Sammendrag
As compared to the more classical frameworks which have been used in the past to deal with Swahili/Kiswahili grammar, this book aims at an alternative approach. I call it "Linguistic Empirical Grammar" (LEG), whose principles are explanatory relevance, plausibility, verification, confirmability and refutation. On the basis of this principle, the book argues that classes are genetic systems and are not gender and number systems in themselves. As a system, classes operate class projection principles over string constructions ensuring that each class is distinct and non-interchangeable with other classes. Number as a system applies only to lexical members in some classes but not to other classes. For, we find at least four classes which have no number distinctions at all. This leads to the postulation of the AAN (axiom of absolute negation) on the role of number in Bantu. The book rejects translational sentences in Eurocentric grammars as demonstrations of number in Kiswahili and Bantu. The book also rejects the term noun classes since there are severy items in class systems which are not nouns, e. g. prepositions, adverbial items, predication-sentences, all of which generate class agreements like nouns. The book discussses animacy versus animacy and points out that there are inanimate terms in the so-called exclusively animate classes 1/2 in Kiswahili. Allonominality is discussed where by affixes from different classes function as allomorphs of other classes. Lexicality is also discussed as well as the problem of coordinate constructions and agreements. The book discovers that there are double subject constructions in Kiswahili which have never been thought possible in grammars. In addition, there are SOV structures in Kiswahili which confirm the tentative feeling that the Bantu languages were originally SOV before becoming SVO in word order.
Vis fullstendig beskrivelse
