Multidirectional leveraging for computational morphology and language documentation and revitalization

Sylvia L.R. Schreiner, Lane Schwartz, Benjamin Hunt, Emily Chen

Research output: Contribution to journalArticlepeer-review

Abstract

St. Lawrence Island Yupik is an endangered language of the Bering Strait region. In this paper, we describe our work on Yupik jointly leveraging computational morphology and linguistic feldwork, outlining the multilayer virtuous cycle that we continue to refne in our work to document and build tools for the language. After developing a preliminary morphological analyzer from an existing pedagogical grammar of Yupik, we used it to help analyze new word forms gathered through feldwork. While in the field, we augmented the analyzer to include insights into the lexicon, phonology, and morphology of the language as they were gained during elicitation sessions and subsequent data analysis. The analyzer and other tools we have developed are improved by a corpus that continues to grow through our digitization and documentation efforts, and the computational tools in turn allow us to improve and speed those same efforts. Through this process, we have successfully identified previously undescribed lexical, morphological, and phonological processes in Yupik while simultaneously increasing the coverage of the morphological analyzer. Given the polysynthetic nature of Yupik, a high-coverage morphological analyzer is a necessary prerequisite for the development of other high-level computational tools that have been requested by the Yupik community.

Original languageEnglish (US)
Pages (from-to)69-86
Number of pages18
JournalLanguage Documentation and Conservation
Volume14
StatePublished - 2020

ASJC Scopus subject areas

  • Linguistics and Language
  • Computer Science Applications
  • Library and Information Sciences

Fingerprint Dive into the research topics of 'Multidirectional leveraging for computational morphology and language documentation and revitalization'. Together they form a unique fingerprint.

Cite this