AI医学典籍翻译规范探赜——以《饮膳正要》为例

Exploring Translation Norms for AI Translations in Medical Classics—A Case Study of Yinshan Zhengyao

摘要: 随着人工智能软件(AI)的兴起，其在翻译领域的应用日益受到翻译专业人士的青睐。尽管AI软件在辅助翻译实践中展现出显著的优势，但其翻译质量仍然是一个值得探讨的问题。过往的研究表明，AI软件依靠其背后庞大的数据库资源，在翻译非文学文本方面能够提供快速且高质量的输出。然而，当涉及到需要深刻理解人类情感和思想的文学、典籍类文本时，AI软件的表现往往差强人意。本研究使用ChatGPT和Kimi智能助手翻译了我国甚至是世界上最早的食疗医学专著——元朝的药膳典籍《饮膳正要》部分内容，并基于图里翻译规范理论中的“充分性”和“可接受性”对比汉学家译本与两种AI工具的译本，对三个译本的质量进行探析。旨在深入探讨AI软件在处理中医典籍文本翻译时的可借鉴性和局限性，并提出可能的改进方向。

Abstract: With the rise of artificial intelligence (AI) software, its application in the translation field has increasingly gained favor among translation professionals. Although AI software has demonstrated significant advantages in translation practice, its translation quality remains a topic worthy of discussion. Previous research has shown that, relying on its vast database resources, AI software can provide rapid and high-quality output when translating non-literary texts. However, when it comes to literary and classical texts that require a profound understanding of human emotions and thoughts, AI software often performs less satisfactorily. This study employs ChatGPT and Kimi intelligent assistant to translate certain parts of the world’s earliest dietary therapy medical classic, the Yuan Dynasty’s pharmacological diet book Yinshan Zhengyao (A Soup for the Qan-Chinese Dietary Medicine of the Mongol Era As Seen in Hu Sihui’s Yinshan Zhengyao), and compares the translations by these two AI tools with those of sinologists based on the “adequacy” and “acceptability” in Toury’s translation norm theory. The aim is to explore the quality of the three translations in depth, to investigate the advantages and limitations of AI software in translating traditional Chinese medical texts, and to suggest possible directions for improvement.