Word categories in Southern Bantu

Abstract

INTRODUCTION No longer need one study a Bantu language like a cryptogram. The existence of language universals is common knowledge today. Most languages have nouns and verbs, but how they are used depends on their typology. Morphology is accepted again as a distinctive level in linguistics, cf. Halle (1973). He relates it to the lexicon, one of the components of the rule system of grammar according to Chomsky (1982: 4). As any Bantu linguist knows, the occurrence of a single free morpheme as a word is not so common in these languages as in English. A word category is usually a concatenation of morphemes. Chomsky has lately broadened the base of his theory and has brought it closer to the more general principles of grammar as practised by linguists in the past. In his latest work (1982:15) he queries the possibility that phrase markers may be represented by tree structures. This is made clear by what is said by him earlier (op. cit.:9) that in a VP there is only a head and that all other specifications are redundant: this is true for any other type of phrase. For the Bantu language this is a most important statement where so many single phrases are sentences consisting only of a verb, or rather a word belonging to the predicative category. Chomsky (1981:5) has accepted an abstract CASE system for the noun. His abstract CASE theory is developed within the framework of a government theory. CASE and 0 theory are closely related. 9 theory has to do with the thematic relationship assigned to nouns, e.g., agents of action. This does not imply that all thematic relations can be subsumed under CASE. Louw (1972) expressed doubts about this when he dealt with relational semantic features which are basically the same as the thematic relations of Chomsky. Chomsky (1982) gives a full exposition of his componential principles of grammar, but we shall make use of it only now and then, because it is still more of a linguistic hypothesis than a set of well-tested linguistic principles.