基于翻译质量评估模型的人机法律汉英翻译比较研究——以《中华人民共和国民法典》人格权编为例

A Human-AI Comparative Study of Legal Chinese-English Translation Based on Juliane House’s TQA Model: A Case Study of the Book of Personality Rights of the Civil Code of the PRC

传播影响力
本库下载频次:
本库浏览频次:
CNKI下载频次:0

归属院系:

外语学院

作者:

洪秋锋

摘要:

在当前全球化和“一带一路”倡议背景下,跨语言交流成为推动国际合作与发展的关键。法律文本作为国家治理和国际交往的重要材料,其精确翻译至关重要。自人工智能翻译兴起后,一些学者开始对人工翻译和人工智能翻译进行比较研究,但很少对法律文本的人工翻译与人工智能翻译进行比较研究,尤其是对《中华人民共和国民法典》(以下简称《民法典》)的人工翻译和人工智能翻译进行比较研究。朱莉安·豪斯的翻译质量评估模型被广泛应用于文学作品的翻译质量评估,但很少用于法律文本的翻译质量评估研究。鉴于此,本文以朱莉安·豪斯的翻译质量评估模型为理论框架,以《民法典》人格权编的人工英译本和人工智能英译本为研究对象,旨在探讨以下两个研究问题。(1)两个英译本在语场、语旨、语式和体裁方面有何异同?(2)该研究对于法律翻译有何启示?为回答以上两个研究问题,本文运用定性与定量分析相结合的方法,对比分析了全国人大法工委的《民法典》人格权编的人工英译本(以下简称“人工译本”)和ChatGPT4.0的人工智能英译本(以下简称“人工智能译本”)。通过分析发现:(1)经过对比,人工译本的质量较高。在语场方面,人工译本和人工智能译本都保留了源文本作为立法文本的主题和主要内容。然而,人工智能译本在术语准确性和一致性方面存在一些翻译错误,所以人工翻译在原文本的准确性、精确性和一致性方面胜过人工智能翻译。因此,人工翻译在语义功能方面比人工智能翻译更为等效;在语旨方面,人工译本和人工智能译本在语用功能方面都基本实现了与原文等效;在体裁方面,人工译本和人工智能译本在媒介(两种文本均以简单的独白形式呈现)和语言衔接方式都很相似。然而,由于人工智能翻译在部分术语翻译和章节编号形式方面存在不一致,人工翻译在一致性方面胜过人工智能翻译;在体裁方面,人工译本和人工智能翻译都是以英文形式书写的中国立法文本。尽管人工智能译本在一致性和部分术语的翻译方面存在问题,但总体上保留了源文本的主题,正确地翻译了大多数术语,并且保留了立法文本的大部分特征,包括正式性、清晰性和权威性等。(2)虽然人工智能译本在术语准确度、一致性和规范性方面表现较差,但是在大部分术语、句子和语篇方面表现出立法文本的特点。因此,人工智能可以作为辅助人工从事法律翻译工作的工具。本研究对《中华人民共和国民法典》人格权编的人工智能翻译和人工翻译进行对比所得出的结论,可为其他法律文本翻译提供重要的启示意义。

语种:

英文

学科:

外国语言学及应用语言学

提交日期

2024-06-18

引用参考

洪秋锋. 基于翻译质量评估模型的人机法律汉英翻译比较研究——以《中华人民共和国民法典》人格权编为例[D]. 西南政法大学,2024.

全文附件授权许可

知识共享许可协议-署名

  • dc.title
  • 基于翻译质量评估模型的人机法律汉英翻译比较研究——以《中华人民共和国民法典》人格权编为例
  • dc.title
  • A Human-AI Comparative Study of Legal Chinese-English Translation Based on Juliane House’s TQA Model: A Case Study of the Book of Personality Rights of the Civil Code of the PRC
  • dc.contributor.schoolno
  • 20210502112014
  • dc.contributor.author
  • 洪秋锋
  • dc.contributor.affiliation
  • 外语学院
  • dc.contributor.degree
  • 硕士
  • dc.contributor.childdegree
  • 文学硕士学位
  • dc.contributor.degreeConferringInstitution
  • 西南政法大学
  • dc.identifier.year
  • 2024
  • dc.contributor.direction
  • 法律翻译理论与实践
  • dc.contributor.advisor
  • 张绍全
  • dc.contributor.advisorAffiliation
  • 外语学院
  • dc.language.iso
  • 英文
  • dc.subject
  • 朱莉安·豪斯翻译质量评估模型,《中华人民共和国民法典》,人工翻译,ChatGPT4.0,比较研究
  • dc.subject
  • Juliane House's Translation Quality Assessment Model;the Civil Code of the People’s Republic of China;human translation;ChatGPT4.0;comparative study
  • dc.description.abstract
  • 在当前全球化和“一带一路”倡议背景下,跨语言交流成为推动国际合作与发展的关键。法律文本作为国家治理和国际交往的重要材料,其精确翻译至关重要。自人工智能翻译兴起后,一些学者开始对人工翻译和人工智能翻译进行比较研究,但很少对法律文本的人工翻译与人工智能翻译进行比较研究,尤其是对《中华人民共和国民法典》(以下简称《民法典》)的人工翻译和人工智能翻译进行比较研究。朱莉安·豪斯的翻译质量评估模型被广泛应用于文学作品的翻译质量评估,但很少用于法律文本的翻译质量评估研究。鉴于此,本文以朱莉安·豪斯的翻译质量评估模型为理论框架,以《民法典》人格权编的人工英译本和人工智能英译本为研究对象,旨在探讨以下两个研究问题。(1)两个英译本在语场、语旨、语式和体裁方面有何异同?(2)该研究对于法律翻译有何启示?为回答以上两个研究问题,本文运用定性与定量分析相结合的方法,对比分析了全国人大法工委的《民法典》人格权编的人工英译本(以下简称“人工译本”)和ChatGPT4.0的人工智能英译本(以下简称“人工智能译本”)。通过分析发现:(1)经过对比,人工译本的质量较高。在语场方面,人工译本和人工智能译本都保留了源文本作为立法文本的主题和主要内容。然而,人工智能译本在术语准确性和一致性方面存在一些翻译错误,所以人工翻译在原文本的准确性、精确性和一致性方面胜过人工智能翻译。因此,人工翻译在语义功能方面比人工智能翻译更为等效;在语旨方面,人工译本和人工智能译本在语用功能方面都基本实现了与原文等效;在体裁方面,人工译本和人工智能译本在媒介(两种文本均以简单的独白形式呈现)和语言衔接方式都很相似。然而,由于人工智能翻译在部分术语翻译和章节编号形式方面存在不一致,人工翻译在一致性方面胜过人工智能翻译;在体裁方面,人工译本和人工智能翻译都是以英文形式书写的中国立法文本。尽管人工智能译本在一致性和部分术语的翻译方面存在问题,但总体上保留了源文本的主题,正确地翻译了大多数术语,并且保留了立法文本的大部分特征,包括正式性、清晰性和权威性等。(2)虽然人工智能译本在术语准确度、一致性和规范性方面表现较差,但是在大部分术语、句子和语篇方面表现出立法文本的特点。因此,人工智能可以作为辅助人工从事法律翻译工作的工具。本研究对《中华人民共和国民法典》人格权编的人工智能翻译和人工翻译进行对比所得出的结论,可为其他法律文本翻译提供重要的启示意义。
  • dc.description.abstract
  • In the context of globalization and the Belt and Road Initiative(BRI), cross-linguistic communication has become a key driver of international cooperation and development. Legal texts, as pivotal materials for national governance and international communication, demand precise translation. With the rise using of artificial intelligence(AI) in translation, some scholars have begun to conduct the comparative studies between the human and AI translation. However, few comparative studies between the human and AI translation of legal texts have been conducted, especially the human and AI translation of the Civil Code of the People’s Republic of China (hereinafter referred to as the the Civil Code). House’s TQA model has been applied to assessing the translation quality of lots of literary works, but this model has been rarely applied to assessing the translation quality of legal texts. In light of this, the present thesis, with Juliane House’s Translation Quality Assessment (TQA) model as the theoretical framework and the human and AI English translations of the Book of Personality Rights of the Civil Code as the research object, attempts to answer the following two research questions. (1) What are the similarities and differences between the two English translations in terms of Field, Tenor, Mode and Genre?(2)What insights does this study provide for the field of legal translation?To answer these two research questions, this thesis, with the combination of qualitative and quantitative methods, conducts the comparative study between the two English translation versions of the Book of Personality Rights of the Civil Code. Of the two, one is the translation by the Legislative Affairs Committee of the National People’s Congress (hereinafter referred to as the human translation) and the other is the translation by ChatGPT 4.0 (hereinafter referred to as the AI translation). After the analysis, this study obtains the following findings: (1) In terms of the Field, both the human translation and the AI translation maintain the legislative subject and main content of the source text. However, due to some translation errors in term accuracy and consistency, human translation excels in these features, making it more equivalent to AI translation in semantic function; In terms of the Tenor, both the human translation and the AI translation are almost equivalent to the source text, except for Author’s Provenance. Overall, both translations achieve pragmatic function equivalence to the source text; In terms of the Mode, both the human translation and the AI translation are presented as simple monologues and use cohesive devices. However, the human translation surpasses the AI translation in consistency because in the AI translation some terms and chapter numbers are translated inconsistently; In terms of the Genre, both the human translation and the AI translation are Chinese legislative texts in English. They generally preserve the subject matter of the source text and translate most of the terms correctly, maintaining typical legislative text features such as formality, clarity and authority, despite a few errors. (2) While the AI translation shows lower accuracy, consistency and normaltivity in professional terminology use, it reflects the characteristics of legislative texts in most terms, sentences and paragraphs. Therefore, AI can serve as an auxiliary tool for human translators in legal translation work.The findings obtained by the comparative study between the human and AI translations of the Book of Personality Rights of the Civil Code can provide enlightenment to the translation of other legal texts.
  • dc.date.issued
  • 2024-06-18
  • dc.date.oralDefense
  • 2024-05-17
回到顶部