Linguística Computacional e línguas indígenas brasileiras

Abstract

We present the challenges of Computational Linguistics in processing Brazilian Indigenous languages, addressing recent advancements, the main difficulties, and the perspectives on using available technologies in the preservation and revitalization of Indigenous languages. To this end, we provide a brief introduction to the field of Computational Linguistics and a review, based on Mager et al. (2018), of the most promising tools for research. Brazilian Indigenous languages represent a diverse linguistic set, with more than 250 languages spoken by different peoples. This diversity challenges conventional approaches in Computational Linguistics, which have struggled to address the particularities of these languages, as seen in the disconnect between technological solutions and their application to Indigenous languages. The lack of investment and social and political interest in research on native languages adds to the linguistic reality. The oral tradition and the absence of orthographic systems, as well as grammatical structures distinct from those of the more widely spoken languages, complicate the creation of databases necessary for training computational models and demand methodological innovations that consider their specificities. Furthermore, the adoption of Eurocentric views and concepts exacerbates the challenges. Through this exposition, we aim to contribute to the understanding of the current situation between Computational Linguistics and Brazilian Indigenous languages and emphasize the need for closer integration between research on technologies and native languages.