Introduction

The incident described in the quotation above occurred at the start of 2013. It was related to me by a taxi driver who, at the time, was the head of a local anti-gang initiative. In Khayelitsha the words “gangster” or “gangs” refer to young people, boys. This initiative involved a group of taxi drivers who, at the request of teachers and community members, broke up fights between rival gangs. In this instance, the children were armed with both pangas Footnote ¹ and amulets, the latter having been allegedly provided by the “witchdoctor.” The taxi drivers, armed with sjamboks,Footnote ² beat the children, confiscated their weapons and warned them not to fight again. After that 50 taxi drivers, in five mini bus taxis, drove to the witchdoctor's house. They forced him to leave the area.

Background

Khayelitsha, meaning “new home” in Xhosa, is situated approximately 30 km from Cape Town's city centre. Established in 1983, it was the last area of the city to be formally reserved for black people, prior to the repeal of the 1945 Native Urban Areas Act. This legislation required urban authorities to establish separate residential locations (townships) for “natives” and to exercise control over “native immigration” into urban areas. Its peripheral location was characteristic of apartheid planning and it was developed as a “dormitory suburb,” consisting of low-income people who commuted to Cape Town for work. At the time of its establishment a tin hut was erected for each family, along with a bucket toilet, one tap per four plots, high-mast street lighting and, a refuse removal service (Reference O'Regan and PikoliO'Regan and Pikoli 2014: 30–33).

Khayelitsha is ethnically and linguistically homogenous with 98.7 percent of the population being black and speaking isiXhosa. Youth unemployment is more than 50 percent and most residents have not completed grade 12 (ibid.). Almost half of Khayelitsha's households live below the food poverty line and there is a significant difference between income distribution within Khayelitsha vis a vis the Cape Town population.Footnote ³ More people in Khayelitsha live in shacks than in formal dwellings (Reference O'Regan and PikoliO'Regan and Pikoli 2014: 37) and it is more crowded than other townships (Reference FrithFrith 2013).Footnote ⁴ Informal settlements, many of which are situated in low-lying flood prone land, exist alongside established suburbs (Affordable Land and Housing Data Centre, undated: 91). Residents live in tin shacks, with many accessing electricity illegally, sharing communal water taps and relying on inadequate sanitation arrangements (such as outside portable toilets).

The air in Khayelitsha is punctuated by the pungent aroma of meat cooking on outside fires, coupled with the stench of blocked drains and overflow pipes. Thin dogs, seeking out food-scraps, roam the pedestrian crowded streets. Minibus taxis, spaza shops, shebeensFootnote ⁵ and other small businesses (offering services such as tyre repair, hair braiding and fruit vending) add to the sense of hustle and bustle. The infamous porterloos and outside toilets that line the perimeters of the informal settlements, together with mounds of uncollected garbage, contribute to a general sense of vibrant decay, what the Western Cape provincial government refers to as a “zone of poverty and unemployment” (Department of Community Safety 2009: 20).

The three police stations in Greater Khayelitsha—Lingelethu West, Harare and Site B—have far less than the average 283 police personnel per 100,000 residents. At 111.32, Harare has the lowest number of police personnel to population in the Western Cape. Camps Bay and Wynberg, both predominantly white middle to upper class suburbs, have a police to population ratio of 959.51 and 852.57 per 100,000, respectively (Redpath, cited in Reference O'Regan and PikoliO'Regan and Pikoli 2014: 315). In 2012–2013, there were 354 murders in “Greater Khayelitsha”Footnote ⁶—the highest numbers of murders for any township in South Africa. The combined figures for attempted murders, sexual offences, assault with intent to do grievous bodily harm and robbery with aggravating circumstances also topped the charts (De Kock, cited in Reference O'Regan and PikoliO'Regan and Pikoli 2014: 44).

Research Question and Core Argument

This article asks what crime prevention looks like for residents in Khayelitsha's informal settlements. It engages with the idea of vigilantism and hybrid policing formations, analyzing the overlaps and intersections between legal community-based crime prevention initiatives, such as officially sanctioned neighborhood watches, and local “punitive practices” (Reference GrayGray 2013: 518) when residents break the law, or overstep the bounds of legality.

The literature on legal pluralism has engaged extensively with the idea of legality and legal consciousness (Reference de Sousa Santosde Sousa Santos, 1987; Reference Ewick and SilbeyEwick and Silbey 1992; Reference MerryMerry 1988). I use the term “legal” to refer to technologies that are sanctioned by official (state) law. The core argument I make is that, at times, particularly in poor areas where the stateFootnote ⁷ is absent and encourages citizens to take responsibility for their own crime prevention, the boundary between “legality” and coercive illegality collapses in on itself. Flowing out of this, I argue that responsibilizing the community, via the state's discourse of “community based crime prevention,” without appropriate state financial and logistical support, results in a dangerous situation, a governmentality gone wrong. Thus, the notion of voluntarism, that is central to official discourse on crime, is particularly problematic when applied in poor communities with high rates of unemployment and crime. Indeed, as I argue, the very notion of “community” is problematic, particularly in “distressed neighborhoods” (Herbert, cited in Reference DelgadoDelgado 2008: 27) where there is a paucity of civil institutions to counter the absent state (Reference Comaroff and ComaroffComaroff and Comaroff 2004; Reference AminAmin 2005; Reference PillayPillay 2008). Although this case study focuses on South Africa, my findings about the dangers of harnessing the “community” in the service of crime prevention are informed by and, applicable to, other contexts of inequality. Reference GarlandGarland's (1996: 451) seminal definition of “responsibilisation” as “central government seeking to act on crime ….indirectly,” by seeking to activate “action on the part of non state agencies and organisations” is a useful framework within which to situate this “turn to community” in South Africa, except that here, it is coupled with the 1980s rhetoric of grassroots revolutionary populism (when street committees were promoted as tools of popular justice by the liberation movement).

The South African state has adopted this populist responsiblization discourse together with a punitive discourse on crime and criminality. It also deploys a distancing maneuver in regard to those who “take the law into their own hands” by setting up a binary between the law and its other—the unlawful or illegal. This is disingenuous because, as I argue, encouraging citizens to take responsibility for their own safety (including making citizen's arrests), whilst simultaneously promoting harsh crime policies, creates the space for illegal vigilante style actions to emerge in the shadow of legal crime prevention initiatives.

Methodology

This article is based on research conducted between 2012 and 2015. The methodology consisted predominantly of selective in-depth semistructured interviews with 59 people. Interviews averaged an hour and were conducted in English, apart from with four victims of vigilantism (conducted in isiXhosa via a translator).

I followed an inductive process, starting by interviewing five members of the Social Justice Coalition (SJC), an activist organization based in Khayelitsha. The SJC was a driving force behind the establishment of the Commission of Inquiry into Policing in Khayelitsha (hereafter referred to as “the Commission”), established by the Western Cape Premier in 2012. The Commission was mandated to investigate “Allegations of Police Inefficiency and of a Breakdown in Relations between the Community and the Police in Khayelitsha” as allegedly indicated by a “spate of vigilante attacks” (Reference ZilleZille 2012).Footnote ⁸ During this initial research phase I also observed three weeks of proceedings in a High Court case—where six people were charged with kidnapping and murdering four youth who were alleged to have stolen the television set of one of the accused.Footnote ⁹ I supplemented my initial interviews and observations with documentary sources, including newspaper reports on vigilantism and, documentation made available by the Department of Community Safety (DOCS). I also attended SJC meetings around its now defunct mass “Campaign for Safe Communities” and those that it organized to lobby for the Commission. This enabled me to gain a sense of popular discourse on vigilantism. In terms of this discourse, vigilantism occurred when people “took the law into their own hands” due to police incompetence. Therefore, so the argument went, a more efficient police force would solve the problem of vigilantism.

After this initial phase, I adopted a more targeted sampling strategy and refined my interview questions. I used two residents to assist me in tracking down individuals who had either been referred to in previous interviews or from incidents that I had read about in the media. I utilized a snowball approach with my initial analysis determining where to go, and what to look for, in my next round of interviews (Reference MassoudMassoud 2015). Since most incidents of “mob justice” took place in informal settlements I shifted my focus to these locations. Interviews were conducted with members of various informal community patrol groups,Footnote ¹⁰ neighborhood watches,Footnote ¹¹ street and area committees from 10 areas,Footnote ¹² four residents who self-identified as “vigilantes,” four mothers whose children were beaten to death, “ordinary” residents in 10 informal settlements,Footnote ¹³ members of the South African National Civic Association (SANCO) and, a senior member of the Congress of Democratic Taxi Associations. After every interview I gave my telephone number to interviewees and requested them to give me a missed call should they wish to share any further information. I would then call them back and arrange for a follow-up interview. I paid a small stipend in compensation for the time spent talking to me (Reference Maxfield and BabbieMaxfield and Babbie 2010). Interviews were conducted in shacks, at shopping centers, community halls, inside my car, courtyards and restaurants—wherever the interviewee felt comfortable talking to me. In all instances I traveled to Khayelitsha to interview participants.

Interviewees and witnesses consistently referred to street and area committees and their role in resolving (and sometimes exacerbating) local conflict. I, therefore, sought to interview as many street committee members as possible. Street committees have a link, tenuous as it may be, to the African National Congress (ANC) aligned SANCOFootnote ¹⁴ and sometimes street committee members were also executive committee members of SANCO branches. As discussed in the paper street committees are involved in both popular justice initiatives as well as (to some extent) coercive crime prevention and punishment activities. They also represent the vague, amorphous and elusive “‘community” that is central to official discourse on crime and on which community-based crime prevention talk hinges. As such they constitute an important part of the story told here.

I did not record interviews because I wished to allay potential fears about confidentiality and was concerned that recording would undermine the establishment of a rapport with interviewees (Reference Irvine and GaffkinIrvine and Gaffkin 2006, Reference NewburnNewburn 2007). Instead, I sought permission to take notes during interviews.Footnote ¹⁵ I typed these up every evening and coded the responses by reading and rereading my notes, underlining words that consistently appeared and highlighting recurrent issues. Thus, I sought a continuous interplay between data collection and analysis. I identified the ways that interviewees engaged in crime prevention, whether they perceived these to be effective, the interaction with lawful state sanctioned initiatives, the extent to which coercion or violence was deployed and, the organizations or formations involved.Footnote ¹⁶ I identified representative quotes from the interviews as a means of highlighting my main findings and to describe a particular situation in the most graphic way possible (Reference Beckett and EvansBeckett and Evans 2015). The six empirical sections were constructed around the main data that emerged from the interviews and fieldnotes—highlighting key words and issues that consistently emerged and were all, in some way, representative of hybrid combinations of law and its other.Footnote ¹⁷

Analytical Frameworks

The focus of the article is not on the intensely violent spectacle of “mob justice,” where suspects are killed, sparking widespread media coverage and outrage amongst certain sectors of South African society. Instead, it is concerned with more ubiquitous, hybrid formations, that also fall on the “continuum” (Reference Huggins and HugginsHuggins 1991: 8) of vigilantism, and which, to some extent or other, the formal (political) state, and its agents, tacitly tolerates. These community level formations are often overlooked in academic literature, precisely because they are “labile” (Reference AbrahamsAbrahams 1998: 7), ephemeral, difficult to research and, may overlap with legal initiatives.

Vigilantism has been part of township life for generations (Reference BrewerBrewer 1994; Reference KynochKynoch 2005; Reference ShawShaw 2002) and the range of “violent social practices” that fall under its’ umbrella are particularly acute in “frontier” societies (Reference AbrahamsAbrahams 1998: 24), characterized as they are by endemic unemployment and high levels of vulnerability. While I am not arguing that violent forms of community mobilization (and reconfiguration of state law) occur only in “frontier” societies, this blurring of the boundary between the legal and illegal is particularly “visible”Footnote ¹⁸ in marginalized resource strapped areas, where “communities” are urged to be “responsible.” Thus, Reference Abrahams, Pratten and SenAbrahams (2008: 426) use of the term “frontier,” as referring to more than just spatial distance between the center and the periphery, serves as a useful framing device. Frontiers include the boundaries created by cultural differences, by temporal flows such as between night and day, by transitions from one social form to another and, by inequality and severe poverty. In frontier zones the state is often viewed as absent, ineffective and corrupt (Reference Little and SheffieldLittle and Sheffield 1983: 796; Reference PillayPillay 2008; Reference Rodgers, Pratten and SenRodgers 2008: 358).

The ineffective or absent state explanation, which is popular in South Africa, suggests that vigilantism has re-emerged in South Africa's black townships as a result of high crime rates and state failure to provide security for its citizens. This is attributed to inefficiency and corruption on the part of the police and the criminal justice system (Reference Dixon and JohnsDixon and Johns 2001; Reference Lee, Seekings and FeenenLee and Seekings 2002; Reference Schärf, Schärf and NinaSchärf and Nina 2001; Reference Swanepoel, Duvenhage and CoetzeeSwanepoel, Duvenhage, and Coetzee 2011; Tshehla 2005). For the most part these perspectives present vigilantism as a Hobbesian “feature of the state of nature,” as being outside of and in opposition to, “legal order” (Reference LenzLenz 1988: 126) and as constituting a threat to the rule of law and state formation (Reference HäefeleHäefele 2006; Reference Lee, Seekings and FeenenLee and Seekings 2002; Reference MartinMartin 2010; Reference RushRush 2015; Reference ShawShaw 2002; Reference Sekhonyane and LouwSekhonyane and Louw 2002; Reference Swanepoel, Duvenhage and CoetzeeSwanepoel, Duvenhage, and Coetzee, 2011; Reference TabaneTabane 2013). This fails to acknowledge how the state and its agents, both discursively and in practice, shape and produce violent policing activities and, in this sense, enjoy a complicated relationship of complicity. It ignores how the term is more of a “relational phenomenon” (Reference AbrahamsAbrahams 1998: 7) than a thing, often reflecting the political leanings of those who deploy it (Reference SuperSuper 2014), with a lack of clear “conceptual and empirical boundaries” (Reference Buur and JensenBuur and Jensen 2004: 148). Similarly, it ignores the fact that “criminality” is itself a product of power relations that are constituted by race, class and ethnicity (Reference PillayPillay 2008). Reference TankebeTankebe's (2009) argument—that expressions of support for vigilantism are a reflection of “perceived procedural injustice” and, that procedurally fair treatment promotes legitimacy in the eyes of the governed—is more nuanced than the “ineffective policing causes vigilantism” thesis. However, he does not engage with the non-universal meaning of legitimacy (Reference Johnson, Maguire and KuhnsJohnson et al. 2014: 949). This article argues that perceptions of legitimacy shift, depending on the context.

In contradistinction to the above arguments is a body of literature, not limited to South Africa, that theorizes an alignment of state and vigilante objectives. Reference Goldstein, Pratten and SenGoldstein (2008), in the context of Bolivia, points out that, like neoliberal rhetoric, vigilantism encourages private actions by citizens. Similarly, Reference Jarman, Pratten and SenJarman's (2008: 343) work on Northern Ireland suggests that we need to view vigilantism as part of a wider continuum of policing activities, rather than as something beyond the bounds of legitimacy. Reference Pratten, Sen, Pratten and SenPratten and Sen (2008: 3) argue that state responsibilisation of “non-state” actors has resulted in a blurring of the distinction between public and private and, Reference FourchardFourchard (2011: 609) argues that vigilantism and community policing in South Africa are “two sides of the same mobilization process.” This article draws on, and adds to, this literature, arguing that illegal forms of crime prevention have emerged out of state sanctioned lawful crime prevention initiatives, in a specifically post-apartheid manifestation. These forms of vigilantism are not an outright flouting of the law, as argued in the mainstream literature.

I show how the melding of state discourses—on neoliberal responsibilization, communitarian empowerment through the “social” and, on grassroots mobilization- has egregious consequences. This is so not only because it continues the “volunteer tradition of citizen-policeman” (Reference BrewerBrewer 1994: 309), but it also ignores historically violent modes of popular sovereignty and the overlaps between vigilantism and popular justice in poor black townships. Given the blurry boundaries between popular justice and vigilantismFootnote ¹⁹ I argue that it is not surprising that the street committees that emerged in the 1980s slipped between vigilantism and popular justice and continue to do so today, albeit in a modified way (Reference KynochKynoch 2005; Reference PillayPillay 2008). I use the term popular justice, both broadly, to denote its location “on the boundary between state law and indigenous law” (Reference Merry, Merry and MilnerMerry 1993: 35) and also in a more specifically socialist sense. The latter is based on Marxist- Leninist theory that promotes people's courts as vehicles for empowering the masses and creating a new society based on a “revolutionary vision” (Reference Merry, Merry and MilnerMerry 1993: 43). In the next section, I discuss popular justice initiatives in the 1980s in the context of “dual power.”

Street Committees and “Dual Power”

The term “street committees” first arose, during the 1980s, in the townships of the Eastern Cape.Footnote ²⁰ The idea was that each street would form a committee, with an area committee as the second leadership tier, and the local civic as the township's highest representative body (Reference Adler, Steinberg, Adler and SteinbergAdler and Steinberg 2000). These institutions emerged both as a “prefigurative ideal” (Reference AllisonAllison 1990)—to establish counter-hegemony against the state—and, also as a tactic to fill the gap left by activist campaigns to make the townships ungovernable (Reference Adler, Steinberg, Adler and SteinbergAdler and Steinberg 2000).Footnote ²¹ Although in practice there was much unevenness, street committees and the ideology of participatory democracy became hegemonic throughout the South African civic movement (Reference CherryCherry 2000). Then, as now, there were serious organizational problems and the practices perpetrated in the name of this ideal did not always live up to expectations (Reference Adler, Steinberg, Adler and SteinbergAdler and Steinberg 2000; Reference AllisonAllison 1990; Reference CherryCherry 2000; Reference Seekings, Schärf and NinaSeekings 2001).

“People's courts” were justified as being part of the struggle for dual power as well as a prefigurative practice of what justice under a socialist post apartheid dispensation could look like (Reference AllisonAllison 1990; Reference Burman and SchärfBurman and Schärf 1990; Reference HorwitzHorwitz 1992; Reference Schärf, Schärf and NinaSchärf and Nina 2001). In Cape Town, street committees were the primary informal courts in the townships, but there were many other bodies that also set up courts, such as criminal gangs, sports teams and the youth (Reference Burman and SchärfBurman and Schärf 1990: 426). Struggle activists presented “people's power” as the “collective strength of the community” (Reference SisuluSisulu 1986), yet, as is the case today, the community was both a site of contestation and a powerful ideological legitimating device. As Reference Seekings, Schärf and NinaSeekings (2001: 92) points out “the prefigurative ideal was all too often submerged under the sordid dynamics of local politics.” According to Reference Burman and SchärfBurman and Schärf (1990: 722) in Cape Town, where political organizations were not strong enough to control people's courts, they were “diverted from their original goals and used to promote the interests of those who usurped them.” They were also characterized by intergenerational conflict between older men and youth as well as discrimination on the basis of gender (ibid.).

Just as the apartheid state depicted township activists as violent criminals, stripping their acts of a political dimension, so too did the “comrades,” sometimes commit violent acts in the name of politics and accuse the South African government officials of being the true criminals. What might be regarded as gratuitous acts of violence, such as “necklacing” assumed some kind of political salience, due to the fact that the targets of these acts were accused of being apartheid collaborators (Reference SuperSuper 2013). At the same time, although the line between politically motivated, and hence justifiable crime, and unjustifiable, non-politically motivated crime, was a blurred one, as is the case today, crime was regarded as a threat to “community solidarity,” criminals were considered counter-revolutionary (Reference Marks, McKenzie, Schärf and NinaMarks and McKenzie 2001) and, were treated harshly (Reference SuperSuper 2013). Thus, there were clear overlaps between popular justice and vigilantism (Reference Buur and JensenBuur and Jensen 2004; Reference KynochKynoch 2011; Reference Seekings, Moss and OberySeekings 1992; Reference WilsonWilson 2002), even though, at the time, the term “vigilante” was reserved for “violent, organised and conservative groupings” operating against left wing activists, in terms of unacknowledged state political directives (Reference HaysomHaysom 1989).

In the next section, I discuss how street committees work in practice, in the face of a political discourse that deploys the “rhetoric of empowerment and capacity building” (Reference AminAmin 2005: 614). Responsibilization is punted as a form of 1980s grassroots mobilization, promoting street committees and “communitizing” crime prevention.

Mobilizing Communities

In 1999, the Gauteng provincial government launched a public campaign to reignite street committees in an effort to engage communities to fight crime “constructively” (Reference MinnaarMinnaar 2001: 40). President Zuma has also called on township communities to revive street committees in terms of a populist discourse, which harks back to the liberation struggle. This form of governance, what Reference DarracqDarracq (2008: 595) refers to as operating in terms of a “liberation paradigm,” valorizes local level initiatives by constantly seeking to mobilize communities on the ground. It is South Africa's version of the “Third Way.” And, indeed, street and area committees are common in Khayelitsha.

Whilst in theory street committees constitute a type of executive committee, consisting of 15 members who meet once or twice a week, and attend a weekly SANCO branch meeting—with all the other street committees in the branch—meetings are not well attended. Many members attend up to four meetings per day, are unemployed, unpaid, and suffering burnout. General meetings, at times also poorly attended, take place once per week, or more frequently if necessary, and should be attended by all the residents in the area. These are the fora where the “community” makes decisions, the theory being that “in the name of democracy the street committee never makes a decision on its own: it calls the general meeting” (area committee chairperson, “M” informal settlement, December 7, 2014). Although most street committee members are paid-up SANCO members and attend the local SANCO branch meetings, I came across street committees where members were not affiliated and did not attend SANCO meetings.Footnote ²² As such the allegedly cohesive community that the street committee represents, is, as was the case in the 1980s, not as united as it seems.

One might have thought that, with the advent of a democratic dispensation, there would no longer be a need for “people's courts” in South Africa's black townships,Footnote ²³ however, this has not been the case. But, nor is this a simple case of a historical residue from a previous era merely ticking over into a new era, since there have been definite shifts in the way that popular justice is meted out. Whereas in the past there was tacit acceptance of “violent forms of vigilantism” if it was initiated by, or had the consent of street committees (Reference Lee, Seekings and FeenenLee and Seekings 2002: 114), street committees no longer openly impose violent punishments. When street committees ask SANCO for advice they are warned not to “take the law into their own hands, to call the police and not hit/kill” (SANCO office bearer, Enkanini, August 16, 2014), because “due to ubuntu…we can't beat anymore” (area committee deputy secretary, “En,” June 16, 2014). The chairperson of an area committee blamed this for the low attendance at general meetings: “People won't come to the meeting because they can't see the point. They won't come because there's no outcome. Residents aren't satisfied about how we judge some of the cases” (“XX” informal settlement, December 7, 2014).

As Reference FourchardFourchard (2011: 624) points out, “the administration of violence is slowly but nevertheless gradually changing within these organizations,” as evidenced by references to the courts, to ubuntu, and to human rights. This has not canceled out other “codes,” but has resulted in “uneven and unstable mixings of legal codes (codes in a semiotic sense)” (Reference de Sousa Santosde Sousa Santos 1987: 298)—what de Sousa Santos refers to as “interlegality.” The result is a volatile hybrid of human rights, constitutionalism, and violence.

The same chairperson who complained that people no longer attended meetings because matters were inadequately resolved also stated that:

In one instance—where a “skollie”Footnote ²⁴ had murdered someone and stolen a laptop—the male members of the area committee solicited the help of male residents and conducted a search. They found the stolen goods and burnt the shack of the alleged perpetrator. According to my source, who was a SANCO office holder, “they never said anything at the SANCO meeting because SANCO wouldn't agree – that's why they don't tell you” (August 16, 2014). She also differentiated between her official capacity as a SANCO office holder and as a member of the community:

A witness in the court case that I observed, who used to be a street committee member and, whose son had been beaten to death for allegedly stealing a television set, stated that when a suspect refused to tell the truth they were “smacked”:

In this way, then, there is an “outsourcing”Footnote ²⁵ from organizations—street committees and SANCO branches—to “formations” (Reference Buur and JensenBuur and Jensen 2004). Formations are much more labile than organizations, enabling the latter to maintain a form of “public authority” (Reference LundLund 2006), which does not come at the expense of violent forms of popular justice. This entails a process of identity shifting (Reference Cooper-KnockCooper-Knock 2014), or “code switching” (Reference AndersonAnderson 1999: 132): when my interviewee shifts from being a SANCO office bearer, to a member of the “community,” she can participate in “street justice” (Reference AndersonAnderson 1999: 10). This works to maintain a violent specter of order, paying lipservice to state law.

An area committee chairperson, who had received training from the SJC in how to chair a meeting in a procedurally fair way, recognized that he had to “take the accused's side” during “interrogation” (September 22, 2014). However, he acknowledged that at times the females are asked to go outside so that the men can intimidate the “accused” into confessing. This generally worked because “usually when the ladies go outside … the truth comes out because the criminal knows he will get beaten.” At the same time, he also told me that “the community knows that it's illegal to beat, [that] vigilantism is not allowed but we just do it to scare. Mostly this works” (ibid.). He stated that a “confession” is obtained by:

Thus, contrary to mainstream discourse, vigilantes do not function in opposition to the formal criminal justice system, but support it, albeit in terms of a “subversive legal consciousness” (Reference SerbanSerban 2014: 789). The idea of the law is called on—via a distorted form of formal court procedures—to legitimate the obtaining of a confession by the area committee. In their drive to be “legitimate” street and area committee members must be both part of, and separate from the community. In this way then popular justice and vigilantism slide into one another, as do law and its other.

Banishment, discussed in the section below, occurs when a morally problematic individual is coerced into leaving an area. It serves both as a technology of crime prevention and punishment and is sometimes the “sentence” that a general meeting will impose. It is a vivid reminder of the spatial dimension of punishment.

Banishment and the Frontier

Banishing is a common remedy deployed by the “moral community” (Reference Buur and JensenBuur and Jensen 2004: 144) against “dangerous” individuals. It is effected either by informal curfews, in terms of which a group of people (often armed with sjamboks) patrol an area at night, confronting those who look suspicious and instructing them to leave; by talking to the family of a suspected criminal and instructing them to ensure that the person leaves the neighbourhood; by demolishing the person's dwelling (either by burning or destroying it with hammers) and finally; by killing the person (Reference SuperSuper 2014).

Banishment is rendered possible because accommodation arrangements in informal settlements consist of a volatile mix of transient and long-term residents. This volatility is compounded by the links that most residents have with the Eastern Cape: not only are 69 percent of adults born in the Eastern Cape (Reference O'Regan and PikoliO'Regan and Pikoli 2014) but, despite having lived in Khayelitsha for years, many refer to the Eastern Cape as “home” and make regular visits. There is much back and forth between the two areas, with new “immigrants” arriving and old ones returning, for the holidays or more permanently. Simkins refers to this as “oscillating migration.”Footnote ²⁶ It is because there is a space/place—what Reference PalmerPalmer (1998: 95) refers to as a “definite homeland which is outside the territory claimed by the perpetrators” that expulsion is a likely option.

A newspaper article reported on how “angry residents” burnt the shack of a woman who had allegedly abducted a baby from a hospital (Reference NombembeNombembe 2014) and, in this way, there was a purging. The incident of the “witchdoctor” described at the start of the paper was similar, in the sense that someone who had transgressed the moral boundaries of the “community” was instructed to leave. A street committee member, seated in my car which was parked outside the Site B police station, a few blocks away from the “X” informal settlement, told me how, as a crime prevention coordinator, he and other men had destroyed the house of two alleged robbers after they were released on bail: “we broke the house, beat the person and called the police to tell them to take the person. They never came back” (August 16, 2014). There are many more examples.

Banishment serves both as a technology of prevention as well as punishment. It is a mode of guarding the moral and territorial boundaries of informal settlements and it functions as “border control” (Reference AasAas 2014: 521) both literally, and in the sense of defining and defending the boundaries of membership. The primary focus is not on moral censure per se, but on who has the right to be in an informal settlement.Footnote ²⁷ The term “frontier” is particularly salient because banishment is a technology to both police and establish the boundaries of particular communities, boundaries that are not set in stone but are established, and re-established, through contested processes. As an area committee member stated: “if you take the person to the police station he stays for a few days and then he's back so sometimes it's better that he's chased away” (June 16, 2014). An old man living in a shack in Site B RR informal settlement was adamant that there was no vigilantism in his neighborhood but that

To the uninformed, driving past on the highway, informal settlements appear as a cacophony of shacks, one on top of the other. However, existence here is not a “free for all” and the presence of outsiders and insiders is tightly regulated. In informal settlements a prospective resident is required to furnish the street or area committee with an official letter,Footnote ²⁸ which states the reason for leaving the previous residence and, the contact details of two referees. Where a resident wishes to sell the rule is that the shack should first be offered to “locals” (people living in the area). This is done by informing the street committee. Only after it has been on offer for 3 months can it be sold to an outsider, provided that the required credentials are furnished. In this way then the street committee plays a vital role in regulating who lives in informal settlements.Footnote ²⁹ One of the reasons given to me was because “we don't want criminals living in our area” (street committee member, “R” settlement, October 4, 2014). In future research it would be worthwhile looking at the extent to which these rules are adhered to or broken, and the informal political economies at play.

Banishment, in the case of children, under the age of 18, who still reside with their parents, is directly or indirectly imposed. One interviewee told me that when a community member was robbed by a youth he was not “chased away but his parents sent him back to the Eastern Cape” because they knew the consequences: “we can beat the kid or chase both the kid and the parents away” (member of street committee, PJS informal settlement, August 11, 2014). In another instance SANCO called a meeting with the parents of a “problem child” and the decision was that the child had to leave the area: he now lives in another section with his grandmother (member of street committee, R settlement, October 4, 2014). Thus, it matters only that the “criminal” moves away from “here,” rather than the whole township, and the process of banishment involves “negotiation” with the parents or relatives of suspected criminals.

These practices, while formally illegal, are defended by agents of the state and differentiated from incidents of “mob justice.” According to a high ranking police witness at the Commission, whether it was a case of housebreaking or “child molestation” none of these “formal meetings” were accompanied by violence: “these people were just motivated to leave” (Colonel Nel, Commission: 4635). These practices of banishment are a clear example of how legality and coercive illegality collapse into each other in marginalized areas, where residents are largely invisible to the state and where the “right” to reside in shacks lacks formal legal status and is dependent on access to political and social networks. In the next section, I discuss how legal and illegal policing combine in the context of state responsibilization discourse.

Anti-Crime Patrols

Informal settlements in Khayelitsha constitute an example of a “spatial frontier zone” where nocturnal social control is weaker than in daylight (Reference AbrahamsAbrahams 1998: 26). In this section, I discuss how neighborhood watches and their less formalized counterparts (community crime patrols) police these nocturnal spaces. These township patrols are a deeply entrenched tradition in the history of South African townships (Reference BrewerBrewer, 1994; Reference De KlerkDe Klerk 1999; Reference FourchardFourchard 2012; Reference HorwitzHorwitz 1992). They have been actively encouraged by the post-apartheid South African state, across former townships and formerly white suburbs.

The National Development Plan 2030 promotes an active citizenry to “build community participation in community safety” (Lawrence, cited in Reference O'Regan and PikoliO'Regan and Pikoli 2014: 152). Similarly the Western Cape's DOCS has adopted a “whole of society approach”—based on the idea of “making safety everyone's responsibility” (Lawrence, op cit: 150). DOCS seeks to “expand community participation in public safety” by “mobilising communities against crime” and encouraging “community safety structures” (Department of Community Safety 2010: 21.1.1: 34–35). These include neighborhood watches, street committees, street watches, block watches, farm watches, and flat watches (Department of Community Safety 2010: 8). Social crime prevention, which initially sought to reduce the “socio-economic and environmental factors conducive to criminality” (Department of Safety and Security 1998: 44), has been diluted to the extent that it now refers largely to situational crime prevention.Footnote ³⁰ Thus, the South African Police Service measures the success of its “social crime prevention strategy” in terms of the number of crime awareness programmes, neighborhood watches, business forums and street committees that are established to deal with crime (Reference LamoerLamoer 2014).

Neighborhood watch members in the well-off middle-class white suburbs support the police by, for example, purchasing walkie-talkies for them but, in Khayelitsha, members look to the police for support. Whereas private security companies play a key role in assisting neighborhood watches in affluent areas, residents in informal settlements cannot afford these services. They are totally dependent on the state. Interviewees consistently complained about not receiving remuneration, and lacking kit, apart from yellow bibs—which make them look official. As Herbert (cited in Reference DelgadoDelgado 2008) argues in the context of the United States, community policing cannot solve structural problems. Residents in frontier zones lack connections with other social groups because of their structural isolation. Whereas bonding social capital (inward orientation with relatively powerful ties (Reference Browning, Feinberg and DietzBrowning, Feinberg, and Dietz 2004: 525) is present, bridging social capital (connections across social groups and areas)—is not. Neoliberal discourse on ethical voluntarism, which presents local cohesion as a cure for structural ills (Reference AminAmin 2005), overlooks the lack of resources and inequality between areas as well as the “cultural transmission” (Reference AndersonAnderson 1999) of violent community-based norms. My findings in relation to residents in Khayelitsha are applicable to other marginalized communities whose residents lack the time or energy to attend “evening meetings with the police month after month” (Reference DelgadoHerbert, cited in Delgado 2008: 1197), or to consistently engage in anti-crime patrols.

It is not surprising that anti-crime patrols have a very small base of active membership and the fact that their effectiveness is totally dependent on unemployed volunteers, who are already severely stretched, is an important factor in their lability. According to a VPUU employee there were 1800 neighborhood watches in the Harare Policing precinct in 2006 but this had dwindled to less than 100 by 2011 (August 13, 2014). The chairperson of the “Father's Committee”, a community patrol group in “XX” informal settlement, told me that DOCS had trained him in patrolling and that although “in those days we had a radio- now we have nothing” (February 1, 2014). All interviewees complained that neighborhood watches, like Community Policing Forums, were ineffective and sporadic. This lability applies even more so to community patrols, because these groupings do not have the overt support of the police, such as it is. In two instances I heard about groups only because their members had been arrested—in one case for destroying a shack, and in the other, for murder.

The layout of informal settlements makes them difficult to patrol. There is often only one access road into the settlement, sometimes none at all, making vehicular use all but impossible. The footpaths are narrow and, in winter, often become waterlogged and difficult to navigate. There is a lack of working streetlights. The technologies that are available to neighborhood watch members in affluent suburbs, to displace potential threats, are not available in poorer areas. It is not surprising that neighborhood watches and community patrols resort to more coercive tactics to essentially achieve the same result—which is to force potential criminals to leave an area.

Neighbourhood watches and community patrols impose an informal curfew after nine pm, when they conduct patrols.Footnote ³¹ The curfew, which has historical precursors in the 1930s Vigilance Associations is premised on the theory that black men should not be walking around at night (Faul, cited in Reference O'Regan and PikoliO'Regan and Pikoli 2014: 323) and, a clear (re)iteration of the colonial and apartheid constructions of young black young men as dangerous. I was told that: “the nine pm rule is because we are trying to make people not be on the street” (Sector four neighborhood watch member, September 16, 2014), that “after 9 we don't like people walking around – especially youngsters” (community patrol group member, August 16, 2014) and that, in Site B, the decision to enforce a curfew was made by the commander at the Site B police station (Sector four neighborhood watch member, September 16, 2014). If this is indeed the case, and since there is no curfew law in South Africa, then the frontier between vigilantes and the state is “more fictional than we would initially think” (Reference AbrahamsAbrahams 1998: 95).

The figure of the criminal outsider takes different shapes, depending on the context. One interviewee stated that it was problematic that the “Somalian shops’ stayed open after 23h00 and in one instance a neighborhood watch noted “late trading of foreign national” in its incident report (Violence Prevention Through Urban Upgrading, December 2012). Thus, criminals are constructed in terms of a discourse that is suspicious of foreigners—black foreigners who are living in the predominantly Xhosa Khayelitsha and, so populist and state discourse goes, taking away employment and business opportunities from South African citizens. Foreign shop-owners, particularly Somalians, are targeted for not adhering to the by-laws on trading hours (minor administrative transgressions).Footnote ³² Similarly, the curfew is often imposed against those “naughty” people, against whom the community must guard itself, who come from other areas to drink in “our” shebeens and then hang around to do crime (member of informal patrol group, S section, January 25, 2014). According to SJC records a man who was entering an outside toilet at 23h30, was punched and kicked by neighborhood watch members because he was violating the curfew (SJC records, April 4, 2011).

A respectable 60-year-old woman, who was very involved in community activities and a longstanding ANC member, told me, as I sat on the couch in her spotlessly clean shack, that she was explicitly against the violence of the local community patrol group. Yet, she admitted that “even us [the sector Four neighbourhood watch], if you are very naughty and you refuse to spread your legs and hold your arms above your head we will sjambok you” (September 16, 2014). And, indeed, neighborhood watch members have been known to mete out punishment in the name of crime prevention. In one instance a neighborhood watch reported that its members had made a person who “sprayed” them with water do 36 push ups for “punishment” (Commission: 2736) and in another the members had seen some men carrying “big stones”. When they enquired as to the reason for this “the people started swearing and shouting at us and throwing the stones at us. We started throwing the stones back at them. X was hit and fell down. We called a van to check on him.” It is interesting that the state was only called on after the neighborhood watch members threw stones at the alleged wrongdoers, seriously injuring someone, to the extent that he collapsed.

The Neighbourhood Watch Learner Guide, which is given to members when they participate in the training provided by DOCS, discusses exceptions to the general constitutional rule that no one may be arbitrarily arrested and unlawfully deprived of property. Even for an English speaker, with a legal background, its ambivalence is astonishing.Footnote ³³ At best this is a gray area of the law, at worst it might be considered as encouraging citizen's arrests in highly dangerous contexts. No wonder then, that neighborhood watches involve themselves in “stolen goods interventions” (VPUU, December 2012). This is an example of how coercive policing initiatives emerge in the shadow the state's drive to have communities become involved in local crime prevention initiatives. Yet, at the same time, there is a distancing because the state cannot be seen to be complicit in overt violence. This brings to mind Johnson's argument that even lawful self-help groups have the potential to exercise or threaten force—which potential may in and of itself not be unlawful (Reference JohnsonJohnson 1996). Similarly, Reference AbrahamsAbrahams (1998: 7) argues that this instability is not so much because anti-crime patrols “emerge in a legal twilight zone” but because they are “rather labile…manifestations [that] are relatively short-lived and always capable of slipping and sliding in one direction or another”. So something that appears to be positive, because it “creates order out of chaos” (Reference Rodgers, Pratten and SenRodgers 2008: 359) has the potential to rapidly become something negative.

The next section discusses how the “moral community” produces its other, the criminal, and how the social construction of criminality impacts both the targets, and degree, of violence used against outsiders.

Constructions of Criminality

At the opening of Parliament in 1995, President Mandela blamed crime and violence for “eroding the foundation of our democracy”, necessitating a “harsher approach” (Department of Safety and Security 1999: 9). In 1999, the Deputy Minister of Justice boasted that mandatory minimum sentences and restrictive bail laws were “progressive” (Hansard 1999: col. 2756) and in 2001, the Minister of Safety and Security stated that prisons were overcrowded because the police were doing their job and that “all what we need … is to fully rally behind the police and to declare that the fight against crime is our fight” (Hansard 2001: col. 2470). Political leaders have also called on police to “kill the bastards” (The Star 2008), to “teach them a lesson” (The Citizen 2008) by means of the use of lethal force and to show “no mercy” (Burger 2009 cited in Reference BruceBruce 2010: 9). Foreigners are targeted by this punitive discourse. In 1997, the Minister of Justice stated that “the influence of a large number of illegal aliens in South Africa…had a significant impact on criminality” (Hansard 1994, col: 2074). The Minister of Safety and Security has referred to “havens of criminality” (Hansard 2000 question 49: col 892) as being the places where many illegal aliens are netted in cleanup operations and, arbitrary arrests and detentions of asylum seekers are common occurrences (Reference Buur, Jensen, Buur, Jensen and StepputatBuur and Jensen 2007; Reference LandauLandau 2005). Special operations such as Operation Aliens and Operation FielaFootnote ³⁴ aim to arrest and deport “illegal immigrants” because of their alleged criminality. This “hyper-politicisation of penal policy” (Reference Gottschalk, Simon and SparksGottschalk 2013: 217), and the targeting of black people from other countries, frames the field of punishment (and other “self-help” initiatives) in Khayelitsha.Footnote ³⁵

Community crime control initiatives often focus on young men, particularly when they are suspected of committing offences that offend a community's sense of justice. According to a police report, between April and June 2012 there were 78 recorded “vigilante incidents” in Khayelitsha (South African Police Service, unpublished). These all resulted in death: most of the victims were young black men, between the ages of 18 and 30. Half had been caught in the act (or were suspects) of stealing, robbing, or, housebreaking and, 10 had recently been released from prison and/or remand detention. Whereas a youthful skollie will be banished, assaulted, chased away, or maybe even killed, this rarely occurs in the case of drug dealers, because they occupy powerful positions within the community and are believed to be protected by the police.

Although interviewees consistently raised the problems of young drug users who committed crimes, suspected criminals (see also Reference Jensen, Pratten and SenJensen 2008) who were out on bail, and young men returning from shebeens late at night, domestic violence was not considered to be worthy of intervention, nor was vigilantism itself. As the chairperson of the area committee in “M” informal settlement told me: “we can't force a husband to love his wife” (September 22, 2014). His attitude is not unusual (Citizen Surveys and Centre for Justice and Crime Prevention 2011). I was told that “if a woman cries they ignore her but if they shout: ‘it's a robbery’ then everyone comes out” (community journalist, August 19, 2014).

The distinctions that are drawn between various types of criminality impact on the responses. A street committee member in PJS informal settlement told me that in the case of gangsters fighting—“small kids between the ages of 13–15”—they would call the parents and the police because the community was “too scared to break up the fight - we ask the police to come and patrol because we can't catch them by ourselves because they have sharp weapons.” “Robbers,” conversely, are “dealt with by the community” but drug dealers are “too powerful” (former CPF member, August 16, 2014).

Whereas street committee members “look the other way” in the case of street justice—itself constituted by violence and, as such, also a crime—a person who commits theft will be assaulted and/or ostracized. Thus, the politics of identification play off against a politics of distinction (Austin, cited in Reference DelgadoDelgado 2008: 1211). Just as the predominantly white residents in well-off suburbs react to crime by insulating themselves from the mainly black poor who are viewed as potential criminals (Reference PillayPillay 2008; Reference ShawShaw 2002) so too do residents in poor black areas seek to insulate themselves from those who they perceive to be a threat to their security. Both forms of insulation have racial and class overtones but deploy different technologies. Precisely who constitutes the community and what offends its sense of justice are also differently constituted.

In one instance, in a typical distancing maneuver, where a group of young Zimbabwean men were necklaced and burnt alive the police claimed that this was “vigilantism… in retaliation for a bar fight death” and in in “no way linked to xenophobic attacks…” (Reference Evans and WicksEvans and Wicks 2015). In another, a black gardener on his way to work in a white area was sjambokked by a white male who suspected him of trying to break into his car (Geach 2014). Thus, the archetypal young black male is a morally ambiguous and potentially criminal figure in both rich and poor areas. In both contexts, the person must leave the area because he constitutes a threat but the discourse underlying the threat is different: in the one area it is laced with xenophobia and in the other with a more generalized form of blanket racism (against black men)—both are forms of “vigilantism” manifesting within specific local political contexts.

So too, does the “community” “take on different meanings in different conditions of economic and social well-being and in different institutional settings” (Reference AminAmin 2005: 623). Because frontier communities are economically and politically trapped in, and delinked from, “networks of spatial links” (Reference PillayPillay 2008: 150) they are particularly fragmented. Although a SANCO member stated that “SANCO” is the “community” (September 16, 2014), when it comes to retrieving stolen property the meaning of “community” is more restricted. As a community crime patrol member from the same area said:

The line between “community” and “mob” is not an obvious one and, sometimes, the same person is both upright member of community, as well as vilified member of mob. One of my interviewees was a member of the local neighborhood watch, a community crime patrol group and the Khayelitsha Community Police Forum. He had participated in the demolition of three shacks, after a decision to this effect was taken at a general meeting of the street committees in his area (February 1, 2014).

In part, whether or not one survives a vigilante attack, or is attacked at all, depends on how well networked the person and/or his relatives are within the area. I was told that

Similarly the Deputy- Secretary of an area committee in the Endloveni informal settlement told me that the type of punishment inflicted depended on how the person interacted “with the community” (June 16, 2014). This is consistent with Larcom's work in the New Guinea Islands—that harsher and more violent punishments are imposed on those at a “greater social distance” (Reference LarcomLarcom 2015: 183).

A female who was beaten by a community patrol for suspected drunkenness told me that this would not have occurred in her own neighborhood (resident in R settlement, June 16, 2014). She also stated that when the Somalian shopkeepers in her neighborhood were robbed no-one stood up for them but that if she were robbed the community would find and punish the offender (ibid.). Thus, just as state discourse targets foreigners as criminals and for being somehow illegitimate so too are they ignored by the “moral community” when crimes are committed against them. This selective attitude is borne out by research conducted by Reference Gastrow and AmitGastrow and Amit (2012).Footnote ³⁶ In this way then, ethnic identities and spatial differentiation (itself a marker of class: who is there legitimately and who is not) are “crucial markers” (Reference KynochKynoch 2005: 500) in the construction of both criminality and victimhood. The conferring of outsider/insider status functions as a mode of boundary setting, in the same way as the other crime prevention technologies discussed in this article.

The Case Number

This section argues that just as a case number is important for those in affluent suburbs, without which they cannot lodge an insurance claim, so too is it an important tool in the retrieval of stolen property in informal settlements, but for different reasons. In both instances, there is an interaction with official law for purposes not quite contemplated by it.

According to research conducted by Seekings (Reference SeekingsSeekings 2013: 24) a far smaller proportion of Khayelitsha residents said that they would go to the police to solve problems of house-breaking than was the case in white/colored neighborhoods. Instead, a significant minority of respondents said that they would solve the problem locally—through friends, neighbors, or local organizations—rather than the police. In the murder trial that I observed the state eyewitnesses testified that had the accused followed the standard practice of reporting the incident (a stolen television) to the street committee the deaths of the suspects could have been avoided. A young woman living in a tin shack in Enkanini—a shack that was big enough to hold little more than a double bed and a hot plate—told me that when her blankets and hair-iron were stolen during a break-in she did nothing because she did not see the thieves, but that if she had seen them she would have alerted the “community” to assist her in retrieving her goods (October 16, 2013).

Conversely, many interviewees stated that a “case number” is required for non-state actors to assist in the retrieval of stolen property. This represents a distinctive shift from the 1980s, where the police were largely irrelevant to the solving of criminal cases in black townships. In the post-apartheid era, a clear link to the police, and a form of official authorization, is now being claimed: “without the case number there's nothing we can do because everything that we do, the police must know about first” (member of community crime prevention committee, “X” informal settlement, August 16, 2014). Thus, ostensibly competing types of authority—the police and the community—complement each other, when they informally collaborate.

According to my interviewee, the police were aware that they searched, and if necessary, assaulted suspects:

This is consistent with Reference Owen and Cooper-KnockOwen and Cooper-Knock (2015: 8) whose research in Nigeria and Kwamashu led them to argue that the “bureaucratic role of the Police is crucial” insofar as this is a means to both “create records and take action”.

A former member of the Amadlozi, a vigilante organization in Port Elizabeth, whom I interviewed during a break at the Commission hearings, after proudly showing me newspaper clippings about the Amadlozi, explained that there was:

His reply to my question of why he would take the person to the police, after retrieval of the goods, was that prevention was “better than cure,” that “the law” had to “take it's course” and that “the judicial system must do its duty: we already assisted the Investigating Officer. We made the job easier.” This is ironic, considering that in fact the law is patently not taking its course. Instead, what we are seeing, is the balancing of an ideal “law” that Khayelitsha's residents desire—one which involves the police cracking down on criminals, locking people up for more than the requisite 48 hour period,Footnote ³⁷ making communities safer and, returning stolen property—with the one that they expect—that the law will not assist them because its agents are corrupt, inefficient and “on the side of criminals.” (See also Reference Cooper-KnockCooper-Knock 2014). One only has to peruse the record of the Commission to find overwhelming references to police and overall criminal justice system inefficiency. Interviewees (and Commission witnesses) consistently expressed frustration because when “criminals” were arrested they were “immediately” released back into the community. The chairperson of the Father's Committee, who had himself spent 15 years in prison and had allegedly never seen a social worker, stated that: “a criminal is supposed to go behind bars for a long time. He can't stay with people” (February 1, 2014).

Clearly, there is state complicity. Thus, a school principal testifying at the Commission, stated that the police had witnessed him driving with suspects in the back of his truck in order for them to show him where the stolen goods were being sold. The police made no effort to charge the suspects and commence with investigations (Mr. Mjonondwana, January 28, 2014). Similarly, a senior prosecutor from the Khayelitsha Magistrate's court publicly requested residents who were attending a “community meeting” in Enkanini, to assist the state in collecting evidence. As she put it:

By asking residents to take responsibility for collecting evidence, the state creates the space for violent illegality and simultaneously shapes the vilification of criminals by framing it as a “good” people versus “bad” people issue.

Conclusions

In South Africa, both the rich and the poor are expected to buy into the ideas of serving the community through voluntarism. This article has argued that, in a context of deep scarcity, it is dangerous to render the “community” responsible for crime prevention because it overburdens those who are most in need of state support. Not only does the discourse of community crime prevention gloss over macro-socioeconomic conditions; the extent to which historically violent modes of “popular sovereignty” still play out; localized power struggles; levels of violence and; the presence of hated outsiders—such as ex-prisoners, people released on bail and foreigners—but it also creates a fractured and volatile solidarity around crime and criminality. The result is an unstable mix of different forms of control, a blurring of the legal and illegal.

The article has sought to provide insights into the governance of security in Khayelitsha. It has described how the variegated informal arrangements between the police and local actors distort “more visible legal logics” (Reference ValverdeValverde 2008: 10)—such as the principles of penal law (due process, the presumption of innocence, no detention without trial) that are enshrined in constitutional democracies, but only ever incompletely realized in practice. This interaction between “legal” and “illegal,” the configuration and reconfiguration of civic action, of legal consciousness and, of state law was examined in the empirical sections of the paper. Whether via street committees which sit as people's courts, neighborhood watches, informal patrols, banishment, or laying a criminal charge at the police station, the effect is to tolerate violence and to keep it in the space of the local. However, when the violence becomes too spectacular, then the state distances itself from those who “take the law into their own hands.”

Just as there are linkages between the police and the mob (Reference Cooper-KnockCooper-Knock 2014), so too are there linkages between the mob and the community: at times the very same people constitute both and, as such, the term “community” is a problematic concept. I have shown how, when suspected criminals are released on bail, the police and courts are perceived to be acting ineffectively, thus necessitating more permanent forms of expulsion—such as the destruction of shacks, or even death. As such the concept of “effectiveness” is dangerous and the quest for legitimacy in the eyes of a deprived and fragmented community, that coheres around crime, can only be met by permanent expulsion, such long- term imprisonment, itself a form of legal banishment which has deleterious social consequences.

Like power, legitimacy is not something that can be seized, but is constantly negotiated and the paper has described how, in seeking to maintain legitimacy both vis a vis the state and also local residents, street committee members “look the other way” when violent acts of street justice take place. The ANC, too, has sought to maintain its legitimacy as a ruling party. Once it transitioned from being a liberation movement, engaged in popular justice, to becoming the governing party it sought to maintain credibility by framing neoliberal responsibilization rhetoric in the cloak of revolutionary populism. It combined this with a punitive discourse on crime and foreigners, particularly foreigners from other African countries. The article has argued that, given the cultural and historically violent context of popular justice, coupled with inequality, scarce resources and, an absent state, the connections between dual power and vigilantism have, at best exacerbated and, at worst, shaped an already volatile situation. Thus, the state has, via its discourse and practices (including omissions), created the space for legal community based crime prevention technologies to slide into illegal vigilantism.

Given the fuzziness of the discourse of neoliberal responsibilization, communitarianism and grassroots democracy and the way they play out on a multiscalar level in South Africa, it is not surprising that things go wrong during implementation, particularly as all three fail to address inequality. Whereas the rich can outsource to private security companies the poor cannot: they have only their time, which is an already stretched resource. Thus, the violence which underpins all forms of exclusion—both legal and illegal—plays out far more visibly in informal settlements than in affluent neighborhoods, yet, at the same time it is also invisible because it is occurring on the frontier—affecting poor people who live far from the city center. This phenomenon is not limited to South Africa but occurs, particularly in contexts of deep inequality, throughout the world.

The hybrid security assemblages that assemble and disassemble in Khayelitsha's informal settlements, and in other marginal areas, throw the ambiguities of liberal theory and the (non) “singular nature of the rule of law” (Reference ValverdeValverde 2008: 5) into stark relief. It is here, on the local scale, on the frontier, that the incompleteness of the liberal democratic state formation project is most obvious and where future research should be focused. When social crime prevention is watered down and woven into the discourse of the “responsible community,” as has been the case in South Africa, the technologies of community policing and community-based crime prevention become penetrated by violent “forms of [illegal] local justice” (Reference HornbergerHornberger 2013: 12). At the same time, these local forms are also penetrated by state law. Thus, the discourse and practices of human rights, ubuntu, due process and, bureaucratic policing procedures (such as registering a case) are all braided into local security assemblages. State law is not so much absent, as reconfigured, with the spectacles of legal and illegal being graphically interwoven with each other.