This research project, a typology of passive constructions in Bantu, involves several fields of linguistics. It focuses on a particular language family, i.e. Bantu, and addresses the question of variation and micro-variation within the family on a specific morphosyntactic construction, namely the passive. Data collection involves the use of grammatical descriptions and text corpora, but will also resort to elicited fieldwork data. Whereas field data are synchronic, a certain number of grammars and text corpora used in the survey are several decades or even centuries old, and thus add a diachronic component to the study. Based on the existing functional typological literature, the project will contribute to enrich existing typologies and feed into the current debate about canonicity. Finally, by adopting and applying phylogenetic methods on language data, it will provide input to the on-going discussion about stability of specific grammatical features.