查詢結果分析
相關文獻
- AI機器翻譯與專業譯者的語言特徵差異分析--以江戶川亂步《怪人二十面相》中文翻譯為例
- The Unity of Nominal Linking Devices
- 試析閩南語情態詞kann和kam的語意與詞類
- 一種以中文前後文關鍵字為基礎的詞彙切割方法
- From Quantitative to Qualitative Studies: Developments in Chinese Computational and Corpus Linguistics
- 從心理語言學觀點看口吃與詞類的關係
- 建構一個以共時與歷時語言研究為導向的歷史語料庫
- 英語詞類變化修辭法研究
- 客語說詞中夾用國語現象分析
- 子弟書的整理與研究世紀回顧
頁籤選單縮合
| 題 名 | AI機器翻譯與專業譯者的語言特徵差異分析--以江戶川亂步《怪人二十面相》中文翻譯為例=A Comparative Analysis of Linguistic Features between AI Machine Translation and Professional Human Translation: A Case Study of the Chinese Translations of Edogawa Rampo's Kaijin Nijyu Menso |
|---|---|
| 作 者 | 鄧敏君; | 書刊名 | 譒:語言.文學.文化 |
| 卷 期 | 9 2024.10[民113.10] |
| 頁 次 | 頁107-141 |
| 分類號 | 801.47 |
| 關鍵詞 | AI機器翻譯; 專業譯者; 大型語言模型; 詞彙豐富度; 詞類; AI machine translation; Professional translators; Large language models; LLMs; Lexical richness; Part of speech; |
| 語 文 | 中文(Chinese) |
| 中文摘要 | 本研究以江戶川亂步的《怪人二十面相》為原文探討 AI 機器翻譯與專業譯 者中文譯文的語言特徵差異。語料包含五位臺灣專業譯者的譯本與 OpenAI 大型語 言模型生成的四種譯文透過八項詞彙豐富度指標及 46 種詞類進行高階層集群分 析。結果顯示詞類較詞彙豐富度更能區分機器翻譯與專業譯者進一步以獨立樣 本 t 檢定篩選使用頻率有顯著差異的詞類發現機器翻譯較常使用功能詞專業譯 者的內容詞比例較高這反映出兩者在文本建構上的根本差異──專業譯者透過內 容詞豐富文本的語境細節機器翻譯則傾向以功能詞重組訊息。面對 AI 翻譯的崛 起本研究方法可擴大應用於人機語言特徵比較研究並為翻譯實踐與翻譯教學的 後編輯策略提供參考依據。 |
| 英文摘要 | This study explores the linguistic feature differences between AI machine translation and professional human translations of Edogawa Rampo's Kaijin Nijyu Menso. The corpus includes Chinese translations from five professional Taiwanese translators and four AI- generated translations created by OpenAI’s large language models. The study utilizes eight lexical richness indicators and 46 part-of-speech categories, employing hierarchical clustering analysis to investigate the natural grouping patterns of the nine translations. The results indicate that part-of-speech distinctions are more effective than lexical richness in distinguishing between the two groups. An independent sample t-test reveals that AI translations tend to rely more on function words, whereas human translations incorporate a greater number of content words. This highlights a fundamental difference in textual construction: human translators enrich the text with contextual details, whereas AI translations reorganize information through function words. In light of the growing presence of AI translation, the methodology of this study can be applied to further research on human-AI linguistic differences and offer valuable insights for post-editing strategies in both translation practice and pedagogy. |
本系統中英文摘要資訊取自各篇刊載內容。