TAF PUBLISHING

In today's research, pedagogy, and learning, ChatGPT is at the forefront, facilitating users. However, its specific manners and forms of response remain enigmatic, sometimes leaving users uncertain about what informs its replies. Based on Verschueren (1999) Pragmatic Adaptation Theory, this study analyzes contextual adaptation in ChatGPT's language use. Using a pre-experimental comparative qualitative method, it purposively analyzed 28 stratified screenshots of ChatGPT conversations to examine how it adjusts its language to different prompts, how effectively it adapts to various contexts, and how it manages ambiguous or unclear inputs. The analysis identified different contextual factors—device, environment, history, user awareness, etc.—that influence ChatGPT’s responses. Findings also revealed that prompt structure and nature affect its replies. Furthermore, ChatGPT retains ongoing discussions within a session but forgets them upon session closure unless explicitly requested to remember. However, it retains user-shared information, enabling it to address users by name and tailor responses accordingly. The study highlights ChatGPT’s adaptational strengths and weaknesses and recommends that programmers address performance gaps. It is an indispensable resource for students and researchers seeking to understand ChatGPT’s adaptive mechanisms.