Computational Analysis of Kinyarwanda Morphology: The Morphological Alternations

For more than 30 years, there have been renewed interests in computational morphology resulting in numerous morphological tools. However the interest has always been on the politically and economically interesting languages of the world resulting in a wide language divide between the technologically rich and poor languages. Kinyarwanda language, a Bantu language spoken in East Africa is one of those underresourced languages without any language technology tools. The two most essential components of most natural language applications are a morphological analyzer and a machine-readable lexicon. These two components are still lacking for Kinyarwanda and so many other under-resourced languages. The task of developing a morphological analyzer involves two problems: the morphotactics (word formation) and the morphological alternations. In this paper we are mainly concerned with the morphological alternations.