Violence, Autochthony, and Identity Politics in KwaZulu-Natal (South Africa): A Processual Perspective on Local Political Dynamics

Abstract: “Political violence” is seemingly on the rise again in KwaZulu-Natal, South Africa. The violence that began in the 1980s and reached a peak in the transition period resurfaced before the local government elections in August 2016. Krämer provides a processual understanding of local dynamics of violence in the eThekwini Municipality and situates the current episode within the historical trajectory of violent conflict. He examines how exclusionary identities get activated in local elections and argues that underlying the violence between supporters of hostile political parties are conflicting forms of autochthonous belonging and contradictory ideas about what constitutes membership in a community.


Introduction
On August 28, 2016, two funerals were held in Inchanga in the Outer West region of eThekwini Municipality, KwaZulu-Natal; one was for a forty-yearold woman murdered shortly after the local government elections. Her funeral was attended by approximately three thousand South African Communist Party (SACP) supporters, almost all of them dressed in the red T-shirts of the party, along with the outgoing eThekwini mayor James Nxumalo, the SACP KwaZulu-Natal secretary Themba Mthembu, and the SACP deputy national chairperson Thulas Nxesi. At the same time, and at the other end of Inchanga, about one thousand African National Congress (ANC) supporters dressed in ANC colors assembled for the funeral of a thirty-eight-year-old ANC youth league member who was killed in retaliation for the murder of the woman. His funeral was attended by the ANC eThekwini regional secretary Bheki Ntuli, the ANC youth league provincial secretary Thanduxolo Sabelo, and former Ward 4 Councillor Boy Shozi, who was poised to gain an ANC proportional representation [PR] seat in the eThekwini Council in the elections. At the "red funeral," as it became known, the SACP deputy national chairperson called on the ANC to address infighting: "Confront the problem; be honest, open and frank and address the internal divisions. We are where we are because of the internal divisions, gate keeping [sic], and infighting" (SABC 2016b). James Nxumalo told the mourners, "It is the arrogance of the ANC that has got us where we are now. People are involved in activities that are foreign to the ANC but they are left to go on because they happen to be on the 'right' side of the factional divide" (Business Live 2016b).
Violent conflict was and continues to be an important aspect of South African democracy: the xenophobic riots in 2008 and 2015 (Landau 2010;Steinberg 2018), frequent acts of popular justice (Gordon 2004;Kirsch 2010), and the numerous service delivery protests in recent years (Kynoch 2016) are just a few examples. Mark Shaw and Kim Thomas (2016), however, point to the often-overlooked fact that, although homicide figures have consistently declined since 1994 (while nevertheless remaining on a comparatively high level), specific forms of violence such as targeted killings or "hits" have increased across South Africa, albeit unevenly.
One form of violent conflict seemingly on the rise again in KwaZulu-Natal is "political violence." This particular manifestation of violence started in the 1980s, reached a peak during the transition from apartheid to democracy in the early 1990s, and has persisted on a lower scale over the next twenty years. Acts of violence before and during the local government elections in August 2016 were reported in many regions (Mail & Guardian 2016), but according to David Bruce (2013:15), the problem of politically motivated murders was particularly severe in KwaZulu-Natal. More than one hundred local politicians were murdered between 2003 and 2013, with more than half killed between January 2011 and September 2012. As a result, there is renewed interest in documenting and theorizing violent conflict (see Gibbs 2017;Hickel 2015;Jarstad & Höglund 2015;Kelly 2018;Krämer 2015;Mathis 2013;Schuld 2013;Smith 2019;van Baalen & Höglund 2017).
This article contributes to current debates about violence, focusing on local dynamics of violent conflict in the Western periphery of eThekwini Municipality. The objective is to provide a processual understanding of the violence and to situate the current episode within the historical trajectory of violent conflict since the 1980s. In adopting a processual perspective (see Beek & Göpfert 2013;Trotha 1997), I focus on the manner in which violence is unfolding and what factors contribute to its persistence in the postapartheid democratic order. Underlying the violence between supporters of hostile political parties are conflicting forms of autochthonous belonging and contradictory ideas about what constitutes membership in a community. These different layers of conflict reinforce each other in specific circumstances (such as elections), and this mutual reinforcement produces severe and prolonged violent action.
The first part of this article provides a broad overview of violent conflict in KwaZulu-Natal and outlines the most common interpretations of the violence. Part Two describes the historical trajectory of local violent conflict since the 1980s, followed by a detailed account of the most recent phase of local violence in Part Three. The fourth part analyzes the past and current dynamics of violence and examines conflicting forms of identity and belonging. In the conclusion, I reflect on the significance of autochthony discourses in local violent conflict and on the renewed relevance of identity politics in South Africa.

Violent conflict in KwaZulu-Natal since the 1980s
The period of violent conflict which began in the early 1980s is called udlame in isiZulu, which may be translated as "the use of savagery in achieving aims" (Kelly 2012:xiiiff). According to Timothy Gibbs, udlame "conveys social chaos: a violence that 'initiates a spiral of revenge attacks'" (Gibbs 2017:377;quoted from Peires 2000:99). "Political violence" is the most commonly used expression in the academic literature and media reports, but a closer look at the main characteristics of this period of violent conflict reveals that we could also categorize it as a typical "small war" (Klute 2013). According to estimates by organizations monitoring violence in South Africa, about twenty thousand people were murdered and thousands became refugees in the course of the violence (Hickel 2015:2).
Violent conflict in KwaZulu-Natal escalated at the end of the transition period and took its highest toll on the eve of the first democratic elections: in March and April 1994, more than six hundred people were murdered in KwaZulu-Natal (Louw 1995:28). The violence facilitated the gaining of territorial control and the securing of constituencies for the first democratic elections (de Haas & Zulu 1994:442-43). This resulted in so-called "no-go areas" and the expulsion of perceived enemies by force. Karl Von Holdt (2013:592) reminds us, that "the new Constitution for South Africa was born out of the most extreme violence in the history of the country." Although violent conflict has slowly decreased since the mid-1990s, it never ceased completely. In the midst of continuing acts of violence, however, a peace process was started by the ANC and Inkatha Freedom Party (IFP) leadership in about 1996. Another factor contributing to less violent conditions was that neither the IFP nor the ANC could win an absolute majority in the provincial elections in 1999, and they had to form a coalition in the provincial parliament. However, the ANC became the dominant party in KwaZulu-Natal after 2004, and in 2014, the IFP won only 10.86 percent of the vote, whereas the ANC won 64.52 percent.
According to the "political rivalry thesis" (see Krämer 2007:13), violent conflict is the outcome of a rivalry between political parties and their leadership-the ANC and its affiliated supporters of the United Democratic Front (UDF) on the one hand, and the IFP, assisted by the security forces of the apartheid regime, on the other. Hence, power struggles between political parties sparked violent action and eventually led to the division of almost every African community in KwaZulu-Natal between ANC supporters on the one hand and IFP followers on the other (Aitchison 1998;Griggs 1997;Gwala 1992;Johnston 1996;Taylor 2002). In the speeches held at the funerals mentioned at the beginning of this article, the aspect of political struggle becomes clear from the frequent references to "infighting" and "internal divisions" within the Tripartite Alliance. Gibbs (2017), however, points out that udlame had various sources, and he shows that intergenerational conflicts structured violent action in rural areas of the Natal Midlands (see also Carton 2000). Violent conflict spread along migrant routes and "was entwined with and exacerbated by patterns of conflict associated with the criminality of uncontrollable young migrants" (Gibbs 2017:378). Thus, "patterns of political violence were often associated with the juvenile criminality of migrant gangs. It was these shape-shifting forms of violence . . . that made the udlame so brutal" (Gibbs 2017:381). This assessment highlights at least one crucial aspect of violent conflict, namely, that the boundaries between politics, economics, crime, and private matters are usually blurred.
An alternative interpretation points to the increasing significance of patron-client relations within the ANC (Lodge 2014). According to Alexander Beresford (2015), gatekeeping is rife within ANC ranks, that is, the control of access to state resources and their distribution to one's clientele to enhance personal power. Gatekeeping is exclusionary and conflictual but not necessarily violent: "big men" or "strongmen" politics (Mathis 2013;Utas 2012) refers to the explicitly violent side of gatekeeping and illustrates the close relationship between violent action and resource control in local politics. The funeral speakers also raised these issues when they interpreted local violence as a result of gatekeeping within the ANC.
While these interpretations of udlame have their merits and are indeed relevant, the focus of this article is on another facet that often goes unnoticed in public and academic attempts at making sense of the violence: the interplay of political and cultural identities and the rising significance of identity politics in more general with all its perilous implications. Interestingly, the funeral speeches mentioned above are more or less silent about this aspect. However, the case study presented here demonstrates that spatial and cultural notions of belonging (such as autochthony discourses) play an important part in the dynamics of violence. These identities are partly the result of violent conflict and frequently get activated in its course. Therefore, an in-depth understanding of the violence requires a fine-grained analysis of the construction and accentuation of identities that are not only-or maybe even not primarily-political.

The historical context of local violent conflict 1
The eThekwini (Durban) Municipality is one of the fastest growing cities in sub-Saharan Africa, with an estimated three and a half million inhabitants at present. The municipality is divided into 110 wards; its council consists of 219 councillors (110 ward seats and 109 PR seats) (eThekwini Municipality 2016), and the ANC won 56 percent of total votes in the 2016 local government elections (Independent Electoral Commission 2016). Although the ANC hegemony came under attack in the 2016 elections in several municipalities nationwide, eThekwini and KwaZulu-Natal in general remain ANC strongholds in South Africa for the time being (Engel 2016:111).
Inchanga (comprising the main part of Ward 4) is one of the numerous and sprawling periurban settlements on the western periphery of eThekwini, which are contributing immensely to the rapid population growth of the municipality. About 38,000 people lived in Ward 4 in 2011 and about 90 percent of those inhabitants were Zulu native speakers. Sixty-four percent of the population were younger than thirty years old, and the median age was twenty-three. Inchanga is characterized by its chronic instability and insecurity, which results from several factors such as recurring violence, (violent) crime, high unemployment rates (28.4 percent formally employed, 46 percent "not economically active"), poverty (average annual household income ZAR14,600), and extensive immigration, mostly from rural areas (Wazimap South Africa 2016). Divisions between people who perceive themselves as the original inhabitants of Inchanga (so-called "autochthons") and immigrants who have settled in the community during the last thirty years play a significant role in the dynamics of violence.
Violent conflict in Inchanga began in 1985: similar to events in other communities of the greater Durban region at that time, young comrades protested and mobilized against the apartheid regime, the KwaZulu government, and its local political institutions. Schools became a particularly important site for political mobilization. Inkatha's school curriculum was

Map 1. KwaZulu-Natal and eThekwini Municipality
Source: The Cartography Unit, School of Environmental Sciences, University of KwaZulu-Natal rejected by the pupils who became aware of other interpretations of South African political history than Inkatha's Zulu ethnonationalist version. Youth resistance was not only directed at the unjust apartheid system but also at the patriarchal order. Until the youth took over control of Inchanga, young men were told by older males how to behave and what to do or not, both within families (fathers versus sons and older versus younger brothers) and in the public sphere. This patriarchal order was turned upside down after the comrades started to rebel.
Violent conflict was also related to differences in socioeconomic status and to conflicts between landowners and tenants over rent payments. Such conflicts largely coincided with existing political divisions because almost all local councillors were landowners and most of them owned some kind of local business, whereas the majority of comrades originated from landless and poorer families. The political mobilization of the comrades was directed at the political establishment and the small class of successful African entrepreneurs who complied with the political status quo. Local violent conflict was, therefore, not only about political and generational conflicts but also the consequence of conflicting economic interests and income inequality. The Inkatha-aligned local elite called in security forces of the apartheid regime in order to suppress youth protests by violent means, but they were unsuccessful, and in 1989 the ANC-aligned youth gained the upper hand in the local power struggle. After the comrades had seized political control, they launched so-called block committees within demarcated sections of the community. The main functions of these block committees were to regulate conflict and to provide security and order. Block committees were transformed into ANC sub-branches in the transition period, but local identities became firmly established in the post-apartheid era when some sections and their leading "strongmen" started to struggle for political control by violent means.
The dynamics of violence in the post-apartheid era have been intricate and complex. 2 Basically, local violence shifted to fighting between different ANC-affiliated sections after 1994, and violent conflict often escalated around elections and primary elections of party candidates. The sections that got the upper hand in these violent power struggles controlled the local ANC, while those which were defeated often split away from the party.
The following examples of two specific phases of local violence illustrate these complex political dynamics.
In the first example, violent conflict started between young men from two different local sections after a football match in 1995: one section was largely inhabited by immigrants who had fled from violent conflict in other areas of KwaZulu-Natal in the 1980s, while the other was mainly populated by people who claimed to be autochthonous inhabitants of Inchanga. The first-versus-latecomers issue was an important aspect of the violence that followed, and about twenty people were murdered between 1995 and 1996. The violence signified a divide within the local ANC, and the football match was merely the spark for what subsequently became a local power struggle.
Violent conflict escalated after a mass killing in December 1996, when a group of gunmen dressed in camouflage uniforms entered the "immigrants' section" and killed six people execution-style. In the aftermath, the violence between members of the two sections shifted to violent conflict between political parties; the "immigrants' section" openly turned away from the ANC and founded a local IFP branch, allegedly for protection. In December 1999 and after several more killings, the Inchanga ANC and IFP signed a peace agreement that was the culmination of two years of negotiations. Thousands of people from all sections of Inchanga congregated on a sports field to attend the peace rally. A tent was erected, cattle were slaughtered to symbolize the return of peace, and residents of hostile sections and political parties celebrated together.
The peace did not last for long, however. The second example illustrates another change in local violent dynamics. A few months before the local government elections in December 2000, a primary election of ANC candidates took place and the nomination process was accompanied by severe conflicts; two ANC leaders from different sections contested the primaries. One of them was born in Inchanga and referred to himself as an "intsintsi," while the other was an immigrant from a nearby rural area who had come to Inchanga in the course of udlame in the 1980s and was labeled an "ibinca." Intsintsi denotes an "indigenous tree" and thus refers to the supposedly "original" inhabitants of Inchanga, whereas ibinca-a term which is generally used in a derogatory way-could be translated as "traditionalist"; to be a traditionalist, however, also serves as a common and positive self-description among many "allochthons." Therefore, in an already tense political climate, place of origin and cultural identity became crucial issues in local politics. Those who claimed to be "autochthons" argued that the local ANC had been hijacked by "allochthons" and strongly opposed the idea that an "immigrant" could be in political control of the ward. At about the same time, violent action started to intensify. From August until early November 2000, at least thirteen killings took place while only two suspects were arrested.
Eventually, a mass killing in November 2000 signified the peak of the escalating violence: five people were shot dead and one was seriously wounded by a group of armed men dressed in South African National Defence Force uniforms. The perpetrators entered a section predominantly occupied by "autochthons" at night carrying AK-47 and 9mm firearms and shot people at random along their way. All of the victims were grassroots ANC members, one of them a brother of the section leader and current ward councillor (Daily News 2000;Natal Witness 2000;The Mercury 2000). A few weeks after the mass killing, the local government elections took place and the ANC candidate (still labeled an "immigrant" and "traditionalist") received about 77 percent of the votes. However, the rate of voided ballots at the central voting station was extraordinarily high, at 13 percent of all votes (Independent Electoral Commission 2001). Two days after the elections, the ANC candidate was arrested and accused of being the mastermind behind the killings in November. However, he was released on bail in early 2001, sworn in as ward councillor subsequently, and won re-election in 2006 and 2011. The leaders of the "autochthonous" sections withdrew from the local ANC executive committee, and they together with their followers subsequently positioned themselves as SACP supporters in opposition to the local ANC branch.
The divisions in the local ANC have persisted until the present; the strictly separated funerals of August 2016 symbolize this extreme divide. Underlying the hostility between ANC and SACP supporters are conflicting spatial and cultural forms of belonging which have developed in the course of local violence since the early 1990s. These spatial and cultural identities have been reactivated and recombined with political identities at election times. In other words, the division of the ANC and the accompanying violent events in 2000 were a precursor to events that occurred before and after the local government elections of 2016.

Election violence in 2016
The outburst of violent conflict in 2016 is incomprehensible without adopting a historical perspective on local developments. In addition, it is important to include a perspective on broader political processes in KwaZulu-Natal. Interestingly, the patterns of conflict that have affected Inchanga since the turn of the century appeared on the scene in provincial politics some years later.
In November 2015, the provincial ANC chairperson and former Premier of KwaZulu-Natal, Senzo Mchunu, was defeated in the provincial ANC chairperson elections by Sihle Zikalala. In the aftermath, an ally of Zikalala named Zandile Gumede unseated James Nxumalo as ANC chairperson for eThekwini, which is the largest ANC region in South Africa and is thus "regarded as strategic both in terms of numbers and its ability to influence policy direction" in the country (Mail & Guardian 2015). Nxumalo, who hails from and lives in Inchanga, was also replaced as eThekwini mayor by Gumede after the local government elections, although he maintained his office as provincial chairman of the SACP in KwaZulu-Natal. The election process for the eThekwini ANC chairperson was chaotic (to say the least) and rife with accusations of gatekeeping and factionalism. All these regional and provincial power struggles within the Tripartite Alliance were a prelude to the severely contested elections of the ANC President at the National Conference in December 2017. What was even worse was that these supralocal conflicts set the scene for renewed violent action at the local level.
In January 2016, two killings occurred in Inchanga during the nomination of candidates for the local government elections. These murders were reported in the regional and national news against the background of a number of violent incidents across the country in the run-up to the elections. The killings took place during a pre-election rally of the local SACP branch. Eyewitnesses reported that at least four more people were wounded when about fifty gunshots were fired from four cars speeding toward the soccer field where the rally took place. A 68-year-old retiree was executed by the attackers after he tried to flee and fell. The other victim, a 38-year-old man, was the driver of one of the attacking cars; the crowd pulled him out of the vehicle when he tried to speed up, and he was murdered immediately.
At the funeral of the retiree, thousands of mourners voiced their anger and frustration at the ruling ANC, and the deceased was declared a hero and martyr by SACP leaders. They claimed that it was he who took a bullet intended for the outgoing eThekwini mayor and provincial SACP chairman James Nxumalo, or for his relative and local SACP leader Petros Malombo Nxumalo. Five men (two of them related to the then-ANC Ward Councillor Boy Shozi) were arrested but released on bail a few days later, which resulted in a standoff between hundreds of SACP and ANC supporters in front of the court house, the latter celebrating the release of the accused.
Campaigning for local government elections became even more tense in the following months, and high-ranking political figures had a stake in fanning the flames of violence. For example, South African President and ANC President Jacob Zuma and the SACP general secretary and national Minister of Higher Education and Training Blade Nzimande visited Inchanga separately to support their local party candidates (Times Live 2016). During his visit in July 2016, Zuma reportedly said that "independents are treacherous witches . . . [but] they could not be killed because 'we are now living in a democratic system'" (Independent Online 2016b).
On election day, the local SACP leader Petros Malombo Nxumalo won 65 percent of the ward vote as an independent candidate (as compared with 29 percent for the new ANC candidate) and thus became the new ward councillor of Ward 4 (Independent Electoral Commission 2016). A closer look at the election results reveals that the ANC nevertheless won 76.5 percent of the proportional representation vote; since the SACP does not contest elections as a separate party but rather as part of the alliance, its representatives stand for the ANC and are placed on the ANC list. Hence, the overwhelming majority of Inchanga citizens supported the ANC, but the two factions obviously disagreed as to who were the legitimate local representatives of the Tripartite Alliance.
Local violent conflict flared up again in the immediate aftermath of the 2016 elections. On August 21, a 40-year-old woman and staunch SACP supporter was shot in the head, execution-style, outside a tavern in Inchanga. A son of James Nxumalo had been with the murdered woman at the tavern and claimed that he had recognized the drunken and reckless killers who fired a shot in their direction which killed the woman. When Nxumalo fled together with two of his brothers, the killers allegedly shouted "they wanted the Nxumalo boys" (Independent Online 2016c). However, local SACP leaders claimed that the murdered woman was a witness to the murder case of January 2016 and that the killers had wanted to kill her and the Nxumalo son at the same time.
The day after the murder of the female SACP member, an outraged crowd of rioting SACP supporters took revenge and lynched a 38-year-old ANC youth league member who happened to be in front of his house when the crowd passed by; according to a relative, he was first shot and then stoned to death by the mob. The crowd also burned down five houses of ANC supporters in the area, blockaded the roads with burning tires, and chanted SACP slogans (Daily News 2016a). At the end of the day, the South African state had severe difficulties in gaining control and stopping the rioters, although more than one hundred security police officers were deployed in Inchanga. Eventually, forty-one people were arrested by the police for public violence, but they had to be released the next day for lack of evidence (Business Live 2016a; SABC 2016a). Four people were arrested for the murder of the 40-year-old woman in late August and were arraigned in court in mid-September, but no one had been arrested for the murder of the 38-year-old ANC youth league member by the end of 2016.

Making sense of violent conflict: individualized versus collectivized autochthony
At first glance, violent conflict appears to be occurring between supporters of hostile political parties. The political divides seem to be so entrenched that even funerals offer no opportunity for establishing or reaffirming cross-cutting ties. The funerals held in August 2016 were completely separated between ANC and SACP supporters, in terms of participation as well as symbolism. The resentment and mistrust became clear when the proposal by the provincial MEC for Transport, Community Safety and Liaison to hold a joint memorial service was instantaneously rejected by the victims' families: "We can't accept such a stupid and dangerous proposal. We can't be with those people in the same place . . . If we were to be in the same space with the other side, all hell would break loose" (Daily News 2016b).
Looking at it again, however, raises the question of who are "those people" from "the other side." The divisions between SACP and ANC supporters turn out to be dangerous in specific circumstances such as local elections, but underlying these political affiliations are other conflicting forms of belonging that are less visible in public but nevertheless crucial for a comprehensive understanding of the violence. Upon closer examination, two contradicting ideas about community membership and access to political power reappear in most instances of local violent conflict: notions of belonging based on the principle of the place of birth on the one hand and of a shared culture on the other. That is, although basic ideas about what constitutes a community in the first place are contradictory, both notions are expressions of autochthony discourses which flourish in contemporary Africa and beyond (see Ceuppens & Geschiere 2005;Pelling 2009).
Autochthony literally means "born from the earth" (Rosivach 1987:294). The concept originated in classical Athens, where it mattered particularly in the fifth and fourth centuries BCE. Interestingly and comparably with the present, autochthony was (and still is) connected with democracy, "for if we are all born of the same earth, we are all brothers and sisters, all alike" (Pelling 2009:471). It was Plato, who (in his Republic) proposed the 'myth of autochthony' and the idea that Athenians should think of themselves as brothers. "In this sense autochthony can be egalitarian" (Rosivach 1987:303), but only insofar as non-citizens, that is, those who have no chthonic origin and allegedly immigrated later, are excluded from full membership and democratic participation. Vincent Rosivach (1987:305) argues that the idea of autochthony became part of democratic ideology "asserting the political equality of all citizens and the superiority of even the humblest citizen to any non-citizen." Hence, the concept of autochthony was and is exclusive and inherently conflictual. Peter Geschiere (2009:224) maintains that the exclusionary tendency of autochthony discourses and "the quest for purity" render autochthony highly problematic and may open the floodgates to violence. Morten Bøås (2009:34) is more cautious and claims that while autochthony discourses can exclude and promote conflict, they may also have an integrative effect and stimulate reconciliation.
Olaf Zenker (2009Zenker ( , 2011, however, argues that more recent conceptualizations of autochthony are often inconsistent. In some instances, autochthony is referred to as a form of ethnicity (Geschiere & Nyamnjoh 2000); in other instances, autochthony is conceptualized as being different from ethnicity, "an identity with no particular name and no specified history, only expressing the claim to have come first" (Ceuppens & Geschiere 2005:387). Zenker's alternative proposal is to distinguish between two ideal-typical logics: "individualized" versus "collectivized" autochthony. In both ideal-types, three constitutive elements (individual, territory, group) are causally linked, and the distinction is "not based on different ingredients but on different causal directions" of the individual-territory-group triad (Zenker 2011:72).
Individualized autochthony, "links the individual, territory and group in such a way that shared culture and/or descent ultimately follow from place of birth and/or residence within the same present" (Zenker 2011:65). That is, the linkage between individual, territory, and group rests on the notion of the "presence of individuals," and a common place of birth and/or residence constitutes, first, the territory and, second, the group (Zenker 2011:71). In collectivized autochthony, "an individual-through shared culture and/or descent-first causally links up with a group, which at some point in its proclaimed past connected itself-through the establishment of land(ed) rights for its membersto a territory" (Zenker 2011:72). In comparison with individualized autochthony the causality is inverted, and the proclaimed linkage between the three elements is based on the idea of a "past of groups." Shared culture and/or descent link the individual first with the group, and second with the territory.
What does this mean for the analysis of violent conflict in the case study presented here? I contend that what appears at first sight to be a conflict between supporters of political parties is actually based on conflicting forms of belonging on a different level by closer examination. Underlying the dynamics of violence in the past three decades are contradictory ideas about what constitutes membership in a community; these contradicting ideas largely fit with the ideal-type distinction between individualized and collectivized autochthony. An identity based on the notion of place of birth which prevails among "autochthonous" inhabitants of Inchanga is opposed to an identity based on the notion of a shared culture which is more typical for "allochthons" who have immigrated in the course and as a consequence of udlame. The emic terms of intsintsi (indigenous tree) versus ibinca (traditionalist) are indicative of these contradictory ideas and symbolize the conflicting identities at stake: the first term links the individual with a territory which then constitutes group membership; the second term refers to the idea that a shared culture is perceived as constituting a group which then claims access to a territory. In other words, the notion of "indigenous tree" illuminates the causal link between individual and territory through the common place of birth, whereas the notion of "traditionalist" points to the link between group and individual through the perception of a shared culture.
For example, an excerpt from an interview with an "immigrant" who settled in the community in the 1980s illustrates the notion of "collectivized autochthony. In contrast, an excerpt from an interview with a young man who was born in Inchanga points to the perceived differences between "autochthons" and "immigrants" and thus exemplifies the notion of "individualized autochthony." The "very indigenous people" of Inchanga, those who are "born from the earth" in Rosivach's (1987) words, differentiate themselves from the latecomers, who are labeled the cultural "others." Interviewee B: You see, the history here at Inchanga has taught us that in actual fact the people who have been like the indigenous people of Inchanga . . . adhered to these Western or modern values of looking after themselves . . . But now these people ["immigrants"] came from areas under the authority of kings . . . So their ideas were more into tribalism and were more culturally inclined. So I think there's that difference . . . So now at one stage it was then assumed or it was actually consensus that these people, no matter what political orientation or affiliation they are falling under, but at one stage they would come together as a sub-culture or whatever I may call it. They will then come together and then plot against the very indigenous people of this area. 5 The categories presented by interviewee B closely resemble what Philip Mayer (1971) argued about the "School" versus "Red" distinction in the Eastern Cape in the apartheid era, that is, the perceived differences between those who were allegedly keen on adopting and identifying themselves with "modern" norms and values and others who largely resisted these innovations and showed a "conservative impulse" (Marris 1986).
Two other interview excerpts suggest that while conflicting notions of "individualized" versus "collectivized" autochthony are common and have local relevance, they translate into violence in specific circumstances only. The following excerpt is from an interview with two middle-aged men on the significance of cultural identity in local politics after the mass killing in November 2000.
Interviewee C: . . . these people came from different areas with their practices from these areas, where they had lived before. These traditional practices. And as they were in Inchanga, they continued with their practices. Then they were referred to as amabinca.
Interviewee D: And we from Inchanga, we were referred to as amagula, Indians.
Interviewee C: Amagula, that is Indians. The reason was that these traditional people, who came from different areas, they thought that Inchanga people turned away from their traditions. Therefore, they lost their identity . . . Then there was that stigmatization where people were called by names and threw names at each other . . . You know, politics were just used to justify the war, really . . . Somehow, when they were in conflict, then politics started to trickle in as well and the whole thing was done in the name of politics. 6 What becomes clear from these statements is that "stigmatization" between "autochthons" and "immigrants" was common, and that the latter accused the former of a loss of traditions and cultural identity. And in stark contrast to the publicly voiced interpretations of the violence in 2016, it is argued here that politics "trickled in" only at a later stage and were "used to justify" local violence. From this point of view, violent conflict was allegedly "done in the name of politics"-but was actually about the construction of exclusive identities and the politicization of the constructed differences.
This last excerpt, from an interview with two other men, highlights the link between autochthony discourses and democratic elections: Interviewee E: Izintsintsi are not ANC . . . though they are members of the ANC, but they are not part of the structures of the ANC. The people who are part of the structures are the people referred to as amabinca. Right now. And the people who are SACP are people who we refer to as izintsintsi, yes . . . MK: Is the difference between izintsintsi and amabinca very important here?
Interviewee E: It's not that important. It's important when it comes to elections and the key positions because some people now are viewing politics as employment agency . . . If people were employed, they would have another way of viewing things . . . That is why we are saying, comes another election we might have the problem of violence because of these divisions.
Interviewee F: And if there is violence, I think it might be because of if you are an ibinca or you are an intsintsi. It might be like that. 7 The contradiction made by interviewee E is interesting and revealing: "it's not that important" versus "it's important when ...". It suggests that perceived cultural differences are rather insignificant in everyday life, but they turn out be dangerous under specific circumstances, that is, before and during democratic elections, when conflicting notions of autochthony get activated. In other words, when gatekeeping and local power struggles amalgamate with the quest for exclusive cultural identities as "autochthons" versus "immigrants," then violent conflict flares up and becomes vicious and protracted. Violence thus has different layers that reinforce each other: it is based on the perceptions of a division (in terms of identity) between "indigenous" versus "traditionalist"-which is more or less hidden and not articulated in public-and the (political) divide between SACP versus ANC supporters-which is the publicly voiced interpretation of violence.

Conclusion
A processual understanding of violent conflict helps resituate local election violence in 2016 against the historical trajectory of udlame in KwaZulu-Natal since the 1980s. When research incorporates the interplay between different forms of belonging and how these identities get activated in specific circumstances, we find that what is usually perceived as "political violence" in public discourse includes different and intricate layers of conflict. In other words, the mutual reinforcement of exclusive forms of local belonging and hostile political identities generates protracted violent conflict, and an in-depth understanding of the violence requires a fine-grained analysis of the construction and accentuation of these identities. In the case study presented here, underlying the violence between local ANC and SACP supporters are contradictory ideas about autochthonous belonging (individualized versus collectivized autochthony) that are particularly perilous before and during democratic elections. These autochthony discourses are often linked with the notion of purification, and quite frequently, demands to track down "malicious elements" and to "cleanse" the community are voiced. It is sure instructive that the SACP National Deputy Secretary told the mourners at the funeral of the pensioner in January 2016 that "[i]t is our duty to clean the ANC. It is rotten. It is us who should clean it" (Independent Online 2016a).
In more general terms, such particularistic and exclusionary forms of belonging are exemplary for what has been described as the rise of the local since the early 1990s (Trotha 2005) and illustrate that "the massive movement everywhere is to go back to identity politics" as Bruno Latour (Latour et al. 2018:594) claims in a recent interview. According to Jason Myers (2008:84), "modern forms of identity politics" are often rooted in universalistic principles, but there are other forms "in which the universalistic demand for an end to particularistic exclusion is gradually transformed into a particularistic claim to exclusive rights on the basis of identity." 8 Localism and ethnonationalism are examples for exclusionary and particularistic identity politics which link political representation with specific cultural or ethnic identities. The notion of autochthony as "shared culture" (as in collectivized autochthony), which is one important feature of local violent conflict discussed in this article, fits to the general reorientation of the national ANC during the Zuma presidency. Under his leadership, the ANC shifted more and more to identity politics on an ethnic and cultural basis, and moral legitimation was increasingly sought from neotraditionalist representations of culture, in Tom Lodge's words (2014: [1][2]. This strategy is reminiscent of the ethnic mobilization by the Inkatha movement from the late 1970s onwards (see de Haas & Zulu 1994;Maré 1993;Waetjen 1999) and with its shift to ethnic and cultural identity politics, the ANC was able to outdo and replace the IFP in many rural areas of KwaZulu-Natal; this, however, came at the expense of alienating a substantial part of its following in urban areas of South Africa. Whether this may be the beginning of a new and politicized urban-rural divide, as Ivor Chipkin (2016:223) suggests, and may provoke further violent conflict in KwaZulu-Natal remains to be seen.