A study on the translation performance of Chatgpt 4.0 and Deepseek Sanqu songs of yuan dynasty from the perspective of Skopos theory— A case study of tune sheep on the hill: a lament at tong pass
目的论视角下ChatGPT 4.0与DeepSeek元代散曲翻译表现研究——以《山坡羊·潼关怀古》为例

摘要

近年来，以对话式大模型为代表的人工智能技术发展迅猛，为各行各业的高效赋能提供了新方案，同时随着全球化进程不断加快，对外文化交流日益频繁，作为艺术瑰宝的中国文学作品亟待传承和发扬，这对中国文学的翻译水平提出了更高的标准和要求。本文将从翻译目的论的视角出发，以元代散曲小令《山坡羊·潼关怀古》为研究对象，对当下最先进的AI大模型ChatGPT 4.0和DeepSeek为翻译工具的译本表现展开研究，从词汇、韵律和情感三个维度进行分析和讨论，以期今后将AI大模型作为翻译工具应用于中国文学作品提供借鉴和参考。

Abstract

In recent years, with the rapid development of artificial intelligence technologies, particularly conversational large models, these technologies have provided efficient solutions across various industries. At the same time, the accelerating globalization process and increasing cultural exchanges have highlighted the need to preserve and promote Chinese literary works as artistic treasures, raising the standard for translation quality. This paper, from the perspective of Skopos theory, takes the Tune Sheep on the Hill: A Lament at Tong Pass as a case study, examining the translation performance of advanced AI models, such as ChatGPT 4.0 and DeepSeek, as translation tools. The analysis focuses on three dimensions: lexical accuracy, rhythm fidelity, and emotional resonance. The study aims to offer insights and references for the future application of AI large models in translating Chinese literary works.