Argumentação e (im)polidez em comentários avaliativos acadêmicos: o uso da inteligência artificial no ajuste de sentidos

Abstract

The phenomenon of (im)politeness as an argumentative strategy allows us to reflect on ways to orient a point of view (PDV) and construct meaning. Linked to the production conditions of face-threatening acts directed at the listener, this phenomenon can serve as a pathway to understanding reality, centered on an intrinsic argumentative dimension in text and discourse studies. In this study, we aim to analyze the emergence of (im)politeness in academic evaluative commentary as a mechanism for activating critical and pathos-driven PDVs, as well as the intensification of (im)politeness force, using a GPT (Generative Pre-trained Transformer) artificial intelligence model as a tool. The corpus consists of an “evaluative commentary” regarding the rejection of an academic article, a situation that involves evaluative practices and the mobilization of PDVs often marked by tensions between empathic and aggressive criticism. We begin with the description and contextualization of this genre within the academic sphere to test a chatbot, observing the strategies of (im)politeness leveling in the orientation of PDVs based on versions generated from researcher-defined prompts. The results of the analyses of these versions demonstrate the use of the tool as an innovative technique for generating derived discourses and/or for expanding or maintaining the argumentative orientation in the text.