Thesis (MA)--University of South Africa, 2002. / ENGLISH ABSTRACT: This study focuses on nasal strengthening and nasal assimilation processes in Tshivenda. Two phonological models are applied in order to present credible descriptions and explanations of these phenomena. After having described the core components of, respectively, the classical Transformational Generative (TG) model of Chomsky & Halle (1968), and the more contemporary feature Geometry (FG) model, analyses of the phenomena are presented. It appears as if the TG model can handle the idiosyncrasies of the language more elegantly, with a sustained high level of credibility. The FG model is unable to render the same results. / AFRIKAANSE OPSOMMING: Hierdie studie fokus op nasaalversterking en nasaalassimilasie in Tshivenda. Twee fonologiese modelle word aangewend om hierdie verskynsels te beskryf en te verklaar. Die klassieke versie van liniêre fonologie, die sogenaamde SPE model van Chomsky en Halle (1968) is eerste aan die orde gestel. Daar is gevind dat dit tot In groot mate daarin slaag om die onderskeie prosesse te verklaar. Nadat die kernkomponente van die sogenaamde FG model verduidelik is, is hierdie modelook toegepas. Verskeie gebreke in die toepassing van hierdie model het na vore gekom. Ten slotte word bevind dat die klassieke TG model die verskynsels meer omvattend kan hanteer as die FG model.