Describing Verbs in Disjoining Writing Systems

Abstract

Many Bantu languages, especially in Southern Africa, have a writing system, where most verb morphemes preceding the verb stem and some suffixes are written as separate words. These languages have also other writing conventions, which differ from the way they are written in other related languages. These two systems are conventionally called disjoining and conjoining writing systems. Disjoining writing can be considered simply as an under-specified way of writing, but for computational description it is a challenge, especially if the system allows only continuous sequences of characters to be recognised as units of analysis. In order to reduce unnecessary ambiguity, verb morphemes should be isolated from such strings of characters that are real words.