Truth About the Truth Commission: [V] The Need For Factual Evidence

The Truth About the Truth Commission

by Anthea Jeffery



Foreword by John Kane-Berman,
SA Inst. of Race Relations (SAIRR)
Johannesburg, 1999

1. Foreword
   
2. Overview
   
3. Introduction
   
4. Publication Of The Current TRC Report
   
5. The Need For Factual Evidence
   
6. The Need For Comprehensive Findings
   
7. The Need For Violations To Be Contextualised
   
8. The Need To Accord With Established Legal Principles
   
9. Findings Based On A 'Balance Of Probabilities'
   
10. Appendix

V. The Need For Factual Evidence


The commission's founding legislation obliged it to compile its report on the basis of 'factual and objective information and evidence' received by it or otherwise placed at its disposal. How well did the TRC discharge this mandate-especially in making findings regarding culpability?

According to the TRC report, the commission 'based its conclusions on the evidence brought before it'. This evidence included:

• the statements of victims regarding gross violations of human rights (described in this study as victim statements);
  
• the statements made by applicants for amnesty (here identified as amnesty statements);
  
• the material gathered to corroborate both victim and amnesty statements;
  
• the historical documentation compiled by the commission; and
  
• the submissions made to the TRC by political parties and other organisations.

Of these various sources of information, victim and amnesty statements were accorded particular weight by the TRC. The evidentiary value of these statements thus merits an in-depth evaluation.



1. Victim statements


In drafting its report, the TRC gave 'priority attention' to the primary data, including the victim statements, that it received. Victim statements were regarded as particularly important in informing the work of the Committee on Human Rights Violations-which, in turn, was primarily responsible for investigating and making findings on the gross violations committed in the past. The commission as a whole, moreover, clearly regarded victim statements as vital to its understanding of the past. It considered these statements invaluable in revealing 'the truth about human rights abuses'. It also described the taking of victim statements by the human rights violations' committee as the 'primary information gathering activity of the commission'.

For present purposes, the key issue is the evidentiary value of these victim statements-which totalled some 21 300 in number and gave details of almost 38 000 alleged gross violations.

Various factors are relevant in this regard.


No oath required

Victim statements were taken by a number of statement takers, who were employed either by the commission itself or by certain non-governmental organisations (NGOs.) (See The gathering of victim statements, below.) At an early stage in the statement-taking process, 'it was decided to remove the demand for the statement to be made on oath'. This was because 'there was a potential for error in the process of its being written down by a statement taker'. Most victim statements were therefore not recorded under oath at the time that they were made.

Moreover, the volume of victim statements made it impossible for at least 90% of victims to testify in person at public hearings. Only those who did appear at such hearings had an opportunity to repeat their testimony under oath. It is also uncertain if as many as 10% of victims-amounting to about 2 130 individuals-were indeed able to testify in this way. (Some 70 public hearings were held for victims (see The focus of public hearings for victims, below) and lasted, in total, for about 185 days. On this basis, between 11 and 12 victims would have had to testify on every day of every relevant public hearing. It may not always have been possible for the commission to keep up this pace.)

Even if it is assumed that 10% of victims were able to give oral evidence and that all took the oath in doing so, it follows that the remaining 90% of victim statements were not made under oath. This means that at least 19 170 of the victim statements received by the commission had an evidentiary status lower than that of an affidavit.


Little, if any, cross-examination allowed

According to the TRC, it faced a fundamental dilemma in seeking to assemble evidence of gross violations. On the one hand, it was enjoined to offer victims of gross violations of human rights a cathartic opportunity to tell their stories of past suffering. On the other, it was obliged to base its report on 'factual' information and evidence.

The commission decided to resolve this dilemma by giving priority to its therapeutic role. It did not want to subject individuals who had already suffered greatly-many of whom who were disclosing for the first time the trauma they had experienced-to the indignity, the scepticism, and the hostile probing implicit in cross-examination. Instead, the TRC took pains to ensure that 'the interaction of the vast majority of victims with the commission was a positive and affirming experience'.

This meant that it made no attempt to cross-examine those victims who gave oral testimony before it, and generally accepted the veracity of their evidence unless 'there were glaring inconsistencies and falsehoods' in it.

This approach generated further difficulties, however. In particular, it required that the TRC have prior knowlege of an incident before it could assess whether a victim statement contained such flaws. On occasion, such knowledge was readily at the commission's disposal. At many other times, however, it was not.

Where an incident was well known, the commission's capacity to identify flaws in witness testimony was much enhanced. In April 1996, for instance, a witness told the TRC that she had counted no fewer than 175 graves after the police had opened fire on protesters in Langa township outside Cape Town in March 1960. Her evidence could not be squared, however, with other information regarding this police shooting-for a judicial commission of inquiry, held in its wake, had found that the police had killed two people and had injured a further 26.

In this instance, the discrepancy was easy to discern. The witness was not subjected to cross-examination, and her dignity remained intact. No harm was done, and the TRC found that three-rather than 175-people had been killed by the police in this incident. (This finding was also partially inconsistent with the earlier judicial inquiry which had found two people killed, and the TRC did not explain its reasoning in this regard. Elsewhere in its report, moreover, it recorded the number of deaths as two.)

In another instance-where the incident in issue was also well known-it is less clear whether the TRC succeeded in detecting and rejecting inaccuracies in victim statements. In 1985, a commission of inquiry under Mr Justice Donald Kannemeyer had probed a further police shooting-this time at Langa township on the outskirts of Uitenhage, near Port Elizabeth.Following an extended inquiry, Judge Kannemeyer had found that the police had killed 20 people and injured 27 when they opened fire, with sharp ammunition, on a crowd marching to a funeral. He had also rejected evidence that even more people had been killed and their bodies then concealed.

The TRC thus had an informed basis for assessing the testimony of witnesses who asserted that 34 people had been killed.

The commission nevertheless appears confused as to the number of fatalities. At one point in its report, it describes the shootings as having resulted in 20 deaths-thereby echoing Judge Kannemeyer's finding. At another point, however, it asserts that 43 people were killed in this same incident-a conclusion presumably reflecting what victims had recounted.

In many instances, however, the commission was unlikely to have had any prior knowledge of an incident. In circumstances such as these, the TRC's capacity to discern inaccuracies in victim statements seems far from clear.

An example may serve to illustrate the point. In October 1996, two witnesses-both family members of the deceased-told the TRC that a Northern Cape farmer, Mr Kobus Hanekom, had poisoned one of his workers, Mr Piet Scheffers. (Mr Scheffers had died in February 1993 after drinking a considerable amount on Mr Hanekom's farm.) An earlier inquest into Mr Scheffers' death was re-opened in apparent response to the TRC hearing, and the witnesses who had appeared before the commission were called to testify again. The presiding magistrate found they made a bad impression under cross-examination and contradicted themselves a number of times. He rejected their testimony as unreliable, and found the evidence too inconclusive to point towards the culpability of any person.

It is unlikely that the commission would have had any independent knowledge of an incident such as the death of Mr Scheffers. Hence, the TRC would not have been in a position, from its own prior information, to determine whether there were inaccuracies in the accounts of these two witnesses.

Relevant too, in this regard, was the sheer volume of the statements made to the commission. The 21 300 victim statements obtained by the TRC, as earlier noted, covered some 38 000 incidents extending over a period exceeding 30 years. This must have made it all the more daunting and all the more difficult for the commission-without resort to cross-examination-to detect the flaws, if any, contained in these many thousands of separate accounts.

Important too was the fact, as earlier noted, that only about 10% of victims were able to appear before the commission at public hearings. In 90% of instances, thus, the commission had no opportunity to observe the demeanour of these individuals as they gave their oral testimony. Nor could it probe, in even the gentlest way, for further details of their allegations. This, too, reduces the likelihood of the TRC having been able to detect inaccuracies in victim statements that proper cross-examination might have brought to light.

How many victim statements, if any, were subjected to cross-examination remains uncertain, for the commission provides no clear data in this regard. The right of alleged perpetrators to cross-examine victims at public hearings was, however, canvassed by the Appellate Division of the Supreme Court (now the Supreme Court of Appeal) at an early stage in the commission's operation. The case arose from the TRC's first victim hearing, held in East London on 15th April 1996. Two former police officers, who were to be implicated in the torture and assassination of an activist, Mr Siphiwe Mthimkulu, applied for an interdict preventing these allegations from being heard by the commission until it had given them full and proper notice of all evidence that might implicate them in gross violations of human rights. (See Audi alteram partem, below.) The Appellate Division ruled that they were entitled to such notice, and added that the commission might also be under a duty to 'permit immediate cross-examination' of the witness responsible for making the allegations.

The TRC remained anxious to avoid placing victims under this kind of pressure, however, and it seems doubtful if cross-examination was allowed in more than a handful of instances, if any. In general, moreover, victims and perpetrators were not invited to give evidence at the same hearing, and perpetrators thus had little opportunity to cross-examine their accusers. (The hearings into the Bisho shootings in September 1992, and into the conflict surrounding the incorporation of Moutse into the KwaNdebele homeland, were exceptions in this regard, for at these both perpetrators and victims were present.)


A 'low level' of corroboration

The commission took pains to emphasise that victim statements-though generally untested under cross-examination-were nevertheless corroborated through the seeking out of information from other sources that would tend to confirm their truth.

The human rights committee, for example, stated that it was 'the corroborated allegations of gross violations of human rights contained in 21 000 statements that formed the basis for its conclusions about the nature of past conflict'. The TRC as a whole, moreover, emphasised that 'all findings were made on duly corroborated evidence'. The reality, however, seems to have been rather different from what the commission averred.

Corroboration was primarily the responsibility of the Investigation Unit, and was secured in a variety of ways. Particularly important was 'a standard list of corroborative pointers'. Such pointers were defined as 'pieces of information or evidence concerning a particular act or event which might assist the Human Rights Violations Committee in establishing that the information provided by victims in their statements was true'. In practice, the pointers used included court records, inquest documents, death certificates, and newspaper clippings.

Various questions arise as to the adequacy of these methods of corroboration. A death certificate, for example, might confirm a fatality and record the medical reason for the death. In general, however, it would give little insight into the circumstances in which the death had occurred, and would certainly not identify the perpetrator of a killing. Newspaper clippings might confirm that a killing had occurred, but be unable to cast adequate light on the identity of the wrongdoer(s).

Inquest findings, in addition, might be inconclusive as to culpability. (For example, the inquest into the death of Mr Steve Biko, a leader of the Black Consciousness Movement who died in police custody in September 1997, found that 'there was no proof that his death had been brought about by an act or omission involving an offence by any person'.)

Corroboration was also obtained in other ways, but these seem equally problematic. It was sought, where possible, from interviewing the individuals who had made the statements in issue (the deponents) as well as other witnesses. It was also obtained by reference to the records of the former government and other archival material, the databases kept by other organisations, the transcripts of investigative hearings conducted by the Investigation Unit (under section 29 of the commission's founding legislation), and the submissions made to the TRC by political parties and other organisations. On occasion, it was acquired through consultations with 'organisations of the state and civil society'.

Interviewing deponents, however, might result in little more than a repetition of the information earlier provided in written statements. Corroboration normally requires confirmatory information from another source. The records of the former government might be incomplete in key regards, as might archival material.

Consultations with governmental bodies and NGOs might yield no more than hearsay. So too might the section 29 hearings generally held behind closed doors. Databases compiled by other organisations might be partial in their coverage and misleading in equal measure.

So great were the difficulties implicit in securing full corroboration that the commission, for the most part, did not even attempt this. Instead, it used two 'levels' of corroboration. A 'high' level would provide confirmation, by other witnesses present at the time, of 'the identity of the actual person committing the gross violation of human rights'. By contrast, 'a low level of corroboration would arise where the witness confirmed the event but not the identity of the perpetrator'.

In corroborating victim statements, the commission was generally content with securing a low level of corroboration. According to its Johannesburg regional office, 'the concept of low-level corroboration gained popularity [towards the end of 1996] as a way of fulfilling the commission's promise to do some investigation on every [victim] statement'. The Investigation Unit stated, too, that 'the commission required only a relatively low level of corroboration' as regards victim statements.

The purpose of low-level corroboration, the unit continued, was to enable the commission 'to make a finding that a person was a victim of a gross human rights violation, as described in the act'. Such a finding was important for two reasons.

First, it would help establish the 'extent' of the gross violations of human rights committed in past conflicts-a matter on which the commission was obliged to report.

Secondly, it would show that the individual in question was indeed a victim of a gross violation, as defined by statute, and would qualify, accordingly, for reparations in due course. Low-level corroboration was relevant, in short, to demonstrating victim status-not to identitying perpetrators.

Even a low level of corroboration, moreover, proved extremely difficult to obtain.

This stemmed from various factors, including:

• the sheer volume of victim statements coupled with the limited time available for their corroboration;
  
• a delay (until March 1997) in appointing the staff needed to work on corroboration, with the result that a 'tremendous backlog' developed; and
  
• the fact that many victim statements simply 'told a story' of past suffering, and contained 'no supporting documentation or other evidence'.

In addition, both the Investigation Unit and the Research Department, which played the primary role in corroboration, were burdened with many other responsibilities.

Skilled investigators were also in short supply, and had to cover incidents spanning more than 30 years and committed both within the country and abroad. Moreover, though the commission recognised (by February 1997) that it needed to give far more time to corroborating the 19000 or so victim statements that would never be heard at public hearings, it also found it difficult to shift its focus from convening such hearings to working behind the scenes on corroboration. This was partly because arranging hearings had become a familiar activity with its own inertia against change. It was also because of the commission's 'considerable concern that it would [then] become driven by technical rather than moral considerations'.

What complicated matters even more was that 'there were different understandings and conceptions as to what was meant by the term "low-level" corroboration. It was not clear exactly what level of information the commission needed in order to make a finding that a person was a victim of a gross human rights violation as described in the act'. This problem, according to the Investigation Unit, was overcome in time through the 'development of some corroboration "pointers"' and the provision of training.

Whether this difficulty was in fact resolved seems less clear than the Investigation Unit would allow. Confusion seems clearly to have continued-at least within the four regional offices of the commission-as to what low level corroboration entailed. The Cape Town office reported that 'despite the call for standardised national procedures, interaction between the regions was poor, and each region tended to develop its own system'. Moreover, 'because the national Human Rights Violations Committee did not set guidelines on levels of corroboration, the process of investigation devolved on the regional investigation units themselves'. This gave rise to 'regional variations in the investigative process'.

The regional offices-applying as they did their varying approaches to corroboration-nevertheless played a crucial role in making the commission's findings of accountability. Regional offices were responsible for making 'pre-findings' in this regard. These were based on victim statements, as well as the 'corroborative material gathered by the investigators and the background research material provided by the researchers'.

'After a pre-finding had been made at a regional level, it was ratified at national level.' In general, such ratification took place without additional checking or verification. Only 10% of regional pre-findings, chosen on a random basis, 'went through a national check, to ensure that regions were operating on the same criteria so that findings would be uniform, and also to double-check for possible mistakes'.

In 90% of instances, regional pre-findings were accepted at face value-even though the victim statements on which they were primarily based might have been corroborated in an inconsistent manner, and would generally have received a low level of corroboration only.

These various difficulties led the TRC to emphasise the 'enormity of the [corroboration] task'.So great were the obstacles, in fact, that the commission was also compelled to acknowledge that it had encountered 'virtually insurmountable practical difficulties' in corroborating victim statements.

These problems, it continued, had served to 'crystallise' what it regarded as a 'profound dilemma'. On the one hand, it was 'a legal institution with the responsiblity for making defensible findings according to established legal principles'. This was essential 'both to safeguard the credibility of its final report and to ensure that those who received reparations were genuinely victims as defined in the act'. On the other hand, it also 'embodied a moral and therapeutic process that aimed at acknowledging suffering and giving victims an opportunity to tell their stories'. 'This aspect of its work', it said, 'would have been greatly diminished had the findings process been approached in too technical a manner, focusing narrowly on rules of evidence and requirements of proof.'

The commission fudges the issue of how it resolved this apparent dilemma. 'In general,' it stated, 'it sought to be both therapeutic in its process and rigorous in its findings, but sometimes the effort to satisfy one objective made it more difficult to attain the other.' The TRC implies thereby that, on the whole, it managed to do both. What is more likely, however, is that it sacrificed the rigour necessary for making findings in order to spare the victims of past abuses the indignity of cross-examination or of having in other ways to substantiate their statements. The commission acknowledges, moreover, that it was willing, at times, to make a finding 'whenever the circumstances allowed this, and even where available information was extremely scanty'.

The commission's approach, to some extent, is readily understandable. 'Nobody who listened to the victims' hearings could fail to be moved by the testimony of people who had been abused or tortured or of families who had lost a son, a daughter, a brother, a sister, a husband or a wife.' The tales told at these hearings were heartfelt and heart-rending. They may not, however, have been wholly accurate in identifying perpetrators nor in assessing their accountability for previous wrong-doing. Moreover, no matter how well-founded a deep-rooted sympathy for victims might have been, it could not obscure the fact that the dilemma described by the commission was one largely of its own making.

The TRC's primary obligation under its founding statute was to base its findings on 'factual and objective' evidence. If it did not want to subject victims to cross-examination or other means of objectively verifying their statements, it should have left victim statements out of account in making its findings of accountability. The proper solution to its apparent dilemma (which it failed to follow) was not to discount the crucial process of verification but rather to omit from its assessment of culpability any evidence that was not demonstrably 'factual and objective'.

The commission seems, thus, to have missed a crucial point. It may have had a sufficient basis-based on a low level of corroboration of victim statements and sometimes on 'extremely scanty' information-for a finding that a particular incident qualified as a gross violation of human rights under the relevant definition contained in the act. It may then also have had a sufficient basis for a related finding that the individuals who had suffered this violation were victims entitled to reparation. However, since low-level corroboration did not encompass the identity of perpetrators-let alone their culpability for alleged wrong-doing-it is doubtful in the extreme whether the commission could validly use victim statements as a foundation for findings of accountability for conduct as heinous as torture and premeditated murder. Cursorily corroborated victim statements might have sufficed for the first two purposes. They could not properly be used for the third as well.


Hearsay not excluded

A further key question is how many of the 21 300 victim statements received by the commission were based on hearsay testimony. Some witnesses made it clear that they would talk only of their personal experiences, and would tell the commission the truth 'as they had seen it'. Others-according to a minority report submitted by Mr Wynand Malan (a former NP politician who later became a co-leader of the Democratic Party)-were 'often not present at the actual violations to which they testified and their stories were accounts of what they had been told'. These accounts might well have seemed quite accurate. They could equally well, however, have incorporated significant misunderstandings and mistakes.

The dangers in hearsay are legion. For example, the original eyewitness-whose experience is now being relayed-might not have been able to see the alleged perpetrators very clearly, especially at night or from a distance. He might have been mistaken as to a complex sequence of events, while the lapse of time might have blurred his memory yet further. The overall surrounding circumstances-relevant to motive and intent-might not have been apparent to him at the time, especially if he were caught up in the immediate events. He might have made deductions as to culpability without access to the further information that would have shown these to be unfounded.

Hearsay is subject to all these difficulties, and many more. Established legal principles thus incorporate strict rules against the admission of any evidence that is not based on personal experience and cannot be tested through cross-examination. Exceptions to the general rule that hearsay evidence must be excluded are relatively few.

The TRC, however, does not acknowledge that it might often have relied on hearsay-and that this might have distorted its understanding of the past. It is only in the minority report of Mr Malan that the issue is raised. According to Mr Malan, hearsay evidence was admitted 'often'. The rest of the commission, in its rejoinder to Mr Malan, takes issue with certain of his criticisms but is silent on this point.

The TRC itself elsewhere implicitly acknowledges that hearsay may well have been common in victim statements. According to the report, deponents testified not only about human rights violations suffered by themselves but also about those experienced by others. In about 18 500 cases, deponents told of their own experiences of violations. In some 17 500 instances, however, deponents reported to the TRC gross violations committed against people other than themselves.

Since more women came forward to give statements than men-and since more men had suffered gross violations than women-it so happened that many of the deponents who gave evidence about the violations committed against others were women. The TRC puts it thus: 'Men were the most common victims of violations. Six times as many men died as women, and twice as many survivors of violations were men. Hence, although most people who told the commission about violations were women, most of the testimony was about men. Most men who came to the commission reported violations they [themselves] had experienced, whereas women tended to talk about violations experienced by others.'

Many of the individuals who gave evidence about the gross violations experienced by others may simply have been recounting what they had heard about these events. This possibility is illustrated by the example of Mrs Sylvia Dlomo-Jele, whose youthful son, Sicelo Dlomo, was killed on the outskirts of Soweto in January 1988. A well-known UDF activist, he had earlier told his mother that he was likely to die soon, as the police were after him. When his body was discovered with a bullet hole in the head, she remembered his words and concluded that the police had killed him. She gave evidence to this effect to the commission in 1996. Her evidence was clearly based on hearsay, however, and not on what she herself had witnessed.

In 1999 she was shown to have been mistaken. Four former ANC cadres, all comrades of her son and frequent guests in her home, applied for amnesty to the commission. They explained at an amnesty hearing in February 1999 that it was they-and not the police-who had shot Sicelo dead as a suspected informer. Mrs Dlomo-Jele's hearsay testimony, proffered in all sincerity to the commission, had clearly been inaccurate.

Many of those who testified about the 17 500 or so violations experienced by people other than themselves may have been recounting hearsay too. Many may also have been eyewitnesses to the events in question. But the character of the violations in issue must also have made eyewitness testimony comparatively rare. Torture, by its very nature, would have taken place beyond the sight of relatives, friends, or other witnesses. When killings occurred, eyewitnesses would not necessarily be present.

Even when they were, it would still, in many instances, have been difficult to identify the perpetrators.

Some examples of the many ways that people died may illustrate the point. People died, for example, in confrontations with the police in which perpetrator identity might be readily ascertainable. But people also died from petrol bombs lobbed through bedroom windows at night by assailants who were never seen. People died in 'drive-by' shootings where the assassins could scarcely be glimpsed before they disappeared. People were killed as 'collaborators' or 'informers' by crowds of people so large and so enraged that it was difficult to know who had done what, or with what level of personal accountability. People died in massacres, where attackers came suddenly out of the darkness and as quickly disappeared again.

Individuals who told the TRC about the deaths of others, in situations such as these, could offer no accurate eyewitness evidence as to the perpetrators of these killings. Whatever they said in this regard was likely to be based on hearsay, rather than on personal knowledge. The commission should have recognised this likelihood. And when it came to assessing accountability for gross violations, it should have treated testimony about these 17 500 or so incidents with greater circumspection than it seems to have evinced.

Hearsay was also clearly used by the TRC at times to substantiate its findings regarding culpablity. It was used, for example, to found its conclusion that the security forces had assassinated Mrs Victoria Mxenge in August 1985. The evidence it gathered (from witness statements, it would seem) was that:

• a former askari, Mr Jimmy Mbane, had said in a statement to the TRC that another askari, Mr Thabiso Sphamla, had confessed to him, while drunk, that he (Mr Sphamla) and three further askaris had killed Mrs Mxenge;
  
• Mr Pat Hlongwane, who had been imprisoned in the ANC's Quatro camp in exile, had stated that his cell mate, Mr Marvin Sefako (alias Mr Bongani Malinga), had told him that he had killed Mrs Mxenge 'on orders from Captain Dirk Coetzee of Vlakplaas';
  
• the ANC, in its second submission, had asserted that Mr Sefako had been recruited by the security police in March 1985 or earlier, and was in training in the use of firearms and poisons from March 1985 to September 1985; and
  
• Mr Sefako had confessed-while imprisoned in Quatro, it would seem-that he had shot Mrs Mxenge five times and thereafter 'followed her with an axe and chopped her next to her dining room door'.

None of this evidence would stand up in a court, either civil or criminal. Mr Sefako himself, moreover, could not be questioned for he had been killed in 1991 following the lifting of the bans on the ANC and his return to South Africa. (His killing is attributed to the IFP at one point in the commission's report, and to the ANC at another.)


The TRC's finding on Mrs Mxenge's death is as follows:

The commission finds that Mrs Victoria Mxenge was killed by, or on the orders of, unknown members of the security forces, and that her death was a gross human rights violation which entailed deliberate planning on the part of the said security forces.

Mrs Mxenge may in fact have been killed by the security forces. But the commission could not properly make a finding to this effect on the basis of this hearsay (and contradictory) evidence. The TRC, shows no hesitation, however, in basing its finding on so shaky a foundation. It even acknowledges that a key aspect of the hearsay evidence in issue (the alleged confession) may have been obtained under duress-but it accords this scant regard as well.


The gathering of victim statements

The commission averred, at various points in its report, that the statements it had received from victims were 'self-selected'. They were made, it declared, by those who had sought it out in order to recount the violations they had suffered. Such self-selection could clearly have coloured the commission's understanding of the past. The commission acknowledged this, conceding that a victim sample confined to individuals who had sought it out would have tended to exclude:

• those who lived very far away from any of the commission's offices;
  
• people who were too old, sick, or depressed to make the effort to depose;
  
• those who were already dead;
  
• individuals with no access to the media and hence no knowledge of the commission's work; and
  
• people 'from constituencies hostile to the commission'.

The commission recognised as well that the last of these factors had resulted, in KwaZulu-Natal, in its receiving many times more statements from ANC supporters than from members of the IFP. It said this had 'created the impression that the violations suffered by the UDF/ANC outnumbered those suffered by Inkatha by five to one'. It added that it was 'unable to establish the degree to which the disparity was a reflection of the IFP's rejection of the commission, or a reflection of the actual experience of violations'. (It did not, however, allow its uncertainty in this regard to hinder it from subsequently finding the IFP the principal perpetrator of politically motivated killings in KwaZulu-Natal and elsewhere.)

The problem goes deeper, however, than a 'self-selection' of victim statements. The commission is also less than frank in this regard. For victim statements, it appears, were often less 'self-selected' than pre-selected by the TRC itself. Such statements were not solely spontaneously volunteered to the commission, as it takes pains to emphasise. Often they were sought out by statement takers who had been briefed by the Research Department as to what incidents to cover-and as to whom to track down and interview.

This emerges clearly from the TRC's report on the activities of its Research Department. This department began its work by holding a series of workshops in early 1996, in the 'first months of the commission', in the geographical areas covered by the TRC's Cape Town, Durban, East London, and Johannesburg offices. These events 'brought together a range of community-based people, historians, journalists, human rights activists, and others'. The purpose of these workshops was to 'identify gross violations of human rights that had occurred in the area-including events both well-known and documented, as well as lesser known events in danger of being lost to public memory'. The identity of those who attended these workshops is not disclosed in the commission's report, but would clearly have had an important bearing on the understanding of past violations that emerged from these discussions-especially if the majority of those present shared the view that a police/IFP 'third force' was primarily to blame for the political killings committed in the past.

The outcome of these workshops, said the Research Department, 'was the beginning of a national chronology and four regional chronologies'. These provided a 'preliminary overview of the 34 years under review by the commission'.'The chronologies � provided a framework for the information gathering work of the commission.' They were 'substantially developed', thereafter, as a result of 'statement taking, human rights violations hearings, and amnesty applications' and were used in 'the corroboration and investigative phases of the commission's work as well as in the findings process'.

Particularly important for present purposes is that these chronologies were also used to brief the statement takers appointed by the commission to record the statements of victims. In the Cape Town office, for example, the Research Department 'supplied statement takers with a chronology of political events and a brief account of documented cases of gross human rights violations-giving them a useful point of entry. In addition, workshops were held for statement takers and local NGOs [non-governmental organisations] and CBOs [community-based organisations] before statement takers worked in a sub-region. These workshops helped further familiarise statement takers with political events and with the people in the community who had been involved in these events'.

Towards the end of 1996, moreover, statement takers adopted 'a more pro-active strategy', in which they sought out potential statement makers rather than waiting to be approached. CBOs were thus, for example, 'asked for information in order to locate potential deponents', while 'the voters' roll was used to try and establish their current whereabouts'.

In other regions, statement takers also shifted from the passive receipt of statements to the overt canvassing of these. The Eastern Cape office adopted a 'more proactive strategy' after it became clear that 'few people were coming forward to make statements'. In the Johannesburg region, various local organisations contributed statement takers who 'proved invaluable in identifying and reaching victims in both urban and rural areas'. In the region encompassing KwaZulu-Natal and the Free State, the same shift apparently took place-some individuals 'approaching the commission of their own accord', while others were sought out by the statement takers 'deployed across the province'. (The commission implies that statement takers in this region sought to approach both ANC and IFP supporters, but were inhibited in their endeavours to be even-handed by the political hostility they encountered in some areas of KwaZulu-Natal.) Overall, what proportion of victim statements was obtained through pre-selection is not explained by the commission.

Statement takers, it thus appears, were not only told what events were particularly significant, but were also in time mandated to seek out deponents who would give evidence regarding such events. Moreover, as discussed below (see National and regional chronologies), the national chronology at least left out a number of significant events-including various massacres of which IFP supporters were the victims. The seeking out of certain deponents to testify about events reflected in this chronology would clearly have influenced the content of the statements received by the TRC, as well as the thrust of its subsequent findings, and would have done so far more than any 'self-selection' would have done.

The commission is also less than frank in this regard. It adverts to (and glosses over, especially in KwaZulu-Natal) the problems implicit in self-selection. But the more disturbing difficulties arising out of pre-selection are not acknowledged at all. The fact that statements were frequently sought out can be gleaned, moreover, only from the odd snippets the commission lets fall from time to time, in different parts of its report. Only a most careful reading of its 3 500-page account-a process handicapped by the absence of any index-makes it possible to piece together the story of how statements were in fact sought out from chosen individuals.


The focus of public hearings for victims

The Committee on Human Rights Violations conducted a considerable number of public hearings for victims. Most of these hearings were intended to give individuals who had experienced gross violations of human rights the opportunity to 'relate their own accounts' of what they they had endured. They were commonly referred to by the TRC as 'victim' hearings. Other hearings involved the testimony of victims in their capacity as witnesses to particularly important events-and were described by the commission as 'event' hearings. They 'focused on specific events in which violations had occurred', and their purpose was to act as 'window cases' that provided 'detailed insights into particular incidents that were representative of broader patterns of abuse'. 'Theme' hearings were also held, to canvass the sufferings of victims as representatives of wider groups-women, youth, and those subjected to compulsory military service. ('Institutional' hearings were conducted in public too. These focused primarily on 'organisations, rather than the individuals within those organisations'.)

Leaving aside these institutional hearings, some 70 public hearings were held for victims (either as individuals recounting their own stories, or as people caught up in significant 'window cases', or as representatives of wider groups). The focus of these public hearings merits examination. This focus, as described below, suggests that the commission may have been selective in the incidents it canvassed-demonstrating a particular interest in bringing certain events to public attention, and less concern about probing other incidents that seemed equally significant.


'Victim' hearings

'Victim' hearings, as they became known in TRC parlance, were intended (as noted above) to allow individuals to tell their own stories of their sufferings. Though they were conducted all over the country, they were organised by each of the commission's four regional offices.In the Western Cape, victim hearings canvassed, among other things, 'the ambush and killing of the "Guguletu Seven" by the South African Police (SAP); the death of the first detainee to die in police custody; indiscriminate shooting by police of civilians in the towns surrounding Kimberley; human rights violations committed by kitskonstabels or police assistants; the torture of young teenage activists (who 'had their testicles, penises, or breasts slammed in drawers'); actions of the Amasolomzi vigilantes, supported by municipal police, in the Boland; the 'recurrent shooting and killing of youths by the police and torture in police cells' in various Boland towns; the killing of an Umkhonto cadre, Mr Ashley Kriel; a further instance of torture in which a young woman's breast was slammed in a drawer in the Beaufort West area; a further incident of torture in the Karoo; and the killing of activists by the security forces in a 1985 cross-border raid.Victim hearings in this region also canvassed the St James' Church massacre, perpetrated by cadres of the Azanian People's Liberation Movement (Apla), the armed wing of the PAC; the killing of a community councillor near George; the burning of residents in Beaufort West who 'did not support the comrades'; the killing of a policeman in Upington and the trial of the 'Upington 26'; attacks on informers, community councillors, and police officers in the Colesberg area of the Karoo; and 'clashes between the UDF and the Azanian People's Organisation (Azapo)' in the Boland.

Although other issues were canvassed as well, the preponderance of the commission's focus fell on the conduct of the former security forces, particularly the SAP. The Cape Town office acknowledged, moreover, that it had been criticised for 'showing a bias towards investigations and hearings on violations committed by the security forces, rather than those committed by the liberation movement'. It justified this on the basis that '90 per cent of statements had demonstrated the involvement of security forces in human rights violations'. This does not seem sufficient an explanation, however-especially since many of the victim statements in issue were, by the TRC's own admission, not spontaneously received but rather deliberately sought out by the statement takers it deployed. (See The gathering of victim statements, above.)The same pattern seems evident in other regions too. In the Eastern Cape, the main focus was on killing, torture, and shooting by the SAP and other homeland forces. In the Johannesburg region, the primary emphasis was on shootings and other killings by the SAP and homeland security forces; by vigilante groups such as the Mbokotho (which had been formed in 1986 in the former KwaNdebele homeland to 'deal with people who enforced boycotts and handle "troublemakers"'); and by politically motivated gangs such as the 'IFP-aligned' Toaster Gang in Tembisa (on the east Rand). Hearings in this region also canvassed the killing of activists through booby-trapped hand grenades; the role of the former security forces and the IFP in hostel and train violence; and conflict between the ANC and the IFP.

In KwaZulu-Natal, hearings canvassed primarily the culpability of the IFP and/or the former security forces in violence, including the killing of 15 people at a memorial service for Mrs Victoria Mxenge; the murder of prominent trade unionists in Mpophomeni (near Howick); the massacre of 11 people by IFP supporters at the Hlobane mine; the assassination of Dr Rick Turner; the killing by the IFP of 'hundreds' of people near Pietermaritzburg in the 'Seven Days War'; and the deaths, at the hands of IFP supporters, of some 35 ANC township residents in two massacres in Bruntville (outside Mooi River in the Midlands region). Particular emphasis was placed on the 1986 training, allegedly as 'hit squads', of 200 IFP supporters in the Caprivi strip in Namibia by the former South African Defence Force (SADF), as further described in due course. (See Ignoring other rulings too, below.)

Other issues were also canvassed in the course of these regional hearings. They focused, for example, on the role of an ANC leader, Mrs Winnie Madikizela-Mandela, and her Mandela United Football Team in violence in Soweto in the late 1980s.

Overall, however, the preponderance of attention was clearly placed on the former security forces and their alleged allies, the IFP.


'Event' hearings

A similar pattern emerges as regards the 'event' hearings that were intended, as noted above, to provide a window on to 'particular incidents that were representative of broader patterns of abuse'. Again, event hearings were organised by each of the commission's four regional offices.

In the Western Cape, these hearings focused on:

• the 'ambushing and killing of the Guguletu Seven cadres', with a special emphasis on 'the Vlakplaas connection'-the evidence that police officers from the Vlakplaas unit had been involved in the planning of the incident;
  
• the 'Trojan Horse' incident in which three youths had been killed by police concealed in the back of a van, and in which 'evidence was led to show that the police were not reacting, but deliberately set out to provoke unrest in order to kill'; and
  
• events in KTC, an informal settlement near Cape Town, where the emphasis was placed on 'witdoek and police complicity in an attack on the KTC community'.

According to the commission's report, these event hearings were particularly significant because they 'confirmed long held beliefs about the role of the state in fomenting violence (as in KTC), the involvement of the police in provoking unrest in order to kill (as in the "Trojan Horse" incident), and the involvement of security forces based at Vlakplaas in the Western Cape (as in the "Guguletu Seven" incident)'. They thus 'provided a window into understanding human rights violations' in the 1980s.

The commission implies that these incidents were indicative of all the human rights violations committed in the 1980s. These cases, however, all focus on security force involvement in violations. They could provide a 'window' on to wrongdoing of this kind only. They offered little insight into the possible role of the liberation movements in political violence. And, by being posited as representative of all abuses from this time, they helped to obscure and conceal this latter aspect of past conflict.

Other event hearings conducted by the commission canvassed, among other things, the Soweto revolt in 1976; the 'six day' war in Alexandra (outside Johannesburg) in 1986; the conflict that attentended the incorporation of the Moutse district into the KwaNdebele homeland in the mid-1980s; the killing of farmers in the former Transvaal; the 1990 'Seven Days' War' in Pietermaritzburg; the training of IFP supporters by the SADF in the Caprivi in 1986; the Pondoland rebellion in 1960; and the Bisho massacre in 1992.

Of these eight further event hearings, almost all pointed to the culpability of the former government, the former security forces, and/or the IFP. Only one, the hearing into the killing of farmers in the former Transvaal, seemed likely to throw light on the possible role in political violence of the liberation movements.


'Theme' hearings

'Theme' hearings canvassed the sufferings of victims, as representatives of particular groups. The hearings focused on women; children and youth; and compulsory military service. Hearings for women canvassed the extent to which they had suffered sexual and psychological abuse, as well as other forms of torture. They were also intended to 'end the silence around the gendered nature of apartheid atrocities', and to show, for example, that the widows of activists had been arrested and harassed as well.

Hearings for children and youth probed their detention, torture, and killing at the hands of the former security forces. They also canvassed the psychological effects on them of exposure to apartheid and violence. In Durban, for example, children 'affected by violence were given the opportunity to express themselves through art and drama workshops'.


The overall focus of these hearings

In conducting its public hearings for victims falling into these three categories, the commission seems to have devoted considerable energy to examining and re-examining certain issues, such as the Trojan Horse incident and the violence in KTC, or the training of 200 Inkatha supporters in the Caprivi as 'hit squads' against the ANC alliance. Its public hearings appear to have omitted, however, an equivalent focus on other important developments-all of which would seem essential to understanding the conflicts of the past. These included, for example, attacks on black policemen, the targeting of black councillors, the coercion that accompanied mass action campaigns, the necklace executions of hundreds of individuals, and the deaths of hundreds more in bomb attacks of various kinds. Submissions made to the commission had drawn the TRC's attention to all these issues, yet public hearings canvassed them only in outline-conveying very little of the impact they had had on township life. (See also Events and issues not investigated, below.)



2. Amnesty statements

The key question here is the extent to which amnesty statements were corroborated, cross-examined, screened for inadmissible hearsay testimony, and verified in general. Each of these issues merits separate consideration.


Inadequate corroboration

According to the TRC report, amnesty statements were corroborated. For this purpose, 'the Investigation Unit was asked to obtain police dockets and other relevant information from institutions like the National Intelligence Agency, the South African Police Service, and the Department of Justice. In certain instances, evidence leaders and analysts interviewed individuals, applicants, and/or victims to corroborate information contained in particular submissions'. In addition, 'use was made of information gathered by the Research Department and the Investigation Unit or contained in submissions made by political organisations and liberation movements. The section 29 in camera hearings were another source of information used to verify and corroborate information provided in applications'.

The commission does not explain what level of corroboration-high or low-was either sought or obtained. Nor does it acknowledge that these methods of corroboration may not have been sufficient.

The statements in police dockets, for example, may not even be sworn affidavits and are not necessarily true. They cannot be accepted at face value without further verification. Police records, moreover, might prove prior criminal conduct, but would not demonstrate culpability in a further specific incident. Interviewing deponents and other individuals would clearly be no substitute for proper cross-examination, while the submissions made by political organisations might reflect no more than hearsay or opinion. The secret testimony given at section 29 hearings might also be mistaken, while its accuracy would be immune from public scrutiny and public confirmation.

A further problem, said Mr Malan in his minority report, was that many amnesty applicants implicated or put the blame for their wrongdoing on individuals who were already deceased. This meant that the most important corroborative evidence of all-that of an alleged co-perpetrator-was frequently beyond all reach. The commission, in its rejoinder to Mr Malan, takes issue with his views on many other matters but is silent on this point.


Insufficient opportunity for cross-examination

Amnesty statements were different from victim statements in a crucial respect. In terms of the commission's founding legislation, those who sought amnesty for gross violations of human rights could be cross-examined on the content of their amnesty statements. This was to be achieved through public amnesty hearings, where victims and others with an interest in the matter could challenge what was said by such applicants, and adduce their own conflicting evidence as well. The veracity of these amnesty statements could therefore be tested in much the same way as in court. This should have made amnesty statements a particularly important source of tested evidence on which the commission could draw in making findings of accountability.

Various factors, however, served to limit the extent to which amnesty statements were actually cross-examined.

First, comparatively few of the amnesty statements received by the commission qualified at all for public hearing-and hence for the cross-examination that this would entail. In all, 7 127 amnesty applications were received by the commission. Most of these, however, were not eligible for the granting of amnesty at all, or did not involve gross violations of human rights, as defined in the TRC's founding legislation. Some 1 700 did deal with gross violations, but it was not clear at the time the commission compiled its report that all would qualify for public hearing. (Some 350 were still awaiting further particulars that might, once received, have excluded the applications from further consideration because, for example, no political motive for misconduct was disclosed.) Thus, at the relevant cut-off date, only 1 341 applications had been identified as qualifying for public hearing. Of these, only 102 had proceeded through a public hearing, while 1 239 remained still to be heard.

In these 1 239 instances, the statements made by applicants had not yet been subjected to cross-examination and could not provide the tested evidence required. It was only the 102 amnesty statements that had proceeded through a public hearing-and had seemingly been confirmed as accurate through the granting of amnesty-that could be taken to contain the substantiated evidence required for conclusions regarding culpability. These 102 amnesty statements comprised no more than 1.4% of the total applications received, and less than 8% of those qualifying for public hearing.

The dangers of relying on the untested and unsubstantiated evidence of an amnesty applicant are clear, of course. In the criminal courts, such evidence would constitute 'accomplice' evidence-and would be treated with particular caution. This is for three main reasons. First, an accomplice, by definition, 'knows what happened when the crime was committed and can give a credible sounding account of it. By the same token, however, it is easy for the accomplice to swop role players and to attribute to one individual deeds that were possibly committed by someone else or even by the accomplice himself'.

Secondly, the accomplice witness has an 'obvious temptation to tell the police what he perceives they want to know'. Thirdly, accomplices are 'self-confessed criminals who are attempting to gain some benefit by testifying'.

Amnesty applicants before the commission, like accomplice witnesses in the criminal courts, could use their knowledge of past violations to implicate others, particularly those who had allegedly instructed the commission of such abuses. The applicants could do so with seeming credibility, though not necessarily with honesty or accuracy. They would also have an interest in telling the TRC's investigators what they thought those investigators might want to hear, for this would increase their chances of obtaining amnesty. Moreover, the primary interest of the applicants would clearly be to escape the punishment that otherwise would lie in store for them.

Many applicants, too, sought amnesty for crimes as serious as murder. They had already, by their own admission, demonstrated scant respect for human life. They might thus find the telling of a plausible lie a relatively trivial matter-especially if in doing so they could save themselves from life or other long-term imprisonment. Some were also likely, as the TRC itself acknowledged (albeit in a different context), to be 'highly skilled operatives, trained in the art of concealing their crimes'.

The commission also conceded that, when amnesty statements were put to the test in public hearings, discrepancies sometimes emerged between what the applicants had originally stated and the oral testimony they subsequently gave. It noted, thus, that 'perpetrators recounted versions of events that were sometimes different'. (How often this occurred is not explained.) It glossed over the significance of this, criticising these discrepancies mainly because they 'led to confusion and anger on the part of victims' families and the wider public'. The more disturbing implication is that these amnesty applicants were likely to have been lying on one or both occasions.

The commission admitted, too, that some of the individuals who claimed to possess important information regarding past violations (and who, on this basis, then sought the benefits of its witness protection programme) were nothing but 'confidence tricksters'. 'Often motivated by financial enrichment,' said the TRC, 'these [potential witnesses] wanted to mislead the commission by falsely professing knowledge of cases under investigation. Such misrepresentation was easily achieved because of the media publicity accorded the cases over the years, the absence of independent eyewitnesses, and the destruction of official documentation.' The commission seems confident that all these people were identified and dealt with. It fails to acknowledge the risk that some amnesty applicants might also have been 'confidence tricksters'-who would have been assisted in putting forward false testimony by the very same factors.

The 1 239 untested amnesty statements that remained unheard on 30th June 1998 constitute, accordingly, a most uncertain basis for any findings of accountability.

The commission seems nevertheless to have taken pains to ensure that these untested allegations would be available to it when it came to compiling its report. Thus, when it realised that these amnesty applications could not be finalised before the cut-off date, it instructed the Research Department to embark on 'Operation Capture'. 'This involved reading all amnesty applications with a view to categorising these into themes and identifying and accessing relevant material for the final report.'

How many of the unheard amnesty statements encompassed within Operation Capture were used to make findings of accountability is difficult to gauge. It seems clear, however, that the TRC relied on at least some of this untested data to support its findings. This is evident, for example, as regards the amnesty statement submitted by Colonel Eugene de Kock, a former police officer and erstwhile commander of a 'counter-revolutionary' police unit stationed at Vlakplaas, outside Pretoria.

Col de Kock had been convicted in 1996 on charges that included five counts of murder, one count of culpable homicide, and 61 of fraud. He had been sentenced to two life sentences and a further 212 years' imprisonment. At his trial, Col de Kock had given no evidence on oath. Instead, he had made a lengthy statement from the dock in mitigation of his sentence. In this, he had placed the blame for his nefarious activities on various former police generals as well as two former state presidents, Messrs P W Botha and F W de Klerk. Much of his court statement had consisted of hearsay. He had also told the trial judge that he was contemplating applying to the TRC for amnesty, and had stated: 'My only strategy is to keep alive. I find myself in a steel vault with no passages and no turns.'

Col de Kock subsequently submitted to the TRC an amnesty application running to 4 000 pages and encompassing some 140 incidents. His application had not been heard at the time the commission compiled its report. The TRC nevertheless seems to have relied, in making certain of its findings regarding the accountability of the SAP for extra-judicial killings, not only on the untested content of his amnesty statement but also on the hearsay allegations contained in his autobiography, A Long Night's Damage.

The commission cites, for example, a passage from this book in which Col de Kock 'stated that he was instructed to "make a plan" in respect of Mr Dirk Coetzee by Brigadier Nick Janse van Rensburg'. (This prompted Col de Kock to post Mr Coetzee a Walkman music casette player containing a concealed bomb. This, in February 1991, ultimately instead reached and killed a Johannesburg lawyer, Mr Bheki Mhlangeni.)

Col de Kock's statement implicating Brig van Rensburg is hearsay. So too is his amnesty statement that it was 'the head of the Komatipoort security police' who requested him to help dispose of the body of another activist, Mr Johannes Sweet Sambo.

So too was his further statement that he was 'asked "to make a plan"' about an askari, Mr Johannes Temba Mabotha, whose loyalty had become suspect (and whom he then killed). So too was his evidence that Colonel Andy Taylor had instructed him to kill another suspected double agent, an askari called Mr Goodwill Neville Sikhakane. So too was his allegation, again contained in his autobiography, that two senior police officers, General 'Bertus' Steyn and General 'Krappies' Engelbrecht, had also authorised the assassination of Mr Sikhakane.

The commission's report provides no indication of how these hearsay allegations against various senior police officers were tested or verified. The TRC seems simply to have assumed their veracity-even though Col de Kock might have had reason to make false or exaggerated allegations against the officers concerned. For Col de Kock's application could not succeed unless he could show that he had been following the orders of his superiors (and had thus been acting within the course and scope of his authority as a police officer)-and unless he could also satisfy the commission that he had made a full disclosure as regards the relevant chain of command.

The commission nevertheless used Col de Kock's amnesty statement, together with similar and equally untested allegations by other amnesty applicants, to conclude that 'the SAP, in the post-1990 period, continued to carry out extra-judicial killings and attempted killings', by way of parcel bombs and other means.

The implication is that the SAP, as a whole, was responsible for such killings as part of official police policy. Clearly, such killings were carried out by policemen in this period, as Col de Kock's own conviction on various counts of murder and culpable homicide makes plain. Col de Kock and other policemen guilty of such crimes might have been abusing their powers for their own purposes. They could also have been acting on the instruction of senior police officers and in pursuance of official police policy. Which of these options applied was a key question for the TRC to address.

It is also true, of course, that the SAP had long implicitly encouraged extra-judicial killings by failing to probe or put an end to the mysterious deaths of the government's political opponents. The TRC's finding goes further than this, and indicates that the SAP itself was directly responsible for such executions in the early 1990s. This might well have been the case. But the TRC's finding to this effect would carry greater weight if it were not based on hearsay and untested allegations. The commission needed verified and substantiated evidence on which to found a conclusion of this kind.

The commission refers also to the untested amnesty statements of various other individuals-including Mr Willie Nortje, Mr 'Brood' van Heerden, Mr W Mentz, Mr Derek Rausch, Mr Johann Verster, and Mr Douw Willemse. It uses their statements, for example, to describe the alleged supply of weapons to the IFP by the South African Police (SAP) in the early 1990s:

According to the amnesty application of Mr Derek Rausch, he assisted Vlakplaas members Lionel Snyman and Snor Vermeulen to make home-made explosive devices.

Rausch, a precision engineer and an ex-Rhodesian police officer, had an engineering shop next to Mechem, a subsidiary of Armscor, and frequently worked for them. Rausch brought the material and Lionel Snyman and Snor Vermeulen provided the explosives to build the explosive devices from Vlakplaas stores � Snyman and Vermeulen � approached Rausch to assist them in making home-made shotguns. Joe Verster of Mechem assisted with this project and Snyman told him that Basie Smit [a general in the SAP] approved of the project. They made approximately 200 shotguns. According to Verster, Snyman later told him that General Le Roux was present when the prototype was tested and was very happy with the results. Both Rausch and Venter were told that the guns were intended for Inkatha. In his amnesty application, Douw Willemse stated that he assisted Snor Vermeulen and Lionel Snyman to test home-made weapons, on the instruction of de Kock.

These allegations are untested, it would seem. They also abound in hearsay. The extent to which they have been corroborated-except by similar testimony from other amnesty applicants who might also have had reason to falsify their evidence-remains unclear.

The commission implicitly acknowledges the difficulties of relying on this testimony when it states: 'The amnesty applications relating to the supply of weapons by the SAP to the IFP have at this stage not been heard and the commission is thus unable to make a finding on this issue.' In its very next sentence, however, the TRC appears to abandon this caution. It continues: 'However, sufficient evidence is available for the commission to make a finding that former SAP operatives provided substantial amounts of unlicensed heavy weaponry, explosives, and ammunition to senior members of the IFP in the post-1990 period.' This finding appears to be based on the amnesty statement of Col de Kock, who (at the time he supplied weapons to the IFP) was no longer a member of the police. The commission fails to explain why Col de Kock's untested evidence-hearsay against all individuals other than himself-should have sufficed to prove a conspiracy among former policemen to provide the IFP with weapons. (Col de Kock had himself been found to have supplied weapons to the IFP, but this did not necessarily prove the wider culpability the TRC asserted.)

To recap, thus, of the 7 127 amnesty applications received by the commission, only 102 had been heard and upheld (through the granting of amnesty) by the time the TRC compiled its report. These 102 statements were only a tiny fraction of the amnesty applications received. They, at least, should have constituted a safe source of tested, substantiated, and reliable evidence on which the commission could properly draw in making findings of accountability. Whether this was always so is difficult to gauge. A careful scrutiny of one key amnesty application suggests, however, that it might not have been.


Unexplained oddities in a key amnesty statement

An application for amnesty was put forward to the TRC by Captain Brian Mitchell, a former police officer. Capt Mitchell had been convicted in 1992 on 11 counts of murder arising out of a massacre at Trust Feed near New Hanover (in the KwaZulu-Natal Midlands) in December 1988, and had been sentenced to death-a punishment commuted in April 1994 to 30 years' imprisonment.

The TRC report describes the Trust Feed massacre as follows. In the early hours of 3rd December 1988, gunmen opened fire on a house in the Trust Feed community, killing 11 and wounding two. The attack was perpetrated by four special constables, acting on the orders of Capt Mitchell.

According to the commission, the attack had earlier been planned at a meeting involving various Inkatha leaders, including Mr David Ntombela. The police had agreed that they would initiate a 'clean-up' operation in Trust Feed one morning, so as to disarm and round up UDF suspects. (According to the trial record, this operation was decided upon after 'a vehicle had been apprehended on 24th November 1988 in which there were eight UDF supporters who were apparently on a mission to attack Inkatha leaders and were armed with firearms and petrol bombs'. The planned operation in Trust Feed was described as a normal crime prevention operation, and was notified as such to the district commissioner of the SAP in Greytown.)

According to the TRC, the underlying intention was that the police would then withdraw-leaving Inkatha members and the special constables to launch an attack on remaining UDF supporters that evening. On 2nd December, the plan was put into operation. About 30 to 40 policemen rounded up known UDF members and detained them under emergency regulations. The police were then withdrawn. At midnight, Capt Mitchell went to Trust Feed to see how the operation had gone. Disappointed that only a building had been burnt and no one killed, he instructed the special constables to burn a shop belonging to a UDF supporter and to attack a particular house. During the latter assault, 11 people were killed. A mistake was made in identifying the house to be attacked, however-and those killed were all Inkatha members, attending a vigil (or wake) for a relative who had died of natural causes.

In the investigation that immediately ensued, police involvement in the massacre was covered up. Senior police officers also tried to thwart a subsequent investigation by Colonel Frank Dutton, who ultimately assembled the evidence that saw Capt Mitchell and the special constables convicted on 11 counts of murder in April 1992.

Capt Mitchell subsequently applied for amnesty to the TRC. His amnesty application raises a number of important questions as to the extent to which amnesty evidence before the commission was, in fact, tested and verified.

Capt Mitchell had twice applied for indemnity in the past. On the first occasion, this was refused because he was awaiting execution rather than serving a prison sentence. After his death sentence had been commuted to 30 years' imprisonment, he applied again-this time under the Further Indemnity Act of 1992, under which the four special constables had been released. (See Indemnity without disclosure, below.) His application was again refused. Capt Mitchell then decided, 'in the spirit of reconciliation and for the purpose of applying for amnesty' to the TRC, that he was 'prepared to make disclosures of the events leading up to the Trust Feed incident and further revelations'. For this purpose, he was due to 'speak about matters not relevant to his application with the broader commission and its investigators', and meetings were being convened towards this end. He had refused to disclose this information in the past, he said, because he had regarded himself as a soldier captured in warfare-and had also expected the police to come to his assistance.

The amnesty panel hearing his application comprised Advocate Chris de Jager, Mr Justice Hassen Mall as chairman, Ms Sisi Khampepe, Mr Justice Bernard Ngoepe, and Mr Justice Andrew Wilson. Judge Wilson had also been the presiding officer in the trial of Capt Mitchell. He offered to recuse himself from the amnesty panel, but this was ruled unnecessary.During the trial of Capt Mitchell, Judge Wilson had noted on more than one occasion the contradictory statements made by the accused-and had commented that these 'highlighted Capt Mitchell's complete disregard for the truth'. During his amnesty hearing, however, Capt Mitchell's credibility was more assumed than challenged. This was despite the fact that Capt Mitchell-having twice been denied indemnity in the past-was utilising his last opportunity to escape a prison term of 30 years, and may have had considerable incentive to say whatever he thought would best secure his release.

Capt Mitchell's amnesty hearing was supposed to have commenced on 15th October. It was postponed till the following day to allow the victims of the massacre and their families to obtain legal representation. When the hearing resumed the next day, however, the amnesty panel was told that the victims no longer wished to give evidence in the matter. This was because all their civil claims against Captain Mitchell, save one, had been settled by the government that morning, through the intervention of the relevant state attorney. Ministerial permission was required for the one that remained outstanding, but it had been promised that this would be 'positively considered'. The victims thus had no objection to the granting of amnesty to Capt Mitchell, and were content to leave the matter to the discretion of the amnesty committee. On this basis, the hearing proceeded-and did so without legal representatives of the victims cross-examining Capt Mitchell. Counsel for the TRC did not oppose the application and hence did not cross-examine him either.

In support of his application for amnesty, Capt Mitchell lodged with the amnesty panel of the TRC an affidavit accompanied by various documents. These included SAP briefing documents and training material that described the 'total strategy' the government had initiated against the 'total onslaught' being mounted by the ANC and its internal ally, the UDF; the role within this of the National Security Management System; and the part being played by the 'special constables', in particular. One SAP document, compiled by a Major General Steenkamp in February 1987, stated that the total strategy included the training of special constables, who were 'to be taken out of the community and placed back within the community', and who were to provide a 'physical force or wedge against the tyranny of the Comrades, the UDF/ANC'. The special constables were to be attached to the SAP's riot units, and to be used in black areas where the UDF/ANC had made substantial gains in recent years, so as to prevent the alliance making any further gains. (It is unclear how new this evidence was. It seems to echo what Capt Mitchell had stated in his trial, where he had also asserted that the special constables were to be 'taken from the community and trained and placed back in the community as a physical force or wedge against the tyrannies of the comrades'.)

Capt Mitchell also attached to his amnesty affidavit a copy of a 1988 research paper, compiled by the Catholic Institute for International Relations, and entitled 'Everyone is afraid: the changing face of policing in South Africa'.

According to this document, 16 000 special constables or municipal policemen had been deployed as a third or auxiliary force to the SAP, which then numbered 48 000. They had thus increased the numerical strength of the SAP by a third. They had also made it possible to depict the violence as 'black-on-black', while also being used to 'push up the violence suddenly in given areas'.

In the case of Trust Feed, continued Capt Mitchell, 'the effect was to weaken the opposition and the area was left in the hands of pro-Security Force people and pro-Inkatha and pro-Government people. The houses of activists were set alight and attacked. Hence the special constables, as a third or auxiliary force, played an important role'.

Capt Mitchell also referred to further SAP documentation in the form of notes and a manual. This included a section describing police strategy for combating unrest in small rural black townships. The document identified the difficulty, in such areas, of using members of the local police in 'clandestine operations'-for such individuals would be well known to the community. It advocated 'making use of members from elsewhere to operate secretly' in such localities.

Capt Mitchell said that he had thus used an unmarked kombi to bring four special constables, all dressed in civilian clothes, to Trust Feed. He left them at the home of the local Inkatha leader, Mr Jerome Gabela. He was instructed to do so, he continued, by Major Deon Terreblanche, the commander of the riot unit stationed at Pietermaritzburg.

The special constables' role, he stated, was to 'render assistance to Inkatha on the evening of the attack on UDF supporters in Trust Feed'. Capt Mitchell testified further that he had 'seen this as falling within the strategy contained in the [SAP] documents'. Maj Terreblanche had also told him to bring the local Inkatha Youth Brigade leader, as well as other Inkatha leaders, to Morava House in Pietermaritzburg, to be addressed by Mr Ntombela, 'the warlord for Inkatha'.

Thereafter, 'the specials were to come to Trust Feed and the attack would be launched that night against the UDF activists within the area. The attack would be conducted by the Inkatha Youth Brigade and Inkatha members, with the assistance of the special constables'. The Pietermaritzburg riot unit was to go into Trust Feed that morning 'to clean up the area of weapons and to make the resistance less against the offensive that was going to take place that evening'.

Capt Mitchell appeared confused as to who had given the order for this attack on UDF activists. At various times-in both his documents and his oral evidence to the amnesty committee-he stated that the attack had been ordered by Mr Ntombela.

Questioned on this point, he stated that it was Mr Gabela who had needed to be schooled into using 'a strong hand in the area'. Mr Gabela, he continued, had told him on the way back to New Hanover after the meeting with Mr Ntombela, that he (Mr Gabela) 'had been told to launch an attack against the UDF after the police operation on 2nd December. Mr Gabela had been told this by David Ntombela, and he had also been told that offensive assistance would be provided in the form of armed special constables'. Capt Mitchell went on to deny, however, that it must then have been Mr Gabela who ordered the attack, and said: 'I think the presence of Maj Terreblanche and myself and David Ntombela � I think it wasn't something we could have told him, and said to him "You must go and attack". It had to come from within his own party.'

At other points in his testimony, however, Capt Mitchell emphasised that the order to use the special constables against the UDF had come-not from anyone within Inkatha-but rather from his immediate superior, Maj Terreblanche. This order, he first denied and then confirmed, had been given to him in the course and scope of his duties. (Under the commission's founding legislation, one of the requirements for the granting of amnesty to a former security force member is that he should have been acting in 'the course and scope of his duties' and 'within the scope of his express or implied authority'.)

Towards the end of Capt Mitchell's evidence it emerged that he had not himself attended the alleged planning meeting at Morava House in Pietermaritzburg. The meeting between Messrs Ntombela, Gabela, and other Inkatha leaders had been private, he said. 'Myself and Maj Terreblanche stood outside. We never attended the meeting personally.' This statement seemed to take the amnesty panel by surprise. Questioned Judge Wilson: 'But you've given evidence about what was said at that meeting?' Capt Mitchell responded: 'No sir, it was things said to me by Maj Terreblanche and it was things said to me by Mr Gabela when we left there on our way to New Hanover.'

This part of Capt Mitchell's testimony was no more than hearsay, accordingly. So too were the documents from both the SAP and the Catholic Institute that he tendered as evidence of the 'offensive' purpose allegedly underlying the deployment of the special constables in the Pietermaritzburg area and elsewhere. Other weaknesses and oddities became evident in his testimony as well:

• Capt Mitchell could not remember how the UDF targets to be attacked by the special constables were to be identified until Judge Wilson reminded him that he (Capt Mitchell) had prepared a list of names which had also been used by the police in their clean-up operation;
  
• the special constables, all strangers to the area, were not given a map of the house in which the activists were likely to have gathered, but were dropped off in darkness at a nearby shop-and pointed by Capt Mitchell in the direction in which the house lay. They were supposed to find a house where 'comrades' had hidden, possibly in a concealed underground cellar;
  
• Capt Mitchell said it was difficult to explain how the special constables were going to find the house. Judge Wilson reminded him that, during the trial, it had emerged that the police had expected the UDF activists to gather together to discuss the morning raid and to have a light burning while they did so. Hence, the fact that there was a light on in the house where the wake was being conducted had led to confusion. Capt Mitchell denied, however, that the special constables had been instructed to attack any house in which a light was showing. He said the location of the house in relation to the shop where the four had been dropped off had been explained to the special constables, but that they must have gone too far in the dark. He also conceded that 'it was very poorly planned and very stupidly done', while Judge Wilson explained to the rest of the amnesty panel that the area in issue was a rural one, where 'the houses were not close together but down this road or that road'.

Judge Mall queried, among other things, the content of the police documents cited by Capt Mitchell as proof that the special constables were intended by the SAP to be used in attacks on UDF supporters. These documents, said Judge Mall, indicated rather that the special constables were 'to be used within the terms of the law and not to commit crimes'. Capt Mitchell conceded that this was so and said that it was 'eventually in practice that things went wrong'. Judge Mall also asked if the idea of using the special constables to kill UDF activists had originated with Maj Terreblanche, or with SAP officers of higher rank-and Capt Mitchell explained that Maj Terreblanche's death in the interim had made it difficult to answer this.

Questioned further by Judge Ngoepe, Capt Mitchell's counsel confirmed that his client had first given evidence of Maj Terreblanche's pivotal role in the massacre only after Maj Terreblanche had been killed.

Capt Mitchell further conceded that the police documents disclosed 'no official justification' for using the special constables to attack and kill UDF supporters.

He explained that it was hard to recall how fraught the situation had been in 1988, and said 'it was the despair and the difficulties facing the security forces in trying to control the situation' that had led to 'stupid, stupid mistakes being made by us'. Capt Mitchell added that police lectures had spoken of hitting hard and 'fighting fire with fire' and it was this that had informed his understanding.

Judge Mall responded that he had read the documentation too and 'understood it to say that the police must hit very hard but within the parameters of the law'. Capt Mitchell had no answer to this. He also conceded that he was not aware of any similar operations having been conducted by other police station commanders in the Pietermaritzburg district. Judge Ngoepe noted further that, whether or not the SAP had authorised the use of the special constables to attack the UDF, it would still have tried to cover up police involvement in the Trust Feed killings because this would have been so embarassing and damaging.

Capt Mitchell also repeatedly told the amnesty panel that he had not been present when the special constables found the house and began shooting at its inmates. This seems to have been accepted by Judge Wilson and his colleagues on the amnesty panel.

During the earlier criminal trial, however, Judge Wilson had expressly rejected Capt Michell's evidence to this effect. He had found that Capt Mitchell had 'not only pointed out the lighted house [to the special constables] but had also given the signal for the attack to start � by firing two shots into the house'. This finding had also been particularly important, it appears, in justifying Capt Mitchell's conviction on charges of murder (rather than any lesser offence), as well as in giving him eleven death sentences.

Notwithstanding the oddities and the deficiencies in Capt Mitchell's testimony, he was granted amnesty by the TRC in December 1996. This was ordered on the basis that he had 'made a full disclosure of all relevant facts'. Moreover, said the commission, his offences were clearly 'part of the counter-revolutionary onslaught against the ANC and UDF activists' and 'he had acted within the course and scope of his duties as an officer in the SAP'. The amnesty panel did not explain why it was satisfied that he had made a full disclosure when his amnesty evidence (denying his presence at the house at the time of the attack) had expressly been rejected by the trial court as untrue.

A further important question that arises is the reliablity of Capt Mitchell's evidence in identifying the other perpetrators of the massacre. The TRC seems to have relied on Capt Mitchell's evidence to find Mr Ntombela accountable for the killings. According to the commission: 'Mr Ntombela's actions constituted gross human rights violations, including conspiracy to kill, attempted killing, and killing.' The basis for this finding is not explained by the TRC. Yet the only evidence that Capt Mitchell could have provided against Mr Ntombela was his hearsay testimony of what had allegedly been said at a meeting in Pietermaritzburg which he (Capt Mitchell) had not attended.

The commission made no finding against Maj Terreblanche himself, despite the key role he had allegedly played in the massacre. The TRC did, however, make findings against various other SAP officers. The commission found, for example, that 'the actions of Sergeant Neville Rose constituted a gross violation of human rights in that he was an accessory after the fact to the killing of the persons who died at Trust Feed, and defeated the ends of justice by failing to take any steps to ensure that the persons responsible were charged and prosecuted'.

In the Trust Feed trial, however, Sgt Rose had been charged with murder but acquitted. The principal allegations against him-that he had helped to spirit the special constables out of Trust Feed after the killings and had also replenished the ammunition they had used in the attack-were canvassed at length by Judge Wilson and found to be unreliable. The commission cites no new evidence of Sgt Rose's wrongdoing and gives no reasons for disregarding his earlier acquittal. Its basis for finding Sgt Rose an accessory to murder is not explained in any way.The commission's findings against Mr Ntombela and Sgt Rose raise further questions. It is not clear whether these individuals were even present during Capt Mitchell's amnesty application-nor whether they were accorded the opportunity either to cross-examine him or to adduce their own conflicting evidence. Moreover, the fact that the TRC cites neither the evidence nor the reasoning supporting its findings against them makes it difficult to assess the accuracy of its conclusions.

A particularly important issue is whether Capt Mitchell's flawed amnesty evidence is indicative in general of the reliability of the amnesty testimony put before the TRC. Were the defects in Capt Mitchell's testimony the norm or the exception? If they were exceptional, it raises questions as to why this former police officer should nevertheless have qualified for amnesty. If they were the norm, it would indicate that amnesty statements were not a reliable source of information-and were likely to abound in hearsay and unresolved conundrums.



3. Other evidence of perpetrator identity

Was the other evidence relied on by the TRC to identify the perpetrators of gross human rights violations then of a higher calibre? This needs also to be assessed.

According to the commission's report, the TRC relied on various kinds of evidence in identifying perpetrators. The first was 'identification through court records, confessions, statements implicating people in court dockets, police inquests, and/or previous applications for indemnity'. Each of these merits brief consideration.

The only part of a court record that can be used to identify perpetrators is the ultimate ruling regarding culpability. Such rulings might well have been useful in this regard, but would not necessarily be so. Take again the example of Sgt Rose in the Trust Feed trial. The court record showed him to have been acquitted on all charges. The TRC nevertheless found him an accessory to the killing of 11 people.

The court record, far from supporting the TRC's conclusion, contradicted it. Nor was this an isolated example, for the commission often repudiated earlier court findings (see TRC findings vis-à-vis earlier judicial rulings, below).

Confessions are also not necessarily to be accepted as the truth. Where they reflect accomplice evidence, they are subject to the three key problems identified above and must be treated with particular caution. The statements in police dockets may, or may not, be true. 'Police inquests' (presumably the commission meant court inquests, conducted either by magistrates or judges) are intended to probe the cause of an individual's death, and may be inconclusive regarding the identity of any killer. In addition, 'previous applications for indemnity' may not have been helpful either. (Applications for indemnity had been made under earlier legislation, but either concerned minor offences or were made without a full disclosure.) (See Indemnity without disclosure, below.)

The second kind of evidence used was that arising from the commission's own investigations-particularly its section 29 hearings and its other 'investigative and research work'. According to its report, the TRC used evidence of this kind to identify perpetrators wherever its investigations had generated a 'high level of corroboration', with 'a witness confirming the identity of the actual person committing the gross violation of human rights'.

Most section 29 hearings, however, were held in camera-and this raises the key question of whether secret testimony can or should be accepted as sufficient proof of culpability for gross violations as serious as torture and murder. Say that a witness at an in camera section 29 inquiry had identified Sgt Rose (to use this example once again) as the police officer who had primarily covered up the police role in the Trust Feed killings. Should this evidence suffice to find Sgt Rose an accessory to murder? Reliance on secret testimony is reminiscent of a medieaval inquisition. The sufficiency of such evidence can only be assumed-not demonstrated or assessed. Even if a second witness had thereafter told a TRC researcher that Sgt Rose had played this role, this would not necessarily confirm the policeman's culpability-for both witnesses might have been mistaken, or might have been recounting no more than hearsay.

(Secret testimony may indeed have been used to find Sgt Rose and others accountable for the Trust Feed killings. The commission indicates, at one point in its report, that a former SAP general and commissioner of the KwaZulu-Police, General Jac Buchner, had given testimony in camera on the Trust Feed killings during a section 29 hearing convened by the Investigation Unit. It notes that Gen Buchner had 'claimed the involvement of not just one or two individuals, but many'. This 'evidence'-not explained in any way-may have been used by the TRC to underpin its findings against Sgt Rose and others.)

The third way in which the commission identified perpetrators was through the repetition of their names. 'Where names consistently recurred in the statements of people making allegations concerning gross violations of human rights', this was considered a sufficient pointer to culpability, it appears. This basis for identifying perpetrators is extraordinary, for the repetition of what might initially have been false or mistaken does not turn it into truth.

The commission adds that, even where the repetition of their names pointed to the guilt of particular perpetrators, it still did not identify them as such without first notifying them that they were to be implicated in gross violations. It was obliged to do this, of course, after the former Appellate Division (now the Supreme Court of Appeal) had ruled that the commission was under a duty to comply with the audi alteram partem principle of justice. There are questions, however, as to how adequately the TRC discharged its obligations in this regard. (See Audi alteram partem, below.)



4. 'Justification' as a key factor


The commission appears to have overlooked a further vital issue. It acknowledged at one point in its report that a perpetrator-even if properly identified-could not be held accountable 'if the conduct in question was legally justified'. Justification would be present, it noted, if a person had killed in self-defence. Explaining this further, it said 'the use of force is justified in defence of persons, property, or other legal interest against an imminent, unlawful attack, provided that the defence is directed against the attacker and is not excessive'. By contrast, it added, defence against 'an anticipated future attack or a completed attack is not justified'.

Having thus acknowledged the importance of self-defence as a possible justification, the commission gives little indication of having taken this factor into adequate account. All it says in this regard is that 'the legitimacy of self-defence is often difficult to establish' and that the commission experienced particular problems in this regard because it 'had to deal with large numbers of cases in a limited period and had limited information at its disposal on many specific instances'. It was also especially difficult, it stated-in the context of the conflict between the ANC and the IFP-to ascertain 'who was "innocent" (defending) and who was "guilty" (attacking)'. (The commission does not, however, seem to have allowed this difficulty to deter it from finding the IFP a primary perpetrator of gross violations in KwaZulu-Natal and elsewhere, and from concluding that the IFP had killed people at 3.5 times the rate the ANC had done.)



5. Four different kinds of truth

The commission's founding legislation, as earlier noted, obliged it to compile its report on the basis of 'factual and objective information and evidence' received by it or otherwise placed at its disposal. It seems, however, that the TRC took account not only of 'factual' truth but also of three other kinds of 'truth'.

In describing the 'concepts and principles' that had guided its work, the TRC said it had drawn a distinction between four different kinds of truth. It identified these as 'factual or forensic truth', 'personal and narrative truth', 'social or dialogue truth', and 'healing and restorative truth'.

Factual truth, it said, connoted 'the familiar legal or scientific notion of bringing to light factual, corroborated evidence, and of obtaining accurate information through reliable (impartial, objective) procedures'. Seeking this kind of truth, it said, had 'featured prominently in the Commission's findings process', and had involved 'an extensive verification and corroboration policy to make sure that findings were based on accurate and factual information'. This assertion is hard to square, however, with the many weaknesses in the verifi- cation and corroboration processes earlier described. Factual truth, moreover, was not all that the commission sought.

'Personal or narrative truth, said the TRC, had been particularly important for the victims of gross violations of human rights. Each had been 'given a chance to say his or her truth as he or she sees it'. These 'personal truths', widely communicated to all South Africans through the media, had provided an important potential not only for healing the individuals concerned but also for 'the creation of a narrative truth' that 'captured the widest possible record of people's perceptions, stories, myths, and experiences' and thus facilitated the 'restoring of memory and humanity'.

The commission also saw 'social or dialogue truth' as particularly important to its work. Its understanding of this kind of truth had been informed by Mr Justice Albie Sachs, a prominent participant in the debates preceding the establishment of the TRC and now a judge of the Constitutional Court. Judge Sachs had drawn a distinction between 'microscrope truth' and 'dialogue truth'. 'The first,' he said, 'is factual, verifiable, and can be documented and proved. "Dialogue truth", on the other hand, is social truth, the truth of experience that is established through interaction, discussion, and debate.' In recognising the importance of this kind of truth, said the commission, 'its goal was to try to transcend the divisions of the past by listening carefully to the complex motives and perspectives of all those involved'.

It thus 'made a conscious effort to provide an environment in which all possible views could be considered and weighed, one against the other'. (How adequately the commission in fact considered all views-especially in contextualising the gross violations of the past-is canvassed elsewhere, see The Need for Evidence to be Contexualised, below.)

The commission also 'rejected the popular assumption that there are only two options to be considered when talking about truth-namely factual, objective information or subjective opinions'. There is also, it continued, 'healing' truth-'the kind of truth that places facts and what they mean within the context of human relationships, both amongst citizens and between the state and its citizens'. This kind of truth, it said, was 'central' to its work. Particularly important in this context was the healing brought about by the acknowledgement of earlier wrongdoing. 'It is not merely the actual knowledge about past human rights violations that counts; often the basic facts about what happened are already well known, at least by those who were affected. What is critical is that these facts be fully and publicly acknowledged.' It is this, it said, which restores the dignity of victims.

The commission asserts that it used only 'factual or objective truth' in arriving at its findings. Its emphasis on other forms of truth, it indicates, was intended merely to help heal the victims of past violations by giving them an opportunity to tell their stories, gain a public acknowledgement of their suffering, and participate in a 'dialogue' that sought to build a common understanding of what had happened in the past.

Being 'victim-centred' in its approach and providing a cathartic release from previous trauma were indeed vital parts of the commission's reconciliatory function.

They did not, however, remove from it the responsibility to ensure that its findings- particularly its findings of accountability-were factual and objective (as indeed demanded by its founding legislation). Too great a focus on other forms of 'truth' may have detracted from the accuracy of its conclusions regarding culpability.

The risk of this occurring was particularly acute in situations where the TRC believed that 'the basic facts about what had happened were already well known,' at least to the victims-and that the only need was to heal the victims through narrative, social, and restorative truth. In such instances, the commission may not have done enough to seek a 'factual' truth regarding accountability. It may too readily have accepted that the victims rightly knew who was to blame. The personal beliefs of individuals, however-no matter how strongly held-provide no proper basis for conclusions regarding culpability.

For the purpose of making findings of accountability, there was only one form of truth on which the commission could rely-the factual or objective truth termed 'microscope' truth by Judge Sachs. In making such findings, the commission was called upon to divorce all other 'truths' from its purview, and to focus solely on evidence that had been tested, corroborated, and adequately verified.

Mr Malan, in his minority report, expresses a concern that the commission did not sufficiently maintain this necessary discipline. 'The report,' he states, 'offers a good exposition of different concepts of truth, especially of factual truth and narrative truth and then of social or interactive truth.' The difficulty, he continues, is that 'the distinction is not sustained'. Instead, 'in arriving at findings, all is accepted as evidence, an ingredient of the factual truth'. The rest of the commission, in its rejoinder, rejects this view-stating that it reflects 'a complete lack of understanding of the findings process' on the part of Mr Malan. The commission provides no explanation, however, of how the necessary distinction between factual truth and what it claimed were other forms of 'truth' was in fact upheld.

[I] Foreword [II] Overview [III] Introduction [IV] Publication Of The Current TRC Report [V] The Need For Factual Evidence [VI] The Need For Comprehensive Findings [VII] The Need For Violations To Be Contextualised [VIII] The Need To Accord With Established Legal Principles [IX] Findings Based On A 'Balance Of Probabilities' [X] Appendix

» » » » [Truth About Truth Commission (PDF)]