Until very recently international organisations and many Western states rejected engagement with Eurasia’s de facto states (Nagorny Karabakh, Abkhazia, South Ossetia and Transnistria) as legitimating ethnic cleansing and undermining the primacy of territorial integrity in Western responses to the formation of post-Soviet states. As a result, the region’s de facto states have rarely been looked at through the same approaches of transition and democratisation applied to the region’s de jure states. Instead of being seen as political environments in their own right, de facto states tend to be seen only in the context of their interactions with external actors and peace processes. This omission has been challenged by some de facto states, which have increasingly used the language of democratisation to further their claims to independence. This strategy appears to have resonated with some Western observers. Since 2003, Karabakh Armenian politicians have drawn attention to Nagorny Karabakh’s assessment as ‘partly free’ in Freedom House’s ‘Freedom in the World’ index. Although the Freedom House index presents a highly simplified system for grading democratic practices, Karabakh crucially scored higher than Azerbaijan, while rivalling Armenia.

If, as is sometimes suggested, democratisation is a prerequisite of conflict resolution we need to engage with de facto states as political systems in their own right and as participants, albeit in the margins, of the broader processes transforming the post-Soviet space. Furthermore, withholding support for democratic processes in de facto states ultimately encumbers the development of genuinely participatory and pluralist politics, on which, as most observers seem to agree, any future settlement must be predicated. Thus engaging de facto states could and should be seen as consistent with support for democratic governance, rather than as necessarily inconsistent with adherence to the principle of territorial integrity. The developments outlined above invite a number of questions. How important are recognition and membership of the international state system to democratisation outcomes? Can we distinguish between the practice and the rhetoric of democracy? If so, how has the discourse of democracy been incorporated by de facto states into their quest for legitimacy in the eyes of the international community? To consider these questions is not to a priori legitimate the existence of de facto states, but a necessary corollary to any attempt to engage them in peace processes.

Theories of transition suggest that starting points are critical to the success or failure of democratisation. In the case of Karabakh, the unfavourable legacy of violent conflict was mitigated by a number of factors. First, unlike Chechnya, Karabakh’s armed forces did not fragment into competing warlord armies rooted in ‘clannish’ affiliations. Although a power struggle did develop between the de facto president, Arkady Ghukasian, and Minister of Defence Samvel Babayan, this struggle was contained and resolved in 2000 with Babayan’s imprisonment on charges of an attempted assassination. Second, Armenia’s wide-ranging support, in economic and other terms, has mitigated the impact of war losses. Armenia (and beyond, the Armenian diaspora) provides more than half of Nagorny Karabakh’s budget, as well as a wide range of goods and services in military, energy and other spheres. Third, although unification with Armenia was the central tenet of the ‘Karabakh movement’ that emerged in 1988, it also rapidly became identified with a dissident discourse promoting democratic values. As a result the discourse (at least) of democracy was and is accorded a central place in Karabakh Armenian political culture; this provides an important resource for opposition, reflected in references to the contemporary relevance of the political values associated with the Karabakh movement.

Against these factors a number of structural drawbacks need to be considered. Although Karabakh avoided clan wars of the sort that have so debilitated Chechnya’s bid for independence, its political culture is nonetheless highly militarised. A regime of martial law is still technically in place, renewed yearly by presidential decree, while political partnerships between key individuals forged in war remain above critical reflection. While this has not forestalled the emergence of civil politics, continued martial law allows the regime to seclude certain key types of information, such as population statistics, crucial for transparent electoral processes. Second, Armenia’s support is a double-edged sword so far as democratisation is concerned. While attenuating hardship for the broader population by providing a source of externally derived resources, it alleviates the need for the regime in Stepanakert to negotiate a social contract with local society. This to some extent lends the Stepanakert regime a ‘rentier’ profile, that is, a regime maintained by resources external to the society over which it rules and before which it is thus less accountable.

A key result of the war was the ethnic homogenisation of Nagorny Karabakh through the removal of the Azeri population. This removed a key political cleavage, enabling a certain core consensus on the existence and purpose of the resulting de facto state. However, it has also closed off a potential avenue for the articulation of a civic rather than ethnic sense of membership; although multiethnic Abkhazia has yet to reap genuine democratic dividends from incorporating minorities, their existence in itself creates political space for competing visions of Abkhazian statehood. In Karabakh ethnic homogenisation has shut off debates on the nature of political membership and underpinned the effacement of a Karabakh Azeri identity.

How then do the above factors interact with the context of non-recognition to influence democratic outcomes in Nagorny Karabakh? Since declaring itself independent in 1991 Nagorny Karabakh has held three presidential elections and four sets of parliamentary elections. Although they increasingly attract observers from a range of international non-governmental organisations, as well as the Commonwealth of Independent States (CIS), they are not recognised by the international community as a whole. The ruling elite in Karabakh emerged out of the wartime leadership, which since the accession of Ghukasian as president has increasingly turned towards outwardly civic politics ostensibly rooted in multiparty politics. Opposition in Karabakh has developed in response to specifically local factors but also as a function of wider Armenian politics. The main ‘opposition’ party, the Armenian Revolutionary Federation (ARF, or Dashnaks) forms part of the ruling coalition in Armenia and is a major political force in the Armenian diaspora (see profiles); previously allied with the government in Karabakh, the ARF went into opposition after a dispute over its representation in government cadres. Other opposition parties have been formed by reform-minded intellectuals, successors to the communist tradition and allies of the ruling elite seeking to fragment the opposition vote. Elections in Karabakh nonetheless revolve less around concrete issues or policy choices than raw questions of power: who gets the impunity conferred by political office?

Regime candidates generally dominated elections until the victory of a candidate for Movement-88, a new reformist party, in the 2004 elections to the mayoralty in Stepanakert. Expectations that oppositional success would be repeated at the June 2005 parliamentary elections proved false, however. The main opposition bloc, composed of the ARF and Movement-88, won only 3 of 33 mandates with 25 per cent of the vote; the regime-backed parties Democratic Artsakh and Free Homeland dominated the vote with 64 per cent between them. Some 130 international observers, including representatives of the CIS, the British Helsinki Human Rights Group and a number of United States policymakers, observed the elections; their assessment was almost exclusively positive. Independent media and civil society representatives, however, articulated complaints regarding the conduct of the pre-election campaign, especially the alleged provision of economic incentives to vote for regime-backed parties, and changes to the electoral code removing the ‘50 per cent plus’ requirement to win in one round and the second round run-off system where this is not achieved. In sum, observers and opposition representatives did not question their validity or conduct, yet by further entrenching the incumbent regime the elections appeared to move Karabakh no nearer to a genuinely participatory politics.

What explains the divergence between the apparently assiduous conduct of the elections and their failure to act as a mechanism for internal political transformation? At least in part this may be explained by the paradoxical ways in which non-recognition structures the legitimacy of de facto governments. The withholding of recognition in a context of permanent insecurity and a homogenised population allows the Stepanakert regime to be a single-issue government embodying the quest for sovereignty. Non-recognition thus locates the internal legitimacy of the de facto state in its mere existence, rather than its adherence to democratic principles or responsiveness to society. The reification of ‘stability’ as the cornerstone of Karabakh politics reflects a tacit consensus across the political arena on the parameters of dissent under conditions of constant ‘siege’. The disparagement shown by government and opposition alike to revolutionary events in Georgia, Ukraine and Kyrgyzstan reflects this consensus. This suggests an important distinction between the rules of the game in Karabakh politics as internal players perceive them, and criteria of democracy perceived by external observers.

Nonetheless, the government of Karabakh goes to great lengths to demonstrate compliance with international expectations of democratic states. It has voluntarily implemented a number of international standards applying to de jure states. As the 2005 parliamentary elections demonstrated, it also takes great care to ensure procedural and technical regularity in its electoral processes, far more so than many regimes in the region’s de jure states. Certainly, this strategy may be seen as contributing to the regime’s internal legitimacy, yet it may also be seen as a response to Western agendas of democratisation, where the presence of certain ‘markers’ such as regular elections and multiparty politics is taken to indicate a healthy transition. This is a rational response where Western policymakers have framed the issue of recognition in terms of ‘standards before status’, as they have done in Kosovo.

In this context it seems appropriate to speak of contrasting internal and external vectors of democratisation. Internally we are witnessing an uneven and highly contested process of liberalisation not dissimilar to that of de jure states but where reformists are encumbered by the peculiar conditions of non-recognition. Externally we see the projection of democratic statehood to the outside world in support of Karabakh’s claim to sovereignty. The outcome of the 2005 elections may be better understood from this perspective. In their internal function the elections did little to channel the recent emergence of greater pluralism, serving instead to entrench the incumbent regime and fortify it from challengers. Externally, however, the elections successfully projected the ideal of a pluralistic, participatory process worthy of a functioning democratic state.

What levers exist, then, for society to shape the politics of the de facto state? Unlike civil societies in de jure states empowered by substantial external support, civil society in Nagorny Karabakh faces far greater difficulties in influencing the state. Of about 80 registered non-governmental organisations (NGOs) in Karabakh, only a tenth are thought to be active. The key problem for NGO development is the absence of resources, a corollary of the fact that until very recently apart from humanitarian assistance international organisations have been unwilling to engage in single-community programming in Karabakh (unlike Abkhazia and South Ossetia, where the United Nations, organisation for Security and Cooperation in Europe, European Union and a wide range of international NGOs implement activities – in part due to easier access). Civil society actors in Karabakh have been caught between the chairs of inclusion in region-wide initiatives, logistically and politically fraught bilateral initiatives with Azerbaijani partners and the overpowering influence of diaspora-funded and -run activities.

For international donors and the government in Baku single-community programming for civil society development in Karabakh presents a dilemma. On the one hand the implication of a poorly developed civil society is that it will not have sufficient capacity to mediate between society and state in the case of a peace settlement. Any sustainable peace settlement will demand a certain level of consensus within the societies party to it, which can only be reached through the participation of civil society in channelling different agendas, articulating public concerns and establishing the parameters of what is acceptable. On the other hand, single-community programming can and is often seen as capacity building for separatism. Yet a decade of isolation of de facto states has not brought the metropolitan states any closer to reincorporating them. It is therefore for international donors and the Azerbaijani state to ponder the wisdom of, respectively, ruling out and obstructing funding for single-community programming in Karabakh.

It is now clear that ethnic ideologies of secession have failed to garner international legitimacy for the de facto states, unlike the earlier secession of union republics from the Soviet Union. Much to their chagrin the titular nationalities of the de facto states have not succeeded in convincing the world that Georgia, Azerbaijan and Moldova in their Soviet-era boundaries are imperial states as deserving of fragmentation as the Soviet Union. However, the renewed emphasis placed by leading Western powers on notions of ‘democracy’ and ‘freedom’ has opened new rhetorical spaces receptive to the advocacy of sovereignty by and for de facto states. In this context we have begun to witness what could be termed ‘competitive democratisation’, the attempt to demonstrate indicators of democracy superficially recognisable to Western observers in advance of a significant other, in this case, the metropolitan state.

Is the ‘democratisation-for-recognition’ strategy working? Increasing interest across a range of non-governmental Western actors towards political processes in Karabakh may suggest chinks in the wall of non-recognition. However, against this it is evident that external state actors continue to make territorial integrity a precondition of conflict resolution, and, furthermore, accept the legitimacy of de jure boundaries regardless of what political conditions for democracy exist within them. This would suggest that any argument linking democratisation to recognition is a fallacy.

Implicit in competitive democratisation is a claim that beyond historical grievances or ethnic differences, the Armenian-Azerbaijani conflict is ultimately a conflict of values, structured by the adherence of Karabakh Armenians to ‘Western’ democratic values and the incompatibility of these values with Azerbaijani political culture. This is suggestive of a disturbing elision of democracy and identity, drawing upon orientalist East/ West stereotypes to transplant the ‘clash of civilisations’ thesis to the realm of possible political orders and their compatibility with Armenian and Azerbaijani identities. However, confusing political values with identity carries the danger of elevating these values ‘above’ politics. Karabakh Armenian society currently stands in thrall to the value of independence, an ideal that has become a mission and ‘higher’ vision absolving the regime in Stepanakert of the need to engage in real politics across a range of issues. In this visionary politics ‘democracy’ is being configured as a means to the end of independence, rather than a set of universally binding principles and procedures capable of transforming politics in Karabakh and structuring the inevitable future of Armenian-Azerbaijani relations. This leaves the question: for how long can the promised afterlife of sovereignty outweigh the deficit between the procedural façade favoured by the regime and a genuinely participatory politics?