Background
The move from hostels into informal settlements, clustered around the mines across the platinum belt, has complicated union (and management) relations with workers. Mine hostels themselves, however, like all complex social institutions, were never passive structures dominated by white managers and black indunas.1 They were hives of social activity, myriad informal networks, differentiated and overlapping in manifold ways. There were friendship networks of ‘home-boys’ (often cutting across mines), drinking buddies, room groups, work groups, ethnic organisations, sports teams, criminal gangs, church congregations, ethnic associations, stokfel savings clubs, people on the make and people on the take.
The informal organisation of one particular hostel was not necessarily reproduced at another. Union organisers had to proceed strategically. They were obliged to recruit along the lines of already existing informal networks, as Donald Donham (2011) has shown in his recent fascinating ethnography of conflict on ERPM mine. Moreover, the hostel structure served as a sort of ‘echo chamber’ for protests over grievances. This enabled ready union recruitment but it also meant that the union was always skating on the surface of potential populist and sometimes violent expressions of worker outrage. Ethnic differences required close attention to avoid conflict. Organising mine workers was never easy.
Taking advantage of available spaces, recognising members’ informal attachments and building on them, but, above all, consulting with workers, staying in touch with their needs and desires, always negotiating decisions with them, was the secret of the National Union of Mineworkers’ (NUM) incredibly rapid growth in the 1980s. Without any question, the NUM in its first decade was a social movement union, riding on the already existing informal solidarity of its members.2
Any union must remain close to its members, however. Losing touch with workers, for whatever reason, ultimately spells disaster. As the NUM now has to contemplate what it has lost in Rustenburg, this factor remains fundamental, not only for the NUM but for any successor union on the platinum mines. Two articles in this issue, by Crispen Chinguno and Luke Sinwell, address directly the NUM's recent loss of support on Impala and Anglo Platinum (henceforth Amplats) mines in particular. This article will describe two cases of informal worker insurgency in the rise of the NUM and its early troubles on the platinum mines in the 1990s. The NUM in its origins at Impala Platinum mine in the early 1990s was without doubt a social movement union with strong African National Congress (ANC) networks. At Anglo Platinum (until 1995 called Rustenburg Platinum Mine – henceforth RPM) in Rustenburg and Northam, however, the NUM was obliged to deal with a hostile insurgent worker movement that eventually gave rise to a rival union.
In this article, I provide an analysis of (sometimes violent) informal solidarities in two complex and important events on the South African platinum mines in the 1990s. First, I address the NUM's struggle for recognition at Impala Platinum, based on informal worker committees in uneasy alliance with mine management but in the face of fierce opposition from the Bophuthatswana ‘government’. Second, I describe the rise of worker committees opposed to the NUM on Anglo Platinum (henceforth Amplats) mines in the years after the 1994 elections. This movement eventually gave rise to an alternative union before it petered out in 1999.
In both cases, I describe what G. H. Mead (1932, 23) called ‘an emergent event’, which, in his language, refers to ‘the occurrence of something which is more than the processes that have led up to it and which … adds to later passages a content they would not otherwise have possessed’. Such events, as I describe them, following Mead, typically ‘emerge out of the conditions of their occurrence’ and yet are explainable in structural terms only in retrospect. In order to understand insurgency, then, I argue that we need to focus not only on formal organisation but also, always, on informal relations established by social actors immersed in the ‘common sense’ (to borrow a term from Antonio Gramsci) of their everyday lives.3
There is thus a built-in uncertainty to the outcomes of emergent ‘insurgent’ events. I operate with Holston's (2008, 34) definition of insurgence, as a ‘counterpolitics that destabilizes the present and renders it fragile, defamiliarizing the coherence with which it usually presents itself’. I concur with Holston that, while insurgent movements may take similar emotional and organisational forms, they ‘are not necessarily democratic or just, socialist or populist … . They bubble up in … places where present circumstances seem propitious for an irruption.’ While the causal effects of insurgency may be obvious in retrospect, its impact brings a fundamental ambiguity to analysis of the present.
Conflict, class compromise and the politics of management at Impala Platinum
General Mining (controlled by the Afrikaner company Federale Mynbou) and Union Corporation became one company, Gencor, in 1980. They struggled for several years to develop a coherent managerial policy although in the mining division the manager, Johan Fritz, established a reputation for tough top-down control. His main antagonist in the company was Naas Steenkamp, who reported to him but headed industrial relations at the company and represented it at the Chamber of Mines. Steenkamp, the most liberal member of the Wiehahn Commission,4 was an ardent proponent of a negotiated approach to labour issues.
Impala Platinum was one of Gencor's most successful mining ventures but it was located entirely within the confines of the Bophuthatswana Bantustan state, which, under Lucas Mangope, had a relatively liberal constitution but totally draconian labour legislation – often at odds with one another. Bophuthatswana's industrial relations policies created few problems for Johan Fritz. An old-style mine manager, Fritz, as he told the Financial Mail (9 September 1985), firmly believed that ‘blacks have a different cultural approach. They want to know exactly where you stand otherwise they lose confidence in you.’ As a result, and for reasons of underground safety, one needed to ‘maintain a structured approach toward mining methods and discipline’. He concluded: ‘It is important to remain consistent with black employees and not to vacillate.’
Not that Fritz was particularly open to suggestions from his white subordinates either. Already in 1984, his intervention, in the face of alternative recommendations from virtually every other managerial participant, had prevented resolution of an NUM-initiated strike at Impala Refineries.5 The entire workforce was eventually fired. A similar story could be told about the strike at Marievale gold mine in 1985, where the Industrial Court, in a precedent-setting ruling, ultimately decreed that unfair dismissals had taken place and obliged the mine to rehire the entire labour force of dismissed workers.
Most important, and much closer to home for our purposes, was the dismissal on 7–8 January 1986 of over 25,000 workers from Impala's platinum mines. Impala had refused access to the NUM, citing the Bophuthatswana Industrial Conciliation Act, which forbade South African unions from organising in Bophuthatswana. A worker's committee, late in December 1985, presented management with a list of grievances covering wages and working conditions (accompanied by a brief work stoppage) and demanded recruiting facilities for the NUM. When there was no management response, the entire Impala workforce went on strike early in January 1986. Fritz called in the Bophuthatswana police to break the strike. NUM officials who came to the mine were refused a meeting with management.
The largest dismissals from one mine hitherto in South African history followed. President Mangope of Bophuthatswana was profoundly satisfied with the outcome, but for Impala the whole business was a disaster. Wildebeestfontein North, which had not dismissed its entire workforce, was back to full production in two weeks. The other three Impala mines took six to nine months to return to normal. Indeed, Johan Fritz himself was likely a victim of the fallout from these dismissals.
In April 1986, three months after the Impala mass dismissals, Derek Keys was appointed as Chairman of the Board of Gencor. Fritz, who apparently thought he should have had the job, resigned in disgust. Naas Steenkamp, who had been about to be fired by Fritz, set out to reform Gencor's industrial relations policies. In the Chamber of Mines, Gencor moved sharply 180 degrees from the hard-nosed Goldfields camp into what was perceived to be the ‘soft’ Anglo camp. Brian Gilbertson moved from RPM to head up the mining division at Gencor, and Michael McMahon (also from RPM) took over management of Impala Platinum. Gideon Albertyn, a protégé of Steenkamp, was appointed to head industrial relations at Impala.
The territory mined by Impala was held in trust by the ‘independent’ state of Bophuthatswana. It was, however, registered ‘in trust’ to the Bafokeng tribal authority under Chief Edward Lebone Molotlegi, who was also a major political rival of President Mangope in Bophuthatswana politics. Two rival jurisdictions – the homeland state and the tribal authority – thus competed for royalties from Impala mines.6
In 1988, a successful coup attempt against Mangope was put down by the South African Defence Force. Lebone Molotlegi was held for a while by the Bophuthatswana security forces. Once released, in poor health and fearing for his life, he fled to Botswana. Mangope immediately arranged for Chief Lebone's brother, George Molotlegi, to become chief of the Bafokeng. Trustee state and tribal authority swung into line. George Molotlegi promptly signed a lease with Impala that guaranteed (under Bophuthatswana law) a further extension of their mining rights to a section called ‘The Deeps’ which they had been eyeing for years and which ensured their survival as a platinum mining company.
Lebone Molotlegi's wife, Semane, continued to lead opposition to Mangope from within the Bafokeng tribal authority, however. She was so effective at this that she was eventually banished from Bophuthatswana. Steenkamp, at Gencor, promptly offered her a job in the company. Gencor also provided scholarships for her and Lebone Molotlegi's children. Deep divisions within Gencor/Impala over how to deal with the Bafokeng and about who was entitled to benefit from Impala's royalty payments overlaid the already simmering dispute between representatives of the Bafokeng and the Bophuthatswana government. This is a complex matter to be dealt with elsewhere, except to note that his conflict with the Bafokeng certainly fed into Mangope's vehement hostility to McMahon's new regime at Impala as the ANC was unbanned in South Africa.
Events took a new turn in June 1991.7 As required by Bophuthatswana law, Impala negotiated with Mangope's sweetheart union, BONUME (representing, at most, 7% of the Impala workforce), a roughly 10% raise for its workers. NUM, meanwhile, was negotiating for approximately 13% at Rustenburg Platinum Mines across the ‘border’ in South Africa on top of already higher wages. At the beginning of July, as soon as the BONUME-negotiated wage settlement was announced, Impala workers at Bafokeng South mobilised informal networks of outraged workers and marched ‘from the bar’ to a mass meeting at the stadium where they elected a workers' committee of 10.8 Bafokeng South workers staged a one-day strike on 4 July.
After meeting with Rustenburg NUM regional officials on 15 and 22 July, Albertyn reported (Memo, 1 August 1991, NIS archive) that the NUM ‘claim[s] to represent between 30% and 40% of the workforce. We made it clear that we were not meeting to negotiate. But when a major dispute is looming, we do not refuse to listen.’ In the midst of these informal negotiations, on 16 July, Bafokeng South ‘dismissed a strike leader for intimidation’. According to Albertyn's account, this triggered a strike at Bafokeng South on 17 and 18 July.
Once the workers were back at work, the employee was reinstated, however. The reason was that there had been ‘a lack of evidence of intimidation’. Albertyn reported ‘few incidences of violence’ overall, although, in his opinion, ‘intimidation did play a role throughout and the situation is tense … . Younger, local workers are providing the leadership “in the field” and at least some of the leaders are NUM supporters.’ Indeed, according to press reports, ‘two vehicles belonging to hostel dwellers were set ablaze on Thursday and an attempt was made to burn one of the administration buildings' (Citizen, 20 July 1991).
In fact, a double-headed new informal worker regime was being established in the mines at Impala.9 While there were nascent informal NUM union structures, ANC marshals had been established by Isaac Mayoyo ‘after observing how they operated in COSATU’. According to the report of the Pretorius commission, at Bafokeng North (Mayoyo's mine),
the marshals operated broadly in parallel with the official structures of the formally unrecognized NUM. The marshals were typically younger and more militant mineworkers who attempted to put their stamp on the way work and leisure was conducted at Bafokeng North (and elsewhere). (Pretorius 1992)
It is my impression, from management evidence in the NIS archive, that, intending to correct the hard-nosed despotism of the Fritz regime, Albertyn had used guarantees of the right to free association in the Bophuthatswana constitution to permit mass meetings in the compounds ‘as long as they were not violent’. Workers on the compounds were given freedom to organise their own lives as long as this did not affect production. Similar experiments in ‘worker control’ of the mine compounds were also being conducted at some of the gold mines during this period.
I have already mentioned, citing the work of Donald Donham (2011), that the particular interests of powerful informal networks on the mines sometimes conflicted with general worker interests. At Bafokeng North, some of the violence by marshals seems to have been exercised on behalf of Isaac Mayoyo and his associates’ business interests. According to the Pretorius commission (1992, 14), in early July 1991 ‘facilities linked to the hostels such as the bar and canteens as well as an adjacent trading store’ were closed by ‘worker pressure’.
While boycotting liquor outlets and mine stores was a common way of confronting perceived management interests on the gold mines, at Bafokeng North ‘one of the results of these closures and, arguably, one of the objectives, was that business in the relevant goods and services was shifted to an informal trading area [adjacent to the mine] known as the “Mayoyo Sun”.’ Certain employees of the mine (obviously including Mayoyo himself) ‘were alleged to have had a financial interest in the various businesses plied at the Mayoyo Sun’ (Pretorius 1992, 14–15). Among the 500 pages of internal company documents provided to the Pretorius commission were management memos noting laconically that the Mayoyo Sun traded in ‘the three Ds’ – ‘Drank, Doos en Dagga’.10
There can be no question that the establishment of marshals was a major reason for the meteoric rise of support for the NUM on the mine. Nonetheless, the tendency for marshals to use violent methods of recruitment and control also led to ‘considerable tension between the two chains of authority’. Albertyn's suspicions of ‘intimidation’ were certainly not without foundation.
South African political ferment after Mandela was released and African political movements were unbanned in 1990 thus had a direct effect on industrial relations on the mines. ‘Marshals’ had also emerged on some of the gold mines. They seem to have represented what was earlier called the ‘comrade element’, young, enthusiastic union members who marched around the compounds toyi-toying and singing political songs. They were always an important recruitment medium for the NUM but, without strong union leadership, they also could get out of hand. In the case of Impala, at Bafokeng North, witnesses to the Pretorius commission described being beaten by those they described as ‘comrades’. Local Tswana-speaking marshals had indeed taken over leadership in the hostel, replacing the induna system in handling complaints.
Marshals had also intervened in conflicts underground between workers and team-leaders. When ‘go-slows’ underground in July and August started to affect production at Impala, it was not at all clear whether they had been initiated by the informal worker committees or by the marshal organisation or whether the two somewhat uneasily overlapped. Certainly the union was tarred with the marshal brush. There were rumours that white miners and supervisors were taking guns underground. In response (Pretorius 1992, 16), ‘marshals began commandeering Mine vehicles and also searched persons and vehicles entering the hostel area, ostensibly for weapons.’
The NUM's own evidence to the Pretorius commission at Bafokeng North confirmed that ‘the political climate contributed to the leaders losing control and discipline over certain marshals.’ This, it was claimed, was at least partly a function of ‘lack of access and legitimacy for NUM within the mine’. Nonetheless (Pretorius 1992, 32), the ‘NUM conceded that the marshals had gone out of control and stated that it had in fact called for the dismantling of the system’. There is little evidence that this call from the union leadership was ever heard, let alone implemented.
Immediately after the work stoppages around the BONUME agreement, when Impala management consulted with the NUM regional office, they urged the NUM to apply for recognition under the Bophuthatswana Industrial Conciliation Act. Such recognition for an unregistered union would have enabled it to set up an office in Bophuthatswana, submit its constitution and operate in Bophuthatswana without being registered. Registration of ‘foreign’ (i.e. South African) unions was, of course, not allowed. The NUM, however, adamantly refused to grant even such limited recognition to Mangope's regime.
On 15 August, short shifts were initiated across all the Impala mines except Bafokeng South. The night of 16 August, eight people were injured in collective conflict near the Wildebeestfontein North hostel. Rumours abounded about the nature of this conflict. Workers believed white managers had shipped in workers from other mines to disrupt the go-slow movement. It is possible that conservative white miners or low-level managers had organised this fight. Similar accusations were certainly made earlier on the gold mines but convincing evidence always seems to be lacking.
On 18 August, in response both to the go-slows and the 16 August violence, Impala's top management team met in Rustenburg to discuss strategies to deal with the ongoing unrest and industrial action. One senses from the notes of this Sunday meeting managers struggling for firm ground on the shifting sands of the general situation. While worker takeover of the hostels might have been tolerated, interference with underground work was not. Mass dismissals were not considered an option. This is the point at which, for the first time, management decided on direct confrontation with violent informal ‘comrade’ cadres, while at the same time conceding that they had no option but to negotiate with the NUM.
At this point, James Motlatsi, President of the NUM, showed up to address a huge mass meeting at Impala and soon afterward, the NUM approached Impala with 26,000 completed membership forms. Wage negotiations with ‘worker committees’ (with professional assistance from NUM head office) immediately commenced and ultimately led to a substantially improved agreement signed in September. Such wage improvements were, however, conditional on improvements in worker productivity. Whatever ‘order’ might have been created by the marshals in the compounds, management was clear that their own authority needed to be predominant underground.
Matters came to a head at the very point at which Impala management signed the improved wage agreement with the NUM worker committees. Within a day or two of the agreement, marshals at Bafokeng North dragged a Basotho dagga-dealer (not even a worker) onto the tennis courts at the mine, doused him with petrol and set him alight. Apparently they were acting on behalf of Tswana dealers at the Mayoyo Sun, whose sales were being undercut by the Basotho. Insurgent marshal networks were proving beyond doubt that they were loose cannons. The ‘foreign’ Basotho at Bafokeng North fought back and nine men were killed.
According to documents in Steenkamp's private archives, Gilbertson and McMahon were summoned to a meeting on 1 October with Mangope and his immediate cabinet. Mangope dressed them down. They were instructed in no uncertain terms that the way to deal with industrial unrest was to get rid of the NUM, preferably by mass dismissals. They were sent off with a demand that they come up with a plan to ‘fix their mess or face direct government action’. McMahon's advice to Gilbertson after the meeting was that it was not Impala's responsibility to take action against employees who broke the law. That was the government's job although the mine would cooperate with Mangope's police.
He did add, and here is where the impact of the marshals on production was finally gaining management attention, that they had been ‘considering closing a shaft or two as a “disciplinary” measure anyway to regain management control of the workforce and to attempt to recapture former efficiencies’. They could do that, he said, and ‘sell it to the Bophuthatswana government as an action in response to their demands’.
In a memo to Impala top managers on 14 October, Albertyn made a clear distinction between the workers’ committees and the marshals:
In our mines the worker representative committees have established themselves. It is possible to curtail the operations of the marshals, even with the help of the committees. It would not, however, make sense to declare war on the committees. To end all discussions with the NUM means declaring a war on the committees. It cannot be done. A warning was issued to all employees on 9 October: shafts that do not reach their objectives could be closed for at least two weeks.
At this point, divisions between the worker committee and the marshal group at Wildebeestfontein North mine11 handed an issue to management on a platter. Steve Makalela (a leader of the marshals on this mine) was ‘summoned to a disciplinary hearing for absenting himself from work’. After the hearing, Makalela ‘accused the workers’ committee of working with management and refused to attend the next meeting’. On 15 October, while members of the workers’ committee were attending an NUM shaft steward training course, Makalela called a meeting of 8000 workers at Wildebeestfontein North and ‘verbally attacked committee members’. On the night shift the next day, 7000 workers staged a sit-in. Management sent a car to Rustenburg to fetch the workers’ committee. The NUM regional secretary spent three hours underground talking the workers out of the mine.
At this point, management took the opportunity to announce that it was closing Wildebeestfontein North for 11 days, adding that it might close other shafts permanently as labour unrest and low platinum prices eroded profitability. On 17 October, 8000 Makalela supporters (no doubt summonsed by ‘his’ marshals) gathered in the main hall. Close by, NUM officials and the workers’ committee met to try find some way to sort things out.
According to Zikalala (1992, 34), ‘the situation was explosive.’ Members of the workers’ committee and the NUM officials
decided that Makalela had almost destroyed the gains made by unions and workers, but ‘for unity's sake' the different parties in the committee were asked to put personal differences aside. They decided once again to request management to open the mine.
This was not done until 27 October, however. Meanwhile, mine security began systematically arresting and dismissing members of the marshal group.
The workers’ committee, or what was left of it, ‘organized a third coordinated strike action’ on 28 October across all the Impala mines. At Bafokeng North, Robby Drummond, the general manager, was attacked when he entered the hostel to address workers gathered outside the kitchen area. Mayoyo intervened to save his life. In the midst of the chaos, ‘NUM officials called in by management to assist were arrested by the Bophuthatswana police force, and later released.’ At this point, according to Zikalala's account:
Management agreed to open the mines and asked for union assistance in getting workers back to work. They also agreed that everyone dismissed without proper hearings could appeal and be represented by a Committee member. Hearings would begin after five days of normal production. Reinstated workers would receive back-pay. ‘This was a breakthrough for us’, commented the NUM's regional secretary.
It was also a breakthrough, or so it seems to me, for reform-minded managers, who had been able to use the threat of mine closings and dismissal hearings to get on top of both the violence and wildcat industrial actions and involve the worker committees in dealing with both – and at the same time to demonstrate to the state that they were capable of decisive action.
On 31 October, Impala management met yet again with Mangope, who insisted it ‘forget NUM altogether’. This was something Impala management was clearly not willing to do. Late in the year, McMahon met one night with Marcel Golding of the NUM and agreed to appoint the NUM as ‘Labour Relations Advisers' to Impala at a consultancy fee equivalent to membership dues for their 26,000 members. In exchange Golding agreed to apply to the Bophuthatswana government for recognition (not registration). At Impala, a class compromise had been sealed not only because the union was strong but also because the NUM and Impala management had formed an alliance against Mangope.
Meanwhile, in March 1993, a further bout of factional violence led management and the NUM to close ranks against the marshals. Indeed, Mayoyo and Makalela together addressed a huge mass meeting condemning violence against ‘foreigners’. Basotho workers who had fled returned to the mine. The ‘Mayoyo Sun' settlement was bulldozed to the ground and its denizens scattered across the Bushveld. Marshal militance ended.
When Mangope's regime finally fell in 1994, for all the turbulence we have seen in this account, the NUM was in a better position at Impala than it would ever be at Rustenburg Platinum Mine (RPM, eventually Angloplats) or any of the gold mines. As a result of Golding's settlement with McMahon, for instance, the union had a substantially larger proportion of management-paid full-time shaft stewards than any other mine in South Africa.
Management's relative moderation after the departure of Johan Fritz, along with the insurgence of informal worker committees (initially with marshal backing), had made possible class compromise in the face of the intransigence of Mangope's despotism. This had created what seemed to be a very strong industrial relations system that successfully survived for almost two decades despite ongoing turbulence at neighbouring Amplats mines.
Generous management remuneration for full-time shaft stewards at Impala meant that they were able to live in decent housing. Union representatives began to lose touch with workers in the hostels and informal shack settlements whom they claimed to represent. There were nonetheless years of relative industrial peace before things turned very sour again. But that is another story, addressed in this issue by Crispen Chinguno.
Problems at Amplats: Five Madoda and the rise of the Mouth Peace Workers Union
In 2001, David Radibotsing, an NUM organiser, told Gavin Capps in interview that management at Impala was ‘weak’ and that that had ensured relatively decent conditions for members of the union. The problems, he said, were much greater at Amplats. This is the other case to which we now turn our attention. Again, our focus will be on informal worker organisation, this time against, rather than on behalf of, the NUM.
In March 1994, when Mangope's regime finally collapsed, all the miners employed at Impala were paid out their contributions to SEBO, the Bophutatswana government pension fund. Half the workers at RPM's Union section, located in the Bakgatla-ba-Kgafela tribal authority area (see Capps and Mnwana in this issue) up in the north, had also fallen under Mangope's government. Workers on that half were paid out their pensions. Their fellow-workers on the South African side were (of course) paid out nothing. Retirement benefits on the South African side of the border had been vested in a provident fund to which workers had access through the NUM. Moreover, in 1995, RPM changed its name to Anglo Platinum.12 Workers, to whom the unbundling process was not properly explained, understandably saw it as an opportunity to demand payment from the NUM-instituted provident fund.
An informal group of leaders, who became known as the Five Madoda, approached mine management at the Union section, initially simply for a payout of their contributions to the provident fund. This committee was made up of traditionalist Transkeian migrants. The Five Madoda took the Amplats mines to the north of Rustenburg out on strike and management on both the Union and Amandelbult sections of Amplats, who seem to have had extraordinarily wide discretion, paid out workers’ accumulated provident fund benefits. It is difficult to know why these managers decided to pay. Perhaps they just wanted to keep the peace, but perhaps they also wanted to discredit the NUM.
Whatever the case, the provident fund payout encouraged the Five Madoda to up the ante. By paying out the provident fund, management had provided the Five Madoda workers’ committees, by now consisting of close to 100 workers, with all the ammunition they needed.
On about 6 June 1996, they called workers out on strike again, demanding additional refunds of ‘death benefit payouts; long service bonuses, platinum bonuses, PAYE contributions [and] UIF [unemployment fund] contributions’. None of these demands, of course, could be met. For three weeks, no-one showed up for work.13
In 2013, Siphiwe Mbatha, conducting research for Luke Sinwell on contemporary workers’ committees, happened upon two men in the Sefikile area who had been at the centre of the original Five Madoda movement. Their account casts a baleful light on worker committee attitudes to the NUM. ‘We asked them to represent us as workers to the company that was leaving to help us get our money from the company,’ they said. ‘They failed to talk to management. They never came back with anything to us.’ His companion added: ‘We realized that NUM was not coming with anything for us. Instead they bought new cars.’
After three weeks, on 25 June 1996, Amplats fired all 28,000 striking workers. Many dismissed workers hung around in informal settlements and villages around the mines. This was most notably the case for Sefikile village, clustered up against a rocky outcrop, adjacent to Amplats’ Union Section, where the trouble had all begun. This ‘mountain’ became the headquarters of Five Madoda workers’ committee from which they assayed forth to meetings. Late in July, the Five Madoda organised a march of about 4000 workers to the Union Buildings in Pretoria, where they demanded a meeting with Nelson Mandela. He told them to approach the NUM president, James Motlatsi, whom ‘he trusted’, and he also called in Tito Mboweni, the Minister of Labour at the time, to help resolve the strike.
Many of the striking workers returned to the kopje at Sefikile where they were kept under close surveillance by local mine security forces. On 31 July 1996, Cornelius de Beer, a security officer who was filming worker activity on the kopje, was stoned to death. De Beer's death seems to have galvanised negotiators into action. James Motlatsi , on 1 August 1996, ‘at a meeting held in Rustenburg at NUM's offices … tried to persuade the leadership of the Committee to reapply for their jobs … . No progress was made.’ Tito Mboweni was called in. Eventually, ‘after two days of intensive negotiations, a deal was struck on 3 August' (SA Labour Bulletin, October 1996, 51).
From the workers’ point of view, it was an exceedingly bad deal. Because of worker committee hostility, the NUM had held back from Mboweni's negotiations. Now it had to bear the brunt of renewed hostility occasioned by the agreement. Workers returning to the mines were told they had to apply for work anew. The Labour Bulletin article cites one miner's shock: ‘How can they do that to us? I have been working at this mine for the last eight years, now I am going to be re-employed as a new person. All that I have achieved, promotion, high grading has been taken away. It is not fair.’ Moreover, not only did the workers have to sacrifice seniority, they also had to reapply for union membership. NUM membership at Amplats mines plummeted from 12,838 in August 1996 to 4497 in September. It did not fully recover until July 1999.
The rump of the Five Madoda movement, a few hundred workers and many of the leaders, refused to reapply for their jobs and retreated to the outcrop at Sefikile. Their forays into the neighbouring countryside eventually led to full-scale police and military action against the dismissed workers ‘on the mountain’ behind Sefikile village, including use of a helicopter and a coordinated invasion of the outcrop area from several different directions. Accounts of deaths vary between four and six. Siphiwe's informants said that many bodies were not counted and some were discovered only later by women gathering wood in the copses on the hillside. Shades of events at Marikana in 2012 come irresistibly to mind.
Leadership of the Five Madoda movement had meanwhile split into several groups. Some came back to the NUM (especially at Amandelbult section) but the majority hived off to found a new union that would continue to pursue their ‘five demands’. They consulted with a lawyer, Caesar Bungane, who suggested that they come to his base in Carletonville and set up an office. There, with Bungane and his buddies, brothers Piet and Matt Joubert, was founded the Mouth Peace Workers Union (MPWU). It was formally registered under the Labour Relations Act on 17 June 1997. Meanwhile, the NUM found itself embattled. At union report-back meetings from March 1997, NUM shaft stewards were confronted by vociferous groups of workers whom the MPWU had picked up from the insurgent Five Madoda movement.
A statement in the NIM (1997) documents, dated 19 June 1997, by Lucky Ngqeleni from Amplats Bleskop shaft in Rustenburg, carries ironical echoes of the reign of the marshals at Impala. He was coming back from his home in the Eastern Cape on Sunday 8 June, he reports, when ‘a group of workers calling themselves Workers Mouth Peace’, returning from an MPWU meeting, ‘assailed my house in the nearby squatter camp armed with an assortment of dangerous weapons’. One of them, a certain Masithela, entered the shack and, without a word, kicked the leg out from under a table so that precious glass objects fell on the floor and broke.
He then told my wife that they are going to get me and kill if I do not help them to their death benefits … which they claim are held by the company. They left threatening to come back and set my shack on fire.
The next day, on his way to work, Lucky met an acquaintance, Samuel, ‘who is also a leader of the mouth peace. Samuel told me that it had come to his attention that his marshals have been to my house and damage was caused.’ Lucky asked Samuel to come to his house to inspect the damage. The story continues:
He agreed to heed my suggestion but added that he will arrange for two more persons to accompany him for safety. I asked him whether their marshals are controlled by their Committee or by himself and the leaders of the Workers Mouth Peace. He answered that they themselves have direct control over their marshals. Whilst at my house during lunch with me present, he inspected the damage and thereafter advised me to go to the police and tell them that damage was caused by unknown persons and that I am still investigating who did same. I told him that I will give the name of Masithela to the police as my wife had mentioned that name. We then parted ways with Samuel. I was advised by my own Committee of Kanana squatter camp that they will take the matter up themselves and I should not rush to go to the police. To date I have not heard from the said Committee.
Note the similarities between Lucky's story and what we have learned of informal worker action at Bafokeng North. At Impala, however, damage to property tended to involve damage to mine property. In the informal settlements, damage was to private property. There are several accounts of shacks being burned down by vigilante ‘marshals’. The ‘squatter camp committee’ referred to by Lucky seems to have been independent of the MPWU. Soon after Lucky wrote his statement, however, things would get much worse.
Because of an NIM interview with NUM members at Amandelbult, our information is most complete for that mine. Adjacent to Amandelbult, an informal settlement called Smasher Block was dominated by Xhosa-speaking migrants. Apparently, Smasher Block residents were obliged to belong to the MPWU and had to attend union meetings on Sundays. The MPWU/Five Madoda committee met at the home of a certain Phedla.
In May 1997, Phedla was dismissed because he had clocked in for work and then not shown up for duty. According to NUM interviewees:
WMPU [sic] supporters mobilized and tried to prevent NUM from going to work. They stood at the gates to the hostels armed with traditional weapons. During the strike, NUM workers were assaulted and the NUM committee approached the mine manager asking for more security.
On the second day, a large armed security contingent accompanied by the police arrived, searched the hostels and disarmed the MPWU. Next day everyone reported for work. No doubt the NUM was perceived to be in the pocket of management.
Matters at Amandelbult finally reached a crisis on 22 June, when, at a meeting in Smasher Block, Phedla reportedly ‘said they would kill Eric Rixi [and Clifford Metsing] before close of business on the 23rd. MPWU workers returned to the hostel singing songs about Eric and Clifford saying they were going to die.’ Next day, on 23 June, Eric, who was an instructor at the training centre, was at the change room. He heard someone calling him, looked out of the change room and was shot by men wearing balaclavas who were seen running away.
Rixi was the first in a spate of hit killings, ranging from Rustenburg and into the Transkei. Bruce (2001) identifies 13 hit killings in a period of three months, eight more in the next eight months and then eight more in the four months leading up to the end of 1998. Not all were NUM members, but a substantial majority were, many of them shaft stewards. Such murders are reminiscent of methods of the Mfelandawonye from the Tsolo/Qumbu area of the Transkei. ‘Although the organisation was launched in Gauteng’, the Conflict Resolution Consortium (CRC 2000, 18) asserts, ‘it has branches in the Tsolo/Qumbu villages and other areas outside the Eastern Cape, including the mining areas of Rustenburg.’ In a context of general rural poverty, migrant mineworkers from the Transkei were able to use their earnings to accumulate stock at home. Stock theft by men who chose not to migrate was a major problem for them. The Mfelandawonye responded with a swathe of violent killings.
Xhosa migrants to the platinum mines, however, tended to be from the area around Mqanduli, with not many from Tsolo at all. Xhosa-speaking groups in the Rustenburg area, according to the CRC (2000, 36), ‘also organized themselves into home-guards or vigilante groups, subscribing to the same ethos as the Mfelalandawonye, but calling themselves Iliso Lomzi’. This organisation, we are told, took it upon itself to watch over the interests of Xhosa migrants, stretching out from informal settlements on the platinum belt to sending areas in the Transkei. Members of the organisation employed on the mines contributed R50 a month to its coffers. It is difficult to believe that the original Five Madoda were not associated with this group.
Whatever their organisational base, killings and counter-killings continued until early 1999, when, on 21 April, 18,000 workers at Amplats’ Rustenburg and Union sections came out on a one-day strike because management was insisting on the NUM's inclusion in discussions about MPWU demands that the provident fund be paid out and a new one set up under their auspices.
Siphiwe's old Five Madoda interviewees were revealingly irate about this move, charging:
Caesar Bungane sold us … because we said that now that we have a union let us go and fetch our money from JCI. So we found out that the money was still there, but they [the MPWU] said the money must be in their accounts, then they will be the ones who are going to give us … . When they took the money, they said they want the money to be transferred to the Mouth Peace account. We then said, no, we cannot allow our money to be in a lawyer's account, so that is when not seeing eye to eye started and some of the people left their jobs. They came to the mass [meeting] and told us what was going on and some were arrested. Caesar got them arrested and then we decided we are going to take them out and we went to look for another place in Johannesburg.
After the abortive April 1999 strike, the MPWU broke up into several fragmented union groups. In any event, the hit killings abated and the rump of the MPWU began to take on a more conventional trade union stance. NUM membership at Amplats rose from 5832 in October 1998 to 12,989 in July 1999 (NUM statistics n.d.).
Bruce (2001) argues that the MPWU (or was it the old Five Madoda movement, or perhaps Iliso Lomzi?) continued to dominate informal settlements around the Amplats shafts, running a vigilante organisation that effectively maintained ‘order’ in those totally unserviced and unpoliced areas at least through 2001. Certainly, for future union organisers, informal settlements around the mines continued to be either an asset, if they could link up with the groups who maintained order in them, or a threat, if they could not. In important ways, organising focus tended to move from hostels to informal settlements and in the attempt to bridge the two (see Bezuidenhout and Buhlungu 2011).
South Africa entered the twenty-first century with the NUM seeking to make up lost ground on the platinum mines by militant action. On 17 September 2000, Amplats workers came out on strike for a 10% wage increase. When an agreement was reached with the smaller unions for 9% on 20 September, Archie Palane, deputy general secretary of NUM, who was the NUM head office official directly responsible for Rustenburg, kept the NUM out on strike. At this stage, according to the Mail & Guardian (29 September 2000) the NUM represented barely 30% of the Amplats workforce. By that time, according to Business Day (29 September 2000) the rump of the MPWU represented fewer than 20%. As late as 2004, however, Paul Stewart (2006, 13) reports that 50% of the drill operators on Frank Shaft in Rustenburg were still members of the MPWU. Only 30% belonged to NUM.
It was not until 17 October that NUM members went back to work – without an improved offer. Things did not bode well for the union. By this time, however, Amplats management had apparently got the message. According to Stewart (12), they finally launched a full-scale restructuring process to negotiate an Employee Relations policy that was eventually signed off on ‘by all the trade unions, staff associations and management stakeholders’ in 2002. Palane had worked hard to achieve this agreement, which, on his account (fin24, 10 September 2012), seems to have included a 30% plus 1 threshold for union recognition.
The establishment of a more regularised industrial relations system, further consolidated at Impala, seems to have excluded important independent worker groups, however. These predictably consisted of older Xhosa-speaking migrant men who, since the demise of Tswana ANC marshals at Impala, had become the most insurgent element on platinum. In March 2003, for instance, Impala negotiated a new funeral benefit with the NUM. Eighteen thousand workers (virtually the entire workforce) came out on a ‘Five Madoda-type’ strike that soon expanded into a demand that the company pay out their accumulated dues from the provident fund as well. Eventually Impala agreed to return to the previous funeral benefit and to discuss the issue of the provident fund (Mining Weekly, 28 March and 4 April 2003). Insurgent worker committees thus continued to exist on both Impala and Amplats mines. They dragged the NUM into confrontations that frequently disturbed the established industrial relations system. For an informed discussion of contemporary worker committees at Impala and Amplats, see the contributions of Chinguno and Sinwell to this special issue.
Conclusion
This historical account of two emergent events of worker insurgency on the South African platinum mines seeks to delineate the importance of informal networks among different groups of workers and also in management and ‘government’ circles. The concatenation of these events coalesced over time into different patterns of worker struggle and management and state intervention to produce radically different outcomes in each case.
We need then to return to the two points I made at the outset. First, that analysis of social conflict must focus not only on formal organisation but also, always, on informal relations constituted by social actors in their everyday lives. This of course applies as much to managers and political functionaries as to workers. Neither ‘capital’ nor ‘the state’ act in unmediated ways.
Second, it is important to recognise that ‘insurgency’ serves the interests of participants in sometimes contradictory ways. In the case of Impala in the early 1990s, on the one hand, the NUM took off at least partly because of insurgent agency from marshals, but worker committees and the union also paid a price for marshal violence, finally establishing a class compromise in the face of government intransigence that enabled both improved conditions and stellar production through the later 1990s and into the early 2000s before running aground on a new pattern of insurgency. In the case of the original Five Madoda movement, combined with Amplats mine managements’ extraordinary independence of centralised authority, on the other hand, insurgency disrupted overall worker solidarity in pursuit of the material interests of certain groups of migrant workers. Class conflict seemed to be looming within the union itself. Chinguno's contribution to this issue addresses this question.
This article does not argue for a direct causal connection between the two cases here described and contemporary events. Despite charges to the contrary, there is no evidence of any link between the MPWU and AMCU, although both were indeed rooted in insurgent worker committees. Insurgent unionism on platinum produced an unpredictably mixed impact on unions and management alike. As with contemporary insurgent uprisings on the platinum mines, however, the cases described in this article were emergent events indeed.