Worker consciousness, strikes and conflict on the mines and shop floor¶
How should African worker consciousness, faction fights and strikes be explained, and what is the relative weight of organised vs everyday resistance?
Works engaging the debate: 22
Positions¶
- Situational and structural explanations of faction fights vs tribalist accounts
- Everyday resistance as revealing consciousness (Phimister and van Onselen)
- Moral economy vs proletarianisation (Moodie vs O'Meara on the 1946 strike)
- Economic vs political character of strikes
Evidence base¶
The debate rests on two main bodies of evidence, the 1940s and the 1970s, with the decades between them barely covered. For the 1946 African mineworkers' strike and its wartime precursors, the economic case comes from Transvaal Chamber of Mines Annual Reports for 1936-50, the Gold Producers' Committee's evidence to the Lansdown Commission (UG 21/1944) and Association of Mine Managers papers (James (1987)). From these James builds national series showing a 36-39 per cent fall in domestic African labour supply, a 69 per cent wartime rise in working costs against a fixed gold price, and maize rations cut from about 660 to 550 lb. The consciousness case comes from Native Affairs Department NTS Series 332 'disturbance' reports, police and Chamber spy reports of AMWU meetings (NTS 7248), the J.B. Marks interviews and the author's 1984 interviews with Mpondo ex-miners (Moodie (1986)). Duncan's single-strike reconstruction of the 1944 milling dispute adds the departmental files NTS 199/332 and ARB LC1052/409 and press counts: 1,200-1,500 strikers and 646 arrests (Duncan (1989)). Moodie extends the method backwards to the 1913 Crown Mines Inquiry transcripts, Buckle Commission evidence, the 1919 and 1930 low-grade mines commissions and Barlow Rand company files (Moodie (2005)). The 1970s mine-conflict literature rests on three different kinds of record. First is a press-derived 'Diary of Events' of 81 incidents between October 1972 and December 1979, with 205 deaths and 1,168 injuries, cross-checked against the 1974 Police Commissioner's report, SAIRR Surveys, Hansard questions, TEBA annual reports and the Inter-Departmental (McKay) Committee report (Horner and Kooy (1980a)). The evidence ledgers (Horner and Kooy (1980b)) and Horner and Kooy (1980c) are versions of the same work, not independent corroboration. Second is McNamara's database of 282 violent collective conflicts for 1973-82, coded from Anglo American Gold Records Division reports, TEBA Liaison Division unrest files and diaries, and Chamber COMGU files, supplemented by hostel participant-observation and Human Resources Laboratory interviews (McNamara (1985)). Third is Massey's use of WNLA Annual Reports for 1921-77, Wilson's real-wage series, 116 Mine Labour Organisations blacklist records from Mafeking and 123 interviews with Botswana miners (Massey (1983)). Kynoch carries the mine record into gang networks using CAD police files, NRC and TEBALD reports, the Goldstone and Bregman commissions and oral interviews from 1998-99 (Kynoch (2000)).
Outside the mines, the evidence for worker consciousness comes from author surveys. Schlemmer surveyed 532 Durban industrial workers in 1983 and compared them with a 1982 national migrant sample of over 600 and a rural Natal sample of 856 (Schlemmer et al. (1983)). Goldin drew on a 383-respondent Cape Peninsula hostel survey and a 20-firm employer sample (Goldin (1984)). Sitas used a 14-firm Durban survey, the 1976 Manufacturing Census and the 988-interview Black Women in Industry study (Sitas et al. (1984)). Minkley worked from Border Chamber of Industries files, company archives, worker interviews from 1987-88 and unpublished census figures for East London (Minkley (1990)). Aggregate strike measurement is confined to Beittel's comparison of the government strike series, incomplete before 1917, with World Labor Research Working Group counts of 2,306 Times (London) and 487 New York Times index mentions (Beittel (1995a); Beittel (1995b)). A separate financial strand uses Business Day strike announcements for 1984-93 matched to JSE price tapes for about 50 firms (Bhana (1997)). The syntheses (Lewis (1990); Phakathi (2012); Higginson (1992)) add almost no primary data. Phakathi's desertion claim rests on a single 1910 estimate taken from Johnstone.
Contested evidence¶
On faction fights, the same document is read in opposite ways. Horner and Kooy quote the McKay Committee report as evidence of an official tribalist ideology, then reject it on the basis of their own press-derived incident counts (Horner and Kooy (1980a)). Massey instead uses the unpublished riot commission as a source for the post-1975 real-wage rise and the fall in the foreign labour share from about 80 to 42 per cent (Massey (1983)). The structural thesis in Horner and Kooy (1980a) rests on illustrative pairings of Xhosa-Basotho clashes and a long-service Basotho observation, with no mine-level data on job grade or origin. McNamara supplies some of that test from internal records. For the 1974 Free State clashes he uses a job-distribution sample (Table 13) and assault tables showing Lesotho nationals dominating better-paid jobs. He also finds that the 1982 strikes divided strikers from team leaders along occupational, not ethnic, lines. Yet he still allows a role for BCP politics and for hostel and induna dynamics (McNamara (1985)). Kynoch complicates any purely situational reading by documenting Marashea gangs acting as Basotho ethnic enforcers, for example in the 1991 President Steyn killings (Kynoch (2000)). The two sides also rest on incompatible source bases. McNamara shows that only about a third of the conflicts in internal records reached the press (McNamara (1985)). That undercuts the coverage of press-built catalogues while supporting Horner and Kooy's claim that visible riots are only peaks of wider resistance, a claim they assert without measuring it. In the Moodie-O'Meara exchange, Moodie's rebuttal of proletarianisation rests on oral testimony that the best-organised East Rand strikers were rural-tied migrants. The evidence ledger codes this claim as asserted, because no per-mine data on worker origins, tenure or urban ties are presented (Moodie (1986)). James's Chamber series describe the cost squeeze and food grievances that both camps invoke, but they say little about the miners themselves (James (1987)). Even the size of the strike is contested, with police and NAD counts ranging from 73,000 to 75,000 strikers across 21 of 47 mines. On whether strikes were economic or political, Schlemmer's survey finds pragmatic, wage-driven motives, with only a third of workers willing to answer a stay-away call (Schlemmer et al. (1983)). Beittel shows that the answer depends on the source. The official series exclude politicised stay-aways and the homelands, whereas Times-based counts pick up 1957 and 1959-60 but miss 1919-20 and 1941-46 (Beittel (1995a)).
Underused and missing sources¶
The source that nearly every evidence ledger names as decisive but no work uses is mine-level workforce data by origin, job grade, wage, contract length and hostel allocation. It is needed to test the structural account of faction fights in Horner and Kooy (1980a) and McNamara (1985), Moodie's denial of proletarianisation in 1946 (Moodie (1986)), and Kynoch's membership claims (Kynoch (2000)). The recruiting returns could test whether the origin and job composition of the workforce on each mine tracks the inter-group conflict counts for the 1970s. The 1944-62 wage files cover the Lansdown award and its aftermath, which bear directly on the wage-gap grievance that James places behind the strike (James (1987)). Sources rarely or never used include AMWU, TAMWU and CNETU union records, which would recover worker voices for 1944-46 (Duncan (1989); James (1987)). Compensation, hospital and court records could quantify the underground assault trend argued in Moodie (2005). Rural household and remittance surveys could measure the 'secondary exploitation' that Massey concedes cannot be measured (Massey (1983)). Systematic coding of South African press and police events would give an independent benchmark for Beittel's partly circular 'stylized facts' (Beittel (1995b)). Finally, the strike event studies need their underlying strike lists and dating documented before they can be linked to causes or to labour data (Bhana (1997); Estimation notes first estimations on... (n.d.)).
Strength of key claims¶
| Basis | Share of key claims |
|---|---|
| Systematic data | 8% |
| Partial data | 61% |
| Illustrative or anecdotal | 21% |
| Asserted | 9% |
Comparative figures for all debates are on the Literature overview.
Best-evidenced works¶
McNamara (1985), James (1987), Massey (1983), Moodie (1986), Beittel (1995a), Kynoch (2000)
Most-used primary sources¶
Holdings are matched by machine against the archive's file manifest; a dash means no match was found, not that the project certainly lacks the source.
Works and their positions¶
| Work | Position |
|---|---|
| Beittel (1995a) | Notes that labour history focuses on organised protest and strikes, while social history captures informal protest outside the workplace. Treats both as necessary and builds an indicator meant to capture politicised stay-aways as well as strikes. |
| Beittel (1995b) | Sees the richness of South African labour history as coming from combining both approaches, and notes that official strike statistics miss political stay-aways. |
| Bhana (1997) | Finds partial anticipation (negative drift before the event, and a reaction that varies with strike length) but systematic underestimation of costs. Concludes the results are inconsistent with the efficient market hypothesis. |
| Bhana (1997) | Treats strikes as costly negative investments for shareholders and as a lasting loss. Also acknowledges the view that strikes are integral to collective bargaining and can raise productivity. |
| Bhana (1997) | Presents employer and union views (Boyd vs Patel) without adjudicating. The study measures only share-price effects. |
| Duncan (1989) | Supports Giddens: unrecognised African strikers generated a minor degree of change through collective withdrawal of labour. |
| Goldin (1984) | His survey found no significant perceived division, challenging the 'insider/outsider' divide-and-rule hypothesis. |
| Higginson (1992) | Holds that worker militancy and class consciousness did not arise spontaneously. They were shaped by peasant millenarian ideologies, mutual-aid societies, piece-rate work and urban conditions, so workers were active participants rather than passive victims. |
| Horner and Kooy (1980a) | Rejects the Committee of Inquiry's tribalist explanation as 'crude racism'. Locates the fights in the hierarchy of jobs, differential contracts and hostel allocation by area of origin, while acknowledging that the evidence is too thin for firm conclusions. |
| Horner and Kooy (1980a) | Follows Phimister and van Onselen in treating strikes as 'high water marks' and desertion, absenteeism, avoidance of mines and similar acts as the more telling expressions of consciousness. Notes the theoretical problems in simply transplanting their approach. |
| Horner and Kooy (1980a) | Argues that ethnic mixing of the workforce is a deliberate control device that can backfire when national consciousness becomes the basis of solidarity, as in the Basotho deferred pay strikes. |
| Horner and Kooy (1980b) | Rejects the official Committee's tribalist account as 'crude racism'. Locates the clashes in job hierarchy, differential contracts and access to jobs, and compound and hostel segregation by origin. |
| Horner and Kooy (1980b) | Follows them in stressing day-to-day and unorganised resistance (desertion, avoidance, absenteeism) as revealing worker consciousness. Acknowledges their conceptualisations are not unproblematic. |
| Horner and Kooy (1980b) | Ethnic and national divisions serve management by keeping strikes partial, but can backfire when national consciousness fuels worker solidarity, as with the Basotho deferred-pay strikes. |
| Horner and Kooy (1980c) | Rejects the official tribalist explanation as 'crude racism'. Attributes clashes to differential job access, contract lengths and mine management's division of workers by origin. |
| Horner and Kooy (1980c) | Follows Phimister and van Onselen: strikes and riots are 'high water marks', and desertion, avoidance of mines, absenteeism and loafing are equally important evidence of consciousness. |
| James (1987) | Accepts the strike's importance but criticises O'Meara for giving it an enormous burden of historical responsibility with little documentation of gold mining's own social and economic processes; supplies that political-economy account of origins. |
| Kynoch (2000) | Kynoch builds on Bonner but argues that the 1950s Rand era of Russians working in secondary industry was an aberration. The 'shift from migrant to immigrant status' was never completed, and the mineworker-loafer relationship explains the gangs' survival. |
| Kynoch (2000) | Agnostic and sceptical. Management and security approval is plausible, but no documented proof exists. The Bregman Report and Goldstone found no substantiation, and Russians mostly acted to defend their own networks and Basotho solidarity. Some NUM claims appear to have used the Russians as convenient scapegoats. |
| Kynoch (2000) | Treats Russian involvement as ethnically motivated (as Basotho) and economically motivated (protecting customers and income). Their participation hardened ethnic boundaries in the compounds. |
| Lewis (1990) | Defends a distinct labor history focused on production and industrial relations, but urges that the division between it and social history be eroded. |
| Massey (1983) | Attributes it to both political change in supplier states and rising worker militancy and consciousness, not to outside agitators (the riot commission found none). |
| McNamara (1985) | Finds that both apply: workers lacked unions, but hostel institutions facilitated mobilisation, and a J-curve of 'progressive deprivation' after the early-1970s wage rises explains the strikes. |
| McNamara (1985) | Argues the Marxian approach gives the structural background but cannot predict when conflict erupts; perceptions of injustice (Barrington Moore) explain timing, since conflict followed improvements rather than immiseration. |
| McNamara (1985) | Rejects official and popular tribal-savagery explanations; attributes clashes to labour turnover, hostel tensions, job competition, induna power struggles, and political events, with ethnicity as outward form. |
| McNamara (1985) | Both matter: foreign-labour conflicts reflect regional politics, but strikes originate largely in internal wage and service-condition changes in an isolated institutional setting. |
| Minkley (1990) | Argues that race 'fused' with class. The workplace experience of black workers promoted a racial consciousness largely devoid of tangible class content, so the working class made itself in racial form. |
| Minkley (1990) | Holds that proletarianisation was partial and contested, with pre-industrial work habits, kinship and rural ties shaping resistance. Rural decline plus state coercion eventually stabilised labour. |
| Moodie (1986) | Rejects O'Meara's reductionist proletarianisation account as inadequate. Argues that resistance was rooted in a localised, defensive moral economy and in informal compound networks, and that migrant workers with rural stakes were the most effective strikers. |
| Moodie (1986) | The AMWU initiated and gave a common focus to wage demands (leaflets, mass meetings). The East Rand strike itself was carried out through indigenous room networks, with little union involvement or unionist ideology. |
| Moodie (2005) | Disputes Breckenridge's emphasis on masculinity, Fordist piecework and the allure of violence. Treats cultural factors as background and locates the cause in the maximum average wage system and the absence of incentives. |
| Phakathi (2012) | Sides with agency-centred accounts (Moodie, Allen, Cohen, Sitas): control 'could never be total' and covert and cultural resistance underpinned unionism. |
| Schlemmer et al. (1983) | Argues strikes are instrumental and economic, not political or ideological, and that political consciousness has not invaded the workplace. Allows that political factors can indirectly raise the intensity of action. |
| Schlemmer et al. (1983) | Finds union members only marginally different from others, and slightly more favourable to management than non-unionised migrants. Future radicalisation is more likely to come from rising education than from unions. |
| Sitas et al. (1984) | Attributes them to the breakdown of cheap labour, urbanisation, inflation eroding real wages, and capital concentration with greater employment stability and a collective consciousness. |