About this site

This resource is hosted by the Nelson Mandela Foundation, but was compiled and authored by Padraig O’Malley. It is the product of almost two decades of research and includes analyses, chronologies, historical documents, and interviews from the apartheid and post-apartheid eras.

X. Findings Based On A 'Balance Of Probabilities'

The commission made it clear that it was not required to meet the standard of proof pertaining in criminal law-namely, proof beyond a reasonable doubt. It emphasised that it operated in terms of a different standard that required merely a balancing of probabilities. When it was confronted with different versions of events, its task was thus to 'decide which version was the more probable, reasonable, or likely, after taking all the available evidence into account'. This, it pointed out, was 'the standard criterion used in civil litigation' and was the basis on which it came to its conclusions regarding accountability.

While the relevant standard of proof was thus only 'a balance of probabilities', that balance had nevertheless to be established in the light of all the relevant evidence. How adequately did the TRC discharge this obligation?

The lack of proper reasoning in the report makes it difficult to answer this question. So too does the length of the report and the number of the violations canvassed in it. There are various instances, however, in which the work of the TRC was preceded by judicial inquiries-through which a significant amount of relevant background information had earlier emerged.

In these judicial inquiries, evidence would have been properly corroborated and tested under cross-examination. Hearsay would have been excluded. A comprehensive account of the surrounding facts, as thus found, would have been provided. Reasons for any conclusions regarding culpability would have been adduced and properly explained. These rulings provide, accordingly, a good means of assessing how adequately the TRC weighed the balance of probabilities in coming to its findings.

1. TRC findings vis-à-vis earlier judicial rulings

Five such incidents-in each of which the conclusions of the TRC can be compared and contrasted with earlier judicial rulings-are described below. For ease of reference, they are dealt with in chronological order.


In 1990 a UDF committee operating in the Vaal region (south of Johannesburg) decided to hold a march through the central business district of Vereeniging on 26th March to protest, among other things, against the housing shortage and the education crisis. Permission for the march was sought on 22nd March, and was refused the following day. The organisers then developed alternative plans for a number of local marches to local police stations in the Lekoa townships. During such a march in Sebokeng, a number of protesters were killed and injured when police opened fire on the marchers. A commission of inquiry, chaired by Mr Justice Richard Goldstone, was appointed to investigate the shootings.

According to Judge Goldstone's report, a crowd of about 50 000 gathered in Sebokeng on the morning of 26th March 1990 and began their march. Some of the organisers of the protest met the local police commander, a Colonel Mazibuko, at the local police station and asked him to receive a petition there. In the meantime, however, the marchers passed the police station and moved on towards Vereeniging. The commanding officer of the local reaction unit (responsible for public order policing) came across the protesters and believed they were intent on marching on the town. To stop them, he formed a police line near the gate of a local brewery. The marchers came to a halt some 70m away.

Col Mazibuko and the organisers moved up to the police line and agreed that the petition would be handed over at this point, while the march would go no further. They communicated this to the commander of the reaction unit, a Captain du Plooy. The organisers then addressed the people, explaining that the petition had been handed over and that they should go home. Many in the crowd were still speaking of marching on Vereeniging, however. They were armed with sticks of various kinds and were singing freedom songs, such as 'Shaye M'Boere' ('hit the Boers'). They were apparently not aggressive. As the organisers were speaking, those at the back of the crowd surged forward to hear. Some also tried to outflank the police line. The crowd moved to within 40m of the police line, but there was very little stoning and no immediate threat to the police of being overrun.

Firing began without an order to shoot. A white constable fired a teargas canister, and set off a chain reaction in the police line. Shooting lasted for between 10 and 20 seconds, during which period some 60 rounds of ammunition were fired.

Judge Goldstone found that five people had been killed in the shootings and 161 wounded, 84 from behind. He also found the shootings unjustified, as 'the force used was quite immoderate and disproportionate'. He strongly criticised Capt du Plooy for failure to control the police line. He also criticised the undisciplined behaviour of the policemen, many of whom had loaded their firearms without orders-and some of whom had opened fire without orders too.

Judge Goldstone further criticised the organisers of the march for having decided at the last moment to hold separate and unauthorised marches, with no attempt to communicate this to the police; for having failed to tell the protesters that the Sebokeng march was no longer planned for Vereeniging (this failure having been a key reason for the crowd continuing to press forward towards the police line); and for having provided too few marshalls and megaphones with which to control the 50 000-strong crowd.

Judge Goldstone rejected any criminal conduct on the part of Capt du Plooy, but recommended that the actions of certain policemen in the police line be referred to the attorney general for investigation. Nine of these policemen were subsequently prosecuted, and six of them were charged with murder.

The TRC's finding on the Sebokeng shootings is as follows:

On 26 March 1990, police opened fire on a crowd of 50 000 people who were marching from Sebokeng to Vereeniging, killing 13 people and injuring over 400. Many of the injured people were shot in the back, indicating that they were fleeing when police opened fire. The Goldstone commission found that the gathering was peaceful before the police intervened and that the police did not give an order to disperse before opening fire with live ammunition. In reviewing the information on gross human rights violations committed at Sebokeng on 26th March 1990, the commission finds that the SAP and commanding officer W du Plooy were directly responsible for the deaths and injuries that occurred.

Elsewhere in its report, the TRC states that 17 people died in this incident, while a few pages thereafter it says that 'at least 13 people died'. In its national chronology of events, moreover, it gives the number of people killed as eight, and the number injured as 'over 300'. The commission adds that Judge Goldstone had recommended that 'police be prosecuted', but that 'no action was taken'. In making this last assertion, it ignores its own acknowledgement (contained in another volume of its report) that nine policemen had been put on trial in the Vereeniging Circuit Court, six on charges of murder, in August 1993.

The TRC contradicts the findings of Judge Goldstone in various respects. While Judge Goldstone found that five people were killed and 161 injured, the TRC states that 13 (or 17, or eight) individuals were killed and over 400 (or 300) people injured. It gives no reasons for its varying conclusions regarding the number of fatalities and injuries. The TRC holds Capt du Plooy 'directly responsible' for the shootings, and makes no mention of the Goldstone finding rejecting any criminal conduct on his part. The TRC states that 'no action' was taken against the policemen who opened fire without orders, whereas prosecutions were in fact instituted against nine of these police officers (six of them being charged with murder). Judge Goldstone, moreover, had criticised the organisers of the march as well and had made it clear that they had contributed in various ways to the incident (for which the ultimate culpability lay nevertheless with the police). The TRC makes no reference to the conduct of the organisers.

The TRC does not cite the evidence on which it relied. It does not demonstrate how it weighed the probabilities, or what testimony tipped the balance against the findings previously made by Judge Goldstone. The TRC's claim that 'no action' was taken against the policemen in question shows the criminal justice system in a very poor light, but is simply incorrect.

SADF shootings in Sebokeng in September 1990

On 4th September 1990, some 40 people were killed when IFP supporters attacked a hostel in Sebokeng, south of Johannesburg. A judicial inquest into these and certain further deaths was conducted by Mr Justice E H Stafford, who handed down his findings in March 1991.

According to Judge Stafford, the attack was mounted by IFP supporters who had earlier been evicted from the hostel by the ANC and were aggrieved by this. 'Inkatha supporters,' stated Judge Stafford, 'had been removed from, chased away from, frightened away from, had left their hostel dwellings, whatever you like, and were living in tents with charitable aid, and other residences, including one belonging to the KwaZulu government This was where it all started: the Inkatha supporters, evicted from the hostels and aggrieved thereby, and the mainly ANC supporters who lived in Sebokeng and those that still remained in the hostels.' Judge Stafford further found that the number of people killed in the IFP attack was 38.

What then occurred, continued Judge Stafford, was that '137 red head-banded Inkatha supporters became trapped inside a part of the hostel. The crowd [surrounding the hostel] grew to 5 000 people and was baying for their blood. They, the crowd, firmly believed that the 137 supporters of Inkatha, inter alia or solely, were responsible for the dead and fatally wounded scattered all round the hostel'. The SADF was called in after some hours to help the police 'remove and arrest and disarm the 137 members of Inkatha in order to avoid a bloodbath and to carry out their duties'. The information that had been given to the SADF commanding officer was that 'the police were cornered in the hostel with Inkatha members'.

Shortly after the army contingent arrived at the hostel, a soldier fired a shot and killed one person within the crowd. A further 160 or so rounds were fired within the next 20 seconds. No order to fire was given. Judge Stafford found that the soldier in question, aged 22 and of limited education, may have panicked at the sight of the 'huge, noisy, and hostile crowd'. He also found his behaviour, and that of the other SADF members who had fired without orders, to be inexcusable.

In conclusion, Judge Stafford found that a total of four people had been killed by the SADF. He expressed the hope that this finding would lay to rest rumours and press reports that 11 people had been shot dead by the army.

The TRC describes these events thus:

On 3 September 1990, Inkatha members carried out an attack in the early hours of the morning on the Sebokeng Hostel. Twenty-three people were killed in this initial attack. A further fifteen people died when the SADF opened fire on the crowd.

The commission subsequently expands its description, stating (among other things) that:

Ø. the Inkatha attack began at 1am on 3rd September 1990, and was carried out by Inkatha supporters armed with guns, hand grenades, home-made bombs, spears, and axes;

Ø. the attack was 'an attempt by those who had been evicted from the hostel in July to regain their former residence';

Ø. it took the police four hours to arrive at the scene, even though an SAP police station was situated less than 500 metres from the scene; and

Ø. the SADF opened fire without provocation on the crowd of Sebokeng residents who had gathered outside the hostel, trapping the group of attackers inside.

The TRC, again, effectively repudiates important aspects of an earlier judicial finding. While Judge Stafford had found that 38 people had been killed in the initial IFP attack, the TRC concludes that 23 people had died in this way. While Judge Stafford had ruled that four people had been shot dead by the SADF, the TRC finds that 15 people were killed by the army. The commission cites no evidence and provides no reasons for its conclusions regarding the number of people killed in either incident. Nor does it deal in any way with the hope expressed by Judge Stafford that his finding (that four people had been killed by the SADF) would put an end to rumours of a much higher death toll. The commission does, at various points in its report, acknowledge that '350 Zulu-speaking people' had been expelled from the Sebokeng hostel complex, and that the IFP attack was an attempt to 'regain their former residence'. The significance of this 'antecedent' factor is not explored, however. Elsewhere in its report, the TRC entirely discounts this factor, moreover-stating merely that 'Inkatha had allegedly tried to lay siege to and occupy the Sebokeng hostel'.

Deaths in Tokoza (east Rand) in September 1991

Eighteen hostel residents on their way to a meeting at the Tokoza stadium on the east Rand were assassinated on 8th September 1991. The Goldstone commission was mandated to investigate both these (and other) killings in the area, and delivered its report in July 1992.

The commission found that there was a high level of conflict in the Tokoza township between the residents of a squatter camp called Phola Park, and those who lived within the hostels. This enmity had earlier resulted in the demolition by the residents of Phola Park of a large hostel complex adjacent to their settlement. Tensions had been further fuelled by several attempts to remove or relocate the Phola Park residents, by the establishment by the Phola Park residents of a self-defence unit (SDU), and by ongoing incidents of violence in adjoining shack settlements.

At the beginning of September 1991 the Tokoza Hostel Dwellers' Association decided to hold a meeting of hostel residents on Sunday, 8th September, at the Tokoza stadium. On the Sunday morning in question, a small group of hostel residents from more distant hostels gathered at the southern entrance to the stadium, awaiting the arrival of a bigger crowd of hostel dwellers from three hostels in Tokoza itself. The residents of these three hostels began moving down a road called Khumalo Street towards the stadium. Their conduct was apparently not provocative and was reasonably orderly. When some of these hostel dwellers had already passed house 2044 on Khumalo Street, near the north east corner of the stadium, three men armed with AK-47 rifles opened fire on them at very close range from the front garden of the house. Sixteen hostel dwellers were killed and 13 injured.

Immediately the firing of automatic weapons commenced at house 2044, the small group of hostel dwellers that had already gathered at the south entrance to the stadium came under attack as well. This attack was initiated by a group of men located at the south west corner of the stadium. Handguns were fired, and the hostel residents were assaulted with spears, pangas, assegais, and axes. A further two hostel dwellers were killed, bringing the death toll to 18.

Three men armed with AK-47 rifles were also seen at the north west corner of the stadium. They were not seen to have fired any shots, however. There was also some evidence, found inconclusive, that a fourth group had been stationed at the south east corner of the stadium.

The Goldstone commission found that the Phola Park SDU had held a meeting some time before the march began. At that meeting, 'it was arranged to mount an ambush on the hostel dwellers by locating small units of probably not more than three men each at the four corners of the stadium'. Shortly before the march, the SDU had called a meeting of its section leaders at a water tank at Phola Park. At that meeting the section leaders were advised of the impending gathering of the hostel dwellers, and Phola Park went on to a defensive as well as a 'war footing'-many of the residents 'cloaking themselves with blankets, which constituted an accepted "military uniform"'.

The three assassins who opened fire from house 2044 constituted the unit delegated to the north east corner of the stadium. The attack on hostel dwellers at the southern entrance to the stadium was executed by the unit delegated to the south west corner of the stadium. (It was clear, moreover, that some blanketed Phola Park residents who were not members of the SDU had also been involved in this attack.) The three men with AK-47 rifles seen at the north west corner of the stadium constituted the unit delegated to that sector.

The attack at house 2044 on Khumalo Street, said the Goldstone commission, was 'executed with a high degree of professionalism-in the sense that the AK-47 rifles were handled competently and having regard to the number killed or wounded, as was intended by the assassins'. The attack at the southern entrance of the stadium which followed immediately thereafter was not 'mere spontaneous violence', moreover-for both its timing and the evidence of an ambush plot militated against this.

The police and defence force, added the commission, were 'effective in bringing the whole situation rapidly under control, and in restoring relative calm to the area. There were no retaliatory attacks or counter-attacks despite the very high level of tension in the area following the attack on the hostel dwellers'.

In the course of the subsequent police investigation, the Goldstone commission continued, a suspect and three other people who happened to be residing at the same place were taken to a farm called Vlakplaas. This was not a police station, and normal police records were not kept there. Two minor brothers of the suspect were arrested the same morning. All but the suspect were released the same day. 'An allegation of assault on the suspect was not proved, but the lack of records detracted from the police case,' the commission stated.

The Goldstone commission declined to name any individual as having been guilty of the shootings at house 2044 on Khumalo Street. The police investigating officer, despite extensive efforts, had been unable to find sufficient evidence to lay charges against any individual-and the commission considered itself incompetent to apportion blame to named people in such circumstances.

During the commission's investigation, it emerged that one section leader in the Phola Park SDU, Mr Mncugi Ceba, was a police informer. It also emerged that Mr Ceba had 'actually led the coup' in terms of which 'the Phola Park SDU had ousted the Phola Park Committee'. (This event was not further explained.) It was further argued before the commission that, 'by virtue of having informers in key positions, the police probably knew of the planned attack on the hostel dwellers on 8th September'. It was not contended that the police had instigated the attack in any way.

The commission found that 'no offence on the part of the police was proved'. It added that the use by the police of 'informers in positions such as that held by Ceba was certainly not conducive to improving the already tense relations between the security forces and the communities in question'.

The TRC reports the incident as follows:

On 8 September 1991, a three-month period of relative calm was shattered days before the signing of the National Peace Accord. Approximately 300 members of the Hostel Dwellers Association on their way to a peace meeting at the Thokoza stadium on the East Rand were sprayed with gunfire by three AK-47 wielding gunmen, killing at least 23 people. By the following night, 42 people were dead and at least 50 injured in retaliatory attacks that swept Katlehong, Tembisa, and Johannesburg. Both the ANC and Inkatha later stated that they believed the killing was provoked with the aim of derailing the peace process. Members of the Political Violent Crime Unit based on a farm in Katlehong called Vlakplaas arrested and allegedly tortured a number of SDU members in response to the attack.

The Goldstone Commission found in 1992 that this attack had been planned and carried out by a police informer, Mr Mncugi Ceba, who posed as the head of an ANC SDU in Phola Park. One of the participants in this attack, ANC member Mr Michael Phama, who is currently serving a life sentence for his involvement in the incident, applied to the Commission for amnesty for the killings. He stated in his amnesty application that he was ordered by his SDU commander to shoot "because IFP members might attack our people as they always attack when they have a rally".

The commission's finding on the incident is as follows:

The Commission finds that, on 8th September 1991, 23 people were killed at the Thokoza stadium on the eve of the signing of the National Peace Accord on the east Rand. The Commission finds that 42 people died and at least 50 people were severely injured in violence that broke out between supporters of the ANC and the IFP in the two days that followed. The Commission finds that the initial attack was initiated by one Michael Phama, a member of an ANC SDU, acting on the instructions of his commander, Mr Mncuzi (sic) Ceba. The Commission finds that Ceba was a police informer in the pay of the SAP.

The discrepancies between the Goldstone commission report and the TRC report are significant. Goldstone found 18 people killed, the TRC found 23. Goldstone made it clear that the attack had all the flavour of a professionally-executed ambush, which had involved the stationing of SDU units at three corners of the stadium at minimum. The TRC, by contrast, indicates that it was three lone gunmen (those positioned at house 2044, presumably) who hadopened fire on the hostel residents.

Goldstone said one suspect might have been assaulted by the police at Vlakplaas, though this had not been proved. The TRC states that the police 'allegedly tortured a number of SDU members' after the attack. Goldstone says rapid police and army action prevented any retaliatory attacks after the killings. The TRC says 42 people were killed and 50 injured in such attacks within the next two days.

The Goldstone commission, moreover, professed itself incompetent to name either Mr Ceba or any other individual as having been guilty of the shootings from house 2044. It made it clear that it could not do this in the absence of sufficient evidence. The TRC evinces no such constraint.

What the TRC states about Mr Ceba is entirely inaccurate, moreover. The Goldstone commission discovered that Mr Ceba, who was a section leader in the Phola Park SDU, was also a police informer. The Goldstone commission further stated that Mr Ceba had led 'a coup' against the Phola Park Committee. Goldstone also made it clear that there were a number of section leaders within the Phola Park SDU, and that the attack had been planned by the SDU as a whole-not by a single section leader. The TRC ignores all this, and states that Goldstone 'found that the attack had been planned and carried out by a police informer, Mr Mncugi Ceba'.

The TRC thus misrepresents Goldstone. It also provides no evidence and no reasoning to show why his findings should be repudiated. It further fails to demonstrate how it weighed the balance of probabilities, or how it found these to support its conclusions.

The Boipatong massacre in June 1992

On 17th June 1992, 45 people were killed in Boipatong, a township in the Vaal triangle south of Johannesburg, and in a neighbouring informal settlement called Slovo Park. Many others were injured. The ANC, which had earlier called for a renewed campaign of mass action to force the government from power, accused NP leaders and the police of having been party to the killings. The Goldstone commission was asked to investigate, and called in a team of policing experts from the United Kingdom for this purpose. This team comprised Dr Peter Waddington (director of criminal justice studies at the University of Reading), as well as Commander Tom Laidlaw and Detective Superintendent David Gon (both of the London Metropolitan Police). Dr Waddington presented his report to the Goldstone commission on 20th July 1992.

According to Dr Waddington, 'there was no evidence that the police had any forewarning of an impending attack in Boipatong'. There were suggestions that some serious event would take place somewhere in the Vaal Triangle. However, to the extent that any area was specifically mentioned, it was Sebokeng and not Boipatong.

Dr Waddington summed up the police role in events at Boipatong as follows:

At around 10.00pm on 17 June police began receiving calls reporting shootings, assault, and damage from the Boipatong township. Two Casspirs were deployed under the command of their respective sergeants. The sergeants discovered evidence of murder, arranged for ambulances, and asked for detectives to investigate. The 'scene-of-crime' detective made a necessarily superficial examination of the murder scenes during two visits lasting from midnight to 3.00am, and from 7.30am to mid-morning. The ISU (Internal Stability Unit) supported the detective, protecting him whilst he made his investigations, and remained on patrol throughout the day. However, they took no part in the investigation itself (for example, by interviewing witnesses) and saw their role as merely preventing further disorder and violence. Throughout the night the mood of the township residents was agitated but not hostile to police. Hostility towards the police began to grow after daybreak, possibly as a result of rumours of police involvement in the massacre. Violence escalated and police fought a running battle with rioters throughout much of the day.

The subsequent police investigation, continued Dr Waddington, concentrated on the KwaMadala Hostel, which police visited on several occasions the day after the killings. As a result of repeated interviewing, a number of suspects were then arrested and detained. Inquiries in Boipatong itself, among potential witnesses to the massacre, were obstructed throughout by the hostility and non-co-operation of residents. This, stated Dr Waddington, was 'apparently at the behest of the ANC'.

Dr Waddington criticised the response of the police to the massacre, as well as the methods used in investigating the killings. He found, among other things, that the police had not made the best use of their limited manpower, that their intelligence gathering had been inadequate, that their contingency planning was defective, that their investigation was insufficiently co-ordinated, and that they had not tried hard enough to win the trust of the Boipatong community in the immediate aftermath of the killings. Dr Waddington also made it clear, however, that 'no evidence had been found of direct police complicity in the massacre itself'.

In August 1992 the Goldstone commission began its own hearings into the massacre. Little, if any, credible evidence of police culpability emerged. In particular, the testimony of a special constable who had claimed he had seen gunmen 'climbing into police armoured vehicles' was discredited, for an inspection in loco made it clear that he 'could not have seen what he claimed to have seen'. In November, Judge Goldstone accordingly announced that 'it was impossible to make a finding'.

Some 30 residents of the KwaMadala Hostel were subsequently tried for murder for their alleged part in the killings. During their trial, the accused contended that they had not been involved in the attack at all-and alleged that the police had been responsible instead. The police role in the massacre thus became a critical issue for Mr Justice J M C Smit to decide. Three accomplice witnesses from KwaMadala Hostel who gave evidence for the state all denied that the police had been involved. Some 120 witnesses from the Boipatong and Slovo Park communities gave evidence as well, and denied that police vehicles had assisted the attackers. The principal witnesses who testified to police involvement in the massacre, Messrs Joseph Sello and Abednego Mabuza, were unable to explain a number of material inconsistencies and contradictions in their evidence. Mr Sello was found particularly 'dishonest and unreliable', while Mr Mabuza-though less obviously a liar-was far from credible either.

In the light of all the testimony before him, Judge Smit concluded that the police had not in any way participated in the killings. The allegations by Messrs Sello and Mabuza demonstrated, he continued, how rumours of police culpablity had been spread-but there was no truth in those rumours whatsoever.

Moreover, though the tapes of transactions in the control room of the ISU had been found to have been erased when the Goldstone commission began its investigations, there was nothing sinister in this. According to Judge Smit, 'the erasure of the tapes was the result of incompetence rather than a deliberate attempt to hide evidence of police complicity in the attack'. He came to a similar conclusion regarding other evidence, in the form of eight bullet shells, that had also been inadvertently destroyed by the police. (Further bullet shells had, it seems, been removed from Boipatong by 'comrades' and unnamed persons in the aftermath of the massacre-at a time when the ANC had apparently instructed residents not to co-operate with the police. This may have impeded proper investigation to an equal extent.)

The TRC was aware of the court's decision, for it cited Judge Smit as having 'unequivocally stated that, in the light of the testimony he had heard, there was no evidence to support the allegation that the police in any way participated in or were involved in the Boipatong massacre'. It cited as well his further conclusion that 'the erasure of the tapes was the result of incompetence rather than a deliberate attempt to hide evidence of police complicity'. It noted his similar conclusion regarding the bullet shells that had been destroyed. It referred as well to Judge Goldstone's statement that 'he had not received any evidence that led him to conclude that the police were involved in the attack'. And it cited the Waddington inquiry too, and noted its conclusion that the police had been guilty of 'inefficiency and incompetence' but not more.

The TRC's findings regarding the Boipatong massacre are, however, as follows:

The commission finds that 45 people were killed and 22 severely injured in Boipatong on 17 June 1992 in an attack perpetrated by residents of the KwaMadala hostel, who were primarily supporters of the IFP. The commission finds that KwaMadala residents, together with the police, planned and carried out an attack on the community of Boipatong and the surrounding informal squatter settlement, Slovo Park, on 17 June 1992. The commission finds that the police colluded with the attackers and dropped them off at Slovo Park. The commission finds that white men with blackened faces participated in the attack. The commission finds further that, despite the presence of armoured vehicles in the township, the police failed to intervene and stop the attackers, despite calls by the residents of Boipatong and Slovo Park to do so. The commission finds that the police were responsible for destroying crucial evidence in that they erased the tapes of transactions in the control room of the ISU

The commission finds the KwaMadala residents together with the SAP responsible for the massacre, which resulted in the deaths of 45 people and the injury of 22 others. The commission finds the commissioner of police, the minister of law and order, and the IFP responsible for the commission of gross violations of human rights.

The TRC also uses, among other things, the inadvertent erasure of the ISU tapes and destruction of eight bullet shells to buttress a further finding that:

The police, in their approach to the prevention and investigation of political violence, were biased in favour of the IFP and their failure properly to investigate such violence led to large numbers of gross violations of human rights, and strengthened the prevailing culture of impunity. The SAP was thus accountable for the gross violations of human rights that resulted from their actions.

The TRC fails to explain how it reconciles its view that police investigations were biased in favour of the IFP with the fact that 17 residents of a hostel that primarily housed supporters of the IFP were successfully prosecuted for murder.

Nor does the TRC explain its rejection of the conclusions reached by both Dr Waddington and Judge Smit: viz, that the police had not been involved in the killings. It also fails to explain its reasons for discounting Judge Smit's finding that the erasure of the ISU tapes and destruction of the eight shells had been the result of incompetence rather than anything more sinister. In its strictures against the police for inadequate investigation of the Boipatong massacre, moreover, it makes no mention of the ANC's apparent instruction to residents not to co-operate with the police-and the likelihood that this would have increased the difficulty of mounting a proper investigation.

The TRC cites as 'evidence' of police involvement in the massacre the testimony of various victims and other residents of the area, most of whom remain unnamed. It does not, however, describe their evidence in any depth. Nor does it explain how such evidence was tested or substantiated. It gives no reasons why the untested allegations put before it should have prevailed over the conclusions of the trial court. Those conclusions, furthermore, had been based on the fact that three accomplices and some 120 residents of Boipatong had all testified that the police had not played any part in the attack. Moreover, the witnesses who had alleged the opposite had been shown, under cross-examination, to be dishonest and unreliable.

Whether the commission had any substantial new evidence before it to justify its findings is also unlikely. According to Mr Jan-Ake Kjellberg, a Swedish policeman serving with the TRC, the commission conducted no real investigation of the massacre. It found no new witnesses, and elicited no novel or compelling testimony to cast fresh light on the killings.

If anything, the TRC seems to have lifted its evidence, virtually verbatim, from a report by the Human Rights Commission, or HRC (which in turn had relied on a monitoring organisation called Peace Action). This HRC report was compiled within a few weeks of the massacre at most and before the allegations against the police had been put to any test. This Peace Action/HRC 'evidence' has effectively been recycled as a finding of the TRC.

The commission's obligation, when confronted with different versions of events, was a very different one. It was not to give an official sanction to monitors' reports but to decide which version of disputed facts was 'the more probable, reasonable, or likely, after taking all the available evidence into account'. This-'the standard criterion used in civil litigation'-was, it stated, the basis on which it came to its conclusions regarding culpability.

In the context of Boipatong, however, there is no indication that the TRC made any attempt to weigh which version of events was 'the more probably, reasonable, or likely'. On the contrary, it seems simply to have spurned the ruling of Judge Smit, as well as the earlier finding by Dr Waddington.

The commission was specifically enjoined, moreover, to probe the motives and perspectives of all perpetrators of gross violations. Yet it ignored a seeming pattern of earlier attacks on IFP supporters in the area as well as the fact that the KwaMadala Hostel had become a refuge for Inkatha supporters driven from their hostels or township homes by ANC-supporting 'comrades' and SDUs. It also ignored attacks on IFP supporters that, it seems, had immediately preceded the massacre. These attacks, as described by two journalists, Mr Rian Malan and Mr Denis Beckett, had begun some four days prior to the massacre:

The first person to die was a woman named Nomvula, whose sin was a romantic involvement with a Zulu hostel dweller. A crowd cornered her on Nkgomo Street and necklaced her-burnt her alive. When police tried to intervene they were attacked with stones. Thirty minutes later, a mob torched the home of David Mbele, a former schoolteacher and Inkatha member. A fire engine arrived, only to be driven off by gunshots and stones. Mbele ran for his life, but the mob caught and killed him. The charred body of a third Inkatha member, Mr D L Khumalo, was found near Boipatong cemetery the following afternoon.

The commission ignored not only this account-which would seem to have merited further investigation-but also what was said by Judge Smit on the question of prior provocation. Having rejected police culpability in the massacre and found 17 of the accused guilty on various counts of murder, Judge Smit was obliged to weigh the aggravating and the mitigating circumstances in order to determine an appropriate sentence. Having done so, he concluded that 'the mitigating factors outweighed the aggravation'. Describing these factors, Judge Smit stated that the houses of IFP supporters and their families had been burnt down, while they themselves had been 'murdered and mutilated and driven from their communities'. Since 1990, the KwaMadala Hostel had become the only place of refuge for the IFP members thus forced to flee their homes. (Judge Smit found the death penalty an inappropriate punishment and imposed long prison sentences on the accused instead.)

The Shell House shootings in March 1994

On 28th March 1994 thousands of Zulu loyalists armed with traditional weapons marched through Johannesburg in solidarity with a call by the Zulu monarch, King Goodwill Zwelithini, for a renewal of his sovereignty over KwaZulu and Natal. Shooting broke out when the marchers were assembled at the Library Gardens, and ten people were killed-eight of whom were Zulu demonstrators. In addition, eight Zulu marchers were shot dead by ANC security guards from the roof of the organisation's national headquarters, Shell House.

Police obtained a warrant to enter Shell House, but were denied entry to the building following the intervention of Mr Nelson Mandela, then president of the ANC. Mr Mandela told a press conference some days later that 'he had refused the police permission to enter Shell House to gather evidence on the killing of the eight Zulus'. Only after the police had demonstrated their impartiality by raiding all the hostels in the Johannesburg area, he continued, could they also raid Shell House. The ANC promised, however, to co-operate in the police investigation and to hand over to the police all relevant firearms.

It was only in June 1995, however, that Mr Mandela first disclosed that he, personally, had instructed the ANC's security guards to protect Shell House from any attack, and to use lethal force as well if this were necessary. It was only in July 1996, moreover-more than two years after the event-that the minister for safety and security, Mr Sydney Mufamadi, told Parliament that the last batch of weapons from Shell House had been handed to the police that month. (It nevertheless remained unclear whether relevant weapons were still outstanding-for in December 1996 Mr Mufumadi allegedly told the newspaper Rapport that 99 firearms had yet to be delivered to the police.)

A judicial inquest into the shootings was initiated in 1997. Mr Justice Robert Nugent and two assessors, Professor R C Whiting and Mr J S Baloyi, handed down their findings in December that year. Judge Nugent said that his task had been complicated 'by the failure of the authorities to conduct a thorough and prompt investigation into the killings soon after their occurrence, when the trail of evidence was still intact'. He thus relied substantially on relevant video footage for his assessment of events.

Judge Nugent dismissed the IFP's contentions that the march had been organised by Zulu loyalists and indunas without party political connections. He found it had been organised by the IFP itself and in the party's name. Moreover, when permission for the protest had been sought from the relevant authorities, it was 'the IFP that provided an indemnity for any damage which was caused'.

Judge Nugent then turned to the ANC's contentions that IFP leaders and rogue police officers had conspired to attack Shell House, that the marchers had been the first to open fire, and that the guards had 'fired back in defence of their lives, their headquarters, and the ANC leaders who were there on the day'. Judge Nugent rejected the evidence proffered by the ANC, finding (among other things) that 'it was fabricated after the event so as to bolster the explanations that had been put forward for the shooting which had occurred at Shell House'.

Judge Nugent's concluding words regarding the Shell House shootings were as follows:

Prima facie the evidence does not show that Shell House and its occupants were about to come under attack nor could it reasonably have been believed at the time that it was about to come under attack. Prima facie there was no justification for shooting at the crowd at all. Moreover, the barrage of fire was in any event grossly excessive. We do not accept that any warning was given It is clear, too, that when the shooting started the crowd immediately disintegrated and fled. To have continued firing at them went far beyond what would be permitted in legitimate defence.

The TRC's description of the Shell House shootings is as follows:

On 28 March 1994, approximately 50 people were killed and more than 300 injured during violence associated with a march through Johannesburg in support of the Zulu king. The violence occurred after Transvaal indunas (traditional leaders) called on Zulus in the PWV region to stay away from work and gather at the Library Gardens in central Johannesburg to demonstrate their support for the Zulu sovereign, King Goodwill Zwelithini. After the events of 28 March, the IFP leadership was at pains to emphasise that the march was an independent initiative of the 'Zulu people', rather than a political gathering organised by the IFP. However, senior IFP leadership was present at the gathering and involved in its organisation.

From the start, information about the proposed gathering was confused. Many people believed that the marchers intended gathering at the offices of the Independent Electoral Commission in order to demonstrate their opposition to the elections. However, such a march did not take place. It subsequently emerged that senior IFP leadership had received permission from the Johannesburg magistrate to hold a gathering at the Library Gardens, but had not sought permission for a march of any kind. No organised march did in fact happen. Instead, armed groups of men launched a series of 'offensives' against ANC offices in the city centre. The first three such offensives focused on the ANC regional offices. When the marchers moved on the ANC headquarters at Shell House, ANC security guards responded with automatic gunfire, killing eight people. Several of the ANC security guards who opened fire at Shell House applied to the Commission for amnesty. These hearings were ongoing at the time of reporting.

The TRC does not refer to the judicial inquest. It ignores the finding by Judge Nugent that claims of impending or actual attack on Shell House were untrue. It also omits to mention Judge Nugent's finding that there was no 'justification at all' for the shootings at Shell House. It leaves out the judge's conclusion that the 'barrage of fire was in any event grossly excessive'. It also makes no attempt to explain its view that the IFP marchers were engaged in an 'offensive' against Shell House-a contention that Judge Nugent had not only expressly rejected but had also found to have been 'fabricated after the event'.

2. A possible pattern in the TRC's approach

The errors, omissions, and (on occasion) misrepresentations in the TRC's account of these five incidents are serious. Should they be dismissed, however, as random and relevatively insignificant inaccuracies? Or do they reflect a pattern in the TRC's approach? If a pattern can be discerned, moreover, what is its effect? In assessing these issues, each incident merits brief recall.

In describing the Sebokeng shootings in March 1990, the TRC doubles (or even triples) the number of fatalities at the hands of the police. It virtually doubles the number of people injured; ignores the fact that the police officer in charge was found by Judge Goldstone not to have acted criminally; ignores the behaviour of the march organisers; and wrongly claims that none of the policemen in question was prosecuted.

In describing further incidents in Sebokeng in September 1990, the TRC downplays (or omits) the context in which the IFP attack on the hostel occurred; states that the SADF killed nearly four times as many people as Judge Stafford had found; ignores the judge's explicit call for his finding to put an end to rumours of a higher death toll; and states that the IFP killed 23 people when the correct figure was 38. Since 38 minus 23 is 15, the implication is that the TRC failed to do its homework. Having ruled that the IFP had killed 23, it may have attributed 15 killings to the army on the basis of a simple, but incorrect, subtraction.

In describing the attack on hostel residents in Tokoza, the TRC omits salient evidence of a well-planned and three (or four) pronged ambush. It misrepresents what Judge Goldstone had said about the police informer, Mr Ceba; ignores Judge Goldstone's further finding that the police had committed no offence; and effectively convicts the police of premeditated murder. At the same time, it exonerates the Phola Park SDU, and this despite the evidence of the seminal role that the SDU had played in the ambush and the killings.

In the context of the Boipatong massacre, the TRC omits the attacks on IFP supporters immediately preceding the killings as well as the way in which IFP followers had earlier been assailed and driven from their homes-forcing them to find refuge in KwaMadala Hostel. It spurns the findings of a UK policing expert, Dr Waddington, and repudiates the ruling of Judge Smit in the trial of the Boipatong accused. It uses unexplained, untested, and unsubstantiated allegations to convict the police, in effect, of 45 killings.

In the context of Shell House, the TRC effectively puts the blame on the IFP for provoking the shootings through its 'offensives' on ANC offices in Johannesburg.It omits all reference to Judge Nugent's inquest; ignores his finding that the ANC had lied about an impending attack; and disregards his further conclusion that the shootings had been unjustified and 'grossly excessive'.

The effect of the TRC's approach, in each of these five incidents, is to heap the blame for violence on the former police, the former army, and/or the IFP. At the same time, any possible culpability of the ANC is downplayed or ignored.

3. An unconvincing explanation

The commission's own explanation of why its findings differed sometimes from those of earlier judicial rulings merits mention. The TRC confined its explication to judgements concerning police misconduct in 'riot' situations. Within this context, it stated, some of its findings had differed from earlier judicial rulings for two reasons. First, it was sometimes 'presented with new and compelling evidence (for example, corroborated statements by victims or witnesses)'. Secondly, it regarded 'the use of lethal force as justified only in extreme situations', whereas earlier judicial rulings had been based on an 'uncritical application' by the bench of the Criminal Procedure Act of 1977, which had given the police 'very wide powers to use lethal force'.

Neither reason stands up to scrutiny. As regards the first, victim statements-even if 'new and compelling'-were neither tested under cross-examination nor sufficiently corroborated. They did not have the evidentiary stature to support the repudiation of earlier judicial rulings.

The second reason is also flawed. The Criminal Procedure Act of 1977 does not deal with the use of lethal force by the police in riot situations. The relevant statute, within the commission's mandate period, was rather the Internal Security Act (ISA) of 1982 (which repealed and replaced the virtually identical provisions of the Riotous Assemblies Act of 1956). The ISA states that the use of force by the police must always be 'moderated and proportionate to the circumstances'. It adds that lethal force may not be used unless there is an actual or imminent threat of death or injury to any person, or of destruction or serious damage to valuable property. Even then, the force used must be applied with 'all reasonable caution, without recklessness or negligence, and so as to produce no further injury to any person than is necessary' to protect life, limb, or valuable property. The TRC is thus misleading in implying that earlier legislation gave the police virtual carte blanche to use lethal force in wide-ranging circumstances. The effect of its misrepresentation is again to hold up the police, the former government, and the judiciary to unjustified opprobrium.

Even if the TRC's two reasons were to be accepted at face value, they would still not sufficiently explain the discrepancies, in the five instances described above, between the commission's findings and earlier judicial rulings. For, in three of these incidents, the use by the police (or army) of lethal force in riot situations was not what was in issue at all. Moreover, there is little indication (for example, as regards the Boipatong massacre) that the commission had any new evidence at its disposal.

4. Ignoring other rulings too

The five incidents outlined above are not the only ones, moreover, in which the TRC has departed from earlier commission or judicial findings-and done so without explanation and in a manner likely to exacerbate the culpability of the IFP or the former security forces. Some of these further incidents may briefly be summarised as follows.

The Ngoye incident in 1983

In October 1983 five people died on the campus of the University of Zululand at Ngoye (KwaZulu). A commission of inquiry, chaired by Professor A J Middleton, a law professor at the University of South Africa, found that a group of students had taunted and attacked Inkatha supporters visiting the campus for a commemoration ceremony. The visitors retaliated, a fierce battle ensued in the men's hostel, and three students were killed. Later that day students, in apparent revenge, killed an Inkatha supporter who had arrived late and had taken no part in the earlier conflict. The following day a fourth student died of renal failure, probably caused by a combination of exhaustion (from running) and obesity. (According to an ANC leader and lecturer at the university, the student had heard a rumour that Inkatha was about to attack the campus for a second time, and was trying to run away.)

According to the TRC:

Ø. four students were killed by a group of 500 Inkatha supporters;

Ø. the four died when Inkatha 'attacked the students' residences, breaking down doors and pulling students out from where they were hiding', and then 'assaulting them and stabbing them with traditional weapons'; while

Ø. an Inkatha supporter was also killed in this clash.

Elsewhere in its report, the TRC states that 'Inkatha aligned "warriors" killed five students' in this incident, and injured many more.

The commission does not explain the reasons for its finding on the Ngoye incident-in which calculated killings are implicitly found proven and relevant contextualisation is disregarded. In its description of the incident, the TRC ignores the substantial body of tested evidence assembled by the Middleton commission, and lends credence to the version of events that has consistently been put forward by the ANC alliance.

The ANC's response to the incident, in an article in Sechaba in February 1984, was to accuse Inkatha supporters of having acted like 'Nazi youth' in their allegedly vicious and premeditated assault on students. ANC leaders have continued to echo this theme, generally accusing Inkatha of having murdered five innocent students. In doing so, they have simply ignored the Middleton findings to the contrary. So too, now, has the TRC report.

The KwaMakhutha massacre in 1987

On 21st January 1987 a house in KwaMakhutha, south of Durban, was attacked and 13 people were killed. According to press reports published at the time, a number of Inkatha leaders and supporters in the township had been attacked earlier in the month, and one had been killed. Those who died in the massacre on 21st January included the owner of the house, Mr Willie Ntuli-a member of Inkatha-and four of his children.

In 1996 General Magnus Malan, a former minister of defence, was charged (together with 19 co-accused) on 13 counts of murder arising out of the KwaMakhutha massacre. The prosecution alleged that the intended target of the attack had been Mr Ntuli's son, Victor. It described Victor, a 21-year-old UDF activist, as the owner of the house and said he had been planning to hold a UDF meeting there on the night of the attack. (A contemporaneous report in The Weekly Mail said Victor had not been staying at home for 'several weeks-ever since the start of clashes between the UDF and Inkatha'.) The prosecution further alleged that the killings had been executed by five Inkatha supporters, who had been trained by the SADF in the Caprivi strip (in Namibia) in 1986. It argued that the five (all co-accused with Gen Malan) had been trained to act in 'hit squads' against the ANC and UDF, and had used their training in the KwaMakhutha operation-which had been planned and initiated by Military Intelligence (MI).

The court acquitted all the accused. Mr Justice Jan Hugo found the three prosecution witnesses, all alleged accomplices in the massacre, unreliable. The state's main witness, the MI officer who had ostensibly planned the operation, was found to be a 'lying witness in certain respects and an unreliable one in others'. Overall, Judge Hugo found his testimony 'often contradictory, improbable, and absurd'. The court also indicated (without ruling to this effect) that the police Investigation Task Unit-responsible for marshalling the evidence against the accused-had coached witnesses, transferred portions of one witness statement to another by a computer 'cut and paste' method, and seemingly 'inveigled' a KwaMakhutha resident into giving evidence supporting the prosecution case. It had also, said Judge Hugo, sought to mislead the court on an important issue and 'probably deliberately' so.

The court was never presented, it seems, with the possibility that the attack might in fact have been aimed at Inkatha (as earlier attacks in KwaMakhutha had apparently been). The trial judge said that, though the five trainees accused of the killings were entitled to acquittal, the attack must have been conducted by some of the other Caprivi trainees, acting under the command of MI. This was no more than an obiter dictum, a statement made in passing. It was not a judicial finding on the culpability of the Caprivi trainees or the SADF. At another point in his judgement, Judge Hugo made this quite explicit. He said he was merely assuming-in order to weigh the state's contention of a broader conspiracy (see below) on the strongest basis it could possibly have-that 'the KwaMakhutha murders were committed by elements of the Caprivi trainees under the command or guidance of MI officers'. An assumption of culpability, made for the purpose of weighing potential guilt, is very different from a conclusion that guilt has been established.

According to the TRC, the court in the Malan trial 'found' that Caprivi trainees had been responsible for the attack on 'the home of UDF leader Mr Bheki Ntuli'. The commission presents one statement made in passing (which is in fact contradicted by another) as if it were a considered judgement. Relevant information regarding both the contradiction and the obiter nature of these statements is not provided. No reference is made to the numerous weaknesses in the prosecution case to which the trial judge had drawn attention.

The Caprivi training in 1986

Gen Malan and his co-accused were also charged with having conspired to murder supporters of the ANC and UDF in KwaZulu and Natal. The prosecution alleged that the Caprivi trainees had been given 'hit squad' instruction to equip them to attack the ANC alliance. As evidence of this the state cited, in particular, a number of secret military documents regarding the 'offensive' nature of the Caprivi training. The court found that the documents were capable of an innocent interpretation as well-and ruled that the prosecution had failed to prove its case beyond a reasonable doubt.

The TRC's finding regarding the Caprivi training is that 'the SADF conspired with Inkatha to provide it with a covert, offensive paramilitary unit (hit squad) to be deployed illegally' against the enemies of the former government and Inkatha.

The commission gives few reasons for its finding, stating that full reasons are contained in a 'lengthy document' available from the state archives. (No such document has been lodged with the state archives, however, and comprehensive reasons for the commission's ruling are thus not publicly available.) In its report, the TRC explains its effective rejection of the trial court's conclusion on the basis that:

Ø. the trial was 'based primarily on one incident' (the KwaMakhutha massacre);

Ø. 'no evidence was led to support the general conspiracy charge' that was added at a late stage;

Ø. the prosecution called insufficient witnesses (the court finding, among other things, that three named individuals should have been called as well);

Ø. the 'cut off date of the conspiracy charge excluded some of the most incriminating documents'; and

Ø. the accused were all 'poorly cross-examined'.

These reasons for rejecting the court's ruling merit assessment. During the trial, the conspiracy charges featured as much as the 13 counts of murder arising out of the KwaMakhutha massacre. Numerous documents allegedly revealing the offensive purpose and nature of the training were put before Judge Hugo and were dealt with by him at length. Indeed, Judge Hugo's description of these documents, the points arising from cross-examination on their contents, and his reasons for concluding that they were ambiguous as to the 'offensive' nature of the Caprivi training, ran to some 55 pages of his judgement.

In addition, the court made no finding that additional witnesses should have been called. According to the TRC, Judge Hugo was 'critical of the [prosecution's] failure to call a military expert and lead detailed evidence as to the nature of the training'. If the TRC is to be believed, Judge Hugo expressly 'found that Mr Luthuli, [another Caprivi trainee] Colonel van den Berg and Colonel Blaauw, [both SADF officers] should have been called'.

Judge Hugo made no such assertions. Instead-in analysing the weaknesses in the prosecution's case-Judge Hugo pointed out that he could have drawn an inference against the state for its failure to call witnesses who were available to it and who should have been able to buttress its contentions. He did not do so, however. He found this unnecessary because the prosecution's case, for a variety of reasons, was in any event too flimsy to succeed.

Moreover, the cut off date for the conspiracy charge in the trial was June 1989, while the TRC identified the 'time of the conspiracy' as April 1986. It is therefore difficult to understand how the cut off date usedin the trial could have excluded consideration of relevant testimony. That the accused in the trial were 'poorly cross-examined' seems unlikely, too. The judgement reveals that the accused were cross-examined at length. (In general, they proved more credible than the prosecution witnesses, it seems.)

Once more, the TRC has effectively repudiated an earlier judicial finding. It has failed to make public its full reasons for doing so-while the reasons it has cited do not stand up to scrutiny.

The assassination of Chief Maphumulo in 1992

Chief Mhlabunzima Maphumulo (a former president of the ANC-aligned Congress of Traditional Leaders of South Africa or Contralesa) was gunned down in the driveway of his home in February 1992. Thereafter, allegations that a 'hit squad' had assassinated him were widely publicised, and were buttressed by an apparent confession by one of his killers. A judicial inquest was conducted by Mr Justice N S Page. The 'hit squad' allegations were placed before Judge Page, and were found to be fabricated and untrue. Judge Page found that Chief Maphumulo had many enemies, any one of whom might have killed him.

According to the TRC, Chief Maphumulo (together with other community and political leaders) was 'targeted for attack in a planned hit-squad operation'.

The commission ignores Judge Page's earlier finding. It gives no evidence or reasoning for coming to a conclusion that, again, points to 'third-force' culpability.

Train violence on the Reef in the early 1990s

Murders on commuter trains on the Reef in the early 1990s stood out from many other political killings because of the terror they evoked. The attacks began in 1990 and persisted until 1993, striking panic into millions of commuters with no other choice of transport. So great was the fear engendered that people often jumped to their deaths from moving trains at the first sign of an impending attack. Providing full protection for the trains was beyond the resources of the police, while two lengthy inquiries by the Goldstone commission failed to identify the perpetrators of the attacks.

In July 1992 the Goldstone commission published an interim report on train violence on the Reef. It reported that-in the course of about 100 attacks on trains and stations-138 people had been killed and 261 injured in a ten-month period from July 1991 to April 1992.

Goldstone noted that most attacks took place during peak hours when trains were severely overcrowded and it was impossible to maintain proper access control or conduct searches for weapons. The police and other security services were too overstretched to exercise proper control over all trains at all times, and successful prosecutions had been extremely rare-in part because witnesses were reluctant to come forward. The commission found itself 'unable on the evidence before it to apportion blame' for the attacks.

In May 1993 the Goldstone commission published a final report on train violence. Since its previous report, 107 more people had been killed and 126 injured in train violence in a two-week period from late August 1992 to mid September 1992, while further attacks had also been mounted in October and November 1992. Since December 1992, however, train violence had declined.

The commission noted that increased security measures were being introduced on trains and at stations, with positive effects. It noted, too, that the police had deployed an additional 1 100 members, assisted by the SADF, to combat train violence-and that visible policing had been much improved as a result. However, 'it remained impossible for the police to be everywhere at the same time, especially during peak hours when attacks usually occurred'. The ANC alliance urged Goldstone to find the SAP at fault for its approach to train violence. The commission, however, found itself 'unable to support the contention that the police were not serious in their endeavour to put a stop to train violence'.

The Goldstone commission also noted the testimony put to it in camera by a former hostel resident. This individual alleged that 'attacks on trains were planned in the Nancefield hostel', and provided details of two such attacks, allegedly mounted in revenge for earlier killings of IFP supporters. Goldstone found some of his evidence reliable, but also ruled that it was not 'acceptable in all respects'.

Goldstone further noted, once again, the difficulties in countering, or marshalling evidence against, the perpetrators of train violence. The attacks were carried out at peak hours, when searching commuters was impossible. Attacks were swift and sudden and resulted in 'utter chaos', 'rendering the police, even if present on the scene, ineffective'. Witnesses were untraceable or unwilling to testify in court. The community had little confidence in the police, while some organisations were alleged to have discouraged their followers from making statements to the police.

Overall, the commission's findings on culpability for train violence were inconclusive. It reiterated that 'train violence could not be separated from the ongoing violence in the townships'. It noted that 'political rivalry between the ANC and the IFP for support at grass roots level had resulted in distrust, intolerance and enmity between their followers'. This, in turn, had 'led to violent confrontation' between hostel residents (regarded as IFP supporters) and township residents (seen as ANC supporters). There was, however, 'no evidence that either the ANC or the IFP propagated violence as a policy to achieve their ends'. Moreover, there was also 'no evidence that any organisation actively encouraged the perpetration of violence on trains'. Further, the commission was 'unable to establish whether train violence was aimed at achieving any political goal'.

The Goldstone commission also recalled a statement it had made in its interim report that 'when a group of attackers was identified, they turned out to be hostel residents'. It emphasised that there was 'no foundation for any finding that hostel residents were mainly re-sponsible for the attacks on commuters'. On the contrary, it was 'clear that attacks emanated from hostels as well as from surrounding townships'.

Goldstone also referred to allegations that a 'third force' was responsible for train violence. The commission had invited people with relevant information in this regard to come forward, but no one had done so. It therefore made no finding on the matter, one way or the other.

According to the TRC, about 600 attacks on train commuters took place between 1990 and 1993, and resulted in 572 fatalities. The commission noted that 'supporters of all political parties' fell victim to train violence, and said this 'seemed to suggest that train volence might have been aimed predominantly at causing general terror, rather than at achieving a clear, direct, political objective'.

The TRC went on, however, to assert that 'both local and regional IFP leadership were centrally involved in the authorisation and planning of train violence.' It based this conclusion on the in camera testimony earlier supplied to Goldstone, which had dealt in detail with two attacks alone. As further evidence it cited allegations by:

Ø. witnesses to an attack in which 62 people had been killed, who said the assailants had ran off 'towards' an IFP-supporting hostel and had spoken Zulu;

Ø. Mr Xolani Mnguni (who had been convicted of murdering a train commuter in 1991) and who said he had been acting on the orders of an IFP official;

Ø. an east Rand resident, Mr Paulos Nkondo, who said the assailants in a train attack he had survived had 'spoken Zulu'; and

Ø. a former askari who said 'hostel dwellers from the Nancefield Hostel were used at times in train attacks'.

The TRC also cited allegations that 'special forces' had orchestrated train violence; that policemen had thrown 'about five people' from trains in 1988; and that attacks had been planned and ordered from Vlakplaas, with train killers being paid 'R1 000 after successful operations'.

None of these allegations would qualify as corroborated and admissible testimony in a civil (let alone a criminal) court. Of the 600 or more attacks cited in TRC statistics, they encompass ten specific incidents at most-five of which occurred in 1988, outside the relevant period. On this basis, the TRC finds as follows:

The commission finds that train violence was initiated by groupings opposed to a democratic transition and the possibility of an ANC led government. The commission finds that, whilst train violence was not official IFP policy, a number of individuals and leaders within IFP structures were involved in train attacks. The commission finds that between 1990 and 1993, 572 people died in more than 600 incidents of train violence The commission finds that, in a number of incidents, IFP supporters collaborated with members of the SADF's special forces and members of Vlakplaas in planning train violence attacks The commission finds the IFP, SAP, and the SADF responsible for the killings that took place during train violence attacks and thereby the commission of gross human rights violations.

In concluding that train violence was 'initiated by groupings opposed to a democratic transition', the commission ignores its own earlier acknowledgement that victims came from all political persuasions and that attacks seemed to have no 'clear' political objective. Despite the absence of any new and compelling evidence, the TRC repudiates Goldstone's conclusion that 'there was no evidence that any organisation was actively encouraging the perpetration of violence on trains'. It overlooks his specific recognition that train attackers came from surrounding townships as well as from the hostels. And it rejects his finding that train violence was a spillover from general township violence which stemmed, in turn, from a variety of factors-including the 'political rivalry' between the ANC and IFP and the 'distrust, intolerance, and enmity between their followers' that this had generated.

These further examples could also, of course, be viewed as additional exceptions to the general accuracy of the TRC's report. The pattern that emerges from these incidents seems too clear, however, to be dismissed as insignificant, random, or coincidental. There are many other indications, too, as earlier described, of deep flaws that pervade the TRC's report.

Far from being 'strong on truth', as the commission has claimed, it has produced a report which distorts as much as discloses the truth. The full story about gross human rights violations in South Africa, and the violence that intensified as political and constitutional reform gathered momentum, has yet to be written.

This resource is hosted by the Nelson Mandela Foundation, but was compiled and authored by Padraig O’Malley. Return to theThis resource is hosted by the site.