Tonal variation in Zambian Chinyanja varieties: An autosegmental analysis.

The language varieties examined in this study are Zambian Chicewa, Chinsenga and Chingoni (abbreviated as Ce, Ns and Ng respectively throughout the thesis). The three varieties are here referred to collectively as Zambian Chinyanja. The question to be addressed is that of internal classification, to establish, how close or how different these varieties are. Particular emphasis is placed on their tonal relationships, which is a topic usually neglected in dialect studies. The model adopted for the study is that of Autosegmental Phonology originated in 1976 by Goldsmith and - with particular relevance for Bantu tonology - developed in Clements and Goldsmith (1984), a work which included contributions from a number of people working in this field. (For example: Hyman and Byarushengo, Kisseberth, Laughren, Massamba and Odden). It is found that Autosegmental Phonology has advantages over linear models of Phonology such as the traditional Generative Phonology generally so far adopted by dialectologists Such as Brown (1972), Newton (1972), and Van Vliet (1981). In contrast to linear models, Autosegmental Phonology treats tone separately from the segmental elements that bear it. This allows the model to handle tonal variation more adequately. Furthermore, the traditional models of Generative Phonology were found to be too abstract and unconstrained. In particular, in Generative Dialectology, underlying forms which were constructed tended to be so abstract that they were really historical forms having no direct justification from the contemporary systems, of the individual dialects. In this study, synchrony and diachrony are sharply distinguished in terms of the Alternation Condition (Kiparsky 1968). Although the thesis is mainly concerned with tonal variation, it includes a lexicostatistical analysis and an overview of the Phonology and Morphology of the varieties. On the basis of these preliminary analyses only, three varieties of Zambian Chinyanja would need to be recognized. Examination of tonal data, on the other hand, reveals that there are five varieties, viz; Ce1, Ce2, Ns1, Ng and Ns2. Moreover, it has been established that Ns1 and Ng are tonally identical although this is not the case when lexicostatistics are considered. The overview of the morphology is a necessary inclusion. Firstly, throughout the tonal analysis, there is substantial reference to Phonological and Morphological processes. Secondly, the description of the morphology includes accounts of Ns and Ng, which, in contrast to Ce, have not previously received any scholarly attention. It is concluded that it is significant In dialect studies of some languages to include tonal variation, for on the basis of vocabulary correspondences and other segmental information alone, the classification of language varieties or dialects may be overly simplistic.