基于语料库的《阿Q正传》英译中人类与机器翻译的文化表征比较研究

A Comparative Study on Cultural Representation in Human and Machine Translation of The True Story of Ah Q Based on Corpora

摘要: 随着人工智能的快速发展，机器翻译在语言流畅性和信息传递等方面取得了显著进步，但其在富含文化要素的文学作品翻译中的表现仍有待进一步考察。本文以鲁迅的《阿Q正传》为研究对象，基于平行语料库构建其人类经典译本和AI译本的分析框架，在奈达的文化理论分类指导下，对生态文化、物质文化、社会文化、宗教文化和语言文化要素的翻译进行比较分析。通过对文化词项的语料筛选、词频统计和实例分析，探讨《阿Q正传》的人类翻译和AI译本在文化表征上的翻译差异，并进一步分析这些差异所反映的文化意义重构能力。研究表明，人类译文在文化含义呈现、社会制度表达等方面具有更强的文化意义重构能力；AI译文则更注重语义对应和语言流畅性，在处理深层文化意义时倾向于简化复杂的文化元素。两者之间的差异并非简单体现为翻译质量的优劣，而更多表现为文化意义重构能力的层级差异。本文认为，人类翻译和AI翻译并非简单的替代关系，而更适合作为优势互补的人机协同模式，该案例也为语料库方法分析两者差异提供了实证。

Abstract: With the rapid development of artificial intelligence, machine translation has made remarkable progress in linguistic fluency and information transfer. However, its performance in translating literary works rich in cultural elements remains to be further examined. This study regards Lu Xun’s The True Story of Ah Q as research object, constructs an analytical framework based on a parallel corpus of classic human translations and AI-generated translations, and conducts a comparative analysis of the translation of ecology, material, social, religious, and linguistic cultural elements under the guidance of Nida’s cultural classification theory. By corpus searching, frequency statistics, and case analysis of cultural terms, the study explores the differences in cultural representation between human and AI translations of The True Story of Ah Q, and further analyzes the cultural meaning reconstruction reflected in these differences. The result shows that human translators demonstrate stronger capabilities in reconstructing cultural meaning, particularly in conveying cultural implications and representing social institutions. By contrast, AI-generated translations place greater emphasis on semantic correspondence and linguistic fluency, tending to simplify culturally embedded expressions when dealing with deeper cultural meanings. This paper shows that human and AI translation are not in a simple substitutional relationship; rather, they are better understood as complementary approaches within a human–AI collaborative translation framework. The case also provides empirical evidence for the application of corpus-based methods in analyzing the differences between them.