Thesis (MA)--Stellenbosch University, 2002. / ENGLISH ABSTRACT: This study focuses on two phonological processes occurring in Venda, i.e. on palatalization and labialization. Two phonological models are applied to describe these phenomena: a traditional (linear) phonological model, the Transformational Generative model of Chomsky and Halle (1968), and a non-linear Feature Geometry model of Clements (1985). This was done in order to ascertain which model would be more effective in its coverage of these sound changes. The core concepts of each model were described and the sound system of Venda was analysed in terms of distinctive features. The application of the two models led to the conclusion that a linearly structured model seem to be more effective to account for both these phenomena in Venda than its nonlinear counterpart. / AFRIKAANSE OPSOMMING: Hierdie studie fokus op twee fonologiese prosesse in Venda, tewete palatalisasie en labialisasie. Twee fonologie modelle word toegepas om hierdie verskynsels mee te beskryf: 'n tradisionele liniêre model, die sg Transformasioneel Generatiewe model van Chomsky an Halle (1968), en 'n nie-liniêre model van Clements (1985). Die doel hiermee was om te bepaal welke model hierdie verskynsels die beste kan beskryf. Die hoofkomponente van elke model is vireers bespreek waarna die klanke van Venda ontleed is in terme van distinktiewe kenmerke. Uit die toepassing van die twee modelle het dit duidelik geblyk dat die liniêre model "n veel beter beskrywing moontlik maak van die verskynsels as die meer kontemporêre nie-liniêre model.