Summary It became clear at the conference that anthropologists and historians in Southern Africa have a shared need to examine their past, their problems and their priorities. They need to tackle the writing of an intellectual history which neither foreshortens nor lumps discrete scholars together under a convenient heading - 'structural functionalism', Ronnie Frankenberg remarked, 'exists only as something to distinguish yourself from'; they need in humility to understand how their predecessors were manipulated both by the state and by their African informants and how they were blinkered by the limitations of current 'scientific' paradigms; they need to trace both the continuities of influence and polemic within an intellectual tradition and the discontinuities triggered by socio-economic and political change; in short, they need to replace the crudities that have hitherto passed for the history of anthropology or for historiographical analysis with something altogether more complex, more founded and pertinent to their own situation. Some participants responded to the challenge by anticipating the emergence of a single discipline. John Sharp foresaw the demise of both social anthropology and volkekunde as autonomous disciplines in South African universities, to be incorporated into sociology and African history. John Comaroff advocated a 'truly dialectical historical anthropology'. Archie Mafeje wrote: 'the more anthropological history is, the better'. But others asserted that there would continue to be clear distinctions of method. A historian might produce, said William Beinart, a local extended case study by listening to the actors as they spoke through the archival record, but the skills he had to employ in doing so were highly specialised and very time-consuming; it was almost impossible to imagine a single scholar combining these with the equally specialised and time-consuming skills of participant observation. Nor is it accurate to identify oral history, with all its fieldwork components, with classical anthropological field techniques. Two different approaches to the 'signals from within the society' are involved. If Beinart is right the need is not for a fused discipline but for continual discussion of shared problems; for fuller comprehension of each other's methods, theories and findings; for a full awareness of recent developments in each discipline so that radical historians, when they turn to anthropological insights, are aware of radical anthropology and not only of the 'classic' tradition; and for mutual criticism so that neither anthropologists nor historians are allowed to get away with a so-called 'radical' scholarship which is as abstracted and manipulative as anything in the past. The conference represented a beginning of all this. There were notable omissions, such as a more balanced evaluation of successive historiographical enterprises, and any proper appreciation of the very strong Rhodes- Livingstone tradition of urban anthropology. But there emerged a sense of constructive advance on several fronts. For example, the explicit contextualisation of particular anthropologists' work allowed both historians and anthropologists to achieve a better understanding of significant 'moments' in the development of both disciplines and also of their continuing links with the less fashionable intellectual crevices of their own past. The 'unit of study' was to be understood in terms of the sorts of questions posed rather than in terms of discovering alternative bounded entities, and the conference opened up a number of possibilities for further enquiry into the dynamic interaction between macro- and micro-level social processes. The phenomenon of ethnicity had to be taken seriously, neither as a self-evident 'given', nor merely as an idiom by reference to which individuals construct and manipulate their identity in different situations, nor as a variety of 'false consciousness'; but specifically as a cultural and political product which is part of the material conditions of existence but also has a 'life of its own' within a changing political economy. In these ways, and in other ways, the conference represented the beginnings of a dialogue that is immensely worthwhile. We hope that anthropologists will be sufficiently stimulated by the present issue of JSAS vigorously to continue this dialogue in the pages of the Journal.