Cameroon

This thesis will explore the vowel system of Moloko, a central Chadic language spoken in the Far North province of Cameroon, Africa. An analysis of the literature on Chadic languages of the central (or Biu-Mandara) branch shows that the key issue at stake is the relationship between surface and underlying vowel inventories. These languages have a wide range of surface phonetic forms which can be reduced to a smaller number of underlying phonemic vowels. This can be explained by an analysis of two key factors. Firstly the prosodies (in the Firthian sense of nonsegmental features extending over a series of segments) of palatalisation and labialisation, which affect all vowels and some consonants in words bearing one of these prosodies at the level of the morpheme. Secondly vowel patterns, involving the presence or absence of a vowel at a particular level, whether phonetically or morphologically motivated. The analysis of Moloko in Bow (1997b and 1997c) indicated that at a certain level of abstraction, only one underlying vowel was needed in this language. The structure of the language was marked by the presence or absence of this vowel /a/, its absence being marked by the insertion of an epenthetic vowel to conform to the syllable structure of the language. An alternative analysis gives this epenthetic vowel the status of a phoneme, thus positing two phonemes in the underlying vowel inventory of the language. While both analyses succeed in accounting for the surface forms generated, the goal of the present work is to determine which analysis is preferable. The theoretical framework selected is Optimality Theory, where the ranking of universal, violable constraints gives an account of generated output forms from a given input. In this case however, rather than seeking the optimal output form for a given input, there is a choice between two input forms which give the same output, requiring the strategy of lexicon optimisation. An optimality-theoretic account of the application of prosodies is also required, before any decision can be reached on the optimal vowel inventory of the language.