No abstract

The present article examines whether Late Holocene climate-induced vegetation changes in the Central African forest block may have facilitated the Bantu Expansion. This is done through a body of evidence that is not commonly used for the reconstruction of vegetation dynamics, i.e. language data. The article focuses on common Bantu vocabulary for three pioneer species abundantly present in the Central African pollen record between ca. 2500 and 2000 BP: Musanga cecropioides, Elaeis guineensis, and Canarium schweinfurthii. The geographical distribution patterns of the vernacular names for these pioneer trees add weight to the hypothesis according to which the rainforest contraction that emerged in the first millennium BC had an impact on the way Bantu languages dispersed.

Umlaut is a rather uncommon sound change in Bantu. However, it is prolific in a number of closely related Bantu languages spoken in the Kwilu District of DR Congo, also known as the 'Bantu B80' languages. This shared phonological innovation is not only diagnostic of the genealogical unity of this language cluster, but variation in the way it is realized is also significant for their internal classification. In some languages, umlaut leads to an increase in phonemes through the creation of a new set of rounded front vowels. In others, umlaut triggers vowel change without impact on the number of vowel qualities. This paper presents a detailed descriptive and comparative study of this specific kind of vowel harmony and the phonological contexts in which it is triggered.

Popular belief has it that the Bantu Expansion was a farming/language dispersal. However, there is neither conclusive archaeological nor linguistic evidence to substantiate this hypothesis, especially not for the initial spread in West-Central Africa. In this chapter we consider lexical reconstructions for both domesticated and wild plants in Proto-West-Coastal Bantu associated with the first Bantu speech communities south of the rainforest about 2500 years ago. The possibility to reconstruct terms for five different crops, i.e. pearl millet (Pennisetum glaucum), okra (Hibiscus/Abelmoschus esculentus), cowpea (Vigna unguiculata), Bambara groundnut (Vigna subterranea) and plantain (Musa spp.), indicates that by that time Bantu speakers did know how to cultivate plants. At the same time, they still strongly depended on the plant resources that could be collected in their natural environment, as is evidenced by a preliminary assessment of reconstructible names for wild plants. Agriculture in Central Africa was indeed "a slow revolution", as the late Jan Vansina once proposed, and certainly not the principal motor behind the early Bantu Expansion.

Cet article étudie un phénomène bien connu dans les langues du monde, mais rare dans les langues bantu. Ce fait, la diphtongaison, est pourtant répandu dans quelques langues bantu des groupes B70 et B80 parlées au sud-ouest de la RD Congo dans le district du Kwilu. L’article permet de comprendre clairement les différents conditionnements et les réalisations variables de ce phénomène. Certaines voyelles sont beaucoup plus productives de diphtongues que d’autres. Les divers types de diphtongues sont regroupés en trois grandes catégories selon l’aperture d’origine de la voyelle diphtonguée : la diphtongaison des voyelles mi-ouvertes (* e, * o), la diphtongaison des voyelles mi-fermées (* ɩ, * ʊ) et la diphtongaison de la voyelle ouverte et centrale (* a). Dans tous les cas, la diphtongaison entraîne la formation d’une semi-consonne qui précède la voyelle rompue. La diphtongaison la plus fréquente est attribuable à la longueur vocalique.