An Experimental Analysis of Pragmatic Competence in Human-ChatGPT Conversations

Abstract

The main purpose to conduct this research was to evaluate the pragmatic capability of ChatGPT in human interaction and analyze how does ChatGPT, as an advanced language modal, handle linguistic features; irony, metaphor, and indirect requests during conversation with humans. Besides, this study followed a mixed method approach by conducting interactive conversation with three participants to determine the pragmatic competence of ChatGPT through speech act theory and language analysis using irony, metaphor, and indirect requests as linguistics techniques. Thus, ChatGPT has proved to work effectively in human interactions so far by understanding the linguistic context of content and generating responses accordingly yet, it appears more efficient when dealing with tone recognition and complex linguistic constructs like irony and metaphor respectively. To achieve accurate responses in different type of conversations, this research suggests a progressive approach towards the betterment of the responsive capabilities of ChatGPT by incorporating information from various linguistic datasets and users’ feedback.