The Deepening Politics of Fragmentation in Uganda: Understanding Violence in the Rwenzori Region

Abstract In November 2016, Uganda’s armed forces raided the Rwenzururu kingdom palace in Kasese Municipality, arresting and detaining the king and other kingdom officials on treason and other charges. This was the climax to a puzzling wave of violence that was then unfolding in the Rwenzori Region. We consider this violence an unintended consequence of the deepening politics of fragmentation, which takes two forms: “kingdomization” and “districtization.” Through fragmentation, Uganda’s ruling elites seek to weaken subnational concentrations of power, resources, and legitimacy wielded by otherwise coalesced, potentially strong, subnational authority structures and sociopolitical groups. Fragmentation fractures preexisting intra-regional unity, generates new conflicts, and reopens old wounds, leading to violent encounters at the sub-national level, between regional sub-groups, and with the central state. This unfolding of violent encounters involving both state and non-state actors has important ramifications for managing national security within socially fragile contexts and a politically fragmented polity.


Introduction
The violence which occurred during the period from 2012 to 2016 in western Uganda's Rwenzori Region climaxed in the attack on the palace of the Obusinga bwa Rwenzururu (Rwenzururu Kingdom), and the arrest of Omusinga (King) Charles Wesley Mumbere Irema-Ngoma in November 2016. This phenomenon reflects Uganda's deepening politics of subnational fragmentation through the proliferation of subnational power centers. While it is antithetical to national cohesion and social harmony, fragmentation serves to provide opportunities for localized, subnational contestations which reduce pressure on central-state elites; and to incentivize central-state elites to retain control over subnational spaces. In a bid to consolidate power and control through fragmentary subnational power structures and spaces, central-state elites engender conflicts between and within ethno-regional authorities and social groups, and by extension between the state and various subnational structures.
The dynamics of fragmentation incubate conditions which are conducive for national-subnational political contestations and social tensions that lead to violence. This outcome is not what central elites intend, but it is a logical outcome of the overall strategy of fragmentation as a technology of rule. In the specific context of Uganda and President Yoweri Museveni's rule, a violent political environment and security crisis play perfectly into his selfpositioning as a security president and as the best-placed leader to manage conditions of insecurity and instability. Museveni has been adept at using conflict episodes to project power and signal to his opponents that militarism is central to his response to any challenges to his power (Kagoro 2016:157).
The existing literature attributes outbreaks of violence in Africa to a variety of causes: state weakness (Reno 1998;Rotberg 2003;Herbst 2000), fusion between ethno-nationalism, irredentism, and secessionism (Brubaker & Laitin 1998;Saideman & Ayres 2000), ethnicity (Horowitz 1985), and poor governance (Bates 2008;Young 2012;Williams 2016). This article seeks to contribute to the literature by focusing on how intra-national violence results from the political machinations of the elites, which they employ to monopolize power and maintain their stranglehold over societies. Besides fragmentation, which includes the multiplication of local government units and traditional institutions by fragmenting preexisting ones, the elites' strategies of control also include electoral machinations (Schedler 2002;Dhizaala 2020), manipulation of security services (Kagoro 2016), repression of political opposition (Abrahamsen & Bareebe 2016;Khisa 2019;Grasse et al., 2021), and patrimonial governance, which also entails the multiplication of public institutions (Golooba-Mutebi & Hickey 2016). Fragmentation strategies unintentionally undermine peace and security between and within communities, and between subnational authorities and the central state.
Fragmentation breeds subnational violence by fracturing the social fabrics of different ethno-regional communities and local power centers which mediate tensions between subnational groups, and by exacerbating existing fissures and tensions through rekindling micro-differences within broader social groupings. In fragile subnational regions, such as Uganda's Rwenzori region, fragmentation as an elite strategy of power and control tends to breed violent competition between the central state and subnational authority structures. This is because the exercise of power, legitimacy, and access to critical resources become destabilized when subnational authorities try to resist central-state erosion of their power. Understanding this phenomenon is key to grasping the underlying causes of violence in the Rwenzori region from 2012 to 2016, as well as other similar conflict episodes in Uganda and across Africa.
This article, therefore, foregrounds the role of local fragmentation in fueling civil violence. We demonstrate this relationship through an indepth analysis of the violence which took place in Uganda's Rwenzori region between 2012 and 2016. This violence, which differs from previous armed struggles in the region, is situated within a broader frame of how the strategies of control used by central-state elites feed and intensify subnational power contestations and social tensions that build up into violence. We use a combination of informal field interviews, the authors' close knowledge of Uganda's broader national political landscape, and secondary sources including news reports, drawing on these sources to reconstruct empirical narratives and theoretical insights that augment a growing body of literature on the politics of violence in the Rwenzori region (see Reuss & Titeca 2017;Sseremba 2020;Syahuka-Muhindo & Titeca 2016;Tshimba 2020a). This article potentially opens avenues for thinking about future indepth research and policy measures on the region's peace and security landscape.
This empirical analysis covers controversies and contestations in three kingdom entities that presently constitute the Rwenzori Region: Obukama bwa Tooro, Obusinga bwa Rwenzururu, and Obudingiya bwa Bwamba. While the 2016 attack occurred on the Rwenzururu palace, it represented the climax of a wider phenomenon of violence across the region. The Rwenzururu and Bwamba are recent kingdom-creations. Within these kingdoms, new district and sub-district local governments were created, some controversially. The region's location at the border with the Democratic Republic of Congo (DRC) and its security implications (which is addressed in other works, e.g. Titeca & Vlassenroot 2012), is beyond the scope of this article. Instead, we stress the political motives behind the creation of new kingdoms and districts, how these considerations have affected intra-society and statesociety relations, and how they in turn have led to waves of low-scale violence culminating in the November 2016 deadly attack on the Rwenzururu kingdom palace in the town of Kasese.
The article proceeds in four sections. In the next section, we outline the puzzling features of the violence in Rwenzori. Section two covers the theoretical underpinnings of our analysis. Section three contextualizes local fragmentation, from colonial divide-and-rule to post-1993 kingdomization and districtization. Section four examines this dual process, showing that kingdomization and districtization ignite intra-regional contentions and demands for new subnational authority structures in otherwise cosmopolitan areas, leading to multiple conflicts and violent state responses. The concluding section addresses some theoretical and empirical implications.

Puzzling Civil Violence in the Rwenzori Region
The violence in the Rwenzori Region reached a climax in November 2016, when the Uganda People's Defense Forces (UPDF) and Uganda Police Force (UPF) jointly attacked the Rwenzururu kingdom palace in Kasese, killed almost one hundred people, and arrested the Omusinga. The incident raised national, regional, and global concerns. The Uganda government accused the Rwenzururu kingdom of breeding the violence, alleging that its palace hosted treasonable conspirators and an armed militia. Kingdom officials and local politicians denied the treason charges and blamed government for the worsening insecurity in the region, whose political-administrative districts-Ntoroko, Bundibugyo, Kabarole, Kamwenge, Kyegegwa, Kyenjojo, and Kasese, with an estimated three million people-variously suffered bouts of insecurity during this period. While the region experienced violent episodes and ethnicized political struggles from just before independence in 1962 until the early 1980s (Scorgie 2011;Titeca & Vlassenroot 2012;Reuss & Titeca 2017;Dornbos 2017), the round of violence analyzed here is quite different both in cause and form.
Violent attacks against communities and security installations broke out in the districts of Bundibugyo and Ntoroko in 2012, in Kabarole in 2014, and in Kabarole and Kasese in 2016. Between June and October 2012, clashes erupted between groups suspected to be aligned with the Tooro, Rwenzururu, and Bwamba kingdoms. The region's minority nationalities, the Basongora and the Banyabindi, on their part, have demanded their own autonomous kingdom statuses (KRC and RFPJ 2012), citing potential marginalization under the Rwenzururu kingdom. In July 2014, region-wide violence erupted in Bundibugyo, Kasese, and Ntoroko, after the installation of the Omudingiya (King) of Bwamba in a ceremony presided over by President Yoweri Museveni. During these attacks, "unidentified people," armed with machetes, spears, and arrows, attacked Kasese, Bundibugyo, and Ntoroko districts, killing more than 70 people ( This Rwenzori-region violence is puzzling for several reasons. First of all, it seems to have been systematic, not isolated or uncoordinated. Organization and planning seem to have preceded the attacks. There were accusations and counteraccusations between the government and the respective local kingdom institutions, indicating real or perceived prior planning and organization (Tshimba 2020b;Republic of Uganda 2016). Second, notwithstanding the apparent prior planning, these attacks did not evolve into a full-scale civil war, as it is technically defined (Sarkees & Wayman 2010). Nor did the attackers or the state link these attacks to preexisting armed groups, particularly the Allied Democratic Forces (ADF). Third, the media and the government associated the perpetrators with specific ethnolinguistic identities, although traditional/cultural leaders in the region denied responsibility (ACODE 2014;KRC 2016;Republic of Uganda 2016). Finally, while the region has experienced transnational armed conflicts involving ADF rebels and the National Army for the Liberation of Uganda (NALU), the waves of violence during the 2012-2016 period seemed to be unrelated to, and to differ from, the ADF or any other rebel group activities (Salehyan 2009;Titeca & Vlassenroot 2012;Tshimba 2020a).

Theoretical Framing
As a technology of rule, fragmentation is a deliberate political process of dividing ethno-regional groups within a country to erode subnational concentrations of power, resources, and legitimacy, hence weakening competing, coalesced, and united groups. Ethnic and regional concentrations and constellations of social power have the capacity to strengthen subnational authority structures. This renders local power centers capable of challenging the central state, and even threatening regime survival. This is especially true under conditions of authoritarian rule where decentralized governance is manipulated to benefit central-state elites (Aalen & Muriaas 2017).
Fragmentation involves, among other methods, foregrounding microdifferences within broad ethno-regional groupings and using these microdifferences to subdivide otherwise coalesced regions. This by no means assumes that heterogeneous subnational regions are not prone to conflicts which are not driven from the center or to localized and subnational instrumentalization of identity differences by rural elites. Rather, central elites have political incentives to fracture regions and groups to weaken their ability to contest central-state power. While fragmentation sows seeds of conflict by fracturing the social fabric that holds ethno-regional societies together, subnational weakness serves the political interests of central elites by precluding local resistance, providing opportunities for neopatrimonial domination and the extension of patron-client relationships (Green 2010).
Local fragmentation is a political survival strategy against coalesced groups that would constitute a challenge to central-state power holders. These groups can be socio-linguistic, such as the Temne and Mende in Sierra Leone; they may consist of shared historical and class consciousness, such as Americo-Liberians in Liberia and Creoles in Sierra Leone (Sesay, Ukeje & Gbla 2009); they may be religious, such as Mandigo Muslims in Liberia, the Kurds, Hindus vs. Muslims in India, or Jews in pre-Hitler Germany; they can be occupational and professional (such as armed forces) and/or organized as professional associations and trade unions; and, finally, they can be geographically concentrated, like Afro-Christians in pre-2005 Sudan (LeRiche & Arnold 2012) or Tamils in Sri Lanka. Groups may be transcendentally coalesced by combining their religious, regional, linguistic, class, professional, and occupational aspects. When coalesced groups challenge the central state, they threaten the status quo and become a legitimate target for demobilization. Geosocial group coalescence forces ruling elites to devise means of containing the group; apart from fragmentation of subnational authority structures, co-optation helps to tame threats posed by coalesced groups such as armed forces (Rwengabo 2013).
It follows that ethnicity-a people's claim to and feeling of a shared historical, sociocultural, ancestral, and geosocial identity "that is formed in extrinsic/intrinsic contexts and social interaction" (Baumann 2004:14)can be used to breed fissures within a subnational region where different ethnic and sub-ethnic groups compete against each other for support and/or recognition from the central state, exercise decentralized subnational power, and have conflicts over resources at subnational levels. Ethnic identity can also incubate group coalescence and the ability to resist the central state, or the dominating ethnic group. For instance, in Uganda the Bakonzo and Bamba once resisted domination by the Batooro, which led to the eruption of the Rwenzururu Liberation Movement, the result of which was the fragmentation of Tooro district and the creation of Kasese and Bundibugyo districts, along with further demands for recognition of new cultural institutions (Doornbos 2017).
Spatially concentrated and historically marginalized minority ethnic groups may hope to benefit from decentralized governance, such as Uganda's devolution, which theoretically allows the design and implementation of policies and programs that are germane to their conditions and needs (Tranchant 2008). Yet ethnic-based decentralization "provides no permanent resolution of ethnic conflict" but rather serves to satisfy one group or generation of leaders, while evoking further ethnic and sub-ethnic demands and appeals to grievances that a given generation of leaders may have previously ignored (Larmour 1990:24). It can also evoke pressures to incorporate traditional and other cultural leaders and customary rules into the structure of local government. What is more, it can ignite resistance against ethnic hegemony, as the Rwenzori region historically demonstrates. These possibilities create incentives for central-state elites to fragment subnational regions and districts along ethnic, sub-ethnic and geographical lines, to render an otherwise assertive regional group less problematic for the center. In some circumstances, political actors use ethnic markers, while in other circumstances religion or class matters (Larmour 1990:26). The relative political salience of ethnicity may explain why ethnopolitics permeates countries such as Kenya and Nigeria but not Tanzania.
Fragmentation of territory and authority is an exercise of the tactic of divide-and-rule, but it differs from it in several ways. First of all, divide-andrule is generally considered to be foreign-power driven in an occupied country, historically associated more with imperial-colonial strategy than with domestic politics. Also, divide-and-rule is mainly based on clear-cut differences, such as sociolinguistic identity or religion, with limited appeal to micro-differences within geosocial spaces because the dividing power may be incognizant of intra-society micro-differences. Divide-and-rule is often impervious to a nation-building project. But fragmentation is counterintuitive; the dividers are concurrently rhetorical promoters of national unity, anti-sectarianism, and patriotic consciousness. The craftsmen of subnational disintegration create a façade of subnational development whereby fragmentation processes are legitimized through legal and political frames such as decentralization. Decentralization is presented as a constitutional and local-governance reform effort, but in reality it may involve such elements as splitting hitherto powerful and coalesced ethnic and/or regional groups; constructing and promoting competing social groups; exacerbating pre-existing intra-and inter-group differences within a given region or district; fragmenting existing authority structures; decentralization without transfer of power (decentralization in form, not in substance); and reversal of processes that might implant shared and collective national consciousness in heterogeneous countries.
The above-outlined political manipulation is typical of, though not unique to, Uganda's ruling National Resistance Movement (NRM) and incumbent President Museveni, who seek to frustrate or scatter subnational political coalescence among different regional-political groups. In recent years, this strategy has taken two interrelated forms, beginning with the creation of new districts as local government administrative units, a process called districtization. Known as District Local Governments (DLGs), the units are composed of elected executive leaders and councils operating under the decentralization system of government (Republic of Uganda 1995. Contrary to the promise that decentralization was an exercise in democratization and good governance, in practice it serves to extend neo-patrimonial relations between central-state elites and local political actors, both as extensions of institutional multiplicity and as opportunities for fragmenting otherwise powerful centers of power and geosocial-political influence (Makara 2018;Awortwi 2010;Golooba-Mutebi & Hickey 2016). The central government has tended to interfere with local governments, for example retaking initially devolved functions and powers. Second, what we call kingdomization involves the restoration of kingdom entities which had been abolished in 1967, in addition to the creation of hitherto nonexistent kingdoms and other traditional cultural institutions, some being breakaways from the old kingdoms.
Both districtization and kingdomization run counter to social cohesion and are antithetical to nation building. On the one hand, districtization engenders demands for ethnic-based districts and other lower-level local government structures. These demands become inexorable and engender an endless cycle where one concession triggers the next demand. Sociolinguistic attachments, fluid social customs and norms, and locally crafted claims generate a multiplicity of sub-identities that are used as the basis for demand of often unviable local government units. On the other hand, kingdomization opens up avenues for demanding new kingdoms by groups living under old kingdoms or established chiefdoms. Unity in older kingdoms is raptured by new sub-identity-based breakaways from pre-existing kingdoms. This breeds what Arlene Davilla (1997) calls "sponsored identities," in which "local-level cultural politics" is rationalized and cultural institutions claim to play an important role in shaping definitions of collective identity and local governance. Taken together, the dual processes of districtization and kingdomization intensify ethnic and sub-ethnic consciousness, self-assertion, and competition. The resulting intra-and inter-group struggle to control subnational authority structures, together with evocation of historical memories, breed intense ethno-political assertiveness and ultimately violence.
Fragmentation arouses intense competition. It fuels the new cultural politics of difference which challenge universal and mainstream conceptions such as modern-day nation-building and nationality and their legitimizing power, in preference for difference, diversity, multiplicity, group insularity to the exclusion of others, and specificity for the self (West 1990). Once local fragmentation starts, it acquires path dependency features; some groups seek to separate from other groups under whose authority they may have hitherto existed. Others resist belonging to another group's authority. The process also evokes memories of inter-group tensions that may have waned over time or been rendered ineffectual in a regional coalescence process. Competition and memorial revival give rise to conflicts between groups, touched off by the government's readiness to assuage these interests with offers of a new district or kingdom, both of which come with prospects of access to material resources from the central state. The resulting ethnic assertiveness is linked to civil violence within a region in three ways.
First, fragmentation revivifies groups, creates sub-group consciousness within coalesced groups, and leads to the assertion of sub-group interests as against other groups. Group and sub-group interests clash, leading to ethnic and sub-ethnic tensions which can end in violent conflict. Second, the process re-ignites the underlying tensions between groups, some of which may have waned. Old concerns, for instance about forced re-eviction of Basongora pastoralists from DRC who had been evicted from Uganda during colonial demarcations of Queen Elizabeth National Park become resurrected. Group antipathies erupt when identity fissures are re-engineered, regardless of the actors' intentions.
Finally, national fragmentation erodes trust in the central state's nationbuilding promise and rekindles cultural politics of identity. The state inadvertently abdicates responsibility to create nation-ness and tasks so-called traditional cultural institutions with acculturation in a heterogeneous country. Consociational theory posits that democratic stability is possible in otherwise deeply plural societies because contending subcultures and subnational entities can devise coalitional, cooperative, and mutually rewarding arrangements (Andeweg 2000;Dix 1980;Barry 1975). But a different outcome plays out under conditions of weak central state institutions, selfish interests of central-state elites, and ongoing contestations over a unifying national project. Here, identity and sub-identity acculturation is likely to grow as national consciousness falters. Competing cultural values and ethos may resurface under new districts or kingdoms, gnawing at the marrow of centralstate legitimacy, reviving old forms of identity politics and self-negating politics of state destruction (Horowitz 1985;Kabwegyere 1995). In the next section, we contextualize Uganda's fragmentation before turning to its manifestation and the resulting social violence.

Local Fragmentation: Colonial Antecedents and Postcolonial Deepening in Uganda
Many Africa countries have witnessed the proliferation of subnational administrative units in a process symptomatic of national fragmentation, some created under the guise of decentralization, others intended to assuage subnational political interests and demands. In Uganda, District Local Governments (DLGs) increased from 33 in 1986 to 135 in 2020. Despite criticism, government continues to establish more DLGs and kingdoms, with various justifications (Ayeko-Kummeh 2014; Olowu & Wunsch 2004). This trend has been ably addressed in the existing literature (e.g., Lewis 2014; Awortwi & Helmsing 2014;Grossman & Lewis 2014;Green 2010Green , 2008Muhumuza 2008). This article contributes to this literature in several ways, first by analyzing the way that social violence can be linked not only to decentralization but also to the proliferation of localized traditional structures of power. Second, the article shows how a strategy of fragmentation creates conditions for both subnational and national insecurity, when the interests of local and central-state actors clash, thus undermining overall nation-building and political stability. Empirically, a detailed analysis of the Rwenzori region helps inform our understanding of broader trends of fragmentation and conflict outcomes in Uganda and beyond, and the theoretical and empirical insights illuminate the complicated and contested state-society relations writ large.

Colonial Maneuvers
From the 1890s to 1962, British colonialism in Uganda relied on a "divideand-rule" approach, which purposefully implanted ethnic prejudices and prevented colonized peoples from coalescing against colonial conquest. Simultaneously, colonialism brought disparate socio-political groups into a single colonial polity in which Buganda kingdom enjoyed semi-autonomous status while other parts of the polity were under direct control of British governors and district commissioners who represented the colonial state. Alongside divide-and-rule, disproportionate overrepresentation of minority ethno-regional groups into armed forces rendered Buganda and other relatively coalesced kingdom-areas incapable of violently resisting the colonial power (Olowu & Wunsch 2004). During the 1950s, Andrew Cohen, whose arrival as governor in 1952 "coincided with the development of nationalism and political parties," brought new policies that countered demands for a unitary state (Pratt 1961:160). In 1953, it was announced that representation for Africans in the Legislative Council (LEGCO), which had been a preserve of minority Caucasians, would be increased "to provide an institutional means of achieving national unity" (Engholm 1962:16). The Baganda (people of Buganda kingdom) had to this point enjoyed preferential treatment. The LEGCO elections of 1956 were initially only held in Buganda, ostensibly to serve as an example for the rest of Uganda (Kiwanuka 1970). Despite criticism, LEGCO reforms were welcomed as epitomizing "tolerance and combined effort" (Ingham 1963:39).
The seeds of anti-Buganda resentment were sown by perpetuating ethnoregional discrepancies between Buganda and the rest of Uganda. Governor Cohen failed to reverse these long-entrenched discrepancies and divisive structures. Buganda acquired federal status under the 1962 independence constitution because pre-1962 developments made a federal constitution inevitable (Dinwiddy 1981). Independence followed amid ethno-regional inequalities and "contradictory ethnic bases" of power (Mutiibwa 1992:24). Districts had been created along ethnic lines, at least in British eyes, sowing seeds of ethnic consciousness, for instance in Acholi, Karamoja, and Teso (Apter 1959). The British saw African ethnolinguistic identities as pristine and unchanging, which meant that new districts were very rarely created (Green 2008:439). Postcolonial districtization maintained and followed sociolinguistic identities and micro-identities, but the fragmentation strategy would change the dynamics in later years.
Buganda was a federal state within a unitary country. So, the president and kabaka of Buganda, Edward Muteesa II, and the prime minister, Milton Obote, clashed. Buganda's interests clashed with those of Uganda. Maintaining a Buganda quasi-state within a state became untenable. Thus, Obote ousted Muteesa in 1966, abolished Buganda and other kingdoms in hasty changes which were constitutionalized in 1967, and Obote assumed the position of president. Idi Amin, however, overthrew Obote in 1971, reintroduced ten provincial governments in 1974 under military governors, and increased the number of districts to 37 (Jørgensen 1981:309). After Amin was overthrown in 1979, the Uganda National Liberation Front (UNLF) government (1979)(1980) reduced the number of districts to 33 (Green 2008:439).
Loyalty to subnational identities meant that political behavior reflected linguistic and identity interests with limited national convergence (Mudoola 1996:92). The 1962 constitution had granted federal status to Buganda, semifederal relationship to other kingdoms, and district status (unitary government) to non-kingdom areas in a "parcel of contradictions" (Mutiibwa 1992:25;Dinwiddy 1981:514). Accordingly, Buganda "developed an attitude of complacent arrogance toward other people of Uganda" (Sathyamurthy 1972(Sathyamurthy :2123. The resulting "suspicion and hostility" precluded national unity (Glenworth & Hancock 1973:240), as Buganda's separatist tendencies persisted. The resulting tensions engendered crises between 1962 and 1967, resulting in two decades  of political instability punctuated by Idi Amin's military dictatorship. When Museveni took over in 1986, he promised a unified Uganda with non-sectarian politics.

Post-1986 Musevenic National Fragmentation: Districtization and Kindomization
After Museveni came to power in January 1986, six new districts were created within six years, increasing the number from thirty-three to thirty-nine by 1991 and to forty-five in 1997. In 2007, the number increased again to 80, then to 115 in 2016, 128 in 2018, and 135 in 2020. Kingdoms and other precolonial traditional authority structures, which had been abolished in 1967, were restored starting in 1993. New ones were created by fragmenting the old and/or forging them where they had never previously existed. This dual process of fragmentation eroded the power and resource base of subnational units. The hitherto powerful districts of Bushenyi, Mbale, Luwero, Gulu, Arua, and Soroti were ghostified. Buganda was fragmented to create the chiefdoms of Bunyala, Kooki, and Buruuli. These structures internally compete with the Mengo-based Kabakaship.
A 1987 Commission of Inquiry into the Local Government System saw no value in creating more districts and observed that Amin had used districtization to evade responsibility for service-delivery failures (Green 2008:440). The Commission advised against further districtization, and called for a reversal of the process, which "would undoubtedly result in a large number of the newly created districts losing their existing status" (Republic of Uganda 1987:117-23). Museveni had inherited a dilapidated polity and economy-serious resource gaps warranted caution about costly public administration-and his regime, which followed outright rebel victory, was better placed to resist further districtization.
Uganda's political history was reflected in these latter-day machinations. By the late 1800s, the area now called Uganda consisted of different kingdoms, chieftaincies, and acephalous communities at different levels of sociopolitical sophistication, each with its own leadership system, whether hierarchical or non-hierarchical. In 1900, an Agreement between the British and Buganda kingdom was signed. As part of colonial subterfuge, the British moved to weaken other kingdoms, especially Bunyoro-Kitara, which militarily resisted colonial conquest. Formal agreements with Bunyoro-Kitara kingdom were signed in 1933 and 1955 (Briggs 1998). In other areas of Uganda, "governance was highly decentralized, and power lay in the hands of elders who practiced a form of democratic government" based on precolonial political institutions (Quinn 2014:37).
This was the situation in 1967, when Milton Obote abolished the kingdoms, banishing states that had existed as well-organized and sovereign entities before colonial incursions. Obote sought to reduce competition between the central state and traditional authorities (Briggs 1998). After Museveni assumed power in 1986, the Baganda demanded rediscovery of a past torn between precolonial evolution, colonial favoritism, and post-colonial crisis. On April 3, 1992, Museveni persuaded the National Resistance Army (NRA) Council sitting in Gulu to allow the return of traditional sites to traditional authorities, "provided that this does not interfere with the security of the country" (Republic of Uganda 1993).
Hoping that kingdoms would confine themselves to cultural functions (Kayunga 2001), Article 246 of the 1995 Constitution allowed "the institution of traditional leader or cultural leader" to exist "in any area of Uganda in accordance with the culture, customs and traditions or wishes and aspirations of the people to whom it applies." It called for resolution of "the issue of traditional or cultural leader" by the "community concerned using a method prescribed by Parliament." The 2006 Uganda National Cultural Institutions Policy recognizes that communities' traditional or cultural institutionskingdoms, chiefdoms, clans, and the family-provide identity to their communities, support culture, and mobilize people for development. The 2011 Institution of Traditional or Cultural Leaders Act operationalized Article 246 and the 2006 policy (Republic of Uganda 1995Uganda , 2006Uganda , 2011. In addition to pre-existing kingdoms and chiefdoms, Museveni oversaw the creation of new local divisions which had previously never existed: Acholi, Tieng Adhola, Alur, Buruuli, Lango, Kooki, and Teso chiefdoms. Rwenzururu and Bwamba kingdoms were recently recognized in the Rwenzori region. A UPDF officer, Col. Martin Kamya, was installed as the king/Omudingiya of Bwamba. Buruli, Kooki, and Bunyala were created from fragments of Buganda Kingdom. Busongora and Bunyabindi, not yet recognized, are possible future fragments out of Rwenzururu. This kingdomization phenomenon is not without controversies. In Obugabe bwa Nkore (Nkore kingdom), Museveni's home area, Prince John Barigye, heir to the throne, was crowned Omugabe on November 20, 1993, but Museveni "nullified" the appointment. Nkore kingship remains contentious because its people, the Banyankore, had been divided between pastoralists and cultivators in an evolving ethnic hierarchy that allowed minority pastoralists to dominate majority cultivators (Doornbos 2001).

Contentious Fragmentation and the Making of Conflict in the Rwenzori Region
Kingdomization and districtization are central to the tensions and violence in the Rwenzori region. This is a region which had experienced struggles against domination for many years (Syahuka-Muhindo, 1991). The revival of kingdoms in 1993 was a central-government process. The constitutionalization of kingdoms in 1995, the 2006 policy, and the 2011 law, demonstrated that central-state elites had a vested interest in this process. This reopened old wounds and fueled fresh ones. Associating kingdoms and districts with specific ethnic markers engenders conflicts by driving power contests and social tensions. This typified the 2009 clashes involving the Banyala and Baganda in central Uganda; tensions that followed the creation of Rwenzururu and Bwamba kingdoms; refusal to restore Nkore kingdom; and fractures within Buganda kingdom.
Since its recognition in 2009, the Obusinga bwa Rwenzururu-a fragment from Tooro kingdom-became a center of contention around which social tensions and conflicts built up. Different ethnicities opposed the monoethnic (Bakonzo) kingdom and demanded to be allowed to secede (Ayaa 2012). The creation of Rwenzururu kingdom bred new anti-Bakonzo/Rwenzururu tensions. In non-Bakonzo eyes, the kingdom was akin to the old Tooro-like subjugation against which the Bakonzo had fought. As tensions mounted, the government further fragmented the initial Rwenzururu kingdom by tacitly sponsoring or at a minimum supporting other kingdom creations aimed not at managing intra-Rwenzururu tensions but in the service of fragmentation motives.
Concurrently, intra-RLM factions adopted different positions. One RLM wing, led by Isaaya Mukirania, declared independence and sought UN recognition as a sovereign state in 1962. Other factions, identified with Kawamara and Mupalya, insisted on a negotiated settlement where the kingdom could remain part of the newly independent Uganda. In 1974, Idi Amin carved Rwenzori and Semuliki districts out of Tooro to mollify these groups. This move fragmented the old Tooro into new Tooro (which was Kabarole district but now included Kabarole, Kamwenge, Kyenjojo, Kyegegwa, and Bunyangabo districts), Rwenzori, and Semuliki districts. The Bakonzo and Bamba respectively administered their two new districts.
Thanks to politicians from the region, especially Amon Bazira (who led NALU until 1993), RLM fighters, under the now-Omusinga Mumbere, Mukirania's eldest son, laid down their arms on August 15, 1982(RLP 2014Tshimba 2020a:138). Earlier in 1979, President Godfrey Binaisa renamed Rwenzori and Semuliki districts as Kasese and Bundibugyo (the latter now divided into Bundibugyo and Ntoroko) districts respectively (KRC 2014). For the most part, the region's people, despite their ethnic differences, remained peaceful and combined efforts against the ADF rebellion (1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005). But RLM veterans resumed their demands during the Musevenica kingdomization era, given the incentive structure created by the new politics of fragmentation. In April 2005, a Ministerial Committee chaired by Muganwa Kajura recommended recognition of the Rwenzururu in accord with Article 246 (1) of the Constitution (Republic of Uganda 2005). While this assuaged an age-old RLM struggle, it deepened political contestations and social tensions, giving way to subnational social violence to which we turn.

Intense Competition: Toward Violent Conflicts
Recognition of Rwenzururu as a kingdom did not erase the earlier Mukiranialed struggle to declare independence in 1962. When the government installed Mumbere as the omusinga/king of Rwenzururu on October 19, 2009, the Bamba and Babwisi in Bundibugyo resisted, and the Basongora and Banyabindi demanded their own kingdoms. On July 3, 2012, the Basongora privately crowned Rwigi IV Rutakirwa Agutamba-Kabumba Ivan Bwebale as their king (Kajubu 2012). The government did not expedite the kingdomization process for the Basongora, Bamba, Babwisi, and Banyabindi, and other potential groups, possibly in order to buy time while handling the reactions of different groups to the post-Rwenzururu developments. In Kasese district, the Basongora and Banyabindi claim that the post-2005 kingdomization placed them under further Rwenzururu marginalization (CCFU 2014;Thembo 2014).
In Bundibugyo, the Bakonzo are a minority against a majority Bamba-Babwisi. They claim marginalization under the Bwamba kingdom that broke away from the Rwenzururu kingdom. These claims and counterclaims had almost disappeared, or at least significantly subsided, before the Rwenzururu kingdom was formally established in 2009. The post-1993 fragmentation, through runaway districtization and kingdomization, is a central driver of sociopolitical contests and a contributor to the social violence in the Rwenzori region because it weakened the glue that had hitherto unified different sociolinguistic communities. While the politics of fragmentation dates back to colonial rule and immediate post-independence politics, the manner in which it became central to the technology of rule and control under Museveni's reign contributed to incubating and fueling subnational social violence. During a relatively short period of time following the recognition of the Rwenzururu and Bwamba kingdoms, there were spates of violent confrontations and attacks across the Rwenzori sub-region in a manner that was truly unprecedented.
Clashes erupted in Bundibugyo in June 2014 and spilled over to Ntoroko and Kasese districts, as armed men attacked residents with guns, machetes, and bows and arrows, killing up to 98 people, including five UPDF soldiers and five police officers. In Bundibugyo, they attacked Kicho Police Post, Kanyamwirima military barracks, and the Bwamba kingdom palace (The New Vision 2014b). In Ntoroko, another armed group tried to attack Karugutu police station, but the UPDF intercepted them. Further afield in Kasese, attackers killed civilians in Ibuga division, Kasese municipality, and burned several houses. Another group attacked and killed six people, including a UPDF soldier and his three children in the Bigando Division. Gunmen killed two police officers guarding the Uganda National Roads Authority (UNRA) weighbridge at Katunguru along the Kasese-Mbarara highway, commandeering twenty-two guns. In response, police arrested over 130 suspects; they charged them with treason and murder in the military Court Martial and recovered sixteen of the guns the attackers had taken.
The state's coercive machinery deployed heavily around Lake Edward, Mount Rwenzori, and Queen Elizabeth National Park. The government set up two reception centers in Kasese municipality, at Kasese Council Hall (Nyakasanga), and in Bundibugyo at Buhundu Primary School, to receive surrendering attackers. About 214 reportedly surrendered (The New Vision 2014b). This suggests that many attackers had not been sufficiently socialized in the underlying motives and mission for such violence or were sporadic participants mobilized and deployed by elites seeking negotiation through violence. A group claiming links to Omusinga's royal family, led by Christine Nyamukama, moved around Ntoroko, Bundibugyo, and the highland subcounties of Kasese district on a supposed peace-building mission. They held meetings in Karugutu, Ntoroko, but they were shunned throughout Bundibugyo except at Kasitu.

Government-Rwenzururu Clashes
Control over a kingdom comes with access to Museveni's patronage network with accompanying benefits to the kingdom elites. The Rwenzururu kingdom was recognized toward the 2011 general elections in a calculated move to swing Kasese district from the opposition Forum for Democratic Change (FDC) to Museveni and his NRM. Still, the election result did not go Museveni's/NRM's way. Subsequently, Museveni allowed the Bwamba kingdom to secede from Rwenzururu on May 23, 2014, which led to conflicts and intensified the demands of the Busongora and Banyabindia for their own kingdoms. The Banyabindi kingship, apparently located in north Kasese district and headed by Isebantu (king) Elisa Mugisa-Entare, arose after a vain attempt by the Basongora and Banyabindi to oppose state recognition of Rwenzururu. The recognition of Rwenzururu kingdom did not take into account the local apprehensions against ethnicized kingdomization, the divisions that would be set in motion, and the possible political-security cost. As was typical of the government's fragmentation logics, the creation of neither Rwenzururu nor Bwamba, nor of any other new kingdom had been preceded by a referendum in the area where it had been created to have cross-cutting popular voice and local consensus; there was no regard for views and demands which cut across ethnic and subethnic groups.
In a downward, inexorable spiral, the creation and recognition of new kingdoms in the Rwenzori sub-region weakened extant institutions and further fragmented the social terrain and group identities in the area. When the Rwenzururu kingdom sought to push back against the proliferation of competing kingdoms which chipped away at its power and territorial control, it clashed with the central government, coming to a head in the deadly encounter that took place in November 2016.
Bottom of Form On July 6, 2012, a group of Bamba and Babwisi had rioted in Bundibugyo, protesting an earlier visit of King Mumbere to Bundibugyo on June 30, 2012, to mark the RLM's 50th anniversary. Tensions continued to increase, beginning when Rwenzururu officials announced plans to hold the anniversary celebration, which Uganda's then-prime minister, Amama Mbabazi, attended. As part of the celebrations, the Omusinga erected a shrine in Kirindi village in Bundibugyo district (Asinja 2012). The government "advised" Omusinga Mumbere to restrict his visits to Bundibugyo, a place he claims to be his original birthplace and part of his kingdom. At this point, the government realized that he was consolidating his power over a vast geopolitical space; he was based in Kasese but claimed birthright and loyal subjects in Bundibugyo and other districts, an area covering the largest part of Rwenzori region.
The conflict situation in Rwenzori also had a districtization dimension. In Bundibugyo, the Bakonzo are minority, and yet they were accused of violence in 2014. Toward the 2016 elections, hoping to secure Bakonzo vote, Museveni promised a Bakonzo-dominated district, broken away from Bundibugyo, made up of Bughendera County (The Observer 2016b). Omusinga Mumbere's brother and long-time opposition politician, Christopher Mbalibulha Kibanzanga, quite dramatically defected from the opposition FDC and joined Museveni's ruling NRM. He successfully contested in Bughendera County with Museveni's support (The Observer 2016a). The promise of a Bakonzo-dominated district and Kibanzanga's victory heightened tensions in Bundibugyo. The majority Bamba and Babwisi viewed the Bakonzo minority's self-assertion in Bundibugyo as an expansion of the Rwenzururu kingdom's influence beyond Kasese. Museveni's chameleon-like collusions and counter-collusions seemed to support this process; the promise of a new district and support to Kibanzanga were political tactics in the heat of campaigns intended to benefit Museveni and his ruling NRM. But this heightened existing tensions, leading to violent clashes in Bughendera, where thirty people were killed, hundreds of houses were razed, and thousands of residents were displaced. Here, the elite-driven fragmentation bred bloodshed.
Earlier, in 2010, the Kasese District Council controversially resolved to split Kasese to create Lhubiriha and Rwenzori districts. Area MPs, a majority of them in opposition, resisted the proposal. While meeting with district leaders, Museveni later proposed to split Kasese into four more districts, possibly to forestall the growing combined strength of both the large district and kingdom. Six MPs from Kasese opposed a 2012 motion (which was tabled by the cabinet) in parliament on the split (The New Vision 2016). Museveni may have hoped that splitting Kasese district would open up electoral advantages to the NRM in local electoral races, weaken Kasese's collective voice in national politics, enfeeble the kingdom of Rwenzururu, and facilitate the extension of his patronage reach.
But in 2016, the opposition-dominated District Council revoked the 2010 Council Resolution to split Kasese. In August 2018, district NRM party leaders reignited fragmentation calls, reinforcing the government's insistence on fragmenting Kasese district and reflecting the overall modus operandus of the political establishment at the national and local levels. King Mumbere, on his part, "expressed dissatisfaction" with the leaders seeking to split Kasese district (Wambuzi 2018). On November 1, 2018, councilors from the district petitioned the Speaker of Parliament, opposing the proposed split of Kasese into Bwera, Mubuku, and Nyamugasani districts. They alleged that the 2010 resolution had been reached in bad faith, had ignored citizens' wishes, and had been made in error by flouting Local Government Councils' Rules of Procedure (Nile Post 2018).
The plan to divide Kasese district-which is a demographic, geographic, strategic, and resource-endowment heartland of Obusingwa bwa Rwenzururu-was part of the broader and ongoing politics of fragmentation. Divided along ethnic lines (catering to the Basongora and Banyabindi groups, for instance), it would weaken and reduce Obusinga's subjects and erode their natural resource base. Fragmenting Kasese would be justified under decentralized local governance and ongoing districtization.
Yet, beyond widening and deepening Museveni's "electoral patronage" (Green 2008:442), the "creation of LG jurisdictions in Uganda neither conforms to the policy objective of bringing services closer to the people nor to promoting participatory democratic governance" (Awortwi & Helmsing 2014:766). Instead, Museveni's patrimonialism thrives on fragmentation where the ethnicization of districts and kingdoms negates genuine devolution of power to local governments, durable nation building, and the extension of services to the people. The fragmentation strategy crystalizes identity and majority-minority tensions in given areas, diverts local elites from centralstate failures to local squabbles, and renders the otherwise ebbing fires of communal tensions explosive (Green 2010;Reuss & Titeca 2017). What happened in Kasese in November 2016 is emblematic of how the fragmentation strategy fuels mounting tensions and animosities that build up into a deadly climax, to which we turn next.

The Obusinga Palace Attack
Between Saturday, November 26, and Sunday, November 27, 2016, the UPDF, under the then-Second Division Commander Major-General Peter Elwelu, together with the police, raided the Buhikira Royal Palace of the Obusinga bwa Rwenzururu in Kasese municipality. The Omusinga and about 150 subjects (the police report claims 139) were arrested and detained in Kasese on charges of murder, terrorism, and other transgressions. The Omusinga was thereafter airlifted to Kampala and detained by the Special Investigations Unit (SIU) at the infamous Nalufenya police station (Daily Monitor 2017; Tshimba 2020b). He was later released on bail, but his movements remained restricted to the Kampala metropolitan area.
After the attack, fifty-two bodies remained unclaimed. The military buried them at Kihara army barracks in Kasese Municipality. Many details, including the death toll, remain shrouded in mystery. More than 100 people are believed to have died, but the police claimed that 46 Royal Guards were killed and 139 others arrested. Access to the area for journalists and investigators was blocked, and families were not permitted to collect the bodies of their relatives. One year after the attack, the palace remained a crime scene, still cordoned off. The palace attack occurred in 2016, exactly 50 years after the 1966 attack on Buganda palace, but unlike the 1966 event, it epitomized the then-unfolding attacks and counterattacks between state security agencies and individuals who claimed to belong to the Bakonzo ethnic group (HRW 2016).
On November 29, 2016, the Minister of Internal Affairs, General Jeje Odongo, claimed that the raid on the palace had taken place in response to "attacks conducted by cultural guards 'Kirumira Mutima' said to be his [Mumbere's] creation and under his command." Kirumira Mutima (loosely translated: "the hard-hearted") was a suspected secessionist militia seeking to recreate the Yira Republic stretching into the DRC (Tshimba 2020b; Doornbos 2017), but the Rwenzururu kingdom officials vehemently denied such motives. General Odongo listed attacks between 2014 and 2016, stating that on November 26, 2016, the group had attacked security forces in Kasese town, and thereafter conducted multiple and simultaneous attacks in six other places in the district, where they killed sixteen police officers and took six guns.
General Odongo mentioned multiple causes of the conflict: secession/ Yira Republic claims, Rwenzururu-Bwamba tensions over jurisdictional and control issues, land/livelihood conflicts, and influence of selfish politicians. Save for secession claims, the other "causes" speak to the argument that fragmentation breeds tensions over space, power, influence, and access to patronage opportunities. Odongo claimed that the Kirumira Mutima militia had established camps in Kamabale village, Kabarole district; Ihandiro subcounty, Kasese district, and Kakibuta and Kakimara, both in Kasese district; and that a regional security meeting, held on November 21, 2016, in Fort Portal, Kabarole district, had resolved to dismantle these camps (Republic of Uganda 2016). Area leaders and international organizations called for an independent investigation of these killings (Daily Nation 2016; UN 2016; HRW 2017), some of whose bodies remained unclaimed (URN 2017), but no known progress has since been made.
The conflict situation in the Rwenzori region, as it unfolded from 2012, and its culmination in the deadly assault on the Rwenzururu palace in Kasese in November 2016 was by no means an isolated case. The NRM policy and practice of subnational fragmentation produced conflict flare-ups and violent confrontations elsewhere in Uganda as well. In central Uganda, the Buganda riots of September 2009 were related to recognition of Banyala kingship in Bugerere, which Buganda kingdom claims as its territory. There were territorial disputes over splitting Tororo district, pitting the Ateso against the Japadhola ethnic groups. Additionally, there were controversies over district boundary demarcations between Acholi and West Nile communities. All these controversies and conflicts have resulted in deadly violence, thus undermining national peace and stability and curtailing the nationbuilding project.

Conclusions: Theoretic and Practical Implications
Violence in Rwenzori region was symptomatic of the instrumental fragmentation which hatches self-assertion by groups envisaging potential power and access to central state patronage. What appears as ethnic emancipation, through creation of new districts and kingdoms split from preexisting ones, results in ethnic resurgence thus escalating ethnic contestations (Sseremba 2020). The Rwenzori experience shows that the politics of fragmentation can breed multilevel threats to national security and social harmony as ethnic violence does not remain isolated in specific subnational locales but tends to spread across spaces. To retain control over mushrooming authority structures, particularly the presumed traditional institutions, there are contestations between control (by the central state) and self-assertion (by intersubnational authorities) which incubate conditions that eventually lead to deadly clashes. Years of center-driven fracturing sowed seeds of conflicts, the violent evolution of which culminated in the deadly assault on the Rwenzururu palace in November 2016.
These events are mainly a consequence of the deepening politics of fragmentation. If disagreements and conflicts between kingdom entities instigated attacks and counter-attacks, as the government claimed, then kingdomization bred conditions for violence. Rwenzururu kingdom had conflicts with Tooro kingdom over territorial jurisdiction and claims of historical ethnic marginalization. The government ought to have anticipated similar conflicts between Rwenzururu and Bwamba on precisely similar grounds. The events in Rwenzori-and indeed across Uganda-show that subnational, territorial conflicts and fragmented subnational entities not only lead to local social violence but can also threaten national unity and security.
Kingdom, chiefdom, and district clamors in Rwenzori and other parts of Uganda show that districtization and kingdomization provide the structural and political context in which competing demands and mobilization unfold and deepen. Thus, the violence in the region between 2012 and 2016 shows that the divisive strategies of the ruling elites eroded intra-and inter-communal peace, fractured social harmony, widened existing fissures, and rekindled micro-differences (Daily Monitor 2016). The resulting group antagonisms threaten local security, undermine nation building, and expose the ineptitude of national elites with regard to guaranteeing national security. Subnational fallouts provided ample justification for central-state actors to deploy excessive force. It is instructive that the attack on the Rwenzururu palace occurred 50 years after the Buganda palace attack, underlining the state's failure to coexist with-or erase-traditional authority structures while retaining the military as the central political interlocutor.
One broad implication of our analysis relates to the state-society dependency logic; the state threatens its foundational survival when it avoids societal engagements that enhance nation building and social harmony within its territorial domain. The state must survive in order to protect society. The state cannot survive when society collapses. But society must thrive for the state to survive. The state can collapse-Uganda nearly collapsed by 1986-but society survives. A fragmentary landscape in which any group can create an authority structure and/or threaten another underlies the looming crisis of national unity and state survival, because contestations between central-state control and subnational self-assertion breed multi-level political-security threats. Political elites seem not to grasp this state-society (inter)dependence logic or avoid it for political expediency. More research needs to explore the possibility of resolving standoffs between the state and fragmentary authority structures, and among and within these structures. The government's peacebuilding efforts in the region have come off as little more than symbolic (Muhumuza 2020), which is perhaps unsurprising, considering that the very fragmentation logic leading to these conflicts serves the interests of the ruling elites. Subnational violence is an unintended consequence of fragmentation, but since Museveni has always positioned himself as a security president, it is arguable that he relishes the political instrumental use of insecurity and instability to retain his stranglehold over state institutions.