Improved Finite-State Morphological Analysis for St. Lawrence Island Yupik Using Paradigm Function Morphology

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 2676–2684 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 2676 Improved Finite-State Morphological Analysis for St. Lawrence Island Yupik Using Paradigm Function Morphology Emily Chen, Hyunji Hayley Park, Lane Schwartz University of Illinois Urbana-Champaign Champaign, IL {echen41, hpark129, lanes}@illinois.edu Abstract St. Lawrence Island Yupik is an endangered polysynthetic language of the Bering Strait region. While conducting linguistic fieldwork between 2016 and 2019, we observed substantial support within the Yupik community for language revitalization and for resource development to support Yupik education. To that end, Chen and Schwartz (2018) implemented a finite-state morphological analyzer as a critical enabling technology for use in Yupik language education and technology. Chen and Schwartz (2018) reported a morphological analysis coverage rate of approximately 75% on a dataset of 60K Yupik tokens, leaving considerable room for improvement. In this work, we present a re-implementation of the Chen and Schwartz (2018) finite-state morphological analyzer for St. Lawrence Island Yupik that incorporates new linguistic insights; in particular, in this implementation we make use of the Paradigm Function Morphology (PFM) theory of morphology. We evaluate this new PFM-based morphological analyzer, and demonstrate that it consis- tently outperforms the existing analyzer of Chen and Schwartz (2018) with respect to accuracy and coverage rate across multiple datasets. Keywords: computational morphology, morphological analysis, linguistic resource, Yupik languages, language revitalization 1. Introduction St. Lawrence Island Yupik (hereafter Yupik) is an endan- gered polysynthetic language spoken on St. Lawrence Is- land, Alaska and on the Chukotka Peninsula of Russia. While conducting linguistic fieldwork between 2016 and 2019, we observed substantial support within the Yupik community for language revitalization and for resource de- velopment to support Yupik language education. Yupik ex- hibits highly productive derivational and inflectional mor- phology (§2.), having been observed to allow up to seven derivational morphemes per word (de Reuse, 1994, p.53). Because of this high degree of morphological complex- ity, a high-coverage morphological analyzer is a neces- sary enabling technology that is a prerequisite for the de- velopment of basic language resources and technologies such as a searchable electronic dictionary or a mobile text completion system. Chen and Schwartz (2018) presented the first such morphological analyzer for Yupik (§3.). In this work, we apply the Paradigm Function Morphology (PFM) theory of morphology (§4.) to Yupik, and present a re-implementation of the Chen and Schwartz (2018) finite- state morphological analyzer. To our knowledge, this pa- per represents the first attempt at using PFM to inform the design of a morphological analyzer for Yupik, although some previous work exists in this regard for other lan- guages, namely Lingala (Karttunen, 2003). We find that incorporating PFM into our implementation results in a higher-coverage second-generation morphological analyzer for St. Lawrence Island Yupik (§5.). 2. Yupik Language Description The most thorough sources which describe St. Lawrence Island Yupik are the Jacobson (2001) reference grammar and the Yupik-English dictionary of Badten et al. (2008). Other sources include a …

Content truncated to 500 words