The Creation of Meaning: A Festschrift in Honour of Randy J. LaPolla 意义创造论——罗仁地先生七秩寿庆文集

意义创造论 罗仁地先生七秩寿庆文集 黄成龙、杨旭 主编 日本地理言語学会 ISSN 2436-6471 意义创造论 罗仁地先生七秩寿庆文集 黄成龙、杨旭 主编 日本地理言語学会 2025 『地理言語学研究』モノグラフシリーズ No. 11 ISSN 2436-6471 黄成龙、杨旭 主编 《意义创造论——罗仁地先生七秩寿庆文集》2025 Cover image: Portrait of Randy J. LaPolla © 2025 Yuan Jie doi: https://doi.org/10.5281/zenodo.17274638 © 2025 by Authors. All rights reserved. 発行日： 2025 年10 月6日 発行者： 日本地理言語学会 ウェブサイト： https://geolinguistics.sakura.ne.jp/ 所在地： 東京都渋谷区渋谷 4-4-25 青山学院大学 G1305 研究室 iii 罗仁地先生小传 罗仁地（Randy LaPolla）是当代国际著名语言学家、澳大利亚人文学院院士、功 能语言学派的重要代表人物之一。自 20 世纪 80 年代以来，罗仁地在汉语和中国境内 少数民族语言研究方面作出了开创性贡献；同时，他在语用学、历史语言学、语言接 触、田野语言学、记录语言学、类型学等语言学核心领域也造诣深厚，尤其对汉语、 羌语、独龙语、日旺语以及菲律宾塔伽洛语的研究见解独到。自 2005 年担任中国教 育部“长江学者”讲座教授以来，罗仁地持续推动类型学研究在中国的发展，并就类 型学方法论、描写与分析范式、比较与分类动因等关键问题发表了一系列重要论著。 罗仁地祖籍意大利那不勒斯市，美国公民，1955 年出生于纽约。他本科就读于纽 约州立大学石溪分校（SUNY Stony Brook），主修亚洲历史、哲学与政治。在本科期 间，他通过英译版中国先秦文献对中国哲学产生浓厚兴趣。为深入理解文献原意，他 决心学好汉语。1976 年春天，罗仁地应中国政府邀请以中美人民友好协会积极分子 的身份首次访华，实地了解新中国的发展。这次短暂访问使他萌生长期驻华研究的想 法。当时，赴华长期居留的主要途径是从事英语教学，为此他专门攻读英语教学硕士 学位。1980 年获得学位后，他立即赴湖南长沙担任暑期英语志愿教师，后转至上海华 东师范大学夜大学任教。1981 年，他以进修生身份进入北京大学中文系语言专业学 习。北大求学经历为他开启了汉语语言学和中国少数民族语言学研究的大门，从此与 这些领域结下不解之缘。 在北大期间，他不仅系统修读了汉语语法、汉语史、古代汉语、中国通史、《现 代汉语八百词》、北京话调查、中国诗歌等专业课程，更有幸聆听戴庆厦教授、孙宏 开教授、陈其光教授等专家学者特别开设的“汉藏语概论”课程。这些启蒙课程和导 师的言传身教，激发了他对中国少数民族语言研究的持久兴趣。 1983 年，罗仁地赴美国加州大学伯克利分校（UC Berkeley）攻读语言学博士学 位，师从 James A. Matisoff、王士元、Robert Van Valin 和Johanna Nichols 等教授，专 攻汉藏语系、历史语言学、语言类型学和语法理论。期间他还赴斯坦福大学跟随 Joseph Greenberg 教授研习语言类型学。在伯克利期间，他结识了李方桂教授及其家人。当 时李夫人邀请伯克利地区口述历史办公室（Regional Oral History Office）为李教授录 制口述历史，由罗仁地负责组织工作并担任访谈人。在持续数周的访谈过程中，他得 以与李教授深入交流。罗仁地常表示，这段经历不仅让他汲取了李教授渊博的语言学 知识，更学习了大师的为人处世之道。1990 年，他以论文《Grammatical Relations in iv Chinese: Synchronic and Diachronic Considerations》（汉语语法关系的共时与历时研究） 获得博士学位。该论文奠定了他对语言结构问题的独特见解的基础，尤其在汉语语法 关系和语序研究方面提出了创新性观点。 1990 年博士毕业后，罗仁地先后任职于：台湾中研院历史语言研究所（1990-1996， 为该所首位非华裔研究员）；香港城市大学语言学与翻译学系（1996-2004）；澳大利亚 拉筹伯大学，担任语言学讲座教授（2004-2012，负责全校语言学事务，2008-2010 年 兼任语言类型学研究中心主任）；新加坡南洋理工大学语言学与多语言研究系（20102021，同时在中文系、哲学系指导博士生）；北京师范大学珠海分校人文和社会科学 高等研究院（2021-2023）。 罗仁地的语言学研究以理解人类行为及其认知功能为核心。他认为语言本质上是 一种行为，通过交际行为可以窥见思维方式的差异，因为不同语言反映不同的思维模 式。罗仁地高度重视语言多样性，致力于探究不同语言的共性与个性，主张基于亚洲 语言特点修正类型学和语法理论中的印欧语系偏见，从而拓展人类对语言类型的认知。 多年来，他秉持以经验主义为核心的语言观，对类型学方法论、语言描写规范、语序 表现动因等核心问题提出了独到见解，并系统阐述了解决方案。 （潘露莉博士） v 目 录 罗仁地先生小传 iii 1引言 黄成龙 1 2罗仁地先生意义创造论导读 杨旭 20 3罗仁地先生对汉藏语研究的贡献 黄成龙、马王储 31 4专名的一类非专名化历程：从“胡搅蛮缠”说起 刘丹青、萧泳凯 50 5 Postverbal modification, information focalization, and meaning creation: With a specific case study of Chinese 在zài-phrases in translation ZHU Chunshen 65 6 Abductive inference in Mandarin interactions ZHENG Wuxi 91 7现代汉语超常规三重否定：形成机制、类型与动因 李姝姝 120 8 Empirical foundations of sound correspondence and probabilistic evaluation WANG Feng 136 9类型学框架下的彝语诺苏话关系从句 胡素华 144 10 知识论视角下语气词“哈”的认识与道义调节 周士宏、段晓臻 159 11 Honorific forms of personal pronouns in Tibeto-Burman languages and their typological significance LIU Na and LU Jun 172 12 胶辽官话文荣方言的疑问结构 姜静 199 vi 13 维西县塔城傈僳语的施受结构 李翔 214 14 苗语松桃话等比结构的类型及其特征 麻秀芝 226 15 萝卜寨羌语“给”字结构的类型学考察 王保锋 239 16 南部羌语名词组特征的比较 高韬 252 1 1 引言 *黄成龙 中国社会科学院民族学与人类学研究所 罗仁地（Randy J. LaPolla）先生是国际知名的语言类型学和汉藏语系研究学家，其学 术贡献涵盖历史语言学（historical linguistics）、语言类型学（linguistic typology）、社会语 言学（sociolinguistics）、田野语言学（field linguistics）、语用学与语言发展（pragmatics and language development）、汉藏语普通语言学（Sino-Tibetan general linguistics）、藏缅语普通 语言学（Tibeto-Burman general）以及汉语、羌语、独龙/日旺语、他加禄语（Tagalog）等 领域。罗仁地先生几十年的语言学研究，一直秉持经验主义语言学观，运用归纳法分析范 畴、构式、语言类型、语用等复杂系统构成的语言系统的模式及其演化趋势。他的研究一 方面为语言类型学研究，如独龙语的反身态/中间态、汉语作为话题型语言、原始藏缅语 无形态、名词小句修饰名词等提供了重要的类型特征；另一方面，突破了汉藏语研究受传 统语法和结构主义语言学的严重束缚，开创了汉藏语系研究的新方法、新范式。这里仅举 罗仁地先生语言学观的几个方面及其方法论。 研究方法是一个学科的灵魂，罗先生认为语言研究方法既要重视自然语料，又要重 视归纳法。不应事先预设某个语言有特定范畴，而是应从自然语料里看该语言呈现了什 么范畴。该语言中呈现的范畴应予以深入描写，没呈现的范畴则可以不提。当然也可以提 自然语料没有呈现出的范畴，如声调、示证、中间态等。最好不要说整个语言没有这些范 畴，但可以说所调查的语料没有呈现它们。这是实事求是的做法。每个学者应谨记跨语言 的范畴是不存在的，每一门语言的范畴都是独特的，要按照特定语言的事实来确定范畴。 所以说，要摆脱印欧语系的影响也很容易，就是按照自己的语言来分析。他认为不光研究 汉语要摆脱印欧语系的影响，研究中国少数民族语言也要摆脱汉语框架的影响。他提及， 中国有很多学者学了汉语的那些句法形式，然后当他们在分析少数民族语言时，就把汉 语的那些范畴，如补语、状语等，硬套在少数民族语言上。这种做法是不对的。 传统语法认为，所有语言都有主语和宾语。不及物动词只有主语，及物动词有主语和 宾语，双及物动词有主语、直接宾语和间接宾语，这是早期语法研究的共识。但是，随着 非印欧语研究和跨语言研究的不断深入，研究者们发现主语和宾语不是语言的共性，世 界语言中还有其他的语言类型。罗仁地先生反对把“主语”和“宾语”概念套用到汉语甚 至汉藏语。在研究语法关系时，他认为要区分三种不同的功能。第一种是语义。语义是一 种真实世界的事情，比如说“我打你”，“我”是施事，“你”是受事，这个是不能变的， 跟语言的表达方式根本没有什么关系，用被动、施动都是一样的。很多人认为施事是主 黄成龙. 2025. 引言. 黄成龙、杨旭（主编）《意义创造论：罗仁地先生七秩寿庆文集》，pp. 1–19.  8  The Sino-Tibetan Languages (2nd Edition)[M]. Graham Thurgood & Randy J. LaPolla (eds.), London & New York: Routledge, 2017. [2018 年]  Sino-Tibetan Linguistics: Critical Concepts in Linguistics[M]. Randy J. LaPolla (ed.), London & New York: Routledge, 2018. [2024 年]  ChatGPT 来了——语言科学如何看待 ChatGPT[M]. 杨旭, 罗仁地主编. 上海: 上海 教育出版社, 2024.  Ethnolinguistic contact across the Indo-Myanmar-Southwest China mountains. Special issue of Asian Languages and Linguistics 4[M]. Alexander R. Coupe, Randy J. LaPolla & Hideo Sawada (eds.), Amsterdam & Philadelphia: John Benjamins, 2024.  LaPolla, Randy J., David Sangdong. Rawang-English-Burmese Dictionary[Z]. http://rawang.webonary.org/ 二、单篇文章 [1987 年]  LaPolla, Randy J. Dulong and Proto-Tibeto-Burman[J]. Linguistics of the Tibeto-Burman Area, 1987, 10(01): 1-42. [1988 年]  LaPolla, Randy J. A Computer Application of Matthew Chen’s ‘From Middle Chinese to Modern Mandarin’[J]. Journal of Chinese Linguistics. 1988, 16(02): 339-344.  LaPolla, Randy J. Topicalization and the question of lexical passives in Chinese[C]//Marjorie K. M. Chan & Thomas Boyden (eds.), Proceedings of the Third Annual Ohio State University Conference on Chinese Linguistics. Columbus: Ohio State University, 1988: 1-17.  LaPolla, Randy J. Subject and referent tracking: Arguments for a discourse-based grammar of Chinese[C]//Joseph Emonds, P.J. Mistry, Vida Samiian & Linda Thornburg (eds.), Proceedings of the 7th West Coast Conference on Linguistics. Irvine: University of California, 1988: 160-173.  LaPolla, Randy J. The Phonetic Development of Tibetan: A problem set[J]. Linguistics of the Tibeto-Burman Area. 1988, 11(02): 93-97.  LaPolla, Randy J. A Computer Application of Matthew Chen’s ‘From Middle Chinese to Modern Mandarin’[J]. Journal of Chinese Linguistics . 1988, 16 (02): 339-344. [1989 年]  LaPolla, Randy J. Verb Agreement, Head-marking vs. Dependent-marking and the ‘Deconstruction’ of Tibeto-Burman Morphosyntax[C]//Kira Hall, Michael Meacham & Richard Shapiro (eds.), Proceedings of the Fifteenth Annual Meeting of the Berkeley Linguistics Society. Lexington: Linguistic Society of America, 1989: 356-366. 9  LaPolla, Randy J. Dulong and Proto-Tibeto-Burman[J]. Linguistics of the Tibeto-Burman Area 1987, 10(01): 1-42.（中文译文版：罗仁地著, 乐赛月译. 独龙语和原始藏缅语比 较研究[J]. 民族语文研究情报资料集. 1989, 11(00): 49-78, 12(00): 19-42.） [1992 年]  LaPolla, Randy J. Anti-ergative marking in Tibeto-Burman[J]. Linguistics of the TibetoBurman Area. 1992, 15(01): 1-9.  LaPolla, Randy J. On the Dating and Nature of Verb Agreement in Tibeto-Burman[J]. Bulletin of the School of Oriental and African Studies. 1992, 55(02): 298-315. （中文译 文版：罗仁地著, 罗仁地译. 藏缅语族动词的人称标志：其性质及起源时代[J]. 民族 译丛. 1993(04): 42-52.） [1993 年]  LaPolla, Randy J. Arguments against ‘subject’ and ‘direct object’ as Viable Concepts in Chinese[J]. Bulletin of the Institute of History and Philology, 1993, 63(04): 759-813.  LaPolla, Randy J. Book Review of Contributions to Linguistic Research from China, ed. by Alfred F. Majewicz, Poznan, 1991[J]. Languages of the World. 1993, 7(02): 49-51. [1994 年]  LaPolla, Randy J. An Experimental Investigation into Phonetic Symbolism as It Relates to Mandarin Chinese[C]//Leanne Hinton, Johanna Nichols & John Ohala (eds.), Studies in Sound Symbolism. Cambridge: Cambridge University Press. 1994: 130-147. （中文译文 版：罗仁地著, 杨旭译. 现代汉语的语音象征性[J]. 实验语言学. 2023(02): 12-20.）  LaPolla, Randy J. Parallel Grammaticalizations in Tibeto-Burman Languages: Evidence of Sapir’s ‘Drift’[J]. Linguistics of the Tibeto-Burman Area. 1994,17(01): 61-80.  LaPolla, Randy J. Variable Finals in Proto-Sino-Tibetan[J]. Bulletin of the Institute of History and Philology. 1994, 65(01): 131-173.  LaPolla, Randy J. Book Review of The Classical Tibetan Language, by Stephen V. Beyer, SUNY Press, 1992[J]. Language 1994, 70(01): 195-196. [1995 年]  LaPolla, Randy J. On the Change to Verb-medial Word Order in Proto-Chinese: Evidence from Tibeto-Burman[C]//Kitamura Hajime, Nishida Tatsuo & Yasuhiko Nagano (eds.), Current Issues in Sino-Tibetan Linguistics. Osaka: National Museum of Ethnology. 1995: 98-104.  LaPolla, Randy J. Book Review of Understanding Utterances: An introduction to pragmatics by Diane Blakemore, Oxford, UK & Cambridge, USA: Blackwell, 1992[J]. Language. 1995, 71(01): 173-176.  LaPolla, Randy J. Pragmatic relations and word order in Chinese [C]//Pamela Downing & Michael Noonan (eds.), Word Order in Discourse. Amsterdam: John Benjamins. 1995, 30: 297-329. （中文译文版：罗仁地著, 詹卫东译. 语用关系与汉语的词序. 载北京大学 10 汉语语言研究中心《语言学论丛》编委会编, 语言学论丛. 2004, 30: 334-368. 北京: 商 务印书馆.）  LaPolla, Randy J. Ergative marking in Tibeto-Burman[C]//Yoshio Nishi, James A. Matisoff, & Yasuhiko Nagano (eds.), New Horizons in Tibeto-Burman Morpho-syntax (Senri Ethnological Studies 41). Osaka: National Museum of Ethnology, 1995: 189-228.  LaPolla, Randy J. On the utility of the concepts of markedness and prototypes in understanding the development of morphological systems[J]. Bulletin of the Institute of History and Philology. 1995, 66(04): 1149-1185. [1996 年]  罗仁地，杨将领. 独龙语/日旺语动词的返身态和中间态标志[C]// 戴庆厦等 编. 中国 民族语言论丛. 北京: 中央民族大学出版社, 1996: 13-34.  LaPolla, Randy J. Book Review of Advances in Role and Reference Grammar (Current issues in linguistic theory 82), ed. by Robert D. Van Valin Jr., John Benjanims Publishing Co., Amsterdam[J]. Sprachtypologie und Universalienforschung. 1996, 49: 177-181.  LaPolla, Randy J. Middle voice marking in Tibeto-Burman[C]//Maria Teresa Lino, Maria Francisca Xavier, Fátima Ferreira, Rute Costa & Raquel Silva (eds.), Proceedings of the Fourth International Symposium on Languages and Languages and Linguistics: PanAsiatic Linguistics. Lisbon: European Language Resources Association. 1996, 5: 1-15. [1997 年]  LaPolla, Randy J. Book Review of Language Atlas of China (Pacific Linguistics, Series C, No. 102), ed. by S.S. Wurm et al. Hong Kong: Longman Group (Far East) Ltd, 1987[J]. International Review of Chinese Linguistics. 1997, 1(01): 92-100. [1999 年]  Word Order Patterns in Sino-Tibetan and Their Significance to Theories of Explanation In Typology[C]//潘悟云 编. 东方语言与文化. 上海: 东方出版中心, 1999: 204-237. [2000 年]  LaPolla, Randy J. Valency-changing derivations in Dulong/Rawang[C]//R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Changing Valency: Case Studies in Transitivity. Cambridge: Cambridge University Press, 2000: 282-311.  LaPolla, Randy J. Book Review of A Grammar of Meithei: Shobhana L. Chelliah, (Mouton Grammar Library, 17). Berlin: Mouton de Gruyter, 1997. XXV+ 539 pp[J]. Lingua. 2000,110(04): 299-304. [2001 年]  LaPolla, Randy J. The Role of Migration and Language Contact in the Development of the Sino-Tibetan Language Family[C]//R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Areal Diffusion and Genetic Inheritance: Case studies in language change. Oxford: Oxford University Press, 2001: 225-254. （中文译文版：罗仁地著, 沈瑞清译. 民族迁徙和语 11 言接触在汉藏语系发展过程中的作用[J]. 东方语言学, 2011(01):84-107.）  LaPolla, Randy J. Dulong Texts: Seven fully analyzed narrative and procedural texts[J]. Linguistics of the Tibeto-Burman Area. 2001, 24(02): 1-39. [2002 年]  潘露莉, 罗仁地. 信息传达的性质与语言的本质和语言的发展[J]. 中国语文. 2002(03): 203-209. [2003 年]  LaPolla, Randy J. An Overview of Sino-Tibetan Morphosyntax[C]//Graham Thurgood & Randy J. LaPolla (eds.), The Sino-Tibetan Languages. London & New York: Routledge, 2003: 22-42.  LaPolla, Randy J. Evidentiality in Qiang[C]//Alexandra Y. Aikhenvald & R.M.W. Dixon (eds.), Studies in Evidentiality (Typological Studies in Language). Amsterdam: John Benjamins, 2003: 63-78. （中文译文版：《羌语的示证》（郑武曦译），未公开出版，见 罗仁地教授个人网站。）  LaPolla, Randy J. English-Qiang Glossary by Semantic Field[C]//Ikeda Takumi (ed.), Basic Materials on Minority Languages in East and Southeast Asia (Endangered Languages of the Pacific Rim series, A03-004). Suita: Faculty of Informatics, Osaka Gakuin University, 2003: 153-181.  LaPolla, Randy J., Dora Poa. Texts in the Qugu variety of Northern Qiang[C]//Kitano Hiroaki (ed.), Descriptive and Theoretical Studies in Minority Languages of East and Southeast Asia. (Endangered Languages of the Pacific Rim Project Series). Suita: Faculty of Informatics, Osaka Gakuin University, 2003: 77-94.  Why languages differ: Variation in the conventionalization of constraints on inference[C]//David Bradley, Randy J. LaPolla, Boyd Michailovsky & Graham Thurgood (eds.), Language Variation: Papers on Aariation and Change in the Sinosphere and in the Indosphere: In honour of James A. Matisoff. Canberra: Australian National University, 2003: 113-144.  LaPolla, Randy J. Qiang[C]//Graham Thurgood & Randy J. LaPolla (eds.), The SinoTibetan Languages. London & New York: Routledge, 2003: 573-587.  LaPolla, Randy J. Dulong [C]//Graham Thurgood & Randy J. LaPolla (eds.), The SinoTibetan Languages. London & New York: Routledge, 2003: 674-682. [2004 年]  LaPolla, Randy J., Huang Chenglong. Adjectives in Qiang[C]//R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Adjective Classes: A cross-Linguistic typology. Oxford: Oxford University Press, 2004: 306-322.  LaPolla, Randy J. On nominal relational morphology in Tibeto-Burman[C]//Ying-jin Lin, Fang-min Hsu, Chun-chih Lee, Jackson T.-S. Sun, Hsiu-fang Yang, and Dah-an Ho (eds.), 12 Studies on Sino-Tibetan Languages: Papers in Honor of Professor Hwang-cherng Gong on his Seventieth Birthday. Taipei: Institute of Linguistics, Academia Sinica, 2004: 43-74. [2005 年]  LaPolla, Randy J. Typology and Complexity[C]//James W. Minett and William S-Y. Wang (eds.), Language Acquisition, Change and Emergence: Essays in evolutionary linguistics. Hong Kong: City University of Hong Kong Press, 2005: 465-493.  LaPolla, Randy J., Yang Jiangling. Reflexive and middle marking in Dulong/Rawang[J]. Himalayan Linguistics 2005, 2: 1-13.  罗仁地, 潘露莉. 焦点结构的类型及其对汉语词序的影响[C]//徐烈炯, 潘海华 编. 焦点结构和意义的研究. 上海: 外语教学与研究出版社, 2005: 57-78.  LaPolla, Randy J. The inclusive-exclusive distinction in Tibeto-Burman languages[C]//Elena Filimonova (ed.), Clusivity: Typology and Case Studies of the Inclusive–exclusive Distinction. Amsterdam: John Benjamins, 2005: 291-311.  罗仁地. 第二语言习得对第一语言的影响[C]//戴庆厦, 贾益民 主编. 第四届国际双 语学研讨会论文集. 广州: 暨南大学出版社, 2005: 50-57. [2006 年]  LaPolla, Randy J. Sino-Tibetan Languages[C]//Keith Brown (ed.), Encyclopedia of Language and Linguistics (2nd Edition). London: Elsevier, 2006: 393-397.  LaPolla, Randy J. Copula constructions in Rawang[C]//Ho Dah-an, H. Samuel Cheung, Pan Wuyun & Wu Fuxiang (eds.), Linguistic Studies in Chinese and Neighboring Languages: Festschrift in Honor of Professor Pang-hsin Ting on His 70th Birthday, Taipei: Institute of Linguistics, Academia Sinica, 2006: 1059-1076.  罗仁地. 历史语言学和语言类型学[J]. 北京大学学报. 2006, 43(02): 27-30.  LaPolla, Randy J., Dora Poa. On Describing word order[C]//Felix Ameka, Alan Dench, & Nicholas Evans (eds.), Catching Language: The Standing Challenge of Grammar Writing. Berlin: Mouton de Gruyter, 2006: 269-295.  LaPolla, Randy J. On grammatical relations as constraints on referent identification[C]//Michael Noonan (ed.), Voice and Grammatical Relations: In honor of Masayoshi Shibatani and Taro Kageyama, 2006: 139-151. [2007 年]  Aikhenvald, Alexandra Y., Randy J. LaPolla. New Perspectives on Evidentials: A view from Tibeto-Burman[J]. Linguistics of the Tibeto-Burman Area, 2007: 30(02): 1-16.  Poa, Dora, Randy J. LaPolla. Minority languages of China: Factors affecting language maintenance[C]//Osahito Miyaoka & Michael E. Krauss (eds.), The Vanishing Languages of the Pacific. Oxford: Oxford University Press, 2007: 337-354.  罗仁地, 杨将领. 通过亲属语言、方言的比较了解语言的历史发展[J]. 东方语言学, 2007(01): 11-18. （Translated version in English: Understanding the historical development 13 of a language by comparing it with related languages and dialects (translation by Nathan Straub), unpublished.）  罗仁地. 语言类型学/功能语言学派视野下的语言学田野调查[J]. 语言学论丛, 2007,36: 42-56.  LaPolla, Randy J., Huang Chenglong. The Copula and Existential Verbs in Qiang[J]. Bulletin of Chinese Linguistics. 2007, 1(02): 233-248. [2008 年]  LaPolla, Randy J. Nominalization in Rawang[J]. Linguistics of the Tibeto-Burman Area. 2008, 31(02): 45-66. （中文译文版：《日旺语的名物化》（郑武曦译），未公开出版， 见罗仁地教授个人网站。）  LaPolla, Randy J. Language Contact and Language Change in the History of the SinoTibetan Languages[C]//the Academic Steering Committee of the Beijing Forum, Peking University (ed.), The Harmony of Civilizations and Prosperity for All: Diversity in the development of human civilizations (Selected papers from the Beijing Forum, Peking University and Great Hall of the People, Beijing, 2-4 November, 2007). Beijing: Peking University Press, 2008: 144-162.  LaPolla, Randy J. Relative Clause Structures in the Rawang Language[J]. Language and Linguistics 2008, 9(04): 797-812.  LaPolla, Randy J. ‘Transitivity harmony’ in the Rawang language of northern Myanmar[C/E]//Louise de Beuzeville & Pam Peters (eds.), From the Southern Hemisphere: Parameters of language variation—E-Proceedings of the 2008 Conference of the Australian Linguistics Society. Sydney: University of Sydney, 2008: 1-9. [2009 年]  LaPolla, Randy J. Chinese as a topic-comment (not topic-prominent and not svo) language[C]//Janet Xing (ed.), Studies of Chinese Linguistics: Functional Approaches. Hong Kong: Hong Kong University Press, 2009: 9-22. （中文译文版：《汉语作为话题述题型（非话题突出型/非SVO 型）语言》（李姝姝译），未公开出版，见罗仁地教授 个人网站。）  LaPolla, Randy J. Causes and effects of substratum, superstratum and adstratum influence, with reference to Tibeto-Burman languages[C]//Yasuhiko Nagano (ed.), Issues in TibetoBurman Historical Linguistics (Senri Ethnological Studies 75). Osaka: National Museum of Ethnology, 2009: 227-237. （中文译文版：《底层、表层和附层影响的动因和结果— —以藏缅语为例》（郑武曦译），未公开出版，见罗仁地教授个人网站。）  罗仁地, 潘露莉. 菲律宾塔伽洛诺（Tagalog）的词类范畴[J]. 语言学论丛. 2010, 41: 114. [2010 年]  LaPolla, Randy J. Hierarchical person marking in the Rawang language[C]//戴昭铭 编. 14 第40 届国际汉藏语言暨语言学会议论文集. 哈尔滨: 黑龙江大学出版社, 2010: 107113.  LaPolla, Randy J. Language Contact and Language Change in the History of the Sinitic Languages[J]. Procedia-Social and Behavioral Sciences, 2010, 2(05): 6858-6868. [2011 年]  LaPolla, Randy J. On Transitivity in Two Tibeto-Burman Languages[J]. Studies in Language. 2011, 35(03): 637-650.  LaPolla, Randy J., František Kratochvíl & Alexander R. Coupe. On Transitivity[J]. Studies in Language. 2011, 35(03): 469-492. （中文译文版：《论及物性》（杨旭译），未公开出 版，见罗仁地教授个人网站。） [2012 年]  LaPolla, Randy J. Comments on Methodology and Evidence in Sino-Tibetan Comparative Linguistics[J]. Language and Linguistics. 2012, 13(01): 117-132. [2013 年]  LaPolla, Randy J. Eastern Asia: Sino-Tibetan linguistic history[C]//Emmanuel Ness (ed.), The Encyclopedia of Global Human Migration (Prehistory Volume), Chapter 25. Chicester: John Wiley & Sons, Ltd., 2013.  LaPolla, Randy J. Subgrouping in Tibeto-Burman[C]//Balthasar Bickel, Lenore A. Grenoble, David A. Peterson & Alan Timberlake (eds.), Language Typology and Historical Contingency: In honor of Johanna Nichols. Amsterdam: John Benjamins, 2013: 463-474.  LaPolla, Randy J. Arguments for a construction-based approach to the analysis of SinoTibetan languages[C]//Peng Gang & Shi Feng (eds.), Eastward Flows the Great River— Festschrift in Honor of Professor William S-Y. WANG on his 80th Birthday. Hong Kong: City University of Hong Kong Press, 2013: 127-141.  LaPolla, Randy J. Subgrouping in Tibeto-Burman: Can an individual-identifying standard be developed? How do we factor in the history of migrations and language contact?[C]//Balthasar Bickel, Lenore A. Grenoble, David A. Peterson, & Alan Timberlake (eds.), What’s Where Why? Language Typology and Historical Contingency. Amsterdam: John Benjamins, 2013: 463-474.  LaPolla, Randy J. Arguments for a construction-based approach to the analysis of Chinese[C]//Tseng Chiu-yu (ed.), Human Language Resources and Linguistic Typology, Papers from the Fourth International Conference on Sinology. Taiwan: Academia Sinica, 2013: 33-57. [2014 年]  LaPolla, Randy J. Constituent structure in a Tagalog text[J]. Language and Linguistics. 2014, 15(06): 761-774.  Tournadre, Nicolas, Randy J. LaPolla. Towards a New Approach to Evidentiality: Issues 15 and directions for research[J]. Linguistics of the Tibeto-Burman Area. 2014, 37,(02): 240263. [2015 年]  LaPolla, Randy J. Sino-Tibetan Syntax[C]//William S-Y Wang & Chaofen Sun (eds.), The Oxford Handbook of Chinese Linguistics. Oxford: Oxford University Press, 2015: 45-57.  LaPolla, Randy J. On the logical necessity of a cultural and cognitive connection for the origin of all aspects of linguistic structure[C]//Rik de Busser and Randy J. LaPolla (eds.), Language Structure and Environment: Social, cultural and natural factors. Amsterdam: John Benjamins, 2015: 31-44. （中文译文版：《语言结构与文化、认知的逻辑必然联系》（杨 旭译），未公开出版，见罗仁地教授个人网站。） [2016 年]  LaPolla, Randy J. Review of The Language Myth: Why Language is Not an Instinct, by Vyvyan Evans (Cambridge University Press, 2014)[J]. Studies in Language. 2016,40(01): 235-252.  LaPolla, Randy J. On Categorization: Stick to the facts of the languages[J]. Linguistic Typology. 2016, 20(02): 365-375.  LaPolla, Randy J. Once Again on Methodology and Argumentation in linguistics: Problems with the arguments for recasting Sino-Tibetan as ‘Trans-Himalayan’[J]. Linguistics of the Tibeto-Burman Area. 2016, 39(02): 282-297. （中文译文版：罗仁地著, 黄成龙, 宋佳 译. 再论语言学的方法论与论据——对“汉藏语系”改为“跨喜马拉雅语系”的质疑 [J]. 汉藏语学报. 2019, 11: 1-14.）  LaPolla, Randy J. On Scholarship in Sino-Tibetan Linguistics: Review article on Studies in Chinese and Sino-Tibetan Linguistics[C]//Richard VanNess Simmons & Newell Ann Van Auken (eds.), Journal of the American Oriental Society, 2016. [2017 年]  LaPolla, Randy J. An Overview of Sino-Tibetan Morphosyntax[C]//Graham Thurgood & Randy J. LaPolla (eds.), The Sino-Tibetan Languages (2nd Edition). London & New York: Routledge, 2017: 40-69.  LaPolla, Randy J. Noun-Modifying Clausal Constructions in Sino-Tibetan Languages[C]//Yoshiko Matsumoto, Bernard Comrie, & Peter Sells (eds.), NounModifying Clause Constructions in Languages of Eurasia: Reshaping theoretical and geographical boundaries. Amsterdam: John Benjamins, 2017: 91-103. （中文译文版：《汉 藏语的子句修饰名词构式》（郑武曦译），未公开出版，见罗仁地教授个人网站。 ）  LaPolla, Randy J. Causation as a Factor and Goal in Typological Comparisons[J]. Linguistic Typology. 2017, 21(03): 547-554.  罗仁地. 尊重语言事实 提倡科学方法——21 世纪语言学刍议[N]. 光明日报, 201712-03(12). 16 [2019 年]  LaPolla, Randy J. The Origin and Spread of the Sino-Tibetan Language Family[J]. Nature. 2019, 569(7754): 45-47.  LaPolla, Randy J. On the Structure of the Clause in Proto-Sino-Tibetan and its Development in the Daughter Languages[J]. Journal of Chinese Linguistics Monograph Series, 2019, 29: 122-144.  LaPolla, Randy J. Arguments for seeing theme-rheme and topic-comment as separate functional structures[C]//J.R. Martin, Y.J. Doran, & G. Figuedero (eds.), Systemic Functional Language Description: Making meaning matter. London: Routledge, 2019: 162186.  LaPolla, Randy J. A Note on the History of the Term ‘Pronomenalisation’[J]. Linguistics of the Tibeto-Burman Area. 2019, 42(01): 143-147. [2020 年]  LaPolla, Randy J. Forward to the Past: Modernizing linguistic typology by returning to its roots[J]. Asian Languages and Linguistics. 2020, 1(01): 146-166. （中文译文版：《追根 溯源、鉴古知今 ——实现语言类型学的现代化》（杨旭译），未公开出版，见罗仁地 教授个人网站，郑武曦撰写的长摘要见 https://mp.weixin.qq.com/s/VJIrd2708GXGHQGZdb_xfg。）  罗仁地. 历史语言学、语言类型学和语言接触学[J]. 汉藏语学报, 2020(00): 1-12. [2022 年]  罗仁地. 非结构主义语言学. 实验语言学. 2022, 11(03): 1-6.  LaPolla, Randy J. The Chinese expansion and language coexistence in modern China[C]//Salikoko S. Mufwene & Anna María Escobar (eds.), The Cambridge Handbook of Language Contact: Volume 1: Population Movement and Language Change. Cambridge University Press, 2022: 64-83.  LaPolla, Randy J. On the Use of Rhymes in Determining the Structure and Interpretation of Old Chinese Texts: The Case of the DàoDé Jīng《道德經》[J]. Bulletin of Chinese Linguistics. 2022, 15: 1-9.  罗仁地, 潘露莉. 焦点结构的类型及其对汉语词序的影响[C]//徐烈炯, 潘海华 编. 焦点结构和意义的研究(增订版). 上海: 上海教育出版社, 2022: 38-52. [2023 年]  罗仁地. 以人为中心——交叉研究的必然走向[J]. 语言战略研究. 2023, 3: 93-96.  罗仁地. 汉藏语系修饰指称词子句的类型和历史发展——从日旺语（独龙语）说起[J]. 民族语文. 2023(01): 3-17.  LaPolla, Randy J. Grammatical Relations[C]//Delia Bentley, Ricardo Mairal Usón, Wataru Nakamura & Robert D. Van Valin Jr. (eds.), Cambridge Handbook of Role and Reference Grammar. Cambridge: Cambridge University Press, 2023: 269 - 291. 17  罗仁地著, 姚洲译. 人工智能要考虑到溯因推理[J]. 长江学术. 2023(04): 109-113.  LaPolla, Randy J. How understanding population movements influences our understanding of genetic relationships and language history: The case of Mǐn Chinese[C]//彭刚, 孔江平, 沈钟伟, 汪锋 编. 高山仰止——王士元先生 90 华诞庆祝文集. 北京: 北京大学出版 社, 2023: 81-92.  罗仁地. 赵元任先生的句子和论元结构概念[J]. 实验语言学. 2023, 12(04): 4-7. [2024 年]  LaPolla, Randy J., Keita Kurabe. Manifestations of Jinghpaw influence among Rawang speakers [J]. Asian Languages and Linguistics. 2024: 4(02): 273-290.  LaPolla, Randy J. On the Overlooked Diversity of Clause Structures and Argument Structures in Non-Indo-European Languages[J]. Languages. 2024, 9(04): 135-143.  罗仁地, 卢琳. 21 世纪的田野调查—基于自然语料、归纳法和目的语母语人的认知范 畴[J]. 语言学论丛. 2024(04): 14-19。 三、未公开出版的学术会议论文及教学讲义 [1996 年]  On the Place of Chinese in the Typology of Grammatical Relations Systems, lecture given at Linguistics Seminar of the Dept. of Chinese, Translation, and Linguistics, 1996. (Available on personal website of Prof. Randy J. LaPolla.) [1997 年]  Grammaticalization as the Fossilization of Constraints on Interpretation: Towards a single theory of communication, cognition, and the development of language, lecture given at City University of Hong Kong Seminar in Linguistics, 1997. (Available on personal website of Prof. Randy J. LaPolla.)  Language as Culture: The conventionalization of constraints on inference, submitted to Oxford University Press as partof a volume on Ethnosyntax edited by Nick Enfield, 1997. (Available on personal website of Prof. Randy J. LaPolla.) [2006 年]  The How and Why of Syntactic Relations, invited plenary address and keynote of the Centre for Research on Language Change Workshop on Grammatical Change at the Annual Conference of the Australian Linguistics Society, University of Queensland, 7-9 July, 2006. (Available on personal website of Prof. Randy J. LaPolla.)  Clause linking in Dulong-Rawang, invited paper presented to the Workshop on the Semantics of Clause Linking Research Centre for Linguistic Typology, Institute for Advanced Study, La Trobe University, 24 May, 2006. (Available on personal website of Prof. Randy J. LaPolla.) [2007 年]  Word-class-changing Derivations in Rawang, invited paper presented to the Workshop on 24 图2 语言和非语言行为没有太大差别，都是服务于交流格式塔的。罗先生给了神经 科学的证据：功能磁共振成像研究表明，语言和非语言交流的处理位于大脑同一区 域，包括被称为“布罗卡区”和“维尼克区”的区域。罗先生进一步假设，如果我 们设计一个心理学实验，来验证语言交流是否特殊，那么会发现语言与非语言交流 并无差别。这也与认知语言学所强调的领域通用（domain-general）认知能力的观点 相一致：即没有专门用来处理语言的认知能力，用于处理语言的认知能力也在其他 认知领域中发挥作用，语言不过是诸多领域通用认知能力的延伸。我有一门叫作“语 言与符号”的课，会讲到姓名、脏话、唱歌、表情符号（Emoji）这些不被视为语言 的例子，我发现通过这种交流格式塔的视角，学生更容易对语言学产生兴趣。 在交流行为的格式塔中，语言和非语言的角色并列，都是用来制约对交流意图 的解读的，只不过语言更为明确或精细，即“非语言交流和语言交流之间的区别是 工具或方式的区别，进而导致精度的差异，就像手撕面包和刀切面包的区别一样” （LaPolla 2021）。需要注意的是，明确或精细是相对的，比如英语“I have had lunch” （午饭已经吃了）和“I have been to the mainland”（我到过大陆），都使用了强制的过 去完成标记，制约得已经足够明确（我们不可能理解为现在或未来），但是过去是 多久，只有结合环境才能知道：前者是发生在一个小时左右，后者是发生在更久的 过去。更典型的例子是：我们会字斟句酌地制定法律，目的是明确，但仍然需要法 院或法官来解读法律，这也给了坏人在法律灰色地带游走的可乘之机。 正是在上述背景下，罗先生提出“一切都是语用学（It is all pragmatics）”这一 在很多人看来有些激进的说法。既然一切都是语用，那么与意义创造的主观性一样， 交流行为也具有不确定性。交流不像过去认为的，是一种编码和解码的过程，做不 到像邮寄包裹一样，邮寄的是什么，接收到的就是什么。交流主要基于具有概率性 的溯因推理，其中语言相对于非语言固然提供了更多的制约，但永远不可能做到百 25 分之百制约。 2我从罗先生的论述中领会到了辩证法：人类交流既体现出理性，因为 合作原则的核心即是假设人是理性的；也体现出非理性，因为意义创造具有主观性， 交流行为具有不确定性。 三、语言是一套涌现出的文化习俗 2019 年，我在新加坡南洋理工大学访学，我的合作导师林静夏老师推荐我去听罗先 生的课。说实话，我当时没怎么听说过罗先生，因此并没有要去见大师的那种兴奋。只用 了一节课，罗先生便征服了我，听他讲“语言何以不同、为何不同”（HOW & WHY LANGUAGES DIFFER），我有一种一见如故的感觉。抛开人格魅力不谈，这种一见如故 很大程度上是由语言观决定的，因为我们都信奉：语言是一个动态的复杂适应系统。 我在读硕士研究生的时候，就已经跟随王仁强老师学习了相关理论，还翻译了两篇论文 （杨旭、王仁强 译 2015；王仁强、杨旭、李娜 译 2018）。在新加坡听罗先生讲课，我又 找到了一位知己。 还是拿路的例子来说明比较好——这也是复杂科学和罗先生最常用到的例子 ——但不是类似于柏油路那样的路，而是草坪里面众人踩出来的路，英文叫作“欲 望之路”（desire path）、“牛道”（cow path）或“羊道”（goat track）。人们在穿越草坪 时，没有人想着要创造一条路，只是想着两点之间直线最短（即捷径）；一开始也 只是踩踏的痕迹，但是当越来越多的人加入后，量变发生质变，一条路就“涌现” 出来了，或者被人们识别出来了。这让我对鲁迅的名言“其实地上本没有路，走的 人多了，也便成了路”有了不同的认识。语言就像这条路，在前语言时代，没有人 想着要发明一门语言， 3但是当很多人尝试与同类沟通或互动时，语言便慢慢地形成 了。这种形成没有目的或方向，而且永远不会停止，因此语言不是静态的事物，而 是动态的演变，我们所看到的静态的词典、语法书，不过是动态演变的一个快照， 语言唯一不变的是变异和变化。 罗先生认为，语言就像烹饪/饮食或者穿衣一样，本质上是一种文化习俗，其功能 在于制约人类交流中的解读环境；所谓语言知识或规则，不过是我们以前如何使用词语 和结构以达到某种目的的经验。既然是文化，那么就可以施行“拿来主义”。罗先 生举到的例子是英文中的 gungho（“有极大的精神或热情”），来自中文“工合” （“工业合作社”），英文“拿来”之后便完全按照自己的文化来使用。好比东南亚 虽然“拿来”了西方的叉子和勺子，但是使用的方式却非常不同。这些例子显示了 语言和非语言文化之间的类似。 2 也正是在这个意义上，我们可以理解罗先生为什么要用“制约”这个词，即是从反面来排除歧义的， 或者说缩小可供推测的范围，而不是从正面促进“解码”得到语言中“编码”的意义的。 3 当然，在语言诞生之后，人们具备了独立创造一门语言的能力，比如历史上有很多尝试创造世界通用 语言的实践（如世界语），如今为影视剧创造的人造语言也是非常多（如克林贡语）。 26 语言既然是文化，那么也具有惯性特征。罗先生都是从个人角度论述这个特征 的，但我认为应该区分一下个人习惯和集体习惯：个人在习得语言之后便形成习惯， 比如把某些语音归为同一个音位，而把其他语音视为其他音位，在学习第二语言时 想改变这种习惯非常之难，因此口音是很难避免的；集体在涌现出语言之后也表现 出惯性，即语言虽然一直处于动态演变中，但不是极速或动荡的演变，否则很难作 为一种有效的交流工具，与之前强调的语言之“变”相比，集体语言又表现出稳定 性或一致性。 把语言视为一种文化，就不得不谈及语言与文化的关系了，这是“语言与思维” 领域所关心的话题。罗先生一直重视浪漫主义传统（LaPolla 2020），因此对语言与 文化或思维的关系多有论述： 人们的文化和认知（例如，人们如何描述事件）会影响语言，但是一旦成 为语言中的惯例，就会传递给后代，并且会影响人们如何思考这些事情。…… 语言编码了我们对世界的看法，同时也影响了我们对世界的看法（因为我们从 祖先那里学习了这些概念）。（LaPolla 2015） 看得出，罗先生基本上是主张一种互相影响的关系。首先，语言反映出人们的 文化和认知构建。以中文和英文对比为例（来自罗仁地 2017），“汤”和 soup，“南 瓜”和 pumpkin、squash，“有”“在”和 have，它们之间没有绝对的一一对应关系， 因为不同语言反映了不同社群的认知范畴。需要注意的是，这种“反映”可能随着 时间推移而消失，罗先生举了非语言的例子，比如戴领带、在英国喝茶前倒牛奶、 公路上的拟人交警；也举了很多语言的例子，比如“回车键”（carriage return）的叫 法来源于过去打字机纸架回位，“拨电话”（dial a phone）来源于过去的拨号盘，Goodbye 是对“God be with ye”的简化，这些例子都是过去自有其理据，但现在理据消失了， 习俗化的行为留下了。 其次，语言也会影响人们的思维。《话镜》作者盖伊·多伊彻（Guy Deutscher） 说，空间、性别和色彩是迄今为止最有说服力的展示语言对思维有影响的领域 （Deutscher 2010：236）。但对我而言，语言学家不举语言本身的例子是说不过去的， 罗先生就用了语言本身的例子：一个以英语为母语的人真的会把普通话中“北京” 中的开头声音听为/b/，他们也不会听到普通话中四个声调之间的区别。此外还举 了自身的例子：罗先生在美国长大，认为小虾、大虾和龙虾是三种非常不同的动物， 因为它们在英文中有非常不同的名字（shrimp、prawns、lobsters），但是在中文里，它们 都叫“虾”，只不过在大小上有所区别（“小虾、大虾和龙虾”）。 27 四、语言何以不同、为何不同？ 前面讲过，我上过罗先生的一门课，叫作“语言何以不同、为何不同”。正是在这门 课上，我体会到罗先生对一门具体的语言，可以了解得有多么细微。当我们阅读罗先生的 理论文章时，我们觉得他是一个哲学家；但是当我们阅读他的案例文章时，我们发现他是 一个跋山涉水做过很多田野调查的科学家。 语言何以不同？关键词还是“制约”：语言之间的差异不在于能表达或能理解什么， 而是必须表达或必须理解什么；每种语言都有自己独特的发展历史，因重复使用而发生 习俗化之后，会选择对那门语言使用者来说最重要的概念领域进行制约。罗先生把这种 “制约”呈现为 3 条公式，三者之间呈现递进关系（更多的例子参 LaPolla 2003）： （1）语言是否制约对某一特定概念领域的理解？比如，汉语没有标记时的结构，而 英语有好几种标记时的结构。 （2）如果制约了某一特定概念领域的理解，那么程度如何？比如英语将来时的解读 分为过去、现在、未来，而日语分为过去和非过去，日旺语（缅甸克钦邦）分为非过去时 和四种程度的过去时（取决于事件发生在过去多长时间）。 （3）如果制约了某一特定概念领域的理解，那么是如何制约的？比如说：一个说英 语的人想说 He is washing his hair，必须用一个领属结构来制约对头发所有者的识别。在 同样的情况下，日旺语也需要制约头发所有者的身份，但使用的是反身语和中动结构。 在这个框架下，罗先生对语言差异提出更理智的看法：不能简单地说某种语言的语 法化程度更高或对解读的制约更强，只能说某些语言在特定概念领域（如时态、共指）上 选择了更强的制约，而在其他领域可能更为开放。这种多样性揭示了语言适应不同文化、 认知和环境需求的灵活性。 至于语言为何不同，罗先生认为，语言的差异源于环境、社区规模、认知模式和文化 接触的共同作用。语言不是静止的，而是通过无数个体的反复互动和适应逐渐演变的“无 形之手”的产物。其结构和词汇反映了说话者如何理解世界，以及他们面临的沟通需求。 尽管某些语言特征的最初动机可能不再明显，但它们的存在本身就是历史和文化适应的 证据（详参 LaPolla 2003）。 五、倡导整合主义的研究方法 罗先生倡导过很多种研究方法： 跨学科研究。既然语言本质上不是一种固化的形式，而是灵活的文化行为、交流行 为，那么语言学就是一门行为科学，所有有关行为的科学都跟语言学有密切关系，都是一 体的（罗仁地 2017）。这也是为何罗先生会在“2021 北京师范大学珠海校区交叉学科前 沿论坛”的演讲中说：“我的研究领域是语言学，……但是，当有人问我做什么样的研究 的时候，我说我研究人类行为和认知。”（罗仁地 2023a） 28 归纳式方法。罗先生说，语言研究应基于真实语料，观察重复模式及其功能，依语言 自身特性定义其范畴，而不是先入为主地把所谓跨语言范畴套用到未描述的语言身上， 这样才能摆脱掉印欧语系或汉语框架的影响，确保每种语言的分析独立、客观，揭示其现 象多样性（杨旭、卢琳 2021；罗仁地 2022）。罗先生呼应了石锋先生和李宇明先生（石锋 2025），倡议：“新时代的语言学研究，要着重以人为中心，呼吁大家走向社会！走向田野！ 走向实验室！走向互联网！”（罗仁地 2023a） 不区分共时和历时。罗先生指出，21 世纪的语言学研究观点认为，语言并非固定体 系，也无“共时”与“历时”之别，而是通过人与人之间的交互不断涌现。这种视角拓宽 了语言学研究范围，将研究对象扩展至所有交际行为，包括非语言交际和动物交际，而不 仅仅局限于脱离语境的语言形式（罗仁地 2017；罗仁地 2024）。 具体的提倡还有很多，如果要一言以蔽之的话，那就是“整合主义”。罗先生把他所 主张或认同的理论称为“非结构主义”或基于使用（usage-based）的方法，包括意义创造 论、整合语言学（Integrational Linguistics）、极点构式语法（Radical Construction Grammar） 等。与结构主义的还原主义、成分主义和分布主义相比，整合主义以实际的言语单位（语 调单位、构式）为基本单位，并从交际互动的格式塔来看待言语行为，不仅包括语境和语 调，还包括手势、眼神以及交际互动的其他方面（罗仁地 2022）。 罗先生的方法论是如何形成的？你可以说罗先生之所以提倡这样的方法论，是因为 抱有前述认识论；但也可以说，是因为罗先生按照这样的方法论来展开研究，才形成了前 述认识论。这种循环因果也体现在罗先生的论述里边： 语言是一种社会性现象，是一种复杂自适应系统，是很多人进行交际时涌现出 来的副产品。没有一种简单的方法能够解释其所以然。20 世纪时由于结构主义语言 学理论的影响，有不少人认为在研究语言的时候，可以脱离人类行为和认知的有关 方面而单独研究语言的形式，但这种研究方法并没有带给我们有用的结果，由此可 见，我们需要从交际互动的格式塔（完形）以及影响交际互动发生和运作的因素进行 研究。 …… 遵循以上讨论的融合性方法，我对交际的本质以及语言的作用和本质有了新的 理解，而这种新的理解与传统的看法很不一样：我了解到交际不是像电脑一样通过 编码和解码进行的，交际的关键机制是溯因推理法，就是推测对方做出交际行为的 目的。（罗仁地 2023a） 最后做个总结吧。我想说，罗先生不单单是一个语言学家，而是一个富有人文主义精 神的科学家。也许正是这种人文精神，才是我特别喜爱罗先生的深层原因。我现在正在武 汉大学代“人文社科经典导引”的通识课程，讲述仁爱、理性、正义、生死、爱恨、审美、 悟性这些貌似很抽象的话题。这不是一门语言学的课，但我在课上也常引用罗先生的观 点。比如讲到生死问题时，我会引用罗先生跟我们说过的，人生没有什么意义，是我们为 29 之赋予的意义，追求爱与被爱是意义之一种。尤其是在第一课，我在讲为何要设置人文通 识课程时，会引用到罗先生那篇常读常新的演讲。所以，这篇文章以罗先生的引文开头， 最后还是以罗先生的原话结束吧： 人类如果不想被机器取代，就必须回到使我们成为人的本能（get back to what makes us human）。这包括合作学习、创造力、批判性思维和同理心——而这些正是 与艺术和人文教育相关的技能。……。人文学科的伟大作品之所以伟大，是因为它们 拥有一种明显但难以捉摸的能力，照亮我们共同的人性，照亮不同历史情况下的不 同人的生活。这些作品以某种方式超越了自己创造的条件——针对其时代说话，但 也超越其时代。（罗仁地 2023a） 参考文献 罗仁地 2007 语言类型学/功能语言学派视野下的语言学田野调查，《语言学论丛》第 36 辑。 罗仁地 2017 尊重语言事实 提倡科学方法——21 世纪语言学刍议，《光明日报》2017 年 12 月03 日 12版。 罗仁地 2022 非结构主义语言学，《实验语言学》第 3期。 罗仁地 2023a 以人为中心：交叉研究的必然走向，《语言战略研究》第 3期。 罗仁地 2023b 人工智能要考虑到溯因推理，《长江学术》第 4期。 罗仁地 2024 动物交际的机制，“实验语言学+云上论坛”，2024 年03 月26 日，南开大学 （线上）。https://mp.weixin.qq.com/s/obopPfSRuHIPS-HVdc8rNg 罗仁地、卢 琳 2024 21 世纪的田野调查——基于自然语料、归纳法和目的语母语人的 认知范畴，《语言学论丛》第 4期。 罗仁地、潘露莉 2002 信息传达的性质与语言的本质和语言的发展，《中国语文》第 3期。 石 锋 2025 王士元之问——AI 与人类习得语言的方式如何不同，微信公众号“今日语 言学”，2025 年01 月23 日。https://mp.weixin.qq.com/s/JQFbbARW_Fh6bFKB-6ujBA 王仁强、杨 旭、李 娜 译 2018 作为复杂适应系统的自然语言，《英语研究》第 1期:133142。 杨 旭 2024 一项宣言：语言是平的，微信公众号“摩登语言学”，2024 年12 月3日。 https://mp.weixin.qq.com/s/FzmR3lfllkkGEqglMYgnAA 杨 旭 2025 构式网络视角的汉语修辞研究，“构式网络与语法修辞”扁平工作坊，2025 年03 月29-30 日，中国·武汉，武汉大学。 杨 旭、卢 琳 2021 从功能视角看汉语研究若干问题——著名语言学家罗仁地访谈录， 《南开语言学刊》第 1期。 杨 旭、王仁强 译 2015 语言乃是一个复杂适应系统:立场论文，《英语研究》第 2期。 30 Deutscher, Guy. 2010. Through the Language Glass: Why the World Looks Different in other Languages. Metropolitan books. Gigerenzer, Gerd. 2007. Gut Feelings: The Intelligence of the Unconscious. Penguin. Kahneman, Daniel. 2011. Thinking, Fast and Slow. Farrar, Straus and Giroux. LaPolla, Randy J. 2003. Why languages differ: Variation in the conventionalisation of constraints on inference. In David Bradley, Randy J. LaPolla, Boyd Michailovsky & Graham Thurgood (eds.), Language variation: Papers on variation and change in the Sinosphere and in the Indosphere in honour of James A. Matisoff. Pacific Linguistics. LaPolla, Randy J. 1994. An experimental investigation into sound symbolism as it relates to Mandarin Chinese. In Leanne Hinton, Johanna Nichols & John Ohala (eds.), Sound Symbolism, 130–147. Cambridge University Press. 中译版：杨 旭 译 2023 现代汉语 的语音象征性，《实验语言学》第 2期。 https://mp.weixin.qq.com/s/VFYrq5dJFytvylicl2UqJA LaPolla, Randy J. 2015. On the logical necessity of a cultural and cognitive connection for the origin of all aspects of linguistic structure. In Rik De Busser & Randy J. LaPolla (eds.), Language structure and environment, 31–44. John Benjamins. 中译版：杨 旭 译 2022 语言结构与文化、认知的逻辑必然联系。 https://mp.weixin.qq.com/s/SUqUFW8InuW1Atgania9dg LaPolla, Randy J. 2020. Forward to the past: Modernizing linguistic typology by returning to its roots. Asian Languages and Linguistics 1(1): 147–167. 中译版：杨 旭 译 2022 追根 溯源、鉴古知今——实现语言类型学的现代化。https://mp.weixin.qq.com/s/3XDviv7ooIJBMxhpog40A LaPolla, Randy J. 2021. The Creation of Meaning, 05-21-2021, University of Sydney. https://www.bilibili.com/video/BV1DB4y1T7Qa/?from=search&seid=757983767988051 4828&spm_id_from=333.337.0.0&vd_source=cc2dffce0723096c58949c3055dabe58 Ungerer, Friedrich & Schmid, Hans-Jorg. 2006. An Introduction to Cognitive Linguistics (2nd ed.). Routledge. 31 3 罗仁地先生对汉藏语研究的贡献 黄成龙 马王储 中国社会科学院民族学与人类学研究所 内容摘要：罗仁地（Randy J. LaPolla）先生是国际知名的汉藏语系研究学者， 其学术贡献涵盖历史语言学、语言类型学、描写语言学、田野语言学、语法 理论及语言接触研究等领域。罗仁地先生通过实证研究和跨学科方法，对汉 藏语系的溯源、演化和结构特征提出了很多创新性观点。在方法论上，他强 调必须避免主观臆断，采用自下而上与自上而下相结合的方式，结合人口迁 徙史和语言接触情况，开展汉藏语溯源研究。 一、罗仁地先生的学术生涯 罗仁地（Randy John LaPolla）先生 1955 年10 月生于美国纽约。1981 年进入北京大 学中文系研修语言学专业，这段经历成为罗仁地先生致力于汉语语言学和少数民族语言 学研究的开端。罗仁地先生 1983 年进入加利福尼亚大学伯克利分校（University of California, Berkeley）攻读博士学位，师从马提索夫（James A. Matisoff），跟随王士元、万 瓦林（Robert D. Van Valin, Jr）、Johanna Nichols 等先生系统学习语言学。1990 年获加利 福尼亚大学伯克利分校语言学博士学位。1983 年至 1998 年期间，罗仁地先生先后任 Eurasia 出版社（Eurasia Press, New York）编辑、作者、译员及研究与发展部主任（Director of Research and Development）、华东师范大学继续教育学院教师（Instructor）、加利福尼亚 大学伯克利分校东方语言系（Department of Oriental Languages）教师（Instructor）、研究 助理（Research Assistant）。1990 年至 1997 年在中央研究院（Academia Sinica, Taipei）历 史语言研究所人助研究员（Research Assistant Fellow）、副研究员（Associate Research Fellow）。1997 年至 2004 年在香港城市大学（City University of Hong Kong）任副教授 （Associate Professor）。2004 年至 2012 年在拉筹伯大学（La Trobe University, Melbourne, Australia）先后任/兼任首席教授（Chair Professor）、语言学系主任（Head of Linguistics Programme）、语言类型学研究中心主任（Director, Research Centre for Linguistic Typology）。 2012 年任中央研究院首席教授。2012 年至 2021 年在南洋理工大学（Nanyang Technological 黄成龙、马王储. 2025. 罗仁地先生对汉藏语研究的贡献. 黄成龙、杨旭（主编）《意义创造论：罗仁地 先生七秩寿庆文集》，pp. 31–49.  32 University, Singapore）任/兼任人文与社会科学学院语言学与多语学习系主任（Head, Division of Linguistics and Multilingual Studies, School of Humanities and Social Sciences）、 教授（Professor）。2021 年退休后，罗仁地先生进入北京师范大学珠海校区人文和社会科 学高等研究院语言科学研究中心任特聘教授（2021 年至 2023 年）。罗仁地先生至今仍笔 耕不辍，为世界汉藏语言文字研究作出突出贡献。 罗仁地先生自 1987 年以来，一直从事世界汉藏语言文字研究工作。罗仁地先生尤其 关注汉藏语系藏缅语族，特别是羌语和独龙语/日旺语，研究领域涉及语言描写、语言比 较、语言起源与原始语言构拟、语言亲属关系、濒危语言问题、民族语言长篇语料记录、 词典编纂等方面，出版专著 20 余部，发表论文 100 余篇，是世界汉藏语研究领域的知名 学者。本文拟就罗仁地先生汉藏语源流的研究和学术影响进行简要介绍。 二、汉藏语溯源研究 汉藏语溯源（构拟）历来是学者们关注的核心和热点问题之一。罗仁地先生在一系列 论著和论文中对汉藏语溯源的方法论、汉藏语系的形成和演变、构拟原始汉藏语形态句 法特征等方面均提出了重要的学术观点。 2.1 汉藏语研究的方法论 研究方法的科学性决定汉藏语研究进展方向和研究结果的正确性，罗仁地先生在很 多论著、论文中都对汉藏语研究的方法论问题展开探讨。 1. 汉藏语溯源的方法 汉藏语溯源的研究必须要在比较科学的方法论基础上采取自下而上与自上而下相结 合的方法（黄成龙、马王储 2025）。罗仁地先生在多篇论文中对原始汉藏语构拟时应当采 取的方法问题进行论证。 （1）汉藏语溯源与民族人口迁移史研究、语言接触研究相结合 罗仁地先生（2000[2013a]，2001，2006，2012a，2012b，2019）引用人口统计学的证 据，指出要正确理解汉藏语的历史发展，必须先了解语言使用者的历史，特别是迁徙历 史。这是因为语言使用者的迁徙历史可以帮助我们理解该语系为什么是现在的样貌。 罗仁地先生（2000[2013a]，2001，2006，2012a，2012b，2019）由此进一步指出汉藏 语系语言的使用者在历史上经历了数次迁徙，并在迁徙的过程中不断与其他语言文化相 接触，这些接触是汉藏语系形成的主要影响因素之一。因此，他强调语言接触不仅应该纳 入考虑范围，而且必须视为语言形成的基本要素。共同的发生学来源、分离性的人口迁移 （也就是这些分离之后发生的创新）和语言接触是形成今天的汉藏语系的三项主要因素。 （2）汉藏语溯源的计量研究方法 罗仁地先生（2012a）指出，在原始语言构拟的过程中，有些学者会把他们最熟悉的 33 语言当作是那个语系最保守、最原始的语言，因此会把所构拟的语言构拟成很像他们所 熟悉的语言（即 Teeter 定律）。这种现象的核心问题在于缺乏一套经得起统计学评价的公 认标准。对此，罗仁地先生（2000[2013a]，2012a）提出一种替代性研究方法——通过采 用包括人称标记、反身态/中动态标记及其他显著特征在内的类词性变化集合（paradigmlike sets, sets of features）来构拟原始汉藏语的面貌。 2. 语言同源关系判定的方法 除用以构拟原始汉藏语面貌外，类词性变化集合也应该作为判定语言同源关系的基 本方法（罗仁地 2000，2012a）。罗仁地先生（2000[2013a]，2003）指出，以往常见的研 究方法是使用随机挑选的词汇列表建立语族中的亲缘关系，这种做法得到的结果往往经 不起统计学检验。罗仁地先生建议，要运用历史比较法确立谱系关系，应使用新语法学派 的方法论（Nichols 1996，另参看 Ringe 1992，1995，1996，1999；Campbell 2003），即找 到统计意义上极不可能偶然发生的对应特征。要实现这一目标，关键在于将同源成分（词 汇或形态标记）组织成类词性变化集合，并将这些集合——而非单个元素——严格运用 于亲缘关系判定。 2.2 汉藏语同源词溯源 汉藏语同源词是追溯汉藏语历史的重要证据，但同时也是学者们分歧较大的问题之 一（黄成龙、马王储 2025）。在藏缅语族的很多语言中都可以找到形式明显相关但有一个 音段不同的词（罗仁地 1994b）。罗仁地先生（1994b）指出，有些词的变体形式（variation） 明显受规则支配（rule-governed）（例如，藏语中的齿音后缀可将动词名物化或致使化）， 这种有规律的变体形式有助于原始汉藏语的构拟。 但罗仁地先生（1994b）也强调，即使面对是有规律的变体形式，研究者也必须确认 严格齐整的对应关系。例如，即使我们并不知道藏缅语中*la 和*lap 为变体形式，但由于 古汉语*rap“（树）叶”和藏缅语*la“（树）叶”的声母和韵母均齐整对应，并且词义相 同，因此我们也可以确定古汉语*rap“（树）叶”和藏缅语*la“（树）叶”是同源词。但 需要注意的是，如果必须对每个音段的变体都做出解释，那么除非每个音段的变体均能 找到多个个案的对应关系，否则不该强行构拟同源词。 同时，只有在同语族的语言和方言中找到相同类型的变体，才能构拟原始语言中的 词。如果某种语言不具有特定形态性特征，那么该语言可能为特例。另外不同语支的语言 可能具有不同类型的变体。当无法确定哪种变体形式出现更早时，必须构拟该语族中所 有可能的变体形式。 2.3 藏缅语形态句法特征溯源 罗仁地先生还就原始藏缅语形态句法特征构拟发表了一系列论文。他指出藏缅语中 能够找到语言间平行发展的例证。特定类型的语法化现象往往出现在该语族不同支系中， 有时甚至使用（词源上）相同的语素。该语族众多语言之所以频繁语法化出相同类型的语 法范畴，并惯用相同语素实现这一过程，实则是受到“语言本质特性”的演变规律影响所 40 罗仁地先生（LaPolla 2008）以自然语篇材料为基础，探讨了缅甸北部藏缅语族日旺 语中的关系子句和名词补足语结构的类型及功能，并结合藏缅语族整体特征，分析了日 旺语关系子句的历史发展轨迹。罗仁地先生在研究中发现了两种前置型关系子句。其中， 一类带有名物化标记，另一类无名物化标记。由于名物化结构本身具有名词性属性，因此 也可通过同位结构后置于中心词。研究还结合藏缅语系整体特征， h. 及物性和谐 在藏缅语族不同语言中，动词及物性在不同语言系统中的重要性存在较大差异。罗 仁地先生（LaPolla 2008）对缅甸北部日旺语的动词及物性进行考察，发现该语言所有动 词均通过形态明确区分及物性，并拥有丰富的增价/减价词缀。日旺语中的一个重要现象 是“及物性和谐”，即所有助动词本质均为及物动词，与句中的主要及物动词连用时，助 动词直接接在主要及物动词后，与主要及物动词共享一套及物性标记形态；若主要动词 为不及物动词，则助动词必须通过反身态/中动态后缀转换为不及物形式以实现和谐。此 规律在主要动词为名物化形式时依然成立。罗仁地先生指出“及物性和谐”应当被看作一 项类型学特征。 i. 人称标记 罗仁地先生（LaPolla 2010）发现，日旺语的谓语动词上需要添加人称标记，且人称 标记呈现出层级性（hierarchical）特征。第一人称通过多种方式（后缀、词尾辅音变化、 元音延长等，且单个动词复合体内最多可出现五次标记）获得优先标记；第二人称则采用 逆标记机制，即当存在第一人称参与者但非施事者，且第二人称参与时才会显现。罗仁地 先生指出，这种标记系统与语义角色标记系统（如羌语）或语法关系标记系统（如英语） 存在本质差异。 四、罗仁地先生的语言学观 罗仁地先生对世界汉藏语语言学研究的贡献不仅体现在对汉藏语源流和演化的探索， 也体现在田野语言学、描写语言学、词典与长篇语料编纂等方面。此外，罗仁地先生也具 有前沿的语言学观和创新性的语言学研究视野。 4.1 以语言事实为优先，以类型学与功能为导向 罗仁地先生反对预设普遍语法范畴，并在研究中多次强调语言学研究必须尊重语言 事实（Van Valin & LaPolla 1997；LaPolla 2002，2006a-d，2008a，b，2016a-c）。首先，罗 仁地先生呼吁在类型学和历史比较研究中坚持实证和逻辑严谨性，避免主观臆断。例如， 罗仁地先生坚持同源词判定需满足音义对应规律，构拟需基于多语言变体证据，避免孤 立分析，并提倡通过形态-句法特征集合（如人称标记、示证系统）验证语言亲缘关系。 其次，罗仁地先生反对结构主义的先入为主，强调对具体语言的研究必须从真实自 41 然语料出发，发掘语言的特异性特征（LaPolla 2006a, 2007, 2016a-c）。例如，罗仁地先生 批评了 SVO、SOV 等分类忽略了语言事实，主张描述语言中决定词序的实际原则（LaPolla & Poa 2006），强调信息结构（如话题-述题）对汉藏语句法的塑造作用，提倡以构式为基 础分析汉藏语句法（Van Valin & LaPolla 1997；LaPolla 2008b，2013b）。 4.2 语言演化的多维驱动视角 罗仁地先生（LaPolla 2000 [2013a]，2001，2003，2011，2012a，）在研究中反复强调 语言演化的多维动因。他指出，语言演化是发生学遗存、人口迁徙和语言接触共同作用的 结果，并特别强调接触语言学视角在谱系树模型研究中的主要性。 4.3 跨学科整合 罗仁地先生在语言学研究中融合了认知科学、生物学、计算机科学、历史学等学科的 观点与方法，深化对语言现象的解释。例如，罗仁地先生（LaPolla 2001）强调语言与文 化、认知密不可分。当文化接触的程度足够深，接触会慢慢改变借用者对某些事项的概 念。青海五屯话（陈乃雄 1982）是一种深度藏语化的汉语方言。五屯话不使用两个词分 别表示“寡妇”和“鳏夫”，而是和藏族人一样，用汉语“寡妇”同时表示“寡妇”和“鳏 夫”。再如，罗仁地先生在一些列论文中指出汉藏语系承载着诸多历史东亚族群（包括汉 族、藏族和缅族等）的迁徙历史，并指出汉藏语系的历史与中国新石器时代和青铜时代的 发展密不可分，并通过一项精确的计算分析方法证实，汉藏语系起源于中国北方，其分支 分化始于约 5900 年前（参看 LaPolla 2001，2006，2012b，2019，2022）。又如，罗仁地 先生通过声学实验证明，与很多语言相似，汉语中也存在用高音质音段（acute segments） 关联“小”范畴的词汇，以低音质音段（grave segments）关联“大”范畴的词汇（LaPolla 1994）。 五、结语 罗仁地先生（Randy J. LaPolla）的研究涵盖了汉藏语系的历史语言学、语言类型学、 语法理论、语言接触与演变、田野调查方法等多个领域，其学术思想具有鲜明的跨学科性 和批判性。罗仁地先生的研究以实证为基础，融合历史语言学、语言类型学和功能语法， 推动了汉藏语系研究的科学化。其批判性思维和对语言多样性的尊重，为语言学方法论 提供了重要反思。罗仁地先生曾提问：“语言何以不同、为何不同”，这个问题仍需语言学 者们不断探索。 42 参考文献 黄成龙、马王储. 孙宏开先生的汉藏语源流研究及其学术影响[J]. 中国藏学, 2025(01):179192+220. 罗仁地. 2007. 语言类型学/功能语言学派视野下的语言学田野调查. 语言学论丛 36.0: 4256. 罗仁地. 2020. 历史语言学、语言类型学和语言接触学. 载戴庆厦编, 汉藏语学报, 1-12. 北 京: 商务印书馆. 罗仁地. 2023. 汉藏语系修饰指称词子句的类型和历史发展——从日旺语（独龙语）说起. 民族语文 1: 3-17. 罗仁地. 尊重语言事实 提倡科学方法——21 世纪语言学刍议[N]. 光明日报, 2017-1203(12). 罗仁地, 潘露莉. 2005. 焦点结构的类型及其对汉语词序的影响. 载徐烈炯, 潘海华编. 焦 点结构和意义的研究, 57-78. 广州: 暨南大学出版社. 罗仁地，杨将领. 1996. 独龙语/日旺语动词的返身态和中间态标志.载戴庆厦等编. 中国民 族语言论丛, 13-34. 北京: 中央民族大学出版社. 罗仁地, 杨将领. 2007. 通过亲属语言、方言的比较了解语言的历史发展[J]. 东方语言学 1: 11-18. 潘露莉、罗仁地. 信息传达的性质与语言的本质和语言的发展[J]. 中国语文. 2002(03): 203-209. Bauman, James J. 1979. An historical perspective on ergativity in Tibeto-Burman. In Frans Plank (ed.), Ergativity: towards a theory of grammatical relations, 419-433. London: Academic Press. Campbell, Lyle. 2003. How to show languages are related: Methods for distant genetic relationship. In Brian D. Joseph & Richard D. Janda (eds.), The Handbook of Historical Linguistics, 262-282. Oxford: Blackwell Publishing Ltd. DeLancey, Scott. 1989. Verb agreement in Proto-Tibeto-Burman. Bulletin of the School of Oriental and African Studies 52.2: 315-333. Dryer, Matthew S. 1986. Primary objects, secondary objects, and antidative. Language 62: 808845. Herring, Susan C., John C. Paolillo. 1993. Focus position in SOV languages. In P. Downing and M. Noonan (eds.), Word Order in Discourse, 163-198. Amsterdam/Philadelphia: John Benjamins Pub. Co. Hetzron, Robert. 1975. The presentative movement, or why the ideal word order is V.S.O.P. In Li C. N. (ed.), Word Order and Word Order Change, 345-388. Austin: University of Texas Press. Huang Chenglong. 1994. Qiangyu xingrongc yanjiu (Study on the Adjectives of the Qiang 43 Language). Yuyan Yanjiu,1994: 2.181-9. Lambrecht, Knud. 1986. Topic, Focus, and the Grammar of Spoken French. PhD. dissertation, UC Berkeley. Lambrecht, Knud. 1987. Sentence focus, information structure, and the thetic-categorial distinction. Proceedings of the 13th Annual Meeting of the Berkeley Linguistics Society, 13:366-382. Lambrecht, Knud. 1988. There was a farmer had a dog: Syntactic amalgams revisited. Proceedings of the 14th Annual Meeting of the Berkeley Linguistics Society. Lambrecht, Knud. 1989. When subjects behave like objects: A markedness analysis of sentence focus constructions across languages. (Revised and expanded version of 1987 LSA paper). LaPolla, Randy J. 1987. Dulong and Proto-Tibeto-Burman[J]. Linguistics of the Tibeto-Burman Area 10.1: 1-42. LaPolla, Randy J. 1988a. Topicalization and the question of lexical passives in Chinese. In Marjorie K. M. Chan & Thomas Boyden (eds.), Proceedings of the Third Annual Ohio State University Conference on Chinese Linguistics, 1-17. Columbus: Ohio State University. LaPolla, Randy J. 1988b. Subject and referent tracking: Arguments for a discourse-based grammar of Chinese. In Joseph Emonds, P. J. Mistry, Vida Samiian & Linda Thornburg (eds.), Proceedings of the 7th West Coast Conference on Linguistics, 160-173. Irvine: University of California. LaPolla, Randy J. 1989. Verb agreement, head-marking vs. dependent-marking and the ‘deconstruction’ of Tibeto-Burman morphosyntax. In Kira Hall, Michael Meacham & Richard Shapiro (eds.), Proceedings of the Fifteenth Annual Meeting of the Berkeley Linguistics Society, 356-366. Lexington: Linguistic Society of America. LaPolla, Randy J. 1990. Grammatical relations in Chinese: Synchronic and diachronic considerations. PhD dissertation, University of California, Berkeley. LaPolla, Randy J. 1992a. On the dating and nature of verb agreement in Tibeto-Burman. Bulletin of the School of Oriental and African Studies 55.2: 298-315. LaPolla, Randy J. 1992b. Anti-ergative marking in Tibeto-Burman. LTBA 15.1: 1-9. LaPolla, Randy J. 1993. Arguments Against ‘Subject’ and ‘Direct object’ as Viable Concepts in Chinese[J]. Bulletin of the Institute of History and Philology, 63.4: 759-813. LaPolla, Randy J. 1994a. Parallel grammaticalizations in Tibeto-Burman: Evidence of Sapir’s‘drift’. LTBA 17.1: 61-80. LaPolla, Randy J. 1994b. Variable finals in Proto-Sino-Tibetan. Bulletin of the Institute of History and Philology 65.1: 131-173. LaPolla, Randy J. 1994c. An experimental investigation into phonetic symbolism as it relates to Mandarin Chinese. In Leanne Hinton, Johanna Nichols & John Ohala (eds.), Studies in Sound Symbolism, 130-147. Cambridge: Cambridge University Press. 44 LaPolla, Randy J. 1995a. Ergative marking in Tibeto-Burman. In Yoshio Nishi, James A. Matisoff, & Yasuhiko Nagano (eds.), New horizons in Tibeto-Burman morpho-syntax (Senri Ethnological Studies 41), 189-228. Osaka: National Museum of Ethnology. LaPolla, Randy J. 1995b. On the utility of the concepts of markedness and prototypes in understanding the development of morphological systems. Bulletin of the Institute of History and Philology 66.4: 1149-1185. LaPolla, Randy J. 1995c. An experimental investigation into phonetic symbolism as it relates to Mandarin Chinese. In Leanne Hinton, Johanna Nichols. & John J. Ohala (eds.), Sound Symbolism, 130-147. Cambridge: Cambridge University Press. LaPolla, Randy J. 1995d. On the change to verb-medial word order in proto-Chinese: Evidence from Tibeto-Burman. In Kitamura Hajime, Nishida Tatsuo & Yasuhiko Nagano (eds.), Current Issues in Sino-Tibetan Linguistics, 98-104. Osaka: National Museum of Ethnology. LaPolla, Randy J. 1995e. Pragmatic relations and word order in Chinese. In Pamela Downing & Michael Noonan (eds.), Word Order in Discourse, 297-329. Amsterdam: John Benjamins. LaPolla, Randy J. 1996. Middle voice marking in Tibeto-Burman. In Maria Teresa Lino, Maria Francisca Xavier, Fátima Ferreira, Rute Costa & Raquel Silva (eds.), Proceedings of the Fourth International Symposium on Languages and Languages and Linguistics: Pan-Asiatic Linguistics, 1-15. Lisbon: European Language Resources Association. LaPolla, Randy J. 2000. Valency-changing derivations in Dulong/Rawang. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Changing Valency: Case Studies in Transitivity, 282-311. Cambridge: Cambridge University Press. LaPolla, Randy J. 2001a. Dulong Texts: Seven fully analyzed narrative and procedural texts[J]. Linguistics of the Tibeto-Burman Area 24.2: 1-39. LaPolla, Randy J. 2001b. The role of migration and language contact in the development of the Sino-Tibetan language family. In R. M. W. Dixon & A. Y. Aikhenvald (eds.), Areal diffusion and genetic inheritance: Case studies in language change, 225-254. New York & Oxford: Oxford University Press. LaPolla, Randy J. 2003a. An overview of Sino-Tibetan morphosyntax. In Graham Thurgood & Randy J. LaPolla (eds.), The Sino-Tibetan languages, 22-42. London & New York: Routledge. LaPolla, Randy J. 2003b. Evidentiality in Qiang. In Alexandra Y. Aikhenvald & R.M.W. Dixon (eds.), Studies in Evidentiality (Typological Studies in Language) , 63-78. Amsterdam: John Benjamins. LaPolla, Randy J. 2003c. English-Qiang Glossary by Semantic Field. In Ikeda Takumi (ed.), Basic Materials on Minority Languages in East and Southeast Asia (Endangered Languages of the Pacific Rim series, A03-004) , 153-181. Suita: Faculty of Informatics, Osaka Gakuin University. 45 LaPolla, Randy J. 2003d. Why languages differ: Variation in the conventionalization of constraints on inference. In David Bradley, Randy J. LaPolla, Boyd Michailovsky & Graham Thurgood (eds.), Language Variation: Papers on Aariation and Change in the Sinosphere and in the Indosphere: In honour of James A. Matisoff, 113-144. Canberra: Australian National University. LaPolla, Randy J. 2003e. Qiang. In Graham Thurgood & Randy J. LaPolla (eds.), The SinoTibetan Languages, 573-587. London & New York: Routledge. LaPolla, Randy J. 2003f. Dulong . In Graham Thurgood & Randy J. LaPolla (eds.), The SinoTibetan Languages, 674-682. London & New York: Routledge. LaPolla, Randy J. 2004. On nominal relational morphology in Tibeto-Burman. In Ying-jin Lin, Fang-min Hsu, Chun-chih Lee, Jackson T.-S. Sun, Hsiu-fang Yang, and Dah-an Ho (eds.), Studies on Sino-Tibetan Languages: Papers in Honor of Professor Hwang-cherng Gong on his Seventieth Birthday, 43-74. Taipei: Institute of Linguistics, Academia Sinica. LaPolla, Randy J. 2005. The inclusive-exclusive distinction in Tibeto-Burman languages. In Elena Filimonova (ed.), Clusivity: Typology and Case Studies of the Inclusive–exclusive Distinction, 291-311. Amsterdam: John Benjamins. LaPolla, Randy J. (Luo Rendi). 2006a. Lishi yuyanxue he yuyan leixingxue (Historical linguistics and typology). Journal of Peking University (Philosophy and Social Sciences) 43.2 (March 2006): 27-30. LaPolla, Randy J. 2006b. Sino-Tibetan languages. In Keith Brown (ed.), Encyclopedia of Language and Linguistics, 2nd Edition, 393-397. London: Elsevier. LaPolla, Randy J. 2006c. The how and why of syntactic relations. Invited plenary address and keynote of the Centre for Research on Language Change Workshop on Grammatical Change at the Annual Conference of the Australian Linguistics Society, University of Queensland, 7-9 July, 2006. LaPolla, Randy J. 2006d. On grammatical relations as constraints on referent identification. In Tasaku Tsunoda and Taro Kageyama (eds.), Voice and grammatical relations: Festschrift for Masayoshi Shibatani (Typological Studies in Language), 139-151. Amsterdam & Philadelphia: John Benjamins Pub. Co. LaPolla, Randy J. 2006e. Copula constructions in Rawang. In Ho Dah-an, H. Samuel Cheung, Pan Wuyun & Wu Fuxiang (eds.), Linguistic Studies in Chinese and Neighboring Languages: Festschrift in Honor of Professor Pang-hsin Ting on His 70th Birthday, 10591076. Taipei: Institute of Linguistics, Academia Sinica. LaPolla, Randy J. 2008a . Nominalization in Rawang[J]. Linguistics of the Tibeto-Burman Area 31.2: 45-66. LaPolla, Randy J. 2008b. Questions on transitivity. Keynote presentation to open the Workshop on Transitivity, Research Centre for Linguistic Typology, La Trobe University, 21 August. 46 (Revised version published as LaPolla, Kratockvíl & Coupe 2011.) LaPolla, Randy J. 2008c. Relative Clause Structures in the Rawang Language[J]. Language and Linguistics 9.4: 797-812. LaPolla, Randy J. 2008d . Language Contact and Language Change in the History of the SinoTibetan Languages. In the Academic Steering Committee of the Beijing Forum, Peking University (ed.), The Harmony of Civilizations and Prosperity for All: Diversity in the development of human civilizations (Selected papers from the Beijing Forum, Peking University and Great Hall of the People, Beijing, 2-4 November, 2007) , 144-162. Beijing: Peking University Press. LaPolla, Randy J. 2008e . ‘Transitivity harmony’ in the Rawang language of northern Myanmar. In Louise de Beuzeville & Pam Peters (eds.), From the Southern Hemisphere: Parameters of language variation—E-Proceedings of the 2008 Conference of the Australian Linguistics Society, 1-9. Sydney: University of Sydney. LaPolla, Randy J. 2009a. Chinese as a topic-comment (not topic-prominent and not svo) language. In Janet Xing (ed.), Studies of Chinese Linguistics: Functional Approaches, 9-22. Hong Kong: Hong Kong University Press. LaPolla, Randy J. 2009b. Causes and effects of substratum, superstratum and adstratum influence, with reference to Tibeto-Burman languages. In Yasuhiko Nagano (ed.), Issues in TibetoBurman Historical Linguistics (Senri Ethnological Studies 75) , 227-237. Osaka: National Museum of Ethnology. LaPolla, Randy J. 2010a. Language Contact and Language Change in the History of the Sinitic Languages[J]. Procedia-Social and Behavioral Sciences 2.5: 6858-6868. LaPolla, Randy J. 2010b. Hierarchical person marking in the Rawang language. 载戴昭铭编. 第40 届国际汉藏语言暨语言学会议论文集, 107-113. 哈尔滨: 黑龙江大学出版社. LaPolla, Randy J. 2011. On Transitivity in Two Tibeto-Burman Languages[J]. Studies in Language 35.3: 637-650. LaPolla, Randy J. 2012a. Comments on methodology and evidence in Sino-Tibetan comparative linguistics. Language and Linguistics 13.1: 117-132. LaPolla, Randy J. 2012b. Once again on person-marking in Tibeto-Burman: A reply to DeLancey 2010. Paper presented at the 45th International Conference on Sino-Tibetan Languages and Linguistics, Nanyang Technological University, 26-28 October. LaPolla, Randy J. 2013a. Subgrouping in Tibeto-Burman: Can an individual-identifying standard be developed? How do we factor in the history of migrations and language contact?. In Balthasar Bickel, Lenore A. Grenoble, David A. Peterson, & Alan Timberlake (eds.), What’s Where Why? Language Typology and Historical Contingency, 463-474. Amsterdam: John Benjamins. LaPolla, Randy J. 2013b. Arguments for a construction-based approach to the analysis of Chinese. 47 In Tseng Chiu-yu (ed.), Human Language Resources and Linguistic Typology, Papers from the Fourth International Conference on Sinology, 33-57. Taiwan: Academia Sinica. LaPolla, Randy J. 2015a. Sino-Tibetan Syntax. In William S-Y Wang & Chaofen Sun (eds.), The Oxford Handbook of Chinese Linguistics, 45-57. Oxford: Oxford University Press. LaPolla, Randy J. 2015b. On the logical necessity of a cultural and cognitive connection for the origin of all aspects of linguistic structure. In Rik de Busser and Randy J. LaPolla (eds.), Language Structure and Environment: Social, cultural and natural factors, 31-44. Amsterdam: John Benjamins. LaPolla, Randy J. 2016a. Review of The Language Myth, by Vyvyan Evans. Studies in Language 40.1: 235-252. LaPolla, Randy J. 2016b. Once Again on Methodology and Argumentation in linguistics: Problems with the arguments for recasting Sino-Tibetan as ‘Trans-Himalayan’. Linguistics of the Tibeto-Burman Area 39.2: 282-297. LaPolla, Randy J. 2016c. On categorization: Stick to the facts of the languages. Invited position paper for special issue of Linguistic Typology on descriptive vs. comparative categories, 20.2 : in press. LaPolla, Randy J. 2016d. On scholarship in Sino-Tibetan linguistics: Review article on Studies in Chinese and Sino-Tibetan linguistics, ed. by Richard VanNess Simmons and Newell Ann Van Auken. Journal of the American Oriental Society 136.3: . LaPolla, Randy J. 2017a. An Overview of Sino-Tibetan Morphosyntax. In Graham Thurgood & Randy J. LaPolla (eds.), The Sino-Tibetan Languages (2nd Edition), 40-69. London & New York: Routledge. LaPolla, Randy J. 2017b. Noun-Modifying Clausal Constructions in Sino-Tibetan Languages. In Yoshiko Matsumoto, Bernard Comrie, & Peter Sells (eds.), Noun-Modifying Clause Constructions in Languages of Eurasia: Reshaping theoretical and geographical boundaries, 91-103. Amsterdam: John Benjamins. LaPolla, Randy J. 2017c. Causation as a Factor and Goal in Typological Comparisons. Linguistic Typology 21.3: 547-554. LaPolla, Randy J. 2019a. A Note on the History of the Term ‘Pronomenalisation’. Linguistics of the Tibeto-Burman Area 42.1: 143-147. LaPolla, Randy J. 2019b. The Origin and Spread of the Sino-Tibetan Language Family. Nature 569.7754: 45-47. LaPolla, Randy J. 2019c. On the Structure of the Clause in Proto-Sino-Tibetan and its Development in the Daughter Languages. Journal of Chinese Linguistics Monograph Series 29.0: 122-144. LaPolla, Randy J. 2019d. Arguments for seeing theme-rheme and topic-comment as separate functional structures. In J.R. Martin, Y. J. Doran, & G. Figuedero (eds.), Systemic 48 Functional Language Description: Making meaning matter, 162-186. London: Routledge. LaPolla, Randy J. 2020. Forward to the Past: Modernizing linguistic typology by returning to its roots[J]. Asian Languages and Linguistics 1.1: 146-166. LaPolla, Randy J. 2022. The Chinese expansion and language coexistence in modern China. In Salikoko S. Mufwene & Anna María Escobar (eds.), The Cambridge Handbook of Language Contact: Volume 1: Population Movement and Language Change, 64-83. Cambridge University Press. LaPolla, Randy J. 2023. How understanding population movements influences our understanding of genetic relationships and language history: The case of Mǐn Chinese. 载 彭刚, 孔江平, 沈钟伟, 汪锋编. 高山仰止——王士元先生 90 华诞庆祝文集, 81-92. 北京: 北京大学出版社. LaPolla, Randy J. 2024. On the Overlooked Diversity of Clause Structures and Argument Structures in Non-Indo-European Languages. Languages 9.4: 135-143. LaPolla, Randy J., Dora Poa. 2003. Texts in the Qugu variety of Northern Qiang. In Kitano Hiroaki (ed.), Descriptive and Theoretical Studies in Minority Languages of East and Southeast Asia. (Endangered Languages of the Pacific Rim Project Series), 77-94. Suita: Faculty of Informatics, Osaka Gakuin University. LaPolla, Randy J., Dora Poa. 2006. On Describing word order. In Felix Ameka, Alan Dench, & Nicholas Evans (eds.), Catching Language: The Standing Challenge of Grammar Writing, 269-295. Berlin: Mouton de Gruyter. LaPolla, Randy J., František Kratochvíl & Alexander R. Coupe. 2011. On Transitivity[J]. Studies in Language., 35(03): 469-492. LaPolla, Randy J., Huang Chenglong. 2004. Adjectives in Qiang. In R. M. W. Dixon & Alexandra Y. Aikhenvald (eds.), Adjective Classes: A cross-Linguistic typology, 306-322. Oxford: Oxford University Press. LaPolla, Randy J., Huang Chenglong. 2007. The Copula and Existential Verbs in Qiang[J]. Bulletin of Chinese Linguistics 1.2: 233-248. LaPolla, Randy J., Yang Jiangling. 2005. Reflexive and middle marking in Dulong/Rawang[J]. Himalayan Linguistics 2.0: 1-13. Nichols, Johanna. 1996. The comparative method as heuristic. In Mark Durie & Malcolm Ross (eds.), The Comparative Method Reviewed: Regularity and irregularity in language change, 39-71. New York: Oxford University Press. Poa, Dora, Randy J. LaPolla. 2007. Minority languages of China: Factors affecting language maintenance. In Osahito Miyaoka & Michael E. Krauss (eds.), The Vanishing Languages of the Pacific, 337-354. Oxford: Oxford University Press. Ringe, Donald A. Jr. 1992. On Calculating the Factor of Chance in Language Comparison. Transactions of the American Philosophical Society, New Series, 82.1: 1-110. 49 Ringe, Donald A. Jr. 1995. Nostratic Factor Chance. Diachronica, 12.1: 55-74. Ringe, Donald A. Jr. 1996. The Mathematics of ‘Amerind’. Diachronica, 13.1: 135-154. Ringe, Donald A. Jr. 1999. Language classification: scientific and unscientific methods. In Bryan Sykes (ed.), The Human Inheritance: Genes, language and evolution, 45-74. Oxford: Oxford University Press. Sapir, Edward. 1921. Language: An introduction to the study of speech. New York: Harcourt, Brace and Company. Shibatani, Masayoshi. 1985. Passives and related constructions: a prototype analysis. Language, 61: 821-848. 56 (31) 因看诸人下象棋，曰:“凡事不得胡乱轻易了，又不得与低底下，后遇敌手便惯 了，即败。狮子捉象捉兔，皆用全力。”（《陆九渊集·语录·包扬显道所录》470 页） (32) 此人称是“东岳急脚子”，胡乱打人，不伏收领。（《续资治通鉴长编》8582 页） 四、“蛮”的族群义项 “蛮”原指南方的少数民族。《说文》“蛮，南蛮。它种，从虫，䜌声。”（《说文解字》， 《说文解字注》1170 页）“它种”，即“蛇种”，是解释从“虫”的由来，即以蛇为文化标 签的种族，可能与蛇图腾等文化习性有关，这与“闽”字从虫相似。古代对华夏族周边民 族，常用带动物类偏旁的汉字指称，其实华夏族群体本身也有以动物命名者，这些多与图 腾等文化特性有关。如“狄”与北方和西北民族的狼图腾有关，“羌”与牧羊为生有关。 黄帝本身叫有熊氏，还曾“教熊罴羆貔貅驱虎，以与炎帝战于阪泉之野”（《史记·五帝本 纪》3页），这些动物名可能都代表了一定的图腾。所以族名及其汉字的动物要素本身不 一定有褒贬色彩。段玉裁对《说文》“蛇种，从虫”解释道：“说从虫之所由。以其蛇种也。 蛇者，虫也。蛮与闽皆人也，而字从虫，故居部末。如貉之居豸末，狄之居犬末，羌之居 羊末焉。”（《说文解字注》1170 页）意思就是这些带动物类部首的字指的都是特定群体属 性的人，不是非人的动物，所以（作为该部首不典型的所指）排列在该部的末位。 “蛮”有时有比南方少数民族更广义的用法。李学勤主编的《字源》载“‘蛮’旧指 我国古代的民族，也泛称一切少数民族”（见李学勤主编 2012：1166，“蛮”条）。古人有 时也会直接将特定语境中的“蛮”训释为某一更具体的族群，包括并不在狭义的“蛮”范 围内的族群。《史记·吴太伯世家》“太伯之奔荆蛮”，司马贞注“蛮”为“闽也，南夷之 名”“亦称越”。更常见的用法则是介于各少数民族泛称和特定民族专称两者之间的概念 ——南方少数民族，也即《说文》所释的本义。如《论语·卫灵公》“虽蛮貊之邦行矣”， 朱熹注“蛮”为“南蛮也”。《逸周书·职方》“四夷八蛮”，孔晁注“蛮”为“南方曰蛮” （均见宗福邦等主编 2003《故训汇纂》2039 页“蛮”条）。“蛮”主要指南方少数民族， 这基本上已成共识，但对其语源则学者各执一词。如李永燧（1983）认为上古汉语“蛮” 字是古苗瑶语民族的自称音译，与今天很多瑶族支系的称呼 Mien（勉）和苗族的称呼 Hmong 等同源，郑张尚芳（1998）认同李永燧“蛮”“苗”“勉”语言上同源的观点，并 从音韵学角度做了补充论证；李连进（1995）则认为“蛮”是壮族先民的一个部落的族称 “板”[ʔban]的音译，而朱文旭（1999）认为古汉语中的“蛮”实则为藏缅语“人”的汉 语记音。从“蛮”的文献用例看，“蛮”与当时生活在华中两湖等地的苗瑶系民族关系最 密切，读音上也与 Mien 等最近。在文献中，南蛮常常与北狄西戎东夷及泛指北方西北方 少数民族的“胡”相对并称，有时与东南部“（百）越”及“闽”（闽越也是百越之一）即 壮侗族群的先民也作区分，但未见“蛮”和“苗、瑶”相区分的。而藏缅族系的主要古称 则为羌和氐，郑张尚芳（1998）认为藏缅古称是“戎”。“蛮”的词源义不太可能是夷、越 -壮侗、藏缅等族系的族称或自称。但在“蛮”泛指南方少数民族时，会兼指非苗瑶的各 57 南方少数民族，还可包括属于南亚语系的百濮族群。 总之，“蛮”早期应带有专有名词属性，最狭义可能指苗瑶系族群，常用义项指以苗 瑶为代表的南方少数民族。而《字源》所说的“泛称一切少数民族”则为更宽泛的后起之 义。南方地区本土作物及外来作物都较为丰富，后向北传播时或被冠以蛮名。如蛮姜，即 高良姜，《本草纲目》载其产于高良郡，并以此得名。高良郡，即今广东湛江、茂名一带。 再如蛮瓜，即丝瓜，原产于印度，后传入我国南方。《本草纲目》记“始自南方来，故曰 蛮瓜”（1315 页）。这些“蛮”都没有脱离“蛮”的南方地域指向性。 在先秦到六朝时期，“蛮”主要作为专有名词使用，指南方少数民族或与之相关的事 物。以下列举部分例子并加以简单分析。 (33) 故兵一出而大功十二。故东夷、西戎、南蛮、北狄、中国诸侯，莫不宾服。（《管 子·小匡》393 页） (34) 职方氏掌天下之图，以掌天下之地，辨其邦国、都鄙、四夷、八蛮、七闽、九 貉、五戎、六狄之人民，与其财用九谷、六畜之数要，周知其利害。（《周礼·职方氏》， 《周礼注疏》1861 页） (35) 象胥掌蛮、夷、闽、貉、戎、狄之国使，掌传王之言而谕说焉，以和亲之。（《周 礼·象胥》，《周礼注疏》1944 页） (36) 荆轲顾笑武阳，前为谢曰：“北蛮夷之鄙人，未尝见天子，故振慑，愿大王少假 借之，使得毕使于前。”（《战国策·燕策》1016 页） (37) 《任》，南蛮之乐也。纳夷蛮之乐于大庙，言广鲁于天下也。（《礼记·明堂位》， 《礼记正义》3226 页） (38) 不告朔于诸侯，玉瑞不行诸侯，力政不朝于天子，六蛮、四夷交伐于中国。（《大 戴礼记·用兵》，《大戴礼记解诂》211 页） (39) 昔轩辕不为涿鹿之师，则蚩尤之妖不灭。唐尧不兴丹水之陈，则南蛮之难不平。 （《伐吴诏》，《全上古三代秦汉三国六朝文·全三国文》1079 页） (40) 且当六国之时，天下殷炽，秦既兼之，不修圣道，乃构阿房之宫，筑长城之守， 矜夸中国，威服百蛮，天下震竦，道路以目。（《三国志·魏书·辛毗杨阜高堂隆传》716 页） (41) 雍州刺史王奂启录小府长史刘兴祖，虚称“兴祖扇动山蛮，规生逆谋，诳言诽 谤，言辞不逊”。（《南齐书·王奂传》849 页） 以上例子大致代表了先秦至六朝“蛮”的使用情况，大多以南方民族泛称出现，个别 的泛指少数民族，如例(36)“北蛮夷之鄙人”肯定不指“南蛮”。用法上，“蛮”既可以单 用也可以加方位词、处所名词或数词进行修饰。如加方位词的“南蛮”，加处所名词的“山 蛮”，进一步指明蛮族所在的位置。再如加数词的“百蛮”“八蛮”“六蛮”，做蛮族不同群 体的合称；数词有时并非表示确数，只是反映南蛮族的族系繁多。“蛮”常用来与其他各 地区族名国名并称，如“东夷、西戎、南蛮、北狄、中诸侯国”“蛮、夷、闽、貉、戎、 狄”等，凸显了“蛮”在当时的族名属性，“蛮夷”并称尤为多见，而“夷蛮”的说法亦 存。以上的“蛮”都是作为名词语的核心或定语标记“之”前的领属定语出现。 58 “蛮”也可以出现在不带“之”的复合词或类词紧密组合中做定语，表示带有蛮人蛮 族或蛮地特征的事物。如： (42) 凡封国若家，牛助为牵徬，其守王宫与其厉禁者，如蛮隶之事。（《周礼·司隶》， 《周礼注疏》1909 页） (43) 《莫愁乐》，亦云蛮乐，旧舞十六人，梁八人。（《古今乐录》，《乐府诗集》698 页） (44) 郝隆为桓公南蛮参军。三月三日会，作诗，不能者罚酒三升。隆初以不能受罚， 既饮，揽笔便作一句云：“𡜯隅跃清池。”桓问：“𡜯隅是何物？”答曰：“蛮名鱼为𡜯隅。” 桓公曰：“作诗何以作蛮语？”隆曰：“千里投公，始得蛮府参军，那得不作蛮语也！”（《世 说新语·排调》，《世说新语校笺》432 页） (45) 秋七月，武陵蛮寇临沅，遣谒者李嵩、中山太守马成讨蛮，不克，于是伏波将 军马援率四将军讨之。（《后汉书·光武帝纪》76 页） (46) 齐有宦者内参田鹏鸾，本蛮人也。（《颜氏家训·勉学》124 页） (47) 夜有蛮兵数千斫平前垒，矢及平帐，平坚卧不动，俄而乃定。（《魏书·李平传》 1452 页） 其中，“蛮隶”“蛮寇”“蛮人”“蛮兵”皆指蛮地蛮族的人。“蛮乐”即源于蛮地的乐 曲。 “蛮语”“蛮名”指的是蛮族蛮人所使用的语言及其词汇。至于“蛮府”，则为主管少 数民族的官署。以上各例皆与“蛮”本义有关，指向其明显的专有名词属性。 大致来看，“蛮”作为族名其本身的感情色彩是中性的。当然，封建时代，中央政权 和南方某些族群的矛盾冲突也不时发生，“蛮寇”之类称呼也难避免。而“蛮隶”，据《周 礼·秋官司寇·司隶》，“蛮隶，百有二十人”，郑玄注“征南夷所获者”（《周礼注疏》1875 页），可见亦是军事冲突下的产物。至于“蛮乐”，该例中实指《莫愁乐》，是汉魏六朝清 商曲的一个重要部分。“蛮乐”之称，似乎不减对其的赞美之意。前面例(37)还提到“纳 夷蛮之乐于大庙”，给予很高的地位。考察表明，这些正面或负面的组合并未对“蛮”的 非专名化起到决定性的作用。真正起决定性作用的是“蛮”与泛指处所名词的组合，以及 反映对“蛮”的刻板印象的一些常见话语和语境。 五、从“山蛮”、“野蛮”到“蛮横”、“蛮缠” 负面的语境确实影响着“蛮”后续发展的语义方向。历代文献中，泛称的“蛮”经常 被赋予较为负面的形象，如： (48) 若时得为人，蛮狄无义理，痴骏无孔窍，跛躄哑不语。（《阿难问事佛吉凶经》， 《阿难问事佛吉凶经讲记》5页） (49) 诸将欲召溆州酋长苻彦通为援，行逢曰：“蛮贪而无义，前年从马希萼入潭州， 焚掠无遗。吾兵以义举，往无不克，乌用此物，使暴殄百姓哉！”（《资治通鉴·后周纪》 9483 页） 59 (50) 晟之所为，虽蛮夷不足以论理，而人伦之内实所忍闻焉。（《五国故事》94 页） 以上部分反映了古人对于蛮族的刻板印象，将“蛮”与“无义理”“贪而无义”这样 的负面评价相联系。其中(48)(49)是直接给出的负面评价，(50)本身是谴责五代时南汉 残暴皇帝刘晟的，说刘晟的作为即使用蛮夷的道理也讲不通，是比蛮夷更不讲道理的人。 此句将蛮夷不讲道理当作常识放在了预设中。这些“蛮”虽然仍作为专有名词出现，但在 人们的刻板印象下，时常被置于负面评价的话语和语境中，就容易浮现出贬义感情色彩， 为“蛮”超越专名义，走向非专名化的形容词、副词奠定了语义基础。 唐代诗歌大兴，对仗尤盛。诗人往往会采用对偶等修辞手法以达到形式上的优美，这 为“蛮”的词性变化增加了动力。如： (51) 瘴地难为老，蛮陬不易驯。土民稀白首，洞主尽黄巾。（《送客春游岭南二十韵》， 《白居易集》353 页） (52) 瘴雨蛮烟朝暮景，平芜野草古今愁。酣歌欲尽登高兴，强把黄花插满头。（《九 日》，《全唐诗》5567-5568 页） (53)风雨瘴昏蛮日月，烟波魂断恶溪时。岭头无限相思泪，泣向寒梅近北枝。（《到恶 溪夜泊芦岛》，《全唐诗》5397 页） 例(51)中“蛮”与“瘴”并提。“瘴”修饰“地”，“瘴地”指空气湿热容易生病的地 方。与之对应，此处的“蛮”虽仍为专有名词解读，但已是属性形容作用大于指称作用。 类似地，例(52)中的“蛮”也是在与“瘴”并提中获得了属性作用。同样是“蛮”和“瘴” 并提，例(53)中“蛮”的专名意味进一步减弱，属性形容功能进一步增强。在这些例子中， “蛮”都带上了属性形容的作用，并且与“瘴”一样带有负面评价，描述令北人不习惯的 南方特有的凄风苦雨的环境。当然，这些“蛮”主要形容的还是地域自然环境，还不是形 容人的品行，但与有关蛮族的刻板话语相呼应，共同推动着贬义色彩的浮现。 “蛮”从专名化转向非专名化的一大关键，是“蛮”与表泛指的处所名词组合。由此， 人们直接以地域来指某一群体的特点，最后实现了“蛮”的非专名化。突出个例是“山 蛮”： (54) 授辅国将军、宋安安蛮二郡太守，迁南义阳太守。累破山蛮，郡境清谧。（《梁 书·马仙琕列传》280 页） (55) 累迁司卫监、洛州刺史。抑强扶弱，山蛮畏威，不敢为寇。（《北史·苟颓列传》 929 页） (56) 或庵既领住持，士庶翕然来归，衲子传至虎丘。瞎堂曰：“遮个山蛮杜拗子，放 拍盲禅，治你那一队野狐精。”或庵闻之，以偈答曰：“山蛮杜拗得能憎，领众匡徒似不曾。 越格倒拈苕帚柄，拍盲禅治野狐僧。”瞎堂笑而已。（《禅林宝训》，《禅林宝训讲释》376 页） “山蛮”字面上指居住在山上的蛮族，这个“山”描写性大于区别性，因为南蛮族群 的居地本以山区为主。“蛮所恃者山林耳，若战平地，安能败我”（《资治通鉴·唐纪》8526 页），表明山是大多数蛮族所仰仗的地方。这个“山”是泛指的，并非特定地名，这使“山 蛮”整体上在作定语时也是指别作用小于属性形容功能。当“山蛮”跟与人相关的名词组 合，“山蛮”的形容性意义就可能进一步突显。如“山蛮杜拗子”，“杜拗子”是用来称呼 60 言语不通、说话难懂之禅僧，引申指妄自尊大、不明事理者。“杜拗子”前加的“山蛮”， 已只有属性义，没有指称义，即像山蛮一样的、具有山蛮特性的（不讲道理的），而“杜 拗子”本身已含不明事理义，定语“山蛮”在此主要是起点强化作用。此例中“山蛮杜拗 子”形容的是唐代来自浙江台州的罗姓或庵禅师，与山蛮族群无关。可见，此例中“山蛮” 或“蛮”的专名属性已消失，意义已偏向属性形容和负面评价。 “山蛮”作为形容词的非专名化用法没有得以保存，但类似的有后起的“野蛮”一词。 “野”跟“山”可以构成近义并列复合词“山野”，泛指远离都市的荒野自然环境，“野” 后来还发展出“山”所没有的典型形容词义项（很野~*很山）。与“山蛮”结构平行、意 义类同的“野蛮”也从指人群发展出属性形容的功能，时间上比“山蛮”的形容功能出现 得晚些，但其非专名化的用法一直沿用至今。如： (57) 大赕，周回百余里，悉是野蛮，无君长，地有瘴毒。南赕人至，中瘴者十死八 九。阁罗凤尝遣使筑城于彼，管制野蛮。（《云南志》，《云南志补注》148 页） (58) 玳安叫掌起灯来，骂道：“贼野蛮流民，他倒问我是那里人！”（《金瓶梅》277 页） (59) (上江）好野蛮的百姓。（张白）你们回衙理事，听候旨意，再为定夺。（《因果 报总讲·四本》，《清车王府藏戏曲全编》711 页） 一开始“野蛮”作为专有名词，是指生活在郊野的蛮族。在人们对蛮族的刻板印象影 响下，“野蛮”也就逐渐被用来类指这种群体的特性，逐渐浮现出形容词功能，如(58)， “野蛮”用在骂人话中做表鄙视的定语，事件环境是山东，被骂者是“两个带白毡帽子的 酒太公”（造酒匠），与少数民族无关。(59)是“野蛮”作为成熟的形容词的用例，这类用 法沿用至今，而“野蛮”的内部结构也从定中结构被重新分析为并列结构了。 除了与处所语素结合，“蛮”本身因其长期形成的刻板印象，自身的词性与语义也在 朝属性形容的方向靠拢，如： (60) 蛮帅梅芳攻陷扬州，而广被杀。王时年十五，芳纳之。俄于暗室击芳，不中， 芳惊起曰:“何故反邪？”王骂曰：“蛮畜！我欲诛反贼，何谓反乎？”（《晋书·王广女 传》2520 页） (61) 陈敬瑄以扈从之人骄纵难制，有内园小儿先至成都，游于行宫，笑曰：“人言西 川是蛮，今日观之，亦不恶！”（《资治通鉴·唐纪》8245 页） (62) 有物死在阶下，身如狗，项有九头，皆如人面，面状不一，有怒者，喜者，妍 者，丑者，老者，少者，蛮者，夷者，皆大如拳，尾甚长，五色。（《太平广记·妖怪三》 2868 页） 当“蛮”常出现于负面话语、“山蛮”等又浮现出负面描述评价义项后，“蛮”在单用 时就容易出现两解的情况。例(60)中的“蛮畜”本是指蛮族的畜生，“蛮”带有专名特征。 但在此责骂梅芳的话语和语境中，“蛮畜”也可以理解为“不讲道理的畜生”。例 ( 61)中的 “是”作为判断动词和强调动词，其后既可以接名词，也可以接形容词。后接名词，即表 达“西川是蛮族”，后接形容词，即是“西川（的人）是不讲道理的”。后面说“亦不恶”， 隐含了对“蛮”——不讲道理的转折。“是”后接形容词的用法，同样见于宋代其他文献， 61 “此说话固是好，只是他地位未说得这般话。”（《朱子语类·先进篇·子贡问师与商也章》 1015 页）。况且所言的西川是宋代成都，并非是蛮族聚居之地，此例的“蛮”更像是形容 词用法。至于例(62)，此处的“蛮”也倾向作形容词解，因为与之平行的“怒”“喜”“妍” “丑”“老”“少”均为形容词或心理动词。 总的来说，“蛮”的非专名化发展可以归因于两方面，其一，人们长期对蛮族持有刻 板印象，“蛮”多见于负面话语中，因而“蛮”吸收了语境中的情感色彩。其二“蛮”常 与表泛指的名词“山”、“野”组合，本身的专名意味被削弱，在“山蛮”“野蛮”整体做 定语时浮现出形容词功能。 “蛮”从形容人的整体特性基础上，还进一步扩展到修饰人的动作、话语等，浮现出 状貌形容词和方式副词的用法，如： (63) 说长道短是和非，只与他蛮缠胡搅歪厮占。（《群音类选》，《群音类选校笺》1855 页） (64) 唐卿忙拦住道：“菶如，你且不用先怒，这不是蛮干得来的事。嫂嫂请我们来， 是要给她想个两全的办法，不是请我们来代行家长职权的。”（《孽海花》216 页） (65) 周老爷把嘴凑在他耳朵边，将此事始末缘由，他如何为难，统领如何蛮横，现 在想赖这笔银子的话，说了一遍。（《官场现形记》177 页） (66) 又见王先生拿袖子卷了两卷，把条腿夹在夹肢窝里，想用蛮劲再把这条腿扳过 来。（《官场现形记》433 页） (67) 我道：“这个奇了，那里闹出这么一句蛮话来？”（《二十年目睹之怪现状》159 页） 以上列举的“蛮 N”“蛮 V”等组合，进一步巩固了“蛮”作为贬义形容词、副词的 非专名化用法。此外，还有并列复合词“X 蛮”的组合，如“刁蛮”、“霸蛮”等，也进一 步强化深化了“蛮”的贬义色彩，助力“蛮”完成非专名化。其中最有意义的是已成并列 式俗语的“胡搅蛮缠”（亦作“蛮缠胡搅”，如例(63)），因为两个贬义方式副词“胡”和 “蛮”都来自地域少数民族的泛称，分别指向北方和南方，这个常用俗语集中展示了专有 名词非专名化的历程和结果。但现代汉语使用者，早已感受不到其中的“胡”和“蛮”与 其专名义的关系。在民族平等成为共识、地域偏见不得人心的现代社会背景下，对特定民 族的负面刻板形象也难以在公众话语中存在。“胡” “蛮”一类专名的非专名化历程，是特 定历史时期人们的意识偏见的产物。今天之所以“胡搅蛮缠”等还会常用，就因为“胡” “蛮”这类贬义词已经与其专名义项发生了库藏裂变，语言心理上已不再视为同一个词 的不同义项，而接近语义上无关的同音词了，虽然词典释义还没有完全处理为不同的词 条，如《现代汉语词典》中“胡”“蛮”的专名义和非专名义是在同一条目下的不同义项。 这也有一定的合理性，因为它们还存在着语源上的历史联系，如本文所展示的。但是，从 纯共时角度看，库藏裂变后的不同义项是可以处理为不同条头的同音词的。 62 六、小结 本文通过对“胡”和“蛮”这两个古代区域少数民族泛称的考察，展示了汉语中专有 名词非专名化的一种语义演变历程，它们由北方或南方的少数民族泛称，在历史上的民 族刻板印象、带有评价和感情色彩的话语和语境、特定的搭配关系和句法位置（尤其是定 语和状语）等因素影响下，实现了非专名化，成为表示某种属性的形容词并产生方式副词 的用法。语言学界熟知方言地理学先行者席叶龙(Julies Gilliéron)的一句名言“每个词都 有它自己的历史”。如果说这句话对所有词都适用，那么它对专有名词的非专名化尤其适 用，因为专有名词是个性最强的词语，它们的语义演变史，更多展示词语的个性而非共 性。即使是“胡”和“蛮”这两个南北高度对称的民族泛称专名，非专名化过程也呈现出 各自的个性，例如“胡”的非专名化最早发生在与言语词语的组合，“蛮”就无此过程。 但是对于语言学来说，从个性中发掘共性或一般规律是它的天职。我们从专名非专名化 的考察中，仍能看到一些共同的机制，如对特定群体的刻板印象和话语感情色彩的语境 吸收导致的色彩赋义，特定句法位置催生新意义的作用等。我们在今后对专名非专名化 的考察中，将继续关注这种个性中的共性。 参考文献 汉语大词典编纂处（华建敏主编）2023《汉语大词典（第二版）》上海辞书出版社 李连进. 1995.“蛮”为古代壮族族称再考.广西师院学报 1:101-107. 李学勤 2012《字源》天津古籍出版社 李永燧. 1983. 关于苗瑶族的自称——兼说“蛮”. 民族语文 6:16-22. 刘丹青. 2019. 汉语中的库藏裂变. 语言教学与研究 5:34-50. 刘丹青. 2021. 语言单位的义项非独立观. 世界汉语教学 2:147-165. 叶晓锋. 2024. 胡人考——从古文字谐声材料看胡人的语源.“古文字与上古音整合研 究”国际学术研讨会论文.西湖大学. 郑张尚芳. 1998.“蛮、夷、戎、狄”语源考.扬州大学中国文化研究所集刊第 1 辑，江苏 古籍出版社。收入《郑张尚芳语言学论文集》，中华书局 2012. 朱文旭. 1999.“蛮”语义及其文化现象.西南民族学院学报(哲学社会科版)4:38-41. 宗福邦,陈世铙,萧海波 2003《故训汇纂》商务印书馆 语料引用历代文献 周-秦 (东汉)郑玄注,(唐)贾公彦疏《周礼注疏》阮元校刻 2009《十三经注疏》本中华书局 李山,轩新丽译注 2019《管子》中华书局 许维遹集释,梁运华整理 2009《吕氏春秋集释》中华书局 63 两汉 (东汉)安世高译,肖卫主编 2013《阿难问事佛吉凶经讲记》团结出版社 (东汉)班固著,(唐)颜师古注,中华书局编辑部点校 1962《汉书》中华书局 (西汉)桓宽撰,王利器校注 1992《盐铁论校注》中华书局 (西汉)刘向编,缪文远,缪伟,罗永莲译注 2012《战国策》中华书局 (西汉)司马迁撰,(南朝宋)裴骃集解,(唐)司马贞索隐,(唐)张守节正义中华书局编辑部 点校 1982《史记》中华书局 (清)王聘珍撰,王文锦点校 1983《大戴礼记解诂》中华书局 (东汉)许慎撰,(清)段玉裁注,许惟贤整理 2007《说文解字注》凤凰出版社 (东汉)郑玄注,(唐)孔颖达正义《礼记正义》阮元校刻 2009《十三经注疏》本中华书局 魏晋南北朝 (西晋)陈寿撰,(南朝宋)裴松之注,陈乃乾点校 1982《三国志》中华书局 (西晋)崔豹撰,牟华林校笺 2015《〈古今注〉校笺》线装书局 (东晋)常璩撰,刘琳校注 1984《华阳国志校注》巴蜀书社 (东晋)干宝著,马银琴译注 2012《搜神记》中华书局 (东晋)郭璞注,王贻梁,陈建敏校释 2019《穆天子传汇校集释》中华书局 (南朝宋)范晔撰,(唐)李贤等注,中华书局编辑部点校 1965《后汉书》中华书局 (南朝宋)刘义庆著,徐震堮校笺 1984《世说新语校笺》中华书局 (南朝梁)陶弘景撰,赵益点校 2011《真诰》中华书局 (南朝梁)萧子显撰,中华书局编辑部点校 1972《南齐书》中华书局 (南朝梁)释僧祐撰,苏晋仁,萧鍊子点校 1995《出三藏记集》中华书局 (北齐)魏收撰,中华书局编辑部点校 1974《魏书》中华书局 (北齐)颜之推,檀作文译注 2011《颜氏家训》中华书局 唐-宋 (唐)白居易撰,顾学颉校点 1979《白居易集》中华书局 (唐)樊绰撰,向达原校,木芹补注 1995《云南志补注》云南人民出版社 (唐)房玄龄等撰,中华书局编辑部点校 1974《晋书》中华书局 (唐)李德裕撰，吴企明点校 2012《次柳氏旧闻》中华书局 (唐)李石等编著,裴耀卿语释 2004《司牧安骥集语释》中国农业出版社 (唐)李延寿撰,中华书局编辑部点校 1974《北史》中华书局 (唐)释道世著,周叔迦,苏晋仁校注 2003《法苑珠林校注》中华书局 (唐)魏征等撰,中华书局编辑部点校 1973《隋书》中华书局 (唐)杨筠松《天玉经》陈政儒编著 2003《堪舆正经》西苑出版社 (唐)姚思廉撰,中华书局编辑部点校 1973《梁书》中华书局 (清)彭定求等编 1960《全唐诗》中华书局 64 (宋)陈彭年,丘雍重修,蔡梦麒校释 2021《广韵校释》中华书局 (宋)郭茂倩编 1979《乐府诗集》中华书局 (宋)黄震撰,曾枣庄,刘琳主编 2006《全宋文》上海辞书出版社安徽教育出版社 (宋)净善辑,弘学讲释 2006《〈禅林宝训〉讲释》四川出版集团 (宋)李昉等编 1961《太平广记》中华书局 (宋)黎靖德编,王星贤点校 1986《朱子语类》中华书局 (宋)李焘撰,上海师范大学古籍整理研究所,华东师范大学古籍整理研究所点校 2004《续 资治通鉴长编》中华书局 (宋)林希逸著,周啓成校注 1997《庄子鬳斋口义校注》中华书局 (宋)陆九渊著,钟哲点校 1980《陆九渊集》中华书局 (宋)欧阳修、宋祁等撰,中华书局编辑部点校 1975《新唐书》中华书局 (宋)司马光编著,(元)胡三省音注,标点资治通鉴小组校点 1956《资治通鉴》中华书局 (宋)佚名撰,张剑光,孙励整理 2019《五国故事》大象出版社 元明清-现当代 李修生主编 1998《全元文》江苏古籍出版社 隋树森编 1964《全元散曲》中华书局 (明)胡文焕编,李志远校笺 2018《群音类选校笺》中华书局 (明)兰陵笑笑生著 2005《金瓶梅》内蒙古人民出版社 (明)李时珍著 1999《本草纲目》中国档案出版社 (清)李伯元著,高书平注 2018《官场现形记》崇文书局 (清)吴趼人著 2018《二十年目睹之怪现状》华文出版社 (清)严可均编 1958《全上古三代秦汉三国六朝文》中华书局 (清)曾樸著,张国权注 2018《孽海花》崇文书局 黄仕忠主编 2013《清车王府藏戏曲全编》广东人民出版社 65 5 Postverbal modification, information focalization, and meaning creation: With a specific case study of Chinese 在zài-phrases in translation ZHU Chunshen The Chinese University of Hong Kong (Shenzhen) https://orcid.org/0000-0002-7225-5446 Abstract: In Chinese, zài-phrases like 在教室裡溫習 (‘be-in the classroom review [the lesson]’) are grammatical, but *溫習在教室裡 (‘review [the lesson] be-in the classroom’) is not, due to the ungrammatical use of the postverbal modifier. However, the phrase 沐浴在春風裡 (‘bathe be-in the spring breeze’), which also includes a postverbal modifier, is felicitous with the meaning of ‘basking in the spring breeze’. Interestingly, 在春風裡沐浴 (‘be-in the spring breeze bathe’), which mirrors the structure of ‘be-in the classroom review [the lesson]’, should be grammatically correct but is infelicitous, as it conveys a meaning that defies common sense. This paper explores this discrepancy by linking it to the “discursive propensity” of postverbal modification and its role in information focalization that shapes the experience of meaning creation. The study takes the issue into the domain of translation studies for a sharper and comparative perspective, construing translation as a process of meaning (re-)creation. By examining pairs of texts that aim to convey the same meaning but produce different effects, the paper characterizes meaning creation as a two-stage discursive experience: an initial private phase of meaning-making, followed by a communal phase of sense-making. It argues that, despite variations in discursive experiences among individual readers of the same text, which are influenced by different social contexts, such experiences are textually accountable in terms of information focalization. As such, the felicity of an utterance, beyond grammatical considerations, is grounded in an experiential underpinning sustained by the discursive propensity originated by focalized information. Zhu, Chunshen. 2025. Postverbal modification, information focalization, and meaning creation: With a specific case study of Chinese 在zài-phrases in translation. In Chenglong Huang & Xu Yang (eds.) The Creation of Meaning: A Festschrift in Honour of Randy J. LaPolla, pp. 65–90.  72 4.1 Informational focus can be identified in every sentence. Our review in Section 1 affirms that “topic and focus, along with related phenomena, are universally represented” (Schwabe and Winkler 2007: 4) in many languages, including Chinese. Specifically, every sentence must have a focus, which is closely tied to the presence of new information and serves as a prerequisite for its communicative validity (Xu 2005: 17). 4.2 Focus is marked by morphosyntactic means, although the marking may differ from language to language. The word order of a text facilitates its background-foreground arrangement of information, determining when and how to convey its specific elements in focus. In both Chinese and English, focus is managed through means such as positioning, modification (see Zhu 2022: e.g., Section 5.4 for a discussion on the implications of modification for focalization and attention management), intonation, and language-specific devices. For example, in Chinese, the shì...de (是…的) and de shì (的是) constructions (Zhu 1996: Section 4; see also Xu 2001: 117–118), along with adverbs like jiù (就), yě (也), cái (才), and dōu (都), as well as progressive parallelisms (Zhang 2006: 200), play an important role. 4.3 The semantic sense of a sentence is not affected by the placement of information focus. This is noted in Xu (2005: 23), that is, which constituent of the sentence the focus is placed on does not affect its semantic sense or truth value. So, theoretically, focus shift in translation will not affect the semantic sense the translation shares with its source text. What it affects is the reader’s discursive experience of meaning creation. 4.4 What is put in focus is new, but not necessarily brand new, information. Focus is highly related to “new” information, but it does not mean the information has to be “brand new” (Xu 2005: 14, with a mention of, e.g., Schwarzschild’s proposal not to associate focus with new information). A succinct overview of the issue can be found in Butler (2003: Section 2.5.1.2.1 What is focus?), from which we may conclude that an entity already mentioned in the discourse, hence given and old, can also be re-introduced and placed in focus if it is pragmatically presupposed to be non-recoverable from the immediate context or situation, or if it is deemed so important for the discursive progression that its reiteration and focalization become desirable. 4.5 The sentence/clause-initial and sentence/clause-final positions are normally the most significant positions of informational prominence. This proposition has backing from psychology and psycholinguistics, which have found that, among other things, the subjects remembered the first and the last words much better than 73 the centre-positioned ones in the clause (Downing 1995: 16, mentioning Jarvella’s 1979 experiments; see also Butler 2003: 171 and Section 2.5.2.1 Initial position), irrespective of the designation of the information component in question as topic or focus. In light of the fact that syntactic development from sentence-initial to sentence-final entities represents a cumulative progression of information—commonly observed in the unmarked right-branching structure in English, as opposed to the marked front-heavy left-branching and compressive mid-branching structures (Tufte 1971: e.g., 156)—Butler (2003: Section 2.5.2.2.2, What meanings are associated with placement in ‘final’ positions?, referencing P. H. Fries) notes that “clause-final position in the written language is the most usual site for the placement of newsworthy information”. In Chinese, according to Zhang, sentence-final position is for the “conventional” and “natural” context/situation-independent “static” focus. The position of the marked, “dynamic” type of focus, on the other hand, is dependent on the communicative situation to register a contrast, and can thus be more mobile: at sentence-final, it overlaps the static end-focus; at sentence-initial, it becomes the “topical focus”. And what is put in the topical (focus) position has much to do with viewpoint setting and adjustment. (See Zhang 2006: esp. 181 fig. 1, 182– 183, 194, and 198.) This makes the distribution of “old” and “new” information more accountable in that the viewpoint determines what should be given as old information and what should be foregrounded and focalized as new information, i.e., as the key element that justifies the saying and that propels the progression of the discourse. Similarly, focalization has served as a point of reference in the discussion of EnglishChinese translation, although different terminologies may have been employed, e.g., sentenceinitial functionally extraposed focus versus unmarked, sentence-final focus and the stylistically highly marked sentence-medial one (Liu 1992: 167; see also Shan 1990: 353 and Li 1998: 254 for similar attention to focality). Therefore, while being aware of the function of the entity in the sentence-initial position (e.g., viewpoint-setting), it seems advisable to apply the principle of end-focus, in conjunction with the old-before-new ordering of information in both English and Chinese, as the unmarked benchmark for information (re-)structuring in translation, to facilitate, among others, the observation of any shift of a non-focal entity in the source text via certain target language-specific markings to a focal position in the target text, or vice versa, to comply with its grammar or to manipulate the reader’s response for ideological purposes, for instance. 4.6 The significance of focus (as opposed to topic) should be appreciated not only within the clause/sentence but also in terms of discourse development. Barthes, inspired and informed by generative and functional grammars in his 1960s studies on literature and discourse (Barthes 1987/2004, 1977: 79–124), has identified a comparable, 74 “homological relation between sentence and discourse”.8 This is one of the basic assumptions in his discussion of “the variations of the meanings engendered and … engenderable by works” and the reasons for their acceptability (1987/2004: 29). It also underpins his conception of an “obviously ... linguistic” model of “science of literature (or of writing)” (p. 28, italics original), in which literature is construed as discourse. As “a science of discourse”, the model is designed to integrate the examination of “those signs which are shorter than the sentence” (which underpin “all of the characteristics of literary language as a whole”) with the examination of those “signs which are longer than the sentence, … from which one can infer a structure of the narrative, of the poetic message, of the discursive text, etc.” (p. 31), to “allow [the critic/reader] to explain how chains of symbols are established along certain lines” and see “an ever-increasing unity” (p. 34) in a work of literature. This enables him to stratify on an operational level of structural analysis the discourse (or narration) in layers and project it into the world or situation from which only can it “receive its meaning” (Barthes 1977: 115). This homological relation and the notion of the world as the source of meaning can be further substantiated by later developments in linguistic research, especially with regard to intersentential thematization where the rheme of the previous sentence contains an element (usually in its focus) that becomes (part of) the theme of the subsequent sentence. Insofar as the beginning of a discourse implies a there + be declaration (“Once upon a time, there was…”), it features a “sentence-focus” in Lambrecht’s terms in that the entire sentence is to introduce new information to start the story, or, if it occurs in the middle of the discourse, to “introduce a new discourse referent or … to announce an event involving a new discourse referent” (Lambrecht 1994: 14, also 222). Such there + be type of sentence must be a focus in its entirety at the text level, because it cannot find an entity as its topic to justify its focality. In this way it plays “a hybrid function … of topicality and focality” for “introducing future Topic” to advance the discourse (GómezGonzález 2001: 151, 153). It follows that the significance of a focalized entity—whether in a “static” or “dynamic” focal position (Zhang, previously)—does not reside in its status as either topic or focus within the clause or sentence. Instead, its importance lies in its contribution to inter-sentential thematization, where the sentence receives its full textual meaning from the world of discourse as its focal information links it to subsequent sentence(s). This power of connection we call the discursive propensity of the information focus, which should be an active quality of 8 “…the most reasonable thing is to posit a homological relation between sentence and discourse insofar as it is likely that a similar formal organization orders all semiotic systems, whatever their substances and dimensions. A discourse is a long ‘sentence’ (the units of which are not necessarily sentences), just as a sentence, allowing for certain specifications, is a short ‘discourse’.” Also, “Structurally, narrative shares the characteristics of the sentence without ever being reducible to the simple sum of its sentences: a narrative is a long sentence, just as every constative sentence is in a way the rough outline of a short narrative.” (Barthes 1977: 83, 84) 75 Cameron’s “force of description” (previously) to propel the sentence into a broader context with a prospective topic. On the discourse level, this explains the significance of the “end focus” of a discourse, or the ending of its story, for relating the discourse to a historical or social world to receive its contextual meaning. A telling example can be found in the essay “Of Cunning”,9 where Francis Bacon remarks how a “cunning” person would put the most material matter in the postscript of a letter or postpone the truly intended point until the end of a speech, as if it were an afterthought. Cunningly or artfully delayed as a punchline so to speak, the focal sentence acquires a propensity to catapult the person’s discourse into some worldly business of his genuine concern beyond the text. More specifically, given the grammatical fuzziness surrounding the concept of sentence,10 observing the discursive propensity of a syntactically focalized entity will particularly help trace out the contribution a constituent of the discourse makes to orchestrating an experience of meaning creation, in which the text offers itself for meaning-making and the world beyond conditions the sensibility of the meaning made. The assertion that “each sentence [in a text] can achieve its end only by aiming at something beyond itself” (Iser 2006: 17, overviewing Ingarden) applies not only to literary texts but to all types of texts, as long as their sentences are correlated to refer to a real or fictional world. By relating meaning to the world, meaning creation acquires a hermeneutic dimension worth exploring in the subsequent section. 4.7 There is a need to extend the perspective on focus from the speaker’s intention to include both discourse and the hearer’s discursive experience as shaped by the text. Since a discourse receives its meaning from the world (Barthes, previously), it should receive its meaning from the speaker’s world to justify its being said in such a way in the first place. But it should be in the hearer’s world that its meaning is made and its sensibility judged. That is, it is the hearer who creates the meaning by going through the discourse as it textually proceeds and by relating it to their world through interpretation. As such, the hearer’s interpretation entails a world substitution. According to Eco (1992: 26–27), there are “two basic interpretative attitudes”, “that is, two ways of deciphering either a text as a world or the world as a text”. The two are correlative in that reading a text as a world is a projection of one’s life experience (with its intrinsic biases and blind spots) of being in this world that one has read as a text onto that world which one is seeing emerging from the text that is being read as a world. And the experience of projecting oneself as being in that world—by participatorily reading the text as a world—is supposed to make one more perceptive in reading this world as a text. When a text is read as a world, by being a “world”, its topography of “words in this order” (Cameron, 9 “Of cunning” in Francis Bacon Essays. Ware: Wordsworth Editions Limited, 1997; p. 62. 10 As Favretti et al. (1999: 23) have noted, “Standard sentence grammar is, in general, a grammar of the clause, and would thus be inadequate to investigate the structure of the larger units of discourse” (italics added). 76 previously), or “chains of symbols … established along [this] line” (Barthes, previously), is bound to throw up certain of the features of what it says into relief via when and how to say it. Such when and how, as Wittgenstein sees it, is the most fundamental character of the social use of language as manifested through (literary) texts; and it is with this concern with the when and how that grammar becomes meaningful in “show[ing] in which contexts words can be used and how they can be combined with other words” (see Huemer 2004: 5). In this connection, we see a close interrelation between linguistics and hermeneutics in meaning creation. Indeed, in the study of language and communication, a hypothesized author/speaker intention is commonly referred to as a deterministic factor for focus identification, as seen in such remarks as the focus carries “the information in the sentence that is assumed by the speaker not to be shared by him and the hearer” (Jackendoff cited in Lambrecht 1994: 207, italics added). Similarly, Xu (2001: 118, also 141) claims that determination of a focus is not entirely dependent on specific grammatical marking devices but on whether the information is intended by the speaker to be the key message. Other assertions of speaker-centrality can be found, e.g., in Sidiropoulou (2004: 94), Gómez-González (2001: 256), and Butler (2003: 157, 160). It goes without saying that it is the speaker who initiates communication with an intention. As Halliday (1994: 299) puts it, both “listener-oriented” Given + New and “speaker-oriented” Theme + Rheme segmentations are, “of course, speaker-selected”. However, with the physical absence of the author in written communication, interpretation of an authorial intention is largely based on the reader’s empathy with a notional author whom they have perceived via the text (alongside paratexts perhaps) in such and such a situation. Yet, since the discursive experience the same text may engender varies from reader to reader, even from reading to reading by the same reader, no meaning made by an individual reader on a specific occasion is the final and absolute interpretation of the author’s intention or, for that matter, an exhaustive explication of the text’s meaning potential. Thus, the inaccessibility of a definitive authorial intention is compounded by a hermeneutic tension between the necessity and the unreliability of reader intuition due to the openness of an individual’s discursive experience. This openness is illustrated, on the one hand, by the difficulty in characterizing the reader (see MacKenzie 2002: 38 for various descriptions); and on the other hand, by the position of the author “between the intention of a given human being and the linguistic intention displayed by a textual strategy” (Eco 1992: 69, referencing Mauro Ferraresi). 11 Moreover, there is uncertainty in the relation between the author’s 11 In text linguistics, intentionality is one of the seven properties of textuality ascribed to a human-produced text, but it cannot be confused with a specific authorial intention that is impossible to verify in the practice of text interpretation (see MacKenzie 2002: 202, note 6 on Hirsch), despite the fact that even in the extreme case of poetic discourse, “critics are generally unwilling to accept that a poet did not consciously or unconsciously will 77 biographical context and the text, besides the “impossibility of emptying our minds of our own prejudices and prejudgements in order to duplicate an author’s historical and linguistic context” and the impossibility of “guarantee[ing] that our understanding of an author’s words will allow us to re-experience his or her mental and emotional state” (MacKenzie 2002: 40–41, 45). This means that world substitution is not a simple replacement of the author’s world with the reader’s own world. Instead, it leads to the emergence of a “new” world from the text, which is activated and charted by “these words in this order” (Cameron, previously). Meaning is made as this new world takes shape. In short, the world a reader interprets from a text is not solely the author’s creation. It is a text-induced and textually accountable construct that the reader finds meaningful and sensible enough to “live in” and to share with others, for example, through translation. It is therefore of interest to observe meaning creation from a dual perspective of “positive” and “negative” hermeneutics reviewed in MacKenzie (2002: Chapters 3 and 4). Positive hermeneutics upholds a definite, unchangeable “textual meaning”, claiming that “[t]o relate textual meaning to any context beyond itself is not to create meaning but only significance”, and that “[i]f a reader responds differently to a text at two different times, it is the response that changes, not the meaning”; that is, “positively”, a text establishes its meaning, as “willed or intended by the author” (see MacKenzie 2002: 34, with reference to E. D. Hirsch, italics original), independently of the significance or meaning that a reader may create by referring it to a real or fictional world. Thus, when a text’s significance changes with the reader’s response, it changes because the text is being responded to as signifying something in the real or a fictional world now perceived somewhat differently by the reader beyond the author’s intention. Admittedly, in immediate oral communications—where interlocutors are present and available for mutual verification if necessary—or in professional communications, where interlocutors typically share a common stock of working knowledge about the subject matter, the world perceived by the hearer may closely align with the world referred to by the utterance, making consensus on meaning possible. This convenience, however, cannot be assumed in written communication, or in artistic communication where the text refers to an imaginary world. Furthermore, since no definite, unchanging “meaning” can be identified as referential counterpart to “significance” to trace out the latter’s changes, Hirsch’s “significance” is actually the meaning made by the reader via world substitution and deemed sensible in their current context. With changes in its significance, a text’s history of reading is a history of the expansion of its potential for meaning creation, with all the meanings created converging to contribute to “an ever-increasing unity” of its “poetic message” embedded in its “structure of the narrative” in Barthes’s (previously) terms. Hermeneutics thus turns “negative” and maintains that the text is essentially “a resource or intend whatever it is that their analysis reveals” (see MacKenzie 2002: 58, reporting this remark from Jean Starobinski in the conclusion of his 1971 book on Saussure’s unpublished manuscripts). 78 rather than a source: a heuristic stimulus for the production of new meanings involving elements absent from or even in contradiction with the textual surface” (MacKenzie 2002: 46, italics added). As a “resource” prior to reading, the text does not have an “old”, unchangeable meaning that could serve as a yardstick to compare any newly created meaning. The reader is therefore given a more active and decisive role to play than in Schleiermacher’s and Venuti’s binary models reviewed previously. This does not mean that the reader is a free meaning creator. As reading dictates, there is always the need to engage with the text’s “surface” to discern its focal message. This involves surveying its topography to verify the nuance and sensibility of the meaning in relation to one’s own context at that moment. The meaning thus verified and created does not appear “new” vis-à-vis a fixed authorial meaning. Rather, it is new in comparison to the meanings created by the same reader in their previous readings or by their predecessors. All in all, regardless of how the paradigm of reading evolves, the text remains a constant resource, its topographic “surface” unchanged in its these-words-in-this-order arrangement—a textual mise-en-scène where some information elements provide a topical viewpoint while others take on a focal position. The nuances of such focalization fine-tune—and thereby account for— the readers’ contextual creation of “[new] meanings”. In translation as meaning (re-)creation, every choice of word and word order permitted by the target language—once materialized in the target text, whether made by an AI program or a human agent—is textually accountable not only for its role in meaning-making but also for its contribution to sense-making within a social or communal context. 4.8 Identification of the focus of a clause/sentence is not always a straightforward exercise. Given such hermeneutic uncertainties, critics tend to resort to two means to help ascertain focus identification: intonation and native speaker intuition. But these two approaches may not prove to be sure solutions. Regarding intonation, such observations as the following are pertinent: “it is argued that, when the Focus is marked morpho-syntactically, this category needs (sic) not be marked prosodically, and if it is so marked, it gets a weaker prominence (Gómez-González 2001: 143, referencing Dik), and “what is really at issue is the information structure of the clause, the tone group not being an informationally independent unit” (Butler 2003: 162, referencing Fronek). Since our study is on translation of written texts, we shall take morpho-syntactic devices as the primary marking of focus. As to native speaker intuition, or an intuitive feel for language (語感), it is a linguistic kind of common sense nurtured by the speaker’s or hearer’s exposure to or experience of a language in use in their communitiy. About the openness of this personal “taste and style”, Downing’s (1995: 22) remark is very much to the point: it is “very difficult to quantify, may have a significant impact on word order, making absolute predictability an unachievable goal”. And in Chinese, the same utterance can be parsed differently according to the intuition of different 79 individuals (Pan 1997: 191). Yet, from time to time, critics may still have to resort to intuition in explaining away a delicate case, and controversy can thus ensue.12 In translation research and criticism, the possibility of impressionistic or even selfconflicting interpretations arising from this intuition-based approach is amplified. For example, Liu gave the following as a counterexample to what he described as a general tendency of endweight: (1) 在胡適先生說來，高於一切、指引一切的，是他的母親的精神薰陶 (Liu 1992: 169 (a)) ‘as to Mr. Hu Shi, above all, guiding all, de shì [‘is’] his mother’s spiritual influence’. According to Liu, the “weight” falls on the sentence-medial ‘above all, guiding all’, and an English translation was offered accordingly as follows: (1.1) In Mr Hu Shi’s case, a mother’s hallowed influence has been the guiding star of his life. A linguistically informed awareness of end-focus, however, would lead to a different observation, namely that ‘above all, guiding all’, like the intensifier above all in English, accentuates the entity that follows it, bringing it into focus, rather than directing focus onto itself. This interpretation confirms rather than counters the so-called “general tendency of end-weight” in Chinese and is morpho-syntactically marked by the focus-indicating …de, shì… (“…的， 是…”) pattern. By keeping ‘his mother’s spiritual influence’ in end-focus, the same information but structured differently in the following translation may give rise to a different discursive experience: (1.2) As for Hu Shi, the most important moral principles guiding his life can be traced to his mother’s spiritual influence. Observations in this section make Hypothesis 4.5 all the more meaningful for both intra-lingual interpretation and cross-lingual translation. 12 An example seems in order here. In Chinese, there is a special bǎ (把) construct, which enables the fronting of the object to a preverbal position, i.e., turning the unmarked S+V+O into a marked S+bǎO+V. Xu, in contesting Tsao’s conception of the fronted object as the Secondary Topic, bases his observation on his intuition (“語感”) and claims that the fronted object should be the focus (Xu 2001: 140). However, if we follow Xu’s observation made later in the same section (p. 142), that the sentence-final constituent tends to become the focus, we may come up with a different description, i.e., STop+bǎOSecTop+VFocus, and detect a more general tendency of postponing the verb to the end-focal position, as also seen in the passive bèi (被) construction. 80 4.9 Information focalization can be managed to account for different discursive experiences by which a meaning is created. An unmarked sentence proceeds from initial topic to final focus. Given the compatibility in topic-focus sequence between English and Chinese, as our review has so far indicated, our study focuses less on the differences in their syntactic means of information distribution and more on the distribution facilitated by these means in the source text, as well as the means available in the target language to ensure a similar distribution, if desired. The translator may of course choose to organize the source-text information in a different pattern for a specific purpose, with or without explicitly declaring such an intention. In either case, the source text can serve as a benchmark for examining the information distribution in the target text. This examination can focus on its effectiveness in achieving the translator’s declared purpose, or, if the purpose is not made public, on its potential effect on the reader’s discursive experience. Let us take a few examples from Fang (1999) for closer examination. The following two sentences were presented in Fang (p. 207) to substantiate his observation about sentence focus. Specifically, he notes that the weight of a sentence is often postponed in Chinese texts but can be fronted in English translation: (2) 我原來打算今年 1月訪問中國，後來不得不推遲了 (2.1) I had to postpone the visit which I had intended to pay to China in January; (3) 他5歲生了一場病，變成了聾子 (3.1) He became deaf after an attack of illness at 5. (Fang’s examples, emphases original) The focal information in the Chinese texts is fronted to become the topical information. Both sentence-initial and final positions are important in information distribution (Hypothesis 4.5). However, the key issue lies in the discursive experience that the change of the focal constituent may engender. This is particularly relevant in such cases where the change is optional rather than grammatically obligatory: (2.2) 我不得不把訪問推遲了，原來是打算今年 1月來中國的 ‘I had to have the visit postponed, originally I planned to come to China in January’ (3.2) 他聾了，因為 5歲時生了一場病 ‘he became deaf, because of an attack of illness at five’. In light of Hypothesis 4.6, we see that the end-positioned postponement of the visit in (2) and of the protagonist’s becoming deaf in (3) are to foster a propensity to arouse the reader’s interest in 81 the cause of the schedule change and the consequences of the hearing impairment respectively. In (2.2) and (3.2), the distribution is reversed. By delaying the mention of the original schedule to the end focus, the flashback in (2.2) encourages an expectation for additional information about the visit originally planned for January, which could develop into a future topic. Similarly, the flashback in (3.2) tends to prioritize explaining the cause of the illness over describing the consequences of the deafness as a prospective topic of the discourse. This discursive propensity is what we have found in (2.1) and (3.1), the English translations provided by Fang, though realized through different syntactic means. It is true that different languages may display different preferences for placing certain information constituents in focal or topical positions. And it is tempting to generalize such seemingly prevalent preferences as an inherent cultural characteristic of the language community. For instance, Sidiropoulou (2004: 98) claims that “[t]he fact that the preference for deictic specificity with respect to time and space [to appear at the beginning of a sentence in modern Greek] overruns genres … and literary translation … [which] makes it a better candidate for being part of Greek cultural heritage” (italics added). Incidentally, Chinese grammar shares the tendency to put time and space constituents at the sentence-initial or preverbal position to set the circumstantial scene for the event. Against the tendency, postponements may appear marked or have to be marked by a comma, as illustrated in the following example: 他們都跑過來了，從 教室裡，從操場上，從宿舍裡 (‘they all run over, from the classrooms, from the playground, from the dormitories’, Zhang 2010: 381; see also Lu Xun in Section 1). But there is a risk of over-generalization in relating such grammatical preference to cultural predilection. For example, Fang (1999: 208) in a similar spirit claimed that the difference in word order between (4) and (4.1) and between (5) and (5.1) reflects cultural discrepancies: (4) 接到你們的賀函，我十分愉快 ‘receiving your message of congratulations, I am very happy’ (4.1) I am very happy to receive your message of greetings (5) 現在有機會互相介紹彼此對各種問題的觀點，對此，雙方認為是有益的 ‘now having the opportunity to introduce to each other their views on a variety of issues, this, both parties regard as beneficial’ (5.1) The two sides found it beneficial to have this opportunity to present to one another their views on a variety of issues. According to Fang, the fact that the expression of personal attitude or opinion, rather than reporting the event or fact, takes the sentence-initial position reflects an extrovert Western culture, as seen in (4.1) and (5.1). In contrast, in Chinese, the initial position is typically given to the event, as shown in (4) and (5), which reflects a generally more reserved cultural tendency. 88 accountability in translation as well as in other modes of communication. It calls for renewed attention to linguistic principles to enhance the rigour and depth of translation studies. In doing so, it opens new pathways for understanding how texts mediate between languages, cultures, and human experiences, whether these texts are produced by humans or AI programs. References Barthes, Roland. 1977. “Introduction to the structural analysis of narratives.” In Barthes, R. Image Music Text, selected and translated by Heath, Stephen, pp. 79–124. London: Fontana Press. Barthes, Roland. 1987/2004. Criticism and Truth, translated and edited by Keuneman, Katrine Pilcher. London and New York: Continuum. Butler, Christopher S. 2003. Structure and Function: A guide to three major structural-functional theories, Part 2: From clause to discourse and beyond. Amsterdam/Philadelphia: John Benjamins. Downing, Pamela. 1995. “Word order in discourse: By way of introduction.” In Word Order in Discourse, edited by Downing, Pamela and Michael Noonan, pp. 1–27. Amsterdam and Philadelphia: John Benjamins. Eco, Umberto. 1992. Interpretation and Overinterpretation (With Richard Rorty, Jonathan Culler, Christine Brooke-Rose), edited by Collini, Stefan. Cambridge: Cambridge UP. Fang, Mengzhi [方夢之]. 1999.《翻譯新論與實踐》 (Translation Studies and Practice). Qingdao: Qingdao Chubanshe. Fang, Yuqing [房玉清]. 1992.《實用漢語語法 (‘a practical Chinese grammar’)》. Beijing: Beijing Yuyan Xueyuan Chubanshe. Favretti, Rema Rossini, Giorgio Sandri, and Roberto Scazzieri. 1999. “Translating languages: An introductory essay.” In Incommensurability and Translation: Kuhnian perspectives on scientific communication and theory change, edited by Favretti, Rema Rossini, Giorgio Sandri, and Roberto Scazzieri, pp. 1–29. Cheltenham, UK and Northampton, USA: Edward Elgar. Gómez-González, María Ángeles. 2001. The Theme-Topic Interface: Evidence from English. Amsterdam/Philadelphia: John Benjamins. Halliday, M.A.K. 1994. An Introduction to Functional Grammar. London: Arnold. Halliday, Michael A.K. 1999. “The grammatical construction of scientific knowledge: The framing of the English clause.” In Incommensurability and Translation: Kuhnian perspectives on scientific communication and theory change, edited by Favretti, Rema Rossini, Giorgio Sandri, and Roberto Scazzieri, pp. 85–116. Cheltenham, UK and 89 Northampton, USA: Edward Elgar. Halliday, M. A. K. 2004. An Introduction to Functional Grammar (Third Edition). Revised by Matthiessen, Christian M. I. M. London: Hodder Education. Huemer, Wolfgang. 2004. “Introduction: Wittgenstein, language and philosophy of literature.” In The Literary Wittgenstein, edited by Gibson, John and Wolfgang Huemer, pp. 1–13. London and New York: Routledge. Iser, Wolfgang. 2006. How to Do Theory. Malden, Oxford, and Carlton: Blackwell. Lakoff, George and Johnson, Mark. 1980. Metaphors we Live by. Chicago: The University of Chicago Press. Lambrecht, Knud. 1994. Information Structure and Sentence Form: Topics, focus, and the mental representations of discourse referents. Cambridge: Cambridge UP. LaPolla, Randy J. 2015. “On the logical necessity of a cultural and cognitive connection for the origin of all aspects of linguistic structure.” In Language Structure and Environment: Social, cultural, and natural factors, edited by De Busser, Rik and Randy J. LaPolla, pp. 31–44. Amsterdam and Philadelphia: John Benjamins. Li, Linding [李臨定]. 1984/1992.〈究竟哪個「補」哪個？——「動補」格關係再議 (‘which complements which?—rethinking “Verb-Coplement” constructs’)〉. In 《現代漢語補語 研究資料 (‘research data on complements in Modern Chinese’)》, edited by Beijing Yuyan Xueyuan Yuyan Jiaoxue Yanjiusuo, pp. 495–503. Beijing: Beijing Yuyan Xueyuan Chubanshe. Li, Yunxing [李運興]. 1998.《英漢語篇翻譯 (‘English-Chinese text translation’)》. Beijing: Tsinghua UP. Liu, Miqing [ 劉宓慶]. 1992. 《漢英對比與翻譯 (‘Chinese-English comparison and translation’)》. Nanchang: Jiangxi Jiaoyu Chubanshe. Lodge, David. 1966/2002. Language of Fiction: Essays in criticism and verbal analysis of the English novel. London and New York: Routledge. MacKenzie, Ian. 2002. Paradigms of Reading: Relevance theory and deconstruction. Basingstoke and New York: Palgrave Macmillan. Munday, Jeremy, Sara Ramos Pinto, and Jacob Blakesley. 2022. Introducing Translation Studies: Theories and applications (fifth edition). London and New York: Routledge. Pan, Wenguo [潘文国]. 1997.《汉英语对比纲要 (‘an outline of comparative Chinese-English studies’)》. Beijing: Beijing Yuyan Wenhua UP. Schwabe, Kerstin and Susanne Winkler. 2007. “On information structure, meaning and form: Generalizations across languages.” In On Information Structure, Meaning and Form: Generalizations across languages, edited by Schwabe, Kerstin and Susanne Winkler, pp. 1–29. Shan, Qichang [單其昌]. 1990.《漢英翻譯技巧 (‘techniques of English-Chinese translation’)》. 90 Beijing: Foreign Language Teaching and Research Press. Sidiropoulou, Maria. 2004. Linguistic Identities through Translation. Amsterdam and New York: Rodopi. Snell-Hornby, Mary. 2006. The Turns of Translation Studies: New paradigms or shifting viewpoints? Amsterdam and Philadelphia: John Benjamins. Tufte, Virginia. 1971. Grammar as Style. New York: Holt, Rinehart and Winston. Inc. Venuti, Lawrence. 1995/2004. The Translator’s Invisibility: A history of translation. London and New York: Routledge / Shanghai: Shanghai Foreign Language Education Press. Venuti, Lawrence. 1998. The Scandals of Translation: Towards an ethics of difference. London and New York: Routledge. Weick, Karl E. 1995. Sensemaking in Organizations. Thousand Oaks, London, New Delhi: Sage Publications. Wu, Jingcun [吳競存] and Liang Boshu [梁伯樞]. 1992.《現代漢語語法結構與分析 (‘modern Chinese: syntactic structures and analyses’)》. Beijing: Yuwen Chubanshe. Xu, Jie [徐傑]. 2001.《普遍語法原則與漢語語法現象》(Grammatical Principles and Grammatical Phenomena). Beijing: Peking UP. Xu, Liejiong [徐烈炯]. 2005.〈幾個不同的焦點概念 (‘varied conceptions of focus’)〉. In 《焦 點結構和意義的研究 (‘studies on focal structures and meaning’)》, edited by Xu, Liejiong and Pan, Haihua, pp. 11–33. Beijing: Foreign Language Teaching and Research Press. Xu, Liejiong [徐烈炯] and Pan Haihua [潘海華], eds. 2005. 《焦點結構和意義的研究 (‘studies on focal structures and meaning’)》Beijing: Foreign Language Teaching and Research Press. Preface, pp. i–iv. Zhang, Bin [張斌], ed. in chief. 2010.《現代漢語描寫語法》(‘a descriptive grammar of Modern Chinese’). Beijing: The Commercial Press. Zhang, Yufeng [張豫峰]. 2006.《現代漢語句子研究 (‘studies on modern Chinese sentences’)》. Shanghai: Xuelin Chubanshe. Zhu, Chunshen. 1996. “Syntactic status of the agent and information presentation in translating the passive between Chinese and English”. Multilingua: Journal of cross-cultural and interlanguage communication 15:4, pp. 397–417. Zhu, Chunshen. 2022. Fathoming Translation as Discursive Experience: Theorization and application. London and New York: Routledge. 91 6 Abductive inference in Mandarin interactions ZHENG Wuxi South China Normal University Abstract: This paper describes projection in the form of brief interruptions in Mandarin interactions and looks at how interjections are based on projection from specific grammatic or lexical features. These projections manifest the role of abductive rather than deductive inference in communication. Individuals’ knowledge involves more than the situation (context) in the process of abductive inference. Abductive projection is, essentially, a hypothesis that is not deterministic. The cases investigated here show that the interruptee in a conversation judges the truthfulness of the interrupter’s abduction in the immediately following utterance. The projections in the form of interruptions are generally unsuccessful and single cooperative interruptions. The aim of the interruptee is not to take the floor but to encourage or help the interrupter to continue talking by showing interest.  Keywords: projection; interruption; abductive inference; Mandarin 1. Introduction Auer (2005:8) defined projection as “an individual action or part of it foreshadows another. In order to understand what is projected, interactants need some kind of knowledge about how actions (or action components) are typically (i.e., qua types) sequenced, i.e., how they follow each other in time.” The mental process part of “projection” is talked about in psychology as anticipation or prediction, and it has been argued to be the main thing our brains to, so Clark (2013) argued that “the brain is a prediction machine”. Projection involves the ability to guess what is coming next in an interaction, and grammatical mechanisms assist the hearer in making such guesses. The ability to make these guesses is referred to as abductive inference. Peirce (1965b: 99) distinguished three categories of reasoning, namely, deduction, induction, and abduction, using statements about beans as examples. Induction proceeds from the case (These Zheng, Wuxi. 2025. Abductive inference in Mandarin interactions. In Chenglong Huang & Xu Yang (eds.) The Creation of Meaning: A Festschrift in Honour of Randy J. LaPolla, pp. 91–119.  92 beans are from this bag) to the result (These beans are white) to the rule (All the beans from this bag are white). The conclusion is a rule because it represents a generalization about all the beans in the bag. Deduction proceeds from the application of a general rule (All the beans from this bag are white) to a particular case (These beans are from this bag) to a result (These beans are white); for Peirce (1965a: 374), “all deduction is merely the application of general rules to particular cases.” Abduction proceeds from the rule (All the beans from this bag are white) and the result (These beans are white) to the case (There beans are from this bag). The basis of his classification, then, is that “deduction proves that something must be; induction shows that something actually is operative; abduction merely suggests that something may be” (1965b:106). Peirce defined abduction as “the process of forming an explanatory hypothesis” (1965b:106), considering it prior to induction and deduction—in other words, first, a hypothesis is formed, and then it is tested through deduction and induction. He considered abduction an extremely valuable scientific methodology, declaring that “Every single item of scientific theory which stands established today has been due to abduction” (1965b:106). Scholars have various ideas about the factors that play in communication. Sperber and Wilson (1986:94) offered a general outline of the deduction system used for spontaneous inference and the comprehension of normal utterances, arguing that the function of this system is to analyze and manipulate the conceptual content of assumptions (1986:103). LaPolla (1997, 2015, 2019) argued that abductive inference, rather than deductive inference, plays a role in communication, with communicators applying abductive inference to infer a speaker’s future intention. Sperber and Wilson (1986:3) argued further that the code model (the notion that communication involves encoding and decoding messages) and the inferential model (the notion that communication involves producing and interpreting evidence) are not incompatible and can be combined in various ways. From their perspective, language is a code that pairs phonetic and semantic representations of sentences, but there is a gap between the semantic representations of sentences and the thoughts actually communicated by utterances that is filled, not by more coding, but by inference (1986:11). LaPolla (2019:166) argued to the contrary that communication involves, not coding and decoding, but ostension and abductive inference,17 and, further (2015:33) that a person performs an ostensive act to show the intention to communicate something, while the other party to the conversation must infer (guess) the communicative intention behind the ostensive act. From this perspective, then, communication is achieved, not through coding, but through ostension and abductive inference. If communication is achieved through the encoding and decoding of messages, the meaning would be deterministic, while the 17 LaPolla’s (1997, 1998, 2015, 2016, 2019) use of the term ostension follows that of Sperber and Wilson (1986), but his understanding of the concept diverges in many important ways from Relevance Theory. 93 meaning achieved through abductive inference would be non-deterministic. Interruption is a communicative strategy associated with projection. According to the sequential model, interruption or overlap describes instances in which more than one speaker attempts to communicate at the same time (Sacks et al. 1974), for instance, when one speaker begins to talk before another has finished speaking (Weatherall and Edmonds 2018). Previous studies have classified interruptions based on various characteristics. Mishler and Waxler (1968) distinguished between two types of interruption, successful, in which the interrupter prevents the interruptee from completing a statement, and unsuccessful, in which the interrupter fails to take the floor before the interruptee has completed a statement. Roger et al. (1988), by contrast, focused on the number of interruption attempts, in particular, comparing one attempt (a single interruption) with multiple attempts (a complex interruption), while Murata (1994) distinguished between cooperative and intrusive interruptions.18 A “disagreeing interruption” denies the content of the previous interlocutor’s turn, presents opposite arguments, raises doubts, or questions another’s opinion—for the fact is that not all interruptions are intended to disrupt the previous interlocutor’s turn and take the floor. Rather, the aim of some interruptions is to show involvement and sustain the interruptee’s topic. The present study describes shallow, unsuccessful, single interruptions, whether cooperative or expressing disagreement as a form of projection. Typologically distinct languages show various types of projection (LaPolla 2019:167). Communication involves anticipating what is to come, so languages provide clues regarding what to expect. What’s more, projection is quite important to communication. As Hopper (2012: 208) explained, “Projection is what makes verbal communication an open and collaborative affair; as participants develop a sense of where the discourse is going, they tacitly mould it, allow it to continue, harmonize with the speaker’s goals, interrupt it with their own contribution, offer supportive tokens of various kinds, or predict when their turn will come”. During interaction, projection may lead to interruption. The hearer infers what the speaker is about to say and interrupts with a contribution. In this study, I identify clues that help speakers to make projections in Mandarin conversations (Section 3.1). This paper is organized as follows. In Section 2, I describe the data collection. In Section 3, I present the analysis of correct inferences, including the point at which projection occurs and the motivations for it. In Section 4, I discuss wrong inference and my analysis of the participants’ incorrect inferences. In Section 5, I present my conclusions. 18 Murata (1994) further distinguished intrusive interruptions as either topic-changing, floor-taking, or disagreement; the focus here is on the latter type. 94 2. Data The source of the data for this study is a corpus of 10 conversations among six participants that I recorded and transcribed in 2022 amounting to nearly 10 hours. These six native speakers were about 20 years old at the time and fluent in Standard Mandarin (Putonghua). They are undergraduates from the same university. The conversations are naturally-occurring. Each conversation involves two participants, who are classmates or close friends. Among the collections, I identify 103 cases of projection as the basis for a description of the characteristics of the phenomenon in that language. Among the 103 cases, 91% of the instances of projection involve correct abduction and 9% involve incorrect abduction. The recorded conversations show that, in most cases, the interrupter’s projection is immediately confirmed by the interruptee (i.e., dui ‘correct’). This result proves that the interactants in this study have the ability to make a correct projection in most situations. However, some situations are unexpected, and result in incorrect projection. I discuss the correct inferences in Section 3 and the incorrect inferences in Section 4. In practice, projection in the form of interruption and overlap is fairly infrequent, occurring on average 10 times during a half-hour-long conversation. An interactant may infer or guess what comes next during an interaction but may not always express the inference explicitly. Brief interruptions can serve this purpose, expressing a participant’s projection, which was inferred using abductive inference. 3. Correct inferences The discussion here addresses two questions regarding correct abductive inferences, the point at which projection occurs (§ 3.1) and the motivations for it (§ 3.2). 3.1 The point at which projection occurs The 103 cases of projection analyzed in this study include three common types: a modifier projects its modified head, a main clause projects a subclause, and a copula projects a complement. 3.1.1 Modifier projects its modified head The first type of interrupter’s projections occurs after modifiers such as a numeral/demonstrative-classifier (example 1), noun (example 2), adjective, or adverb in the 95 interruptee’s utterance. In Chinese these modifiers all preceded the head noun. During communication, the hearer makes an abductive inference about the referents based on the description in the speaker’s ostensive act. Irrespective of the speaker’s ostensive act, the hearer may make an inference based merely on personal experience. Accordingly, projection is more likely to occur after a modifier. The position of the head noun at the end of the phrase is a point where projection often occurs. Thus, the head noun can often be omitted because of this. In (1), the projection shexiang ‘camera’ occurs in the utterance of the speaker WM (Line 8). In the preceding context, the speaker RM introduces two facts: first, the students made a video recording of the teacher’s singing performance; second, cell phones were prohibited on campus (Lines 2–6). There is a short pause after the modifier na-ge ‘that’ (Line 7). The hearer WM makes a projection in the form of an interjection after the speaker RM’s short pause. The projection is an abductive inference (guess) that is confirmed to be true by the speaker RM (Line 9). Abduction thus proceeds from the rule (Cell phones were prohibited on campus) and the result (The students make a video of the teacher’s singing performance) to the case (The singing performance was recorded by video equipment, but not cell phones). In fact, video cameras and cell phones are the most common but not the only form of video equipment. So, if cell phones were forbidden, a video camera would, presumably, have been used to record the teacher’s singing performance. (1) [Conversation 9_344，00: 30: 06] 1 WM:你们唱的什么歌？ nimen chang-de shenme ge? 2 PL sing-NOM what song ‘What song did you sing?’ 2 RM:我和我的祖国，然后就是你就可以看到 wo he wo-de zuguo, ranhou jiu shi ni jiu keyi kandao 1 SG and 1SG-GEN country then just COP 2SG just can see ‘Me and My Motherland. You could see 3 他们各个摄像头一直在那里拍拍拍拍拍。 tamen gege shexiangtou yizhi zai nali pai~pai~pai. 3 PL every camera ceaselessly at there take.video they kept videoing.’ 4 WM:也许。 maybe 96 ‘Maybe.’ 5 RM:然后，就是现在的学生跟我们当时 jiu shi xianzai-de xuesheng gen women dangshi just COP now -ASSOC student compare 1PL that.time 6 的很不一样，就是他们(.)都不能用 de hen bu yiyang, jiu shi tamen(.) dou buneng yong NOM very not same just COP 3PL all cannot use ‘They(.) are quite different from us at times like that.’ 7 手机，但是他们都会有那个(.) shouji, danshi tamen dou hui you (na-ge) telephone but 3PL all can have that-CL ‘They were not allowed to use cell phones, but all of them had one (.)’ 8→ WM:摄像。 shexiang. camera ‘Camera.’ 9 RM:摄像机。 shexiangji. camera ‘Camera.’ 10 WM: 哦，摄像机。 wo, shexiangji. oh camera ‘Oh, camera.’ 11 RM : 对。他们就平常就蛮记录 dui. tamen jiu pingchang jiu man jilu correct 3PL just usual just very record ‘Correct. They usually record 97 12 他们的生活的。 tamen-de shenghuo de. 3 PL-GEN live PART their lives.’ Another example of the speaker projecting a head noun is presented in (2). The projection gaokaoti ‘entrance examination’ (Line 7) is the head noun of the modifier gaozhong-de na-zhong wenyanwen-de ‘the classical Chinese tested in high school’ (Line 6). The hearer XY makes an abductive inference based on the speaker ZT’s description of the examination in the preceding utterance (Lines 2–6). In this context, the punctuation, content words, and function words are not tested only in the entrance examination, whereas many examinations are administered during the three-year high school studying period. Therefore, the speaker’s inference regarding gaokaoti ‘entrance examination’ may be true rather than must be true. The speaker XY’s hypothesis is proved to be correct by the speaker ZT (Line 7). The abductive inference is based on personal experience. Many examinations are conducted during high school, but the entrance examination is the most important and memorable. Both of the interlocutors have experienced the entrance examination. In communication, speakers are more willing to discuss events or entities that are familiar to all interlocutors, for excessive new information makes it difficult for hearers to follow a conversation. (2) [Conversation 14_117，00: 08: 05] 1 XY: 那古代汉语考吗？ na gudaihanyu kao ma? that Classical.Chinese exam PART ‘Did (you) take the classical Chinese exam?’ 2 ZT: 考，但是他(.)考大概每年就是(.) kao, danshi ta(.) kao dagai meinian jiu shi(.) exam but 3SG exam about every.year just COP ‘Yes, (I) took the exam, but this (.) 3 考一考你的断句， kao-yi-kao ni-de duanju, exam-one-exam 2SG-GEN punctuation exam covered punctuation, 104 and/or disrespectful to the speaker. In Chinese culture, interruption is welcomed in some situations as an indication of active involvement in a conversation or discussion whereas in other situations it is considered impolite. The interpretation of interruptions depends on the relationship between interlocutors and the circumstances surrounding conversations. In (7), the speaker WM introduces her experiences of taking part in the national civil servants exam. Because the interactant WM has obtained a postgraduate recommendation, she does not need to work as a civil servant, meaning that she simply wanted to try doing so; the results of the test were not important for her. In the interruptee RM’s utterance, the projection is the summarized information shi yi xia ‘have a try’ (Line 7) rather than the less important information in the exchange. In the subsequent context, the speaker WM immediately confirms RM’s inference and then introduces more details to explain why she took the exam. (7) [Conversation 7_16，00: 01: 21] 1 WX:你参加过国考吗？ ni canjia-guo guokao ma? 2 SG join-EXP national.civil.servants.exam PART ‘Did you take part in the national civil servants exam?’ 2 WM: 对，我参加了一次国考。 dui, wo canjia-le yi-ci guokao. correct 1 SG join-PFV one-times national.civil.servants.exam ‘Yes, I took the test once.’ 3 因为，(.)事情其实是这样的。 yinwei, (.) shiqing qishi shi zheyang de. because thing actually COP this.way PART ‘Because, in fact, the situation was like this.’ 4 就是其实那个时候，嗯， jiu shi qishi na-ge shihou, en, just COP actually that-CL time um ‘In fact, at this time, 5 保研(.)情况已经出来了嘛。 baoyan(.) qingkuang yijing chulai-le ma. postgraduate.recommendation result already come.out-PFV PART 105 the results of the postgraduate recommendations had come out.’ 6 已经上岸了。 yijing shangan-le. already land-PFV ‘(I) was already successful.’ 7 RM:对。 dui. ‘Correct.’ 8 但是我爸妈还是想让我就(.) danshi wo bama haishi xiang rang wo jiu(.). but 1SG parents still want let 1SG just ‘But my parents still hoped that I would (.)’ 9→ RM: 试一下。 shi yi-xia. try one-times ‘Have a try.’ 10 WM: 对，试一下。他们说你就尝试一下， dui! shi yi-xia. tamen shuo ni jiu CHANGSHI yixia, correct try one-times 3PL say 2SG just try one-times ‘Correct! Give it a try. They said, ‘You just have to try.’’ 11 你就不要有压力呀。什么。 ni jiu bu-yao you yali ya. shenme. 2 SG just NEG-will have press PART w h a t ‘Don’t feel stressed! Something like this.’ 12 随便去体验一下。 suibian qu TIYAN((glaring)) yi-xia. feel.free go experience one-times ‘Go experience (the test) casually.’ 13 然后我就报了那个国考。 106 ranhou wo jiu bao-le na-ge guokao. then 1 SG just apply-PFV that-CL national.civil.servants.exam ‘Then I applied for national civil servants exam.’ 3.2.2 Cooperative interruption to fill a moment of hesitation During interactions, when a speaker slows or pauses briefly, a conversational partner may perceive the hesitation and supply what the speaker is searching for. Murata (1994:387) described this situation, in which a conversational partner, with no intention of taking the floor, encourages a speaker to continue by showing interest, listenership, and participation, as “cooperative interruption.” In (8), the interactant RM cannot think of the word for beidan ‘sheet’, and says mianbei ‘quilt’ with a rising intonation to express uncertainty (Line 2). Another interactant, WM, interrupts RM after a short pause and makes abductive inferences, xizi ‘mat’ and beidan ‘sheet’ based on the attributive expression bao ‘thin’ and the classifier zhang. The interactant WM guesses that the missing word could be yi-zhang xizi ‘a mat’ or yi-zhang beidan ‘a sheet’ (Line 4). If the classifier is tiao rather than zhang, WM would guess that the missing word is yi-tiao tanzi “a blanket.” This example illustrates that phonological features such as intonation and pauses can communicate a speaker’s hesitation or uncertainty, and a classifier can pragmatically help the hearer make an inference. The interactant WM is, then, cooperating with another interactant, RM, in supporting the flow of the conversation by supplying the word that RM is trying to find, beidan ‘sheet’. (8) [Conversation 9_32，00: 03: 31] 1 RM: 我当时就因为没有(.)没有带那个睡垫， wo dangshi jiu yinwei meiyou meiyou dai na-ge shuidian, 1 SG then just because NEG-have NEG-have bring that-CL mattress ‘Just because I didn’t take that mattress 2 然后只带了一张那种(.)棉被(?)。 ranhou zhi dai-le yi-zhang na-zhong(.) mianbei(?), then only bringPFV one-CL that-kind quilt but only took a quilt. 3 也不是，就很薄的一张(.) ye bu-shi, jiu hen bao-de yi-zhang(.) 107 also NEG-COP just very thin-ASSOC one-CL No, it wasn’t a quilt, exactly, it was a very thin (.)’ 4→ WM: 席子， 被单。 xizi? beidan? mat sheet ‘Mat? Sheet?’ 5 RM: 一张被单，然后就铺在上面睡。 yi-zhang beidan, ranhou jiu pu zai shangmian shui. one-CL sheet then just cover at above sleep ‘It was a sheet. Then (we) just slept on (the sheet).’ Another example of filling a moment of hesitation occurs in (9). In the utterance of the interactant LC, a short pause occurs after the numeral-classifier yi-ge “one” (Line 3). She wants to say “it is a secret” or “it is cheating behavior,” and her hesitation indicates a negative attitude toward a computer store’s deceptive action. The interactant LC feels that it is inappropriate to discuss such illegal behavior.19 This example illustrates the use of a phonological pause to express the interlocutor’s stance. At the moment of the pause, the hearer LC captures the turn timing and says dou dong de “anyone can understand” to indicate that she perceives what the speaker ZC intends to say, that is, to demonstrate agreement. In fact, the hearer LC does not say the head noun of the numeral-classifier yi-ge “one,” but the speaker ZC still says dui “correct” in response, and then she describes the cheating behavior in detail. (9) [Conversation 2_479，00: 22: 14] 1 ZC: 他说你怎么可以放在人家店里？ ta shuo ni zenme keyi fangzai renjia dian-li? 3 SG say 2SG how can put other store-in ‘He said, ‘Why did you leave (your laptop) at someone’s store?’’ 2 他说他们这些其实是就是(.)， ta shuo tamen zhexie qishi jiu shi(.) 3 SG say 3PL these actually just COP ‘He said, “They actually were (.)’ 19 LC may have been unsure whether it was appropriate to say such things when her words were being recorded. 108 3 嗯，可能也是他们这一行的一个(.) en, keneng ye shi tamen zhe-yi-hang-de yi-ge(.) en maybe also COP 3PL this-one-trade-ASSOC one-CL ‘Uh, it may be their business’s (.)’ 4→ LC: 都懂的。 dou dong de. all understand PART ‘Everyone knows that (the store is secretly cheating).’ 5 ZC: 对，就是他会给你动一点手脚的， dui. jiu shi ta HUI gei ni dong yidian shoujiao de, correct just COP 2SG can give 2SG do a.little cheat PART ‘Correct. That means he could cause some problems on (your laptop).’ 6 然后可能本来你有一点问题， ranhou keneng benlai ni you yidian wenti, then maybe originally 2SG have a.little problem ‘Maybe your (laptop) just has a little problem.’ 7 他会让你的问题更严重一点。 ta hui rang ni-de wenti geng yanzhong yidian. 3 SG can let 2SG-GEN problem more serious a.little ‘He would make the problem worse.’ 3.2.3 Interruption as a repair device When an interlocutor makes an incorrect projection, the interrupter may rectify it, in which case interruption functions as a repair device for correcting errors in conversation. Correction is an abductive inference based on the preceding context and personal experience. Murata (1998:390) called this type of interruption, which occurs when the next speaker feels the need to disagree with what the current speaker says, “disagreement interruption”. Thus, in (10), the speaker ZC says that the stores in tall buildings are similar to producers and factories, but the hearer LC believes that ZC has spoken wrongly because producer is equivalent to factory. ZC’s utterance is interrupted, though she hasn’t finished speaking the word gongchang ‘factory’ (Line 5). The projection is xiaoshou ‘seller’ (Line 6). Based on her personal knowledge, the stores in a 109 computer plaza not only produce computers but also may sell them. LC’s correction took place instantly, resulting in an interruption. (10) [Conversation 2_424，00: 19: 11] 1 ZC: 因为他们不是那种什么 yiwei tamen bu shi na-zhong shenmen because 3PL NEG COP that-kind what 2 高楼大厦里面那些店铺。 galoudasha limian naxie DIANPU, tall.buildings inside those store ‘Because they are not like those stores in tall buildings.’ 3 LC: 对。 对。 dui dui ‘Correct. Correct.’ 4 他们算是，又是像生产端， tamen suanshi, you shi xiang shengchanduan, 3 PL regard.as also COP like producer 5 又像是工(.) you xiang shi gong(.) also like COP factory ‘They are like a producer and factory.’ 6→ LC: 销售。 xiaoshou. seller ‘Seller.’ 7 ZC: 对，又像是销售， dui. you xiang shi xiaoshou, correct also like COP sell ‘Correct. (They) also resemble a seller.’ 110 8 LC: 对。 dui. ‘Correct.’ 9 然后也有维修什么的。 ranhou ye you weixiu shenme de. then also have repair what PART ‘Then (they) also can repair (computers).’ 10 ZC: 对，比较全面。 dui. bijiao quanmian. correct relatively comprehensive ‘Correct. (They) are fairly comprehensive.’ 3.2.4 Further exploring the preceding turn In communication, an interlocutor may make a conclusive statement based on inference from personal experience, encyclopedic knowledge, or the preceding context. Communicative projection is, to a large extent, driven by an interlocutor’s existing knowledge and a speaker’s ostensive acts. In (11), the speaker ZT’s utterance is interrupted by the hearer XY (Line 7). ZT is introducing her personal experiences regarding the choice of a school for the postgraduate examination (Lines 1–6). Both interlocutors know that the admission rate of Chinese Studies is so low that many students are afraid to choose this major. This fact is known by both interlocutors. Therefore, when the speaker ZT says “I checked the previous year’s admission rate,” the hearer XY interrupts to make a projection: “You were scared, right?” and ZT immediately confirms the projection. (11) [Conversation 14_22，00: 01: 37] 1 ZT: 我是其实，我本来首先是想 wo shi qishi, wo benlai shouxian shi xiang 1 SG COP actually 1SG originally firstly COP want 2 考学语的,学科语文， kao xueyu-de xueke yuwen, exam Chinese.discipline-ASSOC discipline Chinese ‘Actually, I originally wanted to take the exam of Chinese Studies. 111 3 然后我想的是考本校嘛， ranhou wo xiang-de shi kao benxiao ma, then 1 SG want-NOM COP exam this.school PART Then I wanted to take our school’s exam. 4 后来，后来就是那个初试。 houlai, houlai jiu shi na-ge CHUSHI, later later just COP that-CL first.test Later, that was the first test. 5 他们一考完初试之后， tamen yi kao wan chushi zhihou, 3 PL once exam completely first.test after 6 我就查那个往年的报录比。 wo jiu cha na-ge wangnian-de baolubi. 1 SG just check that-CL previous.years-ASSOC admission.rate After they finished the first test, I checked the admission rate.’ 7→ XY: 吓到了，是吗？ xiadao-le, shi ma? scared-PFV COP PART ‘Were you scared?’ 8 ZT: 对，我被吓到了， dui. wo bei xiadao-le, correct 1SG PASS scared-PFV ‘Correct. I was scared.’ 9 我觉得自己感觉不是那么的(.)有把握。 wo juede ziji ganjue bushi name-de(.) you bawo. 1 SG think self feel NEG COP such-ASSOC have confidence ‘I don’t feel very confident.’ 112 4. Incorrect inferences Unlike deductive inference, abductive inference is non-deterministic. One person performs an ostensive act, and another infers the communicative intention using abductive reasoning. A hypothesis is not definitely true. In the recorded conversations, 10 of the 103 cases involve incorrect inferences (projections). In (12), the participants discuss their internship experiences in middle and high school. In her utterance, ZC slows down the pronunciation of bijiao-de jiu to show her difficulty finding an appropriate word to describe the performances of the students in middle school. The interactant LC interrupts ZC’s utterance to guess that the students are unwilling to answer the teachers’ questions, having formulated the hypothesis through abductive inference that “not willing to answer teachers’ questions was one of the negative characteristics of the classroom.” The speaker LC’s inference regarding the utterance of ZC is immediately proved wrong: rather than refusing to answer their teachers, the students are simply “inattentive in the classroom.” Not answering the teachers’ questions is considered a manifestation of bad behavior in the class. This example shows that projection or inference in communication is based largely on the hearer’s existing knowledge. Context is important for the addressee, but personal experience plays a larger role in helping interlocutors to make an inference. (12) [Conversation 1_725，00: 41: 02] 1 ZC: 嗯，嗯，我觉得可能高中课堂会好一点， en en wo juede keneng gaozhong ketang hui hao yes yes 1SG think maybe high.school classroom would good 2 初中的话，还是(.)<会比较的>，就。 yidian, chuzhong dehua, haishi(.) hui <bijiao-de> jiu. a.little middle.school as.for still would relative-ASSOC j u s t ‘Yes, I think the (situation) in the high school’s classrooms would be a little better. As for middle school, (the situation) would be <relatively> (.)’ 3→LC: 不回答。 bu huida NEG answer ‘(The students would) not answer.’ 113 4 ZC: 也不是不回答，就是他们没那么，<嗯> ye bushi bu huida, jiu shi tamen mei name, <en>, also NEG.COP NEG answer just COP 3PL NEG so um ‘It is not the case that they wouldn’t answer (their teachers’ questions).<um>, 5 就是，上课可能没那么集中， jiu shi, shangke keneng mei name JIZHONG, just COP attend.class maybe NEG so concentrate They may not be, concentrating in class. 6 你知道吗？他们会，他们会， ni zhidao ma? tamen hui, tamen hui 2 SG know Q 3PL would 3PL would You know, they may be, 7 就是会分神。 jiushi hui fenshen. just COP absent-minded They may be inattentive.’ In (13), the abduction proceeds from the rule (different hobbies may cause conflict) and the result (they quarreled) to the case (they quarreled because of hobbies). The hearer XY was not very familiar with the individual under discussion, Longjie, so she simply guesses why the individuals in question quarreled. Though the hearer XY does not know the reason for the quarrel, she interrupts the speaker ZT and makes a projection, aihao “hobby.” This example illustrates the important role that abductive inference plays in interactive communication. (13) [Conversation 16_143，00: 11: 36] 1 ZT: 而且她平时会跟龙杰 erqie ta pingshi hui gen Longjie, in.addition 3SG usual would with person.name ‘In addition, she would usually 2 她看很多动漫嘛。 kan henduo dongman ma. look many cartoon PART 120 7 现代汉语超常规三重否定：形成机制、类型与动因 李姝姝 北京语言大学国际中文教育研究院 内容摘要 双重否定表肯定是汉语无标记的一种否定用法，而语言实践中 存在着一种超常规否定用法——三重否定表肯定。本文结合其他语言类型 学研究成果，并从认知加工、交际互动、言语产出等方面分析了这种超常规 否定用法的形成机制与动因，指出这种三重否定构式是两个音、形、义均有 高度相似性的双重否定构式的整合。从认知加工机制来看，否定层级的叠加 显著提升处理负荷，但听者在知觉整体性驱动下，会优先提取语境化整体语 义而非精确计算否定算子数量。这种“意义完形优先”原则导致三重否定构 式在实际交际中被重构为特殊强调手段，其理解过程印证了 LaPolla（2015、 2021）的“意义创造论”，即语言交际本质上是语境驱动的动态协商而非形 式符号的静态推导。 一、前言 自然语言不完全等同于逻辑语言，在日常交际中，人们的话语不总是遵循着逻辑的 规定来进行构建。譬如，过去曾受到关注的羡余否定就无法用“负负得正”“正负得负” 的数学逻辑来运算。除了比较常见的羡余否定（如“好不热闹”）外，有一种三重否定表 肯定的超常句法现象散布于网络与语料库中。例如： (1) 这两句话，如果当作笑话听，那么没有一个女人不会不承认它说得有理。（《十 日谈》，BCC 语料库，1987 年20） (2) 没有什么产品不会不过时，无论这种产品多么先进，多么拥有市场，都是临时 的。（《玻璃企业：从突破口突破》，中国知网 1998 年） (3) 他有着让人心动的美丽容颜，恐怕世界上没有女孩子不会不对他动心吧。（《 零 之使魔》，BCC 语料库，2004 年） (4) 失败很正常，没有人不会不失败。重点是我们千万要放平心态，不要被无数的失 李姝姝. 2025. 现代汉语超常规三重否定：形成机制、类型与动因. 黄成龙、杨旭（主编）《意义创造论： 罗仁地先生七秩寿庆文集》，pp. 120–135.  20 为了让读者对三重否定表肯定的超常句法现象出现的时间有一个大致的认识，（1）-（4）的语例中标 注了时间，但是由于本文不是历时研究，后续的例子不再标注语例的时间。 121 败所击败。（《年轻人需要具备的几点能力》，网络语料，2016 年） 以上例子中，每个下画线小句都含有三个否定词。如果严格按逻辑分析，以上画线 句因为其中的两个否定互相抵消，只剩一个否定，最终应该表达否定义。但是根据语境 推理，说话者通过画线小句表达的却是肯定性命题义，如“女人会承认它说得有理”“产 品会过时”“世界上女孩子会对他动心”“人会失败”。 除了书面媒介的语料中出现三重否定用例之外，自然口语会话中也存在类似现象。 下面这则会话片段来自《锵锵三人行》，主持人和嘉宾讨论的话题是三聚氰胺毒奶粉事 件。06 行曹红辉想要表达的意思是肯定有人知道三聚氰胺这个事儿。但是曹红辉脱口而 出“不可能没有人不知”这样的一个没有说完的三重否定句，只不过在“不知”说出后， 说话人启动修复（repair），将“不知”修复为“知道”。 (5) 《锵锵三人行》会话片段 01 曹红辉：那三聚氰胺从哪儿来的呢？ 02 叶 檀：对呀。 03 曹红辉：到底谁-谁买的呢？谁采购、谁放进去呢？ 04 叶 檀：嗯 05 窦文涛：嗯 06→曹红辉：不可能没有人不知-呃知道这个事儿吧， 07 窦文涛：嗯 08 曹红辉：所以这个话是说不过去的，道理上。 09 叶 檀：对。 在汉语方言中，也存在类似现象，如粤语网络语料中，存在“冇人唔會唔 VP”“无 人唔会唔 VP”“唔會冇人唔 VP”（其中“冇、唔、无”是否定词）等三重否定表肯定的用 法，例如： (6) 如果想报学琴, 可以考虑 Crystal Chow,佢经验丰富, 而且仲曾于外地考取音乐科 硕士学位,……在琴行里, 冇人唔会唔识佢,大把人争住想跟佢。（网络语料，201921） (7) 去完韩味无穷，行完挥春街买嘢，就嚟到公正路行大笪地，讲起大笪地可以话喺 佛山无人唔会唔知道嘅灯光夜市，从晚上六点半开市，一直到深夜 11 点……（网 络语料，2019） (8) 相信 70-90 年代亚洲电影界唔会冇人唔识佢！（网络语料，2019） 在汉语作为第二语言的中介语语料中，也有三重否定表肯定的用例，例如： (9) 中国市场是那么大，怎么可能没有人不重视中国这个国家呢？（HSK 动态作文 语料库） 例（9）中，反问结构“怎么可能……呢”以及否定词（组）“没有”“不”都含否定 义，是个三重否定，如果按照逻辑来分析，这个句子表达的意思是“没有人重视中国这个 国家”， 但是按语境上下文来看，这位汉语学习者想要表达的意思却是“人们都很重视中 国这个国家”。 以下是一个正常的三重否定表否定的例子，例如： 21 （6）-（8）中的 2019 指的是检索时间，作者当时正在新加坡南洋理工大学访学，因此收集的一部分 语料是通过谷歌搜索引擎获得的，原文使用了繁体字，本文引用时转成了简体字。 122 (10) 我不是不生气， 不是不会不开心，只是觉得是小事情，没必要把情绪摆到面上 来。（网络语料） 例（10）中的画线句“不是不会不开心”的意思是“是会不开心”，这是一个三重否 定表否定的符合逻辑运算的用例。 例（1-9）中的三重否定表肯定的语言使用现象，如果从逻辑和规定性语法的角度来 看，其中的一个否定词是羡余的，这也是为什么本文称之为“超常”的原因。但是从基 于描写和使用（usage-based）的语言观来看，超语法、超常组合并不奇怪，它们是语法和 修辞研究的重要对象之一（陈满华，2015）。在世界语言中，一个句子出现三个否定词的 现象是较为少见的，对它的分析和理解也会更复杂（刘丹青，2017）。为什么汉语语言交 际中会产生三重否定形式表达肯定的用法？三重否定在结构类型上是否能够呈现出某种 特点？这背后有哪些方面的因素在起作用？本文将对这些问题进行分析和调查。文章具 体分为如下几个部分：相关研究回顾、三重否定超常构式的结构类型及形成机制、三重否 定超常构式产出的动因、余论。 二、研究回顾 2.1 关于否定的研究 在语法和语义分析中，否定表示“对句子意义作部分或全部的反驳”（Crystal， 1997/2000：237），它反映的是矛盾或对立关系，对于语义的理解作用十分重大（袁毓林， 2018：141）。赵元任（1979：348）在《汉语口语语法》一书中，就谈到汉语的否定问题： 汉语“一个短语或小句之中有两个否定词的时候，互相抵消，不是像俄法或法语那样互相 支援”。俄语中的否定性无定代词（negative indefinite pronoun，类似于英语中的 nobody, nothing, nowhere 等）总是和谓语的否定词共现（Haspelmath，2013）；法语书面语中的 “ne… pas” 结构中，“ne”与“pas”表达的两个否定义不抵消，而是进行联合否定（Horn， 1989/2001: 455），也被作为否定协和（negative concord）的代表。否定协和指的是句内 两个否定成分共现但不表达肯定义的现象 （Horn，2010） 22。以下是几则具体的语例（Horn， 2009：406；Haspelmath，2013；St-Amour，2009:436）。 (11) Nikto ne prišel. （俄语） Nobody NEG came. ‘Nobody came.’ (Haspelmath，2013) ‘没有人来。’ (12) Milan ne vidi nikoho.（捷克语） Milan NEG.sees nobody 22 “否定协和”也称“否定一致”，主要用于指与否定性无定代词（如 no, nobody, nothing, none 等）有 关的一种否定现象，即句子中有一个否定性无定代词，但与句中另一个否定词在语义不互相抵消，整 个句子表示否定义。西班牙语被认为是较为典型的“否定协和”型语言。 123 ‘Milan doesn’t see anybody’ （St-Amour，2009：436） ‘米兰没看到任何人。’ (13) Well, really, how can I keep from not worrying? （英语口语） ‘Well, really, how can I keep from worrying?’ （Horn，2009：406） ‘我怎么能不担心呢？’ 与否定协和相对的是遵循负负得正的双重否定，不过这只是一种方便称呼的大致分 类，它和否定协和并不是二元对立的，自然语言不总是严格遵循这种分界，两种类型的否 定可能同时存在于同一种语言的不同方言或不同语体中，如非洲裔美国人口语（Labov， 1972：785）、澳大利亚托雷斯海峡的英语克里 尔语奥（Shnukal，1988：72）、南非荷 兰语口语（van der Wouden，1997：192）、德语方言（Haspelmath，1997：205）中存在 否定协和现象。在有的语言中，否定协和是标准语语法的一部分，如匈牙利语、希腊语、 罗马尼亚语和斯拉夫语等（de Swart，2010：21）。 除了两个否定之间存在否定协和的现象之外，语言中还存在着“三重否定迷惑” （Triplex negatio confundit）（Horn，2009）。实际上，“三重否定迷惑”这类现象从根 本上说还是涉及到否定协和的问题，因为其中两个含否定义的词互相不抵消，它们联合 加强否定后，再与最后一个否定进行抵消，整个构式义表达肯定义。下面是英语口语中的 两则三重否定表肯定的例子： (14) There isn’t a man there who doesn’t think they can’t do it. (Horn, 2009: 408) 意译：没有一个人觉得他们不能这么做。 (15) No head injury is too trivial to be ignored. (Wason & Reich, 1979) 意译：不能因为头部创伤轻微就不重视 上述两则例子，均含有三个表达否定意义的成分，如例（14）中是“isn’t ” “doesn’t ” “can’t ”，例（15）中是“no”“too……to”“be ignored”，其中“ignored”是一个隐性否定 词23。根据 Horn（2009：408）的分析，英语母语者在阅读上述例句时，并非按“负负负 得负”的逻辑来进行否定理解， 而是作肯定理解。 24关于例（15），Wason & Reich（1979） 曾调查过英语母语者的理解情况，结果显示被调查者倾向于认为“逻辑上错误”但“语用 上合理”。也有研究者认为“No X is too Y to Z”表达肯定义的现象不能分析为逻辑不正 确或者归因于加工错误，它是一种结合了形式和意义的规约化了的否定构式，具有特殊 的修辞功能，反而是它所对应的所谓正确表达，如“No head injury is too trivial to pay attention to”无法实现同等修辞效果（Fortuin，2014：249-250）。 以上关于世界语言中的两个或多个否定语义的研究表明自然语言中的否定表达的复 杂性、多样性，以及两个否定语义成分共现共同加强否定义这种现象分布的广泛性。世 23 Horn（2009:407-408）指出，在英语口语中，当其中一个否定义由本身具有否定意义的谓语动词如 “surprised, avoid, deny, doubt”表达时，三重否定表示肯定的倾向尤其普遍。 24 在此补充另一个实验结果：对于英语否定性无定代词和谓语否定词共现的句子的解读，研究者通过眼 动实验发现英语母语人在阅读包含两个句法否定词的句子时，会同时产生肯定和否定的解读，并且当 句子的宾语为否定形式时，如“She didn’t answer nothing in that interview”，被试更容易产生强化否定 （否定协和）的解读（Blanchette & Lukyanenko，2019）。 124 界语言中关于否定表达的多样研究打破了语言否定一定会遵循逻辑否定这种假设。 汉语研究否定的文献也很丰富（如石毓智，1993；沈家煊，1994；袁毓林，2000、 2018）。不过，汉语学界讨论较多的羡余否定25大多不是指两个显性否定词之间彼此不抵 消的情况，主要是指句中显性否定词（通常只有一个）删去不影响句子命题义的表达这种 情况，如“差点儿没VP、好不 VP、小心别VP、后悔没 VP、在没有 VP 之前”。吕叔湘 （1985）称这类否定现象为“否定作用的模糊化”。 有学者认为应当为上述所说的羡余否定确定一个研究范围，将“口误”和“笔误” 排除在否定研究的范围之外，例如张谊生（2005；347）认为有些羡余否定式“似乎”（原 文表述）接近于口误或笔误，不宜将这类情况纳入羡余否定的研究范围。张文举的例子 是： (16) 谁也不能否认这些戏没有教育意义吧。（张谊生，2005） (17) 个别单位拒不执行有关这方面的规定。（同上） 本文认为像例（16）和（17）这样的语言使用现象也是可以纳入研究范围的。描写语 法（descriptive grammar）研究的目标不是评判和决定什么是好的语言，什么是不好的语 言，什么是正确的、合语法的结构，什么是不正确的、不合语法的结构（Leech et al.，2005； Lakoff，1973；Newmeyer，2003）；描写语法的研究兴趣在于发现语法在不同语境中以及 不同目的驱使下实际上是如何被应用的（Hinke，2018：1）。本文描写和分析超常规三重 否定构式，目的是跳出正误判定的窠臼，将其置于一个更广的世界语言的图景中，结合人 类交际、认知特点，来展示和分析否定在使用过程中表现出的特殊、复杂现象。 实际上，并不是只有我们关注到三重否定超常构式。陈 青（ 2018）也关注到了该现 象，文中提供了一定数量的语料，这些语料与我们自己搜集的语料形成本文的语料基础。 陈文的研究既包括含有三个显性否定词的超常规三重否定，也包括含有一个隐性否定词 两个显性否定词的情况，陈文还从语言主观性、认知隐含脱落、认知象似性等角度对超常 规三重否定现象作出解释。不过，陈文主要是就汉语自身来进行分析，没有纳入语言类型 学的研究成果，这使得其对超常规三重否定的分析视角比较局限。为了让研究对象相对 聚焦，本文的研究范围限定在含有三个显性否定词的三重否定表肯定的现象上。 2.2 句式整合26 本文要研究的三重否定超常构式的形成机制是构式整合，因此这里要对学界以往关 于句式整合的研究成果做一介绍。关于句式整合的研究，可以分为两个分支，一支是以沈 家煊（2006）为代表，沈先生受到心理空间理论和概念合成理论（Fauconnier & Turner， 1998、2000、2002）的启发，分析了汉语中的句子整合现象，沈文所说的“整合”强调的 是概念或意义的整合；另外一支以叶建军（2010、2013、2014、2020）为代表，主要关注 25 “羡余”（redundancy）“冗余”等词有“多余”之义，这是就逻辑的角度而言的，如果从语用效果的 角度讲，多余的否定词又是有功用的，不过考虑“羡余”这个称呼的通行性，本文在转述汉语文献观 点时还是会使用“羡余”一词。 26 本文将句式看作是构式的一种，构式的范围更大。在进行文献综述时，尽量延用文献中的表述。 125 的是两个语义相同或相近的句式在形式上的整合，叶建军（2020）认为构式整合（原文称 句式糅合）是汉语特殊句式的重要生成机制，叶文将“句式糅合”定义为“两个语义相 同或相近的句式 A与B，因某种语用目的主要通过删略重叠成分合并成一个新的句式 C 的过程或现象”。汉语句式糅合的现象有很多出现在六朝至近代的文献中，口语化程度高 的文献越容易出现句式糅合现象（叶建军，2010：328）。 无论是上文所说的“句式糅合”还是本文所说的“构式整合”，名称不同所指现象基 本相同，并且都需要满足一定的条件，对此学者们已有一些论述，如江蓝生（2008：484、 487）认为构式整合多发生于两个核心语义基本相同的构式之间。叶建军（2013）提出了 句式糅合的三个基本原则——语义相近原则、时代先后原则和成分蕴含原则（或语义蕴 含原则）。语义相近原则指的是“源句式 A 与 B 在语义上必须相同或相近”；时代先 后原则指的是“语义相同或相近的源句式 A 与 B 必须先于糅合句式 C 而存在，或与 糅合句式 C 同时存在”；成分蕴含原则指的是“糅合句式 C 必须蕴含源句式 A 与 B 的主要成分，甚至是全部成分”。 三、三重否定超常构式的结构类型及形成机制 3.1 三重否定超常构式的结构类型 基于本研究收集到的总计约 110 多个三重否定表肯定的用例，我们总结出两大类七 小类三重否定超常构式的常见类型。 第一类是基于否定词“不”“没”形成的三重否定构式，具体小类和对应的用例如下： A类：没有[ ]类属名词不[ ] 助动词/心理动词不[ ]动词谓语 (18) 没有什么产品不会不过时，无论这种产品多么先进，多么拥有市场，都是临时 的。（科技文献，BCC 语料库） (19) 没有人不敢不给这攻下伊谢尔伦的勇士一点面子。（《银河英雄传说》，BCC 语 料库） B类：[ ] 类属名词没有不[ ] 助动词/心理动词不[ ]动词谓语 (20) 其实学生没有不想不好好学习的。谁都想门门功课考一百分，谁都想学习好出 人头地，受人夸奖。只是摸不到学习的门径而已。（网络语料） C类：不[ ] 助动词/心理动词没有[ ] 类属名词不[ ] 动词谓语 (21) 刘德华，这个闪耀的名字，相信不会没有人不认识吧!!! （网络语料） (22) 在哪一个天涯海角举行婚礼，只要他肯请，不愁没有人会不来。（同上） D类：不是不[ ] 助动词/心理动词不[ ] 动词谓语 (23) 顾衣轻笑了一声，“我知道，你们跟着我在端居未必是最好的出路，若是你们二 人想要寻个别的出路，我又不是不会不放人。（网络语料） 第二类是基于具有文言色彩的否定词“无”“非”以及普通否定词“不”形成的三重 否定，具体小类如下： E类：并非不是不[ ] 动词谓语 (24) 另外，麦克梅纳明还强调，科比现在并非不是不想打得更加团队一些，如果他能 126 够实现得分超越乔丹这一壮举，那么他将马上回归团队篮球的表现。（《重庆晨 报》） F类：不是不无[ ] 抽象名词 (25) 雷德立克之所以相信打倒巨人就等于打倒了威诺，看来也不是不无道理。（《 罗 德岛战记》，CCL 语料库） G类：并非不无[ ] 抽象名词 (26) 倘若生逢其时，他被拥立登基为王也并非不无可能。（雪乃纱衣《彩云国物语》） 至此，我们可以用图示的方式再次归纳出前文所列举的七种三重否定超常构式的结 构类型，它们的共同点是都是基于显性的否定词形成的。 图1 显性三重否定的结构类型 上述三重否定的高频组合模式可从结构层级与命题辖域两个维度解析：其一， 在 句法线性序列中，第二与第三否定词（N2 和N3）通 过两种方式构建紧密的双重否定单 元——或于否定词间嵌入情态助动词（如“会、能”）及心理动词（如“敢、想”）， 形成“否定词+[情态助动词/心理动词]+否定词”的嵌套式框架（如“不敢不”“不能 不”），或由第二、三否定词（N2 和N3）直接词汇化为固定结构（如“不无”），该双 重否定单元进一步受首位否定词（N1，如“没有”“并非”）统辖，构成“N1+（N2+N3+X）” 的三重否定。其二，在语义辖域层面，首位否定组合（如“没有”“不是”“并非”及 “不会”）具备高层否定功能，通过将内嵌的双重否定结构整体纳入其否定辖域，形 成三重否定。 超常规三重否定内部的否定成分之间的语义关联并非严格遵循线性序列或单一逻辑 层级，而是受语境驱动与交际意图调控，允许否定在句法表层灵活组合。具体而言，三个 否定词（N1、N2、N3）可以形成不同层级的否定关联：（1）N1 与N2 优先绑定：首位否 定词 N1 可与紧邻的 N2 组合，构成双重否定单元（如“并非不无可能”中，“并非”可以 与“不可能”组合，整体表“有可能”）；（2）N1 与N3 跨层关联：N1 亦可绕过 N2，直 接与 N3 形成语义呼应（如“并非不无可能”中，“并非”可解读为对“无可能”的否定， 组合后整体等价于“有可能”）；（3）N2 与N3 关联：N2 与N3 也可组合构成双重否定（如 127 “不无可能”≈“有可能”）。 上述特征表明， 三重否定并非简单的线性否定叠加，而是由不同层级的否定成分构 成的复杂整合。 三重否定也可以基于一些显性否定词和隐性否定词（如“防止、杜绝、制止、避 免、否认、后悔”等）的组配而形成，如例（27）（28）所示。 (27) 早晨，有个什么人邀请阿尔卡季·伊凡诺维奇喝茶，他无法拒绝不去。（陀思妥耶 夫斯基《脆弱的心》） (28) 最后的结局不外是：有情人终成眷属。我们并不否认现实生活中没有这样的事 例。有，但却不像小说中塑造的那么多，那么强烈……（裘小龙《心理分析学派 试探》） 不过，为了研究的聚焦性，本文目前只重点分析基于三个显性否定词而产生的三重 否定表肯定的现象。 3.2 三重否定的形成机制 本研究认为三重否定超常构式是一种多源构式（multiple source construction）（Van de Velde et al.，2013），它是由两个不同的双重否定构式经过句法整合（syntactic blends） 而成的，它符合叶建军（2013）提出的构式整合的三个基本原则：首先，三重否定句所对 应的两个源构式的语义是相近的。以三重否定超常构式“没有 NP 不会不 VP”的形成为 例。从逻辑上讲，被整合的构式之一“没有 NP 不VP”是否定存在构式，另一个被整合 的构式“NP 不会不 VP”是类指构式，两者都表示全称肯定，逻辑恒等。但是二者又有 细微的差异，例如：否定存在构式“没有女孩子会不喜欢他”的语义逻辑是：在由语境限 定的“所有女孩子”形成的集合中，不存在一个这样的元素，这个元素可以归入由具有 “会不喜欢他”这种属性的元素构成的集合中。而类指构式“女孩子不会不喜欢他”的语 义逻辑是：在由语境限定的“所有女孩子”形成的集合中，全体元素毫无例外地都不可以 归入由具有“不会喜欢他”这种属性的元素构成的集合中，也就是说所有的女孩子都会喜 欢他。因此，“没有 NP 不VP”与“NP 不会不 VP”具有一致的核心语义，这是二者进 行构式整合的基础之一。其次，在共时层面，“没有 NP 不VP”和“NP 不会不 VP”的 出现早于三重否定超常构式“没有 NP 不会不 VP”；最后，三重否定超常构式“没有 NP 不会不 VP”蕴含了源构式“没有 NP 不VP”和“NP 不会不 VP”的全部部分。 在同义构式整合的过程中，会发生重复项归并的“噬同”现象（沈家煊，2006），即 把两个相同的重复成分进行归并。用 Coppock（2010）的话说，就是位于平行句法位置上 的两个源构式中的成分在句法整合时会形成竞争，最终保留一个。以双重否定句“没有女 孩子不喜欢他”和“女孩子不会不喜欢他”为例，在构式整合过程中，“女孩子”以及“不 喜欢他”分别发生归并（连线所示），保留其一。 128 图2 构式整合过程示意图27 下面表 1中，最右侧的三重否定超常构式，它是由左侧的双重否定构式 1和双重否 定构式 2整合形成的。 表1 双重否定句糅合而成的三重否定句 双重否定构式 1 双重否定构式 2 三重否定构式 没有人不犯错误 + 人不会不犯错误 → 没有人不会不犯错误 学生没有不想好好学习的 + 学生不想不好好学习 → 学生没有不想不好好学习的 不会有人不知道 + 没有人不知道 → 不会没有人不知道 我又不是不放人 + 我又不会不放人 → 我又不是不会不放人 并非不想打得更加团队一些 + 不是不想打得更加团队一些 → 并非不是不想打得更加团队一些 不是无道理 + 不无道理 → 不是不无道理 并非无道理28 + 不无道理 → 并非不无道理 如何确定一个三重否定构式的源构式究竟是哪几个？下面将结合具体例子来阐述。 本研究将三重否定超常构式“没有 NP 不会不 VP”的来源构式拟定为“没有 NP 不VP” 与“NP 不会不 VP”的整合，而没有选择与“没有 NP 不VP”形成竞争关系的双重否定 构式“没有 NP 不会 VP”“没有 NP 会不 VP”来作为源构式之一。理论上，“没有 NP 不 VP”“没有 NP 不会 VP”“没有 NP 会不 VP”这些近乎同义的双重否定构式都有可能是 “没有 NP 不会不 VP”的源构式，但是考虑到“没有 NP 不VP”在语料库中的使用频次 是最高的（详见表 2、表 329），我们会更倾向于将“没有 NP 不VP”看作是源构式之一。 其余的一个源构式“NP 不会不 VP”是确定的，没有其他构式与其形成竞争关系。 27 该图的绘制受 De Smet（2013）的启发。 28 在现代汉语中，“并非无 NP”与“并非没有 NP”都有使用语例，如“社会主义有市场，资本主义也 并非无计划”“但这种模式并非无后顾之忧”“美国经济向借贷消费增长模式的转型并非没有争议”。 29 语料来源于北京语言大学 BCC 语料库，统计时已经剔除掉非目标语料，如表示让步条件关系的“没 有爱情不会死人的”，因为这不是我们要研究的“没有 NP 不会 VP”。此外，这些数据反映的语料是下 载量限制条件为 1000 时的使用频次倾向性。 没有女孩子不会不喜欢他 129 表2 以“没有 NP”为首的三种双重否定表达式的出现频次30 表达式 文学 报刊 多领域 微博 科技 篇章 没有 NP 不VP 112 291 470 400 150 0 没有 NP 会不 VP 6 6 80 55 9 0 没有 NP 不会 VP 0 3 16 17 0 0 表3 以“没有人”为首的三种双重否定表达式的出现频次 表达式 文学 报刊 多领域 微博 科技 篇章 没有人不 VP 75 174 467 262 102 0 没有人会不 VP 3 1 55 47 5 0 没有人不会 VP 0 2 14 16 0 0 四、三重否定超常构式产出的动因 三重否定超常构式形成的语言基础是存在两种音近、形近、义近的双重否定构式，除 此之外，它们的形成还受到如下交际、认知以及言语产出方面的动因驱使： 4.1 传递极性义和强烈情感 重复是一种增强和凸显语义强度的手段（Jespersen，1924/2010；Martin，2008）。 Jespersen（1924/2010：518-519）曾指出强化语义和情感的动因对于反复使用否定的重要 影响： “在强烈的感情影响下，说话者想要绝对保证否定的意义得到充分理解，因此，他不 仅把否定词加在动词上，而且加在句中任何易于构成否定的成分上：事实上他给整个句 子蒙上了一层否定色彩而不是把否定局限在一个地方。” 系统功能语言学家 Martin（2008）在论述人际意义时，也表达了与 Jespersen（1924/2010） 类似的看法，他认为否定协和的语义效果是将极性（polarity）意义传递到它所管辖的整 个小句上，例如（29）中反复出现的“no”就起到了这样的作用。 (29) If you don’t get no publicity you don’t get no people at the fight, ...If you don’t get no bums on seats you don’t get paid . . . Anyway I enjoy it. (Martin 2008) 相较于常规的双重否定构式，三重否定超常构式通过将两个双重否定整合，也具有 在整个句子中加强传递双重否定所表达的主观肯定极性义的效果。 4.2 音形义相近构式的联想和激活扩散 不少关于词语或句法整合（syntactic blends）的研究发现，语音或概念相似的词语或 句法会在使用者的大脑中形成联想关系（Carl，1998；Baars et al.，1980；Coppock，2010）， 或者用认知心理实验的描述就是，当一个词语出现在被试大脑中时，另一个语音或概念 30 数据统计时间为 2019 年。 136 8 Empirical Foundations of Sound Correspondence and Probabilistic Evaluation WANG Feng Center for Chinese Linguistics, Peking University Abstract: Sound correspondence is the cornerstone of historical comparative linguistics. Determining sound correspondence involves distinguishing between historical divergence, language contact, and accidental resemblance. From a probabilistic perspective, identifying sound correspondences must first account for the interference of chance similarities. By reviewing previous research of tonal reconstruction of Proto-Miao-Yao based on experience and incorporating the probabilistic calculation of tone comparison across Yao languages, this paper constructs a probability-based algorithm grounded in actual distributions to assist in evaluating accidental correspondences, thereby strengthening the probabilistic foundation for sound correspondence identification. Some mathematic tools or convenient software are expected to develop to evaluate the relevant probability of sound correspondence under the guide of historical experience.  1. Introduction It is generally agreed that the speech of William Jones (1786) is the beginning of historical comparative linguistics. He said: “The Sanskrit language, whatever be its antiquity, is of a wonderful structure; more perfect than the Greek, more copious than the Latin, and more exquisitely refined than either, yet bearing to both of them a stronger affinity, both in the roots of verbs and in the forms of grammar, than could possibly have been produced by accident; so strong indeed, that no philologer could examine them all three, without believing them to have sprung from some common source, which, perhaps, no longer exists: there is a similar reason, though not quite so forcible, for supposing that both the Gothick and Celtick, though blended with a very different idiom, had the same origin with the Sanskrit, and the old Persian might be added to the same family…[Emphasis added]”. (Cannon 1991:31) However, the claim of “some common source” or “the same origin’’ for Sanskrit, Greek, Latin, Gothick, Celtick, and Persian, Wang, Feng. 2025. Empirical Foundations of Sound Correspondence and Probabilistic Evaluation. In Chenglong Huang, Xu Yang (eds.) : The Creation of Meaning: A Festschrift in Honour of Randy J. LaPolla, pp. 136–143.  137 is purely based on William Jones’ experience or belief, which helped him to exclude the possibility of accidence. After more than 100 years, the arbitrariness of linguistic signs was first put forward by Ferdinand de Saussure (1857—1913) in his course of general linguistics32, that is, there is no direct connection between the signal and its meaning. Thanks to this presupposition, Antonine Meillet (1866-1936) explicated the comparative method in his cornerstone book La méthode comparative en linguistique historique (The Comparative Method in Historical Linguistics). Otherwise, if a sound has a nature connect with its meaning, all languages would be related. In this book, numerals in languages like French, Italian, Spanish and etc., are listed as examples to show the non-arbitrariness across these languages. The key point is to find the so-called sound correspondence between these languages. For instance, the pattern of ui (French huit)-o (Italian otto)-o (Spanish ocho) detected in the numeral eight is repeated in some other words like night, cook, milk, and fact. According to Meillet, Such pattern could not be due to accidence, and suggests the historical connection between languages. The reason why sound correspondence could not be resulted in by chance has not been discussed in Meillet (1926). The systematic examination and the followed procedure of reconstruction seems to support such judgement. The complex inflections in Indo-European languages were taken as strong evidence for their common ancestral language, whereas lack of such morphological phenomena in Asian languages like Chinese, Vietnamese, and etc. may be a huge handicap for the reconstruction of their protolanguages. In spite of Meillet’s alarm, the Sino-Tibetan hypothesis has been proposed. The detailed history may be refereed to van Driem (2001). The most influential proposal in China could be attributed to Li (1937). The late professor Li Fang-Kuei put Chinese, Kam-Tai, Miao-Yao and Tibeto-Burman into a language family named Indo-Chinese. This paper was reprinted in the inaugural issue of Journal of Chinese Linguistics with only some minor modifications in 1973. The editor added the note: “A condensed version of this article first appeared in Chinese Year book, Shanghai,1937. Since that time, it has been regarded as the standard reference on this subject. Although research over the past three decades has brought us a more refined understanding of the individual dialects, Li’s broad outline remains essentially accurate and useful.” Three characteristics, namely, tendency towards monosyllabism, tendency to develop a tonal system, and unvoicing of the original voiced initial consonants, are taken as major evidence, aside from “many items in the lexicon which are shared in common in this family”. However, this argumentation was criticized as these phonetic characteristics should be seen as typological similarity rather than genetic resemblance. 32 The content was compiled a famous book, Cours de linguistique générale (Course in general linguistics), and published in 1916. 138 Interestingly, Baxter (2002) examined the history of the comparative method, and found that “the technique of using regular phonological correspondences among genetically related languages to reconstruct aspects of the phonology, lexicon, and morphology of their common ancestor” as the standard method of historical linguistics is quite recent. Following Nichols (1996), LaPolla (2012) agreed to consider the evidence employed for genetic relatedness probabilistically, which would better to be “whole systems or subsystems with a good deal of internal paradigmaticity, ideally multiple paradigmaticity, and involving not only categories but particular shared markers for them” (Nichols 1996:48). As it was well known, most Sino-Tibetan languages are monosyllabic morphology-poor, so that such standard is very difficult, though not impossible. LaPolla (2012) suggests to look for “particular elements (words and/or morphological markers) to be organized into paradigm-like sets and applied rigorously to determine relatedness. That is, the sets are treated as if they were paradigms and used as individual-identifying evidence of relatedness, because the particular combination of independent elements as an internally-structured set would give us the level of statistical significance we need.” Notably, such proposal is to highlight the importance of probabilistic evaluation, however, the actual probabilistic calculation has not been conducted. Maybe, it is still experience-based. 2. Probabilistic calculation of sound correspondence Looking into the term of sound correspondence again. It may be illustrated as follows: If one set of words shared one sound X in language a, at the same time, they shared one sound Y in language b, then “X” and “Y” are sound correspondences between language a and language b. In fact, sound correspondence indicates a recurrent pattern, not necessarily with phonetic similarity. Thus, at least two examples are needed to support a sound correspondence. Under the condition of given words (indicating a constraint of particular meaning), sound correspondence suggests a non-arbitrary relation between sound and meaning of these words, and it must be due to some reasons. Excluding onomatopoetic words and a few words with sound symbolism, the reason can be attributed to language relatedness. However, chance resemblance may cause the problem. For instance, is the sound correspondence sp- <> phbetween English and Chinese in the following table due to genetical relatedness or chance resemblance? (Examples are quoted from Prof. Chen Baoya’s course.) English Chinese split 劈[phi] spill 泼[pho] spread 铺[phu] 139 Here, maybe three examples are not enough. “Are 500 enough, then? And if not, are 501 sufficient? Nobody can give a satisfactory answer to these questions. And this is no doubt the reason that linguistics may disagree over whether a particular proposed genetic relationship is sufficiently supported or not.” (Hock and Joseph 1996:493) The problem cries out for a probabilistic evaluation. A null hypothesis should be made before the probabilistic calculation. In this case, it would be postulated that sound correspondence between two languages without genetical relationship could be generated. As far as we know, the number of phonemes in a particular language is quite limited. For Mandarin Chinese, a syllable can be divided into initial, final, and tones. There are 22 initials, 35 finals and 4 tones, but actual syllables in use are 1298 at most. (Su and Lin, 2006) Therefore, if the number of morphemes or words in the comparative work are more than the number of syllables, the chance resemblance would be unavoidable. However, the probability of chance resemblance could be evaluated, and the threshold for the sound correspondence not due to chance could be calculated then. 3. A case study of Miao-Yao tonal correspondences Previous studies on the tonal correspondences between Miao-Yao languages are illuminating. The eight tones in Proto-Miao-Yao were reconstructed in Chang (1947). Records of 6 Miao-Yao dialects were selected as the basis, and 8 sets of tonal correspondences were detected. Each set was supported by several examples, three at least or nine at most, as follows. Tone examples Notes 1 water, two, three, five, four, deep, snake, monkey Lack ‘monkey’ in Yongcong 2 flower, leaf, door, nine 3 full, fruit, house, tail, long, dog, road, bone, blood 4 fish, fire, village Lack ‘village’ in Yongcong 5 name, moon, iron 6 die, eye, rain, bird Lack ‘rain’ ‘bird’ in Yongcong 7 wing, smile, to drink 8 see, tongue, bean, ten, eight It is notable that several words (called characters in this paper) do not fit any of these 8 sets, though they look quite similar across these dialects as follows. (Chang 1947:101) 140 Since each tonal pattern of them could not be supported by one more example, these examples were regarded as exceptions. Another way to view this reconstruction based on experience is to check whether a priori assumption may have made. As is well known, the four tones in Middle Chinese split into eight based on voicing of the initials. Based on the eight tones and their association with voiced or voiceless initials, the earlier four tones were further reconstructed for Proto-Miao-Yao. In other words, the develop of tonal system of Miao-Yao languages seems to be a parallel of Chinese. Anyway, as LaPolla (2012:120) suggested: “The solution to these epistemological problems is to be aware of the principle of interpretation and be aware of how assumptions can influence our interpretations, and try to work as inductively as possible, and make explicit arguments for all generalizations and assumptions.” We did an experiment of Miao-Yao tonal comparison. From a vocabulary collection of Miao-Yao languages, which was edited by the research group of Miao-Yao languages in Minzu University of China, a (Chuan-qian-dian) Miao language and a Yao (Mien) language were selected. Their morphemes are not equivalent. Among 1443 Miao morphemes and 1593 Yao morphemes, the distribution of their tones is listed as follows. Tone Miao Yao 1 251 335 2 142 252 3 194 214 4 130 85 5 256 276 6 186 169 7 150 176 8 134 86 To evaluate the probability of a set of tonal match under such an uneven distribution, an algorithm has been proposed in Wang and Zhang (2021). Once the collection of morphemes of two languages under comparison is given, it is possible to quickly calculate the upper bound of a set of matched random distributions under a given confidence level using a Matlab program. 141 The probability principles and procedures are as follows: Assume there are N matches in total, the two languages have m and n Tone 1, respectively, so if these matches are randomly scrambled, the resulting number X of 1<>1 is a random variable following the hypergeometric distribution H(m,n,N), whose probability mass function is described in the following formula: P󰇛Xk 󰇜   C  A software application can be employed to derive a unidirectional confidence interval for the aforementioned hypergeometric distribution, which can then be compared with the figure corresponding to a 1<>1 correspondence obtained via semantic match. Correspondence Examples 95% 99% 99.9% 1 1<>1 109 66 71 75 2 3<>3 61 35 38 42 3 7<>7 34 22 24 28 4 5<>5 72 56 60 64 5 2<>2 53 28 31 34 6 4<>4 20 12 15 17 7 8<>8 20 12 14 16 8 8<>6 32 22 25 28 9 6<>6 30 27 30 33 With different confidence levels, the tone matching that reaches the upper limit fluctuates. The first eight matches are robust. However, set 9 (6<>6) exceeds the upper limit of 95%, equals the upper limit of 99%, and falls below the upper limit of 99.9%. The eight-tone system of MiaoYao languages based on sound corresponds has been recognized as an unchangeable theory since it was first established by Chang Kun (1947). According to the table above, it can be seen that the 6<> 6-tone correspondence has developed to the edge of randomization with the increase of the separation time of the Miao language and the Yao language. According to the development of this speed, the earliest randomization and the phonetic correspondence that cannot be determined by probability calculation is this correspondence. From this point of view, the matching of 8<>6 is especially worthy of attention, because it does not conform to the traditional eight-tone system of Miao-Yao. However, this problem is not difficult to explain, because the nine sets of matches are calculated at a 95% confidence level and are non-random matches, that is, they can be regarded as sound correspondences in historical linguistics, and thus may be caused by genetic differentiation or language contact. Instead of explaining the correspondence of 8<>6 from the perspective of language split, it can be explained from the perspective of late 142 language contact. It does not have to be the contact between the two languages after split, but more likely it is the contact between the two languages after split with the other language, for example, Chinese. Nevertheless, in any case, given the same probability evidence, the match of 8<>6 should be treated like any other sound correspondence. If we adhere to the most rigorous probabilistic assessment (excluding random matches), with 99.9% as the standard, we must pay some price that some sound correspondences (e.g., tone correspondence 6<>6) due to language split are not recognized. In this sense, strict requirements of sound correspondence can ensure the stable foundation of reconstruction to a greater extent, but at the cost of recognition. The significant probability can only show that the formation of the correspondence is not accidental, and the cause may be genetic relationship, contact, or a mixture of the two. Therefore, other criteria are needed to distinguish the properties of these non-random corresponding results. 4. Conclusion From the perspective of language evolution, if two related languages diverge from a single ancestor, their earlier phonological matches and supporting examples gradually show a random matching state with the respective replacement of common words. When the two languages differentiate for a long time, the non-random state of phonetic correspondence will eventually collapse. In this spirt, rethinking the position of sound correspondence in historical comparison, it is indeed the fundamental basis of any other steps. The question is how to define it and how to make use it properly. The experience has been accumulated in the comparisons of Indo-European languages and other languages, through about two hundred years. Based on this experience, some mathematic tools or convenient software are expected to develop to evaluate the relevant probability of sound correspondence. With the coming of AI era, this hope may come true in the near future. Acknowledgement: This research was funded by a grant from Ministry of Education of China (No. 22JJD740002). References 苏新春、林进展 2006 普通话音节数及载字量的统计分析——基于《现代汉语词典》注音 材料[J]. 中国语文, 2006(3):274-284. 汪锋、张原 2021 一种评估偶然对应的概率算法—以瑶语声调比较为例. 载汪锋、宋作艳、 143 李子鹤（主编）《理论照亮材料》.昆明：云南大学出版社：21-29. 张琨 1947 苗瑶语声调问题.《中央研究院历史语言研究所集刊》16:93-110. 中央民族学院苗瑶语研究室 编 1987《苗瑶语方言词汇集》. 北京：中央民族学院出版社. Baxter, W. H. 2002. ‘Where does the ‘comparative method’ come from?’, in Cavoto, F. (ed.), The linguist’s linguist: A collection of papers in honour of Alexis Manaster Ramer. Munich: Lincom Europa, pp. 33–52. Cannon, G. 1991. Jones’s “Sprung from Some Common Source”: 1786-1986. Pp.23-47 in Sprung from Some Common Source: Investigations into the Prehistory of Language. S.M.Lamb and E.D.Mitchell, eds. Stanford University Press. van Driem, George. 2001. Languages of the Himalayas: an ethnolinguistic handbook. Leiden; New York; Koln: Brill. Hock, H.H. & B.D. Joseph, 1996. Language History, Language Change, and Language Relationship: an Introduction to Historical and Comparative Linguistics. Berlin: Mouton de Gruyter. LaPolla, Randy. 2012. Comments on methodology and evidence in Sino-Tibetan comparative linguistics. Language and Linguistics 13.1:117-132. Li, Fang-Kuei. 1937. Languages and dialects, in The Chinese Year Book, Shanghai. Reprinted in Journal of Chinese Linguistics, 1973.1:1-13. Meillet, Antonine.1925. La méthode comparative en linguistique historique. Oslo: H. Aschehoug & Co. (Reprint, Paris: Honoré Champion, 1966). Nichols, Johanna. 1996. The comparative method as heuristic. The Comparative Method Reviewed: Regularity and Irregularity in Language Change, ed. by Mark Durie & Malcolm D. Ross, 39-71. New York: Oxford University Press. Ringe, D. 1992. On Calculating the Factor of Chance in Language Comparison. Transactions of the American Philosophical Society, vol. 82, no. 1. Philadelphia: American Philosophical Society. 144 9 类型学框架下的彝语诺苏话关系从句 胡素华 中央民族大学 内容摘要 关系从句（relative clause）是绝大多数语言共有的句法结构，即 用关系化手段提取名词性成分，这么做是因为主句中名词成分和从句中的 名词是一种替代关系（substitution），即关系化或同指。彝语的关系从句尚 未进行过全面、系统地研究。本文从以下几个方面对彝语的关系从句进行描 写：有无核心名词、核心名词的位置、关系化的策略、关系从句性质及标记、 关系从句的名词短语可及性序列、关系从句与其他句法范畴之间关系。研究 表明彝语的关系从句包括有核和无核两种类型，有核类型中关系从句都是 外置于核心名词的，有关系从句后置型（NRel）和前置型（RelN）两种语 序，其中后置型是主要的，前置型是受汉语影响在书面语长句中新发展出来 的类型，早期表示处所和时间的前置型关系从句已演变为小句名物化结构， 另一种无标记的前置型则发展为固定搭配、不可扩展的词组。从关系从句的 性质看，有限定性关系从句是主要的，非限定性关系从句比较少和不典型。 关系从句标记有用名物化标记 su³³（包括量词与 su³³结合的短语）、量 词（ 包 括指示词与量词结合的短语）两种类型。彝语名词短语可及性中主语、直接 宾语、间接宾语、旁格论元、领有者、比较参项（比较的主语）、比较基准 （比较的宾语）都可以被关系化。关系化有留空、宾格代词复指、保留领属 代词、介词悬空四种策略。 关键词 彝语；关系从句；关系化策略；名物化；名词短语可及性 一、引言 在跨语言视野中，关系从句（也称为关系小句或关系子句）（relative clause）是绝大 多数语言共有的句法结构，即使用关系化手段提取名词性成分，这么做是因为主句中名 词成分和从句中的名词是一种替代关系（substitution），即关系化或同指。关系子句就是 子句所修饰的中心名词在子句中也有一个句法位置，也称之共元（common argument）。 胡素华. 2025. 类型学框架下的彝语诺苏话关系从句. 黄成龙、杨旭（主编）《意义创造论：罗仁地先生 七秩寿庆文集》，pp. 144–158.  145 Comrie（1981）“一个关系小句必定包含一个中心名词和一个限制性小句。中心名词本身 的所指对象有某个潜在的范围，而限制性小句用一个命题来限制这个范围”。关系从句和 补足语从句与主句间的关系最紧密，其间语调无间隙，属于嵌入式子句（embedded clauses）。 在关系从句结构特点的认识上，基本达成共识（Dixon）（2010）：1.由一个主句和一 个子句构成一个句子，这个句子只有一个语调。2.主句和子句共享同一个论元，即共元 （common argument），它既是主句的论元，也是子句的论元。在具体的语言中既可能两个 小句中都出现共元，也可能只出现在一个小句中，或是在两个小句中都不出现。在子句中 常常由代词来指代。3.在句法上，关系子句作为主句共元的修饰成分；在语义层面上，关 系子句可以是提供相关信息来表示共元的聚焦（focusing）-或是限制（restricting）-或是 指称（reference）；也可以是，当共元为人称代词或专有名称时，子句提供与此相关的背 景信息。前者为限定关系子句(restrictive RC)，后者为非限定关系子句（non-restrictive RC）。 4.关系子句必须具有一个基本小句的结构——谓语和谓语的核心论元（core argument）。 在有些语言中，RC 可以包括外围论元（peripheral argument），有些语言却不可以包括。 一般不能用主句的语法范畴——如时、体、情态等来标记。 类型学的研究发现，不同语言中关系子句的差异主要表现在如下这些参数上（孙天 心2006）：1.关系子句结构本身（是否限定，名物化与否）；2.中心词与关系子句之间的相 对位置，关系子句在中心词之内还是之外，若在之外是前置还是后置；3.名词关系化手段 （strategies of revitalization），是留空还是保留完整的名词，或是用关系代词或复指代词； 4.哪些论元可被关系化，即名词短语可及性等级序列；5.关系子句中动词的形态（即名物 化成分）所属的范畴；6.关系子句本身差异所属的范畴。 彝语诺苏话（圣乍音）属于彝语北部方言，分布在四川、云南北部，他们自称“诺苏”， 是彝语保留得最好和使用人口最多的方言区，其关系从句的个性和关系化策略却有着其 独特点，尚未在类型学框架下系统地描写过的彝语诺苏话（圣乍音）（下文简称彝语）的 关系从句。 二、彝语关系从句的类型 可以从几个不同的角度对彝语关系从句进行分类：第一，从有无核心名词（head noun） 来看可以分为有核关系从句（headed relative clause）和无核关系从句（headless relative clause）。第二，从核心名词和关系从句的位置来看，有关系从句后置型（NRel）和有关系 从句前置型（RelN），其中关系从句后置型是主要的。第三，从关系从句的性质看，可以 分为限定性关系从句和非限定性关系从句，前者是基本的。 2.1 有核关系从句和无核关系从句 彝语以有核心名词的关系从句为主，也有不少无核关系从句。 1、 有核关系从句 248 在如下这些情况才会强制出现： 4.1.1 受事的生命度高于或等于施事的生命度 当受事的生命度高于或等于施事的生命度时，萝卜寨羌语须在受事之后加标记-to55 以区别施事和受事。例句（15）受事 qɑ³¹“我”的生命度比施事 khu³¹“狗”的生命度高； 例句（16）受事 nә³¹“你”的生命度和施事 qɑ³¹“我”的生命度相同：它们的受事均须加 受事标记-to55。 （15） khu³¹-sә³¹ qɑ³¹-to55 ʑe³¹-dʐe³¹ jɑ³¹. 狗-施事 我-非施事 向里-咬 语气 狗咬了我。 （16） qɑ³¹（-sә³¹） nә³¹-to55 me 55liɑ55 lo³¹ mɑ³¹-ɕou55-sɑu³¹. 我（-施事） 你-非施事 找 来 否定-敢-实然:1单:亲见 我不敢来找你。 4.2.2 突显受事 在萝卜寨话里，非施事者标记-to55 除了具有区分施事者与非施事者以消除歧义的功 能，还具有语用上的对比或突显的功能，如例句（17）和（18）的受事 ʁu55“碗”是句子 的焦点，其后加非施事者标记-to55 起到突显受事的功能。 （17） sɑ 55-sә³¹ ʁu55-to55 sә³¹-khʐe³¹ nә³¹？ 谁-施事 碗-非施事 向外-打碎 疑问 谁把碗打碎了？ （18） thә³¹-sә³¹ ʁu55-to55 sә³¹-dʐәi³5. 他-施事 碗-非施事 向外-打碎 他打碎了碗。 4.2 配置类型 从跨语言角度来看，一种语言可以有不止一种双及物结构，不同语言双及物结构在 表达形式上有很大的差异，Malchukov，Haspelmath & Comrie（2010）指出，世界语言双 及物结构式有三种配置类型：（T-P）-R 式、（R-P）-T 式和 T-P-R 式。在不同的配置类型 中，与事和客体的语序及其跟基本语序的关系又是双及物结构关注的焦点。（陆丙甫、罗 天华 2009；Heine & König 2010） 萝卜寨羌语跟彝语西部方言腊罗话（胡素华、周廷升 2018）一样，受事格标记-to55 的使用是有条件的，受施受事生命度的制约。在单及物动词句中，受事 P的生命度高于 或等于施事者的生命度并有施事能力时，要用受事格标记以区别于施事者，而当受事 P的 生命度低于施事者的生命度或没有施事能力时不用标记，只是信息结构上受事成分为焦 点时，受事格标记-to55 兼做焦点标记才强制使用。在萝卜寨羌语“给”字结构中，T和R 的顺序相对灵活，主要受两个名词性论元是否带标记的影响。在格标记结构中，“给”字 结构以 R-T 优势语序，接受者 R论元由-to55 标记，客体 T论元无标记；当“给”字结构 以T-R 出现时，接受者 R论元和客体 T论元均须带标记-to55。因此萝卜寨羌语根据单及 物句受事论元 P的特点和 RT 语序分裂为四种组合类型。（1）在 以 R-T 为语序的“给”字 249 结构中，①当 P为非生命体时，其编码同双及物句中的 T，即（T-P）-R 式，属于双及物 类型中的间接宾语组配型；②当 P为生命体时用-to55 标记，其编码同双及物句中的 R， 即（R-P）-T 式，属于次要宾语组配型。（2）在以 T-R 为语序的“给”字结构中，①当 P 为非生命体时，其编码有别于双及物句中的 R和T，即（R-T）-P 式；②当 P为生命体时 用-to55 标记，其编码同双及物句中的 R和T，即 T-P-R 式，属于双及物类型中的中和配 置型。而连动策略“给”字结构的配置有所不同，不论其语序为 R-T 还是 T-R，连动策略 “给”字结构的接受者 R论元要带标记-to55，客体 T论元不带标记。故连动策略“给” 字结构的配置跟以 R-T 为语序的“给”字结构相同，根据单及物句受事论元 P的特点分 类为（T-P）-R 式和（R-P）-T 式。 五、结语 Machukov et al.（2007）的统计结果显示，在不同类型语言的双及物结构中，T和R 的顺序主要受两个名词性论元是否带标记的影响。萝卜寨羌语是一个以 SV/APV 为基本 语序的语言，其双及物结构（不论是无标记的双宾构式，还是附属语标记结构）以 R-T 为 优势语序，这跟陆丙甫、罗天华（2009）和柳俊（2019）的关于东南亚语言区域优势语序 的统计结果一致，即动词居尾的语言以 R-T 为优势语序。 在萝卜寨羌语中，格标记结构是“给”字结构最重要的编码形式。这符合 Greenberg （1966：96）的 U41，即 SOV 语言倾向于有格系统。萝卜寨羌语的格标记是语义基础上 的标记，其功能主要是为了辨别其宿主的语义关系。在单及物句中，受事者标记-to55 常用 来标记有生命的受事论元，一般不标记无生命的受事论元。因为格标记的存在，萝卜寨 “给”字结构 R和T的语序相对灵活。在“给”字结构中，以 R-T 为优势语序的格标记 结构经常用-to55 标记 R论元，T论元无标记。因此萝卜寨羌语根据单及物句受事论元 P 的特点分裂为两种类型：当 P为非生命体时，其编码同双及物句中的 T，即（T-P）-R 式， 属于双及物类型中的间接宾语组配型；当 P为生命体时用-to55，其编码同双及物句中的 R，即（R-P）-T 式，属于次要宾语组配型。在“给”字结构中，T论元也可以出现在 R 论元之前，此时以 T-R 为语序的格标记结构中的 T论元和 R论元均须带受事者标记-to55。 格标记结构以凸显 R论元的 hR-TV 为优势结构，而 R、T均凸显的 hT-hRV 结构则表达 强烈的处置义。 双及物动词“给”通过添加致使标记-ʂә³¹而构成双及物动词的致使结构，其典型语 序为 NP1+NP2+NP3+NP4+VP，其中 NP1为使役双及物动词的致使者，在句中充当施事论 元A，可带施事者标记；NP2为使役双及物动词的被致使者，在语义上其仍然是致使者的 受事P，同时也是双及物动词ɦɑ³¹-ʐdɑ55“给”的施事者，其后要强制带非施事者标记-to⁵⁵； NP3作客体论元 T，其后的非施事者标记是非强制使用的；NP4为接受者 R论元，其后要 强制带非施事者标记-to55。萝卜寨羌语双及物动词“给”已经发展出连动式双及物结构， 250 其语序以 T-R 为主，R论元后要强制带非施事者标记-to55，而 T论元后的非施事者标记 是非强制使用的。 参考文献 胡素华、赵镜：《彝语诺苏话双及物及其他三个论元结构》，《语言暨语言学》，2019 年 第3期。 胡素华、周廷升：《彝语方言受事格标记及基本语序类型比较》，《语言科学》，2018 年第 2期. 黄布凡、周发成：《羌语研究》，四川人民出版社，2006年。 黄成龙：《蒲溪羌语研究》，民族出版社，2007年。 黄成龙：《羌语的非施事者及其相关标记》，《语言学论丛》第 41 辑，商务印书馆，2010 年。 黄成龙：《羌语的施事者及其相关标记》，《语言暨语言学》，2010年第2期。 刘丹青：《汉语给予类双及物结构的类型学考察》，《中国语文》，2001 年第 5期。 刘光坤：《麻窝羌语研究》，四川民族出版社，1998 年。 柳俊：《东南亚语言区域的双及物结构类型及其成因》，《外国语》，2019 年第 4期。 陆丙甫、罗天华：《中国境内语言的双及物结构语序》，《汉藏语学报》第 3期，商务印 书馆，2009 年。 孙宏开：《羌语简志》，民族出版社，1981 年。 王保锋：《萝卜寨羌语语法研究》，中央民族大学博士学位论文，2017 年。 Haspelmath, M. 2015. The serial verb construction: Comparative concept and cross-linguistic generalizations. Retrieved from https://www.academia.edu/10652772/The_serial_verb_construction_Comparative_concept_a Heine, Bernd & Christa König. 2010. On the linear order of ditransitive objects. Language Sciences 32:87-131. LaPolla, Randy J. 1992. Anti-ergative marking in Tibeto-Burman. Linguistics of the TibetoBurman Area 15(1): 1-9. LaPolla, Randy J. 2004. On nominal relational morphology in Tibeto-Burman. 林英津等编《汉 藏语研究——龚煌城先生七秩寿庆论文集》，43-74。台北：中研院语言所。 LaPolla, Randy J. with Huang Chenglong. 2003. A Grammar of Qiang, with Annotated Texts and Glossary. Berlin: Mouton de Gruyter. Malchukov, Andrej, Haspelmath, Martin, & Comrie, Bernard. 2010. Ditransitive constructions: a typological overview. In Andrej Malchukov, Martin Haspelmath & Bernard Comrie (eds.) Studies in Ditransitive Constructions: A Comparative Handbook, 1-64. Berlin: De Gruyter 251 Mouton. Margetts A , Austin P K 2007. Three-participant events in the languages of the world: towards a crosslinguistic typology. Linguistics. Newman, J. 1993. The semantics of giving in Mandarin : Conceptualizations and Mental Processing in Language. Conceptualizations & Mental Processing in Language. Newman, J. 1996. Give: A Cognitive Linguistic Study. Berlin and New York: Mouton de Gruyter. Newman, J. 1997. The linguistics of Giving. Amsterdam: John Benjamins. Newman, J. 1998. Recipients and 'give' constructions. In van Langendonck W, van Belle W(eds.). Dative : Volume 2: Theoretical and contrastive studies. London: John Benjamins. Silverstein, Michael. 1976. Hierarchy of features and ergativity, in Dixon, R. M. W. eds., Grammatical categories in Australian languages [C], Canberra: Australian Institute of Aboriginal Studies, 1976. Zheng Wuxi（郑武曦） 2016. A Grammar of Longxi Qiang , PhD thesis, National University of Singapore. A typological investigation of the "give" character structure in Luobozhai variety of the Qiang Language Wang Baofeng School of Chinese Ethnic Minority Languages and Literatures, MinZu University of China [Abstract] From the perspective of typology, this article focuses on the types of the “give” character structure in Luobozhai variety of the Qiang Language. This article first describes the argument encoding method and strategy of the “give” character structure in Luobozhai variety of the Qiang Language, and summarizes the configuration types of the “give” character structure. Then, it discusses the word order of different types of “give” structures T and R, as well as their interaction with nominal argument markers. [Keywords] Qiang language “give” character, structural typology, morphological syntax 252 16 南部羌语名词组特征的比较 高 韬 成都大学 文学与新闻传播学院 内容摘要：本文主要描写比较了南部羌语三个方言点名词组的特征。单音 节普通名词所占据的比例其内部一致性比较高，但由于元音高化等因素， 导致了三个方言点普通名词的形式有所差异，如 u 对应 ue。专有名词中地 名的表达一致性较高，人名中姓氏的表达手段比较有特点，如，用羌语的 动物“獐”对应姓氏“张”，用羌语的性状形容词“高”对应姓氏“高”。亲 属称谓名词一部分用本民族语表示，一部分则用汉语借词来表示，内部一 致性比较高。时间名词的表达方式基本相同，其差异性主要表现在部分读 音发生了变化，也能看出具有共同的来源。数词与量词的表达内部一致性 也比较高，差异主要表现在序数词上，龙溪两河口“第一”、“第二”的表示 方法已经消失，而是借用汉语的表达方式。方位词的表达存在一些差异， 其中最典型的差异体现在“左边”和“右边”的表达形式上。 关键词：南部羌语 名词组 比较 一、引言 1990 年代以来，以罗仁地（Randy J. LaPolla）、孙宏开、刘光坤、黄成龙、Jonathan Evans、潘露莉、郑武曦等学者为代表，对羌语进行了深入的描写研究，已有了丰硕的学 术成果。孙宏开（1981：177-178）把羌语分为南部方言和北部方言。南部方言分为五个 土语：大歧山土语、龙溪土语、绵池土语、黑虎土语。当地人对内主要使用本族语羌语交 流，对外则多使用四川方言进行交流。Randy J. LaPolla & Huang Chenglong（2003：5）指 出当地羌族人与非羌族人主要用汉语交流，口语方面主要采用西南官话的川西方言，书 面语则采用现代汉语形式。由于南部羌语与汉语长期接触，这就导致了南部羌语中出现 了很多借用汉语的现象。本文在前人成果的基础上，主要对南部羌语名词词组的特征进 行描写和比较，选取的方言点为阿坝州汶川县萝卜寨、龙溪两河口、绵虒46。在南部羌语 中没有与汉语所描述的“词”完全对应的“词”。黄成龙（2007：46）把蒲溪羌语的词定 义为“新成分的一个最小的自由形式，包括停顿、重音；在正常的语言里词的内部不能插 高韬 2025. 南部羌语名词组特征的比较. 黄成龙、杨旭（主编）《意义创造论：罗仁地先生七秩寿庆文 集》，pp. 252–268.  46 本文语料来源为田野调查所得，其它引用语料均会注明出处。 253 入任何东西。” 孙宏开（1981：59）指出：“羌语的词大部分是单音词和由单音词组成的合成词，多 音节的单纯词比较少。”南部羌语中的词可以是一个音节，如“水”ʦu³¹萝卜寨 ʦu³¹龙溪两河口 ʦue³¹ 绵虒，“雨”meʴ³¹萝卜寨 mei³¹龙溪两河口 me³¹龙溪两河口；也可以是两个音节，如：“太阳”mu³¹ni³¹萝卜寨 mu³¹ni³¹龙溪两河口 me³¹se55 绵虒，“鞋”be³¹ʦu³¹萝卜寨 ʦua³¹χa³¹龙溪两河口 be³¹ʦә³¹绵虒。 除此之外，南部羌语还有一些后粘附词（enclitics），它们主要附着在名词短语后。包 括：格标记、性标记、比较标记、话题标记。它们不能单独出现，必须与名词短语或小句 一起出现，例如：te³¹-kә³¹朝上走萝卜寨 i³¹-ʨʰyәu55-pu³¹戳龙溪两河口 i³¹-qa³¹煮绵虒；“母猪”pia³¹-mia55 萝卜寨 pie55-mia³¹龙溪两河口 pia³¹-mia55 绵虒。本文所描写比较的名词组主要有以下几类：名词、数 词与量词以及方位词，下面分别描写其特征。 二、名词 2.1 普通名词 普通名词是名词中最常见的一个次类。在所调查的 1051 个名词中，萝卜寨的单音节 名词总共 126 个，占 11.99%，龙溪两河口的单音节名词总共 115 个，占 10.94%，绵虒的 单音节名词总共 120 个，占 11.41%。可以看出，相互间的比例一致性比较高。如表 2-1 所 示： 表2-1 南部羌语三个方言点普通名词 例词 萝卜寨 龙溪两河口 绵虒 狗 kʰu³¹ kʰu³¹ kʰue³¹ 水 ʦu³¹ ʦu³¹ ʦue³¹ 鸡 y³¹ y³¹ i³¹ 树 pʰu³¹ pʰu55 pʰo³¹ 从表 2-1 可见，这些普通名词具有很高的一致性，因为普通名词属于基本词汇，基本 词通常很稳定，不易发生变化。 当然，即使是基本词，也会发生变化，如龙溪两河口的“天”mu³¹to55 就与萝卜寨和 绵虒存在差别。但如果是语言内部的变化，一般是可以解释的。如“水”这一词，在萝卜 寨和龙溪两河口读为 ʦu³¹，但在绵虒读为 ʦue³¹，单元音发生了双元音化，而且还能找到 更多 u→ue 发生裂化的例证，具体对比如下： 254 表2-2 南部羌语 u→ue 裂化情况 例词 萝卜寨 龙溪两河口 绵虒 水 ʦu³¹ ʦu³¹ ʦue³¹ 狗 kʰu³¹ kʰu³¹ kʰue³¹ 烟子煤~ mu 55kʰu55 mu 55kʰu55 mu³¹kʰue 55 铁锈 ɕi³¹ɡu³¹ ——47 ɕe³¹ɡue³¹ 斑鸠 ku³¹ku³¹ —— kue³¹kue³¹ 真~假 ʁu³¹ —— ʁue³¹ 单元音变为复元音的过程被称为元音的裂化。根据朱晓农48（2004）裂化分为：前显 高化裂化和后显低化裂化。南部羌语中 u→ue 属于后显低化裂化，这一变化是由“低化” 造成的。在汉语方言中存在类似的变化，y＞yә/yә/ye（朱晓农 2004: 447）。元音 u是一个 高元音，高元音在变化中会走上一条“高顶出位”的道路，也就是说 u作为最高的一个元 音，还在继续发生高化。从下面的图我们可以看出，u首先是走上继续高化的道路，继而 发生了高顶出位。 图2.1 元音 u的高化 2.2 专有名词 专有名词主要分为两类：地名和人名。在专有名词部分，本文主要比较一些常见的 词，侧重描写三个方言点之间的差异性。 2.2.1 地名 南部羌语中用羌语表达的地名，大多是附近地区的地名，如果是近期影响进入的地 名，则直接借用汉语来表达，如“上海”ʂaŋ²¹³xai5³等不属于当地或邻近当地的地名。表 23是常用地名读音的对照。 47 “——”这一符号表示没有本族语羌语的说法，或者只能说汉语，如有需要会标出汉语的说法，否 则不标。后文同。 48 朱晓农. 2004（5）.汉语元音的高顶出位[J] 中国语文. 255 表2-3 南部羌语常用地名读音对照表 例词 萝卜寨 龙溪两河口 绵虒 威州 kʰua 55ʦa55 kʰo 55ʦo³¹ ʦʰu55ʦʰa³¹ 成都 i³¹to55ko55 i³¹to³¹ tie³¹da55 都江堰 do³¹ko55to55 lo³¹ko55to³¹ ta³¹ku³¹ 茂县 moŋ²¹³ɕiã²¹³（借汉） ke³¹niәu55 ke³¹nio55 松潘 soŋ 55pʰã55（借汉） ʐә³¹tʂә55 soŋ 55pʰan55（借汉） 理县 li 5³ɕiã²¹³（借汉） po³¹ʦu55 li 5³ɕian²¹³（借汉） 龙溪 loŋ³¹tɕʰi55（借汉） ŋe³¹mo³¹ʨe 55 loŋ³¹tɕʰi55（借汉） 绵虒 miã³¹sә55（借汉） kʰa³¹su55 mian³¹sә55（借汉） 索桥村 so³¹tɕʰiau³¹（借汉） kʰa³¹pia³¹χe55 so³¹tɕʰiau³¹（借汉） 2.2.2 人名 南部羌语中人名表达的手段比较特殊，主要体现在姓氏的表达。有一部分姓氏的表 达是直接借用汉语，例如“王”直接用汉语的读音来表示；还有一部分姓氏的表达是采用 与汉语中的姓氏词义相同的羌语词汇来表达。其中的对应关系是汉语的某一姓氏等同于 羌语中的某一类表达，这一类可以是动物，可以是无生命的事物等，甚至可以是描述性状 的词。如： 表2-4 南部羌语常用姓氏对照表 例词 萝卜寨 龙溪两河口 绵虒 姓张 tʂʰә³¹li³¹mo³¹uei³¹ —— —— 姓杨 ʨʰe³¹mo³¹uei³¹ —— —— 姓高 —— bo55mo³¹ua³¹ bu55mu³¹ue³¹mu55 姓何 —— ʦu³¹kʰua³¹mo³¹ua³¹ ʦue³¹ue³¹mu55 萝卜寨方言中，“姓张”一词中的“tʂʰә³¹li³¹”是“獐子”的意思，因为“獐子”的“獐” 和汉语的“张”是同音词；“姓杨”中的 ʨʰe³¹是“羊”的意思，与汉语中的“杨”同音。 龙溪两河口中“姓高”的 bo55 是羌语中的性状形容词“高”，借用来表示姓氏的“高”； “姓何”中的“ʦu³¹”在羌语中是“河水”的意思，借用来表示姓氏的“何”。绵虒羌语 “姓高”中的“bu55”在当地羌语中是“高”的意思，与汉语姓氏的“高”对应，“姓何” 中的“ʦu³¹”是“河水”的意义，与汉语表姓氏的“何”对应。 256 2.3 亲属称谓名词 亲属称谓名词有一部分用本民族词表示，有一部分则用汉语借词来表示，例如： 表2-5 南部羌语常用称谓词对照 例词 萝卜寨 龙溪两河口 绵虒 父亲 pa³¹pa55 bi 55 ʁue³¹ia55 母亲 a 55ma55 mia³¹ ʁue³¹mia55 父母 pa³¹pa³¹a³¹ma55 bi 55mia³¹ ʁue³¹ia55ʁue³¹mia55 儿子 ʨi³¹ ʨi³¹/ŋa²¹³（少说） ʦә³¹ 女儿 ʨe³¹ ʨe55 ʨe³¹ 儿女 ʨi³¹nia³¹ʨe³¹ ʨi³¹na³¹ʨe55 ʦә³¹nia55ʨe³¹ 哥哥 tu³¹ba55 tu 55pu³¹ ʁue³¹po55 弟弟 tu³¹ʦu55 tu³¹ʦu55 tәu³¹ʦu55 弟兄 tu³¹ʦu55tu³¹pa³¹ tu³¹ni³¹ a³¹tu³¹nә55 姐姐 io³¹pa55 sa 55pa³¹/ʁua³¹ʦi55 a55ʨe³¹（面称） ʁue³¹ʨe55 妹妹 io³¹tʂa³¹ io55/sa55ʦu³¹ a55me³¹（面称） liәu³¹ 姊妹 mu³¹nia³¹io³¹ sa³¹ni³¹ a³¹sa³¹nә55 我们选取了一些具有代表性的亲属称谓名词，并且根据属性分为了四组，第一组“父 亲、母亲、父母”，第二组“儿子、女儿、儿女”，第三组“哥哥、弟弟、弟兄”，第四组 “姐姐、妹妹、姊妹”。 第一组：三个方言点的“父亲”和“母亲”的表达形式都不相同，萝卜寨羌语的“父 亲”和“母亲”是借用汉语的形式来表达，即“爸爸”和“阿妈”，其中“阿妈 a55ma55” 可能比“爸爸 pa³¹pa55”借入的时间更早，因为“阿妈 a55ma55”中的词头“阿”是属于比 较早的现象。龙溪两河口羌语的“父亲”和“母亲”分别用“bi55”和“mia³¹”表示，“mia³¹” 除了具有“母亲”的意思以为，还具有[+女性、雌性]的语义特征，在构词中经常使用。 绵虒和龙溪两河口羌语的表达不一样，其中“父亲”用“ʁue³¹ia55”表示，“母亲”用“ʁue³¹mia55” 表示，“ʁue³¹”有可能是词头的一类，而“mia55”也是具有[+女性、雌性]的语义特征，表 明了南部羌语在这一特征上的一致性。“父母”一词的构词方式三个点基本一致，都是把 “父亲”和“母亲”采用联合的形式构成，中间没有使用连词。 第二组：从发音来看，本组中三个方言点的“儿子”读音基本一致，除了绵虒发生了 变化，读为 ʦә³¹，其余两个方言点都读为 ʨi³¹，只是发音部位发生了变化，从发音方法上 来看都是塞擦音。根据 Evans 对南部羌语“儿子”的构拟*ʨi L（2001: 301）可知，萝卜 寨和龙溪两河口保留了原始的读法，而绵虒发生了变化。龙溪两河口羌语的“儿子”有两 种说法，第二种说法 ŋa²¹³很少使用。三个方言点的“女儿”除了声调有差异以外，元音、 辅音都一致。“儿女”的表达则都是采用联合的构词方式，“儿子”加上“女儿”来表达， 257 并且中间都用了连词，这一点和前面“父母”的构词方式有所不同。其中的原因可能是 “父母”一词本来也是具有连词的，由于受到汉语的影响使得原本的连词消失了，变成了 现在看到的形式。而“儿女”一词尚未受到汉语更深入的影响，所以连词仍然存在。根据 “类推”原则，我们猜想，“儿女”一词中的连词 nia/na 最终可能也会消失。 第三组：“哥哥”的表达在三个方言点都不相同，根据构词理据，萝卜寨和龙溪两河 口的“tu”应该是“弟兄”的意思，而“ba、pu”都表示“大”的意思。Evans 对原始南 部羌语“哥哥”的构拟为*tu H brɑ L，其中*brɑ L 是“大”的构拟，而绵虒为什么读为 “ʁue³¹po55”暂时没有找到变化的内在联系。“弟弟”的表达，萝卜寨羌语和龙溪两河口 羌语都是“tu³¹ʦu55”，而绵虒羌语是“tәu³¹ʦu55”，其原始构拟形式为*tu H ʦuә L (Evans 2001: 232)，很明显“tәu³¹ʦu55”中的“tәu”是由于“u”高化之后再双元音化而来的。“弟兄” 的构词方式三个点有所不同，萝卜寨羌语采用联合法的构词方式，没有使用连词。而龙溪 两河口和绵虒羌语则采用另外的构词方式，并且都采用了同样的手段，“tu³¹ni³¹”和“tu³¹nә55” 中后一个成分可能是“二”的意思，而“tu”则有可能是表示“兄弟”意义的词头，组合 在一起之后就表示“两兄弟”的意思。 第四组：三个方言点的“姐姐”的表达都不一样，萝卜寨羌语是“io³¹pa55”，绵虒是 “ʁue³¹ʨe55”，这一形式有可能是一个合璧词，“ʁue³¹”是一个本族语的前缀，“ʨe55”则 是 借自汉语“姐”的读音。龙溪两河口有两种情况，第一种情况是羌语的说法，sa55pa³¹和 ʁua³¹ʦi55，而用于面称的 a55ʨe³¹，可能是借用的汉语“阿姐”49。“妹妹”的表达三个方言 点都不相同，萝卜寨是“io³¹tʂa³¹”，其中“io³¹”是表示“妹妹”的词根义，而“tʂa³¹”是 表示“小”的意思。龙溪两河口是“io55”和“sa55ʦu³¹”两种说法，和萝卜寨方言一样， 其中“io³¹”是表示“妹妹”的词根义，而“sa55ʦu³¹”这一形式中，“sa³¹”可能是表示“姊 妹”的词头，“ʦu³¹”表示“小”的意思，面称则是借用汉语的“阿妹 a55me³¹”。绵虒是 “liәu³¹”，目前尚不知其来源。“姊妹”的构词方式和“兄弟”类似，萝卜寨用的是联合法 构词“mu³¹nia³¹io³¹”，龙溪两河口和绵虒采用相同的构词方式，表示“姊妹”义的 “sa³¹”， 后面加上表示“二（ni 和nә）”的数词，构成了“姊妹”一词。其中，绵虒的“姊妹”还 多了一个词头“a³¹”。 2.4 时间名词 时间名词是指时间的各种形式，或者指相对于目前的时间点（黄成龙，2007：50）。 南部羌语三个方言点时间名词的表达方式基本相同，不同的地方主要在于读音有所差异。 2.4.1 四季的表示 三个方言点中都没有像汉语那样表示“春”、“夏”、“秋”、“冬”四季的词，而只有“热 天”和“冷天”两种表达方式，分别对应于汉语的“春夏”和“秋冬”。这可能与当地非 49 Evans（2001：299）认为“ʨe”不是借词。他的理据是雅都方言读为 ɛ ʨɛ，其对应的汉语意义是 “兄长的妻子”，所以不是来自汉语的“姐”。 264 表3-10 南部羌语三个方言点形状量词对照表 例证 萝卜寨 龙溪两河口 绵虒 一条裤子 soŋ³¹to55a³¹ɚ²¹³ kʰu³³ʦә³¹a³¹ɚ³¹ ze³¹tie55a³¹ɚ³¹ 一张纸 i³¹ʣә³¹a³¹ʦʰe55 zi³¹te55a³¹ʦʰe³¹ ze³¹ʣә³¹a³¹ʦʰi55 一块石头 ʁu³¹pia55 ã³¹ku³¹ ʁo³¹pie³¹a³¹o³¹ ʁo³¹pa³¹a³¹ko³¹ 3.2.2 度量词 度量衡单位量词包括长度单位、时间单位、重量单位和容量单位。根据黄成龙 （2005:18），度量词作为度量的单位，羌语中的度量词可以分为个体度量词（individuated measures words）和群体度量词（group measures words）。 （1）个体度量词。个体度量词主要包括时间单位度量词、尺寸单位度量词、重量单 位度量词等。 ①时间单位度量词 表3-11 南部羌语三个方言时间单位度量词对照表 例证 萝卜寨 龙溪两河口 绵虒 一天 a³¹ɕi³¹ a³¹ɕi³¹ a³¹se55 一个月 a³¹lә³¹ a³¹le³¹ a³¹lә³¹ 一年 a³¹pu³¹ a³¹pu a³¹pu³¹ 从上表中可以看出，三个方言点的时间度量词具有很高的一致性，其中绵虒的“一 天”a³¹se55 和萝卜寨、龙溪两河口羌语的“a³¹ɕi³¹”语音上存在对应关系。 ②尺寸度量词 表3-12 南部羌语三个方言点尺寸度量词对照表 例词 萝卜寨 龙溪两河口 绵虒 一寸 a 55ʨʰy55 a³¹ʦʰuәn²¹³ a³¹ʦʰuei55 一尺 a 55tʂʰi³¹ a³¹tʂʰә³¹ a³¹ʦʰe³¹ 一丈 a³¹dʐa55 a³¹ʐa55 a³¹tʂa55 一庹 a³¹pʰe³¹ a³¹ie55 —— 一拃 ʁue³¹ʨʰy55 a³¹ka³¹ —— ③重量单位度量词 表3-13 南部羌语三个方言点重量单位度量词对照表 例词 萝卜寨 龙溪两河口 绵虒 一斤 a³¹ʨi55 a³¹ʨi55 a³¹ke55 一两 a³¹ʥo³¹ a³¹ʑo55 a³¹tʂәu³¹ 一钱 a³¹ʨe³¹ a³¹zi55 a³¹tʂe55 一石 a³¹ta³¹ a³¹ta³¹ a³¹tiәu55pu³¹ 一升 a³¹ʂi55 a³¹tu³¹ a³¹ɕe55 265 三个方言点的尺寸度量词大都是借用汉语。例如“庹”，萝卜寨羌语中的“庹”的本 义是指成人两臂左右伸直的长度，“一庹”中的“pʰe³¹”是借四川方言的“pʰai5³”，借入之 后发生了音变，即，ai＞e。根据黄成龙（2005），尺寸度量词和重量度量词有标准度量词 和非标准度量词的区分，“丈”、“尺”、“寸”、“斤”、“两”都是标准度量词，其中“丈、 尺、寸、斤”绝大部分都是汉语借词，只是在借入之后发生了音变。“两”则还保持羌语 的读音。 （2）群体度量词。常见的群体度量词主要包括“家、群、堆、行、排、组、套”等， 是由多个个体单位共同构成的度量词。 表3-14 南部羌语三个方言点群体度量词对照表 例词 萝卜寨 龙溪两河口 绵虒 一家 a³¹ʨi³¹pia55 a 55ko55pie55 a³¹ka³¹ 一窝 a³¹pu³¹ a³¹pu³¹ a³¹pu³¹ 一群 a³¹kʰue55 a³¹bo55 a³¹pʰe³¹ 一排 a³¹pʰai³¹ a³¹pʰe³¹pʰe³¹ a³¹pʰai³¹ 一套 a³¹tʰau³¹ a³¹tʰau²¹³ a³¹tʰau²¹³ 一盒 a³¹xo³¹ a³¹xo55 a³¹xo³¹ 群体度量词也分为了固有词和借词两类，“家、窝、群”都是当地羌语本身的表达， 而“排、套、盒”则是借用四川方言来表达。 3.2.3 动量词 动量词是表示动作行为数量和延续时间长短的量词。南部羌语中的动量词不多，但 是使用的范围却很广，有自身专门的动量词，比如“一次”、“一趟”、“一下”、“一觉”等。 例如： 表3-15 南部羌语三个方言点动量词对照表 例证 萝卜寨 龙溪两河口 绵虒 一次 a³¹ɕi55 a³¹tu³¹ a³¹tәu55 一回 —— a³¹tu³¹ a³¹tәu55 一趟 tʰã¹³ —— a³¹tәu55 一遍 a³¹pʰa55 a³¹tu³¹/a³¹tau¹³ a³¹tәu55 一觉 a³¹tye³¹ a³¹zuei³¹ a³¹tua³¹ 一顿 a³¹dua55 a³¹tua55 a³¹tua55 一步 a³¹ua55 a³¹ua55 a³¹tʂʰa³¹tʂʰe³¹ 一脚 a³¹ʨyo³¹tʰәu³¹ a³¹tʂʰu³¹ a³¹ʦʰu³¹ 三个方言点的动量词中的“一次”、“一回”、“一趟”的表达基本是一致的，在当地羌 民族的感知中，“次、回、趟”没有区别，这与汉语的感知（perception）是有差异的，所 以出现了在表达上一致的情况。萝卜寨羌语的“趟”是借自四川方言，说明其感知 （perception）受到汉语的影响而发生了变化。龙溪两河口羌语中“一遍”有两种表示方 266 法，固有词和汉语借词，其中固有词的使用频率更高，而绵虒“一遍”的表达仍然和“一 次（回、趟）”一样。“一顿”则都是用动词词根“打”来表示。萝卜寨羌语的“一脚”是 借用四川方言的“一脚头”来表示，而龙溪两河口和绵虒羌语则是用动词“踢”来表示。 这也可以看出，羌族和汉族认知方式的差异。“一步”这一形式中，萝卜寨和龙溪两河口 的表达是一致的，而绵虒羌语是另外的表达形式。三个方言点的“一觉”互不相同，其中 绵虒羌语的“一觉”和“一顿”相同，萝卜寨的 a³¹tye³¹和龙溪两河口的 a³¹zuei³¹尚不清楚 其构词理据。 四、方位词 方位词是用来表示方向和位置的词，羌语中除了有方向格之外，还有多种方位词。羌 语中并没有固定的表示“东”、“南”、“西”、“北”的词，但是有表示“上”、“下”、“前”、 “后”这样的词。南部羌语表示方位的词比北部羌语的数量少了很多，三个羌语方言点中 表示方位的词已经没有了根据水流方向、靠山方向而进行区别的方位词。三个方言点的 方位词见下表： 表4-1 南部羌语三个方言点方位词对照表 方位词 萝卜寨 龙溪两河口 绵虒 上面 qa 55ɕa55 ke 55ti³¹ ke55te³¹ 下面 qʰe³¹ɕa55 qa 55to55 kʰei 55tәu³¹ 前面 qә³¹ie55 ko³¹la55kә³¹ te³¹ke³¹ 后面 ta³¹qa55to55 ta55to55 ko³¹ko55 里面 do³¹qә55 ŋo 55to55 ke³¹zei55 外面 ʁa³¹ɕa55 bia²¹³ ʁaɹ³¹ 左边 tʂe³¹ʐә³¹ ʦo5³ kʰua5³ʦә³¹ ɡue55a³¹ɕa³¹kʰua³¹ 右边 tʰә³¹ʐә³¹ iәu5³kʰua5³ʦә³¹ da55a55ɕa³¹kʰua³¹ 从表 4-1 中可以看出，南部羌语三个方言点中“左边”和“右边”的表示方法差异较 大。萝卜寨分别用“这边”和“那边”来表示“左边”和“右边”，龙溪两河口羌语则是 借用汉语的“左胯子”和“右胯子”来表示，绵虒的表达比较复杂，其中的 kʰua³¹与龙溪 两河口的汉语借词中的一个语素形式相同，但前面的 ɡue55a³¹ɕa³¹和da55a55ɕa³¹似乎又是羌 语的表达方式。 267 五、小结 本文描写比较了南部羌语三个方言点名词组的特征，既具有一致性，也存在内部差 异性。在最常见的普通名词方面，单音节结构的比例一致性比较高。其差异主要表现在读 音方面，这些差异的形成，不少是元音系统的自身演变造成的，如元音高化、双元音化等。 专有名词方面，附近地区的地名中大多用羌语表达，如果是近期影响进入的地名，则直接 借用汉语来表达。南部羌语中人名的表达手段比较特殊，主要体现在姓氏的表达上。有一 部分姓氏的表达是直接借用汉语，例如“王”；还有一部分姓氏的表达是采用与汉语中的 姓氏词义相同的羌语词汇来表达。如用“獐”对应汉语的姓氏“张”，用性状形容词“高” 对应汉语的姓氏“高”。亲属称谓名词有一部分用本民族词表示，有一部分则用汉语借词 来表示，内部一致性比较高。南部羌语三个点的时间名词的表达方式基本相同，不同的地 方主要在于部分读音发生了变化。南部羌语数词与量词的表达内部一致性也比较高，最 大的差异表现在序数词的表达上，龙溪两河口羌语中序数词“第一”、“第二”的表示方法 已经消失，而是借用汉语的表达方式。南部羌语三个方言点中方位词的表达存在一些差 异，其中最典型的差异体现在“左边”和“右边”的表达形式上。 附记：本文部分内容受到成都大学校基金重点项目（2023XJJO1）的资助。 参考文献 [1] 黄成龙. 2005.《羌语的名量词》，《民族语文》第 5期. [2] 黄成龙. 2007.《蒲溪羌语研究》，北京：民族出版社. [3] 林向荣. 1993.《嘉绒语研究》，成都：四川民族出版社. [4] 孙宏开. 1981.《羌语简志》北京：民族出版社. [5] 孙宏开. 1988.《论羌族双语制：兼谈汉语对羌语的影响》，《民族语文》第 4期. [6] 严木初. 2014.《嘉绒语词汇研究》中央民族大学出版社. [7] 朱晓农. 2004.《汉语元音的高顶出位》，《中国语文》第 5期. [8] Jonathan P. Evans. Introduction to Qiang Lexicon and Phonology Synchrony and Diachrony[M], Tokyo, 2001. [9] Randy J. Lapolla.& Huang Chenglong. A Grammar of Qiang[M]. Mouton de Gruyter, 2003. [10] Zheng Wu-xi. A Grammar of Longxi Qiang [D]. National University of Singapore, 2016. 268 The Comparison of Noun Phrase Characteristics in Southern Qiang Language Gao Tao College of literature and journalism of Chengdu University [Abstract] The article primarily describes and compares the characteristics of noun phrases in three dialects of Southern Qiang language. The proportion of monosyllabic common nouns shows a high degree of internal consistency; however, due to factors such as vowel raising, there are differences in the forms of common nouns across the three dialects, such as the correspondence between 'u' and 'ue'. The expression of place names in proper nouns is relatively consistent, while the representation of surnames in personal names has distinct features, for example, using the Qiang word for the animal 'muntjac' to correspond to the surname 'Zhang', and using the Qiang adjective for 'tall' to correspond to the surname 'Gao'. Kinship terms are partly expressed in the native language and partly through Chinese loanwords, with a high degree of internal consistency. The expression of time nouns is largely the same, with differences mainly appearing in some pronunciation changes, indicating a common origin. The expression of numerals and measure words also shows a high degree of internal consistency, with differences primarily found in ordinal numbers; the expressions for 'first' and 'second' in Longxi Lianghekou have disappeared, adopting instead the Chinese way of expression. There are some differences in the expression of directional terms, most notably in the forms used for 'left' and 'right'. [Keywords] Southern Qiang language noun phrase comparison