Skip to content

The Rise and Fall of the South African Peasantry: A Critique and Reassessment

Literature › Works

Jack Lewis (1984). The Rise and Fall of the South African Peasantry: A Critique and Reassessment. Journal of Southern African Studies, Vol. 11, No. 1. Taylor & Francis.
Type journal article
Scope Wider apartheid literature
Method mixed
Period studied 1848–1882
Regions Eastern Cape, Ciskei, Transkei, Thembuland
Themes agriculture land, origins segregation, mining migrant labour, inequality poverty unemployment, comparative theory, methods data sources
Citation key lewis1984

Argument

Lewis criticises Colin Bundy's thesis that an Eastern Cape African peasantry prospered c.1840-1870 by responding to the market and was then destroyed by mining capital. He argues that Bundy lacks a theory of the pre-capitalist mode of production and wrongly treats Xhosa society as egalitarian. The productive unit was the stratified homestead (umzi), not the household. Census and market data show sharply skewed holdings of cattle, ploughs and land, and for most households market sales meant selling part of their subsistence. The brief 1865-75 expansion owed more to political and demographic factors (the Mfengu move to the Transkei, which left land for those who stayed) than to a positive response to the market. Impoverishment was a long-run trend traceable almost from conquest, and differentiation inside the reserves helps explain the origins of migrant labour and the persistence of the reserves.

Evidence and method

Cape Archives colonial office and Kaffrarian records: the 1848 Xhosa census, the 1854-55 Mfengu censuses, the 1857-58 native market returns, the 1861 Ox-kraal/Kamastone census, the 1869 Healdtown census, the 1864 location returns, the 1875 Cape census, and the 1882 Thembuland Commission. It also uses magistrates', missionaries' and travellers' accounts, the Keiskamma Hoek Rural Survey, and Bundy's own cited figures. Based on Lewis's UCT PhD research, with statistical help from Prof. Juritz.

Place in the debates

Findings

  • The 1848 Ngqika census shows large, hierarchically organised homesteads rather than egalitarian household units. (p. 6)

    In 1848 there were, in the important Ngqika division of the Xhosa under the Chief Sandille, 1 090 homesteads, comprising 5 765 households and 20 928 people.

  • Cattle ownership in 1848 was highly unequal, so poor households depended on the herds of richer ones. (p. 7)

    While 80 per cent of the households had less than 5 cattle (of which 31.5 per cent had none), 20 per cent were in possession of large herds numbering up to 160 head.

  • Before the cattle-killing there was little sign of the technical innovation that Bundy's thesis assumes, even among the Mfengu. (p. 9)

    But in 1854, of 1 592 households in Peddie, only 39 owned a plough, 42 owned wagons and there was a total of 201 sheep.

  • The 1857-58 grain sales were a mass of small sales by many households, yielding roughly enough to pay the hut tax. Lewis reads them as survival marketing during the cattle-killing famine, not as prosperity. (p. 13)

    If we assume that every Mfengu household participated to some extent in the production of grain for the market this would mean that in the period July 1857 - March 1858 each household marketed 79 lbs of grain.

  • Bundy's own Healdtown figures imply that sales cut into subsistence, so the marketed 'surplus' was not a true surplus. (p. 14)

    The Keiskamma Hoek Rural Survey suggested that adequate nutrition over a year required the consumption of 4 000 lbs. of grain by a family of six.

  • In 1861 cultivated land was concentrated in a minority of households. (p. 15)

    the top 25 per cent with eight or more acres each, accounted for 1 638 acres or 56.9 per cent of the total

  • The skewed distribution of cattle limited plough ownership, and therefore arable expansion, to a minority. (p. 19)

    In real terms this meant that the majority of households could not muster enough cattle to put together a team of oxen to draw a plough.

  • The 1865-75 boom resulted from political and demographic circumstances rather than from a response to the market as such. (p. 21)

    this prosperous period which can be exactly located from 1865 to 1875 was the outcome of a particular combination of political and demographic factors

  • Migration to the Transkei left land for those who stayed, so output rose through extensive expansion by a smaller population, not through higher productivity. The gains were lost when population density rose again and drought struck. (p. 23)

    based not so much on a general increase in productivity - the more intensive exploitation of limited land resources - but on the expansion of agricultural activity over a greater area by a reduced population

  • Bundy's Thembuland Commission illustration is unrepresentative: the three well-off cases cited cover only a small share of households. (p. 24)

    The three cases quoted in Rise and Fall are in fact representative of only 10 or 15 per cent of the households.

  • Impoverishment of the majority was long-term, dating from conquest, rather than a late consequence of mining capital. (p. 25)

    impoverishment in real terms, a declining level of domestic consumption for the majority of households, has been a very long term phenomenon traceable almost from the moment that effective conquest was achieved.

Evidence ledger

Sources used

Source Kind Period How used
1848 Census of the Xhosa (British Kaffraria), complete census roll, C.O. 6155, Cape Archives population census 1848 Quantifies polarisation of Xhosa homesteads and households by size, polygamy and cattle holdings to show pre-colonial stratification.
Census of Fingoes Residing in the District of Peddie, 31st July 1854 (G.H. 8/24) state archives 1854 Shows that few Peddie Mfengu households owned ploughs, wagons or sheep before the cattle-killing.
Calderwood's census: Statistical Returns of Fingoe Locations in Victoria, Fort Beaufort and Queenstown, July 1855 (G.H. 8/27) state archives 1855 Supplies household, plough, wagon and sheep counts and 2.5 acres per household used to estimate Mfengu grain marketed per household and per acre.
Weekly market returns by Verity, Line Drift, Ebb and Flow Drift and Fort Montgomery Williams markets, C.O. 2948/49/50 and C.O. 2951 state archives July 1857-March 1858 Daily grain sales, prices and wagon use are used to estimate volume and value of Mfengu grain sales during the cattle-killing and per-household marketing.
Brownlee and Maclean correspondence (G.H. 8/27, B.K. 90) on grain prices, ploughs and mission stations state archives 1854-1855 Gives the 6 shillings per muid ration price and Brownlee's observations that ploughs and wagons were bought mainly on mission stations.
King William's Town Gazette and Border Intelligence, 25th April 1857 newspapers periodicals 1857 Source for the April 1857 grain price of £1 12s per muid at King William's Town.
Cape of Good Hope Blue Book 1860, Money, Accounts, Weights, Measures official statistical series 1860 Defines the muid (180 lbs) for converting grain sales.
Magistrates' and missionaries' correspondence in B.K. series (Tainton, Campbell, Lucas, Chalmers to Maclean), 1856-58 state archives 1856-1858 Anecdotal evidence that Mfengu sold grain they needed and became destitute, and of grain sold up the Kei.
Shepstone's census of Ox-kraal and Kamastone Fingo locations, C.O. 3026, 9th April 1861 population census 1861 Source for the distribution of cultivated acres, ploughs, sheep, cattle, wives and children across 512 households (Table 1) and the polygamy-cultivation correlation.
Healdtown Mfengu census, C.O. 3163, Verity to Colonial Secretary, 16th February 1869 population census 1869 Source for the cattle and sheep distribution (Table 2), plough ownership and polygamy among 513 households.
British Kaffrarian returns of Xhosa locations, B.K. 109, Government Notice No. 19 of 1864 official statistical series 1864-1865 Gives population, households, cattle, sheep, ploughs and wagons for Xhosa locations in 1864, showing 18 households per plough.
Brownlee to Secretary to Lt. Governor, 17th May 1865 (B.K. 73) state archives 1865 Contemporary testimony that stock was unequally distributed and most people had only grain to pay taxes.
Census of Fingoe districts, N.A. 152, Blyth to Southey, 2nd May 1874 population census 1874 Source for the nearly 44 000 Mfengu in Fingoland after the 1865 migration from the Ciskei.
Cumming and Verity correspondence, C.O. 3140 and C.O. 3163, 1868-69 state archives 1868-1869 Supports the claim that established, more prosperous households migrated to the Transkei and leased out their lots to younger households.
General Census of the Cape of Good Hope 1875 (G42-1875) and enumeration papers Tam 7/1, Tam 9/2, Tam 9/3 population census 1875-1876 Provides harvest, household, population, sheep and plough figures to assess economic performance and grain adequacy at the height of the boom.
Returns of Inspectors of Native Locations, A100-1880 official statistical series 1879 Source for 1879 plough numbers (1 209) and sheep decline, set against 1865 and 1875 figures.
Report and Proceedings of the Thembuland Commission of 1882 (G.H. 66-83) commission or official report 1882 Author re-tabulates cattle, sheep and plough ownership for about 700 households to show stratification Bundy overlooked, and uses it for Gacelo's land sales.
Houghton and Walton, The Economy of a Native Reserve, Keiskamma Hoek Rural Survey Vol. II (1952) household or labour survey microdata 1950s Benchmark of about 4 000 lbs of grain per family per year to judge adequacy of grain retained after sales.
Henderson survey of 16 families in Victoria East, 1929 (reported in D. Wilson, South African Outlook, 1931) author survey or fieldwork 1929 Corroborates the grain consumption estimate for a family.
Contemporary observers' accounts of Xhosa society (Alberti 1807, Shaw 1860, Morgan 1833, Dohne 1837, Dundas 1827, Collins 1809, Smith Kaffir Notes) secondary literature compilation 1807-1860 Qualitative evidence on homestead size, monogamy, wealth concentration and chiefly redistribution.
Bundy, Rise and Fall of the South African Peasantry (1979), figures cited as data (e.g. Healdtown 30 000 bushels; Thembuland Commission cases) secondary literature compilation 1858-1882 Bundy's own figures and quotations are re-analysed to show the household stratification and subsistence shortfall he does not explore.

New data

Author compiled and statistically tabulated household-level data from colonial censuses of Ciskei Xhosa and Mfengu locations (1848 Ngqika census of 5 765 households; 1861 Ox-kraal/Kamastone, 512 households; 1869 Healdtown, 513 households; Thembuland Commission households) and market returns of 1857-58, with statistical help from the UCT Mathematical Statistics department.

Inference

Quantification: Descriptive tables.

Critical reassessment of Bundy's thesis by re-analysing archival household censuses and market returns: frequency distributions and cross-tabulations of cultivated acres, ploughs, cattle and wives, back-of-envelope calculations of marketed grain per household against subsistence needs and hut tax, plus theoretical argument about the pre-capitalist Xhosa mode of production.

Key claims

Claim Basis Strength
Pre-colonial Xhosa society was highly polarised, with a minority of large homesteads commanding much of the population and cattle, rather than egalitarian. (p. 5-6) 1848 Xhosa census (C.O. 6155) tabulated for the Ngqika division, plus missionary and traveller accounts. Systematic data
Before the cattle-killing there was little evidence of improved agricultural techniques or productivity among the Mfengu and Xhosa. (p. 8) 1854 Peddie census and 1855 Calderwood returns on ploughs, wagons and sheep, plus Brownlee's observations. Partial data
Grain sales during 1857-58 were made by a mass of small producers each marketing roughly 50-100 lbs, mostly to pay hut tax, and some sold grain needed for subsistence. (p. 12) Verity's market returns, household and acreage counts from the 1855 census, hut tax rates, and magistrates' reports (Peelton, Chalmers). Partial data
Marketed 'surplus' in Bundy's data (e.g. Healdtown 1858) was inadequate once subsistence needs are considered, so market participation could mean declining consumption. (p. 14) Bundy's cited Healdtown figures, household counts from Calderwood and the 1861 return, and Keiskamma Hoek Rural Survey nutritional benchmark. Partial data
Cultivation, plough ownership and stock were highly unequally distributed among Mfengu households in 1861-69, tied to household labour, polygamy and cattle for oxen spans. (p. 14) Ox-kraal/Kamastone 1861 census (Table 1) and Healdtown 1869 census (Table 2), cross-tabulated. Systematic data
The 1865-75 prosperity resulted from the Mfengu migration to the Transkei and demographic and political circumstances, not a positive market response, and expansion was uneven and short-lived. (p. 20-23) Blyth's 1874 census, 1865/1875/1879 location returns and the 1875 census, and magistrates' correspondence on land left behind. Partial data
Even at the height of the boom, households in the Tamacha locations had little grain margin and were vulnerable to drought. (p. 22) 1875 Cape census harvest, household and population figures and 1875-76 Tainton correspondence. Partial data
Bundy's own Thembuland Commission evidence shows the majority of households were not clearly better off than their grandparents. (p. 23) Author's re-analysis of the 1882 Thembuland Commission report cattle, sheep and plough data against Bundy's three quoted cases. Partial data

The author's caveats

Acknowledges that household-based censuses may obscure pooling of resources between fathers and married sons, that household age structure may affect labour estimates, that the market statistics capture only official market sales (not wagon sales up the Kei or to King William's Town), and that the per-household figure assumes every Mfengu household participated.

Gaps in the evidence

The argument rests on a few snapshot censuses from selected locations, so it cannot show trends in stratification over time or establish whether households were better or worse off than before. Direct household income, consumption, grain yield and sales data over multiple seasons would be needed; these mostly never existed, though more magistrates' and mission records may survive in the Cape Archives. The Xhosa census analysis is promised for a separate paper and the Ph.D. thesis, so key tabulations are not shown here.

Replicability: Data described but not public.