This study examines the role of herding, especially cattle, in the growth of complex societies in southern Africa from the 10th century AD. Interpretations of the significance of cattle in the economies of Mapungubwe, Toutswe, and Great Zimbabwe as regional centers in the prehistoric economies of the southern African region have often concluded that cattle were an important resource in the emergence of ranking or hierarchical structures within these farming societies. In addition, since cattle have long been associated with male power, it is implied in the literature that women were subordinated both to male authority and to their role as reproductive labor, valued (or devalued) within the formal and informal exchange systems of resources, including cattle. This idea of the subordination of women rests on the interpretation of women’s contribution to the economy as based primarily on their treatment as social capital. In other words, women were significant as daughters and mothers who could reproduce and bear male children as heirs to wealth and authority.
Interest in the sociocultural organization of southern African societies within archaeology was generated by a wide body of literature on the form and structure of settlements in 19th century AD and contemporary southern African societies (Livingstone 1858; Schapera 1935; Kuper 1980). Early missionary writers observed the layout of settlements and the significance of these settlement patterns within society. They noted that not only did settlements serve as places to live, they were in fact spatial expressions of networks of social, economic, and political relations within these societies. In the Sotho-Tswana communities where missionaries like Moffat and Livingstone worked, these early recorders noted that the layout of a village and its associated settlement places was very much influenced by the king or chief and his royal court. Similar observations were made by writers working within Shona-speaking and Nguni societies living nearby (see Walton 1956). This cross-cultural consistency in the use of space and architecture, when considered along with the linguistic and other cultural indicators of related origins, enabled historians, cultural anthropologists, and archaeologists to make comparative studies of settlement systems and their meaning and to draw from these studies models of social and spatial organization (Huffman 1986).
Within the context of the settlement systems and culture systems of these southern African societies, the main determinants and indicators of wealth and status in society were patrilineal descent in relation to the royal clan, as well as access to and control of such resources as land, water, and food. However, the most important commodities—particularly in the 19th century AD, when an early European presence was being established in the region—were cattle and trade goods in the form of metals and animal products. The economic significance of cattle in these early societies has been documented extensively (Hall 1987; Smith 1992). Cattle were more than sources of wealth, however; they also served social and political functions that were expressed in other ways as well, including the material realm (use of space). The significance of cattle in southern African economies can be traced to the introduction of livestock into the region by early farming societies from the beginning of the 1st millennium AD. It would appear that by AD 200, cattle and other stock were present in southern Africa as far as the Cape (Walker 1994). Herding was adopted by Khoi communities, resulting in the rise of Khoi pastoralism in the region (Smith 1992). By the 7th century AD, most of the early farming societies in areas that are now known as Zimbabwe, Botswana, and South Africa had incorporated cattle herding into their economies. Their settlements were mainly nucleated small villages with subsistence-based production. Transformations in the nature of herding economies in the region appear to have started toward the end of the 1st millennium AD. This was also the period when long-distance trade goods began appearing in the material culture of the interior (Denbow 1986; Hall 1987). Most archaeologists studying societies forming in southern African concur that significant transformations in regional economies started around the 10th century AD (Denbow 1986; Hall 1987; Huffman 1978, 1982).
The economies of the societies in the Limpopo-Shashe basin and the eastern Botswana hardveld were principally mixed farming economies. Cultivation of food crops, consisting of cereals, legumes, and other vegetables, was carried out alongside the activity of herding livestock, comprising cattle, sheep, goats, and chickens. The domestic dog was probably kept both for hunting and protection of the domestic residences. Most of these villages were situated on open plains, with considerations of water and access to pasture probably playing an important part in choice of area (Denbow 1983; Segobye 1994). Although subsistence production was based mainly on domestic labor, it is likely that clientship relations existed between those who had more resources and those who had little or none. Further, ranking within the societies was probably made on the basis of age, patrilineage, and gender. It is likely that political leadership was ranked on the basis of kinship relations to the royal clan or clans. All these factors must have influenced the structuring of the economies of these societies, and particularly the area of production. This chapter focuses on the significance of labor resources, power, and social organization in the development of these societies from simple mixed farming economies to more complex ones with multiple activities carried out by different members of society and differential access to resources. Other studies have focused on how these changes affected the relations between the hunting and gathering societies that were founded by the incoming farming societies (Denbow 1984; Denbow and Wilmsen 1986). These studies have argued, variously, that the dominant nature of the farming economies over the foraging ones resulted in the relegation of foraging communities to lower ranks within society. The absence of formalized political leadership systems within foraging communities and mobile or semipermanent settlement further contributed to the low ranking of these societies within the broader scope of regional settlement systems. This study, however, examines the settlement systems of farming societies as the physical expressions of networks of social and economic relations that had direct bearing on political authority. Emphasis is further placed on the organization of labor within the production sphere, particularly the role played by cattle in terms of production. This chapter will examine the literature advancing the model that cattle were the dominant resource in these early complex societies and more or less determined the power relations within society, particularly relations of production and reproduction. It has been argued that the settlement systems of places such as Mapungubwe and Great Zimbabwe reflected this division, which one can call gendered division of labor, production, and space. The chapter will argue that while, broadly speaking, this may be true, models that have been proposed to date have focused too much on the role of cattle and by and large have ignored the significance of other activities and activity areas where women could (and probably did) exercise considerable power.
Following research at the Toutswemogala site and other related sites, a model of settlement and social organization was proposed by Denbow that ranked the Toutswe type settlements by size, location, and depth of midden deposits (Denbow 1982, 1983; Lepionka 1977). In brief, the model argues that between the 7th and 13th centuries AD an incipient state system developed in east-central Botswana. This comprised several culturally related but independent polities located in an area approximately 100 by 150 km. The main distinguishing element of these polities was the differential size of settlements and their associated middens, representing kraal (cattle byres) that were also differentiated by location. The largest and most important settlements were placed on prominent hilltops and numbered fewer than five. Denbow saw these settlements as the capitals of these polities and therefore designated them as Class 1 settlements (Denbow 1982). In the intermediate level were more numerous but smaller settlements, also located on hilltops. These, however, had less sizable middens and were placed on smaller hills and kopjes. These settlements, numbering about 15 and thus designated as Class 2 settlements, were located at a distance from the large settlements and obviously played an important role as intermediate settlements with the third class of settlements. This third category, Class 3, comprised small settlements that were mostly, if not exclusively, located on low-lying ground and contained relatively small kraal middens. These clustered around the Class 1 and Class 2 settlements and were thought to be residences for commoners and dedicated mainly to agricultural produce (Denbow 1983; van Waarden 1987). In brief, this model suggests that the early farming societies of eastern Botswana were structured on the basis of ownership of cattle that gave their owners both economic and political power. Owners were further able to exercise control over the use of space for settlement and other activities, such as agriculture. This model has also been used to explain the process of settlement change and resource use at Mapungubwe in the northern Transvaal (Voigt 1983; Hall 1987).
More recent research in eastern Botswana (Segobye 1994; Kiyaga-Mulindwa and Widgren 1983) has documented many more settlements in the Toutswe settlement area. Broadly speaking, in terms of distribution and settlement patterning, they more or less conformed to the pattern observed by Denbow. Variations, however, were noted in the internal arrangement of activity areas at both the intersite and intrasite levels. In addition, the distribution of material culture in these sites suggested that there were probably more fluid relations between the different types of settlements than had been proposed by Denbow’s model. One important observation was that while in general these polities may have ceased to be regionally important after the 13th century AD (Denbow 1986; Voigt 1983; Hall 1987), they were not abandoned completely and in fact continued to be occupied until the 17th century AD (Segobye 1994; Kiyaga-Mulindwa 1992). It has been suggested that the main reason for their abandonment around the 13th century AD may have been environmental, brought about by increased aridity resulting in poor pasture for the large herds the people maintained (Denbow 1986; Voigt 1986). Given that crops are generally more sensitive to moisture conditions, it is likely that agriculture productivity would have declined more sharply than herding, if indeed there was a regional decrease in rainfall patterns. This in itself offers an interesting scenario for modeling changing relations of production during this period. If these climatic oscillations prevailed over the period of the peak of these early state systems, they would have had implications on the relationships of those involved in production, particularly those between male and female producers and ways in which they defined their domains. It is argued that rather than remaining unchanged over this seven-century period, gender relations—like other relations in the production-reproduction arena—must have undergone changes, some of which can be inferred from clues in the archaeological record. These changing relations have, however, been subsumed and, to an extent, ignored by models that projected the process of socioeconomic and political change as having been dictated by ruling patriarchs in these early state systems.
Archaeologists reading the settlement history of southern Africa have often worked on the assumption that, dating from the earliest expressions of mixed farming societies in the region, cattle have always been the most important resources within these societies. It is argued that while this may be true, the argument, like that of the “man the hunter” thesis in Stone Age research, underscores the importance of cultivation and foraging as food procurement strategies. In fact, dietary patterns of some of these prehistoric groups suggest that meat resources from cattle (as opposed to small stock and wild animals) constituted a smaller proportion of the overall food intake. This implies that other sources of protein were consumed. These food sources were most probably derived from agriculture and gathered vegetal foods. Both these activities, which were labor consuming, were undertaken by women. Therefore, in terms of relations of production and the domain of daily subsistence activities, women played a significant role in these early societies.
The main influence on the reading of the significance of cattle in the sociospatial organization of the early farming societies has been the rich ethnography of southern Bantu-speaking peoples, including Shona, Nguni, and Sotho-Tswana groups (Huffman 1986; Huffman and Hanisch 1987; see Hall 1984). Most early anthropological studies of 19th century AD Nguni and Sotho-Tswana societies made observations concerning the internal arrangement of these societies’ residential places and other related activities areas (Schapera 1935, 1952; Walton 1956; Kuper 1980). Basically, scholars had noted the consistency in the use of concentric shapes and circles in the arrangement of homesteads, wards, ward clusters, and villages. In other words, from the smallest unit to the largest, circular forms were chosen. The significance of these was mainly as an expression of social (kinship) relationships within and between the different members of the society. As a result, the layout of the settlement was often a direct reflection of the kinship relations between different peoples. Because these societies were patrilineal, these relations were also expressed in terms of the ranking in order of descent between different male heads.
Interpretations of the social meaning of the use of space and the gender roles within these societies were also drawn from structuralist theory that postulated discrete dichotomies between male and female roles and other spheres of social relations. Within the context of the prehistory of early farming societies, assumptions about continuity in the settlement and culture history of the region’s farming societies also strengthened the premise for modeling the social use of space in these societies (Huffman 1986). As a result, the tendency has been for studies to infer broad spatial and temporal continuities in relations of production, specifically gender relations within communities having mixed farming economies whose settlement systems were thought to conform to the model proposed by Huffman, namely the Bantu Cattle Pattern (also known as the Central Cattle Pattern) (Huffman 1986; Huffman and Hanisch 1987).
This paper does not attempt to present an alternative model to explain sociocultural behavior in these prehistory societies. Instead, it highlights the observation that some of the assumptions previously held to be absolute may need further research—particularly in light of both new research in the region and gender perspectives within archaeology. This chapter comes to the conclusion that postulating the process of complex social formation as an intrinsically male-directed process ignores the potential contribution women made to this process and, more specifically, the diversity in forms of social organization and social relations over time and through space. Relegating women and their activities to the level of secondary factors in the process of social and economic complexity in some ways assumes that women were passive commodities, to be traded, distributed, and stored in a process of wealth- and power-seeking by male heads. The idea that women were valuable only insofar as they were able to reproduce surplus labor and to bear daughters (who could, in turn, be married off for creating social or political alliances) oversimplifies relations between different members of society who in their daily interactions probably had to negotiate power and social identity in more ingenious ways.
In the Mogkware hills, where research was undertaken between 1988 and 1996, sites identified as Toutswe tradition settlements were found in different locations, including hilltops, slopes, and valleys. Although these settlements conformed to the pattern Denbow identified in the other sites in the hardveld area, it was noted that some sites had a greater diversity of activity areas, including pottery-making and metal-working places. In addition, some sites were located farther from areas with fertile soils suitable for agriculture and access to water resources. In light of these observations, it was concluded that the relations of production and access to resources in the area must have also included communal or shared access, because of the factors of location of settlements and resources in their immediate vicinity. In addition, the problem of recurring drought conditions and aridity made the entire production process risky. This could have had one of two contrasting effects. First, political leaders (male) could have increased their control over their diminishing resources and made more elaborate rituals concerning the use of these resources, thereby increasing their power. Second, communities could have been drawn together more, with less differentiation on the basis of ethnicity, gender, or other ranks. More-fluid rules of incorporation into different social categories and ethnic groups could have been preferred as a means for facilitating access to scarce food resources. This seems to have been the pattern that 19th century AD communities opted for, following the Mfecane wars. However, as mentioned above, it is harder to generate such a model without adequate information about the range of activities or directions of trade within and between the communities that occupied this region from the 10th century to around the 17th century AD.
This chapter has tried to identify the areas of input that can be made from a gendered reading of the prehistory of southern African farming societies prior to the arrival of known ethnic groups. Two main points are made. First, because research has not focused specifically on this issue in the past, current models have tended to assume that the process of socioculture change was more or less solely directed by male activities—primarily herding. Second, it has been assumed that a certain degree of continuity existed in the occupation history of the region that was reflected in similarities in settlement systems over an extensive area. As a result, the role of women in the societies has been subsumed under the dominant model of male-directed herding economies that were further enriched by long-distance trade after the 13th century AD when the Great Zimbabwe empire reached its peak. What this chapter calls for is a reexamination of this broad model, to take into account the changes within individual societies or polities where women’s contribution to production could have led to power relations’ being constituted differently among the different social groups. One area worth consideration is the reexamination of the role of craft production, including pottery and cultivation, in sustaining the wealth base of a homestead or village. Although the data may not be immediately obvious in the current archaeological literature nor serve as tools to investigate these issues, different approaches need to be considered, rather than science’s simply resting in the comfortable assumption that one general model explains such a process over a long period of time and throughout such a diverse area.