Language offers a lens through which we can discern how humans classify the world around them. Theories of language that are based solely on data from well-documented languages can distort our understanding of the potential universal properties of languages and the full extent to which languages can vary. The analysis of data from under-documented languages is thus essential for testing the limits of linguistic theories and how well they account for the linguistic reality. The study of verbs in Bantu languages has already exposed striking variation in the verbal system compared to other language families and, in turn, impacts the way we view, for example, how speakers navigate divisions in the temporal space. The goals of this project are to (i) document and analyse how situations expressed by verbs unfold in time (i.e., actionality) in Kutu and Kwere, two closely related under-documented and endangered East Ruvu Bantu languages spoken in central Tanzania, (ii) assess the extent to which theories of actionality account for the range of systems in these and other Bantu languages, and (iii) develop a data collection tool that can be used for comparative research on actionality not only in Bantu languages, but in languages across the world. This project will be the first comprehensive analysis of actionality in East Ruvu languages. Bringing under-documented languages to the forefront of this research area will help shape our understanding of the possibilities of human language.