One of the least explored areas of the Nagorno-Karabakh conflict is the process of engagement of ordinary citizens and the gradual breakdown of inter-societal relations, which was occurring simultaneously with major political and social transformations. Yet the demonization of the “other” and the creation and maintenance of the enemy image remain key obstacles to successful conflict management in Nagorno-Karabakh. This paper briefly reports the results of an ethnographic investigation conducted by the author in Armenia, Azerbaijan, and Nagorno-Karabakh between July 2008 and March 2009. A total of 50 interviews were conducted, related to various aspects of the development and radicalization of the Nagorno-Karabakh conflict between the late 1980s and the early 1990s. The interview sample included a combination of elites and ordinary people. A full account and detailed conclusions of the investigation will be developed in a PhD project by the author.

In the course of the Nagorno-Karabakh conflict accepted norms of coexistence were changed, as many members of non-elites turned nationalist only in the course of collective action, not before it had been initiated. Without denying the importance of elite influence and nationalist intellectuals that feature so prominently in many academic accounts of nationalism, this short paper attempts to argue that the transformation of interpersonal relations on a micro-social level depends not so much on elite manipulation but on the continuous redefinition of past and present situations through interaction with other members of the “in-group” and partly through the continuous retelling of stories contained in rumors and other sources of unverified spoken information, which are filtered through collective memories of past events.

The impact of deliberate elite orchestration — the perception of the conflict as organized “from above” by manipulative politicians — is explicitly recognized in many popular post-conflict accounts on both Armenian and Azerbaijani sides:

We used to have very amicable relations with each other… Whenever I am told that the conflict is complex, I always respond that in fact everything is quite simple. We did not want any enmity between the two peoples… interpersonal communication is very important in the regulation of the conflict…  (author interview with an Azerbaijani activist, January 15, 2009, Baku).

We had good relationships with the Azerbaijanis… The elites are to blame for everything… They initiated all this… Ordinary people have got nothing to do with it (author interview with an Armenian woman, July 15, 2008, Yerevan).

Perhaps the most popular explanation of the conflict is what a key scholar of ethnic conflict, Stuart J. Kaufman, calls “manipulative leadership,” according to which the conflict was ignited by elites striving to preserve their positions of power in changing circumstances [Kaufman, 2001, pp. 5-7).

Yet this article attempts to suggest that the “elite manipulation” account does not tell the full story. Many of my interviewees on both sides recognized the centrality of unverified oral narratives to individual participation in communal riots and the ways in which they were taken to reveal or confirm very different truths. To the best of my knowledge, none of these stories had ever been documented or verified. For example, twenty five out of twenty seven informants I have interviewed in Armenia and Nagorno-Karabakh mentioned the story of an Armenian boy’s murder by an Azerbaijani school principal in the 1960s as fundamental to their perception of insecurity. The boy’s body was allegedly later deformed by the principal. The story of this terrible cruelty was retold in Armenian families and passed though generations as an example of a killing motivated by the ethnic otherness of the victim.

As polarization between the communities became more entrenched, these narratives increasingly provided an ethnic frame for the codification of acts of violence as ethnically targeted, ethnic in their causal structure and motivation. The Armenian Genocide provided the most obvious framework within which these developments could be analyzed. Given the centrality of the genocide to Armenian historical experience, many Armenians tended to perceive a close affinity between the policies started by Turks at the end of the 19th century and the atrocities perpetrated by Azerbaijanis.  In this context, anti-Armenian atrocities were made more comprehensible when viewed as part of the historical continuum of persecution by Turks.

In the course of my interviews in Baku, Azerbaijani political activists have also acknowledged how similar stories made engagement in violent riots seem desirable, even necessary:

…In one accident four corpses of Azerbaijanis were found, but where is the proof that they were killed by Armenians? Russians or Azerbaijanis themselves could have done it as well. However, everyone was already speculating about taking revenge on the Armenians… Another example — an Armenian man took the floor at a rally and said that in Nakhichevan his grandfather’s tomb had been defiled and an Armenian had been beheaded. But no one had ever gone there to verify… whether such a thing had really occurred. Or vice versa, an Azerbaijani man in his speech asserted that in one region three Armenians had killed an Azerbaijani national. But no one knew whether it was true or not. That’s how it all began, and then it grew bigger like a snowball (author interview with an Azerbaijani activist, 6 January 2009, Baku).

Although certain individuals clearly contributed to presenting some relatively isolated disputes of the past as ethnic in nature, treating the emotional appeal of oral narratives as a simple instrument deployed by the political elite tends to misrepresent the complexity of the situation on the ground. Thus, the opposition between the rational calculations of the political provocateur and the hot emotions of the masses seems rather simplistic.

In Armenia, Nagorno-Karabakh, and Azerbaijan, rumors appear to have filled a genuine information gap where people knew that important events (such as Sumgait) were occurring but lacked reliable sources of information given the lack of data in official Soviet media. In this context individuals tended to speculate about issues that worried them and repeated the stories that confirmed their fears, largely excluded from the formal discourse through which both sides represented their cases to outside observers.

Murat Somer (2001), an expert on interethnic relations in Yugoslavia, describes the social distancing as a cascading process that changes behavior and attitudes and, once begun, is very difficult to stop. In his words, cascades are:

…self-reinforcing processes that change the behavior of a group of people through interpersonal dependencies… Cascade models explain situations in which the individual’s incentives for taking an action, holding a belief, or conforming to a norm depend significantly on the behavior of others (p. 129).

Thus, the hardening of boundaries between the communities is best explained as a collective interactive process where the framing attempts of political entrepreneurs coalesce with incidents of conflict spreading across the population in a cascading fashion.

This is a point that is often obscured in traditional studies of nationalism, which assume that the link between thought and action is relatively straightforward: nationalist outcomes result from ideas, identities, and interests formed prior to action. However, nationalism is not only a cause of action, but also its effect (Beissinger, 2002, pp. 100-101). The Nagorno-Karabakh conflict effectively highlights that strategies of mobilization are not only cultivated but also take on a momentum of their own that ends up enveloping most actors. It is misleading to argue that the leaders of the respective (Armenian and Azerbaijani) movements had long devised plans and strategies to give the political environment a strong ethnic slant. The increasing prevalence of highly exclusivist nationalist themes was at least partly the outcome of the snowballing spread of mobilization throughout society. On the whole, the Nagorno-Karabakh case supports the hypothesis that ethnicity may only become significant as a result of conflict, not prior to it. The war largely facilitated the transformation in the perceptions of the “others” as being alien, and the redefinition of relationships with the other as dangerous and untrustworthy.


Beissinger, M. (2002). Nationalist mobilization and the collapse of the soviet state. Cambridge: Cambridge University Press.

Gagnon, J. (2004). The myth of ethnic war: Serbia and Croatia in the 1990s. Ithaca: Cornell University Press.

Kaufman, S. (2001). Modern hatreds: The symbolic politics of ethnic war. Ithaca: Cornell University Press.

Snyder, J. (2000). From voting to violence: Democratization and nationalist conflict. New York: Norton and Company.

Somer, M. (2001). Cascades of ethnic polarization: lessons from Yugoslavia. The Annals of the American Academy of Political and Social Science, 573(1), pp. 127-151.

Voronkova, A. (forthcoming). Assessing the radicalization of ethnonational activism in deeply divided societies: From conflict to violence. Unpublished PhD Thesis, Queen Mary, University of London.

The research for this article was partially funded by a grant from the Central London Research Fund, University of London. The author would like to thank the Fund for its generous support.