Rauch, J. & Storey, D. (1998). The Policing of Public Gatherings and Demonstrations in South Africa 1960-1994. Paper commissioned by The Commission on Truth and Reconciliation (TRC) Research Department, May.
Janine Rauch & David Storey
This paper was commissioned by the Truth Commission Research Unit, which provided financial support for the research and preparation, May 1998.
Janine Rauch is an independent consultant.
David Storey is a former Research Associate of the Centre for the Study of Violence and Reconciliation.
The creation of a specialised riot control function within South Africa's policing agencies was essentially a reaction to the disorder and political unrest associated with resistance to apartheid. Although the name and structures of the units tasked with this specialist function changed a number of times during the three decades under examination, and the functions were devolved to the various other policing agencies in homeland and self-governing territories, their essential roles remained the same - the enforcement of apartheid laws, the suppression of political protest and the prevention of 'unrest, intimidation and unrest-related crimes'.
Since the nature of their task was inherently public, the police units tasked with riot control played a prominent role as frontline 'enforcers' of apartheid policies, and were viewed with a mixture of fear and loathing by the communities in which they served. Although accurate figures are not available, it is likely that the riot control (and similar) units were responsible for a majority of police killings during the apartheid years.
The units were para-military in nature (by way of training, operational understanding and culture), and brutal in the enforcement of bans on political protest. They operated within a policy paradigm that accepted and supported the lethal use of force. This, combined with the authorities' complete intolerance of protest action, meant that they frequently used maximum force. As the external environment in which they operated took on the character of a low-intensity civil war, their training, equipment and methodology became increasingly militarised.
The South African police agencies were isolated from the international advances that were being made in the field of public order, and were unaware of the new strategies and techniques that other police services were implementing.
By the late 1980s, their policing of areas where violence had become endemic attracted great controversy and numerous allegations of bias and brutality. This reputation, and their inability to fully adjust to the demands of the new political environment, led to calls from numerous communities and political movements for the total disbandment of the riot police or Internal Stability Units under a new policing dispensation.
In 1994 the Government of National Unity inherited a riot control capacity of approximately seventy two (72) units which were illegitimate, over-centralised, fragmented (into SAP and homeland forces), unaccountable, incident driven, and ill-equipped and not trained for the public order challenges which would face them during South Africa's transition to democracy.
Historical Conceptions of Crowd Control
Chronology of Major Incidents
The period under investigation by the Truth and Reconciliation Commission (TRC) is framed, at each end, by events involving the policing of crowds - the march and subsequent police shootings at Sharpeville in 1960, and the inauguration of President Mandela in 1994. A juxtaposition of these two events symbolises the changing demands on the police in respect of gatherings, and the shifts in police methods for managing crowds in South Africa.
Between them, thousands of incidents of public disorder took place; of which, symbolically, the most significant have been the occasional mass killings by police, which act as period markers in the history of South African state violence. The following is a brief list of some such incidents:1
Sharpeville 1960 - a crowd of black residents, marching in protest against Pass Laws to a police station, were dispersed with police gunfire (after harassment by low-flying Sabre jets and police baton charges), resulting in the deaths of 69 people.
Soweto 1976 - hundreds of youths protesting against the use of Afrikaans as a medium of instruction in schools were killed by police and army fire over a period of weeks.
Western Cape 1980 - Schoolchildren boycotted school and engaged in protest action in support of their demands for elected Student Representative Councils. Various confrontations between police and demonstrators took place, and many lives were lost as a result of police action.
Vaal Triangle 1984 - The South African Catholic Bishops Conference estimated that 142 people were killed as a result of police action during four months of township "disturbances". These "disturbances" revolved around the UDF campaign against the tricameral constitution, and involved rent boycotts, schools boycotts & strikes. The state's response to the uprising was a joint army and police operation titled "Operation Palmiet".
Uitenhage 1985 - In Langa township near Uitenhage in the Eastern Cape, 20 people were killed and 23 injured when police opened fire on the march. Most of the dead were shot in the back as they were fleeing from the police.
Trojan Horse, Athlone 1986 - A number of youths were killed when police opened fire on a crowd of demonstrators from where they had been hiding on the back of a lorry.
Bophutatswana 1986 - homeland police killed eleven people and injured up to 200 others when they opened fire on an open-air meeting in the Winterveld.
Sebokeng 1990 - A march of approx. 50 000 people organised by UDF to protest against housing shortages and the education crisis, was fired on by a line consisting mainly of members of the SAP "Reaction Unit"; after marchers handed a petition to the Sebokeng Station Commander, with whom the leaders of the march had agreed to terminate and disperse the march. The firing of a teargas canister by one white constable began without an order to shoot, sparking a chain reaction in the police line which left five people dead and 161 wounded.
Bisho 1992 - 29 people were shot dead at an ANC protest march into the Ciskei homeland when the Ciskei security forces did not abide by recommended new procedures for crowd control.
Shell House 1994 - A day of marches and scuffles involving supporters of the Inkatha Freedom Party (IFP) protesting at the ANC's Johannesburg Headquarters, was notable for the lack of police preparation and presence. Over forty people died in various parts of the Reef, some of them killed by police fire.
Conceptions of Crowd Control in Apartheid South Africa
The banning of political parties and imprisonment of leaders at the beginning of the 1960s effectively reduced levels of domestic public protest against apartheid for the rest of the decade. This was one reason why no specialised anti-riot unit was established inside the SAP in the early years of apartheid. The main "threat" to the political system was deemed to be coming from the liberation struggles and nascent democracies in neighbouring states; resulting in large numbers of policemen being posted for duty on South Africa's borders and beyond (another reason). This was derived from the belief that the liberation struggle in South Africa would follow a similar pattern to that in Rhodesia. Early on, then, the policing of protest was equated with war - an ideological confluence which continued until the 1990s, as demonstrated by the assertion in a UNISA Police Science Study Guide in use in the early 1990s makes that the principles of [crowd] control are "based on principles of warfare".2 In its first version, riot control was seen as part of "counter-insurgency" - which was required to defend the apartheid government from perceived external enemies. Protests were believed to be organised by "communist agitators". Once again, UNISA text demonstrates this view:
urban terrorists work in a number of ways … (including) inciting general public unrest. The task of the police in combatting terrorism is considerably complicated by the expansion of this new phenomenon3
It was these beliefs which further facilitated a punitive style of policing public gatherings, which was already embedded in the racist philosophy of apartheid. The harsh approaches to dealing with crowds were acceptable principally because it was assumed that the crowds in question would be composed of black people; and that black people did not enjoy rights.
In the 1970s, as resistance to apartheid grew, the police began to see the need for a specialised capacity to "control unrest". The SAP embarked on international research and introduced new training methods and full-time riot control units. The approach continued to be informed by the experience of war, with new training methods aimed at creating "shock troops" that could open the way for other police operations. The SAP were not equipped to deal with a domestic uprising of the scale that they faced in June 1976. The policemen who faced massive protest marches at that time were ordinary police officials drawn from nearby stations, possessing no special skills or training in handling crowds. Their lack of capacity was reflected in their tendency to use maximum force. The absence of a rehearsed response to such situations, besides making use of apartheid urban planning and locking entry and exit points to the township, made it necessary for the police to rely on the army to finally quell the uprising.
The late 1970s saw the police re-conceptualising and re-organising to deal with growing "unrest" - one plank of this preparation involved the enhancement of the "riot control" function, others (dealt with elsewhere in the TRC) involved the strengthening of the security police and the formation of death squads. The late 1970s and early 1980s also saw the creation of the bantustan police forces, each of which followed the SAP in establishing some sort of specialised riot control capacity.
The 1980s saw a new wave of large-scale unrest, again involving schoolchildren; which escalated into the popular campaign against black local authorities and the tricameral constitution and saw the launch of the UDF in 1984. Increased levels of political mobilisation and resistance fed into an escalating cycle of action and reaction by the security forces. The role of the police was increasingly defined as controlling riots and unrest (which included arson, intimidation, public violence and killings in black townships) and defending the interests of the white minority, which meant that traditional police functions in respect of crime were neglected, and police resources were concentrated on the enforcement of apartheid policies and repression of dissent. "The primary task of the police is to prevent group formation and concomitant rioting".4
The riot policing function was expanded and removed from "normal" police work, and resources were concentrated around this increasingly central role of the police. This was in direct contrast to reforms being made to public order policing methods elsewhere in the democratic world at this time. The Nationalist government's "Total Strategy" to tackle resistance to its policies was introduced in 1984, the first State of Emergency declared in 1985, and the National Security Management System established in 1986. This provided a new context for riot control, which involved greater centralisation of police resources and decision-making, in line with a highly militarised and coordinated approach by the state to what was being described by the liberation movement as the "people's war".
The introduction of troops to quell unrest in the townships in 1984 was a key moment in the history of crowd control in South Africa, unequivocally signifying a military approach to public protest. The deployment of troops, to quell unrest provided a context for more heavy-handed policing in the townships, enabled the police to follow in the more repressive footsteps of the army. The army has few options in its range of tactics other than the use of lethal force, whereas the police should have an extensive range of alternatives.
It was in this period of high crisis, that arguments for a new approach to the policing of gatherings began to emerge. This new approach was characterised by its authors5 as a "3rd force" model,6 based on the CRS approach taken in France, and it proposed the creation of a new force or agency to deal with large-scale public disorder:
neither the police nor the army are really suited to the task of controlling long-term unrest. While either force can be employed as a short-term measure, any extended employment in this role will bring a number of potentially serious problems. Perhaps the best course would be to raise a specialised force for this task - this force would only be employed to deal with serious public violence, not demonstrations or minor disturbances (which would be dealt with by the police).7
The motivations for a "third force", which would deal specifically with unrest and counter-insurgency operations, included arguments that such a force:
- would avoid the danger of politicising the police or the army, which politicisation would hamper the performance of their normal duties;
- would allow the police to concentrate on their primary (crime-related) task;
- would allow the army to deal with civil-war-type-insurrection and insurgency, which are more in line with its traditional defence function;
- would create a force that is optimised for unrest control and would have appropriate training and equipment for such work, and therefore would be more effective.
The National Party government first considered the proposal for this type of 3rd Force in 1985 and 1986, and then again in 1990. At both times, the proposal was rejected by the security forces, who wielded considerable policy influence. An amended version of the "3rd Force" idea was manifest in the creation of the Internal Stability Division/ISD in 1991 - a separate division of the police specifically tasked with public order functions, instead of a separate force outside of the police or army. The creation of the ISD as a specialised public order component separate from the rest of the SAP was premised on the naive ideas that the rest of the SAP was uncontaminated by its own history of enforcing apartheid (and could thus be "returned" to a normal policing role); and that the performance of the ISD would not impact on the image and capacity of the rest of the SAP to carry out other police functions. The decision to create the ISD was announced in a 1991 speech by the then-Minister of Law and Order, Adriaan Vlok, who said that he expected the Division to grow to a strength of 17500 men by 1997/98 - an indication that unrest was seen as a long-term feature of the South African policing landscape.8
The ISD was perceived in some areas to be aligned, biased, and ineffective in policing township conflict. In a paradoxical turn of history, some township residents begin to articulate a preference for troops to patrol the townships, rather than the ISD. In other areas, exactly the opposite calls were made. The varied responses to the ISU's were evidence of the lack of control and accountability over units. The units had no structural relationship with local police commanders and conducted their activities with no regard to local communities. The ISD did, however, develop a general reputation for abuses of power and unaccountable behaviour, which began to embarrass even senior police managers during the Peace Accord period.
The experience of the ISD provided a context for the revival of the "3rd force" argument in the context of policing the elections and police re-structuring thereafter. In some ways, it was that argument for a "new" force which presaged the establishment of the National PeaceKeeping Force (NPKF) as an alternative to both the ISD and the SADF for the policing of the election. As history would show again (with the failure of the NPKF) the creation of any sort of "special" force to deal with public disorder in South Africa was unlikely to succeed, in the context of high levels of public violence being policed by illegitimate security forces.
The Legal Framework for Crowd Control
The Regulatory Framework 1960 - 1982
The regulatory framework prior to the introduction of the Internal Security Act 74 of 1982 was extensive, providing little or no space for South Africans to exercise freedom of Assembly. The Riotous Assemblies Act 17 of 1956 and the Suppression of Communism Act 44 of 1950 were the two central pieces of legislation enabling state authorities to prohibit and criminalise marches, gatherings and demonstrations. If read together, these Acts gave the Minister of Justice unrestricted power to control gatherings of all kinds.
Riotous Assemblies Act 17 of 1956
This Act conferred wide powers on magistrates and on the Minister of Justice to control and prohibit public gatherings. Section 2(1) of the Act enabled a magistrate, with the authorisation of the Minister of Justice, to prohibit a public gathering if s/he was of the opinion that it represented a serious threat to public peace. The Minister of Justice had wide (and practically unchallengeable) discretion to prohibit a particular public gathering from taking place, or to prohibit a particular person from attending a particular gathering. Further sections of the Act enabled the Minister to impose blanket bans on gatherings in any public place for such period as he specified.
Once prohibited, mere attendance at such a prohibited gathering was not an offence, but all actions relating to the organisation of a prohibited gathering were criminalised. The only real limitations on magisterial and ministerial authority was that the Act defined a gathering as comprising "twelve or more persons" and such a gathering had to take place in a "public place", the definition of which seemed to exclude meetings in halls and buildings.
Suppression of Communism Act 44 of 1950
Section 9(3) of this Act gave the Minister of Justice absolute administrative powers to prohibit a particular gathering, or to ban gatherings generally in any area and for any period specified by him. The size or location of the gathering (including indoor meetings) did not limit the Act's application.
Other Relevant Legislation
In addition, there were numerous other statutory restrictions. The Internal Security Act 44 of 1950; the Public Safety Act of 1953; related legislation such as the Gatherings and Demonstrations Act 52 of 1973 (prohibiting gatherings and demonstrations in the precincts of Parliament); the Dangerous Weapons Act No 71 of 1968; and subordinate legislation such as municipal bye-laws and the Natal Code of Zulu Law all served to encroach on the rights of citizens to assemble.
In 1976 the Minister of Justice imposed a blanket ban on all outdoor gatherings without permission of the Minister or a magistrate, effectively inverting the onus of proof from the state having to show why a gathering should be prohibited, to citizens needing to show the state why a gathering should not be prohibited. This ban was renewed annually until April 1991.
The Regulatory Framework 1982-1994
In addition to consolidating the regulatory framework that had restricted the right to assemble prior to 1982, security legislation governing the later period expanded Ministerial powers to prohibit gatherings. It granted sweeping discretionary powers, closed any loop-holes that had been previously been exploited and extended the range of criminal offences associated with gatherings, while reducing the restrictions on the use of force to prevent and disperse crowds. Furthermore, the already belligerent approach adopted by state authorities to violations of the regulations worsened, and this was reflected in the manner by which crowds were dispersed.
Internal Security Act 74 of 1982
This Act gave limited powers to magistrates, and wide powers to the Minister of Law and Order, to control and prohibit public gatherings. Section 46(1) of the Act gave magistrates the right to prohibit all gatherings in their district for a period of forty-eight hours if they believed that the gathering would endanger public peace. Alternatively the magistrate could allow a gathering to take place, but impose conditions on how it took place. Judicial precedent deemed these magisterial powers to be an exercise of objective rather than subjective discretion. Under section 46 (3), the Minister had the power to prohibit any gathering, if he deemed it necessary or expedient. This was held by the Appellate Division to be a discretion without bounds - provided that the power was exercised for the purposes of the section, the Minister's view as to the necessity or expediency of the prohibition was regarded as conclusive, and could not be challenged on objective grounds.9
The Minister used Section 46 (3) to re-issue annual government notices (first issued in 1976 under the Riotous Assemblies Act) prohibiting all outdoor gatherings except bona fide sporting and religious activities. Until 1991 when this notice was lifted, it served to allow state authorities to selectively ban and police outdoor gatherings using public safety as a disguise for political strategy.10
In addition to active participation in convening, promoting or organising a prohibited gathering, mere attendance at such a gathering constituted a criminal offence.
Other Relevant Legislation
Although the Internal Security Act served as the primary legislative tool to restrict political activity and freedom of assembly, other legislation also played a role. The Demonstrations In or Near Court Buildings Prohibitions Act 71 of 1982 was introduced to prohibit gatherings and demonstrations in or near court buildings and was specifically introduced to prevent protests during political trials and against the treatment of persons held under security legislation. The Gatherings and Demonstrations Act 52 of 1973 preventing gatherings and demonstrations in a specified open area surrounding Parliament was also still in effect. In addition, other 'technical' legislation such as the National Roads Act 54 of 1971, the Trespass Act and by-laws introduced under the Black Local Authorities Act No 102 of 1982 were all utilised by state authorities to ensure that protest was as restrictive as possible in all spheres of life including the use of pickets during labour disputes. Various City Councils passed by-laws requiring, in addition to magisterial permission, council permission (frequently requiring punitive warrants of indemnity) for marches and gatherings.
Declarations under Emergency Legislation
Under the Public Safety Act 3 of 1953 provision was made for particular areas to be declared "an unrest area" by the Commissioner of Police. Such a declaration, which became standard form during the 1980s gave the Commissioner subjective powers to ban gatherings in particular areas.11
The Use of Force on Crowds and Gatherings
The Importance of the Legislative Framework
The South African Police's historical tendency to the early use of force frequently had lethal consequences. The legislative framework within which police operated is crucial in understanding this tendency. State authorities passed laws with the express intention of criminalising public protest, and ensuring the legal space for the application of repressive and lethal methodologies. However, although the legislative provisions were designed to "give the police the capacity to enter the grey area between legality and illegality without fear of the consequences",12 they did not prescribe that the police do so.
A fuller explanation for police predilection towards the use of force can be found when the results of an examination of the legal framework are coupled with three other factors - the primary purpose of the specialist riot function; the training, equipment, strategies and tactics of the police; and the prevailing attitudes of state authorities and concurrent politicisation of police and judicial officers.
The Primary Purpose of the Specialist Riot Control Function
The South African Police expended a lot of energy defending their approach to crowd control by detailing the numerous 'unique' features of the South African situation which, they claimed, restricted their tactical choices.13 In fact, the primary feature distinguishing them from their international counterparts was their purpose, rather than their context. In western democratic countries, police function to prevent public disorder during assemblies and gatherings; and upholding of the law requires managing a crowd to avoid disorder. The primary role of South African policemen fulfilling a riot control function was the enforcement of apartheid restrictions on the freedom to assemble and protest. Upholding and enforcing these laws essentially required preventing gatherings from taking place, and dispersing them when they did take place. Framed within the context of the illegitimacy of the regime they acted for, and against which large-scale political mobilisation was aimed, prevention and dispersal of gatherings inevitably led to the use of force.
Strategies and Tactics and the Use of Force
As other sections of this report will show, the training and equipment of riot police, and the deployment ratios of these policemen relative to the size of the crowds that they confronted, were all based on the assumption that crowds would be controlled and dispersed through the use of force. Crowd control approaches other than dispersal were all reactive and products of the self-imposed constraints of insufficient numbers, inadequate training, and an over-reliance on collective rather than individual personnel protection (i.e. armoured vehicles rather than protective apparel). These 'strategies' never amounted to anything other than a reliance on armoured vehicles for personnel protection, the establishment of a distance between the crowd and the police, and the ability to bring to bear superior firepower against a crowd that broached this 'distance', disobeyed an instruction, or resorted to violence.
In other words, policemen resorted to force because their training was based on it, their equipment was inappropriate for much else, and their strategies and tactics afforded few other options than force, when the threat of force failed to work on its own.
State Attitudes to Protest
Apartheid authorities were threatened by public gatherings and gave unequivocal support to the police in their struggle to repress them. Sympathetic commentators at the time gave a range of possible explanations for the over-reliance on the use of force. However, they failed to explain the fact that many objective conditions for such an over-reliance (see Strategies and Tactics above) were actively created by police management.14 The SAP Generals created the specialist riot control function in the image of their own attitudes to gatherings, and were not unhappy with the result.
The fact that the victims of harsh police action were almost exclusively black served to stifle any moral qualms held by officers in what was, essentially, a racist police force. Furthermore, the race of the victims assisted in justifying the police's use of force to the white public who were taught by the state-controlled media that a crowd of black people equated to a 'mob', and a mob of black people was by nature barbaric and therefore likely to engage in violence.15 Other submissions to the TRC have indicated that political and police authorities actively encouraged the use of harsh and punitive methods and many police officers felt obliged to use maximum force.16 When police were accused of overstepping their extremely wide legal boundaries, they defended and covered-up their actions. In addition to the indemnity provisions prevailing during the state of emergency, policemen accused of using excessive force could rely on the full support of their superiors, the silence of their peers (who would often be used as witnesses) and the indulgence of security-conscious judicial officers. Finally, the absence of the media during dispersal and unrest situations (due to restrictions on the press - see other TRC reports) served to further inoculate police actions from public censure.
Legal Provisions for the Use of Force on Crowds
Riotous Assemblies Act 17 of 1956
If a prohibited gathering did take place, the police were allowed to disperse the gathering by force. Force could only be used if a policeman above the rank of head constable had given three warnings, to those persons gathered, that force would be used if they did not disperse. In addition, the force used should be proportionate to the need, and lethal force could only be used if other means had not proved successful or unless death, serious injury or damage to valuable property had occurred, or seemed likely to occur.
Internal Security Act 74 of 1982
Section 48 of the Internal Security Act authorised a police officer to order a crowd to disperse and to use force to compel obedience to this instruction.17 This authority applied equally to legal gatherings if violence occurred or appeared imminent to the police officer in command. The dispersal command was supposed to be given in both English and Afrikaans (the official languages at that time) and was supposed to contain a time period within which the crowd must disperse before force would be used. Section 48 (2) stated that provided a valid instruction was given, and it was disobeyed (failure to disperse within the time limit), the police officer "may order the police under his command to use proportionate force, including … firearms and other legal weapons". Section 49 provided that firearms and other lethal means "shall not … be used to disperse a gathering", unless lesser means had proved ineffective, and actual or imminent violence by the crowd to persons or property was likely.
The enforcement provisions18 were severely flawed in two ways. First, they allowed "an excessive latitude in the use of firearms: in short, [they] fail[ed] to limit the use of firearms to situations in which the use of force is proportionate to the interests to be protected".19 Second, the enforcement provisions of the Internal Security Act regarding the use of force were very restrictive. They did not allow for the adoption of means other than dispersal to deal with a prohibited gathering, and did not specify the use of other international public order norms such as negotiation to encourage crowds to disperse. Certainly the experience during that period suggests that the police interpreted the provisions narrowly, and did not feel that they had the legal authority to negotiate or adopt any proactive approaches.20
Public Safety Act 3 of 1953
Section 5A(1) of this Act provided indemnity against civil or criminal prosecution for members of security forces acting in good faith. This meant that, in areas covered by emergency regulations, even the flimsy restraints of the Internal Security Act were removed, and, for practical purposes, police and military officers were placed above the law. In effect, the "regulations amount[ed] to a conferring of a broad discretion to kill or injure without legal consequences. The onus on the victim to show that a policeman acted in bad faith when he fired recklessly into a crowd, or failed to fire a warning shot, or neglected to use less drastic forms of 'coercion', is nearly insuperable".21
In addition to the enforcement provisions of security legislation authorising the use of lethal force by police officers, it is important to note that lethal force in the context of crowds and gatherings could also be justified by two other legal grounds - the common law right to protect person or property; and the right to use deadly force to effect an arrest under section 49(2) of the Criminal Procedure Act 51 of 1977. Policemen accused of overstepping the meagre restraints of the Internal Security Act would often use these legal grounds to justify their actions. Since the prohibitions on gatherings criminalised mere attendance at such a gathering, arrests at gatherings were justified, and the use of force in such situations was used as a defence. Similarly, the use of force by the police often provoked the use of violence by members of the crowd, thus enabling the self-defence argument to apply. It is common cause that, relative to international legal standards, (and certainly to the post-1994 constitutional dispensation), the latitude given by South African courts to the use of these two types of defence, particularly when the victim was black, was unacceptable and disproportionate.
Placing the Legislation in Context
As noted in paragraph 4.1, the degree to which the South African Police used force cannot only be ascribed to the legislative framework. The legislative provisions relating to the degree of force to be used were, in fact, permissive, rather than prescriptive. The provisions "enable[d] the police to order the dispersal of an illegal or violent gathering, but [did] not compel them to do so. Similarly they enable[d] the use of firearms and other lethal weapons to enforce the order to disperse, but they [did] not oblige the police to adopt this course."22
However, when seen in the light of the de facto purpose of riot policing in South Africa during the period of this report (to disperse gatherings), the training, equipment and strategies they employed, and the attitudes of state authorities at the time, it is little wonder that the excessive force was used as a matter of course. This unhappy convergence of law, structural organisation, practice and attitudes led one commentator to remark that "… the unnecessary or reckless use of firearms by the police… [amounts to the] widespread systematic use of lethal and violent weapons principally on South Africa's black citizens."23
The Organisation of Riot Policing Capacity
In the 1960s the SAP established Divisional Anti-Riot Units throughout South Africa. This involved members of the SAP doing part-time training and being mobilised into platoons to deal with unrest when necessary. Each Division was more-or-less left to its own methods of crowd control, based on local conditions and knowledge. Equipment was supplied, on demand, by the SAP's central quartermaster. Riot control procedures became part of the police counter-insurgency training programme at Maleoskop. The Divisional Units were phased out at the end of the decade due to competing demands for manpower for the police counter-insurgency operations in (then) South West Africa and Rhodesia. From 1974, young white conscripts were posted to the police as well as to the Defence Force, to assist in addressing these manpower demands.
In the months leading up to the protests of June 1976, a re-organisation of riot policing began within the SAP. This was based on research that the SAP conducted in the US, Western Europe, Britain and Rhodesia; which, at that time, all maintained militarised approaches to public order policing. The SAP established 18 full-time (specialised) riot units and began to train officers in a new approach to the policing of crowds. The approach which they adopted was a colonial one, based largely on public order policing styles in England, Northern Ireland and Rhodesia. By the time the June uprising began, very few members of the SAP had been trained in riot control. Even when the new riot units were fully manned and trained in the latter part of the 1970s, they were not capable of dealing with large-scale demonstrations, but could be used as strike- and boycott-breakers, road blockers and shock troops for Security Branch raids in the townships.
The widespread protest that marked the 1980s was met, from the state, by a concerted attempt to consolidate its repressive machinery. The government established interdepartmental Joint Planning Centres in all regions, which were the precursors of the Joint Management Centres of the National Security Management System introduced in 1986. The main Joint Planning Centre established a Small Tactics Committee to investigate new methods of riot and crowd control. This committee, chaired by Maj-Gen De Swardt of the SAP, went all over the country to advise in "trouble spots". One of its outcomes was to recommend that all levels of SAP personnel should receive riot control training, in order that local police could cope with public order incidents without having to wait for a national police unit to be called. However, this recommendation was never fully implemented.
It was at this time, the mid-1980s, that a memo proposing the creation of a "third force" to deal with riot control was sent to the then-Chief of the Defence Force, General Constandt Viljoen; who passed it on the State Security Council for investigation. The idea of a separate 3rd force was rejected by the SAP; who set about establishing a separate "Riot Unit" at a national level; drawing on the people who had been trained at local level and asking for volunteers.
The creation of this separate, centralised "Riot Unit" was a critical moment in the organisation of public order policing in South Africa. It was controlled from the SAP national Headquarters, but was based in the provinces. The name of this Unit changed periodically - it was a sometimes called the Riot Control Unit or the Anti-Unrest Unit. During this period, the international comparative models from which the SAP drew were limited to those in repressive countries, friendly to the apartheid state.24
Throughout the 1980s, the homeland police forces had also been establishing some sort of specialised, separate riot control agencies, with approximately 30 units having been established by 1993. Although many of the homeland forces inherited personnel and procedures directly from the SAP, they were largely left to their own devices when it came to dealing with riots. Support from the SAP waned, and the SAP did not share its advantages in terms of international research and training with homeland forces. The riot units in the homelands were, as a result, largely less organised, less skilled, and even more brutal than those of the SAP. On occasion, when it appeared that homeland forces were unable to contain a particular incident, the SADF (not the SAP) were deployed to assist.
In 1990, SAP Lieutenant-General Du Toit met with the then-President De Klerk to propose a new investigation into the viability of the 3rd force option to deal with continuing levels of unrest. This led to the establishment of the Coetzee committee to investigate the proposal, which was rejected again. In its strategic planning for 1991, the SAP leadership realised that the repressive policing style which they had adopted in relation to unrest had "contaminated" the corporate image of the SAP and also reduced its capacity for crime prevention and control, because of the numbers of police officials required for dealing with unrest. As a result, it was decided to create a new National Division of the South African Police Service (SAPS) - the Internal Stability Division - to deal with unrest and to stabilise "racial conflict" in South Africa. The ISD would still be part of the police, but would be organisationally more autonomous than any of the other Divisions - it developed its own logistics, legal and training capacities. Its management was based in separate premises from SAPS Headquarters, and members drove different vehicles and wore a different uniform. As funds for the planned new uniform were not available, camouflage uniforms in government stock were used, with resulting public outcry. The ISD subsequently reverted to ordinary police uniform.
Forty Internal Stability Units, with over 7000 members, were operating throughout South Africa by 1994. One of the largest of these (with 1200 members) was "Unit 19", the special national unit which was based in Pretoria for rapid deployment to unrest focal points anywhere in the country. The remaining units were spread across the country, but concentrated around flashpoints for unrest in the PWV, Natal, Western and Eastern Cape. An additional thirty-seven similar units had been established in the various homeland police forces.
For the pre-election transitional period, a special force, known as the National Peacekeeping Force, was created to assist with the maintenance of Public Order. This report will not deal with the experience of the NPKF.
Methods for Crowd Control
Traditional public order policing philosophy in South Africa was based on the notion of "control". It was essentially a reactive doctrine, which entailed either preventing crowds from forming, or controlling crowds if they did form. Control was achieved through the threatened, or actual, application of force.
The traditional repressive approach to crowd control relied on the construction and maintenance of a suitable distance between the police and those in the crowd. Police personnel were heavily armed and reliant on armoured vehicles for their protection. Very few other protective measures were put in place. Thus, in instances where the distance between the police and the crowd was breached, the police would be faced with three tactical options: - to withdraw; to remain protected within the vehicles and do nothing else; or to use force to disperse the crowd and recreate the distance (or a combination of all three).
Within the context of militarized policing in South Africa, the riot control units were the "hard edge" of paramilitarism. They operated within a policy paradigm that accepted and supported the lethal use of force. This combined with the authorities' complete intolerance of protest action, meant that they frequently used maximum force on crowds and gatherings. As repression and resistance took on the character of a low-intensity civil war, their training, equipment and methodology became increasingly militarised, and conducive to the use of lethal force on a regular basis. The Kannemeyer Commission25 found that there was "in some police circles the prevailing view that birdshot and teargas were not effective enough and that as violence escalated, a decision was taken somewhere to use stronger measures". The Kannemeyer Commission found that the Riot Unit deployed at Langa carried no teargas, rubber bullets or birdshot, but were only armed with R1 rifles and shotguns loaded with SSG.
A police science training manual26 from the University of South Africa in use in the early 1990s still suggested the possible tactic of "[s]elective shooting aimed at eliminating leaders; a leaderless group is usually disorganised and the mob spirit evaporates" in order to control a crowd.
Police use of force did not manifest only in disproportionate use of weapons on crowds, but in all their behaviours in relation to political protest. In the early 1980s, the Catholic Bishops Conference catalogued the following types of unacceptable police behaviour:
- reckless or wanton violence (including the indiscriminate use of firearms and indiscriminate beatings);
- damage to property by the police;
- provocative, callous, or insensitive conduct by the police;
- indiscriminate or reckless use of teargas by the police;
- inappropriate police conduct at funerals (of unrest victims).
The South African police agencies were isolated from the international advances that were being made in the field of public order, and were unaware of the new strategies and techniques that other police services were implementing. They continued using militarised, "heavy" crowd control techniques long after such methods had been abandoned in Western democracies - "generally the police do not appear to have approached the unrest as 'temporary' riot behaviour by excited mobs, where the universal technique is simply to contain the crown and to calm it by a show of force, with personnel deployed as non-punitively as possible, Instead they have on many occasions appeared to act punitively against the crowds, probably on the assumption that the unrest is sufficiently focussed and goal directed to occur over and over again and to expand in scope unless immediate and painful disincentives are applied".27
Police Training for Riot Control
Specialised training of South African police officers for crowd and riot control took a variety of forms and underwent a number of changes after its introduction in 1975.
The first specialised training was conducted at a national level on a centralised basis at the SAP training base at Maleoskop, a farm in Mpumalanga.28 The first courses were introduced in 1975 to deal with the growing demand that crowd control was making on police resources.
The training was conducted on a 'train-the-trainer' basis and it was anticipated that the content of the courses would be devolved throughout the SAP by using the trainees to train other members around the country. The planned implementation of decentralised training was pre-empted by the Soweto uprisings in June 1976.
In the period 1975 to 1990, Maleoskop provided Counter-Insurgency training to police officers deployed in the specialised riot control units. The name of this training course later changed to "Internal Security" training. Training was provided for trainers who were stationed at units (for refresher and drilling purposes) as well as for the officers and members who made up these units. Maleoskop also provided training for units formed within the TBVC homelands. During the 1980s, all SAP recruits attended internal security training for two weeks after completing their basic training. The 1980s also saw the introduction of a seven-hour module (of an eight-week training course) in "unrest duty" for the Special Constables (known as Kitskonstabels) at the Koeberg training centre. The Internal Security training course for Riot Unit personnel took place over a period of three weeks and was buttressed by additional training for officers. The training courses were updated at various points and some in-service liaison and training did take place at unit level. However, by the 1980s, the SAP was hopelessly out of sync with international developments, and the only professional exchanges it made in this sphere were with similarly isolated, repressive states.29 Therefore the fundamental thinking underpinning the course was unchanged, and course content differed only insofar as new weapons were introduced or new tactics employed against protestors.
The content of the training course was broad, including everything from 'tracking terrorists' to sniper training. It focussed almost exclusively on the use of force, and the majority of training time was concentrated on training police officers in the use of the weapons and equipment they would be issued with in the riot units. In addition to weapons training, the trainees were given lessons on the legal framework related to gatherings and the enforcement provisions of these laws (with a particularly narrow interpretation of the laws). In short, the aim of the training was to ensure that a member would emerge from the course with the knowledge of what gatherings should be dispersed, when to disperse them and how to use weapons to disperse them.
In 1990, following the inquiry into the Sebokeng shootings of March that year, the period of training at Maloeskop was increased from three to six weeks and an improved system of local in-service training was introduced. In 1991, coinciding with the formation of the Internal Stability Division and the relatively more relaxed approach adopted by the De Klerk government to gatherings, a slightly more sophisticated version of the training courses (now called "Internal Stability") was introduced. This course focussed more on crowd control and began to introduce some 'soft skills' such as negotiation and crowd psychology.30 Unfortunately, the political slant contained in the content of these 'soft skills' detracted from any value they may have added. Various specialised courses such as video operation were also introduced for selected members. The main focus of the training, however, remained the use of force, and how to apply the law as it stood, rather than how the government during the political transition interpreted it.
In addition to the course at Maleoskop, the 'theory' of crowd control was taught in various university Police Science courses. These theoretical modules bore the same political bias and anachronistic attitudes as the practical training. Citations from UNISA Police Science study guides elsewhere in this report demonstrate their approach to crowd control.
Police Equipment for Riot Control
As stated in other sections of this report, the SAPS's approach to crowd control relied on the creation of a distance between police personnel and members of the crowd. In the absence of large numbers of personnel who are individually (through appropriate safety apparel) and collectively (through formations and reserve armour) protected, this 'distance' could only be created and maintained through the retention of a highly potent offensive capability. Until 1994 personnel were almost entirely reliant on armoured vehicles for protection (and therefore rarely ever operated on foot) and were heavily armed - it was not uncommon for personnel to carry a variety of automatic weapons in addition to shotguns and handguns as standard issue during the policing of public collective action.
The South African Police, when asked to defend the type of offensive equipment they carried, were often at pains to distinguish between lethal and non-lethal weaponry. The experience of many Africans, however, was that depending on how so-called non-lethal weapons were employed, this distinction applied only in theory. The truth is that almost every piece of equipment carried by specialised units could, and did, kill, if employed in an inappropriate fashion. An armoured vehicle is a lethal weapon when driven into a crowd. A baton, used with intent against the head or several other vital organs, can cause death with a single blow. Teargas inhaled within a closed environment (especially by children and frail persons) will cause death by asphyxiation and, when fired in canister form, can kill on impact at short range. Almost all the forms of ammunition used in shotguns will kill if fired at vulnerable body parts at short range or if 'doctored' to inflict more damage.31
Although the equipment used by specialised riot control units changed over the years, the changes were mainly for the purpose of sophistication. In other words, different implements were updated or changed but there was always a weapon of some sort that fulfilled a similar function. For example, if police charged a crowd they would, over the years, have used wooden sticks, truncheons, sjamboks, plastic quirts and plastic batons, but their purpose remained fundamentally similar.
The list below includes the main equipment used by police in the 1980s.32 The list does not include some of the more expensive, but rarely used pieces such as water cannons and 'sneeze-machines'.
the TONFA - a plastic baton/rod approximately 50cm long and 30mm in diameter;
teargas - used to incapacitate people or disperse them from a particular location. it is either expelled from an aerosol can at short distances or fired in the form of a 37mm round from a special gun, or by a 12 bore grenade launching round from a Musler shotgun;
the 'stopper gun' - a short, stubby, pipe-like 37mm gun used specifically for firing tear-gas rounds or large rubber bullets. It has a variable charge thus allowing the gunner to adjust his/her range. 37mm Stopper rounds are capable of causing severe injury at short range;
12 bore Musler shot-guns - used to fire a range of ammunition. The ammunition is colour coded and includes:33
training ammunition (yellow plastic covering);
lessened 12 bore double ball rubber - reduced to enable firing at close range during unrest situations (pink plastic covering);
standard 12 bore double bore rubber - potentially lethal at close range but used to inflict bruising and pain during dispersals (blue plastic covering);
No 5 shotgun ammunition containing between 270 and 280 grains per round. The pellets have little penetrative capacity over 40m although they can still cause serious injuries - indeed the plastic covering itself is potentially lethal at short range (green plastic covering);
AAA ('Triple A') shotgun ammunition containing approximately 40 grains per round - should only be used in self defence since the shot may be lethal at up to 25m (red plastic covering);
SSG shotgun ammunition containing approximately 18 grains per round - is a full strength shotgun round that is lethal and should only be used for protection (black plastic covering).
Uzi hand machine-carbine (9mm - either single or automatic shot with effective range of 100m);
R1 Rifle (7,62mm x 51mm - either single or automatic shot with effective range of approximately 150m);
R5 Rifle (5,56mm x 45mm hypervelocity [tumbling bullet] - either single or automatic shot with effective range of 150m);
LMG - vehicle mounted light machine gun (7,62 x 51mm - only automatic shot, fast firing capacity and effective to 200m) - rarely ever used in Africa by police;
Mortars - these were not carried by police but as late as 1994 Internal Stability Division units still held them as part of their arsenal;
Razor wire - used to prevent access by a crowd across a police line or to an identified 'key-point' - quickly re-usable when deployed using a winch situated on a specially adapted trailer towed behind an armoured vehicle;
Personal safety apparel - including bullet-proof jacket, fire extinguisher, gas mask and helmet;
Grenades - explosive, smoke and stun grenades were carried by Internal Stability Division members - mainly used to effect hostile house entries but stun grenades can be thrown or shot at crowds; and
Mine resistant armoured vehicles - the Casspir and Nyala armoured vehicles were used to carry and protect personnel, move or remove obstructions and barricades, and carry and store equipment.
Reforms to Public Order Policing in the Post-1994 Period
In April 1993, the Goldstone Commission proposed a Draft Bill to address the civil liability of organisers of gatherings, the prevention and prohibition of gatherings, and demonstrations near courts, Parliament and the Union Buildings. The Commission said that legislation was desirable even before the elections, because mass demonstrations and marches were matters of urgency. Passed in 1993, the Goldstone Bill became the Regulation of Gatherings Act, but it was only enacted after the election (1996.11.15).
This Act provides a new legal framework, which entrenches the right to peaceful public expression and peaceful assembly and the right to State protection in the enjoyment of these rights. It sets down a new procedure for the management of public gatherings, of which a cornerstone is the appointment of a "convenor" of the gathering who must notify a responsible officer in the local authority that a gathering is planned. Negotiations may then be instituted with other interested parties, notably the police, to ensure the peaceful progress of the gathering and the protection of both participants and non-participants. The circumstances under which dispersal and use of force may occur are laid down in the Act. Provision is made for civil liability and for other penalties for infringement of regulations and agreements.
The first Minister of Safety and Security in the democratically elected government faced a Constitutional requirement to create a national Public Order Policing unit as part of a new, integrated South African Police Service. To ensure that this requirement was met, and to effect a speedy process of reform to the policing of crowds, the new Police management established a Technical Committee on Public Order policing in April 1995. Its brief was to conduct research and generate proposals for: the transformation of the existing Internal Stability Division and the Riot Control Units of the various other police agencies in South Africa; and the introduction of a new system of public order policing based on the interim Constitution, the Minister's Policy Document, and the draft South African Police Services Act. The committee completed its work and made a set of policy and organisational recommendations in July 1995.
The new policy moved away from "crowd control" to "crowd management" as a conceptual framework for the role of the police with respect to public gatherings. It was designed to complement the new legal framework, and to take into account the rights of all citizens. The new crowd management policy:
developed a clear set of POP goals applicable to public order situations;
established a clear set of principles pertaining to the management of crowds including -
- the legal aspects of crowd management,
- the situational appropriateness of public order strategies and tactics,
- the optimum utilisation of suitable means and available resources, and
- the proportionality of the response to actions taken by participants of public collective actions;
- emphasised the importance of preparation for crowd management operations;
clarified levels of responsibility, command and control;
spelt out detailed pre-planning steps;
introduced planning and operational committees;
outlined the appropriate manner by which operations should be executed (including methods, use of force, media relations, and SANDF liaison); and
expanded on the role and functions of debriefing.
The National Commissioner and the Minister of Safety and Security approved the new crowd management policy in September of 1995. Implementation of the new policy approach, and the accompanying structural reforms, began immediately.
During November 1995, the ISD was amalgamated with the Riot Control Units of the homeland police agencies as part of the amalgamation of the police agencies and the creation of the new South African Police Service. All ISD and Riot Control Unit personnel (approximately 9 000 in total, distributed among 72 units) and equipment were placed under the command and control of the Divisional Head, National Safety Services of the SAPS. The day-to-day deployment of members and the administration of existing units by Provincial Commissioners was maintained, subject to orders from the National Commissioner that all existing units would be kept in place, and that no members, vehicles or equipment could be assigned to any other component until the implementation of the new policy was complete.
These amalgamation and status-quo orders succeeded in consolidating the riot control capacity of the Government of National Unity for the interim period until such time as a new Public Order Police component could be established. The enactment of section 17 of the South African Police Service Act 68 of 1995 provided the final, full, legislative confirmation of this amalgamation, and officially created the Public Order Policing component of the new SAPS. In 1997, the Division within which Public Order is located was re-named "Crime Prevention and Reaction Services".
Possible Recommendations by the TRC
The TRC could make the following recommendations in respect of the policing of crowds in South Africa in future:
ensure that all members of the Police Service dealing with public order incidents have received the required training and that training standards with regard to the policing of crowds are maintained to international standards in future (increased training will lead to greater confidence on the part of those policing crowds and gatherings, which should result in less reliance on force);
decentralisation and integration of public order policing into the normal structures of the SAPS (as per the new policy) should be pursued. The tendency to centralise and militarise this function should be avoided in future. This will entail the reduction in size of the specialised POP units over time;
Public education (and training of local authorities) as to the Regulation of Gatherings Act, and rights and responsibilities in respect of gatherings, should be continued and expanded, especially for elections, public holidays and other regular public events;
prosecutions of any person or party who contravenes the Regulation of Gatherings Act should be pursued to ensure compliance with new approach;
the police should establish an internal inspectorate at POP Headquarters (or in the Secretariat for Safety and Security) and at Provincial level to check adherence to new national standards, and to investigate the policing of particular incidents;
the police should be issued with new equipment and apparel, which will improve their safety and protection - the more protected the police official feels, the less likely they are to use force or act aggressively;
the police should enhance internal procedures to ensure accountability for the storage and use of weapons and ammunition. Post-incident investigation of all cases where force was used should be standard procedure;
the police should create, equip and train camera teams, who would film every incident where the police are deployed for crowd management. This would be a useful device for improved accountability, as well as for debriefing and training of the police officials themselves;
the legislation concerning the carrying of weapons in crowds should be fully enforced at all times in future;
the public order policing component of the SAPS should be encouraged to implement a meaningful affirmative action strategy to improve representivity at the senior levels of the component;
case studies of past incidents of bad practice by the former SAP and homeland police in the policing of crowds and gatherings should continue to be used in POP training materials, to ensure that future public order police officials are aware of the bloody history of their unit, and that such practices are never again tolerated by the police service or the public.
2 para 6.8.2 University of South Africa (UNISA) Police Science 201 Study Guide.
5 An early reference to this approach is found in an article by Simon Baynham in Indicator SA 1985. It was also referred to by former State President F W De Klerk in the National Party submission to the TRC.
6 3rd Force here refers to a "force" which is established between the Army (the 1st force) and the Police (the 2nd force). It does not carry the same meanings as the notion of a "3rd force" which developed currency in South Africa in the 1990s, which tends to imply covert use of violence.
10 "The police do not disperse each and every gathering that takes place outdoors in South Africa, but when patrolling the townships they may select and disperse with force such gatherings as they choose. In this light the right to use firearms to disperse gatherings is converted into a power to enforce, by firearms, an informal indoor curfew by day and night." - N Haysom, Licence to Kill Part 1: The South African Police and the Use of Deadly Force 1989 South African Journal on Human Rights (SAJHR) 3 on p 23.
13 The SAP would seek to justify their emphasis on offensive capacity by virtue of: "the proportion of rioters to police; the terrain; the climate; and the degree of violence by rioters". See interview with Major-General Bert Wandrag, Chief of the special Counter-Insurgency Unit, in Indicator SA Vol 3 No 2 Spring 1985.
15 For further information on the way in which state-controlled media fostered a positive impression of police action 'in response to aggressive crowd behaviour' see Simpson, G and Vogelman, "The Myths of Mobs and Marches" Sunday Star 1 October 1989, p. 35.
16 Informal discussions between the authors of this report and a senior policemen who served in riot control units between 1976 and 1998 reveal that operational commanders would often encourage brutality and promise indemnity. Also see a telex from the Senior Deputy Commissioner of Police on 19 March 1985 stating that "When acid or petrol bombs are thrown at police or private vehicles and/or buildings an attempt must be made under all circumstances to eliminate the suspects" found in N Haysom, Licence to Kill Part 1: The South African Police and the Use of Deadly Force 1989 SAJHR 3 on p 26.
17 Note that this equates to a considerable lowering of the training and authority of the person allowed to give an order to disperse relative to the Riotous Assemblies Act (Head Constable). The authority to disperse would be further lowered in declared 'unrest areas' under Emergency provisions.
18 These provisions enabled a more junior-ranking officer to give to give the order to disperse; required that the warning be given only once, and no longer required the officer to warn that force might be used in dispersing the gathering.
20 see "Memorandum by Policing Bodies to the Goldstone Commission of Inquiry, in Connection with the Handling of Gatherings and Processions, with regard to the press statement released by the Honourable Mr Justice R.J. Goldstone, Chairman of the Commission of Inquiry into the Prevention of Public Violence and Intimidation" (undated).
22 Jeffery, AJ, Riot Policing in Perspective South African Institute of Race Relations, Johannesburg, 1991, p. 120.
28 Before 1975 the training base was used for counter-insurgency training that equipped police officers for deployment in the Rhodesian and South West African wars. It continued to give such training but since 1975 its primary focus was on 'internal security'. Other SAP training bases that provided crowd or riot control training, or aspects thereof (for example SWAT training to reaction force members) include Verdrag and Jakkalsdans.
30 This focus on psychology coincided with the introduction of psychometric testing of recruits to the Internal Stability Division (ISD). The testing was understood by commentators to have been initiated as a result of the severe judicial criticism received during inquiries into shooting incidents. However, conversations between the authors and members of the SAP's Institute for Behavioural Sciences seem to indicate that it had more to do with the concern in police management circles that the paramilitary divisions may be producing criminals, after publicity around the fact that a notorious Johannesburg serial rapist was, in fact, a serving member of the ISD.
31 It is well known within police circles that SAP members would alter ammunition in order to inflict more damage on their victims. The best known example of this was the practice of placing batteries inside rubber stoppers, thus raising the likelihood of lethal consequences if they hit a victim on the head.
33 For further information see Moolman, T "Standardisation of Shotgun Ammunition" in Servamus April 1998.
© Centre for the Study of Violence and Reconciliation