This paper is based on research on vigilantism in Khayelitsha, a black township, on the outskirts of Cape Town, South Africa. Meaning ‘new home’ in isiXhosa, Khayelitsha was established in 1983 in terms of the Native Urban Areas Act.1 It covers an area of about 47 square kilometers, is situated 30 kilometers from Cape Town’s city centre, and is the fastest growing and third largest township in South Africa, with a population estimated to be between 400,000 and 750,000.2 It consists of both formal and informal settlements.

Almost half of Khayelitsha’s households live below the food poverty line3 and there is a significant difference between income distribution within Khayelitsha vis-a-vis the Cape Town population.4 Poverty is widespread, with the majority of Khayelitsha’s residents living cheek by jowl in overcrowded shack settlements, accessing electricity illegally, sharing communal water taps, and relying on grossly inadequate sanitation arrangements (such as outside portable toilets).5 Despite overall poverty levels, some areas — such as Lingelethu West — contain parts that are relatively prosperous (Seekings 2013: 20).

As a ‘frontier society,’ Khayelitsha poses a particular problem for the administration of criminal justice and state structures enjoy little legitimacy (Little & Sheffield 1983: 796).6 Identified by the Western Cape Provincial Government as a ‘zone of poverty and unemployment,’ Khayelitsha has the second highest number of murders in the province and — together with the poor black townships of Nyanga, Gugulethu, and Harare — reports the highest number of murders in the Western Cape (DCS 2009: 20; Lancaster 2013).7

This paper argues that despite the fact that South Africa attained formal democracy in 1994 and has one of the most progressive constitutions in the world, the ‘nature and character of public participation’ is problematic, non-deliberative, and a reflection of blocked opportunities, poverty, and inequality (Miller 2013: 18). I argue that the discourse and rhetoric on ‘community’ is a shallow and dangerous form of incorporation that does not live up to its promises of full participation. Not only is ‘community’ used as a tool to connect the state to the population but vigilantes also claim to act in the name of the ‘community.’ As such, the absence of deliberative democracy is linked to violent community-based punishment in Khayelitsha and there is an elective affinity between this pathological form of grassroots democracy and vigilantism.8

After a brief section on ‘Background and Methodology,’ I discuss the relationship between democracy and imprisonment, noting that South African democratization has witnessed a dramatic increase in long-term prison sentences. I argue that the government’s embracing of a restorative prison has enabled it to double back on its parsimonious, pre-1994 stance on imprisonment. As a result, the prison and penal punitivism have become an important part of state building in newly democratic South Africa. In ‘Populist Politics and People’s Power,’ I discuss the historical relationship between violence and justice and argue that both ‘criminality’ and ‘community’ are highly contested terms that are subject to violent definitions and redefinitions. The section on ‘Politics and Crime’ discusses how the post-apartheid criminal has been de-linked from a structural and political context. I argue that the Khayelitsha Commission of Inquiry into Policing was itself a politicized vehicle that — thanks to its narrow framing — was bound to focus on ‘police inefficiency’ rather than the socio-economic conditions that generate crime. In the section on ‘Partnerships and Grassroots Democracy,’ I discuss how the government has encouraged partnerships to deal with crime and how its policies are filtered and refracted through street committees and Community Police Forums. The section on ‘Banishment and the “Moral Community”’ argues that banishment is utilized as a widespread technology for dealing with crime in Khayelitsha and, although it is often presented as a voluntary departure, there are distinct undertones of coercion. In ‘Violence to Retrieve Property,’ I discuss how violence often occurs in the retrieval of stolen property and that the theft of consumer electronics assumes great importance in a context of deep scarcity and endemic societal inequality. The penultimate section, ‘Outsiders on the Edges’ argues that ‘mob justice’ is in fact less spontaneous than it initially appears to be. Informal political institutions — like their more formal counterparts — are less responsive to those on the margins of mainstream society, particularly foreigners, refugees, and suspected criminals. As such, the power differentials between those who belong to the ‘community’ and those who do not result in ‘differential mobilization’ around crime and the policing of the community’s ‘moral boundaries’ (Miller 2013: 10).

Background and Methodology

Prior to April 2002 vigilantism was not officially recorded but, after an incident in which three suspected criminals were necklaced,9 the police management of the Khayelitsha precinct started to register vigilante incidents and the Khayelitsha Police Crime Intelligence Analysis Centre (CIAC) began to report on the phenomenon (Häefele 2006).

In August 2012, the Premier of the Western Cape — acting on a complaint received from six human rights organizations — established a Commission of Inquiry into Policing in Khayelitsha. The Commission was mandated to investigate ‘Allegations of Police Inefficiency and…a Breakdown in Relations between the Community and the Police in Khayelitsha’ which was supposedly indicated by a spate of vigilante attacks reported in the media (Zille 2012).

It should of course be noted that the term ‘vigilante’ is politically contested and its usage reflects the political leanings of those who deploy it. Thus for example, officials of the National Union of South African Mine workers recently referred to rival union members as being ‘vigilantes’ (Tabane 2013). There is also no certainty as to exactly what vigilantism is and the term has been used to describe a wide range of activities, perpetrated by a variety of actors, e.g. errant police, street committee and neighbourhood watch members, large mobs of people and individuals acting against perceived lawbreakers (Buur 2003; Buur and Jensen 2004; Harris 2001; Huggins 1991; Johnson 1996).

My research methodology included 38 selective in-depth interviews, supplemented by documentation and attendance at various meetings.10 Interviews were conducted with members of community patrol groups in two different informal settlements;11 fourteen members of the Social Justice Coalition (SJC)12; a community journalist who covers vigilante incidents; a film director making a documentary on ‘mob justice’; fourteen street committee members; four members of the ANC aligned South African National Civic Association (SANCO); four residents who self-identified as ‘vigilantes’; four mothers whose children had been beaten to death; a woman who had been assaulted by male members of a neighbourhood crime patrol; four Enkanini shack-dwellers; Gideon Morris – the Chief Director of Civilian Oversight in the Western Cape Provincial Government’s Department of Community Safety; and three advocates who were representing clients accused of vigilantism.

Apart from interviews, I also attended eight meetings, organized by the SJC, around the Khayelitsha Commission and the SJC’s ‘Campaign for Safe Communities.’ I attended the first ten days of the Khayelitsha Commission and three weeks’ worth of proceedings in a High Court case where six people were charged with the kidnapping and murder of four youths alleged to have stolen the plasma television set of one of the accused.13 I supplemented this ethnographic work with documentary sources including newspaper reports on vigilantism, affidavits of witnesses testifying at the Khayelitsha Commission, judgments in court cases that dealt with vigilantism, and extensive documentation made available by the Department of Community Safety.

Democracy and Imprisonment

One might have expected the advent of democratic rule to result in prisons being replaced by a more humane criminal justice system that emphasized non-custodial rather than custodial sentencing.14 This has not been the case. Although the numbers in custody have decreased from an all time high in 2004, democratization has brought with it a dramatic increase in long-term prison sentences ranging from seven years to life. In 2013, the number of people serving life sentences was 11,000 as opposed to just 400 in 1994 (JICS 2013). In 2014, 63 per cent of an overall 154,648 inmates were serving sentences in excess of seven years (JICS 2014: 41). Between 2000 and 2010, sentences of life imprisonment increased by 572 per cent whilst those in excess of ten years increased by 128 per cent (JICS 2010: 25).

Ironically, the 1995 Constitutional Court ruling that the death penalty was unconstitutional also opened the door for greater reliance to be placed on the prison.15 It did so by relying extensively on the concept of ubuntu, as supporting a ‘reformative theory’ of punishment. In so doing it helped to cement the shift to long-term imprisonment by embracing the notion of ‘a reformative prison’ as a replacement for the death penalty (Super 2011, 2013). For example, in 2002 — at the precise time at which the South African rate of imprisonment had almost peaked — the Department of Correctional Services introduced a restorative justice approach that ‘aimed at facilitating the mediation and healing process between offenders, victims, family members and the community’ (Department of Correctional Services 2001). However, as Michel Foucault and others have pointed out, prisons are constantly failing institutions if measured in terms of their capacity to rehabilitate (Foucault 1995; Dumm 1987; Rothman 1980; Meranze 1996).

South African prisons are a stark reminder of the edges of democratic rule insofar as they highlight the gaps in the human rights speak that is central to ‘bourgeois legality’ (Fitzpatrick, cited in Merry 1988). The very professionals — social workers, psychologists, and other professionals — who are key to rehabilitation are in short supply and — in overcrowded institutions — prisoners only have 1.2 square meters in which to eat, sleep, and spend 23 hours of the day (JICS 2010: 25; 2013: 81; 2007: 16). In this sense then, the embracing of restorative justice in the context of imprisonment, was a ‘discursive maneuver’ (Garland 1985: 172) which enabled the new African National Congress (ANC) government to reinvent the prison as being restorative rather than repressive. This reinvention enabled the ANC to deploy it as a vital cog in its state-building project, despite previous pronouncements that it did not support imprisonment because of its association with apartheid rule (Super 2010, 2011).16

Populist Politics and People’s Power

Whether it is referred to as ‘mob justice,’ ‘vigilantism,’ ‘non-state punishment,’ or something else entirely, the phenomenon of people taking the law into their own hands is not new to South Africa. Indeed, the relationship between violence and justice has deep historical roots and — whether legal or illegal — violence has had an ubiquitous and quotidian presence in South African history (Minnaar 2001; Seekings 1989; Seekings 1995; Harris 2001; Thomas 2012; Wilson 2002; Van Onselen 1984; Allison 1990; Kynoch 2008; Pavlich 1992). Both the National Party government and the exiled ANC liberation movement used the death penalty against their enemies and the radical traditions of people’s power and ungovernability sometimes resulted in violent punishment (Super 2013; Kynoch 2011; Machava 2011; Burman and Schärf 1990; Jensen 2008; Seekings 1992; Wilson, 2002).

During the 1980s, left-wing activists presented ‘people’s power’ as the ‘collective strength of the community’ (Sisulu 1986). Whereas some criminal acts were overtly politicized and excused on the basis of having a political justification, non-political, purely criminal acts, were subject to severe punishment. The complex process of constructing and defining criminality had many aspects to it: not only did the comrades17 justify ostensibly criminal acts as being political, but they were, additionally, virulently anti-criminality per se (Harris 2001). People’s power was regarded as a particularly good form of ‘crime control’ since it was — according to struggle stalwart Zwelakhe Sisulu — ‘disciplined, democratic and an expression of the will of the people […and] in the areas where people [were] taking control, crime [was] being wiped out’ (Sisulu 1986: 17–18).

According to Sisulu, the key distinction between ‘people’s power’ and ‘coercion’ was whether a ‘democratic mandate from the community’ existed or not. Thus:

When bands of youth set up so called “kangaroo courts” and gave out punishments, under the control of no-one with no democratic mandate from the community, this is not people’s power […but] a “crime” (Sisulu 1986: 17).

On the other hand, when:

disciplined organised youth, together with other older people participate in the exercise of people’s justice and the setting up of people’s courts; when these structures are acting on a mandate from the community and are under the democratic control of the community, this is an example of people’s power (Sisulu 1986: 17).

This stance against criminality did not only play out in townships but also in some of the ANC training camps.18 The 1981 campaign against dagga smoking is one such example. In 1981, Moses Mabhida — head of the Revolutionary Council in Angola — blamed dagga smoking in the camps on the apartheid regime, stating that ‘the chief aim of the apartheid government was to corrupt our people and introduce bad morals’ (Ngculu 2009: 161). Because of this, he declared war against those found ‘sneaking out of the camps in search of dagga, or those found to be smoking dagga’ and camp security were given orders to shoot people caught leaving the camp without permission (Ngculu 2009: 161). Offenders were tied to a tree or locked in a windowless container as punishment. Some were severely beaten by the Angolan Regional Command, which resulted in a number of deaths (Ngculu 2009; Skweyiya et. al. 1992).

Whereas the National Party government depicted township activists as violent criminals and terrorists in an effort to deny them political legitimacy, comrades sometimes committed violent acts in the name of politics, and accused government officials and their lackeys of being the true criminals.19 When gangs looted trucks driving into the townships, some political activists presented this as part of the struggle (Jensen 2008). Similarly, what might be regarded as gratuitous acts of violence — such as ‘necklacing’ — assumed a form of political salience given that targets were accused of being spies or apartheid collaborators and thus on the wrong side of the just war. In fact both designations from the state and political movements had some substance. Amongst and alongside the ‘comrades’ there emerged activities and people who took advantage of these township struggles to wreak violence for their personal gain – hence the emergence of the label ‘comtstotsis,’ meaning criminals masquerading as ‘comrades’ (Harris 2001; Minnaar 2001; Morris and Hindson 1992; Super 2010; Jensen 2008).

The community was both a site of contestation, as well as a powerful ideological legitimating device, with armed factions legitimating their violent actions via rhetorical appeals to the ‘community’ (Wilson 2002:181; Ferndale, Malakane and Schärf 1994; Burman and Schärf 1990). It was also, as is the case today, not exactly clear what ‘the community’ stood for, nor what its boundaries — moral, physical, and geographic — actually were. Then, as now, the ‘community’ was a moving resultant in a shifting field of power relations.

Politics and Crime

In post-apartheid South Africa crime has become increasingly problematized and politicized, subjected to better measurement, parliamentary debates and a consensus on the need to treat criminals harshly (Simpson 2004; Rauch 2007; Van Zyl Smit and Van der Spuy2004; Super 2010, 2011, 2013).

Shortly after assuming power, the ANC government was faced with the problem of how to prove it was in control of the crime problem, whilst at the same time appearing to act within the liberation framework of a revolutionary organization. It therefore sought to simultaneously legitimize the previously vilified police and prove that it was not soft on crime by embarking on an eclectic range of strategies. This entailed a discursive embracing of the concept of community policing whilst at the same time uncoupling criminals from a political and social context, presenting them as a threat to the country’s young democracy (Super 2014).

At the opening of Parliament in 1995, President Nelson Mandela blamed crime and violence for ‘eroding the foundation of our democracy,’ thus necessitating a ‘harsher approach’ (DSS 1999: 9). In 1999, the Deputy Minister of Justice boasted that mandatory minimum sentences and restrictive bail laws were ‘progressive’ (PRSA 1999). 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 [sic]…is to fully rally behind the police and to declare that the fight against crime is our fight’ (PRSA 2001). 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 Bruce 2010: 9). This was a far cry from the 1992 ANC discussion document on ‘Crime and Crime Control’, which — to give one example — ascribed gang formation to ‘structural and political reasons’ and presented gang members as having ‘legitimate economic needs’ (ANC 1992: 8–9).

The Khayelitsha Commission was also mired in politics. The fact that it was established by the Premier of the Western Cape – the only province that was not run by the ANC – at the request of leftwing activists who were keen to paint it as a ‘people’s commission,’ was, from the outset, a recipe for a political spectacle. Because the Commission was specifically tasked to investigate whether vigilantism was caused by inefficient policing, it tapped into the type of mainstream discourse (Häefele 2006; Minnaar 2001; Sekhonyane and Louw 2002) that called on the state to assert its authority in ‘poor urban neighbourhoods, informal settlements and deep rural areas such as the former homelands’ (Häefele 2006: 10). This considerably narrowed the framework of inquiry, thus excising a more political approach.

The SJC embarked on an activist campaign to mobilize support for what it referred to as a ‘people’s commission.’ It arranged marches, pamphlets, booklets, t-shirts and group meetings aimed at targeting as many people as possible. The leadership urged branch members to ‘spread the message’ especially locally in Khayelitsha.20 It also embarked on a ‘Campaign for Safe Communities,’ which was launched at the University of Cape Town’s Jameson Hall in April 2013 and attended by at least 4000 people, most of whom were bussed in from the townships. The state, on the other hand, argued that the Premier had established the Khayelitsha Commission as a cheap political ploy to discredit the government and attempted to have the proceedings ruled unconstitutional. The matter was ultimately resolved by the Constitutional Court in October 2013,21 more than a year after the establishment of the Commission.

As already noted, the provincial government sought to restrict the ambit of the commission’s findings and terms of reference. Its legal counsel referred to the ‘terrible spate of vigilante killings due to the lack of trust in the police’ and ‘low per capita ratios of police to residents in Khayelitsha.’22 As far as the Province was concerned, the ‘essential focus’ of the commission was to be ‘on the functioning of the South African Police Service in the community and how to improve it’ given that the ‘police are on the frontline of the war against crime.’23 The focus was clearly not to be on the causes of crime, since as their counsel put it: ‘[we would] need to sit for five years and not five weeks if the focus was on the causes of crime.’24

Counsel for the police blamed apartheid policing for having played an active role in fomenting gangs, drugs, and vigilantes whilst stressing that it was in control of the situation and that Khayelitsha was ‘not a lawless society.’25 It argued that although policing was taking place, it was impossible to have ‘normal policing in an abnormal society…characterized by unemployment, a lack of housing […and where residents were] angry at poor service delivery and lack of sanitation.’26 This was a thinly veiled critique of the inadequate public services provided by the Democratic Alliance-controlled Cape Town Municipality. For its part, the Cape Town Municipality blamed ‘vandalism and theft’ as the reason for poor services, thus tapping into a criminological explanation based on a ‘culture of criminality.’27

The common thread running through these arguments was a lack of emphasis on public policies that address broad social inequities. Instead, the focus was narrowed down to police services and how to make the criminal justice system more efficient (Scheingold 1984; Miller 2013: 19). The fact that the government, both national and local, and most political parties in South Africa, blame crime on socio-economic conditions, whilst also supporting harsh punishment is indicative of a superficial approach that seeks not so much to explain crime as to make practical suggestions for making the criminal justice system more effective. This avoids a serious discussion by shunting responsibility to other state agencies. Similarly, calling for a partnership between the police and the community, in terms of policies of responsibilisation that place responsibility for crime prevention on the shoulders of those communities whose capacities are already severely stretched, sidesteps a longer-term and more in depth approach.

Partnerships and ‘Grassroots’ Democracy

The apartheid government first embraced the concept of community-based crime prevention during the 1970s as part of a neo-liberal partnership discourse, which rooted the ‘community’ in the notion of creating a ‘strong stable urban middle class’ (PRSA 1982; Super 2013). The state began to call on the ‘community’ and the individual to take responsibility for preventing and protecting against crime as well as managing the crime risk, acknowledging that the government could not solve such problems on its own.28

The new ANC government has also embraced a partnership discourse in its approach to crime but justifies its appeal to the community by citing the liberation struggle. This form of governance — one that operates in terms of a ‘liberation paradigm’ — valorises local-level initiatives by constantly seeking to mobilise communities on the ground (Darracq 2008). In 1992, the ANC stated that it was ‘the community who [was] largely responsible for prosecutions [and…] not the police alone who combat crime’ (ANC 1992: 8–9). Community policing — based on ‘partnerships and…sharing responsibility’ — was presented as the midwife for the ‘rebirth’ of law enforcement whilst Community Police Forums (CPFs)29 were presented as vehicles of ‘transformation’ (DSS 1996: 1; Nathan and Colin 1993).

In 1999, the Gauteng provincial government launched a public campaign to establish street committees — reminiscent of those active during the 1980s — in an attempt to engage communities to fight crime ‘constructively’ (Minnaar 2001: 40). President Jacob Zuma has also called on townships to revive street committees in terms of a very populist discourse, which harks back to the liberation struggle. Indeed, street committees do exist on most streets in Khayelitsha. Whereas during the 1980s street committees represented alternative sites of grassroots democracy, in post-apartheid South Africa most fall — however loosely — under the umbrella of the ANC-aligned South African National Civic Association. However, as Boyane Tshehla has noted, not all bodies identifying as ‘street committees’ are aligned with SANCO and — in some instances — these proto-democratic forums are merely a code for a gathering of people (Tshehla 2002).

I was told that in informal settlements there are no formally constituted ‘committees’ because of the lack of streets.30 Instead, the ‘community’ elects an ‘area committee’ consisting of about 15 members.31 These committees deal with the problems that are referred to them by residents, including crime. Larger community meetings are usually held in open spaces with attendance solicited by a loudhailer. According to one interviewee, these meetings are often run chaotically with lots of shouting and very little listening.32 One of the SANCO meetings that I attended in the formal settlement of Greenpoint started three hours late and was aborted, due to lack of a quorum.33 Leadership engaged attendees with ANC songs and dances in an effort to evoke the atmosphere of the liberation struggle. On the day that I went to interview residents in Enkanini, I drove past a group of about 30 people assembled in an open space, having a community meeting. The person that I was interviewing, sitting in the car with me, became very nervous and did not want to be seen by this group, because she was with me. She said that she feared being regarded as an impimpi.34

Initiatives such as neighbourhood watches are sporadic with many people referring to them as being a thing of the past and complaining about a lack of funding.35 At one point the Department of Community Safety offered a week of free residential training to neighbourhood watch members, as a condition for donations of equipment. 36 This led to expectations that neighbourhood watches had some official status and that volunteers would be remunerated. Yet, the consistent call is for residents in Khayelitsha to volunteer without remuneration. Similarly, community patrols are also not well-supported, partly because it is dangerous to patrol the streets of Khayelitsha at night but also because Khayelitsha’s population, particularly in the informal settlements, is largely transitory (Seekings 2013). 37

Witnesses testifying at the Khayelitsha Commission had either not heard of CPFs or testified that they were weak structures, meeting irregularly, with poor attendance, and were badly funded (Khayelitsha Commission Record 2014: 2419, 2681, 2701, 2708). Yet, despite the fact that CPFs appear to be largely cosmetic structures with acute resource problems, official discourse refers to them as being ‘elected by an open and democratic process’ and as an important ‘conduit between the police and the community’ (Lamoer 2012). The 2002 SAPS Vigilantism Prevention Strategy for the Western Cape was premised on improving the performance of CPFs and effective implementation of Sector Policing. This was to be coupled with ‘community based crime prevention by means of partnership policing’ (SAPS Vigilantism Preventative Strategy cited in Häefele 2006: 11). This striving to reform and improve what may well prove to be a chimera, a constantly failing project of the ‘community,’ is twinned with a responsibilising discourse that ‘the community’ must take responsibility for solving crime on its own. This is not necessarily a good thing and often results in exclusion as I discuss in the next section.

Banishment and the ‘Moral Community’38

Banishment as a means to deal with crime is widespread.39 It takes the form of lawfully imposed prison sentences, detention as an awaiting trial prisoner as well as unlawfully imposed sanctions. The latter include informal curfews, the demolition of individual residences, instructing family members to send the alleged wrongdoer away from the area,40 instructing homeowners in formal settlements to sell their houses, and finally, killing.

I was shown vacated shacks in Enkanini after the ‘community people’ had asked four alleged drug dealers to move and ‘to take their materials and clothes.’41 Another person told me that he was a member of three different committees, all with varying relationships with the state: the local neighbourhood watch, the Mayitshe Group, and the Khayelitsha Community Police Forum. He had also been a participant in the demolition of three houses in an informal settlement after a decision to this effect was taken at a ‘general council’ of local street committees.

It appears that street committees give permission for people to move into an informal settlement and that this permission can be revoked in the case of wrongdoing. The situation in a formal settlement is different since the owner of a house will presumably have a title deed to the property. In this case, the person is ‘asked’ to leave.42

A police colonel testified before the Khayelitsha Commission that, apart from ‘mob justice incidents,’ he was aware of formal meetings that had resulted in a decision to evict people from their homes due to a crime (Nel 2014: 4635). As he put it:

It might be a child molestation case or it might be a housebreaking cases where these formal meetings have taken place but none of these were accompanied by violence. These people were just motivated to leave, which they then did.

A senior member of the Greenpoint SANCO told me that the street committee’s job is to interview the suspect, ‘take all the details,’ and — in the event of ‘non co-operation’ — remove them from the area, but only as a ‘last resort measure.’43

A member of the area committee in Site B told me that if the accused in a ‘crime of stabbing’ was released from custody, the ‘community’ would always ask him to leave the area. He insisted that there were never any incidents of violent vigilantism in his neighbourhood but that ‘when the community gets mad, so you will move, you will do what they say.’44

I was also told that a young man staying alone would not be ‘banned’ but ‘asked to leave nicely.’45 If he was staying with his parents then ‘we speak to them and they decide what to do: maybe they send him back to the Eastern Cape.’ In instances where he was staying with a grandparent, who was too old to play a supervisory role, then, if he refused to leave, the community would either destroy the home (bash it down with hammers) or threaten to sell it: ‘all I can say it depends how the person interacts with the community.’46

One witness at the Khayelitsha Commission testified that when the community demanded that her nephew leave the area due to his alleged ‘criminality,’ the family did not argue and was unable to think of any other options. As she put it: ‘Our main concern was that he should leave the house so he wouldn’t be harmed. We could see the mood of the residents and it appeared that they would do something.’47

The power implications are plain to see and the degree of social capital that the wrongdoer enjoys at particular points in time obviously plays an important role in the decision that is made by ‘the community’ and how it is reached. Clearly, although the threat of violence might not be overt, there is a form of coercion, a subtle/implicit threat of the violence that will occur should the person not leave ‘voluntarily.’ It would certainly be worthwhile to do more research on the right to residence in informal and formal areas and how these interplay with civil society organisations such as street committees.

Violence to Retrieve Property

As is typical of most townships, Khayelitsha does not rank high on the property-related crime scale. In fact, property-related crime is mostly recorded in suburban and Central Business District station precincts (DCS 2009: 11). This is borne out by research conducted by Jeremy Seekings, who found that a far smaller proportion of Khayelitsha residents said that they would report house-breaking to the police than in white/coloured neighbourhoods (Seekings 2013: 24). Instead, a significant minority of respondents said that they would solve the problem locally — through friends, neighbours, or local organisations — rather than contact the police. Research commissioned by the Department of Community Safety found that Khayelitsha residents were more ‘supportive of one another with regard to their fight against crime’ than in respect of ‘one another’s personal dilemmas’ (DCS 2011: 93).

I was told by a mother that the community beat her son to death because he was always in trouble and was suspected of having stolen a cellphone.48 A teenager living in a tin shack in Enkanini told me that had she seen the thieves who broke into her home, she would have alerted the ‘community’ to assist her in retrieving her goods.49 An executive street committee member stated that in certain instances the ‘community’ would closely observe a suspect (i.e. ‘each and every step he takes’) to make sure that he did not commit further crimes.50

In the trial that I observed, state eyewitnesses testified that had the accused followed the standard practice of reporting the incident — in this case, a stolen TV — to the street committee rather than the police, the deaths of the suspects could have been avoided. Although the police are called in serious cases such as murder — or cases involving bloodshed more generally — in the case of theft, the street committees prefer to sort the matter out; this stems not from a mistrust of the police, but rather a sense of maintaining ‘harmony ‘and teaching people to ‘live politely.’51

In general there is a high degree of community support for violent reprisals against suspected robbers and thieves. Even SJC members support the beating of ‘criminals’ by ‘community members’ as a technique to retrieve stolen goods, although they claim not to participate in violent activities. This might seem surprising given the SJC’s leftwing human rights stance but it is also indicative of how theft of a cellphone or plasma TV assumes great importance in a context of deep scarcity and endemic societal inequality. One person told me that the rate of crime had decreased in his area because the troublemakers had been ‘taken out.’52 He stated that ‘smacking’ is not a punishment but qualified this by stating that ‘if we take law into our own hands we are promoting apartheid’.53 Another told me that although ‘the community’ is ‘civilized,’ due to a ‘lack of justice people will lose their patience’ and ‘violence is always lurking’ with ‘decent people…driven to acts of desperation.’54 Thus, on the one hand there is complicity and, on the other, there is outrage.

According to a 2012 police report between April and June 2012 there were 78 recorded ‘vigilante incidents’ in Khayelitsha (SAPS unpublished).55 These all resulted in death with most victims being young men aged between 18 and 30. At least half were caught — or suspected of — stealing/robbing/housebreaking and ten had previously been released from prison or remanded to detention.

In the court case that I observed, one former street committee member testified that certain individuals had been tasked with administering ‘lashes’ (euphemistically referred to as ‘lectures’) in order to elicit a confession from a suspect and retrieve stolen property.56 This ties in with Rebekah Lee and Jeremy Seeking’s observations that there appears to be a tacit acceptance of ‘violent forms of vigilantism’ if it is initiated by, or has consent of, street committees or other local institutions (Lee and Seekings 2002: 114).

However, I was also told that, ‘due to ubuntu,’ street committees no longer authorised corporal punishment against suspects.57 Instead, if the suspect denied responsibility or refused to give the goods back — and if there was sufficient evidence — then a charge would be laid with the police.58 However, in those instances where there was an admission of theft, the responsible party would either have to buy a replacement or make financial restitution. Sometimes the community would go to the wrongdoer’s home and decide what to take in lieu of monetary payment.59 Despite the fact that overt violence was denied, it would seem that there is still some form of investigation and/or coercion with — at the very least — the threat of an official charge being laid with the police.

A former member of the Amadlozi,60 now living in Khayelitsha, explained that there was:

…nothing wrong with the victim going to the police and getting a case number — just to make sure we are covered — because for us it’s important to have a case numbe …as members of the community we want to be on the safe side. Because when we find the suspect it’s important to get goods back before they get taken to the Eastern Cape where they get sold and you never see them again. We follow the lead and get the suspect – if the suspect is willing to talk (because all is pointing to him) there is no need for a massage and if he doesn’t talk yes he will get a little massage and he talks and we find the goods and it’s only then we take him to the police.

His reply to the question of why he would take the person to the police — even after retrieval of the goods — was that prevention was ‘better than cure,’ that ‘the law’ had to ‘take its course,’ and that ‘the judicial system must do its duty: we already assisted the Investigating Officer. We made the job easier.’

Outsiders on the Edges

In part, whether or not one survives a vigilante attack — or is attacked at all — depends on the social capital that the person and their relatives have within the area. I was told that:

…if I stole something from my own community — in my own area — I would be beaten by the community but I wouldn’t be killed, unless someone came there drunk and beat me in my head.61

As the above interviewee put it: ‘It would seem that there is always more chance that someone from another community would be killed.’62

A female who was beaten by a community patrol for suspected drunkenness told me that this wouldn’t have occurred in her own neighbourhood. She also stated that when the Somalian shopkeepers63 in her neighbourhood were robbed, no one stood up for them; conversely, if she were robbed the community would find and punish the offender.64 This interviewee referred to a mugging incident that had occurred in her street in April 2014. It resulted in community members chasing the two suspects and catching one, in an adjacent informal settlement, where he was severely assaulted with spades, sjamboks, sticks and stones. According to her, he broke his leg: ‘he was a mess but he survived.’ Others are not so ‘fortunate.’

It seems, therefore, that even spontaneous ‘mob’ justice, of the deadly type, is not totally unstructured but targeted at those who are perceived to be outsiders and as such functions as a mode of boundary setting. This ties in with Bruce and Komane’s (1999: 44) findings that:

The actions of vigilantes should not necessarily be seen to represent a generalised intolerance of criminality but rather a selective hostility to the criminality of ‘outsiders.’

These outsiders are also constructed as such in terms of official (state) discourse on crime. In 1994, the Minister of Justice stated that ‘the influence of a large number of illegal aliens in South Africa…had a significant impact on the incidence of criminality’ (PRSA 1994). Official statistics also distinguish between citizens and non-citizens in the context of arrests for commercial crimes, thus contributing to the alienation and criminalization of foreigners (Super 2013). Moreover, official discourse has also de-linked criminals from their political and social context, thus contributing to their outsider status and exacerbating perceptions of their otherness.


Just as Douglas Hay (1975) argued that in 18th century England the criminal justice system entrenched Whig power, so too has the South African Constitution and the ideology of grassroots democratic struggle made it possible for the ANC to govern within a contradictory and deeply contested macroeconomic framework. In the ‘frontier society’ of Khayelitsha, the governance of crime manifests as a ‘pathological’ form of grassroots democracy with organized and unorganized interests all pressing for more punishment and a more effective criminal justice system (Garland 2013: 508). The fact that poverty is unevenly distributed in Khayelitsha — with pockets of relative affluence — makes it a good breeding ground for what Jock Young has described as the ‘relative deprivation’ that characterises ‘incomplete meritocracies’ (Young 2007: 37). This sense of being deprived occurs in both directions: those who are less well-off feel deprived vis-à-vis their more well-off counterparts whilst the latter feel unfairly encroached upon by the less well-off.

Clearly, the specific historic conditions of high crime rates, social insecurity, political underrepresentation, as well as under-enforcement (stemming from segregationist apartheid crime control policies), have impacted contemporary penality in Khayelitsha. Paying greater attention to crime has the potential to address the deeper criminogenic conditions that give rise to it, but — given the ease with which a repressive form of penal power may be imposed — it is understandable that punitive populism has arisen together with demands for the alleviation of criminogenic conditions. Of course, the criminal justice system could be improved upon but first the debate must be broadened to address both inequality and poverty.

This paper has argued that there are clear precursors in the punitive treatment of suspected criminals that played out during the struggle to end apartheid. Indeed, punishment in South Africa has historically been relatively unconstrained by the minimalist considerations associated with liberalism and its democratic tradition has been aligned with — rather than against — harsh punishment. Prevailing ideology in South Africa romanticizes the discourse of the liberation struggle by deploying the ‘community’ as a node of public participation and CPFs as democratic political institutions. Despite the political rhetoric about ‘community,’ there are many fractures and divisions within those very communities that are most in need of cohesion. The result is that the multivalent ‘community’ is deployed in various ways to justify harsh action against criminals, by a whole range of actors in terms of a discourse that weaves the notion of democracy into the discourse of partnership. When ‘civic engagement’ (Barker 2013: 134) in and with crime occurs in the context of blocked opportunity and relative deprivation, the ‘community’ can become a place of exclusion, giving rise to popular punitivism. Those who are marginalized by poverty and inequality live in conditions of extreme risk and the politics of resource allocation in townships such as Khayelitsha are masked by the labels of ‘mob’ justice or ‘vigilantism.’

Because of the distorted institutional dynamics that not only shape the nature and character of public participation in South Africa but also form a crucial component of the relationship between democracy and punishment, it is crucially important to research these dynamics. In particular we should, as Huggins (1991:16) suggests, analyze the dynamics between the state and social organizational structures, between various local organizations and between various government agencies (local, provincial and national) that shape and promote vigilantism in particular and punitive populism in general. In sum, we should focus on the ‘structural conditions’ that enable ‘collective violence’ (Garland 2010: 36). These do not only encompass macro-socioeconomic conditions (such as inequality), although this is obviously of crucial importance, but also factors such as the extant ‘popular sovereignty,’ the power struggles between ‘local actors,’ group relations, levels of violence, and the presence of ‘despised low-status outsiders’ (Garland 2010: 36) – such as ex-prisoners, people released on bail and foreigners. Instead of presenting vigilantism as a form of ‘mob justice’ (Minnaar 2001; Häefele 2006; Sekhonyane and Louw 2002), as a scourge, as inimical to ‘civil’ society, and as being somehow outside of and opposed to it, we should acknowledge how vigilantes, or at least their supporters, are in fact part of ‘civil society.’