A corpus-driven account of the noun classes and genders in Northern Sotho

This article offers a distributional corpus analysis of the Northern Sotho noun and gender system. The aim is twofold: first, to assess whether the existing descriptions of the noun class system in Northern Sotho are corroborated by information provided by the analysis of a large electronic corpus for this language, with specific reference to singular-plural pairings, and second, to present a number of novel visualisation aids to characterise a noun class system (in a radar diagram) and a noun gender system (using a two-directional weighted representation) for Northern Sotho in particular, and for any Bantu language in general. The findings include the discovery of two new genders in Northern Sotho (i.e. class pairs 1/6 and 3/10), and also indicate that the Northern Sotho noun class system, and by extension any one for Bantu, should be seen as dynamic.