Linguística de corpus: o processo de transcrição da variedade da libras da grande Maceió

Abstract

This research aims to discuss Libras documentation methodologies and mechanisms, in addition to developing consistent theoretical-methodological strategies for the registration of this language, establishing and analyzing the linguistic description from the Libras documentation referring to the Maceió variety. The fundamentals of Corpus Linguistics regarding the documentation of languages, especially sign languages, are presented (Quadros and Karnopp (2004), (Crasborn and Sloetjes, 2008), (Johnston 2009), (Hereweghe and Vermeerberge 2012), (Leite and Quadros 2014), (Fenlon, Schembri, Johnston and Cormier 2015) and (Quadros 2016) Methodologically, we focused on the problems encountered in the transcription process of the Corpus de Libras - Maceió project. For this, we interviewed deaf transcribers to understand what difficulties they had during the transcription process and what strategies they developed to deal with these difficulties. In addition, we list the signs that the deaf had difficulty in trying to understand what problems could be behind these difficulties. We mapped the problems and categorized them into three types: a) social problems, linked to the social context surrounding the research; b) structural problems, which concern the structure of the Corpus project; and c) methodological, linked to the guidelines described in the manual and the use of SignBank. The methodological problems were divided into four types: modality effect problems, phonological variation, lexical variation and vocabulary limitation of the signal bank. As an alternative to solve these problems, we proposed guaranteeing the entry and permanence of deaf people in the university and in research projects, the creation of a working group for transcribers to support and validate the new signs and glosses, specific training to learn linguistic concepts that facilitate the distinction between DV (visual description), E (gestures) and stable signs, the creation of a specific code for phonological variation, namely: the addition of letters in alphabetical order for each new variant; the creation of a regional filter in SignBank to add regional lexical variants in each region, and the constant feeding of new signs in the corpus from constant meetings between project transcribers and with projects from other Brazilian states.