The Pragmatics of Emotion in Artificial Intelligence: Theoretical and Applied Perspectives
Abstract
This Special Issue of Corpus Pragmatics explores how emotion shapes and is shaped by language in interaction with artificial intelligence.As human communication increasingly involves intelligent systems, emotion emerges as both a linguistic signal and a computational construct.Based on Romero-Trillo's (2008) vision of corpus pragmatics as a bridge between quantitative and qualitative inquiry, this issue examines how affect is expressed, modelled, and negotiated across natural languagebased interfaces.This methodological flexibility is particularly suited to address two demands of current research: the need to handle large amounts of linguistic data and the need to preserve contextual and interactional detail, essential when studying AI-mediated discourse, where billions of words generated by humans and machines coexist and where pragmatic meaning often resides in subtle affective and inferential processes.Language is central to human adaptation.It is through language that we negotiate meaning, express emotion, and build the social structures that sustain interaction.Verschueren's (1999) stance was that using language is a continuous process of making linguistic choices.In this sense, Silva and Mey (2021) coincide in assuming that pragmatics "studies the use of language in human communication as determined by the conditions of society" (Mey, 2001, p. 6), and take adaptability to be "a crucial concept for understanding the social bond that makes language possible" (Silva & Mey, 2021, p. 2).In the early twenty-first century, language has also become the principal medium through which humans interact with artificial intelligence.Voice assistants, chatbots, and large language models now populate the communicative landscape, transforming how we use language and how language is produced and interpreted.In this new environment, emotion has become a modelled object as well as a modelling agent.The pragmatics of emotion in artificial intelligence therefore explores
// Source
Authors: Eva M. Mestre-Mestre
Institutions: Universitat Politècnica de València