During 1988-1995, Armenian policy mainly concentrated on the issue of Nagorno-Karabakh, an Armenian-populated enclave in the Azerbaijani Soviet Republic. Before the collapse of the Soviet Union, Armenians living in Nagorno-Karabakh and Armenia claimed their right to be reunited, organizing mass protests. The nationalist tide soon replicated across the Soviet Union. With the independences¹, the unsolved issue turned into the bloodiest open conflict of the former Soviet Union, except for Chechnya. In 1994, a cease-fire brokered by Russia halted the bloodshed without bringing peace to the region. Since then, Armenia, Karabakh and Azerbaijan found themselves in a “no-war no-peace” situation that not merely froze conflict on the ground, as is commonly said, but also constrained political evolutions of the involved states.

In Armenia, the foundation of a sovereign nation-state was fed by the Karabakh issue and its movement that engulfed other important issues such as struggle against communist ideology or the re-definition of national identity; when present, these two aspects of the national democratic revolution were dependant on the Karabakh issue, to such an extent that it presided over national perceptions.. As it became obvious that perestroika and democratization were failing the test in Armenia and Karabakh, determination of the Armenian National Movement (ANM) leaders grew as well, and the issues of sovereignty, democratization, and then independence were perceived as conditions for the success of the Movement in the issue of Karabakh and more broadly in all issues redefined according to “national interests”. A strategy of matching the Karabakh issue with national interest – conceived as the interest of the Armenian people - embodied in a popular Movement and then in state structures – led to a change in the term used to refer to the conflict: between 1988 and 1990, it shifted from “irredentism” coupled with “self-determination” to “self-determination” only. The Azerbaijani diplomatic stance on the other hand was to follow Moscow and treat the issue as an unmotivated claim running counter the Soviet motto of the brotherhood of people, and after the Soviet collapse as an international war provoked by an independent state’s aggression and supported by a separatist population on its territory. Since Nagorno-Karabakh has no enforceable legal status, the international community chose to treat the conflict as one opposing two principles of international law: territorial integrity versus right to self-determination. Instrumental in the evolution of the qualification of the Karabakh conflict in Armenia was the “new political thinking” of the Armenian National Movement. The ANM that led the mass movement of the late 1980’s and then the newly independent state insisted on the need for Armenia to be realistic and to establish normal relations with all neighbours, thus rejecting that Armenia had any claim towards Azerbaijan; Armenia was involved in the conflict between Nagorno-Karabakh’s inhabitants and Azerbaijan only as an interested third party preoccupied by the fate of its compatriots in Nagorno-Karabakh, ran the argument, and any solution satisfying the Karabakhtsis would be adopted by Armenia as well.

A mere empirical examination of the Armenian political landscape during the late 1980’s and the mid 1990’s calls on a surprising observation: a single team of political actors took charge of the national movement during 1988-1991 and then of the ruling functions in the country all throughout 1998, making political stability the main characteristic of this war-tormented country. This observation induces a rapprochement with Charles Tilly’s famous formula that war made states, which, in spite of some a priori reservations, may be an interesting prism through which one should understand the formation of the Armenian state². These reservations hold on a number of levels: first, the object of Tilly’s analysis is the European states from the Middle Age to the contemporary period within which military mobilisation and financial resources came to be centralized and monopolized in the hands of an institutionalized ruling group. However, the American theorist leaves little room in his major Capital and Coercion for application to recently formed states. The last chapter he devotes to states emerged from the decolonization especially in Africa and Latin America; he points out major diversions between the state-formation process that characterized Europe and the one from which new states have been emerging. The second reservation is the specificity of the transition former Soviet Union states have been undergoing: more systemic than that of Latin American states, the transition was simultaneously supposed to be political, economic and social, under heavy patronage of the international financial organizations and subjected to periodic ratification by the international community. At first sight, little room would be left within this process for accumulation of coercion. The third reservation is that Tilly situates his reflection is not only in a different location (Europe, including Russia), but also in a different time, over a millennium, whereas the history of statehood in the FSU (Russia excluded) is either recent or discontinuous in terms of time or territory, or both in the case of Armenia. In spite of these reservations, it can be argued that the Tillyan model may still be relevant to analyse state-formation in the case of emerging states, – following scholars such as Amelie Blom for Pakistan³- where war was, and often still is, a road companion of state formation.

Compared to its ex-Soviet neighbours, Georgia and Azerbaijan, and to the majority of former Union Republics, Armenia, in spite of a dramatic economic breakdown and a privatization process impeded by war⁴, displayed surprising political stability during the first post-independence years⁵. Levon Ter Petrossian, who joined the Karabakh Committee in May 1988, soon appeared to be a charismatic leader taking Armenia towards independence through a constitutional process. Whereas in Georgia and Azerbaijan, no government managed to hold office for a full term until 1993, the ANM legally ran the Republic for a decade, and laid the foundations of the Armenian state. How an economically devastated country, with inexperienced political leaders acting in a collapsed environment, can build itself a state while being at war? War is seen here as not only a destructive process but to a certain extent as a creative dynamic, a resource in state-making. The hypothesis of the present study, relying on a Tillyan model of state-making, conceives of the incentive to state formation that is beheld in the Karabakh conflict. This is not to say that conflict exerted an unequivocal influence leading to the establishment of a strong state, but rather that the political will which emerged from the conflict was enough to sustain stability and some efficiency - even at the expense of other political goals such as rapid democratization and rule of law. This study focuses on the active phase of military operations and on the period immediately following the cease-fire.

The process analysed in Europe by Tilly, with its many leaps backward and small steps forward, eventually resulted in the formation of national states, the organizational structure of domination that has become the most common format, casting aside city-states and empires as less efficient structures of domination in terms of monopoly of coercive and financial resources. A national state thus is defined by Tilly in a Weberian fashion as a “relatively powerful, centralized and differentiated sovereign organization”, which can be complemented by this second definition: the governance of “multiple, contiguous regions and their cities by means of centralized, differentiated, and autonomous structures”⁶. Contrary to the European model then, the common wisdom on post-colonization states would hold war responsible for the further destruction of already unstable state structures. Understandably, war is never the only factor determining state institutions. However, one has to wonder about the correlation of war – understood as preparation to the possibility of armed conflict, armed conflict by itself, and post conflict period with no peace⁷ - and the stable re-formation of state: can we draw from a particular case where state reformation was concomitant with the three situations – preparation from 1989 to 1990, war from 1991 to 1994, cease-fire from 1994 onwards- one more argument in favour of the introduction of the Tillyan scheme for decolonized and post-cold war state formation? In this paper the argument goes that with some minor adjustments, the Tillyan model proves relevant to understand state-formation in Armenia. However it is necessary to precise that the tension war poses to the political objective of almost all FSU states between democratization and military power is a known fact, and that the positive role that war seems to be endorsed almost unequivocally here will be questioned in a more extensive work.

A sense of existential threat, and thus a pressing security need, was Armenia’s main feature as soon as 1988. Still an opposition movement to Communist rule, the Armenian National Movement’s concern for security was directed at the Armenian people, whether in Armenia, in Karabakh or Azerbaijan. Once in power, security was conceived more widely as security of the state. Two adversaries embodied two distinct threats. The first potential adversary was the former colonizer and still foremost player in the region, Russia. Armenia’s perception of Russia as an antagonist was high before independence, yet reverted in 1992 when Armenia came to terms with the fact that Russia alone was susceptible to fill in the security vacuum provoked by the collapse of the USSR. A second adversary came into play, after the Soviet collapse, when Azerbaijan turned into an independent state. Already before the collapse, the antagonism between communist leaders in Baku and ANM leaders was growing in the light of mutual resentment starting after the Sumgayt pogrom of February 1988⁸. Attempts at dialogue existed but Moscow support to complacent Azerbaijani authorities until the brutal intervention in Baku in January 1991, nullified those attempts. The militarization of the conflict from April 1991 onwards meant that the numeric and material superiority of an independent Azerbaijan was felt to existentially threaten Armenia. The re-foundation of the Armenian state immediately comprised a need to secure itself and its people; to secure the preservation of territory and population, in a situation where these two do not coincide and where the latter is indeed endangered: “Armenia's perception of its own security, or rather its perceived threats to its security, our geographic position and our long history have greatly contributed to our desire to define our security needs in terms which take into account very carefully the behaviour and the intentions of our immediate neighbours”⁹.

Reservations about the Tillyan model as well as the adjustments required by the Armenian case study can be addressed. The acute perception of a concrete threat and the state of war dictated the terms of state-building. Tilly situates his reflection on a long historical perspective: at least a millennium, during which European states gradually emerged and eventually converged from different types of states (city-states, Empires) to national states, now the dominant type. State-making involves an aggregation of socio-political practices, often conflicting practices, which over time result in a state; as opposed to state-building which implies an intentional top-down project implemented by political elites, who are supposed to have enough resources to sustain it. Obviously, this opposition is theoretically useful, but as any archetype, it is not an exactly replicated model. When studying the formation of states over a millennium, state-making is an appropriate angle¹⁰. Yet, for new emerging states, tempo and conditions are different. In the last chapter, Tilly insists on the weight of external factors on contemporary emerging states: colonial legacies, influence of a particular power in institutional setting, importance of the international community to get recognition¹¹. New states are strongly influenced by what has been for more than a century, the prevailing state model of national state. The interdependence of states and the influence of the world system imposed by consolidated democracies - to put it roughly - leave no emerging state isolated. Political leaders of all new states, especially since the end of the cold war, have to make clear which way of state development they chose for their country. This is not to say that they have all factors under check, but that there is something like a model of a national-state and, more often than not, an explicit project to go in that direction - not withstanding social, cultural or economic specificities. State-building then can be conceived as a particular moment within the state-making process, whereby a top-down process is voluntarily aimed at and implemented to a certain extent. For reasons of available resources, this study does cover exclusively these processes. A complementary study of bottom-up processes would be necessary to understand the overall process of state formation, studying for example the effects of partial military and social demobilization from 1994 onwards, which resonate on state-society relations and shed light on recent evolutions of the Armenian state. To give but a few elements, these are characterized by a tension between growing concentration and centralization of coercion and capital in the hands of a small political and economic elite, and growing social discontent in reaction to oligarchic accumulation of financial resources and social exclusion, a common phenomenon in the FSU where Armenia holds the sad privilege of high-ranking¹². The present study is a focus on a short period of time over which new political leaders in Armenia were conceiving of their political objectives with two –possibly contradicting- requirements in mind: the first one was dictated by Armenia’s specific situation of war with Azerbaijan over Nagorno-Karabakh, and the entanglement of all forces into the conflict. The second requirement was liberalization in the economic and political realms, which was both a requirement indirectly pushed by global powers –the US, the EU and their financial institutions- that Armenian political leaders reframed to make it fit the Armenian specific needs and objectives. One of the arguments of Tilly to dismiss state-building is that “rarely did Europe’s princes have in mind a precise model of the sort of state they were producing, and even more rarely did they act efficiently to produce such a model state”¹³. This argument may be adjusted in the post cold-war period where triumphant liberalism was imposing its mark on a majority of emerging states, and even when only the “least worse of all regimes” or just a by default necessity to be integrated in global economic structures, this adhesion to economic and political liberalism strengthens “mutual influence of state relations in shaping each other”. On the other hand, international competition also plays a role within war situations, with Armenia and Azerbaijan engaged in an unequal race of armaments, giving priority to military spending over all others. The constant increase in capitalised coercion goes at the expense of other functions that the state is expected to perform such as justice, education and health, feeding social discontent. This leads us to the question of how the “positive” dynamic of war reverses and impedes development of the state, a question that we shall however reserve for a different study.

Overall we consider the seven-year period under review as a time of intense state-building within a wider process of state-making. This period in not taken as a snap-shot extracted from the longue durée. The Armenian state as it was in 1991 would be an abstraction would it not be resituated in a genealogical perspective, however shorter than that of European state formation. As Tilly himself emphasizes the accumulation of coercion and capital does not follow a linear process of growth. Some states experienced deflations of one or the other; some others disappeared momentarily or for good in the process. Accumulation was no more than a long-term tendency in Europe, eventually favouring the formation of national states. Similarly, the modern history of Armenian statehood is not linear. Modern Armenian state reappeared in 1918 on the territory of Eastern Armenia, formerly part of the Russian Empire, after centuries of non-existence. The Armenian state that reappeared in 1918 could not be said to govern a clear-cut territory, as it was involved with external wars on three of its four boarders (Turkey, Azerbaijan, and intermittently Georgia). The Dashnak government ruling the first Republic mobilized all forces to sustain heavy fighting against the Turkish armies managing in the process to hold on the rural capital town of Yerevan, which for the first time in its modern history, acquired political meaning as a centre of power and national state. The government also proceeded to land distribution and some limited tax collection in order to sustain the army, although the dramatic economic situation followed by bolshevization in 1920 cut the attempt short. Under the seventy years of Soviet rule, the Armenian Soviet Republic (only formally a state in its own right) was not the locus of monopoly of coercion; however part of its nationals –along with all other titular nationals of the Soviet Union- benefited from military formation in the Soviet army and the Military academies, with a proportion higher than in other non-Slav Republics of high-rank officers. Similarly, industrialization and urbanization of Armenia were planned in Moscow and there was no capital belonging properly to the Soviet Republics; however, the local elites in Yerevan, starting from the Brezhnev period, were successful in obtaining allocation of capital from Moscow fuelled into projects deemed as of national importance (such as the construction of the Yerevanian subway). In addition, the capital-city of Armenia, which virtually tends to substitute itself to the rest of the Republic, is already a figure of center of power in the administrated Republic, contributing to differentiation of the structure and of the communist local elites. Without going further into details of these hypotheses over the genealogy of an Armenian “proto-state” before 1991 to borrow the term of Taline Ter-Minassian¹⁴, it is necessary to point out the fact that the renaissance of the state between 1988 and 1991 was not its birth-day. The situation that emerged after 1991 was characterized by an overall sharp decrease in capital due to the tremendous economic collapse of the FSU, whereas means of coercion were benefiting from whatever financial resources available, resembling what Tilly calls state formation by capitalised coercion. In terms of national state structures, centralization and differentiation can be seen as a product of Soviet domination, whereas autonomy was almost completely inexistent.

Tilly insists on the fact that, in the long-run, war and the acquisition of coercive means strongly shaped national identities making the majority of a population identify with state objectives. The Nagorno-Karabakh issue, then war, indeed contributed to a strong unity around a cause among the Armenian people. However, it did not produce Armenian national identity, which, in its modern version, dates back to the third quarter of the 19^th century¹⁵. According to Tilly’s definition of a national-state, the USSR qualified as one, but failed short of being a nation-state, that is “a state whose population shares a strong linguistic, religious and symbolic identity”. To emphasize the difference between national state and nation state, Tilly summons the example of “militant nationalities of Estonia, Armenia and Georgia” that made this difference palpable to the Soviet Union nearly everyday before it collapsed. In seventy years time, the Soviet system had most unwillingly travelled from federation with national form- socialist content to federation with socialist form and strong national content. Each Union Republic was supposed to represent a national minority in the USSR and a clear majority on its territory. Already in the formative years of the USSR, series of clashes with Azeris living in Armenia (especially in Yerevan and in the southernmost parts of Armenia) turned the demographic balance clearly in favour of the titular nationality. By 1991, Armenia was the most homogenous Republic of the USSR, with a population of around 95% Armenian. Union Republics were given nearly all formal attributes of state; moreover, Armenia (like Georgia or Lithuania for example), had a core of strong national cohesion, that had been strengthened by the Soviet state system. A crucial element that was lacking for a stable political construction of the nation however was statehood, and the Soviet state in Armenia, despite being politically and economically dependent on Moscow, ensured some existence to the idea of statehood. It gave a fixed territory to the jurisdiction of local elites, although its boarders were at times resented as unfit to the Armenian nation or to its history. If one wanted to put it shortly, all Armenia needed was to give itself a sovereign and useable state, and the issue of Karabakh was a vector to touch upon this requirement. Indeed, when the Karabakh issue re-emerged in February 1988 in Karabakh and Armenia, the intelligentsia alone was aware of the persistence of this bone of contention between Armenians and Azeris from Tsarist times through Soviet times. However, the issue of Karabakh resonated with a number of issues that were far more strongly embedded in the Armenian “imagined community” to speak like B. Anderson, first among them the genocide issue of 1915: clichés of loss, suffering and exile evoked by the Karabakh issue and soon bestowed by anti-Armenian pogroms in Azerbaijan are among the factors explaining the extent of social mobilization in Armenia. To the dissidents that reformed a Karabakh Committee at the spring of 1988, Karabakh is perceived as a window of opportunity, a chance to seize, no matter how committed they may have otherwise been to the issue in itself. As proof of this wider resonance of the issue at the beginning of the movement, suffice it to point at the tactic sorting out that took place after a few months: the mixing of other issues such as Nakhichevan, were abandoned after having served as illustrations to the movement in order to strengthen the claim over Karabakh; to the same token, the irredentist argument shifted in favour of a right to self-determination.

Leaders of the national movement of the late 1980’s explicitly undertook a reflection on statehood, in the context of modern Armenian history (since the end of the 19^th century), and then in the particular moment of perestroika. Levon Ter-Petrossian in particular has written numerous texts and speeches, including as a scholar prior to the start of the national movement, where he explicitly deals with that notion. In a private discourse written in 1983¹⁶, he presents his evaluation of Armenian modern political history, linking three elements: self-defense capacity in case of extreme peril, re-apparition of a national state in 1918, and a correct evaluation of perceptions and objectives of other actors involved in the Armenian course. Five years later, Ter-Petrossian still considered these elements to be relevant to the situation prevailing in Armenia: mass movement, support of perestroika, claim on Karabakh in Armenia and Karabakh, new security dilemma of Armenians as made obvious by the pogroms in Azerbaijan. The existence of a national structure’s own forces is seen as the key to security. The idea of statehood reappeared in the Soviet context of discussion on a renewed Union, envisioning broader prerogatives for the Republics, however excluding a control of armed forces by the Republics. Opposed to this view, the ANM claimed that statehood must be supported by national forces, not in order to launch external aggression although this might prove necessary, but in order to sustain sovereignty and to defend the structure in itself. This is for Tilly among the vital functions a state must perform, proving existence by sustenance.

(We propose a chart to capture the political objectives of the ANM: See chart 1 in Annex)

The ANM advocated democratization not only as the “least worse of all regimes” or as the winner of the moment, but worked to give democratization a particular resonance linking it to the Karabakh issue. Without democratization they argued, Armenia had no chance of making its argument on the issue prevail. At the time of perestroika, the Karabakh question was presented as a test of Moscow’s commitment to liberalization; when the national movement explicitly affirmed sovereignty as a political objective of the movement, democratization was then necessary at the national level as the only argument that Armenia could put into the balance for international actors. Armenia has nothing else to trade than its good will and respectability.

What this chart does not capture however is the implicit tension between these objectives; they can be grasped by a system of inductive and deductive arrows in chart 2 (See chart 2 in Annex).

The actual priority was dictated by the economic collapse along with the war with Azerbaijan. Having barely re-established an independent ruling of the national state structures inherited from the Soviet Republic, Armenian decision makers were imperatively committed to sustaining this structure, by providing for external and internal safety. The concentration of men and material into a Defense ministry was further facilitated by the fact that Armenia was opposed to an external adversary, and that mobilization against this adversary prevailed over internal turfs and disputes within the administration. This does not mean that the Armenian political leaders had not their share of rivalry and opposition. If Armenia did not experience civil war (as in Georgia, leading to the collapse of the first government), or extreme political instability and high level of conflict between political and military elites (as in Azerbaijan where a number of successful and unsuccessful coups were both a factor and a consequence of that country’s difficulties during the war), fissures appeared in the Armenian political landscape as soon as 1992, with the scission between Vazgen Manukian and Levon Ter Petrossian. Still, mobilization against Azerbaijan and for Karabakh was high enough to make a last collaboration possible, with Manukian assuming the Defense Minister office in 1992-1993, and scoring important victories on the battlefield, before definitely rallying the opposition.

Tilly’s innovation is to focus his analysis of state-making on the organization of coercion and preparation of war: coercion is the first variable that accounts for state-making. States were shaped through mutual intercourse, especially one of war and competition for trade of capital. On the domestic front, war depended on the quantity and accessibility of capital. Capital is the second major variable, itself involved in coercion and in relations between states. Over a millennium coercion and capital did not continuously accumulated, yet this was the general tendency. Comparing this analysis to our case, a couple of methodological remarks are in order. First, there is very few information on Armenia’s accumulation of capital and how it was invested into the war effort between 1991 and 1994. For reasons linked to the unresolved status of the conflict, accurate information is not yet available for this formative period. Furthermore, as mentioned earlier, all official figures linked to this matter are to be considered carefully. With a collapsed economy, a production virtually to a standstill, no raw material that could be sold easily, no sea access, and two closed borders out of four, one cannot help but wonder: where did the money to wage war come from¹⁷? On how the organization of the army was funded, or arms acquired, we have only partial information. First, Armenia, as well as Azerbaijan, inherited equipment according to the size of Soviet Army units stationed on their respective territories, the 4^th unit in Azerbaijan and the 7^th unit in Armenia. Thus, Armenia started with a strong disadvantage, somewhat compensated by the role Diaspora communities played in providing equipment: boots, jackets, light artillery, especially from Lebanon and Syria. A number of political actors acknowledged, during interviews, that the Diaspora had helped in providing equipment¹⁸, but they would not specify the amount of help, merely saying that it was substantial¹⁹. Stories of how ordinary people in Armenia would enlist in volunteers’ detachments, bringing along their car, a spare tank of fuel or anything else that might be of use, are commonly heard. On a state level, the Defense Ministry was the main recipient of all state revenues during war –and still a major recipient today- but how much exactly? Interviews of political actors and archives confirm that every single production or resource generated by services (like hotels) was fuelled towards Karabakh, but no figures are available. As early as 1992, the power ministries -comprising Defense, Interior and National Security- acquired important privileges in order to organize the war effort. At this date, Ter Petrossian reckoned that the Ministry of Defense was already the most prominent Ministry in terms of budget, personnel and programs. Again, we have suspicions, but actually no extensive information.

The Stockholm International Peace Research Institute gives the following estimates for four countries. Interestingly enough, the higher ratios for Armenia compared to Georgia speak for the tremendous costs that the Karabakh war imposed on the Armenian economy.

Figures for Armenia do not include military pensions. For 2004-2006 these amounted to 9979, 1113 and 12440 b. drams respectively*

Figures for Georgia from 2002 are for the budgeted expenditure. The budget figure for 2003 is believed to be an underestimation of actual spending because of the political turmoil during the year.

Figures for Ukraine are for the adopted budget for the Ministry of Defense, military pensions and paramilitary forces. Actual expenditure was reportedly 95-99% of that budgeted for 1996-1999.

US$ m. = Million US dollars; - = Empty cell; ... = Data not available or not applicable; ( ) = Uncertain figure; [ ] = SIPRI estimate

The following figure from the UNDP gives us an indication of the proportion of the military spending in state spending, and shows that Armenia and Azerbaijan are devoting the better part of their revenues to the military, but is still incomplete for the beginning of the 1990’s.

As Tilly succinctly put it: “Roughly speaking, rulers had three main ways of acquiring concentrated means of coercion: they could seize them, make them or buy them”²⁰. In the European scenario, rulers progressively chose tax collection in money as the main source of revenues to acquire those means, and in turn used part of them in order to improve the efficiency of tax collection. In the case of Armenia, one can exclude making means of coercion. Arm smuggling in the FSU has been studied, especially from Ukraine which was the second place of Soviet arms left over, but no information is available to the best of my knowledge as to whether some of them reached Armenia. Armament was left over by the disbanding Soviet Army in 1991, but certainly not enough to wage three years of intense fighting. Others were seized from the Azeri army, when Armenian troops entered some combat headquarters hastily abandoned in Nagorno-Karabakh. But again, in spite of the usefulness of seizing a couple of tanks, it is trifle compared to war time needs. Buying them was hardly possible. Armenia certainly received significant armament from Russia at a very advantageous price in 1993, although this has never been confirmed²¹.

Therefore, at the current time of research, the dynamic of coercion alone can be borrowed from the Tillyan conceptual framework. High support of the population committed to the war effort made the gradual monopolization of coercive resources easier²². The partiality of this analysis, in the absence of reliable information about institutions exercising coercion, is inevitable. To give but one consequence: since financial accountability of the Ministry of Defense is unverifiable, it can not be concluded that war allowed for an efficient monopolization of capitalized coercion; if corruption highly permeated Defense, and blurred lines between private and public use of the Ministry resources to such a point that they could be diverted of their intended use, state institutions would not be said to be differentiated and autonomous. Given the current state of research on Armenia, all one can do is formulate a hypothesis that correlates empirical observations: security was a sufficiently high priority to make decision makers active in laying the foundations of a useable –if not optimally efficient- state such as defined by Tilly.

It comes as no surprise that the two conflicting parties in no time attempted to create their national armies; although several factors contributed towards the weakening of this attempt in Azerbaijan, in Armenia it was the first state apparatus set up with success. Azerbaijan started with a marked advantage in terms of men and left over material by the Soviet army. However, Azerbaijanis were underrepresented in the top ranks of the Soviet armed forces, and “many Azerbaijani conscripts were assigned to construction battalions, in which military training was minimal and the troops carried out non combat duties”²³. What made the strongest difference in 1993 –the pitch of war- was poor discipline and lack of individual motivation, with forced recruitment reported and high desertion rates. At the end of 1993, the Military Procuracy issued a call for all deserters to surrender or “be severely punished”²⁴. This situation eventually allowed Armenians to advance into Azerbaijani territory outside Nagorno-Karabakh proper. Between 1991 and 1993, Azerbaijani Presidents ousted more than one Defense minister for alleged incompetence, and in 1997 ex-Defense minister Husseinov was charged with organizing a state coup in 1994, 1995 and 1996²⁵.

Although some of these problems undoubtedly existed in the Armenian army as well, they did not reach proportions high enough to plague its mission. In spite of being one of the poorest countries of the ex-USSR, it was rapidly able to lay the foundations of a national army. “High national morale, feeling of imminent threat from the Azeris […] and substantial Russian help (including generals sent as military advisors)”²⁶ after 1992, are the main reasons for this. In 1990 and 1991, self-defense groups burgeoned in the Republic, in reaction to anti-Armenian pogroms in Azerbaijan and violence against a handful of Armenian villages in Nagorno-Karabakh. Immediately after his election at the Supreme Soviet, Levon Ter Petrossian had to cope with these armed battalions, making sure that their loyalty would go to the new government and that their armament was under check. Trying to undermine the newly elected non-communist chairman, Gorbachev issued a decree on July 25 giving Ter Petrossian one month to disarm all groups. Fearing that the decree might legalize a military intervention in the Republic, Ter Petrossian manoeuvred to obtain patience from Moscow, and indeed moved to control those groups. With the mediation of Vazgen Sargsian, he disbanded the so-called Armenian National Army which refused to hand arms over and to recognize the new government’s uncontested legitimacy. Other groups such as Armenian Kond, Tigran the Great, Armenian self-Determination Union and Jermuk militia, pledged allegiance to the new government, and were allowed to resume their defense activities in Karabakh, where they had been for more than a year. The best elements of these fedayi-style battalions²⁷ were merged to form the basis of a national army: on September 20, 1990, a Special Regiment was created, regrouping all militias eager to be loyal to the new leadership, consisting of 26 platoons from Yerevan and provinces, for a total of 2300 militiamen²⁸. In October 1990 the Soviet universal military draft was suppressed (the law dated back to 1967); instead, the law on the draft of Armenia’s citizens was adopted in 1990 as well as a law on Armenians serving outside Armenia. Whereas non-Slav nationalities were poorly represented in the Soviet Army, Armenians were an exception to the rule, therefore numerous soldiers and officers came back to serve in the Republic. One of the military chief architects was General Norat Ter Grigoriants, former Soviet deputy chief of staff who became overall commander of the new national army.

This was the bulk of the future regular army, which however would not be completed until after the cease-fire. From Ter Petrossian’s point of view, this was necessary to ensure that the country stayed aloof of civil war and that a unique political center emerged, uncontested by self-instituted warlords. He made clear that soldiers were not meant as a vain show of force to sustain independence but as a mean to protect the newly re-established state²⁹, precisely a sign that a state is performing one of its vital functions. However, the government did not proceed to the complete dismantling of militias before the cease-fire: the dramatic episode of Khojaly in 1992, when Armenian militiamen killed Azeri civilians on seizing the village, is proof of the existence of incompletely controlled militias in Karabakh. For the needs of war, the government left a number of them operating, but after Khojaly, proceeded to check them requiring the intervention of an undisputed charismatic figure. Vazgen Sargsian, a second-generation ANM militant at the time in charge of coordinating all military forces of the Republic, reunited volunteers in the largest paramilitary group of Armenia, the Yerkrapah (Guardians of the land) in 1993 that was later legitimized. Formation of a regular army, obeying state authority and not a charismatic leader is time and money consuming, yet the grounds of this process were laid early on.

As early as 1992-1993, the issues of threat and security guarantees prevailed in all ex-Soviet Republics. This was all the more true in Armenia, with its delicate geographic situation and entanglement in open war. After Armenia signed the Collective Security Treaty of the CIS in May 1992, another wave of Armenian officers serving in the ex-Soviet army volunteered in Armenia. At about the same time, when Abulfaz Elchibey came to power in Baku, he directly threatened Armenia proceeding to the shelling of border villages, thus giving a new dimension to Nagorno-Karabakh’s security: from then on, the need for security both in Armenia and Karabakh was on a par. Protecting the entire south-eastern part of Armenia, Karabakh is the focal point of the country’s defense system. In the years following the cease-fire, Armenian leaders worked toward making the national army a small, well-balanced, combat-ready defense force. Two political leaders played a prominent role in building the army in Armenia, and two others in Nagorno-Karabakh. In Armenia, Vazgen Manukian, from the Karabakh Committee, Defense minister from October 1992 to June 1993, and Vazgen Sargsian, from the ANM, head of the Defense Committee from May 1991 to January 1992, when the Defense Ministry was established and entrusted to him. Sargsian left over his post in October 1992, being then state minister for coordination of all security forces of the Republic before taking the charge again in 1995, until his assassination in 1999. In Karabakh, two men were instrumental: Robert Kotcharian and Serge Sargsian (no relation to Vazgen). Their action during the war propelled them to high political offices in Armenia proper (Premiership and Defense minister), paving the way for their ascension to presidency. According to Defense Ministry data, paratroopers were the first unit to be constituted, in January 1992, followed by artillery brigade, air defense force in June, a tank division in October, as well as a missile unit. One cannot help but notice the concordance between these steps and seizure of armaments from Azerbaijanis in the aftermath of victorious battles in Karabakh. Vazgen Sargsian set up the bridge uniting the two armies. A charismatic ex-writer, Vazgen Sargsian knew how to convey enthusiasm and belief in him as well as in his grand ideas. Addressing people from the ground, where he liked to be, he worked hard to gather followers during and after the war: “Everywhere I am looking for friends, for loyalists, for companions. Thus both I and the army are getting stronger”³⁰. He gave the image of someone involved with ordinary people, having its share of hardships and bringing its contribution to the (military) success of the country: “This is not a country you can love with words or look at from afar. […] Borders here are traced with blood and asserted with sweat. […] Laws are born at the borders, they are born in need, and they enter parliament to get dressed, simply”³¹.

As expressed by the military establishment during the first stages, the objective of the military was to ensure defensive self-sufficiency, enough to repel an attack from Azerbaijan at least and virtually from Turkey too. In a more aggressive form, this doctrine would envision high degree of readiness to inflict crippling losses during initial phases of an attack, similar to the Israeli army doctrine of defending a surrounded land. Both doctrines emphasized small, highly mobile, well-trained units. Some figures are eloquent even if they must be considered with precaution³²: the initial goal of 30 000 soldiers on active duty was surpassed in early 1994 by 5,000, in spite of a growing desertion rate in 1993 –the bloodiest year of war for both sides³³. In 1996, the total number of soldiers in the army was between 50 000 and 60 000 (including reserve) for a total population of 3.5 million, according to official and optimistic figures. The Karabakh army as well was formed along the same lines. Back in 1993, the decisive year of the war, defeats that Armenians inflicted on Azerbaijan were attributed largely to the self-defense forces, although regular Armenian forces were involved too. In 2005, with 18 500 soldiers in active duty (8 500 from Karabakh and 10 000 from Armenia), and at least as many reservists, for a total population of 140 000, Karabakh is a highly militarised society³⁴. In addition to the Russian military alliance, the national army in Armenia and Karabakh appeared as the only guarantors of a fragile security.

The need for a security system at all costs led Armenia to seek Russia’s protection, despite a suspicion towards it and the will to lessen ties with Moscow. In 1990-1991, an important factor weighing on Armenia’s decision to distance itself from Moscow and to reconsider its relations with other countries, primarily with Turkey³⁵, was the attitude of Moscow towards Armenians in Karabakh, Azerbaijan and Armenia during perestroika. In none of the pogroms of Sumgayt (1988), Kirovabad (1988) and Baku (1990), did Moscow intervene timely; in addition, units of the 4^th army stationed in Azerbaijan and Azeri OMONs were used in “Operation Ring”³⁶, to empty a number of Armenian villages in Nagorno-Karabakh in April 1991. At the beginning of May, Ter Petrossian reckoned that Armenia was in a state of undeclared war with the Union³⁷. The Soviet credo that Russian rule was a necessary protection against Turkish threat was seriously undermined³⁸. The situation reversed when, in the wake of the collapse of the USSR, Russia got involved in discussions regarding a cease-fire, this time on a governmental basis. In addition to the cordial relations between Russian president Boris Yeltsin and his Armenian counterpart, Russian diplomatic involvement in the conflicts on its southern periphery changed the nature of relations between the two countries. The convergence of interests between them was soon expressed in a bilateral Treaty of Friendship signed in 1991, which, for the first time in Russo-Armenian history, recognized de jure equality and sovereignty of the “little brother”. Armenia’s military dependence upon Russia began as early as 1992, with the capture of Shushi, -the historic centre of Nagorno-Karabakh- and has been growing ever since. From the Collective Security Treaty, signed in May 1992, to the 1997 bilateral Treaty of Friendship, Cooperation and Mutual Assistance, Ter-Petrossian’s administration tried to build a security alliance that would guarantee Armenia’s protection³⁹. In 1994, an essential agreement was signed providing for borders protection: the Turkish-Armenian border and the Iranian-Armenian border were to be guarded by mixed troops of Armenians and Russians⁴⁰. What does the alliance with Russia indicate in terms of state monopoly over coercive means? Going back to Tilly’s definition of the state as “an organization wielding coercion” and exercising incomplete but still predominant control over coercive apparatuses, it is possible to conceive of a non-hegemonic state still sovereign enough to be the prevailing force controlling coercive resources. The military alliance with Russia is indeed capital for Armenia, but concerns two of its borders: the border with Turkey and the border with Iran. The aim is to deter any possible incursion of Turkish troops in Armenia and of Azerbaijanis to Nakhichevan, as relations with Iran have been cordial all through the period. Along with ensuring security, agreements with Russia provide for assistance in military organizational development and education in Armenia, besides a possibility for Armenian officers to complete their education in Russia. Therefore, strong links with Russia are also a way to pursue army building and improve the formation of militaries. This move towards Russia was dictated by the heavy burden of war costs on a devastated economy that made civilian life hardly bearable. Political leaders however having delegated part of their control over a fraction of their external forces (border guards and customs), moved to reform the internal security apparatuses. The dependence on Russia was reinforced after 1996 by extensive Russian involvement in the Armenian economy, first with massive participation in the privatization process and, under president Kocharian, with acquisition of strategic enterprises in the energy sector. To my opinion this economic involvement is more decisive in terms of loss of national sovereignty: here, we come to suspect the limits of conflict as a major incentive to state-building: when no solution is found, state-building may become hostage of conflict, conflict may dictate the evolution of the state.

The ANM’s entry in the Supreme Soviet in May 1990, then Levon Ter-Petrossian’s election as first President naturally led to an ANM-composed government and National Assembly. Yet, with Armenia being de facto at war, the power Ministers soon acquired a fundamental weight on foreign policy and economic policy, as well as in the coercion apparatuses. In 1992, the Nagorno-Karabakh Parliament gave all powers to a State Committee for Defense, presided by Robert Kocharian, who was designated by the ANM itself as one of its most devoted partisans in Nagorno-Karabakh. As early as 1993, the government of the ANM and the Defense Committee in Karabakh, assumed political offices in Stepanakert after sidelining Dashnak elements. In September of the same year, a movement signalling the stronghold of the periphery over the center started with the appointment of the director of the Committee of Self-Defense Forces of Karabakh, Serge Sargsian, as Defense minister in Armenia by Ter-Petrossian. The new minister also had the status of special negotiator in the search of a cease-fire. In December 1994, Kocharian was elected President by the Parliament of the unrecognized Republic of Nagorno-Karabakh, while still assuming the Commander-in-Chief position.

As the parliamentary elections of summer 1995 drew closer, it became more obvious that positions of power relied within the military and security forces. The Armenian President was facing a silent confrontation within the ruling elite between first generation ANM associated figures and Karabakh associated figures. Exemplifying the mute tension was the reshuffle of the National Security department (former KGB) in 1994, headed by David Shahnazaryan, a long-time associate of Ter Petrossian, and again in May 1995, this time entrusted to Serge Sargsian⁴¹. Meanwhile, the President adopted a number of measures trying to reduce power ministries’ weight and to keep public politics aloof from their presence. In 1994, their special economic privileges were abolished. With the perspective of parliamentary elections, a law on legislative election stipulated that members of the cabinet (i.e. government), Interior Ministry officers and army officers were not eligible in Parliament. This did not prevent the Yerkrapah (Guardians of the land), a volunteers group now union of Karabakh war veterans, to gain a number of seats in the 1995 election. This block, associated to the Republican faction (Hanrapetutyun) - a coalition group comprising the ANM and political groups supporting the ANM - was loyal to Vazgen Sargsian, recently appointed Defense minister. The Republican faction got a majority in the new National Assembly, but the coalition supporting the ANM was nearly as strong as it looked heteroclite. Possibly uneasy with the new design of the political landscape, at the 7^th congress of the ANM in December 1995, Levon Ter Petrossian called for an alliance of the “liberal democratic forces” and “right wing forces”, excluding Yerkrapah from the list⁴².

However, one should be cautious with the Yerkrapah phenomenon. Yerkrapah is not an offset of the army, and its entry in parliament did not signal the establishment of a military regime in Armenia:

Back in 1995, Yerkrapah had two main goals: with the resumption and acceleration of economic reforms, privatization and foreign investments, Yerkrapah wanted to protect economic interests of veterans and their families, and on an ideological level, to unite a patriotic discourse to a social one. Defense minister Vazgen Sargsian was their spiritual leader, and at the same time the head of the army. He conceived of Yerkrapah’s role in public life as a bridge between different segments of society and always praised unity of the people as the supreme value. However, he was outspoken on the non interference of the army or the Defense minister in public matters independently from the institutional hierarchy. In other terms, only a direct order from the President could bring about such an interference; which happened in the turmoil following Presidential elections in 1996, when the military was deployed to stop the opposition from violently seizing power. A year later however Vazgen Sargsian got publicly involved in non-defense matters, this time on his own account: in September 1997 a disagreement over the Karabakh negotiation process was made public. It opposed the President to Prime Minister Kocharian allied with National Security chief Serge Sargsian. Defense Minister Vazgen Sargsian oscillated between the two, before eventually rallying the “Karabakh clan” for ideological reasons: to preserve unity of the Armenian people. This ideological trend, along with a desire from war veterans to be politically and economically rewarded for their commitment in what is perceived as a successful war, characterizes a Karabakh syndrome that translates the political and ideological weight of the issue after the cease-fire. The aftermath of the cease-fire had made public the organization of power and the growing influence of the power ministries within the Armenian state. Early in the 1990’s the situation of war accounted for the tendency towards the concentration of decision-making in a few hands. The unresolved status of the conflict made this the ordinary state of affairs.

Internal political life - be it in its most heated debates, or, more unexpectedly, in the circulation of its elites – has long been mostly conditioned by the Nagorno-Karabakh issue⁴⁴. As a recurring source of political legitimacy, Karabakh is both the main issue of political processes as well as the main factor that orientates them. The consequences of this issue weighed substantially on the change of leadership that resulted in the constitutional coup of 1998. The first leading team came to power supported by the unity of a vast majority of people around the Karabakh issue primarily. One could say that these intellectual, charismatic leaders were national leaders, in the sense that they carried an array of aspirations that were echoing in the desire of a majority, at least during the revolutionary phase of the Armenian National Movement. The second team is a direct produce of war, although it is not constituted of professional militaries assuming civilian political offices. For this reason, Armenia cannot be typified as a military regime that is a regime where political offices are the reserved domain of superior officers of the army, security forces have extra-judicial authority and armed forces are not under check of the political centre⁴⁵. This second team is one of managers linked with the military in Yerevan and Nagorno-Karabakh, a movement signalling the “Karabakhisation” of Armenian political elites. In the winter of 1997-1998, a coalition between security forces of the administration pushed president Ter-Petrossian to resign. The public reason was a disagreement over the negotiation format proposed by the mediators on the Karabakh issue. After a few months of internal struggle, the pressure put forward by Prime Minister R. Kocharian and National security chief Serge Sargsian, eventually rallied by Defense minister Vazgen Sargsian, resulted in Ter-Petrossian’s resignation in February 1998. The assassination of Vazgen Sargsian in a terrorist attack on Parliament in 1999⁴⁶ shaking the entire country left former allies of the first government, Robert Kocharian and Serge Sargsian, occupying the decisive political functions of the country.

After gaining independence, conflicts became a common feature in all three South-Caucasian states. Nonetheless, the picture was somewhat contrasted. All three countries developed an anti-colonial rhetoric but in the case of Armenia alone did this lead to a durable withdrawal of Communist leaders from the political hierarchy. In Azerbaijan, the Azerbaijani Popular Front was not successful in defeating the Communist Party and whereas the success of the dissident movement was stronger in Georgia, stability came back only with the return of Edouard Shevardnadze, former Foreign Minister under Gorbachev. Second, in terms of cohesiveness of national identity, they stood on an unequal foot. In Azerbaijan, national identity was very much a by-product of the Soviet federal system and was thus strongly linked with the idea of an Azerbaijani statehood; whereas in Armenia and Georgia a sense of national identity existed prior to Sovietization but was reshaped by the Soviet state. For Armenians, the non-matching between nation and state was strongest, both spatial (with the Diaspora) and temporal (deprivation from national state for centuries). Therefore the Soviet regime was perceived in different ways. For Azerbaijanis, sovietization had been a way of gaining formal statehood and territorial aggrandizement, making some sense of a nation emerge. Thus, the Karabakh issue appeared as a threat to the Azerbaijani state, one that might trigger other separatist claims from ethnic minorities. The Karabakh issue not only united Azerbaijanis but also crystallized their national construction in opposition-competition with Armenians. The Georgian national narrative remembers sovietization as being forced upon them, yet demographically speaking, it incompletely ensured Georgian majority on their territory, thus paving the way for nationalism turned against other inner nationalities. For Armenians, the link to sovietization was a difficult mix of political failure and national grievance that paradoxically strengthened and clearly cut the political boundaries of the Armenian nation.

A three-year war changed the orientation of the Republic and shaped a relatively stable state, yet at the expense of rapid democratization and rule of law. Had the 1994 cease-fire been quickly transformed into a peace accord, a constraining burden in the building of the state would have been alleviated. The importance of security suffocated Armenia, reduced her options, and contributed to the current narrow share of political power. The priority given to security, and determined by the Karabakh conflict, soon led the political actors of the ANM to share the decision with the military involved in the conflict. A tension rose in the Armenian political system: while, after the cease-fire proved lasting, the Karabakh issue became a question among others on the foreign policy agenda, the internal centre of decision got increasingly linked with Karabakh. Concerning the Karabakh issue, one can see two competing streams in institutions and actors. The first stream concerns actors institutionally or officially responsible for this question: presidential advisors in foreign affairs – under Ter-Petrossian’s administration they used to play a significant role - and the Ministry of Foreign Affairs, both involving numerous diplomats and western experts. The second stream brought late partisans of the ANM close to military circles and/or from Karabakh to key posts in defense and national security; this movement was initiated by the nomination of Serge Sargsian, head of the Self Defense Committee of Nagorno-Karabakh, as Defense Minister of Armenia in 1993. Paradoxically, while for Armenian decision makers the Nagorno-Karabakh issue has long been one point among others on the agenda, domestic political developments may still be read as an offspring of the unsettled conflict.

An overview of the following years shall corroborate the present analysis of the capacity of the Armenian state to monopolize means of coercion and then economic resources of the country, agglomerating so-called oligarchs around the ruling political elite since 1999, and more decisively after 2003. This trend, which is no peculiar to Armenia but commonly found in a number of FSU states⁴⁷, aroused important contestations in Georgia in 2003 and Ukraine in 2004, leading in both cases to the replacement of newly elected incumbents by their challengers. In Armenia, where election contestation has been the rule since the 1996 presidential elections however, incumbents never retired to the benefit of their opponents. The only time when the institutional calendar got disrupted was in February 1998, when then President Ter-Petrossian resigned in what a constitutional coup. Interestingly, this disruption was not brought by grass-root social opposition but by a political cabale from within the administration. Each and every national election since then has been followed by sustained protests never resulting in what Tilly calls a “revolutionary outcome”. A discussion of the various factors as to why this outcome was never produced is not in order here. We want to emphasize that stability of the political elite and their reliance on the power institutions (defense, interior, national security) as well as the subjection of the judicial apparatus (the procurator institution after 1999) were undoubtedly instrumental in maintaining the revolutionary situations under check. The broadest social movement in a decade that followed the Presidential elections of February 2008 had to be crushed by violent intervention of the coercive apparatuses of the state in March, leaving a dozen people killed. As a result of the monopolization of the 1990’s the Armenian state has become the fiefdom of a small political and economic elite that perpetuate itself, and is shut down to social bargains. Paradoxically, the high mobilization of the late 1980’s and first half of the 1990’s contributed to the emergence of stable state power institutions but reverted into estrangement from society vis-à-vis the state. Ironically, this situation cast heavy doubts on the capacity of such a state to re-mobilize its society in case of imminent threat –the Nagorno-Karabakh conflict being unsolved- whereas permanent insecurity dilemma is precisely the pretext summoned by Armenian authorities to justify the priority given to stability of the regime.