A second poster titled “Historical-Linguistic Approach to the Subsistence of the First Bantu Speakers South of the Rainforest: The Banana Case” was presented by Sifra Van Acker and co-authored by Sara Pacchiarotti and Koen Bostoen. This poster presented preliminary results of Sifra’s ongoing PhD research on the reconstruction of subsistence-related vocabulary in Proto-West-Coastal-Bantu as part of the project’s wider historical-comparative linguistic research. The banana case is highly relevant because it is still not clear when, how and by whom bananas were introduced in Africa and which role they played in the diet of early Bantu speakers and in the expansion of their languages through Central Africa.
Location of the 4 study areas of the September-November 2018 fieldwork in the Kongo Central (formerly Bas-Congo) province of the DRC
In close cooperation with the Institut des Musées Nationaux du Congo (IMNC), archaeological and paleo-environmental fieldwork was carried out from September 20 to November 24, 2018 in the Kongo-Central province of the DRC. The mission was carried out by Prof. Bernard Clist (mission leader, Ghent University), Clément Mambu Nsangathi (IMNC), Isidore Nkanu Nsasa (driver), and Suzanne Bigohe Mugisha (student from Kinshasa University).
The objectives of the mission were: (i) to gather more information from the Neolithic and Early Iron Age archaeological sites excavated in 1951, 1984 and 2015 by M. Bequaert, P. de Maret and B. Clist, respectively; (ii) to investigate possible ancient shell middens along the Atlantic Ocean coast and in the ‘Parc Marin des Mangroves’ along the Congolese banks of the Congo river; (iii) to carry out surveys to locate new sites to be excavated; (iv) to sample a maximum of ancient village pits to identify crops by carrying out flotations; and (v) to extract soil samples from as deep as possible to understand past vegetation changes. All of these objectives aim at a better understanding of the social and economic dynamics of the earliest villages settling the Kongo-Central province between the Atlantic coast and Kinshasa.
During the fieldwork period, we traveled 3,267 km, identified and documented 64 new archaeological sites dating from the Middle Stone Age to the Late Iron Age in the 4 selected study zones (Figure 1), and excavated 10 ancient hunter-gatherer, village and iron working sites. We were oriented to 4 sites by previous research (M. Bequaert, P. de Maret, B. Clist). However, new large-scale surveys around Kinkenge, Muanda and Nduizi (formerly Kongo dia Vanga) enabled us to excavate 6 new settlements. The surveys were strengthened by systematically interviewing farmers we met while walking through the fields and villages, after we showed them samples of broken pottery. These interviews sometimes led to the identification of promising sites, e.g. Muanda 6 on the coast.
Small scale archaeological excavations were carried out at Sakuzi (zone 1), Kindu and Mantsetsi (zone 2), Nduizi, Nguemba 1, Mbanza 2A, and Mbanza 2B (zone 3), and Boma1, Katala, Muanda 6, and Muanda 13 (zone 4). Soil sampling was conducted at the Kindu, Mantsetsi, Mbanza 2, and Muanda 6 sites, and pit sampling at the Kitala, Mbanza 2, Nduizi, Nguemba 1, and Sakuzi sites.
Feature n°6 at the Sakuzi site, completely eroded pit with potsherd concentration at its base.Soil sampling at the Kindu site for paleo-environmental studies.Mbanza 2 site, finding both nice mushrooms for supper and intact Early Iron Age Kay Ladio pots to study.Nguemba 1 site, iron processing feature excavated under the rain.Excavation in the Nduizi village, mainly Early Iron Age material.
Stone Age: Very interesting data related to the Stone Age was collected at Boma 1. It consists mainly of quartz flakes, blades and cores found at -6.20 meters in a large borrow pit by the Boma-Muanda road. Stone artefacts, probably younger, were found stratified at Kindu and Mantsetsi down to -2 meters. Charcoals collected in situ, Carbon-13 (13C) analysis of soils, and the study of artefacts will give new insights on this epoch.
Neolithic and Ancient Iron Age: Most of the finds relate to this period due to the orientation of the BantuFirst project, roughly between 2,400 and 1,500 BP. Excavations at Kindu, Mantsetsi, Nduizi, and Sakuzi aimed at obtaining new data after prior excavations dating back to 1951 and 1984. Crucially, new pottery styles were identified in several pits at Nduizi and Sakuzi. Their study will certainly expand our preliminary understanding of the first villages to have settled the Kongo-Central Province.Apart from a poorly documented archaeological mission by the IMNC in 1986, we were the first to survey and excavate the area near the coast, west of Matadi, specifically between Boma and Muanda. At Muanda 6, systematic dry sieving yielded fish and small mammal bones associated to shell beads. Preliminary on-site analysis suggests that the Muanda style pottery is the oldest found in the lower level of Katala, at Muanda 6, and near the border to Cabinda at Muanda 8. It is followed by the Katala style of pottery found stratified under the present-day village and above the Muanda style level. We are now waiting for Carbon-14 (14C) dates to understand where the Muanda and Katala styles stand within the pottery cultural sequence of the Kongo-Central province.
Late Iron Age: The analysis of surface-collected pottery in the four surveyed areas will enable us to develop local cultural sequences. Such sequences will be informative in isolating interesting Ancient Iron Age sites to be excavated in 2019.
Joseph Emboto and Flore Bollaert (Mbankana, DRC)Linguistic fieldwork in the little town of Mbankana (DRC)Collecting Bwala data with Joseph Maba (Mbankana, DRC)
Flore Bollaert (MA African Studies, UGent) carried out linguistic fieldwork in the region of Kinshasa (DRC) in August 2018 as part of her MA research. She is carrying out a comparative study of the noun class systems of the Nzebi (B50), Mbete (B60) and Teke (B70) languages of West-Coastal/ West-Western Bantu. The objective of her fieldwork was to collect new data on some Teke languages, since these are underrepresented in the existing literature. She collected the bulk of her data in Mbankana, a small town approximately 150 km east ofKinshasa. She primarily worked with Joseph Emboto, who was her main consultant, but she also met and spoke to several other people. In Mbankana, Flore collected data on Bwala, an undescribed and undocumented Teke variety. Additionally, she collected in Kinshasa some data on Wumbu (B78), another Teke variety, in collaboration with native speaker Edouard Mbabu.
The project’s aims and first results have been presented at the 15th Congress of the PanAfrican Archaeological Association for Prehistory and Related Studies in Rabat (Morocco) on September 12, 2018. Bernard Clist discussed “The first villages of Central Africa before 1500 BP” while Dirk Seidensticker presented “New archaeological research on the earliest villages south of the Central-African rainforest”. The two presentations were part of a special session organized by Bernard Clist, Pierre de Maret and Thomas Huffman covering “Early villages and farming through African humid forests, from Central Africa southwards”.
The presentation of Bernard Clist gave an overview of Pre- and Early Iron Age groups in the western parts of Central Africa, especially in Gabon and the coastal areas of the two Congo states. The presentation included a detailed discussion of the known remains of past subsistence strategies. Dirk Seidensticker presented the first results of the project’s 2018 archaeological fieldwork campaign by himself and Katharina Jungnickel, including the first excavation of the site of Mukila since 1952 and the first-ever survey along the road from Kinshasa to Bandundu. Due to the virtually total lack of archaeological research within the Mai-Ndombe, Kwango and Kwilu provinces of the Democratic Republic of the Congo, these new findings will be pivotal for establishing the basis for future research.
During the EAA meeting of 2018 Dirk Seidensticker, together with Florian Thiery, Allard Mees (both Römisch-Germanisches Zentralmuseum Mainz) and Clemens Schmid (Kiel University) presented their research on “RDF based modeling of relative and absolute chronological data: Examples from the central african rainforest and roman periodisation”. The method presented untangles hidden assumptions that are part of the way archaeologists build chronologies. Chronological intervals are sometimes based on absolute “certain” datings (e.g. C14) or derived from relative dating or combinations of both. Chronological relations between entities are modeled using Allen`s interval algebra within a RDF graph in order to produce a machine readable and processable semantic representation. The presentation can be accessed via Zenodo and the code is available via GitHub.
Progress on the historical linguistic side of the project was presented by Sara Pacchiarotti and Koen Bostoen at the 7th International Conference on Bantu languages (SINTU 7) in Cape Town and at the 48th Colloquium on African Languages and Linguistics (CALL) 2018 in Leiden.
The talk at SINTU presented preliminary results on final vowel loss in West-Coastal/West-Western Bantu languages belonging to Guthrie’s groups B50, B60, B70 and B80. Through historical comparative work, we believe that this sound change occurred systematically in some B70 and B80 languages, but not in B50 and B60. Currently, we posit that B80 languages spoken in the eastern part of Congo DRC, such as Yans (B85), Nsong (B85d), Ding (B86), Lwel (B862), Mpiin (B863), Mbuun (B87) and Nzadi (B865), were the center of innovation for final vowel loss. This historical sound change then petered out through contact towards some B70 languages spoken in the western part of Congo DRC (e.g. Teke Tio Bali B75, some East Teke B76 varieties and some South Teke B78 varieties), some languages belonging to the Kikongo Language Cluster (e.g. Hungan H42, Samba L12a), and possibly some Central Western languages, such as Bushoong (C83).
The talk at CALL presented a preliminary assessment of diagnostic sound changes for the internal classification of West-Coastal/West-Western Bantu languages belonging to Guthrie’s groups B50, B60, B70 and B80. We observe that most Proto-Bantu consonants in C(onsonant) 1 position are maintained in these groups. In C(onsonant) 2 position, however, several consonants undergo mergers, simplification or are completely lost in some B70 and B80 languages. This tendency to lose/reduce segments in C2 position results in very eroded reflexes of reconstructed Proto-Bantu forms and is additionally “aggravated” by the loss of final vowels in some languages within these two groups. Tentatively, B50-60 languages appear to be more conservative than B70 and B80. They retain consonants in C2 position which are lost or undergo mergers in B70-80. Moreover, B50-60 languages were not affected by systematic final vowel loss. Guthrie’s B70 group appears to be fragmented, with West Teke B73 often patterning with Guthrie’s B50-60 and South and East Teke (B77, B78 and B76 respectively) patterning with B80. Lastly, Guthrie’s B80 appears to be at least two separate groups on the basis of shared phonological innovations (i.e. B81-B83 except Boma Yumu B80z vs. B85-87+ Boma Yumu B80z).
Some aspects of this new research have been incorporated in an interactive map that can be found here: West-Coastal Bantu Interactive Map.
The first archaeological field mission, carried out by Dirk Seidensticker and Katharina Jungnickel, in close collaboration with Clement Mambu, Jeanine Yogolelo and Roger Kidebua (IMNC) during the summer of 2018, covered the western half of the former Bandundu province (current Kwango, Kwilu and Mai-Ndombe provinces) of the Democratic Republic of the Congo, an area largely unexplored by archaeologists. The aim of this mission was to conduct an initial survey between the Congo, Kwango and Kasai Rivers.
Excavation at the school in Mukila.
To compensate for the lack of any systematic and well-documented research in the area, the fieldwork was preceded by a perusal of the archives at the Royal Museum for Central Africa in Tervuren, which highlighted the rich potential of one specific site: Mukila (Kwango province). Maurice Bequaert excavated this site located some 250 km to the east of Kinshasa on the southern margins of the Bateke plateau in the early 1950s. There, he uncovered both stone tools and ceramics, possibly pointing towards the coexistence of hunter-gatherers and early sedentary communities. As Bequaert’s excavation was only poorly documented (multiple sketches and some photographs) and not published in any detail, we chose Mukila as our first priority target. Thanks to the available photos from the early 1950s and with the help of locals, we relocated two of Bequaert’s trenches. Our first test trench had to determine the quality of finds as well as the visibility and occurrence of features. We subsequently test-cored the site and opened a second 1.5 by 4.5 m large trench on a spot close to Bequaert’s richest excavation where we found promising finds during the coring as well. We excavated it up to 3.6 m, using 75x75x20 cm units. In total 126 units were systematically dry-sieved for small finds. Finds consisted of a rich mixture of pottery and stone tools. At the bottom of the profile we added another 3 m by extracting two additional cores resulting in a 6.6 m deep sequence. We systematically sampled the cores and the profile for paleo-environmental remains.
Recovery of a thin layer of lithic flakes and chips close to Bandundu.Survey on agricultural fields within the area of Bandundu.
After our excavations at Mukila we conducted a detailed and first-ever survey along the dirt roads between the towns of Mongata, Masia-Mbio and Bandundu, a 300 km long stretch between the Congo and Kwango rivers. We surveyed 62 distinct areas corresponding to roughly 11 hectares in total. Besides surveying areas currently used for agriculture, mainly the cultivation of manioc, we also looked at borrow pits created for the construction of the dirt roads. These borrow pits were of special importance as they allowed us to review existing profiles and thus uncover cultural layers invisible to surface surveying only. Two very large borrow pits in the vicinity of Bandundu town yielded a vast collection of stone tools both on the surface and in the extant profile. A variety of half-finished and elaborately done bifacial points as well as large amounts of flakes and chips were scattered within the pits, sometimes forming small concentrations. At the second pit some potsherds were scattered on the surface as well. We extracted botanical samples for paleo-environmental research from the extant profiles of both pits and retrieved charred remains through flotation. While doing so, we discovered a thin layer of lithic flakes and chips, which might be the remains of one single knapping event.
Finally, we identified several fields rich in pottery finds close to the town of Bandundu. These ceramics represent quite some stylistic and technological diversity, possibly of both local and external origin.
Sifra Van Acker presented a talk titled “Contact-induced change in Kisamba (Bantu, L12a), a poorly described Congolese language” at the 9th World Congress of African Linguistics in Rabat (Morocco) on August 25, 2018. Kisamba is a West-Coastal Bantu language belonging to the Kikongo Language Cluster, more specifically the Kikongoid subgroup. Being spoken by small speech communities that are scattered over the Kwilu province of the Democratic Republic of the Congo, Kisamba underwent several distinctive changes through contact with other West-Coastal Bantu languages, such as the B80 languages, and South-West Bantu languages, such as Kimbala (H41) and Kwezo (L13). This talk focused on three of these contact-induced changes: the voicing of coronal and velar stops, final vowel loss and double negation.
On November 27, 2018, on-going research of the BantuFirst team was presented in Tervuren at the 18th Meeting of Partners of the Congo Basin Forest Partnership, which focused on the environmental dynamics of the Congo Basin.
