Applying Finite-State Methods to the Swahili Language

Herein, we explore the current finite-state methods that exist for analyzing English grammar and decide whether they can be applied to the Swahili language and Swahili syntactic patterns. Further, we to explore the differences between Swahili grammar and English grammar to see if it is possible to accommodate these finite-state methods to the Swahili language. In the end, the objective is to deliver a recognition device for identifying well-formed Swahili sentences and generating new Swahili strings that have their components (i.e. noun phrase, prepositional phrases, etc) enclosed in parentheses. To the best of our knowledge, no work has ever been done on Swahili language processing. Also, the structure of Swahili sentences is substantially different than that of English sentences. One example is the fact that there are no equivalent words for the English determiners (DET), “the” and “a”, in Swahili. We hope that our work can contribute to developing better Swahili to English and English to Swahili dictionaries, translators, Swahili grammar checkers, speech recognition devices, and the like. We anticipate, our work in this field can bridge the gap between native Swahili speakers and native English speakers and lead to approaches for other dialects as well.