The ‘Bantu Expansion’ is one of the most controversial issues in African History that has sparked intense debate across several scientific disciplines since the 1950s. The Bantu language family branched off from the rest of the Niger-Congo phylum not earlier than 5,000 years ago. Despite being remarkably young, it is by far Africa’s largest language family, in terms of number of speakers, number of languages and geographical spread. It is widely agreed that this apparent paradox can only be accounted for by a rapid spread of Bantu languages from their West African homeland in the Nigerian-Cameroonian borderland to southern Africa. To do so, Bantu-speaking communities may have followed temporary savannah corridors through the rainforest, formed during severe climate anomalies. This enabled Bantu speakers to settle in a new homeland south of the rainforest around 2,500 BP. These first settlers introduced the Bantu language ancestral to the present-day ‘West-Western’ or ‘West-Coastal’ clade. This putative Bantu homeland has been situated in the interior to the northeast of Malebo Pool.
However, the patterns and the driving forces of this migration are hotly debated. Two widely accepted paradigms are that [1] it was a single migratory macro-event; and [2] it was a farming/language dispersal. Sadly, these popular paradigms are based on very limited evidence, also within the project’s study area: [1] archaeological research on the homeland of the first Bantu speakers south of the rainforest is still non-existent, [2] there is no convincing evidence that these communities were indeed farmers and [3] it is still very poorly understood how the different language subgroups of the West-Western clade genealogically relate to each other.
Therefore, the BantuFirst project aims at a new paradigm-shifting narrative on the Bantu Expansion, by developing a pioneering and comprehensive cross-disciplinary approach. The project aims at untangling the interconnections between human migration, language spread, climate change, subsistence and early farming in the wider area of the putative West-Western Bantu homeland. The core scientific expertise of the project team consists of historical linguistics, archaeology and archaeobotany, while palaeoenvironmental, archaeozoological and genetic data is regularly integrated through inter-university collaboration.
Linguistic Research
The lexicon-based phylogenetic research done within the KongoKing project indicates that West-Western Bantu languages form a discrete clade within the Bantu family tree. This clade breaks up in three genealogical subgroups: (1) the Nzebi-Mbete-Teke languages, still spoken today in the vicinity of the Bateke plateau, (2) the Yanzi languages, stemming from an ancestor language that moved east of the Congo River somewhere in between the Kwango and Kwilu rivers in the DRC’s Bandundu Province, and (3) the Kikongo Language Cluster (KLC), a vast cluster of regiolects spoken in the wider Lower Congo region, whose ancestor language spread in southwestern direction towards the Atlantic Ocean. Although the phylogenetic subgroups appear to be robust, their internal relationships remain uncertain.
The first objective of the project’s linguistic research is to integrate a comprehensive set of West-Western Bantu data from the Nzebi-Mbete-Teke and Yanzi subgroups into the historical linguistic research carried out on the KLC during the KongoKing project within a lexicon-based phylogenetic approach. This historical research will be pushed in new directions by factoring in phonological and grammatical data and by exploring new ways of time-calibrating the language phylogenies, building on extant preliminary work with linguistic time trees. Compared to the KLC, languages from the Nzebi-Mbete-Teke and Yanzi subgroups show peculiar phonological, morphological and syntactic features, which are absent from the KLC and are rather atypical from a more “common” Bantu point of view. These include: rare vowel harmonies, umlaut effects, final vowel loss, 9+ vowel systems, fusions of verb suffixes producing abnormal verbal bases, rare polysemies (e.g. causative/applicative syncretism) and absence of passive morphology. The relative isolation of the Nzebi-Mbete-Teke and Yanzi subgroups in the transition zone between the equatorial rainforest and the southern savannahs may have played a decisive role in the development of their distinctive linguistic profile.
The second objective of this project is to shed new light on another factor which may have contributed to the distinctive linguistic profile of the Nzebi-Mbete-Teke and Yanzi subgroups, but has never been closely examined before. This factor is the substratum influence in the Nzebi-Mbete-Teke and Yanzi subgroups from non-Bantu languages, possibly spoken by autochthonous hunter-gatherers groups with which the ancestors of these two groups interacted in the vicinity of their putative homeland, after Proto-West-Western Bantu had already fallen apart in distinct subgroups. Available genetic data indicate that the maternal gene pool of several western Bantu speech communities is characterized by the significant presence of mtDNA haplogroups which are omnipresent in Central African ‘pygmy’ groups. These genetic findings will be assessed from a linguistic point of view by systematically comparing the many peculiarities of the Nzebi-Mbete-Teke and Yanzi languages with those reported in Bantu languages of present-day Pygmy communities.
Finally, the third objective is to expand the existing specialized reconstructed lexicon of West-Western Bantu, especially in the domains of subsistence and land use strategies of ancestral West-Western Bantu speakers. These reconstructions will be compared to the archaeological and ethnobotanical findings brought about by the project.
Archaeological Research
The archaeological research within the BantuFirst project aims at conducting systematic surveys and small-scale excavations along the Bateke Plateau, between Kinshasa, Bandundu and Kikwit in the Democratic Republic of the Congo as well as Brazzaville and Djambala in the Republic of the Congo. The study area is not only the putative homeland of the West-Western Bantu languages, but also and above all a blank spot on the archaeological maps of the Early Iron Age of Central Africa. The paucity of empirical, genuine historico-cultural data renders interdisciplinary synthesis of the Bantu Expansion difficult.
Our research program aims at establishing a better understanding of the appearance of the first villages, the expansion of Bantu speech communities and the spread of iron metallurgy south of the rainforest, and their interrelationship with their environment itself changing within a severe palaeoclimatic dry episode. Within the perimeter of the current distribution area of the West-Western Bantu languages, our (poor) archaeological knowledge concerning the earliest phase of settlement by pottery producing communities that presumably had a sedentary and food producing lifestyle is mainly limited to the Lower Congo Province of the DRC, to the southern coast and inland of Congo and to south-east Gabon. For the time being, the earliest relevant evidence dates back into the middle of the first millennium BC.
In order to establish a better understanding of the archaeology of the West-Western Bantu homeland, targeted prospections and excavations will be carried out on strategically selected and well preserved sites in three distinct zones where archaeological fieldwork has never been carried out to a larger extent: 1) along the Congo River in between Malebo Pool to the South and Kwamouth/Ngabé in the North; 2) on the Bateke Plateau on both banks of the Congo River; 3) along the road axis connecting Kinshasa with Kikwit and around Kikwit itself. This involves fieldwork within the départements Plateaux as well as Pool of the Republic of Congo and the provinces Mai Ndombe, Kwilu and Kwango of the Democratic Republic of the Congo.
In order to test whether ceramic traditions contemporaneous or older than Tchissanga ware occur in the western part of the study area additional dedicated fieldwork will be carried out along the coastline and adjoining Congo river banks, the Mayombe and specific areas of the Kongo Central and Kinshasa provinces of the DRC, the Cabinda, Zaire and Uige provinces of Angola and in the Kouilou Department of the Congo. The archaeology of most of these areas is virtually unexplored. Such fieldwork will also enable us to study more closely how these western ceramics relate both typologically and chronologically to the earliest ceramics found further East in the vicinity of the putative West-Western Bantu homeland.
An important part of the archaeological research will thus consist of an integrated comparative approach to early pottery in order to connect with neighbouring areas and to identify local cultural continuities and discontinuities and indications for population movements.
The only evidence so far for human subsistence in the sites linked with the first Bantu-speaking settlers south of the rainforest can be derived from carbonized remains of fruit-bearing trees, which are easily retrieved through regular archaeological fieldwork. However, within the study area, no archaeological fieldwork has ever been specifically dedicated to the recovery of archaeobotanical evidence. The planned dedicated archaeobotanical onsite sampling will pay specific attention to the retrieving and recording of plant remains. During fieldwork along the Atlantic coast in Congo, the DRC, and Angola, special attention will be dedicated to mangrove forests where Early Iron Age communities are known to have specialized in fishing and shellfish collecting thanks to the discovery of animal bones which are better preserved in shell middens. While archaeologists will focus on the means of subsistence that have left retrievable remains in Central African soils, historical linguists will additionally – but not exclusively – reconstruct the vocabulary for those plants and animals that are now archaeologically invisible.
Archaeobotanical Research
The archaeobotanical research within the BantuFirst project will lead to significant progress in the understanding of the diversified subsistence economy of early Bantu-speaking societies. In the known Neolithic and Early Iron Age sites in our research area, no single piece of direct evidence for food production has yet been found (or searched for). The Ngovo cave gives the best insight into the subsistence of what were probably the region’s earliest Bantu speakers: fishing, hunting and the exploitation of fruit trees, mainly oil palm (Elaeis guineensis) and bush candle (Canarium schweinfurthii). To gain insights in the interrelationships between the natural habitat and the evolving subsistence strategies of early village communities, the project will consider carbonized remains from archaeological excavations, but also from pedoanthracological (= non-archaeological) contexts in the vicinity of excavated sites, which is unprecedented in Central African archaeology. This research will shed a light on the question whether particular species were targeted during (fuel)wood collection and if these preferences were maintained over time. Furthermore, together with project partners, we will actively look for remains of possible crops such as pearl millet (Pennisetum glaucum) or Bambara groundnut (Vigna subterranea) through dedicated sampling (sieving and flotation of sediment samples).
Furthermore, the archaeobotanical research will deal with the question whether major environmental changes occurred in the wider Lower Congo region. Pollen records from Cameroon indicate that a climatic anomaly affected the equatorial rainforest during the Holocene. Climate change induced the retraction of rainforest and the extension of savannas, first at the periphery and later also in the core of the rainforest (around 2500 BP). This climate change is believed to have facilitated rapid eastward and southward expansion of early Bantu-speech communities. However, it is yet unclear to what extent climate change affected the composition of the vegetation in the putative West-Western Bantu homeland. Vegetation reconstructions will be carried out using identification of charcoal fragments, both from archaeological sites and from nearby pedoanthracological sites. The advantage of charcoal as compared to other commonly used proxies (e.g. pollen), is that charcoal is found in nearly all sites (while pollen require hydromorph soils), and charcoal is an inert material, so charcoal records are little affected by species-specific post-depositional degradation bias.
Finally, the archaeobotanical research will shed new light on the question whether these environmental changes induced human migration or, on the contrary, whether the vegetation changes were caused by human settlers through intense farming. The relationship between human populations and their environment is still very poorly understood. The presence of charcoal remains in almost any pit ever dug in Central Africa has been repeatedly attributed to slash-and-burn farming techniques associated with Bantu-speaking ‘farmers’. However, these studies systematically fail to explain why much older charcoal deposits (> 5000 BP), anterior to Bantu Expansion, also frequently occur in soil profiles without any artefact that indicates human activity. Nevertheless, fire periods throughout the Holocene are associated with dry climate anomalies such as the 8200 BP event, the 4000-2500 BP ‘rainforest crisis’ and the Medieval Climate Anomaly. During these periods, the rainforest was fragmented and the open patches were prone to forest fire ignited by lightning. Although the remarkable rise of fire events during the last 1000 years strongly suggests that slash-and-burn techniques became increasingly commoner during the Late Iron Age, this project will for the first time try to distinguish between charcoal deposits attributable to human activity and those that result from natural forest fires during the Late Stone Age and Early Iron Age by systematically comparing archaeological charcoal assemblages with pedoanthracological assemblages sampled in the vicinity of excavated sites.