Article 'They should have destroyed more': the destruction of public records by the South African state in the final years of apartheid, 1990-941 Veme Harris Introduction Under apartheid the terrain of social memory, as with all social space, was a site of struggle. In the crudest sense this was a struggle of remembering against forgetting, of oppositional memory fighting a life-and-death struggle against a systematic forgetting engineered by the state. The realities, of course, were a little more complex. Forgetting, for instance, was an important element in the struggles against apartheid - forgetting the half- truths, the distorted interpretations, the lies, of the apartheid regime. And the notions of 'oppositional memory' and 'state memory' themselves are problematic. They are artificial constructs, obscuring the sometimes fierce internal contestation in both spaces. Then there is the question of memory and imagination. Memory is never a faithful reflection of process, of 'reality'. It is shaped, reshaped, figured, configured, by the dance of imagination. So that beyond the dynamics of remembering and forgetting, a more profound characterisation of the struggle in social memory is one of narrative against narrative, story against story. Nevertheless, it is true to say that the tools of forgetfulness, of state- imposed amnesia, were crucial to the exercise of power in apartheid South Africa. The state generated huge information resources, which it secreted jealously from public view. It routinely destroyed public records in order to keep certain processes secret. More chilling tools for erasing memory were also widely utilised, with many thousands of oppositional voices being eliminated through media censorship, various forms of banning, TRANSFORMATION 42 (2000) ISSN 0258-7696 Verne Harris detention without trial, imprisonment, informal harassment and assassination. And, as this essay recounts, the tools of forgetfulness were also important to the transfer of power-between 1990-94 the state engaged in a large-scale sanitisation of its memory resources designed to keep certain information out of the hands of a future democratic government. Soon after the initiation in 1990 of the process towards a negotiated settlement in South Africa, a number of individuals and structures in opposition to the state began to express fears that such a sanitisation would takeplace. By 1994 it was clear that these fears had been well founded.2 Not surprisingly, then, when the Truth and Reconciliation,Commission (TRC) was established in 1995 to shine a light into the apartheid system's darkest caverns, one of its specific mandates was 'to determine what articles have been destroyed by any person in order to conceal violations of human rights or acts associated with a political objective'.3 The mandate provided the basis for a focused investigation into the destruction of public records by the state. Given the complexity and extent of the apartheid state, adequate coverage by the investigation of all state structures and records systems proved impossible, and the TRC decided to limit the investigation to state structures subject to national archival legislation,4 thus excluding parastatals, statutory bodies which had not voluntarily submitted to the operation of the Archives Act, 'privatised' bodies and 'homeland' structures. The 'homelands' were responsible for the management of their own records, in some cases in terms of their own archival legislation. The investigation further concentrated its energies on the activities of the security establishment - preliminary research made it clear that initiatives for systematic destruction of public records originated and were felt most acutely here.5 This essay relies heavily on the work and findings of the TRC investigation, thus reproducing in large measure both its emphases and its limitations.6 From the TRC's inception late in 1995,1 carried responsibility for liaison between it and the National Archives. When the investigation into records destruction got underway, I was released to become an integral part of the investigative team, an involvement which endured from late in 1996 until mid-1998. During 1998 I was contracted by the TRC to collate information gathered by the team and to draft sections of the final report dealing with the destruction of records. The essay also draws on my own interrogation of National Archives' documentation of records destruction up to 1994 (all of which was made available to the TRC) and of subsequent 30 'They should have destroyed more' follow-up investigations by the National Archives. I begin with an account of state record-keeping, official secrecy and the destruction of records under apartheid, before detailing the pre-election purge of 1990-94. The question of accountability is then explored, and in the conclusion I offer an assessment of the purge's impact - broadly on social memory, more specifically on the TRC' s work — and an outline of lessons to be learnt from it by a democratic state. State record-keeping and official secrecy Apartheid's bureaucracy was huge, complex, and intruded into almost every aspect of citizens' lives. Controls over racial classification, employment, movement, association, purchase of property, recreation and so on, all were documented - usually in a multi-layered process - by thousands of state offices across the country. This was supplemented by the record of surveillance activities by the Security Police, the Department of Military Intelligence (DMI), the National Intelligence Service (NIS) and numerous other state bodies, including those of the homelands. And large quantities of records were confiscated from individuals and organisations opposed to apartheid. An army of bureaucrats - servicing registries, strongrooms and computer systems—managed this formidable information resource. It is tempting to focus on the unique aspects of information gathering and record-keeping by the apartheid state. But they need to be seen in a broader, international context. One of the distinctive features of the late twentieth century state - and globalisation is rapidly creating a universal pattern - is its massive accumulation of information, particularly about its own citizens. It does this through both programmes with a service provision rationale and the activities of bodies charged with various surveillance mandates. The 'new' information technologies — the pace of their development means that they are always new - provide the state with a capacity for such accumulation which is growing exponentially. What the state does with all the information at its disposal, and how accessible that information is to citizens, are key issues. Under apartheid the state's memory resources were horded with a pathological attention to detail. While all governments are uncomfortable with the notion of transparency and prefer to operate beyond the glare of public scrutiny, in apartheid South Africa state secrecy was a modus operandi. Interlocking legislation restricted access to and the dissemination of information on vast areas of public life. These restrictions were manipulated to secure an extraordinary degree of opacity in government, and the country's formal 31 Verne Harris information systems became grossly distorted in support of official propaganda. The fundamental guideline for public access to public records was provided by the 1962 Archives Act (which was amended in 1964, 1969, 1977 and 1979). The 1962 Act-the forerunners of which were the 1922 Public Archives Act and the 1953 Archives Act - established that access was a privilege to be granted by bureaucrats unless legislation recognised the right of access to specific categories of record. The number of record categories covered by such legislation was insignificant - for instance, records older than thirty years in the custody of the State Archives Service (SAS)7 and deceased estate files in the custody of Masters of the Supreme Court. On the other hand, the discretionary power enjoyed by bureaucrats was severely circumscribed by a range of legislation containing secrecy clauses. Even within state structures, the management of information was framed by an obsession with secrecy. Every bureaucrat was graded in terms of a rigorous security clearance procedure, the grading level determining an individual's right of access to information. This procedure meshed with a pervasive system of information grading - commonly referred to as 'classification' - defined by perceived security risks. The Protection of Information Act, and various legislative forerunners, promised severe punitive action against individuals defying the system. The Archives Act charged the Director of Archives (the chief executive official of the SAS) with '... the custody, care and control of archives ...'. 'Archives' were defined as '... any documents or records received or created in a government office or an office of a local authority during the conduct of affairs in such office and which are from their nature or in terms of any other Act of Parliament not required then to be dealt with otherwise than in accordance with or in terms of the provisions of this Act'. So the SAS had wide-ranging powers over the management of public records at central, provincial and local government levels from the moment of their creation or acquisition. However, the words 'from their nature', as I elaborate on below, left the boundaries around the term 'archives' far from clear. Also unclear was who should determine the records which by their nature should not fall under archival legislation. Other provisions of the Act elaborated on specific aspects of records management - the physical care of records, their management in terms of approved 'filing systems', their conversion into microform, their accessibility, inspection and ultimate disposal. Comparison with the archival legislation of other countries 32 'They should have destroyed more' reveals that the powers enjoyed by the SAS over the active records of the state were amongst the most extensive of any national archives service in the world. The legal disposal of public records involved either their transfer into the custody of a SAS repository or their destruction in terms of a disposal authority. Until 1979 it was the responsibility of the Archives Commission, a statutory body appointed by the responsible cabinet minister, to authorise the destruction of public records. However, while this authority had been vested with the Commission since 1926, by the 1960s the Commission had become a rubber-stamp for recommendations made by the Director of Archives. A 1979 amendment to the Archives Act recognised the de facto situation by empowering the Director of Archives to authorise destruction. The Act made it a criminal offence to damage wilfully a public record, or to remove or destroy such a record otherwise than in terms of the Act or any other law. As with all national archives services, the SAS was obliged by limited resources to select only a small proportion of public records for archival preservation. To date no study has been made of the impact on the archival record of the Service's selection programme. What is clear is that state secrecy ensured that the programme was neither transparent nor accountable to the public, and that the programme was sustained by bodies (the Commission and the Service) reflective of the apartheid system and shaped by its ideology. Needless to say, efficacy in implementation is the most important test of powerful legal instruments. In practice the Service was hampered by inadequate resources and by its junior status in government. Empowered legislatively for the first time in 1922, the Service had undergone a number of name changes and been moved successively from the departments of the Interior, Union Education, Education/Arts/Science and finally to National Education. As with all the Service's staff members, the Director of Archives occupied a public service position and was appointed through the standard public service mechanisms and procedures. Only a small proportion of government offices were reached effectively by the Service's records management programme. The inspection function, crucial to the auditing mandate of the Archives Act, was no more than a token gesture. This, combined with the state's disregard for accountability and the Director of Archives' relatively junior ranking in the public service hierarchy, rendered the Service almost powerless to resist state organs obstructing its legitimate activities and flagrantly ignoring or defying its legal instruments. Especially 33 Verne Harris problematic were bodies located within the security establishment. With the exception of the South African Defence Force (SADF) and the Department of Prison Services, these bodies' records systems were not subjected to professional supervision by the Service. Indeed, there is no evidence of pre-1990 professional liaison between the Service and other branches of the security establishment. It is not clear as to whether this abrogation of responsibility was the result of orders from higher authority or was simply the result of the Service' s leadership being intimidated by the security establishment's powerful position. The consequence was that the establishment was a law unto itself in terms of the management of its own records. Also of crucial importance - and devastating in its consequence — was the vulnerability of the Archives Act's definition of 'archives' to divergent interpretations of the words 'from their nature'. It is not clear what the Act's drafters intended to exclude from the definition by these words, although in his speech of 31 January 1962 to the Senate, the Minister of Education, Arts and Science indicated that the words were designed to accommodate requirements for secret records. The authority of the Act over various categories of public record was challenged unsuccessfully on this basis in the period immediately after the Act's passage into law. However, until 1991 the status of classified (in terms of security grading) records in relation to the Archives Act received no legal scrutiny. In that year it emerged that the NIS had destroyed the sound recording of a meeting between imprisoned ANC leader Nelson Mandela and State President P W Botha. The SAS challenged the legality of the destruction on the grounds that the Director of Archives had not authorised it. On December 10, 1991, the State President's office secured a state legal opinion (299/1991) indicating that 'sensitive' documents -those requiring secrecy—were in their nature not 'archives' and therefore not subject to the Archives Act. Subsequently, the NIS also acquired a state legal opinion (308/1991, December 17, 1991) which produced a similar finding. The legal scrutiny underpinning these opinions revealed that the security establishment had since the Archives Act's inception regarded classified records as falling outside the Act's ambit and had implemented a government-wide policy for the routine destruction of such records. Records destruction up to 1990 In the period 1960-94, first the Archives Commission and later the Director of Archives issued a total of over 4 000 record-disposal authorities to state 34 'They should have destroyed more' offices. As I indicated earlier, it remains to be assessed to what extent the interests of the apartheid state were accommodated in this selection process. Within budgetary and other constraints, the SAS monitored implementation of these disposal authorities to ensure that public records were destroyed with archival authorisation and only after the lapsing of appropriate retention periods. Numerous cases of alleged or actual unauthorised destruction were investigated. Most involved disasters such as fires and flooding, and in some it was clear that negligence had played a role. However, in not a single instance was the SAS able to identify sinister motivation, for instance the deliberate destruction of documentary evidence. Over many years a dispute was sustained with Central Statistical Services (CSS) over the latter's routine destruction of census returns and related records without archival authorisation. CSS's legislative mandate required the agency to ensure the confidentiality of such records, and they adopted the position that only destruction could achieve this. The loopholes in the Archives Act's definition of 'archives' gave CSS the space to outmanoeuvre the State Archives Service successfully. Incredibly, the Service's monitoring activities did not detect a government-wide policy for the destruction of classified records until 1991. It is not clear when this policy was first implemented, but it was certainly in place by 1978. In that year all government departments received guidelines for the protection of classified information, signed by the Prime Minister and empowering department heads to authorise destruction outside the ambit of the Archives Act. The guidelines did not explicitly challenge the Archives Act's ambit. They simply authorised destruction without mentioning the Archives Act at all.8 This was in direct conflict with a standing directive of the SAS which indicated that all classified records were to be regarded provisionally as archival until such time as they had been physically appraised by the Service. The guidelines were updated in 1984 by the NIS under the State President's signature.9 How widespread or stringent was their implementation by state offices remains unclear. Certainly within the security establishment they were implemented rigorously. The SADF utilised a similar parallel set of guidelines from at least 1971. Like their civilian counterparts, military archivists in the SADF Archives appear not to have been aware of their existence. The great majority of the records generatedby the security establishment were classified and therefore subject to the guidelines' provisions for 35 Verne Harris destruction. In essence, the guidelines on the one hand obliged agency heads to destroy certain categories of record in the interests of security, and on the other gave discretionary power to destroy records which had lost their functional usefulness. The TRC investigation revealed evidence of widespread implementation, particularly rigorous in structures of the National Security Management System (NSMS), the NIS, the Security Police and the SADF. The NSMS was set up in the late 1970s to co-ordinate state action against anti-apartheid activities. It was headed by the State Security Council (SSC), ostensibly subordinate to Cabinet but by the mid- 1980s supreme on issues relating to security. The Council ran a huge network of sub-structures reaching into every part of the country, relying mainly on security establishment resources but drawing in almost all organs of the state. When the public debate on the destruction of classified records occurred in 1993 (recounted later in this essay), the head of the SSC Secretariat maintained that a full set of NSMS records were being preserved and that only duplicate copies were being destroyed. However, the official responsible for the management of these records from 1980-90 was later to inform the TRC that the guidelines for destruction were fully implemented throughout that period. Not surprisingly, the documentary residue of the NSMS contains numerous and substantial gaps. NIS was established in 1980, inheriting the functions of the Bureau of State Security (1968-78) and the Department of National Security (1978- 80). The systematic routine destruction of NIS records began at least as early as 1982. On December 1, 1982, the SAS's top management adopted a set of guidelines (Directive 0/01) which authorised divisional heads and regional representatives to destroy records no longer possessing security relevance on an annual basis. It proved impossible for the TRC investigation to determine records disposal procedures in the pre-1980 era, but the evidence suggests that NIS procedures were applied to any records which had survived. The Security Police was a branch of the South African Police (SAP). With the approval of the Director of Archives, they managed their records in terms of records systems approved by the Director for use throughout the SAP but in physically separate record sets classified as secret or confidential. Standing SAP instructions indicated that no secret or confidential records could be destroyed without written authorisation from the Director of Archives. In the period 1960-94 no such authorisation was given. The TRC investigation determined that throughout this period Security Police records 36 'They should have destroyed more' were routinely destroyed in accordance with internal retention/disposal arrangements. In the main this seems to have applied to support function records rather than operational records. Huge volumes of operational records were generated, at head office, regional and local levels. To cope more effectively with them, a microfilming project was initiated, probably in the 1970s. Originals of microfilmed records were apparently destroyed, but not on a systematic basis. From 1983, a computerised database of operational records was implemented. Again, it appears- as if certain original records were destroyed after the core data had been captured on the database. Nevertheless, in 1990 the Security Police retained huge quantities of operational records in locations throughout the country, a large proportion still in paper form. The SADF enjoyed a special status within the framework ofthe Archives Act. It managed its own archives repository and from the late 1960s, provided its own records management service (through the SADF Archives) to SADF structures. Both functions were supervised by the SAS. Standing orders required that records only be destroyed in terms of authorities signed by the Director of Archives, and that destruction certificates be submitted to the SADF Archives. However, as I have already indicated, from at least 1971 conflicting standing orders authorised the routine destruction of classified records without reference to the SADF Archives, the Director of Archives, or the Archives Act. The evidence suggests that substantial volumes of records were destroyed in this way without any archival intervention. There is also evidence of large-scale destruction of records generated by bodies related in one or other way to the SADF. The South West Africa Territory Force was a joint South African/Namibian force established to operate in conjunction with SADF operations in Namibia. Its records were subjected to systematic appraisal in an exercise initiated in December 1988. Decisions on which records were to be destroyed were authorised by the commanding officer. There was no consultation with the civilian archives repository in Windhoek, the SADF Archives, or the State Archives Service. Records which survived this exercise were placed in the custody of the SADF Archives. The Civil Co- operation Bureau was a special unit established to disrupt or eliminate persons considered to be enemies of the state. It reported to the SADF's Special Forces division. The Harms Commission of Enquiry into Certain Alleged Irregularities, which reported in 1989, revealed that all the Bureau's records had either been destroyed or illegally removed. The records of 37 Verne Harris Koevoet, the notorious police counter-insurgency unit which operated out of Namibia, were reported as having all disappeared in transit between Windhoek and Pretoria. Between 1960-90, through its appraisal function and the monitoring of state offices, the SAS had sought to control the destruction of public records and to ensure the preservation of records with archival value. Nevertheless, by 1990 there was a well-established practice within state structures of routinely destroying classified records outside the ambit of the Archives Act. Within the security establishment there was an ethos in the management of its own records characterised by almost complete autonomy from the intervention of the SAS. Nevertheless, throughout the state substantial and archivally rich classified information resources were being maintained. Particularly in the security establishment, a prevailing sense of being in control supported the preservation of records which in more uncertain circumstances would surely have been destroyed. The pre-election purge and subversion of the Steyn Commission Uncertainty for state structures was heralded by the February 1990 unbanning of the ANC and numerous other organisations, and the subsequent initiation of formal negotiations towards the dismantling of apartheid. Apprehension about certain public records passing out of the then government's control became prevalent. There was particular concern about such records being used against the government and its operatives by a future democratic government. The first state agency to act decisively was the NIS. In 1990, it decided to replace its 1982 guidelines for records destruction with a far more rigorous process to be managed by an inter- divisional Standing Re-evaluation Committee. New guidelines were given to the Committee in October 1991. The guidelines required the destruction of paper-based records unless there were very good reasons for their retention. 'Security-relevant' records were to be kept on microfilm or electronic form, where they were most secure and easier to destroy/erase quickly. Continued retention was to be reviewed on an annual basis. In addition, documentation of covert operations was to be categorised according to sensitivity and security relevance criteria, with references to the most sensitive documentation to be removed from the electronic information retrieval system. None of this documentation was to be kept for longer than six years. Top management elaborative guidelines (see TRC 1998, appendices 2 and 3, vol 1, ch 8) issued in February 1992 make 38 'They should have destroyed more' it clear that one of the purposes of this exercise was to sanitise the image both of the government and of NIS in a new political environment. Initially the new guidelines did not accommodate Treasury requirements for the management of financial records. However, in 1992, after conferring with the Auditor-General and the Director of Archives, the NIS Director- General requested ministerial approval for the destruction of financial authorisations, vouchers andrelated documentation. The Minister of Justice and National Intelligence gave his approval on July 3, 1992. Implementation of the new NIS policy seems to have gained momentum in 1992 but to have reached its greatest intensity in 1993. Mass destruction of records, embracing all media and all structures, took place. In a six to eight month period in 1993, NIS headquarters alone destroyed approximately 44 tons of paper-based and microfilm records, utilising the Pretoria Iscor furnace and another facility outside Johannesburg. The evidence suggests that many operatives took the opportunity to 'clean up' their offices, irrespective of the guidelines. Systematic destruction exercises continued until late in 1994, with many of the surviving minutes of chief directorate, directorate and divisional meetings and most administrative records covering the period 1989-94 being destroyed at this late stage. NIS' s own requirements for the preparation of destruction certificates were seldom complied with. The result was a massive purging of NIS's corporate memory. This was supplemented by the unauthorised ad hoc removal of documents by individuals for their own purposes. Any attempt to quantify this phenomenon was beyond the resources of the TRC investigation. Very little pre-1990 material survives in the paper-based, microform and electronic systems, and the documentary residue for the period 1990-94 has been substantially sanitised. The one seemingly intact records series is minutes of senior management meetings which covers the period 1980-94. In 1992, the Security Police followed the example set by NIS. In March of that year an instruction emanating from their head office ordered the destruction of all operational records, including non-public records confiscated from individuals and organisations. The TRC investigation was unable to determine either the precise source of this instruction or its precise content. The evidence suggests that it was received verbally at both regional and local levels. The instruction embraced all media and required the destruction not only of records but of all documentation of the records. In the months following the issuing of the instruction, massive and systematic destruction of records took place. In some cases records were 39 Verne Harris removed to head office for destruction. In others destruction took place on- site. In yet others private companies like NAMPAK (a manufacturer of cardboard containers) and SAPPI (a paper and board manufacturer) were utilised. With few exceptions, it would appear that Security Police offices implemented the instruction to the letter. In fact, some offices destroyed most, if not all, support function as well as operational records. But there were exceptions. The investigation revealed that certain operational records were not destroyed by 11 regional and local offices. Several thousand files also survived in what was the Security Police head office, although most of them post-date 1990. Elevenback-up tapes of the head office computerised database were located, seven of them still readable. And contrary to the March 1992 instruction, three offices kept lists of files forwarded to head office for destruction in terms of the instruction. As early as 1990, NIS's top management expressed the need for co- ordinated government-wide action in the destruction of records. The first step taken in this direction by it related to the records of the NSMS, which was rapidly dismantled after February 1990. NIS was made the official custodian of NSMS records. On November 29, 1991, a circular instruction was sent to all government departments requiring them to transfer to NIS all NSMS-related records in their custody. While the stated purpose of the exercise was to enable the SSC Secretariat to assemble a complete set of these records, it was clearly designed to facilitate systematic sanitisation. The exercise was less than successful, and in July 1993 the head of the Secretariat, with explicit Cabinet approval, sent another circular to all government departments recommending that they destroy all classified records which had been received by them from other sources, with the exception of those constituting authorisation for financial expenditure or 'other action'. Special mention was made of documentation related to the NSMS. The impact of this circular was immediate and severe. Across the country government officials began purging the classified records series under their care. At the time I was an archivist in the records management programme of the SAS. I had professional contacts in numerous government offices, and some of the more conscientious amongst them alerted me to the danger. When I briefed the Director of Archives, I discovered that he knew about the circular and 'had the matter in hand'. But when nothing was done over the next week to stem what was clearly a massive government-wide destruction exercise, I leaked the information to the ANC, the press and 40 'They should have destroyed more' Brian Currin, then National Director of Lawyers for Human Rights. In the public furore which followed, the state maintained that the step was merely designed to eliminate unnecessary duplicate copies of classified records, that all originals would be preserved, and that in any case classified records fell outside the ambit of the Archives Act. Currin then challenged the circular's validity in the Supreme Court, identifying the respondents as the State President, the Minister of National Education, the Director of Archives and the Director-General of NIS. In his application, Currin argued that state legal opinions 299/1991 and 308/1991 were'wrong', and that the nature of 'sensitive' records, including classified material, did not exclude them from the operation of the Archives Act. On September 27, 1993, all the parties reached agreement that from then on no public records would be dealt with otherwise than in terms of the Act 'simply by virtue of the fact that they are classified, or they are classified into a category denoting some degree of confidentiality'.10 Two days later, the Minister of Justice issued a media statement in which he stated that 'Cabinet is of the view that state documentation should be dealt with in terms of the Archives Act' (author's translation of Afrikaans text). Hopes that the loophole in the Archives Act had been removed proved vain. The settlement had not incorporated Currin's broader arguments, and the state exploited this to continue its 'legal' destruction of records outside the operation of the Archives Act. The 1984 guidelines for the destruction of classified records were not withdrawn. In fact, as late as November 1994, after the installation of South Africa's first democratically elected government, NIS issued an updated version of the guidelines which still ignored the Archives Act. This was a direct violation of the Currin settlement. The Director of Archives challenged NIS accordingly, and the guidelines were revised appropriately and re-released in February 1995. For the benefit of the media and oppositional groupings in the wake of the Currin settlement, the state staged a charade of abiding by its provisions. A second circular was sent out to government departments qualifying the contents of the first. An inter-departmental working group was established to prepare guidelines for government offices on which categories of public record fell outside the ambit of the Archives Act. When the group produced draft guidelines, the Director of Archives (through the Director-General of National Education) sought a state legal opinion on their validity. The opinion (220/93, November 2,1993), without even referring to the Currin settlement, simply affirmed the findings of opinion 299/91. However, the 41 Verne Harris opinion did assert that decisions on destruction should not be left to individual department heads and recommended that an advisory mechanism should be created. This was never done. The full extent of the Cabinet's duplicity only emerged during the TRC investigation. Unbeknownst to either Currin or the SAS, on June 2, 1993, a month before the July Security Secretariat circular, Cabinet had approved a new set of guidelines for the disposal of 'state-sensitive' documentation. These guidelines (see TRC 1998, appendix 1, vol 1, ch 8) had their origin in meetings of NIS top management in 1990 and 1991, where it was decided to use NIS's own destruction guidelines as a point of departure for the preparation of government-wide guidelines. The proposal was taken to the SSC which adopted the guidelines in May 1993, subject to a NIS investigation of comparative practice internationally. There is no evidence that NIS conducted such an investigation. The following month the SSC proposed the guidelines to Cabinet, which duly approved them. They empowered ministers to authorise the destruction of financial and related records outside parameters laid down by the Treasury, and heads of departments to authorise the destruction of all 'state-sensitive' records meeting certain loosely-defined criteria. The guidelines were distributed to all government departments. Carrying the weight of the highest authority in the land, their impact was severe. The SAS's own parent body, the Department of National Education, for instance, promptly destroyed most of the files in its security-related filing system, despite the fact that the system was subject to a SAS disposal authority which had earmarked the great majority of the files for archival preservation. Nevertheless, the evidence suggests that implementation was extremely uneven and shaped directly by offices' positioning in relation to the coercive aspects of apartheid administration. It is unclear to what extent subsequent destruction exercises were in response to or shaped by the Cabinet-approved guidelines of June 1993. But clearly senior managers in state structures regarded themselves as having been given the green light to sanitise records in their care. No records of the KwaZulu Intelligence Service (KWAZINT) survived. KWAZINT existed between 1986-91 as a special NIS project managed in co-operation with the KwaZulu homeland. All project records were either sent to or managed by NIS. During 1995, the remaining former homeland intelligence services were integrated into the new civilian intelligence services, the National Intelligence Agency (NIA) and the South African 42 'They should have destroyed more' Secret Service. It seems that before then very little records destruction had been effected by these services. However, between April-October 1995 a NIA Chief Directorate Research and Analysis Co-ordinating Committee subjected some of the records inherited from these services to a thorough re-evaluation process. Working both on-site and with records which had been transferred to NIA headquarters, the Committee was mandated to identify for preservation records of value to NIA from both operational and historical perspectives. The TRC investigation revealed that less than five per cent of the records were identified for preservation, almost none of thempre-dating 1990, and that in practice the sole criterion for preservation seems to have been security relevance. The remaining records were subsequently destroyed, the last destruction exercise taking place as late as November 1996. This episode revealed the resilience of attitudes and values from the past. Not only did NIA, ostensibly a structure of the new democratic South African state, implement the sanitisation policy of the apartheid state, in doing so it ignored the State Archives Service and defied moratoria on the destruction ofpublic records introduced in 1995.n After completion of the re-evaluation process, large volumes of additional records were secured at NIA headquarters from the offices of the former Bophuthatswana, Transkei and Venda intelligence services. The periods covered by these records are as follows: Bophuthatswana Intelligence Service (1973-95), Bophuthatswana National Security Council (1987-94), Transkei Intelligence Service (1969- 94) and Venda Intelligence Service (1979-94). The SADF responded decisively to the Cabinet-approved guidelines. In 1992, Lieutenant-General Steyn, the then SADF Chief of Staff, was appointed to investigate SADF intelligence activities. On November 23, 1992, all SADF structures were informed that from then on records were only to be destroyed with the express approval of Steyn. However, on receipt of the Cabinet-approved guidelines, the Chief of the SADF ordered their immediate implementation, thus effectively repealing General Steyn's instruction. Two joint teams consisting of inspector-general and counter- intelligence personnel were appointed to visit all units and to identify records for destruction. A countrywide destruction exercise followed. By and large this exercise failed to produce the required destruction certificates, making analysis of its impact extremely difficult. The TRC investigation was forced to seek a sense of the impact through probes into what it regarded as hot-spots: 43 Verne Harris • although subjected to close scrutiny during the 1993 destruction exercise, a surprisingly large volume of Military Intelligence files have survived. As one of the South African National Defence Force's (SANDF) legal team commented to me during the investigation: 'They should have destroyed more'. Another instance of being confronted by a ghost from the past. Three discrete file groups were identified at the SANDF Archives: group number 14, comprising 299 boxes of files covering the period 1977-87; group number 21, comprising 254 files covering the period 1975-87; and group number 30, comprising 529 boxes of files covering the period 1976-96. However, significant gaps were identified. For instance, no record accumulations of the Directorate Special Tasks or the Directorate Covert Collection could be found, and only a small accumulation of Contra-mobilisation Projects (COMOPS); • no record accumulation related to the Civilian Co-operation Bureau could be found; • spot checks revealed that not all personnel files could be made available, raising the question of whether or not such files had been destroyed; • spot checks suggested that substantial documentation of cross border operations in neighbouring countries had survived; • very little NSMS documentation managed by the SADF has survived. The only significant accumulation comprises 54 boxes of files (now in the SANDF Archives) generated in the Eastern Cape and preserved for use in the inquest conducted into the death of political activist Mathew Goniwe (the Goniwe Inquest). However, some other NSMS documentation was identified in each of the three DMI file groups described above; • a task group authorised by the Chief of the SANDF in June 1994 managed the acquisition by the SANDF Archives of all extant records of the former defence forces of Transkei, Bophuthatswana, Venda and Ciskei. These forces had been amalgamated with the SADF and non- statutory forces to form the SANDF in April 1994. Apart from the 1 544 boxes offiles secured from the Bophuthatswana Defence Force, relatively insignificant documentary traces were secured: 80 boxes of files from the Transkei, 115 from the Ciskei and 331 from Venda. Excluded from these figures are personnel files, which were integrated with the SANDF' s personnel file series. Clearly, then, huge volumes of records generated by the defence forces of the former homelands had been destroyed. 44 'They should have destroyed more' By May 1994, when the new democratically-elected government took office, a massive deletion of state documentary memory had taken place. This enforced amnesia was concentrated, for obvious reasons, in the security establishment. Unlike their counterparts in the former East Germany, Kampuchea and other countries, South Africa's apartheid leaders had had plenty of time in which to do the job thoroughly. Despite this, surprising pockets of public records survived the process, even within the security establishment. Some I have already detailed. There were others. From the perspective of documenting resistance to apartheid, two are of particular interest. First, the Department of Prison Services, despite routinely destroying classified records in the pre-1990 period and acting on all the 1993 government-wide guidelines, preserved intact two significant file series: case files opened for every security/political prisoner; and case files opened for every prisoner under sentence of death. Second, the comprehensive accumulation of records generated by the Department of Justice's Security Legislation Directorate. The Directorate was established in 1982 and endured until 1991. Its predecessor was the Internal Security Division, andbefore thatthe function was performed (beginning in 1949) by various individuals in the Department. Its function was to make recommendations to the Ministers of Justice and Law and Order concerning the administration of security legislation- for instance should an individual or organisation be banned? Should an individual be restricted? Should a certain gathering be allowed? Legislation falling within its ambit included the Suppression of Communism Act, Internal Security Act, Affected Organisations Act, Terrorism Act, Unlawful Organisations Act and the Public Safety Act. It made recommendations on the basis of investigations initiated by the Security Police. Recommendations were supported by information gathered on its behalf by the Security Police, NIS and DMI. The evidence suggests that the Directorate's records management was impeccable. Records were kept in accordance with SAS and departmental directives, with disposal being performed in terms of disposal authorities issued by the Archives Commission and the Director of Archives. While the Directorate did routinely destroy classified records received from other state offices in terms of the NIS guidelines, they ignored all the 1993 disposal guidelines. The Directorate's extant records, kept in excellent condition by the Ministry of Justice, comprise the following: a series of case files for individuals, spanning the period 1949- 1991; series of case files for organisations and for publications (the series 45 Verne Harris for organisations includes files inherited by the directorate dating back to the 1920s); and policy, administrative and other subject-based correspondence files. Despite the large-scale destruction of records which had been taking place during the negotiation process, as the April 1994 election loomed President De Klerk and his Cabinet became anxious about what remaining public records the new government would inherit. Late in 1993 the President' s office asked the Chief State Law Advisor whether representatives of De Klerk's government could retain custody of certain records after April 1994. A draft memorandum leading to the formal request cited an obscure British precedent and indicated that one of the motivations was to 'keep this information out of the hands of future co-governors' (author's translation). The records referred to were 'gebruiksdokumentasie'- working documentation - including Cabinet minutes and the minutes of Cabinet committees, ministers' committees and the SSC. At the time, none of these records had been transferred into the custody of the SAS, on the grounds that their 'sensitive-nature' excluded them from the operation of the Archives Act. In his opinion 207/1993 of December 22, 1993, the Chief State Law Advisor indicated that such records could not be removed from the state's custody. Also in December 1993, President De Klerk referred the same question to Advocate SA Cilliers for an opinion. Advocate Cilliers responded on January 13, 1994, confirming the Chief State Law Advisor's opinion, and going further by disagreeing with opinion 299/ 1991 and its affirmation of the legality of the destruction of' state sensitive' records on the authorisation of departmental heads.12 Subsequently, Cabinet and Cabinet committee records were transferred to the SAS, albeit with a Cabinet-imposed ten-year embargo on access. The embargo was ignored by the Service, with access being managed from the outset in terms of the Archives Act's access provisions. In 1995 and 1997, the surviving residue of SSC and related records was also transferred into archival custody. Why, one must ask, did De Klerk and his Cabinet not simply destroy these records? With approval already given for the destruction of numerous other records categories, why the fastidiousness over these? I suspect that the answer is twofold. On the one hand, they were high- profile records which both the media and the new government would be anxious to see after April 1994. On the other, the destruction of these records would involve Cabinet directly. It would be impossible to blame junior officials for misinterpreting disposal guidelines. 46 'They should have destroyed more' Accountability The routine destruction of classified public records outside the parameters of the Archives Act had begun well before 1990. Sanctioned by the head of state, the process was concentrated in South Africa's security establishment. Between 1990-94, this process was broadened into a systematic endeavour authorised by Cabinet, reaching into all sectors of the state and embracing categories of record designated as ' state-sensitive'. At the time and subsequently, those responsible maintained that the endeavour was designed simply to protect intelligence sources and the legitimate security interests of the state. The evidence demonstrates that it went far beyond this, constituting a systematic sanitisation of official memory resources ahead of transition to democracy. Those responsible also maintained that the endeavour was entirely legal. They pointed to the state legal opinions secured by the State President's office, NIS and the Director-General of National Education in 1991 and 1993, which ruled that 'state-sensitive' public records fell outside the definition of records which were subject to the Archives Act. This argument is deeply flawed. First, the legal opinions were disputed by the SAS, Advocate SA Cilliers and the National Director of Lawyers for Human Rights, Brian Currin. The basis of Currin's successful legal intervention in 1993 was a rejection of the two 1991 opinions. Second, the public position adopted by Cabinet itself was that all public records should be dealt with in terms of the Archives Act. Third, the state used the legal opinions selectively. For instance, the 1993 opinion's recommendation that an 'advisory mechanism' on records destruction be created was never implemented. Fourth, Cabinet's approval of the destruction of financial records outside requirements laid down by the Treasury was of dubious legal validity. Finally, the legal opinions begged the question 'in terms of what law are "state-sensitive" records to be destroyed?' Several officials involved in such destruction pointed to the Protection of Information Act, but this Act makes no reference to the destruction of documents. Ultimately the question of legality is a non-issue. On the one hand, apartheid was characterised by 'official lawlessness' (Merrett 1994), with rules and actions which were perfectly legal but lacked legitimacy and bore little or no relation to the rule of law. On the other, it is clear that the sanitisation of official memory resources would have taken place irrespective of legal constraints. As Brian Currin said of the 1993 settlement which followed his legal intervention, the only way to enforce it would have been 47 Verne Harris to 'tie up their [government's] hands and confiscate all the relevant machinery they can use to destroy documents' (TRC 1998, vol 1, ch 8:232). Given its legislative mandate, the SAS was the principal state agency responsible for acting against the destruction of public records without archival authorisation. In the 1990-94 period of mass destruction, intervention by the Service achieved nothing. It followed up by correspondence every allegation of illegal records destruction, engaged the security establishment in debate around the issue, registered its disagreement with the 1991 and 1993 legal opinions, and forced revision of NIS's 1994 Guidelines for the Protection of Classified Information. However, it was hamstrung by the apartheid system's disregard for accountability, by inadequate resources, by its junior status in government, and by a leadership which was intimidated by the security establishment and lacked the will to act decisively. •I was a member of staff in the Service throughout this period, and remember well how I and some of my junior colleagues pushed for such action while the leadership chose to sit on the fence. Earlier in this essay I recounted the inadequacy of leadership's response to the 1993 SSC Secretariat circular authorising the destruction of certain categories of classified record. To cite another instance, in June 1992 the Department of Foreign Affairs (DFA) requested authority to destroy certain special projects files. When the Director of Archives indicated that they should be transferred into SAS custody, DFA withdrew their application and claimed that the files were in fact merely empty file covers. My calls for an investigation were refused by the Director. More damning was the Director's collusion with NIS in 1992, cited earlier in this essay, to secure authorisation for the quick destruction of that agency's financial and related records. Specific instances aside, not once in the period 1990-94 did the Director authorise an investigative inspection of an office suspected of destroying records illegally. Not once did the Director undertake a face-to-face meeting with a suspected perpetrator. And not once was the Archives Act used to institute an investigation of possible criminal charges in terms of the Act. What about intervention by the liberation movements? I joined the ANC in 1990, and was appointed to its Archives Committee (a sub-committee of its Commission on Museums, Monuments and Heraldry) in 1992. Within the Committee and other structures I was involved in there was an acute awareness of the danger that the apartheid state was planning a mass 48 'They should have destroyed more' destruction of public records. The experience of Zimbabwe in the months preceding that country's independence, when huge quantities of public records were destroyed by the outgoing regime, was frequently cited in discussions. It was felt to be imperative that the issue be put on the agenda during negotiations with the apartheid government, and that the ANC's leadership should call for a moratorium on the destruction of public records with immediate effect. The first formal recommendation for such a moratorium was made at a meeting of the Commission on Museums, Monuments and Heraldry in March 1992, and at the ANC's 1993 Conference on Culture and Development it was resolved that 'there should be an immediate cessation of the destruction of all State records regardless of existing policy'.13 However, it proved impossible to mobilise leadership behind the issue. It was not put on the table during the multi-party negotiation process. Support for Currin's 1993 legal intervention was limited to a media release backing his endeavour. When the Transitional Executive Council (TEC) was established in 1993, the liberation movements which participated failed to ensure that the enabling legislation addressed the question of a moratorium. Moreover, the TEC failed to take any action in the wake of the Currin settlement - the Council was, in the words of Currin, 'just paralysed and didn't respond' (TRC 1998, vol 1, ch 8:232). Action was only to take place in 1995. In June 1995, the National Intelligence Co-ordinating Committee introduced a moratorium on the destruction of all' intelligence documents'. In November 1995, Cabinet decided on a moratorium which applied to all records of the state, irrespective of their age and irrespective of whether or not the Director of Archives had authorised their destruction. This blanket moratorium endured until completion of the TRC's work in 1998, whereafter it was narrowed to the records of the security establishment. It will only be lifted when the amnesty process, begun by the TRC but not completed, is concluded. These moratoria, of course, came too late. It is also not clear how effectively the moratoria were communicated to and enforced within security establishment structures. Certainly NIS and later NIA, as I pointed out earlier in this essay, continued destroying records after their introduction, until as late as November 1996. It could be argued that more decisive intervention by the ANC and the other liberation movements would not have prevented, nor even curbed, the mass destruction. Nevertheless, this was a lever which sadly was not utilised. Verne Harris In its findings on the destruction of public records in the period 1990- 94, the TRC distinguished between culpability and accountability (1998, vol 1, ch 8: 235-6). The former implies wrongdoing, the latter shortcoming or negligence. Identified as culpable were: • Cabinet and the SSC, for sanctioning, from at least 1993, a government- wide purging of official memory resources; • NIS, for: beginning its purging exercise before Cabinet sanction was secured; initiating the process which led to the adoption of government- wide destruction guidelines in 1993; defying the terms of the Currin settlement by failing to revise its Guidelines for the Protection of Classified Information; and of supervising, or at least of failing to prevent, the purging of NSMS records; • the Security Police, also for beginning its purging exercise before Cabinet sanction was secured; • the numerous individual state officials and operatives who used the cloak provided by the destruction endeavour to destroy or remove documents without authorisation. Also found culpable were the NIA officials directly responsible for the destruction of records until as late as November 1996, in defiance of the two government moratoria. NIA's top management were held accountable for not preventing this destruction. The TRC assigned accountability as follows: • the head of the SSC Secretariat, for the consequences of his July 1993 circular to all government departments recommending the destruction of certain categories of classified record; • the SAS for 'the indecisive and ineffective steps it took to halt the destruction endeavour'; • the liberation movements, for failing 'to exercise all the leverage at their disposal in acting against the endeavour'. Conclusion It is far too early to come to any conclusions about the impact of the 1990- 94 purge on social memory in South Africa. Our knowledge of the purge relies heavily on the TRC investigation into records destruction, an investigation severely constrained by a number of factors. It operated with limited resources within extremely tight timeframes. Of necessity, it had to rely on highly selective probes into hot-spots, and in doing so was 50 'They should have destroyed more' dependent to a greater or lesser degree on resources and co-operation made available by state structures still in the initial stages of transformation. While for the most part levels of support were excellent, there were cases of obstruction.14 In my view, the TRC investigation gave us a sound grasp of the broader processes of records destruction - the big picture - and considerable insight into the impact of those processes within the security establishment. It remains for the National Archives and private researchers to extend these boundaries. What we can say at this stage is that the evidence suggests a considerable impact on social memory. Swathes of official documentary memory, particularly around the inner workings of the apartheid state's security apparatus, have been obliterated. Moreover, the apparent complete destruction of records confiscated from individuals and organisations over many years by the Security Police has removed from our heritage arguably the country's richest accumulation of records documenting the struggles against apartheid. The overall work of the TRC suffered substantially as a result. In seeking to reconstruct and understand the past, so many pieces of that past's puzzle were missing. As the TRC itself indicated, 'the destruction of state documentation probably did more to undermine the investigative work of the Commission than any other single factor' (1998, vol 1, ch8:204). For the most part, the big picture, the fundamental shape and pattern of process, was as clear as any interrogation of the past can be clear. But so often the details, the nuances, the texture, the activities and experiences of individuals, was absent. On the other hand, TRC investigation teams were often surprised by records accumulations which survived the purge. One has to ask why they survived. Imperfect central control over what was a vast bureaucracy? The presence of individuals with consciences in the lower reaches of the state? Determination to preserve information which could compromise the leadership of the new government? During the course of the investigation I saw several files which could create severe difficulties for people now prominent in the public and other sectors. At one point I remember one of my TRC colleagues turning to me with the comment: 'Perhaps it would have been better if all these files had been destroyed'. More edifying has been the discovery of extensive accumulations of records detailing the apartheid state's dispossession of individuals' and communities' rights to land. The National Archives and the Department of Land Affairs have worked closely with the National Commission for the Restitution of Land Rights to identify substantial records series in state 51 Verne Harris offices around the country which the Commission is using to investigate land claims. Clearly, then, much work remains to be done before we have a comprehensive picture of the scale and consequences of the 1990-94 purge. Imperfect as our understanding of the purge might be, we know enough to have learnt crucial lessons from it. Perhaps the most important is the necessity for transparency and accountability in government. As the transition to democracy has gathered momentum, 'openness' and 'disclosure' have become watchwords both within the state and in broader societal processes. This emphasis is underpinned by the new Constitution's recognition of the public right of access to information, particularly that held by the state.15 However, it remains to be seen how well this lesson has been learnt. Already evident is a strong counter-current, fed by state officials and structures who are finding themselves blinded by all the light. There is now awareness within the state - honed by the impact of the records destruction moratoria - that no state has the resources to preserve indefinitely all the information in its systems. Selection procedures - choosing what to remember and what to forget - are essential. This to support both efficacy, in the longer term, of archival programmes, and protection of legitimate interests in confidentiality. But beyond the determining of memory's outer boundaries, the state is also becoming adept at crafting the hidden places - the 'official secret' - within that memory. Take the TRC, torchbearer of disclosure, as an example. Some of its hearings were held in camera. Its records of protected witnesses were secret. Information on certain decision-making processes and of internal tensions and disputes was jealously kept out of the public domain. Its archive is subject to various access restrictions. Some democrats accept that a measure of official secrecy is desirable; most accept it as unavoidable. But there are disturbing signs in South Africa that official secrecy is beginning to be embraced as a point of departure. In March 1999, the SANDF demanded the return of certain 'top secret' documents it had submitted to the TRC - this followed the detention by the police of a Swiss journalist for being in possession of a copy of one of these documents, which he had been given by the TRC, The state has effected no meaningful changes to the inherited systems of information classification and staff security clearance. Increasingly the media are running into government communications officials who constitute brick walls rather than gateways. At the same time, heat directed against 52 'They should have destroyed more' media freedom by the state is gathering fuel. And 'open democracy' legislation - which will provide for freedom of information, the protection of personal information, and the protection of whistleblowers in the state - has been through a protracted gestation cloaked in secrecy. '6 It seems that for South Africans, particularly lawyers, journalists and activists, learning how to wrestle effectively with the 'official secret' will be essential. The degree to which we are successful will be a crucial measure of South Africa's democratisation. The purge also highlighted the need for a democratic state to take appropriate measures to prevent the sanitising of official memory resources. The cornerstone for such measures is the provision of suitably powerful legal instruments to a state agency responsible for the auditing of public record-keeping and, ideally, for managing public archives services as well. In many respects, the 1962 Archives Act constituted such an instrument, but it possessed four fatal flaws. First, many state offices were excluded, wholly or partially, from its operation. Second, its definition of 'archives' (public records) contained loopholes which the apartheid state was able to exploit ruthlessly. Third, the penalty for conviction on a charge of destroy ing a public record without archival authorisation was a laughable fine of R200.'7 This did not constitute a deterrent. Finally, it provided no mechanisms for ensuring accountability and transparency in the selection of public records for preservation by the Director of Archives. All these flaws have been rectified by the National Archives of South Africa Act of 1996, and the national government has put in place mechanisms to ensure that archival legislation passed by the provinces follows the same model. Needless to say, a powerful legal instrument without appropriate executive action is nothing more than a dead letter. This was recognised by the TRC (1998, vol 5, ch 8:346) in three of its recommendations: • government should provide the National Archives with the resources it requires to give life to the legislation. The power to inspect governmental bodies, for instance, is rendered meaningless if the resources to exercise it are not made available. Current budgetary allocations to the National Archives are woefully inadequate; • government should take steps to ensure that the positioning of the National Archives within the state supports its function as the auditor of government record-keeping. Currently, as with the State Archives Service in the past, the National Archives is positioned as a junior sub-component of a non-central national department and lacks both the status and the 53 Verne Harris autonomy it requires to perform the auditing function. Unfortunately, this TRC recommendation loses its force through contradictory elaboration - at one and the same time it advocates independent agency status (the ideal) and positioning within either the office of the President or that of the Deputy President; • the security establishment should not be allowed to escape the operation of the National Archives of South Africa Act. While the Act brings security bodies firmly within its ambit, it does allow for various exclusionary options.18 It is conceded that a special status for such bodies appropriate to the sensitivity of the records they generate would be legitimate, but that they should remain fully subject to the professional supervision of the National Archives. The TRC (1998, vol 5, ch 8:346) also made several recommendations related to redressing the imbalances imposed on official memory resources by the purge. A number related to ensuring that the National Archives secures control over the records of the security establishment which survived the purge. In addition: • the security establishment should make every attempt to locate and retrieve documents removed without authorisation by operatives of apartheid security structures; • the South African government should acknowledge that, in terms of internationally-recognised archival principles, the extant records of the South West Africa Territory Force (currently in the custody of the SANDF Archives) properly belong in Namibia and must be returned to the Namibian government. It was noted that an agreement between South Africa and Namibia covering equivalent civilian records was already in place; • the National Archives should be given the necessary resources to take transfer of, process professionally and make available to the public, the TRC's own records (which fill many of the gaps in official memory resources); • the National Archives should be given the necessary resources to fill the gaps in official memory resources through the collection of non-public records and the promotion of oral history projects. I find the TRC recommendations compelling. It remains to be seen what the state makes of them. The signs are not good. To date there has been no formal government response to the TRC recommendations. The National 54 'They should have destroyed more' Archives has sought either to promote them or, where its mandate and resources allow, to implement them. However, in every case the National Archives is reliant on higher authority to give full effect to the recommendations. Notes 1. A substantially similar version of this paper was delivered at a conference on The TRC: commissioning the past, co-hosted by the History Workshop, University of Witwatersrand and the Centre for the Study of Violence, June 1999. Likewise, a similar version is to be published in a forthcoming edition of the South African Archives Journal (SAAJ). Transformation is grateful to the editors of the SAAJ for agreeing to allow us also to publish this paper. 2. Between 1988-94, I was an archivist in the Pretoria records management division of the SAS. Rumours were rife within the public service, and by early 1993 I had enough evidence from sources in various governmental bodies to know that destruction was widespread. When it became clear that the SAS was unable or unwilling to act decisively, I began leaking information on the destruction to the ANC, other oppositional structures and the media. The 1993 Currin case dealt with in this essay pushed the issue firmly onto centre stage in the.media. The Harms and Goldstone commissions of enquiry as well as the Goniwe inquest also revealed substantial evidence of systematic records destruction. 3. Promotion of National Unity and Reconciliation Act (Number 34 of 1995), section 4(d). 4. The Archives Act (Number 6 of 1962). See in particular the definition of a 'government office' in section 1. 5. Each of the security establishment's structures was subjected to close scrutiny by a joint team comprising representatives of the structure concerned, the TRC, the Human Rights Commission and the National Archives. 6. The work of this investigation is reflected in the following sections of TRC 1998, vol 1, ch 8 and vol 5, ch 8 paragraphs 62, 66, 67, and 100-08. 7. In terms of the National Archives of South Africa Act (Number 43 of 1996), the SAS became the National Archives on 1 January 1997. 8. The guidelines were referenced as EM 9-12. The relevant paragraphs were 31 and 32 (20-21). 9. The updated guidelines were referenced as SP 2/8/1. 10. Case No 19304/93, Supreme Court of South Africa, Transvaal Provincial Division. 11. The moratoria are dealt with later in this essay. 55 Verne Harris 12. Advocate Cilliers' opinion is dated January 13,1993, but this is clearly a dating error. 13. I quote this from my own Conference notes. 14. The TRC (1998, vol 1, ch 8:202-4, 216) acknowledged excellent support received by it from the National Archives and various security establishment structures, but noted obstruction encountered in work with the SANDF. Numerous minor instances of obstruction were not noted by the TRC. It is worth recording here that the reliability of the investigation into the records of the Security Police was placed in question by the emergence in 1999 of the Security Police file on former ANC leader Sifiso Nkabinde - at the Police office in Pietermaritzburg, which had reported that all Security Police files had been destroyed. 15. Section 32(1) of he new Constitution also recognises the right of access to information held by persons other than the state 'that is required for the exercise or protection of any rights'. 16. The Open Democracy Bill was under consideration by the state for about five years. At the last moment it was decided to exclude provisions for the protection of personal information and of whistleblowers - both to be dealt with in separate legislation. In January 2000, the legislature passed the Promotion of Access to Information Act, which defines the right of access to information held both by the state and by persons other than the state. 17. The National Archives of South Africa Act of 1996 has changed the penalty to 'a fine or imprisonment for a period not exceeding two years or both such fine and imprisonment' (section 16(1)). 18. The 1996 Archives Act allows for certain categories of public record identified by the National Archivist to remain in the custody of the creating agency rather than be transferred into the custody of the National Archives. It allows for public records to remain in the custody of the creating agency if another Act of Parliament requires this. And it allows for a governmental body to be exempted from any provision of the Act with the concurrence of the National Archivist, the National Archives Commission and the responsible Minister. References Very little has been written on the events described in this essay. The detailed account provided in the 1998 TRC Report is obviously essential reading (Truth and Reconciliation Commission of South Africa Report. Cape Town: Government Printer). For an overview of the culture of censorship within which the destruction of records took place, see Christopher Merrett (1994) A Culture of Censorship- secrecy and intellectual repression in South Africa. Cape Town and Pietermaritzburg: David Philip and University of Natal Press. 56