SOAS Research Online

Abstract

The thesis is a syntactic study of the Inflectional constituent (INFL) in Mundani, a Grassfields Bantu language of the Mbam-Nkam subgroup, spoken in the S.W. Province of the Republic of Cameroon. Chapter 1 briefly introduces the Mundani language and people; chapter 2 summarizes relevant aspects of the modular system of Universal Grammar known as Government and Binding (GB) theory that forms the basic theoretical framework of the study. In order to place the INFL constituent within its wider syntactic context, chapters 3 and A outline Mundani clause structure, including interrogatives, negation, and interesting deviations from basic SVO word order in Topic and Focus constructions. The INFL constituent itself is introduced in chapter 5, with two possibilities for syntactic analysis: (i) as a single split constituent; (ii) as two separate syntactic categories: TENSE projecting to TP, and AQR projecting to AgrP. The evidence favours the second approach, which is adopted as a basis for discussion. Chapters 6 and 7 detail the content and syntactic properties of the category TENSE, including the licensing of nodes in complex TENSE constituents composed of several elements; the content and syntax of the category AGP are dealt with in chapters 8 and 9. In the case of AGR, two distinct approaches can be adopted: one based on a GB account of Switch Reference (SR) languages; the other in terms of Control Theory. Although the latter offers a more satisfactory account of AGR in this instance, the fact that both approaches can be applied to the Mundani data provides insights into the parallelisms between SR and Control: notably, the binding relationships between two INFL components, the links existing between INFL, its Spec (subject) position and COMP, and the obligatory subject control resulting from these relationships. The concluding chapter 10 summarizes evidence to show that INFL is not a single functional category, but rather a complex of different kinds of functional category, each of which forms the head X of its own XP projection. Two problems remain partially unresolved. Firstly, there are difficulties in accounting satisfactorily for the "spread" of imperfective marking across "complex TENSE" and a following main verb. Secondly, the proposed analysis of Mundani INFL is an obstacle to a coherent account of realis/irrealis modal marking, which falls under either TENSE or AGR across three different construction types.