Simpson, G. & Rauch, J. (1993). Political Violence: 1991. In Boister, N. & Ferguson- Brown, K.(eds), Human Rights Yearbook 1992, First Edition. Cape Town: Oxford University Press.
Graeme Simpson & Janine Rauch
In Boister, N. & Ferguson-Brown, K. (eds), Human Rights Yearbook 1992, First Edition, pp. 212-239. Cape Town: Oxford University Press, 1993.
Graeme Simpson is a founder and former Executive Director of the Centre for the Study of Violence and Reconciliation.
Janine Rauch is an independent consultant.
Introduction: The deregulation of social control
The year 1991 witnessed the continuation of trends in political violence which had been set in motion in the wake of the State President's speech in February of that preceding year.1 The redefinition of the boundaries of the political playing field implicit in the supposed dismantling of legislative apartheid was presented to the international community as the essential prerequisite for the reconstruction of the 'New South Africa'. However, the deconstruction of formal apartheid simultaneously implied the deregulation of repressive forms of social control at precisely the time at which the process of political transition heralded an intensified political contest, in which the key political players jockeyed for position in preparation for anticipated negotiations.
The unshackling of the political process thus had the unintended consequence of 'deregulating the existing mechanisms of social control' without effectively replacing them with consensus-based alternatives. The highly repressive means of control which resulted in a substantially over-regulated society prior to 1990 were necessarily withdrawn to facilitate the envisaged process of negotiation. This served to erode existing perceptions of authority (whether legitimate or not) and rendered the institutions historically responsible for implementing those mechanisms of social control even less effective than had previously been the case. In particular, the state's security forces, confronting their own politically-orientated legacy, were rendered ineffective and incapable of maintaining their authority (in the absence of repressive controls) within the changing political terrain. The historical resistance to these coercive mechanisms of social regulation, their inability to adapt to new social demands, and the failure of any alternative (democratic) regulatory authority mechanisms, has resulted in high levels of social, political, economic and ideological dislocation within the society.
We refer to this process as the 'deregulation of social control' in the post-1990 period. Simply put, this amounts to the piecemeal deconstruction of social control under apartheid, without the immediate generation of viable alternatives. This process operates both institutionally and ideologically, serving to undermine historically dominant ideological and institutional authority within the society. It also operates materially, as competing claims to limited resources within the society become less constrained by the regulatory and repressive mechanisms, which have historically entrenched limitations on people's expectations, social mobility and control over resources.
This process of social deregulation, rather than being entirely new and signifying a dramatic shift in the post-1990 period, corresponds closely to the trajectory of historical trends set in motion in the preceding decade. Repressive controls during the 1980s could not disguise the extent to which resistance to apartheid had destroyed any pretensions of legitimacy of the system and the institutional pillars upon which it rested. In the process, bread and butter issues such as education, housing, the delivery of services, wages, etc. all became politicised and economic impoverishment became all too explicitly etched in racial terms. De Klerk's reform initiatives of February 1990 simply threw open the doors on a tide already on the move.
Our view is that this 'deregulation' is almost inevitably associated with such a transitionary period, heralded by rather haphazard negotiations. The very process of transition demands that historical mechanisms of control in this disintegrating yet over-regulated and (politically) over-policed society, be dismantled to facilitate the change. Yet the alternative forms of consensus-based regulatory institutions and ideological formations (such as new nationhood or national identity) are slow in coming, and are, indeed, the very subject of much of the negotiation process. The material expectations of political transformation, rooted in the historic politicisation of 'bread and butter' issues, are also largely beyond the delivery capacity of a society prior to its reconstruction. The result is an inevitable 'window period' in which social control is utterly 'deregulated'.
These are trends which characterise transitional processes in changing societies around the world. However, in the South African context, this dramatic social insecurity, bred of a sense of transitional disintegration, has articulated particularly destructively with the racial, class-based and ethnic identities and hostile stereotypes generated or reinforced by decades of apartheid. Furthermore, the historically socially-sanctioned resort to violence as a means of both maintaining political power (on the part of the NP government) and as a means of resistance (on the part of the liberation movements) served to embroider the entire political culture with violence as a means of resolving conflict in the context of the simultaneous politicisation of a wide range of social, political, criminal and domestic violence in the society. This has been accompanied by desperate, yet often sluggish and politically constrained attempts to negotiate legitimate interim mechanisms of social control. The consequence has been with the coincidence of dramatic levels of social and political violence with the haphazard process of negotiation.
It is evident that one of the central problems which plagues any analysis of political violence in South Africa is the competing 'labels' through which the violence is described. This problem of 'labelling' is especially evident in the rhetoric of the various political interest groups themselves, as the political contest is waged in terms of who is more responsible for the escalating political violence.
This places great emphasis on control over ideological apparatuses, and nowhere is the effect of such 'labelling' more evident than in the role of the media. Through shaping popular perceptions of causation and explanation, the commercial media - especially in an environment of suspicion and distrust in which rumour easily becomes rampant - arguably plays an important role in shaping the trajectory of the violence itself. The selective choice of labels by the press is itself seldom politically untainted. More importantly, commercial media shorthand often dictates a search for (readable) mono-causal explanation. This is, itself, often reflective of the dominant ideological discourse.2
In this regard, it is instructive to note the shifting terms in which political violence has been described, particularly by the vehicles of the media ministering to the white population. Initially, descriptions of 'black on black' violence were largely functional to the NP's concern to both distance itself from any responsibility for growing conflict, while simultaneously pointing to the ANC's inability to control its support base. This gradually developed into descriptions of 'ethnic conflict' between Zulus and Xhosas, reflective of the shifting concerns to both maintain government's distance, whilst establishing the Inkatha Freedom Party's (IFP) indispensability to the negotiation process. Indeed, whether wittingly or otherwise, this particular approach serviced, very directly, the specific strategy of ethnic mobilisation which lay at the root of the IFP's political agenda.
In addition, multi-lateral attempts at dealing with the violence, as well as the central negotiation process itself, have served to politicise most forms of violence even further. This trend has developed out of the dual dynamics of criminalisation of political violence and the politicisation of crime during the 1980s.3 Violence which has at its roots competition over material resources, such as conflicts within township communities or between permanent urban residents and migrants, is increasingly labelled as ANC/Inkatha or Xhosa/Zulu conflict even when political affiliation or ethnic identity are only peripheral to the conflicts. It has long been recognised that much of the violence, explained and construed as political, is frequently criminally, rather than politically, motivated. All the political players share responsibility for this trend, as point-scoring in the crucial run-up to the first election is increasingly based on responsibility for, or victimisation in, the series of violent events.
The overall effect of such discursive descriptions has also been to entrench many of the fears and insecurities of white South Africans. These very fears and insecurities have traditionally serviced right-wing interests and dominated the ideological discourse within white politics. This has played an important role in shaping the terms of the negotiation debates around such issues as the constitutional entrenchment of ethnically defined minority rights and vetoes.
It is not possible here to document, analyse, or even fully describe political violence in South Africa in all its complexity. What should become clear, however, is that the search for mono-causal explanation is fruitless. The convenient terms in which the violence has been labelled by politicians and the commercial media often do more to disguise complex causation than to explain it. The violence has been variously labelled as ethnic conflict, conflict between hostel dwellers and squatters or township residents, conflict between ANC and IFP supporters, conflict between the police and the residents, conflict between the poor and the very poor, conflict generated by government or by a 'third force', etc. None of these descriptions is completely inaccurate. Yet none, on its own, will properly explain this complex situation. It is only when we accept that we are dealing with a host of overlapping causal factors that we can begin to understand or address the problems constructively.
Furthermore, the breakdown of mechanisms of social control outlined above is clearly not limited in its effects to the formal political arena. In reviewing political violence in 1991, this fact demands that we attempt to disaggregate the resultant violence, distinguishing between different forms and types of violence in order to more effectively trace the roots thereof. However, the independent momentum and dynamic of violence makes this extremely difficult and imposes some critical definitional problems. In particular, the easy slide between political and criminal violence and the dramatic increase in both the quantitative and qualitative levels of violence in South Africa during this period tend to make nonsense of any reliance on statistics as a means of understanding the carnage. The problems posed by 'labelling' are also visible in the differing interpretations of the statistics which are generated in order to explain political violence.
In 1991, according to the SAP Crime Investigation Department, there were 14 693 reported murders countrywide, 22 765 reported rapes, 129 103 reported incidents of serious assault and 45 103 armed robberies.4 However, according to the Commissioner of Police, General van der Merwe, of the 11 764 murders committed between January and August 1991, only 6,8 per cent (806) were 'unrest-related'.5 The truth or otherwise of these claims remains entirely unverifiable and this reflects a serious problem with regard to the lack of information necessary to draw a distinct line between various forms and definitions of violence. At most, it is possible to point to the complex interaction of different forms of violence. This was hinted at by Duma Nkosi, secretary-general of the ANC Thokoza branch, who was quoted as saying:
Political violence and crime sleeps under one blanket. It's easy for anyone to employ criminals to carry out acts of violence for political motives. The desire is for our people to be governed by fear, to become inactive and so not participate in the peace process.6
A further problem with pure statistical descriptions of political violence is the generally arbitrary or highly subjective means of categorising the figures. The monthly or annual body counts of victims of political violence seldom (if ever) offer an accurate prism on causation and police-generated definitions of 'unrest-related' deaths or injuries are frequently ill-conceived.7 Even the political affiliation of victims and/or perpetrators is seldom a conclusive indicator of motive in a society dominated by high levels and arbitrary patterns of violent crime.
A final problem with statistical descriptions of violence is that, despite claims of objectivity, most of the statistics manufacturers are themselves either alleged to be pursuing particular political agendas or are at very least perceived in this way. Some of the producers of information around violence, such as the South African Police, are also perceived as actual protagonists in the violent political battles being described.
In the final analysis, the vast range of incidents including clashes between rival political groups, ethnic conflicts, train killings, attacks by and on hostel dwellers, bombings, stabbings, shootings, etc. and the simple arbitrariness of much of the violence leaves us with limited intelligible information about the selection of victims and a great deal of ambiguity about the specific motives behind many of the particular incidents. The body-count statistics would seem, therefore, to offer us very little.
In view of the problems associated with these statistics and their analyses, the breakdown that follows should be treated as descriptive rather than analytical.8
According to the Human Rights Committee (HRC), there were 2 582 'politically-related' deaths in the course of 1991. Of these, 77,9 per cent were allegedly 'vigilante-related', 3,9 per cent resulted from 'security force action', 2,3 per cent from 'hit squad actions' and 0,8 per cent from 'right wing actions'. Of these deaths 15,1 per cent were classified as being of uncertain origin.9
Broken down regionally, 46,4 per cent of these deaths occurred in the PWV region (approximately 100 per month) mostly in the East Rand and Soweto. A further 40,9 per cent of the deaths occurred in Natal, which averaged 88 deaths per month.10
According to the South African Police (SAP), 2 239 people were killed in 'unrest-related' incidents in 1991. Of these, 63 were killed by 'security forces', 2 121 were killed by 'riotous people and others' and 55 members of the security forces were killed whilst on duty.11 A police report on the Reef violence, covering an 18 month period from February 1990, blamed the ANC for 86 per cent of attacks and the IFP for 12 per cent.12
According to the South African Institute for Race Relations (SAIRR), 'political violence' during 1991 claimed 2 672 lives, approximately 28 per cent lower than the 3 699 killed in the preceding year and bringing the total number of fatalities since September 1984 to approximately 11 910.13 The SAIRR analysed 1 236 violent incidents countrywide in the first eight months of the year and concluded that in 85 per cent of the cases the aggressors could not be identified. The weapons used in the incidents were firearms (30 per cent), explosives and incendiary devices (25 per cent), instruments capable of inflicting hack and stab wounds (16 per cent) and stones (15 per cent). Common assault and burnings accounted for 8 per cent of the incidents and in 6 per cent of the incidents it remained unclear what types of weapons were used.14
According to the Community Agency for Social Enquiry (Case),15 research based on media accounts, in the 12 months to August 1991, out of 601 reported instances of violence, in which responsibility was attributed by the media in 257 of the instances, Inkatha was reported as responsible for 51 per cent of the acts of violence; the SAP for 23 per cent and the ANC for only 4 per cent. In that period, a total of 2 271 people were reported as killed, 87 per cent of whom were 'general members of the community'. In addition, in 46 of the 257 confirmed reports, 'direct collusion between members of the SAP and IFP supporters' was reported. On the question of identifying the victims of this political violence, the Case report claims that:
the target of violence on the Reef has not mainly been the ANC, rather, the broad black community have suffered massive casualties … as they were attacked at home, travelling, or in groups such as at vigils, in beerhalls or waiting at taxi ranks.16
The 'unrest-related deaths' statistics as supplied by the HRC, the SAIRR, Case and the SAP, although differing on the actual numbers killed in any particular month or year, present remarkably similar trends in terms of the month on month increases throughout 1990 and 1991. Although usually backed by different analyses, in each case the high points of political violence as indicated by the numbers of deaths roughly coincide with key developments in the peace process or in the development of either local-level or national political negotiations.17 This simple fact serves to establish the key agenda for interpreting political violence in 1991 - the negotiation context.
The negotation context
According to Everatt and Sadek:
The Reef violence is a product of political transition in South Africa. It began in July 1990, as the ANC suspended armed struggle, was drawn into negotiations and began to create a legal infrastructure inside South Africa. Since then, the violence has provided a constant shadow to the negotiations process. Its peaks and troughs mirror the waxing and waning fortunes of the National Party government, as it seeks to negotiate a future South Africa which retains minority control of economic and political power.18
In terms of this general analysis, they go on to correlate the high points of violence with key developments in the negotiating process as follows:19
In March 1990, (approximately 460 deaths, 360 of them in Natal) immediately after the release of Nelson Mandela and following two 'anti-sanctions' rallies held by Inkatha and funded from SAP sources.
The explosion of the 'Transvaal War' in August 1990 (over 670 deaths) coincided with the signing of the Pretoria Minute - the first clear step in the negotiation process.
November 1990 (approximately 310 deaths), immediately preceding the ANC's first conference inside the country since the organisation's unbanning.
Between March (approximately 350 deaths) and May of 1991 (approximately 310 deaths), the period in which indemnity was granted to ANC guerillas and for all political crimes.
In September 1991 (approximately 280 deaths), immediately prior to the drafting and signing of the National Peace Accord.
November 1991 (approximately 275 deaths), in the month leading up to Codesa I.
March 1992 (approximately 215 deaths), in the month of the white referendum.
On the basis of this outline, Everatt and Sadek claim that the violence peaks when it is most likely to damage the ANC and dies down most dramatically when it threatens to harm the NP government.20 Thus they claim that not only can the high points be linked to key developments in the negotiation process, but low points of violence must also be linked to the specific needs of the NP government at those times. In particular, they suggest that the low levels of violence (as measured by the number of people killed) must be linked to de Klerk's visit to Denmark and Ireland in October 1990, to the ANC/IFP accord in which Mandela was finally pressurised into meeting with Buthelezi in February 1991, and to the opening of parliament in February 1991. Everatt and Sadek conclude that the pattern is visible throughout 1991 and into 1992 and that the 'violence appears to be switched on and off at strategic moments.'21
In analysing this, they argue that the Reef violence appears to represent the
domestication of low intensity conflict which the South African government perpetrated in Mozambique, Angola and Namibia. As the formal apparatus of South African repression has been dismantled, its function - to hamstring and cripple political opposition - has been carried out by the violence.22
A similar analytical viewpoint was adopted by Simpson et al in their review of political violence in 1990,23 where they describe the unbanning of the ANC, PAC and SACP in February 1990 as the beginning of a process of 'internal militarisation' and the transformation of the liberation movements from an 'external enemy' to an 'enemy within'. There it was argued that this shift in state strategy was reflected by the decrease in defence spending and the parallel increase in funding to the SAP - the internal arm of the repressive state apparatus.
However, it is difficult to make these assessments based on 'unrest' death statistics. At no point in 1991 did the death toll reach the dramatic proportions of the 600 dead in August 1990 nor even the 350 dead in the month following that. Yet the cumulative picture of 1991 was of regular and almost perpetual killings with only the respite of occasional decreases, never a halt, as the country crawled towards a promise of negotiated transition - supposedly heralded by Codesa I - in December 1991. Throughout the year, these trends not only continued, but were slowly shifting and developing, as they became more and more intrinsic to the climate of political uncertainty and as the dividing line between criminal and political violence became less and less distinguishable.
The simple resort to body counts would suggest that 1991, on the whole, was considerably less violent than the previous year. Yet the consistency of the violence, its increasingly arbitrary and occasionally random selection of victims, the increasing levels of brutality and the pervasive influence of rumour, all reflect the intensified rather than decreased levels of destabilisation of the political environment during 1991. The violence had, in the course of the year, developed a momentum and dynamic of its own. Thus, whilst it is clearly possible to argue that disruption of the ANC may have been a comprehensive motive for 'switching on' the violence, it is less easily arguable that the ANC's opponents' objectives were served at other times by simply 'switching it off'. Indeed, it seems clear that, despite the regular allegations by government and Inkatha that the ANC could not control its support base (an argument which at times cannot be denied), the same is unquestionably true of both the government and Inkatha. To suggest that they simply control the process which they may well have set in motion, if anything, allows them to escape liability for provoking a separate but interwoven process of destabilisation which they actually could not control.
The independent momentum of the violence is fostered by the mobilisation of defensive ethnic identities and by the survival of an organisational culture within the state security forces in which covert activities for political goals were sanctioned. Recognition of this independent momentum therefore demands a sophistication in the analysis of political violence which goes beyond merely correlating the body counts with key points in the negotiation process.
This is not to suggest an accidental coincidence between escalation of violent attacks and these key developments in the negotiation process although the precise relationship between increased deaths and such events as the ANC's first conference inside the country or the granting of indemnity for political offences seems to be assumed rather than explained. Rather, it is argued here that to properly understand the growing climate of fear, as well as its effect on the negotiation timetable, it is necessary to evaluate the escalating violence in a context wider than that framed by simple labels of the leading political interest groups. It is also argued below that it is precisely because attempts to resolve the violence operate primarily in the formal political sphere that the achievements of initiatives such as the Peace Accord are severely limited.
The urban conflicts of the 1990s are materially and ideologically rooted in apartheid, its deconstruction, and the consequent breakdown of authority and social regulation. In part, the consequence of this is an intensified power struggle between the competing political interest groups, but the destructive might of these political conflicts can only be fully explained through a complex understanding of their articulation with the wider interests of an impoverished and oppressed urban population - on the trains and taxis, in the hostels, the squatter camps and in the townships themselves. In examining the central protagonists, incidents and contexts of violence in 1991, it is therefore to these wider interests that we must attempt to give more attention.
Nowhere were the trends and consequences of deregulation of social control better illustrated than in the critical area of policing of the transitional process in 1991.24 While police on the street experienced the whittling away of their traditional power, the half-hearted and politically clumsy attempts at reform within the force failed to generate any confidence-inspiring structures of accountability which could serve as an alternative source of authority.25 This proved to be central to the spiral of violent crime and informal retribution as people, distrustful of the SAP, increasingly took the law into their own hands to remedy the wrongs they had suffered - whether political in origin or not.
This was the inevitable result of the political legacy of apartheid policing that had, during the preceding decades, placed the law enforcement agencies of the apartheid state at the centre of political control and struggle. The post-February 1990 shift in the locus of repressive state control from the SADF to the SAP, coupled with the sinister spectre of the NP's strategy of internal destabilisation, maintained the centrality of the security agencies as the primary subjects as well as the central objects of the negotiation process.
In the course of restructuring of the SAP, in April 1991, the notorious Security Branch was officially 'disbanded' in an attempt to demonstrate a shift in the focus of police action. However, this disbanding and the increased challenge to the authority and control of the 'security forces' did not mean that tried and tested vehicles of state violence were simply abandoned during 1991.
The majority of reports of police abuse of power revolved not around the use of excessive force or the violations of human rights per se, but centred on the claims of security force bias in the policing of 'political conflict'. Examining the role of the police in the escalating township conflicts, Case claimed that the SAP was an 'identifiable aggressor' in 23 per cent of the reported instances between July 1990 and August 1991. This included noteworthy allegations of police complicity and both passive and active support for IFP attacks, many of which were made prior to the 'Inkathagate' revelations of SAP funding for Inkatha and its trade union wing Uwusa.26
The crisis of police credibility, deepened by the allegations of collusion in violence which accompanied virtually every major incident during the year, was further exacerbated by the revelations in The Weekly Mail proving police support of Inkatha rallies well after the reform process has ostensibly begun. Furthermore, television viewers across the world bore witness to Cape Riot policemen saying that they could not accept the political changes and saw them drunkenly boast on a BBC documentary about how many people they had killed.27 SABC television viewers also saw an SAP member joke with armed Inkatha supporters about 'killing Mandela'.
In this context, it is hardly surprising that policemen themselves became targets of the political violence. In 1990, 107 policemen were killed on duty and in 1991, a further 137 were killed on duty countrywide,28 55 of those in 'unrest-related' incidents.29
On May 9, ANC President Nelson Mandela issued an ultimatum to de Klerk's government linking the continuation of the negotiation process to the dismissals of the Minister of Law and Order, Adriaan Vlok, and the Minister of Defence, Magnus Malan. Within a few months, both lost their jobs as part of the process of securing the bumpy road to Codesa I. However, their successors were to be no less controversial.
The policing of the violence must be understood within the context of rapid political change, the enormous challenges of transformation which confronted the SAP, and its recent history of counter-insurgency warfare. In the process of change, the SAP confronted the legacy of its own historically established organisational culture - politically sanctioned covert activities geared towards the suppression of popular support for political opponents of the apartheid government. This resilient organisational culture, in the absence of any effective multi-party control over the 'law enforcement agencies', meant that the operations of these state apparatuses were, at very best, beyond the control of government and, at worst, continuing to pursue a long-established political agenda.
Furthermore, the 'ethnic' or 'cultural' explanations of the violence critiqued above have also tended to mask the role of effective law enforcement as a potential solution to the problems.30 The police subculture has tended to perpetuate these sorts of explanations and this is, in large part, responsible for police inactivity and inefficiency in pursuing cases involving black victims.
The SAP's internal reform programme that developed in 1991 out of the strategic planning process, implemented after de Klerk's political initiative in 1990, was complemented by large sections of the National Peace Accord which attempted to create a new framework for police conduct.31 However, despite a new rhetoric of reform that emphasised the improvement of police-community relations and professionalisation of the police service, Amnesty International concluded their investigations thus:
The circumstances and views encountered by Amnesty International's representatives during their visit to South Africa in December 1991 left them with an impression of the enormity of the gap between the intentions and pronouncements by officials, on the one hand, and the conduct on the ground of the security forces, on the other.32
Violence and the struggle for power
The increase in violence between competing political interest groups in the black communities must be seen both in the context of the ongoing battle for political support heralded by the shift to negotiation as the means of transition, and in terms of the alienation which this has produced between the leadership and membership of various political parties.33 This has been particularly pronounced because of the irregular and unstable national negotiation process which results in confusion and disenchantment on the ground and in turn increases the potential for violence.
Each of the parties struggles with internally contradictory approaches to the violence. On one hand, the leadership is engaging in public negotiation processes aimed at ending the violence (such as the National Peace Accord and Codesa). On the other hand, the membership base, which is excluded from this process, often views peace as practically unattainable while the competing political forces exist and organise in their areas, and for as long as the police are perceived to be partisan and incompetent. If the solution to violence is political, then the logical conclusion is that to achieve peace, political supremacy must be attained, particularly at the local level. Thus, the negotiation process begins to be perceived merely as a vehicle for working out the details of the transfer of power, while local political struggles, which may involve violence, become a means of attaining meaningful political control.
The key factor in the ANC's response to the violence in 1991 was its interpretation of the violence as a concerted political strategy to weaken the organisation. This interpretation determined that the ANC's response to violence would invariably be political. While there is no doubt that many of the solutions to the problem of violence do lie in the formal political process, this type of reaction limited the ANC leadership's ability to connect with, and empower, its grassroots membership.
Similarly, Inkatha's perception that the current levels of violence are a result of the mass democratic movement's strategies of armed struggle and ungovernability in the 1980s produces the conviction that there will be no solution to the violence until the ANC and its allies are politically isolated and weakened. The most significant moment for the IFP in 1991 was not directly related to violence, but was, rather, the 'Inkathagate' funding scandal which conclusively proved government and police support for Inkatha. It was also alleged that the Directorate of Military Intelligence provided both military training and funding for several hundred Inkatha members from Natal in the Caprivi Strip until at least 1989. These revelations served to discredit Inkatha in the eyes of the international community and of many white South Africans who had seen the organisation as a vehicle for 'moderate' hopes. The result was that Inkatha determinedly sought to demonstrate its independence in the latter part of 1991 which lead to disruptions to the national negotiation process and an escalation in violence perceived to be initiated by the IFP.
As the escalating violence dominated the political process, the ANC confronted demands from its membership that it prioritise the provision of defence (or at least the potential for defence) for a militant support base which considered itself under attack and defenceless.34
This was particularly true of the ANC's youth support base who had occupied centre state as the frontline of an organised resistance to apartheid in the 1980s. The era of negotiations saw them further marginalised by the shifting locus of political struggle - moving from the streets to the negotiation table - and increasingly conducted, not by the youth themselves, but by a range of grey-haired expatriates. But more important, this youth constituency, perceiving itself as under attack and unprotected, remained extremely sceptical of the benefits of the negotiation process. This scepticism translated into frustration as the negotiation process hiccuped and stalled. The result has been a constant struggle by the ANC to convince this militant support base to buy into the negotiation process. The organisation's relative failure to do so has been intimately related to the failure of the negotiation process to deliver meaningful agreements quickly enough.
The ongoing violence has also resulted in the loss of regulation and control by the organisation's leadership over its supporters whose expectations of the fruits of political change were not realised by the negotiation process. The expectations of meaningful change in the material conditions of township life are rooted in the legacy of political struggle of the 1980s. Issues such as housing, education and health care were politicised both by the racist administration and regulation of these areas of life, and by the political challenge to apartheid authority.
What has developed since the political reform process began in 1990 is the breakdown of the apartheid structures responsible for administering various material aspects of township life, without their being replaced by legitimate or effective alternatives. The resultant 'deregulation' of civic life, the effects of which have been exacerbated by devastating economic recession, has not only frustrated the aspirations of the voteless black majority but has also contributed to a significant decline in material standards of living. The combination of frustration and worsening socio-economic conditions could only increase the potential for violence.
To understand the complex nature of political violence, we need to understand the vehicles through which political identities are articulated and the aspirations which are attached to them. The solutions to the violence thus need to move beyond the political, into processes of socio-economic reconstruction and re-regulation of social life. A successful negotiation process and resultant constitutional change are but the first steps towards solving the problems of violence in the 'new' South Africa.
The hostel trend
As noted in the review of 1990,35 hostels have occupied centre stage in the Reef violence since its outbreak in mid-1990. Indeed, it has been argued that the first battle in the Transvaal war was the battle for control of the hostels. In the year to August 1991, Case claims that Inkatha-controlled hostels have been the launch pad for massive attacks on surrounding community squatter camps. These attacks account for at least 40 per cent out of the 2 271 deaths during that period.36 In 75 attacks from Soweto-based hostels, 315 people were killed. A further 491 were killed in 48 attacks emanating from East and West Rand hostels. A further 61 deaths resulted from 12 attacks emanating from hostels in Alexandra township.37
Whilst these statistics indicate the centrality of the hostels to much of the township violence in the period under review, they do not adequately explain the significance or complexity of hostel dwellers' involvement. Beyond the structural violence arguably inherent in these single-sex institutions, the complex political culture of migrant hostel dwellers is only intelligible as a combination of institutional, ethnic, economic and generational factors. Thus, Segal suggests that the political consciousness of hostel dwellers on the 'Rand is best analysed through a "kaleidoscope of explanations" which acknowledges all of these features as combining to shape this complex political culture'.38
In particular, a proper conception of the importance of hostel dwellers' involvement in the violence demands an understanding of the migrant consciousness of many hostel dwellers, rooted in the historically established social and economic concern to preserve their dual existence in both rural and urban areas. The relatively cheap hostel lifestyle was often perceived as essential to this dual existence. Unlike their neighbours in the squatter camps which burgeoned in the wake of the abolition of influx control, the hostel provided its inmates with structured access to minimal resources such as water points, ablution facilities and housing - a minimal urban base. The growing calls for the abolition of the hostels in the midst of conflict between hostel dwellers and township residents or squatter communities, threatened the very existence of this lifestyle and had the ironic effect of consolidating the defensive and volatile disposition of hostel communities.
The largely urban-centric trajectory of township politics throughout the 1980s saw migrant hostel dwellers increasingly politically marginalised from the wider township communities within which they existed. This was in spite of the greater inter-penetration between hostels and townships which resulted from the slackening of bureaucratic control over hostels in the wake of the abolition of influx control in 1986. However, the significance of this marginalisation was only to become manifest in the wake of February 1990 as the IFP implemented a drive for support in the historically politically hostile urban centres of the Transvaal, in an attempt to overcome the constraints of its regional powerbase in rural Natal. Zulu migrants were the obvious springboard for Inkatha into the urban Transvaal. The organisational potential of the hostels, with the 'captive political audience' which they offered, was a natural powerbase.
The effect was the political manipulation of this defensive community which was easily mobilised along ethnic and generational lines and quickly transformed into the central destabilising force in urban township communities. In the course of the year, hostel dwellers were frequently linked to attacks on trains and to massacres at vigils and funerals. The Swanieville attack on 12 May 1991 was one of the most prominent attacks involving hostel dwellers. The Swanieville squatter camp was attacked by as many as 1 000 men armed with guns, pangas and spears - 29 died and 80 shacks were destroyed. Amidst allegations of direct state involvement and complicity in the killings, Inkatha acknowledged its members' responsibility for the attack and claimed that it was in retaliation for the abduction of two hostel dwellers.39
The contention that much of this violence can be explained in terms of ethnic conflict requires further discussion. It would indeed be surprising if the ethnic orientation of forty years of apartheid did not, in some way, shape the identities of people living under the system. However, it is really where racial and ethnic identities correlate with conflicts over material resources or political allegiances that their potential as vehicles of violence becomes explicable. It is argued here that ethnicity does not provide a meaningful explanation of the violence in its own right but must be viewed in the context of the trends discussed above.
The real danger resides in strategies of political organisations (such as Inkatha and the range of extreme right-wing organisations) which deliberately mobilise support in ethnic or cultural terms. Once mobilised, such ethnic-based political identities become particularly volatile, defensive, and difficult to control. This is especially the case where such political tactics are used as the rationale for claiming that certain groups should have the right, based on 'cultural heritage', to carry dangerous weapons - and this becomes one of the political footballs central to the negotiation process in 1991.40 The implication of this analysis of hostel dwellers' key role in the urban township violence of 1991 is that it demands greater complexity in the search for solutions. Quite apart from dealing with the hostels as institutions which seem to generate violence, it is essential that the process of political and economic reconstruction takes account of the material consciousness which underpins the susceptibility of migrant hostel dwellers to mobilisation within the context of violent and divided townships.
'Third force' explanations
Just over 2 000 deaths were recorded by the HRC during 1991, as well as 2 500 known injuries. The PWV accounted for 50,4 per cent of all 'vigilante-related' deaths, with Natal following closely with 44,3 per cent.41 These statistics have been used as a basis for the claimed existence of a 'third force', fomenting violence and destabilising the peace process. However, once again, these statistics cannot tell the full tale. To explain the interface between political and material interests as a critical mechanism in understanding the growing violence has become a central theme of this paper. It was also central to arguments relating to the involvement of a 'third force' in the escalating conflict in the review of 1990.42 A sophisticated analysis of 'third force' involvement in destabilisation must involve a distinction between politically motivated interests in disrupting the peace process on the one hand, and the 'war-based' materially-rooted interests in maintaining high levels of violent conflict on the other. Thus, while it is clear that groupings within the ultra right-wing, the security establishment and even some of the key political organisations have a political interest in disruption (and this underpins the arguments of the Case studies cited above),43 these forces are reliant on the active engagement of materially-rooted interest groups as well.
In poverty-stricken communities confronting economic recession and contracting job markets, violent conflict quickly generates its own sub-economy - based on the trade in arms, assassinations and protection rackets. The insecure climate, coupled with the quest for defence and the resultant proliferation of private armies, generates ready-made foot-soldiers in the conflicts described. Candidates are readily found among disillusioned youth, criminal gangs and within the defensive and insecure hostel populations.
The arguments about politically motivated 'third force' involvement are most easily sustained in relation to the assassination of political activists, particularly those strategically placed in the structures of the unfolding peace process. In this regard, the HRC claims that at least 60 political activists were eliminated by the 'professional hit squads' in the course of the year, as compared with 21 during 1990.44 In March 1992, the Independent Board of Inquiry into Informal Repression (IBIIR) was reported to have linked the assassination or intimidation of at least 20 individuals to their visible involvement in such peace initiatives. Members of the Local Dispute Resolution Committees reportedly admitted that they felt even more threatened serving on peace bodies than they had done beforehand.45
However, one of the significant trends in the violence in the period since February 1990 has been its increasingly arbitrary and random nature as victims have been less and less politically selected. This suggests the growing involvement of less explicitly 'politically directed' killers, such as criminal youth gangs, hostel dwellers or warlords with a more material interest in this process of destabilisation. The first obvious example of this process has been the frequent attacks on mourners at funerals and vigils, and massacres at beerhalls in the PWV region. The Case report on killings at funerals and night vigils documents 13 such attacks in the year from July 1990 to July 1991 in which 89 people were killed.46 In the much publicised attack on mourners at the Sebokeng night vigil of ANC member Christoffel Nangalembe, 39 people were killed by an AK-47-wielding gang on 12 January 1991. Nangalembe had been an active member of the local ANC anti-crime campaign, and Everatt suggests that members of a notorious gang, The Five Star Gang, along with Inkatha supporters, were responsible for the attack.47
Another significant example was the attack at the night vigil of Jane Ramakgola on March 27 in Alexandra. The pre-dawn attack which left 15 people dead involved the use of pangas and automatic weapons and took place 800 metres from the Alexandra police station. It was followed by claims of police incompetence or complicity. The significance of this attack was that it signalled the intrusion of the violence into Alexandra township which had remained virtually unaffected by the Reef conflict in 1990.48
An even more striking illustration of this 'disruptive violence' has been the arbitrary massacres of train commuters. In 1990, 69 people were killed and 277 injured in 11 incidents of attacks on trains. In August and September of that year alone, nine of these incidents took place resulting in 58 deaths and 229 injuries. In 1991, the number of attacks more than doubled to 23, although fewer deaths (38) and injuries (208) resulted. In the first month of 1992, there were 14 recorded attacks resulting in 15 deaths and 72 injuries - the highest in any month since September 1990.49
Once again, these attacks have been accompanied by allegations of police ineptitude50 as well as allegations of direct involvement of Inkatha members and hostel dwellers in particular.51 Young and old, men and women, as well as members of all ethnic groups, have been victims of these atrocities. Any implication that there is a clear-cut political motivation for this violence must be quickly dispelled. If anything, what defines this reign of terror is the apparent arbitrariness in the selection of the victims.
A description of the dynamic of train attacks would be incomplete if it did not refer, at least in passing, to the devastating effect of rumour. This has resulted in the deaths of terrified commuters who, fearing imminent attack, have leapt from moving trains. One consequence of this pervasive fear has been the desertion of the trains by those who can afford other forms of transportation. However, the arbitrary nature of this violence has not been restricted to the trains but has followed township residents and commuters into taxis and back home to beerhalls and parties.
None of this detracts from the arguments about the active involvement of a 'third force' element during this particular period, but it certainly does suggest that there is a need for a slightly altered notion of how this third force is constituted politically. The cruelty and skill with which the attacks are carried out points to a possibility that these assailants are highly trained and do not identify in any way with their victims.52 For such mercenary-type operators, the generation of fear and increased conflict serves a direct interest whilst being simultaneously functional to the processes of destabilisation and the further deregulation of social control. The cumulative effect has clearly been a contribution to the unhinging of an already fragile negotiation process.
The violence perpetrated by ultra-right groupings must be seen in the context of the proliferation of both mainstream and fringe groupings in the wake of the unbanning of the liberation movements in 1990. The choice to remain outside of the negotiation process has also locked these political groups into a highly militarised and conflict-oriented political discourse which has been further serviced by the popular ethnic-based, racist explanations of township conflict.53 The overwhelming defeat of parliamentary right-wing groupings by the Nationalists in the all-white referendum of 17 March 1992 has probably served to consolidate this position.
The HRC reports that the number of deaths caused by organised right-wing violence decreased slightly from 26 in 1990 to 21 in 1991, but injuries inflicted increased from 138 to 178.54 However, the key to understanding the role of the right-wing resides in identifying the high media profile which they began to occupy (eg. the balaclava-clad 'ystergarde') and the extent to which they thereby began to provide a political home for deranged murderers in the image of Barend Strydom. In the course of the year, this paved the way for both the Middelburg massacre and the Ladysmith killings, both committed by individual white men who randomly shot people in the streets. Their victims were racially selected and the perpetrators articulated themselves in terms of the ideological discourse of the extreme right-wing groupings.
This explains the HRC conclusion that:
Right-wing actions during 1991 were again a mixture of organised premeditated attacks, and of impulsive and irrational outbursts against innocent victims.55
One of the most significant developments in the course of 1991, however, was the growing conflict between right-wing groupings and the government. The high point was probably the clash between the SAP and Afrikaner Weerstandsbeweging (AWB) supporters on 9 August in the Western Transvaal town of Ventersdorp.56 The clash, resulting from an attempt to disrupt a meeting at which State President de Klerk was due to speak, left several policemen injured and two right-wingers dead as well as scores injured. One black passerby was also killed. Later that same month, another right-winger was shot dead by police near the same town when demonstrators stormed a black squatter settlement on the nearby farm Goedgevonden. They were protesting against the re-occupation of the farm by black people who had previously been forcibly removed by the government.57
For perhaps the first time in its history, the NP government had to focus some of the attention of its intelligence-gathering and 'state security' apparatuses on the activities of white right-wing subversives, such as the fugitive Piet 'Skiet' Rudolph, who were involved in the theft of arms from SADF bases as well as in several acts of sabotage.
The National Peace Accord
Three agreements dealing with violence were signed in the course of the year. The ANC and the IFP signed an agreement in Durban in January 1991. The government and the ANC signed an agreement at DF Malan Airport in Cape Town the following month. On 14 September all three organisations signed the National Peace Accord in Johannesburg. The National Peace Accord was born out of the failure of the State President's Summit on Violence. It was a significant step, not only to address the increasing violence, but also in the national negotiation process. For the first time, a process of multi-lateral negotiation - initially between the IFP, ANC and NP, but later with a broader range of political groups - was focused around issues of peace and security. As such, it is arguable that, at its inception, the National Peace Accord was more significant as a symbol of the negotiation and consensus required to build the new South Africa than as a concrete mechanism to stop the violence.58
The Peace Accord consists of a wide range of mechanisms and principles. These include:
The Commission of Enquiry regarding the prevention of Public Violence and Intimidation (the Goldstone Commission). This commission is unique in so far as the scope, modus operandi and manner of appointment of commissioners is negotiated on a multi-lateral basis and decided by Judge Goldstone. However, like many structures of the Accord, it suffers from being crisis-driven and has no power to ensure that its recommendations are implemented.
The Code of Conduct for Political Parties. This section is a significant statement of agreement which might have the potential to form the basis of a new culture of political tolerance if it were more widely promoted and adhered to.
The Code of Conduct for the Police. This was supposedly signed by all members of the SAP by the end of the year. The code has had, in the absence of a public complaints procedure and other workable mechanisms of public accountability, less effect than might have been expected.59
Mechanisms for the scrutiny of the police. At a policy level this takes the form of a multi-party civilian police board which had not been appointed by the end of 1991. With regard to police misconduct, Police Reporting Officers are supposed to oversee investigations, but were also not appointed.
Local and regional dispute resolution committees. Only a very small number had been established by the end of 1991.
There are aspects of the Peace Accord which deal explicitly with socio-economic reconstruction and this represents an important recognition by the signatories of the fact that a one-dimensional political process cannot be expected to solve the problem of violence. However, the socio-economic upliftment mechanisms were to be limited to areas and occasions where the process of building political trust is successful.
The Peace Accord suffers from a basic tension in that it attempts to encompass both a conciliatory and a prosecutorial function in a single structure. Hence complaints are lodged at a local level about 'breaches of the National Peace Accord' which are then passed up to national level for resolution. This process relies on the parties taking symbolic action against their own supporters while at the same time attempting to maintain control over their members and encouraging them to create local peace. However, the political premium placed on appearing blameless means that the key political parties are reluctant to admit specific breaches on the part of their members.
Despite the breadth of the National Peace Accord, it has not had the positive effect that many expected. The impact of the Accord is limited to violence which is traceable to 'organised political constituencies'. Where violence is the result of agents provocateurs and shadowy 'third force' elements, not only is the Peace Accord ineffective, it is likely to lose support as well. The Peace Accord structurally entrenches the notion that the conflict is between organised political groupings and encourages party leadership to defend the actions of their supporters. It has thus retarded the process of 'recriminalisation of violence' which has, in turn, fed into the view on the part of the government and police that effective criminal investigations (which are their responsibility) are not essential for the violence to be halted.
However, an even more pessimistic perspective also presents itself and is in some respects illustrated by the developments in the taxi war in the Western Cape. At origin a conflict over relatively lucrative transport routes in the deregulated transport sector, the Webta/Lagunya taxi war rapidly translated into a conflict of political affiliations. Ironically, the effect of the Peace Accord has been to occasionally compound this problem through framing such conflicts and the resolution of them in terms of politically defined interest groups and parties. The result is that violence emanating from a range of ostensibly non-political sources rapidly becomes politicised.
Conclusion and postscript
At the time of writing, there is a good deal of scepticism in all constituencies about the National Peace Accord - this despite the fact that Local and Regional Dispute Resolution Committees (LDRCs and RDRCs) are now functioning in many parts of the country. The reality is that the forging of trust at a local level is a slow and difficult process. When this process falters, local disputes become imbued with national significance, making them that much more difficult to resolve without political loss of face. However, as developments in 1992 illustrate, the one-dimensional labelling of conflict as 'political' (which the Peace Accord substantially reinforced) increasingly allowed violence, or the lack of violence, to determine the nature and progress of multi-lateral negotiations.
The structures of the National Peace Accord cannot, in the absence of broader political settlements, substitute for the all-but-absent mechanisms of social authority and regulation that have disintegrated in the course of the past two years. The vicious circle is completed by the fact that constitutional negotiations themselves will not be able to address the materially rooted grievances which underpin much of the violence. Less still is an erratic process of negotiation able to sustain the confidence of black South Africans impatient for meaningful change in their lives. The end result is that violence has become the central stumbling block to the negotiation process set in motion at Codesa I at the end of 1991.
Despite the fact that Codesa 2 stumbled on questions of a constitutional nature, it was the massacre at Boipatong in June 1992 which finally saw the process collapse - as the politics of negotiation were finally overtaken by the politics of violence.
1 See G Simpson, S Mokwena, L Segal 'Political Violence in 1990: The Year in Review' in M Robertson and A Rycroft (eds) South African Human Rights and Labour Law Yearbook 1991, OUP (1992) 193-219. The present article is a development and update of the analysis in the article cited.
4 SAP Crime Investigation Department, Pretoria, August 1992. Official crime statistics are notoriously unreliable. In particular, police statistics reflect only reported incidents. In cases of political violence where the police are perceived to be partisan, the reporting rates are probably drastically skewed.
8 Indeed, it is arguable that this offers more insight into the problems associated with gathering such information and processing it, than it does into the causes and explanations of the violence itself.
29 SAP Internal Stability Division HQ, Pretoria. It is worth noting that only 29 policemen have been killed in Britain in the last 10 years. The Sunday Times, 24 November 1991; Sunday Star, 15 December 1991.
30 On this point, see E Marais 'Police-Community Relationship', In Glanz, L. (ed), Managing Crime in the New South Africa: A practical and affordable approach, Pretoria: HSRC Publishers. Proceedings of the Human Sciences Research Council conference, Pretoria, 4-6 August 1992 (1993).
31 See J Rauch 'A Preliminary Assessment of the Impact of the Peace Accord Code of Conduct on Police Behaviour', Paper presented at the Centre for Criminal Justice conference, Policing in the New South Africa II, Durban, August 1992.
39 See note 32 at 50-1. The attack took place within 24 hours of the imposition of a curfew in the squatter camp and the simultaneous granting of emergency powers to the police in the area. On 4 December 1991, a similar massive attack involving several hundred Inkatha supporters from nearby hostels took place at Bruntville township in the Natal Midlands. Here 15 residents were killed in what was the third attack within 2 hours; see note 15 at 53-61.
40 This danger is as equally illustrated by the rise of the AWB's 'Ystergarde' as it is by the centrality of 'traditional' rural-based migrant residents of single-sex hostels within township violence on the Reef over the past two years.
43 This sort of destabilisation thesis fits well with the analysis of 'low intensity conflict' methodology in which existing points of social friction (such as between gang members and other organisations or between taxi associations, etc.) are exploited and exacerbated.
44 See note 9 at 17. Amongst the most prominent of these were the murders of Civic Association of Southern Transvaal (CAST) executive member Sam Ntuli and the assassination of human rights lawyer Bheki Mhlangeni, who was killed by an explosive device supposedly intended for notorious SAP defector, Dirk Coetzee.
47 See note 46. This should be read in the context of reports to the Goldstone Commission as well as reports in The Weekly Mail relating to the alleged working relationship between the IFP and the 'Black Cats Gang' of Wesselton. The latter were allegedly involved in assassination of ANC figures in the Ermelo area.
48 It is argued by Everatt that this was at least in part because the Alexandra Civic Association as well as the Alexandra branch of the ANC had organised within the hostels as well as in the township. In March 1991, however, amidst allegations of the arrival of busloads of rural-based Inkatha supporters, a battle for control of the Alexandra hostels ensued. Everatt points out that: 'The attack on the funeral vigil took place at a time of heightened tension in Alexandra, which was an acknowledged ANC stronghold being violently infiltrated by Inkatha.' See note 46 at 5.
50 Despite promises of increased security and offers of rewards for information, little was done by the SAP. In October 1991, an SAP spokesman released a statement in which it was stated that it was 'too risky' for uniformed police to patrol the trains at rush hour and logistically impossible for them to search commuters on a regular basis. To date, despite a number of arrests, no convictions have taken place for any of the train massacres discussed here. See note 49 at 6-17.
51 According to the IBIIR, the involvement of IFP members in the attack of October 21, in which nine people were killed and 36 injured, was acknowledged by Themba Khoza, IFP Youth Brigade leader for the Transvaal, although he claimed that the IFP members were acting in self-defence. Many survivors of the train attacks have claimed that their attackers spoke Zulu and most of the attacks have taken place in close proximity to hostels. There have also been numerous reports by eye-witnesses or survivors of their attackers fleeing to the hostels once the attack was concluded. See note 49 at 3-4. On 19 July 1991, allegations were made by an ex-member of the SADF's 5 Reconnaissance Commando that members of this unit had been responsible for carrying out attacks on workers travelling on the trains from Soweto to Johannesburg. In particular, he alleged their participation in the Benrose train massacre on 13 September 1990.
52 The Weekly Mail reported that a group of 1 000 IFP members were trained by members of the SADF at a base called 'Hippo' in the Caprivi Strip in Namibia. One hundred were allegedly given 'recce' training and the rest 'contra-mobilisation' training. Training was done by 5 Reconnaissance Commando under the orders of the Chief of Staff Intelligence. The Weekly Mail 21 September 1991.
© Centre for the Study of Violence and Reconciliation