Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, AFRICAN HISTORY ( (c) Oxford University Press USA, 2020. All Rights Reserved. Personal use only; commercial use is strictly prohibited( for details see Privacy Policy and Legal Notice).

date: 28 October 2020

Interdisciplinary Perspectives on Precolonial Sub-Saharan African Farming and Herding Communitiesfree

  • Paul LanePaul LaneDepartment of Archaeology and Ancient History, Uppsala University
  •  and Anna ShoemakerAnna ShoemakerDepartment of Archaeology and Ancient History, Uppsala University


Agricultural practices on the African continent are exceptionally diverse and have deep histories spanning at least eight millennia. Over time, farmers and herders have independently domesticated different food crops and a more limited range of animals, and have effectively modified numerous ecological niches to better suit their needs. They have also adopted “exotic” species from other parts of the globe, nurturing these to produce new cross-breeds and varieties better adapted to African conditions. Evidence for the origins of these different approaches to food production and their subsequent entanglement is attested by diverse sources. These include archaeological remains, bio- and geo-archaeological signatures, genetic data, historical linguistics, and processes of landscape domestication.


  • African History
  • African History

Researching Precolonial Food Production

Research on the origins, spread, and evolution of food-production systems across sub-Saharan Africa has intensified in recent decades. Advances in archaeological science, population genetics, palaeo-ecological reconstructions, and historical linguistics, coupled with fresh discoveries, changing archaeological excavation and sampling strategies, and theoretical developments, have all contributed to this upsurge in cross-disciplinary interest. Despite these positive trends, several of which are reviewed below, the number of scholars engaged in such research, particularly among those physically based on the continent, remains low relative to most other regions, especially Europe and North America. One obvious consequence of this is that data sets and their temporal and geographical coverage remain fragmented, with some areas still very much archaeological terra incognita. These observations should be kept firmly in mind, as it is quite probable that current scenarios may change in the near future in the light of new data and innovative analytical approaches.

These caveats notwithstanding, sub-Saharan Africa has much to offer researchers interested in the origins and spread of food production, not least because of the potential to incorporate multiple strands of evidence (figure 1), typically material, biological, linguistic, and genetic, but also for certain time frames and research questions, oral historical, and palaeoenvironmental records as well. There is a long tradition of such inter-disciplinary studies, and a corresponding routine awareness among researchers of advances in areas outside their own particular disciplinary specialism and a readiness to consider alternative perspectives.

We discuss this range of evidence and approaches in the following sections, drawing examples principally from areas south of the contemporary climatic and ecological boundaries between the Sahara and Africa’s Sudanic/Sahelian zones. However, because the location of this boundary has shifted over the millennia as global climates have changed, we periodically draw on examples from sites and landscape that today lie north of this boundary, especially where these places retain key information about the timing, actors, and processes involved in the transition to and adoption of food-producing strategies that subsequently spread to other parts of the continent.

Figures 1a, 1b, 1c, 1d, 1e, and 1f show examples of different types of material archaeological evidence documenting precolonial food-production strategies in Africa. Figure 1a shows an abandoned grinding stone from a Late Iron Age Moloko-period settlement, near Ranaka, SE Botswana.

Photo by Paul Lane, 2014.

Figure 1b. Faunal remains recovered from an excavated context at an 18th- and 19th-century caravan trade halt, Ngombezi, NW Tanzania.

Photo by Paul Lane 2009.

Figure 1c. Abandoned 15th–16th-century pre- or proto-Herero well, Otjozondema, NW Namibia.

Photo by Karl-Johan Lindholm, 2004.

Figure 1d. Stone footings of a possible grain store, Renjuk, Central Equatoria, South Sudan.

Photo: Paul Lane, 2009.

Figure 1e. Iron axe blade, from a Late Iron Age Moloko-period settlement, near Ranaka, SE Botswana.

Photo by Paul Lane, 2014.

Figure 1f. Possible human management indicators that together may be linked with farming and grazing: (a) Cereal pollen grain from domesticated rice (Oryza sativa type), (b) dung-loving fungal spores from Sordariaceae (right) and Coniochaeta lignaria (upper right), (c) microfossil charcoal. Pictures from the Ilaika Marais sequence from Antananarivo, Madagscar. Photographed in 1000x magnification and using oil immersion.

Photos by Anneli Ekblom, 2016.

One striking aspect regarding trajectories of food production in Africa is that south of the Sahara, evidence for animal domesticates predates that for plant domesticates.1 This contrasts with the sequence documented in most other parts of the world, where, typically, crop domestication preceded livestock domestication. Plant domesticates also seem to emerge at a much later date across most of sub-Saharan Africa than on any other continent. One suggested explanation for this anomaly cites the abundance of intermittently distributed wild plant and animal resources that are characteristic of savanna environments and capable of sustaining mobile pastoral or hunter-gatherer-fisher populations.2 Along this line of thinking, the success of mobile livelihoods and the opportunistic procurement of wild resources from African savannahs and forests may have rendered the transition to sedentary, larger-scale, crop-based farming unappealing. However, the actual means by which early hunter-gatherer-fishers and herders engaged with “wild” plant resources, and so may ultimately have domesticated certain species, is largely unknown. It also needs to be acknowledged that in some parts of sub-Saharan Africa, despite clear evidence for the intensification of exploitation of wild resources and efforts to stimulate and manage productivity, as documented, for example in southern Africa, no formal morphological domestication of either plant or animal species occurred.3 The eland, for instance, is a relatively docile herd animal that can be easily driven. Faunal assemblages indicate that eland were widely exploited in southern Africa, and were certainly of symbolic importance, as attested in the region’s rock art (and in the mythology and traditions of contemporary and historical San populations), yet it has never been domesticated. As Peter Mitchell has noted, several of the other commonly exploited species including zebra and Cape buffalo were probably too aggressive to even try to tame. In terms of plant species, Mitchell also suggests that “the long maturation periods and possibly high processing costs of the nut- and fruit-bearing trees of the Savanna and Miombo Biomes,” may have been one of several factors that deterred eventual domestication.4 More research is required to address such questions about the processes that led to the adoption of certain plant and animal domesticates over time in Africa.

A further point of note is that across Africa livestock herding and crop cultivation developed along unique and quite variable trajectories, typically resulting in mosaic landscapes in which different forms of crop cultivation and livestock management coexisted, while often also retaining socio-ecological niches for hunting/gathering/fishing communities.5 Understanding the transition to food production and the spread of domesticates and their evolution thus requires as close an engagement with the coeval hunter-gatherer-fisher archaeological record as with that of the early herders and farmers themselves.

Other emerging themes in the study of the archaeology of African farming and herding traditions include the processes, causes, and consequences of agricultural intensification,6 landscape domestication and the ecological footprints of different food-production systems7 the early adoption of African domesticates such as pearl millet, finger millet, sorghum, and cowpea in the Indian subcontinent and the dating and circumstances of the introduction of Southeast Asian domesticates such as rice, banana, and chicken into Africa,8 and the later integration of American plants into African farming traditions.9 All of these have the potential to inform broader discussions about future food security on the continent,10 and the role of archaeological evidence in meeting the United Nations’ Agenda 2030 Sustainable Development Goals.

Defining Domestication, Farming, and Herding

There is no one standard definition for domestication, though it may be broadly described as a set of processes that give rise to genetic, morphological, and (in the case of animals) behavioral changes in certain plant and animal species as a consequence of selective pressures arising from human intervention.11 In Southwest Asia, where much archaeobotanical research was pioneered, the presence of morphologically “domesticated” plants (usually cereals and pulses) has long been taken to indicate agriculture.12 For cereal crops and legumes, “domestication traits” include an increase in grain size and yield and a reduced ability to disperse seeds without human intervention.13 Other morphological changes induced by domestication in noncereal crops may include a decrease in physical defenses, for instance the loss of spines in yams (Dioscorea).14 Identifying farming exclusively from the presence of domesticates is inappropriate in African contexts, however, where perennial root crops or fruit trees may be propagated through vegetative cloning and where species may exhibit only subtle genetic and morphological change despite having been cultivated for millennia. Classic examples include yams (Dioscorea cayenensis and related species),15 and oil palm (Elaeis guineensis),16 which can be intensively managed and owned without achieving “full morphogenetic domestication.” Furthermore, evidence from archaeologists, geneticists, and evolutionary biologists indicates that humans and plants may be engaged in protracted and complex relationships of domestication for centuries, even millennia, before morphogenetic changes occur.

The term “farmer” is thus used here to refer not only to people who cultivate morphogenetically domesticated plants, but also to people who pursue ongoing cooperative relationships with certain plants, having a significant impact on the care and reproduction of particular species to the advantage of the farmer as well as the plant. The actual plant resources being farmed may undergo dramatic morphogenetic changes to adapt to new agro-ecosystems resulting in a rich diversity of domesticated land races and cultivars, or species may exhibit continuity, producing ideal yields in varied environments without ever becoming domesticates in a morphological sense.17

Livestock domestication has often been defined as the deliberate taming and capture of particular individuals of selected animal species that exhibit distinct behavioral characteristics, which in turn make them more amenable to interactions with humans, and the removal of these individuals from their wild environment for the purposes of controlled breeding for mutual benefits.18 Over the long term, human intervention in animal breeding and mobility results in morphological and genetic changes that can be used to differentiate domesticated species from their wild progenitors. In contrast to plant domestication, behavioral modifications were likely more important than genetic changes, at least in the early stages of domestication. The emphasis on changing animal behavior also has a tendency to generate more overt material traces of these early stages of domestication compared with plant domestication, as most material markers of crop domestication typically only occur after reliance on cultivated plants is well established.19

Additionally, recent approaches to animal domestication acknowledge the possibility of agency on the part of the animal species involved, and that they rather than humans may have been more instrumental in initiating a change in how humans interact with them as individuals. This is perhaps most evident in the processes by which wolves became domestic dogs, although the concept can be extended to other species as well. There has also been a movement away from hard-and-fast characterizations of “wild” and “domesticated” as mutually exclusive states,20 and their replacement with models that acknowledge that these states can be manifest to different degrees in different species—such as in the contrast that can be drawn between domestic beef cattle, on the one hand, and domestic reindeer, on the other.

Methods for Understanding the Archaeology of African Farming Communities


Archaeobotanists study the remains of plants from archaeological contexts to understand the nature of people–plant relationships in the past. Plants may be recovered from the archaeological record in multiple forms. Macro-remains of plant parts such as seeds, fruits, and woody tissue are most often preserved through burning (carbonization) and sometimes by desiccation, waterlogging, or mineralization. Proxy evidence for plants may also be preserved in the form of impressions in pottery fabrics21 and other earthen materials such as mudbrick22 and even in iron-smelting slag.23 Plant micro-remains such as phytoliths (silicified plant cells), pollen, lipids (oily organic compounds insoluble in water, such as fatty acids and waxes), and starch grains are also recoverable from archaeological contexts.

Archaeobotanical analysis entails the identification of plant remains to the highest taxonomic level possible, and recording the plant parts represented. Analysis of archaeological plant remains is dependent on access to comprehensive collections of modern botanical reference material. The relative abundance of particular plant remains in different archaeological contexts may be compared to investigate how, when, and where different wild or farmed resources were produced, procured, distributed, stored, processed, and utilized, and the impacts of these activities on individuals, societies, and landscapes.24 Archaeobotanical studies may further address questions about why plant resources were consumed, whether they were valued for instance as sources of fuel, food, or medicine, or perhaps for their symbolic or economic significance. The identification of species with known ecological preferences in archaeological contexts can also provide insights on environmental conditions in the past. Not all plants are preserved, recovered, and identified in equal proportion, however, creating a bias against the “missing foods” of the archaeological record.25

Figure 2. Samuel Ntimama and Ntisipa Tinayo share their knowledge on the ecological properties and local use traditions of wild edible fruits in Amboseli, Kenya. Pictured here is a Grewia tenax plant (eirri in Kisongo KiMaa).

Photo by Anna Shoemaker, 2015.

Ethnobotany, the study of human–plant relationships, is of critical importance in the interpretation of archaeobotanical assemblages (Figure 2). Engaging with local botanical knowledge, oral historical accounts, and archival sources enables a more nuanced and often revelatory perspective of African farming traditions.26 Experimental archaeology, wherein historic or prehistoric techniques for agricultural harvesting, storing, and processing are replicated in controlled environments to test hypotheses about methods of cultivation, has also yielded insights.27


Zooarchaeologists study animal (which here includes birds, reptiles, fish, and shellfish) remains and traces of their byproducts (dung, milk, blood, etc.) recovered from archaeological contexts to understand the different dimensions of past human–animal relationships and their environments.28 In most settings only bones and other hard tissues such as horn, shell, and ivory survive for archaeological recovery and analysis. These are often referred to as “faunal remains,” and in certain contexts, especially where soils are fairly acidic, preservation may be poor to nonexistent. In other contexts, such as exceptionally arid conditions or waterlogged deposits, traces of soft tissues such as hide, wool, and hair may also survive. A wide range of other pre- and post-depositional taphonomic processes,29 as well as sampling and recovery strategies, typically influence the composition of faunal assemblages and so introduce biases that need careful consideration during different stages of analysis. Primary and secondary butchery practices, and consumption and discard patterns, are well-known examples, and ethnoarchaeological and experimental research on such topics among modern-day communities have played a significant role in raising awareness of such issues and formulating analytical approaches for addressing them.30

Primary concerns include determining the range of animal species exploited in the past by different communities, reconstructing their patterns of animal food procurement, and subsistence activities; past hunting and herd-management practices; the exploitation and distribution of different animal products (meat, milk, wool, bone, etc.); determining the origins and processes of animal domestications and the relative balance of “wild” versus “domestic” animal foods in human diets; the social and symbolic roles of animals in past societies; and the value of animal remains as possible indicators of past environmental conditions.31

Key components of any zoo-archaeological study include taxonomic identification; documentation of the different skeletal parts, age and sex of individual animals represented in an assemblage, and construction of age and sex profiles by species; documentation of the degree of fragmentation and signs of human modification (e.g., cut marks from different butchery stages, burning, splitting for marrow extraction) and their position on individual bones; signs of post-depositional damage (such as gnawing by dogs and other scavengers, weathering due to exposure, damage from trampling); and reconstruction of the relative proportions of different species represented. The latter is typically addressed through calculation of Minimum Number of Individuals (MNI), Minimum Number of Elements (MNE), Minimal Animal Units (MAU), and/or Numbers of Identifiable Specimens (NISP) values.32

Quantification of these data facilitates inferences regarding kill-off patterns, the relative contributions to diets of different livestock species, herd-management systems, the possible use of livestock for “secondary products” (e.g., for milk, wool, or traction), the distribution of different body parts to different individuals, and the relative importance of domestic versus wild animals at a site, among other topics.33 Further contextual analysis of the distribution and associations of faunal and other material at an individual site can also yield important insights into the cultural significance of livestock (and other animals), especially when coupled with ethnoarchaeological observation.34 Common examples of this include evidence for differential distribution of body parts/cuts of meat according to age, gender, and/or social rank, and material evidence (such as from burial and/or rock art) for the symbolic and ritual values accorded to particular domestic species. On a seemingly more mundane level, but often of significant cultural importance, evidence may be recovered for the economic and cultural importance of domestic species, such as donkey, used for purposes other than as a food source.35 Additionally, faunal remains may be examined macroscopically, via radiography and/or chemical analyses for stress- or disease-related pathologies. For example, the prevalence of enamel hypoplasia, that is, interruptions or disturbances in tooth-enamel formation, can be indicative of climate- and/or environment-induced stress,36 while evidence for osteoporosis may be a sign of long-term protein or nitrogen deficiencies related to poor grazing conditions.37 Analysis of microscopic wear traces on animal teeth may also provide some insights into the type of grazing routinely exploited, from which inferences about herd health and seasonal mobility patterns, potentially, can be drawn. Analysis of the species composition of microfauna, and especially domestic rodents (mice, voles, rats, etc.) recovered from archaeological contexts, can also provide information about the local environment, settlement mobility, and food storage strategies, among other matters.38

DNA Studies

The study of modern and ancient plant and animal DNA has generated considerable insight into farming and herding practices, and has made major contributions to our understanding of the earliest appearance of domesticates, their movement through time and space, and their genetic diversity.39 DNA analysis is being used to reconstruct evolutionary histories and identify the wild progenitors of morphogenetically domesticated crops and animals. Knowing the progenitor species of domesticates helps isolate the regions and environments in which people could have first exploited them, and their subsequent trajectories of dispersal.40 For example, the identification of tropical West or northeastern Africa as likely areas of domestication for cowpea, a legume with a widely distributed wild progenitor, has been made possible with DNA research.41 Research into the genetics of contemporary staple crops and innovations in ancient plant DNA analysis has also led to the identification of molecular markers of domestication, again having applications for studying the origins and spread of agricultural species. The pre-Columbian dispersal of the African bottle gourd (Lagenaria siceraria) to Asia and the Americas is one such example.42

Farming is a practice of nurturing and selecting plant resources across time and space, driving responses in plants that differ in part according to varied techniques of human manipulation. DNA studies can offer insights on the geographic, ecological, and cultural processes that have maintained and enhanced the genetic structures of African crop populations. For example, researchers have recently demonstrated a relationship between ethnolinguistic groups and sorghum genetic diversity in Africa, suggesting that sociocultural and ecological factors have been involved in generating and continuing crop landraces on the continent.43 There is a real need for such studies that build understandings of the distribution, extent, and maintenance of genetic diversity in African plant species as landraces most in need of conservation are identified.44 The value of aDNA studies is further underscored as lost genetic diversity information can be recovered from plant remains preserved in archaeological contexts.

The development of modern DNA amplification and sequencing techniques, particularly high-through-put next-generation-sequencing (NGS), has also had a significant impact on the understanding of domestic livestock variability across the continent. For instance, early research raised the possibility of an independent center of cattle domestication in North Africa, although more recent work suggests that hybridization of European taurine cattle via cross-breeding with local wild cattle better explains the DNA configuration.45 Other DNA research has facilitated identification of the geographic origins and dispersal routes across the continent of a range of domestic species from sheep and goats to chickens, along with possible evolutionary bottlenecks and introgressions (consistent hybridizations) between related species.46 Studies of Y chromosome, autosomal, and mitochondrial DNA (mtDNA) in African cattle, for example, all suggest that the process of introgression of zebu cattle (Bos indicus), which were first domesticated in Asia, was primarily a male one, although the reasons for this remain unclear,47 while analyses of mtDNA have shown that chickens (originally domesticated in Southeast Asia) were introduced to the continent multiple times.48

The use of aDNA sequencing in the analysis of faunal assemblages from African sites is still in its infancy, although it clearly has potential, as suggested by characterization studies of sheep in southern Africa.49 It may prove to be an effective means of verifying taxonomic identifications. This can also be attempted by another recently developed method known as ZooMS, or Zooarchaeology by Mass Spectrometry, involving mass fingerprinting of peptides in faunal specimens. This is a nondestructive, high-throughput, biomolecular method of determining species based on the principle that the amino-acid sequences of different species result in distinctive peptide profiles.50 Thus far, however, ZooMS has only been used in Africa as a means to differentiate elephant and other ivories.51 The need for more sustained research is nonetheless highlighted by some of the current conflicting results between taxonomic and aDNA identifications.52 Additional challenges are posed by variable aDNA preservation in samples and the great diversity exhibited by many of the domesticated species on the continent. Similar challenges are associated with plant aDNA. Beyond issues of preservation and contamination, many crops are polyploidy (having two or more chromosomes) and exhibit considerably convoluted genetic heritage. Furthermore, millennia of interspecific hybridization and frequent gene flow between crops and conspecific wild or weedy populations can make it very difficult to distinguish domestication using DNA.53 The combination of genetic and directly dated archaeobotanical evidence is often necessary to understand past plant–human relations in agricultural societies.

The recent advances in human genetics are also transforming understanding of the demographic and migration histories of the African continent and, through correlation with archaeological and historical linguistic information, are advancing knowledge of the potential contributions of the speakers of different language families to the spread and consolidation of herding and faming. Detailed review of this rapidly changing field is outside the scope of this article, but see Calafell and Comas,54 and references therein, for recent overviews.

Proxy Indicators

It is increasingly recognized that there is no singular suite of material culture or “Neolithic package” indicative of the presence of early herders and farmers in Africa. Rather, the character of farming and herding communities across the continent has been highly individualized and must be evaluated as such. However, there are a host of proxy indicators that can inform on human–plant and human–animal relations in the past, indicators that are most powerful when combined with other strands of evidence.

Material Culture Facilitating Herding and Farming

Tools that may be used for agricultural pursuits such as hoes, digging stick weights, and sickles are items of material culture suggestive, though not irrefutably indicative, of farming. The diverse morphology of African agricultural tools is relatively unknown by archaeologists.55 There is scope for building understandings of the relationship between particular tool categories and agricultural practices, research that could illuminate regionally and temporally specific farming histories. For instance, in sub-Saharan West Africa around 5,000 years ago, an increase in small grinding-stone tools in conjunction with evidence for interregional trade, ceramics, and domesticates, and a decrease in residential mobility, is taken as an indication for intensification in Kintampo plant food use.56 In southern Africa, grinding-stone typologies have been used as proxy evidence for the processing of either sorghum and millet or maize during the Iron Age.57

Food-storage facilities encountered at archaeological sites can provide further insights into the significance of certain plant resources to food-producing societies in the past. In the southern Egyptian Western Desert, for example, pits containing abundant remains of wild sorghum dating to circa 10,000 bp, many millennia before the appearance of evidence for domesticated sorghum, indicate a considerably long history of cereal exploitation predating domestication.58 On Pemba Island, Tanzania, evidence for the storage of rice during the 11th century ce suggests not only the cultivation of this grain, but also a major socioeconomic shift entailing the specialized production of Asian crops, and the onset of a period of urbanization and Islamization.59

However, identifying and interpreting agricultural tools and food-storage systems is not always straightforward. Ethnographic studies of grinding-stone tools in Africa, for instance, reveal the essential multi-functionality of these items of material culture with the same grinding-stone tool forms frequently being used for more than one purpose (Figure 3)60. Similarly, it is problematic to assume that certain archaeological features were once used for the storage of particular plant species in the absence of supporting evidence.61

Figures 3a, 3b, and 3c illustrate varied applications for Marakwet (Kenya) grinding-stone tool kits. It is common for the same grinding-stone tool kit in Marakwet to perform all of these functions, and more. Figure 3a shows pounding maize kernels.

Photo by Anna Shoemaker, 2013.

Figure 3b. Marakwet (Kenya) grinding-stone tool kits: grinding finger millet.

Photo by Anna Shoemaker, 2013.

Figure 3c. Marakwet (Kenya) grinding-stone tool kits: processing ochre.

Photo by Anna Shoemaker, 2013.

Analysis of food residues and biomarkers (such as lipids and starch remains) along with use-wear analysis and other lines of evidence are increasingly being used to determine the function of vessels and tools recovered in archaeological contexts.62 Applications are not limited to understanding human utilization of plants. A recent study of ceramics from the Takarkori rock shelter, in the southwest Fezzan section of the Libyan Sahara,63 has demonstrated widespread use of pottery for storing and serving milk (or milk-based foods), indicating a reliance on livestock herding. These results broadly align with the inferences drawn from the faunal assemblages from this and other sites in the vicinity, the isotopic dietary signatures of contemporary human burials, and the extensive rock-art evidence for livestock herding. Recent overviews of ethnographic material from southern and eastern Africa have also emphasized the social and ritual importance of milk and milking in many herding societies, and their gendered nature.64 Unanticipated results may also emerge through the use of such techniques, however, as illustrated by the analysis of late 1st-millennium ce spouted wares from the site of Kasteelberg D East, Western Cape, South Africa. This indicated that these ceramics were used for marine-based foods rather than milk, as previously assumed on the basis of the faunal remains and vessel form.65

The material evidence for herding and farming practices is, however, not limited to portable items or storage features. Structural evidence, in the form of livestock pens, droveways, fields, and irrigation canals, all provide direct testimony of such practices, and may also be the primary sediment traps where plant and animal remains survive. A diverse range of forms revealed by archaeological excavation are known from across the continent, and where stone was used extensively in their construction these traces may extend across vast areas, many of which have been subject to detailed mapping and targeted excavation (figure 4).66 These include the extensive remains of settlements, livestock enclosures, pits, and terracing and associated landesque capital in the Nyanga Highlands, eastern Zimbabwe, covering at least 5,000 sq km and dated, minimally, to between around 1300 ce and 1800 ce.67 Other smaller, but no less impressive, fossil landscapes have been similarly documented, such as the complex of irrigation canals, field systems, and village sites at Engaruka, northwest Tanzania,68 that were in use at least between the 14th and mid- to late 18th century ce. Ethnoarchaeological studies of similar functioning systems, such as that employed by Kofyar communities in central Nigeria69 and those found in Konso, Ethiopa,70 and among Pokot and Marakwet communities along the Cherengani Escarpment in Kenya,71 have also provided valuable insights into the processes and drivers of intensification, landscape management, and formation processes and the social relations of their construction and management,72 among other topics.

Figures 4a, 4b, 4c, 4d, and 4e present examples of African archaeological field systems. Figure 4a shows abandoned agricultural terraces near Tokombéré, northern Cameroon.

Photo by Scott MacEachern, June 1986.

Figure 4b. Stone-line demarcated fields, Engaruka, northeast Tanzania.

Photo by Daryl Stump, October 2016.

Figure 4c. Irrigation canal, Engaruka.

Photo by Daryl Stump, October 2016.

Figure 4d. Stone boundary lines produced by field-clearance activities, Laki, central Equatoria, south Sudan.

Photo by Paul Lane, 2009.

Figure 4e. Terraced fields, Nyanga, Zimbabwe.

Photo by Mats Widgren, 1995.

Salt, an edible mineral and crucial ingredient for food preservation, is also entwined with the history of farming and herding societies in sub-Saharan Africa, as access to salt is important to societies that manage food surpluses73 or rely heavily on agricultural produce. Salt occurs naturally in animal products such meat, blood, urine, and milk in sufficient quantities to meet the average requirements for humans (c. 1 kg per person annually). Root crops, cereals, and most vegetables, on the other hand, are generally salt deficient. One consequence of the adoption of agriculture in Africa, as elsewhere, was an increased demand for salt as a dietary supplement, and a variety of different sources came to be exploited. Salt can be obtained through the evaporation of saline water, collected from naturally occurring salt crusts and brine springs, mined from rock salt deposits, or extracted from plant ash. In the Great Lakes region of East Africa, archaeologists have excavated the material remains of salt production dating to as early as the 6th century ce, including ceramic evaporating vessels and fire pits where saline solutions were reduced.74 Other examples of archaeological sites where perforated pots that could have been used in the domestic production of salt include the 5th-century ce site of Kapwirimbe in Zambia, the 8th-century ce site of Dakawa in east-central Tanzania, various sites around Lake Malawi including the 8th- to 11th-century ce sites at Namaso Bay, and at Baleni in South Africa, a natural brine spring exploited by early farming groups for roughly three hundred years from around 350 ce.75 We also know of the foundational role salt played in trans-Saharan trade throughout the 2nd millennium ce from historical sources and from the existence of salt-mining centers such as Teghaza, where the buildings themselves were constructed of rock salt.76 How the production and trade of salt have supported food-producing societies across the continent continues to be an area warranting further research.

Rock art, whether as paintings or engravings, can also be informative despite generic problems associated with their absolute dating. Although few, if any, images can be read literally, depictions of livestock, ploughing teams, agricultural implements, and cultivated crops clearly attest to their cultural significance in particular settings. The most extensive examples of this kind of imagery occur in the Sahara and northeast Africa, where large panels featuring domestic animals, especially cattle, pastoralist encampments, and routine activities (including milking) are common (figure 5a). The better-documented areas include the Tassili-Tadrart-Acacus Mountains of southeast Algeria/southwest Libya and nearby Messak Plateau (southwest Libya) in the central Sahara,77 and various parts of the Horn of Africa, including several densely decorated rock shelters at Laas Geel, northeast of Hargeisa, Somaliland (figure 5b).78

Figure 5a. Example of Saharan and northeast African pastoralist rock art, showing an engraved milking scene, Tikastin, Messak Plateau, Libya. Line drawing by Victoria Waldock, 2009.

Figure 5b. Example of Saharan and northeast African pastoralist rock art, showing a cattle herd, Laas Geel rock shelter, northern Somalia.

Photo by Jorge de Torres, 2015.

This art has been subject to multiple interpretations, and debate over the chronology of much of the art remains unresolved. Despite this, recent analyses do highlight the potential to infer various aspects of pastoralist practices and beliefs not accessible from other material traces from the art. Augustin Holl, for example, has suggested that variations in the form and patterning of the rock art in Tassili, Algeria, conveys information about the different life-cycle stages of the pastoralists who produced the art.79 Stefano Biagetti and Savino di Lernia have similarly used iconographic details (especially concerning herd size and the gender of the livestock depicted) and locational information at sites in the Acacus Mountains to infer transhumant strategies.80 Victoria Waldock’s recent doctoral study of Saharan pastoralist rock art (c. 7500–3000 bp) on the Messak Plateau, southwest Libya, greatly extends these earlier studies through a sophisticated, theoretically informed analysis of the movement of herders through these landscapes and the complex relationships with their livestock and ontologies of paths and places that emerged.81 Rock art from other parts of the continent similarly reveals information about food production, including sites in southern Africa where the occurrence of images of fat-tailed sheep, cattle, and horses in different contexts has been used to infer some of the sociocultural and economic consequences of their introduction.82

Isotopic Analysis

Isotopic analysis is another field of research that can be used to understand many things about people in the past, including their food. The ratio of naturally occurring stable isotopes (δ13C, δ15N, δ18O/δ‎16O and 87Sr, 88Sr) in dietary resources can vary predictably between environments, making the isotopic ratios found in individuals an expression of factors such as diet, climate, altitude, geological and soil composition, and the longevity of residence of an individual in a particular place.83 Stable isotope analysis in Africa has generated insights on the diets of Iron Age agriculturalists,84seasonal crop-rotation strategies,85 the antiquity of cattle-based pastoralism,86 livestock-breeding management,87 and the mobility of people and their animals.88 Stable isotope analysis on archaeological materials is also contributing to high-resolution palaeoenvironmental reconstructions in Africa.89

Over the past several decades, there has also been a growing awareness of the complexities of how isotopes are distributed in ecosystems and food webs and subsequently incorporated and preserved in the tissues of different plants and animals.90 Research questions must be designed with a thorough understanding of the strengths and limitations of isotopic analysis, and more robust and precise sampling of contemporaneous plant and animal tissues than currently exists is required to better interpret African “isoscapes.”91 Stable isotope research in Africa has a distinguished record, however, and African landscapes are well suited to experimental studies that will improve understandings on factors impacting isotopic variation.92

Linguistic Evidence

In Africa, research on the history of languages, and the systemic ways in which words change, has long been applied to the study of precolonial farming communities.93 Verbs related to tending of plant foods as well as names of plant foods that appear in human populations removed from areas of “natural” plant distribution are of particular interest for building models of the origins and spread of agriculture.94 Comparative analysis of technical terms for different physical features associated with farming and/or herding activities can also be informative regarding the origins and mechanisms of their introduction and adoption, as for example has been attempted for the Tangale-Waja Uplands, northern Nigeria.95 Detailed comparative analysis of terms for different food-processing and -preparation activities among different Bantu-language speakers has been even more informative. Using a “words-and-things” approach, Birgit Ricquier, for example, has tracked the changing linguistic history of the terms for “porridge” in Proto-Bantu and early Bantu languages (figure 6), their associations with different material items such as mortars and cooking pots, and the likely routes and mechanisms through which these terms, and the practices and technologies they imply, spread southward and eastward from the Proto-Bantu homeland.96

Figure 6. Map showing distribution of terms for “mashed” and “flour” porridge in Western and Eastern Bantu languages.

Source: Birgit Ricquier, “The History of Porridge in Bantuphone Africa, with Words as Main Ingredients,” Afriques: Débats, methods et terrains d’histoire 5 (2014): 2, map 1.

While datasets on African language histories are immense and not as well studied as those for Indo-European or Austronesian languages,97 modelling the expansion of Bantu languages has been an area of particular focus. In many places in Africa, the appearance of early farming had been hypothesized to be the result of the arrival of Bantu-speaking agriculturalists possessing sorghum, pearl millet, and knowledge of iron working.98 Straightforward presentations of the “Bantu agricultural package” dispersal have since been contested by linguists, archaeologists, and geneticists alike,99 while the archaeobotanical record, by and large, is too underdeveloped to meaningfully engage with such lines of evidence.

As Koen Bostoen100 has noted recently with reference to the food-crop lexicons of Benue-Congo speakers of west central Africa, the only crops with indisputable Proto-Bantu linguistic origins are yams and possibly cowpea (Vigna unguiculata) and Bambara groundnut (Vigna subterranean), which raises the question as to whether early Bantu-language speakers might be better characterized as “forest cultivators,” rather than farmers in the more traditional sense of the word.101 The likely long history of both forest and savannah manipulation prior to, and after, the inception of food production also highlights the scope for using linguistic studies to clarify the history of plant resources that are not morpho-genetically domesticated, or tend not to preserve in the archaeological record. This remains challenging, however. For instance, Roger Blench102 attempted to reconstruct the origins and subsequent anthropogenic distribution of the baobab tree (Adansonia digitata) using linguistic evidence, postulating an eastern-to-western Africa dispersal. However, genetic research indicates that the most primitive haplotypes for this species are restricted to West Africa, and there is low genetic diversity of baobab in eastern Africa, suggesting that baobab moved across the continent in the opposite direction, from west to east.103 Archaeobotanical evidence, though limited, corroborates a later appearance of baobab in East Africa compared with West Africa.104

A major challenge for archaeologists and others trying to make sense of linguistic models dealing with large geographical areas and time spans is the lack of consensus among linguists themselves.105 At the same time, archaeologists have not always kept abreast with developments in historical linguistics and consequently may rely on outmoded genealogical reconstructions. A case in point being the continuing assumption that South Cushitic is one of the primary branches of Cushitic and hence of considerable antiquity (which in turn encouraged putative associations with the archaeological traces of early herders in East Africa), despite Robert Hetzron’s106 arguments that it is a younger sub-branch of East Cushitic. These and other methodological issues107 need to be considered when drawing on linguistic evidence to support hypotheses about precolonial farming communities, and interdisciplinary perspectives of the kind now being undertaken in Central Africa108 are crucial for their evaluation and reworking.

Landscape Perspectives

Using the landscape as a unit of analysis is often quite rewarding when thinking about people and livelihoods across vast temporal-spatial scales. The act of cultivating plant species or introducing livestock into certain areas can have profound impacts on vegetation structures, burning regimes, soil-nutrient compositions, and erosion rates. Recently, the notion of landscape domestication, paralleling ecological ideas of niche construction, has also been introduced. Broadly speaking, domesticated landscapes are landscapes where humans have created environmental niches, not only supporting human society, but also affecting many other species.109 Sub-Saharan African examples include the formation of African Dark Earths (AfDEs) in the West African rainforest zone110 and the relationships between human-settlement practices and baobab-tree (Adansonia digitata L.) recruitment in West African parkland.111 The open patches of closely cropped grass within a wider mosaic of woody and/or bushy vegetation found in semi-arid savanna landscapes in eastern Africa provide another example. Often known as “glades,” these have diverse origins and complex histories, and various natural processes can contribute to their creation and maintenance. Research has also shown that many such glades mark the site of former pastoralist settlements (figure 7) that after abandonment can remain within the landscape as distinct, nutrient-rich “hotspots” for generations, and even centuries,112 with beneficial consequences on local biodiversity at all trophic levels.

Figure 7a. Excavations at the Maasai Plains site, Mugie Ranch, Laikipia Plateau, Kenya, in 2005; the low mound in the foreground represents the remains of a dung-ash-and-refuse midden. This “glade” marks the location of a large former pastoralist encampment that may have remained open for upward of 600 years.

Photo by Paul Lane, 2005.

Figure 7b. Aerial views of a sequence of former pastoralist settlements, Amboseli, south Kenya.

Photo by Paul Lane, 2009.

Paleo-environmental reconstructions based on evidence from, for example, pollen, phytoliths, fossilized leaf impressions and microcharcoal, and geochemical analysis also allow for the creation of models of landscape and climate change through time. Untangling signatures of landscape change or continuity realized through anthropogenic and other forces is complex. Interdisciplinary perspectives are necessary to identify the varied processes through which landscapes have responded to farming and herding activities, and on what scales. Pollen cores can directly indicate fluctuations in plant species at the landscape level coinciding with fluctuations in the ubiquity and abundance of plant species in regional archaeological contexts.113 The analysis of charcoal assemblages recovered from archaeological contexts can provide information about fuel-wood and building-material choices and, if linked with ethnohistoric and other landscape data, may also provide insights into sociocultural perceptions and categorizations of different landscape types, as documented ethnoarchaeologically for Fang villages in Equatorial Guinea.114 With landscape vegetation reconstructions, it can be informative to not only examine evidence for species that are cultivated or tended by people but also to look for plants and other organisms that colonize and thrive in anthropogenically modified soils. For example, indirect indicators of pastoral activities may include increases in dung fungal spores recovered from sediments thought to have formed around the same time livestock proliferated.115 Different cropping regimes can also be inferred from the range of weedy taxa recovered from archaeobotanical samples.116

Changing landscapes must be considered when hypothesizing about the spread of animal and plant resources and farming traditions in the past. Edmond De Lange117 has suggested a reasonable hypothesis for the route of diffusion of banana across Africa around 3500 bp through patchy forest networks that are no longer in existence, but once linked the Usambara and Pare Mountains in northern Tanzania, the northern extent of the equatorial rainforest and West Africa. Such a proposal is only possible through paleo-environmental reconstruction efforts. As argued by Diane Gifford-Gonzalez,118 as early herding communities advanced southward through eastern and southeastern Africa, they are likely to have encountered various livestock diseases, such as trypanosomiasis and contagious bovine pleuropneumonia. These would have hindered their exploitation of particular areas, until the etymology of these diseases had been learned and the vegetation structure had been modified. The presence of these diseases and their vectors, and the need to learn how to manage habitats and avoid others so as to minimize their effects on livestock herds, may explain why the spread of livestock herding took over two millennia to reach the southern tip of the continent after their first appearance around Lake Turkana, Kenya, although it is also evident, as suggested by recent isotopic studies on faunal material from sites around Lake Victoria, that this was not always the case.119 Recent discoveries in northern Tanzania of occurrences of domestic taxa and associated Pastoral Neolithic material culture dating to around 4000–2900 bp at Luxmanda, further challenge the hypothesis that a disease-associated frontier delayed the southward spread of herding, although it is possible that disease did prevent the early southward spread of donkeys.120 It is evident also that pioneer herding communities often interacted with autochthonous hunter-gatherer-fishers, sharing knowledge, livestock, and technologies in often quite complex ways that caution against using material indices alone to infer a change in subsistence orientation.121

Geo-archaeological studies of soils and sediments can also provide information about the type of cultivation practices employed, the rate and frequency of agriculturally initiated soil erosion, water management, and evidence for soil enrichment or conservation practices, such as by manuring or mulching.122 Common techniques include examination of soil chemistry, lipid biomarkers, and soil micromorphology, coupled with the use of radiometric dating. When linked with palaeoenvironmental evidence and data on archaeological site distributions, it may be possible to reconstruct landscape dynamics over long time frames,123 and so establish the relative and variable contributions of both natural and anthropogenic processes to contemporary environmental problems.124

Discussion of the Literature

African farming and herding systems, despite an apparent reliance on “simple” technologies, are complex and remarkably adaptive. Combined archaeological, linguistic, palaeoenvironmental, and genetic evidence is also demonstrating the diverse transformations in these systems across Africa over the last 8,000 years or so as different animal and crop species were crafted to fit new environments. These novel herding and farming practices and agro-ecosystems ultimately placed certain selective pressures on crop and animal species giving rise to distinct cultivars, landraces, and hybrids, as well as different cultural approaches to food production.

As work on the origins and implications of this variability intensifies, and as more securely dated and more comprehensive data sets become available, understanding of these multiple pathways to food production, agricultural intensification, plant and animal introgressions and diaspora, and the corresponding evolution of distinctive agricultural and culinary practices will undoubtedly deepen and change. While regional sequences are now fairly well established, and the broad parameters of both plant125 and animal domestication126 are also understood, for many areas there is virtually no firm evidence for when either farming or herding first began, where and how the different exploited crops and animals originated, or even which of these species were preferred. Reconstructions of regional histories thus rely on relatively small bodies of well-dated physical evidence from a handful of sites, coupled with associated genetic and/or linguistic reconstruction, inferences drawn on data from neighboring areas, and the widespread use of the presence of certain material evidence (especially ceramics and in some areas iron-smelting remains) as proxies for the inception of food-producing economies. For some critical areas, there are virtually no relevant data. These include areas thought to have been key corridors along which different domesticates spread with knowledge about their propagation (e.g., the southern end of the western Rift Valley, the coastal belt of southern Tanzania, and northern Mozambique), and some of the localities believed to have been centers of initial domestication (such as parts of the Ethiopian highlands).

The reasons for these large gaps are numerous. They include the low numbers of professional archaeologists in most sub-Saharan African countries, the recent origins (relative to other more intensively investigated continents) of research on Africa’s Holocene archaeology, internal and external research agendas that have directed archaeological attention toward other topics, and the limited funding available to support archaeological research, especially for local scholars. Other factors are of a more practical nature; for example, sampling for archaeobotanical remains on excavations and the use of flotation techniques for their recovery are still the exception rather than the rule, and most studies have been predominantly focused on Egypt, the Sahara, and West Africa, and to a lesser degree East Africa, while central and southern Africa are significantly underrepresented.127 This lack of attention to plant remains is partly due to the shortage of trained archaeobotanists on the continent, and possibly also because of a widespread belief that botanical remains are unlikely to survive in the majority of Africa’s tropical soils owing to a combination of taphonomic processes and depositional and crop-processing practices (e.g., figure 8).128

Figure 8a. Figures 8a, 8b, and 8c show taphonomic aspects to finger millet processing in Marakwet, Kenya. Using Marakwet processing techniques, macroscopic remains of millet plants do not come into contact with fire until after they are rendered into fine flour, and are thus unlikely to be preserved by charring. Figure 8a: Millet panicles and unintentional weedy seed inclusions are pounded in a mortar to break the husks. Photo by Anna Shoemaker, 2013.

Figure 8b. Mortar contents are winnowed to remove fragmented chaff and small weeds from the millet grains. Photo by Anna Shoemaker, 2013.

Figure 8c. Clean millet grains are ground into flour. Photo by Anna Shoemaker, 2013.

Yet, while preservation issues are certainly important,129 as recent targeted sampling for plant remains at sites along the East Africa coast130 and similar work elsewhere131 has attested, most of the current gaps in knowledge are as much likely due to a combination of methodological shortcomings and the specific biological characteristics of African domesticates rather than poor preservation alone. The situation from the perspective of faunal remains is only marginally better. While the greater visibility of animal bones (although not necessarily fish bones) has enhanced their recovery, many tropical soils are not conducive to their long-term survival. Also, there are only a few trained zoo-archaeologists on the continent, suitable reference collections are rare, and most studies have been undertaken by foreign researchers whose own work and access to collections are also subject to pressures of funding and other external factors.132

Evaluating the appearance of agricultural practices and people through space and time is not just important for understanding the past; it also has more contemporary relevance, for instance in contributing data to global climate-change projections where none yet exists.133 However, there must also be the creation of histories of farming and herding communities in Africa and indeed globally that “do not homogenize diversity or collapse difference, but remain open to the contingency and place-specificity of practices in the past.”134 Lack of understanding of the long history of human transformation of African landscapes, mediated through such practices as low-intensity farming, must be also addressed.135 Conservation efforts in Africa frequently result in the displacement of people in the belief this will return landscapes to healthier conditions.136 Yet archaeological (and historical) research has come to demonstrate that areas once revered as “pristine” have had much longer histories of farming occupation than previously conceived.137 With the shift in ecological thinking that has embraced non-equilibrium ecosystems,138 anthropogenic disturbance has come to be perceived as a force that does not unequivocally degrade landscapes, but rather is an important decider of environmental health.139 Further study of processes of landscape domestication informed by detailed understanding of the long-term archaeologies of herding and farming are necessary for developing environmental management practices that are historically contingent and sustainable.

Primary Sources

Regarding the material evidence for food production, archaeological excavation, and survey reports, archaeological and ethnographic museum collections, site archives, and online databases are the main repositories of primary sources (other than archaeological sites and landscapes themselves). Nineteenth century and early to mid-twentieth-century published ethnographies, along with accounts written by European and North American explorers, missionaries, and colonial officials, are also a good source of information about practices, products, and equipment in the relatively recent past. For linguistic evidence, dictionaries are perhaps the most important kind of source regarding the terminologies of food production and practices. These are too numerous to list in detail here, however.

Further Reading

  • Beaujard, Philippe. Histoire et voyages des plantes cultivées à Madagascar avant le XVIe siècle. Paris: Karthala, 2017.
  • Bostoen, Koen. “Pots, Words and the Bantu Problem: On Lexical Reconstruction and Early African History.” The Journal of African History 48.2 (2007): 173–199.
  • Cappers, René. Fields of Change: Progress in African Archaeobotany. Groningen: Barkhuis and Groningen University, 2007.
  • Davies, Matthew. “Economic Specialisation, Resource Variability, and the Origins of Intensive Agriculture in Eastern Africa.” Rural Landscapes: Society, Environment, History 3 (2015): 1–18.
  • Fahmy, Ahmed G., Stefanie Kahlheber, and A. Catherine D’Andrea. Windows on the African Past: Current Approaches to African Archaeobotany. Frankfurt: Africa Magna Verlag, 2011.
  • Harlan, Jack R., Jan M. J. de Wet, and Ann B.L. Stamler. Origins of African Plant Domestication. The Hague: Mouton, 1979.
  • Jousse, Hélène, and Joséphine Lesur. People and Animals in Holocene Africa: Recent Advances in Archaeozoology. Frankfurt: Africa Magna Verlag, 2011.
  • Koponen, Juhani. People and Production in Late Precolonial Tanzania: History and Structures. Monographs of the Finnish Anthropological Society 23. Jyväskyla and Uppsala: Finnish Society for Development Studies and Scandinavian Institute of African Studies, 1988.
  • Lane, Paul J. “Early Agriculture in Sub-Saharan Africa to Ca. Ad 500.” In Cambridge World History, volume 2: A World with Agriculture. Edited by Graeme Barker and Candice Goucher, 736–773. Cambridge: Cambridge University Press, 2015.
  • Marshall, Fiona, and Elisabeth Hildebrand. “Cattle before Crops: The Beginnings of Food Production in Africa.” Journal of World Prehistory 16.2 (2002): 99–143.
  • Mitchell, Peter, and Paul J. Lane. The Oxford Handbook of African Archaeology. Oxford: Oxford University Press, 2013.
  • Stevens, Chris, et al. Archaeology of African Plant Use. Walnut Creek, CA: Left Coast Press, 2014.