Archeological Papers of the American Anthropological Association

Changing Responsibilities and Collective Action: Examining Early North African Pastoralism



This chapter uses the concept of an economic transition point to examine and challenge assumptions about static gender roles in pastoralist societies. The transition from a hunting-and-gathering economy to an agro-pastoralist economy would have forced North Africans to perform tasks related to both economic systems until the pastoralist way of life could be firmly established. The greater need for labor in these transitional societies would have discouraged the formation of strict gender divisions like those seen in today's pastoralist societies. The changing needs of the economy would have necessitated a flexible division of labor. To examine these changing needs, this chapter examines evidence from rock art, archaeological sites, and modern pastoralist ethnography.

The words African pastoralism evoke not only a subsistence system based on the herding of cattle or sheep but also an ideology that emphasizes the social superiority of cattle-owning males over females who merely perform routine household tasks. The prevalence of this ideology can be traced to historical processes, which include the colonial powers' tendency to interact only with male authority figures and the production of ethnographies by male ethnographers who imposed the Western idea that male activities were the only “important” ones onto the activity patterns in pastoral societies. Dorothy Hodgson terms this historically produced narrative “the myth of the ‘patriarchal pastoralist’“and argues that”‘ownership’ is itself a historical category, with associated presumptions about private property, alienability, and individual control that differ from more communal and cooperative notions of rights and responsibilities” (Hodgson 2000a:11).

This chapter questions the assumptions that regard women's activities in pastoral societies as invariant and largely irrelevant by examining the “rights and responsibilities” involved in the initial domestication of cattle by early Holocene agro-pastoralists. Agro-pastoralism, which includes elements of agriculture and animal herding, gradually replaced hunting and gathering across many parts of North Africa in the early Holocene. An examination of this shift allows us to look at the feedback between gender and economy at an economic transition point, a point at which gender roles have been “shaken up” by the introduction of new tasks and the consequent need for extra labor. In this situation, some members of the workforce must take on new tasks and others must compensate by taking over old ones, at least until new subsistence patterns or technologies have made the old patterns obsolete.

It can be helpful to think of tasks as becoming metaphorically ungendered and then re-gendered as one subsistence system combines with and eventually supersedes another. This type of thinking, though, can also lead to the problem of gender attribution, which is the assumption that each gender has assigned tasks within a certain subsistence system that are just waiting to be “discovered” (Brumfiel n.d.; Conkey and Gero 1991:11). A cursory look at most modern ethnographic work will show that although some duties may be ideologically associated with people of one gender within a certain group, in practice people of another gender often perform these same duties if it is expedient or necessary (Brettell and Sargent 1997; Murdock and Provost 1973).

Assessing ideological associations within a prehistoric society is always one of the most challenging aspects of archaeological work. In addition to the archaeological record, two other tools exist for analyzing the gender ideology of Holocene agro-pastoralists: ethnographic analogy and rock art. However, two major problems can occur when using modern ethnography to interpret ancient art. First, ethnographic insights may be projected uncritically onto the art; this projection generates the false notion that society has remained static since the rock art was created. Second, analyzing art through an ethnographic lens may lead to the assumption that the art represents everyday life. When rock art is analyzed instead as a culturally specific collection of ideals about tasks and relationships (Barich 1998; Quinlan and Woody 2003), both its limitations and its value for interpreting archaeological data become clear.

A successful re-gendering of archaeological data, then, considers both whatever ideological data are available and the archaeological evidence for tasks and task combinations and creates one or several likely hypotheses about human actions based on this synthesis. These hypotheses should recognize the flexibility of the actors and the existence of variability across time and space. As Conkey and Tringham (1997) have noted, some ambiguity will always be inherent in archaeological interpretations of gender. However, Brumfiel argues for selecting the “subset of most likely pasts” in order to consider women's (or other “disadvantaged genders”) contributions to past societies (Brumfiel n.d.:10). She makes a convincing argument that negating current views of past gender relations is unlikely to convince scholars or the public to abandon their views unless new formulations are provided. Therefore, gendered activities may be most productively viewed as flexible in any given social situation but not as totally ambiguous.

This chapter will focus on the re-gendering of tasks during the economic transformation of Site E-75-6 at Nabta Playa, an Al Jerar phase (ca. 7700 b.p.: 14C) site in southwestern Egypt. Archaeological evidence suggests that the inhabitants of Site E-75-6 were in the early stages of domesticating cattle and sorghum. However, most residents of this site continued to rely on hunting and gathering as their major source of food. During this transitional period, all the resources of a large group would have been required just to perform the daily tasks of subsistence. Rock art across the Sahara indicates that pastoral tasks involved men, women, and children, sometimes acting alone and sometimes in concert. This chapter will explore tasks and task complexes at Nabta Playa by using evidence from rock art and ethnography in addition to the archaeological evidence from Nabta Playa itself.

The Problem of Ethnographic Analogy

Ethnoarchaeological studies have proven useful for studying the adaptation of gendered practices and ideologies in the recent past and especially for examining those changes wrought on pastoralist societies by colonialism and climate alterations. These studies, such as Smith and Webley's (2000) study of the Khoekhoen, have challenged the gendered accounts of current pastoralist societies, which assume that women are simply subordinate adjuncts to their patriarchal, cattle-owning husbands and that these gender relations also existed in precolonial times (Gifford-Gonzalez 1998:118). Ethnoarchaeology counters these assumptions by revealing women's sources of power in the past and by showing how some of these past sources of power persist to the present.

This type of ethnoarchaeology is useful for studying groups in the distant past, even those of the early Holocene, insofar as it prevents archaeologists from uncritically applying modern gender stereotypes in their investigations. However, no modern or even recent precolonial group provides an adequate ethnographic representation of the transition from hunting and gathering to agro-pastoralism, simply because the process of domestication is already complete for the most common species used in food production. Most groups today can specialize in one particular form of subsistence because of these domesticates (Banks 1984). Another key ethnographic difference stems from the increase in territorial wars among pastoralist groups. These wars often allow men to claim privileged status as warriors (Earle, personal communication 2006). No evidence for war is found at Nabta Playa or most other early pastoralist sites, and the low population of the area would probably have made territorial conflicts unlikely.

This does not mean that ethnography is useless for analyzing gender relations in the past. Gifford-Gonzalez suggests, “We must employ ethnographic evidence to move us beyond facile gender and culture stereotyping, to locate those enduring and universal facts of pastoral life to which all groups and households engaged in the keeping of herds and flocks must respond” (Gifford-Gonzalez 1998:123). Ethnographies can thus establish a possible universe of subsistence requirements that human actors can fill in varying manners. Human agency appears in the tasks that the people at a particular site chose to perform in order to cope with environmental and social pressures, and it is in this choice that archaeologists can locate information about gender.

Societies undergoing transitions from one subsistence pattern to another require this conception of a possible “universe of tasks” in order to make sense of the combination of strategies that today would be undertaken by separate groups. Modern pastoralists tend to trade for (or buy) agricultural goods from neighboring agricultural societies, leaving the pastoralists free to follow the herds throughout the wet and dry seasons (Banks 1984:213). In early agro-pastoralism, agricultural labor would have been added to herding duties, and there would have been a continuing need to hunt and gather for the majority of resources. The “universe” of duties thus becomes larger than in more specialized subsistence systems. This implies that each person would have performed a more varied set of these duties than he or she might in a more specialized system. Stereotypes such as “Woman the Gatherer” or that of the pastoralist woman who only tends her husband's home become impossible to uphold in this type of “mixed economy,” as Barich (1998:113) terms it. There would simply be too much work left over for the men to perform if women's duties were restricted to one particular sphere (Barich 1998:114).

The following case study examines the site of Nabta Playa E-75-6 in the Egyptian Sahara. It is meant to closely examine some of the site's characteristics that indicate early steps in the domestication of animals and plants and to theorize about the task patterns that could have helped to create these particular features.

Sorghum and Cattle at Nabta Playa

Site E-75-6 at Nabta Playa is one of the earliest known sites that exhibit evidence of a probable shift from hunting and gathering to agro-pastoralism. Nabta Playa lies in the eastern Sahara, within the borders of present-day Egypt (see Figure 6.1).

Figure 6.1.

Location of Nabta Playa. Courtesy of the University of Texas Libraries, The University of Texas at Austin.

The site was occupied during the El Nabta/Al Jerar phase as defined by Wendorf and Schild (2001), with radiocarbon dates clustering at approximately 8000 b.p. (Banks 1984:66; Krolik and Schild 2001:111). Archaeological evidence indicates a semisedentary population that may have been domesticating both sorghum (Sorghum bicolor) and cattle (Bos primigenius f. taurus) (Banks 1984:236). This constellation of semisedentism, domestication of animals, and domestication of plants certainly represents a departure from the mobile patterns of hunting and gathering that had been common.

During the El Nabta phase (ca. 8000–7770 b.p.: 14C), Site E-75-6 consisted of two rows of huts organized around a basin in which water would have pooled (Krolik and Schild 2001:111). The site was apparently reoccupied during dry seasons for some time, though whether the occupations occurred every year is unknown. The same hut areas were reoccupied frequently, as shown by midden deposits from various phases interspersed with sterile deposits. This suggests that the occupant group remained fairly consistent year by year, though new huts were added that may have either replaced some of the old ones or accommodated a growing population. Major features at the site include huts, bell-shaped storage pits within and outside these huts, and four wells that were probably dug during separate occupations (Krolik and Schild 2001:117–120).

Hunting and gathering remained the primary means of subsistence for the residents of E-75-6, as shown by the assortment of faunal remains and various wild grain species found at the site. Domestication is indicated by rarities or anomalies within this hunting-and-gathering system that are nevertheless significant enough to require explanation. Barich notes that “[t]hese situations [early Holocene adaptations] can be better defined as transition contexts, in which prototypes appeared for the first time, that were later destined for domestication, through repetitive gathering and selection processes” (Barich 1998:107). This provides the important reminder that domestication comprises these “gathering and selection processes,” at the end of which animals or plants are said to be “domesticated”; but domestication itself is a process and not a state. The early stages of this process would therefore look much like the previous subsistence system.

Bearing this in mind, irregularities like that found in the distribution of sorghum at E-75-6 seem suddenly very important. Archaeologists at Nabta found a large number of charred cereal caches in storage pits, which all date to the El Nabta phase. Analysis of the cereals led to the interesting conclusion that the caches belonging to various huts fit into two types. One type, found in several huts, contained a large amount of sorghum (up to 93.9 percent of the total sample) and relatively few other types of cereals (Wasylikowa 2001:580). Another type, also found in several huts and in most of the bell-shaped storage pits outside the huts, contained a wider variety of cereal taxa but far less sorghum (as low as 0.8 percent of the sample) (Wasylikowa 2001:572).

Various models have been proposed to account for this uneven distribution, but it seems fairly evident, as Wasylikowa notes, that some family or household groups were cultivating or preferentially gathering sorghum while others were not. The sorghum all appeared wild, but Wasylikowa believes that its similarity to wild sorghum could simply be an indicator that it was in the early stages of cultivation. She suggests that this division may reflect either preferential gathering for human consumption during years when sorghum was common or that some kin groups who were familiar with sorghum may have begun to cultivate it (Wasylikowa 2001:581).

The “preferential gathering” hypothesis alone cannot adequately explain the decrease in plant diversity among the sorghum-rich deposits. A preferential-gathering explanation would argue that once more sorghum was available, people would eat fewer of the less desirable plants that they had gathered before. This model, while often a valid explanation of varying plant diversity, does not adequately explain the extreme range of variation in the plant remains at E-75-6. Unless the environment swung between two poles—being either so strained that almost no sorghum was available or so rich that it allowed for a sudden sorghum explosion and collection of samples with up to 90 percent sorghum—the preferential-gathering model is unlikely to be valid. The stark distinction between the sorghum-rich and sorghum-poor households therefore suggests that yearly variation in the environment cannot have been solely responsible for these gathering patterns.

Another explanation for differential sorghum gathering or cultivation may be that the sorghum was produced in conjunction with another task in which only some of the community was involved. Some archaeologists, such as Magid, have suggested that sorghum was not native to Egypt and was instead brought there by cattle herders from the Sudan to ensure a reliable food source for their herds (Magid 1995). Sorghum today is commonly used for feeding livestock in arid regions, but in order to provide its full nutritional value to cattle it must be processed first by grinding or some other method to remove the bran shell. The huts containing a greater amount of stored sorghum therefore might belong to family groups who were involved somehow in raising or maintaining cattle. It is possible that the sorghum was gathered or cultivated independently of cattle grazing and stored for cattle (or human) feed—or that it was gathered while the cattle were grazing and transported back to the site. This would suggest that the two separate tasks of cultivating or gathering sorghum and cattle herding were connected to one particular group of people with similar subsistence practices.

A small number of cattle bones were discovered at E-75-6. Gautier (1980, 1987, 2001) has identified these bones as those of domesticated cattle, though this identification is questioned by Smith (1992). However, as Banks points out, the relationship created between animals and humans by domestication has both a biological and a social component. In Banks's view, the biological component of domestication comprises the physical effects of domestication on the animals, while the social component is made up of the changes in behavior of both humans and animals as a result of this new relationship (Banks 1984:203). Taking a social view of the relationship between cattle and humans in the early stages of domestication may produce a different opinion about whether cattle were “domesticated” than taking a biological view.

Given that the presence of early domesticates is always likely to be contested, in order to talk about the effects of early agro-pastoralism one must examine a place where there is some contention about whether domestication had actually occurred. Thus, the topic of the Nabta cattle's domestication is not entirely settled: some articles (e.g., Smith 1992, 2003) have questioned whether the Nabta cattle were domesticated. However, Banks (1984), Gautier (1987), Wendorf et al. (1989), and Wendorf and Schild (2001) all provide strong arguments as to why humans must have brought the Nabta cattle to the playa. These arguments include the fact that the Nabta site, as one of the most arid parts of the desert, would be unlikely to contain any cattle at all unless humans had brought them there. Relatively complete skeletal remains of cattle show that the cattle themselves were not killed and butchered elsewhere but were physically present at the site before being killed (Gautier 1987).

Gautier (1987:177) argues that the remaining fauna at the site, mostly very small animals, do not match the ecological profile of an area where cattle and other large animals would have been able to survive on their own. Therefore, the cattle, while they may or may not have taken on some biological characteristics of domestication, were almost certainly dependent on human intervention for their survival. Banks's criteria for social domestication are thus met. For the purposes of this discussion, the arguments for cattle domestication at Nabta are considered valid.

Wendorf and Schild (2001) suggest that “task groups” were responsible for the pasturing of the E-75-6 herds at the beginning of the rainy season (see also Barich 1998:110). This hypothesis is based on the existence around the large sites like E-75-6 of smaller sites that were apparently only occupied for short periods; these sites are thought to be cattle camps. Banks notes that “the presence of ceramics, grinding implements, and a full array of tool types at both site types signifies that a full complement of maintenance tasks took place at both site types” (Banks 1984:239). The smaller “camps” would therefore have been occupied by a large number of people for a lengthy period, compared to modern practices in which one or two males follow the herds and camp in different places every day or every few days. The groups who lived at the smaller sites during the wet season might have had primary responsibility for the maintenance of the herds during the dry season as well.

If the groups who cultivated or gathered sorghum were the same groups who owned cattle, then it would seem that “task groups” were not divided by gender but by family or kin group, since the sorghum-rich caches were found primarily in certain huts and not at all in others. Entire kin groups might have moved to cattle camps during the wet season and then returned to E-75-6 during the dry season with their cattle and their surplus grain for storage. Even if some duties such as pasturing the cattle, milking them, and gathering their feed were typically divided by gender, all those who worked with cattle would have shared the pastoralist social identity.

There is no reason to assume that those who gathered, cultivated, ground, or stored grain for cattle feed would be perceived as less involved with cattle than those who actually took the cattle to pasture. The activities that women perform among modern pastoralists, such as milking cattle and marketing the milk, as well as supplementing these dairy products by gathering, growing, or buying crops, are just as integral to the workings of a pastoral society as the cattle themselves (Hodgson 2000a). Modern pastoralists do not primarily use their animals for meat; they use by-products such as milk and blood much more frequently, and they use the animals as resource or wealth storage. Therefore, the people who milk cattle produce the cattle's key contribution to the food resources of the community. The low number of cattle remains at E-75-6 indicates that the residents of the site may have followed this same practice (Banks 1984), although there is admittedly little supporting evidence that dairying had yet become a significant part of pastoralist culture.

Gifford-Gonzalez warns that “we must not refrain from investigating important areas of resource use and site formation simply because they are in the hands of persons whom our culture—or even the culture under study—depreciates” (Gifford-Gonzalez 1998:119). Although pastoral societies and ethnographies often marginalize milk and grain control because these types of food do not always enhance the social status of their female possessors, neglecting the role of these tasks in early agro-pastoralist societies would be a grave error.

Task Combinations and Gendered Relationships

The possible interrelationship of sorghum cultivation and cattle domestication at E-75-6 suggests wider associations between tasks, gender, and social status. Using ethnographic analogy to apply stereotyped gender binaries at E-75-6 would provide an inadequate and ineffective analysis, because a process like domestication requires the contributions of an entire group. Gender attribution obscures the strategic relationships between particular activities and suggests a rigid social order in a society that would have to be flexible to survive in a harsh environment. It is more productive to look at all the activities involved in achieving a certain goal or fulfilling a particular need and then to consider what activities might have been combined by one particular person, whether male or female. In the case of sorghum and cattle domestication at Nabta Playa, the “mixed economy” of hunting, gathering, possible cultivation of crops, and cattle raising would have ensured that women and men performed many different types of tasks.

Even though certain groups at E-75-6 may have specialized in cattle herding, the possible cattle camps distributed around the playa still contain bones of wild game (Gautier 2001). Hunting would have been a major food source if cattle were primarily used for milk products. Grinding stones provide evidence for either cultivation or gathering activities; even if the people at Nabta were cultivating sorghum, they must have gathered much of their other wild food (Krolik and Schild 2001). The need to hunt, gather, and watch the cattle either would have necessitated a division between those people who watched cattle and those who hunted and gathered or else it would have required someone to perform these tasks while simultaneously moving to new grazing areas. In this case, it would make more sense for the gatherers to accompany the cattle, given that appropriate vegetation would probably be available in the same areas where the cattle were grazing. Hunting would be extremely constrained if the hunters had to remain with slow-moving cattle and were unable to track or follow fleeing game.

Gathering while herding would also provide the gatherer with information about the herd's eating habits. Knowledge about cattle's nutritional needs is essential to domestication processes, and finding ways to provide these nutritional requirements would have been a major reason for people to change their settlement and subsistence patterns (Banks 1984:234). The practice of pasturing herds during the wet season would be one adaptation to this system. During the dry season, the residents of E-75-6 may have used stored food like the caches with large amounts of sorghum to supplement whatever food the cattle could find around the playa lake. The people who gathered or cultivated that food for storage would have to have had just as much specialized knowledge as herders about what cattle eat and how much of it they eat, as well as knowledge about which key plants were likely to be unavailable in the winter.

The task combination of gathering plants and tending a small herd seems to make sense. However, a more likely scenario avoids even these rigid task divisions. Hunting and gathering may not always have been separate from one another, and herding was probably not strictly distinct from either of these tasks. Although hunters would almost certainly not take the cattle on their hunting trips, a person herding cattle would probably not hesitate to kill some small animal if possible, and a hunter would not hesitate to gather some choice plant if it did not interfere with the hunt. Those whose primary tasks were hunting and gathering would then supplement these takings. This type of flexibility would have enabled people to carry out all the necessary tasks more efficiently and productively than maintaining rigid divisions between those who could and could not perform a set duty. The supposedly rigid roles of men and women in pastoral societies today (but see Hodgson 2000a and the rest of Hodgson's edited volume [2000] for many critiques of this concept) would simply not make sense for a small group that had to procure all its own resources.

Storage, Choice, and Social Distinction

If sorghum cultivation or preferential gathering at E-75-6 was associated with groups who also owned cattle, then what of the groups who did not have large amounts of sorghum in their households? The above model suggests that sorghum was perceived as cattle feed and that groups who did not own cattle would therefore have no reason to gather or cultivate the plant. The small size of the community would imply that knowledge of how to recognize or grow sorghum could not have been effectively restricted. Therefore, the people who stored little sorghum probably did so with just as much intention as the people who stored a great deal.

If the groups who stored sorghum were the only ones who owned cattle, then the community would have been composed of both hunter-gatherers and agro-pastoralists who had chosen to follow different subsistence systems and still live in the same place, which leads to a variety of questions about how these two groups interrelated. Each household appears to have stored food for its own use, and there is no visible evidence for social stratification on the basis of the structure of the households, the amount of food stored, or the ownership of domestic animals (Krolik and Schild 2001). The wells that were dug at the site may indicate a degree of social cooperation in carrying out necessary tasks. The absence of any other form of social hierarchy based on cattle ownership could indicate either that every group at the site owned cattle or else that at this early stage of domestication cattle were simply perceived as an important but optional addition to the hunting-and-gathering lifestyle.

There are two alternative possibilities to consider. The first is that people stored or cultivated sorghum to feed themselves, unrelated to the ownership of cattle. The second is that all the groups at E-75-6 owned cattle but fed them different types of diets. Neither of these alternatives can offer a definite reason as to how separate groups chose a proportion of sorghum to include in their stored grain. Both of them, however, emphasize the fact that the gatherers in each residential group consciously chose the types of plants that they wanted to store, rather than responding blindly to environmental dictates. Gathering may even have been a means of social differentiation from other residential groups who gathered different plant types or cultivated sorghum.

In any of these scenarios, gathering or cultivation would have been a central activity in the lives of residential groups. Gathering or cultivating plants for storage to feed either cattle or humans might have required agreement from the entire household on which plants were best to store that year, and problems with grain storage would have had an adverse effect on any kin group, possibly resulting in hunger or starvation for the people and their cattle. The acts of gathering, cultivation, and storage would thus have been essential to survival for the pastoralists of E-75-6.

Rock Art and Gender Ideology

Rock art has often been used to legitimate claims about gender roles, especially in terms of gathering, hunting, and herding, because it is the main window that archaeologists and art historians have into the ideology of early Holocene societies. This makes rock art a valuable source for considering gender and social change. Using rock art independent of substantive archaeological or historic evidence is, however, a problematic endeavor. Subjective ideological interpretations have often turned rock art into a simple tool for projecting current ethnographies into the past. The people in the pictures become classified as “male” or “female” by the activities that they perform, not by their appearance. This creates a circular argument: a figure is male because it is hunting and a hunter must therefore always be male.

Interpretations of the rock art scene in Figure 6.2 have suffered from this type of fallacy. The scene comes from the Gilf Kebir, which lies in southern Egypt near Nabta Playa. The figures to the left are almost certainly female, because male figures at the same site always have male genitalia. Some of the women in this artwork are holding “sticks.” When males perform this activity within rock art scenes, they are said to be herding the cattle that are near them. The original caption for this picture stated that it depicted a group of women holding “ritual sticks” (Fliegel Jezerniczky Expeditions 2005). Although the women in this picture may or may not be herding the nearby cattle, the double standard is clear. Men holding sticks near cattle are assumed to be herding; women holding sticks near cattle are doing something indeterminate.

Figure 6.2.

Women with sticks and cattle in a rock-art scene from southern Egypt. Redrawn from Figure 37, L'art rupestre nord-africain, p. 63 (Archaeological Institute of Human Paleontology).

The scene in Figure 6.2 belongs to the Pastoral or Bovidian phase of Saharan rock art (ca. 5500–2000 b.c.e.), which is associated with both cattle and ovicaprid (sheep or goat) pastoralism (Brooks et al. 2003:5). This is one of the two major “phases” of Saharan rock art that might reveal some of the preoccupations and practices of the population at E-75-6. The other is the Round Head period (ca. 8000–6000 b.c.e.), which has often been associated with hunter-gatherer groups in the Tadrart Acacus and Tassili n'Ajjer. The gap between these two phases falls squarely into the hypothetical transition period at E-75-6. Rock art can only be assigned dates in the very few cases when it is directly associated with datable archaeological deposits: for example, at Uan Telocat rockshelter in the Tadrart Acacus, deposits that had built up atop Bovidian paintings on the wall dated to 6754 b.p. (14C), so the paintings themselves dated to approximately 7000 b.p. or earlier (Smith 1992:130). However, the vast majority of rock art is subjectively dated using its content to relate it to either hunter-gatherer societies or pastoralist societies, and most rock art is considered to be part of the ideologies of established hunter-gatherer or pastoral systems.

This view tends to make most of the analyses based on these phases useless for analysis of transition stages like that of E-75-6. One exception is found in the work of Barich (1998), who approaches the transition between the hunter-gatherer and pastoral phases by looking at the relative frequency of women appearing in paintings of both the Round Head and the Pastoral phases. She notes that women appear much less frequently in paintings of the Pastoral school than in those of the Bubaline phase, which includes the Round Head period. She then hypothesizes that women had lost social status during the transition to pastoralism.

Although this conclusion may be valid, it also may reflect pastoralists' tendency to represent gender in different ways than hunter-gatherers did. Women and men in Round Head art are often clearly gendered through depiction of breasts and penises. In pastoralists' art, gender distinctions are not always as clear. Holl (2004) analyzes a set of pastoral paintings from the Dr. Khen shelter in the Iheren region of the Tassili n'Ajjer in which neither men nor women are readily identifiable through representations of breasts, penises, or general body shape. He identifies men as the figures with bare upper bodies and women as the figures wrapped in layers of cloth. His basis for this identification relies on the different activities that the bare-upper-body figures and the wrapped figures perform; he argues that the bare figures are the ones who herd cattle and make decisions about the movements of the group, while the wrapped figures perform more peripheral tasks like tending children or being married off by their male guardians. This interpretation leads to a circular argument: because wrapped figures perform stereotypically female tasks, they are female, and because these “females” do not perform stereotypically male tasks, there is clearly a division of labor between males and females. The division between males and females in this case appears to be an artifact created by the structure of Holl's analysis.

Holl's example highlights some of the problems that can occur when trying to differentiate genders within pastoralist art. His identification of male and female figures is quite possibly correct, but his allocation of decision-making, marriage negotiation, and child care to specific genders relies largely on ethnographic analogy. This automatically reproduces the biases of modern ethnography in the ancient artwork. Indeed, although women are shown playing an active part in the majority of the scenes that Holl analyzes, in his final analysis of the Iheren paintings women are reduced to a mere “good matrimonial alliance” for a “young and intrepid herdsman” (Holl 2004:137) or relegated to even greater invisibility in the phrase “Saharan herdsmen and their families and chattel” (Holl 2004:139). This type of assessment reinforces the belief that women become less important in pastoral paintings than in Round Head paintings, although this belief may be an artifact of ethnographic analogy.

Holl himself calls attention to another problem with ethnographic analogy as applied to rock art. Taking an ethnographic approach to artworks can leave the impression that they represent scenes from everyday life. However, Holl (2004) argues, on the basis of the locations of North African rock-art sites relative to the communities that housed the artists, that rock art in the Sahara functioned as a documentation of pastoralist ideology rather than a series of scenes from everyday life. The communities are located in low-lying areas, while the art lies in high, mountainous areas, on the sides of cliffs, or in caves. These locations indicate that rock art probably had some special ritual or symbolic meaning; it did not simply depict the life of a pastoralist community (Holl 2004:2). The art thus may reveal some typical (or atypical) activities of pastoralist women, men, and children, but it cannot by any means provide comprehensive information on “who did what.”

Ethnographic analogy still offers a valuable tool for analyzing rock art, when it is used with caution and with a clear recognition of the flexibility of early pastoralist societies. For example, women appear to be associated with milk production throughout the clearly gendered Round Head period. In Round Head rock art, women themselves sometimes produce milk from their breasts, and the animals they interact with often appear to be lactating as well (for examples, see Sansoni 1994). This shows an ideological association between women and milk that may have developed into women's control over milk products during the pastoral phase. Modern ethnographies of pastoralist societies, like that of the Maasai, verify that women often have (or had, until recently) rights over the marketing of cow's milk (Hodgson 2000b:110).

Conclusion: Re-Gendering the Past and Present

Although some tasks may have been linked with women and others with men, the people at E-75-6 had to perform a wide variety of different activities in order to survive, including some that may have been ideologically linked to people of a particular gender. The transition situation in which early agro-pastoralists functioned demanded a great deal of labor from small task groups, and people would have had to combine tasks in creative ways in order to keep up with the workload.

The best way to differentiate task groups at E-75-6 does not appear to be through gender but by looking at those households that chose to cultivate sorghum and domesticate cattle and those that chose to follow a more traditional hunting-and-gathering lifestyle. At an early point in the domestication process, it makes sense that some households would place more emphasis on hunting and gathering while others experimented with new methods of subsistence. This is exactly what the difference in stored grains within households would seem to indicate. Those households with a wider variety of stored grains probably relied more on gathering various wild plants for food, while the households with more sorghum probably were somehow invested in using sorghum in particular as a domesticate or as a high-calorie source of dry cattle feed.

If households were divided from one another in their methods of subsistence, then ideological statements of difference would be more likely to focus on the tasks that distinguished cattle owners from hunter-gatherers rather than on tasks that distinguished men from women. This may be responsible for the proliferation of pastoralist rock art soon after domestication occurred. Pastoralists may have been trying to set themselves apart and extol the virtues of their lifestyle through ritual artwork. The pastoralist household would have been perceived as a coherent economic unit and then placed in opposition to the households that used hunting and gathering as their primary mode of production.

Modern ethnographies can provide hints as to some of the tasks that the E-75-6 residents would have dealt with, and rock art can hint at how they felt about those tasks. The model presented in this chapter questions assumptions about the past and present roles of women among pastoralists. The broad spectrum of artifacts present at early “cattle camps,” for example, blurs the distinction between those who herd and those who milk, tend, and grow crops and challenges modern notions of cattle ownership. Strategic gathering and agriculture, tasks that are demeaned among pastoralists today, may have been essential to survival in the past, and both men and women would have been heavily involved.

The process of generating hypotheses to re-gender past and present societies should be an ongoing one, supported by ethnographic, historical, and archaeological evidence. African pastoralism is by no means the only type of society in which gender roles have become essentialized and deny women their proper authority. In order to continue the process, additional villages and sites of everyday life should be investigated and new arguments generated that provide alternatives to existing stereotypes about gender in the past and present.