FST Morphological Analyser and Generator for Mapudüngun
This work addresses a domain-specific problem for linguistic researchers and communities interested in Mapudüngun, representing an incremental application of existing finite state methods to a new language dataset.
The authors tackled the problem of computational morphological analysis and generation for Mapudüngun by developing a finite state morphological analyser and generator based on Smeets' grammar, resulting in a system capable of classifying and tagging all components in Mapuche word forms, with evaluation results presented.
Following the Mapuche grammar by Smeets, this article describes the main morphophonological aspects of Mapudüngun, explaining what triggers them and the contexts where they arise. We present a computational approach producing a finite state morphological analyser (and generator) capable of classifying and appropriately tagging all the components (roots and suffixes) that interact in a Mapuche word form. The bulk of the article focuses on presenting details about the morphology of Mapudüngun verb and its formalisation using FOMA. A system evaluation process and its results are also present in this article.