Centre for the Study of Violence and Reconciliation

"As for Violent Crime that's our Daily Bread":
Vigilante violence during South Africa's period of transition

by
Bronwyn Harris

Violence and Transition Series, Vol. 1, May 2001.

Bronwyn Harris is a former Project Manager at the Centre for the Study of Violence and Reconciliation.

Acknowledgments

We would like to thank all the people we spoke to during the course of our research for making the project possible. We are also indebted to Brandon Hamber and Piers Pigou for their editorial support, to Elrena van der Spuy for her comments on the report in draft form and to Wendy Meyer for making the contacts that needed to be made. Also, thank you to Wardie Leppan, Project Officer, International Development Research Centre (IDRC), for his ongoing support of, and insight into, the Violence and Transition Project.

This booklet was funded by International Development Research Centre (IDRC).

The Violence and Transition Series is a product of an extensive research project conducted by the Centre for the Study of Violence and Reconciliation (CSVR) into the nature and extent of violence during South Africa’s transition from apartheid rule to democracy. This series comprises a set of self-contained, but interrelated reports, which explore violence across the period 1980 to 2000 within key social loci and areas, including:

While each report grapples with the dynamics of violence and transition in relation to its particular constituency all are underpinned by the broad objectives of the series, namely:

Through these objectives, the Violence and Transition Series aims to inform and benefit policy analysts, government officials and departments, non-governmental and civic organisations, and researchers working in the fields of:

As a country emerging from a past characterised by violence and repression South Africa faces new challenges with the slow maturation of democracy. Violence today is complex, dynamic and creative in form shaped by both apartheid and the mechanisms of transition itself. In order to understand - and prevent - violence during transition in South Africa and abroad an ongoing action-research agenda is required. Through this series the Centre for the Study of Violence and Reconciliation offers an initial and exploratory, yet detailed, contribution to this process.

The Violence and Transition Series is funded by the International Development Research Centre (IDRC).

Work on the project focusing on the Truth and Reconciliation Commission was supported by the Embassy of Ireland, the Charles Stewart-Mott Foundation and the Scottish Catholic International Aid Fund (SCIAF).

Series Editors
Brandon Hamber
Piers Pigou
Bronwyn Harris

Contents:

Executive Summary

This report explores vigilante violence in South Africa over the period 1980-2000. Based on interview and media material, it traces the spectrum of vigilantism across this timeframe in terms of history, form, method and motivation. The similarities and differences in vigilante patterns, manifestations and causes are also analysed. These are contextualised within South Africa's formal transition process, particularly as they link to the changing criminal justice system. Explanations that go beyond the formal criminal justice system are similarly explored and vigilante violence is interpreted as symptomatic of a 'culture of violence' within South Africa. The findings of this report are summarised below.

This study finds that vigilante methods across the two-decade period have remained fairly stable and that they generally share the following qualities:

Violence, especially in an extreme form of corporal punishment, is an integral feature of vigilante methodology. In the majority of vigilante incidents examined in this study, violence most commonly resulted in death. This suggests that death is often the intention of vigilantism and is built into the form that the vigilante act adopts. Alongside the commonalities of vigilante methodology, these actions are also analysed for the functions that vigilante violence itself serves. On an immediate level, violence serves a punishing function. In addition to this punishing, often fatal function, vigilante punishment also serves a symbolic role which extends beyond the immediate victim to intimidate those witness to the vigilante spectacle. Contained within this symbolism is a preventative function which serves to warn, and thereby control, the future actions of both victim and audience. This section of the research concludes that the efficacy of vigilante violence pivots on fear and operates through the very public, visible nature of vigilantism.

While vigilante methods have generally remained consistent across the period 1980-2000, definitions and explanations of what constitutes vigilante violence have altered considerably. This study suggests that in the pre-1994 period, vigilante actions were defined by recourse to politics and political intention. Generally, the vigilante label was applied to violent actions conducted in support of the apartheid state, rather than those carried out against it. This analysis notes that during the pre-1994 period, 'vigilantism' was a negative term applied to illegitimate and violent actions that were seen as being politically motivated. Its meaning also depended largely on the perspectives adopted by the users of the term.

These features of vigilantism indicate that the pre-1994 definition was descriptive of 'conservative vigilantism' and that it served to disguise violence conducted by roleplayers with different motives e.g. criminal, financial or personal; motives which might today qualify resultant actions as 'vigilante' violence. In this report, findings of the South African Truth and Reconciliation Commission (TRC) are analysed and are seen to support the political basis of the pre-1994 definition. However, this study also recognises that in the post-1994 era, there has been a selective expansion in the description of past vigilante violence to include certain activities conducted by groups resistant to apartheid e.g. violent internal policing practices of the self-defence units (SDUs). This retrospective re-labeling coincides with a general concern about crime during the post-1994 period.

In the post-1994 era, vigilantism has been predominantly defined by recourse to 'fighting crime'. This research project explores the relationship between crime-fighting vigilantism and the formal criminal justice system (CJS). It finds that vigilante violence is frequently justified as 'filling a policing gap' due to police inefficiency, corruption and compliance with criminals; practical failings in the CJS. Additionally, the apartheid history of the CJS is seen to linger in community mistrust and fear, thereby providing a justification for extra-legal violent actions. Vigilantism, as a practice modeled on instant justice, is also analysed within the dynamics of South Africa's political transition. In this explanation, it is located within a 'transitional gap' created in the shift away from repressive state control and heavy-handed vigilante-style policing methods, to police 'transparency' sans draconian powers. Here, vigilantism is seen to represent a privatisation of the old-style policing function.

This study also considers a related 'transitional argument' that locates vigilante actions within the legitimacy of the political order. From this perspective, vigilante violence is defined as 'new' to the South African landscape because the current political dispensation, unlike its predecessor, is seen to be legitimate. In this explanation, the perceived legitimacy of the political order is seen to co-exist in an inverse relationship with vigilante actions: vigilantism occurs in a legitimate political climate, while 'crime-fighting' originates under an illegitimate order.

This analysis also suggests that South Africa's transition to democracy has changed perceptions of crime, as well as crime patterns. Crime, for some respondents, is represented as fallout of apartheid, as something that, for many, remains 'okay' in the current order. However, this study suggests that for various respondents, when crime turns in on itself and starts to occur within certain communities, rather than beyond them, the foundations for vigilantism (as a response to crime) are established. Other respondents are shown to suggest that vigilantism is more a product of expectations about the political transition, specifically, disappointed expectations about the political change, than a response to changed patterns of crime. On one level, this reflects the degree to which due process remains an alien concept for many who have yet to benefit from the constitution. It similarly implies a lack of education about due process, particularly bail laws. However, this research project explains that vigilantism is not solely reducible to disappointed expectations and practical failings of the criminal justice system. Rather, on another level, this analysis suggests that the human rights framework of the formal criminal justice system is fundamentally divergent from the model of instant, popular justice. This is partly because vigilantism occurs for many reasons other than fighting crime. These reasons include justifications that draw on:

This research also indicates that the crime-fighting justification for vigilante violence is frequently used to disguise motives of personal gain and 'pure crime'. More generally, it strives to emphasise the link between vigilantism and crime.

To accommodate the range of explanations and possible causes of vigilantism, this project moves beyond them to analyse vigilante violence as part of a broader 'culture of violence' within South Africa. In this framework, violence is upheld as the normal and legitimate solution to problems. Vigilantism is explored as a symptom of violence in transition. Using this as a model for contextualising contemporary vigilante violence, the 'way forward' is plotted out in general terms for intervention and prevention. These terms are set within a framework that aims to delegitimise violence and inculcate a culture of human rights. They include:

Underpinning the 'way forward' is the recognition that further research is needed into general trends and specific manifestations of vigilantism. This report suggests that vigilante violence is not one singular phenomenon but that rather, it occupies a spectrum of forms and methods. It is not static but exists in the context of transition, as a symptom of South Africa's changing culture of violence. In this light, the report recommends a holistic approach that is complemented by localised interventions and ongoing, in-depth research.

Introduction

Vigilantism is a blanket term for activities that occur beyond the parameters of the legal system, purportedly to achieve justice. It covers a wide range of actions and involves an eclectic assortment of perpetrators and victims. Vigilantism can exist as an isolated, spontaneous incident or as an organised, planned action. It traverses local communities, as well as regional areas. Broadly, achieving justice is the apparent motivation for, and desired outcome of, vigilante actions. For this reason, vigilantism is often defined in relation to the formal criminal justice system and the law. However, justice itself takes on different meanings in different contexts. Similarly, South African history testifies to the changing nature of legislation, as well as its potential for injustice. Vigilantism thus subsumes a number of dimensions. This makes the exact meaning of the term very difficult to pin down and analyse. Rather, contemporary vigilantism in South Africa must be viewed as taking on a spectrum of different forms, methods, motivations, and causes. This spectrum itself must be contextualised in the history of the country because the meaning that 'vigilantism' has today cannot be separated from apartheid politics.

The intentions of this research report are threefold:

This analysis comprises media reports on vigilantism and primary research material derived from in-depth interviews and focus groups held in four specific localities; namely, Alexandra (in Johannesburg), Mamelodi (in Pretoria), and Guguletu and Khayelitsha (in the Western Cape). The material is supplemented by key informant information gathered in Johannesburg, Cape Town, Tembisa, and the Northern Province.1 Although these areas are united by certain similarities in the expression of 'vigilantism', they are also differentiated by specific geo-political factors. It goes beyond the scope of this project to detail and compare these differences. Rather, location-specific factors will be introduced to highlight the role played by unique features in combination with the more common components that produce a vigilante configuration. As such, this report is intended to provide an overview of vigilantism in South Africa. For more detailed, case-specific knowledge, this research makes use of CSVR reports on related topics such as: Mapogo a Mathamaga (von Schnitzler et al 2001,); Pagad (Dixon & Johns, 2001); and taxi violence in the Cape Peninsula (Dugard, 2001).

Background and context: the vigilante spectrum

History

Broadly, the term 'vigilantism' takes on certain meanings across two periods of time in South Africa: pre-1994 and post-1994. Literature reveals that during the pre-1994 era, the term is generally descriptive of politically motivated violence. Between the 1980s and the four years from 1990 to 1994, there are subtle shifts in how such violence is described but, in essence, the term remains politically based. During the 1980s, for example, the term is closely aligned with conservative or right-wing motives and it is used to describe actions conducted in support of apartheid (cf. Haysom, 1986; Bruce & Komane, 1999). During the early 1990s, the phrase continues to describe political motives but it is expanded beyond 'conservative vigilantism' to explain a wider range of actions in terms of 'beneficiaries' and 'victims'. In this period, 'vigilantism' was used as a catch-all phrase to describe violence that was 'unexplainable' (cf. Jeffery, 1992; Coleman, 1998). By contrast, during the post-1994 period, vigilantism, like most other forms of violence, has (largely) become associated with crime rather than political motives. The term 'crime-fighting' represents the most common banner of justification for violent vigilante action in this post-1994 period.

The following section offers a closer interrogation of the term's history. It aims to capture the differences and similarities in vigilantism across this history and to show that commentators and history itself constantly define the past and shape the present understanding of what vigilantism means in South Africa.

Pre-1994

The notion of vigilantism occupies a particular space in South African history. It is rooted in violent actions that were generated specifically in relation to the apartheid system by specific roleplayers. In the apartheid context, the term 'vigilantism' generally conveyed violent actions that were political, or interpreted as such. In other words, vigilante actions were politicised, even although they were not always directed at political targets (This point is explored in detail below). During the 1980s, an explicitly political form of violent action predominated as vigilantism. This took the form of what has been termed right-wing, or conservative, vigilante action (Haysom, 1986; Bruce & Komane, 1999). Haysom (1986) documents this form of vigilantism, suggesting that three common features united a range of different vigilante groups across the country at that time.

Firstly, in most cases, the vigilantes emerged in the latter half of 1985. Secondly, the vigilantes shared a common target group, members or leaders of groups associated with resistance to apartheid or homeland rule. Thirdly, the vigilantes operated brazenly, apparently believing that they enjoyed police support, and in some cases did allegedly enjoy such support. (Haysom, 1986, p. 1)

The role of the apartheid state in supporting and/or actively initiating such conservative vigilante action has been explored by the South African Truth and Reconciliation Commission (TRC). The final TRC report finds that:

[while] not all vigilante activity was a product of state engineering … . Commission investigations produced evidence of a range of levels of endorsement, support and management of vigilante groupings by different security arms of the state. (TRC Report, 2 (3), p. 571)

To a large degree, the TRC history book has therefore relegated pre-1994 vigilantism to right-wing activity, often with overt state sanction.

It is not only 'official' history that interprets the past in this way. Many interview respondents offer a similar definition of vigilantism during the period. Consider the following responses from residents of Alexandra:

We did have vigilante groups in Alexandra, especially at the height of political violence. They were state-sponsored vigilantes, who painted themselves with black paints, pretending to be part of the warring factions, but in fact they are vigilante groups and they've clubbed with the boers. (AWH)
The IFP was used by the police to fight against the community. (AWS)
The police during the time of violence in Alexandra were the puppets of the government to kill the people of Alex. (AWS)
[During the past] there was a lot of conflict [between hostel dwellers and the Alexandra community] because the hostel dwellers were assisting the police to fight against the community. In fact they were used by the government. This was a political conflict engineered by the apartheid government. (AWS, p. 8)

These responses locate vigilantism within the hands of the apartheid state. They represent vigilantes as state agents, taking the form of the police, the Inkatha Freedom Party (IFP), and hostel dwellers. For these respondents, the conflict is clearly connected to politics at both a narrow level, where violence is ascribed to particular political parties such as the IFP, and at a more general level, by reference to the government and 'the boers'. In this understanding, it is the violent actions conducted in support of the apartheid government that are considered to be vigilante.

An analysis of contemporary vigilantism in post-apartheid South Africa must recognise the particular connotations attached to the term 'vigilante' during the 1980s and early 1990s. This is important because the past meaning of vigilantism is specifically descriptive of the political underpinnings of actions, rather than the actions per se. At the time, this had the significant consequence of excluding other actions and actors (such as those resistant to apartheid) from being labelled 'vigilante'. A labelling asymmetry is therefore implied in the application of the term. This asymmetry persists in the final report of the TRC. For example, in chapter 2, section 3, apartheid resistance actions are described as part of a 'Campaign against Vigilantes, Kitskonstabels and Municipal Police' (paragraph 258, emphasis added). These resistance actions include necklacing and corporal punishment methods not dissimilar from those actions that are ascribed to vigilantes (see Appendix A for TRC text on this). All that differentiates them is their political motivation within a particular context. Indeed, when these actions occur today, they are usually framed as vigilantism.

What such a specifically applied label does is create a direct link between motive (that of apartheid-endorsed politics) and vigilantism, rather than between action and vigilantism. This link is clearly evident in the definition utilised by the then Human Rights Commission to monitor vigilante violence during the early 1990s:

the HRC states that 'vigilante groups are reactionary forces that arose out of the attempts of homeland administrations and black local authorities to defend their vested interests against their rejection by the communities in which they are located. Such 'private armies' concentrate their attacks on community structures that press for the dismantling of these apartheid organs, but of late have considerably widened their scope to destabilise township communities at large. In reporting on vigilante-related actions, the HRC includes not only attacks by vigilante groupings, but also retaliatory or pre-emptive measures taken by the affected community in a vigilante-initiated situation'. (Jeffrey, 1992, p. 43)

As this extract reveals, even when vigilante actions were not directly targeted at apartheid's political opponents but rather destabilised a range of victims, the definition ultimately rests on political motivation. 'Retaliatory or pre-emptive measures' are included in an understanding of vigilantism but only because of 'a vigilante-initiated situation' (Jeffrey, 1992). While broader in scope than vigilantism in the 1980s, the phenomenon in the early 1990s retains its conservative underpinnings through definitions of the HRC kind. This is because the destabilisation campaign was largely understood in terms of beneficiaries and victims, which in the apartheid context, allowed for a politically based analysis. Alexandra residents reiterate this point, suggesting that vigilantes in the township created 'havoc' for an end that was ultimately political:

R1: And some of the people who were [taking the law into their own hands], they were getting paid to do that by the whites …
R2: To [beat] their own brothers.
R1: Ja, or to steal or to do whatever, to cause the havoc they needed. They were getting paid. They paid. We were having these people, we call them impimpi. (AVSG).

In this extract, the source of pre-1994 'havoc' is located within 'impimpi' or (state) informers. Their crimes of beating, stealing and 'whatever', are interpreted as a means to an end rather than an end in themselves. And this end is political. Linked to the politically motivated, rather than action-based, definition presented by many commentators, it appears that the label for pre-1994 vigilante violence is applied according to criteria of illegitimacy. For the respondents above, certain actions during this period were illegitimate because they were understood to support the state. Due to this perceived illegitimacy, they were termed 'vigilante actions'. This definition is based on the illegitimacy of violence motivated by reactionary interests. Similarly, these actions remain illegitimate from the retrospective vantage point afforded by the TRC and, therefore, they retain a 'vigilante' label. However, other perspectives, such as that offered by Jeffery (1992), do not restrict vigilantism during the pre-1994 era to state-endorsed actions alone. For example, she challenges the ways in which vigilantism is defined by the then Human Rights Commission (HRC). She notes that the HRC definition is 'so broad that both victims and perpetrators can effectively be tarred with the same brush' and that 'on [the basis of this definition] the HRC is able to attribute 86% of deaths in violence to surrogate state forces and effectively to exonerate the ANC and its allies from any blame' (p. 44).

Jeffrey's (1992) critique rests on the notion of responsibility. For her, it is not political illegitimacy that defines vigilantism during the early 1990s but rather agency and agent accountability for violent actions. And these agents, according to her, cannot be fully reduced to 'surrogate state forces'. Implicitly, her argument is for the term to include the ANC rather than 'exonerate [it] and its allies from any blame'. In this way, she advocates a broadening of the 'vigilante' label beyond conservative, state-supported actions. It is difficult to assess Jeffrey's (1992) own political agenda within the confines of this report and her argument is presented only to reveal that the term was contested during the pre-1994 era, even although most commentators aligned it closely with conservative, right-wing violence. (Any future genealogical studies of vigilantism might consider a deeper analysis of the reasons for Jeffrey's (1992) argument).

Lack of consensus about the term 'vigilantism' suggests that its meaning reflects the political views of the people who use it. For example, it would be interesting to analyse the interpretation of 'vigilantism' from the perspective of the apartheid government. Such an interpretation would doubtlessly differ from that given in the literature utilised here.2 The contestation of the term also suggests that the popular image of 'Robin Hood-type' vigilantes, who gallop in to save the day and ensure that justice prevails, is not one that was utilised during the pre-1994 period in South Africa; that is, vigilantism is not defined by recourse to justice in this period. No-one wanted to take ownership of the term then; it was purely negative. By suggesting that the term 'vigilantism' is relative to who defines it, the intention is not to relativise any moral arguments about actions against apartheid. Rather, it is to reveal the political expedience served by the term at the time. It strengthened anti-apartheid efforts by revealing the illegitimacy of conservative forces, and thereby the state itself. However, at the same time, it allowed the state to divorce itself from certain actions by appealing to their illegitimacy too. This is because during the pre-1994 period, 'vigilantism' was a negative term applied to illegitimate and violent actions that were seen as being politically motivated. Its meaning also depended largely on the views of the users of the term.

The contingency of the meaning of 'vigilantism' on 'who' defines it, 'when', and 'why' is also revealed by retrospective commentary about pre-1994 vigilantes. What can be termed 'historical relativism' comes into play with the distance of time and a changed social order. For example, recent analyses do not restrict past vigilantism to conservative political groups and state forces alone. Rather, they also apply the label to certain operations conducted against apartheid. Thus, Minnaar (1999) explains that:

During the late 1980s and early 1990s, years when political violence was at a high level, and prior to the April 1994 elections, there occurred many incidents where township residents took the law into their own hands. Much of this community justice was labelled political and took the form of People's Courts (in some areas they were known as Disciplinary Committees which were often no more than kangaroo courts in action) and 'necklace' executions of political opponents. However, in reality a large portion was nothing more than certain people using community anger for their own purposes to get rid of political opponents or rivals or taking the law into their own hands for revenge and to impose control over certain communities through intimidation. (p. i)

Similarly, Bruce & Komane (1999) include certain actions by the self-defence units (SDUs) under the vigilante heading. They explain that the SDUs first formed during the early 1980s and 'justified their existence on the grounds that communities were under attack from hostile 'external' forces' (Bruce & Komane, 1999, p. 41). For Bruce & Komane (1999), as for many commentators and TRC amnesty applicants, an explicitly political agenda directed against apartheid does not qualify the SDUs for a 'vigilante' label. However, they explain that the SDUs 'also became involved in imposing "justice" within communities' (Bruce & Komane, 1999, p. 41). It is this internally focused function of self-administered justice that earns the SDUs a 'vigilante' label for Bruce & Komane (1999). The internal 'justice' function of the SDUs and their connection to vigilantism is also alluded to in the TRC Final Report. Although the label 'vigilante' is not explicitly utilised, many findings on internal SDU 'policing' actions are framed as such and the notion of 'revenge attacks' is similarly introduced to condemn certain SDU actions. For example, in Volume 2, Chapter 7, the TRC finds that:

While the ANC leadership has argued that its members were acting in self-defence, it is the Commission's view that at times the conflict assumed local dynamics in which proactive revenge attacks were carried out by both sides. This situation was exacerbated by high levels of political intolerance among all parties, including the ANC. Further, the Commission contends that the leadership should have been aware of the consequences of training and arming members of SDUs in a volatile situation and in which they had little control over the actions of such members. (Volume 2, Chapter 7).

Gear (draft report) provides a detailed overview of the internal policing functions of SDUs. She explains that they adopted a 'crime-fighting' role in response to the policing vacuum which developed in the townships during the late 1980s. Bruce & Komane (1999) similarly note that 'as state control over the townships collapsed, the SDUs were able to step into the gap as alternative policing structures' (p. 41). Their internal policing function was thus premised on a failing and illegitimate political order.

A respondent from Mamelodi contextualises the SDU policing function:

In the late 80s, I was a comrade myself … . Now what transpired then, there was this sort of rowdy element. At first we thought we will sort of control crime, you know, housebreaking etc. etc., when this whole thing started. So we started rounding these guys up, warning them not to do that but it went completely out of control because some criminal elements took over the structures and started operating them, like their own little spaza shops or little cafes, where they were actually benefiting. (ME).

This extract suggests that the initial policing intention of the SDUs was to control crime committed by 'rowdy' community members. However, as the respondent suggests, the SDUs themselves were vulnerable to 'criminal elements' who 'took over the structures' for their own personal benefit. An Alexandra resident notes a similar phenomenon:

There were people [who took advantage of the political situation and became comtsotsis and advanced criminal activities]. It was good for them because it was not very clear whether they were doing criminal activities in pursuit of political objectives or not … . [Unlike real comrades] comtsotsis would steal the properties and use them for their own personal agrandisement. (AWS)

Officially, the SDUs distanced themselves from such 'criminal elements', or 'comtsotsis', suggesting that the latter were roguish, rather than 'real SDUs'. This rejection was couched in political terms:

We had marshalls [who would go and] speak to the criminal elements, to say to them: Look, what you are doing is wrong. You know we have to focus our attention on specific targets, and our energies, but because of this, crime is also bringing the police into the townships. (ME)

Gear (draft report) notes that this was a typical response amongst SDUs at the time; a response that rested on the politicisation of crime.

In response to the policing gap which existed in black townships [during the 1980s] the 'comrades' took on a policing role during the struggle. But criminal activity also became politicised as the 'comrades' perceived it to weaken the base of the liberation struggle.

When appeal to political solidarity failed to reign in criminal activity, the SDU response was, according to the Mamelodi respondent, one of 'barbaric justice':

But the criminals just could not listen to that [appeal to political unity] and it became worse with the involvement of the ANC, when people felt, in a way, on top [morally superior] and started dishing out this barbaric justice to the people. And the whole thing [of kangaroo courts] picked up with the unbanning of the ANC. (ME)

For this respondent, the 'barbaric justice' meted out by SDUs/ANC comrades approximates vigilantism. He recognises that it links to political objectives, within a particular political context, but also comments that many victims of SDU kangaroo courts were 'criminal elements', rather than direct political opponents (even although their criminal activities were phrased as political opposition). The subtle distinction between politicised criminals also known as 'comtsotsis' and political sellouts and informers, or impimpis, is important. This distinction, which is made retrospectively by the respondents, suggests that SDU violence was differentially motivated according to the criminal or political identities of the victims. By separating victim identity into these two categories, it becomes possible for the respondents to apply the 'vigilante' label to SDU violence that was directed at 'criminal elements'. In contrast, the label is not applied to SDU actions against impimpis.

With hindsight, the respondents also separate those SDUs motivated by political objectives from those whose actions led to personal gain. Here, the distinction is between 'criminal elements' within SDUs and 'real SDUs'. Retrospectively, actions perpetrated by SDUs for self-benefit qualify for the 'vigilante' label. Not only are actions against comtsotsis retrospectively defined as vigilantism, the comtsotsis themselves are also given the 'vigilante' label.

An additional area in which vigilantism is retrospectively ascribed to SDUs during the pre-1994 period, is the domestic sphere. Consider the following comments:

[During the 1980s and early 1990s, people's courts in Alexandra, would go into domestic violence issues and this] was actually okay in the beginning because it had good intentions but, as it was growing, the community experienced problems with it. Sometimes you find that the husband and wife are fighting and the presiding officer at the people's court is a young, inexperienced man, who cannot deal with complex issues. That's where it started losing direction. (AWH)
What normally happened was for instance if an elderly man did not sleep at home or did not support his family, he would be called before the wife who complained to these guys [of the kangaroo court] … and the old man [who was old enough to be their grandfather] would be hauled before them and they would start questioning him and he would be lashed. There was no format on how this was structured. Everybody, particularly the accused, had no rights whatsoever because he was not given the opportunity to state his own case. (ME)

For these respondents, the vigilante line is drawn between politics and domestic or personal issues. Central to such a division is the age of the participants. The youthfulness of those in charge of the kangaroo courts is a key reason why they are granted the 'vigilante' label in these examples. This ties into a range of issues, including perceptions about respect and tradition, as well as the age of the respondents. The link between vigilantism and the domestic/sexuality sphere is discussed in more detail below. Here, it suffices to point out the retrospective labelling of SDU vigilantes for their intrusion into matters that are interpreted as non-political.

It appears that with increasing political distance from the violence of the 1980s, interpretations and definitions of what constituted vigilantism then have also altered. However, the involvement of SDUs and non-conservative forces in vigilantism during the pre-1994 period remains a contested terrain. The definition slips along dimensions that include political-criminal activities, political-criminal motives, and political-domestic divisions. In terms of these dimensions, the 'vigilante' label emerges when actions cannot be justified politically, when they no longer appear to be legitimate, particularly with hindsight. This is not always the case, however, and there are those commentators who maintain politics at the heart of their definition. For example, a respondent from Cape Town comments:

In the past, the people's courts were basically politically based. It didn't really have that much to do with criminal matters as such. My experience was that it was basically a political set up to get people out of a community … . It was a form of, I don't want to use the word, 'intimidating' people. The people in the townships thought like that too … . Those community courts, or people's courts as they were known then, kangaroo courts … I think they were specifically there to counter the formal government courts … . Probably in some matters they dealt with criminal matters but I think mostly they dealt with matters [political]. (C1)

Given South Africa's history, it is difficult to divorce politics from the 'vigilante' label. Although this seems to be a retrospective strategy for describing certain non-conservative actions as vigilante, it is also important to recognise that during the pre-1994 period roleplayers such as the police, homeland administrators and black local authorities were not only politically motivated to act beyond the confines of the law. For example, Bruce & Komane (1999) point out that

it is not necessarily the case that official violence was more severe in dealing with perceived political threats than it was in dealing with alleged criminals. Where police violence exceeded the bounds of the law, whether this was officially approved of or not, it amounted to a form of vigilantism. Extra-legal violence was also associated with 'kitskonstabels' and municipal police who worked with town councils in black townships. (p. 41)

Bruce & Komane (1999) also explain that actions conducted by the makgotla, or tribal courts, were a form of vigilantism during the pre-1994 period. They conceptualise the makgotla 'as a community-based response to the problem of crime, as well as representing the continuity (or re-establishment) of traditional structures of authority in black urban areas' (Bruce & Komane, 1999, p. 40). Haysom (1986) notes that these courts 'imposed savage punishments and generally directed their activities towards the rebellious or disrespectful' (p. 4). The makgotla thereby lent indirect support to state repression, but they cannot be considered a solely politically motivated form of vigilantism.

The retrospective analysis of vigilantism offered by commentators such as Bruce & Komane (1999), Minnaar (1999), Gear (draft report), and certain respondents in this research project, broadens the phenomenon beyond political motive to focus equally on actions. This is not surprising given the contemporary emphasis away from politics towards crime (and the specific actions that define crime) as a tool for explaining present violence in South Africa. However, as the history generated by the TRC and certain other research respondents suggest - by retaining a definition based on political motivation - vigilantism in the past remains contested and complex. This affords a valuable lesson not only for defining the past vigilante spectrum but also for understanding the term in the present. Without the distance of time but with the weight of history, vigilantism in contemporary South Africa takes on different meanings for different commentators. The following section describes a range of definitions and explanations for post-1994 vigilantism as offered by interviewees, focus group participants and informants key to this research.

Post-1994
What is vigilantism in the new South Africa?

For most respondents contemporary vigilantism is defined in relation to the law. This is captured by the following responses to the question: 'What does vigilantism mean to you?'

you take the law into your own hands. (MSA)
the law of the country does not allow it. (MSA)
It's when a group of people [operate] alone, separate from the justice system and the police. (MJ)
the actions of people embarking on practices which are against the law, people who take the law into their own hands, trying to deal with crime (MSAPS).
vigilantes are most organised people who don't want to work within the framework of the law (AWH).

As these definitions reveal, vigilantism is set up alongside and beyond the parameters of the law, as something that is not bound by the formal justice system but rather that works separately from it. For some, this means that vigilantism is directly opposed to the law and is, therefore, illegal. Vigilante actions are defined as criminal, as these extracts suggest:

we have seen people who are against crime busy committing crime to those who are already suspected of doing crime. [They are vigilantes]. (MSA)
I think the perpetrators [of crime] are … all who are actually jobless, criminals, together with people who are actually conducting vigilantism. (MSA)

For these respondents, vigilantism is equated with crime.3 For others, vigilantism borders on being innocuous. While it is defined as working beyond legal parameters, it is not necessarily perceived as criminal or wrong. This is because the criminal justice system itself is a contested area that is not always accepted as positive. Indeed, for many, vigilantism is defined as a response to perceived failings of the formal criminal justice system. A Mamelodi respondent explains that,

the people who take part, those vigilantes, much of them are the victims of crime, people who've been raped, people who are not satisfied with the justice system, whom their cases were not properly solved, whom their cases were not taken up. (MJ)

The role of the criminal justice system in generating and justifying certain vigilante actions is discussed in detail under the section on explanations below. Here, it is important to note that vigilantism is not always seen as negative, even although it is defined in relation to the law. Similarly, by defining vigilantism in its relationship to the criminal justice system, it is set up as a response to crime. Unlike vigilante activities during the pre-1994 era, vigilantism now is overtly linked to crime and 'crime-fighting', rather than politics. Through this shift in definition, vigilantism is commonly explained as crime-motivated, rather than politically driven. This opens the term to an ambiguity in interpretation that it did not previously enjoy. Consider the following reaction:

One of the reasons the community courts [run by vigilantes] are popular, [is because] the community feel that they are getting help from these courts. I know a number of cases that have been addressed by these community courts … there are a number of suspects who are apprehended by these community courts and they are summoned to the police. The only wrong in that process is that they usually sjambok them before, which is illegal according to the current constitution. (MJ)

As this extract illustrates, the community courts run by vigilantes are presented as largely positive because they do 'help' the community by apprehending criminal suspects. They are interpreted as negative only because of their violent method of sjambokking suspects. And this is interpreted as 'wrong' as it stands in opposition to the constitution, rather than in terms of any moral or human rights argument. Indeed, it is seen as 'illegal', rather than 'wrong'. An extended form of this argument suggests that 'criminals operate with impunity and that they have too many rights anyway'. This argument supports vigilante methods of corporal punishment as an effective alternative to dealing with criminality.

Criminals are said and believed to be having too many rights with the legal aid offering them 95% protection. This is also accompanied by the comfort they are said to enjoy in prison. Vigilantism is regarded as effective and swift to deal with crime. (HRCK)
And even criminals, they are proud of what they've done; they can tell you right, straight to your face that 'I am going to do anything and I won't get arrested'. (MJ)

Not all community responses to vigilantism are positive. For example, the following respondents comment that vigilantism is problematic and negative:

Mainly the problems are with the modus operandi. It does not take much for them to beat you up. [Also] it's like the first one who goes [to them] is [assumed to be] right. [Anyone who goes subsequently is] presumed guilty. (MM)
The community does not support the vigilantes … . They cannot support something that victimises them. (ACFP)

Although these respondents object to vigilantism, the phenomenon is not viewed unequivocally. Unlike their predecessors, vigilantes in contemporary South Africa are not entirely defined, or rejected, as illegitimate. Rather, in certain quarters, the term co-exists with a sense of legitimacy. For some, there is a degree of ambivalence, and support, towards vigilantes. This is frequently compounded by widespread negativity about the police and a sense that there is 'no other option' but to administer 'justice' of this nature.4 However, this suggestion must be read cautiously. Although illegality does not automatically convey illegitimacy, lack of vocal opposition to vigilantism may not necessarily reflect ambivalence or tacit support for the phenomenon. Silence may also indicate intimidation or powerlessness within the community to address vigilante violence (Pigou, in personal communication). The situation is not as clear-cut as it seemed during the 1980s. Yet, the complexity and range of responses to contemporary vigilantism are interesting because vigilante actions and methods have not fundamentally altered between the pre- and post-1994 periods.

Vigilante methods

Vigilante methods in South Africa are commonly associated with violence and corporal punishment. An overview of vigilantism in 1105 media articles spanning 1991-2000 reveals violence in all but a tiny minority of cases (4/110=3%) (see Appendix B for a tabular summary of the media articles). In the four 'non-violent' cases, violence was intended; it was only through 'last minute' police or community intervention that death and/or corporal punishment were avoided. For example, consider the following newspaper extracts:

Members of the crowd began shouting for the [man accused of stabbing a girl] to be flogged [but, the presiding 'judge' of the people's court] … 'Judge Ninja' ordered the man to pack up his belongings and leave the area. (Sunday Times, 15 September 1991, V46)
The timely arrival of a police patrol saved nine victims of 'people's courts' from grisly deaths over the weekend … [they] found five tyres piled near the accused – they believe they were to be executed by the gruesome 'necklace' method. (Natal Monitor, 26 November 1991, V47)
A young woman sentenced to death by a 'people's court' was saved when police came across a group who were deciding where to execute her. (The Star, 10 August 1992, V97)

Of the violent cases sampled in the media, physical punishment occurred along a spectrum of severity, with death representing the most extreme outcome. Death was also the most common result, occurring in just over half of the vigilante incidents reported (57/106=54%). For example:

A woman was ordered to throw the first stone at her alleged rapist before a crowd stoned him to death this week. (Sunday Times, 24 September 1995, V114)
A man was 'necklaced' and another burnt to death when about 500 people stoned and attacked them. (Citizen, 17 July 1992, V95)
[A youth accused of murder was sentenced to] death by 'necklacing'. Those present tied his hands behind his back with wire and it was decided that he be stabbed once by each man present and stoned by the women. He was then doused with petrol and set alight. (Sunday Times, 13 November 1994, V108)
[A youth] who allegedly robbed and terrorised residents … was beaten to death after he was apprehended by an angry mob at the weekend. (Sowetan, 22 February 1999, V112)

In many cases, victims died during the act of vigilantism, rather than after the vigilante experience (for example, in hospital as a consequence of physical punishment). This suggests that death is often the intention of vigilantism and is built into the form that the vigilante act adopts. The 'necklacing' method is a clear example of this. Minnaar (1999) explains that '[t]he necklace method of execution, which involves placing a petrol-filled tyre (the petrol assists the tyre to burn fiercely) around the neck of the victim and then setting this alight, is a particularly South African activity' and there is a 'particular sense of finality about this method' (p. 28).6

In the media incidents reviewed, 'necklacing' emerged as a specific form of vigilante action across the duration of the decade; it was not restricted to the political climate of the early 1990s, but rather continued into 1999. What has shifted across this time is the profile of necklace victims who were described in largely political terms at the beginning of the decade (as impimpis) but who, in the late 1990s were seen as criminals, including 'drug dealers', 'gangsters', and 'foreigners'. For example, a media report notes the necklacing of Mozambicans accused of committing various crimes in Tembisa:

Vigilante mob necklaces 2 Mozambicans
About 400 vigilantes burnt two men to death and seriously injured three others in Tembisa … [The mob] rounded up a group of Mozambicans whom they accused of rape, housebreaking, theft and generally terrorising residents [and then burnt them, killing two]. (The Star, 17 January 1999, V113a)

Similarly, Dixon & Johns (2001) comment on the burning of Rashaad Staggie, following a march by the People Against Gangsterism and Drugs (Pagad), a group that has come to be closely associated with post-apartheid vigilantism:

a leading member of the powerful Hard Livings gang, Staggie, was brutally lynched in London Road, Salt River on the night of 4 August 1996. The incident was reminiscent of the 'necklacings' of the struggle years and was widely presented in the media as symbolising the escalating violence that threatened to engulf South Africa's new democracy. (p. 17)

From the media reports sampled, it emerges that even when death is not achieved, the intention to kill or cause grievous physical injury, is often still there:

Evaton youths hacked two brothers suspected of killing a local businessman with pangas and garden spades before leaving them for dead. (Saturday Star, 16 December 1995, V111)
A man alleged to be a rapist was cornered and severely beaten by an angry knife-wielding crowd … . He was bleeding from his mouth and head, and was terrified as he cried and pleaded to be set free. Teenagers held knives and screwdrivers at his throat and grasped him by the neck of his throat. (Citizen, 4 August 1997, V97)

Violence with the intention to kill or inflict serious bodily harm is also central to a number of incidents reported by respondents, incidents that have not necessarily been reported in the media. Consider the following extracts:

[A young girl was raped and murdered and the community] went mad. They had no-one to pin the murder to. That was when they started beating [my son and another boy]. The ex-mayor [of township B] came, looked at me and said 'we've found the murderers' … . [The crowd was] beating those kids. They were nearly half dead. My son's head was like this big … . The other boy had broken ribs and was also a mess. [The police] got them out. They were taking them to the hospital. The people were angry because they hadn't killed them. They were very angry. (C2, emphasis added)
R: The poor victim [is] lashed a number of lashes, according to [the judges' liking], and then the victim is taken to a back room where water is poured over them and then they receive another set of lashes. When the victim emerges, he is unrecognisable … . People would just assume who you were.
I: How does the victim get punished? By being punched, lashed, or what is used?
R: No! They use sjamboks and pick handles. (MF)

Case Study 1 below reveals the violent tactics deployed by taxi drivers working in Guguletu in the Western Cape. During early 1999, they received a great deal of publicity for the violence of their methods. They also came under the media spotlight because they were 'hired' by residents to solve specific cases, for example to retrieve stolen goods, and so received payment for their actions. The publicity was both positive and negative, with some condemning the violence of their actions, and others supporting their 'fight against crime'. Pigou (in personal communication) points out that the publicity surrounding this form of vigilantism contributed to reducing the violence, representing an example of successful activist journalism.7 (For a detailed overview of taxi-related violence in the Western Cape, see the CSVR report by Dugard, 2001).

Case Study 1

'People can't rob our workers in the community'
As told by a victim of armed robbery in Guguletu, whose stolen property was retrieved following vigilante intervention by local taxi drivers.

[I was robbed at gunpoint on a Friday at 12pm. I reported the incident to the police and returned to work]. Round about 4pm, we just heard tyres screeching and cars pulling up into the back area [of work] … Those guys were the taxi drivers. The [leader], his shirt was full of blood and his eyes were like wild and I knew he had been drinking something … I saw this guy sitting on the back seat. They pulled him out. They had handcuffed his hands behind his back. His face was like this, like totally swollen, scratches, he had been beaten in his face, his eyes were swollen, scratches over his head, pieces of flesh out of his arms. His t-shirt, his pants torn, blood … within minutes, we had masses of people standing there, and my colleague and I were the only two white people there. There were, I think, more than a hundred people from the community and they were all shouting and saying, 'yes, we've caught the thief' and 'he will pay. People can't rob our workers in the community', 'we don't want people to be chased away from here'. I said: 'okay, you've got the guy, he says he's got the stuff, just cool it now' but I was scared. I was more scared when the taxi driver spoke to me than I was with the guy with the gun at my head and I had this heroic idea of pulling the guy out of the car and keeping him inside and telling everyone to go away but the way they were psyched up and ready for the punishment, I don't know. I thought maybe they would do something to me because I tried to stop them … . The drivers said they would cool it [but after they left me] they held him from the Friday to the Sunday. They had been bashing him up. Early on the Sunday morning, they just drove up to [his] grandfather's house and threw him out of the car while they were still driving and he was very, very badly bashed up. They had to take him to hospital, emergency … . [On Monday, the taxi drivers told me to collect my belongings. I asked them why they had continued to beat him]. All they said is: 'you don't understand, you're not from here, you are white, don't bother'. They were very proud of themselves, of the fact that they are doing their bit, that they are good citizens … . They told me that they don't want qualified people who really want to work in the community being driven away because of gangsterism and stuff like that and that they had made an example of me. [They didn't ask me for money for their services although I think that someone paid them in my name]. … When I told everyone that I had gone to the police station [and that] I didn't know why we had to have this taxi thing happening, they said: 'no, you don't go to the police, they're useless'. Every person I spoke to in the community told me: 'you're wasting your time, don't go, they won't do anything. The taxi drivers are the only people the gangsters will listen to' … I don't know what is happening here and most probably if I lived in the village, I would be happy … I wouldn't want to fear being raped walking down the street or want my house to be robbed every second week, or fear that my child is going to be run over by a drunk driver.

Clearly, violent vigilantism is more 'newsworthy' than less violent methods of informal justice and violent incidents may well be over-represented in the media. A range of other actions, including intimidation, fines and community service are also part of the vigilante ambit. However, as various respondents indicate, violence is a core component of vigilantism even when it is not reported to the media. Similarly, the threat of violence is never far from other vigilante methods, as the following extracts reveal:

Somebody will come from nowhere and start slapping you and saying 'but tell the truth'. So there is a lot of intimidation, harassment and complete abuse of your personal dignity and esteem so that you end up cowering in a corner. The guys have guns, AK-47s, sjamboks, knobkierries, pangas. You are defenceless in that type of situation. You are completely at their mercy … . It's normally a large group … about a hundred people [in a school hall] and you are put in the centre, they will be surrounding you. (ME)
[They burnt my house down], my friends were intimidated, [the vigilantes] went to my work, so … I had to change workplaces. [I moved to another township.] So, in a way, it has actually killed me, it has killed my family, it has destroyed us, destroyed our home, our name. (C2)

Violence also works hand in hand with financial 'punishments' that are frequently meted out by vigilantes. Violence can operate either as a threat to elicit payment or as a gratuitous addition to monetary fines, as the following media comment suggests:

The alleged leader of a gang of hijackers was given at least 180 lashes and fined R3 600 after being tried by a 'people's court' this week. (Sunday Times, 20 December 1991, V104)

The inter-relationship between violence and financial remuneration is complex. For some respondents, violence and money co-exist as a way to instil justice and, importantly, ensure compensation for the complainant. The immediacy of vigilante 'justice' coupled with a perception that vigilantes are able to return belongings or extract financial compensation (via violence) is an argument commonly made in support of their methods. Consider the following extract:

[Vigilantes in Mamelodi are successful because] you find the community saying the police can't do nothing, even if maybe my property was stolen and the person was found guilty by the court, they don't get a way of compensation, they don't get back what was stolen or whatever damage was sustained. They don't get paid back so now if they take their case to the kangaroo court, judgment is passed and then people are paid back and so on. (MSAPS)

This respondent bases community support for vigilantes on their apparent ability to restore belongings and compensate those who turn to them. Pigou (in personal communication) explains that this may be particularly pertinent to communities where material possessions are scarce, a factor which may account for certain regional variations in vigilante activities. The 'restoration of material belongings' as a justification for vigilantism is important because it highlights a fundamental difference between the formal criminal justice model and vigilantism. This difference is not reducible to practical failings of the former, but rather suggests an elementary distinction between the premises of each approach (rehabilitative justice versus compensatory, retributive justice). The fundamental difference behind each approach is additionally confused by the perception that vigilantes 'help' in the wake of police failure,

I won't blame [people for taking the law into their own hands]. You know those people of the kangaroo courts [should] be paid. Those people are helpful. The kangaroo courts help. If the police say these people should stand back where as they the police fail? These people are helpful. (MJ)

The different underlying tenets of each model, as well as the practical failures of the criminal justice system are further explored in the section on explanations for vigilantism below.

Other respondents adopt a more cynical approach to the financial aspect of vigilantism, suggesting that self-gain is a crucial motivation for those who turn to vigilantes. For example:

It seems this lady [who lodged a complaint against me because she wants to get my house] is related to one of the kangaroo [court] people [and that she would get my house and thereby benefit by them finding me guilty]. (MI)

Additionally, self-gain is commonly cited as a central motive for vigilantes themselves. There are two complementary levels through which vigilantes may profit materially from their actions. Firstly, they may benefit by charging for their 'services'. This is particularly pertinent to groups that are organised and coherent rather than spontaneous, because remuneration for 'services' requires planning and structure. Hence, groups such as Mapogo a Mathamaga and PEACA in Khayelitsha fall into this category.

[With PEACA] I understand that you pay by case because if they want to go and pick up somebody … you have to give them money for petrol. That's the way they act. (K1)
I think [vigilantes] have commercial motives. I mean Mapogo has a base in Pretoria, somebody is got to be paying for it … It's like some protection rackets you get in the taxi industry. I don't think they are politically motivated. Someone will tell you this boss supports IFP, that's rubbish, we've been through that. All they do is shoot for money. That's gonna happen to Mapogo, it certainly has happened to Pagad. (JDP)

The second level through which vigilantes may receive financial reward is via their victims, who are frequently forced into 'paying up', not just to the complainant (if at all) but to the vigilantes too. This is a common complaint from a range of respondents:

[One of the kangaroo court members] told me that they were making money. They only make money for themselves. And it's true and clear that they were only interested in making money, for instance, they fined me R450 in a short space of time [without even listening to my side of the story]. (M1)
[Over the last 20 years] vigilantism has changed in a sense of saying that it's in a manner of people are making profit out of vigilantism. (MSA)
[Vigilantes] are actually thugs, are people who are pushing crime. Whereas they 'vie' to town or around town, looking for money and doing crime, committing crime. And, if things are bad their side, and then they come back to that [vigilante] structure so to carry out their duties of their particular vigilantism so that they can go and get money. (MSA)
In one section, you'll find there's a group of about 10 or 20 people involved in [kangaroo courts] and they're benefiting because they are intimidating the people they are living amongst. They will go into a shebeen, demand free liquor, you know, maybe offer protection or demand protection fees … . It's just a handful of people who are doing this and they don't enjoy the support of the community. (ME)8
The efficacy of vigilante violence

The motives of vigilantism, whether financial or not, are usually mediated and achieved through violence. These motives must thus be understood in relation to the functions performed by violence during the vigilante process. Consider Case Study 2.

Case Study 2

A woman needs sex
As told by a man who was beaten by a kangaroo court in Mamelodi

[I brought my girlfriend, who is also the mother of my child, from Pietersburg to Mamelodi. She started seeing other guys and so I ended the relationship. After a few months, she came back to me and asked for money so that she could go home to Pietersburg. I agreed to give her money at the end of the month]. She opted to go to the comrades. She told them that I was not supporting her and that I was not buying food for her. The comrades then wrote me a letter asking me to report to their offices. [I went there and] … . They said that, 'your charge is that you took your girlfriend from Pietersburg and brought her here and now you don't want to support her. We find you guilty for this … . Do you know that a woman is supposed to wash, to eat and she needs sex. Do you know that?' I said, 'yes, I know'. They were now insisting that I hear them properly. The same comrade told me that, 'Before you leave here. You should know that you're being fined … you should then choose between 1000 lashes and R170'. I told them that I'm afraid of being beaten; it would be better for me to choose to pay. They then told me that I have to pay another R50 for petrol in order to fetch my girlfriend [who was already there] and 'now that you have fooled with us, you have to receive 10 lashes before you leave' … they told me to lie down on the bench … . They gave me the first lash and told me that they won't count it. They then asked one of them to bring a twenty litre bucket full of water to wet my pants, so that my trousers should stick to my flesh or skin. I was then given another 10 lashes … . They then instructed me to kiss my girlfriend and take her to my place. 'You will have to see the other lady whom you are presently seeing at a later stage [until this woman decides to go back home]. For now, we want to hear that she is supported, she gets food and sex from you. We don't want to hear anymore complaints'.

In the vigilante process captured by Case Study 2, violence serves at least three purposes. Firstly, it is offered as a choice 'between 1000 lashes and R170'. In this way, violence functions as a threat to elicit financial payment from the victim. Secondly, once the victim commits to paying the fine, violence is actualised through the 11 lashes that he receives. This serves to 'punish' the victim for 'fooling with us'. As a punishing function, violence not only inflicts pain, it plays a symbolic role, representing a certain power and strength about the vigilante process. Corporal punishment renders this symbolic function visible to both the victim and the vigilantes themselves (as well as any bystanders). This function is reinforced by the apparent gratuitousness of the violence, especially because it reneges on the 'either-or' nature of the earlier choice. Thirdly, violence in Case Study 2 serves a function of intimidation. It underlines the verbal warning that is issued to the victim as 'we don't want to hear anymore complaints'. Here, violence strives for prevention and control, by instilling fear within the audience. It warns the victim and onlookers that their future actions will be violently punished if they do not follow the vigilante 'rules'. This preventative function reinforces the symbolic purpose of the corporal punishment and it generalises beyond the victim, to exert control over the audience. It also generalises beyond the actual incident to future possibilities. The following extract notes the symbolic and preventative functions served by violence, as well as its actualisation in terms of witnesses to the vigilante spectacle:

You see their punishment is not a good one. Before being found guilty, you find that you've already been beaten up … . If you could witness how victims are punished, you wouldn't want to witness it again. You would feel the pain. Another thing, those who scream amongst the audience, also get punished. They would leave the victim and beat the screamer. (MF, emphasis added).

The visible and immediate nature of vigilante violence, as well as the extreme pain that is inflicted on victims, serves as a deterrent for both the victim and members of the audience. According to respondents, it is common for vigilantes to turn on any voices of dissent, subsuming them directly into the realm of violence by 'leaving the victim and beating the screamer'. Because this happens, or threatens to happen, vigilante violence seems less about addressing a specific crime or the administration of justice, than the assertion of complete, repressive power. Working alongside the sense of such omnipotence is fear. Vigilantism, by virtue of its public and violent nature, pivots on fear. Fear is frequently cited as a reason for the persistence of vigilantism in communities. For example, consider the following comments made by police representatives:

They administer some kind of jungle justice … if you use that particular type of justice, you are bound to reduce crime [because you] instil fear in the minds of the people, who to an extent, would be frightened to commit crime. But whether in the long term, you would succeed to contain crime in such a fashion is quite debatable. (ASAPS)
[Vigilantes] feel at the end of the day that they have done something and also that they want to set an example to the rest of the community … not to commit a certain crime or crimes. (JSAPS)

Fear is generated and sustained through the public nature of vigilante actions. Central to vigilante methodology is the role of witnesses and public participants, as well as the visibility of the punishment.9 For example, consider the following interview extract:

[At around 4pm] I heard this loud noise, [two men] were breaking the window [of my house] … I took my firearm … and tried to shoot one [but I missed him] and he ran away. [I ran after the other guy and] I managed to grab him. He was very tall and he was tough [and he was a makwerekwere, or foreigner. Some of the neighbours came out and we dragged him to my house]. Then we assaulted him. We assaulted him very, very, very badly so to say. [I felt so angry because this was the third attempted break-in at my house and the police had not done anything about it]. So we assaulted this guy until 8am … . There was this crowbar that [we] used. We beat him on the toes with that, we beat him on the head with that. And then … we removed all his clothes, then we painted him and we let him stand on top of the electric box so that people could see him … see that he is a criminal … . [Of the participants] some were men and some were women, [even although the women came at a later stage]. Mainly they were from the South African National Defence Force (SS).

This particular incident reveals the public nature of vigilantism. The neighbours are key participants in the process and the victim is 'put on display' for non-participants to see. By painting the victim, he is visibly separated from the community and his criminal status10 is publicly monumentalised by placing him on top of the electricity box. This creates a physical and symbolic distance between him and the rest of the community. Ironically, the victim's isolation rests on him remaining fully in the public eye. In this way, the vigilante model contrasts to the prison model, which rests on removing the criminal from the public gaze; the two models are founded on fundamentally different tenets. Painting the bodies of suspected criminals is a violent method that seems to be on the increase in South Africa, with various such incidents reported in 1999. Many of these appear linked to racist vigilantism, although not all such incidents contain a 'race-dimension'.

Broadly, vigilante methods in contemporary South Africa share the following characteristics:

These causes must be explored in order to understand the specific conditions that generate vigilantism and to develop strategies for solving the phenomenon.

Explaining vigilantism

Fighting crime

Vigilante methodology is closely linked to its underlying motive or cause. This may translate literally in instances of 'eye for an eye' type justice, where the vigilante method of punishment imitates or corresponds to the alleged crime. For example, Minnaar (1999) cites a case of castration-for-rape,

A young man, accused of raping an eight-year-old girl, was arrested but no charges – due to lack of evidence – were laid against him and he was released by the police. This apparently angered the community and a group of women … caught him, beat him up, tore down his pants and then castrated him with a broken bottle. (p. 8)

Here, the vigilante method of castration follows a 'punishment fits the crime' type approach. This is not just restricted to spontaneous incidents but may take on a formalised, almost institutionalised, format. Minnaar's (1999) documentation of a 'Code of Punishment' drafted by an Ivory Park People's Court exemplifies this:

Adultery: 500 lashes for the man and banishment for the woman
Murder: Necklacing or execution at gunpoint
Rape: Paraded naked before receiving 400 lashes or execution
Child Abuse: 380 lashes and banishment
Motor vehicle hijacking: Death for repeat offenders and lashes or execution for first-timers
Theft: 50 lashes
Burglery: 200 lashes for first offence and if items not returned to owners, extra 300 lashes
Assault: 90 lashes
Assault by a man of his wife: 50 lashes
Contempt of court: An additional 40 lashes and a two-year banishment from the area (p. 30)

In this extract, particular punishments have been agreed on and formalised as a code of punishment. This suggests a certain degree of consensus and planning within the people's court. Codification of punishment is similarly described by de Villiers (2000) in an overview of what he terms 'chief's courts' or 'customary law'. He outlines common punishment etiquette within this field and contrasts the predetermined code with 'recent judgments from approximately 100 courts of chiefs and headmen in Transkei'. Some of these judgments include:

Attempted assault: Guilty and fined R30
Abusive language: Six lashes
Rape of plaintiff's daughter by defendant's son-in-law: Ordered to pay R800 to plaintiff
Proposed love to school girl by force: 5 lashes (pp. 4-5)

These judgments are drawn from a loose typology of punishments for certain offences. Unlike the written codes, however, the applied judgments do entail corporal punishment in the form of lashes. It is beyond the scope of this project to delve into what have been termed 'traditional', 'indigenous', and 'customary' laws and authorities other than to comment that South Africa's constitution provides for their legal recognition 'subject to the Bill of Rights and other relevant legislation' (de Villiers, 2000, p. 1). In this regard, corporal punishment is not legal and therefore certain of the judgments listed above could earn the 'vigilante' label. (The relationship between 'tradition' and vigilante violence is further explored as an explanation for vigilantism below, as well as in von Schnitzler et al's (2001) report on Mapogo).

The 'punishment typologies' and 'punishment to fit the crime' methodologies described here presuppose a crime. Vigilante methods applied along these lines are thus motivated, or at least justified, as a response to crime. Popular contemporary discourse supports this idea, with certain media images glorifying public 'crime fighters' and bringing vigilantism closer to a 'Robin Hood' or 'Wild West' type ideal.11 The vigilante 'crime-fighting' motive cannot be explained without reference to the criminal justice system. Indeed, for many, its existence is explained by the criminal justice system, by a 'policing gap' left by failing authorities. The police are central to this explanation.

Filling the policing gap
We do have quite a number of people who are taking the law into their own hands and what forces them to do that is because there is too much of this reluctance of the police [to get involved]. (ACPF, emphasis added)
A suspect is arrested, then two or three days [later] the suspect is out and then if they report the case, then the police don't take action and that's where they take the law into their own hands. (AVSG)

The police don't help us at all. We report and say here is violence and criminal activities and they don't even bother to check. (K3)

The criminal justice system, the police, their service is not right for the community of Mamelodi (MJ).

These responses suggest that the police are reluctant to address crime, that they are inefficient and unhelpful. For these reasons, the respondents explain, people are forced to take the law into their own hands, to deal with crime. Rather than being viewed as a considered choice, vigilantism is presented as a necessary and inevitable reaction to police lethargy. This argument takes the line that police inaction creates vigilantism. As one respondent notes:

[There is a] perception in society that criminals are being allowed to go free. This is supported by statistics. (MISS)

And another observes that the presence of 'informal justice' groups, such as PEACA in the Western Cape, is evidence of a policing gap, wrought by police failure.

I think [the police] are failing because of the mere fact that there is PEACA, it's only because the police are failing. (K1)

Other respondents go one step further to suggest that the police themselves are an active part of the crime 'problem'. They are seen to commit crime and collude with criminals, as the following extracts reveal:

Those people we call policemen, they are friends of the criminals outside. Even where I'm staying, there are a lot of criminals, they are robbing some people and all that. But I cannot report them here because their friends are inside the police force. (K3)
Most of the time, the police are favouring the accused. In so much that most of the offenders like armed robbery, there are policemen who are involved in armed robbery. (K3)
We are the ones who don't work with the police because the police think that some of us in the township are very stupid, they are the clever ones because they go along with the criminals. (K3)
You're much more likely to be a criminal if you are a policeman than you are as an ordinary member of the public. (JP)
What the people should mostly complain about is the police themselves, the corruption within the police, the lack of transparency within the police division … and there are a number of guys who are corrupt within the justice system itself, apart from the police. (ME)

For these respondents, the police themselves are criminals. They are held to participate in a range of criminal activities, from armed robbery to corruption. They are also perceived to support and befriend other (non-police) criminals, thereby colluding with an entire criminal system. Through their criminal actions, the police are set up in opposition to the law. They are portrayed as not merely obstructive to justice but actively defiant of it. To the respondents, this constitutes an abuse of power and presents vigilantism as the only alternative to fighting crime, both within and beyond the police service.12 From this position, the vigilante alternative is accompanied by a sense of morality or 'righteousness'.

Not all respondents elevate vigilantism to a level where it becomes the only moral solution however, even within an argument that conflates police actions with criminality. Vigilantism remains a contested term. Consider these comments:

Basically the police support these [vigilante] guys. They are not arrested, they are not taken to court, nothing is done. People go there and report the matter but no steps are taken. (MM)
[The vigilantes] and the police are in cahoots. What they do there, the police are also involved. (K3)
One of the PEACA gang is known to the police and he is always at the police station and this is not a secret and these people who assault victims are known at the police station. (K3)

For these respondents, police actions are represented as illegitimate. However, this illegitimacy is defined by police support of, and participation with, vigilantes. In this argument, vigilantism is presented as yet another aspect of crime. It is seen as illegitimate and any perceived police support of vigilante groups is similarly tarred. This paints the police with a 'vigilante' label themselves. In effect, the tension between vigilante legitimacy and illegitimacy boils down to how the term is defined. From one perspective, vigilantism is seen as an alternative to police criminality; from another perspective, it is seen as a form of crime that contributes to criminality in the police. What is clear, however, is that in both arguments the police are represented as illegitimate. This is reminiscent of the manifestation of political vigilantism during the 1980s.

The conflation of the police with crime is reinforced by media images and specific incidents, such as the Special Assignment report screened on the SABC 3 television channel in November 2000 of six white policemen setting their dogs on three Mozambicans, and the 1999 incident, filmed by the BBC, of South African flying squad members beating suspected criminals. Actions and images such as these earn the police an illegitimate and, indeed, 'vigilante' label. This impacts on levels of community trust, as the following respondents note:

[Last year a journalist was beaten up by the police and he lost an eye. He] was badly beaten up at the police station in Guguletu. That obviously made world headlines and, I don't know, maybe the community has serious problems with what was happening at the police station. (C2)
We know that a lot of people don't trust the police … and they see in the newspapers daily policemen being charged with corruption and other crimes … policemen being accused of racism. … it's not helping other efforts being made to build this trust, reconciliation with the community. (JSAPS)

The argument that is made here suggests that community mistrust fuels perceptions about police illegitimacy and opens a space for community members to take the law into their own hands. For many, community mistrust and general relations with the police cannot be fully understood without recourse to the past and South Africa's apartheid roots.

Apartheid history
We come from a culture where the police are seen as instruments of the state against which we were fighting. (MM)
One can see from the relationship that [the police] have with the community that they haven't changed from the past police, meaning that they are not friendly enough, the communication level is not there. (MJ)
As soon as you bring the police into the picture, the people start to close up … . Historically, they don't want to work with the police. They see the police as the enemy, and that's going to take a long time to effect change. (C1)

For these respondents, South Africa's policing history is offered as the reason for poor community relations with the police. This history, they explain, creates a space for vigilantism today because people 'don't want to work with the police', and would rather take the law into their own hands. More generally, apartheid is represented as a key factor in explaining the 'crime-fighting' activities of contemporary vigilantes, as the following extracts suggest:

I: Why do you think vigilantism as a way of dealing with crime in the first place did come up?
R: Because of apartheid. Because when [black people] came to report a case … [the police] don't take action, even the media did not report such cases about black people … . So that's why the people didn't trust the police by that time. They used to take the law into their own hands. And vigilantism, it was happening in black communities, not in white suburbs. We're dealing with the people ourselves here in Alexandra. Because we're having the problems and we know the people who're doing it. (AVSG, emphasis added)
The apartheid regime has caused vigilantes [to develop]. (ACPF, emphasis added)

The role of past police-community relations and apartheid's impact on contemporary vigilantism is acknowledged by members of the police service themselves. For example, consider these comments:

The history of our country, of actually looking for a victim, plays a role [in ongoing vigilante action]. There are people who simply don't trust the police. That is not something that we are going to change overnight. When we started with community policing, we looked at it as a 10-15 year project and that was in 1994. I think we are getting there. (JSAPS)
You must also understand that the police in the past had extraordinary powers to deal with offences against the state … and that they also had extraordinary powers, almost [draconian] powers to deal with crime itself … and all of a sudden you have this transparency and police have been pulled up, they don't have these powers anymore. (C1)

Both of these extracts locate contemporary vigilantism in relation to apartheid. For these respondents, this link is not only about community perceptions and trust of the police. It is also about policing strategies and changes during a political transition. As C1 suggests, the police themselves have had to change their approach to crime and can no longer engage in vigilante type actions with 'extraordinary powers'. Vigilantism, in this argument, is located in the mechanisms of transition. It arises to 'fill a policing gap' but rather than a gap created by contemporary police ineptitude or criminality, this gap is seen as one created in the shift away from repressive state control and heavy-handed vigilante-style policing methods, to police 'transparency' sans draconian powers. In this argument, vigilantes are conceptualised as directly filling the 'crime-fighting' void left by the transition. Here, vigilantism represents a privatisation of the old style policing function.13 It is rooted in apartheid but cannot be explained in terms of a static past. Rather, contemporary vigilantism is explained as a product of South Africa's political transition, tied as much to the current political dispensation as it is to the apartheid past.

South Africa's political transition

Various respondents explain vigilantism in terms of South Africa's transition from apartheid to democracy. This is done in a number of ways. For example, the police respondents above locate vigilantism in transitional policing practices and diminished police 'powers'. Another respondent situates the phenomenon in relation to changing perceptions of state legitimacy.

I: Is [vigilantism] a new phenomenon in Mamelodi?
R: Yes, it is a new phenomenon, though the history of Mamelodi tells us something, for instance, the period before the elections, the democratic elections, Mamelodi at some stage was a no-go area for the police. People were dealing with crime themselves, so, as a result, people have it in their mind that if they themselves can deal with crime privately, they can bring the crime level to zero level.
I: So has there been any change regarding the nature of vigilantism since around the elections?
R: What actually, I've mentioned before is that at that point in time, one cannot call that vigilantism, before the elections. It was just a community dealing with crime. But at this stage, what has happened with this kangaroo court, that is the phenomenon today. (MSAPS, emphasis added)

For this respondent, vigilantism is defined as something new, as something that has originated since the democratic elections. His argument suggests that through the transition, certain actions have come to be termed 'vigilante'. He considers the same actions – of fighting crime – to be legitimate in the past because of the 'no-go' relationship between the police and the community. In the present, however, these actions attract a 'vigilante' label. While the argument appears to rest on semantics, it must be contextualised alongside perceptions of state legitimacy. What are termed 'vigilante actions' today are 'new' because the current political dispensation, unlike its predecessor, is seen to be legitimate. Here, the perceived legitimacy of the political order co-exists in an inverse relationship with vigilante actions: vigilantism occurs in a legitimate political climate, while 'crime-fighting' originates under an illegitimate order.

The transitional period is also put forward to explain vigilantism in terms of changing perceptions of crime. Consider the following comments:

[Vigilantism] is a situation of those years of the community condoning criminal activity by certain people committed outside that community. That's unfortunately I think one of the hangovers of the previous system, that you have a generation that maybe feels it's okay to do crime. (C1)
Historically [those who live in section 4 of Mamelodi] were involved in the struggle as young people. Now they are in their thirties and forties and they are buying houses and so on. So they have that it has been a very politically active section. They used to have the street committees, things like that. They would patrol during the night, you see. Historically, [it] has always been like that in that section. Then ultimately, it culminated into full-blown vigilantism [when they started protecting their new houses and belongings]. (MM)

For these respondents, South Africa's transition to democracy has changed perceptions of crime, as well as crime patterns. Crime is represented as fallout of apartheid, as something that, for many, remains 'okay' in the current order. However, these respondents argue, when crime turns in on itself and starts to occur within certain communities, rather than beyond them, the foundations for vigilantism (as a response to crime) are established. MM suggests that vigilante conditions are particularly ripe in those communities that were highly politicised and active during the apartheid years; for him, vigilante actions are interpreted as an extension of street committee activities. He also links vigilante conditions to a change in the demographic profile of previously politicised communities. The 'youth' of the struggle are represented as middle-aged today. For this respondent, their priorities have shifted with time and political change, to include property acquisition and protection. And these factors of age and property ownership are seen to fuel vigilante reactions to crime.

Others suggest that vigilantism is more a product of expectations about the political transition than actual property acquisition and protection.14 For example, consider the following comments:

What [causes] community mob attacks is truly a factor of expectation [linked to] the 1994 election. People had hope and expected acceleration of change, that's what happened and it's not taking place. People hoped for better jobs, better houses, free education, free medical care. It's not happening. Now people want to go back to the culture of controlling themselves, people want to go back to the culture of taking leadership of their own life. (TCPF)
I think people had a lot of expectations obviously when the whole political structure changed in this country … . People most probably vote under the illusion that things are going to happen overnight and when things did not happen overnight – this is my opinion - people started becoming frustrated and disillusioned and maybe at that point, started to take the law into their own hands (C1)

In these extracts, vigilantism is conceptualised as a consequence of expectations about democracy, specifically, disappointed expectations about the political change. Vigilantism is portrayed as a manifestation of emotion, motivated by 'frustration and disillusionment'. It is also seen as a form of empowerment, as a way to 'take control'. Vigilantism in this argument is imbued with nostalgia. It is offered as a way for people to 'go back to the culture of controlling themselves'. Such nostalgia serves two functions. Firstly, it creates a particular image of the past, an image that glorifies a lifestyle of 'self-control', unfettered by state intervention and policy. This lifestyle also assumes community coherence and blanket support for 'self control' actions, without recognising that past actions were not uniformly embraced. Secondly, by yearning for such a lifestyle, the nostalgic explanation of vigilantism comments on the present political order and generates a distance between 'the (contemporary) state' and 'the people'. Thus, from this perspective, vigilantism is not only about nostalgia but also represents an expression of contemporary dissatisfaction, as well as disappointment, with the political dispensation.

The sense of disappointment about contemporary governance is commonly located in the criminal justice system. For various respondents, this is based on perceived and real failings of the system, coupled with a sense that 'we have been let down'. It is in this light that they understand vigilante actions. For example,

[But] I don't think the problem lies with the bail as such but it lies with the greater criminal justice system, of how a person is brought before court and how the trial is conducted and brought to conclusion. It goes back to prosecutors … if they were well trained, [defence lawyers] wouldn't get away with [some of the things they do]. (MM)
People are first of all impatient with the duration of the law [with the] time it takes for a person to be convicted … . Secondly, you find cases where people walk free from the courts because of lack of evidence or lack of proper investigation. [Also] people don't trust the justice system and they want to deal with crime in their own way. (JSAPS)
R1: The jails are full … some people don't report these cases even although they have seen the crime being committed because as a person, you don't have protection. Because the criminal would come and terrorise you.
R2: If [the jails] are full, bring the criminal to us and we will sort them out. (K3)
The criminal justice system is corrupt. The police are corrupt. If I go and report a crime to the police or I merely inform them about an act of crime, the perpetrator will be knowing very shortly that I went there to the police to report the crime and as a result, my life becomes in danger. (AHI)

For these respondents, vigilantism is a product of criminals 'getting away with it' in the new order due to negligence, overcrowded jails, badly trained prosecutors, corruption and poor investigations. They see vigilantism in light of practical failings of the criminal justice system. A respondent sums up this perspective thus:

There's a lot of talk, of course, with South Africa's new constitution, [about] access to justice … [but its] meaningless when you don't have effective remedial structures and that is essentially the problem: that wonderful constitution, wonderful bills in many respects, but no capacity to implement, protect or uphold rights. (JP)

This perspective culminates in a general sense that 'justice is not being served'. Case Study 3 indicates a form of vigilante action that was generated out of such a perception.

Case Study 3

We paid his bail so we could kill him:
Murder accused beg to stay in jail after necklace killing V. Khupiso & J. Hennop (Sunday Times, 04 July 1999)

Residents of an East Rand squatter camp each paid R22 this week to bail a man accused of murder out of prison. Then they killed him. Johannes Manamela left jail thinking relatives had paid his R4 000 bail. Instead, he walked into the hands of the residents of the Winnie Mandela squatter camp in Tembisa – who dragged him before a people's court and sentenced him to die. He was stabbed and doused with petrol before a tyre was placed around his neck and set on fire. His horrifying death is part of a frightening new trend in vigilante justice spreading fear in jails as accused men beg to stay behind bars, safe from revenge attacks … Johannes Motaung, a recent victim of crime in Tembisa, was one of those who paid Manamela's bail. 'I donated R22 when I heard what was going to happen to him. The idea of paying bail for criminals and then killing them is liked by everyone. We want them to know that we will get them in the end', he said. 'We are always told that there is a shortage of police. We will protect ourselves' Phumzile Madondo, another resident, said: 'The police must come and arrest us all because we all donated money towards his [Manamela's] bail so that we could administer our own justice. We don't trust the police and the justice system.'

Case Study 3 represents an incident of revenge violence motivated by perceived failings of the criminal justice system. It is initiated as a response to 'police shortages' and by a sense that the criminal justice system cannot be trusted to achieve justice. The incident illustrates an organised, planned set of activities that culminate in the violent death of the accused criminal. Vigilantism here is systematic, not spontaneous, and it entails a complex process of premeditated murder, from gathering bail money to necklacing the accused.

This incident is not only motivated by an apparent failure of the criminal justice system on a practical level. While participants in the process cite policing shortages and general mistrust as key reasons for their actions, they are also motivated to 'administer our own justice'. This suggests a different form of justice – 'ours' – divergent from that administered by the formal criminal justice system. In this argument, 'our justice' is equated with violent death. As a participant comments, 'the idea of paying bail for criminals and then killing them is liked by everyone. We want them to know that we will get them in the end.' This is a popular attitude, one that is similarly reflected in widespread support for the return of the death penalty (see below). With violence and very often death as key vigilante objectives, this form of justice jars against the human rights approach of South Africa's post-apartheid constitution. At heart, the two approaches are not compatible. In this context, vigilantism is not solely reducible to disappointed expectations about South Africa's transition and practical failings of the criminal justice system. Rather, vigilantism appears as a phenomenon generated through fundamental misunderstandings and disagreements about democratic principles and processes.

Due process: misunderstandings and disagreements
Right now, I should say the community themselves, they feel that it is unfair that suspects … should be released with minimum sentences, [unfair that] they should be granted bail. So as a result, these community members came to support these community courts. So, I should think it's because of the constitution which grants us human rights … these actions [in the past, under apartheid] were not so much intensified but now that the present government recognises people have certain rights, it is now people are becoming dissatisfied [that] these criminals walk the streets. (MJ, emphasis added)
Most people are not aware as to how the criminal justice system functions. If you can be arrested today and granted bail tomorrow … people think that the policeman is responsible and certainly they will assume that [the perpetrator] will never stand trial again. And in some way they should try to avenge what he did themselves. (ASAPS, emphasis added)
People think that when a person has got bail, he is off. (MM)
You find a lot of problems with vigilantism. There are a lot of them with good intentions. But they arrest somebody on the basis of rumour. They conduct a citizen's arrest. Then they come to the police and they hand the suspect over. Then there's no evidence to hold that person and then the police are obliged to release that person. They release him back into the same community. That raises a problem. The problem we had in the beginning is now worsened because now the community does not only blame the suspect but they blame the police as well. (JSAPS)

For these respondents, vigilantism is explained in relation to misunderstandings about due process. In particular, bail is represented as a key factor behind vigilante actions. It is interpreted as a space in which criminals are able to return to their communities and flaunt their 'freedom', as well as to intimidate their victims and re-continue with criminal actions. As a respondent notes,

The first thing that brings violence is that the criminal offender commits crime and comes back and brags to the victim. I feel that the victim must be consulted before the perpetrator gets bail. Then we feel the police are not doing their duties. So we feel that the law of the jungle is best. Because if I feel like killing somebody with the gun, I can kill him because I know that tomorrow I will be back here. That is why there is so much crime. (K3)

Bail is seen to serve the interests of criminals rather than their victims and the community at large. It is not understood as a mechanism to protect human rights and uphold the 'innocent 'till proven guilty' doctrine. Instead, it is perceived as a stumbling block to justice, although, as Case Study 3 ironically indicates, bail can create a window of opportunity in which vigilante actions themselves occur. Alongside perceptions that it creates a space for continued crime, bail is also rejected as an impediment to instant justice. Consider these extracts:

Now the community, especially [in] the black townships, don't understand [the courts]. They still believe … 'if we arrest someone, we bring him here to you, you [the police] must keep him, we don't want to see him again. He must serve six, seven years'. For instance, [the taxi people last year] blamed the police. [The community] go to the taxi people and say 'can you help me to solve my problem? [because the police won't].' So the [taxi] people were willing and then this [is a] problem, of the people going out there and going totally out of hand. (GSAPS)
The reason for people to be assaulted is that when I do criminal offences and get arrested and the same night I am back here with you. If we take into consideration like in the olden days, we said 'let us make our own god here because Moses took long in the mountain'. That is why people go to the taxis and say 'make us a god which we can worship now because Moses has forgotten us' … . So people want to see things happening very fast. (K3, emphasis added)

These respondents explain that for vigilantes, the process of formal justice is problematic because it delays punishment. 'People want to see things happening very fast'. This suggests that vigilantism is driven by a lack of understanding and limited appreciation of due process. Clearly, South Africa's past plays a significant role in explaining this response. During apartheid, due process was not followed, 'justice' was used to violate human rights, and corporal punishment was an integral part of a racist system. Consequently, most South Africans have not benefited from bail, fair hearings or balanced sentences and these remain abstract, alien concepts. Consider the following comments about the death penalty:

[I think we should have the death penalty today and the judges who are still mostly white, must determine who should go to the hangman]. That's how they've been doing it before. Because they were happy before because they were not sentencing their own brothers. Now they know that their own [white] brothers can also be hanged, [that's why there is no longer a death penalty] … . Because before they used to do it to blacks only. There was no white man being hanged. (AVSG)
When a person has hijacked a car and shot the driver, that person, in my view, must be given the death penalty. There is no need for the government to imprison that kind of person, it's a waste of our money. He has killed, therefore, he must also be killed. I think the reason why we were against the death penalty during the days of apartheid, was mostly because black people were being hanged and for wrong reasons. But now we have a democratic government in place, I think the death penalty must be brought back. (AB)
The death penalty must not be reinstated until such time that our criminal justice system is transformed. Because there are still those old white judges who look at skin colour when they sentence people. (AB)

These respondents explain that during apartheid the death penalty was applied as a racist and unfair instrument of 'justice' to black people. On these grounds, it is contested for its racist application at the time. It is not, however, contested on the basis of a human rights argument. That the respondents welcome the notion of a death penalty in democratic South Africa, suggests that they have no human rights-based objections. Rather, they are appealing for a racially fair application of capital punishment. This is a common contemporary response to the death sentence. For example, in an assessment of human rights knowledge among the general South African population, Pigou, Greenstein & Valji (1998) found that 73% of their respondents (drawn from a sample of 3 600 participants) 'maintained it was a mistake to take away the death penalty in South Africa, and only 14% disagreed with it' (p. vii). They note that,

The abolition of the death penalty by the Constitutional Court in 1994 has prompted repeated calls for its return from a wide cross-section of South African society, across the political spectrum and racial divide. This attitude reflects a widespread public perception that the government has gone 'soft' on crime and criminals. This reflects the false notion that the death penalty acted as a deterrent in the past, and that its removal has led directly to an increase in the crime rate. (Pigou, Greenstein & Valji, 1998, pp. 95-96)

Popular support for the death penalty reveals a fundamental difference between the constitutional premises governing post-apartheid South Africa and popular public discourses about justice. In this difference rests an explanation for vigilante violence. Vigilantism is explained, not as a misunderstanding of the criminal justice system but as a divergence from, and disagreement with, a human rights framework.

'Our tradition'

Various respondents contest South Africa's human rights framework on the basis of 'culture' and 'custom'. They present corporal punishment as an 'indigenous' and fundamentally 'African' practice of justice. Here, 'tradition' is set up in opposition to a 'western' system, resulting in an 'old' way of life gaining privilege and authentication over the 'new' way. Because it is based on a 'way of life', this argument is not about explaining vigilantism as a particular phenomenon. Rather, violent methods that constitute a 'vigilante' label from South Africa's constitutional perspective are merely absorbed into a broader framework of 'our way'. For example, certain vigilante groups such as Mapogo a Mathamaga15 justify their violent methods as 'the African way':

'This is the African way of stopping crime. The criminal must lie on the ground, and we must work on his buttocks and put him right' – Monhle Magolego, president of the vigilante group Mapogo a Mathamaga. (Mail & Guardian, 23 December 1999).

Statements of this nature appeal to a sense of tradition. This sense of tradition is left vague and generalised as 'the African way'. However, it relies on the weight of heritage and history to contest the 'western' way, which, in this context, includes human rights, constitutionality and non-corporal punishment. A similar justification for violent vigilante actions emerges in the following focus group interchange:

R1: We as black people have never built jails. Jails were for white people … once a person commits a crime in our presence, our jail is a grave yard … if you make trouble in Xhosa [tradition] we can take the law into our own hand because we grew up without jails in our communities … . We sit at the round table and if the man is guilty, we hit him with stones until he dies … our jail is a stick and a stone, we hit you until you die.
R2: Violence is not the same, though in Xhosa, a sin is a sin, there is no small sin. [Now] you find that a person does a horrible thing like raping a two-year old child. The law takes it very easy and the person gets out of jail easily. If you break a matchstick, then you are fined more years. That thing surprises us. Where does the law work and how does it work? If the government could leave these criminals to us, I think it will come to an end because we will carry out our tradition and there will be no more rapes here, never.
R3: During the time when people were hanged, there were no rape cases … and we black people never used to discuss anything with a rapist, we used to pull him with a rope and beat him up or hang him straight away. (K3, emphasis added)

The respondents in this interchange justify violent and instant actions against criminals in terms of 'black' and 'Xhosa' culture. They set 'tradition' against 'the government' and 'the law'. In this way, they appeal for a return to 'tradition' and, by implication, violence, as a way to eradicate rape and crime in general: 'if the government could leave these criminals to us, I think it will come to an end because we will carry out our tradition'.

Various other respondents similarly draw on 'tradition' as an explanation for contemporary vigilantism. Even if they do not advocate instant justice, they understand the phenomenon as rooted in a 'way of life'. Consider the following extracts:

We look at the way in which the indigenous law system was functioning. It's quite different from the western type of legal system, the way it functions. For instance, in indigenous law … when they committed a crime, he was supposed to be tried as soon as possible … . So the matter is dealt with now and it is resolved now. So, to some extent, that has created a situation where people who were exposed to that type of legal system began to think that whenever one commits a crime, actions should be taken now. (ASAPS, emphasis added)
In customary law, you are tried by your own peers. You appear before the very people you live with. They judge you on the basis of the standard which they feel is acceptable to them … . I think [vigilantes] maybe try to revert to that sort of a thing. If you trace it well, the whole concept of community forums and what you have is very old in the township. It has always been there. In the 80s, it went into maComrades, then I know them as maKgotla … this is an ongoing pattern which has always existed. (MM, emphasis added)
The African person has his own way, historically and traditionally, of dealing with certain matters and these type of community courts [that have been set up in the townships] give him a foothold into solving some problems that they maybe can solve. (C1)
If you look at vigilantes' modus operandi, they are going to get support from most of the people who are ignorant of how the criminal justice system functions. People who normally have, were exposed to, the functioning of the indigenous legal system would want to see the action being taken now. (ASAPS)

It goes beyond the scope of this research project to provide a detailed analysis of these statements and their social and practical implications. However, it is important to recognise that a 'traditional' argument can also be made to contest contemporary vigilantism. For example:

[Mapogo a Mathamaga] traces their method to African customary tradition where the offender is healed by being given corporal punishments and made to drink some medicine. The rural courts however, did not operate with such intimidation and brutality. (HRCK)
People have forgotten that they are people. They think they are animals and because we have got this strong concept of ubuntu, now has been taken away from us. That is where restorative justice comes from … we've got a duty to perform, to teach even our courts … so that they adapt to what we have got to say … they mustn't come up with a mix of Roman and Dutch law, and English law and impose it on us because we, in our township, 99% are Africans. You go to court and you are a European because everything is done the European way. That is also one of the factors [causing vigilantism], I would say. (G5)

As these extracts suggest, there is no one perspective on the role of 'culture'. While advocates of violent corporal punishment may utilise this discourse to serve their own ends (as, for example, Mapogo a Mathamaga does16), others may also draw on 'tradition' to advance a human rights approach.

Emotions, prejudice and revenge

All of the explanations offered for contemporary vigilantism thus far are underscored by the basic assumption that vigilantism is motivated by 'fighting crime'. However, the following section suggests that vigilante action does not only revolve around issues of 'crime'. Or rather, the ways in which 'crime' is defined by various vigilante groups are not restricted to activities such as theft and murder. Like vigilantism, crime itself represents a contested terrain. Often, what is considered to be 'crime' expands beyond the common legal categories to include personal disputes and domestic relationships, as well as sexuality. As the incident in Case Study 2 above illustrates, vigilante violence circulates around a complaint about 'food and sex',17 rather than any tangible 'crime' per se. Consider also the following interview extracts:

The involvement of the kangaroo court in domestic issues antagonises the communities. According to African culture, unless it's in the court of law, I can't ask an elderly person about his love life, you know like these [young] guys are operating and that's why the community wants nothing to do with it. (ME)
Vigilantes have already taken [on] an extra-ordinary process of which they now intervene to solve the cases of the husband and wife and of which they don't understand very clearly about the law. (MSA)
[PEACA] don't build the community but they fight against us and with the very same people who go to them. For example, you have a dispute with your wife and they fight you and assault you [but] they should refer this matter [elsewhere] because they are not marriage counsellors. They fight with people, not with crime. (K3, emphasis added)

For these respondents, vigilante intervention in domestic issues is negative and unrelated to any crime-fighting activities. It is seen to 'antagonise' and destroy, rather than build, the community. 'They fight with people, not with crime'. It is this characteristic that earns groups such as PEACA a 'vigilante' label for the respondents. Vigilantism, for them, is about executing control over communities and meddling in personal affairs, rather than solving crime. Others agree:

Sometimes you find that people may have a grudge against a certain individual and they may persuade a lot of other people to help them in dealing with this person in a sense of accusing this guy of committing some other crime. They form vigilante groups to get rid of this person in the name of justice. (JSAPS)
The person that actually came to burn my house doesn't have a house. It's jealousy. (C2)

For these respondents, vigilantism is less about crime-fighting than settling personal scores. As JSAPS suggests, in certain incidents crime may be manufactured and used as a front to disguise other motives, such as jealousy or a personal grudge against the victim. In other vigilante situations, the 'crime-fighting' motive, whether real or manufactured, does not feature. Incidents are framed in terms of emotions, such as revenge or hatred. For example, a respondent explains that her son openly seeks revenge, following a vigilante attack. Although he has not acted on his 'hatred', his rhetoric is about emotion-driven action, rather than crime.

[My son who was badly beaten by a mob] says he hates everything, he hates the police, he hates interviews. We went to a