Análise exploratória de conectivos em redações do/para o ENEM por meio do ChatGPT

Abstract

With the focus on teaching Portuguese, this research deals with the use of connectives in essays. The theoretical basis is based on Lexicology (Barbosa, 2009; Koch, 2010; Antunes, 2018) and Corpus Linguistics (Hunston, 2022). Our general objective is to monitor the use of connectives and seek a pattern of use in essays for/about ENEM. The specific objectives are as follows: 1) Identify the connectives recommended by 4 reference works (Antunes, 2005; Bechara, 2018; Cunha, 2017; Koch, 2010), 2) Compare the use of connectives extracted in 115 essays with maximum Enem grade and 20 essays by second-year high school students from a Full-Time State School (CEPI) in the interior of Goiás and detect what the discrepancies are. 3) Assess whether what raises the grade in this skill is linked to the frequency of use of connectives or just their presence. 4) And promote awareness about the teaching and practice of this textual genre, more specifically the cohesive elements. For this quantitative investigation, the Corpus is composed of 135 essays, 20 of which were produced by second-year high school students (EM) and 115 essays with a score of 1000 from ENEM editions between the years 2014 and 2023. For linguistic processing, we used ChatGPT (4.0), a version of AI used mainly for more robust data analysis and processing. In the ENEM essays, we found a higher average use of connectives and a preference for comparative, additive and conclusive connectives. In general, both groups used the same connectives, but with different frequencies. The data enables teachers to become aware and serves as an evidence-based reference, directly contributing to the praxiologies of textual production.