|
Modern Linguistics 2023
人工智能、机器与人工翻译的语言特征——以小说《活着》译本为例
|
Abstract:
在AI不断发展的时代,ChatGPT作为大型交互式语言模型,具有高智能化、交互性特点,被广泛用于自然语言处理,给人工翻译和机器翻译带来了冲击和挑战。本研究以中文小说《活着》为例,使用CLAWS7、WordSmith和AntConc语言研究工具,对小说《活着》的AI翻译、人工翻译与机器翻译(有道翻译)这三个译本进行研究,旨在发现三个译本的语言特征,特别是词汇特征。研究发现,1) ChatGPT翻译词汇量最大、变化大,阅读难度大,句子所承载信息最多;2) 人工翻译用词更符合英语特点,对文化负载词的翻译最准确;3) 人工翻译人名前后一致,采用拼音翻译,有道和ChatGPT采用拼音翻译或直译,前后翻译不一致。机器翻译和人工翻译是相辅相成的,两者不可能被其中一方取代,人类应借助机器来实现更高效的翻译。
In an era of continuous AI development, ChatGPT, as a large interactive language model, possesses high intelligence and interactivity characteristics, and it is widely used in natural language pro-cessing, posing both impact and challenges to human and machine translation. This research, using the Chinese novel To Live as an example, primarily explores the linguistic features of AI translation compared to human translation and machine translation (e.g., Youdao translation). The study found that: 1) ChatGPT translation has the richest vocabulary and variation, making it more challenging to read, while it carries the most information within sentences; 2) Human translation uses words more in line with English characteristics and provides the most accurate translations for culturally loaded terms; 3) Human translation maintains consistency in translating names, employing phonetic translations, while Youdao and ChatGPT translations use phonetic translation or direct translation, resulting in inconsistent translations. Machine translation and human translation complement each other, and neither can entirely replace the other. Humans should leverage machines to achieve more efficient translation processes.
[1] | 胡开宝, 李翼. 机器翻译特征及其与人工翻译关系的研究[J]. 中国翻译, 2016, 37(5): 10-14. |
[2] | 冯志伟, 张灯柯, 饶高琦. 从图灵测试到ChatGPT——人机对话的里程碑及启示[J]. 语言战略研究, 2023, 8(2): 20-24. |
[3] | 胡健, 范梓锐. 机器翻译视角下的翻译本质[J]. 当代外语研究, 2023(2): 90-96. |
[4] | Popel, M., Tomkova, M., Tomek, J., Kaiser, ?., Uszkoreit, J., Bojar, O. and ?abokrtsky, Z. (2020) Transforming Machine Translation: A Deep Learning System Reaches News Translation Quality Comparable to Human Professionals. Nature Communications, 11, Article No. 4381. https://doi.org/10.1038/s41467-020-18073-9 |
[5] | 郑鑫, 陈海龙, 马玉群, 等. 融合依存句法和LSTM的神经机器翻译模型[J]. 哈尔滨理工大学学报, 2023, 28(3): 20-27. |
[6] | 胡泽林, 高翊, 李淼, 等. 基于字符级语言建模的汉蒙神经机器翻译方法研究[J]. 昆明理工大学学报(自然科学版), 2023, 48(3): 85-92. |
[7] | 仝亚辉. 翻译技术时代下的翻译主体性研究[J]. 解放军外国语学院学报, 2022, 45(6): 132-140. |
[8] | 乔晶, 李鹤元. 一种海图英语地名机器翻译方法[J]. 海洋测绘, 2022, 42(5): 73-77, 82. |
[9] | 李奉栖. 人工智能时代人机英汉翻译质量对比研究[J]. 外语界, 2022(4): 72-79. |
[10] | Almahasees, Z.M. (2017) Assessing the Translation of Google and Microsoft Bing in Translating Political Texts from Arabic into English. International Journal of Languages, Literature and Linguistics, 3, 1-4.
https://doi.org/10.18178/IJLLL.2017.3.1.100 |
[11] | 张法连. 法律翻译中的机器翻译技术刍议[J]. 外语电化教学, 2020(1): 53-58+8. |
[12] | 贺文照, 李德凤. 英语关系从句机译汉语评价——以谷歌机器翻译为例[J]. 中国科技翻译, 2019, 32(3): 30-34. |
[13] | 屈亚媛, 周玉梅. 机器翻译还是人工翻译?——浅析《黄帝内经?素问》双字格养生术语机译错误人工评测[J]. 医学争鸣, 2016, 7(4): 50-53. |
[14] | 蒋跃, 张英贤, 韩纪建. 英语被动句人机翻译语言计量特征对比——以《傲慢与偏见》译本为例[J]. 外语电化教学, 2016(3), 46-51, 63. |
[15] | University Centre for Computer Corpus Research on Language of Lancaster University (n.d.) Free CLAWS Web Tagger. https://ucrel-api.lancaster.ac.uk/claws/free.html |
[16] | 王克非. 语料库翻译学探索[M]. 上海: 上海交通大学出版社, 2012. |
[17] | Cambridge University Press & Assessment (n.d.) Get. English Grammar Today. Cambridge Dictionary.
https://dictionary.cambridge.org/us/grammar/british-grammar/get |
[18] | Pusparini, N.M.D.U., Fitriari, D.A.C., Kasni, N.W. and Susanthi, I.G.A.A.D. (2022) The Use of the Word “Get” in English. KnE Social Sciences, 2022, 383-391. https://doi.org/10.18502/kss.v7i10.11307 |
[19] | Zhou, X. and Hua, Y. (2021) Culture-Loaded Words and Transla-tion Equivalence. Theory and Practice in Language Studies, 11, 210-215. https://doi.org/10.17507/tpls.1102.14 |
[20] | Qu, W. and Li, R. (2015) Translation of Personal and Place Names from and into Chinese in Modern China: A Lexicographical History Perspective. International Journal for the Semiotics of Law, 28, 525-557.
https://doi.org/10.1007/s11196-015-9414-0 |