Vigilantes: A contemporary form of repression

Vigilantes: A contemporary form of repression

Haysom, H. (1989). Vigilantes: A contemporary form of repression. Paper presented at the Centre for the Study of Violence and Reconciliation, Seminar No. 4, 25 May.


Presenter: Nicholas Haysom

Nicholas Haysom is the former Deputy Director of the Centre for Applied Legal Studies.

Date: 25 May 1989

Venue: University of the Witwatersrand, Johannesburg, South Africa


Trade unionists in the Pietermaritzburg area refer to the nearby Edendale Valley as the 'valley of widows'. It has more widows, they claim, than any other valley in South Africa. Since vigilante violence erupted in that region, over 1 200 persons have died, thousands have been subjected to violence against their person or property and an estimated 30 000 persons have become internal refugees.1 In April 1989 the 'Pietermaritzburg monthly death rate' dropped to 34, which was considered a distinct improvement on the levels of violence over the preceding eleven months.2 Yet this figure represents more than one violent death every day. To place these figures in context, more than twice as many persons died as a result of vigilante and counter-vigilante violence in the greater Pietermaritzburg area than died in Beirut.3 Yet in 1988, South African television viewers were regularly exposed to the human tragedy that is Beirut. They saw little visual footage of the Edendale Valley, if indeed they saw any at all.4

Township residents in South Africa are now far more likely to die as a result of vigilante violence than they are as a result of confrontations with the South African police. By October 1988, over ninety percent of unrest-related deaths were caused by vigilante and counter-vigilante violence in South Africa.5 For these residents, the vigilante phenomenon has become the most terrifying manifestation of a conflict-ridden society. How is it that such extreme violence can be tolerated? How can it occur with such little official concern, such limited local and international media attention?

To answer this question it is necessary to review not only the emergence of the vigilante phenomenon in South Africa, but the emergence of similar groups in other parts of the world or, more properly, the roots of this phenomenon, not in policing practices, but in the contemporary theory of counter-insurgency known as 'low intensity conflict'.

Low Intensity Conflict

This paper suggests that the operation of vigilante groups in South Africa's black areas since 1985 is an expression of the militarisation of South Africa, that the prevalence and operation of these groups should be seen as the internal equivalent of the strategy of destabilisation of neighbouring states. It is a low intensity civil war which appears to be conducted an arm's length away from an aggressive state. In fact, however, the state benefits in a variety of ways from the conflict it licenses, indeed more so that it would from direct intervention.

As commentators have noted, there is a strong parallel to be drawn between vigilante violence in South Africa and that in El Salvador and the Philippines.6 In those countries, violence by vigilante and civilian units has become a central component of the mode of repression adopted by the governing regimes.7 Modern counter-insurgency theory lays stress on "total war" – which incorporates a 'winning of hearts and minds' (WHAM) component. The necessity of destroying popular insurgency without appearing to be directly waging war on the populace is the dilemma that has led to the US-sanctioned 'low intensity' civil wars of Central America, the vigilante movements of the Philippines and the destabilisation of states in Southern Africa. The logic of this theory involves the clandestine creation of surrogate armed forces, but which organisations appear to emerge 'spontaneously' from the 'people' themselves. It is then claimed that the Contras, Unita, or the Philippine vigilante groupings are an expression of popular support, or popular rebellion, as the case may be.

The explicit adoption of low intensity conflict strategy by South Africa's security establishment appears to fall into two phases.8 The adoption and implementation of the military establishment's current strategic blueprint coincided with the entrenchment of the influence of the military establishment through the accession to power of P W Botha and General Magnus Malan. This blueprint is a direct application of the military theories of the French General Andrew Beaufre.9 Beaufre, who was a general in the Algerian civil war, argued in his book 'Introduction to Strategy' for a military approach that acknowledged the existence of an extended battlefield. In Beaufre's theory, the battlefield must be extended to encompass all aspects of a civil society, particularly social and ideological spheres, such as the radio and the classroom. According to Beaufre, the proper concern of the military should be extended to co-ordinating all aspects of a civil society.

The dissolution of the boundaries between military and civil society as Beaufre proposed has now passed into South African political lexicon – 'total strategy' in response to 'total war'. The clearest adoption of Beaufre's recommendations, and equally the clearest expression of the influence of the military establishment in South African politics is the elaborate co-ordinating security structure known as the Joint Management Security System.10 It should be mentioned that a young South African lieutenant, Magnus Malan, served as a military observer in Algeria in the very regiments under the command of General Beaufre.11 Beaufre's book has long been prescribed reading at the South African Military Academy.12

In the mid-1980s, South African strategists appeared to be swinging towards the more practical theories of Colonel J. J. McCuen, who developed his theories of counter-insurgency warfare in Vietnam. McCuen's writing belongs to the genre known as low intensity conflict theory. This school of thought is now dominant amongst United States counter-insurgency theories, particularly over the 'Westmoreland strategy' applied in Vietnam. General Westmoreland's approach to counter-insurgency was to make maximum military use of technologically superior resources and firepower to smash a third world enemy. The Westmoreland school believed in 'asphalting Vietnam'.13

Sarah Miles in her informed review of low intensity conflict in Central America reveals some of the distinctive features of low intensity conflict.14 Low intensity conflict is not less violent than conventional warfare. It simply operates on the premise that there should be no direct security force intervention. If the security forces have to intervene in a conflict situation they have already lost the strategic initiative. Military involvement can win battles but it will lose counter-insurgency wars. Low intensity conflict postulates that there are neither battlefields nor victories. Low intensity conflict is a war of destruction and manoeuvre in which optimum circumstances for strategic response are created. Victory is measured in terms of 'avoidance of outcomes' and by 'attitudinal changes' in a target group.15

This redefinition of victory means that even where there is no outright victory, enough physical and political damage will ensure that the revolutionaries cannot win either. Low intensity conflict is a war of counter-organisation. The new generation of counter-insurgency experts recognise that 'pacification' has population not territory as its objective.16 In the Southern African scenario, this is illustrated by the apparent willingness of the backers of the NMR or Unita to allow that group to actually seize 'power' in Mozambique.

Low intensity conflict, according to its theorists, is total war at the grassroots level. In its offensive (Nicaragua, Angola) and defensive (El Salvador, Philippines) positions, it seeks to utilise or create apparently popular groups (composed of grassroots civilians) using 'revolutionary' violence. The role of the military is confined to co-ordinating the distribution of economic aid to target social groups and the supply of logistic and informational support to surrogate groups. 'Foreign' forces, e.g. the U S in Central America or the SADF in the townships, cannot 'win over' the local population. According to Miles, the new U S military strategists are less concerned with winning battles than with separating the enemy from its people base, and neutralising their social structure.17 Apart from the level of violence it introduces, low intensity conflict is 'dirty war'. "It penetrates into homes, families, the entire fabric of grassroots social relations; there are no 'civilians' in a low intensity conflict … it is a science of warfare whose goal of controlling the qualitative aspect of human life merits the term 'totalitarian'".18

In 1986 General Meiring, the former general of the South West Africa Territory Force, expressed his preference for McCuen's theories over the abstractions of Beaufre. They are more practical, more explicit on the particular 'hard war' steps such as the creation of counter-revolutionary groups, and 'soft war' (WHAM) steps including electrification of townships the military should undertake in its WHAM strategy. In late 1986, McCuen's theory had been precised to a 75-page document entitled 'The Art of Counter-Revolutionary Warfare' and distributed throughout the Management Security System.19

What is apparent in McCuen's theory, and in the speeches of its South African proponents, is that the creation of a political solution requires not a commitment to political bargaining, not even top-down reform, but a bottom-up reconstruction of political forces. The move from total strategy to (active) low intensity conflict is the subtle move from controlling dissent to reorganising politics. Phillips and Swilling date the shift as occurring in 1085/6, the same time that vigilantes emerged.20


It is necessary to begin by describing the operation and emergence of vigilante groupings in South Africa. This paper records only the patterns and implications of vigilante activities, rather than providing specific details of the various vigilante groupings in South Africa. A full survey of these groups can be found elsewhere.21

The term 'vigilante' is itself a source of confusion. In South Africa the term 'vigilantes' connotes violent, organised and conservative groupings operating within black communities, which, although they receive no official recognition, are politically directed in the sense that they act to neutralise individuals and groupings opposed to the apartheid state and its institutions. These features, and the fact that they are alleged to enjoy varying degrees of police support, is all that links the A-Team, Ama-Afrika, Pakatis, Mabangalala, Amadoda, Witdoeke, Amasolomzi, Amabutho, Mbhokhoto and the Green Berets.

Vigilantes are not an entirely new phenomenon in South Africa. For example, vigilantes supervised by the Ciskeian authorities terrorised the inhabitants of Mdantsane during the course of a bus boycott in that town from June to October 1983.22 However, 1985 saw a sudden proliferation of such groups as well as the emergence of their more complex urban counterparts.

In 1986, a survey of 13 communities which had experienced vigilante violence revealed a distinct pattern in this new phenomenon.23 Subsequent reports have confirmed this pattern.24

Firstly, as community leaders from the Cape to the Transvaal reported, nationwide vigilante activity in the form of violence against members of anti-apartheid organisations commenced in 1985. The intimate connection between the emergence of vigilante activity and the more general political crisis in South Africa is evident from the fact that the vigilantes emerged in 1985 as the political crisis in South Africa deepened, and from the fact that as the crisis of control over black areas extended geographically, so did the incidence of vigilante activity.

Secondly, the composition of both the vigilante leadership and the victim groups were broadly the same in all regions. The target groups were those perceived to be resisting apartheid institutions whether they be students campaigning against 'Bantu education', community leaders creating alternative black municipal structures or communities resisting the jurisdiction of homeland authorities. Vigilante leadership is comprised mostly of functionaries in the homeland governments (including chiefs), and in the urban areas members of the state and local state organs (police and community councillors), or members of an "embryonic middle class with an interest in stability and a natural inclination to conservatism".25

The third feature in the pattern of vigilante violence is that the vigilantes appear to enjoy police support, operating brazenly as if there are no legal consequences to their extra-legal violence.

It can not be proven that all vigilante groups have received direct sanction or open support from the security forces – although they allegedly have done in several areas.26 Direct support is not necessary for the generation of vigilante conflict. A mere reluctance to curb vigilante activity or a failure to intervene in conflict in the townships allows one favoured group a substantial advantage over the other. The effect is much the same whether the police actively sanction and support the vigilantes or whether they merely appear incapable of or reluctant to curb vigilante activities. This is particularly the case where the vigilante group has access to firearms. The police's passivity while a vigilante gang killed community leader Mayise in Leandra (Transvaal), an impi of Inkatha supporters marched into Lamontville (Natal), or the Mbhokhoto leaders pursued an intensive regional campaign of intimidation in KwaNdebele must be contrasted with the police's vigorous dispersal of anti-apartheid organisations or trade unions. When the victim communities or organisations attempt physical contest with the vigilantes, police intervention has been supportive of the vigilantes.

The vigilantes' use of township council facilities (notably in Thabong and Ashton) and resources provided by homeland governments (in KwaNdebele and Ciskei), reveals that support for vigilantes may take a variety of forms. A copy of minutes of a meeting between a senior police officer and black traders in the Vaal triangle area on 13 November 1985 suggests that police attitudes could have actively prompted vigilante formation in some areas. At this meeting, the police officer offered to arm the traders and encouraged them to form a self-protection organisation.27 It should be mentioned that it is nearly impossible for a black South African to acquire a gun licence without police approval. In Natal many of the vigilante warlords openly carry firearms, and there is evidence to suggest that the police have armed some of these warlords, or tolerate the bearing of firearms when they know that the warlord has no permit to carry a firearm.

Finally a distinctive feature of vigilantism is the extreme and brutal nature of the violence. Thus vigilante violence is associated with the brutalisation of the body of the victims including the dismemberment or decapitation of the victims. Vigilante violence is extreme and symbolic terror.

Although the detailed composition and operation of these groupings varies from region to region, the face of the vigilante phenomenon is well-illustrated by the following random but representative incidents described in Mabangalala28 and elsewhere.

In April/May 1985, a vigilante grouping calling themselves the Phakatis emerged in Thabong township in the Orange Free State. The grouping, openly using the facilities of the municipal authority, embarked on a campaign of indiscriminate assaults on youths whom they believed were involved in the school boycott. One night they apprehended a boy on the streets, David Mabenyane, and whipped him so severely that he died. After whipping the boy, and while he was still alive, they dropped him at the local police station.

During August 1985 armed gangs of Amabutho took to searching houses in Umlazi, Natal, claiming they were looking for United Democratic Front (UDF) 'troublemakers'. Mr B M, a UDF supporter, was at home one night when the Amabutho arrived at his house, surrounded it and set it alight. His brother, Michael, attempted to flee with his infant niece, but was shot in the head. His elder sister Florence was also shot as she tried to escape the flames. Mr B M recognised one of the three armed men as a member of the KwaZulu government.

In 1985 the KwaNdebele homeland leader, Simon Skosana, launched a vigilante organisation called Mbhokoto. On 1 January 1986, a large group of Mbhokoto vigilantes from KwaNdebele abducted over 400 men from the Moutse district, a district resisting the jurisdiction of the KwaNdebele homeland authorities, who were taken to a community hall in the capital of KwaNdebele. There they were ordered to strip and were severely beaten for several hours before being released. The then Prime Minister of KwaNdebele supervised the assaults. Some of the victims identified their assailants to the police. Police have not as yet apprehended any of the assailants.

In KwaNobuhle, a township bordering on Uitenhage in the Eastern Cape, a vigilante group calling itself AmaAfrika emerged in the latter part of 1986. On 4 January 1987 a mob marched through the township destroying the houses of 14 activists and killing two persons in a 12-hour attack. All the houses they attacked were in the area now known as Old KwaNobuhle, and in particular in the very oldest section known as Angola. Most of these houses were the homes of members of the United Democratic Front-affiliated organisations such as the Uitenhage Youth Congress or area committees. A report on the incident commented:29

One of the most disturbing aspects of the day's events was that the KwaNobuhle Municipal Police and the South African Police apparently made little effort to curtail the violence and destruction. It appears to have been pre-arranged by the two parties with the understanding that the destruction and assaults would be perpetrated by the vigilantes with the police monitoring the events to ensure the safety and success of the AmaAfrika. … They also used it as an opportunity to take more detainees.

Leandra illustrates many of the elements of vigilantism. In Leandra, a black township on the East Rand, the township residents had been involved in a grassroots campaign to improve their living conditions and to prevent the forced removal of the inhabitants of the township. To this end, they had formed an alternative civic structure to the officially approved municipal council. By late 1985 the organisational strength of the alternative Leandra Action Committee (LAC) was such that the authorities were compelled to negotiate with one of its leaders, Chief Ampie Mayise, and not the officially recognised black councillors. Shortly thereafter vigilantes began brazenly attacking members of the Leandra Action Committee. The attacks culminated in a mob assault on Chief Mayise's house on 11 January 1986, during which Mayise was publicly hacked to death. A policeman alerted to the attack by Mayise's call for assistance to the nearby police station was ordered not to intervene. Shortly thereafter youths in the township were forced to flee the area for fear of their lives. The leader of the LAC's house was later attacked and the Leandra Action Committee collapsed.

In January 1986 vigilantes, calling themselves 'Witdoeke', emerged in four squatter camps in the Cape Peninsula referred to here as Crossroads and KTC. The communities numbering nearly 70 000 persons had been engaged for several years in a struggle with the authorities over their right to live in these squatter settlements. They had persisted in their campaign despite detentions, threats and intrigue by the authorities. Vigilantes, with the police allegedly intervening only to assist the Witdoeke, tore through the camps in May and June, destroying and burning the houses and driving the inhabitants out. Inhabitants of the squatter camps allege that police assisted the Witdoeke on 17 May at the camps by breaking up groups of resisting residents and clearing the way for Witdoeke to penetrate the camp. Police watched the illegal destruction of property and the assaults. They allegedly intervened whenever the Witdoeke were under attack. During the attack on KTC by the Witdoeke which commenced on 9 June, the police were alleged to have broken the defence line formed by resisting residents. Witdoeke also took 'prisoners' and tortured them without police intervention.30 In the two attacks 53 people were killed and 7 000 shacks demolished. Yet the government has denied any responsibility for the tragedy. The vigilantes accomplished in a few weeks what the state had failed to do in 10 years. The 70 000 refugees were compelled to seek refuge in other townships including the government-designated option – Khayelitsha.

Why Vigilantes?

A description of vigilante activities does not answer the two most important questions – why did vigilantism come to the fore as a form that the repression of popular organisations took, and why did vigilantes only emerge in 1985?

To understand why vigilantes emerged in 1985 we must review briefly the intensity and the scale of apartheid's political crisis during 1984-1986.

In 1984 the United Democratic Front and other organisations launched a nationwide campaign of opposition to the exclusion of blacks in the new Constitution. This opposition was also directed at what came to be known as the 'Koornhof Bills'. The success of the campaign was shown by the very low percentage polls in the Indian and coloured elections and the elections for community councillors.31 The consequence of the campaign was the rapid politicisation of these communities.

At about the same time, the economic climate in South Africa began to deteriorate as a recession began to make itself felt and unemployment reached extremely high levels. In the urban areas, short-sighted parsimony by the authorities for black local government led to pressure on the community councils to be self-financing. When the community councils sought to impose increases in rentals and service charges in order to meet their budgets, a number of civic associations began to orchestrate protest against these bodies which were perceived to be corrupt and self-serving. The protests took the form of marches, attacks on the councillors, and a rents boycott which involved up to 300 000 persons and cost the state over R188-million.32 During the same period South Africa experienced an increasing number of national and local wage disputes.

At approximately the same time, late 1984, school pupils commenced a nationwide campaign against inadequate education and were confronted with an authoritarian approach to their representations. A school boycott spread throughout the country so that by late 1984 even areas in the bantustans had been affected.

This confluence of political and economic factors resulted in intense, widespread and diverse protests in the townships in the Vaal Triangle area from September 1984 and subsequently in other townships throughout South Africa. The existing administrative structures collapsed in the majority of urban areas in the Transvaal and in the Eastern Cape. By 1985 only five of the 38 black local authorities could operate effectively, and community councillors and policemen who had resigned were being housed, for their own safety, outside the black townships.33 In late 1984, one of the largest stayaways in South African history took place when 800 000 trade union members reacted in response to the call of township residents and schoolchildren complaining of inferior schooling, township conditions and police brutality.34

The first response to the civil rebellion was 'maximum force' policing policy, later reinforced by emergency powers and military personnel. Police were unable to control the rebellion which took a myriad different forms. In one area they were called upon to deal with a strike, in another a consumer boycott, in another a stayaway, in another a demonstration, in another a march on municipal council offices, in another a school boycott. The rebellion was the converse of total strategy – it was total resistance. In order to re-impose and bolster government institutions, particularly the community councils, the police were supplemented by units from the South African Defence Force. Regular security force patrols seemed to confirm township residents' fears that they were under siege by an occupying army The mushrooming rate of clashes between security forces and township residents came to a head in July 1985 when the government declared a state of emergency. By the end of 1985 it was estimated that over 10 000 people had been detained under laws relating to the security of the state or homeland security legislation.35 Over 1 000 persons had died in unrest-related events since 1984.36

The distinctive features of the 'maximum force' policing policy was its attempt to control and deter the rebellion by means of fire-power. The incidents which characterise the policing practices of 1985 are the Langa massacre (19 funeral mourners shot dead), the Mamelodi massacre (12 unarmed protesters shot dead), the Winterveld massacre (9 peaceful protesters shot dead), and the Trojan Horse incident (3 shot dead when hidden police fired on youth stoning a truck). These incidents showed that the police were willing to use lethal force in an indiscriminate or disciplinary manner, more specifically in the form of SSG shot, R4 bullets, to dissuade township residents from participating in the process.

It is in the context of a crisis of control in black areas and the failure of the police to restore 'order' that vigilantes emerged and attempted to alter the balance of power in the townships – albeit under the slogan of 'law and order'.

By mid-1985 it was apparent that the first response – maximum force policing – was failing. That grisly index of 'unrest', the daily death rate, had not been affected by the state of emergency. In August the death rate had climbed to 5.3 unrest deaths per day.37 The policing methods had further politicised the black community and created a degree of social cohesion amongst township residents. What was needed was an additional disorganising strategy. Directly in line with low intensity conflict theory, vigilantes emerged to perform this role. Vigilante groups were specifically suited for such a purpose.

Firstly, the South African Police (SAP) and South African Defence Force (SADF) are limited by potential publicity and hindered by legal considerations in perpetrating the extreme terror and violence needed to combat popular organisations. They cannot publicly hack a popular leader to death and avoid prosecution or civil claims. Vigilantes are better able to escape both.

Secondly, however brutal these vigilante groups may be, by purporting to act under the banner of 'restoring law and order', they could legitimate their actions, elicit at least some sympathy from the white public and, in some cases, play on a broader disgruntlement with the disruption caused by the unrest (particularly where vanguardist youth groups had been undemocratic in imposing community boycotts).

Thirdly, security forces cannot coerce support for community councils nor can they administer the townships. A Soweto councillor clearly expressed his frustration with conventional responses to township opposition in a speech to a Sofasonke party rally. He stated that he had asked the government to allow councillors to avenge attacks on their homes, and to embark on full-scale vigilante activities aimed at rooting out 'abosiyayinova' (troublemakers) in the townships. "Hopes that security forces wold bring peace were a pipe dream", he said. This statement implied a realisation that councillors must appear to their own constituency as having their own muscle.

Fourthly, unlike the random and indiscriminate violence which takes place when the police confront popular organisations, vigilante terror is more accurately targeted at the leaders of the organisations. The terror comes from within the community, thus rupturing community confidence and cohesion. It becomes difficult for popular organisations to function in the open. In such cases, it is not only the vigilante interests that are served but also those of the official agencies that the popular organisations challenged. Vigilantes in magisterial districts not covered by the state of emergency in 1985 disrupted civic and youth organisations to a greater extent that did the detentions and actions of the security forces in townships under emergency regulations.

Fifthly, vigilantes are cheap. Lennox Sebe, Prime Minister of Ciskei, rationalised the formation of the Green Berets by saying that the vigilantes were acting as low-cost police auxiliaries. In this sense the vigilante phenomenon represents a Reaganomic tendency in one of South Africa's growth industries. It is the privatisation of repression.

Sixthly, as the proponents of privatisation argue, privatisation has ideological benefits. Vigilante violence, reported locally and internationally as 'black on black' violence, seems to justify the police presence in the townships on the ground that the residents have a racial propensity to internecine strife. More importantly, the destruction of organisational opposition to the state is accomplished without the intervention or presence of the security forces, and without the same public outrage that would follow brutalities committed by the security forces. The persistent use of the label 'black on black' violence by the Department of Information and the media has as much heuristic value as describing World War II as 'white on white' violence. The label does, however, obscure the connections between the violence and apartheid structures.

Seventhly, a cowed and disorganised community allows for a political vacuum which may be filled by community councils or homeland political parties, and which in turn allows these bodies to impose their will on the communities. This is so particularly for communities who, without access to direct political power, have only collective organisation as their weapon.

Communities which are victims of vigilante action are faced with a dilemma as to how they can best resist the vigilantes. Should they go on the offensive and commit a pre-emptive violent assault on the vigilantes? There is no doubt that they would, in such an event, face the full might of the very law and order machinery which had failed to protect them. Furthermore, vigilantes are in most cases better armed, and defence may prove fatal. In the final analysis, victims may create their own causes, wage their own wars, but they do so guaranteed by the State's resources. Just how effective the vigilante phenomenon is, and how difficult it is to counter a spiral of vigilante violence can be seen in Pietermaritzburg township.

The Edendale valley near Pietermaritzburg has since 1987 become the site of bloody and violent clashes between residents who are supporters of Inkatha and those who are perceived to be its opponents, clashes which have claimed a staggering 1 200 lives. By March 1988 nearly two-thirds of persons killed nationwide in unrest-related events were killed in the Pietermaritzburg district.38 By March 1989 over 1 200 persons had been killed. In other words, vigilante violence and counter-vigilante violence in Pietermaritzburg is now the single most important cause of death in South Africa under the state of emergency. Explanations of the violence by Inkatha spokesmen solely in terms of structural poverty do not explain why conflict erupted in Pietermaritzburg, or why it erupted when it did.

Since 1980 there had been occasional violent clashes between supporters and opponents of Inkatha in Natal but these clashes did not extend to the Pietermaritzburg district until after 1985.

In 1985 the situation in Pietermaritzburg began to change, when UDF-affiliated youth congresses began to establish themselves in the Pietermaritzburg district. A series of minor skirmishes involving Inkatha supporters as well as the youth groups opened hostilities. In the same year the United Workers' Union of South Africa was formed as the trade union wing of Inkatha and competition between it and other unions sharpened. After its formation physical attacks commenced on members of its rival union confederation, Cosatu. For example, in December 1985 in nearby Howick a mob led by nine Inkatha members dragged two Cosatu worker leaders and a young woman out of a car and executed them. (To this day the nine members identified by an inquest magistrate as the culprits have not been charged or prosecuted.) In 1986 solidarity action in support of Cosatu-affiliated Sarmcol workers in Howick was heeded by a significant number of residents in the greater Pietermaritzburg area. Inkatha began to lose ground.

On 5 and 6 May 1987, UDF/Cosatu called a stayaway in protest against the parliamentary elections. The stayaway was opposed by Inkatha. It was ninety percent successful, but Inkatha attributed this success to the solidarity action by bus drivers, members of the Cosatu-affiliated Transport and General Workers' Union. Thereafter Inkatha commenced a recruitment campaign in earnest. This drive was associated with threats of violence and actual violence. UWUSA mounted a campaign to compel bus drivers to resign their membership of the Transport and General Workers' Union. In the course of the campaign four Transport and General Workers' Union drivers were shot dead. The residents of Vulindlela were ordered to join Inkatha by 4 October 1987. This drive was met with increasing physical resistance and attacks.39 By January 1988 the conflict had reached its bloodiest level when the monthly death rate reached 162.40

In some cases the violent recriminations against the opposition factions were clearly focussed. In general, however, much of the violence has been indiscriminate and regionally based. As the conflict has escalated, so Pietermaritzburg has become politically and geographically divided. If one lives in an Inkatha area one is deemed a member of Inkatha. Many 'members' are ignorant of the policies of their respective groups. Certainly many of those who claim to be UDF members are not aware of that organisation's policies, have never attended political meetings and can cite no affiliate to which they belong.

Undoubtedly, the central catalyst for the violence is the conflict over political support or superiority in the region – Inkatha blames an 'attempt to render the areas ungovernable' and UDF/Cosatu allege that it is the result of Inkatha's determination to root out any alternative political presence in its own back door. However, the intensity and scale of the violence can only be explained by reference to a third party's role – that of the SA police. Cosatu's first response was to attempt to force the police to intervene, arrest, charge and prosecute perpetrators of the violence. They believed that it was police passivity which had encouraged the brazen violence perpetrated by Inkatha supporters. Cosatu has stated that it believes that had the police mounted 4 or 5 vigorous and effective prosecutions in September, October or November 1987, well over a thousand lives might have been saved.41 To this end legal assistance was rendered in the form of statement-taking, tracing witnesses, delivering potential witnesses to S A P charge offices and identification parades, and assisting the attorney general by means of preparation and research.

Disillusionment with this strategy set in and Cosatu has since March 1988 accused the police of siding with Inkatha. In January 1989 Cosatu attorneys, believing that their informants' statements revealed a pattern of police collaboration with the Inkatha warlords, researched all incidents of violence in Imbali Stage I. The lawyers also recorded police responses to the 29 incidents monitored.42 Their report alleged that the majority of attacks were perpetrated by Inkatha members, yet the majority of persons arrested after such attacks were Cosatu or UDF supporters. The police have shown a differential response to suspects by attempting first to identify whether a suspect was a member of the UDF/Cosatu or of Inkatha. In this regard they have used Inkatha personalities to assist them in identifying troublemakers. Furthermore, the police have since 1987 detained well over 1 076 UDF or Cosatu members and detained only a handful of Inkatha members despite claiming that the detentions were related to the 'unrest'.43 The report suggests further that Inkatha members were less likely to be held in custody if apprehended. The police also tolerate the illegal carrying of arms by Inkatha personalities. Yet they raid non-Inkatha communities precisely on the pretext that they are searching for firearms.

Cosatu's lawyers have cited incidents in which 'UDF' complainants were discouraged from laying charges, and have stated that when charges were laid the investigations have failed inexplicably to proceed to any finality. Lawyers have pointed out that the investigation and lacklustre prosecution of the alleged perpetrators of the violence in Pietermaritzburg stand in sharp contrast to the vigorous and zealous prosecution of UDF activists in the area. They also point out that the same personalities appear again and again in statements describing the violence.44 The fact that only a handful of prosecutions have followed the 1 200 murders speaks for itself.

Cosatu spokesmen have insisted that the conduct of the police should not be seen as the result of lazy or rogue policemen. Cosatu refers to the Minister of Law and Order's comments in which he claims that Inkatha is innocent and that blame for the conflict should be laid squarely at the door of Cosatu.45 The role of senior policemen in attempting to give Inkatha a positive image when it was being labelled the culprit by even conservative bodies is indicative of policy to assist Inkatha taken at the highest level. On 23 May a final order was granted restraining the police from carrying out a pattern of unlawful assaults on the residents of Mpophomeni (a Cosatu-supporting community), and from failing to protect them from the attacks of nearby Inkatha communities.46 The State President's refusal to accede to requests by Lawyers for Human Rights to appoint a commission of inquiry into the role of the police would support Cosatu's claims.

The second attempt by Cosatu to limit the violence was the launching of nine Supreme Court urgent interdicts. In this way Cosatu tried to police the vigilantes by means of the civil courts. Five applicants or witnesses to these proceedings have been killed or assassinated and none of the murderers in these killings have been prosecuted to date. Cosatu's lawyers were left in the unenviable position of persuading witnesses and applicants to come forward knowing that legal proceedings may expose such applicants to violent retaliation but could not afford them protection. By August 1988 the situation had become so serious that lawyers were questioning whether the courts could offer any effective protection to persons who sought its assistance.47

For example, Johannes Mthembu and his four sons, living in Imbali near Pietermaritzburg, had clashed with local Inkatha personalities in late 1987. Mthembu brought interdict proceedings in January 1988 after a clash in which one of his sons, Elphas, was shot. After service of the court papers one of the Inkatha leaders cited in the interdict arrived at Mthembu's house. After an incident (which is the subject of current litigation) two other sons, Smallridge and Simon, were shot, one of whom died. In July 1988, shortly before the case was due to reopen, Mthembu's fourth son, Ernest, was shot dead by an unknown assassin when he opened the front door of his house.48 The family has had to flee the area. It is certainly possible that the legal proceedings initiated by Johannes Mthembu aggravated his position and exposed him to danger.

In September 1988 Cosatu entered into a peace pact with Inkatha in which both organisations committed themselves to a public acknowledgement of the right of freedom of association. The declaration was to be enforced by an informal expedited arbitration procedure whereby a complaints adjudication board (CAB) chaired by former Supreme Court judges, would determine whether members of the respective organisations had breached the code. It would then be left to the organisations each to discipline its own members in accordance with the findings of the CAB. The CAB stratagem was based on a recognition that whereas the courts and the police were, at best, incapable of restraining the violence, the very organisations which did have effective power to control the residents of Pietermaritzburg's townships should shoulder that responsibility.

Inkatha respondents have subsequently refused to participate in CAB inquiries alleging that their testimony may be used in evidence against them in subsequent criminal proceedings. In May this year, a Cosatu witness, Nicholas Duma, was assassinated three days after having testified against a warlord. In the same month, Jabulani Ndlovu, a NUMSA shop steward who publicly alleged police collusion with Inkatha warlords, had her house fire-bombed. Her husband and daughter were killed during the attack and she herself died several days later from burns. Cosatu has now suspended its participation in the CAB.

Cosatu has stressed that even if Inkatha/UDF have the political will to embark upon genuine peace initiatives, partisan police will undermine any peace pact.49 Indeed, the police have been accused of directly subverting the peace process; this is manifest in their attempts to discharge current negotiations.50 Minister Vlok, for example, has openly condemned peace initiatives, and has restricted UDF participants in previous peace talks.

From Vigilantes to Community Guards

A national trend which has caused concern amongst human rights activists and victim communities alike is the induction of vigilantes into the state's formal law and order machinery. The incorporation of many of the Queenstown vigilantes into the Queenstown Commando is one such example. A more prevalent form of this process is taking place through the appointment of community guards, a form of municipal police under the control of the community councillors.51 It has already been reported that erstwhile vigilantes have made application to join the community guards.52

The municipal police ('greenflies') and police auxiliaries ('kitskonstabels') have indeed made a special contribution to converting the mood of the townships from protest to fear. Human rights groups from the Eastern Cape to the Transvaal report numerous incidents in which the municipal police have assumed the methods of the vigilantes. Complaints include torture, beatings, thefts and forcible evictions. Their responsibilities have less to do with crime prevention and more to do with pacification.53 More importantly, the Greenflies' increased involvement with the policing of the townships has gone hand in hand with a withdrawal of the security forces from these areas. The use of municipal police and police auxiliaries is in line with a militarisation of township society while appearing to demilitarise the government of the townships.

Vigilantes and Repression

It remains to place the vigilante phenomena within the overall strategy of repression and reform, within the current imperative to reconstitute a black politics dependent on and amenable to government policies.

The vigilantes have become part of a multi-faceted stratagem to create a moderate black politics. Vigilantism operates in tandem with other repressive practices. As an editorial in the South African Journal on Human Rights points out.54

The state of emergency provides a legal regime for the pacification and administration of the townships.

The mass detention of leaders and members of popular organisations and the intense police control over the freedom of assembly and expression results in an organisational stasis or vacuum in black areas.

Vigilantes attack surviving popular organisations. Thereafter vigilantes together with municipal councillors and their police, Kits Konstabels, or organisations such as Mbokhota, take charge of the townships by moving into the vacuum so created. Activists returning to the area are isolated and victimised.

Upon this platform black local authorities are given both the financial and policing resources to administer or upgrade the area. Here, the joint management councils, convened and chaired by members of the SADF play a crucial role in co-ordinating resources so as to win the 'hearts and minds' of township residents. It is equally clear that the JMCs co-ordinate not only the upgrading of townships but also their downgrading and make devolution of power dependent on the existence of vigilantes.

This stratagem has been powerfully effective. In 1986 police could not enter Alexandra except in armoured vehicles. In 1989 attorneys representing Moses Mayekiso, the community and trade union leader, were encountering difficulties in persuading previously militant residents to testify on behalf of Mayekiso in open court. The police now stroll through the township, unarmed.

Paradoxically, the very withdrawal of troops and police from the townships and their replacement by vigilantes, municipal police, planners and engineers is directly in line with increasing military influence over civil society. The new military theory which lays greater importance on the manipulative but invisible role of the military in co-ordinating a comprehensive strategy involving economic, political, psychological and social factors. The logic of modern counter-insurgency theory requires that the opposition be destroyed without alienating township residents and without the security forces being seen as directly involved in a war on the masses.

However, the prospect of this comprehensive stratagem actually winning the 'hearts and minds' of township residents is remote. The vigilantes have not been able to create popular support for their politics. Instead they have created communities which are either sullenly cowed or violently vengeful. These communities are destined to remain in a state of violent stability. While there are certainly short-term gains for the state, in the long term the vigilante phenomenon will erode any possibility of a popular 'rights discourse'. Township residents will, with good reason, understand only the language of power. Even for vigilante organisations the gains are short-term. As a commentator on the vigilante phenomenon in the Philippines has pointed out:55

The campaign of the [vigilantes] may bring some short-term results but eventually it must only redown to the gain of the [insurgents] and considerably help their recruitment and propaganda campaigns. Pointing to a [vigilante stronghold] they ask the people is this the type of government you want?

… Vigilantes are themselves victims and objects of an inequitable society. They cannot be the solution. If indeed they are a formula for peace then my country can only be headed for more war.


1 COSATU, 1989.

2 The Star, 23 May 1989.

3 COSATU, 1989.

4 Tomaselli, 1988, reviews the one television feature programme on the Natal violence and suggests that it was misleading.

5 The Star, 10 November 1988.

6 Introduction to Conference on Death Squads and Vigilantes: Block to Third World Development, Catholic Institute for International Relations (CIIR), London, May 1988.

7 Diokno, 1988; Caceres, 1988.

8 See generally Swilling and Phillips, 1989.

9 Ibid; The Star, 10 November 1989.

10 Paper by Andrew Boraine at this conference. (See also Phillips, 1988).

11 Swilling and Phillips, 1989p.

12 Cawthra, 1987p.

13 Miles, 1986, pp 21-3.

14 Miles, 1986.

15 Ibid, p.19.

16 Ibid.

17 Ibid, p.29. See also the interview with General Charles Lloyd, Weekly Mail, 23 June 1989.

18 Ibid, p 45.

19 Swilling & Phillips, 1989.

20 Swilling & Phillips, 1989, pp.22-3.

21 Haysom, 1986; Lawyers Committee for Human Rights, 1987; CIIR, 1988 (b); Phillips, 1988, Aitcheson, 1988.

22 Lakob, 1984; Haysom, 1983.

23 The communities examined in the survey included Crossroads, Ashton and KTC (Western Cape), Queenstown and Fort Beaufort (Eastern Cape), Huhudi (Northern Cape), Thabong (Orange Free State), Umlazi, Inanda and Lamontville (Natal), Leandra, Moutse and Ekangala (Eastern Transvaal), and Soweto (Transvaal); Haysom, 1986.

24 CIIR, 1988 (b), pp 67-163.

25 Seekings, 1985, p.27. In the Pietermaritzburg area the warlords appear to come from both homeland functionaries and the members of such an 'embryonic middle class'.

26 Crossroads, KwaNobuhle and Queenstown; CIIR, 1988 (a), pp 10-14.

27 Haysom, 1986, pp 8-9.

28 Haysom, 1986.

29 Monitor, June 1988, p.46.

30 Lawyers Committee for Human Rights, 1987; CIIR, 1988 (b) p.69-82; Phillips, 1988.

31 South African Institute of Race Relations (SAIRR), 1984: 127.

32 Sunday Tribune, 31 August 1986.

33 SAIIR, 1985, p.89.

34 Human Rights Index, 1985, p.85.

35 Detainees' Parents' Support Committee, 1986.

36 House of Assembly Debates, 25 April 1986, cols.11423-6.

37 'Human Rights Index', South African Journal on Human Rights, 1986, p.110

38 Seekings, 1985: 28.

39 Indicator South Africa, 1988.

40 See the papers in Zondo v Inkatha, NPD Unreported Case – September 1988.

41 Aitcheson, 1988. The average monthly death toll for the area since July 1987 is 51.

42 Cosatu, 1989.

43 Cosatu Report, 1989.

44 Merrit, 1988.

45 Dontzin, 1989, Cosatu Report, 1989.

46 Mbambo v Minister of Law and Order, NPD unreported Case No 79/1989, 23 May 1989.

47 Dontzin, 1989.

48 Natal Witness, 5 July 1988. See court papers in Mthembu v Zuma, NPD Unreported Case No. (proceeding).

49 Cosatu, 1989.

50 Weekly Mail, 15 June 1989.

51 In 1986 an initial R26-million had been allocated for the training of 5 000 guards. Work In Progress (WIP), 29, 1986: p.40; CIIR 1988 (b) Black Sash, 1988.

52 In some areas these agencies have simply taken up where the vigilantes left off. Black Sash, 1988; CIIR, 1988; (b) 31-66.

53 Ibid.

54 Phillips, 1987, 4, South African Journal on Human Rights

55 Diokno, 1988, p.41.


Aitcheson, J. Numbering the Dead: Political Violence in the Pietermaritzburg Region, Centre for Adult Education, U N P, Pietermaritzburg, 1988.

Anonymous Paper, "Vigilantes and the New Policing in South Africa". Paper presented to Catholic Institute of International Relations Conference on Death Squads and Vigilantes: Block to Third World Development, London, May 1988.

'Greenflies': Municipal Police in the Eastern Cape, Black Sash, 1988.

Caceres, G. "Violence, National Security and Democratisation in Central America". Paper presented to Catholic Institute of International Relations Conference on Death Squads and Vigilantes: Block to Third World Development, London, May 1988.

Catholic Institute for International Relations, "Now Everyone is Afraid", in The Changing Face of Policing in South Africa, 1988 (forthcoming).

Cawthra, G. Brutal Force, International Defence and Aid Fund, London, 1988.

Centre for Adult Education, Pietermaritzburg Briefing, University of Natal, Pietermaritzburg, 1989. In 1988, 312 persons were killed as a result of conflict in Beirut. In the same period 682 persons died as a result of violence in Pietermaritzburg.

Cheadle, Thompson & Haysom Attorneys. Memorandum to the Natal Law Society on the Police Role in the Violence in the Pietermaritzburg Area, Johannesburg, May 1988.

Cole, J. Crossroads: The Politics of Reform and Repression, 1976-1986, Ravan, Johannesburg, 1987.

Pietermaritzburg Unrest Statistics and Selected News Reports, Cosatu, 29 March 1989.

Role of the Police in Vigilante Violence in the Pietermaritzburg Area, Cosatu Press Release, 29 March 1989.

Cosatu Report on Mbali Stage I, Cosatu, 29 March 1989.

Review of 1985, Detainees' Parents' Support Committee, 1986.

Diokno, M.S. "Self Styled Guardians of Democracy: Vigilantes in the Philippines". Paper presented to Catholic Institute of International Relations Conference on Death Squads and Vigilantes: Block to Third World Development, London, May 1988.

Dontzin, M.S. Crisis in the Administration in South Africa: A Case Study of Vigilante Violence, Police Partiality and Prosecutorial Incompetence in Pietermaritzburg, Natal. Unpublished, March 15, 1989.

Ebersohn, "How South Africa Fights the Total Onslaught", The Star, 10 November 1988.

Evans, M & Phillips, M. 1988, "Intensifying Civil War: The Role of the South African Defence Force". Frankl, P, et al (editors), State Resistance and Change in South Africa, London.

Gwala, N. (1988) Inkatha, Political Violence and the Struggle for Control in Pietermaritzburg. Paper presented at the Centre for Adult Education Seminar on Violence, University of Natal, Pietermaritzburg.

Haysom, N. Mabangalala, Centre for Applied Legal Studies, Johannesburg, 1986.

Haysom, N. Ruling with the Whip, Centre for Applied Legal Studies, Johannesburg, 1983.

Copy of House of Parliament Debates, 24 April 1989. Transcript of Tape 169 disc 255 Take C5.

"Human Rights Index", South African Journal on Human Rights, 1985 (1).

"Human Rights Index", South African Journal on Human Rights, 1986 (2).

Human Rights Trust, Monitor, Port Elizabeth, June 1988.

Indicator South Africa, Vol. 5, No.3, 1988.

Lakob, M. Human Rights in South Africa's Homelands: The Delegation of Repression, Fund for Free Expression, New York, 1984.

Lawyers Committee for Human Rights, Crisis in Crossroads, New York, 1987.

Lund, J. (1988) The Law and its Limitations. Paper presented at Centre for Adult Education Seminar on Violence, University of Natal, Pietermaritzburg.

Mare, G. and Hamilton, G. (1988) Policing "Liberation Politics". Paper presented at the Centre for Adult Education Seminar on Violence, University of Natal, Pietermaritzburg.

Merrit, N. Detentions and the Crisis in the Pietermaritzburg Area, 1987-1988, Centre for Adult Education, U N P, Pietermaritzburg, 1988.

Miles, S. "Low Intensity Conflict in Central America", Report on the Americas, May-June 1986.

Natal Witness.

Phillips, M. Exploiting Divisions: Vigilantes and State Objectives in Crossroads. Paper presented to Catholic Institute of International Relations Conference on Death Squads and Vigilantes: Block to Third World Development, London, May 1988.

Phillips, M. & Swilling, M. The Politics of State Power in the 1980s, Centre for Policy Studies, University of the Witwatersrand, July 1988.

Rightwing Vigilantes and US Involvement, Report of a US-Philippine Fact-Finding Mission to the Philippines, May 20-30, 1987. Ramsay, Clarke et al, Phara, July 1987.

Seekings, J. "Probing the Links" in Work in Progress (WIP), 1985.

Seekings, J. "Vigilantes and the State", Work in Progress 40, 1986.

South African Institute of Race Relations, Race Relations (SAIRR) Survey, Johannesburg, 1984 & 1985.

Tomaselli, R. (1988) Restoring the Dignity of the Local Community: A Case Study of Impartiality, SABC-Style. Paper presented at the Centre for Adult Education Seminar on Violence, University of Natal, Pietermaritzburg.

+ posts

CSVR is a multi-disciplinary institute that seeks to understand and prevent violence, heal its effects and build sustainable peace at the community, national and regional levels.

Related Content

Disaster-Appropriate Policing in South Africa: Protests and State Violence in the COVID-19 Era

Torture Victims Have a Right to Rehabilitation: A Guide for Service Providers to Assist Victims of Torture in Securing their Right to Rehabilitation

Photo: Getty Images

The State Is Failing in Its Duty to Foster Inclusion and Safety of the LGBTIQ+ Community

Photo: Joseph Chirume

Mob Justice Is a Language in South Africa

Tunisia: The Colonial Legacy and Transitional Justice

Leave a Reply

Your email address will not be published. Required fields are marked *

Translate »