Presentation: A synchronic approach to the Chichewa nominal classification system

November 30, 2016
Faculty of Humanities
Senior Common Room Lounge
December 8, 2016 - 3:00 pm
December 8, 2016 - 4:00 pm


In this presentation, I intend to propose a new synchronic approach to the one and half century old problem of the Bantu noun classification system. Nominal classification is a widespread grammatical feature estimated to be found in approximately 44% of world languages (Corbett, 2013: 127). I will discuss this feature as it is manifested in Bantu languages, with specific focus on Chichewa. The generally accepted account of the Bantu noun class system was first proposed by Wilhelm Bleek (1862, 1869) and revised by several subsequent studies such as Carl Meinhof (1899, 1906), Corbett & Mtenje (1987), etc. Despite such revisions, the Bleek-Meinhof framework has remained the standard. This general framework is generally based on a set of prefixes of what is assumed to be the antediluvian parent Bantu language (Proto-Bantu).

In this presentation, I will show that this approach does not account for a range of empirical facts from Chichewa. Working with an electronic database of Chichewa noun classes, I will show that the conventional analysis is based on a biased sample of nouns, specifically those that are obtained through derivational processes. For example, the so-called class 1 is based on agentive nouns which only contribute 14.4% of the total number of nouns in this class. I will also show that the Bleek-Meinhof framework has confounded two distinct systems of Chichewa grammar, namely nominal derivation/inflection and nominal classification. Only when these have been separated do we get a clearer picture of facts related to the two grammatical systems. The results emerging from the proposed analysis show that in Chichewa, the noun class system is mainly based on phonetic-phonological and semantic features, giving rise to a two-tiered system. The consequence of this is that the noun class system is much smaller than previously argued for. Besides making accurate predictions about noun classification, the proposed analysis is also supported by findings in Bantu language acquisition studies.

Peter Kondwani Msaka (PhD Student, Stellenbosch University)