二零二五年第十九卷第一期Vol.19 No.1 2025ISSN-1994-4926MACAU UNIVERSITY OF SCIENCE AND TECHNOLOGYHUMANITIES AND SOCIAL SCIENCES

版權所㈲ 翻㊞必究 All rights reserved. No part of the publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying or otherwise, without the prior consent of the publisher. 《澳 門 科 技 大 ㈻ ㈻ 報》（㆟ 文 ㈳ 會 科 ㈻ 版） 第 ㈩ ㈨ 卷 第 ㆒ 期 （季 刊） JOURNAL OF MACAU UNIVERSITY OF SCIENCE AND TECHNOLOGY (HUM. & SOC. SCI.) VOL.19 NO.1 主 辦 方：澳門科技大學 Distributor：Macau University of Science and Technology 主 席：李行偉 Chairman：Lee, Joseph Hun Wei 編 輯 方：澳門科技大學學報（人文社會科學版）編輯部 Editorial：Editorial department, Macau University of Science and Technology 編 輯：曾幸麒 Editor：Chang, Henry Hang Kei 出 版 者：澳門科技大學 Publisher：Macau University of Science and Technology 地 址：澳門氹仔偉龍馬路 Address：Avienda Wai Long, Taipa, Macau 出版年月：二零二五年三月 Issued date：March, 2025 鳴 謝：澳門基金會資助出版 Acknowledgement：Publication sponsored by the Macau Foundation 聯絡電話(Phone)：(853) 8897-3932 電 郵(Email)：publication@must.edu.mo 印 刷(Print run)：200本 規 格(Size)：21cmｘ14cm 定 價(Price)：澳門幣四十元 期 刊 號(ISSN)：1994-4926 期刊網址(Website)：https://www.mustjournal.com/CN/home

澳門科技大學學報編輯委員會 !" Chief Editor !"# $%&'()*+ LEE, JOSEPH HUN WEI President, Macau University of Science and Technology ,-./01223 Fellow of the Royal Academy of Engineering #$"% Executive Editor 4 5 $%&'()6*+789:22+ PANG, CHUAN Vice-President and Dean of School of Graduate Studies, Macau University of Science and Technology "%&' Editorial Board Members !"# $%&'()*+ LEE, JOSEPH HUN WEI President, Macau University of Science and Technology ,-./01223 Fellow of the Royal Academy of Engineering ;<= $%&'()6*+ TAM, PAUL KWONG HANG Vice-President, Macau University of Science and Technology >?&)223 Member of the Hong Kong Academy of Sciences @AB $%&'()6*+7CDEF89-/GHIJKLM JIANG, ZHIHONG Vice- President and Director of State Key Laboratory of Quality Research in Chinese Medicines, Macau University of Science and Technology 4 5 $%&'()6*+789:22+ PANG, CHUAN Vice-President and Dean of School of Graduate Studies, Macau University of Science and Technology NAO $%&'())PQRSTKLM LIN, ZHIJUN Administrative Officer, Academic Accreditation Office, Macau University of Science and Technology UVW $%&'()*+XYZ[(AACSB) SO, JACKY YUK CHOW Senior Advisor to the President (AACSB), Macau University of Science and Technology \]^_ $%&'()`abc)89dd+ LUI, FRANCIS TING MING Director, The Institute of Development Economics, Macau University of Science and Technology e f $%&'()g)22+ FANG, QUAN Dean, Faculty of Law, Macau University of Science and Technology h-i $%&'()jklmnop)22+ GOH, KOK BENG Dean, Faculty Hospitality and Tourism Management, Macau University of Science and Technology

The Editorial Board List for Journal of Macau University of Science and Technology qAr $%&'()stuP)22+ ZHANG, ZHIQING Dean, Faculty Humanities and Arts, Macau University of Science and Technology qvw $%&'()-x)22+ ZHANG, HONGMING Dean, University International College, Macau University of Science and Technology N

!""#" " " Contents 論莎士比亞戲劇中的法語 France and the French Language in Shakespeare’s Plays 彭建華 Peng, Jianhua 1 當代漢字設計路徑探究 Design Routes of Contemporary Chinese Typography 譚穎妮、王雪青 55 Tan, Ying Ni; Wang, Xue Qing 規則銜接與創新： 中國內地與澳門數據傳輸 SCCs 制度建構—以橫琴深合區為切入點 Rule Convergence and Innovation: The Construction of SCCs regime for Data Transfer between Mainland China and Macau—Taking the Hengqin In-depth Cooperation Zone as the Entry Point 李浩 Li, Hao 77 Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 她們是朋友?托尼·莫裏森小說《慈悲》中的跨種族女性情誼 曹樂怡、高亞娟 121 Cao, Leyi; Gao, Yajuan

Risk Definers and Media Coverage of GM Foods: A Comparative Analysis of the People's Daily and the New York Times 轉基因食品的風險定義者和媒體報道— 《人民日報》和《紐約時報》的比較分析 王迪、李茵韻、茅知非、賀夢凡、韓子天、劉子鴿 147 Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 人工智慧聊天機器人研究：進展和趨勢的文獻計量分析 陳仁章、趙海霞 177 Chen, Renzhang; Zhao, Haixia 世漢㈻會國際㆗文教育研究專欄 (由教育部㆗外語言交流合作㆗心㈾助設立) 國際中文傳播視域下的金庸作品：共賞、共情、共話 Research on Jin Yong’s Martial Arts Works (wuxia) from the Perspective of International Chinese Communication: Shared Appreciation, Shared Empathy, Shared Dialogue 樂琦、鄭博文 217 Yue, Qi; Zheng, Bowen 隱性教學對英語母語者習得語氣詞「吧」的影響 The Effect of Implicit Instruction on the Acquisition of the Modal Particle Ba 汪敏鋒、姚鈺銀 237 Wang, Minfeng; Yao, Yuying

澳門科技大學學報(人文社會科學版)第十九卷第一期 二零二五年三月，頁 1-54 DOI: 10.58664/mustjournal.2025.03.001 1 論莎士比亞戲劇中的法語* 彭建華 （福建師範大學文學院副教授） 摘要：法蘭克方言和盎格魯—薩克森語都是源自波羅的海沿岸地區的日耳曼語方言。古法語是日耳曼語法蘭克方言與拉丁語、原初高盧—羅曼語接觸而後形成的混合語言，九世紀來自北方的諾曼人佔領塞納河下游和科唐坦半島，帶來日耳曼語諾曼方言，高盧—諾曼語是一種長期經歷了語言接觸和語言融合的拉丁語化方言。十二世紀以後法蘭克王國北部出現多伊爾語，十五世紀多伊爾語擴展到法蘭西南部地區，南部、北部高盧—諾曼語諸方言的混合產生了中古法語，十六世紀法蘭西王國出現了早期現代法語。1066 年諾曼第征服改變了英格蘭的社會文化、語言和文學，英格蘭開始多語言時期，而盎格魯—諾曼語成為宮廷語言和官方語言。拉丁語成為教會、教育的書面語言，盎格魯—薩克森語成為本土人的交流口語。盎格魯—諾曼語是古法語的一種方言。莎士比亞戲劇中，從《約翰王》到《理查三世》、《亨利八世》，包含一些法語詞語，法語語句，英語法語混合語句，甚至提出了國王英語、純正英語等早期現代英語的觀念，《李爾王》中的法蘭克國王是一種虛構的歷史假像。英格蘭約翰王到亨利六世、理查三世的時期，屬於法蘭西的多伊爾語、中古法語時期，亨利八世時期的法語屬於早期現代法語。但是在莎士比亞戲劇中沒有表現出不同時期的法語演變狀況，劇中作為外語的法語可能近似於十六世紀後期倫敦商人口中的法語，這是一種簡陋的、隨意的混合外語。作為明顯的口語法語，莎士比亞戲劇中的法語在拼讀上有較多的錯誤。 關鍵詞：盎格魯—法蘭克語言、高盧—羅曼語、盎格魯—薩克森語、多語言、混合語言 * 收稿日期：2023 年 09 月 07 日；通過日期：2025 年 02 月 17 日。此文是國家社科基金一般項目「莎士比亞戲劇的早期版本研究」（18BWW082）課題論文。

彭建華 2 France and the French Language in Shakespeare’s Plays Peng, Jianhua (Associate Professor, College of Chinese Language and Literature, Fujian Normal University) Abstract: Old French was a mixed language formed by the contact between the Germanic Frankish dialect and Latin, as well as the original Gaul Roman language. In the 9th century, the Normans from the north occupied the lower Seine River and the Cotentan Peninsula, bringing with them the Germanic Norman dialect. Gaul Norman was a Latinized dialect that had undergone long-term language contact and fusion. After the 12th century, Doyle language emerged in the northern part of the Frankish Kingdom, then expanded to the southwestern region of France in the 15th century. The blending of various Gaul Norman dialects in the southern and northern regions gave rise to the Middle French. In the 16th century, early modern French came into being. The Normandy Conquest in 1066 completely changed the social culture, language and literature of England, which began a multilingual period in England. Anglo-Norman turned out to be the court language and official language, Latin became the written language of the church and education, and Anglo Saxon the spoken language of the natives. Anglo-Norman was a dialect of Old French. In Shakespeare's plays, from King John to Richard III to Henry VIII, there are French words and sentences, English-French mixed sentences, and even such early modern English concepts as King English and pure English. The King of Frank in King Lear is a fictional historical illusion. The periods from King John of England to Henry VI to Richard III belonged to the Doyle and Middle French periods of France, while the French language of Henry VIII period belonged to early modern French. However, in Shakespeare's plays, there is no manifestation of the evolution of French in different periods. The French language used as a foreign language in the play may be similar to that spoken by London merchant population in the late 16th century, which was a crude and arbitrary mixed foreign language. As a clear example of spoken French, there are quite a few spelling errors in Shakespeare's plays. Keywords: Anglo-Frankish language; Gaul-Romance; Anglo-Saxon language; Multilingual; Hybrid language

論莎士比亞戲劇中的法語 3 一、引言 每⼀種語⾔都有⾃⼰的起源，⾃⼰的辭彙，⾃⼰的詞語形態，通常也有⾃⼰的句式—語法結構，⾃⼰的語法/語法哲學以及習慣表達⽅式。⼀般的，語⾔接觸後產⽣的語⾔混合在語法和表達⽅式更容易趨向簡化，將保留多語⾔的共同要素。由於歷史的氣候變化和北海盆地的地理變化，⼈們很難考證⼈類何時⾸次遷徙到不列顛島，由於便利的海上交通，不列顛東南部、南部⼀直與歐洲⼤陸的伊比利亞、⾼盧、比利時保持著密切聯繫。地中海世界很早就出現了古代⽂明的廣泛交流，古代埃及⽂獻記載了西元前 1000年多前的古利比亞、努⽶底亞王國，⼈們對它們瞭解甚少，1 西元前 800-700 年腓尼基⼈已經在伊比利亞南部建⽴殖民地迦太基，2 從伊比利亞到英格蘭、愛爾蘭的⼤西洋沿岸早期⽂化傳播的考古證據表明，來⾃歐洲⼤陸（或來⾃伊比利亞）的移民已經在不列顛和愛爾蘭開採⾦礦和銅礦，威爾⼠北海岸蘭杜德諾附近的⼤奧姆斯海岬有銅礦開採，在不列顛西南部的康沃爾半島開採錫礦和銅礦，這意味著青銅時代的開始。3 從原初凱爾特語（Proto-Celtic）來看，早期不列顛可能與北非迦太基的腓尼基⼈或者努⽶底亞⼈ Numidian及其⽂字有關。4 在希臘、羅⾺的古典⽂獻中，甚至在盎格魯—薩克森語、丹⿆語⽂獻中，不列顛是⼀直獨⽴於歐洲⼤陸的、多國家的區域。這是⼀個地理現實。羅⾺帝國滅亡之後歐洲形成了諸多民族國家，通⾏各種民族語⾔。國家作為⼀種政治、經濟、軍事綜合性實體，是⼀個在明確界定的領⼟上壟斷、合法使⽤武⼒的組織。中世紀歐洲封建國家是⼀種基於⼟地分封制度、嚴格的等級制度，並具有防禦制度的組織︔其中⼼是設防城市（城堡），它們分散在各戰略要地，為封地上的農民和⾃由民提供保護。城堡鄰近的⼈⼜通過婚姻或接受附庸地位與設防城市建⽴了更緊密的聯繫，領主與附庸關係特徵的是對領主效忠、佃農的法律和兵役等。羅⾺教廷、⼗字軍東征、 1 Hitchner, B. ed. A Companion to North Africa in Antiquity, (Oxford: Wiley-Blackwell Companion, 2022), 41-63. 2 Miles, R. Carthage Must Be Destroyed: The Rise and Fall of an Ancient Mediterranean Civilization, (London: Penguin, 2010), 10. 3 Ashton, N., Lewis, S., Stringer, Ch. eds. The Ancient Human Occupation of Britain, (Elsevier: Academic Press, 2011), 39-51. 4 Ball, M. J., Müller, N. The Celtic Languages, 2nd ed, (London: Routledge, 2009), 3-21.

彭建華 4 騎⼠團體作為外在於封建王國的強⼤⼒量，暗⽰歐洲內部的複雜的內在交流活動。 1066年諾曼第征服之後，英格蘭作為⼀個中世紀歐洲的封建王國，諾曼王朝的國王對歐洲⼤陸的諾曼公爵領地有宗主權，安茹王朝的國王擁有更廣⼤的法語地區，這改變了地理分隔的觀念，直到 1558年加萊被法軍攻陷並歸屬於法國，⽽英格蘭王國失去在歐洲⼤陸最後的⼀個城市，便終結了諾曼—安茹王國開始的英格蘭王國的觀念。在莎⼠比亞的 11個歷史劇中，排除了這⼀短暫傳統的歷史觀念。他寫到安茹王朝的約翰王和亨利六世（Henry VI, r. 1422-1461; 1470-1471）。因為與布列塔尼公爵亞瑟的王位繼承戰爭，1214年約翰王成為諾曼—安茹王國以來第⼀個完全失去⼤陸法語地區的英格蘭國王，1215 年因為王國內部男爵的叛亂⽽被迫簽署《⼤憲章》（the Magna Carta），並導致了 1216-1217年卡佩王國親王路易（Louis Le Lion, 1187-1226）短暫佔領了倫敦和英格蘭東南部。亨利六世在百年戰爭結束後，除開加萊，也幾乎完全失去⼤陸法語地區。雖然蘇格蘭王國是獨⽴的，都鐸王朝主導的國家觀念強調了⼀個英格蘭、威爾⼠、愛爾蘭完整統⼀的英格蘭王國。莎⼠比亞《亨利五世》鮮明地表達了英格蘭、威爾⼠、愛爾蘭和蘇格蘭共同構成獨⽴於⼤陸的民族國家（不列顛），即⼀種孤⽴的英格蘭民族觀念。這是都鐸王朝普遍接受的不列顛歷史觀念和民族⽂化觀念，它也廣泛地被歐洲⼤陸的歷史學家所接受。在莎⼠比亞戲劇的第⼀對折本（F1, 1623）中，七次提到了「歐洲」（Europe），三次使⽤希臘神話⼈物「歐羅巴」（Europa），僅有《亨利六世 第⼀部》中指歐洲⼤陸，「他們的⾎腥⾏為將使整個歐洲震動，」（Whose bloody deeds shall make all Europe quake.）。可能由於埃⽂河畔的斯特拉福德保存了羅⾺⼈遺跡，莎⼠比亞有意突出了英格蘭的羅⾺傳統。 二、不列顛的語言接觸和語言混合 從印歐語⾔學，尤其是從羅曼語語⾔學（Romance Languages）來看英格蘭⽂藝復興時期的英語、法語接觸、混合會更好⼀些，法蘭克語（原初法語）、盎格魯—薩克森語原本都是中古⽇爾曼語，⽽不必把法語、英語對⽴比較。西元前八至六世紀，伊比利亞半島包括伊比利亞⼈、巴斯塔尼⼈ Bastetani、塔爾特斯⼈ Tartessian、圖爾德塔尼⼈ Turdetani、凱爾特⼈、希臘移民和腓尼基移民。5 以弗所的⽶南德記載西 5 Semple, E. Ch. The Geography of the Mediterranean Region: Its Relation to Ancient History, (New York:

論莎士比亞戲劇中的法語 5 元前 816-813 年左右泰爾⼈建⽴迦太基（Qart-hadasht, Carthago），西元前五至三世紀赫⽶爾科（Himilco）和哈諾（Hanno）的漫長探險促進了歐洲和非洲⼤西洋沿岸的聯繫，古伊比利亞語是⼀種由於語⾔接觸⽽形成的⼀種非印歐語系的混合語⾔。6 不列顛位於古代地中海世界的邊緣，⽽最早活躍在⼤西洋沿岸的可能是來⾃北非（迦太基）的移民，可能使⽤某種古⽼的伊比利亞字母（或許是利比亞提非納字母或者腓尼基字母）。7 來⾃歐洲⼤陸的不列顛移民先後歷經史前時期、凱爾特時期、羅⾺時期、盎格魯—薩克森時期、維京時期和諾曼—安茹時期的六次語⾔接觸和語⾔混合。8 在西元前五世紀後期凱爾特⼈來到之前，不列顛已經⽣活著來⾃歐洲⼤陸的某些部族，這些渡過海峽來到不列顛的早期移民的語⾔未知。不列顛⼈可以相互交流的通⽤語⾔是在眾多移民的語⾔接觸和語⾔混合的複雜情況下逐漸形成的。9 現代考古學（Gordon Childe, Christopher Hawkes, Jacquetta Hawkes, Graham Clarke, Stuart Piggott, Gildas, Barry Cunlife）表明，西元前 4000-3500年新⽯器時代早期，英格蘭（威爾特Wiltshire、格洛斯特 Gloucestershire）和愛爾蘭各地似乎都有船隻登陸，⼤約同⼀時間在東英吉利亞、蘇格蘭東部和斯萊⼽海岸的地⽅已經建⽴了農業，這些史前時期移民語⾔未詳。10 在蒙茅斯的傑弗裏《不列顛列王紀》（1155）的傳說中，最早的不列顛⼈是來到巴爾幹—阿爾巴尼亞的特洛伊⼈或者安納托利亞⼈（Trojans, Anatolian），這可能是⼀個源⾃⼩亞細亞的凱爾特語（Gaelic, Keltic, Celtic languages）群體。11 凱爾特是⼀種具有共同信仰體系和共同語⾔的⽂化，巴裏·坎利夫認為，凱爾特語接近義⼤利的伊特魯斯坎語，除了⼀些銘⽂外，凱爾特⼈沒有完全的書寫系統。12 在萊茵河、盧瓦爾河、⾺恩河、摩澤爾河和易北河等中上游地區，西元前七世紀奧地利薩爾茨卡默古特地區的哈爾施塔特（Hallstatt）、德國南部的霍內堡（Heuneburg）成為凱爾 H. Holt and company, 1931), 284. 6 Payne, S. G. A History of Spain and Portugal, (London: University of Wisconsin Press, 1973), 2. 7 Hoyos, D. The Carthaginians, (London, New York: Routledge, 2010), 6-12. 8 Hogg, R. M. The Cambridge History of the English Language, Vol. 1: The Beginning to 1066, (Cambridge: Cambridge University Press, 1992), 409-451. 9 Dyer, J. Ancient Britain, (London, New York: Routledge, 1990), 30. 10 Hayes, A. Archaeology of the British Isles, (London: B.T. Batsford Ltd, 1995), 90. 11 Geoffrey of Monmouth, Hammer J. Historia Regum Britanniae: A Variant Version Edited from Manuscripts, (Cambridge, Massachusetts: Mediaeval Academy of America, 1951), 25-26. 12 Cunliffe, B. The Ancient Celts, (Oxford: Oxford University Press, 2018), 21-25.

彭建華 6 特⼈的中⼼居住區，西元前五世紀遷移到瑞⼠納沙泰爾地區的拉泰納（La Tène），受到阿爾卑斯山南部的希臘⼈和伊特魯斯坎⼈（Etruscan）的影響。愛奧尼亞地區⽶利都的赫卡泰烏斯（Hecataeus Milesius, Heroologia/Genealogia）較早記載西元前 494-450年羅納河⾕的凱爾特⼈（Keltoi, Celtae）。13 ⽽後凱爾特⼈向西遷移並控制了整個⾼盧地區，甚至來到了伊比利亞北部、中部和西部（加利西亞），進⽽向西越過海峽來到不列顛，即海島凱爾特⼈，他們在語⾔上區別於⼤陸凱爾特語，⼤陸凱爾特語包括山南凱爾特⽅⾔（Keltoi）和⾼盧⽅⾔（Galatae）。14 希臘阿卡迪亞的學者波利比烏斯《歷史》（Polybius’ History）記載，西元前 325年希臘航海家⾺薩利亞的⽪提阿斯（Pytheas of Massalia）第⼀次提到不列顛群島，他從盧瓦爾河⼜的科爾比洛向北航⾏，（也可能是從⾺賽出發，經過直布羅陀海峽，沿著⼤西洋沿岸向北航⾏），經過凱爾特⼈居住的布列塔尼（Armorica）和不列顛 Prettanic的肯特（坎蒂翁）、康沃爾（⾙爾裏奧）和奧爾卡斯，然後向北前往圖勒島 Thule或者斯堪的納維亞。15 斯特拉波《地理學》質疑⽪提阿斯的「海上航⾏」（On the Ocean, Peri tou Okeanou）的真實性，但確認了不列顛的地理位置。希臘⼈和羅⾺⼈最早記錄了不列顛的⽣活，這些⽂獻記載有較多的誤解。 西元前四世紀不列顛居民主要是從比利時、⾼盧移民⽽來的凱爾特⼈，他們可能給不列顛帶來了鐵冶煉技術。比利時的移民已使⽤青銅、⾦幣或有固定重量標準的鐵條。凱爾特語問題並不像⼈們想像的那麼簡單，它可能是⼀種來⾃中歐的印歐語⾔，塔西陀注意到不列顛語⾔和⾼盧語⾔的相似性，聖傑羅姆說加拉太語讓他想起了特雷維裏的⾼盧⽅⾔。16 第⼆次布匿戰爭（前 218-201年）之後，羅⾺⼈入侵⾼盧。西元前 56 年尤利烏斯·愷撒征服⾼盧的布列塔尼（Bretagne, Breton, Brittany, Armorica）。17 珍妮特・博德與科林・博德認為，西元前 55年當羅⾺軍隊⾸領尤利烏斯·愷撒的⼀⽀軍隊⾸次登陸不列顛東南海岸時，這⽚⼟地上幾乎沒有重⼤的堅固建築。18 13 Klausen, R. H. ed. Hecataei Milesii Fragmenta. Scylacis Caryandensis Periplus, (Berolini: Impensis G. Reimeri, 1831), 13. 14 MacAulay, D. The Celtic Languages, (Cambridge: Cambridge University Press, 1993), 1. 15 Cunliffe, B. The Extraordinary Voyage of Pytheas the Greek, (New York: Penguin, 2002), 93. 16 Ball, M. J., Müller, N. The Celtic Languages 2nd ed. (London: Routledge, 2009), 3-4. 17 Ireland, S. Roman Britain: A Sourcebook, (London: Routledge, 2009), 15. 18 Bord, J., Bord, C. Ancient mysteries of Britain, (London: Grafton, 1986), 120.

論莎士比亞戲劇中的法語 7 西元 42年不列顛國王⾟⽩林（Cunobelinus, Cymbeline）死後，其流亡羅⾺的長⼦阿德⽶尼烏斯（Adminius）請求羅⾺軍隊入侵不列顛，43年羅⾺皇帝克勞迪斯佔領了卡穆洛杜努姆（Camulodunum, Colchester）和倫敦，並建⽴了不列顛⾏省，於是拉丁語成為官⽅語⾔，不列顛⽅⾔開始出現了羅曼語化。410 年前後羅⾺⼈從不列顛撤離，羅⾺的正式權⼒基本結束，羅⾺⼈在英格蘭、威爾⼠和蘇格蘭留下了許多軍事堡壘、守望塔、糧倉和民⽤建築。19 羅⾺統治時代（Roman Britain）結束，短暫恢復了不列顛⽅⾔。⽽歐洲⼤陸的⾼盧地區，⾃西元前五世紀至凱撒征服⾼盧，整個⾼盧通⾏凱爾特語（Gaulish/Celtic language），羅⾺統治時期，凱爾特語出現拉丁語化，逐漸形成了⾼盧—羅曼語，即⼀種通俗拉丁語（Vulgar Latin）。尤利烏斯·愷撒《⾼盧戰記》寫道：「整個⾼盧分為三部分，其中⼀部分是比利時⼈（Belgae），另⼀部分是阿奎坦⼈（Aquitani），那些⽤⾃⼰的語⾔被稱為凱爾特（Celts）的⼈們，在我們的⾼盧⼈中，是第三部分。所有這些在語⾔、習俗和法律上都各不相同。加隆河將⾼盧⼈和阿奎坦⼈分隔開來︔⾺恩河和塞納河將他們與比利時⼈分隔開來。」20 其他羅⾺⼈（例如，塔西陀、迪奧・卡西烏斯 Dio Cassius、希羅提安 Herodian）的記載認為，不列顛和愛爾蘭的凱爾特⼈都與⾼盧—比利時（Gallia Belgica）有密切的關係。不列顛北部的喀裏多尼亞⼈（Caledonians, Caledoni，Kalēdōnes）或者⽪克特⼈（Picts），抵抗入侵的羅⾺⼈，他們說的是古布列塔尼語。21 但是不列顛可能還有從波羅的海、尼德蘭移民⽽來的原初⽇⽿曼⼈。塔西陀、康斯坦丁・克洛烏斯（Constantine Chlorus）卻暗⽰⼆者可能是⽇⽿曼⼈，「喀裏多尼亞⼈的紅褐⾊頭髮和粗壯的肢體表明他們的德意志⾎統」。22 19 Shotter, D. Roman Britain, (London, New York: Routledge, 2004), 110-112. 20 Julius Caesar. Gallic Wars, trans. by Edwards, H. J. (London: William Heinemann, 1919), 3. 21 Russell, P. An Introduction to the Celtic Languages, (London, New York: Routledge, 1995), 111-112. 22 Cornelius Tacitus. Agricola. Germania. Dialogus, trans. by Hutton, M., Peterson, W., Ogilvie, R. M., Warmington, E. H., Michael Winterbottom, (New York: The Macmillan co., 1914), 187.

彭建華 8 五至六世紀盎格魯、薩克森與朱特人遷移示意圖 *此圖為作者依據印歐語言學、古英語史繪製。 中世紀⽇⽿曼⼈對不列顛的入侵加速了語⾔接觸和語⾔混合。400-450年，曾經作為羅⾺⼈的雇傭軍，來⾃丹⿆、弗裏西亞、尼德蘭的朱特⼈、薩克森⼈、盎格魯⼈的入侵，（三世紀薩克森⼈可能居住在威悉河東部），為不列顛帶來了朱特⽅⾔、盎格魯⽅⾔和薩克森⽅⾔，盎格魯—薩克森⽅⾔是⽇⽿曼語的弗裏斯（Frisian, Frysk, Fries）⽅⾔，使⽤近似希臘字母的中期茹尼字母，盎格魯⽅⾔近似西弗裏斯語，薩克森⽅⾔近似東弗裏斯語（流⾏於現今德國下薩克森州）。吉爾達斯《不列顛的顛覆與征服》（Gildas, De excidio et conquestu Britanniae, 570）較早記載入侵者發起的可怕戰爭。彼得《英格蘭⼈教會史》（Bede, Historia ecclesiastica gentis Anglorum）記載，450 年前後來⾃丹⿆（Jutland）的朱特⼈居住在肯特郡和懷特島、漢普郡的部分地區，後來建⽴了肯特王國。23 七世紀維京⼈的語⾔是⼀種原始⽇⽿曼語的北⽅⽅⾔（Old Norse），他們使⽤近似希臘字母的中期茹尼字母（Runes）。24 隨著北海—波羅的海海上貿易路線的擴⼤， 23 Beda Venerabilis, The Ecclesiastical History of the English People, trans. by Bertram Colgrave, Judith McClure, Roger Collins, (Oxford: Oxford University Press, 1999), 27. 24 Byock, J. L. Viking Language 1: Learn Old Norse, Runes, and Icelandic Sagas, (Barnsley: Jules William

論莎士比亞戲劇中的法語 9 750年以來維京⼈的入侵為英格蘭北部地區帶來了北⽅⽅⾔，⼀些北⽅⽅⾔詞語（⼈名、地名、法律、戰爭）融入到盎格魯—薩克森語中，例如，anger, berserk, cake, call, crave, die, egg, fellow, fro, gift, give, guest, happy, hence, husband, ill, knife, law, low, meek, muck, rotten, ugly, scant, seemly, sister, score, skin, skull, sky, take, thence, they, thrall, thrive, though, thrust, till, trust, want, whence, wish等。870年以後丹⿆⼈已經在英格蘭北部定居下來，⽽後入侵英格蘭中部和南部，西薩克森（威塞克斯）國王阿爾弗雷德成功抵抗了丹⿆⼈的入侵，在不列顛南部保留了盎格魯—薩克森語⾔。980年之後英格蘭王國經受了丹⿆⼈的嚴重威脅，1013-1066 年陸續出現四位丹⿆⼈國王，即斯溫（Sweyn Forkbeard, r. 1013-1014）、卡努特（Canute, r. 1016-1035）、哈洛爾德（Harold I, Harefoot, r. 1035-1040）、哈德卡努特（Hardecanute, r. 1040-1042）。斯堪的納維亞–丹⿆語在英格蘭北部可能發⽣了更廣泛、更深入的影響。25 二、高盧–法蘭克的語言接觸和語言混合 在拉丁語化的羅曼語中，每⼀種語⾔都有⾃⼰發展的歷史，語⾔接觸與語⾔混合往往是在其發展的某個階段發⽣，至少是在特定的語⾔和時期內。比利⽜斯山脈和阿爾卑斯—孚⽇山脈形成了⾼盧西南與東部的⾃然邊界，其中部是中央⾼原，周圍是丘陵和盆地，⾼原之間通過微裂隙⽽彼此互通。西元前 500-300 年凱爾特⼈遷移到⾼盧，取代了原初定居於此的部族，伊比利亞⼈（Iberians）被驅逐到西南部，利古裏亞⼈（Ligurians）被驅逐到東南部。羅⾺征服⾼盧之後，⾼盧—羅曼語是凱爾特語漸進拉丁語化的結果，⾼盧語仍是普通民眾⽇常使⽤的語⾔。六世紀之前通俗拉丁語（Vulgar, Roman rustic language）、⾼盧—羅曼語 Gallo-Roman與凱爾特語、⽇⽿曼語、希臘語（⾺賽、⼟倫、尼斯、薩沃納等）長期共存。26 ⾸先談談⾼盧—法蘭克語（Gallo-Franks）。法蘭克⽅⾔與弗裏西安語、盎格魯—薩克森⽅⾔都是萊茵河—威悉河—易北河地區鄰近部族的⽇⽿曼語⽅⾔，它們有經常性的語⾔接觸。法蘭克⼈說的是⽇⽿曼語法蘭克⽅⾔，希爾德裏克⼀世（Childeric Press, 2013), 78-79. 25 Hogg, R. M., Denison, D. A History of the English Language, (Cambridge: Cambridge University Press, 2006), 15-16. 26 Rickard, P. A History of the French Language, (London: Routledge, Unwin Hyman, 2014), 2.

彭建華 10 I, r. 458-481）建⽴薩利安法蘭克王國之後，⾼盧—法蘭克語成為⼀種拉丁語化的⽇⽿曼語法蘭克⽅⾔，作為⼀種新的混合語⾔，⾼盧—法蘭克⽅⾔深刻地轉向了拉丁語俗語，最初它是法蘭西島（Ile de France）和盧瓦爾河周邊地區居民的通⽤語⾔。27 在羅⾺帝國的歐洲東部邊境有數⼗個說不同民族語⾔的部族，法蘭克⼈是指在萊茵河、美因河、摩澤爾河、圖林根地區的⽇⽿曼部族。早期拉丁語⽂獻中沒有提到法蘭克⼈，240年之後美因茲（Mayence）的法蘭克⼈才出現在拉丁語⽂獻中。28「法蘭克⼈最初來⾃潘諾尼亞（Pannonia），最初在萊茵河畔殖民。然後，他們越過萊茵河，穿過圖林根（Thuringia）定居下來，在每個鄉村地區和每個城鎮中，從他們種族的最重要、最⾼貴的家族中選出長髮的國王。」29 沃⽪斯庫斯《奧瑞利安傳》（Vopiscus, Life of Aurelian）記載奧瑞利安（Aurelian, 215-275）在美因茲打敗入侵的的法蘭克⼈，這可能是法蘭克河畔⼈（Ripuarians）。30 289年查⾺維⼈（Chamavi）被認為是法蘭克⼈，357年薩利⼈（Salii or Salians）被認為是法蘭克⼈，尼德蘭北部海岸的弗裏斯⼈似乎常常與法蘭克⼈混淆。31 358 年法蘭克薩利⼈南遷，穿過帝國的下⽇⽿曼尼亞省 Germania Inferior，進入比利時—⾼盧的圖爾奈（Tournai），該地位於默茲河（Meuse River）與謝爾德河（Schelde River）之間。358-458年薩利⼈的語⾔逐漸實現了拉丁語化。32 法蘭克墨洛溫王朝（Merovingian dynasty, 476-750）統⼀了薩利、利普裏安和包括巴黎的整個⾼盧北部，⽽後薩利⼈的統治達到了上⽇⽿曼尼亞（Germania Superior），其都城從圖爾奈（Tournai）遷移到魯昂（Rouen）。克洛維⼀世（Clovis I, r. 509-511）後期法蘭克王國採⽤⾼盧羅曼語和天主教，法蘭克王國是⼀個多語⾔王國。33 496-1539年拉丁語⼀直是法蘭克王國宗教、法律和教育的主導語⾔與官⽅ 27 Joëlle, D., Olivier, S. L'ancien et le Moyen Français, (Paris: Presses Universitaires de France, 2012), 3. 28 Perry, W. C. The Franks, from Their First Appearance in History to the Death of King Pepin, (London: Longman, 1857), 39. 29 Gregory of Tours. A History of the Franks, trans. by Lewis Thorpe, (London: Penguin, 1974), 125. 30 Howorth, H. H. “The Ethnology of Germany, ”The Journal of the Anthropological Institute of Great Britain and Ireland, Vol.13 (1884): 213-237. 31 James, E. The Franks, (New York: Blackwell Publishers, 1988), 35. 32 Sergeant, L. The Franks, From Their Origin As a Confederacy to the Establishment of the Kingdom of France and the German Empire, (New York: G.P. Putnam's sons, 1898), 75,101. 33 Drew, K. F. The Laws of the Salian Franks, (Philadelphia: University of Pennsylvania Press, 1991), 28.

論莎士比亞戲劇中的法語 11 書⾯語⾔，也是⼀種共同的習語，允許具有或多或少個性化⽅⾔的民族之間進⾏交流，⽽各地⽅⾔或者拉丁語俗語（vulgaire, ou vernaculaire）是法蘭克王國的民眾通⽤⼜語，九世紀才出現書寫的古法語⽂獻。34 其⽇⽿曼語版本的《斯特拉斯堡誓⾔》（Serments de Strasbourg, 842）是現存最早的羅曼—法蘭克語⽂獻，表現出⾼度拉丁語化的特徵，它是加洛林王朝時期使⽤的法律語⾔的⼀個樣本，不同於 880-882 年普魯登修斯《聖尤拉利亞祈禱詩》（Aurelius Clemens Prudentius, Séquence de sainte Eulalie），後者是⼀個混合拉丁語、⾼盧—法蘭克語⽪卡第⽅⾔的讚美詩，包含複雜的句法結構和古語序模式。35 加洛林王朝（Carolingian dynasty, 750-887）時期法蘭克⽅⾔得到了極⼤的推廣，拉丁化的⾼盧—法蘭克語是國家致⼒推廣的公共事業、交往⾏為的語⾔。查理曼（Charles I, Charlemagne, r. 768-814）促進了加洛林⽂藝復興，781 年以來阿爾昆（Alcuin）、教會執事保羅（Paul Diacre）和埃⾦哈德（Eginhard）等當代最偉⼤的學者聚集在宮廷裏，查理曼建⽴免費、開放的學校，並⿎勵收集和抄寫古代⽂獻，以便提⾼拉丁語識字率，並形成新的「加洛林⼩寫字體」（Carolingian minuscule）書寫體系。36 800年強⼤⽽繁榮的加洛林國王查理曼被加冕為神聖羅⾺皇帝（À Charles Auguste, couronné par Dieu grand et pacifique empereur des Romains）。813年圖爾議會要求牧師⽤通俗羅曼語（rusticam romanam linguam）和忒奧提斯坎⽅⾔（Theotiscam）佈道，rusticam romanam linguam sive theotiscam以便民眾理解。37 莎⼠比亞《亨利五世》簡略敘述了法蘭克王國的律法，包括⼜頭流傳的《薩利法》（the Salike Law），這是⽤拉丁語寫下來的法蘭克薩利⼈的部落法典。⼗世紀⾼盧—法蘭克語⼿抄本《喬納斯⽚段》（Fragments de Jonas）是⽤北⽅⽅⾔寫作的，《基督受難記》（La passion du Christ）和《聖萊格爾⽣活》（La Vie de saint Léger）則混合了北⽅和南⽅⽅⾔。38 卡佩王朝（Capetian dynasty, 987-1328）時期，⾼盧—羅曼語有數百種形式和⽅ 34 Ayres-Bennett, W. A History of the French Language Through Texts, (London, New York: Routledge, 1995), 4. 35 Gasté, A. Les Serments de Strasbourg; Étude Historique, Critique et Philologique, (Paris: E. Belin, 1888), 19. 36 Allieres, J. Formation de la Langue Francaise, (Paris: Presses Universitaires de France, 1996), 11. 37 Peignot, G. Essai Analytique sur L'origine de la Langue Française, (Dijon: V. Lagier, 1835), 23-24, 34. 38 Devillard, E. Chrestomathie de L'ancien Français, IXe-XVe Siècles: Texte, Traduction & Glossaire, (Paris: C. Klincksieck, 1887), 6-9.

彭建華 12 ⾔，西部、北部、東北部、東部和中部的⼈們使⽤各種語⾔、⽅⾔和地區⽅⾔，如⾼盧—法蘭克語、歐西坦語（Occitanie, Occitania）、布列塔尼⽅⾔（Breton）、普羅旺斯語（Provençal）、⽪卡第⽅⾔（Picard）、瓦隆⽅⾔、勃根第⽅⾔等。39 ⼗至⼗四世紀學者和教⼠普遍使⽤拉丁語。⼗⼀至⼗四世紀⼤約以盧瓦爾河為界，法蘭克北⽅出現了基於拉丁語化的⽇⽿曼語的奧伊爾語 la langue d'oil。奧伊爾語是⼀種混合了拉丁化的⾼盧—凱爾特語、⽇⽿曼語的法蘭克⽅⾔，是通⾏的官⽅語⾔，它是社會上層、商⼈和部分城市⼈⼜的語⾔。南⽅出現了基於拉丁語化的凱爾特語的奧克語 la langue d'oc。40 腓⼒⼆世（Philippe II Auguste, r. 1180-1223）和路易八世（Louis VIII, Le Lion, r. 1223-1226）達到卡佩王朝的強盛時期，奧伊爾語成為宮廷⽂學的語⾔。 ⼗⼀世紀法國出現了⽤奧伊爾語寫作的騎⼠詩歌和宗教散⽂，例如，《羅蘭之歌》（La Chanson de Roland）是⽤奧伊爾語寫成的，保留了傳統的北⽇⽿曼語語⾳。41 在這⾸歌頌功績之詩（Le Chansons de geste）的原始語⾔中，⼀些辭彙和術語在中世紀封建價值體系中具有特別重要的意義。42 由於⼗字軍東征的財富掠奪和來⾃拜占庭的古典學者和古典⽂獻，⼗⼆世紀以巴黎、蘭斯、奧爾良等地的⼤教堂和修道院為中⼼促進了拉丁語古典⽂化（⽂學、法學、哲學和歷史學、科學）的復興，法蘭西王國北部出現了⼀些學校和⼤學，其圖書館收藏了⼤量的圖書，⼈⽂學者翻譯了希臘語和阿拉伯語的著作。43 1180-1185年巴黎的亞曆山⼤ Alexandre de Paris編撰了《亞曆山⼤傳奇》（Roman d’ Alexandre），其哥特字體的⼿抄本是根據偽卡利斯提尼斯（Pseudo-Callisthènes, L’Iter ad Paradisum）、阿爾⾙裏克·德·⽪桑松（Alberic de Briançon）、尤斯塔什（Eustache）和蘭⾙爾·勒—科爾（Lambert Le Cort）⽤⽪卡第⽅⾔、普羅旺斯⽅⾔創作的騎⼠敘事詩彙編⽽成，這些騎⼠敘事詩篇章（La prise de Defur, Voeux du paon, Li restor du paon, Le voyage d'Alexandre au paradis terrestre, Venjance Alixandre）包含了⼤量從瓦萊裏烏斯（Valerius）的拉丁語⽂本翻譯，該詩 39 Jacques, Ch. Histoire de la Langue Française, (Paris: Presses Universitaires de France, 2006), 16. 40 Jacques, Ch. Histoire de la Langue Française, 19. 41 Littré, É. Histoire de la Langue Française: Études sur les Origines, L'etymologie, La Grammaire, les Dialectes, La Versification et les Lettres au Moyen âge, (Paris: Didier, 1878), 129, 325. 42 Clédat, L. La Chanson de Roland, Traduction Archaïque et Rythmée, Accompagnée de Notes Explicatives, (Paris: E. Leroux, 1887), v-xiv. 43 Haskins, Ch. H. The Renaissance of the Twelfth Century, Cambridge, (Massachusetts: Harvard University Press, 1971), 11, 48.

論莎士比亞戲劇中的法語 13 的亞曆山⼤詩⾏標誌著古法語詩歌的成熟。44 France（法蘭西）⼀詞來源於 Francs（法蘭克），Frank⼀詞演變為 Franc，可能發⽣在法國瓦盧瓦王朝前期，法語普羅旺斯⽅⾔把 frank（拉丁語寫作 Francus, 盎格魯—薩克森語寫作 Franca）寫作 franc。1300年前後《⾺克西⽶安的懺悔》（Le regret de Maximian）⼀詩中較早使⽤了該詞，Of herte ich wes wel liȝt, Soþliche wiis and briȝt, And franc mon of honde.（MS. Maximian 157 in Anglia III. 280）。45 其次，討論有關瓦盧瓦王朝（Valois Dynasty, 1328-1589）的中古法語（1330-1500）。從 1328年瓦盧瓦伯爵腓⼒（Philip VI of Valois, r. 1328-1350）繼承了法蘭克王位，瓦盧瓦王朝從此開始。查理七世（Charles VII, r. 1422-1461）、路易⼗⼀（Louis XI of Valois, r. 1461-1483）致⼒於強化中央王權、削弱各封建領主的特權，基本統⼀了法蘭西王國，1350年之後在⾏政、司法/法庭、教育上⽀持與⿎勵使⽤母語⾼盧—法蘭克⽅⾔，尤其是向東部勃根第⽅⾔、南部普羅旺斯語、奧克語地區的強制推⾏奧伊爾語，這有益於中央王權的強化，因為拉丁化的⾼盧—法蘭克⽅⾔⼀定程度上象徵著話語者的統治地位與（優先）權⼒。46 拉丁語俗語、法蘭西奧伊爾語、奧克語的語⾔混合，羅⾺法學和希臘哲學作品被翻譯到⾼盧—法蘭克⽅⾔，形成了⼀種與拉丁語競爭的⾼盧法語 lingua gallica/franco-gallica，這標誌著中古法語的開始。47 中古法語放棄了拉丁語式的完全依賴詞語變形，古法語詞形態結構部分消失，轉⽽選擇了介詞和詞語順序，語序發⽣了根本性的改變。字母 S成為複數的標記，Y很流⾏，K 和 W 基本被淘汰，拉丁語俗語、奧伊爾語和奧克語的語⾔混合較⼤地促進了語法上的簡化。48 為了反抗神聖羅⾺帝國和羅⾺教廷的外在壓⼒，瓦盧瓦王朝的⽬標⼀直是削弱拉丁語及其在官⽅⾏為中的使⽤。1490 年 12 ⽉，查理⼗八（Charles VIII, r. 1483-1498）根據 1483之〈莫蘭法令〉（L’ordonnance de Moulins）中規定，在朗格多克的刑事調查中使⽤「法語或母語」（langage françois ou maternel）。1510年，路易⼗⼆ 44 Cruse, M. I. Illuminating the ’Roman d’Alexandre’: Oxford, Bodleian Library, MS Bodley 264: The Manuscript as Monument, (Oxford: Boydell & Brewer, 2011), 1-12. 45 Brown, C. ed. English Lyrics of the XIIIth Century, (Oxford: The Clarendon press, 1932), 97. 46 Ueltschi, K. Petite Histoire de la Langue Française, Le Chagrin du Cancre, (Paris: Imago, 2015), 58. 47 Schacht, L. De Elementis, Germanicis Potissimum, Linguae Franco-gallicae, (Berolini: F. Dümmler, 1853), 5-6. 48 Brunot, F. Histoire de la Langue Française, des Origines à 1900, (Paris: A. Colin, 1933), 45.

彭建華 14 （Louis XII, r. 1498-1515）根據 1498規定在所有成⽂法地區，刑事審判將以「（拉丁語）俗語和地區語⾔」進⾏（vulgaire et langage du païs）。1535 年，弗朗索瓦⼀世（François d’Angoulême, r. 1515-1547）根據 1515要求普羅旺斯的刑事審判「使⽤法語，或至少使⽤地區的俗語」langue vulgaire。1539年 8⽉國王弗朗索瓦⼀世在〈維耶爾—⼽特萊敕令〉（Ordonnance de Villers-Cotterêts）中要求所有法律和公證⾏為都將以法語書寫和發送。「在任何情況下都沒有將法語作為王國⼈民/公民的語⾔……它只是說，法律和公證⾏為是法語，⽽不是拉丁語。」「由於此類事情經常發⽣，關於所述判決書中包含的拉丁詞語的理解，我們現在希望，今後所有判決以及所有其他程式，無論是來⾃我們的主權法院和其他下級法院，還是來⾃登記冊、調查、合同、委託書、裁決、遺囑，以及其他任何司法⾏為和⾏為，或依賴於這些⾏為和⾏為，都將被宣佈、登記並交付給雙⽅，⽤母語（langage maternel françois），⽽不是其他語⾔」，這標誌著現代法語的開始。（Et pour ce que telles choses sont souvent advenues sur l'intelligence des mots latins contenus dans lesdits arrêts, nous voulons dorénavant que tous arrêts, ensemble toutes autres procédures, soit de nos cours souveraines et autres subalternes et inférieures, soit de registres, enquêtes, contrats, commissions, sentences, testaments, et autres quelconques actes et exploits de justice, ou qui en dépendent, soient prononcés, enregistrés et délivrés aux parties, en langage maternel et non autrement.）49 接下來，早期現代法語，尤在⼗六世紀初法國⽂藝復興的開始激發了⼈們對希臘、羅⾺古典⽂學和古典語⾔的熱情，古典⽂學⼿稿的數量迅速地增長。尼斯（Nicaea）、昂蒂布利（Antipolis）、阿格德（agatha）、⾺賽（Massaliotes）、阿維尼翁等原初希臘殖民城市為古典希臘語和古希臘⽂學提供了良好的學習條件。50 最初出現在德意志美因茨的活字印刷術傳入法國，是⼀種⾰命性的因素。紐倫堡印刷商安東・科伯格（Anton Koberger, 1440/45-1513）的國際圖書貿易擴展到法國的⾥昂、巴黎等，1470年的索邦⼤學的校長兼圖書管理員邀請三位德意志印刷商在巴黎建⽴第⼀家出版社。1480年代卡昂、魯昂、布列塔尼 Breton等地出現了早期的印刷商，布 49 Jourdan, A.J., Decrusy, A.À., Isambert, F. Recueil Générale des Anciennes Lois Françaises, Depuis l'an 420 Jusqu'à la Révolution de 1789. Tome 12, (Paris: Belin-Le-Prieur, 1828), 592-594. 50 Stevens, L. C. “How the French Humanists of the Renaissance Learned Greek”, Modern Language Association of America (PMLA), 65.2 (1950.03): 240-248.

論莎士比亞戲劇中的法語 15 列塔尼印刷的書籍在英格蘭開拓了廣⼤市場。51 1538年弗朗索瓦⼀世命令巴黎印刷商羅伯特·埃斯蒂安將其印刷的每⼀本希臘語書籍都交送皇家圖書館。1539 年羅伯特・埃斯蒂安《法語—拉丁語詞典》（Robert Estienne, Dictionaire Francoislatin）提供了⼗六世紀早期法語（vernacular作為俗語）的共時記錄。52 普遍的古典語⽂學教育和⼤量的印刷書籍有利於現代法語的統⼀，迅速發展的法語將成為宮廷、司法、⾏政、教育的通⽤語⾔。1549年法國⽂藝復興時期詩⼈杜倍雷在《保衛和發揚法蘭西語⾔》中宣導「法語的純潔運動」，提倡⼈們努⼒使拉丁語的短語和說話⽅式盡可能接近⾃然，使優雅的、⾃然的法語戰勝拉丁語和那些「被濫⽤的」法語，保護法語免受拉丁語和希臘語的侵害︔不要乞求外語，在堅持法語的拉丁語化原則之下，以任何⽅式創造新詞來⾰新現代法語。53 1635年黎塞留（Richelieu et duc de Fronsac, 1585-1642）建⽴了法蘭西學院，其中語⾔學者（⾺萊伯Malherbe、沃格拉·薩沃亞爾Vaugelas savoyard 等）的任務是編撰法語詞典，1694 年《法蘭西學院詞典》（Dictionnaire de l’Académie française）初版刊印，法語由此減少了外省⽅⾔的使⽤，1681-1685 年法語逐漸⾛上現代標準化之路。54 波旁王朝（Dynastie des Bourbons, 1589-1792）之初法蘭西並沒有出現與英格蘭之間的和平關係，⽽拉丁化的現代法語與盎格魯—諾曼語、英格蘭—奧伊爾語的差異卻更加擴⼤了。 三、英格蘭王國諾曼—安茹王朝的古法語 掌握了國家或地區權⼒的⼈們所通⾏語⾔往往為主導性之官⽅語⾔。⼀些歷史事件使得盎格魯—薩克森語和諾曼—法蘭克⽅⾔發⽣了語⾔接觸和語⾔混合。793-1042 年維京⼈多次入侵諾森布裏亞（Northumbria）、⿆西亞、東盎格魯、肯特和西 51 Booton, D. E. Publishing Networks in France in the Early Era of Print, (Lindin, New York: Routledge, 2018), 1-9. 52 Shaw, J. F. Contributions to a Study of the Printed Dictionary in France Before 1539, (Toronto: University of Toronto, 1997), 170. 53 Person, É. La Deffence et Illustration de la Langue Francoyse par Ioachim Du Bellay, (Paris: Librairie L. Cerf, 1892), 56. 54 Picoche, J. Christiane Marchello-Nizia. Histoire de la Langue Française, (Paris, Nathan: Vigdor, 1999), 30.

彭建華 16 薩克森，阿爾弗雷德成功抵抗丹⿆⼈的入侵。55 919-954 年，斯堪的納維亞⼈入侵約克郡，斯堪的納維亞—丹⿆的北⽇⽿曼語部分影響了盎格魯⽅⾔，對南部和東南部語⾔的影響較⼩。56 1066年，諾曼公爵威廉⼀世（William I, r. 1035-1087）在⿊斯廷斯戰役打敗哈羅德⼆世⽽成為英格蘭國王，通⾏於羅曼語化的諾曼—法蘭克⽅⾔成為圍繞威廉⼀世的英格蘭貴族的通⽤語⾔。57 諾曼貴族和新移民通⾏盎格魯—諾曼語，它深刻影響了南部和東南部的語⾔，但盎格魯—諾曼語最初四⼗年未能在法律或⾏政中佔有主導地位。教會中的⽅濟各會和多明尼加會繼續使⽤拉丁語、盎格魯—薩克森語。58 1066-1154 年，英格蘭王國通⾏的盎格魯—諾曼語是⼀種新的混合⽅⾔，且諾曼⽅⾔原本與斯堪的納維亞—丹⿆的北⽇⽿曼語近似。59 200-600年斯堪的納維亞⽊制或⾦屬器物和⽯碑銘⽂採⽤早期茹尼字母（futhark），保留了印歐語⾔的非重⾳母⾳。60 ⾸先談談古法語的諾曼—法蘭克⽅⾔（Norman-french, Anglo-french, Anglo-Norman）。諾曼⽅⾔是⼀種北⽇⽿曼語⽅⾔，⾼盧—諾曼語是⼀種長期經歷了語⾔接觸和語⾔融合的拉丁語化⽅⾔。61 從羅⾺統治⾼盧時期至維京時期（l’ère viking），諾曼第地區的⽅⾔經歷了凱爾特⼈、羅⾺⼈、法蘭克⼈、維京⼈等入侵所帶來的長期演變。西元前 56年尤利烏斯・愷撒征服了塞納河和歐爾河流域之間的凱爾特⼈，其語⾔逐漸出現了拉丁語化。62 486 年前後該地被法蘭克墨洛溫王國統治，進⽽混合了拉丁語化的法蘭克⽅⾔，產⽣了羅曼—法蘭克語。63 諾曼⼈（Normans, or 55 Collingwood, W. G. Scandinavian Britain, (London: Society for Promoting Christian Knowledge, 1908), 43. 56 Townend, M. Language and History in Viking Age England: Linguistic Relations between Speakers of Old Norse and Old English, (Turnhout: Brepols Publishers, 2002), 43. 57 Bates, D., Curry, A. England and Normandy in the Middle Ages, (Hambledon, London: Hambledon Press, 1994), 37. 58 Kibbee, D. A. For to Speke Frenche Trewely: The French language in England, 1000-1600. Its Status, Description and Instruction, (Amsterdam: John Benjamins Publishers, 1991), 12. 59 Cleaver, L. Writing History in the Anglo-Norman World: Manuscripts, Makers and Readers, C.1066-C.1250, (Rochester: York Medieval Press, 2018), 7. 60 De Gorog, R. P. The Scandinavian Element in French and Norman, (New York: Bookman Associates, 1958), 128. 61 Vising, J. Anglo-Norman Language and Literature, (Oxford: Oxford University Press, 1923), 8. 62 Holmes, U. History of the French Language, (New York: Farrar & Rinehart, 1938), 26. 63 De Cassagnac, A. G. Histoire des Origines de la Langue Française, (Paris: Firmin Didot Frères, Fils, et cie., 1872), 29.

論莎士比亞戲劇中的法語 17 Norsemen）是北⽅維京⼈的⼀⽀，據聖昆廷的杜多（Dudonis Decani S. Quintini, Dudone Sancti Quintini）、薩克索・格拉瑪提庫斯（Saxo Grammaticus）、弗洛多阿德（Flodoard）、羅⾙爾・費古松（Robert Ferguson）記載，820-876年塞納河⾕地和巴黎多次遭到維京⼈的襲擊和入侵，911 年法蘭克加洛林王朝的國王查理三世將魯昂周圍和塞納河河⼜的領⼟（西法蘭克王國，the Kingdom of West Francia）割讓給維京⼈⾸領羅洛（或羅伯特）（Robert/Rollo of Normandy, 860-930），⽽後諾曼⼈的領地擴展到科坦廷（Cotentin）。64 諾曼第的⼤部分移民是丹⿆⼈，西北部和塞納河河⾕也有挪威移民（比例不詳），⼆者主要說西部北⽅⽅⾔（Old West Norse）。65 同樣，以丹⿆⼈為主的入侵者在不列顛建⽴了諾森布裏亞王國（Northumbria）和東盎格魯王國。羅洛（或羅伯特）是⼀名來⾃丹⿆（Dacia）的貴族，曾率領六條船從斯堪的納維亞（Scanza）出發，前往阿爾斯特姆斯（Alstelmus, Æthelwold）統治的東盎格魯王國，⽽後在弗利西亞（Frisia）作戰，掠奪了比利時的聖希梅爾特魯德教堂（St. Himeltrude）。876年之後，他襲擊了塞納河盆地的城鎮和聖維達斯特教堂（St. Vedast），並在 900年前後佔領魯昂。66 他與維京⼈⾸領古特魯姆（Guthrum, Æthelstan, ?835–890）、丹⿆國王斯溫・福克彼爾德（Sweyn Forkbeard）是朋友，後⼆者都是丹⿆的貴族。67 960-966 年諾曼戰爭之後，新來的維京⼈掠奪了西法蘭克王國（Francie occidentale, 843-987）的西北地區，這些非天主教群體延續了北⽅⽅⾔在諾曼第的顯著地位。九至⼗世紀這些來⾃丹⿆和斯堪的納維亞的移民（即維京⼈）採⽤了早已形成的羅曼—法蘭克語，其⼤多數接受了羅⾺天主教，使得⽇⽿曼語北⽅⽅⾔與羅曼—法蘭克語發⽣了語⾔接觸與語⾔混合。68 諾曼—法蘭克語明顯受到盎格魯—撒克森的影響，尤其是採⽤了部分盎格魯—薩克森語詞語，⽽盎格魯—薩克森語不再 64 Pohl, B. Dudo of Saint-Quentin's Historia Normannorum: Tradition, Innovation and Memory, (Rochester: York Medieval Press, 2015), 130. 65 Steenstrup, J. Ch. Normandiets Historie under de Syv Første Hertuger, 911-1066, (København: A. F. Høst & Søn, 1925), 37-40. 66 Douglas, D. C. “Rollo of Normandy.” The English Historical Review, LVII.CCXXVIII, (1942.10): 417-436. 67 Howorth, H. “A Criticism of the Life of Rollo As Told By Dudo of St. Quentin.” Archaeologia, 45.2, (1880): 235-250. 68 Fouracre, P., Ganz, D. ed. Frankland: The Franks and the world of the early middle ages, (Manchester: Manchester University Press, 2012), 152.

彭建華 18 ⽤於書寫，僅僅被⼤多數本地民眾⼜頭使⽤，便逐漸趨於它的終結。69 總之，1066年諾曼征服深刻影響了諾曼⼈和英格蘭⼈之間的關係，作為⽅⾔集合的盎格魯—薩克森語顯然受到了諾曼—法蘭克語（古法語⽅⾔）的影響，諾曼—法蘭克語在英格蘭與拉丁語、盎格魯—薩克森語⼀起使⽤，在宮廷、⾏政⾏為中獲得了統治地位。70 1050-1350 年盎格魯—諾曼語成為英格蘭宮廷、法律、議會、市政的官⽅語⾔，通⾏於社會中上階層。古法語成為英格蘭貴族的語⾔，拉丁語依然是外交、教會和學習（學校、⼤學）的語⾔。71 其次談談英格蘭諾曼王朝（1066-1154）的盎格魯—諾曼⽅⾔。波西・范戴克・雪萊認為，1066-1100年英格蘭通⾏拉丁語、法語和盎格魯—薩克森語，由於維京⼈的入侵和英格蘭沒有形成統⼀的民族國家，英格蘭⼈對諾曼⼈並不完全是敵視的，甚至並不把他們看作外來者，⼆者存在不容忽視的聯繫與交往，盎格魯—薩克森語與羅曼—法蘭克語之間的差異比現在所想像的要⼩很多。72 諾曼王朝時期的英格蘭出現了多語⾔社會（拉丁語、盎格魯—諾曼語、盎格魯—薩克森語），到諾曼統治末期，百分之三⼗的英語辭彙源⾃法語。⽽凱爾特語⽅⾔仍通⾏蘇格蘭、威爾⼠和康沃爾等地。73 盎格魯—諾曼語（Anglo-Norman）早在國王愛德華（Edward the Confessor, r. 1042-1066）時期就被引入英格蘭宮廷。⼤衛・喬治認為，盎格魯—諾曼語從⼜頭古法語⽅⾔轉變為書寫的⽂學語⾔，古法語（franceis）的語⾔結構發⽣了較⼤變化，「法語是榮譽的、騎⼠的甚至司法的語⾔，⽽更具男⼦氣概和表現⼒的盎格魯—薩克森語／⽅⾔則被拋棄」。74 在⼗⼆世紀和⼗三世紀對於中層社會法語和英語似乎是可以互換的，至少 1250年前後在英格蘭有⼀些法語教科書、法語詞典和論⽂是為了「向在學校學習過拉丁語並且知道孤⽴法語的⾼級學⽣教授正確的⼤陸 69 Tahkokallio, J. The Anglo-Norman Historical Canon: Publishing and Manuscript Culture, (Cambridge: Cambridge University Press, 2019), 4. 70 Clover, B. The mastery of the French Language in England from the XIth to the XIVth Century, (New York: Corning & Co., 1888), 69-70. 71 West, C. B. Courtoisie in Anglo-Norman Literature, (New York: Haskell House, 1966), 2. 72 Shelly, P. V. D. English and French in England, 1066-1100 (thesis), (Philadelphia: University of Pennsylvania, 1921), 74-92. 73 Mugglestone, L. The Oxford History of English, (Oxford: Oxford University Press, 2006), 61. 74 Georgi, D. Language made visible: The invention of French in England after the Norman Conquest (thesis), (New York: New York University, 2008), 3.

論莎士比亞戲劇中的法語 19 法語（巴黎的奧伊爾語，⽽不是盎格魯—諾曼語）」。75 盎格魯—薩克森語發⾳的變化在很⼤程度上也要歸功於法語。古英語有清⾳擦⾳[f]、[s]、[θ]（thin）和[ʃ]（shin），⽽法語的影響有助於區分它們的濁⾳對應物[v]、[z]、[ð]（the）和[ʒ]（mirage），也有助於形成雙母⾳[ɔy]（boy）。76 接著談談英格蘭安茹王朝（1154-1399）的法語。1154年安茹伯爵、諾曼第公爵亨利（Henry II, r. 1154-1189）成為英格蘭國王，這⼀新建⽴的安茹帝國（Angevin Empire）包括英格蘭和⼤陸法語地區的諾曼第、美茵、布列塔尼、安茹、普瓦圖、阿奎坦、圖盧茲和加斯科涅等，《約翰王》提到了⼤陸法語地區：普瓦圖（Poyctiers）、安茹（Aniowe）、都蘭（Torayne）、美茵（Maine）。法蘭克北部通⾏的官⽅語⾔是奧伊爾語，南部的圖盧茲、加斯科涅通⾏的是奧西坦⽅⾔（奧克語）。奧伊爾語也在英格蘭通⾏。⾃安茹公爵威廉九世（William IX, Duke of Aquitaine, 1071-1127）以來，阿奎坦宮廷積極⿎勵奧克語／普羅旺斯語的⾏吟詩⼈（troubadour）的騎⼠抒情詩創作（Gaucelm Faidit），這提⾼了奧伊爾語的⽂學地位。77 ⼗⼆世紀宮廷的保護促使⽂藝繁榮，阿奎坦的埃莉諾（Eleanor of Aquitaine, 1122-1204）原是法蘭克路易七世的王后（1137-1152），⽽後是英格蘭亨利⼆世的王后（1154-1189），她對宮廷⽂化、騎⼠⽂學影響深刻巨⼤，其母語是奧伊爾語。78 在埃莉諾的宮廷中，復興了拉丁古典⽂學（包括奧維德），詩⼈們創作了早期的騎⼠傳奇（romance）和寓⾔故事，八⾳詩⾏敘事詩（Lais）是這時期最流⾏的奧伊爾語詩體。托⾺斯・切斯特《朗法爾爵⼠》（Thomas Chestre, Sir Launfal, or Launfalus Miles）。英格蘭的托⾺斯（Thomas of England, 1100-1150）創作了《特裏斯坦》（Tristan and Isolde, 1170），79 法蘭西的瑪麗（Marie de France）創作了 12⾸八⾳詩⾏敘事詩（Lais）《聖帕特利茲的淨界》（L’Espurgatoire Seint Patriz）和《伊利杜克》（Eliduc），80 特 75 Lloyd, C. Semantics and Word Formation: The Semantic Development of Five French Suffixes in Middle English, (Oxford: Peter Lang, 2011), 14. 76 Marckwardt, A. H. Historical Outlines of English Sounds and Inflections, (Ann Arbor: George Wahr Publishing Co., 1951), 19. 77 Bourguignon, A. Grammaire de la Langue D'oïl (Français des XIIe et XIIIe Siècles), (Paris: Garnier Frères, 1873), 401. 78 Wheeler, B., Parsons, J. C. eds. Eleanor of Aquitaine: Lord and Lady, (New York: Palgrave Macmillan, 2003), 55-76, 337-368. 79 Legge, M. D. Anglo-Norman Literature and its Background, (Oxford: Clarendon, 1963), 259-261. 80 Harf-Lancner, L., Warnke, K. Les Lais de Marie de France, (Paris: Le livre de poche, 1990), 7-19.

彭建華 20 洛瓦的克雷蒂安（Chretien de Troyes, flourished 1165-1180）創作了 5部亞瑟王傳奇⼩說《埃勒克》（Erec）、《克裏熱》（Cligès）、《蘭斯洛》（Lancelot, ou Le Chevalier à la charrette）、《伊萬》（Yvain, ou Le Chevalier au lion）、《帕西法爾》（Perceval, ou Le Conte du Graal）和《紀堯姆・德・安格爾特爾》（Guillaume d’Angleterre），溫塔杜爾的伯納德（Bernard de Ventadour）創作了較多宮廷愛情詩。81 ⾙諾伊（Benoît de Sainte-Maure）出⽣於法蘭西的普瓦捷，曾經活躍在阿奎坦的埃莉諾宮廷。埃莉諾的女兒⾹檳伯爵夫⼈瑪麗（Marie de Champagne, 1145-1198）在巴黎繼續⿎勵了宮廷愛情詩的創作。82 莎⼠比亞《約翰王》寫到了充滿權⼒欲的埃莉諾，她是約翰王（John Plantagenet, r.1199-1216）的母後，但該劇沒有⽤法語作為其身份認同的標誌。 封建政治的變化往往容易引發語⾔的民族觀念、對各種⽅⾔（語⾔變體）的看法、語⾔政策和語⾔形式本身的改變。英法百年戰爭是從卡佩王朝的王位繼承和對安茹領地宗主權爭奪開始的，1337年英格蘭國王愛德華三世對瓦盧瓦王朝發起了戰爭。莎⼠比亞的歷史劇《愛德華三世》象徵性地表現了最初的戰爭和克雷西戰役。此外，1348年以來⿊死病的流⾏，城市⼈⼜的⼤量死亡可能加速了中古法語與盎格魯—奧伊爾語的分離。⼗⼆至⼗五世紀英格蘭留下了較多盎格魯–諾曼語⼿稿，尤其是來⾃英格蘭本篤會教堂的盎格魯—諾曼語⼿稿，包括宗教⽂學、宮廷愛情詩和亞瑟王傳奇等。83 四、英格蘭金雀花王朝的法語：外語或者方言 ⼗四世紀中期⿊死病的流⾏和英法百年戰爭的發⽣，英格蘭民族意識的興起，英格蘭有意強調了新的民族語⾔，以倫敦英語⽅⾔（東中部地區⽅⾔）為基礎的中古英語迅速興起，1350年之後新的拉丁語—法語化的盎格魯—薩克森語（中古英語）重新成為法院的官⽅語⾔，⾦雀花王朝在語⾔政策上有意加強了中古英語作為通⾏ 81 Kibler, W. W. ed. Eleanor of Aquitaine: Patron and Politician, (Austin: University of Texas Press, 2014), 35. 82 Boyd, D. Plantagenet Princesses: The Daughters of Eleanor of Aquitaine and Henry II, (Barnsley: Pen & Sword Books, 2020), 93. 83 Green, V. The Medieval Author in Medieval French Literature, (New York: Palgrave Macmillan, 2006), 35.

論莎士比亞戲劇中的法語 21 語⾔的地位，法語的重要地位出現顯著的下降。⽤盎格魯—撒克森語柴郡⽅⾔抄寫的《⾼⽂爵⼠與綠⾊騎⼠》（Sir Gawain and the Green Knight）、《珍珠詩篇》（Pearl Manuscript）等深受中古法語騎⼠敘事⽂學的影響。84 ⾸先談談英格蘭⾦雀花王朝的法語。1250-1400年間，法語辭彙句法對盎格魯—薩克森語的影響最為強烈，1380年代倫敦開始使⽤新的盎格魯—薩克森語（中古英語）。85 ⾦雀花王朝（1399-1485）前期依然是普遍的多語⾔⽂化，沒有改變法語作為官⽅語⾔的地位。⽽且亨利五世、亨利六世⼀直擁有說法語的⼤陸屬地，來⾃法蘭西瓦盧瓦王室的王后也使得宮廷中依然流⾏法語。⼗四世紀末英格蘭詩⼈（例如，Pierre de Peccham, Robert Grosseteste, William of Wadington等）往往是⽤英語、法語和拉丁語三種語⾔創作的。其中，喬叟、⾼渥對莎⼠比亞的詩歌和戲劇有明顯的影響。雖然⼗四世紀的英語倫敦⽅⾔是傑弗裏・喬叟（Geoffrey Chaucer, 1343-1400）的母語，但他是⼀個多語種（法語、拉丁語、希臘語和義⼤利語）作家，其作品是以抄寫本流傳的，1476年威廉・卡克斯頓在倫敦⾸次印刷出版。由於法蘭西奧伊爾語和拉丁語的影響，喬叟詩作的辭彙也表現出多樣性。就喬叟本⼈⽽⾔，義⼤利語（即義⼤利—羅曼語）也可以被添加到這些更普遍的外國影響來源之中。1367-1368年喬叟開始⽤法蘭西奧伊爾語寫詩，法語詩⼈尤斯塔斯·德尚（Eustace Deschamps, 1346-1406）對喬叟的早期詩歌作品產⽣了重⼤影響，他⽤英語翻譯了荷⾺《伊利亞特》（古希臘語）、洛裏斯的紀堯姆、默恩的讓《玫瑰傳奇》（中古法語）、波伊修斯《哲學的慰藉》（拉丁語），並為中古英語引入了許多外來語借詞，豐富了英語韻律。86 喬叟在《維納斯的抱怨》（Compleynt of Venus）的末尾宣稱，格蘭森（Granson）是最傑出的法國詩⼈，讀眾應該熟悉這位法國作家，他盡可能地從格蘭森的法語版本精⼼翻譯了這⼀詩作，卻只是翻譯成平淡無奇的英語。《坎特伯雷故事集》（The Canterbury Tales）主要是⽤中古英語創作的，赦罪神⽗在傳教時使⽤拉丁語，⾃由農、法庭召喚⼈都可以說拉丁語，⽽女修道院院長講法語，這表現出鮮明的多語⾔⽂學特徵，也表現出明顯的混合⽅⾔的多樣性。其中《鄉鎮法警／司法官的故事》（Reeve’s Tale）使⽤的北部⽅⾔，是為了追求特定的⽂學表達效果︔⽽出現在他的 84 Strohm, P. Middle English, (Oxford: Oxford University Press, 2007), 39. 85 Price, H. Th. Foreign Influences on Middle English, (Ann Arbor: University of Michigan Press, 1947), 34-38. 86 Burnley, D. The Language of Chaucer, (Basingstoke, London: Palgrave Macmillan UK, 1983), 103.

彭建華 22 作品中更多的是因為它更早之前就存在於現代倫敦英語⽅⾔（東中部⽅⾔）中，儘管可能有助於他突出其詩歌韻律特徵。87 約翰・⾼渥（John Gower, 1325/1330-1408）是喬叟最親密的朋友，他分別⽤法語、英語和拉丁語創作，其英語混合了肯特和斯塔福德（Stafford）⽅⾔，他深刻的受到法語⽂學傳統的影響，中古法語的抒情詩、謠曲和愛情⼩調（dits amoreux）為⾼渥的詩作帶來了靈活的⽅式。88 ⾼渥⽤法語創作的謠曲《冥想的奇跡》（The Speculum meditantis, or Mirour de l’omme）和敘事詩《泰爾的阿波羅尼烏斯》（Apollonius of Tyre），前者在韻律上是成熟⽽獨特的，混合了盎格魯—奧伊爾語和盎格魯—撒克森語⽅⾔。⾼渥主要是⼀位敘事詩⼈和道德家，《情⼈的懺悔》（Confession Amantis）以神的傳譯者（auctors）和詩⼈之間的對話形式寫成，部分是從法蘭克奧伊爾語詩⼈聖莫雷的⾙諾伊《特洛伊傳奇》（Benoît de Sainte-Maure, Roman de Troie, 1160）改寫的。89 1437 年之後，由於與⼤陸法語地區幾乎完全隔離（除加萊外），英格蘭逐漸失去了與⼤陸法語⽂化的直接聯繫，法語在英格蘭王國逐漸⾛向消失。法語作為外語或者⽅⾔，是⼀個難以判斷的問題。90 因為玫瑰戰爭結束的時候，以法語為主要語⾔之英格蘭貴族急劇減少，以盎格魯—羅曼語為主要語⾔之爵⼠、騎⼠階層⼤增，英語倫敦⽅⾔逐漸成為被尊崇的、普遍使⽤的⽅⾔，並將逐漸成為民族語⾔。由於講⽅⾔的⼈群主要是從中部地區、南部地區來的移民，因⽽倫敦⽅⾔是⼀種具有多樣性的混合語⾔。91 都鐸王朝（Tudor Dynasty, 1485-1603）結束玫瑰戰爭和中世紀的混亂，開始了英格蘭⽂藝復興，⽅興未艾的活字印刷業促進了盎格魯—羅曼語（中古英語）向早期現代英語的轉變，並越來越遠離法語。1512-1525年，亨利八世（Henry VIII, r. 1509-1547）與西班牙王國⼀起反對法國。1532 年，亨利八世的國家（宗教）改⾰卻加劇英格蘭與羅⾺教廷和整個⼤陸分離，早期現代英語⾛上單⼀語⾔之民族化道 87 Vising, J. Anglo-Norman Language and Literature, (Oxford: Oxford University Press, 1923), 36. 88 Yeager, R. F. John Gower: The French Balades, (Kalamazoo: Western Michigan University, 2011), 1-4. 89 Smith, J. J. Studies in the Language of some Manuscripts of Gower’s Confessio Amantis (Dissertation), (University of Glasgow, 1985), 27. 90 Moore, S. Historical Outlines of English Phonology and Middle English Grammar, (Michigan: G. Wahr, 1919), 80-81. 91 Horobin, S., Smith, J. An Introduction to Middle English, (Oxford: Oxford University Press, 2002), 29.

論莎士比亞戲劇中的法語 23 路。92 在英格蘭多語⾔社會中，羅曼—盎格魯語（中古英語）往往以特定的⽴場/策略介入到宗教改⾰運動中，1380年約翰·威克利夫（John Wycliffe）從拉丁語俗語翻譯的英語聖經⼿稿是已知最早的英語聖經，1535 年之前英語聖經⼀直受到羅⾺教廷、英格蘭國王和主教的壓制，⽽且對英語本身的發展影響甚微。93 英語翻譯的⽅式是宗教改⾰的⼀個根本問題，亨利八世廢除英格蘭各地的羅⾺天主教修道院，極⼤削弱了拉丁語作為宗教語⾔的地位。1496年約翰・柯爾特（John Colet）從希臘語翻譯了英語《新約》，1525年威廉・廷代爾（William Tyndale）從伊拉斯慕斯的希臘語版本翻譯了英語《新約》，1535 年邁爾斯・科弗代爾（Myles Coverdale）翻譯了英語《舊約》。雖然這些英語《聖經》⼤都遭受了國王和主教的嚴厲打壓，卻也培養了廣⼤的英語讀者。1537年約翰・羅傑斯（John Rogers）刊印了從希伯來語和希臘語翻譯的完整的英⽂聖經。1539年托⾺斯・克蘭默（Thomas Cranmer）主持編寫的英語《⼤聖經》（the Great Bible）被廣泛發⾏和被授權公開使⽤，為了更好地被「沒有完全學會拉丁語」的讀者理解和接受，該譯本採⽤了簡潔明瞭的英語，較好的平衡了與拉丁語彙在字⾯上的適合／對應。1549年聖公會《通⽤祈禱書》（Book of Common Prayer）使⽤了簡明規範的英語，⽇內瓦—湯姆遜《聖經》和這些英語祈禱書有助於早期現代英語的標準化。94 1571 年，《三⼗九條宗教綱要》（Thirty-nine Articles of Religion）最後確認了英格蘭宗教改⾰的結果，尤其對英語規範上的成果，「在宗教集會中⽤⼈們所理解的語⾔說話。公眾禮拜要⽤當地⽅⾔進⾏。」1572年主教《聖經》（Bishops’ Bible）對莎⼠比亞風格和引⽤典故的重要性不容低估，其劇作顯⽰出他非常熟悉星期天在教堂中所閱讀的教會經典《智慧篇》（Ecclesiasticus）的段落，這是宗教改⾰以後出現的次經作品。95 其次談談英格蘭都鐸王朝的早期現代英語。⼗六世紀中期英格蘭王國完全失去了歐洲⼤陸的法語地區的宗主權，成為⼀個海島國家。由地理、歷史事件、內外勢 92 Nevalainen, T. An Introduction to Early Modern English, (Oxford: Oxford University Press, 2008), 1-5. 93 Work, E. Wh. The Bible in English literature, (New York: Fleming H. Revell company, 1917), 122-124. 94 Norton, D. A History of the English Bible as Literature, (Cambridge: Cambridge University Press, 2004), 36. 95 Noble, R. Shakespeare's Biblical Knowledge and Use of the Book of Common Prayer, as Exemplified in the Plays of the First Folio, (London: Society for promoting Christian knowledge, 1935), 99-105.

彭建華 24 ⼒和宗教團體塑造的英格蘭⽂化即是「英國民族性」（Englandness）。在法蘭西國王查理八世（Charles VIII, 1483-1498）的幫助下，亨利七世（Henry VII, r. 1485-1509）打敗了約克家族的理查德三世（Richard III, r. 1483-1485），從⽽建⽴了都鐸王朝。亨利七世開啟與威爾⼠（亨利出⽣於威爾⼠彭布洛克的彭布洛克城堡）的和解與融合，他的改⾰始形成「英國民族性」的意識。96 由於對法蘭西布列塔尼公爵領地的爭奪，1488, 1489-1492年亨利七世 2次發起了對法國的戰爭，最終放棄了布列塔尼的宗主權。1510-1513, 1521-1526年亨利八世 2次參加了歐洲同盟對法國的戰爭，即使 1546-1550 年英格蘭短暫奪取了布洛涅城，1557-1560 年英格蘭、西班牙對法國的戰爭之後英格蘭失去加萊，英格蘭王國的領⼟退回到不列顛。與法國的長期敵對狀態，法語及其⽂學加速退出了英格蘭王國。97 同時，法蘭西王國完成了對⾼盧—法蘭克地區的統⼀，形成「法國民族性」（Franceness）意識，法語作為民族語⾔，持續加強了拉丁化傾向。 都鐸王朝時期，⼈⽂主義的⽂藝復興與活字印刷術深刻改變了普通教育與英語本身。作為現代的民族國家，英語是英格蘭顯性的語⾔標誌，⽽法語是法蘭西顯性的語⾔標誌。莎⼠比亞在《亨利五世》中宣稱英語是亨利五世的語⾔，⽽勃根第公爵宣稱「世界上最好的花園，我們肥沃的法國」。雖然常常因為戰爭⽽中斷，英格蘭與法蘭西的貿易和⽂化交流還是使英國受益，⼗五世紀後期法國印刷的書籍⼤量進入英格蘭國內。法國也是義⼤利⽂藝復興時期古典語⽂學、藝術從歐洲⼤陸傳到英格蘭的「促進者」。98 1066-1400年，盎格魯—薩克森語是社會下層⼈們通⾏的交流語⾔（⼜語），⾃喬叟以來，倫敦英語已經成為優雅的⽂學語⾔。 英法⽂藝復興早期，英語法語並不是處於對⽴的地位，有法語學者聲稱，早期現代英語的詞彙 80%來源於法語。原初英語、法語，法蘭克語（原初法語）、盎格魯—薩克森語原本都是中古⽇爾曼語，法蘭克語與⾼盧羅曼語接觸混合，並進⼀步羅曼語化，⽽盎格魯—薩克森語與拉丁化的不列顛諸⽅⾔接觸混合，並進⼀步羅曼語 96 Haigh, Ch. English Reformations: Religion, Politics, and Society under the Tudors, (Oxford: Oxford University Press, 1993), 87. 97 Smith, J. J. Essentials of Early English: An Introduction to Old, Middle, and Early Modern English, (London, New York: Routledge, 2022). 98 Gunn, S. Henry VII’s New Men and the Making of Tudor England, (Oxford: Oxford University Press, 2016).

論莎士比亞戲劇中的法語 25 化。1066-1500年，諾曼第征服為英格蘭帶來了諾曼第法語，另⼀種羅曼語化的⽇爾曼—法語混合⽅⾔，區別於巴黎多依語，此為北部之法語⽅⾔。 五、戲劇中職業小丑、喜劇性人物及其法語 英格蘭⽂藝復興早期，英格蘭（尤其是倫敦）依然處於通⾏英語、法語、拉丁語的多語⾔社會境況，莎⼠比亞戲劇中的語⾔是莎⼠比亞、國王劇團演員、倫敦的戲劇觀眾共同分享的、⽇常的通俗語⾔，法語、拉丁語本身並不是莎⼠比亞刻意使⽤的個性化語⾔。莎⼠比亞作為劇作家、演員，熟悉倫敦市民⼜中流傳的簡易的法語，尤其是泰晤⼠河南岸商業區流傳的簡易的、隨意的法語⽅⾔。1542年安德魯・波爾德《知識導論入⾨》（Andrew Borde, The fyrst boke of the Introduction of knowledge）指出，作為英格蘭國王的臣民，語⾔（英語、法語、威爾⼠⽅⾔、康沃爾⽅⾔、愛爾蘭⽅⾔、蘇格蘭⽅⾔等）及其特徵是身份認同的重要⽅式。99 作為⼀個商業劇團的演員和劇作家，莎⼠比亞是⼀個嚴格的英格蘭民族主義者，基於都鐸王朝的政治意識形態，他忽視了中古時期通⾏多語⾔的英格蘭⽂化。⼈們不知道莎⼠比亞在公⽴學校或者在倫敦是否接受過正式的法語教育，他可能沒有直接學習過來⾃巴黎的早期現代法語。倫敦有較多來⾃法國的商⼈，莎⼠比亞更容易接受倫敦商⼈⼜中的法語。亞裏⼠多德《詩學》論及喜劇時說，喜劇是對更低微/低下的⼈們的模仿，低微/低下是指道德的，或者社會地位的，⽽喜劇性（誇張可笑）⼈物表現得很糟糕，荒誕可笑的是⼀種不涉及痛苦或毀滅的錯誤或羞恥，荒誕可笑的是⼀種不光彩（不符合道德）的東西。100 只有在英語成為國家的官⽅語⾔之後，倫敦英語才具有比別的英格蘭⽅⾔更⾼的地位。正是因為早期現代英語作為英格蘭王國純正標準的單⼀語⾔，法語和各種英格蘭王國的⽅⾔才成為莎⼠比亞戲劇中被嘲笑/取笑的對象。101 ⾸先談談莎⼠比亞戲劇中 2個友善的法國⼈物。除開頻繁的⽂化交流，英格蘭都鐸王朝與歐洲⼤陸的西班牙帝國、神聖羅⾺帝國保持了友好關係，整個伊莉莎⽩ 99 Loewenstein, D., Mueller, J. ed. The Cambridge History of Early Modern English Literature, (Cambridge: Cambridge University Press, 2002), 203. 100 Lucas, D. W. Aristotle’s Poetics (Clarendon Greek Text and English Commentary), (Oxford: Oxford University Press, 1968), lxii. 101 Alexander, C. M. S. Shakespeare and Language, (Cambridge: Cambridge University Press, 2004), 18.

彭建華 26 時代（Elizabeth I, r. 1558-1603），倫敦從來都不缺說法語的法國⼈，宮廷的時尚深受西班牙和法國風格的影響，但法語幾乎完全從英格蘭的官⽅語⾔中退出。1603年莎⼠比亞曾租住在⽼倫敦城法國珠寶商蒙喬伊（Christopher Mountjoy）的房屋，這個胡格諾派的僑民可能擁有法語書籍。102 莎⼠比亞戲劇，尤其是歷史劇，較多寫到了法國⼈及其所說的法語。1066-1500年，由於法語是英格蘭宮廷、⾏政、法律和教育的官⽅語⾔，莎⼠比亞戲劇中說法語的⼈物往往是英格蘭社會中上階層⼈⼠及其侍從。莎⼠比亞將英格蘭王國與法蘭克王國長期歷史的各個⽅⾯戲劇化，除開《托⾺斯・莫爾爵⼠》（Anthony Munday, Henry Chettle, Thomas Dekker, Thomas Heywood合作），主要包括《約翰王》（King John）、《愛德華三世》（Edward III）、《理查德⼆世》（Richard II）、《亨利四世》第⼀、⼆部（Henry IV, Part I, II）、《亨利五世》（Henry V）、《亨利六世》第⼀、⼆、三部（Henry VI, Part I, II, III）、《理查德三世》（Richard III）和《亨利八世》（Henry VIII）等，《李爾王》、《如願》、《仲夏夜之夢》、《溫莎的風流娘們》、《愛的徒勞》等。莎⼠比亞戲劇的法國來源極少，《暴風⾬》中對理想共和國的諷刺描述可以追溯到蒙⽥的散⽂。《溫莎的風流娘們兒》第⼀場第 4幕快嘴女管家（Mistris Quickly）談論到「國王的英語」（Kings English），這是 1553年出現的⼀個新詞，顯然正確純潔英語是早期現代英語的語⾔現象。 在《亨利五世》中，法語程度副詞Ma foy（好的，確實）, 1600年第 1四開本（Q1）6次，1623年第 1對折本（F1）1次。103 莎⼠比亞描述了 2個友善的法國⼈物。《李爾王》中的法蘭西國王和《亨利五世》中的凱瑟琳公主，是莎⼠比亞戲劇中為數不多的獲得讚譽的法國⼈物。《李爾王》（His True Chronicle Historie of the life and death of King LEAR and his three Daughters, 1608）是根據女王劇團在玫瑰劇院演出的《萊爾國王的真實編年史》（The True Chronicle History of King Leir）⼀劇改編的，《李爾王》⼀劇中沒有出現法語詞語本身。劇中的法蘭西國王（King of France, Lords of France, Princes France）僅僅出現在第⼀場第 1幕中，他娶了李爾王最⼩的女兒科迪莉婭 Cordeilla，其形象更接近九世紀之後的歐洲騎⼠，「於他⼈的冷漠之中，/我的愛竟然燒成熾熱的尊敬。/你的無妝奩的女兒既屬我有，/她便是我的法國臣民的王 102 Edmondson, P., Wells, S. The Shakespeare Circle, (Cambridge: Cambridge University Press, 2015), 175. 103 Shakespeare, W. A Most pleasaunt and excellent conceited Comedie, of Syr Iohn Falstaffe, and the Merrie Wiues of Windsor, (London: Printed by Thomas Creede for Arthur Iohnson, 1602).

論莎士比亞戲劇中的法語 27 后」。that from their couldst neglect,/ My loue should kindle to inflam'd respect,/ Thy dowreles daughter King throwne to thy chance,/ Is Queene of vs, of ours, and our faire France:雖然第五場第 1幕出現了科迪莉婭率領法國軍隊登陸不列顛，法蘭西國王卻未再次出現。李爾王是羅⾺⼈入侵之前（43 BCE）的不列顛國王，法蘭克王國出現在 481年之後，西薩克森統治英格蘭王國時期才出現阿爾班尼、肯特和康瓦爾等封建領主，法蘭克與勃根第的競爭關係暗⽰了玫瑰戰爭時期的國外背景，因此該劇包含⼀些歷史知識的錯誤。霍林希德《英格蘭、蘇格蘭和愛爾蘭編年史》寫到：「世界紀元 3105年，巴爾杜德（Baldud）的兒⼦莱爾/李爾（Leir）被承認為不列顛⼈的統治者，當時喬阿斯（Joas）統治著猶太地區。萊爾是⼀位舉⽌⾼貴的君王，統治著他的⼟地和臣民，擁有巨⼤的財富。他建⽴了萊爾鎮（Caerleir），現在稱為萊切斯特（Leicester），坐落在索爾河上（River of Sore）。」104 在《李爾王》中，除開⼀些拉丁語和法蘭西的地名和⼈名等專有名詞，這些劇中的⼈物對⽩沒有法語詞語。 由於諾曼—安茹王朝的君主與貴族幾乎全來⾃法蘭克王國，因⽽莎⼠比亞並沒有低估法蘭克王國本身的價值，對來⾃法蘭西王國的凱瑟琳公主（Catherine of Valois, 1401–1437）表達了崇敬和讚美，這位亨利五世的遺孀也是亨利七世的祖母。在《亨利五世》中，凱瑟琳公主問亨利：我有可能愛上法蘭西的敵⼈嗎？“Kate. Is it possible dat me sall/ Loue de enemie de France.” 亨利回答說：「不，凱特，你不可能愛法蘭西的敵⼈，但愛我就應該愛法蘭西的朋友︔因為我如此熱愛法蘭西，以至於我不會放棄它的⼀個村莊。」（Harry. No Kate, tis vnpossible/ You should loue the enemie of France:/ For Kate, I loue France so well,/ That Ile not leaue a Village,/ Ile haue it all mine: then Kate,/ When France is mine,/ And I am yours,/ Then France is yours,/ And you are mine.） 其次談談眾多滑稽可笑的法國⼈物。亨利七世、亨利八世持續地與法蘭西王國的敵對狀態與戰爭，較⼤地影響了莎⼠比亞對法國⼈不友善的⽴場。莎⼠比亞戲劇中的法國⼈物常常浮誇傲慢，當然也很敏感。伏爾泰認為，法國沒少受夠侮辱，被愚弄和嘲笑。（France has not insults, fool’s-caps, and pillories enough for such a 104 Holinshed, R. Chronicles of England, Scotland, and Ireland, ed. by W.G. Boswell–Stone, (London: Chatto and Windus, 1907), 1–2.

彭建華 28 scoundrel.）105 《溫莎的風流娘們》（A pleasant conceited Comedie, of Syr Iohn Falstaffe, and the merry Wiues of VVindsor）是⼀個散⽂體的複雜鬧劇，劇中僅有少量的詩體。這也是⼀個多語種的劇作，1602年第 1四開本（Q1）中沒有法語，1623年第⼀對折本（F1）增添了⼀些法語對⽩。第三場第 1 幕長襪褲酒店（Garter Inn）主⼈提到 4 種⽅⾔Gallia and Gaule, French & Welch,（⾼盧⾏省語和⾼盧語，法蘭克語和威爾⼠語），Gallia可能指威爾⼠曾經通⾏的羅⾺帝國不列顛⾏省的羅曼—⾼盧語，“Gaule”指⾼盧的凱爾特語。 《溫莎的風流娘們》⼀劇中表現出多語⾔⽂化的社會階層分離，法語被看作為更⾼地位⼈們的交流語⾔，莎⼠比亞常常使⽤語⾳近似的雙關語。第五場第 5幕快嘴女管家說出「長襪褲騎⼠團」的法語箴⾔“Hony Soit Qui Mal-y-Pence”,（⼼有邪念者才是可恥的）。法官夏婁（Iustice Shallow, Robert Shallow Esquire）可能也是來⾃法蘭克的貴族，因為他宣稱他的家族 300年來是有紋章的紳⼠稱號（and haue done any time these three hundred yeeres.）。福爾斯塔夫爵⼠（Falstoffe, Sir Iohn Falstaffe, the Knight Sir Iohn）無疑是來⾃法蘭西的貴族，他談到了法蘭西宮廷（Let the Court of France shew me such another. III.3）。第⼀場第 3幕他⾛向愛情誤會的陷阱時宣稱「如果⽤確切的英語來說，」（to be english'd rightly）⽪斯托對此評論道：「把她的⼼願很好地譯成英語」，（translated her will: out of honesty, into English）。106 ⽽後福爾斯塔夫說他將學習這個時代的風尚「法國式的節儉」（Falstaffe will learne the honor of the age, French-thrift,）。第⼆場第 1幕暗⽰隊長尼姆（Corporall Nim）似乎不能機智地說出英語（heere's a fellow frights English out of his wits.） 法國醫⽣卡優斯（Caius the French Doctor/French Physician）是另⼀個喜劇性⼈物，他⼗分接近社會上層（de Earle, de Knight, de Lords, de Gentlemen）他的朋友在宮廷中很有勢⼒（his friends Potent at Court），並暗⽰他⽇常是⽤法語交談的。他的英語發⾳錯誤，他對詞語含義的誤解，以及他⽤法國⼜⾳念錯英語詞語的⽅式，法語和英語混雜的、意味略顯粗魯的語句，這些都產⽣了搞笑/鬧劇性的效果。第⼀場 105 Paulin, R., Poole, A., Holland, P. Voltaire, Goethe, Schlegel, Coleridge: Great Shakespeareans: Volume III, (London: Bloomsbury Publishing Plc, 2010), 17. 106 莎士比亞，《莎士比亞全集 溫莎的風流婦人》，梁實秋譯，（北京：中國廣播電視出版社，2002），37。

論莎士比亞戲劇中的法語 29 第 4 幕寫到住在溫莎的卡優斯，他的對⽩使⽤了法語“vn boyteene verd; Court la grand affaires; Ouy mette le au mon pocket, de-peech quickly”。第⼆場第 3幕嘲諷地寫到襪帶酒店主⼈、卡優斯醫⽣ 2⼈⽤混雜的法語和英語的⽣硬表達。店主：「對不起，貴客法官。嗯，便溺先⽣。」（Host. Pardon, Guest-Iustice; a Mounseur Mocke-water.） 卡優斯：「便溺，什麼意思？」（Caius. Mock-vater? vat is dat?）店主：「便溺，在我們英語中的意思是『勇敢（霸氣）」。（Host. Mock-water, in our English tongue, is Valour(Bully.)）卡優斯：「哦，我和英國⼈⼀樣『勇敢』：下流的傢伙–狗牧師：哦，我要割下他的⽿朵。」（Caius. By gar, then I haue as much Mock-vater as de Englishman: scuruy-Iack-dog-Priest: by gar, mee vill cut his eares.）卡優斯：「哦，你這麼說，我得謝謝你：哦，我愛你。」（Caius. By-gar, mee dancke you vor dat: by gar I loue you:） 第三場第 1至 2幕描述卡優斯醫⽣被誘導進入與修爵⼠在弗洛格莫爾原野（the Field near Frogmore）決⾾的玩笑⾏為，店主取笑⼆⼈傷害了英語（hack our English.）。同樣，在此突出了卡優斯醫⽣的混合語⾔，卡優斯：「你為什麼不與我決⾾？」（vherefore vill you not meet-a me?）卡優斯：「哈，我沒有預想到這個（作弄），你是作弄了我們？」（Caius. Ha' do I perceiue dat? Haue you make-a-de-sot of vs,） 卡優斯：「我呃，⼩姐愛的是我：我的快嘴女管家是這樣給我說的。」（Caius. I be-gar, and de Maid is loue-a-me: my nursh-a-Quickly tell me so mush.）第四場第 5幕再次寫到了卡優斯醫⽣，卡優斯：「襪帶酒店主⼈在哪裡？」（Caius. Ver' is mine Host de Iarteere?）卡優斯：「我不懂你說的什麼：但我聽說你在準備盛⼤接待⼀位德意志公爵，我給你說⼀下真實，沒有公爵到宮廷裏來，我好意告訴你：再⾒。」（Caius. I cannot tell vat is dat: but it is tell-a-me, dat you make grand preparation for a Duke de Iamanie: by my trot: der is no Duke that the Court is know, to come: I tell you for good will: adieu.） 第五場第 3, 5幕寫到卡優斯醫⽣與假扮的安·佩吉去教堂結婚，107 卡優斯：佩吉太太在哪裡？噢我被騙了，我要與少年結婚，⼀個男孩，⼀個鄉下⼈，哦。Caius. Ver is Mistris Page: by gar I am cozoned, I ha married oon Garsoon, a boy; oon pesant, by gar. 卡優斯：我唉，是⼀個男孩，噢，我將讓所有溫莎⼈震驚。Caius. I bee gar, and 107 Martina, E. “The Use of Dialects and Foreign Languages in Shakespeare’s King Henry V.” English Studies 100.7 (2019): 767-784.

彭建華 30 'tis a boy: be gar, Ile raise all Windsor.108 在《亨利四世》第⼀、⼆部中，除開⼀些拉丁語和法蘭西的地名和⼈名等專有名詞，這些劇中的⼈物對⽩沒有法語詞語。《亨利六世 第⼀部》可能是莎⼠比亞及其合作者創作的⼀部歷史劇，劇中頻繁論及英格蘭⾦雀花王朝在法蘭西的領地，例如，奧爾良、阿登等，Neuer so needfull on the earth of France, 109 和法蘭西的專有名詞和⼈名，例如，聖丹尼斯是法國的守護神。「服從多爾芬親王？這只是⼀個法語詞語：我們英國戰⼠不在乎它的意義。」Submission Dolphin? Tis a meere French word: We English Warriours wot not what it meanes.⼗⼆世紀以來，維耶諾瓦 Viennois的領主被稱為多爾芬尼（Dauphiné, or delphinate）。1349年法蘭西王室領地維耶諾瓦被分封給查理（Charles V, Charles Le Sage, 1338-1380）1364，查理被稱為多爾芬尼親王Dauphiné。 《亨利六世 第⼆部》也是⼀個多語⾔的劇作，除開拉丁語，其中包含很多法蘭西的專有名詞和⼈名。根據 1420年《特洛瓦協約》，亨利六世的完整頭銜是「英格蘭國王，和法蘭西國王」（King of England, King of France），其王后是來⾃安茹—梅恩公爵、那不列斯、西西⾥、耶路撒冷國王瑞尼耶的女兒瑪格麗特（Margaret of Anjou）。反叛者傑克·凱德則宣稱「法國⼈是我們的敵⼈」（The Frenchmen are our enemies:），莎⼠比亞喜劇性的描述了瑪格麗特王后，親法的薩福克公爵威廉·德·拉—珀爾（William de la Pole first Duke of Suffolke）被⼩丑化，他被指責向法國出賣了安茹、梅恩，沒有處理好對諾曼第的宗主權（By thee Aniou and Maine were sold to France./ The false reuolting Normans thorough thee,）。劇中僅僅⽀持瑪格麗特王后的⾺奇伯爵莫蒂默、克利弗德勳爵的對⽩中 2 處包含法語，「莫蒂默：再⾒」Mortimer. Dieu.「克利弗德：結果賦予所作之事的最終價值」Clifford. La fin Corrone les eumenes.這是拉丁成語 finis coronat opus 的法語翻譯。在《亨利六世 第三部》（The true Tragedie of Richard Duke of Yorke, and the death of good King Henrie the Sixt）⼀劇中，除開拉丁語和⼀些法蘭西的專有名詞和⼈名，⼈物對⽩中沒有出現法語詞語。《約翰王》（The life and death of King Iohn）也是如此，僅僅第三場第 1幕末尾布蘭琪的對 108 莎士比亞，《莎士比亞全集 溫莎的風流婦人》，201。 109 Womersley, D. “France in Shakespeare's ‘Henry V’.” Renaissance Studies 9.4, France in the English and French Theatre of the Renaissance (1995.12): 442-459.

論莎士比亞戲劇中的法語 31 ⽩包含法語詞語，「⼣陽被⾎⾊的雲遮住了：再⾒了美好的⽩天。」Blanch. The Sun's orecast with bloud: faire day adieu.110 六、歷史劇《亨利五世》中的法語分析 亨利五世（Henry V, r. 1413-1422）時期的宮廷語⾔依然是法語，法語和拉丁語依然是英格蘭通⾏的書寫語⾔，亨利五世通過特洛瓦協議獲得了法蘭西王位的繼承權，但他未能成為法蘭西國王。111 ⾦雀花王朝時期英格蘭還是多語⾔⽂化，英格蘭的中上階層是會說法語的，但法國⼈不會說英語，這是歷史事實。盎格魯—羅曼語（中古英語）的辭彙包含了⼤約⼀萬個法語詞語（直接或間接的），其中約四分之三仍在現代英語中使⽤，較⼩部分是法語成語/諺語的簡化形式，Allot（註定）, Artire（關鍵）, Armour（盔甲）, Bawcock（好⼩夥）, Ballow（拿、給）, Boitine（⼩盒⼦）, Cardecue（銀幣）, Consigne（交付）, Endeuors（努⼒）, Salute（歡迎）, Roy（王）等。112 《亨利五世》⽣動地表現了發⽣在法蘭西王國的歷史事件，法語作為外語，劇中第三場第 4, 5, 7幕、第四場第 1, 2, 4, 5幕、第五場第 2幕較多的使⽤了法語對⽩，作為喜劇性的語⾔要素。 (1). 《亨利五世》第三場第 4幕描述了⼀個喜劇性場景，即 1420年法蘭西王國的凱瑟琳公主（Catherine of Valois,1401-1437）試圖學習有關身體和衣物的九個盎格魯—羅曼語（中古英語）詞語。1600年第 1四開本（Q1）中的法語對⽩共計 30個印刷⾏，其拼讀是基於⼜語發⾳本身，其中有⼀些法語表達錯誤，parte [parler], se pella [s’appeler], tude [cudie], Aloues [aller]，是明顯的排印錯誤︔francoy是⼗六世紀通⾏法語 francoyse的誤印，1623年第 1對折本（F1）改為 Francois。Rehearser是1300年前後的盎格魯—諾曼語（古法語）新詞，它源⾃奧伊爾語詞 Rehercier (12c.)。1623年第 1對折本（F1）共計 56個印刷⾏，與前者（Q1）有⼀些明顯的差異，修正了前者的⼀些錯誤，也產⽣了新的錯誤，例如，Apprendre，Apoandre（1600Q1）和Apprins 110 莎士比亞，《約翰王》，劉昊譯，（北京：外語教學與研究出版社，2015）， 57。 111 Jacob, E. F. Henry V and the Invasion of France, (London: Praeger, 2008), 8. 112 Turnbull, W. B. The visions of Tundale; Together with Metrical Moralizations and other Fragments of early Poetry, Hitherto Inedited, (Edinburgh: T.G. Stevenson, 1843), 109.

彭建華 32 des（1623F1）︔Oublier，Oblye/obloye（1600Q1）和Oublie（1623F1）。113 Quato1, 1600 (Modern-spelling) Kate. Allice venecia, vous aues cates en, Kate. Alice, venez ici, vous avez quarante ans, Vou parte fort bon Angloys englatara, vous parlez fort bon l'anglais d'Angleterre. Coman sae palla vou la main en francoy. Comment s’appelez vous la main en anglais? Allice. La main madam de han. Allice. La main, madame? De han. Kate. E da bras. Kate. Et le bras? Allice. De arma madam. Allice. De arma, madame. Kate. Le main da han la bras de arma. Kate. Le main, da han; le bras, de arma. Allice. Owye madam. Allice. Oui, madam. Kate. E Coman sa pella vow la menton a la coll. Kate. Et comment s’appelez vous le menton et le col? Allice. De neck, e de cin, madam. Allice. De neck, et de cin, madame. Kate. E de neck, e de cin, e de code. Kate. Et de neck, et de cin. Et le coude? Allice. De cudie ma foy Ie oblye, mais Ie remembre, Allice. Le coude? Ma foi, j'oublie, Mais je remember Le tude, o de elbo madam. le coude, oh! De elbo, madame. Kate. Ecowte Ie rehersera, towt cella que Iac apoandre, Kate. Ecoutez: je rehearse tout celle que j'ai appris: De han, de arma, de neck, du cin, e de bilbo. de han, de arma, de neck, du cin, et de bilbo. Allice. De elbo madam. Allice. De elbo, madame. Kate. O Iesu, Iea obloye ma foy, ecoute Ie recontera Kate. O Jesu, j'ai oublié, Ma foi! Ecoutez; je raconterai: De han, de arma, de neck, de cin, e de elbo, e ca bon. de han, de arma; de neck, de cin; et de elbo. Est-ça bon? Allice. Ma foy madam, vow parla au se bon Angloys Allice. Ma foi, madame, vous parlez aussi bon anglais Asie vous aues ettue en Englatara. As[comme] si vous aviez étudié en Angleterre. 113 Loiseau, A. Histoire de la Langue Française: Ses Origines et son Développement Jusqu'à la fin du XVIe siècle, (Paris: E. Thorin, 1882), 162-163.

論莎士比亞戲劇中的法語 33 Kate. Par la grace de deu an pettie tanes, Ie parle milleur Kate. Par la grace de Dieu, en petit temps, je parle meilleur. Coman se pella vou le peid e le robe. Comment s’appelez vous le pied et la robe. Allice. Le foot, e le con. Allice. Le foot,et le con. Kate. Le fot, e le con, ô Iesu! Ie ne vew poinct parle, Kate. Le fot, et le con? ô Jesu! Je ne veux point parler Sie plus deuant le che cheualires de franca, ce plus devant les chères chevaliers de France Pur one million ma foy. pour un million! Ma foi! Allice. Madam, de foote, e le con. Allice. Madame, de foot, et le con. Kate. O et ill ausie, ecowte Allice, de han, de arma, Kate. Oh! Est-il aussi? Ecoutez, Alice: de han, de arma; De neck, de cin, le foote, e de con. de neck, de cin; le foot, et de con. Allice. Cet fort bon madam. Allice. C'est fort bon, madame. Kate. Aloues a diner. Kate. Allons-y a dîner. 在第 1四開本（Q1）中，⾸句是⼀個糟糕的法語句式，“Kate. Allice venecia, vous aues cates en, Vou parte fort bon Angloys englatara, Coman sae palla vou la main en francoy.”（凱特：愛麗絲，請過來，你（說英語）四⼗年了，你的英國英語說得非常好。）它可能近似於倫敦南岸區的商⼈⽤語，由於英語南部⽅⾔的發⾳習慣⽽產⽣較多的發⾳錯誤，（venecia➱venez ici; vous aues cates en englatara➱Vous avez quarante ans; parte, palla➱parler; englatara➱Angleterre; Coman sae palla vou➱comment appelez-vous）。sae palla, sa pella, se pella[s’appelez]的動詞變位是不規範的，但 la main卻是正確的表達。冠詞是希臘語和⽇⽿曼語中的詞類，九世紀羅曼—法蘭克語既已使⽤定冠詞。在 1600年第 1四開本（Q1）中，“bras”（拉丁語 bracchium）是⼗⼀世紀通俗拉丁語詞語，⽽且拉丁語中沒有冠詞，⼗⼀世紀拉丁語指⽰代詞 ille/illa（那、那個）可能語義弱化為定冠詞 le/la，⼗三世紀奧伊爾語普遍使⽤定冠詞 le/la，la bras可能是 le bras的誤⽤︔詩⼈⾼渥在其法語詩歌中已經出現了對（陰、陽）詞性及其詞形變化的忽略︔bien（bene）是⼗世紀通俗拉丁語詞語，fort bon原本是 fort bien。Recontera（raconter）是古法語（羅曼—法蘭克語）詞語。與 ausi... que不同，au se ... Asie是⼀個英語化的⽚語。劇作者誇張地突出了[e]普遍發⾳為[a]，[a]有時發⾳為[e]，

彭建華 34 [e]有時發⾳為[i]，[o]有時發⾳為[a, u]。以下是 1623年第 1對折本（F1）第三場第4幕中的法語對⽩，其話語場景更為完整/合理。 Kathe. Alice, tu as este en Angleterre, & tu bien parlas le Language. 凱特：愛麗絲，你曾去過英格蘭，那裏的語言你也說得很好。 Alice. En peu Madame. 愛麗絲：知道一點，小姐。 Kath. Ie te prie m'ensigniez, il faut que ie apprend a parlen Comient appelle vous le main en Anglois? 凱特：我請求你，來教我，我應該學會說它。le main（手），用英語怎麼說？ Alice. Le main il & appelle de Hand. 愛麗絲：le main（手），這讀作 de Hand（手）。 Kath. De Hand. 凱特：de Hand（手）。 Alice. E le doyts. 愛麗絲：還有 le doyts（手指）。 Kat. Le doyts, ma foy Ie oublie, e doyt mays, ie me souemeray le doyts ie pense qu'ils ont appelle de fingres, ou de fingres. 凱特：le doyts（手指），我還差點忘了 doyt（手指）。但我已經想起來了，le doyts（手指）讀作 de fingres，是 de fingres（手指）。 Alice. Le main de Hand, le doyts le Fingres, ie pense que ie suis le bon escholier. 愛麗絲：le main（手），de Hand（手）；le doyts（手指），de fingres（手指）。我想，我是一個好老師。 Kath. I'ay gaynie diux mots d'Anglois vistement, coment appelle vous le ongles? 凱特：我很快學會了 2個英語詞語。le ongles（指甲）怎麼說？ Alice. Le ongles, les appellons de Nayles. 愛麗絲：le ongles（指甲），讀作 de Nayles（指甲）。 Kath. De Nayles escoute: dites moy, si ie parle bien: de Hand, de Fingres, e de Nayles. 凱特：de Nayles（指甲），聽我讀，告訴我讀得好不好。de Hand（手），de Fingres

論莎士比亞戲劇中的法語 35 （手指），de Nayles（指甲）。 Alice. C'est bien dict Madame, il & fort bon Anglois. 愛麗絲：讀得很好，小姐。是很好的英語。 Kath. Dites moy l'Anglois pour . 凱特：告訴我，le bras（手臂）怎麼說。 Alice. De Arme, Madame. 愛麗絲：De Arme（手臂），小姐。 Kath. E de coudee. 凱特：de coudee（手肘）怎麼說？ Alice. D'Elbow. 愛麗絲：D'Elbow（手肘）。 Kath. D'Elbow: Ie men fay le repiticio de touts les mots que vous maves, apprins des a present. 凱特：D'Elbow（手肘）。我現在把你教我的詞語全部重複一遍。 Alice. Il & trop difficile Madame, comme Ie pense. 愛麗絲：我想，這很難的，小姐。 Kath. Excuse moy Alice escoute, d'Hand, de Fingre, de Nayles, d'Arma, de Bilbow. 凱特：沒關係，愛麗絲，你聽我讀吧：d'Hand（手）, de Fingre（手指）, de Nayles（指甲）, d'Arma（手臂）, de Bilbow[誤讀]。 Alice. D'Elbow, Madame. 愛麗絲：是 D'Elbow（手肘），小姐。 Kath. O Seigneur Dieu, ie men oublie d'Elbow, coment appelle vous le col. 凱特：哦天主呀！我忘了是 d'Elbow（手肘）。le col（頸、脖子）怎麼說？ Alice. De Nick, Madame. 愛麗絲：De Nick（頸、脖子），小姐。 Kath. De Nick, e le menton. 凱特：De Nick（頸、脖子）。le menton（頦、下巴）怎麼說？ Alice. De Chin. 愛麗絲：De Chin（頦、下巴）。

彭建華 36 Kath. De Sin: le col de Nick, le menton de Sin. 凱特：De Sin[誤讀]。le col（頸、脖子）讀作 De Nick（頸、脖子）；le menton（頦、下巴）讀作 De Sin[誤讀]。 Alice. Ouy. Sauf vostre honneur en verite vous pronouncies les mots ausi droict, que le Natifs d'Angleterre. 愛麗絲：是的。真誠的說，你讀出來的這些詞語，就像英國人讀的那樣地道。 Kath. Ie ne doute point d'apprendre par de grace de Dieu, & en peu de temps. 凱特：我不懷疑我可以學會的，感謝天主，在短時間裏就學會。 Alice. N'aue vos y desia oublie ce que ie vous a ensignie. 愛麗絲：你是不是要忘了我教你的[詞語]？ Kath. Nome ie recitera a vous promptement, d'Hand, de Fingre, de Maylees. 凱特：沒有[忘記]，我立刻背給你聽。d'Hand（手）, de Fingre（手指）, de Maylees[誤讀]。 Alice. De Nayles, Madame. 愛麗絲：是 De Nayles（指甲），小姐。 Kath. De Nayles, de Arme, de Ilbow. 凱特：de Nayles（指甲）, d'Arma（手臂）, de Ilbow[誤讀]。 Alice. Sans vostre honeus d'Elbow. 愛麗絲：對不起，是 d'Elbow（手肘）。 Kath. Ainsi de ie d'Elbow, de Nick, & de Sin: coment appelle vous les pied & de roba. 凱特：好的，d'Elbow（手肘），De Nick（頸、脖子）讀作 de Sin。les pied（腳）和 de roba（禮袍）怎麼讀？ Alice. Le Foot Madame, & le Count. 愛麗絲：小姐，讀作 Le Foot（腳），和 le Count（禮袍）。 Kath. Le Foot, & le Count: O Seignieur Dieu, il sont le mots de son mauvais corruptible grosse & impudique, & non pour le Dames de Honeur d'vser: Ie ne voudray pronouncer ce mots deuant le Seigneurs de France, pour toute le monde, fo le Foot & le Count, neant moys, Ie recitera vn autrefoys ma lecon ensembe,

論莎士比亞戲劇中的法語 37 d'Hand, de Fingre, de Nayles, d'Arme, d'Elbow, de Nick, de Sin, de Foot, le Count. 凱特：Le Foot（腳），和 le Count（禮袍）？天主呀，這些是糟糕的、粗野的、極不雅致的、不合宜的詞語，對於我們⾼貴的女⼠來說。總之⾯對任何法國的貴族，我是不會說出 Le Foot（腳）和 le Count（禮袍）這些詞語的。我還是把已學過的所有詞語再讀⼀遍：d'Hand（⼿）, de Fingre（⼿指）, de Nayles（指甲）, d'Arme（⼿臂）, d'Elbow（⼿肘），De Nick（頸、脖⼦），de Sin[誤讀]，Le Foot（腳），和 le Count（禮袍）。 Alice. Excellent, Madame. 愛麗絲：很好，小姐。 Kath. C'est asses pour vne foyes, alons nous a diner. 凱特：一次就學這些。我們去吃飯吧。 在第三場第 5幕中，除開法國的地名和⼈名，多爾芬親王、布列塔尼公爵、法軍元帥的對⽩都使⽤了混雜法語、英語的語句，便於突出法蘭西宮廷的場景，例如，“Dolphin. O Dieu viuant: Shall a few Sprayes of vs,”（多爾芬親王：天主永在，我們有了幾個新⽀脈）。“Britaine. ...Mort du ma vie, if they march along Vnfought withall,” （布列塔尼：讓我死吧，如果讓他們恣意前⾏⽽不戰⾾阻⽌）。﹝1600 年第 1 四開本寫作 Constable. Mordeu ma via:﹞。“Constable. Dieu de Battailes, where haue they this mettell?”（法軍元帥：戰神呀，他們的這種勇氣是從哪裡來的？）在第三場第5 幕中，多爾芬親王的對⽩使⽤了混雜法語、英語的語句，“Dolphin. ...ch'ha: he bounds from the Earth, as if his entrayles were hayres: le Cheual volante, the Pegasus, ches les narines de feu.”（啊哈！它從地上躍起，彷彿身體輕盈如毫髮：佩加蘇斯的會⾶天⾺，它的⿐孔將呼出⽕元素）。在第三場第 7幕中，“Dolphin. Le chien est retourne a son propre vemissement est la leuye lauee au bourbier: thou mak'st vse of any thing.”（多爾芬親王：狗會轉回到它吐出的東西︔變得乾淨的豬會再入泥塘。⼀切⾏事你都能便宜⽤之）。前句是來⾃《聖經・新約・彼得後書》中的諺語“le chien est retourne à ce qu'il avait vomi lui-meme, et la truie lavee, à se vautrer au bourbier.”（⼀隻狗會轉回到它吐出的東西，⼀個傻瓜也會回到他的瘋狂。），但劇中對⽩稍有語誤。 (2) 在第四場第 1 幕中，亨利五世⽤英語、法語混合語句對厄平漢勳爵表達了感謝，“King. God a mercy old Heart, thou speak'st chearefully.”（亨利王：謝謝上帝，

彭建華 38 ⽼夥計，你說得令⼈振奮）。英軍營長⽪斯托的⼜令“Pistoll. Che vous la?”（「來的是誰」）是⼀句糟糕的法語。⽽喬裝的亨利五世的回答包含法語詞語“Harry le Roy”（哈利王）。在第四場第 2幕中，為了突出阿⾦泰爾的法軍營地這⼀場景，其中多爾芬親王、奧爾良公爵的對⽩包含混雜英語、法語的語句，和完全的法語語句。 Dolphin. Monte Cheual: My Horse, Verlot Lacquay: Ha. 多爾芬親王：騎上馬，我的馬！侍從，馬夫：快呀！ Orleance. Oh braue Spirit. 奧爾良公爵：哦，勇敢的精靈。 Dolphin. Via les ewes & terre. 多爾芬親王：快走，水和土元素。 Orleance. Rien puis le air & feu. 奧爾良公爵：可能還有氣和火元素？ Dolphin. Cein, Cousin Orleance. 多爾芬親王：還有天（泰空 aether），奧爾良老弟。 在第四場第 4幕中，皮斯托、英軍侍童與一個法國戰士（French Souldier）在阿金庫爾戰場上的對白包含法語語句，這三人都是劇中的小丑，法語作為外語，是喜劇性元素，尤其是突出了皮斯托的糟糕法語，後者還混合了愛爾蘭–凱爾特語。114 French. Ie pense que vous estes le Gentilhome de bon qualitee. 法國戰士：我想，你是一位出身高貴的紳士。 Pistol. Qualtitie calmie custure me. Art thou a Gentleman？ 皮斯托：高貴出身？蘇爾河畔的姑娘（Callin o custure me, Callino casturame, Cailín ó chois tSiúire mé，愛爾蘭–凱爾特語），你是一位紳士？ French. O Seigneur Dieu. 法國戰士：啊天主！ Pistol. O Signieur Dewe should be a Gentleman: 皮斯托：啊！「田主」應該是一位紳士。 French. O prennes miserecordie aye pitez de moy. 114 莎士比亞，《莎士比亞全集 亨利五世》，梁實秋譯，（北京：中國廣播電視出版社，2002），169。

論莎士比亞戲劇中的法語 39 法國戰士：給一點憐憫，可憐我吧！ French. Est il impossible d' eschapper le force de ton bras. 法國戰士：不可能逃脫你的控制力呀？ French. O perdonne moy. 法國戰士：啊！請饒恕我。 Boy. Escoute comment estes vous appelle? 英軍侍童：聽好，你叫什麼名字？ French. Mounsieur le Fer. 法國戰士：先生，勒–菲爾（鋼鐵）。 French. Que dit il Mounsieur? 法國戰士：先生，他說什麼？ Boy. Il me commande a vous dire que vous faite vous prest, car ce soldat icy est disposee tout asture de couppes vostre gorge. 英軍侍童：他讓我告訴你，你要快點準備（贖金），否則這位軍爺要立刻割斷你的咽喉。 Pistol. Owy, cuppele gorge permafoy pesant, 皮斯托：是的，割斷你的咽喉，蠢貨。 French. O Ie vous supplie pour l'amour de Dieu: ma pardonner, Ie suis le Gentilhome de bon maison, garde ma vie, & Ie vous donneray deux cent escus. 法國戰士：呵，為了天主的憐愛，我乞求你啦。我是來自高貴家族的紳士，確保我的生命，我給你 200金幣。 French. Petit Monsieur que dit il? 法國戰士：小先生，他說什麼？ Boy. Encore qu'il et contra son Iurement, de pardonner aucune prisonner: neant-mons pour les escues que vous layt a promets, il est content a vous donnes le liberte le franchisement. 英軍侍童：他說，饒恕任何戰俘，違背了他的誓約；但為了你許諾的金幣，他願意讓你自由，把你釋放。 French. Sur mes genoux se vous donnes milles remercious, et Ie me estime heurex

彭建華 40 que Ie intombe, entre les main d'vn Cheualier Ie peuse le plus braue valiant et tres distinie signieur d'Angleterre. 法國戰士：我將屈膝跪下，1,000次向你致謝，我認為，我是幸運的，落在一位騎士的手裏；我想，他是英格蘭最勇敢最英勇最傑出的領主。 Boy. Saaue vous le grand Capitaine? I did neuer know so full a voyce issue from so emptie a heart: 英軍侍童：你跟隨著這個大營長吧。我從未聽過這樣充實的聲音從一個虛空的心中發出。 被俘的法國貴族的對⽩，在理想的情境中，應該是優雅的中古法語（以巴黎⽅⾔為基礎的多伊語），然⽽它們卻顯得較為笨拙。對比 1600年第 1四開本中簡陋的法語對⽩，1623年第 1對折本中的法語對⽩更接近早期現代法語的通⾏表達⽅式，這些顯著的改變可能不是莎⼠比亞作為劇作家修改和增衍的。 Henry V (Quarto 1, 1600) Henry V (Folio 1, 1623) O Monsire, ie vous en pree aues petie de moy. O Seigneur Dieu. O prennes miserecordie aye pitez de moy. Est il impossible d' eschapper le force de ton bras. O perdonne moy. Monsier Fer. Mounsieur le Fer. Qui dit ill monsiere. Ill ditye si vou ny vouly pa domy luy. Que dit il Mounsieur? O Iee vous en pri pettit gentelhome, parle/ A cee, gran capataine, pour auez mercie/ A moy, ey Iee donerees pour mon ransome/ Cinquante ocios. Ie suyes vn gentelhome de France. O Ie vous supplie pour l'amour de Dieu: ma pardonner, Ie suis le Gentilhome de bon maison, garde ma vie, & Ie vous donneray deux cent escus. Petit Monsieur que dit il? Sur mes genoux se vous donnes milles remercious, et/ Ie me estime heurex que Ie intombe, entre les main d'vn Cheualier Ie peuse le plus braue valiant et tres distinie signieur/ d'Angleterre.

論莎士比亞戲劇中的法語 41 在第四場第 5幕中，法軍元帥、奧爾良公爵、多爾芬親王的對⽩中混合了英語、法語，或者是有語誤的法語語句，“Constable. O Diable.”（法軍元帥：噢，惡魔。）“Orleance. O signeur le iour et perdia, toute et perdie.”（奧爾良公爵：天主呀，現在失敗了，完全失敗了。）“Dolphin. Mor Dieu ma vie, all is confounded all, Reproach, and euerlasting shame Sits mocking in our Plumes. O meschante Fortune, do not runne away.”（多爾芬親王：我死定啦。⼀切全完了，⼀切！永恆的恥辱⾼居在頭盔的⽻飾上嘲笑。噢可惡的命運，不要逃離⽽去。） (3). 在第五場第 2幕中，亨利、凱瑟琳、宮廷貴婦的對話包含法語語句和英語、法語混合語句，這是⼗⼆世紀以來英格蘭王國的宮廷愛情，（常⾒於騎⼠敘事詩），可以追溯到奧維德《愛的藝術》所描述的古典場景。亨利笨拙地試圖⽤法語表達⾃⼰的戀愛之情，也籲求凱瑟琳公主⽤英語表達，“me vnderstand well”是⼀個法語化的英語語句。 Kath. Pardonne moy, I cannot tell wat is like me. 凱特：原諒我，我說不出口「什麼是愛你」 King. An Angell is like you Kate, and you are like an Angell. 亨利王：你像一個天使，凱特，你像一個天使。 Kath. Que dit il que Ie suis semblable a les Anges? 凱特：你是說「我像一個天使」。 Lady. Ouy verayment (sauf vostre Grace) ainsi dit il. 宮廷貴婦：是的，確實這樣，（請你原諒），他是這樣說的。 King. I said so, deare Katherine, and I must not blush to affirme it. 亨利王：我是這樣說的，親愛的凱瑟琳，我不會因為這樣說而羞赧。 Kath. O bon Dieu, les langues des hommes sont plein de tromperies. 凱特：啊天主，這個人的話是（機智的）甜言蜜語。 Lady. Ouy, dat de tongeus of de mans is be full of deceits dat is de Princesse. 宮廷貴婦：是的，公主的話是說，男人的嘴巴說出的滿是騙人的話。 Kath. Sauf vostre honeur, me vnderstand well. 凱特：請你原諒，我很理解你的意義。 Kath. I cannot tell wat is dat.

彭建華 42 凱特：我沒有明白所說的是什麼。 King. ...Ie quand sur le possession de Fraunce, & quand vous aues le possession de moy. (Let mee see, what then? Saint Dennis bee my speede) Donc vostre est Fraunce, & vous estes mienne. It is as easie for me, Kate, to conquer the Kingdome, as to speake so much more French: 亨利王：當我擁有法蘭西，而你也擁有了我，（讓我想想，接著怎麼說？聖丹尼快來幫我），於是法蘭西是你的，而你是我的。征服這個國家和說這些法語，凱特，對我來說還是很容易的。 Kath. Sauf vostre honeur, le Francois ques vous parleis, il & melieus que l'Anglois le quel Ie parle. 凱特：請原諒，你說法語比我說英語要容易千百倍。115 King. How answer you, La plus belle Katherine du monde mon trescher & deuin deesse. 亨利王：怎樣回答你？世界上最美的凱瑟琳，我親愛的、神聖的女神。 Kath. Your Maiestee aue fause Frenche enough to deceiue de most sage Damoiseil dat is en Fraunce. 凱特：尊貴的君王在用十分糟糕的法語來欺騙法國最聰慧的姑娘。 King. Now fye vpon my false French: by mine Honor in true English, I loue thee Kate.... 亨利王：好了，我的糟糕的法語真笨拙，我以名譽為誓，用純正的英語來說，我愛你，凱特…… Kath. Dat is as it shall please de Roy mon pere. 凱特：你所說的，可能會讓我的父王樂意。 Kath. Laisse mon Seigneur, laisse, laisse, may foy: Ie ne veus point que vous abbaisse vostre grandeus, en baisant le main d'une nostre Seigneur indignie seruiteur excuse moy. Ie vous supplie mon tres-puissant Seigneur. 凱特：不要，尊貴的君王，不要，不要這樣！我不願你因為吻了你的卑下的臣下而降低身份。原諒我，我把你尊崇為最有權力的君王。 115 莎士比亞，《莎士比亞全集 亨利五世》，219。

論莎士比亞戲劇中的法語 43 Kath. Les Dames & Damoisels pour estre baisee deuant leur nopcese il net pas le costume de Fraunce. 凱特：女士和小姐在結婚之前與人接吻，法國沒有這種風俗。 Lady. Dat it is not be de fashon pour le Ladies of Fraunce; I cannot tell wat is buisse en Anglish. 宮廷貴婦：說的是法國的女士沒有這種風俗。我不知道 buisse（baiser之誤，接吻，kiss）在英語中怎樣說？ Lady. Your Maiestee entendre bettre que moy. 宮廷貴婦：尊貴的君王，你知道的比我更多。 Lady. Ouy verayment. 宮廷貴婦：是的。確實這樣。 Exeter. ...in French: Nostre trescher filz Henry Roy d'Angleterre Heretere de Fraunce: and thus in Latine; Praeclarissimus Filius noster Henricus Rex Angliae & Heres Franciae. 厄克塞特：……法語寫作：我的女婿亨利，英格蘭國王，法蘭西王國繼承人。拉丁語寫作Praeclarissimus Filius noster Henricus Rex Angliae & Heres Franciae. 中古英語採⽤了⼤量法語辭彙，早期現代英語詞語並不容易從法語區分出來。在莎⼠比亞戲劇中，使⽤法語詞語 sans（=no）是常⾒的，例如，⑴. A confidence sans bound. ⑵. I doe sans question. ⑶. Sans teeth, sans eyes, sans taste, sans euery thing.等︔法語介詞 à 也是常⾒的，例如，take-a-your Rapier, Peace-a-your tongue, send-a you, tell-a-me, vherefore vill you not meet-a me? loue-a-me等。此外，caught, raught, rought, sought, straught, taught等都保留了法語拼寫形式。 七、結語 ⼀種語⾔總是特定⼈群說的語⾔，⼈們的活動範圍就是語⾔得以存在的空間。印歐語⾔可能是由遷徙到亞洲和歐洲的諸多部落所通⽤的語⾔，併發展成不同的⽅⾔，例如，梵語、希臘語、⽇爾曼語、凱爾特語、波羅的海沿岸⽅⾔等。詞語的發⾳、拼寫、意義和語法、⽤法對語⾔外部社會和歷史⼒量總是敏感的。在法蘭克、

彭建華 44 英格蘭封建王國中，法律、⾏政、教育和商業活動多語⾔的使⽤，⼀直是顯著的語⾔現象。1066年諾曼征服改變了英格蘭⼈們的⽣活、語⾔、⽂化藝術、風俗和建築⾯貌，由此英格蘭進入了多語⾔時代，宮廷和社會中上階層常常在法庭、官⽅和⽂學中使⽤法語，教會、教育、外交則使⽤中古拉丁語，⽽原初英格蘭本地⼈⼤多數說各種各樣的英語⽅⾔。每⼀個社會的官⽅語⾔轉變，往往需要較長的時間。英格蘭王國的語⾔改變需要數⼗年，甚至更長時間，盎格魯—撒克森語、諾曼—法蘭克語需要很長時間來實現混合，東區⽼倫敦城的「考尼克⽅⾔」（cockney）可能是受到法語影響的本地⼜⾳。⼗⼆世紀形成盎格魯—諾曼語（古法語）。 1400年亨利四世（Henry IV, r. 1399-1413）的宮廷是最早講英語的。亨利八世的宗教改⾰促進了英語作為單⼀民族語⾔的快速轉變，國王英語、純正英語是早期現代英語⾛向英語純潔運動的出發點。顯然，莎⼠比亞把都鐸王朝的歐洲、王國/王權、民族國家、民族語⾔觀念分散地表達在其許多戲劇中，例如，《約翰王》《亨利五世》《亨利六世》（第三部）與《亨利八世》等，法語作為⼀種外語，往往是法蘭西⼈的顯性語⾔標誌。莎⼠比亞常常使⽤英格蘭王國多語⾔社會的雙關語、語⾳近似、笨拙的外語等作為喜劇性元素。 英格蘭⽂藝復興早期，英格蘭（尤其是倫敦）依然處於通⾏英語、法語、拉丁語的多語⾔社會境況，莎⼠比亞戲劇中的語⾔是莎⼠比亞、國王劇團演員、倫敦的戲劇觀眾共同分享的、⽇常的通俗語⾔，法語、拉丁語本身並不是莎⼠比亞刻意使⽤的個性化語⾔。莎⼠比亞時期作為中古多語⾔社會的延續，殘餘了較多拉丁語、法語、威爾⼠語、蘇格蘭—愛爾蘭語（凱爾特語），不必突出英語法語的對⽴地位，英語、法語⼆者是⾼度混合的羅曼語內的現代語⾔，各⾃的後續演變愈來愈分開⽽已。⼈們不知道莎⼠比亞在公⽴學校或者在倫敦是否接受過正式的法語教育，他可能沒有直接學習過來⾃巴黎的早期現代法語。116 事實上，倫敦有較多來⾃法國的商⼈，莎⼠比亞更容易接受倫敦商⼈⼜中的法語。莎⼠比亞作為劇作家、演員，熟悉倫敦市民⼜中流傳的簡易的法語，尤其是泰晤⼠河南岸商業區流傳的簡易的、隨意的法語⽅⾔。莎⼠比亞戲劇中的法語，不再表現為中世紀英格蘭的多語⾔社會的宮廷語⾔和官⽅書⾯語⾔，也並不準確表現為諾曼—法蘭克語作為拉丁化的⾼盧—法蘭克語的⼀種⽅⾔，或者與中古法蘭克王國奧伊爾語的⼀種⽅⾔，甚至早期現代法 116 Schoenbaum, S. Shakespeare’s Lives, (Oxford: Oxford University Press, 1991), 88.

論莎士比亞戲劇中的法語 45 語。相反，莎⼠比亞戲劇中的法語只是⼀種形態模糊的外語，可能近似於⼗六世紀後期倫敦商⼈⼜中的法語，這是⼀種簡陋的、隨意的混合外語。作為明顯的⼜語法語，莎⼠比亞戲劇中的法語在拼讀上有較多的錯誤，即使與 1100-1500 年英格蘭法語⽂獻比較，例如⼗⼆世紀英格蘭本篤會的聖詩集，劇中的法語並不是好的⽂學語⾔。

彭建華 46 徵引書目 [1] Hitchner, B. ed. A Companion to North Africa in Antiquity. Oxford: Wiley-Blackwell Companion, 2022. [2] Miles, R. Carthage Must Be Destroyed: The Rise and Fall of an Ancient Mediterranean Civilization. London: Penguin, 2010. [3] Ashton, N., Simon, L., Stringer, Ch. eds. The Ancient Human Occupation of Britain. Elsevier: Academic Press, 2011. [4] Ball, M. J., Müller, N. The Celtic Languages, 2nd ed. London: Routledge, 2009. [5] Semple, E. Ch. The Geography of the Mediterranean Region: Its Relation to Ancient History. New York: H. Holt and company, 1931. [6] Payne, S. G. A History of Spain and Portugal. London: University of Wisconsin Press, 1973. [7] Hoyos, D. The Carthaginians. London, New York: Routledge, 2010. [8] Hogg, R. M. The Cambridge History of the English Language, Vol. 1: The Beginning to 1066. Cambridge: Cambridge University Press, 1992. [9] Dyer, J. Ancient Britain. London, New York: Routledge, 1990. [10] Hayes, A. Archaeology of the British Isles. London: B.T. Batsford Ltd, 1995. [11] Geoffrey of Monmouth, Hammer J. Historia Regum Britanniae: A Variant Version Edited from Manuscripts. Cambridge, Massachusetts: Mediaeval Academy of America, 1951. [12] Cunliffe, B. The Ancient Celts. Oxford: Oxford University Press, 2018. [13] Klausen, R. H. ed. Hecataei Milesii Fragmenta. Scylacis Caryandensis Periplus, Berolini: Impensis G. Reimeri, 1831. [14] MacAulay, D. The Celtic Languages. Cambridge: Cambridge University Press, 1993. [15] Cunliffe, B. The Extraordinary Voyage of Pytheas the Greek. New York: Penguin, 2002. [16] Ball, M. J., Müller, N. The Celtic Languages 2nd ed. London: Routledge, 2009.

論莎士比亞戲劇中的法語 47 [17] Ireland, S. Roman Britain: A Sourcebook. London: Routledge, 2009. [18] Bord, J., Bord, C. Ancient mysteries of Britain, London: Grafton, 1986. [19] Shotter, D. Roman Britain. London, New York: Routledge, 2004. [20] Julius Caesar. Gallic Wars, trans. by Edwards, H. J. London: William Heinemann, 1919. [21] Russell, P. An Introduction to the Celtic Languages, London, New York: Routledge, 1995. [22] Cornelius Tacitus. Agricola. Germania. Dialogus, trans. by Hutton, M., Peterson, W., Ogilvie, R. M., Warmington, E. H., Michael Winterbottom, New York: The Macmillan co., 1914. [23] Beda Venerabilis. The Ecclesiastical History of the English People, trans. by Bertram Colgrave, Judith McClure, Roger Collins. Oxford: Oxford University Press, 1999. [24] Byock, J. L. Viking Language 1: Learn Old Norse, Runes, and Icelandic Sagas. Barnsley: Jules William Press, 2013. [25] Hogg, R. M., Denison, D. A History of the English Language. Cambridge: Cambridge University Press, 2006. [26] Rickard, P. A History of the French Language. London: Routledge, Unwin Hyman, 2014. [27] Joëlle, D., Olivier, S. L'ancien et le Moyen Français. Paris: Presses Universitaires de France, 2012. [28] Perry, W. C. The Franks, from Their First Appearance in History to the Death of King Pepin. London: Longman, 1857. [29] Gregory of Tours. A History of the Franks, trans. by Lewis Thorpe. London: Penguin, 1974. [30] Howorth, H. H. The Ethnology of Germany. The Journal of the Anthropological Institute of Great Britain and Ireland, Vol.13 (1884). [31] James, E. The Franks. New York: Blackwell Publishers, 1988. [32] Sergeant, L. The Franks. From Their Origin As a Confederacy to the Establishment

彭建華 48 of the Kingdom of France and the German Empire. New York: G.P. Putnam's sons, 1898. [33] Drew, K. F. The Laws of the Salian Franks. Philadelphia: University of Pennsylvania Press, 1991. [34] Ayres-Bennett, W. A History of the French Language Through Texts. London, New York: Routledge, 1995. [35] Gasté, A. Les Serments de Strasbourg; Étude Historique, Critique et Philologique. Paris: E. Belin, 1888. [36] Allieres, J. Formation de la Langue Francaise. Paris: Presses Universitaires de France, 1996. [37] Peignot, G. Essai Analytique sur L'origine de la Langue Française. Dijon: V. Lagier, 1835. [38] Devillard, E. Chrestomathie de L'ancien Français. IXe-XVe Siècles: Texte, Traduction & Glossaire, Paris: C. Klincksieck, 1887. [39] Jacques, Ch. Histoire de la Langue Française. Paris: Presses Universitaires de France, 2006. [40] Littré, É. Histoire de la Langue Française: Études sur les Origines, L'etymologie, La Grammaire, les Dialectes, La Versification et les Lettres au Moyen âge. Paris: Didier, 1878. [41] Clédat, L. La Chanson de Roland, Traduction Archaïque et Rythmée, Accompagnée de Notes Explicatives. Paris: E. Leroux, 1887. [42] Haskins, Ch. H. The Renaissance of the Twelfth Century, Cambridge, Massachusetts: Harvard University Press, 1971. [43] Cruse, M. I. Illuminating the ’Roman d’Alexandre’: Oxford, Bodleian Library, MS Bodley 264: The Manuscript as Monument. Oxford: Boydell & Brewer, 2011. [44] Brown, C. ed. English Lyrics of the XIIIth Century. Oxford: The Clarendon press, 1932. [45] Ueltschi, K. Petite Histoire de la Langue Française, Le Chagrin du Cancre, Paris: Imago, 2015.

論莎士比亞戲劇中的法語 49 [46] Schacht, L. De Elementis, Germanicis Potissimum, Linguae Franco-gallicae. Berolini: F. Dümmler, 1853. [47] Brunot, F. Histoire de la Langue Française, des Origines à 1900. Paris: A. Colin, 1933. [48] Jourdan, A. J., Decrusy, A. À., Isambert, F. Recueil Générale des Anciennes Lois Françaises, Depuis l'an 420 Jusqu'à la Révolution de 1789. Tome 12. Paris: Belin-Le-Prieur, 1828. [49] Stevens, L. C. “How the French Humanists of the Renaissance Learned Greek”, Modern Language Association of America (PMLA), 65.2 (1950.03). [50] Booton, D. E. Publishing Networks in France in the Early Era of Print. Lindin, New York: Routledge, 2018. [51] Shaw, J. F. Contributions to a Study of the Printed Dictionary in France Before 1539. Toronto: University of Toronto, 1997. [52] Person, É. La Deffence et Illustration de la Langue Francoyse par Ioachim Du Bellay. Paris: Librairie L. Cerf, 1892. [53] Picoche, J. Christiane Marchello-Nizia. Histoire de la Langue Française. Paris, Nathan: Vigdor, 1999. [54] Collingwood, W. G. Scandinavian Britain. London: Society for Promoting Christian Knowledge, 1908. [55] Townend, M. Language and History in Viking Age England: Linguistic Relations between Speakers of Old Norse and Old English. Turnhout: Brepols Publishers, 2002. [56] Bates, D., Curry, A. England and Normandy in the Middle Ages. Hambledon, London: Hambledon Press, 1994. [57] Kibbee, D. A. For to Speke Frenche Trewely: The French language in England, 1000-1600. Its Status, Description and Instruction. Amsterdam: John Benjamins Publishers, 1991. [58] Cleaver, L. Writing History in the Anglo-Norman World: Manuscripts, Makers and Readers, C.1066-C.1250. Rochester: York Medieval Press, 2018.

彭建華 50 [59] De Gorog, R. P. The Scandinavian Element in French and Norman. New York: Bookman Associates, 1958. [60] Vising, J. Anglo-Norman Language and Literature. Oxford: Oxford University Press, 1923. [61] Holmes, U. A History of the French Language, New York: Farrar & Rinehart, 1938. [62] De Cassagnac, A. G. Histoire des Origines de la Langue Française. Paris: Firmin Didot Frères, Fils, et cie., 1872. [63] Pohl, B. Dudo of Saint-Quentin's Historia Normannorum: Tradition, Innovation and Memory. Rochester: York Medieval Press, 2015. [64] Steenstrup, J. Ch. Normandiets Historie under de Syv Første Hertuger, 911-1066. København: A. F. Høst & Søn, 1925. [65] Douglas, D. C. “Rollo of Normandy.” The English Historical Review, LVII.CCXXVIII, (1942.10). [66] Howorth, H. “A Criticism of the Life of Rollo As Told By Dudo of St. Quentin.” Archaeologia, 45.2, (1880). [67] Fouracre, P., Ganz, D. ed., Frankland: The Franks and the world of the early middle ages. Manchester: Manchester University Press, 2012. [68] Tahkokallio, J. The Anglo-Norman Historical Canon: Publishing and Manuscript Culture. Cambridge: Cambridge University Press, 2019. [69] Clover, B. The mastery of the French Language in England from the XIth to the XIVth Century. New York: Corning & Co., 1888. [70] West, C. B. Courtoisie in Anglo-Norman Literature. New York: Haskell House, 1966. [71] Shelly, P. V. D. English and French in England, 1066-1100 (thesis). Philadelphia: University of Pennsylvania, 1921. [72] Mugglestone, L. The Oxford History Of English. Oxford: Oxford University Press, 2006. [73] Georgi, D. Language made visible: The invention of French in England after the Norman Conquest (thesis). New York: New York University, 2008.

論莎士比亞戲劇中的法語 51 [74] Lloyd, C. Semantics and Word Formation: The Semantic Development of Five French Suffixes in Middle English. Oxford: Peter Lang, 2011. [75] Marckwardt, A. H. Historical Outlines of English Sounds and Inflections. Ann Arbor: George Wahr Publishing Co., 1951. [76] Bourguignon, A. Grammaire de la Langue D'oïl (Français des XIIe et XIIIe Siècles). Paris: Garnier Frères, 1873. [77] Wheeler, B., Parsons, J. C. eds. Eleanor of Aquitaine: Lord and Lady. New York: Palgrave Macmillan, 2003. [78] Legge, M. D. Anglo-Norman Literature and its Background. Oxford: Clarendon, 1963. [79] Harf-Lancner, L., Warnke, K. Les Lais de Marie de France. Paris: Le livre de poche, 1990. [80] Kibler, W. W. ed. Eleanor of Aquitaine: Patron and Politician. Austin: University of Texas Press, 2014. [81] Boyd, D. Plantagenet Princesses: The Daughters of Eleanor of Aquitaine and Henry II. Barnsley: Pen & Sword Books, 2020. [82] Green, V. The Medieval Author in Medieval French Literature. New York: Palgrave Macmillan, 2006. [83] Strohm, P. Middle English. Oxford: Oxford University Press, 2007. [84] Price, H. Th. Foreign Influences on Middle English. Ann Arbor: University of Michigan Press, 1947. [85] Burnley, D. The Language of Chaucer, Basingstoke. London: Palgrave Macmillan UK, 1983. [86] Vising, J. Anglo-Norman Language and Literature. Oxford: Oxford University Press, 1923. [87] Yeager, R. F. John Gower: The French Balades. Kalamazoo: Western Michigan University, 2011. [88] Smith, J. J. Studies in the Language of some Manuscripts of Gower’s Confessio Amantis (Dissertation). University of Glasgow, 1985.

彭建華 52 [89] Moore, S. Historical Outlines of English Phonology and Middle English Grammar. Michigan: G. Wahr, 1919. [90] Horobin, S., Smith, J. An Introduction to Middle English. Oxford: Oxford University Press, 2002. [91] Nevalainen, T. An Introduction to Early Modern English. Oxford: Oxford University Press, 2008. [92] Work, E. Wh. The Bible in English literature. New York: Fleming H. Revell company, 1917. [93] Norton, D. A History of the English Bible as Literature. Cambridge: Cambridge University Press, 2004. [94] Noble, R. Shakespeare's Biblical Knowledge and Use of the Book of Common Prayer, as Exemplified in the Plays of the First Folio. London: Society for promoting Christian knowledge, 1935. [95] Haigh, Ch. English Reformations: Religion, Politics, and Society under the Tudors. Oxford: Oxford University Press, 1993. [96] Smith, J. J. Essentials of Early English: An Introduction to Old, Middle, and Early Modern English. London, New York: Routledge, 2022. [97] Gunn, S. Henry VII’s New Men and the Making of Tudor England. Oxford: Oxford University Press, 2016. [98] Loewenstein, D., Mueller, J. ed. The Cambridge History of Early Modern English Literature. Cambridge: Cambridge University Press, 2002. [99] Lucas, D. W. Aristotle’s Poetics (Clarendon Greek Text and English Commentary). Oxford: Oxford University Press, 1968: lxii. [100] Alexander, C. M. S. Shakespeare and Language. Cambridge: Cambridge University Press, 2004: 18. [101] Edmondson, P., Wells, S. The Shakespeare Circle. Cambridge: Cambridge University Press, 2015. [102] Shakespeare, W. A Most pleasaunt and excellent conceited Comedie, of Syr Iohn Falstaffe, and the Merrie Wiues of Windsor. London: Printed by Thomas Creede for

論莎士比亞戲劇中的法語 53 Arthur Iohnson, 1602. [103] Martina, E. “The Use of Dialects and Foreign Languages in Shakespeare’s King Henry V.” English Studies 100.7 (2019). [104] Holinshed, R. Chronicles of England, Scotland, and Ireland, ed. by W.G. Boswell–Stone, London: Chatto and Windus, 1907. [105] Paulin, R., Poole, A., Holland, P. Voltaire, Goethe, Schlegel, Coleridge: Great Shakespeareans, Volume III, London: Bloomsbury Publishing Plc, 2010. [106] 莎⼠比亞，《莎⼠比亞全集 溫莎的風流婦⼈》，梁實秋譯，北京：中國廣播電視出版社，2002。 [107] Womersley, D. “France in Shakespeare's ‘Henry V’.” Renaissance Studies 9.4, France in the English and French Theatre of the Renaissance (1995.12). [108] 莎⼠比亞，《約翰王》，劉昊譯，北京：外語教學與研究出版社，2015。 [109] Jacob, E. F. Henry V and the Invasion of France. London: Praeger, 2008. [110] Turnbull, W. B. The visions of Tundale; Together with Metrical Moralizations and other Fragments of early Poetry, Hitherto Inedited, Edinburgh: T.G. Stevenson, 1843. [111] Loiseau, A. Histoire de la Langue Française: Ses Origines et son Développement Jusqu'à la fin du XVIe siècle. Paris: E. Thorin, 1882. [112] 莎⼠比亞，《莎⼠比亞全集 亨利五世》，梁實秋譯，北京：中國廣播電視出版社，2002。 [113] Schoenbaum, S. Shakespeare’s Lives. Oxford: Oxford University Press, 1991.

彭建華 54 作者簡介 彭建華，福建師範大學文學院副教授，主要從事歐洲文學研究、翻譯研究。1999-2024 年發表學術論文 100篇，已出版《梵語佛經漢譯的傳統》、《文學翻譯論集》、《文學翻譯研究概論》、《歐洲文學經典及其經典化》、《當代比較文學》、《現代中國作家與法國文學》、《現代中國的法國文學接受》等專著 7 種，在研課題：國家社科基金一般項目「莎士比亞戲劇的早期版本研究」（18BWW082） 通信地址：福建省福州市倉山區上三路 8號福建師範大學文學院 通訊電郵：fennuyulin@sina.com

澳門科技大學學報(人文社會科學版)第十九卷第一期 二零二五年三月，頁 55-76 DOI: 10.58664/mustjournal.2025.03.002 55 當代漢字設計路徑探究* 譚穎妮1，王雪青 （1.澳門科技大學人文藝術學院博士候選人） 摘要：本文從漢字的起源和造字原理出發，分析漢字的結構特徵與基本的設計方法。立足於當代設計的語境，闡述基於電子計算機介入設計所帶來的技術革命，從而導致當代文字設計在理念與方法上之全新升級。同時使用這種理念與方法從形與意的結合、字形本身的塑造、字詞句的組合、文字與輔圖的結合等幾個方面對漢字設計的路徑進行探究與解析。 關鍵詞：當代文字設計、漢字設計、設計路徑 * 收稿日期：2023年 09月 25日；通過日期：2025年 02月 17日。

譚穎妮，王雪青 56 Design Routes of Contemporary Chinese Typography Tan, Yingni1; Wang, Xueqing (1. PhD Candidate, Faculty of Humanities and Arts, Macau University of Science and Technology) Abstract: In this study, the structural characteristics and fundamental typographical elements of Chinese characters were investigated on the basis of the origins and principles of formation of these characters. This study also discussed how the technological reform following the introduction of electronic computers in typography elevated the philosophy and method of contemporary typography. Building on this new philosophy and method, this study explored and dissected various design routes of Chinese typography from various aspects as follows: combination of visual and semantic elements; visual design of characters; combination of characters, words, and sentences; and combination of text and graphics. Keywords: Typography; Chinese typography; Design route

當代漢字設計路徑探究 57 一、引言 漢字是記錄漢語的符號體系，是世界上最古⽼的⽂字之⼀，是中華⽂化瑰寶。從刻畫符號到甲骨⽂再到今天，漢字的發展⼀脈相承，雖歷經數千年的歲⽉洗禮，卻都保留著甲骨⽂演變⽽來的⽅塊字形式，形成了漢字⾃身獨特的結構特徵、創造邏輯以及形式語⾔。⽴⾜於當代⽂字設計視角，漢字設計受到全球設計思潮的影響，同時也受到技術與設備的變更，在這個過程中，漢字當代的設計意識有了哪些改變？漢字設計呈現怎樣的樣貌？漢字設計又有哪些設計路徑？正是本⽂所關注的。本⽂將從漢字的起源和造字⼿法出發，探索中國漢字⽂化的源流，結合當代的設計思潮的影響及技術的變更，從⽽尋找當代漢字的設計路徑。 二、漢字的起源及特徵 對世間萬物的象徵性摹寫是⼈類⽂明起源時構造⽂字符號的共同⽅式，全世界很多民族古代⽂字都從象形開始的，這種將客觀物體⽤抽象圖形表達的⽅式，也是漢字圖形化意識的起源。根據⽬前研究的漢字起源八千年說觀點，中國發現最早的刻畫符號是在河南舞陽賈湖遺址出⼟，發現典型刻畫符號⼗餘例，時代距今約 8000年左右，刻於龜甲、陶器、骨器、⽯器等載體上。1 甲骨⽂是刻在龜甲和獸骨上的⽂字，是迄今為⽌中國發現的年代最早的成熟的⽂字系統，發展得較為完備的甲骨⽂主要發現於河南安陽殷墟，距今已 3000多年。2 ⽬前已發現的甲骨約⼗五萬⽚，經科學家考古發掘的有三萬五千逾⽚，單字數量已逾四千字。3 賈湖符號中的有些構型與殷墟甲骨⽂中的字形較為接近，這是迄今所知年代最早的⼀種可與漢字起源命題相聯繫的刻畫符號。 1 廖志林，〈20世紀中國文字起源論說評述〉，《北方論叢》，第 4期（2007），63-66。 2 唐建，〈賈湖遺址新石器時代甲骨契刻符號的重大考古理論意義〉，《復旦學報（社會科學版）》，第3期（1992），94-107。 3 央 廣 網 ，「 我 國 甲 骨 文 發 現 總 計 約 十 五 萬 片 ， 單 字 數 量 逾 四 千 字 」，https://baijiahao.baidu.com/s?id=1749451398854262655&wfr=spider&for=pc，瀏覽於 2024 年 02 月26日。

譚穎妮，王雪青 58 （㆒）、漢字造字法 六書作為中國古⼈總結出來的⽂字分類⽅法，為分析解說漢字構形原理起到了巨⼤的理論指導作⽤，最早對六書作出解釋的是西漢劉歆，班固曾在《漢書》中轉引劉歆的話：「古者八歲入⼩學，故周官保氏掌養國⼦，教之六書，謂象形、象事、象意、象聲、轉注、假借，造字之本也。」不僅指出了六書的具體內容，⽽且還把它當做造字的理論依據。許慎作《說⽂解字》，對六書的名稱和次序做了微調，他把六書的內容概括為象形、指事、會意、形聲、轉注、假借，並對每⼀種類別舉例解說，較為通俗易解影響非常深遠。 其中，在視覺傳達領域的影響及延續較為深遠的主要是象形、會意、形聲三⼤類造字法，是漢字藝術及設計⼀直傳承至今的常⽤⼿法的開端，比如抽象圖形化、意象化、重構漢字、合⽂等。 象形造字法是指通過對客觀對象的象徵性摹寫，以構成⽂字符號的⽅式，許慎的《說⽂解字》說：「象形者，畫成其物，隨體詰詘，⽇、⽉是也」，意思是說⽤線條或筆畫把要表達物體的外形特徵勾畫出來，筆畫隨著物體輪廓的曲折⽽變化，這樣的⽂字就是象形字。相對於世界上很多民族的古⽂字，漢字相對來說抽象程度更⾼，象形程度更低。象形字在漢字的數量中沒有優勢，雖古漢字時期還保持著較⾼的象形程度，但根據⽤字數量上的需求越來越⼤，漢字的象形程度也在隸變以後逐漸消失，甚至有些內容已完全符號化，⼤多數象形漢字最後都發展成了組成漢字的基本構件或者部⾸。 會意造字法是把兩個或兩個以上表⽰不同意義的單字組合起來，表達⼀個新的意義的構字⽅法，有著顯著的直觀性特點。孫海波在《中國⽂字學》中指出：「⾃然之物可以象形圖之，⼈為之事必因形以標識之，至意有隱微，不可得⽽指實者，不得已則合⼆字之義以⽰其意，視之謂會意。」比如圖 1中「⾾」像⼆⼈搏⾾的樣⼦，「及」字表⽰追到⼈把他抓住，基本上都可以⽤⼀幅圖畫來表⽰，讓⼈腦海中⽴即浮現出⼀幅幅對應的圖畫，⼗分⽣動。

當代漢字設計路徑探究 59 圖1：會意字「鬥」和「及」4 形聲造字法是在象形字、指事字、會意字的基礎上形成的，屬於合體造字法，由兩部分組成及形旁和聲旁，又稱義符和聲符，形旁表⽰字的意思或類屬，聲旁則表⽰字的相同和相近發⾳。形旁⼀般由象形字或指事字充當，聲旁可以由象形字、指事字、會意字充當。作為記錄語⾔的符號，⽂字天然需要具有傳聲達意兩重功能，漢字也不例外，在具體實踐中，⼈們發現⼤量抽象的⾳符既無「事」可「指」，又無「形」可「象」，更無「意」可「會」， 形聲造字法就是在解決這⼀⽭盾，⽽且是最⾼產的造字形式，佔《說⽂解字》中收錄漢字比例的 80%以上。例如圖 2 中「漁」字形旁是三點⽔，表⽰與⽔相關，聲旁是⿂，表⽰它的發⾳與⿂字⼀樣。 圖2：形聲造字法「漁」5 （㆓）、漢字的格局㈵徵 漢字是以形表意的⽂字，基本上每個漢字都各具含義，拉丁⽂則是由不同的字母組合⽽成的⾳標復合詞。在此，筆者將兩種⽂字體系從「⾳、形、意」上進⾏對比，以尋找漢字基於拉丁⽂字的獨特之處，有助於在進⾏漢字設計時更加符合其⾃身特點。 4 中國文字博物館，「紀綠片《中國文字發展史》第八集：漢字構形 會意」， 2020，https://mp.weixin.qq.com/s/ROagiWl1n-5ZFIWiG0wEJA，瀏覽日期 2023年 09月 17日。 5 中國文字博物館，「紀綠片《中國文字發展史》第十一集：漢字構形 形聲」，2020，https://mp.weixin.qq.com/s/FLW80eB48ncRSMuqFDtwwQ，瀏覽日期 2023年 09月 17日。

譚穎妮，王雪青 60 從「形」的視覺效應角度來看，漢字為⽅塊字、拉丁⽂是字母組合⽂字，漢字顯得更加獨⽴、有⼒、堅實，⽽拉丁⽂字通過不同數量的字母組合，在形式上更加具有「節奏」與「韻律」︔「⾳」⼆者都兼⽽有之，此研究從視覺設計角度切入則不做過多贅述︔「義」的層⾯，兩者差別最⼤，拉丁⽂字的字母本身不具字意，需要通過字母間的組合形成字意，漢字獨特之處是⼤部分漢字單字就有意義，這也是從漢字起源過程中的象形、會意、形聲造字法就已經鋪好的積澱。 從傳達的含義角度來看，⼤部分漢字是既具有形也具有意的，在設計層⾯上比西⽅拉丁字產⽣了多⼀種可能性，我們可以利⽤形意的區別，不僅能從形態上做⽂章，還可以通過形與意的穿插、借鑒、結合等⽅式作為設計的表現形式，這是漢字相比較拉丁⽂字來說得天獨厚的先決設計條件，同時也是漢字設計中⼀條獨特可以被利⽤的思路和路徑。 從排版⽅式來看，橫豎書寫排版⾃如是漢字在設計上的⼀⼤優勢，⽽拉丁⽂只能橫寫橫排。幾千年來漢字都是豎寫豎排的，從甲骨⽂到鐘鼎⽂再到鐫刻、⼿寫或雕版印刷，從簡牘到⼿卷再到冊⾴、書籍等，⽂字的安排都是以字⾃上⽽下，⾏⾃右⽽左為基本原則。近代以來，特別是新⽂化運動之後，順應西學東漸的形式和需要，遂與西⽂⼀致改為橫排，規則改為字⾃左⽽右，⾏⾃上⽽下。⽽與此同時，漢字豎排的特點並未被摒棄，因此⽂字竪排也是中國傳統書籍排版中⼀個顯著的特點。 三、文字設計從傳統到當代的意識變遷 （㆒）、㆗華文化發展㆗的漢字圖形化意識 漢字作為⼀種獨特的中華⽂化符號載體，在數千年的發展歷程中被廣泛應⽤於政治、⽣活、宗教、藝術等領域當中。中國⾃古有「書畫同源」⼀說，漢字的圖形化設計意識在漢字的起源、造字的邏輯、發展的演變、當代的設計層⾯，都有著重要的體現，不同時期和朝代的漢字圖形化意識有著不同的理念認知和實踐⽅法。除了⾦篆隸楷宋⿊這條正體漢字以及書法的發展之外，漢字的藝術性、圖案化的特徵是與⽣俱來的，從商朝青銅器上的氏族徽號，戰國時期的⿃蟲書，到唐代的花⿃字、⾶⽩書等，再到服飾⽤品、民間⼯藝品、建築景觀等領域的裝飾性漢字設計，形式

當代漢字設計路徑探究 61 豐富載體多樣。眾多傳統漢字的表現形式是中國⼈⽣活中的⽂化符號，充分展現漢字集合象形、會意、表⾳於⼀體的獨特之處，當代漢字設計也能從傳統漢字中源源不斷地汲取設計⽅法、展現形式等。 合⽂也稱合書（⾒圖 3），指由兩個以上的⽂字緊湊地組合成⼀個相對獨⽴的結構，形式上類似製造了⼀個新的⽂字，這種設計⽅法可以追溯到漢字創製初期的會意、形聲造字法，在追求氣勢和筆⼒的傳統書法中所⾒不多，但在民間藝術中很受歡迎，並且在⼆⼗⼀世紀⼈類的⽇常⽣活中仍能時常⾒到它們的身影。當代視覺傳達設計中的 Logo標誌設計、廣告設計、海報設計、⽂創設計中，也常常⽤到「合⽂」這種⽅法，比如⾦⾨縣政府的 logo（⾒圖 4），將「⾦」和「⾨」⼆字相結合，共享其中⼀些筆畫，結合在⼀個⽅塊字中，最後⽤顏⾊區分的⽅式加強識別性，是當代漢字設計中較為常⽤的⼿法之⼀。 圖3：合文組合吉祥語「黃金萬兩」和「招財進寶」6 圖4：金門縣政府logo7 6 陳楠，《中國漢字設計史》，（武漢：湖北美術出版社，2021），189。 7 花瓣網，金門縣政府 logo設計，https://huaban.com/pins/1699011054，瀏覽日期：2024年 02月 26

譚穎妮，王雪青 62 中華民國時期是中國字體設計有所突破及發展的年代，尤其是其中的美術字。隨著國內崇尚科學與民主的新⽂化運動影響，設計藝術家們敢於打破原有規則，將國外的藝術設計及觀念引入到漢字設計中，形成了具有中國特⾊的現代化⽂字設計形態。魯迅的 《吶喊》在 1926年第四版印刷時（⾒圖 5），以隸書書寫設計了封⾯字體，並誇⼤了三個⼜字，以⽰吶喊之意。1928年《進⾏曲選》（⾒圖 6）的封⾯漢字筆畫構成中，設計師融合了管樂器的局部造型和⼀些羅⾺字母，在形式表現上體現了民族化和西洋化風格的交集和共存。 圖5（左）：魯迅《吶喊》封面設計8 圖6（中）：錢君匋為《進行曲選》設計的封面9 (㆓)、數碼時㈹前的文字設計技術演變 在數碼時代尚未來臨之前，⽂字設計的⼿法是受到限制的。比如，對於字體的選擇，⽂字⼤⼩、粗細、間距、⾏距以及圖⽂位置的調動等等，在當代看起來再簡單不過的設計動作，卻在三、四⼗年以前需要花費⼤量的時間與精⼒，這些動作在 日。 8 魯迅，《吶喊》第四版，（北京：北新書局，1926.05），封面。 9 白蕊先，《進行曲選》，（上海：開明書店初版，1928），封面。

當代漢字設計路徑探究 63 「⼿⼯時代」都需要經過複雜的技術操作，這些設計細節也均不是設計師可以獨⾃⾃由把控的。 印刷⼯藝與視覺傳達設計的關係是相輔相成的，設計師做每⼀件設計都必須考慮視覺效果和印刷再現等多⽅⾯的因素。視覺傳達設計作為⼀⾨實⽤藝術，必須依靠技術體現，設計中的各種形式、⽂字圖案都從最開始的徒⼿，發展成為借助印刷⼯藝這⼀媒介最終呈現。 印刷⼯藝的發展要從中國古代四⼤發明之⼀的印刷術開始，經北宋慶歷間（1041年至 1048 年）的畢昇發展和完善，產⽣了泥活字。1448 年，德國匠⼈約翰尼斯·古登堡（Johannes Gultenberg）發明瞭鉛活字印刷術，相對於徒⼿抄⼤量的⽂字更加經濟實⽤且效率⾼，書籍的⼤量出版促進了⽂化知識的傳播，西⽅現代設計意識開始萌發，同時也推動了⼗九世紀西⽅⼯業⾰命的發⽣。⼆⼗世紀六⼗年代，乾轉印字體（dry-transfer type）是設計⾏業的重要設計⼿段之⼀，比起⼿繪或是等印刷坊的試印紙條，使⽤刮字紙⼤⼤節省了設計時間，在當時，刮字紙基本上成為設計師必備物品之⼀，在標題設計上的運⽤較為廣泛。1924年，⽇⽂寫真植字機在⽇本被發明，⽂字間距的調整、字號的放⼤縮⼩、字形的寬扁之類的變形都是利⽤光學原理來進⾏，雖然仍然帶有很⼤的局限性，但已經在原有基礎上有了好⼤的提⾼，因此寫真植字在當時那個年代（1960～1970）對於平⾯設計屆的影響是很⼤的。隨著機械⽂明的快速發展，寫真植字機漸漸消失在了歷史中被電腦所替代。 四、現代主義視覺語境下的文字設計 （㆒）、技術的變遷 技術的變遷直接影響著⽂字設計⽅式和設計觀念，在 80年代末 90年代初，電腦技術的引入徹底改變了視覺傳達設計領域，這⼀變⾰也成為新舊⽂字設計的分界線。電⼦計算機技術的介入從根本上顛覆了傳統的設計⽅法，設計師能夠完全掌控從設計到排版再到印刷之前的所有過程，將設計的「命運」牢牢地掌握在⾃⼰的⼿中。這場視覺設計「⾰命」從⼯具的改變過渡到思維的改變，再到⽂字設計中的圖形化意識崛起，對⽂字的視覺表現的最⼤化追求成了⼀種新的「設計時尚」。⽂字設

譚穎妮，王雪青 64 計的地位也從最初的輔助地位慢慢地從後排⾛向前排，並逐步形成了「⽂字就是圖形」的新的設計理念。張⼤鵬在其論⽂《中國國際海報雙年展視覺敘事的差異與演變研究》中將 2003-2019年入選的中國國際海報雙年展的所有作品分為五個類型（⾒表 1）並進⾏了統計與對比，其中純⽂字海報的量在 2019年時已在五種類型中達到第⼀（⾒表 2）。 表1：中國國際海報雙年展入選作品類型10 表2：中國國際海報雙年展入選作品類型絕對數量比較11 10 張大鵬，《中國國際海報雙年展視覺敘事的差異與演變研究》，（浙江：中國美術學院博士論文，2020），69。 11 張大鵬，《中國國際海報雙年展視覺敘事的差異與演變研究》，68。

當代漢字設計路徑探究 65 電⼦計算機字庫產業也隨著電⼦計算機的普及發展得越來越完善，豐富的字庫也為⽂字設計領域帶來超乎想像的可能性。《字體設計與應⽤》⼀書中說到：「在今天的設計中，涉及字體設計的問題時，主要包括兩個⽅⾯的知識能⼒。⼀是對⽂字的創造，即『造字』，包含了運⽤各種⼿段與⼯具（含對電⼦計算機的使⽤）去創造新的⽂字形態：新的⽂字關係︔⼆是合理地使⽤電⼦計算機字庫中豐富的字體，對其進⾏合理的，尤其是怡如其分的應⽤，即『⽤字』。」12 ⽤字，看似簡單，卻能夠為我們的設計帶來多樣化的可能性，更能夠為我們拓展設計思維疆界，毫無疑問，字庫字體的發展為設計師提供了⼀個更加寬闊的設計游刃空間。 計算機軟件的發展也非常迅速。1982年，剛登上 Times雜誌封⾯的史蒂夫·喬布斯（Steve Jobs）⾼價收購⼀家剛成⽴不久的打印軟件公司 Adobe的兩成股權，接下來的六年⾥，這家公司發佈了兩款劃時代的軟件：Illustrator 和 Photoshop—當今設計師們的必備技能，⾃此，不斷發展和更新的計算機軟件也為設計師的⼯具提供了極⼤的便利與更多的可能性。 ⼆⼗⼀世紀以來，隨著新媒體、⼈⼯智能、數字隨機⽣成等技術的發展，⽂字設計的⼿段與內容變得多元化，傳播環境也從單落點、單形態、單平台的形式向多平台、多落點、多形態轉變，使傳統媒介和新媒體以⼀種相互融合的⽅式實現多元共同發展，也為閱讀⾏為改變提供了環境與契機。13 近⼗年出現的動態⽂字標誌、動態⽂字海報、多維可變字庫、互動⽂字裝置等越來越受推崇。以「⼈⼯智能」「機器學習」為代表的科技發展，能讓計算機通過對⼤數據的分析捕捉規律並學習創造，14 讓計算機幫助設計師完成基礎性、重復性的⼯作，在⼤體量的漢字字庫設計領域得到很好的應⽤。 (㆓)、設計觀念的迭㈹ 現代主義設計理念是在西⽅的⼯業社會發展起來，西⽅現代設計意識的萌芽可追溯到古登堡⾦屬活字印刷術的廣泛應⽤，該技術的推動使得⼤量書籍的出版和傳播成為可能，為⽂字設計的現代主義進程打下基礎，進⽽推動了⼗九世紀西⽅第⼆ 12 王雪青，《字體設計與應用（升級版）》，（上海：上海人民美術出版社，2017），22。 13 李恆、楊波，〈復合媒體背景下的文字設計〉，《數字通信世界》，第 9期（2022），191-193+196。 14 陳楠，《中國漢字設計史》，（武漢：湖北美術出版社，2021），405。

譚穎妮，王雪青 66 次⼯業⾰命的興起。在受到「新⽂字設計運動」和「包豪斯」等思潮的影響下，審美標準和⽣活⽅式對設計提出了與機械時代相適應的要求，逐漸形成了以美觀和實⽤相結合的「現代」⽣產和⽣活⽅式。這種趨勢強調設計的功能性和美學價值的融合，以滿⾜當代社會對於設計的多重需求，中國的⽂字設計也是從⼗九世紀中葉開始向西⽅現代化轉型。15 ⼆⼗世紀⼆⼗年代的西⽅現代主義設計浪潮提出了新字體設計的觀念，主張字體由功能決定形式。「索緒爾式語⾔學」也在⼆⼗世紀前期逐漸成熟，為六七⼗年代的符號學發展提供牢固的根基。⼆⼗世紀三⼗年代後，研究者們將格式塔⼼裡學的⽅法具體運⽤在美學領域中，促進了具有格式塔傾向的美學研究。中華民國時期的設計家們吸收了新字體設計觀念，在漢字設計時將筆畫概括為簡約的幾何圖形，字形也更加簡練鮮明。新⽂化運動之後，以魯迅為代表的⼀批現代⽂藝觀念作家帶動了中國⼤量的⽂獻翻譯與書籍⽣產，中國的書籍裝幀逐漸向書籍設計的概念轉換，認識到⽂字設計不單只是對幾個字的簡單裝飾，有了版式設計的整體意識。 ⼆⼗世紀五⼗至六⼗年代，現代設計發展到瑞⼠設計階段，也就是「國際主義風格」，在排版設計上的主要特點是使⽤無襯線字體和簡潔的幾何圖形，以及遵循⼀定比例的網格系統，各種⽂字變化關係中的技巧和⼿段被充分開發並靈活運⽤。發展至⼆⼗世紀七八⼗年代，西⽅開始向過於嚴謹理性的現代主義發起挑戰，新浪潮設計湧現，對圖像、⽂字和排版進⾏了實驗性的探索，其中⽂本被解構，閱讀⽅式呈現出更複雜、層次豐富和開放性的特徵，從⽽提升了畫⾯的趣味性。同時期的中國，在 1987年印刷業告別「鉛與⽕」迎來「光與電」的漢字傳播時代，九⼗年代初開始漢字設計出現新變化，許多學校⽤「字體設計」課程代替「美術字」課程，開始進⾏系統的字體設計教學及研究。16 ⼆⼗⼀世紀的當代，計算機技術對於設計的介入如同⼀場⾰命，使得⽂字設計發⽣重⼤的變⾰，更⾼效的設計流程和可塑性更強的軟件⼯具為設計師給帶來了無限的可能，也從根本上真正的顛覆了以往的⼯作⽅式。更加重要的是，這場技術⾰命為⽂字設計所帶來的不僅是技術上的便利與進步，新的技術、新的⼯作⽅式也無 15 盧西，《迭代與變異—文字設計觀念與視覺效應研究》，（杭州：中國美術學院博士論文，2021），34。 16 王受之，《世界平面設計史》，（北京：中國青年出版社，2002）。

當代漢字設計路徑探究 67 形中促發了⼈們設計理念與設計思維的改變，思想與技術的結合使得⽂字設計⾛向了更加讓⼈⽿⽬⼀新的境地。 五、當代漢字設計路徑 儘管設計的⼯具可以為我們帶來無窮的可能性，然⽽，主導設計的仍然是⼈的⼤腦與思維，設計師應該針對設計的需求，⽤智慧和審美結合技術⼿段，去解決設計的問題，基於⽂字設計，則應該根據需求，尋找最為有效的表現⽅式。設計是⼀項解決問題的過程，針對不同的⽂字需求、⽂字數量，我們在設計中會尋找不同的應對策略。無論是⽤漢字設計標誌、符號，還是⽤漢字設計海報亦或是設計書籍封⾯、內⽂排版等，根據⽂字數量不同都會有不⼀樣的設計⽅法路徑。 歷史的積累給予當代漢字設計積澱與靈感，我們在擁抱漢字歷史的同時，也要接納國際當代設計觀念，借鑑當代設計理論及⼿法做好漢字設計，讓中國的漢字設計更加國際化的同時保留本⼟⽂化，讓漢字設計發揮其⾃身最獨特的特性。根據漢字造字原理、結構特徵以及當代設計語境，⼀般情況下，我們可以將漢字設計路徑分為以下幾個⽅⾯：形與意的結合︔字形本身的塑造︔字詞句的組合︔⽂字與輔圖的結合，並運⽤適宜的設計⼿法調整漢字的⼤⼩、粗細、⾊彩、形狀、肌理、⽅向、明暗、空間等。 （㆒）、形與意的結合 ⾃漢字造字之初，⼤部分漢字都具有⼀字⼀意的特點，這個特點是漢字所獨有的，也為設計師帶來了很多設計新思路。在理解漢字含義的基礎上探尋⼀些設計元素和靈感，選擇相應的字形視覺語⾔進⾏設計，使設計更加巧妙且具有深層意義。結合符號學的理論，17 漢字符號的能指則是「形」，所指則是「意」，在瞭解漢字含義的前提下，形與意的結合漢字設計，讓漢字符號的能指與所指、外延與內涵能更 17 符號學鼻祖索緒爾將符號界定為由能指（signifier）和所指（signified）相連接所產生的整體。在語言中能指指的是「音響形象」， 所指指的是「概念」，二者缺一不可。李思屈，《廣告符號學》，（四川大學出版社，2004），45。

譚穎妮，王雪青 68 加完美的匹配，從⽽滿⾜⼈們視覺、⼼理、應⽤及審美的各種需要。 比如漢字設計「日漸肥胖」（見圖 7）就將四個字豎排並逐步增加⽂字粗細，調整⽂字字重，到「胖」字時⿊⾊⾯積⼤到已經看不到筆畫無法辨識，但在漢字⽂化語境中能⾃然⽽然辨識出詞組，將「⽇漸肥胖」四個字的含義（能指）與越來越來胖的視覺感受（所指）相匹配，從視覺、⼼理層⾯都做到了恰到好處的含義吻合。書籍封⾯標題設計中的「隔間」⼆字（⾒圖 8）在⿊體的基礎上巧妙運⽤灰度的差別將⽂字視覺空間隔開，通過⽂字的語意進⾏設計表達。當代設計師梅數值在 2021年為「西戲」劇場設計的導視系統（⾒圖 9），延續了漢字「合字」的⼿法，「取票」⼆字將「取」字左右拆分，中間放入「票」字，傳遞從中取票的寓意，「⼜」中放「出」，巧妙傳達出此處就是出⼜的含意。可以看到，形與意巧妙結合的⽂字設計為設計增加了更多的內涵與趣味。 圖7（左）：文字設計「日漸肥胖」18 圖8（右）：蔡佳豪為《隔間》設計的封面19 18 2019年中國美術學院文字設計課程作業 19 薩瓦爾（Niki Saval）著、黃涓芳/王年愷譯，《隔間：我們如何從 19 世紀陰暗帳房走到 21 世紀Google人性化辦公空間》，（台北：貓頭鷹出版社，2016年）。

當代漢字設計路徑探究 69 圖9：「西戲」劇場導視系統，2021年 設計：702design （㆓）、字形本身的塑造 在設計實踐過程中，常常遇到⽤少數幾個字做設計的情況，比如品牌商標、標題⽂字、產品包裝名稱等，這種情況下，除了對字意的關注，更多的應該對⽂字本身的形態塑造，並尋求最為合理、有趣並具有視覺效應的形態設計，無論是對空間、構圖、⾊彩等要素的安排和組織，或是對字本身的、字與字之間的、字與載體之間的和諧關係均是如此。基於字形本身的漢字字設計⼀般來說「造字」的成分更是⽅法的主體。除了基礎的視覺傳達設計⼿法之外，還可以結合格式塔⼼理學理論中焦點、簡單、接近、相似、連續、封閉六個原則，20 將字形塑造成更⼤的形式意味或刺激⼒，這樣觀者在觀察過程中就會調動⾃⼰的經驗、情感、思根等去聯想、去填補，得出⾃⼰對圖形的獨特⾒解，這樣必然會發現以前未曾感受到的新的認識和新的形態。 例如中國中⾞的新商標（⾒圖 10），紅⾊的「⾞」字與⽩⾊的「中」字與巧妙融合，運⽤了圖底正負的設計⼿法，結合焦點原則，讓觀者第⼀眼看到時只看到「⾞」字，仔細看時便能看到其中的負形「中」，對稱均衡，四通八達。「設計⼒量」（⾒圖11）四個字的設計將可讀性幾乎降到了最低，在粗⿊體基礎上將筆畫轉換為⼤⾯積 20 格式塔（Gestalt）由三位德國心理學家威特海默（Max Wertheimer）、考夫卡（Kurt Koffka）和科勒（Wolfgang Kohler）所創立，認為人類的視覺感知具有整體性，整體是大於個體總和的。庫爾特·考夫卡 Kurt Koffka，《格式塔心理學原理（上下）（20世紀心理學通覽）》，（浙江教育出版社，1999年）。

譚穎妮，王雪青 70 的幾何⾊塊，讓畫⾯的視覺衝擊⼒達到最強，也將⽂字⼒量感體現得淋灕精緻。海報中「半⼼半技」（⾒圖 12）四個字的設計，在宋體的基礎上，運⽤了省俭成分的設計⼿法，省略了較細的橫筆畫但依然保持著⽂字的辨識度，因為封閉原則的應⽤使得觀者將⾃動在腦海中補全⽂字筆畫內容，右上角「半」字的⿈⾊點則是⽤了相似原則，將⿈⾊與灰⾊部分分離開，成為了畫龍點睛的視覺中⼼。 圖10（左）：中國中車標誌，2015 圖11（中）：中國美術學院設計藝術學院畢業展視覺設計，2018 圖12（右）：永井裕明退任展海報「半心半技」，2022 （㆔）、字詞句的組合 ⼀般來說，字詞句的組合設計指的是較多⽂字同時出現的狀況，當⼀個視覺載體中需要處理⼤量漢字時，需要考慮字詞句之間的排列組合關係，安排其中⽂本的主次層級，通過對字體、字號、粗細、疏密、空間、排列、佈局、⾊彩等相關因素的視覺組織決策，結合格式塔⼼理學理論的相關原則完成設計。比如信息量⼤的海報設計、書籍內容版式設計等，這⼀類漢字設計⼤部分屬於「⽤字」的設計範疇，同時需要對網格系統深入瞭解及靈活運⽤。值得單獨提出的是「⽂字反應堆」也包含其中，當兩種及以上不同的⽂字信息聚集在⼀起時，設計師就需要通過協調⽂字間的關係，突出⽂字集團化、層級化及差異化的視覺形態，在形象設計、⾴⾯設計、

當代漢字設計路徑探究 71 海報設計等情況下時常使⽤。21 如杉浦康平的中國展海報（⾒圖 13），排列組合關係處理得很有代表性，⽂字信息層級明確，「杉浦康平」為⼀級，「疾風迅雷」和展覽⽇期為⼆級，其他展覽主信息為三級，詳細介紹為四級，運⽤橫豎結合的排列形式，結合網格系統，使海報理性地傳遞出需要傳達的所有信息，留出的⽩⾊虛空間也平衡了畫⾯整體的視覺虛實關係。中國美術學院在 5·12汶川⼤地震時設計的「⽉祭」海報中（⾒圖 14），考慮這 16個字之間的組合關係，從上到下由⼩到⼤排列，和地⼼引⼒⽅向相統⼀，⽂字變⼤的同時字距逐漸縮⼩，逐步增強畫⾯的⾊塊重量，緊密的⾏距再加出框的⽂字，使⽂字在畫⾯中延伸，展現⼀種沈重⽽充滿⼒量的視覺語⾔。 2018年喜茶⾸次入駐⾹港，製作了⼀批港味濃厚的海報、貼紙、環保袋等，其中對漢字的設計就突出了「⽂字反應堆」效應（⾒圖 15），畫⾯中的各級元素通過⼤⼩、橫豎排列、顏⾊分類並相互配合，以組合形態呈現，突出⽂字集團化、層級化和差異化的視覺形態。故意打破常規也是讓設計變得有趣的⼿法，例如梅數植在2019年獲 Tokyo TDC Prize的作品「紅⽯板農貿市場」品牌形象和導視設計（⾒圖16），在⽂字橫豎排列上故意打破常態的閱讀順序，故意製造了⼀種閱讀上的「障礙感」從⽽獲得⼀種新鮮的閱讀體驗。字詞句組合的⽂字設計的要點是對多⽂字進⾏層級的區別，對⽂字的⼤⼩、粗細、空間、韻律之間的良好協調合作，來達到有效傳達信息，增強設計魅⼒的⽬的。 21 盧西，《迭代與變異—文字設計觀念與視覺效應研究》，（杭州：中國美術學院博士論文，2021），150.

譚穎妮，王雪青 72 圖13（左）：「疾風迅雷—杉浦康平雜誌設計半個世紀中國展」海報，2007 圖14（右）：2008年中國美術學院5·12汶川大地震「月祭」海報22 圖15：香港首家喜茶海報（局部）2018 圖16：梅數植 「紅石板農貿市場」品牌形象和導視設計獲Tokyo TDC Prize 2019 22 王雪青，《字體設計與應用（升級版）》，93。

當代漢字設計路徑探究 73 （㆕）、文字與輔圖的結合 ⽂字與輔助圖形相結合的漢字設計，是在以漢字為⽂本的基礎上，加入⼀些輔助圖形，從⽽加強設計的視覺效果。比如⼀條線、⼀個⾊塊或是⼀個放⼤的標點符號，都可以作為輔助圖形，從⽽加強視覺對比、豐富畫⾯形象。輔圖的樣式也可以從⽂化積澱中汲取養分，比如中國傳統建築屋頂、園林窗⼾、傳統吉祥圖案等，又或者是民國風、港風等以前出現過的平⾯設計形式，都可以拿來再設計。 如圖 17中的漢字設計案例，加入粗細位置都不固定的⽅框作為輔助圖形，與各種不同內容、數量、⼤⼩、橫豎排列的⽂字相結合，使輔圖和⽂字在畫⾯中產⽣不同且有趣的聯繫。圖 18的海報設計中，以標點符號作為輔助圖形，將放⼤的雙引號穿插在⽂字當中，豐富視覺畫⾯。輔助圖形式介入⽂字設計，是為了達到⽂字表達更加具有設計感的⽬的，協調好輔圖與主體⽂字的關係是至關重要的。從中國傳統元素中汲取的輔助圖形，可以讓漢字設計更加有⽂化底蘊。比如喜茶與上海美妝品牌百雀羚的聯名，以民國復古風為設計基調，設計了⼀系列联名款周邊，喜雀礼盒中的贈飲券（⾒圖 19）就從中華民國時期的設計中常⽤的花邊底紋、邊框、彩帶等輔助圖形元素，以中軸對稱的排版⽅式，呈現規矩且典雅的民國復古風。 圖17：中國美術學院學生作品：杭州火柴廠紀念版火柴23 23 王雪青，《字體設計與應用（升級版）》，134-135。

譚穎妮，王雪青 74 圖18（左）：海報設計 中國美術學院 學生作品 2009年 圖19（右）：喜茶百雀羚聯名款之贈飲券 2018年 六、結語 漢字從起源到在歷史長河中⼀脈相承的發展，形成⾃身獨特的造字與視覺特徵。在當下，⽴⾜於當代⽂字設計視角，漢字設計不僅受到技術⾰命的相助，也關係到設計師的能⼒及審美素質⾼下的影響，設計師的智慧與先進技術的推助使今⽇的漢字設計在視覺上呈現出百花齊放的樣貌。本⽂中基於當代漢字設計路徑的探究：形與意的結合、字形本身的塑造、字詞句的組合、⽂字與輔圖的結合是對漢字設計⽅法的初探與淺析。可以肯定在對全新的⽂字設計風格視覺效果最⼤化追求的⽬標下，設計道路更加寬廣與多擇，可謂「條條⼤路通羅⾺」，然⽽如何更好的瞭解漢字的形態，瞭解漢字的⽂化內涵，瞭解新時代⽂字設計的理念與⽅法，以及瞭解設計的⽬的與需求，則是我們能將漢字設計的⽔準推向更⾼境界的法寶。

當代漢字設計路徑探究 75 徵引書目 [1] 王受之，《世界平面設計史》，中國青年出版社，2022。 [2] 王雪青，《字體設計與應用（升級版）》，上海人民美術出版社，2017。 [3] 高秦艷，《身份的轉換—當代文字設計與圖形語言建構研究》，杭州：中國美術學院出版社，2020。 [4] 李恆、楊波，〈復合媒體背景下的文字設計〉，《數字通信世界》，第9期（2022）。 [5] 唐建，〈賈湖遺址新石器時代甲骨契刻符號的重大考古理論意義〉，《復旦學報（社會科學版）》，第3期(1992)，94-107。 [6] 孫海波，《中國文字學》，文求堂書店，1941。 [7] 許慎，《說文解字新訂》，中華書局，2002。 [8] 陳楠，《中國漢字設計史》，湖北美術出版社，2021。 [9] 張大鵬，《中國國際海報雙年展視覺敘事的差異與演變研究》，中國美術學院博士論文，2020。 [10] 廖志林，〈20世紀中國文字起源論說評述〉，《北方論叢》，第4期（2007），63-66。 [11] 盧西，《迭代與異化—文字設計視覺效應與觀念研究》，中國美術學院博士論文，2020。 [12] 中 國 文 字 博 物 館 ， 紀 綠 片 《 中 國 文 字 發 展 史 》， 2020 ，https://mp.weixin.qq.com/s/FLW80eB48ncRSMuqFDtwwQ，瀏覽日期2023年09月17日。

譚穎妮，王雪青 76 第㆒作者簡介 譚穎妮，澳門科技大學人文藝術學院設計學博士候選人，研究領域為視覺傳達設計、文字設計 通訊地址：澳門科技大學人文藝術學院 通訊電郵：313396256@163.com 通訊作者簡介 王雪青，澳門科技大學人文藝術學院特聘教授、博士生導師、全國模範教師。研究領域為視覺傳達設計與設計基礎教學，曾任法國埃克斯 (AIX-EN-PROVENCE) 高等美術學院教授，中國美術學院上海設計學院院長、教授，中國美術學院設計藝術學院院長、教授 通訊地址：澳門科技大學人文藝術學院 通訊電郵： xqwang@must.edu.mo

澳門科技大學學報(人文社會科學版)第十九卷第一期 二零二五年三月，頁 77-120 DOI: 10.58664/mustjournal.2025.03.003 77 規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點* 李浩 （澳門大學法學院博士研究生） 摘要：自〈通用數據保護條例〉（General Data Protection Regulation , GDPR）及〈施雷姆斯案〉（Schrems Ⅱ）生效後，寰球數據流動進入標準合同條款時代。促進內地與澳門特區的跨境數據傳輸，是建設「電子灣區」以及與葡語國家數字貿易港的應有之義。SCCs 以其便捷性與實踐性，系現有法律工具箱中可推動先導性創新並打破制度壁壘的首善之選。「9+1」傳輸結構僅為起點，SCCs 的規則銜接應指向「1+1+1」的制度未來。兩地合同條款在包括有關概念、適用前提、具體內容及程序等諸多方面存在差異。橫琴深合區具備頂層設計及「自下而上」的動態規範結構，在立法、行政、司法上保留著相當的制度彈性，為兩地的合同條款規則銜接提供試驗平台；其現有的體制改革經驗，如「一地兩檢」、法定機構的分工等，亦可作為 SCCs 系統協同的參考。在充分認識到兩地制度性差異的前提下，以粵澳深度合作區的制度創新為契機，可奠定標準合同條款基本原則與框架，消彌專業術語分歧，然後通過協議推出樣本合同，在執委會內設機構負責監督執行，從而構建「規則共商、平台共建、機構共管、成果共享」的聯動發展局面。如此可期實現對粵澳乃至區域信息產業與居民福利的提升，同時有效保障數字人權，為跨境數據流動提供帶有中國特色的「大灣區方案」。 關鍵詞：數據跨境傳輸、標準合同條款、規則銜接、橫琴深合區、制度創新 * 收稿日期：2023年 12月 21日；通過日期 2025年 02月 17日。

李浩 78 Rule Convergence and Innovation: The Construction of SCCs Regime for Data Transfer between Mainland China and Macau — Taking the Hengqin In-depth Cooperation Zone as the Entry Point Li, Hao (PhD Candidate, Faculty of Low, University of Macau) Abstract: After the effectiveness of GDPR and the Schrems II case, the global data flow seems to have entered the “era of Standard Contractual Clauses”. Promoting cross-boundary data transmission between the mainland and the Macao Special Administrative Region is essential for building an “electronic bay area” and a digital trade port with Portuguese speaking countries. SCCs, with its convenience and practicality, are the top choice in the existing legal toolbox that can drive leading innovation and break institutional barriers. The “9+1” transmission structure is only the starting point, and the rule convergence of SCCs should point to the future of the “1+1+1” system. There are differences in terms of contractual clauses between the two regions, including relevant concepts, applicable prerequisites, specific content, and procedures. Hengqin In-Depth Cooperation Zone has a top-level design and a “bottom-up” dynamic normative structure, retaining considerable institutional flexibility in legislation, administration, and judiciary, providing a experimental platform for the rule convergence of contract clauses between the two regions; Its existing experience in institutional reform, such as “one place, two inspections” and the division of functions among statutory bodies, can also serve as a reference for SCCs system harmonization. On the premise of fully recognizing the institutional differences between the two regions, taking the institutional innovation of the Guangdong-Macao In-depth Cooperation Zone as an opportunity, the general principles and framework of standard contractual clauses can be established, and distinctions in professional terminology can be eliminated. Then, sample contracts can be launched through agreements, and institutions within the executive committee can be established to

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 79 supervise and execute, thus building a interactive development status of “rule co-consultation, platform co-construction, institution co-management, and achievement co-sharing”. This can achieve an improvement in the information industry and residents' welfare in Guangdong, Macao, and even the region, while effectively safeguarding digital human rights, and providing a “Greater Bay Area Scheme” with Chinese characteristics for cross-boundary data flow. Keywords: Cross-boundary data transfer; Standard Contract Clauses; Rule convergence; Hengqin In-depth Cooperation Zone; Institutional innovation

李浩 80 一、問題的提出：標準合同條款的域際張力 在進入資訊數位化收集與轉移的階段後，1 數據即被稱為「原材料」2、「⽣產要素」，3 以及應納入貿易政策規管的「第五項⽬」。4 鑒於數據權擬表內容的內在緊張，實現信息流動的利⽤⽬的與保障程度的再平衡成為各法域數字治理的基本⽬標。5 連接全球「2+N」的格局中兩⼤代表性法域的〈隱私盾協議〉於 2020年的〈施雷姆斯案〉（Schrems Ⅱ）中被廢⽌，⽽標準合同條款（Standard Contractual Clauses, SCCs）在結合⼀系列的補強措施后卻被宣佈依然有效。6〈歐美數據隱私框架〉的⾯世亦無改其在全球跨境數據傳輸的主流地位。7 SCCs作為職能部⾨供給的「⼀攬⼦程序」，隨著限制的持續放鬆將在區域數據傳送中扮演更為重要的角⾊。制度設計者即以其充當粵港、粤澳雙向信息流動合作⽇程表的「排頭兵」。8 不局限於「9+1」， 1 在此之前，其主要傳遞形式依次為口耳相授及紙質媒介。參見楊崇蔚、廖志漢、廖志聰，《澳門個人資料保護制度》，（北京：社會科學文獻出版社，2014），258-259。 2 See Kunner C, European Data Privacy Law and Online Business. (Oxford: Oxford University Press, 2003), 9. 3 參見中華人民共和國中央人民政府，「中共中央 國務院關於構建更加完善的要素市場化配置體制機制的意見」，https://www.gov.cn/zhengce/2020-04/09/content_5500622.htm，瀏覽日期為 2020年 4月 9日。 4 See Dan Ciuariak and Maria Ptaskhina. “The Digital Transformation and the Transformation of International Trade.” Accessed July 27, 2023. http://www.iberglobal.com/files/2018-2/Digital_Transformation_Trade.pdf. 5 具言之，是為「人權 VS 財產權」的二元矛盾結構。參見於興中，《數字素養：從演算法社會到網路 3.0》，（上海：上海人民出版社，2022），75-76。 6 「2+N」即歐盟與美國，以及其他國家與地區的數據傳輸規範。參見廣東外語外貿大學粵港澳大灣區研究院課題組，〈數據要素跨境流動與治理機制設計—基於粵港澳大灣區建設的視角〉，《國際經貿探索》，第 10期（廣州：2021.10），87-88；美國由於實行監控計畫，被認為未能提供符合 GDPR與 CFR要求的保障水準；同時其巡查員機制（Budsman Mechanism）也被判定為妨礙司法救濟。基於此，歐盟法院在該案的判決中否定了歐委會的隱私盾決定（Privacy Shield Decision），同時對以 SCCs 為基礎的數據傳輸適用更嚴格的標準。數據控制者（Data Controllers）或操作者（Operators）通過簽署標準合同向外傳輸數據，要保證數據主體的權利得到與歐盟法實質性同等標準的保護，如有必要需採取額外措施填補第三國的法律漏洞。See European Parliament. “The CJEU judgment in the Schrems II case” Accessed October 30,2024. https://www.europarl.europa.eu/RegData/etudes/ATAG/2020/652073/EPRS_ATA(2020)652073_EN.pdf. 7 參見李艷華，〈隱私盾案後歐美數據的跨境流動監管及中國對策—軟數據本地化機制的走向與標準合同條款路徑的革新〉，《歐洲研究》，第 6期（北京：2021.12），36-37。 8 作為對 2023年 6月簽訂的合作備忘錄宏觀任務的首次回應（具體化），國家網信辦與香港創新科

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 81 該「明星⼯具」應更廣泛地在中國內地與澳⾨特區之間的數據傳輸中發揮作⽤。9 轉移雙⽅同為⼤陸法系且同受歐盟數據法影響，所存制度壁壘較易被打破。法律協同（legal harmonization）既是相對不同國家⽽⾔，也存在於同⼀國家內部多個法域。10 内地與澳⾨雖同屬中華⼈民共和國，然分別實⾏不同的法律制度，在數據傳輸⽅⾯仍存在⼀定差異。兩地合同條款交互非⽌需法律協同，且要達到「規則銜接、機制對接」的程度，將涉及實體法、衝突法規範及司法（執法）協助多個層⾯，故亟待進⼀步討論。11 近年來粵港澳⼤灣區以及橫琴粵澳深度合作區的建設為兩地合同條款規則提供了難得的制度創新風⼜。為了加強內地與港澳的合作，促進供給側改⾰與對外開放，粵港澳⼤灣區作為國家戰略被提上⽇程。橫琴深合區被認為是前者的補充與⼦項⽬，致⼒於構建「澳⾨—珠海」西岸極點。12 不僅相較⼤灣區內的試點如南沙、 技及工業局、個人資料隱私專員公署於 12月 13日聯合推出《粵港澳大灣區（內地、香港）個人信息跨境流動標準合同實施指引》及附件。又因該文件對個人信息處理者與接收方註冊地的限制（廣東省大灣區九市或香港特區），故而被部分評論家稱為「粵港版 SCCs」。另，為落實備忘錄中關於「共同制定粵港澳大灣區個人信息跨境標準合同並組織推動實施，加強個人信息跨境標準合同備案管理」的合作措施，國家網信辦與澳門特區政府經濟及科技發展局、個人資料保護局於2024年 9月 10日聯合發佈《粵港澳大灣區（內地、澳門）個人信息跨境流動標準合同實施指引》。該文件僅適用於廣東省大灣區九市與澳門特區，亦被稱為「粵澳版 SCCs」。 9 區別於已發佈的粵澳版 SCCs，本文研究的命題是實現中國內地與澳門特區 SCCs規則銜接在橫琴「試點突破」的前景。兩者雖存在重疊之處，但在實行地域、合作路徑與對接程度等都有很大差異。詳論之，前者僅在大灣區內實行，後者包括內地其他省市；前者以部門合作為主導，後者通過試點實行制度創新；前者以行政互動為主，後者還包括立法與司法對接。是以為實現此目的，需要對兩地的數據傳輸相關法律規定做系統性比較。尤其值得注意的是，本文的比較對象是內地法與澳門法，並不包括兩地聯合發佈的法律檔如粵澳版 SCCs。 10 See Muruga Perumal Ramaswamy, “Diversity in Data Protection Standards Governing Cross-border Commercial Transactions and the Role of GDPR as a Potential Harmonization Model” (Proceedings, International Conference on Technological and Social Innovations, Ambassador Bangkok Hotel and GrandPresident Hotel, Bangkok, 2019), 1. 11 「銜接」一詞被認為比「借鑒」、「吸收」等更強調內在的協調統一。參見李可、唐曉晴，〈橫琴粵澳深度合作區：理念創新與制度構建〉，《港澳研究》，第 1期（北京：2022.03），15；在廣東省政府印發的三年計畫中，有關部門專門使用了「實現大灣區數字化規制規則有效對接」的表達，突破了原來民商事規則的範疇，體現了包括 SCCs 在內的數據治理制度的複雜性。參見廣東省人民政府，「廣東省人民政府辦公廳關於印發「數字灣區」建設三年行動方案的通知」，http://www.gd.gov.cn/zwgk/gongbao/2023/31/content/post_4287722.html?version=slh，瀏覽日期為2023年 11月 7日。 12 參見中華人民共和國中央人民政府，「中共中央 國務院印發『粵港澳大灣區發展規劃綱要』」， https://www.gov.cn/xinwen/2019-02/18/content_5366593.htm#1，瀏覽日期為 2023年 7月 23日；參見國務院港澳事務辦公室，「中共中央 國務院印發『橫琴粵澳深度合作區建設總體方案』」，

李浩 82 前海與河套，甚至國內其他⾃貿區如上海臨港新⽚區，橫琴在合作廣度及創新深度⽅⾯皆處領先地位，可望成為兩地數據傳輸合同規則銜接的「試驗⽥」。與其他跨境數據傳輸⼯具如充分性認定、有約束⼒的公司規則、⾏為準則以及個⼈信息保護認證等不同，SCCs⾯臨的爭議與利益糾紛更少、實⽤性更強，同時在灣區內有著良好的實踐基礎（兩版標準合同及指引）。然「標準合同條款+橫琴」的議題並未如願受到數據法學家與⾏政管理學者的關注，13 顯⽰出學術研究對於實踐需求及法律變⾰的滯後性。標準合同條款呈現公私合體之特徵：形式上為合同條款，內容上由國家預定⽽強制納入，兼具個別規範與國家法規範雙重屬性。14 是以其協同⽴法不僅涉及民商事規則銜接，還需就⾏政法與部⾨規章變通進⾏討論。陸澳 SCCs ⽅案的落地，必然需要更豐富的制度想像⼒、更務實的設計思路與更精細的⽴法技術。基於此，本⽂並不聚焦於橫琴深合區提供的孵化平台對於兩地數據傳輸創新「何以能」，因該命題已在上⽂予以闡述︔⽽是著⼒於解決其「如何能」。具體⽽⾔，即回應以下問題：中國內地與澳⾨特區的合同條款在規範結構以及適⽤過程中，包括數據處理原則、專業術語、條款內容、監管機構等⽅⾯，存在何種不同？以上這些差異如何阻礙了兩地合同條款的規則銜接與數據的跨境轉移？橫琴深合區具備哪些獨特的制度框架與先試先⾏經驗，能夠為標準合同條款的發展貢獻何種可能性？粵澳版SCCs既「珠⽟在前」，應如何吸收其經驗又繼續突破，變「9+1」為「1（內地）+1（橫琴）+1（澳⾨）」？藉由深合區建設之機，怎樣根據兩地之制度特點與具體情況實⾏規則改進，推出⾏之有效的數據轉移合同「橫琴模版」，從⽽形成具有本⼟特⾊的「⼤灣區⽅案」，為全球區域合作及數據治理提供參考？ 二、内地與澳門特別行政區合同制度比較研究 由於歷史原因，中國內地與澳⾨遵循不同的規範演進路徑，發展出各具特⾊的 https://www.hmo.gov.cn/zcfg_new/gafg/gfxwj/202211/t20221125_23891.html，瀏覽日期為 2023年 7月 23日。 13 如用以「橫琴 標準合同條款」為主題在中國知網進行搜索，僅得到少量結果，且與橫琴深合區無關；以“Standard Contractual Clauses Hengqin”為關鍵字在 Google Scholar進行檢索，其結果多與橫琴的經濟建設與跨境合作有關，鮮見關於數據傳輸乃至標準合同的文獻。 14 參見金晶，〈作為個人信息跨境傳輸監管工具的標準合同條款〉，《法學研究》，第 5期（北京：2022.10），21-22。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 83 數據保護體系。內地已構建比較完善的數據法系統，包括憲法、民刑基本法、專⾨法、規章條例等，具有鮮明的社會主義成⽂法特點。⽽澳⾨規範系統涵蓋憲制性⽂件（特區基本法）、民刑基本法、專⾨法及執⾏細則，基本上沿⽤歐洲模式又結合⾃身的實際情況。15 概括⽽⾔，內地⽴法較遲，充分考慮了新技術如⼤數據、⼈⼯智慧的發展，呈現出後發優勢︔澳⾨則因為持觀望態度，儘管其⽴法⼀度有「亞太地區最強的個⼈資料保護法」的美譽，如今已有與現實脫節之虞。16 以合同條款作為跨境數據傳輸的合法性依據，主要⾒於兩地的專⾨性⽴法中，即内地的〈個⼈信息保護法〉（2021）與澳⾨的〈個⼈資料保護法〉（2005）。以下將從有關原則與概念、性質與適⽤前提、具體內容、程序及監管機構四個⽅⾯對兩地數據傳輸合同進⾏對比，以求對制度改進有所裨益。 （㆒）、㈲關原則與概念 兩地的數據規範效⼒具有全域覆蓋及輕微的外溢特徵，其中包括適⽤原則與概念。17 又因內地與澳⾨均將數據傳輸歸納為數據處理⾏為，故⽽其上位法對於數據處理⾏為的⼀般規定同樣適⽤於數據傳輸。關於數據傳輸的原則性規定多散佈於民事基本法以及專⾨法中，是整部法律的「精神標桿」，約束其後⾯的細則規定︔同時處於法律體系的中層，起到承上啟下的作⽤。現將雙⽅數據處理的基本原則呈列如下： 15 參見楊崇蔚，〈網路空間的個人資料保護—從監管當局角度分享澳門經驗〉，《網絡信息法學研究》，第 2期（北京：2019.12），79-80。 16 參 見 個 人 資 料 保 護 辦 公 室 ， 「 2012 年 度 年 報 」 ， https://www.gpdp.gov.mo/file/Annual_Report/2012_files/1_all.pdf，瀏覽日期為 2023年 7月 28日。 17 內地的數據立法除有特別規定進行變通外，當然施行於各省、自治區、直轄市，以及其下屬經濟特區、自貿區等；澳門則因享有高度自治權，其立法會所創設數據規治法律作用於特別行政區全境與授權管理區域（如澳門大學橫琴校區及橫琴口岸澳方管轄區等）。而頻繁的立法機構、法律實務界與學術界的交流，以及對外數據傳輸的硬性要求使兩地的數據規範產生了一定的域外效應。

李浩 84 表一 中國內地與澳門特區的數據傳輸原則比較 中國內地 澳門特區 民事基本法 《民法典》第一千零三十五條 《民法典》第七十九條（個人資料之保護）、第八十一條（個人資料真實權） 專門法 〈個人信息保護法〉第五條、第六條、第七條 〈個人資料保護法〉第二條、第五條、第六條 縱向⽽⾔，内地〈個⼈信息保護法〉的原則性規定是對其《民法典》的細化，⽽澳⾨〈個⼈資料保護法〉的⼀般性規定則使⽤了比其《民法典》更為約括的語⾔。橫向⽽⾔，內地之數據處理被納入民法內單獨的⼈格權篇章以予探討，隱私權與個⼈信息保護為交叉狀態︔在澳⾨則涉及隱私權的下位概念個⼈信息保護權（被定義為資訊⾃決權）。18 兩地的數據處理原則有不少相似之處，例如内地〈個⼈信息保護法〉第六條的⽬的相關原則與澳⾨《民法典》第七⼗九條第⼆款的⽬的限制原則、内地〈個⼈信息保護法〉第七條的公開透明原則與澳⾨〈個⼈資料保護法〉第⼆條的透明原則、内地〈個⼈信息保護法〉第八條的品質原則與澳⾨《民法典》第八⼗⼀條的個⼈資料真實原則等。然則以上並非完全相同，更多的是交叉重疊關係，如内地〈個⼈信息保護法〉第八條的品質原則與澳⾨《民法典》第八⼗⼀條的個⼈資料真實原則，前者強調的是個⼈信息的準確、完整，後者則關注個⼈資料的真實性。準確固然含有真實之義，⽽完整則對資訊處理提出了更⾼的要求。「真實但不完整」的資訊同樣可以導致對資料當事⼈的不利後果。此外，兩地數據處理原則也存在細微差異，如内地《民法典》第⼀千零三⼗五條的合法、正當、必要原則與澳⾨《民法典》第七⼗九條第⼀款的當事⼈查閱、更正與更新原則，第三款的第三⼈查閱許可原則。然此種區別並非絕對，如澳⾨《民法典》第七⼗九條第⼀款的當事⼈查閱、更正與更新的權利雖未列入内地數據處理原則，卻與内地《民法典》第⼀千零三⼗七條有異曲同⼯之妙。總體來看，内地⽴法中的數據處理（傳輸）原則範圍比澳⾨ 18 參見唐曉晴、吳奇琦，〈澳門地區民法上個人資訊保護權的體系定位〉，《國家檢察官學院學報》，第 5期（北京：2021.10），76。 法域 法律

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 85 ⼤。原則性規定不同之處，則源⾃兩地⽴法思路、規範結構與法律⽤語的差異。儘管有流於寬泛之嫌，上位法所載之原則為數據處理（傳輸）提供了基本的框架，對標準合同條款的制定與施⾏起到指導作⽤。兩地原則性規定在⽴法結構與實質性的差異，如缺乏恰當的應對措施，恐在⼀定程度上導致 SCCs規則銜接與運⽤的混亂。 相比於數據傳輸原則，其間所涉及的概念（法律術語）對標準合同的理解與合意形成有著更為重要的意義。同⼀主體、客體或⾏為，因為⽂化傳統以及⽤語習慣的不同，或為不同的法律術語所形容。數據傳輸的有關概念多於專⾨法中作特別說明，以達促進法律適⽤便利性與確定性之⽬的。現將兩地規範中對數據傳輸涉及概念的定義列於下表： 表二 中國內地與澳門特區的數據傳輸概念比較 中國內地用語 定義 澳門特區用語 定義 個人信息 《民法典》第一千零三十四條 ；〈個人信息保護法〉第四條 個人資料 〈個人資料保護法〉第四條第一款第（一）項 敏感個人信息 〈個人信息保護法〉第二十八條 敏感資料 無明確規定，但可從〈個人資料保護法〉第七條第一款推導而出 重要數據 〈數據出境安全評估辦法〉第十九條 暫未發現對應類似定義 關鍵資訊基礎設施 〈網絡安全法〉第三十一條國家 關鍵基礎設施 〈網絡安全法〉第二條第一款第（三）項 關鍵基礎設施營運者 〈網絡安全法〉第二條第一款（四）項 個人信息的處理 〈數據安全法〉第三條 個人資料的處理 〈個人資料保護法〉第四條第一款第（三）項 個人信息跨境提供 無明確規定，但被包含於〈數據安全法〉第三條與〈個人信息保護法〉第四條，同時可從〈個人信息保護法第三十八條推導而出 個人資料轉移 無明確規定，但被包含於〈個人資料保護法〉第四條第一款第（三）項，同時可從〈個人資料保護法〉第十九條第一款推導而出

李浩 86 個人信息主體 〈個人信息出境標準合同辦法〉第一條第（四）款 資料當事人 〈個人資料保護法〉第四條第一款第（二）項 個人信息處理者 〈個人信息保護法〉第七十三條第（一）款 負責處理個人資料的實體 〈個人資料保護法〉第四條第一款第（五）項 接受委託處理個人信息的受託人 〈個人信息保護法〉第五十九條 次合同人 〈個人資料保護法〉第四條第一條第（六）款 暫未發現對應類似定義 第三人 〈個人資料保護法〉第四條第一條第（七）款 境內接收方暫未被提及 資料的接收者 〈個人資料保護法〉第四條第一條第（八）款 境外接收方 〈個人信息出境標準合同辦法〉第一條第（二）款 監管機構 〈個人信息出境標準合同辦法〉第一條第（七）款 公共當局 〈個人資料保護法〉第四條第一條第（十一）款 由之可⾒，與數據傳輸（處理）有關的法律術語涉及多個領域以及數據⽣命全週期中的部分，常⾒的包括但不限於個⼈信息/個⼈資料、敏感個⼈信息/敏感資料、個⼈信息跨境提供/個⼈資料轉移、個⼈信息主體/資料當事⼈、監管機構/公共當局等。就中⽂語境⽽⾔，雙⽅所使⽤的術語存在整體對應與部分對應的關係。整體對應之術語中，又可分為差異較⼩者（⽤語部分相同，如個⼈信息/個⼈資料、敏感個⼈信息/敏感資料等）以及差異較⼤者（⽤語完全不同，如個⼈信息主體/資料當事⼈）。部分對應之術語，則如境外接收⽅/資料的接受者，其⼀⽅包含或被包含於另⼀⽅。同時尚有部分術語，如内地的「重要數據」、澳⾨的「第三⼈」等，並未在對⽅所使⽤的術語中找到對應者。⽽除法律有明確定義的外，仍有術語之概念依靠對有關條款的推導⽽闡述。總體來說，相當⼀部分術語彼此之間具有較強的互譯性，即便未經專⾨的法律解釋也不妨礙雙向理解。換⾔之，雙⽅法律術語之差別不⼤，未⾜以構成嚴重程度的溝通障礙。雖則如此，其間存在之種種間隙，包括涵義未完全對應的術語，未有對應概念之術語以及無明確定義之術語等，依然可導致理解的偏誤。更有甚者，内地使⽤的「個⼈信息處理者」術語，因與澳⾨特區及其它法域通⽤的

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 87 控制者理念存在較⼤差異，在導致制度對話問題與解釋成本上升之餘，還造成後續傳輸結構單⼀。19 ⽽在⼤灣區的數據傳輸願景中，傳輸⽅還包括⾹港以及灣區外國家，故⽽法律術語除中⽂外還需提供葡⽂、英⽂以及其他語⾔版本，理解的偏差將在翻譯中進⼀步被放⼤。法律術語之通暢理解，乃雙⽅意思表達、形成合意以及締結條約的前提。缺少此基礎，雙⽅之溝通則無異於「雞同鴨講」，將極⼤增加企業合規成本進⽽降低數據轉移效率。另者，除術語之內涵差異外，其於雙⽅法律⽂本中所處之位置亦有不同。内地多將本法涉及名詞之解釋以附錄形式列於⽂本之末，⽽澳⾨則往往以單獨⼀章置於本法較前部分對定義進⾏闡明。此兩種⽴法習慣之迥異或在相當程度上造致實務中的困惑與不便。 （㆓）、性質與㊜用前提 追踵歐盟 95指令及〈通⽤數據保護條例〉（General Data Protection Regulation, GDPR）有關規範，内地及澳⾨在其⽴法中構建了比較完善的跨境數據傳輸體系。為實現數據的有效跨法域轉移，雙⽅的規則體系皆嘗試提供多元化的法律⼯具（正當性基礎），包括充分性保護⽔準（adequate protection level）、有約束⼒的公司規則（binding corporate rules）以及豁免情形。可⾒標準合同條款並非唯⼀的數據跨境傳輸唯⼀的法律⼯具，⽽系諸多法律⼯具之⼀。合同的性質及適⽤前提均以其在數據傳輸體系中的位置為判斷要件，是以將內地及澳⾨的有關規定列於下表： 19 參見姚佳，〈論個人資訊處理者的民事責任〉，《清華法學》，第 3期（北京：2021.06），45；參見金晶，〈作為個人信息跨境傳輸監管工具的標準合同條款〉， 24。

李浩 88 表三 中國內地與澳門特區的數據傳輸合同性質及適用前提比較 中國內地 澳門特區 〈網絡安全法〉第三十七條 〈個人信息保護法〉第三十八條、第四十條 〈數據出境安全評估辦法〉第四條、第五條第（五）款、第六條第（三）款、第八條第（五）款、第九條、第十四條 〈個人信息出境標準合同辦法〉第四條、第八條 〈規範和促進數據跨境流動規定〉第一條、第四條、第五條、第七條、第八條、第十三條 〈個人資料保護法〉第十九條、第二十條 上述條款顯⽰，中國內地以數據的種類為基礎實⾏不同的傳輸制度，為 SCCs制度的適⽤設置了禁區。其〈網絡安全法〉第三⼗七條從數據屬性的角度（關鍵資訊基礎設施的運營者在境內運營中收集和產⽣的個⼈信息和重要數據），與〈個⼈信息保護法〉第四⼗條從數據處理者角度（關鍵資訊基礎設施運營者和處理個⼈信息達到國家網信部⾨規定數量的個⼈信息處理者），對較為重要與敏感的資訊適⽤「原則本地化+安全評估」的傳輸路徑。以上規定亦被認為是內地「數據硬本地化」的表現，從⽽招致部分國際評價機構的誤解與批評。20 在具體規則中，《數據出境安全評估辦法》第四條則進⼀步細化了需進⾏安全評估的個⼈信息處理者的處理數量⾨檻。〈個⼈信息出境標準合同辦法〉第四條所規定的 SCCs 適⽤範疇，明顯是對需進⾏安全評估情況的排除。值得注意的是，即使是在安全評估制度中，雙⽅擬簽訂包括數據傳輸相關合同在內的法律⽂件依然佔有重要位置。21 譬如《數據出境 20 以政策產生之後果進行劃分，有直接性的物理性數據存儲要求的為硬數據本地化，如印度、中國內地以及俄羅斯等；無直接要求而造成本地儲存效果的為軟數據本地化，如歐盟。參見李艷華，〈隱私盾案後歐美數據的跨境流動監管及中國對策—軟數據本地化機制的走向與標準合同條款路徑的革新〉，32；如 ITIF 曾把中國內地列為對數據限制最嚴的國家，其次為印尼與俄羅斯。See Nigel Cory and Luke Dascoli. “How Barriers to Cross-Border Data Flows Are Spreading Globally, What They Cost, and How to Address Them.” Information Technology & Innovation Foundation. Accessed September 3,2023. https://itif.org/publications/2021/07/19/how-barriers-cross-border-data-flows-are-spreading-globally-what-they-cost/. 21 安全評估中雙方簽訂的法律文件與 SCCs（即使是廣義的）仍屬兩個概念，其原因有三：一、法律文件除了相關合同外還有其他條約、協議等，其範圍明顯比 SCCs 大。二、法律文件僅為安全評估之一要件，不足以構成數據傳輸基礎，而 SCCs 則可以。三、法律文件中的合同並非由有關部門制定的標準合同，與 SCCs 尤存在差距。而安全評估中某些合規程度比較高、通過率比較可觀

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 89 安全評估辦法》第五條第五款將其列為數據出境風險⾃評估的重點事項，第六條第三款規定其為申報安全評估所需提交之材料，第八條第五款將其包括於安全評估重點事項，第九條對其所約定之責任義務提出了要求，第⼗四條以其變更為重新申報評估之條件等。2024年 3⽉發佈的《規範和促進數據跨境流動規定》第⼗三條強調了本規定的優先性，然其具體內容僅對⼀定類別數據、特定事由及數量管控⾨檻進⾏微調，並未根本改變原有規範的框架。由此可⾒，內地 SCCs 制度是次位性、補充性的，同時具有公共機構規制（由有關部⾨發佈、審核）以及私法合同的特徵。且其與其他傳輸制度如安全評估存在相似的組成部分，即以簽訂合同作為要件，是以有⼀定的重疊性與交錯性。22 ⽽無論從規則締造時序還是內容來看，內地實⾏的都是「關鍵（重要）數據—⼀般數據」的傳輸體系。23 與之相對應，SCCs制度的適⽤對象是不涉及公共利益的普通數據，也即無需經過安全評估的數據。儘管其重要性與敏感度較低，然確系⽬前跨境傳輸的基本⽅向與主要內容。 ⽽澳⾨雖也存在數據分類管理機制，然在數據跨法域流動⽅⾯表現不明顯。2019年頒⾏的〈網絡安全法〉雖致⼒於構建本⼟互聯網防護體系，確未對關鍵基礎設施營運者收集與產⽣個⼈資料轉移到特區之外的地⽅做出特別規定。更兼澳⾨並無與內地「重要數據」相對應的概念，其對不同種類數據之傳輸更傾向⼀視同仁。〈個⼈資料保護法〉第⼗九條規定了資料轉移的⼀般原則，也即公共當局所認證的「適當的保護程度」。該法第⼆⼗條則列舉了⼀系列排除適⽤的情況，其中保障基本權利及其⾏使的合同條款屬於第⼆款，是為「例外中的例外」。澳⾨的合同條款制度與内地的 SCCs ⼀樣，在理論以及⽴法地位⽅⾯是補位性質的，即未能滿⾜第⼗九條以及第⼆⼗條第⼀款之原則下的應⽤，同時為第⼆⼗條第三款國際法⽂書與區際協定的規制所排斥。與此同時，澳⾨〈個⼈資料保護法〉第⼆⼗條第⼆款之規定在賦予公共當局較⼤的⾃由裁量權之餘，亦導致合同制度有著較強的綜合性。澳⾨合同制度的適⽤前提並無數據類型限制，卻要根據不同的場景進⾏判斷。比較⽽⾔， 的合同範本是否會逐漸演變成「行業（民間）SCCs」則有待進一步觀察。 22 在某種程度上來講，安全評估可以視為合同制度在關鍵數據上的投射，SCCs則為合同制度在一般數據上的應用。而根據數據重要性的差異，兩種傳輸工具被搭配與其他不同的構成要件與程序。 23 從順序上看，內地《網絡安全法》發佈於 2016年 6月，《數據安全法》發佈於 2021年 6月，《個人信息保護法》發佈於 2021年 8月，《數據出境安全評估辦法》發佈於 2022年 7月，《個人信息出境標準合同辦法》發佈於 2023年 2月，基本遵循「先重點數據、後一般數據」的原則進行上位法創建與下位規則跟進。

李浩 90 兩地的 SCCs 制度的性質比較接近，⽽其適⽤前提差異較⼤。從數據種類⽽⾔，澳⾨合同條款的應⽤範圍無疑廣於内地︔然從使⽤場景⽽⾔，內地 SCCs 的覆蓋效能則遠勝澳⾨。 （㆔）、範本內容 為使合同制度「落入塵埃」具有現實意義，兩地有關部⾨皆以不同⽅式發佈了指導性範本。內地網信辦於 2023年初採⽤「辦法+範本」的結構推出了第⼀版標準合同條款，為數據傳輸⽅提供了較為詳盡的層級應⽤規範。同年 3⽉份，澳⾨個資辦在其官網推出新版通知及許可申請表格。其中許可申請表格不僅適⽤於個⼈資料的跨境轉移，還可應⽤於敏感資料處理、信⽤與償付能⼒處理以及資料互聯的情形。24 現將兩地發佈官⽅表格的重要內容條項呈列如下以作比較： 表四 中國內地與澳門特區的數據傳輸合同範本內容比較 中國內地 澳門特區 文件名稱 〈個人信息出境標準合同辦法〉〈個人信息出境標準合同〉 〈個人信息出境標準合同〉附錄一〈個人信息出境說明〉 〈個人信息出境標準合同〉附錄二〈雙方約定的其他條款（如需要）〉 〈個人資料處理許可申請書【首次申請】（不適用於個人資料互聯的申請）〉 〈個人資料處理許可申請書附表 VI（關於轉移個人資料至澳門特區以外其他地方）〉 〈個人資料處理許可申請書【申請續期/修改許可】（不適用於個人資料互聯的申請）〉 〈個人資料保護辦公室個人資料處理許可申請書【申請廢止許可】（不適用於個人資料互聯的申請）〉 數據傳輸雙方信息 〈個人信息出境標準合同〉開頭部分 〈個人資料處理許可申請書【首次申請】（不適用於個人資料互聯的申請）〉第一、二、六條 24 參 見 個 人 資 料 保 護 辦 公 室 ， 「 負 責 處 理 個 人 資 料 的 實 體 須 知 」 ， https://www.gpdp.gov.mo/zh_tw/law_routine_3.html，瀏覽日期為 2023年 8月 8日。 法域 項目

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 91 傳輸基本情況之說明（如數據種類、方式等） 〈個人信息出境標準合同〉附錄一《個人信息出境說明》 〈個人資料處理許可申請書附表 VI（關於轉移個人資料至澳門特區以外其他地方）〉第一、二、三、四、五、七條 定義 〈個人信息出境標準合同〉第一條 無 個人信息處理者的義務 〈個人信息出境標準合同〉第二條 〈個人資料處理許可申請書附表 VI（關於轉移個人資料至澳門特區以外其他地方）〉第六條、第八條第二、三款 境外接收方的義務 〈個人信息出境標準合同〉第三條 〈個人資料處理許可申請書附表 VI（關於轉移個人資料至澳門特區以外其他地方）〉第六條、第八條第二、三款 接收方法律環境對合同履行之影響 〈個人信息出境標準合同〉第四條 無 個人信息主體的權利 〈個人信息出境標準合同〉第五條 〈個人資料處理許可申請書附表 VI（關於轉移個人資料至澳門特區以外其他地方）〉第八條 救濟、合同解除（終止）與違約責任 〈個人信息出境標準合同〉第六、七、八條 〈個人資料處理許可申請書【申請續期/修改許可】（不適用於個人資料互聯的申請）〉 〈個人資料保護辦公室個人資料處理許可申請書【申請廢止許可】（不適用於個人資料互聯的申請）〉 補充 〈個人信息出境標準合同〉附錄二〈雙方約定的其他條款（如需要）〉 〈個人資料處理許可申請書附表 VI（關於轉移個人資料至澳門特區以外其他地方）〉第八條第五款 由上可⾒，兩地推出之有關合同條款的內容存在著較⼤差異。⾸先是範本性質的不同，中國內地以部⾨規章的形式公佈的〈個⼈信息出境標準合同辦法〉及其附件，系以提供官⽅格式化的合同範本為中⼼的，屬於狹義的 SCCs 制度，具有很強的實務指導意義。⽽澳⾨公共當局所發佈的〈個⼈資料處理許可申請書【⾸次申請】（不適⽤於個⼈資料互聯的申請）〉系列⽂件，並非圍繞合同條款⽽展開，實系為跨境數據轉移乃至數據處理服務。以合同⽅式提供⾜夠保障僅為〈個⼈資料處理許可申請書附表 VI（關於轉移個⼈資料至澳⾨特區以外其他地⽅）〉第八條所列諸多情況之⼀種。其次是有關款項的具體規定與位置的差別。除數據傳輸雙⽅信息、傳輸基本情況之說明等項⼆者比較接近外，其餘的⼤體皆為内地規定比澳⽅更為廣泛

李浩 92 與詳細。如定義與接收⽅法律環境兩項，即為內地規範有⽽澳⾨無，當然這很⼤程度上也是因為澳⾨在其他法律中對其做出了闡述。再如個⼈信息處理者的義務、境外接收⽅的義務、個⼈信息主體的權利等項，内地均以相當篇幅做出了詳細規制，⽽澳⽅規範則流於空泛。 具⾔之，〈個⼈信息出境標準合同〉列舉了包括告知、採取技術與管理⼿段（如匿名化）以及個⼈信息保護影響評估在內與個⼈信息處理者義務相關的⼗⼀項規定定，包括在⾃動化處理、轉委託第三⽅處理數據情景在內與境外接收⽅義務相關的⼗三項規定，包括知情權、決定權在內與個⼈信息主體權利相關的五項規定等。25 與之相對應，〈個⼈資料處理許可申請書【⾸次申請】（不適⽤於個⼈資料互聯的申請）〉則採取了「空格待填」，由數據傳輸⽅申報的⽅式。最後是⼆者的保障程度與確定性。內地之系列規範⽂件已為數據傳輸⽅提供詳盡明確之適⽤標準與範本，個⼈信息處理者僅需機械照搬其條款，即可獲得數據跨境傳輸的合法性。由此論之，認為陸⽅規範為「傻瓜條款」亦未嘗不可。然則澳⾨所發佈之規範，在至為關鍵之傳輸基礎（即〈個⼈資料處理許可申請書附表 VI（關於轉移個⼈資料至澳⾨特區以外其他地⽅）〉第八條）⽅⾯語焉不詳，甚至將證明義務交由申請⽅。⽽合同條款僅為附表 VI 第八條之⼀項考察因素，且未有具體條款可供參考。傳輸⽅之獲許與否，仍有賴於個資局之⾃由裁決，確存較⼤之不穩定性與不可預⾒性。然個資局公布相關的意⾒書以及許可結果，其中包括對合同條款之評價，或為申請⽅從中窺之⼀⼆，以作澳⾨ SCCs的「判例法」參考。26 如第 02/A/2022/GPDP號許可判定雙⽅所簽外判服務協議可提供合同條款保障，27 第 01/A/2009/GPDP 號許可認為當事⼈簽訂的服務合約系屬「適當的安全措施」，28 第 0016/P/2018/GPDP號意⾒書對合同 25 〈個人信息出境標準合同〉第三條第九款規定，轉委託第三人處理個人信息需滿足三個條件：一、事先取得個人信息處理者的同意。二、要求第三方不得超出約定的目的、方式處理個人信息。三、對第三方處理活動進行監督。 26 與之相對應，中國内地、香港特區以及歐盟的做法更似 SCCs 的「成文法」。然唯以成文法之清晰、詳盡規定，不足以稱狹義的 SCCs。雖「判例法」之安慰式構詞似可為澳門辯護一二，其實際操作則成本極高，對於數據傳輸方極不友好；且其所謂許可示範，亦未必為澳方個資局本意。 27 參見澳門特別行政區個人資料保護辦公室，「廣發銀行股份有限公司澳門分行基於委託次合同人（廣發銀行股份有限公司）提供名單庫系統、資料儲存與資訊技術及支援服務，而將澳門公權力機關協查函或查封函內所涉及的自然人之個人資料轉移到澳門特別行政區以外的地方」，https://www.gpdp.gov.mo/file/Permission/02-A-2022-GPDP-TC.pdf，瀏覽日期為 2023年 9月 2日。 28 參見澳門特別行政區個人資料保護辦公室，「關於 A銀行申請與香港 B銀行以『互聯』方式處理其『客戶資料』」，https://www.gpdp.gov.mo/file/Permission/01-A-2009-GPDP-TC.pdf，瀏覽日期為

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 93 之特定條款作出評估並提供修改意⾒等。29 毋庸置疑，兩版本的巨⼤差異必將導致數據傳輸的不效率，從⽽損害⼤灣區的數字融合藍圖。其根本之原因，乃在兩地⽴法邏輯與背景的不同。內地致⼒於以提供準確性極⾼的樣本合同來為數據傳輸提供基礎，⼀定程度上吸收了最新的〈通⽤數據保護條例〉的成果，然未能像歐盟、東盟及⾹港地區⼀樣提供多個傳輸⽅向的範本︔⽽澳⾨則因⽴法較早，並未過多關注合同條款之作⽤，⽽僅將其作為⾏政許可的考量要件之⼀。兩地合同條款作為 SCCs 制度的關鍵部分，其有效性關係整個數據傳輸以及保護體系的品質，是以尤待進⼀步彌合。 （㆕）、程序及㈼管機構 兩地合同條款制度之⾏使，均按循⼀定的程序並受有關部⾨監管。此類事項於⽴法之初已為規則締造者所矚⽬，在專⾨法中對其細節進⾏規定，並在後續規範中不斷予以補充。今為了⽅便理解，將內地及澳⾨的有關條⽂列出於下表： 表五 中國內地與澳門特區的數據傳輸合同程序及監管機構比較 中國內地 澳門特區 程序 〈個人信息保護法〉第三十八條 〈個人信息出境標準合同辦法〉第三、五、六、七、八條 〈個人信息出境標準合同辦法備案指南（第一版）〉 〈個人資料保護法〉第十九、二十條 〈個人資料處理許可申請書【首次申請】（不適用於個人資料互聯的申請）〉 〈個人資料處理許可申請書附表 VI（關於轉移個人資料至澳門特區以外其他地方）〉 監管機構 〈網絡安全法〉第八、十四、五十二條 〈數據安全法〉第六條 《民法典》第七十九條第三款 〈個人資料保護法〉第四條第十一款 〈第 83/2007 號行政長官批示〉 2023年 9月 4日。 29 參見澳門特別行政區個人資料保護辦公室，「關於澳門電訊有限公司將數據中心業務擴展至香港而 涉 及 的 澳 門 機 構 轉 移 個 人 資 料 到 香 港 事 宜 及 相 關 的 通 知 與 許 可 程 序 」 ， https://www.gpdp.gov.mo/file/Opinion/0016-P-2018-GPDP-TC.PDF，瀏覽日期為 2023年 9月 4日。 比較事項 法域

李浩 94 〈個人信息保護法〉第六章 〈黨和國家機構改革方案〉第十四條 〈「數字灣區」建設三年行動方案的通知〉第四條 〈網絡安全法〉第六、七、八、九條 〈第 42/2023 號行政法規〉 在具體程序上，內地 SCCs制度採⽤的是「合同+備案」形式。其〈個⼈信息保護法〉第三⼗八條將訂⽴合同與安全評估、專業機構認證並列為向境外傳輸數據應具備其⼀即可之條件。嗣後發佈的〈個⼈信息出境標準合同辦法〉則覆蓋了 SCCs適⽤的「全週期」。在此之中，第三條對適⽤程序的基本構成部分做了說明，第五條詳細闡述了個⼈信息保護影響評估的內容，第六條包括了對合同條款的遵守、調整以及⽣效事宜，第七條列舉了備案應準備的材料，第八條則涉及重新訂⽴合同與再次備案的情況。〈個⼈信息出境標準合同辦法備案指南（第⼀版）〉對備案⽅式、流程、材料等做出了具體要求，同時還提供了包括⾃評報告在內的⽂件範本。與之相對的，澳⾨的合同條款以⾏政許可的⽅式進⾏。〈個⼈資料保護法〉第⼆⼗條對應當通知與申請許可的前提做出規定。個資局的官網所發佈的《負責處理個⼈資料的實體須知》提供了更為具體的指引，以合同⽅式保障基礎權利的做法需要向公共當局申請許可。30 申請⼈需填寫〈個⼈資料處理許可申請書【⾸次申請】（不適⽤於個⼈資料互聯的申請）〉及其附表 VI並提交，⽽後靜候個資局審批。比較⽽⾔，内地 SCCs 的適⽤程序更具有操作性，表達出了對傳輸⽅締約意志的尊重，然其事後備案制度過於複雜，在⼀定程度上增加了數據傳輸的成本。 在監管機構上，內地可謂「⼀龍牽頭，九龍治⽔」，也即由網信部⾨統籌協調，其他部⾨分區負責。〈網絡安全法〉第八條規定國家網信部⾨全盤負責網絡安全及相關監督管理⼯作，第⼗四條明確可處理有關舉報的部⾨為「網信、電信、公安等」，⽽第五⼗⼆條則賦予了關鍵資訊基礎設施有關部⾨在安全預警領域的管理權。2018年發佈的〈國務院關於機構設置的通知〉規定，國家互聯網資訊辦公室與中央網路安全和信息化委員會辦公室，⼀個機構兩塊牌⼦，列入中共中央直屬機構序列。其後的〈數據安全法〉第六條確定了各地區、各部⾨（包括⼯業、⾦融、教育等⾏業） 30 參見個人資料保護辦公室，「負責處理個人資料的實體須知」，https://www.gpdp.gov.mo/ zh_tw/law_routine_3.html，瀏覽日期為 2023年 8月 8日。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 95 對其⼯作過程中收集與產⽣數據的安全負責的原則。〈個⼈信息保護法〉的第六章對「履⾏個⼈信息保護職責的部⾨」做出了闡釋，並詳細規定了其職責與可採取措施。最新〈黨和國家機構改⾰⽅案〉（2023年 3⽉）第⼗四條所決定組建的國家數據局會承接原網信辦與發改委的部分職能，為數據治理提供更多的可能性。31 在區域治理層⾯，廣東省政府印發的〈「數字灣區」建設三年⾏動⽅案的通知〉（2023年 11⽉）第四條第⼀款提出建⽴專責⼩組負責全盤統籌，並由省政務服務數據管理局承擔⽇常⼯作。32 反觀澳⾨，其《民法典》第七⼗九條第三款與〈個⼈資料保護法〉第四條第⼗⼀款皆提及「負責監察個⼈信息資料之收集、貯存及使⽤之公共當局」，然在〈第 83/2007號⾏政長官批⽰〉頒⾏之前，該機構並無名稱與實體。 2007年設⽴的個⼈資料保護辦公室（Gabinete para a Protecção de Dados Pessoais, GPDP）實為在特⾸監督下獨⽴運⾏之臨時專案組，負責監察、協調對第 8/2005 號法律之執⾏。33 ⽽其〈網絡安全法〉第六、七、八、九條則規定了特區網路安全體系的組成實體，即網路安全委員會、網路安全事故預警及應急中⼼與網路安全監管實體，及其相應職責。依據〈第 42/2023號⾏政法規〉（2024年 2⽉⽣效）建⽴的個⼈資料保護局（Direcção dos Serviços da Protecção de Dados Pessoais, DSPDP）繼受了原個資辦的⼤部分職能，為澳⾨政府屬下的公共部⾨。通過比較可知，内地的監管機構比較分散，主要監管機構兼有政黨部⾨性質，同時為將來制度改⾰預留了可觀的空間，並且在不斷探索、調整之中︔⽽澳⾨個資辦公室為臨時機構，雖職能集中，運⾏有效，但直到 2024年才改制「轉正」，對體制變⾰持保守態度。 三、歷史性機遇：橫琴粵澳深度合作區作為試點 標準合同條款確為兩地乃至⼤灣區內特定類型資訊傳輸的有效路徑，然上述規 31 參見中華人民共和國中央人民政府，「中共中央 國務院印發『黨和國家機構改革方案』」，https://www.gov.cn/gongbao/content/2023/content_5748649.htm，瀏覽日期為 2023年 8月 9日。 32 參見廣東省人民政府，「廣東省人民政府辦公廳關於印發「數字灣區」建設三年行動方案的通知」， http://www.gd.gov.cn/zwgk/gongbao/2023/31/content/post_4287722.html?version=slh，瀏覽日期為2023年 11月 7日。 33 參 見 澳 門 特 別 行 政 區 ， 「 第 83/2007 號 行 政 長 官 批 示 」 ， https://www.gpdp.gov.mo/file/Laws%20and%20Regulations/第 83_2007號行政長官批示_TC.pdf，瀏覽日期為 2023年 8月 9日。

李浩 96 範差異⾜已成為其「攔路虎」。如何打破兩地的數字制度壁壘，向來是粵澳區域治理者與法學家的重要命題。橫琴深合區，作為國家級戰略平台與「特區外的特區」，為該問題的解決提供了⼀個獨特的角度。實現內地與澳⾨特區 SCCs 的規則銜接，深度合作區至少可以從兩個⽅向出發，即「⾃上⽽下」的頂層設計與「⾃下⽽上」的底部創新。 （㆒）、頂層設計㆘的「放權賦能」㊝勢 2019年，黨中央、國務院在〈粵港澳⼤灣區發展規劃綱要〉（以下簡稱〈規劃綱要〉）第⼗章第三節即提出要「建設粵港澳深度合作⽰範區」，其中包括粵澳信息港的計畫。34 三年後，〈橫琴粵澳深度合作區建設總體⽅案〉（以下簡稱〈總體⽅案〉）出台，進⼀步提出要創設共商共建共管共⽤的新體系。35 ⼆三年初廣東省⼈⼤常委會通過〈橫琴粵澳深度合作區發展促進條例〉（以下簡稱〈促進條例〉），對治理體系、⼀體化發展與法治保障做出了更為詳細的規定。36 從中央到地⽅的比較完整的規範體系堅持頂層設計的基本原則，為橫琴深合區的建設「放權賦能」。其中有三處可圈可點，能夠以圖⼀所⽰路徑為粵澳合同條款規則銜接與創新提供制度驅動。37 34 參見中華人民共和國中央人民政府，「中共中央 國務院印發『粵港澳大灣區發展規劃綱要』」，https://www.gov.cn/xinwen/2019-02/18/content_5366593.htm#1，瀏覽日期為 2023年 7月 23日。 35 參見國務院港澳事務辦公室，「中共中央 國務院印發『橫琴粵澳深度合作區建設總體方案』」，https://www.hmo.gov.cn/zcfg_new/gafg/gfxwj/202211/t20221125_23891.html，瀏覽日期為 2023年 7月 23日。 36 參見廣東省人民政府，〈橫琴粵澳深度合作區發展促進條例〉， http://www.gd.gov.cn /gdywdt/zwzt/ygadwq/zxzc/content/post_4091516.html，瀏覽日期為 2023年 8月 9日。 37 圖中所描述之對應關係僅為主要貢獻，並不排除有關制度或經驗在其他領域的作用。如「分線管理」一項，其所創造的原則與法律術語解釋空間屬於「平臺共建」，而適用前提的調整（對上位法變通）又涉及「規則共商」。前者為該制度之突出建構，故在圖一中以予顯示。由於各項深合區制度或經驗對於 SCCs 的參與呈複合交叉的樣態，採用簡化指示更為適宜。上述展示慣例將沿用於後續圖二及圖三中，為方便理解特此說明。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 97 圖一 「自上而下」的規則銜接與創新路徑 其⼀，深合區實⾏「分線管理」，構建數據傳輸先⾏試驗區，落實 SCCs制度的「平台共建」。橫琴合作區為〈總體⽅案〉第三條所規定與〈促進條例〉第⼆條所再度確認的「⼀線」和「⼆線」之間的領域。38 儘管⽬前政策將分線管理的焦點放於貨物、物品、運輸⼯具與⼈員，但同時對跨境數據傳輸表⽰⽀持，如〈總體⽅案〉第⼗九條、〈促進條例〉第五⼗五條。與對貨物（「⼀線」放開，「⼆線」管住）與⼈員（合作查驗、⼀次放⾏通關）明確的分區設計與改⾰⽅向不同，在資訊跨境⽅⾯利⽤兩線差異構建制度縱深較為困難，且有關合作路徑的表達極為模糊。這是由於數據性質本身的複雜性：既在標的與主給付義務⽅⾯可比擬被運輸的貨物，又是個體⽣命檔案化的主要載體，然不少學者仍對「分線管理」對數據傳輸的開拓 38 其中，橫琴與澳門特區之間設為「一線」；橫琴與內地之間設為「二線」。需注意深合區並不包括澳門大學橫琴校區和橫琴口岸澳門管轄區（由全國人大常委授權特區政府管理）。 頂層設計 橫琴深度合作區 分線管理 一線：不同於澳門 二線：不同於內地 數據=貨物？人員？ SCCs試驗自主權 治理體系改革 部門規章與特區立法權 橫琴辦 VS 執委會SCCs 規則締造與執行 多元化糾紛解決模式 SCCs 合同糾紛：民商事訴訟、仲裁、調解 SCCs 備案：行政復議、訴訟 數字財產權與數字人格權保障 平臺共建 規則共商 機構共管 成果共享 產業發展 區域融合

李浩 98 意義持樂觀態度。39 事實上，「分線管理」所締造的特殊區域不僅可以突破內地標準合同的效⼒覆蓋，還能承接澳⾨合同許可的溢出，為 SCCs「實驗室」的建設提供了空間。「數字深合區」將承載著「求同存異」的歷史任務，打造兩地標準合同條款的共識平台、話語平台與交互平台。合作區規則締造者可解構兩線差異，充分利⽤區域改⾰⾃主權，在瞭解內地與澳⾨法律、社會⽂化習慣的前提下，結合科技發展潮流與本地利益對數據及其流通進⾏再定義，協商制定 SCCs 適⽤原則，發佈通⽤或至少趨同的法律術語版本，構建合同願景與話語基礎︔發揮平台結構化溝通與創新作⽤，處理好與兩地上位法以及其他法域規範的關係，尋求制度融合點與出發點，營造規範、經濟與社會共同體。 其⼆，治理體系的改⾰，通過創設數據監管主體落實 SCCs制度的「規則共商」與「機構共管」。〈總體⽅案〉第五章及〈促進條例〉第⼆章提出要建設的粵澳新治理體制，是橫琴合作區的「深度」與亮點所在。治理體制的改⾰有望打破內地基層政府的固有權⼒結構，為區域開放甚至融合提供經驗。⽬前深合區的治理體系包括兩⼤主體：廣東省派出機構（橫琴辦）、橫琴深度合作區管理委員與執⾏委員會。省派出機構負責黨的建設、國家安全、刑事司法以及社會治安等保留事項。管委會為深合區最⾼管理與議事機構，執委會則是管委會的⽇常⼯作機構，依法履⾏國際推介、招商引資、產業導入等公共服務職能。執委會下設⾏政事務局、法律事務局、財政局等九個機構，分別負責深合區不同領域的⼯作。 兩地合同條款制度的銜接與創新，既涉及新規範的締造，也包括對規則的執⾏與監督。無論是內地的標準合同範本還是澳⾨的許可樣本，本質都是專⾨法的下位規範，即⾏政機關的部⾨規章及施⾏細則，同時其變通適⽤又與特區⽴法權有關。是以〈促進條例〉第五⼗七、五⼗八條規定的法規建議權在此或非無「⽤武之地」。具體規則的構建則有賴兩地職權機構，即內地國家網信辦與澳⾨個資局協商進⾏。⽽中央機構與特區部⾨就某⼀事項合作並達成協議並不乏先例，如國家網信辦與⾹ 39 參見金晶，〈作為個人信息跨境傳輸監管工具的標準合同條款〉， 25；參見藍江，〈生命檔案化、演算法治理和流眾—數字時代的生命政治〉，《探索與爭鳴》，第 9期（上海：2020.09），107；如中國（深圳）綜合開發研究院常務副院長郭萬達指出，「分線管理」政策利於數據跨境安全的把控。參見張雅婷，美婷，「橫琴探路數據跨境流動，打造數據治理創新樣本」，https://m.21jingji.com/article/20210909/0052dd35ee00dd29ba0987a32a353f51.html，瀏覽日期為 2023年 8月 10日。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 99 港創科及⼯業局 2023年 6⽉簽署的〈促進粵港澳⼤灣區數據跨境流動的合作備忘錄〉。40 事實上，國家網信辦也嘗試通過制定部⾨規章賦予⾃貿區在數據出境事宜上更多的⾃主權，如〈規範和促進數據跨境流動規定〉第六條之規定。⼀⽅⾯，內地網信辦兼有政黨部⾨性質，不僅負責 SCCs 的備案⼯作，還涉及出境數據安全評估等業務，其職能當然落入省派出機構的範疇。橫琴數據監管機構應屬省派出機構、執委會，抑或兩者兼領之（職權劃分），⽬前尚不明確。另⼀⽅⾯，鑒於內地對適⽤ SCCs 的限制，以合同為基礎的數據傳輸顯然不屬於省派出機構管轄事項。執委會下設九個機構職能並未明確包括網路安全以及個⼈信息保護，也即缺乏與內地網信辦或者澳⾨個資局相對應的部⾨。唯有負責統籌改⾰創新以及制定政策法規的法律事務局與優化科技資源配置的經濟發展局與其有部分交叉。41 ⽽是否可⽤各職能局中的「兜底條款」涵蓋數據傳輸，尤待進⼀步討論。42 其三，涉外糾紛解決模式探索，致⼒於保障數字權益，落實 SCCs 制度的「成果共享」。〈促進條例〉第六⼗⼀條提出，「要推動在合作區建⽴完善國際商事審判、仲裁、調解等多元化商事糾紛解決機制」。2023年 2⽉，廣東省⾼級⼈民法院指定橫琴法院集中管轄應由珠海市其他基層法院管轄的標的 4000 萬元以下的⼀審涉外民商事案件。同年 4⽉份，橫琴國際仲裁中⼼、橫琴國際商事調解中⼼入駐琴澳國際法務⼤廈。43 早在 2017年，珠海市仲裁委員會即發佈了具有⾸創意義的〈橫琴⾃由貿易試驗區臨時仲裁規則〉，使內地企業參與臨時仲裁（ad hoc arbitration）向前邁出了⼀⼤步。44 ⼀⽅⾯，數據傳輸的合同條款制度多數情況下具有民商事規 40 參 見 政 府 新 聞 網 ， 「 大 灣 區 數 據 流 動 合 作 備 忘 錄 簽 訂 」 ，https://sc.news.gov.hk/TuniS/www.news.gov.hk/chi/2023/06/20230630/20230630_120748_916.html ，瀏覽日期為 2023年 8月 10日。 41 參見橫琴粵澳深度合作區，「職能介紹法律事務局」， http://www.hengqin.gov.cn/macao _zh_hant/hzqgl/zzjg/znjs/content/post_2990404.html，瀏覽日期為 2023年 9月 3日；參見橫琴粵澳深 度 合 作 區 ， 「 職 能 介 紹 經 濟 發 展 局 」 ， http://www.hengqin.gov.cn/macao _zh_hant/hzqgl/zzjg/znjs/content/post_2990403.html，瀏覽日期為 2023年 9月 3日。 42 此處系指各局職能的最後一條「負責完成執委會交辦的其他工作」。該條為執委會因應形勢，對未有明確規定之職權進行「個案交辦」，進一步完善深合區治理機構提供了空間，是以有此名稱。參見李可、唐曉晴，<橫琴粵澳深度合作區：理念創新與制度構建>， 17。 43 參見肖皓方，李美琪，「從『一院』拓展到『兩中心』，珠海國際仲裁院落子『集聚區』」， http://www.hengqin.gov.cn/macao_zh_hans/ztjh/wqzc/dtzx/content/post_3532564.html，瀏覽日期為2023年 8月 11日。 44 See Liu Lu, Qi Qi, “Determining the Validity of Ad Hoc Arbitration Agreements in China: Past, Present, and Future.” Frontiers of Law in China 14.1 (Beijing: 2019.03):139.

李浩 100 則的性質，故仲裁、調解與訴訟同為〈個⼈信息出境標準合同辦法〉第九條第四款包含的糾紛解決途徑。另⼀⽅⾯，其又需向⾏政主體的備案或申請許可，是以又涉及⾏政復議與⾏政訴訟。儘管〈促進條例〉第六⼗條對⾏政復議與⾏政訴訟做出了規定，然鑒於治理主體職能劃分的模糊，其實踐仍⾯臨較⼤的不確定性。雖然⽬前存在種種缺陷，但多元化趨勢為 SCCs 糾紛解決機制後續的變⾰與發展提供了⽀撐點。SCCs作為公私複合型合同，本身即涉及商業秘密、知識產權、個⼈隱私等法益︔⽽以其為⼯具傳輸數據所帶來的社會福利，如經濟發展與科研進步，亦會導向數字紅利分配的公平困境。上述問題會在試驗階段被指數型放⼤，在穩定運⾏之後也將持續存在。深度合作區對「訴、裁、調」結合的糾紛解決體系的探究，被寄望於提供司法與⾏政救濟對抗「數字權⼒」，規避 SCCs 創新風險，有效保障作為第四代⼈權的「數字⼈權」，促進 SCCs改⾰成果惠及兩地。45 （㆓）、先行創新與可借鑒經驗 〈促進條例〉第五條提出，合作區可在改⾰開放關鍵領域與重要環節⼤膽嘗試、先試先⾏、⾃主探索。同時，為降低創新風險，剪除後顧之憂，容錯免責機制也應隨之建⽴。46 該制度設想為深合區的改⾰經驗供應提供了「安全閥」。然在〈促進條例〉頒⾏前，琴澳之間的互動合作與體制創新已呈⽣機勃勃之勢。具體⽽⾔，有兩個⽅⾯可以圖⼆之路徑為 SCCs制度所鏡借。 45 參見蔡立東，〈為什麼「數字人權」是第四代人權〉，《數字法治》，第 3期（北京：2023.06），3-6。 46 凡符合深合區建設需求的，決策过程符合法定程序，未謀取私人利益或惡意串謀危害公共利益的，相關單位與個人不需承擔責任。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 101 圖二 「自下而上」的規則銜接與創新路徑 其⼀，既有的跨境數據傳輸實踐提供「案例法」參考，補充 SCCs制度的「規則共商」。技術應⽤先於法律規制是資訊時代的常⾒現實，⽽互聯網無政府主義的破產呼籲更為有序的數據活動。47 該種情況同樣存在於中國內地與澳⾨的跨境數據傳輸中。在兩地出台專⾨規範之前，資訊的跨邊境、超法域流動已經作為事實存在。隨著未來橫琴與澳⾨產業合作與社會交流的深入，雙⽅的資訊傳送需求亦將呈量級增長。在動態更新、持續完善的數據傳輸框架內，琴澳兩地的數據治理主體等進⾏了多領域、深層次、求合作、謀發展的資訊傳輸實踐，如澳⾨⼤學、澳⾨科技⼤學等⾼校與橫琴科研機構的數據交流，新冠疫情期間兩地公共衛⽣部⾨的資訊互通與通過平台驗證模式實現粵澳之間個⼈數據的可攜帶權等。48 以上這些數據傳輸⾏為， 47 此思想潮興於上世紀 90年代後期美國對於科技與法律的討論。業界的聲音頗具左派色彩，對構建無政府的網路烏托邦持積極態度，因而降低政府干預在相當一段時間內成為「政治正確」。參見田旭，〈歐盟個人數據保護法的全球影響成因與啟示〉，《江西財經大學學報》，第 4期（南昌：2020.08），2020（04），135。 48 參見李恩漢，吳偉傑，「粵澳跨境數據驗證平臺上線試運行，探索個人數據跨境流動新模式」， https://m.thepaper.cn/baijiahao_17470477，瀏覽日期為 2023年 9月 2日。 穩定有效 雙向互動 底部創新 橫琴深度合作區 規則共商 機構共管 數據傳輸實踐 行業慣例 習慣法 SCCs 的不成文規範 行政機構合作 法律適用 專業判斷 溝通管道 行為準則 SCCs 的實操模範

李浩 102 或由於相關法律的急劇變遷⽽出現合法性瑕疵（即前傳輸⾏為不符合後法），然其在保持了兩地數據傳輸的穩定進⾏之餘，還為將來之規範改進供給了某種⾏業慣例與習慣法借鑒。尤其是在澳⾨沒有公佈普適性合同範本的情況下，獲得其個⼈資料轉移許可的合同條款，如第 06/A/2022/GPDP號許可，第 02/A/2022/GPDP號許可以及第 12/A/2021/GPDP 號許可等，可作為將來的個⼈資料轉移至內地或⼤灣區的規則⽰範。49 不成⽂規則中的合理內核，經由特定程序如學術研究、⽴法意⾒徵詢進⾏提煉後，避免了機械的「拿來主義」，有機構成兩地⽴法者以及政策制定者的話語權，推動 SCCs具體內容的持續完善。 其⼆，兩地的⾏政機構合作奠定治理基礎，補充 SCCs制度的「機構共管」。標準合同的規則創建與執⾏監督同樣依賴⾏政機關的合作，尤其是橫琴將來成⽴的數據監管部⾨、數據交易試點與澳⾨個資局的溝通。兩地⾏政部⾨就促進⽣產要素的跨境流動⽅⾯具有相當經驗。50 在⼈員的通關模式上，繼港珠澳⼤橋之後，橫琴⼜岸於 2020年 8⽉成功實現「⼀地兩檢」，解決了兩地在法制、管理與技術等存在的差異所帶來的問題。51 「⼀地兩檢」將兩套程序合⼆為⼀，不僅涉及兩地的數據交互，還需要⼯作⼈員的有效配合。在⾞輛的通關模式上，2022年末啟⽤的「澳⾞北上」申請系統涉及廣東省商務廳、公安廳、政數局，52 備案則需經過交警、海關、 49 參見澳門特別行政區個人資料保護辦公室，「中國銀行（澳門）股份有限公司基於委託次合同人（中國銀行股份有限公司）提供銀行電腦系統營運服務而將自然人客戶之個人資料轉移到澳門特別行政區以外的地方」， https://www.gpdp.gov.mo/file/Permission/06-A-2022-GPDP_CN.pdf，瀏覽日期為 2023年 9月 2日；參見澳門特別行政區個人資料保護辦公室，「廣發銀行股份有限公司澳門分行基於委託次合同人（廣發銀行股份有限公司）提供名單庫系統、資料儲存與資訊技術及支援服務，而將澳門公權力機關協查函或查封函內所涉及的自然人之個人資料轉移到澳門特別行政區以外的地方」，https://www.gpdp.gov.mo/file/Permission/02-A-2022-GPDP-TC.pdf，瀏覽日期為2023年 9月 2日；參見澳門特別行政區個人資料保護辦公室，「關於星展銀行（香港）有限公司基於委託次合同人提供資訊技術支援及資料儲存服務，而將黑名單系統中的澳門法院公告刊登被查 封 的 人 士 之 個 人 資 料 轉 移 到 澳 門 特 別 行 政 區 以 外 的 地 方 」 ， https://www.gpdp.gov.mo/file/Permission/12-A-2021-GPDP-TC.pdf，瀏覽日期為 2023年 9月 2日。 50 此處引述兩地行政機關在生產要素跨境方面的經驗，不僅是因為數據與人員、車輛所承載的勞動力、技術等同為生產要素，而且是出於同擬性的考量：數字中包含著「電子人員」，且在跨境傳輸中起到類似於車輛的載具作用（對於「電子人員」以及「賽博財產」而言）。更重要的是，人員與車輛的流通是以數據互聯為前提的。 51 參見中華人民共和國中央人民政府，「粵澳宣佈開通橫琴口岸新旅檢區域」，https://www.gov.cn/xinwen/2020-08/18/content_5535518.htm，瀏覽日期為 2023年 8月 12日。 52 參 見 公 安 部 ， 「 『 澳 車 北 上 』 政 策 實 施 『 五 問 』 」 ， https://baijiahao.baidu.com/s?id=1752774318292935597&wfr=spider&for=pc，瀏覽日期為 2023 年 8

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 103 邊檢等部⾨。53 ⽽橫琴⼜岸⼆期⾞道採⽤「聯合⼀站式」通關，涉及兩地的海關、邊檢，以及澳⾨衛⽣局 5個單位。54 以上這些⾏政機關的合作，因政務數位化的緣故無⼀不涉及數據的跨境傳輸。雖其並未局限於跨境數據流動的某⼀法律⼯具，然兩地有關部⾨就各個領域、基於實務⽬的的資訊轉移合作，已然積累了相應的經驗。對於 SCCs ⽽⾔，主要包括是否適⽤以及如何適⽤（執⾏與監管）兩個⽅⾯。前者涉及⾏政機關執法⼈員判斷在何種合法性前提下進⾏操作、傳輸數據的技術⽀持、數據種類的判斷及安全保障等。因政務數據涉及公共利益，⼀般被排除於 SCCs 的適⽤範圍之外，然鑒於數據性質邊界的模糊性，具體場景尤需經驗豐富者（甚至跨部⾨或非⾏政機構職員）進⾏個案判斷。後者則關係到公共當局的有效對話與溝通管道與⾏業默認的⾏為規範，如對不同⾏業的數據傳輸要求的判斷（法律意⾒），對在跨境數據傳輸範本合同之外被頻繁添加條款的認識等。橫琴志在打造區別於內地其他城市的數據監管機構，是以其合作不僅局限於雙向（琴—澳）關係，還有可能呈現三角結構（琴—澳—陸）。為應對此複雜局⾯，對已有的機構合作經驗進⾏總結歸納尤為重要。在此基礎上打破觀念壁壘，推陳出新，推動「橫琴經驗」轉變成「SCCs智慧」，建⽴常態化綜合數據執法機制，才能促進兩地個⼈信息的「⼀地（機構）兩檢」、「澳數北上」以及「陸數南下」。 四、平台化創新基礎上的兩地標準合同條款方案 實現兩地合同條款的規則銜接與創新，達致以同⼀ SCCs 有效跨境傳送數據之⽬的，系建設數位化⼤灣區西岸驅動點的關鍵環節。欲克服現存之原則、法律術語、性質、適⽤前提、範本內容、程序及監管機構⽅⾯的壁壘，需借助橫琴深合區的試點優勢。鑒於兩地已建⽴成熟及穩定運⾏的數據傳輸制度，民事基本法或專⾨法修改路徑顯然不適合。是以應在深度合作區的制度容量之內，同時利⽤頂層設計賦予 月 13日。 53 參 見 魏 蒙 ， 「 拱 北 海 關 完 成 首 單 『 澳 車 北 上 』 車 輛 海 關 備 案 手 續 」 ， http://gd.people.com.cn/n2/2022/1222/c123932-40239713.html，瀏覽日期為 2023年 8月 13日。 54 參 見 澳 門 特 區 發 佈 ， 「 黃 少 澤 ： 橫 琴 口 岸 二 期 車 道 擬 於 8 月 底 試 運 行 」 ， https://mp.weixin.qq.com/s/Y53Y8Kz4zFrVQFktv6RWdw，瀏覽日期為 2023年 8月 13日。

李浩 104 的改⾰發展⾃主權與吸收底部⽣產要素跨境流動經驗，推出「上位相容，下位協同」55（如圖三所⽰）的新式標準合同條款。 圖三 「雙管齊下」的平臺化創新路徑 （㆒）、原則與術語闡明 在理論及實踐上，兩地在處理（傳輸）數據⽅⾯的原則呈現交錯狀態，⽽其差異很⼤程度來⾃⽴法結構以及理解角度的不同。雖則如此，此種不協調仍會導致法律適⽤的混亂。然⼀般性的原則皆由民事基本法以及專⾨法規定，其銜接與修改涉及中央與特區層⾯的溝通，且實現成本極其⾼昂。⽽由 SSCs 這類施⾏細則倒逼上位法改⾰，似亦墮於荒誕，鮮有先例。故宜在不悖犯已有「⼤原則」的基礎上，在 55 此規則銜接與創新的基本框架是指在不違反雙方上位法（指民事基本法、行政組織法與專門法）的前提下，主要由有關職權部門進行溝通與對接，以部門規章或者實行細則的形式實現 SCCs 的統合；如確需與上位法抵觸，或涉及地方（廣東省、珠海經濟特區）與澳門特區立法權，則提議雙方立法機構進行協商進行變通。 底部創新：數據傳輸實踐 不成文規範與「行業先行」 補次性及梯級適用 （SCCs 的平臺共建與規則共商） 琴澳版 SCCs 上位相容 下位協同 頂層設計：分線管理 區域性、試驗性、突破性 頂層設計：治理體系改革 「職能分領」 橫琴辦：黨建與數據安全 執委會下屬機構：SCCs 的起草、備案與監管 頂層設計：多元化糾紛解決模式 指定管轄、臨時仲裁、程序協助 底部創新：行政機構合作 雙方、多方人員參與 大灣區內部交流與海外聯繫 原則與術語 （SCCs 的平臺共建） 統一樣本合同 （SCCs 的規則共商） 聯合執法與司法協助 （SCCs 的機構共管與成果共用）

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 105 新 SCCs 指導性規範中闡釋更有領域性、更具體的「⼩原則」。事實上這也是內地〈個⼈信息出境標準合同辦法〉第三條的內容，⽽此處「⼩原則」與其不同之處在於聯合協商規範。協商主體應為雙⽅數據監管機構，其內容至少包括以下兩個⽅⾯：⼀、申明該 SCCs指引規範及其範本系屬合作性試驗，⽬的在於促進兩地數據流通。得益於橫琴實⾏的「分線管理」制度，該標準合同條款的實施範圍可以僅限於兩線之間，實⾏「⼀線放開，⼆線管住」的政策。數據向內地其他區域轉移仍需遵守有關法律。⼆、闡釋 SCCs 適⽤基本不違反兩地上位法規定的前提，以合同條款傳輸數據的要求與程序為基本框架，同時實現相對於粵澳版 SCCs 的差異化發展。對於狹義的原則表達，應在充分考慮兩地法律傳統以及社會背景的情況下，由雙⽅數據監管機構協定。其所⽤語⾔，當以清楚、簡潔為要，以便為兩地使⽤者提供指導。 兩地在數據傳輸領域所使⽤不同的法律術語同樣會導致對 SCCs 的誤解於適⽤困難。在漢語環境中，儘管⼤部分⽤語在結合專⾨的法律解釋後均可為雙⽅暢達理解，但仍有少量術語缺乏對應或完全對應概念，又或者在法律中無明確定義。為解決此問題，兩⽅數據監管機構應在對比琴澳法律傳統與⽤語習慣後，聯合提供合同條款中所涉及的專業術語標準並對之做出闡釋。⽬前可供選擇之路徑有三：⼀即兼採內地與澳⾨的⽤詞習慣，推出新的法律術語版本︔⼆則以內地法律術語為主，對與澳⾨對相應語作補充性解釋︔三為以澳⾨法律術語為主，對與內地相應術語作補充性解釋。由於內地擁有海量互聯網企業及個⼈信息，系主要的數據傳輸⽅，故⽽宜適採路徑⼆。在此前提下，對於缺乏對應或完全對應概念之術語（尤其係數據處理主體），應以適當解釋厘清關係︔⽽對於無直接定義的術語，則應增添其明確定義。具體可參考粵港個⼈信息保護認證機制、56 粵港版 SCCs與粵澳版 SCCs，然亦需注意其地域局限性。與此同時，為⽅便其⼤灣區以及跨國適⽤，還應提供官⽅的英⽂、葡⽂以及其他外語版本。考慮到兩地已有之規範習慣，定義部分應於標準合同範本較前位置以單獨章節予以闡釋為宜。 56 該文件由全國信息安全標準化技術委員會於 2023年 11月 1日發佈，內容涵蓋了數據傳輸的原則、術語等項，為奠定區域數字治理共識做出了有益的嘗試。然而其宣稱適用範圍為「9+1」，即廣東9個城市與香港特別行政區，似乎專為建設香港數據傳輸中心服務。

李浩 106 （㆓）、補次性及梯級㊜用 以合同條款作為合法性基礎傳輸低重要性、低敏感度數據，不僅是內地與澳⾨現⾏法的規定，亦是國際社會的慣例。SCCs雖主導⼤部分數據的跨境傳輸，然由於其所傳輸數據的次要性，常被認為是⼀種補充措施。琴澳聯合推出的標準合同條款將延續這⼀傳統，與其他國家與地區接軌，是以應為跨境數據傳輸的補次性措施，僅在不滿⾜上層條件的情況下降梯適⽤。鑒於此做法恰為內地所實⾏，⽽在澳⾨亦有相當基礎，故在協商過程中並不會⾯臨較⼤阻⼒。 關於適⽤的前提，儘管兩地合同條款都採⽤了「負⾯清單」，即對不可適⽤的情形做出列舉，然其出發角度不同。內地〈個⼈信息出境標準合同辦法〉第四條是以數據的性質與數量為標準，⽽澳⾨〈個⼈資料保護法〉第⼗九條、第⼆⼗條則系以傳輸利益為考量。如需消彌雙⽅規範差異，SCCs的梯級適⽤在堅持「負⾯清單」形式之餘，還要對已有的條款進⾏整合與突破。中國內地的排除條件主要涉及國家安全以及公共利益，⽽澳⾨的則強調已有其他合法的傳輸基礎，主要為重要公共利益以及個⼈締約要求。基於此，琴澳版 SCCs 的適⽤應在排斥涉及國家安全與公共利益外，應考慮個⼈締約權益與重⼤利益，實現個⼈信息受保護權的客觀法功能︔並⼀定程度放寬限制範圍，以追求「-1+-1<-2」的效果。57 具體⽽⾔，對於重要數據以及關鍵資訊基礎設施運營者在⼯作中收集與產⽣的數據，禁⽌通過 SCCs 對外傳輸︔對涉及重⼤公共利益、個⼈重要權益或關鍵合同利益的數據，原則上無需通過SCCs對外傳輸︔⽽在對個⼈信息以及敏感資訊的數量⾨檻⽅⾯，則可以考慮適當提⾼或者取消。數量閾值過低將標準合同條款之適⽤局限於中⼩企業與臨時傳輸，與橫琴建⽴數字貿易港與發展⾼科技產業的定位不符。58 為了促進琴澳數據流通，尊重澳⾨法律傳統，突出深合區的制度創新性，取消對處理、傳輸個⼈數據或者敏感數據達到⼀定數量的個⼈信息處理者適⽤ SCCs 的限制無疑更為恰當。鑒於《規範和促進數據跨境流動規定》已表現出以上修改趨勢，粵港版 SCCs與粵澳版 SCCs都採⽤了「法無禁⽌皆可為」的默⽰⽅式，橫琴的 SCCs 嘗試或與之呈現「創新競 57 參見王錫鋅，〈個人信息國家保護義務及展開〉，《中國法學》，第 1期（北京：2021.02）2021（01），151-153。 58 參見金晶，〈作為個人信息跨境傳輸監管工具的標準合同條款〉， 23。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 107 賽，局部突破」以及「重複⾰新」的局⾯，使得區際的法律預景更加顯錯綜複雜。其間涉及可能與上位法抵觸的地⽅，59 需致⼒於理順頂層設計關係，解決許可權問題，60 為在「共商共建」體制下實體法的協同提供基礎。更準確來說，要雙⽅⽴法機構，包括全國⼈⼤、廣東省⼈⼤、珠海市⼈⼤與澳⾨特區⽴法會參與協商，以求在深合區中獲取區域性的施⾏權。善⽤國家網信辦試圖賦予⾃貿區的權⼒，在經省級部⾨批准以及向中央報備後，實⾏具有本地特⾊的數據清單管理機制。在總結現有部⾨經驗的基礎上，討論 SCCs與數據可攜帶權的兼容性問題：⼀⽅⾯，以 SCCs完善各部⾨、⾏業的數據可攜帶權法律基礎︔另⼀⽅⾯以可攜帶權的實踐補充 SCCs的適⽤條件，尤其是對醫療、海關等領域數據性質的判斷。利⽤橫琴分線管理之優勢，創建獨特的「數字特區」，發展、完善新型信息交易合同條款，發揮與葡語國家數據貿易中轉站及樞紐港的作⽤，為內地企業通過港澳實現數據出⼜提供制度便利。61 （㆔）、琴澳統㆒樣本合同 SCCs 制度之核⼼為官⽅通過正式管道提供或推薦的範本合同條款。中國內地公佈之〈個⼈信息出境標準合同辦法〉及其附件已於 2023年 6⽉ 1⽇⽣效，向標準合同本⼟化邁出了重要的⼀步。⽽澳⾨特區個資局並未發佈普適性樣板合同，僅將合同所提供保護程度作為審批許可的關鍵考量，是以其所⾏制度並不屬於嚴格意義上的標準合同條款。又因公共當局在其官網公佈意⾒書及許可結果涉及對合同條款的評價細節，可在⼀定程度上視為半個 SCCs。為提⾼數據傳輸效率與法律的可期待性計，對⼤部涉及私法的條款內容，應在雙⽅就不同點於法律內所達成共識之基礎上出台琴澳版標準合同，⼒求實現「1+0.5>1.5」的聯動創新效應。62 59 因為 SCCs 與安全評估的「整體聯動」，涉及的部門規章變通不僅包括《個人信息出境標準合同辦法備案指南（第一版）》及其附件合同，還有《數據出境安全評估辦法》以及相關網路數據的分類評級辦法。 60 參見唐曉晴、李可，〈橫琴粵澳深度合作區經濟協同立法：動機、問題與路徑〉，《澳門法學》，第 1期（澳門：2022.03），10。 61 See Fengan Jiang, “China’s legal efforts to facilitate cross-border data transfers: a comprehensive reality check.” Asia Pacific Law Review 2 (Hong Kong: 2023.06) :12-13. 62 參見呂冬娟、田兆麟，〈橫琴粵澳深度合作區民商事規則銜接的路徑探究〉，《澳門法學》，第2期（澳門：2022.04），175。

李浩 108 範本應適採陸⽅的「機構提供，傳輸⽅遵守」的形式，⽽非澳⽅將敘述權交由傳輸⽅的路徑。⽽至於具體條款，鑒於內地款項的覆蓋範圍比澳⾨要廣泛，因⽽應以內地的範本為主要參考。其至少包括數據傳輸雙⽅信息、傳輸基本情況之說明（如數據種類、⽅式等）、定義、個⼈信息處理者的義務、境外接收⽅的義務等基本項。對雙⽅共有的條項，應在充分考慮兩地的科技需求及民事習慣後，如數據傳輸各⽅的義務、數據保護措施等，修改為能提供更⾼程度保護的條款。然在雙⽅同意之前提下，亦可適當降低標準以促進數據流通，具體應在粵港版 SCCs 與粤澳版 SCCs減輕接收⽅義務的基础上再⾏突破。63 對⼀⽅獨有，另⼀⽅無的條項，應在新版SCCs中加入，如定義、次級處理者的使⽤等項。儘管有關術語的定義已在澳⾨〈個⼈資料保護法〉中有詳細闡釋，然在琴澳版標準合同中其將被賦予不同的角⾊，如在上述討論中所確定的，應採⽤内地為主，澳⾨為輔的解釋⽅法。尤應吸收澳⾨「案例法」之經驗，將有關條款的合理內核成⽂化、規範化。在結構以及⾏⽂習慣上，應當兼顧雙⽅，致⼒於減少溝通障礙，便於理解。標準條款禁⽌變更，應⼀體化適⽤，以收「固定」保護標準之效。64 在標準條款之外，締約雙⽅允許根據合意添加補充條款，然不得與上項抵觸。在性質上，合同範本宜被認定為橫琴深合區的實在（試驗）法，同時也是相對澳⾨特區及⼤灣區其他城市⽽⾔的⽰範法。此外，還需要研究域外經驗，尤其是具有版塊結構⽽便於移植的歐盟規範，65 嘗試提供多個⽅向如C-C、C-P、P-P、P-C等的傳輸合同範本。不同於上述原則、術語、適⽤前提等項，標準合同範本的銜接與創新⼯作量⼤、極其繁瑣，同時又至關重要，既需要琴澳⽴法機構建⽴長效機制進⾏定期溝通，又要求雙⽅數據監管部⾨、乃至其他有關部⾨充分發揚在「⼀國兩制」框架下團結合作，互助共贏的傳統，以認真負責、謙虛謹 63 參見中華人民共和國網絡信息辦公室，「粵港澳大灣區（內地、香港）個人信息跨境流動標準合同實施指引」，http://www.cac.gov.cn/2023-12/13/c_1704042786237103.htm，瀏覽日期為 2023年 12月 13日；參見中華人民共和國國家互聯網信息辦公室，「粵港澳大灣區（內地、澳門）個人信息跨境流動標準合同實施指引」， https://www.cac.gov.cn/2024-09/10/c_1727567893741986.htm，瀏覽日期為 2023年 12月 13日。 64 參見金晶，〈個人數據跨境傳輸的歐盟標準—規則建構、司法推動與範式擴張〉，《歐洲研究》，第 4期（北京：2021.08），98。 65 See M.P.Ramaswamy, “Legal Standards Governing Data Protection in Europe and Potential Sources of Legal Transplants in Macau SAR.” International Journal of Research and Innovation in Social Science 5 (Delhi: 2021): 649.

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 109 慎，又謀求突破的態度推進。可建⽴「⾏業先⾏」之模式，66 優先發展琴澳重點⾏業（類似做法亦可⾒於⾹港特區）67 如科創、⽂旅會展的數據傳輸合同，68 探索⾏業特有規律與權利義務特點，為繼續完善條款內容積累經驗。69 對因制度嘗試所導致損失之情形，倘程序正當且非謀求私利，應為〈促進條例〉第五條容錯機制所覆履，不得追究單位或個⼈的責任，以免打擊創新熱情。 （㆕）、聯合執法與司法㈿助 ⾏政機關執法以及司法部⾨的救濟是 SCCs成為「活法」的關鍵。省政府的「數字灣區」專責⼩組僅有統籌⽽無地⽅執法職能，負責⽇常⼯作的政務服務數據管理局側重於公共數據，是以仍存在深化機構改⾰的必要。法定機構即執委會的創⽴為深合區體制改⾰拉開了序幕，但同時也導致了部分職能部⾨如數據監管模糊與空⽩。鑒於網信辦的複合性質，此處宜採取「職能分領」的路徑：與黨建與數據安全有關的任務，由省派出機構負責︔⽽促進數據流通及虛擬經濟發展的權能，由執委會下屬專⾨數據監管機構⾏使。執委會內部與之有交叉關係的，為法律事務局與經濟發展局。對在執委會內設專⾨的數據監管機構，⽬前有三套⽅案：⼀、以「兜底條款」將其添設於法律事務局之下，以經濟發展局為業務關聯與顧問機構。⼆、以「兜底條款」將其添設於經濟發展局之下，以法律事務局為業務關聯與顧問機構。三、在法律事務局與經濟發展局之外另設獨⽴的數據監管局，對執委會直接負責。⽅案三雖考慮到數字治理的重要性與獨特性，表現出非凡的前瞻視野，然則會導致執委會權⼒結構的尷尬：如所有空⽩職能均需新設部⾨以填充，則原來下屬機構的數量將⾯臨規模擴張的局⾯。⽅案⼀與⽅案⼆則分別著眼於數據跨境傳輸的穩定性與發展 66 參見劉俊敏、郭楊，〈我國數據跨境流動規制的相關問題研究—以中國（上海）自由貿易試驗區臨港新片區為例〉，《河北法學》，第 7期（石家庄：2021.07），87-88。 67 特區政府將採取系列措施，促進銀行、征信與醫療行業數據在大灣區範圍內流通，並視情況推行至其他界別。參見中華人民共和國香港特別行政區政府創新科技及工業局，「香港促進數據流通及保障數據安全的政策宣言」，https://www.itib.gov.hk/assets/files/Policy_Statement_Chi.pdf，瀏覽日期為 2023年 12月 12日。 68 參見中央人民政府駐澳門特別行政區聯絡辦公室，「國家發展改革委關於印發『橫琴粵澳深度合作區鼓勵類產業目錄』的通知」，http://www.zlb.gov.cn/2023-04/04/c_1211964421.htm，瀏覽日期為2023年 4月 4日。 69 參見唐曉晴、李可，〈橫琴粵澳深度合作區經濟協同立法：動機、問題與路徑〉，5。

李浩 110 性，為決策者展現了經典的⼆元難題。筆者以為，應當根據主要⽭盾的變化，辯證地建⽴動態數據監管機構。在琴澳 SCCs 制度草創以及試⾏之初，其主要問題是內地的上位法變通以及兩地的規則銜接、條款創新與系統對接，涉及機制的穩定運⾏，因⽽應將數據監管機構設於法律事務局下。⽽在標準合同制度成熟以後，其⾯臨的難題則轉變為如何有效促進數據流通以及兩地虛擬經濟的發展，此時數據監管機構應轉隸經濟發展局。又因我國法定機構具有類似新加坡「⼀機構⼀⽴法」之形式，且有賴⽴法對其職權進⾏廓清，是以即需橫琴執委會⾏使法規建議權請求珠海特區⽴法，又需中央網信辦在具體規則適⽤上以予授權。70 在⼈員構成上，橫琴數據監管機構應作比例考慮，在充分吸收澳⾨特區⾏政管理⼈員及專業⼈⼠之餘，尤需制定詳細⽅案延請⾹港地區、葡萄牙、歐盟與其他國家的海外⼈才，以參考不同法系的有益經驗。在職權⾏使上，橫琴數據監管機構不僅負責與澳⾨個資局協商出台標準合同條款指導性⽂件以及樣本，同時還負責規則的執⾏與監督。對於 SCCs的適⽤程序，為了降低數據傳輸成本與增加法律確定性，宜採⽤「標準合同條款+備案」的形式，備案機構從省網信辦下移至橫琴數據監管機構。備案所需提交的材料當僅限於標準合同，不包括個⼈信息保護影響評估報告。⾃評報告事實上是將部分合規檢查義務倒置於個⼈信息處理者，要求其在⾯對規範信息差與不斷變換的政策環境時承擔巨⼤風險。在備案中排除⾃我評估報告，在使監管機構履⾏審查職能之外，還可讓⾃評步驟失去必要性基礎，從⽽通過環節控制提⾼⾏政效率。此後仍需繼續簡化以至取消備案程序，減少對資訊轉移的限制。在保持與政府、企業與社會組織「適距控制」之餘，71 為實現促進數據在更廣泛的區域內流動，創造「1+1>2」的聚合監管效應，該機構還應在琴澳職能機構已有良好合作關係的基礎上，參考「⼀地兩檢」、「澳⾞北上」的跨部⾨協助經驗，⼀⽅⾯加強與珠海市其他區，⼤灣區八⼤城市，廣東省內其他城市以及內地其他省份有關部⾨的交流，打造「數據內循環」的有效路徑︔另⼀⽅⾯促進與澳⾨特區、⾹港特區、葡語系國家與歐盟的公共當局對話，搭建「中國—葡語國家數字圈」、「數據⼀帶⼀路」發展平台。 70 參見葉必豐，〈論我國的法定機構〉，《中外法學》，第 3期（北京：2022.06），631-632。 71 參見石亞軍、馬擎宇，〈我國法定機構的再思考：理念更新與制度改進〉，《中國行政管理》，第 8期（北京：2022.08），22-23。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 111 除數據監管部⾨的合作外，司法協助構成了數據傳輸權利救濟與糾紛解決的「底線」，也是琴澳 SCCs 制度穩定運⾏的信⼼所在。標準合同條款雖為官⽅提供版本，然其本質上⼤部分屬民事合同，在琴澳民商事規則銜接的範疇內。⽬前橫琴法院已獲廣東⾼院集中管轄原珠海其他基院負責的標的 4,000萬以內的案件。依照此思路，涉及琴澳版 SCCs 糾紛的案件，包括以橫琴數據監管機構為對象的⾏政訴訟，也可通過⾼院授權的⽅式由深合區法院管轄。此外，在橫琴深合區建⽴的多元化糾紛解決試驗區，其業務涵蓋審判、仲裁、調解等。關於適⽤標準合同條款的數據傳輸糾紛，因涉及境外主體以及法律適⽤⽽較為複雜，善⽤仲裁、調解等模式可以有效節約成本，提⾼糾紛解決质量與效率。尤其是關於臨時仲裁在 SCCs 糾紛解決上的適⽤，應回應逐漸開放（承認）之總體趨勢，實現局部規則的超前引領。72 ⽽為了促進多元糾紛解決模式的適⽤，兩地的取證、送達、判決互認、執⾏協助均需提上⽇程。具體則可由最⾼⼈民法院與澳⾨特區共同協商，逐步推進。 五、結語 跨境數據傳輸關係到區域產業發展與社會進步，系信息時代法治環境與制度競爭⼒的關鍵所在。SCCs雖為補次性傳輸⼯具，卻承擔著現今主要的數據流動任務，亦被寄予未來重構國際數據治理秩序、改善數字⽣態之厚望。有鑒於此，如何根據中國內地與澳⾨特區的法制特點，實現兩地數據傳輸中合同條款的規則銜接與創新，推出突破性的指導性⽂件與合同範本，確為長期困擾區域⾏政管理學者與數據法專家的⼀⼤難題。然則橫琴粵澳深度合作區的構想，尤其是其作為改⾰「試驗⽥」與「孵化器」的定位，則為此問題之解決提供了難得的歷史性機遇。 就⽬前⽴法⽽⾔，中國內地與澳⾨的數據傳輸合同條款制度存在系統性差異。前者已推出適⽤指導與合同樣本，可以之為獨⽴的跨境數據流動基礎，為典型的SCCs形式︔後者尚未推出普適性（不⽌於灣區）的合同版本，⽽僅以⾜夠提供相當權利保障之合同為個⼈資料轉移許可的考量條件。具體⽽⾔，兩地之數據處理原則、法律術語、合同條款性質、適⽤前提、範本內容、程序以及監管機構皆有不同，⾜ 72 參見王玫黎、陳雨，〈海南自貿港探索建立臨時仲裁規則問題研究〉，《澳門法學》，第 2期（澳門：2023.06），140-141。

李浩 112 以影響數據傳輸的品質與效率。橫琴深合區則從「⾃上⽽下」的頂層設計，包括「分線管理」、法定機構的設⽴與多元糾紛解決機制等︔以及「⾃下⽽上」的底部創新，也即既有的數據傳輸經驗和兩地⾏政機關的合作兩個維度出發，消除制度壁壘並推出橫琴版標準合同條款。在原則與法律術語⽅⾯，藉由「分線管理」闡釋其區域性與特殊性，推動 SCCs的「平台共建」︔在性質、適⽤前提、範本內容⽅⾯，利⽤治理體系改⾰進⾏雙⽅協商確定，同時根據數據傳輸的經驗予以調整，推動 SCCs 的「規則共商」︔在權利救濟與糾紛解決⽅⾯，設置新的法定機構進⾏管理，吸收兩地⾏政機構合作的經驗，繼續促進多元糾紛解決機制的發展，推動 SCCs 的「機構共管」與「成果共享」。如此⽅案雖未盡善，然畢竟系基於內地與澳⾨數據傳輸需求與制度銜接的⼀點想像，對於兩地之將來發展或有⼀⼆智⼒貢獻。⽽至於如何吸收境外先⾏經驗，研究琴澳本⼟情況，內向完善 SCCs 改⾰細節︔又如何有效運⾏樹⽴⽰範效應，使之成為灣區特⾊制度及中國軟實⼒的⼀部分，外向參與國際話語權建構，仍有賴進⼀步研究予以釋發。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 113 徵引書目 ㆒、㆗文書目與期刊 [1] 楊崇蔚、廖志漢、廖志聰，《澳⾨個⼈資料保護制度》，北京：社會科學⽂獻出版社，2014。 [2] 於興中，《數字素養：從演算法社會到網路 3.0》，上海：上海⼈民出版社，2022。 [3] 廣東外語外貿⼤學粵港澳⼤灣區研究院課題組，〈數據要素跨境流動與治理機制設計—基於粵港澳⼤灣區建設的視角〉，《國際經貿探索》，第 10期（廣州：2021.10），86-98。 [4] 李艷華，〈隱私盾案後歐美數據的跨境流動監管及中國對策—軟數據本地化機制的⾛向與標準合同條款路徑的⾰新〉，《歐洲研究》，第 6期（北京：2021.12），25-49+5-6。 [5] 李可、唐曉晴，〈橫琴粵澳深度合作區：理念創新與制度構建〉，《港澳研究》，第 1期（北京：2022.03），13-24。 [6] ⾦晶，〈作為個⼈信息跨境傳輸監管⼯具的標準合同條款〉，《法學研究》，第 5期（北京：2022.10），19-36。 [7] 楊崇蔚，〈網路空間的個⼈資料保護—從監管當局角度分享澳⾨經驗〉，《網絡信息法學研究》，第 2期（北京：2019.12），79-83。 [8] 唐曉晴、吳奇琦，〈澳⾨地區民法上個⼈資訊保護權的體系定位〉，《國家檢察官學院學報》，第 5期（北京：2021.10），75-88。 [9] 姚佳，〈論個⼈資訊處理者的民事責任〉，《清華法學》，第 3 期（北京：2021.06），41-54。 [10] 藍江，〈⽣命檔案化、演算法治理和流眾—數字時代的⽣命政治〉，《探索與爭鳴》，第 9期（上海：2020.09），105-114+159。 [11] 蔡⽴東，〈為什麼「數字⼈權」是第四代⼈權〉，《數字法治》，第 3期（北京：2023.06），1-7。 [12] ⽥旭，〈歐盟個⼈數據保護法的全球影響成因與啟⽰〉，《江西財經⼤學學報》，

李浩 114 第 4期（南昌：2020.08），2020（04），135-147。 [13] 王錫鋅，〈個⼈信息國家保護義務及展開〉，《中國法學》，第 1期（北京：2021.02）2021（01），145-166。 [14] 唐曉晴、李可，〈橫琴粵澳深度合作區經濟協同⽴法：動機、問題與路徑〉，《澳⾨法學》，第 1期（澳⾨：2022.03），1-12。 [15] 呂冬娟、⽥兆麟，〈橫琴粵澳深度合作區民商事規則銜接的路徑探究〉，《澳⾨法學》，第 2期（澳⾨：2022.04），174-186。 [16] ⾦晶，〈個⼈數據跨境傳輸的歐盟標準—規則建構、司法推動與範式擴張〉，《歐洲研究》，第 4期（北京：2021.08），89-109。 [17] 劉俊敏、郭楊，〈我國數據跨境流動規制的相關問題研究—以中國（上海）⾃由貿易試驗區臨港新⽚區為例〉，《河北法學》，第 7期（⽯家庄：2021.07），76-90。 [18] 葉必豐，〈論我國的法定機構〉，《中外法學》，第 3期（北京：2022.06），626-642。 [19] ⽯亞軍、⾺擎宇，〈我國法定機構的再思考：理念更新與制度改進〉，《中國⾏政管理》，第 8期（北京：2022.08），18-25。 [20] 王玫黎、陳⾬，〈海南⾃貿港探索建⽴臨時仲裁規則問題研究〉，《澳⾨法學》，第 2期（澳⾨：2023.06），139-150。 ㆓、英文書目與期刊 [1] Kuner C. European Data Privacy Law and Online Business. Oxford: Oxford University Press, 2003. [2] Muruga Perumal Ramaswamy, “Diversity in Data Protection Standards Governing Cross-border Commercial Transactions and the Role of GDPR as a Potential Harmonization Model” Proceedings, International Conference on Technological and Social Innovations, Ambassador Bangkok Hotel and GrandPresident Hotel, Bangkok, 2019. [3] Liu Lu, Qi Qi, “Determining the Validity of Ad Hoc Arbitration Agreements in China:

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 115 Past, Present, and Future.” Frontiers of Law in China 14.1 (Beijing: 2019.03): 116-140. [4] Fengan Jiang, “China’s legal efforts to facilitate cross-border data transfers: a comprehensive reality check.” Asia Pacific Law Review 2 (Hong Kong: 2023.06): 1-21. [5] M.P.Ramaswamy, “Legal Standards Governing Data Protection in Europe and Potential Sources of Legal Transplants in Macau SAR.” International Journal of Research and Innovation in Social Science 5 (Delhi: 2021.05): 639-650. ㆔、電子網路㈾源 [1] 中華⼈民共和國中央⼈民政府，「中共中央 國務院關於構建更加完善的要素市 場 化 配 置 體 制 機 制 的 意 ⾒ 」 ， https://www.gov.cn/zhengce/2020-04/09/content_5500622.htm，瀏覽⽇期為 2020年 4⽉ 9⽇。 [2] Dan Ciuariak and Maria Ptaskhina. “The Digital Transformation and the Transformation of International Trade.” Accessed July 27, 2023. http://www.iberglobal.com/files/2018-2/Digital_Transformation_Trade.pdf. [3] European Parliament. “The CJEU judgment in the Schrems II case” Accessed October 30,2024. https://www.europarl.europa.eu/RegData/ etudes/ATAG/2020/652073/EPRS_ATA(2020)652073_EN.pdf. [4] 廣東省⼈民政府，「廣東省⼈民政府辦公廳關於印發『數字灣區』建設三年⾏動 ⽅ 案 的 通 知 」 ，http://www.gd.gov.cn/zwgk/gongbao/2023/31/content/post_4287722.html?version=slh，瀏覽⽇期為 2023年 11⽉ 7⽇。 [5] 中華⼈民共和國中央⼈民政府，「中共中央 國務院印發『粵港澳⼤灣區發展規劃綱要』」， https://www.gov.cn/xinwen/2019-02/18/content_5366593.htm#1，瀏覽⽇期為 2023年 7⽉ 23⽇。 [6] 國務院港澳事務辦公室，「中共中央 國務院印發『橫琴粵澳深度合作區建設總 體 ⽅ 案 』 」 ，

李浩 116 https://www.hmo.gov.cn/zcfg_new/gafg/gfxwj/202211/t20221125_23891.html，瀏覽⽇期為 2023年 7⽉ 23⽇。 [7] 個 ⼈ 資 料 保 護 辦 公 室 ， 「 2012 年 度 年 報 」 ， https://www.gpdp.gov.mo/file/Annual_Report/2012_files/1_all.pdf，瀏覽⽇期為2023年 7⽉ 28⽇。 [8] Nigel Cory and Luke Dascoli. “How Barriers to Cross-Border Data Flows Are Spreading Globally, What They Cost, and How to Address Them.” Information Technology & Innovation Foundation. Accessed September 3,2023. https://itif.org/publications/2021/07/19/how-barriers-cross-border-data-flows-are-spreading-globally-what-they-cost/. [9] 個⼈資料保護辦公室，「負責處理個⼈資料的實體須知」， https://www.gpdp.gov.mo/zh_tw/law_routine_3.html，瀏覽⽇期為 2023年 8⽉ 8⽇。 [10] 澳⾨特別⾏政區個⼈資料保護辦公室，「廣發銀⾏股份有限公司澳⾨分⾏基於委託次合同⼈（廣發銀⾏股份有限公司）提供名單庫系統、資料儲存與資訊技術及⽀援服務，⽽將澳⾨公權⼒機關協查函或查封函內所涉及的⾃然⼈之個⼈ 資 料 轉 移 到 澳 ⾨ 特 別 ⾏ 政 區 以 外 的 地 ⽅ 」 ，https://www.gpdp.gov.mo/file/Permission/02-A-2022-GPDP-TC.pdf，瀏覽⽇期為2023年 9⽉ 2⽇。 [11] 澳⾨特別⾏政區個⼈資料保護辦公室，「關於 A銀⾏申請與⾹港 B銀⾏以『互聯』⽅式處理其『客⼾資料』」，https://www.gpdp.gov.mo/file/Permission/01-A-2009-GPDP-TC.pdf，瀏覽⽇期為 2023年 9⽉ 4⽇。 [12] 澳⾨特別⾏政區個⼈資料保護辦公室，「關於澳⾨電訊有限公司將數據中⼼業務擴展至⾹港⽽涉及的澳⾨機構轉移個⼈資料到⾹港事宜及相關的通知與許可程序」，2023年 9⽉ 4⽇，https://www.gpdp.gov.mo/file/Opinion/0016-P-2018-GPDP-TC.PDF，瀏覽⽇期為 2023年 9⽉ 4⽇。 [13] 個⼈資料保護辦公室，「負責處理個⼈資料的實體須知」，https://www.gpdp.gov.mo/zh_tw/law_routine_3.html，瀏覽⽇期為 2023年 8⽉ 8⽇。

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 117 [14] 中華⼈民共和國中央⼈民政府，「中共中央 國務院印發『黨和國家機構改⾰⽅案』」，https://www.gov.cn/gongbao/content/2023/content_5748649.htm，瀏覽⽇期為 2023年 8⽉ 9⽇。 [15] 澳 ⾨ 特 別 ⾏ 政 區 ， 「 第 83/2007 號 ⾏ 政 長 官 批 ⽰ 」 ， https://www.gpdp.gov.mo/file/Laws%20and%20Regulations/第 83_2007號⾏政長官批⽰_TC.pdf，瀏覽⽇期為 2023年 8⽉ 9⽇。 [16] 廣東省⼈民政府，〈橫琴粵澳深度合作區發展促進條例〉， http://www.gd.gov.cn /gdywdt/zwzt/ygadwq/zxzc/content/post_4091516.html，瀏覽⽇期為 2023年 8⽉9⽇。 [17] 張雅婷，美婷，「橫琴探路數據跨境流動，打造數據治理創新樣本」，https://m.21jingji.com/article/20210909/0052dd35ee00dd29ba0987a32a353f51.html，瀏覽⽇期為 2023年 8⽉ 10⽇。 [18] 政 府 新 聞 網 ， 「 ⼤ 灣 區 數 據 流 動 合 作 備 忘 錄 簽 訂 」 ，https://sc.news.gov.hk/TuniS/www.news.gov.hk/chi/2023/06/20230630/20230630_120748_916.html，瀏覽⽇期為 2023年 8⽉ 10⽇。 [19] 橫 琴 粵 澳 深 度 合 作 區 ， 「 職 能 介 紹 法 律 事 務 局 」 ， http://www.hengqin.gov.cn/macao_zh_hant/hzqgl/zzjg/znjs/content/post_2990404.html，瀏覽⽇期為 2023年 9⽉ 3⽇。 [20] 橫 琴 粵 澳 深 度 合 作 區 ， 「 職 能 介 紹 經 濟 發 展 局 」 ， http://www.hengqin.gov.cn/macao_zh_hant/hzqgl/zzjg/znjs/content/post_2990403.html，瀏覽⽇期為 2023年 9⽉ 3⽇。 [21] 肖皓⽅，李美琪，「從『⼀院』拓展到『兩中⼼』，珠海國際仲裁院落⼦『集聚 區 』 」 ， http://www.hengqin.gov.cn/macao_zh_hans/ztjh/wqzc/dtzx/content/post_3532564.html，瀏覽⽇期為 2023年 8⽉ 11⽇。 [22] 李恩漢，吳偉傑，「粵澳跨境數據驗證平台上線試運⾏，探索個⼈數據跨境流動新模式」， https://m.thepaper.cn/baijiahao_17470477，瀏覽⽇期為 2023年 9⽉ 2⽇。 [23] 澳⾨特別⾏政區個⼈資料保護辦公室，「中國銀⾏（澳⾨）股份有限公司基於

李浩 118 委託次合同⼈（中國銀⾏股份有限公司）提供銀⾏電腦系統營運服務⽽將⾃然⼈ 客 ⼾ 之 個 ⼈ 資 料 轉 移 到 澳 ⾨ 特 別 ⾏ 政 區 以 外 的 地 ⽅ 」 ， https://www.gpdp.gov.mo/file/Permission/06-A-2022-GPDP_CN.pdf，瀏覽⽇期為2023年 9⽉ 2⽇。 [24] 澳⾨特別⾏政區個⼈資料保護辦公室，「關於星展銀⾏（⾹港）有限公司基於委託次合同⼈提供資訊技術⽀援及資料儲存服務，⽽將⿊名單系統中的澳⾨法院公告刊登被查封的⼈⼠之個⼈資料轉移到澳⾨特別⾏政區以外的地⽅」， https://www.gpdp.gov.mo/file/Permission/12-A-2021-GPDP-TC.pdf，瀏覽⽇期為2023年 9⽉ 2⽇。 [25] 中華⼈民共和國中央⼈民政府，「粵澳宣佈開通橫琴⼜岸新旅檢區域」，https://www.gov.cn/xinwen/2020-08/18/content_5535518.htm，瀏覽⽇期為 2023年 8⽉ 12⽇。 [26] 公 安 部 ， 「 『 澳 ⾞ 北 上 』 政 策 實 施 『 五 問 』 」 ， https://baijiahao.baidu.com/s?id=1752774318292935597&wfr=spider&for=pc，瀏覽⽇期為 2023年 8⽉ 13⽇。 [27] 魏蒙，「拱北海關完成⾸單『澳⾞北上』⾞輛海關備案⼿續」， http://gd.people.com.cn/n2/2022/1222/c123932-40239713.html，瀏覽⽇期為 2023年 8⽉ 13⽇。 [28] 澳⾨特區發佈，「⿈少澤：橫琴⼜岸⼆期⾞道擬於 8 ⽉底試運⾏」， https://mp.weixin.qq.com/s/Y53Y8Kz4zFrVQFktv6RWdw，瀏覽⽇期為 2023 年8⽉ 13⽇。 [29] 中華⼈民共和國網絡信息辦公室，「粵港澳⼤灣區（內地、⾹港）個⼈信息跨境 流 動 標 準 合 同 實 施 指 引 」 ， http://www.cac.gov.cn/2023-12/13/c_1704042786237103.htm，瀏覽⽇期為 2023年 12⽉ 13⽇。 [30] 中華⼈民共和國國家互聯網信息辦公室：「粵港澳⼤灣區（內地、澳⾨）個⼈信息跨境流動標準合同實施指引」， https://www.cac.gov.cn/2024-09/10/c_172756789374198htm，瀏覽⽇期為 2023年 12⽉ 13⽇。 [31] 中華⼈民共和國⾹港特別⾏政區政府創新科技及⼯業局，「⾹港促進數據流通及 保 障 數 據 安 全 的 政 策 宣 ⾔ 」 ，

規則銜接與創新：中國內地與澳門數據傳輸 SCCs制度建構—以橫琴深合區為切入點 119 https://www.itib.gov.hk/assets/files/Policy_Statement_Chi.pdf，瀏覽⽇期為 2023年 12⽉ 12⽇。 [32] 中央⼈民政府駐澳⾨特別⾏政區聯絡辦公室，「國家發展改⾰委關於印發『橫琴粵澳深度合作區⿎勵類產業⽬錄』的通知」，http://www.zlb.gov.cn/2023-04/04/c_1211964421.htm，瀏覽⽇期為 2023年 4⽉ 4⽇。

李浩 120 作者簡介 李浩，澳門大學 2022級法學博士。研究方向為數據法、區域治理、法理學與法律文化 通訊地址：澳門氹仔大學大馬路W32宿舍 通訊電郵：1042146034@qq.com/yc27220@um.edu.mo

The Journal of Macau University of Science and Technology (Hum. & Soc. Sci.) VOL.19 NO.1 March 2025, pp.121-145 DOI: 10.58664/mustjournal.2025.03.004 121 Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy* Cao, Leyi1; Gao, Yajuan2 1 Lecturer, University International College, Macau University of Science and Technology 2 School of Applied Foreign Languages, Zhejiang International Studies University Abstract: From a black feminist perspective, especially Narayan’s theory on “epistemic privilege”, the paper examines the cross-racial female relation in Toni Morrison’s A Mercy, focusing on two pairs of friends – Lina and Rebekka, and Florens and Daughter Jane. In each relation, the two women, who are in different racial and class backgrounds but are both outcasts of the hierarchic society in the colonial new land, make an attempt to form a bond with each other. However, the relations head to opposite directions and end differently. The friendliness between Lina and Rebekka does not turn into a political solidarity because they are outsiders of the knowledge of “epistemic privilege” they respectively own, and their relation is greatly influenced by patriarchy and supremacy. On the contrary, Florens and Daughter Jane develop an interdependent and reciprocal relation as they bear witness to each other’s untold misery, which leads to their denial of being defined by the dominators and assimilated by supremacist notions. Their friendship bridges the racial differences, functioning as the healing power to help them seek for subjectivity and political solidarity against the same social oppression upon them. Through these two pairs of friends, Morrison illustrates the difficulty as well as the possibility of forming cross-racial sisterhood in slavery. The openness to the women of other races and the willingness of building connection with the other may not only give birth to a female community on the basis of equality and mutual benefits but also contribute to their personal growth and identity construction. Keywords: A Mercy; Epistemic privilege; Cross-racial female relation; Solidarity; Identity construction * Submission Date: 05 November 2023; Acceptance Date: 17 February 2025.

Cao, Leyi; Gao, Yajuan 122 1. Introduction Toni Morrison’s ninth novel A Mercy is about a pre-racial time in American history when slavery is at dawn. 1 Capitalism takes roots and different cultures start to be assimilated into one new American culture at that time. The protagonist is a young black woman called Florens, who is traumatized by her mother’s abandonment and then is brought back to a farm by a white landowner Jacob Vaark as an act of mercy. The loss of maternal love and its psychological impact on the young girl is a continuation of the thematic exploration –the mother-daughter relation – that Morrison repeatedly revisits in many of her writings, especially her earlier novel Beloved in which a mother kills her daughter to protect her from a life of violence in slavery. Florens’ story is supplemented and extended by the narration of other characters that are traumatized in similar ways: Lina, a Native American woman who is orphaned by the diseases brought by the colonists and then enslaved to the white; Sorrow, a hybridized orphan that becomes the victim of sexual abuse; Rebekka, a mail-order bride from England who is given up by her parents and later is widowed after the death of her husband; Jacob, an European orphan and a landowner who aspires to accumulate capital and establish a reputation as a member of the elite but fails due to his sudden death from smallpox; Scully, an indentured white man, who is sold into slavery by his father, and whose servitude, as well as Willard’s (another indentured white male servant), is always stretched and becomes endless. All these people are orphans being marginalized in a small farm, which they take as a refuge. But without the legal recognition from the colonial society, the bond among them is so fragile that when the veil of a seemingly supportive community is removed these people are left separately alone to concern about their own future. As Anderson points out, “because of slavery and indentured servitude, all social life is compartmentalized into a strict hierarchy, and every boundary is 1 Morrison, T., A Mercy. (New York: Vintage, 2008).

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 123 policed ... (thus) there is no value for human connection.”2 All these characters experience isolation and marginalization in A Mercy. In fact, Morrison indicates that one of the themes of the book is to reflect “American individuality” and to show the dangers of it.3 The haunting trauma of slavery undermines every relation among these people, especially women. It is true that the bonding among black women and its healing power is important to the survival of black women in Afro-American culture, which is exemplified in most of Morrison’s writings, such as Beloved, Sula, Paradise, Song of Solomon and Home. In A Mercy, this kind of bonding does not exist and the colored women in the farm are separated individuals heading for different directions for their own freedom.4 But it does not mean that the importance of female community is undermined. In fact, the author changes the focus and extends the scope to the bonding among women of different races in this novel. In an interview, Morrison explains that her “assumed responsibilities” as a novelist is to make writing “thinkable, possible, doable, for others –other African Americans and other women”5 so that “our emotional understanding of those marginalized women who appear in history only incidentally, as a line in a ship’s log, a slaveholder’s inventory, or a letter home” can be deepened.6 She not only gives a voice to the white women to unveil their own suffering, but also bonds them with the colored women to show the possibility of forming a supportive and nurturing female community across races. By doing so, she also explores the significant role of cross-racial female friendship in personal growth and identity construction. Morrison once points out the importance of women’s friendship as a theme of her writing in an interview. ... the world knows that women don’t choose each other’s acquaintanceship. They 2 Anderson, M. R., Spectrality in the Novels of Toni Morrison. (Knoxville: University of Tennessee Press, 2013). 3 Morrison, T., Toni Morrison discusses A Mercy: interview with Lynn Neary in 2008. Accessed September, 8, 2023. https://www.npr.org/templates/story/story.php?storyId=95961382. 4 Anderson, Spectrality in the Novels of Toni Morrison. 5 Fultz, L. P., “Introduction: The Grace and Gravity of Toni Morrison.” in Toni Morrison: Paradise, Love, A Mercy. Ed. L. P. Fultz. (London: 2013): 1-19. 6 Logan, L. M., “Thinking with Toni Morrison’s A Mercy.” Early American Literature 48.1 (2013): 193-199.

Cao, Leyi; Gao, Yajuan 124 choose men first, then women as second choice. But I have made women the focal point of books in order to find out what women’s friendships are really all about.7 Morrison brings into the novel two pairs of friends – Rebekka and Lina, and Daughter Jane and Florens. Both involve a white woman and a colored woman, and they are all outcasts of the society due to patriarchy and supremacy. But the relations head to opposite directions and end differently: one becomes hostile while the other becomes friendly. What the friends experience together shows to the readers the difficulties and the possibility of establishing an interdependent and reciprocal relationship between white and colored women, as well as how the relations either undermine or strengthen their resistance to oppression. Female bonding across races is important as it allows women to recognize and appreciate the unique struggles faced by women of different racial backgrounds so that they can work together to address them. From a feminist perspective, my reading focuses on the relation between white women and colored women – the relation between Rebekka and Lina, and Daughter Jane and Florens – in order to identify the cross-racial female bonding and the consequences of such relation. It argues that there is a way for all women, no matter what colors of their skins are, to form solidarity against hierarchy in class and race, and the friendship can contribute to the women’s personal growth and identity construction. 2. Epistemic privilege and cross-racial female friendship Sisterhood, an important feature of black women’s culture,8 is “generally understood as a nurturant, supportive feeling of attachment and loyalty to other women”.9 It has been considered as an essential interpersonal relation among colored women as they are linked by the shared experience of falling victim to the dual oppression from the white and the 7 McKay, N., “An Interview with Toni Morrison.” Contemporary Literature 24.4 (1983): 413-429. 8 Collins, P. H., “Learning from the Outsider Within: The Sociological Significance of Black Feminist Thought.” Social Problems 33.6 (1986): 14-32. 9 Dill, B. T., “Race, Class, and Gender: Prospects for an All-Inclusive Sisterhood.” In 100 Years of Women’s Suffrage: A University of Illinois Press Anthology. Ed. D. Durante. (Illinois: 2019): 177-196.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 125 patriarchal society. Because the collective identities of African-American women “have largely been formed around issues of race”,10 sisterhood hardly extends beyond racial boundaries, especially in the relation with white middle-class women who enjoy economic and political privilege that excludes black women and other women of color. The unequal relation between white women and colored women is most apparent in domestic household where white women depend on colored slave women for sustenance of status as mistress by exerting power over them.11 There is a phenomenon that in feminist movement we cry for the union and liberation of all women, while at the same time the solidarity is confined within racial groups.12 From a feminist point of view, it is important to achieve mutual understanding among women of different races, especially between white women and colored women, because their experiences are inevitably shaped by the intersection of their gender, race, ethnicity, and other social identities. Recognizing and understanding the challenges faced by women with different racial backgrounds can create solidarity and empowerment to challenge systemic inequalities. Narayan, a prominent feminist and philosopher known for her work on cultural identity, ethics, and the intersections of gender and race, emphasizes the need for a more reciprocal understanding among women from different racial groups so that the insights and contributions of all women can be valued, regardless of their geographical or cultural background.13 However, mere resolution cannot bond the two since the hostility has existed for quite a long time. Narayan, pointing out that “good-will is not enough”, uses a feminist notion of the “epistemic privilege of the oppressed”, which is rooted in standpoint theory firstly developed by Martstock and Harding in the 1980s,14 to explain 10 Dill, B. T., “Race, Class, and Gender: Prospects for an All-Inclusive Sisterhood.” 177-196. 11 Broeck, S., “Property: White Gender and Slavery.” Gender Forum: An Internet Journal for Gender Studies 14 (2006): 1-12. 12 Dill, B. T., “Race, Class, and Gender: Prospects for an All-Inclusive Sisterhood.” 177-196.; Hooks, B., Writing beyond Race: Living Theory and Practice. (New York: Routledge, 2013). 13 Narayan, U., “Sisterhood and ‘Doing Good’: Asymmetries of Western Feminist Location, Access, and Orbits of Concern.” Feminist Philosophy Quarterly 5.2 (2019): 1-26.; Narayan, U., Dislocating Cultures: Identities, Traditions, and Third World Feminism. (New York: Routledge, 1997). 14 Hundleby, C., “Feminist Standpoint Theory as A Form of Naturalist Epistemology.” (PhD diss. The

Cao, Leyi; Gao, Yajuan 126 the difficulties for women of different races, especially the more advantaged group (white women) and the more disadvantaged group (colored women) to bond.15 By recognizing the problems of dialogue between women of heterogeneous backgrounds, where women can be insiders to “specific forms of oppressive social structures – racism, sexism, compulsory heterosexuality, etc.” and outsiders to other forms of oppression, Narayan explains that “members of an oppressed group have a more immediate, subtle and critical knowledge about the nature of their oppression than people who are non-members of the oppressed group”.16 Because the oppressed insiders’ epistemic privilege is not shared by the outsiders, they tend to understand oppression in very different ways which may lead to failure of communication between members of the two groups. This idea is shared by hooks who argues that racism is a barrier to solidarity between white women and colored women, especially between black and white female relationships, as they have different cultural backgrounds and political experiences. 17 Because of race privilege, white women discriminate against and oppress black women so it is very difficult for the exploiters and the exploited to fight side by side as a solidarity. Racism is not the only factor according to hooks. Another barrier is sexism. “While sexism teaches women to be sex objects for men, it is also manifest when women who have repudiated this role feel contemptuous and superior in relation to those women who have not.”18 Male domination not only victimizes all women, but also causes hatred among women themselves. However, it does not mean that solidarity between white women and colored women is unachievable. Narayan argues that solution lies in the awareness of the “epistemic University of Western Ontario, 2001), v, https://must.primo.exlibrisgroup.com/permalink/853MUST_INST/4n97dq/cdi_proquest_journals_304771672. 15 Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” Hypatia 3.2 (1988): 31-47. 16 Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” 31-47. 17 Hooks, B., “Sisterhood: Political Solidarity between Women.” In Feminism and Community. Eds. P. A. Weiss & M. Friedman. (Philadelphia: 1995): 293-315. 18 Hooks, B., “Sisterhood: Political Solidarity between Women.” 293-315.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 127 privilege” of both insiders and outsiders. She elucidates four aspects of the concept that are prone to misinterpretation, thereby deriving implications for the solution.19 First, insiders may have the immediate knowledge of everyday life under oppression, but being oppressed does not guarantee a thorough understanding of cause and the system behind it. Second, outsiders, with great efforts and sympathy, can understand the experiences and knowledge of insiders. Third, insiders’ epistemic privilege does not rule out the possibilities that their knowledge of oppression is wrong. Four, insiders should speak for themselves whether they have epistemic privilege or not because it indicates autonomy, identity and self-respect. Therefore, cross-racial female solidarity requires mutual efforts. Outsiders should practice “methodological humility” and “methodological caution” when communicating with insiders, while insiders should realize the difficulties outsiders may have in understanding the oppression, and help them get to know the epistemic privilege of the oppressed.20 Similarly, hooks proposes mutual understanding and accommodation as the ways to overcome the barriers to achieve solidarity.21 She believes that all women, white or colored, should affirm their diversity by exploring “various ways to communicate with one another cross-culturally”,22 including genuine understanding “rooted in a rejection of dominator culture” and mutual dialogue with “a spirit of resistance ... when women dare to speak truth together” to “create the foundation for a true and lasting feminist-based sisterhood”.23 Lugones questions the use of sisterhood to describe the relationship between white women and colored women as she points out that sisterhood only exists in an egalitarian relationship.24 While there are hierarchies and inequalities among women of different 19 Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” 31-47. 20 Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” 31-47. 21 Hooks, B., Writing beyond Race: Living Theory and Practice.; Hooks, B., “Sisterhood: Political Solidarity between Women.” 293-315. 22 Hooks, B., “Sisterhood: Political Solidarity between Women.” 293-315. 23 Hooks, B., Writing beyond Race: Living Theory and Practice. 24 Lugones, M., “Sisterhood and Friendship as Feminist Models.” In The Knowledge Explosion: Generations of Feminist Scholarship. Eds. C. Kramarae. & D. Spender (New York: 1992): 406-412.

Cao, Leyi; Gao, Yajuan 128 races based on factors such as race, class and privilege, inclusive sisterhood can never be achieved. Thus, another concept to refer to the bonding among women across races is proposed – friendship. “Friendship is a kind of practical love that commits one to perceptual changes in the knowledge of other persons” and it emphasizes mutual understanding, particularities of each person, and conditional love.25 The key element is plurality, which requires women of different races to understand the racial, cultural, and political differences between them and the others so that “a multivocal communication, a dialogue among multiple selves” can be constructed.26 As it is shown that though hooks and Lugones use different terms (“sisterhood” / “friendship”) to refer to the solidarity between white and colored women, the ideas share similarities – they all underline the importance of mutual understanding and cross-racial communication. That is why this paper, overlooking the definitional differences between the two, uses “sisters” and “friends” interchangeably to mean a positive and supportive relationship in the discussion of solidarity between white women and colored women. 3. Lina and Rebekka Morrison in A Mercy describes two paired of cross-racial friends – Lina and Rebekka, and Florens and Daughter Jane – through which the author shows to the readers the effect of racism and patriarchy on women during the early slavery era, and the difficulty as well as the possibility of establishing a positive and reciprocal relation between white women and colored women. Rebekka and Lina are linked within a domestic household, which reflects a stereotypical relation between a white mistress and a colored servant within the colonial plantation where their statuses are closely connected.27 In the story, their relationship changes from hostility to friendliness, and then back to hostility. Lina is a spiritual and 25 Lugones, M., “Sisterhood and Friendship as Feminist Models.” 406-412. 26 Lugones, M., “Sisterhood and Friendship as Feminist Models.” 406-412. 27 Spatzek, S., Unruly Narrative: Private Property, Self-Making, and Toni Morrison’s A Mercy. (Boston: De Gruyter, 2022).

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 129 timeless maternal figure that represents cultural tradition and memory,28 and she resembles many earthy Native American women and Black women in literature,29 such as Eva in Beloved, and Pilate in Song of Solomon. After being purchased by Jacob, and before the arrival of Rebekka, she is the one that takes care of the household and farming. “Together they (Jacob and Lina) minded the fowl and starter stock; planted corn and vegetables. But it was she who taught him how to dry the fish they caught; to anticipate spawning and how to protect a crop from night creatures.”30 However, what she does to the farm is not appreciated by the wife in the first place. When Rebekka arrived at the farm, “the health and beauty of a young female already in charge annoyed the new wife; while the assumption of authority from the awkward Europe girl infuriated Lina.”31 This “hostility” between the two women is not personal, but the result of patriarchal and colonial influence. For Lina, Rebekka belongs to the white supremacist group (“Europe”), representing the oppression that Lina has witnessed and experienced since the arrival of colonism (“authority”). But this racial privilege owned by a European is not the only reason that causes the hostility between the women. Rebekka sees Lina as a potential sexual temptation (“health and beauty”) to her husband Jacob, and a threat to the power she is assumed to possess as the mistress of the place (“already in charge”). Lina challenges her status that is bestowed upon her by Jacob, the owner of the farm. She even points it out plainly later that one of the reasons for them being companions is that “both had to please one man” .32 Rebekka is not reading Lina and herself as independent women, but appendages of men. How Lina and Rebekka think of each other when they first meet reflects the two barriers proposed by hooks that stop white and colored women from forming a positive relationship – the racial exploitation and male supremacist values. 28 Zhou, Q., “Cultural Memory and Ethnic Identity Construction in Toni Morrison’s A Mercy.” Journal of Black Studies 50.6 (2019): 555-568. 29 Montgomery, M. L., “Got on My Traveling Shoes: Migration, Exile, and Home in Toni Morrison’s A Mercy.” Journal of Black Studies 42.4 (2011): 627-637. 30 Morrison, T., A Mercy. 31 Morrison, T., A Mercy. 32 Morrison, T., A Mercy.

Cao, Leyi; Gao, Yajuan 130 Though it does not take a long time for the two to abandon the feeling of hatred and become what they call themselves “friends”, this change is not based on mutual understanding and recognition of class and race hierarchies, but a practical need. It is a necessity for survival because of the harsh farming life, as Lina says that “the fraudulent competition (between her and Rebekka) was worth nothing on land that demanding”,33 and Rebekka states that it is “the result of the mute alliance that comes of sharing tasks”.34 It is true that gradually the two develop some sort of intimacy, and they become “company for each other and by and by discovered something much more interesting than status”.35 However, this personal intimacy does not lead to solidarity in the interest of challenging the class and race hierarchies that isolate the two women. First, politically the unequal social status between the two remains unchanged. Though Rebekka claims that Lina is an important person “whose understanding she trusted and whose judgment she valued”,36 after the death of “the best husband”,37 Rebekka believes that she is left alone in the farm with “no one except servants”,38 including Lina, the one she claims that she values as a friend. To Rebekka, the personal friendliness does not change the hierarchical differences, and she believes that she is, no matter what, superior to a slave woman. Second, Lina’s “epistemic privilege” is not shared with Rebekka. In one scenario when Rebekka takes a bath outside the house with the assistance with Lina and Florens, she is frightened by the sudden appearance of a moose at the edge of the wood. Though the animal does not show any interest in these people, Rebekka runs into Jacob’s arms immediately for comfort. What is she fearing, I (Florens) ask. Nothing, says Lina. Why then does she run to Sir? Because she can, Lina answers. Sudden a sheet of sparrows fall from the sky and settle in the trees. So many the trees seem to sprout birds, not leaves at all. Lina 33 Morrison, T., A Mercy. 34 Morrison, T., A Mercy. 35 Morrison, T., A Mercy. 36 Morrison, T., A Mercy. 37 Morrison, T., A Mercy. 38 Morrison, T., A Mercy.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 131 points. We never shape the world she says. The world shapes us.39 Rebekka’s first reaction to the potential danger shows the male supremacist value that dominates the new world, indicating “a masculinist framework that privileges males as architects of the American domain”.40 Male represents power and authority while female is mere subordinate under a man’s domination. This notion affects all women, white or colored. However, white women still possess privilege over the colored women. A white woman is protected within this masculinist framework as long as she yields to the domination. “She can” implies that women like Lina “can’t” – the colored women are vulnerable to different forms of violence and oppression. As Florens’ mother repeats four times in her chapter, “there is no protection”41 for slave women. This miserable situation of the colored women is represented by symbolic images of sparrows and trees in this scene. These women are like trees being shaped by the outside forces, and they can neither reject nor change what has happened or will happen on them like the sudden arrival and departure of sparrows. This is Lina’s “epistemic privilege”, the knowledge that she learns through her experience as a slave and is not shared with her white mistress. What's more, as Spatzek indicates, she does not dare to share because it might overstep the hierarchic bounds and challenge her white owners, especially her mistress, that she takes as a protection.42 That is why even though she sees the “folly” and the danger of a seemingly stable community in the farm because “pride alone made them (Jacob and Rebekka) think that they needed only themselves”, she chooses to be silent.43 Rebekka, on the other hand, is an outsider of this “epistemic privilege”, who does not understand the class and race hierarchies and the difference those hierarchies create between her and Lina. Third, emotionally, Rebekka does not make any effort to have empathy for Lina’s 39 Morrison, T., A Mercy. 40 Montgomery, M. L., “Got on My Traveling Shoes: Migration, Exile, and Home in Toni Morrison’s A Mercy.” 627-637. 41 Morrison, T., A Mercy. 42 Spatzek, S., Unruly Narrative: Private Property, Self-Making, and Toni Morrison’s A Mercy. 43 Morrison, T., A Mercy.

Cao, Leyi; Gao, Yajuan 132 situations. It is reflected in another scenario when Rebekka asks Lina about her past love relation, Lina reveals her mental and physical suffering: “Not good. Not good, Miss. ... I will walk behind. I will clean up after. I will not be thrashed. No.”44 But the misery does not arouse Rebekka’s empathy for her friend, but only reminds her of her own happiness – she has a loving husband, a healthy daughter, and a still living baby boy, and what’s more, “other than her mother, no one had ever struck her”.45 When explaining the ways to form solidarity between advantaged and disadvantaged groups, Narayan emphasizes the responsibility of both groups in building up positive cross-racial communication. For the oppressed, the insiders who has “epistemic privilege” about their oppression, they should “educate” the outsiders to help them understand the oppression they endure, as outsiders have no other ways to learn about the situations of insiders; but it is more important for concerned outsiders to “actively seek out and acquire such knowledge” instead of seeing the oppression as the insiders’ problem.46 Rebekka is a concerned outsider because she is the one who takes initiative to ask Lina for her past experience, and Lina, the insider, reveals this oppressive past to her. Unfortunately, this communication leads to neither mutual understanding of the two, nor political bonding against patriarchy because Rebekka fails to stand in Lina’s shoes to comprehend the oppression. Instead, it becomes a reference to highlight her own happiness and then strengthen her superior position. Thus, the communication between these two friends is actually ineffective and the hierarchies lie between them remain unchanged. It foreshadows the end of the story: the relation of the two heads back to hostility. After the death of Jacob, Rebekka recovers from smallpox that almost kills her, realizing that she is now in a dangerous situation: without a white man taking charge of the farm, she has no one (except colored people) to count on and the farm may be taken away from her. She chooses to rejoin the white community as a way to save herself, and gradually 44 Morrison, T., A Mercy. 45 Morrison, T., A Mercy. 46 Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” 31-47.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 133 becomes hostile towards the colored servants, especially Lina. Worse is how Mistress is to Lina. She requires her company on the way to church but sits her by the road in all weather because she cannot enter. Lina can no longer bathe in the river and must cultivate alone. I am never hearing how they once talk and laugh together while tending garden.47 Though what alters Rebekka’s attitude towards Lina is not explained explicitly in the story, the hypocritical sisterhood between the two may be a possible way to comprehend this change. There might be personal intimacy in the past, but it never forms a solidarity that truly units the women of two different races. Rebekka’s hostility is not a result of a sudden change, but an extension of what she always believes in – the colored and the white women are hierarchically different. In face of the crisis of being taken away the limited privileges she possesses and becoming illegal without a man, Rebekka turns to the spiritual community of white, which she has always been remotely connected with. She abandons her friend for material advantage, who she has always taken as a loyal servant and nothing more. Lina’s “epistemic privilege” that is not shared by Rebekka is one of the reasons why the two cannot reach a mutual understanding. Another reason is that Rebekka fails to undertake the attempt to meet the requirement of “methodological humility” and “methodological caution”. According to Narayan, “methodological humility” means that the outsider should always sincerely assume that she misses something and does not take it as a mistake on the part of the insider; “methodological caution” means that the outsider should sincerely avoid denigrating or dismissing the validity of the insider’s perceptions.48 These two are essential if members of two groups that are hierarchically different want to form a solidarity against supremacy and patriarchy. Unfortunately, Rebekka, as an outsider 47 Morrison, T., A Mercy. 48 Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” 31-47.

Cao, Leyi; Gao, Yajuan 134 of Lina’s “epistemic privilege”, meets neither requirement. Instead, she internalizes the supremacist belief that breeds hierarchy in class and race, placing herself in a higher social position in a female community with colored women, which makes her “a representation of white female claims to co-mastery at the New World colonial scene”.49 Thus, the so-called “friendship” comes to an end, or to be more specific, has never existed. 4. Florens and Daughter Jane Though both Lina and Rebekka claim that the other is a “friend”, they ultimately remain isolated. This failure of establishing friendship or sisterhood shows the difficulties of forming cross-racial female bonding. However, Morrison inserts another pair of friends in the story to represent the hope. Jane is the daughter of Widow Ealing who provides shelter to Florens when the girl is on her way to find the blacksmith for her dying mistress. Like Rebekka, Jane is a member of the dominator as the family owns a pasture in a Christian village. She is marginalized by the mainstream white society because of her wayward eye: “one of her eyes looks away, the other is as straight and unwavering as a she-wolf’s.”50 Florens plays the role of a domestic servant like Lina in the relation. After breakfast with Widow Ealing and her daughter, she voluntarily cleans all the bowls and spoons. But unlike Rebekka who not only falls victim to the patriarchal society but also internalizes the oppressive social beliefs, Jane refuses to be assimilated. Though her physical anomaly is read by the townspeople as a demonic sign, and she is forced to have her legs whipped by her mother constantly in the hope that her blood can be the proof of her humanity, Jane rejects the others’ interpretation on who she is. She turns to a different way to understand her body and her identity. Bowen states that Jane uses blood to relate human with animal because bleeding is a feature shared by human with animals but not with spiritual beings like a demon.51 If both animal and 49 Spatzek, S., Unruly Narrative: Private Property, Self-Making, and Toni Morrison’s A Mercy. 50 Morrison, T., A Mercy. 51 Bowen, L., “Learning to Read Ecologically: Disability, Animality, and Metaphor in Toni Morrison’s A Mercy.” ELH 88 (2021): 525-550.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 135 human being have blood running in their bodies, they must be similar as creatures. The boundary between human and animals disappear in this new interpretative mode, and in this way, a vulnerable girl is changed into a “she-wolf”. This figuration provides her the power to define herself. More importantly, it indicates harmonious relation and equality of human and animals. It is the opposite to the hierarchic belief that human beings are superior to animals, as shown in Lina’s story about a man crying “mine, mine, mine” when he faces the nature, claiming everything as his own property.52 If the boundary between human and animals can be removed, so can the racial differences. Roseneil states that “friendship is never outside the relations of power which shape the social world, neither is it ever fundamentally contained or defined by the core social institutions of family, work, and nation.”53 It shows the multiplicity of friendship and the possibility of connecting differences. Although racial hierarchy is a barrier in Rebekka and Lina’s relation, it does not hinder the development of sisterhood between Jane and Florens. Jane never looks at Florens from a superior position. She shows no interest in the color of Florens’ skin and does not care to whom she belongs. Instead, Jane takes Florens as another outcast of the society just like herself and shares with her knowledge that is essential to understand the origin of oppression and their present situation. She not only voluntarily exposes her own vulnerability – the wounds on her legs – to a black girl she meets for the first time, but also shares with her (though it is done unintentionally) the real reason of her being cast out of the community. Florens overhears the conversation between the mother and daughter at night when Jane says that the townspeople condemn her as a demon because they crave the pasture of the family. This knowledge of the white patriarchal society that prioritizes money and property enlightens Florens, as she confesses that “... it is a good thing to know ... these things”.54 What’s more, being a victim of the religious and white supremacist dogmatism, Jane also stops Florens from doing the morning praying. 52 Morrison, T., A Mercy. 53 Roseneil, S., “Foregrounding Friendship: Feminist Pasts, Feminist Futures.” In Handbook of Gender and Women’s Studies. Eds. K. Davis, M. Evans, & J. Lorber (London: 2006): 322-341. 54 Morrison, T., A Mercy.

Cao, Leyi; Gao, Yajuan 136 “At the end (of whispering the prayer) I raise my hand to touch forehead and catch Daughter Jane’s frown. She shakes her head meaning no.”55 It is the second knowledge that she passes to Florens – the hypocrisy of religion. In this colonial context, religion is manipulated and used to justify exploitation by the dominator. Under the guise that should advocate love and kindness to everyone indiscriminately, there are oppression and deception that perpetuate slavery and colonialism. Being someone with “epistemic privilege” does not imply that the oppressed know the causes of their oppression.56 “Our relation to social power produces forms of blindness just as it enables degrees of lucidity.”57 They have the first-hand experience, but they may not be able to develop these experiences into explanatory theories because most of the oppressed do not have the access to education and theory production. Florens is an example. Though she is a victim of racial oppression as she is traded in exchange for a debt, forced to be separated from her mother, and deprived of subjectivity and freedom, she does not know the political and economic reasons behind all these. She knows she is a slave; but she does not know what this identity means. This is illustrated in the incident at the beginning of her journey when she travelled with a group of indentured men and women. “They are certain their years of debt are over but the master says no. He sends them away, north, to another place, a tannery, for more years. I don’t understand why they are sad. Everyone has to work.”58 It is the first time she goes outside of the seemingly safe community in the farm. Without the knowledge of slavery, she is ignorant of the oppression and inequality that plagues people like her. The encounter with Jane and the knowledge Jane passes to her is critical, enabling her to look inwards – her identity as a slave, her experience, and the “epistemic privilege” – from a new perspective. Playing the role of an “Outsider Within” 55 Morrison, T., A Mercy. 56 Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” 31-47. 57 Mohanty, S. P., “The Epistemic Status of Cultural Identity: On Beloved and the Postcolonial Condition.” Cultural Critique 24 (1993): 41-80. 58 Morrison, T., A Mercy.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 137 according to Collins’ black feminist standpoint theory,59 Jane enlightens Florens about the complexities of power, morality, and identity in a colonial context. Compared to the so-called friendship between Rebekka and Lina in which they are still positioned in hierarchically different social classes after what they have experienced together, Jane and Florens share a relation that is more like a solidarity. It should be noted that what Jane does is more than just sharing knowledge. She also takes a risk to save the girl from a potential danger. There is a traumatic moment that has a great impact on Florens’ psychological development. A group of white villagers visit Widow Ealing’s house to scrutinize whether Jane is indeed a demon, but their attention is immediately turned to Florens when she enters the room. A great distress is caused by her “black” and being “Afric”, which is amplified by a little white girl’s “shaking and moaning”.60 This uneasiness is similar to Rebekka’s first impression on Lina which is affected by her racial traits – “... she (Rebekka) bolted the door at night and would not let the raven-haired girl with impossible skin sleep anywhere near”61. Unfortunately, though Rebekka gradually develops a personal closeness towards Lina, she remains her superior social position and makes efforts to keep it that way. However, Jane challenges this supremacist interpretation on a human being and then tries to break the hierarchical boundaries. First, she understands Florens’ “epistemic privilege” as an outcast because she has a similar experience of being marginalized, and more importantly, she witnesses Florens’ desperation in this traumatic moment – she is in the same room when Florens is identified by the white villagers as Satan’s minion. Florens is later taken to a closet and shamefully examined nakedly by the white women, which is interpreted as the moment when Florens is “dehumanized, and objectified – differentiated, not only from them (the white villagers) but from all living species”.62 It is very likely that 59 Collins, P. H., “Learning from the Outsider Within: The Sociological Significance of Black Feminist Thought.” 14-32. 60 Morrison, T., A Mercy. 61 Morrison, T., A Mercy. 62 Strehle, S., “‘I Am a Thing Apart’: Toni Morrison, A Mercy, and American Exceptionalism.” Critique: Studies in Contemporary Fiction 54.2 (2013): 109-123.

Cao, Leyi; Gao, Yajuan 138 Jane knows what happens in the closet too because she is there when all the “quarreling” and discussion concerning Florens occur afterwards. This first-hand experience of witnessing Florens being dehumanized and herself being dehumanized by the same white supremacist notion enables Jane to make an emotional and political connection with the girl of different race. Thus, Jane becomes a reliable ally. She is the only one that voices defiance by making “a small sound” while the white villagers are all “quiet” thinking about how they would deal with Florens.63 She is also the only one that takes a risk to set Florens free. A reciprocal solidarity between these two women is formed. I (Florens) say thank you and lift her hand to kiss it. She (Jane) says no, I thank you. They look at you and forget about me. She kisses my forehead then watches as I step down into the stream’s dry bed. I turn and look up at her. Are you a demon I ask her. Her wayward eye is steady. She smiles. Yes, she says. Oh, yes. Go now.64 How Florens and Jane are condemned as demons reflects a dark history of Salem witchcraft trials in the 17th century in which many innocent women were accused of witchcraft and then executed.65 The shared experience gives birth to a sense of empathy and interdependence that transcend all barriers, making the two girls each other’s saviors. Florens’ appearance removes, at least temporarily, the vicious attention on Jane; Jane helps Florens to escape, pointing out the way for her to complete her mission (finding the blacksmith) and more importantly triggering her psychological evolution – it is after the encounter with Jane that Florens’ “inside dark (that) is small, feathered and toothy”66 starts to grow. 63 Morrison, T., A Mercy. 64 Morrison, T., A Mercy. 65 Bross, K., “Florens in Salem.” Early American Literature 48.1 (2013): 183-188. 66 Morrison, T., A Mercy.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 139 5. Cross-racial female friendship and identity construction By depicting the cross-racial female relationships between characters such as Lina and Rebekka, and Florens and Daughter Jane with complexity and nuance, the author challenges and expands the readers’ understanding of race and sexuality in early America, where unique American culture and hierarchies are not yet fully formed. More importantly, the theme of interdependence suggests the fluidity of identity construction for the marginalized women. From the perspective of “epistemic privilege”, one’s cultural identity is defined through the unique insights and perspectives that marginalized individuals gain from their social positions.67 As Mohanty indicates, “identities are ways of making sense of our experiences ... that enable us to read the world in specific ways.”68 However, although insiders of “epistemic privilege” have the first-hand experience in oppression, this experience might be misunderstood or misinterpreted, and as a result, they may get lost in dominant narratives, as it is illustrated in these women’s stories. In A Mercy, Morrison proposes that the mutually beneficial friendship between women from different racial groups offer them a way to get to know the society and their positions in it in a more comprehensive way, empowering them to understand and challenge social inequalities from intersecting perspectives, which in turn aids in the establishment of their own identities. For Daughter Jane, the witness of Florens being dehumanized drives her to confront the white community by rescuing the black girl who had been abandoned by almost everyone. Although the story’s ending leaves Jane’s future unspecified, the act of salvation implies that she will probably put an end to the physical abuse she receives from her mother, and stop using scars and blood to prove her humanity. In addition, she will probably refuse to let others – the dominant white culture – to define who she is. When she confirms her identity to Florens as a “demon”, it is not the evil one condemned by the 67 Narayan, U., Dislocating Cultures: Identities, Traditions, and Third World Feminism. 68 Mohanty, S. P., “The Epistemic Status of Cultural Identity: On Beloved and the Postcolonial Condition.” 41-80.

Cao, Leyi; Gao, Yajuan 140 dominant but a free, strong and confident creature defined by herself.69 Similarly, this relationship enables Florens’ transformation from a timid black girl into a strong black woman, from an abandoned orphan to a self-loved “Precious”, and from “an ass” into “a she-lion”.70 More importantly, the shared epistemic privilege with Daughter Jane triggers Florens’ construction of her identity as one of the African-American women. It is reflected the writing of her traumatic memory on the floor of Jacob’s mansion, an embodiment of slavery culture marked by suffering and oppression. 71 By adding narratives from the perspective of an insider, who not only has the epistemic privilege of the oppressed but also is enlightened by an outsider’s insights, to the dominant racial and patriarchal sociological knowledge, Florens situates the oppression that she has experienced in relation to hegemonic social structure which finally enables her to define who she really is culturally.72 Unfortunately, the relationship between Rebekka and Lina does not bring about similar personal transformation. Rebekka continues to seek refuge within the hierarchy of white supremacy and patriarchy, willingly forsaking the opportunity to independently manage a farm and become a self-reliant woman. From Willard and Scully’s conversation at the end of the story, we know that she will probably marry again, to someone provided by the white community. Lina, on the other hand, remains in a state of servitude to her mistress as a slave, offering no resistance even when faced with Rebekka’s hostility. They are still trapped in the roles of the oppressed, unable to break free from the confinement. 6. Conclusion A Mercy depicts a world where anyone can be enslaved and victimized by different overarching social forces. It is also about a time when a women’s community, white or 69 Cao, L. Y., “A Black Woman’s Standpoint in Toni Morrison’s A Mercy.” The Journal of Macau University of Science and Technology 18.2 (2024): 65-88. 70 Morrison, T., A Mercy. 71 Zhou, Q., “Cultural Memory and Ethnic Identity Construction in Toni Morrison’s A Mercy.” 555-568. 72 Cao, L. Y., “A Black Woman’s Standpoint in Toni Morrison’s A Mercy.” 65-88.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 141 black, has not yet been formed and everyone is individualized with separate concerns for his/her fate. All four women – Lina, Rebekka, Florens and Jane – are disconnected as outcasts because of their inabilities to fight the constraints of a society that is hierarchized by sex, class and race. To seek support in order to survive this hostile society, they automatically bond with the women that they are readily close to: a white mistress/a colored domestic servant. But from the relation between Lina and Rebekka, and the relation between Florens and Jane, Morrison shows the difficulties as well as the possibility of forming such cross-racial sisterhood, and its importance in constructing one’s identity. Although Lina and Rebekka intimately call each other “friend”, their friendliness never develops into a solidarity that truly units the two women of different races. The knowledge of Lina’s “epistemic privilege” is unknown to her white mistress, who at the same time shows no interest in it. Instead, Rebekka dismisses the possibilities of bonding with another woman of different race by clinging to her superior position as the dominator in this relation and thus making Lina remain a dispensable appendage of her property so that her limited legal privileges are protected. Nothing is changed and no one is saved in this relation. On the contrary, Florens and Jane develop a relation beyond the demarcations of race and class. They both are witnesses of how each other is treated inhumanly and rejected by the dominating society, which provides them the first-hand experience of the “epistemic privilege” they respectively own. This kind of sisterhood bridges the racial differences, functioning as the healing power to help both of them seek for subjectivity as well as political solidarity against the social oppression upon them. It indicates that the openness to the women of other races and the willingness of building connection with the other may give birth to a female community on the basis of equality and mutual benefits. The study on the portrayal of cross-racial female friendships in A Mercy is significant because it offers a new perspective on the exploration of feminist topics in Morrison’s works, through which it helps foster understanding and respect between women with different racial and cultural backgrounds. What’s more, it enriches the theory of epistemic privilege by suggesting that the knowledge and insights gained from supportive and reciprocal friendship, like the one between Florens and Daughter Jane, can provide a unique

Cao, Leyi; Gao, Yajuan 142 and valuable understanding of societal issues that is not accessible through dominant cultural lenses. However, the study has its limitations. It mainly focuses on the interpersonal dynamics between white and colored women in the novel, so the discussions on the complexity of each individual female character and other female relationships are either insufficient or ignored. There is also a lack of in-depth analysis of the broader social context. Future research will delve into these aspects for a more comprehensive exploration.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 143 References [1] Morrison, T. A Mercy. New York: Vintage, 2008. [2] Anderson, M. R. Spectrality in the Novels of Toni Morrison. Knoxville: University of Tennessee Press, 2013. [3] Morrison, T. Toni Morrison discusses A Mercy: interview with Lynn Neary in 2008. Accessed September, 8, 2023. https://www.npr.org/templates/story/story.php?storyId=95961382. [4] Fultz, L. P., “Introduction: The Grace and Gravity of Toni Morrison.” in Toni Morrison: Paradise, Love, A Mercy. Ed. L. P. Fultz. (London: 2013): 1-19. [5] Logan, L. M., “Thinking with Toni Morrison’s A Mercy.” Early American Literature 48.1 (2013): 193-199. [6] McKay, N., “An Interview with Toni Morrison.” Contemporary Literature 24.4 (1983): 413-429. [7] Collins, P. H., “Learning from the Outsider Within: The Sociological Significance of Black Feminist Thought.” Social Problems 33.6 (1986): 14-32. [8] Dill, B. T., “Race, Class, and Gender: Prospects for an All-Inclusive Sisterhood.” In 100 Years of Women’s Suffrage: A University of Illinois Press Anthology. Ed. D. Durante. (Illinois: 2019): 177-196. [9] Broeck, S., “Property: White Gender and Slavery.” Gender Forum: An Internet Journal for Gender Studies 14 (2006): 1-12. [10] Hooks, B. Writing beyond Race: Living Theory and Practice. New York: Routledge, 2013. [11] Narayan, U., “Sisterhood and “Doing Good”: Asymmetries of Western Feminist Location, Access, and Orbits of Concern.” Feminist Philosophy Quarterly 5.2 (2019): 1-26. [12] Narayan, U. Dislocating Cultures: Identities, Traditions, and Third World Feminism. New York: Routledge, 1997. [13] Hundleby, C., “Feminist Standpoint Theory as A Form of Naturalist Epistemology.”

Cao, Leyi; Gao, Yajuan 144 PhD diss., The University of Western Ontario, 2001. https://must.primo.exlibrisgroup.com/permalink/853MUST_INST/4n97dq/cdi_proqust_journals_304771672. [14] Narayan, U., “Working Together across Difference: Some Considerations on Emotions and Political Practice.” Hypatia 3.2 (1988): 31-47. [15] Hooks, B., “Sisterhood: Political Solidarity between Women.” In Feminism and Community. Eds. P. A. Weiss & M. Friedman. (Philadelphia: 1995): 293-315. [16] Lugones, M., “Sisterhood and Friendship as Feminist Models.” In The Knowledge Explosion: Generations of Feminist Scholarship. Eds. C. Kramarae. & D. Spender. (New York: 1992): 406-412. [17] Spatzek, S. Unruly Narrative: Private Property, Self-Making, and Toni Morrison’s A Mercy. Boston: De Gruyter, 2022. [18] Zhou, Q., “Cultural Memory and Ethnic Identity Construction in Toni Morrison’s A Mercy.” Journal of Black Studies 50.6 (2019): 555-568. [19] Montgomery, M. L., “Got on My Traveling Shoes: Migration, Exile, and Home in Toni Morrison’s A Mercy.” Journal of Black Studies 42.4 (2011): 627-637. [20] Bowen, L., “Learning to Read Ecologically: Disability, Animality, and Metaphor in Toni Morrison’s A Mercy.” ELH 88 (2021): 525-550. [21] Roseneil, S., “Foregrounding Friendship: Feminist Pasts, Feminist Futures.” In Handbook of Gender and Women’s Studies. Eds. K. Davis, M. Evans, & J. Lorber (London: 2006): 322-341. [22] Mohanty, S. P., “The Epistemic Status of Cultural Identity: On Beloved and the Postcolonial Condition.” Cultural Critique 24 (1993): 41-80. [23] Strehle, S., “‘I Am a Thing Apart’: Toni Morrison, A Mercy, and American Exceptionalism.” Critique: Studies in Contemporary Fiction 54.2 (2013): 109-123. [24] Bross, K., “Florens in Salem.” Early American Literature 48.1 (2013): 183-188. [25] Cao, L. Y., “A Black Woman’s Standpoint in Toni Morrison’s A Mercy.” The Journal of Macau University of Science and Technology 18.2 (2024): 65-88.

Are They Friends? Cross-Racial Female Relation in Toni Morrison’s A Mercy 145 作者簡介 曹樂怡，澳門科技大學國際學院講師，教育學博士，主要研究領域為英美文學、英語教學與學習 通訊地址：澳門科技大學國際學院 通訊電郵：lycao@must.edu.mo 通訊作者簡介 高亞娟，浙江外國語學院應用外語學院副教授，研究方向為教育現象學、高校教師專業發展、高校外語教育、課堂教學改革與創新、翻譯理論與實踐 通訊電郵：yjgao@zisu.edu.mo

The Journal of Macau University of Science and Technology(Hum. & Soc. Sci.) VOL.19 NO.1 March 2025, pp.147-176 DOI: 10.58664/mustjournal.2025.03.005 147 Risk Definers and Media Coverage of GM Foods: A Comparative Analysis of the People's Daily and the New York Times* Wang, Di1; Li, Yinyun1; Mao, Zhifei2; He, Mengfan2; Hon, Chitin3; Liu, Zige3 1 Faculty of Humanity and Art, Macau University of Science and Technology, Macau, China 2 School of Journalism and Communication, Shenzhen University, Shenzhen, China 3 Faculty of Innovative Engineering, Macau University of Science and Technology, Macau, China Abstract: Literature on GM foods reports often examined article tones in general, omitting the differences in various social actors' attitudes. To explore such differences, we conducted a comparative framing analysis of news reports on genetically modified (GM) foods in the United States and China, two large markets of GM foods, to examine (1) the use of frames on GM foods; (2) who has more power in defining the risks of GM foods and (3) the tones towards GM foods of the sources in each country. We used both article and source assertion as the unit of analysis. By content analyzing 267 news articles on GM foods from 2014–2023, including 126 from the People’s Daily (PD) and 141 from the New York Times (NYT), we identified the frames, sources, and tones of each article. We also identified 1,496 source assertions to examine the tones of each source. Results showed PD used more treatment responsibility frame than NYT, while the latter used more conflict and causal responsibility frames than the former. While both publications primarily published neutral articles, NYT published more anti-GM food articles than PD. The NYT used a broader range of sources and more anti-GM food sources than PD, while PD cited more pro-GM food sources. The two newspapers both held neutral attitudes toward GM foods. By reporting both sides of the opinion on the issue, most news articles framed it as uncertain rather than having harm or benefit to human society, which may lead to public confusion about GM foods. Keywords: GM foods; Framing; Tone; Source; Comparative study * Submission Date: 10 October 2023; Acceptance Date: 17 February 2025.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 148 1. Introduction Genetically modified foods (GM foods), which are those foods “produced from or using GM organisms”,1 were first approved for commercial sale in 1994.2 Today, more than two decades later-human societies still have not reached a consensus on the biosafety of those products. Indeed, people’s concerns about the risk of GM foods appeared with the birth and development of genetic engineering.3 Although the scientific risk assessments of some GM foods (e.g., soybeans, rice, maize, and wheat) show that they are as safe as their traditional (i.e., non-GM) counterparts,4 people’s concerns about GM foods remained unchanged.5 This indicates that GM foods are more than a scientific issue. It is also a social and political issue that was formed by different actors from different interest groups. Thus, to understand the debates about the risks and benefits of GM foods, and the power relations behind such debates, we need to review not only the scientific research but also the social arguments of relevant issues. 2. Literature review 2.1 GM Foods in China and the United States In this research, we focused on the media coverage of GM foods in China and the United States (U.S.) to see how they were socially constructed and who was in charge of 1 World Health Organisation. Food, genetically modified. Last modified May1,2014. https://www.who.int/news-room/q-a-detail/FAQ-genetically-modified-foods. 2 Schauzu, M. “The concept of substantial equivalence in safety assessment of foods derived from genetically modified organisms.” AgBiotechNet 2.044 (2000): 1-4. 3 Frewer, L. J., Scholderer, J. & Bredahl, L. “Communicating about the risks and benefits of genetically modified foods: The mediating role of trust.” Risk Analysis: An International Journal 23.6 (2003): 1117-1133. 4 Domingo, J. L. “Safety assessment of GM plants: An updated review of the scientific literature.” Food and Chemical Toxicology 95 (2016): 12-18 5 Bawa, A. S. and Anilakumar, K. R. “Genetically modified foods: safety, risks and public concerns-a review.” Journal of Food Science and Technology 50.6 (2013): 1035-1046.

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 149 mediating them. These two countries differ culturally and politically, but they have similar interests in GM foods. We applied this comparative study based upon Mill’s most different systems design (MDSD), which compares countries that do not share many features in common apart from the political outcome to be explained and a few explanatory factors that are important for the outcome. This comparative logic may help us uncover the major features that drove China and the U.S., the two different countries, to embrace GM foods regardless of their risk argument. 6 According to the U.S. Food and Drug Administration (FDA), GM foods make up a very large percentage of the crops planted in the U.S. (e.g., soybeans, cotton, canola, corn). For instance, according to the FDA, genetically modified organisms (GMO) soybeans “made up 94% of all soybeans planted, and 92% of corn planted was GMO corn” in 2018.7 China, on the other hand, is more cautious with planting GM crops locally. The only two crops that China allowed for commercial production were cotton and papayas.8 The country, however, was not as cautious when importing GM foods, to the point that over 90% of maize and corn that China imported was genetically modified.9 Overall speaking, the U.S. seems to hold an open attitude toward GM food, while China is being cautiously optimistic toward GM foods. Scholars have tried to understand the reasons why the two dissimilar countries both adopted GM foods and the public reactions to the adoption in the respective countries. The majority of the research focused on consumers' perceptions of GM foods in China and the U.S.,10 while some studies compared the laws and policies governing GM foods in the two 6 Mill, J. S. “A system of logic.” In Arguing About Science. Routledge (2012): 243-267. 7 U.S. Food and Drug Administration. GMO Crops, Animal Food, and Beyond. Last modified March5,2024. https://www.fda.gov/food/agricultural-biotechnology/gmo-crops-animal-food-and-beyond. 8 Ministry of Agriculture and Rural Affairs of China. China did not allow the commercial production of main crops. Last modified March5, 2014. http://www.moa.gov.cn/ztzl/2014lhjj/2014nian/201403/t20140305_3803881.htm. 9 Lucht, J. M. “Public acceptance of plant biotechnology and GM crops.” Viruses 7.8 (2015): 4254-4281. 10 Frewer, L. J., Miles, S., & Marsh, R. “The media and genetically modified foods: evidence in support of social amplification of risk.” Risk Analysis: An International Journal 22.4 (2002): 701-711;

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 150 countries.11 Also, some researchers have studied scientists’ perceptions of GM foods.12 By examining these studies, we can find that different social actors' attitudes towards GM foods varied from time to time, struggling between the desire for commercial benefits and the concerns of health risks. One limitation of those studies is that they examined the perceptions of GM foods from different social actors in a parallel way. Few studies have examined the power relations between policymakers, the public, the GMO industry, and scientists in supporting or questioning GM foods. The arguments from those actors were not expressed as isolated political speeches; rather, they formed social debates in China, the U.S., and other countries, competing for dominance in defining GM foods and their risks. To explore such debates and the power relations behind them, we examined the very embodiment of such debates, the media coverage of GM foods in China and the U.S. Different arguments and thoughts were entangled, negotiated, and confronted in the news media, sharpening social perceptions of a controversial technology and its risk, and at the same time being influenced by the social atmosphere regarding the same topic.13 As a controversial technology, GM foods and the media coverage of them were not exceptional .14 We would like to compare and contrast the major concerns and focuses of different social actors and their attitudes toward GM foods in the media; Most importantly, we explored which stakeholders dominated the social construction of GM foods in China and the U.S., and we examined which stakeholders’ voices were marginalized in the two countries’ media. According to sociologists, whoever seizes the power of defining the risk of a technology like GM foods can take advantage of that domination by defining the 11 Wong, A. Y. T. and Chan, A. W. K.. “Genetically modified foods in China and the United States: A primer of regulation and intellectual property protection.” Food Science and Human Wellness, 5.3 (2016): 124-140. 12 ilbeck, A., Binimelis, R., Defarge, N., Steinbrecher, R., Székács, A., Wickson, F. and Wynne, B. “No scientific consensus on GMO safety.” Environmental Sciences Europe 27.1 (2015): 1-6. 13 Cottle, S. TV news, lay voices and the visualisation of environmental risks. (In Environmental risks and the media2013), 29-44. 14 Vilella-Vila, M., & Costa-Font, J. “Press media reporting effects on risk perceptions and attitudes towards genetically modified (GM) food.” The Journal of Socio-Economics 37.5 (2008): 2095-2106.

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 151 technology to satisfy their own interests.15 By examining the media coverage of GM foods in the two countries and the salience of social actors in mediating the topic, we may understand the driving forces of the GM foods market and the obstacles against it. To reach this goal, we will apply framing analysis to approach the media coverage of GM foods in China and the U.S. 2.2 Media Framing of GM Foods: Angles, Sources, and Tones According to Entman, journalists use different frames as devices to define problems, diagnose causes, suggest remedies, and make moral judgments.16 Through the usage of frames, media could shape public opinion on various social issues,17 including GM foods. Scholars have applied framing analysis to examine the social construction of GM foods in different countries.18 The literature shows that the media coverage of GM foods has varied in focused angles, tones, and sources. These three aspects can help us understand (a) the major concerns about GM foods presented by the media, (b) the media's attitudes toward GM foods, and (c) who did and did not dominate the media coverage of GM foods. In this study, we apply Semetko and Valkenburg’s research framework to examine how journalists framed GM foods in the U.S. and China. Semetko and Valkenburg found that journalists used different angles (i.e., frames) to highlight some specific angles of a social issue,19 including, but not limited to: • Economic consequences (illustrating the economic impact of a certain issue), • Conflict (portraying the conflicts between different interest groups), 15 Beck, U., Giddens, A. and Lash, S. Reflexive Modernization: Politics, Tradition and Aesthetics in the Modern Social Order. Stanford, Calif: Stanford University Press,1994. 16 Entman, R. M. “Framing: Toward clarification of a fractured paradigm.” Journal of Communication 43.4 (1993): 51-58. 17 Entman, R. M. “Framing bias: Media in the distribution of power.” Journal of Communication 57.1 (2007): 163-173. 18 Dibden, J., Gibbs, D. and Cocklin, C. “Framing GM crops as a food security solution.” Journal of Rural Studies 29 (2013): 59-70. 19 Semetko, H. A., & Valkenburg, P. M.. “Framing European politics: A content analysis of press and television news.” Journal of Communication, 50.2 (2000): 93-109.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 152 • Attribution of responsibility (discussing who shall take responsibility for fixing a problem), • Human interest (using an emotional angle or bringing a human face to the discussion of an issue), and • Morality (making moral judgments on institutions or individuals Scholars have used those frames in a flexible way to fit with the media data. For instance, some scholars used the frame of human impact (discussing the impact of certain social issues on human beings) instead of human interest as a core frame.20 Some others divided the responsibility frame into the causal responsibility frame (blaming whom for causing the problem) and the treatment responsibility frame (how to solve the problems;).21 Despite the differences in naming, the frames that were originally identified by Semetko and Valkenburg are useful for examining the different focuses and angles journalists used to depict GM foods.22 Based upon such a framework, we may examine, for example, which aspects of GM foods the media were more likely to highlight, the impact of GM foods on human health, or the profits of the GMO industry. We can also analyze the conflicts among interest groups, moral judgments of GM foods, and the attribution of responsibility concerning the industry, which were illustrated in the newspapers. In addition, by comparing the frames used by journalists in China and the U.S., we can understand the differences and similarities between the two countries' media in shaping the social reality of GM foods. Some new frames may emerge from the data, and we will examine this in a pilot study. The first step of our research, therefore, is to identify the major frames that the journalists used to frame GM foods in China and the U.S. and to answer this question: RQ1: How did the framing of GM foods differ between the media of China and the U.S.? 20 Linstrom, M., and Marais, W. “Qualitative news frame analysis: A methodology.” Communitas, 17 (2012): 21-38. 21 Wang, D. and Mao, Z. “From Risks to Catastrophes: How Chinese Newspapers Framed the Coronavirus Disease 2019 (COVID-19) in its Early Stage.” Health Risk and Society. (2021) 1-19. 22 Semetko, H. A., & Valkenburg, P. M. “Framing European politics: A content analysis of press and television news.” Journal of Communication, 50.2 (2000): 93-109.

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 153 2.3 Sources: Who Dominated the Communication of GM Foods, and Whose Voices Were Marginalized? Previous studies have shown that consumers were exposed to conflicting information about food from multiple sources.23 Similarly, in news articles, journalists may also cite multiple conflicting sources in their efforts to give a balanced report. In this study, therefore, we will examine the sources of information about GM foods under different frames to further understand the power relations between different social actors (e.g., scientists, policymakers, the public, journalists, and the GM food industry). We can systematically evaluate whose voices were louder under each topic of GM foods and, conversely, whose voices were marginalized. We will then further compare the sources that the media used in China and the U.S., examining the differences and similarities in constructing GM foods. Hence, we raise the second research question: RQ2: How did the sources used by journalists differ between the media of China and the U.S. in the coverage of GM foods? 2.4 Tones: Attitudes Toward GM Foods Scholars usually use surveys to evaluate public attitudes toward GM foods;24 however, few studies have examined the media’s attitudes toward GM foods. In this study, facilitated by framing analysis, we shall analyze the attitudes toward GM foods in three ways. First, we will take news articles as the unit of analysis and compare the tones of the media in China and the U.S. toward GM foods. After portraying a general picture of the media’s attitudes, we will then treat sources as the second research unit by examining the attitudes of different social actors in China and the U.S. toward GM foods. RQ3: How did the tone of the articles differ between China's and the U.S.'s media coverage of GM foods? 23 Rendahl, J., Korp, P., Ekström, M. P. and Berg, C.. “Adolescents’ trust in food messages and their sources.” British Food Journal, 119.12 (2017): 2712-2723. 24 Kajale, D. B. and Becker, T. C.. “Determinants of consumer support for mandatory labeling of genetically modified food in India.” British Food Journal, 115.11 (2013): 1597-1611.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 154 RQ4: How did the tone of the sources differ between China's and the U.S.'s media coverage of GM foods? 3. Method 3.1 Data This study used content analysis to examine how news frames on GM foods were developed by different sources and what assertions different sources made on GM foods, focusing on newspaper coverage of GM foods in mainland China and the U.S. from 1 January 2014 to 31 December 2023. This study focused on the elite press in the two countries since it could reflect the serious opinions about GM foods in the countries. The People's Daily was chosen to represent the Chinese elite newspapers as it not only has the largest circulation in China but also is the mouthpiece of the central government of China. The New York Times was chosen to represent American elite newspapers as it has been recognized as an authoritative source of information on issues of public policy in the U.S.25 A search was conducted in the official database of People's Daily using the keyword search string (“Zhuan Ji Yin (genetically modified)” OR “Zhuan Ji Yin Shi Wu (genetically modified foods)” OR “Ji Yin Bian Ji (gene editing)”) from January 1, 2014 to December 31, 2023. A total of 274 news articles were found. Similarly, a search was conducted with keywords “genetically modified”, “genetically modified foods”, “GM foods”, and “gene editing” on the official website of the New York Times after paying for its subscription from January 1, 2014, to December 31, 2023. A total of 460 news articles were found. After excluding non-GM foods, novels, and other unrelated articles, 126 articles from the People’s Daily and 141 articles from the New York Times were selected. In total, 267 articles were examined, featuring 1496 source assertions. 25 Friel, H., Falk, R. A., Falk, R. A. and Falk, R. “The Record of the Paper: How the New York Times Misreports US Foreign Policy.” London, New York: Verso. 2004.

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 155 3.2 Coding procedure The full text of each article was examined for coding. Images and videos were not included in the analysis. We used a combination of deductive and inductive approaches to identify frames. Based on previous literature and a pilot examination of a sample of news articles, we identified six frames: conflict, human impact, causal responsibility, treatment responsibility, economic consequences, and information. We divided the attribution of responsibility frame into the causal responsibility frame and the treatment responsibility frame, which makes it easy to distinguish who is the cause and who should provide the treatment. The human interest frame was replaced with the human impact frame as more articles talked about the impact of GM foods on people's health and environment, rather than using an emotional angle to discuss GM foods. The morality frame was included in the coding book but as no articles used this frame, it was excluded from the coding book and the analysis. Based on the pilot coding, we also identified two more frames that are commonly used in the two newspapers, the human impact frame and the information frame.26 Two coding books were revised multiple times during pilot coding before the final version was completed. One coding book used articles as a unit of analysis (Coding Book 1), while the other coding book used source assertion as a unit of analysis (Coding Book 2). For details, see Table 1. The first round of coding used a source assertion as the unit of analysis. A source was defined as a specific named person or organization. Source categories used for this study were domestic government officials, foreign government officials, domestic scientists and experts, foreign scientists and experts, the scientific community without mentioning nationality, non-governmental organizations (NGOs), GM foods manufacturers/research and development companies, media practitioners, community/ordinary citizens and farmers/distributors. Source assertions were coded as one sentence related to GM foods quoted from a specific source. The source tone referred to 26 Neuman, W.R., Neuman, R.W., Just, M.R. and Crigler, A.N. Common Knowledge: News and the Construction of Political Meaning. (Chicago, IL: University of Chicago Press,1992), 22-44.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 156 sources’ attitudes towards GM foods. The second round of coding used an article as the unit of analysis. Coders chose one core frame for each of the articles based on the major frame found in the article: • The conflict frame refers to the conflict between traditional agriculture and GM technology; the conflict between biodiversity and transgenic technology and the conflict interests between farmers and GM crop-developing companies that control the patents of GM seeds. • The human impact frame refers to the impact of GM foods on people’s health as well as the risks of transgenic plants to the natural environment. • The causal responsibility frame focuses on the parties that caused the problems of GM foods such as individuals’ lack of understanding of transgenic technology, academic fraud of scientists and institutions, illegal GM foods production, false media reports, etc. • The treatment responsibility frame focuses on how to solve the problems of GM foods such as individuals' learning knowledge of GM foods; scientific progress of GM foods; laws and regulations on GM technology; and supervision and management of illegal acts of GM foods. • The economic consequences frame focused on the economic interests of the state and enterprises on GM foods. • The information frame refers to the science popularization of transgenic technology. We calculated inter-coder reliability by double-coding a random subsample (n = 39 or 17%) of the articles. Krippendorf’s alpha ranged from .83 to 1.0 for all the variables, higher than the .80 alpha value suggested by Krippendorff. 27 27 Krippendorff, K. Content Analysis: An Introduction to its Methodology (2nd ed.). (Thousand Oaks, CA: SAGE Publications 2004).

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 157 Table 1. Coding book. Variables Codes Examples Domestic government officialsa For the People’s Daily, Chinese government officials=1，else=0. Liao Xiyuan said that the State Council approved the establishment of a major special project on genetically modified organisms in 2008, which has made great progress in the past eight years. For the New York Times, US government officials =1, else = 0. A federal judge in Hawaii has struck down a local ordinance that would have restricted or regulated the use of pesticides and genetically modified crops on the island of Kauai, saying the measure was pre-empted by state law. Foreign government officialsa For the People’s Daily, government officials other than Chinese government officials = 1, else =0. The drafter of the report on the law, Frederick Riyes, a member of the European Parliament from Belgium, stated that the law grants member states wishing to ban the cultivation of genetically modified crops autonomy, which is what many EU citizens hope for. For the New York Times, government officials other than US government officials =1, else = 0. But genetically modified rice has been spreading illegally for years in China, officials have admitted. Domestic scientists or expertsa For the People’s Daily, Chinese scientists or experts =1，else=0. “With the support of major transgenic projects, China's overall research and development of transgenic breeding has reached the international advanced level.” According to Wan Jianmin, academician of the CAE Member and chief technical engineer of the project. For the New York Times, U.S. scientists or experts =1, else = 0. “Just as many on the political right discount the broad scientific consensus that human activities contribute to global warming, many progressive advocacy groups disregard, reject or ignore the decades of scientific studies demonstrating the safety and wide-reaching benefits” of genetically engineered crops, Pamela Ronald, a professor of plant pathology at the University of California, Davis, wrote on the blog of the nonprofit Biology Fortified. Foreign scientists or expertsa For the People’s Daily, scientists or experts other than Chinese scientists or experts = 1, else =0. Belgian food and nutrition engineer Encore Snell told our reporter that there is still a large amount of animal feed containing genetically modified ingredients imported into the European Union, but in Belgium, the use of any genetically modified ingredient in animal feed is illegal. For the New York Times, scientists or experts other Unapproved GM food can be found elsewhere in China's food supply, said Jiajun Dale Wen, an energy

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 158 than US scientists or experts =1, else = 0. and environment researcher at Renmin University. Scientific community without mentioning nationalitya Mentioned=1 ， not mentioned=0 Other scientists, including two Nobel Prize winners, wrote an opinion article for the journal Science last fall titled “Standing Up for G.M.O.s.” NGOsa Mentioned=1 ， not mentioned=0 Cynthia Palmer of the American Bird Conservancy, which is a member of an E.P.A. pesticide advisory committee, said farmers who did not use the genetically modified soybean seeds were being hurt. GM foods manufacturers/research and development companiesa Mentioned=1 ， not mentioned=0 “G.M.O. acceptance is exceptionally low in Europe,” said Liam Condon, the head of Bayer’s crop science division, in an interview the day the Monsanto deal was announced. Media practitionersa Mentioned=1 ， not mentioned=0 To the industry, shifting crucial crops like corn, soybeans, cotton, and rapeseed almost entirely to genetically modified varieties in many parts of the world fulfills a genuine need. Community/ordinary citizensa Mentioned=1 ， not mentioned=0 Amy Harmon argues that scientific evidence shows that foods with genetically modified organisms are as safe as nonmodified products. GM foods farmers/distributorsa Mentioned=1 ， not mentioned=0 “I’m choosing on yield capabilities and plant characteristics more than I am on G.M.O. traits” like bug and poison resistance, he said, underscoring a crucial point: Yield is still driven by breeding plants to bring out desirable traits, as it has been for thousands of years. Source tone of each assertionb Positive=1 The European Union has authorized the import of 58 gene-altered varieties for food and feed from the rest of the world, including the United States. Negative=2 “We support Ecover’s determination to move away from using unsustainable palm oil, but would ask your company to reconsider the false solution of using ingredients derived from the new genetic engineering — synthetically modified organisms,” the groups wrote in a letter to the company. Neutral=3 But unlike the approval process required for new drugs and even many food additives like artificial sweeteners, the review process for new G.M.O. plant foods is voluntary. Article tonec Positive=1 We Need G.M.O. Wheat Negative=2 Genetically Modified Foods: Questions Remain Neutral=3 Opposed to G.M.O.s? How Much Do You Know About Them? The core frame of the Conflict=1 The fight between the competing academics is not focused on questions about the safety of genetically

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 159 articlec engineered seeds themselves. The sides are fighting mainly over the safety of herbicides used in so-called genetically modified organisms, or G.M.O., crops. Human impact=2 The Food and Drug Administration said on Friday that genetically engineered non-browning apples and bruise-resistant potatoes were as safe and nutritious as their conventional counterparts. Causal responsibility=3 However, a review of the pros and cons of G.M.O.s strongly suggests that the issue reflects a poor public understanding of the science behind them, along with a rebellion against the dominance of food and agricultural conglomerates. Treatment responsibility=4 G.M.O. labeling is already required in 64 countries, including those of the European Union; Russia; Japan; China; Australia; Brazil; and a number of countries in Africa, where despite rampant food scarcity and malnutrition, American exports that could save millions of lives have been rejected because the crops contained G.M.O.s. Economic consequences=5 The issue goes back to 2020, when President Andrés Manuel López Obrador of Mexico, in a likely bid to increase Mexican corn production and appeal to his working-class base, made a decree to phase out the import of genetically modified corn by 2024. Information=6 Plant scientists are using it to create new crops. Some researchers are even trying to use Crispr to bring species back from extinction. Note: The examples in the People’s Daily were translated from Chinese to English. a. Used in Coding Book 1 and 2; b. Used in Coding Book 2; c. Used in Coding Book 3. 4. Results We used “article” as the unit of analysis to answer RQ1-RQ3. In response to RQ1, Table 2 showed that in the People’s Daily, the information frame (n = 47, 37.3%) was the most prevalent frame, followed by the treatment responsibility frame (n = 42, 33.3%), the human impact frame (n = 17, 13.5%), the economic consequences frame (n = 11, 8.7%), the causal responsibility frame (n = 6, 4.8%), and the conflict frame (n = 3, 2.4%). In the New York Times, the information frame (n = 44, 31.2%) was also the most prevalent frame, followed by the treatment responsibility (n = 30, 21.3%), the causal responsibility (n =18, 12.8%), the conflict frame and the economic consequences (n = 17, 12.1%), and the human

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 160 impact frame (n =15, 10.6%). Chi-square tests and Fisher’s exact tests showed that the New York Times used significantly more conflict frame and causal responsibility than the People’s Daily while the latter used significantly more treatment responsibility frame than the former. Other frames were not used differently by the two newspapers. Table 2. Percentage of News Articles Featuring Specific Frames in the People’s Daily and the New York Times-related “Genetically Modified Food” Coverage, 2014-2023. Type of frame The People's Daily (%; n=126) The New York Times (%; n =141) χ2 Fisher’s exact testa Conflict 3 (2.4%) 17 (12.1%) N/A 8.99** Human impact 17 (13.5%) 15 (10.6%) .51 N/A Causal responsibility 6 (4.8%) 18 (12.8%) 5.21* N/A Treatment responsibility 42 (33.3%) 30 (21.3%) 4.91* N/A Economic consequences 11 (8.7%) 17 (12.1%) .78 N/A Information 47 (37.3%) 44 (31.2%) 1.10 N/A Note: **: p < 0.01. N/A: not applicable. a: A Fisher’s exact test was performed where the value of the cell expected was lower than 5. 4.1 Differences in Sources between the People’s Daily and the New York Times Table 3 outlines the sources used in each newspaper. The sources the People’s Daily cited most often when reporting GM foods were media practitioners (47.6%), followed by domestic government officials (38.1%), domestic scientists and experts (12.7%). Other sources were mentioned in less than 10% of the articles. The sources the New York Times cited most often when reporting GM foods were also media practitioners (80.1%), followed by domestic government officials (30.5%), community/ordinary citizens (26.4%), NGOs (24.8%), scientific community without mentioning nationality (19.9%), domestic scientists and experts (19.1%), GM foods manufacturers/research and development companies (15.7%), foreign scientists and experts (15.6%), and foreign government officials (11.3%). In response to RQ2, Chi-square tests, and Fisher’s exact tests showed that the New York

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 161 Times used significantly more sources such as foreign scientists and experts, the scientific community without mentioning nationality, NGOs, GM foods manufacturers/research and development companies, media practitioners, community/ordinary citizens, than the People’s Daily. Table 3. Percentage of GM Foods Related News Articles Featuring Specific Sources in the People’s Daily and the New York Times, 2014-2023. Types of sources The People's Daily (%) The New York Times (%) χ2 Fisher’s exact testa Domestic government officials 48(38.1%) 43 (30.5%) 1.71 N/A Foreign government officials 7(5.6%) 16 (11.3%) 2.84 N/A Domestic scientists and experts 16(12.7%) 27 (19.1%) 2.05 N/A Foreign scientists and experts 5(4.0%) 22 (15.6%) 9.91** N/A Scientific community without mentioning nationality 6 (4.8%) 28 (19.9%) 13.65*** N/A NGOs 4 (3.2%) 35 (24.8%) 25.00*** N/A GM foods manufacturers/research and development companies 0(0) 22(15.7%) N/A 21.43*** Media practitioners 60 (47.6%) 113 (80.1%) 30.85*** N/A Community/ordinary citizens 7 (5.6%) 37 (26.4%) 20.93*** N/A GM foods farmers/distributors 3 (2.4%) 26 (18.4%) N/A 17.72*** Note: **: p < 0.01. ***: p < 0.001. N/A: not applicable. a: A Fisher’s exact test was performed where the value of the cell expected was lower than 5. The percentages add up to more than 100% because the coders can choose more than one source for each article. 4.2 Differences in Article Tone As shown in Table 4, neutral articles (50.8%) were most prevalent in the People’s Daily, followed by pro-GM foods articles (30.2%) and anti-GM foods articles (19.0%). In the New York Times, neutral articles (45.4 %) were also the most prevalent, followed by anti-GM foods articles (34.0%) and pro-GM foods articles (20.6%). In response to RQ3, Chi-square tests showed that the New York Times (34.0%) were more likely to use an anti-GM foods tone than the People’s Daily (19.0%; χ2 = 13.04, p <.001).

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 162 Table 4. Percentage of GM Foods Related News Articles Featuring Specific Source Tone in People’s Daily and New York Times, 2014-2023. Article tone The People’s Daily (%) The New York Times (%) χ2 Pro-GM foods 38 (30.2%) 29 (20.6%) 3.26 Anti-GM foods 24 (19.0%) 48 (34.0%) 7.60** Neutral 64 (50.8%) 64 (45.4%) .78 Note: **: p < 0.01. 4.3 Differences in Source Tone To answer RQ4, we used “source assertion” as the unit of analysis and we examined each source’s tone in each newspaper. Table 5 shows the source tone in the People’s Daily. Post hoc analysis following a significant Chi-square test showed that domestic government officials were significantly less likely to show an anti-GM foods attitude than expected (actual count = 12, expected count = 26.3) and more likely to show a neutral attitude (actual count = 106, expected count = 92.1); domestic scientists and experts were more likely to show a pro-GM foods attitude than expected (actual count = 18, expected count = 8.3); foreign scientists and experts were significantly more likely to show an anti-GM foods attitude than expected (actual count = 5, expected count = 1.2) and less likely to show a neutral attitude (actual count = 0, expected count = 4.0); scientific community without mentioning nationality more likely to show an anti-GM foods attitude than expected (actual count = 6, expected count = 1.2). Other sources' tones are distributed evenly among the three tones. Table 5. Source Tone in the People’s Daily, 2014-2023. Types of sources Pro-GM foods Anti-GM foods Neutral χ2 Fisher’s Exact Testa Domestic government officials 42 (42.9%) 12 (19.4%) 106 (48.8%) 17.18*** N/A Foreign government officials 4 (4.1%) 3 (4.8%) 5 (2.3%) 1.35 N/A Domestic scientists and experts 18 (18.4%) 1 (1.6%) 13 (6.0%) 17.83*** N/A Foreign scientists and experts 2 (2.0%) 5 (8.1%) 0 (0) 17.24*** N/A Scientific community without 0 (0) 6 (9.7%) 1 (0.5%) N/A 15.44***

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 163 mentioning nationality NGOs 3 (3.1%) 1 (1.6%) 0 (0) N/A 6.23* GM foods manufacturers/research and development companies 0 (0) 0 (0) 0 (0) N/A N/A Media practitioners 25 (25.5%) 25 (40.3%) 82 (37.8%) 5.39 N/A Community/ordinary citizens 2 (2.0%) 4 (6.5%) 4 (1.8%) N/A N/A GM foods farmers/distributors 1 (1.0%) 4 (6.5%) 3 (1.4%) 6.74* N/A Note: *: p < 0.05. The percentages add up to more than 100% because the coders can choose more than one source for each article. N/A: not applicable. a: A Fisher’s exact test was performed where the value of the cell expected was lower than 5. Table 6 shows the source tone in the New York Times. Post hoc analysis following a significant Chi-square test showed that foreign scientists and experts were significantly more likely to show a pro-GM foods attitude than expected (actual count = 45, expected count = 29.5); the scientific community without mentioning nationality were more likely to show a pro-GM foods attitude than expected (actual count = 27, expected count = 17.2); NGOs were more likely to anti-GM foods attitude than expected (actual count = 41, expected count = 27.9); GM foods manufacturers/research and development companies were more likely to show a pro-GM foods attitude than expected (actual count = 32, expected count = 20.2), and less likely to show an anti-GM foods attitude than expected (actual count = 9, expected count = 21.4); media practitioners were more likely to show a neutral attitude than expected (actual count = 181, expected count = 154.2); Community/ordinary citizens were less likely to show a positive attitude than expected (actual count = 13, expected count = 25.1) and more likely to show a negative attitude than expected (actual count = 44, expected count = 26.4). Other sources’ tones are distributed evenly among the three tones. Table 6. Source Tone in the New York Times, 2014-2023. Types of sources Pro-GM foods Anti-GM foods Neutral χ2 Fisher’s Exact Testa Domestic government officials 44 (11.5%) 54 (13.3%) 34 (10.3%) 1.67 N/A Foreign government officials 3 (0.8%) 5 (1.2%) 5 (1.5%) N/A .93 Domestic scientists and experts 18 (4.7%) 19 (4.7%) 13 (3.9%) .31 N/A Foreign scientists and experts 45 (11.7%) 23 (5.7%) 18 (5.5%) 13.42** N/A

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 164 Scientific community without mentioning nationality 27 (7.0%) 12 (3.0%) 11 (3.3%) 9.06* N/A NGOs 19 (4.9%) 41 (10.1%) 17 (5.2%) 10.43** N/A GM foods manufacturers/research and development companies 23 (8.3%) 9 (2.2%) 18 (5.5%) 14.77*** N/A Media practitioners 168 (43.8%) 174 (43.0%) 181 (54.8%) 12.42** N/A Community/ordinary citizens 13 (3.4%) 44 (10.9%) 16 (4.8%) 20.23*** N/A GM foods farmers/distributors 20 (5.2%) 29 (7.2%) 17 (5.2%) 1.82 N/A Note: *: p < 0.05; **: p < 0.01; ***: p < 0.001. The percentages add up to more than 100% because the coders can choose more than one source for each article. N/A: not applicable. a: A Fisher’s exact test was performed where the value of the cell expected was lower than 5. 5. Discussion This study showed the differences in the frames, sources, and tones used in news coverage of GM foods in the People’s Daily and the New York Times. By examining the different use of frames, sources, and tones, we found the different perspectives the two papers used to report GM foods, who has more power in defining the risks of GM foods, and the attitude of each newspaper towards GM foods. 5.1 Frames: Different perspectives on GM foods In terms of the frames used by the two papers, we found while the treatment responsibility frame and the information frame were both the most frequently used frames by the two newspapers, the New York Times used the conflict frame and causal responsibility frame significantly more than the People’s Daily, while the People’s Daily used the treatment responsibility frame and information frame more than the New York Times. The information frame was used most often by both papers. As GM foods are a scientific topic as well as a controversial topic, both papers published a substantial number of articles on the popularisation of transgenic technology and GM foods. Both papers used this frame to tell readers what GM foods are and the pros and cons of GM foods. The treatment responsibility frame focused on the solutions to the problems of GM

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 165 foods. While both the People’s Daily and the New York Times used this frame heavily, the former used it significantly more than the latter. The People’s Daily used the treatment responsibility frame to describe relevant laws on GM technology formulated by the government and the punishment of illegal and criminal behaviors related to GM foods. For example, the Chinese Ministry of Agriculture claimed that they will strengthen supervision over the research, development, and production of GM agricultural products, including the disposal of illegally imported GM foods following the law. As China has passed laws on the mandatory labeling of GM foods in Article 28 of the Regulations on the Safety Management of Agricultural Genetically Modified Organisms as early as 2001,28 labeling GM foods was not mentioned heavily in the treatment solution frame in our study period. In contrast to China, the U.S. did not enforce a mandatory regulation on labeling GM foods until 2020. In 2023, the U.S. National Bioengineered Food Disclosure Standard required food manufacturers, importers, and certain retailers to disclose information about whether foods are GM or contain GM ingredients. The implementation of this standard took several years: the standard was passed in 2016, implemented in 2021, and finally mandated in 2022.29Accordingly, in our study period, the New York Times often reported the success or failure to pass laws and regulations that would require the labeling of foods containing genetically engineered ingredients using the treatment responsibility frame. Other reports using this frame showed that some food companies labeled their products “G.M.O.-free” to appeal to consumers. The New York Times used the conflict frame significantly more than the People's Daily. Using the conflict frame, the New York Times often pitted consumers and industry advocates against each other. For example, the New York Times reported that the Consumers Union believed that GM food developers were using a technical loophole so that some genetically engineered crops and organisms could escape regulation. The 28 Zhu, R. Regulations on the Safety Management of Agricultural Genetically Modified Organisms. Last modified May23,2001. http://www.gov.cn/gongbao/content/2001/content_60893.htm?ivk_sa=1024320u%20. 29 U.S. Department of Agriculture. BE Frequently Asked Questions-General. Last modified March22,2022. https://www.ams.usda.gov/rules-regulations/be/faq/general.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 166 People’s Daily only had three articles using the conflict frame. One article entitled “Whether You Should Eat GM Foods”, was about the questions of the public and the media raised to the scientists. In this article, science experts enumerated the many benefits of GM products and the safety thereof, while the public and the media seriously doubted it: they thought that GM products are unnatural and dangerous, even speculating that there are interest groups and conspiracies behind GM foods. The People’s Daily held a neutral tone towards GM foods in this article. In terms of the human impact frame, the two newspapers did not differ significantly. By using the human impact frame, the New York Times reported the impacts of GM technology on the natural environment and human health, as well as the risks of super antibody plants caused by GM crops. For example, the use of certain herbicides has no impact on GM crops but has a significant impact on conventional crops. The causal responsibility frame focused on the parties that caused the problems of GM foods. The New York Times used the causal responsibility frame significantly more than the People's Daily. The New York Times used this frame to describe the illegal production by businessmen. For example, the New York Times reported that GM rice has been spreading illegally for years in China. One article also criticized 17 European countries' ban on the cultivation of genetically modified crops, claiming that scientists would not spend years developing GM crops knowing that there is a great chance that they will be outlawed by governments. Another article reported that the consumers who sought out the “all natural” label were misled by the label. In fact, “all natural” does not exclude GM organisms, hormones, pesticides, and artificial ingredients. The economic consequences frame refers to articles that focus on the economic interests of the state and enterprises on GM foods. The two newspapers did not differ significantly in the use of this frame. Using an economic consequences frame, the New York Times discussed the interests of GM-crop developing companies such as the patenting of seeds. It also reported that nearly 20 million farmers around the world invested in GM seeds in the last two decades because farmers are seeing better harvests. Articles in the People's Daily using an economic consequences frame reported the trade agreement on GM

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 167 foods between different countries, such as Canada, China, the U.S., and those in the European Union. One article mentioned that GM soybean seeds are being controlled by multinational companies and that countries such as Argentina tried to independently develop their soybean seeds. 5.2 Diversity in Source and Source Tone: Who Has More Power in Defining the Risks Of GM Foods News sources serve as the agenda builders of newspapers (Berkowitz, 1987). When journalists cite multiple sources in one news article, they may be engaging in cross-checking to include more divergent and balanced views. By doing so, they may also avoid the trap of relying on convenient sources, such as privileged parties or press releases.30 In the People’s Daily, fewer sources were cited, with media practitioners and government officials as the dominant sources. Wang et al. argued that the People’s Daily, as the official party publication, is primarily concerned with interpreting the world from the perspective of the Communist Party of China.31 The reliance on government officials as dominant sources served this function. In comparison, the New York Times exhibited more diverse news sources. Even though media practitioners were also dominant sources, they cited many other sources. These included the scientific community without mentioning nationality, NGOs, GM foods manufacturers/research and development companies, community/ordinary citizens, and GM foods farmers/distributors. These sources may have provided more context, information, and expertise for the discussion of GM foods. They may also reflect the opinions of different interest groups. Those who manufacture and conduct research and development on GM foods may have more vested interests in supporting GM foods, whereas ordinary citizens may not support GM foods to avoid the risks. As in the previous findings, it is evident that the New York Times also 30 Tiffen, R., Jones, P. K., Rowe, D., Aalberg, T., Coen, S., Curran, J., Hayashi, K., Iyengar, S., Mazzoleni, G., Papathanassopoulos, S., Rojas, H. and Soroka, S. “Sources In The News.” Journalism Studies (2013): 1-18. 31 Wang, H., Sparks, C., and Huang, Y. “Measuring differences in the Chinese press: A study of People’s Daily and Southern Metropolitan Daily.” Global Media and China 3.3 (2018): 125-140.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 168 favored government sources over other types. However, in comparison to the People’s Daily, it had a greater range of source diversity. Although citing diverse sources does not necessarily mean citing different views, in our study, this is the case. In comparing the two newspapers, we found that sources in the New York Times were not only divergent, but also their views were evenly distributed among three tones, including pro-GM foods, anti-GM foods, and neutral. In contrast, the sources in the People’s Daily were not limited in diversity, but the views were also mainly leaning toward pro-GM foods. Though both the U.S. and Chinese governments made it legal to produce and distribute certain GM foods, the New York Times cited more anti-GM food sources than the People's Daily. Also, we discovered an interesting finding that different players served distinct roles at the two newspapers. In the People’s Daily, the pro-GM foods parties were Chinese government officials, scientists, and experts. However, in the NYT, they were foreign scientists and experts. The anti-GM food parties in the New York Times were media practitioners, while no source was significantly more likely to show an anti-GM foods tone in the People’s Daily. Sources that were significantly more likely to hold neutral tones toward GM foods in the New York Times were foreign government officials and GM foods farmers/distributors. 5.3 Article Tone: More Neutral Reports Than Pro- and Anti-GM Foods Reports The risk of GM foods is different from other food safety issues which brought about only negative consequences. GM foods are a combination of risk and opportunity. They may have potential risks, but as biotechnology products, their emergence also offers benefits and opportunities. They are conducive to national science and technology development in solving grain production problems and reducing pesticide use.32 Therefore, when the media in China and the U.S. reported on GM foods, they preferred to hold neutral attitudes (about half of the articles in both the People’s Daily and the New York Times used 32 Gurrs, J, R. “Engineering plants with increased disease resistance: What are we going to express?”. Trends Biotechnol, 23 (2005): 275-282.

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 169 a neutral tone), citing both sides of the opinion from news sources rather than showing clear support or opposition. By reporting both sides of the opinion on the issue, most news articles framed it as uncertain rather than having harm or benefit to human society. Corbett and Durfee’s experiment showed that news stories that framed global warming as controversy increased readers’ perceptions of its uncertainty.33 Similarly, by reporting GM foods in a neutral tone and citing two-sided opinions, the two papers may lead to public confusion about GM foods. 6. Conclusion Following the logic of Most Different System Design, we aimed to explain the similar “outcome”, or phenomenon, of the two very different societies in embracing GM foods by examining the few similarities in their media coverage of them. We found that, besides multiple differences, the absolute dominance of the two governments in defining GM foods and their risks or benefits might explain the adoption of these controversial technologies in the two societies. The Chinese government's open attitude was “cautiously optimistic”, with lots of neutral opinions and fewer supportive expressions. The U.S. is ambiguous, with a mixture of supports, oppositions, and neutral opinions of GM foods. Their voices were louder than the experts, the ordinary citizens, the farmers, and other interest groups of GM foods. Such dominance in discourse, we argue, is correspondingly reflecting the dominance of the two governments in the governance of GM foods. Indeed, both newspapers were not reluctant to report the negative opinions from different interest groups, yet the two countries still launched pro-GM food policies. This means the power relations behind the GM food policies were up-to-bottom instead of bottom-up—the two governments launched the policies based on their determinations and judgments, instead of responding to the public will. This is especially interesting given that the U.S. is a liberal government. Indeed, regarding GM foods policies, the U.S. government was acting as a 33 Corbett, J. B., and Durfee, J. L. “Testing public (un) certainty of science: Media representations of global warming.” Science Communication 26.2 (2004): 129-151.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 170 big government just as China is. In general, the New York Times seemed to hold an open attitude toward the treatment of GM foods. The People’s Daily, however, seemed cautiously optimistic. The media framing of GM foods in China was consistent with implemented policies. This finding was understandable since the Chinese media reflected the ideology of policymakers. Interestingly, we found a separation between the policies in the U.S. and the attitudes of social discourses over GM foods. U.S. society was far from reaching a social consensus on GM foods. The New York Times, as a newspaper in a liberal society, reflected the conflicts and questions of different social actors. The scientists and the manufacturers supported the development of GM foods, while the NGOs and ordinary citizens expressed their suspicions about the products. Overall, the American media maintained a neutral to negative position about GM foods. However, the policies and development of the GM foods industry in the U.S. indicated that the policymakers stood firmly with scientists and manufacturers, embracing GM foods regardless of potential controversies. Apart from the above findings, another contribution of the current study is in the methodology part. The previous literature on media content analysis usually focused on articles as a unit analysis. Thus, previous researchers can only examine the tone of each article. As different sources may have divergent attitudes toward an issue, this method of coding cannot reveal the divergent opinions of sources. For analysis, this study used both articles as a unit and source assertions as a unit. By doing so, we can reveal each article’s attitude and each source’s attitude toward GM foods. The limitation of our study is that we focused only on the social construction of GM foods through traditional media. Future studies could examine the social construction of GM foods related to social media. These studies may tell us, for example, whether government officials and scientists still play leading roles in constructing GM foods on social media. They may also reveal whether the tones and concerns about GM foods on social media are different from those found in traditional media. Another limitation is that we only examined the sources of press media, which can partly reflect the power dynamics regarding the social construction of GM foods, but not all of them. We did not include other

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 171 influential sources that did not appear in the media in our study, like the political talks in a congress hall, a scientific discussion in a laboratory, or the talks between two farmers about GM foods. Future studies could include more sources in different platforms to gain a comprehensive picture of the issue. Finally, by including articles in both Chinese and English, the potential language biases could be introduced.

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 172 References [1] World Health Organisation. Food, genetically modified. Last modified May 1, 2014. https://www.who.int/news-room/q-a-detail/FAQ-genetically-modified-foods. [2] Schauzu, M. “The concept of substantial equivalence in safety assessment of foods derived from genetically modified organisms.” AgBiotechNet 2.044 (2000): 1-4. [3] Frewer, L. J., Scholderer, J. & Bredahl, L. “Communicating about the risks and benefits of genetically modified foods: The mediating role of trust.” Risk Analysis: An International Journal 23.6 (2003): 1117-1133. [4] Domingo, J. L. “Safety assessment of GM plants: An updated review of the scientific literature.” Food and Chemical Toxicology 95 (2016): 12-18. [5] Bawa, A. S. and Anilakumar, K. R. “Genetically modified foods: safety, risks and public concerns-a review.” Journal of Food Science and Technology 50.6 (2013): 1035-1046. [6] Mill, J. S. “A system of logic.” In Arguing About Science. Routledge (2012): 243-267. [7] U.S. Food and Drug Administration. GMO Crops, Animal Food, and Beyond. Last modified March5,2024. https://www.fda.gov/food/agricultural-biotechnology/gmo-crops-animal-food-and-beyond. [8] Ministry of Agriculture and Rural Affairs of China. China did not allow the commercial production of main crops. Last modified March 5, 2014. http://www.moa.gov.cn/ztzl/2014lhjj/2014nian/201403/t20140305_3803881.htm. [9] Lucht, J. M. “Public acceptance of plant biotechnology and GM crops.” Viruses 7.8 (2015): 4254-4281. [10] Frewer, L. J., Miles, S., & Marsh, R. “The media and genetically modified foods: evidence in support of social amplification of risk.” Risk Analysis: An International Journal 22.4 (2002): 701-711. [11] Wong, A. Y. T. and Chan, A. W. K.. “Genetically modified foods in China and the United States: A primer of regulation and intellectual property protection.” Food Science and Human Wellness 5.3 (2016): 124-140.

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 173 [12] Hilbeck, A., Binimelis, R., Defarge, N., Steinbrecher, R., Székács, A., Wickson, F. and Wynne, B. “No scientific consensus on GMO safety.” Environmental Sciences Europe 27.1 (2015): 1-6. [13] Cottle, S. TV news, lay voices and the visualisation of environmental risks. (In Environmental risks and the media2013), 29-44. [14] Vilella-Vila, M., & Costa-Font, J. “Press media reporting effects on risk perceptions and attitudes towards genetically modified (GM) food.” The Journal of Socio-Economics 37.5 (2008): 2095-2106. [15] Beck, U., Giddens, A. and Lash, S. Reflexive Modernization: Politics, Tradition and Aesthetics in the Modern Social Order. Stanford, Calif: Stanford University Press,1994. [16] Entman, R. M. “Framing: Toward clarification of a fractured paradigm.” Journal of Communication 43.4 (1993): 51-58. [17] Entman, R. M. “Framing bias: Media in the distribution of power.” Journal of Communication 57.1 (2007): 163-173. [18] Dibden, J., Gibbs, D. and Cocklin, C. “Framing GM crops as a food security solution.” Journal of Rural Studies 29 (2013): 59-70. [19] Semetko, H. A., & Valkenburg, P. M.. “Framing European politics: A content analysis of press and television news.” Journal of Communication 50.2 (2000): 93-109. [20] Linstrom, M., and Marais, W. “Qualitative news frame analysis: A methodology.” Communitas 17 (2012): 21-38. [21] Wang, D. and Mao, Z. “From Risks to Catastrophes: How Chinese Newspapers Framed the Coronavirus Disease 2019 (COVID-19) in its Early Stage.” Health Risk and Society (2021): 1-19. [22] Semetko, H. A., & Valkenburg, P. M. “Framing European politics: A content analysis of press and television news.” Journal of Communication 50.2 (2000): 93-109. [23] Rendahl, J., Korp, P., Ekström, M. P. and Berg, C.. “Adolescents’ trust in food messages and their sources.” British Food Journal 119.12 (2017): 2712-2723. [24] Kajale, D. B. and Becker, T. C.. “Determinants of consumer support for mandatory

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 174 labeling of genetically modified food in India.” British Food Journal 115.11 (2013): 1597-1611. [25] Friel, H., Falk, R. A., Falk, R. A. and Falk, R. “The Record of the Paper: How the New York Times Misreports US Foreign Policy.” London, New York: Verso. 2004. [26] Neuman, W.R., Neuman, R.W., Just, M.R. and Crigler, A.N. Common Knowledge: News and the Construction of Political Meaning. (Chicago, IL: University of Chicago Press,1992), 22-44. [27] Krippendorff, K. Content Analysis: An Introduction to its Methodology (2nd ed.). (Thousand Oaks, CA: SAGE Publications 2004). [28] Zhu, R. Regulations on the Safety Management of Agricultural Genetically Modified Organisms. Last modified May23,2001. http://www.gov.cn/gongbao/content/2001/content_60893.htm?ivk_sa=1024320u%20. [29] U.S. Department of Agriculture. BE Frequently Asked Questions-General. Last modified March22,2022. https://www.ams.usda.gov/rules-regulations/be/faq/general. [30] Tiffen, R., Jones, P. K., Rowe, D., Aalberg, T., Coen, S., Curran, J., Hayashi, K., Iyengar, S., Mazzoleni, G., Papathanassopoulos, S., Rojas, H. and Soroka, S. “Sources In The News.” Journalism Studies (2013): 1-18. [31] Wang, H., Sparks, C., and Huang, Y. “Measuring differences in the Chinese press: A study of People’s Daily and Southern Metropolitan Daily.” Global Media and China 3.3 (2018): 125-140. [32] Gurrs, J, R. “Engineering plants with increased disease resistance: What are we going to express?”. Trends Biotechnol 23 (2005): 275-282. [33] Corbett, J. B., and Durfee, J. L. “Testing public (un) certainty of science: Media representations of global warming.” Science Communication 26.2 (2004): 129-151.

Risk definers and media coverage of GM foods–a comparative analysis of the People's Daily and the New York Times 175 Authors Information Wang Di, Ph.D., is an associate professor at the Faculty of Humanities and Arts of Macau University of Science and Technology. Her main research interests are health communication, new media communication, and media effects. Address: Room R322, Macau University of Science and Technology, Weilong Road, Taipa, Macau, China. Email: dwang@must.edu.mo. Li Yinyun is a Ph.D. student at the Faculty of Humanities and Arts of Macau University of Science and Technology. Her main research interests are new media communication and media effects. Address: Room R306, Macau University of Science and Technology, Weilong Road, Taipa, Macau, China Email: stella.liyy@outlook.com Mao Zhifei, Ph.D., is an associate professor at the School of Journalism and Communication, Shenzhen University. Her main research interests are risk communication, environmental communication, and new media study. Address: 3688 Nanhai Avenue, Nanshan District, Shenzhen, Guangdong Province, China Email: feimyks@gmail.com He Mengfan is a M.A. student at the School of Journalism and Communication of Shenzhen University. Her main research interests are new media communication and media effects. Address: 3688 Nanhai Avenue, Nanshan District, Shenzhen, Guangdong Province, China Email: hoxiyayumeto@163.com

Wang, Di; Li, Yinyun; Mao, Zhifei; He, Mengfan; Hon, Chitin; Liu, Zige 176 Hon Chitin, Ph.D., is a professor at the Faculty of Innovative Engineering of Macau University of Science and Technology. His main research interests are epidemic prediction and early warning, and social network analysis. Address: Room A418, Macau University of Science and Technology, Weilong Road, Taipa, Macau, China Email: cthon@must.edu.mo Liu Zige is a Ph.D. student of Engineering Science at the Faculty of Innovative Engineering of Macau University of Science and Technology. His main research interests are machine learning and data analysis Address: Room A412, Macau University of Science and Technology, Weilong Road, Taipa, Macau, China Email: zgliu@must.edu.mo Corresponding author Mao Zhifei, Ph.D. Assistant Professor. Her research interests are risk communication, environmental communication, and new media communication. E-mail: feimyks@163.com Tel: 0755-86581096 Funding Information This research was funded by the Macau University of Science and Technology Faculty Research Grant (FRG-22-039-FA) to DW, the Science and Technology Development Fund of Macau SAR (005/2022/ALC) to DW, and the Major Project of Guangzhou National Laboratory (GZNL2024A01004) to DW.

The Journal of Macau University of Science and Technology (Hum. & Soc. Sci.) VOL.19 NO.1 March 2025, pp.177-213 DOI: 10.58664/mustjournal.2025.03.006 177 AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends* Chen, Renzhang1; Zhao, Haixia2 1 Modern Educational Technology Center of Zhuhai Campus, Jinan University, 2 University International College, Macau University of Science and Technology Abstract: The meteoric rise of ChatGPT has ignited the so-called “chatbot tsunami,” propelling AI chatbots into the vanguard of technology research and development. This study conducts a comprehensive review of AI chatbot research from 2000 to 2023, harnessing bibliometric and content analysis methods facilitated by VOSviewer software. Quantitative and visual analyses of 1,236 relevant documents retrieved from the Web of Science in May 2023 unveil AI chatbots as an interdisciplinary research domain. The most influential authors are identified from three distinct fields: human-computer interaction, human-related, and computer-related research. Key research foci encompass AI technologies, digital health and education, conversational AI, COVID-19 applications, customer service, and sentiment analysis. Building upon the bibliometric findings, we delineate the developmental trajectory of AI chatbot research and propose a conceptual framework to guide future endeavors. Furthermore, we outline prospective research avenues for AI chatbots, including advancements in AI capabilities, optimization of human-computer interaction design, and fostering interdisciplinary synergies. This comprehensive review elucidates the AI chatbot research landscape, offering valuable insights to steer both academic pursuits and industrial applications in this burgeoning field. Keywords: Chatbots; Human-Computer Interaction; Conversational agents; Deep Learning; Bibliometric analysis * Submission Date: 28 August 2024; Acceptance Date: 17 February 2025.

Chen, Renzhang; Zhao, Haixia 178 1. Introduction With the exponential growth of Internet data and advancements in artificial intelligence technology, communication methods have undergone significant transformations,1 particularly in the realm of human-computer interaction (HCI) systems.2 Chatbots, which offer effective HCI, have garnered considerable attention in both academia and industry.3 The term “chatbot” is derived from “chatterbot” and was initially introduced by Michael Mauldin in 1997 to describe a conversational robot for human interaction.4 According to the Lexico Dictionaries,5 a chatbot is defined as “a computer program designed to simulate a conversation with a human user, particularly over the Internet.” Currently, two primary types of chatbots exist.6 The first type, referred to as rule-based bots, operates on predefined scripts and utilizes keyword matching to generate responses.7 These systems resemble expert systems from the 1980s, which relied on a fixed set of rules to address specific queries. Rule-based bots are limited to scenarios they have been programmed for, making them less adaptable in conversation. The second type, known as AI chatbots or artificially intelligent conversational agents,8 employs natural language processing (NLP) techniques to facilitate more dynamic interactions. Unlike their rule- 1 Allison, DeeAnn, “Chatbots in the Library: Is It Time?,” Library Hi Tech 30, no. 1 (2012): 95-107. 2 Mocanu, Bogdan-Costel et al, “Odin IVR-Interactive Solution for Emergency Calls Handling,” Applied Sciences 12, no. 21 (2022): 10844. 3 Miklosik, Andrej, Nina Evans, and Athar Mahmood Ahmed Qureshi, “The Use of Chatbots in Digital Business Transformation: A Systematic Literature Review,” IEEE Access 9 (2021): 106530-106539. 4 Deryugina, O. V, “Chatterbots,” Scientific and Technical Information Processing 37, no. 2 (2010): 143-147. 5 “Chatbot: Definition of Chatbot in English,” Lexico Dictionaries. 2022. Accessed November 30, 2022. https://www.lexico.com/en/definition/chatbot. 6 Nirala, Krishna Kumar, Nikhil Kumar Singh, and Vinay Shivshanker Purani, “A Survey on Providing Customer and Public Administration-Based Services Using AI: Chatbot,” Multimedia Tools and Applications 81, no. 16 (2022): 22215-22246. 7 Al-Rfou, Rami et al, “Conversational Contextual Cues: The Case of Personalization and History for Response Ranking,” arXiv preprint arXiv:1606.00372, 2016. 8 Moussawi, Sara, Marios Koufaris, and Raquel Benbunan-Fich, “How Perceptions of Intelligence and Anthropomorphism Affect Adoption of Personal Intelligent Agents,” Electronic Markets 31, no. 2 (2020): 343-364.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 179 based counterparts, AI chatbots can learn from user interactions, enabling them to understand context and provide more relevant responses. This flexibility allows AI chatbots to engage in more fluid conversations and handle a wider range of inquiries. AI chatbot has witnessed extensive utilization, with notable examples including Apple Siri, IBM Watson, Google Assistant, Microsoft Cortana, and Amazon Alexa.9 Furthermore, AI chatbot has found applications in diverse domains such as education and training,10 business and e-commerce, 11 as well as entertainment, 12 among others. In 2019, researchers began referring to this phenomenon as the “chatbot tsunami”,13 signifying its widespread impact. Subsequently, on November 30, 2022, OpenAI unveiled ChatGPT, an AI chatbot product that further intensified the ongoing “chatbot tsunami.” ChatGPT swiftly garnered substantial attention, amassing over 1 million registered users within its first week of global launch, and within a mere two months, it amassed a user base of 100 million monthly active users. 14 According to market research reports, the AI chatbot market demonstrates significant potential. 15 Presently, researchers have an unprecedented 9 Park, Namkee et al, “Use of Offensive Language in Human-Artificial Intelligence Chatbot Interaction: The Effects of Ethical Ideology, Social Competence, and Perceived Human likeness,” Computers in Human Behavior 121 (2021): 106795. 10 Essel, Harry Barton et al, “The Impact of a Virtual Teaching Assistant (Chatbot) on Students’ Learning in Ghanaian Higher Education,” International Journal of Educational Technology in Higher Education 19, no. 1 (2022): 57; Guo, Kai, Jian Wang, and Samuel Kai Wah Chu, “Using Chatbots to Scaffold EFL Students’ Argumentative Writing,” Assessing Writing 54 (2022): 100666. 11 Jiang, Hua, Yang Cheng, Jeongwon Yang, and Shanbing Gao, “AI-Powered Chatbot Communication with Customers: Dialogic Interactions, Satisfaction, Engagement, and Customer Behavior,” Computers in Human Behavior 134 (2022): 107329. https://doi.org/10.1016/j.chb.2022.107329. 12 Garcia-Mendez, Silvia et al, “Entertainment Chatbot for the Digital Inclusion of Elderly People without Abstraction Capabilities,” IEEE Access 9 (2021): 75878-75891. 13 Grudin, Jonathan, and Richard Jacques, “Chatbots, Humbots, and the Quest for Artificial General Intelligence,” In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems (2019). 14 “What Is ChatGPT and Why Does It Matter? Here's What You Need to Know,” ZDNet. December 7, 2022. Accessed December 9, 2022. https://www.zdnet.com/article/what-is-chatgpt-and-why-does-it-matter-heres-what-you-need-to-know/;“ChatGPT Has Gained 100 Million Users in Just 2 Months after Being Released,” TechJuice. February 2, 2023. Accessed February 9, 2023. https://www.techjuice.pk/chatgpt-has-gained-100-million-users-in-just-2-months-after-being-released/. 15 “Global Chatbot Market - Growth, Trends, Covid-19 Impact, and Forecasts (2022 - 2027),” Research and Markets. 2022. Accessed September 4, 2022. https://www.researchandmarkets.com/reports/4622740/global-chatbot-market-growth-trends-covid-19.

Chen, Renzhang; Zhao, Haixia 180 opportunity to conduct large-scale real-world studies on AI chatbots.16 However, there is a scarcity of data-driven quantitative studies that comprehensively address the current state and future trajectory of AI chatbot research as a whole and provide an objective evaluation alongside visual representations of the field’s scientific knowledge.17 Existing literature review studies primarily focus on the application of chatbots within specific domains or utilize narrative reviews and systematic literature reviews to evaluate and examine existing literature. 18 To holistically review research progress in the field of AI chatbots and anticipate future development directions, this paper employs bibliometric methods such as co-citation analysis, co-word analysis, and cluster analysis to quantitatively analyze and visually represent the existing research on AI chatbots based on data gathered from the Web of Science literature survey. By quantitatively consolidating the knowledge base and identifying research hotspots in AI chatbot research, our aim is to comprehend the overall development status, uncover research frontiers, and discern development trends. We believe that this endeavor will assist researchers and practitioners in conducting more effective theoretical research and practical applications in the domain of AI chatbots. This study aims to address several key research questions within the field of AI chatbots: What are the prevailing research hotspots? What are the primary trends in the development of AI chatbot technology? What are the principal directions for their 16 Følstad, Asbjørn et al. “Future Directions for Chatbot Research: An Interdisciplinary Research Agenda.” Computing 103, no. 12 (2021): 2915-2942; Rapp, Amon, Lorenzo Curti, and Arianna Boldi, “The Human Side of Human-Chatbot Interaction: A Systematic Literature Review of Ten Years of Research on Text-Based Chatbots,” International Journal of Human-Computer Studies 151 (2021). 17 Liu, Li, and Vincent G. Duffy, “Exploring the Future Development of Artificial Intelligence (AI) Applications in Chatbots: A Bibliometric Analysis,” International Journal of Social Robotics (2023): 14; Donthu, Naveen et al, “How to Conduct a Bibliometric Analysis: An Overview and Guidelines,” Journal of Business Research 133 (2021): 285-296. 18 Kuhail, Mohammad Amin et al, “Interacting with Educational Chatbots: A Systematic Review,” Education and Information Technologies 28, no. 1 (2022): 973-1018; Ramesh, Aishwarya, and Vaibhav Chawla, “Chatbots in Marketing: A Literature Review Using Morphological and Co-Occurrence Analyses,” Journal of Interactive Marketing 57, no. 3 (2022): 472-496; Caldarini, Guendalina, Sardar Jaf, and Kenneth McGarry, “A Literature Survey of Recent Advances in Chatbots,” Information 13, no. 1 (2022): 41; Niicolescu, Luminița, and Monica Teodora Tudorache, “Human-Computer Interaction in Customer Service: The Experience with AI Chatbots-A Systematic Literature Review,” Electronics 11, no. 10 (2022): 1579.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 181 application? Additionally, this study will investigate the challenges and opportunities encountered in the ongoing development and deployment of AI chatbot technologies. To address these questions, the study employs a bibliometric analysis approach, making several significant contributions to the field. First, it offers a comprehensive examination of the historical evolution and current trends in AI chatbot research, providing a foundational understanding of the trajectory of their development. Second, it establishes a structured research framework that delineates key themes and identifies potential future research directions. Finally, through an extensive analysis of the data, the study proposes prospective research pathways and offers valuable insights to guide both academic exploration and the practical advancement of AI chatbot technologies. 2. Research Methodology 2.1 Data Source and Processing This paper was retrieved on May 7, 2023, from the Web of Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI), and Arts & Humanities Citation Index (AHCI), encompassing publications from 2000 to the present. The selected literature spans multiple disciplines-SCI-EXPANDED, SSCI, and AHCI-aiming to provide a comprehensive understanding of the research status and development trends in the field of AI chatbots. The chosen literature covers various aspects, such as technological advancements, application scenarios, user experience, and ethical considerations, which are all closely related to the research questions and themes. The research period is defined as 2000 to 2023 for several compelling reasons. Firstly, 2000 marks a significant period of rapid development in internet and artificial intelligence technologies, which laid the groundwork for the emergence and growth of AI chatbots. Following this year, there has been a steady increase in publications related to chatbots, ensuring a robust sample size for in-depth analysis. Additionally, the evolution of chatbot technology from an exploratory phase to one of accelerated development since 2000 makes this timeframe particularly suitable for capturing advancements in the field. It's important

Chen, Renzhang; Zhao, Haixia 182 to acknowledge that while chatbot technology existed prior to 2000, exemplified by the ELIZA chatbot developed at MIT in 1966, the ensuing years have seen a transformative shift in both the technology and its applications. The search was limited to articles, review articles, and early access materials. To identify relevant literature on chatbots, the search terms utilized were “chatbot,” “conversational agent”, “dialog system”and “relational agent” as suggested by Camilleri and Troise.19 Moreover, following the recommendations of Hwang and Tu,20 the search terms for AI encompassed “artificial intelligence”, “machine intelligence”, “intelligent support” , “intelligent virtual reality”, “machine learning” ,“intelligent agent*”, “expert system*”, “neural network*”, “natural language processing” and “intelligent system”. By formulating search formulas based on these two sets of search terms, a total of 1236 relevant documents were retrieved from the Web of Science. The full record and cited references of these documents were then obtained and utilized as the research data for this study. 2.2 Analytical Methods and Tools Bibliometric methods play a significant role in evaluating research in scientific and applied fields.21 The utilization of bibliometric software has become crucial in assessing and analyzing the research performance and efficiency of scientists, universities, and countries.22 By employing bibliometric software, quantitative analysis of the retrieved literature data offers a comprehensive understanding of research hotspots, evolutionary trends, and research frontiers within the field.23 In this study, we initially conducted a 19 Camilleri, Mark Anthony, and Ciro Troise, “Live Support by Chatbots with Artificial Intelligence: A Future Research Agenda,” Service Business 17, no. 1 (2022): 61-80. 20 Hwang, Gwo-Jen, and Yun-Fang Tu, “Roles and Research Trends of Artificial Intelligence in Mathematics Education: A Bibliometric Mapping Analysis and Systematic Review,” Mathematics 9, no. 6 (2021): 19584. 21 Ellegaard, O., and J. A. Wallin, “The Bibliometric Analysis of Scholarly Production: How Great Is the Impact?,” Scientometrics 105, no. 3 (2015): 1809-1831. 22 Moral-Muñoz, José A. et al, “Software Tools for Conducting Bibliometric Analysis in Science: An Up-to-Date Review,” El Profesional de la Información 29, no. 1 (2020): e290103. 23 Yan, Chunlai et al, “Knowledge Mapping of Research Data in China: A Bibliometric Study Using Visual

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 183 statistical analysis of the 1236 relevant literature sources obtained from AI chatbot research. Subsequently, we employed VOSviewer software (version 1.6.19), a commonly used bibliometric software, to import the complete records of the relevant literature and cited references. This facilitated descriptive and in-depth analyses, which formed the basis for subsequent discussions and conclusions. The research framework is illustrated in Figure 1. Figure 1. The structure of this paper Analysis,” Library Hi Tech (2022): 19.

Chen, Renzhang; Zhao, Haixia 184 3. Bibliometric Analysis 3.1 Publication Year Analysis Figure 2 depicts the publication trends in the field of AI chatbot research over the past two decades. The data reveals a remarkable increase in the number of publications, with six papers published in 2000 and a notable surge to 323 papers in 2022. This signifies a more than 53-fold rise in annual publications during this period (2000 vs. 2022). From 2000 to 2016, the annual publication count ranged from 2 to 21 in the AI chatbot subject area. Conversely, starting in 2017 (with 23 papers), the number of annual publications experienced a substantial and rapid growth trend. By 2022, the annual publication count had increased more than 14-fold compared to 2017. As of May 7, 2023, a total of 168 related papers have already been published. Overall, the publication analysis indicates that AI chatbot research has undergone an exploration phase (2000-2016) followed by a period of rapid development (2017-2023). After over two decades of progress, the cumulative number of publications in AI chatbot research has surpassed 1200, signifying a substantial academic accumulation in the field. Figure 2. Statistic of publications

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 185 3.2 Geographical Distribution Analysis Table 1 presents the top 10 countries or regions in AI chatbot research. Over the course of the past two decades, AI chatbot research has involved 86 countries or regions across the globe. Among the countries with the highest number of documents, the United States leads the way with 285 papers, constituting 23.058% of the total. Following closely is China with 222 papers, representing 17.961% of the total. Additionally, England, South Korea, Australia, India, and other countries have also made notable contributions. These top ten countries or regions collectively account for 90.858% of the total publications on AI chatbots, indicating a significant concentration of research activity. Table 1. The geographical distribution (Top 10) Country/region Count Percentage Country/region Count Percentage USA 285 23.058 INDIA 76 6.149 PEOPLES R CHINA 222 17.961 SPAIN 74 5.987 ENGLAND 115 9.304 GERMANY 64 5.178 SOUTH KOREA 90 7.282 TAIWAN 63 5.097 AUSTRALIA 79 6.392 JAPAN 55 4.45 3.3 Cited Sources Co-Citation Analysis Figure 3 illustrates the mapping of cited sources using co-citation analysis conducted with VOSviewer. Co-citation analysis, a well-established method in bibliometrics, was introduced in 1973 and has since become widely employed in empirical studies examining the structure of scientific communication.24 The aim of co-citation analysis is to identify the most influential literature, the authors responsible for its publication, and the sources 24 Gmür, Markus, “Co-Citation Analysis and the Search for Invisible Colleges: A Methodological Evaluation,” Scientometrics 57, no. 1 (2003): 27-57.

Chen, Renzhang; Zhao, Haixia 186 that have published it by examining co-citation relationships and their frequencies. Consequently, source co-citation analysis focuses on books, journals, conferences, or other sources that disseminate research findings. In this study, sources with a citation frequency exceeding 20 were selected, resulting in 494 out of the 22,610 sources in the dataset meeting the established criteria. Figure 3. The mapping of cited sources co-citation analysis Figure 3 reveals the diverse range of sources within the AI chatbot research domain, encompassing journals, conferences, books, and arXiv. Notably, Computers in Human Behavior, Journal of Medical Internet Research, and Lecture Notes in Computer Science occupy central positions within this research area and display close proximity to other sources. The visualization in Figure 3 highlights the interdisciplinary nature of AI chatbot research, which draws upon a variety of disciplines including Computer Science, Experimental Psychology, Business, and others, indicating significant cross-fertilization between these fields. 3.4 Cited Authors Co-Citation Analysis Figure 4 depicts the co-citation analysis of cited authors using VOSviewer. The analysis focuses on authors within the published literature. For this study, authors who received more than 50 citations were considered, resulting in a selection of 59 authors out of the total 35,359 authors in the dataset. Figure 4 illustrates that co-citations among authors in the AI chatbot research field can be classified into three primary categories: human-computer interaction researchers, human-related researchers, and computer-related

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 187 researchers. Figure 4. The mapping of cited authors co-citation analysis The first category pertains to human-computer interaction researchers, primarily represented by the blue area in Figure 4, which includes notable Figures such as Turing and Weizenbaum. HCI refers to the interaction between machines and humans, enabling machines to better comprehend human intentions and engage in natural language communication. Turing's seminal work on the Turing test laid the foundation for the study of HCI.25 In 1966, the Massachusetts Institute of Technology developed and introduced the world's first chatbot named ELIZA, a significant milestone in this field.26 Since then, chatbot technology has undergone numerous iterations and advancements alongside the progress of artificial intelligence, attracting continuous research and exploration from scholars throughout its extensive history. These HCI researchers have furnished the theoretical groundwork for AI chatbot research. Simultaneously, theoretical investigations into HCI will play a pivotal role in establishing an autonomous academic community dedicated to AI chatbot research. The second category encompasses human-related researchers, predominantly represented by the red area in Figure 4, including Nass, Venkatesh, Huang, Araujo, and others. Notably, in the 1990s, Professor Nass pioneered the examination of human social responses to computers from psychological and sociological perspectives, revealing that 25 Turing, Alan M., and J. Haugeland, “Computing Machinery and Intelligence,” In The Turing Test: Verbal Behavior as the Hallmark of Intelligence, 29-56. 1950. 26 Weizenbaum, Joseph, “ELIZA-A Computer Program for the Study of Natural Language Communication between Man and Machine,” Communications of the ACM 9, no. 1 (1966): 36-45.

Chen, Renzhang; Zhao, Haixia 188 individuals unconsciously apply social rules and expectations to computer interactions.27 Venkatesh et al. delved deep into the theory of user acceptance and consumer acceptance of information technology. 28 In 2018, Huang et al. proposed the theory of AI job substitution. Araujo explored anthropomorphic design cues and communicative agency framing for AI chatbots in the same year. 29 Araujo's publications have emerged as highly influential contributions in the field of AI chatbot research. It is evident that human-related researchers in diverse humanities and social science disciplines such as psychology, sociology, communication, management, and business have conducted research related to AI chatbot users, thereby establishing the humanistic foundation for contemporary AI chatbot research. The third category comprises computer-related researchers, primarily represented by the green area in Figure 4, including Serban, Young, Williams, Walker, and others. Notably, Young et al. presented a principled mathematical framework in 2010 for modeling uncertainty in spoken dialogue systems using partially observable Markov decision processes (POMDPs). In 2012, Walkerdeveloped a reinforcement learning-based spoken dialogue system to acquire optimal dialogue strategies through interaction with human users. 30 Serban et al. addressed the multi-round dialogue problem in 2016 using a 27 Nass, Clifford, and Youngme Moon, “Machines and Mindlessness: Social Responses to Computers,” Journal of Social Issues 56, no. 1 (2000): 81-103; Nass, Clifford, Jonathan Steuer, and Ellen R. Tauber, “Computers Are Social Actors,” In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 1994. 28 Venkatesh, Viswanath et al, “User Acceptance of Information Technology: Toward a Unified View,” MIS Quarterly 27, no. 3 (2003): 425-478; Venkatesh, Viswanath, James Y. L. Thong, and Xin Xu, “Consumer Acceptance and Use of Information Technology: Extending the Unified Theory of Acceptance and Use of Technology,” MIS Quarterly 36, no. 1 (2012): 157-178. 29 Huang, Ming-Hui, and Roland T. Rust, “Artificial Intelligence in Service,” Journal of Service Research 21, no. 2 (2018): 155-172; Araujo, Theo, “Living up to the Chatbot Hype: The Influence of Anthropomorphic Design Cues and Communicative Agency Framing on Conversational Agent and Company Perceptions,” Computers in Human Behavior 85 (2018): 183-189; Young, Steve et al, “The Hidden Information State Model: A Practical Framework for POMDP-Based Spoken Dialogue Management,” Computer Speech & Language 24, no. 2 (2010): 150-174. 30 Walker, Marilyn A, “An Application of Reinforcement Learning to Dialogue Strategy Selection in a Spoken Dialogue System for Email,” Journal of Artificial Intelligence Research 12 (2000): 387-416.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 189 hierarchical seq2seq architecture.31 Williams et al. introduced the Hybrid Code Network HCN in 2017, incorporating domain knowledge modules and behavioral templates to achieve comparable or superior performance with less training data, which can be optimized through supervised learning, reinforcement learning, or a combination of both. 32 Collectively, the diverse research endeavors by computer-related researchers in computer science and artificial intelligence over the past two decades lay the technical foundation for contemporary AI chatbot research. 3.5 Author Keywords Co-Occurrence Analysis Co-word analysis is a content analysis method introduced in 1983, which utilizes keywords that describe the content of scientific literature as fundamental components for mapping or tracing patterns and trends in terminological relevance.33 It is commonly employed in conjunction with cluster analysis, wherein multivariate statistical techniques are applied to establish keyword clusters. This approach helps identify themes or research areas within the scientific literature, enabling the exploration of hotspots and frontiers for future research. In this study, co-word analysis of author keywords was conducted on a dataset consisting of 1236 documents in the field of AI chatbot research using VOS viewer. Prior to the analysis, data cleaning and term merging were performed in VOS viewer. For instance, different expressions such as “chatterbot”, “conversational agent”, “dialog system” and “ChatGPT” were standardized as “chatbot”, while various expressions related to AI were unified as “artificial intelligence”. Subsequently, author keywords with a co-occurrence frequency of at least 8 were selected. Out of the 3321 keywords in the dataset, 31 Serban, Iulian et al, “Building End-to-End Dialogue Systems Using Generative Hierarchical Neural Network Models,” Proceedings of the AAAI Conference on Artificial Intelligence 30, no. 1 (2016). 32 Williams, Jason D., Kavosh Asadi, and Geoffrey Zweig, “Hybrid Code Networks: Practical and Efficient End-to-End Dialog Control with Supervised and Reinforcement Learning,” arXiv preprint arXiv: 1702.03274, 2017. 33 Wang, Zhong-Yi, Gang Li, Chun-Ya Li, and Ang Li, “Research on the Semantic-Based Co-Word Analysis,” Scientometrics 90, no. 3 (2011): 855-875.

Chen, Renzhang; Zhao, Haixia 190 64 keywords met the specified criteria. The resulting co-occurrence analysis graph of author keywords is depicted in Figure 5. Figure 5. The mapping of author keywords co-occurrence analysis Figure 5 illustrates the outcomes of the co-word analysis conducted on author keywords in the domain of AI chatbot research, revealing six distinct clusters denoted by different colors. Within each color cluster, the keywords share close associations. Table 2 summarizes the primary concepts and key phrases identified within these six clusters based on the findings depicted in Figure 5. Table 2. Research concepts and hot topics Cluster Concept Nodes(n=64) 1 AI technologies Attention mechanism, BERT, convolutional neural network, data models, decoding, deep learning, emotion recognition, feature extraction, history, human-AI interaction, intent detection, LSTM, natural language, natural language generation, natural language processing, natural language understanding, neural network, question answering, recurrent neural network, reinforcement learning, semantics, slot filling, social networking, speech recognition, spoken dialog system, task analysis, training, transformer, visualization (n=29)

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 191 2 Digital Health and Education Artificial intelligence, chatbot, communication, depression, digital health, education, ehealth, ethics, mental health, mobile health, mobile phone, physical activity, systematic review, technology, trust, virtual assistant, virtual reality, voice assistant (n=18) 3 Conversational AI Conversational AI, emotion, human-machine communication, knowledge graph, ontology, systematic literature review (n=6) 4 COVID-19 Big data, COVID-19, healthcare, internet of things, machine learning (n=5) 5 Customer Service Anthropomorphism, customer experience, customer service, human-computer interaction (n=4) 6 Sentiment Analysis Affective computing, sentiment analysis (n=2) Figure 5 and Table 2 delineate the six distinct research focal points within the AI chatbot domain, as revealed by the high-frequency author keyword co-occurrence analysis conducted using VOS viewer. These research hotspots encompass AI technologies, digital health and education, conversational AI, COVID-19, customer service, and sentiment analysis. 3.5.1 Research Related to AI technologies Cluster 1 (Red): This cluster, which comprises 29 keywords, stands as the largest cluster in our study. Notably, all these keywords revolve around AI technologies. Consequently, we have designated “AI technologies” as the primary cluster name and research direction within the AI chatbot domain. In this regard, we conducted a comprehensive analysis encompassing speech recognition, natural language processing (NLP), and deep learning. First and foremost, speech recognition is a technology that employs machines to recognize, comprehend, and convert speech signals. In certain real-world intricate speech environments, such as scenarios with distant sound sources, low signal-to-noise ratio, room reverberation, echo interference, and multiple-source signal interference, the speech recognition task becomes exceedingly challenging. 34 Presently, prevalent speech 34 Graves, A., and N. Jaitly, “Towards End-to-End Speech Recognition with Recurrent Neural Networks,” In

Chen, Renzhang; Zhao, Haixia 192 recognition technologies primarily rely on deep neural networks for model development and training. Furthermore, existing neural networks are continually enhanced to cater to distinct application scenario demands and data characteristics. Secondly, Natural Language Processing (NLP) is widely employed in HCI and encompasses various techniques, including syntactic analysis, information extraction, machine translation, semantic analysis, and text mining. These techniques serve as effective methods for analyzing, processing, and comprehending natural language.35 At present, the primary focus of NLP revolves around researching and applying enhanced models based on GPT and BERT, such as GPT-3, GPT-4, LaMDA, and Instruct GPT.36 Thirdly, deep learning is a machine learning technique that employs multilayer neural networks to emulate the cognitive processes of the human brain, enabling the performance of intricate learning tasks using extensive data. Within AI chatbots, deep learning is employed for various tasks, including semantic analysis, natural language generation, intent recognition, and conversation state tracking, resulting in promising outcomes.37 Presently, with the widespread application and advancement of GPT and BERT, featuring Transformer as the architecture in the NLP community, the utilization of deep learning techniques allows for the creation of more intelligent chatbots that offer users more convenient services. Proceedings of the International Conference on Machine Learning, Beijing, China, June 22-24, 2014; Chorowski, J. et al, “Attention-Based Models for Speech Recognition,” In Proceedings of the 29th Annual Conference on Neural Information Processing Systems (NIPS), Montreal, Canada, December 7-12, 2015. 35 “Improving Language Understanding with Unsupervised Learning,” OpenAI Blog. June 11, 2018. Accessed February 9, 2023. https://openai.com/blog/language-unsupervised/; Devlin, Jacob et al, “BERT: Pre-Training of Deep Bidirectional Transformers for Language Understanding,” arXiv preprint arXiv:1810.04805, 2018. 36 Brown, Tom et al, “Language Models Are Few-Shot Learners,” Advances in Neural Information Processing Systems 33 (2020): 1877-1901; OpenAI, “GPT-4 Technical Report,” arXiv preprint arXiv:2303.08774, 2023; Thoppilan, Romal et al, “LaMDA: Language Models for Dialog Applications,” arXiv preprint arXiv:2201.08239, 2022; Ouyang, Long et al, “Training Language Models to Follow Instructions with Human Feedback,” Advances in Neural Information Processing Systems 35 (2022): 27730-27744. 37 Shumanov, Michael, and Lester Johnson, “Making Conversations with Chatbots More Personalized,” Computers in Human Behavior 117 (2021): 106627.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 193 3.5.2 Research Related to Digital Health and Education Cluster 2 (Green) comprises 18 keywords primarily associated with the implementation of AI chatbots in digital health, education, and related functionalities. First, AI chatbots have the potential to revolutionize the healthcare industry by offering personalized solutions and enhancing patient prognosis. In recent years, significant advancements have been made in the application of AI chatbots in digital health, mental health, and depression. Specifically, the application of AI chatbots in digital health encompasses mental health, healthy eating, physical activity, and various physical disorders.38 It is worth noting that the healthcare industry possesses unique characteristics such as data sensitivity and life-safety considerations, leading to scholars dedicating their efforts to researching, designing, and developing AI chatbot systems in the realms of digital health and mental health.39 In conclusion, the aim of applying AI chatbots in the healthcare industry is not to disrupt traditional medical relationships but to serve as an effective complement and enhancement, particularly in areas like patient self-care and medical consultation services. However, due to the challenge of limited access to AI algorithm training data in the healthcare industry, the application of AI chatbots in this sector will remain a focal point and a challenging aspect of future research. Second, AI chatbots hold significant potential in the field of education and can aid students in comprehending learning materials while assisting teachers in delivering enhanced guidance.40 Presently, AI chatbots find widespread application across various educational disciplines, encompassing early education, STEM education, and language education. Their implementations include collaborative learning, mobile/game-based 38 Chakraborty, S. et al, “Large-Scale Textual Datasets and Deep Learning for the Prediction of Depressed Symptoms,” Computational Intelligence and Neuroscience (2022): 5731532; Chew, H. S. J, “The Use of Artificial Intelligence-Based Conversational Agents (Chatbots) for Weight Loss: Scoping Review and Practical Recommendations,” JMIR Medical Informatics 10, no. 4 (2022): e32578. 39 Hassan, A. et al, “Development of NLP-Integrated Intelligent Web System for E-Mental Health,” Computational and Mathematical Methods in Medicine (2021): 1546343. 40 Pérez, José Quiroga, Thanasis Daradoumis, and Joan Manuel Marquès Puig, “Rediscovering the Use of Chatbots in Education: A Systematic Literature Review, ” Computer Applications in Engineering Education 28, no. 6 (2020): 1549-1565.

Chen, Renzhang; Zhao, Haixia 194 learning, distance learning, and affective learning.41 These AI chatbots can autonomously assume instructor roles or collaborate with human instructors.42 Moreover, experimental findings have indicated that students exhibit increased learning outcomes and academic performance when engaged with AI chatbots compared to traditional learning methods. Consequently, the utilization of AI chatbots in education possesses the potential to revolutionize conventional learning and teaching practices. Nevertheless, the implementation of AI chatbots in education is still in its nascent stage, and various ethical and practical challenges persist. The current scarcity of empirical studies on AI chatbots in education exemplifies these challenges.43 In conclusion, AI chatbots have emerged as promising tools in education, offering ample opportunities for further research and innovative practices within the field. 3.5.3 Research Related to Conversational AI Cluster 3 (Blue): Comprised of 6 keywords, “Conversational AI” has been chosen as the name of the third cluster and as a research direction for AI chatbot studies. Conversational AI refers to the technology that encompasses speech and text recognition, natural language processing, intent and emotion understanding, and enables interactions resembling those between humans and machines. It serves as a fundamental component of AI chatbots and encompasses various research areas, such as human-machine communication, sentiment modeling, knowledge graph integration, and ontology development. Achieving a deeper understanding of human-computer communication is vital for creating engaging chatbot experiences, ontologies contribute to knowledge organization and retrieval within chatbots, knowledge graphs provide structured 41 Chen, X. L. et al, “Artificial Intelligent Robots for Precision Education: A Topic Modeling-Based Bibliometric Analysis,” Educational Technology & Society 26, no. 1 (2023): 171-186. 42 Chen, Lijia, Pingping Chen, and Zhijian Lin, “Artificial Intelligence in Education: A Review,” IEEE Access 8 (2020): 75264-75278. 43 Hwang, Gwo-Jen, and Ching-Yi Chang, “A Review of Opportunities and Challenges of Chatbots in Education,” Interactive Learning Environments (2021): 1-14. https://doi.org/10.1080/10494820.2021.1952615.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 195 representations of information and enhance chatbots’ conversational capabilities, and emotional intelligence plays a crucial role in effective human-robot communication.44 In summary, Conversational AI technologies possess extensive research potential to enhance machines’ comprehension, processing, and response to human language and intentions. 3.5.4 Research Related to COVID-19 Cluster 4 (Yellow): Comprising of 5 keywords, we have chosen “COVID-19” as the designated cluster name and research focus for AI chatbot research. The global outbreak of COVID-19 has resulted in a significant number of confirmed cases, and AI chatbots have been utilized to carry out specific pre-hospital triage tasks in medical response to COVID-19, the New Coronary Pneumonia Screening System, and the New Coronary Pneumonia Diagnostic Model. 45 Moreover, COVID-19 has also contributed to a surge in mental health disorders worldwide, and AI chatbots have shown promising applications in supporting mental health during the COVID-19 pandemic.46 Additionally, the COVID-19 crisis has witnessed an influx of information, often 44 Ho, A., J. Hancock, and A. S. Miner, “Psychological, Relational, and Emotional Effects of Self-Disclosure after Conversations with a Chatbot,” Journal of Communication 68, no. 4 (2018): 712-733. https://doi.org/10.1093/joc/jqy026; Tian, Junrui et al, “Intention Model-Based Multi-Round Dialogue Strategies for Conversational AI Bots,” Applied Intelligence 52, no. 12 (2022): 13916-13940. https://doi.org/10.1007/s10489-022-03288-8; Yoo, SoYeop, and OkRan Jeong, “EP-Bot: Empathetic Chatbot Using Auto-Growing Knowledge Graph,” Computers, Materials & Continua 67, no. 3 (2021): 2807-2817; Firdaus, Mauajama et al, “EmoSen: Generating Sentiment and Emotion Controlled Responses in a Multimodal Dialogue System,” IEEE Transactions on Affective Computing 13, no. 3 (2022): 1555-1566. 45 Lai, L. et al, “Digital Triage: Novel Strategies for Population Health Management in Response to the COVID-19 Pandemic,” Healthcare (Amsterdam, Netherlands) 8, no. 4 (2020): 100493; de Campos Filho, A. S. et al, “Content and Usability Validation of an Intelligent Virtual Conversation Assistant Used for Virtual Triage During the COVID-19 Pandemic in Brazil,” Computers, Informatics, Nursing 40, no. 11 (2022): 779-785; Ur Rehman, Muneeb et al, “IoT & AI Enabled Three-Phase Secure and Non-Invasive COVID-19 Diagnosis System,” Computers, Materials & Continua 71, no. 1 (2022): 423-438. 46 Jiang, Q., Y. Zhang, and W. Pian, “Chatbot as an Emergency Exit: Mediated Empathy for Resilience via Human-AI Interaction During the COVID-19 Pandemic,” Information Processing & Management 59, no. 6 (2022): 103074; Nelekar, Shreeya et al, “Effectiveness of Embodied Conversational Agents for Managing Academic Stress at an Indian University (ARU) During COVID‐19,” British Journal of Educational Technology 53, no. 3 (2021): 491-511

Chen, Renzhang; Zhao, Haixia 196 accompanied by disinformation on the internet, posing significant challenges to public health. In this context, AI chatbots have played a prominent role in managing COVID-19-related information.47 To summarize, the COVID-19 pandemic has exposed a notable disparity between healthcare resources and patient needs, and AI chatbots have emerged as valuable tools in assisting individuals with their healthcare requirements during this challenging period. 3.5.5 Research Related to Customer Service Cluster 5 (Purple): Comprised of 4 keywords, “Customer Service” has been chosen as the fifth cluster name and research direction for AI chatbot studies. There is significant commercial value in implementing AI chatbots in the field of customer service, as they have found wide-ranging applications in replacing or augmenting manual customer service in the business sector. Numerous studies have focused on anthropomorphism, exploring the human-like attributes of chatbots. 48 Additionally, a substantial body of research has examined customer experience, satisfaction, and other factors closely associated with customer service.49 In summary, AI chatbots have become practical tools in the business realm, increasingly replacing traditional human customer service. Moving forward, attention should be given to human-machine interaction and anthropomorphic design to enhance customer service experiences and generate greater 47 Pandey, R. et al, “A Machine Learning Application for Raising WASH Awareness in the Times of COVID-19 Pandemic,” Scientific Reports 12, no. 1 (2022): 810. 48 Lu, Lincoln et al, “Measuring Consumer-Perceived Humanness of Online Organizational Agents,” Computers in Human Behavior 128 (2022): 107092; Crolic, Cammy et al, “Blame the Bot: Anthropomorphism and Anger in Customer–Chatbot Interactions,” Journal of Marketing 86, no. 1 (2022): 132-148. 49 Pillai, Rajasshrie, and Brijesh Sivathanu, “Adoption of AI-Based Chatbots for Hospitality and Tourism,” International Journal of Contemporary Hospitality Management 32, no. 10 (2020): 3199-3226; Ruan, Yanya, and József Mezei, “When Do AI Chatbots Lead to Higher Customer Satisfaction Than Human Frontline Employees in Online Shopping Assistance? Considering Product Attribute Type,” Journal of Retailing and Consumer Services 68 (2022): 103059; Lee, Crystal T., Ling-Yen Pan, and Sara H. Hsieh, “Artificial Intelligent Chatbots as Brand Promoters: A Two-Stage Structural Equation Modeling-Artificial Neural Network Approach,” Internet Research 32, no. 4 (2022): 1329-1356.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 197 business value. 3.5.6 Research Related to Sentiment Analysis Cluster 6 (Turquoise): Comprising of 2 keywords, “Sentiment Analysis” has been chosen as the sixth cluster name and research direction within AI chatbot research. In the AI chatbot domain, the application of affective computing and sentiment analysis is primarily focused on personalized customer service, mental health support, and user sentiment feedback. Researchers have explored effective computing approaches and investigated the specific application of sentiment analysis in AI chatbots.50The convergence of affective computing, sentiment analysis, and AI chatbots represents a rapidly evolving field with significant potential to transform HCI. 4. Discussion 4.1 AI chatbot Research Trend Under the Perspective of Time Examining the present state of AI chatbot research and the trends in publications, it is evident that AI chatbot research has been consistently increasing over the past two decades, with a particularly notable surge in the last five years. The temporal evolution mapping of keywords in the six AI chatbot clusters, along with the average year of publication values for each keyword analyzed in VOSviewer, is illustrated in Figure 6, following the clusters and their key phrases depicted in Figure 5. As depicted in Figure 6, the evolution of AI technologies has garnered significant attention since 2000. Beginning in 2017, researchers have not only focused on AI 50 Cambria, Erik et al. “SenticNet 7: A Commonsense-Based Neurosymbolic AI Framework for Explainable Sentiment Analysis.” In Proceedings of the Thirteenth Language Resources and Evaluation Conference, Marseille, France, 2022; Shuang, Kai et al, “Interactive POS-Aware Network for Aspect-Level Sentiment Classification,” Neurocomputing 420 (2021): 181-196; Jain, Rishit et al, “Real-Time Sentiment Analysis of Natural Language Using Multimedia Input,” Multimedia Tools and Applications (2023): 16; Kim, Taekyung et al, “Robots, Artificial Intelligence, and Service Automation (RAISA) in Hospitality: Sentiment Analysis of YouTube Streaming Data,” Electronic Markets 32, no. 1 (2022): 259-275.

Chen, Renzhang; Zhao, Haixia 198 technologies but also dedicated considerable effort to the design of HCI in real-world commercial applications. In summary, the research trajectory of AI chatbot can be characterized as a progression from traditional chatbots to AI chatbots and subsequently to large-scale commercial applications. This progression has been facilitated by the combined contributions of AI technology, HCI design, and commercial applications, ultimately leading to the current phenomenon of the ChatGPT tsunami. In terms of the temporal evolution of specific keywords, two main trends can be observed: (1) the continuous enhancement of AI chatbot intelligence, encompassing keywords such as speech recognition, emotional computing, machine learning, natural language processing, and deep learning; (2) a focus on the harmonious symbiosis of AI development, HCI design, and commercial applications, as indicated by keywords like anthropomorphism, customer services, customer satisfaction, customer experience, and customer acceptance. It is evident that future AI chatbots will undergo further upgrades to become more intelligent, enabling them to collaborate more effectively with humans and assist in solving increasingly complex problems through learning and reasoning. Simultaneously, the future development of AI chatbots will be characterized by highly intelligent systems integrated into large-scale commercial applications, representing significant features of their advancement. For more research trends and directions, see Section 4.3. 4.2 Towards a Conceptual Framework for AI chatbot Research Utilizing the product development model of AI chatbot, the analysis in this paper focuses on the interpretation of the 64 high-frequency author keywords associated with AI chatbot research. These keywords are categorized into three levels: AI development, HCI design, and application scenarios. By refining this analysis, a conceptual framework encompassing the key aspects of AI chatbot research is presented, as illustrated in Figure 7.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 199 Figure 6. The temporal evolution mapping of keywords Figure 7. Conceptual framework for AI chatbot research

Chen, Renzhang; Zhao, Haixia 200 Conversational AI encompasses a wide range of AI technologies that facilitate and enhance “intelligent” interactions between computers and humans. These technologies include speech recognition, affective computing, natural language processing, deep learning, reinforcement learning, as well as various other methods of intelligent analysis related to HCI, such as text, speech, image, touch, and gesture. Among these AI development-related keywords, natural language processing and deep learning hold significant importance in the context of conversation management for AI chatbots. Specifically, natural language processing and deep learning techniques need to address research tasks such as semantics, sentiment analysis, natural language generation, intent detection, and task analysis. The success of AI chatbots in various human-related activities largely depends on the level of HCI design. HCI design factors play a crucial role in imitating human interaction and meeting user expectations. Among these factors, anthropomorphism is the core keyword associated with imitating human interaction. Considering its correlation with customer service and the temporal trend, anthropomorphism is expected to be a prominent research area and frontier in the field of AI chatbot HCI design. Meeting users' expectations is another critical aspect of HCI design. The core keywords associated with meeting users' expectations are customer experience and trust. The ability of AI chatbots to perform artificial replacement tasks effectively relies on their continuous improvement of customer experience and their ability to gain users' trust. Consequently, these keywords serve as the foundation and focal point of HCI design in the context of AI chatbots. The success of ChatGPT is poised to catalyze increased investment in AI chatbot research and development by global IT companies, consequently accelerating the widespread commercial deployment of various AI chatbot products across diverse industries. Based on the distribution of current high-frequency keywords, the primary application scenarios for AI chatbots presently encompass customer service, digital health, mental health, COVID-19, and education. It is anticipated that future AI chatbot application scenarios will encompass nearly all human-related activities.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 201 4.3 Suggestions on Future Research of AI chatbot AI chatbots have gained extensive utilization globally. To foster the ongoing and profound advancement of AI chatbots, this paper draws upon the findings of the literature analysis to propose the following recommendations for future research in this field. 4.3.1 Continuously upgrade the intelligent technology Firstly, enhancing the natural language processing capabilities and deep learning techniques of AI chatbots is crucial to facilitate a better and more precise comprehension of user language and intent, thereby enabling more accurate service provision. One of the key challenges for AI chatbots in the future lies in comprehending conversation context and generating emotionally rich and personalized responses. Secondly, augmenting the machine learning capabilities of AI chatbots enables them to effectively anticipate user intentions based on their interaction history, leading to faster service delivery. Additionally, incorporating reinforcement learning techniques allows AI chatbots to continuously learn from multiple iterations, thereby enhancing service efficiency. Furthermore, leveraging big data techniques can enhance the data mining capabilities of AI chatbots, improve their conversation management systems to enhance semantic understanding and response accuracy, and elevate their sentiment analysis capabilities to enhance user communication experiences. Research indicates a substantial focus from numerous authors on the AI capabilities of chatbots, 51with the continuous enhancement of AI chatbot intelligence technology remaining a prominent area for future investigation. 4.3.2 Focus on the HCI design Chatbots, particularly those employed for customer service, have garnered a 51 Cui, F., Q. Cui, and Y. Song, “A Survey on Learning-Based Approaches for Modeling and Classification of Human-Machine Dialog Systems,” IEEE Transactions on Neural Networks and Learning Systems 32, no. 4 (2021): 1418-1432.

Chen, Renzhang; Zhao, Haixia 202 reputation for being unreliable among consumers. 52 On one hand, to advance future research on chatbot HCI design, scholars can integrate chatbot user experience models, measurement, and evaluation methods to conduct comprehensive studies aiming to enhance service quality and user satisfaction.53 On the other hand, despite providing round-the-clock service in various domains such as sales, support, and marketing, chatbots still face low customer acceptance.54 Therefore, a key area for future research lies in influencing and enhancing the design of chatbot HCI based on current research to continuously improve user acceptance, engagement, experience, satisfaction, and trust in AI chatbots from a design-oriented perspective. 4.3.3 Strengthen the promotion of interdisciplinary research and practice AI chatbots exhibit a notable interdisciplinary nature, as evident from citations and keywords stemming from diverse fields in related studies. Previously, AI research primarily focused on the realm of computer science.55 However, the rapidly expanding research on chatbots currently embraces significant interdisciplinary aspects, encompassing fields such as informatics, management and marketing, media and communication sciences, linguistics and philosophy, psychology and sociology, engineering, design, and HCI. Looking ahead, further interdisciplinary efforts are essential to advance AI technologies, enhance human-intelligence interaction systems, and continually enhance the applicability of AI chatbots across various domains and scenarios, thereby maximizing their value. 52 Kaplan, A. D. et al, “Trust in Artificial Intelligence: Meta-Analytic Findings,” Human Factors 65, no. 2 (2023): 337-359. 53 Mostafa, Rania Badr, and Tamara Kasamani, “Antecedents and Consequences of Chatbot Initial Trust,” European Journal of Marketing 56, no. 6 (2021): 1748-1771. 54 Ashfaq, Muhammad et al, “I, Chatbot: Modeling the Determinants of Users’ Satisfaction and Continuance Intention of AI-Powered Service Agents,” Telematics and Informatics 54 (2020): 101473; Rese, Alexandra, Lena Ganster, and Daniel Baier, “Chatbots in Retailers’ Customer Communication: How to Measure Their Acceptance?,” Journal of Retailing and Consumer Services 56 (2020): 102176. 55 Aoki, Naomi. “An Experimental Study of Public Trust in AI Chatbots in the Public Sector.” Government Information Quarterly 37, no. 4 (2020): 101490.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 203 5. Conclusion This paper provides a comprehensive review of the AI chatbot research literature from 2000 to 2023, utilizing bibliometric and content analysis methods such as co-citation analysis, co-word analysis, and cluster analysis. Through a visual examination and in-depth interpretation of the references and author keywords from 1236 related articles, a systematic analysis was conducted. The statistical analysis reveals that AI chatbot research has progressed through an initial exploration phase (2000-2016) and a subsequent phase of rapid development (2017-2023). Through the examination of cited authors using co-citation analysis, it was observed that influential authors in the field of AI chatbot can be categorized into three groups: human-computer interaction, human-related, and computer-related researchers. These three types of authors contribute theoretical, humanistic, and technical foundations to AI chatbot research. Additionally, co-citation analysis of cited sources demonstrates the interdisciplinary nature of AI chatbot as a research field, highlighting cross-fertilization between various disciplines. Through an in-depth analysis of co-word analysis, it is evident that the research focus of AI chatbot centers around AI technologies, digital health and education, conversational AI, COVID-19, customer service, and sentiment analysis. Building upon the findings of the bibliometric analysis, this study summarizes the developmental trends of AI chatbot, proposes a conceptual framework for AI chatbot research, and provides suggestions for future research directions in the field. This study has certain limitations that should be addressed in future research. Primarily, the analysis is based on data from the Web of Science database, which may not cover all relevant AI chatbot studies. Additionally, the bibliometric and content analysis methods used here, including keyword-driven hotspot analysis with VOSviewer software, may overlook some essential literature and limit the theoretical depth of the analysis. The study’s timeframe, ending in November 2023, further restricts it from reflecting the latest advancements and emerging areas in the fast-evolving AI chatbot field, which has seen accelerated growth since 2017. To address these limitations, future research could adopt a combination of systematic reviews and meta-analyses, thereby providing a more

Chen, Renzhang; Zhao, Haixia 204 comprehensive and in-depth examination of AI chatbot developments.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 205 Bibliography [1] Allison, DeeAnn. “Chatbots in the Library: Is It Time?,” Library Hi Tech 30, no. 1 (2012): 95-107. [2] Mocanu, Bogdan-Costel et al. “Odin IVR-Interactive Solution for Emergency Calls Handling.” Applied Sciences 12, no. 21 (2022): 10844. [3] Miklosik, Andrej, Nina Evans, and Athar Mahmood Ahmed Qureshi. “The Use of Chatbots in Digital Business Transformation: A Systematic Literature Review.” IEEE Access 9 (2021): 106530-106539. [4] Deryugina, O. V. “Chatterbots.” Scientific and Technical Information Processing 37, no. 2 (2010): 143-147. [5] “Chatbot: Definition of Chatbot in English.” Lexico Dictionaries. 2022. Accessed November 30, 2022. https://www.lexico.com/en/definition/chatbot. [6] Nirala, Krishna Kumar, Nikhil Kumar Singh, and Vinay Shivshanker Purani. “A Survey on Providing Customer and Public Administration-Based Services Using AI: Chatbot.” Multimedia Tools and Applications 81, no. 16 (2022): 22215-22246. [7] Al-Rfou, Rami et al. “Conversational Contextual Cues: The Case of Personalization and History for Response Ranking.” arXiv preprint arXiv:1606.00372, 2016. [8] Moussawi, Sara, Marios Koufaris, and Raquel Benbunan-Fich. “How Perceptions of Intelligence and Anthropomorphism Affect Adoption of Personal Intelligent Agents.” Electronic Markets 31, no. 2 (2020): 343-364. [9] Park, Namkee et al. “Use of Offensive Language in Human-Artificial Intelligence Chatbot Interaction: The Effects of Ethical Ideology, Social Competence, and Perceived Humanlikeness.” Computers in Human Behavior 121 (2021): 106795. [10] Essel, Harry Barton et al. “The Impact of a Virtual Teaching Assistant (Chatbot) on Students’ Learning in Ghanaian Higher Education.” International Journal of Educational Technology in Higher Education 19, no. 1 (2022): 57. [11] Guo, Kai, Jian Wang, and Samuel Kai Wah Chu. “Using Chatbots to Scaffold EFL Students’ Argumentative Writing.” Assessing Writing 54 (2022): 100666. [12] Jiang, Hua, Yang Cheng, Jeongwon Yang, and Shanbing Gao. “AI-Powered Chatbot Communication with Customers: Dialogic Interactions, Satisfaction, Engagement,

Chen, Renzhang; Zhao, Haixia 206 and Customer Behavior.” Computers in Human Behavior 134 (2022): 107329. https://doi.org/10.1016/j.chb.2022.107329. [13] Garcia-Mendez, Silvia et al. “Entertainment Chatbot for the Digital Inclusion of Elderly People without Abstraction Capabilities.” IEEE Access 9 (2021): 75878-75891. [14] Grudin, Jonathan, and Richard Jacques. “Chatbots, Humbots, and the Quest for Artificial General Intelligence.” In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems (2019). [15] “What Is ChatGPT and Why Does It Matter? Here’s What You Need to Know.” ZDNet. December 7, 2022. Accessed December 9, 2022. https://www.zdnet.com/article/what-is-chatgpt-and-why-does-it-matter-heres-what-you-need-to-know/. [16] “ChatGPT Has Gained 100 Million Users in Just 2 Months after Being Released.” TechJuice. February 2, 2023. Accessed February 9, 2023. https://www.techjuice.pk/chatgpt-has-gained-100-million-users-in-just-2-months-after-being-released/. [17] “Global Chatbot Market - Growth, Trends, Covid-19 Impact, and Forecasts (2022 - 2027).” Research and Markets. 2022. Accessed September 4, 2022. https://www.researchandmarkets.com/reports/4622740/global-chatbot-market-growth-trends-covid-19. [18] Følstad, Asbjørn et al. “Future Directions for Chatbot Research: An Interdisciplinary Research Agenda.” Computing 103, no. 12 (2021): 2915-2942. [19] Rapp, Amon, Lorenzo Curti, and Arianna Boldi. “The Human Side of Human-Chatbot Interaction: A Systematic Literature Review of Ten Years of Research on Text-Based Chatbots.” International Journal of Human-Computer Studies 151 (2021). [20] Liu, Li, and Vincent G. Duffy. “Exploring the Future Development of Artificial Intelligence (AI) Applications in Chatbots: A Bibliometric Analysis.” International Journal of Social Robotics (2023): 14. [21] Donthu, Naveen et al. “How to Conduct a Bibliometric Analysis: An Overview and Guidelines.” Journal of Business Research 133 (2021): 285-296. [22] Kuhail, Mohammad Amin et al. “Interacting with Educational Chatbots: A Systematic Review.” Education and Information Technologies 28, no. 1 (2022): 973-1018.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 207 [23] Ramesh, Aishwarya, and Vaibhav Chawla. “Chatbots in Marketing: A Literature Review Using Morphological and Co-Occurrence Analyses.” Journal of Interactive Marketing 57, no. 3 (2022): 472-496. [24] Caldarini, Guendalina, Sardar Jaf, and Kenneth McGarry. “A Literature Survey of Recent Advances in Chatbots.” Information 13, no. 1 (2022): 41. [25] Nicolescu, Luminița, and Monica Teodora Tudorache. “Human-Computer Interaction in Customer Service: The Experience with AI Chatbots—A Systematic Literature Review.” Electronics 11, no. 10 (2022): 1579. [26] Camilleri, Mark Anthony, and Ciro Troise. “Live Support by Chatbots with Artificial Intelligence: A Future Research Agenda.” Service Business 17, no. 1 (2022): 61-80. [27] Hwang, Gwo-Jen, and Yun-Fang Tu. “Roles and Research Trends of Artificial Intelligence in Mathematics Education: A Bibliometric Mapping Analysis and Systematic Review.” Mathematics 9, no. 6 (2021): 19584. [28] Ellegaard, O., and J. A. Wallin. “The Bibliometric Analysis of Scholarly Production: How Great Is the Impact?” Scientometrics 105, no. 3 (2015): 1809-1831. [29] Moral-Muñoz, José A. et al. “Software Tools for Conducting Bibliometric Analysis in Science: An Up-to-Date Review.” El Profesional de la Información 29, no. 1 (2020): e290103. [30] Yan, Chunlai et al. “Knowledge Mapping of Research Data in China: A Bibliometric Study Using Visual Analysis.” Library Hi Tech (2022): 19. [31] Gmür, Markus. “Co-Citation Analysis and the Search for Invisible Colleges: A Methodological Evaluation.” Scientometrics 57, no. 1 (2003): 27-57. [32] Turing, Alan M., and J. Haugeland. “Computing Machinery and Intelligence.” In The Turing Test: Verbal Behavior as the Hallmark of Intelligence, 29-56. 1950. [33] Weizenbaum, Joseph. “ELIZA—A Computer Program for the Study of Natural Language Communication between Man and Machine.” Communications of the ACM 9, no. 1 (1966): 36-45. [34] Nass, Clifford, and Youngme Moon. “Machines and Mindlessness: Social Responses to Computers.” Journal of Social Issues 56, no. 1 (2000): 81-103. [35] Nass, Clifford, Jonathan Steuer, and Ellen R. Tauber. “Computers Are Social Actors.” In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, 1994.

Chen, Renzhang; Zhao, Haixia 208 [36] Venkatesh, Viswanath et al. “User Acceptance of Information Technology: Toward a Unified View.” MIS Quarterly 27, no. 3 (2003): 425-478. [37] Venkatesh, Viswanath, James Y. L. Thong, and Xin Xu. “Consumer Acceptance and Use of Information Technology: Extending the Unified Theory of Acceptance and Use of Technology.” MIS Quarterly 36, no. 1 (2012): 157-178. [38] Huang, Ming-Hui, and Roland T. Rust. “Artificial Intelligence in Service.” Journal of Service Research 21, no. 2 (2018): 155-172. [39] Araujo, Theo. “Living up to the Chatbot Hype: The Influence of Anthropomorphic Design Cues and Communicative Agency Framing on Conversational Agent and Company Perceptions.” Computers in Human Behavior 85 (2018): 183-189. [40] Young, Steve et al. “The Hidden Information State Model: A Practical Framework for POMDP-Based Spoken Dialogue Management.” Computer Speech & Language 24, no. 2 (2010): 150-174. [41] Walker, Marilyn A. “An Application of Reinforcement Learning to Dialogue Strategy Selection in a Spoken Dialogue System for Email.” Journal of Artificial Intelligence Research 12 (2000): 387-416. [42] Serban, Iulian et al. “Building End-to-End Dialogue Systems Using Generative Hierarchical Neural Network Models.” Proceedings of the AAAI Conference on Artificial Intelligence 30, no. 1 (2016). [43] Williams, Jason D., Kavosh Asadi, and Geoffrey Zweig. “Hybrid Code Networks: Practical and Efficient End-to-End Dialog Control with Supervised and Reinforcement Learning.” arXiv preprint arXiv:1702.03274, 2017. [44] Wang, Zhong-Yi, Gang Li, Chun-Ya Li, and Ang Li. “Research on the Semantic-Based Co-Word Analysis.” Scientometrics 90, no. 3 (2011): 855-875. [45] Graves, A., and N. Jaitly. “Towards End-to-End Speech Recognition with Recurrent Neural Networks.” In Proceedings of the International Conference on Machine Learning, Beijing, China, June 22-24, 2014. [46] Chorowski, J. et al. “Attention-Based Models for Speech Recognition.” In Proceedings of the 29th Annual Conference on Neural Information Processing Systems (NIPS), Montreal, Canada, December 7-12, 2015. [47] “Improving Language Understanding with Unsupervised Learning.” OpenAI Blog. June 11, 2018. Accessed February 9, 2023. https://openai.com/blog/language-unsupervised/.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 209 [48] Devlin, Jacob et al. “BERT: Pre-Training of Deep Bidirectional Transformers for Language Understanding.” arXiv preprint arXiv:1810.04805, 2018. [49] Brown, Tom et al. “Language Models Are Few-Shot Learners.” Advances in Neural Information Processing Systems 33 (2020): 1877-1901. [50] OpenAI. “GPT-4 Technical Report.” arXiv preprint arXiv:2303.08774, 2023. [51] Thoppilan, Romal et al. “LaMDA: Language Models for Dialog Applications.” arXiv preprint arXiv:2201.08239, 2022. [52] Ouyang, Long et al. “Training Language Models to Follow Instructions with Human Feedback.” Advances in Neural Information Processing Systems 35 (2022): 27730-27744. [53] Shumanov, Michael, and Lester Johnson. “Making Conversations with Chatbots More Personalized.” Computers in Human Behavior 117 (2021): 106627. [54] Chakraborty, S. et al. “Large-Scale Textual Datasets and Deep Learning for the Prediction of Depressed Symptoms.” Computational Intelligence and Neuroscience (2022): 5731532. [55] Chew, H. S. J. “The Use of Artificial Intelligence-Based Conversational Agents (Chatbots) for Weight Loss: Scoping Review and Practical Recommendations.” JMIR Medical Informatics 10, no. 4 (2022): e32578. [56] Hassan, A. et al. “Development of NLP-Integrated Intelligent Web System for E-Mental Health.” Computational and Mathematical Methods in Medicine (2021): 1546343. [57] Pérez, José Quiroga, Thanasis Daradoumis, and Joan Manuel Marquès Puig. “Rediscovering the Use of Chatbots in Education: A Systematic Literature Review.” Computer Applications in Engineering Education 28, no. 6 (2020): 1549-1565. [58] Chen, X. L. et al. “Artificial Intelligent Robots for Precision Education: A Topic Modeling-Based Bibliometric Analysis.” Educational Technology & Society 26, no. 1 (2023): 171-186. [59] Chen, Lijia, Pingping Chen, and Zhijian Lin. “Artificial Intelligence in Education: A Review.” IEEE Access 8 (2020): 75264-75278. [60] Hwang, Gwo-Jen, and Ching-Yi Chang. “A Review of Opportunities and Challenges of Chatbots in Education.” Interactive Learning Environments (2021): 1-14. https://doi.org/10.1080/10494820.2021.1952615.

Chen, Renzhang; Zhao, Haixia 210 [61] Ho, A., J. Hancock, and A. S. Miner. “Psychological, Relational, and Emotional Effects of Self-Disclosure after Conversations with a Chatbot.” Journal of Communication 68, no. 4 (2018): 712-733. https://doi.org/10.1093/joc/jqy026. [62] Tian, Junrui et al. “Intention Model-Based Multi-Round Dialogue Strategies for Conversational AI Bots.” Applied Intelligence 52, no. 12 (2022): 13916-13940. https://doi.org/10.1007/s10489-022-03288-8. [63] Yoo, SoYeop, and OkRan Jeong. “EP-Bot: Empathetic Chatbot Using Auto-Growing Knowledge Graph.” Computers, Materials & Continua 67, no. 3 (2021): 2807-2817. [64] Firdaus, Mauajama et al. “EmoSen: Generating Sentiment and Emotion Controlled Responses in a Multimodal Dialogue System.” IEEE Transactions on Affective Computing 13, no. 3 (2022): 1555-1566. [65] Lai, L. et al. “Digital Triage: Novel Strategies for Population Health Management in Response to the COVID-19 Pandemic.” Healthcare (Amsterdam, Netherlands) 8, no. 4 (2020): 100493. [66] de Campos Filho, A. S. et al. “Content and Usability Validation of an Intelligent Virtual Conversation Assistant Used for Virtual Triage During the COVID-19 Pandemic in Brazil.” Computers, Informatics, Nursing 40, no. 11 (2022): 779-785. [67] Ur Rehman, Muneeb et al. “IoT & AI Enabled Three-Phase Secure and Non-Invasive COVID-19 Diagnosis System.” Computers, Materials & Continua 71, no. 1 (2022): 423-438. [68] Jiang, Q., Y. Zhang, and W. Pian. “Chatbot as an Emergency Exit: Mediated Empathy for Resilience via Human-AI Interaction During the COVID-19 Pandemic.” Information Processing & Management 59, no. 6 (2022): 103074. [69] Nelekar, Shreeya et al. “Effectiveness of Embodied Conversational Agents for Managing Academic Stress at an Indian University (ARU) During COVID‐19.” British Journal of Educational Technology 53, no. 3 (2021): 491-511. [70] Pandey, R. et al. “A Machine Learning Application for Raising WASH Awareness in the Times of COVID-19 Pandemic.” Scientific Reports 12, no. 1 (2022): 810. [71] Lu, Lincoln et al. “Measuring Consumer-Perceived Humanness of Online Organizational Agents.” Computers in Human Behavior 128 (2022): 107092. [72] Crolic, Cammy et al. “Blame the Bot: Anthropomorphism and Anger in Customer–Chatbot Interactions.” Journal of Marketing 86, no. 1 (2022): 132-148.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 211 [73] Pillai, Rajasshrie, and Brijesh Sivathanu. “Adoption of AI-Based Chatbots for Hospitality and Tourism.” International Journal of Contemporary Hospitality Management 32, no. 10 (2020): 3199-3226. [74] Ruan, Yanya, and József Mezei. “When Do AI Chatbots Lead to Higher Customer Satisfaction Than Human Frontline Employees in Online Shopping Assistance? Considering Product Attribute Type.” Journal of Retailing and Consumer Services 68 (2022): 103059. [75] Lee, Crystal T., Ling-Yen Pan, and Sara H. Hsieh. “Artificial Intelligent Chatbots as Brand Promoters: A Two-Stage Structural Equation Modeling–Artificial Neural Network Approach.” Internet Research 32, no. 4 (2022): 1329-1356. [76] Cambria, Erik et al. “SenticNet 7: A Commonsense-Based Neurosymbolic AI Framework for Explainable Sentiment Analysis.” In Proceedings of the Thirteenth Language Resources and Evaluation Conference, Marseille, France, 2022. [77] Shuang, Kai et al. “Interactive POS-Aware Network for Aspect-Level Sentiment Classification.” Neurocomputing 420 (2021): 181-196. [78] Jain, Rishit et al. “Real-Time Sentiment Analysis of Natural Language Using Multimedia Input.” Multimedia Tools and Applications (2023): 16. [79] Kim, Taekyung et al. “Robots, Artificial Intelligence, and Service Automation (RAISA) in Hospitality: Sentiment Analysis of YouTube Streaming Data.” Electronic Markets 32, no. 1 (2022): 259-275. [80] Cui, F., Q. Cui, and Y. Song. “A Survey on Learning-Based Approaches for Modeling and Classification of Human-Machine Dialog Systems.” IEEE Transactions on Neural Networks and Learning Systems 32, no. 4 (2021): 1418-1432. [81] Kaplan, A. D. et al. “Trust in Artificial Intelligence: Meta-Analytic Findings.” Human Factors 65, no. 2 (2023): 337-359. [82] Mostafa, Rania Badr, and Tamara Kasamani. “Antecedents and Consequences of Chatbot Initial Trust.” European Journal of Marketing 56, no. 6 (2021): 1748-1771. [83] Ashfaq, Muhammad et al. “I, Chatbot: Modeling the Determinants of Users’ Satisfaction and Continuance Intention of AI-Powered Service Agents.” Telematics and Informatics 54 (2020): 101473. [84] Rese, Alexandra, Lena Ganster, and Daniel Baier. “Chatbots in Retailers’ Customer Communication: How to Measure Their Acceptance?” Journal of Retailing and Consumer Services 56 (2020): 102176.

Chen, Renzhang; Zhao, Haixia 212 [85] Aoki, Naomi. “An Experimental Study of Public Trust in AI Chatbots in the Public Sector.” Government Information Quarterly 37, no. 4 (2020): 101490.

AI Chatbot Research: A Bibliometric Analysis of Advancements and Trends 213 Author Information Renzhang Chen received his M.S. degree in Computer Technique from the School of Information Science and Technology, Jinan University, Guangzhou, China, in 2008. He obtained his Ph.D. degree from the School of Computer Science and Engineering, Macau University of Science and Technology, Macau, in 2023. He is currently an Associate Director at the Center for Modern Education Technology, Jinan University, Zhuhai, China. His research interests include computer vision, deep learning, and large language models. Haixia Zhao is an Assistant Professor at the University International College, Macau University of Science and Technology. Her research interests include literary and cultural criticism, creative writing theory, cultural heritage preservation, and newspaper literature studies. She is the corresponding author of this paper. Email: hxzhao@must.edu.mo

世漢學會國際中文教育研究專欄 The Column of the International Society for Chinese Language Teaching 專欄為教育部中外語言交流合作中心、世界漢語教學學會與《澳門科技大學學報》（人文社會科學版）編輯部合作設立，受中外語言交流合作中心資助，特此鳴謝。 The column is established in collaboration with the Center for Language Education and Cooperation of the Ministry of Education, the International Society for Chinese Language Teaching, and the Editorial Department of the Journal of Macau University of Science and Technology (Humanities and Social Sciences Edition), funded by the Center for Language Education and Cooperation. We hereby express our gratitude.

教育部㆗外語言交流合作㆗心 簡 介 中外語言交流合作中心（簡稱「語合中心」，英文名稱“Center for Language Education and Cooperation”，簡稱 CLEC）隸屬於中國教育部，是發展國際中文教育事業的專業公益教育機構，致力於為世界各國民眾學習中文、瞭解中國提供優質的服務，為中外語言交流合作、世界多元文化互學互鑒搭建友好協作的平台。 語合中心的主要職能為發展國際中文教育與促進中外語言交流合作提供服務，統籌建設國際中文教育資源體系，參與制定國際中文教育相關標準並組織實施；支持國際中文教師、教材、學科等建設和學術研究；組織實施國際中文教師考試、外國人中文水準系列考試，開展相關評估認定；運行漢語橋、新漢學、獎學金等國際中文教育相關項目；開展中外語言交流合作等。

世界漢語教㈻㈻會 簡 介 世界漢語教學學會（簡稱「世漢學會」，英文名稱“The International Society for Chinese Language Teaching”）成立於一九八七年八月十四日，是經中華人民共和國民政部登記註冊的國際社會組織和非營利性民間學術團體，主管單位為中華人民共和國教育部，秘書處設在教育部中外語言交流合作中心（北京市西城區德勝門外大街一二九號四零一）。二零一一年十月與聯合國教科文組織建立合作關係，會員遍佈全球七十九個國家和地區，主要由世界各地從事漢語教學、研究和推廣的人士及相關機構組成。 學會宗旨是遵守中華人民共和國憲法、法律、法規和國家政策，遵守社會道德風尚；促進國際漢語教學、研究和推廣；加強世界各地漢語教學與研究工作者之間、機構之間的聯繫。自二零二零年起，學會相繼成立「教師發展專業委員會」、「標準與認證工作委員會」、「語言理論與語言教育研究分會」、「課程與教材專業委員會」、「學術交流與出版專業委員會」、「智慧教育分會」、「中小學教育研究分會」、「專門用途中文教育專業委員會」等八個分支機搆。 學會理事會為議事決策機構。歷任會長為朱德熙、呂必松、陸儉明、許嘉璐等，現任會長為天津師範大學校長鍾英華教授（二零二二年十二月當選）。第十二屆理事會由來自二十八個國家和地區的國際中文教育學術團體、各國高等院校及中文專業院系、著名漢學家、具有較高聲望的學術帶頭人、世漢學會創會會員、永久會員和普通會員等二十一位理事（單位）組成。

澳門科技大學學報(人文社會科學版)第十九卷第一期 世漢學會國際中文教育研究專欄 二零二五年三月，頁 217-236 DOI: 10.58664/mustjournal.2025.01.007 217 國際中文傳播視域下的金庸武俠研究：共賞 共情 共話* 樂琦 1、鄭博文 （1. 中國傳媒大學漢語國際教育中心教授） 摘要：金庸武俠作品在開啓中國故事本土化與國際化的敘事流動空間中，承載著豐富的中國故事，內嵌著獨特的武俠精神，以富有魅力的人物語言、高深莫測的中國功夫、生動立體的英雄形象將中國語言文化傳遞給世界。探尋金庸武俠作品如何從粵港澳大灣區出發，面向全球，成為世界理解中國的一扇窗。首先，立足國際中文傳播與國際中文教育研究交叉重疊的學術邏輯，釐清國際中文傳播視域下金庸武俠作品的理論背景與研究價值。其次，從「共賞：語言、文化、歷史、習俗的薈萃」「共情：英雄情結、俠義精神與文化內核」「共話：新秩序的追求與新世界的想象」等方面進行調查研究，深入挖掘金庸武俠作品的獨特人文內涵與跨文化溝通作用。最後，以文明互鑒為旨歸，探討國際中文傳播視域下運用金庸武俠作品進行國際中文教育的有效策略：在語言文化教學中促進跨文化理解、在文化交流實踐中增進文化符號解讀、在融媒體傳播中增強中華文化影響力，進而以金庸武俠作品為載體，建構一個國際中文傳播的「想象共同體」，探索推動國際中文傳播的實踐進路。 關鍵詞：金庸、國際中文傳播、武俠文化、共賞、共情、共話 * 收稿日期：2024年 12月 27日；通過日期 2025年 02月 17日。本文受 2024年度中國外文局科研項目「國際中文傳播的現狀與發展前景研究」（項目號：24GCRCYJ-02）以及教育部中外語言交流合作中心 2024年度「新漢學計劃」學術支持。

樂琦、鄭博文 218 Research on Jin Yong’s Martial Arts Works (wuxia) from the Perspective of International Chinese Communication: Shared Appreciation, Shared Empathy, Shared Dialogue Yue, Qi1; Zheng, Bowen (1. International Chinese Education Center, Communication University of China) Abstract: Jin Yong's martial arts fiction (wuxia) works play a pivotal role in opening the narrative space for the localization and internationalization of Chinese stories. These works encapsulate vivid Chinese tales, embodying unique martial arts spirit (wuxia), through charismatic character dialogues, profound martial arts, multidimensional heroic figures, they convey Chinese language and culture to the world. This study explores how Jin Yong's wuxia works, Originating from Guangdong-Hong Kong area, reaching global audiences and has become a window for the world to understand China. First, based on the overlapping academic perspectives of international Chinese communication and international Chinese education, the theoretical background and research value of Jin Yong's wuxia works in the context of international Chinese communication are clarified. Secondly, the study investigates the unique humanistic connotations and cross-cultural communication functions such as: “shared appreciation” (a blend of language, culture, history, and customs), “shared empathy” (heroic ideals, chivalric spirit, and cultural core essence), and “shared dialogue” (the pursuit of new order and the imagination of a new world). Finally, the research explores effective strategies for utilizing Jin Yong's wuxia works in international Chinese education under the framework of cultural exchange. These include fostering cross-cultural understanding in language and culture education, enhancing the interpretation of cultural symbols in cultural exchange practices, and strengthening the influence of Chinese culture through multimedia communication. Therefore, using Jin Yong's wuxia works as a medium, the study aims to build a “dream community” of international Chinese communication and also explore practical approaches to promoting international Chinese communication globally. Keywords: Jin Yong; International Chinese communication; Martial culture; Shared appreciation; Shared empathy; Shared dialog

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 219 一、引言 從 1955年⾦庸發表的第⼀部武俠⼩說《書劍恩仇錄》，到 1972年的最後⼀部武俠⼩說《鹿鼎記》，加上早期的短篇⼩說《越女劍》，⾦庸給世⼈呈現了⼗五部武俠⼩說。1 ⾦庸武俠作品在開啓中國故事本⼟化與國際化的敘事流動空間中，承載著豐富的中國故事，內嵌著獨特的武俠精神，以富有魅⼒的⼈物語⾔、⾼深莫測的中國功夫、⽴體⽣動的英雄形象將中國語⾔⽂化傳遞給世界，成為中國作品海外傳播的典範。 國際中⽂傳播是指建⽴在世界各國對中⽂需求的基礎之上，中⽂遵循語⾔傳播規律，從中國⾛向世界的語⾔傳播現象。2 國際中⽂傳播不僅傳播中⽂語⾔本身，還傳播中國國情、歷史、⼈⽂、價值觀等深層次的內容。⾦庸⼩說在開啓中國故事本⼟化與國際化的敘事流動空間中，將中國語⾔⽂化傳遞給世界，為國際中⽂傳播提供了可延展的⽂化空間。⾦庸武俠作品如何從粵港澳⼤灣區出發，⾯向全球，成為世界理解中國的⼀扇窗？如何實現不同語⾔、不同⽂化背景讀者的「共賞」「共情」「共話」？如何運⽤⾦庸作品進⾏有效的國際中⽂教育，進⽽以⾦庸武俠作品為載體，建構⼀個國際中⽂傳播的「想象共同體」？探索與思考以上問題，確⽴了本⽂國際中⽂傳播與國際中⽂教育交叉重疊的研究視角與學術邏輯。 二、文獻梳理與問題的提出 由於⾦庸⾃身的影響⼒以及其武俠作品的廣泛傳播，⾦庸武俠獲得國內外諸多學者的關注。⼆⼗世紀七⼗年代以來，國內外學者對⾦庸及其作品展開研究，形成 1 金庸本名查良鏞，著名作家，1924年生於浙江省海寧市，1948年移居香港，並於 1955年開始在《新晚報》連載武俠小說《書劍恩仇錄》，此後直至 1972年，金庸陸續完成了 15部武俠小說，廣受歡迎。2009年，金庸榮獲影響世界華人終身成就獎；2018年 10月 30日下午，金庸在中國香港逝世，享年 94 歲；2019 年，新修《辭海》（第七版）在中國文學新增重要人物條目中，也將包括金庸在內的近幾年去世的名家列入其中；1970年 1月，初次發表於《明報晚報》。 2 吳應輝，〈國際中文教學學科建設及漢語國際傳播研究探討〉，《語言文字應用》，第 3期（北京：2010.08），35-42。

樂琦、鄭博文 220 「⾦學」。3 從國際中⽂傳播視域出發，相關研究主要聚焦以下三個⽅⾯。 ⼀是⾦庸作品中的⽂化內涵與價值。有⽂獻認為，⾦庸武俠是傳統與現代、東⽅與西⽅、中原與異域、古典精神與現代精神相融相⽣相混成的新武俠⼩說，⼩說中，⾦庸⼀⽅⾯刻畫了外國⼈物群像，並通過加⼯使之著上中國⾊彩，如：⼩昭、⾦⽑獅王、天竺僧、神雕等，另⼀⽅⾯借鑒使⽤了外國戲劇、電影和⼩說的技法，如開放式結局等。4 有⽂獻指出，⾦庸武俠⼩說是⼀種既有師承又有創新的現代 「⽂化⼩說」，它既蘊含著眾多中國古代⽂化因⼦，也順應了中國當代⽂化思維和思想價值取向，還汲取了西⽅現代⼩說的創作技巧，將中國傳統俠客故事與西⽅騎⼠⼩說融會貫通。5 ⼆是⾦庸武俠⼩說的翻譯與跨⽂化傳播。有⽂獻從翻譯角度出發，關注到《射雕英雄傳》英譯本在國外⼤受歡迎的原因：歷史的元素、奇幻的武功、俠義的精神、⾦庸的作品，並從⽂學翻譯的角度對中國⽂化如何⾛向世界提出建議。6 有⽂獻指出，東亞地區對於武俠⼩說的⾼度接受是由於東⽅⽂化圈所共有的⽂化共鳴，以及各國本身具有的俠義⼩說創作傳統，在此基礎上重新評估武俠⼩說的當代價值。7 有⽂獻⽴⾜全球化視角，認為⾦庸的「江湖」在上世紀八⼗年代不僅「提供了豐富多彩的幻想世界」，也「滿⾜了⼀種新秩序的追求與新世界的想象」，是⼀個「具有異質性的烏托邦」。8 三是⾦庸武俠作品的改編與跨媒介傳播。有⽂獻注意到⾦庸武俠⼩說的影視化改編，指出這種改編對⾦庸武俠⼩說的經典化起到了巨⼤推動作⽤。9 有⽂獻認為商業化、經典化、產業化是解讀⾦庸⼩說無法回避的三種⽂化現象，⾦庸的武俠⼩說有⼀種神奇的⽂化娛樂功能，是當下社會⾥ 3 金庸學的簡稱，由於金庸的武俠小說深受歡迎，引發讀者、學者的廣泛討論，形成金庸小說研究這一研究類目。 4 謝文興，〈論金庸武俠小說的外國印跡〉，《寧波廣播電視大學學報》，第 12卷第 2期（寧波：2014.06），23-27。 5 彭海雲，〈跨文化視域下金庸小說的藝術創新〉，《嘉興學院學報》，第 28卷第 1期（嘉興：2016.01），20-24。 6 徐雪英、張菁，〈從金庸《射雕英雄傳》英譯看中國文化如何走向世界〉，《浙江學刊》，第 3期（杭州：2020.05），42-53。 7 湯哲聲，〈中國百年武俠小說的價值評估與俠文化的現代構建〉，《文藝理論研究》，第 4期（上海：2023.07），182-193。 8 李瑋，〈「盛世江湖」與漫長的「九十年代」—從金庸、「後金庸」到純武俠的衰落〉，《小說評論》，第 1期（西安：2023.01），62-69。 9 原小平，〈改編的通俗性與文學經典的建構〉，《東北大學學報》，第 1期（瀋陽：2010.01），79-84。

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 221 ⼀種獨特的⽂化資源。10 有⽂獻以《射雕英雄傳》第⼗回中⼀段加插回憶的對打情節為例，說明⾦庸直接⽤類似電影的分鏡頭把武打場⾯的共時性分開。11 可⾒，⾦庸武俠⼩說⾃誕⽣時便具有可視化的特點，改編成影視作品亦是順理成章。 綜上，⽬前關⾦庸及其創作的相關研究較為豐富，然⽽，從國際中⽂傳播視域出發，以在華國際學⽣為關注對象，對⾦庸武俠展開的研究屈指可數。本⽂從國際中⽂傳播視域關照⾦庸武俠⼩說，以在華國際學⽣為調查對象，從「共賞：語⾔、⽂化、歷史、習俗的薈萃」「共情：英雄情結、俠義精神與⽂化內核」「共話：新秩序的追求與新世界的想象」等⽅⾯展開研究，調查⾦庸武俠在來華國際學⽣群體中的傳播與接受，深入挖掘⾦庸武俠作品的獨特⼈⽂內涵與跨⽂化溝通作⽤，進⽽探討運⽤⾦庸武俠開展國際中⽂傳播的提升路徑與傳播策略。 三、共賞：語言、文化、歷史、習俗的薈萃 「語⾔是活⽣⽣的，並且正是在這裡歷史地形成的，在具體的⾔語交際中，⽽不是在抽象的語⾔學的語⾔體系和說話者的個⼈⼼理之中形成。」14 巴赫⾦的超語⾔學注重語⾔的歷史性和社會性，內在性和外在性的交融，綜合主觀性與社會性、共時性與歷史性。⽂字是語⾔的重要載體，⾦庸武俠⼩說為武俠⽂化注入了新的⽣命⼒，也為國際中⽂傳播打開⼀扇窗。 （㆒）、關於㈮庸武俠小說的跨文化接受情況調研 ⽂字是語⾔的重要載體，⾦庸武俠⼩說在奇妙的武俠世界中，以豐富的⼈物語⾔開啓了中國語⾔⽂化傳播的⼀扇窗。⼆⼗世紀下半葉，⾦庸武俠⼩說及其武俠江湖在東南亞⽂化圈受到了廣泛關注，進入⼆⼗⼀世紀以來，越來越多的西⽅⼈開始關注中國的⽂化傳統和思維⽅式，⾦庸武俠⼩說、電影也隨之登陸歐美主流⽂化市 10 張欣，〈論金庸小說的三種文化現象〉，《雲南大學學報（社會科學版）》，第 2期（昆明：2016.03），103-110。 11 李歐梵，〈金庸與武俠小說傳統〉，《明報月刊》，（香港：2018.12），111。 14 巴赫金 M.M. Bakhtin 、錢中文（譯），〈馬克思主義與語言哲學〉，《巴赫金全集》第二卷，（石家莊：河北教育出版社，1998），448。

樂琦、鄭博文 222 場，並被歐洲學者與法國⼗九世紀作家⼤仲⾺的作品進⾏類比，這在西⽅主流⽂化圈中是非常⾼的評價。 本研究從國際中⽂傳播視域出發，以調查問卷與深度訪談相結合的形式，通過問卷星形式⾯向在華國際學⽣發放調查問卷，共計發放調查問卷 150份，收回有效問卷 112份，調查數據真實有效。受調者基本情況如下： 圖表 1 受訪者性別統計 圖表 2 受訪者所屬大洲統計 在收回的 112份有效問卷中，男性為 49⼈，女性為 63⼈。⼤部分受調者來⾃49!43.75%63!56.25%!"70!63.39%7!6.25%24!21.43%10!8.93%#$%$&$'$

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 223 亞洲國家（佔比 63.39%），與⾦庸武俠⼩說的跨⽂化傳播路徑吻合。其次是非洲國家（佔比 21.43%）︔然後依次為美洲（8.93%）、歐洲（6.25%）。需特別指出的是，本次調查沒有收集到來⾃⼤洋洲的有效問卷。調研問題及結果如下： 問題 1.你是否看過武俠⼩說或相關影視作品？Have you read martial arts novels or watched related movies?（多選題） 圖表 3 受調者是否看過中國武俠小說或相關影視作品 問題 2.你瞭解⾦庸嗎？ Do you know JinYong? （多選題） 圖表 4 受調者對金庸的瞭解程度 22!19.64%12!10.72%78!69.64%()*)*+,-./)*+01./33!29.46%30!26.79%38!33.93%11!9.82%23-4567*89:;45&<45

樂琦、鄭博文 224 問題 3.你知道幾部與⾦庸相關的武俠作品？ How many martial arts works about Jin Yong do you know？ （多選題） 圖表 5 受調者對金庸武俠作品的瞭解情況 問題 4.你從哪些渠道瞭解到⾦庸的武俠作品？ From which channels did you learn about Jin Yong's martial arts works？（多選題） 圖表 6 受調者接觸金庸武俠作品的渠道 051015202530350-1= 2-3= 4-6= 7-9= 10=>?@05101520253035404550ABCDECFGHIJDKDLMNOPQHPRSTMUVWHXYWZ[H\[H]^_`abcTde

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 225 調查結果顯⽰，80.36%的在華國際學⽣接觸過中國的武俠⼩說，其中，44.64%的學⽣⾃認為「比較瞭解⾦庸」，且均認為⾦庸可以作為中國武俠⼩說的代表作家。在選擇「瞭解⾦庸」的 72名國際學⽣中，⼤多數學⽣接觸過至少 2部⾦庸武俠相關的⽂藝作品，以 2-3部居多（佔比 38.89%）。國際學⽣依次是通過電影（佔比 21.1%）、電視劇（佔比 19.7%）、紙質書（佔比 15.2%）、電⼦書（佔比 9%）、網絡社交平台（佔比 8.1%）等途徑接觸到⾦庸武俠作品，多媒體對⽂學作品的傳播推動⼒由此可⾒⼀斑。 問題 5.通過⾦庸武俠⼩說，你瞭解到中國哪些⽅⾯的情況? What aspects of China have you learned through Jin Yong's martial arts novels ? （多選題） 圖表 7 受調者在金庸武俠作品中獲取的關於中國的認知 調查顯⽰，通過閱讀⾦庸的武俠⼩說，國際學⽣可以瞭解到中國的⽂化（57.14%）、歷史（46.43%）、語⾔（40.18%）、地理（26.79%）、政治（22.32%）等多⽅⾯內容，進⽽對中國國情與⽂化形成更加具體的認知。 010203040506070fg hi Oj kl mn

樂琦、鄭博文 226 問題 6.你喜歡⾦庸武俠中的哪些⽅⾯? What aspects of Jin Yong's martial arts works do you like? （多選題） 圖表 8 受調者喜愛金庸武俠作品的原因 通過上表可⾒，國際學⽣在⾦庸武俠⼩說中主要瞭解了中國的⽂化和歷史、語⾔等⽅⾯。此外，通過對⾦庸武俠喜好原因的調查，可知國際學⽣群體最喜歡的是⾦庸作品中的武俠想象，其次是故事情節、歷史⽂化和⼈物形象。 問題 7.你認為⾦庸武俠世界中包含了哪些中國⽂化元素? What Chinese cultural elements are included in the world of Jin Yong's martial arts? （多選題） 圖表 9 受調者關注的小說中的文化元素 010203040506070hiop qVrs tuvw fgOj xyzs {|v} de01020304050607080~ ÄÅ ÇÉ xÑ ÖÜ áR àP qV âV de

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 227 如圖所⽰，在閱讀⾦庸武俠⼩說的過程中，受調者⾸先關注到的是書中關於中國武術（61.61%）的描寫，其次分別是⼈物（45.54%）、服飾（42.86%）、詩詞（41.07%）等。總體來看，⾦庸在武俠世界中融入了⼤量中華⽂化元素，為在華國際學⽣閱讀和瞭解中華⽂化提供了有效⽂本。 （㆓）、㈮庸小說㆗的㆗國文化元素 「如果有誰要我介紹有關中國傳統藝術⽂化的入⾨書的話，我會毫不遲疑地向他推薦錢鐘書的《談藝錄》和⾦庸的⼗四部⼩說。」12 ⾦庸武俠⼩說是中國語⾔、⽂化、歷史、宗教、習俗的薈萃，⽤中⽂講中國故事，具有濃郁的中國特⾊。⾦庸武俠⼩說中的中華傳統⽂化元素體現在多個層⾯，種種「潤物細無聲」的⽂化嵌入，在潛移默化中促進了中華⽂化在全球範圍內的傳播和接受。 ⾸先，⾦庸的武俠世界中，⼈物的姓名、稱號、武功招式均有深意。作為⾦庸⼩說的核⼼⽂化符號之⼀，⾦庸的武俠精神主要體現為對正義、忠誠和俠義的追求。⼩說中的俠客不僅具備⾼超的武功，同時也在江湖中維護正義和秩序。在⾦庸作品中，不僅主角的名字引經據典，短暫出場的配角形象也有豐富的⽂化意蘊。如在《笑傲江湖》中，男女主⼈公的名字分別為令狐衝、任盈盈，正對應了《⽼⼦》中的「⼤盈若衝，其⽤不窮」︔⽽「梅莊四友」的⼈物塑造，則與中國傳統雅趣「琴棋書畫」相呼應，通過這些雅趣，表達出四位的藝術愛好與⼈⽣追求。⾦庸作品的⽂化積澱不僅體現在⾏⽂中，也蘊含在充滿詩情畫意的武功招式中。從黯然銷魂掌的以情動⼈到降龍⼗八掌的雄姿英發，再到六脈神劍的武學境界，這些招式共同參與建構了⼀個絢麗多彩的⾦庸武俠世界，內嵌著⾦庸對於武俠精神的深刻理解。⼩說所體現的武俠精神深刻反映了中國傳統的道德觀念和社會倫理，傳達了對個⼈品德和社會責任的⾼度重視。 ⾦庸作品體現了豐富的中醫藥元素與中醫智慧。例如，「九陽真經」「九陰真經」等不僅增強了中國功夫的神秘⾊彩，也展⽰了中醫藥理論在中華傳統⽂化中的重要地位。「東邪、西毒、南帝、北丐、中神通」這五位⾼⼿的名字與特質，與中醫五⾏的「⾦、⽊、⽔、⽕、⼟」息息相關。通過對這些⽂化元素的描寫，⾦庸武 12 王曉明、王海渭、張寅彭編，《胡河清文存》，（上海：三聯書店，1996），199。

樂琦、鄭博文 228 俠⼩說將各種⽂化符號融入到⼩說的故事情節和⼈物中，許多⼈物的健康狀況、疾病治療和傷病恢復都依賴於中醫藥的理論和實踐。 時間與空間、歷史與地理縱橫交錯的⼈⽂背景，是⾦庸作品的又⼀特⾊。⾦庸武俠⼩說在時間上跨越了宋、元、明、清這段千年歷史，在地域上囊括了中國⼤江南北。雖然⾦庸⼩說的故事⼤多是虛構的，但許多故事情節和⼈物都與真實歷史事件和⼈物密切相關。從遼寧的神龍教到浙江的桃花島，從中原的少林寺到陝西的華山派，⾦庸⼩說中的每個地⽅都有蘊含著獨特的⽂化與故事背景。從春秋時期的越女劍到出⽣於清朝的韋⼩寶，⼩說中的等歷史背景增強了故事的真實感，也為讀者提供了深入瞭解中國歷史⽂化的契機。 四、共情：英雄情結、俠義精神與文化內核 「共情」是⼀個⼼理學概念，指的是⼀種個體能夠準確理解他⼈情感，並在特定情景下做出準確情感反應的能⼒。共情傳播是「個體在⾯對群體的情緒情景時參與信息接收、感染和表達以及傳遞分享的⾏為過程」，展現了⼀種個體與群體、群體與群體之間的關係模式。13 共情傳播有兩層遞進關係：⾸先，個體在對他⼈傳遞的情緒、情感信息認知的基礎上，產⽣同理的情緒或情感，引起⼀定的評估、反饋⾏為︔在此基礎上，個體又將這些情緒、情感在⼀定群體內傳遞和擴散，引發群體的情感共鳴和反應。語⾔⽂化理解障礙造成的「⽂化折扣」困境，使其亟需尋找有助於彌合語⾔⽂化鴻溝的⽅法。14 在華國際學⽣來⾃不同的國家，擁有不同的⽂化背景，但他們的情感內核相通且共享。 （㆒）、言「情」：㈮庸武俠㆗刻畫的豐富情感 從共情傳播的視角關照⾦庸武俠，不難發現，⾔「情」正是其中⼀個鮮明元素。⾦庸武俠中的「情」多樣⽽真切：陳家洛和霍青桐、郭靖和⿈蓉、楊過和⼩龍女、 13 劉海明、宋婷，〈共情傳播的量度:重大公共衛生事件報道的共振與糾偏〉，《新聞界》，第 10期（成都：2020.10），11-21+31。 14 樂琦、陳晨，〈國際中文教育品牌的實踐進路〉，《陝西師範大學學報》（哲學社會科學版），第 6期（西安：2024.12），29-39。

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 229 周伯通和劉瑛姑、令狐衝和任盈盈等⼈之間的愛情，江南七怪、全真七⼦等⼈的友情，⿈藥師對⿈蓉的舐犢之情，⽯破天對梅芳姑的反哺之情，⿈藥師和曲凌風的師徒情，武當七俠的兄弟情，康熙和韋⼩寶的主僕情……這些情感或上升到家國⼤義，或夾雜著個⼈私慾，涉及個⼈與個⼈、個⼈與國家、個⼈與社會關係的探討。 武俠⽂化以中華歷史⽂化為底蘊，⾦庸武俠作品中傳達出的美好情感具有普世價值。「東海西海，⼼裡攸同︔南學北學，道術未裂。」雖然各國⼈民在跨⽂化交流中不可避免會有⼀定的差異，但是⾦庸作品中諸如英雄情結、愛情、奉獻與執著等情感內核，所傳達的情感是共通的，⼈類共有的情感體驗使得⾦庸武俠跨越時空，調動起不同⽂化共感的情緒，形成武俠⽂化情感上的共情，並激發⼈們對武俠世界的想象。 （㆓）、傳「情」：㈮庸武俠㆗發掘的情感共性 在前述問卷基礎上，本研究繼續深入探究在華國際學⽣對⾦庸武俠中所傳達出的價值觀的認同程度，探析⾦庸武俠所承載的情感驅動與⽂化傳播意義。 （接前）問題 8.你喜歡下⾯哪句話（或幾句話）? Which of the following sentences do you like? （多選題） 序號 選項 小計 比例 1 為國為民，俠之大者。He that serves the country and the people is the greatest hero. "#$%&'() 48 42.86% 2 男子漢大丈夫，第一論人品心腸，第二論才幹事業，第三論文學武功。A real man is first judged by his character , his talent and career, and then his understanding of literature and martial arts.（《天龍八部》） 41 36.61% 3 你瞧這些白雲聚了又散，散了又聚，人生離合，亦復如斯。Look at these white clouds, they gather and disperse, and gather and disperse again. Life is like this.（《神雕俠侶》） 53 47.32%

樂琦、鄭博文 230 4 世上最寶貴之物，乃是兩心相悅的真正情愛，決非價值連城的寶藏。The most precious thing in the world is the true love between two hearts, not the priceless treasure.（《雪山飛狐》） 41 36.61% 5 只要有人的地方就有恩怨，有恩怨就會有江湖，人就是江湖。As long as there are people, there will be rights and wrongs. Where there are rights and wrongs, there will be jianghu. People are jianghu.（《笑傲江湖》） 39 34.82% 6 是非只因強出口，煩惱只為強出頭。Rights and wrongs arise from trying to stand out, and troubles arise from trying to stand out.（《笑傲江湖》） 22 19.64% 7 他強任他強，清風拂山崗，他橫由他 橫，明月照大江。Let him be strong, the breeze blows over the hills, let him be arrogant, the bright moon shines over the river.（《倚天屠龍記》） 30 26.79% 8 書到用時方恨少，肉到肥時方恨多。You only regret not having enough books when you need them, and you only regret having too much meat when you are fat.（《鹿鼎記》） 28 25% 圖表 10 受調者對於金庸作品經典語句的認知 本題選取了⾦庸武俠原著中部分經典語句，分別涉及個⼈追求、⼈⽣態度、情感觀念、為⼈處世等⽅⾯。調查結果顯⽰，國際學⽣的偏好依次為「你瞧這些⽩雲聚了又散，散了又聚，⼈⽣離合，亦復如斯。」（佔比 47.32%）、「為國為民，俠之⼤者。」（佔比 42.86%）、「男⼦漢⼤丈夫，第⼀論⼈品⼼腸，第⼆論才幹事業，第三論⽂學武功。」、「世上最寶貴之物，乃是兩⼼相悅的真正情愛，決非價值連城的寶藏。」。其中，排在⾸位的「你瞧這些⽩雲聚了又散，散了又聚，⼈⽣離合，亦復如斯。」出⾃《神雕俠侶》中多才多藝的女⼦程英之⼜，程英將對楊過的愛深埋於⼼底，由此發出對於⼈⽣離合的慨嘆。「為國為民，俠之⼤者」出⾃《神雕俠

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 231 侶》中的俠義之⼠郭靖之⼜，「為國為民」既是他對俠者所承擔的責任的解說，也是對後輩楊過的期盼。「男⼦漢⼤丈夫，第⼀論⼈品⼼腸，第⼆論才幹事業，第三論⽂學武功」是對個⼈的價值的追求，與中華傳統⽂化中的「修齊治平」相呼應。「世上最寶貴之物，乃是兩⼼相悅的真正情愛，決非價值連城的寶藏」出⾃《雪山⾶狐》的胡⼀⼑之⼜，體現的是年輕⼈的愛情觀，即真正的愛情勝過⼀切珍寶。以上語句中所傳達出的都是普遍意義上的正向情感。與此相對，《笑傲江湖》中的「是非只因強出⼜，煩惱只為強出頭」則是⼀句俗語，意思是好強⽽多管閒事往往會招惹是非，⽣出煩惱，勸⼈忍得⼀時之氣，明哲保身。從得票結果看，此話並不為⼤多數國際學⽣所接受。 可⾒，在華國際學⽣在追求個⼈發展、追尋真摯情感、實現社會價值這些觀念上與⾦庸武俠所傳達的價值觀是基本⼀致的。⾦庸作品中的書劍恩仇、兒女情長、與俠義江湖所承載的，是「為國為民，俠之⼤者」的情懷，和「⼈⽣離合，亦復如斯」的悲憫。 作為華語⽂化的瑰寶，⾦庸作品以其深厚的中華⽂化底蘊和民族情感深受全球華裔⼈群的喜愛。他們通過對⾦庸作品的共賞，從⽽產⽣強烈的共情，進⽽構築共話之橋。當然，顯⽽易⾒，對於非華裔⼈群⽽⾔，由於其母語不是中⽂，很可能是通過翻譯作品接觸⾦庸⼩說，因此，語⾔障礙、跨⽂化問題會影響他們對⾦庸作品的接受程度，可能無法達到在華裔讀者、甚至在華國際學⽣⼼中的深度和廣度，但是他們對於作品中關於真、善、美的追求是⼀致的。 五、共話：新秩序的追求與新世界的想象 本尼迪克特·安德森曾提出，民族主義「是⼀種想象的政治共同體，並且，它是被想象為本質上有限的、同時享有主權的共同體。」15 這種共同體之所以是想象的，是因為屬於這個民族的成員並不能認識全部的同胞，同胞主要借由意象存在於成員的內⼼中。安德森基於其所在的社會發展狀況，認為民族作為想象的共同體的主要實現途徑是印刷技術，⽽民族觀念正是借助種種現代媒介技術，塑造著公共空 15 本尼迪克特·安德森，吳數人譯，《想象的共同體：民族主義的起源與散布》，（上海：上海人民出版社，2003），6。

樂琦、鄭博文 232 間中的想象共同體。全球化背景下，「想象的共同體」概念在現代社會獲得了廣泛呼應。不同⽂化背景的⼈們在交流和互動中，由於共同的價值觀、審美、興趣⽽跨越地理界限，形成虛擬的共同體。當然，任何「共同體」都是個體想象和建構，進⽽在⼼理上形成的⼀種聯結，即⽂化認同。對於⾦庸武俠作品⽽⾔，其蘊含的，在共賞、共情、共話的過程中，構築了跨⽂化傳播的橋梁，跨越了不同民族和⽂化，拓展了跨媒介傳播的渠道，其內涵和外延不斷豐富。 （㆒）、㈮庸武俠構築跨文化傳播橋梁 判斷⼀個跨⽂化共同體形成與否，關鍵在於個體對所屬的全球性趣緣團體是否產⽣了跨⽂化認同。⾦（Kim）通過整合已有的跨⽂化適應及身份認同理論，明確提出了「跨⽂化認同」（intercultural identity）這⼀概念。16 她認為，個體經歷跨⽂化轉型的後果之⼀即產⽣跨⽂化的認同感。具體可分為兩個階段，⼀是個體在適應陌⽣⽂化環境的過程中，通過與⽂化他者的互動習得全⾯的跨⽂化交際能⼒︔⼆是在此基礎上，個體能夠主動理解⽂化間的差異性以及每⼀種⽂化的相對價值，從⽽形成⼀種包容並蓄、願意與不同⽂化的成員友善共處的超越性身份認同。 ⾦庸武俠在近七⼗年的傳播過程中，曾在世界各地多次掀起「⾦庸熱」，吸引了海內外眾多「⾦迷」。其筆下的江湖有豪情壯志，也有纏綿悱惻，成為眾多讀者⼼中嚮往的武俠世界。他在《⾦庸作品集》新序（2002）中曾提及⾃⼰作品希望傳達的主旨：「愛護尊重⾃⼰的國家民族，也尊重別⼈的國家民族︔和平友好，互相幫助︔重視正義和是非，反對損⼈利⼰︔注重信義，歌頌純真的愛情和友誼︔歌頌奮不顧身的為了正義⽽奮⾾︔輕視爭權奪利、⾃私可鄙的思想和⾏為。武俠⼩說並不單是讓讀者在閱讀時做『⽩⽇夢』⽽沈湎在偉⼤成功的幻想之中，⽽希望讀者們在幻想之時，想象⾃⼰是個好⼈，要努⼒做各種各樣的好事，想象⾃⼰要愛國家、愛社會、幫助別⼈得到幸福，由於做了好事、做出積極貢獻，得到所愛之⼈的欣賞和傾⼼。」17 國際學⽣閱讀⾦庸武俠⼩說的過程，同時也是瞭解中國歷史⽂化、促 16 Kim Y Y, Becoming Intercultural：An Integrative Theory of Communication and Cross-Cultural Adaptation. (Thousand Oaks：Sage Publications:2001): 192-194. 17 金庸，《金庸作品集（新序）》，（廣州：廣州出版社，2002），3。

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 233 進跨⽂化理解的過程。 在媒介全球化時代，擁有共同喜好的⼈較以往⽽⾔，能突破傳統地⽅性的限制，更加便捷地找尋到彼此，聚合成「跨⽂化粉絲圈」。借助影視等媒介的傳播推動，「⾦迷」演化為⼀個全新意義上的跨⽂化互動體系。當今的中國正秉持⼈類命運共同體理念，堅守和平、發展、公平、正義、民主、⾃由的全⼈類共同價值，與世界各國建⽴平等⽽密切的夥伴關係，包括⾦庸武俠⼩說在內的諸多華⽂⽂學作品是華⽂⽂學、中華⽂化⾛向世界的橋梁，為外國瞭解中國打開⼀扇視窗。 （㆓）、㈮庸武俠拓展跨媒介傳播渠道 從⽂本到影像的流動，最早可以追溯到電影《南茜·塞克斯之死》，這部電影改編⾃英國作家狄更斯的⼩說《霧都孤兒》，作為⼩說中最耀眼的女性形象，南茜從⽂本⾛進了⼤屏，受到了更多的關注。18 在世界範圍內，從⼩說⽂本到影像的改編，⼀直是持續不斷又引⼈矚⽬的⽂化現象。但是，影像藝術持續青睞⼀個作家，並將其作品不斷進⾏視覺化呈現，並不多⾒。⾦庸武俠⼩說⾃誕⽣之時起就被影視藝術界不斷改編、翻拍、再創造。作為華⼈世界最暢銷的作家之⼀，⾹港第⼀部新派武俠⼩說雖然並非⾦庸寫就，但是他的武俠⼩說卻最早獲得電影開發商的青睞。⾃ 1958年第⼀部改編⾃⾦庸武俠⼩說的⿊⽩電影《射雕英雄傳》誕⽣至今，⾦庸武俠⼩說與各種傳播媒介相結合，形成了⼤眾⽂化中經久不衰的 IP（Intellectual Property）。 由於⾦庸武俠⼩說連載的特殊時空背景，早期的影視改編、錄制多集中在⾹港地區，無法完全具備再現⼩說中描繪的地理風景、地⽅風物的條件。直到《書劍恩仇錄》（1987）才實現到中國內地取景拍攝。影⽚中，陳家洛與乾隆相交於杭州西湖，與⾹⾹公主相戀於回疆（屬今新疆）……桃花島、風陵渡、少林寺、襄陽城……⾦庸武俠中涉及了⼤量實際存在的⾃然與⼈⽂景觀，因此，觀眾在閱讀⽂字的同時，也可以透過影視劇欣賞，亦可實地遊覽，更加真切地領略中國各地區獨特的風⼟⼈情，加深對中國不同地域⽂化的瞭解。 18 約翰·Ｍ．德斯蒙德、彼得·霍克斯，李升升譯，《改編的藝術：從文學到電影》，（北京：世界圖書出版公司，2016），110。

樂琦、鄭博文 234 視聽傳播時代，⽂字已然融入⼀個「⽴體和智能的視頻語⾔體系」，視頻傳播將成為國際傳播的流量⽣產⼒和秩序轉換者。 ⽂學借助包括影視在內的各類媒介可隨時隨地為⼈所知，線上、線下的多⽅聯動讓武俠⽂化愈發清晰可感，這也使⾦庸武俠成為中國⽂學、⽂化國際傳播中的⼀道靚麗風景線。當下，⾦庸武俠作品在契合⼤眾娛樂審美需求的同時，展⽰了世⼈共求的崇⾼⼈格、情操與理想，甚至吸引包括在華國際學⽣在內的外國⼈⾃發開啓從影像到⽂字的逆向閱讀，形成國際中⽂傳播的正向循環。 六、 結語與展望 ⾦庸的武俠⽂學創作開闢了對後來者影響深遠的「新武俠⼩說」，其作品也隨著多種媒介的傳播為越來越多的外國友⼈喜聞樂⾒。本研究從「共情」「共賞」「共話」三個維度，提出國際中⽂傳播視域下建構⾦庸武俠跨⽂化共同體的可能，發掘⾦庸武俠⼩說的獨特內涵與傳播價值，揭⽰其作為華⽂⽂學、中華⽂化載體進⾏跨⽂化傳播的有效路徑。 要⽽⾔之，⾦庸為讀者們奉獻了⼤量富於⽂化內涵、美好情感和充滿奇異想象的優秀作品，賦予華⽂⽂學新的語⾔魅⼒和⽂化價值，也促進了不同⽂化背景讀者的跨⽂化認同，凝聚成具有武俠情結的共同體。這種⽂學、⽂藝、⽂化的交流，有助於民族國家間擺脫意識形態的偏⾒，從⽽最⼤程度地增強合作機制、理念、政策的開放性和包容性，實現從跨⽂化到轉⽂化的新型傳播路徑，拓展中華⽂化國際傳播的實踐進路。

國際中文傳播視域下的金庸武俠研究：共賞 共情 共話 235 徵引書目 ㆒、㆗文著作 [1] 吳應輝，〈國際中⽂教學學科建設及漢語國際傳播研究探討〉，《語⾔⽂字應⽤》，第 3期（北京：2010.08），35-42。 [2] 謝⽂興，〈論⾦庸武俠⼩說的外國印跡〉，《寧波廣播電視⼤學學報》，第 12卷第 2期（寧波：2014.06），23-27。 [3] 彭海雲，〈跨⽂化視域下⾦庸⼩說的藝術創新〉，《嘉興學院學報》，第 28卷第 1期（嘉興：2016.01），20-24。 [4] 徐雪英、張菁，〈從⾦庸《射雕英雄傳》英譯看中國⽂化如何⾛向世界〉，《浙江學刊》，第 3期（杭州：2020.05），42-53。 [5] 湯哲聲，〈中國百年武俠⼩說的價值評估與俠⽂化的現代構建〉，《⽂藝理論研究》，第 4期（上海：2023.07），182-193。 [6] 李瑋，〈「盛世江湖」與漫長的「九⼗年代」—從⾦庸、「後⾦庸」到純武俠的衰落〉，《⼩說評論》，第 1期（西安：2023.01），62-69。 [7] 原⼩平，〈改編的通俗性與⽂學經典的建構〉,《東北⼤學學報》，第 1期（瀋陽：2010.01），79-84。 [8] 張欣，〈論⾦庸⼩說的三種⽂化現象〉，《雲南⼤學學報（社會科學版）》，第 2期（昆明：2016.03），103-110。 [9] 李歐梵，〈⾦庸與武俠⼩說傳統〉，《明報⽉刊》，（⾹港：2018.12），111。 [10] 巴赫⾦ M.M. Bakhtin 、錢中⽂（譯），〈⾺克思主義與語⾔哲學〉，《巴赫⾦全集》第⼆卷，（⽯家莊：河北教育出版社，1998），448。 [11] 王曉明、王海渭、張寅彭編，《胡河清⽂存》，上海：三聯書店，1996，199。 [12] 劉海明、宋婷，〈共情傳播的量度:重⼤公共衛⽣事件報道的共振與糾偏〉 ，《新聞界》，第 10期（成都：2020.10），11-21+31。 [13] 樂琦、陳晨，〈國際中⽂教育品牌的實踐進路〉，《陝西師範⼤學學報》（哲學社會科學版），第 6期（西安：2024.12），29-39。

樂琦、鄭博文 236 [14] 本尼迪克特·安德森，吳數⼈譯，《想象的共同體：民族主義的起源與散布》，上海：上海⼈民出版社，2003，6。 [15] ⾦庸，《⾦庸作品集（新序）》，廣州：廣州出版社，2002，3。 [16] 約翰·Ｍ．德斯蒙德、彼得·霍克斯，李升升譯，《改編的藝術：從⽂學到電影》，北京：世界圖書出版公司，2016，110。 [17] 廖祥忠，〈視頻天下：語⾔⾰命與國際傳播秩序再造〉，《現代傳播（中國傳媒⼤學學報）》，第 1期（2022.01），1-8。 ㆓、西文著作 [1] Kim Y Y, Becoming Intercultural：An Integrative Theory of Communication and Cross-Cultural Adaptation. Thousand Oaks：Sage Publications, 2001, 192-194. 作者簡介 樂琦，中國傳媒大學漢語國際教育中心教授、博士生導師，研究方向為國際中文傳播、國際中文教育、跨文化傳播、世界華文文學與華文教育 通訊電郵：paoapaoqq@163.com 通訊作者簡介 鄭博文，中國傳媒大學人文學院博士生，研究方向為世界華文文學與跨文化傳播、國際中文教育 通訊電郵：ysddz1120@163.com

澳門科技大學學報(人文社會科學版)第十九卷第一期 世漢學會國際中文教育研究專欄 二零二五年三月，頁 237-261 DOI: 10.58664/mustjournal.2025.01.008 237 隱性教學對英語母語者習得語氣詞「吧」的影響* 汪敏鋒、姚鈺銀 （福建師範大學海外教育學院副教授） 摘要：語氣詞「吧」一詞多能，屬於意義複雜的語法結構，其內部功能類型又存在較大的難度差異，在二語教學實踐中，往往進行無元語言規則講解的隱性教學。基於這一情況，本研究遵循 Housen & Pierrard（2005）隱性教學的 6 個特徵，採用無元語言規則講解的隱性輸入教學方式對 4 名英語母語者被試進行為期半年的跟蹤實驗，實驗包括基於情景導向的主題訪談式隱性教學輸入和情景再現的自然對話測試。實驗結果表明：（1）以情景為導向的隱性教學對意義複雜的語氣詞「吧」的習得是有效的，「吧」的運用能力總體呈非線性提升；（2）不同水準的英語母語者隱性教學的效果存在差異性，隱性教學對語氣詞「吧」不同功能項的影響也具有選擇性。 關鍵字：隱性教學、語氣詞、「吧」、習得效果 ! * 收稿日期：2025年 01月 04日；通過日期 2025年 02月 17日。本文為國家社科一般項目「面向國際中文教育的漢語語氣成分研究及知識庫建設」（24BYY047）的階段性成果。曾在第 14 屆國際中文教育大會（2021年 12月 18日，天津）上報告，袁博平教授提出寶貴的指導意見，特此致謝。

汪敏鋒、姚鈺銀 238 The Effect of Implicit Instruction on the Acquisition of the Modal Particle Ba Wang, Minfeng1; Yao, Yuying (1. Associate Professor, International College of Chinese Studies, Fujian Normal University) Abstract: The modal particle ‘ba’ is a multi-functional grammatical structure with complex meanings; the difficulty of its internal functional types varies greatly. The modal particle ‘ba’ is frequently taught in second-language teaching using implicit teaching, which omits the explanation of metalinguistic principles. Based on this circumstance, we conducted a trace experiment on four native English speakers using a large number of implicit input teaching without elaborating on meta-linguistic principles, adhering to the six features of implicit instruction identified by Housen & Pierrard (2005). The experiment is divided into two parts: situation-oriented thematic interview implicit teaching input and scene reproduction natural dialogue test. The findings reveal that: (1). Situation-oriented implicit teaching is successful for the acquisition of the modal particle ‘ba’ with complex meaning, and the learners' application capacity of ‘ba’ has improved nonlinearly. (2). The effect of implicit teaching is different for native English speakers with varying Chinese language proficiency, In addition, the influence of implicit teaching on each functional item of modal particle ‘ba’ is also diverse. Keywords: Implicit teaching; Modal particle; ‘Ba’; Acquisition effect

隱性教學對英語母語者習得語氣詞「吧」的影響 239 一、引言 語氣詞是漢語⼜語中使⽤頻率⾼、個性又強的⼀類多功能詞，能否得體使⽤語氣詞是衡量漢語學習者漢語⽔準的重要指標，漢語語氣詞的運⽤能⼒是漢語語⽤能⼒的重要組成部分，與漢語習得效率和漢語交際能⼒密切相關。當前，母語者的語氣詞研究成果頗為豐富，與之形成鮮明相比的是，漢語⼆語學習者的語氣詞功能系統研究顯得⼗分不⾜。已有研究為數不少是建⽴在仲介語語料庫基礎上的偏誤分析，基於調查的習得順序研究，也有⼀部分是從本體角度對學習者仲介語中某（幾）個語氣詞進⾏考察，1 ⼤多採⽤的還是語⾔學的描寫法、對比法，⽅法設計的恰當性和有效性留有爭議，2 ⽽且針對語氣詞獲得⽅式的動態跟蹤研究和習得的認知研究還沒引起⾜夠重視，尤其是在漢語教學界，鑒於語氣詞是不影響資訊傳遞的「⼩詞」，⽽且意義空靈，難以把握，難以描寫，在⼆語教學實踐中往往採取隱性教學或者弱化模糊的處理⽅式，但是隱性教學效果到底怎麼樣，它對「形式簡單但功能複雜」的語氣詞習得過程會產⽣哪些影響，學界還缺少相關的實證研究成果。近⼗年來，實證性的量化研究、實驗研究⽇益受到漢語教學界的重視，以漢語習得資料為出發點的動態實證研究將是漢語學習者語氣詞動態習得研究的必然⽅向。3 本研究以互動語氣詞「吧」的 6 個典型功能項為語⾔觀察點，採⽤微變化研究法（Micro-genetic method）在半年內⾼頻率對英語母語學習者進⾏情景導向的隱性教學輸入，觀察學習者漢語語氣詞「吧」隱性習得變化的動態過程，通過實驗實證檢驗隱性教學的效果和「吧」獲得的動態認知軌跡。 1 董妍，《對外漢語教學中的語氣助詞教學》，黑龍江大學，2011；湯茗儀，《對外漢語常用語氣詞教學策略研究》，蘇州大學，2018。 2 馮麗萍、孫紅娟，〈第二語言習得順序研究方法述評〉，《語言教學與研究》，第 1期（2010），9-16。 3 郝美玲，〈留學生漢字正字法意識的萌芽與發展〉，《世界漢語教學》，第 1期（2007），29-39+2-3；靳洪剛，〈任務複雜度及其互動、輸出效應〉，《Journal of the Chinese Language Teachers Association》，第 2 期（2010），101-135；陳默、王建勤，〈漢語聲調教學的實驗和電腦類比研究〉，《語言教學與研究》，第 1 期（2011），11-16；靳洪剛、章吟，〈對外漢語糾錯回饋研究的單位界定及過程分解分析〉，《Journal of the Chinese Language Teachers Association》，第 3期（2014），87-116；袁博平，〈計算複雜性與第一語言遷移—以漢語第二語言態度疑問句為例〉，《世界漢語教學》，第 1期（2017.31），85-104。

汪敏鋒、姚鈺銀 240 二、關於隱性教學 隱性及相關概念的研究最早來⾃認知⼼理學，後來被⼆語習得領域所接受。4隱性教學相對於顯性教學，是⼈類雙認知系統中的⼀個重要系統，⼀直是⼼理學界、腦神經科學界和⼆語習得領域共同關注的重點問題之⼀。5 在⼆語教學有效性⽅⾯，⽬前的研究有單獨研究並證實顯性教學的有效性的︔6 也有對比隱性顯性教學有效性的，這有三種結論：⼀種是實驗證實顯性教學效果優於隱性教學效果︔7 ⼀種是隱性教學效果好於顯性教學效果︔8 還有⼀種是各有千秋，都有效果，不過教學效果與語法複雜度及複雜類型有關。例如Robinson以「主謂倒裝」和「假性嵌套句」為例進⾏對比實驗，結果表明隱性教學在複雜結構上比顯性教學有優勢，顯性教學並不是都有效，具有選擇性，9 wang以形式複雜度較⾼的「定語從句」和意義複雜度較⾼的「不/沒有」進⾏實驗，證明隱性教學和顯性教學對「不/沒有」都有效，但隱性教學對「定語從句」無效，10 Liao對比隱性顯性教學對致歉功能的影響 4 Reber A S., “Implicit learning of artificial grammars.” Journal of Verbal Learning and Verbal Behavior 6.6 (1967): 855-863. 5 Reber A S., “Implicit learning of artificial grammars.” Journal of Verbal Learning and Verbal Behavior 6.6 (1967): 855-863; Anderson J R, Lebiere C J., The atomic components of thought. Psychology Press, 1998; Wallach D, Lebiere C., “Conscious and unconscious knowledge: Mapping to the symbolic and subsymbolic levels of a hybrid architecture.” (2003); Hazeltine E, Ivry R B., “Neural structures that support implicit sequence learning.” (2003); Ellis R. 2., Measuring Implicit and Explicit Knowledge of a Second Language//Implicit and explicit knowledge in second language learning, testing and teaching. Multilingual Matters, 2009: 31-64; Richards J C, Schmidt R W. Longman dictionary of language teaching and applied linguistics. Routledge, 2013; 戴曼純，〈二語習得的「顯性」與「隱性」問題探討〉，《外國語言文學》，第 2 期（2005），101-111；靳洪剛、侯曉明，〈漢語作為第二語言實證研究縱觀：顯性與隱性學習、知識、教學〉，《世界漢語教學》，第 3期（2016.30），379-400。 6 Harley B., “Functional grammar in french immersion: A classroom Experiment1.” Applied linguistics 10.3 (1989): 331-360; Taft M, Chung K., “Using radicals in teaching Chinese characters to second language learners.” Psychologia 42.4 (1999): 243-251; Spada N, Jessop L, Tomita Y, et al., “Isolated and integrated form-focused instruction: Effects on different types of L2 knowledge.” Language Teaching Research 18.4 (2014): 453-473; 孫曉明，〈任務為導向的留學生伴隨性詞彙學習研究〉，《漢語學習》，第 4期（2012），91-96。 7 Kuo L H., Improving implicit learning and explicit instruction of adult and child learners of Chinese. Brigham Young University, 2013. 8 洪煒，〈漢語作為第二語言的近義詞教學實驗研究〉，《世界漢語教學》，第 3期（2013），424-432。 9 Robinson P., “Learning simple and complex second language rules under implicit, incidental, rule-search, and instructed conditions.” Studies in second language acquisition 18.1 (1996): 27-67. 10 Wang J. The effect of implicit vs. Explicit instruction on learning form-based vs. Meaning-based

隱性教學對英語母語者習得語氣詞「吧」的影響 241 實驗，顯⽰「教學有明顯效果，但隱性和顯性教學組無統計差別」，11 Peng以簡單語法結構「是非問句」和複雜結構「了」為例，卻證明「顯性教學的即時效果優於隱性教學，且對相對複雜的語法結構更有優勢」。12 可以看出，對於意義複雜的語法結構，例如「了」「不/沒」等隱性教學的有效性還沒有形成統⼀結論。造成這樣局⾯的原因，可能是與顯性和隱性教學的區分不夠清楚、效果量化測量不夠科學統⼀有關。13 如何厘定、區分和設計隱性教學事關實驗結果的科學性。隱性教學是通過教學設計吸引（attract）學習者將「注意」（attention）投入到⽬標成分上，説明學習者下意識地⾃主習得語⾔規則。14 其中輸入加強（text enhancement）是常⽤的隱性教學實驗⽅法，通過⽬的知識的「⼤劑量輸入（input flooding）」⾼頻重複和⼤量範例吸引學習者注意和學習。同樣，「意識」和「注意」在隱性教學區分中起著重要的鑒別作⽤。⽬前學界常⽤的隱性教學法主要有交際法和任務法（ task-based teaching）。15 Housen & Pierrard曾⽴⾜語⾔形式教學（Form-Focused Instruction，FFI)指出隱性教學的 6個特徵：（1）吸引注意資源到⽬標結構的應⽤︔（2）與交際活動同步的隨機性教學︔（3）以不打斷交流為原則，對意義交際的⼲預為最⼩限度︔（4）語境中含有⽬標結構但不突出︔交際場合為重，結構使⽤為輔︔（5）不使⽤如規則講解等任何元語⾔解釋︔（6）⿎勵⽬標結構的⾃由使⽤。16 語氣詞「吧」⼀詞多能，屬於意義複雜的語法結構，其內部功能類型又存在較⼤的難度差異，本⽂將遵循Housen & Pierrard隱性教學的 6個特徵，採⽤⼤劑量、無 language features. University of Pittsburgh, 2014. 11 Wang J. The effect of implicit vs. Explicit instruction on learning form-based vs. Meaning-based language features. University of Pittsburgh, 2014. 12 Peng F., “The Effectiveness of Explicit Instruction Versus Implicit Instruction Method on Chinese Grammar Acquisition.” (2015). 13 靳洪剛、侯曉明，〈漢語作為第二語言實證研究縱觀:顯性與隱性學習、知識、教學〉，《世界漢語教學》，第 3期（2016.30），379-400。 14 Ellis R., “Learning a second language through interaction.” Learning a Second Language through Interaction (1999): 1-295. 15 靳洪剛、侯曉明，〈漢語作為第二語言實證研究縱觀：顯性與隱性學習、知識、教學〉，《世界漢語教學》，第 3期（2016.30），379-400。 16 Housen A, Pierrard M., “Investigating instructed second language acquisition.” Investigations in instructed second language acquisition (2005): 1-27.

汪敏鋒、姚鈺銀 242 元語⾔規則講解的隱性輸入教學⽅式對 4名英語母語者被試進⾏實驗。17 實驗分為兩個部分，第⼀部分是基於情景導向的主題訪談式隱性輸入教學，情景導向可以吸引被試將注意資源投入⽬標語氣詞「吧」上︔第⼆部分是隱性教學後各有⼀次情景再現的⾃然對話測試，在⼀定時間週期內，跟蹤記錄 6次。擬解決的問題是： 問題 1：英語母語者語氣詞「吧」隱性教學的學習效果及動態認知軌跡如何?難度等級不同的功能項呈現哪些習得差異? 問題 2：漢語熟練度與「吧」的隱性學習效果之間呈何種關係?是不是漢語⽔準越⾼，隱性輸入的學習效果越好? 研究設計 2.1 觀察對象 本⽂ 4 名被試是母語為英語的男性漢語學習者，年齡在 25-35 歲之間，美國 1名、英國 2名、澳⼤利亞 1名，都在中國⼯作⽣活，處於⽬的語環境中。被試確定標準：（1）以在⽬的語環境下⾃然習得為主，未接受系統完整的顯性教學︔（2）在前測訪談中無法陳述使⽤語氣詞「吧」的規則或知識。根據評估，我們將被試分為中級組（Lz）和⾼級組（Lg），各 2名，中級組在中國旅居時間 1-2年︔⾼級組在 3年以上。具體資訊如表 1： 表1 被試具體資訊 信息 被試 L1 愛倫 L2 麥克 L3大牛 L4 威廉 國籍 英國 澳大利亞 英國 美國 在華時間 2020.2-2021.5 2020.1-2021.2 2019.1-2021.5 2017.8-2021.6 漢語水準 中級（Lz） 中級（Lz） 高級（Lg） 高級（Lg） 職業 外貿公司職員 調酒師 博主 外企高管 17 Housen A, Pierrard M., “Investigating instructed second language acquisition.” Investigations in instructed second language acquisition (2005): 1-27.

隱性教學對英語母語者習得語氣詞「吧」的影響 243 2.2 實驗材料 實驗材料選取語氣詞「吧」六個功能項，並根據使⽤頻率和語境依賴程度分為⼀⼆三級：⼀級為使⽤頻率最⾼，最典型的語氣詞功能項，對語境依賴程度較低,包括F1「揣測」和F2「求證」︔ ⼆級為使⽤頻率次⾼，次典型的語氣詞功能項，對語境依賴程度較⾼，包括F3「建議」、F4「⿎勵」、F5「勉強同意、認同」等︔三級為使⽤頻率較⾼，較典型的語氣詞功能項，對語境依賴程度⾼，主要指「吧」的⼈際意向功能：F6「緩和」。18 詳⾒表 2。 表2 實驗材料 功能項 話輪位置 語類 例句 F1 語境依賴度較低（一級功能） 揣測 應答話輪 陳述句 大概有 10斤吧。 F2 求證 起始話輪 疑問句 有一年吧？ F3 語境依賴度較高（二級功能） 建議行為 起始/應答 祈使句 晚上七點半去吧。 F4 鼓勵行為 起始/應答 祈使句 你說吧。 F5 勉強同意認同 應答話輪 祈使句 好吧。 F6 語境依賴度高（三級功能） 緩和 起始-應答 祈使句 陳述句 讓我先說吧。 她這個人吧，喜歡錢。 2.3 實驗過程 採⽤微變化研究法，⾃ 2020年 12⽉ 6⽇至 2021年 6⽉ 17⽇對觀察對象進⾏6 次以周或⽉為間隔單位的主題式線上訪談錄⾳，創設「吧」各功能運⽤的典型語境，通過情景導向訪談進⾏隱性輸入，全過程無元語⾔規則講解。材料舉例⾒表 3。 表3 實驗句示例 功能類型 情景導向 隱性輸入目標句 物件水準 時間/次數 F1揣測 ①揣測情景 ②揣測構式或詞彙標記 ③應答話輪 A：這個月胖了多少？ B：大概五六斤吧。 中級水準 2020.12.15（第 1次） ①揣測情景 ②揣測構式或詞彙標記 ③應答話輪 A：你在中國這麼多年，有考慮在中國發展下去嗎？ B：還不知道，可能吧。 高級水準 2020.12.14（第 1次） 18 汪敏鋒，《語氣詞的人際語用功能研究》，中國社會科學出版社，2022。

汪敏鋒、姚鈺銀 244 訪談結束後通過情景線索再現，通過提供語氣詞啟⽤的語境線索，要求被試在20分鐘內完成⾃然對話測試，6個功能項各 3次為⼀組，6次共 18組。按句長、句中其他詞語的難度（均控制在甲級和⼄級詞彙中）以及⽬標詞上下⽂語境的強弱等因素進⾏了匹配。然後將這些句⼦隨機呈現。句義準確但遺漏語氣詞「吧」計 1分，準確使⽤語氣詞「吧」計 2分，不正確計 0分。 三、實驗資料 3.1 語氣詞「吧」各功能㊠使用描述統計結果 四位英語母語者（L）經過 6 次隱性輸入後，在⾃然對話測試中，「吧」的各功能項得分描述統計結果⾒圖 1和圖 2。 012345 52 24 41 1324320 0 0 0 00123456前測 第一次 第二次 第三次 第四次 第五次 第六次得分數測驗次數L1揣測 求證 建議 鼓勵 勉強同意、認同 緩和0 0 024 450123456前測 第一次 第二次 第三次 第四次 第五次 第六次得分數測驗次數L2揣測 求證 建議 鼓勵 勉強同意、認同 緩和

隱性教學對英語母語者習得語氣詞「吧」的影響 245 圖1 被試各項功能習得表現 以F1「揣測」功能為例，採⽤單因素重複測量⽅差分析⽅法，判斷在實驗週期內隱性輸入對研究物件 6個⽅⾯得分情況的影響。可以通過主體效應檢驗，發現第1 次成績到第 6 次成績的差異具有統計學意義，由於不滿⾜球形假假定，讀取校正34 4 45652126425 541413 30 02 2 201234567前測 第一次 第二次 第三次 第四次 第五次 第六次得分數測驗次數L3揣測 求證 建議 鼓勵 勉強同意、認同 緩和2 2345 5434 464512 2630 0 0201234567前測 第一次 第二次 第三次 第四次 第五次 第六次得分數測驗次數L4揣測 求證 建議 鼓勵 勉強同意、認同 緩和

汪敏鋒、姚鈺銀 246 後F（2.523，7.568）=14.163，P<0.001。 表4 「揣測」功能的描述統計19 平均值 標準差 個案數 揣測 1 2.00 1.414 4 揣測 2 2.25 1.708 4 揣測 3 3.25 .957 4 揣測 4 4.50 .577 4 揣測 5 5.25 .500 4 揣測 6 4.25 .957 4 表5 「揣測」功能的主體內效應檢驗 源 III 類平方和 自由度 均方 F 顯著性 time 假設球形度 33.833 5 6.767 14.163 .000 格林豪斯-蓋斯勒 33.833 2.523 13.413 14.163 .002 辛-費德特 33.833 5.000 6.767 14.163 .000 下限 33.833 1.000 33.833 14.163 .033 誤差（time） 假設球形度 7.167 15 .478 格林豪斯-蓋斯勒 7.167 7.568 .947 辛-費德特 7.167 15.000 .478 下限 7.167 3.000 2.389 具體通過LSD的多重比較可以看出，第 4 次和第 1 次的成績相差 2.5 分，顯著性P值為 0.05，第 5次和第 1次的乘積相差 3.25分，顯著性P值為 0.007，第 6次和第 1次的乘積相差 2.25分，顯著性P值為 0.037。 19 為了節省篇幅，其他各功能具體的描述統計不再呈現。

隱性教學對英語母語者習得語氣詞「吧」的影響 247 表6 成對比較 （I）time （J）time 平均值差值（I-J） 標準誤差 顯著性 b 差值的 95% 置信區間 b 下限 上限 1 2 -.250 .479 .638 -1.773 1.273 3 -1.250 .479 .080 -2.773 .273 4 -2.500* .500 .015 -4.091 -.909 5 -3.250* .479 .007 -4.773 -1.727 6 -2.250* .629 .037 -4.252 -.248 為了得到更可靠的組間比較結果，我們逐⼀對各功能進⾏「估算邊際平均值」（EMMs）計算，結果總體呈現如下： 1 ….092 ns.016 *.012 * .001 **.012 *1.002.003.004.005.006.001 3 5估算邊際平均值time「求證」的估算邊際平均值1 ns.638 ns.080 ns.015 *.007 **.037 *2.003.004.005.006.001 3 5估算邊際平均值time「揣測」估算邊際平均值1 ns.080 ns.016 *.003 **.005 **.002 **0.001.002.003.004.005.001 3 5估算邊際平均值time「建議」的估算邊際平均值1 ns.182 ns.182 ns .0 ***.005 **.007 **1.002.003.004.005.006.001 3 5估算邊際平均值time「鼓勵」的估算邊際平均值

汪敏鋒、姚鈺銀 248 圖 2 「吧」各功能項邊際平均值統計 四位被試 6次隱性學習後測結果顯⽰，在實驗週期內，語氣詞「吧」的運⽤能⼒都有了明確提⾼，平均值顯⽰「吧」的各功能項的習得效果也呈總體上升態勢。具體來說，6次訪談中，受試者F1揣測功能的成績逐步提升，⼀直到第 5次測試成績達到最⾼，平均分值為 5.25 分。F2 求證功能⼀直到第 4 次測試成績達到最⾼，平均分值為 4.5分，之後幾次測試平均成績穩定，保持在 4.25分。F3建議功能到第5次測試成績達到最⾼，平均分值為 4.25分，⽽第 6次測試中成績出現下滑，平均分值為 4分。被試F4⿎勵功能成績先抑後揚，在第⼆次測試中有所下滑，平均分值為 1分，到第 3次測試以後逐步上升，⼀直到第 6次測試成績達到最⾼，平均分值為 4.75分。F5勉強同意和F6緩和功能雖有反復，但邊際值總體向好。 3.2 語氣詞「吧」各㊠功能的變化軌跡 採⽤單因素重複測量⽅差分析⽅法，判斷 70天週期內的隱性輸入對研究物件語氣詞「吧」各功能項習得情況的影響。 nonenone.137 ns.050 *.184 ns.050 *0.000.501.001.502.002.501 3 5估算邊際平均值time「緩和」的估算邊際平均值1 ns.801 ns.157 ns.263 ns.263 ns.289 ns2.502.702.903.103.303.503.701 3 5估算邊際平均值time「勉強同意」的估算邊際平均值

隱性教學對英語母語者習得語氣詞「吧」的影響 249 表7 「吧」各項功能的配對檢驗 配對差值 序號 項目 平均值 標準差 T值 自由度 效應量 Sig 顯著性水準 配對 1 揣測 6 — 揣測 1 2.250 1.258 3.576 3 1.79 .037 顯著 配對 2 求證 6 — 求證 1 3.500 1.291 5.422 3 2.71 .012 顯著 配對 3 建議 6 — 建議 1 2.750 .500 11.000 3 5.50 .002 顯著 配對 4 鼓勵 6 — 鼓勵 1 3.250 .957 6.789 3 3.40 .007 顯著 配對 5 勉強同意 6 — 勉強同意 1 1.000 1.414 1.414 3 0.71 .252 不顯著 配對 6 緩和 6 — 緩和 1 2.000 1.633 2.449 3 1.22 .092 不顯著 雖然隱性學習對語氣詞「吧」的習得有積極的幫助作⽤，但學習者各功能類型運⽤能⼒的發展路徑呈現差異：⾸先，各功能類型測試成績達到最⾼值的時間不同。F5 在第 3 次測試時成績達到最⾼，之後出現了反復，F2、F6 在第 4 次測試時成績達到最⾼，F1、F3 在第 5 次測試時成績達到最⾼，⽽F4 在第 6 次測試時成績才達到最⾼。其次，各功能類型測試成績並不是直線上升的，反復、下滑現象時有發⽣。 通過主體效應檢驗，F2 求證功能主體內效應讀取校正後F（1.682，5.045）=11.219，P<0.001，求證功能配對顯著性P值為 0.012<0.05，說明第 6次訓練後求證得分與第 1次訓練後求證得分存在顯著的統計學差異，從得分來看第 6次訓練後求證得分⾼於第 1次訓練後求證得分。從發展軌跡來看，第 3次和第 1次的成績相差2分，顯著性P值為 0.016︔第 4次和第 1次的顯著性P值為 0.012，第 5次和第 1次的顯著性P值為 0.001，第 6次和第 1次的顯著性P值為 0.012，「吧」的F2求證功能在第三次隱性學習後率先習得。F3建議功能主體內效應讀取校正後F（1.793，5.378）=10.857，P<0.001，建議功能配對顯著性P值為 0.002<0.05，具體通過LSD的多重比較可以看出，第 3次和第 1次的成績相差 2分，顯著性P值為 0.016︔第 4次和第 1次的成績相差 2.5分，顯著性P值為 0.003，第 5次和第 1次的成績相差 3分，顯著性P值為 0.005，第 6次和第 1次的成績相差 2.75分，顯著性P值為 0.002。F4⿎勵功能主體效應檢驗，讀取校正後F（2.000，6.005）=25.711，P<0.001，⿎勵功能配對顯著性P值為 0.007<0.05，第 1次成績到第 6次成績的差異顯著，具體通過LSD的多重比較可以看出，第 5次和第 1次的成績相差 3分，顯著性P值為 0.005，第 6次和第 1次的成績相差 3.25分，顯著性P值為 0.007。但F5勉強同意功能（F（2.530，7.590）=1.000，P>0.001）和F6緩和功能（F（2.207，6.082）=2.942，P>0.001），6

汪敏鋒、姚鈺銀 250 次隱性學習後的成績表現不顯著，勉強同意功能配對顯著性P值為 0.252>0.05，緩和功能配對顯著性P值為 0.092>0.05。由此可知，兩個⽉的隱性學習對語氣詞「吧」的揣測功能、求證功能、建議功能和⿎勵功能的習得有顯著效果，⽽對勉強同意功能、緩和功能的習得效果不明顯。 從以上討論我們發現，⼀定頻次的隱性學習以後，受試者對語氣詞「吧」的掌握達到⼀個恒定值，我們將隱性學習的最佳頻次定為nopt，語氣詞「吧」各個功能平均成績最⾼值定為Umax，具體數值⾒表 8。 表8 6項功能類型的Umax和nopt值 功能 F1 F2 F3 F4 F5 F6 Umax 5.25 4.5 4.25 4.74 3.5 2 nopt 5 4 5 6 >6 >6 最佳頻次nopt顯⽰，「吧」的⼀級功能項（F1 揣測功能和F2 求證功能），⼆級功能項（F3建議功能）經過 4-5次的隱性輸入教學後即可獲得，⽽⼆級功能項（F4⿎勵功能、F5 勉強同意）和三級功能（F6 緩和功能）的獲得需要經過 6 次以上典型語境的隱性輸入教學。 3.3 ㆗級與高級隱性教㈻習得效果 根據被試漢語熟練度，中級和高級兩組的隱性教學效果，統計結果如下圖 3。 圖3 中高級6項成績對比圖 35 3227 2931451464942 44220102030405060F1 F2 F3 F4 F5 F6中级⾼级

隱性教學對英語母語者習得語氣詞「吧」的影響 251 各功能項總分顯⽰，⾼級組和中級組隱性教學習得效果路徑⾼度相似，但⾼級組隱性教學的總體效果要優於中級組。我們採⽤獨⽴樣本t檢驗，進⼀步判斷中級組和⾼級組語氣詞「吧」6 個功能專案 6 次總得分的具體差異。研究資料不存在顯著異常值。⾒下表 9。 表9 中高級水準使用「吧」各項功能的描述統計表 分組 個案數 平均值 標準差 t sig F1 中級 2 17.50 3.536 2.263 .152 高級 2 25.50 3.536 F2 中級 2 16.00 1.414 4.950 .038 高級 2 23.00 1.414 F3 中級 2 13.50 2.121 6.957 .020 高級 2 24.50 .707 F4 中級 2 14.50 3.536 2.600 .122 高級 2 21.00 .000 F5 中級 2 15.50 .707 13.000 .006 高級 2 22.00 .000 F6 中級 2 2.00 .000 9.000 .012 高級 2 11.00 1.414 表 9 結果顯⽰，在F2 上，中級組得分成績（16 ± 1.414）低於⾼級組得分成績（23 ± 1.414），差值為 7分。獨⽴樣本t檢驗結果顯⽰，t= 4.950，P<0.05，說明中級組和⾼級組得分存在統計學差異，⾼級組平均得分⾼於中級組。在F3上，中級組得分成績（13.5 ± 2.121）低於⾼級組得分成績（24.5 ± 0.707），差值為 11分。獨⽴樣本t結果顯⽰，t=6.957，P<0.05，說明中級組和⾼級組得分存在統計學差異，⾼級組平均得分⾼於中級組。在F5上，中級組得分成績（15.5 ± 0.707）低於⾼級組得分成績（22 ± 0.000），差值為 6.5分。獨⽴樣本t檢驗結果顯⽰，t= 13.000，P<0.05，說明中級組和⾼級組得分存在統計學差異，⾼級組平均得分⾼於中級組。在F6上，中級組得分成績（2 ± 0.000）低於⾼級組得分成績（11 ± 1.414），差值為 11分。獨⽴樣本t檢驗結果顯⽰，t= 9.000，P<0.05，說明中級組和⾼級組得分存在統計學差異，⾼級組平均得分⾼於中級組。在F1、F4 上，根據獨⽴樣本t結果顯⽰，P>0.05，說明中級組和⾼級組得分不存在統計學差異。 由此可⾒，語氣詞「吧」不同功能類型隱性教學效果在中⾼級組中呈現差異，

汪敏鋒、姚鈺銀 252 F2（求證）、F3（建議）、F5（勉強同意）、F6（緩和）四個功能類型，⾼級組的平均得分均⾼於中級組，隱性學習效果更加顯著︔F1（揣測）、F4（⿎勵）兩個功能類型，中⾼級組得分差異不明顯，隱性教學效果無顯著差異，這可能與該功能項的語境依賴程度有關。 四、討論 （㆒）、語氣詞的隱性教㈻的㈲效性 隱性教學對語氣詞「吧」的習得有顯著效果，通過前後測配對實驗可以發現：經過 432次的隱性學習輸入，語氣詞「吧」的運⽤能⼒總體有了明顯的提⾼，這不僅體現在語氣詞使⽤的準確性上，也表現在語氣詞「吧」各功能運⽤的廣度上。經過隱性教學，英語母語漢語學習者語氣詞「吧」運⽤變得更豐富多樣，總體成績表現不斷進步。訪談中，在有導向的情景中頻繁出現⽬標句能夠吸引學習者將注意資源轉移到語氣詞「吧」上，使其⾃覺分析、推導語⾔規則，對比各功能項之間的差異，從⽽下意識習得語氣詞「吧」各個功能。訪談結束後進⾏⾃然對話測試，通過情景再現可以吸引學習者根據情景線索有效輸出包含語氣詞「吧」各個功能類型的句⼦，這進⼀步促進了語氣詞「吧」的習得。總體來看，隱性教學對意義複雜的語氣詞「吧」的習得是有效的。 （㆓）、隱性教㈻對語氣詞功能類型影響的差異性 語氣詞「吧」具有很強的動態性和語境適應性，陸儉明先⽣曾指出，不論「吧」字句作為問話或答語，語⾳形式⼀樣，什麼時候看作問話，什麼時候看作答語或平敘述語，完全是由上下⽂（即語境）決定的。20 因此語境決定了語氣詞「吧」在交際中發揮什麼功能。根據實驗可知，語氣詞功能類型語境依賴程度越⾼，語義複雜度就越⾼，習得難度就越⼤，習得效果最弱，已經程式化，功能複雜度較低，語境依賴較弱的功能習得難度較⼩，隱性教學效果較好。原因在於對語境依賴程度越⾼，整體意義越模糊越不透明，學習者無法從其本身瞭解語義。這與普通話兒童習得語 20 陸儉明，〈關於現代漢語裡的疑問語氣詞〉，《中國語文》，第 5期（1998），4。

隱性教學對英語母語者習得語氣詞「吧」的影響 253 氣詞的實驗結果基本⼀致。語氣詞的基本語氣⽤法兒童習得最早、產出最多，主要承擔⼜氣強弱調節功能的語氣詞，對兒童的語⾔使⽤能⼒要求更⾼，習得更晚。21 另外從記憶的角度來看，需要根據語境推測意義的功能類型，不利於學習者長期記憶，⽽語境依賴程度較弱、已經程式化的功能類型則相反，Schmid & Günther認為⼀個詞彙或表達具有認知上的顯著，其中之⼀便是語境⾃由的固化引起的顯著：符合長期記憶中知識儲備的預期。22 因此已經程式化、語境依賴程度較弱的功能類型，其語義能夠脫離語境的推導，作為⼀種知識儲存在長期記憶中，23 ⽅便學習者識別、記憶和使⽤，因此語境依賴程度越⾼習得難度越⼤。這也可能是F5（勉強同意）和F6（緩和）中⾼級組習得效果不顯著的原因。同時，難度比較低的功能項⽬隱性教學效果不存在顯著差異，體現了隱性教學對語氣詞「吧」不同功能項影響也具有選擇性。 （㆔）、語氣詞「吧」獲得過程的動態變異性 通過隱性教學，英語母語漢語學習者語氣詞「吧」的語⽤能⼒發展路徑呈動態複雜、非線性的發展樣態。語⾔系統就是⼀個非常典型的動態系統，⾸先語⾔系統的發展是非線性的，學習過程並不是簡單的直線，⽽是有進步、又倒退。其次，語⾔發展也具有變異性，學習者之間存在著個體差異，⽽同⼀個體的學習過程在不同時間、不同情況下也會表現出不同的發展特點。24 實驗發現，學習者經過情景導入的隱性教學後，語氣詞「吧」的運⽤發⽣了以下變化：從變化速率看，經過四次隱性學習輸入和測試後，學習者總體在後續兩次訪談中會主動使⽤語氣詞⼀⼆級功能類型，爾後兩次測試維持在相對穩定的⽔準，說明前四次共 288頻次的隱性輸入對使⽤頻率⾼且語境依賴不強的語氣詞功能類型習得有階段性轉變效果，變化速度和廣度發展都較快，⽽使⽤頻率次⾼、語境依賴程度強的功能類型習得速度則較慢。但學習者L1 在F5 勉強認同功能項上出現反復，第 5 次隱性輸入後測表現不及第 3 21 高亮，〈普通話兒童句末語氣助詞的習得研究〉，《中國語文》，第 6期（2019），675-692+766-767。 22 Schmid H J, Günther F., “Toward a unified socio-cognitive framework for salience in language.” Frontiers in psychology 7 (2016): 1110. 23 李強，〈習語構式「多的是」的動態語義浮現〉，《語言教學與研究》，第 4期（2021），56-67。 24 周丹丹、王文宇，〈動態系統理論與微變化研究法在二語習得中的應用〉，《江淮論壇》，第 4期（2013），163-168+125。

汪敏鋒、姚鈺銀 254 次輸入後測。在變化的差異性上，儘管所有被試隱性習得語氣詞「吧」效果都有所提升，但F1 揣測功能第 5 次隱性教學後測和第 1 次的成績相差 3.25 分，第 6 次和第 1次的成績相差 2.25分，出現倒退，語氣詞「吧」的功能項表現出⼀定的變異性。 （㆕）、漢語熟練度對語氣詞隱性教㈻效果的影響 漢語熟練程度與語氣詞「吧」運⽤能⼒總體上是⼀同發展，漢語熟練程度與語氣詞「吧」功能類型習得存在差異（F(1，53)=159.44，p＜0.001）。⾼熟練程度的英語母語漢語學習者通過隱性輸入的學習效果要⾼於⽔準低的學習者，能更多更準確地運⽤語氣詞「吧」做出反應和回復，⽽中等熟練程度的學習者隱性學習效果與⾼熟練程度的學習者存在顯著差異（F（4.53）= 160.73，p＜0.001）。關於隱性教學的效果很多學者都做過相關研究，⼤多數研究中語⾔分析能⼒被證實與⼆語外顯知識具有較⾼正相關，⽽張潤晗、陳亞平經過實證研究驗證了Robinson、Ranta及Roehr的發現，證明語⾔分析能⼒對內隱知識的發展能夠產⽣影響。25本實驗通過情景導向式訪談隱性輸入語氣詞「吧」的功能項，過程中只提供「⼤劑量輸入」，不提供元語⾔知識，因此受試者必須依賴語⾔分析能⼒推斷並概括語⾔規則，⾼熟練度的英語母語漢語學習者知識儲備較豐富，語⾔分析能⼒較強，在接觸⽬標語的過程中能有效投入「注意」，⾃覺發現其規律，構建語⾔規則，從⽽提⾼語氣詞「吧」的習得效果︔26⽽漢語熟練度較低的學習者，其語⾔分析能⼒較弱，對情景再現線索敏感度較低，不能夠很好的構建語⾔規則，因此隱性學習效果低於前者。但是兩者與母語使⽤者相比（中級組得分率 0.366:1︔⾼級組得分率 0.588:1），在語氣詞「吧」使⽤上都存在較⼤差異，回避和使⽤不⾜現象比較普遍，即使漢語熟練程度⾼、在⽬的語環境沉浸時間較長，語氣詞運⽤能⼒與本族語者仍然存在較⼤差距。 25 張潤晗、陳亞平，〈個體差異因素與二語內隱知識的關係研究〉，《現代外語》，第 3期（2020.43），377-388；Robinson P., “Individual differences and the fundamental similarity of implicit and explicit adult second language learning.” Language learning 47.1 (1997): 45-99; Ranta L., “The role of learners’ language analytic ability.” Individual differences and instructed language learning 2 (2002): 159; Roehr K., “Metalinguistic knowledge and language ability in university-level L2 learners.” Applied Linguistics 29.2 (2008): 173-199. 26 蘇建紅，〈顯性/隱性教學與語言分析能力對二語知識習得的交互作用〉，《現代外語》，第 4 期（2012.35），385-392+437。

隱性教學對英語母語者習得語氣詞「吧」的影響 255 這說明雖然隱性教學對語氣詞「吧」的習得有⼀定效果，但效果是有限的。 五、結語及啟示 本研究對採集的資料進⾏統計分析，其中語氣詞「吧」的功能類型、隱性輸入頻次為被試內變數︔漢語⽔準（中級、⾼級）為被試間變數。對話測試結果採⽤ SPSS22.0 進⾏重複測量（Repeated Measure）⽅差分析。研究運⽤微變化研究法，以互動語氣詞「吧」的 6個典型功能項為語⾔觀察點，採取情景導向式隱性輸入教學和情景再現的⾃然對話後測，觀察學習者學習語氣詞「吧」習得變化的整個過程，通過實驗實證檢驗隱性教學效果和語氣詞的「吧」動態發展軌跡。研究發現：（1）英語母語漢語學習者語氣詞「吧」隱性教學的效果顯著，語氣詞運⽤能⼒總體得到提升，主要體現在語氣詞使⽤的準確性和語氣詞功能運⽤的廣度上︔（2）通過隱性教學，英語母語學習者語氣詞運⽤能⼒發展路徑呈動態、複雜、非線性的發展樣態︔（3）不同漢語⽔準的英語母語漢語學習者隱性教學的效果、不同語氣詞功能類型習得效果均存在差異，學習者漢語熟練程度越⾼，隱性教學的效果越顯著，語氣詞「吧」的功能類型語境依賴程度越⾼，習得難度越⼤。 我們從研究中得到啟⽰：隱性教學在⼀定程度上能夠幫助學習者習得漢語互動語氣詞，提升語氣詞「吧」運⽤能⼒，但隱性教學的效果受到語法知識難度、學習者個體差異的影響，因此教學中要考慮多⽅因素實施有針對性的教學。⼀⽅⾯，充分考慮到漢語語氣詞的特點，將教學重點著眼於提升學習者語氣詞語⽤能⼒上。語境依賴程度⾼的功能類型需要結合⼤量的真實語境，引導學習者在交際中把握語義、⽤法。另⼀⽅⾯，充分考慮學習者的個體差異。對漢語熟練度⾼、語⾔分析能⼒強的學習者，提供真實語境和⼤量會話實例，説明其發揮⾃身優勢進⾏隱性學習。對於漢語熟練度較低的學習者，在提供語⾔實例的基礎上結合顯性教學，對語⾔規則進⾏講解，盡可能彌補他在語⾔分析能⼒上的不⾜，達到最佳學習效果。 本研究也具有⼀定的局限性。受到「沒有接受顯性課堂教學的⾃然習得者」的條件限制，且能對其進⾏跟蹤調查，符合實驗要求的被試規模客觀上受到較⼤限制。有些異常資料，例如表 9中F4 ⾼級組、F5 ⾼級組以及 F6 中級組的標準差都為 0，說明在這幾個⼦組中，各被試的得分完全⼀致，沒有差異。這通常意味著樣本量偏

汪敏鋒、姚鈺銀 256 ⼩，若要進⼀步確認其實際差異，有待更⼤規模樣本樣進⼀步的檢驗。今後，將進⼀步擴⼤被試規模和實驗週期，探討其他語氣詞的隱性學習效果，並與顯性教學進⾏對照實驗，為語氣詞的教學⽅法選取和獲得⽅式提供更多的來⾃漢語的科學證據。

隱性教學對英語母語者習得語氣詞「吧」的影響 257 徵引書目 [1] 陳默、王建勤，〈漢語聲調教學的實驗和電腦類比研究〉，《語⾔教學與研究》，第 1期（2011），11-16。 [2] 戴曼純，〈⼆語習得的「顯性」與「隱性」問題探討〉，《外國語⾔⽂學》，第2期（2005），101-111。 [3] 董妍，《對外漢語教學中的語氣助詞教學》，⿊龍江⼤學，2011。 [4] 馮麗萍、孫紅娟，〈第⼆語⾔習得順序研究⽅法述評〉，《語⾔教學與研究》，第 1期（2010），9-16。 [5] ⾼亮，〈普通話兒童句末語氣助詞的習得研究〉，《中國語⽂》，第 6期（2019），675-692+766-767。 [6] 郝美玲，〈留學⽣漢字正字法意識的萌芽與發展〉，《世界漢語教學》，第 1期（2007），29-39+2-3。 [7] 洪煒，〈漢語作為第⼆語⾔的近義詞教學實驗研究〉，《世界漢語教學》，第 3期（2013），424-432。 [8] 靳洪剛、侯曉明，〈漢語作為第⼆語⾔實證研究縱觀:顯性與隱性學習、知識、教學〉，《世界漢語教學》，第 3期（2016.30），379-400。 [9] 靳洪剛，〈任務複雜度及其互動、輸出效應〉，《Journal of the Chinese Language Teachers Association》，第 2期（2010），101-135。 [10] 靳洪剛、章吟，〈對外漢語糾錯回饋研究的單位界定及過程分解分析〉，《Journal of the Chinese Language Teachers Association》，第 3期（2014），87-116。 [11] 陸儉明，〈關於現代漢語裡的疑問語氣詞〉，《中國語⽂》，第 5期（1998），4。 [12] 蘇建紅，〈顯性/隱性教學與語⾔分析能⼒對⼆語知識習得的交互作⽤〉，《現代外語》，第 4期（2012.35），385-392+437。 [13] 孫曉明，〈任務為導向的留學⽣伴隨性詞彙學習研究〉，《漢語學習》，第 4期（2012），91-96。 [14] 湯茗儀，《對外漢語常⽤語氣詞教學策略研究》，蘇州⼤學，2018. [15] 汪敏鋒，《語氣詞的⼈際語⽤功能研究》，中國社會科學出版社，2022.

汪敏鋒、姚鈺銀 258 [16] 袁博平，〈計算複雜性與第⼀語⾔遷移—以漢語第⼆語⾔態度疑問句為例〉，《世界漢語教學》，第 1期（2017.31），85-104。 [17] 張潤晗、陳亞平，〈個體差異因素與⼆語內隱知識的關係研究〉，《現代外語》，第 3期（2020.43），377-388。 [18] 周丹丹、王⽂宇，〈動態系統理論與微變化研究法在⼆語習得中的應⽤〉，《江淮論壇》，第 4期（2013），163-168+125。 [19] 李強，〈習語構式「多的是」的動態語義浮現〉，《語⾔教學與研究》，第 4期（2021），56-67。 [20] Anderson J R, Lebiere C J., The atomic components of thought. Psychology Press, 1998. [21] Ellis R. 2., Measuring Implicit and Explicit Knowledge of a Second Language//Implicit and explicit knowledge in second language learning, testing and teaching. Multilingual Matters, 2009: 31-64. [22] Ellis R., “Form-focused instruction and the measurement of implicit and explicit L2 knowledge.” Implicit and explicit learning of languages (2015): 417-441. [23] Ellis R., “Learning a second language through interaction.” Learning a Second Language through Interaction (1999): 1-295. [24] Harley B., “Functional grammar in french immersion: A classroom Experiment1.” Applied linguistics 10.3 (1989): 331-360. [25] Hazeltine E, Ivry R B., “Neural structures that support implicit sequence learning.” (2003). [26] Housen A, Pierrard M., “Investigating instructed second language acquisition.” Investigations in instructed second language acquisition (2005): 1-27. [27] Kuo L H., Improving implicit learning and explicit instruction of adult and child learners of Chinese. Brigham Young University, 2013. [28] Liao Y F., The effect of explicit and implicit instruction and native language exposure for advanced L2 learners in Chinese pragmatics: apologies. Brigham Young University, 2014. [29] Peng F., “The Effectiveness of Explicit Instruction Versus Implicit Instruction Method

隱性教學對英語母語者習得語氣詞「吧」的影響 259 on Chinese Grammar Acquisition.” (2015). [30] Ranta L., “The role of learners’ language analytic ability.” Individual differences and instructed language learning 2 (2002): 159. [31] Reber A S., “Implicit learning of artificial grammars.” Journal of verbal learning and verbal behavior 6.6 (1967): 855-863. [32] Richards J C, Schmidt R W. Longman dictionary of language teaching and applied linguistics. Routledge, 2013. [33] Robinson P., “Individual differences and the fundamental similarity of implicit and explicit adult second language learning.” Language learning 47.1 (1997): 45-99. [34] Robinson P., “Learning simple and complex second language rules under implicit, incidental, rule-search, and instructed conditions.” Studies in second language acquisition 18.1 (1996): 27-67. [35] Roehr K., “Metalinguistic knowledge and language ability in university-level L2 learners.” Applied Linguistics 29.2 (2008): 173-199. [36] Schmid H J, Günther F., “Toward a unified socio-cognitive framework for salience in language.” Frontiers in psychology 7 (2016): 1110. [37] Spada N, Jessop L, Tomita Y, et al., “Isolated and integrated form-focused instruction: Effects on different types of L2 knowledge.” Language teaching research 18.4 (2014): 453-473. [38] Taft M, Chung K., “Using radicals in teaching Chinese characters to second language learners.” Psychologia 42.4 (1999): 243-251. [39] Wallach D, Lebiere C., “Conscious and unconscious knowledge: Mapping to the symbolic and subsymbolic levels of a hybrid architecture.” (2003). [40] Wang J. The effect of implicit vs. Explicit instruction on learning form-based vs. Meaning-based language features. University of Pittsburgh, 2014.

汪敏鋒、姚鈺銀 260 附件： 前測問卷： 請問下列各句要不要⽤「吧」，並嘗試說⼀說⽤或不⽤的原因。 A.這個有多重？B.⼤概有 10⽄（ ）。 A.你來中國有⼀年（ ）？B. 差不多。 A.我們什麼時候去游泳？B.晚上七點半去（ ）。 A.這個問題我不會？B.沒關係，說說（ ）。 A.你能陪我⼀下去嗎？B.好（ ），下次你⾃⼰去了。 A.你⾛（ ），我不想和你說了。 後測問卷⽰例： 情景：談論⽇常⽣活習慣，被試估算玩遊戲時長。 A. 每天玩遊戲玩多長時間？ B. 兩三個⼩時（ ）。 情景：誇獎對⽅漢語不錯，應該學了⼀段時間。 A. 你漢語說得很不錯，學了⼀段時間了（ ）？B. 學了⼀年了。 情景：⼤家商議明天的活動，提建議，⼀起爬⿎山。 A. 好久沒運動了，我們去爬⿎山（ ）。B. 好啊好啊。 8. 情景：⿎勵朋友勇敢嘗試臭⾖腐。 A. 這道菜很好吃，吃（ ），你會喜歡的。B. 這氣味不是很好。 情景：朋友約你打籃球，但是你有會議，經過朋友多次勸說，最後答應了。 A. 會議能不能改⼀下時間，就差你了。B. 好（ ），你們等⼀下。 情景：看到朋友不注意環保，亂扔垃圾，你覺得這樣不好，想提醒他。 A. 亂扔垃圾不好（ ）。B. 不好意思，我撿起來。

隱性教學對英語母語者習得語氣詞「吧」的影響 261 作者簡介 汪敏鋒，男，博士，福建師範大學教育學院副教授，世界漢語教學學會會員，研究方向為語言學及應用語言學、國際中文教育 通訊電郵：best0301@163.com 作者簡介 姚鈺銀，女，福建師範大學海外教育學院碩士研究生，研究方向為國際中文教育 通訊電郵：1281080006@qq.com

262 《澳門科技大學學報》（人文社會科學版）稿 約 （2025年01月修訂） 1. 《澳門科技大學學報》為澳門科技大學出版之綜合性學術刊物，下設「人文社會科學版」﹝以下簡稱《科大學報》（人文社科版）﹞，內容涵蓋人文藝術、歷史學、社會科學、商學、經濟學、管理科學、語言學、國際中文教育研究等各領域專業論文。《科大學報》（人文社科版）歡迎海、內外不同專家學者賜稿，尤其具有問題意識，資料豐富，論證堅實，文字精鍊，具原創性之論文，以促進專業知識交流。 2. 《科大學報》（人文社科版）之當然主編為校長，副主編為分管學報副校長，下轄編輯委員會若干成員與編輯部，負責處理稿件審查、對外聯絡、排版與出版等事宜。 3. 《科大學報》（人文社科版）論文形式以繁體中文或英文為主。字數方面以不超過三萬字為原則，唯特約稿不在此限。《科大學報》（人文社科版）為季刊，於每年三月、六月、九月下旬及十二月中旬前出版。來稿採「隨到隨審」制，由《科大學報》（人文社科版）編輯委員會初審後，再送交海內外專家學者評審。評審人寫出評審意見書後，再由編委會逐一討論是否採用。稿件審查採「雙向匿名」方式，作者與評審人之姓名互不透露，文中請勿出現足以辨識作者身分之資訊。《科大學報》（人文社科版）當儘速通知作者審查結果，然恕不退還來稿。 4. 請確保提交之稿件為最終版本。若要修正或重新提交稿件，編輯部有權拒絕接受相關稿件，並紀錄在案。 5. 編輯委員會保留對來稿採用之決定權，或作技術與文字之修改。 6. 請勿一稿兩投。來稿以未曾發表者（含網路發表或專書）為限，會議論

263 文請查明該會議無出版論文集計畫；若為政府機構補助之研究計畫成果，亦請註明。若經查來稿已發表或一稿兩投屬實，將逕行退稿。 7. 所投稿件須或推薦稿件必須為作者獨立研究完成之作品，並充分尊重他人智慧財產權（勿侵害第三人權利），且無任何違法、違紀和違反學術倫理規範之內容。如發現存在學術造假或學術不端等行為，概由投稿者負擔法律責任，且按照本刊規定嚴肅處理。 8. 論文中牽涉版權部分（如圖片及較長之引文），請事先取得原作者或出版者書面同意，《科大學報》不負版權責任。 9. 寫作時請參照《科大學報》（人文社科版）格式(請隨時留意網站更新版本)撰寫，並附中英文篇名；論著及研究討論加附中英文關鍵詞及各五百字左右的中英文摘要。 10. 《科大學報》（人文社科版）以電腦排版，除特殊情況外，請作者檢附電子檔案，並使用Microsoft Office-Word繁體中文版。 11. 《科大學報》（人文社科版）收到稿件後，約十個工作天內進行初審，如通過者有以下兩種情況：一、稿件顯示初審狀態，即已通過初審入庫；二、如透過電郵回覆，則會以電子郵件回覆。上述程序完成後，將視乎稿件多寡進入雙匿審程序。一般而言，程序約為五十至六十個工作日（仍需視乎評審專家實際狀況）。如投稿超過六個月（一百八十個工作日）沒有通知，作者可逕行處理稿件。 12. 刊稿不付稿酬，但送作者抽印本十五份、當期四本；但約稿則不受此限制。 13. 凡登載於《科大學報》（人文社科版）之文稿，其出版權即屬《澳門科技大學學報》（人文社會科學版）所有，除取得《澳門科技大學學報》（人文社會科學版）同意外，不得自行轉載。惟作者仍保有集結出版、教學

264 及個人網站等無償使用之權利。 14. 《科大學報》（人文社科版）現已全面採用線上投稿系統，敬請閣下至本刊網站註冊作者系統後即可投稿（電郵與紙本投稿皆不受理）。

265 《澳門科技大學學報》（人文社會科學版）撰稿格式 （2025年1月修訂） 一、說 明 《澳門科技大學學報》為海內外綜合性學術期刊，下設「人文社會科學版」（以下簡稱《科大學報》（人社科科版），收錄人文哲學、歷史學、社會學、經濟金融、法學、語言學、教育學及國際中文教育研究等各項領域之文章。故此，本刊特以《芝格哥格式手冊》（第十八版）﹝The Chicago Manual of Style (18th version)﹞為基礎，提供撰稿格式之基本參考，敬請投稿者務必遵照。 二、撰稿規範 1. 《科大學報》（人社科科版）撰稿以本格式為規範。除從缺或特別格式需求，可參照自身領域期刊規範外，本刊一律以此規範和《芝格哥格式手冊》（第十八版）﹝The Chicago Manual of Style (18th version)﹞為基礎。 2. 請用新式標點符號。「」用於平常引號，『』用於引號內之引號；《 》用於書名與期刊，〈 〉用於論文、篇名與法令；英文書名需用斜體(Italic); 論文篇名用“ ”。 全文與徵引書目符號之使用與區別：中文一律使用全形符號，英文一律使用半形符號，切勿混合使用。 3. 獨立引文每行低三格（楷書）；不必加引號。 4. 西元年代、統計數字，請使用阿拉伯數字。 5. 中、西文混合使用時，標點符號請以符號前一字為依據。 6. 圖表照片請注明資料來源，並以阿拉伯數字編號，引用時請注明編號，萬勿使用如前圖、見右表等表示方法。 7. 請勿在註釋/腳中使用：同上、同前引書、同前書、同前揭書、同注幾引書，ibid., ; Op. cit.,; loc. cit.,; idem等專有引用詞。 8. 如有未盡之處，請務必參照《芝加哥格式手冊（第十八版）》內之相關規定。

266 (一 )、 版 面 格 式 規 範 1. 請以橫式左至右書寫。 2. 請勿胡亂更動本刊撰稿格式（Word檔），並注意字體之繁簡轉換。 3. 論文字型主要為新細明體，英文則為 Times New Roman，全文必須左右對齊。以下為論文各部分字型與字體大小之規定： (1) 論文之中文題目為華康標楷體W7 16pt，英文則為 Times New Roman 16pt. (2) 作者姓名、職位、服務單位與簡介等資料為標楷體（中文）與Times New Roman（英文）10.5pt。 (3) 中英文摘要字體為標楷體與 Times New Roman 11pt，內文為新細明體 11pt。 (4) 論文標題分為四種，依次為：一、（一）、1、（1）、對照為大標、中標、小標、小小標。除大標題為標楷體 14pt外，其餘標題皆為華康魏碑體W713pt。 (5) 表格標題為標楷體 10pt，內容為新細明體 11pt，表說明則為標楷體 10pt。 (6) 所有插圖標題為標楷體 10.5pt，圖說為標楷體 10pt。 (7) 致謝辭、徵引書目、附錄為標楷體 14pt，內容皆為新細明體/Times New Roman 11pt。 4. 作者必須清晰標明資料來源，引用專書或論文，請依序注明作者、書名(或篇名)、出版項。一般來說，參考文獻僅指參考材料之來源，可不用在文章中引用；徵引書目/文獻則指文章中所引用之來源，必須每筆引註。故此，本刊一律採用徵引書目以區別參考文獻，俾能明暸作者據實引用何種著作與材料。 5. 本刊之表格、圖片和附錄，引註格式與下列相同。 (二 )、註釋 /腳與徵引書目格式 (1). 中文書目之引用 一、專書 1.1 作者，《書名》，（出版地：出版社，年份），頁數 初引：孫隆基，《殺母的文化：二十世紀美國大眾心態史》，（臺北：臺大出版中心，2009），357。 1.2 作者，《書名》，頁數 再引：孫隆基，《殺母的文化：二十世紀美國大眾心態史》，367。 1.3 作者，《書名》，出版地：出版社，年份。

267 徵引書目：孫隆基，《殺母的文化：二十世紀美國大眾心態史》，臺北：臺大出版中心，2009。 ◎電子書引用方式亦同，但必須在最後附上相應網址或 DOI碼。 二、 論文與篇章 1.4 作者—〈篇名〉—《期刊》—第?期—（出版地：年/月份），頁數。 初引：高斯華著、田一言譯，〈1895年澳門流行病腺鼠疫報告〉，《澳門學》，第 1期（澳門：2022.03），298。 1.5 作者—〈篇名〉—頁數。 再引：高斯華著、田一言譯，〈1895年澳門流行病腺鼠疫報告〉，287-328。 1.6 作者—〈篇名〉—《期刊》—第？期—（出版地：年/月份），全部頁數。 徵引書目：高斯華著、田一言譯，〈1895年澳門流行病腺鼠疫報告〉，《澳門學》，第 1期（澳門：2022.03），287-328。 研討會或論文集，未經正式出版之成果，引註如下： ◎作者—篇名—研討會/論文集名稱—錄入《》/「研討會名稱」—參與地：參與場所，日期—頁數。 初引： 1. 曾幸麒，〈帝國與婚姻—二十世紀不列顛王權的形象塑造〉，「第四十三回臺灣世界史討論會—帝國與島嶼學術研討年會」，（臺北：中央研究院史語所，2018），156。 2. 米健，〈現今台港澳法學教育的若干問題〉，錄入米健、方泉、謝耿亮編，《澳門法律改革與法制建設》（北京：社會科學文獻出版社，2011），56。 再引： 1. 曾幸麒，〈帝國與婚姻—二十世紀不列顛王權的形象塑造〉，156。 2. 米健，〈現今台港澳法學教育的若干問題〉，56。 徵引書目： 1. 曾幸麒，〈帝國與婚姻—二十世紀不列顛王權的形象塑造〉，「第四十三回臺灣世界史討論會—帝國與島嶼學術研討年會」，臺北：中央研究院史語所，2018，145-168。 2. 米健，〈現今台港澳法學教育的若干問題〉，錄入米健、方泉、謝耿亮編，

268 《澳門法律改革與法制建設》，北京：社會科學文獻出版社，2011。 三、碩博士學位論文 1.7 作者—《書名》—（畢業地：畢業校/系所/碩博士論文，年份）—頁數。 初引：曾幸麒，《二十世紀不列顛王室的婚姻與憲政關係》，（臺北：政治大學歷史學系研究所碩士論文，2022），56。 1.8 作者—《書名》—頁數 再引：曾幸麒，《二十世紀不列顛王室的婚姻與憲政關係》，56。 1.9 作者—《書名》—畢業地：畢業校/系所/碩博士論文，年份。 徵引書目：曾幸麒，《二十世紀不列顛王室的婚姻與憲政關係》，臺北：臺大出版中心，2022。 四、引用原版或影印版古籍，請注明版本與卷頁 1.10 作者—《書名》—（出版地：出版商，年份）—冊/卷數—頁數。 初引：劉芳輯、章文欽校，《葡萄牙東波塔檔案館藏—清代澳門中文檔案彙編》（澳門：澳門基金會，1999）上冊，卷 1，300。 再引：劉芳輯、章文欽校，《葡萄牙東波塔檔案館藏—清代澳門中文檔案彙編》上冊，卷 1，300。 徵引書目：劉芳輯、章文欽校，《葡萄牙東波塔檔案館藏—清代澳門中文檔案彙編》上冊，澳門：澳門基金會，1999。 五、檔案與報紙 1.11 作者/機構—〈檔名/報章標題〉—《檔案/報章名稱》期/卷數—檔號—日期—頁數。 初引：澳門特別行政區行政長官辦公室，〈第 20/2000號行政命令〉，《澳門特別行政區公報》第 13期，2000年 3月 27日，240。 再引：澳門特別行政區行政長官辦公室，〈第 20/2000號行政命令〉，240。 徵引書目：澳門特別行政區行政長官辦公室，〈第 20/2000號行政命令〉，《澳門特別行政區公報》第 13期，2000年 3月 27日，232-244。

269 六、網路資源： 1.12 作者/機構(如有)—「資料名稱/標題」—資源出處—網址—瀏覽日期。 初引、再引與徵引書目： 香港電台(RTHK)，「Travelling with Water: Joseph LEE (I)」(與水同游/遊：李行 偉 教 授 ) ， Our Scientists (English Version) ，https://podcast.rthk.hk/podcast/item.php?pid=1344&eid=105927&lang=zh-CN，瀏覽日期為 2022年 6月 23日。 (2)英文書目之引用(Bibliography) † 英文徵引書目之引用需留意作者/譯者/編者名字先(First Name)而姓氏後(Last Name)。如遇多位作者，只需調整第一位作者即可。如出現四位或以上作者，於註釋第一作者後以 et al.表示，但在徵引書目必須詳列所有作者。 I. 西文專書(Books) Ø 作者─書名─版次 ed.─(出版地點: 出版公司, 出版年分)─頁碼。 初引：Nathan Henry, The History of Fashion, 1500-1900, 2nd ed. (Hongkong: University of Hongkong Press, 2020), 50-80. 再引：Nathan, The History of Fashion, 1500-1900, 50-70. 徵引書目：Henry, Nathan. The History of Fashion, 1500-1900, 2nd ed. Hongkong: University of Hongkong Press, 2020. Ø 作者—書名—, ed./trans.編⁄譯者姓名—(出版地點: 出版公司, 出版年分)—頁碼。 初引：Walter Bagehot, The English Constitution, ed. Miles Taylor (Oxford: University of Oxford, 2001), 32-80. 再引：Bagehot, The English Constitution, 20-25. 徵引書目：Bagehot, Walter. The English Constitution. Edited, Miles Taylor. Oxford: University of Oxford, 2001.

270 Ø 章節作者—“章節標題”—in 書名—,ed(s).編者—(出版地: 出版者, 年份)—頁碼。（譯者亦如是—,trans譯者） 初引：Ina Zweiniger-Bargielowska, “Prince Philip: Sportsman and Youth Leader,” in The Man Behind the Queen: Male Consorts in History, eds. Charles Beem and Miles Taylor. (London: Palgrave Macmillan, 2014), 223-239. 再引：Ina Zweiniger-Bargielowska, “Prince Philip: Sportsman and Youth Leader,” 223-239. 徵引書目：Zweiniger-Bargielowska, Ina. “Prince Philip: Sportsman and Youth Leader.” in The Man Behind the Queen: Male Consorts in History. Eds. Charles Beem and Miles Taylor, 223-239. London: Palgrave Macmillan, 2014. †只需調整單一章節作者即可，編者不需調整姓氏與名字次序。 II. 期刊論文(Journals and Articles) Ø 作者—“篇名,” —期刊名稱—期.卷數—(出版地: 年份/月份)—: 頁數。 初引： Yan, Hao-Chen, Lam, Man-Yue, Lee, Joseph Hun-Wei, “Field Measurements and Numerical Modeling of Hydraulic Transients in HDPE Pipeline with PRV Interaction,” Journal of Hydraulic Engineering 147.6 (ASCE: 2021.06): 67. 再引：Yan, Hao-Chen, Lam, Man-Yue, Lee, Joseph Hun-Wei, “Field Measurements and Numerical Modeling of Hydraulic Transients in HDPE Pipeline with PRV Interaction,” 67. 徵 引 書 目 ： Hao-Chen, Yan, Lam, Man-Yue, Lee, Joseph Hun-Wei, “Field Measurements and Numerical Modeling of Hydraulic Transients in HDPE Pipeline with PRV Interaction.” Journal of Hydraulic Engineering 147.6 (ASCE: 2021.06): 1-67. 研討會或論文集，未經正式出版之成果(Unpublished manuscripts, Lectures and Working papers)，引註如下： ◎作者—“篇名”—(研討會名稱/參與地: 參與場所, 日期)—頁數。 初引： 1. David G. Harper, “The Several Discoveries of the Ciliary Muscle” (PowerPoint

271 presentation, 25th Anniversary of the Cogan Ophthalmic History Society, Bethesda, MD, March 31,2012), 24. 2. Deborah D. Lucki and Richard W. Pollay, “Content Analyses of Advertising: A Review of the Literature” (working paper, History of Advertising Archives, Faculty of Commerce, University of British Columbia, Vancouver, 1980), 46. 再引： 1. David G. Harper, “The Several Discoveries of the Ciliary Muscle”, 24. 2. Deborah D. Lucki and Richard W. Pollay, “Content Analyses of Advertising: A Review of the Literature”, 46. 徵引書目： 1. David G. Harper, “The Several Discoveries of the Ciliary Muscle” PowerPoint presentation, 25th Anniversary of the Cogan Ophthalmic History Society, Bethesda, MD, March 31, 2012. 2. Deborah D. Lucki and Richard W. Pollay, “Content Analyses of Advertising: A Review of the Literature” Working paper, History of Advertising Archives, Faculty of Commerce, University of British Columbia, Vancouver, 1980. III. 電子書(Electronic Books) Ø 作者—書名—(出版地: 出版社, 年份)—電子書類型—頁數. 初引：John Jowett, et al, The Oxford Shakespeare: the Complete Works, 2nd ed, (Oxford: Clarendon Press, 2005), EPUB, 60. 再引：John Jowett, et al, The Oxford Shakespeare: the Complete Works, 2nd ed, EPUB, 60. 徵引書目：Jowett, John, William Montgomery, Gary Taylor, Stanley Wells. The Oxford Shakespeare: the Complete Works, 2nd ed. Oxford: Clarendon Press, 2005, EPUB. IV. 報紙(Newspapers) Ø 作者/機構—“標題”—報刊名稱—月日, 年/期數—版數(如有)—URL/資料庫(如有).

272 初引：Georgina Rannard, “Five planets to line up in rare planetary conjunction” BBC News Climate & Science, June 24, 2022, https://www.bbc.com/news/science-environment-61910977. 再引：Georgina Rannard, “Five planets to line up in rare planetary conjunction”, BBC News Climate & Science, June 24, 2022. 徵引書目：Rannard, Georgina, “Five planets to line up in rare planetary conjunction” BBC News Climate & Science, June 24, 2022, https://www.bbc.com/news/science-environment-61910977. V. 碩博士學位論文(thesis and dissertations.) Ø 作者—“論文名稱” —章節—(碩 /博士論文 , 畢業大學 , 年份)—頁數—URL(如有). 初引：Melanie Subacus, “Duae Patriae: Cicero and Political Cosmopolitanism in Rome,” abstract (PhD diss., New York University, 2015), v, http://pqdtopen.pro quest.com/pubnum/3685917.html. 再引：Melanie Subacus, “Duae Patriae: Cicero and Political Cosmopolitanism in Rome,” abstract, v. 徵引書目：Subacus, Melanie, “Duae Patriae: Cicero and Political Cosmopolitanism in Rome.” PhD diss., New York University, 2015, http://pqdtopen.proquest.com/pubnum/3685917.html. VI. 檔案與政府文件(public and Legal documents) Ø 作者—檔案匯編名稱/卷/冊數—(出版地: 出版商, 年份)—頁數. Ø 作者(如有)—檔案/文件名稱—日期(如有)—檔案/文件編號—收藏機構(如有)—頁數(如有). 初引： 1. Arthur Christopher, Viscount Esher, The Letters of Queen Victoria: A Selection from Her Majesty’s Correspondent between the Years 1837-1861, Vol. III(London: John Murray, 1908, Published by Authority of Her Majesty the King), 163-178. 2. Act of Settlement, 1701,12 & 13 Will. 3, c. 2.

273 再引： 1. Arthur Christopher, Viscount Esher, The Letters of Queen Victoria: A Selection from Her Majesty’s Correspondent between the Years 1837-1861, Vol. III, 163-178. 2. Act of Settlement, 1701,12 & 13 Will. 3, c. 2. 徵引書目： 1. Christopher, Arthur, Viscount Esher. The Letters of Queen Victoria: A Selection from Her Majesty’s Correspondent between the Years 1837-1861, Vol. III. London: John Murray, 1908, Published by Authority of Her Majesty the King. 2. Act of Settlement, 1701,12 & 13 Will. 3, c. 2. VII. 其他網絡資源(Websites resources) Ø 作者/機構/其他名稱—“標題” 網站名稱—瀏覽/最後修改日期—網址. 初引、再引與徵引書目： 1. Alliance for Linguistic Diversity, n.d. “Balkan Romani.” Endangered Languages. Accessed April 6,2016. http://www.endangeredlanguages.com/lang/5342. 2. Google. 2016. “Privacy Policy.” Privacy & Terms. Last modified March 25,2016. http://www.google.com/policies/privacy/.

JOURNAL OF MACAU UNIVERSITY OF SCIENCE AND TECHNOLOGYHUMANITIES AND SOCIAL SCIENCES