AI vs. Human Communication: A Pragmatic Analysis of Implicature and Gricean Maxims in EFL Contexts

Abstract

The importance of Gricean maxims in online text communication stems from the lack of prosodic features and non-verbal cues in AI tools that may lead to digital misunderstanding. As AI tools like ChatGPT continue to be integrated into English as a Foreign Language (EFL) classrooms, learners increasingly interact with AI interlocutors whose pragmatic abilities, patterns of adherence to Gricean maxims, and implicature generation may differ fundamentally from those of human speakers. This mixed-methods explanatory study compares AI-generated and human EFL learner communication with respect to adherence to Gricean maxims and the generation and interpretation of conversational implicature. The study sample comprised 84 intermediate-to-advanced EFL undergraduates at a Saudi university who completed a Discourse Completion Task (DCT), an Implicature Interpretation Task (IIT), and a 32-item questionnaire. A standardized prompting protocol was used to prompt ChatGPT to deliver responses to the same DCT scenarios as human participants. Analysis of quantitative data indicated that the AI responses complied with Grice’s maxims of quantity and manner but exhibited productive flouting, reflecting a diminished capacity to engage in the contextually sensitive maxim flouting necessary to generate conversational implicatures. Meanwhile, EFL learners flouted the maxims more intuitively - particularly in relation and quality - generating richer implicature, but their comprehension of AI-generated implicit meaning was poorer than that of human-generated implicature. The study concludes with theoretically informed suggestions for pragmatics-integrated EFL teaching and AI tool development, and proposes adopting generative AI systems tuned to pragmatic depth in language.