人工翻译与机器翻译应用的对比分析——以散文翻译为例

A Comparative Analysis of the Application of Human Translation and Machine Translation—A Case Study of Prose Translation

摘要: 人工智能翻译技术的迅猛发展显著提升了文学翻译的工作效率。本研究选取散文文体作为研究对象，采用对比分析法，从词汇层面系统考察了传统人工翻译与当前主流AI翻译工具(包括有道、百度等平台)的翻译质量差异。研究发现机器翻译技术存在显著的局限性，其译文往往过于拘泥于原文句子的字面意义，导致翻译过程倾向于直接的字面对应和句子结构的机械转换，而非深入理解语境与意图；代词解析和词汇不一致，对成语、谚语的理解不到位，因此其译文不能完整准确地表达原文本的意思，这些局限是机器翻译面临的挑战。AI翻译略有优势，尤其是它支持实时互动对话功能，能够及时解答译者的疑问与困惑，从而助力译者更高效、准确地完成文本翻译工作。但在文学典故、词汇理解等方面仍有待改进。相比较而言，人工翻译(张培基、徐新磊)在翻译质量上是最高的，两位译者都能合理地调整语序并精准地翻译原文所表达的内容。从翻译时间来讲，机器翻译和AI翻译耗时短，几秒钟就能得出翻译结果；人工翻译耗时最长。因此，译员可以结合机器翻译和AI翻译的优点，进行译后编辑，从而使之成为提高自己工作效率的有力工具。

Abstract: The rapid development of artificial intelligence (AI) translation software has, to a certain extent, enhanced the efficiency of translating literary works. Taking prose as an example, this paper compares human translation with popular AI translation tools, as well as Youdao and Baidu translation software, conducting an analysis from the perspective of vocabulary translation to summarize their respective strengths and weaknesses. The research reveals significant limitations in machine translation technology, which tends to adhere overly to the literal meanings of the original sentences, resulting in a process that favors direct word-for-word correspondence and mechanical conversion of sentence structures, rather than a deep understanding of context and intention. Issues such as improper pronoun resolution, inconsistent vocabulary usage, and inadequate comprehension of idioms and proverbs contribute to translated texts that fail to fully and accurately convey the meaning of the original text. These limitations pose challenges for machine translation. AI translation shows slight advantages, particularly in its real-time interactive dialogue feature, which promptly addresses translators’ queries and concerns, facilitating more efficient and accurate text translation. However, it still requires improvement in areas like literary allusions and nuanced vocabulary understanding. In contrast, human translation (exemplified by the works of Zhang Peiji and Xu Xinlei) boasts the highest quality, with both translators adept at adjusting sentence order and precisely capturing the intended meaning of the original text. In terms of translation time, machine and AI translations are notably faster, often producing results within seconds, while human translation takes the longest. Consequently, translators can leverage the strengths of machine and AI translation by incorporating post-editing, transforming these tools into effective aids for enhancing their work efficiency.