You are here

Fiona Marshall & Elisabeth Hildebrand
Cattle Before Crops: The Beginnings of Food Production in Africa

Journal of World Prehistory, Vol. 16, No. 2, June 2002. pp. 99-143

The importance of day-to-day predictability and scheduled consumption

Day-to-day predictability and scheduled consumption. Prehistoric farmingSince Lee's research of the 1960s, which suggested that hunter-gatherers often work less than farmers (Lee, 1968; but see Hawkes and O'Connell, 1981), archaeologists have emphasized the productivity of farming (it can support many people per unit area) as the major advantage of agriculture over hunting and gathering. Current theories for agricultural origins focus on population growth, environmental deterioration, and social or economic gain (reviews in Cowan and Watson, 1992; Price and Gebauer, 1995); despite diverging specific emphases, they share the premise that domestication occurred because humans wanted larger quantities of certain resources. A number of scholars have emphasized risk and the need for a predictable food supply (Flannery, 1986; Redding, 1988; Smith, 1998; Wills, 1995; Winterhalder and Goland, 1997). But the more common view, well expressed in current textbooks, is that understanding the reason for increased food yields is key to understanding the origins of agriculture (Crabtree and Campana, 2001, p. 242). Surprisingly little research has focused on individual human motivations during early stages of the domestication process, however. The end results of agriculture—visible today as larger yields, higher carrying capacity, denser stands of crops, larger seeds and seed heads, or greater animal productivity—were not necessarily realized during the earliest phases of the domestication process. Human motivations for initial manipulation of plants and animals thus remain an open question.

Historic, anthropological, and ethnobotanical accounts of tending and transplanting wild plants by hunter-gatherers are potentially valuable sources of information on this issue (e.g. Steward, 1938; Turner and Kuhnlein, 1982; reviews in Harris and Hillman, 1989; Lourandros, 1997). In many cases, however, domestication was not the focus of research, and specific reasons for intensification are unclear (Hayden, 1990, versus Keeley, 1995). Ethnoarchaeological research seems the most promising source of new data on processes of domestication. Studies have identified modes of selection that lead to domestication, such as the use of sickles to harvest cereals (Hillman and Davies, 1990) or breeding only the most tamable individual animals (Trut, 1999). Researchers recognize that planting selected plants and captive breeding of desired animals are essential to maintaining selection and to the domestication process. Yet, ethnoarchaeological research has not focused on the precise contexts in which people living in small-scale societies find it advantageous to capture and breed plants or animals.

In this section, we use ethnoarchaeological data to address the question of why, during the domestication process, people perform the critical steps of sowing, relocating, or capturing plants and animals from the wild. We follow Harris's conceptual framework for degrees of intensification in interactions between humans and plants and animals (Harris, 1989, 1996b; see also Hillman and Davies, 1990). Our ethnoarchaeological studies focus on motives for domestication among hunter-gatherers and extensive farmers currently adopting wild resources into domestic contexts. Marshall (2001) lived among Okiek, hunter-gatherers of Kenya during 1989–90, many of whom were farming at the time. Hildebrand (2001) studied shifting cultivators of forests and grasslands of southwest Ethiopia; she lived with the northeast Sheko during 1998–2000, and collected data on local crops and land use in nearby Gura Ferda and northern Dizi areas. In both studies, we found that people want to be able to schedule use of certain resources, and so manipulate plants and animals to ensure predictable access.

We think that the motivations we observed for the capture and breeding of plants and animals are relevant (sensuWylie, 1985) to those of any group of people contemplating the initial manipulation of wild resources. Knowledge of horticulture does not negate the relevance of these decisions, and presentday motivations for plant manipulation can inform considerations of early herding. This is true because, although decisions to capture or manipulate plants and animals may lead to domestication or adoption of domesticates, domestication is not necessarily the goal envisaged at the time of initial manipulation. Small, discrete steps in the domestication process reflect more general decisions about resource intensification. These arise in particular contexts (time, place, and social setting) of resource use, and are made by hunter-gatherers, horticulturists, and pastoralists alike. Understanding such contexts is key to understanding the reasons for capture and breeding of plants and animals by ancient people.

The Case Studies

Sheko and northern Dizi farmers who harvest yams (Dioscorea cayenensis Lam. complex) in wild areas often leave them to regrow, but transplant them to home gardens in contexts where the need for yams is unusually great, or yams are more difficult to find. In northeast Sheko, spontaneous yams are mostly transplanted by bachelors who value them as an easily prepared food. People with predictable access to food cooked by female kin transplant spontaneous yams much less frequently. In nearby northern Dizi areas, families historically transplanted yams from sparsely populated lowlands, where spontaneous yams are abundant, to denser upland settlements, where they are rare.Yams growing spontaneously near upland settlements are left in place. Northern Dizi yam transplantation thus focuses on making resources distant from settlements more accessible. Adoptive transplantation is almost never practiced among the people of Gura Ferda, a third subgroup of the Sheko–Dizi ethnic entity. Demands for increased predictability at Gura Ferda are low because spontaneous yams are abundant and easy to find; furthermore, residential mobility in Gura Ferda is much higher in terms of both frequency and distance of moves, so that transplantation would not make access to yams much easier or more predictable in the long term. These examples show that the need for predictable access to certain foods can motivate people to capture and propagate plants; this is especially true for individuals who have an exceptional need for the plant, live in locations where the desired plant is relatively scarce, or invest in a long-term occupation site. Moving either a plant or an animal near a house guarantees predictable access, so that people can monitor the resource and schedule its consumption.

Piik ap Oom Okiek also manipulate the distribution of wild plants, and plant some of them in domestic settings. Okiek families eat wild and weedy greens every day, but it is not the most desired greens that are sown or transplanted. Instead, effort is put into manipulating only those plants that are uncommon or not conveniently located (Marshall, 2001). Basella, for example, is an edible wild climbing shrub most common in steep ravines. Families often transplant one or two plants to the doorways of their homes. Similarly, Gynandropsis is a weed of disturbed ground, which families may sow if it does not spontaneously reseed near houses. Some less mobile Piik ap Oom families now also transplant a tree (Dracaena sp.), a creeper (Periploca sp.), and a high-altitude grass (Eleusine sp.) used in initiation and marriage ceremonies. All cases of planting and manipulation of these resources affect distribution more than abundance. Planting edible wild plants where they are accessible is a form of living storage. Through day-to-day monitoring of the depredations of pests, predators, and neighbors, individuals know the status of the plant and its availability for consumption or use. Concern is for predictable availability of plants eaten regularly or used at important ceremonies, not for increased yield. Manipulation is likely in contexts where families are staying in one place for a long period, where the distribution of desired plants is not convenient, or where ceremonies are about to take place.

The need for precise scheduling also shapes Okiek choices of animal species for different consumption events. Okiek modes of meat acquisition (trapping, hunting with spears, or use of domestic animals) differ in the predictability of their outcome. When Piik ap Oom families eat meat, it is usually wild (giant forest hog or bushbuck), but at ceremonies or gatherings, domestic sheep, goat, or cattle are used exclusively. Domestic animals provide no more meat than wild animals, and have no ritual significance. In fact, the “wild” is especially important in Okiek ceremonies (Kratz, 1994), but domestic stock is favored for ceremonies and gatherings because of its predictable availability.

Okiek use of cattle provides examples of contexts in which close control over animals may be extremely desirable. Thus cattle, although domesticated, help us to understand potential circumstances in which wild animals are likely to have been domesticated prehistorically. Scheduled ceremonies create a clear need for easily and predictably available meat, which is best supplied by penned animals. The relation between scheduled consumption events, an increased need for predictablity, and capture of wild animals is supported by recently reported practices of the Conibo-Shipibo of the Peruvian Amazon, who capture, pen, and feed manatees, peccaries, and monkeys in preparation for feasts (DeBoer, 2001).

Despite great differences in socioeconomic organization and environment among these ethnoarchaeological cases, they reveal similar contexts and motivations for manipulating resources. People often manipulate plants and animals in situations where the need for predictable access or scheduled consumption is especially great. We argue that these needs shaped prehistoric decisions to manipulate resources in ways that led to domestication. Although higher yield is a popularly cited advantage of food production and does affect later phases of agricultural innovation, the need for predictable access may have been a more important catalyst in many domestication events.

The Model

Intensification and domestication are structurally linked to the need for predictable day-to-day access and scheduled consumption. Understanding this connection makes it possible to move beyond the particulars of our studies to model a broader range of contexts in which predictable access to specific resources can become an especially prominent concern.We begin by identifying contexts that make predictable access to resources desirable, and go on to examine the relationship between these contexts and the large-scale social, economic, and ecological processes that create them. This approach fits the large anthropological literature on the risk of going without food in spatially and temporally variable environments (Cashdan, 1990; Halstead and O'Shea, 1989;Kelly, 1995). It also echoes the archaeological literature on risk and domestication cited earlier, but differs in scale and in emphasis.We focus on day-to-day, rather than seasonal, uncertainty about the availability of key resources. We are concerned more with the initial risk and social cost of not having what you want when you want it, than the subsequent danger of starvation.

Day-to-day predictability of wild resources may become a major concern in settings where sedentism is increasing, plant abundance is decreasing, or prey mobility is increasing. Increasing sedentism can make distant resources difficult to collect, so that access becomes unpredictable. Transplanting, sowing, or otherwise moving the resource to the locus of human occupation restores the compromised predictability of that resource. Decreased abundance of certain plants, or increased mobility of prey can also bring about a crisis in predictability. In both situations, a formerly dependable resource becomes difficult to rely on, and humans may respond by focusing more on other resources (if available), or by finding ways to better manage either the resource itself or their access to it. In the case of plants, this may mean transplanting or sowing the plant in an especially favorable location where it can be weeded or tended. In the case of animals, it may mean allocating a certain portion of the human group to follow or guide stock along optimal routes for grazing, to find water, and to protect the herd from predators or competing human groups.

These proximate factors (changes in sedentism, plant abundance, or prey mobility) may be caused by a number of broader social or ecological conditions. Sedentism is often thought to arise where key resources are abundant and concentrated. Decreased or patchy distribution of plants may be due to natural ecological differences, environmental degradation, overexploitation, or novel patterns of predation or disease. Prey mobility can increase as rainfall becomes lower and more variable, and the distribution of plants becomes more spatially and temporally stochastic. Any of these highly varied circumstances may raise search costs for key resources, so that knowledge about the location or state of a resource becomes less certain. One potential response to this crisis in predictable access is to manipulate the distribution of the resource to satisfy human needs for day-to-day or ceremonial events. If people knowingly or inadvertently cause morphological selection in tandem with manipulation of plant or animal distributions, then domestication may result.

Domestication is linked to broad social or ecological processes, proximate contexts in which predictability becomes crucial, and manipulation practices meant to restore or enhance predictable access and scheduled consumption. We think the utility of this scheduled consumption–predictability model for domestication is broad: it can be applied to plants or animals in pristine settings of domestication, and to circumstances in which domesticates were adopted from elsewhere. In Africa, this model can explain much about the domestication and spread of cattle, the patchy spread of food production, and the late domestication of plants.