Before turning to Macmillan, let us consider Cobbing's claim that what Walker and later professional historians said about the Mfecane was integral to their view of South African history. Like Theal, Walker showed himself interested enough in the history of Africans to include a brief passage on 'the Mfecane' in a general history of the country – more Eurocentric historians omitted the topic – but as with other historians of his time, and many of his successors, he had little interest in precolonial history, where the sources were so problematic and it was so difficult even to provide a chronology. For Walker and other early liberal historians, the dominant theme of South African history was the advancing frontier and the white racism that emerged there and dominated the political life of the country after the frontier closed. The Mfecane received relatively little attention in their work, and was hardly central to it.
This is also true for later liberal historians. The first chapter of the most detailed recent history of the country, the successor to Walker's History, begins with a chapter which takes us 'From the Dawn of History to the Time of Troubles', but the balanced interpretation of the Mfecane which T.R.H. Davenport offers there is relatively brief and hardly integral to his vision of South African history as a whole. The same applies to Leonard Thompson's interpretation – very different from Walker's, as we shall see – in his shorter History.25 In all these general histories, the Mfecane receives relatively minor attention, and in the collection of essays on 'The Shaping of South African Society' edited by two leading specialists on early South African history, no significant reference is made to the Mfecane at all.26
Macmillan and De Kiewiet
Though Macmillan in his classic Bantu, Boer, and Briton: The Making of the South African Native Problem (1929) followed Theal in writing of 'the Chaka wars' and their effect 'even on tribes far away in the interior', he went on to anticipate, in outline, central features of Cobbing's argument in 'The Mfecane as Alibi'. For Macmillan the 'great upheaval among the Bantu' in the Natal/Zulu kingdom area and the interior was not unrelated to the process of colonisation. In fact, he linked it specifically to the slave trade, while at the same time making clear that the lack of source material posed a serious problem to any analysis of what had happened. Macmillan wrote as follows:
How far this great upheaval among the Bantu must be attributed, in Bishop Stubbs' words, to the 'generally unsettled state of all tribes bordering' on European conquests,27 can never be fully known. While from the nature of the case the effects of the frontier wars on the remoter tribes are not directly evident, the suggestion that there was a connexion is not wholly to be dismissed. It is significant that the rise of Chaka came at the very moment when things were moving towards a climax on the Cape frontier . . .
Further, to meet the demands of European planters, slave-traders had not only raided on their own account for a hundred years past, but set tribe against tribe in such ruthless fashion that if the consequences were often bloody it is not for Europeans to cast a stone. There is no reason to believe that the slave trade left the southern part of Africa unaffected.28
Macmillan then cited various pieces of evidence in support of the idea that slave-raiding was important: that Lord Charles Somerset had in 1823 considered annexing Delagoa Bay as a check on the slave trade; that John Philip of the London Missionary Society had reported that the Tswana-speaking people in the interior feared the ravages of the slave trade; and that Mzilikazi was said to have met slave- raiders from Portuguese ports before he headed west, and that explains why he did not go north. Macmillan went on to quote Philip as saying both that Mzilikazi's people had to 'maintain an incessant struggle against the Portuguese slave-traders', and that 'To Farewell's establishment at Port Natal we are to trace the devastations of Chaka.'29 Writing in the Cambridge History of the British Empire a few years later, Macmillan cited references in Theal's Records of South East Africa to Delagoa Bay being used by slavers, before adding: 'the effects of the slave trade upon the natives in what is now called Zululand have never been considered. Nor is it possible to gauge the repercussions of the check administered on the Fish River to the coast tribes lying to the west of Chaka's sphere of action.'30
Macmillan's brilliant student C.W. de Kiewiet also rejected Theal's version of the Mfecane. Under Macmillan's guidance, De Kiewiet had written a thesis at the University of the Witwatersrand – no copy survives – on the Cape northern frontier in the period just after the Mfecane. When he came to write about the upheaval itself in the 1930s, first in a chapter in the Cambridge History of the British Empire and then in his History of South Africa Social and Economic, De Kiewiet, though aware that 'the causes of these events can never be adequately investigated', advanced an ecological interpretation. Conflict had occurred because of 'an intense competition amongst the tribes for sowing and grazing land'. There was 'much reason', he asserted, to think that what had been written about 'the devastation of the Zulu, Matabele and Mantati "hordes" was very greatly exaggerated'. What had happened in the 'confusion of the eighteen-twenties' was 'much displacement', after which people 'poured back into their lands'. And in his History he added: 'Amongst the causes of this singular crisis that smashed tribes, scattered others, and dashed the fragments into new combinations, the halting of the Bantu vanguard on the Eastern Frontier probably had much influence.'31
For John Philip in the 1830s, then, as for Macmillan and De Kiewiet a century later, the upheavals in Natal and the interior were by no means divorced from the process of colonisation. Only in a paragraph concerning the Ndebele in 'The Mfecane as Alibi' does Cobbing, in passing, acknowledge that Macmillan, 'slightly deviant here as in some other respects', linked the ejection of Mzilikazi to the slave trade and that 'this hypothesis is better than anything we have'.32 Elsewhere in his article Cobbing suggests a line of interpretation running from Theal through Macmillan to Omer-Cooper and beyond. As these historians in fact advanced vastly different interpretations, we must ask why the suggestive insights which Macmillan and De Kiewiet made in the 1920s and 1930s were not followed up. The loss of the Philip papers in the fire at Wits in 1931, and Macmillan's departure from the country – De Kiewiet had already left – is part of the explanation. Macmillan and, to a lesser extent, De Kiewiet were regarded as radicals, and their work was marginalised within the country because it was viewed in ruling circles as politically dangerous. But there are also other reasons. The very few leading South African professional historians of the 1940s and 1950s believed precolonial history was appropriately the terrain of anthropology, and devoted their attention to more traditional themes for historians, such as the imperial connection and constitutional issues.33 Such graduate students as there were in those decades were directed to similar topics. No one did significant work on the Mfecane.
What of Cobbing's claim that the idea of 'cataclysmic black-on-black destruction' was an 'alibi'? He speaks of historians attributing 'the land distribution of 1913 . . . to a black-on-black holocaust in the period 1815–35', and continues: 'Since the Second World war, the stress of the alibi has been on the natural "pluralism" of black societies and how they self-sequestered themselves into proto-Bantustans in the era of Shaka, leaving the whites merely the task of surveying and recognition.' The Mfecane was 'a characteristic product of liberal history used by the apartheid state to legitimate South Africa's racially unequal land division'.34 This is a serious charge against the liberal historians.
Macmillan was in fact a leading critic of the unequal distribution of land, and of racial segregation in general. His first public statement on African affairs, made to a Dutch Reformed Church conference in 1923, started with a critique of the 1913 Land Act and its effects.35 Macmillan was also firm in his rejection of any idea of a separate African history. De Kiewiet did not return to South Africa in large part because