This talk aims to show how diachronic scenarios can be a powerful tool for discovering patterns of affixation, making sense of their synchronic properties, and contributing to the elusive goal of reconstructing aspects of the morphological profile of a proto-language. I will illustrate this by means of a comparative study of relative clause constructions in the Bantu languages. Agreement on relative verb forms in the Bantu languages has long been known to show an intriguing amount of variation, including in the domain of affixation (see e.g. Meeussen 1967; 1971; Nsuka-Nkutsi 1982; Henderson 2007). On the one hand, there is variation in the paradigm of agreement prefixes used on relative verb forms, traditionally described as a choice between a paradigm called verbal prefixes and a paradigm called pronominal prefixes. On the other hand, there is variation in the controller of agreement, which, still according to the literature, can be the subject of the relative verb, the relativised NP or both. There is a statistical correlation between the type of agreement controller and the position of lexical subjects preceding or following the relative verb, for which a synchronic explanation in terms of “inversion constructions” has been proposed (e.g. Henderson 2011; Marten & van der Wal 2014). I will argue that previous accounts have been hampered by reductionist approaches to comparative data (Van de Velde to appear) and that a diachronic scenario - in this case the Bantu Relative Agreement (BRA) cycle - allows us to find a much wider range of synchronic variation than was previously recognised, while making it easier to understand this variation and to predict its limits. The BRA cycle consists of three simple steps, viz. (i) the emergence of an agreeing relativiser, (ii) the integration of this relativiser in the relative verb as a prefix and (iii) its merger with a pre-existing agreement prefix (Van de Velde 2021). The BRA cycle predicts that, contrary to what is claimed in the literature, we should find the same agreement variation in subject relatives as in non-subject relatives. It also predicts the existence of relative verb forms that agree twice with the relativised NP, and that there must be languages with a dedicated paradigm of relative agreement prefixes. All of these predictions are borne out. Moreover, the last prediction (based on the third step in the cycle) provides a template for the emergence of new paradigms of agreement prefixes in other contexts too. Finally, the BRA scenario forces us to reconsider Meeussen’s (1967) reconstruction of relative clause constructions in Proto-Bantu in a way that has consequences for the morphological template of Proto-Bantu verbs, which must have allowed a smaller number of prefixes than verbs in contemporary Bantu languages (Van de Velde 2022). References Henderson, Brent. 2007. The syntax of agreement in Bantu relatives. In Frederick Hoyt, Nikki Seifert, Alexandra Teodorescu & Jessica White (eds.), Morphosyntax of underrepresented languages. (Texas Linguistics Society 9), 167–184. CSLI Publications. Henderson, Brent. 2011. Agreement, locality, and OVS in Bantu. Lingua 121(5). 742–753. Marten, Lutz & Jenneke van der Wal. 2014. A typology of Bantu subject inversion. Linguistic Variation 14(2). 318–367. Meeussen, Achille Emile. 1967. Bantu grammatical reconstructions. Africana Linguistica 3. 81–121. Meeussen, Achille Emile. 1971. Relative clauses in Bantu. Studies in African Linguistics supplement 2. 3–10. Nsuka-Nkutsi, François. 1982. Les structures fondamentales du relatif dans les langues bantoues (Annalen - Reeks IN-8° - Menswetenschappen). Tervuren: Koninklijk Museum voor Midden-Afrika. Van de Velde, Mark. 2022. Agreement on Proto-Bantu relative verb forms. In Koen Bostoen, Gilles-Maurice de Schryver, Rozenn Guérois & Sara Pacchiarotti (eds.), On reconstructing Proto-Bantu grammar (Niger-Congo Comparative Stucies), 463–492. Berlin: Language Science Press. Van de Velde, Mark L. O. to appear. Alternatives for reductionist approaches to comparative Bantu grammar. LLA – Linguistique et Langues Africaines 9(2). Van de Velde, Mark L.O. 2021. The Bantu Relative Agreement cycle. Linguistics 59(4). 981–1015. https://doi.org/10.1515/ling-2021-0113.