【作 者】：
张 垒， 黄 琰， 喻安楠
【关 键 词 】：
ChatGPT； 对话式学习； 请求言语； 扎根理论； 思维活动
【栏 目】：
课程与教学
【中文摘要】：
生成式人工智能ChatGPT在教育领域蕴藏着巨大潜力，如何合理有效应用ChatGPT以支持学习却充满挑战。学生与ChatGPT交互以“对话式学习”模式展开，学生“请求言语”是对话的“驱动器”，为解开其所内含的学生思维活动“黑箱”，回答“学什么”与“怎么学”的核心问题，文章借助扎根理论与内容分析法，从“请求言语内容”“请求言语类型”“请求言语层次”三个维度构建了ChatGPT支持的对话式学习请求言语分析框架。结果表明：请求言语内容有知识、分析、成果、编辑、评价和思维六种；请求言语类型可分为提问、指令与反馈；请求言语层次有初次请求、重复请求与追加请求。文章运用编码矩阵分析了三个维度的潜在联系，进而揭示出三个维度之间相互交叉影响的复杂特征，为深入分析和有效引导学生与ChatGPT开展的对话式学习提供了更具针对性的理论参考与实践指引。
【英文摘要】：
ChatGPT, a generative artificial intelligence, has great potential in the field of education, but how to apply ChatGPT reasonably and effectively to support learning is full of challenges. The interaction between students and ChatGPT is carried out in the mode of 'dialogue-based learning', where students' "request speech" is the "driver" of the dialogue. In order to unlock the "black box" of students' thinking activities it contains and answer the core questions of "what to learn" and "how to learn", this paper, with the help of grounded theory and content analysis method, constructs a dialogue-based learning request speech analysis framework supported by ChatGPT from three dimensions of "request speech content", "request speech type" and "request speech level". The results show that there are six types of request speech content: knowledge, analysis, outcome, editing, evaluation and thinking. The types of request speech can be classified as question, instruction and feedback; and the request speech level include initial request, repeated request and additional request. This paper uses the coding matrix to analyze the potential connection of the three dimensions, and then reveals the complex characteristics of the three dimensions that cross-influence each other, which provides a more targeted theoretical reference and practical guidance for in-depth analysis and effective guidance of students' dialogue-based learning with ChatGPT.