of April 2016. These destabilising years saw the intensifying collision of two strategies of coercion: the embedding of Armenian deterrence and a gradual shift in Azerbaijan's policy from strategic patience to compellence. Increased Line of Contact violence in 2014-16 effectively suggested 'land for security', a strategy aimed at Armenian territorial concessions in return for basic security, along a revised Line of Contact. This reading of 'land for peace' n:akes it easy to portray Azerbaijan as the party driving the coerCIve turn. As Sergey Minasyan observes, 'Deterrence is typically a strategy of preservation, while compellence is a strategy for change. ' 76 Yet both deterrence and compellence are strategies of coercion, and indeed the occupation of territory has long been seen as part of an Armenian calculus compelling Azerbaijan to submit to secession. 304 'Land for Peace' It can be argued, as many Armenian commentators have, that compellence is unlikely to succeed as a military strategy. The dynamic of truncated asymmetry discussed in Chapter 6 offers some support to this view. Yet this conclusion under-estimates the political utility of violence to authoritarian elites confronting domestic political challenges and popular mobilisation. Violence enacted at such times can be targeted just as much against the society in whose name it is committed, as it is against its direct victims.77 Where such violence succeeds in demobilising domestic challengers and homogenising political space, military objectives may be secondary. The coercive turn culminating in April 2016's explosion of violence gave political cover to a wide range of illiberal outcomes in its aftermath. These included referendums that in different ways secured and prolonged the rule of incumbents in Azerbaijan and Nagorny Karabakh. New military doctrines and legislation were introduced, such as Armenia's Nation-Army concept discussed above, and an Azerbaijani law strengthening presidential authority over other units with men-at-arms, such as the internal ministry troops, border guards and civil defence units under the Ministry of Emergency Situations,?8 Despite slow growth, continual public protest on socio-economic issues and continuing exposes of elite corruption, in 2017 both states incre~sed their military budgets. In Armenia the overhanging ~ecunty th~eat was taken to give cover to Serzh Sargsyan's renegmg of a pnor commitment not to assume the country's 'new' lead post of prime minister. This was a landmark in Armenia's steady regr~ssion from its early democratic promise, and was swiftly pumshed by an outraged citizenry in April 2018. Authoritarian conflict strategies have provided networked regimes with cover for the indefinite deferral of democratic transition. It is telling that the solution on which Armenian and Azerbaijani leaders have come closest to agreement, the plans for a territorial swap of 1999-2001, avoided any need for transitions to new institutions or power-sharing arrangements challenging the flows of networked power. Authoritarian conflict strategies have also supplied leaders with otherwise scarce connective tissue with reservoirs of genuine popular sentiment, committed to the status quo in the case of Armenians, and unreconciled to it in 305 Armenia and Azerbaijan the case of Azerbaijanis. In this sense, these strategies are hardly 'authoritarian', but benefit from broad-based popular sUppOrt for illiberal approaches to conflict. In Armenia, the 2018 Velvet Revolution demonstrated the limits to their effectiveness in deflecting the citizen's gaze from the domestic encroachments of an authoritarian state. Yet a central problem confronting the new Armenian leadership was whether Armenia could be liberalised while still upholding the practices of an illiberal peace, or whether only a limited variety of 'garrison democracy', undermined by continued reliance on islands of authoritarian practice to sustain rivalry, was possible. In conclusion, despite uninterrupted dialogue, myriad peace proposals, the efforts of highly skilled negotiators, the intermittent attention of global leaders, the diversion of economic resources, and the continuing loss of life along the Line of Contact, the parties' positions a quarter-century after the ceasefire on the core issue at stake - Karabakh's political status - remain as implacably opposed today as they were in 1988. This is all the more remarkable given that the same basic components of a solution have been discussed for at least a decade, sometimes two. The Madrid Principles linger on the negotiating table, neither accepted nor rejected, a meta-proposal for peace that serves the sole - if still important - purpose of justifying continued dialogue. This outcome is often interpreted in terms of absence and insufficiency: of political will, preparation of societies, statesmanship, peacebuilding impact, and so on. In this chapter I have argued that mediation failure should be understood not only in terms of the absence of enabling conditions, but in terms of actively pursued strategies dissolving the basis for the liberal peace on which the current mediation approach is based. Rather than the convergence between a liberal peace and democratic transitions that informed thinking about conflict resolution in the 1990s, a dynamic emerged in Armenia and Azerbaijan that was more or less its opposite: stably non-democratic regimes developing in a co-constitutive dynamic with strategies harnessing the context of rivalry to authoritarian power. Control, communalisation and coercion form a coherent model completely at odds with the normative assumptions of a liberal peace based, inter alia, on 306 'Land for Peace' inclusive negOtIatlOns, the acknowledgement and expression of grievances, electoral mechanisms, power-sharing arrangements, and the desecuritisation of politics. They substitute the political earthquakes of compromise with a mythology of irreconcilable difference. 307 Afterword: Rivalry Unending? ~ began thinking about this project in 2013 convinced of the madequacy of 'frozen conflict' as an analysis of the ArmenianAze~baijani conflict. Over the following three years a new dynamic of vIOlence, culminating in the major escalation of April 2016 confirmed that conviction. To meet the analytical challenge of this ev~lving context, this book has argued for a new reading of Armeman-Azerbaijani conflict as an enduring rivalry. While it shares several formative aspects with other conflicts in the former Soviet. Union, the Armenian-Azerbaijani rivalry has increasingly n:ore .m common with long-term militarised and violence-prone nvalnes elsewhere in the world. It shares with the India-Pakistan and Arab-Israeli rivalries features such as territorial contestation, inconclusive strategic interactions, diffusion across fractured reg~onal environments, the involvement of great powers, and natIOn- and state-building processes under conditions of longterm, competitive militarisation. A key implication is that the Armenian-Azerbaijani enduring rivalry cannot be understood through single-factor analysis. Rather, its persistence needs to be explained by the convergence of international, strategic, domestic and leadership factors. At the international level, the enduring rivalry framework questions explanations of post-Soviet conflicts extrapolating causalities from the wider state of Eurasian geopolitics. Since their emergence in the early 1990s, understandings of the nature of these conflicts, and the terminology used to describe them, have taken their cues from over-arching scripts of geopolitics in Eurasia. An evolving terminology described conflicts first as residual ('ancient hatreds', 'post-Soviet conflicts') in the aftermath of 308 Afterword the Soviet collapse, then as inactive ('frozen conflicts', 'no war, no peace') thro.ugh the period of relative detente to 2008, and more recently as mstrumental ('Putin's frozen conflicts') in the era of confrontation that followed. This study acknowledges the roles of outside actors, above all Russia, in lending Armenia the necessary powe~ to maint~~n the rivalry despite the growing power asymmetry WIth AzerbaIJan. But the problem with geopolitical and great p.ower-centred explanations is that the Armenian-Azerbaijani nvalry has outlasted several distinct geopolitical conjunctures and regional shocks, including Russian-Western rapprochement in the 1990s, 9/11 and the global 'war on terror', uprisings and civil war in the Middle East from 2011, the Georgian and Ukrainian crises of 2008 and 2014 respectively, and the Russian-Turkish crisis of 2014-16. None of these has been sufficient to alter the configuration of rivalry between Armenia and Azerbaijan. To explain this, this study has argued that an instrumental or competitive attitude towards the rivalry among outside actors is less important than the peculiar balance of power that it sustains among ~he~. An:ong Eurasia's conflicts it remains unique in the scale of Its dIffUSIOn across both regional and global contexts and ~s a post-S?viet theatre where an external consensus on pre~entmg escalatIOn has remained solid. Russia, an aspiring regional hegemon and a global entrepreneur of authoritarian conflict management, is embedded within the deep structure of the rivalry because of the power asymmetry. As the only external state with treaty obligations in the event of all-out war however Russia is also a key stakeholder in the tactical con~ensus with EuroAtlantic partners on deterring renewed Armenian-Azerbaijani war. Yet that consensus appears incapable of conversion into a strategic partnership to bring about positive peace, and would surely be tested if Armenian-Azerbaijani negotiations were to move i~ the. direction of a negotiated agreement - particularly one takmg lIberal form. The prospects of this are dim however O\:in? to the frac~uring of the global policy landsca~e dealin~ WIth mternal conflIcts, and the emergence of authoritarian models of conflict management as a rival to the liberal peace. The liberal peace is in retreat across the world, and across post-communist Eurasia in particular. The Armenian-Azerbaijani rivalry is 309 Armenia and Azerbaijan -consequently embedded within a wider regional context where liberal schools of conflict resolution are likely to recede further. The OSCE's Minsk Group, for now, quietly works around these contradictions. The regional policy landscape towards territorial conflict is highly fractured. There are inconsistent approaches by both Russia, which recognises some - but not other - de facto states as independent states, and the Euro-Atlantic powers, which enact sanctions in support of some parent states - but not others. This inconsistency intersects with the quite distinct projects in hegemonic regionalism pursued by the European Union and the Eurasian Union (EAEU), neither of which appears likely to offer inclusive regional ties capable of influencing the ArmenianAzerbaijani rivalry in the foreseeable future. Europeanisation, once considered plausible as a route to resolving Eurasian con- flicts,l confronts both a lack of appetite for membership perspectives in either rival and the wider retreat of the liberal peace. The EAEU, meanwhile, comprises a security community of illiberal states invested in authoritarian models of managing conflict that is deeply unsympathetic to territorial revisionism. There is consequently no meta-region or security community bridging the security and normative priorities of both rivals within which the rivalry could be embedded and de-escalated. Also at the regional level, the absence of connective infrastructure and the truncated power asymmetry are central to the rivalry's persistence. Economic interdependence is a commonly cited variable in reducing conflict between states, by establishing common interests in peace, cross-border flows and stability. The absence of economic relations between Armenia and Azerbaijan is a direct result of the fracturing impacts of the rivalry. Owing to its oil and gas reserves, Azerbaijan enjoys a high level of structural autarchy. Armenia's isolation from regional projects and opportunities for development have in turn driven alternative strategies of dependency on Russia and on remittance communities, again overwhelmingly located in Russia. Yet isolation has never been sufficient to force Armenia to accept the opening of communications and borders as a substitute for its desired political status for Karabakh. Conversely, there are few grounds to convince 310 Afterword Azerbaijan that Armenia would become more politically tractable should its economic isolation lessen. Across the rivalry alternatives to regional trade, whether in the form of petro-dollars, diasporic funds or interstate credits, have been effective in supporting regimes while bypassing societies. This inhibits economic diversification and the development of socio-economic classes and interests autonomously from the state. There is consequently no basis for interdependencies to develop among plural and diversified actors and groups across the rivalry. The truncated power asymmetry across the rivalry is a critical factor in its persistence. As the parties both understand, especially after April 2016's 'four-day war' did not escalate into a wider conflagration, this dynamic makes the chances of a major war slim. The low-intensity conflict over recent years indicates instead the presence of a stability-instability paradox, whereby major war is unlikely but there is increasing frequency of minor skirmishes and contained escalations. These take place, however, in a context marked by multiple and overlapping deterrents Armenian and Russian - that have distinct strategic goals and targets. The risks of recursive, low-level violence in this context are high. Yet because of the truncated power asymmetry, neither rival has a logic for concessions. Armenia has developed what is certainly an uneven and unpredictable but still functional deterrent against an Azerbaijani blitzkrieg, diminishing the prospect of an existentially threatening war. Azerbaijan meanwhile continues to see its resource profile and development prospects as leading eventually to an overwhelming preponderance. The truncated asymmetry dynamic has two important implications. The first is Russia's strategic insertion into the rivalry, underscoring the fact that Russia's greatest single source of leverage over Armenia and Azerbaijan is the rivalry between them. The second is that while Azerbaijan's capacity to coerce Armenia outright is limited, for as long as Armenia must devote a substantial share of its resources to the rivalry and shape its geopolitical alliances accordingly, Azerbaijan effectively holds significant veto power over its future.2 At the domestic level, it needs to be constantly stated and restated that Armenian-Azerbaijani rivalry is not enduring because of fundamental cultural - still less religious - differences between 311 Armenia and Azerbaijan Armenians and Azerbaijanis. Rather, this study has sought to contextualise the Armenian-Azerbaijani rivalry against the contingency of the territorial homelands inherited from imperial and Soviet rule. In their Soviet templates Armenian and Azerbaijani homelands generated compelling perceptions of incompleteness that first cultural, then political elites found to be irresistible sources of social capital and power. Territorial unity became a compelling icon of national identity in both Armenian and Azerbaijani geopolitical cultures. But while irredentist ideas were more marginal in Azerbaijani geopolitical culture, they became hegemonic in its Armenian counterpart. The component of irredentism, hardwired into Armenia's political culture by the advent of sovereignty, military victory and the post-war migration of leadership from Nagorny Karabakh to Armenia, embedded a political elite defined by the idea of unification. This entrenched an irredentist ideology within Armenia's domestic politics to an unparalleled degree in any post-Soviet conflict. An imaginary parallel would be the capture of the Russian presidency by natives of Crimea for twenty years. Conversely, rather than the unity of the ethnic nation, it was the integrity of the territorial state that came to define resistance to Armenian irredentism in Azerbaijan. Historically, states pursuing irredentism as a geopolitical creed have often been punished. For some scholars, indeed, irredentism and self-destructive behaviour go hand in hand.3 Within Armenian geopolitical culture, however, affective commitments to territorial revisionism are perceived as positive and worthy ideals: the aspiration to self-rule, the right to self-determine, the obligation to prevent genocide, and the desire to overcome the burdens of an exceptionally traumatic twentieth-century history. The unrecognised Nagorno-Karabakh Republic (NKR) is the geopolitical embodiment of these ideals. But as is the case elsewhere in the post-Soviet space, the pursuit of these ideals has also been a 'will-to-power disguised as virtue'.4 The NKR, particularly in its maximalist boundaries, silently incorporates the same practices of ethnic cleansing and exclusion to which it is also a response. An alternative project, which I have termed 'compliant Armenia', to normalise Armenian geopolitical culture - including through an interrogation of irredentist ideas - was a significant casualty of this 312 Afterword will-to-power. In the process Armenia conceded wider concerns over statehood, relations with neighbours and the country's place in the international state system to the politics of sovereignty over a peripheral and contested borderland. (last accessed 3 July 2018). The 'four-day war' is covered in detail in Laurence Broers, The Nagorny Karabakh Conflict: Defaulting to War (London: Royal Institute for International Affairs, 2016). 2. This trope headlined an op-ed on an earlier round 0fff Ar,meni~n,Azerbaijani skirmishing in August 2014. Brenda Sha er, Russia s Next Land Grab' The New York Times, 9 September 2014, (last accessed 12 July 2018). 19. Helsinki Additional Meeting of the CSCE Council, 24 March 1992, Summary of Conclusions, Article 9. 20. OSCE, Lisbon Summit Declaration, Articles 20 and 21. 21. Armenia's objections to these formulas were registered III a separate annex. 379 Armenia and Azerbaijan 22. Moorad Mooradian, 'The OSCE: Neutral and Impartial in the Karabakh Conflict?', Helsinki Monitor 2, 1998, pp. 5-17, p. 14. 23. On this distinction see Gerard J. Libaridian, The Challenge of Statehood: Armenian Political Thinking since Independence (Watertown, MA: Blue Crane Books, 1999), pp. 55-7. 24. This discussion draws on P. Terrence Hopmann, 'Minsk Group MediatIOn of the Nagorno-Karabakh Conflict: Confronting a? "Intractable Conflict"', Institut fur Friedenforschung und Sicherheitspolitik an der Universitat Hamburg (IFSH), OSCE Yearbook 2014 (Baden-Baden: IFSH, 2015), pp. 167-79. 25. Remler, Chained to the Caucasus, pp. 70-1. 26. Ali Abasov and Haroutiun Khachatrian, Karabakh Conflict. Variants of Settlement: Concepts and Reality (Baku: Areat and Yerevan: Noyan Tapan, 2006), p. 155. 27. 'Land for a liberal peace' similarly underpinned other plans put forward by practitioners and academics in the 1990s. Two examples are the peace plan devised by John Maresca in 1994 and a thirteen-point framework for a solution put proposed by s~holars David Laitin and Ronald Grigor Suny in an article published in 1999. Both envisaged the formation of a Republic of NagornoKarabakh within the preserved boundaries of Azerbaijan, compensated by wide powers of self-government. John J. Maresca, War in the Caucasus: A Proposal for Settlement of the Conflict over Nagorno-Karabakh (Washington DC: United States Institute of Peace, 1994); David D. Laitin and Ronald Grigor Suny, 'Armenia and Azerbaijan: Thinking a Way Out of Karabakh', Middle East Policy 7: 1, 1999, pp. 145-76. 28. Yevgeny Primakov, Russian Crossroads: Toward the New Millennium, trans. Felix Rosenthal (New Haven and London: Yale University Press, 2004), p. 201. 29. Interview Azimov. 30. See Carey Cavanaugh's comments in Carey Cavanaugh, Hamlet Isaxanli, Ronald Suny and Brenda Shaffer, Negotiations on Nagorno-Karabagh: Where Do We Go from Here? (Event Summary) (Cambridge, MA: Caspian Studies Program, Harvard University, 2001), p. 10. 31. Tatul Hakobyan, Karabakh Diary: Green and Black (Antelias: n.p., 2010), pp. 257-8. 32. Personal conversation, Vafa Guluzade, Baku, February 2005. 33. See Hakobyan, Karabakh Diary, pp. 260-1; Thomas de Waal, 380 34. 35. 36. 37. 38. 39. Notes Black Garden: Armenia and Azerbaijan through Peace and War, 2nd edn (New York: New York University Press, 2013), p. 276. Author's interview with Vartan Oskanian, former Foreign Minister of Armenia 1998-2008, Yerevan, 10 March 2015. Author's notes, meeting with former Azerbaijani official, Baku, February 2005. Levon Ter-Petrossian, 'War or Peace? Time for Reflection', Respublika Armenii, 1 November 1997. Interview with Arkady Ghukasian, Parts of a Circle, Film Three: The Search for Peace (London: Conciliation Resources, 2015). See Tabib Huseynov, 'Mountainous Karabakh: New Paradigms for Peace and Development in the 21st Century', International Negotiation 15 (2010), p. 16. The full text of the Madrid document is available on the website of the ANI Armenian Research Center, (last accessed 12 July 2018). 40. Interview Oskanian. 41. Interview Azimov. 42. Ibid. 43. For competing Armenian and Azerbaijani readings of five of the key Madrid Principles see the series of briefs produced by the Karabakh Contact Group, a civil society and expert-level format supported by peacebuilding NGO Conciliation Resources, (last accessed 12 July 2018). 44. Interview Oskanian. 45. For details see Remler, Chained to the Caucasus, pp. 99-111. 46. Ambassador James Warlick, 'Nagorno-Karabakh: The Keys to a Settlement', 7 May 2014, (last accessed 12 July 2018). 47. For an overview of these debates see Susanna Campbell, David Chandler and Meera Sabaratnam (eds), A Liberal Peace? The Problems and Practices of Peacebuilding (London and New York: Zed Books, 2011). 48. See 'Special Issue - OSCE Mediation and Conflict Management: Unraveling Complexities in OSCE Mediation', Security and Human Rights 27: 3-4,2016. 49. See the statement by the United States Mission to the OSCE, 381 Armenia and Azerbaijan 'Closure of the OSCE Office in Yerevan', PC.DEL/579/17, 4 May 2017. 50. Oliver P. Richmond and Jason Franks (eds), Liberal Peace Transitions (Edinburgh: Edinburgh University Press, 2009); Oliver P. Richmond, A Post-Liberal Peace (London and New York: Routledge, 2011); Roger Mac Ginty, International Peacebuilding and Local Resistance (New York: Palgrave Macmillan, 2011); Oliver P. Richmond and Audra Mitchell (eds), Hybrid Forms of Peace Resistance (New York: Palgrave Macmillan, 2012). 51. David Lewis, John Heathershaw and Nick Megoran, "'Illiberal Peace?" Authoritarian Approaches to Conflict Management', Cooperation and Conflict, 23 April 2018, DOl: 10.117710010 836718765902. 52. For perspectives on authoritarianism that look beyond regime classification see Marlies Glasius, 'What Authoritarianism Is . .. and Is Not: A Practice Perspective', International Affairs 94: 3, 2018, pp. 515-33. 53. Christina Stenner, 'Understanding the Mediator: Taking Stock of the OSCE's Mechanisms and Instruments for Conflict Resolution', Security and Human Rights 27: 3-4,2016, pp. 256-72, pp. 265-9. 54. I draw here on arguments presented by Lewis, Heathershaw and Megoran in their discussion of discursive practices as a component of authoritarian conflict management. Lewis et aI., "'Illiberal Peace?"'. 55. Interview Azimov. 56. Maresca, Helsinki Revisited, p. 156. 57. Personal conversation, Western diplomat, May 2011. 58. Personal conversation, Azerbaijani official, 2014. 59. Author's notes, senior Azerbaijani official, Baku, 8 May 2012. 60. Ilham Aliyev, 'Threats to international peace and security caused by terrorist acts: Strengthening international cooperation in the implementation of counter-terrorism obligations', Statement to the United Nations Security Council, 4 May 2012. Reproduced in Agshin Mehdiyev and Tofig F. Musayev, The Republic of Azerbaijan in the United Nations Security Council 201212013, pp. 40-5, p. 40 (New York: Liberty Publishing House, 2014). 61. Author's notes, meeting with senior de facto official, Stepanakert, 19 March 2012. 62. Author's interview with senior de facto official, Stepanakert, 12 March 2015. 63. Ibid. 382 Notes 64. See 'List of Foreign Citizens [Who1 Illegally Visited Occupied Territories of the Republic of Azerbaijan' on the website of the Azerbaijani Ministry of Foreign Affairs, (last accessed 26 April 2018). There is no agreed or advertised procedure for visitors to the territory to notify the Azerbaijani authorities in advance. 65. For background see Shahin Rzayev, 'Lapshin's Case: What, How and Why?', JAMnews, 10 February 2017, (last accessed 12 July 2018); for an interview in Russian with Aleksandr Lapshin after his release see BBC Azeri, 'Aleksandr Lap?in: "Ozumu asmaml?am . .. maskalt adamlar m,mi kamerada bogub"', (last accessed 12 July 2018). 66. Radio Free Europe/Radio Liberty, 'U.S. Envoy Barred from Armenian Cemetery in Azerhaijan', 22 April 2011, (last accessed 12 July 2018). 71. BBC News, 'Azeri Killer Ramil Safarov: Concern over Armenian Anger',3 September 2012, (last accessed 12 July 2018). 72. For a collection of articles about the Stone Dreams scandal and an interview with Akram Aylisli in which he discusses his motives in releasing the novella see the special section 'Dreams in the Black Garden: Literature and the Nagorny Karabakh Conflict', Caucasus Survey 2: 1-2,2014, pp. 41-78. 73. BBC News, 'Azeri Writer Akram Aylisli Hounded for "ProArmenian" Book', 15 February 2013, (last accessed 12 July 2018). 74. Hratch Tchilingirian, 'New Structures, Old Foundations: State Capacities for Peace', in Laurence Broers (ed.), The Limits of 383 Armenia and Azerbaijan Leadership: Elites and Societies in the Nagorny Karabakh Peace Process, Accord 17 (London: Conciliation Resources, 2005), p. 66. 75. 'Deported Pianist Burak Bedikyan Thanked Turkish MFA for Immediate Reaction', Today.Az, 27 December 2006, (last accessed 5 December 2018); 'Grazhdanku Estonii s armyanskoy familiey deportirovali iz Baku', Turan, 28 March 2018, (last accessed 12 July 2018). 76. Sergey Minasyan, 'Coercion in Action: Deterrence and Compellence in the Nagorno-Karabakh Conflict', PONARS Eurasia Policy Memo 242, George Washington University, September 2012, p. 1. 77. V. P. Gagnon Jr., The Myth of Ethnic War: Serbia and Croatia in the 1990s (Ithaca and London: Cornell University Press, 2004), p.8. 78. Zaur Shiriyev, 'Azerbaijan'S New Law on Status of Armed Forces: Changes and Implications', Eurasia Democratic Security Network, 23 January 2018, (last accessed 12 July 2018). Afterword 1. Bruno Coppieters, Michael Emerson, Michel Huysseune, Tamara Kovziridze, Gergana Noutcheva, Nathalie Tocci and Marius Vahl, Europeanization and Conflict Resolution (Gent: Academia Press, 2004). 2. This dynamic echoes preponderant India's veto power over Pakistan's future development, as argued by Walter Russell Mead in an interview with Nasir Jamal, 'India Enjoys Veto Power over Pakistan's Progress', The Dawn, 15 August 2010. 3. Stephen M. Saideman, 'At the Heart of the Conflict: Irredentism and Kashmir', in T. V. Paul (ed.), The India-Pakistan Conflict: An Enduring Rivalry (Cambridge: Cambridge University Press, 2005), pp. 202-24,p. 224. 4. Gerard Toal, Near Abroad: Putin, the West and the Contest over Ukraine and the Caucasus (Oxford: Oxford University Press, 2017), p. 300. 5. Warren Zimmermann, Origins of a Catastrophe: Yugoslavia and its Destroyers, rev. edn (New York: Times Books, 1999), p. vii. 384 Index Abiyev, Safar, 192 Abkharia,4,7,19,28,124,203,204, 280 Abrahamyan, Ara, 224 'affair of the seven', 71 affective dispositions, 130-41 conceptualisation of, 126 effects of, 129 sources of, 126-9 Agabekyan, Eduard, 265, 268 Agaev, Rasim, 104 Agdam,29,37,38,98, 146, 198,283 Aghayev, Ahmad bey, 51 Albania (Caucasian), 87, 107-9, 109n,117 Caucasian Albanian school in Azerbaijan, 106-11, 114 Albanians (Caucasian), 58 Aliyev, Abulfaz see Elchibey, Abulfaz Aliyev, Heydar, 36, 231, 285, 288 association with Azerbaijanism, 64, 113-14,115 claims regarding Soviet-era policy towards NKAO, 138-9, discusses territorial swap plan with Robert Kocharian, 288-90, 292 dismissal from Politburo, 36, 132 interactions with patronage networks, 36-7, 145, 169, 171-2,173 returns to power, 38, 63 subordinates resign in 1999, 159 Western-oriented foreign policy of, 216 Aliyev, Igrar, 57 Aliyev, Ilham, 1, 114, 115, 159, 188, 216,218,219,220-1,299, 300 extends rule through referendums, 171,178 increases military spending, 185, 189 interactions with patronage networks, 170-3, 220 references to Armenia, 115, 117 Aliyeva, Mehriban, 172 All-Armenia Fund, 234, 258 Alma-Ata,33 Amasiya, 135, 137 Amnesty International, 14, 143 Arab states, 11 Ararat mountain, 73, 75, 118 plain, 20, 21, 74 Republic, 69 Aras (river), 20, 50, 52, 53, 54, 58, 59,64,82,85,111,116 Ardabil, 115 Ardahan, 66, 69, 71, 73 Armavir,74 Armenakans,68 Armenia 1 March 2008 violence, 159, 164-5,180,229 'Armenian cause', 233 Armenian oblast, 67, 82 Armenian Soviet Socialist Republic see Armenian Soviet Socialist Republic 'asset-for-debt' arrangements with Russia, 223 385