Histories and critiques of anthropology usually deal mainly with what anthropologists say, including what they say they do. However, like their informants, anthropologists are accustomed to saying one thing and doing another. The emphasis in this paper is on the nature of anthropological data and methods, and thus on what anthropologists really do. The argument is that what anthropologists do is often productive and sensible, but that it has very little relationship to what many anthropologists say they should be doing. Indeed those anthropologists who take such theoretical directives seriously labour under a considerable, self-inflicted handicap.

Thinking about the future, we are obliged to rethink conventional accounts of our history. Alternative histories challenge received ideas about the development of British social anthropology. Some have put in question the precise nature of the Malinowskian revolution. Others reopen debates about the relationship between anthropology and colonialism (or, more broadly, about the relevance of social anthropology to policy questions). And the oral tradition insistently reminds us that private histories may undermine what we think we know about the great figures of the past. As we consider the prospects for European social anthropology, other questions impose themselves. One has to do with the institutional rivalries between social anthropology and sociology, and between social anthropology conceived of as a social science and the cultural anthropology dominant in the United States.

From its late eighteenth century beginnings until the early twentieth century, the House of Rothschild was the greatest of the European banks. Its success was due not only to the financial skills of the partners but also to their innovative strategies in the sphere of kinship and marriage. The structure of the bank was dependent on the family's unique rules of succession and marriage. It was organised into five branches, each in a different country, originally directed by one of the five sons of the founder, Mayer Amschel. The branches were linked by ties of fraternity, reinforced by repeated endogamous marriages. These marriages were normally contracted between the children of partners. Given the rule, established by Mayer Amschel, that only the founder's sons and, later, their sons and sons sons could become partners, the most common pattern was for a man to marry his father's brother's daughter.

This interview with Professor Adam Kuper of Brunel University took place in London in July 2000. It is divided into three sections. In the first, Professor Kuper outlines the background to and summarises the main arguments of his most recent book, Culture. The Anthropologists’ Account (Kuper 1999b), relating these to the issues of postmodernism, biological determinism and the ‘intellectually impoverished heritage’ of social anthropology in Britain. He then moves on to discuss in turn the relationship between theory and ethnography, ethical issues in fieldwork and the nature of anthropology as a cross-cultural, comparative discipline. Finally, Kuper considers the role played by EASA in the development of a sense of ‘common project’ among European anthropologists, addressing in particular the history and place of British social anthropology within this enterprise, as well as the question of language.

The rise of the Zulu power in the early nineteenth century has conventionally been treated as the outstanding example of a contemporary southern African process of ‘state-formation’, which was associated with revolutionary social changes. This paper advances an alternative view, that there were strong continuities with established forms of chieftaincy in the region, and in particular that the Zulu political system was based on a traditional, pan-Nguni homestead form of organization.

The Zulu homestead was divided into right and left sections, each with its own identity and destiny. This opposition was mapped into the layout of ordinary homesteads and royal settlements. It was carried through into the organization of regiments. The homestead and its segments provided both the geographical and the structural nodes of the society. The developmental cycle of the homestead ideally followed a set pattern, creating a fresh alignment of units in each generation. The points of segmentation were provided by the ‘houses’, constituted for each major wife and her designated heir. Each of these houses represented the impact, within the homestead, of relationships sealed by marriage with outside groups, whose leaders threw their weight behind particular factions in the political processes within the family.

Some years ago one of the very few structuralist anthropologists to contribute to Southern African studies complained that ‘Aucune étude n'a encore été consacrée à l'organisation spatiale et aux relations qu'elle implique chez les Bantu du Sud-Est’ (Roumeguère- Eberhardt 1963: 77). This was not strictly accurate. For example, Holleman (1940) had provided an important analysis of the Zulu homestead, borrowing something of the methods and concerns of the Leiden school's studies in Indonesia; and one cannot ignore Walton's (1956) wide-ranging if unsystematic treatment of the Southern African village. Yet the topic has been largely neglected despite the highly-ordered nature of the Southern Bantu homestead and the legal and symbolic significance of the traditional homestead plan.

The debate on the ‘Bantu expansion’ has been notable for the relative absence of contributions by social and cultural anthropologists. Murdock's rather extravagant contribution (1959) has assumed a quite inordinate significance, as the only direct attempt to discuss the sociology of large linguistically defined areas. This anthropological silence is particularly noteworthy in view of the close collaboration between anthropologists, linguists and archaeologists in the Americas. It is fair to say that nothing in the Africanist literature approaches the sophistication of, say, Dyen and Aberle's Lexical reconstruction: the case of the proto-Athapaskan kinship system (1974). Studies like that are the outgrowths of a living tradition stemming from Boas and Kroeber and Sapir. There is nothing comparable in Africa—H. Baumann, the only candidate, lacked sociological interests and did not systematically relate his findings and hypotheses to those of the linguists. Yet it is surely obvious that an anthropological contribution is needed here, if only in a critical role. The present paper outlines a few of the ways in which social and cultural anthropologists might contribute; and begins with a critique of traditional ‘culture area’ approaches, since these are generally thought of as the kind of thing that anthropologists can do, however poorly.