Towards computer-assisted word frequency studies in Northern Sotho

Preventing the erroneous omission of essential words, and selecting the ‘correct’ corpus of lexical entries for a specific dictionary are two of the lexicographer's worst nightmares! The aim of this paper is to offer a solution to both these problems by means of a computerized option for word frequency studies in African languages developed at the University of Pretoria (WFS-Project)2. The latest results obtained from an analysis of different types of data corpora exceeding 155 000 words will be interpreted for Northern Sotho and discussed in more detail. These findings enable lexicographers for African languages to commence word frequency studies in any African language without having to rediscover the wheel, and could contribute to the ultimate goal of bringing the compilation of dictionaries and word frequency studies in African languages on a par with internationally appreciated projects such as the British Lancaster-Oslo/Bergen (LOB) Corpus and its American counterpart, the Brown Corpus.