Spellcheckers for the South African languages, Part 2: The utilisation of clusters of circumfixes

Abstract

The aim of this article is to introduce the utilisation of “clusters of circumfixes' as a new strategy to increase the lexical recall in spellcheckers for especially African languages. This strategy is particularly useful in the compilation of spellcheckers for conjunctively written languages such as those in the Nguni group, where even extensive lexica render lexical recall values of only 90%. It is not disputed that a full morphological decomposition, for instance by means of finite-state analysis, is a good strategy for obtaining improved lexical recall, especially for the Nguni group, and that it should be pursued as an excellent solution in the long run. The utilisation of clusters of circumfixes, however, offers a less comprehensive alternative or in-between strategy to increase the lexical recall. These clusters can furthermore be compiled in a short space of time and with standard word processing and database software.