EUROPE-ASIA STUDIES Vol. 59, No. 4, June 2007, 567 – 590

Framing Ethnic Minority Mobilisation in Central Asia: The Cases of Uzbeks in Kyrgyzstan and Tajikistan

MATTEO FUMAGALLI

Abstract This article seeks to understand why, although widely predicted, very little ethnically motivated political mobilisation has occurred in post-Soviet Central Asia. Building on the contribution of frame analysis, the article examines the case of Uzbek communities in Kyrgyzstan and Tajikistan. Different frames have been articulated by the elites and have been used to mobilise and demobilise the community. The more successful ones (those that resonated) were less confrontational, and ultimately led the Uzbek communities towards a non-confrontational path with the authorities. The article concludes that frame analysis sheds important insights on the process of the mobilisation of ethnic minorities in Central Asia, and also contributes to the literature on ethnic mobilisation.

IN JUNE 1990 TENSIONS ORIGINATING FROM PERCEPTIONS OF UNFAIRLY distri- buted land and housing erupted in a sudden and short-lived, but particularly bloody series of riots in the south of Kyrgyzstan.1 Concentrated in the city of Osh2 and the nearby town of Uzgen, the Kyrgyz and Uzbek population clashed, leaving about 200 people dead and many more injured, in what seemed to announce an era of ethnic conflict in the region (Tishkov 1999, p. 581). The conflict was not an isolated case in Central Asia, as the clashes can be located in a broader trend of insurgencies and riots which marked the dusk of the Soviet period. Nevertheless the Osh conflict was one of the largest and most violent of its kind (Tishkov 1995, p. 134). Two years later, the Soviet-era arrangements between regional networks over the distribution of power and

The author would like to thank Laura Adams, Allan Craigie, Nick Megoran, Paolo Sartori and Tommaso Trevisani, and the anonymous reviewers of the journal for their helpful comments on earlier drafts of this article. The author is also grateful to the British Academy Committee for Central and Inner Asia, the British Association for Slavonic and East European Studies, and the British International Studies Association for their generous support of this research. 1‘Riot’ here is meant as a kind of conflict involving ‘no structured armed forces and organised long term fighting’ (Tishkov 1995, p. 148, fn. 1). 2I follow the international spelling of all names of locations indicated in the article.

ISSN 0966-8136 print; ISSN 1465-3427 online/07/040567-24 ª 2007 University of Glasgow DOI: 10.1080/09668130701289869 568 MATTEO FUMAGALLI resources in Tajikistan collapsed, precipitating the country in a five-year civil war, which cost the lives of at least 50,000 people and led to the displacement of many more. Although the Tajik conflict was not ethnic in nature,3 ethnic groups re-aligned themselves during the conflict, with Uzbeks mostly supporting the ruling faction challenged by the democratic-Islamic opposition.4 This article asks how these two conflicts, or better the memory of them, have been framed by the leaders of the Uzbek community in Kyrgyzstan and Tajikistan and what impact they have had on the course of Uzbek mobilisation.5 By examining the frames adopted by Uzbeks in both countries, the article aims to show how Uzbeks have adjusted to living in post-Soviet Kyrgyzstan and Tajikistan.6 It will argue that an understanding of what issues have been framed and how widely they have resonated across the Uzbek community sheds light on the broader mobilisational process. The main contribution this article seeks to make is to understand the political behaviour of Uzbek ethnic minorities in Central Asia, a topic traditionally indicated as carrying significant implications for both state and regional stability, but rarely subject to empirical investigation (Bohr 1998; Horsman 1999; Lubin & Rubin 1999; Melvin 2000; Olcott 1994, 1995; Tabyshalieva 1999). The article also aims to contribute to the discussion on ethnic mobilisation through a focus on the role of frames in the mobilisational process. Building on the expanding literature on cultural framing, this article identifies what frames were available to the Uzbek communities in the two countries since independence, and which resonated more widely (Gorenburg 2003; Johnston & Noakes 2005; Renwick 2006; Schatz 1999; Snow et al. 1986; Snow 2004; Ukudeeva-Freeman 2003; Zald 1996). Following Gorenburg, frames are defined and understood as ‘interpretive schemes that condense and simplify a person’s experience by selectively highlighting and encoding certain situations, objects, events and experiences’ (2003, p. 11). In order to resonate, frames need to be familiar and hence build on the material as well as cognitive context. Frames do not emerge naturally, but are an essential tool used by political elites to mobilise or demobilise the broader community. Finally this article calls for more attention to be paid to the relational aspect of frames. Rather than being seen as fixed sets of ideas, perceptions and beliefs, the article shows that the same frame can be used to mobilise and demobilise the community.7 A brief caveat is

3For a discussion of the causes of the Tajik conflict see Abdullaev and Barnes (2001). 4This is of course a simplification. The conflict ran through the country’s multiple cleavages. 5Unlike most definitions of mobilisation that focus on claims and protests, my understanding of mobilisation follows Nedelman’s definition which refers to ‘the actors’ attempt to influence the existing distribution of power’ (1987, p. 181). This allows me not to assume a predetermined outcome (the movement from passivity to activity), or focus only on the ‘noisy’ part of mobilisation (claims and protests), but to gain an understanding of a ‘quiet’ phase too (e.g. electoral support for the authorities). 6I am not implying that the Uzbek communities have recently settled there. In fact, Uzbeks have lived in that part of the Ferghana Valley even during Soviet times, but the absence of ‘real’ borders made their being formally resident in Soviet Tajikistan and Kyrgyzstan immaterial, given their physical proximity to Uzbekistan. The establishment of new interstate borders changed all this. 7For the purpose of this article I rely on Charles Tilly’s definition of demobilisation as ‘the process by which a group of active participants in public life becomes a passive collection of individuals’ (Tilly 1978, p. 54). Unlike mobilisation, demobilisation remains an unexplored process both empirically and theoretically. In most cases it refers to the process following a violent conflict through which former ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 569 necessary here. The article adopts the year 2003 as the cut-off point. This means that events which occurred after that year are not considered. The balance of power in Tajikistan has not been altered since then. The situation appears much more in flux and convoluted in Kyrgyzstan and any attempt to present an analysis of the impact of the fall-out of the 2005 events that led to the ousting of the regime of former President Akaev would risk being superseded by the current unfolding of events.8 I will briefly address their fallout on inter-ethnic relations in the later part of the article. This article is divided in the following manner. First I discuss the significance acquired by framing theory in recent explanations of ethno-political mobilisation. Next I provide a concise historical background of the case studies and discuss methodological issues. I then present the empirical findings. Concluding remarks then follow.

The role of frames in explanations of ethnic mobilisation Structural explanations have traditionally dominated the debate in mobilisation studies.9 The concept of political opportunity structure (POS), first introduced by Peter Eisinger (1973), has traditionally been central to structural explanations.10 POS refers to the ‘dimensions of the political environment that provide incentives for people to undertake collective action by affecting their expectations for success or failure’ (Tarrow 1998, p. 41). Building on Tarrow’s work (1998), POS theorists tend to assign a crucial role to the state and state policies which would expand or constrain political opportunities.11 While structural factors, most notably the salience of the Soviet experience and the elite structure, certainly play an important role in explanations of Uzbek political mobilisation, changes within or across mobilisational processes are ascribed to factors that are external to the group itself, de facto depriving it of autonomous agency. combatants relinquish the use of violence and are integrated in the political process (see Walter 1999 for example). To capture the dynamics at play in Central Asia I find it more appropriate to follow the way the concept is used by Gagnon (2004), Remmer (1980) and Hipsher (1996) with regard to the strategy adopted by political elites in authoritarian systems to demobilise the population to prevent possible challenges and consolidate their position. 8Askar Akaev’s appearance on the Kyrgyz political scene is due to the fact that in October 1990 the two candidates running for the country’s presidency (Apas Jumagulov and Absamat Masaliev) did not obtain sufficient votes in the first round of voting in the Supreme Soviet and were thus prevented from re-running. He was then selected as a compromise candidate between the competing factions. Originally from the northern region of Chuy, Akaev spent some considerable time abroad, in St. Petersburg (then Leningrad), where he trained as a physicist and began his academic career. He returned to his native country in 1977. On the origins and implications of the so-called ‘Tulip Revolution’ see E. Marat, The Tulip Revolution: Kyrgyzstan One Year After (Washington, Jamestown Foundation), 2006, available at: http://www.jamestown.org/images/pdf/Jamestown-TulipRevolu- tion.pdf, accessed 15 April 2006. 9For examples of approaches emphasising the salience of context as enabling and/or constraining opportunities for mobilisation, see Gurr (1993), Kitschelt (1986), Koopmans (2000) and McAdam (1982). For a critical review of mobilisation studies see Kriesi (2004) and McAdam et al. (2001). 10See also Tarrow (1991, p. 14). 11In his recent work Tarrow has expanded the concept of POS to incorporate the influence of transnational processes within the study of the mechanisms of contention (2005). 570 MATTEO FUMAGALLI

Unlike structural explanations, framing theory privileges the ‘strategic framing of injustice and grievances, their causes, motivations, and associated templates for collective action’ (Zald 1996, p. 261) over the political opportunity structure. The study of cultural framing represents a welcome addition in the study of ethnic mobilisation for a number of reasons. First, the study of the frames adopted by the group leaders to mobilise the Uzbek community appears particularly useful to explain how structure and agency interplay as well as to account for variations within and across cases. Frames, in short, operate as the interface between background structural factors and the contingent choices of the elites. Second, the study of framing processes emphasises the nexus between the frames and the agents, those that actually do the framing. Ukudeeva-Freeman (2003) and Gorenburg (2003) have shown that frames do not come out of nowhere. Because they need to be popularised among the target population, they need be familiar to resonate: ‘To be successful, nationalist leaders had to frame their demands in language and imagery that could resonate with the population’ (Gorenburg 2003, p. 12). When looking at post-Soviet mobilisation this inevitably means taking into full account the legacy of 70 and more years of the Soviet experience, which throughout the twentieth century ‘decisively moulded the perceptions, beliefs, and identities of minority ethnic group members’ (Gorenburg 2003, p. 12). This points to both the importance of agency (of the elites) and the connection between elites and the rank and file. As scholars of ethnic and national mobilisation have emphasised, for a movement to be successful, ‘it is not enough for it to acquire organisational resources and a cadre of dedicated activists’ (Gorenburg 2003, p. 11). Third, we should not assume frames to be static or even fixed. Frames vary, compete with each other and change. While a focus on resources and structural constraints may be sufficient to explain the emergence or absence of a national movement (Ukudeeva- Freeman 2003, p. 99), it does not provide convincing evidence to make sense of variations in levels of mobilisation across and within cases. By contrast, the study of frames and framing appears particularly useful for comparative analysis. To this end, what needs to be taken into account is: (1) the extent to which the movement (or its leaders) is able to create an idea or set of ideas among its target population that it has been an object of past injustices, or that taking part in the movement is worthwhile because of future positive payouts; and (2) the extent to which the interpretation of ideas, perceptions, and beliefs (frames) resonate across the group itself and are appropriated. In an insightful study of political contestation and mobilisation in late Soviet Kazakhstan, Edward Schatz has shown how focusing on the way grievances are framed accounts for the variation between nationalist and non-nationalist mobilisa- tion in the late Soviet era (1999). The case of eco-internationalism in Kazakhstan (that is, where ecological mobilisation did not take a nationalist tone, but emphasised ‘internationalist tropes’) is particularly interesting as it allows us to make sense of an expected case of nationalist mobilisation that did not happen. Emphasis on both the changing nature of frames as well as the role of political agency emerges from Jamilya Ukudeeva-Freeman’s study of national-democratic movements in late Soviet Azerbaijan and Kyrgyzstan. Ukudeeva-Freeman has convincingly applied the concept of mobilising ideas as ‘a useful analytical tool to explore the extent to which an idea or a set of ideas are shared by members of a group, ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 571 who are then ready to mobilise around that very same idea’ (2003). Given that the presence of grievances or the opening of an opportunity in the structural context does not in itself generate political action, the leader has to ‘present a concept, vision, or an idea, that resonates among the people and gets people to be sympathetic with the movement goals’ (2003, p. 99). Hence, she defines a mobilising idea as ‘an action- oriented set of beliefs that unifies people around itself for a common goal. [It] inspires people for action, legitimates leaders’ actions, and expresses conviction about how things should be’ (2003, p. 99). A mobilising idea is both an action plan and also a solution to problems, which makes it similar to Gorenburg’s understanding of cultural frames. The movement would be over, Ukudeeva-Freeman contends, when leaders cannot recruit people anymore because their ideas and goals do not resonate among people. In this article I also propose to take the study of ethnic mobilisation further by introducing the concept of a demobilising idea. By demobilising idea I refer to a type of frame strategically developed by elites which leads the target community from mobilisation to demobilisation.12 The concept of a demobilising idea also shows how frames are quintessentially relational, thereby overcoming the risk that a master frame might be viewed as something static and pre-determining the trajectory of mobilisation. The same frame can simultaneously act as a powerful mobiliser in one direction, and demobilise the very same group in another.

Uzbek communities in Kyrgyzstan and Tajikistan Uzbeks constituted the largest Muslim and the largest non-Slavic community in the Soviet Union and even after independence they represent the largest ethnic group in post-Soviet Central Asia. On the basis of the first census conducted after the Soviet collapse,13 Uzbeks composed approximately 78% of the population in Uzbekistan, 15.3% in Tajikistan, 13.8% in Kyrgyzstan, 2.4% in Kazakhstan and 9.2% Turkmenistan.14 While Uzbeks occupied a privileged position in Soviet Uzbekistan, they played a secondary role in the life of the neighbouring republics. Because they already enjoyed titular status in the Uzbek Soviet Socialist Republic (SSR), Kyrgyzstan’s Uzbeks, or in fact all Uzbeks outside the Uzbek SSR, were not granted territorial autonomy. This has severely hampered the pool of resources from which Uzbeks could draw to

12I am not assuming that mobilisation, demobilisation, or even inertia may be a natural condition of a community. The community’s condition and the shift of condition are products of social (elite) construction. 13Uzbekistan has not conducted a census. One reason for this could arguably lie in the purpose the post-independence censuses served in the other republics, namely officially signalling the demographic dominance of the titular group; in countries like Kazakhstan, and to a minor extent Kyrgyzstan, these constituted either a plurality or a bare majority. Uzbekistan’s situation was remarkably different because Uzbeks have never faced the possibility of being outnumbered by other ethnic groups. Data are taken from Alisher Ilkhamov’s Etnicheskii Atlas Uzbekistana (Ethnic Atlas of Uzbekistan), published in 2003. It is based on empirical work conducted by him and other Uzbekistan-based scholars and I consider it to be a reliable source of information, accurate, though unofficial. 14Perepis Naseleniya Turkmenistana 1995 (1997), Kratkie Itogi Perepisi naseleniya 1999 (1999), Naselenie Kirgizskoi Respubliki 1999 (1999), and Naselenie Respubliki Tadzhikistan 2000 (2002). 572 MATTEO FUMAGALLI mobilise.15 As Marianne Kamp argues, Uzbeks ‘could expect access to positions and opportunities within Uzbekistan but did not find doors opened wide within the larger boundaries of the Soviet Union’ (2002, p. 273). Because different rules and practices applied outside Uzbekistan, this suggests that in the end some form of boundary existed between Union republics, though this had little impact on the population’s everyday life. A significant change after independence was that whereas during the Soviet era ‘pilgrimages to Tashkent’ represented a relatively easy opportunity for Uzbek co-ethnics to escape from republics where other groups were privileged, the establishment of five independent states in Central Asia progressively cut off Uzbek co-ethnics from this opportunity (Kamp 2002, p. 273) (see Figure 1).

Uzbeks in Kyrgyzstan Following the principles of Soviet ethno-federalism, non-titular groups (co-ethnics living outside their alleged homeland) were not endowed with any form of special protection. This meant that, for example, while Uzbeks living outside the Uzbek SSR would enjoy cultural rights as individuals, they would not be granted the series of privileges that titular groups typically enjoyed (e.g. territorial autonomy, language protection). Commonalities of cultural traits and convergence of political interests (limiting the influence of northern power groups) did not prevent the outbreak of riots between Uzbeks and southern Kyrgyz in the Kyrgyzstani towns of Osh and Uzgen in June 1990. Towards the end of the 1980s a conjuncture of socio-economic crisis, decreasing living standards, and political destabilisation, led to the eruption of inter- group tensions over the competition for resources (land lots), lack of housing and control over power structures (Tishkov 1995, p. 135).16 Mass riots broke out on 4 June and were ‘extinguished’ on 10 June 1990 only after armed forces intervened and a state of emergency was declared in the republic (Tishkov 1995, p. 135). One should also note the absence of organised movements implicated in the preparation or unfolding of events, although loose and informal organisations, namely ‘Osh Aimagy’ (Osh Land/Region) among the Kyrgyz, and ‘Adolat’ (Justice) among the Uzbeks, were indeed agitating and inflaming spirits. While it is safe to argue that the conflict had essentially socio-economic causes, it manifested itself along ethnic lines, and demands started to take on ethnic tones, including requests for recognising Uzbek as an official language and for incorporating parts of Kyrgyzstani territory to Uzbekistan (Spector 2004, p. 6). The Osh conflict brought to the attention of political elites the centrality of inter-ethnic stability for the viability of the country. It also provided a window of opportunity for Askar Akaev, then a political outsider, to emerge as president of the Kyrgyz SSR in October 1990. Akaev, who was ousted during the so-called ‘Tulip Revolution’ in spring 2005, managed to build a broad coalition, uneasily inclusive of both more nationalist

15There is an established consensus in the literature on ethnic mobilisation on the importance that ethnic institutions played as resources for minority mobilisation in the late Soviet and post-Soviet era (Beissinger 2002; Gorenburg 2003; Hirsch 2005; Roeder 1991; Slezkine 1994). 16The possible triggering factors indicated as causes of the riot are several, including the KGB’s covert role and links between political factions and mafia-like economic groups. TNCMNRT OIIAINI ETA SA573 ASIA CENTRAL IN MOBILISATION MINORITY ETHNIC

Source: Reliefweb, Ferghana Valley: Reference Map, 2003, available at: http://www.reliefweb.int/rw/fullMaps_Sa.nsf/luFullMap/69CEE2A40B9DE34B85256DCE00622796/ $File/rw_ferghana291003.pdf?OpenElement, accessed 15 April 2006.

FIGURE 1. THE FERGHANA VALLEY REGION 574 MATTEO FUMAGALLI TABLE 1 ETHNIC COMPOSITION OF KYRGYZSTAN

1926 1926 1959 1959 1989 1989 1999 1999 Group (no.) (%) (no.) (%) (no.) (%) (no.) (%)

Kyrgyz 668,700 66.8 836,800 40.5 2,229,663 52.4 3,128,144 64.9 Russians 116,800 11.8 623,600 30.2 916,558 21.5 603,198 12.5 Uzbeks 106,300 10.6 218,900 10.6 550,096 12.9 664,953 13.8 Ukrainians 64,200 6.4 137,000 6.6 108,027 2.5 50,441 1.0 Germans 4,300 0.4 39,900 1.9 101,309 2.4 21,472 0.4 Tatars 4,900 0.5 56,300 2.7 70,068 1.6 45,439 0.9 Uighurs 8,200 0.8 13,800 0.7 37,318 0.9 46,733 1.0 Kazakhs 1,700 0.2 20,100 1.0 36,928 0.9 42,657 0.9 Dungans 6,000 0.6 11,100 0.5 36,779 0.9 51,766 1.1 Tajiks 7,000 0.7 15,200 0.7 33,518 0.8 42,636 0.9 Other 13,600 1.4 93,400 4.5 137,491 3.2 125,499 2.6 TOTAL 1,001,700 100.0 2,066,100 100.0 4,257,755 100.0 4,822,938 100.0

Source: Narodnoe khozyaistvo Kirgizskoi SSR (1982); Naselenie Kirgizskoi Respubliki 1999 (2000).

TABLE 2 ETHNIC COMPOSITION OF TAJIKISTAN

Group 1979 (%) 1979 (no.) 1989 (%) 1989 (no.) 2000 (%) 2000 (no.)

Tajiks 58.8 2,237,048 62.3 3,172,420 79.9 4,898,382 Uzbeks 23.0 873,199 23.5 1,197,841 15.3 936,703 Russians 10.4 395,089 7.6 388,481 1.1 68,171 Kyrgyz 1.3 48,376 1.3 63,832 1.1 65,515 Lakay ––––0.851,001 Congrats ––––0.315,102 Catagans ––––0.14,888 Yuz––––0.01,053 Barlos ––––0.13,743 Semiz ––––0.01 Kesamir ––––0.013 Others 6.5 252,508 5.3 270,029 0.1 148,436 TOTAL 100.0 3,806,220 100.0 5,092,603 100.0 6,127,493

Source: Naselenie Respubliki Tadzhikistan 2000 (2002). factions as well as ethnic minorities. In return for his perceived role as guarantor of ethnic peace Akaev could count on the loyalty of the country’s many minorities (Table 1), Uzbeks among them, which manifested itself in elections, parliamentary and presidential (Huskey 1997).

Uzbeks in Tajikistan Uzbeks constitute the largest non-Tajik group in Tajikistan (ca. 15 – 25%,17 see Table 2). To understand the place of Uzbeks in today’s Tajikistan it is

17This depends on which census is considered: censuses conducted before the latest showed a large Uzbek population (greater than 20%), whereas in the most recent (2000) the Uzbek share of the overall population had shrunk to about 15%. ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 575 necessary to recall the formative period of the Tajik republic during the national delimitation process (1924 – 36). This played a crucial role in shaping contemporary Tajikistan. Not only did it define the boundaries which have remained unchanged since 1929— when the country’s status changed from autonomous republic (ASSR) within the Uzbek SSR to full-fledged Union Republic (SSR)—it also constituted the only experience Tajikistan has had as a national-territorial state. The decision to include strongly ‘Uzbekified’ areas such as the western Ferghana Valley, but to leave within Uzbekistan cities like Bukhara and Samarkand—where the Persian language and culture were dominant—were highly controversial and ultimately made Tajikistan an ‘incomplete country’, strongly dependent on its stronger western neighbour and Russia. Furthermore the unofficial balance of power assigned the northern province of Leninabad, home to a significant Uzbek population, a hegemonic role in the republic’s life.18 Even towards the end of the Soviet era when nationalist sentiments began to emerge and occasional clashes were sparked off,19 the position of minority groups rarely seemed threatened. In fact, Muriel Atkin observes that ‘the late Soviet regime in Tajikistan made several conciliatory gestures towards the republic’s Uzbek minority’ (1997, p. 299). Bookstores offering Uzbek language publications were opened in the south of the country, a new Uzbek weekly was launched, alongside two other newspapers already publishing in the Uzbek language, and programmes in Uzbek were broadcast on the radio. Schools provided Uzbek-language tuition in areas of compact Uzbek settlement (1997, p. 299). This relatively positive environment, alongside Leninabad’s traditionally close links with Tashkent, helps to explain Uzbek behaviour during the civil war of 1992 – 97. The roots of the civil war are complex and multi-faceted and a mono-causal explanation fails to capture the different dynamics operating at different levels. Although the war never took on ethnic tones, it was also about defining what being Tajik meant. Overall, a critical reason for the outbreak of hostilities lay in the failure of government and opposition to moderate and accommodate their respective demands. Gretsky (1995), in particular, is adamant that Soviet Tajikistan remained deprived of any autonomous agency for a large part of its history. He argues that Soviet Tajikistan acted as Tashkent’s appendix through the action of the Leninabadi faction. Khujand/Leninabad remained outside Dushanbe’s writ for several months in 1992, though it may also be suggested that there was a lack of political will to seek outright secession (Human Rights

18The Leninabad elite’s hegemonic role in Soviet Tajikistan can be ascribed to economic, socio- cultural and geographic reasons, including the fact that the region was the only industrialised area in an otherwise rural or inhospitable land. Its geographical location, integrated in the Ferghana Valley region and with Uzbekistan, made Tajikistan’s economic system viable, though in the end the country depended on Moscow’s subsidies. The north was also culturally more exposed to contacts with Russia and the rest of the Soviet Union, compared to the more provincial areas in the south. It should be pointed out nonetheless that ‘regions’ should not be reified as if they were discrete units and that internal differences and competition existed (e.g. between Khujand, Leninabad’s administrative province and the areas around Istaravshan and Penjikent). 19This was the case in Isfara (southern Kyrgyzstan) where Kyrgyz and Tajiks clashed over water allocation, and in Dushanbe (Tajikistan), where protests against alleged favourable housing allocation to Armenian refugees took place. 576 MATTEO FUMAGALLI

Watch 1998; Martin 1997).20 Overall, as Akiner notes, Uzbeks sided with the ‘pro- government’ forces and in the southern areas have joined the Popular Front, a pro- government militia. The motivation lay in the attempt by the ruling elite to protect its existing privileges from the perceived threat coming from the Islamic and democratic opposition. Tajikistan Uzbeks have thereafter predominantly supported the incumbent in Tajikistan, Rakhmonov, since 1994 for his role in stabilising the country.21

Methodological issues This article adopts a comparative strategy. It controls for context (a common Soviet legacy and the political opportunity structure) and explores variations in terms of frames adopted by the groups in the two cases. In so doing it hopes to partly address one of the ‘pathological’ conditions of comparative research, that of ‘too many variables, not enough cases’ (Lijphart 1971; Ragin 1987). Of course, a two-case comparative study does not have ambitions to generate statistically relevant conclusions, but can offer insights deriving from an in-depth focused comparison. Strong similarities exist between the two cases under investigation. First and foremost both share a legacy of 70 and more years of Soviet history.22 For both Tajikistan and Kyrgyzstan, and all the Central Asian republics for that matter, modern state formation originates in the Soviet period, with the establishment of the Tajik Soviet Socialist Republic in 1929 and the Kyrgyz SSR in 1936 within their current boundaries, smaller boundary adjustments aside. Second, Kyrgyz and Tajik societies are particularly fragmented and sub-national allegiances have long been stronger than national affiliations. Clan and/or regional politics have traditionally dominated Tajik and Kyrgyz politics, and the Uzbek minorities in both countries have often played a significant role in the intra-ethnic power struggles. Finally, even the political development of the two countries since the late Soviet period is broadly comparable. Before the Soviet collapse, the Tajikistani and Kyrgyzstani societies were among the more pluralistic in Central Asia. Some degree of political competition was allowed, informal organisations proliferated and candidates could compete in

20This view was also confirmed by the Khujand-based researcher Fatimakhon Ahmedova (Centre of Democratic Transformation, August 2003). 21Emomali Rakhmonov, originally head of a state farm (sovkhoz) in the province in Kulyab, was little known until 1992 when he was first named Chairman of the Executive Committee of the Kulob province, and later named Chairman of the country’s Supreme Soviet in 1992. He was elected president in 1994 (the presidency had been abolished between 1992 and 1994) (Nourzhanov 2005, p. 129, fn. 57). Though he owed his sudden rise in national politics to his patron, Sanjak Safarov, he later tried to consolidate his grip on power by curtailing the influence of local warlords on Tajikistan’s politics. To be sure, there were Uzbeks among the ranks of the opposition, particularly within the ranks of the exiled Islamic opposition that had found refuge and some degree of support when fleeing Uzbekistan in the early 1990s. Elements of the IMU (Islamic Movement of Uzbekistan) closely co-operated with factions of the Tajik opposition, though this tactical alliance (anti-government and anti-Uzbekistani) between them waned in the aftermath of the 1997 Peace Accords, leading to power sharing between government and opposition. For an analysis of Uzbekistan’s role in the Tajik civil war see Horsman (1999). 22For various studies on this complex process of societal re-organisation and its consequences see Hirsch (2005), Martin (2002), and Suny and Martin (2001). ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 577 elections. This trend continued in Kyrgyzstan for the first part of the following decade, to the point that in an excess of hyperbole the country came to be referred to as ‘Central Asia’s island of democracy’. After initial reforms however, state authorities in both countries resorted to authoritarian practices to consolidate their positions in power, relying less on democratic procedures of accountability and responsibility or participatory governance. On the other hand, one should not overlook the differences between the two, most notably the fact that Tajikistan descended into a bloody civil war from 1992 until 1997. Kyrgyzstan also experienced brief, communal infighting during the 1990 Osh conflict, but state and societal breakdown was prevented despite the fact that the odds were stacked against it (regional, ethnic sub-regional and clan cleavages, social and economic deprivation). This article relies on a combination of individual in-depth interviews conducted with 60 and 30 members of the local Uzbek community in Kyrgyzstan and Tajikistan, respectively, and a small-scale survey (136 respondents in Kyrgyzstan and 137 in Tajikistan), conducted between the summer of 2003 and 2005.23

Framing Uzbek ethnic mobilisation The legacy of conflict: memory as a mobilising idea In the immediate aftermath of the 1990 Osh conflict the informal organisations that were more actively involved in the riots (Osh Aimagy and Adolat) were disbanded and ethnically based political movements banned from public life. The main Uzbek political actors since independence have been cultural organisations. The two Uzbek cultural organisations (the Uzbek National-Cultural Centre, O’zbek Milliy Madaniyat Markazi, UNCC, and the Society of Uzbeks, O’zbeklar Jamiyati) depict themselves as cultural and explicitly non-political. However, the line between culture and politics in this case appears very much blurred, especially in non-democratic contexts where dealing with cultural issues is often a veil to dealing with sensitive political questions, and open discussion of them would be otherwise not permitted. A new party was established in the southern town of Jalalabad in 1999: the Party of National Unity and Concord (Partiya narodnogo edinstva i soglasiya). Although the leaders of this party (including Azamjan Azimov, the party’s founder, and Adikjan Abildov, the Osh party

23Interviews and surveys were carried out in the areas of greater Uzbek settlement. The sample included 71 respondents from the Osh province, 56 from Jalabat and Batken, and nine from Chuy and the city of Bishkek. Interviews were held in Osh (44), Jalalabad (seven), Bishkek (eight) and Tokmok (one). In Tajikistan the survey sample included 39 respondents from Khujand, 63 from the surrounding Sughd province and 35 from Dushanbe. Twenty-three interviews were held in Khujand and seven in Dushanbe. Sampling was purposive. This was done to overcome the political sensitivity of the topic. Potential respondents were identified by asking any respondent, at the end of an interview or completion of the questionnaire, to indicate additional persons who may have been interested in participating in this research. I was aware of the risk of obtaining a skewed sample, as respondents might have been inclined to refer to others with similar views. However, given the variety of the responses it seems safe to argue that this was rarely the case. 578 MATTEO FUMAGALLI representative and former head of the UNCC) vigorously deny that it is an Uzbek party, the leaders themselves admit that 95% of its membership is actually Uzbek.24 The fear that the 1990 conflict might repeat itself appears to be a constant theme and a real concern for Kyrgyzstani Uzbeks. The memory of the Osh conflict has in fact acted as a mobilising idea. What makes this frame peculiar, however, is the trajectory towards which mobilisation develops. The literature on ethnic mobilisation suggests that grievances and memories of past injustices would be used as resources from which nationalist leaders draw to build support for their organisations. In Kyrgyzstan the concern that such a conflict may repeat itself led Uzbeks to lend their allegiance not to nationalist leaders, but to actors perceived to be capable of preserving inter-ethnic stability. Former President Akaev made ethnic harmony one of the cornerstones of his tenure in office and has therefore greatly benefited from the political support of minority groups. It is important to note how the frame emphasising the necessity to support whoever could preserve social and inter-ethnic stability, and more generally speaking, the way the Osh conflict was remembered, were neither in any way ‘natural’ and nor was their emergence pre-determined. They were the product of the Uzbek elites’ strategic choices. The leaders of the Uzbek community acted both as patrons of the community and also as intermediaries between their clients and their own patrons, the state elites (in particular Akaev). Mutual support led not only to the creation of a system of patronage, but also to mutual dependence. What is important here is to note that this frame (memory as mobiliser) resonated widely across the Uzbek community. The resonance was evident in the light of the general consensus that existed within the Uzbek community with regard to the then president’s dealing with ethnic issues: ‘We support Akaev; we have nothing to do with the opposition’;25 ‘Akaev may be weak but is better than others [Kyrgyz nationalists]’; ‘What happens if Akaev goes? Who knows who will come after him?’.26 Overall, one could detect two sentiments underlying Uzbek support: the perception of a lack of alternatives and a serious concern for the future successor in his post. Akaev, in this sense, was not supported for the success or efficacy of his policies: an interview with Hokimjon, an Osh-based journalist, well captures the Uzbek mood by pointing out how Akaev ‘does [did] not have real power to implement his decisions’,27 but was preferred because of the lack of an alternative.28 The question that Uzbeks kept asking themselves (‘Who will come after Akaev?’) revealed strong concern that a more nationalist-oriented political figure might emerge and replace the current president. For this reason the frustrations and grievances among the Uzbek population were often left aside when it came to participating in parliamentary and presidential elections. Indeed, concern about the future shaped a self-reinforcing myth among Uzbeks that Akaev’s successor would be more nationalist. The case of Tajikistani Uzbeks follows a similar pattern. In the early part of the 1990s, Uzbeks aligned themselves with the ruling faction. Uzbek mobilisation went

24Interview with Mr Abildov (Osh, July 2003). The party had very little impact in the subsequent electoral campaigns and remained marginal in Kyrgyzstani politics. 25Interview held with local activists at the Osh UNCC (Osh, 2 July 2003). 26Interviews with members of the local Uzbek community (Jalalabad, 10 July 2003). 27Not his real name (Osh, 25 June 2003). 28On this see also Khamidov (2000). ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 579 beyond mere political support, and some Uzbeks joined the Popular Front, a pro- government militia, during the civil war. Uzbeks maintained their support throughout the post-independence period. According to Muhtar,29 a local Uzbek working for an international organisation, this was because Rakhmonov was perceived by the local population as the actor that brought an end to the civil war: ‘We support the president because he brought an end to the war. Of course there are problems, but he keeps [Tajik] nationalists under control’.30 This does not neglect the fact that with Rakhmonov came the ‘Kulyabisation’31 of the country and the monopolisation of all power positions by a close group of the president’s allies. This meant a marginalisation of minority groups, particularly of those associated with the northern region, left out from the 1997 peace agreement. Similarly to the Kyrgyzstani case, Tajikistani Uzbeks also remained fearful of any alternative to the incumbent regime. Again, crucial to shaping the pro-government position of the Uzbek community has been the local cultural organisation (‘Society of Uzbeks’). The amplification of this frame was made possible by the close connections between the members of the Society of Uzbeks and the local (city and provincial) administration in the Sughd province. Meetings of the organisation were often held in the offices of the local administration’s weekly newspaper, Leninobod Khaqiqati (The Truth of Leninabad).32 This ultimately enabled the frame presented by the local Uzbek elites to resonate across the Uzbek population. The publication may well have had a limited reach, but the fact that both the media and the official representative of the community aired the same message contributed to reinforce the frame’s strength compared to alternative options.

Memory as a demobilising idea As noted in the previous section, the memory of the Osh conflict and the Tajik civil war were used by the leading figures of the Uzbek community to rally support for state authorities and as such to legitimise their position (as intermediaries). This was not the only way in which the memory of previous conflict was perpetuated. The leaders of the local Uzbek community warned against the risks of ethnic outbidding whereby local figures within the community would seek to capitalise on the grievances of the larger population for their own personal advantage. This, it was argued, would politicise ethnicity and destabilise inter-ethnic relations and, therefore, was to be prevented. One way to do so was by demobilising the Uzbek community by convincing ordinary

29Not his real name. Where indicated, the interviewees’ names have been changed to preserve their anonymity. On particular occasions, the informants insisted that their real name be published. 30Interview conducted in Khujand on 11 August 2003. This is a broadly shared view, also acknowledged in the scholarship (Akiner 2001). 31Kulyabisation refers to the fact that the near totality of the members of the power structure in Tajikistan come from Kulyab, a region in south-western Tajikistan. In reality, however, this grossly neglects the fact that except for the areas around the villages of Farkhad and Dunghara, even the rest of Kulyab is under-represented. 32Often the same person serves as a representative of the organisation and as a journalist or even editor for the publication. 580 MATTEO FUMAGALLI

Uzbeks of the consequences of creating an ‘Uzbek question’ in the country. As Alisher, an activist at the UNCC in Jalalabad observes:33

Why should we get involved in politics? Uzbek problems should be solved within the community. People are afraid that by becoming active in politics, political struggles become a matter of ‘us against them’, of Uzbeks against Kyrgyz. As soon as we get involved, they will all unite and divert their problems against us. This is why during the Aksy riots,34 we told our people: Do not go there, stay home. Don’t go, they will blame us.35

Uzbek behaviour can be ascribed to a conscious strategy of Uzbek political entrepreneurs to demobilise the community after the Osh conflict and the Tajik civil war. ‘Aralashma!’ (do not get involved!), urged leading representatives of the Uzbek community. As a result, even those who had headed towards Aksy literally turned their cars around and returned to their home towns and villages.36 Uzbek leaders framed the Aksy events as an entirely ‘Kyrgyz affair’. Besides grievances regarding the difficulties for Uzbeks concerning employment in the army, judiciary, and secret services (both in terms of recruitment and career advancement), many also lamented the way the Uzbek community was being framed by some elements of the Kyrgyz media and political elite: ‘They [Kyrgyz] think we are all members of Hizb-ut Tahrir! They do not trust us; they think we cannot be loyal. But we are citizens of Kyrgyzstan, this is our homeland!’.37 Being framed as potentially disloyal, Uzbeks have responded by retreating to other sectors such as retail trade, the police (militsiya), and mass media. As noted earlier, this ‘retreat’ has been encouraged by the emphasis of the Uzbek elites on the incapacity of the state to address their grievances and demands.38

Soviet frames: indigenous nation or diaspora? The Soviet experience has continued to shape the way Uzbeks view and frame their position within the post-Soviet state. A study of the way Uzbeks relate themselves to conceptions of diasporic identities sheds light on the significance of Soviet frames and their implications for post-Soviet state-building and nation-building. Uzbek presence in both countries was articulated not as a temporary or recent presence, but as a permanent one. A plurality of respondents in Kyrgyzstan (45.5%)39 rejected the idea

33Interview held in Jalalabad on 12 July 2003. 34In March 2002 demonstrations were organised in the southern town of Aksy by the local population to protest against the arrest of the local deputy Azimbek Beknazarov, and they were viewed as politically motivated. State authorities reacted by firing against the demonstrators killing at least five of them and arresting hundreds. For more on the ‘Asky events’, as they have become known, see Radnitz (2005). 35Interviews held with member of the Uzbek population in Kyrgyzstan (June and July 2003; July and August 2005). 36This account was confirmed by a large number of interviews with local Uzbeks in both Osh and Jalalabad (June and July 2003). 37Interview with Tohirjon, an Uzbek postgraduate student at Osh State University and local journalist (Osh, 8 July 2003). 38For a more detailed discussion of Uzbek grievances and state response see Fumagalli (2005). 39These and the following data derive from the survey the author conducted in Kyrgyzstan and Tajikistan in 2003, unless otherwise acknowledged. ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 581 that the Uzbek community might be categorised as a national minority, whereas a slightly lower percentage seemed instead to share this thesis (40.2%), with one in six respondents undecided (14.4%). The label national minority was rejected by representatives of the Uzbek community, arguing that this label applies to ‘very small groups’. ‘National minority’ implicitly refers to a diminished status which Uzbeks find difficult to reconcile themselves to, especially in areas of compact Uzbek settlement in the south. Referring to local Uzbeks as a ‘diaspora/diasporic community’ was considered more or less a term of abuse. Respondents overwhelmingly rejected any association between the Uzbek community and the label diaspora (only 3.3% of the sample consider Uzbeks to be a diasporic community). One in four respondents (23.6%) assigned the label a negative connotation, more or less the same proportion of the sample was indifferent to the question (25.5%), and one in three (36.4%) did not have any opinion in this regard. I then explored what meaning respondents attached to the phrase ‘Uzbek diaspora’ in order to understand the rationale behind the massive rejection of the term. This overwhelmingly negative attitude toward the term is not well captured from survey data, but the follow-up interviews seem to give a better grasp of the extent to which Uzbeks did not wish to be identified as such. On no occasion, in individual interviews, could I find an Uzbek willing to consider his/her community of belonging as diasporic. Reactions tended to include comments such as the following: ‘We are not a diaspora!’;40 ‘Nobody should refer to Uzbeks as diaspora’;41 ‘We did not arrive from anywhere. We are not like Germans or Koreans. They are small peoples, we are more than a million’.42 Across the post-Soviet space the terminology of diaspora remains not only politically charged, but firmly rooted in Soviet understandings of nationalities. This appears clearly when exploring the meanings and perceptions associated with the term. Typically, respondents had difficulties in explaining what one should understand by the concept. However, they were able to provide plenty of examples of diasporic communities, such as the Chechens, Koreans, Poles, and Germans, that in the Stalinist period were considered ‘enemy peoples’ (read: anti-Soviet and disloyal), and on these grounds deported. What made things ‘worse’ from the Uzbek perspective, was that these peoples were small in size and not indigenous, and therefore lacked a key pre- requisite for aspiring to status and territory and the resources and prestige associated with the claim to membership of national territories. Similar considerations were drawn with regard to the term ‘minority’ (10.7% of respondents referred to the Uzbek community as such). The fact that a few miles away from the border Uzbeks constituted the titular nation only accentuated the sensitivity of the question. What appeared to be the most popular form of self-categorisation was ‘historical nation’ (67.2%). The adjective ‘historical’ was used as synonymous for indigenous (‘korennoi’ in Russian): this fits well with what was said above about the ideas that Uzbeks attach to the term diaspora. The Uzbek views behind their use of the concept stood in stark contrast to attempts from some nationalist elements in the state

40Interview with Alisher Sabirov, deputy at the national parliament and one of the most well known figures in the local Uzbek community (Bishkek, 14 June 2003). 41Interview with local Uzbek politician (Osh, 26 June 2003). 42Interviews held at the Osh branch of the UNCC (Osh, 25 June 2003). 582 MATTEO FUMAGALLI administration to pass legislation (later vetoed by Akaev) that would have favoured ethnic Kyrgyz in land ownership thereby implying that Kyrgyzstani territory was first and foremost Kyrgyz land (Spector 2004, p. 12). Similar to their Kyrgyzstani kin, even Tajikistani Uzbeks rejected a diasporic identity. About one in two Uzbeks did not perceive themselves as constituting a national minority in Tajikistan (50.4%). The question as to whether or not one community may be categorised as a national minority was more a matter of status than mere statistics. No-one even advanced the idea that on a strictly numerical level Uzbeks might not constitute a minority. What was subject to dispute were the implications of this labelling. This entailed a diminished status vis-a`-vis the Kyrgyz. A minority status might also lead to a smaller role in public affairs. Finally, it might generate an encirclement syndrome, a feeling of being isolated from kin in Uzbekistan, and from fellow citizens, because of a categorisation which divided the citizenry into de facto first and second class citizens. The idea of being seen as a national minority was largely rejected on the basis that Uzbeks were the largest community in the country after the titular group, but also constituted the majority in some areas. Negative feelings associated with the term diaspora were also conspicuous (36.8%), to emphasise even further the claim to indigenousness and the fact that Uzbeks ‘did not migrate here [Tajikistan] from anywhere’. In northern Tajikistan the majority of Uzbeks perceived themselves as a historical nation (60.8%),43 and as a result, they advanced a claim of indigenousness and symbolic ownership of the territory (locale). Comments such as the following were common within the Uzbek community: ‘Uzbeks have been here forever’;44 ‘We did not come from anywhere! This is our land’.45 Opting for one term (e.g. diaspora) rather than another (e.g. member of the Kyrgyzstani nation) reveals how the new state and the new state identity were perceived by potentially marginal groups.46 Here is not the place to go into the detail of a comparative analysis between Uzbek perceptions and those of other minority groups (Russians, Germans, Koreans), but it is interesting to note that, despite a common Soviet background, different groups showed varying attitudes towards the territory where they currently resided. Koreans, for example, seemed to view the Central Asian republics as their adoptive homeland, whereas Germans saw Germany as their homeland (Diener 2006), and the views of Russians seemed to be particularly diverse (Gorenburg 2000).

The missing frame: autonomy for Uzbeks The 1990 Osh conflict did not seem to bode well for the integrity of the Kyrgyz SSR. Requests for land swaps with Uzbekistan, or even autonomy if not outright secession of the Uzbek-populated areas, escalated before and in the immediate aftermath of the tragic events in June that year (Asankanov 1996). The Uzbeks’ geographic

43The term historical nation is understood in similar ways in both Kyrgyzstan and Tajikistan. 44Interviews held with academic staff and postgraduate students at the Department of Uzbek Philology at Osh State University (Osh, 28 June 2003). 45Interview with Kadyrjan Batyrov and other members of the Jalalabad UNCC, held at the Katyrov University (Jalalabad, 12 July 2003). 46For an insightful study of locality and space, and how these have been affected by post-Soviet transformations, see Liu (2005). ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 583 concentration and size, along with their various grievances made the emergence of an autonomist frame a possibility. However, despite the fact that the issue of autonomy emerged before the conflict and was promoted by the Adolat organisation, autonomy soon faded as a mobilising frame, giving way to the perception of the local Uzbek population that their present and future would presumably lie within the boundaries of Kyrgyzstan. Pragmatism prevailed and the autonomist frame has virtually disappeared from the agenda of Uzbek organisations and more broadly from the Uzbek mindset. Outright secession has virtually turned into a taboo topic, and a similar fate was reserved for autonomy as well. The number of Uzbeks who openly support the idea of institutionalising Uzbek autonomy appears marginal at best. The following comments were made repeatedly during my visits in southern Kyrgyzstan (June – July 2003 and August 2005) and represent widely held views on the topic from members of the Uzbek community: ‘We do not need autonomy’; ‘Autonomy is dangerous. Autonomy leads to secession’.47 Uzbeks have been traditionally wary of making any move that might be interpreted from part of the Kyrgyzstani political forces as separatist or even autonomist. Even those like the Uzbek businessman and politician Davron Sabirov48 have gradually softened their inflammatory tones in recent years. In Tajikistan, Leninabad (renamed Sughd in the late Soviet era) remained unscarred by the hostilities. The possibility that the northern province would opt for outright secession has haunted the country’s post-Soviet transition. Muriel Atkin (1997, p. 295) recalls that in the very early years of the 1990s (in 1992 particularly) Leninabad appeared on the verge of secession. Officials banned transmissions of radio and television programmes from Dushanbe, and took a range of institutions out of central control following the power struggle that had embroiled centre – north relations. In addition, a separate provincial militia was established that year. Leninabad’s obstructionism lasted until the end of 1992. The role of ethnicity, in the form of the links between cross-border Uzbeks and the role of Uzbekistan, has been the subject of some dispute, but as Schoeberlein (1996) has noted, Leninabad’s secessionist bid may be explained by its refusal to accept Kulyabi domination rather than by its alleged loyalty to Tashkent. A second wave of separatist tendencies emerged in the second half of the 1990s, when elements of the northern opposition coalesced under Abdullajonov’s Movement for National Revival, and appeared to join forces with a renegade of the Tajik army, Col. M. Khudoiberdiev. Khudoiberdiev was commonly referred to as an ethnic Uzbek, despite being a Lokay.49 On several occasions from 1996 to November 1998 disturbances and uprisings took place in different locations of the Sughd province. The final uprising was meant to enable Khudoiberdiev to take control of the whole province, however it ended up with him being forced to retreat into Uzbekistani territory, where the insurgent militia had allegedly crossed into Tajikistan. The insurrection failed, and Abdullajonov and Khudoiberdiev were expelled from the country in a seemingly definitive way.

47Interviews held with members of the Uzbek community in southern Kyrgyzstan (June – July 2003 and August 2005). 48An Uzbek businessman, Sabirov was head of the Osh-based ‘Society of Uzbeks’. 49This is an important distinction. The latest population census aside (2000), Lokays, as well as other Uzbek-speaking communities in Tajikistan, were grouped within the larger Uzbek category. The 2000 census disaggregated the category into smaller ones, thereby leaving Uzbeks with a far smaller share of the overall population. 584 MATTEO FUMAGALLI

Various respondents in Khujand and Dushanbe, regardless of their ethnic identity, noted how ethnicity had little to do with Khudoiberdiev’s actions, and emphasised how looking at his ethnic background does not advance an understanding of his actions and the motives behind them. In addition, many recalled the outcome of his last appearance in Khujand in 1998 when, according to Makhsud Ahmedov,50 Khudoiberdiev met fierce resistance from the Tajik republican army. A direct consequence of these secessionist bids and/or uprisings is represented by the caution with which people, especially Uzbeks, have addressed the question of autonomy for the Sughd province. In fact, autonomy was rarely indicated among the demands of the Uzbeks. Overall, although Leninabad was actually more integrated with Soviet Uzbekistan, in the post-independence era the idea of a ‘republic of Leninabad’ appeared untenable to a large segment of the population (Martin 1997). As the head of the Centre for Democratic Transformations in Khujand explained:51

In the first half of the 1990s there seemed to be a possibility for Leninabad to break away from the rest of Tajikistan. Once the country descended into chaos, some were left wondering about the future prospects. From a practical point of view the region could be self-sufficient, but politically no-one could imagine an independent Leninabad. Realistically speaking, Uzbekistan would expect it to be absorbed within its borders, and Tajikistan would never be keen to lose the engine of the country. Rather than becoming a satellite of Uzbekistan, people prefer to stay within Tajikistan.

Whatever separatist sentiments may have been present then, these faded in the following years. A separatist and even an autonomist agenda now appear not only unrealistic politically, but the very idea of establishing an independent state, being annexed to Uzbekistan, or even just institutionalising autonomy, are options rarely put forward in the north.

Uzbeks under siege? The use of nationalist frames during Kyrgyzstan’s 2000 parliamentary campaign

O’Zing haqqsan, O’Zing khalqsan, O’Zing mezon—menim khalqim!52

The slogan above was used by Davron Sabirov, head of the Osh-based ‘Society of Uzbeks’, during the 2000 parliamentary electoral campaign. While the message appears no more or less threatening than any other populist slogan, what caused alarm among the local population at the time was the style of the poster distributed across Osh. The slogan, in particular the insisted repetition of the letters ‘O’ and ‘z’ was a close reminder of the word O’zbek, with some even seeing a reference to Uzbekistan (O’zbekiston), from where Sabirov was reported to be receiving money and political support. These two allegations have always remained unsubstantiated. None-the-less

50Not his real name (Centre for Democratic Transformations, Khujand, 16 August 2003). 51Interview held in Khujand on 20 August 2003. 52‘You yourselves are right, you yourselves are the people, you yourselves are the scales—my people!’ ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 585 the question of the political backing which Sabirov would be receiving has been a matter of dispute throughout the 1990s.53 Sabirov campaigned in electoral district no. 34 in the north-eastern outskirts of Osh, a deeply impoverished area, fully inhabited by Uzbeks, and commonly known as ‘Amir Temur’. Sabirov lay at the centre of a highly debated controversy during the run-up to the 2000 elections. His candidacy was suspended and re-instated a number of times by the electoral authorities because he was being sued for inciting ethnic hatred. The spark that initiated this was a brief electoral advertisement, broadcast on the local Osh TV station, in which a man wearing a traditional Uzbek hat was seen being mobbed and robbed by a group of people riding horses and wearing the black and white Kyrgyz hat (kalpak). The metaphor of Uzbeks being attacked by Kyrgyz did not require a sophisticated interpretation. Towards the end of the advertisement one could hear a voice reading Mr Sabirov’s electoral slogan. Finally allowed to run, Sabirov eventually won 65% of votes in his district. Sabirov’s electoral campaign did not come as a total surprise to the local population. From 1997 to 2000 he had systematically framed the Uzbek question in (ethnically) antagonistic terms. He made continuous references to Uzbekistan and proposals to shift to a Latin script, for example, which were not met with favour by many even among the Uzbek community.54 The editor of Jalalobod Tongi (the Dawn of Jalalabad) dismissed Mr Sabirov’s campaign as ‘sheer nationalism and populism: Sabirov represents himself and no-one else in the Uzbek community’.55 The 1997 – 2000 period signalled the heightening of inter-ethnic tensions in the south of the country,56 and one of the possible explanations behind this was the framing of the Uzbek question in more radical terms by Sabirov. In particular Sabirov insisted on adopting a more critical stance towards the Kyrgyzstani leadership, given the perceived inefficacy of the regime. Also, Sabirov pushed for the Uzbek organisation to opt for a new denomination: Society of Uzbeks as opposed to ‘national-cultural centre’, a denomination which, as shown in the previous section, many Uzbeks saw as appropriate for Koreans or Russians, but not for themselves. Within a few years (from 1997 to 2000) the Society of Uzbeks became an established and relatively popular organisation in Osh (reportedly comprising 10,000 members57). The organisation benefited from support in the media. The Mezon (Scale) newspaper and Mezon TV58 had, until the 2000 elections, constituted a less complacent voice towards state authorities. Many would regard them as vehicles of nationalist

53The broadcast was very controversial at the time and local Uzbeks and Kyrgyz perceived it in radically opposite ways. I am grateful to Nick Megoran for pointing out this aspect to me. 54The Uzbek language is written using the Cyrillic script in the countries neighbouring Uzbekistan. Uzbekistan, however, switched to the Latin alphabet in 1993 in a move that emphasised the state leadership’s will to depart from Russian influence. 55Interview with the editor of Jalalobod Tongi (Jalalabad, 12 July 2003). 56See Khamidov (2000), and N. S. Megoran, ‘Elections and ethnicity in the south of Kyrgyzstan’, Eurasianet, 29 March 2000, available at: http://www.eurasianet.org/departments/insight/articles/ eav032900.shtml, accessed 12 February 2007. 57Interview with Davron Sabirov (Osh, 17 July 2003). 58Though he did not own the newspaper and the TV station officially, his being in charge was one of the worst-kept secrets in Osh. In 2004 Mezon TV was sold and the close link with Sabirov was severed. Mezon newspaper started to be published again in the run up to the controversial 2005 parliamentary elections where it promoted a pro-Sabirov agenda. 586 MATTEO FUMAGALLI propaganda, others as instruments of Sabirov’s own ambition. Overall, Sabirov’s rhetoric yielded contrasting results. On the one hand his use of the term ‘Jamiyat’ (Sabirov’s organisation’s name in Uzbek59) as opposed to ‘national-cultural centre’,60 emphasising how Uzbeks should not be regarded as a minority ‘like any other smaller nationality’, resonated widely across the Uzbek population and was certainly effective in rallying Uzbek support behind himself and his candidacy. On the other hand, his instrumental use of inflammatory frames during the 2000 elections was short-lived. While Uzbeks favourably received Davron Sabirov’s attempt to present a less complacent stance towards the authorities, they also grew increasingly sceptical of Sabirov’s and the whole Uzbek leadership’s political conduct, seeing it as more interested in promoting their own status than in addressing the community’s demands. Over the years the activities of the O’zbeklar Jamiyati have become less regular.61 And Sabirov’s tones have become more conciliatory towards the authorities.

The Tulip Revolution and its impact on inter-ethnic relations in Kyrgyzstan Although still supportive of the Akaev regime until its final days, Uzbeks grew increasingly discontent with the lack of representation in governing bodies and the lack of response to their demands (Spector 2004, p. 12). As already noted, the situation in Kyrgyzstan is still in flux. While the term ‘revolution’ is certainly misplaced, a re- alignment of political actors has nonetheless followed the ousting of the old elites. The Uzbek leadership, particularly the UNCC, has been discredited by its association with the previous regime. The extent to which the Uzbek population played an active role in the events is highly disputed.62 The initial involvement of southern political elites and of Uzbek leaders (Anvar Artykov was briefly named governor of the Osh region from March to December 2005) seemed to suggest a new era in state – community relations. Instead the situation in 2006 seemed to have taken a turn for the worse. Artykov was forced to leave his post in January 2006; and the Jalalabad Uzbek Cultural Centre (headed by local businessman Kadyrjan Batyrov) sent a petition to the state authorities pointing to the rise in anti-Uzbek discrimination following the so- called revolution (Rotar 2006). It is far too early to speculate whether this points to a new rift in Kyrgyz – Uzbek relations and about the implications this may have on inter-ethnic relations. One should recall, however, that this is not the only cleavage in the country and that turmoil has stemmed more from competition between factions and individual figures than from inter-ethnic differences.

Conclusion This article has shown that several frames have competed for the attention of the Uzbek public: the memory of earlier conflicts and concern for inter-ethnic stability;

59O’zbeklar Jamiyati. 60Milliy Madaniyat Markazi (or Natsional’nyi Kul’turnyi Tsentr’ in Russian, as it is more commonly referred to). 61The half-finished building close to the city library in Osh is testimony to this (I am thankful to Nick Megoran for reminding me of this). 62For contrasting views see Saipjanov (2005) and Saipov (2005). ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 587 autonomist/separatist frames; and nationalist frames. Overall, more moderate messages—two exceptions aside, 1990 – 92 and 1997 – 2000—seemed to prevail and resonated more widely across the Uzbek community in both cases under investigation. The empirical findings presented in this article seem to support four main arguments. First, a study of the frames that emerged before and after independence among Uzbeks in both Tajikistan and Kyrgyzstan suggests that the way the Uzbek question was framed went through several phases. It is possible to associate one or more frames with each phase. The changes in the political structure at Union level and the reconfiguration of the power structure across regions opened a window of opportunity for the formation of new ideas. Whatever separatist or autonomist bids there might have been, the actions taken by state actors prevented them from combining with the grievances into a potentially far more radical mobilising idea. To this end state actors and individual leaders created a frame that widely resonated across the community for the entire post-independence period: supporting the incumbent would be the only means to maintain inter-ethnic stability. As there was no counter-proof this became a sort of self-reinforcing myth. The population was demo- bilised, leading to the exclusion of the Uzbek community from politics. Alongside these ideas a more nationalist one developed framed by Sabirov in the period 1997 – 2000. His softening of tones in 2000 greatly contributed to decreasing the level of tension. By contrast, Tajikistani Uzbeks did not seem to develop a similar set of competing ideas. The lack of any articulated vision appears to be one of the factors behind the lack of political action among Uzbeks living in the northern region. The perceived change in the political opportunity structure in 1990 – 92 with the unfolding power battle between regions contributed to the development of a possible separatist idea which was much more regional than national in essence. The wave of uprisings from 1996 to 1998 owed more to power politics than to separatist bids, and ethnicity played little role in those events. Second, the prevalence of more moderate frames points to the significance of the common Soviet legacy. The language adopted by the Uzbek elites to mobilise the community and to frame Uzbek issues was rooted in the Soviet experience and discourses on internationalism and inter-ethnic harmony. They also built on the legacy of the Soviet nationality policies where Uzbeks expressed utter rejection of labels such as diasporas and minorities. In Soviet times these terms were associated with diminished cultural rights at the very best, and possibly with the experience of the ‘enemy peoples’. Claims were made instead on the basis of Uzbek indigenousness. Being local, in short, meant being able to aspire to a higher status and access to resources. Being ‘from somewhere else’, by contrast, was associated with marginality. Third, data also suggest that frames are quintessentially relational. As it has been shown, the same set of ideas (memory of earlier conflicts, most notably) can be used to either mobilise or indeed demobilise the population. For this reason I argued that the concept of mobilising ideas should be accompanied by that of demobilising ideas. The concept of demobilising idea not only disputes the commonly held expectation that political mobilisation is by its nature confrontational with authorities, but has also helped emphasise how depending on the use political actors decide to make of it, a particular set of frames can simultaneously act as a powerful mobiliser in one direction, but can also demobilise the very same group in another respect. The 588 MATTEO FUMAGALLI example of past experiences of conflict clearly illustrates the way mobilisation and demobilisation can be activated by the very same idea. Fourth and finally, by showing the variation of how frames emerged within and across cases, this article contributes to an understanding of the internal heterogeneity of the Uzbek community in both countries. This has implications for the way ethnic mobilisation is studied as it cautions against the assumption of internal cohesiveness of a community simply out of common ethnic bonds, and calls for a metaphorical de- construction of the internal dynamics and debates with the community itself. Although inter-ethnic relations have remained generally stable in both countries since independence, the episodic emergence of nationalist frames and even autonomist/ separatist frames in limited periods shows that there is a potential constituency that might appropriate a more confrontational frame. Because Kyrgyzstan’s ‘post- revolution’ instability has opened a window of opportunity for social and political actions of various kinds (legal and illegal), preventing the political conflict from slipping out of control and addressing any long-standing grievances appears to be of particular urgency.

University College Dublin

References Abdullaev, K. & Barnes, C. (eds) (2001) Politics of Compromise. The Tajikistan Peace Process (London, Conciliation Resources). Akiner, S. (2001) Tajikistan. Disintegration or Reconciliation? (London, Royal Institute of International Affairs). Asankanov, A. (1996) ‘Ethnic Conflict in the Osh Region in Summer 1990: Reasons and Lessons’, in Rupesinghe, K. & Tishkov, V.A. (eds) (1996) Ethnicity and Power in the Contemporary World (New York, United Nations University Press), pp. 116 – 125. Atkin, M. (1997) ‘Thwarted Democratization in Tajikistan’, in Dawisha, K. & Parrott, B. (eds) (1997), pp. 277 – 311. Beissinger, M.R. (2002) Nationalist Mobilisation and the Collapse of the Soviet State (Cambridge, Cambridge University Press). Bohr, A. (1998) Uzbekistan: Politics & Foreign Policy (London, Royal Institute of International Affairs). Dawisha, K. & Parrott, B. (eds) (1997) Conflict, Cleavage and Change in Central Asia and the Caucasus (Cambridge, Cambridge University Press). Diener, A. (2006) ‘Homeland as Social Construct: Territorialization among Kazakhstan’s Germans and Koreans’, Nationalities Papers, 34, 2, pp. 201 – 235. Eisinger, P.K. (1973) ‘The Conditions of Protest Behaviour in American Cities’, American Political Science Review, 67, pp. 11 – 28. Fumagalli, M. (2005) The Dynamics of Uzbek Ethno-Political Mobilisation in Kyrgyzstan and Tajikistan (1991 – 2003), PhD dissertation, University of Edinburgh. Gagnon, V.P. (2004) The Myth of Ethnic War: Serbia and Croatia in the 1990s (Ithaca, NY, Cornell University Press). Gorenburg, D. (2000) ‘Not with One Voice: An Explanation of Intragroup Variation in Nationalist Sentiment’, World Politics, 53, 1, pp. 115 – 142. Gorenburg, D.P. (2003) Minority Ethnic Mobilisation in the Russian Federation (Cambridge, Cambridge University Press). Gretsky, S. (1995) ‘Civil War in Tajikistan: Causes, Developments and Prospects for Peace’, in Sagdeev, R.Z. & Eisenhower, S. (eds) (1995). Gurr, T.R. (1993) Minorities at Risk (Washington, DC, United States Institute of Peace). Hipsher, P.L. (1996) ‘Democratization and the Decline of Urban Social Movements in Chile and Spain’, Comparative Politics, 28, 3, pp. 273 – 297. ETHNIC MINORITY MOBILISATION IN CENTRAL ASIA 589

Hirsch, F. (2005) Empire of Nations. Ethnographic Knowledge and the Making of the Soviet Union (Ithaca, NY, Cornell University Press). Horsman, S. (1999) ‘Uzbekistan’s Involvement in the Tajik Civil War 1992 – 97: Domestic Considerations’, Central Asian Survey, 18, 1, pp. 37 – 48. Human Rights Watch (1998) ‘Tajikistan. Leninabad: Crackdown in the North’, 10, 2, April. Huskey, E. (1997) ‘Kyrgyzstan: The Fate of Political Liberalization’, in Dawisha, K. & Parrott, B. (eds) (1997), pp. 242 – 276. Ilkhamov, A. (ed.) (2003) Etnicheskii Atlas Uzbekistana (Tashkent, Institut Otkrytoe Obshchestvo). Johnston, H. & Noakes, J.A. (eds) (2005) Frames of Protest. Social Movements and the Framing Perspective (Lanham, Rowman & Littlefield). Jones Luong, P. (2002) Institutional Change and Political Continuity in Post-Soviet Central Asia (Cambridge, Cambridge University Press). Kamp, M. (2002) ‘Pilgrimage and Performance: Uzbek Women and the Imagining of Uzbekistan in the 1920s’, International Journal of Middle Eastern Studies, 34, 2, pp. 263 – 278. Khamidov, A. (2000) ‘Osh Anniversary Events—A Source of Discontent in Kyrgyzstan’, Eurasianet, 19 October. Kitschelt, H. (1986) ‘Political Opportunity Structures and Political Protest: Anti-Nuclear Movements in Four Democracies’, British Journal of Political Science, 16, pp. 57 – 85. Koopmans, R. (2000) ‘Migration and Ethnic Relations as a Field of Contention: A Political Opportunity Structure Approach’, in Koopmans, R. & Statham, P. (eds) (2000) Challenging Immigration and Ethnic Relations Politics. Comparative European Perspectives (Oxford, Oxford University Press), pp. 13 – 54. Kratkie itogi perepisi naseleniya 1999 goda v Respublike Kazakhstana (1999) (Almaty, Agentstvo Respubliki Kazakhstana po Statistike). Kriesi, H. (2004) ‘Political Context and Opportunity’, in Snow, D.A. et al. (eds) (2004), pp. 67 – 90. Lijphart, A. (1971) ‘Comparative Politics and Comparative Method’, American Political Science Review, 65, pp. 682 – 693. Liu, M.Y. (2005) ‘Hierarchies of Place, Hierarchies of Empowerment: Geographic of Talk about Postsocialist Change in Uzbekistan’, Nationalities Papers, 33, 3, pp. 421 – 438. Lubin, N. & Rubin, B.R. (1999) Calming the Ferghana Valley. Development and Dialogue in the Heart of Central Asia (Washington, DC, The Twentieth Century Fund). Martin, K. (1997) ‘Welcome to the Republic of Leninabad?’ Central Asia and the Caucasus, 4, 10. Martin, T. (2002) An Affirmative Action Empire: Nations and Nationalism in the Soviet Union, 1923 – 39 (Ithaca, NY, Cornell University Press). McAdam, D. (1982) Political Process and the Development of Black Insurgency 1930 – 70 (Chicago, Chicago University Press). McAdam, D., Tarrow, S. & Tilly, C. (2001) Dynamics of Contention (Cambridge, Cambridge University Press). Megoran, N. (2002) The Borders of Eternal Friendship? The Politics and Pain of Nationalism and Identity along the Uzbekistan – Kyrgyzstan Ferghana Valley Boundary, 1999 – 2000, PhD disser- tation, University of Cambridge. Melvin, N.J. (2000) Uzbekistan. Transition to Authoritarianism on the Silk Road (Amsterdam, Harwood). Narodnoe khozyaistvo Kirgizskoi SSR (1982) (Frunze, Komitet po Statistike). Naselenie Kirgizskoi Respubliki 1999 (1999) (Bishkek, Komitet po Statistike). Naselenie Respubliki Tadzhikistan 2000 (2002) (Dushanbe, Gosudarstvennyi Komitet Statistike Respubliki Tadzhikistan). Nedelmann, B. (1987) ‘Individuals and Parties – Changes in Processes of Political Mobilization’, European Sociological Review, 3, 3, pp. 181 – 202. Nourzhanov, K. (2005) ‘Saviours of the Nation or Robber Barons? Warlord Politics in Tajikistan’, Central Asian Survey, 24, 2, pp. 109 – 130. Olcott, M.B. (1994) ‘Nation-building and Ethnicity in the Foreign Policies of the New Central Asian States’, in Szporluk, R. (ed.) (1994) National Identity and Ethnicity in Russia and the New States of Eurasia (London, M.E. Sharpe). Olcott, M.B. (1995) ‘Ethnic Violence in Central Asia: Perceptions and Misperceptions’, in Sagdeev, R. & Eisenhower, S. (eds) (1995), pp. 115 – 125. Perepis Naseleniya Turkmenistana 1995 (1997) (Ashgabad, Goskomitet Turkmenistana po Statistike). Przeworski, A. & Teune, H. (1970) The Logic of Comparative Social Inquiry (Malabar, Krieger). Radnitz, S. (2005) ‘Networks, Localism and Mobilisation in Aksy, Kyrgyzstan’, Central Asian Survey, 24, 4, pp. 405 – 424. 590 MATTEO FUMAGALLI

Ragin, C. (1987) The Comparative Method. Moving beyond Qualitative and Quantitative Strategies (London, University of California Press). Remmer, K.L. (1980) ‘Political Demobilization in Chile, 1973 – 78’, Comparative Politics, 12, 3, pp. 275 – 301. Renwick, A. (2006) ‘Why Hungary and Poland Differed in 1989: The Role of Medium-Term Frames in Explaining the Outcomes of Democratic Transition’, Democratization, 13, 1, pp. 36 – 57. Roeder, P. (1991) ‘Soviet Federalism and Ethnic Mobilisation’, World Politics, 43, 2, pp. 196 – 232. Ro’i, Y. (1991) ‘Central Asian Riots and Disturbances 1989 – 90: Causes and Context’, Central Asian Survey, 10, 3, pp. 21 – 54. Rotar, I. (2006) Uzbeks appeal to Bakiev, claiming ethnic discrimination. Jamestown Foundation Eurasia Daily Monitor, 3(29), 10 February 2006, available at: http://www.jamestown.org/ publications_details.php?volume_id¼414&issue_id¼3615&article_id=2370760, accessed 30 March 2007. Sagdeev, R. & Eisenhower, S. (eds) (1995) Central Asia: Conflict, Resolution and Change (Chevy Chase, CPSS Press). Saipjanov, A. (2005) ‘Uzbeks in Kyrgyzstan View the Revolution with Caution’, Eurasianet, 31 March. Saipov, A. (2005) ‘Myth of Pro-Akaev Uzbeks Shattered’, IWPR, 364, 30 March. Schatz, E. (1999) ‘Notes on the ‘‘Dog that Didn’t Bark’’: Eco-Internationalism in Late Soviet Kazakstan’, Ethnic & Racial Studies, 22, 1, pp. 136 – 161. Schoeberlein-Engel, J.S. (1996) ‘Some Thoughts on Tajikistan Events’, Dialogue on CenAsia List, February 1996, available at: http://www.hartford-hwp.com/archives/53/116.html, accessed 30 March 2007. Slezkine, Y. (1994) ‘The USSR as a Communal Apartment, or How a Socialist State Promoted Ethnic Particularism’, Slavic Review, 53, 2, pp. 414 – 452. Snow, D.A. (2004) ‘Framing Process, Ideology, and Discursive Fields’, in Snow, D.A. et al. (eds) (2004), pp. 380 – 412. Snow, D.A. & Benford, R. (1992) ‘Master Frames and Cycles of Protest’, in Morris, A.D. & McClurg Mueller, C. (eds) (1992) Frontiers in Social Movement Theory (New Haven, CT, Yale University Press). Snow, D.A., Rochford Jr., E.B., Worden, S.K. & Benford, R.D. (1986) ‘Frame Alignment Process, Micromobilization, and Movement Participation’, American Sociological Review, 51, pp. 464 – 481. Snow, D.A., Soule, S.A. & Kriese, H. (eds) (2004) The Blackwell Companion to Social Movements (Oxford, Blackwell Publishing). Spector, R.A. (2004) The Transformation of Askar Akaev, President of Kyrgyzstan (Berkeley, Post- Soviet Program). Suny, R.G. & Martin, T. (eds) (2001) A State of Nations: Empire and Nation-Making in the Age of Lenin and Stalin (Oxford, Oxford University Press). Tabyshalieva, A. (1999) The Challenge of Regional Co-operation in Central Asia. Preventing Ethnic Conflict in the Ferghana Valley (Washington, DC, USIP Press). Tabyshalieva, A., Alisheva, A. & Shukurov, E. (1999) Social Reality in South Kyrgyzstan: Ethnic Situation (Bishkek, Institute for Regional Studies). Tarrow, S. (1991) ‘‘‘Aiming at a Moving Target’’: Social Science and the Recent Rebellions in Eastern Europe’, PS: Political Science and Politics, 24, 1, pp. 12 – 20. Tarrow, S. (1998) Power in Movement: Social Movements and Contentious Politics, 2nd Edn (Cambridge, Cambridge University Press). Tarrow, S. (2005) The New Transnational Activism (Cambridge: Cambridge University Press). Tilly, C. (1978) From Mobilization to Revolution (New York, Random House). Tishkov, V. (1995) ‘‘‘Don’t Kill Me, I’m a Kyrgyz!’’: An Anthropological Analysis of Violence in the Osh Ethnic Conflict’, Journal of Peace Research, 32, 2, pp. 133 – 149. Tishkov, V. (1999) ‘Ethnic Conflicts in the Former USSR: The Use and Misuse of Typologies and Data’, Journal of Peace Research, 36, pp. 571 – 592. Ukudeeva-Freeman, J.T. (2003) Collective Action Problem: Mobilization of National-Democratic Movements in Azerbaijan and Kyrgyzstan, PhD dissertation, University of California, Riverside, CA. Walter, B.F. (1999) ‘Designing Transitions from Civil War. Demobilization, Democratization, and Commitments to Peace’, International Security, 24, 1, pp. 127 – 155. Zald, M.N. (1996) ‘Culture, Ideology, and Strategic Framing’, in McAdam, D., McCarthy, J. & Zald, M.N. (eds) (1996) Comparative Perspectives on Social Movements. Political Opportunities, Mobilising Structures, and Cultural Framings (Cambridge, Cambridge University Press), pp. 261 – 274.