A Cross-Linguistic Semantic Study with Large Language Models: Translating Core Concepts and Historical Names in <i>The Analects</i>

The Analects poses considerable challenges when translated into English, owing to the cultural specificity of its core concepts and the diverse naming conventions of historical figures. This study leverages natural language processing and large language models to create a framework for assessing cross-linguistic semantic similarity. It examines five prominent English translations of The Analects , focusing on core conceptual words and personal names. By quantitatively analyzing the semantic shifts or loss of these essential elements, the research evaluates how well the translations preserve the original meanings across languages. The findings reveal notable differences in the translation of Confucian concepts such as Junzi and Ren , as well as the names of historical figures. These discrepancies reflect the diverse translation strategies, shaped by the translators’ cultural and historical contexts. This study underscores the intricacies of translating culturally embedded terms and demonstrates the impact of translation choices on conveying Confucian philosophy accurately. It offers a systematic, quantitative approach to evaluating translation variations, providing valuable insights and guidance for improving the consistency and accuracy of future translations.