Diachronic Perspectives and Synchronic Variation in Southern Min [1 ed.] 036731357X, 9780367313579

Citation preview

Diachronic Perspectives and Synchronic Variation in Southern Min

Diachronic Perspectives and Synchronic Variation in Southern Min aims to address a range of grammatical phenomena in Southern Min. The Sinitic languages show divergence not only in phonology but also in grammar. Together with Hakka, Yue, and part of Wu, Min forms the two major Southern groups of Far Southern and Southeastern languages. There is a range of grammatical phenomena in Southern Min addressed here; the themes and theoretical issues covered in this book touch on a wide range of grammatical patterns of Southern Min from both synchronic and diachronic perspectives including comparatives, obligative and dynamic modals, formation of coordinate conjunctions from the comitative marker, the benefactive marker, the rise of the continuative aspect marker, grammaticalization of the verb of saying into a complementizer, and purposives in Southern Min. This book is aimed at researchers and scholars working on and interested in Chinese linguistics. Chinfa Lien (連金發) is Chair and Emeritus Professor at the Graduate Institute of Linguistics, National Tsing Hua University, Taiwan. He publishes widely from a synchronic and diachronic perspective focusing on functional categories such as demonstratives, pronouns, modals, negation, and aspect as well as grammatical constructions such as interrogatives, exclamatives, and imperatives in Southern Min. Alain Peyraube is Emeritus Director of Research at the Centre National de la Recherche Scientifique (CNRS, Paris, France) and Chair Professor of Chinese Linguistics at the École des Hautes Etudes en Sciences Sociales (EHESS, Paris, France). As a specialist in Chinese historical syntax and more recently on linguistic typology of Sinitic languages, he has authored five books and around two hundred articles. His latest research has been done within a broadly functional and cognitive framework from a cross-linguistic perspective.

Routledge Studies in Chinese Linguistics Series editor: Hongming Zhang (张洪明)

Partition and Quantity Numerical Classifiers, Measurement and Partitive Constructions in Mandarin Chinese Jing Jin Mandarin Loanwords Tae Eun Kim Intensification and Modal Necessity in Mandarin Chinese Jiun-Shiung Wu The Architecture of Periphery in Chinese Cartography and Minimalism Victor Pan Focus Manifestation in Mandarin Chinese and Cantonese A Comparative Perspective Peppina Po-lun Lee Prominence and Locality in Grammar The Syntax and Semantics of Wh-Questions and Reflexives Jianhua Hu Prosodic Studies Challenges and Prospects Edited by Hongming Zhang and Youyong Qian Diachronic Perspectives and Synchronic Variation in Southern Min Edited by Chinfa Lien and Alain Peyraube For more information about this series, please visit: www.routledge.com/ languages/series/RSICL

Diachronic Perspectives and Synchronic Variation in Southern Min Edited by Chinfa Lien and Alain Peyraube

First published 2020 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN and by Routledge 52 Vanderbilt Avenue, New York, NY 10017 Routledge is an imprint of the Taylor & Francis Group, an informa business © 2020 selection and editorial matter, Chinfa Lien and Alain Peyraube; individual chapters, the contributors The right of Chinfa Lien and Alain Peyraube to be identified as the authors of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging-in-Publication Data Names: Lian, Jinfa, editor. | Peyraube, Alain, editor. Title: Diachronic perspectives and synchronic variation in Southern Min / edited by Chinfa Lien and Alain Peyraube. Description: New York: Routledge, 2020. | Includes index. Identifiers: LCCN 2019046903 (print) | LCCN 2019046904 (ebook) | ISBN 9780367313579 (hardback) | ISBN 9780429316517 (ebook) Subjects: LCSH: Southern Min dialects—Grammar, Historical. Classification: LCC PL1703 D53 2020 (print) | LCC PL1703 (ebook) | DDC 495.17/9512—dc23 LC record available at https://lccn.loc.gov/2019046903 LC ebook record available at https://lccn.loc.gov/2019046904 ISBN: 978-0-367-31357-9 (hbk) ISBN: 978-0-429-31651-7 (ebk) Typeset in Times New Roman by Apex CoVantage, LLC

Contents

List of figures and mapsvii List of tablesviii List of contributorsx Abbreviationsxi Acknowledgementsxiii  1 Introduction

1

CHINFA LIEN AND ALAIN PEYRAUBE

  2 Comparatives of inequality in Southern Min: a study in diachronic change from 15th to 21st centuries

9

HILARY CHAPPELL, ALAIN PEYRAUBE AND SONG NA



Appendix I: Li Jing Ji 荔鏡記 and Li Zhi Ji 荔枝記 operas (5 versions)

39

Appendix II: Seven further operas: Măn Tiān Chūn 滿天春, Jīn huā nǚ 金花女, Yù Yán Lì Jǐ 鈺妍麗 錦, Băi Huā Sài Jǐn 百花賽錦, Yáng Guăn Gē 楊管歌, Sū Liùniáng 蘇六娘, Tóng chuāng qín shū jì 同窗琴書記

43



Appendix III: Doctrina Christiana

46



Appendix IV: 174 examples of pi2 比 in all functions

47

  3 The emergence of the obligative modal tioh8 in Southern Min: a change induced by semantic-pragmatic factors

49



TING-TING CHRISTINA HSU

  4 Negation of dynamic modals with DIT 得 in Hainan Min HUICHI LEE

76

vi  Contents   5 Word change and language change: a case of 共 as a coordinating conjunction from Archaic Chinese gòng 共 to ka7 共 in Taiwanese Southern Min

102

JANG-LING LIN

  6 Exploration of the benefactive marker kang7 共 in Ming Qing Southern Min script

125

CHIAN-TANG SU

  7 Taiwanese Southern Min hoo7 and its counterparts in the Southern Min varieties in Quemoy and Quanzhou

144

CHIA-YIN HU

  8 The etymology and grammaticalization of the continuative aspect marker le(h)4: a survey from the historical documents

157

MANJUN CHEN

 9 Kong2 as a verb for saying ‘on the move’ in Taiwanese Southern Min

197

CHINFA LIEN

10 Purposives in Taiwanese Southern Min

217

CHINFA LIEN AND MAY WANG

Index

234

Figures and maps

FIGURES 3.1 Conceptual structure for maximally subjectivized offstage situation in (4c) 3.2 Continuum of force of imposition 3.3 Change of zhuó from verb to preposition 3.4 Reanalysis of zhuó in the construction [zhuó NP V] 3.5 Grammaticalization through revival of the previous cline 3.6 From causative verb to deontic modal 3.7 Diachronic Continua for subjectification in grammaticalization (Traugott 1995:47) 3.8 The importance of a pivotal argument in the change 3.9 Grammaticalization through revival of the previous cline 6.1 Competition between kang7 共 and kah4 合 6.2 The concept of construction B 6.3 The concept of construction C 6.4 Different possessive relations at mismatch’s comparison 8.1 The evolvement from 著 to 嘞 (Wu 2008:21) 8.2 The development of continuity (a) 8.3 The development of continuity (b) 8.4 Grammaticalization process of the marker 咧

52 57 63 65 67 69 69 70 71 126 135 136 137 162 163 165 193

MAPS 7.1 Tendency and distribution of the markers used in double object constructions, dative constructions, and causative constructions 7.2 Tendency and distribution of the markers used in passive constructions

151 152

Tables

1.1 2.1 2.2 2.3 2.4 2.5 2.6 2.7 3.1 3.2 4.1 4.2 5.1 5.2 5.3 5.4 6.1 6.2 6.3 6.4 6.5 6.6 6.7

Patterns of comparatives in Chapter 2 Comparatives in Early Southern Min A new comparative in contemporary Southern Min Four types of comparative structures in the folk operas Lì Jìng Jì 荔鏡記 and Lì Zhī Jì 荔枝記 The four comparative structures in seven local operas from the Ming and the Qing periods The four comparative structures in the Doctrina Christiana (1607) Four functions of pi2 比 in in the folk operas Li Jing Ji 荔鏡記 and Li Zhi Ji 荔枝記 Four functions of pi2 比 in seven local operas from the Ming and the Qing periods The functions of tioh8 in modern SM (from Lien 2001) Negative strategies of SM deontic modals Frequency of the nine types of [Neg X] forms Five basic patterns of negative forms Historical grammatical change of gòng/ka7 共 – throughout 3,500 years The use of ka7 共 and kah4/kap4 合 as AND-CONJ in the 20th century in Taiwanese TV series Sìchóngzòu ‘Quartet’ Functions of Southern Min ka7 共: from ESM (14th – 19th c.) to TSM (20th century) The functions of gòng/ka7 共 for 3500 years The pronunciation of kang7 共 and kah4 合 The conceptual frame, schemas, and constructions of benefactive marker kang7 共 The characteristics of construction A The characteristics of construction B The characteristics of construction C Comparison of the characteristics of the constructions of A, B, and C The conceptual frame, schemas, and constructions of benefactive marker kang7 共

3 10 11 13 14 17 47 48 53 56 90 98 103 111 118 119 126 130 132 134 135 138 140

Tables ix 7.1 Paradigm of TSM ditransitive verbs, causative verbs/markers, dative markers, and agent markers 7.2 Paradigm of the ditransitive verbs, causative verbs/markers, dative markers, and agent markers in Quemoy Southern Min 7.3 Paradigm of the ditransitive verbs, causative verbs/markers, dative markers, and agent markers in Quanzhou Southern Min 7.4 Markers in Double Object Constructions, Dative Constructions, Complex Ditransitive Constructions, Causative Constructions, and Passive Constructions in Min sub-dialects 7.5 The choices of markers in Dative Constructions, Complex Ditransitive Constructions, and Causative Constructions by Speakers from various Min sub-dialects 10.1 Feature distinctions between finite and non-finite clauses

145 148 148 149 154 221

Contributors

Hilary Chappell École des Hautes Études en Sciences Sociales Manjun Chen Jimei University Ting-ting Christina Hsu Chung Yuan Christian University Chia-yin Hu National Taipei University Huichi Lee National Cheng Kung University Chinfa Lien National Tsing Hua University Jang-Ling Lin Hautes Études Commerciales de Paris Song Na Institut National des Langues et Civilisations Orientales Alain Peyraube Centre National de la Recherche Scientifique Chian-Tang Su National Tsing Hua University May Wang National Tsing Hua University

Abbreviations

1 2 3

First person Second person Third person abl Ablative ach Achievement adv Adverb, Adverbializer alt Alternative conjunction am Agent marker appl Applicative art Article asp Aspect marker assum Assumptive ben Benefactive caus Causative clf Classifier cm Comparative marker comi Comitative comp Complementizer conj Conjunction cop Copula counter Counter-intuition dat Dative deg Degree dem Demonstrative deon Deontic dim Diminutive dm Discourse marker dsp Disposal marker esm Early Southern Min excl Exclusive ext Existential foc Focus fut Future

xii  Abbreviations gen Genitive goal Goal

inch Inchoative incl Inclusive

int Interjection

ins Instrumental loc Locative mw

Measure word

neg Negation

nmlz Nominalizer om

Object marker

pass Passive perf Perfect pl Plural pm

Patient marker

prep Preposition

prog Progressive pron Pronoun prt Particle

recp Reciprocal s Subject

sbjv Subjunctive sfp

Sentence-final particle

sem Semelfactive sg Singular stat Stative top Topic tsm

Taiwan Southern Min

v Verb

vol Volitive

? Unknown

Acknowledgements

The research reported on in this book is based on a three-year (2009–2011) ­Taiwan-France cooperative research project Diachronic Change in Southern Min, a Sinitic Language (NSC 98–2923-H-007–001-MY3 and ANR-08-BLAN-017401). We are deeply indebted to the Agence Nationale de la Recherche (ANR), France, and to the Ministry of Science and Technology (MOST) (formerly National Science Council (NSC)), Taiwan, for the funding support. The project saw three workshops for the exchange of ideas and the mutual stimulation of theoretical proposals: (1) ANR/NSC Workshop on Diachronic Change in Southern Min Syntax (Le Centre National de la Recherche Scientifique (CNRS), L’École des Hautes Études en Sciences Sociales (EHESS), Paris, March 29 – April 5 2010), (2) the ANR/NSC workshop in conjunction with the 7th International Symposium on Ancient Chinese Grammar (CNRS and EHESS, Roscoff, France, September  17–18, 2010), and (3) Reflections of Diachronic Change Mirrored in Early Southern Min Texts, (Graduate Institute of Linguistics, National Tsing Hua University, Hsinchu, Taiwan, November 26–27, 2011). Thanks are due to both ANR and NSC/MOST for financial support. During the gestation and execution of the project we were blessed with the opportunity to encounter and benefit immensely from the input and inspiration of a huge number of friends, colleagues, and students. While running the risk of missing the names, we would like to mention in alphabetic order a list of people for appreciation: Chingchuan Chang, Chun Chang, Celine Y-C Chang, Hsinchang Chen, I-hsuan Chen, Liying Chen, Po-ching Chen, Ying Cheng, Minhua Chiang, Chengchun Hsiao, Peiyi Hsiao, Hanchun Huang, I-long Huang, Weiju Kuo, Huiling Lai, Liching Lin, Liwei Liu, Anne Saisi, Kuanming Teng, Youtsih Yan, and Huiling Yang. We would like to thank, in particular, Hilary Chappell, for her ungrudging concern and perceptive comments and suggestions at every stage of the fine-tuning of the manuscripts. We are grateful to Hongming Zhang for his interest and encouragement to include this book in a series of books on Chinese linguistics for which he is the chief editor. We would like to thank the anonymous reviewers for their positive evaluation. We owe a great deal to Andrea Hartill, the publisher, for her help and

xiv  Acknowledgements enthusiasm in making the project of publishing this book a reality. We are especially indebted to Yu-ju Yang for his patience and meticulous attention in the final stage of preparing the text for publication. Chinfa Lien, Alain Peyraube Hsinchu & Paris

1 Introduction Chinfa Lien and Alain Peyraube

The Sinitic languages show great divergence not only in phonology but also in grammar. Past efforts have been mostly focused on the mainstream and official language, Standard Mandarin, or almost overwhelmingly on phonological aspects of Sinitic.1 Yet Mandarin does not equal the Sinitic languages on their own, and phonology only reflects one aspect of the linguistic sign featuring the composite whole of sound, meaning and syntax. The Sinitic languages comprise major subgroups such as Mandarin, Xiang, Gan, Hui, Wu, Pinghua, Jin, Min, Hakka, and Yue.2 Together with Hakka, Yue and part of Wu, Min forms part of the two major Southern groups of Far-Southern and Southeastern (Chappell 2015). These groups can be distinguished from the Northern groups by a range of grammatical features. First, for negation, Min (Southern Min included), along with other Southern groups such as Hakka and Yue are characterized by sources in the ordinary negative adverb with the onset traceable to the bilabial nasal *m-唔 rather than voiceless bilabial stop (*p-) 不.3 Negative markers possessing the onset of the reconstructed voiceless bilabial stop (*p-) in these languages must be regarded as an alien element borrowed from Northern Sinitic. However, one should be careful not to overgeneralize this prima facie evidence. For some other grammatical aspects, Southern Min may better construed as patterning with Wu as opposed to Hakka and Yue (Chappell 2015). The distribution of third-person pronouns also shows a distinction of Northern (or rather Mandarin) and Southern (or rather non-Mandarin) groups. The Northern group features tā 他 in Mandarin dialects, whereas Southern groups involve pronominal forms including *g- 渠/佢, *g- 其 or *0- 伊, the syllables with the velar or zero onset, in Yue, Hakka, Gan, and Min as well as Wu and Xiang. However, some varieties of Wu and Xiang experience various degrees of onslaught from the Mandarin pattern (Norman 1988). Second, the permissive causative verb and the agentive marker in passives are a result of grammaticalization of ditransitive verbs meaning ‘give’ such as khit4 乞 or thoo7 度 in Southern Min Lien (2001). By contrast, the link between causatives and passives is expressed by jiào 叫 as a directive causative in Mandarin. Kah4 教, a commonly used function word showing the imprint of glottalization, in the Quanzhou variety of Southern Min, is used as a directive causative verb and has nothing to do with the agentive marker in the passives.

2  Chinfa Lien and Alain Peyraube Third, bǎ 把 as a differential object marker in the disposal construction in Mandarin is expressed by another functional word, tsiong1 將, which underwent the same type of grammaticalization but in an earlier chronological stratum,4 and also by a comitative, ka7~kang7 共 in the native stratum. Fourth, the most common conjunctions are hé 和 and gēn 跟 in Mandarin, whereas their counterparts in Southern Min are kap4/kah4 合, tsham1 參 or an earlier form kang7 共, which evolved from an adverbial function. Fifth, the postverbal element tit4 得 has developed into a modal in Mandarin, whereas modals in Southern Min come from main verbs such as sai2 使 ‘cause’, ing7 用 ‘use’, and tsue3 做 ‘do’ sandwiched by preverbal modal e7 解 and the postverbal tit4 得 yielding the complex lexicalized modals such as e7sai2tit4 解使得, e7ing7tit4 解用得, and e7tso3tit4 解做得 in Southern Min. All the three lexicalized elements meaning ‘can’ function as preverbal deontic modals. These five features are only the ‘tip-of-the-iceberg’ phenomena in Southern Min. There exist many more telltale grammatical patterns and concomitant theoretical issues that still remain unexplored. As a research result of the Taiwan-France cooperative project entitled Diachronic Change in Southern ­ Min, a Sinitic Language, or DiaMin, this volume aims at filling in some of the gaps in our understanding of Sinitic grammar, at the same time reflecting the state of the art for studies on Southern Min languages with the caveat that other southern Sinitic languages still remain unexplored (Chappell 2019). Even within the confines of the languages spoken in the region of Southern Fujian (福建, Hokkien), the homeland of Southern Min, the research represented in this book barely scratches the surface, and the range of Southern Min dialects studied is still underrepresented, let alone the other ‘Southern Min’ varieties distributed over substantial areas of Guangdong and Hainan Island, as well as the Hokkien Southern Min prevalent in Southeast Asia found in far-flung areas such as the Philippines, Singapore, Malaysia, and even in Indonesia. Evidence for such a variety of Diaspora Hokkien Southern Min can already be attested in extant texts dateable to the sixteenth century at the earliest, when the maritime Spaniards witnessed its golden age.5. This book touches on a wide range of grammatical patterns of Southern Min from both synchronic and diachronic perspectives including comparatives, obligative and dynamic modals, formation of coordinate conjunctions from the comitative marker, the benefactive marker, the rise of continuative aspect marker, grammaticalization of the verb of saying into a complementizer, and purposives in Southern Min. In what follows we provide a rough outline of each chapter. In ‘Comparatives of inequality in Southern Min: a study in diachronic change from 15th – 21st c.’, Hilary Chappell, Alain Peyraube, and Song Na address the topic of comparatives of inequality from a diachronic perspective (see also Peyraube 1989). Any comparative construction will involve the basic grammatical categories such as the target of comparison, the degree adverb, the predicate of comparison, the standard marker, and the standard of comparison. Typological and diachronic change for comparatives involve a basic issue concerning the

Introduction  3 order of the standard of comparison relative to the predicate of comparison, in particular determining: 1 2 3 4

Whether the standard of comparison occurs before the predicate of comparison or not; Whether the standard of comparison is marked or not; Whether the predicate of comparison is modified by a degree adverb; If the answer is affirmative to (2), what marker is chosen?

This approach yields the patterns of comparatives captured in Table 1.1: Table 1.1  Patterns of comparatives in Chapter 2

a b (i) b(ii) c (i)

c(ii)

1

2

3

4

√ * * *

√ √ √ *

√ * * *

pi2 比/ phing7 併 kue3 過 ju5 如/ su7 似 0

*

*

√

0

The tick and the asterisk correspond to a positive or negative answer to the question posed for the first three parameters. The fourth parameter, a parameter relating to the second question, is concerned with the lexical selection of the marker of the standard of comparison. Markers of SC may be subject to lexical variation. Some lexical choice may have to do with subdialectal variation. For example, phing7 併 alongside pi2 比 seems to be uniquely found in the Quanzhou variety of Southern Min. The three patterns of construction outlined above may be viewed in a diachronic perspective. Pattern (a) differs from (b) and (c) in that the standard of comparison is obligatorily marked by a standard marker which occurs before the predicate of comparison. By contrast, (b) and (c) feature the occurrence of the predicate of comparison before the standard of comparison. Type (a) is only found in Modern Chinese and is said to arise from a serial verb construction. Grammar changes through time. Early Southern Min playscripts preserve and attest to the existence of (b) and (c) that suggest more conservative patterns of comparatives. The structural change affecting comparatives reflects the general tendency of grammatical changes in Sinitic languages. Importantly, the prepositional phrase comprising the standard noun phrase in comparatives that used to be placed post-verbally in Old Chinese has come to occupy a preverbal position in Modern Standard Chinese (see Peyraube 1989; Lien 1999; Chappell and Peyraube 2015). In the second study, ‘The emergence of obligative modal tioh8 in Southern Min: a change induced by semantic-pragmatic factors’, Ting-ting Christina Hsu tackles two routes of grammaticalization for the lexeme tioh8 in Taiwanese Southern Min: (1) as a verb of attachment to a postverbal preposition and in the form of (2) causative verbs to obligative modals. First, tioh8 著 as a verb

4  Chinfa Lien and Alain Peyraube of placement was reanalyzed as a preposition heading a post-verbal locative prepositional phrase in Early Medieval Chinese (2nd – 6th c. c.e.). Later in the Late Medieval period, it developed into a preposition heading a preverbal instrumental phrase. Second, tioh8 著 as a verb of attachment was reanalyzed as a postverbal complement denoting ‘attach, destination’ in Early Medieval Chinese and evolved into the postverbal phrasal complement denoting stative continuation or completion in the Late Medieval period. It took on the role of a causative verb, as attested in the Lǎo Qǐ Dà 老乞大, a text dating from Yuan Dynasty in the first half of the 14th century, in the Pre-Modern period (see Peyraube 2017 for the different periods in the history of Chinese). This provides a link to its functioning as a deontic modal in Southern Min. As encaptured in Figure 3.7, the author provides a stepwise account of the development of the causative verb tioh8 著 as a deontic modal: (1) subjectivizing the causer, (2) providing a slot for the causee, and (3) reanalyzing the pivotal construction into a monoclausal construction. The contribution ‘Negation of dynamic modals with DIT 得 in Hainan Min’ by Huichi Lee is concerned with the negation of four types of construction featuring the postverbal dynamic modal ɗit55 得 in Hainan Min: (1) V得, (2) V得O, (3) V得C, and (4) V得OC where V=verb, O=object, and C=complement. The negation of each construction simply involves attaching the negative element bo22 無 in a position preceding the modal complex: (1) 無V得, (2) 無V得O, (3) 無V得C, and (4) 無V得OC. The author argues that these unique constructions show conservative features and bear the effect of analogy. Hainan Min, alongside of Southern Sinitic languages such as Hakka, Cantonese, and Southern Min, shares the important constructional features of Neg V 得 O C where the post-verbal ɗit55 得 and C are separated by O, the object (cf. Lien 2011). By contrast, the complement, C, in Northern Sinitic languages is attracted to the position immediately after 得. In the chapter on ‘Word change and language change: a case of 共 as a coordinating conjunction from Archaic Chinese gòng 共 to ka7 共 in Taiwanese Southern Min’, Jang-Ling Lin proposes that the adverb 共 evolves into the WITH/AND preposition/conjunction (Liu and Peyraube 1994; Peyraube 1996; Xu and Peyraube 1997). It is well-known that kang7 共 functioned as a verb in the Archaic Chinese period. As attested in Shì Shuō Xīn Yǔ 世說新語 ‘A New Account of the Tales of the World (5th c. c.e.)’, an early Medieval work, it emerged as an adverb. It was not until the Late Medieval period as attested in the Dūnhúang Biànwénjí 敦煌 變文集 ‘Transformation texts (8th – 11th c. c.e.)’ that it took on the dual function of adverb and preposition/coordinating conjunction. Its adverbial function was kept alive in Mandarin vernacular literature such as Xī Yóu Jì 西遊記 ‘Travels to the West (16th c. c.e.)’ and Hóng Lóu Mèng 紅樓夢 ‘Dream of the Red Chambers (18th c. c.e.)’. In early Southern Min texts kang7 共, dating back to the 17th century c.e. at the earliest, kept its function of a conjunction that was inherited from its Late Medieval precursor, but went on to develop its prepositional function serving as a marker of a range of semantic roles such as source, goal, and benefactive. It is quite noteworthy that the function of conjunction in kang7 共 has been taken over by another conjunction kap4 合 in modern Taiwanese Southern Min (see Lien 2015).

Introduction 5 ‘Exploration of the benefactive marker kang7 共 in Ming Qing Southern Min script’ contributed by Chian-Tang Su deals with three constructions (simple benefactives, causative benefactives, and focused benefactives) underscored by three schemas featuring either an event wherein an agent executes an action or two successive events in which the agent does a preparatory act for the purpose of carrying out the next act. Be it a single event or two consecutive events, the event or events are for the benefit of an affectee. The simple benefactive performs an action for the benefit of an affectee. The causative benefactive denotes a causative event in which the causer, overt or covert, impels a causee into executing an act for the benefit of an affectee. The focused benefactive comprises a sequence of two related events; the first event serves as the backdrop against which the event in focus is placed. The affectee is always attached to the event in focus. In general, the three constructions all involve the execution of an event or events for the benefit of the affectee which most often is realized as the speaker or the addressee in interpersonal interaction. In the chapter, ‘Taiwanese Southern Min hoo7 and its counterparts in the Southern Min varieties in Quemoy and Quanzhou’, another important contribution to this volume, Chia-yin Hu explores the multiple functions of ditransitive verbs in Taiwanese Southern Min, Quemoy Southern Min from Quemoy (alias Jinmen, an island off Southern Fujian) and Quanzhou Southern Min which is essentially the homeland of Southern Min, using data based on her fieldwork. Previous studies show that, apart from being a ditransitive verb, hoo7 boasts a range of functions such as the marker of datives (i.e. the recipient in double object constructions), the causative verb in pivotal constructions and the agent marker in passive constructions. There are two types of passives featuring hoo7. In commonly used passives hoo7 functions to mark the agent or instrumental noun phrase, whereas in atypical passives it may denote an adversative event by which the speaker, often not explicitly spelt out but implied, is affected (Lien 2008). Fieldwork shows that there are many points of discrepancy in the choice of ditransitive verbs of giving as well as in their functional extensions between Taiwanese and its counterparts in Quemoy and Quanzhou. First, in Quanzhou Southern Min there is more lexical variation to be found for the exponents of ditransitive verbs, causative verbs, dative markers, and agent markers, including thoo7 度 and tng5 傳 as well as hoo7 予, whereas only hoo7 予 is used in Taiwanese Southern Min. Second, the agent marker can be realized as kheh4 乞 (#x003C; khit4) alone in Quemoy Southern Min and Quanzhou Southern Min, whereas it can only be expressed as part of the synonymous compound kheh4hoo7 乞予 (#x003C; khit4hoo7) in Taiwanese Southern Min.6 From the distribution of the functional words relating to the construction types under discussion, we can pinpoint a spatial continuum concerning the diversity in the use of different exponents. In ‘The etymology and grammaticalization of the continuative aspect marker le(h)4 a survey from the historical documents’, Manjun Chen addresses the issue of grammaticalization turning a light locative noun te3 處 or a locative prepositional phrase ti7te3 在處 into a continuative aspect marker. Previous literature claims that the modern continuative marker leh4 咧 evolving from tioh8 著 as a verb or a preposition is the source from which leh4 咧 as a marker of accomplishment,

6  Chinfa Lien and Alain Peyraube progressive or continuative aspect emerges. Drawing on proposals concerning the rise of continuative aspect from a single source view, Chen proposes that there are two routes of grammaticalization: (1) the demonstrative + te3 處 started off as a location-denoting content expression and then developed into a stage which was ambiguous between spatial and temporal readings with the shedding of the demonstrative. The construction featuring te3 處 alone involves the non-­ existential meaning and is responsible for the rise of imperfect leh0 咧. In particular, it underwent phonological reduction and took on the sole temporal function as an aspectual marker, and ultimately yielded the function of a sentence-final particle in connection with the expression of a mild imperative mood. For another route of grammaticalization, the source is a verb-object construction comprising the existential verb ti7 在 and the determiner phrase (a demonstrative +light locative noun te3 處). This carries an existential meaning. With the shedding of the demonstrative, ti7te3 在處 became a unit wavering between locative and temporal readings. The postverbal ti7te3 may assume a role of a continuative aspect marker, but if it is followed by a locative phrase then it becomes a preposition. Chinfa Lien’s ‘Kong2 as a verb for saying “on the move” in Taiwanese Southern Min’ explores the process of grammaticalization of kong2 講 based on a body of colloquial Taiwanese Southern Min folktales. As a lexeme, kong2 講 is a threeplace predicate encompassing the semantic role of speaker, addressee, and theme denoting the content of speech. Apart from being a lexeme, it has also taken on a range of grammatical functions such as quotative marker, complementizer, topic marker, parenthetical, and sentence-final particle. ‘Purposives in Taiwanese Southern Min’, co-authored by Chinfa Lien and May Wang, aims at teasing out the constraints on purposives in Taiwanese Southern Min. There are three kinds of purposives: (1) purposives with 來 lai5/khi3 去, (2) purposives with ti7leh4 佇咧, and (3) purposives with hoo7 好. This study is highly pertinent to the issue regarding the distinction between finite and non-finite clauses and the grammaticalization of the deictic particles 來 lai5/khi3 去, as well as the scalar adjective hoo7 好 as markers of purposive clauses. NOTES 1 See Kwok 2018 for a fine-grained phonological treatment of Southern Min from a comparative perspective. 2 See Chappell 2001a, 2001b, 2015 for overviews of the Sinitic languages in particular with respect to their much-neglected vibrancy and diversity. For surveys of southern dialects like Yue, Wu and Min as well as Mandarin dialects Norman 1988 and Wang 1991. For Hakka dialects see Hashimoto 1973 and Chappell and Lamarre 2005. 3 Demiéville 1950; Dobson 1966; Norman 1988; Mei 2013; and Chappell and Peyraube 2016. For negation in Southern Min, see Teng 1992. 4 Peyraube 1996; Sun 2015. 5 See the pioneer work of van der Loon 1966, 1967 as well as subsequent works such as Yue-Hashimoto 1999; Chappell 2000; and Klöter 2011.

Introduction  7 6 Hoo7 予 may well evolve from thoo7 度 by spirantizing the stop onset of the syllable, regardless of whether or not it can be taken as a descendant of 與 in Archaic Chinese. See Mei 2007. Note that 予 is a demotic character for the lexeme hoo7.

REFERENCES Chappell, Hilary. 2000. Dialect grammar in two early modern Southern Min texts. Journal of Chinse Linguistics 28:247–302. Chappell, Hilary. 2001a. Language contact and areal diffusion in Sinitic languages: Problems for typology and genetic affiliation. In Areal diffusion and genetic inheritance: Problems in comparative linguistics, ed. by Alexandra Aikhenvald and R.M.W. Dixon, 328–357. Oxford: Oxford University Press. Chappell, Hilary, ed. 2001b. Sinitc grammar: Synchronic and diachronic perspectives. New York: Oxford University Press. Chappell, Hilary. 2015. Linguistic areas in China for differential object marking, passive and comparative constructions. In Diversity in Sinitic languages, ed. by Hilary Chappell, 13–52. Oxford: Oxford University Press. Chappell, Hilary. 2019. A grammatical sketch of Southern Min. In The Mainland Southeast Asia linguistic area, ed. by Alice Vittrant and Justin Watkins, 176–233. Berlin: Mouton de Gruyter. Chappell, Hilary, and Christine Lamarre. 2005. A grammar and lexicon of Hakka, historical materials from the Basel Mission Library. Paris: CRLAO, École des Hautes Études en Sciences Sociales. Chappell, Hilary, and Alain Peyraube. 2015. The comparative construction in Sinitic languages: Synchronic and diachronic variation. In Diversity in Sinitic languages, ed. by Hilary Chappell, 134–154. Oxford: Oxford University Press. Chappell, Hilary, and Alain Peyraube. 2016. A  semantic typology of negation in Sinitic languages: Synchronic and diachronic views. In New horizons in the study of Chinese: Dialectology, grammar and philology - studies in honor of Professor Anne Yue, ed. by Pang-hsin Ting, S. Cheung, S. Tang, and A. Chin, 483–533. Hong Kong: T.T. Ng Chinese Language Research Center, Chinese University of Hong Kong. Demiéville, P.  1950. Archaïsmes de prononciation en chinois vulgaire. T’oung Pao 40:1–59. Dobson, W.A.C.H. 1966. Negation in Archaic Chinese. Language 42.2:278–284. Hashimoto, Mantaro J. 1973. The Hakka dialect: A linguistic study of its phonology, syntax and lexicon. New York: Cambridge University Press. Klöter, Henning. 2011. The language of the Sangleys: A Chinese vernacular in missionary sources of the seventeenth century. Leiden, Boston: Brill. Kwok, Bit-chee. 2018. Southern Min: Comparative phonology and subgrouping. New York: Routledge. Lien, Chinfa. 1999. The sociolinguistic dimensions of comparative constructions in Taiwan Southern Min. In Contemporary studies on the Min, ed. by Pang-hsin Ting, 204–224. Berkeley: Project on Linguistic Analysis, University of California. Lien, Chinfa. 2001. Grammatical function words 乞, 度, 共, 甲, 將 and 力 in Li4 Jing4 Ji4 荔鏡記 and their development in Southern Min. In Papers from the third international conference on Sinology: Linguistic section. Dialect variations in Chinese, ed. by Dah-an Ho, 179–216. Taipei: Institute of Linguistics, Preparatory Office.

8  Chinfa Lien and Alain Peyraube Lien, Chinfa. 2008. Special types of passive and causative constructions in TSM. In Chinese linguistics in Leipzig, ed. by Redouane Djamouri, Barbara Meisterernst, and Rint Sybesma, 223–237. Paris: Centre de Recherches Linguistiques sur l’Asie Orientale. École des Hautes Études en Sciences Sociales. Lien, Chinfa. 2011. Interface of modality and the 得 constructions in Southern Min: A case study of their developments from earlier Southern Min in the Ming and Qing to modern Taiwanese Southern Min. Language and Linguistics 12:723–752. Lien, Chinfa. 2015. The condition and change of 共 vis-à-vis 合 in Southern Min with a sidelight on intra-dialectal variation. Journal of Chinese Linguistics 43:1–33. Liu, Jian, and Alain Peyraube. 1994. History of some coordinative constructions in Chinese. Journal of Chinese Linguistics 22:179–201. Mei, Tsu-lin. 2007. Mǐnyǔ yú yù liǎngzì yángdiào h- shēngmǔ de láiyuán 閩語“輿”、“浴”兩 字陽調h-聲母的來源 [An etymology for Southern Min hoo7 ‘to give’ and related matters]. Fāngyán 方言 4:289–295. Mei, Tsu-lin. 2013. Fǒudìngcí ‘bù’ ‘fú’ zài Hànyǔ fāngyán lǐ de fēnbù jí qí yǎnbiàn 否定 詞“不”“弗”在漢語方言裡的分布及其演變 [The distribution of ordinary negatives 不 bù and 弗 fú in Chinese dialects past and present]. Fāngyán 方言 1:1–10. Norman, Jerry. 1988. Chinese. New York: Cambridge University Press. Peyraube, Alain. 1989. History of the comparative constructions in Chinese from the 5th century bc to the 14th century ad. In Proceedings of the 2nd international conference on Sinology, 589–612. Taiwan: Academia Sinica. Peyraube, Alain. 1996. Recent issues in Chinese historical syntax. In New horizons in Chinese linguistics, ed. by C-T. James Huang and Y.H. Audrey Li, 161–213. Dordrecht: Kluwer Academic Publishers. Peyraube, Alain. 2017. Periodization. In Encyclopedia of Chinese language and linguistics, ed. by R. Sybesma. Vol. III, 346–349. Leiden: Brill. Sun, Chaofen. 2015. The grammaticalization of the bǎ construction: Cause and effect in a case of specialization. In The Oxford handbook of Chinese linguistics, ed. by Willaim S-Y Wang and Chaofen Sun, 429–442. New York: Oxford University Press. Teng, Shou-hsin. 1992. Diversification and unification of negatives in Taiwanese. In Chinese languages and linguistics. Vol. 1, Chinese dialects, ed. by Institute of History and Philology Academia Sinica, 609–629. Taipei: Academia Sinica. van der Loon, Piet. 1966. The Manila incunabula and early Hokkien studies (Part 1). Asia Major 12:95–186. van der Loon, Piet. 1967. The Manila incunabula and early Hokkien studies (Part 2). Asia Major 13:1–43. Wang, William S-Y., ed. 1991. Languages and dialects of China. Journal of Chinese Linguistics Monograph Series. Hong Kong: The Chinese University of Hong Kong Press. Xu, Liejiong, and Alain Peyraube. 1997. On the double-object construction and the oblique construction in Cantonese. Studies in Language 21:105–127. Yue-Hashimoto, Anne Oi-kan. 1999. The Min translation of the Doctrina Christiana. In Contemporary studies on the Min dialects, ed. by Pang-Hsin Ting, 42–76. Berkeley: Project on Linguistic Analysis.

2 Comparatives of inequality in Southern Min A study in diachronic change from 15th to 21st centuries* Hilary Chappell, Alain Peyraube and Song Na 1  INTRODUCTION A comparative construction involves a grading process: two objects are positioned along a scale or continuum with respect to a certain property. One object or person can have either more, less or an equal degree of the given dimension or quality when judged against a second object or person. There are consequently comparatives of inequality (including comparatives of superiority and comparatives of inferiority) and comparatives of equality. Examples: 1 2 3

Carla is taller than Nicolas (inequality/superiority subtype) Carla is less tall than Nicolas (inequality/inferiority subtype) Carla is as tall as Nicolas (equality)

We will single out the comparative of inequality (chābĭjù 差比句 in Chinese) as our target in this research on the diachronic evolution of the construction in Southern Min. In other words, we will not deal with comparatives of equality (known as either píngbĭjù 平比句 or dĕngbĭjù 等比句 in Chinese). As there are no comparatives of inferiority involving a comparative morpheme like the ­English ‘less Adj than’ or the French ‘moins Adj que’, either in Standard Mandarin or in Southern Min, this means that our study will concentrate on comparatives of inequality for the superiority subtype.1 Chappell and Peyraube (2011, 2015) claim that synchronically two main comparative construction types predominate in Sinitic languages: Type I: Compare type – dependent marked: NPA-CM-NPB-Verb Type II: Surpass type – head marked: NPA–Verb-CM-NPB [where CM=comparative marker; NPA=comparee; NPB=NP acting as standard; Verb=adjective or verb which codes the dimension of the comparison].

10  Hilary Chappell, Alain Peyraube and Song Na Examples and a brief description for each of these two types follow:

Type I: Standard Mandarin (4)

她 比 我 高 tā bǐ wǒ gāo 3sg cm 1sg tall ‘She is taller than me’.

The preposition bĭ 比 has its origin in a verb ‘to compare’, used in serial verb constructions from the period of Early Medieval Chinese (3rd – 4th c. ce). See ­Peyraube 1989. Dependent-marking structures like Type I are often considered to represent the most prototypical of all comparative constructions (Dixon 2012:343ff; Heine 1997:110).

Type II: Hong Kong Cantonese (5)

佢 高 過 我 khøy13 kou55 kuɔ33 ŋɔ13 3sg tall cm 1sg ‘She is taller than me’.

The second structural Type II is commonly represented as the Surpass schema in Sinitic languages because the comparative marker has its origin in a verb meaning ‘surpass’ or ‘exceed’. It is also called the ‘Action’ schema in Heine 1997:112 due to its transitivity: ‘she tall surpasses me’. Nevertheless, cross-linguistic research on the major comparative strategies and cognitive schemata points to the existence of more than just these two types. See the six different structures distinguished by Stassen (1985, 2005, 2011) and the eight cognitive schemata identified by Heine (1997), which are compared in Chappell and Peyraube 2015. Chappell (2015) observes that at least seven of these recognized types can be identified for Sinitic languages. Our analysis will investigate four syntactic structures in Southern Min for these comparatives of inequality which are associated in their turn with five cognitive schemata, as follows, with A representing the comparee and B, the standard of comparison: Table 2.1  Comparatives in Early Southern Min Syntactic Structure

Cognitive schema

Type I: Type II(a): Type II(bi/ii): Type III: Type IV:

Compare comparative Surpass comparative Surpass-Similarity comparative Zero-marked comparative Adverbial comparative

A - CM比 - B - Verb A - Verb - CM過 – B A - Verb - CM如/似 – B A - Verb - B A - CM較 - Verb - B

Comparatives of inequality in Southern Min  11 Table 2.2  A new comparative in contemporary Southern Min Hybridized comparative

Type V: A - CM(i) 比 - B - CM(ii) 較 - Verb

We have divided Type II into two subtypes, depending on whether the postverbal marker is guò 過 or rú 如/sì 似, which belong to two different cognitive schemata, identified as Surpass and Similarity in Heine 1997. Finally, in a brief examination in §4 of the synchronic situation, we will point out which of the comparative structures in Table 2.1 are still used in contemporary Southern Min and also consider a newly arisen hybridized comparative form which combines Type I with Type IV, i.e. Type V. Note that the labels, A=NPA and B=NPB, representing respectively the ‘comparee’ and the ‘standard’ of the comparison, are not necessarily NPs in Chinese languages. They may be pronouns, demonstratives, or even VPs or clauses. Furthermore, in our survey of all these constructions in the given texts, we did not count any examples where either the comparee (A) or the standard of the comparison (B) were missing from the immediate context, even if mentioned in the preceding text, and certainly not, if they were implicit. Type I is dependent-marking, Types II and IV are head-marking, whereas Type III has no marking whatsoever and thus reveals itself to be a rather unusual comparative structure. Type V will be shown to combine both head-marking and dependent-marking strategies. Examples of these four types (and five schemata) found in Early Southern Min are given next while the entire corpus may be consulted in the appendices:2

Type I: Compare comparative: A+比+B+Verb (6)

阮 不 比 王魁 負心 gun2 m7 pi2 ong5khue1 hu7sim1 1sg neg cm Wang.Kui unfaithful ‘I am not as unfaithful as Wang Kui’.3

(29.191 Jiājìng 嘉靖)

Type II (a): Surpass comparative: A+Verb+過+B (7)

你 只 計 賽 過 孫 吳 li2 tsit8 ke3 sai3 ke3 sun1 ngo5 2sg dem strategy be.better cm Sun.Tzu Wu.Qi ‘Your strategy is better than those of Sun Tzu or Wu Qi’.

(23.054 Jiājìng 嘉靖)

Type II (bi/ii): Surpass-Similarity comparative: A+Verb+如/似+B (8)

一 場 姻愛 勝 如 春夢 tsit8 tiunn5 ian1ai3 sing3 ji5 tshun1bang7 one clf love.affair be.better cm romantic.dream ‘A love affair is better than a romantic dream’.

(27.019 Wànlì 萬曆)

12  Hilary Chappell, Alain Peyraube and Song Na

Type III: Zero-marked comparative: A+Verb+B   (9)

人 lang5 people

說 sueh4 say

一 tsit8 one

某(ㄙ) bo2 wife

強 kiong5 better.than

十 tsap8 ten

被, phe7 quilts

十 tsap8 ten

被 甲 也 寒 (5.047 Jiājìng 嘉靖) phe7 kah4 a7 han5 quilts cover still cold ‘People say having a wife is better than ten quilts, ten quilts cover you but you’re still cold’.4

Type IV: Adverbial comparative: A+可+Verb+B (10)

官人 kuann1lang5 gentlemen

許 hi2 dem

前頭 tsing5thau5 front

一 tsit8 one

陣 tin7 group

娘仔， niu5a2 woman

生得 senn1tit4 be.born.to

句 可 親淺 伊 (8.069 Jiājìng 嘉靖) koh4 khah4 tshin1tshian2 i1 even cm beautiful 3sg ‘There are beautiful women in front of the gentleman who are even more beautiful than her’. The Type III Zero-marked comparative, while clearly pan-Sinitic in nature, is a structure apparently not found in other languages of the world (see Haspelmath and the Comparative Constructions Consortium 2013), whereas Type IV is rare not only in Sinitic but also cross-linguistically (Chappell 2015). While giving brief descriptions of all four types (and five cognitive schemata) in Southern Min, we will concentrate our study on the two subtypes of Surpass, illustrated in (7) and (8) since they turn out to be common comparative construction types in Early Southern Min, as well as the Adverbial type in (10) in order to discuss their diachronic development. 2  THE DIACHRONIC CORPUS FOR EARLY SOUTHERN MIN We have consulted a large corpus of works on Early Southern Min, including principally the operas known under the titles of both Lì Jìng Jì 荔鏡記 and Lì Zhī Jì荔枝記 (see the English translations below) for which five extant versions are in existence spanning three centuries (see Table 2.3 below). We have also made use of a late 16th century religious text, the Doctrina Christiana and a 1620–1621 sketch grammar of Southern Min written in Spanish, in addition to seven other local operas (see Tables 2.3 and 2.4 below).

1

深tshim1(2) 重tang7(1)

強 kiong5(2) 勝 sing3(1) 賽 sai3 (2)

負心 hu7sim1 (1)

1

4

24

5

1

Total

4 強kiong5(3) 勝 sing3(1) 賽 sai3 (1)

1 負心 hu7 sim1(1)

光緒 (1884)

37

6

6 深tshim1(2) 大tua7(2) 緊kin2(1) 重tang7(1)

3 深tshim1(2) 大tua7(2) 緊kin2(1) 重tang7(1)

5

强 kiong5 (2) 賽 sai3 (1) 勝 sing3(1)

負心 hu7 sim1(1)

4

強 kiong5 (2) 賽 sai3 (1) 澀 siap4(1)

0

11 勝seng3(1) 富hu3(2) 光kong1(1) 緊kin2(1) 冷ling2(2) 大tua7(3) 悶bun7(1) 11

6

道光 (1831)

順治 (1651)

萬曆 (1581)

短tuan (1) 緊kin2 (1) 甜tinn1(1) 悶bun7(3) 大tua7(1) 重tang7(2) 遠hng7(2)

強kiong (2) 賽 sai3(3) 澀 siap4(1)

負心 hu7 sim1 (1)

嘉靖 (1566)

2

Surpass-Similarity

II(bi) A+Verb+如+B

Surpass 5

II(a) A+Verb+過+B

A+比+B+ Verb

Compare

Type II

Type I

Text

Schema

4

0

0

0

1

勝sing3(1)

3

薄pok (1) 深tshim1(1) 忙bang5(1) 8

II(bii) A+Verb+似+B

Table 2.3  Four types of comparative structures in the folk operas Lì Jìng Jì 荔鏡記 and Lì Zhī Jì 荔枝記

36

强 kiong5 (2) 賽 sai3 (2) 勝 sing3 (1) 出 tshut4 (1) 超 tshiau1 (1) 7 強 kiong5 (2) 賽 sai3 (2) 出 tshut4 (1) 超 tshiau1(1) 勝 sing3 (1) 7

5

強 kiong5 (3) 賽 sai3 (2)

8

強 kiong5 (5) 賽 sai3 (1) 出tshut4(1) 超 tshiau1(1)

9

21

5

5 重 tang7 (1) 強 kiong5 (4)

強 kiong5 (4) 重 tang7 (1)

7

強 kiong5 (6) 重 tang7 (1)

0

4

強 kiong5 (2) 重tang7 (1) 親淺 tshin1 tshian2 (1)

強 kiong (4) 勝sing3 (1) 弱 jiok8 (1) 賽 sai3 (3) 5

Adverbial

A+可+Verb +B

Type IV

Zero-marked

A+Verb+B

Type III

126

24

23

21

24

34

Total

鈺妍麗錦 1572–1620 Yù Yán Lì Jǐn

金花女 (明萬 曆) 1572–1620 Jīn huā nǚ

滿天春1604 Măn Tiān Chūn

Text

Schema

0

0

0

寬khuann1(1) 大tua7 (1) 怒no.7 (1) 准(幼)iu3 (1) 嚴giam5 (1) 痛thong3 (1) 亂luan7 (2) 焦悴tsiau5tsui7 (1) 重tang7 (1) 緊kin2 (1) 11 幼iu3 (1) 勝sing3 (1) 冷ling2 (2) 細se3 (1) 薄pok8 (1) 6 紛紛hun1hun1 (1) 重tang7 (1) 疼thiann3 (1) 深tshim1 (1) 遠hng7 (1) 5

強 kiong5 (6)賽 sai3 (5) 勝 sing3 (1)

勝 sing3 (1) 賽 sai3 (1) 高kuan5 (1)

賽 sai3 (1)

1

3

12

Surpass-Similarity

II(bi) A+Verb+如+B

Surpass

II(a) A+ Verb + 過+B

A+比+B+ Verb

Compare

Type II

Type I

0

0

0

II(bii) A+Verb+ 似+B

Table 2.4  The four comparative structures in seven local operas from the Ming and the Qing periods

0

1

强 kiong5 (1)

5

賽 sai3 (2) 重 tang7 (1) 超tshiau1 (1) 勝 sing3 (1)

Zero-marked

A+Verb+B

Type III

0

0

0

Adverbial

A+可+ Verb +B

Type IV

6

10

28

Total

Total

同窗琴書記(清 乾隆) 1782 Tóng chuāng qín shū jì

蘇六娘(明萬曆) 1572–1620 Sū Lìuniáng

楊管歌 1572–1620 Yáng Guăn Gē

百花賽錦 1572–1620 Băi Huā Sài Jǐn

Text

Schema

10

29

0

41

7

大tua (1) 深tshim1 (2) 緊kin2 (1) 急kip4 (2) 6

3

賽 sai (9) 強 kiong5 (1)

2

0 大tua7 (1) 冷ling2 (2) 細se3 (1) 幼iu3 (1) 重tang7 (1) 急kip4 (2) 8

0

5

勝 sing3 (2)

1

疼thiann3 (1) 深tshim1 (1) 遠hng7 (2) 緊kin2 (1)

賽 sai3 (1)

0

0

0

0

Surpass-Similarity

II(bi) A+Verb+如+B

Surpass

II(a) A+ Verb + 過+B

A+比+B+ Verb

Compare

Type II

Type I

0

0

0

0

0

II(bii) A+Verb+ 似+B

10

1

賽 sai (1) 3

0

0

0

0

強 kiong (1) 重 tang7 (1) 2

0

0 5

Adverbial

A+可+ Verb +B

Type IV

1

賽 sai3 (1)

Zero-marked

A+Verb+B

Type III

80

17

10

2

7

Total

16  Hilary Chappell, Alain Peyraube and Song Na The relevant information on the corpus is presented below, while details on these sources including available reprints are to be found in the reference list at the end: I The folk operas of the Lì Jìng Jì 荔鏡記 and the Lìzhī Jì 荔枝記 in ­Chinese-character text:   1 Lì Jìng Jì 荔鏡記 (The romance of the mirror and the litchi) 1566 Jiājìng 嘉靖 edition but presumably written under the Yuan dynasty of 14th century or at the latest in the beginning of the Ming (14th – 15th c.). Supposedly written in both Cháozhōu and Quánzhōu dialects.  2 Lì Zhī Jì 荔枝記 (The romance of the litchi) 1581 Wànlì 萬曆 edition. Written in the Cháozhōu dialect.   3 Lì Zhī Jì 荔枝記 (The romance of the litchi). 1651 Shùnzhì 順治 edition. Written in both the Cháozhōu and Quánzhōu dialects.   4 Lì Zhī Jì 荔枝記 (The romance of the litchi). 1831 Dàoguāng 道光 edition. Written in both the Cháozhōu and Quánzhōu dialects.  5 Lì Zhī Jì 荔枝記 (The romance of the litchi). 1884 Guāngxù 光緒 edition. Written in both Cháozhōu and Quánzhōu dialects. II Other local operas in Southern Min   6 Jīn huā nǚ 金花女 (Lady Jinhua) (1572–1620). Written in the Cháozhōu dialect.   7 Sū Liùniáng 蘇六娘 (Dame Su Liuniang) (1572–1620). Written in the Cháozhōu dialect.   8 Yù Yán Lì Jǐn 鈺妍麗錦 (1572–1620) (Elegant brocades of precious beauty).   9 Băi Huā Sài Jǐn 百花賽錦 (1572–1620) (Exquisite brocades of a hundred flowers). 10 Yáng Guăn Gē 楊管歌 (1572–1620) (Yangguan music). 11 Măn Tiān Chūn 滿天春 (1604) (All-embracing spring). 12 Tóngchuāng Qínshūjì 同窗琴書記 (1782) (The Romance of the lute and the classmates). Written in the Quánzhōu dialect. III Spanish missionary materials 13 Doctrina Christiana en letra y lengua china (The Christian doctrine in the alphabet and language of China), blockprinted, ca.1597–1605, Manila.5 14 Arte de la lengua Chiõ Chiu (Grammar of the Chiõ Chiu language), 1620–1621, manuscript.6

2.1  The data In the five extant versions of the Lì Jìng Jì 荔鏡記 and its associated text, the Lì Zhī Jì 荔枝記, we find 126 instances of the four types of comparative constructions,

Comparatives of inequality in Southern Min  17 Table 2.5  The four comparative structures in the Doctrina Christiana (1607) Schema Type I

Type II

A+比+B+ Verb

II(a) A+ II(bi) A+Verb+ Verb + 如+B 過+B

Compare

Surpass

0

3

勝sing (6)

II(bii) A+ Verb+似+B

Type III

Type IV

A+Verb+B

A+可+ Verb +B

Surpass-Similarity

Zero-marked Adverbial

0

0

0

0

6

noting that the only four examples of Type I with bĭ比 occur in negated sentences, which we discuss next. In another seven local operas recounting a different set of stories from the Lì Jìng Jì, there is a total of 80 comparatives, distributed in the following manner: Finally, the comparative of inequality was investigated in the religious text, the Doctrina Christiana. Over all thirteen texts (12 operas and the Doctrina Christiana), there were 212 examples of the four different structural types of comparatives: Type I, Compare, had just 4 examples (which is actually the same example in four different versions) but these were all negated (§ 3.1); Type II was in fact the most numerous with 141 examples, 59 belonging to the Surpass cognitive schema with guò 過 and 82 belonging to the Surpass-Similarity subtype with rú 如 or sì 似 (§ 3.2). Type III, Zero-marked, had 46 examples (§ 3.3) and Type IV, Adverbial, had just 21 examples (§ 3.4). In the next section, we discuss each of the comparative structures one-by-one. 3 A DIACHRONIC EVALUATION OF THE HISTORICAL DATA FOR THE FOUR COMPARATIVE STRUCTURES

3.1 Type I: dependent-marking-Compare comparative: A + 比+ B + Verb (4/212) The status of the Type I construction, A+比+B+Verb, proves to be highly marginal, as although there are 174 examples of sentences in our corpus of thirteen Early Southern Min texts containing pi2 比 (=Mandarin bĭ), the majority prove to be examples of its lexical use as a full verb or as a noun, if not, of some other more minor functions, including its function as a preposition. Used as a verb, it has the meanings of ‘to compare with/to’ or ‘to give an example’ (87/174), as a noun including use in nominal expressions (63/174) and as a preposition ‘with’ or ‘according to’ (16/174) in addition to the four examples (4/174) where it is used as a comparative marker and three examples (3/174) that were excluded (see

18  Hilary Chappell, Alain Peyraube and Song Na Tables 2.6 and 2.7 in Appendix IV for more details). Examples of these main uses are provided below: (i) Pi2 continues its use in Southern Min as a full lexical verb meaning ‘to compare’, as it did in Medieval Chinese (3rd – 12th c.): (11) 林大

卜

比

陳三，

林大

不

值

一 文錢 . . . (22.052 Jiājìng 嘉靖) lim5tai7 beh4 pi2 tan5sann1 lim5tai7 m7 tat8 tsit8 bun5tsinn5 Lim.Tai want compare Tan.Sann Lim.Tai neg worth one penny ‘(If one) wants compare Lim Tai with Tan Sann, Lim Tai is not worth a penny . . .’

(ii) Another lexical use of pi2 is found in the form of a compound, bu5pi2 無比 ‘incomparable’, serving as an adjective: (12)

富 貴 是 無比 (4.016 Jiājìng 嘉靖) hu3 kui3 si7 bu5pi2 wealthy noble cop incomparable ‘To be wealthy and noble cannot be surpassed’.

The only examples of pi2 akin to a comparative marker of inequality are just the four examples where it co-occurs with a verb expressing a quality, hu7sim1 負心 ‘unfaithful’. Significantly, all these are found in the identical utterance and context in four of the five different versions of the Lì Jìng Jì 荔鏡記 and the Lì Zhī Jì 荔枝記 and, moreover, one that is negated by m7 不. We reproduce the sole Daoguang example (see also example (6) above from the Jiajing version). (13)

我 不 比 王魁 負心 pi2 ong5khue1 hu7sim1 gua2 m7 1sg neg cm Wang.Kui unfaithful ‘I am not as unfaithful as Wang Kui’.

(31.032 Dàoguāng 道光)

There are no examples of pi2 as a comparative marker found in any of the other seven operas, not to mention in the Spanish grammar, the Arte, nor in the Doctrina Christiana.7 Evidently, our conclusion is that there is no real use of pi2 for expressing a comparative of inequality of Type I in Early Southern Min.8 Hence, it is only in combination with a negative adverb that we find pi2 forming a comparative of inferiority ‘less than’. Since there are only four examples of this use across our thirteen main texts which span three centuries, it necessarily has to be considered an extremely marginal use. In fact, there is nothing surprising about this. The comparative structure with pi2 began to appear sporadically during the Tang dynasty (618–907) (there are, for example, only twelve occurrences in the Tang poems, and only one of these is an example of a real comparative marker, see §3.2.1 below). It only became an

Comparatives of inequality in Southern Min  19 established form in the Chinese language during the Qing period (1644–1911). In fact, we have to wait until the very end of the Ming dynasty for Type I to be seen in use as much as Type II (bi/ii) Surpass-Similarity with rú 如 or sì 似, as in the vernacular novels of Feng Menglong (1574–1645). The Surpass-Similarity comparatives represent the predominant structure during the Tang, Song, Yuan, and Ming dynasties. See Zhang 2004a; Chappell and Peyraube 2015.

3.2 Type II: head-marking comparatives: A + Verb + CM + B (141/212) The second type of comparative in which the verb is modified by a postposed comparative marker, has been divided into two subtypes, according to the cognitive schema involved. These are the Surpass and the Surpass-Similarity comparatives. 3.2.1 Type II (a): head-marking: the Surpass comparative: A + Verb + CM + B (59/141) Type II (a) with the Surpass schema is typically coded by ke3 過 in the Early Southern Min texts (=guò in Standard Mandarin) and represents a common comparative marker that is distributed across Central and Southern Sinitic languages today (see Li 2003). Although there is a large number of comparative examples in our corpus which use the marker ke3 過 (59/212), the great majority (56/59) co-occur with three main verbs: kiong5 強 ‘be better than’ (18 occurrences) #x003C; ‘be strong’, sing3 勝 ‘win’, ‘vanquish’ (13), sai3 賽 ‘compete’ (25), as in the following examples. (14)

孫 吳 (23.054 Jiājìng 嘉靖) li2 tsit8 ke3 sai3 ke3 sun1 ngo5 2sg dem strategy be.better cm Sun.Tzu Wu.Qi ‘Your strategy is better than those of Sun Tzu or Wu Qi’.

(15)

繼祖承宗

勝 過 求 名 (14.013 Jīn huā nǚ 金花女) ke3tso2sing5tsong1 sing3 ke3 kiu5 bing5 succeed.ancestor.succeed.ancestor be.better cm seek renown ‘It’s better to have descendants than to seek renown’.

(16)

正是 tsiann3si7 indeed

你

只

計

腹 pak4 stomach

賽

飢 ki1 hungry

過

添 thinn1 add

一 tsit4 one

斗， tau2 mw

強 kiong5 be.better

過 ke3 cm

十 年 糧粮 (30 Măn Tiān Chūn 滿天春) tsap8 ni5 niu5sit8 ten year food ‘It is said that giving a dou (6 kilos) of grains when one is hungry, is better than ten years of food’.

20  Hilary Chappell, Alain Peyraube and Song Na There are only three examples with a main verb other than these surpass-class verbs. These include siap4 澀‘be puckery, astringent’ (2) and kuan5 高 ‘high, tall’ (1). (17)



飯

心

食，

過 吞 砂 (21.028 Wànlì 萬曆) png7 bo5 sim1 tsiah8 siap4 ke3 thun1 sua1 food neg mood eat be.puckery cm swallow sand ‘I am not in the mood for food; it tastes more astringent than swallowing sand’.

(18) 為 ui7 for



無

著 功名 tioh8 kong1bing5 obtain renown

到 kau3 come

澀

只 tsit4 dem

路 來， lo7 lai5 road comedeictic

山嶺 suann1nia5 mountain

高 過 上 天台 (431.017 Jīn huā nǚ 金花女) kuan5 ke3 tsiunn7 thian1tai5 high cm climb heaven ‘In order to obtain worldly renown, I chose this way to come, but the mountains are higher than climbing up to heaven’.

Let us next digress to discuss the diachronic development of the Surpass comparative in Sinitic languages to provide a background for our analysis of guò in Early Southern Min. In terms of its history, the use of guò as a comparative marker is rare in Late Medieval Chinese. There is only one attested example in a Tang poem, as in (19) below, out of a total of thirteen uses. Peyraube (1989) claims that yú 於 #x003C; ‘at’ remains the most common comparative marker during the Tang period (7th – 10th c.), used in the structure ‘A+V+ yú 於+B. In the following example, both these markers are used in parallel clauses: (19) Type II structure with both Location and Surpass schemata



貧 於 楊子 兩 三 倍 老 過 榮 公 pín yú yángzǐ liǎng sān bèi lǎo guò róng gōng poor cm Yangzi two three times old cm Rong Master 六 七 年 (白居易詩：送劉五司馬赴任硖州兼寄崔使君) liù qī nián six seven year ‘Two or three times poorer than Master Yang, six or seven years older than Mr. Rong’. (Bai Juyi shi, 9th c.)

There are 12 examples of guò found in the Tang poems that could be interpreted as a comparative marker and only one in the Dūnhuáng Biànwén 敦煌變文 (Transformational texts). In this respect, Zhang (2005) claims convincingly that almost all of these guò are indeed verbs meaning ‘to surpass’, with the only exception

Comparatives of inequality in Southern Min  21 being the one above in (19) (Ōta 1958; Peyraube 1989; Peyraube and Lin 2011). This single example of guò as a comparative marker goes almost unnoticed with respect to the 376 examples of Type II which use the markers yú 於, rú 如, or sì 似 in Surpass-Similarity schemata in the same documents from Late Medieval Chinese. Zhang (2005) thus confirms Peyraube’s (1989) analysis and also concludes that guò never grammaticalized from a verb into a comparative morpheme in Medieval or even in Modern Chinese (13th – 18th c.). This is also the viewpoint of Wei (2007) and Wu (2010). Guò 過 meaning ‘to surpass’, or even shèng 勝 ‘to defeat’, would both have been very good candidates to replace yú 於 #x003C; loc as the comparative marker of inequality in Type II structures during the Late Archaic and Early Medieval periods.9 In this postverbal position, they did not undergo grammaticalization early enough to compete with the Type I Compare comparative (with bĭ 比) which had already arisen during the late Tang and was gradually replacing the other variants of Type II, in particular, the Surpass-Similarity cognitive schema, coded by rú 如 and sì 似. In contrast to this, in Southern China, where the Type I structure had not yet infiltrated, Chappell and Peyraube (2015) hypothesize that guò 過 did have ample time to grammaticalize into a comparative marker and so ultimately to replace the other markers used in the same syntactic structure, namely rú 如 and sì 似. Nevertheless, it could also plausibly be proposed that we have a case of ­contact-induced grammaticalization. Southern Sinitic languages, including Early Southern Min, may have borrowed this structure from a neighbouring language, or certain of the Southeast Asian languages could have borrowed it from Southern Sinitic, as Ansaldo (2010) has proposed, inter alia. This is because the Surpass comparative is one of the syntactic features (and cognitive schemata) which identifies a linguistic area comprising Tai-Kadai, Austroasiatic, Hmong-Mien, and many Sinitic languages. Chappell and Peyraube (2015) argue that the necessary historical evidence for establishing contact-induced grammaticalization in either direction is lacking. As observed earlier, only 3/59 examples in our corpus have adjectives preceding guò 過, while all the other examples with guò 過 combine with either kiong5 強, sing3 勝 or sai3 賽, which all mean ‘be better than’ and consequently belong to the lexical field of surpass verbs. The fact that guò 過 forms a verb complex, acting as the second verb attached to one of these three aforementioned verbs and that both V1 and V2 have the same meaning of ‘surpass’ recalls other diachronically significant structures found in Pre-Medieval Chinese (Han period, 206 bc – 220 ce), where V1V2 are both similarly filled by components from the same lexical field. For example, in an early form of the dative, the serial verb construction V1V2 is formed by two verbs of giving (Peyraube 1996), before V2 is later reanalyzed as a grammatical marker introducing the dative noun (or recipient). This seems to suggest for our data that guò 過 has not reached the final stage of conventionalization in the process of grammaticalization as a marker of the Surpass comparative but serves to lexically reinforce the first surpass-class verb in the majority of

22  Hilary Chappell, Alain Peyraube and Song Na 56/59 examples. In sum, the Type II (a) structure with surpass verbs is not a fullfledged comparative of inequality, given that we have only three authentic examples. It represents instead a lexical means of coding this constructional meaning. This contrasts with the use of grammaticalized surpass verbs as comparative markers in many contemporary Central and Southern Sinitic language such as Cantonese and other Yue dialects, as well as Hakka. 3.2.2 Type II (bi) and Type II(bii): Surpass-Similarity comparative: A + Verb + CM + B (82/141) There is a total of 82 Surpass-Similarity comparatives across the sample of thirteen texts, of which 78/82 use the comparative marker rú 如 and 4/82 uses sì 似. An example of this head-marking subtype follows: (20)



人

說

酒

腸

寬 如 海， 果然 (22 Măn Tiān Chūn 滿天春（下）) lang5 sueh4 tsiu2 tng5 khuann1 ji5 hai2 ko2jian5 people say alcohol intestine large cm ocean truly 色胆 大 如 天 sik4tann2 tua7 ji5 thinn1 desire big cm heaven ‘People say that his capacity for drinking is larger than the ocean; his sexual desire is plainly bigger than the sky’.10

These examples all conform to the structure found in Late Medieval Chinese (7th – 12th c.) using either rú 如 or sì 似, which became the foremost markers of the comparative of inequality during the Jin (1115–1234) and Yuan (1206–1368) dynasties (Peyraube 1989; Huang 1992; Zhang 2004b; Jiang 2011): (21) Comparative of inequality using Type II Similarity schema in Early Modern Chinese: npa– verb – cm(rú 如, sì 似) – npb 但 輕 如 你 底 (馬致遠 1250–1321:任風子) dàn qīng rú nĭ dĭ dem load light cm 2sg poss ‘This load is lighter than yours’. (Mă Zhìyuăn: Rèn fēngzǐ, 14th c.) 這 zhè

The majority of the examples in this category use an adjective that denotes a dimension or quality (79/82) while the remaining three (3/82) co-occur with sing3 勝 ‘win, vanquish’. The figures for the 82 relevant examples are listed below: (22) sing3 勝 (3) ‘win, vanquish’, tuan2 短 (1) ‘short’, kin2 緊 (7) ‘fast’, tinn1 甜 (1) ‘sweet’, bun7 悶 (4) ‘depressed’, tua7 大 (11) ‘big’, tang7 重 (8) ‘heavy’, tshim1 深 (11) ‘deep’, hng7 遠 (5) ‘far’, hu3 富 (2) ‘rich’, kong1 光 (1) ‘bright’, ling2 冷 (6) ‘cold’, khuann1 寬 (1) ‘large’, no7 怒 (1)‘angry’, iu3 准(幼) (3) ‘soft’,

Comparatives of inequality in Southern Min  23 giam5 嚴(1) ‘strict’, thong3 痛 (1) ‘pain’, tsiau5tsui7 焦悴 (1) ‘gaunt’, se3 細 (2) ‘thin’, hun1hun1 紛紛 (1) ‘numerous’, thiann3 疼 (2)‘pain’, kip4 急 (4) ‘hurry’, pok8 薄 (2) ‘indifferent’, bang5 忙 (1) ‘busy’, and luan7 亂 (2). (23)

悶 如 海 深 (Băi Huā Sài Jìn 百花賽錦 17) uat8ju2 thinn1tit4 gua2 bun7 tshim1 ji5 hai2 even.more add 1sg depress deep cm ocean ‘With even more added, my depression is deeper than the ocean’. 越自

添得

我

This Surpass-Similarity model (with Type II(bi) and Type II(bii) combined) is thus relatively common in our Early Southern Min texts. It conforms in any case to the same model used most frequently in Pre-Modern Standard Chinese during the Yuan dynasty and up until the end of the Ming, that is, from 13th to the end of the 16th century. Mainly based on the Northern dialects, it represents 61% of comparative constructions of inequality during the Yuan.11 It is only from the 17th century onwards that this model falls into disuse in the North to the advantage of Type I (Compare comparative with bĭ 比). It needs to be pointed out that examples belonging to Type II (bi) and Type II(bii) are often ambiguous. The comparative markers rú 如 and sì 似 may also be used to express the comparative of equality as well as the comparative of inequality (superiority subtype). See example (20) above. Nor is there anything surprising about this, since such was the case in the Late Medieval and Pre-Modern standard language, at least from the 10th century onwards (see Peyraube and Wiebusch 1995). It is probable, moreover, that from the Yuan dynasty (13th c.), the large majority of comparatives with the form (A)+Adj+rú/sì+B were used to express the comparative degree of inequality (superiority subtype), simply due to the fact that the canonical form in Classical Chinese for the expression of this relation of superiority with yú 於, (A)+Adj+yú+B, had clearly fallen into disuse and there was no other means available. On the other hand, to express the relation of equality, the form with the comparative marker between the two terms of comparison, A and B, and preceding the adjective became common; that is, the form, A+rú/ sì+B+Adj, as in the following example: (24) Comparative of equality using the Similarity schema: npa– cm – npb – vp, comparative marker = rú 如: 臉 如 紅 杏 鮮 妍 (Xiăo Sūn Tú 小孫屠, 14th c.) liǎn rú hóng xìng xiǎn yán face cm red apricot fresh beautiful ‘(Her) face is as fresh and beautiful as a red apricot’.

These are indeed the cases which deserve to be discussed, for any ambiguity has entirely disappeared. Furthermore, such examples, attested from the Song (10th – 13th c.) make use of a very limited number of adjectives or verbs including qiáng

24  Hilary Chappell, Alain Peyraube and Song Na 強, shèng 勝, and sài 賽, for which we observed above that they had lost a good part of their original lexical meaning in acquiring that of ‘be better than’, instead being used for unequivocally expressing the comparison of inequality (superiority subtype) in the construction (A‒qiáng/shèng/sài‒rú/sì‒B Verb). These cases are numerous in Late Medieval and Pre-Modern Standard Chinese, as Peyraube and Wiebusch (1995) have pointed out; see also Zhang 2005 for more details. And such is the case in our Early Southern Min texts.

3.3 Type III: Zero-marked comparative: A + Verb + B (46/212) According to Li (2003), the zero-marked comparative with its comparee (A) and standard nouns (B) separated simply by the verb is to be found in all branches of Sinitic. It is also a common pattern in Classical Chinese (Late Archaic period, 5th – 2nd bc) and thus one that has firmly ‘taken root’ in this language taxon. The verb is typically an adjective expressing a physical dimension such as height, size, weight, or age and as a comparative, it can be classified as belonging to the Action schema, which is semantically transitive and also includes the Surpass schema (cf. Heine 1997). This absence of any comparative marker may be linked to the intrinsically comparative nature of adjectives in their absolute form in many languages. Hence, Mandarin tā dà 他大 (3sg-tall) does not only objectively mean ‘He is tall’ but also implicitly that ‘He is tall with respect to someone else or to others’, that is, to some shared benchmark for tallness (see also Huddleston and Pullum 2002:1161–1162 for the case in English). An example follows from contemporary Hui’an Southern Min (Quan-Zhang subgroup): (25)

我 肥 汝 ua3 pui2 lɯ0 1sg fat 2sg ‘I am fatter than you’.

In a majority of Sinitic languages, however, unlike Hui’an Southern Min and unlike Classical Chinese, this comparative structure is accompanied by an obligatory measure phrase: (26) Standard Mandarin NPA Verbstative NPB Q-CL 哥哥 大 我 三 歲。 gēgē dà wŏ *(sān suì) brother old#x003C; ‘big’ 1sg three year ‘My brother is three years older than me’. In the Early Southern Min data, this structure is less frequent than ke3 過 (59), having only 46/212 examples across the texts consulted. Furthermore, it does not

Comparatives of inequality in Southern Min 25 conform entirely to the specifications of this construction in contemporary Sinitic languages: only two examples out of the total of 46 present a case of adjectival main verbs, both formed with the verb tang7 重 ‘be heavy’, ‘be important’ (see (27) and (28). Nonetheless, it could also be considered that some uses of tang7 重 belong to the lexical field of surpass, at least as much as those of tshut4 出 ‘go out’ do, which also falls into this group.12 The two examples in question are reproduced below: (27) 我 gua2 1sg 只 tsit4

只 tsit4 dem

羅巾 lo5kin1

羅巾， lo5kin1 scarf 情 tsing5

爾 li2 2sg 重 tang7

收記 siu1ki3 keep.remember 千 金 tshian1 kim1

在 心， ti7 sim1 at heart (4 Măn Tiān Chūn 滿天 春)

scarf love important thousand gold ‘My scarf, keep it in your heart; the love represented by this scarf is more important than a thousand pieces of gold’. (Măn Tiān Chūn 4)

dem

(28)

救 阮 一 命 重 再生 (Yáng Guăn kiu3 gun2 tsit4 mia7 tang7 tsai3sing1 Gē 楊管歌)13 save 1sg.gen one life be.important reborn ‘Your saving my life is more important (to me) than a rebirth’.

The other 44 are all verbs which belong to the lexical field of surpass once again, including kiong5 強 (18), sing3 勝 (4), sai3 賽 (14), tshut4 出 (3), and tshiau1 超 (4), if not ‘to be worse than’, jiok8 弱 (1). Here are two such examples with nongrammaticalized verbs with this general meaning of ‘to surpass’, or more simply, ‘to be better than’: (29)

咱

今冥

( 14 Măn Tiān Chūn   滿天春 (上)) lan2 kim1mi5 sing3 jip8 tong7pong5 1pl.incl tonight be.better.than enter bridal.chamber ‘For us this night is better than the wedding night’.

(30)

秀才

阿，

勝

救

入

阮

洞房

命

強 再生 (Yáng Guăn Gē 楊管歌) siu3tsai5 ah4 kiu3 gun2 mia7 kiong5 tsai3sing1 bachelor prt save 1sg.gen life be.better.than reborn ‘Bachelor, you saved my life, which is better than being reborn’.

Hence, we conclude that the zero-marked comparative is a marginal structure as far as the Early Southern Min local operas and other historical documents are concerned. The structure as we know it from modern Sinitic languages and also from Archaic Chinese is not at all put into service in these texts but

26  Hilary Chappell, Alain Peyraube and Song Na rather a lexical form is preferred, chosen from within the semantic domain of surpass verbs. Could this in fact be seen as a precursor or early form of the modern zeromarked comparative? Our judgement is negative on this point. It would be ­counter-intuitive to argue that a semantic generalization has taken place whereby the main verb slot expands from just a small number of surpass verbs to include adjectives coding a semantically specific dimension. This would be quite the reverse process for any hypothesis of grammaticalization. Reinforcing our argument is the fact that this structure has existed since the Archaic Chinese period and without any recourse to the use of such a group of surpass verbs.

3.4 Type IV: Adverbial comparative: A + khah4 可 + Verb + B (21/212) Type IV is claimed to be a rare comparative strategy in the languages of the world (see Haspelmath and the Comparative Constructions Consortium 2013) but one that is, nonetheless, attested in Southern Min and in Hakka, among the Sinitic languages. Furthermore, it seems that this syntactic form is only found in these two branches. In Hakka dialects, the preverbal adverb is frequently a cognate of guò 過 ‘more’, while in Min dialects it is typically a cognate of jiào 較 ‘more’, represented by various characters, homophonic to its pronunciation in the given dialect. These include khah4 可 in our historical texts, as well as 恰 in various reference grammars. This fourth type of comparative belongs to the Action schema as well. It makes use of an adverb as its comparative marker with the basic meaning of ‘more’, khah4 較 in Southern Min dialects, which occurs in a transitive clause to code the comparative meaning. An important feature of the adverbial comparative is that, like the Type III Zero-marked comparative, it is mainly used with adjectives such as ‘tall’, ‘old’, ‘fat’, and ‘rich’; see (25) and (26) above. Despite this, the clause has transitive SVO syntax which normally is not permitted with predicative adjectives since the latter pattern in the same way as intransitive: they do not normally take a direct object argument. In any case, these special features allow us to identify the construction as a comparative, namely, the transitive syntax with a main verb that comes typically from the adjective class, used in conjunction with a comparative marker derived from a degree adverb ‘more’: AdverbCM + Verbstative. (31) Southern Min (Hui’an) NPA CMmore Verbstative NPB 伊 恰 富 我 i33 khaʔ4 pu55 ɡua55 3sg morecm rich 1sg ‘She is richer than me’. (Literally: she-more-riches-me) This comparative type can be considered as an identifying feature of the contemporary Min group of dialects, as it is found in most of its subgroups: in Mindong (Eastern Min), in Minbei (Northern Min), and in Minzhong (Central Min) (Wu

Comparatives of inequality in Southern Min  27 2012). In fact, Wu (2012) claims it is a shared innovation of the Min dialects. In the historical documents, the character 可 is predominantly used to represent this morpheme. In spite of this, its source is likely to be khah4 較, an adverb which means ‘more, a bit, fairly’. In the early 17th century Southern Min grammar, the Arte de la lengua Chiõ Chiu, composed by the Spanish missionary Melchior de Mançano, kha3 可 is given as the particle which forms the comparative, by placing it in front of the positive or plain form of the adjective: (32)

汝 可 好 我 lù cǎc hò guà 2sg cm good 1sg ‘You’re better than me’. (eres mejor q̃-yo) (Arte: folio 5b)

Piñol y Andreu (1928:82ff) is also of the view that for the Southern Min dialect of Amoy 廈門 (Xiamen), as spoken in the early 20th century, this form of the comparativo de superioridad ‘comparative of superiority’ is the most common in use, being constructed with the adverb khah ‘more’ preceding the positive form of the adjective, as in Spanish.14 Its comparative and adverbial uses have also been noted by Douglas (1853:258). He cites the following example: (33)

汝 較 肥 我 lí khah pâi guá 2sg cm fat 1sg ‘You are fatter than me’.

We find a small number of different stative verbs that may act as the main verb in the data from the operas, including tshin1tshian2 親淺 ‘beautiful’ (1), tang7 重 ‘heavy’ (4), and kiong5 強 ‘strong’ (3). Here are two such examples: (34)

官人 kuann1lang5 gentlemen

許 hi2 dem

前頭 tsing5thau5 front

一 tsit8 one

陣 tin7 group

娘仔， niu5a2 woman

生得 senn1tit4 be.born.to

句 可 親淺 伊 (8.069 Jiājìng 嘉靖) koh4 khah4 tshin1tshian2 i1 even cm beautiful 3sg ‘There are some beautiful women in front of the gentleman who are even more beautiful than her’. (35)

伊 有 力氣， 向 可 強 我 i1 u7 lat8khi3 koh4 khah4 kiong5 gua2 3sg have strength even cm be.strong 1sg ‘He is young and strong, he is even stronger than me’.

(15.033 Guāngxù 光緒)

28  Hilary Chappell, Alain Peyraube and Song Na Example (35) presents a case of kiong5 強 being used as an adjective ‘strong’ in this construction, rather than as a surpass verb ‘be better than’, as we saw is also possible when it occurs in the Type II Surpass comparative with ke3 過 or in the Type III Zero-marked comparative. There are in fact sixteen such examples of kiong5 強 across the thirteen texts where it co-occurs with khah4 較. However, in only 3/16 examples do khah4kiong5 可強 mean ‘stronger than’ while the remaining 13/16 have the interpretation of ‘be better than’. Therefore, in total, only 8/21 predicates are used with their basic lexical meaning in this comparative type, since the majority, the remaining thirteen (13/21), are represented by precisely these semantically extended uses of kiong5 強, as in example (36) below: (36) 有 人 愛 可 強 無 人 愛 (17.114 u7 lang5 ai3 khah4 kiong5 bo5 lang5 ai3 Shùnzhì have people love cm be.better neg people love 順治) ‘To be loved is better than not to be loved by others’. In terms of lexical main verbs, the Type IV Adverbial construction appears to be the most productive, compared with the other three types, which are either marginal, as for the Type 1 Compare comparative, or in an incipient state of development, as for the Type II Surpass comparative, given that it mainly uses lexical verbs from the surpass class, and similarly for the Type III Zero-marked comparative. Incidentally, Li (1994) independently remarks that between the 16th and the 19th centuries, the Type IV comparative was one of the most commonly used in Southern Min, adding that the adverb khah4 had completely lost its original lexical meaning to become purely a comparative morpheme. He quotes example (34) above to support his claim. Chen (2015) puts forward the very interesting conjecture that the origin of the Adverbial comparative in Southern Min could be the absolutive form of the comparative which includes just the comparee noun, the degree adverb khah4較 and the main verb: (37)

伊 恰 巧 點仔 i1 khaɁ7–8 kiau3 tiəm3–2a3 3sg cm clever a.bit ‘He is a bit cleverer (than me)’. (Chen’s example from the Hui’an 惠安 dialect of Southern Min, Quanzhou county)

One could also claim that Type IV appeared in Early Southern Min under the influence of the common language of the period (gòngtóngyŭ 共同語) which already used a preverbal comparative marker (Zhang’s hypothesis 2005), as in the case of bĭ 比. However, we do not agree with Zhang’s viewpoint, since the use of the bĭ 比-comparative is not found in any of our thirteen texts which span the period from the 16th to the 19th centuries. Furthermore, the comparative morpheme

Comparatives of inequality in Southern Min  29 bĭ 比 derives from a verb which introduces the standard noun phrase in preverbal position, never functioning as an adverbial modifier of the main verb, while this is the precise source of khah4 較.

3.5 Conclusion In this analysis of the comparatives of inequality used in Early Southern Min, four main construction types have been described and discussed. All four express the superiority dimension of ‘more than’ for inequality comparatives, there being no special syntactic form to express the comparative of inferiority: ‘less than’. Structure

Schema

Type I: Type II(a): Type II(bi/ii): Type III: Type IV:

A + CM比 + B + Verb A +Verb + CM過 + B A + Verb + CM如/似 + B A + Verb + B A + CM較 + Verb –+ B

(Compare comparative) (Surpass comparative) (Surpass-similarity comparative) (Zero-marked comparative) (Adverbial comparative).

Among these, we have shown that head-marking comparatives which belong to two main types (including subtypes) are the most frequent and favored forms. Furthermore, these are the syntactic structures for the comparative which harmonize with SVO word order: Type II(a): Type II(bi/ii): Type IV:

A + Verb + CM過 + B A + Verb + CM如/似 + B A + CM較 + Verb + B

(Surpass comparative) (Surpass-similarity comparative) (Adverbial comparative)

Nonetheless, there are some further important points to be noted about all four major comparative types examined. Verbs coding the state of superiority such as kiong5 強 ‘be better than’ #x003C; ‘strong’, sing3 勝 ‘be superior to’ #x003C; ‘vanquish, win’, and sai3 賽 ‘surpass’ #x003C; ‘compete’ are favored above all in the structure corresponding to Type III, the Zero-marked comparative, A + Verb + B, for which we found only two adjectives (2/46) used as the main verb. Therefore, this type cannot be considered as a full-fledged construction during the Early Modern period for Southern Min. Its significance lies in the fact that these lexical verbs are all sources of grammaticalized markers of the comparative in contemporary Sinitic languages and dialects (cf. Huang 1996:678–681). Type II (a), the Surpass comparative with the structure: A + Verb + CM过 + B, has a main verb which, similarly to Type III, belongs to the lexical field of surpass verbs meaning ‘be better than’ or ‘be superior to’. Hence, this verb should also not be deemed to be a fully conventionalized comparative of inequality. Only a few examples are to be had, 3/59 to be exact, where the main verb does not code the notion of surpass itself. Moreover, this comparative model never gained the upper hand, since it has totally disappeared in contemporary Southern Min.15 As

30  Hilary Chappell, Alain Peyraube and Song Na observed above, guò 過 undoubtedly retained its verbal status as V2 reinforcing its preceding surpass-class verb in a serial verb construction V1-V2 and never grammaticalized into a comparative marker, as was the case for other Southern Sinitic languages. Type II (bi/ii) appears to be a form that is linked to the literary register. This construction using rú 如 or sì 似 as a comparative marker clearly adopted the model which was standard for the common language of the time and which prevailed from the end of the Song dynasty (1279 ce). Moreover, we find the same properties in Early Southern Min which characterized this form in the Yuan (1206–1368) and the Ming (1368–1644) – the frequent use of verbs or adjectives which themselves possess the meaning of ‘surpass’, combined with the ambiguity of certain examples for which it is difficult to know if the comparative marker expressed superiority or inequality). Type II (bi/ii) was maintained until the 20th century, side-by-side with Type IV which subsequently prevailed in Southern Min while it continued to evolve, resulting in a marginalization rather than a complete disappearance, of Type II (bi/ii). It can also be incidentally observed that Type II (bi/ii) has been maintained right up until today in the dialects of Shandong. Luo (1992) thus notes that the form with the postverbal comparative marker sì 似 (but also qĭ 起 or de 的) remains used just as much as Type I with the preverbal comparative marker bĭ 比. He also remarks that speakers of these dialects do not use these two forms as simple alternatives. In other words, they are often in complementary distribution, depending on the nature of the verb which precedes the comparative markers. Such a differentiation is not attested in our texts from Early Southern Min. Type IV, the Adverbial comparative, makes use of the largest number of different adjectival main verbs which do not lexically code superiority or the notion of surpass. Hence, even though the number of corpus examples is small, this would seem to suggest that the Adverbial comparative was already well-established in the earliest period for which we have texts for Southern Min. It is also the syntactic structure for the comparative which is described in the early 17th century Spanish grammar of Southern Min, the Arte (see example (32)). Finally, the Type I Compare comparative with the structure A + CM比 + B + VP does not exist in our texts for Early Southern Min, except in negated form for four sentences. It is an entirely marginal structure. 4  COMPARATIVES IN CONTEMPORARY SOUTHERN MIN In this section, we describe two main types used in contemporary Southern Min: the Type IV Adverbial with khah4 較 ‘more’ and a new hybridized form which does not occur in our Early Southern Min texts. The number of different structures for the comparative of inequality in Southern Min dialects may be very high, as Chen (2015) ably demonstrates for the Hui’an variety 惠安話 (Quan-Zhang subgroup). Hui’an makes use of six different forms, noting that the purely Mandarin Type I with pi2 比 is the least frequent. By way of contrast, this is precisely the structure which Lien (1999)

Comparatives of inequality in Southern Min  31 shows to be increasingly favored by younger generation speakers of another variety of Southern Min, Taiwanese, in the highly urbanized Hsin-chu area of northwestern Taiwan. In both these varieties of Southern Min, one of the high frequency comparatives is an unusual structure that proves to be the outcome of inter-dialectal contact: this is the hybridized comparative in Southern Min which uses both markers pi2 比 and khah4 較 ‘more’, that is, double marking. It has the form: NPA [CM(i)比 NPB]CM(ii)較Verb. In other words, it is a blend of the borrowed Northern Mandarin structure which uses the comparative marker bĭ 比 ‘compared to’, a preposition in a prepositional phrase preceding the predicate, while retaining the characteristically Southern Min comparative marker, khah4 較 ‘more’, in its VP constituent where it directly modifies the verb, originally as an adverb of degree, but in this form functioning as a second comparative marker. This results in a new hybrid form, Type V, with double marking, that is, a combination of Types I and IV, the Compare and the Adverbial comparatives. Inter-dialectal contact thus leads to hybridization through the incorporation of the native comparative marker into a new syntactic structure, borrowed from Mandarin. It further reveals the interesting phenomenon of the co-existence of both dependent- and head-marking strategies in one and the same structure. A simple elicited example is provided below to show the structure clearly in (38) and an example from authentic discourse is given in (39): Type V: NPA [CM(i)比 NPB] CM(ii)較Verb Hybridized comparative construction in Taiwanese Southern Min (38)

我 比 伊 較 gua2 pi2 i1 khah4 1sg cm(i) 3sg cm(ii) ‘I am taller than him’.

(39)

有 阮遐 價數 喔， 可能 比 板橋 gun2hia5 ke3siau3 oh kho2ling5 u7 pi2 pang1kio5 1pl.excl-there price prt probably have compare Pang.kio 市裏 較 貴 淡薄 啊。(Jesse’s narratives, lines 485–487) tshi7lai2 khah4 kui3 tam7poh8 a city-inside more expensive a.little prt ‘As for our prices, compared with downtown Pang-kio, they were probably a little more expensive’.

高 kuan5 tall

In a sociolinguistic survey of the Hsin-Chu variety of Taiwanese Southern Min, which has developed on the basis of the mainland Quanzhou dialect, Lien (1999) shows that while two structures prevail, the comparative marker khah4 較 is in the process of disappearing from both hybridized forms, noted by its optionality in the syntactic patterns given in (40).

32  Hilary Chappell, Alain Peyraube and Song Na (40) (i) (ii)

A phing22 并 A pi53 比

B B

(khah4 較) (khah4 較)

predicate predicate

Furthermore, the local marker phing22 并 in this variety of Southern Min is also slowly being replaced by pi53 比. As remarked upon above, younger generation speakers who have received higher education, either to senior high school or university level, favor the purely Mandarin form with only the marker pi53 比 for which tone change is in progress from the Southern Min value pi53, which is the high falling citation Tone 2 or Yángpíng, to pi214, the contour Shăngshēng Tone 3 of Standard Mandarin. Older generation male speakers favour the form with only the marker phing22 并, while older generation female speakers prefer the form with both phing22 并 and khah4 較. Nonetheless, the frequency of use of the second hybridized form with pi53 比 and khah4 較 remains relatively stable, according to Lien’s survey of 150 speakers. In a considerable number of Hakka dialects, we also find cases of the Type V Hybridized Comparative: for example, in Meixian (梅縣) Hakka, the prestige variety spoken in northeastern Guangdong province (see Wu 2012 for a list). The two comparative markers are the Mandarin preposition pi31 比 ‘compared to’, once again, in a position preceding the predicate, and the Hakka adverb kuo53–55 過 ‘more’ directly modifying the verb, a marker which is distinct from that found in Early Southern Min. Its adverbial meaning of ‘more’ is undoubtedly derived from the verbal meaning of ‘surpass’ or ‘exceed’ of kuo53–55 過. (Note however that this is not a case of the Surpass comparative, in which the comparative marker, kuo53–55 過, would necessarily follow the main verb and not precede it.) (41) Hybridized comparative construction in Meixian Hakka NPA [CM(i) NPB] CM(ii)more 這隻 比 那隻 過 e31tsak1 pi31 e53–55tsak1 kuo53–55 dem-clf cm(i) #x003C;‘compare’ dem-clf cm(ii)more#x003C;‘surpass’ ‘This one is better than that one’. (Literally: this-one-compared:to-that-one-more-good)

Verb 好 hau31 good

Li (2003) also remarks upon the fact that in Hakka dialects of the Fēngshùn district in Guangdong province (廣東豐順), there are hybrid forms of the type NPA + CM(i) +V + CM(ii) + NPB using the comparative markers 較 as CM(i) and 過 as CM(ii). For example: (42)

梅縣 較 冷過 湯坑 méixiàn jiào lěngguò tāngkēng16 Meixian cm(i) froid-cm(ii) Tangkeng ‘It’s colder in Meixian than in Tangkeng’.

Comparatives of inequality in Southern Min  33 This could represent a sixth type which hybridizes Type IV Adverbial with Type II Surpass. More research is needed to confirm its distribution and characteristics. The Type V hybridized comparative is thus essentially a feature of Southern Min and some of the Hakka dialects, while Type VI might turn out to be a feature of just certain Hakka dialects.17 We find many examples of the first type in the corpus data from present-day Taiwanese Southern Min but none in the 12 operas which span the late Ming dynasty and most of the Qing dynasty (late 16th-late 19th centuries) and whose texts have been thoroughly scoured for all the comparative structures in Early Southern Min. Nor do we find it in any of the late 16th and early 17th century missionary documents such as the Doctrina Christiana. This seems to suggest that the hybridized form is a very recent development. It remains an important issue for investigation to see if it is at all possible to pinpoint when the hybridized form first began to be observed and noted by scholars in 20th century documents concerning the Southern Min language. In conclusion to our diachronic study, a final word is in order to highlight the clear historical preference for Southern Min languages and dialects to use two major head-marking types of comparative, namely, the Type II Surpass and Type IV Adverbial structures. Both of these can be classified as belonging to the transitive Action schema in Heine’s (1997) typology of comparatives of inequality. The change to the purely dependent-marking structure with pi53 比 of Northern Chinese in the form of the Type I Compare comparative is only a relatively recent phenomenon of the late 20th century, and is potentially taking place via hybridized structures, as hypothesized by Lien (1999), which first retain the native comparative markers before eventually omitting these. Needless to say, the native preference for Type II Surpass and Type IV Adverbial structures evidently corresponds to the comparative structures which preserve harmony with SVO word order, in contrast to the Northern Sinitic form with the dependent-marking bǐ. NOTES * The research leading to this paper has received funding from the French Agence Nationale de la Recherche (ANR) for the blue-sky project entitled Diachronic change in Southern Min – a Sinitic language (DIAMIN) (Programme blanc, ANR-08BLAN-0174), Principal investigator, Hilary Chappell and Co-investigator, Alain Peyraube. We would also like to thank Professor Chinfa Lien and his team members who have been our Taiwanese partners in undertaking this bilateral cooperative project with the National Science Council of Taiwan (NSC)/Ministry of Science and Technology, as well as Professor Zeng Nanyi and Lai Yunfan for their comments and enlightenments on various aspects of the Southern Min texts. 1 To express a comparison of inferiority, Chinese (Mandarin as well as Southern Min) relies on negation of the comparative of superiority. 2 For the corpus of Early Southern Min, we simply provide the figures for each type of comparative structure, as found across the thirteen texts. Necessarily, there is repetition of examples for the five texts based on the Romance of the Litchi or the popular story of Chen San and Wu Niang. We do not aim to carry out a statistical analysis of the corpus examples but rather are interested in patterns of diachronic change reflected in these five texts which span three centuries.

34  Hilary Chappell, Alain Peyraube and Song Na 3 This is not a comparative of inequality but a negation of a comparative of equality. 4 In this Type III, the VP has more than often the meaning of ‘be better (than)’. See §3.3. 5 Different copies of the Doctrina Christiana are held at the Vatican Library and at the British Library. The translation of the Spanish version of the Doctrina Christiana into a Min dialect in character form is attributed by van der Loon (1966) to Dominican missionaries in the Philippines, including Fathers Juan Cobo (ca. 1592) and Miguel Benavides (1550–1605), with the aid of unknown Chinese collaborators. We make use of the reproduction by van der Loon (1966) which collates the character text with two romanized versions. The page numbers for all examples from this document refer to this edition. The Doctrina Christiana was intended for use in missionary work to propagate the Christian faith to the Chinese community in Manila and comprises prayers, articles of faith, the Ten Commandments and the catechism among other items. See also Chappell and Peyraube 2014. 6 This is a handwritten manuscript held in different versions at both the University of Barcelona library and the British Library. The date is given as: 萬曆四十八年 Wàn Lì sìshíbā nián ‘the forty-eighth year of the reign of the Emperor Wan Li, which is 1620 or 1621 when converted to the Gregorian calendar. We thank Professor Mei Tsu-Lin of Cornell University for bringing this to our attention. There are at least two versions of the Arte in existence. Only the University of Barcelona version includes this page and not the British Library copy. Note that we have not counted examples from the Arte in the tables for the simple reason that it is not a prose text but a grammar, just like Piñol y Andreu’s (1928) grammar. 7 There is in fact only one example in the Doctrina Christiana: 無有可比 (no romanization) used to translate the Spanish original sin comparación ‘without comparison’ (Section 23a, van der Loon 1966:175). 8 Lien (1999) and Yue-Hashimoto (1999) have both also observed its absence as a comparative marker for different set of sample texts. The adoption of this comparative form is one of the main topics in Wu 2012 who, relying on Chen 2010, asserts that in works from the late 19th century in the Fuzhou dialect, examples of pi2 are very rare (there is in fact only one occurrence across two texts). 9 We also find several cases where it is rather shèng 勝 that has the function of a comparative marker (there are 23 examples to be found in the Tang poems, according to Zhang 2005), as in: 湖

魚

香

勝

肉， 官

酒 重 於 餳 (劉禹錫：曆陽書事七十韻) hú yú xiāng shèng ròu guān jiǔ zhòng yú xíng lake fish delicious cm meat government alcohol precious cm malt.sugar ‘Fish from the lake is more delicious than meat; alcohol produced by the government brewery is more precious than malt sugar’. (Tang dynasty 唐長慶4年 824 ce). 10 This example could also be interpreted as a comparative of equality: ‘People say that his capacity for drinking is as large as the ocean; his sexual desire is plainly as big as the sky’. 11 See the complete set of statistics, adapted from Huang 1992 in Chappell and Peyraube 2015. 12 The case is less clear for tshut4 出 than for tang7重, since the ‘surpass’ or ‘better than’ interpretation relies even more heavily on the context. We have also included jiok8 弱 in this group which has the basic meaning of ‘weak’ but in the comparative is interpreted as ‘be worse than’. 13 The authors have access to the digitalized version of this opera which explains the lack of a page number here. 14 Original Spanish: ‘. . . directamente con el adverbio khah (más) antepuesto al positivo, como en castellano’.

Comparatives of inequality in Southern Min  35 15 This is typical of Southern Min dialects within Fujian province and Taiwan. Many Southern Min dialects in Guangdong and Hainan provinces do use however the Surpass model, for example, Tunchang 屯昌, Leizhou 雷州, and Shantou 汕頭. In the latter case, this is certainly an areal feature shared with the co-territorial Yue and Hakka dialects. See Li 2003 and Wu 2012. 16 Mandarin pīnyīn in small capitals are used here as the IPA values are not given in the source article. 17 For more details, see Chen 2015, on this double-marked comparative in the Hui’an dialect of Southern Min, and also Lien 1999; Li and Lien 1995 for Taiwanese Southern Min. 18 This document has been attributed to the two missionaries named as authors in the reference list (see van der Loon 1966) who were undoubtedly aided by their Chinese collaborators who remain unknown to us. Van der Loon estimates the period for the completion of this religious text as between the years of 1597 and 1605, according to the historical sources he consulted.

REFERENCES

Primary sources Băi Huā Sài Jǐn 百花賽錦 (1572–1620) (Exquisite brocades of a hundred flowers). http://140.114.116.3/DB/index.php. Benavides, Miguel, and Juan Cobo. 1597–1605. Doctrina Christiana en letra y lengua china [The Christian doctrine in the alphabet and language of China]. Manila: Keng Yong. Blockprint.18 Jīn huā nǚ 金花女 (Lady Jinhua) (1572–1620). Lì Zhī Jì 荔枝記 (The Romance of the Litchi). 1831 Dàoguāng edition. Written in both the Cháozhōu and Quánzhōu dialects in Chinese characters. http://140.114.116.3/DB/ index.php. Mançano, Melchior de. 1620–1621. Arte de la lengua Chiõ Chiu [Grammar of the Chiõ Chiu language]. Ms. N°1027, University of Barcelona Library. Măn Tiān Chūn 滿天春 (1604) (All-embracing spring). http://140.114.116.3/DB/​index. php. Sū Liùniáng 蘇六娘 (Dame Su Liuniang) (1572–1620). Tóngchuāng Qínshūjì 同窗琴書記 (1782) (The Romance of the lute and the classmates). van der Loon, Piet. 1992. Míngkān Mǐnnán xìqǔ guǎnxián xuǎnpěn sānzhǒng 明刊閩南戲 曲管弦選本三種 [The classical theatre and art song of Southern Fukien]. Taipei: SMC Publishing INC. Wu, Shou-Li 吳守禮. 2001a. Míng Jiājìng kān Lìjìngjì xìwén jiàolǐ 明嘉靖刊荔鏡記戲 文校理 [The text of the opera – The romance of the Litchi. Ming Dynasty edition of the Emperor Jiajing, revised and annotated]. Taipei: Tsongyi Publishing. Wu, Shou-Li 吳守禮. 2001b. Míng Wànlì kān Lìzhījì xìwén jiàolǐ 明萬曆刊荔枝記戲文校 理 [The text of the opera – The romance of the Litchi. Ming Dynasty edition of the Emperor Wanli, revised and annotated]. Taipei: Tsongyi Publishing. Wu, Shou-Li 吳守禮. 2001c. Qīng Shùnzhì kān Lìzhījì xìwén jiàolǐ 清順治刊荔枝記戲文 校理 [The text of the opera – The romance of the Litchi. Emperor Shunzhi edition of the Qing Dynasty, revised and annotated]. Taipei: Tsongyi Publishing. Wu, Shou-Li 吳守禮. 2001d. Qīng Guāngxù kān Lìzhījì xìwén jiàolǐ 清光緒刊荔枝記戲 文校理 [The text of the opera – The romance of the Litchi. Emperor Guangxu edition of the Qing Dynasty, revised and annotated]. Taipei: Tsongyi Publishing.

36  Hilary Chappell, Alain Peyraube and Song Na Wu, Shou-Li 吳守禮. 2002a. Míng Wànlì kān Jīn Huā Nǚ xìwén jiàolǐ 明萬曆刊金花女 戲文校理 [The text of the opera –Lady Jinhua. Ming Dynasty edition of the Emperor Wanli, revised and annotated]. Taipei: Tsongyi Publishing. Wu, Shou-Li 吳守禮. 2002b. Míng Wànlì kān Sū Liùniáng xìwén jiàolǐ 明萬曆刊蘇六 娘戲文校理 [The text of the opera –Dame Sū Liùniáng. Ming Dynasty edition of the Emperor Wanli, revised and annotated]. Taipei: Tsongyi Publishing. Wu, Shou-Li 吳守禮. 2003. Qīng Qiánlóng kān Tóngchuāng Qín Shū Jì xìwén jiàolǐ 清乾 隆刊同窗琴書記戲文校理 [The text of the opera –The romance of the lute and classmates. Qing Dynasty edition of the Emperor Qianlong, revised and annotated]. Taipei: Tsongyi Publishing. Wu, Shou-Li 吳守禮. 2007. Míng Wànlì kān Lìjìngjì xìwén fēnlèi cíhuì 明嘉靖刊荔鏡記 戯文分類詞彙 [Thematic lexicon for the Ming Dynasty edition of the Emperor Jiajing of the tomance of the Litchi and the mirror]. Taipei: Tsongyi Publishing. Yáng Guăn Gē 楊管歌 (1572–1620) (Yangguan music). http://140.114.116.3/DB/index.php. Yù Yán Lì Jǐ 鈺妍麗錦 (1572–1620) (Elegant brocades of precious beauty) http://140.114.116.3/DB/index.php.

Secondary resources Ansaldo, Umberto. 2010. Surpass comparatives in Sinitic and beyond: Typology and grammaticalization. Linguistics 48:919–950. Chappell, Hilary. 2015. Linguistic areas in China for differential object marking, passive, and comparative constructions. In Diversity in Sinitic languages, ed. by Hilary Chappell, 13–52. Oxford: Oxford University Press. Chappell, Hilary, and Alain Peyraube. 2011. The comparative construction in Sinitic languages: Synchronic and diachronic variation. Paper presented at the International Workshop on Ecology, Population Movements, and Language Diversity, University of Lyon-2, 26–27 May. Chappell, Hilary, and Alain Peyraube. 2014. The history of Chinese grammars in Chinese and Western scholarly traditions. Language and History 57:107–136. Chappell, Hilary, and Alain Peyraube. 2015. The comparative construction in Sinitic languages: Synchronic and diachronic variation. In Diversity in Sinitic languages, ed. by Hilary Chappell, 134–154. Oxford: Oxford University Press. Chen, Weirong 陳偉蓉. 2015. Comparative constructions of inequality in the Southern Min dialect of Hui’an. In Diversity in Sinitic languages, ed. by Hilary Chappell, 248– 272. Oxford: Oxford University Press. Chen, Zeping 陳澤平. 2010. Shíjiǔ shìjì yǐlái de Fúzhōu fāngyán – Chuánjiàoshì Fúzhōu Tǔbái wénxiàn zhī yǔyánxúe yánjiū 十九世紀以來的福州方言-傳教士福州土白文獻 之語言學研究 [The dialect of Fuzhou in the 19th century – Linguistic research on missionary documents in vernacular language]. Fuzhou: Fujian Renmin Publishing. Dixon, Robert M.W. 2012. Basic linguistic theory. Vol. 3, Further grammatical topics. Oxford: Oxford University Press. Douglas, Carstairs. 1853. Chinese-English dictionary of the vernacular or spoken language of Amoy with the principal variations of the Chang-chew and Chin-chew dialects. Supplement by Thomas Barclay. London: Trübner & Co. Haspelmath, Martin, and the Comparative Constructions Consortium. 2013. The typology of comparative constructions revisited. Paper presented at the Association for Linguistic Typology 10th Biennial Conference (ALT10), The University of Leipzig, 15–18 August.

Comparatives of inequality in Southern Min  37 Heine, Bernd. 1997. Cognitive foundations of grammar. Oxford: Oxford University Press. Huang, Borong 黄伯榮. 1996. Hànyǔ fāngyán yǔfǎ lèibiān 漢語方言語法類編 [Concordance of Chinese dialect grammar]. Qingdao: Qingdao Chubanshe. Huang, Xiaohui 黄曉惠. 1992. Xiàndài Hànyǔ chābǐ géshì de láiyuán jí yǎnbiàn 現代漢語 差比格式的來源及演變 [The origin and the evolution of the comparatives of inequality in Chinese]. Zhōngguó Yǔwén 中國語文 3:213–224. Huddleston, Rodney, and Geoffrey K. Pullum. 2002. The Cambridge grammar of the ­English language. Cambridge: Cambridge University Press. Jiang, Nanxiu 姜南秀. 2011. Yuánběn Lǎoqǐdà zhōng de bǐjiào, bǐnǐ hé xuǎnzé wènjù yánjiū 原本老乞大中的比較, 比擬和選擇問句研究 [The study of comparatives, similes, and disjunctive questions in the oldest edition of the Lao Qida]. MA thesis, Northwestern Normal University, Lanzhou. Li, Chia-chun 李佳純. 1994. Mǐnnányǔ bǐjiàojù – lìshí jí leìxíng de tàntǎo 閩南語比 較句―歷時及類型的探討 [The comparative construction in Southern Min - diachronic and typological investigations]. Master thesis, National Tsing Hua University, Taiwan. Li, Chia-chun 李佳純, and Chinfa Lien 連金發. 1995. Lùn Mǐnnányǔ bǐjiàoshì leìxíng jí lìshí de tàntǎo 論閩南語比較式類型及歷史的探討 [On the comparative constructions in Southern Min - diachronic and typological investigations]. In Papers from the 1994 conference on language teaching and linguistics in Taiwan. Vol. 1, Southern Min, ed. by Feng-fu Tsao and Meihui Tsai, 71–87. Taipei: Crane Publishing Co. Li, Lan 李藍. 2003. Xiàndài Hànyǔ fāngyán chābǐjù de yǔxù lèixíng 現代漢語方言差比 句的語序類型 [Word order typology of comparative constructions in modern Chinese dialects). Fāngyán 方言 3:1–21. Lien, Chinfa. 1999. The sociolinguistic dimensions of comparative constructions in Taiwan Southern Min. In Contemporary studies on the Min dialects, ed. by Pang-hsin Ting, 204–224. Journal of Chinese Linguistics Monograph Series 14. Berkeley: Project on Linguistic Analysis, University of California. Luo, Futeng 羅福騰. 1992. Shāndōng fāngyán bǐjiàojù de lèixíng jí qí fēnbù 山東方言比 較句的類型及其分布 [The typological distribution of the comparative constructions in the Shandong dialects]. Zhōngguó Yǔwén 中國語文 3:201–205. Ōta, Tatsuo 太田辰夫. 1958. Chūgokugo rekishi bunpō 中国語歴史文法 [Historical grammar of Chinese]. Kyoto: Hōyū Bookstore (Translated by Shaoyu Jiang 蔣紹愚 and Changhua Xu 徐昌華. 2003. Beijing: Beijing University Publishers). Peyraube, Alain. 1989. History of the comparative construction in Chinese from the 5th cenury b.c. to the 14th century a.d. In Proceedings of the 2nd international conference on Sinology, 589–612. Taiwan: Academia Sinica. Peyraube, Alain. 1996. Recent issues in Chinese historical syntax. In New horizons in Chinese linguistics, ed. by C-T. James Huang and Yen Hui Audrey Li, 161–214. Dordrecht: Kluwer Academic Publishers. Peyraube, Alain, and Jang-Ling Lin. 2011. Some isues on the comparative construction of superiority in the history of Chinese and in Taiwan Southern Min. Paper presented at the Workshop on Reflections of Diachronic Change Mirrored in Early Southern Min, National Tsing Hua University, 26 November. Peyraube, Alain, and Thekla Wiebusch. 1995. Sur un cas d’ambiguïté en chinois médiéval concernant des formes comparatives. Faits de Langue 5:73–82. Piñol y Andreu, Francisco. 1928. Gramática China del dialecto de Amoy. Hong Kong: Imprenta de Nazaret. Stassen, Leon. 1985. Comparison and universal grammar. Oxford: Blackwell.

38  Hilary Chappell, Alain Peyraube and Song Na Stassen, Leon. 2005. Comparative constructions. In The world atlas of language structures, ed. by Martin Haspelmath, Matthew S. Dryer, David Gil, and Bernard Comrie. Map and Description 121. Oxford: Oxford University Press. Stassen, Leon. 2011. Comparative constructions. In The world atlas of language structures online, ed. by Matthew S. Dryer and Martin Haspelmath. Feature 121A. Munich: Max Planck Digital Library. http://wals.info/feature/121A. Accessed on 13 May 2011. van der Loon, Piet. 1966. The Manila incunabula and early Hokkien studies (Part 1). Asia Major 13:95–186. Wei, Pei-chuan 魏培泉. 2007. Guānyú chābǐjù fāzhǎn guòchéng de jǐ diǎn kànfǎ 關於 差比句發展過程的幾點看法 [Some remarks on the development of the comparative constructions]. Language and Linguistics 8:603–637. Wu, Fuxiang 吳福祥. 2010. Yuèyǔ chābǐshì ‘X + A + guo + Y’ lèixíngxué dìweì粤語差比 式 ‘X+A+過+Y’ 類型學地位 [Typology of the ‘X + A + guo + Y’ comparative constructions of inequality in Cantonese]. Zhōngguó Yǔwén 中國語文 3:238–255. Wu, Fuxiang 吳福祥. 2012. Mǐnyǔ chābǐshì de lìshí céngcì 閩語差比式的歷史層次 [The historical stratum of comparative constructions in the Min dialects]. Líyún Xuékān 勵 耘學刊 1:1–14. Yue-Hashimoto, Anne. 1999. The Min translation of the Doctrina Christiana. In Contemporary studies on the Min dialects, ed. by Pang-Hsin Ting, 42–76. Journal of Chinese Linguistics Monograph No. 13. Berkeley: Project on Linguistic Analysis, University of California. Zhang, Cheng 張赬. 2004a. Tángsòng shíqí ‘bǐ zì bǐjiàojù de yǎnbiàn 唐宋時期 ‘比’ 字比 較句的演變 [The comparatives with bĭ in the Tang-Song period]. Yǔyánxué Lùncóng 語 言學論叢 30:169–187. Zhang, Cheng 張赬. 2004b. Míngdài de chābǐjù 明代的差比句 [The comparatives of inequality in the Ming period]. Language and Linguistics 5:705–725. Zhang, Cheng 張赬. 2005. Cóng Hànyǔ bǐjiàojù kàn lìshǐ yǎnbiàn yǔ gòngshí dìlǐ fēnbù de guānxī 從漢語比較句看歷史演變與共時地理分布的關係 [On the relationship between the historical development and the geographical distribution of Chinese comparative constructions]. Yǔwén Yánjiū 語文研究:43–48.

Appendix I Li Jing Ji 荔鏡記 and Li Zhi Ji 荔枝記 operas (5 versions) 126 examples of comparative structures

TYPE I

 1  2  3  4

Type I: Compare Comparative A+比+B+ Verb

Total: 4

阮不比王魁負心。 我不比王魁負心。 我不比王魁負心。 不比王魁負心。

29.191 嘉靖 18.088 順治 31.032 道光 26.059 光緒

TYPE II

 5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Type II

Total:65

(a) Surpass Comparative A+Verb+過+B

Total: 24

好元宵強過別冥。 阮娘仔伊是千金閨女，都不強過阮奴婢。 娘是千金人閨女，都不強過阮隨婢。 啞娘是千金閨女，夭不都強過阮奴婢。 亞娘是千金閨女，夭不都強過阮奴婢？ 亞娘是千金閨女，夭不都強過阮奴婢。 阮說娘是千金之女，強過阮奴婢。 三人直過藍橋路，賽仙景，強過皇帝都。 一ム強過九領被。 一某強過九領被。 一某強過十領被。 看伊人物爽利，賽過廟裏天妃。 你只計賽過孫吳，許時有乜記號？ 恁今同姒(事)卜相痛，賽過姊妹弟共兄。 桃花面貌，密榴口齒，恰親像許嫦娥，賽過西施。 知阮啞娘昨冥即共三哥一冥恩愛，賽過百日恩情。 共君結托賽過相如配文君。 共君結托，賽過相如配文君。 共君結托賽過相如配文君。 飯今袂食，澀過吞沙。

5.117 嘉靖 28.132 嘉靖 27.089 萬曆 17.181 順治 30.070 道光 25.116 光緒 27.094 光緒 33.009 萬曆 3.073 順治 3.023 道光 2.054 光緒 9.011 嘉靖 23.054 嘉靖 54.024 嘉靖 15.024 萬曆 19.035 順治 33.094 順治 47.041 道光 42.032 光緒 25.052 嘉靖

(Continued )

40  Hilary Chappell, Alain Peyraube and Song Na (Continued)

25 26 27 28

Type II

Total:65

(a) Surpass Comparative A+Verb+過+B

Total: 24

飯無心食，澀過吞沙。 小妹只計勝過孫吳。 小妹你只計勝過孫吳。 小妹只計勝過孫吳。

21.028 萬曆 10.144 順治 21.050 道光 17.065 光緒

(b) Surpass-Similarity Comparative A+Verb+如+B 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

世事短如春夢。 見說廣南遠如天。 光陰相催緊如箭。 荔枝清香甜如蜜。 惹得我悶如江海。 我只處心悶如江海。 悶如長江水。 陳三色膽大如天。 情重如山恩深似海。 崖州一路遠如天邊。 感恩德重如天。 伊人富如石崇。 富如石崇，我不去爭。 一場姻愛勝如春夢。 你只功勞大如天。 陳三色膽大如天。 秋月窗前光如水。 今卜行，緊如箭。 心悶如遭，想著冤家目汁流。 三更過了，腳尾冷如霜。 夫妻人情冷如霜，那畏莫久長。 我兄威勢大如天。 越添得我悶深如江海。 情重如山恩深如海。 情重如山恩深如海。 心茫茫，步緊如箭。 伊今關門落樓去，越添我悶深如江海。 看只賊婢膽大如天。 情重如山，恩深如海。 情重如山，恩深如海。 一場官司大如天。 越添得我悉深如江海。 情重如山，恩深如海。 情重如山，恩深如海。 看只賊婢膽大如天。 一場官府大如天。 步緊如箭。

Total: 37 1.001 嘉靖 10.011 嘉靖 11.001 嘉靖 17.018 嘉靖 17.126 嘉靖 21.036 嘉靖 23.007 嘉靖 28.114 嘉靖 34.021 嘉靖 48.185 嘉靖 53.029 嘉靖 10.126 萬曆 10.127 萬曆 27.019 萬曆 27.038 萬曆 27.081 萬曆 28.111 萬曆 29.003 萬曆 31.005 萬曆 40.041 萬曆 40.137 萬曆 47.040 萬曆 7.117 順治 21.131 順治 21.131 順治 13.017 道光 16.041 道光 24.102 道光 34.069 道光 34.069 道光 37.048 道光 14.056 光緒 29.097 光緒 29.097 光緒 20.155 光緒 32.079 光緒 11.014 光緒

Comparatives of inequality in Southern Min  41 (bii) Surpass-Similarity Comparative A+Verb+似+B   66   67   68   69

人情薄似秋雲。 情重如山恩深似海。 一心忙似箭，兩腳走如飛。 收古記，勝似黃金。

Total: 4 1.002 嘉靖 34.021 嘉靖 47.054 嘉靖 15.041 萬曆

TYPE III Type III: Zero-marked Comparative A+Verb+B   70   71   72   73   74   75   76   77   78   79   80   81   82   83   84   85   86   87   88   89   90   91   92   93   94   95   96   97   98   99 100 101 102 103 104 105

算來讀書強別事。 一ム強十倍，十倍甲也寒。 阮處門風更強恁所在。 新年願卜強舊年。 母豬肉強食笋。 人說潮州稅户，亦袂强阮做官人個。 阮厝門風更強恁厝個。 我乞人飼夭強你。 阿娘，學們夭要強伊，肯不親像伊？ 新年強舊年。 我厝門風，更強恁只所在。 強你。 我厝門風更強恁只所在。 新年強舊年。 新年強舊年。 我厝門風更強恁只所在。 一陣阿妹賽觀音。 一位娘子賽觀音。 明日遊馬賽蓬萊。 三人直過藍橋路，賽仙景。 一位娘仔賽觀音。 高樓上賽觀音人物，都在珠簾底。 高樓上賽西施，人物在珠簾底。 高樓上賽西施。 那因亞娘昨冥共三哥恩愛，賽其再作夫妻。 那因亞娘共三哥恩愛，賽其再做夫妻。 嫁女必勝吾家，娶婦必弱吾家。 願得年年頭大勢，新年便得勝舊年。 新年便得勝舊年。 嫁女必勝吾家，娶婦必弱吾家。 人物出眾，又超群。 人物出眾又超群，面貌又聰俊。 人物出眾又超群。 人物出眾，又超群。 物出眾又超群，面貌又聰俊。 人物出眾又超群。

Total: 36 4.023 嘉靖 5.047 嘉靖 21.023 嘉靖 32.158 嘉靖 07.098 萬曆 18.026 萬曆 18.027 萬曆 19.257 萬曆 24.132 萬曆 2.077 順治 10.037 順治 11.75 順治 21.011 道光 4.024 道光 3.039 光緒 17.008 光緒 6.245 嘉靖 8.040 嘉靖 54.032 嘉靖 33.009 萬曆 4.632 順治 7.056 順治 16.019 道光 14.024 光緒 27.009 光緒 32.010 道光 14.307 嘉靖 33.027 道光 28.040 光緒 14.308 嘉靖 10.112 萬曆 10.022 道光 8.036 光緒 10.112 萬曆 10.022 道光 8.036 光緒

42  Hilary Chappell, Alain Peyraube and Song Na Type IV: Adverbial Comparative A+可+Verb+B 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126

也句可強你睛盲頭(青冥頭)。 啞娘那卜學伊，都不可強伊。 伊後生人好氣好力，句可強我。 會可強你。 許你那帶只處，可強當人命。(只：直) 那卜學伊，都不可強伊？ 有人愛可強無人愛。 得人惜，可強乞人怨。 伊有氣力，向可強我。 那卜學伊，句可強伊。 有人愛可強無人愛。 得人惜可強得人怨。 伊有力氣，向可強我。 那卜學伊，可強伊。 有人愛，可強無人愛。 得人惜可強得人怨。 我打一雙金釵句可重伊人个，乞你。 我冥旦打一雙可重伊個，乞你帶。 我明旦打一對可重伊的乞你帶。 我冥旦打一對可重伊個乞汝帶。 官人許前頭一陣娘仔，生得句可親淺伊。

Total: 21 22.366 嘉靖 26.242 嘉靖 8.106 順治 11.752 順治 14.286 順治 15.239 順治 17.114 順治 17.123 順治 18.020 道光 27.086 道光 30.043 道光 30.047 道光 15.033 光緒 23.128 光緒 25.063 光緒 25.071 光緒 14.111 嘉靖 5.382 順治 10.052 道光 8.088 光緒 8.069 嘉靖

Appendix II Seven further operas: Măn Tiān Chūn 滿天 春, Jīn huā nǚ 金花女, Yù Yán Lì Jǐ 鈺妍麗 錦, Băi Huā Sài Jǐn 百花賽錦, Yáng Guăn Gē 楊管歌, Sū Liùniáng 蘇六娘, Tóng chuāng qín shū jì 同窗琴書記 80 examples TYPE I Type I: Compare Comparative A+比+B+ Verb

Total: 0

TYPE II

127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145

Type II

Total: 70

(a) Surpass Comparative A+Verb+過+B

Total: 29

結佐兄妹十六年，勝过是好姻緣。 繼祖承宗勝過求名。 阿娘只識藥，夭勝過醫官。 見娘身體，伏事勝過生時。 伊人府地官家門楣，兼又人物生得怜俐賽過 潘安標致。 莫說是凡間女子，就是月裡姮娥賽不过他。 岂不知恁丈夫人口甜賽过糖密，都那是眼前花。 又見許紅芍藥花開，粉蝶児穿採成陣，賽过十Ｌ 錦。 原來是山芲岸柳苑囿芲养賽過牡丹芍藥。 玉樓上，賽過觀音寺。 共君結托賽过相如配著文君。 刈淂我障諂賽过清凨劍。 遍處是,賽過蓬萊無二。 遍處是賽過蓬萊無二。 同房舍，又同鄉里，賽過賽過親生兄弟。 師徒情分，賽過親生。 論相惜，賽過親生。 師徒情分，賽過親生。 三年同衾枕，賽過夫妻情。

39 滿天春 14.013 金花女 02.136 蘇六娘 07.015 蘇六娘 34 滿天春 34 滿天春 8 滿天春 8–9 滿天春 29 滿天春 176.017 金花女 31 鈺妍麗錦 8 百花賽錦 4.047 同窗琴書記 4.054 同窗琴書記 8.079 同窗琴書記 9.076 同窗琴書記 9.079 同窗琴書記 9.083 同窗琴書記 13.022 同窗琴書記

(Continued )

44  Hilary Chappell, Alain Peyraube and Song Na (Continued)

146 147 148 149 150 151 152 153 154 155

156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184

Type II

Total: 70

(a) Surpass Comparative A+Verb+過+B

Total: 29

兄弟相見，賽過得金寶。 杭州同學時，恩情賽過親生。 咱今对景消愁，都不強過鬱除相思病損。 共我爹討一官媒，共咱明婚正娶，一來全其節， 又都不強过只招商店內苟合。 招伊結親義，都不強過招商店內只處無人見。 天時地利人和最美，我今為恁說合，強過尔明昏 正娶。 正是腹飢添一斗，強过十年糧粮。 叔做太守兄運使，阮厝門凨強过恁厝个我那不 呾，[女+卜]子你內頭揔不知。 求得文學胸中志，強過掌上珠共金。 為著功名到只路來，山嶺高過上天台。

14.153 同窗琴書記 17.04 同窗琴書記 10 滿天春 11 滿天春

(b) Surpass-Similarity Comparative A+Verb+如+B

Total: 41

共恁㤙爱重如山。 我只心內乱如系。 我身上痛如刀刺。 爹爹見尔衣衫襤縷，発怒如雷。 心頭焦悴如刀刺。 自古道，酒令嚴如軍令。 人說酒腸寬如海，果然色胆大如天。 人說酒腸寬如海，果然色胆大如天。 那亐我独自心頭乱如系。 恨許畨軍催緊如箭。 覩物趣准如面。 爹媽早世，幼如孩。 行過山崗驛遞，雄山峻嶺，勝如登天。 至親情分薄如紙。 忍心絕倖冷如水。 風飛雪白， 細如絲。 我一身冷如冰。 ○○○向般人可　○○○紛紛如醉。 為尔相思一病重如山。 管乜弓鞋短小路遠如天。 越自添得我悶深如海。 只悶人心肝疼如刀刺。 阮身上疼如刀刺。 步□□□□遠如天。 趕浮云飛走緊如箭。 那恨我只弓鞋短細，路遠如天。 越自添得我悶深如海。 床寬被闊，冷如霜。 二頭消息冷如霜。

7 滿天春 15 滿天春 12 滿天春(上) 34 滿天春(上) 39 滿天春(上) 13 滿天春(下) 22 滿天春(下) 22 滿天春(下 24–25 滿天春(下) 31 滿天春(下) 39–40 滿天春(下) 64.005 金花女 577.008 金花女 221.01 金花女 221.01 金花女 510.008 金花女 640.01 金花女 1 鈺妍麗錦 14 鈺妍麗錦 15 鈺妍麗錦 22 鈺妍麗錦 26 鈺妍麗錦 15 百花賽錦 (上) 4 百花賽錦(中) 4 百花賽錦 (中) 5 百花賽錦(中) 17 百花賽錦 01.018 蘇六娘 02.079 蘇六娘

12 滿天春 12 滿天春 30 滿天春 36 滿天春 4.215 同窗琴書記 431.017 金花女

Comparatives of inequality in Southern Min  45

185 186 187 188 189 190 191 192 193 194 195 196

(b) Surpass-Similarity Comparative A+Verb+如+B

Total: 41

頭毛挩落細如絲。 肉今割落幼如麵。 五年恩愛，共伊人，重如山。 事急如火，事急如火。 事急如火，事急如火。 桃花賊婢，罪過大如天。 思深如海。 思量我恨深如海。 階下百拜辭先生，周旋恩德大如天。 只去路塗，心急如箭。 心急如箭。 日月緊如梭。

02.150 蘇六娘 02.151 蘇六娘 06.106 蘇六娘 08.015 蘇六娘 08.015 蘇六娘 09.009 蘇六娘 7.069 同窗琴書記 9.022 同窗琴書記 12.061 同窗琴書記 12.062 同窗琴書記 13.009 同窗琴書記 14.211 同窗琴書記

(bii) Surpass-Similarity Comparative A+Verb+似+B

Total: 0

TYPE III

197 198 199 200 201 202 203 204 205 206

Type III: Zero-marked Comparative A+Verb+B

Total:10

听伊言談，志氣超群。 紅顏胭脂，色賽牡丹，真貴氣。 我只罗巾，尔收記在心，只罗巾情重千金。 咱今冥勝入洞房。 德行如金石，志節賽氷霜。 『三分親強別人』。 《百花賽錦》 秀才阿,救阮命強再生。 救阮一命重再生。 郊外好所在，真個賽蓬萊。

1 滿天春(下) 3 滿天春(下) 4 滿天春 14 滿天春(下) 40 滿天春 347.008 金花女 百花賽錦 楊管歌 楊管歌 1.127 同窗琴書記

TYPE IV Type IV: Adverbial Comparative A+可+Verb+B

Total: 0

Appendix III Doctrina Christiana 6 examples

TYPE II Type II (a) Surpass Comparative A+Verb+過+B 207. 208. 209. 210. 211. 212.

僚氏賜福乞你。勝過衆婦人。 惜僚氏勝過各衆物。 一件惜僚氏勝過各衆物。 僚氏里賜福與汝。勝過衆婦人。 受大敖羅里耶。勝過衆天人。 賜汝大福。勝過衆天人。

Total: 6

嘉靖 (1566) 萬曆 (1581) 順治 (1651) 道光 (1831) 光緒 (1884) Total

Version

Schema

1

0

0

0

7

2

7

5

5

23

5

2

2

1

0

12

4

4

2

0

2

6

4

0

‘according to’

‘compare to’ (metaphor)

‘compare with’

‘give an example’

Preposition

Verb

4

0

0

0

4

0

‘with’

4

1

1

1

0

1

Comparative marker

Table 2.6  Four functions of pi2 比 in in the folk operas Li Jing Ji 荔鏡記 and Li Zhi Ji 荔枝記

174 examples of pi2 比 in all functions

Appendix IV

28

6

6

5

3

8

Noun

9

5

1

1

1

1

Nominal expressions

Nominal expressions

1

1

0

0

0

0

Difficult to interpret

93

24

19

17

11

22

Total

Total

滿天春1604 Măn Tiān Chūn 金花女 (明萬曆) 1572–1620 Jīn huā nǚ 鈺妍麗錦 1572–1620 Yù Yán Lì Jǐ 百花賽錦 1572–1620 Băi Huā Sài Jǐn 楊管歌 1572–1620 Yáng Guăn Gē 蘇六娘(明萬曆) 1572–1620 Sū Liùniáng 同窗琴書記 (清乾隆)1782 Tóng chuāng qín shū jì Doctrina Christiana

Version

Schema

0

0

0

1

0

3

0

4

8

0

10

0

1

37

1

6

5

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

8

15

5

‘according to’

‘compare to’ (metaphor)

‘compare with’

‘give an example’

Preposition

Verb

0

0

0

0

0

0

0

0

0

‘with’

0

0

0

0

0

0

0

0

0

Comparative marker

Table 2.7  Four functions of pi2 比 in seven local operas from the Ming and the Qing periods

16

0

7

1

0

1

1

2

4

Noun

10

0

4

0

0

1

1

2

2

Nominal expressions

Nominal expressions

2

1

0

0

0

0

0

0

2

Difficult to interpret

81

1

20

5

0

5

3

11

36

Total

3 The emergence of the obligative modal tioh8 in Southern Min A change induced by semanticpragmatic factors* Ting-ting Christina Hsu 1  INTRODUCTION Deontic modals are generally regarded as more objective than epistemic modals since they express ‘the necessity and possibility of acts performed by morally responsible agents’ and they are not concerned with ‘matters of knowledge, belief, or opinion’ of speakers, as epistemic modals are (Lyons 1977:681, 823). Nevertheless, deontic modals can also indicate speakers’ attitudes or beliefs towards the expressed proposition, for example, the deontic modal may in the sentence You may smoke here (Palmer 1986:103). In light of the fact that both deontic and epistemic modals involve subjectivity, Traugott and Dasher (2002:113) propose that all modals, be them deontic or epistemic, are more or less subjective, depending on how much they represent the attitudes or opinions of the speaker. Traugott (1989) even defines deontic modals as a kind of ‘mood with will’ as opposed to epistemic modals, following Jespersen’s seminal distinction of modals based on will (1924:320–321). Given the subjective nature of deontic modals, subjectification could be a promising analysis (Langacker 1990) to motivate the change from premodals to deontic modals, despite the fact that subjectivity-driven change is relatively untouched in the literature. In this paper, we will see how subjectification has helped develop the obligative deontic modal tioh8 in Southern Min (henceforth SM). I identify two major historical stages that were involved in the developmental process. During the first stage, tioh8 changed from a verb to a preposition and in the second stage to a causative verb, ultimately becoming a deontic modal. The preposition tioh8 originated from a verb meaning ‘attach to’ or ‘hit’ and it usually selected a locative direct object. The whole PP headed by tioh8 was located to the right of the verb, but later moved to the preverbal position. So, tioh8 was reanalyzed as a causative verb, which dominated the following lexical verb. As a result, the original complement of the preposition tioh8 was then transformed into a pivotal argument that associated the preceding causative verb and the following lexical verb. With the Causer removed from the sentence due to subjectification, the causative verb tioh8 finally developed into a deontic modal. When tioh8 was used as an instrumental preposition, it began to select animate NPs as complements and it obtained a force meaning such that the objects

50  Ting-ting Christina Hsu it selected were used for a certain purpose. Throughout the entire developmental process from preposition to deontic modal, tioh8 preserved the meaning of force in different ways. Therefore, when tioh8 was eventually used as a deontic modal, which is its most common use today, it indicates that the speaker approves of the authority under which the subject feels obligated to act in a certain way. It is worth mentioning that the meaning of force is a distinguishing property of deontic modals. Modals, be they deontic or epistemic, all involve possibility or necessity, that is, ‘the central notions of traditional modal logic’ (Lyons 1977:787) but only deontic modals allow speakers to interact with listeners by implementing force on the listener or a relevant third party. Therefore, to develop obligative modals, a reliable source of force (Talmy 1988; Heine 1993) or authority (Sweetser 1990) is important. Traugott (1989:38) claims that premodals should already possess some semantic properties that are able to later generate the force-related sense that deontic modals require. As we will see in the SM examples, the force meaning of premodal tioh8 is so strong and persistent that, even when the Causee of the causative verb tioh8 was moved out of the verb’s immediate scope to a subject position, the effect of force being put on it is still detectable. This paper is structured as follows. Section 2 reviews the theoretical background of the study. Section 3 introduces the multiple functions of tioh8 and its diachronic distributions. Section 4 outlines the mechanisms of this development. Section 5 concludes the paper. 2  LITERATURE REVIEW

2.1  Grammaticalization and reanalysis The study of grammaticalization, which became popular after Meillet’s famous article on the evolution of grammatical forms (Meillet 1912), is concerned with the change from input lexemes (or earlier lexical forms) to more grammatical formatives. Grammaticalization can introduce new forms or categories into languages (Meillet 1912:133). One example is the Ewe complementizer bé, which developed from a more contentful verb meaning ‘say’, as shown in (1a) and (1b) (Lord 1989:307–308): (1) a  Me-bé me-wɔ 1sg-say 1sg-do ‘I said I did it’.

e 3sg

b Me-dí bé máfle 1sg-want (say) buy.1sg.sbjv ‘I want to buy some dresses’.

awua dress

ɖe-wó some-pl

A broad sense of grammaticalization includes all processes that give rise to forms that are more grammatical in function than their earlier uses (Hopper 1991). Under this view, any type of change that leads to more grammaticalized functions can be seen as an example of grammaticalization, whether there are lexical inputs or not.

The obligative modal tioh8  51 For example, the nominal suffix -ro ‘in’ in the Mande language Vai developed into an aspectual marker signaling durative or iterative meaning (Koelle 1968:90): (2)

Kē ̣rẹ bẹ kí-ro Deer cop sleep-in ‘The deer was sleeping’.

Grammaticalization is a gradual process which does not necessarily include a certain type of input or output. Rather, it is defined by the difference in grammaticality before and after a change. Moreover, grammaticalization involves unidirectional changes. For example, meanings tend to unidirectionally become more and more generalized and specific parts of the original meaning might be lost or bleached (Hopper and Traugott 2003:68). One instance of this type of meaning shift can be seen in the development of the English word car, which came from the Latin carrum meaning ‘two-wheeled Celtic war chariot’ and was generalized to refer to all automobiles in the late 19th century. Since the original meaning of the word has been lost, including specifications that the car is ‘two-wheeled’, the word car now needs modification to refer to a specific kind of two-wheeled car. Similarly, unidirectionality predicts that grammaticalization makes new forms more grammatical than their earlier uses, but not vice versa. This spectrum of grammaticality constrains the types of changes that we can account for. For example, changes that do not necessarily lead to the loss of expressivity, or changes that do not give rise to more grammatical forms, would fail to be accounted for by a theory of grammaticalization. One way to explain how forms become less grammatical is through the process of reanalysis, which refers to ‘changes in the structure of an expression or class of expressions that do not involve any immediate or intrinsic modification of its surface manifestation’ (Langacker 1977:58). For example, in Proto CahuillaCupeno, an Uto-Aztecan language, the reciprocal prefix *nǝ- was reanalyzed as part of the postpositional morpheme *-w, as shown in (3) (Langacker 1977:69): (3)

Ø | nǝ

-

WITH | w

>

WITH | nǝw

Resegmented forms are reanalyzed as having new functions. Though reanalysis is considered the primary mechanism for grammaticalization (Hopper and Traugott 2003:32), it does not necessarily lead to grammaticalization, that is, the loss of lexical autonomy and the increase of dependency (Lehmann 2005:155).1 Whether a process of reanalysis involves a change from a grammatical structure to a lexical structure, which contradicts the unidirectionality of grammaticalization, or not is irrelevant. For example, the preposition up in He wants to up the ante and Life is full of ups and downs is reanalyzed as a more contentful verb or noun (Hopper and Traugott 2003:49). In this sense, the operation of reanalysis can also lead to lexicalization, i.e. the adoption of a unit into the lexicon (Bussmann 1996; Lehmann

52  Ting-ting Christina Hsu 2002; Brinton and Traugott 2005). The change of SM tioh8 from a preposition to a causative verb resembles the change of prepositional up to a verb, indicating that reanalysis plays a role in developing the deontic modal tioh8 in SM.

2.2  Subjectification Another important theory that the present study relies on is subjectification. Subjectification is a pragmatic-semantic process whereby ‘meanings become increasingly based in the speaker’s belief/attitude toward the proposition’ (Traugott 1989:35). Langacker (1990:17) defines subjectification as the ‘realignment of some relationship from the objective axis to the subjective axis’. In other words, an element is subjectivized when it shifts from describing an external situation to reflecting the internal situation of the speaker’s evaluative or cognitive perspective (Traugott 1982). In a relatively objective scenario, all things involved in a proposition can be referred to with on-stage landmarks, while in a more subjective scenario the speaker participates in expressing the proposition, creating landmarks that refer to or are associated with the speaker. Therefore, (4b) is more subjective than (4a) since the speaker contributes to the interpretation of the sentence. In an even more subjective situation, the speaker completely disappears from the sentence, as in (4c). However, the speaker’s perspective is still crucial for interpreting the sentence, since the audience would need to take her standpoint as the reference point to interpret the sentence in (4c). (4) a John is sitting across the table from Mary. b John is sitting across the table from me. c John is sitting across the table. In (4c), the landmark of the preposition across is the speaker herself, even though this is not overtly expressed. This is a case of maximal subjectification, by which the speaker becomes part of the common ground of the conversation, creating a zero form in the sentence (Langacker 1995; Traugott 1995). An illustration of the maximally subjectivized offstage scenario in (4c) is shown in Figure 3.1: The notion of subjectification is specifically crucial in this study in modeling the final change from the causative tioh8 to the deontic modal tioh8, since it explains why the interaction of force between arguments is preserved. Although

Speaker

table

John

Stage [sentence] Figure 3.1  Conceptual structure for maximally subjectivized offstage situation in (4c)

The obligative modal tioh8  53 the highest argument, i.e. the Causer, of the causative tioh8 does not surface with the deontic modal, the original semantic interactions between participants in the causative construction determines how force relates the participants of the modal construction. 3  THE MULTI-FUNCTIONAL TIOH8 IN SOUTHERN MIN Southern Min (SM) is a Chinese dialect used in Fujian and Taiwan. It has a large population of speakers in the southeastern coastal region of China, e.g. Guangdong and Hainan, and Southeast Asian countries such as Singapore, Malaysia, the Philippines (Li and Yao 2008; Hsin 1999). SM is valuable for diachronic studies of Chinese because it preserves many features of Middle Chinese, specifically features of the language spoken in the period of the Tang Dynasty, which spanned from the 6th.c to the 9th.c and was very influential in international business during this period. In this paper we tap both modern and historical texts of SM tioh8 from several sources. First of all, four versions of the playscript Li Jing Ji/Li Zhi Ji,2 datable to the sixteenth century at the earliest, were adopted as the primary sources of diachronic SM. However, the root of SM tioh8 runs deeper and can be traced back even earlier to the old Chinese cognate zhuó 着/著. So the present study also garners Chinese data to track down the routes of early changes. To ascertain the modern developments of SM, we rely on colloquial SM data including SM folk stories and soap operas.

3.1 Polysemous tioh8 in modern Southern Min In modern SM, tioh8 is a frequently used, multifunctional word, including verbal uses and more grammatical functions such as verbal complement, clause connector, adverbial marker, and pragmatic marker. Between the more contentful uses and the more functional uses of tioh8 is the deontic modal tioh8, the most common function of the form in modern SM. The examples in (5) demonstrate these functions and special meanings of those functions are included after each example (Lien 2001). The examples in (6) demonstrate how tioh8 behaves as a deontic modal. Table 3.1 compares all of these functions (Lien 2001). Table 3.1  The functions of tioh8 in modern SM (from Lien 2001) Grammatical constructions Functions A. Verb B. Verbal complement C. Deontic modal D. Adverbial marker E. Clause connector F. Discourse marker

[Reaching or acquiring an entity, turn-taking, perpetrating a bad outcome, metalinguistic operation showing agreement] [phasal function showing achievement] [The imposition of the speaker’s will on the addressee] [Distressed imposition] (rarely used) [Linking temporal, causal, and conditional sequences] [Signal of speaker’s sudden realization or the initiation of an explanation]

54  Ting-ting Christina Hsu (5) a 箭 著 紅心。  [A: reaching or acquiring an entity] tsinn3 tioh8 ang5sim1 arrow hit.the.target red.target ‘The arrow hit the target right in the center’. b 與

我

了

錢

攏 是 汝 著 的。 [A: perpetrating a bad outcome] tsinn5 long2 si7 li2 tioh8 e0 hoo7 gua2 liau2 dat 1sg waste money all cop 2sg cause prt ‘It is all because of you that I lost money’.

c 汝 講 的 著。 [A: metalinguistic operation] li2 kong2 e5 tioh8 2sg speak nmlz correct ‘What you said is correct/right’. d 掠 賊仔 會 著。 [B: achievement] liah8 tshat8a2 e7 tioh8 catch thief-dim can ach ‘(They) can (definitely) catch the thief’. [D: distressed adverb]

e 伊 定著 會 來。 i1 tian7tioh8 e7 lai5 3sg certainly assu come ‘He certainly will come’. f 天

抑

未

光，

thinn1 iah4 be7 kng1 sky yet neg light ‘He woke up before dawn’. g 此 tsit4

個 e5

dem

cl

事志 tai7tsi7 matter

我 gua2 1sg

伊 i1 3sg

是 si7 cop

著 精神 啊。 [E: temporal connector] tioh8 tsing1sin5 a0 conj wake.up prt

帶 著 恁 兄弟仔， tshua3 tioh8 lin2 hiann1ti7a2 consider conj 2pl brother-dim

容允 汝。 [E: causal connector] iong7un2 li2 conj ext pardon 2sg ‘It is only out of respect for your brother that I won’t blame you for this’. 即 有 tsiah4 u7

The obligative modal tioh8 55 h 著

食

此

款

藥仔， 即 有 效。 [E: conditional connector] tsiah4 u7 hau7 tioh8 tsiah8 tsit4 khuan2 ioh8a2 conj eat dem cl medicine-dim conj ext work ‘You can only be cured if you take this kind of medicine’.

i A: 你 今旦日 下晡 哪會 無 來 上 課? kin1na2jit8 e7poo1 na2e7 bo5 lai5 siong7 kho3 li2 2sg today afternoon how.come neg come attend class ‘Why didn’t you attend the class this afternoon’?  B: 阿著 儂 唔 好 貯咧 [F: initiation of an explanation] a1tioh8 lang5 m7 ho2 ti7leh4 dm body neg good ext-prt ‘It was because I was unwell then’ or ‘Well, I was unwell then’. (6) a 汝 著 較 拍拼 咧。 [C: modal indicating imposition] li2 tioh8 khah4 phah4piann3 leh4 2sg should more work.hard prt ‘You should work harder’. b 汝 著 唔 通 講。 [C: modal indicating imposition] li2 tioh8 m7 thang1 kong2 2sg should neg deon say ‘You shouldn’t tell (this) (to anyone)’. c 著 緊 食。 tioh8 kin2 tsiah8 should hurry.up eat ‘You’d better hurry up finishing your meal’. The examples in (6) tell us two interesting things about the behavior of the modal tioh8. First, unlike other deontic modals, tioh8 cannot be negated simply by a negative marker such as m7 or bo5. When speakers need to negate the modal tioh8, they may use another deontic modal to help, such as thang1 shown in (6b), or may be totally substituted by a negative modal bian2, which means ‘exempt from the duty or obligation’, as shown in (7).

56  Ting-ting Christina Hsu (7) a 汝 免 拍拼 啦。 li2 bian2 phah4piann3 lah4 2sg neg.deon work.hard prt ‘You don’t have to work (that) hard’. b 汝 免 食 遐 緊。 li2 bian2 tsiah8 hiah kin2 2sg neg.deon eat deg fast ‘You don’t have to eat that fast’.

(negative counterpart of (6a))

(negative counterpart of (6c))

Unlike other modals the deontic tioh8 cannot be negated, as in Table 3.2 where bian2 is a suppletive form of tioh8. Except for tioh8 and ai3, all deontic modals in SM resorts to suppletion as a means to express negative modality.3 The idiosyncratic behavior of tioh8 in negative formation suggests that the modal tioh8 might not be as old as other modals since it requires an irregular negation strategy. It is indeed the case that tioh8 can express negation by adjoining with negative markers bo5 or m7, but in those cases tioh8 would be interpreted as a predicative use of ‘right’ or ‘hit the target’, instead of a deontic modal. A possible explanation is that when bo5 or m7 was historically used to negate modals, the deontic modal tioh8 did not yet exist, suggesting that the modal use of tioh8 is fairly young. Another interesting thing that we see in example (6) is that the presence of tioh8 can reduce the imposition being put on the listener in imperative sentences. As shown in (6c), the sentence-initial tioh8 can effectively lessen the degree of imposition amounting to some extent of a strong suggestion. Since imperative expressions are a kind of direct interaction between the speaker and the listener, (6c) shows that the deontic modal can effectively reduce the authoritative power in the situation. In other words, deontic modals can be regarded as weaker imperative markers in terms of how strong the force is between the speakers. The close relationship between the imperative mood and the deontic modal is illustrated below in Figure 3.2:

Table 3.2  Negative strategies of SM deontic modals Modals

Character

Meaning

Character

Neg. expression

e7sai2(tit4) e7tit4thang1 e7ing7e/tit4/tsit4 thang1 ing1kai1 ing1tong1 tioh8 ai3

會使(得) 會得通 會用e/得/tsit 通 應該 應當 著 愛

Permissive Permissive Permissive Permissive Obligative Obligative Obligative Obligative

袂使(得) 袂得通 袂會用e/得/tsit 毋通 無應該 無應當 免 免

be7sai2(tit4) be7tit4thang1 be7ing7e/tit4/chit4 m7thang1 bo5ing1kai1 bo5ing1tong1 bian2 bian2

The obligative modal tioh8  57

[Weak force] [Strong force] Deontic modalImperative mood Figure 3.2  Continuum of force of imposition

Figure  3.2 suggests that these two kinds of expressions can shift along the dimension of force. If the speaker’s force increases in a sentence with deontic modals, it might be interpreted as making a request or order; on the contrary, if the speaker’s force reduces in an imperative sentence, it might be perceived that he is making a suggestion. This tells us that in an expression the amount of force is adjustable and words might express various degrees of force in a given context. The functions listed in Table 3.1 do not reflect the diachronic use of tioh8. Many previous uses of tioh8, such as a preposition and a causative verb, became obsolete. On the other hand, some modern uses of tioh8 in SM did not exist until the 20th c. We will explore how these uses evolved in the next section.

3.2  Diachronic development of tioh8 A good diachronic study should use sources from as early in history as possible, but in the case of SM the earliest faithful record we have only dates back 500 years. This short duration of time is not sufficient to build patterns of historical change. Fortunately, tioh8 in SM has a cognate zhuó 著 in Chinese (Mei 2000; Yang 1992). The earliest uses of Chinese zhuó express meanings that are very similar to what we know about tioh8 in SM, such as ‘to attach to’ and ‘to hit the target’. Though the Chinese zhuó is pronounced differently from tioh8 in SM, their identical morphological form, written either as 著 or 着, and their similar meanings and distributions all suggest that they are highly likely to be a single cognate with alternate pronunciations (Yang 1992:366),4 probably due to dialectal or regional differences. The history of Chinese zhuó can therefore help us recover the history of tioh8 in the past when no written record of SM was attested. The diachronic development of tioh8 can be separated into two stages according to the mechanisms of change. The first stage spans from the period of Old Chinese (OC, 1250 bce to 221 bce) to the period of Late Middle Chinese (LMC, 10th-12th. cc), during which tioh8 changed from a lexical verb to a preposition. The second stage spans from the period of Early Middle Chinese (EMC, 5th-7th.cc) to the period of Modern Chinese (ModC, 13th – 20th c), during which tioh8 obtained the meaning of force required for the emerging deontic modal, and eventually in the period of Late ModC it acquired the modal function. Stage one is described in section 3.2.1 and stage two, in section 3.2.2. Note that section  3.2.1 makes use of the Chinese cognate zhuó data, which predates the data available for SM tioh8.

58  Ting-ting Christina Hsu 3.2.1  From verb to preposition The earliest use of zhuó in OC was a verb meaning ‘to attach’ or ‘to contact with others’, as shown in (8a) and (8b) (Wang 1958 [2003]) (8) a 今 jīn now 底

戾 lì stable

久 jiǔ long

矣， yǐ prt

戾 lì stable

著

滯

淫，

誰

久 jiǔ long

將 jiāng fut

底， dǐ bottom

能 興 之? 《Jìnshū Guóyǔ 4; 》 dǐ zhuó zhì yín shéi néng xīng zhī bottom attach stagnation long who can revive 3sg ‘Now the country has been stable for a long time, and when it has been table for a long time there will be severe stagnation. Who can invigorate it’? b

風 行 而 著 於 土。 《Zuǒzhuàn Zhuānggōng Year 22; 4th c bce》 fēng xíng ér zhuó yú tǔ wind go conj attach prep soil ‘The wind blows and falls to the ground’.

In (8a), the object of zhuó implies that there is an unchanging state in the country by metaphorically alluding to something being stuck at the bottom of a container. So, we can literally interpret the verb to mean something like ‘attach to’. In (8b), zhuó is followed by a locative expression signaling the final position of the wind. In both cases zhuó selects its own arguments and is used independently as a verb. Later, during the Han Dynasty (206 bce-220 ce), zhuó began to participate in complex predication with a preceding verb. During this time, the object of zhuó was restricted to locations or destinations, and the verbs that zhuó combined with were restricted to verbs causing a change of location. This indicates that the initial meaning of zhuó still played a role in building semantic relationships because attachment relations imply that there is a relevant location for something to attach to. Therefore, in a sentence that contains only a pronoun as the object of zhuó, the pronoun must be referring to a location, as shown in (9). (9) 今 鐘 皷 無 所 懸 著。 《Lùnhéng Léixū; 80 ce》 jīn zhōng gǔ wú suǒ xuán zhuó now bell drum neg place hang attach ‘Now the musical instruments have no place to hang from’. Later, zhuó began to lose its salience in the verbal complex (Wang 1958) by making less semantic contributions and being less restrictive of which verbs it can

The obligative modal tioh8  59 follow and which object it can select for. In example (10), zhuó doesn’t follow a verb that brings about a change of location, but rather a psych verb tān 貪 meaning ‘to covet’. This change indicates that zhuó was losing its content and initiating the early stage of grammaticalization (Hopper and Traugott 2003:96). (10)

不 bù neg

留 心 於 liú xīn yú keep mind prep

明， 貪 著 世間。 《Dàbǎojījīng 93; 8th c》 míng tān zhuó shìjiān light covet attach world

‘(They do) not concentrate on pursuing light but instead indulge (themselves) in the world’.

This kind of semantic attenuation and generalization of zhuó occurred at the end of the Han Dynasty period (around 100–200 ce) (Cao 1995:27). Despite being weakened, zhuó still put locational restrictions on the following noun phrase, which suggests that zhuó had not lost its verbal status completely during this time. The verbal complex composed of zhuó and the preceding verb became unstable and vulnerable to change in the period of Wei, Jin, Southern, and Northern Dynasties (the WJSN, 260–590 ce), when the interpretation of the sentence began to rely solely on the verb preceding zhuó for lexical content. As a result, the remaining function of zhuó was to simply connect the verb preceding it and the NP object following it. This marks a stage at which zhuó was more like a grammatical functional word rather than a lexical word. The ongoing process of change created some bridging contexts between the verbal complex and the verb-complement construction, as shown in (11) (Jiang 2005:159): (11) 其

dem

rén yǔ person dat

pín poor

人 言: 今 可 脫 《Bǎiyùjīng Pínrénshāocūhéyīyù; 5th c》 rén yán jīn kě tuō person say now able.to take.off

汝 rǔ 2sg

麤 cū rough

衣 yī clothes

著 zhuó put

prep

火 huǒ fire

即 脫 著 jí tuō zhuó immediately take.off prep

火 huǒ fire

中。 zhōng inside

qí

人

人 便 rén biàn person conj

與

褐 hè hemp

貧

於 yú

 . . . 貧 中。 zhōng pín inside poor

‘The person . . . told me, “Now you can take off your rough hemp clothes and throw them into the fire”, so I took them off and put them into the fire immediately’.

The first zhuó in (11) functions as a verb followed by a preposition yú which heads a locative prepositional phrase, whereas the second zhuó is ambiguous due to the absence of the preposition yú. So, the second zhuó can be interpreted as either a verb or a preposition, depending on how salient the meaning of the action ‘put’ is. If the verb tuō ‘take off’ in sentence two is the only action that the speaker

60  Ting-ting Christina Hsu or listener cares about, then the function of zhuó would be reduced to simply connecting the verb tuō to the following locative object huǒ ‘fire’, similar to the function of the preposition yú in the first sentence. In the following examples, zhuó behaves even more like a true preposition. Firstly, zhuó does not contribute to the semantic content of the sentence; the only function of zhuó is to syntactically introduce the following locational NP. Moreover, the verb preceding zhuó takes its own argument, i.e. wǎn ‘bowl’ in (12a) and ‘Changwen’ in (12b), so both have no obligatory sematic relationship with zhuó. (12)  a  雷公 若 二 升 椀， 放 著 Léigōng ruò èr shēng wǎn fàng zhuó duke.of.thunder dem two cl bowl put prep 庭 中。 《Sānguózhìzhù   Wèizhì Cáoshuǎngzhuàn; 5th 》 tíng zhōng yard inside ‘The two-liter bowl of the Duke of Thunder was put into the yard’.    b 長文

尚

小，

載 著 車 中。 《Shìshuōxīnyǔ   Déxìng; 5th c》 chángwén shàng xiǎo zài zhuó chē zhōng Chang.Wen yet little ride prep car inside ‘(Since) Changwen was still young, he rode in the car’.

At this stage the original lexical meaning ‘attach’ of zhuó had been bleached to the point that only the sense of location remained, and it did not select its own argument, making it more like a locative preposition. In general, location NPs are sandwiched between Chinese locative prepositions and a locative word that indicates the orientation of the position, such as zhōng ‘inside’ in (12). This is also the case for zhuó. Along with the grammatical change, zhuó more frequently combined with more locative words, such as qián ‘front’, hòu ‘back’, shàng ‘up’, xià ‘down’, and zhōng ‘middle’ (Wang 1958). This change of distribution and the loss of the semantic content of zhuó indicates that in the WJSN period the lexical verb zhuó was on its way to becoming a more functional preposition. However, the fact that the complement of zhuó was still locational shows that its original meaning still had an impact on the contexts where it could exist (Hopper and Traugott 2003:3). Before the WJSN period, the attachment meaning of zhuó entailed some kind of movement event, as shown in (8b). However, during the WJSN period, the attachment meaning and the movement meaning began to diverge and each developed distinct newer functions (Ma 2002). The motion sense ‘move’ or ‘put’ obtained a locative function, whereas the attachment sense potentially contributed to the emergence of the postverbal complement zhuó, which indicates continuation or completion of the event denoted by the preceding verb (Zhang 2002). This meaning split of zhuó and the prevalence of these new functions in the WJSN period suggests that this was a time period of rapid linguistic changes.

The obligative modal tioh8  61 Although zhuó could be used as a locative preposition in the WJSN Period, it was not a full-fledged preposition until the period of EMC. Three distributions found from the EMC period onward indicate that zhuó became a full-fledged preposition during this stage (i.e. Cao 1995; Zhang 2002). First, the prepositional phrase headed by zhuó can appear in preverbal position, like other typical prepositions in this period, as shown in (13a). Second, zhuó did not constrain the kind of context it could show up in. Therefore, for example, the verb qiú with the abstract meaning ‘to beg’ could occur with zhuó. Finally, and most importantly, the noun phrase that was taken by the preposition zhuó as a complement was extended from locations to instruments, as shown in (13c): (13) a 卻 què

著 言辭 怪 主人。 《Dūnhuángbiànwénjí   58; 8th c》 zhuó yáncí guài zhǔrén conj ins words blame host ‘However, (he) used (his) words to blame the host’. [lit. He orally blamed the host]

b 師 曰 不 著 佛 求。 《Zǔtángjí 4; 10th c》 shī yuē bù zhuó fó qiú mentor say neg to Buddha beg ‘The mentor said “Do not ask (for things) from Buddha” ’. 有 人 將 此 月輪 相 來 問   c 若 ruò yǒu rén jiāng cǐ yuèlún xiàng lái wèn if ext person ins dem moon image come ask 相 中心， 著 牛 字 對 也。《Zǔtángjí 19; 10th c》 xiàng zhōngxīn zhuó niú zì duì yě Buddha’s.image center ins cattle word reply prt ‘If anyone comes to you with the image of the moon asking for the image of Buddha, reply with the word cattle’.

After its emergence, the instrumental use of zhuó continued competing with the earlier locative use of zhuó until the older use was completely replaced, which occurred around the LMC period (Zhang 2002:215). Later, in the beginning of the ModC period, the use of zhuó as an instrumental preposition became dominant, and zhuó was able to select for abstract complements, as shown in (14) (Zhang 2002:225): 體例 (14) 百姓 著 甚麼 行? 《0Xiàojīngzhíjiě   54; 18th c》 bǎixìng zhuó shénme tǐlì xíng civilian ins what criteria behave ‘By what criteria should the civilians behave’? The dominance of zhuó as an instrumental preposition did not prevent it from being replaced by other functions over time, as we will see in the next section.

62  Ting-ting Christina Hsu During this period, zhuó also obtained the function of a postverbal complement. Similar to the preposition zhuó, the postverbal complement zhuó did not contribute to the lexical content of the sentence. Rather, it was generally used to indicate the aspectual continuation or completion of the event (Jiang 2005; Lien 2001). This emergence of the aspectual complement was initiated in the EMC period (around 5th-7th.cc). Examples are as follows: (15) a 如 無 燈 夜 把著 枕子。 《Zǔtángjí   5; rú wú dēng yè bǎzhuó zhěnzǐ 10th c》 similar.to neg light night hold-ach pillow ‘(It is) like holding a pillow in the night without the help of any light’.  b 雨

中 溪 破 無 乾 地， 浸著 床頭 yǔ zhōng xī pò wú gān dì jìnzhuó chuángtóu Rain middle stream break NEG dry land soak-ach bed.front 濕著 書。 《Wáng Jiàn Yǔzhōngjìdōngxīwéichǔshì; 9th c》 shīzhuó shū wet-ach book ‘In a day with heavy rain the stream broke so there was no dry land around. As a result, the front of my bed got soaked and made my book wet’.

 c 自說 孤 舟 寒 水 畔， 不 曾 逢著 zìshuō gū zhōu hán shuǐ pàn bù céng féngzhuó self.say alone boa chill water edge neg ever meet-ach 獨 醒 人。 《Dùmù Zèngyúfǔ; 9th c》 dú xǐng rén sole awake person ‘(The fisherman) said to himself “Since I began drifting alone in a boat on the chilly water, I’ve never seen a single waking person” ’. (15a) exemplifies zhuó functioning as an aspectual marker, indicating the continuation of the event ‘hold’ and in (15b), zhuó signifies that there was a change of state after the completion of the events ‘dip’ and ‘wet’ and that the state persisted. Some might argue that in both examples the original meaning ‘attach’ of zhuó appears to play roles in selecting the verb that zhuó modifies, since the resulting states of the actions involve contact between two things, such as ‘hand’ to ‘pillow’ in (15a) and ‘water’ to ‘book’ in (15b). However, in (15c) there is no real contact, which indicates that the original meaning of zhuó continued to be bleached during this period. The change of Chinese zhuó from verb to preposition is illustrated in Figure 3.3:

The obligative modal tioh8  63 PERIOD

OC~ ~200ad

WJSN 200~600ad

EMC 5th-7th.cc

著zhuó ‘attach’

verb as ‘put’

postverbal preposition + LOC

preverbal preposition + INSTRU

Figure 3.3  Change of zhuó from verb to preposition

3.2.2  From preposition to deontic modal Around 700ad Chinese zhuó had obtained its full-fledged status as a preposition. It could be put in a preverbal position, like most prepositional phrases at that time, and it did not put special constraints, such semantic selectional restrictions, on its complements. The prepositional zhuó selected both concrete and abstract complements, as shown in (13) and (14). Then, zhuó extended its pool of complements to animate entities, as shown in (16): (16) a

恁 三個 裏頭， 著 這 老的 看 者。 《Nogeoldae 25; 14th c》 nín sāngè lǐtóu zhuó zhé lǎode kàn zhě 2sg three-cl among ins dem old-nmlz watch prt ‘Among the three of you, this old man was selected to be the night watch’.

b 到 dào to 箇， gè

這裡的 半夜 前後， 卻 著 兩 bànyè qiánhòu què zhuó zhèlǐde liǎng mid-night front.back conj ins here-nmlz two 替 迴 來。《Nogeoldae 43; 14th c》 tì huí lái cl substitute return come ‘Around midnight, (he) sent two of us there to switch out with them’.

c 文契 著 誰 寫? 《Nogeoldae 66; 14th c》 wénqì zhuó shéi xiě who write word.contract ins ‘Who should (I pick to) write the contract’? The above examples are all selected from the oldest edition of Nogeoldae 老乞 大, a popular book that reflects the colloquial language used in Northern China during the early EMC period (14th – 16th c.) (Takekoshi 2007a). As shown in (16a-c), the force meaning of the preposition zhuó is reminiscent of the meaning expressed by the causative verb zhuó. In both cases, the entity denoted by the

64  Ting-ting Christina Hsu noun that follows zhuó is forced to do the action assigned by the subject, or at least is under the imposition of doing that action. According to Takekoshi (2007b), in the oldest edition of the extant Nogeoldae, zhuó had already been used to introduce a Causee within the construction [zhuó+NP+V], and alternated with another causative verb jiào 教 in later editions of Nogeoldae.5 The alternating behavior and the force meaning that zhuó had during the EMC period indicates that zhuó was on its way to becoming a causative verb. The causative verb zhuó was fairly stable and prevalent by the Early ModC period. We can even find zhuó used with a causative function in colloquial plays during the Early ModC period, such as a range of collections of Yuánqǔ 元曲 ‘Playscripts in Yuan Dynasty’, which was popular around the 13th to the 14th c. (17) a 我 wǒ 1sg

著 你 但 zhuó nǐ dàn caus 2sg dem

去處 qùchù place

誰 著 你 迤逗地 胡亂 shéi zhuó nǐ yǐdòudì húluàn who caus 2sg defer-adv random ‘I sent you there to monitor and guard around arbitrarily’? b 紅娘 看 熱 hóngniáng kàn rè Madame immediately warm

行監 坐守， xíngjiān zuòshǒu execute.monitor hold.guard 行走? 《Xīxiāngjì   4:2; xíngzǒu 14th c》 walk (them), how can you just walk

酒 來! 著 小姐 與 jiǔ lái zhuó xiǎojiě yǔ wine come let girl and

他 哥哥 把盞 者! 《Xīxiāngjì 2:3; 14th c》 tā gēgē bǎzhǎn zhě 3sg old.brother hold.cup prt ‘Warm (a pot of) wine here now, Madame! Send one girl here to enjoy the wine with me’.

c 你

nǐ 2sg

也 yě conj

态 tài too

老實， lǎoshí homebred

員外 yuánwài lord

著 zhuó caus

你 nǐ 2sg

難道 nándào

著 zhuó

counter

caus

你 nǐ 2sg

死， sǐ die

了

不

成?

你 nǐ 2sg

就 jiù conj

跪， guì kneel.down

你

就

死

跪， guì kneel.down

《Shāgǒuquànfū   2; 14th c》

nǐ jiù sǐ le bù chéng 2sg conj die prt neg success ‘You honest folks! When the Lord forced you to kneel down, you kneel down. If he asked you to kill yourself would you kill yourself’?

The obligative modal tioh8  65

d 則 zé conj

合 hé should

帶月披星， dàiyuèpīxīng carry.moon.cover.star

誰 shéi who

著 zhuó caus

你 nǐ 2sg

停 tíng stop

眠 整 宿? 《Xīxiāngjì 4:2; 14th c》 mián zhěng xiǔ sleep all night ‘You are supposed to work all night long, who ordered [allowed] you to stop (working) and sleep all night’? The examples in (17) are all from the scripts of the 13th.c popular dramas, such as Xīxiāngjì ‘Romance of the Western Chamber’. Since these sources are more colloquial, we can get a clearer understanding of the contexts in which zhuó was used. In these examples, the speakers are making requests using the word zhuó, suggesting that the word was becoming more of a directive and more similar to a causative marker than an instrumental preposition. Particularly, in (17c), the action requested is expressed by an unaccusative verb sǐ ‘die’. In this case, the only possible candidate for the sole argument of sǐ ‘die’ is the nǐ ‘you’ introduced by zhuó. If zhuó was simply an instrumental preposition and nǐ ‘you’ was the complement of zhuó, then nǐ ‘you’ would be optional and we would expect that sǐ ‘die’ would have no argument. However, since sǐ ‘die’ does select an argument, zhuó must be interpreted as a causative verb. The emergence of a causative meaning can presumably be attributed to the extension of the NP complements of zhuó from inanimate entities to animate ones. Along with this change, a sense of authority and imposition of that authority also emerged, rendering zhuó more like a causative verb. However, a change from a preposition to a causative verb seems to violate the tendency of unidirectionality (Hopper and Traugott 2003) which claims that if we account for the change based on the theory of grammaticalization then the change should always go from a more contentful item to a more functional item. This problem can be solved if we regard this process not as grammaticalization but as reanalysis, that is, a kind of rebraketing that changes the deep structure without changing the surface (Langacker 1977, 1995; Hopper and Traugott 2003). The process is illustrated in Figure 3.4:

zhuó NP] V] 著 你] 跪] INS you kneel.down ‘(He) used you to kneel down’. [VP[PP [VP[PP

> >

[VP zhuó [VPNP V] [VP 著 [VP你 跪] CAUS you kneel.down ‘(He) forced you to kneel down’.

Figure 3.4  Reanalysis of zhuó in the construction [zhuó NP V]

Reanalysis is often seen in languages without morphological markings, such as Chinese, since it involves creating new interpretations out of an already existing combination of words without making any changes in surface representation.

66  Ting-ting Christina Hsu Examples in (17) show that during the Mongol-ruled Yuan Dynasty (1271– 1368), which was between the LMC period (10th – 13th c.) and the ModC period (14th – 20th c.), the causative use of zhuó was very common. However, we can hardly find instances of causative zhuó before the 13th c. Instead, only instrumental zhuó prepositions can be found in the LMC period (10th – 12th c.) (Zhang 2002). One of such examples is shown in (18): (18) 無

事

時，

便

著

存養

收拾 此 心。 《Zhūzǐyǔlèi 2779; 13th c》 wú shì shí biàn zhuó cúnyǎng shōushí cǐ xīn neg matter time conj inS cultivation purify dem mind ‘When you have time, purify your heart by cultivating yourself’.

Despite the fact that in the LMC period zhuó only possessed the instrumental function and had not yet been able to take animate complements, we did find some occasional uses of an animate object complement at the end of the LMC period. This signals the initiation of the extension from inanimate object to animate object, as shown in (19) (Zhang 2008): (19) 畫鼓 輕 敲， huàgǔ qīng qiāo drawing.tambourine gently beat 著 個 人 zhuó gè rén ins.p cl person ‘The tambourine was beaten gently, people under the spell of the music’.

紅牙 隨 應， hóngyá suí yìng castanet immediately respond 勾喚。 《Yángwújiù   gōuhuàn Yǔzhōnghuālìng; attract.request 12th c》 followed by a castanet, which lured

The use of causative tioh8 (i.e. Chinese zhuó) can also be found in the scripts of Li Jing Ji/Li Zhi Ji ‘Legend of the Litchi Mirror’. An example of causative tioh8 found in the earliest version of Li Jing Ji is shown in (20): (20) 我

著

人

就

去

提來

拷問，

gua2 tioh8 lang5 tsiu7 khi3 theh8lai5 kho2bun7 1sg caus person conj go arrest.come heckle ‘I’ll send someone to arrest and heckle him right now’.

發落 伊。 (51.159 JJ) huat4loh8 i1 dispose 3sg

However, this causative use of tioh8 in Li Jing Ji is rare. Only one case can be found in the earliest edition of Li Jing Ji and no case of causative tioh8 can be identified in later editions of Li Jing Ji. The scarcity of the causative verb tioh8 during this period indicates that the newly-emerged deontic modal tioh8, which was thriving at that time and is still very popular today, was ready to replace this

The obligative modal tioh8  67 old causative use. Some examples of the deontic modal tioh8 used in Li Jing Ji are shown in (21): (21) a 益春， 你 著 關防 ik4tshun1 li2 tioh8 kuan1hong5 Iktshun 2sg have.to be.cautious.of ‘Iktshun, you should be cautious of him’.

伊。 i1 3sg

(14.448 JJ)

 b 做 tso4

簡 都 著 乞 人 使。 (22.137 JJ) kan2 to1 tioh8 khit4 lang5 su2 cop servant conj have.to pass person command ‘Being a servant, you always need to follow others’ orders’.

 c 啞娘， 你 著 去 賠 伊 人 性命 (29.071 JJ) a1niu5 li2 tioh8 khi2 pue5 i1 lang5 senn2mia7 ma’am 2sg have.to go pay 3sg person life ‘Ma’am, you have to pay for his life’. Up to this point, we have gone through the history of tioh8 in SM and zhuó in Chinese. A simple sketch of the development is shown in Figure 3.5, which expands on the story outlined in Figure 3.3. OC~WJSN (~200ad)

WJSN (200~600ad)

EMC (5th~7th.cc)

Early ModC (13th~15th.cc)

Late ModC (17th~20th.cc)

Verb ‘attach’

Preposition + LOC (postverbal)

Preposition + INS (preverbal)

Causative verb

Deontic modal

Figure 3.5  Grammaticalization through revival of the previous cline

4  EMERGENCE OF THE DEONTIC MODAL TIOH8 Thus far, the historical data has supported a grammaticalization path for tioh8 such that it changed from its earliest verbal function to its later prepositional function throughout several stages. After being grammaticalized as a preverbal preposition, it was then reanalyzed as a causative verb. The causative verb tioh8 was later changed into a deontic modal, which continues to be used in today’s SM. One immediate obstacle to motivating the change from a causative verb to a deontic modal is that they have distinct distributions and different relationships to the arguments in a sentence. For one thing, a causative verb can potentially take up to three arguments: a Causer, a Causee, and a Theme; however, a deontic modal does not take any arguments. Further, in a typical causative sentence

68  Ting-ting Christina Hsu the speaker does not intervene in the transmission of force. If this was the case, the sentence would be more like an imperative. However, a deontic modal can express the speaker’s subjective attitude towards a proposition. Hence, when a speaker uses a deontic modal it is often interpreted as them agreeing with the authority of the imposition that the modal introduces. Regardless of these differences, the causative verb and the deontic modal are similar to each other in that they both possess a force-bearing participant that can impose that force on other participants. A causative sentence has a super Agent, i.e. the Causer, that can order and burden the Causee. Similarly, in a sentence with a deontic modal, a speaker that upholds the authority of an obligation is putting pressure on the subject of the sentence. Now, to derive a deontic modal from a causative verb, we need to make sure that there is a good reason for the Causer to be removed from the sentence, while at the same time preserving the force of the imposition for the deontic modal. I argue that the reason for this change is subjectification, through which a speaker becomes more involved in expressing a proposition (Traugott 1989; Langacker 1990). During the development of the causative tioh8, subjectification would have removed the Causer from the sentence, increasing the speaker’s involvement in the proposition. As a result, the relationships between the immediate participants, specifically between the speaker and the subject of the sentence, were strengthened. The change from the causative verb tioh8 to the deontic modal tioh8 is illustrated in Figure 3.6, based on example (20). I repeat example (20) here as ((22)a) for easy reference and the deontic modal counterpart is stated in ((22)b) for comparison:6 (22) a 我

著

人

就

去

提來

拷問，

gua2 tioh8 lang5 tsiu7 khi3 theh8lai5 kho2bun7 1sg caus person conj go arrest.come heckle ‘I’ll send someone to arrest and heckle him right now’.

b 伊 著 去 提來 拷問， i1 tioh8 khi3 theh8lai5 kho2bun7 3sg deon go arrest.come heckle ‘He has to arrest and heckle him right now’.

發落

伊。 (51.159 JJ) huat4loh8 i1 dispose 3sg

發落 huat4loh8 dispose

伊。 i1 3sg

During the first step of this change, the matrix subject, i.e. the Causer, of the causative tioh8, is removed from the sentence, which builds a stronger link between the speaker and the sentence. After the operation of subjectification, the vacant subject position left by the original Causer triggers the original Causee to be raised to the subject position of the sentence. Along with this change, the verbal status of the causative verb tioh8 is weakened, since it has no argument relationship with the new subject. The original embedded verb then becomes the new matrix verb, and the sentence is reanalyzed as monoclausal. Despite this reanalysis, the force that the Causer originally exerts on the Causee is well preserved and

The obligative modal tioh8  69 INPUT

我 gua I

Original structure

著 2

NP1

伊

去

tioh CAUS

i him

khi3 go

V1

[NP2

V2

]

]

8

CAUSER

1

CAUSEE

STEP 1: Subjectivize the Causer (If I=speaker)

Ø

V1

[NP2

V2

STEP 2: Profile the Causee

AGENT NP2

V1

Ø

V2

STEP3: Reanalyze as monoclausal

[NP

Modal

V

OUTPUT

伊

著

去

i

1

he

tioh

8

has to

]

kui7 go

Figure 3.6  From causative verb to deontic modal

continues to exist between the speaker and the new subject of the modal sentence. The arguments that bear force are marked with ovals in Figure 3.6. The output of this change reflects how tioh8 is used in contemporary SM. With subjectification, the original Causer is backgrounded and coindexed with the speaker, and as a result the participants in the sentence make stronger associations with the immediate context. Removing a subject from a sentence by subjectification is one of the documented mechanisms of diachronic change (Traugott 1995:47). Other changes include changing a modal from a non-epistemic to an epistemic one and creating a syntactic subject from a non-syntactic subject, as shown in Figure 3.7 (Traugott 1995:47): a. b. c. d. e. f.

Propositional function Objective meaning Non-epistemic modality Non-syntactic subject Syntactic subject Full, free form

Discourse function Subjective meaning Epistemic modality Syntactic subject Speaking subject Bounded form

Figure 3.7 Diachronic Continua for subjectification in grammaticalization (Traugott 1995:47)

70  Ting-ting Christina Hsu Traugott’s changes d. and e., i.e. from non-syntactic object to syntactic object and from syntactic subject to speaking subject, account for steps 1 and 2 in the development of tioh8 from its causative function to its deontic modal function. These two steps participate in a chain reaction. When the original syntactic subject is removed from the sentence and becomes a speaking subject, a higher degree of subjectification is achieved, as a result the gap left behind by the original subject would then need to be filled. The best candidate for this substitution would be any original lower subject because the matrix sentence can then maintain the argument relationship between the original lower subject and its verb. By raising the lower non-syntactic subject to the position of syntactic subject, we reanalyze the sentence from having two subjects to possessing only one subject, i.e. a monoclausal sentence. The serial changes brought about by subjectification indicate the importance of a pivotal argument in the premodal stage, since the presence of this argument ensures that a non-syntactic subject can later be moved to a syntactic subject position. This pivotal argument is crucial not only syntactically and pragmatically, but also semantically. It preserves the force of imposition that was originally given by the matrix subject, i.e. the Causer, to the non-syntactic subject, i.e. the Causee, even after it raises to the subject position and is no longer a Causee (Talmy 2000). This kind of force, according to the ‘mind-as-body metaphor’, turns into a certain form of psychological pressure that the new subject is obliged to take (Sweetser 1990:28–48). The shift from causative verb to deontic modal supports the notion that the force interactions among their semantic components are structured similarly, even though they differ in the way that force is expressed syntactically. For the causative verb, the psychological force is exerted by the subject, i.e. the Causer, onto the non-syntactic subject, i.e. the Causee. Contrastively, in the case of a deontic modal the psychological force comes from the speaker, which represents, or at least complies with, the authority under which the syntactic subject is obligated. However, crucially, in both cases an argument in the subject position is involved in the force interaction, so this pivotal argument of the causative verb is very important in its change to the deontic modal. This is illustrated in Figure 3.8. In this paper I have presented the relevant historical data for the emergence of the deontic modal tioh8 in SM. We have seen that this function originated from

From causative verb to deontic modal CAUSATIVE

NP1

V1

[

NP2: CAUSEE

V2 ]

NP2: AGENT DEONTIC

[ NP: AGENT

Modal

[Speaker Interaction] Figure 3.8  The importance of a pivotal argument in the change

V ]

The obligative modal tioh8  71 a more lexical verbal use meaning ‘attach’, which developed into a prepositional function, was then reanalyzed as a causative verb, and ultimately became the deontic modal tioh8. Several mechanisms were involved in this development, such as grammaticalization, reanalysis, and subjectification. This trajectory of development, illustrated in Figure 3.5, is reiterated here in Figure 3.9, complete with the mechanisms that were involved at various stages. OC~WJSN (~200ad)

WJSN (200~600ad)

EMC (5th~7th.cc)

Early ModC (13th~15th.cc)

Late ModC (17th~20th.cc)

Verb ‘attach’

Preposition + LOC (postverbal)

Preposition + INSTRU (preverbal)

Causative verb

Deontic modal

-------------GRAMMATICALIZATION----------------------REANALYSIS--------------SUBJECTIFICATION--Figure 3.9  Grammaticalization through revival of the previous cline

5  CONCLUDING REMARKS In this article I have demonstrated the path of development of the deontic modal tioh8 in SM. It originated from a lexical verb meaning ‘attach to’, which then extended to ‘put’, which could select a great variety of locational NPs as its complement. At this stage, tioh8 was also used with other verbs to make complex predicates. The fact that tioh8 was in a less stable position, i.e. the second position, of a complex predicate made it more vulnerable to change and it eventually reduced to a more functional item. After losing its semantic content, the lexical verb tioh8 became a preposition heading a locative PP, with the whole PP being located after the main verb. This PP later moved to a preverbal position and gained an instrumental function, eventually selecting animate objects. From there, a new causative use of tioh8 was generated by reanalyzing the whole preverbal structure. The causative tioh8 ultimately developed into a deontic modal after subjectification. Although tioh8 developed from a lexical verb to a more functional preposition and then back to a causative verb, the unidirectionality of grammaticalization, i.e. to always go from a more contentful item to a more functional item in a grammatical change (Hopper and Traugott 2003:68), is not violated because different mechanisms were involved at different stages in the process of change. During the first stage of the development, the lexical verb tioh8 was grammaticalized into an instrumental preposition. Then, at the second stage, the preposition was reanalyzed as a causative verb, which finally changed into a deontic modal after subjectification. The three mechanisms that were involved in the development of the deontic modal tioh8, namely grammaticalization, reanalysis, and subjectification,

72  Ting-ting Christina Hsu worked independently yet successively to finally generate the deontic modal tioh8 in SM. The interweaving syntactic, semantic, and pragmatic mechanisms thus collaboratively contributed to the birth of the deontic modal tioh8. The fact that the causative verb tioh8 and the deontic modal tioh8 share a force meaning specifically helped in the change from one to the other, although this force meaning operates differently for the two functions. In the causative construction, the force is exerted by the Causer onto the Causee, whereas in a deontic modal sentence the force interaction is between the speaker and the subject. The shared force meaning conceptually links the two different functions of tioh8 together, making the change from one function to the other less costly, since the cognitive schema of the force interaction before and after the change remains the same and the only difference is which participant is a more salient subject to the speaker (Talmy 2000). The deontic modal tioh8 is not the only modal that can be traced back to a causative origin in SM. Another deontic modal ai3愛, which also means ‘should’ or ‘have to’, underwent similar mechanisms of change (Hsu 2013). Interestingly, the two words are both young in their deontic modal functions and the time during which their causative uses began to change into deontic modals was quite similar, presumably around the 16th c. The parallel development of the two modals suggests that speakers of SM at that time were interested in the immediate context of the utterance, so they preferred expressions which associated the participants of an event to the current speech context. It is also likely that people at that time favored expressions which contained a low degree of force meaning, such as a deontic modal, compared to causative verbs. As a result, new forms and functions were created to express less coercive meanings and less forceful interpersonal relationships between the immediate participants of a conversation. In future work we might expect to see more examples from other dialects of Chinese, or even from other languages, where similar mechanisms create functional change during this time period. NOTES * This paper was partially sponsored by the Ministry of Science and Technology (NSC 98–2923-H-007–001-MY3). An earlier version of this paper has been announced at the workshop Reflections of diachronic change mirrored in early Southern Min texts held at National Tsing Hua University in Taiwan in November, 2011. I thank all the audiences for their helpful comments, especially Prof. Alain Peyraube, Prof. Hilary Chappell, Prof. Chinfa Lien, and Prof. Barry C-Y. Yang. I am also thankful for Kristin G. Denlinger’s help and suggestions at the final stage of the paper. All remaining inadequacies are my own. 1 There are at least four arguments against equating reanalysis and grammaticalization according to Lehmann (2005). First, though reanalysis might be related to phonological reduction and the increase of cohesion, it is not necessary for reanalyzed units to have undergone desemanticization or the loss of semantics. Second, though reanalysis always involves category changes, not all category changes are regarded as instances of grammaticalization. Third, reanalysis is concerned with the categorial change of units, which might coincidentally parallel certain grammaticalization clines, however grammaticalization is naturally a gradual unidirectional process. Finally, grammaticalization

The obligative modal tioh8  73 may create brand-new structures and categories in a language, whereas reanalysis cannot create anything that is genuinely new because it is motivated by an analogical model (Lehmann 2005:164–165). 2 The early Southern Min playscripts recording the Li Jing Ji/Li Zhi Ji ‘Legend of Litchi Mirror’ comprise four editions (1566, 1581, 1651, and 1884). See Wu 2001a, 2001b, 2001c, 2001d. 3 The idiosyncratic behavior of tioh8 著 and ai3 愛 in negative formation can be attributed to their developmental history. Hsu (2010) finds that the deontic modal ai3 in SM newly emerged from its earlier volitive use. The long history of the non-modal uses of these two modals has set constraints on their negative formation. For example, the negative marker m7 had already been used with the non-modal uses of tioh8 ‘right’ and ai3 ‘like’ to express ‘not right’ and ‘not like’ and another negative marker bo5 had been used with the verbal uses of tioh8 ‘hit the target’ and the volitive use of ai3 ‘want’ to indicate ‘not at the target’ and ‘not want’. The already existed combinations of negative markers and premodal forms are assumed to prevent identical negative markers from combining again with the later generated modals to avoid ambiguity. 4 The SM tioh8 as ‘hit, acquire, and suffer’ in Yang’s (1992) paper is transcribed as tioq8, which is assumed to be related to the ancient 著 meaning ‘to attach’ 著. In the rhyme book Guangyun (902–1127) it is recorded as an entering tone (入聲) and belongs to the rhyme Yao (藥) /ĭak/, which reflects the pronunciation of 著 in SM instead of the pronunciation of 著 in contemporary Chinese. 5 However, only 1 out of the 10 causative uses of zhuó was replaced by the typical causative verb jiào叫. Nevertheless, the interchangeability between jiào and zhuó still suggests that zhuó had obtained a causative function at that time. 6 Minor changes are made to the original example (20) to facilitate this explanation. For example, the original Causee is an unspecified pronoun, which is not appropriate when it becomes the subject of a deontic modal. Therefore, in (22)b we use a specified pronoun as the subject of the sentence, which, if being recovered, would be equal to a specified third person singular object of the causative verb tioh8. Also, irrelevant details are removed from the figure to keep it clear and focused.

REFERENCES Brinton, Laurel J., and Elizabeth Closs Traugott. 2005. Lexicalization and language change. New York: Cambridge University Press. Bussmann, Hadumod. 1996. Routledge dictionary of language and linguistics. New York: Routledge. Cao, Guang-shun 曹廣順. 1995. Jìndài Hànyǔ zhùcí 近代漢語助詞 [Particles in modern Chinese]. Beijing: Yuwen Chubanshe. Heine, Bernd. 1993. Auxiliaries: Cognitive forces and grammaticalization. Oxford: Oxford University Press. Hopper, Paul. 1991. On some principles of grammaticalization. In Approaches to grammaticalization, ed. by Elizabeth Closs Traugott and Bernd Heine, 17–35. Philadelphia: John Benjamins. Hopper, Paul J., and Elizabeth Closs Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press. Hsin, Aili 忻愛莉. 1999. Modality in Taiwanese Southern Min. Doctoral dissertation, National Tsing Hua University, Hsinchu. Hsu, Ting-ting Christina. 2010. From volition to obligation: Deontic modal 愛 ai3 in TSM. Lecture given in Ling-talk at National Tsing Hua University, Hsinchu.

74  Ting-ting Christina Hsu Hsu, Ting-ting Christina. 2013. Subjectification and the emergence of deontic modal verbs TIOH8, AI3, and E7 in Southern Min. Doctoral dissertation, National Tsing Hua University, Taiwan. Jespersen, Otto. 1924. The philosophy of grammar. London: Allen and Unwin. Jiang, Shao-yu 蔣紹愚. 2005. Jìndài Hànyǔ yánjiū gàiyào 近代漢語研究概要 [The handbook of studies on Modern Chinese]. Beijing: Peking University Press. Koelle, Sigismund Wilhelm. 1968 [1854]. Outlines of a grammar of the Vei languages: Together with a Vei-English vocabulary and an account of the discovery of the Vei of syllabic writing. London: Church Missionary House. Reprinted: Farnborou: Gregg International Publishers. Langacker, Ronald W. 1977. Syntactic reanalysis. In Mechanisms of syntactic change, ed. by Carles N. Li, 57–193. Austin: University of Texas Press. Langacker, Ronald W. 1990. Subjectification. Cognitive Linguistics 1:5–38. Langacker, Ronald W. 1995. Possession and possessive construction. In Language and the cognitive construal of the world, ed. by John R. Tylor and Robert E. MacLaury, 51–80. Berlin: de Gruyter. Beijing: Zhonghua Book Company. Lehmann, Christian. 2002. New reflections on grammaticalization and lexicalization. In New reflections on grammaticalization (TSL 49), ed. by Ilse Wischer and Gabriele Diewald, 1–18. Philadelphia: John Benjamins. Lehmann, Christian. 2005. Theory and method in grammaticalization. Zeitschrift für Germanistische Linguistik 32:152–187. Li, Ru-long李如龍, and Yao, Rong-song姚榮松. 2008. Mǐnnán fāngyán 閩南方言 [Dialects of Southern Min]. Fuzhou: Fujian People’s Publishing House. Lien, Chin-fa. 2001. The semantic extension of tioh8 著 in Taiwanese Southern Min: A construction-based account. Language and Linguistics 2:173–202. Lord, Carol Diane. 1989. Syntactic reanalysis in the historical development of serial verb construction in languages of West Africa. Doctoral dissertation, University of California, Los Angeles. Lyons, John. 1977. Semantics. Cambridge: Cambridge University Press. Ma, Beijia 馬貝加. 2002. Jìndài Hànyǔ jiècí 近代漢語介詞 [Prepositions in modern Chinese]. Beijing: Zhonghua Book Company. Mei, Tsu-lin 梅祖麟. 2000. Hànyǔ fāngyán xūcí lǐ zhe zì sānzhǒng yòngfǎ de láiyuán 漢 語方言里虛詞“著”字三種用法的來源 [On the sources of the three grammatical functions of zhe 著 in Chinese dialects]. In Méizǔlín yǔyánxué lùnwénjí 梅祖麟語言學論文 集 [The collection of papers from Zhulin Mei]. Beijing: The Commercial Press. Meillet, Antoine. 1912. L’évolution des formes grammaticales. Scientia 12:130–148. Palmer, Frank R. 1986. Mood and modality. Cambridge: Cambridge University Press. Sweetser, Eve. 1990. From etymology to pragmatics. Cambridge: Cambridge University Press. Takekoshi, Takashi. 2007a. Cóng Lǎoqǐdà de xiūdìng láikàn zhe de gōngnéng yǎnbiàn從 《老乞大》的修訂來看著的功能演變(上) [On the functional change of 著 from the revision of Nogeoldae (Part 1)]. Kotonoha 58:6–15. Takekoshi, Takashi. 2007b. Cóng Lǎoqǐdà de xiūdìng láikàn zhe de gōngnéng yǎnbiàn 從 《老乞大》的修訂來看著的功能演變(下) [On the functional change of 著 from the revision of Nogeoldae (Part 2)]. Kotonoha 59:1–8. Talmy, Leonard. 1988. Force dynamics in language and cognition. Cognitive Science 12:49–100. Talmy, Leonard. 2000. Toward a cognitive semantics. Vol. 1, Concept structuring systems. Cambridge: MIT Press.

The obligative modal tioh8  75 Traugott, Elizabeth Closs. 1982. From propositional to textual and expressive meanings: Some semantic and pragmatic aspects of grammaticalization. In Perspectives on historical linguistics, ed. by Winfred P. Lehmann and Yakov Malkiel, 245–271. Amsterdam and Philadelphia: John Benjamins. Traugott, Elizabeth Closs. 1989. On the rise of epistemic meanings in English: An example of subjectification in semantic change. Language 65:31–55. Traugott, Elizabeth Closs. 1995. Subjectification in grammaticalization. In Subjectivity and subjectivisation: Linguistic perspectives, ed. by Dieter Stein and Susan Wright, 31–54. Cambridge: Cambridge University Press. Traugott, Elizabeth Closs, and Richard B. Dasher. 2002. Regularity in semantic change. Cambridge: Cambridge University Press. Wang, Li 王力. 1958 [2003]. Hànyǔ shǐgǎo 漢語史稿 [Studies on diachronic Chinese]. Beijing: Science Press. Wu, Shouli 吳守禮. 2001a. Míng Jiājìng kān Lì Jìng Jì xìwén jiàolǐ 明嘉靖刊荔鏡記戲文 校理 [Annotated texts of the romance of Li Jing Ji of Ming Jiajing edition (JJ)]. Taipei: Congyi Workshop. Wu, Shouli 吳守禮. 2001b. Míng Wànlì kān Lì Zhī Jì xìwén jiàolǐ 明萬曆刊荔枝記戲文 校理 [Annotated texts of the romance of Li Zhi Ji of Ming Wanli edition (WL)]. Taipei: Congyi Workshop. Wu, Shouli 吳守禮. 2001c. Qīng Shùnzhì kān Lì Zhī Jì xìwén jiàolǐ 清順治刊荔枝記戲文 校理 [Annotated texts of the romance of Li Zhi Ji of Qing Shunzhi edition (SZ)]. Taipei: Congyi Workshop. Wu, Shouli 吳守禮. 2001d. Qīng Guāngxù kān Lì Zhī Jì xìwén jiàolǐ 清光緒刊荔枝記戲 文校理 [Annotated texts of the romance of Li Zhi Ji of Qing Guangxu edition (GX)]. Taipei: Congyi Studio. Yang, Hsiu-fang 楊秀芳. 1992. Cóng lìshǐ yǔfǎ de guāndiǎn lùn Mǐnnányǔ ‘zhe’ jí chíxùmào 從歷史語法的觀點論閩南語「著」及持續貌 [On Southern Min 著 and its continuative function from historical perspectives]. Hànxué Yánjiū 漢學研究 10:349–394. Zhang, Cheng 張赬. 2002. Hànyǔ jiècí cízǔ cíxù de lìshǐ yǎnbiàn 漢語介詞詞組詞序的歷 史演變 [The historical evolution of the word-order of prepositional phrases in Chinese]. Beijing: Beijing Language and Culture University Press. Zhang, Xiang 張相. 2008 [1953]. Shīcíqǔ yǔcí huìshì 詩詞曲語辭彙釋 [Decipher to the words in poems, lyrics, and plays]. Beijing: Zhonghua Book Co.

4 Negation of dynamic modals with DIT 得 in Hainan Min Huichi Lee 李惠琦

1  INTRODUCTION The present study explores the modal constructions involving the word ɗit55 得 ‘get, acquire’ in Hainan Min, focusing on their negative forms. Moreover, dynamic modality is especially studied. The paper mainly discusses the modality associated with potentiality and ability rather than volition. I present data from a synchronic case study of the word ɗit55 ‘get, acquire’ in Hainan Min, while bringing in historical evidence from other Chinese dialects (Southern Min, Hakka, Cantonese, and Mandarin) to account for the special modal forms in Hainan Min. Issues arise concerning the modality involving the word ɗit55 ‘get’ in Hainan Min. Modality is assumed to refer to the status of the proposition and a ternary distinction is often made to analyze modal systems: epistemic, deontic, and dynamic. According to Palmer (2001), epistemic modality refers to the certainty that the speaker expresses towards the truth of the proposition; deontic modality deals with the force exercised on the subject of the sentence to perform an action and dynamic modality encodes ability and willingness revealed by the subject of the sentence. The division of these three types of modality has recently been adopted by linguists like Tsai (2010) and Lien (2011a, 2011b) investigating Chinese languages. Among the three modal types, the dynamic modality is expressed in English by can and will (ability and volition). This paper particularly pays attention to the dynamic modality corresponding to the can expression. Following Palmer (2001), Lien (2011a) points out that the can dynamic modality can be further divided into two types: inherent and circumstantial dynamic modals. The inherent dynamic modality refers to the subject’s ability to conduct actions; the circumstantial dynamic modality refers to the possibility of the subject’s ability to conduct actions. The former emphasizes the inherent ability of the subject, while the latter focuses on the possibility offered in a given circumstance. Based on the previous studies, this paper categorizes the modal system as follows: epistemic modality deontic dynamic willingness (will) ability (can) inherent circumstantial

Negation of dynamic modals  77 In this modal system, the inherent and circumstantial dynamic modals are of particular concern in the present study. In addition to the dynamic modality, this paper aims to explore the data in Hainan Min involving the word ɗit55 ‘get’. Most Chinese dialects share a common verb indicating ‘get, acquire’, rendered as de in Mandarin, as dak in Cantonese, as tet in Hakka, as tit in Southern Min, and as ɗit in Hainan Min. This ‘get, acquire’ morpheme in Chinese dialects shows a similar range of lexical and grammatical functions: a verb, a part of compound verb, an aspectual marker, a complement marker, and a modal marker. For example, the Mandarin de can serve as a transitive verb (e.g. dé dào dà jiǎng ‘get a big prize’), an aspectual marker (e.g. wǎnfàn zuò dé le ‘have completed the dinner’), a state/degree/result complement marker (e.g. chī de hěn kuài ‘eat very fast’), an infix-like dynamic modal (e.g. ná de dòng ‘be able to take it’), and a post-verbal modal (e.g. chī dé ‘can be eaten’). The de constructions in Mandarin and other Chinese dialects are thus often popular issues for linguists to investigate (e.g. Lü 1944; Chao 1968; Lien 1997, 2011a, 2011b; Lamarre 2001a, 2001b; Liu 2001, 2002; Wei 2004; Wu 2009; Yue 2001a). However, ɗit constructions in Hainan Min have never been studied in previous research. Based on previous efforts and findings on the de constructions in several Chinese dialects, this paper attempts to investigate the syntactic and semantic properties of the Hainan Min ɗit constructions, especially when associated with the dynamic modality. At the same time, we hope the effort spent on the Hainan Min dialect can also help advance the study of the other dialects. 2  OVERVIEW OF DYNAMIC MODAL MARKER DIT Hainan Min is a branch of the Southern Min dialectal group. Hainan Min has been surveyed by several linguists, like Hashimoto (1959, 1960, 1961a, 1961b, 1976), Liang (1986), Chen (1996), Woon (2004), and Liu (2006, 2010). They described the language and mostly investigated its phonological issues. Due to their contributions, we basically comprehend this Southern Chinese dialect which has not been sufficiently studied for decades. Unlike the previous research, this paper aims to inquire into the syntactic structures and specifically into the modal issues in Hainan Min. This section, I present data from a synchronic case study of the polyfunctional verb ɗit55 ‘get’ in Hainan Min, concentrating on its modal usage. There are many researchers who have raised questions concerning the widespread pattern of grammatical polyfunctionality involving the verb ‘get’, with reference to Chinese dialectal data and Southeast Asian languages (e.g. Lü 1944; Chao 1968; Clark 1989; Matisoff 1991; Bisang 1996; Lien 1997, 2011a, 2011b; Lamarre 2001a, 2001b; Simpson 2001; Enfield 2001; Liu 2001, 2002). In Hainan Min, ɗit can be a verb, as in (1); a part of a verb compound, as in (2), (3), a postverbal modal, as in (4) and (5). (1) 伊 个 作文 一定 會 i44 kai22 toh55bun22 dziak33ɗia44 oi42 3sg gen work definitely will ‘His work will definitely win the award’.

得 ɗit55 get

獎。1 tsiang21 prize

78  Huichi Lee (2) 這 dzia11

本 ɓui21

書 無 值得 放 tu44 bo22 dzit33ɗit55 ɓang11 dem clf book neg worth-get put ‘This book is not worth taking the time to read’.

工 kang44 day

看。 mo44 see

(3) 囝 小 你 捨得 去 無？ kia21 niauh55 du21 tia21ɗit55 hu11 bo21 kid little 2sg abandon-get go neg ‘Do you feel okay to leave while the kids are still so young’? (4) 這 dzie11

件 kien42

事 小王 辦 se42 tiau21uang22 ɓan42 dem clf thing little-Wang handle ‘Little Wang can achieve this matter’.

得 ɗit55 get

成。 tsia22 succeed

(5) 伊 可能 贏 得。 i44 ho21neng22 ia22 ɗit55 3sg maybe win get ‘It is possible that he can win’. Mandarin de can serve as a marker of complex complement constructions such in resultative and descriptive expressions. Unlike Mandarin, ɗit in Hainan Min cannot perform these two functions. It cannot occur either in a descriptive expression, like (6), or in a resultative expression, like (7). (6) a *伊 食 得 只顧 i44 tsiah33 ɗit55 na42ku11 3sg eat get very ‘He ate himself very full’.

飽。 ɓa21 full

b 他 吃 得 很 飽。 tā chī de hěn bǎo 3sg eat get very full ‘He ate himself very full’.

(Mandarin)

(7) a *我 肥 得 共 大 gua21 ɓui22 ɗit55 kang44 ɗua44 1sg fat get like big ‘I am so fat that I am like a pig’. b 我 肥 得 像 豬。 wǒ féi de xiàng zhū 1sg fat get like pig ‘I am so fat that I am like a pig’.

豬。 ɗu44 pig

(Mandarin)

Negation of dynamic modals  79 The de constructions in many dialects in Chinese may have different phonetic varieties, but all show a strikingly similar range of functions. One of the common usages is to express the ‘potential’ or ‘possibility’ interpretation. This potential usage was pointed out by Chao (1968:452–457) to syntactically occur postverbally, either in a pattern of [V de C] or in a pattern of [V de]. The [V de C] form is frequently used in Chinese dialects. Its negative counterpart is [V bu C] where the word bu ‘not’ is to negate the potential or possibility of the complement C, exactly corresponding to the function of de. It is noteworthy that Hainan Min does not employ the common pattern [V bu C]. Instead of the [V Neg C] form, Hainan Min only accepts the [Neg V C] form. In addition to the [V de C] form, Hainan Min also uses the [V de] form to express modality and its negative counterpart is like the [Neg V de] form, rather than the common [V Neg de] form. A very interesting issue thus arises concerning these special negative forms of modality in Hainan Min. In this section, I provide a detailed survey of the negative forms of dynamic modality: [Neg V C] and [Neg V de].

2.1 [V 得 C] and [Neg V C] The [V 得 C] ([V de C]) form is largely found in Chinese dialects. It is rendered as [V ɗit C] in Hainan Min and exemplified by (8) and (9). (8)

伊 行 得 i44 kia22 ɗit55 3sg walk get ‘He can walk here’.

(9)

這 dzia11

條 ɗiau22

來。 lai22 come

題 我 ɗoi22 gua21 dem clf question 1sg ‘I can answer this question’.

作 toh55 do

得 ɗit55 get

出。 sut55 exit

As described above, the negative counterpart of the [V ɗit C] form is [Neg V ɗit C]. The negative morpheme bo22 無 precedes the sequence [V ɗit C], as shown in (10) and (11). (10) 糜 無 食 得 了。 mue22 bo22 tsiah33 ɗit55 liau21 rice neg eat get finished ‘The rice cannot be finished’. (11) 公 無 走 得 快。 kong44 bo22 tau21 ɗit55 huai11 grandfather neg run get fast ‘Grandfather cannot run very fast’.

80  Huichi Lee Furthermore, when the main verb in the [V ɗit C] pattern takes an object, the object does not occupy the normal position to form the [V ɗit C O] pattern, like in Mandarin or in most Chinese dialects. (12) Mandarin: [V de C O] a 煮 得 沸 水。 zhǔ de fèi shuǐ cook get boil water ‘can boil the water’ b 吹 得 破 氣球。 chuī de pò qìqiú blow get break balloon ‘can blow up the balloon’ c 吹 得 乾 chuī de kān blow get dry ‘can dry the hair’

頭髮。 tóufǎ hair

In Hainan Min, unlike Mandarin, the object is not acceptable to follow the verbal complement. (13) Hainan Min: *食 得 了 糜。 tsiah33 ɗit55 liau21 mue22 eat get finished rice ‘The rice can be finished’. Mandarin: 吃 得 完 飯。 chī de wán fàn eat get finished rice ‘The rice can be finished’. The object occurs preceding the verbal complement to form the [V ɗit O C] pattern, as shown in (14)-(16). (14) 煮 得 水 沸。 tu21 ɗit55 tui21 ɓui11 cook get water boil ‘The water can be boiled’.

Negation of dynamic modals  81 (15) 歕 得 氣球 擘。 bun22 ɗit55 hui11hiu22 ɓeh55 blow get balloon break ‘(Someone) is able to pop the balloon’. (16) 母 吹 得 頭毛 mai21 sui44 ɗit55 hau22mo22 mother blow get hair ‘Mother is able to dry her hair’.

乾。 kan44 dry

The negative form of the [V ɗit O C] pattern is also very different from the common form [V de C O] which corresponds to its negative form [V bù C O], exemplified by following Mandarin data. (17) Mandarin [V de C O] and [V bu C O] a 看 得 到 kàn de dào see get arrive ‘can see you’

你。 nǐ 2sg

b 看 不 到 kàn bú dào see neg arrive ‘cannot see you’

你。 nǐ 2sg

In Hainan Min, the negative forms do not follow the exchange-rule of de and bù. It shows a particular pattern that the negative morpheme bo always precedes the whole verb phrase, [bo V ɗit O C], rather than [V bo O C]. The examples are shown as follows: (18) 伊 無 搬/(抬) 得 行李 i22 bo22 ɓua22/ka22 ɗit55 hing22li21 3sg neg carry/lift get luggage ‘He is not able to carry the luggage out’. (19) 伊 無 寫 得 書 了。 i44 bo22 tia21 ɗit55 tu44 liau21 3sg neg write get book finish ‘He is not able to finish writing the book’.

出。 sut55 exit

82  Huichi Lee (20) 伊 無 拎 得 起 i44 bo22 ling22 ɗit55 hi21 3sg neg lift get rise ‘He cannot lift the hammer’.

鐵頭。 hih55hau22 hammer

2.2 [V 得] and [Neg V 得] The modal form [V ɗit] in Hainan Min is used to express the possibility or potential of the action conveyed by the verb, as shown in (21)-(23). (21) 爹 現旦 回 得。 de44 ɦien42nua11 ɗui21 ɗit55 father tomorrow return get ‘Father can return tomorrow’. (22) 這 dzia11

種 菜 食 tsiang21 sai11 tsiah33 dem type vegetable eat ‘This kind of vegetable is edible’.

得。 ɗit55 get

(23) 偌 苦 个 工 伊 都 作 得。 ua42 hou21 kai22 kang44 i44 ɗou44 toh55 ɗit55 how bitter clf work 3sg all do get ‘No matter how heavy the workload is, he can do it’. The negative form of the [V ɗit] follows the specific negative rule in Hainan Min in which the negative morpheme bo precedes the whole verb phrase. The negative counterpart of the modal [V ɗit] form is [Neg V ɗit]. Examples are shown as follows. (24) 這 dzia11

件 kien42

事 小王 應該 無 se42 tiau21uang22 eng44kai44 bo22 dem clf thing little-Wang should neg ‘Little Wang may not be able to accomplish this thing’.

辦 ɓan42 handle

(25) 母 氣 無 耐 得。 mai21 hi11 bo22 nai42 ɗit55 mother anger neg endure get ‘Mother’s anger cannot be controlled’. (26) 姊 共 姨嬙 無 合 tsi21 kang44 i22hiang22 bo22 kap33 elder.sister with younger.sister neg get.along ‘The sisters cannot get along with each other’.

得。 ɗit55 get

得。 ɗit55 get

Negation of dynamic modals  83 The [V ɗit] form can take an object which immediately follows [V ɗit] to form the [Vɗit O] sequence, as shown in (27) and (28). (27) 伊 忍 得 氣。 i44 nun21 ɗit55 hi11 3sg tolerate get anger ‘He can control his temper’. (28) 爹 大 个 有 力 只 掠 得 de44 ɗua44 kai22 u42 lat33 na42 liah33 ɗit55 father big gen have power only catch get ‘Father can catch the thief because he is strong enough’.

枚 mo42 clf

賊。 sat33 thief

The negative counterpart of the [V ɗit O] pattern is not surprisingly found to be the [Neg V ɗit O] form, as shown in (29)-(33). (29) 燈 照 無 開 得 目。 deng44 tsio11 bo22 hui44 ɗiet55 mak33 light illuminate neg open get eye ‘The light is so bright that one cannot open one’s eyes’. (30) 這 dzia11

多 話 無 謊 得 toi44 ue44 bo22 uang21 ɗit55 dem many word neg cheat get ‘He cannot be deceived by these words’.

(31) 伊 無 跳 得 過 bo22 hiau11 ɗit55 kue11 i44 3sg neg jump get over ‘He cannot jump over the ditch’.

伊。 i44 3sg

溝。 kau44 ditch

(32) 雞頭 無 飛 得 上 koi44hau22 bo22 ɓue44 ɗit55 tsio42 cock neg fly get up ‘The cock cannot fly up to the sky’. (33) 伊 無 過 得 i44 bo22 kue11 ɗit55 3sg neg pass get ‘He cannot stand the heat’.

過 gue11 over

天。 hi44 sky

熱。 dzuah33 heat

To sum up, the dynamic modal patterns in Hainan Min involving the marker ɗit basically contain two categories: V 得 (O)C [V ɗit (O) C] and V得(O) [V ɗit (O)].

84  Huichi Lee The findings of the dynamic modal distribution in Hainan Min are displayed as follows: (34) Distribution of Hainan Min dynamic modal marker DIT Types

Negation

Examples

V得C [Vɗit C]

無V得C [Neg Vɗit C]

i44 kia22 ɗit55 lai22/ i44 bo22 kia22 ɗit55 lai22 ‘He can/cannot walk here’.

V得OC [Vɗit OC]

無V得OC [Neg Vɗit OC]

tu21 ɗit55 tui21 ɗui11/ Bo22 tu21 ɗit55 tui21 ɗui11 ‘The water can/cannot be boiled’.

V得 [Vɗit]

無V得 [Neg Vɗit]

i44 ia22 ɗit55/ i44 bo22 ia22 ɗit55 ‘He can/cannot possibly win’.

V得O [Vɗit O]

無V得O [Neg Vɗit O]

i44 nun21 ɗit55 hi11/ i44 bo22 nun21 ɗit55 hi11 ‘He can/cannot control his temper’.

The four dynamic forms in Hainan Min shown in (34) are all patterned in the formula [Neg X]. The variable ‘X’ includes [V ɗit (O) C] and [V ɗit (O)]. The negative morpheme bo22 exclusively occurs preceding the whole constituent that it modifies. Unlike most Chinese dialects, it does not display a complementary distribution with the modal marker ɗit. The present study assumes that this unusual negative form in Hainan Min can be explained by diachronic accounts. The following section aims to account for the facts presented above through diachronic claims. In addition, the diachronic accounts will be argued with dialectal association. 3  DIACHRONIC DISCUSSION IN DIFFERENT DIALECTS The [Neg X] form is a rare negative modal form in the Chinese dialects. There are mainly three dialects in which the diachronic data encodes this unique form: Southern Min, Hakka, and Cantonese.

3.1  Southern Min Hainan Min is a branch of the Southern Min group. The ancient data of Southern Min can be traced back to the Ming dynasty. Li Jing Ji (Legend of the Litchi Mirror), a Southern Min play script, is the only linguistic material which can offer the old Southern Min data in the Ming and Qing dynasties (ca. 16th – 19th c.). The [V tit4] form in Li Jing Ji can be used as a resultative complement or a dynamic modal marker depending on the constructions it joins. The data in Li Jing Ji shows that the Southern Min word tit4 得 has developed the modal sense of possibility as early as the Ming dynasty. Lien (2011b) explores four types of the tit4 constructions in Li Jing Ji: (i) the postverbal tit4 construction with the cognitive verbs, e.g. ki3-tit4 ‘remember,’ (ii) the postverbal tit4 constructions with other types of verbs,

Negation of dynamic modals  85 e.g. tsiah8 tit4 ‘can eat,’ (iii) the bue7tit4 V construction, and (iv) the construction featuring e7tsue3tit4/e7ing7tit4/ e7sai2tit4 and their negative counterparts. The four constructions basically express deontic and dynamic modality in Southern Min texts. Constructions (i) and (ii) are the main tit4 constructions in Southern Min texts. The bue7tit4 V construction simply expresses circumstantial modality. Constructions (iv) are rarely used in the diachronic texts, though they occur more frequently in modern Southern Min. Among the four types of tit4 得 constructions, the second type is our major concern, the dynamic modal pattern [V tit]. Like Hainan Min, the negative [V tit4] form also follows the [Neg X] format. The [Neg V tit4] is exemplified as (35)-(39). (35) 亦



共 你 鬥 得2 iah8 kang7 li2 tau3 tit4 also neg with 2sg match get ‘also cannot work together with you’ 袂 be7

(11.296 SZ)3

(Lien 2011b, (6)-(10))

(36) 酒 勿 斟， 我 袂 食 tsiu2 mai3 thin5 gua2 be7 tsiah8 wine neg pour 1sg neg eat ‘Don’t pour the wine. I cannot take it’.

得 tit4 get

(37) 十種， 袂 推 得 了 tsap8tsing2 be7 the1 tit4 liau2 mixed.seed neg push get finish ‘B astard! You cannot get off scot-free’.

(5.627 SZ)

(38) 頭上仔 愛 襪 飼 得 thau5tsiunn7a2 ai3 be7 tshi7 tit4 first-born-dim inevitably neg raise get ‘The first baby inevitably cannot survive’. (39) 阮 句 那 是 guan2 ku3 na7 si7 1pl.excl nevertheless if cop ‘If I am reluctant to do so . . .’

袂 be7 neg

(28.084 SZ)

(14.114 JJ)

捨 sia2 abandon

得 tit4 get

(27.082 GX)

Lien (2011b) points out that ‘no bare V tit4 has survived in modern Southern Min’. This paper considers that, on the one hand, it is true that the [V tit4] form is not commonly used in modern Southern Min. On the other hand, some expressions with the [V tit4] pattern in negative forms still remain in modern Southern Min. For example, be7tan2tit4 ‘cannot wait’ and be7khun3tit4 ‘cannot sleep’ appear in Li Jing Ji, (40), and they are still actively used in modern Southern Min.

86  Huichi Lee (40) [Neg V tit#] a 我 一 時 都 袂 gua2 tsit8 si5 long2 be7 1sg one time all neg ‘I cannot wait any longer’.

等 tan2 wait

得 tit4 get

(45.045 JJ)

b 一 冥 返 去 厝 祙 睏 得 tsit8 me5 tng2 khi3 tshu3 be7 khun3 tit4 one night return go house neg sleep get ‘(I) went back home one night and could not sleep’.

(9.044 JJ)

Moreover, some negative forms of the transitive [V tit4] expressions also exist both in Li Jing Ji and modern Southern Min, for example, be7tsue3tit4tsu2 ‘cannot make decisions about something’. (41) [Neg V tit O] 有 馬 袂 做 得 主 (26.404 JJ) u7 ma7 be7 tsue3 tit4 tsu2 have also neg do get master ‘Even if there is something, (I) cannot make the decision about it’. It is noteworthy that the intransitive [Neg V tit4] form occurs much more frequently than the transitive form. As for the dynamic modal usage, the example (41) is the only example with the transitive form in the text of the Ming dynasty.4 The intransitive and transitive [Neg V tit] forms still co-occur and compete with each other in the Qing dynasty. However, unlike the Ming dynasty, the transitive [Neg V tit4 O] becomes more common than the intransitive [Neg V tit4] in the Qing dynasty. (42) [Neg V tit4 #]



我 袂 困 得 gua2 be7 khun3 tit4 1sg neg trap get ‘I cannot be trapped’.

(3.020 KX)

(43) [Neg V tit4 O] a 人 都 袂 受 得 人 嘮嘈 lang5 long2 be7 siu7 tit4 lang5 lo5tso1 people all neg accept get people preach ‘People cannot stand other people’s preaching’.

(10.115 SZ)

Negation of dynamic modals  87 b 袂 be7

欺 得 我 khi1 tit4 gua2 neg deceive get 1sg ‘(You) cannot deceive me’.

過 kue3 pass

(11.694 SZ)

c 我 袂 顧 得 gua2 be7 koo3 tit4 1sg neg take.care get ‘I cannot take care of you’.

你 li2 2sg

(11.725 SZ)



d 大 樹 都 袂 受 得 金 刀 斧 (17.053 KX) tua7 tshiu7 long2 be7 siu7 tit4 kim1 to1 poo2 big tree all neg accept get gold knife ax ‘Even big trees cannot stand the metal axes’. In the latest version, the Guangxu version, [Neg tit4 S V] started to emerge. Only one example is found. (44) 踗腳攝手， 袂 得 阮 好 (22.102 GX) liam1kha1liap4tshiu2 be7 tit4 gun2 ho2 tiptoe.leg.careful.hand neg get 1pl.excl good ‘(Even if you are) very careful, you cannot achieve more than us’. (45) Variants of V-tit constructions in different dynasties in Southern Min [Neg V tit4] > [Neg V tit4]/[Neg V tit4 O] > [Neg V tit4]/[Neg V tit4 O]/[Neg tit4 S V] Ming > Ming-Qing > Qing The findings of the historical material in Southern Min show that in the Ming and Qing dynasties, Southern Min had the negative forms, [Neg V tit4] and [Neg V tit4 O]. These patterns remain and are still actively used in Hainan Min.

3.2 Hakka The [V tet] pattern of dynamic modals is acceptable in modern Hakka. The [V tet] construction can be construed to be either dynamic or deontic, as shown in (46). However, the [Neg V tet] is not acceptable in modern Hakka, either with a dynamic or deontic reading, as shown in (47). (46) 我 去 得 (Sixian Hakka) ngai5 hi3 tet45 1sg go get ‘I am able to go/I am permitted to go’.

88  Huichi Lee (47) *你 唔 去 得 (Sixian Hakka) ngi5 m5 hi3 tet4 2sg neg go get ‘*You are not able to go/*you may not go’. There are two studies investigating the historical materials of Hakka from the Basel Mission, Chappell and Lamarre 2005 and Zhuang and Huang 2008. The former’s study is based on several manuscripts held in the Basel Mission Library Archives, including dictionaries, grammar and lexicon books published around 1909. They are basically copies of the Kleines Deutsch-Hakka Wörterbuch (Concise German-Hakka Dictionary), Dictionary Hakka-English, Kleine HakkaGrammatik (Concise Hakka Grammar), and Kleines Deutsch-Hakka Wörterbuch für Anfänger (Concise German-Hakka Dictionary for Foreigner, in 1909). The latter study is based on a Hakka textbook published by the Basel Missionary Society in 1879, entitled, Khi3mung2 Tshen3hok56 啟蒙淺學 (First Book of Reading). Chappell and Lamarre’s (2005) book provides Hakka grammar and lexicon from the late Qing dynasty, while Zhuang and Huang’s (2008) paper focuses on the Hakka potential marker tet spoken in New Territories in Hong Kong. This paper discusses the historical data of Hakka from these two previous studies. Zhuang and Huang (2008) offer several types of [Neg V tet] examples. Based on the data in the First Book of Reading from 1879, Zhuang and Huang categorize the [Neg X] forms into nine subcategories: [Neg V tet], [Neg V tet O], [Neg V tet C], [Neg V tet dau], [Neg V tet C dau], [Neg V tet O C], [Neg V tet O dau], [Neg V tet O C dau], and [Neg V tet C O]. (48) a [Neg V tet]: 但係 因為 係 硬 佢 就 than3he4 yin1wui4 he4 ngang4 ki1 tshyu4 but-cop because cop hard 3sg then ‘But, because it is really hard, he cannot eat it’.

唔 m1 neg

食 šit5 eat

得 tet6 get

b [Neg V tet O]: 算法 係 難 過 讀書 寫字， 但係 son4fap6 he4 nan1 ko4 thuk5šu1 sya1sii4 than3he4 math.method cop hard over study.book write.word but-cop 十分 有用， 唔 少 得 佢 šip5fun1 yu1yung4 m1 šau1 tet6 ki2 very useful neg lack get 3sg ‘Arithmetic is harder than studying and writing, but it is very useful and one cannot lack it’. c



[Neg V tet C]: 獅 雖然 有 柬 大 力 都 唔 sui1yen2 yu1 kan4 thai4 lit5 tu1 m1 sii1 lion although have that big strength all neg ‘Although lions have great strength, they still cannot tear (it)’.

扯 tšha3 tear

得 tet6 get

爛 lan4 broken

Negation of dynamic modals  89 d [Neg V tet dau]: 唔 m1 e



救 得 kyu4 tet6 neg save get ‘cannot successfully save it’

倒 tau3 arrive

[Neg V tet C dau]: 但係 咁 箇 鵰 唔 養 than3he4 kan2 kai4 tyau1 m1 yong1 but-cop such clf bird neg breed ‘But, such a kind of bird cannot be raised to be very old’.

f [Neg V tet O C]: 人 唔 多 捉 ngin2 m1 to1 tsuk5 people neg all catch ‘People cannot catch him firmly’. g [Neg V tet O dau]: 唔 m1 neg



‘cannot see its tail’

看 khon4 see

得 tet6 get

得 tet6 get

家 kya1 3gen

得 tet6 get

佢 ki2 3sg

尾 mui1 tail

幾 ki3 how

老 lau3 old

倒 tau3 arrive

唔 m1

當 tong4 serve

穩 wun3 firm

倒 tau3 arrive

h [Neg V tet O C dau]:



所羅門 so3lo2mun2 Solomon

王 wong2 king

箇 kai4

得 tet6 get

百合花 pak6hap5fa1 lily.flower

柬 kan4

clf

dem

榮華 yin2fa2 glory

都 tu1 all

排長 phai2tšhong2 ostentation

倒 tau3 arrive

neg

‘Even Solomon, in all his glory, could not be clothed like the lily flowers’.

i [Neg V tet C O]: 唔 鑽 得 出 鎖鍊， 又 唔 得 脫 m1 tson1 tet6 tšhut6 so3len4 yu4 m1 tet6 thot6 neg drill get out chain and neg get escape ‘(Someone) cannot get out of the chain, and cannot escape from it’. In addition to demonstrating the tet-data, Zhuang and Huang (2008) also count the tokens of the occurrence of tet of which they found the total number of tet in the book to be 341. The frequency of the nine types of [Neg X] forms diverges and is shown in Table 4.1. Compared to the total frequency of tet, 341, the [Neg X] forms do not occur very often. The previous two types, [Neg V tet] and [Neg V tet O], seem to appear

90  Huichi Lee Table 4.1  Frequency of the nine types of [Neg X] forms Type

Vtet

Vtet O Vtet C Vtetdau Vtet C Vtet Vtet VtetOCdau VtetCO dau OC O dau

Token

18

11

2

1

1

6

8

1

2

more frequently than others. However, the present study finds that there are restrictions on the distribution of the two types of negative forms. As for the [Neg V tet] form, the main verb is always monosyllabic, such as (49); as for the [Neg V tet O] form, it is often found to occur in a bi-clausal sentence, for example, in a conditional clause, such as (50). (49) a



上帝 有 箇 本事， 佢 起 得 一 條 城牆 song4ti4 yu1 kai4 pun3sii4 ki2 hi3 tet6 yit6 thyau2 šang2syong2 God have clf ability 3sg rise get one clf wall 圍竟 仇敵， 使 佢 唔 落 得。 wui2kin3 šu2thit5 sii3 ki2 m1 lok5 tet6 besiege enemy make 3sg neg fall get ‘God has the ability. He can build up a wall to besiege enemies and makes them not able to get down’.

b 有



日 佢 落 一 條 村， 忽然間 揭病， nyit6 ki2 lok5 yit6 thyau2 tshun1 fut6yen2kan1 pot6phyang4 day 3sg arrive one clf village suddenly get.illness 行 得。 hang2 tet6 neg walk get ‘One day, he came to a village and suddenly suffered an illness. He was not able to walk’. yu1 have 無 m1



(50) 竟係



雅爸 kin3he4 nga1pa1 if-cop father 工夫。 kung1fu1 labor ‘If my father gets labor’.

老 lau3 old

裡， li1 perf

血氣 hyet6hi4 vigor

衰弱， soi1nyok5 weak

唔 m1 neg

做 tso4 do

得 tet6 get

惹 nya1 2sg.gen

old, his vigor will weaken and will not be able to do your

Data in Khimung Tshenhok (First Book of Reading) show that there are several types of [Neg X] forms, but they appear either under certain restricted syntactic circumstances or very infrequently. Chappell and Lamarre 2005 show that Hakka tet can be used as a dynamic modal to express potential and possibility, such as (51).

Negation of dynamic modals  91 (51) a 佢 去 得 ki2 hi4 tet6 3sg go get ‘He can go’. b 我 聽 得 到 ngai2 thang4 tet6 tau4 1sg listen get arrive ‘I can understand it’. The data also include negative forms of tet, especially the [Neg V tet] form that we are concerned with, as shown in (52). (52) a 佢 唔 脫 得 身 ki2 m1 thot6 tet6 šin1 3sg neg escape get body ‘He can’t arrange for time off (from his business)’. b 你 唔 得 得 柬 大 ngi2 m1 tet6 tet6 kan4 thai4 2sg neg obtain get such big ‘You can’t obtain such good fortune’.

福 fuk6 fortune

倒 tau3 arrive

In addition to the [Neg V tet] form, the [V Neg C] form is also described, as in (53). (53) a 放 唔 落 fong4 m1 lok5 put neg fall ‘cannot put it in’ b 講 唔 出 kong3 m1 tšhut6 speak neg out ‘cannot speak it out’ c 我 尋 唔 ngai2 tshim2 m1 1sg search neg ‘I cannot find it’.

倒 tau3 arrive

Chappell and Lamarre (2005:124) point out that the negative marker can either precede or follow the main verb. However, the negative marker is generally placed

92  Huichi Lee between the main verb and the complement. In other words, the [V Neg C] form occurs more commonly and naturally than the [Neg V tet] form. Since modern Hakka does not have the [Neg V tet] form, the competition of the two negative forms results in the fact that the [V Neg C] form eventually wins the [Neg V tet] form.

3.3 Cantonese Modern Cantonese allows four [Neg X] forms: [Neg V dak O], [Neg V dak C], [Neg V dak O C], and [Neg V dak C O] (cf. Yue 1993, 1995, 2001a, 2001b; Kwok 2005; Wu 2005). They are exemplified as in (54). (54) a [Neg V dak O] 唔 m4



食 得 sik6 dak1 neg eat get ‘cannot eat’

飯 faan37 rice

 b [Neg V dak C] 唔 m4



食 得 sik6 dak1 neg eat get ‘cannot eat’

落 lok6 fall

  c [Neg V dak O C] 唔 m4



打 得 佢 daa2 dak1 keui5 neg hit get 3sg ‘cannot hit him to death’

死 sei2 die

 d [Neg V dak C O] 唔 m4



打 得 死 佢 daa2 dak1 sei2 keui5 neg hit get die 3sg ‘cannot hit him to death’

This paper employs historical data in Cantonese from Yue 2001b in which she clearly describes the data in the 麥仕治廣州俗話書經解義 (Book of Documents Translated into Colloquial Cantonese by Mo Shizhi). This document was published around 1893 and written in the colloquial language to explicate the text of 書經 (the Book of Documents) to the people who could not literally comprehend the classical texts. Yue (2001b) observes that Cantonese in the 19th century contained four [Neg X] forms which exactly correspond to Modern Cantonese: [Neg V dak O], [Neg V dak C], [Neg V dak O C], and [Neg V dak C O].

Negation of dynamic modals  93 (55) a [Neg V dak O] 唔 m4

識 得 呢 sik1 dak1 ni1 neg know get dem ‘cannot know this thing’

件 gin6 clf

事 si6 affair

 b [Neg V dak C]



就 唔 辦 得 成功 zau6 m4 baan6 dak1 sing4gung1 then neg deal get success ‘then cannot do it successfully’

  c [Neg V dak O C] 你地 唔 nei5dei5 m4

使 憂心 話 唔 信 得 我 過 sai2 jau1sam1 waa6 m4 seon3 dak1 ngo5 gwo3 pl neg make worry say neg trust get 1sg pass ‘You don’t have to worriedly say that you cannot trust me’.

 d [Neg V dak C O] 唔 m4

依照 得 正 節令 時候 ji1ziu3 dak1 zeng3 zit3ling6 si4hau6 neg follow get exact seasonal time ‘cannot exactly follow the seasonal time’

From the modern and historical data, it is believed that the [Neg X] form has been actively used in Cantonese for centuries. Yue (2001b) speculates that the [Neg V dak O C] form appears earlier than the [Neg V dak C O] form. Kwok (2005) further argues that the emergence of the [Neg V dak C O] form is derived from the effect of analogy. The earlier positive-negative pair [V dak O C – Neg V dak O C] affects the later positive form [V dak C O] to create its negative counterpart, [Neg V dak C O]. Cantonese thus contain these two related [Neg X] forms. According to Kwok (2005), the semantics of the two forms are very similar, except for the literal/colloquial divergence. Wu (2005) points out that the [V dak O C] form is an inherent form in Cantonese, while the [V dak C O] form is borrowed from northern Chinese dialects. The [Neg V dak O C] form occurs in the Tang dynasty and the [Neg V dak C O] in the Ming dynasty. They both are still widely used in modern Cantonese. However, Li et al. (1995), Yue (2001b), and Kwok (2005) also observe that due to the competition between the two analogous forms for decades, the earlier form gradually converged with the borrowed form. The [V dak O C] form in modern Cantonese always occurs in some restricted syntactic distributions. The [V dak O C] form only occurs either when the verbal complement is directional, like lai4 來 ‘come’, heui3 去 ‘go’, hei2起 ‘rise’, lok6落 ‘fall’, or when the object is pronominal. These restrictions are not obeyed in early Cantonese.

94  Huichi Lee 4  HISTORICAL PERSPECTIVE OF THE [NEG X] FORM In addition to the different historical dialectal material, ancient Chinese data facilitate understanding the development and derivation of the V-ɗit constructions and their negative counterparts in Hainan Min. According to Wang (1980), on the one hand, the potential [V de] form (suffixal de) emerged in the Han dynasty and was full-fledged in the Tang dynasty. Its negative counterpart, the [V bu de] form, emerged in the Late Han dynasty. On the other hand, the verb compound [VC] form also emerged in the Han dynasty and was well developed by the Tang dynasty. Its negative counterpart, the [V bu C] form, emerged in the Tang dynasty. The potential modal pattern of [VC] form, the [V de C] form, emerged in the late Tang dynasty. Furthermore, Wu (2003) investigates the potential de constructions in Southern Chinese dialects. He speculates that the [V de O C] form emerged in the late Tang dynasty and was shared by most Southern Chinese dialects. As for the special negative form, [Neg V de C], Yue (1984) considers it to have occurred in the Song and Yuan dynasties. Based on the findings of these previous studies, the developments of the dynamic de constructions are shown as follows: (56) The development of Mandarin Chinese [Vde] and [VC] patterns Dynasty:

Han [V de] [VC]

Late Han [V bu de]

 . . .

Tang

Song

Yuan

[V bu C] [V de C] [V de OC] [bu V de C]

Notice that the negative form [V bu C] appears earlier than the positive form [V de C]. If the negative form may not be derived from its corresponding positive form, it is possible to assume that positive and negative forms can have different developments. This assumption can explain the fact that there are four types of positive de constructions in Hainan Min, but only one pattern for their corresponding negative forms. In Hainan Min, there are four positive ɗit constructions: [V ɗit], [V ɗit O], [Vɗit C], and [Vɗit O C]. This paper speculates that these four constructions in Hainan Min, like the de constructions in other Chinese dialects, developed themselves no later than the Tang-Song dynasties. However, unlike the other dialects, the negative forms of ɗit-constructions underwent their own distinctive development. The uniform of the negative ɗit-constructions in Hainan Min, [Neg X], is assumed to be a complex of a remnant of early Chinese and an effect of analogy. The previous section examines the diachronic material from three dialects, Southern Min, Hakka, and Cantonese. These dialects have preserved sufficient data and can be traced back to at least the Qing dynasty. Unlike these dialects, historical documents in Hainan Min are rarely discovered and other relative dialects thus need to be searched in order for the language development of Hainan Min

Negation of dynamic modals  95 to be examined. (57) exhibits the result of the historical de-constructions in negative forms in the three dialects discussed in the previous section. In order to make comparisons with the historical data in the former three dialects, modern Hainan Min is also listed in (57). Notice that Hainan Min fully corresponds to Hakka and Cantonese except for the last slot, [Neg V ɗit C O]. (57) Variants of [Neg X] forms in dialects Southern Min

Hakka

Cantonese

Hainan Min

Neg V tit

Neg V tet

Neg V dak

Neg V ɗit

Neg V tit O

Neg V tet O

Neg V dak O

Neg V ɗit O

Neg V tet C

Neg V dak C

Neg V ɗit C

Neg V tet O C

Neg V dak O C

Neg V ɗit O C

Neg V tet C O

Neg V dak C O

Wu (2003) points out that different de-constructions in Southern Chinese dialects may reflect different linguistic stratifications. Wu (2005) further argues that among the variants of de-constructions, the [V de O C] form is shared in most Southern Chinese dialects. This paper agrees with Wu’s theory by accounting for the absent slot of the [Neg V ɗit C O] in Hainan Min. According to Wu (2005), [V de O C] is an inherent form in Southern dialects, while the [V de C O] form is borrowed from Northern dialects. Following this hypothesis, Hainan Min is assumed to retain its intrinsic form and not borrow the [V de C O] form from northern dialects. The negative counterpart of the [V de C O] form is correspondingly not borrowed or has not developed. It is noteworthy that the form [V de C] in Chinese generally matches to the negative form [V Neg C] rather than [Neg V de C]. In some dialects, the two variants of negative forms co-occur in the diachronic materials, for example, Southern Min, Hakka, and Cantonese. Hainan Min is different from any of these dialects since it only has access to the [Neg X] form, instead of [V Neg C]. This idiosyncratic property is attributed to the effect of analogy. Following Yue (1984), the [Neg V de O C] form emerges during the time between the Song and Yuan dynasties. This form is well preserved in Hainan Min and Cantonese. It diffuses its [Neg X] pattern to other ɗit constructions. Hainan Min eventually developed itself to be a language with a full-fledged [Neg X] form. Since the dialects spoken in the areas neighboring Hainan Island, e.g. Cantonese and Hakka, did not undergo the same syntactic analogy, the present study considers this syntactic diffusion occurring language-internally only in Hainan Min. To sum up, the special negative ɗit-constructions in Hainan Min is attributed to historical preservation and the effect of analogy. Along with the [Neg X] form, the [V de] form provides one more piece of evidence to support the assumption that historical association can be shown by cross-dialectal commonness. Like the [Neg X] form, the [V de] form is shared

96  Huichi Lee by Hainan Min ([V ɗit]), Cantonese ([V dak]), and Hakka ([V tet]). From the diachronic data, the [V de] form is found in the texts of Middle Chinese and Early Chinese. The [V de] form can be either intransitive or transitive, as shown in (58)-(61). (58) 一



擊

得

《Lùnhéng, Shùngǔ 論衡/順鼓》 (Han dynasty, Middle Chinese)

yì rén jí dé one person beat get ‘One person can play (the drum)’.

(59) 若



人

有

誦

《Shēngjīng, Dàzhèngzàng 生經/大正藏卷三》 (ca. Weijin Nanbei dynasty, Middle Chinese)

得

ruò yǒu sòng dé if have recite get ‘If there is anybody who can recite (the religious text . . .)’

(60) 是

何

人

誦

得?

shì

hé rén sòng dé which person recite get ‘Who is the person who can recite’?

《5.030.09 Zǔtángjí 祖堂集》 (Wudaishiguo, Early Chinese)

cop



(61) 必



救

得

君 母 之 患 《Sōushénjì, Bàihǎiběnjuàn 搜神記/稗海本卷七》 (Nanbei dynasty, Middle Chinese) bì jiù dé jūn mǔ zhī huàn surely save get king mother gen trouble ‘It can definitely relieve king’s mother’s anxiety’.

Based on the historical material, the [V de] forms in Hainan Min, Hakka, and Cantonese are assumed to be historical linguistic remnants. Simpson (2001) also considers that the data in Middle Chinese is the original source of the dynamic [V dak] form in Modern Cantonese. Ancient Chinese not only attributes the [V de] pattern to Chinese dialects, but also the [V can] form occurring in several East and Southeast Asian languages: (62) Thai (Simpson 2001, (2)) khaw khian 3sg write ‘He can write’.

dai can

Negation of dynamic modals  97 (63) Cambodian (Simpson 2001, (4)) goa’at root-ut 3sg run ‘He can run’.

baan can

(64) Vietnamese (Simpson 2001, (6)) anh-ta den 3sg come ‘He can come’.

duoc can

The above examples in Thai, Cambodian, and Vietnamese are all with the configuration of [V can] and express dynamic modality. While the East and Southeast Asian languages are not of concern in the present study, this shared feature of syntax and semantics in Hainan Min, Cantonese, and Southeast languages may be ascribed to the historical factor. 5  CONCLUDING REMARKS The aim of this study is to explore the dynamic modality in Hainan Min involving the word ɗit ‘acquire, get’. Moreover, this paper is especially concerned with the negative forms of the ‘ability/ potential’ dimension of dynamic modality. A historical approach is taken and cross-dialectal accounts are also considered. Previous research has rarely studied the modal issues in Hainan Min. At the beginning of this paper, the distribution of the modal word ɗit ‘acquire, get’ is clearly introduced in the paper. Like in many Chinese dialects, ɗit can serve as a transitive verb and also as a modal marker. It can occur in four dynamic modal constructions: [Vɗit], [Vɗit O], [Vɗit C], and [Vɗit O C]. Their negative counterparts correspond to [Neg Vɗit], [Neg Vɗit O], [Neg Vɗit C], and [Neg Vɗit O C]. These negative forms can be exactly formulated to be the [Neg X] pattern. The variable X includes the four types of ɗit-constructions. This negative form is very special in Chinese dialects. Even in Southern Min dialects, it is not often found. To investigate the structure, the present study seeks both diachronic and cross-dialectal explanations. Hakka and Cantonese are compared with Hainan Min because they respectively display similarity in the modal structures involving with the word de. This paper makes efforts to examine the negative modal forms in the three dialects with diachronic material: Southern Min, Hakka, and Cantonese. Five basic patterns of negative forms are found in the historical texts and listed in Table 4.2. Diachronic data in Southern Min are gathered from Li Jing Ji; Hakka data are from several manuscripts contained in the Basel Mission Library Archives and Cantonese data are from the Book of Documents Translated into Colloquial Cantonese by Mo Shizhi. These data are from the 19th century or earlier. Diachronic data in Hainan Min have not been found yet and modern Hainan Min data are thus

98  Huichi Lee Table 4.2  Five basic patterns of negative forms

Neg V de Neg V de O Neg V de C Neg V de O C Neg V de C O

Southern Min

Hakka

Cantonese

Hainan Min

 

    

    

   

compared with the historical data in the three dialects. It is clear that Hainan Min is historically more similar to Hakka and Cantonese than it is to Southern Min. However, with regards to the modern data, Hainan Min is closer to Cantonese than to the other dialects. Only some fixed terms in modern Southern Min retain the [Neg X] construction; most [Neg X] forms in modern Hakka have faded away; only modern Cantonese still actively employs the [Neg X] forms. In addition to the negative modal constructions, Hainan Min, Hakka, and Cantonese have other commonalities. For example, the three dialects all have the [V de] pattern expressing deontic or dynamic modality. Unlike these dialects, most Southern Min dialects do not use the modal [V de] pattern. As for the modal structures, Hainan Min syntactically behaves more similar to Cantonese and Hakka, than to other Southern Min dialects. This paper attributes the commonalities between these Southern Chinese dialects to historical factors and dialectal contact. NOTES 1 All the Hainan data presented in this paper are collected from the author’s field work on the Hainan Island. 2 Some Chinese characters in the Li Jing Ji are not real etymons, but phonetic loan characters instead. 3 The capital letters JJ indicates Jiajing, SZ Shunzhi, WL Wanli, and GX Guangxu editions of the Legend of the Litchi Mirror (Wu 2001a, 2001b, 2001c, 2001d). The numeral preceding the dot denotes the ordinal number of acts in the playscripts, and the numeral following the dot refers to the serialized item. 4 Unlike the [V tit4] form, the [tit4 V] form has more transitive forms. For example, 樂毅 未得逢知音 (18.008 JJ) gak8ge7 be7 tit4 hong5 ti1im1 ‘Gak-ge cannot meet a confidant’, 幸然得見娘仔 (19.167 JJ) hing7jian5 tit4 kinn3/kian3 niu5a2 ‘Fortunately, (I) can meet the hostess’, and 祙得瞞你 (25.200 JJ) be7 tit4 mua5 li2 ‘(I) cannot hide anything from you’. 5 This study follows the Manual of Taiwan Hakka Tongyong Romanization System (台灣 客語通用拼音使用手冊) proclaimed by the Ministry of Education in 2003. In the Taiwanese system, tone number 1 corresponds to 陰平 Yīnpíng; 2 to 上聲 Shǎngshēng; 3 to 陰去 Yīnqù; 4 to 陰入 Yīnrù; 5 to 陽平 Yángpíng and 8 to 陽入 Yángrù. However, only modern Hakka follows this Tongyong system. The tonal denotation of historical Hakka will follow its original tonal system. 6 The tone symbols in historical materials of Hakka from the Basel Mission, including Khi3mung2 Tshen3hok5 啟蒙淺學, are very different from Taiwan Hakka system. In the Basel system, tone 1 corresponds to 陰平 Yīnpíng; 2 to 陽平 Yángpíng; 3 to 上聲 Shǎngshēng; 4 to 去聲 Qùshēng; 5 to 上入 Shǎngrù; 6 to 下入 Xiàrù.

Negation of dynamic modals  99 7 The Cantonese Romanization system adopted in this paper follows the Romanization Scheme (Jyutping) proposed by the Linguistic Society of Hong Kong in 1993. Tone number 1 corresponds to 陰平 Yīnpíng; 2 to 陰上 Yīnshǎng; 3 to 陰去 Yīnqù; 4 to 陽平 Yángpíng; 5 to 陽上 Yángshǎng; 6 to 陽去 Yángqù; 1 to 高陰入 Gāoyīnrù; 3 to 低陰入 Dīyīnrù, and 6 to 陽入 Yángrù.

REFERENCES Bisang, Walter. 1996. Areal typology and grammaticalization: Processes of grammaticalization based on nouns and verbs in East and Mainland South East Asian languages. Studies in Language 20:519–597. Chao, Yuen Ren趙元任. 1968. A grammar of spoken Chinese. Berkeley: University of California Press. Chappell, Hilary 曹茜蕾, and Christine Lamarre 柯理思. 2005. A grammar and lexicon of Hakka: Historical materials from the Basel Mission Library. Paris: École des Hautes Études en Sciences Sociales. Chen, Hong-Mai 陳鴻邁. 1996. Hǎikǒu fāngyán cídiǎn 海口方言詞典 [Haikou dialect dictionary]. Nanjing: Jiangsu Education Press. Clark, Marybeth. 1989. Hmong and areal South-East Asian. In Papers in South-east Asian linguistics 11: South-East Asian syntax, ed. by David Bradley, 175–230. Canberra: Australian National University. Enfield, Nick J. 2001. On genetic and areal linguistics in Mainland Southeast Asia: Parallel polyfunctionality of ‘acquire’. In Areal diffusion and genetic inheritance, ed. by Alexandra Y. Aikhenvald and R.M.W. Dixon, 255–290. Oxford: Oxford University Press. Hashimoto, Mantaro 橋本萬太郎. 1959. Kainango oninron 海南語音韻論 [Hainan phonetic and phonology – on Wenchang dialect]. Chugokugogaku 中国語学 91. Reprinted in Hashimoto Mantaro publication collection 2 橋本萬太郞著作集, ed. by Hashimoto Mantaro Chosakusyu Kankokai 橋本万太郎著作集刊行会, 452–458. Tokyo: Uchiyama Books. Hashimoto, Mantaro 橋本萬太郎. 1960. The Bonshio dialect of Hainan. Gengo Kenkyu 言語研究, 38. Reprinted in Hashimoto Mantaro publication collection 2橋本萬太郞 著作集, ed. by Hashimoto Mantaro Chosakusyu Kankokai 橋本万太郎著作集刊行会. Tokyo: Uchiyama Books. Hashimoto, Mantaro 橋本萬太郎. 1961a. Kainango no seicho taikei 海南語の声調体 系 [Tone system in Hainan Min]. Tokyo Shina Gakuhou 東京支那学報7. Reprinted in Hashimoto Mantaro publication collection 2橋本萬太郞著作集, ed. by Hashimoto Mantaro Chosakusyu Kankokai 橋本万太郎著作集刊行会, 462–474. Tokyo: Uchiyama Books. Hashimoto, Mantaro 橋本萬太郎. 1961b. Bilabial and alveolar implosives in a South Chinese dialect. Onsei no Kenkyu 音声の研究9. Reprinted in Hashimoto Mantaro publication collection 2 橋本萬太郞著作集, ed. by Hashimoto Mantaro Chosakusyu Kankokai 橋本万太郎著作集刊行会, 475–482. Tokyo: Uchiyama Books. Hashimoto, Mantaro 橋本萬太郎. 1976. Kainango Bonsho hougen 海南語文昌方言 [On the Wenchang dialect in Hainan Min]. Azia Afurika Gengo Bunka Kenkyu アジア・ア フリカ言語文化研究 11. Reprinted in Hashimoto Mantaro publication collection 2 橋 本萬太郞著作集, ed. by Hashimoto Mantaro Chosakusyu Kankokai 橋本万太郎著作 集刊行会. Tokyo: Uchiyama Books. Kwok, Bit-Chee 郭必之. 2005. Yuèfāngyán néngxìng shùbǔ jiégòu fāzhǎn de jǐdiǎn guānchá 粵方言能性述補結構發展的幾點觀察 [Some observations on the development of the

100  Huichi Lee potential complement construction in Cantonese]. In Proceedings of the 9th international conference on Yue dialects, ed. by Keng-pan Tang 鄧景濱 and Choi-lan Tong 湯 翠蘭, 271–277. Macao: Macau Chinese Language Society. Lamarre, Christine 柯里思. 2001a. Cóng Pǔtōnghuà lǐ gēn ‘de’ yǒukuān de jǐge géshì qù tàntǎo fāngyán lèixíngxué 從普通話裡跟「得」有關的幾個格式去探討方言類型學 [A dialectal typological study of several Chinese DE constructions]. Yǔyán Yánjiū 語 言研究 43:7–18. Lamarre, Christine. 2001b. Verb complement constructions in Chinese dialects: Types and markers. In Sinitic grammar: Synchronic and diachronic perspectives, ed. by Hilary Chappell, 85–120. Oxford: Oxford University Press. Li, Xinkui 李新魁, Jiajiao Huang 黃家教, Qisheng Shi 施其生, Yun Mai 麥耘, and Dingfang Chen 陳定方. 1995. Guǎngzhōu fāngyán yánjiū 廣州方言研究 [A study of the Guangzhou dialect]. Guangzhou: Guangdong Renmin Chubanshe. Liang, You-Gang 梁猷剛. 1986. Hǎinándǎo Wénchāng fāngyán yīnxì 海南島文昌方言音 系 [Sound system of Hainan Min Wenchang dialect]. Fāngyán 方言2:123–132. Lien, Chinfa 連金發. 1997. Aspects of the evolution of tit (得) in Taiwan Southern Min. In Studies on the history of Chinese syntax, ed. by Chaofen Sun, 167–190. Journal of Chinese Linguistics Monograph Series 10. Berkeley: University of California Press. Lien, Chinfa 連金發. 2011a. Mǐnnányǔ qíngtàicí de fǒudìng lèixíng tànsuǒ 閩南語情態詞 的否定類型探索 [Patterns of negative modals in Southern Min]. In Dì Shíèr Jiè Mǐnyǔ Guójì Xuéshù Yántǎohuì Lùnwénjí 第十二屆閩語國際學術研討會論文集 [Proceedings of the 12th conference on Min languages], 233–250. Taipei: Academia, Sinica. Lien, Chinfa. 2011b. Interface of modality and the tit 得 construction in Southern Min: A case study of their development from earlier Southern Min in the Ming and Qing to modern Taiwanese Southern Min. Languages and Linguistics 12:723–752. Liu, Cheng-Hui 劉承慧. 2001. Dòngbǔ ‘DE’ zì jiégòu de lìshǐ fāzhǎn 動補「得」字結 構的歷史發展 [Historical evolution of the verb complement DE]. Táidà Wénshǐzhé Xuébào 台大文史哲學報 54:95–134. Liu, Cheng-Hui 劉承慧. 2002. Hànyǔ dòngbǔ jiégòu lìshǐ fāzhǎn 漢語動補結構歷史發展 [On the historical development of the V-de-C construction]. Taipei: Hanlu Tushu. Liu, Xin-Zhong 劉新中. 2006. Hǎinán Mǐnyǔ de yǔyīn yánjiū 海南閩語的語音研究 [The study of phonology in Hainan Min]. Beijing: Zhongguo Shehui Kexue Press. Liu, Xin-Zhong 劉新中. 2010. Guǎngdōng, Hǎinán Mǐnyǔ rùokān wèntí de bǐjiào yánjiū 廣東、海南閩語若干問題的比較研究 [The comparative study on several issues in Kuangdong Min and Hainan Min]. Kuangzhou: Jinan University Press. Lü, Shuxiang 呂叔湘. 1944. Yǔ dòngcí hòu dé yǔ bù yǒuguān zhī cíxù wùntí 與動詞後「 得」與「不」有關之詞序問題 [Issues of word order relating to the postverbal de and bu]. Zhōngguó Wénhuà Yánjiū Huìkān 中國文化研究匯刊 4:47–57. Matisoff, James. 1991. Areal and universal dimensions of grammatization in Lahu. In Approaches to grammaticalization, ed. by Elizabeth Traugott and Bernd Heine. Vol. II, 383–453. Amsterdam: John Benjamins. Palmer, Frank R. 2001. Mood and modality. Cambridge: Cambridge University Press. Simpson, Andrew. 2001. Focus, presupposition and light predicate raising in East and Southeast Asia. Journal of East Asian Linguistics 10:89–128. Tsai, Wei-Tien Dylan 蔡維天. 2010. Tán Hànyǔ mótàicí qí fēnbù yǔ quánshì de duìyìng guānxī 談漢語模態詞其分佈與詮釋的對應關係 [On the syntax-semantics correspondences of Chinese modals]. Zhōngguó Yǔwén 中國語文 2010:208–221. Wang, Li 王力. 1980. Hànyǔ shǐ gǎo 漢語史稿 [Outline of the history of Chinese]. Peking: Zhonghua Shuju.

Negation of dynamic modals  101 Wei, Pei-Chuan 魏培泉. 2004. Jìndài Hànyǔ néngxìng dòngbǔ jiégòu zhōng bīnyǔde wèizhì 近代漢語能性動補結構中賓語的位置 [Object position in verb-complement potential constructions in early Mandarin]. Language and Linguistics 5:663–704. Woon, Wee-Lee 云惟利. 2004. Yì zhǒng fāngyán zài liǎngdì sāndài jiān de biànyì 一種 方言在兩地三代間的變異 [A dialect variation in the three generations and two locations]. Fujian: Xiamen University Press. Wu, Fuxiang 吳福祥. 2003. Nánfāng fāngyán néngxìng shùbǔ jiégò V dé/bù C dài bīngyǔ de yǔxù lèixíng 南方方言能性述補結構“V得/不C”帶賓語的語序類型 [Types of word order of the potential verb-complement construction V+de+C/V+bu+C with object in the Southern Chinese dialect]. Fāngyán 方言 3:243–254. Wu, Fuxiang 吳福祥. 2005. Yuèyǔ néngxìng shùbǔ jiégò Neg-V de OC/CO de láiyuán 粵 語能性述補結構“Neg-V得OC/CO”的來源 [Origin of the potential verbal-complement constructions “Neg V-得OC(/CO)” in Yue dialect]. Fāngyán 方言 4:306–318. Wu, Fuxiang 吳福祥. 2009. Cóng dé yì dòngcí dào bǔyǔ biāojì Dōngnányǎ yǔyán de yìzhǒng yǔfǎhuà qūyù 從“得”義動詞到補語標記─東南亞語言的一種語法化區域 [From the verb meaning ‘get, acquire’ to the resultative construction marker: A  kind of grammaticalization area in the Southeast Asian languages]. Studies of the Chinese Language 中國語文 2009:195–211. Wu, Shouli 吳守禮. 2001a. Míng Jiājìng kān Lì Jìng Jì xìwén jiàolǐ 明嘉靖刊荔鏡記戲文 校理 [Annotated texts of the romance of Li Jing Ji of Ming Jiajing edition (JJ)]. Taipei: Congyi Workshop. Wu, Shouli 吳守禮. 2001b. Míng Wànlì kān Lì Zhī Jì xìwén jiàolǐ 明萬曆刊荔枝記戲文 校理 [Annotated texts of the romance of Li Zhi Ji of Ming Wanli edition (WL)]. Taipei: Congyi Workshop. Wu, Shouli 吳守禮. 2001c. Qīng Shùnzhì kān Lì Zhī Jì xìwén jiàolǐ 清順治刊荔枝記戲文 校理 [Annotated texts of the romance of Li Zhi Ji of Qing Shunzhi edition (SZ)]. Taipei: Congyi Workshop. Wu, Shouli 吳守禮. 2001d. Qīng Guāngxù kān Lì Zhī Jì xìwén jiàolǐ 清光緒刊荔枝記 戲文校理 [Annotated texts of the romance of Li Zhi Ji of Qing Shunzhi edition (SZ)]. Taipei: Congyi Workshop. Yue, Anne O. 1993. Comparative Chinese dialectal grammar: Handbook for investigators. Paris: Centre de Recherches Linguistiques sur l’Asie Orientale, École des Hautes Études en Sciences Socials. Yue, Anne O. 余靄芹. 1995. Yuèyǔ yánjiū de dāngqián kètí 粵語研究的當前課題 [The current subjects of the Yue language research]. Journal of Chinese Linguistics 23:1–41. Yue, Anne O. 2001a. The verb complement construction in historical perspective with special reference to Cantonese. In Sinitic grammar: Synchronic and diachronic perspectives, ed. by Hilary Chappell, 232–265. Oxford: Oxford University Press. Yue, Anne O. 余靄芹. 2001b. Yuèyǔ de fǒudìngshì – yí ge lìshǐ de kǎochá 粵語的否定 式 – 一個歷史的考察 [Negation forms in Yue: A historical study]. Paper presented at the Conference of Projects on Dialectal Comparison of Southeastern China. Shanghai: Fudan University. Yue, Jun-Fa 岳俊發. 1984. Dé zìjù de chǎnshēng hàn yǎnbiàn 得字句的產生和演變 [The derivation and development of DE sentences]. Yǔyán Yánjiū 語言研究 2:10–30. Zhuang, Chusheng, and Tingting Huang. 2008. Potential constructions in Hakka more than 100 years ago in the New Territories. Bulletin of Chinese Linguistics 2.2:149–173.

5 Word change and language change A case of 共 as a coordinating conjunction from Archaic Chinese gòng 共 to ka7 共 in Taiwanese Southern Min1 Jang-Ling Lin 林征玲 1  INTRODUCTION The paper discussed the historical change of 共 as a coordinating conjunction AND (henceforth, AND-CONJ), spanning 3500 years. The change runs through Mandarin Chinese (Northern Sinitic) to Southern Min and Taiwanese Southern Min (Southern Sinitic languages). The Chinese coordinating conjunction AND is a grammatical category (Liu and Peyraube 1994). They come from verbs via grammaticalization (Kuryłowicz 1965; Heine, Claudi, and Hünnemeyer 1991; Hopper and Traugott 2003; Campbell and Janda 2001; Heine and Kuteva 2002; Brinton and Traugott 2005). It has been attested that the verb gòng 共 was first grammaticalized into an adverb during the 7th – 10th c. (the Tang dynasty) and the adverb gòng 共 (henceforth, 共-ADV) was grammaticalized into a preposition WITH (WITH-PREP) during the 10th c. and the 13th c. (the Song dynasty). The WITH-PREP gòng 共 (henceforth, 共-WITH-PREP) was grammaticalized into a coordinating conjunction gòng 共-AND (henceforth, 共-ANDCONJ), (Liu and Peyraube 1994). The grammaticalization path of gòng 共 in Liu and Peyraube 1994 can be formalized as: V > ADV > PREP > CONJ.2 I argue that polygrammaticalization was involved in the change (Craig 1991; Traugott and Heine 1991). What happened is that both the preposition WITHgòng 共 and the conjunction AND-gòng 共 come directly from the ADV in the 5th century, as attested in Shì Shuō Xīn Yǔ 世說新語 ‘A New Account of Tales of the World’.3 The change can be formalized as follows: V > ADV > PREP V > ADV > CONJ

Word change and language change  103 Table 5.1 Historical grammatical change of gòng/ka7 共 – throughout 3,500 years (Only the main function(s) of each period are noted in this table. See Tables 5.3 and 5.4 for more comprehensive counting)

Jiǎgǔwén, Shījīng - 14th – 5th c. bce (Archaic Chinese) Lúnyǔ, Mòzǐ, Guóyǔ - 5th c. ce (Medieval Chinese) Shì Shuō Xīnyǔ - 5th c. ce (Medieval Chinese) Dūnhuáng Biànwén - 6th – 11th c. (Medieval Chinese) Xīyóujì, Jīnpíngméi - 16th c. (Modern Chinese) Hónglóumèng - 18th c. (Modern Chinese) Early Southern Min (14th – 19th c.) Taiwanese Southern Min (20th c.)

V

ADV

CONJ

PREP

√

√

√

√ √

√

√

√

√

√

√

√

√

√

√

√ √

√ √

√

Finally, toward the end of 16th century and the beginning of the 17th century (the Ming dynasty) the conjunction AND-gòng 共 was replaced by the AND-hé 和 in Mandarin Chinese (Northern language), which is still pervasively used nowadays. Yet, Southern Min is not such a case. Kang7 共 is attested in both Early Southern Min (ESM, the language from the end of the 14th to the 19th centuries; Wu 2001a, 2001b; van der Loon 1992)4 and Modern Southern Min (TSM, 20th c.).5 CONJ and PREP are the main functions of kang7 共 in Southern Min, as in Table 5.1). In ESM, kang7 共 was the only AND-CONJ. It is interesting to note that the functions of kang7 共 were multiplied in ESM and TSM (see section 3.4 and section 4, Tables 5.3 and 5.4). Since the 20th century, in TSM, a second coordinating conjunction kah4/kap4 合 has come to support the function of CONJ by a shift domain, as attested by Lien (2015). The difference between the Chinese Mandarin gòng 共AND (since 5th – 18th c.) and kang7 共-AND is that, the latter can connect clauses and short sentences ESM, in addition to connecting words, NPs, and VPs. The main concern of this paper is the grammatical change from gòng 共-AND to ka7 共-AND. Our discussion focuses on the evolution of the 共-AND-CONJ, 共-WITH-PREP, and the 共-ADV that have a close relationship in their grammaticalization (Hagège 1975, 1982; Lemaréchal 1989).6 In Table 5.1, the functions are shown in a chronological order. 2 HISTORICAL EVOLUTION OF CHINESE-MANDARIN GÒNG 共

2.1  Gòng 共 in Archaic Chinese (14th – 2nd c. bce)7 The gòng 共 in Archaic Chinese (14th – 2nd c. bce) was rarely used, with only 0.03 % on average. The verb, with meanings of ‘offer’ or ‘provide’, is the original

104  Jang-Ling Lin function of gòng 共.8 Soon after, since Shījīng, gòng 共 was used as an adverb ‘together’,9 which has been kept to datefor more than 3000 years.10 In Pre-Archaic Chinese (14th – 11th c. bce), 共 was represented as two hands as or (two hands hold the valuables ) in the Jiǎgǔwén 甲骨文 ‘oracle bone script’. With or without the sign of , the ­original meaning of 共 was ‘offer respectfully (the valuables to gods)’ and was used as a verb, as in gòng yáng 共羊 ‘offer the sheep’, gòng mă 共馬 ‘offer the horses’ (Yao and Xiao 1989; Liu et al. 2005).11 In Early Archaic Chinese (10th – 6th c. bce) the representative work of this period is Shījīng 詩經 ‘Book of Odes’. Gòng 共 has only 9 occurrences (0.03%) that includes 2 occurrences of verbs with the meaning of ‘fulfill (the duty)’, e.g. (1), 2 occurrences of ADV with the meaning of ‘together’, e.g. (2), and 5 occurrences of nouns with the meaning of ‘duty’ or ‘colleague’.12 (1) 靖 共 爾 位 (Xiǎomíng 小明) jìng gòng ěr wèi quietly fulfill 2sg office ‘Quietly fulfill the duties of your offices’. (2) 共 武 之 服 (Xiǎoyǎ 小雅) gòng wŭ zhī fú together military gen service ‘render the military service together . . .’ In Late Archaic Chinese (5th – 2nd c. bce), the verbal gòng 共 acquired new meanings, such as ‘share’, ‘obey’, ‘surround’, and ‘take a bow’.13 It also acquired a new meaning ‘mutually’ as an adverb. Besides, gòng 共 developed a new meaning ‘respect’ as a noun.14 There are only 4 occurrences of gòng 共 among the 12000 characters (0.03%) in the Lúnyǔ 論語 ‘The Analects of Confucius’ (5th c. bce): 3 occurrences of verbs with the meaning ‘share’, e.g. (3), ‘surround’, ‘take a bow’, and 1 occurrence of ADV ‘together’, e.g. (4). (3) 子路 zĭlù Zĭ.Lù



曰:「

願

車

馬、

衣

yuē say

yuàn wish

chē chariot

mă horse

yī clothes

共」 裘， 與 朋友 (Ch. 5 Gōngyě Cháng 公冶長) qiú yŭ péngyǒu gòng fur comi friend share

‘Zi Lu said: I would like to share my chariots, horses and clothes with my friends’.

學」 (Ch. 9 Zǐhàn 子罕) (4) 子曰：「可 與 共 kě yŭ gòng xué can comi together learn ‘(One) can learn together with (someone else)’

Word change and language change  105 There are 22 occurrences of gòng 共 in Mòzǐ 墨子 (5th – 3rd c. bce), including 13 occurrences of verbs with the meanings ‘share’ (5 occurrences), e.g. (5), ‘provide’ (4 occurrences), ‘fulfill’ (2 occurrences), and ‘obey’ (2 occurrences). It also has 6 occurrences of ADV-together, e.g. (6), 2 occurrences of ADV-mutually, and 1 occurrence of nominal function with a new meaning ‘respect’.15 It can be observed that the functions and meanings started to have diverted in Mòzǐ. (5) 選

材 士 有力 者 三十 人 共 船 (Bèishǔi 備水) xuăn cái shì yǒulì zhě sānshí rén gòng chuán choice talent scholar powerful person thirty people share boat



‘(They) choose thirty talents and strong guys to share a ship’.

(6) 共 入 他 署 gòng rù tā shǔ together enter others office ‘enter others’ office together’

(Hàolìng 號令)

There are 47 occurrences in Guóyǔ 國語 ‘Discourses of the States’ (475–221 including 19 occurrences of verbal gòng 共 with the meanings ‘offer’ and ‘share’, e.g. (7), 2 or 3 occurrences of ADV-together, e.g. (8), and 23 occurrences of nouns, notable person names). bce),16

(7) 所以 共 幣帛 suǒyǐ gòng bìbó so.that share money.silk ‘Thus we offer money and silk’. (8) 夫 苦匏 不 材 于 人， 共 濟 fū kǔpáo bù cái yú rén gòng jì thus bitter.gourd neg material goal people together cross 而已。 (Vol. 20) éryĭ nothing.more ‘For the bitter gourd is not a useful material, (but) together, it helps to overcome the difficulty’.

2.2  Gòng 共 in Medieval Chinese period (2nd bce – 13th c. ce) Two crucial documents featuring Medieval Chinese are the Shì Shuō Xīnyǔ 世說 新語 ‘A New Account of Tales of the World’17 (5th c.) and the Dūnhúang Biànwénjí 敦煌變文集 ‘Transformational texts of Dunhuang’ (9th – 11th c.).18

106  Jang-Ling Lin The use of gòng 共 was on the increase in Medieval Chinese: there were 138 occurrences of gòng 共 among 79,800 tokens in Shì Shuō Xīnyǔ. The rate of use of Medieval Chinese was 0.17 %, higher than that of Archaic Chinese (0.03%). The adverbial function ‘together’ of gòng 共 was the most significant in this period, e.g. (9). It reached 134 occurrences.19 We found only 2 occurrences of verbal use, e.g. (10). Surprisingly, there was 1 occurrence of AND-conjunction, e.g. (11) and 1 occurrence of WITH-preposition, e.g. (12). The examples (11) and (12) are the evidence of polygrammaticalization (Craig 1991; Traugott and Heine 1991), as proposed in this paper. (9)



管寧、 華歆 guănníng huàxīn Guăn.Níng Huà.Xīn

共 gòng together

園 yuán garden

中 zhōng middle

鋤 chú hoe

菜， cài vegetable

見 地 有 片 金。 (Déxíng 德行) jiàn dì yǒu piàn jīn see soil have slice gold ‘Guăn Níng and Huà Xīn were hoeing the vegetables together in the garden and spotted pieces of gold in the soil’.

共 之。  (Zhèngshì 政事) (10) 文 王 之 囿， 與 眾 wén wáng zhī yòu yŭ zhòng gòng zhī Wén king gen pen comi people share 3sg ‘King Wén shared his parks with the people’. The first gòng 共-AND-conjunction appeared in Shì Shuō Xīnyǔ, connecting proper names: (11) 王僧彌、

謝車騎 共 王小奴 許 集。 (Yǎliàng 雅量) wángsēngmí xìejūjì gòng wáng xiăonú xŭ jí Wáng.Sēngmí Xìe.Jūjì conj Wáng.Xiăonú there gather ‘Wáng Sēngmí, Xìe Jūjì, and Wáng Xiăonú gathered at Wáng Xiăonú’s place’.

Shì Shuō Xīnyǔ also features the first gòng 共-WITH-preposition in addition to the first gòng 共-AND-conjunction (12) 既 jì since



知 zhī know

不 bù neg

能 néng can

踰 yú surpass

己， jǐ self

稍 shāo briefly

共 gòng comi

諸 zhū 3pl

生 敘 其 短 長。 (Wénxué 文學) shēng xù qí duăn cháng person talk prt short long ‘After he learned that (Lie’s interpretation) is no better than his own, he reviewed its weaknesses and strengths with his disciples’.

Word change and language change  107 The Dūnhuáng Biànwénjí 敦煌變文集 ‘Transformational texts of Dunhuang’ was one of the most important documents of the 9th to the 10th centuries. The rate of use of gòng 共 was 0.07 % among 229 instances. The functions were diversified. In addition to ADV (86 occurrences) and V (25 occurrences), it developed a new function as an adjective (ADJ). The WITH-preposition was in the majority (94 occurrences), e.g. (14). Interestingly, there were 10 occurrences of gòng 共 as AND-conjunction, including 9 which connects humans, e.g. (13–14), and only one occurrence that connects insects and fowls rather than humans, e.g. (15). Note that humans, animals or insects are both animate. In any case the AND-conjunction function was clear during Medieval Chinese period. In example (13), gòng 共 was used as a WITH-preposition, and bìng 並 as an AND-conjunction in Dūnhuáng Biànwénjí: (13) 楚 chŭ Chŭ



莊 zhuāng Zhuāng

王 wáng king

夜 yè night

女， 並 諸 群 nǚ bìng zhū qún lady conj 3pl group ‘The King Chuzhuang dreamed ministers’.

夢 mèng dream

共 gòng comi

後宮 hòugōng harem

美 měi beautiful

臣 飲 酒 (Vol.8) chén yĭn jiǔ minister drink wine that he was drinking with his harem and

The AND-conjunction gòng 共was found in Dūnhuáng Biànwénjí as well and 9 out of 10 occurrences of connected elements were humans: (14) 彌勒 mílè Mílè



共 gòng conj

文殊 wénshū Wénshū

親 qīn in.person

問答， wèndá discussion

因茲 yīnzī therefore

眾會 zhònghuì public

得 聞 經。 (Vol.3) dé wén jīng can listen doctrine ‘Mílè and Wénshū led the discussion in person; therefore the public could listen to the doctrine’.

The connected element of gòng 共-AND is non-human: (15) 蟭蟟



共

鵬

鳥，

如

同

飛

jiāoliáo gòng péng niăo rú tóng fēi conj roc bird as together fly cicada ‘The cicada and the huge bird are like a pair flying together’.

對。(Liújiā Tàizǐbiàn 劉家太子變) duì couple

108  Jang-Ling Lin

2.3  Gòng 共 in modern Chinese period (14th c.–First Half of the 19th c. ce) In Modern Chinese from the 14th c. onwards, our research is based on a set of texts with various time depth such as Gǔběn Lǎo Qǐ Dà 古本老乞大 (14th c.), Xī Yóu Jì 西遊記 (16th c.), Jīn Píng Méi Cíhuà 金瓶梅詞話 (17th c.), and Hóng Lóu Mèng 紅樓夢 (18th c.). At the beginning of the modern period, no examples of gòng 共 appeared in Gǔběn Lǎo Qǐ Dà 古本老乞大 ‘Old Sinologist - original version’ (19,222 tokens), reflecting the language of the 14th century (the Yuan Dynasty: 1271–1368).20 Later, during the 16th century, in the Xī Yóu Jì 西遊記 ‘Journey to the West’,21 9 among 111 occurrences of gòng 共 functioned as AND-conjunction; 86 of them were ADV. However, the function of AND-conjunction is clearer than the previous periods. The short NP connected by gòng 共-AND, i.e. [NP+共+NP] cannot be reanalyzed as prepositions, e.g. (16). This is not the case for Dūnhuáng Biànwénjí because the major function was ambiguous (see section 4.2). (16) 鯨

鰲 並 蛤 蚌， 蟹 鱉 共 魚 蝦。(Ch. 28) jīng áo bìng há bàng xiè biē gòng yú xiā whale sea.turtle conj clam oysters crab turtle conj fish shrimp ‘Whale, sea-turtle and clam, oysters; crab, turtle and fish, shrimp’

Gòng 共 had 106 occurrences, whereas hé 和 had 1412 occurrences in Jīn Píng Méi Cíhuà 金瓶梅詞話 ‘Plums in the Golden Vase’ of the 17th century.22 Although there were only 4 occurrences of AND-gòng 共, e.g. (17), they are significant in that 3 of the 4 occurrences connects inanimate elements. The adverbial function of gòng 共had 89 occurrences and remained important as it was in Archaic Chinese. Besides, there were 9 occurrences of verbs, 1 occurrence of nouns, and 3 occurrences of WITH- gòng 共, e.g. (18). AND-gòng 共 in Jīn Píng Méi Cíhuà connects inanimate elements, such as water and sky: (17) 群 星 與 皓 月 爭 輝， qún xīng yŭ hào yuè zhēng huī group star comi white moon dispute brightness 青 天 同 碧。(Ch. 8) 綠 水 共 lǜ shuĭ gòng qīng tiān tóng bì green water conj blue sky together blue ‘Stars and the moon were vying for radiance and the water and the sky were both green’. (18) 那 nà

dem

燈 dēng lantern

市 shì market

中 zhōng prep

人煙 rényān resident

湊集， . . . còují densely

判官 pànguān magistrate

Word change and language change  109



鐘馗 共 小妹 並 坐。(Ch. 3) 燈 dēng zhōngkuí gòng xiăomèi bìng zuò lantern Zhōng.Kuí comi lady side.by.side sit ‘The streets lined with lanterns were crowded with people.. . . The image of Zhōng Kuí and a maid sitting together was projected on the screen of the judge lantern’.

In the masterpiece of the 18th century, novel Hóng Lóu Mèng 紅樓夢 ‘Dream of the Red Chamber’, gòng 共 was used as an ADV in 109 of 115 occurrences.23 There was only 1 occurrence of AND-conjunction, e.g. (19) and 1 occurrence of WITH-preposition, e.g. (20).24 (19) 嬌 喘 共 細 腰 俱 絕。 (Ch. 21) jiāo chuăn gòng xì yāo jù jué tender gasp conj tiny waist together break.off ‘The sweet moan and the slender waist (all) disappeared’. (20) 若

小姐 共 你 多 情 同 鴛 ruò gòng nĭ duō qíng xiǎojiě tóng yuān if comi 2sg much love lady share mandarin.duck ‘If (I may) share the conjugal tent with the amorous lady (you),’

帳   (Ch. 24) zhàng mosquito.net

2.4 Conclusion–grammaticalization of gòng 共 in Mandarin (Northern Sinitic language) 2.4.1  Grammatical change of gòng 共 In Archaic Chinese (14th – 5th c. bce) gòng 共 was mainly though rarely used as a V. Since Shījīng (10th – 6th c. bce), the ADV-together function appeared.25 In the 5th century gòng 共 began to be used as AND-CONJ and WITH-PREP, as attested in Shì Shuō Xīnyǔ ‘A New Account of Tales of the World’. Our hypothesis is that the 共-AVD-together came from the 共-V-share, which originates from the共-V-offer. The process of this grammatical evolution is as follows 共-V-offer (14th – 11th c. bce) > 共-V-share (5th – 2nd c. bce) > 共-ADV-together The 共-ADV-together was then grammaticalized into the 共-AND-CONJ. At the same time in the 5th century, it was also grammaticalized into the 共-WITHPREP. The change can be accounted for by a new approach of polygrammaticalization as follows (Craig 1991; Traugott and Heine 1991) V > ADV > CONJ V > ADV > PREP

110  Jang-Ling Lin The ADV-together function appeared quite early since the Shījīng period (10th – 6th c. bce). However, it became unequivocal later. From the 5th to 18th centuries, ADV-‘together’ was the primary function of gòng 共 in Mandarin. 2.4.2 Connected elements of conjunctions – A process from animate to inanimate element In brief, at the beginning stage of the grammaticalization of gòng 共-AND in the 5th century, the connected elements were humans only. In the 9th and the 10th centuries, animals came to play as well. After the 17th century, the connected element can even be inanimate. From the 5th century to the beginning of the 17th century, the gòng 共-AND conjunction only connected NP, but the following 2 examples from Jīn Píng Méi Cíhuà ‘Jin Ping and Mei’ (17th c.) are surprising in that 2 clauses, instead of NPs, are connected (the same sentence appeared twice),26 e.g. (21): (21) 世上 萬般 哀 苦 事， 無非 死 shìshàng wànbān āi kŭ shì wúfēi sĭ in.world ten.thousand sadness bitter thing but dead 共 生 離。 (Ch. 15 and Ch.26) 別 bié gòng shēng lí separate conj living separate ‘There is nothing sadder and more traumatic than separation in life and parting at death’. The NP jointed by gòng 共-AND, [NP +共+NP], appeared only as a subject. But 2 examples from Dūnhuáng Biànwénjí ‘Transformational texts of Dunhuang’ (9th – 10th c.) manifested that the phrase [NP+共+NP] appeared as an object. Both examples are the object of the verb kàn 看 ‘see’. Finally, at the end of the 16th century and the beginning of the 17th century, the function of the gòng 共-AND conjunction was replaced by hé 和-AND. 2.4.3 The transition period of the gòng 共-AND – 9th and 10th century The role of the 共-AND conjunction in the 9th and 10th centuries was still ambiguous. In Dūnhuáng Biànwénjí, for the most cases the 10 occurrences can not only be analyzed as AND-conjunction but as WITH-preposition (see Campbell and Janda 2001 for reanalysis), e.g. (21). The ambiguity features a period of transition and has continued up to date (Hagège 1975, 1982; Lemaréchal 1989). (22) 大 dà great

王 wáng king

共 gòng conj

仙者 xiānzhě immortals

入 rù enter

內， . . . (Vol. 3, Ch. 4) nèi inside

Word change and language change  111 (Analysis of coordinating CONJ role: [NP+共+NP], the 共 is the connector AND) CONJ: The king and the immortals entered . . . Great king COMI immortals enter inside . . . (Reanalysis of PREP function: NP+ [共+NP], 共is the comitative WITH) COMI: The king entered with the immortals . . .27 3  KA7 共-THE MAIN AND-CONJ IN SOUTHERN MIN

3.1 Introduction It is important to note that kang7 共 became a very important function word of both CONJ and PREP in Southern Min. Kang7 共 mainly functioned as ANDCONJ in Early Southern Min (ESM) and could connect NPs, VPs, clauses, and short sentences (see section 3.3). Kang7 共 was widely used in ESM playscripts (language from the end of 14th – 19th c.) with 4.38 % in general. In the 20th century,28 in Taiwanese Southern Min (TSM), ka7 共-AND was still used as AND conjunction. In Sìchóngzòu 四重奏 ‘Quartets’ (TV series, 12287 Chinese characters), it had a high rate of 8.57 % (66 occurrences out of 770).29 In the 20th century, in TSM, kah4/kap4 合 has also come to play in the function of the AND-CONJ by a domain shift (Lien 2015),30 52 among 245 appearances in Sìchóngzòu ‘Quartet’. The rate of use of kah4/kap4 合-AND reached 21.22% as in Table 5.2. From the 14th to the 20th centuries, the connected elements of AND-CONJ in both ESM and TSM could be animate elements, e.g. (24), (25), (34), (37), and (40), inanimate elements, e.g. (26), etc., or even abstract elements, e.g. (35), (38), (41), (43) . . ., etc. Both ka7 共-AND-CONJ and kah4/kap4 合-AND-CONJ shows the tendency. See section 3.3. Besides, ka7 共 had an unexpected function of OR-CONJ in ESM, but it has been replaced by ah8si7 抑是 in TSM (Yang 1991). The change is as follows. ka7 共-OR > ah8si7 抑是-OR On the other hand, ka7 共 had multiple functions in Early Southern Min, including an object marker, goal, benefactive, ablative, comparative morpheme, verb, and resultative verbal complement. Table 5.2 The use of ka7 共 and kah4/kap4 合 as AND-CONJ in the 20th century in ­Taiwanese TV series Sìchóngzòu ‘Quartet’ Sìchóngzòu Total characters 12287

Occurrences

Occurrence rate in Total characters

AND use

AND use rate in Occurrences

ka7 共 kah4/kap4 合

770 245

6.27 % 1.99 %

66 52

8.57 % 21.22 %

112  Jang-Ling Lin In TSM, ka7 共 could no longer function as an ADV as in Mandarin, which was the original and the most important function.31 In ESM, only 2 or 3 occurrences of ka7 共 were found.

3.2  Corpus on Southern Min 3.2.1  Corpus of Early Southern Min-language from the 14th century to the 19th century Listed below are a set of playscripts representing the early Southern Min: Lì Jìng Jì 荔鏡記 ‘Romance of litchi and mirror’, 1566 edition (henceforth, LJJ-1566). 7335 Chinese characters. Lì Zhī Jì 荔枝記 ‘Romance of the litchis’, 1581 edition (henceforth, LZJ1581). 4361 Chinese characters. Lì Zhī Jì 荔枝記 ‘Romance of the litchis’, 1651 edition (LZJ-1651). 8208 Chinese characters. Lì Zhī Jì 荔枝記 ‘Romance of the litchis’, 1884 edition (LZJ-1884). 9147 Chinese characters. Jīn Huānǚ 金花女 ‘Lady Jin Huanü’, 1573–1620 edition (JHN-1573–1620). 1469 Chinese characters. Sū Liù Niáng 蘇六娘 ‘Dame Su Liu Niang’, 1573–1620 edition (SLN-1573– 1620). 783 Chinese characters. Tóngchuāng Qínshūjì 同窓琴書記 ‘The romance of the classmates with lute and books’, 1782 edition (TCQSJ-1782). 2007 Chinese characters. 3.2.2  Corpus of Taiwanese Southern Min (TSM)–20th century The discussion of the conjunctions of Taiwanese Southern Min is based on computerized texts of TV Films of 20th & 21st centuries as shown below: Yuèliàng Chūláile 月亮出來了 ‘The moon is up’. 5800 Chinese characters. Sìchóngzòu 四重奏 ‘Quartet’. 12287 Chinese characters. Hòushān Rì Xiānzhào 後山日先照 ‘The Sun Shines First Behind the Mountain’. 11504 Chinese characters. Liǎngshì Tóngxiū 兩世同修 ‘Blessings of Two Generations’.

3.3 共 and 合 in AND conjunctions and other functions ESM and TSM 3.3.1 Ka7 共-AND-CONJ connects NPs The conjunction ka7 共-AND connects NP. The whole connected NPs could appear in subject position, as in (23) and (24) or in object position as in (25) and (26):

Word change and language change  113 The connected NPs [inanimate ‘carriage’+ animate ‘horse’] in (23) function as the subject of the sentence: (23) 車 共 馬 來來去去。 tshia1 ka7 be2 lai5lai5khi3khi3 carriage conj horse come.come.go.go ‘Carriages and horses come and go’.

(TCQSJ-1782)

Likewise, the NPs connected by ka7 共-AND were the subject: (24) 黃蜂 共 尾蝶 盡都 成 ng5phang1 ka7 be2tiap8 to1 sing5 bee conj butterfly all become ‘Bees and butterflies are all in pairs’.

對。 tui3 pair

(TCQSJ-1782)

The NP hong7ka7luan5 鳳共鸞 is the object of the verb phah4suann3 打散: (25) 狂

風 打散 鳳 共 鸞。 (LJJ-1566) kong5 hong1 phah4suann3 hong7 ka7 luan5 fierce wind separate male.phoenix conj female.phoenix ‘The Fierce wind breaks up the male phoenix and the female phoenix’.

Likewise, the connected NP appears after the verb: (26) 過

盡 千 山 共 萬 水。 (LJJ-1566) kue2 tsin7 tshian1 san1 ka7 ban7 sui2 cross totally thousand mountain conj ten.thousands streams ‘Having crossed thousands of mountains and streams . . .’.

3.3.2 Ka7 共-AND-CONJ connected VPs In the following example, the VPs connected by ka7 共 are the object of the matrix verb: (27) 亦 著 學 掃 厝 共 煎 茶。 (SLN-1573–1620) ia7 tioh8 oh8 sau3 tshu3 ka7 tsuann1 te5 also need learn sweep house conj cook tea ‘(I) also had to learn how to clean the house and make tea’. 3.3.3 Ka7 共-AND-CONJ connects clauses and sentences The AND connector can also connect clauses and sentences in ESM and two forms were found in the corpus: a monosyllabic ka7 共in (28) and a disyllabic form ia7ka7 亦共 in (29). The disyllabic form had only one example.

114  Jang-Ling Lin (28) 花 紅 共 柳 綠。(LJJ-1566) hue1 ang5 ka7 liu2 lik8 flower red conj willow green ‘The flower is red, and the willow is green . . .’ (29) 任 汝 山 高 亦共 水 深。 (LZJ-1884) jin5 li2 suann1 ko1 ia7ka7 tsui2 tshim1 even 2sg mountain high conj water deep ‘Even the mountain is high, and the water is deep (as your power)’. However, such a kind of example was rare. Only four examples in total were found,32 and they are attested only in 4 versions of Lì Jìng Jì and Lì Zhī Jì (LJJ1566, LZJ-1581, LZJ-1651, LZJ-1884). The other two examples beside (28) and (29) are (30) and (31): (30) 花 迷 共 酒 醉。 (LZJ-1651) hua1 be5 ka7 tsiu2 tsui3 flower captivate conj wine getting.drunk ‘The flowers are captivating, and one is getting (me) drunk’. (31) 猿 啼 共 鳥 叫。(LZJ-1581) uan5 thi5 ka7 tsiau2 kio3 monkey cry conj bird chrip ‘Monkeys are crying and birds are chirping’. 3.3.4  About the juxtaposition It is known that single juxtaposition (paratactic device)33 is used more often than a hypotactic one in the spoken register. The paratactic is nevertheless the most common from a typological perspective (Meillet 1912; Grévisse 1975; Mithun 1988). The generalization accounts for Early Southern Min as well. There are 2 similar sentences, as shown below, from the 16th century and the 17thcentury respectively. They differ in that the connector 共 is employed in (32), while the connector is absent in (33). (32) 深



深

拜， 拜

謝

天，

乞

靈

共

乞 聖。 (LZJ-1581) tshim1 tshim1 pai3 pai3 sia7 thinn1 khit4 ling5 kang7 khit4 sing3 deep deep bow bow thank sky pray spirits conj pray saints ‘Bow deeply. Bow to the heaven. Pray to the (ancestral) spirits and (pray to) the saints’.

Word change and language change  115 (33) 深

深 拜， 拜 告 天， 乞 靈 聖。 (LZJ-1651) tshim1 tshim1 pai3 pai3 ko3 thinn1 khit4 ling5 sing3 deep deep bow bow talk sky pray spirits saints ‘Bow deeply. Bow to the heaven. Pray to the (ancestral) spirits and the saints’.

3.3.5 AND conjunctions in Taiwanese Southern Min of the 20th century In the 20th century, ka7 共 functioned as AND-CONJ and WITH-PREP in reciprocal construction, e.g. (36) (Lien 2015). 3.3.5.1  Ka7 共-AND in Taiwanese Southern Min–20th century In the 20th century, ka7 共 and kah4/ kap4 合 both functioned as AND-CONJ and connected animate elements, e.g. (34) and (37), or inanimate elements, e.g. (35) and (38). In the TV series Sìchóngzòu 四重奏 ‘Quartets’ (12287 Chinese characters), ka7 共-AND was still the predominant CONJ, with 8.57% (66 occurrences out of 770), as in Table 5.2. The ka7 共-AND connected animate elements are the object of the V: (34) 你



愛

扣

電話

呼

阿雪仔

共

li2 ai3 kha3 tian7ui7 hoo7 a1suat4a2 ka7 2sg must call telephone give A.Suat.a conj ‘You have to make a call to A-Suat-a and Po-lian’!

寶蓮 喔! (Sìchóngzòu) po1lian5 o0 Po.lian prt

The ka7 共-AND connected inanimate elements is the subject: 時間 (35) 一個 人 e 精神 共 甘哪 tsit8e5 lang5 e5 tsing1sin5 ka7 si5kan1 kan1nah2 one person gen energy conj time how 這爾 濟。 (Sìchóngzòu) tsiah4ni7 tsue7 so.many much ‘How could a person have so much energy and time’?

有 u7 have

Like ESM, ka7 共 still functioned as a WITH-PREP in 20th century: (36) 你



擱

倒轉來

共

阿嬤

睏 好無。 (Hòushān Rì Xiānzhào) li2 koh4 to2tng2lai5 ka7 a1ma2 khun3 ho2bo5 2sg again return.come comi grand.mother sleep yes.no ‘(Speaking to her grandchild) come back again and sleep with me, please’.

116  Jang-Ling Lin 3.3.5.2  Kah4/Kap4 合-AND in Taiwanese Southern Min–20th century Moreover, in the 20th century, kah4/kap4 合 acquired the function of AND-CONJ (Lien 2015). Kah4/kap4 合 was mainly used as a verb meaning ‘unite’, ‘gather’, ‘reflect’. Lien (2015) attested AND-CONJ in the reciprocal construction and the WITHPREP function connected by kap4/kah4 合 (and its dialectal variants as tsham1 摻, ham7 含, and kau1 交) in Taiwanese Southern Min. Kap4/kah4 合 was never used as AND-CONJ in Early Southern Min. It developed the function in the 20th century. Besides, at the end of the 19th century kap4/kah4 合 can be used as WITH-PREP, as attested in LZJ-1884 with 11 occurrences). Gradually, the function thrived in the 20th century, e.g. (39). Kap4/kah4 合 as an AND-CONJ is attested in Sìchóngzòu ‘Quartets’ as well: 52 occurrences among 245 (21.22 %). In (37), the connected phrase is the subject of the sentence: (37) 阿爸

合 阿母 非常 疼 自己 的 a1pah4 kah4 a1bu2 hui1siong5 thiann3 tsu7ki2 e5 father conj mother very.much love self gen ‘Father and mother love their own child(ren) very much’.

囝。 (Sìchóngzòu) kiann2 child

Both connected elements in (38) come after the verb as its object: (38) 阿滿



a1buan2 A.buan

伊 i1 3sg

直接 跨過 婚姻 合 tit8tsiap4 hann7ke3 hun1in1 kah4 directly jump marriage conj

享受

著

做

媽媽

e

單純

喜悅。 (Sìchóngzòu) hiong2siu7 tioh8 tsue3 ma1ma1 e5 tan7sun5 kah4 hi2iat8 enjoy asp do mother gen simplicity conj joy ‘Braving marriage and reality, A-Buan is enjoying the simplicity and the joy of being a mother’.

(39) 這 tsit4



現實， hian7sit8 reality

遍 的 感覺 合 pian3 e5 kam1kak4 kah4 dem time gen feeling comi 無 仝款。 (Sìchóngzòu) 遍 攏 pian3 long1 bo5 kang5khuan2 time all no.have same ‘The feeling this time is different from before’.

合

早 tsa2 earlier

幾 kui1 several

3.3.6 Ka7 共-OR alternative conjunction 3.3.6.1  Ka7 共-OR conjunction in Early Southern Min Ka7 共-OR was the only Alternative Conjunction in ESM. It connected elements that are contrastive or opposite:

Word change and language change  117 (40) 管乜

尾蝶 共 黃蜂， 須 待 凰鳳 (LJJ-1566) kuan2mih4 be2iah8 kang7 ng5phang1 su1 thai7 hong7hong5 whether butterfly alt bee have.to wait phoenix ‘Regardless of butterflies or bees, (they) have to wait for the phoenix . . .’

(41) 二 頭 不 知 生 共 死。 (JHN-1573–1620) ji7 thau5 m7 tsai1 sinn1 kang7 si2 two side neg know alive alt dead ‘Two heads (of the dead bodies) do not know whether (they are) alive or dead’. (42) 莫 說 有情 共 無情， . . . boh8 seh8 iu2tsing5 kang7 bo5tsing5 do.not say have.love alt no.have.love ‘Don’t say (we have) love or not . . .’ (43)

(JHN-1573–1620)

益春 見 你 只 模樣， 曉得 你 有 共 無。(LZJ-1651) iah4tshun1 kinn3 li2 tsi2 boo5iunn7 hiau2tit4 li2 u7 kang7 bo5 Iah.tshun see 2sg dem appearance know 2sg have alt not.have ‘From your appearance, (I) know whether you have (the properties) or not’.

(44) 生 共 死， 不 敢 忘 恩義。 (LZJ-1884) sinn1 kang7 si2 m7 kann2 bong5 un1gi7 alive alt dead neg dare forget favor ‘Regardless of life or death, (I) can never forget your favors’. 3.3.6.2  The change of the OR-CONJ in TSM-ka7 共-OR > a7-si7 抑 是/亦是-OR In TSM ka7 共-OR was replaced by the disyllabic form ah8-si7 抑是/亦是-OR (Yang 1991): (45) 你

抑是 畫 是 在寫字 符仔 啦? (Hòushān Rì Xiānzhào) li2 si7 teh4sia2ji7 ah8si7 uei7 hu5a2 la3 2sg cop write.word alt painting talisman prt ‘Are you writing characters or drawing the Taoist charm’?

There was only one occurrence of 亦是-OR: (46) 卜 beh4 want 亦是 ah8si7 alt



回 hue5 return 無? bo5

外家 來 恭賀， 禮物 有 款 gua7ke1 lai5 kiong1ho7 le2but8 u7 khuan2 parents.in.law come congratulate gift have prepare (Liǎngshì Tóngxiū)

neg

‘If (you) want to visit your maiden home for congratulation, have (you) prepared the gift or not’?

118  Jang-Ling Lin 3.3.7 Ka7 共 in prepositions (OM, Goal, ABL, COMP) and other functions In ESM and TSM, ka7 共 is an important functional word, mainly used as CONJ and PREP (see Table 5.1). As a PREP, ka7 共 can function as an object marker like bă 把 in Mandarin, e.g. (47), goal, ablative, and comitative in non-reciprocal constructions (Lien 2015). Ka7 共 was rarely used as an object marker in ESM, e.g. (48). (47) 無啊!

你 是 共 遮 當做 旅館 bo5a li2 si7 ka7 tsia1 tong3tsue3 lu2kuan2 not ptc 2sg cop om here take.do hotel ‘My, oh, do you treat this place as a hotel, don’t you’?

(Yùeliàng Chūláile)

是毋? si7m7 yes.no

(48) 益春，



恁 啞娘 木屐 擺 了， 快 共 iah4tshun1 lin2 a1niu5 bak8kiah8 pai2 loo7 kin2 kang7 Iah.tshun 2sg.gen lady slipper crook prt quickly om 移 正。 (LJJ-1566) 伊 sua2 tsiann3 i1 3sg move good.position ‘Your mistress’s clogs are laid randomly, Iah-tsun. Arrange them properly’.

3.4  Conclusion on Southern Min ka7 共-AND conjunction The vital change from ESM (14th – 19th c.) to TSM (20th c.) are: (1) the categorical status of ka7 共’s functions shifted a lot in the developmenti. In ESM, it was used only as AND-CONJ,34 while in TSM, it is still commonly used as ANDCONJ (see Table 5.2). Another AND-CONJ kah4/kap4 合has also come to play. In general, WITH-PREP ka7 共 is still the most important function in ESM and TSM, (2) ka7 共 could not function either as ADV or as OR conjunction in TSM (see Table 5.3), and (3) ka7 共 as an object marker was very marginal in ESM,35 but quite predominant in TSM.

Table 5.3 Functions of Southern Min ka7 共: from ESM (14th – 19th c.) to TSM (20th century) Southern Min

CONJ

PREP

ADV

RVC

and or COMI Goal BEN ABL OM COMP V together Early Southern √ Min Taiwanese √ Southern Min

√

√

√

√

√

√

√

√ √

√

-

√

√

√

√

√

√

-

√

-

Word change and language change  119 4  CONCLUSION Lien (1994) stated that grammaticalized words will gradually lose their meaning and keep their functions. Only in the transition stage, two roles, full meaning and functioning unit respectively, could co-exist. In the case of the grammaticalization of 共the transition stage was very long. Gòng 共 had been used as V and ADV until the 18th century (in Hóng Lóu Mèng ‘Dream of Red Chamber’). In Modern Chinese Mandarin the ADV-together gòng 共 has changed into a disyllabic form ‘gòngtóng 共同’, e.g. gòngtóng shēnghuó 共同生活 ‘to live together’. In ESM and TSM, only CONJ and PREP were attested (see Table 5.4). Before concluding the paper, here is a short reflection on language change from a broader perspective. Lemaréchal (1989) mentioned that: Genesis or renewal? This is the question that must be arisen in historical grammar.36 In the case of the Chinese 共, it is the case of the genesis of two new categories: 共-WITH preposition and the 共-AND conjunction in 5th century. As to the Chinese AND conjunctions, undoubtedly, they had been renewed constantly since Pre-Archaic Chinese to date. Below is a brief list of AND-CONJ in Mandarin and Southern Min: tà 眔, yòu 又,37 bìng 並, jí 及, yĭjí 以及, lián 連, tóng 同, jiāng 將, yú 于, gòng 共, jì 暨, yŭ 與, hé 和, gēn 跟, ka7 共, and kap4 合. The chance for a word from any category to become a conjunction is equal. The requirement to renew the expression or revive the strength can yield word replacement, as Meillet (1912) argued. The things, the beliefs, the vogue and the society change can provoke the word change (see also Hagège 1985). Therefore, conjunctions that express permanent concepts should not affect change. Secondly, conjunctions are structural word whose grammatical function should be very stable throughout the time. Third, languages from the same family should have similar conjunctions. But in practice, the situation is in opposite; conjunctions can evolve from different categories such as nouns, verbs, etc. A number of Indo-European languages have different conjunctions.38 Likewise, Mandarin (Northern language) and Southern Min (Southern language) also have very different AND conjunctions. Hagège (1982) stated that 19% of languages of the world made no difference between the conjunction AND and the preposition WITH. Chinese is such a case. AND and OR are the two most important conjunctions in Indo-European languages (Meillet 1912). The situation applies to Sinitic languages too. Interestingly the same marker ka7 共 can be used as coordinating conjunction ‘AND’ and as alternative conjunction ‘OR’ in Early Southern Min. Table 5.4  The functions of gòng/ka7 共 for 3500 years

Archaic Chinese (14th – 2nd c. bce) Medieval Chinese (1st bce – 13th c. ce) Modern Chinese (14th – 19th c.) Early Southern Min (14th – 19th c.) Taiwanese Southern Min (20th c.)

CONJ

PREP

V

ADV

N

√ (rare) √ √

√ √ √ √

√ √ √ √ -

√ √ √ √ -

√ √ -

120  Jang-Ling Lin NOTES 1 The previous version of the paper was presented at the international conference Reflections of Diachronic Change Mirrored in Early Southern Min Texts, a workshop of a joint project of National Science Council (NSC) Taiwan (present Ministry of Science and Technology (MOST)) and the ANR France, held at National Tsing Hua University, Taiwan, 26–27 November  2011. I  owe a debt of gratitude to MOST (Taiwan) and ANR (France). I am also grateful to Professors Chinfa Lien (NTHU, Taiwan), Huei-ling Lai (NCCU, Taiwan), and Alain Peyraube (EHESS, Paris) for their insightful comments and suggestions. My heart-felt thank goes to Professor Chinfa Lien and Professor Alain Peyraube for their great encouragements and supports on this study. I am particularly indebted to Professors Chinfa Lien for all of his support and precious comments, including the use of his online Southern Min database. 2 The sign of ‘ >’ means ‘changed into’. 3 See note 17. 4 Modern Southern Min ka7 共was pronounced as câng in the Arte de la Lengua chiõ chiũ by Dominicain missionary Melchor de Mançano compiled in 1620 (the Ming dynasty) in Manila (Philippines). It was pronounced as kā, kiōng (reading style), and kāng (colloquial, vernacular or in Chin-chew dialect) in Douglas 1873. 5 The first version of the Southern Min opera Lì Jìng Jì 荔鏡記 ‘Romance of litchi and mirror’ was republished in the year of Bǐngyín 丙寅 of the era Jiajing 嘉靖, the Ming dynasty (1566) (Wu 2001a); the second version Lì Zhī Jì 荔枝記 ‘Romance of the litchi’ was republished in the year of Xīnsì 辛巳 of the era Wanli 萬曆, the Ming dynasty (1581) (Wu 2001b). According to Wu (2001a, 2001b) and van der Loon (1992), both versions probably reflect the language used at the end of the 14th century or the beginning of the 15th century, as argued in Peyraube and Lin 2010; Lin and Peyraube 2012, 2015. 6 Hagège (1982) notes that 19% of languages make no difference between AND and WITH and Sinitic languages exhibit a similar phenomenon. 7 This paper follows Peyraube’s (2011) periodization. 8 Later, the meaning ‘offer’ or ‘provide’ was replaced by the word ‘供’, pronounced gōng. 9 The duration of the Shījīng ‘Book of Odes’ is from the 10th c. to the 6th c. bce. 10 In addition, since the period of Shījīng, gòng 共 was also used as place name and person name in disyllabic forms, notably in Guóyǔ 國語 ‘Discourses of the States’ (475–221 bce): gònghuá 共華, gòngsì 共賜, gònggōng共工. 11 In Jīnwén 金文 ‘Chinese Bronze Inscriptions (about 1300 bce to 219 ce)’, and Jiǎgǔwén 甲骨文, 共 is a verb and is manifested as, denoting ‘(two hands hold the valuables) to offer (to gods)’. In the Archives of Bronze Images and Inscriptions (about 1300 bce to 222 bce) of Academia Sinica, 65 occurrences are observed, with the functions of V (‘offer; provide; respect’) and N. 12 The meaning of the noun gōngrén 共人 in Shījīng is ‘colleague’. Another possible meaning is the ‘southern barbarians of the Wu 吳 and Yue 越’. 13 Later, the verb ‘surround’ and ‘take a bow’ (in Lúnyǔ) was replaced by the word 拱, pronounced gǒng. 14 Later, the verb ‘obey’ and the noun ‘respect’ (both appeared in Mòzǐ), was replaced by gōng 恭. 15 The 2 occurrences of ADV-mutually appear in a disyllabic form gòngxīang 共相, as in gòngxīang jĭngjiè 共相儆戒 [to warn mutually] (Tiān zhì 天志). 16 The percentage of the use of gòng 共 in Guóyǔ is approximately a little more than 0.05%. 17 Shì Shuō Xīnyǔ, ‘A New Account of Tales of the World’, compiled by Liu I-Ch’ing (403–444), was a collection of anecdotes, short conversations, and pithy observations on people who lived in China approximately between 150 and 420 ce (Liu 2002). 18 Dūnhuáng Biànwén ‘Transformational texts of Dunhuang’ are Buddhist song-tales dated 850–1025.

Word change and language change  121 19 Interestingly, there were 53 occurrences of ADV-together followed by verbs of saying: 簡文 與 許玄度 共 語。 jiănwén yŭ xŭxuándù gòng yŭ Jiăn.Wén conj Xŭ.Xuán.dù together talk ‘Jiăn Wén and Xŭ Xuán-dù was talking together’.

(Qīngdǐ 輕詆)

20 Note that the language in Lǎo Qǐ Dà 老乞大 reflects a kind of Chinese influenced by Mongolian in the 13th and 14th centuries (Kang 1985; Liang 1998). 21 Xī Yóu Jì by Wu Cheng’en, commonly known as ‘The Legend of the Monkey King’ in English-speaking world, is one of the four great classical novels in Chinese literature. 22 The use of 與 in Jīn Píng Méi Cíhuà are 3316 occurrences in total. The majority is prepositional use (with, to, for. . .). 23 In Hóng Lóu Mèng, the use of hé 和 was increasing, with 1599 occurrences, in function of both AND-conjunction and WITH-proposition. 24 Besides, there were one V and one N in use. 25 Otherwise, we are not sure if there was gòng 共-ADV function in Jiăgŭwén period. 26 See Cristofaro 2003; Dixon 2009; Aikhenvald 2009 for discussions on clauses. 27 See Hagège 1975; Zhu 1982; Cheng and Tsao 1995; Hsiao 1995 for the differentiation between Chinese AND conjunctions and WITH prepositions. 28 For details on Southern Min playscripts, see Wu 2001a, 2001b, 2001c, 2001d, 2002a, 2002b, 2003. On the dating of Southern Min playscripts, see van der Loon 1992; Wu 2001a; Lin and Peyraube 2012, 2015. 29 In 20th century, ka7 共 became an object marker (like bă 把 in Mandarin), which was not the case in Early Southern Min (only 1 out of 4 occurrences in the playscripts) (Lien 2015). 30 In ESM, kah4/kap4 合 was never used in AND-CONJ but rarely used as a verb denoting ‘convene’, ‘unite’, ‘gather’, and ‘reflect’. 31 3 occurrences of ADV use were observed in Sìchóngzòu (20th c.), all in literary writing. 32 For clause and sentence discussion, see Cristofaro 2003. 33 The paratactic device involves the connected elements that are joined without the connector AND; on the contrary, the hypotactic one features the connector AND between two or more connected elements. 34 Besides, there are 3 cases of tang5 同-AND in LJJ-1566, 1 case in LZJ-1581, 1 case in TCASJ-1782, 8 cases of kah4 甲-AND in LZJ-1651, 2 cases of ho5 和-AND in LJJ1566 (Mandarin context), 3 cases of u2 與-AND in LJJ-1566, 1 case in LZJ-1651, and 1 case in LZJ-1884. These are due largely to homophony or to Mandarin context. 35 The commonly used functions of ka7 共 are goal and benefactive in ESM as well as in TSM. 36 Translated from French ‘Genèse ou renouvellement? C’est la question qu’il faut souvent se poser en grammaire historique’ (Lemaréchal 1989). 37 In Pre-Achaic Chinese, there were tà 眔-AND (Takashima 1989), which disappeared very soon and yòu 又-AND, which connected numbers only as in wéi shí yòu lìu nián 隹（唯）十又六年 [art ten and six year], the year of 16 (Kè Zhōng 克鐘 ‘Bell Ke’, the end of the Western Zhou: 1046–771 bc). 38 One of the examples: English AND and German UND are of different origin.

REFERENCES Aikhenvald, Alexandra. 2009. Semantics and grammar in clause linking. In The semantics of clause linking – A cross-linguistic typology. New York: Oxford University Press. Brinton, Laurel J., and Elizabeth Closs Traugott. 2005. Lexicalization and grammaticalization all over again. In Historical linguistics 2005, ed. by Joseph C. Salmons and Shannon Dubenion-Smith, 3–19. Amsterdam, Philadelphia: John Benjamins.

122  Jang-Ling Lin Campbell, Lyle, and Richard Janda. 2001. Introduction: Conception of grammaticalization and their problems. Language Sciences 23:93–112. Cheng, Ying鄭縈, and Feng-Fu Tsao 曹逢甫. 1995. Relationship between the roles of Southern Min ka. In Papers from the 1994 conference on language teaching and linguistics in Taiwan. Vol. 1, Southern Min, ed. by Tsao Feng-fu and Tsai Mei-Hui, 23–45. Taipei: The Crane Publishing. Craig, Colette G. 1991. Way to go in Rama: A case study in polygrammaticalization. In Approaches to grammaticalization. Vol. 1, Focus on theoretical and methodological issues, ed. by Elisabeth Cross Traugott and Bernd Heine, 455–492. Amsterdam, Philadelphia: John Benjamins. Cristofaro, Sonia. 2003. Subordination. New York: Oxford University Press. Dixon, Robert M.W. 2009. The semantics of clause linking in typological perspective. In The semantics of clause linking: A cross-linguistic typology, ed. by Robert M.W. Dixon and Alexandra Aikhenvald, 1–55. New York: Oxford University Press. Douglas, Carstairs. 1873. Chinese-English dictionary of the vernacular or spoken language of Amoy with the principal variations of the Changchew and Chinchew dialects. London: Trübner & Co. Grévisse, Maurice. 1975. Le bon usage. Louvain: Duculot. Hagège, Claude. 1975. Le problème linguistique des prépositions et la solution chinoise (avec un essai de typologie à travers plusieurs groupes de langues). Paris: Société de Linguistique de Paris. Hagège, Claude. 1982. La structure des langues. Paris: Presses Universitaires de France. Hagège, Claude. 1985. L’homme de paroles - Contribution linguistique aux sciences humaines. Paris: Fayard. Heine, Bernd, Ulrike Claudi, and Friederike Hünnemeyer. 1991. Grammaticalization: A conceptual framework. Chicago: Chicago University Press. Heine, Bernd, and Tania Kuteva. 2002. World lexicon of grammaticalization. New York: Cambridge University Press. Hopper, Paul J., and Elizabeth Closs Traugott. 2003. Grammaticalization. New York: Cambridge University Press. Hsiao, Su-Ying蕭素英. 1995. Semantic and syntax of coordinating conjunction of Southern Min ‘kah4’, ‘na2’. In Papers from the 1994 conference on language teaching and linguistics in Taiwan. Vol. 1, Southern Min, ed. by Tsao Feng-fu and Tsai Mei-Hui, 141–165. Taipei: The Crane Publishing. Kang, Shizhen 康寔鎮. 1985. ‘Lǎoqǐdà’ ‘Piáotōngshì’ yánjiū: zhū shū zhī zhùchéng jí qí shū zhōng Hànyǔ yǔyīn yǔfǎ zhī xīlùn 「老乞大」「朴通事」研究:諸書之著成及其 書中漢語語音語法之析論 [Study on Laoqida and Piaotongshi: On different versions and phonetic and grammar of the Yuan Dynasty]. Taipei: Xuesheng shuju. Kuryłowicz, Jerzy. 1965. The evolution of grammatical categories. Diogenes 51:55–71. Lemaréchal, Alain. 1989. Les parties du discours – Sémantique et syntaxe. Paris: Presses Universitaires de France (PUF). Liang, Wuzhen 梁伍鎮. 1998. Nogeoldae Paktongsa yeonggu: Haneomun e boi neun eohwi wa munbeop ui tikjing eul jungsim euro 老乞大, 朴通事研究: 漢語文에보이 는語彙와文法의特徵을中心으로 [A study of Nogoldae and Pakt’ongsa: Focusing on the characteristics of vocabulary and grammar appearing in their Chinese sentences]. Doctoral dissertation, Korea University, Seoul. Lien, Chinfa 連金發. 1994. Táiwān Mǐnnányǔ de wánjié shíxiàngcí shìlùn 台灣閩南語的 完結時相詞試論 [The perfective and the temporal terms in Taiwanese Southern Min]. In Paper from the 1994 conference on language teaching and linguistics in Taiwan. Vol. 1, Southern Min, ed. by Tsao Feng-fu and Tsai Mei-Hui, 121–140. Taipei: Crane Publishing.

Word change and language change  123 Lien, Chinfa. 2015. The condition and change of 共 vis-à-vis 合 in Southern Min with a sidelight on an intra-dialectal variation. Journal of Chinese Linguistics 43:1–33. Lin, Jang-Ling 林征玲, and Alain Peyraube. 2012. Individual classifiers in early Southern Min (14th – 19th c.). In Plurality and classifiers across languages in China, ed. by Xu Dan, 101–120. Berlin: Mouton de Gruyter. Lin, Jang-Ling 林征玲, and Alain Peyruabe 貝羅貝. 2015. Lǎoqǐdà hé Lìjìngjì de zhǐshì dàicí zhè yǔ nà 《老乞大》和《荔鏡記》的指示代詞“這”與“那” [The demonstratives of Lao Qida and Li Jing Ji]. In Méizǔlín jiàoshòu bāzhì shòuqìng xuéshù lùnwénjí 梅祖麟教授八秩壽慶學術論文集 [A festschrift in honor of Tsu-lin Mei’s 80th birthday], ed. by Bo Hong 洪波, Fu-xiang Wu 吳福祥, and Chaofen Sun 孫朝奮, 60–76. Peking: Capital Normal University Press. Liu, I-Ch’ing. 2002. Shih-shuo Hsin-yu: A new account of tales of the world. Michigan: University of Michigan. Liu, Jian, and Alain Peyraube. 1994. History of some coordinative conjunctions in Chinese. Journal of Chinese Linguistics 22:179–201. Liu, Zhiji, Zhang Deshao, Takashima Ken-ichi, and Zang Kehe, eds. 2005. Hànyīng duìzhào Jiǎgǔwén jīnyì lèijiǎn 漢英對照甲骨文今譯類檢 [Translation of the oracle bone script in contemporary Chinese-English]. Nanning: Guangxi Jiaoyu Chubanshe. Mançano, Melchor de. 1620. Arte de la lengua Chiõ Chiu. Ms., University of Barcelona. Meillet, Antoine. 1912. Renouvellement des conjonctions. In Linguistique historique et linguistique générale, 159–174. Paris: Champion. Mithun, Marianne. 1988. The grammaticalization of coordination. In Clause combining in grammar and discourse, ed. by John Haiman and Sandra Thompson, 331–359. Philadelphia: John Benjanmins. Peyraube, Alain. 2011. Les langues sinitiques. In Dictionnaire des langues, ed. by Emilio Bonvini, Joëlle Busuttil, and Alain Peyraube, 973–978. Paris: PUF. Peyraube, Alain 貝羅貝, and Jang-Ling Lin 林征玲. 2010. Lǎoqǐdà de gètǐ liàngcí hé yǔyán xúnhuán xiànxiàng zhī guānxī 老乞大的個體量詞和語言循環現象之關係 [Classifiers in the Laoqida and the hypothesis of cyclic change]. Hànyǔshǐ Xuébào 漢 語史學報 10:1–9. Takashima, Kenichi 高島謙一. 1989. Jiǎgǔwén zhōng de bìnglián míngcí lèyǔ 甲骨文中 的並聯名詞仂語 [The nominal conjunctions in oracle bone scripts]. Gǔwénzì Yánjiū 古 文字研究 17:338–356. Traugott, Elisabeth Cross, and Bernd Heine. 1991. Approaches to grammaticalization: Focus on theoretical and methodological issues. Philadelphia: John Benjamins. van der Loon, Piet. 1992. The classical theatre and art song of South Fukien. Taipei: SMC Publishing. Wu, Shou-Li 吳守禮. 2001a. Míng Jiājìng kān Lì Jìng Jì xìwén jiàolǐ 明嘉靖刊荔鏡記 戲文校理 [Annotated texts of the Romance of Li Jing Ji of Ming Jiajing edition (JJ)]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2001b. Míng Wànlì kān Lì Zhī Jì xìwén jiàolǐ 明萬曆刊荔枝記戲文 校理 [Annotated texts of the Romance of Li Zhi Ji of Ming Wanli edition (WL)]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2001c. Qīng Shùnzhì kān Lì Zhī Jì xìwén jiàolǐ 清順治刊荔枝記 戲文校理 [Annotated texts of the Romance of Li Zhi Ji of Qing Shunzhi edition (SZ)]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2001d. Qīng Guāngxù kān Lì Zhī Jì xìwén jiàolǐ 清光緒刊荔枝記戲 文校理 [Annotated texts of the Romance of Li Zhi Ji of Ming Guangxu edition]. Taipei: Congyi Workshop.

124  Jang-Ling Lin Wu, Shou-Li 吳守禮. 2002a. Míng Wànlì kān Jīn Huā Nǚ xìwén jiàolǐ 明萬曆刊金花 女戲文校理 [Annotated texts of Jin Hua Nü of Ming Wanli edition]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2002b. Míng Wànlì kān Sū Liù Niáng xìwén jiàolǐ 明萬曆刊蘇六 娘戲文校理 [Annotated texts of Su Liu Niang of Min Wanli edition]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2003. Qīng Qiánlóng kān Tóngchuāng Qín Shū Jì xìwén jiàolǐ 清乾 隆刊同窗琴書記戲文校理 [Annotated texts of the Romance of Tongchuang Qin Shu Ji of Qing Qianlong edition]. Taipei: Congyi Workshop. Yang, Hsiu-Fang 楊秀芳. 1991. Táiwān Mǐnnányǔ yǔfǎ gǎo 臺灣閩南語語法稿 [Taiwanese Southern Min grammar]. Taipei: Da-An Publishing. Yao, Xiaosui 姚孝遂, and Ding Xiao 肖丁, eds. 1989. Yīnxū Jiǎgǔ kēcí lèizuǎn 殷墟甲骨 刻辭類纂 [Concordance of Yin inscribed oracle bones]. Beijing: Zhonghua Bookstore. Zhu, Dexi 朱德熙. 1982. Yǔfǎ jiǎngyì [Handout of Chinese grammar]. Beijing: shangwu yingshuguan.

6 Exploration of the benefactive marker kang7 共 in Ming Qing Southern Min script Chian-Tang Su 蘇建唐

1 INTRODUCTION In the light of frame semantic (Fillmore 1982; Fillmore and Atkins 1992) and construction grammar (Fillmore, Kay, and O’Connor 1988; Goldberg 1995, 2006), this paper aims to explore the interaction of syntax and semantics of benefactive marker kang7 共 in Ming Qing Southern Min based on Li Jing Ji, compiled by Wu Shou-Li (abbreviated as LJJ). This script is one of the earliest data written in Southern Min at present. In my finding, the mood of these sentences could be divided into two types, imperatives and indicatives and their effects are shown in three respects, Person of the benefactee, the constraint on the movement of causee in construction B (the construction with benefactive marker kang7 共 could be divided into three types, construction A, B, and C), and the percentage of imperatives in different constructions. Secondly, there are cases of ‘mismatch’ (Francis and Michaelis 2000) in the affective construction with kang7 共. Such a mismatch could result from the effect of the feature of ‘inalienable’, and it will be discussed in the following sections. Many scholars (Cheng and Tsao 1995; Hung 1995; Tsao 2003; Chen 2006; Yang 2006, 2011; Hsu 2008; Chen and Lien 2008; Lien 2002, 2009) have studied kang7 共. Most of them mainly discussed modern Southern Min and offered many insightful proposals. Previous research (Cheng and Tsao 1995; Hung 1995; Tsao 2003; Chen 2006; Lien 2002) points out that kang7 共 is a multi-functional preposition. With different types of predicate, kang7 共 could be divided into five types of functions: commutative (reciprocal verb), source (verb of conveying), goal (verb of communication), benefactive (discussed later), and target1 (transitive verb). Generally speaking, benefactive marker refers to an entity benefiting from some action (Gove 1966:203; Radford 1988:373; Haegeman 1994:50). For example: (1) 卓 兄 共 我 顧 too2 hiann1 kang7 gua2 koo3 Zhuo brother kang 1sg look.after ‘Mr. Zhuo, look after these goods for me’.

貨 her3 goods

(4.562 SZ)

126  Chian-Tang Su In fact, both Benefactee and Malefactee are the Affectee which refers to the one affected by the whole event, and they form optional elements beyond the argument structure, i.e. gua2 我. The NP introduced by kang7 共 is an extra-argument (Hole 2005; Tsai 2005, 2007). Since the malefactive function was not found in LJJ, I will focus on the benefactive function in the following section. Differing from the studies above, Chen and Lien (2008) and Lien (2009) have dealt with the historical data and shed new insights. In these two papers, the functions of kang7 共 could be divided into two different interactional ways: reciprocal2 (commitative) and non-reciprocal (including source, goal, benefactive, and target). Lien (2009) discussed the competition between kang7 共 and kah4 合 from a broader perspective. Unlike the kang7 共 in LJJ, the benefactive marker is ka7 in present-day Taiwan Southern Min (abbreviated as TSM) instead. The commutative function of kang7 in TSM may well be pronounced as kah4. Cheng and Tsao (1995) think that both ka7 and kah4 are 合 and propose a route of diachronic development like this: ‘kap4◊ kah4◊ ka7’. However, Lien (2015) proposes that the development should have something to do with phonological development and lexical competition. It boils down to two points: first, ka7 should have developed from kang7 共 after the coda /-ng/ was lost. Second, kah4 should have developed from kap4 through the glottalization of the stop coda. The competition between kang7 共 and kap4 合 is shown in Figure 6.1 below: Lien’s proposal could be proved with these fieldwork data3 shown in Table 6.1 (‘Reciprocal’ is abbreviated as ‘recipro’). These data show that the original pronunciation of kang7 共 is still found in the dialect spoken in Quanzhou.

共

Non-reciprocal Conjunction Conj n unction 共

Non-reciprocal

Reciprocal

Conj n unction Conjunction

合

Reciprocal 合

Figure 6.1  Competition between kang7 共 and kah4 合

Table 6.1  The pronunciation of kang7 共 and kah4 合 Quanzhou [＋Recipr.] kang7 共

Malacca [－Recipr.] kang7

[＋Recipr.] ka7

Taiwan, Kinmen [＋Recipr.] ka7

[－Recipr.] kah4 合

[＋Recipr.] ka7 共

The benefactive marker kang7 共  127 Notwithstanding the achievements of previous studies, there are still some issues awaiting further investigations. First, diachronic data were rarely discussed in previous papers, so they would be explored in this paper. Second, Lien (2009) mainly analyzes kang7 共 in LJJ globally, but he did not focus on individual functions. In my finding, the syntactic features of the benefactive marker kang7 共 are quite robust. So it is worth a detail exploration. Besides these studies above, multi-functional words such as kang7 共 are often described in terms of a number of grammatical categories in traditional dictionaries, such as preposition and conjunction. But the meaning of every function and the relation of these meanings could not be delimited in such a traditional framework. In terms of cognitive semantics, a word should be analyzed not only by syntactic categories, but also by the conceptual frame and schemas based on our experience, belief, practice, and so on. It is crucial, however, that multi-functional words can only be fully understood with consideration of their syntactic position in a construction. From the viewpoint of construction grammar, constructions are also one kind of the vehicle for semantics (e.g. Goldberg 1995, 2006). In summary, grammatical categories, conceptual representations, and constructions should all be considered. The explorations of many facets of kang7 共 will be based on a set of the earliest Southern Min playscripts LJJ or Li Zhi Ji including four editions: Jiajing (JJ), Wanli (WL), Shunzhi (SZ), and Guangxu (GX). Apart from Introduction and Closing Words, the main structure of this paper is organized as follows: (2) ­Theoretical introduction, (3) The conceptual frame and construction of ‘kang7’, (4) Comparison of the characteristics of constructions A, B, C, and D. 2  FRAME SEMANTICS AND CONSTRUCTION GRAMMAR ‘Frame semantics’ proposed by Fillmore (1982) is a model of the semantics of understanding. In such theories the meaning of a word can be understood only with reference to a structured background of experience, beliefs, or practices, constituting a kind of conceptual prerequisite for understanding the meaning (Fillmore and Atkins 1992). Word meanings can be teased out with reference to the background structure of the world including our experience, belief, imagination practice, and so on (Lien 2000). As the definition in Taylor (1995:87), a frame refers to ‘the knowledge network linking the multiple domains associated with a given linguistic form’. A conceptual frame consists of a group of related schemas, and its independent level is motivated to establish a link among multiple senses covered by a lexeme and considerably reduce the overloading of lexical meanings (Lien 2000). Some elements of a frame could be profiled to syntactic representation, and a frame will offer background for understanding these profiled elements. For example, ‘David bought an old shirt from John for ten pounds’. In this sentence, David (the buyer), ten pound (the price), and shirt (the goods) are profiled, but John (the seller) is not. Presupposition refers to these ideas without being profiled. In a good example in Fillmore 1982, ‘chase’ means that two guys are moving in the same course, and the movement of the one in front is presupposition.

128  Chian-Tang Su Besides frame semantics, construction grammar proposed by Goldberg (1995) is also one of cognitive linguistic theories. Different from the thought that there can be no ‘mixed’ representation (Jackendoff 1994:8–9), the approach in question supports the views4 summarized by Goldberg (1996) as shown below. 1 2

3

4

Semantics is based on speakers’ construals of situation, not on objective truth conditions. (Langacker 1985, 1987, 1988; Fauconnier 1985; Lakoff 1987; Talmy 1978, 1985) Semantics and pragmatics form a continuum, and both play a role in linguistic meaning. Linguistic meaning is part of our overall conceptual system and not a separate modular component. Semantics is associated directly with surface form. (Talmy 1978, 1985; Haiman 1980; Lakoff 1987; Langacker 1987) Grammatical constructions, like traditional lexical item, are pairings of form and meaning. They are taken to have a real cognitive status, and are epiphenomena based on the operation of generative rules or universal principles. (Fillmore, Kay, and O’Connor 1988; Lakoff 1987; Wierzbicka 1988; Goldberg 1995) Grammar consists of a structured inventory of form-meaning pairing: phrasal grammatical constructions and lexical items. (Fillmore and Kay 1993; Lakoff 1987; Langacker 1987; Wierzbicka 1988; Goldberg 1995)

According to these views shown above, construction is a pairing of form and meaning, or pairing of form and functions (Goldberg 1995:4). Later, Goldberg (2006:5) takes frequency into account, and proposes the definition below: Any linguistic pattern is recognized as a construction as long as some aspect of its form or function is not strictly predictable from its component parts or from other constructions recognized to exist. In addition, patterns are stored as constructions even if they are fully predictable as long as they occur with sufficient frequency (Goldberg 2006:5) In summary, in terms of frame semantics, understanding the semantics should make use of the conceptual frames and schemas. Different presuppositions could profile as different syntactic representations, which also reveals the change of focus. In fact, these profiled representations could be regarded as one kind of pairing of form and meaning. On the other words, an underspecified word could be put into a profiled construction to be defined. Kang7 共 will be discussed in the light of such idea in the following sections. 3  THE CONCEPTUAL FRAME AND CONSTRUCTION OF KANG7 共 According to the definition of benefactive, ‘an entity benefiting from some action’, the core meaning of benefactive could be set as ‘someone does some acts for another one’s benefit’. Secondly, from the script in question, three types of benefactive manners could be extracted as shown below:

The benefactive marker kang7 共 129

Type 1: The agent benefits the Affectee by doing some act (2) 阮 是 潘 秀才， 來 共 三舍 送 路 (03.010 WL) gun2 si7 phuann1 siu2tsai5 lai5 kang7 san1sia7 sang2 loo7 1sg cop Pan scholar come kang San.Sia give road ‘I am Scholar Pan. I come here to give a farewell dinner in San Sia’s honor’.

(3) 卓 兄 共 我 顧 too2 hiann1 kang7 gua2 koo3 Zhuo brother kang 1sg look.after ‘Mr. Zhuo, look after these goods for me’.

貨。 her3 goods

(SZ 4.562)

Type 2: agent orders the cause to do some act for the benefit of the Affectee (4) 牌頭 哥， 煩動 你 共 阮 叫 pai5thau5 hiann1 huan5tang7 lir2 kang7 gun2 kio3 jailer bother move 2sg kang 1sg caus 出 來。 (40.085 WL) tshut4 lai5 out come ‘Jailer, please call my master to come out here for me’.

阮 官人 gun2 kuan1lin5 1sg master

Type 3: agent does a preparatory act and then an act for the Affectee’s benefit (5) 聊 備 盃 酒 共 恁 起離     (WL 41.011) liau5 pi7 pue1 tsiu2 kang7 lin2 khi2li5 casually prepare cup wines kang 2sg give.a.farewell.dinner ‘I might just as well prepare cups of wine to bid you farewell’. (6) 柴



梳

借

一

個

來

共

官人

tsha5 sue1 tsio7 tsit8 e5 lai5 kang7 kuan1lin5 firewood comb borrow one clf come kang master ‘Fetch a wooden-comb to comb my lord’s hair’.

梳 頭。 (27.145 SZ) sue1 thau5 comb head

Moreover, from the five sentences above, three constructions could be extracted: A) X+ kang7 共 NP2+VP(Vi/Vt+NP3)+(Y), B) X+ kang7 共+NP2+ V1+ (NP3)+ VP2, and C) X+VP1+kang7 共 NP2+VP2(Vi/Vt+NP3). In terms of the idea of frame, a conceptual frame consists of a group of related schemas. I will regard this core meaning of benefactive as the conceptual frame. The ‘benefactive manner’ is viewed as a presupposition and used to divide kang7

130  Chian-Tang Su Table 6.2  The conceptual frame, schemas, and constructions of benefactive marker kang7 共 Conceptual frame Schema A Presupposition Construction Schema B Presupposition Construction Schema C Presupposition

Someone does some acts for another one’s benefit. The Agent benefits the Affectee by doing some act. X+ kang7 共 NP2+VP(Vi/Vt+NP3)+Y The Agent orders the cause to do some act for the benefit of the Affectee. X+ kang7 共 NP2+V1 +NP3+VP2 The Agent does a preparatory act and then an act for the Affectee’s benefit. X+VP1+ kang7 共 NP2+VP2(Vi/Vt+NP3)

Construction

共 into three types of schemas, schema A, B, and C. Furthermore, constructions are also one kind of semantic vehicle. The constructions A, B, and C shown above could be regarded as the profiles of these three schemas above. Table 6.2 is the summary of the conceptual frame, schemas, and constructions of the benefactive marker kang7 共. Based on the summary shown in Table 6.2 above, the characteristics of the constructions of these three schemas will be explored in section 3.1–3.3. The comprehensive comparison will be shown in section 3.4. In my finding, the benefactee (NP2) in the kang7 共 phrase could be divided into speaker (including singular, plural lst Person, or speaker’s name) and non-speaker. Besides, these two kinds of benefactees could influent the interpretation of the targeted sentences, i.e., imperative or indicative moods.

3.1  The characteristics of construction A In construction A, only one verb could be found, and this construction could contain the presupposition of schema A: the Agent benefits the Affectee by doing some act. For example: (7)

(8)

阮

是

潘

秀才， 來

共

送 路 (03.010 WL) gun2 si7 phuann1 siu2tsai5 lai5 kang7 san1sia7 sang2 loo7 1pl.excl cop Pan scholar come kang San.Sia give road ‘I am Scholar Pan. I come here to give a farewell dinner in San Sia’s honor’. 卓 兄 共 我 顧 貨。 too2 hiann1 kang7 gua2 koo3 her3 Zhuo brother kang 1sg look.after goods ‘Mr. Zhuo, watch over these goods for me’.

三舍

(4.562 SZ)

The benefactive marker kang7 共 131 (9)

喬 共 我 扛 伊 kio7 kang7 gua2 kng1 i7 sedanchair kang 1sg carry 3sg ‘Carry the sedan chair carefully for me’.

(10) 鏡 擔 共 我 挑 kiann7 tann7 kang7 gua2 tann1 mirror load kang 1sg carry ‘Carry the mirror load for me’. (11)

你 lir2 you 共

今 kin1 now

將 tsiong1 caus

心 sim1 heart

腹 pak4 stomach

話 ue7 word

好 ho2 good

入 lip4 enter

寫 sia2 write

(32.085 GX)

去 khir3 go

一 tsit8 one

(16.403 GX)

封 hong1 envelop

書， si1 letter

我 gua2 1sg

你 送 去 度 阮 亞 娘 (14.291 SZ) lir2 sang3 khir3 thoo7 gun2 a1 niu5 kang 2sg send go give 1sg dim madam ‘You write down your innermost thoughts and feelings in the letter. I’ll deliver it to my madam for you’. kang7



(12) 我

共 你 𤆬 一 个 泉州 仔 gua2 kang7 lir2 tshua7 tsit8 e5 tsuan5tsiu1 a7 1sg kang 2sg marry one clf Tsuantsiu dim ‘I’ll help find a guy from Tsuantsiu to marry you’.

(13) 你 lir2 2sg 共 kang7

度 你 (9.733 SZ) thoo7 lir2 give 2sg

只 來， 我 一 錢 銀 tsi2 lai5 gua2 tsit8 tsinn5 girn5 here come 1sg one money silver 系 線 (19.177 WL) 我 買 gua2 bue2 si1 suann7 kang 1sg buy silk yarn ‘Come here. I give you one tael to buy the yarn for me’.

乞 khit4 give

你， lir2 2sg

The semantic and syntactic properties of the examples shown above can be captured by Table 6.3 featuring Construction A. As shown in Table 6.3, construction A emphasizes that NP2 will get the benefit after the agent complete VP2, but the event does not involve Agent’s intention. The concepts of construction A and schema A correspond with each other. This construction is the most typical one of kang7 and the percentage is up to 68.9% (177/257). Structures 1–4 share the same syntactic construction, but the elements of the predicate show great diversity and could be divided into four types of categories in terms of different kinds of action. In structure 1, the verb has to be intransitive, whereas those in structures 2–4 are transitive. Furthermore, the differences among structure 2–4 lie in the categories

132  Chian-Tang Su Table 6.3  The characteristics of construction A Construction A

X+共 NP2+VP(Vi/Vt+NP3)+Y

Meaning Features of elements

NP2 will get the benefit after agent did VP X

共 Phrase

Predicate

Lower semantics 1. Do some action

Agent

Y

Theme

＊

benefactee Intransitive verb

Agent

2. Arrange NP3

NP3

Verb of Theme arranging (based benefactee Verb of taking generated) care Theme (topicalized)

ex (1) (2) (6)

ProN. Compl.

(7)

＊

3. Change the location of NP3

Agent

benefactee

Verb of conveying

4. Make NP2 get NP3

Agent

benefactee

Verb of creation Theme Verb of obtain

(Theme)

(PP) (PP)

(8) (9)– (10)

of verbs. In structure 2, they are two-place verbs. In structure 3–4, both are three-place verbs, but the direction is reversed. It is worth having a look at ‘arrange NP3’ in structure 2. When NP3 is in situ (basegenerated), the position Y could be filled by a ‘retained object’ pronoun and couldn’t be filled by other complements (e.g. 10). This situation only occurs in Quanzhou Southern Min, but in TSM it should be 予伊, and the pronoun 伊 is often omitted. However, when NP3 is topicalized to the position X and leaves the slot occupied by a ‘trace’. No pronoun could be filled in the slot. Both topicalization and omission of the theme (e.g. 12) are found in modern Southern Min. Why these two situations are not found in other structures could be related to the effect of text choice. It should be mentioned in passing that I have not tapped other data, such as Jin Hua Nu and Doctrina Christina, written or translated in Southern Min during the same period. Usually the agent in the subject position exhibits the feature [+Human]. However, there are also cases where the object NP could be treated as a topicalized theme, as in (14). (14) 牛 共 我 牽 好 gu5 kang7 gua2 khan1 ho2 ox kang 1sg care good ‘Take care of the ox stably for me’. However, NP2 is [±Animate]. It might be correlated with that there is no restriction on the disposed target. Finally, both the Affectee in this script is [+Human], but it could be [－Human, +Animal]. This situation would result from the generalization of kang7 共. Generalization refers to a semantic change expanding the kinds of referents or distribution of a word (Hopper and Traugott 2003).

The benefactive marker kang7 共 133

3.2  The characteristics of construction B In construction B, two predicates could be found and such situation also reflects the presupposition of schema B, Agent should cooperate with NP3 to complete the benefit for the Affectee. For example: (15) 牌頭哥， pai5thau5hiann1 jailer.bother



煩動 huan5tang7 move

你 lir2 2sg

共 kang7 kang

伊 都牢丈 共 阮 i1 too1loo5tiunn7 kang7 gun2 caus 3sg jailer kang 1pl.excl ‘Ask the jailer to open the door for us’.

(17) 許 hi2

叫 kio3 caus

阮 gun2 1sg

官人 出 來。 (40.085 WL) kuan1lin5 tshut4 lai5 master out come ‘Jailer, please call my master to come out here for us’.

(16) 叫 kio3

阮 gun2 1sg

犯罪人 共 我 叫 huan7tsue7lin5 kang7 gua2 kio3 dem prisoner kang 1sg caus ‘Call the prisoner to come out for me’.

(18) 我 共 恁 叫 出 來 gua2 kang7 lin2 kio3 tshut4 lai5 1sg kang 2sg caus out come ‘I’ll ask him to come out here for you’.

開 門 (27.086 SZ) khui1 merng5 open door

來 lai5 come

(36.124G X)

(36.084 GX)

The semantic and syntactic properties of the above examples can be encapsulated in Table 6.4 showing the characteristics of construction B. In Table 6.4, the presupposition of schema B is that NP2 will get the benefit after the agent order NP3 to do some act. This presupposition is realized as a pivotal construction carrying ‘causative’ meanings. The causative meaning and imperatives correspond with each other, so the percentage of imperatives in this construction is the highest in these three constructions in question. Differing from the constructions of A and C, the main point of construction B is that after the VP1 and VP2 are completed by NP1 and NP3, NP2 will benefit from these two actions. Secondly, the mood of this construction could be divided into two types: imperatives and indicatives, and this situation could also bring some effects to the construction in question. One is the connection between the Person feature of Affectee (NP2) and mood. If the Affectee is 1st person (include singular and plural), this sentence will be easier to be read as an imperative than the non-1st

134  Chian-Tang Su Table 6.4  The characteristics of construction B Construction B Meaning Features of Elements

X+共 NP2+V1 +NP3+VP2 NP2 will get the benefit after the agent makes NP3 do some action Mood

X

共Phrase

Imperative Agent 1st. Indicative

NP3

V1

Causative Causee verb

Agent Non-1st.

(Causee)

VP2

ex:

Transitive verb Directional (3)–(12) verb Directional verb

(14)

person ones. Because the Affectee is ‘the dominator’ in imperatives, but the agent is the ‘dominatee’. Generally the only one which could assign the requirement for Affectee is himself. In fact, this situation could also be found in construction A. In addition, although X could also be the agent or topicalized causee, it should be [+Human] in LJJ. Because the person who could give an order to the causee should not be an element with [－Human]. If not, it could not correspond with our experience. Moreover, ‘叫伊都牢丈’ in (16) should be regarded as a topicalized ‘V1＋NP3’. For example: (19) a 叫 kio3

伊 都牢丈 共 阮 i1 too1loo5tiunn7 kang7 gun2 caus 3sg jailer kang 1pl.excl ‘Ask the jailer to open the door for us’.

b [叫 kio3



開 門 (27.086 SZ) khui1 merng5 open door

伊 都牢丈]i 共 阮ti i7 too1loo5tiunn7 kang7 gun2 caus 3sg jailer kang 1pl.excl ‘Ask the jailer to open the door for us’.

開 khui1 open

門 merng5 door

Finally, only directional verbs found in the VP2 of indicative sentences might also be the effect of text choice as shown in construction A.

3.3  The characteristics of construction C The presupposition of schema C, meaning that the Agent does some act for the Affectee’s benefit, is also profiled as a construction with two predicates. For example: (20) 聊 備 盃 酒 共 恁 起離   (41.011 WL) liau5 pi7 pue1 tsiu2 kang7 lin2 khi2li5 casually prepare cup wines kang 2pl give.a.farewell.dinner ‘I might just as well prepare cups of wine to bid you farewell’.

The benefactive marker kang7 共 135 (21) 柴



借

一

個

來

共

官人

tsha5 sue1 tsio7 tsit8 e5 lai5 kang7 firewood comb borrow one clf come kang ‘Fetch a wooden-comb to comb my lord’s hair’.

(22) 伊



梳

即

甘心

捧

盆

水

共

kuan1lin5 master

梳 頭 (27.145 SZ) sue1 thau5 comb head

恁

掃 廳堂 (23.098 GX) i1 tsiah4 kam1sim1 phang5 phun5 tsui2 kang7 lin2 sau3 thiann1tong5 3sg just willingly hold basin water kang 2pl sweep hall ‘He just willingly holds a basin of water and sweeps the hall for you’.

The semantic and syntactic properties of the above examples can be encompassed in Table 6.6. showing the characteristics of construction C. As shown in Table 6.5, the meaning of construction C is that ‘the agent carries out VP1 in order to do VP2 for NP2’. The meaning corresponds with the presupposition of schema C shown above. Although both predicates are found in constructions B and C, the difference between constructions B and C could be the agent of VP2 shown in Figure 6.2 and Figure 6.3:

Table 6.5  The characteristics of construction C Construction C

X+VP1+共 NP2+VP2(Vi/Vt+NP3)

meaning Semantic type

The agent carries out VP1 for doingVP2 to NP2. X

共Phrase

VP1

V2

1 For doing some acts for NP3

Agent benefactee Acts done for Intransitive doing VP2 verb

2 Take care of NP2’s NP3

Agent benefactee Acts done for Verb of doing VP2 Active verb

VP1

VP2

NP1

NP3

Figure 6.2  The concept of construction B

ex (15)

Theme (4)-(16)

NP2 Benefit

+

NP3

136  Chian-Tang Su VP1+VP2 NP1

NP2 Benefit

Figure 6.3  The concept of construction C

It is noteworthy that the purpose for the agent to do VP1 is to do VP2 as shown in Figure 6.3, so the intention of the agent will be emphasized in construction C. However, this situation and the imperative are in conflict because the agent’s intention in imperatives will be shaded. This conflict leads to the result that the percentage of imperatives in this construction is lowest between these three constructions in question. Moreover, I find that NP2 and NP3 are often in a possessive relation, (a situation that could also be found in construction A). The possessive relation could be also divided into different sub-relations: one is the ‘owner and body-part’, e.g. (23) and the other is the ‘owner and belonging’, e.g. (25). Jackendoff (1987) argues that a single syntactic NP may have multiple thematic roles.5 According to the ‘theta criterion’, normally the relationship between thematic roles and NPs is one to one. If not, it is one kind of ‘mismatch’ (Francis and Michaelis 2000). Such phenomena could also be found after the ellipsis of NP3 here as shown below: (23) Owner and body-part  a 柴梳借一個來共官人Affectee梳頭Theme  b 柴梳借一個來共官人Affectee/Theme.梳 ei ?c 柴梳借一個 ei 來梳頭*Affectee/Theme.

(27.145 SZ)

(24) a 你 毋 通 共 伊Affectee 拍 頭Theme lir2 m7 thang1 kang7 i1 phah4 thau5 2sg neg could kang 3sg hit head ‘You could not hit his head’.  b 你毋通共伊Affectee/Theme.拍ei ?c 你毋通拍頭*Affectee/Theme(Self-made) (25) Owner and belonging  a 伊即甘心捧盆水共恁Affectee掃廳堂Theme. ?b 伊即甘心捧盆水共恁Affectee/*Theme.掃 ei 。 ?c 伊即甘心捧盆水 ei掃廳堂*Affectee/Theme.

(23.098 GX)

The benefactive marker kang7 共  137 In the ‘owner and body-part’ relation, besides ‘Affectee’, NP2 could get another theta role ‘Patient’ after the ellipsis of NP3, e.g. (23b) and (24b),6 and it corresponds with what Jackendoff (1987) argues above. However, such a situation was found in the ‘Owner and belonging’ e.g. (25b). I would like to bring your attention to the possibility that the asymmetry shown above bears on the issue of ‘inalienability’. The NP2 and NP3 are ‘inalienable’ in the ‘owner and body-part’ relation, so NP2 is not only the Affectee, but also the logical Recipient of the act. Therefore, while NP3 could be shaded without reducing the semantics, NP2 should be highlighted in syntactic representation. This argument could be proved in (4c) and (17c): the ellipsis of the real Recipient NP2 will result in the other reading: ‘the agent does VP2 for himself’. Conversely, NP2 and NP3 are ‘alienable’ in the ‘owner and belonging’ relation. NP2 is only the Affectee, and the real recipient is NP3. Both these two elements should be projected to the surface syntactic structure together. No matter which one is omitted, it will lead to reducing the semantics of (16a). This phenomenon will be shown in Figure 6.4 below: Owner and body-part

[[NP2Affectee/Target]

AFFECT

CAUSE

(extra-argument)

(internal-argument)

Owner and belonging [[NP2Affectee] (extra-argument)

[NP1Agt. DO (NP3 ti)]EVENT]

Real recipient AFFECT

CAUSE

[NP1Agt. DO *(NP3Target)]EVENT] (internal-argument)

Figure 6.4  Different possessive relations at mismatch’s comparison

4 COMPARISON OF THE CHARACTERISTICS OF CONSTRUCTIONS A, B, C, AND D Table 6.6 shows the comparison and contrast between characteristics of constructions A, B, and C. Attention should be paid to the four main points in Table 6.6. The first point is that besides the conceptual frame, the three constructions in question share one similar feature: NP in kang7 共+NP should be [+Human], i.e. sentient beings. This situation corresponds with the concept of benefit because only NP with [+Human] could be emotionally influenced. Notice that in LJJ, NP with [－Human, +Animate], which is productive in contemporary TSM, is not found in the subject

138  Chian-Tang Su Table 6.6  Comparison of the characteristics of the constructions of A, B, and C Construction Conceptual frame

共 phrase

Meaning

A

B

C

‘Someone does something for another one’s benefit’ NP2 will get the benefit after agent did VP2

NP2 will get the The agent carries benefit after the out VP1 for agent makes NP3 doingVP2 to NP2 did some action

Theta role

benefactee

benefactee

benefactee

Semantic features

[＋human]

[＋human]

[＋human]

Argument of Verb

－

－

－

1st

±

±

－

60.2%

84.4%

3.7%

Benefactive source

context

context

context

Features of construction

Imperative mood Features of verb

Pivotal construction Serial verb construction

Vt

V1

Vi

V2

Causative verb Vt Vi

V1

Vt Vt

V2

Vi

－

＋

－

－

－

＋

position. The reason is that it might be correlated with the effect of text choice or the ‘Generalization’ in grammaticalization. Secondly, there is a strong tendency about the co-occurrence of the Person of Affectees and the mood of sentences. If the Affectee is 1st person (including singular and plural), the percentage of imperative is up to 97.3% (74/76). The only two exceptions are affected by other functional words, e.g. volitional modal verb khun2 肯: (26) 汝 肯 共 我 送 去。 lir2 khun2 kang7 gua2 sang3 khir3 2sg would kang 1sg sent go ‘You are willing to deliver it for me’.

(8.086 GX)

This reason for the tendency in question would be correlated with our experience. The imperatives must reflect the speaker’s volition. However, the feelings usually differ from person to person. We never know whether someone will gain

The benefactive marker kang7 共 139 the profit from some acts, so no one could make any decision for the other one. These two ideas will crash in logical thinking. In fact, the percentage of imperatives varies from constructions to constructions. As shown in Table 6.6, the scale of the percentage of imperatives between these three constructions is B > A > C. I think that this situation should result from the interaction between construction C and imperatives on semantic logic. As discussed in section 3.2, construction B is a pivotal construction with [+cause]. Both the construction and imperatives involve order assignment, so they correspond with each other. On the contrary, the intention of the Agent would be emphasized in construction C. So the Agent in construction C is the ‘dominator’. However, in imperatives, the Agent is the ‘dominatee’. These two ideas would also crash in logical thinking. The same situation could also be found in (26). Moreover, ‘context’ is often the source of benefit. In fact, I  think that it could make use of Tsao’s (2003) and Hung’s (1995:134–138) discussion to prove the idea in question. Both these two papers followed Jackendoff’s (1987) view that theta relations could be more than in one tier, and explained the ambiguity of ‘kang7 共 phrase’ with the view. The key to the ambiguous ‘kang7 共 phrase’ in constructions A, B, and C is the effect of context. Particularly, they would change meanings after being extracted from the previous context. For example: (27) 伊 共 阿三 教 i1 kang7 a1sam1 ka3 3sg kang Asam teach Benefactee/Goal ‘He taught Asam a secret’.

一 tsit8 one

個 e5 clf

秘訣   (Hung 1995:136) pi3kuat4 secret.method

Finally, both transitive and intransitive verbs could be allowed in these three constructions (only VP2 in constructions B and C). The contexts are the source of benefactive semantics, so there is no constraint on the categories of verbs. However, the VP1s in constructions B and C are not in the same situation. Based on the meaning of pivotal constructions, ‘causation’, only causative verbs can be in the VP1 position in construction B. In construction C, V2 should adopt the entity introduced by V1 in serial verb construction; thus V1 should be a transitive verb. 5  CLOSING WORDS Based on LJJ, this paper examines the characteristics of the benefactive marker kang7 共 in Ming Qing Southern Min script in light of frame semantics and construction grammar. First, the conceptual frame of kang7 could be divided into

140  Chian-Tang Su three types of schemas with different presuppositions, and these schemas could profile as constructions A, B, and C: Table 6.7  The conceptual frame, schemas, and constructions of benefactive marker kang7 共 Conceptual frame Schema A Presupposition Construction Schema B Presupposition Construction Schema C Presupposition Construction

Someone does some acts for another one’s benefit. Agent benefits the Affectee by doing some act. X+ kang7 共 NP2+VP(Vi/Vt+NP3)+Y Agent should cooperate with NP3 to complete the benefit for the Affectee. X+ kang7 共 NP2+V1 +NP3+VP2 Agent does some act for doing another act to benefit the Affectee. X+VP1+ kang7 共 NP2+VP2(Vi/Vt+NP3)

Construction A is the most typical construction of the benefactive marker kang7 共, which only contains a verb. Although both constructions B and C include two verbs, the meanings of these two constructions are different. Construction B is a ‘pivotal construction’. The agents of VP1 and VP2 are not the same and it reveals that the benefit should be done by the Agent and the other people. Construction C is a ‘serial verb construction’. The agent of these two verbs is the same one and the intention of agent could be emphasized in this construction in question. Moreover, the mood of these sentences could be divided into two types, imperatives and indicatives and their effects are shown below. First, when the Affectee is 1st person, the sentences tend to be imperatives. Secondly, the causee with [+Human] in construction B could not be topicalized to X position, or it could yield semantic ambiguity. But such a situation is not found in imperatives. Third, the percentage of imperatives varies from constructions to constructions, and the lowest degree is in construction C. Both these three phenomenon are correlated with the semantics of imperatives. In fact, the possessive relation will influence the affective construction with kang7. When NP2 and NP3 are in the ‘owner and body-part’ relation, a ‘mismatch’ could be found between the thematic role and NP2. Because of the effect ‘inalienable’, NP2 is the real Recipient of the act referred by the verb in this relation. NP2 should be realized in the syntactic structure, but NP3 could be shaded without reducing the semantics. This argument could also be proved in the ‘owner and belonging’ relation. In the secondary relation, NP2 and NP3 are ‘alienable’, so both of them should be profiled to the syntactic structure. Finally, besides LJJ, there are other data, such as Jin Hua Nü and Doctrina Christina, written or translated in Southern Min during the same period and more interesting phenomena would be found therein. However, due to the space limitation, the exploration of these data has to be left to another occasion in the future.

The benefactive marker kang7 共 141 NOTES 1 Some scholars use Object (Chappell, Peyraube, and Wu 2011) or Patient to name this functional marker. Object could be mixed with the syntactic object. ‘Patient’ usually refers to the entity undergoing the action expressed by the predicate (Longacre 1983:155–156; Radford 1988:373; Haegeman 1994:49, etc.), but this term could not include the entity without undergoing the action. e.g. ‘you’ in ‘I love you’. I will use Target to refer to the recipient of the act. 2 In Chen and Lien 2008, the term they use is bidirectional and unidirectional. For avoiding the mistake, I use reciprocal and non-reciprocal in this paper. 3 In February and July 2011, I was sponsored by NSC to do fieldwork in Malacca, Jinmen, and Quanzhou. The research reported in this paper is a part of the result of the fieldwork undertaken. 4 There are seven points advanced in Goldberg (1996). Here we merely cite four of them which bear on the theme of this paper. 5 Jackendoff (1987) proposes that there are two thematic tiers: semantic tier and action tier. Tsao (1991) and Hung (1995) both use such an idea to explain the developmental relationship among the functions of kang7 共. However, I have no direct evidence to prove the development of benefactive function and target function of kang7 共 and it needs to be explored in an independent paper. 6 Kang7 共 in these two examples could not be read as a ‘target’ marker. NP2 is the only one complement of target construction, and no other NP could be reconstructed. Affective marker  (1) a 你毋通共伊拍腳腿。 ‘Do not hit his/her legs’.    b 你毋通共伊拍ei。 ‘Do not hit him/her’. Target marker (2) a 你毋通共伊拍。 ‘Do not hit him/her’.     *b 你毋通共伊拍腳腿。 ‘Do not hit him/her’.

REFERENCES Chappell, Hilary, Alain Peyraube, and Yunji Wu伍雲姬. 2011. A comitative source for object markers in Sinitic languages: 跟 kai55 in Waxiang and 共 kang7 in Southern Min. Journal of East Asian Linguistics 20:291–338. Chen, I-hsuan, and Chinfa Lien. 2008. The interaction between the construction kong + topic and its thematic markers in Taiwanese Southern Min. In Chinese linguistics in Leipzig, ed. by Redouane Djamouri, Barbara Meisterernst, and Rint Sybesma, 175–190. Paris: Centre de Recherches Linguistiques sure l’Asie Orientale, École des Hautes Édudes en Sciences Sociales. Chen, Zeping 陳澤平. 2006. Fúzhōu fāngyán chǔzhì jiècí ‘gòng’ de yǔfǎhuà lùjìng 福州方 言處置介詞“共”的語法化路徑 [The grammaticalization path of the disposal preposition ‘gong’ in the Fuzhou dialect]. Zhōngguó Yǔwén 中國語文 3:233–236. Cheng, Ying 鄭縈, and Feng-fu Tsao 曹逢甫. 1995. Mǐnányǔ ‘ka’ yòngfǎ zhījiān de guānxī 閩南語ka用法之間的關係 [The relationship between the usages of ka in Taiwanese Southern Min]. In Papers from the 1994 conference on language teaching and linguistics in Taiwan. Vol. 1, Southern Min, ed. by Feng-fu Tsao and Meihui Tsai, 23–45. Taipei: Crane Publishing. Fauconnier, Gilles. 1985. Mental spaces. Cambridge, MA: MIT Press. Fillmore, Charles J. 1982. Frame semantics. In Linguistics in the morning calm, ed. by the Linguistic Society of Korea, 111–138. Seoul: Hanshin. Fillmore, Charles J., and B.T.S. Atkins. 1992. Towards a frame-based lexicon: The semantics of risk and its neighbors. In Frames, fields and contrasts: New essays in semantic

142  Chian-Tang Su and lexical organization, ed. by Adrienne Lehrer and Eva Feder Kittay, 75–102. Hillsdale: Lawrence Erlbaum Associates. Fillmore, Charles J., and Paul Kay. 1993. Construction grammar coursebook. Berkeley: University of California. Fillmore, Charles J., Paul Kay, and Catherine O’Connor. 1988. Regularity and idiomaticity in grammatical constructions: The case of let alone. Language 64:501–528. Francis, Elaine J., and Laura A. Michaelis. 2000. Approaches to mismatch: Introduction. In Proceeding of the Berkeley formal grammar conference workshops, ed. by Mariam Butt and Tracy Holloway King. Stanford: CSLI Publications. Online Conference Proceeding. http:// web.stanford.edu/group/cslipublications/cslipublications/LFG/5/bfg00/bfg00-toc.html. Goldberg, Adele E. 1995. Constructions: A construction grammar approach to argument structure. Chicago, London: The University of Chicago Press. Goldberg, Adele E. 1996. Jackendoff and constructions-based grammar. Cognitive Linguistics 7:3–19. Goldberg, Adele E. 2006. Constructions at work: The nature of generalization in language. New York: Oxford University Press. Gove, Philip B., ed. 1966. Webster’s third new international dictionary of the English language. Chicago: Encyclopedia Britannica. Haegeman, Lilian. 1994. Introduction to government and binding theory. Cambridge: Blackwell. Haiman, John. 1980. Dictionaries and encyclopedias. Lingua 50:329–357. Hole, Daniel. 2005. Extra-argumentality ─ affectees, landmarks, and voice. Linguistics 44:383–424. Hopper, Paul J., and Elizabeth Closs Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press. Hsu, Fu-mei 徐富美. 2008. Táiwān Mǐnnányǔ ka hoo hòumiàn míngcí chéngfèn de yǐnxiàn yòngfǎ 台灣閩南語「ka」「hoo」後面名詞成分的隱現用法 [(Co)overt nouns occurring after ka and hoo in Taiwanese Southern Min]. Paper presented at the Seventh International Symposium on Taiwanese Language and Teaching, National Taiwan Normal University, 6–7 September. Hung, Suchen. 1995. A study of the Taiwanese preposition KA and its corresponding Mandarin prepositions. Master thesis, National Tsing Hua University, Taiwan. Jackendoff, Ray. 1987. The status of thematic relations in linguistic theory. Linguistic Inquiry 18:369–412. Jackendoff, Ray. 1994. Lexical insertion in a post-minimalist theory of grammar. Ms., Brandeis University. Lakoff, George. 1987. Women, fire, and dangerous things. Chicago: Chicago University Press. Langacker, Ronald W. 1985. Observations and speculations on subjectivity. In Iconicity in syntax: Proceedings of a symposium on iconicity in syntax, ed. by John Haiman, 109–150. Amsterdam: John Benjamins. Langacker, Ronald W. 1987. Foundations of cognitive grammar. Vol. 1, Theoretical prerequisites. Stanford: Stanford University Press. Langacker, Ronald W. 1988. An overview of cognitive grammar. In Topics in cognitive linguistics ed. by Btygida Rudzka-Ostyn, 127–161. Philadelphia: John Benjamins. Lien, Chinfa. 2000. A frame-based account of lexical polysemy in Taiwanese. Language and Linguistics 1:119–138. Lien, Chinfa. 2002. Grammatical function words 乞, 度, 共, 甲, 將 and 力 in Li4 Jing4 Ji4 and their Development in Southern Min. In Dialect variations in Chinese: Papers from the third international conference on Sinology: Linguistic section, ed. by Dah-an Ho, 179–216. Taipei: Institute of Linguistics Preparatory Office, Academia Sinica.

The benefactive marker kang7 共 143 Lien, Chinfa 連金發. 2009. Rìzhì shídài Táiwān Mǐnnányǔ cíhuì yǔyì yánjiū huígù 日治時 代台灣閩南語詞彙語義研究回顧 [Linguistic perspectives in Taiwan during the period of Japanese rule: A  glimpse at studies of lexical and semantic issues in Taiwanese]. Táiwān Yǔwén Yánjiū 臺灣語文研究 4:79–93. Lien, Chinfa. 2015. The condition and change of 共 vis-à-vis 合 in Southern Min with a sidelight on intra-dialectal variation. Journal of Chinese Linguistics 43:1–33. Longacre, Robert. 1983. The grammar of discourse. New York: Plenum. Radford, Andrew. 1988. Transformational grammar: A  first course. Cambridge: Cambridge University Press. Talmy, Leonard. 1978. The relation of grammar to cognition. In Theoretical issues in natural language processing-2, ed. by David L. Waltz, 14–24. Champaign, IL: Coordinated Science Laboratory, University of Illinois. Talmy, Leonard. 1985. Force dynamics in language and thought. In Papers from the parasession on causatives and agentivity at the 21st regional meeting, ed. by William Eilfort, Paul Kroeber, and Karen Peterson, 293–337. Chicago: Chicago Linguistics Society. Taylor, John R. 1995. Linguistic categorization: Prototypes in linguistic theory. Oxford: Oxford University Press. Tsai, Dylan Wei-Tien 蔡維天. 2005. Tán Hànyǔ de méngshòu jiégòu 談漢語的蒙受結 構 [Affective constructions in Mandarin Chinese]. Ms., National Tsing Hua University. Tsai, Dylan Wei-Tien. 2007. Ins and outs: Evidence from adverbials, applicatives, light verbs, and object fronting in Chinese. Paper presented at International Symposium of Linguistics Consortium, National Tsing Hua University, 14–17 December. Tsao, Feng-Fu 曹逢甫. 1991. On the mechanism and constraints in syntactic change: Evidence from Chinese dialects. In International symposium on Chinese languages and linguistics. Vol. 2, 370–388. Taipei: Academia Sinica. Tsao, Feng-Fu 曹逢甫. 2003. Táiwān Mǐnnányǔ de KA7 zìjù 台灣閩南語的Ka7字句 [Ka7 constructions in Taiwanese Southern Min]. Hànyǔ fāngyán yǔfǎ yánjiū hé tànsuǒshǒujiè guójì Hànyǔ fāngyán yǔfǎ xuéshù yántǎohuì lùnwénjí 漢語方言語法研究和探 索-首屆國際漢語方言語法學術研討會論文集 [The study and exploration of Chinese dialectal grammar: Proceedings of the 1st international symposium of Chinese dialectal grammar], ed. by Zhaoming Dai 戴昭銘, 114–136. Harbin: Heilongjiang People’s Publishing House. Wierzbicka, Anna. 1988. The semantics of grammar. Philadelphia: John Benjamins. Yang, Barry C-Y. 2006. Syntactic structure of ka-construction in Taiwanese Southern Min. UST Working Papers in Linguistics (USTWPL) 2:141–171. Yang, Barry C-Y. 2011. Revisiting the Ka-construction in Taiwan Southern Min. Paper presented at the 12th Conference on Min Languages. Institute of Linguistics, Academia Sinica, 5–7 November.

7 Taiwanese Southern Min hoo7 and its counterparts in the Southern Min varieties in Quemoy and Quanzhou Chia-yin Hu 胡佳音 1 INTRODUCTION The Min language has been one of the focal topics in the research on Sinitic languages. Both the synchronic dialectal diversity and its crucial role in the study of Chinese languages history have been drawing more and more scholarly attention in recent decades. The subgrouping of Min has been a disputed issue, but it is agreed that Southern Min is the subgroup with the most speakers and the widest geographical distribution: southern Fujian province, Taiwan, and other areas in the southern part of Mainland China and overseas with descendants of emigrants. Southern Min can be further sub-classified into a number of variants, including Quanzhou, Zhangzhou, and many others. Hybrids of these dialects are common and two well-known cases are Amoy and Taiwanese Southern Min, in which Quanzhou and Zhangzhou dialects are mixed. In the present study, based on two sets of fieldwork conducted in August 2010 and July 2011, a comparison between Quanzhou and Taiwanese Southern Min as well as the variety spoken in Quemoy is made, focusing on the uses of multifarious exponents: hoo7 予, thoo7 度, tng5 傳, and khit4 乞. The main findings from the field work include gradual linguistic variation exhibited in spatial distribution. Before going into the findings of the field works, the relevant literature is summarized in Section  2 as background knowledge. The reports on two fieldwork trips are introduced in Section 3; results and discussion are provided in Section 4. Finally, Section 5 concludes the study. 2  THE MULTIFUNCTIONAL MORPHEMES It has been well established that in Taiwanese Southern Min (henceforth TSM), the exponent hoo7 予 serves a number of distinct lexical and grammatical functions, including a ditransitive verb, a permissive, a stative causative verb, a dative marker, and an agent marker (Chen 1972; Huang 1977; Cheng 1974, 1998; Tsao 1988, 1997; Cheng et al. 1999; Chen and Lien 2008). From Table 7.1, it can be seen that while there are numerous ditransitive verbs in TSM, there is no other members in the causative and dative paradigms except hoo7 予. As for agent markers, in addition to hoo7 予 in isolation, the compound khit4hoo7 乞予 can also be used as an agent marker1. It is worthy to note that in contemporary TSM, the

Taiwanese Southern Min hoo7 145 Table 7.1 Paradigm of TSM ditransitive verbs, causative verbs/markers, dative markers, and agent markers Ditransitive verbs

Causative verbs/marker(s)2

Dative marker(s)

Agent marker(s)

hoo7 予 ‘give’ sang3 送 ‘send’ tsioh4 借 ‘lend’ . . .

hoo7 予 ----------------

hoo7 予 ----------------

hoo7 予 khit4hoo7 乞予 -----------

word khit4 乞 is mainly used in a verbal phrase composed with a transitive action verb ‘to beg’ immediately followed by an object, such as 乞雨 ‘beg for rain’, or in a common noun 乞食 ‘a beggar’. Its isolated use as an agent marker, on the other hand, is rarely seen (Lien 2008; Chen and Lien 2008:328), which is quite different from its counterparts in Quanzhou Southern Min as we will see later. The multifarious hoo7 予 can be found in a number of syntactic patterns, including ditransitive, dative, causative, and passive constructions. First, it can participate in two types of double object constructions, as pointed out in Cheng et al. 1999. In both types, however, hoo7 予 precedes the nominal indicating the Recipient and another nominal denoting the Theme in a transferring event (cf. Chen and Lien 2008). The distinction between these two types is that hoo7 予 is the main predicate in the first type (Double Object Construction-1), while in the second type (Double Object Construction-2), it immediately follows another ditransitive verb. (1) Double Object Construction-1: NP1+ hoo7+NP2+NP3



我 予 你 三 百 gua2 hoo7 li2 sann1 pah4 1sg hoo 2sg three hundred ‘I give you three hundred dollars’.

箍3 khoo1 dollar

(2) Double Object Construction-2: NP1+ V-hoo7+NP2+NP3



我 送 予 gua2 sang3 hoo7 1sg send hoo ‘I sent him a book’.

伊 i1 3sg

一 tsit8 one

本 pun2 clf

冊 tsheh4 book

Besides double object constructions, hoo7 予 also occurs in dative constructions, where similar to double object constructions, hoo7 予 also introduces the nominal with the thematic role Recipient. (3) Dative Construction: NP1+ V+ NP2 + hoo7 + NP1



我 還 三 百 箍 予 gua2 hing5 sann1 pah4 khoo1 hoo7 1sg return three hundred dollar hoo ‘I returned three hundred dollars to him’.

伊 i1 3sg

146  Chia-yin Hu Moreover, hoo7 予 can be found in one certain typologically-unique syntactic structure. As can be seen from the following example cited from Cheng et al. 1999, this pattern is similar to the dative construction in that hoo7 予 is preceded by the Theme and followed by the Recipient. The sole distinction between these two constructions lies in the verb after the Recipient NP: (4) NP1+ V+(NP2)+ hoo7+NP1+V



我 唱 一 首 歌 kua1 gua2 tshiunn3 tsit8 siu2 1sg sing one piece song ‘I sang a song for you to listen to’.

予 hoo7 hoo

你 li2 2sg

聽 thiann1 listen.to

Researchers from various points of view have given different labels to this pattern: ‘special type of pivotal constructions’ in Tsao 1997:104–106, ‘serial verb construction’ in Cheng et al. 1999, and ‘extended ditransitive construction’ in Lien 2005:708. All these names indicate two important characteristics of this construction. First, this is a serial verb construction in its essence, comprised of more than two verbs in the clause. The term ‘serial verb construction’ may be a little too broad, however; hence it can also be regarded as a pivotal construction. It is noteworthy, as explicitly pointed out by Tsao (1997:104–106), although similar to canonical pivotal constructions, the pattern also features the pivotal nominal between two clauses, significant distinction must remain, however, that unlike common pivotal constructions compatible with verbs without specific collocational restrictions, the targeted construction is closely related to ditransitive and dative constructions and therefore is called ‘special type of pivotal constructions’ in Tsao (ibid.). In other words, it is semantically similar to ditransitive and dative constructions in that they all involve an outward transferring event. Unlike these constructions, however, the construction is a conflation of a preparatory action and a consequence purpose action linked by hoo7 予 (Lien 2005), i.e. the construction denotes an outward transferring event which is to bring about another purposive event. This view is shared by Wang (2008) in suggesting hoo7 予 as a marker in a purpose clause (also see the discussion on the Chinese gěi ‘give’ in Ting and Chang 2004:54; Paul 1988, whose analyses can also be applied to the targeted TSM construction). The close link between this construction and ditransitive constructions motivates the name of ‘extended ditransitive construction’ (Lien 2005:708), which will be adopted in the present study. The extended ditransitive construction is actually related to the next pattern to be introduced presently: the analytic causative construction, where it functions as a permissive or stative marker (Cheng 1974; Cheng et al. 1999; Lien 2008). Considering the syntactic position of hoo7 予 in these two constructions, one would see their similarity in that they connects two clauses which form one complex sentence. Semantically speaking, as is mentioned in the last paragraph that hoo7 予 in the extended ditransitive construction has been viewed as a purposive marker, it essentially carries the causative meaning. Both characteristics are shared with the analytic causative construction:

Taiwanese Southern Min hoo7  147 (5) Causative Construction: S1+ hoo7+NP1+V1



你 無 乖 予 媽媽 真 傷心 li2 bo5 kuai1 hoo7 ma2ma2 tsin1 siong1sim1 2sg neg behave hoo mother really sad ‘That you did not behave well makes Mother brokenhearted’.

Finally, the last function hoo7 予 is as an Agent marker in a passive construction. Significantly, corpus data support Lien’s (2008) analyses that the traditional view of ‘passive constructions’ in Taiwanese Southern Min should be sub-classified into two types of relations: agent-patient relation (Passive Construction-1) and affector-affectee relation (Passive Construction-2). While in the former type, both the nominal phrases preceding and following hoo7 予 are in the argument structure of the main verb, it is not necessarily so in the latter type. For instance, for the example in the following figure, i1 hoo7 ma2ma2 tshu2huat8 伊予媽媽處罰 ‘He was punished by Mother’, there is no doubt that the subject i1 伊 ‘he’ is the patient and the NP following hoo7 予, ma2ma2 媽媽 ‘mother’, is the agent of the predicate tshu2huat8 處罰 ‘punish’. (6) Passive Construction-1



伊 予 媽媽 處罰 i1 hoo7 ma2ma2 tshu2huat8 3sg hoo mother punish ‘He was punished by Mother’.

The second type of passive constructions on the other hand depicts affectoraffectee relation. As shown in the following corpus data, the contextual information indicates though the subject tsit4 king1 bio7 這間廟 ‘this temple’ is indeed the patient of the verb sio1 燒 ‘burn’, the nominal following hoo7 予 or more specifically khit4hoo7 乞予 is not the agent of the verb sio1 燒 ‘burn’, unlike the previous case. Interestingly, the third person pronoun i1 伊 is actually a pronoun referring back to the subject tsit4 king1 bio7 這間廟 ‘this temple’. (7) Passive Construction-2



這 tsit4

个 e5

御林軍， 卜 gu7lim5kun1 beh4 palace.guard want

dem

clf

个

汀州，

e5 clf

ting1tsiu1 Ting.Tsiu

因為 in1ui7 because

汀州 造反 啦！ ting1tsiu1 tso7huan2 la Ting.Tsiu rebel prt

卜 beh4 want

去 khi3 go

拍 phah4 attack

汀州 ting1tsiu1 Ting.Tsiu

的 e5

這 tsit4

間 king1

dem

clf

gen

去 khi3 go

時陣， si5tsun7 time

拍 phah4 attack

這 tsit4 dem

廟 煞 去予 伊 燒 起來。 (4.10 沙鹿) bio7 suah4 khit4hoo7 i1 sio1 khi2lai5 temple unexpectedly khit-hoo 3sg burn inch ‘The palace guards were off to attack Ting-Tsiu due to the rise of rebellion. When they were going to attack Ting-Tsiu, the temple unexpectedly caught fire fire’.

148  Chia-yin Hu Table 7.2 Paradigm of the ditransitive verbs, causative verbs/markers, dative markers, and agent markers in Quemoy Southern Min Ditransitive verbs

Causative verbs/marker(s)

Dative marker(s)

Agent marker(s)

hoo7 予 ‘give’ sang3 送 ‘send’ tsioh4 借 ‘lend’ . . .

hoo7 予 ----------------

hoo7 予 ----------------

khit4 乞 ----------------

Table 7.3 Paradigm of the ditransitive verbs, causative verbs/markers, dative markers, and agent markers in Quanzhou Southern Min Ditransitive verbs

Causative verbs/marker(s)

Dative marker(s)

Agent marker(s)

hoo7 予 ‘give’ thoo7 度 tng5 傳 khit4 乞 sang3 送 ‘send’ tsioh4 借 ‘lend’ bue7 賣 ‘sell’ . . .

hoo7 予 thoo7 度 tng5 傳 --------------------------

hoo7 予 thoo7 度 tng5 傳 --------------------------

hoo7 予 thoo7 度 tng5 傳 khit4 乞 khit4hoo7 乞予 khit4thoo7 乞度 khit4tng5 乞傳 ------

After reviewing the meanings and functions of hoo7 予 in TSM, let’s now turn to the counterparts in Quemoy and Quanzhou Southern Min. To begin with, according to Liu (2008), similar to its counterpart in TSM, QSM hoo7 予 is readily used in ditransitive, dative, and causative constructions and the isolated use of khit4 乞 in passive constructions marks one of the major distinctions between TSM and Quemoy Southern Min. In Quanzhou Southern Min, on the other hand, research has listed the isolated use of khit4 乞 as a passive marker; moreover, in addition to hoo7 予, other markers such as thoo7 度 and tng5 傳 can also be used as a ditransitive verb, causative verbs/markers, and dative markers (Lin 1993:262; Li 1997:123–126; Liu 2008, 2010). Agent markers in the form of compounds including khit4hoo7 乞予, khit4thoo7 乞度, and khit4tng5 乞傳can also be found therein. With the background for the markers in the Southern Min varieties in ­Taiwan, Quemoy, and Quanzhou in mind, we will proceed to the two field works now. 3  FIELD WORK BACKGROUND Twelve informants were interviewed in these two field work conducted in August 2010 and July 2011. Among these informants, 5 of them reside in Quemoy and the others in Quanzhou: 2 from Nan-an (南安), a couple from Yong-chun (永春), the rest from

Taiwanese Southern Min hoo7 149 Shi-shan (詩山), Jin-jiang (晉江), and An-xi (安溪). The interviews were conducted by both elicited sentences and spontaneous story-telling. The present study focuses on the use of the markers in ditransitive, causative, and passive constructions. 4  FINDINGS AND DISCUSSIONS In this section, findings and discussions of the field work will be presented concerning 1) geographical distribution of the markers and 2) the tendency of the choice of markers in dative constructions, causative constructions, and complex ditransitive constructions.

4.1  Geographical distribution The most important findings from the field work concern the spatial variation of the diverse markers used by speakers of Min sub-dialects in different areas. As can be seen from the results in Table 7.4, the choices of the markers by speakers from different areas differ from each other: Table 7.4 Markers in double object constructions, dative constructions, complex d­ itransitive constructions, causative constructions, and passive constructions in min sub-dialects 100%

81%

0%

92%

0%

–

54%

69%

0%

18%

30%

8%

29%

–

46%

31%

0%

0%

70%

0%

71%

–

%0%

0%

0%

0%

0%

0%

0%

–

0%

%0%

100%

92%

7%

56%

0%

–

50%

71%

0%

8%

43%

4%

33%

–

50%

29%

0%

0%

50%

40%

77%

–

0%

0%

0%

0%

0%

0%

0%

–

0%

100%

83%

0%

57%

0%

71%

5%

–

0%

17%

8%

7%

22%

29%

95%

–

0%

0%

92%

36%

88%

0%

0%

–

0%

0%

0%

0%

0%

0%

0%

100%

97%

9%

89%

0%

87%

40%

100%

0%

3%

73%

2%

16%

13%

50%

0%

0%

0%

18%

9%

84%

0%

0%

0%

0%

0%

0%

0%

0%

0%

0%

0%

48%

92%

4%

89%

0%

100%

56%

29%

0%

4%

36%

2%

6%

0%

40%

7%

0%

0%

39%

6%

21%

0%

0%

0%

52%

4%

18%

3%

1%

0%

4%

64%

4%

1% 1% 70%

0%

–

150  Chia-yin Hu Looking Table  7.4 more closely, we would find that the informants from Quemoy used hoo7 予 in all targeted constructions except passive constructions for all the time. They used the isolated hoo7 予 and the compound khit4hoo7 乞予 equally in passive constructions. This is quite interesting if we compare the previous research conducted mainly by Liu (2008). Her study stated that while like the case in TSM, speakers in Quemoy use hoo7 予 for nearly all the targeted constructions except for passive constructions, in which isolated khit4 乞 is chosen. Our field work indicates that in passive constructions, isolated hoo7 予 and the compound hoo7 予 were used equally the time, which is extremely similar to what has been observed in TSM. It is likely, therefore, that due to increasing exposure to TSM via mass media, people in Quemoy gradually take on similar usage to TSM. More in-depth investigation, particularly from more informants of diverse gender, age, social, and education background, is definitely needed in addition to examination of historical data if available. Among the data collected from the informants in Quanzhou, one of the most intriguing findings is the different tendency of the two informants from Nan-an. While one of them Mr. Lin chose hoo7 予 for almost all the time in all constructions, the other Mr. Wang used tng5 傳 more frequently than thoo7 度 and hoo7 予 in the double object constructions and dative constructions. A side note is that the markers in the two types of double object constructions actually are slightly different. In the first type, he used tng5 傳 for 70% of the time and thoo7 度 for the rest. In the second type, he used tng5 傳 for half of the time; thoo7 度 43% and hoo7 予 7%. His clear inclination for choosing tng5 傳 is also observed in dative construction. For causative construction, however, he used thoo7 度 much more frequently than thoo7 度 and hoo7 予. In passive constructions, more interestingly, he prefered the isolated tng5 傳 and thoo7 度, both nearly 40%, but also used the compound khit4thoo7 乞度 and khit4tng5 乞傳, with sole distinction concerning frequency. The issue concerning the distinct distribution of the markers used by these two informants from Nan-an (南安) is worth exploring. Since both of them are male, over 60 years of age; one possible explanation, leaving gender and age aside, would be related to language contact. Turning now to the couple both from Yong-chun, whose data are worth special attention. The wife is from the city, and the husband grew up in the countryside. It is also interesting to see their slightly different choices of markers as well as possible language contact that occurred right in the household. As can be seen from the summarizing Table 7.4, the wife Ms. Zheng used hoo7 予 most of the time while the husband Mr. Lin used hoo7 予 and thoo7 度 almost half of the time except in dative constructions, he used thoo7 度 far more than any other markers. At this point it is necessary to mention certain non-trivial observations during the interview. It was found that self-awareness concerning dialectal distinctions can be easily found from the informants. In the sentence elicitation section, in particular, informants would respond to my questions intended to check possible sentences with different markers by commenting which variant they think the targeted markers are used. For example, the informant from An-xi replied to my confirmation whether she would use sentences like 我傳伊兩本冊 ‘I gave him

Taiwanese Southern Min hoo7 151 two books’. She replied: 喔！「我傳伊兩本冊」，下跤人說。 . . .泉州片，泉 州、晉江的。 . . .這個「傳伊」無，安溪無啦。 ‘Oh, ‘gua2 tng5 i1 nng7 pun2 tsheh4’ is used by people underneath, such as those in Quanzhou and Jing-Jiang. We don’t say ‘tng5 i1’ in An-xi.’ Such self-awareness is also observed during my interview with the wife from Yong-chun, only with more consciences, often accompanied with self-monitoring and self-correction. One clear example is her response: 「伊租一个厝傳我」，啊，不對不對，這個是泉州話 . . . 「伊租 厝度我」，「予我」，「予我」，「伊租厝度我」，「予我」 ‘i1 tsoo1 tsit8 e5 tshu3 tng5 gua2’, oh no no, this is used in Quanzhou. ‘i1 tsoo1 tsit8 e5 tshu3 thoo7 gua2’, ‘hoo7 gua2’, ‘hoo7 gua2’, ‘i1 tsoo1 tsit8 e5 tshu3 thoo7 gua2’, ‘hoo7 gua2’. In addition to her awareness of the uses of tng5 傳, the great effort to replace hoo7 予 with thoo7 度 is also apparent. In fact, this restraining one’s own linguistic intuition is also found in her husband. In my first interview with him in 2010, actually he used thoo7 度 more than hoo7 予. In my second interview with him in 2011, however, somehow he seemed to find out my concerned issue and for some reason

Map 7.1  Tendency and distribution of the markers used in double object constructions, dative constructions, and causative constructions

152  Chia-yin Hu

Map 7.2  Tendency and distribution of the markers used in passive constructions

appeared to correct himself with hoo7 予 instead of thoo7 度. For several times, there are instances like 伊將作業交度 . . .交予我嘮 ‘i1 tsiong1 tsok4giap8 kau1 thoo7 kau1 hoo7 gua2 lo’ or 冊度伊 . . .予伊借行嘮 ‘tsheh4 thoo7 i1 hoo7 i1 tsioh4 kiann5 lo’. These obvious self-awareness, self-monitoring, and self-correction, together with the husband inclination for using thoo7 度 for 95% of the time in dative constructions, there is reason to speculate that in Yong-chun, people choose thoo7 度 as the default marker and may incline to hoo7 予 since it is the the most widely used marker in other areas where Southern Min is spoken. Generally speaking, a geographical diffusion can be perceived more easily by visualization via language maps. In combination with previous research, generally spatial gradualness is exhibited concerning the diversity in the uses of different markers. Specifically, the distributions of markers in double-object constructions, dative constructions, and causative constructions pattern together and distinctly from those in passive construction, as can be seen from Maps 7.1 and 7.2.

Taiwanese Southern Min hoo7 153 As can be seen from the Map 7.2, the most diverse area is around Quanzhou City with different inclination indicated by the size of the markers on Map 7.2. Moreover, discussed previously, language contact and shift from thoo7 度 to hoo7 予 can be speculated in Yong-chun, indicated by the arrow. Moreover, it seems that for Quemoy Southern Min, the use of ditransitive verbs, dative markers, causative markers, and passive markers is similar to the ditransitive verbs, dative marker, causative marker in Taiwanese Southern Min, and is similar to the passive marker in Quanzhou areas.

4.2  Dative or causative? One of the contentious issues involving the polyfunctions of hoo7 予 concerns its meaning and function in the complex ditransitive construction (henceforth CDC). For example, Cheng et al. (1999) specifically suggest the targeted hoo7 予 is a ‘pure’ causative verb with a canonical causative construction embedded under a transitive activity verb. Its close relationship with ditransitive constructions, however, cannot be overlooked, as suggested by Lien (2005). The indeterminacy of hoo7 予 in CDC results from the fact that CDC share common features with both dative construction (henceforth DC) and causative construction (henceforth CC) and differ with them at the same time. First, like DC, in CDC hoo7 予 follows Theme and precedes Recipient: (8) Complex ditransitive construction



阿伯 提 糖仔 予 汝 食 (74.05 嘉義) a1peh4 theh8 thng5a2 hoo7 li2 tsiah8 uncle take candy hoo 2sg eat ‘I’ll give you some sweetmeats for you to eat’.4

(9) Dative construction



伊 無 所費 予 我 啦！ bo5 soo2hui3 hoo7 gua2 lah0 i1 3sg neg money hoo 1sg prt ‘He doesn’t have any money to give me’.

(56.78 彰化)

Unlike DC though, the Theme can be readily omitted as long as it can be inferred or understood in the suitable contexts: (10) 好！ 好！ 我 犁 予 你 看。 ho2 ho2 gua2 le5 hoo7 li2 khuann3 well well 1sg plow give 2sg see ‘Well, well, I’ll show you how I plow the fields’.

(124.01 雲林)

It has been insightfully pointed out by Shao (2009) that rather than an actual transferred Theme, it is an action performed by the Speaker to be shown or ‘virtually’

154  Chia-yin Hu Table 7.5 The choices of markers in Dative Constructions, Complex Ditransitive Constructions, and Causative Constructions by Speakers from various Min sub-dialects

hoo7

100%

83%

0%

57%

0%

71%

5%

thoo7

0%

17%

8%

7%

22%

29%

95%

tng5

0%

0%

92%

36%

88%

0%

0%

khit4

0%

0%

0%

0%

0%

0%

0%

hoo7

100%

90%

5%

75%

0%

86%

29%

thoo7

0%

10%

74%

5%

22%

14%

71%

tng5

0%

0%

21%

20%

88%

khit4

0%

0%

0%

0%

0%

0%

0%

100%

100%

12%

100%

0%

100%

62%

hoo

7

0%

0%

73%

0%

11%

0%

38%

5

0%

0%

15%

0%

89%

0%

0%

4

0%

0%

0%

0%

0%

0%

%0%

thoo tng

0%

khit

7

transferred to the Addressee. In the present data, for example, it is the action of plowing that the Speaker is going to show the Addressee. The object noun for the verb le5 犁 ‘plow’ is backgrounded and thus omitted in this context. The results of the field work, as a matter of fact, contribute to the theoretical exploration of the controversial status of hoo7 予 in CDC. As can be seen from Table 7.5 the markers in CDC tend to be in accordance with those in CC. One of the informants from Nan-an, for example, used tng5 傳 in DC for more than 90% of the time, and used comparatively more thoo7 度 in CDC and CC than tng5 傳. This finding supports the widely accepted view that giving verbs can be semantically decomposed to ‘cause’ plus ‘have’ or ‘possess’. The tendency of speakers from various areas to choose specific markers for different syntactic patterns in fact serves as theoretically informed empirical research to unveil the multifunctionality of the versatile functional words. 5 CONCLUSIONS As part of a contribution to Chinese dialectal syntax, the present paper reports two field studies conducted in August 2010 and July 2011, focusing on the areal diffusion of dative, causative, and passive markers in Quemoy and Quanzhou sub-dialects. Our fieldwork studies found that 1) informants may subconsciously/ consciously monitor their production. Patterned self-correction can shed light on the true nature of the informants’ mental grammar, 2) contact among subdialects leads to ongoing dynamic changes, which can only be observed with a

Taiwanese Southern Min hoo7 155 close examination of the actual uses in each variety, 3) surveys of sub-dialects can provide significant theoretical implications to general research on a language or dialect, and 4) synchronic areal diffusion may reflect the traces of diachronic linguistic development during migration. Generally speaking, the most diverse markers are used in the Quanzhou area, with finer distinctions and preferences for markers in different uses and syntactic patterns. Taiwanese Southern Min, on the other hand, only uses hoo7 予 and khit4hoo7 乞予. For Quemoy Southern Min, the use of ditransitive verbs, dative markers, causative markers, and passive markers, is similar to the ditransitive verbs, dative marker, and causative marker in Taiwanese Southern Min, and is similar to the passive marker in the Quanzhou area. Further discussion is devoted to the long-standing issue concerning the role of hoo7 予 in complex ditransitive constructions, and the asymmetric distribution of the markers preferred by speakers from different areas suggests that the targeted hoo7 予 patterns with that in the causative construction instead of the dative constructions, a piece of finding supporting the view of decomposing giving verbs into ones with causative meanings. NOTES 1 It is argued by Lien (2008) that in passive constructions the isolated hoo7 予 and the compound khit4hoo7 乞予 are actually distinct from each other in that the former does not necessarily denote negative evaluation but the latter tends to express the speaker’s adversative feelings. 2 The causative markers here are limited to permissive and stative ones. Directive (or dynamic) causative markers are not discussed here. 3 The examples here are all taken directly from Cheng et al. 1999:147 with minor modification of characters. The free translation is provided by the author as they are not available in the original reference. 4 Notice that in these data, the speaker refers to himself as ‘uncle’, from the perspective of the addressee, the child.

REFERENCES Chen, Betty Hsiu-ying. 1972. The uses of HO in Taiwanese. In Papers in linguistics in honor of A. A. Hill, ed. by T.C. Charles, Jeffery C.H. Tung, and Anthony Y.T. Wu, 5–31. Taipei: Rainbow-Bridge Book. Chen, Chenju, and Chinfa Lien. 2008. Transfer of possession verbs in Taiwanese Southern Min: A case study of lexical and constructional effects. In Chinese linguistics in Leipzig, ed. by Redouane Djamouri, Barbara Meisterernst, and Rint Sybesma, 191–205. Paris: Centre de Recherches Linguistiques sur l’Asie Orientale, École des Hautes Édudes en Sciences sociales. Cheng, Lisa L-S., C-T. James Huang, Y-H. Audrey Li, and C-C. Jane Tang. 1999. Hoo, hoo, hoo: Syntax of the causative, dative, and passive constructions in Taiwanese. In Contemporary studies on the Min dialects, ed. by Pang-Hsin Ting, 146–203. Berkeley: Project on Linguistics Analysis. Cheng, Robert L. 1974. Causative constructions in Taiwanese. Journal of Chinese Linguistics 2:279–324.

156  Chia-yin Hu Cheng, Robert L. 1998. Tóngyìyǔ xiànxiàng zài Tái-huá duìyì cíkù lǐ de chùlǐ wèntí - tiáojiàn hé cùchéng jiégòu 同義語現象在臺華對譯詞庫裡的處理問題 – 條件和促成結 構 [The problem of processing synonymous expressions in Taiwanese-Mandarin lexicon: The case of conditional and causative construction]. In Dì Èr Jiè Táiwān Yǔyán Yǔyánxué Guójì Yántǎohuì Lùnwén Xuǎnjí 第二屆臺灣語言、語言學國際研討會論 文選集 [Selected papers from the second international symposium on languages in Taiwan], ed. by Shuanfan Huang, 529–564. Taipei: Crane. Huang, Jing-Shin. 1977. Double-object construction in Taiwanese. MA thesis, Fu-Jen Catholic University, Taiwan. Li, Ru-Long 李如龍. 1997. Dòngcí wèiyǔjù 動詞謂語句 [Verb-predicate constructions]. Guangdong: Jinan University. Lien, Chinfa. 2005. Families of ditransitive constructions in Li Jing Ji. Language and Linguistics 6:707–737. Lien, Chinfa. 2008. Special types of passive and causative constructions in TSM. In Chinese linguistics in Leipzig, ed. by Redouane Djamouri, Barbara Meisterernst, and Rint Sybesma, 223–237. Paris: Centre de Recherches Linguistiques sur l’Asie Orientale, École des Hautes Études en Sciences sociales. Lin, Lian-Tong. 林連通. 1993. Quánzhōushì fāngyánzhì 泉州市方言志 [A description of the Quanzhou dialect]. Beijing: Yuwen Chubanshe. Liu, Hsiu-Hsueh 劉秀雪. 2008. Mǐnnányǔ khit hé hoo de bǐjiào yánjiū 閩南語「乞」和 「與」的比較研究 [A comparative study of khit and hoo in Southern Min]. Zhōngguó Yǔwén Yánjiū 中國語文研究 26:27–38. Liu, Hsiu-Hsueh 劉秀雪. 2010. Fúzhōu fāngyán de qǐ yǔ gòng 福州方言的乞與共 [Qi and gong in the Fuzhou dialect]. Paper presented at the 11th International Conference on Min Dialects, Minnan Normal University, 9–11 January. Paul, Waltraud. 1988. The purposive gěi-clause in Chinese. Cahiers de Linguistique Asie Orientale 17:25–65. Shao, Jing-min 邵敬敏. 2009. Cóng V-gěi jùshì de lèihuà kàn yǔyì de juédìngxìng yuánzé 從“V給”句式的類化看語義的決定性原則 [The meaning-determining principle from the categorization of the ‘V gěi’ structure]. Language Teaching and Linguistic Studies 6:1–8. Ting, Jen, and Miller Chang. 2004. The category of gěi in Mandarin Chinese and grammaticalization. Taiwan Journal of Linguistics 2:45–74. Tsao, Feng‑fu. 1988. The function of Mandarin gěi and Taiwanese hou in the double object and the passive constructions. In The structure of Taiwanese: A modern synthesis, ed. by Robert L. Cheng and Shuanfan Huang, 165–208. Taipei: Crane. Tsao, Feng-fu. 1997. Complement and adjunct distribution and the two-place nominal in Chinese. Paper presented at the 9th North American Conference on Chinese Linguistics, University of Victoria. Wang, Hsiao-Mei. 2008. Purposives in Taiwanese Southern Min. MA thesis, National Tsing Hua University, Taiwan.

8 The etymology and grammaticalization of the continuative aspect marker le(h)4 A survey from the historical documents2 Manjun Chen 陳曼君 1  THE HISTORICAL DOCUMENTS REFERRED TO Regarding the references of historical documents, they include Míng Jīajìngkān Lì Jìng Jì Xìwén Jiàolǐ published in 1566 (Annotated Texts of the Romance of Li Jing Ji of Ming Jiajing Edition; henceforth JJ), Míng Wànlìkān Lì Zhī Jì Xìwén Jiàolǐ published in 1581 (Annotated Texts of the Romance of Li Zhi Ji of Ming Wanli Edition; henceforth WL), Qīng Shùnzhìkān Lì Zhī Jì Xìwén Jiàolǐ published in 1652 (Annotated Texts of the Romance of Li Zhi Ji of Qing Shunzhi Edition; henceforth SZ), and Qīng Guāngxùkān Lì Zhī Jì Xìwén Jiàolǐ published in 1884 (Annotated Texts of the Romance of Li Zhi Ji of Qing Guangxu Edition; henceforth GX). These are edited and annotated by Wu (2001a, 2001b, 2001c, 2001d). In addition, we also refer to Jīnhuānǚ (Wu 2002a), Sū Liù Niáng (Wu 2002b); both were published during Ming Dynasty’s Wanli era (1572–1620). Apart from these, Míngkān Mǐnnán Xìqǔ Xuánguǎn Xuǎnběn Sānzhǒng (Three Ming Anthologies of the Theatre and Art Song of South Fukien, henceforth Míngkān Xuǎnběn; van der Loon 1995), Quánzhōu Chuántǒng Xìqǔ Cóngshū, Dì 1–15 Juàn (Quanzhou traditional drama series (1–15); Zheng 1999–2000), a corpus of Yǔyuàn (the texts in the journal published in Taiwan when Taiwan was under Japanese rule) and Táiwān Xiàndài Mǐnnányǔ Gùshìjí (The Stories in Modern Taiwanese Southern Min), several dictionaries and some relevant previous research are also found in the list of references. 2 THE CONTINUATIVE ASPECT MARKER LE(H)4 IN SOUTHERN MIN According to our survey, le(h)4 is employed as the continuative marker in not only Taiwanese Southern Min, but also in other Southern Min dialects, including Amoy, Quanzhou, and Zhangzhou. This continuative marker is relatively common especially in the Quanzhou area, Taiwan, and Amoy. When it is written with Sinograms, several forms are employed respectively. Aside from leh4 咧, people also use 嘞 (Li 1996:204) or 𠼱 (Wang 2010:87). In the following, we will represent it with 咧 in the examples shown in Sinograms, except for citations from other research.

158  Manjun Chen In those aforementioned dialects in which this marker is observed, le(h)4 can occur not only in an existential construction (see (1a)), but also in other kinds of sentences (see (1b)). (1) a 怹 厝 个 門 定定 關 in1 tshu3 e5 mng5 tiann7tiann7 kuinn1 3sg.gen house lnk door usually close ‘The door of his house has been frequently closed’.

咧。 leh4

b 你 碗 共 我 捀 li2 uann2 kang7 gua2 phang5 2sg bowl appl 1sg hold.up.with.both.hands ‘Take this bowl and hold it for me’.

咧。 leh4

(2) a 伊 倒 咧， 無 坐 咧。 i1 to2 leh4 bo5 tse7 leh4 3sg lie leh neg sit leh ‘He is lying there; not sitting there.’ b 你 坐 咧， 袂使 li2 tse7 leh4 be7sai2 2sg sit leh not.allowed ‘Sit up. Don’t lie there’. (3) a 這 tsit4

領 nia2

倒 to2 lie

(Huian)

leh

(Huian)

leh

(Taiwan)

咧。 leh4

(Taiwan)

leh

衫 伊 定定 sann1 i1 tiann7tiann7 dem clf clothes 3sg usually ‘He wears these clothes usually’.

穿 tshing7 wear

咧。 leh4

(Amoy)

leh

b 我 叫 你 下 咧， 你 就 下 咧。 gua2 kio3 li2 khe7 leh4 li2 tioh8 khe7 leh4 1sg ask 2sg put leh 2sg prt put leh ‘I told you to put it there, and you should just put it there’.

(Amoy)

As pointed out by Li (1996:204, among others), the continuative marker le(h)4 denotes the continuation of the state after some activity is done. No object is allowed to follow it. Whenever there is an object, the object has to precede this marker. Regarding the etymology of this marker, there are two prevalent proposals in the literature. The first suggests zhe 著 in Classical Chinese as its origin (Wang 1980:308; Mei 2000:155; Cao 1995:26–37; Wu 2004:17). The other proposes that it was derived from the AB-form in VAB and its simplification (Lü 1984:58–64;3 Liu 1996:26–30, among others). In section 3, we will review the literature with respect to the etymology and the grammaticalization of le(h)4.

Continuative aspect marker le(h)4 159 3  PREVIOUS STUDIES Below we summarize the research of Yang (1992), Wang (2004), Wu (2008), Lin (2006), Lin (2010) respectively. Yang (1992:349–394) provides one of the earliest studies on the marker le(h)4. She suggests that the durative aspect zhe 著 in Mandarin Chinese was derived from the locative preposition 著, which expresses existence. During her research on the transformation of the verb 著 in the historical documents, she found the two function words, the continuative leh4 咧 in Southern Min and the durative aspect zhe 這 in Mandarin Chinese are semantically equivalent. Moreover, leh4 咧 can be reasonably considered as a product of lenition from ti7. Based on these, she claims that both ti7 著 and leh4 咧 share an identical original character, which is 著. Regarding the grammaticalization and sound change from ti7 著 to leh4 咧, Yang suggests the following three routes. The first is based on the consonantal variation between voiced and voiceless onsets, as in pat4/bat4 識 (the experiential aspect marker), i2tiau5/i2liau5 椅條 ‘bench’, and pui5tut4tut4/pui5lut4lut4 肥朒朒 ‘fat’. The second concerns vowel lowering. According to Yang, the vowel [ɪ] has been reduced and shifted to [ɛ] due to a lower tongue position triggered by the grammaticalization process. As for tone, the mid-level tone was neutralized, also as an outcome of the grammaticalization. Based on these conjectures, Yang suggests both the Mandarin durative marker zhe 著 and Southern Min continuative marker leh4 咧 are derived from the locative preposition zhe 著. Nonetheless, she gives no explanation for the difference of pronunciation between zhe 著 and leh4 咧. And her speculation of the consonantal shift is dubious, for the examples provided are synchronic variations of a lexical item, instead of diachronic changes due to grammaticalization. Wang (2004:128–131), in addition to Yang, also claims that the continuative marker leh4 咧/嘞 was grammaticalized from ti7 著. His reconstruction of the historical transition of ti7 著 with respect to its graph, pronunciation, and meaning is as follows. 著 [tɰ4] (existential) → 除 [tɰ0] (continuative; accomplishing) → 處 [tə0] (locative preposition; continuative) → 咧/嘞 [lə0] (continuative) Wang (2010:91) revised his analysis later and came up with the revised reconstruction. 著

除

除

處

𠼱 (咧/嘞)

verb; accomplishment; accomplishment/ continuative/ accomplishment; preposition continuative continuative accomplishment continuative/ progressive Lì Jìng Jì

[tɯ22]→[tɯ０]

Mǎn Tiān Chūn

[tɯ ]

Quanzhou dialect

[tɯ22]

22

[tɯ0]→[tǝ0]

[tǝ0] [lǝ0/lǝ55]

160  Manjun Chen Wang’s reconstruction is persuasive with regard to pronunciation shift. However, it seems suspicious from the semantic and syntactic perspectives. According to Wang (2010:89–90), using tor0 除 to express continuation obviously originated from its early usage which expresses accomplishment, based on their counts in the documents. In Lì Jìng Jì (1566), there are 40 to 50 odd instances in which tor0 除 denotes accomplishment, and only six express continuation. More than 20 examples are found in Mǎn Tiān Chūn 滿天春 (1604) in which tor0 除 is employed to express accomplishment, compared with only one denoting continuation. Below are the seven occurrences of tor0 除 in Lì Jìng Jì and Mǎn Tiān Chūn which indicates continuation.4 (4) [末]



雙親 siang1tshin1 parents

堂上 tong5ting2 parents

老 lau7 old

年紀， ni5ki2 age

功名 kong1mia5 achievement

牽絆 khan1tinn5 tangle

覓 除 伊 (Lì Jìng Jì Act 4 Yùnshǐ Dēng Tú) bih4 tor0 i1 seek tor 3sg ‘His parents are old, but he is obsessed by fame and achievement’.

(5) [外] 都



(6) [旦]



是 你 只 死 虔婆 乘 除 too5 si7 li2 tsit4 si2 khian5po5 sing3 tor0 all cop 2sg dem dead bitch pamper tor 仔兒， 即 會 逞性。 (Lì Jìng Jì Act 14 Zé Méi Tuì Pìn) kann2li5 tsiah4 e7 sai2sing3 son prt will throw.a.tantrum ‘It’s all because of you the damned hag, who pampers your daughter, that she threw a tantrum’. 陳三， tan5sann1 Tan.Sann

你 li2 2sg

到 kau3 enough

薄幸。 pok8hing7 heartless

你 li2 2sg

袂記得 be7ki3tit4 forget

花 hue1 flower

嬌 kiau1 lovable

姿 tsi1 disposition

潤 lun7 gloss

色 sik4 color

之 tsi1

時； si5 time

今旦 kim1tuann3 today

那

謝

了，

便

lnk

提 覓 除。 (Lì Jìng Jì Act 24 Yuán Nèi Huā Kāi) na7 sia7 liau2 pian7 theh8 hinn3 tor0 if wither pfv prt take discard tor ‘Tan Sann, you’re so heartless. You don’t remember those days when the flower was lovable and splendid. Once it withered, you just cast it away’.

Continuative aspect marker le(h)4 161 (7)

[貼]



(8)

[貼]

娘仔 阮 書 拾 去， 連 阮 都 gun2 niu5a2 phue1 khioh4 khi3 lian5 gun2 too3 1sg.gen madam letter pick.up asp even me prt 瞞 除， 不 乞 我 知。 (Lì Jìng Jì Act mua5 tor0 m7 khit4 gua2 tsai1 24 Yuán Nèi conceal tor neg let 1sg know Huā Kāi) ‘Madam got the letter and kept it from me, lest I know what the letter is about’. 伊 i1 3sg 廳 thiann1 hall



(9)

[貼]



(10)

[外]

即 tsiu7 prt

甘心 kam1sim1 willingly

捧 phang5 hold

盆 phun5 basin

水， tsui2 water

邊。 pinn1 margin

忍 lim2 tolerate

除 tor0

志氣， tsi3khi3 ambition

受 恁 siu7 lin2 suffer 2sg

tor

掃 sau3 sweep

一 嘴 苦氣。 (Lì Jìng Jì Act 26 Wǔniáng Cìxiù) tsit8 tshui3 khoo2khi3 one mouth resentment ‘He therefore became willing to get a basin of water to serve and to clean the hall room. He left his dignity behind and suffered a great deal of your resentment’. 覓 bih4 hide

除 tor0 tor

爹媽 在 tia1bu2 ti7 father.mother in

後頭， au7thau5 home

思量 si1liong5 think

起來 khi2lai5 asp

目滓 流。 (Lì Jìng Jì Act 41 Lǚguǎn Xù Qíng) bak8sai2 lau5 tear low ‘She left her parents behind at home. Whenever she thinks of this, she can’t help but cry’. 只 tsit4 dem

一 tsit8 one

孜娘人 tsi1niu5lang5 lady

多少 心事， gua2tsue7 sim1si7 how.many worry

明 bing5 obviously

知 tsai1 know

阮 gun2 1pl.excl

只處 tsit4te3 here

坐， tse7 sit

不   看。 m7  khuann3 neg  see

故意 thiau1kang1 intentionally

162  Manjun Chen



不免 put4bian2 unavoidably

力 liah8 hold

扇 sinn3 fan

連 lian5 even

面 遮 bin7 lia1 face cover

假意

過去，

伊

必然

是

除， tor0 tor

來

留  我。 (Mǎn Tiān Chūn) ke2gui2 kue3khi3 i1 pit4lian5 si7 lai5 lau5  gua2 feigningly go.there 3sg surely cop come keep 1sg ‘This lady has a load on her mind. She obviously knows that I am sitting here, but pretends that she is not watching me. Consequently, she uses her fan to cover her face. If I feign to go away, she will surely ask me to stay’.

Among these examples, four of them are of the structure ‘V+除+O’ (see (4), (5), (8), and (9)). The others have a verb preceding tor0 除 without a following object (‘V+除’). As has been pointed out, syntactically, whenever a verb is succeeded by the continuative marker, no object is allowed to follow (except for the locative object). And neither in the playscripts dated from the Ming dynasty to the Qing dynasty nor in today’s Southern Min dialects do we find exceptions to it. As for (4), (5), (8), and (9), in which a non-locative object follows tor0 除, we should note that tor0 除 in these sentences had not been completely grammaticalized yet. The readings of the sentences tell us that tor0 除 used in those examples is better interpreted as a resultative complement than as a continuative marker. Regarding the remaining three examples, tor0 除 in (6) does not function as a continuative marker. Rather, it is a content word that means

Figure 8.1  The evolvement from 著 to 嘞 (Wu 2008:21)

Continuative aspect marker le(h)4 163 ‘away’. Although we do have (7) and (10) in which tor0 除 can be interpreted as a continuative marker, the context suggests that a more reasonable understanding is a resultative complement which denotes steadiness. In these two examples, the speakers obviously put emphasis on the result of the action, instead of the continuity of the action. Based on these observations, I argue that tor0 除 in Lì Jìng Jì and Mǎn Tiān Chūn does not denote continuity and should not be deemed as a continuative marker. Wang’s reconstruction of ti7 著 is, therefore, unpersuasive. Regarding the etymology of this continuative marker, Wu has a proposal similar to that of Wang. According to Wu (2008:18), the origin of this marker is 著 and it has been grammaticalized via a transition stage of ti7 佇. Figure 8.1 is the reconstruction of the diachronic evolvement from the content word ti7 著 to the continuative marker leh4 嘞 in the Quanzhou dialect proposed by Wu (2008:21). Wu does not provide enough evidence for her reconstruction and, therefore, her proposal seems doubtful. In addition to Wang (2004) and Wu (2008), Lin (2006:14) also reconstructs the diachronic development of the continuative marker in Quanzhou dialect as follows:

V在X

→

V在(只/許)處

(A) S1

→

(B) S2

V在處

→

(C) S3

V處 (D)

Figure 8.2  The development of continuity (a)

The following examples for illustration are from Lin (2006:14):5 (11) 幸

hing3 fortunate



然 lian5 manner

見 kinn3 see

君， kun1 2sg

有 u7 have

緣 ian5 predestined.affinity

相逢 siong1hong5 meet

在 只 花 邊。 (Liú Kuí Huì Yún Yīng 劉奎會雲英) ti7 tsit4 hue1 pinn1 at dem flower side ‘I am so lucky to have seen you. We are predestined to meet by the flowers’.

(12) 我

gua2 1sg

曉得， hiau2tit4 know

必定 娘子 個， pit4ting7 niu5a2 e5 definitely mistress poss

我 gua2 1sg

不免 共 put4bian2 kang7 have.to prep

伊 i1 3sg

失落 sit4loh8 lose

在  只處， ti7   tsit4te3 prep  dem-te

卻 起來， 掛   只  花 khioh4 khi2lai5 kua3  tsit4  hue1 pick up.come hang  dem  flower

164  Manjun Chen 樹

上，

待

tshiu7 ting2 tree above



小妹， 門 掩 上 sio2be7 mng5 am1 siong7 sister door close up

入來

便

lip8lai5 pian7 enter-come just

開

sak4 push

khui1 open

prep

許處， 你 hit4te3 li2 dem-te 2sg

卜 beh4 want

入來。 (Liú Kuí Huì Yún Yīng 劉奎會雲英) lip8lai5 enter-come

只處， 我 句 惜惜 你。 (Liú Kuí Huì Yún Yīng 劉奎會雲英) niu5a2 li2 tua3 tsit4te3 gua2 ku2 sioh4sioh4 li2 mistress 2sg stat dem-te 1sg indeed cherish 2sg ‘Dear, you stay here. I do cherish you’.

(15) [丑]



搡

在 ti7

‘My little sister, the door over there remains closed. If you want to come in, just push and open it and come in’.

(14) [生]



尋 便了。 (Liú Kuí Huì Yún Yīng 劉奎會雲英) niu5a2 lai5 tshe7 pian7looh4 mistress come look.for then-prt

tai7 wait

來

‘I know. This belongs to her and she lost it here. I have to pick it up and hang it on the blossoming tree. It will be easier for her to find it after she comes’.

(13) [生]



娘子

娘仔

你

不 m7

是 si7

neg

娘子， niu5a2 mistress

帶

阿 ah4

是 si7

cop

你 li2 2sg

prt

cop

都 too3

是 si7

許 hit4

prt

cop

dem

一 tsit8 one

我， gua2 1sg 個 e5 clf

無 bo5 neg

廉恥。 liam5thi1 sense.of.shame

劉 lau5 Lau

厝 tshu3 house

店 在 處 (Liú Kuí Huì Yún Yīng 劉奎會雲英) tiam3 ti7 te3 stay prep te ‘If it were not you, then could it be me?! Don’t you have any sense of shame? Dear lady, it is Lau who is here’.

(16) [外] 不卜

m7beh4 neg-want [旦] 你 li2 2sg

跪處 kui3te3 kneel-te 都 too3 prt

賴 lua7 hang.on 不可 m7thang1 neg-can

人。 lang5 person 跪處 kui7te3 kneel-te

陪 pue5 accompany

阮 gun2 1sg

一下。 tsit8e7 awhile

Continuative aspect marker le(h)4 165 [外] 今 kim3 now



跪處 陪 kui7te3 pue5 kneel-te accompany

你 li2 2sg

好了。 ho2loo1 good-prt

(Mǎn Tiān Chūn)

‘You can’t kneel here to pester him./Can’t you also kneel down to accompany me?/OK, now I’ll kneel here to be your company’.

Based on the observations above, Lin (2006:14–15) claims the grammaticalization process has three stages (from A to D in Figure 8.1). Stage A is about the comprehensive description of an event, including the specific location of the action. In this stage, ti7 在 serves as a preposition and X indicates the specific position. These are exemplified by the phrase hue1pinn1 花邊 in (11). In stage B, te3 處 then evolved into a general postposition and only allowed tsti4 只 and hit4 許 to collaborate with it to form compounded demonstratives such as tsit4te3 只處 (see (12)) and hit4te3 許處 (see (13)). Lin suggests that it is this development that endowed te3 處 with the potential to further become a real postposition. However, the construction found in stage B is still about the comprehensive description of an event. Tsit4te3 只處 and hit4te3 許處 are locative arguments which cannot do without a preceding preposition. Nonetheless, the requirement of a preposition is not observed in (13). In (13), tua3tsit4te3 帶只處 is a counterpart of zhànzàizhèlǐ 站在這裡 ‘(literally) stand at here’ in Mandarin. The postverbal tsit4te3 只處 in this stage needs no preposition anymore and has transformed into a postposition. In stage C, tiam3 店 denotes a standing action. The action itself is the foreground and the location of this action submerges. That is to say, the action becomes the sole focus of description. In addition, in C, which is derived from B, ti7 在 and te3 處 were agglutinated and precede the verb. As for stage D, te3 處 has become a real postposition and no preposition is allowed to precede it. This is exemplified by the V 處 sequence in (16). Regarding his reconstruction, Lin (2006:15) admits that there is a problem with respect to stage C: (15) is the single example that illustrates the sequence ti7te3 在 處 in the text during Ming dynasty. We cannot preclude the diachronic development after stage B is S3 instead of S2 in Figure 8.3. At least S3 is more probable than S2, according to Lin.

V 在處（C）

V在X （A）

S1

V 在（只/許）處 （B）

S3 V 處（D） Figure 8.3  The development of continuity (b)

166  Manjun Chen Lin’s (2006) reconstruction sheds some light on the etymology of the continuative marker and has some critical contributions. However, he fails to distinguish the grammaticalization process of the existential construction from that of the non-existential construction. As a result, the, in fact, necessary phase, S2, is wrongly played down. In addition to the research reviewed so far, the latest study on this topic comes from Lin (2010) and his study provides us with a relatively different view. According to him (2010:386, 389), the continuative marker is a product of the grammaticalization of the phrase: ‘preposition+locative demonstrative pronoun’, but not from the verb ti7 著. In other words, the marker has its origin in the phrase ti7hia1 佇遐 ‘prep there’, which occurs as a complement of the verb in a sentence. In order to support his claim, Lin (2010:392) further exemplifies his argument with the pronunciations of this phrase in Zhangzhou, Amoy, and Taiwanese dialects and comes up with the reconstructed diachronic transition in syntax, semantics, and phonetics as follows. 1 Prepositional phrase as a complement [ti33hia55]→[ti33ia55]→[ti33a55] Phonetically, contraction process observed Semantically, the location reading gradually bleached 2 Weakend prepositional phrase; disyllabic continuative marker [ti33a55]→[ti33a0]→[ti33e0] Phonetically, vowel raising and tonal neutralization observed Semantically, the location reading being further impoverished, and the continuity reading emerging 3 A monosyllabic continuative marker from further contraction [ti33e0]→[te0] Phonetically, further contraction observed Semantically, the location reading nearly vanishing 4 The new weakend monosyllabic continuative marker [le0]→[le0] Phonetically: The consonant becoming voiced Semantically, a real continuative marker has materialized According to Lin (2010:392), the extent of grammaticalization differs among Southern Min dialects and, therefore, the phonetic forms also vary. In Zhangzhou dialect, it surfaces with a relatively less grammaticalized disyllabic form 佇咧 [ti33 e0], compared with the further weakened monosyllabic form 咧 [le0]/嘞 [lə0] in Amoy, Quanzhou, and Taiwanese dialects. It is noteworthy that Lin’s reconstruction was built on a synchronic data set. An instance of this is the pronunciation of the phrase 佇遐 [ti33–21 hia55] that he

Continuative aspect marker le(h)4  167 presumes is the one in today’s Zhangzhou dialect. Nonetheless, 遐 may not be pronounced as [hia55] in the past, and may not even be monosyllabic6. Consequently, Lin’s reconstruction may fail to reflect the true grammaticalization process. Moreover, the disyllabic form employed to denote continuity is not merely found in Zhangzhou dialect. Nowadays we still observe 佇咧 [ti33 le0] marking continuity in the Quanzhou dialect in sentences such as i55 uah4 ti33 le0 伊活佇咧 ‘(literally) he is living’. The Quanzhou disyllabic form poses a difficulty for the phonetic transition in the reconstruction proposed by Lin. Based on the discussion so far, I suggest that the way to approach the real grammaticalization process of this marker must be based on the historical documents, in order to pinpoint the date of its emergence and where it evolved from. 4 THE RISE OF THE CONTINUATIVE MARKER IN SOUTHERN MIN The earliest extant Southern Min historical document is Míng Jiājìngkān Lì Jìng Jì Xìwén Jiàolǐ published in 1566 (Annotated Texts of the Romance of Li Jing Ji of Ming Jiajing Edition; JJ). Since there is no continuative marker found in this document, it seems that this marker had not come into existence at that time. The full title of JJ is Chóngkān Wǔ Sè Cháo Quán Chā Kē Zēng Rù Shīcí Běiqǔ Gōulán Lì Jìng Jì Xìwén Quánjí 重刊五色潮泉插科增入詩詞北曲勾欄荔鏡記 戲文全集. The term Cháo Quán Chā Kē in the title indicates that this playscript is a compilation of two playscripts of Lì Jìng Jì in Chaozhou dialect and Quanzhou dialect respectively. However, for today’s Southern Min speakers, the language in JJ reflects nothing but the characteristics of the Quanzhou dialect, except for several obviously Chaozhou expressions. Some people, therefore, suggest JJ was written simply in the Quanzhou dialect. I cannot agree with this. The reason is that we should not neglect the fact that the Quanzhou dialect and the Chaozhou dialect had not yet been differentiated into two dialects when Lì Jìng Jì was written. The rise of a different Chaozhou dialect was a relatively late development. Since the available versions of the Lì Jìng Jì were all republished ones, Wu (2001a:5) speculates that the playscript was first written in the 15th century. Based on Wu’s conjecture, the Southern Min dialects, including Quanzhou and Chaozhou, still did not have any continuative marker before the 15th century. In the existent Southern Min playscripts published in the Ming Dynasty, diachronically, the first continuative marker was found in WL (1581). Among the 9 examples in WL, 3 of them are in the form of ti7te3 在處, 6 in the form of te3 處. The items bearing the continuative marker ti7te3 在處 include verbs and verbal phrases (among them, one verb-complement construction and one verb-object construction). All the verbs involved are action verbs, such as hong1 封 ‘to seal’, am1 掩 ‘to cover’, and sinn1 生 ‘to give birth’. On the other, some of the items bearing the marker te3 處 are adjectives, in addition to verbs. The only one verb that bears te3 處 is pang3 放 ‘to put’ and the two examples of adjectives are of the same lexical item: khuan1 寬 ‘wide’. We know that, in the beginning, the continuative marker was postverbal and it denoted the sustained state after an action was

168  Manjun Chen accomplished. The post-adjectival occurrence of this marker indicates that it had been employed in the language for quite a period of time and its emergence must be earlier than when WL was published. In conclusion, for today’s Southern Min speakers, it is not difficult to read WL in the Chaozhou dialect. And, therefore, we suggest Quanzhou and Chaozhou had not evolved into two relatively independent dialects at that time. According to Lin (2001:304), the Chaozhou and Shantou speakers living in eastern Guangdong are descendants of the immigrants from Fujian during the period from the Tang dynasty to the Qing dynasty. The Chaozhou dialect had not become an independent dialect before the Ming dynasty (Lin 2001:310). However, these two dialects did differ in some respects. By way of example, the two continuative markers ti7te3 在處 and te3 處 that we found in Míngkān Xuǎnběn seem to still be in a relatively early stage of their development, compared with the markers found in an earlier version of the Lì Jìng Jì in the Chaozhou dialect. We consequently conjecture that the Chaozhou continuative marker emerged earlier than its Quanzhou counterpart. Though we still have no clue on the specific time of the emergence of the Southern Min continuative marker, we can be sure that it must be later than the writing of Lì Jìng Jì. 5  THE ETYMOLOGY OF THE CONTINUATIVE MARKER Remember in the last section, there are two continuative markers, ti7te3 在處 and te3 處, in WL. The question then arises, ‘is one of the two the proto-form of the other’? Based on the Southern Min historical documents, we suggest a more complex picture for the etymology of the continuative marker. And we believe that a better understanding of the usage of this marker in existential constructions and non-existential constructions will help us clarify the relationship between 在處 and 處 and, furthermore, to reveal where the continuative marker 咧 comes from. In existential constructions, both ti7te3 在處 and te3 處 are found as continuative markers. We suggest they have their own origins respectively. As for the non-existential construction, it only employs te3 處 as the continuative marker. And we also maintain etymologically it does not derive from ti7te3 在處. Instead, we suppose that historically the first continuative marker in Southern Min was te3 處. The continuative usage of te3 處 is a product from the grammaticalization of the content word te3 處, which denotes ‘place’ or ‘locus’. Lin (2006:12) claims that localizers, such as siong7 上 and lai7 內, denote specific positions. In cases when position specification is not needed, grammaticalization would be triggered. The result is a general localizer te3 處, which is relatively abstract and bleached semantically. We do not concur with Lin. Unlike Lin, we figure that te3 處 had been used to indicate some specific locus7 with a demonstrative. The relative bleached meaning is a later development of grammaticalization. Below are the examples in which te3 處 was used as a content word before it derived into a continuative marker.

Continuative aspect marker le(h)4 169 (17) 三兄 在 值處？ 阿娘 叫 使 你。(24.164 WL) sann1hiann1 ti7 ti7te3 a1niu5 kio3 sai2 li2 Sann.hiann prep prep-te mistress call order 2sg ‘Sann-hiann, where are you? Madam is looking for you to run an errand’. (18) 簡 在 只處。 (26.111 JJ) kan2 ti7 tsit4te3 maid prep dem-te ‘I am here at your service’. (19) 畜生！ 踢 伊 倒 許處， 夭 句 無 打。 (14.261 JJ) tsing1sinn1 that4 i1 to2 hit4te3 iau2 koh4 bo5 phah4 livestock kick 3sg lie dem-te still again neg beat ‘Bastard! Kick him and throw him there. He still denied that he beat her’. 只處。 (14.245 JJ) (20) 小七， 跪 sio2tshit4 kui7 tsit4te3 Sio.tshit kneel dem-te ‘Sio-tshit, kneel here’. (21) 既



不

願

便

返去， 乜苦

卜 帶 只處。 (1. 26 Míngkān Xuǎnběn) ki3 m7 guan7 pian7 tng2khi3 mih8khoo2 beh4 tua3 tsit4te3 since neg willing then return why.bother will stay dem-te ‘If you are unwilling, then you should just return. Why do you bother to stay here’?

As shown above, te3 處 occurs not only in the existential construction but also in the non-existential construction. When it is found in an existential construction and preceded by demonstratives like tsit4 只, hit4 許 or interrogative pronoun ti7 值, it can either be the object of the verb of existence and appearance 在 (see (17) and (18)) or the object of an action verb (see (19)). In the latter case, it seems that the verb 在 is omitted. On the other hand, when te3 處 occurs in an non-existential construction, its combination with a demonstrative can only be the object of the verbal predicate and no locative preposition ti7 在 is allowed to show up (see (20) and (21)). The insertion of ti7 在 in (20) and (21), such as sio2tshit4 kui7ti7tsit4te3 小七，跪在只處 and ki3m7guan7pian7­tng2khi3 ­mih8khoo2beh4tua3ti7tsit4te3 既不願便返去，乜苦卜帶在只處, will turn the non-­existential meaning into an existential one. Consequently, the sentence would become perfect instead of imperfect and in conflict with the imperative force of the speaker.

170  Manjun Chen As we have seen, te3 處 is always preceded by a demonstrative. And, based on this, we suggest that it unexceptionally indicates some specific location. In other words, the demonstrative underlies the locative denotation of te3 處. However, it is worth noting that the conspicuousness of this locative denotation depends on where the focus of the sentence is placed. The paragraph below is cited from Zhang and Fang (1996:73). The focus of a sentence is where the core meaning of this sentence is carried. Usually the older the information is, the more to the front it occurs in a sentence. In that vein, the closer a sentential part is to the end of the sentence, the newer the information it conveys is. And we often consider the sententialfinal part as the sentential-final focus. And this kind of focus is of normal focus. On the other, we also have the so-called contrastive focus when the relevant sentential part does not introduce new information. On the contrary, the information has been brought into the context either directly or indirectly, and it is emphasized on purpose of contrasting with by the speaker. What is contrasted with can be either mentioned previously, known in the context, or presupposed by both the speaker and the addressee. Now go back to the examples. In (17) to (21), it is the sentence-final phrase, demonstrative+處, that is under normal focus. And the locative meaning of 處 emerges well. However, focus other than normal focus may prevail. As a result, focus position can be somewhere except for sentential-final position. Note that constituent questions like (17) must be considered separately. For constituent questions, focus is always on the interrogative pronoun, such as ti7te3 值處. What concerns us here are those cases that involve focus, excluding normal focus and are not constituent questions. For instance, in (18) to (21), the ­sentential-final phrase of ‘demonstrative+處’ is not focused and consequently, this phrase is interpreted as background information (in contrast to the foreground information). With its background status, its content meaning is bleached and gradually grammaticalized. As a result, the preceding demonstrative becomes more and more pointless and finally gets dropped. Hence, te3 處 begins to occur independently and its locative meaning has been bleached. At this stage, postverbal te3 處 denotes only the sustaining state after the action of the verb is done. Here are some examples: (22) 阿媽 叫： 捧 入 內廳 去 食， a1ma2 kio3 phang5 lip8 lai7thiann1 khi3 tsiah8 old.lady shout hold.with.both.hand into inner.hall to eat 眾 群 媽 在處。 (10.54–55 WL) tsiong3 kun5 ma2 ti7te3 many group old.lady prep-te ‘The old lady shouted, ‘Take the food to eat it in the room inside. Here are many elderly ladies’.

Continuative aspect marker le(h)4  171 (23) 哇， gua7 prt



值 ti7

個 e5

q

clf

什種 tsap8tsing2 bastard

食 tsiah8 eat

西瓜 si1kue1 watermelon

嗾 我。 (9.31 WL) 放處 3 3 8 pang te tshuh gua2 put-te slip 1sg ‘Wow! Which son of a bitch ate watermelons and left the rind here to make me slip and fall’.

(24) 小妹， 放處， 請 自在。 sio2be7 pang3te3 tshiann2 tsi7tsai7 little.sister put-te please be.at.ease ‘Miss, just put it down. Make yourself at home’. (25) 阿娘， 花 a1niu5 hue1 madam flower



皮， phe5 peel

今 kim1 now

正 開 tsiann3 khui1 right.now open

放處 乞 pang3te3 khit4

不 m7 neg

拗 au2 turn.and.break

來 lai5

插， tshah4

come

stick.in put-te

let

yellow.bee white.butterfly

採 tshai2 gather

謝 sia7 wither

夭 iau2 very

可惜 年。   (20.24–25 WL） khoo2sioh4 ni0 a.pity sfp

年， ni0 sfp

黃蜂 ng5phang1

(17.40 WL)

粉蝶 hun2tiap8

‘Madam, the flowers are blooming. If we don’t pick the flowers to arrange them in a vase, these flowers would only serve to be pollen-collected by bees and butterflies before they wither. What a pity’!

These sentences can be separated into two subgroups: (22) and (23) are existential and (24) and (25) are non-existential. Regarding the existential construction, there are objective existential sentences and subjective existential sentences. Objective existential sentences denote some fact depicted by the speaker without any personal feelings and opinions. For instance: in (22), it is all about the fact that ‘there are many elderly ladies’. As for subjective existential sentences, they differ in that speaker’s personal feelings and opinions are involved. In (23), the fact ‘leaving the watermelon rind here’ is linked with the speaker’s negative personal opinion of ‘making me slip and fall’. Therefore, sentences like (23) are uttered in order to express speaker’s standpoint, viewpoint, and attitude. Since subjectivity is incompatible with objective existence which tends to be brought in by 在, consequently, sentences such as (23) do not employ te3 處 together with ti7 在. On the other hand, non-existential sentences can also be classified into two categories: imperatives (e.g. (24)) and non-imperative imperfect sentences (e.g. (25)). We suggest that te3 處 in the existential sentences, (22) and (23), was

172  Manjun Chen grammaticalized from ‘demonstrative+處’ in (18) and (19) respectively. And the two cases of te3 處 in non-existential (24) and (25) figure as grammaticalized products from ‘demonstrative+處’ in (20) and (21) respectively. Among the four kinds of sentences exemplified in (18) to (21), te3 處 in objective existential sentences and imperatives is probably the one that underwent grammaticalization first, promoted by the linguistic context where they occur in. In WL, three examples of 在verb+處 are found, including (22). Here are the other two: (26) 丈夫 值 個 在處？ tsa1poo1 ti7 e5 ti7te3 male q clf prep-te ‘Which man is here’?

(25.7 WL)

(27) 不免

入 來 去 遞 貼。 值 個 put4bian2 lip8 lai5 khi3 tu1 thiap4 ti7 e5 unavoidable enter come to deliver invitation q clf



在處？ (47.93 WL) ti7te3 prep-te

‘I have to go in to deliver the invitation. Someone at home’?

All the phrases ti7te3 在處 are derived from ti7tsit4/hit4te3 在只/許處 by grammaticalizing tsit4/hit4te3 只/許處 into te3 處 in the linguistic context of sentences, such as (22), (26), and (27). In other words, the pre-grammaticalized predecessor of ti7te3 在處 in these sentences is supposed to be either ti7hit4te3 在許處 or ti7tsit4te3 在只處. Note that either hit4te3 許處 in (22) or tsit4te3 只處 in (26) and (27) carry old information; hit4te3 許處 in (22) denotes the room inside (the inner hall) previously mentioned and tsit4te3 只處 in (26) and (27) points to exactly where the speaker is. Due to their old information status, these phrases cannot be focused. What is focused by the speaker in (22) is the fact of the presence ti7 在 of many elderly ladies, tsiong3kun5ma2 眾群媽. With regard to (26) and (27), it is the (absence/) presence of someone or anyone that concerns the speaker (the interrogator). Among these three instances, tsit4te3 只處 in sentences of the kind of (26) and (27) is more prone to be grammaticalized into te3 處 because the venue where the speaker is restricts the possible denotations of the demonstrative phrase tsit4te3 只處, and, as a result, the phrase tends to be interpreted as trivial background information. Consequently, te3 處 serving as background information gradually undergoes grammaticalization. Now let’s turn to the grammaticalization of te3 處 in imperatives. The examples in the following demonstrate the context which facilitates this process. (28) [春白]



待 tai7 wait

我 gua2 2sg

去 khi3 go

捧 phang5 hold.with.both.hand

來。 lai5 come

師父 sai1hu7 master

在 只。 (17.39–40 WL) 水 tsui2 ti7 tsit4 water prep dem ‘Wait me a moment. I will bring some water. Master, here is the water’.

Continuative aspect marker le(h)4  173 [生白]



小妹， 放 處， 請 sio2be7 pang3 te3 tshiann2 little.sister put te please ‘Lady, just put it here. Don’t bother’.

(29) [外]

甲 kah4 order ‘Kneel here’. [爭]



你 li2 2sg

只處 tsit4te3 prep dem-te ‘I will stand here’. 那 na3

跪 kui7 kneel

只處。 tsit4te3 dem-te

自在。 tsi7tsai7 be.at.ease

(5.581–585 SZ)

立。 khia7 stand

[外]



跪 處！ kui7 te3 kneel te ‘Kneel down’!

Based on the context in (28), we can easily understand that te3 處 in pang3te3 放 處 has its pre-grammaticalized counterpart as pang3hit4te3 放許處 ‘to put it there’. The demonstrative phrase hit4te3 許處 clearly points to tsit4 只 in the preceding sentence sai1hu7tsui2ti7tsit4 師父水在只. By the same vein, we know that kui7te3 跪處 in (29) indicates kui7tsit4te3 跪只處 ‘to kneel here’ before it was semantically bleached due to grammaticalization. And tsit4te3 只處 corresponds to hit4te3 只 處 in na3tsit4te3khia7 那只處立. In other words, either hit4te3 許處 in pang3hit4te3 放許處 or tsit4te3 只處 in kui7tsit4te3 跪只處 is a repetition of old information. Information like this, though it is not new, still works more or less for either the one given an order or the issuer of a directive. For the issuer of a directive, the specific location of the action instructed is relatively unimportant, compared with the action which is enjoined. Furthermore, in those examples above, these actions are also employed as a response to the action, phang5 捧 and khia7 立 respectively, in a preceding sentence. Consequently, we know that phrases such as pang3hit4te3 放許處 and kui7tsit4te3 跪只處 have their focus on the actions, khng3 放 and kui7 跪, instead of the locative hit4te3 許處 and tsit4te3 只處. The focus of this kind of sentences is shared by not only the issuer of a directive but also the receiver. Such kind of tacit and mutual understanding, therefore, further makes the old information parts, hit4te3 許處 and tsit4te3 只處, even more trivial. To some extent, the trivial part loses all its effectiveness. As a result, pang3tsit4te3 放許處 and kui7tsit4te3 跪只處 simply show up as pang3te3 放處 and kui7te3 跪處, and te3 處, which is void of content word meaning, then ends up in a trajectory of grammaticalization. Regarding the grammaticalization of te3 處 in objective existential sentences and imperatives, being-at-the-scene and high frequency of use are the two major

174  Manjun Chen motivations. Moreover, the respective grammaticalization of te3 處 in each of these two kinds of sentences also have influence on each other. After the grammaticalization mentioned above, te3 處 in subjective existentials and non-imperative imperfect sentences followed suit. This is what we can observe in (23) and (25). From the big picture, te3 處 in WL is halfway along the grammaticalization process. Therefore, in an identical context, it can either be interpreted as a continuative marker or with locative meaning. Relatively, the priority of the continuative reading verifies the ongoing grammaticalization. It is worth noting that the other continuative marker ti7te3 在處 in WL emerged later than te3 處. It is grammaticalized from ti7te3 在處 in objective existentials, such as (22), (26), and (27). In addition, unlike te3 處, ti7te3 在處 got grammaticalized only in some specific syntactic and semantic structures. This can be demonstrated by contrasting the usage of ti7te3 在處 before its grammaticalization with the usage after it went through the grammaticalization process. (22) 阿媽

a1ma2 old.lady



叫： kio3 shout

捧 phang5 hold.with.both.hand

入 lip8 into

內廳 lai7thiann1 inner.hall

去 khi3 to

食， tsiah8 eat

眾 群 媽 在處。 (10.54–55 WL) tsiong3 kun5 ma2 ti7te3 many group old.lady prep-te ‘The old lady ordered, ‘Take the food to eat it in the room inside. Here are many elderly ladies’.

(30) 李婆，

阮 li2po5 gun2 Madam.Li our

厝 tshu3 house

豬母 生 ti1bu2 sinn1 female.pig give.birth

年 ni0 perf

七 tshit4 seven

八 pueh4 eight

在處， 畏 無 向生 物。 (25.43–44 WL) ti7te3 kiann1 bo5 hiunn7sinn1 mngh8 clf prep-te afraid neg that.kind thing ‘Madam Li, my sow has had seven to eight pigs. We do have that kind of thing’. 個 e5



In (22), the semantic core lies in ti7 在, the verb as the predicate, which is unlikely to be grammaticalized with te3 處. Whereas, if ti7te3 在處 follows a verb predicate, then it becomes possible for ti7te3 在處 to be interpreted as background information, as we can see in (30). The shift of semantic core from ti7 在 in (22) to somewhere else than ti7 在 in (30) illustrates the possibility to use ti7 在 as a function word, instead of a content word. Based on this line of reasoning, we suggest that ti7te3 在處 underwent a transformation from ‘verb+continuative marker’ (in (22)) into an independent continuative marker (in (30)). Although these two both mark continuity, they differ in that te3 處 denotes nothing but temporal continuity but ti7te3 在處 additionally denotes a sense of spatial existence. We believe both of these two are still in an ongoing process of grammaticalization (see §6.2).

Continuative aspect marker le(h)4  175 6 TRANSFORMATION OF THE SOUTHERN MIN CONTINUATIVE MARKER Based on the discussion so far, the earliest example of Southern Min continuative marker is found in WL. In fact, at that time, it was still not a pure continuative marker as it still conveyed additional meaning. The emergence of a pure ­continuative marker took quite a long period of time.

6.1  The transformation of te3 處 The equivocality of te3 處 is not only observed in WL. In fact, between ce 1581 to 1884, from the publication of WL to GX, te3 處 was in an intermediate state of grammaticalization. The reason why the process took a relatively long time is pertinent to its syntactic position and the semantics of it. During this period of more than 300 years, the change with respect to the syntax and semantics of the continuative marker te3 處 is very limited. In most of the cases, this element was employed in an imperative and occurred sentencefinally, with a preceding monosyllabic bare verb, such as, pang3 放 ‘to set aside’, to2 倒 ‘to lie down’, kui7 跪 ‘to kneel’, he7 下 ‘to put down’, na3 那 ‘to be lost in thought’. The combinations of one of this kind of verb with te3 處 were mostly used as a predicate. If the subject was overtly realized, it was usually an Agent. Some examples are listed below. (31) 不用， 倒 處。 (11.59 SZ) m7bian2 to2 te3 not.necessary lie.down te ‘You don’t have to get up. Just keep on lying there’. (32) 汝 跪 處， 我 叫。 (16.393 GX) li2 kui7 te3 gua2 kio3 2sg kneel te 1sg yell ‘If you just kneel there, I will yell’. (33) 甲 汝 下 處。 kah4 li2 he7 te3 order 2sg put.down te ‘I asked you to put it down’.

(18.57 GX)

In this kind of sentence, sometimes we will find the adverbial tsit8e7 一下 denoting a short period of time. (34) 林婆， 你 且 住 處 一下。 (3.33 Sū Liùniáng) lim5po5 li2 tshiann2 tua3 te3 tsit8e7 madam.Lim 2sg temporarily stay te in.a.while ‘Madam Lim, please stay here temporarily’.

176  Manjun Chen Occasionally, te3 處 occurs between two structurally simple verb phrases between which a comitative relationship is held. By way of example: (35) 阿嫂， a1so2 sister.in.law



你 li2 2sg

且 tshiann2 please

去 入 來 lai5 khi3 lip8 enter come to ‘Dear sister, please sit and stay embroidery’.

那 na2 stay

處 te3

坐， tse7 sit

te

繡花 siu3hue1 embroider here. I will

我 gua2 1sg

一下。 (180 Jīnhuānǚ) tsit8e7 in.a.while go in to the room to do some

In a non-imperative imperfect sentence, the syntactic constituent ‘verb處’ is used as a subject (e.g., (36)), or an adverb preceded by a theme or a negative (e.g., (37) and (38)). Alternatively, this constituent can be followed by a temporal adverbial (as in (39)). (36) 我 gua2 1sg



想 siunn7 think

一下， tsit8e7 in.a.while

可 khah4 relatively

平宜？ pian3i5 worthwhile

可 平宜？ (16.394 GX) 跪 處 kui7 te3 khah4 pian3i5 kneel te relatively worthwhile ‘Wait a moment. Well, which is better? To be a ? or just to kneel down there?’

(37) 呼 hoo7

十公 tsap8kong1 ?

掃帚 放 sau3tshiu2 pang3 dsp broom put ‘Put down the broom’.

處。 te3

(17.029 GX)

te

(38) 簡

留 伊 不 帶 處， 亞娘 汝 去 留 伊。 (22.63 SZ) kan2 lau5 i1 m7 tua3 te3 a1niu5 li2 khi3 lau5 i1 maid keep 3sg neg stay te mistress 2sg go keep 3sg ‘I, as a maid, asked him to stay, and he refused. Madam, please ask him to stay’.

(39) 亞娘 a1niu5 mistress



那 na2 if

不 m7 neg

來 lai5 come

牽， khan1 hold.hands

跪 kui7 knee

處 te3 te

三 sann1 three

冥 mi5 night

五 日， 亦 不 起來。 (23.310 GX) goo7 lit8 ah8 m7 khi2lai5 five day still neg get.up ‘If Madam doesn’t come to pull me up, I will keep on kneeling here even for days’.

Continuative aspect marker le(h)4  177 In a subjective existential sentence, the phrase of ‘verb處’ occurs either in a purposive, as in (23), or between two verbs. For instance: (40) 我 gua2 1sg



處 te3 te

洗 sue2 wash

面， bin7 face

陳三 tan5sann1 Tan.sann

口店 tiam3 stay

處 te3 te

立 khia7 stand

看 khuann3 watch

阮， 甲 伊 立 開 去。 (11.87 SZ) khia7 khui1 khi3 gun2 kah4 i1 1sg ask 3sg stand away go ‘I am washing my face, and Tan-sann is there watching me. Ask him to go away’.

In examples from (31) to (40), the verbs preceding are location-related action verbs which are all semantically [+action] and [+location]. In other words, these verbs involve actions that cannot do without locations. Moreover, te3 處 began to follow an adjective in this stage, as previously mentioned, and in some cases it even occurs after a non-location-related action verb. See the examples below. (41) 阿娘 寬 處， 益春 就 去。 (26.29 WL) khuan1 te3 iah4tshun1 tsiu7 khi3 a1niu5 mistress be.at.ease te Iah.tshun just go ‘Madam, don’t worry. I, Iah-tshun, will go right away’. (42) 馬 亦 著 飼 處 度 三爹。 be2 ah8 tioh8 tshi7 te3 thoo7 sann1tia1 horse also have.to feed te for Sann.tia ‘We also have to keep the horse for Sann-tia’. (43) 別 pat8 other

物 mih8 thing

趕 kuann2 hurry

城 siann5 town

內 lai7 inner

去， khi3 go

去 khi3 go

(8.137 SZ)

買 bue2 buy

添 thiam1 add.on

也 罷。 (313.11 Jīnhuānǚ) ah8 suah4 te also finish ‘As for the other things, we can just buy them in the town. No big deal’! 處 te3



Recall that we said te3 處 was halfway along the grammaticalization process. This is exemplified by the sentences in which te3 處 can either be interpreted as a grammaticalized continuative marker or a locative content word. As for examples in which te3 處 follows an adjective such as (41), they are few and the adjective is unexceptionally khuan1 寬. Examples in which te3 處 co-occurs with a nonlocation-related verb are even scanty and (42) and (43) are the only two. In these sentences, te3 處 is a true continuative marker. Due to the small number of these

178  Manjun Chen examples, we cannot deny the semi-grammaticalized status of te3 處 at that time. Note in (31) to (40) there is no locative adverbial except for te3 處. This demonstrates that te3 處 still more or less has a locative sense in these sentences and consequently the grammaticalization process was deferred. Not until finally in the Qing dynasty did te3 處 go through the last rite of passage. The first example of te3 處 to co-occur with a locative adverb, to our knowledge, is in GX. Additionally, some similar examples can also be seen in Zheng 1999, Vol. 4:137, 214). They are listed in the following examples. (44) 諒 liong7 guess



去 khi3 go

手 tshiu2 hand

未 be7 neg

若 lua7 how

遠， hng7 far

句 koh4 still

在 ti7

許 hit4

prep

dem

戳 tshak8 stab

去， khi3 asp

不免 put4bian2 can’t.help.but

放 pang3 put

手袖 tshiu2ng2 sleeve

住 處。 (Yǐn Xíngyì Tóujǐng 尹行義投井) 來 5 7 3 lai tshu te come pad te ‘My hand was stabbed. I can’t help but put down my sleeve to have my hand on it’.

(46) 大同關 tai7tong5kuan1 stronghold.Tai.tong



亦 ah8 yet

內， 居 處 坐。 (39.25 GX) 戲房 hi3pang5 lai7 khu5 te3 tse7 theatre inside crouch te sit ‘I guess he has not gone far yet. He is probably still squatting in the theatre’.

(45) 阮 gun2 1sg.gen



伊 i1 3sg

正 tsiann3 exactly

是 si7 cop

漢 han3 Han

的 e5 gen

所在， soo2tsai7 place

通 那 處。 (Liú Wénlóng Mǎi Shāmào 劉 thang1 na7 te3 文龍買紗帽) can stay te ‘The stronghold, Tai-tong-kuan, is under the control of Han dynasty. We can stay there’.

As shown above, in (44) to (46), in addition to te3 處, there are locative adverbs including ti7hit4hi3pang5lai7 在許戲房內 ‘in that theatre’, tshiu2ng2 手袖 ‘sleeve’, tai7tong5kuan1 大同關 ‘the stronghold called Tai-tong-kuan’. Since the locations of the actions in these sentences are specified by these adverbs, the locative sense of te3 處 is deprived. And this development led to the further semantic bleaching of te3 處. Consequently, te3 處 gradually became a pure function word denoting continuity.

Continuative aspect marker le(h)4  179 Furthermore, based on the texts in Zheng 1999, more and more verbs irrelevant to a location came to co-occur with te3 處 as time went by. Below are some such kind of examples. (47) 你



li2 2sg

那 na2 if

歡喜， 阮 huann1hi2 gun2 happy 1sg

歡喜，

阮

就

穿 處， tshing7 te3 wear te

那 na2 if

不 m7 neg

脫 度 恁， 成乜 稀罕。 (Sū Qín Luò Cítáng 蘇秦落祠, Zheng 1999, Vol. 5:127) huann1hi2 gun2 tsiu7 thng3 thoo7 lin2 siann5mih8 hi1han2 happy 1sg.gen just take.off give.to 2sg what rare ‘If you like it, I’ll just wear it. If not, I’ll take it off and give it to you. Why do you think that I care’?

(48) 咱



而 bi7 then

lan2 1pl.incl

只遭 打 tsit4pai2 phah4 dem-time hit

伊 . . . 伊 i1 i1 3sg 3sg

必定 無通 諒 pit4ting7 bo5thang1 liong7 definitely cannot yield

咱， lan2 1pl.incl 二

咱 lan2 1pl.incl 人

會 e7 will 處

諒 liong7 yield 打。

必定 pit4ting7 definitely 鉗

無通 bo5thang1 cannot 就

伊， i1 3sg

(Lǚ Méngzhèng Rèn Jiǎojī 呂蒙正認腳跡, Zheng 1999, Vol. 5:238) lng7 lang5 khinn5 te3 tsiu7 phah4 two people grab te then hit ‘If I hit her this time . . ., she definitely cannot yield to me and vice versa. We’ll just grab and then fight’.

The increasing number of sentences like (47) and (48) apparently correlates with the multiplying of the kind including (44) and (46). The prevalence of the latter kind directly facilitates the further grammaticalization of te3 處. And the emergence of more and more sentences such as (47) and (48) is a result of this accelerating process, and also serves as evidence of it. During late Qing dynasty, in some copies of playscripts, a few cases in which te3 處 is replaced by leh4 咧 began to appear. Though we do not know precisely when these copies were made, it is quite clear that the use of leh4 咧 began later than the publication of GX (1884) for no such examples can be found in GX. Note that leh4 咧 usually co-occurs with a locative phrase when it follows a verb relevant to a location. (49) 廳

thiann1 hall

角 度 你 蹛 咧 立 好好， kak4 thoo7 li2 tua3 leh4 khia7 ho2ho2 corner let 2sg stay leh stand good

無 bo5 neg

因 in1 reason

180  Manjun Chen 嚷切切，



噪 阮 耳。 (Sū Qín Jiǎ Bú Dì 蘇秦假不第, Zheng 1999, Vol. 5:102) liong2tshiat4tshiat4 tsau3 gun2 hi7 yell.loudly    to.disturb.by.making.a.noise 1sg.gen ear ‘We let you stay well in the corner of the hall. However, it’s so intolerable for you to yell loudly without a reason’.

Only when there is a specific context, the concurrent locative phrase is dropped. Here’s an instance: (50) 阮 gun2 1sg.gen



小姐 sio2tsia2 lady

合 kap4 and

我 gua2 1sg

帶 tua3 stay

咧 leh4 leh

咱 lan2 1pl.incl

都 too7 all

有 u7 have

鬧熱 (Lǚ Méngzhèng Rèn Jiǎojī 呂蒙正認腳跡, Zheng 可 1999, Vol. 5:244) khah4 lau7liat8 more bustling ‘With my wife staying with me, everyone is having more fun here’.

Moreover, the continuative marker leh4 咧 also began to follow a verb irrelevant to a location, as previously mentioned. (51) 是 si7 cop



阮 gun2 1sg.gen

事， tai7 business

媒人 hm5lang5 matchmaker

錢 tsinn5 mondy

即 tsiah4

共 kang7

prt

appl

你 li2 2sg

留 咧。 (Zhū Mǎichén 朱買臣, Zheng 1999, Vol. 4:55) lau5 leh4 keep leh ‘This is my business. I’ll keep the matchmaker honorarium for you’.

Transformed in written form from te3 處 to leh4 咧, the grammatical function of the marker has correspondingly shed all its remaining locative sense and been turned into a pure continuative marker. This is demonstrated in (49) to (51), in which leh4 咧 has no connection to a location any more. Leh4 咧 emerged during the late Qing dynasty and was competing with te3 處 until leh4 咧 totally replaced it. And the final stage of this grammaticalization process came to an end after a long run. In the 34 issues of Yǔyuàn 語苑, which were published between 1908 to 1941 in Taiwan under Japanese rule, only leh4 咧 is found and te3 處 has totally disappeared.

6.2  The transformation of the continuative marker ti7te3 在處 To our knowledge, the continuative marker ti7te3 在處 had appeared as early as in the Southern Min play scripts published during the Wanli era in the Ming dynasty.

Continuative aspect marker le(h)4 181 In the play scripts published during the Ming dynasty, there are six examples of ti7te3 在處. Among these, five are in the Lì Zhī Jì and Jīnhuānǚ written in the Chaozhou dialect, and only one is found in Míngkān Mǐnnán Xìqǔ Xiánguǎn Xuǎnběn Sānzhǒng 1995, which is in the Quanzhou dialect. Comparing the usages in Chaozhou and Quanzhou, we found the degrees of grammaticalization of ti7te3 在處 are not in tandem. In the one example in Quanzhou, ti7te3 在處 is preceded by a location-related verb. Additionally, it does not co-occur with a locative phrase. Compared with the examples in Chaozhou, ti7te3 在處 in Quanzhou has a lower degree of grammaticalization. (52) 娘仔，



都

是 許 一 個 劉 厝 店 在處。 (Lǚ Yúnyīng Huāyuán Yù Liú Kuí 呂雲英花園遇劉奎) niu5a2 too7 si7 hit4 tsit8 e5 lau5 tshu3 tiam3 ti7teh4/te3 lady all cop dem one clf Lau home stay ti-te ‘Dear, it’s always Lau who’s been staying here’.

Although ti7te3 在處 in this example is a continuative marker that denotes the continuity of Lau’s staying, it also denotes some locative meaning, which may be interpreted as ‘here’. Now let’s turn to the five examples in Chaozhou. Among them, there’s also one in which ti7te3 在處 is preceded by the location-related verb tiam3 店 ‘stay’. Whereas, we find a locative phrase hit4te3 許處 ‘there’ in this sentence, too. The presence of the locative phrase implies that ti7te3 在處 in this case denotes more about the continuity of the two lambs’ staying there. As a side note, ti7te3 在處 in this sentence not only functions as a continuative marker but also as a progressive marker. Due to economy, another ti7te3 在處 which serves as a progressive aspect marker is omitted. (53) 阿嫂， a1so2 sister.in.law



許處 hit4te3 dem-te

夭 iau2 still

兩 lng7 two

個 e5 clf

羊子 iunn5a2 lamb

店 tiam3 stay

在處 ti7te3 ti-te

食 水 在 許。 (174.6 Jīnhuānǚ) tsiah8 tsui2 ti7 hit4 eat water prep dem ‘Dear sister-in-law, two lambs are still staying and drinking water over there’.

As for the other three examples, the main verbs preceding ti7te3 在處 are nonlocation-related ones, including hong1 封 ‘to seal’, am1 掩 ‘to half close’, and siu3 繡 ‘to embroider’. Among the four verbs, only am1 掩 occurs in a bare form, and the others all take either a complement or an object.

182  Manjun Chen (54) 是

小妹， 門 掩 在處， 挨 開 來。 (21.19 WL) sio2be7 mng5 am1 ti7te3 ue1 khui1 lai5 cop little.sister door half.close ti-te push open come ‘It’s my little sister. The door stands ajar. Just push it open’. si7



(55) 小七，

你 sio2tshit4 li2 Sio.tshit 2sg 呾： tann3 say



許 hit4 dem

內， lai7 inside

共 kang7

包 pau1 pack

銀 gin5 silver

封 便 hong1 pian7 hairpin handy

appl

恁 媽 lin2 ma2 2sg.gen grandma 在處， ti7te3 ti-te

持 出來 插 媒姨 花 一下。 (10.44–46 WL) theh8 tshut4lai5 tshah4 hm5i5 hue1 tsit8e7 take out-come stick.in matchmaker pretty for.a.while ‘Sio-tshit, go in there and tell your grandma: “There’s a pack of silver hairpins in the room. Take them out for the matchmaker to decorate herself” ’.

(56) 我 gua2 1sg



一 tsit8 one

去 khi3 go

繡 有 siu3 u7 embroider have

春 夏 秋 冬 在處， tshun1 ha7 tshiu1 tang1 ti7te3 spring summer autumn winter ti-te

今 正 起頭。 (151.12 WL) tann1 tsiann3 khi2thau5 now just begin ‘I’ve embroidered “spring, summer, autumn, and winter” over here. I just began the work for a moment’.

Regarding these three sentences, (54) emphasizes that the door is in a half-opened state; unlike (54), there’s a locative phrase hit4lai7 許內 ‘in there’ in (55) modifying the silver hairpins, consequently ti7te3 在處 focuses on the existence of the packed hairpins; in (56), the location where the embroidered cloth is not specified, but from the context we know the conversation happens outside of the embroidering room and it is the embroidered cloths with patterns of spring, summer, autumn, and winter, that have their attention. We can, therefore, conclude that ti7te3 在處 was highly grammaticalized in Chaozhou play scripts published during Ming dynasty. Not until play scripts published during Qing dynasty can we find ti7te3 在處 to co-occur with a tsuan5 泉-locative phrase. And the co-occurrence indicates the relatively higher degree of grammaticalization of this marker at that time. Below are some examples: (57) 你



都 驚畏 袂記得， 各 記 上 簿 在處。 (Zhū Wén Sù Diàn Diǎn Dēng 朱文宿店點燈, Zheng 1999, Vol. 3:317) li2 too7 kiann1ui7 be7ki3tit4 koh4 ki3 siong7 phoo7 ti7te3 2sg foc afraid neg-remember even note up book ti-te ‘You even were afraid of forgetting it, and noted it on a book’.

Continuative aspect marker le(h)4 183 The verb ki3 記 ‘remember’ in the sentence above is a resultative verb. Unlike the location denoting noun phoo7 簿 ‘book’, the sentence-final ti7te3 在處 only indicates the continual existence of what is noted down, but not any locative sense. It is noteworthy that there are two different routes with respect to the grammaticalization from ti7te3 在處 to leh4 咧. The first is through the weakening of the spatial existential meaning of ti7te3 在處 in order to facilitate the dropping of ti7 在 and, then, the remaining te3 處 was further grammaticalized into leh4 咧. Examples in which ti7te3 在處 got grammaticalized into te3 處 began to show up in Quanzhou play scripts published during the Qing dynasty. And they are found in two kinds of syntactic contexts: one has a non-location-related verb with the continuative marker and the other has a location-related verb with the continuative marker. Here are some instances of the former: (58) 庵主，



笑處 在 許 塗腳 處 看 (Fàn Jū Rèn Zǐ 范雎認子, Zheng 1999, Vol. 5:323) am1tsu2 li2 khuann3 tshio3te3 ti7 hit4 thoo5kha1 te3 abbess 2sg see laugh-te prep dem floor te ‘Abbess, see! She is rolling on the floor laughing’. 你

(59) 更 深 了， 花園 門 都 開 (Yǐn Xíngyì Tóu Jǐng 尹行義投井, Zheng 1999, Vol. 4:137) kinn1 tshim1 looh4 hue1hng5 mng5 too7 khui1 two-hour.period deep perf garden gate all open ‘The night is late, and the gates of the garden are all left open’.

處。 te3 te

The verbs in (58) and (59), tshio3 笑 ‘laugh’ and khui1 開 ‘open’, are a one-place intransitive verb and a two-place transitive verb respectively. Both of them are non-location-related verbs, but there are differences between these two. Note that the continuative marker does not denote the continuity of the action of the verb tshio3 笑, even though it is a duration verb. Instead, the marker merely signifies the state brought about by the action continues. This state needs a spatial domain for it to be revealed which is specified by the locative phrase ti7hit4thoo5kha1 在 許塗腳 ‘on the floor’. As a result, the location-related preposition ti7 在 in ti7te3 在處 becomes redundant and is finally dropped. On the other hand, what the verb khui1 開 is pertinent to is an object, but not a spatial domain. Therefore, (59) has nothing to do with the spatial existence of the opening of the garden gates, but its temporal continuity. In the same vein, the redundant 在 in ti7te3 在處 is then consequently dropped. Now let’s turn to the objective existential construction. In this construction, the grammaticalization of the marker ti7te3 在處 into te3 處 can also be triggered when the verb modified by this marker is a location-related verb and there is a locative phrase in the sentence. By way of examples:

184  Manjun Chen (60) 我



娘仔 一 人， 帶 許 邊頭 處 啼 (Sū Qín Jiǎ Bú Dì 蘇秦假不第, Zheng 1999, Vol. 5:105) gua2 niu5a2 tsit8 lang5 tua3 hit4 pinn1thau5 te3 thi5 1sg lady one person stay dem side te cry ‘My wife is staying there and crying alone’.

That the predicate of the sentence in (60), tua3 帶, is followed by a locative phrase hit4pinn1thau5 許邊頭 indicates the locative sense pertinent to the action tua3 帶 is well realized. As a result, the locative sense of the preposition ti7 在 in ti7te3 在處 is suppressed and gets dropped. Moreover, te3 處 in (60) differs from the ones in (44)-(46) in the degrees of grammaticalization. Note the locative phrases in (44)(46) all precede the verb+處 combination, whereas the locative phrase in (60) is inserted between the verb and te3 處. Here is what we suggest: once the locative phrase precedes the verb+處 combination, it is relatively separated from te3 處. Consequently, it has a weaker interaction with te3 處. Even so, te3 處 still more or less retains its locative sense. However, so long as the locative phrase is embedded in the combination, it is adjacent to te3 處 and its locative sense is completely deprived. That being so, te3 處 tends to specialize in denoting continuity, instead of being simultaneously locative. After ti7te3 在處 evolved into te3 處, it underwent further transformation to become leh4 咧 in the objective existential construction. A handful of examples have appeared in Quanzhou play scripts during late Qing dynasty. (61) 引公， 你 免 驚， 伊 跪 咧 咯。 (Liáng Hào Bàigōng Xúnzǐ 梁灝拜公訓子, Zheng 1999, Vol. 5:178)8 in2kong1 li2 bian2 kiann1 i1 kui7 leh4 looh4 grandpa 2sg not.have.to be.afraid 3sg kneel leh prt ‘Grandpa, you don’t have to be afraid. He’s already kneeling there’. Aside from the route of grammaticalization described so far, we also found ti7te3 在處 being grammaticalized into leh4 咧 via another course by retaining ti7 在 and evolving into ti7leh4 在咧, followed by a later development which produced leh4 咧. While this is rarely found in the Quanzhou play scripts during the Qing dynasty, quite a number of this kind of example is observed in Yǔyuàn, published in Taiwan under Japanese rule. Below are some examples of ti7leh4 在咧: (62) 藏 真 多 書 tsue7 tsheh4 khng3 tsin1 store degree many book ‘Many books are stored here’.

在咧。 ti7te3/leh4 ti-te

(2.33.8 Yǔyuàn 18)

Continuative aspect marker le(h)4 185 (63) 請 tshiann2 please 的 e5

入來 內面， 三 號 與 四 號 lip8lai5 lai7bin7 sann1 ho7 kah4 si3 ho7 enter inside three number and four number 箱 開 開 在咧。 (6.6.5–6 Yǔyuàn 16) siunn1 khui1 khui1 ti7te3/leh4 lnk box open open ti-te ‘Please come in. Both box number 3 and box number 4 are open’.

船仔 (64) 門腳口 泊 mng5kha1khau2 pok8 tsun5a2 doorway anchor boat ‘There’s a boat out of the door’. (65) 老





弱

日日新

的

報社，

竟然

(8.20.1 Yǔyuàn 7)

活 在咧 (3.45.6–7 Yǔyuàn 5) lau7 e5 liok8 e5 king3jian5 uah8 uah8 ti7te3/leh4 old lnk weak lnk unexpectedly live live ti-te ‘Unexpectedly, the old and the weak are still alive’.

(66) 現時

的、

在咧 ti7te3/leh4 ti-te

活

有

幾

個

記者 在咧。 (1.39.11 Yǔyuàn 19) hian7si5 jit8jit8sin1 po3sia7 u7 kui2 e5 ki3tsia2 ti7te3/leh4 now Jit.jit.sin newspaper.office have several clf reporter ti-te ‘Now there are several reporters in the Jit-jit-sin News office’.

The environments where ti7leh4 在咧 appears in Yǔyuàn can be categorized into four groups. In the first, ti7leh4 在咧 occurs to follow a location-related verb phrase including an object and no locative phrase appears (see (62)). Otherwise, ti7leh4 在咧 may follow a bare non-location-related verb or a verb phrase including a non-location-related verb and an object. In this case, there would be a locative phrase that either precedes the subject (see (63)) or serves as a locative subject (see (64)). Apart from the two groups, ti7leh4 在咧 may also follow a verb of existence and appearance without an accompanying locative phrase, as shown in (65). Lastly, ti7leh4 在咧 may be found to follow the verb phrase containing the verb of existence and appearance u7 有 and an object, with a locative phrase leading the whole sentence (demonstrated in (66)). To summarize, all these sentences are characterized by emphasizing the continual spatial existence in addition to the temporal continuity of the state resulting from the execution of the action of the verb or the state regarding the verb itself. Note ti7 在 in ti7leh4 在咧 can indicate the continual spatial existence of a certain state and, on the other hand, leh4 咧 signifies the temporal continuity of it. As long as there is no other locative phrase

186  Manjun Chen in the sentence, as seen in (62) and (65), the occurrence of ti7leh4 在咧 meets the need of the semantic requirement. With respect to the kind of sentences in (63), (64), and (66), a distance is kept between the locative phrase and the continuative marker and, relatively, the continuative marker is less constrained semantically. Whenever there is a need to emphasize the spatial existence of a person or an object in the sentence, 在 in ti7leh4 在咧 becomes the assistant. The closer the noun of the person/object is to ti7leh4 在咧, the stronger the spatial existential meaning is, and vice versa. This is exemplified by the contrast between (64) and (66) as a group and (63). In the former, the spatial existential meaning of the boat and reporters is stronger than the box in the latter. After ti7te3 在處 was grammaticalized and became ti7leh4 在咧, it usually appears post-verbally as a temporal and spatial continuative marker, with ti7 在 pertinent to the spatial sense and leh4 咧 to the temporal. However, when ti7leh4 在咧 is followed by a locative phrase, its function would change. When a locative phrase follows ti7leh4 在咧 in a sentence, semantically ti7leh4 在咧 will not connect with the preceding verb, but connect with the following locative phrase instead. Syntactically, ti7leh4 在咧 in this case is not a continuative marker for the verb. It has been transformed into a preposition and constitutes a prepositional phrase with the locative phrase. By way of examples: 囝仔 (67) 彼個 查某 tsa1boo2 gin2a2 hit4e5 dem-clf female child ‘That girl is sitting there’.

坐 tse7 sit

佇咧 ti7leh4 ti-leh

彼 hia1 there

(120.15 Changhua 7)

Since ti7leh4 在咧 in this case combines with the locative phrase first, only ti7 在 still has a semantic function. As for leh4 咧, it becomes a semantically bleached suffix. Besides its evolving into a prepositional element, the continuative marker ti7leh4 在咧 also further developed some functions which are even more abstract. In the following examples, the degree of grammaticalization of ti7 在 is even higher. (68) 羊



不過

是

有

淡薄仔

孝心

在咧 而已 (1.19.3 Yǔyuàn 13) iunn5 put4ko3 si7 u7 tam7poh8a2 hau3sim1 ti7te3/leh4 nia7nia7 goat nothing.but cop have a.bit.of filial.piety ti-te only ‘What goats have is nothing but a little filial piety’.

(69) 主顧 吩咐 的 工課， 攏 是 有 tsu2koo3 huan1hu3 e5 khang1khue3 long2 si7 u7 client instruct lnk task all cop have 在咧 (4–13–10, 11 Yǔyuàn 16) 日子 jit8tsi2 ti7te3/leh4 date ti-te ‘The tasks required by our clients all have their deadlines’.

定 tiann7 designate

Continuative aspect marker le(h)4  187 The filial piety in (68) only exists in our mind and is intangible. And what ti7 在 denotes is the existence of such a kind of intangible thing. Compared with its usage in (62) to (66), the locative sense involved here is more abstract, not pertinent to a concrete space anymore. Therefore, what the continuative marker ti7leh4 在咧 mainly denotes is the temporal continuity coded by leh4 咧. Now let’s turn to (69). In this sentence jit8tsi2 日子 indicates a time point. And ti7leh4 在 咧 denotes that the deadline is always there since it has been designated. Hence, what is relevant here can only be temporal continuity and ti7 在 in ti7leh4 在咧 has been totally grammaticalized. It is this development which facilitates ti7leh4 在咧 to further become leh4 咧, forfeiting their locative sense and keeping the temporal continuity sense only.

6.3 The development after te3 處 and ti7te3 在處 were grammaticalized into leh4 咧 To recapitulate, the grammaticalization from te3 處 to leh4 咧 started as a bi-­ functional locative and temporal continuative marker and developed towards a pure temporal marker. By way of contrast, another continuative marker ti7te3 在 處 went via two different routes in its grammaticalization into leh4 咧. Both of them start as a marker which can either denote the locative sense or both locative and temporal senses. On the first route, this marker first became ambiguous without exclusively specializing in either the locative or temporal function and then became a pure temporal continuative marker. The development stages are: ti7te3 在處→ te3 處→ leh4 咧. The same marker, however, evolved into a bifunctional temporal and locative marker on the second route and followed by a stage in which it was grammaticalized into either a preposition or a pure temporal continuative marker. And the markers were transformed stage by stage as ti7te3 在 處→ ti7leh4 在咧持續→ ti7leh4 在咧介 or ti7te3 在處→ ti7leh4 在咧→ leh4 咧. As a consequence, both imperfect and perfect leh4 咧 are markers denoting temporal continuity. Typically, this marker follows a verbal or an adjectival predicate immediately which has nothing to do with existence or appearance, without a locative phrase or an object. Nonetheless, when the predicate is a verb or adjective phrase of existence and appearance, or when the verbal predicate is modified by a locative phrase or an object, in addition to leh4 咧, the meaning of leh4 咧 will be weakened depending on the elements present in the sentence and its syntactic position. By this reasoning, we suggest that the grammaticalization of the continuative marker leh4 咧 is a multi-path development. No matter whether it is imperfect or perfect, leh4 咧 undoubtedly has become a marker denoting the temporal continuity of a certain state. When it follows a predicate that is irrelevant to existence and appearance, it is deemed a typical continuative marker due to maintaining of the state, which has either resulted from the finishing of the action or simply belongs to the nature of the action itself and is impossible to be conveyed without leh4 咧. For instance, kuinn1 關 ‘to close’ is only an action. When it combines with leh4 咧, the state maintained after the completed action is expressed. This can also be exemplified with the predicate

188  Manjun Chen ang5 紅 ‘red’. The predicate denotes a characteristic. Only when this predicate combines with leh4 咧, the meaning of a continual state of being red existing can be conveyed. On the other hand, once what is followed by leh4 咧 is a predicate of existence and appearance, the meaning of leh4 咧 will get weakened. Below are some examples: (70) 雷公 卻 lui5kong1 khiok4 thunder.god but



咧，

有 u7 have

咧， leh4 leh

雷母 lui5bu2 thunder.goddess

不曾 m7bat4 never

講 (9.33.6 Yǔyuàn 9) kong2 say of thunder, but have never heard of the goddess of

伊 做堆。 (10.17.8 Yǔyuàn 13) iah8bue7 leh4 kin1ni5 tang1 bue2 tio7 beh4 hoo7 i1 tso3tui1 not.yet leh this.year year end foc will let 3sg be.together ‘Not yet. But we will let her be together at the end of this year’.

(72) 伊 i1 3sg



cop

聽 人 thiann1 lang5 hear people ‘I know about the god thunder’.

(71) 尚未



是 si7

今年

本身 pun2sin1 self

冬

腳 kha1 foot

尾， 即

就 tio7

無 bo5

foc

neg

要

給

好 ho2 good

這個 溝仔 底 安呢 行 落 an2ne1 kiann5 loh8 tsit4e5 kau1a2 te2 this.way walk down dem-clf ditch bottom ‘His feet are bad and he dared not walk into the ditch’.

咧， leh4 leh

毋敢 m7kann2 not.dare

(30.01 Xinshe Township 2)

Regarding the three examples above, each of them has a verb of existence and appearance (u7 有 and bue7 未) or adjective of existence and appearance (bo5ho2 無好). These verbs and adjectives denote the continuity of a certain state without the assistance of leh4 咧. However, the continuity of a state expressed by these words is relatively implicit. Only if leh4 咧 occurs with them, the continuity reading becomes conspicuous. Such being the case, we suggest leh4 咧 in these sentences not only denotes continuity but also serves as an intensifying function. That is to say, leh4 咧 at this stage is ready to be further grammaticalized into a mood particle. As discussed previously, after ti7te3 在處 and te3 處 have been grammaticalized into leh4 咧, the marker leh4 咧 denotes only temporal continuity and has no more pertinence to spatial existence. Whereas, it is quite common for the action of a verb

Continuative aspect marker le(h)4 189 to involve some specific location, and this brings about the co-occurrence of a locative phrase and the continuative marker leh4 咧. Two possible positions are available for the locative phrase following a verbal predicate when leh4 咧 is also employed in the sentence: one is to precede leh4 咧 and the other is to follow it. For example: (73) 若



紙面 咧 (4.43.6 Yǔyuàn 21) na2 beh4 sia2 ji7 tioh8 sia2 tsua2bin7 leh4 if want write word have.to write paper.surface leh ‘If you want to write, you have to write on a paper’.

(74) 假使



要

寫

字，

著

寫

佇 牛稠 咧 (20.18 Changhua 19) ka2su2 am3si5 gu5 tio7 long2 pak8 ti7 gu5tiau5 leh4 if night bull foc all tie prep bull.pen leh ‘During the night, bulls are all tethered in the pen’. 暗時， 牛

就

攏

縛

(75) 此個 明明 是 阮 丈夫 喇， 因何 tsit4e5 bing5bing5 si7 gun2 tiong7hu1 lah4 in1ho5 dem-clf evidently cop 1sg.gen husband prt why 咧   (2.82.8–10 Yǔyuàn 21) 此 能 死 在 e7 si2 ti7 tsia1 leh4 will die prep here leh ‘This is evidently my husband. Why did he die here’? (76) 啊， 有影， 我 來 僻 咧 門扇後 (Liu QuanJin)9 ah4 u7iann2 gua2 lai5 phiah4 te3/leh4 mng5sinn3au7 prt true 1sg come hide teh door.plank.back ‘Ya! You’re right. Let me hide behind the door’. (77) 啊 ah4



去 歇 khi3 hioh4 prt go rest 尋， 伊 來 lai5 tshue7 i1 come look.for 3sg ‘When he rests under a escapes’.

咧 leh4

樹仔腳 咧， 啊 人 tshiu7a2kha1 leh4 ah4 lang5 leh tree.foot leh prt person 就 閃 咧。 (120.08 Changhua 7) tio7 siam2 leh4 then escape leh tree and someone comes to look for him, he just

As the degree of grammaticalization of the continuative marker leh4 咧 got even higher, leh4 咧 began to follow a locative phrase more often, and the sentences where it occurred are mostly imperfect (e.g., (73)). Sometimes it also appeared

190  Manjun Chen in an objective existential construction. This being the case, the locative phrase co-occurring was usually constructed as a prepositional phrase, 在+locative noun. With regard to the verbal predicate, it may be stative or non-stative, as shown in (74) and (75) repectively. Note that leh4 咧 following a locative phrase could vary in its degree of grammaticalization based on the preceding locative phrase. We observe the degree of grammaticalization of leh4 咧 in (75) is higher than the one in (74) and leh4 咧 in (74) is higher than in (73). What is pivotal here is the extent of interaction between leh4 咧 and the verb phrase and the semantic role played by leh4 咧. For the constituent ‘verb+locative phrase’ in sentences like (73), being with or without 咧 can make a big difference for its meaning. A bare constituent ‘verb+locative phrase’ only signifies where a certain action is carried out, without specifying the result state after the action is done. Whereas, when the constituent is combined with leh4 咧, the continuity of the post-action state is denoted. This kind of usage of leh4 咧 is a typical continuative marker. Unlike (73), (74), and sentences of its kind, the constituent ‘verb+在+locative phrase’ without leh4 咧 still denotes the existence of the state after the action, even though the continuity sense is not strong. However, when leh4 咧 is added to it, the continuity of the state becomes salient, and the emphasis is shifted from the existence of a time point to the existence of a time interval. Therefore, leh4 咧 in this kind of sentence is not only a continuative marker, but also a mood particle. As for (75), this is of another additional kind. No meaning difference with respect to continuity is observed between the presence or absence of leh4 咧 accompanying the ‘verb+在+locative phrase’ constituent. Adding leh4 咧 merely brings about a different mood. We maintain that the combination of the verb and the prepositional phrase denotes a continuing state after a specific action in an implicit way. It is the presence of leh4 咧 that makes the continuity meaning obvious. That’s why we suggest the main function of leh4 咧 in this kind of sentence is pertinent to mood, instead of marking continuity. From (73) to (75), we observe a weakening of the interaction between leh4 咧 and the preceding verb phrase. Diachronically, the continuative marking function of leh4 咧 is gradually lost and the data demonstrates a spectrum of a continuative marker developing into a pure mood marker. In addition to its following position, we also find leh4 咧 inserted between the verb and the locative phrase, resulting in a sequence as ‘verb+咧+locative phrase’. This kind of construction is in fact a later development. Wang (2010:87) proposes that leh4 咧 serves as a locative preposition in this construction and joins the following locative phrase to form a prepositional phrase, which in turn becomes the complement of the verb. We disagree with Wang’s analysis. For a parallel sequence in Mandarin Chinese, Wang’s claim seems to be reasonable. Whereas, for Southern Min speakers, leh4 咧 is intuitively more like an element with the verb, but not a prepositional head introducing the following locative phrase. In other words, the sequence should be considered by taking the verb and leh4 咧 as a constituent. In our opinion, the verb interacts not only with leh4 咧 but with the locative phrase. The adequate analysis of the sequence ‘verb+咧+locative phrase’ is a bipartite structure composed with ‘verb+咧’ and ‘verb+locative phrase’. By way of example: leh4 咧 in (76) is combined

Continuative aspect marker le(h)4 191 with the verb phiah4 僻 directly, instead of the locative phrase mng5sinn3au7 門扇後. Therefore, the chunk phiah4leh4mng5sinn3au7 僻咧門扇後 should be decomposed into ‘phiah4 僻 + leh4 咧’ and ‘phaih4 僻 + mng5sinn3au7 門扇後’. In Southern Min, leh4 咧 has been used as a continuative marker following a location-related action verb. It conveys the meaning that the state continuously exists somewhere after the action is done. It is the occurrence of a locative phrase that semantically retains 咧. Additionally, the position of leh4 咧 which precedes the locative phrase associates leh4 咧 with the prepositional function and diminishes its continuity-marking denotation. In examples like (77), the dilution of the continuative aspect denotation of leh4 咧 even triggers the use of an extra leh4 咧 in order to make up for the semantically incapacitating effect of the original one. Comparing the two leh4 咧 in (77), the second one obviously has a stronger sense as a continuative marker. Moreover, the second one is also a mood particle. In cases where a transitive verb takes the continuative marker leh4 咧 and a non-locative object simultaneously, leh4 咧 can only occur after the verb phrase containing the object. From the historical documents we know in the beginning that verbs in these cases can merely be volitional verbs. Non-volitional verbs becoming possible is a later development. Even so, the degrees of grammaticalization of leh4 咧 differ depending on the kind of verbs it co-occurs with. For instance: 咧。 (9.4.3 Yǔyuàn 5) (78) 尚 汝 著 與 我 包 紙 li2 tioh8 kap4 gua2 pau1 tsua2 leh4 ia2 also 2sg have.to and 1sg wrap paper leh ‘And you also have to help me wrap it with paper’. (79) 耳孔 那 塞 破布 咧， 准 無 hinn7khang1 na2 that4 phua3poo3 leh4 tsun2 bo5 earhole like stuff rag leh behave.as neg 聽見。 (3.66.5 Yǔyuàn 24) thiann1kinn3 hear ‘He behaved as if his earholes were stuffed with rags and could not hear’. (80) 啊 ah4 prt



今 tann1 now

也 a7 also

是 si7 cop

阮 gun2 1sg.gen

老父 lau7pe7 father

娶 tshua7 marry

某， boo2 wife

生 sinn1 give.birth

二 個 子 咧 (36.08 Shigang 1) lng7 e5 kiann2 leh4 two clf son leh ‘Then it was my father who got married again. They even have two kids now’.

192  Manjun Chen (81) 啊 ah4

後面 擱 有 果園 au7bin7 koh4 u7 kue2hng5 prt dem back additive have orchard ‘There is even an orchard there at the back’. 彼 hit4

咧 leh4

(10.12 Shalu 2)

leh

In the sentences containing a sequence of ‘verb+object+咧’, as exemplified above, if the verb is a volitional verb, the verb would denote an action no matter whether it takes an object or not. On the other hand, the (im-)perfectivity determines whether leh4 咧 is a typical continuative marker or not. Only in an imperfect sentence would leh4 咧 serve as a typical continuative marker, as seen in (78). In (78), what leh4 咧 denotes in pau1tsua2leh4 包紙咧 is identical in a phrase without the object, pau1leh4 包咧: the state after the wrapping continually exists. However, if the sentence is perfect (including irrealis perfect sentences), the continuity meaning of leh4 咧 would be weakened, as shown in (79). The reason is that the perfect aspect signifies the relevant action has happened and, moreover, that the involved object would be in a certain state after the action of the transitive verb is done. Both imply a certain state is already in place. And leh4 咧, in addition to making the existence of that state conspicuous, would add the continuity meaning to the state. In cases like this, leh4 咧 mainly functions as a continuative marker, but it also serves as a mood particle. It is in Xiàndài Táiwān Mǐnnányǔ Gùshìjí (Modern Southern Min Stories in Taiwan) that we find non-volitional verbs begin to precede leh4 咧. When leh4 咧 is preceded by a non-volitional verb, its degree of grammaticalization is even higher. By way of example: leh4 咧 in (80) is more grammaticalized than the one in (79). When leh4 咧 immediately follows sinn1 生 ‘to give birth’, it does not denote the state of giving birth continuously exists. Instead, it indicates the children born remain continuously there. Obviously, leh4 咧 occurs here as a continuative marker and nothing else. Whereas, when leh4 咧 follows a verb phrase with an object, as we can see in (80), the object lng7e5kiann2 兩個子 ‘two kids’ partly takes the place of the function of tsiam2 佔 with respect to state denoted. And leh4 咧 takes on extra function regarding a tone of confirmation. Consequently, leh4 咧 is turned into a bifunctional element as not only a continuative marker but also as a mood particle. It is noteworthy that the degree of grammaticalization found for leh4 咧 may even vary depending on individual verbs. For leh4 咧 that occurs right after a verb phrase including an object and a verb of existence and appearance, such as u7 有 and bo5 無, it enjoys the highest degree of grammaticalization and has evolved into a pure mood particle. This is illustrated in (81). In (81), with the phrase u7 有 + kue2hng5 果園 ‘have+orchard’ it is more than sufficient to express the continuity of existence. Once leh4 咧 co-occurs, it at most works as an emphasizer. We, therefore, propose leh4 咧 in this kind of sentence is further grammaticalized into a mood particle. 7  CONCLUDING REMARKS In this study, we propose the Southen Min continuative marker leh4 咧 has its origin in the continuative markers te3 處 and ti7te3 在處. These two markers were

Continuative aspect marker le(h)4 193 already employed in WL, which was published during the Wanli era (1572–1620) of the Ming dynasty. The continuative marker te3 處 was produced from the grammaticalization of the content word te3 處. After the grammaticalization, te3 處 still kept its original local sense and sometimes could even denote a relatively solid meaning of a specific location. On the other hand, ti7te3 在處 came into existence from the structure of ti7 在verb + te3 處. It evolved via the grammaticalization of ti7 在verb + te3 處 followed by another grammaticalization process targeting the whole structure. Just like its former counterpart, it could also denote a relatively solid meaning of a specific location. The transformation of these two markers took a long period of time before they became pure continuative markers. The emergence of leh4 咧 is a turning point for the grammaticalization of te3 處 to arrive at its culmination. Leh4 咧 began to be used in the late Qing dynasty. It competed with te3 處 and then gradually substituted for it. We also suggest that the grammaticalization of ti7te3 在處 into leh4 咧 has gone through two different routes. Both of them started as a bifunctional marker denoting either only a locative sense or both the locative and temporal senses. For one route, the marker evolved into a marker merely with a temporal sense, with its stages of development as ti7te3 在處→ te3 處→ leh4 咧. For the other, the marker evolved into a marker denoting both locative and temporal senses before it was grammaticalized into a pure marker with respect to time. Here are the stages it went through: ti7te3 在處→ ti7leh4 在咧→ leh4 咧. Additionally, ti7te3 在處 also developed into a preposition, via several stages in the order of ti7te3 在處→ ti7leh4 在咧continuative→ ti7leh4 在咧preposition. After te3 處 and ti7te3 在處 are grammaticalized into leh4 咧, the process was not halted. Leh4 咧 kept on its path of transformation until its continuity denotation was gradually weakened. In some cases, it becomes a pure mood particle. We reconstruct the emergence and grammaticalization process of the marker leh4 咧 in Figure 8.4. (S: spatial, T: temporal, M: mood) To sum up, the Southern Min continuative marker leh4 咧 should neither be deemed as a grammaticalized product from ti7te3 在處 (the AB-form) nor as the

Demonstrative+處

處 S or T

在+(Demonstrative+處)

在 verb+處 S

在 prep+處/ 處/ 在處 S and T S

處 S or T 在 S and T Prep.

Figure 8.4  Grammaticalization process of the marker 咧

T

T and M

M

194  Manjun Chen subsequent simplified form respectively. Contrary to this view, we suggest it has different origins in the existential construction and the non-existential construction. In non-existential sentences, the imperfect leh4 咧 evolved from B (te3 處), and this B must not be wrongly deemed as a simplified AB. On the other hand, the leh4 咧 denoting objective existence in an existential construction originated from AB (ti7te3 在處). Here leh4 咧 is not a mere simplified form of AB, either. In fact, there is a process of evolution which should be distinguished from the AB form. Lastly, when leh4 咧 is used to express personal emotions, it has its origin in B and does not co-occur with A. That is to say, the grammaticalization process of leh4 咧 involves different routes depending from where and what it began with. NOTES 1 As pointed out by Liu (1996:26), A in this AB-form can be used solely as a locative verb and preposition and it is synonymous to zài 在 in Mandarin Chinese. On the other hand, B in this AB-form is a locative suffix which cannot be used alone. B is roughly equals to lǐ 裡 and shàng 上 in Mandarin Chinese. For example: zhè lǐ 這裡 ‘here’, nà lǐ 那裡 ‘there’, fángjiān lǐ 房間裡 ‘in the room’, xīn lǐ 心裡 ‘in mind’, běnzi shàng 本子上 ‘on the notebook’, lìshǐ shàng 歷史上 ‘in the history’. 2 This study is granted by Agence Nationale de Recherche under the project DIAMIN N0 ANR-08-BLAN-0174. An early version was presented in the conference Reflections of Diachronic Change Mirrored in Early Southern Min (November 26–27, 2011 at National Tsing Hua University) and the author is grateful to Hilary Chappell, Alain Peyraube, and Chinfa Lien for their useful comments and suggestions. All errors remaining are mine. We are very grateful to Mr. Seng-hian Lau 劉承賢 of National Tsing Hua University for his excellent translation into English of the original Chinese. 3 Though Lü dubbed zài . . . lǐ 在. . .裡, which corresponds to the AB-form, and its simplification as particles, it is in fact used to express continuation. 4 The romanization and glosses are mine. 5 The Romanization and glosses are mine. 6 I speculate that 遐 [hia55] is a contracted form from 迄搭 [hit4 taʔ4] ‘that place’. 7 This kind of usage is still found in today’s Southern Min. For example: the phrase 阮處 ‘our place’, or the sentence such as 到處甲汝拍電話 ‘I will call you after I arrive there (in that place)’ are heard in Shishi, Quanzhou (泉州石獅), and Xinglin, Amoy (廈門杏 林). 處 in these cases indicates some place. 8 Quánzhōu Chuántǒng Xìqǔ Cóngshū (1999, Vol. 5:178). 9 Quánzhōu Chuántǒng Xìqǔ Cóngshū (1999, Vol. 7:287).

REFERENCES Cao, Guang-shun 曹廣順. 1995. Jìndài Hànyǔ yǔzhùcí 近代漢語語助詞 [The particles of modern Chinese]. Beijing: Language Press. Li, Ru-rong 李如龍. 1996. Quánzhōu fāngyán de tǐ 泉州方言的體 [The aspects of Quanzhou dialect]. In The series on comparative research of China’s southeastern dialects. Vol. 2, The aspect of verbs, ed. by Shuang-Qing Zhang 張雙慶, 195–224. Hong Kong: T.T. Ng Chinese Language Research Centre. Lin, Lunlun. 林倫倫. 2001. Lùn Yùedōng Mǐnnán fāngyán yǔ Fújiàn Mǐnnán fāngyán de guānxī 論粵東閩南方言與福建閩南方言的關係 [On the relation between Southern Min dialects in East of Guangdong and Southern Min dialects in Fujian]. In Mǐnnán fāngyán Zhāngzhōuhuà yánjiū 閩南方言漳州話研究 [The studies on Zhangzhou

Continuative aspect marker le(h)4 195 dialect: A Southern Min dialect], ed. by Bi-jia Chen 陳碧加, 303–315. Beijing: China Federation of Press. Lin, Songyu 林頌育. 2010. Shì lùn Mǐnnánhuà chíxùtǐ biāojì de láiyuán 試論閩南話持續 體標記的來源 [Discussion on the original word of the durative aspect marker]. Yǔyán Kēxué 語言科學 2010:386–393. Lin, Tiansong. 林天送. 2006. Quánzhōu fāngyán yǔfǎ sìbǎinián de yǎnbiàn 泉州方言語法 四百年的演變 [Grammatical changes of Quanzhou dialect in 400 years]. Master thesis, Xiamen University, Xiamen. Liu, Dangqing 劉丹青. 1996. Dōngnán fāngyán de tǐmào biāojì 東南方言的體貌標記 [The aspect markers of Southeast dialects]. In The series on comparative research of China’s southeastern dialects. Vol. 2, The aspect of verbs, ed. by Shuang-Qing Zhang 張 雙慶, 9–33. Hong Kong: T.T. Ng Chinese Language Research Center. Lü, Shu-xiang 呂叔湘. 1984. Shì Jǐngdéchuán Dēnglù zhōng zài zhe èrzhùcí 釋景德傳燈 錄中在、著二助詞 [The explanation of two particles ‘zai’ and ‘zhe’ in Jingde Chuandeng recorded]. In Hànyǔ yǔfǎ lùnwénjí 漢語語法論文集 [Papers on Chinese grammar], ed. by Shu-xiang Lü 呂叔湘, 58–72. Beijing: The Commercial Press. Mei, Tsu-lin 梅祖麟. 2000. Hànyǔ fāngyán lǐ xūcí zhe zì sānzhǒng yòngfǎ de láiyuán 漢 語方言裡虛詞「著」字三種用法的來源 [The source of three usages of the function word ‘著’ in Chinese dialects]. In Méizǔlín yǔyánxué lùnwénjí 梅祖麟語言學論文集 [Mei Tsu-Lin’s papers on linguistics], 155–187. Beijing: The Commercial Press. van der Loon, Piet. 1995. Míngkān Mǐnnán xìqǔ xuánguǎn xuǎnběn sānzhǒng 明刊閩南戲 曲弦管選本三種 [Three Ming anthologies of the theatre and art song of South Fukien]. Beijing: Chinese Theater Press. Wang, Jianshe 王建設. 2004. Cóng Míngqīng fāngyán xìwén kàn zhe de yǔfǎhuà guòchéng 從明清閩南方言戲文看「著」的語法化過程 [On grammaticalized process of ‘著’ from the dramatic script of Southern Min dialect during Ming and Qing dynasties]. Huáqiáo Dàxué Xuébào 華僑大學學報 3:128–132. Wang, Jianshe 王建設. 2010. Zàilùn Quánzhōuhuà wánchéngtǐ he chíxùtǐ zhùcí l↔0 de ­láiyuán 再論泉州話完成體和持續體助詞[lə0]的來源 [A further discussion on the source of durative particle ‘lə0’ in Quanzhou dialect]. In Hànyǔ Fāngyán Xīntànsuǒ: Dìsìjiè Hànyǔ Fāngyán Yǔfǎ Guójì Yántǎohuì Lùnwénjí 漢語方言新探索: 第四屆漢 語方言語法國際研討會論文集 [New studies on the grammar of Chinese dialects: The papers of the 4th international conference on the grammar of Chinese dialects], ed. by Hua-Dong Lin 林華東, 87–92. Xiamen: Xiamen University Press. Wang, Li 王力. 1980. Hànyǔ shǐ gǎo 漢語史稿 [Chinese historical manuscript]. Beijing: Zhonghua Book Company. Wu, Fu-xiang 吳福祥. 2004. Yě tán chíxùtǐ biāojì zhe de láiyuán 也談持續體標記「著」 的來源 [On the source of the durative aspect marker ‘zhe’]. Hànyǔshǐ Xuébào 漢語史 學報 4:17–26. Wu, Qiuli 吳秋麗. 2008. Quánzhōu fāngyán jiècí yánjiū 泉州方言介詞研究 [The study on the prepositions in Quanzhou dialect]. MA thesis, Fujian Normal University, Fujian. Wu, Shou-Li 吳守禮. 2001a. Míng Jiājìng kān Lì Jìng Jì xìwén jiàolǐ 明嘉靖刊荔記戲文 校理 [Annotated texts of the Romance of Li Jing Ji of Ming Jiajing edition (JJ)]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2001b. Míng Wànlì kān Lì Zhī Jì xìwén jiàolǐ 明萬曆刊荔枝記戲文 校理 [Annotated texts of the Romance of Li Zhi Ji of Ming Wanli edition (WL)]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2001c. Qīng Shùnzhì kān Lì Zhī Jì xìwén jiàolǐ 清順治刊荔枝記 戲文校理 [Annotated texts of the Romance of Li Zhi Ji of Qing Shunzhi edition (SZ)]. Taipei: Congyi Workshop.

196  Manjun Chen Wu, Shou-Li 吳守禮. 2001d. Qīng Guāngxù kān Lì Zhī Jì xìwén jiàolǐ 清光緒刊荔枝記戲 文校理 [Annotated texts of the Romance of Li Zhi Ji of Ming Guangxu edition]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2002a. Míng Wànlì kān Jīn Huā Nǚ xìwén jiàolǐ 明萬曆刊金花 女戲文校理 [Annotated texts of Jin Hua Nü of Ming Wanli edition]. Taipei: Congyi Workshop. Wu, Shou-Li 吳守禮. 2002b. Míng Wànlì kān Sū Liù Niáng xìwén jiàolǐ 明萬曆刊蘇六 娘戲文校理 [Annotated texts of Su Liu Niang of Ming Wanli edition]. Taipei: Congyi Workshop. Yang, Xiufang 楊秀芳. 1992. Cóng lìshǐ yǔfǎ de guāndiǎn lùn Mǐnnánhuà zhe jí chíxùmào 從歷史語法的觀點論閩南話「著」及持續貌 [On ‘著’ and durative aspect of Southern Min dialect from the point of view of the historical grammar]. Hànxué Yánjiū 漢學 研究 10:349–394. Zhang, Bojiang 張伯江, and Mei Fang 方梅. 1996. Hànyǔ gōngnéng yǔfǎ yánjiū 漢語功能 語法研究 [Function studies of Chinese grammar]. Nanchang: Jiangxi Education Press. Zheng, Guo-quan 鄭國權. 1999–2000. Quánzhōu chuántǒng xìqǔ cóngshū 泉州傳統戲 曲叢書（第1–15卷) [Quanzhou traditional drama series (1–15)]. Beijing: Chinese Theater Press.

9  Kong2 as a verb for saying ‘on the move’ in Taiwanese Southern Min* Chinfa Lien

1 INTRODUCTION The fascinating grammatical variability of kong2 講 as a verb of saying in Taiwanese Southern Min (henceforth TSM) has been noted in the literature for some time. Lien (1988) touches on the pragmatics of kong2 講 classifying it into two types in terms of mood: 1) the assertive kong2 講 and 2) the directive kong2 講. In particular, the first type is used by the speaker to contradict his own previous thought. The second type as a directive is conducive to the reading of a mild injunction. Based on spoken data Chang (1995) tries to motivate and map out grammaticalization paths of kong2 講 in terms of syntactic analysis, pragmatic inference, and metaphorical transfer. Its basic function as she claims is the expression of counter-belief. Cheng (1997) identifies four grammaticalized functions of kong2 講: adverb, sentence-final particle, clause-taking complementizer, and quotative marker beyond its status as a lexical verb. Cheng claims that the adverbial kong2 講 may be used to express hearsay and condition. Chappell (2008) (Cf. Tseng 2008) distinguishes lexical verbs uses and grammatical uses of kong2 講. She considers its quotative use to be lexical, and uncovers five grammatical functions: 1) complementizer, 2) topic, 3) conditional, 4) purposive, and 5) discourse marker. Two important observations emerge in this paper. First, speech act verbs and kong2 講 are more loosely bound than cognition verbs and kong2 講 in the light of frequency of token in the spoken corpus. Second, the complementizer kong2 講 will shed all of its syntactic trappings. The paper mainly deals with grammaticalization of kong2 講 as a verb of saying in TSM texts.1 The change of kong2 講’s grammatical function has a close relationship with its syntactic structure. As a result of grammaticalization, kong2 講 develops from a three-place lexeme to markers taking on a range of grammatical functions. The present study heavily relies on the spontaneous and colloquial data for extracting relevant examples. The theoretical proposals are based on the evidence adduced from the data. In terms of conceptual structure, kong2 講 involves three semantic roles: speaker, addressee, and the content of speech. When functioning as a main verb of saying, kong2 講 as a three-place predicate involves three thematic roles: speaker, addressee, and theme denoting the content of speech. Kong2 講 as a complementizer is preceded by main verbs such as

198  Chinfa Lien tsai1iann2 知影 ‘know’, siunn7 想 ‘think’, thiann1 聽 ‘listen’, mng7 問 ‘ask’, in3 應 ‘answer’, ma7 罵 ‘scold, rail’, and uan3than3 怨嘆 ‘complain’. Kong2 講 shows multiple functions: other than verb, kong2 講 ‘say’ can function as quotative marker, complementizer, topic marker, parenthetical, and sentence-final particle. It can also form lexicalized collocates. In terms of argument structure kong2 講 is a three-place predicate taking three arguments. The three arguments carry the semantic role of speaker, addressee, and the content of speech. Let X, Y, and Z stand for speaker, addressee, and content of speech respectively, the construction featuring kong2 講 can be schematized as follows: X + Preposition + Y + kong2 講 + Z In terms of grammatical function, as shown in examples (1) to (6), X can be realized as subject, whereas Y can be realized as the object of a preposition and Z, the object of the verb of saying. The preposition can be realized as hiong3 向 ‘to’, tui3 對 ‘to’, or ka7 共 ‘with’. (1) 就 tsiu7 then



dm

二个 nng7e5 two-clf

人 lang5 person

啦：「 lah0

講 kong2 say

prt

您 lin2 2pl

師傅 sai1hu7 mentor

就 tsiu7 then

向 hiong3 to

徒弟 too5te7 apprentice

講：「 kong2 say

你 li2 2sg

佇 ti7 at

遮 tsia1 here

擱 做 二個 月， (210.07 新社一) koh4 tso3 nng5koo3 gueh82 again do two-clf month ‘The mentor talked to his apprentice, ‘You work here for two more months’.

(3) 伊 i1 3sg



in1 in1 3pl

二个 人 一个 人 一 碗， (144.11 新社一) nng7e5 lang5 tsit8e5 lang5 tsit8 uan2 two-clf person one-clf person one bowl ‘(The boss) said to the two persons, ‘Only one bowl for one person’.

(2) 啊 ah0



向 hiong3 to

對 tui3 to

母親 bo2tshin1 mother

講：「 kong2 say

母親 bo2tshin1 mother

妳 li2 2sg

有 u5 have

啥 物貨 無？」 (32.02–03 新社一) siann2 mih8hue3 bo5 what thing neg ‘He talked to his mother, “Mom, what would you want to eat” ’?

愛 ai3 love

食 tsiah8 eat

Kong2 as a verb for saying ‘on the move’ 199 (4) 我 對 你 講 恭禧。 gua2 tui3 li2 kong2 kiong1hi2 1sg to 2sg say congratulation ‘I said ‘congratulations’ to you’. (5) 這

tse1 dem



菜 tshai3 vegetable

的 e5 nmlz

就 tsiu3 then

共 ka7 ka

先生娘 sian1sinn1niu5 scholar-wife

講： (90.05 沙鹿) kong2 say

‘The vegetable vendor talked to the scholar’ wife’.

(6) 我 gua2 1sg



賣 bue7 sell

(160.16 清水一)

就 tsiu3 then

共 ka7 ka

閻羅王 giam5lo5ong5 Lord.of.Hades

講：『 kong2 say

我 gua2 1sg

呼， honn0 top

若 na7 if

有 u7 have

就 好 啦， (142.2 沙鹿二) 錢 tsinn5 tsiu7 ho2 lah0 money then good SFP ‘I talked to the King of Hell, ‘It would be much better that if I were rich’.

In the remaining parts of the paper I will first deal with kong2 講 as a quotative marker in Section 2. It remains as a lexical verb. Section 3 tackles kong2 講 as a complementizer. Section 4 looks at it as a topic marker. Section 5 examines it as a parenthetical. Section 6 is concerned with it as a sentence-final particle. Section 7 discusses it as a grammaticalized elements in lexicalized collocates and 8 is devoted to the treatment of the bridging context before concluding the paper in Section 9. 2 KONG2 講 AS A QUOTATIVE MARKER Kong2 講 can occur after verbs of saying such as mng7 問 ‘ask’, in3 應 ‘answer, reply’, huan1hu3 吩咐 ‘give instruction to sb for him to do sth’, ma7 罵 ‘scold’, uan3than3 怨嘆 ‘complain’. In these examples, kong2 講 can function as a quotative marker of a quoted speech.3

2.1 Mng7kong2 問講 What is quite intriguing of this construction in TSM is that the quotative marker kong2 講 that follows the verb mng7 問 ‘ask’ has to be followed by an interrogative clause featuring various wh-words like siann2mih8 啥物 ‘what’ in (7), an2tsuann2 安怎 ‘how’ in (8), na2 那 ‘why’ in (9), given that any expression can appear in quotation. Kong2 講 in this construction seems to exhibit the burgeoning state of becoming a bona fide complementizer.

200  Chinfa Lien (7)

(8) (9)



來 問 講 是 啥 物 lai5 mng7 kong2 si7 siann2 mih8 come ask say cop what thing ‘(Someone) came to ask what the factor was’.

原因 guan5in1 factor

啦。 lah0

(72.18 新社一)

prt

表兄 問 講 卜 安怎 害？ (68.09 沙鹿) piau2hiann1 mng7 kong2 beh4 an2tsuann2 hai7 dem cousin ask say want how frame ‘The cousin asked how to set a trap’. 這 tse1

啊 ah0

問 講 你 這个 所在， 那 mng7 kong2 li1 tsit4e5 soo2tsai7 na2 dm ask say 2sg dem-clf place why 「打狗」？  (108.19 大甲) 號做 hoo7tso2 ta2kau2 be-called Ta-kau ‘(Someone) asked you why this place was called “Ta-kau” ’?

會 e7 can

One can drive home the obligatory clause-taking nature of the complementizerlike element kong2 講 in construction with mng7 問 by contrasting the mng7 kong2 問講 construction with kong2 講-less mng7 問 construction. Apart from the cases where the verb mng7 問 can be followed by an interrogative clause, as in (10), mng7 問 can be followed by a noun phrase denoting reason, as in (11) and (12). (10) 請 問 嘿 到底 啥麼 關係，(158.03 嘉義五) tshiann2 mng7 he1 tau3te2 siann2mih8 kuan1he7 please ask dem what.on.earth what relation ‘(The magistrate) asked what on earth was the relation with it’. (11) 我 來 問 gua2 lai5 mng7 1sg come ask ‘I came to ask you’.

你 li2 2sg

啦 lah0

嘿， he04

prt

prt

(12) 來 問 這个 原因 啦。 lai5 mng7 tsit4e5 guan5in1 lah0 come ask dem-clf reason prt ‘He came to ask for the reason (why)’.

(168.12 新社鄉二)

(8.02 新社鄉一)

2.2 In3kong2 應講 Kong2 講 immediately following the verb in3 應 functions as a quotative marker, as illustrated in (13) where ho2lah0 好啦 ‘okay’ is a direct quote.

Kong2 as a verb for saying ‘on the move’ 201 (13) 共 ka7

伊 應 講 「好 i1 in3 kong2 ho2 ka 3sg answer say okay ‘(He) answered ‘Okay’ to her’.

啦」。 lah0

(12.09 嘉義五)

prt

Absent kong2 講, in2 應 ‘answer’ can take an object noun phrase realized either as an addressee, as in (14), or a theme denoting a general speech content, as in (15), or it can simply go alone, as in (16). (14) 安呢 應 因 老父仔 呼 an2ne1 in3 in1 lau7pe7a2 honn05 this.way answer 3sg.gen old.father-dim prt ‘(The oldest daughter) answered her father this way’. (15) 這个 tsit4e5 dem-clf



飼 tshi7 keep

羊 iunn5 sheep

袂 be7 cannot

顧 koo3 take.into.account

咧 leh0 prt

合 伊 應 話 (114.06 沙鹿二) kah4 i1 in3 ue76 with 3sg answer word ‘The shepherd can’t spare time for answering him’.

(16) 啊 ah0

囝仔 gin2a2 child

(164.130 南投一)

阮 毋 就 gun2 m7 tioh8 dm 1sg.gen neg then ‘Wouldn’t I answer you’?

共 ka7 ka

你 li2 2sg

應， in3 answer

(94.21 苗栗)

2.3 Huan1hu3kong2 吩咐講 The quotative marker kong2 講 immediately after the verb huan1hu3kong2 吩咐 takes a quote, as in (17) (17) (in1 翁) 就 吩咐 講 呼: 「你 就 共 in1 ang1 tsiu7 huan1hu3 kong2 honn0 li2 tsiu7 ka7 3sg.gen husband then instruct say prt 2sg then ka 拭 啦， (90.03 臺南二) tshit4 lah0 wipe prt ‘His husband gave the instructions, “you just wipe (the stature) for Mother” ’. If huan1hu3 吩咐 ‘give instruction’ does not combine with the quotative marker kong2 講, then what is involved is a pivotal construction, as in (18).

202  Chinfa Lien (18) 你

去 共 吩咐 毋通 賣 別 人。 (10138.11 台南) li2 khi3 ka7 huan1hu3 m7thang1 bue7 pat8 lang57 2sg go ka inform cannot sell other people ‘You go and tell (him) not to sell (the pork) to others’.

2.4 Ma7kong2 罵講 Ma7 罵 when followed by the quotative marker kong2 講 always take a quoted clause, as illustrated in (19) and (20). (19) 老師

才 共 罵 講： 「你 赫爾 聖， (46.06 宜蘭二) lau7su1 tsiah8 ka7 ma7 kong2 li2 hiah4ni7 siann3 teacher adv ka scold say 2sg so efficacious ‘The teacher just scolded (the student) that your prediction was so accurate’.

(20) 共

罵 講： 「是 安怎 才 會無當 尋？ (48.11 羅陳) ka7 ma7 kong2 si7 an2tsuann2 tsiah8 e7bo5tang3 tshue7 ka scold say cop why just cannot find



‘ “Why couldn’t I find (my husband)”? railed she at him’.

By contrast, ma7 罵 rid of the quotative marker occurs an object noun phrase, as in (21). (21) 啊 ah0

我 得 罵 彼的 gua2 leh4 ma7 hit4e5 dem 1sg asp scold dem-clf ‘I was scolding that child’.

囝仔， gin2a2 child-dim

(128.23 沙鹿)

2.5 Uan3than3kong2 怨嘆講 Uan3than3 怨嘆 combined with the quotative marker kong2 講 take a quoted clause, as in (22). (22) 啊 ah0

袂使 怨嘆 講 日頭 赫 bue7sai2 uan3than3 kong2 jit8thau5 hiah4 dm neg-can complain say sun that ‘We cannot complain that the sunlight is so strong’.

猛，(76.12 雲林四) bing2 hot

Uan3than3 怨嘆 without being adorned by a quotative marker can be used intransitively, as in (22), or transitively, as in (23). (23) 飼

著 好 囝 痞 囝 免 怨嘆。(60.03 嘉義五) tshi7 tioh8 ho2 kiann2 phai2 kiann2 bian2 uan3than3 raise asp good child bad child need.not complain ‘You need not complain whether you have a good child or a bad child’.

Kong2 as a verb for saying ‘on the move’ 203 (24) 啊 ah0

袂使 怨嘆 別 人。 (76.05 雲林四) bue7sai2 uan3than3 pat8 lang5 dm should.not complain other people ‘He couldn’t blame other people for his fate’.

Besides, the bare verb uan3than3 怨嘆 can also be followed by a clause, as in (25), but it is not a quoted speech. (25) 袂使

怨嘆 人 無 共 伊 照顧。 bue7sai2 uan3than3 lang5 bo5 ka7 i1 tsiau3koo3 should.not complain people neg ka 3sg take.care ‘He shouldn’t complain that no one would take care of him’.

(76.09 雲林四)

3  KONG 2 講 AS A COMPLEMENTIZER When kong2 講 occurs after perceptual verbs such as khuann3 看 ‘look’, thiann1 聽 ‘listen’ or cognitive verbs such as siunn7 想 ‘think’ and tsai1iann2 知 影 ‘know’, it has lost its status of three-place predicate and become a complementizer which takes a factive or interrogative clause as its complement.8 It is the perceptual and cognitive nature of the verbs in this category that spurs the change of kong2 講 from a quotative marker to a complementizer. It is similar to but behaves not exactly like the complementizer ‘that’ in English. An important difference is that the complementizer kong2 講 can be further followed by an interrogative word in TSM. As shown in the following examples, the complementizer kong2 講 that follows perceptual or cognitive verbs is always followed by a finite clause. But cognitive verbs without the complementizer may or may not take a clausal complement. The presence of the complementizer triggers a change of perceptual verbs to pure cognitive verbs. The complementizer kong2 講 in combination with cognitive verbs can be taken as a further development of kong2 講 as a quotative marker, since the complementizer kong2 has shed its sense of saying. In what follows I will provide examples showing the contrast between the V + kong2 construction and the V + kong2-less construction.

3.1 Khuann3kong2 看講 There are two kinds of clauses that can follow khuann3kong2 看講: (1) factive clauses and (2) interrogative clauses. The first kind of clauses that appears after the complementizer kong2 講 denote a real situation visible to eye witnesses, as in (26) and (27). (26) 看

講 有 筍仔 浮 出來， (32.12 新社一) khuann3 kong2 u7 sun2a2 pu5 tshut4lai5 see comp have bamboo.shoot-dim float come.out ‘(Someone) saw that there were bamboo shoots popping up from the soil’.

204  Chinfa Lien (27) 出去 tshut4khi3 out



看 khuann3 go.see

講 kong2

啊 ah0

comp

dm

有 u7 have

一个 tsit8e5 one-clf

所在 soo2tsai7 place

闊闊 啦， (132.1 沙鹿) khuah4khuah4 lah0 spacious prt ‘(Someone) went out and saw that there was a spacious place’.

The second kind of clause that follows the complementizer features an A-not-A question, as in (28) and (29), or a question-word question, as in (30) and (31). (28) 看 講 會 漏 袂？ (98.12 新社二) khuann3 kong2 e7 lau7 bue2 see comp can leak neg ‘(A foreman) tested the pot and saw if water would leak from it’. (29) 看

講 亻因 厝 毋 知 有 要緊 無？(161.04 清水一) khuann3 kong2 in1 tshu3 m7 tsai1 u7 iau3kin2 bo5 see comp 3pl house neg know have important not.have ‘(They) wanted to know whether their house was safe from the disaster or not’.

(30) 啊 ah0



得

看

講

佗

一个

查甫

leh4

khuann3 kong2 to2 tsit8e5 tsa1poo1 dm asp see comp which one-clf woman ‘(Someone) was looking which one was the right woman’.

(31) 來

看 講 lai5 khuann3 kong2 come see comp

這的 tsia1e5 dem-nmlz

草 是 tshau2 si7 grass cop

安呢 呼。 (236.06 大甲) an2ne1 honn0 this.way prt

卜 怎樣 beh4 tsuann2iunn7 want why

會 無去。 (112.71 外埔) e5 bo5khi0 dim can not.have-inch ‘(The official) came to see why the grass would disappear’. 仔 a2



This construction involves the decision-making following an eye-witness’s observation. Thus, it takes on an additional sense of decision-making, namely ‘determine whether or not something is the case’, or ‘find out the identity of an unknown entity or why something is the way it is’ through observation.

3.2 Thiann1kong2 聽講 Hearsay is the source of the information conveyed by the clause in this construction, as illustrated in (32), (33), and (34).

Kong2 as a verb for saying ‘on the move’ 205 (32) 主

考官 聽 講 是 福建 人 (178.05 嘉義五) tsu2 kho2kuann1 thiann1 kong2 si7 hok4kian3 lang5 chief examiner hear comp cop Fujian person ‘It was said that the chief examiner was from Fujian’.

(33) 我 聽 講 你 去 關。 gua2 thiann1 kong2 li2 khi3 kuinn1 1sg hear comp 2sg go imprison ‘I heard that you were imprisoned’. (34) 啊



彼

蛇

聽

講

足

大

(196.19 彰化七)

尾

就

著 啦！ (186.070 外埔) ah0 he1 tsua5 thiann1 kong2 tsiok4 tua7 bue2 to7 tioh8 lah0 dm dem snake hear comp enough big tail adv right prt ‘Rumor had it that the snake was so huge. Yes, that is what I heard’.

3.3 Siunn7kong2 想講 The clause in this construction expresses the content of thought, as in (36), (37), and (40). Though a complementizer kong2 講 still retains a trace of the sense of saying in this connection. That the cognitive verb ‘think’ can take a clausal complement expressing the content of thought is not surprising since thought has to be encoded via language.9 (35) 這个 tsit4e5 dem-clf

福仔 hok4a2 Hok-dim

伊 i1 3sg

心肝 sim1kuann1 heart

想 siunn7 think

講， kong2 comp

好 ho2 okay

一个 紅包， (140.09 新社鄉一) 加 ke1 tsit8e5 ang5pau1 prt add one-clf red.envelope ‘The guy named Hok-a thought that it was okay to add one more red packet’. 啊， ah0



(36) 想

(37)

講

將

這

鴨公

放

siunn7 kong2 tsiong1 tse1 ah4kang1 pang3 think comp pm this drake release ‘(The father) thought that he might let the drake go’. 我 想 講 那 會 gua2 siunn7 kong2 na2 e7 1sg think comp why can ‘I wonder why it would be like this’?

安呢？ an2ne1 this.way

予 伊 去， (134.08 新社二) hoo7 i1 khi3 hoo 3sg go

(74.25 沙鹿)

206  Chinfa Lien

3.4 Tsai1iann2kong2 知影講 The cognitive verb tsai1iann2 知影 ‘know’, when followed by kong2 講, takes a causal complement denoting a fact, as in (38). Even when it is preceded by a question-word, it will not affect the factive nature of the content of proposition denoted by the clause, as in (39). However, if the clause falls under the scope of an epistemic modal denoting necessity such as it4ting7 一定 ‘must’, then the effect of factivity will be lessened, as in (40). (38) 馬



哪會

亻因

主人

去予

知影

講

我

毋

是

lin2 na2e7 tsai1iann2 kong2 gua2 m7 si7 2pl how.can know comp 1sg neg cop ‘How could you know that I was not your aunt’?

(40) 伊 i1 3sg



講

人

be2 tsai1iann2 kong2 in1 tsu2lang5 khi3hoo7 horse know comp 3sg.gen master am ‘The horse knew that its master was killed by someone’.

(39) 您



知影

知影 tsai1iann2 know

講 kong2 comp

一定 it4ting7 definitely

害 啊。 (84.2 雲林三) lang5 hai7 ah010 person hurt prt

您 lin2 2sg.gen

姑婆 啦？ (120.08 新社一) koo1po5 lah0 aunt prt

是 si7

啊 ah0

去予 khi3hoo7

cop

dm

am

鬼 kui2 ghost

安呢 啦。 (48.01 新社一) 嚇 著 1 8 2 1 0 kiann tioh an ne lah frighten ach this.way prt ‘She knew that her husband must’ve been frightened by the ghost’.

4  KONG2 講AS A TOPIC MARKER Kong2 講 in the sentence-initial position can be used to usher in a new topic comprising a determiner phrase featuring number one + classifier often preceded by an existential verb u7 有. What is first brought up for further elaboration is an indefinite noun phrase, as in (41). The numeral it4 ‘one’ can be replaced by a deictic element like tsit4 這, as in (42), contributing an effect of familiarity. Kong2 講 can also be used to bring up a new situation for further discussion, as in (43). (41) 講 kong2 top



有 u7 there 啦， lah0

一位 啊 伊 tsit8ui7 ah0 i1 one-clf dm 3sg (46.02 新社一)

工人 kang1lang5 worker prt ‘There was a worker named Li A-Kui’.

名 mia5 name

叫 kio3 call

李阿桂 li2a1kui3 Li.A.kui

的 e5 nmlz

Kong2 as a verb for saying ‘on the move’  207 (42) 講 這 kong2 tsit4 top



dem

一 tsit8 one

工 啊， kang1 ah0 day prt

阿城仔 a1siann5a2 A.Siann-dim

抵仔好 tu2a2ho2 just.now

對 外面 轉 來， (14.08 新社一) tui7 gua7bin7 tng2 lai5 from outside turn come ‘On that day A-Siann happened to come back from the outside’.

(43) 講 自從 離開 kong2 tsu7tsiong5 li7khui1 top since leave



這个 tsit4e5 dem-clf

了後， 我 在 liau2au7 gua2 ti7 after 1sg at

彼个 hit4e5 dem-nmlz

所在， soo2tsai7 place

過 了 麼真 好 呢 啦 (54.06 新社一) ke3 liau2 ma7tsin1 ho2 ni7 lah0 pass pfv also.very good prt prt ‘Ever since I left, I’ve had a wonderful time there’.

5  KONG2 講 AS A PARENTHETICAL When kong2 講 loses its capacity of taking complements and furthermore sheds its status as a head, then it has become an optional element. That is, its deletion does not affect the grammaticality of the sentence involved, as shown in (44), (45), and (46) (Fraser 1980). But when it is removed, the overall interpretation of the sentence will be affected. In a sense, it will miss some important elements of subjective meaning (Nuyts 2001). The addition of the parenthetical kong2 講 has a mitigating effect of downtoning the force of the speech. As a parenthetical, it has no connection at all with and is disjoined from a subject. Such a mitigating effect is achieved by the fact that the speaker is less committed to the truth of the propositional content of the host clause. The speaker can attribute the utterance of the host clause to an unknown third person, and therefore hedge the duty of the validity of what he says, so to speak. Each of the parentheticals illustrated also serves the bracketing function of marking the boundary of a constituent in a clause. It occurs between the modal and the VP, in (44), and between the object noun phrase and the secondary predicate in (45) and (46).11 (44) 逐个 亦 tak8e5 ia7 every-clf also

neg

敢 講 kann2 kong2 dare say

提起 這个 the5khi2 tsit4e5 mention dem-clf

代誌 啦 呼。 (186.14 新社一) tai7tsi3 lah0 honn0 nmlz thing prt prt ‘Everyone dared not mention the baby’. 的 e5



毋 m7

紅嬰仔 ang5inn1a2 baby-dim

208  Chinfa Lien (45) 叫



講

到

五間

尾

去

睏

kio3 i1 kong2 kau3 goo7king1 bue2 khi3 khun3 call 3sg say to five-clf tail go sleep ‘Asked him to sleep in the minor room in a house complex’.

(46) 得

送 sang3 asp ka send ‘(As if Senior and Heaven’. leh4



伊

共 ka7

伊 i1 3sg junior

講 kong2 say fellow

安呢 啦， (216.05 新社一) an2ne1 lah0 this.way prt

上 天庭 安呢， (176.14 清水二) tsiunn7 thian1ting5 an2ne1 ascend heaven this.way apprentices) were seeing him off going up to the

6  KONG2 講AS A SENTENCE-FINAL PARTICLE When kong2 講 functions as a sentence-final particle, it loses its capacity of taking arguments. It takes on the tone value of high-level combination tone, rather than prepausal isolation tone. It expresses the subjectivity on the part of the speaker. It serves to underscore a statement and call the addressee’s attention to a specific situation, as in examples (47) and (48). It can also be used in an imperative with a tone-down sense, as in examples (49) and (50). It carries an additional function of joint activity. (47) 「阿叔、 阿叔， 你

是

怎樣

啦？」 伊

講： 「呼 (130.08 沙鹿) kong2 honn0 say dm



a1tsik4 a1tsik4 li2 si7 tsainn2iunn7 lah0 uncle uncle 2sg cop how prt ‘Uncle, Uncle. What happened to you’?



一下 講。」(130.09 沙鹿) 我 予 阮 阿兄 驚 gua2 hoo7 gun2 a1hiann1 kiann1 tsit8e7 kong212 1sg am 1pl.excl older-brother scare sem prt ‘I was scared by my older brother’.

(48) 啊 a1

dm

車底 所 載 tshia1tue2 soo2 tsai3 car.bottom Object.pronoun carry

空空空 啊。 khang1khang1 khang1 a0 empty prt



被 hoo7 am

i1 3sg

的 e5 nmlz

物件 mngh8kiann7 thing

亦 ia7 also

押 入 山區 的 ah4 jip8 suann1khu1 e5 escort enter mountain.area nmlz

百姓 無 看 著 (46.25 新社鄉二) peh4sinn3 bo5 khuann3 tioh0 peasant neg look ach ‘The car is completely empty. No folks escorted into the mountain can be seen’.

Kong2 as a verb for saying ‘on the move’ 209



無 kah4 一个 5 bo kah4 tsit8e5 not.have to.the.extent.of one-clf ‘No one appeared’.

出 來 講。 (48.01 新社鄉二) tshut4 lai0 kong2 exit come prt

講：「 唔 (49) 伊 啊 醉醉仔， i1 kong2 m7 a1 tsui3tsui3a2 3sg say neg dm drunk 睏 講。」。 (96.11 清水一) 去 khi3 khun3 kong2 go sleep prt ‘He muttered, ‘I am drunk. Let’s go to bed’. (50) 我



gua2 1sg

昨昏 來 tsa1hng1 lai4 yesterday come

提 theh8 carry

一个 tsit8e5 one-clf

查某囝仔

偌

睏 khun3 sleep

您 lin2 2sg.gen

就 tsiu7 then

來 lai5 let’s

兜 tau1 place

呼， honn7

紅龜粿， 啊 ang5ku1ke2 a1 red.turtle.glutinous.cake dm

變 pian3 change

一个 tsit8e5 one-clf

水

(141.100 雲林一)

咧。」

prt

tsa1boo2kiann2 jua7 sui2 leh0 woman-dim very beautiful prt ‘I came to your home carrying a red turtle cake last night. It changed into a beautiful lady’.

伊 i1 3sg

卜 beh4 want

講：「 我 來 您 kong2 gua2 lai5 lin2 say 1sg come 2sg.gen

兜 看 tau1 khuann3 home look

講。」(141.101–2 雲林一) kong2 prt

‘He said, “Let me take a look at her in your home” ’.

7  KONG2 講 IN LEXICALIZED COLLOCATES Kong2 講 can combine with other words to form lexicalized collocates like soo2i2kong2 所以講 ‘therefore’, as in (51), na7kong2 若講 ‘if’, as in (52), ­na2tshin1tshiunn7kong2 若親像講 ‘appear, look’, tshiunn7kong2 像講‘for example’,13 as in (53), kan2na2kong2 敢若講 ‘look as if, appear, like’, as in (54) and (55), and (khia7) sng3kong2 (企)算講 ‘count as’, (56) and (57).14 All the lexicalized collocates serve to denote the speaker’s subjective sense and fulfill a metalinguistic function.

210  Chinfa Lien (51) 因為

彼時 in1ui7 hit4si5 because dem-time 啦

講

啊 a7 also

是 si7

得 tit4

cop

prog

修理 水圳 siu1li2 tsui2tsun3 repair irrigation-ditch

這个

水圳仔

無

水

lah0



啦， (28.08 新社二) lah0

kong2 tsit4e5 tsui2tsun3a2 bo5 tsui2 prt prt say dem-clf ditch-dim not.have wate ‘Since they were fixing the ditch, there was no water in it (the ditch)’.

(52) 阿嫂

你 若 講 安呢， 我 就 安呢。 (140.15 沙鹿) a1so2 li2 na7 kong2 an2ne1 gua2 tsiu7 an2ne1 sister.in.law 2sg if say this.way 1sg then this.way ‘Sister-in-law, if you prefer this way, I will follow suit’.

(53) 像



講

姊妹仔

翁 啊， (178.01 雲林二) tshiunn7 kong2 tsi2muai7a2 sann1e5 ke3 sann1e5 ang1 ah0 like say sister-dim three-clf marry three-clf husband prt ‘For example, three sisters married three husbands (respectively)’.

(54) 我

三個

干若 講 都 這段 gua2 kan2na2 kong2 to7 tsit4tuann7 1sg as.if say all dem-clf



嫁

時間 si5kan1 time

無 bo5

佇 ti7

咧 (98.12 清水一) leh0

neg

cop

prt

‘Like, I was not here during this period of time’.

(55) 這 tse1

三個

就 干若 講 鎮家之寶， tsiu7 kan2na2 kong2 tin3ke1tsi1po2 dem just as.if say heirloom ‘It was like our family heirloom’.

(106.08 清水一)

(56) 我

這个 翁 算 講 是 破病 死 的 (72.19 沙鹿) gua2 tsit4e5 ang1 sng3 kong2 si7 phua3penn1 si2 e5 1sg dem-cfl husband count say cop sick die nmlz ‘My husband died of illness if I may put it this way’.

(57) 咱 lan2 1pl.excl



古早 koo2tsa2 old

人 lang5 people

企算 講 khia7sng3 kong2 be.considered say

勤儉 (10.12 大安二) 較 khah4 khin5khiam7 more industrious.thrifty ‘We, old people, were considered more industrious and thrifty’.

Kong2 as a verb for saying ‘on the move’ 211 8  THE BRIDGING CONTEXT Kong2kio3 講叫out of context may have two interpretations: (1) It may be a phrase consisting of a sequence of kong2 講 as a verb of saying and kio3 叫as a directivecausative verb and (2) it may be a phrase comprising kong2 講 as a parenthetical and kio3 叫 as a directive-causative verb. The first case is illustrated by kong2kio3 講叫 in (59), a syntagmatic unit since the subject of the verb kong2 講 absent in the present sentence can be recovered from a previous context and identified as dad and mom, as shown in (58). (58) 阿嫂！ 阿爸 a1so2 a1pa5 sister.in.law father

阿母 a1bu2 mother

呼， 講 hoonn0 kong2 dm say

你 li2 2sg

有 囝仔 u7 gin2a2 have child-dim

無 耳 鼻 啦 (146.08 彰化五) 呼， hoonn0 bo5 hi7 phinn7 lah0 prt prt not.have ear nose prt ‘Sister-in-law, dad and mom said you have a kid without ears and a nose’. 啊 a1



(59) 講

叫 我 著 共 你 添 耳 鼻 安呢。 (146.09 彰化五) kong2 kio3 gua2 tioh8 ka7 li2 thiam1 hinn7 phinn7 an2ne1 say call 1sg must ka 2sg add ear nose this.way ‘They (viz., dad and mom) asked me to add ears and nose for you’.

The second case is exemplified by (60) where kong2 講 in the two instances of kong2 kio3 講叫 is two-way ambiguous: it can be construed as a verb of saying or a parenthetical. The subject of the verb kong2 講 ‘say’ is the princess in the first interpretation, whereas kong2 講 as a parenthetical is detached from the host clause and is conducive to a mitigating effect of the illocutionary force in the second interpretation. (60)



啊 ah0

伊 講：「啊 就 無 法度， 彼 皇帝 的 i1 kong2 ah0 tsiu7 bo5 huat4too7 he2 hong5te3 e5 dm 3sg say dm then not.have method dem emperor nmlz 查某子 就 講 叫 我 著 予 伊 招親， 啊 tsa1boo2kiann2 tsiu7 kong2 kio3 gua2 tioh8 hoo7 i1 tsio1tshin1 ah0 daughter just say call 1sg must hoo 3sg marry dm 講 叫 我 袂使 講 阮 兜 有 某， 阮 兜 有 kong2 kio3 gua2 be7sai2 kong2 gun2 tau1 u7 boo2 gun2 tau1 u7 say call 1sg cannot say 1sg home have wife 1sg home have 某 愛 共 掠 去 刣。」 (134.37–39 雲林一) boo2 ai3 ka7 liah8 khi3 thai5 wife must ka arrest go kill ‘He said, “I can’t help it. The emperor’s daughter asked me to be her husband. She said that I shouldn’t have a wife at home. If I do, she should be killed” ’.

212  Chinfa Lien Furthermore, kong2kio3 講叫 can become a lexicalized unit taking on the metalinguistic function of ‘be dubbed as’, as illustrated in (61). (61) 一張

共 貼 e1 彼个 講 叫 蚊罩 (114.138 外埔鄉) tsit8tiunn1 ka7 tah4 e1 hit4e5 kong2 kio3 bang2ta3 one-clf ka paste on dem-clf say call mosquito.net ‘One sheet of the charm is pasted on the one called mosquito net’.

In underscoring the important role of context in grammaticalization Heine (2002) proposes a scenario of grammatical change where a bridging context plays a pivotal role in motivating grammaticalization in that a context-induced new function through inference coexists with the old function. When it reaches the stage at which the old function is shed in favor of the new function then change of function has been conventionalized. In the case of kong2kio3 講叫 we can see the rise of the new function of kong2 講 construable as a parenthetical in coexistence with its older function as a verb of saying. But we are still at the intermediate stage as we have not found any context where the older function has been obliterated.

9 CONCLUSION To recapitulate what has been discussed so far, apart from being a three-place predicate of saying kong2 講 has taken on a range of grammatical functions: quotative marker, complementizer, topic marker, parenthetical, and sentence-final particle.15 As a first step of grammaticalization kong2 講 is used as a quotative marker in the construction Vsaying + kong2 講 to mark a quoted speech. As a second step of grammaticalization kong2 takes on the function of being a complementizer of cognitive verbs and what it marks is de dicto (about expressions) sense rather than de re (about things) sense (Frajzyngier 1991). It carries a metalinguistic function. This expression may be made possible by the thinking-is-saying metaphor, viz., the content of thinking is likened to the content of saying. When a perceptual verb such as khuann3 看 ‘look’ or thiann1 聽 ‘listen’ is followed by a sentential complement without being headed by the complementizer kong2, the complement may function as a factive denoting things in the real world, viz., fulfilling the de re function. But when kong2 講 appears as a complementizer taking a sentential complement, the complement may represent a reported situation fulfilling its de dicto function. When the main verb is khuann3 看 the complementizer kong2 講 may be followed by an indirect question or A-not-A question. This may yield a sequence of complementizer and wh-word. Such a sequence is not tolerated in English, as in *‘see that whether or not John will come’. As a step further in its grammaticalization, kong2 講 as a parenthetical has lost its complement-taking capacity and become an optional element. Its removal will affect the grammaticality of the sentence, but its presence will contribute subjective sense to the overall sentential interpretation. In other words, it is a metalinguistic device to express the

Kong2 as a verb for saying ‘on the move’ 213 speaker’s comment on the content of the sentence proper. When occurring in the sentence-final position kong2 講 has shed all its more concrete lexical senses and its function is closely related to types of mood, an enhanced degree of subjectivity and discourse function. It inevitably expresses the speaker’s subjective stand, functions in interactive/interpersonal situation, and special types of sentence construction like hortatives and question and answer pair. When functioning as a parenthetical and a sentence-final particle kong2 講 is also conducive to a mitigating effect of what is expressed in the propositional content (Fraser 1980). If we test various manifestations of grammaticalization of kong2 講 elaborated on in the paper against the five heuristic principles of grammaticalization: layering, divergence, specialization, persistence, and de-categorization (Hopper 1991), we can see that only three principles, viz. divergence, persistence, and de-categorization, are at work. Divergence is involved since kong2 講 as a lexeme coexists with a wide range of grammaticalized features such as complementizer, topic marker, parenthetical, and sentence-final particle that it develops. Despite its multiple functions, kong2 講 can still be used as a verb for statements. This is in keeping with the principles of persistence. Kong2 講 has also undergone de-categorization in that it takes on some minor grammatical categories like complementizer, parenthetical, topic marker, and sentence-final particle. Since there is no earlier form of complementizer in Southern Min, no layering, viz. the coexistence of earlier complementizer and new complementizer is involved; nor does specialization play a role, since no other synonyms join kong2 講 in the paths of grammaticalization. Thus, not all the five principles of grammaticalization are involved in the case of kong2 講 examined in the paper. NOTES * The present work is partially supported by NSC 98–2923-H-007–001-MY. An earlier version of the paper was presented at the 20th International conference on historical linguistics, July 25–30, 2011, National Museum of Ethnology, Osaka. I would like to express my gratitude to the audience for comments and suggestions. Thanks are due to May Wang for helping collect relevant examples and prepare the paper. 1 The TSM examples are extracted from a data base consisting of a series of folk tales. A coded abbreviation in parenthesis for which the detailed information can be found in the appendix is given at the end of each extracted example. 2 Ah0 啊 functions as a discourse marker to initiate a sequence of utterances. 3 If we follow Ransom (1988) in taking many kinds of grammaticalization as a continuum, kong2 講 as a quotative marker is not a full-fledged but a pseudo-grammatical function word. See Chappell 2008, which takes kong2 講 as a lexeme. The usage foreshadows the emergence of kong2 講 as a complementizer. 4 PRTs as a rule express the speaker’s subjective stance beyond the propositional content of the host sentence. Lah0 啦 conveys an assertive force, whereas he0 嘿 is used to engage the attention of addressee. 5 Honn0 呼 as a sentence-final particle is used to inform the addressee of a new message carried by the host sentence is being conveyed to him. 6 Leh0 咧 is a weakened form of tit4 得, a post-verbal dynamic modal denoting capability. Be7tit4 袂得 ‘not be able to, cannot’ is a circummodal construed as a discontinuous construction comprising a preverbal and postverbal modal at the same time.

214  Chinfa Lien 7 Ka stands for ka7 共 derived from kang7 by shedding its velar coda. Prototypically ka7 共 occurs before a pronoun and the pronoun can be dropped if it is fully recoverable in context. The noun phrase following the chameleon-like preposition ka7 共 takes on a range of context-induced semantic roles such as source, goal, and benefactive (see Lien 2002). 8 I take kong2 講 as a complementizer as an emergent case of grammaticalization as it has experienced semantic bleaching, shedding its lexical sense of saying, and takes on a bundle of morphosyntactic properties. For example, in its capacity of complementizer it can never be followed by phrase marker ke3 過 and preceded by the experiential marker bat4/pat4 捌. However, it has not proceeded to the stage of having phonetic reduction, although it is evident that its tone is always invariably high-level without participating in the lexical tone sandhi. For the discussion of the notion of grammaticalization see Hopper 1991 and Traugott and Dasher 2002. 9 This use of thinking-is-talking is not unlike the quotative function of be like in ­English reporting thought or speech (Romaine and Lange 1991), but unlike English it is not limited to adolescents’ speech. For the intimate relationship between thought and speech see Huang 1982. 10 khi3hoo7 去予 should rather be khit4hoo7 乞與, an agent marker as a compound comprising two synonymous agent markers khit4 乞 and hoo7 與. See Lien 2008. 11 I concur with Urmson (1952) that parentheticals are not a result of dysfuency, but a device of conveying their illocutionary force. There are important syntactic, semantic, and pragmatic properties worth in-depth probe (Dehe and Kavalova 2007:1–22). 12 Hoo7 予 can also be an agent marker in passives. 13 Tshiunn7kong2 像講 can be further preceded by na2 若, ho2 好, and tshin1 親 yielding the lexicalized collocates like na2tshiunn7kong2 若像講, ho2tshiunn7kong2 好像講, and tshin1tshiunn7kong2 親像講. This is a set of synonyms with similar but not entirely identical functions which, due to space constraints, will not be treated here. 14 For the notion of lexicalization see Brinton and Traugott 2005. 15 Heine and Kuteva (2002: 261–269) shows that verbs of saying in some languages in the world tend to take on a wide range of functions such as cause, complementizer, conditional, evidential, purpose, quotative, simile, and subordinator. TSM shares with these traits partially in that kong2 講 develops into a complementizer, an evidential, and a quotative, but if lexicalized collocates are taken into consideration, then it may also participate in the formation of conditional and cause.

REFERENCES Brinton, Laurel, and Elizabeth Traugott. 2005. Lexicalization and language change. Cambridge: Cambridge University Press. Chang, Miao-hsia張妙霞. 1995. Discourse function of Taiwanese kong in relation to its grammaticalization. In Selected papers from the second international symposium on languages in Taiwan, ed. by Shuanfan Huang, 111–127. Taipei: Crane Publishing. Chappell, Hilary. 2008. Variation in the grammaticalization of complementizers from verba dicendi in Sinitic languages. Linguistic Typology 12:45–98. Cheng, Robert L. 1997. The constructional markers shuō ‘say’ and kàn ‘see’ in Taiwanese and Taiwan Mandarin. In Taiwanese and Mandarin structures and their developmental trends in Taiwan II: Contacts between Taiwanese and Mandarin and restructuring of their synonyms, ed. by Robert L. Cheng, 105–131. Taipei: Yuan-Liu Publishing. Dehe, Nicole, and Yordanka Kavalova, eds. 2007. Parentheticals. Amsterdam: John Benjamins.

Kong2 as a verb for saying ‘on the move’ 215 Frajzyngier, Zygmunt. 1991. The de dicto domain in language. In Approaches to grammaticalization. Vol. I, Theoretical and methodological issues, ed. by Elizabeth C. Traugott and Bernd Heine, 219–251. Amsterdam: John Benjamins. Fraser, Bruce. 1980. Conversational mitigation. Journal of Pragmatics 4:341–350. Heine, Bernd. 2002. On the role of context in grammaticalization. In New reflections on grammaticalization, ed. by Ilse Wischer and Gabriele Diewald, 83–101. Amsterdam: John Benjamins. Heine, Bernd, and Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press. Hopper, Paul J. 1991. On some principles of grammaticalization. In Approaches to grammaticalization. Vol. I, Theoretical and methodological issues, ed. by Elizabeth C. Traugott and Bernd Heine, 17–35. Amsterdam: John Benjamins. Huang, Shuanfan 黃宣範. 1982. On the (almost perfect) identity of speech and thought: Evidence from Chinese dialects. In Paper presented at fourteenth international conference on Sino-Tibetan languages and linguistics, ed. by Chauncey Chu, W. South Coblin, and Feng-fu Tsao, 171–186. Taipei: Student Book. Lien, Chinfa 連金發. 1988. Taiwanese sentence-final particles. In The structure of Taiwanese: A modern synthesis, ed. by Robert L. Cheng and Shuanfan Huang, 209–240. Taipei: Crane Publishing. Lien, Chinfa 連金發. 2002. Grammatical function words 乞, 度, 共, 甲, 將 and 力 in Lì Jìng Jì 荔鏡記 and their development in Southern Min. In Papers from the third international conference on Sinology: Linguistic section, dialect variations in Chinese, ed. by Dah-an Ho, 179–216. Taipei: Institute of Linguistics Preparatory Office, Academia Sinica. Lien, Chinfa 連金發. 2008. Special types of passive and causative constructions in TSM. In Chinese linguistics in Leipzig, ed. by Redouane Djamouri, Barbara Meisterernst, and Rint Sybesma, 223–237. Paris: Centre de Recherches Linguistiques sure l’Asie Orientale, École des Hautes Édudes en Sciences sociales. Nuyts, Jan. 2001. Subjectivity as an evidential dimension in epistemic modal expressions. Journal of Pragmatics 33:383–400. Ransom, Evelyn. 1988. The grammaticalization of complementizers. Berkeley Linguistics Society 14:364–374. Romaine, Suzanne, and Deborah Lange. 1991. The use of like as a marker of reported speech and thought: A case of grammaticalization in progress. American Speech 66.3:27–279. Traugott, Elizabeth Closs, and Richard B. Dasher. 2002. Regularity in semantic change. Cambridge: Cambridge University Press. Tseng Minghua 曾明樺. 2008. The multifunction of Taiwanese Southern Min kong2. Master thesis, National Tsing Hua University, Taiwan. Urmson, J.O. 1952. Parenthetical verbs. Mind 61:480–496.

Appendix

Editor

Year

Title

Publisher

Abbreviation

宜蘭縣 胡萬川

2006 1998 1998

胡萬川、黃晴文

1996

胡萬川、黃晴文

1997

新社鄉閩南語故事集一 (台中縣) 新社鄉閩南語故事集二

胡萬川

1994

沙鹿鎮閩南語故事集

胡萬川

2005

沙鹿鎮閩南語故事集二

胡萬川、黃晴文

1996

清水鎮閩南語故事集一

胡萬川、黃晴文

1996

清水鎮閩南語故事集二

胡萬川 胡萬川 胡萬川、王正 雄、張裕宏 胡萬川、賴萬發 胡萬川、陳素 主、葉翠華 胡萬川 胡萬川、陳益源

1994 1998 1998 1995 1995

大甲鎮閩南語故事集 大安鄉閩南語故事集二 外埔鄉閩南語故事 集(台中縣) 彰化縣民間文學集五 彰化縣民間文學集七

宜蘭縣文化局 宜蘭市:宜蘭縣立 文化中心 苗栗縣立文化 中心 豐原巿:台中縣立 文化中心 豐原巿:台中縣立 文化中心 豐原巿:台中縣立 文化中心 豐原巿:台中縣立 文化中心 豐原：台中縣立 文化中心 豐原：台中縣立 文化中心 豐原巿:中縣文化 豐原巿:中縣文化 臺中縣立文化 中心 彰化縣文化局 彰化縣文化局

宜蘭二 羅陳

胡萬川

宜蘭縣民間文學集二 羅阿蜂、陳阿勉故事專 輯 (宜蘭縣) 苗栗縣閩南語故事集

2003 1999

南投縣福佬故事集一 雲林縣閩南語故事集一

南投一 雲林一

胡萬川、陳益源

1999

雲林縣閩南語故事集二

胡萬川、陳益源

1999

雲林縣閩南語故事集三

胡萬川、陳益源

1999

雲林縣閩南語故事集四

江寶釵 胡萬川

2000 2001

嘉義市民間文學集五 臺南縣閩南語故事集二

投縣文化局 雲林縣斗六市： 雲林縣文化局 雲林縣斗六市： 雲林縣文化局 雲林縣斗六市： 雲林縣文化局 雲林縣斗六市： 雲林縣文化局 嘉義市文化局 臺南縣文化局

苗栗 新社一 新社二 沙鹿 沙鹿二 清水一 清水二 大甲 大安二 外埔 彰化五 彰化七

雲林二 雲林三 雲林四 嘉義五 臺南二

10 Purposives in Taiwanese Southern Min Chinfa Lien and May Wang

1 INTRODUCTION In light of the theory of Construction Grammar elaborated in Fillmore, Kay, and O’Connor 1988, this paper aims to provide an analysis of purposive constructions in Taiwanese Southern Min (TSM). The study of Lin and Liao (2005) shows that lai 來-purposives are one of the purposive constructions in Mandarin Chinese and in TSM there exist not only lai-purposives but also khi 去-purposives. These deictic motion verbs can turn into a purposive marker while losing their lexical property of directionality and spatial indication function or still maintain them. Another type of purposive is (ti)leh-purposive which might be a TSM-specific purposive construction and doesn’t occur in Mandarin Chinese. (Ti)leh (佇)得 is a progressive marker and when used a purposive marker, the aspectual property may also disappear. Still another purposive is hoo-purposives in which hoo would indicate the subject as the beneficiary. Those three purposive constructions discussed in this paper are analyzed to bear the following properties: a b c

They are control constructions in which the controller may be in the sentential domain or in the discoursal domain. They possess the feature of temporal sequence. The pre-purposive clause is examined as means or instrument which is used to fulfill the goal described by the purposive clause.

This paper is organized as follows. Section two is the literature review of control construction and temporal sequence. Section three deals with the analysis of lai/khi-purposive, (ti)leh-purposive, and hoo-purposive. Section four discusses the distribution of pronoun i1 伊 and the empty category. The final section is the conclusion. 2  LITERATURE REVIEW Two points are to be examined in purposive clauses in TSM. The first point is that the purposive clauses can be regarded as control constructions which can be divided into two subtypes, the subject-controlled or the object-controlled structure

218  Chinfa Lien and May Wang in which PRO is the subject. The better candidate of the controller should be the nearest one and c-commend PRO. The second point involves the temporal sequence constraint conveying that the word order of the main clause and the purposive clause is under the effect of temporal sequence.

2.1  Control construction On Huang’s (1992) analysis, the resultative constructions in Mandarin Chinese should be regarded as control constructions instead of raising constructions. Control constructions can be divided into subject-controlled and object-controlled constructions. The choice between them is determined by whether it is the matrix subject or the matrix object that is nearest to PRO and c-commend it. So the resultative construction shows that the controller and controlee show the effect of Minimal Distance Principle (MDP). In addition, subject-controlled and objectcontrolled constructions have distinct characteristics. The first one is that passivization is only possible in object-control predicates and subject-controlled predicates cannot undergo passivization. (1) a 張三 把 馬 騎得 [PRO 很 zhāngsān bǎ mǎ qíde hěn Zhangsan ba horse ride-de very ‘Zhangsan rode the horse and got it tired’.

累] lèi tired

b 馬 被 他 騎得 [PRO 很 累] mǎ bèi tā qíde hěn lèi horse by him ride-DE very tired ‘The horse was ridden by him until it got very tired’. (2) a Bill promised [PRO to leave]. b *Bill is promised [PRO to leave]. Sentence (1a) is an example of object-controlled predicate which presents the passivization (1b) in which the object mǎ 馬 has to be moved to the subject position in order to get the nominative case. In D-structure of (1a), the object is the nearest possible controller of PRO, so it controls PRO and signifies the effect of MDP. Since there is no other NP between PRO and the subject Bill, so PRO in (2a) is co-referential with the subject Bill and (2a) is a subject-controlled structure. The ungrammaticality of (2b) shows that passivization is implausible in subject-controlled constructions. The second characteristic of control constructions is that the object can’t be omitted in subject-controlled sentence, but not in object-controlled one. Concerning the subject-controlled predicate (3a), the object is omitted and no other intervening possible controller between Zhāngsān 張三 and PRO, so it is no doubt that Zhāngsān 張三 is the controller. The object-controlled example (3b) reveals when

Purposives in Taiwanese Southern Min 219 the object mǎ 馬 exists, the meaning of (3b) is entirely different from that of (3a). The reading of (3a) is that the one who got tired is Zhāngsān 張三, but in (3b) it is the horse which got tired. So mǎ 馬 not only is the object or instrument of the verb qí 騎, but also the subject or the experiencer of the predicate hěn lèi 很累. Though from the S-structure, mǎ 馬 should be considered the controller of PRO in (3c), in Huang’s analysis the VP qímǎ 騎馬 in D-structure is an adjunct in which mǎ 馬 does not c-command PRO. Therefore that is why (3c) is still regarded as a subject-controlled sentence. The D-structure of (3c) is presented as (4). (3) a 張三 騎得 [PRO 很 zhāngsān qíde hěn Zhangsan ride-de very ‘Zhangsan rode and got tired’.

累] lèi tired



b 張三 騎得 馬 [PRO 很 zhāngsān qíde mǎ hěn Zhangsan rode-de horse very ‘Zhangsan rode the horse and got it tired’.



c 張三 騎 馬 騎得 [PRO 很 zhāngsān qí mǎ qíde hěn Zhangsan ride horse ride-de very ‘Zhangsan rode the horse and got very tired’.

累] lèi tired

累] lèi tired

(4) IP

NP

VP

Adjunct

V1

張三 zhāngsān Zhangsan

V’

NP

騎 qí ride

V2

馬 mǎ horse

Result

騎得 qíde ride-de

PRO

很 hěn very

累 lèi tired

220  Chinfa Lien and May Wang Now that qímǎ 騎馬 is an adjunct, so if the VP qímǎ 騎馬 is deleted, the D-structure of (3c) will be the same as that of (3a) and the meaning of (3c) also shows it is the subject Zhāngsān 張三 got tired instead of the object mǎ 馬. Huang (1992:119) states that ‘the lexical subject of the result clause is represented as the object, not of the matrix verb alone, but of a complex predicate containing the matrix verb and the predicate of the result clause. According to this analysis, Lǐsì in (31a) [(5) in this paper] is not directly represented as the subject of hěn shāngxīn ‘very sad,’ but as the object of the complex predicate kū-de Pro hěn shāngxīn meaning ‘cry so as to mark Pro very sad.’ The D-structure of (5) is presented as (6). (5) 張三 哭得 李四 很 傷心 zhāngsān kūde lǐsì hěn shāngxīn Zhangsan cry-de Lisi very sad ‘Zhangsan cried so much that Lisi got very sad.’ (6) IP

NP

VP

NPi

V’

V

RC

NPi

張三 zhāngsān Zhangsan

李四 lǐsì Lisi

哭得 kūde cry-de

VP

PRO

很 hěn very

傷心 shāngxīn sad

What the thematic roles Lǐsì 李四 displays are the affectee or the patient which is theta-marked by matrix verb kū 哭 and the agent of the result predicate hěn shāngxīn 很傷心. Tang (2000) distinguishes finite and non-finite clauses in Mandarin Chinese in terms of whether they allow overt NP/pro or PRO to be the subject and whether they can have the form of A-not-A questions. Table 10.1 shows the feature distinctions between finite and non-finite clauses.

Purposives in Taiwanese Southern Min 221 Table 10.1  Feature distinctions between finite and non-finite clauses

Finite clause Non-finite clause: Raising clause Exceptional case-marking clause Control clause

(non) temporal

(non) finite

Overt NP/ pro

PRO

A-not-A question

+

+

+

–

+

+ + +

– – –

+ + –

– – +

– – –

Table 10.1 clearly shows that control constructions are non-finite clauses in which they cannot form A-not-A questions and PRO is the only possible covert subject. Sentences (7a) and (7c) show that the nearest controllers of PRO are the subject tā 他 and the object wǒ 我. Therefore tā 他 in (7a) c-commands PRO showing a subject-controlled construction, and wǒ 我 in (7c) c-commands PRO exemplifying an object-controlled construction respectively. Inasmuch both (7a) and (7c) are control structures which are a subtype of non-finite clauses, they definitely cannot turn into A-not-A questions as shown in (7b) and (7d). (7) a 他 打算 [PRO 明天 來 看 tā dǎsuàn míngtiān lái kàn 3sg plan tomorrow come see ‘He plans to come to see you tomorrow’.

你] nǐ 2sg

b *他 打算 [PRO 明天 來 看 不 看 你] tā dǎsuàn míngtiān lái kàn bú kàn nǐ 3sg plan tomorrow come see neg see 2sg ‘He is planning whether to come to see you or not tomorrow’. c 他 要 我 [PRO 趕 快 回 tā yào wǒ gǎn kuài húi 3sg want 1sg hurry fast return ‘He wants me to go (home) immediately’.

去] qù go

d *他 要 我 [PRO 趕 快 回 不 回 tā yào wǒ gǎn kuài húi bù húi 3sg want 1sg hurry fast return neg returen ‘(?) He wants me whether or not I go (home) immediately’.

去] qù go

Chomsky and Lasnik (1993:20) mentioned ‘all PRO are controlled if apparent uncontrolled PRO actually has a hidden controlled. . .’ and it is called PROarb. Therefore when PRO cannot find its antecedent in the sentential domain, the

222  Chinfa Lien and May Wang antecedent of PRO might be an arbitrary coreferent or discourse-related one in other term (Chomsky 1981; Růžička 1999). (8) Ann was just wondering [CP howi [PRO to do it ti]] The controller of PRO in (8) can be the matrix subject Ann, and the interpretation is that Ann was wondering if she could come up with a certain way to have the thing done. Another possible controller might be a juggler or some one else which is indicated by the discourse and the reading is that Ann was wondering how the juggler or someone performed the magic trick.

2.2  Temporal sequence The principle of temporal sequence (PTS) is defined by Tai (1985) as ‘the relative word order between two syntactic units is determined by the temporal order of the states which they represent in the conceptual world’ (Tai 1985:50). This principle also takes effect in serial verb construction (SVC) in Mandarin Chinese. (9)

a 張三 上 樓 睡覺 zhāngsān shàng lóu shuìjiào Zhangsan go.up stair sleep ‘Zhangsan went upstairs to sleep’. b *張三 睡覺 上 樓 zhāngsān shuìjiào shàng lóu Zhangsan sleep go.up stair *‘Zhangsan slept to go upstairs’.

The ungrammaticality of (9b) is because it’s impossible for Zhāngsān 張三 to do the action going upstairs after he sleeps (unless he sleepwalks). So the common knowledge tells us that Zhāngsān 張三 firstly performs the action going upstairs and then continues to complete the second action going to bed. The sequence of action-taking follows the temporal order. In addition to PTS in serial verb constructions in Mandarin Chinese, SVCs in TSM also show the effect of PTS (Chang 1990; Tai 1985; Yen 2004). (10) 提 去 買 豬肉 theh8 khi3 be2 ti1bah4 take go buy pork ‘Take (money) to buy (some) pork’. The first predicate theh8khi3 提去 must precede the second predicate be2ti1bah4 買 豬肉 in word order which reveals the time order of events happening. Because of the restriction of temporal sequence in SVC, the word order can’t be reversed.

Purposives in Taiwanese Southern Min 223 In the discussion of the distinction between finite and non-finite clauses, Tang (2000) also mentioned the control complement clause usually indicates the possible future of irrealis. Therefore, the event indicated by the control complement clause has to follow the event illustrated by the matrix verb in terms of the restriction of PTS.

2.3 Purposives Previous studies have extensively discussed the purposives in Mandarin Chinese. In Lin and Liao 2005, they discussed three types of purposives and offered a structural analysis to distinguish these three types of purposive structures. Hǎo-purposive is a bi-clausal construction, so the matrix clause and hǎo-purposive clause are two conjuncts which are conjoined by a null conjunction marker. Bare purposive is an adjunct and lái-purposive is analyzed as the clausal complement of the matrix verb. These three types of purposive clauses are all control constructions in which the subject gap (PRO) is coreferential with the subject in the matrix clause. (11) a 張三i 買了 一個 漢堡j zhāngsān mǎile yíge hànbǎo Zhangsan buy-perf one-clf burger ‘Zhangsan bought a burger to eat’.

[

來 PROi lái lai

吃 ej] chī eat

b 張三i 買了 一個 漢堡j [ 來 PROi 討好 李四*j/k] zhāngsān mǎile yíge hànbǎo lái tǎohǎo lǐsì Zhangsan buy-perf one-clf burger lai please Lisi ‘Zhangsan bought a burger to please Lisi’. (11) are examples of lái-purposives. There exists an obligatory subject gap in lái-purposive which is presented by PRO and PRO is necessarily controlled by the matrix subject Zhāngsān 張三. Concerning MPD, since there exists an object hànbǎo 漢堡 in (11a) which is much closer to PRO, so Zhāngsān 張三 should not be the proper controller of PRO. However, the actual reading of (11a) prefers PRO to be co-indexed with Zhāngsān 張三 and this might suggest that hànbǎo 漢堡, in D-structure, should not c-command PRO. In the consideration of the object gap, if the object in purposive clauses is coreferential with the object in the matrix clause, the object in lai-purposive must be omitted. However, when the object in lái-purposive clause and the object in the pre-lái-purposive don’t have the same referent, the object in lái-purposive is required to have the overt existence as in (11b). 3  PURPOSIVES IN TAIWANESE SOUTHERN MIN In this section, we will examine three types of purposive constructions in TSM. Each type of purposive can be analyzed as a control construction which can be subdivided into subject-control and object-control constructions. In the purposive construction,

224  Chinfa Lien and May Wang there occurs the relation of temporal sequence between the matrix clause and the purposive clause. The purposive clause features the possible future of irrealis and the time order must be later than the matrix clause. Even though both matrix clause and purposive clause describe the event happening in the past time, the purposive clause in the presentation of temporal sequence is much closer to the present time. Therefore based on the same concept, when both matrix clause and purposive clause present the future time, the matrix clause is much closer to the present time and the time indicated by the matrix clause can be considered the benchmark.

3.1  Lai/khi-purposives The deictic motion verb lai/khi in purposive clauses may lose the deictic function or still retain it. Lai signifies the meaning of ‘from far away to close to the speaker’ and khi presents ‘from close to far away from the speaker (Chen 2006; Lien 2003). The construction of lai/khi purposives is presented in (12). In the purposive construction, the matrix clause presents the means or instrument that the subject agent will adopt in order to achieve the goal which is indicated by the purposive clause. Y presents the object which requires the optional existence. (13) are examples of lai/khi purposives. (12) S V1 O lai/khi PRO V2 Y (ti)leh (佇)得 (13) a 汝 哪會 無 佫 掠 塗虱 來 li2 na2e7 bo5 koh4 liah8 tho5sat4 lai5 2sg why neg again catch milk.fish come ‘Why do you not catch the milk fish to sell again’? lii  liah    thosatj  lai    [PROi  bue    ej] b 日本人 jip8pun2lang5 Japanese



請 tshiann2 invite

咱 lan2 1pl.incl

講 kong2 call

北港 pak4kang2 Pak.Kang

賣？ bue7 sell

媽祖 ma2tso2 Ma.Tso

去 khi3 go

服侍 日本 jip8pun2 hok8sai7 Japan serve ‘Japanese devoutly invited the Goddess that we called Pak-Kang Ma-Tso to Japan for worship’. jip pun langi tshiann pakkang matsoj khi [PROi hoksai ej]

(13a) and (13b) are both subject-control purposives in which the matrix subject and PRO are co-referential. Lai5 來 in (13a) is a simple purposive marker which loses its deictic function, but in (13b) khi3 去 not only is a purposive marker but also presents its spatial and directional function that the use of khi3 去 indicates the motion

Purposives in Taiwanese Southern Min 225 is away from the speaker. Lai5 來 and khi3 去 can extend their meanings that they express the speaker’s attitude toward the event, in other words, whether the speaker shows the enthusiasm about the event or not. The speaker in (13a) truly cares about why the addressee doesn’t sell the fish again and lai5 來 expresses the involvement of the speaker in the event. The speaker in (13b) expresses the idea that Japanese invite the goddess ma2tso2 媽祖 deferentially to Japan for them to worship. In this event, the speaker isn’t involved, so the choice between lai5 來 and khi3 去 would reflect the speaker’s attitude toward the event and might also depend on whether the speaker would be affected by the event or not (Anand and Hsieh 2005; Chen 2006)1. Therefore, the event li2 na2e7 bo5 koh4 liah8 tho5sat4 lai5 bue7 汝無佫掠塗虱來賣 in (13a) will influence the speaker so that if the addressee doesn’t catch fish to sell, the speaker couldn’t buy the big and fresh fish from the addressee. In (13b), the speaker just describes the event performed by Japanese only. The matrix clause and the purposive clause in (13a-b) follow the restriction of temporal sequence that the time indicated by the matrix clause must precede the time expressed by the purposive clause. In the purposive constructions, the matrix clause is regarded as the instrument or means that is used to carry out the goal which is as a matter of course indicated by the purposive clause. In (13a), the speaker takes the means ‘fishing’ in order to fulfill the goal ‘selling fish’ and in (13b) the agent Japanese adopt the means ‘inviting with respect’ in order to carry out their goal ‘worshiping’. When the subject has been mentioned or understood in the context, the purposive seems to be subjectless. But actually in D-structure, a covert subject presented in the form of e occupies the spec of IP position. The context of (14) shows the one who gets the power to arrest the killer and prisons him is a policeman; therefore, the subject of VP liah8 khi3 掠去 is understood as a policeman. We may say that the covert subject is a pragmatic-determined empty form and PRO which is controlled by the subject can be PROarb. The matrix object is moved to the preverbal position because of the patient marker ka7 共 which assigns the patient role to i1 伊. In the purposive clause, PRO is controlled by the matrix subject showing (14) is a subject-controlled construction. Because the mitted object of kuinn1 關 is co-referential with the matrix object i1 伊, according to the obligatory object gap mentioned in Lin and Liao 2005, the object in the purposive clause requires a covert form. The matrix clause presents the means or manner of how the policeman achieves the goal ‘prison the killer’ and the matrix and purposive clauses also follow the temporal sequence restriction. (14) 煞 去 suah4 khi3 unexpectedly go 啊 a2 prt

共 ka7 ka

人 就 lang5 tio7 person then

人 拍死 一个 lang5 phah4si2 tsit8e5 person beat.die one-clf

人 啦， lang5 la2 person prt

共 ka7

啊 a2

ka

伊 i1 3sg

告， ko3 accuse

prt

就 tio7 then

226  Chinfa Lien and May Wang 共 ka7



伊 掠 去 關 啦 i1 liah8 khi3 kuinn1 la2 ka 3sg arrest go prison prt ‘Unexpectedly he killed one person by beating and someone accused him of killing, so a policeman arrested him and sent him to the prison’. ei ka ij liah tj [khi PROi kuinn tj]

TSM also has the object-controlled constructions as (15) shows. PRO in the purposive clause is controlled by t which is the object of liah8 掠. The matrix object tsit4 e5 mo1hi5e5 這個摸魚的 at first is in the postverbal position, but later it is moved to the preverbal position because of ka7 共 which also assigns the patient role to the object. Under the topicalization or in order to get the focus, the matrix object again moves to the sentence-initial position. E, the object of kau1thue3 交替 in the purposive clause, refers to the subject or the agent who performs the action liah8 tsit4 e5 mo1hi5e5 掠這個摸魚的. Notwithstanding the subject is not present in the sentence, the context will tell us e refers to the subject. (15) 「啊 當 a2 tong1 prt when 共 無 bo5 ka7

時間 到 囉， 啊 這 個 摸魚的 si5kan7 kau3 lo3 a2 tsit4 e5 mo1hi5e5 time reach prt prt dem clf fisherman 掠 來 交替」 liah8 lai5 kau1thue3 neg ka catch come replace ‘When it is time (for rebirth), why don’t I just catch the fisherman to replace me’. tsit e mo hi ei KA ti liah ti [lai PROi kauthue ej]

3.2  (Ti) leh-purposives The progressive marker (ti)leh (佇)得 can also be used as a purposive marker and in the construction of (ti)leh-purposives there will be two readings which will be determined by the context. The construction of (ti)leh-purposive and examples of it are in the followings. (16) [S V1 O] (ti)leh (佇)得 [PRO V2 Y] (17) a 伊 安呢 逐日 得 掠 水雞 得 i1 an2ne1 tak8jit8 leh4 liah8 tsui2kue1 leh4 3sg this.way every.day prog catch frog prog ‘He makes a living by catching frogs every day’. ii liah tsuikue le [PROi leh singuah]

生活 sing1uah8 live

Purposives in Taiwanese Southern Min  227 b 阿 a2



阮 寒 到 尾仔 著愛 拍 火把 佇得 烘 gun2 kuann5 kah4 bue2a2 tioh8ai3 phah4 hue2pa2 ti7leh4 hang1 prt 1pl.excl cold so finally need light torch prog warm ‘I was so cold that finally I lighted the torch to warm myself’. guni phah huepa tileh [PROi hang ei]

c 透 早 人 佇得 擔 竹筍 佇得 賣 thau3 tsa2 lang5 ti7leh4 tann1 tik4sun2 ti7leh4 bue7 early morning people prog carry bamboo.shoot prog sell ‘Early in the morning people were carrying bamboo shoots to sell’. langi tann tiksunj tileh [PROi bue ej]

The matrix subject and PRO in (17a) are co-referential and therefore feature a subject-control construction. The first predicate indicates that the subject i1 伊 chooses the means ‘catching frogs’ for the purpose to keep his living. The aspectual function of (ti)leh (佇)得 in this sentence is bleached. The event ‘catching frogs’ is done by the agent every day and the purpose ‘keeping his living’ is a lifelong aim, instead of a temporal one, so (17a) presents the generic or habitual reading. (17b) is also a subject-control construction in which the matrix subject gun2 阮 and PRO share the same referent, but the covert object in the purposive clause does not refer back to the object in the matrix clause, but body parts of the subject which is a discourse-related referent (Her 2006)2. Since the omitted object is not coreferential with the matrix object, it can be overtly presented in the purposive clause such as phah4 hue2pa2 ti7leh4 hang1 tshiu2/sin1khu1 拍火把佇得烘手/身 軀. From the context of (17b), it reveals that the agent only lights a troch in the wintertime especially during the time of cold current, so the event phah4 hue2pa2 ti7leh4 hang1 拍火把佇得烘 only happens during a short time which requires the episodic interpretation and (ti)leh (佇)得 still retains its progressive reading. (17c) is kind of difficult to decide whether it has the general reading or the episodic reading. From the perspectives of the speaker who is a scholar, the event tann1 tik4sun2 ti7leh4 bue7 擔竹筍佇得賣 occurring outside his room only happened in the specific time when he was thinking of the couplet that day and this purposive sentence has the episodic reading. However, for the agent himself, the event tann1 tik4sun2 ti7leh4 bue7 擔竹筍佇得賣 is what he should do every day in order to sell out the bamboo shoots for earning a living which is the ultimate goal suggested from our conceptual world. So (ti)leh-purposive in (17c) has two readings: in the episodic interpretation, (ti)leh (佇)得 still has the function of being a progressive marker, but in the generic reading (ti)leh (佇)得 is a simple purposive marker. (17c) is a subject-control construction in which the matrix and PRO are co-referential and the obligatory object gap is required in the purposive clause because of the condition that matrix object and the object in purposive clause have the same referent. The matrix clauses in (17) generally represent as the means or instrument and the time order of it is regarded to precede the time indicated by the purposive clause. So in (17a), the subject agent i1 伊 must firstly catch frogs and then he can earn a living.

228  Chinfa Lien and May Wang

3.3  Hoo-purposives The last type of purposives in need of discussion in TSM is hoo-purposive. Hoo7 予 has been observed to occur in several constructions: double object construction, dative construction, serial verb construction (SVC), passive construction, and causative construction (Cheng et al. 1999). In the SVC, the first predicate is the means that is taken by the agent to achieve the goal indicated by the second predicate. 唱 一 首 歌j [PROi (18) 我i gua2 tshiunn3 tsit8 siu2 kua1 1sg sing one clf song ‘I sang a song for you to listen to’.

hoo [ hoo7 give

汝 li2 2sg

聽 tj]] thiann1 listen

The SVC is a control construction that the matrix subject and PRO are co-­ referential and the omitted object is co-indexed with the matrix object. The major difference between the analysis of Cheng et al. (1999) and Her (2006) is that Cheng et al. considered (18) to be a subject-control construction, but in Her’s analysis (14) is an object of hoo-control construction. The major concern of this paper is to provide the analysis that hoo in TSM can function as a purposive marker. (19) [Si V1 O] [PROi hoo X V2 Y] 分 一點仔 予 我 (20) a 「 我 麼 無 食， gua2 ma7 bo5 tsiah8 pun1 tsit8tiam2a2 hoo7 gua2 1sg too neg eat divide a bit give 1sg 無？」 食 好 tsiah8 ho2 bo5 eat okay neg ‘I didn’t eat, either. Is it okay that you give me some (food) to eat’? ei pun tsittiamaj [PROi hoo gua tsiah tj] b 無 bo5 neg



看 就 好 哦！ khuann3 tio7 ho2 o0 see just okay prt

‘Wouldn’t it be okay if I open the city gate for you to watch’? guai khui siann mngj ‑[PROi hoo li khuann ek] c 啊 a2



我 開 城 門 予 你 gua2 khui1 siann5 mng5 hoo7 li2 1sg open city gate give 2sg

底 起 厝 予 住。 第三 願， 海 te7sann1 guan7 hai2 te2 khi2 tshu3 hoo7 tua3 prt third wish sea bottom build house give live ‘The third wish is (God) builds a house in the seabed for him to live’. ei hai te khi tshu [PROi hoo tj tua]

Purposives in Taiwanese Southern Min 229 The construction of hoo-purposive is slightly different from lai/khi-purposive and (ti)leh-purposive constructions. In the hoo-purposive, usually the matrix subject is not co-referential with the subject in the purposive clause and the subject of purposive clause can overtly present or be omitted, so X in the construction (19) can be overt or covert NP. The subjects in purposive clauses in (20a) and (20b) are overtly presented and they do not have the same referent with the matrix subjects. Though (20a) seems to be a subjectless sentence, yet the subject can be determined from the context that the interlocutor is the subject of the main clause. Since the object of the purposive clause is co-referential with the matrix object, the object of the purposive clause is obligatory left out. The first clause presents the means that the addressee would adopt to attain the aim that the speaker could eat something. There doesn’t occur the co-referential relation between the matrix subject and the subject of purposive clause in (20b) and the referent of the object in the purposive clause and matrix object isn’t exactly the same one. The object of khuann3 看 is not the city gate but what is in the inside of the city gate and the antecedent of the object of khuann3 看 is a discourse-related one or it is inspired by the conceptual world. The purpose PRO hoo7 li2 khuann3 予你看 is achieved by the event indicated from the main clause. In (20c), the matrix subject who is inferred from the context doesn’t share the same referent with the subject who is not overtly occurred in the purposive clause. The covert subject in purposive clause is also pragmatic-determined. Since the matrix subject is context-determined, the PRO which is under its control can be also PROarb. Therefore, though PRO can’t find its antecedent in the sentential domain, yet it can find its antecedent in the discoursal domain. The context indicates that it is the god who is the subject of the matrix clause and controls PRO. The matrix clause khi2tshu3 起厝 is the means that used by the matrix subject to fulfill the goal hoo7tua3 予住. With regard to the temporal sequence, the time slot presented in the purposive clauses in (20) should be later the time period of the matrix clause. Take (20b) as an example. Although the matrix and purposive clauses are both about the future, the time indicated by the matrix clause is much closer to the present time the time presented by the purposive clause. PRO in (21) is controlled by the matrix subject e which is pragmatic-­determined, so PRO is also considered PROarb. The subject in the purposive clause is not overtly presented in the S-structure, but in D-structure the purposive subject is in an empty form which is co-index with the i1 伊 in the matrix clause. Ka7 共 in (21) is a beneficiary marker instead of a patient marker, and assigns the beneficiary role to i1 伊. We might think t is the agent of the predicate khun3 睏, but at the same time being the beneficiary of the matrix predicate. (21) 共 ka7

伊 創 得 眠床 予 i1 tshong3 leh4 bin5tshng5 hoo7 ka 3sg make leh bed give ‘(Someone) made the bed for him to sleep’. ka ij ei tshong leh bintshng [PROi hoo tj khun]

睏。 khun3 sleep

230  Chinfa Lien and May Wang In the hoo-purposives, lai can coexist with hoo to form a purposive clause. Two purposive markers that coexist can fortify the purposive reading and lai may also indicate the spatial and directional meaning. Lai in (22) reveals that the movement of the objects tsit8 e5 sian1li2 一個仙女 and tsit4 e5 ioh8tshau2 這个藥草 goes from far away or from other place to the place where the purposive subject i1 伊 stays. PRO in (22a) is an arbitrary one whose controller is context-determined, while the controller of PRO, i1 伊 in matrix clause, in (22b) occurs overtly. (22b) is a double-purposive construction in which the first goal of the predicate the theh8 liau2 tsit4 e5 ioh8tshau2 提了這个藥草 is lai5 tsian1 來煎, and lai5 hoo7 i1 tsiah8 來予伊食 is the second goal of the first purposive construction theh8 liau2 tsit4 e5 ioh8tshau2 lai5 tsian1 提了這个藥草來煎. The purposive clauses also follow the temporal sequence. After the first goal lai5 tsian1 來煎 is achieved, the second goal lai5 hoo7 i1 tsiah8 來予伊食 can be performed. Though there are two i1 伊 in (22b), these two i1 伊 have different referents. The first i1 伊 is the agent of the predicate theh8 liau2 tsit4 e5 ioh8tshau2 提了這个藥草 PRO lai5 tsian1 來煎 in which i1 伊 controls PRO. The second i1 伊 not only is the agent of the predicate tsiah8 食 but also the beneficiary of the event i1 tio7 theh8 liau2 tsit4 e5 ioh8tshau2 lai5 tsian1 伊 就提了這个藥草來煎. (22) a 有 u7 have



講 卜 kong2 beh4 say want

送 一 sang3 tsit8 send one

个 e5 clf

仙女 來 予 伊 sian1li2 lai5 hoo7 i1 fairy come give 3sg

做 某 tso3 boo2 make wife ‘(One god) had said he would send one fairy for him to be his wife’. ei beh sang tsit e sianlij [lai [PROi hoo i tj tso bo]]

b 伊 就 提 了 這 个 藥草 來 煎 i1 tio7 theh8 liau2 tsit4 e5 ioh8tshau2 lai5 tsian1 3sg then take asp dem clf herb come decoct 食 來 予 伊 lai5 hoo7 i1 tsiah8 come give 3sg eat ‘Then he took the herb to decoct and gave it for him to take’. ii theh liao tsit e iohtshauj [lai PROi tsian tj [lai PROi hoo i tsiah tj]] 4  DISTRIBUTION OF PRONOUN AND EMPTY CATEGORY When there exists the topicalization, the third person pronoun i1 伊 will not occur in the position following the patient marker ka7 共, but instead what is in the postka position is the empty category.

Purposives in Taiwanese Southern Min 231 (23) a 一 寡 神明 攏 是 共 掠 tsit8 kua2 sin5bing5 long2 si7 ka7 liah8 one few god all cop ka take ‘Some statues of gods were taken to be burned’. b 「啊 a2



當 tong1 when 無 bo5

去 khi3 go

燒。 sio1 burn

時間 si5kan7 time 共 ka7

到 囉， 啊 這 個 kau3 lo3 a2 tsit4 e5 prt reach prt prt dem clf 掠 來 交替 . . . 摸魚的 mo1hi5e5 liah8 lai5 kau1thue3 fisherman neg ka catch come replace ‘When it is time (for rebirth), why not I just catch the fisherman to replace me’.

The object tsit8kua2 sin5bing5 一寡神明 and tsit4 e5 mo1hi5e5 這個摸魚的 in (23a) and (23b) are firstly in the post-verbal position as liah8 tsit8kua2 sin5bing5 掠一寡 神明 and liah8 tsit4 e5 mo1hi5e1 掠這個摸魚的 in D-structure and they are moved to the preverbal position because of the patient marker ka7 共 in the construction of ka7 共 O V in which O and V are object and verb respectively. Objects would further be moved forward forming the topicalization as a result of (23a-b). In the topicalization, only empty category can be in the post-ka position because of the effect of control; therefore, pronoun i1 伊 cannot occur in the position following ka. The control construction is presented as the follows. (24) [Oi ka共 ei V1 ei] lai 來/khi 去 [PRO V2 e] In the matrix clause the empty category presenting as e is under the control of the object and the co-indexation reveals the co-reference relation. PRO in purposive clause is controlled by the subject agent and the subject agent should be discoursed inferred controller as (25a) shows. (25b) reveals PRO might also be co-indexed with the object in the matrix clause and the empty category in the purposive clause refers to the matrix subject. When the object is moved to the post-ka position and the movement is done, what follows ka is usually the pronoun i1 伊 instead of the empty category. (25) a 啊 a2

就 共 伊 掠 去 關 啦 tio7 ka7 i1 liah8 khi3 kuinn1 la3 prt just ka 3sg arrest go prison prt ‘(A policeman) just arrest him and sent him to prison’.

b 啊 a2

煞 共 伊 掠 去 受罪 囉。 suah4 ka7 i1 liah8 khi3 siu7tsue7 lo3 prt just ka 3sg arrest go punishment prt ‘(Someone) just arrested him in order for him to be punished’.

232  Chinfa Lien and May Wang The pronoun i1 伊 in the post-ka 共 position usually refers to the patient bearing the feature [+human], so when i1 伊 appears as an object, the sentence would be kind of odd. If the entity in the position following ka7 共 is an object, then the entity is usually presented as a NP instead of a pronoun. (26) 伊 共 錢 拐 i1 ka7 tsinn5 kuai2 3sg ka money swindle ‘He swindled the money’.

去 khi3 go

彼號 hit4lo7

啊 a7

hitlo

prt

5 CONCLUSION Lai 來/khi 去, (ti)leh (佇)得, and hoo 予 can be regarded as purposive markers. In the purposive constructions, they may still maintain their function or lose the function to become a simple purposive marker. These three types of purposive constructions are all control constructions and exhibit the order of time sequence: the matrix clause must precede the purposive clause in word order or time order and the purposive clause usually indicates the possible future of irrealis. Although lai 來/khi 去, (ti)leh (佇)得, and hoo 予 represent different markers and have distinct functions, but once they enter the purposive construction, they all turn into markers of purposives. NOTES 1 The speaker regarded as the entity x who might be affected by the event of the purpose clauses (Anand and Hsieh 2005:45). a  Bill took a chair to hit John’s statue (which would affect me). b  Bill took a chair to hit John’s statue (which would not affect me). 2 The controller/antecedent of the second predicate can be discoursally determined (Her 2006:1293). The antecedent of e is the part of soup but not the soup in its entirety. 她 放了 很多 佐料j 在 湯i裡, ei 很 tā fàngle hěnduō zuǒliào zài tānglǐ hěn 3sg put-asp lots.of ingredient at soup-indise very ‘She put lots of ingredients in the soup; it (=soup) is nice to drink’.

好喝。 hǎohē good-drinking

REFERENCES Anand, Pranav, and Feng-fan Hsieh. 2005. Long-distance reflexives in perspective. In Proceedings of the 24th West Coast conference on formal linguistics, ed. by John Alderete, Chung-hye Han, and Alexei Kochetov, 43–51. Somerville: Cascadilla Proceedings Project. Chang, Claire Hsun-huei 張郇慧. 1990. On serial verbs in Mandarin Chinese: VV compounds and co-verbial phrases. Ohio State University Working Papers in Linguistics 39:288–315.

Purposives in Taiwanese Southern Min 233 Chen, I-Chun. 2006. A study of semantics and constructions of deictic motion verbs ‘Lai5’ and ‘Khi3’ in Taiwanese Southern Min. Master thesis, Nation Tsing Hua University, Taiwan. Cheng, Lisa L-S., C-T. James Huang, Y-H. Audrey Li, and C-C. Jane Tang. 1999. Hoo, hoo, hoo: Syntax of the causative, dative and passive constructions in Taiwanese. In Contemporary studies on the Min dialects, ed. by Pang-Hsin Ting, 146–203. Berkeley: Project on Linguistic Analysis, University of California. Chomsky, Noam. 1981. Lectures on government and binding. Dordrecht: Foris. Chomsky, Noam, and Howard Lasnik. 1993. Principles and parameters theory. In Syntax: An international handbook of contemporary research, ed. by Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld, and Theo Vennemann, 506–569. Berlin: Walter de Gruyter. Fillmore, Charles J., Paul Kay, and Catherine O’Connor. 1988. Regularity and idiomativity in grammatical constructions: The case of let alone. Language 64:501–538. Her, One-Soon. 2006. Justifying part-of-speech assignments for Mandarin gei. Lingua 116:1274–1302. Huang, C-T. James. 1992. Complex predicates in control. In Control and grammar, ed. by Richard K. Larson, Sabine Iatridou, Utpal Lahiri, and James Higginbotham, 109–147. Boston: Kluwer Academic Publishers. Lien, Chinfa 連金發. 2003. Shíliù shìjì jí xiàndài mǐnnányǔ zhǐshì dòngcí de yǔfǎhuà十 六世紀及現代閩南語指示動詞的語法化 [The grammaticalization of deictic verbs in 16th century and modern Southern Min]. Guójì Zhōngguóxué Yánjiū 國際中國學研究 6:379–410. Lin, Tzong-Hong Jonah, and Wei-Wen Roger Liao. 2005. Purposives in Mandarin Chinese and phrase structure. Ms., National Tsing Hua University. Růžička, Rudolf. 1999. Control in grammar and pragmatics: A  cross-linguistic study. Amsterdam: John Benjamins. Tai, James H-Y. 1985. Temporal sequence and Chinese word order. In Iconicity in syntax, ed. by John Haiman, 49–72. Amsterdam: John Benjamins. Tang, Ting-chi. 2000. Finite and nonfinite clauses in Chinese. Language and Linguistics 1:191–214. Yen, Hsiu-Shan. 2004. Serial-verb construction in Taiwanese Southern Min. MA thesis, National Hsiuchu University of Education, Taiwan.

Index

Note: Page numbers in italic indicate a figure and page numbers in bold indicate a table on the corresponding page. ablative functions 111, 118 action verbs 145, 167 – 169, 177, 191 adverbs 1 – 4; adverbial comparatives 10, 12, 13, 14, 17, 17 – 18, 26 – 33, 42, 45; adverbial comparatives with khah4 可 in 10, 12, 26 – 31, 42, 45; degree 2 – 3; expressions of time and te3 處 175 – 176, 178; as an kang7 共 4 – 5, 102 – 104, 103, 106, 108; kong2 講 as 197; locative 178; negative 1; as an tioh8 著adverbial marker 53 – 54, 53 affective markers 141n6 agent markers 4 – 5; as an hoo7 予144 – 145, 145, 147 – 148, 148, 214n10, 214n12 ah0 啊: as a discourse marker 213n2 ai3 愛 56, 72, 73n3 All-Embracing Spring see Măn Tiān Chūn 滿天春 Amoy 廈門 (Xiamen) dialect 12, 27, 144, 157 – 158, 166, 194n7 Analects of Confucius, The (Lúnyǔ 論語) 103, 104, 120n13 Archaic Chinese 6; comparatives in 25 – 26; coordinating conjunctions in 102 – 106, 108 – 109, 119, 119 Arte de la lengua Chiõ Chiu (Grammar of the Chiõ Chiu Language) 16, 18, 27, 30, 34n6, 120n4 bǎ 把 2, 118, 121n29, 227 Băi Huā Sài Jǐn 百花賽錦 (Exquisite Brocades of a Hundred Flowers) 15, 16, 23, 43 – 45, 48 bat4/pat4 捌: as an experiential marker 214n8

benefactive markers 2, 5; definition of 128; kang7 共 5, 125 – 141, 126, 130, 132, 134 – 135, 138, 140 Blessings of Two Generations (Liǎngshì Tóngxiū 兩世同修) 112, 117 bo22 無 4, 56, 79 – 84, 192 Book of Odes (Shījīng 詩經) 103, 104, 109 – 110, 120n9, 120n10, 120n12 bronze inscriptions (Jīnwén 金文) 120n11 bu5pi2 無比 18 Cambodian language 96 – 97 Cantonese 4; comparatives in 10, 22; negation of dynamic modals in 76 – 77, 84, 92 – 98, 98 causative constructions (CC) 52, 72, 146 – 150, 149, 151, 152, 152 – 155, 154 causative verbs 1 – 5; benefactive marker kang7 共 and 134, 138, 139; directive 1, 211; hoo7 予 as a 144, 145, 148, 148, 153; kong2 講 as a 211; stative 144; tioh8 著 as 49 – 52, 57, 63 – 72, 67, 69, 71, 73n5, 73n6 Central Min (Minzhong) 26 – 27 Cháozhōu dialect: continuative aspect marker in 167 – 168, 181 – 182; corpus of 16, 167 – 168 Chinese language see Gan language subgroup; Hakka; Mandarin; Southern Min; Xiang language subgroup; Yue Christian Doctrine in the Alphabet and Language of China, The (Doctrina Christiana en letra y lengua china) 12, 16 – 18, 17, 33, 34n5, 34n7, 46, 48, 132, 140 – 141

Index  235 clauses: clause connectors 53, 53; clausetaking complementizers 197, 200; comparative 11, 20; conditional 90; factive 203; finite vs. non-finite 6, 203, 220 – 223, 221; interrogative 199 – 200, 203 – 204; purposive 6, 217 – 218, 223 – 232; serial verbs 146; transitive 26 cognates 26, 53, 57 comparative constructions 2, 9 – 12, 10, 11, 13, 14; absence of markers in 24; adverbial 10, 12, 13, 14, 17, 17 – 18, 26 – 33, 42, 45; in Archaic Chinese 25 – 26; clauses 11, 20; in Cantonese 10, 22; Compare comparatives 10, 11, 17 – 19, 21, 23, 29 – 30, 33, 39, 43; in contemporary Southern Min 30 – 33; dependent-marking 10 – 11, 18 – 19, 33; diachronic evaluation of four comparative structures 17 – 30; in Early Southern Min (ESM) 10 – 11, 10, 17, 17 – 30, 39 – 48; of equality 9, 34n10; in Hakka 22, 26, 32 – 33; head-marking 11, 19 – 22, 29 – 31, 33; hybridized 11, 11, 30 – 33; m7 不 in 11, 18; morphemes of 9, 21, 28 – 29, 111; negation in 33n1; nouns in 3, 17, 21, 24, 28 – 29; in the Quanzhou dialect 16, 28, 31; with sì 似 3, 10, 11, 13, 14, 17, 17 – 23, 29 – 30, 40 – 41, 45; subject-verb-object (SVO) word order in 29, 33; of superiority 9, 23 – 24, 27, 29 – 30, 33n1; syntax in 26; in Taiwanese Southern Min (TSM) 31, 33; zero-marked 10, 12, 13, 14, 17, 17, 24 – 26, 28 – 29, 41, 45; see also Surpass comparative; Surpass-Similarity comparative complementizers 2, 6; clause-taking 197, 200; Ewe complementizer bé 50; kong2 講 as a 197 – 200, 203 – 207, 213n3, 214n8, 214n15 complex ditransitive constructions (CDC) 149, 149, 153 – 155, 154 Confucius 103, 104, 120n13 conjunctions 2, 4 see also kah4/kap4 合; kang7 共 construction grammar 127 – 128, 139 – 140, 217 constructions: causative constructions (CC) 52, 72, 146 – 150, 149, 151, 152, 152 – 155, 154; complex ditransitive constructions (CDC) 149, 149, 153 – 155, 154; double object 5, 145, 149, 150 – 152, 151, 152, 228;

existential vs. non-existential 194; serial verb constructions (SVC) 3, 10, 21, 30, 138, 139 – 140, 146, 222, 228; see also comparative constructions; dative constructions (DC); locative constructions; purposive constructions continuative aspect markers 2, 5 – 6, 157 – 194, 162, 163, 165, 193; in the Cháozhōu dialect 167 – 168, 181 – 182; in locative constructions 159, 162, 165 – 166, 169 – 170, 173 – 174, 177 – 193; see also leh4 咧; te3 處; ti7te3 在處 Dame Su Liu Niang (Sū Liù Niáng 蘇六 娘) 15, 16, 43 – 45, 48, 112, 175 dative constructions (DC) 145, 153 – 155, 154; causative constructions vs. 153; choice of markers in 149 – 152, 149, 151, 152; dative markers 5, 144, 145, 148, 148, 153, 155; hoo7 予 in 5, 144 – 146, 145, 148, 153, 155, 228 deontic modals 2, 4, 49 – 50; vs. epistemic modals 49; see also ɗit55 得; tioh8 著 dependent-marking comparatives 10 – 11, 18 – 19, 33 see also pi2 比 diachronic perspectives 1 – 6, 3; interdialectal contact and dynamic change 21, 31, 150, 153 – 155; on negative modals 84 – 98, 98; see also grammaticalization; reanalysis dialects see specific language groups and dialects Diaspora Hokkien Southern Min 2 discourse markers 53, 197, 213n2 Discourses of the States (Guóyǔ 國語) 105, 120n10, 120n16 ɗit55 得 2 – 4; examples of 40 – 42, 43 – 44; negation of dynamic modals with 4, 56, 76 – 99, 90, 98; in (ti)leh (佇)得 217 – 220, 224, 226 – 227, 229, 232 ditransitive verbs 1, 5; complex ditransitive constructions (CDC) 149, 149, 153 – 155, 154; hoo7 予144 – 149, 145, 148, 155 Doctrina Christiana en letra y lengua china (The Christian Doctrine in the Alphabet and Language of China) 12, 16 – 18, 17, 33, 34n5, 34n7, 46, 48, 132, 140 – 141 Dream of the Red Chamber (Hóng Lóu Mèng 紅樓夢) 4, 108 – 109, 119, 121n23

236  Index Dūnhúang Biànwénjí 敦煌變文集 (Transformational Texts of Dunhuang) 4, 20, 105 – 107 Early Medieval Chinese 4, 10 Early Middle Chinese (EMC) 57, 61 – 64, 63, 67, 71 Early Southern Min (ESM) 3 – 4; comparatives in 10 – 11, 10, 17, 17 – 30, 47, 48; corpus for 12 – 17, 13, 14, 112; juxtaposition (paratactic device) in 114 – 115, 121n33; use of kang7 共 in 103, 103, 111 – 112, 115 – 119, 118, 119, 121n35; use of kap4/kah4 合 in 116, 121n30 Eastern Min (Mindong) 26 – 27 Elegant Brocades of Precious Beauty (Yù Yán Lì Jǐn 鈺妍麗錦) 14, 16, 43 – 45 Ewe language 50 existential verbs 6, 206 exponents 5, 144 Exquisite Brocades of a Hundred Flowers (Băi Huā Sài Jǐn 百花賽錦) 15, 16, 23, 43 – 45, 48 Feng Menglong 19 field work 5, 126, 144, 148 – 150, 154 finite clauses 6, 203, 220 – 223, 221 folk operas see Lì Jìng Jì 荔鏡記 (Romance of the Mirror and the Litchi); Lì Zhī Jì 荔枝記 (The Romance of the Litchi) frame semantics 125, 127 – 128, 139 – 140 Gan language subgroup 1 gēn 跟 2, 119 generalization 26, 59, 114, 132, 138 gòng 共 see kang7 共 grammar 1 – 3, 128; construction grammar 127 – 128, 139 – 140, 217 Grammar of the Chiõ Chiu Language (Arte de la lengua Chiõ Chiu) 16, 18, 27, 30, 34n6, 120n4 grammaticalization 6, 50 – 52; contactinduced 21, 31, 150, 153 – 155; of kang7 共 109 – 111; of kong2 講 6, 212 – 213; polygrammaticalization 102, 106, 109 – 110 Guangyun 73n4 Gǔběn Lǎo Qǐ Dà 古本老乞大 (Old Sinologist - oldest version) 108 guò 過 see ke3 過

Guóyǔ 國語 (Discourses of the States) 105, 120n10, 120n16 Hainan Min 2, 4, 56, 76 – 99, 90, 98 Hakka 1, 4; comparatives in 22, 26, 32 – 33; Meixian (梅縣) Hakka 32 – 33; negative modal form in 77 – 76, 87 – 92, 90, 94 – 98, 98; Sixian Hakka 87 – 88 hǎo 好-purposive clauses 223 hé 和 2, 103, 108, 110, 119, 121n 23, 121n34 he0 嘿 200, 213n4 head-marking comparatives 11, 19 – 24, 29 – 31, 33 Hokkien 2; Amoy 廈門 (Xiamen) dialect 12, 27, 144, 157 – 158, 166, 194n7 Hóng Lóu Mèng 紅樓夢 (Dream of the Red Chamber) 4, 108 – 109, 119, 121n23 honn0 呼 213n5 hoo7 予: as an agent marker 144 – 145, 145, 147 – 148, 148, 214n10, 214n12; in causative constructions 147, 145, 148, 154; as a causative verb 144, 145, 148, 148, 153; in dative constructions 145 – 146, 145, 148, 153 – 154, 154; ditransitive verbs and144 – 149, 145, 148, 155; in double object (ditransitive) constructions 145, 154; as an exponent 144; geographical distribution of 149 – 153, 149, 151, 152; in kheh4hoo7 乞予 5, 144, 145, 147 – 148, 148, 150, 155, 155n1; passive constructions with 5, 145, 147 – 153, 149, 151, 152, 155n1, 214n12, 228 – 230; possible origins from thoo7 度6n6; purposive constructions with 6, 228 – 230; in Taiwanese Southern Min (TSM) 144 – 155, 145, 148, 149, 151, 152, 154 Hòushān Rì Xiānzhào 後山日先照 (The Sun Shines First Behind the Mountain) 112, 115, 117 Hui language subgroup 1 i1 伊 1, 217, 230 – 332 imperatives: imperative mood of kang7 共 125, 130, 133 – 134, 134, 136 – 140, 138; imperative mood of leh4 咧 6, 169 – 176; with kong2 講 208; with tioh8 著 56 – 57, 57, 68 inalienability 125, 137, 140 indicative mood 125, 130, 133 – 134, 134, 140

Index  237 instrumental case 4 – 5; as an tioh8 著 instrumental preposition 49 – 50, 61, 65 – 66, 71 interrogative clauses 199 – 200, 203 – 204 Jiǎgǔwén 甲骨文 (oracle bone inscriptions) 103, 104, 120n11 jiào 叫 1, 73n5, 211 – 212 Jin Dynasty 22, 59 Jīn Huānǚ 金花女 (Lady Jin Huanü) 16, 112, 157, 176 – 177, 181 Jin language subgroup 1 Jīn Píng Méi Cíhuà (Jin Ping and Mei) 108 – 110 Jīnwén 金文 (Chinese Bronze Inscriptions) 120n11 jiok8 弱 13, 25, 34n12 Journey to the West (Xī Yóu Jì 西遊記) 4, 108, 121n21 juxtaposition (paratactic device) 114 – 115, 121n33 ka7 共 see kang7 共 kah4 教 1, 64, 139 kah4/kap4 合 2 – 4, 116; kang7 共 and 103, 111 – 113, 111, 116, 118 – 119, 121n30, 126 – 127, 126, 126 kàn/khuann3 看 14, 110, 203 – 204, 212, 229 kang7 共 2, 4 – 5; ablative functions of 111, 118; as an adverb 4 – 5, 102 – 104, 103, 106, 108; as an affective marker vs. a target marker 141n6; from animate to inanimate 110; in Archaic Chinese 103 – 105, 119; as a benefactive marker 5, 125 – 141, 126, 130, 132, 134 – 135, 138, 140; causative verbs and 134, 138, 139; as a coordinating conjunction 4 – 5, 102 – 121, 103, 111, 118, 119; evolution into a coordinating conjunction 102 – 121, 103, 111, 118, 119; grammaticalization of 109 – 111; imperatives with 125, 130, 133 – 134, 134, 136 – 140, 138; juxtaposition (paratactic device) 114 – 115, 121n33; kah4/kap4 合 and 103, 111 – 113, 111, 116, 118 – 119, 121n30, 126 – 127, 126, 126; ke3 過 and 214n8; in Medieval Chinese 105 – 107, 119; during the Ming Dynasty 103, 125, 139 – 140; in modern Chinese 108 – 109, 119; as a noun 104 – 105, 108; as an OR alternative conjunction 116 – 117; as patient markers

137, 141n1, 225 – 226, 229 – 232; in prepositional phrases 118; pronouns and 132, 214n7; during the Qing Dynasty 5; in the Quanzhou dialect 126, 126, 132; thematic tiers of 141n5 ke3 過 3, 20 – 22, 26 – 28, 214n8; see also Surpass comparative khah4 可: adverbial comparatives with 10, 12, 26 – 31, 42, 45 khah4 較 10, 11, 26 – 32 kheh4hoo7 乞予 5, 144, 145, 147 – 148, 148, 150, 155, 155n1 khi3 去: lai5 來 and 6, 113; negation with 93; purposives with 217, 224 – 226, 231 – 232 khi3hoo7 去予 147, 206, 214n10 khit4 乞 1, 5, 144 – 145, 148, 148, 150 khit4hoo7 乞與 214n10 khit4tng5 乞傳 148, 148, 150 khuann3kong2 看講 203 – 204, 209 kiong5 強 12, 13, 14 – 15, 19, 21, 24 – 29, 39, 41 – 45 kong2 講 6, 197 – 199; as an adverb 197; assertive and directive moods of 197; bridging context of 211 – 212; as a causative verb 211; as a complementizer 197, 203 – 207, 213n3, 214n8; grammaticalization of 6, 212 – 213; imperatives with 208; in khuann3kong2 看講 203 – 204; in kong2kio3 講叫 211 – 212; in lexicalized collocates 209 – 211, 214n13; parentheticals with 6, 198 – 199, 207 – 208, 211 – 213, 214n11; as a quotative marker 199 – 203, 212 – 213, 213n3; as a sentence-final particle 208 – 209, 212 – 213, 213n5; in siunn7kong2 想講 205; in thiann1kong2 聽講 204 – 205; in tsai1iann2kong2 知影 講 206; in tshiunn7kong2 像講 209 – 210, 214n13; as a verb of saying 6, 197 – 216 Lady Jin Huanü (Jīn Huānǚ 金花女) 16, 112, 157, 176 – 177, 181 lah0 啦 56, 213n4 lai5 來: khi3 去 and 6, 113; negation with 93; purposives with 217, 223 – 226, 230 – 232 Lǎo Qǐ Dà 老乞大 4, 121n20 Late Middle Chinese (LMC) period 57, 61, 66 leh4 咧: as a continuative aspect marker 5 – 6, 157 – 194, 162, 163, 165, 193; etymology of 158; imperative mood of

238  Index 6, 169 – 176; locative constructions and 159, 162, 165 – 166, 169 – 170, 173 – 174, 177 – 193; te3 處 and 167 – 174, 174 – 180, 187 – 194, 193; in ti7leh8 佇 咧 6, 166 – 167, 186, 210; ti7te3 在處 and 167 – 174, 180 – 187, 187 – 194, 193; as a weakened form of tit4 得 213n6; zhe 著 vs. 158 – 159, 194n1 lexicalized collocates 198 – 199, 209 – 210, 214n13, 214n15 Lì Jìng Jì 荔鏡記 (Romance of the Mirror and the Litchi): comparatives in 12, 13, 16 – 18, 39; continuative aspect markers in 157, 159, 160 – 163, 167 – 168; in the Early Southern Min corpus 12, 16 – 18, 112, 120n5; functions of pi2 比 47; negation of dynamic modals in 84 – 86, 97; see also kang7 共 Lì Zhī Jì 荔枝記 (The Romance of the Litchi): comparatives in 12, 13, 16 – 18, 39; continuative aspect in 157, 181; in the Early Southern Min corpus 12, 16 – 18, 112, 120n5; functions of pi2 比 47 Liǎngshì Tóngxiū 兩世同修 (Blessings of Two Generations) 112, 117 locative constructions: adverbs 178; continuative aspect marker leh4 咧 and 159, 162, 165 – 166, 169 – 170, 173 – 174, 177 – 193; direct objects 49; nouns 5 – 6, 183, 186, 190; prepositional phrases 4 – 5, 59 – 61, 71, 159, 169, 186, 190, 194n1 Lúnyǔ 論語 (Analects of Confucius) 103, 104, 120n13 m7 不 1, 11, 18, 81 Măn Tiān Chūn 滿天春 (All-Embracing Spring): comparatives in 14, 16, 19, 22, 25, 43; continuative aspect in 159 – 166; functions of pi2 比 in 48 Mançano, Melchor de see Arte de la lengua Chiõ Chiu (Grammar of the Chiõ Chiu Language) Mandarin; purposive constructions in 218 – 223, 221; Southern Min vs. 1 – 2, 9 – 10 Mande language Vai 51 markers: absence of markers in comparatives 24; adverbial marker tioh8 著 53 – 54, 53; affective markers 141n6; agent marker hoo7 予 1, 4 – 5, 144 – 145, 145, 147 – 148, 148, 214n10,

214n12; ah0 啊 213n2; causative verbs and benefactive marker kang7 共 and 134, 138, 139; choice of markers in dative constructions 149 – 152, 149, 151, 152; comitative 2; continuative aspect 2, 5 – 6, 167 – 168, 181 – 182; dative 5, 144, 145, 148, 148, 153, 155; differential object marker bǎ 把2, 118, 121n29, 227; discourse markers 53, 197, 213n2; experiential marker bat4/pat4 捌 214n8; local marker pheng22 并 32; negative 1; object 2; patient 137, 141n1; progressive 181, 217, 226 – 227; quotative marker kong2 講 199 – 203, 212 – 213, 213n3; see also benefactive markers; continuative aspect markers Meixian (梅縣) Hakka 32 – 33 Minbei (Northern Min) 26 – 27 Mindong (Eastern Min) 26 – 27 Ming Dynasty 5; comparatives in 14, 19, 23, 30, 33, 48; continuative aspect in 162, 165, 167 – 168, 180 – 182, 193; corpus from the 14, 16, 84 – 87, 157; negation in 84 – 87, 93; use of kang7 共 in 103, 125, 139 – 140; Wanli era 16, 20, 34n6, 120n5, 127, 157, 180 – 182, 193 Minimal Distance Principle (MDP) 218 Minzhong (Central Min) 26 – 27 modals: dynamic modality 76 – 77; epistemic modality 76; negation of dynamic modals with ɗit55 得 4, 56, 76 – 99, 90, 98; negative 84 – 98, 98; obligative 3 – 4; permissive 1, 56, 144, 146; preverbal 2; subjectivity in 49; see also deontic modals; tioh8 著 Modern Standard Chinese 3 mood: assertive 197, 213n4; indicative 125, 130, 133 – 134, 134, 140; particles and 188, 190 – 193; see also continuative aspect markers; imperative mood Moon Is Up (Yuèliàng Chūláile 月亮出來 了) 112, 118 morphemes: comparative 9, 21, 28 – 29, 111; multifunctional 144 – 148, 145, 148; negative 79 – 84; postpositional 51 Mòzǐ 墨子 103, 105 negation 1, 3 – 4; with bo22 無 4, 56, 79 – 84, 192; in Cantonese 76 – 77, 84, 92 – 98, 98; diachronic perspectives on negative modals 84 – 98, 98; of dynamic modals with ɗit55 得 4, 56, 76 – 99, 90,

Index  239 98; in Hakka 87 – 92, 90; idiosyncratic behavior of tioh8 著 in 56 – 57, 56, 57; with lai5 來 93; with m7 不 1, 11, 18, 81; in Southern Min 84 – 87 New Account of Tales of the World (Shì Shuō Xīn Yǔ 世說新語) 4, 102, 103, 105 – 106, 109, 120n17 Nogeoldae 老乞 大 63 – 64 nominative case 218 non-finite clauses 6, 203, 220 – 223, 221 Northern Min (Minbei) 26 – 27 nouns: in comparatives 3, 17, 21, 24, 28 – 29; dative 21; instrumental noun phrases 5; kang7 共 as a 104 – 105, 108; locative 5 – 6, 183, 186, 190; tioh8 著 and 58 – 61 obligative modals see tioh8 著 Old Sinologist - oldest version (Gǔběn Lǎo Qǐ Dà 古本老乞大) 108 operas see Lì Jìng Jì 荔鏡記 (Romance of the Mirror and the Litchi); Lì Zhī Jì 荔 枝記 (The Romance of the Litchi) oracle bone inscriptions (Jiǎgǔwén 甲骨 文) 103, 104, 120n11 paratactic device (juxtaposition) 114 – 115, 121n33 parentheticals: kong2 講 in 6, 198 – 199, 207 – 208, 211 – 213, 214n11 particles: deictic 6; durative 159; mood 188, 190 – 193; see also mood; sentencefinal particles parts of speech see adverbs; conjunctions; markers; nouns; particles; prepositional phrases; pronouns; verbs passive constructions 1, 5; with hoo7 予 5, 145, 147 – 150, 149, 151, 152, 152 – 153, 155n1, 214n12, 228 – 230; passive markers 148, 153 – 155 patient markers 137, 141n1, 225 – 226, 229 – 232 patterned self-correction 151 – 152, 154 pheng22 并 32 phonetics 166 – 167, 214n8 phonology pi2 比 47, 48; in bu5pi2 無比 18; in comparatives 3, 3, 10, 11, 17 – 18, 21 – 23, 28 – 30, 33, 39, 43; in hybridized comparatives 11, 11, 30 – 33 Pinghua language subgroup 1 play scripts 3, 53, 64, 84, 111 – 112, 162, 167, 179 – 184

Playscripts in Yuan Dynasty (Yuánqǔ 元 曲) 64 polygrammaticalization 102, 106, 109 – 110 possessive constructions 135 – 137, 140 postpositions 51, 165 postverbal complements 4, 60 – 62 pragmatics 128, 197 Pre-Modern Standard Chinese 23 – 24 prepositional phrases 3 – 6, 166; in comparatives 31; continuative aspect with 166, 186, 190; kang7 共 in 118; locative constructions 4 – 5, 59 – 61, 71, 159, 169, 186, 190, 194n1; as an tioh8 著 instrumental preposition 49 – 50, 61, 65 – 66, 71 principle of temporal sequence (PTS) 222 – 223 progressive markers 181, 217, 226 – 227 pronouns: continuative aspect markers and 166 – 167; distribution of third-person 1; i1 伊 in purposive constructions 1, 217, 230 – 232; interrogative 169 – 170; kang7 共 with 132, 214n7; in passive constructions 147; tioh8 著 and 58 – 59, 73n6 Proto Cahuilla-Cupeno, an Uto-Aztecan language, 51 purposive constructions: constraints on 6; distribution of pronoun and empty category 230 – 232; with hǎo 好 223; with hoo7 予 6, 228 – 230; with i1 伊 1, 217, 230 – 232; with khi3 去 217, 224 – 226, 231 – 232; with lai5 來 217, 223 – 226, 230 – 232; in Mandarin 218 – 223, 221; principle of temporal sequence (PTS) in 222 – 223; purposive clauses 6, 217 – 218, 223 – 232; syntax of 218, 222 – 223, 232; in Taiwanese Southern Min (TSM) 217 – 232, 221; with (ti)leh (佇)得 217 – 220, 224, 226 – 227, 229, 232; with ti7leh8 佇咧 6, 166 – 167, 186, 210 qiáng 強 12, 13, 14 – 15, 19, 21, 24 – 29, 39, 41 – 45 Qing Dynasty: benefactive marker kang7 共 in 5; comparative structures in 14, 18 – 19, 33; continuative aspect in 162, 178 – 184, 193; corpus of works in 157; negation in 84, 86 – 88, 94; pi2 比 in 48 Quanzhou dialect 1, 3, 5; comparatives in 16, 28, 31; continuative aspect in 157, 159, 163, 166 – 168, 181, 183 – 184;

240  Index fieldwork in 141n3, 149 – 150; kang7 共 in 126, 126, 132; see also hoo7 予 Quartets (Sìchóngzòu 四重奏) 111 – 112, 111, 115 – 116, 121n31 Quemoy (Jinmen) dialect: dative constructions 155; fieldwork in 149 – 150; see also hoo7 予 quotative markers 6, 197 – 198; kong2 講 as 199 – 203, 212 – 213, 213n3; quotative function 214n9 reanalysis 50 – 52, 65, 65, 68, 71, 71, 72n1 Romance of the Litchi, The see Lì Zhī Jì 荔 枝記 Romance of the Lute and the Classmates, The (Tóngchuāng Qínshūjì 同窗琴書 記) 15, 16, 43 – 45, 48, 112 Romance of the Mirror and the Litchi, see Lì Jìng Jì 荔鏡記 Romance of the Western Chamber (Xīxiāngjì) 64 – 65 rú 如 3, 10, 17, 19 – 23, 29 – 30 see also Surpass-Similarity comparative sài 賽 11, 13, 14 – 15, 19, 21, 23 – 25, 29 semantics: frame semantics 125, 127 – 128, 139 – 140; of imperatives 140; loss of (desemanticization) 72n1; pragmatics vs. 128 sentence-final particles 6 – 7, 197 – 199; kong2 講 as a 208 – 209, 212 – 213, 213n5 serial verb constructions (SVC) 3, 10, 21, 30, 138, 139 – 140, 146, 222, 228 shèng 勝 13, 14 – 15, 17, 19, 21 – 25, 29, 34n9, 40 – 41, 43 – 46 Shì Shuō Xīn Yǔ 世說新語 (A New Account of Tales of the World) 4, 102, 103, 105 – 106, 109, 120n17 Shījīng 詩經 (Book of Odes) 103, 104, 109 – 110, 120n9, 120n10, 120n12 sì 似: comparatives involving 3, 10, 11, 13, 14, 17, 17 – 23, 29 – 30, 40 – 41, 45 Sìchóngzòu 四重奏 (Quartets) 111 – 112, 111, 115 – 116, 121n31 Similarity comparative 11; see also Surpass-Similarity comparative Sinograms 157 siunn7kong2 想講 205 Sixian Hakka 87 – 88 Song Dynasty: comparatives in 19, 23 – 24, 30; coordinating conjunctions in 102; negation in 94 – 95 Southern Min 1 – 6; Hainan Min 2, 4, 56, 76 – 99, 90, 98; Mandarin vs. 1 – 2, 9 – 10;

phonetics 166 – 167, 214n8; Zhangzhou dialect 144, 157, 166 – 167; see also Cháozhōu dialect; Early Southern Min (ESM); Quanzhou dialect; Quemoy (Jinmen) dialect; Taiwanese Southern Min (TSM) stative causative verbs 144 Stories in Modern Taiwanese Southern Min, The (Táiwān Xiàndài Mǐnnányǔ Gùshìjí) 157, 192 Sū Liù Niáng 蘇六娘 (Dame Su Liu Niang) 15, 16, 43 – 45, 48, 112, 175 subjectification 52 – 53, 52 subjectivity 49, 171, 208, 213 subject-verb-object (SVO) word order in comparatives 29, 33 Sun Shines First Behind the Mountain, The (Hòushān Rì Xiānzhào 後山日先照) 112, 115, 117 Surpass comparative 9 – 12, 10, 13, 14, 17, 17 – 30, 32 – 33, 34n12, 35n15; examples of 39 – 41, 43 – 45, 46 Surpass-Similarity comparative 10, 11, 13, 14, 17, 17 – 23, 29 – 30; examples of 40, 41, 44, 45 syntax 1; in comparatives 26; purposive constructions 218, 222 – 223, 232; in other Asian languages 96 – 97; subjectverb-object (SVO) word order in comparatives 29, 33 tā 他 1, 24, 221 Táiwān Xiàndài Mǐnnányǔ Gùshìjí (The Stories in Modern Taiwanese Southern Min) 157, 192 Taiwanese Southern Min (TSM) 3 – 6; comparatives in 31, 33; corpus of 112; folktales 6; hoo7 予 and its counterparts in 144 – 155, 145, 148, 149, 151, 152, 154; origins of coordinating conjunction ka7 共 in 102 – 121, 103, 118, 119; purposive constructions in 217 – 232, 221; use of kah4/kap4 合 in 4, 116; see also kong2 講; tioh8 著 Tang Dynasty: comparatives in 18 – 21, 34n9; negation in 93 – 94; Southern Min during the 53, 168 tang7 重 13, 14 – 15, 22, 25, 27, 34n12, 40 – 45 target markers 141n6 te3 處 5 – 6; as continuative aspect marker 159 – 165, 165, 167 – 174; examples of 40 – 44, 147; expressions of time and 175 – 176, 178; leh4 咧 and 167 – 174,

Index  241 174 – 180, 187 – 194, 193; in ti7te3 在 處 5 – 6, 163 – 174, 180 – 185, 187, 192 – 194, 193; transformation of 175 – 184, 187 – 188, 192 – 193 temporal sequence 222 – 225, 229 – 230 Thai language 96 – 97 thiann1kong2 聽講 204 – 205 thoo7 度 1, 5, 6n6; continuative aspect and 177, 179; as an exponent 144; hoo7 予 and 144, 148, 148, 150 – 154, 151, 152, 154 ti7 在 6, 165; in ti7te3 在處 5 – 6, 163 – 174, 180 – 185, 187, 192 – 194, 193 (ti)leh (佇)得 217 – 220, 224, 226 – 227, 229, 232 ti7leh8 佇咧 6, 166 – 167, 186, 210 ti7te3 在處: leh4 咧 and 167 – 174, 180 – 187, 187 – 194, 193; ti7 在 and 5 – 6, 163 – 174, 180 – 185, 187, 192 – 194, 193 tioh8 著 3 – 6, 49 – 50, 53 – 57, 53, 56, 57; as an adverbial marker 53 – 54, 53; as a causative verb 49 – 52, 57, 63 – 72, 67, 69, 71, 73n5, 73n6; as a deontic modal 4, 67 – 71, 69, 70, 71, 71 – 72; diachronic development of 57 – 67, 63, 65, 67; as an instrumental preposition 49 – 50, 61, 65 – 66, 71; imperatives with 56 – 57, 57, 68; negation and 56 – 57, 56, 57, 73n3; nouns and 58 – 61; pronouns and 58 – 59, 73n6; zhuó 着 vs. 53, 57 – 67, 63, 73n5 tng5 傳 5, 144, 148, 148, 150 – 151, 154 Tóngchuāng Qínshūjì 同窗琴書記 (The Romance of the Lute and the Classmates) 15, 16, 43 – 45, 48, 112 tor0 除 159 – 163 Transformational Texts of Dunhuang (Dūnhúang Biànwénjí 敦煌變文集) 4, 20, 105 – 107 tsai1iann2kong2 知影 197 – 198, 203, 206 tsham1 參 2 tshiunn7kong2 像講 209 – 210, 214n13 tshut4 出 13, 25, 34n12, 41 tsiong1 將 2, 119 tsit8e7 一下 175 – 176 u7 有 185 – 188, 191 – 192, 206 uan3than3kong2 怨嘆講 202 – 203 velar coda 214n7 velar onset 1 verbs: of action 145, 167 – 169, 177, 191; deictic motion verbs 217, 224;

existential verbs 6, 206; postverbal complements 4, 60 – 62; preverbal modals 2; of saying 2, 6, 197 – 198; serial verb constructions (SVC) 3, 10, 21, 30, 138, 139 – 140, 146, 222, 228; types of functions 125; verbal complements 53, 53, 80, 93, 111; see also causative verbs; ditransitive verbs; kong2 講 vernacular novels 19 Vietnamese language 96 – 97 vowel lowering/raising 159, 166 Wanli era 16, 20, 34n6, 120n5, 127, 157, 180 – 182, 193 see also Ming Dynasty Wei, Jin, Southern, and Northern Dynasties (WJSN) 59 – 61, 63, 67, 71 word order see syntax Wu Shou-Li see Lì Jìng Jì 荔鏡記 (Romance of the Mirror and the Litchi); Lì Zhī Jì 荔枝記 (The Romance of the Litchi) Xī Yóu Jì 西遊記 (Journey to the West) 4, 108, 121n21 Xiamen (Amoy) dialect 12, 27, 144, 157 – 158, 166, 194n7 Xiang language subgroup 1 Xīxiāngjì (Romance of the Western Chamber) 64 – 65 Yáng Guăn Gē 楊管歌 (Yangguan Music) 15, 16, 25, 43 – 45, 48 Yu language subgroup 1 Yù Yán Lì Jǐn 鈺妍麗錦 (Elegant Brocades of Precious Beauty) 14, 16, 43 – 45 yú 於 20 – 21, 23 Yuan Dynasty: comparatives in 19, 22 – 23, 30; Gǔběn Lǎo Qǐ Dà 古本老乞大 108; Lǎo Qǐ Dà 老乞大 4, 121n20; negation in 94 – 95; Yuánqǔ 元曲 (Playscripts in Yuan Dynasty) 64 Yue 1, 22, 35n15 see also Cantonese Yuèliàng Chūláile 月亮出來了 (The Moon Is Up) 112, 118 Yǔyuàn texts 157, 180, 184 – 191 zero-marked comparatives 10, 12, 13, 14, 17, 17, 24 – 26, 28 – 29, 41, 45 Zhangzhou dialect 144, 157, 166 – 167 zhe 著 158 – 159, 194n1 see also leh4 咧 zhuó 着 53, 57 – 67, 63, 73n5 see also tioh8 著