Pragmatic Implicature Processing in ChatGPT

Introduction

The rapid advancement of artificial intelligence, particularly in the field of natural language

processing, has sparked considerable debate regarding the extent to which large language

models (LLMs) and LLM-driven chatbots, such as ChatGPT, resemble humans in their

language use and cognition (e.g., Chomsky, Roberts & Watumull, 2023; Piantadosi, 2023). As

these models continue to break new ground, the distinctions between human-like language

processing and AI-driven natural language understanding become increasingly nuanced,

prompting researchers to investigate whether LLMs genuinely mirror human-like language use

and cognition or merely simulate it at a superficial level (e.g., Cai et al., 2023; Mahowald et

al., 2023). The current study further examines the potential LLM-human similarities and

distinctions in terms of pragmatic use of language; specifically, we explore whether ChatGPT

resembles humans in its ability to enrich literal meanings of utterances with pragmatic

implicatures.

In recent years,