The southern African regional formation: a geographical perspective¶
Jonathan Crush (1982). The southern African regional formation: a geographical perspective. Tijdschrift voor Economische en Sociale Geografie 73(4). Royal Dutch Geographical Society / Blackwell.
| Type | journal article |
| Scope | Economics and economic history of apartheid |
| Method | mixed |
| Period studied | 1860–1914 |
| Regions | Southern Africa, Lesotho, Swaziland, Natal and Zululand, Orange Free State, Transvaal, Mozambique, Zimbabwe, Botswana |
| Themes | origins segregation, mining migrant labour, agriculture land, bantustans removals influx, comparative theory |
| Citation key | crush1982a |
Argument¶
Crush argues that the southern African regional formation (c.1860-1914) cannot be explained by generalisations about capitalist penetration, coercion and 'labour reserves'. African societies were incorporated unevenly, and the outcome depended on three factors: the form of capital, the internal dynamics of each African social formation, and the local ecology. Drawing on Kimble's work on Lesotho, he claims that Basotho migrant labour and grain production grew out of the existing chiefly social division of labour, not from external compulsion alone. The Swazi case shows reserves being designed by the colonial state to preserve chiefly power, and sized for subsistence, not to push people into wage labour. He proposes a de-mystified, low-level version of 'articulation' concepts that geographers can use to disaggregate migrant labour and commodity production.
Evidence and method¶
Secondary literature on Nguni, Basotho, Pedi, Swazi, Zulu, Pondo and Mozambican societies. Kimble's unpublished thesis and papers on Lesotho (grain exports, migrant numbers). Swaziland National Archives file (carrying-capacity calculations for 32 reserves). Christopher (1976) and Palmer (1977) tables on land expropriation and man-land ratios. Maps of Swazi royal and chiefly locations.
Place in the debates¶
- Articulation of modes of production vs. world-systems and functionalist accounts of capitalist penetration (see Cheap labour thesis and the function of the reserves). Rejects world-systems universalism and the high-level 'mode of production' abstraction. Keeps a modified, de-mystified articulation approach that treats African societies' internal dynamics as active.
- Labour-reserve functionalism (Wolpe, Burawoy, Magubane, Wallerstein & Martin) on the origins and purposes of reserves (see Cheap labour thesis and the function of the reserves). Argues reserves had varied origins and that the functionalist reading fits only some cases. Swaziland shows reserves sized for subsistence and preserving chiefly power, not designed solely to force labour out.
- Coercion vs. internal social structure in explaining migrant labour and peasant commodity production (Bundy's peasant-to-proletarian sequence) (see Agrarian transition to capitalism and the fate of the African peasantry). Holds that the peasant-to-proletarian sequence was not universal. Migrancy and commodity production followed varied sequences, and internal division of labour (e.g. BaKoena chiefs) was a key determinant before 1886.
Findings¶
-
Defines the 'regional formation' as a fluid combination of colonial and African social formations governed by capital. (p. 1)
A regional social formation or regional formation is defined here as a geographically and socially fluid combination of colonial and African social formations integrated and governed by the developmental dynamic of capital.
-
Stresses the need to specify the successive and simultaneous forms of capital, not treat capitalism generically. (p. 2)
Attention is hence directed to the need to specify the successive (and often simultaneous) forms in which capital initiated and consolidated its penetration of African society
-
The peasant-to-proletarian sequence was not universal: Swazi and Zulu households moved straight into labour supply late in the 19th century. (p. 3)
Swazi and Zulu households were wrenched directly from their pre-capitalist status into labour supply only in the closing years of the 19th century.
-
For the Pedi and the Lozi in Mozambique, migrant labour came before commodity production, the reverse of the common sequence. (p. 3)
In the case of the Pedi of the Transvaal and the Lozi in Mozambique, migrant labour actually preceded any notable development of commodity production
-
After 1900 a regional labour market emerged dominated by South African mining capital, with local employers reduced to staging posts. (p. 4)
Nevertheless, particularly after 1900, a regional labour market was forged, dominated by the demands of South African mining capital.
-
Centralised recruiting for the Rand was difficult and incomplete until about 1920. (p. 5)
centralized recruiting for the Rand mines was achieved only with a great deal of difficulty, and then not fully until as late as 1920.
-
Basotho grain exports surged in the 1870s on the back of Kimberley demand, driven by chiefly extraction of surplus labour. (p. 6)
Exports of grain rose from 50,000 muids in 1871 to between 200,000 and 400,000 muids by the end of the decade
-
Basotho migration in the 1870s was largely chief-organised, so the hut tax fell on a society already migrating. (p. 6)
In 1879, an estimated 15,000 Basotho males were engaged in migrant labour, the majority of whom were sent out by the chiefs.
-
Swazi reserves were sized to cover subsistence plus a 50% population increase, contrary to the reserve-as-forced-labour thesis. (p. 8)
A total of 621,907 ha (some 37% of the area of the country) was calculated as necessary to provide subsistence for the existing population and a 50% population increase.
-
Officials expected many Swazi to stay on expropriated land as tenants or wage workers, with taxation inducing labour to the Rand. (p. 9)
Legal provision was made for its repatriation to the reserves, but it was correctly assumed by state officials that many Swazi would remain on expropriated land
-
The colonial state kept customary tenure in reserves and left land allocation with chiefs, which bolstered the ruling lineage. (p. 10)
the colonial state preserved these forms within the reserves through legislation prohibiting private property and individual tenure, and vesting control of land distribution with the Swazi chiefs where it had hitherto been.
Evidence ledger¶
Sources used¶
| Source | Kind | Period | How used |
|---|---|---|---|
| Judy Kimble's unpublished work on Lesotho (MA thesis 1978 and workshop papers 1976, 1977), including grain export figures and 1879 migrant estimates | secondary literature compilation | 1830-c.1885 | Supplies the Basotho case on which the argument about internal social division of labour, grain exports and migrant labour before 1886 is built. |
| Swaziland National Archives, File DO9j2 (carrying capacity of the 32 Swazi reserves) | state archives | c.1907-1910s (land partition) | Source for the Table 3 data on area, population, stock and hectares per household in the 32 demarcated reserves, used to show how the 'native requirement' was calculated. |
| Swaziland colonial tax records (used in the 'native requirement' calculation) | official statistical series | c.1907 partition period | Total population was calculated from tax records as part of the procedure for determining land needed for reserves; the paper reports this as state procedure rather than consulting the records directly. |
| Christopher 1976, Southern Africa: An Historical Geography, p. 117 (land expropriation table) | secondary literature compilation | 1860-1911 | Table 1 data on area and percentage of land in European farms in 1860 and 1911, used to show the extent of land expropriation. |
| Palmer 1977, Land and Racial Domination in Rhodesia, p. 108 (man-land ratios in reserves) | secondary literature compilation | 1914 | Table 2 hectares per head of population in reserves, used to show wide discrepancies in the land Africans retained. |
| Board 1967, 'Southern Africa' in Africa in Transition, p. 332 (map base) | maps spatial data | circa 1914 | Fig. 1 map of labour reserves around 1914, adapted from Board. |
| Swaziland spatial data on royal elements and chieftaincies, 1907 (Figs 2-4; derived from Crush 1980b) | maps spatial data | 1907 | Maps showing reserves 'drawn' around royal households, cattle posts, graves and chiefly homesteads, supporting the claim that the pre-colonial spatial structure was frozen. |
| Unterhalter 1979 study of migration from the Nquthu district, 1879-1910 | secondary literature compilation | 1879-1914 | Provides the three phases of Nquthu migration, used to show temporal fluctuation in labour supply. |
| Ballard 1978 on migrant labour in Natal 1860-1879 | secondary literature compilation | 1860-1879 | Used to show that distance and labour supply were not simply related, since Natal labour routes from the Tsonga areas crossed Zululand. |
Inference¶
Quantification: Descriptive tables.
Historiographical synthesis and comparative case study. It contrasts regional variation across African social formations (Basotho, Swazi, Pondo, Zulu, Pedi) using secondary studies, with two short cases (Lesotho from Kimble's thesis; Swaziland reserves from archive data and the author's earlier work) used to argue against generalised accounts. Argument is theoretical and qualitative, with three descriptive tables.
Key claims¶
| Claim | Basis | Strength |
|---|---|---|
| Paths of African incorporation into the regional capitalist formation varied sharply in time and space; the cultivator-peasant-proletarian sequence was not universal. (p. 203) | Synthesis of secondary case studies (Guy, Bonner, Beinart, Kimble, Webster, Delius); no primary data presented. | Partial data |
| Basotho commodity production and migrant labour before 1886 were driven by the internal social division of labour and chiefly control of labour time rather than by coercion. (p. 205) | Kimble's unpublished thesis and papers, including grain export figures (50,000 muids in 1871 to 200,000-400,000 by 1879) and an estimate of 15,000 migrants in 1879. | Partial data |
| Reserve land varied widely in size per head across the colonies, and the creation of reserves predated their labour-reserve function. (p. 206) | Table 1 (Christopher 1976) and Table 2 (Palmer 1977) on land expropriation and man-land ratios; secondary sources on the origins of reserves. | Partial data |
| In Swaziland the colonial state deliberately allocated generous reserve land (621,907 ha, about 37% of the country) sized for population growth, so land expropriation per se was not the mechanism releasing labour. (p. 207) | Swaziland National Archives File DO9j2 and the described 'native requirement' procedure from tax records; Table 3 on carrying capacity. | Partial data |
| The Swaziland partition preserved the ruling Dlamini lineage's spatial structure while displacing commoners, reinforcing chiefly control over land distribution. (p. 209) | Spatial maps (Figs 2-4) of royal elements and chieftaincies in 1907 and the author's earlier work (Crush 1980b); commoner displacement figure given without source. | Illustrative or anecdotal |
| Labour supply varied over time and was shaped by competition between local and regional employers; a regional labour market dominated by the Rand mines emerged only after about 1900. (p. 204) | Secondary literature (Unterhalter, Van Onselen, Ballard, Etherington, Vail & White, Jeeves). | Illustrative or anecdotal |
The author's caveats¶
The author notes that the Lesotho argument relies on Kimble's unpublished work, and that the Swazi land partition background is treated elsewhere. He calls the carrying-capacity assessment of each reserve 'subjective' on the colonial state's part. No other data limitations are discussed.
Gaps in the evidence¶
The argument rests almost entirely on secondary literature and one unpublished thesis, so the Basotho claims cannot be checked against original mission, trader or colonial records. Swazi claims (55% of commoner households displaced, reserve boundaries around royal sites) need the underlying partition commission records, tax rolls and household or location data, which are in the Swaziland archives. Quantified labour-flow series by origin (e.g. recruiting or pass records) would test the claim of uneven migration, and the relevant records exist in archives today.
Replicability: Data poorly documented.