The semantics of Bantu noun classification: a review and comparison of three approaches

In this paper I review and compare three studies of the semantics of Bantu noun classification. The three studies have been selected mainly on account of the different perspectives they take. The first is Richardson (1967), widely cited as an example of the position that class allocation is arbitrary, and often deemed unworthy of serious consideration for that reason. It is an important study, however, because it takes a broader view, looking not just at an isolated group of words that according to a dictionary belong to a certain class but also to some other language structures relevant to noun classification. The second study is a paper by Palmer & Woodman (2000) on Shona noun class 3. As can be expected from Palmer (cf. Palmer 1996), it has a decidedly cultural linguistic outlook. It is based on dictionary data and introduces the notion of polycentric categories to account for the semantic diversity found. The third study to be reviewed here is a study by Selvik (2001), which analyses three Setswana noun classes as polysemous categories in a Langackerian fashion. Selvik’s paper, in approach for a large part similar to that of Palmer & Woodman, is especially interesting because of the results of a psycholinguistic experiment she carried out to test the predictions generated by her analysis.