“Data is Nice:” Theoretical and pedagogical implications of an Eastern Cherokee corpus

Abstract

This paper serves as a proof of concept for the usefulness of corpus creation in Cherokee language revitalization. It details the initial collection of a digital corpus of Cherokee/English texts and enumerates how corpus material can augment contemporary language revitalization efforts rather than simply preserving language for future analysis. By collecting and analyzing corpus material, we can quickly create new classroom materials and media products, and answer deeper theoretical linguistic questions. With a large enough corpus, we can even implement machine translation systems to facilitate the production of new texts. Although the vast majority of print material in Cherokee is in the Western dialect, this corpus has focused on Eastern texts. Expanding the dataset to include both dialects, however, will allow for comparison and facilitate generalizations about the Cherokee language as a whole. A corpus of Cherokee data can answer second language learners’ questions about the structure of the language and provide patterns for more effective, targeted learning of Cherokee. It can also provide teachers with ready access to accurate representations of the language produced by native speakers. By combining documentation and technology, we can leverage the power of databases to expedite and facilitate language revitalization.