Engendering the archaeological record or the prehistoric past requires that we ask and attempt to answer a number of difficult questions. Most specifically we may ask who collected the shellfish, who painted the rock paintings, and who made the stone artifacts we record and collectively regard as “the record.” Was it men, was it women, was it both, was it some men or some women, or was it anyone who could? More generally we may wonder what it was like to have been a man, a woman, or perhaps an old woman, in some particular prehistoric time and place. And it follows from this that we should wonder how male, female, or other roles were maintained, challenged, and changed in social situations where men and women held different agendas, enjoyed differential access to status and power, and may have led very different lives. In short, engendering the past is a search for resolution in the social dimension and an attempt to reflect the experiences of different categories of person. It is a parallel endeavor to the resolution of space and time that, when taken together, offer a resolved and therefore meaningful past.
During the heyday of ecological archaeology the relations that were considered significant were those between people and their natural environment (for a critique see Mazel 1987). Although the environment was conceived of as complex and as incorporating many interactive components such as rainfall, soils, vegetation, and resources, society was often depicted as unified and simple, apparently devoid of internal structure or divisions. Not surprisingly society, incorporating undifferentiated “people,” generated no latent energy for change, no contradictions to be accommodated; explanations, consequently, focused on environmental or climatic change. Coincident with, and in fact necessary for, a shift to an interest in social explanations for changing archaeological signatures has been the recognition that past societies were far from simple. Thus, the recognition of interest groups with conflicting objectives or competitive visions allows the engine of change to be situated in the social as well as the natural environment. This shift has meant that “people” now have to be resolved into men and women, young and old, initiated and uninitiated, relatives by blood and by marriage, stone tool makers, bead makers, dancers, and painters. I view gender as a component of this interest in greater resolution.
The problem is, clearly, how to achieve the resolution of person, how to recognize the roles of and rules for women or men at particular times in the past. For me it is not simply a matter of problematization, a stage that has been effectively achieved very largely through the momentum of the feminist movement (Conkey and Spector 1984; Gero and Conkey 1991). It now appears to be an issue of method, a need to operationalize our understanding of the roles of men and women by relooking at archaeological evidence through gender-sensitive eyes. My suggestion here is that to find gender, the culturally defined roles of men and women, requires first that we find men and women. It is hard to disentangle the specific question “who collected shellfish?” from the more general “what was it like to have been a woman?” or the issue “how were gender relations constructed, perceived, and changed?”
To pursue this agenda I describe some work being done by archaeologists interested in the history of Stone Age precolonial people in the Western Cape of South Africa. The intention is not to claim that the arguments are conclusive, or even persuasive, but rather to illustrate an approach to seeing men and women in the distant past as actors in a social as well as an environmental arena. My thread is the argument that there is a consistency among very different kinds of evidence, written and material, about gender relations among Southern African hunter-gatherers (often referred to as San, Basarwa, or Bushmen) for the past few millennia, but that this becomes elusive when we look back as far as the terminal Pleistocene. I suggest three sources, from which men and women are clearly visible: ethnographic texts, rock art images, and buried human skeletons. I refer also to some less well-anchored speculations drawn from other parts of the archaeological record such as stone tools and faunal remains. Although spatial patterns in archaeological remains may relate to gendered behavior, I agree with Kent (this volume) that most patterns reflect the distribution of activities rather than gendered space.
Ethnographic accounts of former hunter-gatherer communities in the Kalahari (among them Katz 1982; Lee 1979; Marshall 1976; Silberbauer 1981; Tanaka 1976) obviously contain direct observations of men and women and have been a popular entry point for Western Cape archaeologists interested in various technological and social issues including gender (as examples see Parkington 1972, 1977, 1984, 1988, 1996; Solomon 1989, 1992, 1994). This is not unreasonable because it is widely believed that there are strong genetic, linguistic, and cognitive links between 20th-century people of the Kalahari and late precolonial communities of the Western Cape. More pertinently, perhaps, we have observations made by early travelers and early settlers on the lives and behaviors of hunter-gatherer and pastoralist groups that survived into the 17th or even 19th centuries in the Western Cape, albeit often as dispossessed and subordinate communities (see, for example, the reports in Barrow (1801–04); Sparrman (1785); Thunberg (1795–96)). Even more direct references to the lives of men and women come in the form of the 19th-century accounts of life in the Karoo region of the South African interior collected by Wilhelm Bleek and Lucy Lloyd (Bleek 1924, 1931, 1932a,b, 1933a,b, 1935, 1936; Lloyd 1911). These/Xam people were former hunter-gatherers who were being incorporated into colonial society as farm laborers. It is not the intention here to deny the usefulness of ethnographic or historic observations for suggesting explanations of archaeological remains. Quite the contrary, the implication is that gender roles may not have changed much in the last thousand years before colonialism. Perhaps the problem, though, is well made by returning to some of the questions posed earlier. Who painted, who collected shellfish, who made stone tools? In the Kalahari and the Karoo the answer is, simply but obviously, no one does any longer. Historical and ethnographic accounts incorporate changes resulting from colonial intervention, so that some behaviors had disappeared prior to reliable observations and others had changed.
What, then, does it mean to be informed by the ethnography, more specifically on the issue of gender? Women emerge as the gatherers of plants, providers of carbohydrates, collectors of the major components of the diet; as the makers of ostrich eggshell beads; as the singers and clappers but not normally the dancers at healing dances; as potential “owners” of water holes and central figures in the maintenance of group identity; as users and maintainers, but perhaps not makers, of digging sticks and perforated weight stones; as closely identified with domestic space and duties but not so with wider-scale political roles. Men, by contrast, are meat providers, being solely responsible for the hunting of large game with the bow and poisoned arrows, though this may happen rarely and contribute occasional bounties rather than regular fare. Men also manufacture most hunting and gathering gear, engage in dance and trance more frequently than women at healing occasions, are more likely perpetrators of violence against both men and women, and emerge in interactions with farmers and administrators as spokespersons for former hunter-gatherers.
So much for the economics and technology, but Megan Biesele’s work (1993) on the expressive culture of Kalahari Ju/’hoansi has revealed the deep and pervasive role of gender relations in the language and oral traditions of such groups. What emerges is a complex metaphorical framework whereby men and women articulate the tensions of gender relations, incorporating and focusing on the roles of “man the hunter” and “woman the childbearer.” Encapsulated in the belief system of n!ao (n!ow in Lorna Marshall’s [1957] usage), the economic and procreational levels are conflated so that a common language for sex and hunting is used. McCall (1970) has aptly noted that men “hunt” women as they hunt large game. Equivalents are noted between the penis and the arrow, the semen and the poison, the blood of the kill and the blood of birth. It is also apparent from Biesele (1993) and other texts that these tensions characterize the relations only of adult and sexually active men and women. Girls before first menstruation and boys before their first large game kill—in each case the first shedding of blood—do not engage in the metaphorically phrased struggle, and older men and women—after the last shedding of killing or menstruating blood—are less antagonistically related. These tensions find expression in the notion that harm enters society through women, especially women who are, or expect to be, menstruating, who thus are disqualified from becoming healers and must adopt the supportive role of clapping and singing at the healing dances.
Nor is such a framework limited to the Kalahari of the mid-20th century, though it is best reported here. Clearly the comments of/Xam Bushmen of the 19th century Karoo to Lucy Lloyd and Wilhelm Bleek reflect a version of n!ao belief, although the word is never used. Thomas’s (1950) northern Namibian (then South West African) observations from the first half of the 20th century and Qing’s comments to Orpen (Orpen 1874) may be taken to imply a widespread set of beliefs that framed gender relations through the perceived equivalence of hunting and sex. These beliefs also include the creation of hunters and husbands at male initiation and prey and wives at female first menstruation rituals (Parkington 1996). A man has two wives, his human wife and his eland or gemsbok mate animal. There is, thus, a direct link between the technology of the bow and poisoned arrows, the economics of the division of labor between man the hunter and woman the gatherer, carnivore and herbivore, and the generalized gender roles of aggressive man and supportive woman.
It is obviously tempting to simply import these wonderfully detailed scenarios into our reconstructions of Western Cape late prehistory. But it is important for us to question whether we are learning much about the past by this form of analogical reasoning (Wylie 1982). Even a brief focus on the question of shellfish gathering may persuade us that we are not. There are many changes in the archaeological record of shellfish gathering in the Western Cape during the Holocene. The volumes of shell midden recorded per unit time vary enormously, as do the sizes, numbers, and placement of sites (Buchanan 1988; Jerardino and Yates 1996; Parkington et al. 1988). The frequencies of different shellfish targets also vary considerably and include periods of undoubted focus on either limpets or mussels as well as periods where more diverse sets of shellfish were gathered. There are some Holocene millennia missing from the coastal archaeological record, suggesting the possibility of a lengthy cessation of shellfish gathering on any scale. The density and range of artifacts and animal bones in shell-dominated deposits also varies, as does the extent to which the animal bones point to seasonally restricted collecting times. All of this points to a far from stable strategy of shellfish gathering and an almost certainly variable relationship between coastal visits and overall settlement systems. Reliance on the ethnography might lead us to assume that shellfish collecting is invariably the task of women, whereas this may have been quite variable through the millennia.
Clearly the ethnographic and early traveler accounts provide ideas and detailed scenarios that can be used to formulate hypotheses about past situations. Testing such hypotheses, in this instance ones about male and female roles, requires that we build arguments to link men and women with component parts of the changing patterns of evidence we call the archaeological record. Where, though, is the check on our constructed history? Obviously ethnography alone does not produce a past, so information from these sources needs to be linked to material traces of past systems. We may not have to penetrate far back into the prehistory of the Western Cape before we have to envisage scenarios quite different from those seen and partially described by travelers and ethnographers.
For many viewers the most beautiful rock art images in the Western Cape, and certainly among those into which the most obvious efforts to incorporate natural detail have gone, are the magnificent paintings of large game, most often the eland. These animals are the most likely to have been killed by men with bows and poisoned arrows and they remind us of the “hunter and his gemsbok wife” (Thomas 1950) or the special relationship between the creator mantis/kaggen and his eland (Bleek 1924). The focus on these animals, to the almost complete exclusion of shellfish, plant foods, and small game such as tortoises that could have been collected by women, encourages us to begin to see the art as part of the same expressive vocabulary as Marshall and Biesele have recorded as n!ao in the Kalahari.
Human figures are far more frequently painted than animal figures in the corpus of Western Cape rock art. As a residual art, however, many of the human figures are not readily diagnosed as those of men or women, and instead have to be recorded as of indeterminate sex. My view, not shared by all researchers (Solomon 1996), is that almost all human figures were intended to be read unambiguously as either men or women, and quite probably particular categories of men or women, more specifically related to differences in age and initiation. The regular depiction of penises on men and buttocks on women draws attention to the sexual potential of these people and underlines the relationship between the paintings and sexual tensions and metaphors. Men are painted with recognizable penises not because they walked and hunted naked, but surely because the painter intended the viewer to be able to distinguish between different categories of person, between men and women, and between younger and older men and women. If we are right, then, here is a finely resolved component of the Western Cape archaeological record that allows us to assess the roles, associations, and references implied by male and female human figures.
Constraining the potential of this situation, and certainly limiting the range of gender interpretations, is the social context and both the ritual and symbolic intent of the painters and the painting (Lewis-Williams 1981, 1993; Lewis-Williams and Dowson 1989, 1994; Vinnicombe 1976). No doubt the paintings are far from a simple visual ethnography and refer more directly to the socially constituted roles of men and women, rather than attempting to illustrate the range of activities and contributions made by them. There are clear patterns reflecting distinctions intentionally drawn by the painters between depictions of men and women, and these are of direct relevance to gender roles and relations. The association of men with bows and, less often but no less unambiguously, women with weighted digging sticks, is entirely consistent with ethnographic accounts, as is the observation that most groupings of humans are either predominantly men or women and seldom include both. Men with the finger-spread clapping convention have never recorded on survey. These features of the painted record can be taken as meaning that the society of the painters was similarly structured along gendered lines and shared many of the gendered roles with that of modern Kalahari groups.
In the Western Cape there are many more depictions of men than of women, in some subregions as many as five or seven times as many (Manhire 1981; Van Rijssen 1980). Most of the large groupings of human figures, certainly those with more than 10 figures, are processional and best interpreted as dances—most likely the dancing associated with initiation events for young men or first menstruation events for young women (Parkington and Manhire 1991). Many paintings, including processional ones, and especially those dominated by men, include conventions that are reliably associated with trance, either the occasion of a healing dance or the geometric forms experienced by shamans during an altered state. The initiation events and the shamanistic references clearly link the paintings with custom and belief as expressed in Kalahari and Karoo ethnographies (Yates and Manhire 1991; Yates et al. 1985).
Consistent metaphorical associations between women and the kinds of large game animals hunted by men with bows and arrows are found in recent Kalahari expressive culture, specifically among the Ju/’hoansi (Biesele 1993; McCall 1970), 19th-century Karoo stories (Bleek 1924), and Western Cape rock paintings (Parkington 1996). The paintings, as Solomon has noted (1989, 1992, 1994), relate explicitly, and perhaps primarily, to gender and sexual relations and should continue to be the focus of any research that attempts to uncover the history of gender in Southern African hunter-gatherer society. It is, thus, extremely important to try to date rock paintings so that archaeologists can actively derive changes in subject matter, which may constitute a unique record of the maintenance, challenge, and change in relations between men and women. The paintings at the moment express detailed resolution of place and person, but not of time, with the result that the detection of change through this evidence is difficult or impossible. Recent dates for buried painted wall slabs in the Western Cape (Jerardino and Yates 1996) of over 3,000 years imply that the association of male and female identities with hunting and procreation respectively may be valid for at least much of the later Holocene.
Buried human skeletons, if dated directly, are easily the most resolved component of the archaeological record. Not only can time, place, and person be derived from appropriate analysis, but also some resolution of the age of the person can be gained. What this means is that any associated grave goods, any pathologies, any inferences from anatomy, tooth wear, or injury, and any measurement of trace element or isotopic composition refer to a specific, individual man or woman. There can be little doubt that here lies the most significant potential for establishing patterned differences between men and women at particular times in the past—differences that could provide the basis for gendered interpretations of diet, health, injury, and general behavior.
In the Western Cape the most promising work so far in this direction is that of Judy Sealy and her colleagues in the analysis of stable carbon and nitrogen isotopes (Sealy and van der Merwe 1985, 1986, 1987, 1988; Sealy et al. 1987; Lee Thorp et al. 1989), trace elemental composition (Gilbert et al. 1994; Sealy and Sillen 1988), and tooth wear and tooth pathology patterns (Gilbert et al. 1994) among dated human skeletons, most of them from the coastal strip. Although there is much discussion about the translation of stable carbon isotope readings into specific dietary profiles (Parkington 1986, 1988, 1991), and though the effects of pregnancy and lactation on bone signal turnover are poorly understood as yet, the patterns of male and female isotope values must potentially reflect differences between the diets of men and women as well as shifts in those differences through time.
Interestingly, the most pronounced separation of male and female carbon isotope values, though still perhaps not strictly statistically significant, is that of skeletons dated to between 2,000 and 3,000 years ago (Lee Thorp et al. 1989). Before this time there are relatively few observations, and after it the values for men and women are clearly indistinguishable. What is interesting about this particular millennium is that it is marked in the archaeological record of at least part of the Western Cape coast by the accumulation of “megamiddens,” very large shell accumulations opposite rocky shorelines with extremely low densities of artifacts and animal bones and showing little evidence of domestic features (Buchanan 1988; Jerardino and Yates 1997; Parkington et al. 1988). This is a reassuring convergence of evidence from the archaeological and archaeometric records and implies some connection between the formation of megamiddens and the level of difference between the diets of men and women, more specifically a higher marine food intake by men.
The meaning of the megamidden phenomenon has been and remains controversial. My position is as follows (Henshilwood et al. 1994). The very large volumes and extraordinarily low densities for any materials other than shell and charcoal in the megamiddens (Buchanan 1988) stand in stark contrast with equivalent values for either cave or open site localities in earlier or later millennia (Parkington et al. 1992). One interpretation for this, but not the only one, is that the megamiddens are the residue of logistically planned visits to appropriate rocky shore localities where shellfish, specifically mussels, were gathered, dried, and accumulated for transport elsewhere, leaving only modest evidence for domestic behavior in massive shell and charcoal heaps. Because of the temporal correlation with a more marked differentiation between the stable carbon isotope values of men and women, and because the protein of shellfish would unquestionably be reflected in the carbon isotope values of the collagen measured, it is tempting to identify men as making up the logistical shellfish parties. Regular, if short, visits to shellfish colonies might have exposed men to higher intakes of marine foods than women, even if the dried shellfish were subsequently shared equally among all band members.
Conventionally, the Kalahari hunter-gatherers are not seen as “logistical” in Binford’s (1980) sense, because they move people to resources rather than the opposite. They practice residential rather than logistical mobility. Thus, the shellfish drying and transportation scenario suggested here departs from ethnographic practice in proposing logistical trips to the coast involving spending some days away from a home base. But Kalahari men are occasionally away from camp for a few nights when hunting, whereas women are much less likely to stay away. Logistical parties of male shellfish processors are consistent with some aspects of ethnographic practice. It remains to be shown why such behavior characterized this particular millennium and why there was such a need for protein resources when terrestrial alternatives must surely have existed. It is also clear that, at present, the information derived from the buried human skeletons, while interesting and important, remains somewhat detached from that derived from the rock art and makes little direct reference to gender.
Most Holocene stone tools in the Western Cape, as elsewhere in much of Southern Africa (Deacon 1984b), are made on small flakes or bladelets struck from single- or multi-platformed cores made of fine-grained raw materials such as sil-crete, chalcedony, or quartz. A variety of small scrapers, backed pieces, and drills are made from these blanks, and residual cores are often included in the assemblages. These cores weigh only a few tens of grams and even when capable of producing tens or dozens more flakes would have been easily carried around in a bag for ad hoc use by a stone tool maker. For the most part, then, the assemblages are microlithic and the conventional reduction sequence involves the production of small flakelets and bladelets suitable for mounting or hafting. It is extremely unfortunate that this pattern of toolmaking did not survive long enough to be reliably reported by early travelers, so that we do not know definitively who made the tools.
In these Holocene assemblages there is another tool type, an adze, that was conventionally made on a much larger flake and that was used for woodworking, to judge by both microwear observations (Binneman and Deacon 1986) and wood shaving associations (Mazel and Parkington 1981). Many adzes have been made on older flakes, some of them dating from the Middle Stone Age (Anderson 1990; Kaplan 1987), and were not specifically struck for the purpose. This can be easily seen from the patination difference between the bulbar surface of the adze and the adze retouch or utilization scars. It is quite likely that Later Stone Age stone tool makers recognized Middle Stone Age sites as useful sources of large, fairly rectangular flakes suitable for mounting as adzes. Cores capable of producing a few dozen adze flakes on demand would have weighed several kilograms and would not have been easily transported. The production of adzes thus completely bypasses the conventional stone tool production sequence and in fact does not require any initial involvement in flaking. This is interesting in light of our argument that the primary, but not only, use for the adze would be for making and maintaining wooden digging sticks of the kind recorded in both ethnography and rock art as associated with women. Bows, once made, need almost no maintenance, but digging sticks need to be made and sharpened regularly (Vincent 1985a).
It is obviously possible that men made adzes as often as they made scrapers or drills, simply recognizing the advantages of raiding a Middle Stone Age site for large blanks over carrying heavy cores, even if such large cores were available. For that matter, women may have made all stone tools, though general ethnographic patterns may make that unlikely. The point is, though, that we have suggestive but not conclusive evidence that links plant food collecting, digging stick use and maintenance, woodshavings and adze frequencies, and women’s work with a circumvention of (read “challenge to”?) the mainstream procedure of stone tool making. Whether this implies some tension between men and women emerging in the toolmaking and tool using arena is difficult to say.
In the terminal Pleistocene, stone artifact making patterns were markedly different; adzes much rarer, less patterned, and more frequently made of quartz; and woodworking not necessarily restricted to adzes (Binneman in press; Parkington and Yates in press). The fit between ethnographically recorded patterns and the archaeological record of this time is so far only examinable at Elands Bay Cave, a fairly large cave that currently lies some 200 meters from the shore in the base of a prominent ortho-quartzite cliff. In the terminal Pleistocene it experienced (if caves can be said to experience) a dramatic change of place as the sea level rose from its maximum low of about minus 120 meters at 20,000 years ago to a high of about plus 3 meters some 6,000 years ago, shifting the coastline some 30 to 35 kilometers east. The deposits of the cave contain a very diverse reflection of this change in the form of variable depositional volumes per unit time and a dramatically variable assemblage of artifactual and foodwaste debris. The argument I have put forward (Parkington 1988, in press) is that the nature of cave occupation changed through this time from one characterized by brief but regular visits by male hunting parties to one of longer visits by complete groups of hunter-gatherers once the shoreline had advanced to within exploitable distance of the cave. The evidence I used for what is at least in part a gender argument is as follows.
Shellfish are almost entirely absent from deposits predating 11,000 years ago but dominate all subsequent levels. With the appearance of shellfish, the densities of tortoise bone and ostrich eggshell fragments per unit volume, and thus by extrapolation, per unit time, increase extraordinarily, while the densities of stone tools, but not bone tools, decrease substantially. Bone tools, including bone beads, as well as ostrich eggshell beads, including many unfinished examples, increase after 11,000 years ago, adding to the impression of extraordinary diversity of materials, artifacts, and foodwaste after that time. Burials are unknown before 11,000 years ago, though five were found, including a newborn child, to have occurred between then and 8,000 years ago.
Although greater space would allow a more sophisticated or nuanced presentation, the pattern is for pre-11,000-year-old levels to be fairly rich in mammal bone and an expediently produced quartz-based stone tool assemblage, with modest amounts of tortoise bone and very small numbers of ostrich eggshell fragments. Shellfish, ostrich eggs, and tortoises, all slow-moving or stationary targets not inconsistent with the ethnographic contributions of women, either first appear or peak at 11,000 years ago along with evidence for an across-the-board increase in manufacturing activities in bone and shell. Longer periods are spent at the site, and occasional deaths in residence result in a modest but noticeable pulse of burials. It seems inescapable that at 11,000 years ago the proximity of the shoreline prompted people to begin to use the site as a home base, but does the absence of arguably collected food and arguably female (on ethnographic grounds) activities imply the absence of women beforehand? Other variables such as preservation and the relatively smaller size of our pre-11,000-year-old sample are hard to eliminate, and supporting evidence from other contemporary sites is not yet available. Closure is some way off, but enough evidence is at hand to make it likely that settlement strategies were then more logistical, settlement patterns more dispersed, and gender relations less likely to have remained static.
Terminal Pleistocene settlement patterns in the Cape have been envisaged as involving large group sizes, large “territorial” ranges (Deacon 1976), and infrequent but large-scale movements, contrasting with the Kalahari-style small group residential mobility of the Holocene. If there is any truth in these speculations, there are enormous implications for the roles of men and women, as well as for marriage rules and consequent residential and affinal arrangements. As an educated guess, the length of residential stay would be inversely proportional to the importance of plant foods in the diet, unless some form of storage or food manipulation was practiced, because large parties of women radiating out from one location would soon exhaust local edible plant foods. At Elands Bay Cave the terminal Pleistocene bone tool assemblage is very substantial but includes not a single example of the bone point or linkshaft forms of the later Holocene. Although others disagree with this suggestion (Deacon 1984b; Wadley Chapter 4 this volume), I take this to imply that bow and arrow technology was either nonexistent or radically different at that time. Certainly the density and assemblage composition of terrestrial animals change markedly at about 9,000 years ago, in what I take to be a major shift in settlement strategy. If it is accepted that terminal Pleistocene hunters were operating without the bow and poisoned arrow, then the metaphorical links between hunting and sex could not have existed, at least not as articulated in ethnographic record. Moreover, the remarkable linguistic parallel drawn between the dying eland and the trancing shaman could also not have existed. There is an enormous difference between an eland slowly succumbing to the effects of poison and an eland felled by rapid loss of blood from a heavy arrow or spear wound. Poison, in other words, is crucial to both the technological and the expressive dimensions of recent Kalahari and late precolonial gender relations.
These attempts illustrate the chronological limitations of currently known and ethnographically documented gender roles even as they urge us to envisage and articulate terminal Pleistocene arrangements quite alien to recent Kalahari practice. But they do suggest that there is some connection between the technological, social, and political spheres that can be approached cautiously with archaeological evidence such as stone tools and faunal remains. If the more distant past is to be engendered, archaeologists need to be innovative and imaginative in the ways they relate arguments to evidence. Women may still have made beads and collected plants in the terminal Pleistocene but “what it was like to have been an old woman” may not have resembled the experiences of a recent San grandmother.
“Knowing the past” may be an ambitious goal, in that the closure of an argument through controlled experiment and the ability to focus on one variable at a time are not strictly possible. Archaeological knowledge is a form of triangulation from the territory of the relatively well known into the frontier of the relatively poorly known. Unlikely scenarios can be identified as such, but competitive explanations are likely to coexist for long periods and critical evidence is scarce. The aim is to link ethnographic and historic eyewitness accounts with empirical observations from archaeological excavations, viewing all such evidence as capta (things taken) rather than data (things given), thus recognizing the active agency of the archaeologist (Chippindale 1991).
In recent times Southern African hunter-gatherers have constructed a system of belief, n!ao, that mediates sexual relations by grafting some culturally generated rules onto some biological facts. Whereas childbirth is the realm of women, the hunting of large game is constructed as the realm of men. In an effort to underline the significance of both and to firm up this division of real and symbolic labor, women are likened to the large game animals such as eland, linguistic usages blur the difference between enjoyment of the two “meats,” and women are culturally defined as bringing bad luck to the hunt. Precisely when women are capable of conceiving, they are forbidden to touch the arrows of hunters, husbands are encouraged to focus on one or other wife (meat) at a time, and everyone makes as much, if not more, fuss over the death of an eland as over the birth of a child. A man’s identity is so intensely invested in his hunting ability that he would never consider leaving camp without his bow and a quiver full of poisoned arrows.
The success of this system may be judged from the comments of several anthropologists about the relative egalitarian nature of San society, where women often rise to positions of leadership, are identified as the “owners” of n!ores (territories), and have substantial influence on social issues. In the issues of violence, contacts with external groups, and dominance on ritual occasions such as the public healing dance, men retain far greater power and influence. The basis of male–female relations, however, is clearly the notion of complementarity, whereby men and women have notional equality by performing parts of various tasks with socially constructed status, respect, or value. Competition between men and women is diffused by humorous exchanges from separately defined positions.
Tracking this system back into undocumented prehistory is a great challenge. The Holocene archaeological record generally supports an economic division of labor, but not unambiguously so. Technological continuities such as the making of ostrich eggshell beads; wooden digging sticks with perforated stone weights; reed, bone, and sinew arrows; ochreous paint; vegetable mastics; and fibrous string and cord are suggestive of behavioral continuities, but no more. Burials so far have proven relatively mute on issues of gender, though there are hints that changes in food consumption patterns may be accessible. Stone tools have no documented pedigree and thus can only with imagination be used to construct gender profiles. My assessment of the material from Elands Bay Cave (Parkington in press) is that there are substantial changes in settlement at almost exactly 9,000 years ago. At least the coastal component of settlement changes from relatively long periods of cave use, perhaps combined with longer moves between a smaller set of residential locations, to much shorter visits with many more sites used in a seasonal round. My sense is that this change is a barrier through which it would be dangerous to push ethnographically or historically derived patterns of behavior. It is quite likely that terminal Pleistocene hunter-gatherers related to one another in ways that do not easily fit into Kalahari models.
It is the rock art that has the greatest potential to extend back in time our understanding of gender and sexual relations. The choice of subject matter, the conventional emphasis on the male hunter, the large game animals, the sexual attributes of adult men and women, and the absence or rarity of copulation or kill scenes all imply a reference to the roles of and rules for adults rather than to the depiction of their activities. The absence of paintings of plants, shellfish, or other small game that are more properly gathered than hunted is supportive of this position, but the large number of paintings of small bovids, not n!ao animals, suggests that other explanations have a place. It may be prudent to admit that we will probably never have a satisfactory explanation for many paintings, either because the images are residual or because the meanings are not recoverable without textual parallels.
This does bring up the issue of the dating of the paintings, without which we are incapable of constructing a history of image choice, convention, or emphasis. Despite cries of chronocentrism (Lewis-Williams 1993), we certainly suffer from not knowing the age of the paintings we study. It is quite likely, for example, that the paintings we see in the Western Cape are much older than those we see in the Drakensberg and also that many of the anecdotally noted differences relate to differences in social context and both external and internal social relations. Women are notably rarer in the Drakensberg than in the Western Cape (Lewis-Williams 1981; Manhire 1981), while therianthropes are arguably more common. It is very unlikely that any of what we see in either region was painted in the terminal Pleistocene, which means that any help in reconstructing gender relations that we can get from the rock art is inapplicable to these earlier times.