The papers in this special issue of the Journal were originally presented to a conference held at the School of Oriental and African Studies under the joint auspices of the School and the British Social Science Research Council, in September 1981. The conference, which grew out of an earlier series of seminars run by the S.O.A.S. History Department, arose out of the editors' concern that the history of the family, which had become so lively and important an area of study in Europe and America since the 1960s, was being almost totally neglected in Africa; through the seminars, the conference and now the publication of a selection of the papers in the Journal we hoped to stimulate research on a range of questions which could be fruitfully explored in the African context, and which could possibly also feed into the wider historiography on the family in Britain, America and Europe.

For many of its formative years, the study of history was concerned in the main with the history of the dominant, of ‘great men’ and their institutions, states and government, armies, churches and culture. The history of the dominated, of ‘ordinary people’ was thought to be lost, for the majority of our ancestors left no obvious written record from which their lives could be re-created. From 1929, when Marc Bloch and Lucien Febvre founded the journal Annales d'histoire économique et sociale, this view has been increasingly challenged, and it is fair to say that today the past of ordinary people is at least as significant historiographically as that of famous leaders or powerful institutions.

One of the problems with interdisciplinary work is that the outsider to a discipline so often tends to join its discourse at points which seem irrelevant to its practitioners. The experience is not unlike that of the foreign visitor who watches with amazement and not a little sadness as the ‘natives’ abandon those items of their culture which seem to him to be of most value. And the ‘native’ has only impatience for advice to slow down the pace of change.

African specialists on the history of the family seem to be entertaining a similar break with their past practices and present significant problems of orientation for the outside observer. The chief dissatisfaction appears to be with the legacy of the powerful generation of anthropologists who emerged in the 1930s and 1940s. They dealt with the organization of large corporate groups and interpreted action, belief and feeling in terms of a few principles derived from the structures of such groups.1 It has become an everyday criticism that this great work suffers from its ‘timelessness’, and of course historians have a professional interest in reiterating the point. More serious is the fact that the older constructs no longer seem to assist in analysing either the new problems that excite the historian aware of what his colleagues elsewhere are doing or the actual findings of new research.

To the outside observer – in this case an historian of Europe, whose bedtime reading consists of ethnologies of Africa, Papua-New Guinea, and the like – a confrontation with current research on Africa helps focus issues and problems in his own work and suggests a few points where common discussion might be fruitful.

Using the factors of family structure examined by Audrey Richards in a well-known essay, this article suggests that a more productive concept for the historical study of Central Africa than either the unique tribe or a group of societies identified by their rule of descent may be the lineage mode of production, in the restricted sense developed by P. P. Rey. Analysis of the organization of political, economic and ritual functions among the BaKongo, BaSuku, BaPende and other Zairean peoples shows the complementarity and flexibility of patrilateral and materilateral relationships. It is suggested that the greater ‘quantity’ of social structure exhibited by coastal peoples, as well as their matrilineal development, may result from the prolonged effects of the great Congo trade, especially the trade in slaves, modifying an old and generally bilateral system organized by networks of permanent matrimonial alliance. This system is characteristic of the Congo basin, Zimbabwe and Angola.

The article considers the changing pattern of kinship relations amongst the Kongo south of the Zaïre river and west of the plateau in the region once dominated by the nuclear Kongo kingdom. It argues that the normative pattern of kinship and family relationship was probably established in the early years of agricultural settlement by the ideology of the kanda, the exogamous matrilineal descent groups which controlled access to land. This ideology dominated family and kinship relationships as long as agricultural production was the dominant economic factor. In the fifteenth and early sixteenth centuries the evolution of the trade-based kingdom of Kongo modified the normative pattern established by the kanda, and in the late sixteenth century the acquisition of large numbers of slaves and the use of Christianity as a legitimating ideology effected more profound change. In particular, the elite developed a system of patrilineal descent categories which were used to control trade-based wealth and to organize political relationships. Freedom came to be related more to patrilineal descent category membership and less to kanda membership, whilst the economic and political position of all but the most eminent women deteriorated. When, in the late seventeenth century, changing patterns of trade caused the kingdom of Kongo to disintegrate, the Mwissikongo of the centre adopted a cognatic mode of descent which enabled them to control both agricultural and trade-based wealth. Certain eminent women seized the political opportunities afforded by the crumbling of male-dominated centres of power whilst the definition of slave and free became increasingly problematic. In the north-western province of Sonyo, increased trade-based wealth enabled the dominant patrilineal category to establish itself as a corporate group and to monopolize all positions of power. In the eighteenth century power disintegrated throughout the region and land again became the primary economic asset. The ruling elite sought legitimation in terms of the ideology and descent system of the kanda. The former slave groups sought to establish rights in the same way. In the nineteenth and twentieth centuries all groups legitimized their holding of land, primarily in terms of the ideology of the kanda and secondarily in terms of the concept of Mbanza Kongo.

Records of Ethiopian property dealings provide insights into the affairs of Ethiopian families. They suggest further, dynamic links between property and family. Amhara rules of descent and inheritance are ambilineal. Each person reckons their ancestry equally through their father and mother and inherits property the same way. Yet, without abandoning their profession of the principles of ambilineal descent and of equal partible inheritance, the Abyssinian nobility subverted these norms, thereby creating ‘families’ out of a welter of ramifying lineages, and ‘estates’ out of disintegrating holdings. Their devices included wills and marriage endowments which privileged one sibling at the expense of others. Another device was the alaqenat, an office which functioned as the head of a family and its affairs. The alaqenat, previously unreported in the Ethiopian literature, appears in a number of documents in the later eighteenth and earlier nineteenth centuries, and the article uses it to trace the affairs of one noble family over a period of six generations. Further comments are made about the links between public office and notions of property. Finally, two wider spheres are addressed. In spite of its radical ambilineality Abyssinian society reveals tendencies common to other societies based on plough agriculture, tendencies towards greater class differentiation based on the accumulation of landed property. And, unlike most other historic societies professing Christianity, Abyssinian society is marked by frequent divorce and marital instability.

This article examines the possible relevance to Africa of the recent contribution to the history of industrialization in western Europe made by a group of historians of the family using statistical demographic techniques, especially ‘family reconstitntion’, developed by Peter Laslett at Cambridge. It is suggested that in particular the work of David Levine on eighteenth-century England yields a hypothesis about a link between ‘proto-industrialization’ and increased population growth which may be applicable, in a suitably modified form, to black Africa.

The article focuses particularly on the mechanism of widened income opportunities for young people lowering the age of marriage. Levine concentrates especially on the lowering of the marriage age of women causing enhanced fertility per wife. The modified form of the hypothesis suggested for black Africa would focus more on the reduction of the age of marriage for men and on the transition from polygyny to monogamy, the latter especially producing increased fertility per individual wife.

Early twentieth-century missionaries in South Africa invariably subscribed to an ideology of domesticity for African Christian women. Being a housewife and mother was seen as a full-time spiritual vocation. In Johannesburg, for example, women's domestic role within marriage gave point to the whole range of female mission activity. Housewifery was not only taught at special mission schools, but also in women's hostels, a girls' youth movement and a delinquents' institution. These agencies and the African women's prayer unions also sought to combat premarital pregnancy in adolescent girls, because Christian marriage was regarded as the only proper context for motherhood. Maternity was central as well to female medical mission efforts. But the ideology of domesticity was in practice fraught with contradictions in the South African setting. First, housework training inevitably had an ambiguous status since it often aimed at supplying domestic servants to white households. Secondly, full-time motherhood was impossible for urban African women compelled to supplement their husbands' low earnings. White liberals therefore favoured higher African wages partly to free black women financially to fulfil their ‘natural’ child-rearing role. African women themselves have campaigned for domesticity against the authorities' policies of labour control. Bloemfontein in 1913 and Crossroads squatter camp in the late 1970s provide striking examples of such struggles. The Western feminist may see both the family and the role of housewife as oppressive. For the black South African woman, by contrast, the ideology of domesticity infusing these institutions may represent something positive denied her by the state.

In its ideal form, the Chewa family system is supposed to place considerable emphasis on the mother right, the avunculate, uxorilocal marriage, husband's subordination to wife's kin, and importance of female children as future reproducers of the lineage. In practice, however, there are a number of factors which have tended to mitigate the impact of these tendencies in the system, as far as men or husbands are concerned. Attention has been drawn to the fact that a number of important changes in the marriage contract, family residential patterns, exercise of domestic authority, and control or custody of children, have been the result of historical developments. There is therefore a need to understand these developments, even if the changes in the Chewa family system to which they led have not brought about a complete transformation of the system.

Several of these historical developments have been reviewed, as well as the features of patriliny to which they have given rise in several parts of central Malawi. They include the growth of the slave trade in the nineteenth century and the opportunity which this gave to some lineages to acquire women who could be married virilocally; the intrusion of patrilineal peoples like the Ngoni which increased the incidence of virilocal marriage and patrilineal descent; the spread of certain Christian missionary teachings that are in conflict with matrilineal principles of family organisation; and such developments of the colonial period as labour migrancy and cash-cropping which often impinged on relations between husband and wife and family life in general.

This paper explores some of the major methodological problems associated with the study of the history of the family in Africa. It sets out to explore the problem of the unit of analysis, concluding that the historian must be careful to distinguish between idealized family forms and the reality of family structures. Using both historical and contemporary examples from southern Malawi the paper explores this problem further by analysing the role of the matrilineage vis-à-vis the household over time.

Both oral and written sources specifically concerned with the history of the family tend to emphasize the formal structure of kinship relations and it is difficult to know how these relate to the facts of social and economic organization. Even using present-day evidence it is difficult to integrate cultural perceptions of kinship and family relations with realities – in particular with the economic realities, which may change much faster than cultural norms. In the final section of the paper it is suggested that the nearest we can get to a knowledge of the history of the family, avoiding the problems of ideology and the drawbacks of structural and evolutionary models, is to approach the subject ‘sideways’. By studying other institutions and relationships which impinge on family structures, we may get closer to defining the boundaries of these structures. This approach is illustrated using the example of chinjira - a non-kin-based relationship between women which exists in parts of southern Malawi. A study of chinjira indirectly demonstrates both the strength and the limits of kinship relations.