Stabilization and Local Conflicts: Communal and Civil War in South Sudan

Abstract Scholars have long argued that local conflicts need to be integrated into the analysis of civil war and peacebuilding. Yet, systematic research of the linkages between communal violence and civil war is sparse. This contribution connects communal violence research to the stabilization and peacekeeping debate. To further a more systematic analysis of communal conflicts, I distinguish various types and their linkages to civil war and peacebuilding. In South Sudan, large-scale communal conflicts—communal wars—precede the country’s civil war and are likely to succeed it. Their protracted and fundamentally political nature means that they cannot be addressed as ‘local conflicts’ in isolation from national politics and state institutions. I argue that military force may temporarily stabilize a conflict zone but the horizontal linkages between urban and rural communal conflicts and their vertical linkages to national political processes require concerted efforts on the national and the sub-state level to avoid renewed conflict cycles and contribute to lasting stability.


Introduction
Peacekeepers face both civil war and communal conflicts in a number of countries they are deployed to. In all three UN missions on the African continent that include 'stabilization' in their title-in the Democratic Republic of the Congo (DRC), Mali, and the Central African Republic (CAR)-communal conflicts are recognized as a prime source of the instability that threatens civilians and undermines the political process. The death toll of South Sudan's armed conflicts has been estimated at almost 400,000 people since the civil war broke out in December 2013, of which about 190,000 are reckoned to have been directly killed in fighting (Checchi, Testa, Warsame, & Burns, 2018). There are no reports as to how many people were killed in civil war-related combat versus fighting by communal militias, but some observers estimate that communal conflicts may have killed as many people as fighting between government forces and rebel groups. 1 Since the signing of the Revitalized Agreement on the Resolution of the Conflict in the Republic of South Sudan (R-ARCSS) in September 2018, communal violence is again on the rise (IPI et al., 2019).
South Sudan is not the only country where communal conflicts may at times be deadlier than a parallel state-based conflict. A case in point is northern Nigeria, where fighting between herdsmen and farmers cost an estimated 1,300 people their lives in the first six months of 2018, which is six times the number who died in the Boko Haram insurgency during this period (International Crisis Group, 2018). Such deadly communal conflicts fought by well-organized militias represent one crucial dimension of sub-state or 'local' conflicts that peacekeepers are increasingly expected to address. The mandate of the United Nations Mission in South Sudan (UNMISS) explicitly demands mediating and reducing communal conflicts.
Scholars have long argued that local conflicts need to be integrated into the analysis of civil war and peacebuilding (Autesserre, 2010;Kalyvas, 2006). These insights have increasingly been incorporated into peacekeeping practice. The United Nations Department of Peacekeeping Operations (UN DPKO) has explored how to address local conflicts in the context of peacekeeping missions (O'Bryan, Rendtor-Smith, & Donati, 2017). Yet, to date, a more systematic understanding of local conflicts in the shadow of civil war is lacking. UN missions' Civil Affairs Units often employ ad hoc approaches without strong linkages to the broader and long-term political process (Gorur & Velituro, 2017). Although local conflicts can be significant drivers of violence against civilians, peacekeeping missions have been critiqued for treating them inconsistently and as a marginal issue. Since academic research has rarely explored how to build peace after large-scale communal conflict, there is a dearth of advice for peacekeepers.
Imprecise terminology further compounds the lack of systematic study. Scholars rightly advocate for more attention to be paid to local conflicts but often define 'the local' as everything below the national level, e.g. 'the level of the individual, the family, the clan, the district, the province and the ethnic group' (Autesserre, 2017a, p. 116). This somewhat artificial divide of the 'national' versus the 'local' impedes precise analysis of diverse local conflicts, their interaction with national processes, and tailored intervention strategies. Research would benefit from a more precise definition of local conflicts and a nuanced analysis of their dynamics, linkages to the state, impact, and legacies. Lumping communal conflicts into the broader category of 'local' conflict and ignoring the academic field of communal violence is unhelpful and unnecessarily entrenches academic siloing of peacekeeping, civil war, and communal conflict research.
This contribution connects research on communal violence to the stabilization and peacebuilding discourse. I understand communal conflict as non-state armed conflict between social groups that define themselves along identity lines, such as ethnicity, religion, language, and culture. Such conflicts often precede and succeed a civil war. During civil war, they may run parallel to or interact directly with fighting between government forces and rebel groups, even to the extent that for specific violent incidents the line between communal violence and civil war can be blurred, as the case of South Sudan will illustrate.
Communal conflicts can vary dramatically in scale. Many incidents kill in the double digits but some clashes result in hundreds or even thousands of dead, crossing the death toll of a civil war, as cases from Nigeria, Indonesia, South Sudan or the DRC have demonstrated. To further a more systematic analysis, I distinguish communal conflicts according to scale, type, geography, armed actors, and the national context. When such conflicts kill more than 1,000 victims per year I refer to them as 'communal wars' because such entrenched civilian mobilization and armament requires distinct peacekeeping, intervention, and peacebuilding strategies (Krause, 2018).
Because communal conflicts do not directly challenge the national government, they are often referred to as 'tribal', 'ethnic', and 'apolitical'. However, such conflicts are fought over local government control and political representation; land rights and use; and resources and access to economic opportunities. I argue that communal conflicts are fundamentally political and will demonstrate that such conflicts do not take place in isolation from national politics and state structures. I further argue that addressing this particular sub-set of local conflicts requires a thorough understanding of horizontal and vertical linkages between national and local power arrangements and electoral constellations; state and customary law; urban and rural conflicts; and resource governance on the national and the local level.
Overwhelming military force may achieve a temporal end to fighting in urban areas but beyond towns and cities, a military approach barely achieves stability or peace. In a country as vast as South Sudan, peacekeepers are inevitably spread thin. Urban and rural dynamics of communal violence often interlink both horizontally, between urban centres and periphery, and vertically, between major urban centres and rural areas, because of family and kin networks. Therefore, even urban 'stabilized' areas, such as the Protection of Civilian sites (PoCs) in South Sudan, continue to exist within areas affected by fighting. The legacies of previous atrocities and the continuation of rural violence often keep urban tensions stoked. Elite politics in urban centres can further undermine rural conflict resolution mechanisms and contribute to an escalation in fighting. Consequently, transforming temporary into long-term stability requires a nuanced understanding of the types of communal conflict and their interlinkages with national and regional conflict dynamics.
This contribution analyses communal violence in the context of civil war and its implications for stabilization and peacekeeping. It advocates usage of more precise terminology for analysing how such conflicts interlink with civil war and the national political process, which is imperative for understanding peacemaking in South Sudan (De Vries & Schomerus, 2017). My analysis draws on secondary literature and reports on the conflicts in South Sudan, complemented by interviews conducted with UNMISS and NGO staff in Juba, South Sudan's capital, during two research visits in November 2017 and April 2018.
The next section examines the stabilization discourse and implications for communal conflicts. I then survey the field of communal conflict and discuss dynamics of violence in more detail, drawing on a typology that distinguishes between scale, type of conflicts, geography, armed actors, and the national context. The third part focuses on the case of South Sudan to illustrates how communal conflicts interlink with civil war there and how the legacies of civilian armament have profound implications for the national political process, peacekeeping, and stability. In the conclusions, I reflect on the implications of this analysis.

Stabilization and Local Conflicts
Peacekeepers are increasingly responsible for maintaining peace locally. However, neither peacekeeping research nor practice systematically disaggregate the category of 'local conflicts'. 'Local' generally refers to sub-state dynamics, with little distinction between communal and other types of conflict. There is near consensus that local conflict dynamics do not necessarily mirror national ones in civil war (Kalyvas, 2006), that national peace does not automatically trickle down (Autesserre, 2014;Cheng, Goodhand, & Meehan, 2018;Manning, 2003), and that local conflict resolution is crucial to building sustainable peace (Autesserre, 2017a). Research has further shown that peacekeeping contributes to a reduction in violence when large contingents deploy under robust mandates (Di Salvatore & Ruggeri, 2017), but its effect on local conflicts is less clear. Peacekeepers may contain the deadliness of local clashes but do not prevent local violence (Ruggeri, Dorussen, & Gizelis, 2017), while international efforts at local peacebuilding may yield ambiguous results (Autesserre, 2017a).
Among policymakers, local conflicts are increasingly acknowledged as a significant cause of insecurity, with the capability to derail peace processes. Yet, within the UN system, this recognition comes more in response to the destructive nature of local conflicts and the threat they pose to civilians rather than because they are understood as an integral part of the larger and complex conflict landscape that peacekeeping operations are mandated to tackle. (O'Bryan et al., 2017) How stabilization can address local conflicts remains unclear, not least because the term 'has a wide range of interpretations' and 'usage of that term by the United Nations requires clarification' (UN, 2015). Mandates of the four UN missions that have stabilization in their name (Haiti; DRC; Mali; CAR), emphasize peacekeepers' robust use of force to protect civilians and remove non-state armed actors (Belloni and Moro, this collection). Stabilization prioritizes security over governance and democratization and is meant to be a short-term approach. Its primary security objective has generated criticism for ignoring the underlying sources of conflict and bottom-up community-driven resolution (Mac Ginty, 2012;Newman, 2013).
Elite bargains, i.e. discrete agreements that explicitly re-negotiate the distribution of power and allocation of resources between elites, play an integral role in the stabilization of conflict (Cheng et al., 2018). However, stabilization of the macro-conflict, based on such bargains, may not reduce forms of local, 'embedded' or 'permissive' violence, such as communal conflicts, criminal violence, and gender-based and domestic violence, which generally continue to expose populations to significant levels of harm (Cheng et al., 2018). These forms of local conflict do not only result in potentially high victim numbers but also harbour a reservoir of armed actors available for remobilization. Consequently, stabilization without concerted efforts of transforming elite bargains into national, as well as local, conflict resolution may not be sustainable.
The use of force against non-state actors in the DRC illustrates the dilemma of trying to build localized stability but running the risk of further undermining security in the longer term for the civilian population. In 2013, the UN mission in the DRC, MONUSCO, deployed the Force Intervention Brigade (FIB) to support the government and neutralize non-state armed groups. The objective was to turn targeted areas into 'islands of stability' and 'prevent an immediate relapse of the communities concerned into a cycle of violence after armed groups have freed an area' by applying an integrated multi-dimensional and multi-stakeholder response of at least six months (Vogel, 2014). Yet, the strategy resulted in vast 'swamps of insecurity' because other areas, armed groups, and the underlying power structures remained unaddressed. The failure to properly align the use of robust force with the political activities of the UN mission as a whole, and to connect military activities to conflict resolution processes that take into account the local, national, and regional interdependent conflict drivers account for the overall failure in stabilizing the country (Berdal, 2018). The case illustrates the difficulties of using military force to stabilize local conflicts without a more comprehensive understanding of how various conflict levels and drivers interlink.
International peacekeepers are increasingly mandated to address communal conflicts but there remains a dearth of research on ending such conflicts, not least because the category of 'local conflicts' lacks disaggregation. Proponents of bottom-up peacebuilding argue that resolving local conflicts and strengthening local actors is vital for building a sustainable peace after civil war (Autesserre, 2017b). Autessere identified local communities in the Congo where civilians have addressed violence and poverty on their own, seeking mediation and justice through religious and traditional leaders rather than militias. Consequently, she argued, the international focus should be on supporting such local initiatives, rather than almost exclusively engaging with national elites in the capital Kinshasa, because bottom-up approaches would be at least as effective-if not more so-as top-down peacebuilding and should be seen as a necessary complement to addressing national politics (Autesserre, 2017c). Critics have countered that local conflicts in the DRC remain deeply intertwined with national conflict and power struggles, and that local actors have finite peacebuilding capacity (Stearns, Vlassenroot, Hoffmann, & Carayannis, 2017). They see little evidence that grassroots actors have the power to stand up to powerful militias and the political networks that support them. It would be hard to find a local dispute in the Congo that is divorced from elite politics because 'most of the protracted communal conflicts in eastern Congo and elsewhere in the country have been provoked, instrumentalized, or sustained by regional, national, or provincial political actors' (Stearns et al., 2017).
Debates about how local conflicts link to elite politics and how local peace initiatives can best be supported would benefit from more precise terminology and a disaggregated analysis of communal conflicts. I understand communal conflicts as distinct conflicts that may interact with civil war dynamics. In contrast to the broader category of 'local conflicts', which may include numerous nonviolent conflicts between social groups that only escalate with the arrival of government forces or rebel groups and then impact civilians and war dynamics, communal conflicts usually exist before, during and in parallel, and often after civil war. For some violent incidents the line between civil war violence and communal conflict may be blurred. However, the local population often understands such armed conflict as distinct from the national conflict. For example, in South Sudan, people distinguish between 'the government', and its army and police forces and militarized rebel movements with a national political agenda, and the 'home' community and its militia, which is primarily seen as a defender of rural cattle communities (Pendle, 2015, pp. 412-413).
While civil wars are fought against the state government and over national government control, and rebel organization command full-time recruits, communal conflicts are fought over local government control (particularly in urban areas), political representation, land ownership and use, natural resources, and access to economic opportunities. Even though these conflicts do not directly challenge the national government, and are frequently referred to imprecisely as 'ethnic' or even 'tribal', they are fundamentally political and do not take place in isolation from national political processes and the institutional framework of the state. To further a more systematic analysis of such conflicts and their links to civil war, peacekeeping and stabilization, I draw on a typology and analytical categories that support the study of variation in communal conflicts ( Table 1).
Scale is a fundamental distinction: If a communal conflict reaches the threshold of a civil war (1,000 deaths in a year), I refer to the conflict as a communal war. The dynamics of violence in a communal war differ from a conflict that kills in the double digits. Consequently, a communal war requires different intervention and civilian protection strategies. I first distinguish type of communal conflict. Dynamics of violence differ in one-sided communal violence or pogroms in which a majority attacks a minority, compared to dyadic conflicts or communal clashes fought between similarly strong communal groups. The two types of conflict require different intervention and protection efforts. In both types of conflict, armed groups may perpetrate attacks and massacres. In addition, in dyadic conflicts, similarly strong groups may also face each other in battles, akin to civil war dynamics. When both sides mobilize repeatedly for revenge attacks, such dyadic conflicts may turn into long and protracted cycles of violence.
Second, the geography of communal violence further shapes protection and prevention strategies. Urban violence is often referred to as 'riots' and election-related, while rural violence is often termed 'farmer-herder' conflicts (e.g. Nigeria), cattle conflicts (e.g. South Sudan), or generally pastoralist conflict because such conflicts involve issues of land use and cattle raiding. These descriptions refer to some of the root causes of the conflict but neglect other dynamics and ignore deeper questions of agency and motives for fighting. Rural communal conflicts may not simply be triggered by dire economic conditions. Although climate change, land scarcity, and poverty are root drivers of the conflicts, the breakdown of conflict resolution mechanisms is key for explaining the escalation and deadliness of these conflicts (e.g. Nigeria, see Higazi, 2016). Urban and rural dynamics of violence can interlink and mutually reinforce each other due to overlapping civilian social networks. This means that if localized stabilization efforts ignore such linkages they risk further undermining long-term stability. Third, I distinguish different categories of armed actors. In communal conflicts, the level of organization and military capability can vary vastly, ranging from neighbours who take out kitchen knives, to urban thugs and gangs armed with guns, to communal militias with hundreds of fighters and military training. Battles and massacres are commonly carried out by militia groups, usually with guns at their disposal and some training in military tactics. Once conflicts have escalated to the scale of communal war, state security forces may find it impossible to bring hundreds or even thousands of armed civilians under control. Such levels of escalation require different intervention strategies by peacekeepers and disarmament programmes.
Fourth, the national context is an important dimension. Whether urban clashes take place in the form of one-sided pogroms in the context of elections or militias fight each other in rural areas, locked into a dyadic conflict while the state fights a civil war, represents very different contexts that require distinct intervention strategies. A look into some of the very deadly communal conflicts in Nigeria (Angerbrandt, 2018;Higazi, 2016), Indonesia (Krause, 2018;Tajima, 2014;Van Klinken, 2007), South Sudan (Brosché, 2014;Jok, 2013) or the DRC (Vlassenroot & Huggins, 2005), demonstrates that large-scale communal violence often takes place in the shadow of civil war when the state is unable or unwilling to manage conflicts and halt civilian armament and conflict escalation.
Communal militias may align with parties to the civil war but may not be fully controlled by either the government or rebel groups. Furthermore, their mobilization dynamics differ from rebel groups in that their fighters usually stay deeply connected to their home communities, where they remain embedded as husbands, brothers, fathers, and sons, fighting often with extensive community support (e.g. Stringham & Forney, 2017 on South Sudan). Women and men, children, and the elderly, may support conflicts and provide the logistical preparations necessary for fighting. In South Sudan, women often encourage further cattle raiding, preparing special foods for the men who fight, and shame men into participation, while children aid the militias as combatants and herders (International Crisis Group, 2014; Rolandsen & Breidlid, 2013).
Breaking cycles of communal conflicts requires understanding scale, type, geography, actors, and national context, to analyse how communities mobilize, militarize, and protect themselves. Building local resistance against further killings often needs broad-based community support. Militarization of civilian communities and local governance further produces long-lasting effects for local power structures, access to government positions, markets and economic produce, as well as community and family relations. Hence, the challenges of civilian disarmament are different from those of rebel groups whose fighters have often been part of a more institutionalized armed group with loyalties to national elites.

Communal Conflicts in South Sudan
Communal conflicts in South Sudan have complex historical roots in local practices of cattle herding, but their transformation into deadly communal wars 'was deliberately wrought by political elites in order to mobilize civilian raiders for their own ambitions' (Wild, Jok, & Patel, 2018). The large-scale communal conflicts, particularly those in former Jonglei State and in Upper Nile, in the centre of the country, have received significant academic and policy attention (Brosché, 2014;Brosché & Elfversson, 2012;De Vries & Schomerus, 2017;Johnson, 2016;Jok, 2012Jok, , 2013Laudati, 2011;Pendle, 2018;Rolandsen & Breidlid, 2012Stringham & Forney, 2017;UN Security Council, 2009;Wild et al., 2018;Willems & Rouw, 2011). Jonglei is one of the world's most under-developed regions and home to an estimated 1.3 million people (Johnson, 2016). The largest ethnic groups that reside in this area are the Nuer, Dinka, Murle, and Anuak. The Dinka and the Nuer are the two largest ethnic groups and have long held the most senior government positions: the presidency (Dinka) and the Vice-Presidency (Nuer).
The conflicts have killed thousands and spurred numerous local peace initiatives by the United Nations Mission in Sudan (UNMIS) and later UNMISS, and exposed the limits of civilian protection by peacekeepers when faced with communal militias (Johnson, 2016). Apart from Jonglei, other areas also suffer from communal conflicts, such as deadly intra-Dinka clashes, but have received less research attention. 2 I assume a general underreporting of conflict deaths from communal violence in newspaper-based datasets, such as the Armed Conflict Location and Events Dataset (ACLED) and the Uppsala Conflict Data Program (UCDP), due to the lack of reporting from remote parts of the country, the poor transport infrastructure, and journalists' difficulties in publishing reports of communal conflicts that Western media outlets may find 'too complicated' (see Dawkins, 2019). Therefore, no comprehensive overview of communal conflicts in South Sudan can be offered here.
Large-scale communal conflicts primarily take place in rural and difficult-to-access areas, where some have escalated into communal wars. The reported conflicts are primarily dyadic rather than pogroms of a majority against a minority. They stem from a history of cattle raiding and ethnic group polarization during the civil war with Sudan; discourses of revenge; lack of political representation and security provisions; local tensions over land, cattle ownership and grazing rights; the concentration of cattle ownership among the political elite; and deep distrust against the government. Communal militias engage in battles and attack civilian settlements.
In the shadow of the five decades long civil war with Sudan, cycles of large-scale communal violence have mobilized thousands of men into armed groups to protect communities and cattle and to serve in alliance with civil war actors (Pendle, 2015). These militias have been co-opted and instrumentalized during both the civil war with Sudan (1983Sudan ( -2005 and the South Sudanese civil war (2013-2018). All sides armed cattle herders, which had a lasting impact on the practice of cattle raiding (Pendle, 2015(Pendle, , 2018. Communal conflicts also affect urban environments, for example PoCs. Research reported communal conflicts and crime as major security threats in the PoCs in Juba, Bor, and Bentiu (McCrone, 2016). These clashes often involved hundreds of residents but with the presence of UN police and peacekeepers, they rarely killed people. Disputes and inter-communal tensions were strongly affected by reports of fighting and atrocities taking place in surrounding rural areas, especially because communal militias were often involved in the fighting (McCrone, 2016). Militias have also repeatedly attacked urban areas out of resentment that urban populations received significantly more foreign aid and material wealth than the rural ones, and some militias sought to 'level' these 'islands of development' with their attacks (Stringham & Forney, 2017).
Cattle ownership by national political elites represents a further important link between urban / national political processes and rural dynamics of violence (see also Brosché, 2014). Since the 2005 Comprehensive Peace Agreement (CPA), many politicians have invested their wealth into cattle ownership in their homelands and fund male family members to guard their cattle, including the provision of guns (Pendle, 2015). These cattle herders are part of the cattle raiding conflicts, and access further guns and ammunition through their patrons in the capital, thus worsening the general militarization of rural communities.

South Sudan's Civil War and Communal Conflicts
The legacies of community militarization before and after the 2005 CPA impacted the 2013-2018 civil war. The CPA did not address communal conflicts and militia groups. Instead, it constructed a 'vision' of the South Sudanese government that 'removed legitimacy from non-government armed groups including localised, armed defence forces that protected communities and cattle' (Pendle, 2015, p. 410). The CPA stated that more than 60,000 militia fighters were to be integrated into the army, the police, or the prisons and wildlife services, but many SPLA officers resented this policy and militias mostly remained excluded (International Crisis Group, 2009). The legacies of the civil war and the proliferation of small arms made communal conflicts significantly more deadly and undermined traditional conflict resolution mechanisms that in the past had limited the deadly impact of cattle rustling and land right conflicts (International Crisis Group, 2009).
The 2006 Juba Declaration that followed the CPA was an elite bargain that did bring together leaders and stabilized national-level politics, but which neither ended local-level fighting nor created a situation of security (Pendle, 2018). For some communities, their experience of communal violence remained so intense that the CPA and national peace was meaningless (Jok, 2013). Even independence did not decrease the deadly impact of communal wars, but rather led to further massive killings. State efforts to disarm civilian communities were planned alongside extensive local peace conferences, but the army's use of force resulted in high numbers of civilian casualties and raised international criticism (International Crisis Group, 2009). The SPLA's forced disarmament during the December 2005 to May 2006 campaign targeted the Nuer disproportionately and reportedly killed an estimated 1,600 people (Small Arms Survey, 2007). Subsequently, a better organized disarmament campaign, which relied on support from traditional chiefs, was more successful, partly because chiefs were keen to regain control and reduce the influence of communal militias, in particular the 'White Army' of Nuer youth. In 2007, the Murle in Pibor county were targeted with a disarmament campaign, but only a fraction of guns in civilian hands were obtained (Rands & LeRiche, 2012). The uneven civilian disarmament programmes resulted in disarming the Lou Nuer, but not the Dinka and the Murle, in Jonglei (Brosché, 2014;International Crisis Group, 2009). This contributed to further attacks and a re-armament of civil militias on all sides. The Murle remained among the most marginalized groups and were politically, socially, and economically isolated from the rest of Southern Sudan (Laudati, 2011).
In 2009, communal wars killed an estimated 2,500 people, predominantly in Jonglei, and no reconciliation took place (International Crisis Group, 2009). Communities remained heavily militarized. The government sought to address these conflicts and brokered intercommunal dialogues and a peace initiative between the Lou Nuer and the Dinka. A civilian disarmament programme was again put into place but due to the legacy of violence in the context of disarmament programmes after the 2009 cycle of killings, communities did not trust state authorities to provide security and no real disarmament occurred (International Crisis Group, 2009).
Given the stark tensions within the SPLA and its leaders in the newly independent South Sudan, national elites 'were reluctant to rebuild relationships between South Sudanese communities in case they needed to mobilise a constituency based around historic and ethnic divisions' (Pendle, 2018, p. 29). Many communities remained afraid to restore interethnic relationships and reinstall peaceful mechanisms of local justice as they did not think they had permission to do so from elites (Pendle, 2018, p. 29). After decades of victimization and persistent insecurity, local communities also often lacked the capacity to sustain bottom-up conflict resolution and peace efforts on their own. 3 Furthermore, community militias were not simply controlled by national elites but pursued their own interests and at times rebelled against them (Stringham & Forney, 2017).
During the civil war, former Jonglei State became a site of large-scale counter-insurgency operations and has been among the areas worst affected. The region's overlap between rebels and communal militias also provided further tinder for the war (International Crisis Group, 2014). The civil war broke out in Juba and resulted in mass killings of Nuer civilians and the creation of PoC camps next to the UN base for those who managed to flee their killers. The killings of Nuer in Juba stoked tensions in Jonglei because many had relatives in the rural areas their families came from. Conflict discourses equated the Nuer with rebel supporters loyal to Rieck Machar, the former Vice President, while the government in Juba and the SPLA was seen as a Dinka force. The incorporation of a Dinka militia consisting of former militarized cattle herders loyal to president Salvar Kiir and his homeland, which was stationed near Juba for his protection, furthered the ethnicization of the civil war (Pendle, 2015).
The legacy of large-scale communal conflicts had severe implications for civilians. They further eroded civilian capacity to negotiate neutrality during the civil war because communities with associated militias could not credibly remain neutral and avoid being drawn into the conflict. 4 Some cases of local Dinka and Lou Nuer groups, and Murle and Lou Nuer groups, have reportedly prevented heavy fighting with ad hoc peace agreements, enabled by particular constellations of local to national elites rather than community-based peace efforts (International Crisis Group, 2014). In general, for civilians in Jonglei, ethnicity mostly determined alliance. According to South Sudanese NGO staff members who worked with civilian communities to support protection strategies, even when communities tried to evade the fighting, surveillance by government or rebel forces made this almost impossible, while support from some civilians for government or opposition forces turned entire communities into targets once the frontlines shifted. 5 Given community militarization, men who wanted to evade the fighting had few options and would be harassed and have their cattle and possessions confiscated if not supporting 'their' community in the war. 6 In the midst of the civil war, communal conflict and cattle raiding between the Dinka, Nuer, and Murle communities continued. The UNMISS Civil Affairs Division repeatedly tried to address these conflicts and brokered peace deals between communities. For example, in December 2016, the Bor Dinka and Murle youth signed a peace agreement, which lasted for five months. In May 2017, the government, under the leadership of then Vice-President Taban Deng Gai and with support from UNMISS, brokered another ceasefire agreement (Mold, 2017). In late 2017, fresh fighting took place between communal militias. At the end of 2017, Nuer communities in the rebel stronghold Akobo feared a final offensive by the Dinka-dominated SPLA government forces and new attacks by the Murle militias, while the Murle in turn feared well-armed Dinka militia attacks. 7

Stabilization and Communal Conflicts in South Sudan
UNMISS' mandate included facilitating the prevention, mitigation, and resolution of intercommunal conflict in order to foster sustainable local and national reconciliation as an essential part in preventing violence and long-term statebuilding activity (UN Security Council, 2016). The two missions, UNMIS and UNMISS, continuously tried to support local peace and facilitated negotiations with militias and communities. Some observers lauded UNMISS for having 'successfully threaded the complexity of local conflict dynamics into the broader national analysis and their political strategy to implement the mandate' (O'Bryan et al., 2017).
Before the country's independence, the UNMIS recognized communal violence as a major security problem, particularly in then Jonglei State. In response to the 2009 clashes, UNMIS devised the Jonglei Stabilization Plan in close collaboration with the Government of Southern Sudan and the state governments, in order to enhance civilian protection activities in the region (UN Security Council, 2009). Temporary UNMIS bases were established within Jonglei State in Pibor, a town dominated by the Murle ethnic group, and in Akobo, a Lou Nuer stronghold (UN Security Council, 2009). UNMIS also supported a peacebuilding process in Jonglei led by the Southern Sudan Peace Commission, with plans for a Lou Nuer-Murle peace conference. Since the 2009 fighting, communal militias have resorted to revenge attacks aimed at inflicting maximum damage and conflict narratives refer to wiping out the enemy population (UNMISS, 2012).
However, after independence in 2011, communal wars in Jonglei escalated again, only five weeks after UNMISS was officially established (Johnson, 2016). According to Hilde Johnsson, then head of mission, protecting civilians from communal wars was well recognized as part of the mission's mandate, but its military capacity to do so was 'wholly inadequate' (Johnson, 2016, p. 104). Although the mission initiated another peace process for the conflict-affected communities, mobilization for further attacks continued. For several months, the Murle had raided Nuer communities, killing up to 1,000 people (Rands & LeRiche, 2012). In response, in December 2011, an estimated 6,000-8,000armed youth from the Lou Nuer, the 'White Army' militia, marched against the Murle. UNMISS flew then Vice President Rieck Machar deep into Murle territory to convince the mobilized men not to attack, but he did not succeed (Johnson, 2016). Machar offered no convincing plan of how to prevent Murle raids on Nuer communities and Nuer militia leaders defied him and continued their attack (Stringham & Forney, 2017). UNMISS had also urged the SPLA to deploy and deter an attack on the Murle, and sent its peacekeepers to the area as much as logistically possible. Yet, with the total of UNMISS infantry numbering a mere 3,500, they had little effect against the Lou Nuer militias marching with clear military training and under unified command (Johnson, 2016, p. 112). Their attacks could not be prevented by military force. From December 2011 to February 2012, at least 900 people, and possibly many more, were killed, and some 90,000 people displaced into remote territory (Rands & LeRiche, 2012;UNMISS, 2012). The Murle responded with smaller retaliatory attacks against the Lou Nuer and the Bor Dinka (UNMISS, 2012).
In 2012, a peace process spearheaded by church leaders and mandated by the government aimed at bringing the Lou Nuer and the Murle into dialogue. UNMISS facilitated further local peace initiatives, culminating in the All-Jonglei Peace Conference in May 2012 where the paramount chiefs of all three communities (Nuer, Murle and Dinka) signed a detailed Framework Agreement for peace, in the presence of the President (Johnson, 2016, p. 115). UNMISS then flew the paramount chiefs together to various parts of Jonglei to prepare people for the peace process. Johnsson estimated that UNMISS provided almost 1,000 helicopter flights to support the Jonglei peace process (ibid). In parallel, the government launched another civilian disarmament campaign for Jonglei communities. However, the level of civilian armament and distinct military tactics of the Dinka, Nuer, and Murle militias represented serious challenges to security forces that were not adequately equipped to conduct such wide-scale operations with significant impact (Rands & LeRiche, 2012). Until the outbreak of the South Sudanese civil war, the situation in Jonglei remained the 'greatest headache' for UNMISS (Johnson, 2016, p. 126).
Interlinking national and local dynamics of conflict repeatedly undermined 'local' peacemaking efforts. In 2009, when communal conflicts killed thousands, an UNMIS staff member lamented, 'all our peace conferences and reconciliation efforts at the grassroots level are doomed as long as politicians are whipping up their communities' (International Crisis Group, 2009). Years later, during the civil war, observers in Juba continued to question whether any local peace negotiations could be viable as long as the parties to the civil war did not seriously negotiate peace. 8 Many national politicians own cattle and can order cattle-keeping communities into specific grazing routes. When a local ceasefire has been agreed, politicians may deliberately stoke tensions between cattle and farming communities, even outside areas affected by the civil war fighting, if this serves their political ambitions, leaving local communities with limited agency to avoid and mitigate conflicts. 9 Often, local peace negotiations have only achieved temporary ceasefires. Such ceasefires may be important, as they allow for limited humanitarian assistance to be provided to the civilian population, but to date there is little evidence that these measures significantly reduced the level of armed violence in South Sudan.

Conclusions
Peacekeepers are increasingly mandated to address communal conflicts and to protect civilians locally. The analysis of such conflicts requires precise terminology beyond the binary of 'national' versus 'local' conflicts. This contribution has proposed a typology of communal conflicts for their careful analysis, distinguishing scale, type, geography, armed actors, and linkages to the national context. Communal conflicts often interlink horizontally across urban/rural divides and vertically from the sub-state to the national level and cannot be addressed in isolation. The High-Level Independent Panel on Peace Operations concluded that the fundamental obligation of a peace operation is to support the implementation of a political process to establish a sustainable peace (United Nations, 2015). High-intensity communal violence can undermine this process. It threatens the legitimacy of armed actors' commitment to peace, creates new challenges of civilian protection and IDP flows, and makes large-scale election violence more likely. The case of South Sudan demonstrates how protracted communal conflicts can precede and succeed a civil war and interlink with national political dynamics and elites. Breaking the conflict cycle requires an understanding of how communities mobilize, militarize, and protect themselves in the context of communal conflict and civil war, and often means effectively addressing decades of community militarization.
Long-term stability would require large-scale civilian disarmament on the local level, which in turn demands a stable political settlement on the national level because subnational political elites, and even community leaders, may often remain deeply dependent on national elites. Military force can temporarily establish 'islands of stability', and peacekeepers can restore an environment conducive for local peace negotiations by providing truce enforcement, civilian protection, and reassurance. However, such temporarily stabilized areas remain embedded in the wider urban/ rural and national/ local conflict dynamics due to family and kin, ethnic, and religious networks. Without a peace process that integrates both national elites and local communities, 'stable' areas can quickly turn into sites of renewed fighting on both the local and national level, with dire consequences for civilians.