Automatic morphosyntactic analysis of Light Warlpiri corpus data

Abstract

Morphosyntactic analysis aligns a morphosyntactic tag (‘gloss’) for each word in a given text. Manual morphosyntactic glossing requires significant time and effort to implement on a larger scale, such as for a language corpus. Computational methods of automatic analysis can aid in automating this process. In this thesis, I applied a method of automatic morphosyntactic analysis to a set of Light Warlpiri corpus data (O’Shannessy, 2005). The method used the software tool Computerised Language Analysis (MacWhinney, 2000) to apply rules-based word analysis and syntactic disambiguation to the data. My thesis will describe how this method was adapted to the morphosyntactic properties of Light Warlpiri, as well as its performance on the corpus data. Overall, the method was successfully adapted to the Light Warlpiri data, with some recurring challenges noted. Finally, the thesis will discuss the variables within the workflow that affected the adaptation of the method, with emphasis on practical considerations. Show more