ChatGPT affordance for logic learning strategies and its usefulness for developing knowledge and quality of logic in English argumentative writing

Abstract

Logic learning is essential for developing knowledge and quality of logic in English argumentative writing. Its main strategies include gathering logic instructions (Gathering), understanding logical concepts (Understanding), exercising knowledge of logic (Exercising), analysing logic in authentic writings (Analysing), and revising and creating logical links under guidance (Crafting) – which ChatGPT may afford. However, there has been limited exploration of ChatGPT-based logic learning. To address this gap, we developed a GPT-4-based logic learning bot and engaged 40 EFL university students, exploring ChatGPT's affordance for logic learning strategies and its usefulness for developing logical knowledge and quality of logic in English argumentative writing. We measured ChatGPT affordance by analysing learner-bot conversations and evaluated the usefulness of ChatGPT affordance through pre-post knowledge tests and essay writing tasks. Semi-structured interviews were also conducted. Our findings revealed that ChatGPT afforded 12 sub-strategies of logic learning, with Gathering and Exercising strategies being the most frequently used. ChatGPT-based logic learning strategies, especially Gathering and Exercising strategies, significantly developed knowledge and quality of logic in English argumentative writing. Through analysing the results, we identified ChatGPT features influencing the frequency and effectiveness of logical learning strategies and offered implications for implementing ChatGPT-based logic learning.