Phase 1: Cultural diversification: Project 014|
Linguistic and cultural evolution of Bantu-speakers in sub-equatorial Africa
SUPERVISOR: - Ruth Mace
Clare Holden (Department of Anthropology, University College London)
PROJECT FUNDING: Grant
In this project I am investigating the causes of cultural variation in Bantu-speaking populations in sub-equatorial Africa, focussing in particular on matrilineal descent. A tree of 75 Bantu languages was constructed using maximum parsimony (Fig. 1). The tree reflects the spread of farming across the region (Fig. 2). Bantu language trees are used as a proxy for cultural phylogeny, in order to investigate processes of cultural evolution. Phylogenetic comparative methods are used to control for statistical non-independence among cultures (Galton’s problem).
Bantu language trees can be used to infer the cultural practices of early Bantu-speakers, for example, when did matriliny and patriliny arise? I have also tested for correlated evolution among cultural traits, and tested the direction of cultural evolutionary change (which of two traits changed first). For example, adopting cattle led to a change from matrilineal descent to patriliny or mixed descent in formerly matrilineal Bantu-speaking populations.
Recently I have constructed Bayesian Bantu language trees based on Markov Chain Monte Carlo (MCMC) sampling methods. This provides a sample of possible trees, sampled in proportion to their likelihood. One can then test hypotheses using multiple possible trees, allowing one to make inferences about cultural evolution even though the ‘true tree’ of the Bantu languages remains uncertain.