Trajectories from Autochthony Discourses to Violence

The apparent self-evidence and naturalizing effects of the powerful claim of having been “born from the soil itself” (ancient Greek: auto–self, chtonos–soil, earth), which is at the heart of autochthony discourses, hide the ambiguities, paradoxes, and uncertainties that lurk under the surface. According to Geschiere, the inherent tensions in these discourses are to a large extent related to their “free-floating profile” and the “unfinished quality” (2009:6,34) of the figure of the autochthon and its constitutive “other,” the “foreigner.” Importantly, the notions of “self” and “other” can be applied at various scales (local, national, regional), where they draw from and overlap with other discourses of belonging, like ethnicity (Jackson Reference Jackson2006). Indeed, the discourses of ethnicity and autochthony display a number of similarities, as in both cases the socially constructed nature of their categorizations is camouflaged by their “everyday primordialist” appeal (Fearon & Laitin Reference Fearon and Laitin2000:848).^{Footnote 1} Yet there are also significant differences between them. One of the hallmarks of autochthony is its lack of substance. It ultimately revolves around a dichotomy between strangers and original inhabitants that can be operationalized both within and between pre-existing social groups, including those designated as “ethnic” (Jackson Reference Jackson2006). Ethnicity, by contrast, constitutes a more comprehensive form of social categorization, and is defined by elaborate rules of both membership and content, relating to the specific characteristics attributed to members (Fearon & Laitin Reference Fearon and Laitin2000). While both the rules and content of ethnic membership are subject to change, their specified and historically shaped nature makes the contours of ethnicity less plastic than those of the free-floating trope of autochthony (Geschiere & Jackson Reference Geschiere and Jackson2006). This extreme plasticity nourishes a pervasive uncertainty, which has been identified as a crucial step on the path from discursive to violent social practice (Appadurai 2009; Geschiere Reference Geschiere2009).

Autochthony is recognized as an increasingly important factor in the occurrence of collective violence in Africa and elsewhere (Geschiere Reference Geschiere2009). In a recent volume, Bøås and Dunn find that claims to autochthony in Africa “often provok[e] dramatic expressions of political violence” (2013:2). These instances of what they call “autochthonous violence” commonly do not implicate social groups as a whole, but involve “young men, usually operating within private militias or armed gangs” (2013:25). Bøås and Dunn propose three broad trajectories through which autochthony discourses may lead to violent action. The first trajectory relates to the interplay between certainty and uncertainty, and draws on Appadurai’s (Reference Appadurai1998) claim that the aura of “dead certainty” surrounding “ethnic violence” provides an antidote to uncertainties relating to identification. Similarly, violence remedies the uncertainties stemming from the meandering nature of autochthony’s boundary-drawing and its inherent paradoxes. The communicative nature of violence and the performative dimensions of identification further feed into this turbulence, with violent acts becoming seemingly effective tools in the elusive quest for certainty, truth, and purity. The second trajectory runs via narratives of victimization, marginalization, and exclusion, which nourish justifications of violence. Such legitimizing discourses commonly have pronounced historical dimensions, emphasizing past injustices like massacres, and drawing on ideologies of anticolonialism and anti-imperialism. However, these narratives also focus on contemporary conditions such as poverty and the denial of rights articulated in the framework of the nation-state, including citizenship, political representation, and access to state-provided goods (see also Appadurai Reference Appadurai1998). The third possible trajectory from autochthony to violence is connected to the violent nature of state-making practices, in particular processes of geographical and identity-related boundary-drawing and the quest for social order (see also Comaroff & Comaroff Reference Comaroff and Comaroff2001). As Bøås and Dunn point out, state-making relies on a double maneuver—the production of ontological crises and the propagation of the idea of the sovereign nation-state as a means to overcome these. Where this feeds into political competition framed in antagonistic identity-based terms, violence may be the result (see also Geschiere & Nyamnjoh Reference Geschiere and Nyamnjoh2000).

As this overview suggests, the trajectories described by Bøås and Dunn are broad conceptual frameworks outlining the contingent links between autochthony discourses and violence, rather than detailed causal pathways. Therefore, several of the analytical issues commonly raised in relation to collective violence remain relatively undertheorized, including the need to distinguish analytically between conflict and violence (Brubaker & Laitin Reference Brubaker and Laitin1998) and between immediate precipitants and broader underlying causes (Collins Reference Collins2006); the relations between specific organizations, in this case armed groups, and the wider social networks from which they are drawn (Brubaker Reference Brubaker2004); and the processes by which violence becomes coded post hoc (Brass 1997). While explicitly addressing these issues may not be necessary for a broad descriptive study, it does become important when claims are made about the nature of violence, which occurs when it is labeled as “autochthonous.” This raises a question similar to the one posed by Brubaker and Laitin (1998:427) in relation to “ethnic violence”: how does the adjective autochthonous modify the noun violence? Only an in-depth empirical study of the actual practices of violence that are framed by discourses of autochthony can help to answer this question.

Note on Methods

As Kalyvas (Reference Kalyvas2006) explains, the study of violence presents a number of methodological challenges. Violence is driven by multiple, overlapping, and sometimes contradictory motives that differ per perpetrator and per act, and may be aimed at achieving an equally complex mix of sometimes contradictory objectives. This renders it difficult, for eyewitnesses and analysts alike, to establish why specific acts of violence have been committed. Moreover, interpretations of violent acts are heavily shaped by the grids of the intelligibility, memory, and positionality of those reporting them, leading to numerous biases. These biases often come to inform the interpretations of violence of outside analysts, which are susceptible to a host of methodological and interpretational fallacies. One of these is the tendency to infer motives for violence from the identity or political-ideological orientations of perpetrators and victims, which is a crucial mechanism underlying the coding of violence as “ethnic.” Such undue inferences commonly correspond to a heavy orientation toward perspectives of urban-based elite actors, to the detriment of narratives from rural areas and broad layers of the population. Furthermore, since analysts commonly focus on explanations of violence that are informed by a conflict’s overarching narratives, the data thus obtained tend to be decontextualized and to downplay the importance of local and personal conflicts.

In order to overcome these data problems, the research for this study was conducted in rural areas in Fizi, despite the fact that large parts of the territory are barely accessible due to insecurity, poor road conditions, and difficult terrain. The data were collected by means of ethnographic methods over nine months between 2010 and 2012, in three successive stages. The iterative nature of the study facilitated the establishment of trust with key informants, including field-based monitors of a human rights organization charged with systematically collecting data on violent incidents, members of the civilian and military security apparatus, local customary and politico-administrative authorities, members of the military branch and political wing of PARC-FAAL, and the leadership of a number of other armed groups operating in the area.^{Footnote 2}Aside from the towns of Baraka and Fizi centre, research was conducted in thirty-three villages throughout Fizi territory, including areas accessible only by boat or motorcycle or on foot, and those located within the Mai-Mai’s main zones of operation.

In each village, data were gathered by means of group interviews of around ten inhabitants and/or semistructured interviews with key informants such as the village chief, schoolteachers, the staff of local health care centers, religious leaders, members of community organizations, and the vieux sages (elderly wise men). Interviews also took the form of less formal discussions in the evenings, often with the customary chief and village elders. The collected data were triangulated with and complemented by information obtained via interviews and written reports from members of the security services and politico-administrative apparatus, local media, civil society organizations, various armed groups, agencies of the United Nations, and international nongovernmental organizations (NGOs). However, data collection was not limited to violent incidents. Another main objective of the fieldwork was to develop an in-depth understanding of the histories and dominant discourses of the Babembe (i.e., those of the Bembe ethnic group) and the Banyamulenge, respectively the majority and minority group in Fizi, and the socioeconomic and power relations both within and between them. This contextual information was deemed crucial for detecting possible reporting biases and for drawing inferences on the processes leading to acts of violence.

The Historical Evolution of the Language of Autochthony in Fizi

The political salience of the discourses of autochthony and ethnicity in the Kivus and their implication in collective violence are inscribed in a wider history of ethnically framed political competition and violence in the Great Lakes Region. This history is deeply, although by no means exclusively, shaped by developments in the colonial era, when complex, overlapping, and primarily socioeconomically grounded social categorizations were transformed into ossified hierarchies based on essentialized, biologically defined identities (Lemarchand Reference Lemarchand1994; Newbury Reference Mutambo1988). Drawing on the so-called Hamitic hypothesis, the German and then Belgian colonizers of Rwanda and Burundi came to portray Tutsi (associated with Caucasoid “Hamites” and “Nilotics”) as a distinct and superior invading race that had subjugated the racially inferior Hutu (identified as “Bantu”), thus creating a distinction between “original inhabitants” and “outsiders/foreigners” (Eltringham Reference Eltringham2006). In this way, the trope of autochthony, which resonated with and came to inform local “mythico-historical” (Malkki Reference Malkki1995:54) narratives of origins, was injected into ethnic discourses. This trope became highly salient politically, first to justify the dominant position of those whom the colonizers defined as “Tutsi,” and then to underpin the claims to political rights of those who had been labeled as “Hutu” (Eltringham Reference Eltringham2006; Malkki Reference Malkki1995).

While the Hutu/Tutsi dichotomy became the predominant grid of intelligibility of power struggles and political violence in Rwanda and Burundi, its political salience has varied strongly over time and according to place, social group, and event. One reason for these fluctuations and uneven distributions is that its salience, to a large extent, has been the product of competition driven primarily by regional cleavages, socioeconomic differences, clan and lineage distinctions, and diverging political views (Newbury & Newbury Reference Newbury1999). Furthermore, the discourses of ethnicity and autochthony have been but one element within the complex set of interacting factors explaining the occurrence, forms, and intensity of violence, including the episode of mass violence in Rwanda commonly labeled as “genocide” (see, e.g., Strauss Reference Stearns, Verweijen and Baaz2006).

Although the evolution, signification, political use, and effects of the discourses of autochthony and ethnicity in the Kivus have a distinct character, they have been influenced by these developments in neighboring countries. This influence can be ascribed partly to the shared colonial past, as Belgian colonial authorities and scholars produced and diffused essentializing narratives of ethnicity in the Congo that were similar to those propagated in Rwanda and Burundi. Specifically, they viewed the social stratification of Rwanda, seen in the light of the Hamitic hypothesis, as a useful model for studying the peoples of Kivu as well (Newbury Reference Newbury1978). The presence of Rwandophone populations (i.e., Kinyarwanda-speakers, encompassing Hutu and Tutsi) in the Kivus, and frequent cross-border movements of refugees and combatants fleeing or participating in violence framed as ethnic, added to this influence (Reyntjens Reference Rukundwa2009). Yet the distinct socioeconomic and political dynamics in the Congo, which harbors hundreds of groups defined as “ethnic” and had its own blend of colonial policies, led to important differences in the historical evolution, forms, and political and social roles of the discourses of ethnicity and autochthony.

The history of the Kivus is strongly marked by the colonial system of indirect rule, which fostered a deep attachment to ethnically defined territories ruled by customary “native chiefs.” The latter were granted a de facto monopoly over political representation and the management of communally owned lands. This combination of policies set in motion the twin processes of the territorialization of ethnicity and the ethnicization of local authority (Muchukiwa Reference Muchukiwa2006)—processes that were disadvantageous to groups that were not granted a “tribal homeland” and therefore were excluded from customary authority and land ownership (Mamdani Reference Mamdani1998). In what is now South Kivu, this exclusion befell a group of Tutsi pastoralists who, since the eighteenth or nineteenth century, had migrated in various waves from present-day Rwanda and Burundi.³ For a variety of reasons—including the small size of the group, its seminomadic lifestyle, and its refusal to pay taxes—the colonial authorities decided to deny it a chefferie aggrandie (enlarged customary chiefdom). Consequently, what were at the time known as “Banyarwanda” (those coming from Rwanda) were subjected to the authority of customary rulers from other communities, to whom they had to pay tribute in order to access land (Rukundwa 2006). Settling eventually in the mountainous Hauts Plateaux area at the intersection between Fizi, Uvira, and Mwenga territories, the Banyarwanda lived relatively aloof in a sociocultural sense, although they maintained important economic interdependencies with other communities (Depelchin Reference Depelchin1974). In the territory of Fizi, this aloofness led to friction with the Babembe, who consider themselves the customary owners of the land.

Tensions intensified during the Simba rebellion in the 1960s, when the Babembe fought on the side of the Mulelist rebels, while the Banyamulenge supported the government. Relations continued to sour after the rebellion, when a new generation of educated Banyarwanda elites started to voice demands for political representation and a local administrative entity governed by the Banyamulenge community. They justified these claims by pointing to their status as “authentic Congolese,” arguing that their ancestors had arrived in South Kivu before the start of colonialism. These representations of origins were reflected in their gradual adoption of the name “Banyamulenge,” or “those coming from Mulenge,” a place in Uvira territory where some of the early movers had temporarily settled (Vlassenroot Reference Verweijen2002). The Babembe strongly refuted this new self-labeling, seeing it as an attempt of the Banyarwanda to mask their Rwandan origins (see also Jackson Reference Jackson2006). While frictions between Babembe and Banyamulenge continued during the 1980s, it was only in the 1990s that they acquired a violent dimension, in particular during the two Congo Wars (1996–97, 1998–2002). Both wars were characterized by ethnically framed bloodshed on a massive scale, pitting Bembe Mai-Mai militias against Banyamulenge linked to insurgent movements initiated and directed by the Tutsi-dominated Rwandan government: first the Alliance des Forces Démocratiques pour la Libération du Congo-Zaïre (AFDL) and then the Rassemblement Congolais pour la Démocratie (RCD). Thus the wars confirmed to the Babembe a widely held belief in the existence of a plot to annex the Kivus to a purported “Hima/Tutsi empire,” an idea that draws on antagonisms between the mega-“ethnic” categories of “Bantu” (associated with self-styled autochthonous groups like the Babembe) and “Nilotic” (relating to Tutsi groups, like the Banyamulenge). The ensuing “Balkanization” of the Congo would facilitate the plunder of its resources, and was therefore believed to be backed by a coalition of Western “imperialists.” These narratives also reflect longer-standing conspiracy theories of Rwandan expansionism that regained currency after the outbreak of the Rwandan and Burundian civil wars at the beginning of the 1990s, which triggered a massive influx of refugees and combatants into eastern Zaire. In conjunction with these (sub) regional developments, autochthony discourses obtained a strong orientation toward the national and regional scales, leading to an emphasis on notions such as “national sovereignty” and “territorial integrity” (Jackson Reference Jackson2006).

The war era also heralded other transformations in autochthony discourses, as they came to assume a quasi-official status and were increasingly oriented toward tropes of autodéfense (self-defense). Shortly after the invasion by the RCD in 1998, President Kabila called for a “war of national resistance,” imploring the population to “take up spears and arrows” to crush the (Tutsi) enemy. In a similar vein, his minister of information called upon youth to constitute a force d’autodéfense (self-defense force) (Lanotte Reference Lanotte2003:104). In this way, the concept of self-defense, as linked to the notion of “patriotism,” became part of the official political vocabulary. Self-defense was also a crucial dimension of Kinshasa’s military strategy toward the RCD, as Mai-Mai groups were used as proxies in a counterinsurgency campaign in the rebel-held areas. This auxiliary status was rendered official by the government’s symbolically incorporating the Mai-Mai into the government forces, at the time operating in non RCD-held areas relatively far from the Kivus. Furthermore, it promoted the most important Mai-Mai commanders to the highest ranks. Thus in 1999 Dunia Lwendama Dewilo (referred to in Fizi as “Dunia”), the chief commander of a coalition of Mai-Mai groups in Fizi called Forces d’Autodéfense Populaires (FAP), was appointed colonel in the Congolese army (interview, March 2011). Due to this change in status, representations of the Mai-Mai groups shifted. In contrast to their former image as fragmented and ragtag rural militias they now became portrayed as patriotic heroes who were part of a nationwide politico-military movement. This broadened their appeal and support base, drawing in growing numbers of urban- and diaspora-based intellectuals and politicians (Jackson Reference Jackson2003).

Within Fizi, the language of autochthony acquired a very specific center of gravity. In September 1999 the RCD rebel administration, in which the Banyamulenge held important posts, created a new administrative entity on the Plateaux, which encompassed most of the lands inhabited by the Banyamulenge. Since other communities, including the Babembe, saw themselves as the customary owners of these lands, the creation of what was baptized the “territory of Minembwe” provoked fierce resistance (Muchukiwa Reference Muchukiwa2006). The Babembe would increasingly justify this resistance by framing the creation of the territory as the first stage in the realization of the infamous “Balkanization” plot. In this manner, the longstanding tensions between Babembe and Banyamulenge were recoded in the terms of the overarching conflict narratives of the struggle between the Congolese government and the RCD/Rwanda, highlighting how the interplay between the discourses of autochthony and ethnicity plays out at various scales. While the war between the government and the RCD formally ended in 2002, autochthony discourses as employed in Fizi continued to refer to these wider national and (sub)regional conflict narratives.

The Autochthony Discourses of the Mai-Mai in Fizi

In the wake of the signing of a peace accord in 2002, Dunia’s Mai-Mai brigades were supposed to integrate into the newly created national military, the Forces Armées de la République Démocratique du Congo (FARDC), just like various Banyamulenge armed factions. However, a part of the Bembe Mai-Mai was fiercely opposed to military integration. In January 2007 this group of dissidents, led by battalion commander William Amuri Yakotumba, went underground to found a politico-military movement called Mai-Mai reformé, renamed PARC-FAAL in 2011. Yakotumba justified this move by referring to the continuing threat posed by Banyamulenge rebels, in particular two groups that refused to integrate into the FARDC (interview, December 2011). These rebels advocated the restoration of the territory of Minembwe, an administrative entity that had been suppressed in 2003, when it was declared unconstitutional. But Yakotumba’s dissidence was also nourished by a variety of other factors, including personal grievances about only having been appointed captain in the FARDC. More important, by now General Dunia (who was promoted from colonel to general in 2004) and other elites in Fizi had strongly encouraged the launching of a new rebellion, hoping that it could safeguard the spheres of influence they had carved out during the wars. These more particularistic motivations are little visible in PARC-FAAL’s public discourses, which focus more on political and identity-related issues.

By far the most salient theme of the Mai-Mai’s official political communication is agitation against the Banyamulenge and their alleged desire to resurrect Minembwe territory. In an emblematic speech held at a large peace conference organized in Goma in January 2008, a Mai-Mai representative stated that “erecting Minembwe as territory will lead to a bloody and interminable war of the type KOSSOVO [sic], or IRAN-IRAK” (Assanda Reference Assanda2008:10). In the same speech, the Mai-Mai refuted all claims of citizenship of the Banyamulenge, whom they continued to label as “Banyarwanda” immigrants and refugees, and also announced their two most important objectives: “1. To defend the soil and subsoil of Fizi territory coveted by the Rwandan immigrants[;] 2. To defend the territorial integrity of the Democratic Republic of the Congo against foreign forces” (Assanda Reference Assanda2008:6). These assertions illustrate the connection between the Mai-Mai’s anti-Banyamulenge agenda and the notion of autodéfense, which is a second leitmotif in their discourses. For the Mai-Mai, taking up arms is not a manifestation of bellicose intentions but rather a legitimate act of self-protection needed to ward off the supposed aggression of the Banyamulenge and defend the territorial integrity of both Fizi and the Congo. This emphasis is reflected, for example, in the name of the political branch of the group—PARC. This originally stood for “Parti pour l’Autodéfense et la Reconstruction du Congo,” but it was changed to “Parti pour l’Action et la Reconstruction du Congo” in 2011, when the group submitted a request to be officially recognized as a political party and was advised to remove the radically sounding word “autodéfense” from the name. But autodéfense is not only associated with militancy in the context of the Congo. It also evokes connotations of autoprise en charge (taking care of oneself), an idea that is reminiscent of the débrouillardise, or “fending-for-oneself-ism,” that became an important coping mechanism in the Mobutu era, when public service provision and the official economy crumbled (Lemarchand Reference Lemarchand2002). As such, the notion of autodéfense does not only relate to “foreign aggression” but also draws on longstanding imaginaries of state–society relations.

Another recurring theme in the Mai-Mai’s discourses is victimization, in particular in relation to the marginalization of the Mai-Mai in the power-sharing processes that characterized the transition following the peace accord (2003–6). According to Yakotumba’s group, former Mai-Mai have been systematically denied positions of importance in both the politico-administrative arena and the military, to the advantage of Rwandophones. Accordingly, demands for positions of responsibility for both Mai-Mai politicians and military officers figure prominently in their political communications (see, e.g., Assanda Reference Assanda2008). These demands reflect a wider discontent with the government, in particular President Kabila, who is accused of siding with “Rwanda.” Kinshasa is also perceived as privileging a part of the Fizi Mai-Mai who rallied to the government camp during the transition and who are united in the political party Patriotes Résistants Maï-Maï (PRM). Originating predominantly from the Lulenge sector of Fizi territory, this progovernment faction managed to obtain a number of important political positions in the transitional institutions, driving Yakotumba’s group further toward the opposition. This shows that feelings of marginalization derive not only from intercommunity tensions, but also from divisions among the Babembe. What is further feeding PARC-FAAL’s outspoken antigovernment agenda is a deep disappointment with Kinshasa’s policies and achievements. The preamble to PARC’s policy blueprint (“projet de société”) states that “the Congolese politician is in fact the product of a system that has blocked the creative spirit of the people by . . . rendering the state and its institutions at the service of a mal-positioned political clientelism and nepotism for facile and illicit enrichment to the detriment of the people” (PARC 2007). Raphael Looba Undji, the president of PARC, also claims that the group is committed to combating the pillaging of natural resources, corruption, injustice, the unequal distribution of revenues, and poor management of the national armed forces (interview, December 2011). This last dimension, the malfunctioning of the FARDC, is often linked back to the notion of self-defense, as the failure of the government to protect both the country’s territorial integrity and its citizens would justify the Mai-Mai’s taking up arms.

Criticism of the government forces became increasingly prominent in the Mai-Mai’s political communications in the course of 2009, after hundreds of new FARDC troops arrived in Fizi for military operations. These troops were dominated by freshly integrated former rebel soldiers of the Hutu-led Patriotes Résistants Congolais (PARECO) and to a lesser extent the Tutsi-led Congrès National pour la Défense du Peuple (CNDP), whose leadership was largely composed of former RCD members. The Mai-Mai capitalized on these developments by presenting them as “evidence” for the “Rwandan infiltration hypothesis,” which posits that the FARDC is dominated by Rwandan troops masquerading as Congolese (Eriksson Baaz & Verweijen Reference Eriksson Baaz and Verweijen2013). Thus the Mai-Mai framed the arrival of Rwandophone troops in Fizi as a sign of “foreign occupation,” thereby reinvigorating the “Balkanization” trope. As many of these troops were ill trained and ill behaved, their presence triggered an upsurge in insecurity. The resulting animosities were in part projected onto the Babembe’s “traditional enemy,” the Banyamulenge, who were suspected of manipulating their supposed “Rwandophone brothers” in the FARDC with the aim of reviving Minembwe territory. In this manner, the local competition between Babembe and Banyamulenge became once again closely linked to conflict narratives at supralocal scales, in this case both the antagonism between the Congolese government and the CNDP, and that between Kinshasa and Kigali.

The Resonance of Autochthony

As illustrated in the previous sections, the Mai-Mai frame their practices in terms of a complex blend of the discourses of autochthony and ethnicity, while also drawing on notions of resistance to the government. This emphasis on the rights of the Babembe as “original” inhabitants of Fizi, and the concomitant antagonism toward the Banyamulenge as “newcomers,” is at once an expression of their worldview and a means to garner support. Indeed, the Mai-Mai’s discourses strike a powerful chord among the Babembe. For Geschiere (Reference Geschiere2009), part of the performative power of the autochthony discourse resides in its connection to Foucault’s notion of “subjectification,” relating to the individual’s constitution as a subject or his or her “subjectivity”—both in terms of self-identification and self-knowledge and as being subject to power (Foucault Reference Foucault1982). Similarly, the notion of autochthony as employed by the predominantly Bembe PARC-FAAL is linked to ethnic discourses about the Bembe self that have been formed over the longue durée. The Congo Wars reinforced these forms of identification, connecting them to traumatic memories of violence inflicted by Banyamulenge and other Tutsi soldiers. As a member of a civil society organization explained, “Yakotumba is a psychological safeguard for the Babembe that the Banyamulenge will not take over power and that Minembwe territory will never be erected. . . . Even when they do not agree with the violence, they want to keep him in the bush as a last resort” (interview, November 2010). The Mai-Mai explicitly play upon these collective self-representations by strongly inscribing themselves in Bembe history and by emphasizing what are supposedly core Bembe values. For example, during the Goma conference referred to above, a Mai-Mai delegate stated that “history demonstrates that the Bembe tribe are proven warriors.” (Assanda Reference Assanda2008:3) Further placing the group in traditions of Bembe resistance, in particular to the existing political order, he added that “the Mai-Mai revolution started in 1964 under the name MAI MULELE . . . and had as [its] objective to revive the Lumumbist ideology” (Assanda Reference Assanda2008:5). A similar emphasis on historical continuity is reflected in one of the Mai-Mai’s main adages, which its members ascribe to Pierre Mulele, a leading revolutionary figure in the 1960s: “A calabash which once contained pepper will always retain the smell” (interview with vice-president of PARC, February 2011).

However, Bembe subjectivity is not the only explanation for the mobilizing power of the autochthony discourse in Fizi. What social movement theorists call “frame resonance” provides an additional understanding. According to Snow and Benford (1988:198), social movements “frame, or assign meaning to and interpret, relevant events and conditions, in ways that are intended to mobilize potential adherents and constituents, to garner bystander support, and to demobilize antagonists.” The extent to which the resulting “frames” or schemes of interpretation (Goffman Reference Goffman1974) strike a chord among social agents—or have what Snow et al. (1986:477) call “frame resonance”—is shaped by, among other factors, their “empirical credibility,” or the extent to which they are substantiated by the unfolding events, and their “experiential commensurability,” or how they harmonize with people’s everyday experiences (Snow & Benford Reference Reyntjens1988:208).⁴ Both of these factors explain why the narrative of autochthony “resonates” among the population of Fizi. Crucially, autochthony discourses render social phenomena that people experience on a day-to-day basis intelligible. For example, the insecurity caused by the arrival of new FARDC troops in Fizi in 2009 lent empirical credibility to, and was given meaning by, “Rwandan invasion” scenarios and representations of Rwandophones as evil. Similarly, autochthony resonates with dissatisfactions and conflicts that develop in connection to the seasonal migration of cows (called transhumance). During these movements, cattle, the majority of which are owned by Banyamulenge, often trample upon the fields of Bembe farmers. Discursive framings of the resulting friction in terms of the autochthony narrative—for example, allegations that certain cattle owners refuse to pay the customary transhumance tax because of their status as “foreign aggressors” (Brabant & Nzweve Reference Brabant and Nzweve2013)—render this abstract concept comprehensible and credible. Indeed, a devastated field and the resulting loss of livelihood are likely to affect a farmer much more strongly than the merely hypothetical threat of an impending “Hima–Tutsi empire.”

To conclude, the strong mobilizing capacity of the Mai-Mai’s blend of the discourses of autochthony and ethnicity can be explained by the ways in which these discourses play upon both subjective and everyday experiences, which they frame and co-constitute. Nevertheless, the resonance of these discourses is not the sole explanation of popular support for and collaboration with the Mai-Mai. As we will see below, support and collaboration also derive from shared political orientations that are not primarily linked to forms of social identification, or they can rest on nonideational bases like family or patronage ties.

The Social Embedding of the Mai-Mai

Due to Fizi’s long history of insurgent activity, armed resistance has become a tradition for some Bembe families, meaning that they expect all members to have at least one son in the Mai-Mai. Families also encourage their relatives to enroll because this connection provides them with a certain influence vis-à-vis others in the community. For example, creditors with relatives in the Mai-Mai can mobilize their kin for help in recovering debt. However, family interests appear to be only one factor in motivating combatants to enroll. Peer pressure, socioeconomic marginalization, and the attractions of social identification as a combatant also push people to join the ranks of the Mai-Mai (see also Jourdan Reference Jourdan, Vlassenroot and Raeymaekers2004). For instance, when inquiring about his choice to join, an ex-combatant explained: “I wanted to go to school but there was no money. My father died during the war. There was nothing I could do. Then the Mai-Mai said they were looking for patriots and I joined” (interview, March 2011). A currently serving combatant said, “I just want to be a soldier. We protect civilians and we defend the country. I joined [the Mai-Mai] out of love for the country” (interview, February 2011). Another added, “They said the war was over but we were still in insecurity. The Tutsi are threatening our villages. They killed two of my sisters. So we need to defend ourselves” (interview, December 2011). Some also highlighted their opposition to the current government: “This government is bad. We are fed up with Kabila. We fight with the Mai-Mai because we want change.” These last comments show that although combatants frequently invoke notions of autodéfense or opposition to the government as reasons for joining, these explanations are usually less elaborate than the radical discourses of “othering” and the ethnically framed commitment to armed struggle employed by the leadership.

Similar to the reasons for enrollment, civilians’ motives for collaborating with PARC-FAAL are varied. Such collaboration is crucial to the Mai-Mai, who depend heavily on civilians for food, financial contributions, intelligence, health care, shelter, and hiding places. Furthermore, they need civilians’ cooperation in spreading their legitimizing discourses, attracting new recruits and supporters, conducting business, and exercising political and administrative influence. Civilian collaboration is the product of a variety of modes of reasoning and social mechanisms, which cover a wide range on the coercion–persuasion spectrum. A large part of the population simply see the Mai-Mai as their own offspring whom they have an obligation to help, even when not fully agreeing with their political agenda. Collaboration may also be the result of fear of reprisals, or derive from “sympathy for the cause” and a shared worldview. Another basis is the expected reciprocity inherent in patron–client relations, which involves the exchange of loyalty, protection, and favors (Pitcher et al. Reference Pitcher, Moran and Johnston2009). Collaboration may also derive from more utilitarian concerns, such as the pursuit of business advantages, social status, or in the case of politicians, votes. In most cases, it does not stem from a single factor, but is driven by a variety of overlapping and sometimes contradictory motives that differ according to the individual and shift over time, making it difficult to pin down the exact degree of coercion or persuasion involved (Kalyvas Reference Kalyvas2006). The examples below further illustrate these intricacies.

Customary chiefs are crucial sources of support to the Mai-Mai, more so than government-appointed administrators, since they are closer to the population and often are viewed as having more legitimacy. Many chiefs in Fizi organize donations both in money and in kind, such as cassava flour and goats. In some villages chiefs also play an important role in the comités de base (local committees) set up by PARC, which organize supporters at the village level. A number of chiefs of mineral-rich localities levy taxes for the Mai-Mai among artisanal miners and gold traders, or facilitate the Mai-Mai’s business activities in the minerals sector. However, the levels and types of collaboration and the motives for these activities differ per chief. Not all of them work for the Mai-Mai voluntarily; some simply want to avoid retaliation and protect their communities. Many comply with the Mai-Mai’s demands out of fear. The risks of opposition are not merely hypothetical. For example, in August 2011 the deputy territorial administrator of Fizi, Kashindi Kati, an important member of the progovernment PRM party, was kidnapped by the pro-opposition PARC-FAAL in the run-up to the 2011 elections (interview, December 2011). The Mai-Mai also put pressure on economic actors, often for purposes of extortion. For instance, in July 2011, in the wake of a series of hijackings of big ships on Lake Tanganyika, PARC-FAAL sent a letter to the president of the shipowners’ association in Uvira demanding a contribution of U.S.$500 per ship for every passage (UNSC Reference Strauss2011). In June 2012 a similar letter was sent to representatives of the cell phone operators Vodacom and Airtel, accusing them of cutting off the phone signal before every FARDC attack and demanding a financial contribution to compensate for this alleged “lack of patriotism” (PARC-FAAL 2012).

Yet the level of threat faced by political and economic actors is highly variable, and some maintain cordial, often mutually beneficial, relations with the Mai-Mai. PARC-FAAL is involved in a wide array of economic activities, including the exploitation of and trade in gold, copper, hardwoods, and cannabis. This has allowed them to build up the economic influence to bestow advantages upon befriended businesspersons, such as privileged access to natural resources. Such entrepreneurs include members of the Bembe diaspora in Tanzania and Burundi, where many fled during the wars. This position abroad enables them to play a crucial role in the transborder commercial circuits through which the extracted resources are traded (UNSC Reference Strauss2011). Collaboration with the Mai-Mai is attractive to these economic actors not only because it facilitates their business activities, but also because they sympathize with the movement and its vision, partly based on its character as a Bembe organization. A similar mixture of utilitarian, political, and identity-related considerations push Bembe politicians to support the Mai-Mai. There are longstanding ties between PARC-FAAL and certain politicians active at the national level, who harness the Mai-Mai both to obtain votes within Fizi and to position themselves in the national political arena (UNSC Reference Strauss2011). These politicians partly share the antigovernment outlook and the radical interpretations of autochthony that are propagated by the Mai-Mai. In order to help these allies obtain votes, the Mai-Mai put pressure on community members during the 2011 election campaign, while also promoting Étienne Tshisekedi, a presidential candidate from an opposition party. Allegedly, in the course of a meeting with customary chiefs shortly before polling day, PARC-FAAL declared that it would burn down every village that would not vote for Tshisekedi (interviews with customary chiefs, members of civil society organizations, and representatives from intelligence services, December 2011, January 2012). Although the incumbent, Kabila, was already very unpopular in Fizi, this may have contributed to his gaining only 13.62 percent of the votes, against 75.15 percent for Tshisekedi (CENI n.d.).

Like their relations with political and economic actors, the Mai-Mai’s connections to civil society organizations and churches are diverse and located variously on the coercion–persuasion spectrum. Some pro-Mai-Mai clergy and civil society activists use their professional position to aid the movement, such as by mobilizing support among their congregations or by using their organizations’ motor vehicles for logistical services, including the transport of ammunition. However, others have been more critical, refusing to “send phone credits,” as the regular financial contributions expected by the Mai-Mai are code-named. Furthermore, certain religious authorities have been active in campaigns to sensitize youth not to join the Mai-Mai, while human rights organizations regularly denounce the extortion, pillage, and assassinations committed by PARC-FAAL. These efforts have often led to serious acts of intimidation, including death threats (Durand Reference Durand2012). For example, a pastor involved in a sensitization campaign related that he received a text message saying that the Mai-Mai “would cut off both of his legs” if he did not stop these activities (interview, December 2011).

As these examples show, the ties between the Mai-Mai and civilians are complex and heterogeneous. This is partly a result of variations in the rationalities and attitudes informing them, which, moreover, have shifted significantly over time. Importantly, the popularity of the Mai-Mai has diminished somewhat in recent years, largely as a result of war weariness and a lessening of the “siege mentality” that had been provoked by the wars. There is also widespread disappointment concerning what the Mai-Mai’s struggle has accomplished for the Babembe, leading some to conclude that their grievances can be addressed more successfully through nonviolent means. These doubts are further nourished by the belief that the leadership of PARC-FAAL displays increasingly self-interested behavior, as manifested in their involvement in business activities. As a civil society activist explained, “Yakotumba has deviated [from] his mission. . . . They used to call him commandant-pasteur, but now he is more interested in interfering with the business of the creuseurs [artisanal miners]” (interview, January 2012). While many Babembe continue to sympathize with the Mai-Mai’s anti-Banyamulenge/Tutsi and antigovernment agendas, this sympathy does not translate automatically into financial or other support. For example, although rhetorical approval for the Mai-Mai initially increased after autochthony discourses became more salient following the arrival of Rwandophone FARDC troops in Fizi in 2009, this did not lead to more recruits or financial support. In fact, at the beginning of 2010 the movement was reported to be at its weakest since its creation in terms of troops and military power.

The noncongruence in fluctuations between, on the one hand, the salience of autochthony discourses, and on the other hand, the amounts of practical and financial support to the Mai-Mai thus indicates that these discourses are not a sufficient explanation for the relatively widespread civilian collaboration with the Mai-Mai. The bases of this collaboration must also be located in other discourses such as opposition to the government, in fear and intimidation, in (anticipated) political and economic advantages, and in social relations based on family, patronage, and other ties. These social dynamics are only partly related to the rationalities of autochthony. This puts into perspective the capacity of these discourses to contribute to the legitimization of Mai-Mai violence, as further illustrated below. Autochthony discourses also play only a limited role in the generation of such violence, which is often informed by other considerations and modes of reasoning.

Autochthony Discourses and the Mai-Mai’s Repertoires of Violence

An extensive study of the Mai-Mai’s “repertoires of violence” (Wood Reference Vlassenroot2009:133) showed that these practices fall into four broadly defined, overlapping categories. The first is violence related to clashes with the government forces, which chiefly targets soldiers but may intentionally or unintentionally victimize civilians as well. The second category is violence related to income generation, such as ambushes, road robbery, piracy, and illegal taxation at mining and fishing sites. The third category is violence aimed at maintaining social control over the Bembe community and the settling of conflicts and scores. Practices in this category include the intimidation of political opponents or economic competitors, revenge on suspected collaborators with the FARDC, reprisals against businesspersons perceived to have violated the terms of exchange, and interventions in conflicts between civilians, such as disputes surrounding debts or the payment of dowry. Such interventions occur either on the Mai-Mai’s own initiative or at the request of civilian parties linked to the Mai-Mai through family or patronage ties (see Eriksson Baaz & Verweijen Reference Eriksson Baaz and Verweijen2014). A fourth category is “expressive violence” (Kalyvas Reference Kalyvas2006:24), which is directed primarily at targets for what they represent in terms of identity or beliefs, but is often combined with other motives like revenue generation. Given the symbolic value of cows as life giving to the Banyamulenge, cattle rustling can be seen as falling partly into this category (Brabant & Nzweve Reference Brabant and Nzweve2013).

The relative importance of each of these categories of violence has varied substantially over time and per locale, depending on such factors as the military context (e.g., relations with the FARDC and other armed groups), the Mai-Mai’s overall sources of income, and the political situation (e.g., electoral campaigns, overtures by the government). In the first years of its existence, the FARDC tolerated Yakotumba’s group, as the government’s main policy was to persuade its members, via negotiations, to integrate into the military. This led to various forms of economic collaboration between locally deployed army units and the Mai-Mai. However, in 2009, when the umpteenth attempt at integration failed, serious hostilities broke out, which only became stronger after the arrival of Rwandophone troops. Despite the Mai-Mai’s discursive framings, these hostilities do not appear to have been exclusively provoked by these troops’ ethnic origins, as evidenced by the fact that FARDC casualties showed very little distinction in terms of ethnicity. The upsurge in attacks against the military was also related to the FARDC’s new mission to fight the Mai-Mai and its efforts to expand economic control over Fizi, thus disturbing the power equilibrium established by the previous troops.

The 2009 events also triggered another shift in patterns of violence. The rising insecurity in Fizi had disastrous effects on people’s livelihoods, which undermined their willingness to actively support the Mai-Mai’s military campaigns. The resulting drop in recruits and income contributed to PARC-FAAL’s intensification of links to two foreign Hutu-led armed groups, the Rwandan Forces Démocratiques de Libération du Rwanda (FDLR) and the Burundian Forces Nationales de Libération (FNL). It also reinforced the group’s engagement in transborder trade and violent revenue-producing activities, such as piracy on Lake Tanganyika (UNSC Reference Strauss2011). This created a backlash among the inhabitants of Fizi, many of whom resented the presence of foreign combatants and the growth in violent forms of income generation, increasingly suspected to be for the personal benefit of the group’s principal leaders. The result was a vicious cycle in which diminishing popular support prompted the Mai-Mai to rely on more coercive and violent strategies to exert social control, which in turn further eroded civilians’ appreciation of the group. The electoral campaign of 2011 reinforced this trend, leading to a sharp rise in the intimidation of political opponents, including the above-mentioned kidnapping of Kati. While some opponents were targeted partly for their perceived tolerance toward the Banyamulenge and Minembwe territory, a more important criterion appears to have been their progovernment and anti-Mai-Mai attitude. This indicates that these forms of violence were driven to a large extent by power struggles within the Bembe community as opposed to tensions with the Banyamulenge per se, even though, to some extent, the two inform each other. For instance, intra-Bembe frictions are often discursively framed as relating to diverging attitudes vis-à-vis the Banyamulenge.

The strong interplay between inter- and intra-ethnic competition is further evidenced in other episodes of violence that occurred in 2011. These were prompted partly by “ethnic outbidding” (Brubaker & Laitin Reference Brubaker and Laitin1998:434) among electoral candidates, or their competition in terms of demonstrating who is most “anti-Banyamulenge,” but were also a reaction of the Mai-Mai to their waning popularity. In order to regain popular support and help their radical allies, PARC-FAAL stepped up expressive violence targeting the Banyamulenge, which up to that point had been relatively rare. In mid-2011 PARC-FAAL raided approximately five hundred cows in Mizinga and Nyamukoma in the Ngandja forest, leading to the death of two herders. These actions, while aimed partly at revenue generation, were widely believed to be geared primarily toward exacerbating intercommunity tensions. A much more blatant effort at escalation took place in October 2011 when a vehicle of the humanitarian organization Eben Ezer was ambushed near the town of Fizi centre. Its passengers were first forced out of the car and then separated according to ethnic origin, with only the Banyamulenge being killed (Human Rights Watch 2012).

While the Eben Ezer massacre was generally ascribed to the Mai-Mai, their staunchest supporters refused to believe they were responsible, alleging it was a setup to justify further military operations against PARC-FAAL by Rwandophones in the FARDC. Yet many Babembe were convinced that the massacre had been sponsored by the Mai-Mai. Although some claimed the vehicle had been attacked because it transported ammunition—implicitly suggesting the killings had been a legitimate form of self-defense—others refuted this justification, considering the attack an ignoble and dangerous deed that could provoke retaliation. These divisions clearly show the limits to the capacity of discourses of autochthony and ethnicity to legitimize violence perpetrated by the Mai-Mai

To date the Eben Ezer massacre has remained an isolated event. This indicates that only a relatively small part of the violent practices of the Mai-Mai are directly aimed at harming the “foreign other” or display the clear identity-related performative dimension that Bøås and Dunn (Reference Bøås and Dunn2013) consider an important characteristic of violence informed by autochthony. There have been no direct confrontations between the core group of PARC-FAAL and Banyamulenge or Hutu armed groups, and the targeting of Banyamulenge civilians has been quite rare overall. The clearest examples of violence provoked by antagonism against the “other” are cattle rustling and, according to the Mai-Mai’s claims, the group’s resistance to and targeting of Rwandophone networks in the FARDC—even though, in practice, their attacks lead to casualties among soldiers of all backgrounds. Furthermore, it is precisely among those carrying out the fighting, the rank-and-file, that radical notions of autochthony appear least developed. Rather, the largest share of the Mai-Mai’s violent practices are related to revenue generation and establishing, maintaining, and expanding control over the Bembe community. While these objectives serve a variety of projects—including enabling the Mai-Mai to reproduce themselves and thereby to realize their ethnically framed political visions—they are importantly oriented toward reproducing and enhancing the personal power and resource base of the group’s leadership and the politicians and businesspersons who are their allies.

Conclusion: Can We Speak of “Autochthonous Violence”?

As this article has demonstrated, most of the violence committed by the Mai-Mai in Fizi is not informed directly by the desire to harm or defend against the ethnic or nonindigenous “other,” but is directed toward revenue generation and the (re)production and expansion of control over the Babembe. Discourses of autochthony and ethnicity are not absent from these categories of violence. Violence is the product of multiple, complex, interacting, and overlapping processes, and these discourses may well play a part in certain dimensions without being the primary driving force. For the Mai-Mai, generating resources, fighting the FARDC, and maintaining control over civilians are a form of, or a precondition of, legitimate self-defense. Yet in the acts of violence informed by these objectives, the role of autochthony discourses is frequently indirect and nonlinear, as evidenced by their forms and targets and the importance of other framings, such as antagonism toward the government or political opponents. Furthermore, there are clear limits to the ability of autochthony discourses to mobilize for and legitimize violent action. Many Babembe fail to see how practices such as massacres of civilians, piracy, and lootings of health care centers qualify as autodéfense. Hence, while a substantial number of the residents of Fizi endorse the Mai-Mai based on their shared political vision and forms of belonging, they do not find all of PARC-FAAL’s practices, as framed in the discourses of autochthony and ethnicity, legitimate. Additionally, the resonance of these discourses is but one factor accounting for collaboration with and support for the Mai-Mai, which is also driven by opposition to the government, fear, existing social ties, and utilitarian considerations.

The nonlinear and indirect relations between discourses of autochthony and the violent practices of the Mai-Mai in Fizi require a modification of Bøås and Dunn’s (2013:97) claim that Mai-Mai violence in the Kivus is “all about resistance, these days first and foremost against a Rwandophone conspiracy.” This is not to say that the three trajectories from autochthony to violence identified by Bøås and Dunn are inaccurate—PARC-FAAL’s discursive and (violent) social practices are clearly nourished by the interplay between certainty and uncertainty, narratives of victimization, and the effects of state-making processes. However, these phenomena in themselves, whether alone or in concert, do not directly generate violence. For example, while similar uncertainties, narratives, and effects are at play in the Lulenge sector of Fizi, the heartland of the PRM, few of the Babembe here support or are recruited into PARC-FAAL. Nor does uncertainty resulting from allegations of doubtful origins—to cite another example—always elicit a violent reaction or the desire to support the Mai-Mai. This is evidenced by the fact that such allegations have also been directed against politicians who are not involved with the Mai-Mai, such as the national Member of Parliament Shenila Mwanza. There are also few indications that the forms of uncertainty resulting from a bad security climate directly reinforce the motivation to engage in or support violent action. While the fear and disruption related to military operations in 2009–10 intensified the salience of autochthony discourses, this situation did not lead to an increase in direct assistance to the Mai-Mai. By contrast, it eventually eroded active support, since it fostered a pronounced war-weariness among the population of Fizi.

When analyzing the role of autochthony-related narratives of victimization, a similarly complex and contingent relation to violence emerges. While these narratives became intensified after 2009 due to the perceived privileging of Rwandophones in the military, they did not lead to greater popular approval for acts of violence, nor did they propel more people to take up arms and join the Mai-Mai. While interviews with Mai-Mai combatants indicate that victimization framed in terms of autochthony does play a role in their desire to enroll, this victimization is often related to individual experiences of misfortune, such as the death of family members, as in the case of the combatant referred to above who lost two sisters. Although these experiences are interpreted through and feed into collective narratives of victimization, they cannot be equated with them.

Finally, when looking at ongoing processes of state-making, it is clear that these do nourish the forms of power competition and boundary-drawing in which autochthony discourses thrive. In fact, the birth of the Mai-Mai reformé itself was in part a reaction to efforts by Kinshasa to reinforce central state authority over Fizi during the 2003–06 transition. Another way in which state-making fosters tensions is via elections, which often trigger ethnic outbidding and lead to the deployment of the Mai-Mai for securing votes. Yet state-making practices also take place in parts of the Congo without armed group presence, showing their contingent relation to organized armed violence. Furthermore, increased mobilization during the electoral period may also intensify tensions between self-styled autochthonous communities and related armed groups, as occurred with the Bahunde and Banyanga in North Kivu (UNSC Reference Strauss2011). This evokes the question of whether the mobilization of PARC-FAAL for electoral purposes is not more an expression of the general militarization of power competition in the Kivus than a result of autochthony discourses per se.

In sum, while autochthony discourses play an important role in mobilizing the populace and legitimating the armed activities of the Mai-Mai—in particular via the trope of autodéfense—this role is complex and variable. Even where identity-related boundary-drawing mechanisms create situations of uncertainty and perceived marginalization, they do not generate armed group violence in and of themselves. Creating and maintaining an armed group requires organizational skills and sustained efforts from determined agents, whose personal interests inevitably play a part in their motivation for instigating or supporting armed mobilization. Within the militarized social order of the Kivus, the premise that political and economic influence require a reservoir of armed supporters prompts social actors of all stripes to found or liaise with armed groups, sometimes with little regard for these groups’ worldviews and discursive orientations (Stearns et. al. Reference Snow2013). Certainly, militarized power-mongering is often not the sole rationality driving armed mobilization, as forms of belonging, political ideas, and existing social ties may be at play simultaneously. But even where armed mobilization feeds into and draws upon the politics of belonging, it is crucially driven by these wider dynamics of power. Furthermore, armed mobilization also occurs where the “original belonging” of a group is not the primary discursive maneuver. Two important examples are the (now defunct) groups led by Richard Tawimbi and Baleke Sumaili in the Uvira mountains in 2011–12. Interviews with the leadership of these groups and the analysis of their political communications showed that neither appealed directly to autochthony discourses to frame their violent activities (e.g., MDP 2012a, 2012b).

For these reasons, claims about the relations between the autochthony discourse and violence must be analyzed carefully, lest they obscure either the wider political-economic processes or the narrower particularistic projects that inform violence. Without such a careful analysis, and without detailed empirical studies of the occurrence, forms, perpetrators of, and motives behind violent acts, and the entire spectrum of discourses at play, the labeling of violence as “autochthonous” is difficult to substantiate. Finally, labeling violence as “autochthonous” is not only analytically imprecise, but may have similar essentializing effects as the labeling of violence as “ethnic” (Brass 1997). These observations indicate that while researching how autochthony discourses frame and feed into violence is a valid exercise, this kind of study entails a path of analysis that may be as slippery as autochthony itself.