Scholarlink

<p>The current study investigates Chinese as a second language (CSL) instructors’ perceptions and attitudes toward using ChatGPT as a supportive tool for teaching Mandarin near-synonyms. Using few-shot and chain-of-thought prompting, ChatGPT-generated responses were designed to facilitate CSL learners’ acquisition of the near-synonym pairs <i>yizhi/yixiang</i> ‘always/consistently’ and <i>bimian/yifang</i> ‘to avoid/in case of’. Experienced CSL instructors evaluated the quality of the responses using Likert scale ratings across eight criteria, including vocabulary definitions, limitation explanations, situational context, example sentences, error recognition, error correction, overall accuracy, and overall comprehensibility. In addition, instructors rated their level of support for utilizing ChatGPT as an educational tool and provided written explanations for their evaluations. The results indicate that the near-synonym pair was a more consistent predictor of instructors’ ratings than the prompt-type with the higher-frequency pair <i>yizhi/yixiang</i> ‘always/consistently’ receiving more positive ratings across most criteria, suggesting that instructors’ evaluations were more influenced by the lexical properties of the near-synonym pairs rather than the prompting method. Thematic analysis further elaborated upon these results in that instructors valued accuracy, clarity, comprehensiveness, and authenticity when evaluating ChatGPT-generated explanations. Overall, the findings suggest that while ChatGPT can support the instruction of high-frequency near-synonyms, it faces challenges in explaining low-frequency items, indicating the importance of instructor guidance in AI-assisted language learning.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A study on the application of ChatGPT in teaching Mandarin near-synonyms

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月2日

查看全文
  • Anwei Yu,
  • Huichen S. Hsiao

摘要

The current study investigates Chinese as a second language (CSL) instructors’ perceptions and attitudes toward using ChatGPT as a supportive tool for teaching Mandarin near-synonyms. Using few-shot and chain-of-thought prompting, ChatGPT-generated responses were designed to facilitate CSL learners’ acquisition of the near-synonym pairs yizhi/yixiang ‘always/consistently’ and bimian/yifang ‘to avoid/in case of’. Experienced CSL instructors evaluated the quality of the responses using Likert scale ratings across eight criteria, including vocabulary definitions, limitation explanations, situational context, example sentences, error recognition, error correction, overall accuracy, and overall comprehensibility. In addition, instructors rated their level of support for utilizing ChatGPT as an educational tool and provided written explanations for their evaluations. The results indicate that the near-synonym pair was a more consistent predictor of instructors’ ratings than the prompt-type with the higher-frequency pair yizhi/yixiang ‘always/consistently’ receiving more positive ratings across most criteria, suggesting that instructors’ evaluations were more influenced by the lexical properties of the near-synonym pairs rather than the prompting method. Thematic analysis further elaborated upon these results in that instructors valued accuracy, clarity, comprehensiveness, and authenticity when evaluating ChatGPT-generated explanations. Overall, the findings suggest that while ChatGPT can support the instruction of high-frequency near-synonyms, it faces challenges in explaining low-frequency items, indicating the importance of instructor guidance in AI-assisted language learning.