PresVoc: proposta de vocabulário contrastivo a partir de discursos de presidentes do Brasil

Abstract

This dissertation presents the development of PresVoc (http://presvoc.votec.ileel.ufu.br), a contrastive vocabulary compiled from corpora of speeches delivered by presidents of Brazil, with particular emphasis on the administrations of Luiz Inácio Lula da Silva and Jair Bolsonaro. Guided by Corpus Linguistics, Terminology, and the Communicative Theory of Terminology, the study adopts a data-driven approach that enables the systematic observation of patterns of use, recurrence, and deployment of lexical items in Brazilian political discourse. The corpus comprises speeches from Lula’s second and third terms and from Bolsonaro’s single term, allowing for a contrastive analysis across distinct political periods. All speeches used to build the corpus were retrieved from the Library of the Presidency of the Republic, the Brazilian government’s official repository of presidential speeches, ensuring the reliability and authenticity of the material analyzed. The research encompasses the collection, organization, and processing of the corpora, using WordSmith Tools 9 for lexical analysis to identify recurrent and semantically relevant units. PresVoc is based on the VoTec model (Fromm, 2007), from which it inherits a systemic structure and definitional organization principles, adapted to the specificities of Brazilian presidential political discourse. The contrastive analysis highlights different discursive strategies and ideological projects, showing how particular lexical items are mobilized to construct specific meanings across each administration. As a result, PresVoc provides a methodological and technological proposal that brings together linguistic analysis, public transparency, and political memory, offering a consultation platform that supports a critical understanding of Brazilian presidential discourse.