Study on the application of human-AI collaborative annotation of prompt words in the translation of poetic imageries——A case study on Human-AI collaborative translation of Wang Wei’s poems
人机协作加注提示词在诗歌意境翻译中的应用研究——以王维诗歌人机协同翻译为例

摘要

以王维部分代表性诗歌的英译为研究对象，探讨人机协同加注提示词在翻译中对诗歌意境再现的作用。通过对比分析AI直接翻译、加注提示后翻译与权威人工译本的差异后发现，加注提示词能有效补充文化背景、作者生平及诗歌主题等信息，尤其在处理高语境文化负载词、情感层次及音韵结构等方面表现出较大潜力，能够显著提升AI翻译的准确性与艺术性。此外，“AI+人工协作”模式通过迭代提示与背景注入，不仅可减少误译，还有可能生成更加贴合原文意境的译文，为古典诗歌的跨文化传播提供了一种创新路径。

Abstract

Taking the English translation of Wang Wei’s representative poems as the research object, this paper examines the role of human-AI collaborative annotation of prompt words in reproducing the poetic imageries of classical Chinese poetry. Through a comparative analysis of AI-generated translations, prompt-enhanced AI translations, and an authoritative translator’s translation, it is revealed that annotating prompt words can effectively supplement necessary information related to cultural backgrounds, the poet‘s life, and thematic content, and demonstrates significant potential, particularly in rendering high-context culturally loaded terms, conveying emotional nuances, and preserving phonological structures, thereby markedly improving the accuracy and artistry of AI-generated translations. Furthermore, based on iterative prompting and contextual enrichment, “AI + human collaboration” model not only helps reduce translation errors but also holds the potential to generate translations that can reflect the original artistic conception more faithfully, offering an innovative pathway for the cross-cultural dissemination of classical poetry.