SOAS Research Online

Abstract

The central concern of this thesis is a group of derivational suffixes characteristic of the Bantu languages known as verbal extensions yielding such derived verbs as causative, frequentative, passive, reciprocative. The study is based on a textual corpus from Emakhuwa, a Bantu language of Mozambique, supplemented by the author's native knowledge of the language. The theoretical background is provided by Lexical Functional Grammar (Bresnan (1982)), which provides a means of relating theta roles (agent, instrument, theme etc.) to grammatical functions through the Lexical Mapping sub-theory. After exploring in chapter two the morphology of these suffixes and their suppletive relationship within the lexicon, chapter three examines the syntax of primitive verbs, classified principally as ergative, unergative and unaccusative. In this analysis "objecthood" and "restrictedness" prove difficult to establish, since object c1iticization is largely restricted to human reference, while passivization (chapter five) is applied to all roles below the highest, including roles such as time and manner, normally perceived as adverbial. Furthermore, word order is little constrained and not decisive of function. Certain constructions allow variable mapping of roles to functions and introduction of supplementary objects corresponding to co-referent patients or reason, but without morphological verbal indexation. Focus in chapters four and five is narrowed to thematic extensions adding or dropping roles. The applicative introduces a beneficiary, instrument or goal, interpreted partially according to animacy; the repeated extensions may introduce multiple roles. The causative constructions include the inductive, introducing a reason/instrument role with optional suppression of agent and/or theme. The reciprocative may have a quasi-causative reading introducing an involved but unequal participant. Uses of these extensions with the passive and stative, singly or in combination, are systematically explored. The conclusion casts doubts on the adequacy of theories relying heavily on the traditional morpho-syntactic manifestations of object.