The First Part of ‘Text Analysis’ is ‘Text’: Applying Digital Methods to an Under-Documented Language

Abstract

The digitization and curation of large bodies of text has inspired and encouraged new methods of research into language and literature, but only into those languages for which such corpora have been established. What sort of strategies are available to a researcher wishing to apply these research methods to a language which is not yet represented in a digitized collection? Is construction of a text corpus a feasible task for a researcher more interested in human languages than in programming languages? This talk provides a case study of the creation of a small text corpus for Tohono O’odham (an endangered language of the Southwest), the use of that corpus to investigate questions about the way that verbs are used in narratives, and more broadly, the sometimes unexpected ways in which the development of a text corpus can influence the research process.