HAWAIIAN MORPHEMES: IDENTIFICATION, USAGE, AND APPLICATION IN INFORMATION RETRIEVAL

Abstract

The grammar of the endangered Polynesian language, Hawaiian, has been studied and documented in print starting as early as 1838. Previous studies on Hawaiian morphology analyzed only affixes. Although the significance of roots were repeatedly acknowledged by Hawaiian grammar scholars, there are no previous studies on Hawaiian roots. The lack of a generalizable, comprehensive morphologically annotated lexicon of Hawaiian words that includes a tagged set of roots and affixes impedes the development of morpheme-based information retrieval systems and word-level pedagogical approaches. Historical documents are key for endangered language revitalization. Preservation and revitalization efforts of endangered languages have focused on documenting, digitizing, and archiving historical documents with minimal research on digital information retrieval. A morphologically annotated lexicon has the potential to improve digital information retrieval systems. In this dissertation, the first comprehensive list of Hawaiian roots is identified through the development and implementation of a semi-automated morphological analysis system for Hawaiian. Examination of the identified morphemes includes the first statistical analysis of Hawaiian roots, affixes, and their relationships, producing profiles of Hawaiian morphemes. Finally, a conceptual model for Hawaiian information retrieval optimization is outlined based on the morpheme analyses. This dissertation broadens the knowledge of Hawaiian morphology, adds to digital information retrieval techniques for endangered languages, and contributes a method for morphological annotation of low-resource languages.