Natural language processing and the Mohawk language:creating a finite state morphological parser of Mohawk formal nouns

Presented in this thesis is the design, implementation and evaluation of a finite state morphological parser for Mohawk formal nouns. Utilizing the finite state morphology software designed by Beesely and Karttunen (2003) along with three of the most comprehensive grammars for Mohawk, one from each of the major dialectal regions, a lexicon for a finite state system was created that incorporated a structure I created from cross-referencing the three sources. Since there was no formal coursework in the program providing instruction in computer programming or morphology, these skills were self-taught. In addition to the parser, a taxonomy of Mohawk prefixes and suffixes was developed for the finite state system. The challenges facing the development of Natural Language Processing tools and language-learning technologies for the Mohawk language, as a polysynthetic language are representative of the problems that other related languages in this important group experience. The research shown here demonstrates that using finite state morphology techniques that the Mohawk language’s formal nouns can be successfully described and parsed and that with further study and review that it is possible a finite state system could be expanded to include all of Mohawk nouns, verbs, and particles. This work is looking to provide greater access to the Mohawk language for the speakers and learners of the language by use in an e-dictionary and also to support projects for language tools development for other polysynthetic and under-resourced languages.