Shan, L., Pan, Z., & Weidman, R. (2024). Integrating Task-Based Language Teaching and Generative AI: Design, Implementation, and Evaluation of the CFLingo Platform for Chinese Learning. Journal of Technology and Chinese Language Teaching, 15(2), 1-34.
[单丽梅, 潘子龙, & Rob Weidman. (2024). 任务型语言教学与生成式人工智能的融合：CFLingo中文学习平台的设计、实施与评估. 科技与中文教学 (Journal of Technology and Chinese Language Teaching), 15(2), 1-34.]
Abstract/摘要：
Recent advancements in generative artificial intelligence (GAI) have led to the development of GAI-integrated platforms to enhance foreign language learning. However, such platforms' effective design, development, and evaluation require a robust theoretical framework. This design-based study applies task-based language teaching (TBLT)—specifically the cognition hypothesis (CH) and the triadic componential framework (TCF)—to inform the design and implementation of CFLingo, a GAI-integrated Chinese language learning platform. The study addresses three key inquiries. First, the study examines how the cognition hypothesis can inform task sequencing within the platform. By progressively increasing task complexity, the platform scaffolds learners’ cognitive load, guiding them from simpler to more challenging tasks in a structured and supportive way. Second, it explores the role of the triadic componential framework in enhancing the platform’s adaptability through prompt engineering techniques, which optimize task conditions to address learners’ varying proficiency levels and provide tailored feedback, creating opportunities for meaningful language practice. Third, the study evaluates the platform’s effectiveness through open-ended responses and interviews with 26 college students who used CFLingo over a semester. The findings reveal that task sequencing and adaptive feedback enhanced task authenticity, improved performance, and enriched the learning experience. These insights offer valuable design and instructional implications for future GAI-integrated language learning platforms. 近年来，生成式人工智能（GAI）的快速发展催生了多种旨在提升外语学习效果的GAI集成平台。然而，这类平台的有效设计、开发与评估需要一个坚实的理论框架作为支撑。本项设计型研究运用任务型语言教学（TBLT），特别是认知假说（CH）和三元成分框架（TCF），为GAI集成语言学习平台——智语学伴 (CFLingo)的设计与实施提供理论指导。研究围绕三个核心问题展开探讨。首先，研究探讨了认知假说如何指导平台中的任务序列设计。通过逐步增加任务复杂性，平台帮助学习者合理分配认知负荷，从简单任务逐渐过渡到更具挑战性的任务，实现结构化、支持性的学习进程。其次，研究探索了三元成分框架在平台适应性方面的作用。通过提示工程技术，平台优化任务条件，以适应不同水平学习者的需求，提供个性化反馈，为学习者创造有意义的语言练习机会。最后，研究通过对26名在一学期内使用CFLingo的大学生进行开放式问卷和访谈，评估了平台的有效性。结果显示，任务序列设计与适应性反馈提升了任务的真实性，改善了任务表现，并丰富了整体学习体验。本研究的结果为未来GAI集成语言学习平台的设计与教学提供了宝贵的启示与实践指导。
|
