All Against All: The Tajik Civil War (1991-1997)
As the Soviet Union crumbled in 1991, the West celebrated the liberation of newly minted states from Soviet domination. However, these celebrations did not apply in certain corners of the former Soviet Empire where the legacy of foreign domination ended with the eruption of civil war. In Transdnistria, Abkhazia, South Ossetia, and Tajikistan conflicts began. Of all these conflicts, the civil war in Tajikistan took the most lives, lasted the longest, and held the attention of foreign powers the longest. Conflicts in Georgia (Abkhazia and South Ossetia) and Moldova (Transdnistria) centered on secessionist movements where there were clearer lines between forces and clearly defined goals. For the purposes of this paper, these conflicts fall into the category of “implosive” civil wars. The civil war in Tajikistan is an example of an explosive civil war—a civil war resulting in the collapse of the state and of civil society. While scholars (Atkin, 1992; Rubin, 1993; Akbarzadeh, 1996; Lynch, 2001; Nourzhanov, 2005; Kilavuz, 2009) have approached the Tajik Civil War before, the primary focus of this literature has been identity politics, the existence (or nonexistence) of nationalism, elite-driven power struggles, warlordism, and network-activation. There is not an accounting for why levels of violence varied across space and time during the Tajik Civil War. This lack is important because a first-cut analysis on how violence spread in the Tajik Civil War might serve as a springboard for a larger theoretical framework for the metastasis of civil war violence. In this paper, I present preliminary results in an attempt to explain the sub-national variation in violence in the Tajik Civil War.
Wilder Bullard is a PhD candidate in Comparative Politics at the George Washington University. Mr. Bullard specializes in the relationships between democratization, liberalization, and conflict in the Former Soviet Union. Mr. Bullard has published editorials in The Guardian (London), Hurriyet (Istanbul), and Ekathimerini (Athens). Mr. Bullard is a research at the Program on New Approaches to Research and Security in Eurasia, where is a regular contributor of information briefs on their website on topics ranging from low intensity conflict in post-Soviet Central Asia to power politics in the Russian Federation.
To gain leverage on sub-national variation, I engage with the extant literature concerning the “motivators and sustainers” of civil war. This literature is concerned with what factors directly impact the onset of civil wars (motivators) and which impact the duration of civil wars (sustainers). I use these different motivators and sustainers as independent variables to test how the Tajik Civil War would have appeared (the sub-national variation in violence) if these factors were present. These motivators and sustainers are as follows: ethnic diversity (Sambanis, 2002), greed (Collier & Hoeffler, 2002), geography (Buhaug & Gates, 2002), state capture, and ideology. Using these motivators and sustainers as tests against the empirics of the conflict, I examine the expected outcomes that the above authors posit in order to better understand the underlying drivers of violence in the Tajik civil war.
A War of All Against All: Tajikistan 1991-1997
In studying post-Soviet Central Asia, it is important to note that “the [Central Asian Republics] did not leave the USSR, the USSR left them” (Helf, 2010). In the case of Tajikistan, the departure of Soviet patronage at the end of 1992 created a power vacuum that accelerated the processes which eventually led to the civil war. While it is not the purpose of this paper to explore why the Tajik Civil War unfolded, the reasons motivating the initial conflict are important in understanding how the conflict unfolded. So, while the narrative of the civil war that follows is presented as if in a vacuum, it is important to note that the Tajik Civil War can be seen as epiphenomenal to the larger collapse of the Soviet Union.
Beneath the larger collapse of the Soviet Union, the election of Rahmon Nabiev to the presidency of Tajikistan in November 1991 is an excellent starting point to discuss the subsequent civil war. Nabiev had previously served as the First Secretary of the Communist Party of Tajikistan, but Gorbachev had him sacked in 1985 as part of a USSR-wide purge[i]to replace Brezhnev appointees with those who were more sympathetic with glasnost’ and perestroika. Despite his political revival, Nabiev soon proved inept at handling an independent Tajikistan as local political interests soon drove apart national consensus (Nourzhanov, 2005). In May 1992, when Nabiev formed the GNR[ii] of Tajikistan, it was rejected by Leninobod and Kulob as favoring the regions of Gharm and Badakhshan unfairly. After the formation of the government was rejected by these two regions, accounts soon spread of non-state paramilitary organizations being formed in Leninobod, Kulob and Qurghonteppa. This same week, Nabiev began distributing as many as 18,000 weapons to his supporters so that he could form a presidential guard. Armed clashes broke out in Dushanbe as supporters of Nabiev clashed with supporters of the Democratic Party of Tajikistan, the outlawed Islamic Renaissance Party of Tajikistan, and a group led by ex-prisoner and warlord Sangak Safarov known as the Popular Front.
State power in Dushanbe and across the other population centers of Tajikistan quickly unraveled and warlordism soon prevailed across the country. The majority of intense intra-Tajik fighting occurred between May and December 1992 and led to the deaths of 50,000 and the displacement of 800,000[iii] (ibid). Dushanbe traded hands numerous times as opposition forces pushed out the Nabiev government, only to have hardliners backed by Hissaris, Kulobis and the Popular Front[iv] retake the capital (Rubin, 1993). Regional power struggles exacerbated the political crisis in the capital as regions such as Badakhsan declared sovereignty and other regions refused to accept the Kulobi dominated government elected in 1994. Although the majority of the high-intensity violence had already occurred by 1994, the civil war continued officially until 1997 when a general peace agreement was signed in Moscow. Violence continued across Tajikistan, even after the general peace agreement, and still continues to threaten the state occasionally.
To account for variations in sub-national violence in the Tajik Civil War, I employed a number of methods, both qualitative and quantitative. A mixed-methods approach allowed for greater leverage in teasing out causal relationships, as quantitative analysis quickly became problematic when a high volume of information concerning the conflict was either unavailable or nonexistent.
In addition to exploring the secondary literature concerning Tajikistan and holding informational sessions, I analyzed what the print media reported during the duration of the conflict. Due to the availability and ease of access, I primarily relied on the New York Times online database of articles.[v] I used these articles to develop a micro-history of the conflict and to discern who was perceived to be a major player in the conflict at the international level. More importantly, I used these articles to track where large events were occurring. Understandably, during this time period reporters were undoubtedly constrained in their movements, so their reporting is primarily focused around population centers such as Dushanbe and Gharm. That being said, there is still a significant amount of reporting concerning events that occurred in smaller villages and on the Afghan-Tajik border. So, while there is undoubtedly and urban-rural bias in the reporting of the conflict, the articles I examined did include references to events that took place in the predominantly rural areas of Tajikistan. The Associated Press and ITAR-TASS are both cited within these articles, but are not the primary sources for these articles. In future research, I intend to expand my sources and gain better access to the ITAR-TASS archives as well as examine local media releases during that time period.
The following tables are the raw information and some of the descriptive statistics of my quantitative analysis of the Civil War.
Table 1: Raw information on the Tajik Civil War
|Skirmishes||6||5||1992||Dushanbe||State Security Services versus Opposition||14|
|Skirmishes||7||5||1992||Dushanbe||State Security Services versus Opposition||60|
|Massacre||10||5||1992||Dushanbe||State Security Services versus Opposition||6|
|Battle||24||10||1992||Dushanbe||Muslim-Democratic Coalition versus Hardliners||6000|
|Battle||25||10||1992||Dushanbe||Muslim-Democratic Coalition versus Hardliners||150|
|Massacre||12||11||1992||Yavan||Hardliner Forces against local population||800|
|Battle||10||12||1992||Dushanbe||Muslim-Democratic Coalition versus Hardliners||N/A|
|Battle||18||12||1992||Kofernikhon||Muslim Rebels versus Uzbek Military||19|
|Battle||19||12||1992||Dushanbe||Muslim Rebels versus Uzbek Military||2|
|Battle||7||1||1993||Dushanbe||201st Motorized Rifle Division versus Rebels||N/A|
|Battle||21||2||1993||Komsomolabad||Hardliner Forces versus Muslim Rebels||N/A|
|Battle||21||2||1993||Navabad||Hardliner Forces versus Muslim Rebels||N/A|
|Battle||21||2||1993||Gharm||Hardliner Forces versus Muslim Rebels and Afghan Fighters||1000|
|Battle||11||6||1993||Afghan-Tajik Border||FSB versus Muslim Rebels||25|
|Battle||18||6||1993||Kofernikhon||Hardliner Forces versus Muslim Rebels||45|
|Battle||19||6||1993||Pyandzh||FSB and Hardliners versus Muslim Rebels||2|
|Battle||–||10||1993||Afghan-Tajik Border||FSB versus Afghan Fighters||N/A|
|Assassinations||–||6||1994||Dushanbe||FSB and Government Employees versus Rebels||12|
|Skirmishes||–||8||1994||Afghan-Tajik Border||FSB and Rebels||57|
|Battle||12||4||1995||Khorog||FSB and Rebels||199|
|Battle||23||4||1995||Maymay||Russian Air Force versus Rebels||12|
|Massacre||22||1||1996||Dushanbe||Rebels versus Religious Leaders||4|
|Battle||9||2||1996||Central Tajikistan||Government Forces versus Rebels||N/A|
|Battle||4||12||1996||Gharm||Government Forces versus Rebels||100|
|Battle||26||12||1996||Afghan-Tajik Border||FSB versus Rebels||N/A|
|Battle||10||8||1997||Dushanbe||Government Forces versus Government Forces||N/A|
|Battle||19||8||1997||Kurgan Tyube||Governmenet Forces versus Col. Khudoyberdiev||50|
|Massacre||17||10||1997||Dushanbe||Rebels versus Presidential Guards||14|
|Battle||17||10||1997||Dushanbe||Rebels versus Presidential Guards||N/A|
|Battle||1||5||1998||Dushanbe||Government Forces versus Rebels||20|
|Battle||3||5||1998||Dushanbe||Government Forces versus Rebels||N/A|
Table 2: Participant Frequency, Dyad Frequency, Violence of Dyads
|Participants||Freq(Participants)||Dyads[vi]||Freq(Dyads)||% of Total Violence|
|Nabiev State Security Forces||9.3||MDC/HL||20||72.3|
|Yavan Local Population||3.1||MR/HL||20||N/A|
|Tajik Gov’t Forces||21.87||GF1/CK||6.6||0.51|
Table 3: Location Frequency, Percentage of Violence at Location
|Location||Frequency||Violence||% Total Violence|
The above data helps form a first-cut analysis of the geographical, temporal and political distribution of violence throughout Tajikistan from 1991 to 1998. Also, it provides a means of testing the different logics contained within the motivators and sustainers literature discussed earlier. The motivators and sustainers I chose to analyze are: grievance, greed, geography, state-capture, or ideology. With each motivator/sustainer there is a key assumption that provides the logic that could explain the variation in levels of violence throughout Tajikistan during the civil war. These assumptions are as follows: Conflicts motivated or sustained by…
- Grievances will be most intense in areas high levels of ethnic diversity, as the perceived “enemy” is close by and can be easily mobilized against and targeted.
- Greed will be centered on major sites of industrial production, known trade routes, or agricultural zones.
- Geography will focus on geographical areas that allow for large scale movement and maneuver. Difficult terrain should inhibit high levels of civil war violence.
- State-capture will focus on the taking of the capital city of the state and the political institutions governing the state. It is not so much a struggle over ideology, but over the control of the (in)tangible resources of the state.
- Ideology will focus on taking of the capital city and other key population centers and will seemingly seek to restructure the state.
To test these different theories, I mapped the sites of the different battles from my data onto a map of Tajikistan and then compared it to maps outlining ethnic diversity, economic production, and a topographical relief map. The results and implications of these tests are in the following section.
My central research question sought to examine what accounts for variation in sub-national violence in the Tajik Civil War. Through the use of historiography, information sessions, contemporary print media and my original data set, I constructed a means of testing different theories of how violence would appear in the Tajik Civil War. Below, I take each of the motivators and sustainers chosen for this paper and discuss how they match up against the empirical data of how the conflict actually occurred.
The theory that ethnic diversity served to motivate and sustain civil war in Tajikistan has several assumptions and implications. First is that ethnicity is a clearly defined concept in the region with readily identifiable fault lines. As was discussed in the section devoted to historiography, this assumption appears problematic. It is unclear whether the key motivating factor would be tension between Uzbeks and Tajiks; Tajiks and Russians; Tajiks, Russians, and Tajiks; or an intra-Tajik split between those in different regions of the country that still claim the same identity, such as those Tajiks in Dushanbe against those in Khorog.
The dataset I collected provides insight into the strength of this variable in explaining sub-national variation in violence. If the key assumption of grievances is a clearly defined ethnic or class enemy, then the participants in conflicts should be mobilized along ethnic lines according to that logic. However, when compared to the data collected, it would seem that ethnic grievances might not explain that much of the sub-national variation. The most frequent participant in the conflict was the group Muslim rebels, participating in 50% of all of the events in this dataset.
As Islam is the predominant religious ideology in this region, this seems to provide little explanatory power in terms of ethnic grievance. The most frequent dyad in the dataset was the combination of the FSB and Muslim rebels, accounting for 33.3% of all events. While this frequency might lend some weight to ethnic diversity as an explanatory variable, most of these events occurred on the Afghan-Tajik border, with poorly reported information concerning number of casualties and time of location.
Furthermore, the FSB is a foreign actor that was not necessarily trying to decide the superior ethnic group that would lead a future Tajikistan.[vii]Perhaps the most telling conclusion can be drawn from the fact that the most violent dyad, Muslim Rebels contra Hardliners, accounts for 72.3% of the deaths in my dataset while both were ostensibly ethnically Tajik. When I compared maps of ethnic diversity to the geographical distribution of events in my data set, another potential explanation arose as Dushanbe was both the focus of much of the fighting and a center of high levels of ethnic diversity. This can be excluded from my analysis, however, as the other ethnic group in Dushanbe were Russians who either fled or did not take part in the fighting (Nourzhanov, 2005).
Before the data was examined, it seemed problematic that greed would be a significant factor in explaining the variations in violence in the civil war. Tajikistan’s economy was highly subsidized by the Soviet Union, as domestic industrial output was one of the lowest in the entire USSR and Tajikistan was the poorest Union-level republic. While the dearth of economic opportunities might lend credence to the idea that greed would motivate civil war following the collapse of Soviet authority in 1991, the empirical evidence does not bear out this claim. The central assumption of greed as a motivator and sustainer of civil war is that centers of production should be the focal points for conflict. If certain factions in the civil war can gain access to material resources that are highly valued to the outside world, then they should be able
to sustain their combat readiness.To test if greed has high explanatory value, I examined the frequency of location for events in my dataset and then compared them to the areas of high industrial output. The areas in Tajikistan labeled as centers of industrial output are Regar, Taboshar, Dushanbe and Qurghonteppa.[viii] Dushanbe and Qurghonteppa contained food processing plants and were areas of high conflict in the civil war. Regar is a site for aluminum production, but was not a center for significant fighting in the civil war.[ix]But most telling is the fact that a weapons manufacturing plant is located in Taboshar, but Taboshar was not involved at all in the conflict. It would seem that such a lucrative target would attract the attention of the armed groups party to the conflict, but Taboshar, and the province of Leninabad as a whole, was not highly involved in the conflict. The fact that relatively low-yielding industries such as food processing were located in cities that were the sites of major battles seems to be coincidental and not a motivating factor when two cities with higher value production centers did not experience significant fighting.
The relationship between geography and how violence in a civil war might be distributed seems to be a clear relationship. Buhaug and Gates (2002) explore the various geographical determinants of civil wars. One of their key claims is that civil wars occurring in more mountainous regions will have a larger scope, ceteris paribus (ibid; 422). At first glance, this
seems to explain variations well for the Tajik Civil War, as 90 percent of Tajikistan is mountainous. Buhaug and Gates couch their study in terms of government forces and rebels. While my dataset is also coded along these lines, the reality of the Tajik civil war derived from the historical record shows that it was not a bi-polar conflict, but rather a series of smaller dyadic struggles between warlords, ideological groups, foreign fighters and others. That being said, the concentration and frequency[x] of violence is heavily clustered around Dushanbe, an area of Tajikistan that is less mountainous than others. In future field work, I intend to mark the location of events using GIS and analyze the distribution of violence throughout Tajikistan comparing elevation, presence of forests and rivers.[xi] As I am still collecting data, it does not seem that I can entirely reject geography as a motivator or a sustainer for conflict, but it does not seem to account for the concentration of violence around Dushanbe.
Recorded instances of battles, massacres, assasinations or other events in the Civil War
Of all potential motivators and sustainers discussed in the literature, state capture is the most obvious. It makes sense that the opposition forces in a civil war would want to either capture the state, or destroy the state’s capacity to intervene.[xii]In the case of the Tajik Civil War, this explanation seems to carry some causal weight. As has already been discussed, Dushanbe experienced the most casualties and instances of combat than any other location in the civil war. The capital city changed hands at least twice, with large periods of anarchy and incomplete domination. However, Rakhmon’s[xiii]election to the office of President and the control of Dushanbe by the government of Tajikistan, assisted by the Uzbek and Russian militaries, seemed to mark the end of high levels of intense fighting throughout the country.
This being said, if the capture of Dushanbe was the central goal of opposition participants in the conflict, the narrative of state capture does not explain why violence continued in regions of the country such as Gorno-Badakhshan years after Dushanbe was in the hands of Rakhmon’s government. Further field research is required so that I can gather more data on the nature of the fighting for the capital and the violence that continued in the four years after the capital was firmly taken by Rakhmon and forces allied to the government.
Ideology presents the largest problem in my analysis. This is because many groups had ideological labels, such as the Islamic Renaissance Party, the Democratic Party, and the Communist Party of Tajikistan, but there was never a time during the conflict when the IRPT or the DPT[xiv]were able to implement policy choices that were radically different than those policy choices under Soviet rule. In my dataset, the most violent dyad is ostensibly between two ideological enemies, the Muslim-Democratic Opposition and the Nabiev Hardliners.[xv] However, the dataset does not capture the presence of warlords that fought without clear ideological loyalties. For example, Sangak Safarov had been a prisoner for years during Soviet rule, but still supported a group that consisted mainly of Hardliners from the old regime (Nourzhanov, 2005). More research is needed, especially with members of the IRPT and other oppositionist parties, to discern the ideological nature of the conflict.
The main goal of this paper was to provide a preliminary explanation for the sub-national variations in violence in the Tajik Civil War. To approach this question, I analyzed the historiography concerning the development of the Tajik state and identity, held information sessions with professionals and academics familiar with the conflict and the state of affairs “on the ground,” delved into the archives of contemporary print media, and developed a unique dataset to compare the empirics of the conflict with theories of how civil wars are motivated and sustained. While ethnic grievances, greed and potentially ideology seem weak as explanatory variables for the variation in sub-national violence, state capture and geography are clearly two areas where more work is needed. I have mentioned my future research agenda throughout this paper, and hope to conduct this research as soon as possible.
To revisit a point briefly made in the introduction, it is important that scholars pay attention to how civil wars unfold. As conflict zones like Afghanistan and Western Pakistan exhibit similar characteristics of the conflict in Tajikistan, a theoretical model of civil war violence derived from experience in Central Asia could serve as helpful heuristic to enhance policymakers’ understanding of conflict in the region. Much more empirical work needs to be done on-site in Tajikistan, and the findings of field research, I am confident, will shed greater light on the varying levels of civil war violence from 1991-1997 in Tajikistan and will serve as a springboard to study other conflicts in the former Soviet Union.
Akbarzadeh, S. (1996). Why Did Nationalism Fail in Tajikistan. Europe-Asia Studies, 48(7), 1105-1129.
Atkin, M. (1992). Religious, National, and Other Identities in Central Asia. In J. Gross (Ed.), Muslims in Central Asia (pp. 46-73). Durham: Duke University Press.
Blakkisrud, H. & Nozimova, S. (2010) “History writing and nation building in postindependence Tajikistan”, Nationalities Papers, 38 (2), 173 — 189
Buhaug, H. & Gates, S. (2002). The Geography of Civil War. Journal of Peace Research, 39(4), 417-433.
Collier, P. & Hoeffler, A. (2002). The Centre for the Study of African Economies Working Paper Series. Retrieved Apr. 26, 2010, from Centre for the Study of African Economies, Berkeley, CA. Web site: http://www.bepress.com/csae/paper160.
Helf, G. (2010). Information Session.
Hirsch, F. (2005). Empire of Nations: Ethnographic Knowledge and the Making of the Soviet Union. Ithaca, NY: Cornell University Press.
Khazanov, A.M., (1976). “Characteristic features of nomadic communities in the Eurasian steppes,” The Nomadic Alternative: Modes and Models of Interaction in the African-Asian Deserts and Steppes, W. Weissleder, ed
Kilavuz, I. (2005). The role of networks in Tajikistan’s civil war: Network activation and violence specialists. Nationalities Papers, 37(5), 693-717.
Lynch, D. (2001). The Tajik Civil War and Peace Process. Civil Wars, 4(4), 49-72.
Martin, T. (2001). The Affirmative Action Empire: Nations and Nationalism in the Soviet Union, 1923-1939. Ithaca, NY: Cornell University Press.
Marton, E. (1976). Nationalities and Nationalism in the USSR: A Soviet Dilemma. Georgetown University: The Center for Strategic and International Studies.
Nourzhanov, K. (2005). Saviours or Robber Barons? Warlord Politics in Tajikistan. Central Asian Survey, 24(2), 109-130.
Patai, R., (1951). “Nomadism: Middle and Central Asian,” Southwestern Journal of Anthropology, vol. 7.
Pierce, R A. (1960). Russian Central Asia: 1867-1917. Berkelely: University of California Press.
Rubin, B. (1993). The Fragmentation of Tajikistan. Survival, 35(4), 71-91.
Sambanis, N. & Elbadawi, I. (2002). How much war will we see? Explaining the prevalence of civil war. Journal of Conflict Resolution, 42(3), 307-334.
Spuler, B., (1970). “Central Asia from the Sixteenth Century to the Russian Conquests,” The Cambridge History of Islam, P.M. Holt, A.K.S. Lambton, and B. Lewis, eds., vol. 1, pp. 468-494
Suny, G. (1998). The Soviet Experiment: Russia, the USSR and the Successor States. Oxford: Oxford University Press.
Suny, R. (1993). Revenge of the Past. Stanford: Stanford University Press.
Wheeler, G. (1974). Russian Conquest and Colonization of Central Asia. In T. Hunczak (Ed.), Russian Imperialism from Ivan the Great to the Revolution (pp. 264-299). New Brunswick: Rutgers University Press.
All articles from the New York Times were extracted from the following website between the dates of 4/12/10 and 4/19/10:
All of these maps were gathered from the following resource: http://lib.utexas.edu/maps/tajikistan.html