人机协同下知识建构的认知参与分析模型构建及应用

Construction and Application of a Cognitive Engagement Analysis Model for Knowledge Construction in Human-Computer Collaboration

摘要: 随着数字化转型的深入发展，人机协同学习逐渐成为未来学习新样态，相关研究成为领域热点。基于解释的知识建构模型从阐释学的角度出发，将知识建构视为学习者在持续自我解释与交互解释中逐步构建知识体系的动态过程，这一观点与学习者在人机协同环境中进行互动协作的实践紧密契合。ICAP框架将学习者的学习参与行为划分为被动、主动、建构和交互四种类型，为深入理解知识建构过程中的认知参与提供了坚实的理论基础。本研究在上述理论的支撑下，构建了人机协同下知识建构的认知参与分析模型，并确定了7个可观察的显性指标。通过将该模型应用于本科课程实践，使用内容分析法对学习者的协同对话数据进行编码，探究了人机协同知识建构中的认知参与情况。研究结果表明：1) 人机协同知识建构中的认知参与水平主要为建设性及互动式；2) 采用提示语可以显著提升学习中的认知参与水平；3) 系统化的平台培训有助于学习者快速熟悉平台，进而更好地发挥平台在知识建构和认知参与方面的潜力；4) ChatGPT的生成性对话有助于提高学生在学习中的认知参与程度，增加学习投入。基于以上结论，本研究进一步提出未来研究展望。

Abstract: With the deepening of digital transformation, human-computer collaborative learning is gradually becoming a new mode of future learning, and related research has become a hotspot in the field. The interpretive knowledge construction model, from the perspective of hermeneutics, views know-ledge construction as a dynamic process in which learners gradually build their knowledge system through continuous self-interpretation and interactive interpretation. This perspective closely aligns with the practice of learners engaging in interactive collaboration in human-computer collaborative environments. The ICAP framework categorizes learners’ engagement behaviors into four types: passive, active, constructive, and interactive, providing a solid theoretical foundation for understanding cognitive engagement in the knowledge construction process. Based on these theoretical foundations, this study constructs a cognitive engagement analysis model for knowledge construction in human-computer collaboration and identifies seven observable explicit indicators. By applying this model to undergraduate course practice and using content analysis to code collaborative dialogue data, the study explores cognitive engagement in human-computer collaborative knowledge construction. The results show that: 1) Cognitive engagement in human-computer collaborative knowledge construction is mainly constructive and interactive; 2) Using prompts can significantly enhance cognitive engagement during learning; 3) Systematic platform training helps learners quickly familiarize themselves with the platform, thereby better leveraging its potential in knowledge construction and cognitive engagement; 4) Generative dialogues in ChatGPT contribute to enhancing students’ cognitive engagement in learning and increasing their investment in learning. Based on these conclusions, the study further proposes directions for future research.