The Evolution of Chinese Grammar 1108844057, 9781108844055

Citation preview

THE EVOLUTION OF CHINESE GRAMMAR The Chinese language has the longest well-documented history among all human languages, making it an invaluable resource for studying how languages develop and change through time. Based on a twenty-year research project, this pioneering book is the English version of an award-winning study originally published in Chinese. It provides an evolutionary perspective on the history of Chinese grammar, tracing its development from its thirteenthcentury BC origins to the present day. It investigates all the major changes in the history of the language within contemporary linguistic frameworks, and illustrates these with a wide range of examples taken from every stage in the language’s development, showing how the author’s findings are relevant to contemporary descriptive, theoretical, and historical linguistics. Shedding light on the essential properties of Chinese and, ultimately, language in general, it is essential reading for academic researchers and students of Asian linguistics, historical linguistics, and syntactic theory. yuzhi shi is Associate Professor in the Department of Chinese Studies at the National University of Singapore. He obtained an MA from the University of California, San Diego, in 1995 and a Ph.D. from Stanford University in 1999. His major publications include Motivations and Mechanisms of Grammaticalization in Chinese (Peking University Press, 2006), Chinese Grammar (The Commercial Press in Peking, 2010), and The Historical Morpho-syntax of Chinese, which won the Prize of China Excellent Publications in 2016.

Published online by Cambridge University Press

Published online by Cambridge University Press

THE EVOLUTION OF CHINESE GRAMMAR

YUZHI SHI National University of Singapore

Published online by Cambridge University Press

Shaftesbury Road, Cambridge CB2 8EA, United Kingdom One Liberty Plaza, 20th Floor, New York, NY 10006, USA 477 Williamstown Road, Port Melbourne, VIC 3207, Australia 314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India 103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467 Cambridge University Press is part of Cambridge University Press & Assessment, a department of the University of Cambridge. We share the University’s mission to contribute to society through the pursuit of education, learning and research at the highest international levels of excellence. www.cambridge.org Information on this title: www.cambridge.org/9781108844055 DOI: 10.1017/9781108921831 © Yuzhi Shi 2023 This publication is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press & Assessment. First published 2023 A catalogue record for this publication is available from the British Library. Library of Congress Cataloging-in-Publication Data Names: Shi, Yuzhi, 1963– author. Title: The evolution of Chinese grammar / Yuzhi Shi, National University of Singapore. Other titles: Han yu yu fa yan hua shi. English Description: Cambridge ; New York, NY : Cambridge University Press, 2023. | Includes bibliographical references and index. Identifiers: LCCN 2022039194 | ISBN 9781108844055 (hardback) | ISBN 9781108926119 (paperback) | ISBN 9781108921831 (ebook) Subjects: LCSH: Chinese language – Grammar. | Chinese language – Grammar – History. | Grammar, Comparative and general – Verb phrase. | Grammar, Comparative and general – Syntax – Research – History. | Chinese classics – History and criticism. Classification: LCC PL1103 .S57813 2023 | DDC 495.15–dc23/eng/20220923 LC record available at https://lccn.loc.gov/2022039194 ISBN 978-1-108-84405-5 Hardback Cambridge University Press & Assessment has no responsibility for the persistence or accuracy of URLs for external or third-party internet websites referred to in this publication and does not guarantee that any content on such websites is, or will remain, accurate or appropriate.

Published online by Cambridge University Press

To my Mother, Father, and Crystal, with Love

Published online by Cambridge University Press

The research was supported by the National Science Foundation of the USA (SBR–9818629, 1996–2005). The writing of this book was also supported by the Chinese Fund for the Humanities and Social Sciences (19WYYB008, 2019–2022) and the Wan Boo Sow Research Grant (C–102–000–026–001).

Published online by Cambridge University Press

CONTENTS

List of Figures page xii List of Tables xiii Acknowledgments xvi Conventions Used in the Examples xviii List of Abbreviations and Symbols xix 1

Some Preliminaries 1 1.1 1.2 1.3 1.4 1.5

2

Copular Word and Construction 10 2.1 2.2 2.3 2.4 2.5

3

Introduction 10 The Copular Particle Yě 11 The Focus Marker Yě 14 The Copula Shì 18 Copula and Wh- Movement 24

Focus and Wh- Words 27 3.1 3.2 3.3 3.4 3.5 3.6

4

Readership 1 A Natural Laboratory 2 Periodization of the Chinese Language and the Dating of Texts 5 Historical Texts 7 The Structure of the Contents 8

Introduction 27 The Licensing of Wh- In Situ 29 Focus Constructions and Wh- Questions 35 Topicalization and the Landing Site of Wh- Words 40 The Emergence of the Copula Shì 46 Wh- Words in Preposition Phrases 57

Serial Verb Construction 62 4.1 Introduction 62 4.2 Conjunctions of Verbs and Clauses 62 vii

Published online by Cambridge University Press

Contents

viii

4.3 4.4 4.5 4.6 5

Disyllabification 78 5.1 5.2 5.3 5.4 5.5 5.6 5.7

6

Introduction 106 The Syntax of the Resultative Construction 108 The Origin of the Resultative Construction 116 The Separable Resultative Structure 120 Frequency and Idiomatization 124 Context for Fusion 135 Extension 139

Information Structure 142 7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9

8

Introduction 78 Disyllabic Words 80 Motivations for Disyllabification 81 The Development of Disyllabic Words 84 Effect on the Resultative Construction 86 Fusion of Verb and Resultative 89 Constraint of the Number of Syllables 98

Resultative Construction 106 6.1 6.2 6.3 6.4 6.5 6.6 6.7

7

The Decline of the Co-ordinate Connective Ér 66 Fusion across Clauses 70 The Resultative Chéng 71 The Resultative Bù-Dé 74

Introduction 142 Information Structure 143 The Change in Topicalization 144 The Change in Focusing 147 The Focus Construction 150 Changes in Assigning Definiteness 151 Non-question Functions of Wh- Words 157 The Existential Construction 160 The Principle of Action–Resultative Ordering 164

The Passive Construction 175 8.1 8.2 8.3 8.4 8.5 8.6

Introduction 175 Optionality of the Agent in Passives 177 The Semantics and Syntax of Passives 178 The Agent Noun in the Passive Structure 179 From Optional Agent to Obligatory Agent 183 Passive with an Obligatory Agent 193

Published online by Cambridge University Press

Contents 8.7 Lexical Sources for Passive Morphemes 196 8.8 The Development of Passive Structures 200 9

The Disposal Construction 203 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9

10

Introduction 203 Hypotheses about Its Origins 205 Functions of the Disposal Construction 212 Two Historical Motivations 227 Other Related Constructions 234 Disposal Markers: Ná, Guǎn, and Gěi 237 Language Contact 239 Effect on Other Structures 241 Competition between Jiāng and Bǎ 243

Verb Copying and Reduplication 246 10.1 The Verb-Copying Construction 246 10.2 Verb Reduplication 265

11

The Comparative Construction 273 11.1 11.2 11.3 11.4 11.5 11.6

12

The Ditransitive Construction 296 12.1 12.2 12.3 12.4 12.5

13

Introduction 273 Comparatives in Old Chinese 274 Comparatives in Modern Chinese 279 The Comparative Marker Bǐ 282 Comparatives in Contemporary Chinese 287 Suffixes of Adjectives and Adverbs 291

Introduction 296 Relation to the Instrumental Construction 297 Relation to the Disposal Construction 298 Assimilation by the Resultative Construction 300 Bidirectional Transfer 314

Aspect and Tense 326 13.1 13.2 13.3 13.4 13.5

Introduction 326 Conditions for the Emergence of Aspect Suffixes 330 The Emergence of Aspect Suffixes 334 Inchoative and Continuous Aspects 350 Interaction between Aspect Suffixes 352

Published online by Cambridge University Press

ix

Contents

x

13.6 Diverse Forms across Dialects 356 13.7 Future Tenses 360 13.8 Auxiliary Verbs of Aspect 364 14

Negation 368 14.1 14.2 14.3 14.4 14.5

15

The Boundedness of the Predicate 386 15.1 15.2 15.3 15.4

16

Introduction 411 The Demonstrative Puzzle 413 Changes in Demonstratives 420 From Classifier to Demonstrative 427 The “Semantic Primitive” Hypothesis 437

Distal Demonstratives from Phonological Derivation 439 18.1 18.2 18.3 18.4 18.5

19

Introduction 396 The Emergence of the Classifier System 397 Motivations for Noun Classifiers 400 The Emergence of Verb Classifiers 403 Reduplication of Classifiers 406

Demonstratives from Classifiers 411 17.1 17.2 17.3 17.4 17.5

18

Introduction 386 The Boundedness of Verbal Predicates 387 The Boundedness of the Adjectival Predicate 390 Effects of the Boundedness Tendency 392

Classifiers 396 16.1 16.2 16.3 16.4 16.5

17

Introduction 368 Affirmation and Negation of the Perfective Aspect 368 The Establishment of the System for Negative Markers 371 Characteristics of the Old Negative Systems 373 New Developments of Perfective Aspect Marking 383

Introduction 439 Dialect Variants of Demonstratives 442 Correlations between Proximal and Distal Demonstratives 446 Derivation from Proximal to Distal Demonstratives 448 Iconicity between Acoustic Property and Distance 466

Pronouns, Plurals, and Diminutives 470 19.1 Introduction 470 19.2 Pronominal Systems in Old Chinese 471

Published online by Cambridge University Press

Contents 19.3 19.4 19.5 19.6 19.7 20

Structural Particles 488 20.1 20.2 20.3 20.4 20.5 20.6

21

Introduction 488 Structural Particles in History 489 Classifiers and the Structural Particle De 493 Functions of the Structural Particle De 495 Demonstratives and Structural Particles 498 Grammaticalization of the Structural Particle De 506

Word Order and Relative Clauses 511 21.1 21.2 21.3 21.4 21.5

22

The Emergence of Plural Morphemes 475 Classifiers and Plurals 478 Word Formation for Uncountable Nouns 480 Further Development of the Classifier Gè 481 The Emergence of Diminutive Inflection 483

Introduction 511 Changes in the Constituent Order 514 The Relative Clause and the Head Noun in Old Chinese 520 Typological Change in the Relative Clause Structure 529 Development toward Consistency 538

Conclusions 544 22.1 22.2 22.3 22.4 22.5

The Momentum of Language Change 544 Changes in Cluster 546 Patterns of Grammatical Change 549 Drivers of Language Change 551 Construction Schema 552

References 558 Primary Sources of Texts 575 Index 578

Published online by Cambridge University Press

xi

FIGURES

18.1 18.2

The structure of Chinese syllables page 451 Sonority degrees of proximal and distal demonstratives 468

xii

Published online by Cambridge University Press

TABLES

1.1 3.1 4.1 5.1 5.2 5.3 5.4 5.5 5.6 6.1 6.2 8.1 8.2 8.3 8.4 8.5 8.6 8.7 12.1 13.1 13.2 13.3 13.4 13.5 14.1 14.2

The periodization of the Chinese language page 6 Phonetic forms of wh- words in Old and Contemporary Chinese 54 The decline of the connective ér over time 67 Disyllabic words in Contemporary Chinese 80 Monosyllabic and disyllabic words in Cantonese and Standard Mandarin 83 Disyllabic verbs from the fifth century AD to the twelfth 84 Separable resultative constructions in dialects 88 The distribution of V-sǐ + O in Shui Hu Zhuan 100 Fronting of adverbs in Dun Huang Bian Wen 102 The ratio of adjacent to separate structures in Zhu Zi Yu Lei 137 The ratio of V-liǎo to V-X-liǎo in Zu Tang Ji 138 The wéi passive and the wéi . . . suǒ passive 185 The increase of the bèi passive with agent phrases over time 188 Frequency of the four passive morphemes in Pekingese 190 The most common passive morphemes in dialects 192 The proportion of passives with/without agent phrases 193 Construal and diverse passive markers across languages 197 Types of passive in the history of Chinese 197 “Transfer” verbs and ditransitive constructions in English and Chinese 324 The phonological reductions of aspect suffixes 330 The resultative liǎo from the sixth century AD to the tenth 334 The resultative liǎo adjacent to the verb in Dun Huang Bian Wen 336 The resultative liǎo of monosyllabic verbs in Dun Huang Bian Wen 337 The functions of aspect markers and their lexical sources 354 The possession verbs of Old and Contemporary Chinese 371 The phonological forms of negatives in Old Chinese 373

xiii

Published online by Cambridge University Press

List of Tables

xiv 17.1 17.2 17.3 17.4 17.5 17.6 17.7 18.1 18.2 18.3 18.4 18.5 18.6 18.7 18.8 18.9 18.10 18.11 18.12 18.13 18.14 18.15 18.16 18.17 18.18 18.19 18.20 18.21 18.22 18.23 18.24 18.25 19.1 19.2 19.3 19.4 20.1

Frequencies of demonstrative pairs from the seventh century AD to the tenth 415 Distribution of the demonstratives zhè and gè in dialects 416 Correlations between classifiers and demonstratives 418 The dialect distributions of the general classifiers gè and zhī 428 The rise and decline of the demonstratives zhè and cǐ 431 Differences between the demonstratives cǐ and zhè 431 The functional features of classifiers and demonstratives 437 Phonological correlations between proximal and distal demonstratives 442 Changes and dialectal variations of demonstratives 446 The dialect distribution of demonstratives 446 Demonstratives in the Southern Min dialect 447 Demonstratives in standard Mandarin 448 Demonstratives in the Hanshou and Lixian dialects 448 The proximal zhè and the distal nà in vernacular texts 449 Distal demonstratives from the classifier gè in subdialects 450 The historical phonological forms of distal demonstratives 455 The onsets of distal demonstratives related to the proximal demonstrative [tʂɤ] 456 Patterns of the two onsets with sonority increase 457 Patterns of the two onsets with equal sonority 458 Patterns of the two nuclei with sonority increase 458 Patterns of the two nuclei with equal sonority 458 Demonstratives in the Jin dialect with different onsets 459 The classifier gè realized as a distal demonstrative in the Yue dialect 460 Reverse derivation between proximal and distal demonstratives in Cantonese and Hakka 460 Reverse derivation in the Xiang dialect 461 The [h] and [x] onsets of distal demonstratives 462 Phonological variations of demonstratives in subdialects 463 Different nuclei distinguishing proximal from distal demonstratives 463 Cliticization of demonstratives 464 Use of [u] as a distal demonstrative in the Jin dialect family 466 Nasals and vowels encoding demonstratives in the Xiang family 466 Three exceptions to the phonological derivational rule 466 Singular and plural pronouns in Contemporary Chinese 471 Accurate and fuzzy singular/plural of demonstratives 479 Diminutive of reduplicated form in dialects 486 Diminutive of high-pitched tone in dialects 486 Functions of four structural particles in history 489

Published online by Cambridge University Press

List of Tables 20.2 21.1 21.2 21.3

The rise and decline of the old and new numeral phrases 504 Word order correlations with VO and OV 515 Word order changes in Old and Modern Chinese 535 The recurrent correlations of VO and OV languages 540

Published online by Cambridge University Press

xv

ACKNOWLEDGMENTS

I am grateful to several mentors for their education, advice, inspiration, encouragement, and friendship. Among various linguistic frameworks, the most useful include cognitive linguistics, functionalist linguistics, historical morphology and syntax, and particularly grammaticalization theory. During my graduate studies, I took courses and independent studies in grammaticalization theory with Elizabeth Traugott at Stanford University, in historical morphology and syntax with Paul Kiparsky at Stanford University, in cognitive linguistics with Ronald Langacker and Gilles Fauconnier at the University of California, San Diego, and in functionalist linguistics with Sandra Thompson at the University of California, Santa Barbara. I owe a huge debt to my mentor and friend Charles N. Li, who introduced me to the field of Chinese historical linguistics. As a research fellow I joined the Historical Morpho-syntax of Chinese project, whose principal investigator was Charles N. Li, which was supported by the National Science Foundation of America, at the University of California, Santa Barbara, in 1996. This was the first time that I started to do research in Chinese historical linguistics. It was extremely fruitful, as we coauthored dozens of papers and one monograph that was published by Peking University Press in 2001. I returned to Stanford University as a visiting scholar, with Paul Kiparsky as my mentor, in 2010–2011. In this academic year, I took many courses in linguistics, such as Laboratory Syntax, taught by Joan Bresnan; Linguistic Typology, taught by Paul Kiparsky; and Diachronic Construction Grammar and Constructionalization, taught by Elizabeth Closs Traugott, which were extremely helpful in studying the evolution of Chinese grammar. Additionally, I benefited greatly from personal communications with these professors. Special thanks are due to Tania Kuteva. Her books are always inspiring for my research. More importantly, Tania played the key role in the publication of this manuscript by the prestigious Cambridge University Press. When she was invited to deliver a lecture at National University of Singapore in 2018, Tania recommended my manuscript to the press and offered critical help during every stage of the publication process. Moreover, Tania graciously provided excellent comments and suggestions on an earlier xvi

Published online by Cambridge University Press

Acknowledgments

xvii

version of this manuscript, which were extremely helpful in improving the quality of this book. I am grateful to Professor Tan Eng Chye, the president of National University of Singapore, for giving me generous aid when my family suffered a tragedy, which helped me concentrate on my research and teaching. I am indebted to National University of Singapore for affording me a wonderful working environment. In the past two decades, I have been invited to give talks about historical Chinese linguistics by numerous universities and conferences in mainland China, Japan, South Korea, Canada, and the USA, where I benefited greatly from the audiences. Over the years, I have collaborated with many colleagues to carry out research. I would like to take this chance to express my appreciation for them, including Charles N. Li, Audrey Li, Li Chongxing, and Wang Tongshang. I have been fortunate beyond words to have enjoyed the love, support, wisdom, and laughter of my daughter, Crystal. Finally, my deepest gratitude goes to my parents and sisters, Shi Li (my elder sister), Yupu (my younger sister), Yuluo (my younger sister), and Yuhong (my youngest sister), who managed to take pride and pleasure in my work even though it has meant time away from them. They have all made my dream come true.

Published online by Cambridge University Press

CONVENTIONS USED IN THE EXAMPLES

1 2 3 4

[ ]: a constituent is absent. *: an utterance is structurally unacceptable. The parts that are relevant to the current discussion are in bold or underlined. Glosses and translations: each Chinese example is first given in traditional Chinese characters and then in Pinyin. In the line immediately below the Chinese example, we gloss each Chinese word with the clearest and most literal English equivalent possible. In the following line we offer a translation of the whole utterance into idiomatic English, attempting to preserve the “flavor” of the Chinese utterance as much as possible. 5 Pinyin: the transcription system for examples of Classical Chinese, present-day standard Mandarin and most northern dialects (despite some phonological variations) is pīnyīn (lit. “spell sound”), the official romanization system of the People’s Republic of China, which is also the most widely used system in the media and scholarly writings on Chinese in the West. However, examples of other dialect families are given in IPA. 6 When tone values are relevant to the current discussion, especially for examples of classical Chinese of other dialect families, two or three numbers at the top right corner indicate the tone values of the syllables. 7 The approximate dating of ancient examples is provided in parentheses on the bottom line.

xviii

Published online by Cambridge University Press

ABBREVIATIONS AND SYMBOLS

ACCU Adj Adv Advl ANAP ASP ASSOC AUX BA BEI BEN CAUS CL CON COND CONJ COP CP DAT DEC DEICT DEM DET DIM DIS DUR EMPH EXP FOC

accusative adjective adverb adverbial anaphor aspect associative particle (-de) auxiliary verb disposal marker bǎ passive marker bèi benefactive causative classifier continuous conditional conjunction copula complement dative declarative deictic marker demonstrative determinator diminutive disposal marker durative aspect (-zhe, zài) emphatic experiential aspect (-guo) focus xix

Published online by Cambridge University Press

xx FUT GEN IMP MOD NEG NOM NOMI N NP Num Obj Od Oi OV PART PASS PERF PL PN POSS PP PREP Pro PROG PRT QUE REL RES SG SOV Subj SVO TNS TOP V Vintr VO VP VR

List of Abbreviations and Symbols future genitive (-de) imperative modal negative nominalizer (de) nominative noun noun phrase number object direct object indirect object object–verb word order particle passive marker perfective aspect plural (-men, -xie) personal name possessive preposition phrase preposition pronoun progressive aspect particle question (ma) relative clause resultative singular subject–object–verb word order subject subject–verb–object word order tense topic verb intransitive element verb–object word order verb phrase verb–resultative construction

Published online by Cambridge University Press

List of Abbreviations and Symbols Vtr wh1st 2nd 3rd

transitive element interrogative pronoun first-person singular pronoun second-person singular pronoun third-person singular pronoun

Published online by Cambridge University Press

xxi

Published online by Cambridge University Press

1 Some Preliminaries

1.1

Readership

This book is written for the English-speaking audience in the world who cannot read publications in Chinese, especially those who cannot read texts in Classical Chinese. We have selected the essentials of our own studies from the past twenty years or so to provide a comprehensive picture of the evolution of Chinese grammar over the past three millennia. Thus we have made technical adjustments to increase the readability of the contents, including the following four features. First, traditional Chinese characters and the corresponding Pinyin transcriptions are provided only in the numbered examples; no Chinese characters are provided for Chinese expressions in the narrative parts. Second, for the historical examples and the examples in the northern dialects, only Pinyin forms representing the phonological system of standard Mandarin are provided, though there are apparently phonological variations across dialects. Only when the phonetic forms of the examples are relevant to the analysis are International Phonetic Alphabet (IPA) representations provided. Examples from other major dialect families are given in the IPA form according to the phonological systems of the representative subdialects. Third, only the general titles of the ancient texts from which examples are cited are indicated; the subtitles of these texts are omitted because they are meaningless to Western readers. Fourth, we do not adopt the common practice of dating in Chinese historical linguistics by using the dynasty, such as the Han Dynasty and the Tang Dynasty, as the temporal frame to describe the historical development. Instead, the Western calendar is adopted to date examples and the development of the language. All analyses are made within contemporary linguistic frameworks, mainly grammaticalization theory, topological linguistics, cognitive linguistics, construction grammar, functionalist linguistics, and general linguistics. Our work is neither a simple description of the changes nor an exhaustive list of all changes in the history of Chinese. This book is not written encyclopedically; only parts that are theoretically significant will be discussed in detail. However, all major changes in the history of Chinese are covered in this book.

1

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

Some Preliminaries

2 1.2

A Natural Laboratory

The Chinese language has a well-documented history of approximately 3,300 years that is actually invaluable for studying the languages and civilization of all human beings. Although there are roughly 6,700 different languages in the world, Chinese possesses not only the longest history but also the greatest amount of documentation that accurately and completely records detailed developments at each stage. In this regard, only Greek may be comparable with Chinese. To understand the scientific value of historical documents of the Chinese language, let us consider some parallel examples in the natural sciences. There are two types of law in nature: one type, such as the law of free fall, needs to be proven in a human-designed laboratory; the other type, such as Einstein’s relativity theory, can be tested only in a natural laboratory, namely the natural world. First, we consider how to identify and prove the first type of natural law. Aristotle (384‒322 BC), an ancient Greek philosopher, discussed the law of falling objects in Physics (Book VII), and from then to the sixteenth century AD it was generally assumed that the speed of a falling body was proportional to its weight. That is, a ten-kilogram object was expected to fall ten times faster than an otherwise identical one-kilogram object through the same medium. This “intuitively correct” law was found to be wrong by physicists in the sixteenth century AD, when Galileo (1589–1592) is said to have first discovered the law by dropping two objects of unequal mass from the Leaning Tower of Pisa. However, the law of free fall can be confirmed as a scientific law that is entirely accurate only in human-created ideal conditions of a uniform gravitational field without air resistance. Likewise, identifying the structure of atoms and capturing ions require sophisticated equipment such as electronic microscopes and accelerators that are designed by scientists. Next, we would like to provide an example of a natural law that can be tested only in nature. On the basis of his new theory of general relativity, which was proposed in 1911, Einstein calculated that light from another star should be bent by the sun’s gravity. In 1919, that prediction was confirmed by Sir Arthur Eddington during the solar eclipse on May 29. Additionally, the theory of general relativity proposed by Einstein approximately 100 years ago posited that gravitational waves exist in the universe. To prove this hypothesis, the American government spent a huge amount of money building two laboratories in northern and southern areas. In 2016, scientists in the two laboratories announced that they had made the first direct observation of gravitational waves originating from the merging of a binary black hole system. Likewise, linguistic rules also fall into two types: one type is rules that can be proven by psychological experiments or statistical surveys of contemporary language; the other type is rules that can be discovered and tested only in the historical development of

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

A Natural Laboratory

3

a language. For example, how construal influences grammatical forms may be identified by a psychological experiment, but how construal affects the recruitment of lexical sources for a grammatical category, such as the passive, requires a comprehensive investigation of several millennia of historical data. We can benefit greatly from the history of a language if we view it as a natural laboratory for testing the validity of any theoretical hypotheses. According to our own research experiences, the evolution of a language is especially invaluable for exploring the following linguistic issues. (a)

(b)

(c)

(d)

Exploring the relationship between cognition and language. One of the unresolved yet most challenging questions is why all human languages must constantly change over time and no language can remain the same forever. Answering this question involves interdisciplinary collaboration among linguistics, psychology, neurology, and cognitive science, but any hypothesis can be tested only in the evolution of languages. Testing theoretical hypotheses of different linguistic frameworks. For example, the usage-based model assumes that repetition and frequency are mainly responsible for the innovation of grammatical forms (Bybee 2006: 269‒278). Similarly, cognitive linguistics claims that grammatical structures develop out of the entrenchment and enforcement of certain linguistic expressions (Langacker 1987: 59‒60). Additionally, grammaticalization theory suggests that pragmatic inference is a major factor in triggering grammatical change (Hopper and Traugott 2003: 71‒98). In contrast, generative linguistics believes that language consists of principles and parameters and that the former are immune to change but the latter may change due to discrepancies in language acquisition (Lightfoot 2013). In our view, however, the validity of these hypotheses must ultimately be tested using diachronic evidence. Discovering universal correlations across languages. For example, Greenberg (1966a) identified a universal correlation: if a language takes SVO as its basic word order, its relative clause follows the head noun. After more than 1,000 languages were investigated, the results indicated that Chinese was the sole counterexample to this universal correlation (Dryer 1992, 2007). However, if we examine the diachronic data of Chinese, it is clear that Chinese was consistent with the correlation. In addition, the motivation for causing the change from consistency to inconsistency reveals an essential property of the interaction among different constructions. Proving concrete argumentation about the relation between different grammatical categories. For example, it has been generally accepted that, historically, the subject grammaticalized out of the topic. Logically, this

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

4

Some Preliminaries hypothesis entails that English grammar be more developed than Chinese grammar because the former is subject-prominent and the latter topicprominent (Li and Thompson 1976, Hopper and Traugott 2003: 28). A simple comparison between diachronic and synchronic data clarifies whether this view can be true. (e) Revealing the relationship between different constructions. For example, it is assumed that the passive structure is transformed from an underlying active form, a key point that Chomsky (1957) established in his generative linguistics. Since then, the correlation between the passive and the active has occupied the central position in creating other theoretical frameworks, such as cognitive linguistics, lexical functional grammar, and relation grammar. In our view, the real relationship between the passive and the active can be seen only from diachronic change in the grammar.

The earliest written records of Chinese were the oracle bone inscriptions, which can be dated back to the thirteenth century BC. Since then, Chinese has been well documented in numerous genres, such as poetry, dialogues, novels, books of history, philosophy, politics, and many other subjects that objectively reflect the situation of the spoken language at each stage. Although Chinese is not the earliest language in the world to have been recorded by a writing system, the other earlier languages all ceased to be used at some point in history. For example, the oldest known written language is Sumerian, which dates back to at least the thirty-fifth century BC. The earliest proof that the written Sumerian language existed was the Kish Tablet, which was found in Iraq. Sumerian is older than Egyptian, but it lasted as a spoken language only until the twentieth century BC, when it was replaced by another language, called Akkadian. The longer the history of a language is, the better chance we have of identifying the regularity of the development of the grammar. A grammaticalization process often takes several hundred years, and may take over a millennium. An extreme case is the development of reduplication, which took nearly 3,000 years to reach the completion point (for details, see Section 10.2). Specifically, reduplicated forms of adjectives and adverbs were already commonly used in texts composed around the eleventh century BC, but the reduplicated forms of nouns first appeared in texts around the third century AD, those of classifiers emerged around the eighth century AD, and not until the fourteenth century AD did verb reduplications start to occur. Why the reduplicated forms of different word classes emerged at such different periods is the key factor in understanding the grammatical properties of these forms and the mechanism of language changes.

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

Periodization of the Chinese Language and the Dating of Texts

5

As we know, the writing system of Chinese is not phonetic in nature. It is also by and large not pictographs, which account for merely a small portion of Chinese characters; approximately 80 percent are pictophonetic, with one part of a character indicating the semantic category and the other part roughly identifying the sound. As the phonological system developed, the writing system remained highly stable. Thus this stability enables us to investigate the whole history of the language without much difficulty. The original texts of poems and prose works have been adopted in textbooks of Chinese language teaching even in primary schools, and the greater proportion of Chinese textbooks are those of Classical Chinese. People who have obtained a BA in Chinese language and literature are generally able to understand texts composed in ancient times. This reading ability is crucial for an individual researcher to discover the regularity of the evolution of the Chinese language. Therefore numerous scholars in the circle of Chinese linguistics have studied Classical Chinese, and a tremendous number of books and papers have been published in this area. These collective efforts are very helpful in exploring the development of the language. By comparison, the earliest English texts were composed around the seventh century AD, and the writing system was basically phonetic, although it has changed over time. Only limited scholars who have undergone special training can understand the ancient texts, which still does not guarantee that they will understand texts composed at a different stage. As Barber (1995: 100) pointed out, people today are almost entirely unable to read texts in English that were composed around the twelfth century AD. In contrast, most people who graduate from high school in China find it comfortable and enjoyable to read the poetry, prose, and other literary genres composed after the tenth century AD. Another advantage of investigating the history of Chinese is that the language system has never been fundamentally influenced by other languages due to low language contact, and its vocabulary, phonology, and grammar have been developing on their own path, without disruption by any foreign language contact. In contrast, almost every language of the Indo-European family, such as English, French, German, and Italian, has undergone fundamental changes due to language contact. Thus their development paths were often disrupted and altered by other languages, which interfered with and blurred the regularity of their development. In short, the long, undisrupted, and well-documented history of the Chinese language makes it the best natural laboratory to experiment with scientific problems in relation to human language.

1.3

Periodization of the Chinese Language and the Dating of Texts

There is no generally accepted periodization of the history of the Chinese language in the literature, and various labels are used, such as Archaic Chinese, Old Chinese,

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

6

Some Preliminaries

Classical Chinese, Medieval Chinese, and Modern Chinese, which often refer to quite different time spans. There are several periodizations of the Chinese language, which are based on different criteria, such as syntax, phonology, or lexicon (Wang 1980, Lü 1985, Peyraube 2017). Since this book is targeted at readers in the English-speaking world, we adopt the terminology used for describing the stages of the development of English that is widely accepted in the English literature (e.g. Hopper and Traugott 2003: xx, Kytö and Päivi 2016: 1‒18).1 The primary purpose of dividing the stages of the history of a language is to facilitate the description of its development. According to our research experiences, therefore, the following periodization is most convenient for describing the historical development of the Chinese language. The periodization is further refined by adding either an “early” or a “late” modifier. When these modifiers are added, the term refers to the first or last three centuries of the period; for example, “Early Medieval Chinese” means a period from the second century BC to the first century AD. When periodizing the history of the Chinese language, we have taken into consideration the periodizations of Wang (1980: 35) and Lü (1985: 1), both of which are highly influential in the field. Table 1.1 The periodization of the Chinese language Old Chinese 1300 BC–200 BC Early Old Chinese 1300 BC–800 BC Middle Old Chinese 800 BC–400 BC Late Old Chinese 400 BC–200 BC Medieval Chinese 200 BC–AD 1000 Early Medieval Chinese 200 BC–AD 300 Middle Medieval Chinese AD 300–AD 700 Late Medieval Chinese AD 700–AD 1000 Modern Chinese AD 1000–1900 Early Modern Chinese AD 1000–1300 Middle Modern Chinese AD 1300–1700 Late Modern Chinese AD 1700–1900 Contemporary Chinese AD 1900–present

1

The stages of English are listed as follows: Old English Middle English Early Modern English Modern English Contemporary English

AD 600–1125 AD 1125–1500 AD 1500–1750 AD 1750–1950 AD 1950–present

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

Historical Texts

7

The above four stages are taken into account the timing of the establishments of three major dynasties. Specifically, the boundary between Old and Medieval Chinese is the approximate time when the Qin dynasty was established in 221 BC and the Han dynasty in 206 BC, which signals the beginning of the feudal society in Chinese history. The division between Medieval Chinese and Modern Chinese refers to the transition period from the Tang dynasty to the Song dynasty: the former ended in AD 907, and the latter was established in AD 960. The division between Modern Chinese and Contemporary Chinese represents the end of the Qing dynasty (precisely AD 1912) and the beginning of the new era of the Republic of China (precisely AD 1912). Therefore our periodization is compatible with the temporal frames that a vast majority of scholars have become accustomed to using in their research. By noting these correlations, we do not mean that the development of the grammar has any relation to the rise and fall of different dynasties. The evolution of language always follows its own paths, which are determined mainly by its internal structural factors and the collective cognition of the language community, and changes do not happen at the same pace. As far as the same duration of a period is concerned, in some periods the language was quite stable and no dramatic changes happened, but in other periods many dramatic changes happened, possibly even affecting the overall texture of the language. Theoretically, any change can be dated to a particular time, date, or even minute. However, it is impossible and unnecessary to do so in a historical investigation. Although any language is always in the process of evolution, no change happens in a single day. Change is always gradual, and the language system typically remains stable over quite a long period of time. This book adopts fifty years as a basic time unit in discussing historical development. The dates that are provided after each example are understood to represent the linguistic system around that time rather than specifying the precise time when the text was actually composed. Our way to date the examples is to identify the midpoint of the life span of the author and see whether it is closer to the upper or lower time unit boundary of a fifty-year period. For example, the life span of Confucius lasted from 551 to 479 BC, and the midpoint is 514 BC; hence examples from the Analects are dated as 500 BC. In many cases, there is not enough information to identify both the authors and the time of composition; in those cases, we can provide only a very rough dating. For example, the Classic of Poems is a collection of folk songs that were composed from the eleventh to the seventh century BC, so all we can say is that it basically reflects the language system of Early Old Chinese.

1.4

Historical Texts

To explore the evolution of the grammatical system, the ideal situation is that there are always enough vernacular texts available for each stage of language development to

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

Some Preliminaries

8

accurately and entirely record the spoken language at each particular period of time. However, this kind of text is rare, and most historical texts are actually a mixture of spoken and written and are even infused with dialectal expressions. We deal with these complex situations in the following ways. First, we carefully chose texts of the genres that were most likely to reflect the system of the spoken language at the time, such as folk songs, dialogues, drama scripts, and religious speeches for ordinary people. Second, since we are interested mainly in regular phenomena that had far-reaching effects on the development of the language, we tend to ignore usages that existed only in limited texts with short life spans, some of which might belong to other dialects. More importantly, we view the whole history of Chinese as an integral system when doing our research, and we try to put every phenomenon in its proper position in the overall historical context. In this case, we can easily detect which phenomena were meaningful for the development of the grammatical system. To avoid our being biased by our own theoretical backgrounds and misled by irregular usages, the scope of the investigation must be wide enough to cover all the major texts of each period. We have read most of the representative texts that reflect the grammatical system at different times, which has given us a feeling for the general situation of the system. Additionally, electronic databases of ancient texts have enabled us to conduct a comprehensive statistical survey on any usage, which is crucial in two respects: first, many rules and principles can be revealed in the statistical data; second, these statistics can be used to test many theoretical hypotheses. Over the past two decades or so, we spent a tremendous amount of time reading the original texts, collecting a large number of examples, and manually conducting many statistical surveys. In addition, we have made use of three main important databases of ancient Chinese texts: (a) the Corpus of Ancient Chinese Texts at Academia Sinica of Taiwan, (b) the Chinese Corpus of Peking University, and (c) the Chinese Language Corpus of Beijing Language University. According to our own research experiences, the most useful theoretical frameworks investigating the development of a language are grammaticalization theory, linguistic typology, cognitive linguistics, and construction grammar. However, our general philosophy or methodology is that we learn linguistics from language and our main goal is to discover rules, principles, and any other regular phenomena in the natural language. Thus any ideas and methods can be used in our analyses if they are helpful in realizing our research goals.

1.5

The Structure of the Contents

There are many ways to organize the contents of a book in historical linguistics, which are determined largely by the purposes of the research. For example, The Cambridge History of the English Language (Hogg et al. 2001), which consists of six large volumes,

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

The Structure of the Contents

9

is structured by dividing the language into four periods and two large areas, with one volume dedicated to each. Each of the six volumes contains roughly the same topics, such as phonology, semantics, and syntax, written by different scholars. And different scholars were invited to write about the same topic relative to different periods; for example, the author of the section on syntax during the period from the beginning to AD 1066 is Elizabeth Traugott, while the author on the same topic in the succeeding period from AD 1066 to 1476 is Olga Fischer. Structuring a book in this way is convenient for descriptive historical linguistics, but it is obviously disadvantageous for exploring the motivation, mechanism, and regularity of language development. One reason is that a complete grammaticalization process usually takes several hundred years, and some might even take more than a millennium to reach the completion point of their development. For instance, an extreme case in the history of Chinese is the development of reduplication, which took nearly three millennia to expand from adjectives to verbs. More importantly, the starting and end points of a developmental process will vary from one grammatical category to another. Since the focus of this book is on the motivation, mechanism, and regularity of the evolution of Chinese grammar, mainly exploring the motivations and mechanisms behind language developments, we adopt a unique structure for the contents of this book, which distinguishes it from any other historical linguistics book, regardless of whether it is written in Chinese or any other language. Specifically, we structure and order the contents according to the natural relationship among the grammatical forms, and every change within the same cluster holds some cause–effect relationship. For example, the establishment of the resultative construction motivated the emergence of the disposal construction, the verb-copying construction, the aspect system, and verb reduplication; therefore these topics are arranged in the logical order. That is, they are discussed together within either one section or over several consecutive sections.

https://doi.org/10.1017/9781108921831.001 Published online by Cambridge University Press

2 Copular Word and Construction 2.1

Introduction

The copular construction is one of the basic grammatical apparatuses required by daily communication; hence any language or a language at different periods of time always has some kind of grammatical device to perform the corresponding function.1 Thus we begin with a fundamental change in copula structures from Old to Medieval Chinese, which in turn caused a series of consequent developments that significantly altered the texture of Chinese grammar at the time. The earliest documents show that the basic word order in Old Chinese was SVO, although there were some OV variations with special pragmatic values (for details, see Wang 1989: 198‒216, Yang and He 2001: 784‒813). In comparison, a typical SVO language, English, has a copula (i.e. be) that occurs between the subject and the complement, as ordinary verbs do, which shows personal agreement with the subject and inflects for tense and aspect. However, Old Chinese lacks such a copula verb, as there is no linking verb between the subject and the complement, and the copula construction at the time was obligatorily marked by a sentence-final particle yě that followed the complement, unlike other verbs. The copula construction in Old Chinese can be formulated as in (1). (1)

Subj (zhě), NP yě.

The word zhě in parentheses is a demonstrative, optionally used as an anaphor to refer to the subject; there is a pause between the subject and the complement that is indicated by a comma in written texts, as illustrated below. (2)




弓矢者, 器也。(易經 繫辭下) Gōng shǐ zhě, qì yě. bow arrow ANAP instrument COP “Bows and arrows are instruments.” (Yi Jing, Xi Ci, 800 BC)

1

In contrast, some other constructions may not be essential and they did not exist all the time; some might have disappeared from history, such as the verb co-ordinate construction (for details, see Chapter 4), and some might have been innovated at a certain stage, such as the disposal construction (for details, see Chapter 9).

10

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

The Copular Particle Yě (3)

11




彼, 虎狼也。(左傳 哀公六年) Bǐ, hǔ láng yě. he tiger wolf COP “He is a tiger and a wolf.” (Zuo Zhuan, Ai Gong Liu Nian, 550–400 BC)

The above copula construction was used until Late Old Chinese, when the demonstrative shì grammaticalized into a copula to connect the subject with the complement. As a consequence, the old copula construction was eventually replaced by the innovative shì structure that was consistent with the basic SVO word order of the language. It is cross-linguistically true that the copula is extremely active in further grammaticalizing into other grammatical devices, such as focus, emphasis, contrastive, agent, avertive, conditional, consecutive, and obligation marking (for details, see Heine and Kuteva 2002: 319). In Chinese, one of the further developments of shì is as a focus marker, and shì as a new, overtly formal focus marker gradually replaced the old device for highlighting an element that relied on a change in constituent orders (for a fuller discussion, see Chapter 3). The grammar of Old Chinese was typologically distinct from that of Contemporary Chinese in these respects; for instance, wh- words in Old Chinese had to be moved to precede the predicate verb, as did the focused elements and the emphasized element in negation. Due to the emergence of the copula verb shì “be,” and especially its further grammaticalization into a focus marker, the grammatical device of manipulating constituent orders entirely disappeared in the first half of the Medieval Chinese period. As repeatedly happens in many languages, the copula shì was fused with wh- words due to their frequent co-occurrences, which resulted in the dramatic development of the interrogative pronoun system (Harris and Campbell 1995: 162‒165). There were numerous changes during the transition from Old to Medieval Chinese, and the emergence of the copula shì and its later development are undoubtedly among the most significant changes with far-reaching effects on the texture of Chinese grammar.

2.2

The Copular Particle Yě

Chinese has a grammatical category – mood words, which are constrained to sentencefinal position and thus are called the “sentence-final particle.” Throughout history, mood words have played a crucial part in Chinese grammar by expressing various functions such as “yes/no” questions, uncertainty, exclamation, the imperative, and possibility. There were dozens of these mood particles in Old Chinese, but none have survived to the

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

12

Copular Word and Construction

present day. A new set of these particles that are used in Contemporary Chinese was introduced into the language mainly after the tenth century AD. It is assumed that sentence-final particles are correlated with SOV languages (Greenberg 1966a, Li and Thompson 1981: 16), but whether there is an intrinsic correlation is still unclear; hence the existence of the sentence-final particle is not sufficient to be considered evidence that Proto-Chinese was ever an SOV language.2 For example, the category of sentence-final particles is not found in any languages of the Indo-European family, but their ancestor languages, such as Latin, are said to have been SOV. Sentence-final particles may be a regional linguistic phenomenon; for instance, Japanese and Korean, two SOV languages, also possess rich sentence-final particles. According to Wang (1989: 295), the question particle ma in Contemporary Chinese developed from the negative marker wú “not have” that was used in sentence-final position to form an affirmation–negation question. The factual ne seems to come from a locative preposition phrase zài nàlǐ “in there” when used in sentence-final position. These facts suggest that the existence of sentence-final particles in a language may have nothing to do with its basic word order because they are equally common in SVO and SOV languages. More importantly, they can be grammaticalized from lexical sources in an SVO language, as shown in the case of Medieval Chinese. The particle yě was the most frequently used sentence-final particle, and had the most numerous functions in Old Chinese (He 2004: 412‒416). According to Guo’s (1997) investigation, it did not yet appear in the oracle bone inscriptions, bronze inscriptions, or the Shang Shu (the Book of Documents), probably because of the constraints of technical writing difficulty or genre (e.g. the Shang Shu is about politics and philosophy). However, the particle yě is widely attested in the Shi Jing (Classic of Poems), mainly a collection of folk songs that might reflect the language from the eleventh to seventh centuries BC. That is, even in the earliest texts, yě was already highly grammaticalized, with multiple functions. Although we do not know how it evolved out of a lexical item because of the lack of empirical evidence, we can reconstruct the possible grammatical pathway on the basis of regularity, either diachronic or cross-linguistic. Having conducted a comprehensive quantitative study of the Zuo Zhuan (550– 400 BC), He (2004: 412) revealed that of the seven major functions of the particle yě, such as copula, declarative, explanatory, exclamatory, enumerative, imperative, and interrogative, 46 percent of instances, or roughly half of the total tokens, represent

2

Peyraube (1997a, 1997b) claimed that Proto-Chinese actually showed a regular order of SVO and was indeed more thoroughly SVO than later stages (Early or Late Old Chinese). Thus it is more likely that Old Chinese employed the word order shift of the verb and the object to express certain pragmatic functions such as question, focus, or emphasis, which are also found in other languages like English.

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

The Copular Particle Yě

13

a copular construction. Considering that the particle yě is the necessary marker for the copular construction, as mentioned earlier, it is safe to say that the basic and essential function of the particle yě was to mark a copular construction, which can be called a “copular particle,” and the other functions of yě were all derived from its basic copular usage, via a series of further grammaticalization processes. From this angle, we can see that the seemingly extremely complex meanings of yě become quite neat and regular. More examples of the copula yě are given below.




(4)

制, 岩邑也。 (左傳 隱公元年) Zhì, yán yì yě. Zhi rocky city COP “Zhi is a rocky city.” (Zuo Zhuan, Yin Gong Yuan Nian, 550–400 BC)

(5)

余, 而所嫁婦人之父也。(左傳 宣公十五年) Yú, ér suǒ jià fùrén zhī fù yě. I you NOM marry woman REL father COP “I am the father of the woman who was married (to someone else) by you.” (Zuo Zhuan, Xuan Gong Shi Wu Nian, 550–400 BC)




In Old Chinese, the behavior of yě revealed that it had been grammaticalized from a verb since it still preserved some properties of verbs of the time. Specifically, it could be followed by other sentence-final particles, similar to ordinary verbs, as illustrated in (6). (6)




此亦妄人也已矣。 (孟子 離婁下) Cǐ yì wàngrén yě yǐ yǐ. this also crazy-people COP already PERF “This is also a crazy person.” (Meng Zi, Li Lou Xia, 300 BC)

In (6), the copula yě is followed by the adverb yǐ “already” and the perfective aspect marker yǐ that indicates an action in the past with current relevance. In the grammar of Old Chinese, only ordinary verbs could be followed by such adverbs and aspect markers. In Contemporary Chinese, the verb-like copula shì can also be either preceded by an adverb or followed by an aspectual suffix, e.g. yǐjīng shì-le “already be-PERF.” Guo (1997) found that when two or more sentence-final particles co-occurred, yě was always followed by the others but could not follow others, a sign that it still retained some verbal properties; in other words, the degree of the grammaticalization of yě was relatively lower than that of the aspectual yǐ, for instance.

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

Copular Word and Construction

14

One puzzle remains unresolved: why did the copular marker yě occur after the complement? How did this structure come into existence in an SVO language? Where did it come from? The copular structure in Old Chinese looks quite foreign because the basic word order of the Chinese language has been SVO throughout history. There may be two possible explanations for this phenomenon. First, Proto-Chinese might have been an SOV language and a verb in sentence-final position grammaticalized into a copula. Due to the high frequency of the copular construction, it survived into Old Chinese. As indicated above, this explanation is not quite convincing because a sentence-final particle could develop in an SVO language. Second, also due to the high frequency of the copular word, yě was phonologically reduced to a clitic and was restricted to sentence-final position by treating the complement as its host, though it initially became grammaticalized in the position between the subject and the complement. This hypothesis is based on a similar phenomenon in Polish, in which the basic word order is SVO but the copula has become a clitic and must be suffixed to a verb. The second explanation is more promising, but at this point we would like to leave the question open for further study.

2.3

The Focus Marker Yě

A copula is a functional word that is usually grammaticalized from other lexical sources such as demonstratives or verbs of either sitting or standing (Heine and Kuteva 2002: 330). Furthermore, it often serves as a lexical source for other grammatical morphemes such as focus, obligation, conditional, and avertive marker. Here, we are concerned only with yě as a focus marker, one of its multiple functions in Old Chinese. The development of a copula into a focus marker is common across languages, and the steps are generalized by Heine and Reh (1984: 249) as follows: Stage 1 There is a cleft structure: “NP/PP copula + subordinate clause,” where the sentence-initial constituent is made new and foregrounded information, and the subordinate clause represents given and presupposed information. Stage 2 In the above context the copula is grammaticalized into a focus marker, and is often used to emphasize the inherent focus property of wh- words. As expected from the language universal mentioned above, the copula yě also developed into a focus marker that was already widely used in the earliest texts. Thus, unlike other sentence-final particles that were constrained to their original position (e.g. yǐ “perfective aspect” could occur only in sentence-final position), yě had much more freedom of distribution within a sentence, and it actually could

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

The Focus Marker Yě

15

follow any constituent of a sentence, such as the subject, embedded clauses, or preposition phrases, to highlight the preceding element. One of its major highlighting functions was to make a subject focused, as illustrated below.




(7)

女也不爽, 士貳其行。(詩經 氓) Nǚ yě bù shuǎng, shì èr qí xíng. wife FOC not wrong husband disloyal his behavior “It is not the wife who did something wrong. The husband is disloyal in his behavior.” (Shi Jing, Mang, 1000–600 BC)

(8)

人不堪其憂, 回也不改其樂。 (論語 雍也) Rén bù kān qí yōu, Huí yě bù gǎi qí lè. people not tolerate the hardship Hui FOC not change his happy “Other people cannot tolerate the hardship, but only Hui did not change his happiness.” (Lun Yu, Yong Ye, 500 BC)




In (8), in contrast with the preceding clause, the function of yě in highlighting the preceding subject is obvious. Examples of this sort were so common in Old Chinese that the subject marker yě seemed to exist at the time. Some researchers have mistakenly regarded the above use of yě as a topic marker because the topic refers only to given and backgrounded information, which is exactly the opposite of the semantic feature of the focus. Indeed, Old Chinese had a topicalization construction that can be schematized as follows: (9)

The topicalization construction in Old Chinese TOP, Subj + V + ANAP

In (10), the underlined part is the topic through extraction from its original object position, which is usually filled with the anaphoric pronoun zhī. (10)




夏禮, 吾能言之。 (論語 八佾) Xià lǐ, wú néng yán zhī. Xia rite I can talk ANAP “As for the rites of the Xia Dynasty, I can talk about it.” (Lun Yu, Ba Yi, 500 BC)

At the time, the topic could never be followed by the focus yě because they semantically contradict each other: the former refers to given information and the latter to new information. Since Chinese is regarded as a topic-prominent language (Li and Thompson 1977),

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

Copular Word and Construction

16

instances of topicalization are very common throughout its history, but there is an important development regarding the structures of topicalization between Old and Contemporary Chinese: Contemporary Chinese no longer allows an anaphoric pronoun to occur in the object position; otherwise, the sentence becomes ill-formed. In Old Chinese, another common structure of the focus yě is the copular construction with a cleft clause as the subject, as illustrated in (11) and (12).




(11)

能補過者, 君子也。 (左傳 昭公七年) Néng bǔ guò zhě, jūnzǐ yě. can correct mistake those virtuous-man YE “It is the virtuous men who can correct their mistakes.” (Zuo Zhuan, Zhao Gong Qi Nian, 550–400 BC)

(12)

故為淵驅魚者, 獭也。 (孟子 離婁上) Gù wèi yuan qū yú zhě, tǎ yě. therefore for lake repel fish those otter YE “It is the otters who repel fishes from the lake.” (Meng Zi, Li Lou Shang, 300 BC)




In the above examples, the complementizer zhě introduces a cleft clause that expresses the backgrounded information, and the NPs marked by yě are the focus of the sentences. The constituent ordering is exactly the reverse of the English focusing construction: “it be NP that clause,” where the highlighted NP precedes the cleft clause (Quirk et al. 1985: 10‒54). As a language universal, wh- words inherently possess a “focus” feature that may be realized differently in different languages: some languages use variations of constituent orders (e.g. English) and some attach a focus morpheme to them (e.g. Medieval Chinese, the Southern Min dialect). Both of these methods were used in Old Chinese, but in different contexts. At the time, in sentences with an action verb, the first method was used to realize the inherent focusing feature of wh- words (for details, see Chapter 3). However, if a wh- word is highlighted by the focus yě, their combination can occur only in sentence-final position, as illustrated in (13) and (14), regardless of their original position in the sentence. (13)




追我者, 誰也? (孟子 離婁下) Zhuī wǒ zhě, shéi yě? chase I those who FOC “Who chased me?” (Meng Zi, Li Lou Xia, 300 BC)

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

The Focus Marker Yě (14)

17




君與我此, 何也? (國語 晉語一) Jūn yǔ wǒ bǐ, hé yě? you give I this why FOC “Why did you give me this?” (Guo Yu, Jin Yu Yi, 350 BC)

In (13), the wh- word shéi “who” is the subject, which is supposed to occur in sentenceinitial position; in (14), the wh- word “why” is an adverb, which is supposed to occur between the subject and the predicate. When their inherent “focus” feature is strengthened by adding the focus marker yě, the wh- words must be extracted to sentence-final position. In the grammar of Old Chinese, sentence-initial position is reserved for the topic that expresses old and backgrounded information, and sentence-final position is reserved for the focus that expresses new and foregrounded information. However, this contrast disappeared after Old Chinese. In addition, the position between the subject and the predicate is an unmarked focus position for the object, wh- words, and NPs, an issue that will be discussed in depth in Chapter 3. The focus yě could also highlight a preposition phrase in the way that the preposition phrase was extracted to sentence-final position and followed by the focus yě, as illustrated in (15). (15)




城小穀, 為管仲也。 (左傳 莊公三十二年) Chéng Xiǎo Gǔ, wèi Guǎn Zhòng yě. city Xiao Gu for Guan Zhong YE “It is for Guan Zhong to build a city in Xiao Gu.” (Zuo Zhuan, Zhuang Gong San Shi Er Nian, 550–400 BC)

In a normal declarative sentence, the preposition phrase wèi Guǎn Zhòng “for Guan Zhong” occurs before the matrix verb. In Contemporary Chinese, the preposition is highlighted in its original position, without extraction, being preceded by the focus shì (we will return to this issue in Chapter 3). According to documentary evidence, the yě structure as a canonical copular construction existed for nearly fifteen centuries (from the eleventh century BC to the fifth century AD). The copular word is verb-like, cross-linguistically. As pointed out previously, the yě structure is inconsistent with the basic SVO word order of the language since the particle yě occurs only in sentence-final position. Thus it is possible in this situation that a verb-like copula was introduced into the language and eventually replaced the yě structure. In this situation, the demonstrative shì was motivated to grammaticalize into a copula between the subject and the complement and eventually replaced the old yě structure over approximately 500 years.

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

Copular Word and Construction

18 2.4

The Copula Shì

In Contemporary Chinese, the copula shì must be used to connect the subject with its complement. The earliest examples of the copula shì appeared in texts composed around the first century BC. It is generally accepted that the copula usage of shì grammaticalized from its earlier use as a demonstrative (Li and Thompson 1977, Wang 1989: 183‒197). Indeed, copulas in many languages, such as Ancient Egyptian, Vai, and Sranan CE, developed out of demonstratives. However, a grammaticalization event can happen only in certain very specific contexts, and demonstratives usually have various pragmatic functions. Therefore the following question arises: which function and what context are directly responsible for the emergence of the copula out of its original demonstrative use? Li and Thompson (1977) argued that the copula shì was derived from its resumptive pronominal usage. If this was the case, the demonstrative zhě should also have developed into a copula because it was often used as a resumptive pronoun immediately following the subject, as illustrated in (2); however, zhě did not actually have such a development. Hengeveld (1992) argued that the mechanism of the emergence of a copula out of its demonstrative usage can be identified through reanalyzing the theme-then-clause construction as a subject–predicate construction. However, this is not the story for the development of the copula shì, as we will see below. In the earliest documents such as the Shi Jing (from the eleventh to the seventh centuries BC), shì was already used as a demonstrative with various pragmatic functions. According to Wang (1989: 187), in the early stage of Old Chinese, when the demonstrative shì was used as an object, it often preceded the verb, as did other pronouns, as illustrated in (16) and (17), unlike other object nouns that followed the verb. This might signal that ProtoChinese was an SOV language, and, cross-linguistically, pronouns usually lag behind in the development of the language because of their high frequency. A similar phenomenon is found in French, where pronouns still preserve the SOV word order, a feature of Old Latin, but ordinary nouns must take the SVO word order when used as an object.




(16)

民獻有十夫予翼。(尚書 大誥) Mín xiàn yǒu shí fū yǔ yì. people send have ten man I assist “The people sent ten men to assist me.” (Shang Shu, Da Gao, 1100–800 BC)

(17)

民具爾瞻。 (詩經 節南山) Mín jù ěr zhān. people all you look “All the people look up to you.”




(Shi Jing, Jie Nan Shan, 1000–600 BC)

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

The Copula Shì

19

In Old Chinese, another major pragmatic usage of the demonstrative shì was as an anaphor to refer back to a whole preceding clause, as illustrated in (18), or to a preceding highlighted element, as illustrated in (19).




(18)

昭王南征而不復, 寡人是問。(左傳 僖公四年) Zhāo Wáng nán zhēng ér bù fù, Guǎrén shì wèn. Zhao King south expedition and not return I this inquire “King Zhao did not return from his expedition to the south. This is what I inquire about.” (Zuo Zhuan, Xi Gong Si Nian, 550–400 BC)

(19)

率師以來, 惟敵是求。 (左傳 宣公十二年) Shuai shī yǐ lái, wéi dí shì qiú. lead troop with come only enemy this seek “I led the troops to arrive here. It is only the enemy that we sought.” (Zuo Zhuan, Xuan Gong Shi Er Nian, 550–400 BC)




The contexts depicted above cannot provide the proper circumstance for the grammaticalization of the copula shì because they bear no resemblance to the syntactic configuration of the copular construction. The emergence of the copula shì involved the innovation of a new copular construction whose schema is “Subj COP Complement.” Around the fifth century BC, the proper context emerged for grammaticalizing the demonstrative shì into a copula, with shì occurring in the initial position of the last clause, as an anaphor to refer back to preceding clauses, as illustrated below.




(20)

若棄德不讓, 是棄先君之舉也。 (左傳 隱公三年) Ruò qì dé bù ràng, shì qì xiān-jūn zhī jǔ yě if abandon virtue not yield ANAP abandon ancestor AOSS act COP “If one abandons the virtues and does not yield, this is the act of abandoning his ancestors.” (Zuo Zhuan, Yin Gong San Nian, 550–400 BC)

(21)

敏而好學, 不恥下問, 是以謂之文也。 (論語 公冶長) Mǐn ér hàoxué, bù chǐ xià wèn, shì yǐ wèi zhī wén sensitive and studious not shame blow ask ANAP for call he literate yě. COM “He is sensitive and studious, and does not feel ashamed to seek advice from his subordinates, so he is called literate.” (Lun Yu, Gong Ye Chang, 500 BC)

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press




Copular Word and Construction

20

In (21), the anaphoric shì is the object of the preposition yǐ “for,” whose order is inverted, but ordinary nouns typically follow the preposition, according to the grammar at the time. In this context, shì introduces an assessment of the preceding clauses, with the anaphoric shì and the following text forming a copular instance, as shown in (21). Even so, this context did not trigger the grammaticalization of the demonstrative shì because the preceding materials were clauses rather than nouns, which did not quite match the stereotype of the copular construction, whose subject is usually a nominal phrase. The above discourse structure was increasingly used from the fifth century BC onward. At the time, the elements following the anaphoric shì were either a noun phrase, as in (20), or a clause, as in (21). According to our comprehensive investigation of the texts composed around the fifth century BC, there were no examples in which the following material was a bare noun or a proper name. With the discourse structure of the anaphoric shì developing, an increasing number of bare nouns (typically proper names) appeared following the anaphoric shì. The preceding materials could be called “topic” and the following ones “comment.” If the topic is a proper name or a bare noun with no modifier (the underlined parts of the following examples), there must be an inversion between the comment and the anaphoric shì in the last clause, where the copular marker yě was obligatory at the time, as schematized below. (22)

Topic, Comment shì yě. ANAP COP

The above discourse structure is illustrated in (23) and (24).




(23)

臣聞七十里而為政天下者, 湯是也。 (孟子 梁惠王) Chén wén qīshí lǐ ér wéi zhèng tiānxià zhě, Tāng shì yě. I hear seventy mile and do politics world one Tang ANAP COP “I heard of one who built a country within an area of seventy square miles, who is Tang.” (Meng Zi, Liang Hui Wang, 300 BC)

(24)

天不能死, 地不能埋, 仲尼、子貢是也。 (荀子 榮辱) Tiān bù néng sǐ, dì bù néng mái, Zhòng Ní, Zi Gòng shì heaven not can kill earth not can bury Zhong Ni Zi Gong ANAP yě. COP “The heaven cannot kill them and the earth cannot bury them: those are Kong Zi and Zi Gong.” (Xun Zi, Rong Ru, 250 BC)

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press




The Copula Shì

21

In the above examples, shì was still used as a pronominal word, an anaphor to refer to the preceding clauses. In this context, it was combined with proper nouns to form a copular construction marked by yě. This usage signaled an important development that created a proper context that triggered the development of the pronominal shì into a verb-like copula. More importantly, the topic could also be a noun phrase, as exemplified in (25) and (26). (25)




天下之道, 管是矣。 (荀子 儒效) Tiānxià zhī dào, guǎn shì yǐ. world GIN way govern ANAP COP “The way of the world is government.” (Xun Zi, Ru Xiao, 250 BC)

(26)




百王之道, 一是矣。 (荀子 儒效) Bǎi wáng zhī dào, yī shì yǐ. hundred king GEN principle unity ANAP COP “The principle that hundreds of kings followed is unity.” (Xun Zi, Ru Xiao, 250 BC)

In the above two examples, the comments are simple bare nouns, as shown by the parts that are underlined. At this stage, the anaphoric shì occurred in the following two structures: (27)

(a) Clause, Proper Name + shì + yě. (b) NP, Bare Noun + shì + yě.

In the above structures, the last clause (i.e. after the comma) constitutes a copular construction that is marked by the copula yě. At the time, as analyzed previously, the copula yě had developed out of multiple functions to mark various declarative sentences, expressing moods such as factual, assertive, and narrative. The most common elements that occupy the position of shì in the constructions of (27a) and (27b) are ordinary verbs when they are preceded by a nominal subject. It was in this context that the basic word order of the language initialized the operation of analogy to reanalyze the pronominal shì into a verb-like copula. This hypothesis finds solid empirical evidence in the earliest examples of the copula shì, where the subjects are typically pronouns or proper names: (28)




客人不知其是商君也。 (史記 商君列傳) Kèrén bù zhī qí shì Shāng Jūn yě. guest not know he be Shang Mr COP “The guests did not know that he was Mr Shang.” (Shi Ji, Shang Jun Lie Zhuan, 100 BC)

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

Copular Word and Construction

22 (29)




鬷夷氏是其後也。 (論衡 龍虛) Zōngyí Shì shì qí hòu yě. Zongyi Surname be his descendant COP “Zongyi Shi was his descendant.” (Lun Heng, Long Xu, AD 100)

Note that in the beginning the copula shì was still accompanied by the former copular particle yě, and the old and new forms were used together to mark the copular construction. Whether the sentence-final particle yě was present is a reliable criterion for confirming that shì had already become a pure copula. When shì became a copula, the particle yě was no longer needed to mark the copular construction, so it gradually disappeared from the copular construction, as was evidenced in some of the earliest examples of the copula shì illustrated below. (30)




此是君家果。 (世說新語 言語) Cǐ shì jūn jiā guǒ. this be your family fruit “These are the fruits of your family.” (Shi Shuo Xin Yu, Yan Yu, AD 450)

(31)




我是李府君親。 (世說新語 言語) Wǒ shì Lǐ Fǔjūn qīn. I be Li Fujun relative “I am the relative of Li Fujun.” (Shi Shuo Xin Yu, Yan Yu, AD 450)

Another sign of the acquisition of the verbal properties of the newly born copula shì is that it could be modified by adverbs, similar to other ordinary verbs, as illustrated in (32). (32)




或時是鳳凰。 (論衡 講瑞) Huò shí shì fènghuáng. some occasionally be phoenix “Some of them are occasionally phoenixes.” (Lun Heng, Jiang Rui, AD 100)

In the above example, the adverb shí “occasionally” modifies the copula shì; however, the particle yě is absent. Only from this time can we say that the demonstrative shì had developed into a verb-like copula. It is safe to say that the demonstrative shì grammaticalized into a copula around the first century BC, and it coexisted or competed with the old yě copular construction for

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

The Copula Shì

23

approximately 500 years. By the fifth century AD, the new copular construction with shì became dominant at the expense of the yě structure. The evolution of grammar can be categorized into two types. First, a newly innovated device replaces an old one; that is, the grammatical categories all exist in the language and only the specific forms and markings have undergone changes. The diachronic development of the copular construction belongs to the first type. Second, a newly emerging grammatical device creates a novel grammatical device for the language; that is, this device did not exist before. The disposal construction belongs to the second category (for details, see Chapter 9). According to our observations, the rise and decline of new and old grammatical devices do not happen randomly. The reason why a new form can emerge, gradually mature, and eventually replace an old one is usually that it is favored by the overall properties of the grammar at that stage. Innovative grammatical devices usually reflect the trend of development and are more consistent with the general properties of the grammar. This is why the copular construction with shì was introduced into the language and developed strongly enough to replace the old one with the sentence-final particle yě. As we discussed above, in the proper context the demonstrative shì was reanalyzed as a copula by the operation of analogy with the basic word order, namely the SVO form. Additionally, the copular construction of shì was consistent with the basic word order. For these reasons, the shì construction was empowered by the overall grammatical system to develop from weak to strong and eventually became the new paradigm. The emergence and development of the shì copular construction had a far-reaching impact on the texture of Chinese grammar. First, the firm establishment of the shì construction as a basic and extremely frequent structure to a great extent strengthened the basic SVO word order of the Chinese language. Second, the copula shì further developed into a marker with multiple functions, such as focus, emphasis, comparison, and conjunction. The marker shì is a function word which directly precedes the highlighted or emphasized constituents, which were originally marked by inverting the constituent order in Old Chinese. The disappearance of the method of inverted constituent order caused a series of historical typological changes in Chinese grammar. The history of any grammatical system inevitably consists of two phenomena, death and birth, similar to any biological phenomenon. The birth of the shì structure and the death of the yě structure are two events caused by the same change. Moreover, all the many other usages of the particle yě, one of the most frequently used grammatical markers in Old Chinese, died out in the first half of the Medieval Chinese period. Shì itself finally became a canonical copula around the fifth century AD, entirely depriving it of its early pronominal usages.

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

Copular Word and Construction

24 2.5

Copula and Wh- Movement

During its history, the Chinese language underwent a dramatic typological change from a rigorous wh- movement language to a complete wh- in situ language (for a fuller discussion, see Chapter 3). Generally, in the period of Old Chinese, wh- words used as objects needed to be fronted to preverbal position, though the basic word order was SVO throughout history. However, during the first half of the Medieval Chinese period, whmovements were gradually replaced by wh- in situ. According to He (2004: 265), for instance, approximately 99 percent of the wh- words shuí “who” and hé “what” in Old Chinese were fronted to preverbal position when used as an object, as illustrated in (33). In addition, the object wh- elements of prepositions were fronted to the left periphery of the preposition, as illustrated in (34). (33)




其子焉往? (孟子 離婁上) Qí zi yān wǎng? his son where go “Where did his son go?” (Meng Zi, Li Lou Shang, 300 BC)

(34)




吾誰與為鄰? (莊子 德充符) Wú shuí yǔ wéi-lín. I who with neighbor “With whom do I neighbor?” (Zhuang Zi, De Chong Fu, 300 BC)

In (33), the object of the first clause is the wh- word yān “where” that occurs prior to the verb wǎng “go.” In the first clause of (34), shuí “who” precedes the preposition yǔ “with,” and the whole preposition phrase occurs prior to the verb. However, in certain instances, the object argument wh- words stayed in situ, which was a highly regular phenomenon. For example, if the predicate was a semicopula wéi, the complement wh- words did not undergo movement, as illustrated in (35). (35)




此為何器? (荀子 宥坐) Cǐ wéi hé qì? this be what instrument “What instrument is it?” (Xun Zi, You Zuo, 250 BC)

Obviously, the syntax of wh- words in Old Chinese was different from that in English. As the above translations show, in English, the wh- words used as the complement of the copula be must be moved to sentence-initial position. As analyzed previously, Old Chinese lacked a copular verb, and the canonical copular construction is “NP, NP yě,”

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

Copula and Wh- Movement

25

where the sentence-final particle yě was necessary for marking the copular construction. The verb wéi, which originally meant “do,” was occasionally extended to function as a copula. The typical copula shì emerged around the first century BC and grammaticalized from its demonstrative usage (Wang 1989: 183, Guo 1997), as discussed above. The inherent meaning of the copular verb is unbounded, and it played a key role in motivating the typological change of Chinese from a wh- movement language to a whin situ language. Moro (2006) noted that the wh- movement in copular sentences is different from that in other noncopular sentences. Old Chinese was a rigid wh- movement system, but it was sensitive to the boundedness of verbs. The semicopula wèi in Old Chinese always licensed a wh- in situ. At the beginning of Medieval Chinese, the typical copula shì emerged and licensed wh- in situ. More crucially, the copula shì further grammaticalized into a grammatical morpheme to mark a focused constituent, which ultimately replaced the old device of word order change (i.e. wh- movement and highlighted constituent movement). Wh- phrases immediately follow the subject when used as the subject complement, and they are marked by the sentence-final particle yě. Thus the anaphoric pronoun zhě is generally absent, as illustrated in (36). (36)




此誰也? (戰國策 齊策) Cǐ shuí yě? this who PRT “Who is this?” (Zhan Guo Ce, Qi Ce, 450–200 BC)

Since there was no copula in Old Chinese, wh- words, usually marked by the sentencefinal particle yě, were immediately adjacent to the subject when used as the complement. Due to the emergence of the copula shì, the canonical construction of copular sentences underwent the following change during the transitional period from Old Chinese to Medieval Chinese: (37)

[Subj (+ zhě), NP + yě] > [Subj + shì + NP]

In the history of the language, yě and shì belong to different word classes: the former is a sentence-final particle, and the latter is a linking verb. This change had the potential to affect the texture of the grammar because of the high frequency of copular instances. The course of the transition started as early as the first century BC, which is exactly when the copula shì grammaticalized from its demonstrative source, and this transition was not completed until the sixth century AD. Here, we can sketch the channels and stages of the typological change from wh- movement to wh- in situ. The manner of linguistic

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

26

Copular Word and Construction

replacement between the old and new forms can be formulated as A > A/B > B (Hopper and Traugott 2003: 67), which means that there is always an intermediate stage where the old and new forms coexist for quite a long period. The emergence of the copula verb shì, and especially its further grammaticalization into a focus marker, had a profound effect on the texture of the grammar, which will be discussed in depth in the following chapter.

https://doi.org/10.1017/9781108921831.002 Published online by Cambridge University Press

3 Focus and Wh- Words

3.1

Introduction

Research on wh- movement and wh- in situ has occupied a central place in the development of generative linguistics, leading to many theoretical hypotheses about the nature of syntax (Chomsky 1964, Ross 1967, Rizzi 1990, Huang et al. 2009: 260‒281).1 In the earliest days of generative linguistic studies, the major topic of research centered on the phenomenon of wh- movement, which formed the foundation of important theoretical constructs and principles. Although the properties of wh- in situ had been investigated as early as the 1960s, it did not become a major subject of interest until the late 1970s, when a logical form was proposed by Chomsky (1976). In the logical form, in situ wh- words are considered a kind of quantifier to undergo raising. Aoun et al. (1981) and Huang (1982: 251‒268) first proposed that in situ wh- words undergo covert movement in logical form. As a result of the works of Huang (1982), Aoun and Li (1993), Cheng and Huang (1996), and others, it has now been established that Chinese is a wh- in situ language. In the past three decades, the motivations and mechanisms for distinguishing between wh- movement and wh- in situ have been subjects of major interest in theoretical linguistics. During its history, however, Chinese underwent a dramatic typological change from a rigid wh- movement language to a complete wh- in situ language, as mentioned in Chapter 2. Generally, in the period of Old Chinese, wh- words used as objects needed to be fronted to preverbal position,2 though Chinese was an SVO language similar to

1

2

This chapter aims to explore the diachronic motivation and mechanism for the typological change of Chinese from a wh-movement language to a wh- in situ one. By using the term “whmovement,” we simply refer to the fact that the syntactic position of wh-elements is different from that of the corresponding non-wh- elements. We leave this purely theoretical issue open: what were the derivations in the wh- movement of Old Chinese? In the framework of lexicalfunctional grammar, there is no wh- movement and no derivation at all (see Bresnan 2001, Dalrymple 2001 for a comprehensive introduction; see Mycock 2007 for an analysis of related phenomena in Hungarian). According to He (2012: 178‒193), for instance, more than 99 percent of the wh- words shuí “who” and he “what” in Old Chinese occurred in preverbal position when used as an object.

27

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

28

English.3 However, during the first half of the Medieval Chinese period, wh- movements were gradually replaced by wh- in situ. This change is illustrated in (1) and (2): (1)




客何好? (戰國策 齊策) Kè hé hào? guest what love “What do you love?” (Zhan Guo Ce, Qi Ce, 450–200 BC)

(2)

您喜歡什麼? (現代漢語) Nín xǐhuān shénme? you love what “What do you love?” (Contemporary Chinese)

In (1), the object of the first clause is the wh- word hé “what” that occurs prior to the verb hào “love.” As shown in (2), however, with the Contemporary Chinese translation of (1), no wh- movement is possible, and wh- words must stay in situ. Unlike in English, the landing site of wh- words in Old Chinese is between the subject and the verb rather than in sentence-initial position. This is constrained by the topicalization structure, where sentence-initial position is reserved for the topic. In the literature, many sophisticated hypotheses have been proposed to explain what factors license wh- in situ. All of these studies have been based on cross-linguistic comparison, the semantic properties of wh- words, or the syntactic structures of questions within a particular language. Additionally, there has been some speculation about the motivations for the typological change in Chinese from wh- movement to wh- in situ; for instance, Xu (2006) and Huang (2006) claimed that the cause of this change was the rise of light verbs in Medieval Chinese, which changed Chinese from a synthetic language to an analytical language (Feng 2014). However, they did not identify any specific motivations and mechanisms for the typological change regarding wh- questions. If we examine the issue from a diachronic perspective, however, the truth may be quite straightforward. This chapter assumes that every wh- word contains a focus feature (henceforth +FOC) that must be checked by means of certain grammatical devices. Thus the behaviors of wh- words in a particular language are generally in accordance with the canonical focus expressions. As pointed out in Chapter 2, Old Chinese lacked a copula and a focus marker; hence word order variation was the only device for focusing expressions. In this period, wh- movement was a reflection of these focusing

3

In Chapter 2, we explain that these object-fronted constructions are grammatical devices to highlight those elements.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Licensing of Wh- In Situ

29

expressions. As we saw in Section 2.4, in the Late Old Chinese period, the demonstrative shì developed into a copula and later further grammaticalized into a focus marker that served to check the +FOC of wh- words. As a result, wh- movement gave way to whin situ.4 In addition, wh- movements in Old Chinese were sensitive to the transitivity of verbs. Only content-transitive verbs and prepositions grammaticalized from transitive verbs could trigger the movement of wh- words. However, semantically light verbs and lowtransitivity verbs or prepositions, such as the semicopula wéi and the general verb nài “do,” could not license wh- movement. Naturally, the newly emerging copula shì never triggered wh- movement. Therefore “shì + wh- word” was the only possible grammatical construction, and its high frequency caused the change from “wh- word + V” to “V + wh- word” via analogy.

3.2

The Licensing of Wh- In Situ

In this section, we examine the most influential hypotheses about the licensing of whin situ: (a) the “Q-morpheme” hypothesis, (b) the “quantificational and covert movement at LF” hypothesis, and (c) the “focus marking” hypothesis. Historical empirical data are extremely helpful in judging the merits of these proposals, because they enable us to avoid inadequate analyses that are based mainly on artificial examples. As Lin (2014) has noted, empirical data are the foundation of theory formation, but researchers’ intuitions often distort the truth of the grammar. Thus a diachronic approach is needed to clarify the crucial issues involved.

3.2.1

The “Q-Morpheme” Hypothesis

Many researchers believe that overt wh- movement relates to a Q-morpheme in C°, a notion that was first posited by Katz and Postal (1964) and dominated the discussion of wh- question structure in the late 1960s and early 1970s (for details, see Cheng 2003). This Q-feature can be realized as lexical items such as whether/if, the Subj–AUX inversion, or question particles. Chomsky (1995) considered the Q-feature to be a matter of degree, and only when it is strong enough can it trigger overt movement. This speculation was supported by scarcely any empirical evidence and hence was soon abandoned. It can hardly be said that the Q-feature is stronger in English than in Chinese; hence English is a wh- movement 4

In our view, the typological change discussed in this chapter was motivated by language-internal factors. However, language contact may have caused a similar change in other languages; according to Duguine and Irurtzun (2014). For example, Labourdin Basque underwent a change from obligatory wh- movement to optional wh- in situ due to language contact.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

30

language, whereas Chinese is a wh- in situ language. The obvious counterevidence is that Chinese was also historically a wh- movement language. Chomsky (2000), without mentioning Q-feature strength, admitted that the Q-feature has no relation to either whin situ or wh- movement. Baker (1970) discussed the correlation between wh- in situ and the Q-morpheme. He assumed that English-type languages have a Q-morpheme in sentence-initial position and that a wh- word moves to be adjacent to it. Since Japanese-type languages have a question particle in sentence-final position, their overt wh- movements are accordingly blocked. In line with Baker’s analysis, Kayne (1994) claimed that in languages with sentence-final question particles, the Spec-CP, where a wh- word is supposed to occur, is always occupied. In their views, this explains why wh- words in Japanese do not undergo overt movement. However, the history of Chinese provides a counterexample for Kayne’s hypothesis. Chinese has always possessed sentence-final question particles throughout its history, but wh- words in Old Chinese did undergo overt movement, whereas those in Modern Chinese did not. In addition, in a broad typological survey and detailed comparison of Passamaquoddy and Chinese, Bruening (2007) revealed that there is actually no particular correlation between the positioning of question particles and wh- in situ. Following the “Q-morpheme” hypothesis, Cheng (1991) assumed that in Chinesetype languages, question particles function to determine the type/force of question and therefore render overt wh- movement unnecessary. Some researchers have gone even further and claimed that wh- words in Chinese do not have inherent Q-features and that they are actually “negative-polarity items” (see Huang 1982, Cheng 1991, Li 1993, Lin 1996). Aoun and Li (1993) argued that the relation between wh- words and a Q-operator is a bindee–binder relation. Thus the wh- words in Chinese cannot express a question by themselves but can do so only by means of the question particle ne. However, this observation is entirely incorrect for the following two reasons. First, the sentential particle ne is not a genuine question particle in Modern Chinese (see Dong 2009, Shao 1996); rather, it is a particle that confirms the actuality of an event. Our own investigation shows that 98 percent of wh- questions can raise questions by themselves without the help of the particle ne. When the particle ne is added to a whquestion, the force of the question is weakened; in most cases, wh- questions with this particle are actually rhetorical questions, which means that the speaker already knows the answer and does not expect an answer from the hearer, as illustrated in (3). (3)

誰叫你偷去呢? (紅樓夢六回) Shuí jiào nǐ tōu-qù ne? who ask you steal-go PRT “Who asked you to steal it?” (lit.) = “You should not have stolen it.” (Hong Lou Meng, Chapter 6, AD 1750)

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Licensing of Wh- In Situ

31

Second, Mandarin Chinese has a real question particle ma that cannot co-occur with wh- words in most cases; such a sentence would be ill-formed. When wh- words cooccur with this particle, the wh- words must be interpreted as indefinite or existential, as illustrated in (4). (4)

這是不算什麼的嗎? (紅樓夢一〇八回) Zhè shì bù suàn shénme de ma? this be not account-for what REL QUE “Doesn’t this account for anything?” (lit.) = “This doesn’t account for anything.” (Hong Lou Meng, Chapter 108, AD 1750)

According to Wang (1989: 295), the sentence-final particle ne was not introduced into the language until the twelfth century AD. By then, wh- words had been used for more than 2,500 years. Old Chinese also had a genuine question particle hū, which never cooccurred with wh- words, similar to the situation in Modern Chinese, as illustrated below. (5)




吾誰欺? 欺天乎? (論語 子罕) Wú shuí qī? Qī tiān hū? I who cheat cheat heaven Q “Whom can I cheat? Can I cheat heaven?” (Lun Yu, Zi Han, 500 BC)

In the above example, the first clause with the wh- word shuí was not followed by the question particle hū because the interrogative property of the wh- word makes the question particle unnecessary. However, the second clause, which lacks a wh- word, must rely on the particle hū to form a question. In short, the wh- words in Chinese can be interpreted only as interrogative elements throughout history. More importantly, their quantificational usages were the extensions of their inherent interrogative features, which did not emerge until quite recently, as discussed in the next subsection.

3.2.2

The “Quantificational Movement at LF” Hypothesis

According to Chomsky (1976), there is a level of semantic interpretation called the “logical form.” Based on this proposal, some researchers have considered that in situ wh- words behave like quantifiers that can undergo raising. Aoun et al. (1981) and Huang (1982: 251‒268) first argued that there is a movement of in situ wh- words in the logical form, and parallelism exists between wh- movement and

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

32

wh- in situ, both of which meet the selection requirements. Huang used English as an example to claim that the Chinese wh- word shénme “what” yields the same representation as its English counterpart, as illustrated in (6). (6)

[CP [IP Qiaofeng wen wo [CP shenme [IP Guojing mai-le ti]]]] Qiaofeng ask me what Guojing buy-PERF

Even if there is covert movement of Chinese wh- words, as Huang claimed, wh- words do not necessarily undergo movement as in English, as there are at least two types of overt wh- movement in the languages of the world: (a) the landing site is in the clauseinitial position, as in most Indo-European languages; (b) the landing site is between the subject and the verb, as in Hungarian (Horvath 1986), Hindi (Butt and King 1996), Urdu (Kidwai (2000), and Malayalam (Jayaseelan 1996). As we saw in Chapter 2, whmovements in Old Chinese belong to the second type, which we will discuss in depth in the following sections. The “quantificational hypothesis” is based on the phenomenon that the Chinese whword shénme “what” can be interpreted as a universal reference or an existential quantifier in addition to its interrogative usage. Watanabe (1992) claimed that the potential quantificational interpretations license wh- in situ, and Huang’s (1982: 251‒268) analysis assumed that interrogative wh- phrases are indeed inherent existential quantifiers. However, all the extensions of wh- words, such as indefinite/ existential meanings, universal references, and covariation readings, happened after the tenth century AD (see Lü 1985: 104‒182 for details). Those wh- words in Old and Middle Chinese did not have such quantificational usages. In short, two diachronic changes in history facilitated these extensions: (a) wh- words stayed in situ, and (b) bare nouns in preverbal position were automatically assigned the feature definite, and those in postverbal position were automatically assigned the feature indefinite (for a fuller discussion, see Section 7.6). When interpreted as a universal reference, whwords must occur in preverbal position; when interpreted as an indefinite/existential qualifier, wh- words must occur in postverbal position. That is, these functional extensions of wh- words were actually by-products of the typological change from wh- movement to wh- in situ. The wh- word shuí “who” has had the longest history among all the Chinese whwords. This word was widely used in the Shi Jing, which includes poems composed as early as the eleventh century BC. According to Lü (1985: 104), however, it did not develop out of its universal and existential quantifiers and covariation usage until the sixteenth century AD. Let us consider some of the earliest examples of its quantificational usage.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Licensing of Wh- In Situ

33

(7)

Universal reference 誰也不能佔便宜。 (三俠五義九十一回) Shuí yě bù néng zhàn-piàn-yí. whoever also not can take-advantage “No one can take undue advantage.” (San Xia Wu Yi, Chapter 91, AD 1850)

(8)

Indefinite quantifier 也沒有什麼妨礙著誰去處。 (紅樓夢七十七回) Yě méi shénme fángài-zhe shuí de qùchù. also not somewhat hurt-PROG who de place “Also, it does nothing to hurt anyone.” (Hong Lou Meng, Chapter 77, AD 1750)

(9)

Covariation reading 誰和我好, 我就和誰好。 (紅樓夢二十七回) Shuí hé wǒ hǎo, wǒ jiù hé shuí hǎo. who with I nice I then with who nice “I would be nice to whoever is nice to me.” (Hong Lou Meng, Chapter 27, AD 1750)

In summary, none of the approaches to the properties of Chinese wh- words can pass a simple test based on diachronic empirical data, and they also fail to provide an accurate description of the syntax of Contemporary Chinese. Even within Chinese generative linguistics, there is still no consensus regarding the behaviors of wh- words, and the major hypotheses all need further evidence to be proven or disproven (Lin 2014).

3.2.3

The “Focus Marking” Hypothesis

Working within the framework of generative linguistics, some researchers have noted that wh- movement and wh- in situ are closely related to the marking of focus, a point most relevant to our analysis. Within the generative framework, Albridge (2010) claimed that wh- movement in Late Old Chinese is a clause-internal focus fronting to the edge of vP. This observation is roughly correct, but the phenomenon can be traced back to the earliest Chinese texts around the thirteenth century BC and continued to exist until the sixth century AD. In other words, its existence was far beyond the period of Late Old Chinese, which extended from the third to the first century BC. We do not accept the hypothesis of vP because it is not supported by any empirical evidence.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

34

Cheng and Rooryck (2002) found that focus could also license wh- in situ in the Setubal dialect of European Portuguese. Specifically, “the positions which allow nonfronted wh- elements correlate with positions in which the corresponding non-whelements may be interpreted as having a focus.” In what follows, we will discuss the facts that wh- words themselves are inherently focal and that their distributions are determined largely by the focus constructions of a particular language. As the following analysis shows, in situ wh- words do not have to be licensed by their corresponding nonwh- elements, as Cheng and Rooryck claimed. Moreover, Jayaseelan (1996) pointed out that wh- words are normally an intrinsic focus and thus must undergo movement to Spec-FOC-P for focus checking. For example, in Malayalam, an SOV language of the Dravidian language family, there is a fixed preverbal focus position. When the subject is a wh- element, the word order changes from SOV to OSV, as illustrated in (10). (10)

(a) niṉṉ-e aarǝ aticcu? OSV you-ACC who beat “Who beat you?” (b) *aarǝ aare niṉṉ-e aticcu? SOV (Malayalam)

A similar phenomenon is found in Hungarian, in which the focus position also precedes verbs and wh- words undergo the same movement as focused nouns (Horvath 1986, É Kiss 1987, Lipák 2002). Moreover, there is a strong tendency to keep wh- elements immediately to the left of the verb in many languages, such as other South Asian languages (Jayaseelan 1996), Turkish (Kornfilt 1997: 9‒29), and Hindi/Urdu (Dayal 2014). Jayaseelan attributed this feature to the SOV word order, but our analysis shows that the left periphery of VP for marking focus is cross-linguistically observed, regardless of basic word order. In many languages, the landing site of wh- words is determined by the focus position. In some Bantu languages, for example, nonsubject wh- elements must occur immediately after the verb because that position is the focus position, as Aboh (2006) argued. Lassadi (2003) argued that although both French and Egyptian Arabic are wh- in situ languages, their wh- words can exhibit overt movement when triggered by focus features. In addition, some researchers (Wahba 1991, Cole and Hermon 1998) have classified wh- movement into two types: full movement and partial movement. In their view, whwords in English undergo full movement because they are fronted to sentence-initial position, whereas wh- words in Malayalam and many Bantu languages undergo partial movement because they occur in the left or right periphery. We disagree with this view because the landing sites of wh- words are actually determined by the focus structure of a language. In English, for example, the common device for marking a focused

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus Constructions and Wh- Questions

35

constituent is the construction “it be X that-clause,” where X represents the focused noun phrase (Comrie 1981: 142). According to Heine and Kuteva (2002: 331), a crosslinguistically common grammaticalization path occurs in which a copula, usually together with a sentence-initial pronoun, develops into a focus marker, as in Cora and Haitian. If the phrase “it be” is viewed as a focus marker, we will see a parallel between wh- questions and the focus construction in English, as in (11). (11)

(a) Focused NP + CP: It was a car that John bought yesterday. (b) Wh- word + CP: What did John buy yesterday?

That is, both wh- words and focused constituents occupy the same position, namely sentence-initial position, which is not different from their position in so-called “partial movement” languages, such as Malayalam. In short, the landing site, wh- movement, and wh- in situ (either obligatory or optional) are closely related to the focus marking of a language. This “focus” hypothesis is most closely related to our diachronic exploration, as indicated in the following discussion.

3.3

Focus Constructions and Wh- Questions

Our diachronic investigation is based on the assumption that all wh- words share a uniform semantic feature, namely +FOC, that must be checked by certain grammatical devices. In line with Comrie (1981: 63) and Mycock (2007), we adopt the following view about focus and wh- elements: Wh- elements are the default focus of a sentence. Thus, it is crosslinguistically true that there is a semantic and syntactic parallelism between wh- question structures and focused-constituent constructions. In what follows, we use three pieces of empirical evidence to show the syntactic parallel between wh- questions and focus constructions: typological facts, diachronic data, and synchronic data in Chinese. First, let us begin with typological data. Croft (1996: 168‒170) argued that there is a functional similarity among the following four constructions: (a) the focus of a noun phrase, (b) relativization, (c) wh- questions, and (d) constituent negation. According to Schachter’s (1973) foreground–background analysis, wh- words, focused noun phrases, negated elements, and relativized items are the foregrounded pieces of information, and the rest of the clause is backgrounded information. Additionally, these four types of sentence commonly use the same constructions, typically called “extraction constructions.” For example, in Akan, both the relative clause and the

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

36

focus construction involve positioning the relativized/focused noun phrase before the verb and using a special anaphoric pronoun, ńo. In many Mayan languages, the focus, wh- question, negation, and relative clause constructions all have in common the fronting of the foregrounding noun phrase. In terms of Croft’s analysis (2004: 175), common grammatical structures are motivated by their functional similarity – a kind of iconicity. Second, let us consider the situation in Old Chinese. At this stage, the whquestions and constituent negation used the same construction as the focus construction.5 For the sake of simplicity, we focus on the foregrounded elements that are used as the object of the verb. The extraction constructions have two variations depending on whether the foregrounded element is a simple pronoun or an NP (i.e. containing a content noun). If the foregrounded element is a simple pronoun, the construction is “Subj + Xpro + V”; if it is an NP, the construction is “Subj + XNP + zhī/shì + V,” where an anaphoric zhī or shì must be added.6 Since the subject has always been optional in Chinese throughout history, fronted elements often occur in sentence-initial position. The three types of extraction constructions are illustrated in the following examples. (12)

The wh- question construction (a) 彼且奚適也? (莊子 逍遙遊) Bǐ qiě xī shì yě? he will where go PRT “Where will he go?” (Zhuang Zi, Xiao Yao You, 300 BC)







(b) 王何卿之問也? (孟子 萬章下) Wáng hé-qīng zhī wèn yě? king which-official Pro ask PRT “Which official do you ask about?” (Meng Zi, Wang Zhang Xia, 300 BC)

5

6

In general, Old Chinese lacked a relative clause marker. Xu and Li (1993) may have been the first to note that the wh- movement in Old Chinese relates to focus expression. This idea was adopted by Wei (1999) and Feng (1996), among others. The fronting of a focused element in Old Chinese is sensitive to whether it is a pronominal item; this may be related to cliticization. In French, for example, the pronoun must occur in preverbal position when used as an object (Harris and Campbell 1995: 24): Il vous voit. he you see “He sees you.”

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus Constructions and Wh- Questions

37

In (12b), hé-qīng is pied-piping wh- movement; in this case, the anaphoric zhī is usually used between the fronted wh- phrase and the verb. (13)

The constituent negation construction (a) 余不汝忍殺。 (左傳 昭公七年) Yú bù rǔ rěn shā. I not you have-heart kill “I don’t have the heart to kill you.” (Zuo Zhuan, Zhao Gong Qi Nian, 550–400 BC)







(b) 吾未之見也。 (論語 里仁) Wú wèi zhī jiàn yě. I not it see PRT “I have not seen it.” (Lun Yu, Li Ren, 500 BC) (14)

The focus construction (a) 唯婦言是用。 (尚書 牧誓) Wéi fù yán shì yòng. only woman advice this take “It is the advice of women that he always takes.” (Shang Shu, Mu Shi, 1100–800 BC)







(b) 余惟利是視。 (左傳 成公十三年) Yú wéi lì shì shì. I only profit this seek “What I seek is profit.” (Zuo Zhuan, Cheng Gong Shi San Nian, 550–400 BC) If focused elements are ordinary nouns, the emphasis marker wéi “only” is usually used in the initial position of the focus construction. In this case, the anaphoric pronoun shì occurs between the focused noun and the verb. Note that in Old Chinese, wh- words had to undergo overt movement, but only 40 percent of pronouns in the constituent negation construction were fronted to the left periphery of VP (He 2004: 45). For content nouns, the extraction construction was used only when they needed to be focused. These facts show that wh- words inherently contain +FOC, but other pronominal words and ordinary nouns do not have an inherent +FOC, and their focus features are assigned by some pragmatic factors.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

38

Due to a typological change, Contemporary Chinese employs the light verb shì to mark the focus feature of wh- words,7 which was a further grammaticalization from its copular usage (for details, see Chapter 2). Shì can be added to any nominal element or preposition phrase (containing a noun) in preverbal position to make it focused, including a subject, time words, and other adverbial preposition phrases, as illustrated in (16)–(19). The following example is a normal declarative sentence in which no constituent is focused. (15)

湯姆昨天在家裡修好了一輛自行車。 (現代漢語) Tāngmǔ zuótiān zài jiālǐ xiū-hǎo le yī-liàng zìxíngchē. Tom yesterday at home fix-well PERF one-CL bicycle “Tom fixed the bicycle at home yesterday.” (Contemporary Chinese)

In the above sentence, four phrases have the potential to be focused by adding the marker shì immediately in front of them: Tāngmǔ “Tom,” zuótiān “yesterday,” zài jiālǐ “at home,” and yī-liàng zìxíngchē “a bicycle.” In general, each sentence can have only one focused element, as illustrated in the following examples.

7

(16)

是湯姆昨天在家裡修好了一輛自行車。 (現代漢語) Shì Tāngmǔ zuótiān zài jiālǐ xiū-hǎo le yī-liàng zìxíngchē. FOC Tom yesterday at home fix-well PERF one-CL bicycle “It was Tom who fixed a bicycle at home yesterday.” (Contemporary Chinese)

(17)

湯姆是昨天在家裡修好了一輛自行車。 (現代漢語) Tāngmǔ shì zuótiān zài jiālǐ xiū-hǎo le yī-liàng zìxíngchē. Tom FOC yesterday at home fix-well PERF one-CL bicycle “It was yesterday that Tom fixed a bicycle at home.” (Contemporary Chinese)

(18)

湯姆昨天是在家裡修好了一輛自行車。 (現代漢語) Tāngmǔ zuótiān shì zài jiālǐ xiū-hǎo le yī-liàng zìxíngchē. Tom yesterday FOC at home fix-well PERF one-CL bicycle “It was at home that Tom fixed a bicycle yesterday.” (Contemporary Chinese)

Both the copula shì and the focus shì are light verbs. They still reserve the property of verbs and are involved in building syntactic configurations. In certain cases, it is difficult to distinguish the copula shì from the focus shì because the form may have dual functions. For the sake of simplicity, we use the term “focus marker” to label it.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

39

Focus Constructions and Wh- Questions (19)

湯姆昨天在家裡修好的是一輛自行車。 (現代漢語) Tāngmǔ zuótiān zài jiālǐ xiū-hǎo de shì yī-liàng zìxíngchē.8 Tom yesterday at home fix-well REL FOC one-CL bicycle “It was a bicycle that Tom fixed at home yesterday.” (Contemporary Chinese)

To make the object focused, the VP prior to the object must be converted to a nominal by adding the relativizer de, and the focus marker shì occupies the verb position as in “(VP + REL) + shì + object,” as illustrated in (19). However, once any of the four phrases is replaced by a wh- phrase, only this whphrase can be marked by the focus marker shì in Contemporary Chinese. That is, adding the focus marker to any other non-wh- phrases will produce an ill-formed structure. When the time word zuótiān is replaced by shénme shíhòu “when,” for instance, only this wh- word is qualified to be marked by the focus marker shì, as illustrated in (20). (20)

(a) 湯姆什麼時候在家裡修好了一輛自行車? (現代漢語) Tāngmǔ shénme-shíhòu zài jiālǐ xiū-hǎo le yī-liàng Tom when at home fix-well PERF one-CL zìxíngchē? bicycle “When did Tom fix a bicycle at home?” (b) 湯姆是什麼時候在家裡修好了一輛自行車? Tāngmǔ shì shénme-shíhòu zài jiālǐ xiū-hǎo Tom FOC when at home fix-well yī-liàng zìxíngchē? one-CL bicycle “When did Tom fix a bicycle at home?”

le PERF

(c) *是湯姆什麼時候在家裡修好了一輛自行車? *Shì Tāngmǔ shénme-shíhòu zài jiālǐ xiū-hǎo FOC Tom when at home fix-well yī-liàng zìxíngchē? one-CL bicycle

8

le PERF

In this sentence, the focus marker shì can also be seen as the copula (see Zhu 1982: 105 for details).

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

40

(d) *湯姆什麼時候是在家裡修好了一輛自行車? *Tāngmǔ shénme-shíhòu shì zài jiālǐ xiū-hǎo Tom when FOC at home fix-well yī-liàng zìxíngchē? one-CL bicycle

le PERF

(e) *湯姆什麼時候在家裡修好的是一輛自行車? *Tāngmǔ shénme-shíhòu zài jiālǐ xiū-hǎo de shì Tom when at home fix-well REL FOC yī-liàng zìxíngchē? one-CL bicycle (Contemporary Chinese) Although the wh- phrase is not obligatorily marked by the focus shì, its presence means that other non-wh- elements cannot be focused. Therefore it is evident that wh- phrases are the natural and default focus of a sentence and that they have priority for taking the focus marker. This observation explains why wh- words in Chinese and many other languages have been fused with a focus marker into a single lexical item.

3.4

Topicalization and the Landing Site of Wh- Words

In this section, we discuss the landing site of wh- words in Old Chinese, the interaction between wh- movement and topicalization structure, and the movement of wh- words in preposition phrases. Although they all occur before the verb, the specific landing sites of wh- arguments and wh- adjuncts are not exactly the same. More interestingly, the movement of wh- words in preposition phrases seems to involve two operations: the inversion between the wh- word and the preposition and the fronting of the whole preposition phrase prior to the verb, a phenomenon that has not been found in other languages.

3.4.1

The Landing Site of Wh- Word Arguments

As mentioned earlier, wh- phrases in Old Chinese obligatorily underwent overt movement (Wang 1989: 76‒85, He 2004: 265‒304, Ohta 1987: 125‒126). Thus far, there are two types of movement: English-type languages, in which wh- words must be fronted to sentence-initial position, and Malayalam-type languages, in which wh- words must occur immediately prior to the verb. Old Chinese may contribute a third type: the linear landing site of wh- words can be described as follows.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Topicalization and the Landing Site of Wh- Words (21)

41

S + temporal/negatives + (wh- words) + AUX/PREP + V?

That is, wh- words must occur prior to auxiliary verbs or prepositions (if any) and after temporal words or negative markers (if any). Thus wh- words occur in the leftmost periphery of the VP but below the IP. Since the subject, temporal adverbs, and negative markers (and so on) are often absent, wh- words might occur in sentence-initial position. The following are examples of wh- movements in Old Chinese.




(22)

臣實不才, 又誰敢怨? (左傳 成公三年) Chén shí bù cái, yòu shuí gǎn yuan? I really not able again who dare complain “I am not talented. Who am I to complain?” (Zuo Zhuan, Cheng Gong San Nian, 550–400 BC)

(23)

然則又何以兵為? (荀子 議兵) Ránzé yòu hé yǐ bīng wéi? however again what with army do “If this is the case, then why do we have armies?”




(Xun Zi, Yi Bing, 250 BC) (24)




彼且奚適也? (莊子 逍遙遊) Bǐ qiě xī shì yě? he will where go PRT “Where will he go?” (Zhuang Zi, Xiao Yao You, 300 BC)

Theoretically, there are many possible patterns of multiple wh- questions. In fact, only one type of wh- question is attested in the history of Chinese. In the co-ordinate construction consisting of two VPs, each matrix verb can host a fronted wh- word, as illustrated in (25). (25)




夫何憂何懼? (論語 顏淵) Fú hé yōu hé jù? then what worry what fear “What do I worry about and fear?” (Lun Yu, Yan Yuan, 500 BC)

Indeed, within a co-ordinate verb construction, one matrix verb can host a focused NP, and another can host a wh- word, as illustrated in (26).

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

42 (26)




非夫人之為慟而誰為? (論語 先進) Fēi fū rén zhī wéi tòng ér shuí wéi? not the person zhi for grieve and who for? “For whom should I be grieved if I am not for this person?” (Lun Yu, Xian Jin, 500 BC)

Jayaseelan (1996) pointed out that wh- words are normally intrinsically focal and thus must undergo movement to Spec-FOC-P for focus checking. As mentioned earlier, in Malayalam, an SOV language, the landing site of wh- words is always the position immediately to the left of the verb; hence a wh- word subject must move to follow the object, resulting in the O-Swh- V word order. Since Chinese has been an SVO language throughout its history, subject wh- words stay in sentence-initial position. They do not need to move to follow temporal words if any are present, which can be formulated as follows: (27)

Wh- word + temporal/negatives + AUX/PP + VP?

Superficially, subject and object wh- words occur before the verb but do not actually occupy the same exact position. When used as the subject, wh- words occupy the leftmost periphery of the IP, namely Spec-IP. (28)




誰敢不讓? (尚書 虞書) Shuí gǎn bù rang? who dare not yield “Who dares not yield his position?” (Shang Shu, Yu Shu, 1100–800 BC)

In summary, the landing site of wh- words used as the object is after the “Subj + temporal” part and before the “AUX./PP + V” part, rather than the position immediately adjacent to the verb. Additionally, wh- word subjects do not occupy exactly the same position as wh- objects.

3.4.2

The Extraction of Wh- Words

There are two types of wh- adjunct: a single wh- word and “preposition + wh- word.” When it is a single wh- word, the landing site is after the subject and before the leftmost periphery of the IP, which can be formulated as follows. (29)

S + wh- adjunct + VP?

(30)

子奚不為政? (論語 為政) Zǐ xī bù wéi zhèng? you why not do politics “Why don’t you work for the government?” (Lun Yu, Wei Zheng, 500 BC)




https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Topicalization and the Landing Site of Wh- Words

43

However, if the wh- adjunct is a preposition phrase, an inversion between the preposition and wh- word must first occur because of the relationship between verbs and prepositions. Almost all Chinese prepositions grammaticalized from ordinary verbs; however, many verbs could be used as prepositions in the same period. More interestingly, many preposition phrases in Old Chinese usually occurred in sentence-final position, whereas the whole preposition phrase needed to be fronted to the left of the VP if it contained a wh- word. This point is illustrated by the following pair of examples with the same matrix verbs and the same prepositions found in the same texts. (31)




何以贈之? (詩經 渭陽) Hé yǐ zèng zhī? what with present she “What do I present her with?” (Shi Jing, Wei Yang, 1000–600 BC)

(32)




贈之以芍藥。(詩經 溱洧) Zèng zhī yǐ sháoyào. present she with Chinese-peony “I present her with Chinese peonies.” (Shi Jing, Qin Wei, 1000–600 BC)

In (31), hé yǐ “what + with” must be fronted to the position before the matrix verb zèng “present.” In contrast, in the declarative sentence with the same matrix verb, the preposition phrase occurs in sentence-final position (cf. Section 2.5). The same movement operation can be found in the following example. (33)




天下惡乎定? 定於一。 (孟子 梁惠王上) Tiānxià wū hū ding? Dìng yú yī. world what in settle settle in one “To what can the world be settled? The world can be settled in one.” (Meng Zi, Liang Hui Wang Shang, 300 BC)

The preposition hū in the first clause is synonymous with the preposition yú, and they are exchangeable. Once again, the wh- PP involves two operations: (a) inversion between the wh- word and the preposition and (b) the fronting of the whole preposition phrase. In summary, the whole wh- PP must undergo overt movement to the left of the VP, and the wh- word must also undergo overt movement to the left of the preposition. In the corresponding declarative sentence in (33), the preposition phrase occurs in sentence-final position, and the noun follows the preposition. To my knowledge, this kind of double whmovement has never been reported in the literature. A detailed theoretical exploration is beyond the scope of this book, and we would like to leave this topic for future research.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

44 3.4.3

Topicalization and Wh- Movement

Why the landing site of wh- words is the sentence-initial position in some languages but between the subject and verb in other languages is very puzzling. A possible answer can be found in diachronic evidence from Chinese. We observe that the landing site of whwords in Old Chinese should result from the interaction between topicalization and the focus construction (including wh- questions). There is a functional contrast between the topic and focus: the topic expresses a piece of given information, whereas the focus refers to a piece of new information. Thus the topic and focus are usually marked in contrastive grammatical devices in a particular language. Chinese has been thought of as a topic-prominent language, in contrast to English, which is considered a subjectprominent language (Li and Thompson 1976). Comparing a corpus of spoken Beijing Mandarin and an English speech corpus, Liu (2009: 108‒109) found that the frequency of topic structures is fifty-four times more frequent in Chinese than in English. Old Chinese was also a topic-prominent language, although its topicalization was more constrained than that in Contemporary Chinese. Sentence-initial position in Old Chinese was reserved for the topic, and the trace of the topicalized element was usually overtly indicated by the anaphoric pronoun zhī, which can be schematized as follows. (34)

The topicalization construction in Old Chinese. TOP, Subj + V + ANAPtopic.

Now let us consider some of the earliest examples of topicalization.




(35) 《詩》三百, 一言以蔽之。(論語 子張) Shī sān bǎi, yī yán yǐ bì zhī. poem three hundred one sentence with summarize it “For the 300 poems, one sentence can be used to summarize them.” (Lun Yu, Zi Zhang, 500 BC) (36)




殷禮, 吾能言之。 (論語 八佾) Yīn lǐ wú néng yán zhī. Yin rite I can talk it “The rite of the Yin dynasty I can talk about.” (Lun Yu, Ba Yi, 500 BC)

Old Chinese employed two sets of contrastive devices to distinguish the topic from the focus: (a) the landing site of the topic was sentence-initial position, and the focus was somewhere between the subject and the matrix verb; (b) an anaphoric pronoun was typically used in the original position of the topicalized element but could not occur in the original position of either wh- words or other focused elements. As noted in the

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Topicalization and the Landing Site of Wh- Words

45

previous examples, when the focused element was a noun phrase, the anaphoric pronoun zhī or shì usually occurred before the matrix verb. The topicalization and focus construction can be summarized as follows: (37)

(a) The topicalization construction: TOP + Subj + V + ANAP.9 (b) The focus construction: Subj + FOC + ANAP + V.

In the early stage of Old Chinese, a wh- question was formed by means of two operations: (a) the fronting of the wh- word to the left of the verb and (b) the addition of an anaphor between the wh- word and the verb, as illustrated in (38). (38)




云誰之思? 西方美人。(詩經 簡兮) Yún shuí zhī sī? Xīfāng měirén. say who her miss west beautiful-girl “Whom do you miss?” “A beautiful girl in the West.” (Shi Jing, Jian Xi, 1000–600 BC)

The anaphoric zhī is crucial in avoiding the ambiguity of (38): shuí “who” can be interpreted only as the object of the verb sī “miss.” Otherwise, it can be understood as either the agent or the patient. Logically, if the only element prior to the verb was a wh- word in Old Chinese, there were always two possible interpretations: the wh- word was either the subject or the object. Old Chinese removed this ambiguity by means of the two contrastive structures summarized in (37), as illustrated in the following two examples. (39)

9




然則孰立之? 石碏立之。 (公羊傳 隱公四年) Ránzé shú lì zhī? Shí Què lì zhī. but who establish he? Shi Que establish he “However, who made him a king?” “Shi Que made him a king.” (Gong Yang Zhuan, Yin Gong Si Nian, 400 BC)

This topicalization construction was abandoned at the beginning of Medieval Chinese, as the whmovement phenomenon disappeared. These two diachronic changes should be directly related. In Contemporary Mandarin Chinese, adding an anaphor to an instance of topicalization will produce an ill-formed structure, as illustrated in the following. *書我已經看了它。 (現代漢語) *Shū wǒ yǐjīng kàn le tā. book I already read PERF it Furthermore, there is a debate about whether topicalization is base-generated or undergoes movement (see Xu 2015 for details). The empirical evidence from Old Chinese reveals that topicalization at that time more likely involved movement.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

46 (40)




孰繼? 繼子般也。 (公羊傳 閔公元年) Shú jì? Jì Zǐ Bān yě. who succeed succeed Zi Ban ye “Whom do I succeed?” “You succeed Zi Ban.” (Gong Yang Zhuan, Min Gong Yuan Nian, 400 BC)

In (39), the wh- word shú “who” can be interpreted only as the subject of the first clause because the anaphoric zhī follows the verb lì “establish.” However, the same wh- word shú “who” in (40) must be interpreted as the object because there is no anaphor in the first clause. In Old Chinese grammar, no anaphor was allowed when the wh- word was fronted to the left of the verb. In this case, if the verb was transitive, the wh- word prior to the verb had to be the object of the verb. In summary, the specific positions of wh- words in Old Chinese were actually different, although they all occurred prior to the verb. (41)

(a) wh- word objects: Subj + temporal/quantifier + wh- + AUX/PREP + V. (b) wh- word subjects: wh- + temporal/quantifier + AUX/PREP + V. (c) pure wh- word adjuncts: Subj. + wh- + temporal/quantifier + AUX/ PREP + V. (d) wh- word adjuncts in PPs: Subj. + temporal/quantifier/AUX + wh- + PREP + V.

The positioning of wh- words was constrained by the topicalization construction. According to Simpson and Bhattacharya (2003), the landing site of wh- words is also between the subject and the verb in Bangla, a language in which the subject can mostly be interpreted as a topic. That is, the wh- movement of Bangla is also conditioned by its topicalization construction. Some researchers (e.g. Fanselow 2017) have claimed that in SOV languages, the position to the immediate left of the verb is retained for a focused element, and German is characterized as such a language. However, the specific landing site of wh- words in a particular language is determined by the focus construction in general, and the configuration of the focus construction may be conditioned by the topicalization construction.

3.5

The Emergence of the Copula Shì

As analyzed in Chapter 2, the emergence of the copula shì in Late Old Chinese fundamentally changed the texture of the grammar at that time. Additionally, this factor caused the typological change in Chinese from wh- movement to wh- in situ, which had the following effects. First, even in Old Chinese, wh- movement was sensitive to light

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Emergence of the Copula Shì

47

verbs; unlike ordinary transitive verbs, light verbs did not trigger wh- movement. The emergence of the copula shì, an extremely high-frequency verb, created a considerable number of “V + wh- word” instances so that other nonlight ordinary verbs by analogy appeared in the same constructions. Second, and most importantly, the copula shì underwent further grammaticalization into a focus marker, which became the most common device for marking a focused constituent. Eventually, the “vf + wh- word” structure entirely replaced wh- movement – a word order device for marking focus.

3.5.1

“Subj + Complement” Construction

As we knew, Old Chinese lacked a copular verb to link the subject and its nominal complement. In general, the sentence-final particle yě needed to be used, and the particle zhě was occasionally added to the subject (for details, see Section 2.2). This construction can be formulated as follows: (42)

Subj (zhě) + NP yě.

The following are examples of the “Subj + Complement” construction in Old Chinese:




(43)

王, 人君也。 (戰國策 齊策) Wáng, rén jūn yě. king people administrator YE “The king is the administrator of the people.” (Zhan Guo Ce, Qi Ce, 450–200 BC)

(44)

南冥者, 天池也。 (莊子 逍遙遊) Nán míng zhě, tiān chí yě. south sea ZHE heaven lake YE “The south sea is the lake of heaven.” (Zhuang Zi, Xiao Yao You, 300 BC)




Wh- phrases immediately follow the subject when used as the subject complement, and they are marked by the sentence-final particle yě. Thus the particle zhě is generally absent, as illustrated in (45) and (46): (45)




此誰也? (戰國策 齊策) Cǐ shuí yě? this who YE “Who is this?” (Zhan Guo Ce, Qi Ce, 450–200 BC)

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

48 (46)




其故何也? (周易 繫辭) Qí gù hé yě? its cause what YE “What is its cause?” (Zhou Yi, Xi Ci, 800 BC)

Since there was no verb-like copula in Old Chinese, wh- words, which were usually marked by the sentence-final particle yě, immediately followed the subject when used as the complement.

3.5.2

Light Verbs and Wh- Movement

The term “light verb” is an important concept in the minimalist program of Chomsky (1995). The term is used to refer to quite different things in general generative linguistics (Grimshaw and Mester 1988, Larson 1988, Hale and Keyser 1993) and in Chinese general linguistics (Huang 1994, 1997, Huang et al. 2009, Lin 2001, 2014). For the purpose of diachronic description, we adopt the following definition from Hale and Keyer (1993): “A light verb has semantic contents and denotes an elementary meaning e (i.e., eventive), which contributes to the building of the event structure of the sentence in a configurational way.” In addition, light verbs are a matter of degree. Among the widely accepted light verbs, such as do, have, become, and be (Huang et al. 2009), the copula be may be the lightest because it has little independent meaning and does not physically affect the complement. Compare the following transformations. (47)

(a) (b) (c) (d)

He did his homework. He had an airplane. He became a doctor. He was a nurse.

What did he do? What did he have? What did he become? *What was he be? (What was he?)

As mentioned earlier, light verbs often licensed wh- in situ even in Old Chinese, in which wh- words obligatorily underwent movement. Specifically, the transitivity of verbs played a role in creating light verbs. If the verb was a regular verb, the wh- word had to undergo overt movement. However, for verbs such as zhī “go,” the wh- word could stay in situ, as illustrated in (48) and (49). (48)




寡人將去斯而之何? (列子 立命) Guǎrén jiāng qù sī ér zhī hé? I will leave here and go where? “Where shall I go after leaving here?” (Lie Zi, Li Ming, 400 BC)

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Emergence of the Copula Shì (49)

49




其言歸何? (公羊傳 隱公元年) Qí yán guī hé. he say return where “He said, ‘where did you return?’” (Gong Yang Zhuan, Yin Gong Yuan Nian, 400 BC)

According to our comprehensive investigation of the data in Old Chinese, there are a total of three light verbs that, without exception, licensed a wh- in situ: yún “say,” nài “do,” and wéi “be” (cf. Section 2.5). (50)




子夏云何? (論語 微子) Zǐ Xià yún hé? Zi Xia say what “What did Zi Xia say?” (Lun Yu, Wei Zi, 500 BC)

(51)




子為誰? (論語 微子) Zǐ wéi shuí? you be who “Who are you?” (Lun Yu, Wei Zi, 500 BC)

(52)




事將柰何矣? (戰國策 趙策) Shì jiāng nài hé yǐ? matter will do what PRT “What shall we do about the matter?” (Zhan Guo Ce, Zhao Ce, 450–200 BC)

The verb wéi in (51) was a polysemous word in Old Chinese that could mean “do,” “make,” “become,” “call,” etc. Occasionally, this verb could function as a linking verb to connect the subject and its complement, as in (51). In this case, the verb could be viewed as a semi-copula. When used as a semicopula, wéi always preceded the whcomplement.

3.5.3

The Focus Marker Shì

The demonstrative shì was widely attested in the earliest texts, such as the Shi Jing, which reflected the grammar during the period from the eleventh to the seventh century BC. It was further grammaticalized into a copular verb around the first century BC (for details, see Chapter 2). The example shown in (53) is one of the earliest appearances of the copula shì:

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

50 (53)




其是吾弟歟? (史記 呂不韋列傳) Qí shì wú dì yǔ? he be my younger-brother QUE “Is he my younger brother?” (Shi Ji, Lü Bu Wei Lie Zhuan, 100 BC)

From the beginning, the copula shì always preceded a wh- complement, even when wh- movement was still a rigorous rule during the transitional period from Old to Medieval Chinese. Historically, only the “Subj + shì + wh- word” construction was attested, and no “Subj. + wh- word + shì” ordering was found, as illustrated in (54) and (55): (54)




此是何水? (世說新語 言語) Cǐ shì hé shuǐ? this be which river “Which river is it?” (Shi Shuo Xin Yu, Yan Yu, AD 450)

(55)




第一流復是誰? (世說新語 品藻) Dìyī liú fù shì shuí? first class else be who “Who else is in the first class?” (Shi Shuo Xin Yu, Pin Zao, AD 450)

When the copula shì became obligatory to link the subject and the complement, the sentence-final particle yě was ultimately abandoned. Since expressions of the copular construction were extremely common in daily communication, both in ancient times and in the present, instances of “copula + wh- word” became increasingly popular. This phenomenon had the potential to challenge the old principle of wh- movement in Old Chinese. As mentioned earlier, it is cross-linguistically common for a copula to develop into a focus marker. Likewise, the copula shì also developed into a focus marker, a usage widely attested in the texts edited around the fifth century AD. Let us consider some of the earliest examples of the focus marker shì: (56)




云是卿為其計。 (世說新語 言語) Yún shì qīng wéi qí jì. say FOC you design his plot “As said, it was you who designed his plot.” (Shi Shuo Xin Yu, Yan Yu, AD 450)

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Emergence of the Copula Shì (57)

51




子敬可是先輩誰比? (世說新語 品藻) Zǐ Jìng kě shì xiānbèi shuí bǐ? Zi Jing can FOC older-generation who compare “Whom in the older generation can Zi Jing compare with?” (Shi Shuo Xin Yu, Pin Zao, AD 450)

During the course of grammaticalization, a lexical item usually undergoes semantic bleaching (Hopper and Traugott 2003: 96). Considering that the focus shì grammaticalized from its copular usage, the focus marker shì can be assumed to be even “lighter” than the copula shì. The further grammaticalization of the copula shì into a focus marker eventually caused the disappearance of the three old focus constructions: (a) wh- questions, (b) constituent negation, and (c) focused nouns, all of which involved the same device of word order change. These disappearances happened during the same period, from the first century BC to the fifth century AD. Around the eighth century AD, the demonstrative dǐ grammaticalized into the relativizer de (the phonologically reduced form),10 which combines with the copula/ focus marker shì to form various focus constructions (Shi 2016: 38‒55, Zhu 1982: 105). (58)

我問的是你。 (現代漢語) Wǒ wèn de shì nǐ. I ask REL FOC you “It was you whom I asked for.” (Contemporary Chinese)

(59)

我最後一次見到他是在上海。 (現代漢語八百詞) Wǒ zuìhòu yī-cì jiàndào tā shì zài Shànghǎi. I last one-CL meet he FOC in Shanghai “It was in Shanghai that I met him the last time.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 496)

In Modern Chinese, the focus marker shì is combined with the relativizer de to form various focus constructions.

3.5.4

The Compounding of Wh- Words and the Focus Marker Shì

As stated earlier, because wh- words are the natural focus of a sentence, in many languages they are obligatorily marked by a focus marker. Due to the high

10

This grammatical morpheme has multiple functions, such as serving as a genitive marker, a nominalizer, and an associative marker.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

52

frequency of their co-occurrences, wh- words tend to form compound words with a focus marker. As a result, wh- phrases in many languages contain a focus morpheme. For example, in some Creole languages, such as Papiamentu and Saramaccan, wh- words are affixed by a focus marker (Holm 1988: 188). It is a widely attested grammaticalization path that a copula in a focus cleft clause becomes grammaticalized into a focus marker, and that, simultaneously, wh- words are obligatorily marked for focus (Heine and Reh 1984: 109). A similar phenomenon happened in the history of Chinese, as discussed below. While the copula shì developed into a focus marker and wh- movement was gradually replaced by wh- in situ in the first half of the Medieval Chinese period, the system of wh- words also underwent a fundamental change. Except for the whword shuí “who,” all wh- words that had been used from the very beginning of Old Chinese either disappeared or were replaced by new wh- forms. These replacements of wh- words were closely related to the emergence of the focus marker shì. According to Lü (1985: 104), the wh- word shuí and the focus shì formed a compound-like word and often occurred in the subject position, as illustrated in (60). This compound wh- phrase could be used in the “Subj + Complement” construction to which a semicopula wéi was added, as illustrated in (61):




(60)

是誰教汝? (北齊書 列传卷三十一) Shì shuí jiāo rǔ? be who teach you “Who taught you?” (Bei Qi Shu, Lie Zhuang, Chapter 31, AD 700)

(61)

師為是誰? (出曜經) Shī wéi shì shuí? master be be who “Who is your master?” (Chu Yao Jing, Chu Yao Jing, AD 400)

The above phenomenon is observed in some Contemporary Chinese dialects. According to Huang (1963), for example, in a Southern Min dialect, the whword tsui must be marked by the focus marker tsi. This compound whphrase can be used as an object of transitive verbs, and in the “Subj + Complement” construction the copula tsi must be added, as illustrated in the following:

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Emergence of the Copula Shì (62)

53

Hɪk-pun tsʰeʔ tshe tsi tsui tsioh.11 this-CL book seek FOC who borrow “Whom should I seek to borrow the book from?” (A Southern Min dialect)

(63)

Tsi-ui tsi tsi tsui? this-CL be FOC who “Who is this?” (A Southern Min dialect)

In this Southern Min dialect, the focus marker tsi and the wh- word tsui have firmly formed a compound wh- word that can be used as a single word occurring in the object position. Another major wh- word, he “what,” was replaced by shénme around the sixth century AD. Lü (1985: 122) proposed that the form shénme was the outcome of the fusion of the phrase shì hé “be + what.” This combination could often be interpreted as a single wh- word “what” in the texts of Medieval Chinese, as illustrated in (64): (64)




十六篇中, 是何篇是者? (論衡 謝短) Shíliù piān zhōng, shì hé piān shì zhě? sixteen chapter in FOC which chapter be ZHE “Among the sixteen chapters, which is it?” (Lun Heng, Xie Duan, AD 100)

Therefore the two major wh- words in Modern Chinese, namely, shuí “who” and shénme “what,” have actually incorporated the focus morpheme shì, as reflected by the same initial [ʂ]. Superficially, only the wh- word shuí “who” in Old Chinese survived in Medieval Chinese and has been used to the present day. In fact, this word also contains the focus morpheme shì, simply because they happened to share the same initial sound. Other major wh- expressions include a phrase consisting of shénme “what,” e.g. shénme shíhòu “what time” (equal to when), and wèi shénme “for what” (equal to why). In addition, the wh- expression “how” in Contemporary Chinese is zěnme [ʐǝnmǝ], whose initial is similar to that of the copula shì [ʂʅ]. Table 3.1 contains the phonetic forms of the major wh- phrases in Old and Contemporary Chinese. Due to the fusion of wh- words and the focus marker shì, most wh- words in Modern Chinese share the onset sh- ([ʂ]), which is the same as or similar to the onset of the focus marker shì, as shown in (65).12 11 12

The tones are omitted since they are irrelevant here. This list is based on The Applied Chinese Dictionary (Beijing, The Commercial Press, 2000). Some of these words may be used more often in other dialects. Moreover, some are compound words, e.g. shénme shíhòu “when” = what time, shénme dìfāng “where” = what place, and wéi

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

54

Table 3.1 Phonetic forms of wh- words in Old and Contemporary Chinese

Old Chinese Contemporary Chinese

who

what

when

why

how

ʑwi ʂuei

ɣa ʂǝnmǝ

ɣat ʂǝnmǝ ʂʅxou

ɣa uei ʂǝnmǝ

an tsǝnmǝ

Note: The phonological reconstruction of Old Chinese is based on Li and Zhou (1999).

(65)

誰 什麼 啥 為什麼

shuí “who” shénme “what” shá “what” wéi shénme “why”

咋 咋樣 怎麼樣 什麼地方

zǎ “how” zǎyàng “how” zěnme yang “how” shénme dìfāng “where”

Like wh- words in English, therefore, question phrases in Chinese can be called sh-words. In Medieval Chinese, other common wh- words also emerged that were not phonetically related to the focus morpheme shì. They needed to stay in situ, indicating that the typological change from wh- movement to wh- in situ had been completed. Let us consider some of the earliest examples (Lü 1985: 122):




(66)

持底報郎恩? (萧衍 歡聞歌) Chí dǐ bào láng ēn? take what repay lover kindness “What can I take to repay the kindness of my lover?” (Huan Wen Ge, Huan Wen Ge, AD 600)

(67)

那個瓶老實? (寒山詩) Nǎ-gè píng lǎoshí? which-CL bottle strong “Which bottle is stronger?” (Han Shan Shi, Han Shan Shi, AD 650)

When Chinese underwent the typological change from wh- movement to wh- in situ, the system of wh- words was also replaced by newly emerging wh- elements in the second half of the Medieval Chinese period (from the sixth to the tenth century AD). These new wh- elements actually incorporated the focus morpheme shì, a common phenomenon in many languages that is motivated by the inherent [+FOC] of wh- words.

shénme “why” = for what. There are also two common wh- words that have different initials: nǎgè “which” and nǎlǐ “where.”

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

The Emergence of the Copula Shì 3.5.5

55

From Wh- Movement to Wh- in Situ

The process of replacing wh- movement with wh- in situ took approximately six centuries. This process started as early as the first century BC and finished around the sixth century AD. The emergence of the copula shì and its further grammaticalization into a focus marker were responsible for this typological change. Comrie (1981: 64‒65) indicated that in English, the focus feature of wh- words must be checked in sentenceinitial position by means of a word order device. However, he explained that it is “by no means necessary for a language to exhibit this relevance of focus, for instance, in Mandarin Chinese.” In contrast, my analysis has shown that in Chinese, as in English and many other languages, the focus feature of wh- words must be checked: it is the word order shift in Old Chinese and the addition of a focus marker in Modern Chinese. That is, the difference between English and Chinese concerns the methods of how to check the inherent [+FOC] of wh- words, and, in both, certain devices must be employed to do so. In Contemporary Chinese, every wh- phrase is actually a mini-clause that consists of a light verb (i.e. the focus marker shì) and a wh- word, which can be formulated as follows: (68)

wh- phrase → [vfoc + wh- word]

The above structure is by no means an underlying form but can always be phonetically spelled out. If appearing in the subject position or before the verb, wh- words can be focused by adding the focus marker shì, as the previous examples showed. As Hale and Keyser (1993) noted, light verbs are involved in making syntactic configurations. The positioning of the focus marker shì is subject to the constraint of other grammatical constructions. This light verb shì is not allowed to occur between a verb/preposition and its object. If the PP occurs in preverbal position, the focus marker must be raised to the left of the whole PP; if the focused constituent is an object noun, the rest of the sentence must be turned into a cleft clause, and the focus shì is then immediately added to the object. (69)

(a) Subj + [vfoc [PREP. + wh- word/focused N]] + VP (b) [Subj + V + de] + vfoc + Obj

However, if the PP is in sentence-final position, it must be fronted to the left of the verb and preceded by the focus marker shì, as shown in the following examples. As mentioned earlier, all wh- words have an inherent [+FOC] feature and are the natural focus of a sentence in which they are used. In addition, the major wh- words in Modern Chinese have incorporated the focus marker shì. Even so, the focus features of wh- words can be strengthened by adding the focus marker shì. The relativizer and

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

56

nominalizer de is often combined with the focus marker shì to form various focus constructions. (70)

他是在哪裡買的房子? (北京話相聲) Tā shì zài nǎlǐ mǎi de fángzi? he FOC at where buy DE house “Where did he buy a house?” (Present-day Pekingese Cross-talks)

(71)

你們最缺的是什麼? (北京話相聲) Nǐmen zuì quē de shì shénme? you most need DE FOC what “What do you need most?” (Present-day Pekingese Cross-talks)

(72)

是誰告訴你的。 (現代漢語八百詞) Shì shuí gàosù nǐ de? F who tell you DE “Who told you?” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 496)

If the wh- word is used as a subject and occurs in sentence-initial position, the focus marker shì can be preceded by a demonstrative, as illustrated in (73). This is similar to the focus constructions in English, French, and Lamang (for details, see Heine and Kuteva 2002: 111). (73)

那是誰來了? (現代漢語八百詞) Nà shì shuí lái le? that FOC who come PERF “Who is coming?” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 496)

The development process of the typological change of Chinese from a wh- movement language to a wh- in situ language can be summarized as follows. Stage 1 In Old Chinese, the object wh- elements of verbs and prepositions obligatorily underwent movement to the left periphery of predicates. Then, wh- questions and focus construction were semantically and syntactically parallel. However, the wh- movement was sensitive to the transitivity of verbs; thus certain light verbs with the least transitivity, such as the semicopula wéi, never triggered wh- movement.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Wh- Words in Preposition Phrases

57

Stage 2 In the beginning of the Medieval Chinese period (approximately the first century BC), the typical copula shì grammaticalized from its original demonstrative usage, and it never triggered wh- movement. The high frequency of this copula created a considerable number of wh- in situ instances, which had the potential to influence other wh- questions by analogy. More crucially, around the fourth century AD the copula shì further grammaticalized into a focus morpheme, which ultimately replaced the old focus device, the movement of wh- elements and focused constituents, in Old Chinese. Stage 3 Around the sixth century AD, the typological change reached a completion point; after that, the Chinese language became a typical wh- in situ language. During the course of this transition, a majority of wh- words in Chinese were fused with the focus morpheme shì, and, consequently, all wh- words except shuí in Old Chinese were replaced by new lexical forms. Since grammar is an integral system, any historical changes certainly affect related components. The emergence of the copula shì and its further grammaticalization into a focus marker eventually replaced the old focus constructions, namely the wh- movement, constituent negation, and focused nouns. As a result, the texture of Chinese grammar was altered.

3.6

Wh- Words in Preposition Phrases

In Old Chinese, in general, prepositions overwhelmingly occurred in sentence-final position (Wang 1989: 198‒216, He 2004). As far as specific sentence structures such as passives, comparisons, instrument phrases, and ditransitives are concerned, the related preposition phrases were limited to sentence-final position. When the object of a preposition was a wh- element, however, only the construction of (74a) is attested in the texts, whereas (74b) and (74c) were ungrammatical. (74)

Declarative structure: Subj + V + Obj (PREP NP). Wh- question structure: (a) Subj + (Wh- element PREP) + V + Obj? (b) *Subj + V + Obj + (Wh- element PREP)? (c) *Subj + V + Obj + (PREP Wh- element)?

Compared to the corresponding declarative structure, obviously, the wh- element in a preposition phrase undergoes two movements: first, the wh- element is moved to the left of the preposition; second, the whole PP must be fronted to before the matrix verb,

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Focus and Wh- Words

58

the landing site of the focused constituents. In what follows, we will illustrate this syntactic property by analyzing concrete examples. The two sentences in (33) are found in the same paragraph of the text (the Mengzi): the first clause is a question, and the second clause is the corresponding answer (cf. Section 3.4.2). The “wh- word + PREP” phrase in the question occurred in preverbal position, but the PP in the corresponding declarative sentences occurred in sentencefinal position. There is a division of labor between the two synonymous locative prepositions, hū and yú: the former was often used in questions and the latter more frequently in declarative sentences. For the sake of simplicity, we have chosen the same verbs in the same syntactic construction to illustrate the contrastive distribution between the whquestions and their corresponding declarative instances. In Old Chinese, the canonical ditransitive construction and its corresponding wh- question are formulated as follows. (75)

(a) Declarative form of ditransitive instances: Subj V Oi yǐ Od. (b) Question form: Subj [wh- Od yǐ] V Oi

All the ditransitive instances in Old Chinese obeyed the above formulas. In the following two sets of examples, each uses the same verb but one is a declarative sentence, and the other is a question. The following examples thus illustrate the syntactic behaviors of those prepositions with a wh- element. (76)




(a) 何以贈之? (詩經 渭陽) Hé yǐ zèng zhī? what with send she “With what should I send her?” (Shi Jing, Wei Yang, 1000–600 BC)




(b) 贈之以芍藥。(詩經 溱洧) Zèng zhī yǐ sháoyào. send she with peony “I sent her with peonies.” (Shi Jing, Qin Wei, 1000–600 BC) (77)




(a) 何以報我? (左傳 僖公二十三年) Hé yǐ bào wǒ? what with repay me “With what will you repay me?” (Zuo Zhuan, Xi Gong Er Shi San Nian, 550–400 BC)

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Wh- Words in Preposition Phrases

59




(b) 天報之以福。 (荀子 宥坐) Tiān bào zhī yǐ fú. God repay he with luck “God repays him with luck.” (Xun Zi, You Zuo, 250 BC) Likewise, the following set of examples includes instrumental phrases. In (78a), the wh- element is a “wh- word plus N” phrase that is pied-piped to the left periphery of the preposition. The whole preposition phrase has also moved to preverbal position. (78)




(a) 其將何辭以對? (左傳 隐公三年) Qí jiāng hé-cí yǐ duì. he will what-word with answer “With what words will he answer?” (Zuo Zhuan, Yin Gong San Nian, 550–400 BC)




(b) 對以禮。(禮記 曲禮) Duì yǐ lǐ. answer with ritual “(He) answered with rituals.” (Li Ji, Qu Li, 450 BC) The whole preposition phrase seems to have moved to preverbal position. If there was an auxiliary verb, however, the auxiliary verb had to be inserted between the fronted whelement and the preposition, as illustrated in (79). (79)




孰可以代之? (左傳 襄公三年) Shú kě yǐ dài zhī? who can with replace he “With whom can he be replaced?” (Zuo Zhuan, Xiang Gong San Nian, 550–400 BC)

The above example is striking, revealing that the surface form might be derived in three steps, as follows. (80)

Step 1. The wh- element is moved to the left of the preposition. AUX V Obj [P P Wh-i PREP ti] 可 代 之 孰 以 kě dài zhī shú yǐ can replace it who with

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

60

Focus and Wh- Words Step 2. The PP with the wh- element is moved to the left of the matrix verb. AUX [P P Wh-i PREP ti] V Obj T pp 可 孰 以 代 之 kě shú yǐ dài zhī can who with replace it Step 3. The wh- word is moved to the position prior to the auxiliary verb. Wh-i AUX Ti PREP V Obj 孰可以代之? Shú kě yǐ dài zhī? Who can with replace he “With whom can he be replaced?”

Recall that the landing site of wh- elements in Old Chinese is “Subj Temp Whelement Auxi/Pre VP.” The form, which requires that the auxiliary verb intervene between the fronted wh- element and the verb/preposition, overrides the fronted whPP, producing the surface structure illustrated in (81). However, we must admit that the above derivations are just an intuitively appealing explanation rather than a rigorous treatment in the generativist framework. We leave this question open for future research. It is universal that wh- phrases contain a +FOC semantic feature and that they are the natural focus of a sentence in which they occur. This +FOC must be checked by certain grammatical devices, including the word order shift in Old Chinese and the addition of the focus marker shì in Modern Chinese. In Chinese, wh- in situ has no relation to quantificational properties or sentence-final question particles. Wh- elements have the potential to be extended to express indefiniteness or universal references, but the emergence of these usages is conditioned by the functions of other grammatical constructions. The wh- words in Old Chinese did not have these noninterrogative usages and they did not occur until after the eleventh century AD. The nature of wh- words has not changed throughout history, but devices for checking their +FOC feature have changed. Our diachronic investigation reveals that the factors responsible for blocking whmovement and licensing wh- in situ are quite straightforward. The emergence of the copula shì and its further grammaticalization triggered the typological change in two ways. First, the emergence of the copula created a considerable number of “V + wh- word” instances, in contrast to the “wh- word + V” ordering in Old Chinese. Second, and most importantly, the focus marker shì ultimately replaced the older device of word order change. From typological and diachronic perspectives, we can see that the syntax of wh- words in a language is by no means isolated. Their behaviors are in agreement with other focus expressions, such as relativization, constituent negation, and

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

Wh- Words in Preposition Phrases

61

focused-noun construction. For wh- movements, the landing site may interact with the topicalization construction or other functionally related structures. The evolution of a language can serve as a natural laboratory for linguistic theories, providing a straightforward yet convincing solution to issues involved in complex and sophisticated theoretical hypotheses. Studies of Chinese wh- in situ have played an important role in the development of generative linguistics, according to which many hypotheses have been proposed. For example, studies have observed the following: whphrases in Chinese also undergo movement at the “logic form” level that is exactly parallel to their counterparts in English; by nature, wh- words in Chinese are not interrogative and express several kinds of quantifier that must be licensed by certain operators, and the sentence-final question particle ne is responsible for wh- in situ. If the diachronic facts are taken into consideration, all these issues should be rethought because the synchronic system of syntax is the outcome of many diachronic developments during different periods. The diachronic perspective enables us to successfully identify the uniform semantic features of wh- words and to accurately describe the syntactic behaviors of wh- elements. Thus many meaningless theoretical arguments can be avoided. Wh- movement and wh- in situ have been the two most important pieces of empirical evidence driving the development of generative linguistics over the past six decades; many concepts and hypotheses have been proposed that center on these facts. Almost all such works have been conducted from a synchronic perspective, and because a language is considered either wh- movement or wh- in situ, it is difficult to uncover the factors distinguishing a wh- movement language from a wh- in situ language. However, discovering the truth will be much easier if a diachronic approach is taken. Diachronic research can be helpful in clarifying many issues and can make a significant contribution to theoretical linguistics.

https://doi.org/10.1017/9781108921831.003 Published online by Cambridge University Press

4 Serial Verb Construction

4.1

Introduction

During the transitional period from Old to Medieval Chinese, the disappearance of verb conjunctions resulted in the emergence of serial verb constructions. This event had a profound and far-reaching impact on the evolution of the grammar, which in turn motivated many further developments, such as the establishment of the resultative construction, the aspect system, the disposal construction, and the changes in the passive construction. In Old Chinese, where two or more verb phrases within a single clause (including adjectival and adverbial phrases) had to be connected by the conjunction ér or qiè, serial verb construction did not exist. In this grammatical system, it was impossible for any fusion and reanalysis across verb phrases to take place because the reanalysis had to involve a context consisting of more than one element that occurred adjacently. However, these conjunctions started to decline in Late Old Chinese and eventually died out in the first half of the Medieval Chinese period, which enabled many grammatical changes to happen. Then, the first verb in a serial verb construction became grammaticalized into various grammatical markers.

4.2

Conjunctions of Verbs and Clauses

From Old to Medieval Chinese, the Chinese language underwent a typological change in terms of ways to connect two or more verb phrases within a clause. By definition, serial verb construction refers to a type of grammatical construction in which a sequence of verbs or verb phrases is left unmarked (Crystal 2008: 434). Specifically, the sequence of verbs may share an object, but there is no conjunction to mark their co-ordinate relationship. From a synchronic perspective, Contemporary Chinese undoubtedly belongs to the serial verb type of language because it lacks a conjunction to link two or more verb phrases within a clause. That is, Contemporary Chinese has no equivalent to the conjunction word “and” in English that must be used in a verb co-ordination, e.g.

62

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Conjunctions of Verbs and Clauses

63

I went to a grocery and bought some fruit. Some researchers have even claimed that serial verb constructions are found predominantly in isolating languages such as Chinese or the Kwa languages of West Africa, implying that the existence of serial verb constructions in Chinese is related to its property of being an “isolating language” type (Déchaine 1993, Lord 1993, and Bussmann 1996). Obviously, this view is not supported by the diachronic evidence in the history of the Chinese language. In Old Chinese, there were two conjunctions, qiě and ér, that served to link verbs with a clear division of labor: in general, the former linked verbs or adjectives that were semantically and syntactically parallel, which meant that two or more actions or qualities took place and existed at the same time, and the latter linked those that took place in a temporal sequence or had unequal syntactic statuses, such as a relation between the adverbial and the matrix verb or the matrix verb and the resultative. At that time, the conjunction qiě could juxtapose two adjectives or verbs, as illustrated in (1) and (2): (1)




河水清且漣猗。 (詩經 伐檀) Héshuǐ qīng qiě liányǐ. river clear and rippled “The river is clear and rippled.” (Shi Jing, Fa Tan, 1000–600 BC)

(2)




(趙盾)鬥且出。 (左傳 宣公二年) (Zhào Dùn) dòu qiě chū. Zhao Dun fight and run-away “Zhao Dun ran away after fighting (with them).” (Zuo Zhuan, Xuan Gong Er Nian, 550–400 BC)

In the above situations, qiě was the most appropriate conjunction. However, in the following situations, only the conjunction ér was suitable for linking verbs. First, within a verb co-ordination, when two or more verbs whose actions took place in a temporal sequence shared an object, the action denoted by the first verb happened first. In this case, the conjunction ér had to be used to link the verbs, as illustrated in (3) and (4). (3)




學而時習之。(論語 学而) Xué ér shí xí zhī. learn and often review it “Learn and often review it.” (Lun Yu, Xue Er, 500 BC)

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Serial Verb Construction

64 (4)




予既烹而食之。(孟子 萬章) Yǔ jì pēng ér shí zhī. I already cook and eat it “I have already cooked and eaten it.” (Meng Zi, Wan Zhang, 300 BC)

If two adjectives contrasted with one another, the conjunction ér was often used instead of qiě, although it did not necessarily imply a temporal sequence, as illustrated in (5) and (6):




(5)

君子欲訥於言而敏於行。(論語 里仁) Jūnzǐ yù nè yú yán ér mǐn yú xíng. gentleman want slow in speech and quick in action “The gentlemen should be slow in speech and quick in action.” (Lun Yu, Li Ren, 500 BC)

(6)

宰予之辭, 雅而聞也。 (韓非子 顯學) Zǎi Yǔ zhī cí yǎ ér wén yě. Zai Yu GEN speech elegant and well known PRT “Zai Yu’s speech was elegant and well known.” (Han Fei Zi, Xian Xue, 300 BC)




Second, if there were two clauses within a sentence, which usually implied a temporal sequence, they had to be linked by the conjunction ér, as illustrated in (7). If there were more than two clauses, only the last two clauses needed to be linked by ér, as illustrated in (8).




(7)

魚餒而肉敗。 (論語 鄉黨) Yú něi ér ròu bài. fish spoiled and meat rot “The fish became spoiled, and the meat rotted.” (Lun Yu, Xiang Dang, 500 BC)

(8)

殺雞, 為黍而食之。 (論語 微子) shā jī, wéi shǔ ér sì zhī. kill chicken cook millet and feed him “(The man) killed a chicken, cooked millet, and fed him.” (Lun Yu, Wei Zi, 500 BC)




Third, the conjunction ér was often used to connect an adverbial phrase to the matrix verb, as illustrated in (9) and (10):

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Conjunctions of Verbs and Clauses (9)

65




子路率爾而對。(論語 先進) Zi Lù shuàiěr ér duì. Zi Lu hastily and answer “Zi Lu answered hastily.” (Lun Yu, Xian Jin, 500 BC)

(10)




夫子莞爾而笑。(論語 陽貨) Fūzǐ wǎněr ér xiào. master winsomely and smile “The master smiled winsomely.” (Lun Yu, Yang Huo, 500 BC)

Fourth, the matrix verb and the resultative also needed to be juxtaposed by the conjunction ér, as illustrated in (11): (11)




子與人歌而善。 (論語 述而) Zi yǔ rén gē ér shàn. Confucius and people sing and well “Confucius sang with people and they sang well.” (Lun Yu, Shu Er, 500)

Both ér and qiě could occur within a single sentence, which clearly shows their division of labor. In the following example, the first adjectival phrase bù yì “not righteous” described the action of achieving the statuses fù “rich” and guì “noble.” Thus the conjunction ér was inserted between the first adjective and the last two adjectives because they represent a temporal sequence, but the conjunction qiě was used to link the last two adjectives because they do not imply a temporal sequence. (12)




不義而富且貴。(論語 述而) Bù-yì ér fù qiě guì. injustice ER rich QIE noble “(Someone) did injustices to be rich and noble.” (Lun Yu, Shu Er, 500 BC)

In the grammar of Contemporary Chinese, however, there are no conjunctions equivalent to ér and qiě. This type of change indicated that old grammatical devices had simply died out, and no replacement or alternative apparatus had been innovated to perform the former functions. Additionally, the combination of the functions of both ér and qiě was equal to that of and in English, which can be seen from the English translations. It is

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Serial Verb Construction

66

clear that Old Chinese had a finer method of linking verbs, adjectives, and adverbials within a sentence. However, it is still unclear what motivated the decline and complete disappearance of the two conjunctions ér and qiě. Since the Medieval Chinese period, many important grammatical morphemes, such as the disposal and passive markers, could develop only out of the first verb of a serial verb construction, which are discussed in following chapters. That is, Chinese underwent a typological change from nonserial to serial verb construction, an event that took place during the transition period from Old to Medieval Chinese.

4.3

The Decline of the Co-ordinate Connective Ér

There is an asymmetry between nominal and verbal co-ordination in Contemporary Chinese. There is a nominal conjunction hé in Contemporary Chinese, but it cannot coordinate two verbs or verbal phrases, as illustrated in (13): (13)

(a) *她已經喫飯和看書。 (現代漢語) *Tā yǐjīng chī-fàn hé kàn-shū. she already eat-food and read-book “She has already eaten food and read a book.” (b) *她已經買和喫蘋果。 *Tā yǐjīng mǎi hé chī píngguǒ. she already buy and eat apple “She already bought and ate apples.” (Contemporary Chinese)

In fact, verbal co-ordination is simply impossible in Contemporary Chinese even without an overt conjunction as in *kàn xiě xìn “read + write + letter.” This is actually one effect of the resultative construction on Chinese grammar. However, the situation is exactly the opposite in Old Chinese: verb co-ordination (similar to verb co-ordination in English) was a common construction at that time. In this case, the obligatory conjunction was ér, as illustrated in (14): (14)




豹自後擊而殺之。 (左傳 襄公二十三年) Bào zì hòu jī ér shā zhī. leopard from back strike ER kill he “A leopard struck and killed him from behind.” (Zuo Zhuan, Xiang Gong Er Shi San Nian, 550–400 BC)

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

The Decline of the Co-ordinate Connective Ér

67

Table 4.1 The decline of the connective ér over time Title

Time

Tokens of ér every 10,000 characters

Lun Yu Meng Zi Shi Shuo Xin Yu Bai Yu Jing Liu Zu Tan Jing Shen Hui Yu Lu Lao Qi Da Piao Tong Shi

500 BC 250 BC AD 450 AD 500 AD 700 AD 750 AD 1350 AD 1350

160 171 48 64 9 16 0 1

According to Zhu’s (1958: 22) study of the Zuo Zhuan, the connective ér was obligatory in verb co-ordination at that time. However, this rule was more and more frequently broken over time. For example, in the text of Zhuang Zi (approximately 200 years later than Zuo Zhuan), ér was absent from approximately 20 percent of the examples of verb co-ordination. Finally, this connective was completely abandoned around the tenth century AD and was preserved only in certain fixed written expressions. Table 4.1 shows the decline of the connective ér over time. We selected two texts from each period to investigate the change in the frequency of the connective ér. As Table 4.1 shows, the connective ér declined steadily over time from Old to Medieval Chinese. A “growth–decline” relationship existed between the resultative construction and this connective ér because of the semantic and syntactic incompatibility between them. As ér became optional, the verb co-ordination construction changed accordingly, and the maximum number of verb phrases within a verb co-ordination or serial verb construction was four. The process of the development is formalized in (15). (15)

VP1 ér VP2 → “VP1 ér VP2”/ “£VP1 VP2” → “VP1 VP2.”

When two verb phrases were adjacent (i.e. not separated by a connective), they were subject to compounding into a single verb. Verb compounds of this kind made their first appearance around the first century BC. (16)




射殺靈公。 (史記 齊太公世家) Shè shā Líng Gōng. shot kill Ling Duke “(Someone) shot and killed Duke Ling.” (Shi Ji, Qi Tai Gong Shi Jia, 100 BC)

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Serial Verb Construction

68 (17)




乃收養之。 (論衡 詰驗) Nǎi shōu yǎng zhī. then adopt raise it “Then adopted and raised it.” (Lun Heng, Jie Yan, AD 100)

In Contemporary Chinese shōu-yǎng “adopt–raise” is an unanalyzable compound verb. Although accurate dating of verb compounds is difficult or even impossible, it is clear that only when the connectives became optional could this kind of verbal compounding come into existence. In Old Chinese, clauses were also connected by ér, as illustrated in (18) and (19):




(18)

是何異於刺人而殺之? (孟子 梁惠王上) Shì hé yì-yú cì rén ér shā zhī? this how different-from stab people ER kill they “How is it different from the fact that (you) stab the people and kill them?” (Meng Zi, Liang Hui Wang Shang, 300 BC)

(19)

驅虎、豹、犀、象而遠之。 (孟子 滕文王) Qū hǔ, bào, xī, xiàng ér yuǎn zhī. drive tiger leopard rhino elephant ER distance them “(They) drove away tigers, leopards, rhinoceroses, and elephants and kept them at a distance.” (Meng Zi, Teng Wen Wang, 300 BC)




Due to the decline of the connective ér, conjoined clauses of this kind were simply left unmarked by any overt grammatical device, as illustrated below (cited from Zhu 1958: 25).




(20)

攻鄭敗之。 (史記 趙世家) Gōng zhèng bài zhī. attack Zheng defeat it “(They) attacked the Zheng country and defeated it.” (Shi Ji, Zhao Shi Jia, 100 BC)

(21)

擊李由軍破之。 (史記 曹相國列傳) Jī Lǐ Yóu jūn pò zhī. strike Li You troop break it “(They) attacked the troops of Li You and overcame them.” (Shi Ji, Cao Xiang Guo Lie Zhuan, 100 BC)




https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

The Decline of the Co-ordinate Connective Ér

69

Li (1987) suggested that the separable resultative structure was derived from the structure exemplified in (20) and (21) through dropping the pronoun zhī “it.” In our view, however, this hypothesis is problematic. All of the verbs in the second clauses had to be transitive because they were followed by the pronominal object zhī. In Medieval Chinese, however, the resultative elements of the separable structure were overwhelmingly intransitive verbs or adjectives, and they could not have pronominal objects. Actually, in Old Chinese the second clause connected by ér could also be an intransitive element, which was directly related to the emergence of the separable resultative structure:




(22)

聞王命而遂不果。 (孟子 公孫醜) Wén wáng mìng ér suì bù-guǒ. hear king order and then unsuccessful “(He) heard the king’s order and was unsuccessful.” (Meng Zi, Gong Sun Chou, 300 BC)

(23)

知管仲、晏子而已矣。 (孟子 公孫醜) Zhī Guǎn Zhòng, Yàn Zi ér yǐ yǐ. know Guan Zhong Yan Zi and complete PRT “(They) just knew Guan Zhong and Yan Zi. That’s it.” (Meng Zi, Gong Sun Chou, 300 BC)

(24)

如之何其使斯民饑而死也。 (孟子 梁惠王) Rú-zhī-hé qí shǐ sī mín jī ér sǐ yě. why he make these-people starve and die PRT “Why did he make these people starve and die?” (Meng Zi, Gong Sun Chou, 300 BC)







Actually, examples of this kind developed into the separable resultative construction when ér dropped; for instance, in (24), jī ér sǐ “hungry and die” would become jī sǐ “hungry die,” an example of the separable resultative structure. The separable resultative structure started to occur around the first century BC, and became a syntactic pattern around the fifth century AD. Example (25) is one of the earliest instances of the separable structure we have found: (25)




其子死, 哭之失明。 (史記 仲尼弟子列傳) Qí zi sǐ, kū zhī shīmíng. his son die, cry him blind “His son died and (Zi Xia) cried over him so much he went blind.” (Shi Ji, Zhong Ni Di Zi Lie Zhuan, 100 BC)

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Serial Verb Construction

70

The emergence of the separable structure is one of the consequences of the decline of the verbal connective ér. When two clauses were conjoined without any marker, they were subject to clause combination, which would increase the dependence of one clause upon another.

4.4

Fusion across Clauses

The separable resultative structure came from the combination of two clauses, which in turn was made possible by the absence of the connectives ér and qiě. Now we turn to a theoretical and empirical examination of the pathway to the clause combination. As mentioned previously, the separable resultative structure can be considered a kind of complex sentence consisting of a “nucleus” and a “margin” (adjunct). By definition, a nucleus clause can stand alone and a margin “may not stand alone but nevertheless exhibit different degrees of dependency” (Hopper and Traugott 2003: 177). Historically, the complex sentence was derived from a clause combination that integrates two mutually relevant but separate and autonomous nuclei into one structure. The outcome of clause combination is a more complex structure because it has two subparts. Clause combination is a matter of degree, which may be thought of as a cline with three “cluster points.” The following cline is based on Hopper and Traugott (2003: 179). (26)

Parataxis hypotaxis dependence nucleus margin minimal integration maximal integration

With reference to the separable resultative structure, the three cluster points can be described as follows (Hopper and Traugott 2003: 169). (a) Parataxis. The two independent clauses stand in an “action–result” semantic relation. The resultative part is relatively compositional and often clause-like (e.g. the verbal element used as the resultative has its own subject or object). Thus there is a clause boundary between the action part and the resultative part. (b) Hypotaxis. The resultative part (i.e. clause) can no longer stand alone and is therefore dependent on the nucleus. Typically, these bounded resultatives are monosyllabic items that are occasionally modified by adverbs or negatives. At this stage the clause boundary is weakened, if not entirely lost. (c) Integration. The verb and resultative are integrated into a single constituent; consequently, the originally intervening materials are squeezed out. Like a common verb, verb–resultative phrases can precede a nominal

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

The Resultative Chéng

71

object. In other words, at this stage, the relationship between the verb and the resultative involves the highest degree of fusion, and the resultative element loses its lexical status. The clause combination described above – from minimal to maximal integration – is advanced by several factors. First, as mentioned before, the absence of the connective ér makes the two clauses adjacent to each other, setting the scene for the fusion of the two clauses. Second, the tendency toward disyllabification restricts the possibility of using monosyllabic words as resultatives (for details, see Chapter 5). As a result, monosyllabic resultatives can no longer truly stand alone and thus tend to depend on the preceding clause. Third, the semantic relation also plays a role in the fusion of the verb and resultative. As Givón (1990) said, “The more two events/states are integrated semantically or pragmatically, the more will the clauses that code them be integrated grammatically.” Action and result represent the starting and ending points of an event and are highly integrated in terms of semantics, so much so that the causative relation between the two points is often coded by a morphological device (e.g. Old Chinese and Old English) or even by a lexical item.1 In what follows, we sometimes use examples from the same text to illustrate the cline described above. Our assumption in doing so is that old and new forms might coexist for a long time. It is cross-linguistically true that new forms are continually emerging; in the process the older forms are not necessarily discarded but may coexist with the new form.

4.5

The Resultative Chéng

From the fifth century AD to the present day, the word chéng “complete” has been very commonly used as a resultative that indicates the accomplishment of an action. In Contemporary Chinese, it can occur between the verb and the resultative only if there is an object. Since chéng was originally an intransitive element, its use was quite different from that in Contemporary Chinese.2 (a)

1

2

Parataxis. In the following examples, the resultative chéng has its own subject, which is identical to the object of the verb. Thus the resultative is its own independent sentence.

Old Chinese employed tone shifts to achieve a causative form. In addition, in Old and Medieval Chinese, a single verb often conveyed a causative sense, which is expressed by a “verb + resultative” phrase in Contemporary Chinese; e.g. wū “smear” (Old and Medieval Chinese) corresponds to nòng-zāng “make-dirty” (Contemporary Chinese). Cheng could be a transitive verb when used in the sense of “aid” or “help.”

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Serial Verb Construction

72




(27)

支道林造《即色論》 , 論成。(世說新語 文學) Zhī Dàolín zào Jí-sè lùn, lùn chéng. Zhi Daolin write Ji-se theory, theory complete “Zhi Daolin wrote the theory of Ji-se and the theory was complete.” (Shi Shuo Xin Yu, Wen Xue, AD 450)

(28)

求紙筆作書, 書成。(世說新語 雅量) Qiú zhǐ bǐ zuò shū, shū chéng. request paper pen write letter, letter complete “(He) requested paper and pens to write a letter and the letter was complete.” (Shi Shuo Xin Yu, Ya Liang, AD 450)




(b) Hypotaxis. In the following examples, the resultative chéng is separated from the verb by an object, but the object NPs are not repeated as the subject of the resultative. This resultative is also modified by adverbs, and together they form a disyllabic unit. These resultative parts could behave either like an independent clause or like an adjunct to the nucleus, depending on the complexity of the clause.




(29)

何平叔注 《老子》 始成。 (世說新語 文學) Hé Píngshū zhù “Lǎozi” shǐ chéng. He Pingshu annotate Laozi just complete “Pingshu just finished annotating Laozi.” (Shi Shuo Xin Yu, Wen Xue, AD 450)

(30)

左太沖作 《三都賦》 初成。(世說新語 文學) Zuǒ Tàichōng zuò Sān Dū Fù chū chéng. Zuo Taichong write three capital prose just complete “Zuo Taichong just finished writing The Prose on Three Capitals.” (Shi Shuo Xin Yu, Wen Xue, AD 450)




(c) Integration into a single clause. The resultative in each of these examples is a bare monosyllabic word, and could not stand alone as an independent clause because of the weakened syntactic function of monosyllabic words. In this context, the resultative chéng “complete” became a bound morpheme. That is, this is an example of the separable resultative structure.

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

The Resultative Chéng

73




(31)

袁伯彥作名士傳成。 (世說新語 文學) Yuán Bóyàn zuò Míng-shì Zhuán chéng. Yuan Boyan write eminent-person biography complete “Yuan Boyan finished writing biographies of eminent people.” (Shi Shuo Xin Yu, Wen Xue, AD 450)

(32)

孫興公作天台賦成。 (世說新語 文學) Sūn Xìnggōng zuò Tiāntái fù chéng. Sun Xinggong write Tiantai prose complete “Sun Xinggong finished writing the prose of Tiantai.” (Shi Shuo Xin Yu, Wen Xue, AD 450)




The various uses of chéng “complete” were a manifestation of an ongoing change at that time: a given lexical item started to grammaticalize into a specific construction but remained the same in other contexts. There is always a discourse strategy for expressing the action– result relation: within two neighboring sentences, the first refers to an action and the second to the result. This is called “parataxis.” Nevertheless, the same meaning can also be expressed by two clauses co-ordinated within a sentence, a use of hypotaxis. Due to the decline of the verbal connective ér and the reduced syntactic freedom of monosyllabic words, chéng became a dependent clause when used alone as a resultative. Thus the bound resultative chéng did not develop out of its paratactic use; rather, its emergence was enabled by other syntactic and phonological changes. Once it became dependent, syntactically and phonologically, chéng was likely to become subject to fusion with the preceding verb. Unsurprisingly, chéng and its preceding verb became a single constituent, and cheng lost its lexical status. (d)

(33)

Clitic. Three hundred years later, the resultative chéng was fused with the verb, forming one constituent. At this point, chéng was no longer a lexical item but a clitic, and hence the compound phrase “V-chéng” could be followed by an object, as illustrated in (33) and (34): 織成錦繡麒麟兒。 (遊仙窟) Zhī chéng jǐnxiù qílín-er. weave complete silk unicorn “(She) finished weaving a silk unicorn.” (You Xian Ku, AD 750)

(34)




壘成方丈。 (敦煌變文 維摩詰) Lěi chéng fang-zhàng. build complete Buddhist abbey. “(They) finished building a Buddhist abbey.” (Dun Huang Bian Wen, Wei Mo Jie, AD 800–1000)

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Serial Verb Construction

74 4.6

The Resultative Bù-Dé

The resultative bù-dé “not able” was a compound-like word with two meanings: (a) as a common verb, it meant “not get”; (b) as an auxiliary verb, it meant “cannot.” This phrase often occurred after a clause to indicate the result of the preceding action. The relation can be divided roughly into four types, according to the degree of independence. (a) Parataxis. In the following examples, bù-dé was used as an independent clause to express the failure of the preceding action, as illustrated in (35) and (36): (35)

今正要人恁地理會, 不得。(朱子語類卷七十九) Jīn zhèng yào rén nèndì lǐhuì, bù-dé. now just request people that understand unable “Now (you) requested that other people understand in that way, but you could not (do so), and thought it again.” (Zhu Zi Yu Lei, Chapter 79, AD 1200)

(36)

只是便把光作燈不得。 (朱子語類卷九十七) Zhǐshì biàn bǎ guāng zuò dēng bù-dé. only then DISP light use-as lamp not able “Then (you) used light as a lamp, but it didn’t work.” (Zhu Zi Yu Lei, Chapter 97, AD 1200)

(b) Hypotaxis. In this case, the phrase bù-dé was integrated with the preceding clause to form a single clause, as illustrated in (37) and (38): (37)

終身改口不得。 (朱子語類卷七) Zhōngshēn gǎikǒu bù-dé. lifetime change-mouth unable “You can never take back your promise for a lifetime.” (Zhu Zi Yu Lei, Chapter 7, AD 1200)

(38)

便漏了他底也不得。 (朱子語類卷一一六) Biàn lòu le tā-dǐ yě bù-dé. then leave-out PERF his still not-able “We still cannot leave out his.” (Zhu Zi Yu Lei, Chapter 116, AD 1200)

(c) Integration into a single constituent. If no object intervenes between the verb and bù-dé, neither of them can stand alone. That is, they must rely on each other to form a clause, as illustrated in (39) and (40):

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

The Resultative Bù-Dé

75

(39)

更把捉不得。 (朱子語類卷一百十七) Gèng bǎzhuō bù-dé. even-more capture not-able “(It) is even more difficult to capture.” (Zhu Zi Yu Lei, Chapter 117, AD 1200)

(40)

事多記不得。 (朱子語類卷十六) Shì duō jì bù-dé. thing most memorize not-able “Most things cannot be memorized.” (Zhu Zi Yu Lei, Chapter 16, AD 1200)

(d)

High fusion. At this stage, the whole construction “V + bù-dé” had been fused into a single constituent, and thus could have an object:

(41)

皆曉不得子由意。 (朱子語類卷十八) Jiē xiǎo bù-dé Zi Yóu yì. all understand not-able Zi You intent “No one can understand Zi You’s intent.” (Zhu Zi Yu Lei, Chapter 18, AD 1200)

(42)

他既讀不得書。 (朱子語類卷十七) Tā jì dú bù-dé shū. he already read not-able book “He already cannot read books.” (Zhu Zi Yu Lei, Chapter 17, AD 1200)

These were the earliest examples of the potential form in Modern Chinese. The cline from parataxis to dependence is designed to describe the degrees to which two clauses or elements hold a particular semantic relation – action and result. It is not meant to imply that the dependent combination necessarily developed from the corresponding paratactic one. What we want to say is that parataxis is a kind of loose discourse organization. Presumably, this organization always existed in a language. However, although there may not be an action–result relation, the meanings of loose discourse may be expressed with a new pattern at a later stage just because the dependent elements are fused. For instance, from the sixth century to the first century BC, many co-ordinate verbs underwent the process of compounding due to the disappearance of the connective ér. As a result, the meanings of two independent clauses were replaced by a single verb compound plus an object (He 2005: 212), as illustrated in (43) and (44):

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

Serial Verb Construction

76 (43)




(a) 遂襲虞, 滅之。 (左傳 僖公五年) Suì xí Yú, miè zhī. then raid Yu, destroy it “Then, (they) raided the Yu country and destroyed it.” (Zuo Zhuan, Xi Gong Wu Nian, 550–400 BC)




(b) 襲滅虞。(史記 晉世家) Xí miè yú. raid destroy Yu “(They) raided and destroyed the Yu country.” (Shi Ji, Jin Shi Jia, 100 BC) (44)




(a) 射共王, 中目。 (左傳 成公十六年) Shè Gòng Wáng, zhōng mù. shoot Gong King, hit eye “(He) shot King Gong and hit his eye.” (Zuo Zhuan, Cheng Gong Shi Liu Nian, 550–400 BC)




(b) 射中共王目。 (史記 晉世家) Shè zhòng Gòng Wáng mù. shoot hit Gong King eye “(He) shot and hit King Gong’s eye.” (Shi Ji, Jin Shi Jia, 100 BC) In (43) and (44), each pair of sentences expresses exactly the same event, but the earlier sentences use two independent clauses, while the later sentences use verb compounds. This was a reflection of the language change. In (43a), xī “raid” and miè “destroy” occur in two separate clauses and each has its own object. The object of the verb miè “destroy” is a pronoun anaphoric with the object of xī, a configuration that was ruled out at the time of the establishment of the resultative construction. To conclude, many factors were involved in the process of the formation of the separable resultative structure. First, the temporal sequence was employed to symbolize the action–result relation – action first and result second, a word order reflecting reality. Second, the relation of “action–result” formed a cognitive package or a semantic unit that had the potential to be coded with one syntactic constituent. Third, the disappearance of the verbal connective ér made the verb and the resultative adjacent to each other, a proper context making the resultative dependent on its preceding verb, which became a syntactic pattern. Fourth, the increasing tendency toward disyllabification made it less possible that a

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

The Resultative Bù-Dé

77

monosyllabic resultative could stand alone as an independent clause. All these factors worked together in motivating the emergence of the separable resultative structure, where the resultative and verb rely on each other. The emergence of this separable resultative structure in turn made it possible for the verb and resultative element to undergo further development – fusion into a compound-like word.

https://doi.org/10.1017/9781108921831.004 Published online by Cambridge University Press

5 Disyllabification

5.1

Introduction

In general, Old Chinese was monosyllabic, which means that the overwhelming majority of the words were represented by a single syllable, regardless of whether they were content or function words. In Medieval Chinese, the phonological system was dramatically simplified; the number of consonants and vowels was reduced, and the syllabic structures were simplified. To restore the phonological distinctions of lexical items, the language increased the number of syllables for words, typically by adding one syllable to originally monosyllabic words. This disyllabification tendency has lasted nearly two millennia after that. This new sort of prosodic unit stimulated the fusion of two monosyllabic items, a key factor for the emergence of the resultative construction and other grammatical morphemes. Historically, the phonological change – disyllabification – had profound effects on the lexicon, morphology, and syntax of Chinese. The tendency toward disyllabification, which started as early as Old Chinese and is still at work in Contemporary Chinese, has had a significant impact not only on the lexicon but also on both morphology and syntax, and is considered one of the most important changes in the history of Chinese (Wang 1989: 1‒3). The momentum of this tendency gained more strength than ever during the period from roughly the first century AD to the tenth. Under the influence of this tendency, two monosyllabic words were subject to compounding if they often cooccurred in contexts where they were adjacent. By definition, compounding results from reanalysis in which the boundary between words and morphemes is weakened or entirely lost, a process with double effects – grammaticalization and lexicalization (Hopper and Traugott 2003: 49). Some pairs of verbs and resultatives have been frozen into compound verbs via fusion, which is simply a lexical matter, but some resultatives, typically those with great semantic generality, had become further grammaticalized to express a stable grammatical function. The fusion of verb and resultative has generated many new grammatical devices, including aspect markers, verbal clitics, and the resultative construction.

78

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Introduction

79

Phonological processes often interact with semantic and syntactic factors in the grammaticalization of morphemes and of whole constructions (Lehmann 1995: iii), but the manner of interaction may vary from language to language. Phonological constructions often involved in the onset of grammaticalization are intonation units, including their pitch and duration contours (Zwicky 1982, Langacker 1994, Givón 1991, Croft 1995). The basic phonological unit of Contemporary Chinese is a disyllabic chunk, representing a fundamental change of the phonological system in the past two millennia. Although in Old Chinese the equation “one syllable is equal to one word” was generally valid, in Contemporary Chinese a great majority of words are two syllables. This disyllabic form for the lexicon in fact compensates for the historical simplification of the phonological system (Wang 1980: 268). As we will see in what follows, the tendency toward disyllabification not only involves the formation of words (i.e. morphology) but also influences the grammar. The interface between phonological units and syntactic constructions is evident in many ways. For example, the maximal number of syllables of a Chinese phonological meter is four, which determines the maximal length of idioms and morphological reduplication. Let us use verb co-ordination in Medieval Chinese to illustrate the meter constraint on syntax. At the time, two or more verbs could share an object, and, theoretically speaking, this construction could contain any number of verbs. However, the maximal number of verbs in this construction was in fact limited to four, each of which must be monosyllabic. The examples in (1) and (2) illustrate the longest possible type of verb co-ordination; no longer examples have been attested in the whole history of Chinese.




(1)

盡斬殺降下之。 (史記 匈奴列傳) Jìn zhǎn shā xiáng xià zhī. entirely chop kill subdue vanquish them “(Chinese troops) chopped, killed, subdued, and vanquished all of them.” (Shi Ji, Xiong Nu Lie Zhuan, 100 BC)

(2)

乃遣晃及史溪邀擊破走之。 (三國志 魏書) Nǎi qiǎn Huǎng jí Shǐ Xī yāo jī pò zǒu zhī. then dispatch Huang and Shi Xi rouse hit break expel them “Then, General Huang and Shi Xi were dispatched to rouse, beat, break, and expel them.” (San Guo Zhi, Wei Shu, AD 300)




Each verb and the shared object in both (1) and (2) hold an action–patient relationship; in other words, all of the verbs in the construction must be transitive. Therefore, we can

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

80

disassemble the VP into four separate clauses according to the “distributive law” in mathematics: (a + b + c + d )x = ax + bx + cx + dx. For instance, example (1) can be disassembled as follows: (3)

斬殺降下之 = 斬之, 殺之, 降之, 下之。 zhǎn shā xiáng xià zhī chop kill subdue vanquish them = zhǎn zhī, shā zhī, xiáng zhī, xià zhī chop them, kill them, subdue them, vanquish them

This type of verb co-ordination disappeared before the tenth century AD, and the resultative construction was firmly established by the tenth century AD.

5.2

Disyllabic Words

In Contemporary Chinese, disyllabic words are much more numerous than monosyllabic words or other multisyllabic words. Table 5.1 shows the results of a survey of the book The First 3,000 Common Words in Standard Mandarin (Lü 1963). When all three categories are taken together, the disyllabic words account for more than 75 percent of 3,000 words. Beyond these basic words, there is a greater percentage of words that are disyllabic. Overall, it is estimated that more than 80 percent of Contemporary Chinese words are disyllabic (Hu 1981: 28). In addition to the greater percentage, disyllabic words behave more freely in a syntactic sense than monosyllabic words. According to Lü (1963), a monosyllabic word must have an extra syllable added in order to become a syntactic constituent or to be used as an independent sentence in many circumstances. The extra syllable may be an affix, a suffix, or a content word. This point is elaborated on by Lü and can be summarized as follows. First, to be a syntactic constituent such as a subject or an object or to be used as a form of address, monosyllabic surnames must be made disyllabic by adding the affix lǎo or xiǎo, literally meaning “old” or “young,” respectively. For example, it is quite unacceptable to call somebody Zhào, but either Lǎo Zhào or Xiǎo Zhào sounds natural to the ears of Chinese Table 5.1 Disyllabic words in Contemporary Chinese Syntactic categories

Total

Disyllabic words

Percentage

Nouns Adjectives Verbs

1621 451 941

1379 311 573

85 69 61

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Motivations for Disyllabification

81

people. That is, monosyllabic surnames behave like bound morphemes. In contrast, disyllabic surnames are not bound and can be used alone. However, if these disyllabic words are affixed by lǎo or xiǎo, the resulting phrases will sound unnatural or even unacceptable because they are also not disyllabic. For example, Sīmǎ, a Chinese surname, can be used as a syntactic constituent or a form of address, but is resistant to affixation. Second, monosyllabic country names must be attached by the categorization word guó “country” when used as a sentence constituent or an independent sentence in conversation, for example, Měi-guó “United States” and Yīng-guó “England.” By comparison, disyllabic country names such as Rìběn “Japan” and Nuówēi “Norwa” are generally used without the categorization word; they sound unnatural if the word guò is added. The preference for disyllabic phrases can also be seen in another aspect of compounds containing country names. When a disyllabic country name occurs as a morpheme of a compound, the second syllable is omitted in order to make a disyllabic compound, such as Yīng-bàng “British pound” and Rì-yuán “Japanese yen.” Finally, in the ordinal numbers for dates of the lunar months, the ordinal number affix chū is necessary for the first ten days, whose numbers are all monosyllabic. For example, (zhēng-yuè) chū-wǔ but *(zhēng-yuè) wǔ “(January) fifth.” However, this affix cannot be applied to the dates after the tenth because they are all already disyllabic, and the disyllabic ordinal numbers simply take the same forms as their corresponding cardinal numbers, such as (zhēngyuè) shíwǔ “(January) fifteenth.” The phenomena mentioned above are only a few cases exemplifying that the syntax of disyllabic units is less restricted than that of monosyllabic ones. Hu (1981: 28‒41) even claimed that the disyllabic form is the phonological representation of the lexicon of Modern Chinese. The disyllabification of the lexicon is thought of as one of the most significant changes in Chinese grammar in the past 2,000 years because it is related to the formation of words (Wang 1980: 342). In Chinese, the formation of compound words has many things in common with syntax; for instance, compound words are formed exactly in conformity with the syntax (Zhu 1982: 25). This chapter demonstrates that this phonological change is also a major factor enabling the emergence of the resultative construction.

5.3

Motivations for Disyllabification

One of the most important changes in the history of Chinese is the increase in disyllabic words and the corresponding decrease in monosyllabic words. There never was a stage in which the Chinese lexicon was purely monosyllabic, and even in the earliest documents disyllabic words were already attested. According to Guo (1997), disyllabic words account for approximately 20 percent of the lexicon in Old Chinese. He also pointed out that the methods for creating disyllabic words were in the embryonic stage in the period from the eleventh century to the seventh century BC and were established

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

82

Disyllabification

in Late Old Chinese. The tendency toward disyllabification grew steadily over time and finally became the predominant phonological representation of the lexicon. A question naturally arises: what was the motivation for the development of disyllabic words? One widely accepted hypothesis is that disyllabic words were created as compensation for the (well-attested) simplification of the phonological system over the history of Chinese. Below are two representative statements of this view. However, the simplification of the phonological system does not mean that Chinese has lost something, which has been compensated for by other aspects. The emergence of the neutral tone provides Chinese with a new phonological symbol, and also it often serves as the phonological representation of many new grammatical particles. A large number of disyllabic words have made it possible that Chinese efficiently distinguished phonological representations of different words with its relatively simple phonological system. Therefore, simplification is good for the language. (Wang 1980: 268) Why is there a strong tendency to disyllabify the lexicon of Modern Chinese? It is because there are too many homonyms in Modern Chinese. Due to phonological evolution, many lexical items with different phonological representations have become homonyms in Contemporary Chinese, and disyllabification is in fact a compensation for that change. There are more homonyms in northern dialects, where the tendency of disyllabification is stronger. By comparison, in those southern dialects there are fewer homonyms because of their relative complex phonological systems, and the tendency of disyllabification is weaker. (Lü 1963) With the simplification of the phonological system, an increasing number of words have become homonyms. The addition of syllables is an effective way to preserve sufficient numbers of distinctive sound forms in the lexicon. This hypothesis successfully explains some important differences between the northern and southern dialects of Chinese. First, let us consider the historical simplification of the phonological system. According to Wang (1980: 268), many phonological devices were abandoned in the history of Chinese, including the following phenomena. (a) All three stop consonants at the coda position – [p], [t] and [k] – disappeared (Wang 1980: 133). (b) The distinctive “voiced” and “voiceless” were neutralized, and, as a result, the set of voiced consonants all merged with their corresponding voiceless counterparts; for example, [b] became [p] and [d] became [t] (Wang 1980: 143).

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Motivations for Disyllabification (c) (d) (e)

83

During the transition from Old to Medieval Chinese, the long entering tone (chang ru) merged with the falling tone (Wang 1980: 143). The thirty-five consonants used as initials in Medieval Chinese were reduced to twenty in Modern Chinese (Wang 1980: 268). Regarding the finals, sixteen rime categories ( yun she) were reduced to twelve (Wang 1980: 268).

Clearly, the phonological system had been significantly simplified since Old Chinese, which directly resulted in a dramatic increase in the number of homonyms. Too many homonyms would create serious ambiguity in conversation, though this is less of a problem in written communication because most have different written forms. To resolve this problem, one alternative is to add more syllables to those monosyllabic roots. Another factor in the tendency toward disyllabification may be the fact that an increasing number of new words have to be coined as human activities become increasingly complex, which can also certainly increase the number of homonyms. For the same reason mentioned in the paragraph above, the phonological representations of newly introduced concepts may be lengthened by adding another syllable. That is, the increasing complexity of the lexicon, together with the simplification of the phonological system, provides an endless force to advance the tendency toward disyllabification. There is evidence in Chinese dialects other than Mandarin in favor of the hypothesis that the simplification of the phonological system was the cause of the tendency toward disyllabification. In general, the southern dialects have tended to preserve more phonological properties of Old or Medieval Chinese than Northern ones, particularly standard Mandarin, and the southern dialects accordingly have more monosyllabic words. For example, Cantonese has eleven types of tone and fifty-nine types of final (Li et al. 1995: 25‒41), but standard Mandarin has only four tones plus a neutral and thirty-nine finals. Many disyllabic words of standard Mandarin correspond to monosyllables in presentday Cantonese, as in Table 5.2. Additionally, Cantonese lacks many of the suffixes of standard Mandarin, such as the nominal suffixes -zi, -er, and -tou, which function primarily to make a monosyllabic

Table 5.2 Monosyllabic and disyllabic words in Cantonese and Standard Mandarin Cantonese

Standard Mandarin

hai23 ŋan23 mɪŋ21 mei22

螃蟹 pángxiè “crab” 眼睛 yǎnjīng “eye” 明白 míngbái “understand” 味道 wèidào “taste”

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

84

word disyllabic. A simple explanation is that Cantonese has more phonological devices to distinguish lexical forms and thus does not need as many disyllabic words.

5.4

The Development of Disyllabic Words

The tendency toward disyllabification started as early as 2,000 years ago and grew steadily over time. In the beginning, disyllabification was basically a matter of phonology and of the lexicon. Presumably, once this tendency gained unusual momentum and a significant proportion of lexical items became disyllabic, it could affect the syntax. Regarding the interface between disyllabification and syntax, our goal is to identify how this tendency plays a role in the creation of the resultative construction; our first task is to date the key period for the development of disyllabic words. First, let us consider our own investigation of the dating issue. We selected 124 disyllabic verbs from the twelfth century AD text Zhu Zi Yu Lei that are still used as compound verbs in Contemporary Chinese and dated them to determine when they were introduced into the language. The results are shown in Table 5.3. Most of these disyllabic words were created after the fifth century AD. Of the 124 disyllabic verbs, approximately 60 percent came into the language during the period from the eighth century AD to the twelfth, and roughly 34 percent were created within the period from the fifth century AD to the eighth. Of the 124 verbs, 94 percent were created from the fifth century AD to the twelfth. Only 6 percent were created before the fifth century AD. Under the assumption that this phenomenon was one indication of the development of disyllabic words, we can conclude that after the fifth century AD the tendency toward disyllabification developed very quickly, and the key period for the development of disyllabic words was from the eighth century AD to the twelfth. As we will see in what follows, the period of the quick development of disyllabic words (from the fifth century AD to the twelfth) was also the key period for the formation of the resultative construction, which suggests that there may be a relation between these two events. Table 5.3 Disyllabic verbs from the fifth century AD to the twelfth Periods

Compounds

Percentage

Before the fifth century AD Fifth–eighth centuries AD Eighth–twelfth centuries AD

7 42 75

6 34 60

Note: The texts that we investigated include (a) Shi Shuo Xin Yu, fifth century AD; (b) Dun Huang Bian Wen, eighth–tenth centuries AD; and (c) Zhu Zi Yu Lei, twelfth century AD.

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

The Development of Disyllabic Words

85

Another important piece of evidence for dating the development of disyllabic words is the time when certain nominal affixes and suffixes were introduced. These affixes and suffixes function mainly to make a disyllabic word out of a monosyllabic root. We assume that their emergence was a reflex of the tendency toward disyllabification. According to Wang (1989: 4), the affixes lǎo and ér were introduced into the language during the Medieval Chinese period (e.g. lǎo-pó “wife,” lǎo-shī “teacher,” gài-er “cover”), and the suffixes -zi and -tou appeared in the language within the period from the third century AD to the fifth (e.g. shízi “pebble” and zhǒngzi “seed”). Clearly, these nominal particles originated in the time frame in which disyllabic words were developing most quickly. Disyllabification was related to every syntactic category. There are many ways to disyllabify monosyllabic words: (a) suffixes are added to monosyllabic roots, a method mainly applied to nouns, as illustrated in the above paragraph; (b) monosyllabic words are juxtaposed with synonyms, a device used in all syntactic categories, such as yì → yìsi “meaning” and zhù → zhùshì “annotate”; (c) monosyllabic words are replaced by newly created disyllabic words – that is, earlier monosyllabic words are abandoned, such as yuē → jiǎnyào “concise” and wù → lǐjiě “understand”; (d) monosyllabic words are lengthened via reduplication, such as gē → gēgē “elder brother” and gāng → gānggāng “just minutes ago.” Another major source of disyllabic words is the lexicalization of two monosyllabic syntactic constituents that are linearly adjacent to each other. Almost all aspects of the basic structure of a sentence can be found in the formation of disyllabic words (Zhu 1982: 14–20): (a) (b) (c) (d) (e)

Subject + predicate, for instance, 冬至 dōng-zhì (winter-arrive) “the winter solstice”; 年輕 nián-qīng (age-light) “young.” Verb + object, for instance, 主席 zhǔ-xí (control-seat) “chair”; 動員 dòngyuán (move-member) “mobilize.” Modifier + head, for instance, 飛機 fēi-jī (flying-machine) “airplane”; 重視 zhòng-shì (heavily-see) “value.” Verb + resultative, for instance, 改良 gǎi-liáng (change-good) “improve”; 說明 shuōmíng (say-clear) “explain.” Co-ordination, for instance, 重疊 chóng-dié (overlap-fold) “reduplicate”; 買賣 mǎi-mài (buy-sell), “business.”

The English words in the parentheses are the literal translations of the preceding Chinese compounds, keeping the same order of the two morphemes of each compound. All of these compounds have now become lexicalized. Their meanings cannot be derived from the literal meanings of the two morphemes of every compound. These phenomena have motivated some scholars, such as Zhu (1982: 32–33), to claim that the most prominent feature of Chinese grammar is the identity between sentence structure and the formation

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

86

of compound words. However, as we just saw, some structures are found only in the formation of compound words, such as “root + suffix,” which does not exist in sentential structures. On the other hand, many major structures at the sentence level are absent from compounds, such as the disposal construction, verb copying, and topicalization. Our central concern is the influence of the tendency toward disyllabification on syntax rather than lexicalization itself. We want to draw attention to the fact that under the influence of the tendency toward disyllabification, two monosyllabic words that are adjacent within a syntactic structure are subject to compounding into a disyllabic item. Except for type (d), whose internal structure is “verb + resultative,” all of the other types of sentence structure providing sources for disyllabic words have been in the language since the beginning of its documented history. The “verb + resultative” compounds are merely one by-product of the development of the resultative construction, which was advanced by the disyllabic tendency. Now let us consider the development of the resultative construction. As Zhu (1990: 240) pointed out, “verb + resultative” phrases emerged and increased rapidly within the period from the sixth century AD to the thirteenth. Toward the end of this period, the resultative construction as a syntactic structure was firmly established. The development of “verb + resultative” phrases spanned a long time, and thus it is difficult to determine the exact date of the innovation of this construction. According to Jiang (2005b: 182), different types of “verb + resultative” phrase were introduced into the language at different times. Although the dating of the first appearance for each type of verb– resultative phrase is difficult to determine, there is no doubt that the resultative construction had become widely used by the tenth century AD. The resultative construction came into being in the period in which the tendency toward disyllabification was greatly accelerated.

5.5

Effect on the Resultative Construction

Now we discuss how the tendency toward disyllabification influenced the establishment of the resultative construction. When the tendency toward disyllabification developed up to a critical point where the disyllabic unit had become the basic phonological representation of the lexicon, it was able to exert its effect on the syntax. Influenced by the disyllabic tendency, two monosyllabic words were likely to be assigned a single intonation unit, which potentially causes their boundary to be weakened or lost, if they frequently occur in a context where they are adjacent to each other. This process produces three types of outcome – verb compounds, morphological markers, and a syntactic construction. Following the hypothesis stated in the paragraph above, the steps of the grammaticalization of the resultative construction are as follows.

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Effect on the Resultative Construction Step 1.

Step 2.

Step 3.

87

In Medieval Chinese, there existed a widely used separable resultative construction, a direct consequence of the disappearance of the VP connective ér at that time. The abstract structure of the separable resultative structure is as follows: V + XP + R (XP = object, adverb, or negative). If the verb and the resultative were both monosyllabic, they tended to be assigned a single intonation unit in the context where XP was absent because of the tendency toward disyllabification. Consequently, the boundary between the verb and the resultative was weakened or even lost, and the verb and the resultative were then bound into one syntactic constituent or even a compound word. The combined force of individual “verb + resultative” phrases gave rise to a new syntactic pattern – the resultative construction. When established as an independent syntactic structure, its elements could be multisyllabic; in other words, some “verb + resultative” phrases could have more than two syllables at a later stage.

Now let us first use data from other dialects to test the hypothesis about the relationship between the tendency toward disyllabification and the establishment of the resultative construction. As their phonological systems are more complex than those of the northern dialects, the southern dialects have much fewer disyllabic words (i.e. a weaker disyllabic tendency) than the northern dialects (e.g. standard Mandarin). According to the voluminous reports of Huang (1996: 733–738), the development of the resultative construction in many southern dialects fell far behind that of the northern dialects: the reanalysis of the verb and the resultative has not yet happened or may be halfway to completion, and the verb and resultative in these southern dialects can still be separated by an object, adverb, or negative, reminiscent of the separable resultative structure in Medieval Chinese, as shown in the following examples. (4)

烧伊酥。(上海方言) Sɔ i su. toast it crisp “Make it crisp by toasting.” (Shanghai dialect, Huang 1996: 733)

(5)

晒伊干。(上海方言) Sa i kø. shine-upon it dry “Make it dry by exposing it to sunshine.” (Shanghai dialect, Huang 1996: 733)

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

88

Table 5.4 Separable resultative constructions in dialects Areas

Southern dialects

Northern dialects

Dialects

Shanghai, Changsha, Xiangxiang, Rucheng, Jinhua, Lingxiang, Kejia, Meizhou, Guangxi, Guangzhou, Wenzhou, Xianju, Ningbo, Yangjiang 14

Weifang, Huanggan, Ruqi, Luoshan, Sichuan, all Mandarin dialects 5

Total

(6)

你打渠唔過。 (廣州方言) Nei ta kʰøy mou-kuɔ. you beat he/she not-exceed “You cannot beat him.” (Guangzhou dialect, Huang 1996: 733)

In the above two examples from the Shanghai dialect, the verbs and resultatives are separated by pronominal objects. These pronominal objects can also be attached to the whole verb–resultative phrase. For instance, (5) can be turned into Sa kø i “sunshine-dry it.” This shows that the Shanghai dialect is still in an intermediate stage in the development of the separable resultative structure into the resultative construction. Table 5.4, based on Huang (1996: 733–738), offers an overall picture of the existence of the separable resultative construction. Three of the five northern dialects that still preserve the separable resultative construction are located at the boundary between the northern and southern dialectal families. Thus we assume that the weaker tendency toward disyllabification in the southern dialects is responsible for the slow development of the resultative construction. One advantage of our hypothesis is that it can explain the unbalanced evolution of grammar between the northern and southern dialects without resorting to the theory of language contact. The northern dialects have developed many new grammatical devices since the tenth century AD, such as aspect markers, the disposal construction, the verbcoping construction, and a new structure of topicalization. However, many southern dialects lack these new grammatical devices or have similar devices halfway to the completion of their development. Li (1987) attributed these cross-dialectal differences to language contact. From the third century AD to the nineteenth, China was conquered and governed by nations whose languages belonged to the Altaic family (verb-final languages). Superficially, quite a few of the new grammatical constructions in Chinese serve to move a patient argument to preverbal position. For example, the disposal construction produces seeming verb-final clauses. However, a careful examination reveals that the predicates in these constructions are not truly verb-final and that the predicates must be

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Fusion of Verb and Resultative

89

verb–resultative phrases. The reanalysis of the verb and resultative causes the originally intervening object (i.e. the patient of the verb) to appear elsewhere in a sentence. Some of these patient arguments must occur in preverbal position owing to certain semantic and/ or syntactic constraints. Under our hypothesis, the cause–effect chain has a “stronger tendency toward disyllabification > fusion of the verb and resultative > increase of preverbal patients.” In other words, the faster development of the northern dialects is due to a language-internal force rather than an external one, namely language contact.

5.6

Fusion of Verb and Resultative

The phonological unit, namely the disyllabic chunk, plays an important role in motivating a compounding or fusion of two lexical items. On the one hand, many examples of fusion represent the development of an unambiguous phonological unit via the loss of a boundary separating two morphemes. How to form a phonological unit is largely determined by the phonological system of a particular language. In Chinese, as a tone language, fusion often causes the second morpheme to lose its segmental materials and especially its tone value. On the other hand, the basic phonological unit of a language might vary as its phonological system evolves; for example, Chinese has changed its basic phonological unit from monosyllabic in Old Chinese to disyllabic since Medieval Chinese. When disyllabic units become the basic phonological representation for the lexicon, two monosyllabic words are highly subject to fusion if they frequently co-occur in a certain context. The tendency toward disyllabification, which gained its highest momentum in Medieval Chinese, enabled the fusion of the verb and the resultative. Now we turn to an empirical defense of the hypothesis.

5.6.1

Shimura’s Hypothesis

Shimura (1967) proposed a hypothesis that the tendency toward disyllabification was responsible for the establishment of the resultative construction. He elaborated the process as follows (Jiang 2005b: 187). Step 1.

(7)

The resultative construction developed from two types of serial verb construction, as illustrated in (7) and (8):

Co-ordinate construction. 射而殺之。 (韓非子 似類) Shè ér shā zhī. shoot and kill he “(Someone) shot and killed him.”




(Han Fei Zi, Si Lei, 300 BC)

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

90 (8)

Separable resultative construction. 吹我羅裳開。(子夜四時歌) Chuī wǒ luō-shang kāi. blow my silk-skirt open “(The wind) blew my silk skirt open.” (Zi Ye Si Shi Ge, AD 450)

Step 2. Pairs of verbs became compounded through idiomatization (e.g. mièjìn “kill-entire”). The second verb of the compound was detransitivized. For example, dǎ-shā “beat-kill,” two transitive verbs, became dǎ-sǐ “beat-die” at a later stage. In this example, the second verb was detransitivized from shā “kill” to sǐ “die.” Step 3. Two serial verbs became idiomatized or fixed; for example, miè-jìn behaved like a single lexical item at a later age. Finally, the two verbs lost their original syntactic equality and the second verb was detransitivized. As a result, the “Vtr + Vtr” construction developed into a “Vtr + Vintr” one (e.g. dǎ-shā “beat-kill” = dǎ-sǐ “beat-die”) through the detransitivization of the second transitive verb. Shimura portrayed the large-scale compounding of serial verbs as a reflection of disyllabification. However, he neither argued for this insight nor defended it empirically. He correctly asserted that in the beginning, a particular fusion of two individual verbs was a lexical matter – compounding. However, we disagree with him on the source of the resultative construction. Shimura considered both the verb co-ordination and the separable resultative structure to be sources of the resultative construction. In our view, the verb co-ordination phrase, which typically consisted of two transitive verbs, was not a source but a competing form for the resultative construction. The two transitive verbs usually become lexicalized after the connective ér becomes optional and finally lost. The fusion of the two parallel transitive verbs of the verb co-ordination produced compound verbs, which was a lexical matter rather than a syntactic matter, as exemplified in (9): (9)




郡國守相明檢查之。 (曹操 田租户調) Jùn-guó shǒuxiāng míng jiǎn chá zhī. country prime-minister clearly choose examine they “The prime ministers of our countries should clearly choose and examine them.” (Cao Cao, Tian Zu Hu Diao, AD 200)

In (9), jiǎn “choose” and chá “examine” are co-ordinate verbs sharing an object, but they have become a verb compound in Contemporary Chinese, meaning “check” or “criticize.” In this sense both are bound morphemes.

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Fusion of Verb and Resultative

91

When the “verb + resultative” phrase was firmly established as a stable syntactic pattern, this verb co-ordination construction was gradually ruled out because of its structural incompatibility with the resultative construction. As a result, the verb coordination of Old and Medieval Chinese eventually disappeared from Chinese. The resultative construction was derived only from the separable resultative construction whose second verbal elements were mostly intransitive elements. There was no evidence that the second verb of the verb co-ordination ever underwent detransitivization. However, some uses of the verb co-ordination were replaced by VR phrases. For instance, dǎ-shā “beat-kill,” a verb co-ordination consisting of two transitive verbs, which was widely used before the tenth century AD, was often followed by an object. It was replaced by the verb–resultative phrase dǎ-sǐ “beat-die,” a transitive verb plus an intransitive element, when the resultative construction was firmly established and the verb co-ordination was no longer needed.1

5.6.2

A Criterion for the Resultative Construction

We need to set up a formal criterion to identify the earliest appearances of the resultative construction – a reanalysis of the verb and resultative. According to Langacker’s (1977: 58) definition of “reanalysis,” such a change is invisible at its initial stage. As he defined it, reanalysis is “a change in the structure of an expression or class of expressions that does not involve any immediate or intrinsic modification of its surface manifestation” and involves boundary creation, shift, and loss. This view was challenged by Harris and Campbell (1995: 61) and by Heine et al. (1991: 216). The special development of the resultative construction provided us with a reliable formal criterion for judging when a verb–resultative reanalysis took place. The resultative construction was derived from a so-called separable resultative structure in which the verb and resultative represent two syntactic constituents and thus could be separated by other words at that time. The formula is as follows: (10)

V + XP + R; XP = object, adverbs, or/and negative

When the verb and the resultative become fused, an overt change certainly happens: the fusion of the verb and the resultative can no longer be separated by the words originally occurring in the XP position. These squeezed-out materials have to be expressed in another position of a sentence: the patient argument of the verb usually moved to postverbal position as the object of a whole verb–resultative phrase and adverbs or negatives typically moved to preverbal position. If such a formal change was identified

1

In some dialects, such as the Wu dialect, shā “kill” was often used as a resultative, e.g. dǎ-shā. It may just preserve an old use because this verb–resultative phrase usually had an object in Medieval Chinese.

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

92

for an individual “verb + resultative” phrase, it is safe to say that a fusion of the verb and the resultative had already happened, though the reanalysis presumably happened earlier than the overt change. Logically, a verb–resultative pair could not be fused and separable at the same time.

5.6.3

Degrees of Fusion

Grammatical changes are characterized by both gradualness and continuity (Hopper and Traugott 2003: 233). It is difficult to imagine that the verb and the resultative suddenly became fused without an intermediate stage. A reasonable hypothesis is that the fusion of the verb and the resultative is a matter of degree. At the intermediate stage, the boundary should be perceptible in some ways because it was merely weakened rather than entirely lost. For the sake of our analysis, we set up scales of fusion with only two degrees. (a) A low degree of fusion refers to an intermediate stage in which the verb and the resultative have started to become fused but have not reached the completion point of the development. At this time, the verb and the resultative no longer allow any materials to intervene between them, but the verb–resultative phrase cannot have a nominal object, indicating its phrasal status. The originally intervening materials, such as patient objects and adverbs, are moved to preverbal position. (b) A high degree of fusion refers to the stage in which the verb and the resultative become fused into a single syntactic constituent. In addition to the features of low fusion, namely no intervening material, the verb– resultative phrases that have reached a high degree of fusion can precede a nominal object. To understand the features of the different degrees of fusion, we need to examine the structure of Chinese and the development of the resultative construction. In Chinese, an SVO language, a transitive verb typically precedes its patient argument. In Medieval Chinese, as pointed out previously, verb co-ordination phrases consisting of two or more transitive verbs could be followed by an object, and each of those transitive verbs had to bear an “action–patient” relationship with the object. Intransitive verbs (including adjectives) were not allowed in co-ordinated verb phrases followed by objects. This grammatical rule was gradually broken by the resultative construction, where the second elements were generally intransitive (i.e. intransitive verbs or adjectives) but the whole construction could have an object. We assume that only under the condition of the resultative losing its status as an autonomous word and fusing with its preceding verb into a single verb could the

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Fusion of Verb and Resultative

93

related verb–resultative phrase precede a patient object. Recall that before the fusion the resultative as an intransitive element could appear only after the object. According to our definition, a verb–resultative phrase has accomplished the high degree of fusion when the phrase behaves as a pure verb, namely when it precedes an object. However, if the boundary of the verb and the resultative is merely weakened, their collocation is still strongly phrasal rather than compound-like, and the “verb + resultative” phrase is thus unable to precede an object.

5.6.4

Number of Syllables

After the tenth century AD, the resultative construction developed very rapidly. The number of syllables of early “verb + resultative” phrases reveals how the tendency toward disyllabification advanced the formation of the resultative construction. According to Zhu’s (1990) comprehensive examination of Zhu Zi Yu Lei, a text of the twelfth century AD with more than two million characters, the resultatives were all monosyllabic. Obviously, to form a disyllabic verb–resultative phrase, the resultative must be monosyllabic. Moreover, having checked all verb–resultative phrases provided by Zhu, we find that the verbs are also exclusively monosyllabic. Thus the logical conclusion is that verb–resultative phrases in the twelfth century AD were necessarily disyllabic. On the other hand, those verb–resultative phrases with multisyllabic resultatives underwent only a low degree of fusion because the whole verb–resultative phrase had more than two syllables. This indicates that if the phonological representation of a verb–resultative phrase fit the disyllabic prosodic unit, the phrase would quickly become fused. In fact, the number of syllables in verb–resultative combinations influenced the speed of their development toward the resultative construction. This point is illustrated in the following two examples: (11)

如今都教壞了後生。 (朱子語類卷一〇八) Rújīn dōu jiào-huài le hòushēng. now all teach-bad PERF student “Now all students were taught to be bad.” (Zhu Zi Yu Lei, Chapter 108, AD 1200)

(12)

雖上司約束分明。 (朱子語類卷一〇八) Suī shàngsi yuēshù fēnmíng. even superior discipline clear “Even the superior’s regulations are clear.” (Zhu Zi Yu Lei, Chapter 108, AD 1200)

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

94

The verb–resultative phrase in (11), jiào-huài “teach-bad,” a disyllabic unit, had reached a high degree of fusion at that time, as shown by the fact that it precedes an object. In contrast, the verb–resultative phrase in (12) has four syllables – yuēshù (verb) plus fēnmíng (resultative) “discipline-clear”; thus it underwent only a low degree of fusion, as shown by the patient argument shàngsī “superior” being topicalized to preverbal position. Even in Contemporary Chinese, verb–resultative phrases not fitting the disyllabic unit still remain largely at the low degree of fusion; for example, they cannot have objects. The effect of the number of syllables on the reanalysis of the verb and the resultative can also be seen from the speed of the development of the resultative construction. The fusion of the verb and the resultative happened mostly in the time range from the sixth century AD to the fifteenth. Some took place earlier, some later. The timing was determined largely by the frequency of the co-occurrence of individual “verb + resultative” pairs and by the semantic features of specific resultatives. As a whole, the separable resultative construction had disappeared from the language before the fifteenth century AD. However, a small group of such examples survived as late as the sixteenth century AD. According to our investigation of two texts composed around the fifteenth century AD, Shui Hu Zhuan and Jin Ping Mei, the resultatives of the separable structure that survived up to this point were without exception multisyllabic, as illustrated in (13) and (14): (13)

王婆收拾房裡乾淨了。 (水滸傳二十四回) Wáng Pó shōushí fánglǐ gānjìng le. Wang Lady organize room clean PERF. “Lady Wang has cleaned up the room.” (Shui Hu Zhuan, Chapter 24, AD 1400)

(14)

揩抹得桌兒乾淨。(金瓶梅四十五回) Kāimò de zhuōer gānjìng. wipe DE table clean “(She) wiped the table clean.” (Jin Ping Mei, Chapter 45, AD 1550)

The lagging development of those verb–resultative phrases with multisyllabic resultatives suggests that their fusion happened analogously to that of the earlier disyllabic verb–resultative phrases. Examples of the separable resultative structure had been entirely ruled out not long after the sixteenth century AD. In Contemporary Chinese, some of them are expressed by a VRO sequence (e.g. cā-gānjìng zhuōzi “wipe-clear table”), but most are expressed by the bǎ construction (e.g. bǎ fángjiān shōushí-gānjìng “BA room organize clean”), a device that moves patient arguments to preverbal

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Fusion of Verb and Resultative

95

position. This is because the fusion of the verb and multisyllabic resultatives means that the originally intervening materials, particularly a noun phrase used as an object, have to appear elsewhere. The above analysis naturally leads to the conclusion that the tendency toward disyllabification played a key role in the development of the resultative construction. Within a verb–resultative phrase, if both words are monosyllabic, they are likely to be assigned to one phonological unit because of the effect of the tendency toward disyllabification. The assignment of the unambiguous phonological unit tends to motivate a verb–resultative phrase to undergo either fusion into a single syntactic constituent or lexicalization into a verb compound. Therefore the early appearances of the resultative construction were mostly disyllabic “verb + resultative” phrases. The same hypothesis can also explain why verb–resultative phrases with more than two syllables developed later and why most of them reached only a low degree of fusion.

5.6.5

The Potential Form

The potential form is one subtype of the resultative construction, with modal infixes that express the possibility of verbs realizing certain resultatives. Its formula is as follows: (15)

(a) Affirmative potential: V-de-R + Obj (b) Negative potential: V-bu-R + Obj

Superficially, the establishment of these potential forms involves fusion among three words (at least three syllables) and the loss of two boundaries – between the verb and de/ bu and between de/bu and the resultative, a case that seemingly contradicts the hypothesis of the tendency toward disyllabification. However, it actually does not contradict our hypothesis. Both the affirmative and the negative potential forms took two steps to accomplish their fusion, and the fusion at each step involved only two constituents. Without exception, the first step fused only two monosyllabic words into a disyllabic unit, and this fused disyllabic unit was then assembled with the third element. The two potential forms underwent the same process, though they grammaticalized through different pathways at different times. First, let us consider the development of the affirmative form. The infix -de- was derived from a full verb meaning “get” and “can” (cf. Section 13.3.4). According to Zhu (1990), V-de without exception occurred before an object if there was one during the period from the seventh century AD to the ninth. That is, at this time, V-de was first fused as a single constituent under the influence of the tendency toward disyllabification. After the tenth century AD, the combination V-de (O) R became gradually more

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

96

frequent, with the resultatives being mostly monosyllabic adjectives or monosyllabic intransitive verbs, as illustrated in the following examples: (16)

公不曾看得那物事出。 (朱子語類卷一一四) Gōng bùcéng kàn-dé nà wùshì chū. you not find-DE that thing out “You didn’t find out the thing.” (Zhu Zi Yu Lei, Chapter 114, AD 1200)

(17)

不知怎生唤得它醒? (朱子語類卷九七) Bù zhī zěnshēng huàn-dé tā xǐng? not know how call-DE he awake “I don’t know how to call him awake.” (Zhu Zi Yu Lei, Chapter 97, AD 1200)

Clearly, at the time of the twelfth century AD, the phrase V-de was a disyllabic syntactic constituent and the resultative was another independent constituent, and they could be separated by objects or other elements. For instance, in (16), kàn-dé (V-de) “able to find” and chū (resultative) “out” are separated by nà wùshì (the object of the verb) “that thing.” The affirmative potential form at this stage was in fact an instantiation of the separable resultative structure. After the thirteenth century AD, the “V-de-R O” construction started to occur, suggesting that another reanalysis between the V-de and the resultative had taken place. That is, the three elements of the affirmative potential form constituted a single syntactic constituent. At this time, the tone of the infix -de- was removed, and its final was also reduced to a schwa; thus the infix represented merely a reduced syllable. Considering the phonological reduction of the infix, we know that the fusion of V-de and the resultative still complied with the basic phonological unit. After the establishment of the affirmative potential, the verb and the resultative were no longer limited to being monosyllabic, but monosyllabic verbs and resultatives were still used more freely in the form. The development of the negative form took similar steps, but through a different pathway. From a synchronic point of view, the negative form is derived from its corresponding affirmative simply by replacing the infix -de- with the negative bu, but this does not reflect the diachronic process. Before the thirteenth century AD, unlike -de-, which was bound to the preceding verb, bu always went with the resultative to form a disyllabic unit, both of which were often separated from the preceding verb by objects or by adverbs. The abstract structure is as follows: (18)

V XP bu-R

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Fusion of Verb and Resultative

97

In line with our analysis, it is in this construction that the negative bu and the resultative were first fused into a single constituent via the loss of their boundary, as shown in the following: (19)

便攔他不住。 (朱子語類卷一一六) Biàn lán tā bù-zhù. then block he not-stay “Then cannot block him.” (Zhu Zi Yu Lei, Chapter 116, AD 1200)

(20)

說自家意思不盡。 (朱子語類卷九) Shuō zìjiā yìsi bù-jìn. state self thought not-complete “(Someone) didn’t completely state their own thought.” (Zhu Zi Yu Lei, Chapter 9, AD 1200)

After the thirteenth century AD, V-bu-R examples started to occur and eventually replaced the old separable form around the fifteenth century AD. Examples (21) and (22) are among the earliest instances of this type. (21)

便記不得細注字。 (朱子語類卷一二一) Biàn jì-bù-dé xì-zhù zì. then memorize-not-DE small-annotate character “Then cannot memorize those small annotating texts.” (Zhu Zi Yu Lei, Chapter 121, AD 1200)

(22)

他其實咽不下玉液金波。 (王實甫 西廂記) Tā qí-shí yàn-bùxià yù-yè jīn-bō. she in-fact drink-not-down jade-fluid gold-wave “She in fact cannot drink wine.” (Wang Shi, Fu, Xi Xiang Ji, AD 1300)




Interestingly, even in some modern dialects, we can still see disyllabification at work in creating the potential form. As just pointed out, historically, the negative bu and the resultative first became compounded, but in some dialects bu and the verb first formed a syntactic constituent, as exemplified in (23) and (24): (23)

煮不飯熟。 (湘鄉方言) Tɕy-pu ɣua ɕiʊ. cook-not rice done “Cannot cook the rice well.” (Xiangxiang dialect)

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

98 (24)

聽不英語懂。 (湘鄉方言) Tʰiɒŋ-pu iɛn-ȵy toŋ. listen-not English understand “To be unable to understand spoken English.” (Xiangxiang dialect)

We also notice that the affirmative potential form of these dialects is still at the stage of the separable resultative structure, namely that V-de and R can still be separated. This leads us to believe that the “V-bu O R” construction of these dialects was created analogously to the affirmative form. Compared to standard Mandarin, in these dialects the potential form has not reached the completion point of its development. Either way, the development of the potential form is constrained by the disyllabic phonological unit. Based on the above analysis, we can draw a generalization from the commonalities of the development of the two potential forms. From a synchronic point of view, the potential form involves two boundary changes (loss), which seems to contradict the constraint of the tendency toward disyllabification. However, a diachronic examination reveals that the potential form underwent two steps and that, at each step, only two elements were involved in fusion – two syllables at the first step and two syllables plus a reduced syllable at the second step. Specifically, at the first step, two of the three elements, V-de in the affirmative form and bu-R in the negative form, were fused into a single constituent via boundary loss, which formed a disyllabic unit. Then, at the second stage, each of these first fused constituents was fused with another element, the resultative in the affirmative form and the verb in the negative form. We propose that these trisyllabic forms took two steps to become fused into a single constituent, a process that was determined largely by the basic intonation unit, which was disyllabic.

5.7

Constraint of the Number of Syllables

Having seen the effect of disyllabicity on the grammaticalization of the potential form, we now turn to several concrete cases that support the hypothesis about the influence of the tendency toward disyllabification on the formation of the resultative construction. The resultative construction as a syntactic pattern came from the combined power of many developments of individual verb–resultative phrases. (a)

V- sǐ “action-die”

As stated previously, there was a restrictive rule in Medieval Chinese: only transitive verbs could share an object. Constrained by this principle, sǐ “die” as an intransitive verb had to occur after the object at that time when used as a resultative. The formula is either “V + O + sǐ ” or “O + V + sǐ,” as illustrated below:

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Constraint of the Number of Syllables

99




(25)

發兵攻殺其王及漢使者皆死。(漢書 嚴朱吾丘主父傳) Fā bīng gong-shā qí wáng jí Hàn shǐzhě jiē sǐ. send troops attack-kill the king and Han envoys entirely die “(He) sent troops to attack and kill the king and the envoys of the Han Dynasty who entirely died.” (Han Shu, Yan Zhu Wu Qiu Zhu Fu Zhuan, AD 100)

(26)

百餘人炭崩盡壓死。 (論衡 明義) Bǎi yú rén tàn bēng jìn yā sǐ. hundred more people charcoal explode entirely press die. “More than one hundred people died when the charcoal exploded.” (Lun Heng, Ming Yi, AD 100)




In (25) the patient argument intervenes between the verb and sǐ, and in (26) the patient argument occurs before the predicate. Superficially, the constraint mentioned above was gradually broken and gave way to this new form: “V-sǐ + O.” How could this change have happened? We assume that V and sǐ first became fused as a single constituent, thereby behaving as a single verb. Once fused into a compound-like unit, the V-sǐ phrase could be followed by a patient object without necessarily violating the previous constraint, which applied only to verb co-ordination. According to Mei (1991), the earliest “V-sǐ + O” examples can be dated to as early as the fifth century AD. Only a single example of this type has been found in texts before the tenth century AD, hence its reliability is questionable. However, there is no doubt that “V-sǐ + O” examples started to occur after the twelfth century AD and eventually became a new paradigmatic form. The following are the only two such examples found around the twelfth century AD: (27)

秦時六月皆凍死人。 (朱子語類卷七十九) Qín-shí liù-yuè jiē dòng-sǐ rén. Qin-time June always cold-die people “In the Qin time there were always people who died of frost in June.” (Zhu Zi Yu Lei, Chapter 79, AD 1200)

(28)

蜀中今年殺死了系名色人。 (朱子語類卷一三二) Shǔ zhōng jīn-nián shā-sǐ le xìmíngsè rén. Shu within this-year kill~die PERF Ximingse people “In the Shu area, (they) killed people from Ximingse this year.” (Zhu Zi Yu Lei, Chapter 132, AD 1200)

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

100

Table 5.5 The distribution of V-sǐ + O in Shui Hu Zhuan

Monosyllabic V Disyllabic V

V-sǐ

V-sǐ + O

56 21

143 0

Note that both of these earliest V-sǐ phrases are disyllabic. Until the sixteenth century AD, V-sǐ phrases that could precede a patient object were exclusively disyllabic. That is, the verb had to be monosyllabic, as shown in Table 5.5. Table 5.5 shows that the number of syllables in V-sǐ phrases crucially determines their syntactic behavior. Only monosyllabic verbs plus the resultative -sǐ could be followed by a patient object. Of the 199 tokens with monosyllabic verbs, 72 percent are followed by an object. In contrast, none of the twenty-one examples of multisyllabic verbs are followed by a patient object. In fact, most of their patient objects simply do not appear, as illustrated in the following examples: (29)

定是中毒身死。 (水滸傳二十六回) Dìng shì zhòng-dú shēn sǐ. must be poison physically die “He must have been poisoned to death.” (Shui Hu Zhuan, Chapter 26, AD 1400)

(30)

由你們碎屍而死。 (水滸傳三十四回) Yóu nǐmen suì-shī ér sǐ. let you cut-body and die “Let you cut my body and kill me.” (Shui Hu Zhuan, Chapter 34, AD 1400)

The two verbs in (29) and (30) are themselves disyllabic compounds, each of which forms a unit longer than a disyllabic one. Both objects (i.e. the victims of “poison” and “cut-body”) are absent, probably because they could be determined from the context. This is one common strategy for arranging the patient noun of those “verb + resultative” phrases that could not have an object.2 The resultative part also indicates how the number of syllables in verb–resultative phrases influenced the speed of the development of the resultative construction. In the sixteenth century AD, every V-sǐ phrase that formed a disyllabic unit had reached the high degree of fusion, e.g. they could have objects and allowed no intervening material. 2

Another strategy is to put patients in preverbal position, used as the subject or topic, as we saw in the earlier examples.

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Constraint of the Number of Syllables

101

However, if the resultative part was made multisyllabic by adding a modifier to sǐ, the verb and the resultative were still separated by an object, as illustrated in (31): (31)

輕則打你半死。(水滸傳三十回) Qīng zé dǎ nǐ bàn-sǐ. gently then beat you nearly-die “Even with a light beating, (I) will beat you nearly dead.” (Shui Hu Zhuan, Chapter 30, AD 1400)

The resultative bàn-sǐ “nearly dead” is a disyllabic unit. Therefore the verb and the resultative here are still separated by a pronominal object, a sign that the boundary between the verb and its multisyllabic resultative was still preserved at this time. The process of the reanalysis between the verb and the resultative sǐ clearly shows how the phonological process – disyllabification – operated in the formation of this case. (b)

V-le “V-completion verb/perfective aspect”3

Some inflections resulted from the further grammaticalization of certain resultatives (for details, see Section 13.3.1). One such change is the perfective marker -le, which derives from the ordinary verb liǎo “complete” in the resultative position. Before the tenth century AD, like other intransitive items used as resultatives, liǎo could appear only in the separable resultative construction. The abstract form is “V O liǎo,” as illustrated in (32). This use indicated the completion of an action. After the tenth century AD, however, the “V-le O” form (-le is the reduced form of -liǎo) started to occur, as exemplified in (33): (32)

六祖說偈已了。 (六祖壇經) Liù Zǔ shuō jì yǐ liǎo. Liu Zu explain Buddhist chant already complete “The Sixth Patriarch has completed explaining the Buddhist chant.” (Liu Zu Tan Jing, AD 700)

(33)

南朝已應付了三處。(乙卯入國奏請) Nán cháo yǐ yìngfù le sān chù. Southern government already answer PERF three places “The Southern government has already given three places to you.” (Yi Mao Ru Guo Zou Qing, AD 1050)

The resultative liǎo “complete” in (32) is a full verb that is separated from the main verb shuō “explain” by an object jì “Buddhist chant” and adverb yǐ “already.” However, -le in 3

For a fuller discussion of the mechanism of the development of the aspect marker -le, please see Section 13.3.1.

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Disyllabification

102

Table 5.6 Fronting of adverbs in Dun Huang Bian Wen

V + OBJ + Adv + liǎo Adv + V + OBJ + Adv + liǎo Disyllabic V + Adv + liǎo Adv + disyllabic V + liǎo Monosyllabic verb + Adv + liǎo Adv + monosyllabic verb + liǎo

Tokens

Percentage of fronting examples

401 171 69

2 6 60

(33) is a perfect marker, and V-le precedes an object. The diachronic movement of the “completion” verb from after the object to before the object indicates its status change – from a verb to a grammatical marker, a process that took several hundred years. Here, we focus on how the number of syllables in the verb influences the fusion of the verb and the resultative liǎo. Table 5.6 shows the results of marker investigation in Dun Huang Bian Wen, a text of the eighth century AD that was produced not long before the emergence of the aspect marker -le. Here we use adverb fronting as a formal criterion to judge where the reanalysis between the verb and the resultative first happened. A symptom of the V-liǎo fusion at the initial stage is that those adverbs that originally intervened between the verb and the resultative liǎo were fronted to before the whole V-liǎo phrase. When monosyllabic verbs appear together with the resultative liǎo, of the fifteen examples without an intervening object, 60 percent have the adverb fronted. Considering that these adverbs originally appeared only between the verb and the resultative liǎo, we can say that the boundary between the verb and the resultative liǎo first became weakened in the case of the verb being monosyllabic. In contrast, of the eighteen examples with disyllabic verbs, only one has the adverb fronted, merely 6 percent of the total. In other words, in 94 percent of examples, the verb and the resultative liǎo were still separated by adverbs, a feature showing that the boundary between multisyllabic verbs and the resultative liǎo was still clear. The positioning of adverbs in the separated V-liǎo phrases also suggests the condition of the fusion of the verb and the resultative liǎo. Before the tenth century AD, the patient argument, if present, had to appear between the verb and the resultative liǎo. When the verb and the resultative liǎo were separated by a patient object, the adverbs remained in their original position even with monosyllabic verbs, as illustrated in (34) and (35): (34)




仙人相太子已了。 (敦煌變文 太子成道) Xiān-rén xiàng tàizǐ yǐ liǎo. celestial-being examine Crown-prince already complete “The celestial being has already examined the Crown prince.” (Dun Huang Bian Wen, Tai Zi Cheng Dao, AD 800–1000)

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Constraint of the Number of Syllables (35)

103




子玉讀書已了。 (敦煌變文 唐太宗入冥記) Zǐ Yù dú shū yǐ liǎo. Zi Yu read book already complete “Zi-yu has already read the book.” (Dun Huang Bian Wen, Tang Tai Zong Ru Ming Ji, AD 800–1000)

A “V O liǎo” phrase, as an (at least) trisyllabic unit, is unlikely to become fused; thus adverbs still intervene. This is related to another condition for the fusion of the verb and resultative: they must occur in a construction in which they are adjacent to each other. Once again, the development of the perfective -le demonstrates how the phonological unit affects a grammaticalization process. Other factors are also involved in the development of the perfective. The influence of the tendency toward disyllabification can be seen in two major areas: (a) in the initial stage of the development of the resultative construction, “verb + resultative” phrases were typically disyllabic, representing the basic phonological unit; (b) if the verb and the resultative are monosyllabic, they tend to be more tightly bonded, thereby behaving as a single verb – for example, they can occur quite freely in VRO constructions. At a later stage this syllable constraint was loosened, although it was still at work. For example, some verb–resultative phrases consisting of disyllabic verbs and/or resultatives could also behave as disyllabic verb–resultative phrases, e.g. they could be followed by an object. This is the functional extension of disyllabic verb–resultative phrases through analogy. Let us examine some examples of Contemporary Chinese in which verbs or resultatives are multisyllabic: (36)

disyllabic verbs 我理解錯了那個問題。 (現代漢語) Wǒ lǐjiě-cuò le nà-gè wèntí. I understand~wrong PERF that-CL problem “I misunderstood the problem.” (Contemporary Chinese)

(37)

disyllabic resultatives 我才弄明白他的意思。 (現代漢語) Wǒ cái nòng-míngbái tā de yìsi. I just make~clear he GEN intention “I just understood his intention.” (Contemporary Chinese)

There is a cyclic interaction between the tendency toward disyllabification and the emergence of the resultative construction. The tendency toward

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

104

Disyllabification

disyllabification developed as compensation for the simplification of the phonological system. It was manifested in all syntactic categories. Disyllabic units were formed in two major ways: the first was to add an extra syllable to a monosyllabic root through affixation; the second was to freeze two adjacent and monosyllabic constituents in a syntactic construction such as “subject + predicate” compounds (e.g. nián-qīng “age-light,” “young”). However, the emergence of the resultative construction is much more important than the other syntactic constructions that provided sources for disyllabic words. The other syntactic constructions have been in the language since the beginning of their history and have no syllable constraint. They produce disyllabic words by accident; consequently a limited number of disyllabic words are derived from these syntactic constructions. However, the resultative construction as a new syntactic pattern is highly productive and consists of disyllabic units. Therefore the emergence of the resultative construction dramatically increased the number of disyllabic units, which in turn greatly reinforced the tendency toward disyllabification. The influence of the tendency toward disyllabification extends beyond verbal phrases. The emergence of the resultative construction and its syntactic consequences (for details, see related sections) all belong to changes in verbal phrases. Parallel to the development of the resultative construction are many appearances of morphological markers for nominal phrases, mainly including the following: (i) the establishment of the classifier system (cf. Chapter 16), (ii) the grammaticalization of the plural marker -men (cf. Section 19.3), (iii) the particle for clauses, and adjectival and genitive phrases -de (cf. Section 13.3.4), (iv) the nominal suffixes -zi, -er, and -tou (cf. Section 19.7). All of these markers were introduced into the language during roughly the same period when the resultative construction came into being. They share a common prosodic feature in that they all occupy the “unstressed” slot of the prosodic pattern “stressed + unstressed.” Some nominal morphology is clearly a response to the tendency toward disyllabification; for instance, the primary function of the three nominal suffixes in (iv) is to make disyllabic nouns out of monosyllabic roots (Wang 1989: 13‒17). The relationship between the other changes and the tendency toward disyllabification is worthy of further exploration. Disyllabification is a basic change that had a profound effect on the syntax, morphology, and formation of words in Mandarin Chinese. It plays a critical role in creating

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

Constraint of the Number of Syllables

105

the resultative construction and many verbal morphological markers. This phonological factor is only one of the factors responsible for the creation of the resultative construction. The following chapter will address how co-occurrences of the verb and the resultative advanced their fusion and the idiomatization of early verb–resultative phrases.

https://doi.org/10.1017/9781108921831.005 Published online by Cambridge University Press

6 Resultative Construction

6.1

Introduction

The grammar of Contemporary Chinese is characterized by ubiquitous occurrences of diverse variants of the resultative construction, which did not exist before Medieval Chinese. This chapter discusses the diachronic sources and the synchronic properties of the resultative construction from various angles, including a brief cross-linguistic comparison of similar phenomena; an investigation of the semantic, syntactic, and phonological properties of the resultative construction; and a consideration of its obligatory application in certain syntactic constructions. In the literature of general linguistics, the term “resultative” is used to refer to various phenomena. The definitions generalized from different languages are not entirely applicable to the Chinese case. First, Li and Thompson defined the resultative construction of Chinese as follows: “A two-element verb compound is called a resultative verb compound if the second element signals some result of the action or process conveyed by the first element” (Li and Thompson 1981: 54). This definition characterizes the two major features of the Chinese resultative. First, it is a verb compound consisting of two elements. Li and Thompson did not elaborate on the meaning of “verb compound.” For our present purposes, let us define verb compounds as follows: the two elements do not allow other material to intervene; namely, they bond together to form a single syntactic constituent, and this constituent behaves like a pure verb. For instance, it can precede a patient object as a simple transitive verb does (e.g. kàn-wán shū “read-finish book”). The term “verb compound” is somewhat misleading because it suggests that the Chinese resultative is a lexical matter and thus not productive. The following discussion will show that it is actually a syntactic pattern allowing many kinds of collocation. Second, the two elements stand in an “action–result” relationship, which distinguishes the resultative construction from ordinary verb serialization. The term resultative here can be understood only in a very broad sense that may refer to the state, degree, accomplishment, achievement, or effect of the action. The working definition of the resultative construction is based on that of Li and Thompson, with our slight modification.

106

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Introduction

107

English has a resultative construction somewhat similar to that in Chinese. Goldberg (1995: 193–194) identified some constraints on the English resultative construction as follows: (a) (b) (c) (d)

the two-argument resultative construction must have an (animate) instigator argument; the action denoted by the verb must be interpreted as directly causing the change in state: no intermediary time intervals are possible; the resultative adjective must denote the end point of a scale; the change in state must occur simultaneously with the end point of the action denoted by the verb.

The above features of the resultative construction in English are quite useful for understanding the corresponding Chinese phenomena. However, we would like to point out the following differences between the corresponding constructions of the two languages. First, as just mentioned, in Chinese, a verb–resultative phrase is similar to a compound, which can precede a patient object as other ordinary transitive verbs can. However, in English it cannot do so, although English has the same SVO word order as Chinese. The patient argument usually intervenes between the verb and the resultative. This difference is illustrated in (1) and (2): (1)

The dog barked the children awake.

(2)

老王叫醒了小張。 (現代漢語) Lǎo Wáng jiào-xǐng le Xiǎo Zhāng. Lao Wang call-awake PERF Xiao Zhang “Old Wang called Young Zhang awake.” (Contemporary Chinese)

In the English example above, the object “children” intervenes between the verb “bark” and the resultative “awake,” but in the Chinese example the verb and the resultative first form a constituent (i.e. “call-awake”) that is followed by the object. In fact, the word order of the English resultative is similar to that of the so-called separable resultative construction of Medieval Chinese from which the resultative of Contemporary Chinese was derived. One of our major tasks in this chapter is to explore how the resultative of Contemporary Chinese developed from the separable one of Medieval Chinese. Second, if Goldberg’s generalization is correct, the Chinese case also differs from its English counterpart in that in the Chinese case the change in state denoted by the resultative does not have to occur simultaneously with the end point of the action

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

108

denoted by the verb. For example, chī-pàng “eat-fat” means that somebody gets fat via eating (too much meat), but it is impossible for the state of being fat to occur simultaneously with the end point of the action of eating. Finally, as Goldberg (1995: 192) pointed out, resultatives in English are often collocated with particular verbs. For example, the action eat is most colloquial with the resultative sick. (3)

(a) He ate himself sick. (b) ?He ate himself ill/nauseous/full. (Goldberg 1995: 192)

However, in Modern Chinese the resultative construction is a syntactic pattern rather than a class of idiosyncratic expressions. As with other syntactic constructions, any “verb + resultative” collocation that makes sense is possible. This point can be seen from the following examples which illustrate some combinations of the verb chī “eat” and its possible resultatives: (4)

喫飽 喫腻 喫病 喫胖 喫窮 喫晕 喫瘦 喫累 喫吐

chī-bǎo “eat-full” chī-nì “eat-bored” chī-bìng “eat-sick” chī-pàng “eat-fat” chī-qióng “eat-poor” chī-yūn “eat-dizzy” chī-shòu “eat-thin” chī-lèi “eat-tired” chī-tù “eat-vomit”

Theoretically, any words referring to a state or property that can be caused by the action of eating can be used as a resultative with the verb chī.

6.2

The Syntax of the Resultative Construction

In the above analysis, we have already touched on some syntactic properties of the resultative construction. This section will deal with the issue more systematically.

6.2.1

Degree of Fusion

The following cline can be used to describe the syntactic and semantic features of different types of resultative construction. This is designed to characterize the different degrees of integrity of the two elements.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

The Syntax of the Resultative Construction

109

We do not intend to imply that there is a clear-cut boundary between consecutive categories. These three categories indicate major differences on the continuum. Now, let us elaborate on each category with concrete examples. (a) Syntactic collocation. By definition, the resultative construction is a syntactic pattern, and any two elements can thus occur in it if they hold an “action–result” relation. This construction is as productive as any other syntactic construction. This type is more of a loose syntactic organization than a compound verb, as illustrated in (5): (5)

學怕 看歪 研究窮 說漏嘴

xué-pà “study-fear” kàn-wāi “look-oblique” yánjiū-qióng “research-poor” shuō-lòuzuǐ “say-leak”

In some instances, the meanings of spontaneous collocations are difficult to understand without context. For example, (b) kàn- wāi “look-oblique” may be used when one is standing on a chair and hanging a picture on a wall and needs another person to help judge whether the picture is hanging straight. However, it turns out that the person’s judgment is not accurate and the picture hangs obliquely. Therefore the hanger may complain, Nǐ kàn-wāi le (lit. “You look (the picture) oblique”). (b) Verb + bound resultative. The two components – the verb and the resultative – co-occur so frequently that people view them as a single unit, and the resultative is cliticlike with a fairly general meaning that enables it to combine with numerous verbs. The components are highly integrated so that they behave like a single verb and thus can precede an object as an ordinary transitive verb can. What distinguishes this type from “verb + resultative” compounds is that the meaning can be derived from the meaning of the components plus an “action–result” relation. (6)

喫飽 看完 學會 洗净

chī-bǎo “eat-full” kàn-wán “read-finish” xué-huì “learn-master” xǐ-jìng “wash-clean”

The high frequency of the co-occurrence of those verb–resultative pairs is largely the result of the frequency of the verb and the fact that the verb has a most natural result; for instance, “eat” is one of the most frequent verbs and bǎo “full” is its most natural result. (c) Compound verb. Resultatives of this type behave like unanalyzable single verbs, and people cannot sense that their internal structure is “verb plus resultative” unless they have knowledge of Classical Chinese. Dictionaries usually provide entries for these

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

110

compounds. Moreover, their meanings cannot be derived from their components plus the “action–result” relation between the two morphemes, as illustrated in (7): (7)

看開 說明 拿定 抓紧

kàn-kāi shuō-míng ná-dìng zhuā-jǐn

Lit. see open = to understand Lit. say light = to explain Lit. take settle = to make up one’s mind Lit. grasp tight = to make the best of time

This group of verbs has the same syntactic properties as ordinary verbs. For example, they can be freely followed by an object. However, the number of this type is quite limited. What the above three types of resultative construction have in common is that they do not allow any intervening material, though this situation was exactly the other way around in Medieval Chinese. A clarification is necessary here. As we will find occasionally, the Chinese language also has a potential form with the formula V-de-R (affirmative) and V-bu-R (negative). Superficially, these forms are created by entering the infix -de- or -bu- into a “verb + resultative” phrase. However, this view finds no support in either diachronic process or synchronic function. From a functional point of view, the overall meaning is not that of the “verb + resultative” form plus a negative sense. Rather, the potential sense is added. Therefore it is better to think of the potential form as another independent grammatical device than as something derived from a regular resultative construction (for a fuller discussion, please see Section 13.3.4). In short, we do not think that the potential form is an exception to the generalization we have just made. All of the types of resultative that we have thus far seen are simple words. However, the resultative could in fact be a clause, and in this case the verb must be suffixed with the resultative marker -de, whose formula is “V-de + clause,” as illustrated below: (8)

她說得王先生笑了。 (現代漢語) Tā shuō-dé Wáng Xiānshēng xiào-le. she say-DE Wang Mr. smile-PERF “She made Mr. Wang smile by saying (something funny).” (Contemporary Chinese)

(9)

他學得眼睛都近視了。 (現代漢語) Tā xué-dé yǎnjīng dōu jìn-shì le. he study-DE eye even short-sighted PERF “He became short-sighted by studying (too much).” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

The Syntax of the Resultative Construction 6.2.2

111

Underlying Grammatical Relationship

The underlying grammatical relationships of resultatives are very important for historical exploration. In the surface structure, the resultative must occur immediately after the verb, but semantically it can be related to any constituent of a sentence. The following formula represents one of the typical structures of a complete sentence in Contemporary Chinese: (10) Subj + PP (instrument, locative, etc.) + V-R + Obj Note that we employ two sets of terms with reference to different features. “Subject” and “object” refer to syntactic positions: the former is preverbal and the latter is postverbal. “Agent” and “patient” indicate semantic roles in relation to the matrix verb; although the agent tends to be the subject and the patient tends to be the object, the picture is quite messy in Contemporary Chinese. For instance, many patient nouns are used as subjects, and agents can also be objects. Although the verb and the resultative form a syntactic constituent in the surface structure, the resultative may bear a grammatical relationship to other constituents in the underlying structure, as illustrated in (11) (based on Ma and Lu 1996): (11) (a) (b) (c) (d)

Agent: Instrument: Action: Patient:

砍累 砍钝 砍完 砍掉

kǎn-lèi kǎn-dùn kǎn-wán kǎn-diào

“cut-tired” “cut-blunt” “cut-finish” “cut-down”

(the person) (the knife) (the activity) (the tree)

In (11a), the resultative lèi “tired” is actually associated with the agent of the action. In the grammatical relationship of (11a), the resultative is the predicate of the agent (with reference to their underlying relationship). The rest can be understood in the same way. Another related fact is that a “verb + resultative” phrase is constrained only by whether it makes sense in reality. That is, there is no lexical restriction on which verbs or resultatives can co-occur in the resultative construction. According to Ma and Lu (1996), all monosyllabic adjectives can occur as resultatives. In contrast, in English, “the type of adjective that occurs as a resultative is fairly limited” (Goldberg 1995: 195). The time span over which the resultative construction was established is nearly 1,000 years. The resultative construction can be classified into several types, according to the grammatical relationship between the verb and the resultative, and the particular grammatical relationship manifested influenced the speed of development of the different subtypes of the resultative construction. Some types developed earlier, some developed later, and some are still undergoing the process in Contemporary Chinese.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

112 6.2.3

The Transitivity of the Resultative Construction

Only some verb–resultative phrases can be followed by an object, and they are considered “transitive.” According to whether they can have an object, these phrases can be divided into “transitive” and “intransitive” (Zhu 1982: 126). This section aims to draw generalizations regarding the transitivity of resultative constructions. The transitivity of resultative constructions is determined not by either the verb or the resultative alone, but mainly by the grammatical relationships of the resultative with other sentence constituents and the “verb + resultative” combination as a whole. Even if both elements are intransitive, for example, the verb–resultative phrase can also take an object, as illustrated below: (12)

她哭哑了嗓子。 (現代漢語) Tā kū-yǎ le sǎngzi. she cry-hoarse PERF throat “She cried her throat hoarse.” (Contemporary Chinese)

(13)

小孩哭醒了隔壁奶奶。 (現代漢語) Xiǎohái kū-xǐng le gébì anǎinai. child cry-awake PERF next door grandmother “The child cried the grandmother next door awake.” (Contemporary Chinese)

In (12), neither kū “cry” nor yǎ “hoarse” is transitive, and neither can govern the object sǎngzi “throat.” Thus the patient role is assigned by the whole “verb + resultative” phrase rather than by any individual constituent. A similar phenomenon is found in English: She shouted herself hoarse (Goldberg 1995: 181), where herself is not the patient of either shout or hoarse. According to Goldberg, the resultative construction conveys a causal meaning that contributes transitivity and thus can add a patient argument to it. Even if the verb is transitive, the object is not necessarily the patient argument of the verb, as illustrated below: (14)

弟弟跑丢了一只鞋。(現代漢語) Dìdì pǎo-diū le yī-zhǐ xié. young-brother run-lose PERF one-CL shoe “My young brother lost one shoe while running.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

The Syntax of the Resultative Construction (15)

113

他吃圓了肚子。 (現代漢語) Tā chī-yuán le dùzi. he eat-round PERF stomach “He made his stomach round by eating.” (Contemporary Chinese)

Clearly, there is no “action–patient” relationship between “run” and “shoe” in (14) or between “eat” and “stomach” in (15). Once again, the patient role is assigned by the “verb + resultative” phrases as a whole. Ma and Lu (1996) drew several generalizations to predict when a verb–resultative phrase can take an object. Their predictions work fairly well, though none of them is without exception. The following are their generalizations with some slight modifications. Generalization 1. If the resultative is the underlying predicate of a subject (agent), the “verb + resultative” construction cannot be followed by an object: (16)

*他看病了書。(現代漢語) *Tā kàn-bìng le shū. he read-sick PERF book “He got sick by reading books.” (Contemporary Chinese)

(17)

*他喫胖了肉。(現代漢語) *Tā chī-pàng le ròu. he eat-fat PERF meat “He became fat through eating meat.” (Contemporary Chinese)

If the objects are deleted, (16) and (17) will become well formed. The only grammatical construction for this kind of expression is the verb-copying construction, a construction in which the first verb introduces a direct object and the copied verb introduces a resultative (for a discussion of the diachronic development, see Section 10.1). For example, (18) can be expressed by means of the verb-copying construction as follows: (18)

他喫肉喫胖了。 (現代漢語) Tā chī ròu chī-pàng le. he eat meat eat fat PERF “He became fat through eating meat.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

114

Generalization 2. If the resultative is the predicate of a verb, the construction cannot take an object, as illustrated below: (19)

*他打重了孩子。 (現代漢語) *Tā dǎ-zhòng le háizi. he spank-heavy PERF child “He beat his child too heavily.” (Contemporary Chinese)

(20)

*他喫晚了飯。 (現代漢語) *Tā chī-wǎn le fàn. he eat-late PERF meal “He ate late.” (Contemporary Chinese)

The resultative of (19), zhòng “heavy,” refers to the degree of the action dǎ “beat,” and wǎn “late” in (20) expresses the temporal structure of chī “eat.” Again, other linguistic forms are available to express the intended meanings of (19) and (20), such as the disposal construction in (21) or topicalization in (22): (21)

老王把孩子打重了。 (現代漢語) Lǎo Wáng bǎ háizi dǎ-zhòng le. Lao Wang DISP child spank-heavy PERF “Old Wang spanked his child too heavily.” (Contemporary Chinese)

(22)

飯她喫晚了。 (現代漢語) Fàn tā chī-wǎn le. meal she eat-late PERF “She ate a meal late.”

(Contemporary Chinese)

Generalization 3. If the resultative is the predicate of the patient noun (i.e. object), the “verb + resultative” phrase usually has an object, as illustrated below: (23)

她哭湿了枕头。 (現代漢語) Tā kū-shī le zhěntou. he cry-wet PERF pillow. “He cried the pillow wet.” (Contemporary Chinese)

(24)

她笑疼了肚子。 (現代漢語) Tā xiào-téng le dùzi. she laugh-pain PERF stomach. “She laughed her stomach painful.”

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

(Contemporary Chinese)

The Syntax of the Resultative Construction

115

Verb–resultative phrases of this type have the greatest freedom to take an object. The verbs do not even have to be transitive, such as kū “cry” in (23) and xiào “smile” in (24), which are intransitive. Generalization 3 is not a sufficient condition for a “verb + resultative” phrase to have an object. In fact, many “verb + resultative” phrases whose resultatives refer to the patient argument still cannot take an object. There is a great deal of idiosyncrasy involved in the collocations of the verb and the resultative in Medieval Chinese. Once the two elements were fused into a compound word, the above generalizations could be broken. For example, we have found two exceptions for generalization (a): chī-bǎo fàn “eat-full food” and hē-zuì jiǔ “drink-drunk wine.” Different types of resultative construction were introduced into the language over a great time span. Some started to occur more than two millennia ago and were well established by the twelfth century AD. However, some are still in the early stages of their development even in Contemporary Chinese. The rates of development of different verb–resultative phrases are to a great extent determined by the semantic features of the resultative.

6.2.4

Types of Resultative Construction

The term resultative as it is used in the literature on Chinese generally includes phenomena other than those that form the focus of this study (Zhu 1982: 125‒139): (25)

(26)

(27)

V + PP 放在桌子上。 (現代漢語) Fàng zài zhuōzi shàng. put on table above “Put (it) on the table.” Adjectival intensifier 暖和多了。 (現代漢語) Nuǎnhuo duō le. warm much PERF “It is much warmer.”

(Contemporary Chinese)

(Contemporary Chinese)

Verb classifier 他看了兩遍。 (現代漢語) Tā kàn le liǎng biàn. he read PERF two-time “He has read it twice.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

116 (28)

Directional verbs 搬起來。 (現代漢語) Bān qǐlái. take rise “Take something up.” (Contemporary Chinese)

Historically, the above phenomena are related to the development of the resultative construction as defined in this chapter: they either interacted with the development of the resultative construction or underwent a change parallel to that of the resultative construction.

6.3

The Origin of the Resultative Construction

In the above section, we addressed the syntactic and semantic features of the resultative construction in Contemporary Chinese grammar. Now, we turn to identifying its historical origins. The resultative construction as a syntactic pattern was not firmly established until the tenth century AD, and its function before this time was expressed by another syntactic structure, a subtype of verb serialization whose two verbal elements stood in an “action–result” relation. Since the two verbal elements of this structure form two separate syntactic constituents, and between them are syntactic positions for other constituents such as object, adverb, or negation, it is termed the separable resultative structure. This is the direct source of the resultative construction. We will briefly discuss how this direct source came into being in Medieval Chinese.

6.3.1

Proper Context

As mentioned above, “verb + resultative” phrases resemble compound verbs in several major respects; for example, they possess lexical integrity (e.g. they cannot have other words intervening) and can also take an object. This kind of organization was simply impossible not only in Old Chinese but also in Medieval Chinese because of a principle that was incompatible with the syntactic properties of the resultative construction in Modern Chinese. The principle, which held that only two (maximally four) transitive verbs could share (precede) an object, can be formulated as follows: (29)

(V1 + V2) + O

The above formula can be called “the verb co-ordination principle.” Each of the verbs bears an “action–patient” relation to the object. Thus the formula can be broken down into “V1 + Obj” and “V2 + Obj.”

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

The Origin of the Resultative Construction

117

However, in Contemporary Chinese, resultatives serve mainly to indicate the result of an action. This means that most resultatives are static and thus do not have an “action– patient” relationship with the object of the whole verb–resultative phrase, violating the above principle. That is, the emergence of the resultative construction means that the verb co-ordination principle was broken. The verb co-ordination principle of Old and Medieval Chinese is well illustrated by a pair of examples. Ohta (1987: 194‒199) discovered this important contrast in terms of a pair of synonymous verbs in Medieval Chinese – sǐ and shā. Sǐ was intransitive, meaning “die,” and shā was transitive, meaning “kill.” When they were used as a second verb in a verb serialization, V + shā could precede an object, but V + sǐ could not, as illustrated in the following examples (cf. Section 5.7):




(30)

岸崩, 盡壓殺臥者。 (史記 項羽本紀) Àn bēng, jìn yā-shā wòzhě. bank collapse, entirely press-kill sleeper “The collapse of the bank crushed and killed all the sleeping people.” (Shi Ji, Xiang Yu Ben Ji, 100 BC)

(31)

百餘人炭崩盡壓死。 (論衡 明義) Bǎi yú rén tàn bēng jìn yā-sǐ. hundred more people charcoal collapse entirely press-die “The charcoal collapsed, and more than a hundred people were crushed and all died.” (Lun Heng, Ming Yi, AD 100)




In (30), both verbs that precede the object “sleeping people” are transitive, and the predicate can be broken down into two phrases: yā wò-zhě “to crush sleeping people” and shā wò-zhě “kill sleeper.” In contrast, in (31), the patient argument of the transitive verb yā “crush” is sentence-initial because the second verb sǐ “die” is intransitive. If the second verb is intransitive in a two-verb serialization, another place to put a patient argument is between the two verbs. In this case, the patient is the object of the first verb and the subject of the second, as illustrated in (32): (32)

1




炊一石豆熟1。 (齊民要術 作豉法) Chuī yī-shí dòu shú. boil one-CL bean cooked “Boil one dan of beans cooked.” (Qi Min Yao Shu, Zuo Shi Fa, AD 550)

Cited from Hu (2005).

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

118

In (32), the patient noun yī-shí dòu “a dan of beans,” appearing between the two verbs, is the patient of “boil” and the subject of “cooked.” In Chinese, the majority of adjectives can be used as resultatives, which can be thought of as intransitive verbs. In Chinese, adjectives and verbs share major syntactic properties; e.g. they can be used as a predicate, which has motivated some scholars to group adjectives and verbs into one syntactic category – “predicate word” (or “verbal element” in our discussion). In Medieval Chinese, the adjectives used as resultatives behave exactly like intransitive verbs in that V + Adj. phrases cannot be followed by objects.

6.3.2

Causative Form

Both Old and Medieval Chinese lacked the resultative construction, but this does not mean that the language at that time did not have the function that is expressed by the resultative construction in Contemporary Chinese. In fact, at that time, there were several options for expressing an “action–result” relation. Old Chinese belonged to one of many languages possessing a causative morphological form, a device for creating a verb form meaning “to cause X to Verb” from a form “X verb” (Spencer 1995: 24). Old and Medieval Chinese often used a single word to function like a verb–resultative phrase, which could be paraphrased as “cause something to have some result.” That is, a single word expressed an “action–result” relationship, and this kind of word could be an intransitive verb (including adjectives). If an intransitive element was used as a causative form, its function would be signaled inflectionally. According to Mei (1980), Li (1993: 24‒27), and Pulleyblank (1995: 118), Old Chinese had a set of inflections to create a causative form, including a suffix -s, an infix -r-, and a shift of voiced initials to their voiceless counterparts. However, all of these devices have left no traces in any of the numerous dialects of Contemporary Chinese; hence it is difficult to find empirical evidence for these suggestions. However, one device of the causative form – tone shift – undoubtedly existed in Old and Medieval Chinese because we can still find relics of it in Contemporary Chinese. The verb yǐn “drink,” for instance, is an ordinary verb when pronounced with a falling–rising tone, meaning “drink,” but when used as a causative, meaning to make somebody drink, its tone shifted to a falling one. This causative form is still actively used in the spoken language of Contemporary Chinese, and one example from Old Chinese is provided in (5). In another pair of examples, an adjective (intransitive) changes to a (transitive) verb by means of the same tone device: hǎo means “good,” and hào means “like” or “love.” This causative form was systematically recorded in ancient

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

The Origin of the Resultative Construction

119

phonological books, such as the Qie Yun composed at the beginning of the seventh century AD, which makes the suggestion that tone shift was a grammatical device of Old and Medieval Chinese more convincing. In general, the pattern is that causative forms of verbs with a nonfalling tone are created by changing to a falling tone and causative forms of verbs with a falling tone are created by changing the tone to one of the nonfalling tones. The following examples illustrate causative usages of verbs:




(33)

晉侯飲趙盾酒。 (左傳 襄公二年) Jìn Hóu yìn Zhào Dùn jiǔ. Jin Duke cause-drink Zhao Dun wine “Duke Jin had Zhao Dun drink wine.” (Zuo Zhuan, Xiang Gong Er Nian, 550–400 BC)

(34)

則修文德以來之。 (論語 季氏) Zé xiū wén-dé yǐ lái zhī. then build culture-virtue to cause-come them “Then (you) made them come by cultivating culture and virtue.” (Lun Yu, Ji Shi, 500 BC)

(35)

甘其食, 美其服。 (老子 八十章) Gān qí shí měi qí fú. sweet their food, beautiful their cloth “(They) regard their food as tasty and regard their clothes as beautiful.” (Lao Zi, Chapter 80, 550 BC)







The meaning of (33) is “Duke Jin causes Zhao Dun to drink.” The verb lái in (34) is normally intransitive and has a falling–rising tone, but the tone was presumably shifted to the falling tone when the verb was used as a causative.2 Likewise, because the two adjectives in (35) are also in causative use, they are both pronounced with the falling tone, but the default tone of gān “sweet” is the level tone and that of měi “beautiful” is the falling–rising tone. However, as Wang (1989: 367) pointed out, the inflectional causative in Old and Medieval Chinese had a functional defect in that it expressed only the result and not the action. For example, xiào zhī “small it” means “to make it small,” but it does not specify what action causes the result. In Old and Medieval Chinese, there were two other major options for expressing both an action and its result.

2

This phenomenon is called pò-dú “variant pronunciation of a Chinese character with a different meaning.” It is well evidenced in the literature, such as Ohta (1987: 68‒73).

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

120 6.3.3

Auxiliary Causative

A surface form of the causative found in many other languages is “AUX + Obj + VP,” as in the English example “He made her smile” (Spencer 1995: 284). The Chinese language has always had such a form since Old Chinese. In Medieval Chinese, for example, to express an action and its result, an auxiliary causative form could be inserted between the verb and the resultative, given by the formulas: (a) “V + AUX + Obj + VP” and (b) “V + Obj + AUX + R,” as illustrated below:




(36)

养令翮成。 (世說新語 言語) Yǎng lìng hé chéng. nurture make feather complete “(He) nurtured the feathers (of the bird) and made it recover.” (Shi Shuo Xin Yu, Yan Yu, AD 450)

(37)

拽弓叫圓。 (敦煌變文 韓擒虎話本) Zhuāi gōng jiào yuán. drag bow make round “He dragged his bow and made it round.” (Dun Huang Bian Wen, Han Qin Hu Hua Ben, AD 800–1000)




Literally, the two auxiliary verbs lìng and jiào mean “make” and “have,” respectively. The above structure conveys a causative sense, emphasizing that the result is the volition of the agent. In Medieval Chinese, another causative form, the separable resultative construction, was more expressive and more common.

6.4

The Separable Resultative Structure

As discussed in Chapter 5, the separable resultative structure was a widely used syntactic pattern in Medieval Chinese. This structure typically consists of a verb plus an adjective or an intransitive verb. The two elements stand in an “action–result” relation and represent two independent constituents. They are often separated either by the patient of the verb or by the modifier of the adjective/intransitive. This structure can be formulated as follows: (38)

V + XP + R (XP = object, adverb, or negative; R = adjective or Vintr)

For the sake of simplicity, we use the capital letter R to refer to the result of the action in both the separable resultative structure and the resultative construction, but the syntactic and phonological properties of resultatives changed in the course of the development of the resultative construction.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

The Separable Resultative Structure

121

Note that the verb and the resultative in Medieval Chinese represent two independent constituents. Between them is another syntactic position, which might accommodate nouns (i.e. the object of the verb), adverbs, and negative elements. Many syntactic categories could be used as resultatives, including intransitive verbs, adjectives, preposition phrases, and quantifiers. This structure was the direct source of the resultative construction in Modern Chinese. They were functionally similar but syntactically distinct. In Contemporary Chinese, the verb and the resultative form a single constituent and allow no intervening material. Considering their diachronic relationship and syntactic contrast, we call the corresponding construction in Medieval Chinese “the separable resultative construction.” The form in (38) represents only the surface string. The constituent structure varies, depending on the syntactic category of the XP. If the XP was a noun phrase, it had to be the patient of the verb, representing the constituent structure “(V + O) + R.” If the XP was an adverb or a negative, it had to be the modifier of the resultative, reflecting the constituent structure “V + (Adv./Neg. + R).” Let us use a group of examples to illustrate the types of separable resultative structure in terms of different syntactic categories of resultatives. (a) R = intransitive verb, as illustrated in (39) and (40): (39)




喚江郎覺! (世說新語 假譎) Huàn Jiāng Láng jué! call Jiang Lang awake “Call Jiang Lang awake!” (Shi Shuo Xin Yu, Jia Jue, AD 450)

(40)

(b)




果震柏粉碎。 (世說新語 朮解) Guǒ zhèn bǎi fěn-suì. really shake cypress break-to-pieces “(Lightning) really shook the cypress tree into pieces.” (Shi Shuo Xin Yu, Shu Jie, AD 450) R = adjective, as illustrated in (41) and (42):




(41)

分肉食甚均。 (史記 陳丞相世家) Fēn ròu-shí shèn jūn. distribute meat-food very even “He distributed meat and food very evenly.” (Shi Ji, Chen Cheng Xiang Shi Jia, 100 BC)

(42)

制街衢平直。 (世說新語 言語) Zhì jiēqú píng zhí. build street flat straight “(He) built streets flat and straight.”




(Shi Shuo Xin Yu, Yan Yu, AD 450)

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

122

The development of the above two types of resultative is the focus of our investigation. Adjectives and intransitives are the core of the resultatives in Modern Chinese. Nevertheless, their behaviors should be viewed in the context of Medieval Chinese. Sentence-final position can be occupied by many other syntactic categories, as illustrated below. (c) R = preposition phrase, as illustrated in (43) and (44):




(43)

種瓜於長安城東。 (史記 蕭相國世家) Zhòng guā yú Cháng ’ān chéng dōng. plant melon at Chang-an city eastern “(He) planted melons in eastern Chang-an city.” (Shi Ji, Xiao Xiang Guo Shi Jia, 100 BC)

(44)

乃說桓公以遠方珍怪物。 (史記 齊太公世家) Nǎi yuè Huán Gōng yǐ yuǎnfāng zhēn guài wù. then please Huan Duke with far-place precious strange thing “Then, (he) pleased Duke Huan with precious and strange things from remote places.” (Shi Ji, Qi Tai Gong Shi Jia, 100 BC)




In contrast to adjectives and intransitive verbs, however, only some of the above categories in sentence-final position serve to indicate the resulting state of the V as a resultative does. Their development is closely related to that of types (a) and (b). All of the types of resultative listed above could be regarded as adjuncts to the main clause. They can be divided into three types according to semantics (the classification is based on Thompson and Longacre 1985: 172 with modification). (a) Complement. Cases where the intervening object and the resultative are in a subject–predicate relation and together form a clause, e.g. types (a) and (b), whose resultatives are intransitive verbs or adjectives. (b) Adverbial clause. Resultatives that function as modifiers of verb phrases or entire sentences, e.g. types (c) and (d), whose resultatives are preposition phrases and time words. (c) Relative clause. Clauses that serve to modify nouns, e.g. type (e), whose resultatives are numeral phrases. All of these examples can be thought of as a complex sentence consisting of a nucleus plus an adjunct (Hopper and Traugott 2003: 169), with the resultatives in an adjunct position. These examples are reminiscent of the “small clause” in

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

The Separable Resultative Structure

123

English, such as We hammered the metal flat and hot, and like its English counterpart the resultative in Medieval Chinese should be considered a sentential constituent rather than an expansion of the nominal object. According to Li and Shi (1997), all of these syntactic categories used as resultatives could occur alone as predicates. In fact, the semantic categories of resultatives listed in (a) and (b) did influence the development of resultatives at a later stage when this adjunct position was gradually eliminated. In all of the examples of the separable resultative structure that we have shown, the verb and the resultative are separated by a nominal object, the patient of the verb. This means that the verb in the structure must be transitive. When the verb is intransitive, the verb and the resultative are still often separated by an adverb or negative, as illustrated in the following examples. (45)




子道, 來何遲? (世說新語 文學) Zǐ Dào, lái hé chí? Zi Dao, come why late “Zi Dao, why did you come late?” (Shi Shuo Xin Yu, Wen Xue, AD 450)




(46)

謝萬石後來, 坐小遠。 (世說新語 雅量) Xiè Wànshí hòu lái, zuò xiǎo yuǎn. Xie Wanshi late come sit slight far “Xie Wanshi came late and sat slightly far away.” (Shi Shuo Xin Yu, Ya Liang, AD 450)

(47)

胡之去已遠。 (世說新語 仇隙) Hú Zhī qù yǐ yuǎn. Hu Zhi go already far “Hu Zhi already went far away.”




(Shi Shuo Xin Yu, Chou Xi, AD 450) The examples in (45) and (46) illustrate one subtype of the separable resultative structure and the source of the resultative construction in Contemporary Chinese. As the preceding chapter pointed out, in Contemporary Chinese a verb–resultative phrase can still assign a patient argument even if its two components are both intransitive. In addition, these examples demonstrate the separability of the structure: in each of the three examples, the verb and resultative are separated, e.g. by a question word hé “why” in (45) and by an adverb yǐ “already” in (47). Once again, the separability shows that the verb and resultative at this time also belong to two independent constituents.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

124

Types (a) and (b) are the focus of our investigation because intransitive verbs and adjectives form the great majority of resultative elements in Modern Chinese. These two types are the first change toward the resultative construction of Contemporary Chinese and serve as the basis of the analogical change that the other types would undergo. By calling the structure used in the examples above “the separable resultative structure,” we emphasize a diachronic relation between it and the resultative construction. However, they have essential differences: in the separable resultative structure, the verb and the resultative have only a semantic relationship (action– result). They do not define a syntactic unit, as verb–resultative phrases do in Contemporary Chinese. As the name suggests, the verb–resultative phrases of the separable structure are often separated by an object, adverb, or negative in Medieval Chinese. We will explore how the two syntactic constituents – the verb and the resultative – fused into one over time and the syntactic and morphological consequences of the fusion. Any linguistic development is typically a chain of changes. An earlier change enables a later change to take place, and the earlier change is made possible by yet another preceding change. The separable resultative structure in Medieval Chinese provides the direct source for the resultative construction in Modern Chinese, but what made this structure come into being? To answer this question, we have to examine the development of the verb co-ordination construction in Old Chinese.

6.5

Frequency and Idiomatization

The former sections showed that when a verb–resultative pair defines a disyllabic unit, it tends to become fused into a single constituent because of the disyllabification tendency. However, not all disyllabic verb–resultative phrases became fused at the same time, and the process from the first verb–resultative fusion to the establishment of the general pattern took several hundred years. The timing of the fusion of a particular verb–resultative phrase is related to the frequency of the cooccurrence of the particular verb and resultative. The more often a monosyllabic verb and a monosyllabic resultative co-occurred, the more likely they were to be subject to fusion. The fusion of the verb and the resultative did not instantly create a syntactic pattern. Initially, the fusion produced only compound-like verbs, and the collocation of verb and resultative was highly lexically constrained. A given verb was typically associated with a particular resultative, which means that their combination was not productive. Therefore we think that the earliest “verb + resultative” phrases were idiomatic or

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Frequency and Idiomatization

125

lexical but not syntactic. The “verb + resultative” syntactic pattern arose out of the accumulation of the tokens of fusion.

6.5.1

Frequency and Syntactic Change

How to view the role of frequency in grammaticalization is an unresolved problem. According to Heine et al. (1991: 38‒39), it has been cross-linguistically attested that frequency is relevant to grammaticalization. A grammatical device generally comes from linguistic expressions in frequent and general use rather than from rare or constrained ones. Can we say that a lexeme has become grammaticalized simply because of a high frequency of occurrence? Some researchers, such as Heine et al. (1991: 38‒39), have argued that high frequency alone is not sufficient to account for grammaticalization but is rather a “concomitant feature of the concepts recruited for this purpose” Heine et al. (1991: 39). We agree that high frequency of use is not sufficient to account for grammaticalization, but disagree that it is merely a concomitant feature of the lexemes recruited for the purpose. High frequency of use is often a trigger for grammaticalization, especially that involving fusion of two items. The term “high frequency of use” can actually refer to several different properties of lexemes. These different properties bear different relationships to the process of grammaticalization. First, the term refers to the general occurrence of a lexeme in the language, which is determined largely by its semantic generality. The more general its semantics, the more frequent its occurrence. Frequency in this sense may not be directly related to grammaticalization. Second, high frequency of use could also mean the occurrence of a lexeme in a particular linguistic context involved in a grammaticalization process. In this case, frequency of use does play a critical role in triggering and advancing certain lexical items to develop into grammatical markers. Third, it could mean that when a lexeme becomes grammaticalized, its frequency increases. In this sense, high frequency of use is truly a concomitant feature of a grammaticalized item, as Heine et al. claimed. For our purpose here, the relevant point is that the more frequently a verb–resultative pair co-occurs, the more likely it is to become fused. In the initial stage of the development of the resultative construction, the frequency of co-occurrence of the verb and the resultative to a great degree determined which VR phrases were first introduced into the language. Frequency can be used as a criterion to judge the appearance of a new grammatical device and the degree of grammaticalization (Bybee 2006). As a lexical item develops into a grammatical marker, it becomes semantically bleached and thus has fewer lexical constraints on what it combines with. As a result, its

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

126

frequency will increase. According to the assumption that “the more frequently a form occurs in texts, the more grammatical it is assumed to be” (Hopper and Traugott 2003: 103), we can detect approximately when a lexical item becomes grammaticalized. The effect of frequency is also found in language contact. Grammar borrowing happens first to high-frequency words. For example, according to Yue-Hashimoto (1993), the extension of the new question form “V-not-V” from the southern to the northern dialects followed the route from more frequent to less frequent verbs: copula > existential/possessive verbs > optative verbs > ordinary verbs.3 This pattern has been attested in various dialects. From a diachronic point of view, lexicalization has much in common with grammaticalization. They both involve “phonological and semantic change” (Lipka 1990: 97, Traugott 1997) and “univerbation as a single unit.” The development of the resultative construction illustrates another relation between lexicalization and grammaticalization: the lexicalization of the verb and the resultative is the precursor of the resultative construction and aspect suffixes. At the initial stage, every verb–resultative pair was in essence idiomatic, and the combination was idiosyncratic. This is also true of English equivalents, in which certain resultatives are apt to collocate with particular verbs. For example, “eat” is most likely to collocate with sick and cry with to sleep (Goldberg 1995: 192): (48)

(a) He ate himself sick. (b) ?He ate himself ill/nauseous/full.

(49)

(a) She cried herself to sleep. (b) ?She cried herself calm/wet.

The above type of collocation between the verb and the resultative is determined by many factors. First, a verb tends to co-occur with its most natural result; for example, “beat” is most likely to collocate with “broken.” Other combinations may be particular to the language. For example, unlike the English pair “eat-sick,” in Chinese the corresponding verb chī “eat” is more likely to be collocated with bǎo “full” than with “sick.” As discussed previously, the resultative construction of Contemporary Chinese behaves like a syntactic pattern that allows combinations only if they hold an “action–result” relation. In the initial development of the resultative construction, however, the collocations of the verbs and resultatives were largely idiosyncratic, 3

Yue-Hashimoto (1993) did not provide statistical work to support this ordering. It is our understanding that the copula and existential/possessive verb are light verbs and thus more frequent than others.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Frequency and Idiomatization

127

and a given verb could be combined only with particular resultatives in pairs that had a natural “action–result” relationship. The collocation of a particular “verb + resultative” pair first became lexicalized into compound verbs because of frequent co-occurrences. The disyllabification tendency was one of the factors motivating the emergence of the resultative construction. When the idiomatized “verb + resultative” phrases fit a disyllabic unit, they were likely to be lexicalized into compound verbs. These compounds could then function like ordinary verbs; for example, they could precede a patient argument. There are similarities and differences between idiomatization and lexicalization with reference to “verb + resultative” phrases in Chinese: (a)

(b)

Idiomatization refers to the formation of idiosyncratic “verb + resultative” pairs that are frequently used. These pairs also tend to have a natural “action-resultative” relationship and high co-occurrence. Their meanings tend to be less derivable from their components. The pairs can be either words or phrases. Lexicalization involves idiomatized verb–resultative pairs whose phonological representation is a disyllabic unit. When a “verb + resultative” pair fits the disyllabic unit, it tends to undergo further fusion into a compound verb. Thus the resulting “verb + resultative” phrases can behave like regular verbs; for instance, they can have objects.

Frequency plays a key role both in idiomatizing “verb + resultative” collocations and in lexicalizing “verb + resultative” compound verbs. There is a formal criterion for evaluating the function of frequency in grammaticalization. Old and Medieval Chinese had a syntactic rule that stated that in the co-ordinate verb construction “V1 + V2 + O” the two verbs must be transitive and each of them must have an “action–patient” relationship with the object. This rule ceased to apply when the resultative construction was firmly established around the twelfth century AD. Superficially, the emergence of the construction “verb + resultative + O” is a violation of this rule because most resultatives are intransitive and do not have an “action–patient” relationship with the following object. One question arises: can the innovation of syntactic devices violate the existing grammar? We assume that grammatical change cannot proceed by violating existing rules. Under this assumption, we hypothesize that the development of the resultative construction involved the following steps: Step 1. Idiomatization of a “verb + resultative” collocation due to their frequent co-occurrence.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

128

Step 2. Compounding of the verb and the resultative due to the combination being a disyllabic unit. Step 3. Innovation of the resultative construction due to the lexicalization of the verb and the resultative. Step 4. Creation of the “verb + resultative” syntactic pattern due to the accumulation of tokens of VR fusion. The case studies in this chapter show that the high frequency of collocation stimulates a separable “verb + resultative” structure to develop into a syntactic pattern. Therefore the resultative construction developed along the following path: (50)

Separable VR structure > Idiomatization > Compounding/Fusion > Resultative construction

This process seems to be inconsistent with the usual development cline from the loosest to the most highly fused means of expression (Bybee et al 1994: 40): “Syntactic > nonbound grams > inflection > derivation > lexical.” However, the orderings here aim to define the degree of integration between two elements. The ordering in (50) is designed to capture an important historical fact: in the initial stage of the development of the resultative construction, the emergence of the resultative construction did not immediately create a new syntactic pattern but reflected cases of compounding or lexicalization. The “verb + resultative” syntactic pattern arose out of the combined power of a vast number of individual “verb + resultative” phrases. The cline in (50) indicates the steps of the development of the resultative construction, but it does not suggest that some lexicalized “verb + resultative” pairs became a syntactic organization. Particular VR pairs, once lexicalized, stayed lexicalized as compounds. Every case of lexicalization involves a boundary loss between a particular verb and a particular resultative. When an increasing number of verb–resultative pairs became fused, the separable resultative structure in Medieval Chinese developed into the resultative construction via the boundary loss between the verb and the resultative as a whole. The VR syntactic pattern is the combinational output of the process, while particular verb compounds and certain inflections are by-products of the process.

6.5.2

The Negative Potential Form

The development of the negative potential resultative construction is another example that illustrates how the VRO form made its earliest appearance and how it spread via lexical expansion (cf. Section 4.6). The negative potential form in Modern Chinese is “V-bu-R” (bu “not,” reduced to a neutral tone), which means “the action is unable to realize the resultative,” as illustrated in (51):

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Frequency and Idiomatization (51)

129

他搬不動那張桌子。 (現代漢語) Tā bān-bu-dòng nà-zhāng zhuōzi. he carry-not-move that CL table “He is unable to move the table.”

This form was created in two steps. First, the negative marker bu “not” and the resultative had to be fused into a single constituent in the post-object position, and the fused bu-R phrase was then further combined with the verb into an inseparable constituent in certain contexts, which gave rise to the negative potential form. The two steps are described as follows: (52)

Step one: V O bu-R Step two: V-bu-R O

The form of step one is one instantiation of the separable resultative structure and that of step two exemplifies the resultative construction. That is, the development of the negative potential form followed the same path as that of ordinary “verb + resultative” phrases. The change from step one to step two started around the tenth century AD. Before this point, bu-R phrases were exclusively limited to the position after the object, where they were bound into a disyllabic constituent. We collected 1,041 examples of negative potential forms in Zhu Zi Yu Lei, a text composed around the twelfth century AD that contained many types of resultative. Of them, 913 were potential instances with the resultative bu-de “not-able,” accounting for eighty-seven of the total examples, and appearing much more frequently. At this point, only the form with the particular resultative bu-de “not-able” could occur in the new construction “V-bu-R O,” as illustrated in (53), and the rest could still be used in the old construction “V O bu-R,” as illustrated in (54): (53)

卻管不得那富貴。 (朱子語類卷三) Què guǎn-bù-dé nà fù guì. but control-not-able that rich noble “(We) cannot control the richness and nobleness.” (Zhu Zi Yu Lei, Chapter 3, AD 1200)

(54)

他也自遏他不住。 (朱子語類卷二十一) Tā yě zì è tā bù-zhù. he also himself contain him not-stop “He also could not contain it.” (Zhu Zi Yu Lei, Chapter 21, AD 1200)

The development of the negative potential form reveals a causal relation between the high frequency of collocation and the earliest appearance of the new form. Not considering lexical constraints on the forms, the old form “V O bu-R” and the new form “V-buR” coexisted simultaneously. However, as with other cases we have considered thus far,

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

130

Resultative Construction

these two forms were used with two distinctive V + bu-R sets. The new form applied only to V-bu-de, whose meaning (i.e. unable to do) was most general and whose frequency was much higher than that of the others. There are 1,255 V-bu-de tokens, but the total of the other four types is only 199. We hypothesize that it was this high frequency of collocation that lexicalized V-bu-de into a compound verb so that the whole phrase could take an object like regular verbs. However, not all of the verbs collocating with bu-de “unable to do” could occur in the “V-bu-R” form in the twelfth century AD. Although only bu-de could occur in the V-bu-R O form, it also occurred in the old form “V O bu-R.” Which form it appeared in was determined by the frequency with which it collocated with the verb in question. According to the statistical work done by Shi (2002a: 119), in Zhu Zi Yu Lei the resultative bu-de “not-able” appeared with 278 different verbs, but only in a small portion of these collocations could it precede an object. Once again, this shows that there is a correlation between the frequency of the collocations of verbs with bu-de and their appearance in the new form: the more frequently the verbs collocated with bu-de, the more likely the resulting V-bu-de phrase was to be able to precede an object. All five verbs that collocated most frequently with bu-de could occur in “V-bu-de O,” accounting for roughly 50 percent of all such uses of the 278 verbs. Those verbs with a low frequency of collocation with bu-de were less likely to occur in “V-bu-de O.” Of the 161 verbs with only one collocation with bu-de, only 2 percent could be used in the construction “V-bu-de O,” and they accounted for only 6 percent of all “V-bude” uses (for details, see Shi 2002a: 117‒120). This bu-de case indicates that frequency had a double function in creating the negative potential form. The first one was related to the resultative. Within a set of resultatives with similar functions, the most frequent one possessed the greatest possibility of occurring first in the new form. The second function of frequency was related to the verb. For a given resultative, the verb that most frequently occurred with it was most likely to occur in the new form. In other words, the high frequency of the use of an element as a resultative enabled it to occur in the “V-bu-R O” construction earlier than other similar bu-R phrases. A high frequency of appearance of a verb with the high-frequency bu-R enabled the verb to occur in the construction earlier than other verbs. Only at the very beginning of the development do we see a particular verb–resultative phrase being used in both the old and new constructions. However, the negative potential form “V-bu-R” was established as a new grammatical device by the fourteenth century AD. Not only could all phrases with bu-de as their resultative be used in the new form, but those using other bu-Rs could take the new form as well, as illustrated below (cf. Section 4.6):

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Frequency and Idiomatization

131




(55)

你咽不下相思這口涎。 (元刊雜劇 霍光鬼諫) Nǐ yàn-bù-xià xiāngsī zhè kǒu xián. you swallow-non-down lovesick DEM CL sputum “You cannot swallow this lovesick sputum.” (Yuan Kan Za Ju, Huo Guang Gui Jian, AD 1300)

(56)

鎖不住心猿意馬。(王實甫 西廂記) Suǒ-bù-zhù xīn-yuán-yì-mǎ. lock-not-stay fanciful imagination “His fanciful thinking cannot be stopped.” (Wang Shi Fu, Xi Xiang Ji, AD 1300)




By roughly the sixteenth century AD, the innovative construction had entirely replaced the old one. Then the Chinese language had a new grammatical device: the negative potential form. Once the new form became a syntactic pattern, the lexical constraint was removed, and any verb and bu-R could collocate as long as their collocation made sense.

6.5.3

The Form “V-sǐ + O”

The verb sǐ “die” was intransitive and was often used as a resultative with verbs denoting violent actions. According to the syntactic rule that only two transitive verbs could precede a patient object, the resultative sǐ was limited to two constructions when the verb had a patient argument in Medieval Chinese: (57)

(a) V + O + sǐ (b) Patient argument + V + sǐ.

The “patient argument” in pattern (57b) could be the topic of a sentence or introduced by the disposal morpheme. However, the patient argument gradually started to occur after the construction “V-sǐ” after the tenth century AD. According to our investigation, the V-sǐ-O form made its first appearance around the twelfth century AD. The text Zhu Zi Yu Lei already had several examples of this type. During the period from the twelfth century AD to the sixteenth, the resultative construction was established. The resultative sǐ could collocate with an increasing number of verbs. However, the distribution of the V-sǐ-O was quite uneven across different verbs. According to the statistical work of Shi (2002a: 107), in Shui Hu Zhuan, a vernacular novel composed around the fourteenth century AD, there were a total of twenty-three different verbs that were followed by the resultative sǐ “die.” Of them, the first two most frequent pairs, dǎ-sǐ “beat-die” and shāsǐ “kill-die,” account for 65 percent of all the VR pairs. The more frequently the verb and the resultative were combined, the earlier they reached a high degree of fusion. All of the

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Resultative Construction

132

first four most frequent pairs with the resultative sǐ, dǎ-sǐ “beat-die” and sha-sǐ “kill-die, ” could be used in the resultative construction. In addition, the most frequent pair, dǎ-sǐ “beat-die,” appeared in the earliest attested example of the “V-sǐ O” form (Mei 1991). Consider two examples of the “V-sǐ O” form: (58)

等酒家去打死了那廝便來! (水滸傳三回) Děng jiǔjiā qù dǎ-sǐ le nà sī biàn lái! wait I go beat-die PERF that guy then come “Please wait. I will go kill that guy and then come back.” (Shui Hu Zhuan, Chapter 3, AD 1400)

(59)

林沖殺死差拔。(水滸傳九回) Lín Chōng shā-sǐ Chà Bá. Lin Chong kill-die Cha Ba “Lin Chong killed Cha Ba.” (Shui Hu Zhuan, Chapter 9, AD 1400)

In contrast, even at this time, the less frequent pairs of V-sǐ could not be followed by the object. In the fifteenth century AD, the resultative construction was firmly established, and the verb and the resultative sǐ could no longer be separated by the object or other materials. However, the frequency of a “verb + resultative” collocation to a great extent determined the degree of verb–resultative fusion. The lower degree of fusion applied to those verb–resultative phrases that could no longer be separated by an adverb, object, or negative, but could not be followed by an object; the higher degree of fusion applied to those that did not allow any intervening material and could have a patient object. It is evident that there was a strong connection between the high frequency of collocation and the degree of fusion; specifically, the more frequently a verb and resultative collocated, the more likely they were to occur in the VRO form, the defining feature of the high degree of fusion. On the other hand, verb–resultative pairs with the lowest frequency tended to stay at the lower degree of fusion. The frequency of a verb– resultative phrase was determined largely by the semantic generality of both the verb and the resultative. The analysis of this chapter has several theoretical implications, including the pathway of transition from old to new forms, the function of frequency of collocation in grammaticalization, the relationship between degree of fusion and unidirectionality, and the place of surface structure in grammaticalization. These issues are addressed in this section. Grammaticalization is characterized by the coexistence of older and newer forms, which is formulated as follows (Hopper and Traugott 2003: 36): A > A/B > B. This formula states that the transition from A to

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Frequency and Idiomatization

133

B involves an intermediate stage in which A and B coexist. It is applied to changes at all levels of a linguistic system, including phonology, lexicon, and syntax. This type of transition especially highlights the nature of syntactic changes. Here the initial stage is already one of variation, and the final exemplified stage may still be in variation. If only linear arrangements are considered, it may seem as if nothing changes throughout the entire process of the development of a new grammatical construction. However, if we employ the following criterion to judge the status of linguistic expressions, we will see that status change is often involved in the process: (a) (b)

Lexical (idiomatic) refers to forms whose collocations are not productive and are lexically governed. Syntactic refers to forms whose collocations are productive and the least lexically governed.

When we consider the properties of lexicalization and syntacticization, we see that two linguistic forms regarded as variants in one functional domain actually belong to different levels of the linguistic system. The development of the resultative construction provides us with an excellent case for observing the transition from the old form to the new. There were two changes involved during the transition from the separable resultative structure to the resultative construction: Change 1: V + Obj + R > V R O; R is intransitive. Change 2: (V1 + V2) + Obj > zero; both V1 and V2 are transitive. The two constructions, “VOR” and “VtrVtrO,” represent two sides of one syntactic rule: two verbal elements preceding an object must be transitive; otherwise, the intransitive one must appear after the object. VRO was not just a new form but also was incompatible with the existing verb co-ordination rule if VRO was initially a syntactic structure. One general question arises: can a new grammatical structure come into being through violating an existing syntactic rule? No matter what our answer is, we face a paradox here. If the answer is yes, the existing rule can no longer be called a rule because it permits exceptions to occur and to increase. However, robust historical data demonstrate that the rule did exist in the language for a very long time, allowing no exceptions. If the answer is “no,” we will conclude that new forms can never come into being. However, this conclusion is evidently wrong. To resolve the paradox, we propose another model of syntactic change: (a) verb set: {V1, V2, V3, . . .}; (b) resultative set: {R1, R2, R3, . . .}

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

134

Resultative Construction Step 1. Idiomatization: if V1 and R1 frequently collocate, they tend to become a fixed expression. Step 2. Lexicalization: if an idiomatized V1R1 phrase is disyllabic (i.e. reflects the basic phonological unit), it is likely to be lexicalized as a single verb. Once the phrase becomes a single verb, it can function like ordinary verbs; for instance, it can precede an object. The earliest V1R1O combination is essentially a lexical entity rather than a syntactic pattern. In the initial stage R1 is again used only with one particular verb. Step 3. Syntacticization: the enabling factors – disyllabification and high frequency of co-occurrence – are continuously at work, which adds an increasing number of similar examples to the V1R1O set. When the number of examples of this type reaches a critical point, a new syntactic pattern will be established and will be generalized to other similar uses.

For the sake of simplicity, we assume that fusion involves only one particular verb and one particular resultative. In fact, the same process may work in multiple dimensions. A particular verb may be idiomatized with several resultatives at once, and vice versa. The present model of the transition from old form to new form has the following advantages. First, new forms are created without violating the existing grammar. Superficially, the appearance of VRO contradicted the verb co-ordination rule which states that only two transitive verbs can have an object. However, initially, the verb– resultative collocation was merely a lexical matter rather than a syntactic pattern. In other words, the linguistic status of VRO is different from that of VOR and VtVtO: the VR phrases are compound-like, but the latter are syntactic patterns and thus are highly productive. VRO developed into a syntactic pattern at the expense of the disappearance of the two old forms. Second, the model captures the lexical characteristics of the earliest appearances of individual VR phrases. When the negative potential form was first introduced in the twelfth century AD, for example, the resultative was limited to bu-de “unable to do,” and the verbs that could collocate with this resultative were only those that frequently co-occurred with it. When xǐng “awake” was used as a resultative in the construction “VRO,” it could be collocated with only a particular verb, jiào “call.” Likewise, when sǐ “die” was used as a resultative in VRO around the tenth century AD, the verb was limited to dǎ “beat.” More interestingly, nouns used as objects may also be involved in the idiomatization process. For example, when chī-bǎo “eat-full” was used in the VRO construction, it could have only fàn “food” as its object noun, which shows that the whole VRO combination had been idiomatized.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Context for Fusion

135

Finally, our model suggests a possible pathway for the innovation of syntactic patterns: occasional idiomatization/lexicalization > large-scale idiomatization/lexicalization > innovation of syntactic pattern. Although the VRO combination was attested as early as the eighth century AD, it did not become a new grammatical pattern until the twelfth century AD. Once it was used as a syntactic pattern, the verb–resultative tokens and types both increased dramatically, gained new syntactic properties, and soon replaced the old VOR form. The present model is similar to “the mechanism of lexical expansion” proposed by Bybee et al. (1994: 51). As they pointed out, auxiliaries were originally used only with active main verbs and then spread to uses with static predicates. This scope expansion is typically found in the further development of a gram. As a lexical item grammaticalizes, it tends to increase its potential for collocation. Our model has something in common with lexical expansion, but the tow models are different in some respects. Our model aims to address how a new form makes its first appearance, how it is extended, and how it finally develops into a syntactic pattern. That is, I am not concerned with further development after the establishment of a syntactic pattern, which is the central concern of the “lexical expansion” theory.

6.6

Context for Fusion

In this section, we discuss how the separable resultative structure provided the context that developed into the resultative construction. In general, the verb and the resultative not only represent two independent constituents but also can be separated by objects, adverbs, and negatives. The transition is caused by the fusion of the verb and the resultative in one variant of the separable resultative structure in which they are adjacent to each other. The fusion of the verb and the resultative has two direct outcomes: the innovation of the resultative construction and the grammaticalization of morphological markers such as aspect markers. Both outcomes are derived from the same origin – the separable resultative structure, which can be formulated as follows: (60)

V + XP + R; X = object, adverb or negative4

Superficially, the verb and the resultative become fused across an intervening constituent – the X elements, including objects, adverbs, and negatives. Intuitively, it is difficult to understand how the fusion takes place if the verb and the resultative are always separated by X materials. However, in fact, X elements are not necessary; the object is often omitted if it can be understood from the context, and adverbs and negatives are also often absent. 4

If the verb is intransitive, the possible XP elements are adverbs and negatives but not objects.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

136

Resultative Construction

On the basis of the presence and absence of X elements, the separable resultative structure can be divided into the following four subtypes: Structure 1: (V + Obj) + (Adv./Neg. + R). Structure 2: (V + Obj) + R. Structure 3: V + (Adv./Neg. + R). Structure 4: V + R. In the separable resultative construction, the object must be the patient argument of the verb and Adv./Neg. must be the modifier of the resultative; in other words, different intervening materials belong to different parts of the structure. The constituent hierarchy is marked with a parenthesis. Structure 1 is the complete one, as it has all of the possible components. Adv./Neg. is absent in structure 2, and the object is absent in structure 3. Nothing intervenes between the verb and the resultative in structure 4, but the two elements still form independent constituents. The object is the most common element intervening between the verb and the resultative in these four structures; thus its behavior is a reference point for investigating the transition from the separable resultative structure to the resultative construction. In structures 1 and 2, the object may be absent, which gives rise to the surface structures of 3 and 4. Note that the verb in structure 4 is not necessarily intransitive. In fact, verbs in this structure are typically transitive because only transitive verbs can cause something to possess some quality. Additionally, the earliest stage of the resultative construction involved “transitive verbs + resultative,” and only when the resultative construction was firmly established could intransitive verbs be freely used in the structure. Therefore we focus on verb–resultative phrases with transitive verbs because we are concerned with how the resultative construction made its first appearance. Our hypothesis is that the verb and the resultative started to become fused in structure 4, where they were adjacent to each other. Obviously, another type of frequency is relevant here. We argued that the high frequency of collocation between the verb and the resultative could trigger their fusion, but did not take into consideration the ratio between separate and adjacent uses for a particular “verb + resultative” pair. If a given “verb + resultative” pair was overwhelmingly used in structures where the verb and the resultative were separated by other material, it would be difficult for the verb and the resultative to become fused, even though their collocations were relatively frequent in comparison to other pairs. In short, there are two types of frequency: one refers to how often a given verb occurred with a particular resultative in comparison with other types of verb–resultative pairs, and the other refers to how often a verb–resultative pair appeared adjacently in comparison to their separated uses. Therefore the frequency of structure 4 is most relevant to our hypothesis.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

137

Context for Fusion Table 6.1 The ratio of adjacent to separate structures in Zhu Zi Yu Lei Types

Adjacent

Separated

Ratio

V + bu-de 不得 “not-able” V + bu-xia 不下 “not-down” V + bu-jin 不尽 “not-entire” V + bu-chu 不出 “not-out” V + bu-zhu 不住 “not-stay”

914 23 46 34 24

279 9 22 19 21

3/1 3/1 2/1 2/1 1/1

Although structure 4 is only one of the four subtypes, it is more frequent than the other three structures together. This conclusion is drawn from an investigation of two representative cases – the negative potential form and the perfective -le. Table 6.1 shows the distribution of the negative potential form in the text Zhu Zi Yu Lei across the different verbs occurring in the construction. At this time the negative potential form had just started to take shape, and only the resultative bu-de could occur in the construction “V-bu-de O”; the rest still took the old form “V + O + bu-R.” In Table 6.1, “adjacent” refers to structure 4, and “separated” refers to the other three structures. As shown in the table, for each pair of V and bu-R, the frequency of adjacent uses is higher than that of separated uses. Therefore it should not be a surprise that structure 4 enables V and bu-R to become fused, creating a paradigmatic example of the negative potential form at a later stage. The emergence of the perfective -le was another major grammatical change in relation to the development of the resultative construction. In Contemporary Chinese, the aspect marker -le is suffixed to a verb to express the completion of an action with present relevance. It grammaticalized from the resultative use of the full verb liǎo “complete” in the structure “V O liǎo.” The resultative liǎo started to occur between the verb and the object around the tenth century AD, indicating that liǎo had already become grammaticalized at that time. Therefore we are interested in structures in which liǎo was used as a resultative before this point. Table 6.2 shows the results of our investigation of the ninth-century AD text Zu Tang Ji (AD 950), which was composed close to the time of the grammaticalization of liǎo. As before, the resultative liǎo and the verbs preceding it appear more often adjacently than separately. In other words, the frequency of structure 4 is much higher than that of the other three structures for the V-liǎo pair. This encouraged the fusion of the verb and the resultative in structure 4. A generalization can be drawn from the preceding discussion: when the verb and the resultative were in a separable resultative structure, their adjacent uses were more numerous than their separate ones. This can be expressed by the following inequality:

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

138

Resultative Construction Table 6.2 The ratio of V-liǎo to V-X-liǎo in Zu Tang Ji Adjacent: V + liǎo

Separated: V + X + liǎo

Ratio of adjacent to separated

77

54

7:5

(61) VR ≥ VXR Thus, even when the principle at work is the separable resultative structure, the adjacent collocations of the verb and the resultative are still frequent enough to give the two constituents the chance to fuse into one. Adjacent occurrences are more frequent than separated ones. As for regular verb serializations, the collocation of the verb and the resultative was made adjacent by the disappearance of the verbal connective ér. As discussed in Chapter 4, in Old Chinese two verbal elements used as a predicate of a sentence had to be connected by the conjunction ér. Under this condition, where the verb and the resultative were not linearly adjacent, it is impossible for them to become fused; in other words, adjacent collocation is a necessary condition for the fusion of the verb and the resultative. The connective ér gradually became optional from the first century AD onward and finally died away around the sixth century AD for reasons that are still unclear. The disappearance of the connective made two or more co-ordinate verbal elements adjacent in the predicate, which made it possible for the verb and the resultative to be adjacent. That is, the necessary precondition for verb–resultative fusion matured around the sixth century AD. The structure itself is not sufficient to trigger the emergence of the resultative construction. The preceding two chapters have dealt with two other factors (the disyllabification tendency and frequency), and the following chapter will address another one (semantic relevance). Together they determine how, when, and where the verb and the resultative first became fused. This section provides an empirical defense of the “adjacency” hypothesis, which states that the verb and the resultative became fused into a single constituent in surface structures where they were adjacent to each other. If this hypothesis is correct, we can predict that the development of the resultative construction was uneven across all four structures and that those verb–resultative pairs first gained the properties of fusion in structure 4. Considering the source of the resultative structure, we know that a verb– resultative pair has become fused when we see the following changes: (a) objects, which originally occurred between the verb and the resultative, follow the whole verb– resultative pair; (b) adverbs or negatives, which originally intervened between the verb and the resultative, are fronted to the whole verb–resultative phrase under the

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Extension

139

influence of analogy; and (c) the resultative undergoes phonological reduction. Since Chinese characters are not sensitive to change in phonological representation, feature (c) is difficult to observe in the initial stage of the development of the resultative construction. Now we employ features (a) and (b) to analyze several concrete cases.

6.7

Extension

As seen before, the formation of individual verb–resultative phrases was motivated by a cluster of factors such as disyllabification, adjacent structure, and high frequency of collocation. At the beginning of the development of the resultative construction, every “verb + resultative” had to undergo the same procedure to fuse into a single element. Although the earliest appearances of VRO phrases were around the eighth century AD or even earlier, these phrases were not established as a syntactic pattern until the twelfth century AD. Only at this point could the resultative construction generalize other related phrases. When the resultative construction became a new syntactic pattern, the old verb co-ordination principle was abandoned, and since then the following organization has been ill-formed: (62)

(Vtr + Vtr. . .) + Obj

This construction was very common in Medieval Chinese, which maximally allowed four transitive verbs to share an object. The firm establishment of the verb–resultative pattern permitted more kinds of verb– resultative collocation. In the previous stage, at least the verb had to be transitive and had an action–patient relationship with the object. However, after the twelfth century AD, the verb did not have to be transitive and it was possible for there to be no direct grammatical relation to the object. In this case, only the whole meaning of a verb–resultative phrase could be understood to have an action–patient relationship to the object, illustrated as follows: (63)

哭損我一雙眼。 (張協狀元) Kū-sǔn wǒ yī-shuāng yǎn. cry-blind my one-pair eyes “(I) cried and caused my eyes to be blind.” (Zhang Xie Zhuang Yuan, AD 1200)

Here, neither kū “cry” nor sǔn “damage” can have yǎn “eye” as its object. The patient argument is assigned by the resultative construction rather than by individual elements. A resultative construction as a whole can assign a patient argument, where neither of the two components has to bear an action–patient relation to the object. In Chinese, this assignment of a patient argument became possible only after the VR pairs became a stable syntactic pattern.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

140

Resultative Construction

The emergence of the resultative construction is the most fundamental syntactic change in the history of Chinese grammar and had a profound effect on the texture of the syntax. This section provides a chronology of the resultative construction, reviews the source and motivation, and summarizes the effects. The development of the resultative construction covers a long time. It is difficult to accurately date when the first example came into being and when the construction was firmly established. Scholars may reach totally different conclusions about the time of the emergence of the resultative construction depending on the criteria they use to identify early examples (Zhou 1958, Shimura 1967, Wang 1989). The criterion used in this study is different from that used in others. I have identified the separable resultative structure as the direct source of the resultative construction. The formula for the separable resultative structure is as follows: (64) V + XP + R; XP = Objects, Adverbs, or Negatives. The formation of the resultative construction is essentially a fusion of the verb and the resultative; that is, the loss of the boundary between them. This change (i.e. reanalysis) does not involve any “immediate or intrinsic modification of its surface manifestation” (Langacker 1977) and thus is invisible in the initial stage. Considering the source, however, we can find a reliable formal criterion to identify when and where the fusion first took place. Logically, the fusion of the verb and the resultative is incompatible with their separability; in other words, once the verb and the resultative become fused, they cannot be separated by other material. The fusion gives rise to two word order shifts, depending on the syntactic categories of the original intervening words: (65) (a) VXR > Adv./Neg. + VR, if the original intervening word is an adverb or negative. (b) VXR > VR + O, if the original intervening word is the object of V. The fusion can be divided into two degrees – low and high. The change in (77a) means that the verb–resultative pair has reached the low degree of fusion, and the change in (77b) means that the verb–resultative pair has reached the high degree of fusion. The fronting of adverbs/negatives can take place as soon as the boundary between the verb and the resultative is weakened (the low degree). However, only when the boundary is entirely lost and the verb and the resultative become lexicalized into a verb compound (the high degree) can the verb–resultative phrase precede an object noun. According to these formal criteria, the chronology of the resultative construction can be stated as follows. In the eighth century AD, the VRO form and the fronting of adverbs/negatives started to appear, signaling the onset of the development of the

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

Extension

141

resultative construction. By the twelfth century AD, the resultative construction as a syntactic pattern had been established. There was a general fusion of the verb and the resultative, and the collocation of the verb and the resultative has been highly productive since then. By the fifteenth century AD, the resultative construction had entirely replaced the separable resultative structure. However, different types of resultative construction came into existence at different times. According to the grammatical relations of the resultative and other syntactic constituents, verb–resultative phrases can be divided into three types, each of which reaches the low or high degree of fusion at quite different times.

https://doi.org/10.1017/9781108921831.006 Published online by Cambridge University Press

7 Information Structure

7.1

Introduction

Changes in constituent orders, constructions, and grammatical markings are easily noticeable and thus usually attract much attention from researchers. However, changes in the information structure are hardly noticeable, and, to our knowledge, few scholars in historical Chinese linguistics have conducted any serious and systematic studies of issues of these sorts about the development of the information structure of a sentence. In fact, however, there were fundamental changes in two types of information structure: one is at the clause level, involving the arrangement of new and given information, and the other is at the predicate level, involving the arrangement of resultative and nonresultative (accompanying) information. The first is cross-linguistically true and can be found in many other languages (Hopper and Traugott 2003: 7); the second may be particular to the language and has thus far been known to exist only in Chinese. As we see throughout the entire history of Chinese grammar, these two principles of information organization are key for understanding how the grammar has evolved over time. On the one hand, both reflect the results of the grammatical developments; on the other hand, they motivated a series of significant changes in history that together have largely shaped the grammatical system of Contemporary Chinese. Information structure involves two key terms – topic and focus: the former expresses given information and the latter new information. When investigating the informationorganizing principle in a language, we must keep in mind that there are two strategies for arranging information: constituent ordering and grammatical marking.1 There is much confusion about the following four types of concept in Chinese linguistics, and different scholars may use the same term to refer to something quite different or even exactly opposite. First, these terms need to be clarified.2

1 2

Pitch is also recruited to indicate focus, as in English the focused element usually bears a high pitch. The definitions here are based mainly on Crystal (2008).

142

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure (a)

(b)

(c)

(d)

7.2

143

Definite and indefinite. The term “definite” refers to a specific, identifiable entity (or class of entities) and is usually contrasted with indefinite. In Chinese, it can be realized either by syntactic position (e.g. sentence-initial position) or by grammatical morpheme (e.g. marked by the disposal preposition bǎ). It is antonymous with “indefinite.” Given and new information. In terms of information structure, “given” information refers to information already supplied by previous linguistic contexts or social circumstances, as opposed to “new” information that has not previously been supplied. Topic and subject. “Topic” refers to given or definite information. The topic and comment form a binary characterization of sentence structures. The topic of a sentence is the entity about which something is said, whereas the further statement about the entity is the comment. It is said that Chinese is a “topicprominent” language, in contrast to “subject-prominent” languages such as English (Li and Thompson 1976). In Chinese, almost every constituent of a sentence which contains a nominal element can be topicalized by moving it to the front of a sentence. Although the topic often coincides with the subject of a sentence, they are not equal in Chinese grammar (for details, see Shi 2010: 165‒177). A subject can be topicalized by adding a mood particle (e.g. ne, ya, mo). There are essential distinctions between the subject–predicate construction and the topic–comment structure; for instance, the former can be used as an embedded clause, but the latter cannot. Focus and emphasis. “Focus” refers to the new, yet most important, information or the center of communicative interest that is assumed by the speaker, as opposed to presupposition. In Old Chinese, a constituent could be focused or highlighted in two ways: if a constituent was postverbal, it could be focused by being moved to the immediately preverbal position, and if it was preverbal, it could be focused by being moved to sentence-final position and by adding the focus particle yě to it. In Contemporary Chinese, a constituent can be highlighted simply by preceding it with the focus marker shì. In contrast, emphasis may not refer to new information, and there is no fixed grammatical morpheme for emphasizing an element in a sentence in Chinese grammar.

Information Structure

It is well known that how to arrange constituents with either given or new information influences not only the synchronic sentence structure but also the diachronic development of a language. Concerning this issue, many linguistic laws, theories, and hypotheses have been proposed in the literature. For instance, one of Behaghel’s (1909: 139)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

144

laws states that the given-information element precedes the new-information element. This idea was developed in discourse analysis by Chafe (1976), Givón (1984: 263), and many others. Faarlund (1990: 58) even claimed that it is a language universal to begin a sentence with old information and end it with new, and further explained the motivation for the word order change in Norwegian (for discussion, see Harris and Campbell 1995: 219). The diachronic evidence from the history of Chinese grammar shows that the syntactic position and grammatical morphemes for expressing new and given information have changed dramatically over time. Before we start to discuss the Chinese phenomena, let us consider the general situation across languages. For new information, there is no universal focus position across languages. For instance, in many languages, such as Hungarian (É Kiss 1987), Mayan languages (Aissen 1992), and Korean (Comrie 1988: 268), focus occurs at the immediately preverbal position. In some languages, such as Breton and Welsh (Dik 1980: 160), sentence-initial position is reserved for the focus.3 In contrast, in languages such as Aghem (Watters 1979), the focus position seems to be immediately postverbal. In the literature, preverbal focus is much more documented than sentence-final position. Focus position in a language tends to be in accordance with the site of wh- words; e.g. English is such a language (cf. Chapters 2 and 3).

7.3

The Change in Topicalization

It is assumed that there is a diachronic relationship between topic and subject, reflecting the degree of grammaticalization. Specifically, subjects are largely grammaticalized from topics (Li and Thompson 1976: 484, Hopper and Traugott 2003: 29). Therefore it is a logical assumption that English is more developed than Chinese because every finite clause in English requires a subject and topicalization in English is very limited. Nevertheless, every nominal constituent of a sentence in Chinese can be topicalized by moving it to the front of a sentence, although the subject is always optional and often omitted. This hypothesis cannot be true if we examine the evolution of Chinese grammar; topicalization was very common in Old Chinese and is even more frequent in Contemporary Chinese due to the complication of the predicate caused mainly by the establishment of the resultative construction. The study of topic and subject is one of the hottest areas in Chinese linguistics. There is much confusion and controversy about their status in Chinese grammar. Some

3

Note that the two terms “immediately preverbal position” and “sentence-initial position” are not equal, at least in Old Chinese, where the former is reserved for focus and the latter for topic, as we will discuss later (cf. Chapter 3).

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Change in Topicalization

145

researchers (e.g. Chao 1979: 45) believe that Chinese has only topics without subjects, but others (e.g. Zhu 1982: 95) hold exactly the opposite point of view. Additionally, others argue that Chinese has both categories, topic and subject, with different syntactic characteristics. Here, we would like to use two simple sentences to illustrate the distinction between subject and topic: (1)

The subject–predicate construction 我看了書。 (現代漢語) Wǒ kàn-le shū. I read-PERF book “I have read a book.” (Contemporary Chinese)

(2)

The topic–comment construction 書我看了。 (現代漢語) Shū wǒ kàn-le. book I read-PERF “I have read the book.” (Contemporary Chinese)

There are significant distinctions between the two structures illustrated above. First, in (1) the noun shū “book” is the object, automatically interpreted as indefinite, and in (2) it is the topic, interpreted as definite. The first constituent wǒ “I” in (1) is the subject, which can be highlighted by the focus marker shì and replaced by a wh- word, and the whole structure of (1) can be used as a relative clause to modify an NP within a sentence. In contrast, the first element shū “book” in (2) is the topic, which cannot be focused or replaced by a wh- word because it refers to given or presupposed information, and the whole structure of (2) can be used only as an independent sentence and cannot be used in an embedded clause. Thus, in our analysis throughout this book, the topic and the subject are treated differently because of these semantic and syntactic contrasts. It is cross-linguistically common to topicalize a nominal constituent by moving it to the front of a sentence. This strategy has been used from Old Chinese to the present day, but there are certain specific formal changes regarding topicalization over time to which attention needs to be paid (cf. Section 2.3). First, in Old Chinese, when an object was moved to the front of a sentence to be topicalized, its original position was usually filled with an anaphor zhī “3rd pronoun,” as exemplified in (3). However, in Contemporary Chinese, adding an anaphoric pronoun will make the sentence ill-formed, as illustrated in (4):

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

146 (3)




吾道一以貫之。 (論語 里仁) Wú dào yī yǐ guàn zhī. my theory one with unify it “My theory can be unified with one (theme).” (Lun Yu, Li Ren, 500 BC)

(4)

*書我看了它。 (現代漢語) *Shū wǒ kàn-le tā. book I read-PERF it “The book, I have read it.” (Contemporary Chinese)

Second, in Contemporary Chinese, the subject can be topicalized by adding a suffixlike particle (e.g. ne, me, or ya, usually occurring at sentence-final position) after it, as illustrated in (5). In Old Chinese, the subject can be topicalized by adding a prefix-like particle fú before the subject, as illustrated in (6): (5)

老王嘛, 早就不吸煙了。 (現代漢語) Lǎo Wáng ma, zǎo jiù bù xīyān le. Lao Wang PRT early then not smoke-cigarette ASP “As for Lao Wang, he stopped smoking a long time ago.” (Contemporary Chinese)

(6)

夫民, 神之主也。 (左傳 桓公六年) Fú mín, shén zhī zhǔ yě. PRT people, god GEN master COP “As for the people, they are the master of the god.” (Zuo Zhuan, Huan Gong Liu Nian, 550–400 BC)




Third, in Contemporary Chinese, when the object of a preposition is topicalized, its original position is usually filled by an anaphoric pronoun, as illustrated in (7). This topicalization structure did not exist in Old Chinese, and is actually one of the most recent developments. (7)

這把刀我用它來切水果。 (現代漢語) Zhè-bǎ dāo wǒ yòng tā lái qiē shuǐguǒ. this-CL knife I use it come cut fruit “As for this knife, I use it to cut fruit.” (Contemporary Chinese)

As we mentioned previously, there is no diachronic relationship between topic and subject. In the past three millennia, we found no tendency for topic to develop into

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Change in Focusing

147

subject in Chinese grammar. In contrast, the topicalization construction in Contemporary Chinese is even more common than before. For instance, in Old Chinese, topicalization was always an optional device with certain pragmatic values, which means that each instance of the topicalization construction had a corresponding declarative sentence with the canonical structure. Due to the structural complications of the predicate, in many situations in Contemporary Chinese the topicalization construction is the only option available to introduce the patient noun, as in the following examples:




(8)

老爺的雨衣你給我放在哪兒啦? (曹禺 雷雨) Lǎoyé de yǔyī nǐ gěi wǒ fàng zài nǎ’er la? lord GEN raincoat you for me put at where PRT “Where did you put the raincoat of our lord?” (Cao Yu, Lei Yu, Contemporary Chinese)

(9)

這地方我絕對不會再住下去。 (曹禺 雷雨) Zhè dìfāng wǒ juéduì bù huì zài zhù-xiàqù. this place I absolutely not can again live-continue “In this place, I absolutely cannot live any longer.” (Cao Yu, Lei Yu, Contemporary Chinese)




In (8), the predicate is the matrix verb plus a resultative preposition phrase, which disallows any following object. In (9), the matrix verb is suffixed by the continuous aspect xiàqù, so no patient object can occur after the predicate. The above phenomenon is probably responsible for the formal distinction of the topicalization construction between Old Chinese and Contemporary Chinse: the topic in Old Chinese typically had an anaphoric pronoun to fill the original position of the fronted object because each topicalization had a corresponding canonical structure; however, in Contemporary Chinese, many topicalizing instances have no corresponding normal declarative structure, which means that there is no position for an anaphoric pronoun. This restriction makes it impossible for an anaphor to follow, which eventually became a new grammatical rule for topicalizing.

7.4

The Change in Focusing

The devices that make a constituent focused underwent a dramatic historical change (for a fuller discussion, see Chapter 3). The chief means in Old Chinese was to move an element to the immediately preverbal position. Then, there was a neat division of labor between sentence-initial position and the immediate position that was somewhere

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

148

between the subject and the predicate: the former was reserved for the topic, referring to given information, and the latter was designed for the focus, expressing new information. This division can be formulated as follows, where the topic and focus never cooccur within a single sentence. (10) (a) The topicalization construction: TOP + Subj + VP (b) The focusing construction: Subj + FOC + VP In Old Chinese, four types of word typically became focused through manipulating constituent orderings. First, wh- words had to be moved to preverbal position when used as the object, as discussed in Chapter 3, but ordinary nouns did not. There is a cross-linguistic discrepancy in the location of the landing site for the movement of wh- words: in English, it is sentence-initial position, and in Chinese, it is preverbal position between the subject and the predicate. The specific landing site for wh- words may be influenced by whether the topic is prominent in the language. As in Contemporary Chinese, the topicalization construction in Old Chinese was very common, so sentence-initial position was specialized for the topicalized element, which forced the focused constituent to occur after the subject instead. In contrast, topicalization is much rarer in English, so sentence-initial position is usually available for the focus. As seen in Chapter 3, in Old Chinese, there was a rigorous rule that a wh- word object must be inverted with the verb or preposition, but an ordinary noun object followed the verb or preposition instead. Cross-linguistically, the syntactic behaviors of wh- words in a language often differ from those of ordinary nouns because wh- words possess an inherent “focus” feature that must be realized in some way. Therefore the landing site of wh- words is usually identical to the position of the focus, a phenomenon that has been well documented across languages. This was also the case in Old Chinese. (11)




子將誰驅? (墨子 耕柱) Zǐ jiāng shuí qū? You will which drive? “Which (animal) will you use to drive (your cart)?” (Mo Zi, Geng Zhu, 350 BC)

If the object was an ordinary noun, it was most often marked by the affix wéi “only” and is also usually followed by the anaphor shì or zhī “this,” as illustrated in (12). That is, when the highlighted element was a noun, two devices, a grammatical morpheme and constituent order change, were employed to perform the highlighting at the same time, differentiating it from wh- words, as illustrated below:

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Change in Focusing (12)

149




吾唯子之怨。 (左傳 文公七年) Wú wéi zǐ zhī yuan. I only you ANAP hate “It is you who I hate.” (Zuo Zhuan, Wen Gong Qi Nian, 550–400 BC)

The above two types of focusing construction have been studied quite thoroughly (e.g. Wang 1989: 198‒210, Yang and He 2001: 784‒809). However, few scholars have realized that there was another type of focusing construction that happened mainly in the preposition phrase marked by yǐ, one of the most common prepositions in Old Chinese, which exhibited multiple functions, such as introducing instruments, directions, or reasons. The yǐ preposition phrase could occur either before or after the predicate. The preposition yǐ had some unusual behaviors; for instance, it could follow the object, unlike other prepositions, seemingly surfacing as an “NP + Postposition” structure. However, scrutiny reveals that this pattern actually reflected the focusing structure, which can be proven by two pieces of evidence: first, the yǐ phrase could occur in two positions, either preverbal or postverbal, but only in the preverbal position could the object of the preposition yǐ be inverted with the preposition, with the NP actually occupying the focus position, as illustrated in (13). Second, when the focused NP was fronted to before the preposition yǐ, an anaphoric zhī was inserted between them, as illustrated in (14).




(13)

君子義以為質。 (論語 衛靈公) Jūnzǐ yì yǐ wéi zhì. noble-man righteousness with be basic “It is righteousness that noble men consider basic.” (Lun Yu, Wei Ling Gong, 500 BC)

(14)

我之不共, 魯故之以。(左傳 昭公三年) Wǒ zhī bù gòng, Lǔ gù zhī yǐ. I NOM not share Lu reason ANAP for “It was for the reason of the Lu country that I didn’t share.” (Zuo Zhuan, Zhao Gong San Nian, 550–400 BC)




The example in (14) was the same operation for highlighting the NP of ordinary verbs in the focusing construction in (12). In addition to the above focusing method of manipulating constituent order, another means existed for the same purpose: moving the focused constituent to sentence-final position and adding the focus morpheme yě to it. In Section 2.2 we discussed the development in Old Chinese of the copular particle yě into a focus

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

150

marker, which often marked a focused element at sentence-final position. The focus particle yě could also highlight a preposition phrase in terms of how the preposition phrase was extracted to sentence-final position and was followed by the focus particle yě, as illustrated in (15): (15)




城小穀, 為管仲也。 (左傳 莊公三十二年) Chéng Xiǎo Gǔ, wèi Guǎn Zhòng yě. city-build Xiao Gu for Guan Zhong FOC “It was for Guang Zhong to build a city in Xiao Gu.” (Zuo Zhuan, Zhuang Gong San Shi Er Nian, 550–400 BC)

In a normal declarative sentence, the preposition phrase wèi Guǎn Zhòng “for Guang Zhong” should occur before the matrix verb. In Contemporary Chinese, the preposition phrase is highlighted in its original position, without extraction, by inserting the focus marker shì in front of it. In Old Chinese, the focusing yě construction was needed by the grammatical system. When a sentence had no verb or preposition (or the noun phrase to be focused was not the object of either a verb or a preposition), and when the focused element was an adverbial phrase whose original position was already between the subject and the predicate, the other device of constituent ordering was incapable of focusing the element; in this case, the yě focus construction became the only option available. In (15), the constituent to be highlighted is a causative preposition phrase whose canonical position was preverbal; hence it had to be moved to sentence-final position to be highlighted.

7.5

The Focus Construction

The above methods of focusing disappeared in the first half of the Medieval Chinese period. The disappearances were caused by a significant historical event in which the demonstrative shì grammaticalized into a copula around the first century BC and further developed as a focus marker. It is quite straightforward to make a constituent focused in Contemporary Chinse simply by adding the focus marker shì before the element that needs to be highlighted. Except for the topic that refers to given information and contradicts the meaning of the focus, in principle every constituent in preverbal position that contains a nominal or pronominal element (including noun phrases, preposition phrases, and time words) can be focused by adding the morpheme shì. In the following example, each of the underlined parts can be focused in this way:

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Changes in Assigning Definiteness (16)

151

我昨天在草地上把自行車修好了。 (現代漢語) Wǒ zuótiān zài cǎodì shàng bǎ zìxíngchē xiūhǎo le. I yesterday in lawn on DISP bicycle fix-well PERF “I fixed my bicycle on the lawn yesterday.” (Contemporary Chinese)

Note that the object cannot be directly highlighted simply by adding the focus marker shì to it. There are two steps to focusing an object: first, use the complementizer de to turn the part before the object into a clause; second, insert shì between the clause and the object, as illustrated in (17): (17)

我昨天看見的是一隻猴子。 (現代漢語) Wǒ zuótiān kànjiàn de shì yī-zhǐ hóuzi. I yesterday see COMP be one-CL monkey “What I saw yesterday was a monkey.” (Contemporary Chinese)

However, when a sentence already contains a wh- word, the wh- word will become the natural focus, as it is the only element that can be focused in the sentence. That is, only this wh- word can be highlighted by the focus marker shì. For example, if the word zuótiān “yesterday” was replaced by shénme shíhòu “when,” then only this wh- word could be focused by adding the copula shì. In the history of Chinese grammar, there was a typological change in the focusing construction that had a far-reaching impact on the texture of Chinese grammar (for details, see Section 3.5.3).

7.6

Changes in Assigning Definiteness

The grammatical devices for expressing definiteness also underwent a dramatic historical change. Most nouns can be made definite by adding determiners such as demonstratives or personal pronouns, and this lexical device marking definiteness has remained the same throughout history. We will focus on grammatical devices that express definiteness, such as syntactic positions, grammatical morphemes (e.g. the disposal marker bǎ), and certain morphological methods (e.g. reduplication). Contemporary Chinese has rich grammatical devices for expressing definiteness that did not exist in Old Chinese.

7.6.1

Principle of Definiteness Assignment by Syntactic Position

Some researchers have claimed that in the Chinese language, noun phrases in preverbal position tend to be definite (e.g. Cheng and Sybesma 1999, Jenks 2018).

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

152

This generalization is problematic and fails to capture the rule of definiteness assigning. “Noun phrases in preverbal position” may refer to very different constituents in various constructions; for instance, is the noun phrase located in the topic position, in the subject position, or between the subject and predicate positions? Which preposition is used to mark the preverbal NP? Is it a bare noun or a noun phrase, and if it is a noun phrase, is its determiner a demonstrative or a “Num + CL” phrase? All of the factors may play a part in determining a definite expression. There is an interplay in assigning a “definite” feature to a noun phrase. To uncover the mechanism behind assigning definiteness, the conditions need to be refined. For bare nouns, there are two exception-free rules for assigning a “definite” feature: the first rule is that if they can occur either before or after the matrix verb, namely that both structures are grammatical, then a noun in preverbal position is automatically assigned the feature definite, and one in postverbal position is naturally given the feature indefinite. (18)

(a) 客人来了。 (現代漢語) Kèrén lái-le. guest come-PERF “The guest has come.” (b) 来了客人。 Lái-le kèrén. come-PERF guest “A guest has come.” (Contemporary Chinese)

(19)

(a) 書我已經還了。 (現代漢語) Shū wǒ yǐjīng huán-le. book I already return-PERF “I have returned the book.” (b) 我已經還了書。 (現代漢語) Wǒ yǐjīng huán-le shū. I already return-PERF book “I have already returned a book.” (Contemporary Chinese)

The second rule is that if the matrix verb is transitive and the whole sentence describes a concrete activity, a bare noun in the subject position of either an active or a passive sentence must be interpreted as definite, as illustrated below:

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Changes in Assigning Definiteness (20)

153

孩子已經睡覺了。 (現代漢語) Háizi yǐjīng shuìjiào-le. child already sleep-PERF “The child has slept.” (Contemporary Chinese)

(21)

房間已經被打掃乾淨了。(現代漢語) Fángjiān yǐjīng bèi dǎsǎo-gānjìng le. room already PASS sweep-clean PERF “The room has been cleaned.” (Contemporary Chinese)

On the basis of the above phenomena, however, we cannot jump to the conclusion that all noun phrases in preverbal position must be interpreted as definite. What we can say is that the default meaning of a bare noun in preverbal position is definite, a feature assigned by the syntactic position, but this assignment by syntactic position can be overridden by the marking of lexical forms. For instance, all the bare nouns in these positions can be made indefinite by adding a function word yǒu plus a “Num + CL” phrase, which is equal to the existential structure there is some in English, as illustrated in (22): (22)

有個客人來了。 (現代漢語) Yǒu-gè kèrén lái-le. have CL guest come-PERF “A guest has come.” (Contemporary Chinese)

(23)

有些孩子已經睡覺了。(現代漢語) Yǒu-xiē háizi yǐjīng shuìjiào-le. have-some child already sleep-PERF “Some children have slept.” (Contemporary Chinese)

The same is true for a bare noun in postverbal position, where its default meaning is indefinite. However, it may be made definite by adding demonstratives, personal pronouns, and the like, as in the following examples: (24)

我已經交了那次作業。(現代漢語) Wǒ yǐjīng jiāo-le nà-cì zuòyè. I already hand-in-PERF that-CL homework “I have already handed in the homework.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

154 (25)

我已經打掃乾淨了那個房間。 (現代漢語) Wǒ yǐjīng dǎsǎo-gānjìng-le nà-gè fángjiān. I already sweep-clean-PERF that-CL room. “I have already cleaned the room.” (Contemporary Chinese)

If the verb is intransitive and the object is actually the agent, the bare noun of the object cannot be made definite by adding those lexical forms; for instance, the following sentence is ungrammatical: (26)

*來了那個客人。 (現代漢語) *Lái-le nà-gè kèrén. Come-PERF that-CL guest “The guest came.”

On the basis of the above analysis, we can draw the following generalizations about the principle of definiteness-assignment by syntactic position. (27)

Principle of definiteness assignment by syntactic position

(a) If a noun is a patient noun and the sentence has a corresponding VO structure, a bare noun in preverbal position, used as either subject or topic,4 must be assigned the feature definite. (b) If the matrix verb is transitive and the sentence depicts a concrete activity, a bare noun in preverbal position, usually used as the subject, must be assigned the feature definite. (c) The above two rules can be overridden by lexical forms; that is, bare nouns in preverbal position can be made indefinite by adding forms with indefinite meanings, such as yǒu “exist.” In addition to the syntactic-position principle, Contemporary Chinese has a major grammatical device, usually called the “disposal construction,” which is specialized to make the object NP definite. Both this principle and the disposal construction were introduced into the language in Medieval Chinese, and a cause–effect relation existed between these two changes, a topic we will discuss in detail in Chapter 9. It has been widely accepted in the literature that the noun phrase that is introduced by the disposal marker bǎ is definite (Lü 1984: 176‒199, Ding et al. 1961: 95‒98, Li and Thompson 1981: 463‒491, Zhu 1982: 185‒186). Whether it is a rigorous rule or just a tendency depends on what empirical data the researchers

4

Bare nouns in the disposal construction are also assigned the feature definite, which will be discussed in detail in Chapter 9.

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Changes in Assigning Definiteness

155

observe and how they handle seemingly complex phenomena. For example, Zhu (1982: 185) claimed that this is a rule, and violating it will create an ungrammatical sentence. However, Li and Thompson (1981: 170) considered it merely a tendency: “Nouns preceding the verb tend to be definite, while those following the verb tend to be indefinite.” To discover the truth behind this phenomenon, we need to separate two things: (a) the default function and (b) the conditioned usages of the disposal construction, which are stated as follows. (b)

(b)

7.6.2

The default function of the disposal construction. If the object of bǎ is a bare noun, namely without any modifiers such as demonstratives or numerals, and if this patient noun can occur in either preverbal or postverbal position (both well formed), then the disposal construction automatically assigns the feature definite to the preverbal noun phrase and the feature indefinite to the postverbal noun phrase. This is a rigorous rule with no exceptions (for details, see Section 9.3.1). The conditioned usage of the disposal construction. The object of bǎ is not necessarily definite and can be indefinite or generic under the following two conditions: first, the patient noun must be modified by a numeral phrase or the like; second, the predicate is structurally complex and cannot be followed by any object.

Assigning Definiteness in History

The principle of assigning definiteness by syntactic position did not, however, exist in Old Chinese, and bare nouns in preverbal position were usually indefinite; therefore some grammatical morpheme such as yǒu “there exist” must be added when translating such constructions into Contemporary Chinese, as illustrated below: (28)




門人問曰: “何謂也?” (論語 里仁) Ménrén wèn yuē: “Hé wèi yě?” student asked say what call PRT “A student asked: ‘What is it?’” (Lun Yu, Li Ren, 500 BC)

(29)




鄉人飲酒。 (論語 鄉黨) Xiāngrén yǐn jiǔ. Villagers drink wine “Villagers drank wine.” (Lun Yu, Xiang Dang, 500 BC)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

156 (30)




人曰: “子未可以去乎?” (論語 微子) Rén yuē: “Zi wèi kěyǐ qù hū?” people say you non can leave Q “People said: ‘Can’t you leave?’” (Lun Yu, Wei Zi, 500 BC)

Around the third century BC, the grammatical morpheme yǒu, which meant existence, like English some, started to appear. In the following example, the verb is preceded by the temporal noun jīn “now,” forming a mini existential clause in the subject position. However, this usage was still uncommon at the time. (31)




今有人日攘其鄰之雞者。 (孟子 滕文公) Jīn yǒu rén rì rǎng qí lín zhī jī zhě. now have people everyday steal his neighbor GEN chicken who “There is someone who steals his neighbor’s chicken everyday.” (Meng Zi, Teng Wen Gong, 300 BC)

By the fifth century AD, both bare nouns and those marked by yǒu “exist” in the subject position, which were interpreted as indefinite, were equally common, as illustrated in (32) and (33). (32)




人問其故。 (世說新語 德行) Rén wèn qí gù. people ask his reason “People asked his reason.” (Shi Shuo Xin Yu, De Xing, AD 450)

(33)




有人止之。 (世說新語 方正) Yǒu rén zhǐ zhī. exist people stop he “Somebody stopped him.” (Shi Shuo Xin Yu, Fang Zheng, AD 450)

It is difficult to judge the accurate date of the emergence of the principle that assigns definiteness by syntactic position. A reliable criterion for doing so may be the frequency increase of the topicalization construction in which the bare noun in topic position must be interpreted as definite, as illustrated in (34): (34)

錢財奴婢用。 (王梵志詩) Qián cái núbì yòng. money treasure servant use “His money and treasure were used by the servants.” (Wang Fan Zhi Shi, AD 650)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Non-question Functions of Wh- Words

157

According to Sun (1997: 136), instances of this kind were rare before the sixth century AD but then became increasingly popular, signaling the gradual formation of the principle in (27). What we can say here is that the tendency might have started as early as the first century BC but was not established as a principle until the seventh century AD.

7.7

Non-question Functions of Wh- Words

The establishment of the principle of assigning definiteness by syntactic position had a far-reaching impact on the grammatical system of Chinese. This principle is summarized as follows: (35)

The default meaning of preverbal position is definite; that of postverbal position is indefinite.

This principle did not exist in Old Chinese because the immediately preverbal position at that time was reserved for the focus, which referred to new information. Although new information might sometimes be definite, none of the bare nouns in preverbal position that are interpreted as definite could be highlighted by the focus marker shì in Contemporary Chinese, signaling that semantic conflict exists between the focusing and definite interpretation. Thus we believe that assigning focus in preverbal position in Old Chinese blocked the possibility of assigning definiteness in the same position in Contemporary Chinese. This might be why the principle could not enter the language until the focusing construction in Old Chinese disappeared around the fifth century AD. One of the direct consequences of the establishment of the principle is that it enabled all Chinese wh- words to develop two non-question usages: (a) universal reference and (b) indefinite reference. From a typological perspective, wh- words in any language have the potential to be extended to express universal and indefinite references, but how and whether to realize the two usages depends largely on the property of the particular grammar. In English, for instance, these references are realized by adding certain modifiers, e.g. whatever/somewhat, however/somehow, wherever/somewhere, where in each pair the left item creates a universal reference and the right one an indefinite meaning. Some English whwords have only the compound word of universal reference, such as whoever versus *somewho and whenever versus *somewhen. Except for shúi “who,” which has survived throughout history, the wh- words in Contemporary Chinese came into existence in the last half of the Medieval Chinese period, a period from the sixth century AD to the tenth (Lü 1985: 104). This means that

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

158

the majority of former wh- words in Old Chinese were abandoned in the same period (for details, see Shi and Li 2001: 33‒53). There was a fundamental distinction between the former and latter sets of wh- words: none of the former set developed the two noninterrogative uses, namely universal and indefinite reference, but those in the latter set all have the two extended meanings. According to Lü’s (1985: 104) investigation, the wh- word shúi “who” was attested in Shi Jing, a text composed from the eleventh century BC to seventh, but it was not extended to express a universal reference before Modern Chinese, as illustrated in the following example: (36)

這種毒藥誰都願意喫的。 (老殘遊記二十回) Zhè zhǒng dúyào shéi dōu yuànyì chī de. this-CL poison everyone all willing eat PRT “Everyone is willing to eat this kind of ‘poison.’” (Lao Can You Ji, Chapter 20, AD 1900)

The universal reference covers every member in a definite domain, namely involving a definite meaning. In contrast to English, which attaches ever to a wh- word to gain a universal reference and some to form a compound with an indefinite meaning, Chinese does not allow wh- words to be modified by any such quantifiers to form a compound. Once the principle of assigning definiteness by syntactic position was established, the wh- words in Chinese could receive a definite interpretation in preverbal position and an indefinite interpretation in postverbal position. As a consequence, in Contemporary Chinese, all wh- words have acquired these two noninterrogative usages, namely universal and indefinite references, but these interpretations are conditioned by their syntactic positions: the interpretation of universal reference can be obtained only in preverbal position, as illustrated in (37), and the interpretation of indefinite reference can be obtained only in postverbal position, as illustrated in (38): (37)

他哪裡都去過。 (現代漢語) Tā nǎlǐ dōu qùg-uò. he where all visit-EXP “He visited everywhere.” (Contemporary Chinese)

(38)

我想喫點兒什麼。 (現代漢語) Wǒ xiǎng chī diǎner shénme. I want eat some what “I want to eat something.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Non-question Functions of Wh- Words

159

It is remarkably characteristic of Chinese grammar that the grammatical functions of certain words can be realized only at particular positions. Regardless of what semantic role a wh- word plays, for example, either agent or patient, the universal-reference use must occur somewhere in preverbal position, either before or after the subject, as illustrated in (39): (39)

(a) 誰我都通知到了。(現代漢語) Shúi wǒ dōu tōngzhī-dào-le. who I all inform-to-PERF I have informed everyone. (Contemporary Chinese) (b) 我誰都通知到了。(現代漢語) Wǒ shúi dōu tōngzhī-dào-le. I who all inform-to-PERF “I have informed everyone.” (Contemporary Chinese)

Additionally, the reduplication of classifiers and certain nouns exhibits exactly the same semantic feature and syntactic behavior as the universal-reference usage of whwords. In Contemporary Chinese, almost all monosyllabic classifiers (including two highly frequent nouns: rén “people” and shì “thing”) can be reduplicated to create a universal reference, but they are restrained to preverbal position, as illustrated below. (40)

(a) 我人人都通知到了。 (現代漢語) Wǒ rén-rén dōu tōngzhī-dàole. I person-person all inform-to-PERF “I have informed everyone.” (b) *我都通知到了人人。 *Wǒ dōu tōngzhī-dào-le rén-rén. I all inform-to-PERF person-person (Contemporary Chinese)

(41)

(a) 個個我都認識。(現代漢語) Gè-gè wǒ dōu rènshí. CL-CL I all know “I know all of them.” (b) *我都認識個個。(現代漢語) *Wǒ dōu rènshí gè-gè. I all know CL-CL (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

160

Like wh- words, the reduplicated forms of classifiers cannot be modified by any determiners such as demonstratives; for the same reason, their interpretation is based solely on their syntactic position. Classifiers as a grammatical category were introduced into the language on a large scale in Medieval Chinese, and they took over the function of nominal reduplication of monosyllabic nouns in terms of marking universal reference. The shift between classifiers and nouns happened in the last half of the Medieval Chinese period (for details, see Section 16.5). Nonetheless, this grammatical feature is still preserved for the two most frequent nouns, namely, rén “people” and shì “thing,” because of their high frequency over time. Cross-linguistically, high-frequency words tend to preserve old grammatical features. However, all other nouns lost the function of being reduplicated for a universal reference.

7.8

The Existential Construction

The rule in (27b) of the principle of assigning definiteness by syntactic position states that the default meaning of the noun phrase in postverbal position is indefinite, a feature closely related to the meaning of the existential construction. In reality, the establishment of this principle caused a significant historical change in the existential structure. Before we discuss the Chinese phenomenon, let us examine the structure in English. There is used as the dummy subject because every finite clause in English requires a subject; the verbs occurring in this structure are very limited, mainly be and exist, and the subject is typically indefinite and is often modified by the indefinite article a/an, some, several, or any. Before the tenth century AD, the existential construction in Chinese closely resembled its counterpart in English, but Contemporary Chinese actually allows all concrete-action verbs to appear in the existential construction, producing expressions that literally mean “There is writing my name,” “There is sleeping a person,” etc., as illustrated in (42) and (43): (42)

書上寫著我的名字。(現代漢語) Shū shàng xiě-zhe wǒ de míngzì. book on write-PROSS I GEN name “My name is written on the book.” (Contemporary Chinese)

(43)

鍋里炒著菜呢。 (現代漢語) Guō lǐ chǎo-zhe cài ne. Pan inside fry-PROSS vegetable PRT “Vegetables are being fried in the pan.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Existential Construction

161

Before the tenth century AD, these existential expressions such as those illustrated above were impossible because no concrete verb could precede an existential subject. The canonical existential structure before Modern Chinese was formulated as follows: (44)

(Yǒu “have” + NP) + PP + VP.

The verb yǒu “exist” marks the indefinite subject, the PP indicates the location, and VP may contain a concrete verb, as illustrated in (45) and (46):




(45)

有蛇自泉宮出。 (左傳 文公二年) Yǒu shé zì Quán Gōng chū. have snake from Spring Palace exit “A snake emerged from Spring Palace.” (Zuo Zhuan, Wen Gong Er Nian, 550–400 BC)

(46)

有一人從橋下走出。 (史記 張釋之馮唐列傳) Yǒu yīrén cóng qiáo xià zǒu-chū. have one person from bridge under walk-out “Someone came out from under the bridge.” (Shi Ji, Zhang Shi Zhi Feng Tang Lie Zhuan, 100 BC)




Although the above structure is still in use in Contemporary Chinese, the more common existential structure is “PP + VP + NP,” where the sentence begins with a locative phrase, has the concrete verb as the center and ends with the existential subject, as illustrated in (47) and (48). This structure did not exist before the tenth century AD. (47)

從泉宮裡爬出一條蛇。 (現代漢語) Cóng Quán Gōng lǐ pá-chū yī-tiáo shé. from Spring Palace inside crawl-exit one-CL snake “A snake crawled out of Spring Palace.” (Contemporary Chinese)

(48)

從橋下走出來一個人。 (現代漢語) Cóng qiáo xià zǒu-chūlái yīgè rén. from bridge under walk-exit one-CL person “A person is walking out from under the bridge.” (Contemporary Chinese)

The key characteristic of the existential construction is that the subject is indefinite, which must be marked in some way, either by lexical markers or by syntactic position, depending on the grammatical system. To do so, English uses the locative adverb there plus the existential verb be. Similarly, Old Chinese used the grammatical morpheme yǒu “exist.” When the principle of assigning

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

162

definiteness by syntactic position became effective, the feature of indefiniteness of the existential subject could be realized in postverbal position because any NP in postverbal position was automatically given the default meaning “indefinite.” Not long after the principle was created, instances of the new type of existential construction started to appear, and it has steadily developed since then. Some of the earliest examples are illustrated in (49) and (50) (Chu et al. 1997):




(49)

面上貼花紅。 (敦煌變文集 下女夫詞) Miàn shàng tiē huā hóng. face one stick flower red “A red flower was pasted on her face.” (Dun Huang Bian Wen, Xia Nü Fu Ci, 800–1000)

(50)

背上馱著一個搭膊。 (錯斬崔寧) Bèi shàng tuó-zhe yīgè dā-bó. back on carry-PROSS one-CL backpack “A backpack was being carried on his back.” (Cuo Zhan Cui Ning, AD 1100)

At a later stage, the existential structure was greatly expanded so that several quite unusual structures were innovated. For example, one of the hottest topics in Chinese linguistics has been the fact that a noun phrase in object position can actually be the agent of the matrix verb, as illustrated in (51) and (52): (51)

萬秀娘死了夫婿。 (萬秀娘仇報山亭兒) Wàn Xiù Niáng sǐ-le fū xù Wan Xiu Lady die-PERF husband “Lady Wan Xiu’s husband died.” (The Vernacular Novels of the Song Dynasty, AD 1100)

(52)

(王冕)七歲時死了父親。 (儒林外史一回) (Wáng miǎn) qī-suì shí sǐ-le fùqīn. Wang Mian seven-year when die-PERF father “Wang Mian’s father died when he was seven years old.” (Ru Lin Wai Shi, Chapter 1, AD 1750)

In the above sentences, the subject is construed as a location, indicating that something has happened to someone. Although the noun phrases in postverbal position are semantically definite and identifiable from the context, they are grammatically treated as indefinite and hence cannot be modified by any definite determiner such as tā-de “her/his”; for instance, the following example is ungrammatical:

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Existential Construction (53)

163

*萬綉娘死了她的夫婿。 *Wàn Xiù Niáng sǐ-le tā de fūxù. Wan Xiu Lady die-PERF she GEN husband “Lady Wan Xiu’s husband died.”

Even in Contemporary Chinese, if the verb is intransitive and the noun phrase in postverbal position is still incompatible with any definite determiners, an ill-formed instance will be created, as illustrated below: (54)

(a) 他們家來了一個客人。 (現代漢語) Tāmen jiā láile yīgè kèrén. their family come-PERF one-CL guest “A guest came to their house.” (b) *他們家來了那個客人。 *Tāmen jiā láile nà-gè kèrén. their family come-PERF that-CL guest “The guest came to their family.” (Contemporary Chinese)

However, if the verb is transitive and bears a VO relation to the postverbal noun phrase, the noun phrase can be made definite by adding definite determiners to it, as illustrated in (55). (55)

書上寫著我的名字。 (現代漢語) Shū shàng xiězhe wǒ de míngzì. book on write-PROSS I GEN name “My name is written on the book.” (Contemporary Chinese)

Any changes typically involve many factors. When we try to explain why a new grammatical device is invented, we must look at the big picture of the grammatical system at the time. In particular, when we are dealing with a synchronic phenomenon, we cannot find a plausible explanation unless we investigate the historical motivation for its emergence. As we have seen, for instance, the syntax of the verb sǐ “die” in Contemporary Chinese is peculiar because it can be followed by an object that is actually the agent of the verb of dying, a usage that did not exist before Modern Chinese and has not been found in other languages. Thus this phenomenon has attracted much attention in the realm of Chinese linguistics, and many researchers have explored it with different frameworks in the past six decades or so (e.g. Xu 1999, Shen 2006), but they have focused on the property of the particular verb sǐ “die” from a purely synchronic perspective. As we have

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

164

seen, the instance of the verb of dying merely manifests the change in the whole existential structure in response to the establishment of the principle of assigning definiteness by syntactic position. The systematic property of grammar becomes more clear in the diachronic process, from which we can observe the relationship among different parts.

7.9

The Principle of Action–Resultative Ordering

There are two principles determining the information structure at work in the grammar of Chinese. The first principle operates at the clause level, at which, as we have discussed above, the linear order of the basic constituents is arranged based on “new” and “given” information and involves topicalization, highlighting, and wh- word movement. This phenomenon is common across languages and has been well documented. The second principle operates at the predicate level only and governs the linear ordering of the constituents on the basis of their semantic features, whether resultative or nonresultative; it involves the distribution of adverbs, adjectives, prepositions, and temporal words. This action–resultative ordering principle has not been found in other languages, but it plays a key role in the grammar of Chinese. Before we address the historical motivation for its emergence, let us first examine how it operates in Contemporary Chinese. At first glance, the principle of action–resultative ordering is very straightforward: the constituent that refers to the resultative of the verb must occur in postverbal position, and the constituent that refers to the accompanying feature (non-resultative) of the verb must occur in preverbal position. This can be schematized as follows: (56)

Principle of action–resultative ordering (a) VP + Resultative (b) Non-resultative + VP

It seems that the linear ordering depicted above reflects iconicity in reality: the non-resultative entity goes first because it exists prior to the action, but the resultative entity comes last because it happens only after the action. In other words, the linear ordering determined by the principle reflects the temporal sequence of the events involved. Indeed, “iconic” versus “symbolic” constitutes a dimension for determining constituent order (Vincent 1979, Haiman 1985: 196‒ 229). Regarding surface value, Contemporary Chinese is more iconic, but Old Chinese was more symbolic. However, the formation of this principle is the byproduct of the establishment of the resultative construction.

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Principle of Action–Resultative Ordering

165

(a) The distribution of time words Time words fall into two types: (a) those referring to a temporal point or location, such as yesterday, last month, and three o’clock, providing information about when, and (b) those indicating duration or times of repetition, such as twice, three hours, and one week, providing information about how many times or how long. Temporal phrases in English generally occur in sentence-final position, where phrases of temporal point or location are usually introduced by certain prepositions such as at, on, or in, and phrases of temporal duration are typically introduced by the preposition for. Some time words in English do not need any preposition to introduce them, and their distribution is quite free, as they can occur either at sentence-initial position or sentence-final position, e.g. “Yesterday I went downtown” or “I went downtown yesterday.” Although prepositions are rich and play a crucial part in the grammar of Chinese, unlike English Chinese employs just syntactic positions and no prepositions to distinguish the two types of temporal expression. Words of time point or location can occur only in preverbal position, as illustrated in (57), but those of temporal duration or repetition are constrained to postverbal position, as illustrated in (58): (57)

他昨天去爬山了。 (現代漢語) Tā zuótiān qù pá shān le. he yesterday go climb mountain PERF “He went to climb the mountain yesterday.” (Contemporary Chinese)

(58)

我已經看過兩遍了。 (現代漢語) Wǒ yǐjīng kànguò liǎng-biàn le. I already read-EXP two-CL PERF “I have read it twice.” (Contemporary Chinese)

Obviously, time points or locations are independent of the action and are not the result of the action of the verb, while the duration or time of repetition is a kind of result of the action. The distributions of these words reflect the action–resultative principle. However, if the action does not happen at all, in the case of complete negation, then the time words of duration or repetition usually occur in preverbal position, as illustrated in (59) and (60), because they are not the result of any action, as there was no action, but rather indicate the temporal frame in which the action did not happen:

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

166 (59)

他八個小時沒有睡覺。 (現代漢語) Tā bā-gè xiǎoshí méiyǒu shuìjiào. he eight-CL hours not sleep “He did not sleep for eight hours.” (Contemporary Chinese)

(60)

他三次沒有來。 (現代漢語) Tā sān-cì méiyǒu lái. he three-times not come He did not come three times. (Contemporary Chinese)

However, the action–resultative principle did not exist in Old Chinese; for instance, temporal words of duration or repetition occurred mostly in preverbal position, as illustrated in (61) and (62): (61)




三歲貫汝。 (詩經 碩鼠) Sān suì guàn rǔ. three year feed you “I have fed you for three years.” (Shi Jing, Shuo Shu, 1000–600 BC)

(62)




季文子三思而後行。 (論語 公冶長) Jì Wénzi sān sī érhòu xíng. Ji Wenzi three think then act “Ji Wenzi thinks three times and then takes action.” (Lun Yu, Gong Ye Chang, 500 BC)

(b) Distribution of preposition phrases In Contemporary Chinese, only preposition phrases that express some kind of result of the matrix verb can occur in postverbal position, and the others must occur in preverbal position. For the same preposition phrase, the interpretation in postverbal position is totally different from that in preverbal position, as illustrated in (63) and (64): (63)

她跳在床上。(現代漢語) Tā tiào zài chuáng-surface. She jump on bed-surface “She jumped upon the bed.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Principle of Action–Resultative Ordering (64)

167

她在床上跳。 (現代漢語) Tā zài chuáng-shàng tiào. she on bed-surface jump “She was jumping on the bed.” (Contemporary Chinese)

The preposition phrase in (63), zài chuáng-shàng “on the bed,” indicates the end point of the jumping, and the action of the jumping happens only once, but the same preposition phrase in (64) refers to the location of the action of jumping, and the action can happen repeatedly. If a preposition phrase refers to the source of an action, it can occur only in preverbal position, such as phrases with the preposition còng “from,” as illustrated in (65). However, the same kind of preposition phrase was usually constrained to postverbal position in Classic Chinese, as illustrated in (66): (65)

他從北京來。(現代漢語) Tā cóng běijīng lái. he from Beijing come “He came from Beijing.” (Contemporary Chinese)

(66)




楚師還自徐。(左傳 昭公二年) Chǔ shī huán zì Xú. Chu troop return from Xu “The troop of the Chu country returned from the Xu country.” (Zuo Zhuan, Zhao Gong Er Nian, 550–400 BC)

The principle of action–resultative ordering started to take effect as early as the first century BC, but it was not firmly established until the twelfth century AD, meaning that the development process took more than a millennium. This principle emerged on analogy with the resultative construction, one of the most significant events in the evolution of Chinese grammar. Both the token and type frequencies of the resultative construction, the semantic structures of which were “action + resultative,” increased greatly over time, having a significant impact on the grammar. As a result, the elements in postverbal position had to be some kind of resultative, and all non-resultative words or phrases disappeared from postverbal position. This historical event caused a fundamental change in the structure of various sentence types, mainly involving passive, locative, comparative, instrumental, and ditransitive constructions (for details, see related chapters). Studies in linguistic typology reveal a language universal: SVO languages are always correlated with SVO PP/Adv, where the PP or adverbial phrase in sentence-final

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

168

position functions to modify the preceding VP in various ways, forming an integral part of many basic sentence structures (for details, see Greenberg 1966a, Li and Thompson 1981, Hawkins 1983, Dryer 1992, 2007). For example, sentence-final position, where various types of preposition phrase in the English grammar appear, is very important, as illustrated in (67): (67)

(a) (b) (c) (d) (e)

He is reading a newspaper in the dining room. The window was broken by a child. He fixed the bicycle with a screw. Mary is even taller than John. He finished the project efficiently.

From the synchronic perspective, Contemporary Chinese dramatically deviates from the language universal depicted above, creating the only exception to some language universals. According to Greenberg (1966b), for example, the comparative structure of an SVO language is “Subj Adjective Marker Standard,” but in Contemporary Chinese it is “Subj Marker Standard Adjective,” which fits no typologically expected pattern (Dryer 2007). However, as the following example shows, the comparative structure in Old Chinese was perfectly consistent with the language universal. The grammar of Old Chinese was greatly consonant with the language universals (for a fuller discussion, see Chapter 11). The deviations are far from random, actually rather highly regular, guided by the principle of action–resultative ordering. In what follows, we sketch the overall changes in the history of these developments; for a detailed explanation, see related chapters.

(a) Locative structure (cf. Section 21.4.1)




(68)

夫人哭於門內。 (左傳 成公五年) Fūrén kū yú mén-nèi. lady cry in door-inside “The lady cried in the doorway.” (Zuo Zhuan, Cheng Gong Wu Nian, 550–400 BC)

(69)

救民於水火之中。(孟子 滕文公) Jiù mín yú shuǐ huǒ zhōng. save people in water fire GEN inside “Save the people from fire and water.” (Meng Zi, Teng Wen Gong, 300 BC)




https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Principle of Action–Resultative Ordering

169

(b) Temporal structure




(70)

賜乎, 舉大事於今之時矣。 (墨子 非儒) Cì hū, jǔ dà shì yú jīn zhī shí yǐ. Ci PRT start big matter in now GEN time PRT “Mr. Ci, it is time to start a big matter.” (Mo Zi, Fei Ru, 350 BC)

(71)

蓄積收藏於秋冬。 (荀子 天論) Xùjī shōucáng yú qiū doing. Accumulate collect in autumn winter “It is to collect and store in autumn and winter.” (Xun Zi, Tian Lun, 250 BC)




(c) Passive structure (for details, see Chapter 8)




(72)

治人者食於人。 (孟子 滕文公) Zhì rén zhě sì yú rén. govern people -er feed by people “Rulers are fed by people.” (Meng Zi, Teng Wen Gong, 300 BC)

(73)

則必見欺於張儀。(史記 楚世家) Zé bì jiàn qī yú Zhāng Yí. and certainly PASS cheat by Zhang Yi “And (you) must be cheated by Zhang Yi.” (Shi Ji, Chu Shi Jia, 100 BC)




(d) Comparative structure (cf. Chapter 11)




(74)

楚弱於晉。 (左傳 襄公十一年) Chǔ ruò yú jìn. Chu weak than Jin “Chu is weaker than Jin.” (Zuo Zhuan, Xiang Gong San Shi Yi Nian, 550–400 BC)

(75)

季氏富於周公。 (論語 先進) Jì Shì fù yú Zhōu Gōng. Ji Shi rich than Zhou Gong “Ji Shi is richer than Zhou Gong.”




(Lun Yu, Xian Jin, 500 BC)

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

Information Structure

170

(e) Instrumental structure (cf. Section 21.4.1) (76)




百工為方以矩。 (墨子 法儀) Bǎi gōng wèi fāng yǐ jǔ. Hundred carpet draw square with ruler “All carpenters draw a square with a ruler.” (Mo Zi, Fa Yi, 350 BC)

(77)




古之君子進人以禮。 (世說新語 方正) Gǔ zhī jūnzǐ jìn rén yǐ lǐ. Ancient ASSO noble-man recommend people with rite “In ancient times, noble men recommended others according to the rites.” (Shi Shuo Xin Yu, Fang Zheng, 450 BC)

(f) Ditransitive structure (for details, see Chapter 12) (78)




報之以瓊瑤。 (詩經 木瓜) Bào zhī yǐ qióngyáo. Repay she with fine-jade “(I) repaid her with fine jade.” (Shi Jing, Mu Gua, 1000–600 BC)

(79)




宋襄公贈之以馬二十乘。 (左傳 僖公二十三年) Sòng Xiāng Gōng zèng zhī yǐ mǎ èr-shí shèng. Song Xiang Gong give he with horse twenty CL “Song Xiang Gong gave him eighty horses (as a gift).” (Zuo Zhuan, Xi Gong Er Shi San Nian, 550–400 BC)

In Old Chinese, the preposition phrases in the passive, comparative, and locative structures were exclusively constrained to postverbal position, but it was equally possible for those in the temporal, instrumental, and ditransitive constructions to occur in either preverbal or postverbal position, which might convey different pragmatic values. When the object of the preposition yǐ (introducing an instrumental noun or direct object) was focused via inversion, for instance, the preposition phrase could appear only in preverbal position (for a fuller discussion, see Chapter 3). Apparently, all the preposition phrases in structures of the above six types did not express any resultative sense in relation to the verb; thus they disappeared from postverbal position under the influence of the principle of action–resultative

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Principle of Action–Resultative Ordering

171

ordering. Superficially, it seems that those preposition phrases in postverbal position were moved to preverbal position; however, in practice, few prepositions that occurred mainly in postverbal position in Old Chinese survived, and they simply died out over time. Their original functions were taken over by a new set of prepositions that grammaticalized in the first verb position of the serial verb construction at a later stage. The speeds of development were different in terms of when the replacement was completed. According to He (2005: 126‒177), preposition phrases referring to locatives and passives were on a large scale constrained to preverbal position in the period from the first century BC to the fifth century AD, whereas a preposition expressing a comparative sense was preserved in postverbal position until the ninth century AD. Therefore the distribution of preposition phrases has undergone a fundamental change in terms of their relation to the matrix verb, whether resultative or nonresultative (accompanying). Resultative PPs remained unchanged throughout history, but those non-resultative PPs were gradually constrained to preverbal position, which can be formalized as follows: (80)

(a) V + Obj + PPresultative > V + Obj + PPresultative (b) V + Obj + PPnon-resultative > PPnon-resultative + V + Obj

The above formulas can be used to explain many significant changes in the evolution of Chinese grammar. This point is briefly illustrated by the following cases; for an in-depth discussion, please see related chapters. For the passive structure, Old Chinese had two passive auxiliaries – jiàn (from the verb of seeing) and bèi (from the verb of suffering) – that preceded the verb and could not introduce the agent; if the agent needed to be indicated, a preposition phrase with yú had to be used in postverbal position. Since this preposition phrase in the passive structure did not express the resultative, it was eliminated from postverbal position. Quite surprisingly, the passive auxiliary bèi developed into a preposition to introduce an agent in preverbal position around the fifth century AD (Wang 1989: 272‒287), but the auxiliary jiàn had no such development and hence was abandoned. Clearly, the biological evolutionary principle “survival of the fittest” was applicable to language. Since then, at least five passive markers have been introduced into the language, and all of them can introduce an agent in preverbal position (for a fuller discussion, see Chapter 8). The distributional change in the preposition phrase had a profound effect on the construction of nominal phrases involving relative clauses. Once again,

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

172

Information Structure

Contemporary Chinese is the only counterexample for the language universal in a sample of 625 languages: all SVO languages adopt the constituent order “NP + REL,” but the structure in Chinese is “REL + NP” (Dryer 1992, 2007). Historically, however, the canonical structure in Chinese was also “NP + REL,” and if this diachronic evidence is taken into consideration, the language universal is without exception, the only one of such types among the forty-four Greenbergian language universals (Greenberg 1966a). In our view, the change in the relative clause was caused by the distribution of preposition phrases. There is a semantic parallelism between the preposition phrase and the relative clause: the preposition phrase is the modifier of VP, and the relative clause is the modifier of NP. There was morphological evidence in Chinese. The associative particle de, which grammaticalized from a demonstrative around the eighth century AD, can link either a relative clause to an NP or an adverbial phrase to a VP, following the same rule. Therefore we propose that the position shift of non-resultative PPs might have affected the redistribution of the relative clause via analogy (for a fuller discussion, see Sections 7.9, 7.20, 7.21). The principle of action–resultative ordering also determined the fates of two competing grammatical morphemes. Around the sixth century AD, the verb jiāng “take” developed into a disposal morpheme, and approximately 200 years later a synonymous verb bǎ acquired the same function. After more than a millennium of competition, bǎ finally won over jiāng and became the only marker for the disposal construction. One major factor in the result was that the preposition jiāng could introduce a patient noun at either preverbal or postverbal position, which obviously violated the principle of the predicate information structure because the preposition phrase was non-resultative. In contrast, the preposition bǎ could introduce the patient only in preverbal position, which obeyed the principle; hence it has survived to the present day. In other words, the bǎ disposal construction is consonant with the principle and hence became the winner (for a fuller discussion, see Chapter 9). Furthermore, the establishment of the principle of action–resultative ordering enabled classifiers to develop into demonstratives. Numeral classifiers as a grammatical category started to emerge in Late Old Chinese and developed steadily in Medieval Chinese. By the end of the Medieval Chinese period, they were widely used nominal phrases. When the number was “one,” it was often omitted, forming a “CL + N” phrase. Under the influence of the principle, this phrase was made definite when it occurred in preverbal position. As a result, the classifier zhī, one of the most general classifiers, grammaticalized as a proximal

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

The Principle of Action–Resultative Ordering

173

demonstrative in the northern dialects of Chinese (including Mandarin Chinese), and another general classifier gè developed into a proximal demonstrative in many southern dialects. Their grammaticalization involved the semantic features of the lexical sources and the default meaning of preverbal position. It is assumed that demonstratives are primitive words of a language, as they needed to be there at the very beginning when the language was invented (Greenberg 1978, Diessel 1999a: 1‒12). In other words, in theory, demonstratives cannot be grammaticalized from other lexical sources, though they are functional words, and this is a challenge for the hypothesis of grammaticalization theory that every functional morpheme grammaticalized from lexical items with concrete meanings (for a fuller discussion, see Chapter 17). Additionally, the ditransitive construction and resultative construction appear to be disparate. In Chinese, however, due to the extremely powerful analogy of the resultative construction, the ditransitive construction is treated as a subtype of the resultative construction because the semantic relation between the verb and the indirect object holds a semantic structure, “action + end point,” that resembles that of VR phrases. Specifically, the verb and the indirect object first form an immediate constituent that governs the direct object. Therefore, in Contemporary Chinese, if the direct object is indefinite, it follows the verb and the indirect object; if it is definite, it must be introduced in preverbal position by the disposal marker bǎ. One type of the earliest disposal instances belonged to the ditransitive construction (Wang 1989: 266‒271, Ohta 1987: 241). As Croft (2001: 32) pointed out, every construction is languageparticular because it is subject to the influence of the related constructions in the language. Some VO constructions were also turned into a kind of VR phrase that can assign an extra patient argument. One type of VO phrase must affect another entity in addition to the object; for instance, bāo-pí “peel-skin” usually involves a whole fruit, such as an orange. Since a VO phrase disallows any following object, to mention the orange, which is licensed by the whole phrase, it must be done by some other means, such as a passive construction, a disposal construction, or topicalization. This phenomenon has become increasingly common since the tenth century AD (for details, see Lü 1984: 200–208). Since the resultative construction took shape in Medieval Chinese, its token and type frequencies increased greatly, and in the present day this construction is ubiquitous. The predicates of many other constructions tend to use a “verb + resultative” phrase, such as the passive and disposal constructions. Some constructions even require

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

174

Information Structure

a “verb + resultative” predicate to be well formed, such as the verb-copying construction (for a fuller discussion, see Section 10.1). Such extensive use of the resultative construction has extreme power to influence other seemingly unrelated constructions by analogy. As a result, the ditransitive construction and even some VO phrases are assigned the syntactic features of the resultative construction.

https://doi.org/10.1017/9781108921831.007 Published online by Cambridge University Press

8 The Passive Construction

8.1

Introduction

The structures and functions of Chinese passives underwent three typologically significant changes: (a) the passive morphemes in Old Chinese could not introduce any agent phrase in preverbal position, but those in Contemporary Chinese must introduce an agent noun to make the passive structure well formed; (b) the agent phrase could be introduced only in postverbal position in Old Chinese but is exclusively constrained to preverbal position in Contemporary Chinese; and (c) the passive morphemes are extremely diverse in Contemporary Chinese, including four markers in Mandarin Chinese and at least sixty-nine markers in the other dialects that have grammaticalized from different lexical sources but have quite uniform structures and functions, where the presence of an agent is generally obligatory. This chapter addresses the motivations and mechanisms behind these changes. We will show that passives and other major active structures actually share the same syntactic schema and that the overall historical morphosyntactic evolution is responsible for determining the functions, structures, life spans, and grammaticalization paths of the passives. Studies of passive structures have played a crucial role in developing linguistic theories since Chomsky (1957), and the present analysis may therefore shed light on current linguistic argumentation. There have been numerous passive morphemes in the history of the Chinese language, and they could combine in various ways to form numerous passive structures (Wang 1989: 272‒288, Yang and He 2001: 667‒697). They differ from one another in their frequencies, constructions, pragmatic functions, and syntactic behaviors. For instance, some passive morphemes, such as jiàn and bèi in Old Chinese, could not introduce an agent phrase; some passive morphemes, such as the wéi passive and the bèi passive in Medieval Chinese, could optionally introduce an agent phrase; and some passive morphemes, such as the yú passive in Old Chinese and both the jiào and ràng passives in Contemporary Chinese, must introduce an agent phrase to make the passive structure well formed. Moreover, in Old Chinese, the earliest passive morpheme, yú, could introduce an agent phrase only in postverbal position, but all the later passive morphemes

175

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

176

could do so only in preverbal position. Nevertheless, all these passive morphemes and structures in Chinese display strong commonalities with passives in other languages. It is cross-linguistically rare, if not an isolated phenomenon, that the passives in Contemporary Chinese, either in Mandarin or in other dialects, require an agent phrase to be well formed. Lü (1999: 463), for instance, pointed out that the agent of the ràng passive must be overtly expressed, whereas those of the other three passive patterns need not be. Moreover, Liu et al. (2001: 760) claimed that the agentive nouns of the jiào and ràng passives are necessary to make the clauses well formed. The two basic passive morphemes, namely jiào and ràng, are illustrative: (1)

(a) 活兒都讓他們幹完了。(現代漢語八百詞) Huó-er dōu ràng tā-men gàn-wán le. task-DIM all PASS they do-complete PERF “The entire task has been finished by them.” (b) *活兒都讓幹完了。 *Huó-er dōu ràng [ ] gàn-wán le. task-er all PASS do-complete PERF (Contemporary Chinese, Lü 1999: 462)

(2)

(a) 那張地圖叫人借走了。 (現代漢語) Nà-zhāng dìtú jiào rén jiè-zǒu le. that-CL map PASS someone borrow-away PERF “The map has been borrowed by somebody.” (b) *那張地圖叫人借走了。 *Nà-zhāng dìtú jiào that-CL map PASS

[]

jiè-zǒu le. borrow-away PERF (Contemporary Chinese, Lü 1999: 304)

Note that in (2a), even the agent is nonspecific and indefinite, namely “someone,” but must be spelled out, whereas the corresponding by-phrase in English is often omitted. The function of this unspecific agent in the Chinese passive is similar to the expletive subject it or there in finite clauses in English, whose obligatory appearance is determined by the well-formed grammatical principle rather than by the motivation to provide information. Moreover, Cao (2008) presented the outcome of a project over ten years in duration in which hundreds of scholars participated in investigating 930 dialects in China. The results revealed that the bèi passive morpheme, which has a history of approximately 2,300 years, has ultimately become obsolete in the spoken languages and that there are sity-nine distinct passive markers in total that have been grammaticalized from various

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

Optionality of the Agent in Passives

177

lexical sources, most of which make it obligatory to express agentive nouns. That is, the obligatory occurrence of agentive nouns in the passive is the commonality among the various passive patterns in Mandarin Chinese and other Chinese dialects. This chapter addresses the motivation and mechanism for the emergence of the Chinese passive with an obligatory agent, an issue of empirical and theoretical importance.

8.2

Optionality of the Agent in Passives

In general, the by-phrase in the English passive is considered marginal or nonessential in very different theoretical frameworks. In the model of Chomsky (1957: 61‒84), the three constituents of the passive in English, i.e. the copula “be,” the past participle, and the by-phrase, are treated as meaningless elements within the transformation process from deep structure to surface form. Here, the insertion of a by-phrase is optional since the agent is “marginal” but the object is “nonmarginal” (Lakoff 1971, Bouton 1973). Likewise, as for the passive, Langacker (1982, 1991, 2013) claimed that every element constituting the passive is meaningful and is responsible for the overall properties of the passive. Specifically, be refers to a processual value unique to the passive, whereas the past participle indicates a stative status resulting from the action involved. Note that Langacker (2013: 117‒120) regarded the by-phrase not as an essential constituent of the passive but as functionally similar to its other prepositional usages, such as expressing the location or instrument, and hence considered it optional. Following Shibatani (1985), where the agent in the passive is regarded as being “downgraded” or “defocused,” Langacker (2008: 299‒231) treated the agent phrase in a passive as a “landmark” or “background” but treated the patient as a “trajector” or “foreground.” According to the findings of linguistic typology, the type of passive that has an obligatory agent should not exist in any language. Keenan (1985b), in an extensive survey of passives in the world, claimed that some languages have no passive construction and that for those with only one type of passive construction, it must be the “basic passive,” as exemplified in “John was slapped,” where the agent is absent. He drew the following generalization: if a language has passives with agent phrases, then it also has passives without agent phrases. This generalization implies that there exist languages whose passives disallow an agent phrase. The following is an explicit statement in linguistic typology and language universals (Kiparsky 2013): if a language has passives with agent phrases, such phrases are optional. The above principle has repeatedly been challenged. For instance, Siewierska (1984: 35) claimed that what Chung (1976) called the Indonesian object-preposing construction is a passive that requires an agent phrase in order to be well formed. However, Kiparksy (2013) dispensed with this so-called counterexample by claiming that it arises from a misanalysis and actually incorporates

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

178

a pronoun. Thus the Chinese passive with an obligatory agent phrase apparently remains the only counterexample to the empirical research on language. In contrast, according to Keenan (1985b), many languages, such as Latvian and Kutenai, are cited as permitting only agentless passives. In other languages, agentless passives are preferred even when agentive phrases are grammatically possible. For instance, native speakers of Turkish are reluctant to accept passives with agent NPs, which is probably a phenomenon induced by language contact with English, Dutch, and other languages (Kirsner 1976).

8.3

The Semantics and Syntax of Passives

Unlike in many other languages, such as English, the verbs of Chinese passives have no inflections for the passive, and throughout history the base form of a verb could be used in both active and passive sentences. Regarding the English passive construction “be + Past Participle + by NP,” Langacker (2013: 118‒123) argued that the copula be and the past participle are the essential parts of the English passive but that the by-phrase is marginal, appearing only when information about the agent has to be expressed. From Langacker’s point of view, the Chinese passive possesses no obligatory morphological markers for its passives. As a result, unmarked passive instances are pervasive in Chinese, both diachronically and synchronically, and are generally labeled “semantic passive sentences” in the literature (Wang 1989: 272), as illustrated in the following examples:




(3)

昔者龍逢斬, 比干剖。 (莊子 胠箧) Xīzhě Lóng Féng zhǎn, Bǐgàn pōu. before Long Feng decapitate Bigan cut-open “In the past, Mr. Longfeng was decapitated and Mr. Bigan(’s belly) was cut open.” (Zhuang Zi, Qu Qie, 300 BC)

(4)

作業已經做了。 (現代漢語) Zuòyè yǐjīng zuò-le. homework already do PERF “The homework has already been done.” (Contemporary Chinese)

In the two examples above, both subjects are the patients affected by the verbs, and no object or other nominal phrases occur in postverbal position. These instances have to be expressed by means of the passive morphology in English but are morphosyntactically unmarked in Chinese. In Chinese, however, the construction “SUBJ + V + Obj,” the

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Agent Noun in the Passive Structure

179

typical schema of active sentences, can produce either a passive or an active interpretation, as illustrated in (5): (5)

(a) 約翰在吹電扇。 (現代漢語) Yuēhàn zài chuī diànshàn. John PROG blow electric-fan “John is being cooled down by an electric fan.” (Contemporary Chinese) (b) 約翰在吹蠟燭。 (現代漢語) Yuēhàn zài chuī làzhú. John PROG blow candle “John is blowing the candles.” (Contemporary Chinese)

Note that (5a) renders a passive reading, but (5b) is an active interpretation. Apparently, both share exactly the same syntactic structure, and in both the verbs take the same base forms. Given that the Chinese passive structure has no compulsory morphosyntactic markers or forms, one may wonder when passive morphemes and constructions in Chinese are used. Some pragmatic factors may play a role in applying passive morphemes, such as emphatic passivation or ambiguity avoidance (Wang 1989: 272‒273). Unlike their counterparts in English and other languages, Chinese passives often convey an adversarial sense, namely “negative affectedness” or “suffering” (Li and Thompson 1981: 492‒508, Wang 1989: 284). This is clear because the passive preposition yú in Old Chinese was used only when an agent had to be indicated. In other words, the passive morpheme yú was absent in contexts in which the agent did not need to be mentioned. As a result, the frequency of passive instances in Chinese was much lower than that in English.

8.4

The Agent Noun in the Passive Structure

In the distinction between short and long bèi passives, it is very important to consider that the preposition phrase introducing the agent phrase in Chinese passives underwent a significant change from Old Chinese to Medieval Chinese. The earliest passive yú, a preposition grammaticalized from its original locative usages, like the English by and Japanese ni, was even attested in Early Old Chinese texts, always occurring in postverbal position to express the agent. Wang (1989: 273) claimed that in this usage, yú was actually a locative preposition rather than a passive marker, akin to English by. The typical passive morpheme jiàn in Old Chinese, which grammaticalized from the verb

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

180

“see,” behaved like an auxiliary verb that immediately preceded the predicate verb to indicate passivity but was always incapable of introducing an agent phrase. The complete passive pattern in Old Chinese, namely the “passive morpheme + PPagent” construction, is illustrated in the following example: (6)




見疑於齊君。 (呂氏春秋 季冬紀) Jiàn yí yú Qí Jūn. PASS suspect by Qin King “(I) am suspected by the king of the Qi country.” (Lü Shi Chun Qiu, Ji Dong Ji, 250 BC)

One overall change in Chinese grammar was responsible for triggering the development of the passive patterns. In general, the syntactic distribution of preposition phrases underwent a fundamental change from postverbal position to preverbal position, a change that started as early as the Late Old Chinese period and was completed in the Medieval Chinese period. This position shift of preposition phrases was motivated by the emergence of the resultative construction, which required by analogy that all nonresultative PPs appear in preverbal position. Thus preposition phrases in passive, comparative, instrumental, and locative structures gradually disappeared from postverbal position, and the functional morphemes that grammaticalized from the first verb of a verbal serialization construction later served to mark these types of structure. These facts account for the systematic changes in the basic syntactic structures, including passive structures and many other active structures. The basic word order of the Chinese language throughout history was always SVO, although there were some variations that may be called SOV (Li and Thompson 1974, Sun and Givόn 1985). In accordance with the language-universal correlations of this type of word order (Greenberg 1966a, Comrie 1981: 92, Croft 1990: 49‒62), PPs in Old Chinese overwhelmingly appeared after the VO, namely in sentence-final position. The configuration of major types of declarative sentence, such as passive sentences as illustrated above, locative sentences, comparative sentences, and instrumental sentences, were “Subj + V + Obj + PP,” exactly like those in English, but their counterparts in Contemporary Chinese have become “Subj + PP + V + Obj,” observing the information-structuring principle (for details, see Chapter 6). Note that there were no movements involved, though the agent phrase occurred first in postverbal position and later in preverbal position. The truth is that a group of new prepositions grammaticalized from the first verbs in the verbal serialization structure. Specifically, the locative yú was replaced by zài, the instrumental yǐ by yòng, the comparative yú by bǐ, and the passive yú by bèi. However, these changes involved no lexical replacements but were independent developments. Later, PPs, denoting location, instrument, agent, and comparison

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Agent Noun in the Passive Structure

181

standard, had to precede predicate verbs and only a few PPs that express the resultative state of the action still remained in postverbal position (Chao 1968: 8.2), as illustrated in (7): (7)

約翰寄了一個包裹給瑪麗。 (現代漢語) Yuēhàn jì-le yī-gè bāoguǒ gěi Mǎlì. John mail-PERF one-CL parcel to Mary “John mailed one parcel to Mary.” (Contemporary Chinese)

In example (7), “Mary” is the destination (end point) of the mail action, a kind of resultative state. Thus the structure in (7) is still grammatical in Contemporary Chinese. The syntactic position changes of preposition phrases described above significantly influenced the development of the passive morphemes in Old Chinese. First, since the passive preposition yú was always unable to introduce an agent noun in preverbal position, it gradually became obsolete in the Medieval Chinese period.1 Second, the auxiliary verb jiàn never acquired the function of introducing an agent phrase in preverbal position and therefore was abandoned during the Medieval Chinese period. Now, let us focus on the developments of the passive morpheme bèi and its associated constructions, as it has the longest history among all the passive markers, from the second century BC to the present day. Undoubtedly, the verb bèi became grammaticalized into a passive marker from its “suffer” verbal usage no later than the second century BC (Wang 1989: 279). Its developments fall into the following major stages. Stage one. When bèi became a passive morpheme in Late Old Chinese, the grammar of the language at that time allowed the preposition yú to introduce an agent phrase only in postverbal position. The passive morpheme bèi then behaved like an auxiliary verb, functionally identical to the earlier passive morpheme jiàn. Both bèi and jiàn immediately preceded the predicate verb and were unable to indicate an agent phrase. In that period, when an agent phrase had to be expressed, the preposition yú had to be used in postverbal position, producing the complete passive pattern “bèi + V + yú + NP,” exactly the same as “jiàn + V + yú + NP,” as exemplified in (6). There was no temporal difference between the emergence of the bèi passive and that of the bèi . . . yú passive; they happened at roughly the same time. The “bèi + V + yú + NP” passive pattern survived through the Early Medieval Chinese period.

1

When we say that some morphological forms disappeared or became outdated or obsolete, we mean that they were no longer being used in the spoken language after the movements occurred. These forms may, however, have survived in written texts or formal speech over a long period.

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

182 (8)

(9)




知友被辱。 (韓非子 五蠹) Zhi-yǒu bèi rǔ. Best-friend PASS insult “(Their) best friend was insulted.”

(Han Fei Zi, Wu Du, 300 BC)




萬乘之國被圍於趙。 (戰國策 齊策) Wàn shèng zhī guó bèi wéi yú Zhào. Ten-thousand chariots ZHI country PASS surround YU Zhao “A country with 10,000 war chariots was surrounded by the Zhao country.” (Zhan guo ce, Qi Ce, 450–200 BC)

Stage two. As the grammatical system gradually evolved, the agent phrases of passive instances needed to occur in preverbal position after the Late Old Chinese period. During the Early Medieval Chinese period, in particular, the dominant passive patterns were the wéi form and its formal variation, e.g. the wéi . . . suǒ form, where wéi could introduce an agent phrase in preverbal position. Under these circumstances, the auxiliary morpheme bèi developed into a preposition that could function to introduce an agent phrase in preverbal position, a new usage that started no later than the fourth century AD and continues today. One of the earliest examples of the bèi passive with an agent phrase is illustrated in (10):2 (10)

果被眾人怪笑。 (百喻經) Guǒ bèi zhòngrén guài xiào. really PASS everyone sardonically laugh “Surely, he was sardonically laughed at by everyone.” (Bai Yu Jing, AD 500)

The so-called long bèi passive and short bèi passive did not actually exist in history. Additionally, it is cross-linguistically true that the expression of the agent in passives is typically optional rather than obligatory in various languages.3 The Chinese passives were no exception in the period from Old Chinese to Medieval Chinese. The real difference between the earlier bèi passive and the later bèi passive is the means they used to express information about the agent. In Old Chinese, the preposition yú served to introduce the agent in postverbal position, and bèi itself could not introduce the agent until the fourth century AD. This functional change in the passive morpheme bèi resulted from the general development of the Chinese grammatical system. From our 2 3

Example (10) is cited from Wang (1989: 281). In Section 9.5, we deal with types of passive structure in Contemporary Chinese that need agent phrases in order to be well formed.

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

From Optional Agent to Obligatory Agent

183

perspective, any plausible explanation for the evolution of passive patterns must be compatible with the global change in the grammatical system.

8.5

From Optional Agent to Obligatory Agent

Concerning passive morphemes in preverbal position, no agent nouns were used in passives in Old Chinese, whereas they were optional in Medieval Chinese and necessary from Modern Chinese to the present. The above section discussed the situation in Old Chinese, and this section sketches the development from Medieval Chinese to Contemporary Chinese.

8.5.1

The Wéi Passive and the Wéi . . . Suǒ Passive

There are two perspectives from which to consider the passive pattern marked by wéi: first, whether the passive morpheme wéi appeared alone or was paired with suǒ; second, whether an agent phrase is present or absent. However, the empirical data show that both the wéi pattern and the wéi . . . suǒ passive belong to the same passive structure with some formal variations. In Early Old Chinese, wéi, a verb meaning “do,” became a multifunctional word. Additionally, it grammaticalized into a preposition with multiple functions; for instance, it could introduce a reason, a benefactive, or a purpose. Additionally, wéi could serve as a passive morpheme in certain contexts. During this period, the morpheme wéi could be used alone only to mark the passive structure, and the earliest use of the passive pair wéi . . . suǒ was attested in texts written around the second century BC. In the first two centuries, when wéi became a passive marker, it had three structural variations: (a) it marked an agentless passive, where wéi immediately preceded a verb; (b) it was paired with the preposition yú, where yú always functioned to introduce an agent phrase in postverbal position, similar to the jiàn and bèi passive patterns mentioned above; and (c) it functioned like a preposition to introduce an agent phrase in preverbal position. These three types of passive wéi instances are illustrated as follows:4 (11)

4




宜其為擒也。 (左傳 宣公元年) Yí qí wéi qín yě. sure he PASS capture PRT “He was surely captured.” (Zuo Zhuan, Xuang Gong Yuan Nian, 550–400 BC)

The following three examples are cited from Tang and Zhou (1985).

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

184 (12)




身死为僇於天下。(墨子 法儀) Shēn sǐ wéi lù yú tiānxià. body die PASS kill YU world “After death, he was killed by the world.” (Mo Zi, Fa Yi, 350 BC)

(13)




吾恐其為天下笑。 (莊子 徐無鬼) Wú kǒng qí wéi tiānxià xiào. I afraid he PASS world laugh-at “I am afraid that he would be laughed at by the world.” (Zhuang Zi, Xu Wu Gui, 300 BC)

Wang (1989: 278) suggested that instances of the wéi . . . suǒ passive pair had already appeared in texts of the first century AD in the Early Medieval Chinese period. However, Tang and Zhou (1985) noted that the wéi . . . suǒ pattern was already attested in texts from the second century BC, part of the Late Old Chinese period. (14)




夫直議者不為人所容。 (韓非子 外儲說) Fū zhíyì-zhě bù wéi rén suǒ róng. PART criticize-er not PASS people SUO accept “The one who speaks frankly is not accepted by others.” (Han Fei Zi, Wai Chu Shuo, 300 BC)

The wéi passive and the wéi . . . suǒ passive were used in the same period. Although the frequency of the wéi . . . suǒ passive was remarkably higher than that of the wéi passive during the Early Medieval Chinese period, the former never completely replaced the latter. Table 8.1 is based mainly on data from the vast texts of Tang and Zhou (1985) and Tang (1987). Considering that their statistical work was carried out manually in the 1980s and might contain errors, I have double-checked the figures via electronic corpuses in order to guarantee the accuracy of the figures. Note that the percentages listed in Table 8.1 reflect only the proportion of instances of the two passive patterns: the wéi passive versus the wéi . . . suǒ passive. In every period, passive patterns existed that might be marked by jiàn, yú, bèi, and others, as illustrated in previous sections. These two passive patterns were largely replaced by the bèi passive after the sixth century AD. According to Tang (1987), of the total of 3,435 passive instances, instances of the bèi passive account for 509 tokens from the third century AD to the sixth; thus it was one of several major passive patterns during that time. More importantly, the wéi . . . suǒ construction was never purely a long passive. According to the data from Tang (1987), a total of seventy-seven instances of the wéi . . .

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

From Optional Agent to Obligatory Agent

185

Table 8.1 The wéi passive and the wéi . . . suǒ passive

Periods

Tokens (percentage) of the wéi passive

Tokens (percentage) of the wéi . . . suǒ passive

Fifth–third centuries BC Second–first centuries BC First–second centuries AD Third–sixth centuries AD

101 (100%) 49 (89%) 123 (25%) 110 (6%)

0 (0%) 4 (11%) 379 (75%) 1821 (94)

suǒ passive are agentless. These agentless instances are attested in sixteen books that were written from the first century BC through the sixth century AD. Although the overwhelming majority of the wéi . . . suǒ passive instances contained an agent phrase, this passive pattern was by nature not a long passive. The following examples illustrate the usage of the agentless wéi . . . suǒ passive:5




(15)

射猛獸, 亦為所傷云。 (史記 李將軍列傳) Shè měngshòu, yì wéi-suǒ shāng yún. shoot beast also WEI-SUO wound PART “(General Li) shot a beast and was wounded (by the beast).” (Shi Ji, Li Jiang Jun Lie Zhuan, 100 BC)

(16)

授不降也, 为所執耳。 (後漢書 袁紹劉表列傳) Shòu bù xiáng, wéi-suǒ zhí ěr. Shou not surrender WEI-SUO capture PART “Shou didn’t surrender and was captured.” (Hou Han Shu, Yuan Shao Liu Biao Lie Zhuan, AD 450)




Although wéi and suǒ occur adjacently in (15) and (16), they are not a compound passive marker, because there is a syntactic position between them that can be filled with an agent noun. When the agent is identifiable in context, the agent phrase can be omitted; for instance, the omitted agent in (15) is the měngshòu “beast,” which was mentioned in the preceding clause. The flourishing of the wéi passive in the period from the first century BC through the sixth century AD was motivated by the global change in the grammatical system in that particular period. As mentioned previously, all preposition phrases that did not express resultatives of predicate verbs could no longer appear in postverbal position. Therefore the wéi passive was favored by the grammatical system at that time, since it was able to introduce an agent phrase in preverbal position. The preposition wéi had multiple functions, as pointed out earlier, which could introduce various nominal phrases such 5

The following two examples are cited from Tang (1987).

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

186

as reason, benefactive, and purpose, and the related preposition phrases always appeared in preverbal position (He 2004: 370‒383). When wéi became one of the major passive morphemes, it might have easily caused ambiguity because of its polysemy and multiple functions. Thus the addition of suǒ served to eliminate possible ambiguity, and all instances of the wéi . . . suǒ pair could be interpreted only as passive sentences. In Section 8.6, I will explain why the morpheme suǒ was selected to pair with the passive wéi in passive instances. The wéi passive and the wéi . . . suǒ passive originated from the same source, representing two structural variants with the same function. Neither of them is purely the long passive in any particular period because agent phrases could always be absent. The flourishing and dominance of the passive patterns with wéi in the Early Medieval Chinese period were promoted by the global change in the grammar.

8.5.2

The Bèi Passive

Simply speaking, the original meaning of bèi was “blanket,” which was extended to express “cover” or “suffer” in later times. The passive morpheme bèi grammaticalized from its “suffer” verbal usage through a reanalysis of “bèi + NP” as “bèi + VP” (Wang 1989: 79‒81). We saw the passive instances in the previous section, and the following example is its “bèi + NP” usage:6 (17)




秦王復擊荊軻, 被八創。 (史記 刺客列傳) Qín Wáng fù jī Jīng Kē, bèi bā chuāng. Qin King again stab Jing Ke suffer eight wound “King Qin stabbed Jing Ke again, and (Jing Ke) suffered eight wounds.” (Shi Ji, Ci Ke Lie Zhuan, 100 BC)

The seminal works of Tang and Zhou (1985) and of Tang (1987, 1988) provided a comprehensive description of every stage of the diachronic development of the passive bèi. At almost the same time, roughly around the second century BC, both the wéi . . . suǒ passive and the short bèi passive emerged in the language. In the first seven centuries or so, the passive morpheme bèi was unable to introduce an agent phrase, and if the agent phrase had to be expressed, the passive preposition yú was employed to do so in postverbal position, as illustrated in the previous section. According to Wang (1989: 281), no later than the fourth century AD the passive morpheme bèi started to be able to introduce an agent phrase by itself in preverbal position. The overall development of the methods of introducing an agent is sketched in (18): 6

Example (30) is cited from Wang (1989: 283).

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

From Optional Agent to Obligatory Agent (18)

187

(a) The second century BC~the third century AD Short passive: bèi + V; long passive: bèi + V + ( yú + NP) (b) The fourth century AD onward Short passive: bèi + V; long passive: bèi + NP + V

The structural changes described above show that the agent phrase in the Chinese passive was always optional in these periods, and the only difference between the different passive patterns was how to introduce the agent phrase and where to place it. Obviously, the loss of the passive preposition yú from postverbal position might be responsible for triggering further grammaticalization of the passive bèi, which made it able to introduce an agent phrase. As a result, bèi became a preposition-like element, a change motivated by developments in the overall property of the grammar at that time. In contrast, the oldest passive morpheme, jiàn, did not acquire the preposition function of bèi and thus became outdated in Medieval Chinese. As mentioned previously, the functional changes in the passive bèi were triggered by the overall development of the grammatical system in that particular period. When bèi became grammaticalized in the Late Old Chinese period, there was a preposition yú whose special function was to introduce an agent phrase in postverbal position. Thus both jiàn and wéi, the only two passive morphemes strictly constrained to preverbal position in Old Chinese, behaved like passive auxiliary verbs that could not introduce an agent in preverbal position. Due to the change in the grammatical system, the preposition yú gradually went out of fashion because the newly established grammatical system no longer allowed any prepositions to introduce an agent in postverbal position. It was in this situation that the auxiliary passive bèi gave birth to a new function that introduced an agent in preverbal position. In fact, all the passive morphemes that were produced after the Medieval Chinese period could introduce an agent phrase only in preverbal position. Therefore we disagree with the viewpoint of some researchers (Wang 1989: 281) that the bèi passive became “more mature” when it was able to introduce an agent phrase. The grammaticalization process of any grammatical marker, which has no expected course of growth, is unlike the growth of organic things, such as humans or plants. The life spans, functions, and meanings of grammatical markers and constructions, we believe, are to a large extent influenced by the overall property of the grammatical system in a particular period. There were obvious motivations behind the changes in passive patterns, particularly in how to introduce an agent phrase. What motivated the passive morpheme bèi to introduce an agent phrase was that the wéi . . . suǒ pattern gradually went out of fashion; hence, to satisfy a pragmatic need, the language needed to find another way to express agent phrases. It was no coincidence that the disappearance

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

188

Table 8.2 The increase of the bèi passive with agent phrases over time Text

Date and size

With agents

Without agents

Shi Ji Shi Shuo Xin Yu Bai Yu Jing Dun Huang Bian Wen Zu Tang Ji Zhu Zi Yu Lei

First century BC, 620,000 characters Fifth century AD, 79,000 characters Fifth century AD, 22,000 characters Eighth century AD, 91,000 characters Tenth century AD, 222,000 characters Twelfth century AD, 942,000 characters

0% (0) 22% (2) 33% (2) 49% (46) 78% (59) 87% (399)

100% (21) 78% (7) 67% (4) 51% (48) 22% (17) 13% (58)

of the wéi . . . suǒ passive and the emergence of the new bèi pattern happened at the same time. The percentage changes in the long bèi passive instances over time reveal the strong tendency of an increasing number of passive instances to occur with an agent phrase. Meanwhile, the bèi passive took over the function and status of the wéi . . . suǒ pattern to become the dominant passive pattern during the long period from the sixth century AD to the seventh. The auxiliary passive bèi further grammaticalized into a preposition to introduce an agent noun in preverbal position. This change was motivated by overall developments of the grammatical system. If this had not happened, the passive bèi would inevitably have been abandoned, as was the fate of the passive jiàn.

8.5.3

The Chī Passive

During the period from the eleventh century AD to the sixteenth, the verb chī “eat” underwent a grammaticalization path quite similar to that of the passive morpheme bèi. That is, it is a perfect example to demonstrate that an independent grammaticalization process of this type can produce two outcomes, the long passive and the short passive, indicating that the agent in this passive pattern is optional. We should immediately note that long passive usages do not need to inherit the function of any other former passive structures. There was a semantic extension of the grammaticalization of the verb chī “eat.” This verb chī was metaphorically extended to mean “suffer”; e.g. chī jiàn (lit. eat + arrow) meant “hit by an arrow,” and chī guānsī (lit. eat + lawsuit) meant “sued by somebody else.” In the context where chī was followed by a verb rather than a noun, it was reanalyzed as a passive morpheme (for details, see Jiang 1989). The chī passives account for approximately 17 percent of all passive instances in the novel Shui Hu Zhuan, a vernacular text written around the fifteenth century AD, whereas bèi passive instances account for 80 percent. Unlike the passive morpheme bèi, which could not

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

From Optional Agent to Obligatory Agent

189

denote an agent by itself in the first seven centuries during which it gained passive usage, as mentioned previously, chī could introduce an agent phrase from the very beginning of its grammaticalization, as illustrated in (19): (19)

(a) 後來便吃殺了。 (水滸傳四十六回) Hòulái biàn chī sha le. late then PASS kill PERF “Then he was killed.” (Shui Hu Zhuan, Chapter 46, AD 1400) (b) 我吃這婆子盯住了。 (水滸傳二十一回) Wǒ chī zhè pó-zi ding-zhù le. I PASS this lady tail PERF “I was tailed by this lady.” (Shui Hu Zhuan, Chapter 21, AD 1400)

In terms of generative linguistics, (19a) is the short chī passive, and (19b) is the long chī passive. Although bèi and chī shared similar grammaticalization paths, namely from “suffer” to passive morphemes, their grammatical functions differed dramatically with regard to co-occurrence with agent phrases. Among the 128 passive instances that I collected from six texts of Early and Middle Modern Chinese (tenth century AD– sixteenth), 107 have agent phrases marked by chī, accounting for roughly 80 percent of all passive instances. The percentages of chī passives with and without an agent phrase are almost the same as the percentages of bèi passives with and without an agent phrase in this Modern Chinese period, according to the data presented by Tang (1988). This coincidence reveals that the hidden factors determining the syntax of the passive morphemes are the overall properties of the grammar in that particular period. In this period, Chinese grammar favored the passive structure that introduced an agent phrase in preverbal position by means of its passive morpheme. As Jiang (1989) pointed out, the sudden disappearance of the chī passive around the sixteenth century AD was quite mysterious. Additionally, the frequency of bèi passive instances was greatly reduced, and it has gradually become restricted to formal speech and written language since then. Meanwhile, three new passive morphemes, jiào, ràng, and gěi, had become grammaticalized, and all of them require an agent phrase in order to make passive clauses well formed. In other words, an agent phrase is compulsory in these new passive structures, representing a typologically rare type of passive. The chī case demonstrates that the “suffer” verbs, including bèi, could naturally develop out of two passive variants, namely the long passive and the short passive, without having to inherit the functions of any other former passive patterns, as Li (2018) claimed. The long chī passive pattern could not inherit any other passive structures

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

190

because the wéi . . . suǒ passive pattern had disappeared 700 years earlier and the coexisting bèi passive was always the major passive pattern through the period of the rise and decline of the chī passive. The cause for the decline of the chī passive was the new trend in the evolution of Chinese grammar.

8.5.4

Passive Morphemes in Standard Mandarin

The usage of the passive morpheme bèi has become restricted to written texts and formal speech in Contemporary Chinese. Wang (1980: 434) pointed out that the passive morpheme bèi is no longer used in the spoken language of Mandarin Chinese, as we confirmed in our statistical survey of the spoken language in Contemporary Pekingese, shown in Table 8.3. In Table 8.3, of all the passive instances, the bèi passives account for only 2 percent, showing that they have become nearly extinct in the spoken language. Instead, the ràng passives have become absolutely dominant in passive expressions. A remarkable commonality is shared by the newly grammaticalized passive morphemes ràng, jiào, and gěi, all of which require an agent phrase in order to make their passive structures well formed, akin to the dummy subjects it or there in English declarative sentences. One principle of English grammar is that every finite clause requires a subject for well-formedness, even when the subject is expressed unnecessarily. Thus English makes extensive use of it or there to fill in the subject position in cases where the subject cannot be identified in certain contexts, e.g. It is raining (*Is raining) and There was a magazine on your table (*was a magazine on your table). Likewise, for the new patterns of Chinese passives, when an agent phrase cannot be identified, nouns referring to general concepts, such as rén “people,” dàjiā “everybody,” and shénme shì “something,” must be employed to occupy the object position of the passive preposition in order to make the passive instances grammatical, as illustrated in (20):

Table 8.3 Frequency of the four passive morphemes in Pekingese

Text

Size of text

The Selection of Cross Talks by Jiang Kun and Liangzuo7

200,000 Chinese characters

7

Bèi passive

Ràng passive

Jiào passive

Gěi passive

2 (4%)

40 (86%)

2 (4%)

3 (6%)

This was published by the Culture and Art Press, Beijing, 1992.

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

From Optional Agent to Obligatory Agent (20)

(a)

小偷讓人綁樹上了。 (北京話相聲) Xiǎotōu ràng rén bǎng shù shang le. theft PASS people truss tree at PERF “The thief was trussed to a tree.” (Cross Talks of Contemporary Pekingese)

(b) *小偷讓綁樹上了。 *Xiǎotōurang [ ] theft PASS (21)

(a)

191

bǎng truss

shù tree

shàng le at PERF

他可能叫什麼事拖住了。 (北京話相聲) Tā kěnéng jiào shénme-shì tuō-zhù le. he probably PASS something drag PERF “He is probably bring dragged by something.” (Cross Talks of Contemporary Pekingese)

(b) *他可能叫什麼事拖住了。 *Tā kěnéng jiào [ ] tuō-zhù le. he probably PASS drag PERF The “meaningless” agents do not need to be spelled out in the corresponding English instances, but once they are removed, the passive instances become ungrammatical, as in (20b) and (21b).

8.5.5

Functional Commonality of Passive Morphemes

Currently, the overwhelming majority of Chinese dialects no longer use the bèi passive and have invented other passive morphemes for use. According to our extensive survey of relevant publications in English, only Yue (2015) briefly mentioned that, in certain dialects, an agentive NP must occur in order to make the passive structure well formed, but she does not provide details in exploring the theoretical and empirical importance of this extraordinarily interesting property of the Chinese passive. Cao (2008: 95) presented a comprehensive investigation of the grammars of 930 Chinese dialects, including passive usage. Based on Cao’s work, we carried out a comprehensive statistical survey of the passive morphemes and their associated constructions, and our findings are shown in Table 8.4. A total of sixty-nine different passive morphemes are used in those dialects, but, for the sake of simplicity, Table 8.4 provides only the four most common passive morphemes in Chinese dialects. Of the 930 dialects investigated, only seventy-one still use the bèi passive, accounting for less than 8 percent of all passive morphemes. Surprisingly, this is the only historical

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

192

Table 8.4 The most common passive morphemes in dialects The gěi passive (from “give”)

The jiào passive (from “allow”)

The ràng passive (from “allow”)

The bèi passive (from “suffer”)

250 dialects

214 dialects

40 dialects

71 dialects

example of such a wide range of lexical items (all of which were originally regular verbs) becoming grammaticalized into so many different passive morphemes at the same time; this phenomenon deserves further exploration. According to the comprehensive investigation by Cao (2008: 95), however, the syntax of these numerous passive morphemes is quite uniform in comparison with their counterparts in Old Chinese, all of which share the same schema: (22)

Subj + (PASS Agent) + VP

In many dialects, the appearance of an agent phrase is obligatory to make a passive structure grammatical, but in other dialects the agent phrase may be optional. However, there are no dialects whose passive morphemes cannot introduce an agent phrase, unlike the jiàn passive and the bèi passive in Old Chinese. Additionally, these dialects do not allow any passive morphemes to introduce an agent phrase in postverbal position, such as the preposition yú in Old Chinese; this feature is determined by the overall property of Chinese grammar. To calculate the percentages of the presence and absence of agent phrases in passives of Contemporary Chinese texts, I investigated a vast number of vernacular texts, and the statistical results are shown in Table 8.5. Agent phrases are purely optional in the bèi passive structure, which has a history of nearly 2,300 years but has been constrained to written texts and formal speech in Contemporary Chinese. The current syntax of the bèi passive is inherited from its old usage. In contrast, almost 100 percent of the other three passive morphemes require an agent phrase in order to make the passive structure grammatical. Additionally, Song et al. (2007) extensively investigated passives with regard to the absence or presence of agent phrases. Their statistical results are very similar to those shown in Table 8.5. Their investigation revealed that within a corpus of novels consisting of approximately a million Chinese characters, there are 107 total instances of the jiào passive, all of which have agent phrases; of the forty-one ràng passives, only one example is agentless. Passive morphemes in preverbal position have undergone a dramatic change throughout history, from one extreme to another: (a) none of them could introduce agent phrases in Old Chinese, and (b) all of them must introduce agent phrases in Contemporary

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

Passive with an Obligatory Agent

193

Table 8.5 The proportion of passives with/without agent phrases Passive morphemes

+ Agent phrase

− Agent phrase

bèi ràng jiào gěi

59 (52%) 139 (98%) 55 (100%) 3 (100%)

55 (48%) 2 (2%) 0 (0%) 0 (0%)

Chinese. Once again, the motivations for these functional distinctions still reside in the overall properties of the grammatical system in particular periods.

8.6

Passive with an Obligatory Agent

Kiparksy (2013) claimed that there are actually no passive-specific constructions because the syntax of passives is predictable from that of active structures. The evolution of Chinese passives was motivated by schematic changes in the grammar, which consists mainly of active constructions. Specifically, the global property of the grammar in a particular period influences and even shapes the configuration and intrinsic constituency of passives, the function and life span of passive morphemes, and even the selection of lexical candidates for grammaticalizing into passive morphemes. From this vantage point, considering the history of the Chinese passive, we realize that no changes in Chinese passives were accidental or isolated, and each was due to analogy with the whole grammatical system. There is a fundamental word order change in declarative sentences with regard to the positioning of PPs, such as the locative, comparative, instrumental, scalar, and passive. A detailed exploration of the schematic change in each type of declarative sentence is beyond the scope of our concern; here, we can only sketch the global changes that all of them have undergone: (23)

Subj V Obj PP > (a) Subj PPnon-resultative V Obj; (b) Subj V Obj PPresultative

Briefly, the motivation for the above schematic change was the establishment of the resultative construction, which in turn triggered the occurrence of the new principle of information structure. Under this information-organizing principle, only preposition phrases denoting resultatives such as an end point or recipient still remain in postverbal position, whereas those adjuncts of preposition phrases expressing non-resultative features could no longer stay in their original postverbal position. Note that a group of verbs grammaticalized in the first verb position of the serialization structure,

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

194

The Passive Construction

functionally taking over the tasks of the old grammatical markers. That is, the schematic change involved no movement at all. The process started as early as the first century BC and was not complete until the eighteenth century AD, lasting nearly 2,000 years. The four major types of declarative sentence in Old Chinese and their corresponding structures in Modern Chinese are exemplified in Section 9.1. In general, the NP within the PP adjunct in preverbal position is obligatory for the well-formedness of structures.8 Now, the passive with an obligatory agent in Contemporary Chinese no longer appears unusual but simply is a property in accordance with the abstract schema shared by all major types of declarative sentence involving both active and passive structures. Nevertheless, the direct impetus for the emergence of the passive pattern with an obligatory agent is perhaps the emergence of the disposal construction, a new grammatical structure that was first introduced into the language in the seventh century AD, developed quickly from the tenth century AD to the thirteenth, and spread widely after the fifteenth century AD (Wang 1989: 266‒271). A very brief history of the disposal construction may be stated as follows. Due to the fusion of the verb and the resultative, certain definite NPs that originally occurred between the verb and the resultative had to be arranged in a preverbal position; thus some types of sentence adopted the SOV word order. This in turn triggered the grammaticalization of jiāng and bǎ, two “take” verbs, as accusative markers indicating a preposed patient (object). Once again, jiāng and bǎ were reanalyzed as the so-called disposal marker in the first verbal position of the serial verb construction. The grammaticalization of jiāng happened around the seventh century AD, 200 years earlier than that of bǎ. After they competed for the disposal construction for more than 1,000 years, bǎ finally won over jiāng in the northern dialects (including standard Mandarin), but jiāng survived in some southern dialects, such as the Wu dialect. The passive morphemes and the disposal morphemes are functionally complementary: the former serve to mark the agent in a preverbal position, and the latter serve to mark the patient in a preverbal position. In addition, both share the same schema, “Subj + PREP-NP + VP.” Within the period from Late Medieval Chinese to Contemporary Chinese, there were four disposal morphemes in total, namely, jiāng, bǎ, ná, and guǎn, all of which have one thing in common: the patient NP is obligatory for the well-formedness of the disposal construction.9 Any native speaker of the Chinese language could readily sense an ill-formed disposal structure if its patient noun were deleted, as discussed in great detail in Chapter 9. 8 9

When it appears alone, the morpheme zài functions like a progressive aspect rather than a locative preposition. Thus this is not a counterexample to our assertion. Li and Thompson (1981: 463‒491) claimed that the patient can be omitted from the disposal structure, but we have found no such examples in our large-scale empirical investigation of numerous texts from the eighth century to the present.

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

Passive with an Obligatory Agent

195

Based on the above analysis, we can draw the following generalizations. (a)

(b)

The motivations for the emergence of the passive with an obligatory agent in Contemporary Chinese. We are now in a position to explain the strikingly unusual direction of the development of passive morphemes in the past roughly 1,000 years. The firm establishment of the disposal construction, which required a patient phrase in order to make it well formed, tended to be analogical to its counterpart, namely passive structures; under this circumstance, all passive morphemes after the sixteenth century AD tended to require an agent noun in order to make passive clauses well formed as well. However, the passive morpheme chī, originating from its “suffer” usage, failed to meet this requirement and hence became extinct around the sixteenth century AD. For the same reason, the dominant bèi passives in early Modern Chinese gradually went out of fashion after the sixteenth century AD and ultimately became restricted to written language and formal speech. The selection of lexical items for the grammaticalizing of passive morphemes. According to Heine and Kuteva (2002: 333), at least nine types of lexical source (e.g. locative, anticausative, reflexive) have the potential to become grammaticalized into passive morphemes in different languages or during different periods within the same language. Which lexical items are recruited for a given functional domain by a particular language is entirely random. However, there are some constraints on the selection of lexical candidates for grammaticalization if we consider the evolution of the language as a whole. When the agent phrases of passive structures became obligatory by analogy with the newly formed disposal construction, which might be regarded as optimization via the overall grammatical priorities during a particular period (Kiparsky 2011), the lexical items in the first verb position of a serial verb construction became favorable for grammaticalizing into passive morphemes. The reason is that the first verb in this structure is always followed by an NP, which shared the same schema with the optimum scheme, that is, Subj + (P + NP) + VP. The verbs that frequently occurred in the first verb positions are most likely to be reanalyzed into passive morphemes. For example, the verbs jiào “passively allow,” ràng “passively allow,” and gěi “give” are most semantically suitable for expressing the grammatical function. Under the influence of this trend in Contemporary Chinese grammar, some unusual grammaticalization paths have also been found in the dialects; according to Cao (2008), there are thirty-nine dialects whose morphemes came from the verb děng “wait,” six

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

196

whose morphemes came from the verb yāng “ask for,” and four whose passive morphemes came from the verb yào “want,” among others. Grammaticalization paths of this type have not yet been attested in other languages, and it was the overall grammatical property in particular periods that enabled so many verbs to undergo the change. In addition, this is why the lexical sources for passive morphemes are so diverse, with sixty-nine different morphemes in the 930 dialects investigated by Cao (2008: 95), as pointed out previously.

8.7

Lexical Sources for Passive Morphemes

Among all grammatical categories, the passive has the greatest degree of diversity or the most variations across languages, with a total of fourteen types that grammaticalized from different lexical sources. Chinese adds one more type to this list; that is, the passive markers in Mandarin Chinese and hundreds of dialects that have evolved from verbs meaning “allow” and “let be” (i.e. jiào and ràng), which are not listed in Kuteva et al. (2019: 484). Additionally, they miss a major type of passive marker that grammaticalized from locative prepositions, a widely attested phenomenon across languages. As far as I know, there are at least sixteen types of passive marker, cross-linguistically, and this number is likely to increase as the typological investigation continues. The reason why passive markers are so diverse is that different language communities construe the passive event from many different perspectives, as summarized in Table 8.6. Table 8.6 reveals some important properties of the passive. One is that the number of possibilities of construal varies from one perspective to another. With regard to passive expressions, the patient has the focal prominence and thus manifests the most possible perspectives of construal, totaling six in Table 8.6. At the same time, the agent is defocused in passive expressions and downgraded as an adjunct, thus manifesting the fewest possible perspectives of construal (i.e. zero). When the agent is viewed as the focal prominence, the expression should be an active sentence, and grammatical devices related to active instances should therefore be employed. In the richly documented history of the Chinese language, there have been more than ten passive markers and various passive structures. For the sake of simplicity, we consider only the dominant passive structures in each major period, which are summarized in Table 8.7. Every time the language community takes a different perspective to construe the passive event, namely selecting different lexical sources for passive markers, it involves a change in the syntactic configuration of the passive. The motivations for shifting perspectives of construal come exclusively from overall change in the grammatical

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

Lexical Sources for Passive Morphemes

197

Table 8.6 Construal and diverse passive markers across languages Perspectives of construal

Lexical sources

I. The location

(a) (b) (c)

The “in” preposition The “by” preposition The “get close to” verb

II. The patient

(a) (b) (c) (d) (e) (f) (g)

The “suffer” verb The “eat” verb The “see” verb The “get” verb The “allow” verb The “let be” verb The “give” verb

III. The change in state

(a) (b) (c) (d) (e)

The “become” verb The “come” verb The “go” verb The “be” verb The “fall” verb

IV. The valence loss of the verb

(a) (b) (c) (d)

Anticausative Nominalizer Person pronoun Past participle

Table 8.7 Types of passive in the history of Chinese Period

Perspective of construal

Example

Old Chinese (1300 BC–100 BC)

(a) (b)

Locative The “see” verb

yú (於) jiàn (見)

Medieval Chinese (100 BC–AD 1000)

(a) (b)

Change of state The “suffer” verb

wèi . . . suǒ (為 . . . 所) bèi (被)

Modern Chinese–Present (AD 1000–2000)

(a) (b)

The “allow” verb The “give” verb

jiào and rang (叫, 讓) gěi (給)

system; that is, the construal perspective of a language community would not change for no reason. Similar to the sentential structures of English, Old Chinese as an SVO language allowed various types of preposition phrase to occur in sentence-final position (i.e. after VO) to introduce locatives, comparatives, instrumentals, and others. For passive instances, the locative preposition yú was extended to introduce an agent noun in sentence-final position if the agent needed to be indicated in certain contexts. There was also a passive auxiliary at that time that grammaticalized

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

198

from the “see” verb; it preceded the main verb but was unable to introduce an agent noun in any situation. Both the preposition yú and the auxiliary jiàn could be used independently to mark a passive clause, and they could also be combined to make a passive instance, as in the following example: (24)




必見驁於民。(商君書 更法) Bì jiàn ào yú mín. definitely PASS respect by people “(He) is definitely respected by the people.” (Shang Jun Shu, Geng Fa, 350 BC)

During the transition period from Old to Medieval Chinese, the language underwent a systematic change with regard to sentential structures. As a result, all preposition phrases in passives, as well as in locatives, comparatives, and instrumentals, could no longer appear in sentence-final position, and they gradually disappeared from that position. In response to this global change, the passive pattern marked by yú collapsed, and since expressing an agent was sometimes necessary, the language community started to recruit a new lexical candidate to take on the task of introducing an agent in preverbal position. At this time, the passive auxiliary bèi, which had been grammaticalized as early as the first century BC but previously could never introduce an agent (like the passive auxiliary jiàn), became the best candidate to fill this gap. As a result, bèi developed a new function of introducing an agent noun, similar to a preposition, as illustrated in (25). In contrast, the passive auxiliary jiàn, which had been widely used in Old Chinese, did not acquire the function and thus was abandoned. (25)




亮子被蘇峻害。 (世說新語 方正) Liàng zǐ bèi Sū Jùn hài. Liang son PASS Su Jun kill “Liang’s son was killed by Su Jun.” (Shi Shuo Xin Yu, Fang Zheng, AD 450)

During the period from late Medieval Chinese to Modern Chinese, one of the significant changes in Chinese grammar was the emergence and development of the disposal construction. The disposal construction not only shares the same schema as the passive, “PP + VP,” but, more importantly, is functionally complementary with the passive pattern. Specifically, the disposal markers serve to introduce a patient noun in preverbal position, and the passive markers function to indicate an agent noun in the same position; the patient and agent are functionally related to each other. From the beginning of the emergence of the disposal construction, the appearance of a patient noun has been obligatory to make a disposal instance well formed, as illustrated in (26). Most

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

Lexical Sources for Passive Morphemes

199

likely because of the analogy with the property of the disposal construction, an agent noun gradually became obligatory for the well-formedness of passive instances. The passive bèi never reached this extreme of necessarily introducing an agent noun, and its usage has greatly decreased since the sixteenth century AD. Among the total of 930 dialects investigated by Cao (2008: 95), only seventy-one still use the passive bèi, and in the vast majority of dialects the passive markers grammaticalized from other lexical sources, mainly the verbs meaning “give,” “allow,” or “let be.” More importantly, all these new passive structures require an agent noun in order to be grammatical, as illustrated in (27), similar to every English sentence with a finite verb needing a subject to be well formed. (26)

(a)

他把房間收拾了一下。 (現代漢語) Tā bǎ fángjiān shōushí-le yīxià. he DISP room clean-PERF one-time “He cleaned the room for a while.” (Contemporary Chinese)

(b) *他把收拾了一下。 *Tā bǎ [ ] shōushí-le he DISP clean-PERF (27)

(a)

(b)

yīxià. one-time

小偷讓人綁樹上了。 (北京話相聲) Xiǎotōu ràng rén bǎng shù shàng le. thief PASS people truss tree at PERF “The thief was trussed to a tree.” (Cross Talks of Contemporary Pekingese) *小偷讓綁樹上了。 *Xiǎotōu ràng [ ] bǎng shù shàng le thief PASS truss tree at PERF

Once again, the grammatical system as a whole played a key role in recruiting the appropriate lexical candidates for passive markers. The verb jiào “allow” and ràng “let be” are often used in a pivotal construction, where the following object is also necessary; otherwise, the pivotal instances will be ill-formed, as illustrated in (28): (28)

(a) 我讓他走了。 (現代漢語) Wǒ ràng tā zǒu I let he go-away “I let him go away.”

le. PERF (Contemporary Chinese)

(b) *我讓走了。 *Wǒ ràng [ ] I let

zǒu go-away

le. PERF

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Passive Construction

200

We want to highlight that the function of a grammatical marker is generally influenced by its lexical source, semantically and syntactically. When the language requires the agent noun to appear in the passive, the language community favors a lexical item whose context for the grammaticalization path must contain a nominal object. Both the “allow” and “let be” verbs satisfied this condition and thus underwent the change to passive markers. One might wonder why the already existing passive marker bèi did not undergo further development to acquire the function of the agent noun obligatorily being present. It is cross-linguistically true that the old features of grammatical forms with high frequency are typically resistant to any analogical change (Bybee and Beckner 2015). In Late Modern Chinese, bèi had been used as a passive marker for more than 1,800 years, and it was never able to introduce an agent noun in the first six centuries or so after it grammaticalized from a “suffer” verb. It was difficult to eliminate the long and highly conventionalized usage of the passive bèi due to its high frequency. Therefore the language community sought new passive markers to satisfy the requirement that the agent must appear in passive instances. Through the above analysis, we can see that the selection of lexical sources for a grammatical marker is determined largely by the social construal of a language community, and the perspective of the social construal is in principle guided by the overall property of a grammatical system at the particular time. Selecting a lexical source is not simply a semantic matter but involves the syntax of the lexical item. The language community tends to recruit a lexical item whose syntactic configuration resembles the structure of the grammatical marker.

8.8

The Development of Passive Structures

The Chinese language has a history of more than three millennia of uninterrupted documentation since the era of the Oracle Bone Inscriptions, having produced vast texts in rich genres such as religious works, poems, dialogues, novels, and many scholarly books. No other language in the world is comparable to Chinese in this regard. Thus the language is a unique and invaluable treasure for studies of historical linguistics and linguistics in general, as it can be considered a natural experimental laboratory for any linguistic hypothesis and theoretical framework. However, cautious and large-scale investigation is needed when researchers try to use historical empirical evidence to support their analyses or propose hypotheses. When carrying out research in historical linguistics, researchers must keep in mind that any theoretical bias may influence the goals of their studies and the accuracy of their investigation. The diachronic changes in the Chinese passives are by no means isolated, and any plausible explanation, from our viewpoint, must take into account the evolution of the general grammatical system. Superficially, many passive morphemes and structures

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

The Development of Passive Structures

201

have been used throughout history; however, their development has been far from random and accidental. The grammaticalization processes, syntactic functions, and even life spans of these passive morphemes and their related patterns are motivated and even governed by the overall property of the ever-changing grammatical system in particular periods. In general, two major changes in history are mainly responsible for the development of the passives. The first is that all non-resultative PPs were eliminated from postverbal position due to the emergence of the new information-organizing principle, which happened during the transition period from Old Chinese to Medieval Chinese. This global evolution directly resulted in structural changes of many types of declarative sentence, involving the locative, instrumental, comparative, and passive. The second is that the disposal construction emerged in the Late Medieval Chinese period, and the appearance of the “PREP. + patient” phrase in this construction was obligatory to make the construction grammatical.10 Passives and disposal structures are functionally complementary: the disposal introduces a patient in preverbal position, and the passive introduces an agent in preverbal position. Probably through analogy with the obligatory usage of the patient in disposal structures, the agent phrase gradually became obligatory in passive structures. This feature of passive structures in Contemporary Chinese, a typologically surprising property, is in sharp contrast to their counterparts in Old Chinese. All passive morphemes and structures throughout Chinese history fall into the following three types, according to two criteria: whether the morphemes by themselves can introduce an agent phrase and where the agent phrase is located, namely in either postverbal or preverbal position. Type 1. The passive morphemes cannot by themselves introduce an agent phrase in preverbal position, and the preposition yú must be used to do so in postverbal position. All passive markers and structures in Old Chinese belong to this type. Note that the preposition yú, which introduces an agent in postverbal position, could by itself mark a passive instance and could also be paired with an auxiliary passive marker such as jiàn, bèi, or wéi to form a complete passive construction. As auxiliary verbs, both jiàn and bèi in Old Chinese could not introduce an agent phrase in preverbal position, and they had to rely on the preposition yú in postverbal position to do so when the agent had to be expressed. The passive morpheme wéi might be an exception in this regard: it could either introduce an agent by itself in preverbal position or rely on the preposition yú to do so in postverbal position. 10

In fact, the disposal morphemes include jiāng, bǎ, and ná, and bǎ is selected as the representative marker for the disposal construction.

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

202

The Passive Construction Type 2. The passive morphemes can optionally introduce an agent phrase in preverbal position only, and no other preposition is needed or allowed to do so in postverbal position. All passive morphemes in Medieval Chinese and Modern Chinese belong to this type, including the late function of the bèi passive (i.e. the usage beginning in the fourth century AD), the wéi passive, the wéi . . . suǒ passive, and the chī passive. Type 3. The passive morphemes must introduce an agent phrase in preverbal position in order to make the passive structures well formed. This type of passive includes the two major types of passive morpheme in Contemporary Chinese, the jiào passive and the ràng passive, as well as the less frequent gěi type. Likewise, numerous passive morphemes in Contemporary Chinese dialects also require agent phrases in order to make passive instances well formed. Typologically, this feature seems to be unique and may be owing to analogy with related constructions, particularly the disposal construction.

Superficially, the passive morphemes in other Chinese dialects appear to be extremely complex, but their functions are quite uniform, which is consistent with Mandarin Chinese. As pointed out previously, there are at least sixty-nine different passive morphemes in Contemporary Chinese dialects that have originated from numerous lexical sources. Their syntactic properties, however, are consistent, and without exception all passive morphemes can introduce agent phrases in preverbal position, either optionally or obligatorily, in conformity with the newborn passive morphemes in Contemporary Mandarin Chinese. The question of why there is such diversity in the lexical sources that became grammaticalized into passive morphemes in Chinese dialects is certainly worthy of further exploration.

https://doi.org/10.1017/9781108921831.008 Published online by Cambridge University Press

9 The Disposal Construction

9.1

Introduction

Synchronically and diachronically, a great deal of research has been conducted on the semantics and syntax of the disposal construction in both traditional Chinese linguistics and general linguistics internationally. There are two main reasons why this construction has been the focus of Chinese linguistics: first, its syntactic configuration makes it appear to be an SOV word order, in sharp contrast to the basic SVO word order of the Chinese language, and second, Chinese seems to be the only case across languages where a superficial typological change from SVO to SOV might have occurred, motivated by internal evolution rather than language contact. Thus numerous proposals have been made about the syntactic functions of the disposal construction and diachronic motivations for its development. Unfortunately, in our view, none of them are successful in characterizing its grammatical properties or identifying its diachronic development. In this chapter, we demonstrate that because of the fusion of the verb and the resultative via reanalysis, the syntactic position for a patient noun between them was eliminated so that the originally intervening patient noun needed to be rearranged somewhere in the sentence on the basis of the following principle: the patient noun was introduced in preverbal position by the disposal marker if it was definite, and was introduced by the first verb of the verb-copying construction if it was indefinite (for details, see Section 10.1). This issue is related to the motivations for the emergence of the verbcopying construction and the unmarked SOV structure, all of which were entirely innovative in Modern Chinese. In the literature, there are two labels for the construction in question: “the disposal construction” and “the bǎ construction,” which have been used interchangeably. However, we prefer the former over the latter because there were at least five grammatical morphemes marking the construction in history: jiāng (from the verb take), bǎ (from the verb hold), ná (from the verb take at a later stage), guǎn (from the verb govern), and gěi (from the verb give), and the last four are still used in Standard Contemporary Chinese, let alone that there are dozens of different disposal

203

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

204

morphemes in other Chinese dialects. We follow the convention of using the term “disposal construction,” although it is somewhat misleading in capturing the common semantics of the construction; for example, no “disposal” sense can be felt in the following examples: (1)

他把老李當作自己的偶像。(現代漢語) Tā bǎ Lǎo Lǐ dāng-zuò zìjǐ de ǒuxiàng. he DISP Lao Li regard-as self GEN idol “He regarded Lao Li as his idol.” (Contemporary Chinese)

(2)

把個北京城走了一多半。 (現代漢語) Bǎ gè Běijīng chéng zǒu-le yī-duō-bàn. DISP one Beijing city walk-PERF one-more-half “(He) walked on more than half of the roads of Beijing city.” (Contemporary Chinese)

The disposal construction appears to be quite special or unusual in Chinese grammar, both synchronically and diachronically, with the syntactic schema “Subj + PREP + Obj + VP” that seemingly resembles SOV word order. Two questions naturally arise: first, why did Chinese, as a typical SVO language throughout history, acquire such a special construction? Second, what is the function of this disposal construction? In the literature on Chinese linguistics, numerous hypotheses have been proposed in attempts to explain where it came from and what its function is, and some of them are briefly reviewed in the subsequent section. Moreover, these issues have aroused much interest among researchers in linguistic typology. On the basis of the development of the disposal construction, Li and Thompson (1974) claimed that Chinese had shifted its basic word order from SVO to SOV, a counterexample to the historical generalization across languages that without external factors (i.e. language contact), an SOV language can evolve into an SVO language, but the reverse process is impossible (Lehmann 1973, Givón 1979, Light 1979, Smith 1981, Sun and Givón 1985). Moravcsik (2010) stated this concept as follows: Universal historical generalization. In all languages, in spontaneous historical change, the source of the direct inversion of verb and object is OV and not VO. For example, English and French, two members of the Indo-European family that historically had the SOV word order, have become rigid SVO languages over time. If the above historical generalization is universal, then Chinese is the only exception to the above universal historical generalization.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Hypotheses about Its Origins 9.2

205

Hypotheses about Its Origins

Researchers, even those who are working in the same theoretical framework, usually hold quite different philosophies about the nature of language, which to a great extent influence their achievements, particularly in making assumptions when approaching certain issues and putting forward hypotheses to explain certain phenomena. Thus, in historical linguistics, there is hardly any consensus on any given issue regarding what could happen and what could not happen. We regard this as an ideal chance to frankly state our own philosophical view on language by critically reviewing other researchers’ explanations for the same phenomenon. In what follows, we will review several influential hypotheses about the motivation for the emergence of the disposal construction. Hypothesis 1. The analogous effect of the SOV language features in Old Chinese When they first proposed that the emergence of the disposal construction signaled a word order change in Chinese from SVO to SOV, Li and Thompson (1974) offered two hypotheses about why this could have happened: first, the verb bǎ “hold” grammaticalized as a disposal marker in the first verb position of a serial verb construction, which collapsed two clauses into a single one. It is undeniable that a grammaticalization process must happen in certain specific contexts. However, critical questions remain unanswered: what triggered the grammaticalization process? More importantly, what propelled development of a major new construction in the grammar? In our view, grammaticalization theory itself cannot provide any direct explanation for these “why” issues in the diachronic investigation of any phenomenon. Second, Li and Thompson (1974: 207–208) argued that the emergence of the disposal construction was motivated by other SOV characteristics that originated from Old Chinese or the prehistorical language: Old Chinese possessed significant SOV characteristics even at the time when word order was strictly SVO. Thus, the structure of the noun phrase in Old Chinese was completely that of an SOV language: relative clause always preceded head noun; genitive was always positioned before the noun; adjectives and other modifiers always occurred before the modified. The existence of these OV characteristics in Old Chinese might very well have provided an impetus for the SVO order to shift to SOV. In other words, we suggest that the presence of OV properties may have had a catalytic effect, inducing the language to shift toward the OV word order. First, the term “Old Chinese” in the literature (equivalent to Old Chinese in our analysis) is reserved to refer to the period of the language before the first century BC, but the disposal bǎ came into existence after the sixth century AD. Obviously, there was

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

206

a 600-year gap between the so-called cause and the effect, and how could this happen? Empirically speaking, in the Chinese language before the sixth century AD, the relative clause marked by the relativizer zhě had to follow the head noun, a widely used type of noun phrase, as illustrated in (3). Similarly, complex adjectival phrases, which were typically marked by the pronominal morpheme zhī, usually followed the head noun, as illustrated in (4) (for details, see Chapter 21):




(3)

子產, 君子之求樂者也。 (左傳 昭公十三年) Zǐ Chǎn, jūnzǐ zhī qiú lè zhě yě. Zi Chan gentleman GEN seek happiness REL PRT “Zi Chan was a gentleman who sought happyness.” (Zuo Zhuan, Zhao Gong Shi San Nian, 550–400 BC)

(4)

孟氏選圉人之壯者三百人。(左傳 定公八年) Mèng Shì xuǎn yǔ-rén zhī zhuàng zhě sān bǎi rén. Meng Shi select prisoner GEN strong REL three hundred person “Meng Shi selected 300 prisoners who were physically strong.” (Zuo Zhuan, Ding Gong Ba Nian, 550–400 BC)




After the sixth century AD, the relative clause and adjectival modifiers were gradually restricted to occurring in prenominal position. Until then, these features had existed in the language for more than a millennium; therefore, even if we accepted their view of the correlation between the basic word order and the “impetus” features, the structure of the noun phrase at that time was in fact inconsistent with Li and Thompson’s claim. In other words, the empirical evidence does not support their hypothesis. Theoretically, typological investigations have revealed that there is no correlation between the constituent order of phrases and the basic word order of clauses (for detailed discussion, see Dryer 1992). From this point of view, it is unlikely that the constituent order of the noun phrase could have served as an impetus for the emergence of the disposal construction, which involved word order change at the clausal level. Although Li and Thompson’s (1974) proposal for the putative word order change is well known in academic society, their explanation is not popular, and few scholars have paid serious attention to it (Moravcsik 2010). Hypothesis 2. Language borrowing from Buddhist scripture translations. Cao and Yu (2000) provided another explanation, surprisingly, that the disposal markers at that time, jiāng and bǎ, may have been used by analogy with an early structure, “qǔ (take) NP + VP,” that was found in Buddhist scripture translations from

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Hypotheses about Its Origins

207

the third century AD to the sixth. They believed that Sanskrit, which was an SOV language, affected the Chinese translation, producing the following structure: (5)

王今取我身體碎如芥子。(出曜經) Wáng jīn qǔ wǒ shēntǐ suì rú jièzǐ. You now take I body crush like mustard “Now you would crush my body into mustard.” (Chu Yao Jing, AD 400)

This phenomenon can at best be called “language borrowing” rather than “language contact” (e.g. Feng 2014) because the latter term usually means that two groups of people who speak different languages live together and communicate with each other every day. The translations of religious documents belonged to written language, and there is no report of this kind of grammatical change induced by translation in written language ever happening in any other language, though the translation of religious documents is extremely common across languages with different word orders. Indeed, some SVO languages, e.g. Bukhara Arabic, indeed changed to SOV owing to extensive language contact (Ratcliff 2005, Moravcsik 2010), but this language contact refers to daily communication among ordinary people, which is totally different from the “language-translating” hypothesis of Cao and Yu. At first glance, we are reluctant to accept the explanation of Cao and Yu because there are two obvious cases in relation to Chinese that immediately invalidate their argument. Historically, Japanese and Korean were heavily influenced by the Chinese language; for instance, they borrowed much of the basic lexicon from Chinese as well as the Chinese writing system. More importantly, these two countries historically adopted the original ancient texts, such as Confucius (i.e. Lun Yu) and Mencius (i.e. Meng Zi) as textbooks in their primary schools. In addition, they sent many young people to study in China in ancient times. However, both of their languages remained rigid SOV languages, with no sign that their basic orders were ever affected by the SVO order of the Chinese language, perhaps because the majority of ordinary people in these two countries historically did not live with Chinese people. Needless to say, language contact can induce a change from SOV to SVO (reflecting the universal historical generalization; see Moravcsik 2010) supposedly much more easily than the reverse word order change; in line with this analysis, it should be more likely that the word order of Chinese might have caused Japanese or Korean to undergo a basic word order change from SOV to SVO, in comparison with the hypothesis of Cao and Yu. Moreover, there are two key issues that Cao and Yu should have been concerned with: why could the verb qǔ not take a further step and develop into a dominant marker for the

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

208

disposal construction? Why was it replaced by other grammatical markers, such as jiāng and bǎ? In fact, at least five verbs at the time – qǔ “take,” jiāng “take,” bǎ “hold,” zhuō “capture” and chí “carry,” which were synonymous or semantically closely related to each other – behaved similarly and could have introduced a patient noun in the first verb position of serial verb construction (Jiang and Cao 2005: 352). In our view, it is better to think of them as content verbs rather than grammatical morphemes at the time, as they were all competing for a grammatical function that was required by the grammatical system at that time, and the final winners for marking the disposal construction were first jiāng and then bǎ. This phenomenon happened repeatedly in the evolution of Chinese grammar; for example, in Medieval Chinese, at least four content verbs – liǎo “finish,” bì “complete,” jìng “fulfill” and yǐ “end” – were competing for the expression of the perfective aspect, but only liǎo (i.e. -le via phonological reduction) eventually won over the others (Ohta 1987: 85‒86, Wang 1989: 102‒121). As Jiang and Cao (2005: 198‒ 225) correctly pointed out, there was actually no replacement of the earlier qǔ by the latter jiāng and bǎ, and all of them had to undergo their own grammaticalization process in specific contexts. Hypothesis 3. The replacement of the preposition yǐ in Old Chinese. With regard to the lexical source of the disposal morpheme, there has been a very popular idea in Chinese linguistic circles that the disposal bǎ came into existence simply via a lexical replacement of the early morpheme yǐ, a common preposition in Old Chinese with multiple functions, such as introducing an instrument, a reason, a time point, and a direct object for a ditransitive construction. It was first proposed by Ohta (1987: 241‒245) and followed by many other researchers (e.g. Chen 1956, Mei 1981, Peyraube 1986, Sun 1996: 51‒81). It is true that some instances of the preposition yǐ could be translated into instances of the disposal bǎ, as illustrated in (6), but translation is by no means a reliable way to identify the cognates between grammatical morphemes. (6)




(a) 天子不能以天下與人。 (孟子 萬章) Tiānzǐ bùnéng yǐ tiānxià yǔ rén. emperor not-can with world give people “The emperor cannot give the world to others.” (Meng Zi, Wan Zhang, 300 BC) (b) 天子不能把天下給別人。 (現代漢語) Tiānzǐ bù-néng bǎ tiānxià gěi biérén. emperor not-can DISP world give other “The emperor cannot give the world to others.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Hypotheses about Its Origins

209

However, a simple investigation would reveal the essential difference between the functions of the two prepositions yǐ and bǎ: the former could introduce an indefinite noun, as exemplified in (7), but the latter has not been able to do so throughout history, which we will discuss in subsequent sections. (7)




復以弟子一人投河中。 (史記 滑稽列傳) Fù yǐ dìzǐ yī-rén tóu hézhōng. again with young-man one-CL throw river-in “They threw a young man into the river again.” (Shi Ji, Hua Ji Lie Zhuan, 100 BC)

Considering the semantic requirement of the patient noun, the functions of yǐ and bǎ were essentially different from one another. In other words, historically, there could not be a lexical replacement relationship between them. More importantly, the distribution of the preposition yǐ was fundamentally different from that of the disposal bǎ in history: the former could occur either before or after the predicate, as exemplified in (8), but the latter could never occur after the predicate. (8)




虞思於是妻之以二姚。 (左傳 哀公元年) Yú Sī yúshì qī zhī yǐ èr Yáo. Yu Si then marry he with two Yao “Then Yu Si married him to the two Yao girls.” (Zuo Zhuan, Ai Gong Yuan Nian, 550–400 BC)

In contrast, throughout history the disposal bǎ phrase could never occur in sentencefinal position. According to Yang (1981: 247), the preposition yǐ had at least eight different functions in Old Chinese, such as introducing reasons, instruments, and time nouns, which the disposal bǎ never had. Even in preverbal position, the object of yǐ could be either omitted, as illustrated in (9), or inverted, as illustrated in (10),1 but the disposal bǎ never had similar functions. (9)

1




御人以告子元。 (左傳 莊公二十八年) Yù rén yǐ [ ] gào Zi Yuán. drive man with tell Zi Yuan “The driver told Zi Yuan the matter.” (Zuo Zhuan, Zhuang Gong Er Shi Ba Nian, 550–400 BC)

The omission of the object of yǐ is conditioned by the fact that the referent of the object was either mentioned in the previous context or was clear from the context. The fronted object of yǐ was the focus of the sentence; in other words, the inversion was pragmatic rather than yi behaving like a postposition.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

210 (10)




一言以蔽之。 (論語 為政) Yī yán yǐ bì zhī. one word with generalize it “(I) generalize it with one word.” (Lun Yu, Wei Zheng, 500 BC)

If we take the preposition yòng into account, the “lexical-replacement” hypothesis seems even more dubious. This preposition grammaticalized as early as the fifth century BC and could also introduce the instrument noun, the direct object for a ditransitive construction, and the like. Additionally, some instances of yòng in Old and Medieval Chinese could be translated into Contemporary Chinese using the disposal construction, as shown below: (11)




(a) 用其姊妻之。 (史記 為政) Yòng qí zǐ qī zhī. take his sister marry him “(He) took his sister to marry him.” (Shi Ji, Wei Zheng, 100 BC) (b) 把他的妹妹嫁給了他。 (現代漢語) Bǎ tā de mèimei jià-gěi-le tā. DISP he GEN younger-sister marry-to-PERF he “(The man) married him to his younger sister.” (Contemporary Chinese)

In the above situation, the prepositions yǐ and yòng are interchangeable, but their historical developments were quite different: yǐ gradually fell out of use in Medieval Chinese, but yòng is still being used as a common preposition in Contemporary Chinese and retains many of its usages from Old Chinese. The reason why they had such different developments is that Chinese grammar underwent a global change from Old to Medieval Chinese: all non-resultatives, including preposition phrases, no longer appeared in postverbal position (for details, see Chapter 7). As mentioned previously, the yǐ phrase could have such usage and thus fell out of fashion, though it was found in written language at a later stage. In contrast, the yòng phrase always appeared in preverbal position, which was in accordance with the global change of the grammar, and hence has survived up to the present day. From the point of view of grammaticalization theory, the “lexical-replacement” hypothesis seems least acceptable because every grammatical morpheme must undergo its own grammaticalization path in a certain context, unlike content words, which can simply be replaced by a newly coined word; for instance, the noun mù “eye” was

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Hypotheses about Its Origins

211

replaced by yǎn, a word that originally meant “eyeball” in Old Chinese. Actually, yǐ in Early Old Chinese meant “do” and then developed into a preposition; that is, it had its own development process. The disposal morphemes jiāng and bǎ, both content verbs at the time, must have undergone a grammaticalization process, independent of yǐ and of each other, to become markers for the disposal construction. In short, it is not possible that the disposal morphemes could have been created just to replace the former preposition yǐ for no good reason. Hypothesis 4. Further grammaticalization from the instrumental usage. Having observed that one of the early usages of both disposal morphemes bǎ and jiāng was to introduce an instrumental noun, as illustrated below, some researchers speculated that the disposal morphemes might have resulted from the further grammaticalization of their instrumental usages (Peyraube 1986, Wu 2003).




(12)

(奴)將帽塞口。 (顏之推 顏氏家訓) (Nú) jiāng mào sāi kǒu. I with hat stuff mouth “I stuffed my mouth with a hat.” (Yan Zhi Tui, Yan Shi Jia Xun, AD 550)

(13)

莫把金籠閉鸚鵡。 (蘇郁 鸚鵡詞) Mò bǎ jīn-lóng bì yīngwǔ. Do-not with golden-cage confine parrot “Don’t confine the parrot with the golden cage.” (Su Yu, Ying Wu Ci, AD 800)




Although the two disposal markers exemplified above both possessed instrumental usages, we cannot jump to the conclusion that the disposal function came from their instrumental usages, because there are two ways for a lexical item to evolve into a grammatical morpheme: (a) linear, where one function occurs after another, and (b) radial, where more than one function emerges from the same lexical source independently. It is cross-linguistically true that a single lexical item develops out of a grammatical morpheme with multiple functions, but the pathways fall into the following two groups, as illustrated by the grammaticalization paths involving demonstratives: (14)

(a) Linear, e.g. demonstrative > copula > focus (b) Radial, e.g. the verb “go” > future > distal demonstrative

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

212

Many factors need to be taken into consideration to establish possible links between different functions of the same grammatical form, in particular typological evidence and grammatical properties. For example, the demonstrative shì in Old Chinese evolved into a copula in Medieval Chinese and then developed a focus-marker function (for details, see Chapter 3). It has been widely attested across languages that a copula is often further grammaticalized as a focus marker (for more examples in other languages, see Heine and Kuteva 2002: 331). Therefore we believe that a linear pathway exists from demonstrative to copula to focus. In contrast, the verb xíng “go” grammaticalized into a future marker in Old Chinese, similar to be going to in English, and the verb zhī “go” developed into a distal demonstrative. Since there is no case of a distal demonstrative pronoun in any language developing out of a focus marker or vice versa, we believe that the verb “go” is equally suitable for future markers and distal demonstratives. According to the functions and diachronic processes of the disposal and instrumental morphemes, we suggest the following pathways: (15)

Radiant manner: the verb take > disposal > instrumental

That is, the two take verbs, jiāng and bǎ, have semantic suitability that enabled them to develop into a disposal or instrumental morpheme; in other words, both of the functions were directly grammaticalized from the verbs meaning “take.” This explanation is strongly supported by diachronic evidence; for instance, the disposal usage of jiāng came into existence as early as the fifth century AD, but its instrumental usage did not occur until the seventh century AD. Apparently it is impossible for the earlier disposal usage to have originated from the later instrumental usage. In addition, other researchers have provided hypotheses about the motivation for the emergence of the disposal construction from different perspectives. For example, Mei (1991) divided the disposal construction into three subtypes, claiming that each of them had different lexical sources. Sun (1996) argued that the disposal construction emerged from the purposive clause marked by lái “come” or qù “go” because they often cooccurred. Feng (2014) assumed that the prosodic structure change in clauses was responsible for the production of the disposal construction. Because these ideas were not supported by any diachronic evidence and hence have been accepted by few scholars, we make no further comments here.

9.3

Functions of the Disposal Construction

An accurate characterization of the functions of the disposal construction is undoubtedly helpful in studies on its diachronic development. Identifying the unique grammatical

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Functions of the Disposal Construction

213

properties of the disposal construction can provide us with a clue to why and where it was first introduced into the language. In this section, we briefly review representative views on the semantics and syntax of the disposal construction and then put forward our analysis, which will be applied to the diachronic investigation in the following sections (cf. Section 7.6.1). The syntactic schema of the disposal construction can be described as follows, with bǎ representing all the disposal markers: (16)

The syntactic schema of the disposal construction2 Subj + (bǎ + NP) + VP.

The subject is optional, but the remaining three elements are essential to make a disposal sentence well formed. In the construction, the patient NP stands out through being marked by bǎ, where the VP constitutes the background. In our view, the semantic feature of the patient noun and, especially, the grammatical relation between the object of bǎ and the structure of the predicate are critical in identifying the status of the disposal construction in the grammatical system of Chinese. First, some terminological issues need to be clarified. In the literature, the disposal markers are labeled by various terms, including “preposition,” “coverb,” “light verb,” and even “verb” (Chao 1979: 330‒338, Li and Thompson 1981: 356‒369, Huang et al. 2009: 153‒196, Shi 2010: 98‒108). These different labels do not necessarily reflect different views on the grammatical properties of the disposal morphemes but may simply indicate the habitual usages of individual researchers or conventions of different theoretical frameworks. However, there are indeed differences in degrees of grammaticalization from a content verb to a preposition in regard to the disposal markings. Roughly, all the major disposal markers that were used before the tenth century AD, including jiāng, bǎ, qǔ, chí, and zhuō, were actually content verbs, and some of them, such as jiāng and bǎ, had probably started to be grammaticalized at the time. However, the others remained ordinary verbs and were only occasionally used in contexts that were semantically similar to a disposal instance. Around the tenth century AD, the disposal markers jiāng and bǎ became fully grammaticalized and were already pure prepositions; meanwhile, both of them lost the major properties of verbs. For instance, they could not be used as the matrix verb, they could not be reduplicated, and they could not be suffixed with aspect markers. The verb ná “take” developed out of a disposal marker only in the eighteenth century AD, and it is still being used mainly as a regular verb in Contemporary Chinese. Similarly, the verb guǎn “control” developed use as a disposal marker only in the past century, and still behaves like an ordinary verb in most other situations. 2

We use bǎ as a representative marker for the disposal construction here. Note that there are other markers for the same function, diachronically and synchronically.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

214

As Wang (1943: 82) put it, if the disposal construction was just a simple inversion between the verb and the object, with the same function as an SVO structure or other structures, there should be no need for it to be innovated and to develop as a major apparatus in the grammar. Like all other constructions, the disposal construction exhibits a wide range of semantic and syntactic properties, from marginal to essential. To discover the motivation for its emergence in history, we must accurately capture the essential characteristics of the disposal construction that successfully distinguish it from other constructions. As mentioned previously, the term “disposal” is somewhat misleading because many instances under this label do not express any “disposal” meaning, and, more crucially, this “disposal” meaning can also be expressed by means of many other declarative SVO sentences. In other words, the term “disposal” cannot successfully distinguish the construction from other constructions. Similarly, there are many other analyses on the functions of the disposal construction. For instance, it increases the transitivity of the clause (Chao 1979: 330‒338, Li and Thompson 1981: 463‒491, Sun 1996: 52‒81), highlights the affectedness of the object (Huang 2006), or indicates a bounded event. However, all these proposals also fail to characterize the semantic and syntactic properties of the disposal construction because all the functions mentioned here can be expressed by an ordinary VO construction, as illustrated below: (17)

他打碎了那個杯子。 (現代漢語) Tā dǎ-suì-le nà-gè bēizi. he beat-broken-PERF that-CL glass “He broke the glass.” (Contemporary Chinese)

(18)

他把北京城走了一大半。(現代漢語) Tā bǎ Běijīng chéng zǒu-le yī dàbàn. he DISP Beijing city walk-PERF one more-than-half “He walked more than half of the roads of Beijing.” (Contemporary Chinese)

In (17), the object nàgè bēizi “the glass” is definite and referential, and the predicate is a VR phrase, dǎ-suì “beat-broken,” expressing the maximal degree of transitivity and affectedness. Additionally, the whole event is temporally bounded. More importantly, the degrees of transitivity and affectedness are not necessary conditions for using the disposal construction; for example, in (18), the object of bǎ is Beijing city, and the main verb zǒu “walk” is intransitive. Thus it is difficult to say that the sentence expresses any meaning of a maximal degree of transitivity and affectedness. Regarding boundedness,

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Functions of the Disposal Construction

215

an overwhelming majority of sentences, regardless of whether they are active or passive, are generally bounded by adding aspect markers, resultatives, quantifiers, and so forth (for details, see Chapter 15). Specifically, the predicate of the passive structure must also be bounded, exactly like the disposal construction. Clearly, the feature of boundedness also fails to distinguish the disposal constructions from others. Moreover, disposal markers are labeled with various names, such as “preverbal transitive morpheme” or “preobject marker” (Chao 1979: 342). These labels are merely a superficial observation rather than a serious analysis. Chinese, as an SVO language, exploits word order to signal important grammatical relations; for instance, the subject precedes the verb, and the object follows the verb. Why does the speaker bother to move an object to preverbal position and to mark it with an extra linguistic form (e.g. bǎ)? In fact, many patient nouns in preverbal position do not need any marker. As Zhu (1982: 188‒189) pointed out, if the subject is absent, all the disposal markers can be removed from the sentences, turning them into “unmarked” passive sentences, as illustrated below: (19)

(a) 把衣服都洗乾淨了。 (現代漢語) Bǎ yīfú dōu xǐ-gānjìng le. DISP clothes all wash-clean PERF “(He) has washed all clothes clean.” (b) 衣服都洗乾淨了。 (現代漢語) Yīfú dōu xǐ-gānjìng le. clothes all wash-clean PERF “All clothes have been washed clean.” (Contemporary Chinese)

9.3.1

Definiteness of the Patient Noun

It has been generally accepted that the NP in the disposal construction must be definite (Lü 1984, Ding et al. 1961: 95‒97, Li and Thompson 1981: 463‒491, Zhu 1982: 185‒ 188). Whether it is a rigorous rule or just a tendency depends on what empirical data the researchers observe and how they handle seemingly complex phenomena of disposal instances. For example, Zhu (1982: 186) claimed that it is a rule, and that violating it will create an ungrammatical sentence. However, Li and Thompson (1981: 170) just thought of it as a tendency: “Nouns preceding the verb tend to be definite, while those following the verb tend to be indefinite.” To discover the truth, we must separate two things: (a) the default function of the disposal construction and (b) the conditioned usages of the disposal construction.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

216

First, let us examine the default function of the disposal construction. If the object of bǎ is a bare noun, namely without any modifiers such as demonstratives or numerals, and if this patient noun can occur in either preverbal or postverbal position (both well formed), then the disposal construction automatically assigns the feature definite to the preverbal NP and the feature indefinite to the postverbal NP, a rigid rule in Chinese grammar (cf. Section 7.6.1), as shown below: (20)

(a) 他把面包喫了。(現代漢語) Tā bǎ miànbāo chī-le. he DISP bread eat-PERF “He ate the bread.” (b) 他喫了面包。(現代漢語) Tā chī-le miànbāo. he eat-PERF bread “He ate some bread.” (Contemporary Chinese)

(21)

(a) 他把書還了。 (現代漢語) Tā bǎ shū hái-le. he DISP book return-PERF “He returned the book.” (b) 他還了書。 (現代漢語) Tā hái-le shū. he return-PERF book “He returned a book.” (Contemporary Chinese)

Now, let us consider the conditioned usage of the disposal construction. In fact, the object of bǎ is not necessarily definite and can be indefinite or generic. In this case, however, two conditions must be satisfied: first, the patient noun must be modified by a numeral phrase or the like, and second, the predicate is structurally complex and disallows a following object, as illustrated in the following examples: (22)

他把金錢看作糞土。 (現代漢語) Tā bǎ jīnqián kàn-zuò fèntǔ. he DIS money view-as manure “He views money as manure.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Functions of the Disposal Construction (23)

217

馬銳把一瓶豆腐摔碎在地上。 (現代漢語) Mǎ Ruì bǎ yī-píng dòufu shuāi-suì zài dìshàng. Ma Rui DISP one-CL tofu throw-smash on floor “Ma Rui smashed a bottle of tofu on the floor.” (Zhang 2001, Contemporary Chinese novel)

In (22), jīnqián “money” is a generic noun, which cannot occur after the verb because the predicate consisting of a “VO” structure blocks any following object. In (23), the predicate is a “V + PP” structure that disallows any following object; therefore the indefinite object must occur in preverbal position. Note that if the “Num + CL” modifier is deleted in (23), the NP will be automatically assigned the feature definite, as illustrated in (24): (24)

馬銳把豆腐摔碎在地上。 (現代漢語) Mǎ ruì bǎ dòufu shuāi-suì zài dìshàng. Ma Rui DISP tofu throw-smash on floor “Ma Rui smashed the bottle of tofu on the floor.” (Contemporary Chinese)

If the predicate is simple and can be followed by an indefinite object, the NP in the disposal construction cannot be indefinite; otherwise, the construction will become illformed, as illustrated below: (25)

(a) 他請來了一位大夫。 (現代漢語) Tā qǐng lái-le yī-wèi dàifū. he invite come-PERF one-CL doctor “They invited a doctor to come.” (b) *他把一位大夫請來了。 *Tā bǎ yī-wèi dàifū qǐng lái-le. he DISP one-CL doctor invite come-PERF (Zhu 1982: 187, Contemporary Chinese)

(26)

(a) 他買了很多書。 (現代漢語) Tā mǎi-le hěn duō shū. he buy-PERF very many book “He bought many books.” (b) *把很多書買了。 *Tā bǎ hěn duō shū mǎi-le. he DISP very many book buy-PERF (Lü 1999: 52, Contemporary Chinese)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

218

There is an interaction between lexical marking and feature assignment by syntactic position. This pair of grammatical categories, “definite” and “indefinite,” can be expressed either by lexical items (e.g. yī gè “one CL”) or syntactic position. In normal circumstances, the NP in the object position of the preposition bǎ must be definite unless the predicate structurally disallows any following object, and the patient noun must occur in preverbal position. The default function of the disposal construction is to assign a definite feature to the NP by means of syntactic position. A definite feature can be expressed by adding demonstratives or other pronominal items, forming a definite noun phrase that is favored by the disposal construction. In Contemporary Chinese, there are two types of ditransitive construction, as formulated in (27): (27)

(a) Subj V + Oi + Od (b) Subj (bǎ Od) + V + Oi

There is a rigorous division of labor between the above two forms: if the direct object is indefinite, the construction in (27a) is chosen; if it is definite, the disposal construction depicted in (27b) must be used. Otherwise, an ungrammatical sentence will be created, as illustrated below: (28)

(a) 他把那本書送了王教授。 (現代漢語) Tā bǎ nà-běn shū song-le Wáng Jiàoshòu. he DISP that-CL book send-PERF Wang Professor “He has sent the book to Professor Wang.” (b) *他送了王教授那本書。 *Tā song-le Wáng he send-PERF Wang

Jiàoshòu nà-běn shū. Jiaoshou that-CL book (Contemporary Chinese)

Some researchers have assumed that every disposal sentence has a counterpart with a VO order (e.g. Huang et al. 2009: 153‒156), and they have even claimed that the disposal construction is a simple inversion of the verb and the object. This point of view fails to capture the essential function of the disposal construction. In fact, many disposal instances do not have a corresponding VO expression, as shown by the complex predicate examples in (23); even if both are equally grammatical, there is always a key semantic difference between the disposal and VO instances: the NP in the disposal instance must be definite, and the NP in the corresponding VO construction must be indefinite. The disposal construction’s need to assign the feature definite is so strong that it can override the default

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Functions of the Disposal Construction

219

“indefinite” interpretation of a “Num + CL” phrase by adding a definite reading to it, an issue to which we will return below. When used in the disposal construction, a classifier actually serves to strengthen the degree of definiteness. As Cheng and Sybesma (1999) pointed out, classifiers have the functions of singularization and individualization. According to Langacker (1991: 308), referring to a single member represents the highest degree of definiteness; for instance, a proper noun is more definite than a common noun. Interestingly, proper nouns in the disposal construction can also be accompanied by the general classifier gè, as illustrated in the following examples: (29)

把個英蓮拖去。 (紅樓夢四回) Bǎ gè Yīng Lián tuō-qù. DISP CL Ying Lian drag-away “(He) dragged Ying Lian away.” (Hong Lou Meng, Chapter 4, AD 1750)

(30)

把個小王聽得入了迷。 (現代漢語八百詞) Bǎ gè Xiǎo Wáng tīng-de rù-le mí. DISP CL Xiao Wang listen-PRT enter-PERF fascination “Xiao Wang is fascinated by the music.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 52)

In addition, unique definiteness, such as the people’s names in the above examples, can be strengthened by adding the classifier gè when the nouns are used in the disposal construction. This finding undermines the claim made by Jenks (2018) that unique definiteness in Mandarin can be realized only as a bare noun. In the following examples, lǎo fùqīn “old father” and rìtóu “sun” are uniquely definite but are modified by the classifier gè: (31)

把個老父親送到北京。 (北京話相聲) Bǎ gè lǎo fùqīn song-dào Běijīng. DISP CL old father send-to Beijing “He sent his old father to Beijing.” (Cross Talks in Pekingese)

(32)

把個日頭擋得嚴嚴實實。(中國日報) Bǎ gè rìtóu dǎng-dé yán-yán-shí-shí. DISP CL sun cover-PRT complete “The birds cover the sun completely.” (China Daily)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

220

Common nouns can also be modified by the general classifier gè when appearing in a disposal construction. In this case, the classifier gè functions like the demonstrative pronoun zhè “this” to emphasize definiteness, as illustrated in (33) and (34):




(33)

把個女兒慣得一點樣兒都沒有。 (曹禺 雷雨) Bǎ gè nǚér guàn-dé yī-diǎn yàngr dōu méiyǒu. DISP CL daughter spoil-PRT a bite shape all lack “You have completely spoiled your daughter.” (Cao Yu, Lei Yu, Contemporary Chinese)

(34)

把個屋子打扮的花里胡哨。 (中國日報) Bǎ gè wūzǐ dǎbàn-dé huā-li-hú-shào. DISP CL room decorate-PRT colorfully “They have decorated the room colorfully.” (China Daily)

The classifier gè was already grammaticalized as a fully fledged demonstrative pronoun in the period from the fifth century AD to the sixteenth, and it is still currently widely used as a demonstrative in many southern dialects (for details, see Chapter 17). Across Chinese dialects, disposal morphemes are without exception used to introduce a definite nominal phrase. In line with our analysis above, it is easy to explain a seemingly unusual phenomenon in the Zunyi dialect, spoken in an area of Guizhou Province, where all nominal phrases in the disposal construction can be followed by a classifier that matches the preceding noun (Huang 1996: 132), as illustrated in (35) and (36): (35)

把你的鋼筆支借一下。 (遵義方言) Bǎ nǐ-de gāngbǐ zhī jiè yīxià. DISP you-GEN pen CL borrow a-while “Please lend me your pen for a while.” (Huang 1996: 132, Zunyi dialect)

(36)

牛把麥子個喫了。 (遵義方言) Niú bǎ màizǐ gè chī-le. ox BA wheat CL eat-PERF “The ox has eaten the wheat.” (Huang 1996: 132, Zunyi dialect)

The classifier class as a whole has been grammaticalized as a kind of suffix to strengthen the degree of definiteness of the disposal argument.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Functions of the Disposal Construction 9.3.2

221

The Definiteness of “Num + CL + N” Phrases

“Num + CL + N” phrases can also be interpreted as definite in the disposal construction, but all researchers in the area have predicted that this interpretation is impossible. For instance, on the basis of their observation that “CL + N” instances in Cantonese can be interpreted as definite but “Num + CL + N” instances cannot, Cheng and Sybesma (1999) proposed that nominal phrases with a numeral cannot be interpreted as definite in any environment. This statement might be true in Cantonese, but it is by no means universal in Mandarin Chinese and its dialects. First, let us examine how the most authoritative grammar book described this usage: (37)

他把兩本書都看完了。 (現代漢語八百詞) Tā bǎ liǎng-běn shū dōu kàn-wán le. he BA two-CL book all read-finish PERF “He has read the two books.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 53)

In parentheses immediately following the above examples, the authors indicated that “there is a prior mention of which two books are in context.” It may be argued that the definiteness of the “Num + CL + N” in (36) is licensed by the adverb dōu “all” (Cheng and Sybesma 1999). However, this is not the case. We collected a total of 1,742 examples of this type from the Peking University Corpus and found that, in the disposal construction, the overwhelming majority of “Num + CL + N” phrases interpreted as definite do not need to be licensed by any particular operator, such as dōu. This finding shows that their definiteness is assigned by the disposal construction rather than by any particular “operator,” such as numeral phrases. Considering that such usage has largely been ignored in the literature, we would like to illustrate this point by providing more examples, all of which are naturally occurring instances:

3

(38)

我們應該把兩個條約合併起來研究。 (現代中國史) Wǒmen yīnggāi bǎ liǎng-gè tiáoyuē hébìng-qǐlái yánjiū.3 we should DISP two-CL contract combine-up research “We should combine the two contracts and study them together.” (Xian Dai Zhong Guo Shi, Contemporary Chinese)

(39)

傑克把四個女兒帶走了。 (讀者) Jiékè bǎ sì-gè nǚér dài-zǒu le. Jack DISP four-CL daughter take-away PERF “Jack took away his four daughters.” (Du Zhe, Contemporary Chinese)

We have modified these examples by omitting irrelevant parts.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

222 (40)

壯士把六個人拖出雪坑。 (林海雪原) Zhuàngshì bǎ liù-gè rén tuō-chū xuěkēng. soldier DISP six-CL people pull-out snow-pit “The soldiers pulled the six people out of a snow pit.” (Lin Hai Xue Yuan, Contemporary Chinese)

(41)

我把五個餃子喫啦。 (傳統相聲選) Wǒ bǎ wǔ-gè jiǎozǐ chī-lā. I DISP five-CL dumpling eat-PERF “I have eaten the five dumplings.” (Collections of Traditional Cross Talks, Contemporary Chinese)

All the above examples represent anaphoric or familiarity definiteness in the terms of Jenks (2018). Thus there is always a prior mention in an earlier context. In (38), for instance, the two contracts – the “Contract between China and England” and the “Maguan Contract” – were specifically mentioned in the previous context. Likewise, the referent of wǔ-gè jiǎozǐ “five dumplings” in (40) is definite and had to be mentioned previously in context. In contrast, the referent is indefinite if this “Num + CL + N” phrase occurs after the verb chī “eat.” To summarize, the disposal construction has the greatest power in assigning a definiteness feature to nominal phrases. All three types of noun phrase, bare nouns, “CL + N” phrases, and “Num + CL + N” phrases, can be interpreted as definite in this construction. Here, even unique definiteness can be realized as a “CL + N” phrase. The degree of definiteness of proper nouns can be strengthened by adding a classifier to them. Interestingly, in the Zunyi dialect illustrated above, all noun phrases in the disposal construction can be suffixed with an appropriate classifier to emphasize definiteness through singularization and individualization. 9.3.3

Obligatory Application

In the section above, we saw that the disposal construction must be used to introduce the patient noun when the predicate is structurally complex. We focus on this issue in this section: in what situation does the disposal construction have to be used? This issue is key to identifying both the status of the disposal construction and the motivation for its emergence in history. When a disposal instance has a corresponding VO expression, as discussed above, their functions always differ critically: the NP in the disposal construction must be interpreted as definite and that in the VO construction can be understood only as indefinite. More importantly, in many situations, the disposal construction is only a grammatical

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Functions of the Disposal Construction

223

device for expressing the meaning, mostly because of the structural complexity of the predicate (Lü 1984, Liu et al. 2001: 731‒752). Of the thirteen types of predicate used for disposal events, twelve of them disallow the appearance of an NP in postverbal position. Liu et al. (2001) observed that in some situations, the bǎ construction must be used, but in other situations it may be optional. In addition, Liu et al. (2001: 731‒752) listed eight types of predicate in which the disposal construction must be used. Here, we discuss only those types of predicate that are most relevant to our diachronic investigation. (a)

(42)

The structural complexity of the predicate. When the predicate consists of a verb and a resultative locative preposition phrase, the patient noun cannot follow the verb and must be introduced in preverbal position by the disposal morpheme, as illustrated in (42): 他把雞蛋放在桌子上。 (現代漢語) Tā bǎ jīdàn fàng zài zhuōzi shàng. he DISP egg put on table above “He put the eggs on the table.”

(b)

The definite direct object of the ditransitive construction. In a ditransitive expression, if the direct object is definite, it can no longer be used in the unmarked construction “V Oi Od,” as mentioned previously; in this case, it is most properly expressed by the disposal construction, as illustrated in (43):

(43)

我把那些書還給圖書館。 (現代漢語) Wǒ bǎ nàxiē shū hái-gěi-le túshū-guǎn. I DISP those books return-to-PERF library “I have returned those books to the library.” (Contemporary Chinese)

In this case, there is an option of using topicalization to introduce the definite direct object, as illustrated in (44), but the pragmatic value is different to that of the disposal construction. The former treats it as the topic of communication; the latter emphasizes it as an activity to cause some effect on the patient noun. (44)

那些書我還給了圖書館。 (現代漢語) Nàxiē shū wǒ hái-gěi-le túshū-guǎn. Those book I return-to-PERF library “Those books have been returned to the library by me.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

224

(c) The equational construction. There is a set of verbs that express the equality of two things in a certain sense, such as jiào “call (name),” dāng “treat,” and V- chéng “act-be.” Of the two nouns involved, only the resultative one can follow the verb, and the other is usually introduced by a disposal marker, either bǎ or guǎn, as illustrated below: (45)

他把王老師當成父親。 (現代漢語) Tā bǎ wáng lǎoshī dāngchéng fùqīn. he DISP Wang Laoshi treat-as father “He treats Teacher Wang as his father.” (Contemporary Chinese)

(46)

他把頭髮染成了藍色。 (現代漢語) Tā bǎ tóufǎ rǎn-chéng-le lán-sè. he DISP hair dye-be-PERF blue “He dyed his hair blue.” (Contemporary Chinese)

(47)

他們管土豆叫山藥蛋。 (現代漢語) Tāmen guǎn tǔdòu jiào shānyào dàn. they DISP potato call Chinese-yam egg “They call potato the egg of the Chinese yam.” (Contemporary Chinese)

Sometimes, the relationship of the two things involved is quite indirect, with one being made out of the other or one being about the other. In this case, the disposal structure does not have a corresponding VO expression, as illustrated below: (48)

他把這件事寫了一篇報導。 (現代漢語) Tā bǎ zhè-jiàn shì xiěle yī-piān bàodào. he DISP this-CL event write-PERF one-CL report “He wrote a news report about the event.” (Contemporary Chinese)

(49)

把衣服改了個樣子。 (現代漢語八百詞) Bǎ yīfú gǎi-le gè yàngzi. DISP clothes change-PERF a shape “(He) changed the clothes into another shape.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 54)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Functions of the Disposal Construction (d)

(50)

225

Verb compounds with an internal VO structure. Mainly due to the disyllabification tendency, there are many disyllabic verbs with the internal structure “V + O” in Contemporary Chinese. Most of them cannot be followed by any object in surface construction. In this case, the disposal construction becomes necessary when the patient noun needs to be indicated, as illustrated in (50) and (51): 他把門上了鎖。 (現代漢語) Tā bǎ mén shàng-le suǒ. he DISP door close-PERF lock “He locked the door.” (Contemporary Chinese)

(51)

他把橘子剝了皮。 (現代漢語) Tā bǎ júzi bō-le pí. he DISP orange peel-PERF skin “He peeled the oranges.” (Contemporary Chinese)

In the above examples, the English simple verbs correspond to the VO compounds in Chinese: lock ~ shàng-suǒ (lit. add + lock) and peel ~ bō-pí (lit. peel + skin). In Old Chinese, these meanings were also conceptualized by simple verbs; for instance, in the Shi Jing, only a simple verb bō was used to mean “peel,” and there was no compounding with the nominal morpheme pí “skin.” From Medieval Chinese to the present day, an increasing number of compound verbs of this type have been created (Lü 1984). (e)

(52)

The VO construction. In addition to the ditransitive construction that involves “transfer” verbs, either give or receive, ordinary verbs can also have two patient objects, one of which is usually a quantifier (a temporary classifier or one holding a part–whole relationship with another patient noun).4 In this case, only the “quantifier” object can follow the verb, and the other one needs to be introduced by the disposal construction, as illustrated in (52) and (53): 把馬打了兩鞭子。 (現代漢語) Bǎ mǎ dǎ-le liǎng biānzi. DISP horse beat-PERF two whip “(He) whipped the horse twice.” (Contemporary Chinese)

4

Chapter 12 deals with the development of the ditransitive construction.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

226 (53)

他把桌子踢斷了兩條腿。 (現代漢語) Tā bǎ zhuōzi tī-duàn-le liǎng-tiáo tuǐ. he DISP table kick-broken-PERF two-CL leg “He broke (by kicking) two legs of the table.” (Contemporary Chinese)

(f) The structural complexity of resultatives. As we have seen in Chapter 6, some sorts of patient object must appear in preverbal position because of the fusion of the verb and the resultative. In particular, when the resultative is structurally complex, which is usually marked by the resultative particle de, the patient noun cannot occur in postverbal position and has to be introduced by the disposal morpheme, as illustrated below: (54)

他把錢看得太重。 (現代漢語) Tā bǎ qián kàn de tài zhòng. he DISP money look RES too important “He puts too much emphasis on money.” (Contemporary Chinese)

(55)

他把話說得很漂亮。 (現代漢語) Tā bǎ huà shuō dé hěn piàoliang. he DISP word speak RES very beautiful “He promised very beautifully.” (Contemporary Chinese)

The above analysis shows that the disposal construction plays a key role in the grammar, far beyond being just a variant of the VO construction. Additionally, it cannot be merely an optional device for expressing some pragmatic function because it has to be used in many situations, as discussed above. For instance, Tsao (1990) argued that the object of bǎ is the secondary topic, but if so the following question arises: given that the initial position of a sentence in Chinese has been reserved for the topic throughout history, why does Chinese need another “secondary topic” after the subject? In fact, the grammar does not allow both the supposed primary topic and the so-called secondary topic to co-occur within a single sentence. According to Shao and Zhao (2005), the disposal morpheme may be a focus marker for highlighting the patient noun, but this suggestion cannot be true because the focus is to highlight the most important new information in a sentence, whereas the patient noun generally refers to given information. Furthermore, Light (1979) argued that OV is a marked, emphatic, contrastive construction in Mandarin Chinese, and Sun and Givón (1985) similarly proposed the idea that OV functions as an emphatic/contrastive discourse device.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Two Historical Motivations

227

These suggestions are actually based on the assumption that the OV construction must have a corresponding VO one; as we have indicated above, however, many disposal instances do not have VO correspondences. One of the properties of the disposal construction needs to be highlighted here: in some situations, the patient noun is the object not of the verb but of the whole predicate consisting of the verb and the resultative, an important fact that calls into question the widely held assumption that the NP is the object of only the verb. In example (56), the verb xiào “laugh” and the patient noun yǎnlèi “tear” do not bear a verb–object relationship. From the point of view of construction grammar, the whole function of the “verb + resultative” phrase can license an extra argument (Goldberg 1995: 180‒198). (56)

她把眼淚笑出來了。 (現代漢語) Tā bǎ yǎnlèi xiào-chūlái-le. she DISP tear laugh-out-PERF “She laughed till she cried.” (Contemporary Chinese)

It needs to be emphasized that among all prepositions in preverbal position, only the disposal morpheme can function to assign the feature definite to the preverbal NP; other prepositions cannot do so even in preverbal position. For example, the passive bèi, jiào, and ràng; the locative zài; and the comparative bǐ are all incapable of assigning the feature definite to their objects. Additionally, the NP in the disposal construction can be modified by demonstratives or personal pronouns, whose inherent meanings are definite. In general, the disposal construction has two essential characteristics: (a) assigning the feature definite to the patient noun and (b) obligatorily introducing an object because of the structural complexity of the predicate. These two characteristics can not only successfully distinguish the disposal construction from other grammatical devices, but also reveal the motivation for its emergence in history.

9.4

Two Historical Motivations

The evolution of grammar constitutes a chain of changes, where one change causes another one, and the latter change may serve as a trigger for further development. Few grammatical changes happened in the history of Chinese in isolation or independently, and none resulted purely from pragmatic inferences. Since a language’s grammar at any period represents an integrated system, every new device added to the system

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

228

must be licensed by the overall properties of the grammatical system of the time. For the disposal construction, a legitimate question needs to be answered: why did this construction come into existence in Late Medieval Chinese rather than earlier or later? Why did it develop from occasional to robust usage and become a major and frequently used device? The answers reside in two motivations. First, a principle was formed at the time that automatically assigned the feature definite to a noun in preverbal position. Second, the verb and the resultative became fused into an immediate constituent by erasing the original syntactic position between them. As a result, the object that originally intervened between the verb and the resultative had to move elsewhere in the sentence. These two conditions are illustrated below. As discussed above, in a sentence with an action verb, a bare noun in preverbal position is automatically assigned the feature definite and that in postverbal position is automatically assigned the feature indefinite, as illustrated below (cf. Section 7.6.1): (57)

(a) 人來了。(現代漢語) Rén lái-le. Person come-PERF “The person has come.” (Contemporary Chinese) (b) 來人了。 (現代漢語) Lái rén le. come person PERF “A person has come.” (Contemporary Chinese)

However, the above principle did not exist before the fifth century AD, as illustrated in (58) and (59), where the two nouns in subject position are still interpreted as indefinite.




(58)

人無遠慮, 必有近憂。 (論語 衛靈公) Rén wú yuan-lǜ, bì yǒu jìn yōu. one lack foresight must have near worry “If one lacks foresight, he must have near worry.” (Lun Yu, Wei Ling Gong, 500 BC)

(59)

友人有疾。 (世說新語 德行) Yǒu-rén yǒu jí. friend have ill “One of my friends was sick.”




(Shi Shuo Xin Yu, De Xing, AD 450)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Two Historical Motivations

229

According to Sun (1997), the above principle was established around the sixth century AD. The evidence is that the new topicalization construction, where a bare noun was fronted to the beginning of a sentence and interpreted as definite, was formed then, as illustrated in (60) and (61). Meanwhile, an anaphor in the original object position was no longer needed, distinguishing it from the topicalization pattern in Old Chinese (for details, see Section 3.4.3). (60)

錢財奴婢用。 (王梵志詩) Qián cái núbì yòng. money treasure servant use “His money and treasures were used by the servants.” (Wang Fan Zhi Shi, AD 650)

(61)

茶錢酒家自還你。 (水滸傳三回) Chá qián jiǔjiā zì huán nǐ. tea money I sure repay you “As for the tea expense, I will surely repay you.” (Shui Hu Zhuan, Chapter 3, AD 1400)

The emergence of the disposal construction was a response to the fusion of the verb and the resultative, which was specialized for repositioning the definite patient nouns that originally occurred between the verb and the resultative. At the very beginning, all the patient nouns in the disposal construction had to be interpreted as definite (Wang 1989: 266‒271). There were two variants of these nouns: first, the nouns were modified by a demonstrative or a definite element, as illustrated in (62) and (63); second, the nouns had no modifier but had to be interpreted as definite, as in (64) and (65): (62)




誰將此義陳? (杜甫 寄李十二白) Shéi jiāng cǐ yì chén? who DISP this meaning state “Who will state the meaning?” (Du Fu, Ji Li Shi Er Bai, AD 750)

(63)




火急將吾錫杖與? (敦煌變文 大目干連) Huǒjí jiāng wú xī-zhàng yǔ? immediate DISP my tin-stick give “Please immediately give my monk’s cane (to him).” (Dun Huang Bian Wen, Da Mu Gan Lian, AD 800–1000)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

230




(64)

將詩莫浪傳。 (杜甫 泛江送魏十八) Jiāng shī mò làng chuán. DISP poem not wildly circulate “Don’t circulate my poems wildly.” (Du Fu, Fan Jiang Song Wei Shi Ba, AD 750)

(65)

只把空書寄故鄉。 (王建 維揚冬末) Zhǐ bǎ kōng-shū jì gùxiāng. only DISP blank-letter mail hometown “(I) only mailed the blank letter to my hometown.” (Wang Jian, Wei Yang Dong Mo, AD 900)




The process of the formation of the resultative construction was essentially the fusion of the verb and the resultative (for details, see Chapter 6). Before this fusion, if the resultative was an intransitive verb, the definite patient object usually occurred between the verb and the resultative, as illustrated in (66a). After the fusion, however, the most appropriate form for expressing the same meaning was the disposal construction (cf. Sections 5.7, 6.3.1), as illustrated in (66b): (66)




(a) 喚江郎覺。 (世說新語 假譎) Huàn Jiāng Láng jué. call Jiang Lang awake “Please wake up Jiang Lang.” (Shi Shuo Xin Yu, Jia Jue, AD 450) (b) 把江郎叫醒。 (現代漢語) Bǎ Jiāng Láng jiào-xǐng. DISP Jiang Lang call-awake “Please wake up Jiang Lang.” (Contemporary Chinese)

In contrast, if the object was indefinite, the corresponding form in Contemporary Chinese is the VRO structure, where the indefinite patient is the object of the whole VR phrase, as illustrated below: (67)




(a) 周仲智飲酒醉。 (世說新語 雅量) Zhōu Zhòngzhì yǐn jiǔ zuì. Zhou Zhongzhi drink wine drunken “Zhou Zhongzhi drank wine and became drunk.” (Shi Shuo Xin Yu, Ya Liang, AD 450)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Two Historical Motivations

231

(b) 周仲智喝醉了酒。 (現代漢語) Zhōu Zhòngzhì hē-zuì-le jiǔ. Zhou Zhongzhi drink-drunken-PERF wine “Zhou Zhongzhi drank wine and became drunk.” (Contemporary Chinese) When the resultative was a preposition phrase to indicate the end point of the verb, before the fifth century AD the patient noun occurred between the verb and the resultative, as illustrated in (68) and (69). Once the verb and the resultative became fused, the patient noun could occur only somewhere prior to the VR predicate. One type of the earliest disposal examples had the predicate with this VR structure, as illustrated below:




(68)

埋玉樹著土中。 (世說新語 傷逝) Mái yù-shù zháo tǔ-zhōng. bury jade-tree in soil-inside “(He) buried the jade tree in the soil.” (Shi Shuo Xin Yu, Shang Shi, AD 450)

(69)

把舜子頭髮懸在中庭樹地。 (敦煌變文 舜子變) Bǎ Shùn Zi tóufǎ xuán zài zhōngtíng shùdì. DISP Shun Zi hair hang on center-yard tree “Shun Zi’s hairs were hung upon the tree in the central yard.” (Dun Huang Bian Wen, Shun Zi Bian, AD 800–1000)




From its emergence to the present day, the disposal construction has never been a simple inversion of a corresponding VO sentence, as we discussed above. Roughly speaking, there are two situations in which the disposal construction is used: first, a bare patient noun can be used in either the disposal or a VO construction, both of which are grammatical, but their meanings are different. Specifically, in the disposal construction the noun must be interpreted as definite, and in VO construction it must be indefinite. Second, the bare noun can occur only somewhere prior to the predicate because of the structural complexity of the predicate. These early characteristics of the disposal construction reveal the motivation for its emergence. We have seen some examples in the above analysis. Let us examine the syntactic characteristics of the disposal construction as follows. When the predicate was a simple transitive verb, the patient noun could occur either in the disposal construction or in an active structure, which produced different interpretations. It had to be interpreted as definite in the disposal construction, as illustrated below.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

232 (70)

官員將玉觀音反覆看了。 (碾玉觀音) Guānyuán jiāng yù-guānyīn fǎnfù kàn-le. official DISP jade-Guanyin repeatedly look-PERF “The officials looked at the jade Guanyin repeatedly.” (Nian Yu Guan Yin, AD 1100)

At this time, the ditransitive construction had already been divided into two subtypes: if the direct object was indefinite, it adopted the traditional unmarked structure “Subj V Oi Od”; if the direct object was definite, the traditional structure could not be used, and the disposal was the best choice to express the meaning, as illustrated below:




(71)

和尚莫將境示人。 (祖堂集 雪峰和尚) Héshàng mò jiāng jìng shì rén. monk not DISP mirror show people “The monk did not show the mirror to others.” (Zu Tang Ji, Xue Feng He Shang, AD 950)

(72)

卻將家信託袁師。 (呂溫 临洮送袁七) Què jiāng jiā-xìn tuō Yuán Shī. but DISP family-letter consign Yuan Shi “But (I) consigned my family letter to Yuan Shi.” (Lu Wen, Lin Yao Song Yuan Qi, AD 800)




If the matrix verb was followed by another nominal item, which indicated a quantifier, a resultative object, or the parts of the patient noun, the patient noun could no longer follow the predicate and had to be moved elsewhere (Lü 1984: 176‒199), as discussed in the above section. The disposal construction was the best choice if the patient noun was definite, as illustrated in the following examples: (73)

The postverbal nominal is a quantifier. 把郭立打了五十背花棒。 (碾玉觀音) Bǎ Guō Lì dǎle wǔshí bèi-huā-bàng. DIP Guo Li beat-PERF fifty back-broken-stick “(They) beat Guo Li with a strong stick fifty times.” (Nian Yu Guan Yin, AD 1100)

(74)

The postverbal nominal is part of the preceding patient noun. 把妮子縛了兩隻手。 (簡帖和尚) Bǎ nízi fùle liǎng-zhī shǒu. DISP girl tie-PERF two-CL hand “(They) tied the two hands of the girl.” (Jian Tie He Shang, AD 1100)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Two Historical Motivations

233

(75)

The postverbal nominal refers to the resulting object made of the patient noun. 把這情由細細寫了個稟帖。 (儒林外史三十八回) Bǎ zhè qíng-yóu xìxì xiěle gè bǐngtiě. DISP this reason carefully write-PERF a report “(They) carefully wrote a report about the reason.” (Ru Lin Wai Shi, Chapter 38, AD 1750)

(76)

The postverbal nominal forms a compound verb with the preceding verb. 將角門皆上鎖。 (紅樓夢七十四回) Jiāng jiǎomén jiē shàng suǒ. DISP corner-door all close lock “(They) locked all the corner doors.” (Hong Lou Meng, Chapter 74, AD 1750)

If the resultative was structurally complex, normally marked by the resultative particle de, the patient noun could no longer follow the matrix verb. In this case, the disposal construction was usually the best choice, as illustrated in (77): (77)

又把寶玉的一碗茶也砸得碗碎茶流。 (紅樓夢九回) Yòu bǎ Bǎoyù de yī-wǎn chá yě zá again DISP Baoyu GEN one-CL tea also pound de wǎn suì chá liú. RES bowl smash tea flow “Again, (she) pounded Baoyu’s tea bowl so strongly that the bowl broke and the tea flew everywhere.” (Hong Lou Meng, Chapter 9, AD 1750)

Before the sixth century AD, there was an equal structure, “V X zuò Y,” where the verb could be those meaning “make,” “treat,” “call,” or “name,” corresponding to the English pattern “regard X as Y,” as illustrated in (78). Later, the X element could no longer occur between the verb and the Y element. Then the best way to express the meaning was the disposal construction, as illustrated in (79): (78)




煮豆持作羹。 (曹植 七步詩) Zhǔ dòu chí-zuò gēng. boil bean make-be soup “Boil beans to make soup.” (Cao Zhi, Qi Bu Shi, AD 250)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

234 (79)

有人把椿樹喚作白梅檀。 (寒山詩) Yǒurén bǎ chūn-shù huàn-zuò bái-méi tán. someone DISP toon call-be white-plum sandalwood “Some people call toon white-plum sandalwood.” (Han Shan Shi, AD 650)

The development of the disposal construction was also related to the change in the comparative structure. As discussed in Chapter 11, the equal comparative structure in Old Chinese was “Subj bǐ O yú Standard,” and until the first century AD the preposition yú was no longer obligatory under the operation of the principle of action–resultative ordering, giving rise to the structure “Subj bǐ O Standard.” After the sixth century AD, these two variants of the comparative developed a subtype of the disposal construction, as illustrated in (80) and (81): (80)

Subj bǐ O yú Standard > Subj bǎ O bǐ yú Standard 將世比於花。 (寒山詩) Jiāng shì bǐ yú huā. DISP world compare with flower “Compare the world with flowers.” (Han Shan Shi, AD 650)

(81)

Subj bǐ O Standard > Subject bǎ O bǐ Standard 莫將邊地比京都。 (王縉 九日作) Mò jiāng biān-dì bǐ jīngdū. not DISP remote-area compare capital “Don’t compare the remote area with the capital.” (Wang Jin, Jiu Ri Zuo, AD 700)




In conclusion, the emergence of the disposal construction was far from simply adding another option to the existing VO structure. It was motivated by the establishment of the resultative construction and the formation of the principle of assigning definiteness by syntactic position, two changes that were closely related to each other. Just these strong motivations propelling its development could have enabled the disposal construction to become a major grammatical device.

9.5

Other Related Constructions

Not long after the disposal construction emerged, the Chinese language acquired another important sentential structure whereby the patient noun simply occurred between the subject and the predicate of the action verb without any marking. This

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Other Related Constructions

235

innovative structure served mainly to express an activity. This syntactic construction, which did not exist before the thirteenth century AD, is formulated below: (82)

Subj + NP + VP.

According to our investigation, the earliest examples of this structure are attested in the texts of the thirteenth century AD, as illustrated below: (83)

我昨日冷酒喫多了。 (老乞大新釋) Wǒ zuórì lěng-jiǔ chī-duō-le. I yesterday cold-wine drink-much-PERF “I drank too much cold wine yesterday.” (Lao Qi Da Xin Shi, AD 1300)

Within the above SOV construction, the predicates of early examples were usually complex, typically VR phrases, as exemplified in (83). If the predicate was a simple verb, a normal VO construction was used, as illustrated in (84): (84)

連飲三杯冷酒。 (五代秘史三十回) Lián yǐn sān-bēi lěngjiǔ. consecutively drink three-CL cold-wine “He consecutively drank three cups of cold wine.” (Wu Dai Mi Shi, Chapter 30, AD 1400)

Since then, this construction has been used to the present day, and the following examples reflect the usage of this SOV construction in Contemporary Chinese: (85)

他每餐饭尽量喫饱。 (清代宫廷艳史二十六回) Tā měi-cān fàn jìnliàng chī-bǎo. he every-CL meal to-the-greatest-extant eat-full “He eats as much as he can at every meal.” (Qing Dai Gong Ting Yan Shi, Chapter 26, AD 1930)

(86)

小姐昨夜的酒沒喝醉么? (留東外史續集四十一章) Xiǎojiě zuóyè de jiǔ méi hē-zuì me? Miss last-night GEN wine do-not drink-drunk QUE “Didn’t Miss drink wine and become drunk last night?” (Liu Dong Wai Shi Xu Ji, Chapter 41, AD 1930)

The SOV structure that was first attested in the thirteenth century AD could be viewed as a variant of the disposal construction. That is, it can be regarded as the unmarked disposal construction. The examples could hardly be transformed into a marked disposal expression by adding the bǎ morpheme, which is probably determined by the semantic

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

236

feature of the NP after the subject. As the above examples show, this NP is indefinite. In other words, a division of labor exists between marked and unmarked constructions in linguistic expressions: the NP in the disposal construction must be definite, and that in the SOV construction generally must be indefinite or generic. We should be cautious not to confuse this construction with a sentence whose predicate is another “Subj + VP” structure, forming a “Subj1 + Subj2 + VP” structure that had already existed since Old Chinese. In this traditional structure, the second subject typically holds a possessive relationship with the first subject, and the VP was usually an adjectival phrase describing a static quality. More importantly, the second subject could not be the object of the matrix verb, which crucially distinguished it from the newly emerged SOV structure. The following example represents the so-called double subject structure that has been used throughout history: (87)




新婦神色卑下。 (世說新語 賢媛) Xīnfù shénsè bēixià. new-wife complexion lowly “The new wife looks lowly (of station).” (Shi Shuo Xin Yu, Xian Yuan, AD 450)

Our analysis finds strong support in many Chinese dialects. Different dialects usually develop at different rates and also possibly in different directions, which can be seen in the development of the resultative construction and the disposal construction. As discussed above, one of the two major motivations for the emergence of the disposal construction was the fusion of the verb and the resultative via reanalysis. Even now, this process has not been completed in many dialects, especially in the southeastern region of China, and a sign of this lagging is that the verb and the resultative can still be separated by the object, for instance in the Shanghai dialect, as illustrated below. Accordingly, the development of the dialect’s disposal construction lags far behind that of the standard Mandarin dialect (basically northern Chinese), and the use of the disposal construction is much less frequent in these dialects than in standard Mandarin. When the predicate is a “verb–resultative” phrase, the patient noun simply appears between the subject and the predicate in exactly the same new structure as described above, as illustrated below: (88)

我書抄好勒。 (上海話, 黃伯榮 1996) Ngu sɿ tsʰɔ-hɔ ləʔ. I book copy-complete PRT “I have copied the book completely.” (Huang 1996: 656, Shanghai dialect)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Disposal Markers: Ná, Guǎn, and Gěi (89)

237

我衣裳汰清爽勒。 (上海話, 黃伯榮 1996) Ngu i-zɦɔ̃ dæ-tsʰinsɔ̃ ləʔ. I clothes wash-clean PRT “I have washed the clothes clean.” (Huang 1996: 656, Shanghai dialect)

In the Shanghai dialect, the patient noun can occur in three positions: (a) between the subject and the predicate, (b) between the verb and the resultative, and (c) after the verb and the resultative. According to Huang (1996: 733‒738), however, in other dialects the patient noun can only precede the “verb + resultative” predicate, as in the Shaoxing dialect, the Jinhua dialect, and the Wenzhou dialect, all of which belong to the Wu dialect family. Additionally, these dialects have a considerable number of the unmarked disposal constructions, namely the SOV construction. In particular, the Jinhua dialect has not even developed a marked disposal construction and only has an unmarked disposal construction. Once again, these facts reveal that there is a cause– effect relation between the resultative construction and the emergence of the disposal construction. 9.6

Disposal Markers: Ná, Guǎn, and Gěi

In the development process of the disposal construction, there were successive lexical replacements that were caused by the grammaticalization of the verb take into a disposal marker. As has been widely assumed, verbs meaning “take” started their grammaticalization processes in the first verb position of the serial verb construction, as schematized below: (90)

Subj + VP1 + VP2

Among a set of synonymous words, the member with the most general meaning always has the advantage of being recruited to serve a grammatical purpose. For example, of movement verbs in English, including go, walk, run, and stroll, only go became an immediate future marker, “be going to,” because its meaning is most general (Hopper and Traugott 2003: 102). As mentioned above, at least four verbs (i.e. jiāng “take,” qǔ “fetch,” chí “grasp,” and zhuō “capture”) were used in the first verb position to express a “disposal” sense, but only jiāng evolved into a disposal preposition through semantic bleaching,5 resulting in the loss of its

5

No phonetic erosion was involved in this initial position of the phrase, which could happen only after a host such as the aspect marker le, zhe, and guo (their tonal value was lost, and the vowels were neutralized to schwa [ǝ]).

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

238

verb status. However, the concept of the verb take was needed in daily communication; thus the language community coined another verb bǎ to take over the task around the seventh century AD. Due to its semantic suitability, the verb bǎ was later further grammaticalized into a disposal preposition because it often occurred in the VP1 position. When bǎ became fully grammaticalized, its verb task was taken over by another newly coined verb, ná, around the twelfth century AD. Recently, ná has also developed into a disposal morpheme (and an instrumental marker) for the same reason, as illustrated below: (91)

又拿我當個人。 (紅樓夢六回) Yòu ná wǒ dāng gè rén. also DISP I treat a human “(He) also treated me as someone.” (Hong Lou Meng, Chapter 6, AD 1750)

(92)

他故意拿我開玩笑。 (現代漢語八百詞) Tā gùyì ná wǒ kāi-wánxiào. he purposely DISP I play-joke “He played a joke on me purposely.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 392)

Now, ná can express only a small portion of the various functions of bǎ, for instance when used in the “treat-as” pattern, which can be used by bǎ. However, ná has acquired a unique function, as exemplified in (93), and it cannot simply be replaced by bǎ when the predicate is an idiom with a VO structure. The disposal use of ná was first attested in texts in the eighteenth century AD. In Contemporary Chinese, the variant of the disposal construction with the lowest degree of transitivity, namely the equal structure “X be called Y” (one type of the earliest examples of the construction, attested around the seventh century AD), can be expressed by a newly grammaticalzied morpheme, guǎn, whose lexical meaning is “manage,” as illustrated below: (93)

山西人管土豆叫山藥蛋。 (現代漢語) Shānxī rén guǎn tǔdòu jiào shānyào dàn. Shanxi people DISP potato call Chinese-yam egg “The Shanxi people call potato the egg of Chinese yam.” (Contemporary Chinese)

Only in this specific use can guǎn function like a disposal morpheme, which can also be marked by the disposal bǎ.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Language Contact

239

The above analysis shows that the verb ná “take” has started to be grammaticalized into a disposal morpheme and can partially function like the disposal bǎ. According to the lexical replacement between jiāng and bǎ, we may reasonably expect that ná may eventually replace bǎ as the major disposal morpheme, a process that may take several centuries or even longer. That is, the effect of grammaticalization is twofold: giving rise to new grammatical devices and triggering lexical replacement.

9.7

Language Contact

As pointed out earlier, some researchers have advocated the idea proposed by Li and Thompson (1974) that the Chinese language underwent a word order change from SVO to SOV (e.g. Hashimoto 1976, Huang 1978, Tai 1985). Hashimoto (1976, 1984) made an even more radical suggestion that Chinese languages exhibit an order continuum from north to south, whereby the northern regions acquire many SOV features that are the result of contact with an Altaic substratum. Sun and Givón (1985) denied these assumptions in a statistical work that showed that OV instances in Contemporary Chinese accounted for less than 10 percent of all the instances they investigated. However, our analysis demonstrates that the emergence of the disposal construction signaled not a typological word order change from SVO to SOV in the Chinese language, but rather a new grammatical device, internally motivated by the reanalysis of the verb and the resultative. As a result, the use of the disposal construction (including the unmarked one) is conditioned by two key factors: the definiteness of the patient noun and the structural complexity of the predicate, two characteristics that are uncommon in a real SOV language. Thus we do not believe that the so-called word order change truly happened in the Chinese language, and there is no sign that the language has gained the momentum to develop toward this basic SOV word order. Nevertheless, due to intensive language contact, many Chinese dialects, especially in the northwestern region of the country, have acquired many features that SOV languages typically exhibit. Specifically, the effect of language contact has been made possible by the emergence of the disposal construction that gave the Chinese dialects a linear order that is truly similar to an SOV language, though they are essentially different. Historically, the northwestern areas of China were inhabited by many nationalities or ethnic groups, mainly Mongolian, Uygur, and Tibetan, whose languages all adopted the basic SOV word order. Ethnic Han Chinese people and those of other nationalities lived together and communicated with each other every day. Owing to this intensive language contact, the disposal constructions in the

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

240

Chinese dialects of these areas have indeed developed quite a few remarkable features that are characteristic of SOV languages, some of which are briefly discussed as follows. In the Lanzhou dialect, the scope of using the disposal construction is much wider than in standard Mandarin; for instance, the verb can be simple and low in transitivity, as illustrated in (94) and (95), which may be judged to be ungrammatical by a native speaker of Mandarin Chinese. (94)

我把他們的名字知道。 (蘭州話) Wǒ bǎ tāmen de míngzì zhīdào. I DISP they GEN name know “I know their names.” (Huang 1996: 656, Lanzhou dialect)

(95)

他把我想了。 (蘭州話) Tā bǎ wǒ xiǎng-le. he DISP I miss-PERF “He is missing me.” (Huang 1996: 656, Lanzhou dialect)

In the disposal construction of the Weinan dialect, the patient noun is often indefinite (e.g. a “Num + CL” phrase) when the predicate is a simple verb, as illustrated in (96) and (97). As discussed above, sentences of this type are considered ill-formed in Standard Chinese (Zhu 1982: 187). (96)

把一個雞死了。 (渭南話) Bǎ yīgè jī sǐ-le. DISP one-CL chicken die-PERF “(They) had a chicken that died.” (Huang 1996: 656, Weinan dialect)

(97)

把一本書丟了。 (渭南話) Bǎ yī-běn shū diū-le. DISP one-CL book lose-PERF “(He) lost a book.” (Huang 1996: 656, Weinan dialect)

It is quite unusual among the Chinese dialects that the Qinghai dialect has developed a postnominal marker to signal the object status of the patient noun in preverbal position, a morphological device that is typically found in SOV languages (e.g. Japanese) in order to avoid possible ambiguity between the subject and the object.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Effect on Other Structures

241

In the following examples, the particle ha indicates that the preceding noun phrase is the object: (98)

我開水哈喝了。 (青海話) Wǒ kāi-shuǐ hā hē-le. I boil-water ACCU drink-PERF “I have drunk the boiled water.” (Huang 1996: 656, Qinghai dialect)

(99)

我你哈沒見。 (青海話) Wǒ nǐ hā méi jiàn. I you ACCU do-not see “I didn’t see you.” (Huang 1996: 656, Qinghai dialect)

According to Huang (1996: 656), in the Qinghai dialect the above examples can also be expressed with a bǎ construction, but the morpheme bǎ must precede the patient noun. It is unlikely that the postnominal accusative case hā grammaticalized from the internal system of the Chinese language because its basic word order is still SVO and the grammatical morphemes are typically grammaticalized in the first VO phrase of a serial verb construction. The case marker ha may not be directly borrowed from the grammar of Tibetan, which is spoken by roughly half the population in the county. We can reasonably assume that the development underwent two steps: first, the number of SOV instances dramatically increased in the Qinghai dialect due to language contact, and then the postnominal object marker emerged from the internal grammatical system of the dialect to satisfy the requirement of communication, similar to what happened in other SOV languages.

9.8

Effect on Other Structures

As discussed in Chapter 8, the direct impetus for the emergence of the passive pattern with an obligatory agent was perhaps the emergence of the disposal construction, a new grammatical structure that was first introduced into the language in the seventh century AD, developed quickly from the tenth century AD to the thirteenth, and spread widely after the fifteenth century AD (for details, see Wang 1989: 266‒271). We are now in a position to explain the strikingly unusual direction of the development of passive morphemes in the past millennium. The firm establishment of the disposal construction, which required a patient phrase in order to make it well formed, tended to be structurally similar to its structural counterpart, namely the passive structure. Under this circumstance, all passive morphemes after the sixteenth century AD tended to require an agent

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

242

noun in order to make passive clauses well formed. In contrast, the passive morpheme chī, originating from its “suffer” usage, failed to meet this requirement and hence became extinct around the sixteenth century AD. For the same reason, the dominant bèi passives in early Modern Chinese gradually went out of fashion after the sixteenth century AD, and ultimately became restricted to written language and formal speech in Contemporary Chinese. In addition, due to analogy with the disposal construction, the passive marker bèi developed into a nominative case marker to signal the subject in preverbal position. In this case, the patient noun still stayed in its original position following the verb, a structure that was never seen before the emergence of the disposal construction, as illustrated in (100). When both the agent and the patient occurred in preverbal position, interestingly, the agent noun was marked by the passive bèi, like a nominative marker, and the patient noun by the disposal bǎ, like an accusative marker, as illustrated in (101). This phenomenon is typically found in an SOV language.




(100)

每被老元偷格律。 (白居易 戲贈元九李二十) Měi bèi Lǎo Yuán tōu gélǜ. often NOM Lao Yuan steal rhythm “Lao Yuan often copied my poetic rhythms.” (Bai Ju Yi, Xi Zeng Yuan Jiu Li Shi Er, AD 850)

(101)

被武松把兩個都殺了。 (水滸傳三十二回) Bèi Wǔ Sōng bǎ liǎng-gè dōu shā-le. NOMI Wu Song ACC two-CL all kill-PERF “Wu Song killed the two (girls).” (Shui Hu Zhuan, Chapter 32, AD 1400)

The above phenomenon is regarded as a deviation from the normal development of the passive structure (Wang 1989: 282). A reasonable explanation for this deviation is that due to the fusion of the verb and the resultative and large-scale use of the disposal construction, a considerable number of patient nouns, especially those with definite meaning, appeared in preverbal position. Some of these were marked by the disposal marker, but some remained unmarked, as discussed above. When both the agent and patient nouns occurred in preverbal position, ambiguity was unavoidable because the word order failed to identify which was the subject and which was the object. To avoid possible ambiguity, therefore, a passive marker was recruited to mark the agent noun, and the disposal marker was specialized to mark the object. That is, here communicative function played a part in triggering the grammatical change, which was actually linked with the bigger picture of the development of the grammatical system at that time.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Competition between Jiāng and Bǎ 9.9

243

Competition between Jiāng and Bǎ

The emergence and development of the disposal construction was one of the major changes in the evolution of Chinese grammar. The verb bǎ “take” grammaticalized into a disposal marker around the eighth century AD, approximately 200 years after the verb jiāng underwent the same change. In the period from the eighth century AD to the twelfth, the disposal jiāng was used more frequently than bǎ, and in many contexts they were interchangeable. After more than a millennium of competition, however, bǎ finally won over jiāng and became the only disposal marker in Standard Contemporary Spoken Chinese at the expense of jiāng around the eighteenth century AD. It was quite rare that there existed only one lexical candidate for development into a given grammatical morpheme. It was diachronically much more common that more than one lexical item competed within a grammatical domain, and only one eventually became the ultimate winner. In other words, typically, two or more grammatical morphemes from different lexical sources competed for the same function and coexisted for several centuries or even longer. In the end, one of them won at the expense of the others. Superficially, this kind of competition appears random, and the winner seems to be determined purely by chance. In reality, however, the competition is generally regulated by two factors: first, the inherent semantic and syntactic properties of the grammatical morphemes from their original lexical sources, and second, the overall property of the grammatical system in the particular period of time. All researchers who have studied the disposal construction from a diachronic perspective have treated jiāng and bǎ equally (e.g. Wang 1989: 266‒271). As analyzed previously, prior to their grammaticalization, they were synonymous when used as a verb, roughly meaning “take.” When used as a verb, however, their meanings were not exactly the same, and a crucial subtle difference existed between them: jiāng meant “lead (take) something from one place to another,” implying a distance change, and bǎ usually referred to a static action more similar to hold in English. This semantic contrast influenced their functions after their grammaticalization. We find that in the same paragraph of texts where both the disposal jiāng and bǎ were used, jiāng was preferred if the patient involved a location change, as illustrated in (102), but ná was favored if there was no such location change, as illustrated in (103). The following examples are quoted from two adjacent sentences of the same text composed in the thirteenth century AD. (102)




他將駿馬牽著。 (元雜劇 馬丹陽三度任風子) Tā jiāng jùnmǎ qiān-zhe. he DISP horse lead-PROS “He was leading the horse.” (Yuan Za Ju, Ma Dan Yang San Du Ren Feng Zi, AD 1350)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

The Disposal Construction

244 (103)




把性命耽饒。 (元雜劇 馬丹陽三度任風子) Bǎ xìngmìng dānráo. DISP life risk “You risked your life.” (Yuan Za Ju, Ma Dan Yang San Du Ren Feng Zi, AD 1350)

Because the original lexical meaning of jiāng implied a location change, it often combined with directive verbs such as lái “come” to appear in the first verb position of a serial verb construction, a verb usage coexisting with its disposal use, as exemplified in (104). In this use, jiāng was still an ordinary verb rather than a disposal marker. (104)




將酒來我喫。 (關漢卿 鄧夫人苦痛哭存孝) Jiāng jiǔ lái wǒ chī. bring wine come I drink “Bring the wine for me to drink.” (Guan Han Qing, Deng Fu Ren Ku Tong Ku Cun Xiao, AD 1300)

The above example shows that jiāng preserved more verbal properties than bǎ in Modern Chinese, even though the former grammaticalized earlier. Sun (1996) argued that the directive verb lái introduced a purposive clause after the disposal marker, which was a trigger for the grammaticalization of the disposal construction as a whole. There are two problems with Sun’s analysis: first, the directive lái “come” was usually combined with jiāng but not with bǎ; thus how could it trigger the grammaticalization of the disposal bǎ? Second, a structural hierarchy mistake was made in Sun’s analysis; that is, the directive verb lái was combined with the preceding jiāng to form a VR phrase rather than to introduce a following purposive clause. More crucially, the factor that determined the fate of jiāng was that it also often introduced a patient NP in postverbal position in Modern Chinese, either definite, as in (105), or indefinite, as in (106), but the disposal bǎ could never do so. (105)

當時崔寧買將酒來。 (碾玉觀音) Dāngshí Cuī Níng mǎi jiāng jiǔ lái. then Cui Ning buy DISP wine come “Then Cui Ning bought some wine and brought it here.” (Nian Yu Guan Yin, AD 1200)

(106)

玄德叫拖將張飛來。 (三國志通俗演義卷二) Xuán Dé jiào tuō jiāng Zhāng Fēi lái. Xuan De order drag DISP Zhang Fei come “Xuan De ordered (them) to drag Zhang Fei here.” (San Guo Zhi Tong Su Yan Yi, Chapter 2, AD 1400)

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

Competition between Jiāng and Bǎ

245

The above uses of jiāng, a very common phenomenon in early Modern Chinese, may have originated from its historical contexts. At the time when jiāng grammaticalized around the fifth century AD, the operation of the principle of action–resultative ordering was still relatively weak; thus the disposal jiāng as a non-resultative preposition could occur in either position, preverbal or postverbal. As the principle eventually became a powerful rule, the postverbal use of the disposal jiāng was eliminated from the grammar. In other words, the postverbal use of the disposal jiāng was inconsistent with the grammatical system at the time, which put it at a disadvantage in the competition with bǎ (always consistent with the general development tendency of the grammar). By the eighteenth century AD, the competition was over: jiāng had been completely replaced by bǎ in Mandarin Chinese. However, jiāng survives in some dialects, such as the Suzhou dialect and the Shanghai dialect, whose resultative construction lagged behind that of northern Chinese; thus the principle has not been finally formed. The semantics and syntax of a verb prior to its grammaticalization usually determine its functions as a grammatical morpheme. In turn, these inherent features may influence its competition with other related grammatical morphemes that originated from different lexical sources.

https://doi.org/10.1017/9781108921831.009 Published online by Cambridge University Press

10 Verb Copying and Reduplication 10.1 The Verb-Copying Construction Grammatical changes fall into two types: alteration and innovation. The term “alteration” means that a grammatical category already existed at an earlier stage, but its specific structures, markings, and functions might vary over time. For instance, the passive structure and the ditransitive construction belong to this sort of development and were the oldest grammatical apparatuses to undergo serial changes in markings, structures, and functions in history. In contrast, the term “innovation” means that the grammatical category did not exist in the language before a certain time but was created sometime in history; for example, the disposal construction, the plural marker, and the classifier system were completely new in markings, structures, and functions for the language at the time when they emerged. Like the disposal construction, the verb-copying construction was innovated approximately 300 years ago and is a relatively young construction that did not have any equivalent before the fifteenth century AD. Unlike the disposal construction, which is marked by a grammatical morpheme, bǎ, the verb-copying construction represented an abstract syntactic structure without any specific lexical marking, as formulated in (1), where Vx stands for the identical (copying) verb: (1)

Subj + (Vx + O) + (Vx + R)

The two verbs are identical, as exemplified below: (2)

他做飯做累了。 (現代漢語) Tā zuò-fàn zuò-lèi-le. she cook-meal cook-tired PERF “She became tired from cooking.” (Contemporary Chinese)

246

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction

247

The characteristics of the verb-copying construction are elusive; hence it is not easy to generalize the conditions of how to use it. Each of the four elements comprising the predicate depicted in (1) must be present in order to make it well formed. Li and Thompson (1981: 442‒450) stated that the verb-copying construction refers to a grammatical process in which a verb is “copied after its object,” that the object must be nonreferential and cannot be modified by any definite word (e.g. demonstratives) or any numeral phrases, and that only the copied verb is finite and can be suffixed with aspect markers. Although their observation is basically correct, it is still unsuccessful in capturing the necessary condition of its application because they fail to define the property of the resultative,1 a key factor in determining when the verb-copying construction must be used. Once again, the innovation of the verb-copying construction, like the emergence of any new grammatical form, must have been conditioned by two factors: first, it should have been favored or even required by the overall properties of the grammatical system in that particular period of history, which motivates a new grammatical device to appear and propels it to develop a new grammatical paradigm; second, it could develop only out of a certain context, which guarantees that the newly innovated form would be structurally compatible with the existing system of the grammar.

10.1.1 The Syntax of the Verb-Copying Construction In general, the application of the verb-copying construction could be either optional or obligatory, depending on the semantic features of the object and the resultative. First, let us consider the optional situation, where the object could be introduced either by the verb-copying construction, as illustrated in (3); by the topicalization construction, as illustrated in (4); by the SOV structure, as illustrated in (5); and even by the VRO structure, as illustrated in (6). Each of these might have its own pragmatic values (Liu et al. 2001: 533‒643). (3)

The verb-copying construction 他喫肉喫腻了。 (現代漢語) Tā chī-ròu chī-nì le. he eat-meat eat-bored PERF “He has become bored with eating meat.” (Contemporary Chinese)

1

Li and Thompson (1981) termed the element after the copied verb “adverbial.” In our analysis, we distinguish the term “adverbial” from “resultative”: the former refers to the accompanying feature of the verb, such as manners, locations, and instruments; the latter indicates some kind of result of the verb, such as end points and resulting states. This distinction is critical in understanding the evolution of Chinese grammar as well as the grammatical system of Contemporary Chinese.

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

248 (4)

The topicalization construction 肉他喫腻了。 (現代漢語) Ròu tā chī-nì-le. meat he eat-bored PERF “As for meat, he has become bored (with it).” (Contemporary Chinese)

(5)

The SOV structure2 他肉喫腻了。 (現代漢語) Tā ròu chī-nì-le. he meat eat-bored-PERF “He has become bored with eating meat.” (Contemporary Chinese)

(6)

The VRO structure 他喫腻了肉。(現代漢語) Tā chī-nì-le ròu. he eat-bored-PERF meat “He is sick of eating meat.” (Contemporary Chinese)

Now let us examine when the use of the verb-copying construction is obligatory. First, the object noun must be indefinite and nonreferential; second, the underlying subject of the resultative must be the subject of the sentence. When these two conditions are met, the verbcopying construction is the sole choice for expressing the meaning with a single clause. In this situation, it is impermissible to use topicalization or the VRO construction, as illustrated below. (7)

(a) The verb-copying construction 他喫肉喫胖了。 (現代漢語) Tā chī-ròu chī-pàng-le. he eat-meat eat-obese-PERF “He has become obese from eating meat.” (b) The topicalization construction *肉他喫胖了。 *Ròu tā chī-pàng-le. meat he eat-obese-PERF

2

In the literature, the OV phrase in this structure is viewed as another “Subj + Predicate” construction, where the object noun is labeled “small subject” or “secondary topic” (Zhu 1982: 106‒108).

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction

249

(c) The VRO structure *他喫胖了肉。 *Tā chī-pàng-le ròu. he eat-obese-PERF meat (Contemporary Chinese) In example (7a), the underlying subject of the resultative pàng “fat” is the subject of the sentence, and the object ròu “meat” is indefinite and nonreferential. The resultative nì “bored” in (3)–(6) seems to have the subject of the clause as its underlying subject, but in Chinese this resultative can also be used as an adjective to describe the quality of the object “meat” as “greasy,” as in the following example: (8)

這塊肉看起來很膩。 (現代漢語) Zhè-kuài ròu kàn-qǐlái hěn nì. This-CL meat look-up very greasy “This piece of meat looks very greasy.” (Contemporary Chinese)

In contrast, the resultative pàng “obese” cannot be used in the same way as nì “greasy.” Therefore the syntactic behavior of the V-nì “greasy” phrase is different from that of V-pàng “obese.” The application of the verb-copying construction is closely related to the rule governing which “verb + resultative” phrases can or cannot take an object. The transitivity of a “verb + resultative” phrase is determined by the semantic features (i.e. the underlying subject) of the resultative. Syntactically, the verb and the resultative form an immediate constituent. Semantically, however, the resultative may relate to the subject, the verb, or the object. The constituent that is related to the resultative is called the underlying subject of the resultative in our analysis. From this angle, we can draw the following two generalizations. Generalization 1. If the resultative describes the patient noun, the whole “verb + resultative” structure is transitive and can be freely followed by the patient noun as its object, even when the verb is intransitive or bears no direct relation to the object, as illustrated below: (9)

他笑疼了肚子。 (現代漢語) Tā xiào-téng-le dùzi. he laugh-painful-PERF stomach “He laughed (so much) his stomach hurt.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

250 (10)

他喫窮了一家子。 (現代漢語) Tā chī-qióng-le yī jiā-zi. he eat-poor-PERF one family “He ate (too much expensive food) and caused his family to be poor.” (Contemporary Chinese)

(11)

他打碎了一个碗。 (現代漢語) Tā dǎ-suì-le yī-gè wǎn. he beat-broken-PERF one-CL bowl “He broke a bowl.” (Contemporary Chinese) Generalization 2. When the resultative describes either the subject or the verb, the whole verb–resultative phrase becomes intransitive and cannot be followed by any object, as illustrated in (12), where the resultative wǎn “late” modifies the verb. In this case, the verb-copying construction is the only available device to express the meaning with a single clause, where the first verb introduces the object and the copied verb introduces the resultative. We have seen the examples of the verb–resultative phrase whose underlying subject is identical to that of the sentence in (7), and in the example below the underlying subject is the verb:

(12)

(a) 他喫飯喫晚了。 (現代漢語) Tā chī-fàn chī-wǎn-le. he eat-food eat-late-PERF “He ate late.” (b) *他喫晚了飯。 *Tā chī-wǎn-le fàn. he eat-late-PERF food

However, in Contemporary Chinese, there are two exceptions to Generalization 2 whose resultatives relate to the subject of the sentence but can be followed by an object: chī-bǎo “eat-full” and hē-zuì “drink-intoxicated,” as illustrated by the naturally occurring examples that we investigated: (13)

夏先生喫饱了飯。 (作家文摘) Xià Xiānshēng chī-bǎo-le fàn. Xia Mr. eat-full-PERF food “Mr. Xia has eaten enough food.” (Zuo Jia Wen Zhai, Contemporary Chinese)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction (14)

251

他在外面喝醉了酒。 (努尔哈赤傳) Tā zài wàimiàn hē-zuì-le jiǔ. he in outside drink-intoxicated-PERF wine “He drank outside and became intoxicated.” (The Biography of Nu Er Ha Chi, Contemporary Chinese)

Superficially, the above examples seem to violate the second generalization, but they are actually formed at the lexical level – word formation. In other words, they are not generated by any active grammatical rule but are idiomatic due to their high-frequency co-occurrences. Scrutiny reveals that the usages, namely, chī-bǎo fàn “eat-full food” and hē-zuì jiǔ “drink-intoxicated wine,” are purely idiosyncratic without any productivity, and no elements can be replaced by other lexical items. For example, if the object fàn “food” in (13) is replaced by miànbāo “bread” or jiǔ “wine” in (14) is replaced by pí-jiǔ “beer,” the sentences will become ill-formed. The above two so-called exceptions came into existence approximately 100 years ago. The mechanism of their emergence was the same as that for the establishment of the resultative construction in Medieval Chinese. Due to the high-frequency co-occurrences of the verb and the resultative, they became fused into compound-like verbs and hence could be followed by the object. According to the diachronic development of the resultative construction that caused the disappearance of the verb co-ordination construction, we might expect that the increasing exceptions might finally rule out generalizations 1 and 2, but this process may take several hundred years or even longer to reach completion. By understanding the above rule determining transitivity, we can avoid misjudgment of the grammaticality of the related usages, which is useful even for native speakers. For example, Li (1993) claimed that the following sentence has three interpretations: (a) the subject Taotao is tired, (b) the object Youyou is tired, and (c) both Taotao and Youyou are tired: (15)

涛涛追累了友友。 (現代漢語) Tāotāo zhuī-lèi-le Yǒuyǒu. Taotao chase-tired-PERF Youyou. Interpretation 1: “Taotao’s chasing made Youyou tired.” Interpretation 2: “Taotao became tired from chasing Youyou.” Interpretation 3: “Both Taotao and Youyou were tired because of the chasing.”

Li’s judgment is problematic according to the grammatical rule, which predicts that only when the resultative lèi “tired” describes the state of the patient “Yaoyao” (i.e. the

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

252

underlying subject) can the “verb + resultative” phrase be followed by the object. Among the above three interpretations suggested by Li, in reality only interpretation 1 is possible, and the other two interpretations are ruled out. In the above examples of the verb-copying construction, the resultatives are adjectives (e.g. pàng “obese”) or a complex structure marked by the resultative particle -de. Additionally, the resultative can be a preposition phrase or a duration quantifier, as illustrated in (16) and (17): (16)

他看書看到凌晨兩點。 (現代漢語) Tā kàn-shū kàn dào língchén liǎng diǎn. he read-book read to morning two o’clock “He had been reading until two o’clock in the morning.” (Contemporary Chinese)

(17)

他等車等了很長時間。 (現代漢語) Tā děng-chē děng-le hěn cháng shíjiān. he wait-bus wait-PERF very long time “He waited for a bus for a very long time.” (Contemporary Chinese)

As discussed in Chapter 6, the development of the resultative construction involved the fusion of the verb and the resultative (including preposition phrases). Consequently, the object could no longer occur between the verb and the resultative. This fusion had many by-products, one of which was the emergence of the verb-copying construction, as analyzed in the following section. 10.1.2 The Stages of Development of the Verb-Copying Construction The synchronic characteristics of the verb-copying construction, which were discussed above, hint at its diachronic origin. The copied verb must be followed by a resultative, and the semantic property of the resultative determines when to use the verb-copying construction. These facts suggest that the emergence of the verb-copying construction was closely related to the establishment of the resultative construction. Like the disposal construction, in fact, the verb-copying construction was an outcome of the development of the resultative construction. As we discussed previously, the process of the development of the resultative construction was actually the fusion of the verb and the resultative, a process of fusing two independent constituents into a single constituent, where the boundary between the verb and the resultative was erased and the syntactic position between them was eliminated accordingly. Due to the elimination of the syntactic position, the elements that originally occurred in the position between the verb and the resultative, mainly

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction

253

including the object, adverbs and negative markers, had to be rearranged elsewhere in the sentence. The rearrangement of the patient noun was guided by the following two factors: (a) the semantic property of the patient noun, either definite or indefinite and either referential or nonreferential, and (b) the underlying subject of the resultative, the subject, the verb, or the object. This process resulted in three major innovative constructions, which can be outlined as follows: (a)

(b)

(c)

The first type of rearrangement by means of the disposal construction is VOR > bǎ O VR, when the object is definite and the underlying subject of the resultative is the patient noun. The second type of rearrangement by means of the VRO structure is VOR > VRO, when the object is indefinite and the underlying subject of the resultative is the patient noun. The third type of rearrangement by means of the verb-copying construction is VOR > VO VR, when the object is either nonreferential or indefinite and the resultative describes either the subject or the verb.

Having discussed the types of VOR structure that developed into disposal or VRO constructions (for details, see relevant chapters), we focus here on VOR instances that developed into the verb-copying construction. Before Modern Chinese, temporal adverbs and time nouns could freely occur after the object to indicate the duration or repetition of the action. These structures became ungrammatical due to the fusion of the verb and the resultative, and their corresponding expressions in Contemporary Chinese usually take the form of the verb-copying construction, as illustrated in (18) and (19): (18)




(a) 待君久不至。 (世說新語 方正) Dài jūn jiǔ bù zhì. wait you long not come “I waited for you for a long time, but you didn’t show up.” (Shi Shuo Xin Yu, Fang Zheng, AD 450) (b) 我等你等了很久。 (現代漢語) Wǒ děng nǐ děng-le hěn jiǔ. I wait you wait-PERF very long “I waited for you for a quite long time.” (Contemporary Chinese)

(19)

(a) 一僧長念法華經已多年。 (入唐求法巡禮記) Yī sēng-zhǎng niàn Fǎ Huá Jīng yǐ duō-nián. One monk-master read Fa Hua Jing already many year “A master of monks had read the Fa Hua Sutra for many years.” (Ru Tang Qiu Fa Xun Li Ji, AD 850)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

254

(b) 僧長念法華經已經念了很多年了。 (現代漢語) Sēng zhǎng niàn Fǎ Huá Jīng yǐjīng niàn-le hěnduō nián. monk master read Fa Hua Jing already read-PERF very many year “A master of monks had read Fa Hua Jing for many years.” (Contemporary Chinese) Note that if the resultative described the temporal property of the verb, the object could be definite or referential, as exemplified in (19), where the object is the title of a sutra. At the time, an adjectival phrase could also occur after the object to describe the manner or the quality of the verb, and their corresponding linguistic form in Contemporary Chinese is also the verb-copying construction, as illustrated in (20) and (21): (20)




(a) 戴安道中年畫行像甚精妙。 (世說新語 巧藝) Dài Ᾱndào zhōng-nián huà xíngxiàng shén jīngmiào Dai Andao middle-age draw portrait very skillfully “Dai Andao drew portraits very skillfully.” (Shi Shuo Xin Yu, Qiao Yi, AD 450) (b) 戴安道中年畫行像畫得十分精妙。 (現代漢語) Dài Ᾱndào zhōng-nián huà xíngxiàng huà de shífēn jīngmiào. Dai Andao middle-age draw portrait draw RES extremely skillful “Dai Andao drew portraits very skillfully.” (Contemporary Chinese)

(21)




(a) 而談琴書愈妙。 (世說新語 雅量) Ér tán qín-shū yù miào. But talk zither-book more beautiful “But (he) was talking about the theory on zither more beautifully.” (Shi Shuo Xin Yu, Ya Liang, AD 450) (b) 他談琴書談得更加漂亮了。 (現代漢語) Tā tán qín-shū tán de gèngjiā piàoliang le. he talk zither-book talk RES more beautiful PERF “But (he) was talking about the theory on zither more beautifully.” (Contemporary Chinese)

Furthermore, when the object is nonreferential and the resultative is marked by the resultative morpheme de, the corresponding linguistic form in Contemporary Chinese

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction

255

can only be the verb-copying construction, as illustrated in (22) and (23). Thus the structure changed from “V de O R” to “Vx O Vx de R,” as illustrated below: (22)

(a) (顏沖)只是中得毒深。 (大慧普覺禪師書) (Yán Chōng) zhǐshì zhōng de dú shēn. Yan Chong just capture RES poison deep “Yan Chong was just seriously poisoned.” (Da Hui Pu Jue Chan Shi Shu, AD 1150) (b) 顏沖中毒中得很深。 (現代漢語) Yán Chōng zhòng-dú zhòng de hěn-shēn. Yan Chong capture-poison capture RES very-deep “Yan Chong was seriously poisoned.” (Contemporary Chinese)

(23)




(a) 然做得官好。 (朱子語類 訓門人) Rán zuò de guān hǎo. But do RES official-position well “He just carries out the official position well.” (Zhu Zi Yu Lei, Xun Men Ren, AD 1200) (b) 他做官做得很好。 (現代漢語) Tā zuò guān zuò de hěn-hǎo. he do official-position do RES very-well “He carries out the official position well.” (Contemporary Chinese)

The establishment of the resultative construction took more than a millennium to reach its completion point. The earliest “verb + resultative” examples of this type can be dated back to the third century AD or even earlier. From then to the fifteenth century AD, the old form and the new form coexisted, and the new form finally entirely replaced the old one. Even in texts composed around the fifteenth century AD, the VOR structure could still be sporadically seen, as illustrated in (24) and (25), where the verb and the resultative were separated by the object: (24)

鞭打的督郵死。 (關張雙赴西蜀夢) Biān dǎ de dūyóu sǐ. whip beat RES superintendent die “(He) whipped the superintendent to death.” (Guan Zhang Shuang Fu Xi Shu Meng, AD 1350)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

256 (25)

又酒淹得襯衫濕。 (李太白貶夜郎) Yòu jiǔ yān de chènshān shī. again wine splash RES shirt wet “Once again, his shirt was made wet by splashing wine.” (Li Tai Bai Bian Ye Lang, AD 1350)

Around the fifteenth century AD, not long after the VOR structure entirely disappeared, the verb-copying construction came into existence.

10.1.3 The Discourse Structure Origins of the Verb-Copying Construction Real examples of the verb-copying construction are attested in texts from around the fifteenth century AD, as illustrated in the following examples: (26)

請人請到四五次不來。 (盧太學詩酒傲公侯) Qǐng-rén qǐng dào sìwǔ cì bù lái. invite-people invite to four-five times not come “(He) invited the people many times, but they didn’t come.” (Lu Tai Xue Shi Jiu Ao Gong Hou, AD 1650)

(27)

從小兒一處淘氣淘了這麼大。 (紅樓夢五十四回) Cóng xiǎo’ér yīchù táo-qì táo-le zhème dà. from young together play-naughty play-PERF so old “(We) played mischievously from youth to now.” (Hong Lou Meng, Chapter 54, AD 1750)

(28)

我碰頭碰到這裡來。 (紅樓夢四十四回) Wǒ pèng-tóu pèng dào zhèlǐ lái. I bump-head bump to here come “I struggled to come here.” (Hong Lou Meng, Chapter 44, AD 1750)

(29)

一年鬧生日也鬧不清。 (紅樓夢五十一回) Yī nián nào shēngrì yě nào bù qīng. one year celebrate birthday also celebrate not finish “We celebrated birthdays for a whole year long.” (Hong Lou Meng, Chapter 51, AD 1750)

The commonality among the earliest examples of the verb-copying construction was that the objects were in general nonreferential and indefinite. Owing to the disyllabification tendency, many disyllabic verbs with the VO internal structure cannot be followed by another object, such as táo-qì “play naughty” in (27), where

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction

257

the internal object is by nature nonreferential. In this case, the verb-copying construction is often used to indicate some kind of resultative or even an extra object, as illustrated in (30). (30)

未到一年, 站籠站死兩千多人。(老殘遊記三回) Wèi dào yī nián, zhàn lóng zhàn sǐ liǎng qiān duō rén. not at one year close cage close die two thousand more people “Within less than one year, more than two thousand people were killed by enclosing them in cages.” (Lao Can You Ji, Chapter 3, AD 1900)

The verb-copying construction evolved out of a loose discourse organization, where the first clause had a predicate consisting of a VO phrase and the second clause had a VR predicate. As a stable grammatical device, the number of the constituents is fixed and cannot be added to; for instance, only a single VO phrase is allowed within the construction. Additionally, there is a grammatical dependence between the base and copied verbs, and only the copied verb can be finite, being suffixed with an aspect marker or modified by a temporal adverb. These features were gradually acquired during the development process, and before the verb-copying construction took shape, more than one VO phrase could be used within a clause, as illustrated in (31): (31)

官里無貪淫貪欲貪成性。 (輔成王周公攝政) Guān-lǐ wú tān-yín tān-yù tān chéng xìng. government not-have covet-sex covet-desire covet be habitual “The officials have not come to habitually covet sex and other desires.” (Fu Cheng Wang Zhou Gong She Zheng, AD 1300)

Additionally, the diachronic development of the resultative construction involved the combination of two independent clauses into a single clause, a common phenomenon across languages (for details, see Hopper and Traugott 2003: 175‒211). Between a concessive or hypothetic clause and the matrix clause, the two verbs of both clauses could be identical, where the first introduced an object and the second was followed by a resultative. (32)




東城雖說佛家語亦說得好。 (朱子語類 訓門人) Dōngchéng suī shuō Fújiā yǔ yì shuō de hǎo. Dong Cheng although talk Buddhism theory also talk RES good “Although he talked about Buddhism, Dong Cheng still talked about it very well.” (Zhu Zi Yu Lei, Xun Men Ren, AD 1200)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

258 (33)




又如喫飯, 不喫飯在肚子裡。 (朱子語類 訓門人) Yòu rú chī-fàn, bù chī zài dùzi lǐ. again as eat-food not eat into stomach inside “As you eat food, you don’t eat it into your stomach.” (Zhu Zi Yu Lei, Xun Men Ren, AD 1200)

Nevertheless, we do not claim that the verb-copying construction developed directly out of the biclausal sentences described above. What we are trying to say here is that the existence of the above discourse organization to a certain extent influenced what the specific structure of the new grammatical device could look like. In the following examples, the two clauses were combined more tightly, with no pause between them: (34)

死也死得憋脫。 (大慧普覺禪師書) Sǐ yě sǐ de biētuō. die also die RES good-for-nothing “If you die, your death is good for nothing.” (Da Hui Pu Jue Chan Shi Shu, AD 1150)

(35)

打呵打著實處。 (漢高祖濯足氣英布) Dǎ hē dǎ zhuó shí-chù. beat PRT beat at real-place “If (you) beat him, you beat on the key part (of his body).” (Han Gao Zu Zhuo Zu Qi Ying Bu, AD 1350)

The verb sǐ “die” in (34) is intransitive, which is disallowed in the verb-copying construction. Additionally, since the two clauses were still relatively independent, as the examples in (35) show, the object of the first verb was omitted in the above examples, which is impossible in the verb-copying construction. In the following contexts, two adjacent clauses express an activity and its resultative with structures even closer to that of the verb-copying construction: (36)

另寫一個賞單, 勉強寫足了五百貫。 (宋四公大鬧禁魂張) Lìng xiě yī-gè shǎng dān, miǎnqiǎng xiě-zú-le additionally draft one-CL award-check barely draft-enough-PERF wǔbǎi guàn. five hundred dollar “(He) additionally drafted another award check that was merely five hundred dollars.” (Song Si Gong Da Nao Jing Hun Zhang, AD 1350)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction

259

(37) 向前解了拶子, 解的平安兒直聲呼喚。 (金瓶梅詞話三十五回) Xiàng-qián jiě-le zǎnzi, jiě de Píngānr zhí-shēng hūhuàn. Go-ahead untie-PERF stick untie RES Pinganr out-loud cry “(Two soldiers went to untie the torture sticks, and Pinganr cried out loud.” (Jin Ping Mei Ci Hua, Chapter 35, AD 1550) Apparently, the above examples are still not real instantiations of the verb-copying construction because the object in (36) is modified by a numeral phrase and the first verb in (37) is suffixed with the perfective -le, which distinguishes them from the verbcopying construction in Contemporary Chinese. As mentioned before, these features are not allowed in the verb-copying construction. Undoubtedly, the following example was a signal that the verb-copying construction had almost been established, although there was still some difference: (38) 我們認人問姓還認不清呢。 (紅樓夢五十二回) Wǒmen rèn rén wèn xìng hái rèn bù-qīng ne. we recognize people ask surname still recognize not-complete PRT “We ask for their surnames but still cannot entirely recognize who they are.” (Hong Lou Meng, Chapter 52, AD 1750) In the above example, there are two VO phrases with different verbs, but the final part is only a single VR phrase. This feature indicates that the template of the verb-copying construction had nearly taken shape, with the final part being limited to a single VR phrase. There were various types of discourse structure in which two adjacently occurring clauses had the same verbs: the former introduced the object, and the latter carried the resultative. The wide existence of this phenomenon provided the discourse configuration from which the verb-copying construction was innovated. As a result, the new structure naturally entered the language without making the native speaker feel like it was “foreign.” In short, the emergence of the verb-copying construction was favored by the development of the resultative construction and shaped by the existing discourse.

10.1.4 Semantic Diversification between Verb and Object The VO construction has been one of the most basic constructions throughout the history of Chinese, and the canonical word order – namely the verb plus the object – has remained the same. However, the semantic relationships between the verb and the object have changed significantly from Old Chinese to the present day in a direction from simplicity to complexity. In Old Chinese, many noun phrases, when used in postverbal position, needed a preposition to be introduced. That is, they were directly governed by the preposition rather than by the matrix verb. As discussed in Chapter 7,

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

260

these preposition phrases could refer to the locative or instrumental of declarative sentences, the agent noun of passive sentences, the direct object of ditransitive sentences, and so on. When the principle of action–resultative ordering took effect from the beginning of the Medieval Chinese period, no preposition phrases that did not express any kind of resultative of the verb could stay in postverbal position. In response to this change, as we discussed in other chapters, the language employed preverbal prepositions to take over the original functions of those in the passive, locative, comparative, instrumental, and ditransitive constructions, all of which were originally constrained to sentence-final position. In this case, a straightforward response was that some of the old “V + Pre NP” constructions simply changed to “V + NP” by deleting the particular preposition that originally occurred in postverbal position. As a result, the possible relationships between the verb and the object became much more complex; in other words, the semantic relationship between the verb and the object came to be highly diverse. To illustrate this point, let us examine how many kinds of object the verbs of eating, whose forms are shí in Old Chinese and chī in Contemporary Chinese, could have. In Old Chinese, for a period of more than a millennium, the objects of the verb shí “eat” were limited to only those concepts that were edible and bore a stereotypical action–patient relationship, as illustrated in (39): (39)




碩鼠碩鼠, 無食我黍! (詩經 碩鼠) Shuò shǔ shuò shǔ, wú shí wǒ shǔ! big mouse big mouse not eat my millet “Big mouse! Don’t eat my millet.” (Shi Jing, Shuo Shu, 1000–600 BC)

In Contemporary Chinese, however, in addition to the words meaning edible concepts, the objects of the verb chī “eat” can be a location, an instrument, a condition, a manner, or many other options, as illustrated in the following examples. (40)

Object referring to something edible: 他喫了一個苹果。 (現代漢語) Tā chī-le yī-gè píngguǒ. he eat-PERF one-CL apple “He has eaten an apple.” (Contemporary Chinese)

(41)

Object referring to location: 他喫餐館, 我喫家裡。 (現代漢語) Tā chī cānguǎn, wǒ chī jiālǐ. he eat restaurant I eat home “He ate in a restaurant, and I ate at home.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction

261

(42)

Object referring to condition: 他們喫父母, 我喫自己。 (現代漢語) Tāmen chī fùmǔ, wǒ chī zìjǐ. they eat parent I eat myself “Their life relies on their parents, and I make a life by myself.” (Contemporary Chinese)

(43)

Object referring to qualification: 他喫手藝, 我喫智慧。 (現代漢語) Tā chī shǒuyì, wǒ chī zhìhuì. he eat craftsmanship I eat intelligence “He lives on his craftsmanship, and I live on my intelligence.” (Contemporary Chinese)

(44)

Object referring to instrument: 他们喫銀行利息。 (現代漢語) Tāmen chī yínháng lìxí. they eat bank interest “They live on the interest of their bank.” (Contemporary Chinese)

Except for (40), where the object is edible, in Old Chinese a proper preposition needed to be used to introduce “inedible” objects. For example, the locative preposition yú was used in (45) and the instrumental preposition yǐ was used in (46), and the conditional preposition hū was used in (47). For comparison, let us consider some real examples in Old Chinese with the matrix verb shí “eat.” (45)

Preposition yú “at” to introduce a locative: 朝食於株。 (詩經 株林) Zhāo shí yú Zhū. morning eat in Zhu “(They) had breakfast in Zhu.” (Shi Jing, Zhu Lin, 1000–600 BC)

(46)

Preposition zì “from” to introduce a resource: 退食自公。 (詩經 羔羊) Tuì shí zì gōng. Come-back eat from government “They came back and ate food from the government.” (Shi Jing, Gao Yang, 1000–600 BC)







https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

262 (47)

Preposition hū “on” to introduce a condition: 不多食乎力。 (莊子 秋水) bù duō shí hū lì. not many eat on physical-work “Not many people make a living on the basis of physical work.” (Zhuang Zi, Qiu Shui, 300 BC)




The disappearance of the non-resultative prepositions from postverbal position involved a long and complex process. In the early stages, only the preposition was absent, leaving the NP unmarked. According to He (2005: 178‒211), this change started as early as the first century BC, which is illustrated in the following examples from the Shi Ji (composed around the first century BC), whereas the unmarked NPs in postverbal position were introduced by a particular preposition in the Zuo Zhuan (composed around the fifth century BC): (48)

The postverbal locative NP left unmarked: 大破之定陶。 (史記 項羽本紀) Dà pò zhī [ ] Dìng Táo. seriously defeat they Ding Tao “(The troop) seriously defeated them in Ding Tao.” (Shi Ji, Xiang Yu Ben Ji, 100 BC)

(49)

The postverbal agent NP left unmarked: 食其故得幸太后。 (史記 呂太后本紀) Shí Jí gù dé xìng [ ] Tài-Hòu. Shi Ji then get appreciate Emperor-Mother “Then Shi Ji was appreciated by the emperor’s mother.” (Shi Ji, Lü Tai Hou Ben Ji, 100 BC)

(50)

The postverbal resource NP left unmarked: 而胡降者皆衣食縣官。 (史記 平淮書) Ér Hú jiàng-zhě jiē yī sì [ ] xiàn-guān. But Hu surrender-er all wear eat county-official “But the Hu surrenderers all were provided clothes and food by the officials of the county.” (Shi Ji, Ping Huai Shu, 100 BC)

(51)

The postverbal comitative NP left unmarked: 足下自以為善漢王。 (史記 淮陰侯列傳) Zúxià zì yǐwéi shàn [ ] Hàn Wáng. you self think kind Han Wang “You think that you are kind King Han.” (Shi Ji, Huai Yin Hou Lie Zhuan, 100 BC)













https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

The Verb-Copying Construction (52)

263

The postverbal comparative standard NP left unmarked: 所斬捕功已多大將軍。 (史記 衛將軍驃騎列傳) Suǒ zhǎn bǔ gōng yǐ duō [ ] dà jiàngjūn. NOM kill capture credit already more big general “(His) credits in killing and capturing (at war) are more than you (general).” (Shi Ji, Wei Jiang Jun Piao Qi Lie Zhuan, 100 BC)




Due to the above change, the former ditransitive construction, which used the preposition yǐ “with” to introduce the direct object after the indirect object, as illustrated in (53), disappeared. After that, the ditransitive construction became unmarked (cf. Chapter 12), as illustrated in (54).3




(53)

投我以木瓜, 報之以瓊琚。 (詩經 木瓜) Tóu wǒ yǐ mùguā, bào zhī yǐ qióngjū. throw I with papaya repay she with fine-jade “She threw me a papaya, and I repaid her with a piece of fine jade.” (Shi Jing, Mu Gua, 1000–600 BC)

(54)

客從遠方來, 遺我雙鯉魚。 (飲馬長城窟行) Kè cóng yuan-fāng lái, yí wǒ shuāng lǐyú. guest from remote-place come give me two carp “The guest who came from a remote place gave me two carps (as a gift).” (Yin Ma Chang Cheng Ku, AD 100)

In Early Old Chinese, there were two variations on the ditransitive construction. Within one of them, the direct object was marked by the preposition yǐ, as shown above. In the other pattern, the direct object was unmarked, as illustrated in the above two examples. There was a division of labor: when the direct object was concrete, it needed to be introduced by the preposition yǐ “with”; when it was abstract, no preposition was used, and the construction remained unmarked, as illustrated in the following two examples. Later, these two variations converged into a single structure.

3




(55)

誨爾序爵。 (詩經 桑柔) Huì ěr xù jué. teach you rank official “(I) would teach you the rankings of the officials.” (Shi Jing, Sang Rou, 1000–600 BC)

(56)

夫子語我九言。 (左傳 定公四年) Fūzǐ yǔ wǒ jiǔ yán. teacher tell I nine word “The teacher told me nine words.” (Zuo Zhuan, Ding Gong Si Nian, 550–400 BC)




For a fuller discussion, see Chapter 12.

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

264

Around the sixteenth century AD, when the verb gěi “give” developed into a dative marker, two construction variations became available to express a ditransitive event: (a) “Subj Vgěi Oi Od” and (b) “Subj V Od gěi Oi,” as illustrated in (57) and (58), respectively. (57)

我寄給他了一本書。 (現代漢語) Wǒ jì-gěi tā le yī-běn shū. I mail-to he PERF one-CL book “I mailed a book to him.” (Contemporary Chinese)

(58)

我寄了一本書給他。 (現代漢語) Wǒ jì-le yī-běn shū gěi tā. I mail-PERF one-CL book to he “I mailed a book to him.” (Contemporary Chinese)

The above marked ditransitive constructions coexist with the unmarked constructions (for details, see Chapter 12). Within the ditransitive construction in Contemporary Chinese, the preposition phrase “gěi + NP” can be construed either as the end point (as illustrated above) or as a beneficiary. Under the former construal, it occurs in postverbal position; under the latter construal, it can only precede the verb, as illustrated in (59): (59)

(a) 我給她剪頭髮。 (現代漢語) Wǒ gěi tā jiǎn tóufǎ. I for she cut hair “I cut her hair for her.” (b) *我剪給她頭髮。 *Wǒ jiǎn gěi tā tóufǎ. I cut for she hair (c) *我剪頭髮給她。 *Wǒ jiǎn tóufǎ gěi tā I cut hair for she (Contemporary Chinese)

That is, the principle of action–resultative ordering is actively at work in Contemporary Chinese, determining the distribution of preposition phrases in terms of their semantic features, whether resultative or not.

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Reduplication

265

In summary, the semantic diversity of the relationship between the verb and the object is much greater in Contemporary Chinese than in Old Chinese, and this change gradually took place throughout history. It resulted from the elimination of nonresultative preposition phrases in sentence-final position under the influence of the principle of action–resultative ordering.

10.2 Verb Reduplication As a major morphological device in the grammar of Contemporary Chinese, reduplication is manifest with all basic word classes, namely verbs, adjectives, adverbs, nouns, and classifiers. Reduplication is crucially related to all facets of the language, including phonology, morphology, lexicon, and syntax. It is critical to understand that the reduplication of different word classes was introduced into the language in significantly different historical periods. In one of the earliest texts – the Shi Jing, composed around the tenth century BC – reduplicated forms of adjectives, adverbs, and onomatopoeic expressions are already widely attested. The reduplication of nouns emerged around the third century BC and was taken over by classifiers after the sixth century AD, while a number of classifiers grammaticalized out of ordinary nouns. However, the reduplication of verbs did not exist until the thirteenth century AD, more than two millennia later than the reduplication of adjectives and adverbs. It can be said that the phonological system of the Chinese language favors morphological reduplication, which is closely related to the simple structure of its syllables, but whether and when a particular reduplicated form was applied to a particular word class depended largely on the syntax of the word class and the overall property of the grammatical system in the particular period. With verbs, the reduplicated form may be viewed as a type of aspect, which is termed the “delimitative aspect” and belongs to the grammatical category consisting of the perfect, experiential, and progressive aspects (Li and Thompson 1981: 232‒235). The meaning of verb reduplication is “doing the action a little bit or for a short period,” and it implies that the action is casual or for leisure (Zhu 1982: 66‒68). That is, the reduplicated part refers to a kind of state of the progression of the preceding verb, which obviously has nothing to do with the following object. Additionally, its application is subject to a phonological limitation – the number of syllables. The maximal number of syllables for verb reduplication is four. Specifically, a monosyllabic verb can be reduplicated into a disyllabic form, and a disyllabic verb can be reduplicated into a quadrisyllabic form. However, trisyllabic verbs (though quite rare in Chinese) do not have corresponding reduplicated forms because of the syllabic constraint. The reduplicated verb is phonologically reduced, usually losing the tone value, e.g. kàn-kan “take a look.” Among the three major aspects, only the perfective aspect -le

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

266

can be inserted between the base verb and the reduplicated verb, but the experiential and progressive aspects cannot be used in a reduplicated verb form. In standard Mandarin, the reduplicated verb actually occupies the syntactic position of the resultative; hence verb reduplication is incompatible with any other resultatives because there is a constraint in the grammar: one verb can be followed only by a single resultative element. The incompatibility of verb reduplication and the resultative construction hints at a diachronic relationship between its emergence and the establishment of the resultative construction, a point that we will discuss below. Like ordinary verbs, reduplicated transitive verbs can freely take an object, as illustrated in (60) and (61). (60)

他們昨天討論了討論這個問題。 (現代漢語) Tāmen zuótiān tǎolùn-le-tǎolùn zhè-gè wèntí. they yesterday discuss-PERF-discuss this-CL question “They discussed this question for a while yesterday.” (Contemporary Chinese)

(61)

我昨天晚上聽了聽音樂。 (現代漢語) Wǒ zuótiān wǎnshàng tīng-le-tīng yīnyuè. I yesterday night listen-PERF-listen music “I listened to music for a while yesterday night. (Contemporary Chinese)

If we examine it over the whole grammar, the reduplicated form is at best regarded as a morphological process across word classes. Reduplication of different word classes shares the following phonologic, semantic, and syntactic properties (for details, see Shi 2010: 109‒122). (a) The phonological commonality of reduplication. Regardless of which word class it applies to, the reduplication operation is always subject to the constraint on the number of syllables. For verbs and adjectives, the maximal reduplicated form can consist of four syllables, and for nouns and classifiers the maximal form is limited to two syllables, which means that only monosyllabic nouns and classifiers can be reduplicated. (b) The semantic commonality of reduplication. The common meaning of reduplicated forms, regardless of the word class, always refers to some kind of quantification. As we saw above, for verbs, the reduplicated form means doing an action “a little bit” or “for a short period,” indicating the times and lengths of the actions expressed by the verbs. For adjectives or adverbs, the reduplicated form functions to intensify the degree of the quality or property, functioning like a degree word “very” or “quite” to refer to a fuzzy degree of the related quality. For nouns and classifiers, the reduplicated

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Reduplication

(c)

267

form renders a universal reference, meaning “everyone in a definite domain” (for details, see Shi 2010: 109‒122). The syntactic commonality of reduplication. Regardless of the word class, once a base form is reduplicated, its grammatical status has been identified and is also restricted to a certain syntactic position. For example, reduplicated verbs can be used only as the matrix verb and can no longer be nominalized or used in an embedded clause. Reduplications of adjectives also cannot be nominalized, and when used as an attributive they must be associated with the nominal head by the structural particle de, while the base form can directly modify the head without any linker. Reduplicated nouns and classifiers must occur in preverbal position even when used as the object of the matrix verb and are incompatible with any modifier or determiner. Additionally, in factual statements (nonsubjective mood), no reduplicated form can be negated by any negative marker.

10.2.1 Conditions for the Emergence of Verb Reduplication As a morphological device, the emergence of verb reduplication was conditioned by both phonological and syntactic factors, which collectively determined when it could enter the language. As mentioned above, reduplication of adjectives, adverbs, and nouns was already widely used in Old Chinese. By definition, a reduplicated form is derived by replicating the base form once rightward. Thus word classes that are not followed by other syntactic constituents, such as adjectives and nouns, have the freedom to be reduplicated rightward. In other words, the syntactic templates in which adjectives and nouns typically occur have no limitation on their reduplication extension; hence their reduplication happened much earlier than that of verbs. However, the reduplication of transitive verbs has the following skeleton, which was constrained by the X syntactic position between the matrix verb and the object: (62)

V [X] O > Vi Vi O (the second Vi is a reduplicated verb)

Therefore the key factor in licensing the emergence of verb reduplication was the creation of the X position, which could be occupied by an element that described the state of the action denoted by the preceding verb but without bearing any grammatical relation to the following object. This condition clarifies why verb reduplication could not exist until the thirteenth century AD, when the resultative construction and aspect suffixes came into existence. As discussed in Chapter 4, before the third century BC no material could intervene between the verb and the object except for another transitive verb that

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

268

also governed the object argument. At the time, two transitive verbs could share an object, but they had to be connected by the conjunction ér, as illustrated in (63): (63)




吾得而食諸。 (論語 顏淵) Wú dé ér shí zhū. I get and eat it “I got and ate it.” (Lun Yu, Yan Yuan, 500 BC)

Due to the decline and disappearance of the conjunction ér in Late Old Chinese, two transitive verbs (maximally four verbs) could co-occur adjacently and share an object, which is known as “serial verb construction,” as illustrated in (64) and (65):




(64)

武士擊殺盜。 (史記 秦始皇本紀) Wǔshì jī shā dào. warrior strike kill bandit “The warrior struck and killed the bandit.” (Shi Ji, Qin Shi Huang Ben Ji, 100 BC)

(65)

遂禽殺蚩尤。 (史記 五帝本紀) Suì qín shā Chī Yóu. then capture kill Chi You “Then (Huang Di) captured Chi You and killed him.” (Shi Ji, Xiong Nu Lie Zhuan, 100 BC)




Under the operation of the above grammatical rule, before the sixth century AD, intransitive verbs and any other elements referring to the resultative of the action could appear only after the object, producing the VOR form (for a fuller discussion, see Chapter 6), as illustrated in (66): (66)




制街衢平直。 (世說新語 言語) Zhì jiēqú píng zhí. build street flat straight “(He) built the streets flat and straight.” (Shi Shuo Xin Yu, Yan Yu, 450 BC)

At this stage, only intransitive verbs could be reduplicated to express the duration or repetition of the action, as illustrated in (67) and (68), and could not be followed by any object. Note that the meaning of intransitive verbs at this stage was different from that of the reduplicated form of verbs (including transitive and intransitive): the former

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Reduplication

269

indicates the continuation or repetition of an action, and the latter refers to a short duration of the action, as mentioned above.




(67)

黃雀得飛飛。 (曹植 田野黃雀行) Huáng què dé fēi-fēi. yellow bird able fly-fly “The yellow birds can continuously fly.” (Cao Zhi, Tian Ye Huang Que Xing, AD 250)

(68)

行行道轉遠。 (謝惠連 西陵遇風獻康樂) Xíng-xíng dào zhuǎn yuǎn. walk-walk road become far “He kept walking and the road became long (he went a long distance).” (Xie Hui Lian, Xi Ling Yu Feng Xian Kang Le, AD 400)




The timing of the appearance of the reduplicated form of intransitive verbs exemplified above mainly determined the establishment of the serial verb construction. As discussed in Chapter 4, in Old Chinese there was no serial verb construction, and any two or more verbs within a clause had to be linked by the conjunction ér or qiè. In this grammatical system, no form of verb reduplication was possible because it structurally violated the grammatical rule. When the serial verb construction came into existence due to the disappearance of the former conjunctions, it became possible for more than one verb to occur adjacently. This change enabled the reduplicated form of intransitive verbs to emerge in the language because they did not involve any object. Thus we can see that grammatical development was like a chain in which an early change created a condition for a later development. Crucially, the emergence of verb reduplication in Contemporary Chinese was made possible by the creation of the syntactic position between the verb and the object. As we know, the establishment of the resultative construction occurred because a new syntactic position had been created between the verb and the object that could be occupied by an intransitive element referring back to the state of the progression of the preceding verb, without bearing any direct grammatical relation to the following object. As we discussed in Chapter 6, the verb and the resultative were fused into a compound-like unit in the context where the object was absent, under the influence of the disyllabification tendency and the frequency of co-occurrence. The “V + Xintr + O” structure was innovated at the expense of the old “Vtr + Vtr + O” described above, namely verb co-ordination. After the sixth century AD, in the spoken language the co-ordinate construction in which two transitive verbs shared an object became ungrammatical, e.g. *mǎi chī fàn “buy eat food.” More importantly, the formation of the aspect system directly motivated the emergence of verb reduplication. The aspect markers occurred between the verb and the

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

270

object, and all of them were actually suffixed to the preceding verb to form an immediate constituent. Semantically, the aspect suffixes described the state of the progression of the preceding verb and had the following structure: (V le/zhe/guo) + O. As we discussed in the sections above, the perfective aspect emerged around the tenth century AD, and the experiential and progressive aspects came into existence around the twelfth century AD. Not long after these events happened, verb reduplication entered the language, and the earliest examples were attested in texts composed around the thirteenth century AD, as illustrated in (69) and (70):




(69)

我試看看這來文咱。 (關漢卿 包待制三勘蝴蝶夢) Wǒ shì kàn-kàn zhè-lái wén zán. I try look-look this-CL article PRT “Let me take a look at this article.” (Guan Han Qing, Bao Dai Zhi San Kan Hu Die Meng, AD 1300)

(70)

你聞聞這書裡甚麼氣? (元代話本) Nǐ wén-wén zhè shū lǐ shénme qì? you smell-smell this book inside what scent “You smell this book and find out what scent it contains.” (Yuan Dai Hua Ben, AD 1350)

Although the condition became mature at the time, verb reduplication had its own developmental process, which involved another change in Medieval Chinese: the emergence of verb classifiers (Jiang and Cao 2005: 198‒225). Due to this change, the structure “Num + V” in Old Chinese became “V + Num + CL” after the sixth century AD. One common way to express the number of times of the action appearances was to add yī “one” between the base verb and the copied verb. In this case, the object was either absent, as illustrated in (71), or inserted between the verb and the “yī + the reduplicated verb” phrase, as illustrated in (72): (71)

試定精神看一看。 (朱子語類卷九) Shì dìng jīngshén kàn-yī-kàn. try calm spirit look-one-look “(He) tried to calm down and take a look at it.” (Zhu Zi Yu Lei, Chapter 9, AD 1200)

(72)

抬頭看了冉貴一看。 (元代話本) Tái-tóu kàn-le Rǎn Guì yī-kàn. raise-head look-PERF Ran Gui one-look “(He) raised his head and took a look at Ran Gui.” (Yuan Dai Hua Ben, AD 1350)

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Reduplication

271

In the above examples, the reduplicated part is a nominalized verb, similar to a verb classifier. It is plausible to assume that verb reduplication evolved out of the above context. Unlike lexical items that could be freely coined in the mind of language speakers, any grammatical device can be invented only in specific contexts even if it is favored or required by the overall properties of the grammatical system at the time. The diachronic relationship between verb reduplication and verb classifiers can be seen in the fact that verb reduplication does not exist in the southwestern dialects of Chinese; instead, it expresses the meaning of verb reduplication by attaching the verb classifier xià “one time” to the verb; e.g. kàn-xià shū “read CL book” in the Sichuan dialect is equal to kàn-kàn shū in Mandarin Chinese. Here we can see another chain of developments in the grammar: (73)

noun classifier > verb classifier > verb reduplication

That is, there were in fact two lines of change that worked together to innovate reduplicated verb form. In other words, the verb classifier itself might provide merely the proper context for the emergence of verb classifiers, but it was realized by the creation of the new syntactic position between the verb and the object, an outcome of the combined power of the establishment of the resultative construction and the emergence of the aspect system.

10.2.2 Verb Reduplication in Chinese Dialects The resultative construction developed at different rates in different dialects, which resulted in the greatly unbalanced development of verb reduplication across dialects. It has been best developed in the northern dialects (including standard Mandarin), and many southeastern dialects have been left behind. The criteria for judging the stages of the development of the resultative construction are whether the verb and the resultative can be separated by the object or adverbs and whether the dialect possesses the same aspect system as standard Mandarin. One extreme case is the southeastern dialects which simply lack verb reduplication, as pointed out above. Superficially, many of the southeastern dialects also possess verb reduplication, but its functions and scope are remarkably different from those of the reduplicated verb form in standard Mandarin, which can be classified into the following three situations. First, in many of the southeastern dialects, such as the Chenghai, Yongchun, Chaoyang, and Xiamen dialects, verb reduplication means “doing the action continuously, repeatedly, or entirely” and cannot be followed by any object; the patient noun must occur in front of the reduplicated verb (for details, see Huang 1996: 261‒267),

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

Verb Copying and Reduplication

272

namely between the subject and the predicate. The following example is from the Xiamen dialect: (74)

物件提提去。 (廈門方言) mĩʔ-kiɛn tʰe-tʰe kʰu. stuff take-take go “All the stuff has been taken away.” (Xiamen dialect)

Additionally, the Xiamen dialect does not have the aspect system of standard Mandarin, and its resultative construction still remains in the state of Medieval Chinese, which means that the verb and the resultative are often separated by the object. Second, as mentioned above, within a reduplicated form, the reduplicated verb actually occupies the position of the resultative; hence a reduplicated verb form cannot be added to another resultative in any type of sentence in Contemporary Chinese. However, in the Suzhou dialect, the reduplicated verb can be followed by another resultative item, especially in imperative sentences, as illustrated in (75): (75)

快點兒走走開。 (蘇州方言) kʰuɒ tiɪ tsɘɪ-tsɘɪ-kʰᴇ. quick bit go-go-apart “Please go away quickly.” (Suzhou dialect)

Third, as pointed out previously, in quite a few dialects, such as the Mengzi and Zhejiang dialects, verb reduplication has the meaning of the perfective aspect, but it still cannot be followed by an object, as illustrated in (76): (76)

飯食食再去吧。 (浙江吳方言) Va zei-zei tse tɕʰy pa. food eat- eat then leave “Please have a meal before you leave.” (Wenzhou dialect)

The division of labor between different aspects results from the long-term interaction of those aspect markers. During this process, there is often a mixture or confusion of different aspects, and even a certain mismatch has been fixed as a grammatical rule.

https://doi.org/10.1017/9781108921831.010 Published online by Cambridge University Press

11 The Comparative Construction

11.1 Introduction In general, grammatical constructions fall into two types: one type is universal across languages, such as passives and comparatives, and the other type is language-particular, such as the disposal and verb-copying constructions in Chinese. The first type of construction can be viewed as essential for human communication, and the second type is motivated by the particular grammatical system of a language. At the beginning of its documented history, the Chinese language already possessed the comparative structure, but since then it has undergone quite a few fundamental changes in structure and marking. As a result, the comparative structures in Contemporary Chinese appear unusual in terms of language correlation universals and do not fit into any patterns of language universals on the basis of a typological investigation. According to Greenberg (1966a) and Dryer (1991), the comparative structure of an SVO language, such as English, has the form in (1a), and the comparative structure of an SOV language, such as Japanese, has the form in (1b). (1)

(a) The comparative structure of SVO languages: Subj + Adj + (Marker Standard) (b) The comparative structure of SOV languages: Subj + (Standard Marker) + Adj

A sample of more than 700 languages shows that the correlation depicted in (1a) is almost universal with only one exception – Chinese. Although the Chinese language has remained an SVO language throughout its history, the comparative structure in Contemporary Chinese is “Subj + (Marker Standard) + Adj,” where the standard phrase precedes the adjectival predicate, fitting the pattern of an SOV language. However, the marker precedes the standard NP, resembling an SVO language, as illustrated in (2):

273

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

274 (2)

廣州比北京熱。 (現代漢語) Guǎngzhōu bǐ Běijīng rè. Guangzhou COMP Beijing hot “Guangzhou is hotter than Beijing.”

(Contemporary Chinese)

Cross-linguistically, the comparative marker is typically an adposition, and the relative order of the marker and the standard is consistent with the order of the verb and object. Therefore, within the standard phrase, the order is “Marker + Standard” in a VO language and “Standard + Marker” in an OV language. In the Chinese comparative structure, therefore, the constituent order of the standard phrase (i.e. preposition phrase) reflects that of a VO language. However, VO order also entails that the preposition phrase (i.e. the standard phrase) occurs in sentence-final position; hence the standard phrase as a preposition phrase tends to appear after the adjectival predicate in an SVO language. However, this universal correlation has one exception, the comparative structure of Contemporary Chinese (Dryer 1992, 2007). Nevertheless, the comparative structure of Old Chinese did perfectly fit the universal correlation of SVO languages described above, because it had the structure “Subj + Adj + (Marker Standard)” exactly as formulated in (1a). Let us consider two comparative examples from that time:




(3)

苛政猛於虎也。 (禮記 檀弓下) Kē Zhèng měng yú hǔ yě. severe policies fierce than tiger PRT “Severe political policies are fiercer than tigers.” (Li Ji, Chan Gong Xia, 450 BC)

(4)

霜葉紅於二月花。 (杜牧 山行) Shuāng yè hóng yú èr yuè huā. frost leave red than second month flower “Frost leaves are redder than flowers in February.” (Du Mu, Shan Xing, AD 850)




In this chapter, we explore why the comparative structure of Chinese developed from consistency to inconsistency with regard to the universal correlation defined in (1a).

11.2 Comparatives in Old Chinese To understand the development of the comparative structure, we must understand the overall property of the grammatical system in each particular period of its history. In Old Chinese, as an SVO language, sentence-final position was important because it could

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Comparatives in Old Chinese

275

contain numerous types of preposition phrase, such as those in the passive, instrumental, locative, and ditransitive constructions (see relevant chapters for details). Together, all of these types of clause structure formed the syntactic schema “Subj + V + O + PP,” which determined the concrete comparative structure of Old Chinese. In other words, the specific structure of comparison at the time was actually an instantiation of the main abstract syntactic schema in Old Chinese (for details, see Chapter 7). Additionally, as discussed in Section 7.9, due to the establishment of the principle of action–resultative ordering, all of the non-resultative preposition phrases disappeared from sentence-final position and a set of newly grammaticalized markers in preverbal position gradually took over the corresponding functions, a change that dramatically altered the original comparative structure. Comparative structures fall into several subtypes – superlative, superior, equal, and inferior – which have different syntactic schemas and have undergone different development paths, as discussed below. We saw the examples of the superior comparative above; here, more examples are provided to illustrate the other subtypes. Regardless of what the subtype of the comparative is, the syntactic schema at the time always restricted the standard phrase to sentence-final position: (5) (a)

(6)

Subj + Adj + (Marker Standard) Variants of the superior comparative in Old Chinese. At the time, the comparative marker yú could be used to introduce a standard to superior, superlative, equal, or inferior comparisons. In the superior comparison, other markers could be used to replace the preposition yú, such as guò in (6) and bèi in (7), a verb denoting “double the quality”:




而巧過其師。 (列子 湯問) Ér qiǎo guò qí shī. but clever than his teacher “But (he) is more clever than his teacher.” (Lie Zi, Tang Wen, 400 BC)

(7)




富倍季氏可也。 (左傳 襄公二十三年) Fù bèi Jì Shì kě-yě. rich double Ji Shi okay-PRT “It is acceptable that (he) is twice as rich as Ji Shi.” (Zuo Zhuan, Xiang Gong Er Shi San Nian, 550–400 BC)

Although the markings were different lexical items, they always occurred after the adjectival predicate. This comparative structure is still preserved in some southern dialects, such as Cantonese. In contrast, in Contemporary Chinese, the standard phrase can occur only prior to the predicate.

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

276

(b) The equal comparative. In Old Chinese, the superior and equal comparison shared the same structure and the same marker and could be distinguished from one another only on the basis of the meaning of the adjective in the construction. In the following examples, the adjectives are tóng “same” and móu “equal,” giving rise to an equal interpretation:




(8)

自同於季氏。 (左傳 昭公二十七年) Zì tóng yú Jì Shì. self same as Ji Shi “He is the same as Ji Shi.” (He himself agreed with Ji Shi.) (Zuo Zhuan, Zhao Gong Er Shi Qi Nian, 550–400 BC)

(9)

奉邑侔於諸侯。 (史記 趙世家) Fèng yì móu yú zhū-hóu. receive city equal to princes “The cities that he received are equal to those of the princes.” (Shi Ji, Zhao Shi Jia, 100 BC)




The preposition yú could also introduce the standard after the verb bǐ “compare” (which developed into a comparative marker in Modern Chinese, as discussed in the following section), referring to an equal comparison, as illustrated in (10): (10)




今而時大, 比於諸華。 (左傳 昭公三十年) Jīn ér shí dà, bǐ yú zhū huá. now and time large compare with various Chinese “Now they increased their territory and were comparable with the Chinese countries.” (Zuo Zhuan, Zhao Gong San Shi Nian, 550–400 BC)

After the Old Chinese period, the equal comparative construction employed its own markings, unlike the superior comparative. Instead of using the preposition yú, other prepositions grammaticalized in Medieval Chinese, mainly ruò, rú, and sì, which introduced a standard to an equal comparison in sentence-final position, as illustrated in (11) and (12): (11)




勇若孟賁。 (漢書 東方朔傳) Yǒng ruò Mèng Bēn. brave as Meng Ben “(He) is as brave as Meng Ben.” (Han Shu, Dong Fang Shuo Zhuan, AD 100)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Comparatives in Old Chinese

277




(12)

春來江水綠如藍。 (白居易 憶江南) Chūn lái jiāng shuǐ lǜ rú lán. Spring come river water green as indigo “In spring, the water of the Yangtze river is so green it is like indigo.” (Bai Ju Yi, Yi Jiang Nan, AD 850)

(c)

The inferior comparative. In Old Chinese, the inferior and superior comparisons also shared exactly the same structure and marking, and the interpretations were determined by the meaning of the adjective, as illustrated in (13):

(13)

師少於我。 (左傳 僖公十五年) Shī shǎo yú wǒ. troop less than ours “Their troop strength is less than ours.” (Zuo Zhuan, Xi Gong Shi Wu Nian, 550–400 BC)

(d)




The superlative construction in Old Chinese. At the time, a superlative comparison was made by adding the negative pronoun mò “none” to the shared comparative structure, as illustrated in (14) and (15):




(14)

蟲莫知於龍。 (左傳 昭公二十九年) Chóng mò zhì yú lóng. animal none clever than dragon “The dragon is the wisest of all animals.” (Zuo Zhuan, Zhao Gong Er Shi Jiu Nian, 550–400 BC)

(15)

禍莫大於不知足。 (韓非子 解老) Huò mò dà yú bù-zhī-zú. disaster none big than unsatisfied “No danger is worse than greed.”




(Han Fei Zi, Jie Lao, 300 BC) In Old Chinese, a superlative comparative expression could be achieved by using another analytic pattern, adding the negative pronoun to the equal comparative, as illustrated in (16) and (17): (16)




知臣莫若君。 (韓非子 外儲說) Zhī chén mò ruò jūn. known courtier none like monarch “The monarch knows his courtier best.” (Han Fei Zi, Wai Chu Shuo, 300 BC)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

278 (17)




無如季相。 (史記 高祖本紀) Wú rú Jì Xiāng. none like Ji Xiang “Ji Xiang is the best.” (Shi Ji, Gao Zu Ben Ji, 100 BC)

Of special interest is the fact that the expression of a “compare” action also made use of the preposition yú to introduce a standard noun in sentence-final position in Old Chinese, as illustrated in (18) and (19). Here, the matrix verb was bǐ “compare,” which grammaticalized into a comparative marker after the sixteenth century AD.




(18)

若將比予於文木邪? (莊子 人間世) Ruò jiāng bǐ yú yú Wén Mù yé? you will compare I with Wen Mu PRT “Would you compare me with Wen Mu?” (Zhuang Zi, Ren Jian Shi, 300 BC)

(19)

爾何曾比予於管仲? (孟子 公孫醜) Ěr hécéng bǐ yǔ yú Guǎn Zhòng? you ever compare I with Guan Zhong “Did you ever compare me with Guan Zhong?” (Meng Zi, Gong Sun Chou, 300 BC)




The above examples share the same schema with the locative, passive, and instrumental, with the preposition phrase occurring in postverbal position, as mentioned elsewhere. In the Old Chinese equal comparative, the standard phrase could be introduced by the comitative preposition yǔ “with” prior to the adjectival predicate, namely in preverbal position, as illustrated in (20) and (21):




(20)

我諸戎飲食衣服不與華同。 (左傳 襄公十四年) Wǒ zhū róng yǐnshí yīfú bù yǔ Huá tóng. I many Rong food clothes not with Chinese same “The food and clothes of our Rong people are not the same as those of Chinese.” (Zuo Zhuan, Xiang Gong Shi Si Nian, 550–400 BC)

(21)

西北有高樓, 上與浮雲齊。 (古詩十九首 西北有高樓) Xī-běi yǒu gāo-lóu, shàng yǔ fúyún qí. west-north have high-building top with cloud equal “In the northwest is a high building whose top is as high as the clouds.” (Gu Shi Shi Jiu shou, Xi Bei You Gao Lou, AD 100)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press




Comparatives in Modern Chinese

279

Note that only in the equal comparative could the preposition phrase (i.e. standard phrase) occur in preverbal position at the time.

11.3 Comparatives in Modern Chinese All of the subtypes of comparative structures underwent the following change during the transitional period from Middle to Modern Chinese: (22)

Subj + Adj + (Marker Standard) > Subj + (Marker Standard) + Adj

From the surface structures of the above formulas, it seems that the phrase “Marker + Standard” moved from sentence-final position to the left periphery of the adjectival predicate. However, there was actually no movement; rather, the new markers that grammaticalized from a lexical item in preverbal position took over the functions of the old comparative markings. The changes in the comparative structure were motivated by the information structure principle of the predicate (i.e. the principle of action–resultative ordering), which required that no non-resultative preposition phrases could appear in postverbal position; non-resultative prepositional phrases could only appear in preverbal position. As discussed in Section 7.9, under the operation of this principle, the preposition phrases in the passive, locative, and ditransitive constructions underwent a parallel change: they all disappeared from sentence-final position, and their functions were taken over by newly grammaticalized morphemes in preverbal position. However, these changes did not take place at the same rate, and the constructions that expressed high transitivity, including the passive, locative, instrumental, and ditransitive, first underwent the change in the period from the first century BC to the fourth century AD. The reason is that a highly transitive verb and its following constituent represent the stereotype of the relationship between the action and the result. In contrast, the predicate of the comparative structure was usually an adjective, and the relationship between the adjectival predicate and the preposition phrase that introduced a standard was relatively weak. Thus the preposition phrase in the comparative changed nearly one thousand years later than the other clause structures. Even in Late Medieval Chinese, the comparative structure of Old Chinese was still well preserved and was the paradigmatic pattern, as illustrated below (Ohta 1987: 168, Huang 1992): (23)

昔日貧於我。 (寒山詩) Xīrì pín yú wǒ. past poor than I “In the past, (he) was poorer than I.” (Han Shan Shi, AD 650)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

280




(24)

霜叶红于二月花。 (杜牧 山行) Shuāng yè hóng yú èr-yuè huā. frost leave red than second-month flower “The frosted leaves are redder than the flowers in February.” (Du Mu, Shan Xing, AD 850)

(25)

綠苔狂似人。 (貫休 寄令狐郎中) Lǜ tái kuáng sì rén. green moss wild like people “The green moss is wild like people.” (Guan Xiu, Ji Ling Hu Lang Zhong, AD 900)




That is, the comparative construction of Old Chinese survived until the ninth century AD. The collapse of the old comparative structure was caused partially by the decline of the preposition yú. This functional word had numerous usages in Old Chinese; for example, it served to introduce a locative for a declarative sentence and the agent of a passive, but these uses ceased after the first century BC. Due to the decline of yú, some instances of the comparative structure became unmarked after the first century BC, as illustrated in (26) and (27):




(26)

德厚侔天地。 (史記 孝文帝本紀) Dé hòu móu [ ] tiān dì. virtue thickness equal sky earth “The thickness of virtue is equal to heaven and earth.” (Shi Ji, Xiao Wen Di Ben Ji, 100 BC)

(27)

所斬捕功已多大將軍。 (史記 衛將軍驃騎列傳) Suǒ zhǎn bǔ gōng yǐ duō [ ] dà jiàngjūn. NOM kill capture credit already more big general “His credits for killing and capture (in the war) were already more than those of your great general.” (Shi Ji, Wei Jiang Jun Piao Qi Lie Zhuan, 100 BC)




In Old Chinese, however, the square brackets could not be empty but needed to be occupied by the preposition yú. As will be discussed below, the omission of yú from the comparative structure created a necessary context in which the verb bǐ “compare” could be grammaticalized into a comparative marker in the first verb position. There was a gap in the development process of the comparative construction, from roughly the eighth century AD to the fourteenth. When the marker yú was gradually abandoned, no new grammatical morpheme was immediately ready to

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Comparatives in Modern Chinese

281

mark the superior comparison at the time. Therefore, interestingly, the equal comparative markers rú, ruò, and sì were borrowed to mark the superior structure. These three markers grammaticalized from verbs meaning “resemble” or “like,” which supposedly marked the equal comparative structure according to their semantic suitability. After Late Old Chinese, however, they started to mark the equal comparison, as illustrated below:




(28)

小人之交甘若醴。 (莊子 山木) Xiǎorén zhī jiāo gān ruò lǐ. villain GEN friendship sweet like rice-wine “The friendship of villains is sweet like rice wine.” (Zhuang Zi, Shan Mu, 300 BC)

(29)

夏侯初朗朗如日月之入懷。 (世說新語 容止) Xiàhóu Chū lǎnglǎng rú rì yuè zhī rù huái. Xiahou Chu optimistic like sun moon GEN enter mind “Xiahou Chu is optimistic, as if the sun and moon enter his mind.” (Shi Shuo Xin Yu, Rong Zhi, AD 450)




Ohta (1987: 168) first noticed that the equal and superior comparatives became mixed with the superior comparison in this period, as illustrated below: (30)

(a) 直如富過石崇家。 (十二時) Zhí rú fù guò Shí Chóng jiā. just likely rich than Shi Chong family “It was just likely that (he) was richer than Shi Chong’s family.” (Shi Er Shi, AD 800) (b) 更饒富似石崇家。 (十二時) Gèng ráofù sì Shí Chóng jiā. even rich than Shi Chong family “(He) was even richer than the family of Shi Chong.” (Shi Er Shi, AD 800)

Example (30b) is ambiguous because it could be interpreted as either superior or equal. Only when we compare it with (30a) from the same text can we know that (30b) was actually also a superior comparative because the marker guò in the first clause could express only the superior. In other words, the markers guò (always a superior marking) and sì (originally an equal marking) were exchangeable, which tells us that (30b) actually expresses a superior comparison. A similar situation is found in the following examples:

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

282 (31)




(a) 本寺遠於日。 (姚合 贈供奉僧次融) Běn sì yuǎn yú rì. this temple far than sun “The temple is farther than the sun.” (b) 新詩高似雲。 Xīn shī gāo sì yún. new poem high than cloud “The new poem was higher than the clouds.” (Yao He, Zeng Gong Feng Seng Ci Rong, AD 800)

The above two clauses, both of which were rhetorical exaggerations, occurred adjacently in the same poem. The first is the traditional superior comparison marked by yú, from which we know that the original equal comparative sì was actually used as a superior marking.

11.4 The Comparative Marker Bǐ As the grammatical system developed, the old comparative structure could not remain unchanged for too long and had to adapt to the new consistency across grammatical constructions that resulted from the principle of action–resultative ordering. As discussed above, when the preposition yú finally died out, the first change in the superior comparative was to merge with the equal comparative structure. However, this mixture of two different subtypes of comparison often caused ambiguities in communication and hence did not serve communicative purposes well. In this situation, a new grammatical superior marking that was structurally consistent with the contemporary grammatical system was favored. The best candidate for marking the superior comparative was the verb bǐ “compare,” which was already a common verb in Old Chinese but was not grammaticalized into a comparative morpheme until the seventh century AD. As discussed at the beginning of this chapter, the superior comparative structure of Contemporary Chinese is formulated as follows: (32)

Subj + (bǐ Standard) + Adj.

Thus the proper context for bǐ “compare” to be grammaticalized was its use as the first verb within a serial verb construction. The formation of this context underwent several key steps that took more than a thousand years. Once again, the development of the comparative structure demonstrates that the emergence of any new grammatical morpheme must be conditioned by two factors: the requirement of the grammatical system and the suitable context for grammaticalization. The grammaticalization process of the superior structure is outlined below.

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Marker Bǐ

283

In Old Chinese, bǐ “compare” was a content verb and was often used as the matrix verb with the preposition yú to introduce a standard noun, referring to an action of comparing two entities, as mentioned above. More examples are illustrated in (33) and (34).




(33)

爾何曾比予於管仲? (孟子 公孫醜) Ěr hécéng bǐ yú yú Guǎn Zhòng? you ever compare I with Guang Zhong “Did you ever compare me with Guang Zhong?” (Meng Zi, Gong Sun Chou, 300 BC)

(34)

若將比予於文木邪? (莊子 人間世) Ruò jiāng bǐ yú yú Wén Mù yé? you will compare I with Wen Mu PRT “Will you compare me with Wen Mu?” (Zhuang Zi, Ren Jian Shi, 300 BC)




The instances of the above type functioned to express that the two entities were comparable in certain dimensions of property or quality. This structure could only be used as an independent sentence, which meant that the verb bǐ could not occur in the first verb position of a serial verb construction. At the time, the object (i.e. the standard noun) of the verb bǐ could be omitted, as illustrated in (35): (35)




上比於春秋。 (韓非子 奸劫弑臣) Shàng bǐ [ ] yú Chūn Qiū. past compare with Chun Qiu “It could be compared with Chun Qiu in the past.” (Han Fei Zi, Jian Jie Shi Chen, 300 BC)

A key change after the first century BC in the sentence with the matrix verb bǐ was the possible deletion of the preposition yú, which enabled bǐ to occur in the first verb position of a serial verb construction. Initially, the verb bǐ “compare” was used as the matrix verb of the first clause, and the following clause indicated the parameters of the comparison (what they were being compared in terms of), as illustrated in (36) and (37). This comparative structure was very common in Medieval Chinese. (36)




王比使君, 田舍貴人耳! (世說新語 品藻) Wáng bǐ Shǐ Jūn, tiánshè guì-rén ěr! Wang compare Shi Jun country side noble-man PRT “Compared with Shi Jun, you are a noble man in the countryside.” (Shi Shuo Xin Yu, Pin Zao, AD 450)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

284 (37)




阿奴比丞相, 但有都長。 (世說新語 品藻) Ā Nú bǐ Chéngxiàng, dàn yǒu dōuzhǎng. A Nu compare prime-minister only have handsome “Compared with the prime minister, you are just more handsome.” (Shi Shuo Xin Yu, Pin Zao, AD 450)

However, in the following examples, the verb bǐ “compare” was actually used as the first verb of a serial verb construction, there was no pause between the two adjacent verb phrases, and the second VP was typically an adjectival phrase, as illustrated in (38) and (39):




(38)

馬比死多少? (世說新語 簡傲) Mǎ bǐ sǐ duō-shǎo? horse compare death how-many “Compared with the deaths of people, how many horses died?” (Shi Shuo Xin Yu, Jian Ao, AD 450)

(39)

官職比君雖較小。 (白居易 先贈長句) Guān zhí bǐ jūn suī jiào xiǎo. official rank compare you though relatively low “Compared with your official rank, mine is relatively low, though.” (Bai Ju Yi, Xian Zeng Chang Ju, AD 850)




Bǐ in the above examples was still an ordinary verb, and the adjectives could be modified by degree words, e.g. jiào “relatively,” which is disallowed in the superior comparative structure in Contemporary Chinese. Even in the texts composed in the thirteenth century AD, the bǐ comparative structure was not firmly established and could refer to any comparative subtype, such as the superlative in (40), the superior in (41), and the equal in (42). Once established, it could refer only to a superior comparison.




(40)

(咱兄弟) 比外人至親熟。 (元刊雜劇 死生交范張雞黍) (Zán xiōngdì) bǐ wàirén zhìqīn shú. our brother compare outsiders most intimate “Compared with outsiders, our brothers are most intimate.” (Yuan Kan Za Ju, Si Sheng Jiao Fan Zhang Ji Shu, AD 1300)

(41)

(這橋) 比在前十分好。 (老乞大) (Zhè qiáo) bǐ zàiqián shífēn hǎo. this bridge compare before extremely good “Compared with before, this bridge is extremely good.” (Lao Qi Da, AD 1300)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Marker Bǐ (42)

285

(官司) 比咱們這裡一般嚴。 (老乞大) (Guānsī) bǐ zánmen zhèlǐ yībān yán. law compare our place equally strict “Compared with our place, their law is equally restrictive.” (Lao Qi Da, AD 1300)

In the above examples, bǐ refers to an action of comparison that is followed by a specific evaluation of the related quality, which was typically indicated by a “degree word + Adj” phrase. It was in this context that the verb bǐ became grammaticalized into a superior comparison marker. Not until the seventeenth century AD was the comparative morpheme bǐ specialized to express only the superior comparison, a sign that it was truly grammaticalized, as illustrated in (43) and (44): (43)

這些姐姐們, 再沒有一個比寶姐姐好的。 (紅樓夢三十一回) Zhèxiē jiějiěmen, zài méiyǒu yī-gè bǐ Bǎo jiějiě hǎo These sister-PLU again not-have one-CL COMP Bao sister good de. PRT “None of these sisters is better than Sister Bao.” (Hong Lou Meng, Chapter 31, AD 1750)

(44)

這個又比那個亮。 (紅樓夢四十五回) Zhè-gè yòu bǐ nàgè liàng. This-CL also COMP that-CL bright “This one is brighter than that one.” (Hong Lou Meng, Chapter 45, AD 1750)

Therefore it is safe to say that the superior comparative construction marked by bǐ was firmly established in the eighteenth century AD. Additionally, at this time, the system of the comparative structures of Contemporary Chinese was developed, and clear divisions emerged among the following four subtypes of comparative: (45)

(a) (b) (c) (d)

Superior comparative: Subj + (bǐ Standard) + Adj Inferior comparative: Subj + (měi Standard) + Adj Superlative comparative: Subj + zuì + Adj Equal comparative: Subj + (xiàng Standard yīyàng) + Adj

In contrast to Old Chinese, in which all the standard phrases were restricted to sentence-final position, in the four structures described above, the standard phrases occur before the predicate, although their markings are lexically different.

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

286

In the eighteenth century AD, the morpheme zuì that specially marked the superlative comparative had already emerged, as illustrated in (46) and (47): (46)

他回答的最妙。 (紅樓夢二回) Tā huídá de zuì miào. he answer NOM most clever “His answer is cleverest.” (Hong Lou Meng, Chapter 2, AD 1750)

(47)

這賈政最喜讀書人。 (紅樓夢三回) Zhè Jiǎ Zhèng zuì xǐ dúshū rén. this Jia Zheng most love study people “Jia Zheng most loves intellectuals.” (Hong Lou Meng, Chapter 3, AD 1750)

Additionally, the pattern of the equal comparative “PREP . . . yīyàng (same),” where the preposition could be hé “with” or sì “like,” is illustrated in (48) and (49): (48)

老太太也是和鳳姐姐的一樣看待。 (紅樓夢三十五回) Lǎo tàitài yě shì hé Fèng Jiějiě de yīyàng kàndài. old lady also be with Feng sister GEN same treat “The grandmother treats her the same as Sister Feng.” (Hong Lou Meng, Chapter 35, AD 1750)

(49)

身子就像在雲端裡一般。 (紅樓夢六回) Shēnzi jiù xiàng zài yúnduān lǐ yībān. Body just like in cloud inside same “It is like the body in the cloud.” (Hong Lou Meng, Chapter 6, AD 1750)

As will be discussed in Section 14.4.4, the negative méi developed out of its original meaning of “disappearing in water” around the eighth century AD. In the earliest stage, it could only negate nominal phrases meaning “not have,” and it was later extended to negating verb phrases, especially bounded predicates. Around the eighteenth century AD, it developed into a marker for inferior comparison, a structure that was previously lacking, as illustrated in (50): (50)

還沒有咱們這一半大。 (現代漢語) Hái méiyǒu zánmen zhè yībàn dà. still COMP our this half big “It is still not half as big as ours.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Comparatives in Contemporary Chinese

287

This negative verb was used to form a subtype of the comparative, especially to make an inferior comparison. This pattern is equal to the “less Adj than” construction in English. Thus Chinese acquired two structurally distinct constructions to express superior and inferior comparatives, which can be schematized as follows: (51)

(a) The superior comparative construction: Subj + (bǐ Standard) + Adj (b) The inferior comparative construction: Subj + (méi Standard) + Adj

There is a transformation between the two subtypes of the comparative: “X + (bǐ Y) + Adj” > “Y + (méi X) + Adj,” as illustrated in (52). (52)

(a) The inferior comparison: 這道題沒有那道題難。 (現代漢語) Zhè dào tí méiyǒu nà-dào tí nán. This-CL question not-have that-CL question difficult “This question is not as difficult as that one.” (b) The superior comparison: 那道題比這道題難。 Nà-dào tí bǐ zhè-dào tí nán. That-CL question COMP this-CL question difficult “That question is more difficult than this one.” (Contemporary Chinese)

Apparently, the structures and markings of comparison in Contemporary Chinese are much more sophisticated than they were previously. The functions of the markings are more specialized; for example, bǐ can express only the superior, zuì can refer only to the superlative, and méi can represent only the inferior. Additionally, the divisions of labor among the subtypes of comparative structure are neat and fixed.

11.5 Comparatives in Contemporary Chinese Although comparative structures exist in most languages, the subtypes of the structures, and especially the comparable constituents, vary from language to language. Let us consider the difference between Chinese and English. In English, the degree of two adverbs can be compared by means of the paradigmatic structure marked by than, for example: (53)

(a) Mary studies harder than her roommates. (b) Tom runs faster than John.

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

288

In contrast, throughout its history, the comparison between two adverbs has been impossible in Chinese. However, the establishment of the resultative construction provided an alternative for making a comparison between two adverbs. The method is to turn the adverb into an adjectival resultative by inserting the resultative particle de, which appears in variants of the comparative structure, as formalized in (54): (54)

(a) Subj + (bǐ Standard) + (V-de R) (b) Subj + V-de (bǐ Standard) + R

Examples are given in (55) and (56). (55)

(a) 汤姆比约翰跑得快。 (現代漢語) Tāngmǔ bǐ Yuēhàn pǎo de kuài. Tom COMP John run DE fast “Tom runs faster than John.” (b) 汤姆跑得比约翰快。 Tāngmǔ pǎo de bǐ yuēhàn kuài. Tom run DE COMP John fast “Tom runs faster than John.” (Contemporary Chinese)

(56)

(a) 我比他来得早。 (現代漢語) Wǒ bǐ tā lái de zǎo. I COMP he come DE early “I came earlier than he.” (c) 我来得比他早。 Wǒ lái de bǐ tā I come DE COMP he “I came earlier than he.”

zǎo. early (Contemporary Chinese)

That is, the emergence of the resultative construction made it possible for adverbs to be used in the comparative construction by turning them into resultatives. The above usage emerged quite recently, roughly around the seventeenth century AD. Before that time, the predicate of the comparative structure could only be an adjective and the two entities that were compared were limited to the static properties of nouns. However, in the above examples, the compared entities are actually the action in question. This means that the function of the superior comparative was extended to express gradable actions; consequently, the frequency of the usages of comparative instances increased remarkably.

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Comparatives in Contemporary Chinese

289

A subtle change in comparative structures deserves special attention because it reflects the general development of the grammatical system. When comparing two qualities, the degree of difference often needs to be specified. In English, the quantifier or intensifier precedes the adjective as in normal declarative sentences, e.g. much in John is much richer than Tom and three years in James is three years older than Mary. Similarly, before the seventeenth century AD, these sorts of degree words appeared before the adjective in the adverbial phrase to modify the predicate adjective, as illustrated in (57): (57)




官職比君雖較小。 (白居易 先贈長句) Guān zhí bǐ jūn suī jiào xiǎo. official rank COMP you though relatively low “Compared with your official rank, mine is relatively low, though.” (Bai Ju Yi, Xian Zeng Chang Ju, AD 850)

In Contemporary Chinese, however, the intensifiers can be used only as a resultative following the predicate adjective in the comparative structure, where they are treated as a kind of result of the comparison, as illustrated in (58) and (59): (58)

(a) 我的官位比他的低得多。 (現代漢語) Wǒ de guān-wèi bǐ tā-de dī de duō. I GEN official-rank COM he-GEN low DE much “My position is much lower than his.” (b) *我的官位比他多低。 *Wǒ de guānwèi bǐ tā duō dī. I GEN official-rank COMP he much low (Contemporary Chinese)

(59)

(a) 我比他大三歲。 (現代漢語) Wǒ bǐ tā dà sān suì. I COMP he old three year “I am three years older than he.” (b) *我比他三歲大。 Wǒ bǐ tā sān suì dà. I COMP he three year old (Contemporary Chinese)

As (58b) and (59b) show, the predicate adjectives disallow precedence by any modifiers; otherwise, an ill-formed structure will be created. The above comparative structure that is specified with a quantifier became possible only after the establishment of the resultative construction.

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

290

When a grammatical construction has been innovated, both its token and its type frequencies increase noticeably. Meanwhile, the innovated construction may lead to the development of other variant structures. After the resultative construction was firmly established around the tenth century AD, many subtypes of the resultative construction were created by analogy. Here we discuss a subtype of the resultative construction. In the spoken language of Contemporary Chinese, there is a highly productive resultative construction that entered the language only in the past century, as schematized in (60): (60)

Subj + V + O (Quantifier Resultative).

The phrase “quantifier resultative” must be structurally complex; hence, unlike other verb–resultative phrases, the object occurs between the verb and the resultative. This configuration is determined by a pragmatic factor, the weight of the constituent, rather than the remaining feature of the former resultative structure in Medieval Chinese (for details, see Section 6.4). The verb can be any action verb that does not govern the object; the resultative must be overtly present, combining with the verb to govern the object. The function of this type means that “the action causes the object to lose some quantity of a thing (e.g. money, time),” as illustrated in (61) and (62): (61)

這頓飯喫了我八十美元。(現代漢語) Zhè-dùn fàn chī-le wǒ bāshí měi-yuán. This-CL meal eat-PERF I eighty US-dollar This meal cost me eighty dollars. (Contemporary Chinese)

(62)

這封信寫了我一個晚上。 (現代漢語) wǎnshàng. Zhè-fēng xìn xiě-le wǒ yīgè This-CL letter write-PERF I one-CL night “I took all night to write this letter.” (Contemporary Chinese)

In Chinese, verbs and adjectives are syntactically similar; for example, these two word classes could both be directly used as the predicate. Thus the adjective can also be combined with a quantifier to form a resultative construction. This construction can assign an extra patient argument that occurs between the adjective and quantifier. Note that the quantifier is obligatory to make the structure grammatical, as illustrated in (63) and (64): (63)

(a) 約翰大湯姆三歲。 (現代漢語) Yuēhàn dà Tāngmǔ sān suì. John old Tom three year “John is three years older than Tom.”

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Suffixes of Adjectives and Adverbs

291

(b) *約翰大湯姆。 *Yuēhàn dà Tāngmǔ. John old Tom (Contemporary Chinese) (64)

(a) 我高他一頭。(現代漢語) Wǒ gāo tā yī tóu. I tall he one head “I am one head taller than he.” (b) *我高他。 *Wǒ gāo tā. I tall him (Contemporary Chinese)

In Medieval Chinese, the above expressions were created by the following construction: “Subj + Adj + (Marker Standard) + quantifier,” where the preposition yú or guò had to be present, as illustrated in (65): (65)




(a) 貧於楊子兩三倍。 (白居易 送劉五司馬) Pín yú Yáng Zi liǎng sān bèi. poor than Yang Zi two three times “(He) is two or three times poorer than Yang Zi.” (b) 老過榮公六七年。 Lǎo guò Róng Gōng liù-qī nián. old exceed Rong Gong six-seven year “I am six or seven years older than Rong Gong.” (Bai Ju Yi, Song Liu Wu Si Ma, AD 850)

At first glance, the sole difference among the above structures of Medieval and Contemporary Chinese is whether the preposition yú or guò was used. In reality, however, the structural hierarchies have fundamentally changed. In the structure of Medieval Chinese, it was a paradigmatic comparative structure after the quantifier was deleted, but it is ungrammatical without the quantifier in Contemporary Chinese. The comparative pattern “Adj . . . Quantifier” was an outcome of the further development of the resultative construction.

11.6 Suffixes of Adjectives and Adverbs As discussed above, there was a fundamental change in comparative structures: before the tenth century AD, standard phrases were always introduced by a preposition in the

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

292

postpredicate position; then these phrases had to precede the predicate. These changes can be formalized as follows: (66) (a) Subj + Adj + (PREP Standard) > (b) Subj + (PREP Standard) + Adj Historically, there was a particular context in which the prepositions in (66a) grammaticalized into suffixes for adjectives or adverbs through analysis. At the time, the adjective and the following preposition occurred adjacently, which enabled them to fuse into a single constituent via deleting the standard NP and weakening the constituent boundary. Over time, however, the reanalysis became impossible because the preposition had to precede the adjective instead and was also separated from the adjective by the standard NP. Therefore Old Chinese always had a set of highly productive suffixes for adjectives and adverbs that developed out of the prepositions of the equal comparative, namely rú, ruò, and sì, which meant “like” or “resemble,” as illustrated in the following examples: (67)




其葉沃若。 (詩經 碩人) Qí yè wò-ruò. its leave fertile-like “Its leaves are somewhat lush.” (Shi Jing, Shuo Ren, 1000–600 BC)

(68)




與士大夫言, 侃侃如也。 (論語 鄉黨) Yǔ shìdàfū yán, kǎnkǎn-rú yě. with bureaucrat speak eloquent-like PRT “He is quite eloquent when speaking with bureaucrats.” (Lun Yu, Xiang Dang, 500 BC)

English does not have equivalents to the above adjectival suffixes that function to highlight the vividness of the quality or state or to intensify the degree of quality. As we discussed in the previous section, the prepositions rú and ruò “like” or “as” could occur only after the predicate adjective. In Old Chinese, these suffixes could also function to turn a noun or verb phrase into an adjective-like phrase, as illustrated below:




(69)

趣進, 翼如也。 (論語 鄉黨) Qù jìn, yì-rú yě. quick come-in feather-like PRT “(Confucius) quickly came in like a feather.” (Lun Yu, Xiang Dang, 500 BC)

(70)

升堂, 鞠躬如也。 (論語 鄉黨) Shēng táng, jūgōng-rú yě. Walk-up lobby bow-like PRT “(Confucius) walked up to the lobby in a bowing-like manner.” (Lun Yu, Xiang Dang, 500 BC)




https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Suffixes of Adjectives and Adverbs

293

In (69), the adjectival suffix turns the noun “feather” into an adjective and in (70) it turns the verb “bow” into an adjective. From the perspective of the grammar of Contemporary Chinese, the structures in Old Chinese appear unusual and seem oddly incompatible with the grammatical system. For example, the preposition sì “like,” which was often preceded by the sentence-final particle yě, could form an “NP + (yě-)sì” phrase to modify a noun or verb phrase, a very common usage from the tenth century AD to the fourteenth, as illustrated in (71) and (72): (71)

把山海似深恩掉在腦後。 (董西廂) Bǎ shān-hǎi sì shēn-ēn diào zài nǎo hòu. DISP mountain-sea like deep-kindness throw in mind back “(He) forgot the profound kindness like the mountains and the sea.” (Dong Xi Xiang, AD 1200)

(72)

正熟睡呢, 傾盆也似雨降。 (劉知元諸宮調) Zhèng shú-shuì ne, qīng pén yě-shì yǔ jiàng. right sound-sleep PRT pour basin YE-like rain down “Just as I was soundly sleeping, it rained like an overturned basin.” (Liu Zhi Yuan Zhu Gong Diao, AD 1200)

The structure of the “NP + sì” phrase looks like a “NP + PREP” phrase, which could hardly develop out of a VO language as it did in Chinese. Therefore Jiang (1999b) speculates that the structure might be a result of language contact with Mongolian, an SOV language that was spoken by the nation on the northern boundary of China at the time. However, like rú or ruò, the preposition sì historically occurred after the predicate adjective to introduce an adverbial phrase, as illustrated in (73), which indeed provided a proper context for it to become a suffix for an adjective or adverb. In the above two examples, the suffix -sì actually functions to turn an NP or VP into an adjective-like or adverb-like phrase to modify the head verb or noun. (73)

大似一頭驢。 (鎮州臨濟慧照禪師語錄) Dà sì yītóu lǘ. big like one-CL donkey “It is as big as a donkey.” (Zhen Zhou Lin Ji Hui Zhao Chan Shi Yu Lu, AD 800)

Consider the compound suffix yě-sì, where the first morpheme is the sentence-final particle yě and the second particle is the preposition sì. Their reanalysis was most likely to happen in the context illustrated below:

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

The Comparative Construction

294 (74)




其道聽途說也似學。 (抱樸子 名實) Qí dào-tīng-tú-shuō yě-sì xué. his hearsay YE-like learn His hearsay also seems to be learned. (Bao Pu Zi, Ming Shi, AD 350)

(75)




不覺蒼華也似絲。 (徐鉉 柳枝詞) Bù jué cāng-huá yě-sì sī. not notice black-hair YE-like silk “(I) didn’t notice that my black hair is like (white) silk.” (Xu Xuan, Liu Zhi Ci, AD 950)

A plausible explanation is that, through reanalysis, the objects of the preposition sì became the heads, and the original heads prior to yě-sì were viewed as the modifiers. The word class of the gerund form xué “study” was construed as a verb. In addition, this analytical process involved a boundary change from the boundary between yě and sì to that between yě-sì and the head, where yě-sì was combined with the preceding VP in (74) and NP in (75) to form a modifier. A similar case is the suffix rán in Old Chinese. In Old Chinese, the suffix rán was a high-frequency grammatical morpheme that turned a noun or verb into an adjectivelike item, as illustrated in (76), or was attached to an adjective in the predicate or to an adverbial phrase to modify the predicate, as illustrated in (77): (76)




望之儼然。 (論語 微子) Wàng zhī yǎn-rán. watch him dignified-like “He looks dignified.” (Lun Yu, Wei Zi, 500 BC)

(77)




夫子循循然善誘人。(論語 子罕) Fūzǐ xúnxún-rán shàn yòu rén. Confucius patient-like good inspire people “Confucius is patiently good at inspiring other people.” (Lun Yu, Zi Han, 500 BC)

At the time, rán was a demonstrative that referred back to a VP or AP, similar to so in English, as illustrated in (78) and (79): (78)




樂然後笑。 (論語 憲問) Lè rán-hòu xiào. happy so-after laugh “When he was happy, he smiled.” (Lun Yu, Xian Wen, 500 BC)

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

Suffixes of Adjectives and Adverbs (79)

295




歲寒, 然後知松柏之後雕也。 (論語 子罕) Suì hán, rán-hòu zhī sōngbǎi zhī hòu diāo yě. season cold so-after know cypress GEN late wither PRT “When the season is cold, we realize that cypress withers latest.” (Lun Yu, Zi Han, 500 BC)

However, due to systematic changes in history, after the beginning of the Modern Chinese period, the above adjectival or adverbial suffixes gradually faded from use in the spoken language. Instead, many reduplicated suffixes, which function to turn verbs or nouns into adjectives or adverbs, were innovated in the language; these reduplicated forms can also be suffixed to an adjective to intensify the degree of the quality represented (for a detailed discussion of the development of reduplication, see Section 10.2). (80)

Noun + reduplicated suffix 血淋淋 xuè-línlín “bloody” 泪汪汪 lèi-wāngwāng “tearful” 雾蒙蒙 wù-méngméng “misty”

(81)

Verb + reduplicated suffix 笑盈盈 xiào-yíngyíng “smilingly” 哭啼啼 kū-títí “cryingly” 鬧哄哄 nào-hōnghōng “noisily”

(82)

Adjective + reduplicated suffix 白花花 bái-huāhuā “shiningly white” 紅彤彤 hóng-tōngtōng “brightly red” 綠莹莹 lǜ yíngyíng “brightly green”

The above schema “X + reduplicated suffix” is highly productive and widely used in the spoken language, but it did not exist before Modern Chinese. These reduplicated suffixes are idiosyncratic and lexically constrained; for example, -huāhuā can be collocated only with the single adjective bái “white,” and -línlín can be combined only with two nouns – xuè “bloody” and yǔ “rainy.” Their meanings highlight the vividness of the state and intensify the degree of quality, which can hardly be translated into English. In conclusion, the suffixes of adjectives and adverbs have changed dramatically in history, reflecting the developments of sentential structures in which the lexical items became grammaticalized. This phenomenon is a window into how the grammatical system evolved consistently.

https://doi.org/10.1017/9781108921831.011 Published online by Cambridge University Press

12 The Ditransitive Construction

12.1 Introduction The ditransitive construction is an essential grammatical device that has always existed throughout the history of Chinese. However, its semantic and syntactic properties have varied quite dramatically over time due to the influence of the general grammatical system in different periods. Generally, the ditransitive construction of the Chinese language has undergone the following three changes. First, in Old Chinese, there was an interaction or overlap between the ditransitive and instrumental constructions. If the direct object referred to a concrete thing, ditransitive expressions took the instrumental construction, where the preposition yǐ was used to introduce the direct object in sentence-final position. Otherwise, the remaining ditransitive instances made use of the unmarked form “Subj + V + Oi + Od,” which has remained the same throughout history. Second, in Medieval Chinese, there was a division of labor between the disposal and ditransitive constructions: if the direct object was definite, the corresponding expression had to take the disposal construction, and the disposal marker bǎ was needed to introduce the definite direct object in preverbal position. If the direct object was indefinite, the most proper means was the unmarked ditransitive construction described above. Third, from Modern Chinese onward, by analogy with the resultative construction, the relation between the verb and the indirect object has been treated as that between the verb and the resultative in the grammar. As a consequence, the ditransitive construction resembles the resultative structure, semantically and syntactically. In addition, the Chinese ditransitive construction can express bidirectional transfers, which means that the entity indicated by the direct object may move either from the subject to the indirect object or vice versa. This phenomenon is rare across languages; the ditransitive construction in many languages allows a one-way transfer only, moving the entity from the subject to the indirect object. This semantic property is rooted in the conceptualization of Chinese verbs.

296

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Relation to the Instrumental Construction

297

12.2 Relation to the Instrumental Construction To examine the diachronic developments of the ditransitive construction, we need to distinguish the unmarked structure from the marked forms. The unmarked structure has remained unchanged throughout history, relying on the constituent order to indicate which is the direct object and which is the indirect object, as schematized below. (1)

The unmarked ditransitive construction Subj V Oi Od

However, marked ditransitive constructions were different in markings and structure. In Old Chinese, there were two types of marked ditransitive construction, both of which employed the instrumental preposition yǐ or yòng to introduce the direct object, especially when they referred to concrete entities. That is, instrumental and ditransitive expressions simultaneously shared the same syntactic structures, as illustrated in (2) and (3). (2)

The instrumental construction: 無聽之以耳而聽之以心。 (莊子 人間世) Wú tīng zhī yǐ ěr ér tīng zhī yǐ xīn. not listen it with ear but listen it with heart “Don’t listen to it with your ears but with your heart.” (Zhuang Zi, Ren Jian Shi, 300 BC)

(3)

The ditransitive construction: 赠我以琼瑰。(左傳 成公十七年) Zèng wǒ yǐ qióngguī. send me with precious-jade “He sent me a precious jade.” (Zuo Zhuan, Cheng Gong Shi Qi Nian, 550–400 BC)







Alternatively, the instrumental yǐ could also introduce a direct object in preverbal position, especially when it was definite and concrete, as illustrated in (4). Another instrumental preposition entered the language around the sixth century BC, and it could introduce the direct object only in preverbal position, as illustrated in (5): (4)




韓宣子將以叔孫與之。 (左傳 昭公二十三年) Hán Xuānzi jiāng yǐ Shū Sūn yǔ zhī. Han Xuanzi will with Shu Sun send he “Han Xuanzi will send Shu Sun to him.” (Zuo Zhuan, Zhao Gong Er Shi San Nian, 550–400 BC)

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

298 (5)




(翕侯)用其姊妻之。 (史記 匈奴列傳) (Xī Hóu) yòng qí zǐ qī zhī. Xi Hou with his sister marry he “Xi Hou married him to his sister.” (Shi Ji, Xiong Nu Lie Zhuan,100 BC)

The unmarked ditransitive construction was already attested in the oracle bone inscriptions which date back to the thirteenth century BC (for details, see Yang and He 2001: 559‒568). In contrast, this unmarked structure seems to have been to be less restricted and could introduce either concrete or abstract objects, as illustrated below:




(6)

夫人使饋之錦與馬。(左傳 襄公二十六年) Fūrén shǐ kuì zhī jǐn yǔ mǎ. lady permit award he silk and horse “The lady permitted others to award him silks and horses.” (Zuo Zhuan, Xiang Gong Er Shi Liu Nian, 550–400 BC)

(7)

後稷教民稼穡。 (孟子 滕文公) Hòu Jì jiào mín jià-sè. Hou Ji teach people farm “Hou Ji taught the people to farm.”




(Meng Zi, Teng Wen Gong, 300 BC) As analyzed in Section 7.9, the principle of the information structure of the predicate was established in Medieval Chinese under the influence of the emergence of the resultative construction. In this circumstance, all preposition phrases that expressed non-resultatives were restricted to preverbal position; hence the instrumental yǐ disappeared from postverbal position. As a result, the ditransitive construction marked by yǐ collapsed not long after the end of the Old Chinese period.

12.3 Relation to the Disposal Construction In Contemporary Chinese, there are two types of ditransitive construction, as schematized in (8): (8)

(a) Subj V + Oi + Od, if the direct object is indefinite. (b) Subj (bǎ Od) + V + Oi, if the direct object is definite.

There is a clear division of labor between the above two forms: if the direct object is indefinite, the construction in (1a) is used; if it is definite, the disposal construction must be employed. Otherwise, an ungrammatical sentence will be created, as illustrated in (9) and (10).

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Relation to the Disposal Construction (9)

299

The direct object is indefinite: (a) 他送了王教授一本書。 (現代漢語) Tā sòngle Wáng Jiàoshòu yī-běn shū. he send-PERF Wang Professor one-CL book “He gave Professor Wang a book.” (b) *他把一本書送了王教授。 *Tā bǎ yī-běn shū song-le Wáng Jiàoshòu. he BA one-CL book send-PERF Wang Jiaoshou. (Contemporary Chinese)

(10)

The direct object is definite: (a) 他把那本書送了王教授。 (現代漢語) Tā bǎ nà-běn shū sòngle he BA that-CL book send-PERF “He sent the book to Professor Wang.” (b) *他送了王教授那本書。 *Tā song-le Wáng he send-PERF Wang

Wáng Jiàoshòu. Wang Professor

Jiàoshòu nà-běn shū. Professor that-CL book (Contemporary Chinese)

The above situation historically resulted from two major changes in Medieval Chinese: (a) the formation of the principle of assigning definiteness by syntactic position and the establishment of the resultative construction (for a fuller discussion, see Section 7.6). After the verb and resultative became fused into a single constituent, the object that originally occurred between them had to be rearranged somewhere else. If the object was definite, it was typically introduced by the disposal construction. Additionally, another important principle in relation to information structure took effect, which automatically assigned the feature definite to a bare noun in preverbal position and the feature indefinite to a bare noun in postverbal position. As mentioned above, at this time, the ditransitive construction was divided into two subtypes: if the direct object was indefinite, it adopted the traditional structure “Subj V Oi Od”; if the direct object was definite, the traditional structure could not be used and the disposal construction was the best structure to express the meaning.




(11) 卻將家信託袁師。(呂溫 临洮送袁七) Què jiāng jiā xìn tuō Yuán Shī. but DISP family letter entrust Yuan Shi “But (he) entrusted Yuan Shi with his family letter.” (Lü Wen, Lin Yao Song Yuan Qi, AD 800)

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

300 (12)




和尚莫將境示人。(祖堂集 趙州和尚) Héshàng mò jiāng jìng shì rén. monk not DISP mirror show people “The monk does not show the mirror to other people.” (Zu Tang Ji, Zhao Zhou He Shang, AD 950)

Since then, the unmarked ditransitive has been limited to cases of the direct object being indefinite, as illustrated below:




(13)

(單于)遂度與天使弓箭。 (敦煌變文 韓擒虎話本) (Chán yú) suì dùyǔ tiānshǐ gōn gjiàn. Chan Yu then award angel bow arrow “The Chan Yu then awarded the angel a bow and arrows.” (Dun Huang Bian Wen, Han Qin Hu Hua Ben, AD 800–1000)

(14)

傳與你口訣。 (西遊記一回) Chuányǔ nǐ kǒujué. tell you formula “(I) tell you a formula.” (Xi You Ji, Chapter 1, AD 1550)

12.4 Assimilation by the Resultative Construction Superficially, the ditransitive construction and resultative construction appear to be different. In Chinese, however, due to the extremely powerful analogy of the resultative construction, the ditransitive construction is treated as a subtype of the resultative construction, because the semantic relation between the verb and the indirect object holds a semantic structure, say “action + end point,” which resembles that of VR phrases. In the ditransitive construction, the verb and the indirect object first form an immediate constituent that governs the direct object. Thus, in Contemporary Chinese, if the direct object is indefinite, it follows the verb and the indirect object; if it is definite, it must be introduced in preverbal position by the disposal marker bǎ. One type of the earliest disposal instances belonged to the ditransitive construction (Ohta 1987: 244, Wang 1989: 266). As Croft (2001: 32) pointed out, every construction is language-particular because it is subject to the influence of related constructions in the language.

12.4.1 Semantic Similarity The ditransitive constructions of Chinese and English, as SVO languages, have exactly the same syntactic configuration: “Subj V Oi Od.” With the “give” type of verb, the

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Assimilation by the Resultative Construction

301

ditransitive instances of both Chinese and English express the same meanings, as illustrated in (15); for the “receive” type of verb, however, the meanings of the Chinese ditransitive instances are exactly the opposite of the meanings of their English counterparts, as illustrated in (16): (15)

(a) John sent Mary a car. ~ John sent a car to Mary. (b) 約翰送了瑪麗一輛車。 (現代漢語) Yuēhàn song-le Mǎlì yī-liàng chē. John send-PERF Mary one-CLcar “John sent Mary a car.” (Contemporary Chinese)

(16)

(a) (b)

John bought Mary a car. ~ John bought a car for Mary. 約翰買了瑪麗一輛車。(現代漢語) Yuēhàn mǎi-le Mǎlì yī-liàng chē. John buy-PERF Mary one-CL car “John bought a car for Mary.” (Contemporary Chinese)

In the following section, we will show that in Chinese the “give” subtype of the ditransitive construction behaves like the transitive resultative construction and the “receive” type shares the syntax of the intransitive construction. Here, we focus only on the overall properties of the ditransitive construction by ignoring the details involved. Goldberg (1995: 141‒151) generalized the semantics of the ditransitive construction in English as follows: [X causes Y to receive Z]. As mentioned previously, however, the Chinese ditransitive construction has two antonymous meanings: (a) [X causes Y to receive Z]; (b) [X causes Y to give Z to X]. This archetypical function of the ditransitive construction resembles the semantic structure of a verb–resultative construction in the following aspects: (a) (b) (c)

The causative event is triggered by the agent. The patient undergoes movement with the indirect object as an end point. The resultative state is the destination of transferred patient.

We do not attempt to demonstrate that the English ditransitive construction also bears some resemblance to its resultative construction, but semantic similarities do exist between well-accepted resultative instances and the prepositional paraphrases of the ditransitive construction, which have always been a ditransitive structure. (17)

(a) Mary sneezed the napkin off the table. (Goldberg 1995: 3) (b) Messi kicked the ball to the goalkeeper.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

302

Both the examples above involve the movement of an object, napkin and ball, and the remaining difference is that the location of to the goalkeeper is more specific than off the table, which can be ignored in generalizing a syntactic pattern. Before we discuss the common properties of the resultative and ditransitive constructions, let us consider the major types of the verb–resultative construction in terms of their internal structures. Type 1. The verb is transitive and can independently have an object. The resultative element is intransitive, usually an adjective or an intransitive verb. (18)

他修好了電腦。(現代漢語) Tā xiū-hǎo le diànnǎo. he fix-good PERF computer “He has fixed a computer.” (Contemporary Chinese)

Type 2. The verb is transitive but bears no direct grammatical relation to the following object. The resultative element is intransitive, usually an adjective or an intransitive verb. The object can be seen only as the patient of the whole resultative phrase. (19)

他喫壞了肚子。(現代漢語) Tā chī-huài le dùzǐ. he eat-bad PERF stomach “He ate something wrong and his stomach felt bad.” (Contemporary Chinese)

Type 3. Both the verb and the resultative are intransitive, and together they have the following NP as their patient object. (20)

他哭腫了眼睛。(現代漢語) Tā kū-zhǒng le yǎnjīng. he cry-swell PERF eye “His eyes became swollen because of crying.” (Contemporary Chinese)

Type 4. All three types of resultative above are about what happens to the object. If the resultative describes what happens to the subject instead, the verb–resultative phrase cannot be followed by any object.1 1

There are only two exceptions to this regularity: one is chī-bǎo “eat-full,” which can have the noun fàn “food” as its object, and the other is hē-zùi “drink-intoxicated,” which can be followed by the noun jiǔ “wine” as its object. Obviously, the two resultatives bǎo “full” and zùi

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Assimilation by the Resultative Construction

303

In this case, the proper device is the verb-copying construction to introduce the patient: the first verb has the patient as its object and the repeated verb is followed by a resultative element (for a full discussion, see Section 10.1). (21)

他喫烤雞喫胖了。 (現代漢語) Tā chī kǎo-jī chī-pàng le. he eat roasted-chicken eat-fat PERF “He ate (too much) roast chicken and put on a lot of weight.” (Contemporary Chinese)

Bear in mind that the fourth type of resultative cannot be followed by any patient noun. In short, the semantic similarity between the ditransitive construction and the resultative construction makes it possible for them to influence each other and share many syntactic behaviors.

12.4.2 Syntactic Similarity Hudson (1992) designed eleven empirical criteria to show that the direct object of the ditransitive construction is the real object of the predicate, in contrast to monotransitive clauses. However, he did not clarify the relationship between the verb and the indirect object. In what follows, I shall demonstrate that the relationship between the verb and the indirect object in Chinese is actually that between the verb and the resultative. First, let us briefly review the interaction between the indirect object and the resultative phrase, as mentioned in Section 13.2. Within a single clause, the resultative phrase and the indirect object repel each other, which means that only one of them can be licensed by a predicate verb. A similar constraint is found in English. In English monotransitive clauses, an adjective can occur immediately after the object to indicate a resultative, but in the ditransitive construction the addition of the resultative is blocked, as illustrated in (22). Consider the fact that two distinct resultative phrases cannot co-occur, as illustrated in (23). This constraint is applied to prepositional resultatives, as illustrated below: (22)

(a) Joe kicked a suitcase open (Goldberg 1995: 82). (b) *Joe kicked Bob a suitcase open (Goldberg 1995: 82).

(23)

(a) *She kicked him bloody dead (Goldberg 1995: 82). (b) *He wiped the table dry clean (Goldberg 1995: 82).

“intoxicated” are about what happens to the subject. These exceptions are due to lexicalization because of the high frequency of co-occurrence of the verb and the resultative involved. Shi (2002a: 68‒100) provides the statistics and relevant analysis.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

304

To explain these facts, Goldberg (1995: 82‒83) proposed two closely related hypotheses: “unique-path constraint” and “unique-change-of-state constraint.” These two constraints can in fact be combined into a single constraint, reflecting the more general principle proposed by Levin and Rappaport Hovav (1990): resultatives act as delimiters or boundaries of events, and a clause can be delimited only once. When the Chinese facts are taken into account, this principle can be stated as follows: “every verb can be followed by no more than one resultative phrase, regardless of the circumstances.” As we saw above, in English the resultative elements typically occur after the object, with a few exceptions, e.g. John cut short his article, where the resultative short intervenes between the verb and the object. In contrast, the resultative in Chinese can occur only between the verb and the object. As in English, no more than one resultative is allowed in a single clause, as illustrated in (24): (24)

a

他修好了電腦。(現代漢語) Tā xiū-hǎo le diànnǎo. he repair-good PERF computer “He has fixed the computer.”

b

他修完了電腦。 (現代漢語) Tā xiū-wán le diànnǎo. he repair-finish PERF computer “He has finished repairing the computer.”

(c) *他修好完了電腦。 *Tā xiū-hǎo-wán le diànnǎo. he repair-good-finish PERF computer (Contemporary Chinese) In the ditransitive construction in Chinese, no resultative phrase is allowed between the verb and the indirect object, as is illustrated in (25a) and (26a). However, the ditransitive verb can be followed by a resultative phrase when the indirect object is absent, as illustrated in (25b) and (26b): (25)

a

b

*我借丟了他一本書。 *Wǒ jiè-diū le I lend-lose PERF

tā yīběn shū. he one-CL book

我借丟了一本書。 (現代漢語) Wǒ jiè-diū le yīběn I lend-lose PERF one-CL “I lent a book and lost it.”

shū. book (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Assimilation by the Resultative Construction (26)

a

*我寄完了他一個包裹。 *Wǒ jì-wán le I mail-finish PERF

tā he

yī-gè one-CL

b 我寄完了一個包裹。 (現代漢語) Wǒ jì-wán le yī-gè I mail-finish PERF one-CL “I have mailed a package.”

305

bāoguǒ. package

bāoguǒ. package (Contemporary Chinese)

The above contrasts show that the indirect object and the resultative phrase actually occupy the same syntactic position. Additionally, we can conclude that the indirect object has the same syntactic property as the resultative object; thus they cannot cooccur after a single predicate verb. As pointed out previously, the indirect object usually expresses the end point of the transfer event, a kind of resultative. The verb-reduplicated form also cannot occur in the ditransitive construction in Chinese. Verb reduplication in Chinese expresses a short duration or casualness of the action involved. Li and Thompson (1981: 232) call this “the delimitative aspect,” one of the morphological forms enabled by the emergence of the resultative construction in Late Modern Chinese (around the sixteenth century AD) (see Shi 2002a for details). Verb reduplication itself is bounded like a VR phrase; hence it cannot be added to another resultative for the same reason – the uniquepath principle (Goldberg 1995: 82).2 In other words, the reduplicated verb cannot be added as a resultative phrase because the repeated verb functions like a resultative phrase. Consequently, the verb in the ditransitive construction also cannot take the reduplicated form, as illustrated in (27). (27)

The incompatibility between verb reduplication and resultative: (a) 他送完了禮物。 (現代漢語) Tā sòng-wán le lǐwù. he send-finish PERF gift “He has sent out the gifts.” (b) 我送了送禮物。 (現代漢語) Wǒ sòng-le-sòng lǐwù. I send-PERF-send gift “I tried to send gifts (to someone).”

2

In some southern dialects, such as the Wu dialect, reduplicated verb forms can be followed by another resultative phrase, because the meaning, referring to the continuation of the action, is different from that in Mandarin Chinese.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

306

(c) *我送送完了禮物。 *Wǒ song-sòng-wán le lǐwù. I send-send-finish PERF gift (Contemporary Chinese) Specifically, the resultative phrase cannot co-occur with an indirect object. As (27a) shows, those ditransitive verbs can take a resultative phrase if the indirect object is absent. However, the structures are ungrammatical if the indirect object is present. It is evident that the resultative phrase and the indirect object actually occupy the same syntactic position; thus they expel each other from the position between verb and object. In other words, the indirect object is treated as a subtype of the resultative.

12.4.3 Incompatibility with the Co-ordinate Construction Given the unique-path principle proposed by Goldberg (1995: 82), we can explain another constraint on the ditransitive construction in Chinese: a nominal co-ordinate construction cannot occur in the indirect object position but it can occur in either the subject or the direct object position. First, let us consider the related phenomenon in English: (28)

(a) He says that he saw John and me last night. (b) They have invited Mary and Tom to the party.

The test of the co-ordinate criterion can serve to support the proposal by Hudson (1992) that the direct object is the real object of the ditransitive construction but the indirect object is not, because the former can be a co-ordinate NP but the latter cannot, as illustrated in (29) and (30): (29)

(a) John bought Mary a dress and a bicycle. (b) *John bought Mary and Joan a dress.

(30)

(a) Joan baked Jack a potato and a pizza. (b) *Joan baked Jack and Tom a cake.

Similarly, in the ditransitive construction in Chinese, the direct object can be a coordinate construction, but the indirect object cannot, as illustrated in (31): (31)

(a) 我送了湯姆一個本子和一支鉛筆。 (現代漢語) běnzǐ hé yī-zhī qiānbǐ. Wǒ sòng-le Tāngmǔ yī-gè I send-PERF Tom one-CL notebook and one-CL pencil “I sent Tom a notebook and a pencil.”

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Assimilation by the Resultative Construction

307

(b) *我送了湯姆和珍妮一盒巧克力。 *Wǒ song-le Tāngmǔ hé Zhēnnī yī-hé qiǎokèlì. I send-PERF Tom and Joan one-box chocolate (Contemporary Chinese) We saw that the indirect object and the resultative play the same semantic role and actually occupy the same syntactic position. Therefore the co-ordinate NP in the position of the indirect object is equal to the co-occurrence of two resultatives after a single predicate verb, which is ruled out by the unique-path principle. Once again, the above facts prove that the indirect object has semantic and syntactic properties similar to those of the resultative in Chinese.

12.4.4 The Corresponding Disposal Construction Modern Chinese provides an alternative construction – the disposal construction to express a transfer activity. As we saw in Section 9.1, the schema of the disposal construction is [Subj + BA + NPpatient + VP]. Li and Thompson (1981: 483) suggested two generalizations about how to use the disposal construction: (a)

(b)

The patient NP introduced by bǎ is definite, specific, or generic. In most cases, the patient NP is the direct object of the verb, but it may also be a patient directly affected by the disposal event designated by the verb plus the direct object in the VP position. The term “disposal” means that something happens to the patient marked by the bǎ morpheme.

The previous example in (1) illustrates that the predicate (VP) of the disposal construction is typically transitive and must be a verb–resultative phrase in many contexts. As Li and Thompson pointed out, the predicate of the disposal may be another verb–object construction, which produces an effect on the preverbal patient, as illustrated in (32): (32)

我把橘子剝了皮。 (現代漢語) Wǒ bǎ júzǐ bāo-le pí. I BA orange peel-PERF skin “I peeled the orange.” (Contemporary Chinese)

The syntactic configuration of the predicate above is “verb plus direct object,” but the meaning is similar to that of a transitive verb because the event bāo pí “peel skin” must affect a second patient. This patient noun is typically placed in preverbal position by the bǎ morpheme.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

308

As discussed above, the verb–indirect object in the ditransitive construction in fact has the same semantic properties as the transitive verb–resultative construction. Therefore the disposal construction is preferable for expressing a transfer event when the direct object is definite. Note that only the direct object can be extracted to preverbal position marked by bǎ, and the indirect object cannot. The constraint on the extraction of the two objects is illustrated in the following two examples: (33)

a 他把車送了瑪麗。 (現代漢語) Tā bǎ chē sòng-le Mǎlì. he DISP car send-PERF Mary “He gave his car to Mary (as a gift).” b *他把瑪麗送了車。 *Tā bǎ Mǎlì sòng-le chē. he DISP Mary send-PERF car (Contemporary Chinese)

(34)

a 他把禮物給了湯姆。 (現代漢語) Tā bǎ lǐwù gěi-le Tāngmǔ. he DISP gift give-PERF Tom “He gave a gift to Tom.” b *他把湯姆給了禮物。 *Tā bǎ Tāngmǔ gěi-le lǐwù. he DISP Tom give-PERF gift (Contemporary Chinese)

As in English, the indirect object of the Chinese ditransitive construction is usually a pronoun or a proper noun, which is definite and thus meets the requirement of the preverbal patient in the disposal construction. However, the indirect object cannot be extracted to preverbal position, revealing the function of the ditransitive construction in general. As mentioned previously, Goldberg (1995: 142) suggested that the meaning of the English ditransitive construction is [X(subject) causes Y(Oi) to receive Z(Od)], and this formula implies that the recipient indicated by the indirect object is the primarily affected entity in the transfer activity. The extraction constraint in the transformation from the ditransitive construction to the disposal construction comes from the fact that it is the direct object that represents the primarily affected entity. The Chinese case shows that the “disposal” action happens to affect the direct object, resulting in its movement from the subject to the indirect object. Thus the meaning of the Chinese ditransitive construction can be generalized as [Subj Causes Od to move to Oi].

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Assimilation by the Resultative Construction

309

Once again, in expressing a transfer activity, the disposal construction reveals that the verb and the indirect object form a transitive construction like a verb–resultative phrase, which allows definite direct objects to be extracted to preverbal position marked by bǎ.

12.4.5 Passivization of the Indirect Object Whether the indirect object or the direct object can be passivized differs from language to language. Some languages do not permit either of them to appear as the subject of a passive clause. However, in English, either can be used as the subject of a passive clause, as illustrated in (35): (35)

(a) The girl was given a doll. (Quirk et al. 1985: 1208) (b) A doll was given to the girl. (Quirk et al. 1985: 1208)

In Chinese, however, under no circumstances can the indirect object be passivized, but the direct object can. The constraint on the passivization of the indirect object comes from the syntactic rules of Chinese. When the indirect object is extracted to subject position, the remaining predicate is a “verb plus direct object” construction, which cannot function like a transitive phrase to cause the subject to be affected in some way. Unlike English, the indirect object cannot undergo passivization. In contrast, if the preposition gěi “to” is added to the verb, the direct object can be freely passivized, as illustrated in (36). (36)

(a) Active sentence: 他賣了瑪麗一輛自行車。 (現代漢語) Tā mài-le Mǎlì yī-liàng zìxíngchē. he sell-PERF Mary one-CL bicycle “He sold Mary a bicycle.” (b) Passivizing the indirect object: *瑪麗被他賣了一輛自行車。 *Mǎlì bèi tā mài-le yī-liàng zìxíngchē. Mary PASS he sell-PERF one-CL bicycle (c) Passivizing the direct object: 自行車被他賣給了瑪麗。 Zìxíngchē bèi tā mài-gěi-le Mǎlì. bicycle PASS he sell-to-PERF Mary “The bicycle was sold to Mary by him.” (Contemporary Chinese)

When the direct object is passivized, however, the particle gěi “to” can be omitted, as illustrated in (37):

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

310 (37)

這些珍貴的藝術品被他隨隨便便送了人。 (現代漢語八百詞) Zhèxiē zhēnguì de yìshùpǐn bèi tā suísuíbiànbiàn song-le these precious DE art PASS he casually send-PERF rén. people “These precious artworks were casually sent to other people by him.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 225)

The passivization of the indirect object is constrained for the same reason as in the disposal case. The predicate of the passive clause in Chinese is typically a verb– resultative construction that is transitive and indicates that the subject undergoes a certain state change. Therefore the direct object can be passivized because the remaining predicate is a “verb plus indirect object” construction, representing an “action plus resultative” relation. However, once the indirect object is passivized, the remaining predicate is a “verb plus direct object” construction, which does not have a transitive meaning. As a result, the indirect object cannot be passivized.

12.4.6 Constraints on Topicalization of the Indirect Object Li and Thompson (1976) pointed out that Chinese is a “topic-prominent” language, while English is a “subject-prominent” language. In English, the subject is obligatory for every finite clause, but it is often absent in Chinese. In Chinese, however, almost all the nominal elements in a sentence can be topicalized by fronting them to sentence-initial position. In terms of the ditransitive construction in Chinese, both the indirect object and the direct object can seemingly be topicalized, but these two types of topicalization have different syntactic configurations. Specifically, when the indirect object is topicalized, an anaphoric pronoun usually occurs in the original position (the trace) of the indirect object, producing the syntactic configuration, as in (38a). When the patient noun (direct object) is topicalized, however, no anaphoric pronoun is allowed to appear in the original position of the direct object, as in the schema of (38b): (38)

(a) TOPIC (Oi), Subj V Pro anaphoric Od — topicalization of the indirect object (b) TOPIC (Od), Subj V Oi

— topicalization of the direct object

The different configurations above are illustrated in (39) and (40), respectively: (39)

(a) 瑪麗, 湯姆送了她一輛自行車。 (現代漢語) Mǎlì, Tāngmǔ sòng-le tā yī-liàng zìxíngchē. Mary Tom send-PERF she one-CL bicycle “As for Mary, Tom sent her a bicycle.”

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Assimilation by the Resultative Construction

311

(b) 瑪麗, 湯姆送了一輛自行車。 (現代漢語) Mǎlì, Tāngmǔ sòng-le yī-liàng zìxíngchē. Mary Tom send-PERF one-CL bicycle “As for Mary, Tom sent a bicycle.” (Contemporary Chinese) (40)

(a) 自行車, 湯姆送了瑪麗。 (現代漢語) Zìxíngchē Tāngmǔ song-le Mǎlì. bicycle Tom send-PERF Mary “As for the bicycle, Tom sent it to Mary.” (b) *自行車, 湯姆送了瑪麗它。 *Zìxíngchē, Tāngmǔ sòng-le Mǎlì tā.3 bicycle Tom send-PERF Mary it (Contemporary Chinese)

Although both (40a) and (40b) are acceptable, the former is much more common than the latter. Considering the passivization and topicalization constraint on the indirect object, we can see that the indirect object is less likely to be extracted from its original position. In our view, this constraint is rooted in the structural hierarchy of the predicate. Under our assumption, the verb and the indirect object first form an immediate constituent, such as a verb–resultative phrase, and then have the following patient as their real object. This means that the indirect object bears a closer relation to the verb. Compared with the passive structure, the topic construction does not require that the predicate (i.e. comment) be transitive and could be anything about the “topic.” This may be why the indirect object can freely undergo topicalization but not passivization, while the direct object can undergo either topicalization or passivization. 12.4.7 The Syntax of the “Receive” Subtype of Ditransitive Construction As we saw in the introduction to this chapter, the ditransitive construction in Chinese has two opposite meanings: (a) the transfer of the patient from subject to indirect object, which can be labeled the give ditransitive construction, and (b) the transfer of the patient from indirect object to subject, called the receive ditransitive construction. The receive subtype of the Chinese ditransitive construction does not have alternative constructions, such as the disposal, passive, and topic structures, because the indirect object in this subtype refers only to a starting point (accompanying) rather than an end point 3

In many southern dialects of the Chinese language, if the indirect object is a pronoun, it can occur after the direct object. This has nothing to do with topicalization, and the motivation will be discussed in Section 13.5.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

312

(resultative). This means that the indirect object in the receive ditransitive construction cannot be conceived of as a VR phrase because of its semantic properties. The semantic difference between the two subtypes of indirect object of the give and receive ditransitive constructions can be clearly seen from their prepositional paraphrases. As mentioned previously, the information-organizing principle in Modern Chinese means that those resultative elements tend to occur after the verb, and the accompanying elements are usually restricted to preverbal position. For give ditransitive instances, the prepositional paraphrases are typically structured as follows: the preposition gěi “to” introduces the recipient after the direct object, usually in sentence-final position as a resultative phrase. In receive ditransitive instances, however, the prepositional paraphrases are typically structured in another way: the preposition cóng “from” is used to introduce the source in preverbal position as an adverbial phrase to modify the verb. The different paraphrases are illustrated in (41) and (42), respectively. (41)

The “give” ditransitive construction: (a) 約翰賣了瑪麗一輛自行車。 (現代漢語) Yuēhàn mài-le Mǎlì yī-liàng zìxíngchē. John sell-PERF Mary one-CL bicycle “John sold Mary a bicycle.” (b) 約翰賣了一輛自行車給瑪麗。 (現代漢語) Yuēhàn mài-le yī-liàng zìxíngchē gěi Mǎlì. John sell-PERF one-CL bicycle to Mary “John sold a bicycle to Mary.” (Contemporary Chinese)

(42)

The “receive” ditransitive construction: (a) 約翰買了瑪麗一輛自行車。 (現代漢語) Yuēhàn mǎi-le Mǎlì yīliàng zìxíngchē.4 John buy-PERF Mary one-CL bicycle “John bought a bicycle from Mary.” b 約翰從瑪麗那裡買了一輛自行車。 (現代漢語) Yuēhàn cóng Mǎlì nàlǐ mǎi-le yī-liàng zìxíngchē. John from Mary there buy-PERF one-CL bicycle “John bought a bicycle from Mary.” (Contemporary Chinese)

That is, in the receive ditransitive construction, the verb and the indirect object have a “verb–source” relationship, which does not match the semantic relation between verb 4

Note that the English example John bought Mary a bicycle means John bought a bicycle for Mary, which still produces a “give” sense.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Assimilation by the Resultative Construction

313

and resultative. This is why receive ditransitive instances cannot be transformed into constructions such as the disposal, passive, and topic constructions, which require that the predicate be a verb–resultative phrase, as illustrated below. (43)

The disposal construction: *約翰把自行車買了瑪麗。 *Yuēhàn bǎ zìxíngchē mǎi-le Mǎlì. John DISP bicycle buy-PERF Mary (Contemporary Chinese)

(44)

The passive construction (a) *自行車被約翰買了瑪麗。 *Zìxíngchē bèi Yuēhàn mǎi-le Mǎlì.5 Bicycle PASS John buy-PERF Mary (b) *瑪麗被約翰買了自行車。 *Mǎlì bèi Yuēhàn mǎi-le zìxíngchē.6 Mary PASS John buy-PERF bicycle (Contemporary Chinese)

(45)

The topicalization construction: (a) *瑪麗, 約翰買了一輛自行車。 *Mǎlì, Yuēhàn mǎi-le yīliàng zìxíngchē. Mary John buy-PERF one-CL bicycle (b) *自行車, 約翰買了瑪麗。 *Zìxíngchē, Yuēhàn mǎi-le Mǎlì. bicycle John buy-PERF Mary (Contemporary Chinese)

To summarize, we saw that the VOi phrase within the ditransitive construction shares major semantic and syntactic properties with the VR construction. We reached the following conclusions: (a)

(b)

5

The unique-state-change principle states that a verb can be followed by no more than one resultative. The verb in the ditransitive construction cannot be followed by any resultative, which indicates that the indirect object actually plays the resultative role. In the ditransitive construction, the indirect object cannot be a co-ordinate construction consisting of two or more nouns, but the direct object can be a co-ordinate construction. This shows that the occurrences of the indirect

Direct object passivation is not allowed.

6

Indirect object passivation is not allowed.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

314

object observe the unique-state-change principle, just as the resultative construction does. (c) In the disposal construction, the direct object (patient), marked by the morpheme bǎ, can be moved to preverbal position, but the indirect object cannot. Here, the predicate is “verb plus indirect object,” as a transitive verb–resultative phrase similar to other ordinary disposal clauses. (d) In the passive construction, the direct object (patient) can be extracted to appear in subject position, and no anaphoric pronoun is allowed to appear in its trace. When the indirect object is passivized, however, an anaphoric pronoun is frequently used in the trace of the indirect object when topicalized. Once again, the predicate is a construction formed by a VOi phrase that exhibits the property of other ordinary passives whose predicate is typically a VR phrase. (e) The “receive” ditransitive construction has no alternatives as the disposal, passive, and topicalization constructions because the indirect object refers to an accompanying feature that is semantically not in agreement with the resultative. This fact further supports the proposal that the VOi phrase in the ditransitive construction in Chinese shares the semantics and syntax of the verb–resultative construction. The above analysis is grounded exclusively on the synchronic system of Mandarin Chinese. The next section will provide evidence from numerous dialects, revealing the intrinsic relation between the ditransitive and resultative constructions.

12.5 Bidirectional Transfer This section addresses the interaction between the semantic structures of ditransitive verbs and the function of the ditransitive construction in Chinese. The set of archetypical “transfer” verbs in Chinese has semantic schema that are more abstract by one level than those in English. For example, the verb jiè in Chinese is equivalent to the combined meanings of borrow and lend in English. Chinese does not have individual verbs specifically expressing each of the two corresponding English verbs. When used in the ditransitive construction, therefore, ditransitive instances with jiè are ambiguous, giving rise to two opposite interpretations: (46)

約翰借了麥克一千塊錢。 (現代漢語) Yuēhàn jiè-le Màikè yīqiān kuài qián. John borrow/lend-PERF Michael one-thousand CL money Interpretation 1: “John lent Michael one thousand dollars.” Interpretation 2: “John borrowed one thousand dollars from Michael.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Bidirectional Transfer

315

To avoid possible ambiguity, Chinese usually utilizes prepositions (e.g. gěi “to” or cóng “from”) to differentiate the two meanings in certain contexts, but the clause structures may be altered, accordingly, which means that they are no longer the unmarked ditransitive construction.7 The conceptual schema of jiè in Chinese is a manifestation of the language that is widely attested in the set of “transfer” verbs in Old Chinese, Contemporary Chinese, and Chinese dialects (Shi 2004, 2016). We will argue that the semantic schema of this conceptualization in Chinese is responsible for conventionalizing the function of the ditransitive constructions discussed in Section 12.5.2.8

12.5.1 Semantic Features of the Ditransitive Construction The syntax and semantics of the ditransitive construction in Chinese provide empirical evidence for uncovering the nature of grammatical constructions. Unfortunately, such an exploration remains either neglected or misrepresented in the English literature on Chinese. For instance, the two most influential books on Chinese grammar in the English world, Chao (1979: 160‒164) and Li and Thompson (1981: 141), provide only simple descriptions of the ditransitive construction without touching on the interaction between transfer verbs and the construction. Huang et al. (2009) used merely the single example dì-gěi “hand-to,” a phrase consisting of the verb plus the preposition, to exemplify the ditransitive construction, although it is not the archetypical unmarked ditransitive construction. Zhang (1998) was one of the few major papers published in English that dealt particularly with the Chinese ditransitive construction, but her analysis was far from accurate in capturing the characteristics of the ditransitive construction. Her central argument is that the default meaning of the ditransitive construction is transference to the referent of the indirect object in English, transference from the referent of the indirect object in Chinese. There is, in fact, no such contrast between the two languages, and, as mentioned previously, some 7

For example, in (a), the preposition phrase gěi Màikè “to Michael” is used as a resultative and the verb jiè refers to “lend”; in (b) the preposition phrase cóng Màikè “from Michael” functions as an adverbial phrase to modify the predicate and the verb jiè means “borrow.” a. 約翰借給麥克一千塊錢。 Yuēhàn jiè gěi Màikè John lend to Michael “John lent Michael 1,000 dollars.”

yīqiān kuài qián. one-thousand CL money

b. 約翰從麥克那裡借了一千塊錢。 Yuēhàn cóng Màikè nàlǐ jiè-le yīqiān kuài qián. John from Michael there JIE-PERF one-thousand CL money “John borrowed 1,000 dollars from Michael.” 8

The term “conceptualization” is used in different senses in the literature. For instance, Langacker (1999: 45‒72) used it to refer to how to construe or understand words or sentences in the contexts of daily communication. In this chapter, however, we reserve this term to refer to how the speakers of a language coin words when categorizing things or actions in reality.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

316

ditransitive utterances (e.g. the “give” transfer) in Chinese are the same as their counterparts in English. Although she aimed to address the interaction between the meanings of verbs and ditransitive constructions, Zhang did not realize that the conceptual features of verbs are internally linked to their constructional meanings. Concerning verbs whose meanings are “transfer to” (i.e. the direction of the movement of the patient is from the subject to the indirect object), the ditransitive constructions in English and Chinese have exactly the same function. In these cases, English and Chinese translations can be performed word by word without changing the meaning, as illustrated in (47) and (48): (47)

比爾送了瑪麗一本書。 (現代漢語) Bǐěr sòng-le Mǎlì yī-běn shū. Bill send-PERF Mary one-CL book “Bill sent Mary a book.” (Contemporary Chinese)

(48)

比爾賣了瑪麗一輛自行車。 (現代漢語) Bǐěr mài-le Mǎlì yī-liàng zìxíngchē. Bill sell-PERF Mary one-CL bicycle “Bill sold Mary a bicycle.” (Contemporary Chinese)

However, when verbs that mean “transfer from” (the direction of the movement of the patient is from the indirect object to the subject) are used, the ditransitive constructions in English and Chinese have exactly opposite functions. In English, this still means that the patient is transferred from the subject to the indirect object; for instance, John bought Mary a bicycle means that John bought a bicycle and gave it to Mary as a gift. In contrast, in Chinese, the patient is transferred from the indirect object to the subject; in other words, the actor (subject) receives the patient instead, as illustrated in (48) and (49): (49)

約翰買了瑪麗一輛自行車。 (現代漢語) Yuēhàn mǎi-le Mǎlì yī-liàng zìxíngchē. John buy-PERF Mary one-CL bicycle “John bought a bicycle from Mary.” (Contemporary Chinese)

(50)

湯姆贏了傑克一盒巧克力。 (現代漢語) Tāngmǔ yíng-le Jiékè yī-hé qiǎokèlì. Tom win-PERF Jack one-CL chocolate “Tom won one box of chocolate from Jack.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Bidirectional Transfer

317

The commonalities and differences between English and Chinese are important for applied linguistics and theoretical linguistics. When teaching English as a second language to native speakers of Chinese, lecturers might caution students that the ditransitive constructions of the two languages differ in function. Otherwise, Chinesespeaking students are very likely to misunderstand the sentence John bought Mary a bicycle as “John bought a bicycle from Mary”; the same mistakes may be made by native speakers of English who are learning Chinese. We can easily recognize how the function of a construction can alter the meanings of concrete verbs in English, compared to its counterparts in Chinese. In English, the transfer direction of the ditransitive construction is specific: the route must be from the subject to the indirect object. Consequently, verbs that have the meaning “receive” will change to a “give” sense when used in the ditransitive construction, e.g. John bought Mary a bicycle or Bill got Joan a movie ticket. In active and monotransitive clauses, both get and buy mean that the subjects receive something from someone else. For instance, John bought a bicycle suggests that John has a bicycle now; similarly, Bill got a movie ticket means that Bill has a movie ticket. “Receive” verbs such as buy or get in English must change their transfer directions when used in a ditransitive construction. These are ideal examples for illustrating how the function of a construction can affect the meanings of concrete verbs. On the basis of the above discussion, the fundamental differences between the English and Chinese ditransitive constructions are shown in the following figure: A.

The function of the ditransitive construction in English

Single function: the patient is transferred to the indirect object from the subject. B.

The function of the ditransitive construction in Chinese

Double functions: (a) the patient is transferred to the indirect object from the subject; (b) the patient is transferred to the subject from the indirect object. According to our extensive investigation of typological research in the literature, Chinese seems to be the only language in which the ditransitive construction is capable of expressing bidirectional transfers. That is, the ditransitive constructions of all other languages have meanings similar to those in English: the subject is a giver, and the indirect object is a recipient. In Chinese, however, the subject can be either a giver or a recipient, and the indirect object can also be either a giver or a recipient.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

318

As we saw above, the meanings of the Chinese verb jiè subsume those of the two English verbs – lend and borrow. Thus ditransitive instances with jiè have two opposite interpretations, expressing the patient transfer either from the subject to the indirect object or vice versa. Likewise, the verb shàng in an educational context can mean either teach or learn; consequently, example (51) has two opposite interpretations. (51)

我上了他一門中文課。 (現代漢語) Wǒ shàng-le tā yī-mén zhōngwén kè. I teach/learn-PERF he one-CL Chinese course Interpretation 1: “I taught a course in Chinese to him.” Interpretation 2: “I took a course in Chinese from him.” (Contemporary Chinese)

Corresponding to the pair of verbs send and receive in English, there are also two distinct verbs in Chinese: sòng “send” and shōu “receive.” The syntax of shòu in Chinese is, however, different from that of its counterpart “receive” in English. Both Chinese verbs in this antonymous pair can occur in ditransitive constructions, whereas only the “give” member of the English antonymous pair can be used in ditransitive constructions, as illustrated in (52): (52)

(a) 麗薩送了約翰一百塊錢。(現代漢語) Lìsà sòng-le Yuēhàn yībǎi kuài qián. Lisa send-PERF John one-hundred CL money “Lisa sent John 100 dollars.” (b) 麗薩收了約翰一百塊錢。 (現代漢語) Lìsà shōu-le Yuēhàn yībǎi kuài qián . Lisa receive-PERF John one-hundred CL money “Lisa received 100 dollars from John.” (Contemporary Chinese)

12.5.2 Conceptualization of “Transfer” Verbs By definition, conceptualization refers to how people coin words to encode their observation of the world, and it relates to how a language’s lexical system is established. Different languages might employ different methodologies to conceptualize the world, and, as a result, lexical systems differ dramatically from one language to another. In this chapter, different means of conceptualization are critical to understanding the differences between the verbal systems of English and Chinese. Before discussing the differences, I would like to introduce the technical term “vector” from mathematics

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Bidirectional Transfer

319

and physics: a quantity having direction as well as magnitude, especially in determining the position of one point in space relative to another. The salient features of actions, movements, or changes, which are typically encoded by verbs in a language, are temporally and dynamically directional. Here, we are concerned only with the dynamic direction. English and Chinese employ different ways to encode this “vector” feature of verbs; consequently, the number and semantic structures of verbs are systematically different between the two languages. Generally, the Chinese language does not specify the vector direction of its verbs, and typically, a single verb is used to denote the same type of action with two opposite directions. However, English verbs specify the vector of action; thus the language usually has two distinct verbal forms to express opposite directions of the same kind of action. As we saw above, the Chinese verb jiè is equal to the combined meanings of the verbs lend and borrow in English, and the Chinese verb shàng is equal to the combined meanings of the two English verbs teach and learn. That is, Chinese schematizes the concepts of actions at a more abstract level than English does. At a superordinate level, either lend– borrow or learn–teach can be viewed as one type of action. When necessary, Chinese makes use of prepositions or directional verbs to specify the directions of the verbs; e.g. jiè-gěi “borrow-to” means “lend” and jiè-zì “borrow-from” means “borrow.” The methodology of conceptualization in Chinese is quite regular and systematic, synchronically, diachronically, and geographically. First, let us consider more synchronic examples to illustrate this point, including examples from standard Mandarin (Putonghua) and Chinese dialects. In addition to verbs such as jiè, exemplified above, more verbs of this type are provided as follows. (a)

(b)

(c)

(d)

The Chinese verb ná corresponds to three verbs in English: take, bring, and hold. To specify the direction, directive verbs have to be added after the verb: ná-lái (-“come”) means “bring,” ná-zǒu (-“go”) means “take,” and ná-zhù (-“stay”) means “hold.” The Chinese verb dài can mean either borrow money or lend money in English. Similarly, shé can mean either buy on credit or sell on credit. A proper preposition must be used in order to specify the direction of the verb. The Chinese verb fēn can mean either distribute or get one’s share in English. The appropriate morphemes must be added to specify the direction when used in written or spoken language. The Chinese verb mǎi corresponds to buy in English, and the Chinese verb mài corresponds to sell in English. However, buy and sell bear no phonological resemblance in English, whereas mǎi and mài have the same syllable with different tone values to differentiate the direction of action. These two Chinese verbs share the same etymological source in the history

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

320

of Chinese. In some Chinese dialects, these two verbs have exactly the same phonological form; e.g. in the Suzhou dialect both are pronounced mɒ̀ .9 (e) In many Chinese dialects, such as Wuhan, Suzhou, and Nanchang, the verb bǎ with phonological variants can mean either “give” or “receive” in English. (f) The Chinese verb zū can express the meanings of rent and lease in English, the same social activity. Rent means “pay someone for the use of something,” implying a “receive” transfer, while lease means “grant property to someone,” implying a “give” transfer. (g) The Chinese verb tì means either substitute or replace in English. These two English verbs express the same type of action but with different directions. “Substitute” means “a person or thing acting or serving in place of another,” and “replace” means “to take the place of something, or to put something or someone in the place of something or someone else.” All the “transfer” verbs listed above belong to the set of verbs that frequently occur in the ditransitive construction, giving rise to two opposite meanings if not specified by other directive verbs or prepositions. Nevertheless, not all “transfer” verbs have nonspecific directions, as in the examples above. For example, the two verbs gěi “give” and sòng “send” seem to be directionally specific, and both verbs always refer to a movement of the patient from the subject to the indirect object when used in ditransitive or monotransitive clauses. Note that the English ditransitive allows the indirect object to be absent, e.g. John gave a book, but the direct object must be present to make it well formed, e.g. *John gave her. In contrast, in Chinese, neither the direct object nor the indirect object is necessary for a well-formed ditransitive instance. When they occur without indirect or direct objects, sentences with these two verbs are also ambiguous, as illustrated in (53): (53)

傑克已經給了。 (現代漢語) Jiékè yǐjīng gěi-le. Jack already give-PERF Interpretation 1: “Jack has been given (something).” Interpretation 2: “Jack has given (somebody something).” (Contemporary Chinese)

The conceptualization methodology of the transfer verbs above can be traced back to Old Chinese,10 and many “transfer” verbs have semantic structures similar to that of jiè, 9 10

Department of Chinese Language and Literature at Peking University, ed., The Dialect Lexicon of Chinese, Beijing: The Philological Press, 434. The terms “ditransitive verb” and “transfer verb” are used in related, yet distinctly different, senses. “Ditransitive verb” is a syntactic definition referring to a verb that can be followed by two objects, indirect and direct. “Transfer verb” is a semantic definition, meaning that such a verb expresses object transfer.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Bidirectional Transfer

321

which has been capable of expressing the antonymous pair borrow and lend throughout history. Here, we do not intend to offer an exhaustive list of verbs of this kind throughout history; instead, we provide several typical examples to illustrate this point. (a)

In Old Chinese, jiǎ also has two opposite meanings, borrow and lend, exactly like its counterpart jiè in Contemporary Chinese.

(54)

久假而不歸。 (孟子 盡心章句上) Jiǔ jiǎ ér bù guī. Long borrow and not return “(Someone) borrowed it a long time ago but didn’t return it.” (Meng Zi, Jin Xin Zhang Ju Shang, 300 BC)

(55)

唯器與名不可以假人。 (左傳 成公二年) Wéi qì yǔ míng bù kěyǐ jiǎ rén. only instrument and title not can lend people “Only ritual objects and titles cannot be lent to somebody else.” (Zuo Zhuan, Cheng Gong Er Nian, 550–400 BC)

(b)

In Medieval Chinese, shòu had both meanings of the antonymous English verbs “give” and “receive.” Then, this verb had only the single phonological form and the same written form, but later a “hand” radical was added to the written form when expressing “give.” The phonological forms remain the same today. Additionally, two compound verbs have been coined to differentiate the transfer direction: shòu-yú means “give,” and jiè-shòu means “receive,” both retaining the morpheme shòu. The following two examples illustrate the two antonymous meanings of shòu in Classical Chinese:

(56)

權辭讓不受。 (三國志 吳主傳) Quán círàng bù shòu. Name decline not receive “Quan declined and didn’t receive it.” (San Guo Zhi, Wu Zhu Zhuan, AD 450)

(57)

因能而受官。 (韓非子 外儲說) Yīn néng ér shòu guān. base ability and give official “The positions of government officials are given to individuals according to their abilities.” (Han Fei Zi, Han Fei Zi, 300 BC)

(c)

In Old and Medieval Chinese, the verb gū can mean either buy or sell, as illustrated below.













https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

The Ditransitive Construction

322 (58)




當為子沽酒。 (墨子 公孟) Dāng wéi zǐ gū jiǔ. will for you buy wine “I will buy wine for you.” (Mo Zi, Gong Meng, 350 BC)

(59)




當爐自沽酒。 (陸龜蒙 酒垆) Dāng lú zì gū jiǔ. serve stove himself sell wine “He sat next to the stove and sold wine by himself.” (Lu Gui Meng, Jiu Lu, AD 850)

Like the construction with jiè “borrow/lend” in Modern Chinese, the ditransitive instances in Old Chinese are also ambiguous. Let us cite the most frequent verb shòu “give/receive” to illustrate this point. (60)




賓降堂, 受老束錦。 (儀禮 聘禮) Bīn jiàng táng, shòu lǎo shù jǐn. guest step-down hall give/receive elder roll brocade Interpretation 1: “The guest stepped out of the hall and gave an elder a roll of brocade.” Interpretation 2: “The guest stepped out of the hall and received a roll of brocade from an elder.” (Yi Li, Pin Li, 300 BC)

In most cases, the real meanings of ditransitive clauses such as (60) can be determined only by context because they always have two possible readings. Obviously, a perfect parallelism exists between the meanings of the ditransitive verbs and the functions of the Chinese ditransitive constructions. This parallelism can hardly be accidental. Before we explore the motivations for these phenomena, let us examine the historical usage of the ditransitive construction. Yang and He (2001: 559‒568) provided robust evidence that Old Chinese (from the thirteenth century BC to the first) had the same ditransitive construction as Contemporary Chinese, as illustrated in the following examples. Example (61) expresses a “give” transfer, (62) expresses a “receive” transfer, and in (63) the first clause is a “receive” transfer and the second is a “give” transfer: (61)




靜女其孌, 貽我彤管。 (詩經 靜女) Jìng nǚ qí luán, yí wǒ tóng guǎn. quiet girl so beautiful send I red stem “The quiet girl is so beautiful. (She) sent me a red flute.” (Shi Jing, Jing Nü, 1000–600 BC)

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Bidirectional Transfer

323




(62)

公攻而奪之幣。 (左傳 哀公二十六年) Gōng gōng ér duó zhī bì. Gong attack and seize he money “Gong attacked him and seized money from him.” (Zuo Zhuan, Ai Gong Er Shi Liu Nian, 550–400 BC)

(63)

得他一束絹, 還他一束羅。 (王梵志詩) Dé tā yī-shù juàn, huán tā yī-shù luó. Get he one-CL silk return he one-CL satin “(I) got one roll of silk from him and gave him one roll of satin.” (Wang Fan Zhi, AD 650)

Thus far, we have observed that the meanings of the “transfer” verbs and the ditransitive construction have existed in the Chinese language for more than 3,000 years and have also existed in English for a long time. We have excluded one possible explanation that attributes parallelism to some pragmatic factors in Contemporary Chinese or English. To conclude, it is consistent in Chinese for a single verb to be used to encode the same transfer action with opposite directions, leaving the vector directions nonspecific. In contrast, English typically employs two distinct verbs to express the two opposite directions of the same social activities. This difference in conceptualization has profound consequences for the grammars of English and Chinese, which we will discuss in detail in the following section.

12.5.3 Bidirectional Functions Having explored the different functions of the ditransitive constructions and the different conceptualizations of related verbs in English and Chinese, we now turn to explaining the interaction between the verb and the construction. Apparently, there are parallel differences between the meanings of the “transfer” verbs and the functions of the ditransitive constructions in the two languages, as shown in Table 12.1. As we discussed in the preceding sections, in English, borrow and lend serve to conceptualize the same “transfer” activity in opposite directions. The conceptualizing process of borrow is from the receiver (starting point) to the giver (end point), and that of lend is the opposite. Thus the meanings of these two verbs contain an inherently specified vector direction. In contrast, the Chinese language coins a single verb, jiè, to refer to the overall transfer activity, leaving the vector directions nonspecific. Correspondingly, the same distinctions have been found in the functions of their ditransitive constructions. The transfer direction of the English ditransitive construction is specific: the transfer of the patient can only be from subject to indirect object. In

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

324

The Ditransitive Construction

Table 12.1 “Transfer” verbs and ditransitive constructions in English and Chinese Languages

Vector direction of “transfer” verbs

Vector direction of ditransitive construction

English Chinese

Specific (one direction) Nonspecific (two directions)

Specific (one direction) Nonspecific (two directions)

contrast, the vector directions of the ditransitive construction in Chinese are nonspecific, and the transfer of the patient can be in either direction: from the subject to the indirect object or from the indirect object to the subject. Two theoretical frameworks can provide plausible explanations for the phenomena in question: cognitive linguistics (particularly construction grammar) and a usage-based model. Our hypotheses can be sketched as follows. (a) Explanation on the basis of the hypothesis of a continuum between the lexicon and grammar. One of the fundamental hypotheses of cognitive grammar and construction grammar is that a continuum exists between relevant lexical and grammatical constructions. Langacker (1987: 11‒45, 1991: 19‒22, 2008: 80) argued that linguistic units are symbolic by nature, consisting of two poles: phonology and semantics. Within this framework, morphemes, words, and complex constructions are all regarded as kinds of symbol with different degrees of complexity. Goldberg (1995: 1‒5) defined “construction” as any form–meaning (function) pairings whose meanings are somehow idiosyncratic or co-occur sufficiently frequently, from morphemes to complex constructions. Following these theories, “transfer” verbs and ditransitive constructions are by nature the same type of linguistic unit; thus their meanings come from the same human cognitive abilities, conceptualization, and schematization. However, the means of conceptualization are not random but regular and systematic. Individual transfer verbs are used to encode concrete actions of those transfer activities, and the ditransitive construction is employed to encode all events of the transfer activities, typically involving an agent, a patient, and a recipient. When encoding concrete actions and whole events, Chinese tends not to specify the direction of the transfer behavior, whereas English always specifies the direction for a concrete action or a whole transfer event. As a result, parallelism in the semantics and syntax of the transfer verbs and the ditransitive construction emerges in English and Chinese. (b) Explanation on the basis of the hypothesis of the usage-based model. The usagebased model is closely related to cognitive linguistics, which was first proposed by Langacker (1988, 1991, 2008) and fully developed by Bybee (2007). From the perspective of this model, grammatical constructions emerge from frequently used patterns of

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

Bidirectional Transfer

325

lexical combinations, as mentioned at the beginning of this chapter. Bybee developed a finer model called the “exemplar model” to explain where the meanings of grammatical constructions come from. Specifically, she claimed that the meaning of the English ditransitive construction is based on the verb give, which most frequently occurs in the construction. Since the meaning of give is specific in terms of the movement direction, the meaning of the English ditransitive construction is specific in the transfer direction, which is only from the subject to the indirect object. In line with this usage-based model, I believe that it is easy to explain why the transfer directions of the Chinese ditransitive construction are nonspecific. Throughout history, the set of transfer verbs frequently used in ditransitive constructions have always been bidirectional; e.g. shòu in Old Chinese means “give” and “receive,” and jiè in Modern Chinese means “borrow” and “lend.” Because either meaning of these verbs is equally possible, their lexical meanings are given to the ditransitive construction. As a result, the Chinese ditransitive construction also expresses bidirectional meanings. These two explanations seem equally plausible and compatible with each other. However, I prefer the second explanation, grounded in the “usage-based model,” to the first explanation, based on the hypothesis of a continuum between the lexicon and grammar. In my view, it is basically arbitrary that meaningless sounds combine to create the phonological forms of words, but words combine in regular ways to create complex constructions. Words play an active role in constructing complex linguistic units. In Chinese, the bidirectional conceptualizations of those transfer verbs determine the bidirectional function of the ditransitive construction, but not vice versa.

https://doi.org/10.1017/9781108921831.012 Published online by Cambridge University Press

13 Aspect and Tense

13.1 Introduction The Chinese language had no aspect suffixes attached to the verb stem before the tenth century AD. In his seminal work The History of Chinese Grammar, Wang (1989: 99) stated that the emergences of the aspect suffixes le (perfective) and zhe (progressive) were epoch-marking events in the history of Chinese grammar and that their development processes should be seriously studied. To explain the reasons why they emerged at the time, two closely related questions need to be answered: what factors in the earlier grammatical system blocked aspect suffixes from grammaticalizing? What changes in the grammar at the time made it possible for them to come into existence? When studying the motivation and mechanism for their development, we must separate two issues: (a) the overall change of the grammatical system at the time and (b) the concrete grammaticalization process of each aspect suffix. The emergence of the aspectual system actually created a new grammatical category, namely the verb suffix, which was expressed by auxiliary verbs before this time (we will return to this point in the subsequent section). In the evolution of Chinese grammar, grammatical changes can be classified into the following five types: (a) The grammatical category simply died out, which meant that there was no replacement in the grammar at a later stage, such as the co-ordinate conjunctions for verbs ér and qiè in Old Chinese, which fell out of use in the first half of the Medieval Chinese period and have had no equivalents in the grammar since then (for details, see Chapter 4). (b) The grammatical morpheme was replaced by a later form; for example, the disposal jiāng occurred approximately 200 years earlier than the disposal bǎ, and after more than a millennium of competition, the disposal bǎ entirely replaced its predecessor jiāng (for details, see Section 9.9). (c) Some grammatical apparatus was entirely innovative in terms of the grammatical category but not in terms of the word class of its markings.

326

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Introduction

(d)

(e)

327

For instance, the disposal morphemes belong to this type because they mark the disposal construction, which did not exist before the sixth century AD, but the markings were prepositions that existed in the language from the beginning (for details see Chapter 9). The grammatical elements were completely new in terms of grammatical category and the word class of marking, such as classifiers, the plural marker -men, and the diminutive marker -er (see Chapter 19 for details). The grammatical category has always remained in the language but the word class representing this category has changed. This type of change includes the emergence of the aspect suffixes. As will be discussed below, the same grammatical category was represented by auxiliary verbs in Old Chinese, where they preceded the verb, but in Modern Mandarin the category is represented as suffix as a kind of clitic which immediately follows the verb.

Some clarification is needed here. Aspect is easily confused with tense because they are often intertwined in use. Wang (1989: 91) correctly pointed out that the suffixes in question stand for aspects rather than tenses. Aspect refers primarily to the stage at which the action denoted by the verb takes place, focusing on the internal development of an action: whether it has reached completion (with or without current relevance) or is in progress. Despite the same label being used for a type of aspect, such as “perfective,” the meanings and functions of a particular aspect vary greatly from language to language, depending mainly on the division of labor of the aspect categories that the language grammaticalizes. For example, English has two aspects: progressive (e.g. I am eating) and perfect (e.g. I have eaten), but Chinese has a three-way contrast in aspect: perfective, progressive, and experiential. Consequently, the perfective aspect in English must have characteristics that differ from those of its counterparts in Chinese. For convenience of discussion, let us first consider the definitions of the three primary aspect suffixes (for a full discussion, see Li and Thompson 1981: 184‒237, Shi 2010: 229‒247). (a)

Perfective -le, referring primarily to an action in the past but with current relevance or denoting a present state resulting from a past situation. When occurring after a VO phrase or in sentence-final position, it indicates a change of state, which may be labeled le2 in the literature. This perfective aspect focuses on the current situation after the action has happened. This term is labeled “perfective” in Li and Thompson (1981: 185) and describes an action viewed as a simple whole. Their definition accurately characterizes the function of the suffix -le, as the following examples show its contrast with the suffix -guo. Thus, in our reanalysis, we use the label

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

328

“perfective aspect” to refer to -le, including both le1 and le2. Historically, the two perfective markers evolved from the same lexical source. (b) Experiential -guo, referring primarily to an action in the past, without the implication of current relevance, focusing on the end point of the action. The contrast between -le and -guo is evident in the following examples. (1)

(a) 他們兩個去年結了婚。(現代漢語) Tāmen liǎng-gè qùnián jié-le hūn. they two-CL last-year marry-PERF marriage “The two of them got married last year.” (They are still married now.) (b) 他們兩個去年結過婚。(現代漢語) Tāmen liǎng-gè qùnián jié-guò hūn. they two-CL last-year marry-EXPE marriage “The two of them held a wedding last year.” (They are already divorced.) (Contemporary Chinese)

(2)

(a) 他去了北京。 (現代漢語) Tā qù-le Běijīng. he go-PERF Beijing “He has gone to Beijing.” (He is still in Beijing now) (b) 他去過北京。 (現代漢語) Tā qù-guò Běijīng. he go-EXPE Beijing “He went to Beijing before.” (Contemporary Chinese)

If there is no time word in a sentence, the default time reference is the point at which the speech action takes place. Thus the aspect markers -le and -guo often indicate an activity in the past, which makes them appear to be a past tense, but they can also refer to an action in the future if the time reference is set in the future, as in the following examples. (3)

明年這個時候我就拿到了畢業證。 (現代漢語) Míngnián zhège shíhòu wǒ jiù nádào-le bìyè zhèng. Next-year this-CL time I then obtain-PERF graduate diploma “I will have obtained my graduation certificate at this time next year.”

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Introduction (4)

329

明天我喫過了早飯就來。 (現代漢語) Míngtiān wǒ chī-guò-le zǎofàn jiù lái. tomorrow I eat-EXPE-PERF breakfast then come “I will come after I have had breakfast tomorrow.” (Contemporary Chinese)

Note that -le and -guo often co-occur within a verb phrase, as illustrated above, and the order is always V-guo-le, where -guo highlights the action in the past and -le refers to the current relevance. (c)

(5)

Progressive -zhe, referring primarily to an ongoing action at a temporal point or span. This aspect roughly corresponds to the progressive in English, as illustrated in (5): 我喫着午飯呢。 (現代漢語) Wǒ chī-zhe wǔfàn ne. I eat-PROG lunch PRT “I am eating my lunch now.”

Owing to the same motivation and mechanism as the emergence of the above aspect suffixes, and during the same period in history, the language developed other minor aspect markers such as the inchoative -qǐlái, referring to the start of action, and the socalled delimitative aspect – the reduplication verb form, which will also be discussed in subsequent sections. In Contemporary Chinese, the aspect suffixes can be freely combined with verbs and adjectives. Note that, as discussed repeatedly in this book, in Chinese grammar the verb and the adjective are syntactically alike; hence a cover term, wèicí “predicate word,” has been coined to label both of them (Zhu 1982: 40). Specifically, adjectives in Chinese behave like verbs; for instance, they can be directly used as the predicate without a copula such as be in English to link it with the subject. Moreover, the three aspects have been cliticized and undergone phonological reduction due to the effect of grammaticalization; all have lost their tonal values, and their vowels have become a schwa [ǝ]. Their phonological forms in Medieval Chinese and in Contemporary Chinese are listed in Table 13.1.1 The emergence of these aspect suffixes creates a distinction between finite and nonfinite verbs in the grammar, and a sentence is structured around the center of the finite verb. If two or more verbs refer to events that happen at the same time, only one of them can be suffixed with an aspect marker, or another verb may be suffixed with the progressive -zhe (similar to I saw him playing basketball in English), as illustrated in (6): 1

The reconstruction is based on Li and Zhou (1999).

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

330

Table 13.1 The phonological reductions of aspect suffixes Medieval Chinese Perfective -le Progressive -zhe Experiential -guo

(6)

213

liau ȶǐo51 kuɑ51

Contemporary Chinese lǝ tʂǝ kuǝ

(a) 我看过他打篮球。 (現代漢語) Wǒ kàn-guo tā dǎ lánqiú. I see-EXPE he play basketball “I saw him play basketball.” (b) *我看過他打過籃球。 *Wǒ kàn-guo tā dǎ-guo lánqiú. I see-EXPE he play-EXPE basketball (Contemporary Chinese)

As exemplified above, the emergence of the finite clause was a major event in the evolution of Chinese grammar. As a consequence, the language has acquired a morphological method for distinguishing prepositions from verbs. All prepositions in Chinese developed out of ordinary verbs, but before the Modern Chinese period it was difficult to tell prepositions from verbs because of the lack of overt morphological forms. In Contemporary Chinese, some prepositions, such as the disposal bǎ and the passive bèi, have permanently lost their verb status, but some of them, such as the dative gěi, whose verb status “give” is still in use, can still be used as a verb in different contexts. Morphological markings are helpful in structuring a sentence by reducing potential ambiguity, and in this sense this change may be regarded as a progression of the grammar.

13.2 Conditions for the Emergence of Aspect Suffixes In fact, the emergence of aspect suffixes was a by-product of the development of the resultative construction. In other words, it can also be said that the aspect suffices virtually belong to a subtype of resultatives that are just a small group of resultatives with general meanings suitable for describing the internal structure of an action. For instance, the perfective -le grammaticalized from the verb “finish,” the same change found in many languages (Heine and Kuteva 2002: 334). The combined force of the reanalysis of numerous resultative cases created a new syntactic position, labeled R in

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Conditions for the Emergence of Aspect Suffixes

331

the square brackets in the following schema, between the verb and the object, where R could be intransitive and might bear no direct semantic relation to the object (for an indepth discussion, see Chapter 6). (7) V [R] O The creation of the R position between the verb and the object was necessary for the emergence of many major grammatical devices in Modern Chinese, including the resultative construction, the three aspect suffixes, and the reduplication of verbs. The structure depicted in (7) was achieved through two critical changes in the grammar that are outlined as follows. The first change was the disappearance of the verb conjunction ér, an important grammatical morpheme in Old Chinese (for details, see Chapter 4). At the time, instances of the serial verbal construction were ruled out because they were ill-formed in the grammar (Wang 1989: 255‒261), and the grammar then required the conjunction -ér to link any two verbs, adjectives, or even adverbs within a co-ordinate construction. In Old Chinese, the verb and the resultative were also separated by this conjunction, as illustrated below:




(8)

恭己正南面而已矣。 (論語 衛靈公) Gōng jǐ zhèng nánmiàn ér yǐ yǐ. respectful self proper south-face and complete PRT “(Shun) respectfully and properly sat in the ruler’s seat, that is all.” (Lun Yu, Wei Ling Gong, 500 BC)

(9)

如之何其使斯民饑而死也? (孟子 梁惠王) Rú-zhī-hé qí shǐ sī mín jī ér sǐ yě? how he make this people starve and die PRT “How did he make the people starve and die?” (Meng Zi, Liang Hui Wang, 300 BC)




In the above examples, the underlined words refer to the resultatives of the verbs, which were separated by the conjunction -ér. Because the verb and the resultative could not occur adjacently, no reanalysis could happen to make them a compound. Due to the disappearance of the conjunction between the first century BC and the fourth century AD, the serial verb construction became the paradigm structure, which enabled the reanalysis of the verb and the resultative to take place. Then, the original boundary between the verb and the resultative weakened and they eventually fused into a single constituent consisting of “verb + suffix.” The second change was that the co-ordinate verb construction ceased to work. Roughly before the fifth century AD, the co-ordinate verb construction required that all verbs be transitive and each of them was combined with the object to form

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

332

a “verb–object” relation (for a fuller discussion, see Chapter 4), which can be formalized as follows: (10)

(V1 + V2) + O = (V1 + O) + (V2 + O) An example is provided below:

(11)

樓煩輒射殺之。(史記 項羽本紀) Lóu Fán zhé shè-shā zhī. Lou Fan then shoot-kill he “Lou Fan shot and killed him.”




(Shi Ji, Xiang Yu Ben Ji, 100 BC) Under the operation of the rule of verb co-ordination, if the second verb element was intransitive, the constituent order could only be “Vtr O Vintr,” where the resultative element was separated from the first verb by the object. Ohta (1987: 360‒367) was the first to find the contrastive distribution of the two synonymous verbs – shā “kill” and sǐ “die” – when used as a resultative: the former could occur only between the verb and the object because it was transitive, and the latter could only follow the whole VO phrase because it was intransitive. This could be viewed as a rule derived from the general principle regulating the co-ordinate verb construction. This was a rigorous rule by at least the fifth century AD (for details, see Section 6.4), as illustrated in (12) and (13):




(12)

殺其騎且盡。 (史記 李將軍列傳) Shā qí qí qiě jǐn. kill their cavalry almost entire “(He) killed their cavalry almost entirely.” (Shi Ji, Li Jiang Jun Lie Zhuan, 100 BC)

(13)

其日照窗倍明。 (祖堂集 吉靈和尚) Qí rì zhào chuāng bèi míng. the sun irradiate window double bright “The sun’s radiation made the window double bright.” (Zu Tang Ji, Ji Ling He Shang, AD 950)




The structures determined by the two grammatical rules discussed above channeled the path and process of the developments of the resultative construction and the aspect suffixes. Clearly, when the rule of the above verb co-ordination was still in operation, the VRO structure was ruled out by the grammar because resultatives were usually intransitive and did not syntactically govern the object argument. Under this condition, it was impossible for aspect suffixes to emerge.

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Conditions for the Emergence of Aspect Suffixes

333

Thus far, many studies on the emergence of the aspect suffixes have been conducted in the circle of Chinese historical linguistics. The above phenomenon of the grammar at that time is helpful in understanding the shortcomings of previous studies on the subject of the development of the aspect suffixes. Regarding the perfect marker -le, superficially, its emergence involved a change in the following constituent order: (14)

V O liǎo > V-le O, where liǎo is the former verb form of the perfective aspect -le.

Only when -le could occur between the verb and the object can we say that it had been grammaticalized into an aspect suffix. How could the verb liǎo jump across the object in history? Many researchers have proposed that this change happened on analogy with other resultatives such as de “modal suffix,” què “complete,” and zháo “attach,” because this group of intransitive words originally followed the VO phrase, and they moved to the position between the verb and the object at a later stage. However, this explanation fails to answer the ultimate question: what factors triggered the position change of these intransitive verbs? Furthermore, as Jiang (2005b: 136‒154) correctly pointed out, the position changes of the modal suffix de happened even after the perfective -le had changed its position, and so it was impossible for the former to affect the latter. At face value, the emergence of the “V [R] O” structure violated the grammatical rule of the co-ordinate verb construction “(Vtr-1 + Vtr-2) + O” as analyzed above. Historically, these two constructions were indeed incompatible with each other. The establishment of the “V [R] O” structure occurred at the expense of the earlier coordinate verb construction. After the Late Medieval Chinese period, the following construction, which was extremely common in Old Chinese, became impossible (for details, see Chapter 4). (15)




學而知之者。 (論語 衛靈公) Xué ér zhī zhī zhě. learn and understand it PRT “I learn and understand it.” (Lun Yu, Wei Ling Gong, 500 BC)

In Contemporary Chinese, there is no conjunction word such as ér to co-ordinate two verbs or adjectival phrases. Additionally, the structure in which two or more verbs share an object, e.g. mǎi chī fàn “buy eat food,” is ungrammatical. Here, a crucial question arises: how could a new structure enter a language by violating an existing grammatical rule? As we discussed in Chapter 5, the earliest VRO instances came into existence through compounding, and at the beginning they

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

334

Aspect and Tense

belonged only to disyllabic compound verbs. Under the influence of the disyllabification tendency, pairs of verb and resultative that were both monosyllabic with a high frequency of co-occurrence first formed disyllabic compound verbs. The compounding happened in the context where no object occurred between the verb and the resultative. At the early stage, every verb–resultative phrase (including the aspect suffixes) had to undergo this process of compounding, and once the number of these compound verbs had reached a certain point, a new paradigmatic structural schema was innovated that was highly productive and influenced other old usages via analogy. At this point, new members occurring in the R position of the VRO form no longer had to undergo their own compounding processes. As a productive new paradigm, the resultative construction was finally established around the eleventh century AD.

13.3 The Emergence of Aspect Suffixes 13.3.1 The Perfective Aspect -Le Let us first consider the results of the investigation by Li and Shi (1997). The scope of the investigation was twenty pieces of text representative of the vernacular language of the Tang and the Five Dynasties (seventh to tenth centuries AD) from a collection edited by Liu and Jiang (1990). Our method was to divide the related examples into two types according to whether the verb and the resultative liǎo are separated: “V liǎo” and “V O liǎo.” Then, we tried to identify any differences between the two forms in the word order change of adverbs/negatives and the frequency of the intervening material, and discovered a striking difference between them, as illustrated in Table 13.2. The result of the total minus both “separated” and “preceded” is the number of examples with no adverbs/negatives. The numbers reveal that there is an internal connection between object absence (adjacency) and the fronting of adverbs/negatives. When verb and liǎo are separated by the object, adverbs/negatives always appear between the object and the resultative liǎo; that is, no fronting of adverbs/negatives was happening yet in this context, as exemplified in (16) and (17):

Table 13.2 The resultative liǎo from the sixth century AD to the tenth Types

Total

Separated by Adv./Neg.

Preceded by Adv./Neg.

V O liǎo V liǎo

69 43

41 (100%) 2 (28%)

0 (0%) 5 (78%)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes (16)

335

填色未了。 (入唐求法巡禮記) Tián sè wèi liǎo. fill color not complete “(Someone) has not completely filled in the color.” (Ru Tang Qiu Fa Xun Li Ji, AD 850)

(17)

嘆之已了。 (廬山遠公話) Tàn zhī yǐ liǎo. praise it already complete “(He) already praised it.” (Lu Shan Yuan Gong Hua, AD 950)

In (16), the verb tián “fill” and the resultative liǎo “complete” are separated by both an object noun sè “color” and a negative wèi “not”; in (17), the verb and the resultative are also separated by an object and an adverb. This strongly suggests that the verb and the resultative liǎo in the separate structure represent two syntactic constituents without any symptoms of fusion. In contrast, however, in the case where the verb and the resultative liǎo are not separated by an object, the two elements also display another property of fusion: the fronting of adverbs/negatives. Within the object-absent structure, 78 percent of all examples with adverbs/negatives had adverbs/negatives fronted to the whole V-liǎo form. Considering that these adverbs/negatives originally occurred between the verb and the resultative, their fronting is a reliable sign that the two elements had become fused and thus allowed no intervening materials. Note that in the very beginning, adverb/negative fronting is found exclusively in the structure without object insertion. Let us consider two adverb/negative fronting examples: (18)

法既付了。 (六祖壇經) Fǎ jì fù liǎo. doctrine already teach complete “The Buddhist doctrine has already been taught.” (Liu Zu Tang Jing, AD 700)

(19)

早說了也。 (祖堂集 長慶和尚) Zǎo shuō liǎo yě. early say complete PRT “I said it a long time ago.” (Zu Tang Ji, Chang Qing He Shang, AD 950)




In (18), the adverb jǐ “already” occurs before fu-liǎo “teach-complete,” and in (19) the adverb zǎo “early” appears immediately prior to shuō-liǎo “say-complete.” Assume that

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

336

Aspect and Tense

when the verb and liǎo started to become fused, they were resistant to the insertion of adverbs/negatives. Therefore adverbs/negatives are fronted to the whole VR phrase. All of these initial properties of V-liǎo phrases indicate that verb and liǎo had a strong tendency to become fused when no object intervened between them. More empirical evidence for the “adjacency” hypothesis came from the results of our comprehensive investigation of one of the most important vernacular texts composed not long before the tenth century AD, Dun Huang Bian Wen, as displayed in Table 13.3. Once again, Table 13.3 shows a remarkable difference between separate and adjacent structures with regard to the fronting of adverbs/negatives. When the verb and the resultative liǎo are separated by an object noun, nearly 60 percent of the examples are also separated by an adverb or a negative. Only two of them have adverbs/negatives fronted to the whole “V O liǎo” form, accounting for merely 5 percent of the examples modified by adverbs/negatives. By comparison, when the verb and liǎo have no object insertion, only approximately 26 percent are also separated by an adverb or a negative. Additionally, there are nine examples of adverb/negative fronting, accounting for 22 percent of the examples modified by adverbs/negatives, seventeen points higher than the fronting tokens of the separate structures. These phenomena also demonstrate that verb and liǎo first became fused in the structure where they are not separated by an object. From another angle, we can test the assumption that the fronting of adverbs/negatives is a reliable formal criterion for judging the first appearances of verb–resultative fusion. The fusion of the verb and the resultative is also conditioned by the number of syllables of the two elements. According to the disyllabicity of the phonological system, when used within the structure where they are adjacent, the verb and the resultative become fused early if they constitute a disyllabic unit. In Table 13.3, we did not take into consideration the number of syllables of verb–resultative pairs without object insertion. If we further divide them into two groups according to the number of syllables, we observe another striking difference with reference to adverb/negative fronting, as shown in Table 13.4. If the verb and the resultative liǎo form a trisyllabic unit (i.e. the verb is disyllabic), of all the examples modified by adverbs/negatives, only 5 percent have either adverbs or negatives fronted to the whole verb–resultative phrase. However, if the verb and the

Table 13.3 The resultative liǎo adjacent to the verb in Dun Huang Bian Wen Types

Total

Separated by Adv./Neg.

Fronted by Adv./Neg.

V O liǎo V liǎo

68 122

40 32

2 9

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes

337

Table 13.4 The resultative liǎo of monosyllabic verbs in Dun Huang Bian Wen Types

Total

Separated by Adv./Neg.

Preceded by Adv./Neg.

V + liǎo = disyllabic V + liǎo = trisyllabic

39 83

17 (95%) 6 (40%)

1 (5%) 9 (60%)

resultative liǎo make a disyllabic unit (i.e. if the verb is monosyllabic), of all the examples modified by adverbs/negatives, 60 percent have adverbs/negatives fronted to the whole verb–resultative phrase. This contrast also serves to support the hypothesis that the disyllabification tendency triggers the fusion of the verb and the resultative. Now let us illustrate the contrast between disyllabic and trisyllabic V-liǎo phrases by using two pairs of examples. Influenced by the number of syllables of the verbs, the same adverb in each pair appears in different places:




(20)

錦帳已鋪了。 (敦煌變文 下女詞) Jǐn zhàng yǐ pū liǎo. brocade curtain already set complete “The brocade curtain has already been set.” (Dun Huang Bian Wen, Xia Nü Ci, AD 800–1000)

(21)

裝束已了。 (敦煌變文 伍子胥變文) Zhuāng-shù yǐ liǎo. dress already complete “(She) has finished dressing.” (Dun Huang Bian Wen, Wu Zi Xu Bian Wen, AD 800–1000)




The adverb yǐ “already” appears prior to the whole predicate pū liǎo “set complete” in (20), when the verb is monosyllabic, but remains between the verb and liǎo in (21), when the verb is disyllabic. (22)




太子才問了。 (敦煌變文 雙恩記) Tàizǐ cái wèn liǎo. crown-prince just ask complete “The crown prince just asked.” (Dun Huang Bian Wen, Shuang En Ji, AD 800–1000)

(23)




鋪置才了。 (敦煌變文 燕子賦) Pu-zhi cai liǎo. spread just complete “(The bed-curtain) was just spread.” (Dun Huang Bian Wen, Yan Zi Fu, AD 800–1000)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

338

Similarly, in (22), the adverb cái “just” appears before the verb and the resultative liǎo when the verb is monosyllabic but between the verb and liǎo when the verb is disyllabic, as illustrated in (23). Once more, the above phenomenon demonstrates that the fusion of the verb and the resultative arises out of a cluster of factors rather than any single factor. When we consider the phonological dimension, we can refine the “adjacency” hypothesis as follows: verb–resultative pairs that form disyllabic units are more likely to become fused in the structure in which they appear adjacently. According to our own investigation, the form V-le-O made its first appearance in the tenth century AD. (24)

萧禧已受了文字。 (乙卯入國奏請) Xiāo Xǐ yǐ shòu-le wénzì. Xiao Xi already receive-PERF article “Xiao Xi has already received the article.” (Yi Mao Ru Guo Zou Qing, AD 1050)

The perfective aspect -le, the reduced form of the full verb -liǎo, lost its original tone value and its final neutralized into a schwa, a typical phenomenon in grammaticalization. When the “V-liǎo” phrase could be followed by an object, liǎo should have already become grammaticalized or lost its lexical status, which led to the phonological reduction “liǎo > le.” In the discussion above, we employed a formal criterion (adverb fronting) to examine the boundary change between verb and liǎo. The design of this criterion is based on such facts as the V-le phrases permitting no intervening material when the form “V O liǎo” changed to “V-le O.” In other words, the emergence of the “VRO” form and the fronting of adverbs/negatives function equally in judging the presence of VR fusion. Only when the boundary between verb and liǎo has been weakened or lost and only when liǎo has lost its lexical status and fused into a verb compound with the verb can the “V-le” phrases have an object. The above analysis suggests that a given resultative may have a different status in different structures. While the resultative liǎo was still a full verb in the structure separated by an object, for example, it had lost its lexical status in the adjacent structure, becoming more grammatical. Lexical and grammatical statuses may coexist for a fairly long time. Neither the fronting of adverbs/negatives nor the postposing of objects involves innovation of syntactic constructions. Since the word order is VO, a “verb + resultative” phrase can precede an object as any ordinary verb can once it is fused into one constituent. Similarly, in the grammar of Chinese, adverbs/negatives always precede verbal elements; thus the originally intervening adverbs/negatives can naturally occur prior to verb–resultative phrases that have become verb compounds due to fusion.

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes

339

In the above analysis, we have addressed the process of the transition from the separable resultative structure to the resultative construction. Recall that the separable resultative structure has four subtypes. The first step was that the verb and the resultative became fused in structure (25), where the verb and the resultative are adjacent. This step can be formulated as follows: (25)

(V) + (R) > (VR)

The bracketing reflects the fusion of the verb and the resultative into one constituent; that is, the boundary loss between the two elements. This is a case of reanalysis, a change that “does not involve any immediate or intrinsic modification of its surface manifestation” (Langacker 1977: 58). Although we do not know exactly when the fusion took place, it should not have been later than the fronting of adverbs/negatives, the second step of the transition process: (26)

(V) + (Adv./Neg. + R) > Adv./Neg. + (VR)

According to our investigation of the -le case, adverb/negative fronting occurred 200 or 300 years earlier than the “V-le O” form emerged. This temporal ordering is related to the degree of fusion. Fronting will happen as soon as the boundary between the verb and the resultative has been weakened (but not necessarily lost). This is the low degree of fusion. However, only when the verb and the resultative have become lexicalized into a verb compound – that is, have reached the high degree of fusion – can the VR phrase precede an object. Thus the third step is as follows: (27)

V + Obj + R > (VR) + Obj or (V + Obj) + (Adv./Neg. + R) > Adv./Neg. (VR + O)

Through a similar process, different types of VR pairs developed into the resultative construction at different times. The fusion of the verb and the resultative must proceed in a structure in which they are adjacent. This provides us with a new view of the transition from an old form to a new one. A transition process often involves word order shift, where some elements move from one position to another. There are two kinds of word order shift. In the first kind, the new word order merely fits another existing pattern; for instance, when adverbs/ negatives that originally separated the verb and the resultative were fronted to the verb– resultative phrases, the resulting pattern complied with the existing structure “adverbs/ negatives + VP.” The only difference between the source and the resulting structure is whether the verb and the resultative are treated as one or two verbal elements. In the second kind, the word order of the resulting pattern is the product of the analogy of the existing structure. The new word order does not already exist in the language; for

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

340

example, the VRO form was entirely new. The R elements were typically intransitive and thus could not occur prior to a patient argument in the grammar of Old and Medieval Chinese. Obviously, the mechanism for the innovation of the VRO form cannot be analogy alone. What follows is our take on this issue. The effect of structure on grammaticalization shows that syntactic structures can be formed without a need for any “jumping” word order change. This point can be illustrated with the perfective aspect marker -le. It was derived from a full verb, liǎo, but the aspect marker and the full verb occurred in quite different patterns: (28)

V + Obj + liǎo (full verb) → V-le + Obj. Note: liǎo = full verb; -le = aspect.

Superficially, it looks as if -le moves from after the object to before the object. This change has led some scholars (Cheung 1977, Cao 1986, Mei 1981) to propose the following hypothesis. First, the VRO pattern already existed before the tenth century AD. Second, the verb liǎo had been semantically bleached into an R in the postobject position before the tenth century AD. Finally, the bleached liǎo was moved from after the object to before the object by analogy with the existing VRO pattern. There are several problems with this hypothesis. First, how could the first group of VRO examples have come into being? The grammar at that time required that two elements preceding the object be transitive (the verb co-ordination principle), but the resultatives were typically intransitive. Second, no evidence shows when or how the verb liǎo underwent semantic bleaching in the postobject position. Finally, analogy is a rule generalization, and thus “irregular” uses will quickly be ruled out. However, this effect was not attested in the replacement of the V O liǎo form by the “V-le O” form. In fact, in the very beginning (around the tenth century AD), “V-le O” examples were quite rare, and the old and new forms coexisted for many centuries. There is no “jumping” word order change at all according to our hypothesis. Motivated by the disyllabification tendency and the high frequency of collocation, the verb liǎo was first fused with its preceding verb in the structure where intervening material was absent or the object was zero. After the first verb and liǎo became fused into a single verb, the combination could take an object, as ordinary verbs do. That is, there was no word order change in the grammaticalization of -le. When -le lost its lexical status and formed a single constituent with the verb, the V-le construction conformed to the syntactic behavior of ordinary verbs that could take an object. There is plenty of evidence that confirms this pathway for the grammaticalization of -le. For instance, intervening adverbs/negatives had to appear before the V-liǎo phrases where the object was zero, a sign of increasing integrity of verb–resultative phrases. Regarding the perfective aspect, reanalysis between the first verb and liǎo first happened in one of the substructures of the old syntactic pattern (the separable resultative structure), and the fused V-liǎo phrases were then generalized by

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes

341

verbs, producing a new form, “V-le O,” at the expense of the other substructures. This is also the process by which the first group of VRO examples emerged. In the process, no elements underwent any “jumping” word order change. In other words, our hypothesis provides a view on word order change without “jumping.” The “no-movement” theory also means that innovation of syntactic patterns typically does not involve violation of existing grammatical rules. Some novel syntactic structures not only do not exist in the system of the old grammar but also contradict certain existing principles. There are two types of grammatical innovation that are differentiated according to structural properties. One is newly introduced devices that express only a new functional domain but neither create a new structure nor conflict with any existing rule; for example, the bǎ construction serves to mark a preverbal patient, but its structure is one instantiation of the paradigm for preposition phrases. The other is novel devices that conflict with certain existing rules. Here, an example is the resultative construction. The novel structures of the second type eventually bring about competition and are eventually established at the expense of the older structure. The resultative construction was new and incompatible with the existing verb co-ordination principle in Medieval Chinese. Superficially, this principle was violated by the VRO form. It is hard to explain how the “ungrammatical form” could come into being when the principle was at work, and it is even more difficult to imagine how the new form with only rare initial occurrences could finally replace the existing paradigm. In the above section we demonstrated that the earliest appearances of VR(O) examples were essentially lexical rather than syntactic, and their existence did not directly contradict the existing syntactic principles because they were at different levels of the linguistic system. The lexical nature was indicated by a property of the initial examples: a given resultative was typically stuck with one particular verb. Examples of this kind increased over time because of independent motivation such as the disyllabification tendency, structural adjacency, and frequent collocation. Once the numbers of these examples became robust enough to give rise to a new syntactic pattern, it gained momentum to develop. This newborn form was then able to compete with the existing principle. Since the new form was backed by other fundamental phonological changes in the linguistic system, it finally superseded the old principle. This event happened in the period from the twelfth century AD to the thirteenth. Consequently, Modern Chinese does not allow two co-ordinated transitive verbs to share an object.

13.3.2 The Experiential Aspect -Guo The experiential -guo underwent a grammaticalization process similar to that of the perfective -le. The verb guò originally meant “pass,” and could be followed by a locative

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

342

or temporal word as its object, such as guò běi-mén “pass the northern door.” It first became a preposition that often introduced a locative or temporal phrase after the matrix verb, as illustrated in (29) and (30):




(29)

蛤蟆跳過雀兒浴。 (韓愈 贈侯喜) Hámá tiào guò quèer yù. toad jump across sparrow pond “The toad jumped across the sparrow pond.” (Han Yu, Zeng Hou Xi, AD 800)

(30)

杜鵑, 你休得叫過通宵。 (張協狀元) Dùjuān, nǐ xiūdé jiào guò tōngxiāo. Cuckoo you do-not call through all-night “Cuckoo, you cannot call all night long.” (Zhang Xie Zhuang Yuan, AD 1200)

In the above examples, guò as a preposition governed the following object, a property that was inherited from its original verb transitivity. It seems quite natural to assume that the experiential -guo developed directly out of the context illustrated by the above two examples, but this actually was not the case. Here, guò and its object first form an immediate constituent and there was a clear boundary between the verb and this immediate constituent. That is, the structural hierarchy of the preposition guò is different from that of its aspect suffix use, as shown below: (31)

(a) V + (guòpreposition + NP) (b) (V + guosuffix) + NP

Therefore the reanalysis involved a restructuring of the construction which consisted of two steps: first, the boundary between the verb and the preposition guò was deleted; second, a boundary was created between the aspectual guò and the following object. In fact, the experiential -guo did not develop in this way. According to Wang (1989: 86‒101), after the seventh century AD, guò started to be used as an experiential marker but was not followed by any object noun. As in the case of the perfective -le, -guo grammaticalized in contexts where the object noun was absent (but it disappeared after the whole phrase “V-guò”). Before the twelfth century AD, none of the instances of the experiential suffix guò were followed by an object noun, as shown below: (32)

本司檢過。 (入唐求法巡禮記) Běn sī jiǎn-guò. this department inspect-EXPE “The department inspected it.” (Ru Tang Qiu Fa Xun Li Ji, AD 850)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes (33)

343

涉獵看過。(朱子語類卷一) Shèliè kàn-guò. cursorily read-EXPE “(I) cursorily read (it).” (Zhu Zi Yu Lei, Chapter 1, AD 1200)

Until the thirteenth century AD, the verb plus the experiential -guo gradually began to be followed by an object noun, as the following examples show: (34)




又請眾人喫過酒飯方散。 (元話本選輯 玉堂春落難逢夫) Yòu qǐng zhòngrén chī-guò jiǔ fàn fāng sàn. also invite everyone eat-EXPE wine-food only-then disperse “(He) also invited everyone to eat food and drink wine before leaving.” (Yuan Hua Ben Xuan Ji, Yu Tang Chun Luo Nan Feng Fu, AD 1350)

(35)




他吃過了一杯茶。(明代話本 韓秀才乘亂聘嬌妻) Tā chī-guo-le yībēi chá. he drink-EXPE-PERF one-glass tea “He has drunk a glass of tea.” (Ming Dai Hua Ben, Han Xiu Cai Cheng Luan Pin Jiao Qi, AD 1650)

Due to the different rates of development across dialects, in the vernacular novels written around the sixteenth century AD by authors who spoke the Wu dialect, the “V + O + guò” structure was still attested, as illustrated in (36): (36)




我又不曾擔水過的。 (醒世恆言 兩縣令競義婚孤女) Wǒ yòu bù-céng dān shuǐ guò de. I also not-ever carry water EXPE PRT “I didn’t ever carry water (with a shoulder pole).” (Xing Shi Heng Yan, Liang Xian Ling Jing Yi Hun Gu Nü, AD 1650)

The above phenomenon reveals how a grammatical morpheme could come into existence, as was often the result of an interaction between specific contexts and the grammatical system at the time. When guò was used as an aspect marker, it bore no relation, either semantically or syntactically, to the following object noun, but formed an immediate constituent with the preceding verb, denoting the end point of the action. Constrained by the rule governing the co-ordinate verb construction at the time, “V-guoexperiential” could not be followed by an object NP in the early stage of its development as an aspect marker. Once the resultative guò became grammaticalized into an aspect suffix at the expense of its verb status, it behaved like a compound-like verb that could be followed by an object noun. That is, the experiential -guo underwent

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

344

its own complete grammaticalization process rather than its grammaticalization process being analogous to the perfect -le, which had become grammaticalized 200 years earlier.

13.3.3 The Progressive Aspect -Zhe In Old Chinese, the verb zháo (the verb form of the progressive -zhe) meant attach, reach or place. After the first century AD, this verb developed a use as a preposition, usually introducing a phrase expressing a certain locative resultative (i.e. end point) of the verb. This development from verb to preposition resembled that of the experiential -guo, as discussed above. However, the specific contexts for their grammaticalization were different. The verbal and the prepositional uses of zháo were often separated from the main verb by the object of the verb, as illustrated in (37) and (38), but there was no object between the main verb and the verbal and prepositional uses of -guò, as we saw above:




(37)

輒含飯著兩頰邊。(世說新語 德行) Zhé hán fàn zháo liǎng jiá biān. often contain food into two cheek side “He often contained food inside his two cheeks.” (Shi Shuo Xin Yu, De Xing, AD 450)

(38)

埋玉樹著土中。 (世說新語 傷逝) Mái yù-shù zháo tǔ zhōng. bury jade-tree in soil inside “(He) buried the jade tree in the soil.” (Shi Shuo Xin Yu, Shang Shi, AD 450)




Compared with the experiential -guo, the progressive -zhe needed to take an extra step to develop into an aspect suffix: the movement of the intervening object noun to another location. As we saw in Chapter 6, due to the establishment of the resultative construction, the boundary between the verb and the resultative preposition phrase was weakened via reanalysis and the object position between them was finally eliminated. The patient noun that originally occurred between the verb and the preposition phrase had to be introduced by other grammatical devices; for instance, if it was definite, it was typically introduced by the disposal construction (for details, see Chapter 9). The grammaticalization of the progressive -zhe underwent an intermediate stage in which it was used in the existential structure. In the period from the sixth century AD to the ninth, a new pattern of existential structure took shape with the schema “Locative + V + zháo + NP,” which means “there is NP in the location,” as illustrated in (39). However, the noun phrase was often absent, as illustrated in (40):

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes (39)

345

猶掛著唇齒在。 (六祖壇經) Yóu guà zháo chún-chǐ zài. still hang at lip-tooth in “The lip and teeth are still hanging on there.” (Liu Zu Tan Jing, AD 700)

(40)

於西間壁上題著。 (六祖壇經) Yú xī-jiàn bì shàng tí zháo. On west-room wall above inscribe on “(He) again got someone who could write on the wall of the west room.” (Liu Zu Tan Jing, AD 700)

It was in the context shown in (39) that the verb and the preposition zháo became reanalyzed, which made zháo become an aspect suffix. The earliest examples of the progressive were attested in texts composed in the thirteenth century AD, as illustrated below: (41)

擔著一對酒桶。 (宣和遺事元集) Dān-zhe yī duì jiǔ tǒng. carry-PROG one pair wine barrel “They were carrying a pair of wine barrels.” (Xuan He Yi Shi Yuan Ji, AD 1300)

(42)




見他戰篤速驚急烈慌慌走著。 (元刊雜劇 陳季卿悟道竹葉舟) Jiàn tā zhàn-dǔ-sù jīng-jí-liè huāng-huāng zǒu-zhe. see he panic-state hastily walk-PROG “It was seen that he was hastily walking in a panic.” (Yuan Kan Za Ju, Chen Ji Qing Wu Dao Zhu Ye Zhou, AD 1300)

Historically, the progressive -guo emerged approximately three centuries later than the perfective -le, which became an aspect suffix around the tenth century AD.

13.3.4 The Modal and Inchoative Suffixes -De and -Qǐlái Roughly around the time of the emergence of the aspect suffixes discussed above, a modal suffix was also introduced into the language, and all of these changes were motivated by the same diachronic events. The term “epistemic modality” is an important semantic and grammatical category referring especially to the attitudes of speakers toward the factual content of an utterance, e.g. possibility and definiteness. Different languages or the same languages at different periods usually employ different devices to express the modal category. For example, English mainly uses modal auxiliaries, e.g.

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

346

may, can, and shall, to perform this function. Likewise, before the ninth century AD, the Chinese language relied mainly on either auxiliary verbs or sentence-final particles to express various kinds of modality. As the resultative construction developed, the circumstances enabled a morphological form -de to express modality, with grammatical status similar to that of the aspect suffixes.2 Wang (1989: 102‒121) and Jiang (2005b: 194‒204) outlined the development process of the modal suffix -de. Here, we focus on what factors made it possible for this modal suffix to occur. First, let us consider the major functions of the modal suffix -de in Contemporary Chinese (for a detailed description, see Zhu 1982: 133‒137). (a) Possibility and capability. When the modal suffix -de was not followed by another resultative, the corresponding negative form was “V-bu-de” (lit. verb-not-able), as illustrated in (43). When the modal suffix -de was followed by another resultative, the corresponding negative form was “V-bu-R” (V-not-R), where the suffix -de had to be absent, as illustrated in (44): (43)

(a) 這東西喫得。 (現代漢語) Zhè dōngxī chī-dé. this thing eat-able “This thing is edible.” (b) 這東西喫不得。 (現代漢語) Zhè dōngxī chī-bù-dé. this thing eat-not-able “This thing is not edible.” (Contemporary Chinese)

(44)

(a) 他搬得動那張桌子。 (現代漢語) Tā bān-dé-dòng nà-zhāng zhuōzi. he carry-able-move that-CL table “He can move the table.” (b) 他搬不動那張桌子。 (現代漢語) Tā bān-bù-dòng nà-zhāng zhuōzi. he carry-not-move that-CL table “He can’t move the table.” (Contemporary Chinese)

2

This modal suffix and the resultative suffix share exactly the same phonological form, -de, both originating from the same lexical source. However, they are quite different in function. The following section will discuss the resultative one.

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes (b)

(45)

347

Factual resultative suffix -de. The suffix -de can also introduce a resultative clause that describes what is actually caused by the matrix verb. In this case, the resultative clauses are usually structurally complex, similar to normal declarative sentences, as illustrated below: 這篇文章寫得很清楚。(現代漢語) Zhè-piān wénzhāng xiě de hěn qīngchǔ. this-CL article write DE very clear “This article is very clearly written.” (Contemporary Chinese)

(46)

他逗得我們哈哈大笑。 (現代漢語) Tā dòu de wǒmen hāhā dà-xiào. he amuse DE we loud laugh-out “His joking around made us laugh out loud.” (Contemporary Chinese)

All the structures depicted above did not exist before the sixth century AD, and each was introduced into the language at a different time. The whole set of the varying usages of the modal -de took approximately a millennium to develop, and the process can be outlined as follows. In Old Chinese, the verb dé (i.e. the verb form of the suffix -de) was a verb of get or obtain that could be used as the main verb of a sentence, as illustrated in (47), or as the second verb of a co-ordinate verb construction, as illustrated in (48), where it was a transitive verb and bore a VO relation to the following object noun:




(47)

工師得大木。(孟子 梁惠王) Gōng-shī dé dà mù. engineer obtain big wood “The engineer gets a big piece of wood.” (Meng Zi, Liang Hui Wang, 300 BC)

(48)

求牧與芻而不得。 (孟子 公孫醜) Qiú mù yǔ chú ér bù dé. seek animal and grass and not obtain “(He) sought animals and grasses but didn’t get them.”




(Meng Zi, Gong Sun Chou, 300 BC) (49)




堯射得之。(論衡 感虛) Yáo shè dé zhī. Yao shoot get it “Yao shot and got it.” (Lun Heng, Gan Xu, AD 100)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

348

Note that in (46), the verb dé “get” was connected by the conjunction ér, reflecting the grammatical rule of verb co-ordination in Old Chinese (for details, see Chapter 4). In this situation, no reanalysis of the preceding verb and dé could happen. Once the conjunction word disappeared at a later stage, as exemplified in (48), it became possible for the verb dé to be grammaticalized into a modal suffix in the second verb position of the serial verb construction. Although the verb dé “get” developed from an auxiliary verb to express capability or possibility, it was limited to preverbal position, as exemplified in (50). After the second century AD, the modal suffix dé could be used in the second verb position in a serial verb construction, as illustrated in (51):




(50)

使不得耕耨。 (孟子 梁惠王) Shǐ bù-dé gēng-nòu. make not-can cultivate “(He) made them unable to cultivate.” (Meng Zi, Liang Hui Wang, 300 BC)

(51)

使妾摇手不得。 (漢書 外戚傳) Shǐ qiè yáo shǒu bù-dé. make me shake hand not-able “You had me unable to shake my hands.” (Han Shu, Wai Qi Zhuan, AD 100)




The example in (51) reflects the grammatical rule at that time that required an intransitive verb to occur after the matrix verb and the object noun. After the sixth century AD, the first two structures that are described in (43) and (44) in Contemporary Chinese came into existence, as illustrated in (52) and (53):




(52)

亂後誰歸得? (杜甫 得舍弟消息) Luàn hòu shéi guī dé? war after whoever return able “Who can come back after the war?” (Du Fu, De She Di Xiao Xi, AD 750)

(53)

故鄉歸不得。 (杜甫 春遠) Gùxiāng guī-bù-dé. hometown return-not-able “I cannot return to my hometown.”




(Du Fu, Chun Wang, AD 750) The above structure became possible when the resultative construction was underway to its final firm establishment. Only then could the modal suffix, which occurred in

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

The Emergence of Aspect Suffixes

349

postverbal position, come into existence. Soon after the above structure occurred, the “V-de-R” structure emerged, as illustrated in (54) and (55): (54)




旗下依依認得真。 (敦煌變文 捉季布傳文) Qí xià yīyī rèn-dé-zhēn. flag under certainly recognize-able-clear “They are clearly recognizable under the flag.” (Dun Huang Bian Wen, Zhuo Ji Bu Zhuan Wen, AD 800–1000)

(55)

此條記得極好。 (朱子語類卷五十九) Cǐ tiáo jì dé jí-hǎo. this item memorize DE extreme-well “This line was memorized extremely well.” (Zhu Zi Yu Lei, Chapter 59, AD 1200)

The suffix -de in (54) is a modal, referring to capability or possibility, but in (55) it introduces a factual clause to describe the resultative state of the verb. The development of the “V-de-R + O” structure deserves special attention, as it involved two steps of reanalysis. The object noun could be absent, as illustrated in (56). However, if there was an object, the structure could only be “V-de O R,” as illustrated in (57): (56)




氣象四時清, 無人畫得成。 (方幹 處州洞溪) Qìxiàng sì shí qīng, wú-rén huà-dé-chéng. landscape four season beautiful nobody draw-able-complete “The landscape is beautiful in the four seasons. Nobody can draw it completely” (Fang Gan, Chu Zhou Dong Xi, AD 900)

(57)




十三學得琵琶成。 (白居易 琵琶行) Shí-sān xué-dé pípá chéng. thirteen learn-DE lute complete “She completely mastered the lute at the age of thirteen.” (Bai Ju Yi, AD 850)

Around the twelfth century AD, the construction of both “V-not-de + O” and “V-de-R + O” began to emerge, as illustrated in (58) and (59), respectively. At this point, all subtypes of the resultative constructions had been developed. (58)




禁止不得淚, 忍管不得悶。 (黃庭堅 卜算子) Jìnzhǐ-bù-dé lèi. control-not-able tear “He cannot control his tears.” (Huang Ting Jian, Bo Suan Zi, AD 1100)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

350 (59)

須看得見那物事。(朱子語類卷十三) Xū kàn-dé-jiàn nà wùshì. should look-able-see that thing “You should be able to see the thing.” (Zhu Zi Yu Lei, Chapter 13, AD 1200)

13.4 Inchoative and Continuous Aspects As mentioned above, the creation of the syntactic position between the verb and the object that could contain an intransitive element to describe the situation resulting from the action of the preceding verb was one of the most important changes after Late Medieval Chinese. This change exerted a far-reaching effect on the development of Chinese grammar. The position could be occupied by numerous lexical items that may be regarded as some kind of resultative that describes the state resulting from the action of the matrix verb but bears no action–patient relation to the following object. Some lexical items, such as téng “painful” in xiào-téng dùzi “laugh-painful stomach” (the stomach hurts because of laughing too much), were only sporadically used in this position, but some resultatives, such as the aspect suffixes and the modal suffix discussed above, might have developed into stable grammatical devices, due to their semantic suitability in relation to the verb and the frequency of their uses as resultatives. In addition, two minor aspects are also worth mentioning: (a) inchoative aspect -qǐlái (lit. “get up”) and (b) continuous aspect -xiàqù (lit. “go down”). The former means the start of an action, and the latter refers to the continuation of an action; both were introduced into the language after the fifteenth century AD, when the resultative construction and the aspect system were already well developed. First, let us consider two early uses of the inchoative aspect: (60)

恐怕火盆內有小炭延燒起來。(水滸傳十回) Kǒngpà huǒpén nèi yǒu xiǎo tàn yánshāo-qǐlái. afraid brazier in have smallcharcoal burn INCH “I’m afraid there is a small charcoal in the brazier to burn up.” (Shui Hu Zhuan, Chapter 10, AD 1400)

(61)

老僧聞言就喜歡起來。 (西遊記二十回) Lǎo sēng wén yán jiù xǐhuān-qǐlái. old monk hear word then like-INCH “The old monk liked it immediately after he heard the words.” (Xi You Ji, Chapter 20, AD 1550)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Inchoative and Continuous Aspects

351

The earliest instances of the continuous aspect -xiàqù were attested in the nineteenth century AD, nearly five centuries later than the inchoative aspect and almost a millennium later than the perfective aspect. Let us consider one of the earliest examples of the continuous aspect. (62)

便靜靜兒的聽他唱下去。(兒女英雄傳三十八回) Biàn jìng-jìng-er-de tīng tā chàng-xiàqù. then quietly listen she sing-CONT “He quietly listened to her continue singing.” (Er Nu Ying Xiong Zhuan, Chapter 38, AD 1850)

Regarding the words in the X position of the VXO construction, there was a continuum from concrete lexical items to highly grammaticalized markers. This diachronic evidence shows that any boundary between lexical items and grammatical morphemes is by nature arbitrary. As we already know, under the influence of the disyllabification tendency, the resultative construction was able to emerge and develop through the compounding of the verb and the resultative (for details, see Chapter 5). During the process of the development of the resultative construction, the active and productive rule was still verb co-ordination, which regulated the pathways of each “verb + resultative” pair that took place, including the aspect suffixes. In the period from the fifth century AD to the twelfth, the rule worked at sentence level, while the compounding of verb–resultative phrases happened at the word formation level. The VRO construction as a new paradigmatic structure that entirely replaced the verb co-ordination rule was quite a late event, approximately after the twelfth century AD. After that, the co-ordinate verb construction became ungrammatical and was eventually ruled out by the new grammatical principle, and the VRO pattern became a highly productive paradigm that leveled irregular and less frequent verb–resultative forms. For instance, as mentioned earlier, the verb bì “complete” could occur after the VO phrase to refer to the completion of the action represented by the verb, a usage attested as early as the fifth century AD. Since the verb liǎo became dominant in expressing this function, bì was restricted to the written language. In texts composed around the sixth century AD, however, bì could also occur between the verb and the object, as illustrated in (63). This use of bì did not come out of its own development but clearly was due to analogy with the perfective -le and the like. (63)

西門慶喫畢茶。(金瓶梅三十回) Xīmén Qìng chī-bì chá. Ximen Qing drink-finish tea “Ximen Qing finished drinking tea.” (Jin Ping Mei, Chapter 30, AD 1550)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

352

The verb bì “finish” was only a lexical item used in the written register and never developed into an aspect suffix. The establishment of the “VXO” structure, where X could be a suffix of the verb, also made it possible to borrow the schema of word formation from other languages. One of the representative examples in this regard is the suffix -ize and other verb suffixes of English, which function to turn nouns or adjectives into verbs, e.g. modern ~ modernize, nasal ~ nasalize, fossil ~ fossilize. In the past century, Chinese has created a suffix, huà, that is extremely productive in making a verb out of a noun or adjective, and some of the derived verbs even have no counterparts in English, as shown below: 綠化 lǜ-huà “green-ize” 歐化 ōu-huà “Europeanize” 機械化 jīxiè-huà “mechanize” 中國化 zhōngguó-huà “China-ize”

美化 měi-huà “beautify” 污名化 wūmíng-huà “stigmatize” 電子化 diànzǐ-huà “electronicize” 年輕化 niánqīng-huà “youthful-ize”

The derived verbs above behave like transitive verbs, which can be followed by an object noun. In our view, only after the establishment of the aspect patterns could this sort of schema-borrowing of word formation become possible. In other words, in the grammar of Old Chinese, no suffixes were allowed to be attached to the verb so that the structure “V-huà + O” was ill-formed and thus ruled out by the old grammatical system. As far as grammatical schema are concerned, language borrowing cannot be allowed if it violates the grammatical rule of the receiving language.

13.5 Interaction between Aspect Suffixes Cross-linguistically, aspect markers vary from language to language, and within a language the meaning and function of a particular aspect marker depends on the contrast with other aspect markers. In Chinese, aspect suffixes emerged after the tenth century AD, but they did not all come into existence at the same time. As discussed above, the perfective -le became grammaticalized around the tenth century AD, and the experiential -guo and the progressive -zhe were introduced into the language approximately 200 years later. According to Wang (1989: 102‒121), it took approximately eight centuries to set up the division among these aspect suffixes. He observed that there was some confusion of the uses of the perfect and progressive aspects; for example, in (64) the perfective -le was used as the progressive suffix, and in (65) the progressive -zhe was used as the perfective aspect: (64)




太后指了天曰。 (五代史平話 晉史) Tàihòu zhǐ-le tiān yuē. emperor-mother point-PROG sky say “The mother of the emperor was pointing to the sky and said . . .” (Wu Dai Shi Ping Hua, Jin Shi, AD 1100)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Interaction between Aspect Suffixes (65)

353

楊志因等候我了, 犯著這罪。 (宣和遺事元集) Yáng Zhì yīn děnghòu wǒ le, fàn-zhe zhè zuì. Yang Zhi for wait I PERF commit-PERF this crime “Yang Zhi committed this crime because he waited for me.” (Xuan He Yi Shi Yuan Ji, AD 1300)

Additionally, there were some confusing usages of the perfective -le and the experiential -guo. This phenomenon has not been reported in the literature, probably because the difference between these two aspects was too subtle for their mixtures in history to be noticeable. Both referred to the completion of an action in the past, and their characteristics were that the perfective -le emphasized the current relevance and the experiential -guo focused on the end point of the action. In certain contexts even native speakers have difficulty telling one from the other. However, these two aspects were indeed confused during their development. For example, the experiential -guo in (66) is used in an imperative sentence whose action has not yet started, where only the perfective -le can be used according to the grammar of Contemporary Chinese: (66)

武二都記得嫂嫂說的話了, 請飲過此杯。 (水滸傳二十四回) Qǐng yǐn-guò cǐ-bēi. please drink-EXPE this-CL “Please drink this cup.” (Shui Hu Zhuan, Chapter 24, AD 1400)

Furthermore, in Contemporary Chinese, psychological verbs, such as rènshí “know,” xuéhuì “learn” and dǒngdé “understand,” can be suffixed only with the perfective -le and not the experiential -guo because they lack the semantic characteristic – a clear end point that the experiential aspect requires. As exemplified below, however, in texts composed in the nineteenth century AD, the verb rènshí “know” was suffixed by the experiential -guo, where the perfective -le should be used now: (67)

不過在小錢莊時認識過幾個數目字。 (二十年目睹之怪現狀七十九回) Bùguò zài xiǎo qiánzhuāng shí rènshí-guò jǐ-gè shùmù but in small bank time learn-EXPE several-CL numeral zì. character “But he just learnt some numeral characters when he worked in a small bank.” (Er Shi Nian Mu Du Zhi Guai Xian Zhuang, Chapter 79, AD 1900)

If we consider the situation in Chinese dialects, the picture appears extremely messy. Grammatical morphemes typically result from lexical items, and the lexical item must satisfy two conditions: possessing semantic suitability and frequently appearing in

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

354

Table 13.5 The functions of aspect markers and their lexical sources Grammatical category

Function of suffix

Meaning of lexical item

Perfective -le Experiential -guo Progressive -zhe Inchoative -qǐlái Continuous -xiaqu Possible -de

Action in the past with current relevance Action in the past ongoing action Start of an action Continuation of an action Possibility of carrying out action

“finish” “pass” “in, at” “get up” “go down” “get, can”

particular contexts. However, these conditions are applied only to those that really undergo a grammaticalization process. For this sort of aspect suffix, there are always transparent semantic relations between the functions of the aspect suffixes and the meanings of their lexical sources (see Table 13.5). We suppose that all the above grammaticalization events genuinely happened in standard Mandarin (i.e. the northern family of dialects), but not all the dialects actually underwent these developments; instead, they simply borrowed the aspect system from the standard language. As a result, there are many mismatched aspect suffixes with regard to lexical sources. As we know, Chinese has eight dialect families and hundreds of subdialects. For the sake of simplicity, we focus only on how the perfective aspect is expressed in different dialects, and, as we see below, all six types of different aspect device are used to express the perfective aspect, as a stable grammatical device in those dialects. (a) Perfective aspect expressed by the progressive -zhe, e.g. Cantonese, as illustrated below: (68)

渠衝咗涼。 (廣東話) Kʰøy tʃʰʊŋ-tʃɔ lœŋ. he take-PERF shower “He has taken a shower.” (Huang 1996: 175, Guangdong dialect, a variant of Cantonese)

(b) Perfective aspect expressed by the modal suffix -de, e.g. Guiyang dialect, as illustrated below: (69)

他在那家鋪子買得兩包煙。 (貴陽話) Tā zài nà-jiā pùzi mǎi-dé liǎng-bāo yān. he in that-CL grocery buy-PERF two-CL cigarette “He bought two packs of cigarettes in that shop.” (Huang 1996: 175, Guiyang dialect)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Interaction between Aspect Suffixes (c)

(70)

355

Perfective aspect expressed by the inchoative -qǐqù, e.g. Sichuan dialect, as illustrated below: 他把那個人抓起去。 (四川話) Tā bǎ nà-gè rén zhuā-qǐqù. he DIS that-CL man grab-PERF “He grabbed the man.” (Huang 1996: 175, Sichuan dialect)

Quite surprisingly, verb reduplication is also recruited to express the perfective aspect in some dialects. As discussed in the subsequent section, only quite recently (around the thirteenth century AD) was verb reduplication introduced into the language, defined as a delimitative aspect, expressing the short duration of an action. However, in quite a few dialects, particularly those of the Wu family, such as the Zhejiang dialect, the reduplicated verb form can function to express the meaning of the perfective aspect, as illustrated below: (71)

信寄寄就來。 (浙江吳方言) səŋ tsï-tsï ʒiu la. letter mail-mail then come “I will come after I have mailed the letter.” (Huang 1996: 175, Wu dialect in Zhejiang)

Some dialects even adapt the diminutive marker -er to express the meaning of the perfective aspect. As we will discuss in Chapter 19, the word ér “child” started its grammaticalization process toward a diminutive marker, an apparatus attached primarily to nouns, after the sixth century AD. Interestingly, it has been used as a perfective aspect in some dialects, such as the Haiyang and Muping dialects (Huang 1996: 175), as exemplified below: (72)

腊月三十日打儿个兔子。 (牟平方言) Làyuè sānshí rì dǎ-er gè tùzǐ. December thirtieth day shoot CL rabbit “(He) shot a rabbit on the thirtieth day of the lunar month of December.” (Huang 1996: 175, Muping dialect)

It seems that these dialects do not care where the forms come from and what their original functions were; they simply adopt a form to perform the task of expressing the meaning of the perfective aspect, a grammatical category that was developed in the northern dialect and borrowed by many other dialects. Grammatical borrowing occurs

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

356

much more easily across dialects than across languages, creating many interesting phenomena that are theoretically significant.

13.6 Diverse Forms across Dialects It is well known that a grammaticalization process often proceeds along a cline (Hopper and Traugott 2003: 6): (73)

Content item > grammatical word > clitic > inflectional affix > zero

However, this cline should be considered with caution because it reflects only one type of change that has undergone a complete development process. According to Harris and Campbell (1995: 50), there are only three basic mechanisms of syntactic change: reanalysis, extension, and borrowing. If a language or a dialect simply borrowed a grammatical category from another, what would happen to the phonological form? Apparently, the borrowed item does not have to go through all the stages of the grammaticalization process. Because of the vast geographic area and the extreme diversity of dialects in China, grammatical borrowing across dialects, similar to loanwords, happened often. Grammatical borrowing is different from contact-induced change because it does not have to happen between geographically adjacent dialects. In this section, we deal with various forms of grammatical borrowing across dialects. We often find that for a given grammatical category, some dialects employ just a sound feature to express its function, whereas standard Mandarin makes use of a function word; e.g. the plural is expressed by the morpheme -men in standard Mandarin but by a highpitched tone in Cantonese, by reduplication in the Chongqing dialect, and so forth. If we quickly jump to the conclusion that the plural markers in these dialects are more grammaticalized than those in standard Mandarin simply on the basis of their phonological representation, we actually ignore the many possible ways that a language community can acquire a grammatical device. In addition to grammaticalization, a grammatical device can be obtained through phonological derivation (e.g. the emergence of distal demonstratives; see Chapter 18 for details), borrowing from other dialects, or developing it by itself. In what follows, this point is illustrated by the rich phonological representations of aspect markings. Some dialects seem to rely purely on phonological form to express the function of the perfective aspect, such as the addition of a sound, a change in tonal value, or a vowel change. (a) In the Hakka dialect, the vowel [e] is added to the matrix verb to express the completion of the action with current relevance (Huang 1996: 175): (74)

渠食[e] 飯。 (客家話) kʰi sət-e fan. he eat-PERF food “He has had a meal.” (Hakka dialect)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Diverse Forms across Dialects (b)

(75)

357

Some dialects express the function of the perfective aspect by changing the tonal value of the matrix verb. According to Huang (1996: 175), in the Shangxian dialect (spoken in northwestern China), there are four tones, each of which has a lengthened tone with the following tonal contours:3 21~3231, 24~2141, 53~5231, and 55~551. The verb bearing the long tone is equivalent to the verb plus the meaning of the perfective -le in standard Mandarin; thus no other perfect suffix is needed. In the following example, the tonal value of the verb tʂ‘ǝ21 “eat” can be lengthened to tʂ‘ǝ3231 to express the completion of the action. 他喫[tʂ‘ǝ21→3231]飯啦。 (晉方言) He tʂ‘ǝ3231 fan la. he eat-PERF meal PRT “He has eaten a meal.” (Jin dialect)

A similar strategy to express the perfective aspect is found in Cantonese (Huang 1996: 175), a southern dialect that is far from the Shangxian dialect. Specifically, the original tone of the matrix verb is changed to a high-pitched tone value of 35 to express the function of the perfective aspect, as illustrated in (76): (76)

(a) ŋɔ13 sik22 la22. I eat PRT “I am going to eat.” (b) ŋɔ13 sik35 la22. I eat-PERF PRT “I have had a meal.” (Huang 1996: 175, Cantonese)

(c)

3

Altering vowels to express the perfective aspect. According to He (1989: 58), in the Huojia dialect (belonging to the Jin family), the vowel or the rhyme of the syllable of the verb is altered to express the function of the perfective aspect. This dialect has a phonological rule to derive the form of the perfective aspect from its base syllable. Here we provide only two examples to illustrate this point. In the following two pairs of examples, the first example has the base form of the verb, which refers to an action that has not happened yet, and the second example has the derived forms of the verb, which refer to the completion of the action with current relevance.

The numbers stand for the relative pitches of tone values, where 1 represents the lowest pitch and 5 the highest pitch. In addition, two numbers identify the contour of a tonal value.

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

358 (77)

(a) 我要買蔥。 (獲嘉方言) Uaʔ33 mai53 ts‘uŋ33. I buy onion “I am going to buy onions.” (b) 我買了蔥。 Uaʔ33 mɛ53 ts‘uŋ33. I buy-PERF onion “I have bought onions.” (He 1989: 54, Huojia dialect)

(78)

(a) 我要看電影。 (獲嘉方言) Uaʔ33 kan13 tian13-iŋ53. I see movie “I am going to see a movie.” (b) 我已經看了電影。 Uaʔ33 k‘ã13 tian13-iŋ53. I see-PERF movie “I have seen a movie.” (He 1989: 54, Huojia dialect)

It happened from time to time in history that when the language developed to a certain point where the conditions became mature or certain pragmatic devices were ready, the language community then recruited those semantically suitable lexical items to develop the grammatical morpheme that was favored or needed by the grammatical system at the time. The change first occurred in the central dialect of the language, namely standard Mandarin, which was spoken in the vast majority of the country, at a particular period. This situation was something like a general call, with every dialect having to respond to the new grammatical innovation. For example, there were a set of grammatical innovations that did not exist before the fifth century AD, such as the classifier system, the plural, the diminutive, and the system of the aspect suffixes. No dialect could resist these changes and remain in the former state of its own grammar, but the methods of responding to these innovations varied greatly from dialect to dialect. Now, let us summarize how many methods standard Mandarin used to innovate grammatical apparatuses and how the other dialects reacted to represent the innovated grammatical devices. (a) Grammaticalization. This is a major channel for a language to acquire an innovative device. For example, the disposal bǎ grammaticalized from a verb meaning “take”; the structural particle de (marking relative clause,

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Diverse Forms across Dialects

(b)

(c)

(d)

(e)

359

genitive, and associative) from a demonstrative; the plural men from a classifier for family; and the diminutive er from the concept “child.” Phonological derivation. This method is typically applied to two closely related grammatical categories within a grammatical domain. For instance, the proximal and distal demonstratives are crucial for the daily communication of any language community. It is unlikely that each side grammaticalized from different lexical sources, undergoing independent development and finally reaching completion at different times. As we will see in Chapters 17 and 18, the proximal demonstratives zhè in the northern areas and gè in the southern areas developed out of a general classifier in Late Medieval Chinese, and they acquired their distal demonstrative usages simply through a phonological rule. Conventionalization of the function and the phonological form. This method is different from phonological derivation, termed “grammatical borrowing across dialects,” and it is like coining a new word whose meaning and phonological form are purely conventional or entirely arbitrary. This often happened to dialects in isolation from other dialects in ancient times. These dialects simply chose a phonological form to express the function of the newly invented grammatical categories in standard Mandarin. For instance, as we saw in related sections, the plural is expressed by a nasal consonant, the diminutive by a high-pitched tone, the perfective aspect by a vowel shift, and so forth. Adapting the form from the major dialect. This represents a special kind of language borrowing across dialects that often results in a mixture of function and form. For example, the perfective aspect is expressed by verb reduplication, by the modal suffix de, and even by the diminutive marker -er in different dialects. The language communities in these dialects have treated the grammatical categories as loanwords and have often chosen different phonological forms to encode the function of the borrowed grammatical categories. Regional grammaticalization. For a given grammatical category, some dialects choose their own lexical sources to develop a grammatical morpheme. As we saw above, for example, the progressive marker -zhe grammaticalized from the preposition “in” or “to” in Medieval Chinese. Similarly, many dialects have chosen other prepositions, either zài “in/at” or dào “at/to,” to express the function of the progressive aspect. For example, the preposition dào developed into a progressive marker in the Hubei, Sichuan, Guiyang, and Yunnan dialects, as illustrated in (79), and the preposition zài “in/at” grammaticalized into a progressive aspect in the

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

360

Chaoxian, Hefei, Huoqiu, and Southern Min dialects and is still a free word rather than a suffix because it can be separated by the object, as illustrated in (80) (for a detailed description, see Huang 1996: 175): (79)

他喫到飯。 (雲南方言) Tā chī-dào fàn. he eat-PROG meal “He is eating a meal.” (Huang 1996: 175, Yunnan dialect)

(80)

他看書在。 (合肥話) Tā kàn shū zài. he read book PROG “He is reading a book.” (Huang 1996: 175, Hefei dialect)

According to Huang (1996: 204), the progressive use of the preposition can be dated to Late Medieval Chinese, for example: (81)




兩處也應相憶在。 (白居易 郡中閑獨) Liǎng chù yě yīng xiāng yì-zài. two place also should mutual miss-PROG “The couples at different places should be missing each other.” (Bai Ju Yi, Jun Zhong Xian Du, AD 850)

That is, the two synonymous prepositions zháo and zài were competing for the progressive aspect at the time. The former eventually superseded the latter in northern areas, but the latter was still preserved in some southern dialects. The above factors are largely responsible for the great variations and diversity of forms in expressing a given grammatical category across dialects. According to Cao (2008: 95), for instance, there are at least sixty-nine different markers for passive structures. In addition to the grammaticalization process, there are many other channels for a dialect (including Mandarin Chinese) to acquire a new grammatical device, which makes the picture extremely complex.

13.7 Future Tenses Both aspect and tense belong to the major grammatical categories of verbs across languages, but the Chinese language developed aspect suffixes only, without acquiring tense markings during the development process. Why was Chinese unable to develop tense markers? Now we are in a position to answer this question. Tense and aspect are

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Future Tenses

361

closely related and often expressed by the same devices in some languages. For example, in English, tense and aspect markers all take the form of verb suffixes, e.g. the past tense -ed, the past participle -ed and the progressive aspect -ing. Thus, why Chinese did not acquire verb suffixes for these two grammatical categories at the time is a legitimate question. The answer lies in the historical contexts in which the aspect suffixes were created. The development of the resultative construction directly motivated the emergence of the aspect suffixes. From another angle, it can be said that the aspect suffixes virtually belonged to a subtype of resultative element. Under the influence of the resultative construction, the principle of action–resultative ordering, which governed the information structure of the predicate, came into effect, requiring that only resultative constituents could remain in postverbal position and other non-resultative elements had to occur in preverbal position. By definition, aspect markers refer to the state of the action, whether in progression or in completion, and since all these features can be seen only after conducting the action, they belong to a kind of resultative of the verb. Therefore the related resultative words could occur in postverbal position, where they had the potential to develop into verb aspect suffixes. In contrast, however, tense markers mark the time when the action takes place relative to the time of speaking or other deictic center, which is independent of the action. Thus they are non-resultatives of the verb and can appear only in preverbal position according to the principle of information structure, where they might stand a chance of developing a prefix but not a suffix. In other words, the lexical items for developing aspect suffixes correspond with the words of either the number of repetitions or the duration of an action. As discussed in Section 7.9, words of this type could occur only in postverbal position. Meanwhile, those words that specify the time of an action, including those that express tense information, were limited to preverbal position, which blocked their development into verb suffixes. In Chinese, the information about the time of an action is expressed by time nouns (e.g. zuótiān “yesterday,” qùnián “last year”) and temporal adverbs (e.g. céng-jīng “ever,” cóng-qián “before”), all of which are limited to preverbal position, but verb suffixes must be grammaticalized in postverbal position. This is the real reason why the Chinese language could not acquire tense markings during the development of the resultative construction. Nevertheless, the Chinese language does have certain grammatical devices for expressing tenses, though they are not verb suffixes. Now, we briefly discuss the linguistic forms expressing futurity, which has been represented by two means throughout history: adverbs and auxiliary verbs. The former could freely occur either before or between the subject and the predicate, but the latter had to appear immediately before the predicate. Typological investigation reveals that there are twelve lexical items that tend to develop into future markers across languages (for details, see Heine and Kuteva 2002: 331), and four of them have been attested in Chinese. Moreover, Contemporary Chinese

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

362

has contributed a new type of lexical item for developing a future marker, huí-tóu “come back,” which expresses an immediate future in the Pekingese dialect. (a) Verbs denoting the movements of legs. The verbs in this group are the most common lexical sources for developing future-tense markers. Among a sample of more than 500 languages investigated by Bybee et al. (1994) and Heine and Kuteva (2002), there are fifty-seven languages whose future markers originated from this group of verbs, e.g. be going to in English, referring to the near future. This grammaticalization pathway is based on a common pattern of cognition, mapping a spatial distance onto a temporal distance (Langacker 1991: 219–220). There were always quite a few verbs of this type that developed into future markers in the history of the Chinese language, as attested in the earliest texts. For example, the verb xíng “go” in Shi Jing was already used as a future marker; (82) illustrates its verb xíng “go” and (83) its use as an auxiliary for expressing futurity: (82)




我獨南行。 (詩經 擊鼓) Wǒ dú nán xíng. I alone south go “I am going south alone.” (Shi Jing, Ji Gu, 1000–600 BC)

(83)




行與子還兮。 (詩經 十畝之間) Xíng yǔ zǐ huán xī. will with you return PRT “(I) will return with you.” (Shi Jing, Shi Mu Zhi Jian, 1000–600 BC)

Around the fifth century BC, another verb, jiāng, which was synonymous with xíng, also developed as a future marker. In Heine and Ketuva (2002: 331), it is classified as a verb meaning “take” which developed into a future marker, but this classification is problematic. The verb jiāng was polysemous at the time, meaning “go,” “go along with” or “take,” and it referred mostly to the movements of legs, as illustrated in (84). It was also a common future marker in Old Chinese, as illustrated in (85): (84)




之子於歸, 遠於將之。 (詩經 燕燕) Zhī zǐ yú guī, yuǎn yú jiāng zhī. this girl PRT marry far from go-with she “When this girl gets married, (I) will go with her for a long distance.” (Shi Jing, Yan Yan, 1000–600 BC)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Future Tenses (85)

363




諾, 吾將仕矣。 (論語 陽貨) Nuò, wú jiāng shì yǐ. Yes I will be-an-official PRT “Yes, I will be an official.” (Lun Yu, Yang Huo, 500 BC)

The future marker jiāng has survived to the present day, forming a compound word with the verb lái “come,” jiāng-lái “future,” which is the canonical marker for the distant future in Contemporary Chinese. However, the future marker xíng fell out of use after Old Chinese, although it was still sporadically seen in the compound xíng-jiāng in Medieval Chinese. The verbs qù “go” and lái “come” can also refer to the immediate future in certain contexts, but they have not developed into a stable device for the expression of the future tense. Interestingly, the verb compound huí-tóu (lit. “return head”) “come back” has recently developed into a future marker in standard Mandarin, especially in Pekingese, in which it is often reduced to huí-er by omitting the noun and coloring the preceding rhyme with a retroflex feature, as illustrated in (86). (86)

我回頭給你送去。 (北京話) Wǒ huí-tóu gěi nǐ song-qù I be-going-to for you send-go “I am going to send it to you.” (Pekingese)

(b)

Verbs of desire and intention. According to Bybee et al. (1994) and Heine and Kuteva (2002: 331), the second most common lexical sources for future markers are verbs meaning “want” or “desire”; for instance, the auxiliaries will and shall in English belong to this type. Obviously, future markers in a language may originate from different types of lexical item. The verb yù “desire” or “want” became a future marker around the first century AD, and that usage lasted from then to the tenth century AD. About the same time, the verb yào, with the same meaning as yù, developed into a future marker, as illustrated below:

(87)

雞鳴外慾曙。 (古詩為焦仲卿妻作) Jī míng wài yù shǔ. rooster call outside will dawn “It is going to be dawn when roosters call outside.” (Gu Shi Wei Jiao Zhong Qing Qi Zuo, AD 200)

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

364 (88)




人生要死。 (漢書 武五子傳) Rén shēng yào sǐ. people life will die “Everyone will die.” (Han Shu, Wu Wu Zi Zhuan, AD 100)

The future markers, which resulted from different lexical sources, accumulated and entered the language at different times. There were three kinds of development fate: some died out, such as the future markers xíng in Old Chinese and yù in Medieval Chinese; some are still preserved in compounds such as jiāng in jiāng-lái. Those that continue to be used tend to manifest a division of labor among them. For example, the form be going to in English expresses immediate futurity, which distinguishes it from will/shall, which refer to intermediate and distant futures. Likewise, yào, huí-tóu, and jiāng-lái form a three-way division of labor in expressing futurity: yào expresses an immediate future, huí-tóu means an intermediate future, and jiāng-lái refers to a distant future. Thus yào can be modified by adverbs such as lìkè “at once” or mǎshàng “immediately,” but the other two cannot; the future yào in (87) is not interchangeable with the other two future tenses: (89)

車馬上要開了。 (現代漢語) Chē mǎshàng yào kāi-le. car immediately about-to leave-PERF “The car is about to leave.” (Contemporary Chinese)

The finer divisions among future markers in a synchronic system may reflect the increasing complexity of modern life, with too many things that need to be strictly scheduled. The types of grammatical category are to a certain extent influenced by the social activities of the language community. For example, the rich honorific forms in Japanese reflect the life of Japanese people.

13.8 Auxiliary Verbs of Aspect Finally, we should mention how the Chinese language expressed the meanings of aspect before the emergence of the aspect suffixes. Old Chinese primarily used axillary verbs to express the category of aspect. In Old Chinese, there were three methods to express the function of the perfective aspect. The first method was the possession verb yǒu “have” plus the matrix verb, a very common pattern across languages that is found in hundreds of languages (for details, see Bybee et al. 1994, Heine and Kuteva 2002: 334), as illustrated in (90) and (91):

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Auxiliary Verbs of Aspect (90)

365




女子有行。 (詩經 泉水) Nǚzǐ yǒu xíng. girl have marry “The girl has been married.” (Shi Jing, Quan Shui, 1000–600 BC)

(91)




子路有聞。 (論語 公冶長) Zi Lù yǒu wén. Zi Lu have hear “Zi Lu has heard (of some request).” (Lun Yu, Gong Ye Chang, 500 BC)

In standard Mandarin, this pattern existed from the eleventh century BC to the fourth century AD. Today, in many southeastern dialects, the above pattern is still the predominant form for expressing the perfective aspect. Because of the lack of historical documents, it is unclear whether these dialects preserved the old patterns or came up with them through independent grammaticalization. The apparent cause for the disappearance of this pattern was the rise of the resultative “VOR” pattern around the fourth century AD, where the resultative elements might be certain intransitive elements to denote the completion of the verb with a function similar to that of the perfective aspect. As discussed above, the intransitive resultative liǎo “complete” developed aspect suffix uses at a later stage. In Old Chinese, the verb jì “finish” also became an auxiliary to express the meaning of the perfective aspect, as illustrated in (92) and (93):




(92)

春服既成。 (論語 先進) Chūn fú jì chéng. Spring clothes finish make “The spring clothes have already been made.” (Lun Yu, Xian Jin, 500 BC)

(93)

予既烹而食之。 (孟子 萬章上) Yú jì pēng ér shí zhī. I finish cook and eat it “I have cooked and eaten it.”




(Meng Zi, Wan Zhang Shang, 300 BC) There was also a sentence-final particle, yǐ “already,” that was often combined with the adverb yǐ to denote a change of state, as illustrated in (94) and (95):

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Aspect and Tense

366 (94)




道之不行, 已知之矣。 (論語 微子) Dào zhī bù xíng, yǐ zhī zhī yǐ. theory GEN not implement already know it PERF “I have already known that my theory failed.” (Lun Yu, 500 BC)

(95)




今乘輿已駕矣。 (孟子 梁惠王) Jīn shèng yú yǐ jià yǐ. now take carriage already leave PERF “Now your carriage is already prepared.” (Meng Zi, Liang Hui Wang, 300 BC)

Regarding the experiential aspect, in Old Chinese there was an auxiliary cháng, which grammaticalized from the verb “taste,” that expressed an action in the past, as illustrated in (96) and (97) (for details, see Wang and Shi 2019). This grammaticalization path is also found in other languages, such as Wolof (an African language) and Ancient Tibetan (DeLancey 2018, Voisin and Vittrant 2012, Kuteva et al. 2019: 365).




(96)

吾嘗學此矣。 (左傳 昭公十二年) Wú cháng xué cǐ yǐ. I EXPE learn this PRT “I learned it once before.” (Zuo Zhuan, Zhao Gong Shi Er Nian, 550–400 BC)

(97)

昔者吾友嘗從事於斯矣。 (論語 衛靈公) Xīzhě wú yǒu cháng cóngshì yú sī yǐ. before my friend EXPE engage in this PRT “In the past my friend achieved (these five) things.” (Lun Yu, Wei Ling Gong, 500 BC)




After the establishment of the new set of aspect suffixes, the old forms with similar functions, either auxiliaries or adverbs, were not simply abandoned, and some are still preserved in compound adverbs, such as cháng in céng-jīng “once” in Contemporary Chinese (due to phonological reduction, cháng became céng; for a detailed discussion, see Wang and Shi 2019). Regarding the progressive aspect, Old Chinese seemingly lacked a grammatical device to express the function. The preposition yú “in” was extended to take on the function, but it was attested only in the Shi Jing, a text composed around the tenth century BC, as illustrated in (98). The adverb zhèng “just” or “right,” which originated from an adjective meaning “upright,” could indicate an action in progression, as illustrated in (99):

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

Auxiliary Verbs of Aspect (98)

367




燕燕於飛。 (詩經 燕燕) Yàn-yàn yú fēi. swallow PROG fly “Swallows are flying.” (Shi Jing, Yan Yan, 1000–600 BC)

(99)




庾公正料此事。 (世說新語 政事) Yǔ Gōng zhèng liào cǐ shì. Yu Gong PROG think this matter “Yu Gong is thinking of this matter.” (Shi Shuo Xin Yu, Zheng Shi, AD 450)

In Contemporary Chinese, these adverbs often co-occur with the aspect suffixes to express certain meanings, forming the following three patterns: (100)

(a) Perfective aspect: yǐ-jīng “already” + V-le (b) Experiential aspect: céng-jīng “once” + V-guo (c) Progressive: zhèng-zài “in the process of” + V-zhe

The above adverbs in preverbal position reflect the layer before the first century BC, and the suffixes in postverbal position represent developments since the tenth century AD. Old Chinese employed auxiliary verbs or adverbs, both preceding the matrix verb, to express the functions of the three aspects, because there was as yet no resultative construction that provided the specific context for the emergence of the aspect suffixes. The Chinese language as an SVO language allows full verbs in the first verb position to become functional words, the path of development from a full verb to an auxiliary or adverb. This is the reason why VO languages always have the constituent order “AUX + V.” The concrete forms to take the aspect function might vary greatly from time to time, but the three types of aspect remained unchanged over time and have always been manifested throughout the history of Chinese: perfective, experiential, and progressive, though the word class changed from auxiliary to verb suffix.

https://doi.org/10.1017/9781108921831.013 Published online by Cambridge University Press

14 Negation

14.1 Introduction Negative systems have historically undergone several major changes that were caused by the overall properties of the grammatical system in particular periods. The functions of negatives in any period had to be consonant with the predicate structure and, especially, the structure of verbal predicates. This chapter addresses the developments of negatives that are most relevant to the establishment of the negative system of Contemporary Chinese rather than exhaustively surveying all negatives in history. The period from Late Middle to Early Modern Chinese saw three major changes in relation to the structure of the predicate: the establishment of the resultative construction, the formation of the aspect suffixes, and the emergence of verb classifiers. These changes together made the structure of the predicate much more complex than before, and consequently the predicate was generally grammatically bound by certain grammatical forms (for a fuller discussion, see Chapter 15). Specifically, the predicate verb could no longer be simple and followed mostly by some kind of resultative; thus what we call the predicate became bounded in senses such as time, location and quantity. In turn, this fundamental change of the predicate structure enabled the verb méi (the antonym of the verb yǒu “have”) to develop into a new negative marker, another major event in Modern Chinese with a profound effect on the texture of Chinese grammar. In what follows, we discuss the formation of the negative system of Contemporary Chinese, the typological mixture of the negation and affirmation of the perfective aspect in Chinese, the characteristics of Old Chinese, and the motivation and mechanism for the grammaticalization of the negative méi.

14.2 Affirmation and Negation of the Perfective Aspect It is quite unusual for a typological discord to actually exist between the affirmation and negation of the expressions of the perfective aspect in Contemporary Chinese. According to Bybee et al. (1994: 51‒55), the first two most common lexical sources

368

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Affirmation and Negation of the Perfective Aspect

369

for developing perfective aspect marking across languages are the verbs meaning “have” and “finish.” In Contemporary Chinese, the affirmative form of the perfective aspect is “V-le,” where the suffix -le evolved out of a verb of finishing, liǎo, in Medieval Chinese. However, the negative form of the perfective aspect is “méi-V,” where méi developed out of a verb which is the antonym of “have.” The affirmative and negative forms of the perfective aspect are formulated as follows: (1)

(a) affirmative form of perfective aspect marking: V-le + NP, (b) negative form of perfective aspect marking: méi-V + NP.

By “typological discord,” we mean that the affirmative -le represents one of the two major types of lexical source for developing the grammatical category of the perfective aspect, a verb for “finish.” In addition to Chinese, other languages that underwent the same development are Lhara (Lord 1989: 369), Burmese (Park 1992: 16), and Kongo (Heine and Reh 1984: 88). However, the negative méi belongs to the other major lexical source for the grammaticalization processes, the antonym of “have.” There are many other languages with perfective aspect markings that originated from the verb “have,” including English, German, and many Chinese dialects; as discussed below, Old Chinese had the same grammaticalization. In Contemporary Chinese, the position and status of these two perfective aspect markers contrast with each other: -le is a suffix following the verb, and méi- is an auxiliary preceding the verb. Note that these affirmative and negative forms cannot co-occur within a predicate, as illustrated in (2): (2)

(a) 他已經喫了飯。 (現代漢語) Tā yǐjīng chī-le fàn. he already eat-PERF food “He has already eaten.” (b) 他還沒喫飯。 Tā hái méi chī fàn. he still non-have eat food “He still has not eaten.” (c) *他還沒喫了飯。 *Tā hái méi chī-le fàn. he still non-have eat-PERF food

The affirmative and the negative forms of the perfective aspect marking were introduced into the language at quite different times: the affirmative form entered the language around the tenth century AD, and the negative form came into existence around the thirteenth century AD, with a temporal discrepancy of three centuries.

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

370

Before the negative méi emerged, the other negative wèi or the compound wèi-céng (lit. “not ever”) was used to negate the perfective aspect. By comparison, the affirmation and negation of the perfective aspect are symmetric in English; e.g. the negation of I have eaten my lunch can be achieved simply by inserting “not” between have and the past participle of the verb: I have not eaten my lunch. Likewise, in many Chinese dialects in southeastern areas, such as the Southern Min dialect, Cantonese, the Hakka dialect, and the Xiang dialect, each side of the pair of antonymous verbs meaning “have” and “not have,” which are both purely simple morphemes, has developed into the affirmative and negative forms of their perfective aspect markers, respectively. In these Chinese dialects, the affirmative and negative forms of the perfective aspect are perfectly in accord in terms of lexical source, grammatical status, and distribution, as illustrated below: (3)

我有收着汝個批。 (閩南話) ŋɔ̃ iu siu-tu iɯ ko phe. I have receive-in you ASSO letter “I have received your letter.” (Southern Min dialect)

(4)

我冇收着汝個批。 (閩南話) ŋɔ̃ mu siu-tu iɯ ko phe. I not-have receive-in you ASSO letter “I have not received your letter.” (Southern Min dialect)

In the above examples, both the affirmative yǒu (pronounced iu in the Southern Min dialect) and the negative mǎo (pronounced mu in the Southern Min dialect) are auxiliary verbs with the same grammatical status and distribution, and both developed out of their former verbal usages. This phenomenon is in sharp contrast to that in standard Mandarin, which reveals that the dialects may have undergone a grammaticalization pathway different from that of standard Mandarin. The development of the perfective aspect in these dialects was regular and straightforward; what must be explained is the case of the perfective aspect in standard Mandarin (i.e. the northern dialect). Undoubtedly, most grammatical morphemes develop out of lexical items. However, different languages usually use different vocabulary items within a given conceptual domain, mainly because of their language-specific methodologies of conceptualization. In other words, the lexical system of a language is the result of how a language community conceptualizes the world, which varies greatly from language to language. For the concept of “possess,” Chinese coined two independent words to encode the affirmation and negation of possession; both are purely single morphemes and neither is

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

The Establishment of the System for Negative Markers

371

Table 14.1 The possession verbs of Old and Contemporary Chinese

Old Chinese Contemporary Chinese

Affirmation

Negation

yǒu yǒu

wú méi

derived from the other by adding a negative marker, unlike have and do not have in English. In contrast, there is no real pair of antonyms in English regarding the concept “have,” so its negation must be expressed by a periphrastic form, e.g. adding “not”: do not have. In English dictionaries, lack is listed as the antonym of have, but the former is not exactly the negation of the latter. The word lack refers to the fact or condition of not having enough, shortage, or deficiency, implying an absence or insufficiency of something essential or desired. However, have does not have such a semantic connotation. Since the verbs of possession are highly frequent and active in grammaticalizing into function words, the different lexical items of the concept domain “have” inevitably affect the ways in which the related verbs develop into aspect markers.

14.3 The Establishment of the System for Negative Markers According to Yang and He (2001: 320‒329), at least seventeen negatives always existed in the history of the Chinese language, but only two have survived to the present day, namely bù and méi (méi-yǒu). The former was already widely used in the earliest texts composed around the tenth century BC, such as the Shi Jing, and thus has lasted more than 3,000 years. The latter was introduced into the language after the twelfth century AD. The functions of each negative marker vary greatly from one to another and from time to time, and there have historically always been overlaps, mixtures, and even confusion among these varied negative markers. We find that no clear divisions of labor among these negative markers could be identified for any synchronic systems of any particular period in the past. In Contemporary Chinese, however, the division of labor between bù and méi is clear, as follows (for a greatly detailed analysis, see Shi 2010: 202‒219): (5)

(a) bù: a negative marker for unbounded constituents, (b) méi: a negative marker for bounded constituents.

As has been mentioned previously, resultatives and aspect suffixes function to make the matrix verb bounded. Once the predicate is bounded, the verb can be negated only by méi and not by bù, as illustrated in (6) and (7):

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

372 (6)

我没有 (*不) 打破玻璃杯。 (現代漢語) Wǒ méi (*bù) dǎ-pò bōlí bēi. I not hit-break glass cup “I did not break the glass cup.” (Contemporary Chinese)

(7)

我沒有 (*不) 看過那本書。 (現代漢語) Wǒ méi (*bù) kàn-guò nà-běn shū. I not read-EXPER than-CL book “I have not read the book.” (Contemporary Chinese)

When the verb is not overtly bounded by any linguistic form, the two negative forms, “bù + V” and “méi + V,” warrant two distinctive interpretations, as illustrated in (8): (8)

(a) 老虎不喫草。 (現代漢語) Lǎohǔ bù chī cǎo. tiger not eat grass “Tiger does not eat grass.” (b) ?老虎没喫草。 (現代漢語) ? Lǎohǔ méi chī cǎo. tiger not eat grass “The tiger has not eaten the grass.” (Contemporary Chinese)

(8b) is grammatically correct but semantically incorrect because tiger is carnivorous, but the sentence implies that it has previously eaten grass. This implication stems from the function of méi, which automatically construes the action as a bounded event; in other words, méi can negate only a concrete and individual activity. By comparison, when bù is used to negate the verb, the action is virtually construed to be unbounded, hence yielding a complete negation. The functional distinction between bù and méi reveals the motivation for the establishment of the negative system and especially for the grammaticalization of méi. As pointed out above, the three major changes in Late Medieval Chinese and Early Modern Chinese – the establishment of the resultative construction, the emergence of the aspect suffixes, and the grammaticalization of verb classifiers – all made the predicate bounded. The combined power of these changes triggered the verb méi “the antonym of having” to develop into an auxiliary verb.

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Characteristics of the Old Negative Systems

373

14.4 Characteristics of the Old Negative Systems The systems of negative markers varied from time to time, which actually reflected the overall properties of the grammatical system at the time. Compared to Contemporary Chinese, there were three remarkable characteristics that we must address in order to gain a better understanding of the history of the negative markers: (a) in Old Chinese some negative markers seem to have made the following verbs intransitive; (b) in Old Chinese, in negative constructions, pronouns that were used as the object were moved to before the matrix verb; and (c) the copular construction had an asymmetry between affirmation and negation. These three points are briefly discussed in subsequent sections.

14.4.1 Varying Degrees of Grammaticalization In the history of the Chinese language, negative markers often developed out of content verbs through a process of reanalysis in which the related verbs became negative auxiliary verbs. As we will discuss in depth below, the grammaticalization of the negative méi underwent three stages: first, as a content verb with the meaning “not have,” it could be followed only by a noun object; second, the object was extended to the nominalized form of verbs, where the nominalized verb was construed as a nominal item and thus could not carry an object; and finally, the verb of not-having was reanalyzed as an auxiliary, so the following main verb could have its own object. A similar phenomenon is found in the history of English. That is, the auxiliary “have” plus past participle evolved from the construction in which have was the main verb and its following participle was actually a nominalized form; thus, in the early stages, the nominalized form of the verb participle could not be followed by any object (for details, see Hopper and Traugott 2003: 53). This regularity in auxiliarization can be applied to explain a major puzzle of the negatives in Old Chinese. It has long been known that there were two pairs of negatives in Old Chinese, with a division of labor: one of each pair made the following verb intransitive (Ding 1935, Lü 1984: 73‒102). Based on the reconstructions of Wang (1989: 122), the phonological forms are shown in Table 14.2 (the tone values are omitted). In the above table, the first pair were the most general negative markers in declarative sentences, and the second pair were used mainly in imperative sentences. The negative Table 14.2 The phonological forms of negatives in Old Chinese Transitive negatives

Intransitive negatives

pĭuǝ miua

pĭuǝt mĭuǝt

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

374

pĭuǝ (i.e. bù in Contemporary Chinese) could negate either transitive or intransitive verbs; the negated transitive verbs often took an object. In contrast, when the negative pĭuǝt was used, the verbs could not be followed by an object, regardless of whether they were transitive or intransitive. The contrastive functions between these two negatives can be illustrated by the following examples, where the verb in the first clause that is negated by pĭuǝt does not carry an object but the second one negated by pĭuǝ has an object zhǐ “taste” (Wang 1989: 134): (9)




弗食, 不知其旨也。 (禮記 學記) Fú shí, bù zhī qí zhǐ yě. (note: Fú = pĭuǝt) not eat not know its taste PRT “If you don’t eat, you don’t know the taste.” (Li Ji, Xue Ji, 450 BC)

A parallel usage existed between the two subjunctive negative markers, namely, miua and mĭuǝt, which is perfectly exemplified by the following example consisting of two clauses. According to Wang (1989: 122), in Old Chinese the verbs that were negated by miua had to take an object. (10)




毋友不如己者, 過則勿憚改。 (論語 子罕) Wú yǒu bù rú jǐ zhě, guò zé wù dàn gǎi. (note: wú = miua; wù = mĭuǝt) Not friend not as self who mistake but not fear correct “Don’t make friends with those who are not as good as yourself; don’t fear to correct your mistakes.” (Lun Yu, Zi Han, 500 BC)

In the above example, yǒu is the verb “make friend” and the object is the NP bù rú jǐ zhě “those who are not as good as yourself.” Some researchers have suggested that the above phenomena were due to the fusion of the negative marker and the third pronoun zhī (Lü 1984: 73‒102). However, this proposal is hardly convincing because a fusion was usually caused by two factors: first, the two lexical items had to appear together frequently; second, the fused item was phonologically similar to the original two words and typically took the onset of the first item and the nucleus of the second one. For example, the subjective negative béng in Contemporary Chinese is the fusion of the negative bù and the auxiliary yòng “need,” two words that often co-occur, and the fused form takes the onset of bù and the nucleus of yòng.1 According to the reconstruction of Wang (1980), the third pronoun had the

1

In béng, the vowel is altered because of the rule of syllabification; specifically, the onset [p] cannot form a syllable with the rhyme [oŋ] in Contemporary Chinese.

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Characteristics of the Old Negative Systems

375

phonological form [ȶǐə], which did not contain a final consonant [t], although both of the supposedly fused negative markers had this final consonant. Thus the “fusion” theory cannot hold. From the perspective of grammaticalization theory, however, we can provide a straightforward explanation. We argue that each of the two pairs of negatives actually developed out of the same lexical sources but reflected two usages at different stages of the grammaticalization process. These negative markers originated from ordinary verbs, and in the earlier stages they could be followed only by nouns or verb nominalizations. This feature was preserved in the full phonological forms, namely pĭuǝt and mĭuǝt. In this case, the following verbs were actually construed as a nominal item; hence they could not be followed by any object. As a result, these two negatives seem to have made the following verbs intransitive. As the grammaticalization proceeded, the negatives were reanalyzed as auxiliary verbs, and both also underwent phonological reduction by deleting the final [t]. Only at this stage could the following verb take an object. The written system of Chinese did not directly reflect phonological forms; hence different Chinese characters were usually chosen to record the grammaticalized items due to semantic bleaching and phonological reduction, all of which created a great deal of difficulty in uncovering the truth. This analysis is supported by the development of the negative méi in Modern Chinese, a point to which we will return below. In the earliest texts, both bù and wú were used as verbs that meant “not having” and could be followed by a nominal object, as illustrated in (11) and (12):




(11)

君子於役, 不日不月。 (詩經 君子於役) Jūnzǐ yú yì, bù rì bù yuè. gentleman in servitude not-have day not-have month “My gentleman has been in servitude without a limit of time.” (Shi Jing, Jun Zi Yu Yi, 1000–600 BC)

(12)

勿士行枚。 (詩經 東山) Wù shì xíng méi. Not-have soldier walk stick “(They) don’t have soldiers to walk with a stick.” (Shi Jing, Dong Shan, 1000–600 BC)




When these had been well grammaticalized into auxiliary verbs, the feature of making the verb intransitive disappeared in Late Old Chinese.

14.4.2 Inversion between Verbs and Pronouns In Old Chinese, negative markers often caused an inversion of the verb and a pronominal object, a phenomenon that has not been found since then. As we saw in

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

376

Chapter 3, there was then a focus position between the subject and the predicate, which was usually occupied either by wh- words (interrogative pronouns) or by focused constituents that were used as the object. Since wh- words were naturally a focus of a sentence, they had to be moved to the focus position to check their inherent focus feature. Nouns appeared in the focus position only when focused. Similarly, negative markers often caused personal pronouns to be moved to the position immediately after the negative and before the predicate, as illustrated in (13) and (14), but this inversion was a strong tendency rather than a strict rule at the time, unlike wh- words, which were regularly moved to the preverbal focus position: (13)




今予惟不爾殺。 (尚書 多士) Jīn yú wéi bù ěr shā. now I only not you kill “Now I only didn’t kill you.” (Shang Shu, Duo Shi, 1100–800 BC)

(14)




僂句不余欺也。 (左傳 昭公二十五年) Lóu Jù bù yú qī yě. Lou Ju not I deceive PRT “Lou Ju didn’t deceive me.” (Zuo Zhuan, Zhao Gong Er Shi Wu Nian, 550–400 BC)

Note that in the affirmative construction the pronoun remained in postverbal position like an ordinary noun, as illustrated in (15): (15)




我勝若, 若不我勝。 (莊子 齊物論) Wǒ shèng ruò, ruò bù wǒ shèng. I win you you not I win “I won over you, you didn’t win over me.” (Zhuang Zi, Qi Wu Lun, 300 BC)

In the above example, the first clause is affirmative, with the object pronoun ruò “you” following the verb, and the second clause is negative, with the object pronoun wǒ “I” appearing before the verb. What motivated the inversion of the verb and the object pronoun in the negative construction? It has been widely accepted that this phenomenon in fact preserved the SOV word order in prehistorical Chinese (for details, see Wang 1989: 198‒216). However, this viewpoint is problematic for three reasons: first, there is no evidence that prehistorical Chinese ever had been an SOV language; second, the fronted pronouns could occur only immediately after the negative markers, a feature that does not exist in a typical SOV language, where the object generally precedes both the negative marker

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Characteristics of the Old Negative Systems

377

and the predicate; and if it truly preserved the prehistorical SOV feature, it would more likely be found in the affirmative construction rather than the negative one because the former usages are supposedly more frequently used, and high frequency is the key factor for preserving old features of grammar. More crucially, we find that the inversion never happened at the level of embedded clauses, as illustrated in (16) and (17):




(16)

不知我者謂我何求? (詩經 黍離) Bù zhī wǒ zhě wèi wǒ hé qiú. not know I who ask me what pursue “Those who didn’t know me asked what I was pursuing.” (Shi Jing, Shu Li, 1000–600 BC)

(17)

以其不從己而敗楚師也。 (左傳 成公十七年) Yǐ qí bù cóng jǐ ér bài Chǔ shī yě. for his not listen-to him and defeat Chu troop PRT “For his not listening to him and defeating the Chu troops.” (Zuo Zhuan, Cheng Gong Shi Qi Nian, 550–400 BC)




In (16), bù zhī wǒ “not know me” is a relative clause to modify the pronominal head zhě, where the negation did not trigger the inversion of the verb and the pronominal object. In (17), qí bù cóng jǐ is a nominalized clause, with the pronoun jǐ “himself” following the verb. It is cross-linguistically true that the embedded clause is more likely to preserve the older word order of the language. Therefore we believe that the above phenomenon in Old Chinese was a pragmatic apparatus rather than a relic of prehistoric Chinese. In other words, it made use of the device of changing constituent order to achieve certain discourse functions, either highlighting or emphasizing the pronominal object. This pragmatic method was consistent with other features of the grammar at the time, such as the syntactic behaviors of wh- words and focused constituents in Old Chinese (for details, see Chapter 3). If the inversion is regarded as a grammatical device, we can easily understand the change whereby all three types of constituent order shift, which involved wh- words, focused constituents, and object pronouns in negative construction, fell out of use around the same period, roughly the Late Old Chinese period. Once again, the characteristics of the negative system were always to a great extent determined by the overall property of the grammar at the time.

14.4.3 The Negative Copular Verb in Old Chinese As we saw in Section 2.2, the copular construction in Old Chinese was obligatorily marked by the sentence-final particle yě and optionally by the anaphoric zhě that was used immediately after the subject. There was a pause between the subject and the

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

378

complement. The grammar of Old Chinese did not have a copular verb that occurred between the subject and the nominal complement like ordinary verbs. Since the particle yě could not be preceded by any negative marker, the negative copular verb fēi was invented to negate a copular construction and behaved like a negative copular verb to link the subject with the nominal complement. This created a structural asymmetry between the affirmative and negative copular constructions at the time, as schematized in (18): (18)

(a) The affirmative copular construction: Subj (zhě), NP yě. (b) The negative copular construction: Subj fēi NP.

The negative copular construction is illustrated in (19) and (20) (for the affirmative construction, see Chapter 2 for details): (19)




子非魚。 (莊子 秋水) Zi fēi yú. you not-be fish “You are not a fish.” (Zhuang Zi, Qiu Shui, 300 BC)

(20)




白馬非馬。 (公孫龍子 孔丛子) Bái mǎ fēi mǎ. white horse not-be horse “The white horse is not a horse.” (Gong Sun Long Zi, Kong Cong Zi, 250 BC)

In Old Chinese, there were three characteristics of the negative copular construction that distinguished it from the corresponding affirmative form: first, the anaphoric zhě could not be used in the negative copular construction; second, in the negative copular construction, there was no pause between the subject and the complement; and third, the sentence-final particle yě did not need to be used in the negative construction. Note that the copula fēi was a verb-like single morpheme that formed an SVO structure by itself. When the copula shì developed out of its original demonstrative use, an event that occurred around the first century BC, a negative copular expression could be made simply by adding the general negative marker bù to the copular verb. After that, the negative copula fēi ceased to be used. It is mostly true across languages that a negative expression can be achieved simply by adding a negative marker to the affirmative instance (e.g. I am a student versus I am not a student; I have eaten my lunch versus I have not eaten my lunch), but the historical facts were not always as straightforward as in the above case and usually resulted from complex and interacting factors. Another remarkable asymmetry between the affirmative and negative forms is found in the

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Characteristics of the Old Negative Systems

379

perfective aspect of Contemporary Chinese, namely the affirmative suffix -le and the negative auxiliary méi, as discussed above.

14.4.4 The Negative Perfective Aspect Méi As mentioned previously, a common grammaticalization pathway across languages is the evolution of the verb of “have” into a perfective aspect, as in English, German, Old Chinese, and many southeastern Chinese dialects. However, this common pathway is affected by how the concept “have” is lexically encoded in a particular language. In this respect, languages fall into two types: one is like English, which has coined a single verb “have” to encode it and expresses the opposite by the analytical form do not have, and the other is like Chinese, which coins two unanalyzable single-morpheme words to encode the affirmative and negative sides of the concept of possessing. For example, the pairs of antonymous words are yǒu and wú in Old Chinese and yǒu and méi in Contemporary Chinese. This fact is critical in understanding the asymmetric developments of the two verbs of the antonymous pair toward the perfective aspect. The situation in English is quite simple, with the affirmative side “have” first developing into a perfective marker and the corresponding negative expression of the aspect being made simply by adding the negative do not to it (Hopper and Traugott 2003: 57). It is unlikely that, in English, the negative analyzable form, either “not have” or “have not,” first developed into an aspect marker and then the affirmative expression was created by deleting the negative marker, and that the negative analyzable form went in a different direction than its affirmative form. Due to the particular conceptualization in Chinese, however, each of the two members, the affirmative yǒu and the negative wú or méi, might have undergone its own development. For instance, the affirmative yǒu “have” in Old Chinese grammaticalized into a perfect marker, but the negative wú did not have the same development. By comparison, the negative wú “not-have” developed into a general negative marker at the time, which did not happen to the affirmative yǒu “have.” However, in many Chinese dialects, the two could undergo a parallel development; for example, in the Southern Min dialect, yǒu “have” and mǎo “not have,” two verbs consisting of a single morpheme, function to mark the affirmative and negative expressions of the perfective aspect marker respectively (for details, see Section 13.6). Now we turn to the development of the negative perfective aspect. The emergence of the negative méi (antonymous to yǒu “have”) and its further grammaticalization were one of the significant changes in the history of Modern Chinese. Its development underwent three stages, which are described below.

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

380

Stage 1. The original meaning of méi is “sink into water” or “disappear,” a usage that already existed in Old Chinese, as illustrated in (21). After the sixth century AD, it was extended to refer to the concept of not-having and was antonymous to yǒu “have,” as illustrated in (22):




(21)

夢為魚而沒於淵。 (淮南子 俶真訓) Mèng wéi yú ér mò yú yuān. dream become fish and submerge in lake “I dreamed that I became a fish and sank in a lake.” (Huai Nan Zi, Chu Zhen Xun, AD 150)

(22)

深山窮谷沒人來。 (劉商 送王旯樗罩荨妨跎) Shēn-shān qióng-gǔ méi2 rén lái. deep-mountain remote-valley not-have people come “Nobody came to the deep mountain and remote valley.” (Liu Shang, Song Wang La Chu Zhao Xun Fang Tuo, AD 800)




That is, méi replaced wú to express the negation of possessing, probably because wú had been highly grammaticalized into a negative marker at the time. This type of lexical replacement that resulted from grammaticalization often happened in the history of Chinese; for example, after the verb bǎ “hold” fully grammaticalized into a disposal marker around the tenth century AD, the verb ná “hold” was coined to refer to the concept “take/hold” (for details, see Section 9.9). In the early stage, méi was still a content verb and could be followed only by an object noun. Stage 2. Méi was first extended to negate the verb yǒu “have,” forming a compound-like verb méi-yǒu because of their frequent co-occurrences, as illustrated in (23). Then, either méi or méi-yǒu was further extended to negate a verb-nominalized form, as illustrated in (24), and the nominalized verb could not be followed by any object at that time. (23)

如今為沒有賣的。 (老乞大諺解) Rújīn wéi méi-yǒu màide. now be not-have goods “Now they don’t have the goods.” (Lao Qi Da Yan Jie, AD 1300)

(24)

2

莊家田禾沒有收成。 (老乞大諺解) Zhuāngjiā tiánhé méi-yǒu shōuchéng. crop seedling not-have harvest “The crops and seedlings have no harvest.” (Lao Qi Da Yan Jie, AD 1300)

When extended to express the negation of “have,” the pronunciation also changed.

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Characteristics of the Old Negative Systems

381

It was in the above context in (24) that the verb méi was reanalyzed as an auxiliary, as it preceded a nominalized item with the meaning of the following verb. Stage 3. Méi developed into an auxiliary; hence the following verb could be suffixed by aspect markers or followed by an object noun, as illustrated in (25) and (26): (25)

只說三日没喫飯。 (水滸傳六回) Zhǐ shuō sān rì méi chī fàn. only say three day non-have eat food “I only said that I didn’t eat for three days.” (Shui Hu Zhuan, Chapter 6, AD 1400)

(26)

如今方下種, 還沒有發芽哩。 (明話本 蔣興哥重會珍珠衫) Rújīn fāng xià zhǒng, hái méi-yǒu fā yá lī. now just plant seed still not-have sprout burgeon PRT “Now the seeds were just planted but the burgeons had not grown out.” (Ming Hua Ben, Jiang Xing Ge Chong Hui Zhen Zhu Shan, AD 1650)




As indicated previously, the function of this newly innovated negative marker méi was largely influenced by the overall properties of the grammar at the time, in particular the structure of the predicate. Then, the predicate structure was generally bounded grammatically due to three major changes: (a) the establishment of the resultative construction, (b) the emergence of the aspect markers, and (c) the introduction of verb classifiers. All of these changes functioned to make the predicate bounded by overt grammatical forms; furthermore, the high frequency of this type of boundedness resulted in a strong tendency for a grammatically correct sentence generally to require a bounded item to make it sound complete (for a fuller discussion, see Chapter 15). This fundamental change in the predicate structure motivated the emergence of a new negative marker and, more importantly, created the division of labor between the major negative markers in Medieval Chinese. Therefore the negative méi was specialized to negate a bounded predicate, and the negative bù performed the rest of the task of negation (cf. Section 14.3), as illustrated below. (27)

VR phrases can be negated only by méi: 他没有 (*不) 喫飽飯。 (現代漢語) Tā méiyǒu (*bù) chī-bǎo fàn. he not-have eat-full food “He hasn’t eaten enough.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

382 (28)

“V + guo (experiential aspect)” can be negated only by méi: 我沒有(*不)看過那部電影。 (現代漢語) Wǒ méi-yǒu (*bù) kàn-guò nà-bù diànyǐng. I not-have see-EXPR that-CL movie “I didn’t see the movie.” (Contemporary Chinese)

(29)

“V + quantifier” can be negated only by méi: 湯姆沒有(*不)參加一次。 Tāngmǔ méiyǒu (*bù) cānjiā yī-cì. Tom not-have participate one-CL “Tom never participated in it.” (Contemporary Chinese)

14.4.5 The Structural Discord of the Two Perfective Aspect Markers At the beginning of this chapter, we pointed out that there is a structural discord in terms of the markers for the affirmative and negative perfective aspects. In Contemporary Chinese, the affirmative construction of the perfective aspect is “V + le” and the negative is “méi + V,” where -le and méi cannot be used together. Additionally, they were introduced into the language at different times: -le appeared around the tenth century AD, and méi as the auxiliary of the perfective aspect did not appear until the fifteenth century AD. That is, the affirmative form came into existence approximately 500 years earlier than the negative form. The methods of negating the perfective aspect underwent a series of changes along the grammaticalization pathway. At the early stage when liǎo (the full phonological form of the aspect -le) was used as an independent constituent and thus became separable from the matrix verb, the negative wèi was inserted between them to make a negation, as illustrated below:




(30)

嘆之未了。 (敦煌變文 廬山遠公話) Tàn zhī wèi liǎo. sigh it not finish “(He) had finished saying it.” (Dun Huang Bian Wen, Lu Shan Yuan Gong Hua, AD 800–1000)

(31)

言猶未了。 (唐佛語錄 慧照禪師語錄) Yán yóu wèi liǎo. speak still not finish “(He) hadn’t finished speaking.” (Tang Fo Yu Lu, Hui Zhao Chan Shi Yu Lu, AD 800)




https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

New Developments of Perfective Aspect Marking

383

Due to the fusion of the verb and the resultative -liǎo, no intervening materials were allowed to occur between them. As a consequence, the object could occur only after the whole “V-le” phrase, while the negative markers and other adverbials were fronted to before the “V-le” form. Before the emergence of the negative perfect méi, the negative form of the perfective aspect could be either bù-céng or wèi-céng, where bù and wèi were negative markers and céng was an adverb referring to the experiential aspect that dated back to Old Chinese. Later, the function of negating the perfective aspect was undertaken by the negative méi.

14.5 New Developments of Perfective Aspect Marking As we have emphasized repeatedly, it is the totality of the structural properties of the grammar at a particular period that determines a grammaticalization process, including which lexical sources are chosen, what specific contexts trigger the change, what the life spans of the grammatical morphemes are, and what functions they possess. No grammatical change can happen in isolation, immune to the effects of the grammatical system. The diachronic evidence shows that neither pragmatic inferences nor the language acquisition of children ever plays a clear role in the evolution of grammar. In the grammatical system of Old Chinese, auxiliary verbs were preferred for undertaking many important grammatical functions, such as the passive and aspect markers. Then the experiential aspect cháng, which grammaticalized from the verb “taste,” behaved like an auxiliary, preceding the matrix verb (for details, see Section 13.8). Within this grammatical system, the verb yǒu “have” developed a perfective aspect usage (cf. Section 13.8), as illustrated below:




(32)

女子善懷, 亦各有行。 (詩經 載馳) Nǚzǐ shàn-huái, yì gè yǒu xíng. girl kind also everyone have marry “The girls are kind and each of them has got married.” (Shi Jing, Zai Chi, 1000–600 BC)

(33)

有隕自天。 (周易 姤) Yǒu yǔn zì tiān. have fall from sky “It has fallen from the sky.”




(Zhou Yi, Gou, 800 BC) As the resultative construction developed and a set of verbs in postverbal position grammaticalized into aspect markers, the above form of perfective aspect marking

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

Negation

384

gradually became obsolete (for details, see Section 13.8). Since then, the use of suffixes for marking aspect has been favored by the overall properties of the grammatical system. However, the verb yǒu “have” again started to be used to express the perfective aspect over the past 200 years or so, although it is limited mainly to the written language. First, the verb yǒu can be followed by a nominalized verb form (typically disyllabic compound verbs) to express “some change has been made,” as illustrated in (34). Second, in some cases the verb yǒu is interchangeable with the perfective aspect -le, as illustrated in (35) (Shi 1984): (34)

他們的學習有進步。 (現代漢語) Tāmen de xuéxí yǒu jìnbù. They GEN study have progress “They have made progress in their studies.” (Contemporary Chinese)

(35)

他在那家銀行存有一萬塊錢。 (現代漢語) Tā zài nà-jiā yínháng cún-yǒu yī-qiān-kuài qián. he in that-CL bank deposit-have one-thousand-dollar money “He has deposited 1,000 dollars in the bank.” (Contemporary Chinese)

Even in today’s spoken language, the verb yǒu has started to function to express the perfective aspect, behaving like an auxiliary. It seems that there has been a return to the grammar of Old Chinese. This change was caused by the emergence of the affirmative–negative question structure, which raises a question by combining an auxiliary and its negative form prior to the predicate, as schematized and illustrated below. (36)

Subj AUX-not-AUX VP?: 你會不會開車? (現代漢語) Nǐ huì-bù-huì kāi chē? you can-not-can drive car “Can you drive a car?” (Contemporary Chinese)

According to the investigation of Shao (1996), the above question pattern was introduced into the language approximately 100 years ago. Since this question pattern has been increasingly used, the negative form méi (i.e. auxiliary) also appears in this construction, which produces the question structure “yǒu-méi-yǒu + VP.” A naturally occurring example is shown in (37).

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

New Developments of Perfective Aspect Marking (37)

385




我這話有沒有冒犯你? (馮驥才 崇拜的代價) Wǒ zhè huà yǒu-méi-yǒu màofàn nǐ? I this speech have-not-have hurt you “Does my speech hurt you?” (Feng Ji Cai, Chong Bai De Dai Jia, Contemporary Chinese)

If the answer is negative, the form is “méi-yǒu + VP,” a normal sentence; however, if the answer is positive, the structure should be “V-le NP.” In standard Mandarin, using “yǒu + VP” to express the perfect is still generally considered “not fully acceptable,” as it sounds like a non-standard dialect expression. However, the expression “yǒu + VP” (e.g. yǒu màofàn “have hurt”) is often heard, especially in the speech of the younger generation. In this question-and-answer context, this expression is certainly much more acceptable for many people. This case shows that younger generations are readier to accept new changes because they are less subject to the influence of the early system of the older grammar. However, this is not a matter of language acquisition but still an outcome of the effect of the grammatical system.

https://doi.org/10.1017/9781108921831.014 Published online by Cambridge University Press

15 The Boundedness of the Predicate 15.1 Introduction Linguistic forms serve two basic purposes: one is to convey information for the purpose of communication and the other is to meet the requirements of the grammatical system. For instance, in the phrase “three books,” three and book express the necessary information, but the plural morpheme -s here is purely required by English grammar rather than conveying any information because the number three has already specified the quantity of the book, namely more than one copy. In some cases, these two purposes converge; for example, in I bought pencils, the plural morpheme not only indicates “more than one pencil that I bought” but also makes the sentence grammatical. Another example is that English grammar requires that every clause with a finite verb needs a subject in order to be well formed, so even when the subject has been unambiguously identified in the context it still must be overtly spelled out. Otherwise, the sentence will be ungrammatical. An extreme case is the application of the so-called dummy subject it in it is raining, which neither conveys any information nor refers to anything. Similarly, as we discussed in Chapter 8, in Contemporary Chinese, the passive structures marked by jiào, ràng, and gěi require an agent noun in order to make the passive instances well formed even when the agent cannot be identified or does not need to be overtly indicated. This rule is in sharp contrast with the historical passive structures; for instance, the passive auxiliaries jiàn and bèi in Old Chinese could not introduce any agent noun; if the agent had to be indicated, the preposition yú had to be used to introduce the agent in postverbal position. The feature of the passive structure in Contemporary Chinese is a result of the operation of the analogy of the disposal construction, which requires the patient noun in order to make it grammatical. In other words, an agent was introduced into the passive of Contemporary Chinese to satisfy the requirement of the grammar rather than to serve any communicative purpose (for a detailed discussion, please see Chapter 8).

386

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

The Boundedness of Verbal Predicates

387

15.2 The Boundedness of Verbal Predicates Boundedness is ubiquitous and has many manifestations in grammar (for details, see Langacker 2013: 63‒65). This means that the referent of a concept has a limit in certain senses. For example, the referents of nouns are bounded in space, and those of verbs are bounded in time. Boundedness can be viewed from two perspectives: either inherent meaning or overt grammatical form. For nouns, countable nouns represent bounded entities in space and can be suffixed with the plural -s in English and modified by the numeral classifier phrases in Chinese. For verbs, the concept “eat” refers to a bounded action and can be suffixed with a perfective aspect in Russian and with an experiential aspect in Chinese. However, we are not concerned here with the inherent boundedness of words and their grammatical manifestations; instead, our focus is on the boundedness of the predicate which primarily satisfies the requirement of the grammatical system rather than serving any communicative purpose. As we discussed in former chapters, from Medieval Chinese to the present day, the Chinese language has acquired many grammatical constructions and markings, including mainly the resultative construction, the aspect system, verb reduplication, and verb classifiers. All of these grammatical apparatuses have something in common: they are suffixed with the matrix verb to constitute an immediate constituent. As a consequence, they all function to make the matrix verb bounded in some sense. Meanwhile, several major constructions, such as the disposal construction and the verb-copying construction, which either prefer or even require a verb–resultative phrase to be used as their predicates, resulted from the establishment of the resultative construction, where the resultative also serves to make the predicate bounded. In many cases, the bounded elements attached to the predicate carry informative value; for instance, the perfective -le expresses an action in the past with current relevance. Due to the analogy of the combinational power of these newly acquired devices of the grammar, the structure of the predicate in Contemporary Chinese is grammatically required to be bounded by some means; otherwise the sentence may be ill-formed. In Contemporary Chinese, specifically, a sentence sounds incomplete unless aspect markers, time words, quantifiers, preposition phrases, or resultatives are added (Lu 1988, Kong 1994, Huang 1994). However, this grammatical feature did not exist before Modern Chinese, as is evident through a simple comparison of instances at different stages. Before the sixth century AD, a simple verb could be freely used as the predicate, but in the corresponding expression in Contemporary Chinese, a resultative or an aspect marker, or even both, must be added to make it grammatical. To guarantee the accuracy and objectiveness of our comparison, we choose Shi Shuo Xin Yu, a vernacular text composed in the fifth century AD, and its translation into Contemporary Chinese by Xu

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

The Boundedness of the Predicate

388

(1995), to examine the contrast between the structures of the predicate, as illustrated in (1) and (2): (1)




(a) 真人東行。 (世說新語 德行) Zhēn-rén dōng xíng. Taoism-man east go “The Taoist man goes to the east.” (Shi Shuo Xin Yu, De Xing, AD 450) (b) 真人往東方去了。 (許嘉璐 1995) Zhēnrén wǎng dōng-fāng qù-le. Toaism-man toward eastern-direction go-PERF “The Taoist man has gone to the east.” (Xu 1995: 1, translation into Contemporary Chinese)

(2)




(a) 皆以為登龍門。 (世說新語 德行) Jiē yǐwéi dēng Long-mén. all think board dragon-gate “Everyone thought that they had scaled the Dragon Gate.” (Shi Shuo Xin Yu, De Xing, AD 450) (b) 都認為登上了龍門。 (許嘉璐1995) Dōu rènwéi dēng-shàng-le Lóngmén. all think scale-above-PERF dragon-gate “They all thought they had boarded the Dragon Gate.” (Xu 1995: 9, translation into Contemporary Chinese)

In the above examples, the predicates of the sentences in the fifth century AD were both simple verbs, but the aspects or resultatives had to be added for the sentences to be well formed in Contemporary Chinese. The two grammatical categories were introduced into the language after the tenth century AD, as discussed previously. In a serial verb construction, which refers to two or more actions that happen sequentially in time, both verbs could be simple without any bounded element in Old and Medieval Chinese, whereas today they must be bounded by some grammatical means (for the emergence of the serial verb construction, see Chapter 4). Otherwise, the construction will be ill-formed in Contemporary Chinese. The contrast can be seen in the following pair of examples: (3)




(a) 登車攬辔。 (世說新語 德行) Dēng chē lǎn pèi. board carriage take bridle “Board the car and take the bridle.” (Shi Shuo Xin Yu, De Xing, AD 450)

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

The Boundedness of Verbal Predicates

389

(b) 坐上車子, 拿過韁繩。 (許嘉璐 1995) Zuò-shàng chēzi, ná-guò jiāngshéng. board-on carriage take-over bridle “Get in the car and take the reins.” (Xu 1995: 9, translation into Contemporary Chinese) In Contemporary Chinese, a “verb + resultative” phrase must be used to express the meanings of the simple verbs in Medieval Chinese, a sign of the general boundedness of the predicate. Within a serial verb construction in Contemporary Chinese, the first verb can no longer be simple and must be bounded by some grammatical means. The methods include an aspect marker, a reduplicated form, a resultative, and even a classifier, as illustrated in the following examples. All grammatical devices to make the predicate bounded were introduced into the language after the sixth century AD. (4)

Aspect suffix: 喫過飯再走。 (現代漢語) Chī-guò fàn zài zǒu. eat-EXPER food then leave “Please eat before you leave.” (Contemporary Chinese)

(5)

Verb reduplication: 喫喫飯再走。 (現代漢語) Chī-chī fàn zài zǒu. eat-eat food then leave “Please eat something before you leave.” (Contemporary Chinese)

(6)

Resultative: 喫飽飯再走。 (現代漢語) Chī-bǎo fàn zài zǒu. Eat-full food then leave “Please eat enough before you leave.” (Contemporary Chinese)

(7)

Classifier: 喫一碗飯再走。 (現代漢語) Chī yī-wǎn fàn zài zǒu. eat one-bowl food then leave “Please eat one bowl of food before you leave.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

The Boundedness of the Predicate

390

Note that in (7), the “Num + CL” phrase is the modifier of the object noun and can also function to make the matrix verb bounded. Without this phrase, the sentence is ungrammatical.

15.3 The Boundedness of the Adjectival Predicate In addition to the verbal predicates discussed above, the boundedness of the predicate also affects the structure of the adjectival predicate, as it requires that the adjective in the predicate be modified by a degree word or other quantifiers, as illustrated in (8). In Contemporary Chinese, a simple adjective cannot usually be used as a predicate unless there are two adjacent clauses to form a comparison of the degrees of the quality in question (Zhu 1982: 104), as illustrated in (9): (8)

(a) 那輛車很貴。 (現代漢語) Nà-liàng chē hěn guì. that-CL car very expensive “That car is very expensive.” (b) ? 那輛車貴。 ? Nà-liàng chē guì. that-CL car expensive (Contemporary Chinese)

(9)

那輛車貴, 這輛車便宜。 (現代漢語) Nà liàng chē guì, zhè-liàng chē piányí. that-CL car expensive this-CL car cheap “That car is expensive, but this car is cheap.” (Contemporary Chinese)

According to our investigation of Contemporary Chinese, more than 90 percent of adjectival predicates have a degree word to modify the adjectival predicate, and the degree word is usually hěn “very.” The meaning of this degree word has been neutralized so that it serves to make the adjectival predicate bounded and grammatical but does not intensify the degree of the quality of the adjective in any sense. This contrast can be clearly seen from a comparison of a text composed in the fifth century AD and the corresponding translations into Contemporary Chinese. Before the sixth century AD, simple adjectives were freely used independently as the predicate, but the predicates are usually bounded by a degree word in the corresponding translations into Contemporary Chinese, as illustrated in (10) and (11):

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

The Boundedness of the Adjectival Predicate (10)

391




(a) 時論以此多之。 (世說新語 德行) Shí lùn yǐ cǐ duō zhī. Public opinion for it praise he “Public opinion at the time praised him for it.” (Shi Shuo Xin Yu, De Xing, AD 450) (b) 當時的輿論就因非常讚美他。(許嘉璐 1995) Dāngshí de yúlùn jiù yīn yǐ fēicháng zànměi tā. that-time GEN opinion then for it extremely praise he “Public opinion at that time praised him very much for this matter.” (Xu 1995, translation into Contemporary Chinese)

(11)




(a) 今吾處之不易。 (世說新語 德行) Jīn wú chǔ zhī bù yì. now I handle it not easy “This is not easy for me to handle.” (Shi Shuo Xin Yu, De Xing, AD 450) (b) 現在我仍然這樣對待是很不容易的。 (許嘉璐 1995) Xiànzài wǒ réngrán zhèyàng duìdài shì hěn bù róngyì de. now I still so handle be very not easy PRT “It is not easy for me to still handle it this way.” (Xu 1995: 9, translation into Contemporary Chinese)

To avoid misunderstanding, one matter needs to be clarified. In Medieval Chinese or earlier, some boundedness devices were available, but they were used only for communicative purposes and not solely to make the sentence well formed. These boundedness elements were independent of the matrix verb, which could be directly modified by adverbials, as illustrated below. The scope and frequency of the boundedness at that time were much smaller and lower than in Contemporary Chinese. (12)




胡之去已遠。 (世說新語 仇隙) Hú Zhī qù yǐ yuǎn. Hu Zhi go already far “Hu Zhi already went far away.” (Shi Shuo Xin Yu, Chou Xi, AD 450)

(13)




帝甚不平, 食未畢。 (世說新語 汰侈) Dì shén bùpíng, shí wèi bì. emperor very angry eat not complete “The emperor didn’t eat enough because he was very angry.” (Shi Shuo Xin Yu, Tai Chi, AD 450)

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

The Boundedness of the Predicate

392

The above phenomenon belongs to the separate resultative construction in Medieval Chinese (for details, see Section 6.4).

15.4 Effects of the Boundedness Tendency The general boundedness of the predicate in Contemporary Chinese, which was an effect of changes from the sixth century AD, had far-reaching consequences for the properties of the grammatical system. In this section, we discuss two major changes resulting from the boundedness tendency of the predicate: first, the division of labor between the two major negative markers, bù and měi (for details, see Section 14.3), and second, the further grammaticalization of the nominalizer de to mark a bounded event (for details, see Section 15.3). As we discussed in Chapter 14, the negative system underwent several major changes over time that were largely influenced by the overall properties of the grammatical system at a particular period. As the structure of the predicate developed toward boundedness, a new division of labor took shape between the two major negative markers, bù and měi: the former negated nonbounded elements, mainly including adjectives and simple verbs, and the latter functioned to negate all bounded elements. The negative bù already existed in the earliest texts composed before the tenth century BC, but the negative měi became grammaticalized only after the twelfth century AD. It can be said that the grammaticalization of měi was triggered and propelled by the tendency toward the boundedness of the predicate. Generally, all bounded predicates, including “verb + resultative” phrases, “verb + aspect marker” phrases, “verb + Num + CL,” and verb reduplications, could be negated only by měi. The situations fall into two types: the corresponding negations of the structures of “verb + perfect marker” and verb reduplication are “měi + V,” but the perfect marker -le and the reduplicated form cannot occur with the negative měi. In other words, when negated by měi, they must be reduced to the base forms by deleting the perfective aspect -le, as illustrated in (14) and (15): (14)

(a) 他喫了午飯。 (現代漢語) Tā chī-le wǔfàn. he eat-PERF lunch “He has eaten lunch.” (b) 他没喫午飯。 (現代漢語) Tā méi chī wǔfàn. he non-have eat lunch “He has not eaten lunch.” (Contemporary Chinese)

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

Effects of the Boundedness Tendency (15)

393

(a) 我昨天晚上看了看電視。 (現代漢語) Wǒ zuótiān wǎnshàng kàn-le-kàn diànshì. I yesterday night see-PERF-see television “I watched television for a while last night.” (b) 我昨天晚上没有看電視。 (現代漢語) Wǒ zuótiān wǎnshàng méiyǒu kàn diànshì. I yesterday evening not-have see television “I didn’t watch television last night.” (Contemporary Chinese)

In contrast, the experiential marker -guo, resultatives, and other quantifiers can be combined with the negative měi, as illustrated in the following examples. (16)

(a) 他去過美國。 (現代漢語) Tā qù-guò Měiguó. he go-EXPER America “He has been to America.” (b) 他没有去過美國。 (現代漢語) Tā méiyǒu qù-guò Měiguó. he not-have go-EXPER America “He has never been to America.” (Contemporary Chinese)

(17)

(a) 他做完了作業。 (現代漢語) Tā zuò-wán-le zuòyè. he do-finish-PERF homework “He has finished his homework.” (b) 他没有做完作業。 (現代漢語) Tā méiyǒu zuò-wán zuòyè. he not-have do-finish homework “He hasn’t finished his homework.” (Contemporary Chinese)

Motivated by the development of the predicate, the structural particle de underwent a further grammaticalization to a grammatical morpheme to make the predicate bounded (for a fuller discussion, see Chapter 15). Due to the emergence of the classifier system, which associated a numeral modifier with its head noun in Late Medieval Chinese, the demonstrative dǐ (the former phonological form of de) developed a grammatical

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

The Boundedness of the Predicate

394

morpheme that associated the rest of the modifiers (i.e. not numbers), which included relative clauses, possessive phrases, adjectival phrases, and many others, with their head nouns. At a later stage, the particle de was even extended to mark a complex adverbial phrase. After the fifteenth century AD, de evolved into an important function to make the predicate bounded. If the predicate is an adjectival reduplication, this sentence-final particle de is generally needed; otherwise, the sentence sounds incomplete or even ungrammatical, as illustrated in (18) and (19): (18)

(a) 她的眼睛大大的。(現代漢語) Tā de yǎnjīng dà-dà de. she GEN eye big-big DE “Her eyes are big.” (b) *她的眼睛大大。 (現代漢語) *Tā de yǎnjīng dà-dà. she GEN eye big-big (Contemporary Chinese)

(19)

(a) 他的襯衫雪白雪白的。 (現代漢語) Tā de chènshān xuěbái xuěbái de. he GEN shirt snow-white snow-white DE “His shirt was snow white.” (b) *他的襯衫雪白雪白。 (現代漢語) *Tā de chènshān xuěbái xuěbái. he GEN shirt snow-white snow-white (Contemporary Chinese)

Since English does not have the grammatical forms corresponding to the Chinese associative particle de, the accurate meanings of the above sentences can hardly be translated into English. The adjectival reduplication conveys desirable connotation to the subject when intensifying the degree of quality. When the predicate is a verb phrase, the primary function of the sentence-final particle de is to make the predicate bounded, which in turn produces two side effects: first, it emphasizes that the action happened in the past or will happen in the future, as illustrated in (20); second, it behaves like a past tense to indicate an action in the past, as illustrated in (21): (20)

(a) 這本書借來的。 (現代漢語) Zhè-běn shū jiè-lái de. this-CL book borrow-come DE “It is a book that was borrowed from others.”

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

Effects of the Boundedness Tendency

395

(b) 媽媽會來的。 (現代漢語) Māmā huì lái de. mother will come DE “Mom will certainly come.” (Contemporary Chinese) (21)

(a) 他騎自行車去。 (現代漢語) Tā qí zìxíngchē qù. he ride bicycle go “He will go there by bicycle.” [Not yet] (b) 他騎自行車去的。 (現代漢語) Tā qí zìxíngchē qù de. he ride bicycle go DE “He went there by bicycle.” (Contemporary Chinese)

In general, the structure of the predicate in Contemporary Chinese is far more complex than in Medieval Chinese and earlier. Additionally, in Contemporary Chinese, there are many more grammatical means before or after the predicate to express various delicate meanings. The general boundedness tendency of the predicate structure comes from the combinatorial power of many major changes since Medieval Chinese, including mainly the establishment of the resultative construction and, especially, the emergence of the aspect suffixes. This newly innovated grammatical feature distinguishes Contemporary Chinese from Middle and Old Chinese.

https://doi.org/10.1017/9781108921831.015 Published online by Cambridge University Press

16 Classifiers

16.1 Introduction Before Medieval Chinese, there was no such grammatical category as “classifier” in the grammar of the Chinese language, and this “classifier” was historically innovated, unlike nouns, verbs, and adjectives, which existed from the beginning. Individual classifiers started to emerge as early as Late Old Chinese, and the number of classifiers steadily increased over time. However, the rigorous grammatical rule that required a proper classifier for every numeral phrase was firmly established around the fourteenth century AD. Importantly, there was a parallel development of nominal and verbal classifiers; that is, verbal classifiers came into existence not long after the emergence of nominal classifiers. In this chapter, we first focus on the change in nominal numeral phrases, whose structures underwent the following changes: (1)

The structural change of the nominal numeral phrase: Num + NP > Num + CL + NP

Contemporary Chinese is a rigid classifier language, which means that a proper classifier must be used to link a numeral modifier to its noun head; otherwise the phrase will be ungrammatical. This is required by grammatical rule, and in most cases the classifier does not contribute any semantic value to the phrase, as illustrated below: (2)

(3)

三本書 sān běn shū; three CL book “three books” 五台電腦 wǔ tái diànnǎo; five CL computer “five computers”

*三書 (現代漢語) *sān shū three book (Contemporary Chinese) *五電腦 (現代漢語) *wǔ diànnǎo five computer (Contemporary Chinese) 396

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

The Emergence of the Classifier System

397

As exemplified below, there was another way around in Old Chinese and most of Medieval Chinese, in which no classifier was needed to connect the numeral modifier with its nominal head. More importantly, the emergence of the classifier system had a profound effect on the structure of noun phrases in general; for instance, it triggered the development of the demonstrative dǐ into a relative clause marker, an associate and genitive morpheme, and a marker with many other pragmatic functions (for details, see Chapter 20). Furthermore, the two most general classifiers, zhī and gè, further developed into demonstratives at the expense of the old set of demonstratives (for a detailed discussion, see Chapters 19 and 20). The development process of classifiers has been described in great detail in Wang (1989: 18‒40), Yang and He (2001: 204‒212), and many other studies, and the focus of this chapter is on the motivation and mechanism for the emergence of classifiers, particularly their effect on the texture of the grammar. According to Lü (1999: 710‒715), more than 200 common classifiers are actively used in everyday communication in Contemporary Chinese. Some are restricted to particular nouns; for instance, pǐ can be used for mǎ “horse” and luòtuó “camel.” Some are extremely general, such as gè, which can be used for countless nouns, including concrete and abstract ones. The collocation with nouns is sometimes quite idiosyncratic, purely owing to conventionality, such as pǐ for mǎ “horse,” tóu for niú “cattle,” and zhǐ for yáng “sheep.” Most classifiers are merely abstract grammatical markers, and their applications are required by the grammar; their uses, the meanings of which may be traced in terms of etymology, make no sense for native speakers. However, certain shape-based classifiers do contribute their meanings to the related numeral phrases; for example, when used to modify miànbāo “bread,” the classifier piàn means “a thin piece,” kuài means “lump” and gēn means “bar.”

16.2 The Emergence of the Classifier System Note that numeral classifiers are different from measure words, though they have a similar distribution. Both can appear between the numeral modifier and the nominal head. Measure words, such as chǐ “foot,” cùn “inch,” jīn “unit of weight,” liǎng “ounce,” dòu “unit of grain” and shēng “liter,”1 already existed in Old Chinese, but there were no numeral classifiers at the time. Before the tenth century AD, Chinese was generally a nonclassifier language, in which numbers directly preceded nouns. According to Wang (1989: 18), in Old Chinese numerals could either precede or follow the head noun, as illustrated in (4) and (5), respectively: 1

Measure words in Old Chinese semantically correspond to those in English but are not exactly the same measures.

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Classifiers

398 (4)




一人有慶。 (尚書 吕刑) Yī rén yǒu qìng. one person have celebration “One person has a celebration.” (Shang Shu, Lü Xing, 1100–800 BC)

(5)




(齊)喪車五百。 (左傳 哀公十五年) (Qí) sàng chē wǔ bǎi. Qi lose chariot five hundred “The Qi country lost five hundred chariots.” (Zuo Zhuan, Ai Gong Shi Wu Nian, 550–400 BC)

In Chapter 21, we demonstrate that the constituent order between the modifier and the head was not fixed at the time, which distinguished it from Contemporary Chinese. For instance, adjectival modifiers could precede or follow the head noun. The above phenomenon reflected the overall feature of the grammar at that time. In general, however, the “modifier + head” order was much more common than “head + modifier.” The latter nominal structure entirely disappeared in the first half of Medieval Chinese. According to Wang (1989: 20), if a numeral was followed by either a classifier or a measure word, these “numeral plus classifier/measure” phrases could follow only the head noun at the time, as illustrated in (6) and (7). This grammatical rule ceased to work after the sixth century AD. (6)




負服矢五十個。 (荀子 議兵) Fù fú-shǐ wǔshí gè. carry arrow fifty CL “(He) carried fifty arrows.” (Xun Zi, Yi Bing, 250 BC)

(7)




漢王賜良金百鎰, 珠二斗。 (史記 留侯世家) Hàn Wáng cì liáng-jīn bǎi yì, zhū èr dǒu. Han King bestow fine-gold hundred YI pearl two DOU “King Han bestowed on him 100 yi of gold and two dou of pearls.” (Shi Ji, Liu Hou Shi Jia, 100 BC)

At this stage, the so-called classifiers were actually ordinary nouns with concrete meanings. For example, the word gè originally referred to the individual bamboos of which arrows were made; thus it was used to count arrows. It was in the position of the measure words described above that ordinary nouns became grammaticalized into classifiers; they moved to the prenominal position at a later stage. That is, it was not the case that a classifier was inserted between the numeral and the head noun, where an

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

399

The Emergence of the Classifier System

ordinary noun became grammaticalized. This means that the grammaticalization of classifiers involved a position change, which we will explain below. As mentioned above, in the first half of Medieval Chinese, it was a paradigmatic pattern that simple numerals preceded the head noun, but “numeral + CL” phrases followed the head noun. This rule can be clearly seen in a single sentence as illustrated in (8) and (9):




(8)

(阮籍)蒸一肥豚, 飲酒二斗。(世說新語 任誕) (Ruǎn Jí) zhēng yī féi-tún, yǐn jiǔ èr dòu. Ruan Ji steam one fat-pig drink wine two DOU “(Ruan Ji) steamed one fat pig and drank two dou of wine.” (Shi Shuo Xin Yu, Ren Dan, AD 450)

(9)

率民養一豬, 雌雞四頭。 (齊民要術 序) Shuài mín yǎng yī zhū, cí-jī sì tóu. lead people raise one pig female-chicken four CL “(He) led the people to raise one pig and four hens.” (Qi Min Yao Shu, Xu, AD 550)




In example (9), the pure number yī “one” precedes the head noun and the “Num + CL” phrase sì tóu “four CL” follows the head noun. To judge when an ordinary noun became grammaticalized into a classifier, a reliable criterion is that it started to appear between the numeral and the head noun. In the beginning, it seems that numbers (particularly yī “one”) became compounded with frequently co-occurring measure words, and once lexicalized into a compound-like unit, the number and the measure word were used as a single constituent to precede the noun head. According to Wang (1989: 18), the earliest instances of the “Num + CL + NP” pattern involved only measure words whose collocations had the highest frequency in daily communication, as illustrated in (10) and (11):




(10)

一箪食, 一瓢飲。 (論語 雍也) Yī dān shí, yī piáo yǐn. one bamboo dish food one dipper drink “(He just had) one small dish to eat and one dipper to drink.” (Lun Yu, Yong Ye, 500 BC)

(11)

猶以一杯水救一車薪之火也。 (孟子 告子)




Yóu yǐ

yī bēi

shuǐ

jiù

yī chē xīn

zhī

like use one glass water extinguish one cart firewood

huǒ yě.

ASSO fire

PRT

“It is like using a glass of water to extinguish the fire of a cart of firewood.”

(Meng Zi, Gao Zi, 300 BC)

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Classifiers

400

In the above examples, the head nouns refer to “food,” “drink,” and “firewood,” which were the most important elements of daily life at the time, and the words for measuring these things were so frequently used that they eventually formed compound words. Note that in the earliest examples of this pattern, the number was exclusively “one,” the basic unit of enumeration with the highest frequency among all numbers. That is, yī dān “one basket,” yī piáo “one water-dipper,” yī bēi “one glass” and yī chē “one cart” were actually compound words due to their frequent collocation. As the evolution of Chinese grammar has repeatedly demonstrated, no grammatical rule could be invented overnight, and the earliest examples of many grammatical structures were actually manifestations of lexicalization with the least degree of productivity rather than grammatical paradigms. When this kind of lexicalization accumulates to a certain point, a new paradigmatic pattern may be innovated via analogy. At the early stage, every ordinary noun had to undergo this process of lexicalization or grammaticalization (for the relation between lexicalization and grammaticalization, see Brinton and Traugott 2005: 62‒88; for a related discussion, see Section 6.5). From the first century BC onward, “Num + CL” phrases began to precede the head noun, as illustrated in (12) and (13):




(12)

烏孫以千匹馬聘漢女。 (史記 大宛列傳) Wū Sūn yǐ qiān pǐ mǎ pìn Hàn nǚ. Wu Sun use thousand CL horse marry Han girl “Wu Sun used one thousand horses to marry the Han girl.” (Shi Ji, Da Wan Lie Zhuan, 100 BC)

(13)

陸地千足羊。 (史記 貨殖列傳) Lùdì qiān-zú yáng. land thousand-CL sheep “On the land were one thousand sheep.” (Shi Ji, Huo Zhi Lie Zhuan, 100 BC)




As seen time and time again, the transition from the old to the new forms always had an intermediate stage in which the old and new forms coexisted, as described in the formula A > A/B > B (Hopper and Traugott 2003: 67). This numeral structure typically underwent three stages: “N + Num + CL” > “N + Num + CL” or “Num + CL + N” > “Num + CL + N,” a process whereby the item following the “Num” changed its status from an ordinary noun to a classifier.

16.3 Motivations for Noun Classifiers As pointed out previously, the frequency of the co-occurrence of the number and the classifier played a key role in lexicalizing them into compound words at the earliest stage,

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Motivations for Noun Classifiers

401

a change that can be viewed as a process of reanalysis whereby the boundary between the numeral and the classifier word was weakened or entirely lost. Nevertheless, lexicalization by itself did not necessarily lead to the innovation of a grammatical structure. The establishment of the classifier system was propelled by the prominent trends of language development during the period, namely the disyllabification tendency (for a related discussion of the resultative construction, see Chapter 5), which is discussed below.

16.3.1 Disyllabification and the reanalysis of the numeral and classifier In Chapter 5, we discussed how the tendency toward disyllabification gained momentum from the beginning of the Medieval Chinese period onward as compensation for the ongoing simplification of the phonological system. Many monosyllabic words in Old Chinese became disyllabic during the period from the first century BC to the fifth century AD.2 Under the influence of this phonological development, a disyllabic unit became a new type of prosodic unit favored by language speakers, and in this situation two monosyllabic words that frequently occurred together could easily be reanalyzed into a single unit by weakening and eliminating the boundary between them. In the process of compounding, individual morphemes might lose their original lexical meanings. Historically, the morpheme in the second syllable position often underwent phonological reduction (typically losing tonal value), similar to what happens in a grammaticalization process (for details, see Hopper and Traugott 2003: 3). In our view, the tendency toward disyllabification to a great extent enhanced the compounding of a monosyllabic number and a monosyllabic noun, eventually giving rise to the emergence of classifiers as a new grammatical category. According to Ohta (1987: 147‒152), Wang (1989: 18‒40), Yang and He (2001: 204‒212), and many others, within the “Num + CL + NP” construction in the early stages, all numbers and classifiers were monosyllabic, which means that each pair formed a disyllabic unit, as illustrated in (14). This means that the numbers were limited from one to ten because the numbers greater than ten were all disyllabic or even longer. If a number was multisyllabic, it was still used in the old pattern in which it followed the noun head, as illustrated in (15): (14)

2




兩個月秋耕。 (齊民要術 雜說) Liǎng gè yuè qiū gēng. two CL month autumn plow “Two months of plowing in autumn.” (Qi Min Yao Shu, Za Shuo, AD 550)

As we see in other chapters, this phonological change was responsible for the emergence of the resultative construction.

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Classifiers

402 (15)




取醋石榴兩三個。 (齊民要術 種竹) Qǔ cù shíliú liǎng sān gè. take vinegar pomegranate two three CL “Take two or three vinegar pomegranates.” (Qi Min Yao Shu, Zhong Zhu, AD 550)

In the above two examples, the same classifier in the same text had different distributions: when the number was monosyllabic, the “Num + CL” phrase preceded the head, and when it was disyllabic, it followed the head, a sign that the disyllabification tendency was at work. Our hypothesis can to a certain extent be proven by Dai’s (1998: 60‒68) observation, which is based on his investigation of twenty minority languages in China with classifier systems that recently came into existence. He found that if the basic numbers and classifiers are both monosyllabic, together forming a disyllabic unit, then the language has a more developed classifier system, as in the Yi, Hani, Pumi, and Qiang languages. In contrast, if the basic numbers are multisyllabic, the classifiers are much rarer, as in the Jingpo and Luoba languages. These facts show that disyllabification also played a role in motivating the emergence of classifiers in these minority languages.

16.3.2 Change in the constituent order of the nominal structure Although the disyllabification tendency might have played a role in motivating the compounding of a monosyllabic numeral and a monosyllabic classifier, the crucial factor for the change from “NP + Num + CL” to “Num + CL + NP” should have been the trend of development of the grammatical system at that time; that is, all nominal modifiers moved from the postnominal to prenominal position. Without this trend, the structure need not have changed even after the emergence of classifiers. As we saw above, in the early stages, the paradigmatic structure was “NP + Num + CL” in Medieval Chinese, a schema that was licensed by the grammatical system at the time. In particular, structurally complex modifiers preferably followed the head noun; for instance, adjectival modifiers and relative clauses often occurred after the head noun, as illustrated in (16) and (17) (for details, see Chapter 21): (16)




《詩》 所謂人之無良者。 (左傳 宣公二年) “Shī” suǒ-wèi rén zhī wú-liáng zhě. Shi Jing so-call people GEN unscrupulous REL “They are the people who were called ‘unscrupulous’ in Shi Jing.” (Zuo Zhuan, Xuan Gong Er Nian, 550–400 BC)

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

The Emergence of Verb Classifiers (17)

403




今魏地已定者數十城。 (史記 高祖本紀) Jīn Wèi dǐ yǐ dìng zhě shù shí chéng. now Wei area already suppress REL several ten city “The Wei areas that have been suppressed are several dozens of cities.” (Han Fei Zi, Nei Chu Shuo, 300 BC)

The above grammatical regularity explains why simple numeral words preceded the head noun and the “Num + CL” phrase followed the head noun at the time. The reason was that the latter was more structurally complex than the former. This usage also reflected a cross-linguistic principle that prefers that heavy modifiers follow the head even when the prevailing order of the language is “modifier + head” (for details, see Hawkins 1983, Dryer 1992). Therefore, there was a systematic change in the nominal structure in which all postnominal modifiers disappeared.

16.4 The Emergence of Verb Classifiers Historically, the morphology of nouns and verbs always developed in a parallel manner that was nonlinear, unlike analogical changes. Not long after noun classifiers emerged, verb classifiers started to enter the language. To express the duration or iterations of an action, in Old Chinese, numbers directly preceded the verb without any classifier to link the adverbial to its verb head, as illustrated in the following examples, exactly the same as the numeral NPs discussed above.




(18)

子重、 子反於是乎七奔命。 (左傳 成公七年) Zi Zhòng, Zǐ Fǎn yúshìhū qī bēn mìng. Zi Zhong Zi Fan then seven run life “Then Zi Zhong and Zi Fan ran for their lives seven times.” (Zuo Zhuan, Cheng Gong Qi Nian, 550–400 BC)

(19)

吾日三省吾身。 (論語 學而) Wú rì sān xǐng wú shēn. I everyday three inspect my body “I inspect myself three times every day.”




(Lun Yu, Xue Er, 500 BC) (20)




三嚥, 然後耳有聞。 (孟子 滕文公) Sān yàn, ránhòu ěr yǒu wén. Three swallow then ear have hear “(He) swallowed three times and then could hear some sounds.” (Meng Zi, Teng Wen Gong, 300 BC)

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Classifiers

404

However, the above structure was no longer permissible after the beginning of Modern Chinese, and a verb classifier such as cì “times” had to be used after the number. Unlike nominal phrases, however, the “Num + CL” phrase generally follows the verb with the constituent order “VP + Num + CL,” in contrast to that of the numeral NP.3 In Contemporary Chinese there are dozens of verb classifiers, many fewer than noun classifiers, and their collocations with verbs are also largely conventional, such as the classifier dùn for the action of eating, biàn for the action of reading, tàng for the action of walking, and xià for the action of beating. That is, the numeral structure of VPs has historically undergone the following change: (21)

Num + VP > VP + Num + CL

According to Wang (1989: 34), the earliest uses of verbal classifiers are found in texts composed in the fourth century AD, as illustrated in (22) and (23), approximately three centuries later than the earliest examples of noun classifiers:




(22)

繞樹三匝。 (曹操 短歌行) Rào shù sān zā. circle tree three loop “(The bird) circled the tree three times.” (Cao Cao, Duan Ge Xing, AD 200)

(23)

余嘗往返十許過。 (水經注 江水) Yú cháng wǎng-fǎn shí xǔ guò. I ever go-return ten about time “I visited the place about ten times.” (Shui Jing Zhu, Jiang Shui, AD 500)




While noun classifiers developed entirely out of ordinary nouns, verb classifiers grammaticalized mainly from ordinary verbs, for instance, zā in (22) originally meant “surround” and guó in (23) originally meant “pass.” Additionally, some instrumental nouns could be used as temporary verb classifiers, for instance, qiè yī-dāo “cut oneknife” and dǎ yī-quán “beat one-fist.” We can assume that the emergence of verb classifiers was by analogy with noun classifiers for two reasons: first, noun classifiers entered the language nearly three 3

The “Num + CL” phrase may occur before the VP to achieve certain pragmatic values, such as complete negation, for example: 他三次没來。 Tā sān-cì méi lái. he three-CL not-have come “He didn’t come three times.”

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

The Emergence of Verb Classifiers

405

centuries earlier than verb classifiers; second, noun classifiers were much more numerous and much more frequently used than verb classifiers. If there was truly an analogy between noun and verb classifiers, the original power must have come from noun classifiers and could not have been the other way around. Here, another question arises: how did the analogy between the noun and the verb happen? A simple answer is that nouns and verbs share the same quantity feature: both represent discrete units in their prototypical members, nouns in three-dimensional space and verbs in temporal space, and hence can be counted by natural numbers such as one, two, and three. In English, for instance, there are parallel lexical forms to count the two word classes: one, two, three . . . for nouns and once, twice, thrice . . . for verbs. Langacker (1987: 214‒274) revealed that there is a semantic and grammatical parallel between countable versus uncountable nouns and perfective versus imperfective verbs, reflecting another common quantity-related property between the two basic word classes, nouns and verbs. In addition to classifiers, there have been many other parallel developments in the syntax and morphology of nouns and verbs; for example, both developed reduplication and suffixes of various types (for details, see the relevant chapters in this book). Once again, the parallel development of the noun and verb classifiers tells us that the evolution of grammar is highly regular, resembling the law of phonological change, as the neogrammarians claimed more than a hundred years ago. As a consequence, there was a sharp contrast between the distributions of noun and verb classifiers: regarding the “Num + CL” phrases, the noun modifiers underwent a historical position change from postnominal to prenominal, but the verb classifiers stayed in postverbal position. As discussed above, in Old Chinese when no classifiers existed, number words directly preceded both nouns and verbs; namely NPs and VPs shared the same numeral structures. Why did the numeral structures of NPs and VPs differ from each other when classifiers emerged? In the above section, we explained the reason for the diachronic position change of noun classifiers, proposing that it was due to the systematic change in the NP structure in which all types of modifier were limited to prenominal position. Now we explain why the phrase consisting of number and verb classifier continued to follow the VP (i.e. the predicate) during the grammaticalization process. When verb classifiers started to emerge and develop from the fifth century AD onward, the principle of action–resultative ordering that determined the information structure of the predicate took effect, requiring that all resultative elements must follow the predicate (i.e. occur in postverbal position; for a detailed discussion, see Chapter 7). With verb classifiers, their distributions are somewhat different: some occur between the verb and the object, but some can only follow the whole VO construction, depending mainly on the degree of grammaticalization of the classifier. For highly grammaticalized classifiers, the “Num + CL” phrase can optionally occur either between the verb and the object or after the whole VO construction, as illustrated in (24). “Temporary” classifiers

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Classifiers

406

(which are actually ordinary nouns and are occasionally used as an instrument in some contexts) can occur only after the whole VO construction, as illustrated in (25): (24)

(a) 他去了兩趟北京。 (現代漢語) Tā qùle liǎng-tàng Běijīng. he go-PERF two-CL Beijing “He has been to Beijing twice.” (b) 他去了北京兩趟。 (現代漢語) Tā qùle Běijīng liǎng-tàng. He go-PERF Beijing two-CL “He has been to Beijing twice.” (Contemporary Chinese)

(25)

(a) 他打了那條狗三棍子。 (現代漢語) Tā dǎle nà-tiáo gǒu sān-gùnzi. he beat-PERF that-CL dog three stick “He beat the dog with a stick three times.” (b) *他打了三棍子那條狗。 (現代漢語) *Tā dǎ-le sān-gùnzi nà-tiáo gǒu. he beat-PERF three-stick that-CL dog (Contemporary Chinese)

In short, the phrase “number + verb classifier” expressed the time or duration of an action and was regarded as a kind of resultative. Under the operation of the principle of action–resultative ordering, this type of numeral phrase must follow the matrix verb. As a result, the distributions of nominal and verbal classifiers are different in Contemporary Chinese.

16.5 Reduplication of Classifiers First, we would like to review some background information about the development of reduplication, one of the most important morphological devices in the grammar of the Chinese language, which is frequently used to express various meanings of quantification. In Contemporary Chinese, reduplication is applicable to only a couple of nouns and most classifiers, verbs, adjectives, and adverbs. The extensive and frequent use of reduplication is remarkably characteristic of the Chinese language. As discussed in Section 10.2, one of greatest puzzles about diachronic development is that reduplication for different word classes entered the language in significantly different periods: the reduplication of adjectives, adverbs, and onomatopoeic forms is already widely attested

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Reduplication of Classifiers

407

in the Shi Jing, a text composed around the tenth century BC, the earliest use of nominal reduplication is found in texts of the third century BC, and verb reduplication was not introduced into the language until the seventeenth century AD. That is, the language took nearly three millennia to develop reduplicated forms for all word classes. The timings of the emergence of reduplication for different word classes provides an ideal window for appreciating how the overall properties of the grammar in different periods guided the innovation of grammatical devices, an issue we discuss repeatedly in this book. In what follows, we focus on the development of the reduplication of nominal classifiers. Nominal reduplication refers to a morphological process of expressing a universal reference through reduplicating a root noun or classifier, such as gè-gè “everyone” and rén-rén “everybody.” In Contemporary Chinese, almost all monosyllabic classifiers can be reduplicated to make a universal reference, which is a rigorous rule with a maximal degree of productivity. However, only a couple of ordinary nouns, exclusively rén “people” and shì “thing,” can take the reduplicated form, as illustrated in (26) and (27), two naturally occurring examples:




(26)

人人都喜歡你。 (王朔 痴人) Rén-rén dōu xǐhuān nǐ. person-person all love you “Everybody loves you.” (Wang Shuo, Chi Ren, Contemporary Chinese)

(27)

我事事都按照您說的那麼辦。 (王朔 編輯部的故事) Wǒ shì-shì dōu ànzhào nín shuō dì nàme bàn. I thing-thing all follow you say NOM so do “I have done all the things according to your suggestion.” (Wang Shuo, Bian Ji Bu De Gu Shi, Contemporary Chinese)




In addition, two more nouns in Contemporary Chinese, chù “place” and shí “time,” can also take the reduplicated form but only in the written style. The reduplications of these ordinary nouns are in fact idiosyncratic and hence need to be memorized by learners because no other nouns, whether monosyllabic or multisyllabic, can be reduplicated to make a universal reference. Therefore the reduplicated forms – rén-rén “people-people” and shì-shì “thing-thing” – are usually listed as lexical entries in dictionaries, such as The Dictionary of Modern Chinese (The Commercial Press, Beijing, 2018). They are actually the fossils of a productive rule that operated roughly from the third century BC to the fifth century AD. The use of the reduplication of rén “people,” one of the most frequently used nouns, is first attested in the Meng Zi, a text composed around the third century BC, as illustrated below:

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Classifiers

408 (28)




人人親其親。 (孟子 離婁) Rén-rén qīn qí qīn. person-person love their parent “Everybody loves their parents.” (Meng Zi, Li Lou, 300 BC)

After that, an increasing number of ordinary nouns that were monosyllabic were reduplicated to make universal references, as exemplified below: (29)




則國國傳送食。 (史記 大宛列傳) Zé guó-guó chuán sòng shí. then country-country in-turn send food “Then every country in turn sent food (to them).” (Shi Ji, Da Wan Lie Zhuan, 100 BC)

(30)




事事擬學。 (世說新語 文學) Shì-shì nǐ xué. Thing-thing copy learn “This is all copying.” (Shi Shuo Xin Yu, Wen Xue, AD 450)

(31)




器器標題。 (冥祥記 珠林) Qì-qì biāo tí. utensil-utensil mark title “Every utensil was marked with a title.” (Ming Xiang Ji, Zhu Lin, AD 500)

However, as an increasing number of classifiers were introduced into the language via grammaticalization, a division of labor between the noun and the classifier was established: classifiers were specialized to express quantity properties of things, such as numbers, whereas nouns referred particularly to the varying categories of things. This division took place from the sixth century AD to the tenth. Clearly, universal reference belonged to the expression of the quantity of things. When classifiers emerged as a new grammatical category, they naturally took over the function of the reduplicated form that had previously been found with ordinary nouns. That is, the rule of nominal reduplication remained unchanged in history, but the lexical members (i.e. nouns) that were governed by the rule were replaced by classifiers, a historical event that happened in Late Medieval Chinese, as illustrated below: (32)

我見百十狗, 個個毛猙獰。 (寒山詩) Wǒ jiàn bǎi shí gǒu, gè-gè máo zhēngníng. I see hundred ten dog CL-CL hair hideous “I saw 110 dogs, and all of them had hideous fur.” (Han Shan Shi, AD 650)

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

Reduplication of Classifiers (33)

409

嗟見世間人, 個個爱喫肉。 (拾得诗) Jiē jiàn shìjiān rén, gè-gè ài chī ròu. alas see world people CL-CL love eat meat “Alas, (I) see all people in the world love to eat meat.” (Shi De Shi, AD 700)

Now, we turn to why the reduplication of different word classes was introduced into the language at such different periods with temporal intervals of approximately a millennium. Here, we focus on the reduplication of nouns (including classifiers) and adjectives (including adverbs and onomatopoeic forms).4 In our view, the phonological system of Chinese favored or facilitated the morphological device of reduplication because its words or morphemes are mostly monosyllabic and the syllabic structures are relatively simple (e.g. no consonant clusters5). The timing of when the reduplicated forms of different word classes entered the language depended on two factors: the meaning of their reduplicated form and the syntactic position of their distribution. The reduplication of adjectives functions to intensify the degree of quality, semantically similar to “very + Adj,” and their distribution is also similar to “very + Adj,” which can occur in the predicate, as illustrated in (34), or in a noun phrase, as an attributive of the noun, as illustrated in (35): (34)




楊柳依依。 (詩經 采薇) Yáng liǔ yī-yī. poplar willow sway “Poplars and willows were swaying in the breeze.” (Shi Jing, Cai Wei, 1000–600 BC)

(35)




采采芣苡。 (詩經 兔罝) Cǎi-cǎi fóuyǐ. colorful asiatica “Asiaticas are very colorful.” (Shi Jing, Tu Ju, 1000–600 BC)

However, the reduplicated forms of nouns and classifiers were syntactically conditioned: they could occur only in preverbal position, including sentence-initial position (i.e. topic or subject) and somewhere between the subject and the predicate, regardless of whether they were agents and patients. In other words, the reduplicated noun form could not follow the verb even if it was semantically governed by the verb, a rigorous 4 5

See Section 10.2 for the motivation and mechanism for the emergence of the verb reduplication. Some researchers (e.g. Yang 1981) have argued that consonant clusters existed in Old Chinese on the basis of their reconstruction in comparison to Tibetan. However, this argument is unconvincing because of the lack of any solid empirical evidence.

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

410

Classifiers

rule that has never been violated since the earliest instances were attested around the third century BC. This syntactic distribution of nominal reduplication hints at the reason why it entered the language approximately 1,000 years later than the reduplication of adjectives and adverbs. As pointed out above, nominal reduplication expresses a universal reference, covering every entity in a given domain. That is, it implies the feature definite, which must be realized by means of certain lexical or grammatical devices. In Chinese grammar, there are two options available for making a definite reference: one is adding lexical determiners such as demonstratives or personal pronouns, and the other is relying on the syntactic position to achieve the definite meaning. However, the reduplicated form of nouns and classifiers cannot be modified by any determiners; hence the only option for achieving the definite meaning is to rely on the syntactic position; that is, to use the form in preverbal position according to the principle of definiteness assignment by syntactic position (for details, see Section 7.6.1). This is the reason why nominal reduplication must appear in preverbal position. In other words, the principle of definiteness assignment by syntactic position played a key role in determining the timing of the emergence of nominal reduplication. As we discussed in depth in Section 7.6, this principle formed gradually from the third century BC to the fifth century AD, enabling the reduplicated forms of nouns and classifiers to come into existence. As analyzed in Section 9.4, this principle also played a key role in motivating the emergence of the disposal construction. Once again, this demonstrates that grammatical change cannot happen in isolation and must occur under the influence of the grammatical system at that time. Only from a diachronic point of view can we see how a grammatical system is interconnected and how concrete grammatical rules co-operate to produce well-formed instances.

https://doi.org/10.1017/9781108921831.016 Published online by Cambridge University Press

17 Demonstratives from Classifiers 17.1 Introduction In Late Medieval Chinese, two demonstratives, the proximal zhè and the distal nà, entered the language at the expense of the old demonstratives that had already existed for nearly two millennia. The different manners of emergence of these two demonstratives represent two sorts of mechanism: (a) grammaticalization from a lexical source and (b) derivation by a phonological rule, which are addressed in two separate chapters. This chapter focuses on the proximal demonstratives in Standard Chinese and other dialects, and the following chapter will deal with the distal demonstratives. Demonstratives, as a major functional category for every language, are regarded as a “zero stage” or “semantic primitive” category in that they are the lexical sources for many grammatical devices such as definite articles and complementizers, while they themselves cannot be historically derived from other lexical items (Greenberg 1978, Diessel 1999a). This view calls into question one of the central assumptions of grammaticalization theory, the claim that all grammatical markers are ultimately derived from lexical sources. This chapter demonstrates that the demonstratives zhè and gè in Mandarin Chinese and other dialects grammaticalized from two general classifiers in Late Medieval Chinese, both of which originated from ordinary nouns. Based on diachronic facts and dialect data, the present analysis addresses the motivation and mechanism for the course of development from classifiers to demonstratives. This finding may make a significant contribution to historical linguistics. The importance of demonstratives for the study of diachronic and synchronic linguistics cannot be overestimated, since they not only have important grammatical and pragmatic functions such as acting as determiners and anaphors, but also are sources for many grammatical morphemes such as definite articles, relativizers, complementizers, and third-person pronouns (Greenberg 1978, Frajzyngier 1987, 1995). Some researchers have argued that demonstratives are diachronically basic or “semantic primitives” and hence cannot be historically derived from other lexical sources (Plank 1979, Diessel 1999b). Moreover, Greenberg (1978b: 61) postulated that demonstratives

411

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

412

Demonstratives from Classifiers

were the zero stage, followed by the development of the definite article, the nongeneric article, and the noun marker. In Indo-European languages, for instance, the demonstrative pronoun to- is among the highly stable grammatical forms with no lexical origin and was later further grammaticalized into a third-person agreement marker (Heine and Traugott 1991: 10). However, Heine and Kuteva (2007: 84) suggested that there is massive evidence that demonstratives have locative adverbs as their primary diachronic sources and that verbs such as “go” or “see” may also be sources of demonstratives. They noted that in the English-based Tok Posin and other varieties of Melanesian PE, the English adverb “here” grammaticalized into a demonstrative and relative clause marker, ia (or ya). Additionally, the Portuguese-based Angolar Creole has developed a second series of demonstrative attributes derived from the locative adverbs aki (Portuguese aqui “here”) “here.” They also indicated that the demonstrative zhī in Old Chinese was derived from its verbal usage “go,” similar to a case reported in the Chadic language by Frajzyngier (1995). Even though these examples are true, a related question must be answered: are these changes cases of semantic extension or of a grammaticalization process? In the case of Chinese zhī, it is still unclear in what context and through what stages the verb zhī “go” became grammaticalized into a demonstrative, and only some seemingly semantic correlations exist between them. Furthermore, Heine and Kuteva (2007: 299) divided word classes and morphological elements into seven layers. The first three layers are all content words, namely nouns, verbs, adjectives, and adverbs, and demonstratives are the very first elements among all functional words and morphological markers. As we will see in the latter part of this chapter, however, all the Chinese classifiers evolved from ordinary nouns, which in turn served as the sources for demonstratives. That is, classifiers as functional words are somewhere between nouns and demonstratives; hence demonstratives are not in the first place among all functional words, at least in Chinese. The hierarchy proposed by Heine and Kuteva may need to be revised accordingly. Clearly, concerning the properties of demonstratives, the central issue is whether they are diachronically basic or “semantically primitive”; that is, whether they can be derived from other lexical sources. Although a type of semantic or functional similarity exists between demonstratives and locative adverbs or movement verbs, it is not clear whether the emergence of these demonstratives was a simple semantic extension or a grammaticalization process, because any grammaticalization must happen in a certain context. In contrast, semantic extension is not constrained by linguistic constructions. For instance, the verb “see” can mean “perceive visually” and “understand,” a case of semantic extension rather than grammaticalization. In general, the sources of demonstratives remain unknown, as shown in the current literature on grammaticalization. Therefore some researchers have claimed that

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

The Demonstrative Puzzle

413

demonstratives, a class of deictic categories, belong to the basic vocabulary of every language (Plank 1979, Traugott 1982), and thus cannot be traced back to lexical sources (Bühler 1934, Peirce 1955, Ehlich 1979). Diessel (1999b) even argued that demonstratives might present a second source domain from which grammatical markers might have emerged, a finding that undermines one of the central assumptions of grammaticalization theory, that all grammatical morphemes are ultimately derived from lexical sources. The above challenging question can be answered on the basis of a diachronic change in Chinese, a replacement of the old by the new demonstrative systems in Late Medieval Chinese. This chapter demonstrates that the demonstrative zhè in Mandarin Chinese and other northern dialects and the demonstrative gè in the southern dialects grammaticalized from two general classifiers in Late Medieval Chinese. According to Wang (1989), in Medieval Chinese the classifier zhī grammaticalized from a noun that originally meant “one bird,” and the classifier gè grammaticalized from the noun “one bamboo.” Thus the Chinese demonstratives underwent the following steps of development:1 (a) Noun > (b) Classifier > (c) Demonstrative The evidence is robust and includes phonological correlations, semantic/pragmatic suitability, frequency in the specific context, and the overall properties of the grammatical system at that time. My study shows that demonstratives themselves may be derived from other sources, such as classifiers, and that they are not necessarily semantic primitives or the most basic functional words since classifiers actually originated from ordinary nouns. In addition, the development from classifiers to demonstratives in Chinese is undoubtedly a case of the grammaticalization process rather than a semantic extension.

17.2 The Demonstrative Puzzle Chinese historical linguists have long debated the origins of the demonstratives zhè and nà. The interest of those researchers has centered on the issue of which demonstratives in Old Chinese were cognates with the new pair. A related issue is the identification of the relationship between the proximal zhè and the distal nà; in other words, whether they grammaticalized from two independent lexical sources or whether the distal demonstratives were derived from the corresponding proximal demonstrative via inflection.

1

The classifier zhī is the source of the demonstrative zhè, and the motivation for this phonological change will be discussed below. In this chapter, I discuss only the second step of the development, namely from classifier to demonstrative.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

414

17.2.1 The Emergence of the New Demonstrative System As we knew, Chinese has a long-documented history. Some demonstratives, such as zī and zhī, were used even in the earliest texts, the oracle bone inscriptions.2 In addition, in Old Chinese (from the thirteenth century BC to the first), quite a few other demonstratives emerged, such as sī, ruò, shì, cǐ, and bǐ, which might have been dialect variants at that time. Among these demonstratives, zhī was also used as a “go” verb; thus some researchers (Yue 1998, Heine and Kuteva 2007: 77) have speculated that the verbal use of zhī might have been the source of its demonstrative usage. However, it is still unclear whether this was just a case of semantic extension or a grammaticalization process because none of the specific contexts triggering the change have been identified yet. The demonstrative shì evolved into a copula around the first century BC and since then has ceased to be used as a demonstrative. The demonstrative zhī and others gradually went out of fashion by the second century AD. The proximal zhè and distal nà of Contemporary Chinese were introduced into the language from the seventh century AD to the tenth (Ohta 1987: 115‒122, Wang 1989: 65‒75, Lü 1985: 183‒245), but at that time the typical demonstratives were still the proximal cǐ and the distal bǐ that had been used since the Old Chinese period. That is, these two pairs of demonstratives coexisted during this period, as illustrated in the following examples: (1)

(a) 不省這個意。 (拾得詩) Bù xǐng zhè-gè yì. not understand this-CL meaning “Don’t understand this meaning.” (Shi De Shi, AD 700) (b) 那個瓶老實。(寒山詩) Nà-gè píng lǎoshí. that-CL bottle strong “That bottle is strong.” (Han Shan Shi, AD 650)

(2)

(a) 此是崔女郎之舍耳。(遊仙窟) Cǐ shì Cuī nǚláng zhī shě ěr. this be Cui lady Gen house PRT “This is the house of Lady Cui.” (You Xian Ku, AD 750)

2

Note that the demonstrative zhī and the classifier zhī are just homophones in Contemporary Chinese. Historically, their phonological forms were different; e.g. in Medieval Chinese the demonstrative zhī is reconstructed as tɕǐə1 and the classifiers as tɕǐɛk4 (Li and Zhou 1999).

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

The Demonstrative Puzzle

415

(b) 如何知彼? (葉靜能詩) Rúhé zhī bǐ ? how know that “How do you know that?” (Ye Jing Neng Shi, AD 700) In the collection of vernacular texts edited by Liu and Jiang (1990),3 the demonstrative pair cǐ (proximal) and bǐ (distal) are much more frequent than the new pair zhè (proximal) and nà (distal), as shown in Table 17.1. Through a nearly 1,000-year competition between the two pairs, the demonstrative pair zhè and nà eventually dominated over the pair cǐ and bǐ. The zhè–nà pair is currently used in Mandarin Chinese and the overwhelming majority of other northern dialects, while the cǐ–bǐ pair has been ousted and restricted to written texts. Almost at the same time that the demonstrative pair zhè and nà emerged, another demonstrative gè,4 which was originally a classifier, came into being, as illustrated in (3) and (4): (3)




真成個鏡特相宜。 (庾信 鏡賦) Zhēn chéng gè jìng tè xiāngyí really become this mirror surely suitable. “This mirror became really suitable.” (Yu Xin, Jing Fu, AD 550)

(4)




個時空林難獨守。 (駱賓王 贈道士李榮) Gè shí kōng lín nán dú shǒu. this time empty forest hard alone live “During this time, it is hard for me to live alone in an empty forest.” (Luo Bin Wang, Zeng Dao Shi Li Rong, AD 650) Table 17.1 Frequencies of demonstrative pairs from the seventh century AD to the tenth

3 4

Proximal demonstratives

Distal demonstratives

cǐ, 1010 tokens zhè, 188 tokens

bǐ, 110 tokens nà, 99 tokens

This book consists of approximately 237,000 Chinese characters and includes various genres such as dialogues, poems, travel articles, religious texts, and government documents. For the sake of simplicity, only the Pinyin forms of the corresponding Chinese characters are provided in this book unless the phonological forms of the real dialect pronunciation are relevant to the ongoing discussion.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

416

Demonstratives from Classifiers

Table 17.2 Distribution of the demonstratives zhè and gè in dialects Dialect

Demonstrative zhè

Demonstrative gè

Northern dialects (including Mandarin) Jin dialect Wu dialect Gan dialect Xiang dialect Southern Min dialect Hakka dialect Yue dialect (Cantonese)

+ + − − − + − −

− − + + + − + +

Note: For the same morpheme, the real phonetic form varies from one dialect to another. Here, I use the Pinyin forms to represent the phonological variants of the same morpheme. Real phonetic forms will be used when they are relevant to the present discussion.

According to Lü (1985: 183), the demonstrative gè was used mainly in texts authored by those who came from areas where the southern dialects were spoken at that time. It occurred sporadically in the vernacular texts from the sixth century AD to the sixteenth. Although it has been entirely ousted by the proximal demonstrative zhè in the northern dialects today, the demonstrative gè is still currently used as the only proximal demonstrative in an overwhelming majority of southern dialects. Chinese has eight major dialects and hundreds of subdialects, with remarkable differences in phonology, lexicon, and grammar.5 According to my statistical survey of Cao (2008: 12), the demonstrative zhè is now used in a total of 465 subdialects, and the demonstrative gè is used in a total of 271 subdialects. The distributions of these demonstratives in the eight major dialects are shown in Table 17.2. As the table shows, of the eight dialects, three use the demonstrative zhè, and five use the demonstrative gè. In the literature, the origin of the demonstrative zhè is still a major puzzle, but it is quite obvious that the demonstrative gè originated from its classifier usage. As indicated above, this chapter aims to show that the demonstrative zhè actually grammaticalized from another general classifier, zhī, in Medieval Chinese. The empirical evidence for this proposal includes phonological forms in history and dialects, semantic and syntactic properties, and the specific context in which the classifiers zhī and gè grammaticalized.

5

The so-called “eight major dialects” are the northern dialect, the Jin dialect, the Gan dialect, the Xiang dialect, the Wu dialect, the Southern Min dialect (Taiwanese), the Hakka dialect, and the Yue dialect (Cantonese). In this chapter, the term “dialect” specifically refers to any of these eight major dialects, and the term “subdialect” is reserved for smaller dialect areas.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

The Demonstrative Puzzle

417

17.2.2 Lexical Sources for Demonstratives The introduction of the demonstrative pair zhè and nà into the language in Late Medieval Chinese and their ultimate replacement of the old demonstrative system in Modern Chinese (from the eleventh century AD to the nineteenth) is one of the greatest changes in the history of Chinese. Many scholars have tried to identify the sources of these new demonstratives. In Chinese historical linguistics, the dominant methodology is still the attempt to find cognates of this pair in Old Chinese, which is the traditional philological method when lacking an awareness that demonstratives might have evolved from other sources. The two most influential views, which were proposed by the two best-known scholars of Chinese linguistics, deserve to be mentioned here. Lü (1985: 183) speculated that the demonstrative zhè might have come from zhě, a pronominal item that can be traced back to the very beginning of the language. Wang (1989: 65) disagreed with this view, pointing out that the grammatical functions of zhè and zhě are dramatically different. The demonstrative zhè always precedes the head noun, but the pronominal zhě could never do so; instead, zhě always functions as the nominal head, modified by verbs or adjectives, to form a noun phrase; e.g. shì zhě “eat person” means “eater.” Wang proposed that the demonstrative zhè might have been derived from the demonstrative zhī in Old Chinese. However, this view remains problematic for two reasons. First, the demonstrative zhī was not in use in the spoken language around the first century AD, and the demonstrative zhè was not introduced into the language until the seventh century AD, which means that there was an approximately 700year gap between the loss of zhī and the emergence of zhè (Ohta 1987: 117). Second, the functions of zhī and zhè are fundamentally different: zhī can never omit its head noun to replace an entire nominal phrase, but zhè can freely omit its head noun to replace it with an entire nominal phrase, as illustrated in (5) and (6): (5)




之人也, 之德也, 將旁礡萬物。 (莊子 逍遙遊) Zhī rén yě, zhī dé yě, jiāng pángbó wànwù. this person PRT this moral PRT will surround all-thing “This person and this moral will surround all things.” (Zhuang Zi, Xiao Yao You, 300 BC)

(6)




彼中還有這個也無? (祖堂集 石頭和尚) Bǐ zhōng hái yǒu zhè-gè [ ] yě wú? that inside still have this-CL PRT not “Does that place still have this object or not?” (Zu Tang Ji, Shi Tou He Shang, AD 950)

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

418

Demonstratives from Classifiers

Throughout history, the demonstrative zhī has always been an adjective-like demonstrative, and the noun chóng “insect” could never be dropped. In contrast, even in the very beginning, the demonstrative zhè, often together with a classifier, could freely omit the head noun to replace the entire NP. Zhang (2001) conducted a cross-dialect study of the possible source of the demonstrative zhè. She found that hundreds of subdialects in southern areas use the demonstrative gè, which has been the most general classifier for more than 1,500 years. Additionally, she noted that the demonstrative wū in the eighteen subdialects of the Jin dialect is also used as a general classifier and that the demonstrative mò in the Haikou subdialect is also used as a classifier. Her findings and argument are summarized in Table 17.3. Drawing an analogy between the first three pairs, she hypothesized that the demonstrative zhè is also derived from a general classifier, zhī, in Medieval Chinese, pointing out that their different phonological forms resulted from the evolution of the phonological system. Her hypothesis is strongly supported by the following two pieces of phonological evidence. (a) In two ancient phonological books, the Qie Yun, edited in AD 601, and Piao Tong Shi Yan Jie, edited in the fifteenth century AD, the demonstrative zhè and the classifier zhī have exactly the same phonological form. (b) In at least nine contemporary subdialects, the demonstrative zhè and the classifier zhī have the same phonological form. For instance, in the Fuqing subdialect (belonging to the Fuzhou Min dialect), both are tsia21.6 The above hypothesis is brilliant, but Zhang (2001) based it only on phonological forms and failed to prove it by means of semantic and syntactic evidence. Thus her view

Table 17.3 Correlations between classifiers and demonstratives

6

Words

Classifiers

Demonstratives

gè wū mò zhè

+ + + ?

+ + + +

The two numbers on the top right of a syllable represent the tone contours, and a single number represent the tonal type of the syllable.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

The Demonstrative Puzzle

419

seems to have convinced few researchers in the area. For instance, the two most influential surveys of historical Chinese, Jiang (2005b) and Jiang and Cao (2005), do not even mention Zhang’s analysis. Unfortunately, her insightful idea remains largely unknown. In line with Zhang’s hypothesis, this chapter addresses the various factors motivating the classifiers to develop into demonstratives, including contexts, frequency, and semantic suitability, as well as the overall properties of the grammar at that time.

17.2.3 Asymmetry between Proximal and Distal Demonstratives In any language, demonstratives always occur in pairs, namely proximal and distal, and in some languages a medial demonstrative can appear between them.7 The following question arises: what is the relationship between the proximal and distal demonstratives? Are they completely independent of each other, or is one of them more basic and the other derived from that basic one? In other words, can they have different lexical sources? At first glance, for instance, it appears that there are some phonological correlations between the English demonstratives this–that and these–those, revealing that one may be derived from the other. However, no historical evidence shows the specific derivational process. Some scholars have speculated that the distal demonstrative nà in Chinese originated from the former ruò or ěr, which can be traced back to Old Chinese (Ohta 1987: 120, Lü 1985: 183, Wang 1989: 65). If so, the proximal and distal demonstratives have different sources. However, this view is implausible because of the lack of phonological and syntactical evidence. In contrast, robust evidence in the Chinese dialects reveals that there is a derivational correlation between the proximal and distal demonstratives, and the proximal demonstrative is usually the basic demonstrative from which the distal demonstrative is derived. In the book, The Dialect Lexicon of Chinese,8 among the twenty subdialects, only the Guangzhou subdialect is an exception: its general classifier gè grammaticalized into a distal demonstrative, pronounced kɔ35, and its proximal demonstrative is realized as ni55. In the remaining nineteen subdialects, all the demonstratives that originated from classifiers are realized as proximal, and their corresponding distal demonstratives take a derivational form. Concerning the correlation between the proximal and distal demonstratives, there are three types of dialect, as follows.9

7 8 9

It has been reported that some Chinese subdialects have a third demonstrative that refers to a modest distance (see Huang 1996: 465‒505 for details). This book was edited by the Department of Chinese Language and Literature at Peking University and published by The Yuwen Press in 1995 in Beijing. The demonstratives in some dialects may have several phonological variants, but I select only one in the following discussion. The data are cited from Huang (1996: 465).

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

420

Demonstratives from Classifiers Type 1. The proximal and distal demonstratives have exactly the same phonological form, e.g. in the Suzhou subdialect of Wu, both are kɤʔ44. Type 2. The proximal and distal demonstratives have the same syllables, changing tone values to distinguish the proximal from the distal; e.g. in the Meixian subdialect of Hakka, the proximal is kɛ21, and the distal is kɛ52. Type 3. The proximal and distal demonstratives are differentiated by vowel; e.g. in the Danyang subdialect, the proximal demonstrative kiʔ22 takes the front and high vowel [i], and the distal demonstrative kə33 takes the schwa [ə].

These phenomena are worthy of further study. The two members of a demonstrative pair are always syntactically parallel. Therefore this chapter focuses on the proximal demonstratives, but the generalizations are perfectively applicable to the corresponding distal demonstratives.

17.3 Changes in Demonstratives During the Medieval Chinese period, several major changes occurred that worked together to trigger the grammaticalization process from classifiers to demonstratives. First, classifiers gradually grammaticalized from ordinary nouns, which fundamentally altered the structure of nominal phrases. Second, the disposal construction emerged, serving to introduce a definite NP in preverbal position. Third, preverbal positions, mainly the subject or the topic, became able to assign a definiteness feature to bare nouns and “CL + N” phrases.

17.3.1 The Establishment of the Classifier System Mandarin Chinese and other dialects are now all classifier languages,10 which means that a proper classifier must be employed to connect a numeral and a noun. However, Old Chinese was not a classifier language. In brief, classifiers began to grammaticalize from ordinary nouns as early as the first century BC but were not firmly established until the fifteenth century AD, a developmental process that took more than 1,500 years (for details, see Chapter 16). We conducted a large-scale investigation showing that in texts from the Tang dynasty (from the seventh century AD to the tenth), “Num + N” phrases account for 70 percent and “Num + CL + N” phrases account for 30 percent of instances, and that in texts from the Song dynasty (from the tenth century AD to the thirteenth), 10

Here, the term “classifier” is concerned only with sortal classifiers in terms of Croft (1996) or count classifiers in terms of Cheng and Sybesma (1999).

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Changes in Demonstratives

421

“Num + N” phrases account for 40 percent and “Num + CL + N” phrases account for 60 percent of instances. The following three examples are selected from texts of different periods:




(7)

夙夜匪懈, 以事一人。 (詩經 烝民) Sù yè fěi xiè, yǐ shì yī rén. day night not relax for serve one person “I worked day and night, without relaxing, to serve one person (the king).” (Shi Jing, Zheng Min, 1000–600 BC)

(8)

七人常集於竹林之下。(世說新語 任誕) Qī rén cháng jí yú zhú lín zhī xià. seven person often meet in bamboo forest PRT under “The seven people often met in the bamboo forest.” (Shi Shuo Xin Yu, Ren Dan, AD 450)

(9)

每見不過五六人十數人。 (朱子語類卷四十) Měi jiàn bùguò wǔ liù rén shí shù rén. every see just five six person ten several person “I saw five, six or about ten people every time.” (Zhu Zi Yu Lei, Chapter 40, AD 1200)




Note that both the “Num + N” phrases in (7) and (8) are interpreted as definite, but the former is in postverbal position and the latter is in preverbal position. Thus the distribution and interpretation of nominal phrases regarding definiteness were quite different from those in Contemporary Chinese, including Mandarin and its dialects; this issue warrants further examination. The development of classifiers reached its completion point around the fifteenth century AD, and a classifier became obligatory when connecting a numeral (usually including a demonstrative) and a head noun. The emergence of this new grammatical category fundamentally altered the structure of nominal phrases; specifically, the structure underwent the following change: (10) (a) Num/Dem + N → “Num/Dem + CL + N” or “CL + N”11 Correspondingly, the referring devices became much richer, and a classifier was able to express the meanings of the related NP, either alone or in combination with a numeral or a demonstrative to replace the whole NP, namely omitting the head noun, as illustrated in the following examples: 11

When the numeral is “one” or the classifier phrase occurs in subject position, the numeral and demonstrative can be omitted.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

422 (11)

訪覓漢時人, 能無一個在? (寒山詩) Fǎng mì Hàn shí rén, néng wú yī-gè zài? visit seek Han time people, can not one-CL live “I am seeking the people of the Han dynasty, but can it be true that nobody still lives?” (Han Shan Shi, AD 650)

(12)

這個是阿誰不是? (敦煌變文 舜子變) Zhè-gè shì āshuí bù shì? this-CL be someone not be “Is this person the one?” (Dun Huang Bian Wen, Shun Zi Bian, AD 800–1000)

(13)

據見目前無個識。 (呂岩 七言) Jùjiàn mùqián wú gè shí. see now lack CL know “From what I see, there is no one whom I know now.” (Lü Yan, Qi Yan, AD 850)







The head noun rén, “person” or “people,” is omitted in the three above examples, a type of expression that was enabled by the emergence of the classifier system.

17.3.2 Definite Interpretation of Bare Nouns As far as bare nouns are concerned,12 there was no connection between syntactic positions and (in)definiteness interpretation in Old Chinese. Regardless of whether they were in preverbal or postverbal positions, bare nouns could be interpreted as either definite or indefinite depending purely on their context (for details, see Section 7.6). Let us consider some examples. (14)

12




燕國有勇士秦舞陽, 年十三, 殺人, 人不敢忤視。(史記 刺客列傳) Yàn guó yǒu yǒngshì Qín Wǔyáng, nián shísān. Yan country have hero Qin Wuyang, age thirteen shā rén, rén bùgǎn wǔ-shì. kill person person not-dare directly-look-at “The country Yan had a hero named ‘Qin Wuyang.’ He killed a person at the age of thirteen and other people dared not look him in the eye.” (Shi Ji, Ci Ke Lie Zhuan, 100 BC)

Bare nouns and “CL + N” phrases manifest the same behavior in terms of being marked as definite by preverbal position throughout history.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Changes in Demonstratives

423

In (14), the same bare noun occurs twice. The first bare noun rén “person” was indefinite and singular, meaning “a person”; the second bare noun rén “person” was indefinite and plural, meaning “other people.” In Contemporary Mandarin, the first usage of a bare noun in the object position is still well formed, but the second usage is ill-formed. Having considered the above examples, we cannot conclude that bare nouns in postverbal position were interpreted as indefinite because another possibility also existed at that time. Let us consider the following example: (15)




得一柏樹, 數日中果震柏粉碎。 (世說新語 術解) (a) Dé yī bǎi shù, get one cypress tree “(He) got a cypress.” (b)

Shù rì zhōng guǒ zhèn bǎi fěnsuì. several day within really shake cypress shatter “Within several days, the cypress was fully shattered into pieces by lightning.” (Shi Shuo Xin Yu, Shu Jie, AD 450)

The example in (15a) is an earlier sentence in this context that mentioned bǎi “a cypress.” Thus, in (15b), the bare noun bǎi “cypress” is anaphoric and so definite, but is used as the object of the verb zhèn “hit.” According to Jenks (2018), it is impossible for a bare noun to be interpreted as anaphoric in Contemporary Mandarin. Obviously, this was not the case before the fifth century AD. However, bare nouns in subject position were often interpreted as indefinite, as illustrated in (16). (16)




人問: “痛邪? ” (世說新語 德行) Rén wèn: “Tòng yé?” someone ask painful QUE “Someone asked, ‘is it painful?’” (Shi Shuo Xin Yu, De Xing, AD 450)

The bare noun rén “person” in (15) is indefinite, and this expression is ruled out by the grammar of Contemporary Mandarin. To express an indefinite topic or subject, the indefinite affix yǒu, which grammaticalized from an existential verb, must be added to bare nouns. In Contemporary Mandarin, a bare noun in subject position can be interpreted as either definite or generic, as mentioned by Cheng and Sybesma (1999). In texts from the fourth century AD to the twelfth, “CL + N” phrases could also be used in the subject position and interpreted as definite, as in Cantonese usage, illustrated in (17):

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

424 (17)




個事何曾落見聞。 (范成大 東林寺) Gè shì hécéng luò jiàn wén. CL thing ever fall PASS hear “This thing could hardly be heard.” (Fan Cheng Da, Dong Lin Si, AD 1150)

According to the comprehensive investigation by Lü (1985: 183), “CL + N” phrases in subject position were always interpreted as definite at that time. The general classifier gè grammaticalized into a demonstrative and associative morpheme in Mandarin from the sixth century AD to the twelfth (see Cao 1995: 139‒150 for details). Such usage continues in many contemporary southern dialects (see Huang 1996: 546‒553 for details). In northern dialects, however, the demonstrative use of gè was replaced by zhī, which was also grammaticalized from a classifier (Zhang 2001), and the associative morpheme was expressed by dǐ, which was originally a demonstrative. It is still not clear when or why preverbal positions lost the possibility for bare nouns to appear there and still be assigned the feature indefinite. However, it is not later than the twelfth century AD that I find the emergence of the grammatical rule that requires that bare nouns in subject position be interpreted as definite (or generic). One of the early examples is illustrated in (18): (18)

馬喫了這和草。 (老乞大諺解) Mǎ chī-le zhè-hé cǎo. horse eat-PERF this-CL grass “The horses have eaten the grass.” (Lao Qi Da Yan Jie, AD 1350)

There was no connection between the syntactic positions and the definiteness interpretation of bare nouns and “Num + N” throughout most of history. Hence they could be interpreted as either definite or indefinite, regardless of whether they occurred in preverbal or postverbal positions. As in contemporary dialects (e.g. Cantonese), “CL + N” could occur in subject position and be interpreted as definite in Old Chinese.

17.3.3 Motivation for the Disposal Construction The disposal construction began to emerge around the sixth century AD and became fully fledged after the eleventh century AD. This change was motivated mainly by the establishment of the resultative construction. A detailed discussion about the interaction between these constructions is beyond the scope of this chapter. Here, I would like to repeat an earlier example to illustrate the correlation between the two constructions:

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Changes in Demonstratives (19)

425




數日中果震柏粉碎。 (世說新語 術解) Shù rì zhōng guǒ zhèn bǎi fěnsuì. several day within really shake cypress shatter “Within several days, the cypress was fully shattered into pieces by lightning.” (Shi Shuo Xin Yu, Shu Jie, AD 450)

Example (19) represents the paradigmatic structure of the resultative construction in Medieval Chinese (from the first century AD to the tenth). At that time, if the resultative was an intransitive verb or an adjective, the object could occur only between the verb and the resultative phrase.13 Due to reanalysis, the boundary between the verb and the resultative phrase disappeared and they fused into a single constituent. After that, objects could no longer intervene between the verb and the resultative. This reanalysis caused another change: if the patient noun was bare and definite, it had to be moved to preverbal position, and one of the most appropriate structures was the disposal construction. The corresponding expressions in contemporary Mandarin are illustrated in (20). (20)

(a) The disposal construction: 雷電把柏樹震碎了。(現代漢語) Léi diàn bǎ bǎishù zhèn-suì le. thunder lightning BA cypress shake-break PERF “The lightning shook and broke the cypress into pieces.” (b) The passive construction: 柏樹被雷電震碎了。 Bǎishù bèi léidiàn zhèn-suì le. cypress PASS thunder-lightning shake-break PERF “The cypress was shattered into pieces by lightning.” (c) The SVO construction: *雷電震碎了柏樹。 (如果賓語為有定的) *Léidiàn zhèn-suì le bǎishù. (if the object is definite) thunder shake-break-piece PERF cypress (Contemporary Chinese)

The disposal construction represents the structure of active declarative sentences, which is pragmatically different from that of passives. Thus, in certain contexts, this construction 13

This situation is similar to English resultative constructions, such as the dog barked my child awake, where the object intervenes between the verb and the resultative (Goldberg 1995: 180‒198).

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

426

must be used to satisfy a certain pragmatic purpose. Because the bare noun bǎishù “cypress” was mentioned previously,14 it must occur in preverbal position to be assigned definiteness; thus (20c) is ill-formed. If it is modified by “Dem + CL,” the nominal phrase can be used as an object, as illustrated in (21) (Jenks 2018): (21)

他砍到了那棵柏樹。(現代漢語) Tā kǎn-dào le nà-kē bǎishù. he cut-down PERF that-CL cypress “He cut down that cypress.” (Contemporary Chinese)

In response to the firm establishment of the resultative construction, another construction was invented, called the “verb-copying construction” (Li and Thompson 1981: 442‒450), in which the first verb introduces the patient and the repeated verb is followed by a resultative element. The construction is [Subj + Vi + Obj + Vi + RP]. If the argument affected by the resultative phrase is the subject or agent of the main verb and the patient argument is an indefinite bare noun, the verb-copying construction is the best candidate to express the meaning, as illustrated in (22): (22)

他看書看累了。 (現代漢語) Tā kàn shū kàn-lèi le. she read book read-tired PERF “She read too much and became tired.” (Contemporary Chinese)

The direct motivation for the emergence of the disposal construction is the introduction of a definite nominal phrase in preverbal position, a change in response to the fusion of the verb and the resultative phrase. Recall that bare nouns, “Num + N,” and “CL + N” could be interpreted as definite at least prior to the tenth century AD. The fact that three types of nominal phrase can be interpreted as definite in the disposal construction can be easily explained. The emergence of the function of preverbal positions to mark the definiteness of bare nouns and “CL + N” phrases, including the topic, the subject, and the disposal construction,15 was a necessary condition for motivating the grammaticalization from classifier to demonstrative.

14 15

Bǎishù “cypress” is a disyllabic noun in Contemporary Chinese, corresponding to the monosyllabic word bǎi in Classical Chinese. For the sake of simplicity, I mention subject position only when discussing preverbal positions in marking the definiteness of “CL + N” phrases.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

From Classifier to Demonstrative

427

17.4 From Classifier to Demonstrative The necessary condition for the grammaticalization from classifier to demonstrative is that “CL + N” phrases without numerals can occur in preverbal position, where they are automatically assigned a definiteness feature by their syntactic position. By nature, classifiers share commonalities with demonstratives, such as being used for anaphora, individualization, and reference, a point that I will return to later. Thus the general classifiers zhī and gè, which had the highest frequency and widest distribution, permanently acquired the feature definite in preverbal positions and ultimately became demonstratives.

17.4.1 The High Frequency of General Classifiers Grammatical changes are inseparable from the absolute frequency of the forms and the frequency of their co-occurrences with other forms (Hopper and Traugott 2003: 40). The increased frequency of a grammatical morpheme and a construction over time is prima facie evidence of grammaticalization. Furthermore, Bybee (2003: 602) noted that “frequency is not just a result of grammaticalization, it is also a primary contributor to the process, an active force instigating the changes that occur in grammaticalization.” The high frequency of the two general classifiers made them special among all classifiers in terms of being grammaticalized into demonstratives. According to Lü (1999: 710‒715), approximately 270 classifiers are used in Contemporary Mandarin. Their distributions and frequencies vary dramatically. In Mandarin Chinese, gè is the most general classifier, and it can collocate with numerous nouns referring to human beings, animals, and inanimate objects (abstract and concrete). Thus its distribution and frequency are much higher than those of other classifiers. However, to identify the most general classifier, factors such as diachronic change and dialect variation must be considered. A reliable criterion is to determine whether the classifier can modify nouns about humans: a general classifier must be able to modify human nouns because human nouns are among the most frequently used words. First, let us consider the general classifiers for human nouns from a cross-dialect perspective. According to the Dialect Lexicon of Chinese, only gè and zhī can modify human nouns, and their distributions are shown in Table 17.4. Both gè (originating from units of bamboo) and zhī (originating from “one bird”) grammaticalized into classifiers in Medieval Chinese and belong to the group of earliest classifiers. According to Wang (1989: 18), both are among the eight most popular classifiers in Medieval Chinese. By Medieval Chinese, they were already the general classifiers with the highest frequency.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

428

Table 17.4 The dialect distributions of the general classifiers gè and zhī General classifier

Number of subdialects

gè zhī gè/zhī

14 4 2

17.4.2 Grammaticalization from Classifier to Demonstrative In contrast to semantic extension, grammaticalization is defined as the process by which a lexical item becomes a grammatical morpheme in the context of a particular construction. Typically, this process involves a construction that contains particular lexical items becoming grammaticalized. The central point of the present analysis is to identify the context in which grammaticalization from classifier to demonstrative is realized. In what follows, I demonstrate that the general classifiers grammaticalized into demonstratives in preverbal positions. It is widely recognized that bare nouns in subject position are automatically assigned the feature definite, a common property in Mandarin Chinese and other dialects (Chao 1979: 45, Zhu 1982: 95). In some dialects, such as Cantonese and the Wu dialect, “CL + N” phrases can also be assigned the feature definite, though they cannot be used in subject position of Mandarin Chinese and some other dialects. Of the 930 subdialects investigated by Cao (2008: 14), 170 subdialects still allow “CL + N” phrases to occur in subject position, as illustrated in (23) and (24): (23)

tsuəʔ luoʔiəŋtɕi səʔ niŋ nɛʔ-tɕʰy ləʔ? CL recorder what person take-way PERF “Who took away that recorder?” (Huang 1996: 136, Shanghai subdialect)

(24)

kʰuᴇ ȵiaŋ li tɕʰiɪʔ-tʰəu tsᴇ puo. CL allow he eat-finish PERF PRT “He was allowed to eat this piece (of food).” (Huang 1996: 134, Suzhou subdialect)

Note that when no numerals occur, “CL + N” refers to a singular entity. In (23), a “CL + N” phrase occurs in the topic position, and in (24) a classifier alone occurs in subject position; both are interpreted as definite. As we will see later, the same usage could be found in Late Medieval Chinese. Although the above dialect usages are no longer allowed in Mandarin and many other dialects, they were indeed grammatical in the standard Mandarin of the Medieval

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

From Classifier to Demonstrative

429

Chinese period. In the literature, all the earliest examples of the demonstratives gè and zhī can be interpreted ambiguously, exactly like the above dialect usages. In other words, the so-called demonstratives in the literature (e.g. Lü 1985: 183) were actually classifiers that were just temporarily interpreted as definite in subject position. The following are among the earliest examples that are often cited in the literature (Lü 1985: 183, Ohta 1987: 115, Wang 1989: 65), where the so-called earliest demonstratives all occurred in subject position:




(25)

個人諱底? (北齊書 徐之才) Gè rén huì dǐ? CL/ this avoid-as-taboo what “What does this person avoid as a taboo?” (Bei Qi Shu, Xu Zhi Cai, AD 650)

(26)

個是誰家子? (寒山詩) Gè shì shuí jiā zǐ? CL/this be who family girl “Whose girl is this?” (Han Shan Shi, AD 650)




(27)

只言知了盡悲傷。 (敦煌變文 歡喜國王緣) Zhī yán zhī-le jìn bēishāng. CL/this word know-PERF all sad “(When they) understood these words all became sad.” (Dun Huang Bian Wen, Huan Xi Guo Wang Yuan, AD 800–1000)

(28)

只是西方羅漢僧。 (敦煌變文 目連緣起) Zhī shì xīfāng luóhàn sēng. CL/this be Western arhat monk “These are the Western arhats and monks.” (Dun Huang Bian Wen, Mu Lian Yuan Qi, AD 800–1000)




Because of the limitations of ancient texts, we can hardly identify the specific process through which the two general classifiers evolved into demonstratives step by step. Fortunately, the vast data from contemporary dialects provide reliable evidence of the possible grammaticalization process. Eventually, the demonstrative zhī won over gè in the northern dialects (including Mandarin), but gè won over zhī in the southern dialects. One issue needs to be explained. In Mandarin Chinese, the classifier zhī has the phonological form tʂʅ55, but its corresponding demonstrative is pronounced tʂɤ51. Why do they currently have different phonological forms? As mentioned earlier, the classifier and demonstrative zhī have exactly the same phonological form in many contemporary

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

430

subdialects. It is not uncommon for a grammaticalization process to cause phonological differentiation of the involved elements. For instance, in the Suzhou subdialect, the onset sound changed from voiceless to voiced, kɤʔ > gɤʔ, when its classifier gè grammaticalized into a demonstrative. In addition, this process often causes changes in tonal value; e.g. in the Changsha subdialect, the classifier is pronounced ko11, but its corresponding demonstrative is either ko24 or kai24, in which the tone and vowel have changed. 17.4.3 “CL + N” Phrases in the Disposal Construction Three important changes occurred in Late Medieval Chinese: (a) the large-scale grammaticalization of classifiers from nouns, (b) the emergence of the demonstrative pair zhè and nà, and (c) the establishment of the disposal construction. The disposal morpheme bǎ introduces a definite NP in preverbal position, which might also have played a role in triggering the change from classifiers to demonstratives. Even in Contemporary Mandarin Chinese, “CL + N” phrases still often occur in the disposal construction, which contrasts with the views of Cheng and Sybesma (1999), Jenks (2018), and many others. Having investigated the Contemporary Chinese Corpus of Peking University, I find that there are thousands of disposal instances with a “CL + N” phrase, indicating that this is a regular usage. To guarantee accuracy, I use only naturally occurring examples or those from authoritative grammatical books to support my analysis. Common nouns can also be modified by the general classifier gè when appearing in a disposal construction. In this case, the classifier gè functions like the demonstrative pronoun zhè “this” to emphasize definiteness, as illustrated in (29) and (30):




(29)

把個女兒慣得一點樣兒都沒有。 (曹禺 雷雨) Bǎ gè nǚér guàn-dé yī-diǎn yàng dōu méiyǒu. BA CL daughter spoil-PRT a-bit shape all lack “You have completely spoiled your daughter.” (Cao Yu, Lei Yu, AD 1940)

(30)

把個屋子打扮的花里胡哨。 (人民日報) Bǎ gè wūzǐ dǎbàn-dé huā-li-hú-shào. DISP CL room decorate-PRT colorfully “They have decorated the room colorfully.” (People’s Daily, AD 1992)

As mentioned above, the classifier gè already grammaticalized into a fully fledged demonstrative pronoun in the period from the fifth century AD to the sixteenth, and it is

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

431

From Classifier to Demonstrative

still widely used as a demonstrative in many southern dialects (see Lü 1985: 183‒245 and Huang 1996: 546‒553 for details).

17.4.4 Replacement of Demonstrative Pronouns To offer a plausible explanation for any diachronic change, we must examine the change in a larger context. When the demonstrative pair zhè and nà were introduced into the language in late Medieval Chinese, the dominating demonstrative pair was still cǐ and bǐ, which had been used since Old Chinese. These new and old demonstrative pairs competed for nearly 1,000 years, but the zhè–nà pair completely replaced the cǐ–bǐ pair around the sixteenth century AD. For the sake of simplicity, Table 17.5 shows the rise and decline of the two proximal demonstratives. Superficially, the figures reflect the replacement process of the old by the new demonstrative pair. The essential question is why the new demonstrative zhè could gain momentum to develop from weak to strong and was finally able to oust the old cǐ, which had been the dominant demonstrative. The answer is shown in Table 17.6. Although the total number of appearances of the demonstrative cǐ is five times larger than that of zhè in Late Medieval Chinese, the ratio is exactly the opposite when a classifier occurs (i.e. “Dem + CL + N,” “Dem + CL,” and “Dem + CL + Num + N”).

Table 17.5 The rise and decline of the demonstratives zhè and cǐ Period

Size of text

Percentage of cǐ (tokens)

Percentage of zhè (tokens)

7th–10th centuries AD 11th–13th centuries AD 14th–16th centuries AD

237,000* 298,000 177,000

84% (1,010) 58% (1,320) 21% (307)

16% (188) 42% (951) 79% (1,156)

* The figures indicating the sizes of texts are approximate rather than accurate. Note: This survey is based on the three collections of vernacular texts edited by Liu and Jiang (1990, 1992, 1995).

Table 17.6 Differences between the demonstratives cǐ and zhè

Demonstratives

Total

Occurring alone

Dem + N

Dem + CL +N

Dem + CL

Dem + CL + Num + N

cǐ zhè

1,009 139

43% (430) 4% (5)

54% (559) 21% (30)

2% (16) 38% (53)

1% (4) 36% (49)

0% (0) 1% (1)

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

432

Demonstratives from Classifiers

The total of the zhè instances is 103, accounting for 75 percent of all instances, but that of the cǐ cases is merely twenty, accounting for only 3 percent of all instances. This means that the demonstrative zhè served mainly to combine with classifiers. Until the development of zhè, the demonstrative cǐ had been used for approximately 2,000 years. Its two typical usages are alone and directly modifying a head noun without anything between them.16 With the development of the classifier system, a classifier became obligatory in connecting a numeral and its head noun. Correspondingly, a classifier or the plural xiē increasingly occurred between a demonstrative and its head noun. According to my investigation, in Mandarin Chinese these instances account for more than 90 percent of all “Dem + NP” phrases. In many dialects, however, the demonstratives could no longer be directly combined with a head noun, which means that a classifier became obligatory when combining with them. This requirement is the power behind the replacement of the old demonstrative system by the new system. Although the new demonstratives share the basic deictic function with the old ones, their functions are remarkably different because they have different lexical sources and distinct grammaticalization contexts. All the old demonstratives, including cǐ–bǐ and those in Old Chinese (e.g. sī, ruò, shì), lack the following functions or features of the new demonstrative zhè (and the distal nà). (a) Singular/plural contrast. “Dem + CL + N” always refers to a singular noun; e.g. zhè běn shū “this CL book” means “this book.” “Dem + xiē + N” always refers to a plural; e.g. zhè xiē shū “this PL book” means “these books.” Jenks (2018) claimed that demonstratives in Chinese are the most important devices for marking anaphoric definiteness. In the Aari and Gamo languages, nouns may be inflected to indicate case and number, but only when marked as definite (Hayward 1990: 442–445, Corbett 2004: 166). Further grammaticalization from demonstratives to quasi-definite markers may have pushed the Chinese language to develop in the direction of the Aari and Gamo languages. (b) Substitute function. In the very beginning, the demonstrative zhè often combined with a classifier, and this combination was able to replace the whole NP (the head noun was omitted). A “Dem + CL” phrase could even be modified by another “Num + CL,” which shows that the “Dem + CL” phrase was truly similar to an ordinary noun.

16

For the sake of simplicity, I discuss only the proximal demonstratives, but the corresponding distal demonstratives have exactly the same syntactic differences.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

From Classifier to Demonstrative (31)

433




彼中還有這個也無? (祖堂集 石頭和尚) Bǐ zhōng hái yǒu zhè-gè yě wú? that inside still have this-CL PRT not “Does that place still have this object?” (Zu Tang Ji, Shi Tou He Shang, AD 950)

(32)

不是給了一個這個了麼? (郭德綱相聲) Bù shì gěi le yī-gè zhè-gè le me? Not be give PERF one-CL this-CL PERF PRT “Didn’t I already give him one of these?” (Cross Talks of Guo Degang, Contemporary Chinese)

In Contemporary Chinese, the demonstrative zhè has two phonological forms: tʂɤ51 and tʂei51. It is generally believed that tʂei51 is the fusion of the demonstrative zhè and the number yī ([i55]) “one,” probably because the coda of tʂei resembles the nucleus of the number “one” (e.g. Zhu 1982). In my view, this hypothesis is unlikely, and the tʂei51 form is probably a fusion of the demonstrative zhè and the general classifier gè. According to Lü (1985: 183), in their first several hundred years, zhè and gè always occurred together, and gè even behaved as a suffix to the demonstrative zhè. Thus the following historical usage can be attested: (33)

這個一場狼藉不是小事。 (景德傳燈錄) Zhè-gè yī-chǎng lángjí bù shì xiǎo shì. this-CL one-CL mess not be small matter “This mess is not a small matter.” (Jing De Chuang Deng Lu, AD 1000)

In (33), zhè-gè behaves as a single unit to modify another “Num + CL + N” phrase. Throughout history, the co-occurrences of zhè and the classifier gè are much more frequent than those of zhè and the numeral yī. Thus the demonstrative zhè is often used to replace a noun in object position: (34)

他們拿這作原料。 (現代漢語八百詞) Tāmen ná zhè zuò yuánliào. they take this use material “They take this as raw material.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 392)

A similar phenomenon occurs in Nauru, a Micronesian language, in which numeral classifiers are fused with demonstratives in a numeral noun phrase (Kayser 1993: 41, Aikhenvald 2000: 182‒183).

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

434

(c) Possessive marker. In the typology of classifiers, so-called genitive classifiers refer to all classifiers that are used in possessive constructions (Grinevald 2000: 66, Croft 2001: 120‒125). According to Langacker (2008: 401), in Uto-Aztecan possessive expressions, the possessor prefix attaches to a classifier, to which the possessed noun stands in a sort of appositive relation. A similar phenomenon is found in some Chinese dialects; for instance, in the Kaiping subdialect, all classifiers can be used as genitive markers, e.g. tā juàn shū “he CL book” means his book (Yue 1998).17 The demonstrative zhè, which grammaticalized from a classifier, also displayed this function over time, as illustrated in (35) and (36):




(35)

則我這兩條腿打折般疼。 (元刊雜劇 神奴兒大鬧開封府) Zé wǒ zhè liǎng-tiáo tuǐ dǎ-shé bān téng. but I this two-CL leg beat-break like pain “But my two legs are painful as if broken.” (Yuan Kan Za Ju, Shen Nu Er Da Nao Kai Feng Fu, AD 1300)

(36)

屋裡這人是誰? (現代漢語八百詞) Wūlǐ zhè rén shì shuí? room this person be who “Who is the person of (in) the room?” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 656)

(d) Appositive anaphora. Lü (1985: 183) found that classifiers could have been historically used as appositive markers, as exemplified in (37). Likewise, the demonstrative zhè could have had the same function, as illustrated in (38). This commonality reflects the historical interrelationship between classifiers and demonstratives.

17




(37)

不得已央及你個漁父。 (元刊雜劇 朱太守風雪漁樵記) Bù-déyǐ yāngjí nǐ gè yúfū. no-alternative beg you CL fisherman “I have no alternative but to beg a fisherman like you.” (Yuan Kan Za Ju, Zhu Tai Shou Feng Xue Yu Qiao Ji, AD 1300)

(38)

他們這幾個是新來的。 (現代漢語八百詞) Tāmen zhè jǐ gè shì xīnláide. they this several CL be new comer “They are all newcomers.” (Xian Dai Han Yu Ba Bai Ci, Lü 1999: 656)

Tā juàn shū is the Pinyin form of the corresponding Chinese characters rather than the actual phonological form of the dialect.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

From Classifier to Demonstrative

435

(e)

“MOD + Dem + N” construction. Cross-linguistically, demonstratives occupy the determiner [D] position, namely the outermost layer of NPs with multiple attributives. This means that “Dem + NP” cannot be preceded by any other modifiers; for instance, in English, this beautiful girl is grammatical, but *beautiful that girl is ungrammatical. The same is true for all other former demonstratives in the history of the Chinese language, such as cǐ and bǐ. However, “zhè + N” phrases can also be preceded by other modifiers, as illustrated in (39):

(39)

但不知拿獲這人是邬泽不是? (三俠五義八十五回) Dàn bù zhī náhuò zhè rén shì Wū Zé bù shì? but not know capture this person be Wu Ze not be “But I don’t know whether the captured person is Wu Ze or not.” (San Xia Wu Yi, Chapter 85, AD 1850)

In the above discussion, the last three features, namely (c), (d), and (e), share the schema “MOD + DEM + N,” an intrinsic property that reveals the source of the demonstratives. As I suggested above, the demonstratives of Contemporary Chinese grammaticalized from classifiers, specifically “CL + N” in subject position. At that time, the canonical structure for classifiers was always “Num + CL + N,” with classifiers often being preceded by a numeral. Actually, numerals in NPs can be regarded as a type of modification; thus we have the following parallel structures: (40)

(a) MOD + CL + N (b) MOD + DEM + N

It is common for the function of a grammatical morpheme to be influenced by its lexical source and the context in which it grammaticalizes (Hopper and Traugott 2003: 98). For the sake of simplicity, I have not discussed the distal demonstratives and the demonstrative gè in numerous dialects, but they possess exactly the same functions as the demonstrative zhè, all of which distinguish them from the old demonstratives. These facts further strengthen the hypothesis that the demonstratives of Contemporary Chinese grammaticalized from classifiers.

17.4.5 Cross-linguistic Perspectives The present analysis is the first to identify the specific grammaticalization process from classifier to demonstrative in human languages. I believe that the general typological properties of classifiers help readers understand them better. In the literature, five pragmatic functions of classifiers have been identified: (a) deixis, (b) anaphora,

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

Demonstratives from Classifiers

436

(c) individualization, (d) substitution, and (e) indefiniteness. Here, the first four features are also cross-linguistically exhibited by demonstratives. Having investigated hundreds of classifier languages worldwide, Allan (1977) concluded that in all numeral classifier languages, the classifiers are always multifunctional and can occur in anaphoric, deictic, and quantitative expressions. Classifiers also indicate that the noun that they classify must be understood as having a non-generic reference; in other words, classifiers individuate nouns in classifier languages (Lucy 2000: 334). Additionally, numerals may be used anaphorically (Aikhenvald 2000: 98). In addition to their function in numeral noun phrases, classifiers in various languages function as nominal substitutes (Adams and Collins 1979: 2). As Bisang (1999: 21) also noted, in Mandarin Chinese and its many dialects, “CL + N” phrases in postverbal position refer to indefiniteness. I found the following perfect example from Late Medieval Chinese that can illustrate all five features of classifiers listed above: (41)




故歲雕梁燕, 雙去今來只。 (喬知之 定情篇) Gù suì diāo liáng yàn, shuāng qù jīn lái zhī. old year carve beam swallow pair leave now come CL “A couple of swallows lived on the carved beam of our house last year. Both of them left (last year), but only one of them comes back now.” (Qiao Zhi Zhi, Ding Qing Pian, AD 700)

The classifier zhī in (41) is the source of the demonstrative zhè, as I discussed above. Both the classifiers, shuāng “pair” and zhī “single,” refer to the couple of swallows that had lived in the beam of the building rather than having generic meanings. These classifiers occur in the second clause to anaphorically refer to the swallows mentioned earlier, in the first clause. Both classifiers function to individualize the swallows and to replace whole phrases: yī shuāng yàn “one pair swallow” and yī zhī yàn “one CL swallow.” Note that the classifier shuāng “pair” in subject position could be interpreted only as definite, meaning “this couple,” but the classifier zhī in object position can be understood as indefinite, meaning “one of the couple.” Typologically, demonstratives have the following semantic features and pragmatic functions: quantification, anaphora, substitution (pronoun), individualization (genitive), and definite marking (determiner) (Diessel 1999a: 54‒55). In addition, the semantics and syntax of the new demonstratives are discussed in this section. Therefore we can use Table 17.7 to represent the functional features of classifiers and demonstratives. The only difference between classifiers and demonstratives is [±definite], as classifiers are [−definite], but demonstratives are [+demonstrative]. Thus it should not be a surprise that classifiers in certain languages change into a demonstrative-like element.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

The “Semantic Primitive” Hypothesis

437

Table 17.7 The functional features of classifiers and demonstratives Category

Number

Anaphoric

Substitution

Individualization

Definite

Classifier Demonstrative

+ +

+ +

+ +

+ +

− +

Indeed, deictic classifiers exist in North American, South American, and African languages, and in Eskimo, that are considered “demonstrative” or “article classifiers” (Aikhenvald 2000: 176‒183). In addition, in Assamese, classifiers mark definiteness– indefiniteness (Barz and Diller 1985), while in Vietnamese, classifiers signal the definiteness and referentiality of the noun (Simpson et al. 2011, Simpson and Biswas 2016, Simpson 2017). 17.5 The “Semantic Primitive” Hypothesis This chapter is the first to identify the specific grammaticalization process from classifier to demonstrative by providing detailed motivations and mechanisms for development among all the languages in the world. In the literature, demonstratives are considered “diachronically basic,” “semantic primitives,” or “zero-stage” morphemes, and it is generally believed that they cannot be derived from other lexical sources. Although it has been indicated that demonstratives might originate from such verbs as “go” and locative adverbs, it is still not clear whether their changes belong to a grammaticalization event or are merely a case of semantic extension because no specific processes have been identified yet. Thus the present analysis is significant both empirically and theoretically. In the very beginning of the written history of Chinese (around the thirteenth century BC), there were already several demonstratives with sources that remain unknown. In Medieval Chinese, the emergence of the new demonstratives, which grammaticalized from classifiers, was the outcome of interaction among the following three changes: (a) the emergence of the classifier system, (b) the establishment of the disposal construction, and (c) the ability of preverbal position to mark the definiteness of bare nouns and “CL + N” phrases. In addition, the syntax of the old demonstratives and the structure of nominal phrases provided the background for the change. There are hundreds of classifier languages in the world, but the realization of the potential development from classifiers to demonstratives is conditioned by various factors. Further study in the future may reveal that Chinese is not the only language in the world to take this grammaticalization path.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

438

Demonstratives from Classifiers

The emergence of the classifier system fundamentally changed the texture of Chinese grammar. Classifiers themselves belong to a major grammatical category that largely influences the construction of numeral phrases, and they also serve as sources for demonstratives, associative morphemes, genitive morphemes, object markers, and even modal markers. In many southern dialects, the demonstrative gè developed into an associative morpheme, which is equivalent to the associative morpheme de in many northern dialects. This associative morpheme de in Mandarin Chinese and many other northern dialects grammaticalized from the demonstrative dǐ in Late Medieval Chinese, a change that might be analogous to the emergence of classifiers. All these issues are worthy of further study.

https://doi.org/10.1017/9781108921831.017 Published online by Cambridge University Press

18 Distal Demonstratives from Phonological Derivation 18.1 Introduction In Chapter 17, we demonstrated that the proximal demonstratives grammaticalized from a general classifier in Late Medieval Chinese. A question arises: how did its counterpart, the distal demonstrative, develop? This is an ideal window for examining a special type of mechanism for the development of a grammatical form: a phonological derivation from its semantic opposite rather than an independent grammaticalization process from a lexical source. It is very common that grammatical morphemes develop out of lexical sources, such as proximal demonstratives; however, it is extremely unusual that a grammatical morpheme or a functional word is phonologically derived from its counterpart, such as the mechanism of the emergence of distal demonstratives. Considering the empirical and theoretical importance, therefore, we explore the developments of proximal and distal demonstratives in two chapters. Every language has at least two demonstratives, proximal and distal, and some languages also have a medial demonstrative or other additional demonstratives. Demonstratives themselves possess many grammatical and pragmatic functions, and they also serve as major sources for various important grammatical devices, such as copulas, relativizers, definite articles, and complementizers. However, lexical sources for demonstratives remain unknown, as do the mechanisms for the process of their emergence. Based on a database of more than 1,000 subdialects of Chinese, this chapter demonstrates that the distal demonstratives in these subdialects are phonologically derived from the corresponding proximal demonstratives that grammaticalized from classifiers in Late Medieval Chinese. This finding identifies a new type of mechanism for the emergence of grammatical items: within a pair of two closely related grammatical elements, the basic and unmarked element originates from a lexical source and gives rise to the other through certain phonological principles. That is, both the grammaticalization process and phonological derivation are responsible for the emergence of grammatical forms.

439

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

440

Distal Demonstratives from Phonological Derivation

Demonstratives are regarded as “conventional signs” and are by nature symbolic, which is the opposite of onomatopoeia or sound symbolism (Langendonck 2007). However, they are among the very few grammatical items that exhibit a nonarbitrary relationship between phonological form and meaning, similar to onomatopoeias and sound symbolism. Examining twenty-six geographically unrelated languages, for instance, Woodworth (1991) found that there is a strong correlation in terms of proximal and distal demonstratives: proximal demonstratives tend to use the high and front vowel [i], and distal demonstratives use mostly the low and back vowel [a]. Although they share semantic and syntactic commonalities, proximal and distal demonstratives have completely different mechanisms of emergence. In Chinese, proximal demonstratives were first typically grammaticalized from classifiers, and they then gave rise to the corresponding distal demonstratives by means of certain phonological principles involving an iconicity between speech sounds and distance. This was a new type of mechanism for the emergence of grammatical items that stands in contrast to the grammaticalization process and can explain the nonarbitrary relationship between the phonological forms and meanings of the two demonstratives in numerous languages. Compared to other functional categories, demonstratives display certain peculiar properties. Proximal, distal, and medial demonstratives must be phonologically spelled out;1 in other words, they cannot employ the marked and unmarked contrast apparent in singular versus plural forms, such as one book and two books. It is difficult to imagine that the proximal and distal demonstratives in a given language could have developed from two independent lexical sources at exactly the same time because daily communication requires both to be ready for use. Therefore both the grammaticalization process and phonological derivation are necessary for the emergence of grammatical items such as demonstratives. To understand this point of view, let us review different grammatical categories and their marked and unmarked forms. Type 1. Grammatical categories that are isolated and independent, such as the future marker be going to in English and the disposal morpheme bǎ in Chinese. Their emergence does not involve any other grammatical changes. They are generally grammaticalized from lexical sources. Type 2. Grammatical categories with two contrastive grammatical items. For instance, in almost all languages with a singular–plural category, plural nouns are marked but the singular nouns are unmarked, e.g. one book versus two books in English (Corbett 2004: 19). Thus only the marked side has to be grammaticalized from a lexical source, such as the plural marker.

1

Approximately 31 percent of languages in the world have a medial demonstrative (Diessel 2006).

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Introduction

441

Type 3. Grammatical categories with two or more loosely related grammatical items. For instance, the English definite article the originated from the demonstrative that, and its indefinite article a/an originated from the number one (Hopper and Traugott 2003: 119). However, they are not semantically and syntactically parallel to each other; specifically, the can modify singular and plural nouns, but a/an can only precede a singular noun. In addition, they do not have to occur in a language; for instance, many languages have only a definite article (Diessel 2006). For this type of functional domain, the two grammatical items might develop from more than one lexical domain, as the English case shows. Type 4. Grammatical categories with two intrinsically contrastive grammatical items. A typical example is gender. A language cannot have only a feminine gender without a masculine one. Since genders derive from the sex distinction in human beings and animals, they could have two lexical sources. For instance, the feminine gender may originate from the word “women,” as in the Kilivila vivina “woman” > na (Senft 1996: 22), and the masculine gender may originate from the word “man,” as in the Zande ko “man” > kɔ́ (Heine and Reh 1984: 223, Claudi 1985). However, one language may acquire one of its two gender markers through phonological derivation only. In Somali, for example, gender is marked on nouns by means of tonal morphemes. Masculine forms exhibit a high tone on the penultimate vocalic mora, while feminine forms exhibit a high tone on the final vocal mora, e.g. ínan “boy” and inán “girl” (Hyman and Comrie 1981, Saeed 1987). At first glance, demonstratives should belong to the type to which gender belongs, but demonstratives and gender are in essence different. Only a limited number of languages have markers for gender, but all languages must have demonstratives. In addition, feminine and masculine genders mirror concrete things in reality (i.e. sex), but demonstratives do not have to do so. Consequently, unlike gender markers, proximal and distal demonstratives in a language are unlikely to develop from two independent lexical sources. In this case, it is unlikely, for instance, that each of the proximal and distal demonstratives has its own independent sources via separate processes of grammaticalization. The Chinese case shows that the classifiers first developed into proximal demonstratives and drove the emergence of their corresponding distal demonstratives via a rigid phonological rule. Having examined a sample of twenty-six languages, Woodworth (1991) concluded that, cross-linguistically, there is an element of sound symbolism distinguishing proximal and distal demonstratives: the proximal tends to use the high and front vowel [i], but the distal favors the low and back vowel [a]. Following Woodworth, Traunmüller (1996) investigated a wider range of languages and found that 71 percent of languages fall into the pattern shown here, as illustrated in Table 18.1.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

442

Distal Demonstratives from Phonological Derivation Table 18.1 Phonological correlations between proximal and distal demonstratives Language

Proximal

Distal

Function

English French Spanish Guugu Ambula Lezgian Yimas Vietnamese

this celui-ci este yii kéni i k đây

that celui-là aquel nhaa wani a m đãy

singular singular, masculine pronoun pronoun determiner root adverbial adverbial

Traunmüller correctly noted that if there is any sound symbolism in proximal and distal demonstratives, the relationship between sound and meaning would be rather abstract since there are no sounds to imitate in the first place. As we will see in the subsequent section, among vowels, [a] has the highest sonority and [i] has the lowest. This chapter will demonstrate that the sonority of sounds is a key factor in the derivational process from proximal to distal demonstratives in Chinese. There are hundreds of phonological forms for each of the demonstratives, but they originated from two major lexical sources. These numerous phonological forms were generated by a phonological principle. In standard Mandarin, for instance, the distal demonstrative nà [na51] is phonologically derived from the proximal demonstrative zhè [tʂɤ51], and the latter grammaticalized from a general classifier. The focus of the present analysis is on the phonological derivation from proximal to distal demonstratives.

18.2 Dialect Variants of Demonstratives This section provides some background knowledge that is necessary to understand the present analysis, including (a) coverage of the database for the present analysis and (b) the lexical sources of demonstratives. The grammaticalization process from classifier to demonstrative has been discussed in Chapter 17; here, I simply point out what the lexical sources for demonstratives are.

18.2.1 Databases It is difficult to identify how many dialects Chinese has because they are enormously different in phonology, lexicon, and grammar. Even within a dialect family, such as the Wu dialect, people from different areas often cannot communicate by means of their

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Dialect Variants of Demonstratives

443

own dialects. Consequently, there are different views about the division of Chinese dialects (Li 2001: 29).2 For the purpose of the present study, I adopt the so-called eight major dialects as follows: the northern dialect,3 the Jin dialect, the Xiang dialect, the Gan dialect, the Wu dialect, the Min dialect,4 the Hakka dialect, and the Yue dialect.5 In this chapter, I reserve the term “dialect” for each of the eight dialect families and use “subdialect” for smaller dialect areas.6 In reality, the Chinese language possesses more than 1,000 subdialects with remarkable differences in phonology, lexicon, and grammar. This chapter makes no further distinction among them and simply uses the term “subdialect” for them. To guarantee the accuracy and complete coverage of the present study, I established a database on the basis of the following investigations. (a)

Cao (2008). A well-balanced investigation of all eight dialects covering 930 subdialects. (b) Huang (1996). A collection of papers and monographs covering thirtyseven subdialects of the eight dialects. (c) Wu (2000). An investigation of fifty-three subdialects spoken in Hunan Province, belonging mainly to the Xiang dialect but including the Hakka dialect and the northern dialect. (d) Qiao (2000). An investigation of twenty-one subdialects of the Jin dialect. (e) Beida (1995).7 An investigation of twenty representative subdialects of all eight dialects. (f) Li (2001) and Yuan (2001). Two general surveys of Chinese dialects covering all eight dialects. The present analysis is based on the demonstratives of more than 1,000 subdialects, a comprehensive coverage of Chinese dialects. Although there are some overlaps among the above works, the details of their descriptions vary greatly; hence they are all valuable for my study. Cao (2008) and Beida (1995) are dictionaries that provide only partial forms of demonstratives.

2 3 4 5 6

7

The “Hui dialect” is classified in the northern dialect and the “Ping dialect” in the Yue dialect. This dialect includes Mandarin Chinese, standard Mandarin, and Pekingese. This dialect includes Northern Min, Eastern Min, and Southern Min (including Taiwanese). The alternative name is Cantonese. Note that many studies are based on administrative areas such as provinces. Keep in mind that dialects and administrative areas do not neatly match each other. For instance, there are at least three dialects in Hunan Province, the Xiang dialect, the Hakka dialect, and the northern dialect, although a majority of the people in this province speak the Xiang dialect. This is an outcome of a collective work edited by the Department of Chinese Language and Literature at Peking University (see the References for details).

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

444

Distal Demonstratives from Phonological Derivation 18.2.2 Lexical Sources for Demonstratives

A replacement of systems of demonstratives occurred in Late Medieval Chinese, from the seventh century AD to the tenth. This period gave birth to all the demonstratives that are still used in Contemporary Chinese; that is, zhè, gè, and dǐ.8 The origin of these demonstratives has long been debated in the field of Chinese linguistics, especially the origins of the proximal and distal demonstratives zhè and nà (for details, see Lü 1985: 183, Ohta 1987: 115, Wang 1989: 183, Jiang and Cao 2005: 31).9 Here, I mention only the most convincing analyses instead of going into detail about this issue. Jiang (1999a) postulated that the demonstrative dǐ originated from a locative word indicating “bottom,” and I would like to leave this issue open. Zhang (2001) employed solid phonological evidence to argue that the demonstrative zhè in the northern dialects and gè in many southern dialects originated from two general classifiers in Medieval Chinese. I carried out another study to prove Zhang’s proposal from semantic and syntactic perspectives, in which I identified the specific grammaticalization processes. It is less debatable that the demonstrative gè grammaticalized from its classifier usage because they have the same phonological form in many dialects (for details, see Qian 1997, 2014, Lin 2018), but it is problematic to claim that the demonstrative zhè was also derived from the classifier zhī, because their phonological forms are different in standard Mandarin. However, this issue was successfully resolved by Zhang (2001), who provided the following two pieces of phonological evidence. (a) In Qie Yun (a phonological book composed around the sixth century AD) and Piao Tong Shi Yan Jie (a textbook to teach Koreans Chinese composed around the thirteenth century AD), the demonstrative zhè and the classifier zhī have exactly the same phonological form. (b) In at least nine contemporary subdialects, the demonstrative zhè and the classifier zhī have the same phonological form. For instance, in the Fuqing subdialect (belonging to the Southern Min dialect), both are tsia21.10 According to Cao (2008: 10), 131 subdialects have the same phonological forms for the proximal demonstrative zhè and the general classifier zhī. The classifier gè grammaticalized from the noun “one bamboo” and the classifier zhī grammaticalized from the noun “one bird” in Medieval Chinese, and both were among the few earliest classifiers (Wang 1989: 18‒40). The earliest examples of the demonstratives gè, zhè (zhī), and dǐ are illustrated in the following (Lü 1985: 183):

8 9 10

When the Pinyin forms are given, the symbols for the four tones are added. We will discuss the distal demonstrative nà later. The two numbers at the top right of a syllable represent the tone contours, and the single number represents the tonal type of the syllable.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Dialect Variants of Demonstratives (1)

445




個賊住! (太平廣記 佥載) Gè zéi zhù! CL theft stop “This theft, please stop stealing!” (Tai Ping Guang Ji, Xie Zai, AD 1000)




(2)

只個逍遙是謫仙。 (李咸用 題王處士山居) Zhī-gè xiāoyáo shì zhé-xiān. this-CL happy be exiled-fairy Zhī yán zhī-le jìn bēishāng. “This happiness is an exiled fairy.” (Li Xian Yong, Ti Wang Chu Shi Shan Ju, AD 850)

(3)

竹籬茅舍, 底是藏春處。 (驀山溪詞) Zhú-lí máo-shě, dǐ shì cáng chūn chù. Bamboo-fence thatched-cottage, this be store spring place “The bamboo fences and thatched cottages, these are the places that store the spring.” (Mo Shan Xi Ci, AD 1000)

Due to diachronic change and cross-dialect variations, every demonstrative above has many different phonological forms. The consonants in the onset are the most important clues to associate those demonstratives with the same origin. Table 18.2 shows the diachronic changes and cross-dialect variations of the three demonstratives with different origins. The onsets of these three demonstratives are [k], [tʂ] or [t], and their dialect variations are listed as follows. (a) (b) (c)

If the onset consonants of the demonstratives are [tʂ], [ts], [tɕ] or [tʃ], they originated from the general classifier zhī. If the onset consonants of the demonstratives are either [k] or [ɡ], they originated from the general classifier gè. If the onset consonants of the demonstratives are either [t] or [d], they originated from the locative word dǐ.

For the 930 subdialects in Cao (2008: 10), the distribution of the three demonstratives with different origins is shown in Table 18.3. The number of subdialects in Table 18.3 does not reflect the size of the populations that speak dialects. For instance, those who speak the northern dialect account for approximately 70 percent of the entire population, and the speakers of the remaining seven dialects account for merely 30 percent of the population. In addition, the distinction between the investigated subdialects of the southern dialects, such as the Wu dialect

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Distal Demonstratives from Phonological Derivation

446

Table 18.2 Changes and dialectal variations of demonstratives Time/dialect Medieval Chinese Modern Chinese Contemporary Chinese Wu dialect Xiang dialect Gan dialect Hakka dialect Cantonese Eastern Min Southern Min

Variations of zhī (zhè) 4

Variations of gè 3

kɑ kɔ4 kɤ4 kəu4 kɑ4 ko4 kɛ4 kɔ3 kɔ4 ko4

tɕǐɛk tʂi3 tʂʅ1 tsɤh6 tsɿ6 tsak6 tsak5 tʃɛk8 tsieh6 tsik6

Variation of dǐ tiei2 ti3 ti3 ti3 ti3 ti3 ti3 tɐi3 ti3 te3

Note: The reconstruction of the phonological forms is based on Zhou and Li (2001). The single number at the top right of the phonetic form indicates the tone category rather than tonal contour.

Table 18.3 The dialect distribution of demonstratives Originated from zhī

Originated from gè

Originated from dǐ

Unidentified source11

418 subdialects

273 subdialects

60 dialects

179 subdialects

and the Xiang dialect, is much finer than that in the northern dialects, probably because of different complexities in the relationship between them. In the other 179 subdialects, most of the phonological forms are derived from the same principle, as I will discuss in subsequent sections.

18.3 Correlations between Proximal and Distal Demonstratives A vast body of empirical evidence shows that there must be some kind of phonological derivation between proximal and distal demonstratives. Unlike other grammatical markers, the phonological forms of demonstratives in all Chinese dialects are highly variable, and many of them have two or more forms, from which we can clearly see the phonological correlations between proximal and distal demonstratives. The correlations fall into the following five types. Type I. Proximal and distal demonstratives have the same phonological forms. Subdialects of this type are rare, and only the Suzhou subdialect is reported to belong 11

The proximal of this group of dialects may originate from the three demonstratives, but due to diachronic changes, it is difficult to identify which.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Correlations between Proximal and Distal Demonstratives

447

to this type in the literature. In the Suzhou subdialect, both proximal and distal demonstratives share exactly the same form, including onset, nucleus, coda, and tone. Yuan (2001: 97) indicated that gɤʔ23 in this subdialect is a neutral demonstrative and that kᴇ44 can be either a proximal or a distal demonstrative. In other cases, however, a glide [u] is added to the distal use, producing the form [kuᴇ]. This case shows that the proximal and distal demonstratives share a root and were later differentiated phonologically. Type II. Within a subdialect, the proximal and distal demonstratives share the same syllable but are differentiated by tone. In Hakka, for instance, the proximal and distal demonstratives have the same syllabic form and are distinguished by tone. The proximal demonstrative is ke31, and the distal demonstrative is ke52 (Yuan 2001: 171). As in the Suzhou subdialect, once again, the proximal and distal demonstratives likely originated from a single source and were differentiated by tone at a later time. Type III. The rhymes and tones are the same but the consonants in the onset positions are shifted to express proximal and distal demonstratives. According to Yuan (2001: 268), the Southern Min dialect has seven phonological forms for each proximal and distal demonstrative that either express different meanings or occur in different contexts as independent pronouns or locatives or with “Dem + CL” classifiers, plurals, or temporals. With the rhyme and the tones held constant, [ts] in the onset position refers to a proximal demonstrative and [h] in the onset position refers to a distal demonstrative, as shown in Table 18.4. Yuan (2001) claimed that the Min dialect actually employs an inflection strategy to distinguish proximal from distal demonstratives. Type IV. In standard Mandarin, including Pekingese and other northern subdialects, the proximal and distal demonstratives have different onsets and rhymes, sharing only the same tone. Type V. In some subdialects, the proximal and distal demonstratives share no phonological commonality, differing in onset, nucleus, coda, and tone. In Table 18.6, the two subdialects, Hanshou and Lixian, belong to the Xiang dialect. Table 18.4 Demonstratives in the Southern Min dialect Dialect

Proximal

Distal

Function

Southern Min dialect

tsi51 tse51 tsit32 tsia24 tsia51 tsit32

hi51 he51 hit32 hia24 hia51 hit32

Independent use Independent use Preceding classifier e34 Locative Preceding plural e34 Preceding temporal tsun33

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Distal Demonstratives from Phonological Derivation

448

Table 18.5 Demonstratives in standard Mandarin Dialect

Proximal

Standard Mandarin

51

Distal

Function

51

na na51

tʂɤ tʂei51

Pronoun Fusion with a classifier

Table 18.6 Demonstratives in the Hanshou and Lixian dialects Dialect Hanshou Lixian

Proximal 24

tiɛ tiɛ13

Distal no55 la33

For the first three types listed above, the phonological correlations between proximal and distal demonstratives are so close that they must have originated from the same sources and been differentiated by a certain phonological rule. As we will see below, even the last two types of proximal and distal demonstratives are derived from a phonological principle. Superficially, the overall picture of cross-dialect demonstratives in Chinese is extremely complex. In terms of phonological forms, there are at least seventy-six proximal demonstratives and ninety-seven distal demonstratives, according to Cao (2008: 10‒11). Moreover, within a single subdialect, a demonstrative, either proximal or distal, can have up to six different phonological forms,12 a phenomenon that has not been found in any other grammatical category in Chinese. In what follows, I will demonstrate that all these phenomena are related to the derivational rule between contrastive demonstratives and are highly uniform and regular from this perspective.

18.4 Derivation from Proximal to Distal Demonstratives This section is the core of the present analysis, addressing the following eight issues: (a) the asymmetry between proximal and distal demonstratives, (b) phonological forms of twin syllables and cognate words in Chinese, (c) the phonological rule of the derivation from proximal to distal demonstratives, (d) the sonorous asymmetry between the segments of proximal and distal demonstratives, (e) the reversal of the derivational process between proximal and distal demonstratives, (f) the glottal consonants at the

12

For example, the distal demonstrative of the Shanyin subdialect, which belongs to the Jin family, has the following six phonologically derived forms: nã213, nəʔ45, nɛe35, nəu213, nəu53 and na213.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

449

Derivation from Proximal to Distal Demonstratives Table 18.7 The proximal zhè and the distal nà in vernacular texts Text

Era

Size*

Proximal zhè

Distal nà

Zhuzi Yulei Yuan Zaju Hong Lou Meng Dangdai Xiangsheng

12th century AD 13th century AD 18th century AD contemporary

1,975,900 365,700 860,700 478,300

6,381 1,361 7,748 6,919

2,926 1,015 4,876 3,019

* The size of the texts refers to the number of Chinese characters.

onset of the distal demonstrative, (g) the different specifications of the onset and nucleus, and (h) the cliticization of demonstratives.

18.4.1 The Asymmetry between Proximal and Distal Demonstratives Within a grammatical category such as singular versus plural, as mentioned above, the singular side is more basic and remains unmarked and the plural side is less basic and marked. For demonstratives, the proximal ones are more basic and unmarked (default), and the distal ones are derivative and marked. This distinction is confirmed by human cognition and the frequency of occurrence. Proximal demonstratives refer to a shorter distance between the speaker and the referent, and distal demonstratives refer to a longer distance between the speaker and the referent. The order of human cognition is typically from self-center to a farther location. Langacker (1991: 242‒246) related the proximal versus distal contrast in demonstratives to the present–past distinction, where proximal corresponds to present and distal corresponds to past. Likewise, Ariel (1990: 130) argued that proximal demonstratives are less marked than distal demonstratives. These proposals are supported by the different frequencies of the proximal zhè and distal nà throughout history. Clearly, the frequency of the proximal zhè is always greater than that of the distal nà, which shows that the former is more basic than the latter. There are some exceptions to the derivational direction between proximal and distal demonstratives. Of the 930 subdialects in Cao (2008: 10), in thirty-three of them (accounting for 3.5 percent of the total) the demonstratives that grammaticalized from the classifier gè were first realized as distal demonstratives, and then generated the corresponding proximal demonstratives through a phonological principle, as illustrated in Table 18.8. This abnormal realization caused interesting derivations, which will be discussed below. Some subdialects even have more than one proximal demonstrative, with each originating from a different source. For instance, the Shanghai subdialect has two proximal demonstratives, diəʔ12 and geʔ12; the former is the traditional demonstrative,

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

450

Distal Demonstratives from Phonological Derivation Table 18.8 Distal demonstratives from the classifier gè in subdialects Subdialect

Proximal

Distal

Guangzhou Guangning Yongxin Yunan

li/lia ni/ne kə/ki i/ia

kɔ/ko/ku kɔ/ko/ku kɔ/ko/ku kɔ/ko/ku

and the latter is the newly emerging demonstrative (Huang 1996: 565‒505). Obviously, the proximal diəʔ12 grammaticalized from the locative word dǐ, and the proximal geʔ12 grammaticalized from the classifier gè. Thus far, I have mainly discussed proximal demonstratives without touching on the corresponding distal demonstratives. As discussed previously, it has been convincingly argued that two general classifiers plus one possible locative word first developed into proximal demonstratives, but no sources for the distal demonstratives have been found.13 The focus of this chapter is on where the corresponding distal demonstratives come from, and how, which will be discussed in depth in the following section.

18.4.2 Phonological Correlation of Cognate Words Chinese is a tonal language, and tonal contours function to distinguish meanings. Additionally, Chinese is basically a monosyllabic language, and approximately 90 percent of its morphemes, defined as the minimal pairing of meaning and form (usually smaller than or equal to words), are monosyllabic. In all dialects, basic demonstratives in Chinese are monosyllabic, regardless of whether they are proximal or distal. Chinese does not allow any consonant clusters; hence only a single consonant can occur at the onset or coda position. The glides, [j], [w], or [y], may occur at the initial position of the rhyme. The maximal syllable of Chinese is represented in Figure 18.1.14 When a consonant is absent from the onset position, the glides (or semivowels) are raised to occupy the onset position. In this case, the glides may function to distinguish medial from distal demonstratives, as we will see in the next section. In terms of auditory phonetics, the central concept “sonority” in this analysis can be defined from two closely related perspectives: energy and openness. First, it refers to the carrying power of a sound; the more sonorous a sound, the more articulatory effort is 13 14

I will return to this issue in a subsequent section. For possible structures of Chinese, please see Duanmu (2007: 71‒84) for details. The abbreviations used in this chapter are C ~ coda, G ~ glide, O ~ onset, R ~ rhyme, S ~ syllable, T ~ tone, and V ~ vowel.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives

451

51 Figure 18.1 The structure of Chinese syllables

required to produce it (Clark and Yallop 1992: 61). Second, it refers to the relative openness of the vocal tract, which directly determines the relative loudness of a sound. The more open the mouth is, the more sonorous the sound; e.g. among vowels, the low vowels have the highest degree of sonority and the high vowels have the lowest degree of sonority (Zsiga 2013: 333‒334). Syllables are structured in accordance with certain principles in relation to the sonority of sounds. It has been known for over a century that the construction of syllables is guided by the following principle (Kenstowicz 1994: 254): Sonority Sequencing Principle: onsets must rise in sonority toward the nucleus and codas must fall in sonority from the nucleus. Surprisingly, cross-linguistic regularity in constructing a syllable is widely found in the phonological derivation between proximal and distal demonstratives in many Chinese dialects. In other words, the sonority sequencing principle works in a linear fashion to construct a syllable across languages, but it behaves in a nonlinear fashion to derive distal from proximal demonstratives in Chinese. Overall, the sonority of vowels is higher than that of consonants, but different vowels differ in sonority. According to Kiparsky (1979), Kenstowizc (1994), and Gnanadesikan (1997), the sonority scales of vowels are defined as follows: (5)

Low vowels [a, ɑ, ɒ, æ] > Mid vowels [ɛ, ɔ, e, ɤ, o] > High vowels [ɯ, u, i, ʅ, ɿ] > Schwa [ǝ]

Note that among all vowels, the schwa [ǝ] has the least sonority and the low vowel [a] has the highest sonority; the rest are between these extremes. The sonority scales of sounds in the onset position are as follows (Selkirk 1984: 112, Durand 1990, Hulst 1984, Parker 2002):

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Distal Demonstratives from Phonological Derivation

452

(6) Glides [j, w] > Liquids [l] > Nasals [m, n, ŋ] > Fricative [s, f] > Voiced affricatives [v] > Affricatives [tʂ, ts] > stops [t, p] When no consonant occurs in the onset position, the high vowels [i] and [u], originally the medials of the rhyme, automatically take the onset position and are pronounced as glides (semivowels) [j] and [w]). The sonority sequence principle operates within a single syllable in every language. For example, English has single-syllable words such as plain [plein] and plant [plænt], but the syllables cannot be structured as plani or plnat. In Chinese, however, this phonological principle is found not only inside a syllable but also within a group of semantically related words in a cross-syllabic fashion. As mentioned earlier, Chinese is basically a monosyllabic language, which means that a syllable usually stands for a meaningful morpheme that corresponds to a Chinese character. Due to the historical disyllabification tendency, nearly 80 percent of words in Contemporary Chinese are dissyllabic, and their phonological forms are determined by the two meaningful morphemes. For instance, in tɕiao51 ʂʅ55 (teach + master) “teacher,” the two syllables have no phonological constraints between them. Many nouns become disyllabic through the addition of nominal suffixes such as -zi, -er, or -tou, for which the syllables are also independent. However, there is always a portion of disyllabic words for which the two syllables are phonologically constrained. In other words, neither of the two syllables conveys any semantic content; thus they cannot be separated and used independently. These words are called “twin-syllable” words, as illustrated in (8) (Shi 1995). (7)

(a) Nouns: hu55 tiɛ55 “butterfly”; ha35 ma “frog”; tʂaŋ55 laŋ35 “cockroach.” (b) Verbs: ku213 tiao “tinker”; pa55 la “remove”; hu55 you “trick.” (c) Adjectives: phiao51 liaŋ “pretty”; tɕi55 liŋ “clever”; mәŋ35 loŋ35 “dim.” (d) Onomatopoeia: phu55 toŋ55 “pit-a-pat”; phiŋ55 phaŋ55 “rattle”; huŋ55 loŋ55 “rumble.”

Between the two syllables of a twin-syllable word, even the tones are closely related to each other. For instance, almost 100 percent of Chinese onomatopoeias have a high tone, e.g. phu55 toŋ55, phiŋ55 phaŋ55, and huŋ55 loŋ55. When we understand the sonority scales of sounds and the SSP, we can immediately recognize that the construction of twin-syllable words in Chinese obeys a principle similar to what is found cross-linguistically within a single syllable. In the following formula, S stands for syllable, O for onset, N for nucleus, and C for coda; the number at the bottom right of the letters represents whether they belong to either the first syllable or the second syllable. The function fs is defined as the variable of sonority values. (8)

fs (S2) ≥ fs (S1); fs (O2) ≥ fs (O1); fs (N2) ≥ fs (N1); fs (C2) ≥ fs (C1)

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives

453

For example, according to the definition of sonority degree, in the onomatopoeia huŋ55 loŋ55, the onset of the second syllable [l] is more sonorous than that of the first syllable [h], the nucleus of the second syllable [o] is more sonorous than that of the first syllable [u], and the codas of the two syllables are equally sonorous. Before using the sonority principle to explain the derivation between proximal and distal demonstratives, I should emphasize a language-specific method of how Chinese invents phonological forms for words. Phonemes within a syllable are linearly adjacent to each other, so it is not difficult to understand that certain phonological constraints exist between them. However, both proximal and distal demonstratives are independent of each other and rarely occur adjacently. How can they have a phonological correlation? The answer lies in the methodology of the Chinese people for inventing phonological forms for words. In his pioneering study, Wang (1978) discovered that a pair of antonymous words or a group of semantically related words can be assigned a new phonological form by changing the tones, consonants at onset, or vowels in the nucleus of certain etymological roots. There are thousands of words with phonological forms that have been created in this way (for details, see Wang 1997), as illustrated in (10): (9)

(a) Tone shifts, e.g. mai213 “buy”; mai51 “sell.” (b) Consonant and tone shifts, e.g. tʂaŋ35 “long”; tʂaŋ213 “grow.” (c) Nucleus and tone shifts, e.g. pei55 “low-rank”; pi51 “servant girl.” (d) Consonant, nucleus, and tone shifts, e.g. ʂʅ35 “eat”; sɿ51 “feed.”

In contrast, the English verbs buy and sell have no phonological similarities, but their corresponding verbs in Chinese share the same syllable and are differentiated only by tone. The special method for inventing phonological forms in Chinese is crucial for understanding the phonological derivation between proximal and distal demonstratives. In short, phonological correlations and constraints work in a cross-syllabic fashion in Chinese, including in the invention of phonological formations for words that are semantically related. Since proximal and distal demonstratives are deictically contrastive, the phonological correlations between them are perfectly natural in Chinese.

18.4.3 Derivation from Proximal to Distal Demonstratives The phonological rule governing the derivation between proximal and distal is similar to the SSP for constructing syllables. Proximal and distal demonstratives are semantically contrastive and thus can be viewed as pairs of antonymous words. As noted previously, classifiers and locative words were the first to be grammaticalized into proximal demonstratives.15 Instead of recruiting another lexical source, all dialects employ 15

In nearly 4 percent of subdialects, they are first realized as distal demonstratives instead of proximal demonstratives. This issue will be discussed later.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Distal Demonstratives from Phonological Derivation

454

a phonological rule to derive the corresponding distal demonstratives. Specifically, the sonority values of sounds at the onset, nucleus, and coda are remarkably contrastive between the phonological forms of proximal and distal demonstratives, which can be formulated as follows. In the following formulas, fs is defined as a function of sonority scale; Od, Nd, and Cd represent the onset, nucleus, and coda of distal demonstratives, respectively; and Op, Np, and Cp represent the onset, nucleus, and coda of proximal demonstratives, respectively: (10)

(a) fs (Od) ≥ fs (Op) (b) fs (Nd) ≥ fs (Np) (c) fs (Cd) ≥ fs (Cp)

These three formulas do not operate in the same fashion in differentiating proximal and distal demonstratives. Another phonological principle for constructing a syllable, namely the sonority dispersion principle, states that every language prefers to maximize the sonority slope from onset to nucleus and to minimize this value from nucleus to coda (Clements 1990). Similarly, the sonority slopes between the two onsets of proximal and distal demonstratives tend to be maximized and are rarely equal to each other, whereas the sonority slopes between the two nuclei and codas of the two demonstratives tend to be minimized and are often equal to each other. According to the database consisting of more than 1,000 subdialects, 97 percent of subdialects can be explained by means of formula (11). First, let us use the demonstratives of standard Mandarin (Mandarin) to test the principle: (11)

Standard Mandarin: proximal demonstrative tʂɤ51 distal demonstrative na51

According to the sonority scale, [n] ≥ [tʂ]; [ɑ] ≥ [ɤ], and both share the falling tone. As mentioned previously, this pair of demonstratives in standard Mandarin emerged in Late Medieval Chinese. In the literature, the source of the distal demonstrative nà remains unknown, though there has been some speculation about its origin. For instance, Ohta (1987) and Wang (1989) speculated that nà might have originated from one of the two demonstratives ruò or ěr in Old Chinese. This view is problematic for three reasons. First, these two demonstratives ceased to be used around the first century AD, and nà did not arise until the eighth century AD. When nà emerged, the dominant distal demonstrative was bǐ, for which the corresponding proximal demonstrative was cǐ. Second, their phonological forms were remarkably different, as shown in Table 18.9.16 Clearly, it is unlikely that the two Old Chinese demonstratives developed into nà according to the evolution of the phonological system. 16

The reconstruction of historical sounds is based on Li and Zhou (1999).

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives

455

Table 18.9 The historical phonological forms of distal demonstratives

ruò ěr nà

Medieval Chinese

Modern Chinese

Contemporary Chinese

ɽǐak ɽǐe2 na3

ɽiɔ ɽï3 na4

ʐuo4 ɚ3 na4

4

4

Note: The single numbers on the right top of syllables indicate tonal types rather than tonal contours.

Third, and more importantly, their functions differ remarkably. For instance, ruò and ér always behaved like adjectival demonstratives; in other words, they could never be used alone without noun heads. However, the syntactic use of nà is entirely parallel to that of its corresponding proximal zhè; for instance, they are often combined with a classifier as a substitute for the entire NP, omitting the noun heads. Additionally, both zhè and nà can be used as genitive markers, preceded by a possessor, as illustrated in (12) and (13).17 (12)

你知道我這病。 (紅樓夢十一回) Nǐ zhīdào wǒ zhè bìng. you know I this disease “You know my disease.” (Hong Lou Meng, Chapter 11, AD 1750)

(13)

誰喜歡喫你那糕? (紅樓夢六十回) Shuí xīhǎn chī nǐ nà gāo? who like eat you that cake “Who likes to eat your cake?” (Hong Lou Meng, Chapter 60, AD 1750)

Clearly, the distal nà could not have developed out of any other distal demonstrative in Old Chinese. In fact, this distal demonstrative is derived from the proximal demonstrative zhè via a phonological rule, as discussed above. The phonological forms of demonstratives in standard Mandarin are based on Pekingese. Indeed, there are many phonological variants of the distal demonstratives in the northern dialects that are used by approximately 70 percent of the population. The proximal demonstratives that originated from the classifier zhī are realized as [tʂɤ], [tsɤ], [tʂʅ], [tsɿ], [tʂei], [ʦei], [ʦit], and [tɕiaʔ] in different subdialects. For the sake of simplicity, Table 18.10 shows only the distal demonstratives for which the corresponding proximal demonstratives are [tʂɤ].

17

For possible structures of Chinese, please see Duanmu (2007: 71–84) for details. The abbreviations used in this chapter are C ~ coda, G ~ glide, O ~ onset, R ~ rhyme, S ~ syllable, T ~ tone, and V ~ vowel.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Distal Demonstratives from Phonological Derivation

456

Table 18.10 The onsets of distal demonstratives related to the proximal demonstrative [tʂɤ] Examples Onset of distaldemonstrative

Number of subdialects

Subdialect

Proximal

Distal

[n] [l] [m] [u] [ŋ] [h]** [k] [a]

148 31 3 7 6 5 3 2

Libo Jishou Yingcheng Jiujiang Yuexi* Lufeng Jingzhou Huaning

tʂɤ/tsɤ tʂɤ/ʦɤ tʂɤ/tsɤ tʂɤ/ʦɤ tʂɤ/ʦɤ tʂɤ/ʦɤ tʂɤ/ʦɤ tʂɤ/tsɤ

na/ne/nɤ la/le mo/mə/mei ua/uei n/ŋ hi/hia/hiɔ kɔ/ko/ku a/e/ei

* Distal n/ŋ in Yuexi dialect results from cliticization, a point that I will return to in the subsequent section. **In Cao (2008: 11), this [h] is transcribed as [x], but in Yuan (2001), it is transcribed as [h].

There are twenty-one different phonological forms of the distal demonstratives that correspond to the proximal demonstratives [tʂɤ]. Even within a given subdialect, there are typically two or more phonological forms for distal demonstratives. These facts show that it is unlikely that these distal demonstratives originated from any lexical sources. Of the 205 subdialects in Table 18.10, only five seem to be exceptions to the derivational rule; i.e. the onset consonants are [k] and zero. In other words, with respect to the emergence of their distal demonstratives, the phonological rule accounts for 97 percent of subdialects. In phonological theory, [k] is regarded as the least sonorous and is lower in sonority than the affricative [tʂ]. There is one possible explanation for this. The phonological property of the distal and proximal demonstratives suggests that this group of subdialects might have developed the classifier gè in their distal demonstratives and the classifier zhī in their proximal demonstratives. Indeed, some subdialects obtained their distal demonstratives from the classifier gè, which in turn caused the reversal of the derivation process between the two demonstratives. The Huanning subdialect chose the low vowel [a] with the highest sonority as its distal demonstrative, which shows that it is not a real exception because there is still a remarkable increase in the sonority of the nucleus.

18.4.4 Sonorous Correlation of Proximal and Distal Demonstratives According to the derivational rule in (11), the sonority scale of the onsets and nucleus of distal demonstratives must be either higher than or equal to the corresponding segments of proximal demonstratives.18 However, this requirement varies from one segment to another, as stated in the following. 18

It seems that the segments in the coda position do not play a role in distinguishing proximal from distal demonstratives.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives (a) (b)

457

For the segments at the onset, the sonority of distal demonstratives is usually higher than but rarely equal to that of proximal demonstratives. For the segments in the nucleus or coda, it is equally possible that the sonority of distal demonstratives is either higher than or equal to that of proximal demonstratives. This pattern differs across dialects.

The above asymmetry is manifested by the sonority degree principle, namely that a syllable prefers to maximize the sonority slope from onset to nucleus and to minimize this value from nucleus to coda (Clements 1990), as mentioned earlier. This principle can be revised to maximize the sonority slope from the onset of the proximal demonstratives to that of the distal demonstratives and to minimize this value from the nucleus/ coda of proximal demonstratives to those of distal demonstratives. Within the 320 subdialects of the northern dialect,19 311 subdialects (97 percent) have increased sonority from the onsets of proximal demonstratives to those of distal demonstratives. Since the proximal demonstratives mostly originated from the classifier zhī, the onsets of these subdialects are [tʂ] or [ts]. The onsets of the distal demonstratives are mostly [n] and sometimes [m], [l] or [w]. In the nine subdialects exhibiting the equal sonority of the two onsets, the consonants of the onsets of their proximal demonstratives are either nasals or liquids, which have the highest sonority of all consonants. Thus it is impossible to increase the sonority from the first onset to the second. The Baoji subdialect deserves special attention. It is the only subdialect in which the onset of the proximal demonstratives is [k]. This reveals that it originated from the classifier gè, which is typically found in the southern dialects, making it the only exception in the northern dialect. Within this group of 320 subdialects, fifty-nine exhibit a sonority increase from the first nucleus to the second. The vowel of the first nucleus is either [i] or one of its two variants, [ɿ] and [ʅ].

Table 18.11 Patterns of the two onsets with sonority increase

19

Pattern

Proximal

Distal

Subdialect, province

tʂ → n tʂ/ts → l tʂ/ts → w tʂ/ts → m

tʂɤ/tʂei tʂɤ/ʦɤ tʂʅ/ʦɿ tʂɤ/tsɤ

na/nei la/le wa/wei mo/mə/mei

Pekingese Yanzhou, Shandong Longde, Ningxia Yingcheng, Hubei

This statistical survey is based on Cao (2008: 11).

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

458

Distal Demonstratives from Phonological Derivation Table 18.12 Patterns of the two onsets with equal sonority Pattern

Proximal

Distal

Subdialect, province

n→n l→l k→k

ni/ne li/lia kə/ki

na/ne/nɤ la/le kuɛ

Chongqing Enshi, Hubei Baoji, Shanxi

Table 18.13 Patterns of the two nuclei with sonority increase Pattern

Proximal

Distal

Subdialect, province

high vowels → middle/low vowels

tʂʅ/ʦɿ tʂʅ/ʦɿ tʂʅ/ʦɿ

na/ne/nɤ la/le a/e/ei

Wenshan, Yunnan Anlong, Guizhou Dafang,* Huizhou

* In the Dafang subdialect, the onset of the proximal demonstrative is empty due to cliticization, which I will discuss later.

Table 18.14 Patterns of the two nuclei with equal sonority Pattern

Proximal

Distal

Subdialect, province

əʔ → əʔ ei → ei ɤ→ɤ i→i

tʂəʔ/ʦeʔ tʂei/ʦei tʂɤ/ʦɤ ʦi/ʦia

naʔ/nəʔ nai/nei na/ne/nɤ ni/niɛ

Taipusi, Neimenggu Haerbin, Heilongjiang Yangyuan, Hebei Wuzhong, Ningxia

In the remaining 261 subdialects, the distal demonstratives always have an alternative form with the same vowel as that of the corresponding proximal demonstratives. This means that the sonority scales of the two nuclei can be equal. Standard Mandarin has two pairs of phonological forms for its proximal and distal demonstratives: [tʂɤ]/[na] and [tʂei]/[nei]. The first pair has a sonority increase between the two nuclei, but the second pair exhibits equal sonority between the two nuclei. As noted in Section 18.3, in the Min dialect, with the rhymes (i.e. nucleus and coda) remaining the same, the shift between [ts] and [h] refers to the proximal and distal demonstratives. Similarly, in the Jin dialect, some subdialects can have as many as six phonological forms for each of the proximal and distal demonstratives, and the rhymes remain unchanged; the different consonants at the onset thus distinguish the different demonstratives. In short, in distinguishing proximal from distal demonstratives, the sonority scales are maximized from the first onset to the second onset and minimized from the first nucleus to the second nucleus.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives

459

Table 18.15 Demonstratives in the Jin dialect with different onsets Proximal 213

Distal 45

213

53

tʂã tʂəʔ tʂəu tʂəu tʂǝʔ32 tʂar54 tʂǝu54 tiᴀ53 tiɛ33 ti11 tiə11 tə̃ r33 tɐr33

213

Subdialect 45

213

53

nã nəʔ nəu nəu nǝʔ32 nar54 nǝu54 niᴀ53 niɛ33 ni11 niə11 nə̃ r33 nɐr33

Shanyin Datong Jincheng

18.4.5 Reverse Derivation Of the 930 subdialects in Cao (2008: 11), approximately thirty (accounting for 3 percent of the total) first developed the classifiers into distal demonstratives and then derived the corresponding proximal demonstratives by means of the same phonological rule. In other words, the derivational process proceeded in reverse, from distal to proximal demonstratives, and exhibits some very interesting properties. The reverse derivation is found mainly in Cantonese (Yue dialect). A successful explanation for this phenomenon will further prove the hypothesis that a phonological rule guides the derivation between distal and proximal demonstratives. First, let us focus on the consonants at the onset. The classifier gè has the onset consonant [k], which has the least sonority among all sounds. Once it is realized as the distal demonstrative in a particular subdialect, there are only two strategies left to satisfy the formula fs (Od) ≥ fs (Op): (a) find another plosive consonant for the onset of the proximal demonstrative, e.g. [t] in Liannan and Fogang (in this case, the formula becomes fs (Od) = fs (Op)), and (b) empty the onset of the proximal demonstratives, as in Conghua and Longmen (in this case, the formula becomes fs (Od) > fs (Op)). However, quite a few subdialects, such as Hong Kong, Zhongshan, and Boluo, violate the formula by reversing the sonority increase from the onset of the distal demonstratives to that of the proximal demonstratives. An appropriate explanation is that these subdialects have different rankings of the rules: the optimal rule is that there must be a sonority increase between the onsets of proximal and distal demonstratives, regardless of the direction. As a result, nasals and liquids are used as the onsets of proximal demonstratives in Hong Kong, Zhongshan, and Boluo. In the Yue dialect, the phonological form of the classifier gè is [kɔ], with a middle vowel as its nucleus. According to the sonority hierarchy, [ɔ] is more sonorous than the high vowels [i] and [ǝ]. Thus, to satisfy the formula fs (Nd) ≥ fs (Np), it is possible to increase the sonority scale from the nucleus of proximal demonstratives to that of distal demonstratives. All the vowels of the proximal demonstratives are either [i] or [ǝ], as shown in Table 18.16, which satisfies the phonological principle.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

460

Distal Demonstratives from Phonological Derivation Table 18.16 The classifier gè realized as a distal demonstrative in the Yue dialect Proximal

Distal

Subdialects

ni/ne li/lia li/lia ti tə/tei i/ia i/ia

kɔ/ko/ku kɔ/ko/ku kai/kei keʔ/kəʔ kɔ/ko/ku kɔ/ko/ku kai/kei

Hong Kong Zhongshan Boluo Liannan Fogang Conghua Longmen

Table 18.17 Reverse derivation between proximal and distal demonstratives in Cantonese and Hakka Subdialect

Proximal

Distal

Dianbai Gaozhou Lianzhou Lianping Meizhou

kə/ki kə/ki kə/ki kə/ki kə/ki

kɔ/ko/ku kɔ/ko/ku koŋ/gã kai/kei ka/ke/kə

The above generalization perfectly explains the following group of subdialects. The demonstrative gè is realized as both proximal and distal, differentiated by the sonority of the vowels. The nuclei of the proximal demonstratives are either [i] or [ǝ], which are the least sonorous vowels. Furthermore, in the Xiang family, there are three ways to derive proximal demonstratives where the classifier gè is first realized as a distal demonstrative instead, as shown in Table 18.18. First, like the Hong Kong subdialect in the Yue family, the Pingjiang subdialect follows the optimal rule that there must be a sonority increase between the two onsets of proximal and distal demonstratives, regardless of direction. As a result, the liquid [l] occupies the onset position of the proximal demonstrative. Second, like the Liannan subdialect in the Yue family, the Yizhang subdialect finds another plosive consonant [t] for the onset of its proximal demonstrative; in this case, the formula becomes fs (Od) = fs (Op). Third, like the Conghua subdialect in the Yue family, the Miluo and Yanling subdialects empty the onset of proximal demonstratives; in this case, the formula becomes fs (Od) > fs (Op). In summary, the phonological principle works in reverse when the classifier gè is first realized as a distal demonstrative in the subdialects. Once again, this phenomenon

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives

461

Table 18.18 Reverse derivation in the Xiang dialect Subdialect Pingjiang Yizhang Miluo Yanling

Proximal 35

li ti21 i24 i31

Distal ko35 kai41 ko24 kai31

reveals that the derivation between proximal and distal demonstratives is consistently governed by the phonological rule.

18.4.6 The Glottal Sounds of Distal Demonstratives The cross-linguistic commonality of glottal sounds further proves that the phonological principle guides the derivation from proximal to distal demonstratives in Chinese. In the literature, there are two opposite views on the sonority scales of glottal consonants. One view is that glottal consonants are highly sonorous (Pike 1954, Chomsky and Halle 1968: 301, Levin 1985, Parker 2002). In contrast, other researchers have argued that glottal consonants have the lowest sonority, basically the same sonority as obstruents (Heffner 1950, Lass 1976, Zec 1988, Lombardi 1999). In addition, Clements (1990: 322) claimed that glottal consonants behave arbitrarily in terms of how they class with other sounds, effectively having no sonority value. Empirical evidence shows that in sonority distance restrictions, glottal consonants usually behave like highly sonorous elements. For example, Gujarati allows only glides, liquids, and [h] as the second members of onset clusters: [kjal] “opinion,” [krupa] “kindness,” [kleʃ] “fatigue,” and [khǝrǝc] “cost” (Cardona 1965: 31). In contrast, it is rare for [h] to have the same distribution as other fricatives, as is evident, for example, in a comparison of the English [slɪt] “slit,” [flɪt] “flit,” and *[hlɪt]. Likewise, in Chinese dialects, the two glottal consonants [ʔ] and [h] act like nasals and liquids that are highly sonorous. In the Jin dialect, for instance, only [ʔ], [n], [m], and [ŋ] can occur in the coda position. In the derivation process from proximal to distal demonstratives, the glottal [h] behaves like a nasal or liquid in the Min dialect, as discussed below. The consonants in the onset position of distal demonstratives are mainly nasals ([n], [m], and [ŋ]) and the liquid [l]. In Cao (2008: 11), however, 111 subdialects use the uvular [x] as the onset for their distal demonstratives. In this work, no onsets are transcribed as glottal [h]. Both [x] and [h] are fricatives. Distal demonstratives of this type are found mainly in the subdialects of the Min family. However, according to Beida (1995) and Yuan (2001), in the phonological system of the Xiamen subdialect (i.e. the

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

462

Distal Demonstratives from Phonological Derivation Table 18.19 The [h] and [x] onsets of distal demonstratives Subdialect Wenzhou Nanchang Xiamen Chaozhou Fuzhou

Dialect family Wu Gan Min Min Min

Proximal 213

ki kɔ213 ʦit32 ʦi53 ʦi21

Distal he45 hɛ213 hit32 hɯ53 xi21

representative area of the Min dialect), only [h] exists, and [x] is absent. Among the twenty representative subdialects in Beida (1995), there are four with [h] as the onset of the distal demonstrative and only one with [x] as the onset. As Table 18.19 shows, [h] is used as the onset of the distal demonstrative four times more frequently than [x]. In addition, the sound [x] of demonstratives is recorded as [h] in Yuan (2001). Thus I believe that the transcriptions of Cao (2008) are inaccurate and that most instances of [x] at onset are actually the glottal [h]. Logically, a majority of the 111 subdialects should have [h] onsets for their distal demonstratives. In other words, in the Min dialect, [h] functions like a nasal or liquid to increase the sonority scale from the onset of the proximal demonstratives to that of the distal demonstratives. This can explain why the device of shifting between [ts] and [h] is used to distinguish proximal from distal demonstratives in the Min dialect, which Yuan (2000) regarded as a kind of inflection.

18.4.7 Specifications of Onset and Nucleus A remarkable feature of demonstratives is that their phonological forms are variable in almost all Chinese dialects. In Cao (2008: 11), only one of the 930 subdialects has a single form for its proximal and distal demonstratives. The rest have two or more phonological forms for each of their demonstratives. This feature is not found with any other grammatical morpheme. For instance, the associative marker de, the most frequently used morpheme, has only a single phonological form within a particular subdialect, e.g. [tǝ] in standard Mandarin, [kɤʔ] in Suzhou, and [kɔ] in Yangjiang (see Beida 1995 for details). In standard Mandarin, the proximal demonstrative zhè has two phonological forms, [tʂɤ51] and [tʂei51],20 and the distal demonstrative also has two corresponding forms, [nɤ51] and [nei51]. The situation in other dialects is much more complex. Some subdialects even have as many as six different forms for each of their proximal and distal

20

The phonological forms [tʂei] and [nei] are regarded as a fusion of the demonstrative and yi “one” (Zhu 1982).

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives

463

demonstratives. This means that a single subdialect can have more than ten distinctive forms for its demonstratives, as shown in Table 18.20 (Qiao 2000). Clearly, the consonants in the onset position are fixed for proximal and distal demonstratives, but the rhymes are variable. In this case, there is a rigorous correspondence between the rhymes of proximal and distal demonstratives, including medial demonstratives if there are any. In many other subdialects, the consonants in the onset position are the same, and in these cases the vowels in the nucleus must be distinctive in distinguishing proximal from distal demonstratives; typically, the sonority increases from the proximal nucleus to the distal nucleus. If the onsets remain the same, the vowels must be different. In both cases, the consonants in the onset position are specified.

Table 18.20 Phonological variations of demonstratives in subdialects Subdialect Jincheng

Shouyang

Proximal 53

tiᴀ tiɛ33 ti22 tiəʔ22 tə̃ r33 tɐr33 tsɔ21 ʦə̃ 21 ʦəʔ22 ʦei21 æ22

Distal niᴀ53 niɛ33 ni22 niəʔ22 nə̃ r33 nɐr33 uɔ21 uə̃ 21 uəʔ22 uei21 æ22

Table 18.21 Different nuclei distinguishing proximal from distal demonstratives Subdialect

Proximal

Distal

Longyan

xi xia

Maoming

kə ki

Nanhai

li lia

xɛn xaŋ xɔ̃ kɔ ko ku lɔ lu ləu

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Distal Demonstratives from Phonological Derivation

464

Table 18.22 Cliticization of demonstratives Subdialect

Proximal

Distal

Yunan Heshan Loudi Liuyang Xinhua Hengyang Chaling Wuyishan Jianshui Huangmei Yiwu Linyi

a a n i/ia i/ia kɔ/ko/ku ɤ/ɯ/u i/ia tʂʅ/ʦɿ tə/tei n tʂuo/tʃou

u/ɔ a/e/ei n/ŋ kɔ/ko/ku n/ŋ a/e/ei mɛ̃ 52 u/ɔ a/e/ei i/iɛ dɔŋ u\ɔ

18.4.8 Cliticization In some dialects, only one phoneme is used to express demonstratives. These phonemes, typically [a], [i], [u], [n], and [ŋ], all have high sonority, conveying the potential to be syllabic, as shown in Table 18.22. Since the focus of this chapter is on the phonological derivation relation between proximal and distal demonstratives, I have not discussed their syntax. The derivation process may be obscured by phonological reduction from cliticization or fusion with other adjacent elements, which is related to syntax. In this section, I will address the phonological reduction of demonstratives and several special cases for which the derivation processes cannot be identified at present. In many dialects, demonstratives cannot immediately precede a noun and must be connected by a classifier (for details, see Huang 1996: 128‒172). To replace an entire NP, a demonstrative must combine with a classifier. In addition, demonstratives are often combined with locative words to indicate places or adverbial suffixes to modify a VP. These factors may cause demonstratives to cliticize or fuse to the following elements. Their high frequency of occurrence with classifiers may trigger the fusion of demonstratives with the following classifiers or locatives. Ariel (1990: 59) developed an accessibility-marking scale, a cline with the tendency of phonological reduction: (14)

Distal demonstrative > proximal demonstrative > clitic > zero

It is cross-linguistically true that pronominal elements often cliticize to an element in their environment. Likewise, in other languages, demonstratives are phonologically reduced to a single phoneme; for instance, Lezgian has two

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Derivation from Proximal to Distal Demonstratives

465

demonstrative roots, i “proximal” and a “distal,” which must be combined with other elements to be used (Haspelmath 1993: 259). Since they are frequently combined with other elements, demonstratives are subject to cliticization, which in turn causes their phonological reduction. This reduction may occur in the onset or nucleus. In the Liuyang subdialect (Li 2001), for instance, the distal demonstrative n cannot be used alone and must co-occur with the classifier ke, as illustrated in (16): (15) N44-ke ʂʅ11 thua44 lie tsia24-tsi ke. that-CL be he Gen sister PRT “That is his sister’s.” (Liuyang subdialect) When used as an adverb to modify APs or VPs, the distal demonstratives in Liuyang have two forms – koŋ and n – indicating that the reduced form n should be phonologically reduced from koŋ. Likewise, in the Jin dialect, the distal demonstrative u is a reduced form and is usually combined with other elements. For instance, u213.ɛ in the Yangqu subdialect is a fusion of a distal demonstrative and a classifier. The full form in the Yuanping subdialect, its neighboring subdialect, is uai54 (Qiao 2000: 126‒127). Of the twenty-four subdialects of the Jin family (Qiao 2000), six subdialects, Pinglu, Wanrong, Linyi, Xiangfen, Huozhou, and Pinyao, use [u] or [ɔ] to encode distal demonstratives only. Within this dialect family, many subdialects use u alone as the onset of their distal demonstratives. There is robust evidence that this u is a reduced form, as shown in Table 18.23. Here, let us focus on the single phoneme [u] in the Yuanping subdialect in Table 18.23, which refers to a distal demonstrative. There is a neat correspondence between each form of the proximal and distal demonstratives in the two subdialects. With the same rhymes, the shift between [tʂ] and [u] at the onset distinguishes proximal from distal demonstratives. Logically, the distal [u] in the Yuanping subdialect is actually a reduction of [ui].21 Of the ninety-three subdialects of the Xiang family, thirteen use one of the following phonemes to encode their proximal demonstratives. (16)

(a) [i]: Xiangyang, Yanling, Miluo, Pingjiang, Liuyang, Xinhua. (b) [a]: Ningyuan, Jiangyong, Guzhang, Fenghuang, Mayang. (c) [n]: Loudi. (d) [ɤ]: Chaling.

21

The pronunciation of [i] is [ʅ] when following the affricative [tʂ].

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Distal Demonstratives from Phonological Derivation

466

These single-phoneme demonstratives also result from phonological reduction due to cliticization. Within the same dialect family, these vowels often occur in the nucleus of demonstratives, as shown in Table 18.24. Among the 930 subdialects of Cao (2008: 11), there are three exceptions to the phonological rule. All of them have [pi44] as their distal demonstratives, where the onset is [p], the consonants with the least sonority, and the nucleus is [i], the vowels with the least sonority. Their corresponding sounds in proximal syllables are more sonorous. These distal demonstratives have a phonological form similar to that of bǐ [pi213], which emerged in Old Chinese and was replaced around the tenth century AD. It is probable that these three subdialects still retain a relic of the old distal demonstrative. It seems that the phonological rule worked even in Old Chinese. The old pair of demonstratives cǐ [tsi213] and bǐ [pi213] is reconstructed as [tshǐe2] and [pǐai2] (Li and Zhou 1999), with an increase in sonority from the first nucleus [i] to the second [a].

18.5 Iconicity between Acoustic Property and Distance The phonological forms of demonstratives represent a special type of iconicity between the phonological forms and meanings of demonstratives. Iconicity is regarded as being Table 18.23 Use of [u] as a distal demonstrative in the Jin dialect family

proximal demonstratives distal demonstratives

Linfen subdialect

Yuanping subdialect

tʂei tʂǝŋ tʂaŋ uei uǝŋ uaŋ

tʂæɛ tʂʅ tʂər uæɛ u uǝr

Table 18.24 Nasals and vowels encoding demonstratives in the Xiang family Subdialect Guiyang Guzhang Shuangpai

Proximal 41

ki ai33 ʦɤ25

Distal kai41 oŋ33 la25

Table 18.25 Three exceptions to the phonological derivational rule Subdialect

Province

Dialect family

Proximal

Distal

Lanshan Xinfeng Tongshan

Hunan Jiangxi Hubei

Xiang Gan northern dialect

ɔi/uai ɡia kʰɑ

pi44 pi44 pi44

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Iconicity between Acoustic Property and Distance

467

in contrast to arbitrariness and is the opposite of symbolic, according to Peirce (1974). Traditionally, the study of iconicity has been restricted mainly to onomatopoeia and sound symbolism, marginal phenomena in the lexicon of a language (Langendonck 2007). In the past two or three decades, interest has extended to the correspondence between linguistic forms and the contents that these forms express, such as adjacency and isomorphism. Haiman (1985: 196) maintained that iconicity should be sought in the system of grammatical rules for combining words to express complex concepts. It is generally accepted that there is iconicity if something in the form of a sign reflects something in the world, normally through a mental operation. This means that something in the form of a linguistic sign reflects, through its meaning, something in its referent (Mayerthaler 1980, 1988). In this broad sense, demonstratives can be regarded as a kind of iconicity. The phonology and semantics of demonstratives are a puzzle in linguistic theory. Langendonck (2007) claimed that demonstratives such as this and that are “conventional signs” and are by nature symbolic, which is the opposite of onomatopoeia or sound symbolism. As mentioned earlier, however, Woodworth (1991), having examined twenty-six geographically unrelated languages, found a strong correlation between proximal and distal demonstratives: proximal demonstratives tend to use [i], the least sonorous vowel, and distal demonstratives mostly use [a], the most sonorous vowel. She concluded that this establishes a basis for a relation between the vowel sonority of proximal and distal forms and their meaning, cross-linguistically. In line with Woodworth, Traunmüller (1996) investigated a wider range of languages and found that 71 percent fall into the pattern shown here. Nevertheless, Traunmüller correctly noted that if there were any sound symbolism in proximal and distal demonstratives, the relationship between sound and meaning would be relatively abstract, since there are no sounds to imitate in reality. Therefore an unusual phenomenon arises: on the one hand, demonstratives are semantically abstract, like grammatical morphemes; on the other hand, they reflect a correlation between phonological forms and meanings, as onomatopoeia or sound symbolism does. How to explain this peculiar phenomenon is a challenge. Thus far, all the types of iconicity that have been identified in the literature are projections from reality into language, but demonstratives are involved in the reverse of that projection, i.e. from language to reality, specifically through human cognition regarding the acoustic property of sounds. Let us refer to the definition of “sonority” in Crystal (2008: 442): A term in auditory phonetics for the overall loudness of a sound relative to others of the same pitch, stress and duration. Sounds are said to have an “inherent sonority,” which accounts for the impression of a sound’s “carrying further,” e.g. [s] carries further than [ʃ], [a] further than[i].

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

468

Distal Demonstratives from Phonological Derivation

Figure 18.2 Sonority degrees of proximal and distal demonstratives

Thus the most important phonological devices distinguishing proximal and distal demonstratives, namely the sonority scales of sounds, actually reflect an iconicity between the loudness of sounds and the distance that a demonstrative refers to, which is summarized in Figure 18.2. The above principle of iconicity explains why there is a nonarbitrary relationship between the phonological form and meaning of demonstratives in many world languages. English still employs a similar phonological strategy to distinguish its proximal from distal demonstratives: this [ðis] and that [ðæt] share the same consonants at the onset, but the former uses the high vowel [i] as the nucleus, whereas the latter uses a mid vowel [æ]. As noted earlier, the sonority of [æ] is higher than that of [i]. As mentioned earlier, the same device is used by 71 percent of languages to distinguish proximal from distal demonstratives (Traunmüller 1996). Thus the present analysis is applicable to the derivational relationship between proximal and distal demonstratives in many other languages. In short, the formula in (11) actually reflects an iconicity between the acoustic properties of sounds and the referents of demonstratives. The phonological rule that has been identified in this study can explain nearly 99 percent of the phonological forms of proximal and distal demonstratives within a database consisting of more than 1,000 subdialects that comprehensively covers all the major dialect families. The present analysis not only resolves the greatest puzzle regarding the origins of demonstratives but also identifies a new type of mechanism for the emergence of grammatical morphemes. Both the grammaticalization process and phonological derivation are responsible for producing new grammatical items. Our phonological rule can also explain why there are an enormous number of phonological forms for demonstratives in Chinese dialects. The formula “fs (Odistal) ≥ fs (Oproximal)” requires that the sonority of the consonant at the onset of distal demonstratives be equal to or greater than that of the corresponding proximal demonstrative demonstratives. The onsets of the three lexical sources are [k], [tʂ], and [d]; thus there are many options for consonants at the onset of distal demonstratives. As a result, nasals, liquids, glottal consonants, and glides account for approximately 75 percent of the onsets of the distal demonstratives in the more than 1,000 subdialects. Within a subdialect, the onset must be specified by a fixed consonant, but the nucleus may have a range of vowels, resulting in numerous forms of demonstrative. In all subdialects,

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

Iconicity between Acoustic Property and Distance

469

each set of proximal and distal demonstratives has two to six distinctive phonological forms. The typological study of demonstratives shows that there are clear phonological correlations between proximal and distal demonstratives across languages. This suggests that there may also be certain phonological rules that are responsible for the derivation between them. Future study in this direction will definitely provide new insight into the mechanism of the design of human languages.

https://doi.org/10.1017/9781108921831.018 Published online by Cambridge University Press

19 Pronouns, Plurals, and Diminutives 19.1 Introduction Number is one of the most underestimated grammatical categories in Chinese linguistics, but it is mainly characteristic of the grammar of Contemporary Chinese. Today we have stable grammatical devices to signal the number of entities in reality we are talking about, which Old Chinese simply lacked. One reason why this topic has long been neglected in studies on Chinese linguistics is probably that researchers are apt to take for granted the definition of the concept “number” in English or some of the other major Indo-European language, where all countable nouns are marked in two ways, singular and plural, regardless of whether they are definite. Actually, English reflects only one particular type of the wide range of possibilities in marking the number of nouns. Like many other languages in the Indo-European family, English adopts a two-way marking system, as in cat versus three cats,1 which is the most common division across languages. Note that dual (“two”), trial (“three”), and even paucal (“few”) are also found in some other language families. Furthermore, the languages that manifest the category of “number” fall into two groups: first, all countable nouns, whether definite or indefinite, are marked with the singular or plural morpheme, e.g. in English and many other Indo-European languages; second, only definite countable nouns are marked as singular or plural, but indefinite nouns are not marked for number, as in Aari, spoken by Omotic people in the region of Ethiopia, and Kambera, spoken by Austronesian people (Corbett 2004: 278). That the application of number markings is sensitive to definiteness has a cognitive foundation. When people are clear about what the referents are, they are usually aware of the basic quantity of the referents, either one or more. The relation between definiteness and number marking is critical for understanding how the number category works in Chinese grammar. Although the grammar of English is generally insensitive to definiteness in marking numbers, this feature is also indicative of its pronominal system. 1

The grammatical number category also exists in verbs, e.g. in English, he works versus they work. It is applicable only to nouns in Chinese.

470

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronominal Systems in Old Chinese

471

Since the inherent meanings of personal pronouns and demonstratives are definite, they all adopt peculiar lexical forms to distinguish singular and plural, e.g. I ~ we, he/she/it ~ they, this ~ these, that ~ those; but the interrogative pronouns have no such distinctions because they are inherently indefinite; for instance, who, what, and which can refer to either singular or plural. Only from the interplay between definiteness and number marking can we appreciate the importance of the “number” category in Contemporary Chinese, an important grammatical category that was introduced into the language after the eighth century AD.

19.2 Pronominal Systems in Old Chinese The emergence of the classifier system created a new grammatical category that did not exist in Old Chinese. It also caused a series of changes in nominal structure and triggered several significant grammaticalization processes, having largely altered the texture of Chinese grammar. One such change was the innovation of the number category in Chinese grammar. Chierchia (1998a, 1998b) claimed that classifiers are intrinsically incompatible with plural markers and thus tend not to coexist with them in a single language. However, the intrinsic reason for their incompatibility within a single language, whether it is a coincidence or there is some logical ground, is still not clear. It is safe to say that it was the establishment of the classifier system that enabled the emergence of the number category in Chinese grammar. Since pronominal words are most sensitive to the marking of singular/plural due to their definite referents, let us use them to illustrate this point. Contemporary Chinese has a neat pronominal system with regard to number, which started to enter the language after Late Medieval Chinese. In standard Mandarin, the plural -men can be attached only to pronouns and human nouns, such as xuéshēng-men “students,” but it can be applied to animal names in some dialects (e.g. the Shijiazhuang dialect) and even inanimate nouns in other dialects (e.g. the Guizhou dialect). Moreover, the plural marking can even be expressed by a consonant sound or just a tone value (for details, see Huang 1996: 432). These differences across dialects reflect the different rates of grammaticalization across

Table 19.1 Singular and plural pronouns in Contemporary Chinese

First-person pronoun Second-person pronoun Third-person pronoun Proximal demonstrative Distal demonstrative

Singular

Plural

我 wǒ 你 nǐ 他 tā 這 zhè 那 nà

我們 wǒ-men 你們 nǐ-men 他們 tā-men 這些 zhè-xie 那些 nà-xie

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronouns, Plurals, and Diminutives

472

dialects. By comparison, the plural marking of demonstratives is applicable to any nouns, regardless of whether they are animate or inanimate. In contrast, the personal nouns and demonstratives in Old Chinese did not differentiate between singular and plural, which means that all pronouns had the function of referring to both singular and plural, with number interpretations determined mostly by context (Wang 1989: 41), as illustrated below:




(1)

靜女其姝, 俟我於城隅。 (詩經 靜女) Jìng nǚ qí shū, sì wǒ yú chéng yú. quiet girl so beautiful wait I at city corner “The quiet and beautiful girl is waiting for me at the corner of the city.” (Shi Jing, Jing Nü, 1000–600 BC)

(2)

十年春, 齊師伐我。 (左傳 莊公十年) Shí nián chūn, Qí shī fá wǒ. ten year spring Qi troop attack we “In the spring of the tenth year, Qi troops attacked us.” (Zuo Zhuan, Zhuang Gong Shi Nian, 550–400 BC)




In the above example, the first-person pronoun wǒ is used as singular in (1) and as plural in (2), whereas the plural interpretation of the first-person pronoun in Contemporary Chinese must be suffixed by the plural marker -men. Wang (1989: 41‒64) indicated that there were some lexical words, such as chái “peer,” bèi “type,” děng “equivalent,” and shǔ “group,” that could be combined with pronouns to express the meaning “more than one,” but they were by nature lexical items with content meanings rather than grammatical devices. More importantly, these lexical items did not need to be used to indicate plurality; in reality, they were not used in most cases expressing plurality. As will be discussed below, in Contemporary Chinese, demonstratives cannot directly precede the head noun and the number feature must be signaled by certain grammatical devices. Specifically, in standard Mandarin, the singular is expressed by a simple classifier (implying “one”) or zero marking, and the plural must be indicated by the plural -xiē or a “Num + CL” phrase (when the number is two or more). In contrast, there was a set of demonstratives in Old Chinese, such as shì “this/these,” sī “this/these,” cǐ “this/these,” zī “this/these,” and bǐ “that/those,” all of which could directly modify nouns without having to indicate any information about the number feature, as illustrated in (3) and (4): (3)




之子於歸, 宜其室家。 (詩經 桃夭) Zhī zǐ yú guī, yí qí shìjiā. this girl in marriage suitable her family “This girl who was married was suitable for her family.” (Shi Jing, Tao Yao, 1000–600 BC)

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronominal Systems in Old Chinese (4)

473




之二蟲又何知? (莊子 逍遙遊) Zhī èr chóng yòu hé zhī? these two insect also what know “What did these two insects also know?” (Zhuang Zi, Xiao Yao You, 300 BC)

After classifiers were introduced into the language (for details, see Chapter 16), the functions of demonstratives changed dramatically. On the one hand, in Contemporary Chinese, the demonstratives and the head noun had to be linked by a classifier (usually together with a number) or a quantifier, such as the plural -xiē. On the other hand, the demonstrative plus the classifier can be used independently, without the head noun, substituting for the meaning of the whole noun phrase. In this case, the phrase “demonstrative + CL” behaves like an ordinary noun, for example, zhè-gè “this CL” is semantically equal to zhè-gè rén “this CL person” in certain contexts. As a result, phrases of this type can be used as the subject, the object, or the genitive, similar to an ordinary noun (Ohta 1987: 101). However, in Old Chinese, the demonstratives could not be used independently to function like an ordinary noun; they had only a deictic function of making proximal and distal references, which meant that they could be used mostly as nominal modifiers. As we saw in Chapter 17, the establishment of the classifier system caused the former set of demonstratives to be abandoned and triggered the general classifiers zhī in the northern dialects and gè in the southeastern dialects to develop into demonstratives. Despite the lack of the number category, the pronominal system of Old Chinese preserved a relatively complete case-marking system with a distinction between nominative, accusative, and genitive cases. Due to their high frequency, pronouns often lagged behind the development of the rest of the grammar and preserved some earlier features of the language, which is cross-linguistically true. In English, for instance, these three cases are still retained in pronouns, e.g. I ~ me ~ my, he ~ him ~ his, they ~ them ~ their. A similar situation existed in Old Chinese, although there were already some mixtures among these usages (for a fuller discussion, see Wang 1989: 41‒64). For the first-person pronouns, wú was used as the nominative and genitive pronoun, and wǒ could be used for all three cases. Wang (1989) found that when both wú and wǒ occurred within a single sentence, wǒ was used only as the accusative case, but wú was used as either the nominative or the genitive case, as illustrated in (5) and (6). This fact inspired him to hypothesize that wǒ should be the accusative case of the first-person pronoun. (5)




今者吾喪我。 (莊子 齊物論) Jīnzhě wú sàng wǒ. now I lose me “Now I lost myself.” (Zhuang Zi, Qi Wu Lun, 300 BC)

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronouns, Plurals, and Diminutives

474 (6)




吾君殺我而不辜。 (墨子 明鬼) Wú jūn shā wǒ ér bù gū. my lord kill me and not crime “It is not wrong that my lord would kill me.” (Mo Zi, Ming Gui, 350 BC)

In (5), wú is the subject and wǒ is the object; in (6) wú is the genitive case and wǒ is the nominative case. The case-marking contrast described above was highly regular in Old Chinese, and the division became blurred and eventually disappeared after the first century BC (Wang 1989: 45). For the genitive case, some features of Old Chinese deserve particular attention. The pronouns zhèn and nài could be used only as the genitive case of the first-person pronoun in the period from the thirteenth to the seventh century BC. Although there was a genitive morpheme, zhī, in Old Chinese that linked a possessive noun with its head, it could never occur between the pronoun and the head noun. In contrast, in Contemporary Chinese the genitive marker de is necessary to link the genitive with the head, such as wǒ de shū “I GEN book,” as used with ordinary nouns. In other words, today pronouns do not possess an inherent genitive feature. For the third-person pronoun, the case markings were quite neat: zhī was always used for the accusative case, as illustrated in (7), and qí was used for the genitive, as illustrated in (8): (7)




吾不欲觀之矣。 (論語 八佾) Wú bù yù guān zhī yǐ. I not want see it PRT “I don’t want to see it.” (Lun Yu, Ba Yi, 500 BC)

(8)




北冥有魚, 其名為鯤。 (莊子 逍遙遊) Běi míng yǒu yú, qí míng wèi Kūn. north sea have fish its name call Kun “There was a fish in the north sea whose name was Kun.” (Zhuang Zi, Xiao Yao You, 300 BC)

It was quite unusual that Old Chinese simply lacked a third-person pronoun for the nominative case, despite the existence of the accusative zhī. Thus, in the context in which the third participant had to be specified, the preceding noun had to be repeated, as illustrated in (9). The third-person pronouns tā “he/she/it” and tā-men, which are used in Contemporary Chinese, did not enter the language until the eighth century AD (Wang 1989: 41).

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

The Emergence of Plural Morphemes (9)

475




且私許復曹偉, 曹偉告絕於楚。 (左傳 僖公二十八年) Qiě sī-xǔ fù Cáo Wěi. Cáo Wěi gào-jué and secretly-promise return Cao Wei Cao Wei tell-disconnect yú Chǔ. with Chu “And he secretly allowed (him) to return to Cao and Wei. Cao and Wei announced they were no longer part of the country Chu.” (Zuo Zhuan, Xi Gong Er Shi Ba Nian, 550–400 BC)

The case-marking system of Old Chinese collapsed around the first century BC, which in turn caused some pronominal items to become out of date. In the meantime, the new system of pronouns, which conveyed the number of grammatical categories instead, gradually took shape and was finally established around the tenth century AD.

19.3 The Emergence of Plural Morphemes As mentioned previously, there are two plural morphemes in Contemporary Chinese with a restrictive division of labor: (a) -men for pronouns and human nouns and (b) -xie for demonstratives and the indefinite quantifier yī “one.” The earliest uses of these plurals were attested around the eighth century AD, but their widespread use occurred after the tenth century AD (Lü 1985: 54‒103, Wang 1989: 41‒65). The lexical source of the plural -xie is less controversial, as it likely came from its former use as an indefinite quantifier “some/several” (for details, see Lü 1985: 54, 365). Nonetheless, the lexical source of the plural -men was a major puzzle in historical Chinese linguistics because there were many written variants even in the earliest texts that actually hid the truth of its origin (Shi and Li 2001: 326‒339). In Contemporary Chinese, the tonal value of the plural -men has been neutralized and its vowel has become a schwa [ǝ], a phonological erosion due to grammaticalization. In the circle of historical Chinese linguistics, there have been many speculations about the origin of the plural -men. For example, Lü (1984: 1‒37) assumed that it came from the lexical item děng “equivalent,” Norman (1988: 120) argued that it might have resulted from the combination of the two words méi “every” and rén “person,” and Jiang (1995) proposed that it developed out of the word wù “thing.” All these hypotheses face difficulties in phonology, semantics, and syntax, and thus are not truly convincing. In our view, however, the plural -men simply developed out of its former use as a classifier for enumerating family or school, pronounced mén, with classifier use that was actually grammaticalized from the original noun meaning “door.” The strongest evidence for our analysis is that there was a classifier jiā (originally meaning “family”) in Late Medieval Chinese that was synonymous with the

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronouns, Plurals, and Diminutives

476

classifier mén: at that time, they were interchangeable and formed a compound jiā-mén “family.” The classifier jiā grammaticalized into a plural morpheme in many southern dialects such as the Hongdong and Linfen dialects (belonging to the Jin family; see Huang 1996: 432‒464 for details), and the Wujin, Yixing, Jiangyin, and Danyang dialects (belonging to the Wu family; see Lü 1985: 89). Then, the classifier jiā can be collocated with nouns meaning “family” or “school,” as illustrated in (10) and (11), respectively:




(10)

有一家姚姓。 (敦煌變文 舜子變) Yǒu yī-jiā Yáo xìng. have one-family Yao surname “There was a family whose surname was Yao.” (Dun Huang Bian Wen, Shun Zi Bian, AD 800–1000)

(11)

佛家若勝。 (敦煌變文 降魔變文) Fó jiā ruò shèng. Buddhism school if win “If the Buddhism school wins” (Dun Huang Bian Wen, Xiang Mo Bian Wen, AD 800–1000)




The classifier mén had the same usages as jiā at the time, as illustrated in (12) and (13): (12)




一門叔父則有何大、 中郎。 (世說新語 賢媛) Yī mén shūfù zé yǒu Hé Dà, Zhōng Láng. one family uncle but have He Da Zhong Lang “The uncles of the single family also have He Da and Zhong Langa.” (Shi Shuo Xin Yu, Xian Yuan, AD 450)

(13)

佛家有三門。 (朱子語類卷八) Fójiā yǒu sān mén. Buddhism have three school “Buddhism had three schools.” (Zhu Zi Yu Lei, Chapter 8, AD 1200)

Our analysis finds support in three pieces of diachronic evidence. First, phonologically, the form of the classifier mén in Medieval Chinese was [muǝn], with the even tone,2 and in the plural -men the vowel diphthong has been reduced to schwa [ǝ] and the tonal value has been neutralized. Second, semantically, the classifier mén always referred to a family or a school – more than one human being – which perfectly matched its plural meaning that was restricted to the plurality of human nouns. Third, syntactically, the 2

This is based on Li and Zhou (1999).

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

The Emergence of Plural Morphemes

477

classifier mén occurred between the modifier and the head, the proper context for developing a plural morpheme because they both shared the same distribution. At the time, the classifier mén could also follow a pronoun, as illustrated below, in exactly the same position as when it was used as a plural marker: (14)




如何聘向他門。 (敦煌變文 醜女緣起) Rúhé pìn xiàng tā-mén. how marry with he-family “How do you marry with his family?” (Dun Huang Bian Wen, Chou Nü Yuan Qi, AD 800–1000)

Now we can conclude that both the plural -men in standard Mandarin and -jia in dialects developed directly out of their original usages as classifiers. Therefore it is safe to say that the establishment of the classifier system was solely responsible for the emergence of the plural markers in the grammar of the Chinese language. Behind this diachronic interaction was a profound reason for the emergence of the plural morphemes, as the data from Chinese dialects suggest. First, let us examine the rich lexical sources for the plural category in the different dialects. In addition to -men and -jia, there are many other sources in dialects for grammaticalizing into a plural, mainly including the following. (a) (b) (c) (d)

The pronominal word dōu “all,” e.g. the Huojia, Handan, and Anyang dialects (Huang 1996: 432). The noun rén “people,” e.g. the Xixian dialect (Huang 1996: 432). The quantifiers jì-gè “several CL,” as in the Suqian dialect, and nà yī xiē “that one some,” as in the Zibo dialect (Huang 1996: 438, 442). Varying tonal values, a kind of inflection; e.g. in the Xian dialect, the singular versus plural contrast is the first-person pronoun ŋɤ53 ~ ŋɤ21, the second-person pronoun ȵi53 ~ ȵi21, and the third-person pronoun uo53 ~ uo21. That is, the tone value 21 is reserved for indicating plurality. The use of inflection to differentiate between singular and plural is also found in many other dialects, such as the Linyi, Wangrong, and Yuncheng dialects (Huang 1996: 449‒450).

In fact, the specific forms of plurals vary greatly from dialect to dialect, and some even make use of purely phonological apparatus. However, these dialects have one thing in common: each has its own grammatical device to express the distinction between singular and plural, which means that none of the dialects lacks the plural category, a sharp contrast to the grammar of Old Chinese, which did not have such a “number” category. In our view, the general motivation for the emergence of this “number”

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

478

Pronouns, Plurals, and Diminutives

category was that the establishment of the classifier system increased the consciousness of language users about various kinds of quantity when referring to things. At this time, the number category entered the cognition of the language community in all regions, and under this circumstance they were forced to create some kind of apparatus to encode the information about the number of nouns. As a result, no dialect could resist this change, and they had to innovate a plural device for their daily communication. A similar situation happened regarding the emergence of diminutive morphemes, which were introduced into the language in the same period as classifiers and plurals emerged, as discussed below. All these changes did not happen randomly or in isolation but were a response to the general trend of grammatical evolution. As we mentioned previously, the grammaticalization processes of the plural markers in different dialects are at different stages. A criterion for judging the stage is the scope of the extension of the grammatical morpheme in question: the wider its extension, the more developed its grammaticalization is. In standard Mandarin, for example, the plural -men is restrained to pronouns and human nouns, but in the Shijiazhuang dialect it is extended to refer to animal nouns, and in the Guizhou dialect it is further extended to refer to inanimate nouns, quite like the plural -s in English. That is, the plural marker is more grammaticalized in the Guizhou dialect than in the other two dialects.

19.4 Classifiers and Plurals As analyzed above, the emergence of the plural morpheme was motivated by the classifier system; the distribution of the plural morphemes was also influenced by classifiers. According to Chierchia (1998a, 1998b), classifiers are intrinsically incompatible with plural markers; thus they tend not to occur with them within a single language. This prediction seems problematic because Chinese acquired both classifiers and plurals in roughly the same historical period. Truly, the plural -men could never cooccur with a classifier within a nominal structure throughout history, e.g. *sān-gè tóngxué-men “three-CL student PL.” It seems that the phrase “Num + CL” serves to express an accurate number, whereas the phrase “noun-men” is specialized to refer to a fuzzy number. As far as the demonstratives are concerned, there are neat correspondences between the accurate and fuzzy singular/plural in Contemporary Chinese, as illustrated in Table 19.2. First, let us consider the examples of the fuzzy number usages of demonstratives. The phrase refers to singular when there is no numeral or classifier between the demonstrative and the head noun, as illustrated in (15). If the phrase refers to more than one, the morpheme -xie must be used, as illustrated in (16). This contrast did not exist in the language before the tenth century AD.

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Classifiers and Plurals

479

Table 19.2 Accurate and fuzzy singular/plural of demonstratives

Fuzzy Accurate

Singular

Plural

Dem N Dem CL N

Dem xie N Dem Num CL N




(15)

那人為他閃開道。 (王朔 頑主) Nà rénwéi tā shǎn kāi-dào. that person for he open-road “That person gave way to him.” (Wang Shuo, Wan Zhu, Contemporary Chinese)

(16)

這些書都能背下來吧。 (王朔 編輯部的故事) Zhè-xiē shū dōu néng bèi-xiàlái ba. this-PLU book all may memorize-down PRT “(You) can memorize all these books.” (Wang Shuo, Bian Ji Bu De Gu Shi, Contemporary Chinese)




The following examples illustrate the accurate expressions of the number category. When only a simple classifier occurs between the demonstrative and the head noun, the nominal phrase must refer to a singular referent, as illustrated in (17). When expressing accurate plurality, a concrete number must be used to specify it, and the nominal phrase cannot be suffixed with the fuzzy plural -men, as illustrated in (18): (17)

他仍然唱著那首歌。 (現代漢語) Tā réngrán chàngzhe nà shǒu gē. he still sing-PROG that-CL song “He is still singing the song.” (Contemporary Chinese)

(18)

這兩本書跟有意思。 (現代漢語) Zhè-liǎng-běn shū hěn yǒuyìsi. this two-CL book very interesting “These two books are very interesting.” (Contemporary Chinese)

As the example in (15) shows, when used to refer to a singular noun in standard Mandarin, the demonstrative can directly precede the head noun without being linked by any classifier, which can be viewed as “zero marking” for singularity, similar to the situation in English. However, this method of marking singularity is ruled out in many

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronouns, Plurals, and Diminutives

480

dialects, which therefore require an appropriate classifier to link the demonstrative with its head. The dialects of this type include the Lanzhou, Shanghai, Southern Min, and Cantonese dialects (for details, see Huang 1996: 465‒505). Let us consider two examples of the Lanzhou dialect: (19)

這個人 zhè ge rén “this-CL person” 這件衣裳 zhè-jiàn yīshang “this-CL clothes”

*這人 *zhè rén “this person” *這衣裳 *zhè yīshang “this clothes”

The fact that there must be a classifier to link the demonstrative to the head noun reveals another major property of plurality in the grammar: the singular/plural marking is closely related to the expression of definiteness, as mentioned above. Since personal pronouns and demonstratives have an inherent meaning of definiteness, the “number” feature needs to be signaled by means of certain grammatical devices. Thus -men marks the plurals of personal pronouns, and the singular use takes a zero marking in standard Mandarin. For demonstratives, the singular is signaled by a classifier and the plural by -xie or a concrete “Num CL” phrase. In these dialects, both the singular and plural uses of demonstratives must be overtly expressed by a grammatical morpheme, as in some Iranian languages that overtly mark both singular and plural. When combined with human nouns, in Contemporary Chinese, the use of the plural -men does not produce exactly the meaning of the “plural” concept in the same way as in English, which actually means a collective number (typically more than one) in a definite domain (for a discussion of this issue, see Chao 1979: 125). Therefore, the phrase “bare noun-men” is confined to preverbal position unless it is modified by a definite determiner such as a demonstrative, under the influence of the principle of definiteness assignment by syntactic position (for details, see Section 7.6), as illustrated in (20): (20)

(a) 老師們我都通知到了。 (現代漢語) Lǎoshī-men wǒ dōu tōngzhī-dào-le. teacher-PLU I all inform-of-PERF “I have informed all of the teachers.” (b) *我都通知到了老師們。 *Wǒ dōu tōngzhī-dào-le lǎoshī-men. I all inform-of-PERF teacher-PLU (Contemporary Chinese)

19.5 Word Formation for Uncountable Nouns The distinction between countable and uncountable nouns is important in languages with plural marking. In English, for example, some uncountable nouns can be known

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Further Development of the Classifier Gè

481

from the natural properties of their referents, such as water, soil, and weather, which are names of inseparable things that have nothing to do with enumeration; hence they are uncountable nouns. However, some uncountable nouns in English are languageparticular, such as news, furniture, luggage, bread, and knowledge, which are perfectly countable in Chinese (i.e. the corresponding words can be freely modified by “Num CL” phrases). Therefore foreign-language learners must memorize the group of uncountable nouns in order to use the grammar of English correctly. Since Chinese has no equivalent to the plural marking -s in English that is mandatory in many contexts, few researchers have noticed that there is actually a group of uncountable words in Chinese that are generic nouns and cannot be modified by numeral phrases. Here we are particularly interested in a set of uncountable compound nouns that are coined by a rule of word formation: “noun root + classifier,” as exemplified in (21): (21)

The word formation of uncountable nouns 人口 rén-kǒu “person-CL” 書本 shū-běn “book-CL” (generic (population) book) 車輛 chē-liàng “car-CL” (traffic) 馬匹 mǎ-pǐ “horse-CL” (generic horse) 船隻 chuán-zhī “boat-CL” (generic 紙張 zhǐ-zhāng “paper-CL” boat) (generic paper) 槍支 qiāng-zhī “gun-CL” 燈盞 dēng-zhǎn “lamp-CL” (generic gun) (generic lamp)

According to Wang (1989: 31‒32), this kind of word formation started to occur as early as the fourth century AD, when classifiers started to be grammaticalized. Therefore we believe that the emergence of the above uncountable nouns in Chinese was a by-product of the establishment of the classifier system.

19.6 Further Development of the Classifier Gè According to Lü (1999), approximately 260 classifiers are actively used in the daily communication of Contemporary Chinese, and the collocation of classifiers and nouns is largely conventional. Nonetheless, the scope of collocation and the frequency of use vary greatly from one classifier to another. The most frequent classifier is gè, which has the most general meaning and can be combined with a tremendous number of nouns. Additionally, it has now developed into a demonstrative and a major grammatical morpheme in many southeastern dialects. In standard Mandarin, the classifier gè has developed two important functions: a nominalizer and a complementizer (for a discussion of its status, see Ding et al. 1961: 66, Zhu 1982: 121).

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronouns, Plurals, and Diminutives

482

As a nominalizer, when the classifier gè is inserted between the verb and the resultative, the whole verb–resultative construction will become a VO construction. Here, gè functions to nominalize the following adjective or VP, as illustrated below: (22)

他們玩了個痛快。 (現代漢語) Tāmen wán-le gè tòngkuài. They play-PERF NOM joy “They had a good time.” (Contemporary Chinese)

(23)

他們掃了個一乾二淨。 (現代漢語) Tāmen sǎo-le gè yīgān èr jìng. they sweep-PERF NOM clean “They swept it super-clean.” (Contemporary Chinese)

In the above examples, the general classifier gè is functionally similar to the indefinite article a/an in these English phrases, e.g. to give him a push and to give him a call. According to Lü (1984: 145‒175), the above two functions of the classifier gè, namely nominalizer and complementizer, started to occur after the tenth century AD. Its use as a complementizer is illustrated as follows:




(24)

贏個他家偏有。 (辛棄疾 一枝花) Yíng gè tā jiā piān yǒu. win COMP he family especially rich “(He) won so that his family became especially rich.” (Xin Qi Ji, Yi Zhi Hua, AD 1200)

(25)

只沒有個看著你自己作踐了身子呢。 (紅樓夢二十回) Zhǐ méiyǒu gè kàn-zhe nǐ-zìjǐ zuòjian-le shēnzi ní. only not-have COMP see-PROG yourself damage-PERF body PRT “The only thing (I) didn’t see is you damaging your health.” (Hong Lou Meng, Chapter 20, AD 1750)

In the above examples, gè marks the object complement clause, a sentence structure that did not exist before this time. Due to the emergence of the complementizer gè, the language acquired a new device to organize a sentence; hence more complex phrases could occur after the matrix verb. This development of the grammar was propelled by the further grammaticalization of the general classifier gè.

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

The Emergence of Diminutive Inflection

483

19.7 The Emergence of Diminutive Inflection Within the same period, roughly from the fifth century AD to the twelfth, there were three major innovations in Chinese: (a) the classifier system, (b) the plural morphemes, and (c) the diminutive inflection. The “inflection” concept here means that the diminutive marker is extremely productive and can be suffixed to almost all concrete nouns, and that it has reached an extreme end of the grammaticalization cline, losing its syllabic and even sound status manifested only as a phonological feature (i.e. adding a retroflex feature to the nucleus vowel of the root noun). A question arises: are these innovations merely accidental or closely related? In our view, they are intrinsically correlated, manifesting a profound change in the cognitive perspective of the language community for conceptualizing things in reality. Apparently, they reflect different perspectives of the “quantity” dimension of things in reality: (26)

(a) classifier: type of quantity (b) singular/plural: number of quantity (c) diminutive: size of quantity

As stated above, probably due to the emergence of classifiers, language users had to pay more attention to the quantity dimension of nouns in order to select the proper classifier. This kind of daily communication further aroused the consciousness of speakers about the related quantity properties of nouns. In these circumstances, facets such as the number and size of things naturally entered the domain of the conceptualization of the language community, which created devices to encode this kind of information. However, grammar is not a matter of conceptualization, and it involves various levels of abstract schemas at different levels. For instance, the plural and diminutive markers share this skeleton: “stem + suffix,” where the suffix is often a phonologically reduced form. This schema for nominal phrases did not exist in Old Chinese and took place from the first century AD to the fifth as a by-product of the disyllabification tendency. Under the influence of the increasingly stronger tendency, an increasing number of nouns became disyllabic by compounding two monosyllabic stems. Some nouns in the second morpheme position became nominal suffixes via semantic bleaching and phonological erosion, such as -zi, -er, and -tou, all of which were introduced into the language around the fifth century AD (for details, see Wang 1989: 165‒182). These suffixes serve to make a disyllabic word but do not make any semantic contribution to the compound words. Consequently, the combinational force of this word formation created a new schema for noun phrases.

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronouns, Plurals, and Diminutives

484

19.7.1 The Diminutive Inflection In standard Mandarin and some other dialects, especially Pekingese, the diminutive morpheme has been reduced to a phonological feature – the retroflex action, neither a syllable nor a sound, which grammaticalized from ŋiei (with the first tone) “child” in Medieval Chinese. The phonological feature is achieved by causing a retroflexion of the vowel of the preceding noun, and depending on the type of rhyme of the preceding syllable there are several restrictive rules for deriving a diminutive inflection, as stated below. (27)

(a) If the preceding rhyme is [a], [o], [e] or [u], the vowel is featured by the retroflex action. (b) If the preceding rhyme is [ai] or [ei], delete [i] and then add the vowel with the retroflex feature. (c) If the preceding rhyme is [an], delete [n] and add the retroflex feature to it. (d) If the preceding rhyme is [aŋ], delete [ŋ] and add both the retroflex and nasal features to the vowel.

According to the cline defined by Hopper and Traugott (2003: 6), the ultimate end of a grammaticalization is “zero”; hence the diminutive marker is almost at the end point of its grammaticalization, from a full syllable with a tonal value to a retroflex feature by losing its status as an independent syllable. This is the most reduced phonological form, being consonant with its maximal degree of productivity: this suffix can be used with almost any noun to produce a diminutive meaning. The diminutive marking also behaves like a nominalizer, turning verbs or adjectives into nouns,3 e.g. huà “draw” > huàr “picture,” cuò “wrong” > cuòr “mistake.” The morphological marking has been incorporated as an integrated part of compound words. e.g. xiǎohái-r “child.” It is cross-linguistically true that diminutive markers in various languages develop out of the “child” concept, as in Awtuw, Korean, Dagaare, and Baka (for details, see Heine and Kuteva 2002: 65‒67). Likewise, in Chinese, the diminutive marker grammaticalized from ér “child,” and its development was outlined by Wang (1989: 11‒13). In the first stage, roughly from the fifth century AD to the eighth, it was attached to names for children in a form of intimate address, as illustrated in (28). In the second stage, from the eighth century AD to the tenth, it started to occur after an animal name, such as yàn-er “swallow-DIM,” as illustrated in (29). In the third stage, from the tenth century AD onward, it could be used after inanimate nouns, such as chē-er “cart DIM,” as illustrated in (30).

3

Here, the letter “r” represents the retroflex feature of the diminutive marking.

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

The Emergence of Diminutive Inflection (28)

485




小字龍兒。 (南齊書 武帝紀) Xiǎo zì Lóng-ér. young name Long-DIM “He was called Long-er when he was a child.” (Nan Qi Shu, Wu Di Ji, AD 500)

(29)




細雨魚兒出。 (杜甫 水檻遣興詩) Xì yǔ yú-ér chū. small rain fish-DIM jump-out “Small fishes jumped out of water when it was drizzling.” (Du Fu, Shui Jian Qian Xing Shi, AD 750)

(30)




小車兒上看青天。 (邵雍 小車吟) Xiǎo chē-ér shàng kān qīng tiān. Small cart-DIM on look blue sky “I looked at the blue sky on a small cart.” (Shao Yong, Xiao Che Yin, AD 1050)

The grammaticalization process of the diminutive marker underwent the following stages: (31)

stage 1: collocating with human nouns only, stage 2: collocating with all animate nouns, stage 3: collocating with all nouns, animate or inanimate.

Recall that the plural marker -men can be collocated only with human nouns, and it still preserves a complete syllabic form, though its tonal value has been neutralized. Considering that these two morphemes started their grammaticalization processes at the same time, the diminutive marking -ér is more grammaticalized than the plural marker -men. Clearly, there is a correlation between the degree of grammaticalization and phonological reduction.

19.7.2 The Phonological Representation of the Diminutive in Other Dialects As we know, Chinese has eight families of dialects, and each family has various subdialects with remarkable differences in grammar, phonology, and vocabulary. Clearly, in many dialects, the “diminutive” category is not grammaticalized from any lexical source, because the phonological forms simply come from certain sound features of children’s speech. This reveals another significant channel for the innovation of grammatical morphemes. As we saw in Chapter 18, the distal demonstratives in various dialects are phonologically derived from their corresponding proximal ones that were

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

Pronouns, Plurals, and Diminutives

486

indeed grammaticalized from certain lexical sources, namely general classifiers. These diachronic facts tell us that grammatical devices do not have to develop out of any lexical source, and they may be formed by purely phonological principles. Crosslinguistically, diminutive markers are often represented by the following five types of phonological form (Jakobson and Waugh 1979, Moravcsik 1978, Jespersen 1922): (32)

(a) (b) (c) (d) (e)

reduplication, high pitch, nasal sound, front and high vowel, front consonant.

These phonological features are closely related to the speech of children; thus many languages express the diminutive sense by taking a sound feature of children’s speech (Brown and Levinson 1978: 268). In general, the diminutive marking comes out of two major sources: the lexical item “child” and a sound feature of children’s speech. At least two of the total five phonological forms listed in (29) are found in Chinese dialects to represent diminutives, as shown in Tables 19.3 (Huang 1996: 1‒8) and 19.4 (Chen 2002, Yan 1993, Ye and Tang 1982). For a given grammatical category, the phonological forms and lexical sources are highly diverse across dialects. At first glance, they often appear tremendously messy and arbitrary. Nevertheless, behind this picture is a third model of language change in addition to the two traditional models: the wave model and the tree model. To recapitulate briefly, the wave model posits that a linguistic innovation is first made in a specific

Table 19.3 Diminutive of reduplicated form in dialects Dialect

Basic form

Diminutive form

Wenshui Yichang Linli

dāo “knife” pán “plate” zì “character”

dāo-dāo pán-pán zì-zì

Table 19.4 Diminutive of high-pitched tone in dialects Dialect

Basic form

Diminutive form

Ningbo Lichuan Cantonese

tɕi “chicken” miau33 “cat” tʃɛ53 “car”

tɕi35 miau53 tʃɛ55

53

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

The Emergence of Diminutive Inflection

487

area and then spreads to neighboring areas. The tree model holds that one language diverges into separate languages due to immigration or political division. Overall, these two models may work in some areas of the language if only certain specific linguistic forms are concerned, but they fail to explain the complete stories of the history and the situation across dialects in Chinese. It has happened repeatedly in the history of the Chinese language that when some grammatical category was innovated by the central language community, the other dialects had to adapt to this innovation and struggled to invent devices to encode it. For the diminutive marking, some dialects obtained the device through grammaticalizing from a lexical item, but others may have taken a sound feature of children’s speech to encode it. As a consequence, no dialect could lack such a grammatical device to mark the grammatical category “diminutive.” The same is true for the plural markings and many other grammatical morphemes. These grammatical categories, such as the plural and the diminutive, did not exist in Old Chinese, but at a certain time they were demanded by the language and every language community in different regions created a form in its own way in response to the requirement. As a result, it was more likely that there was an “invisible hand” with great power to guide the development of a language, which made the grammar evolve in a regular and consistent fashion.

https://doi.org/10.1017/9781108921831.019 Published online by Cambridge University Press

20 Structural Particles

20.1 Introduction The functional word de is definitely one of the most important grammatical markers in the grammar of Contemporary Chinese; it has multiple functions and is the most frequently used word in both the spoken and written languages. According to the frequency dictionary edited by Beijing Language University,1 it has the highest frequency of use among all words, either content words or functional words; its basic function is to link relative clauses, genitives, and all other modifiers – except numbers – to the head nouns and to link adverbial phrases to the head verbs. Moreover, the nominal head can be omitted, so the morpheme de can be combined with numerous words to form nominal phrases or adjectival or adverbial phrases that can be used independently. Depending on which function is considered, this morpheme is labeled with various terms, such as relativizer, genitive marker, and associate morpheme. Considering that there is no proper cover term for all these functions, we use different labels for it depending on which function is being discussed; sometimes we just call it “the structural particle” as a cover term (for details, see Li and Thompson 1981: 575‒593, Zhu 1982: 142‒145, Shi 2010: 416‒435). However, such a grammatical morpheme was not introduced into the language until Late Medieval Chinese. In this chapter, we address where it came from, how it developed, and what factors triggered it to develop into a grammatical morpheme. The phonological form of this morpheme is [tǝ], in which the tonal value is neutralized and the vowel becomes a schwa, due to the phonological reduction of its grammaticalization process. It grammaticalized from the demonstrative dǐ in Medieval Chinese. Keep in mind that, in the following discussion, we use the full form dǐ to refer to the early lexical form of the grammatical morpheme.

1

The Frequency Dictionary of Contemporary Chinese (Peking Language University Press, 1986).

488

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

489

Structural Particles in History 20.2 Structural Particles in History

Before the particle de was introduced into the language around the eighth century AD, there existed at least three particles, namely zhī, zhě, and suǒ, in the language. They overlapped functionally but each possessed its own grammatical characteristics that distinguished it from the others. All three particles were already widely used in the earliest texts, such as the Shi Jing (composed from the eleventh to the seventh century BC), so we cannot investigate their grammaticalization paths because of the lack of diachronic evidence. However, it is extremely common for functional words to coexist with their lexical forms for a long time, and some have always existed side by side. Therefore we can reconstruct the grammaticalization processes on the basis of data at a later stage. The earliest texts showed that these particles were all already used as demonstratives or other pronominal words at the time, which might reveal their lexical sources for grammaticalization into a structural particle on the basis of historical regularity within Chinese and across languages. A simple comparison of their functions is helpful for understanding how the overall properties of the grammar at a particular time influenced the semantics and syntax of grammatical particles that came from the same type of lexical source. In Table 20.1 “plus head” means that the head must be used to make a well-formed phrase. Table 20.1 shows that de is an all-round particle that possesses all the functions of the earlier three grammatical morphemes. As we will see in Chapter 21, the structure of the relative clause and its head noun underwent a typological change around the fifth century AD. Before then, zhī could link a relative clause to its head in the construction “REL zhī N,” functioning to delineate the head noun with a statement of static quality. Meanwhile, zhě could only introduce a relative clause in postnominal position in the construction “N VP zhě,” delineating the head noun by means of a concrete event or activity. In Late Medieval Chinese,

Table 20.1 Functions of four structural particles in history Function

zhī

zhě

relative genitive nominalizer for agent nominalizer for patient associate plus head head-omitting adverbial

√ √

√

√ √

√

√

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

suǒ

√ √

de √ √ √ √ √ √ √ √

Structural Particles

490

however, the postnominal zhě clause was abandoned, and de introduced a relative clause in prenominal position by taking over the functions of both zhī and zhě. In Contemporary Chinese, except for some expressions of inalienable possession in which the possessive relation can be unmarked, such as wǒ māma “I mother” (“my mother”) and tā shēntǐ “he body” (“his body”), possessive expressions require the particle de to associate the possessor with the head, as in wǒ de chē “my car.” The particle zhī can also mark a genitive phrase in Old Chinese, as illustrated in (1): (1)




夫子之文章可得而聞也。 (論語 公冶長) Fūzǐ zhī wénzhāng kě dé ér wén yě. Confucius GEN article can get and hear PRT “Confucius’ articles can be understood just by listening to them.” (Lun Yu, Gong Ye Chang, 500 BC)

The usages of the particle zhī were related to a language universal. Aristar (1991) pointed out that many languages use the same grammatical marker to indicate both genitives and relative clauses, suggesting that there is an intrinsic relation that originates from an anaphoric strategy. All these structural particles possessed the common function of nominalizing a verb into a nominal item, but the functions of their nominalized forms differed from one another. In Old Chinese, for example, zhī could be inserted between the subject and predicate of a clause to nominalize, as illustrated in (2), a very common usage at the time, but Contemporary Chinese does have an equivalent marker:




(2)

夫子之言性與天道, 不可得而聞也。 (論語 公冶長) Fūzǐ zhī yán xìng yǔ tiān-dào bùkě dé ér wén Confucius NOM talk nature and general-rule cannot get and understand yě. PRT “Confucius’ talking about the nature and the general rule could not be understood just from listening.” (Lun Yu, Gong Ye Chang, 500 BC)

(3)

即患秦兵之來。 (史記 廉頗藺相如列傳) Jí huàn Qín bīng zhī lái. because fear Qin troop NOM come “Because (he) feared the coming of the Qin troops.” (Shi Ji, Lian Po Lin Xiang Ru Zhuan, 100 BC)




In the written language of Contemporary Chinese, the particle de can nominalize a noun–verb phrase by being inserted between them, as in tā de xiào “his laughing”

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles in History

491

and tā de dàolái “his arrival,” which is similar to the usage of zhī in (3). However, de cannot nominalize an SVO phrase, which distinguishes it from the function of zhī described in (2). In Old Chinese, the pronominal zhě could be suffixed to a verb phrase to make a nominal phrase, referring to the agent of the verb, e.g. shí-zhě “eat-ZHE” meant “eater.” In contrast, the pronominal suǒ could precede the verb phrase to make a noun phrase, referring to the patient of the verb, as in suǒ shí “SUO-eat,” which meant “food.” For its nominalizing function, the form “suǒ-verb” was used as a passive marker from the first century BC to the fifth century AD, and the preposition wéi served to introduce the agent (for a full discussion, see Chapter 8). In Contemporary Chinese, de has the functions of both zhě and suǒ; e.g. chī-de “eat-DE” is ambiguous because it can mean either “eater” or “food.” Except for numeral modifiers that were unmarked in Old Chinese and have been linked by a classifier since Medieval Chinese, the modifiers were marked by zhī before the fifth century AD and by de after the eighth century AD. Zhī was always optional in linking a modifier to its head, but de became obligatory in many situations. For instance, if the modifier is an adjectival phrase with a quality intensified by a degree word or through reduplication, the particle de must be used to link the intensified AP to its head noun, as illustrated below: (4)

(a) 很好的朋友 (現代漢語) hěn hǎo de péngyǒu very good DE friend (b) *很好朋友 *hěn hǎo [ ] péngyǒu very good friend (Contemporary Chinese)

Additionally, the particle de is used to link an adverbial phrase to the verb, a function that the former three particles did not possess. Especially when the adverbial phrase is structurally complex, such as reduplicated adjectives and “degree word + adjective” phrases, the particle de is typically or even obligatorily used to link the modifier to the verb (Lü 1999: 161). This rule is similar to that applicable to noun phrases, as discussed previously, for example: (5)

(a) 王教授很熱情地招待了我們。 (現代漢語) Wáng Jiàoshòu hěn rèqíng de zhāodài-le wǒmen. Wang Professor very enthusiastically DE entertain-PERF we “Professor Wang entertained us very enthusiastically.”

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

492

Structural Particles

(b) *王教授很熱情招待我們。 *Wáng Jiàoshòu hěn rèqíng [ ] zhāodài wǒmen. Wang Professor very enthusiastically entertain we (Contemporary Chinese) Some researchers have argued that the adverbial particle de has a lexical source different from the nominal one (for the debates, see Jiang and Cao 2005: 253‒273). However, a diachronic examination reveals that both of them originated from the same lexical source, a demonstrative in Medieval Chinese (for details, see subsequent sections), and developed in parallel. Their functions also resemble each other, with both linking a modifier to its head, regardless of whether the head is nominal or verbal. In Old Chinese, zhī could occur only between the modifier and the head, which means that it was unable to appear without a head noun and occur in the final position of a nominal phrase. However, because zhě could not precede a head noun, the relative clause or adjectival phrase introduced by zhě could only follow the head noun (for a fuller discussion, see Chapter 21).2 In general, the “VP + zhě” phrase was used as an independent NP with no head noun. In comparison, in Contemporary Chinese, the “modifier + de” phrase can freely appear without a head to express the whole meaning of the complete noun phrase, a function that comes by analogy with classifiers (we will return to this point below). In the literature, there has been much debate about the lexical source of the particle de. Before the 1990s, when grammaticalization theory was introduced into the circle of Chinese linguistics in China, it was traditional for researchers to tend to seek the ancestors of recent grammatical functional words in Old Chinese. In their view, the number of functional words had not historically increased or decreased, and ordinary lexical items could not develop into functional words by themselves. For example, Lü (1984: 122‒131) speculated that the particle de originated from zhě in Old Chinese, Wang (1989: 224‒230) argued that it should be a legacy of the particle zhī at that time, and some researchers (e.g. Mei 1988) even tried to demonstrate that it was derived from both zhī and zhě. As we discussed above, despite some functional similarities among them, the particle de is remarkably distinct from either zhī or zhě, semantically and syntactically; needless to say, a large phonological gap existed between them (Jiang 2005b: 163‒176). Probably for this reason, Feng (1990) speculated that the particle de might have come from its original demonstrative usage but presented no further analysis to justify his argument. However, this idea was further pursued by our work (Shi and Li 2001: 305‒325), which was the first study to provide a detailed analysis of the

2

According to Lü (1984), there were some exceptions in ancient texts, but they were found in the written language, which might not reflect their real usage in the vernacular language.

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Classifiers and the Structural Particle De

493

grammaticalization of de from its former demonstrative by analogy with the classifier word class. Disagreeing with the analysis of Shi and Li, however, Jiang (1999a) suggested that the particle de might have come from its former locative usage. It is true that the former particle suǒ also meant “place” and another particle xǔ with a similar function, which was occasionally attested in history, also had a locative meaning, but their locative usages could hardly have evolved into the structural particle de because there were great semantic and syntactic differences between the locatives and the structural particles. The demonstrative dǐ was originally an ordinary noun, meaning “bottom,” which might have been extended to mean “that,” a usage that already existed before the emergence of the particle de. Even today, the demonstrative use of dǐ is still preserved in some dialects, such as the Edong dialect (Huang 1996: 465‒505). Thus the development of dǐ likely involved the following three steps of grammaticalization: (6)

Locative > Demonstrative > Structural particle (relative, genitive, and associate).

Apparently, the use of the demonstrative was necessary for its further development into relative, genitive, or associate markers, as is evidenced by typological, historical, and dialect research. As we will see below, the most important evidence is that the structural particle gè in many southern dialects of Chinese, such as Wu, Xiang, Hakka, and Cantonese, developed out of a demonstrative that came from a classifier (for details, see Chapter 17). 20.3 Classifiers and the Structural Particle De The history of the Chinese language has demonstrated again and again that grammaticalization is subject to the influence of the overall structural change of the grammar in a particular period. In this section, we try to prove that there was a causal relationship between the emergence of the classifier word class and the grammaticalization of the particle de from its earlier demonstrative use. It was quite common in the evolution of Chinese grammar for a group of lexical items with equal semantic suitability for a given grammatical category to develop into morphosyntactic markers in different periods and with distinctive functions. As mentioned previously, for example, the particle zhī in Old Chinese and de in Contemporary Chinese could both mark relative clauses, genitive phrases, and associate phrases. Although they originated from their earlier demonstrative uses, both zhī and de exhibited remarkably different grammatical properties. As discussed above, zhī in Old Chinese was always optional in marking modifiers of the three types of phrase. In contrast, the particle de in Contemporary Chinese is obligatory in connecting relative clauses and most

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

494

Structural Particles

possessors to their heads. However, “modifier plus de” phrases are often used without the head, but zhī did not have this function. The effect of the overall structure of the grammar on grammaticalization generally occurs via analogy. In historical morphosyntax, analogy can be viewed as a process of optimization of grammatical structure, a change that can be channeled by the structure of the language as a whole (Kiparsky 1992). As a language evolves, the structures of a language as a whole constantly and gradually change. When the changes accumulate to a critical point, a new optimal structure may expand at the expense of the old one. This new optimal structure drives the analogical change of certain lexical items and gives rise to new morphological markers through grammaticalization so that these markers are endowed with features in accordance with the optimal structure. Therefore these new markers may have semantic and syntactic properties that differ from those of their counterparts at an earlier stage of the language even though they might be derived from the same class of lexical items. Cross-linguistically, morphological markers in the same functional domain tend to grammaticalize from the same group of lexical items. However, the general structural properties of a particular language determine how far those markers can be grammaticalized. This section demonstrates that analogy was responsible for motivating the grammaticalization of de and extending its functions to other domains. This grammaticalization and functional extension were in line with the optimal grammatical structure of Chinese that emerged at that time. In other words, analogy may also be understood as a historical process that projects a generalization from one set of expressions (model form) to another (target form). In the case of the Chinese classifier system and the structural particle de, the former served as the model form and the latter was the target form. We hypothesize that the grammaticalization of dǐ, a demonstrative/interrogative pronoun, was induced by the new syntactic pattern “Numeral + Classifier + Noun” through analogy. As a result, many features of the particle de were actually inherited from the model form, namely the classifier system. Our view contradicts the widely held assumption that de was derived from the early morphosyntactic particle zhī through phonological evolution (Lü 1984, Wang 1989: 224‒230), as mentioned above. According to our analysis, both zhī and de developed independently from their own original demonstrative/interrogative function. They had different morphosyntactic properties and their diachronic transformation was motivated by different factors. Here we will explore the general question of what motivates a grammaticalization process. This issue needs to be clarified in the theory of grammaticalization. Although semantic suitability, salience, and frequency are among the prerequisites for grammaticalization, they do not necessarily initiate a grammatical change as Traugott (1994) put it. The development of the structural particle de suggests that grammaticalization is most likely to be initiated by the emergence of a new optimal structure. A cluster of

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Functions of the Structural Particle De

495

structural changes in a particular historical period, including innovation of new morphological markers, often reflected the outcome of the analogical leveling based on a newly introduced optimal structure at the time.

20.4 Functions of the Structural Particle De Before investigating its diachronic development, we offer a general picture of the semantic and syntactic properties of the structural particle de in Contemporary Chinese, parts of which have been indicated above (for a discussion on the use of it as a relativizer, see Chapter 21). Schematically, the particle de functions to connect a modifier to its head in five types of construction: relative clauses and genitive, adjectival, associate, and adverbial phrases. The use of de in each construction type is illustrated in the following examples. (7)

Relative clause: 教我數學的老師已經退休了。 (現代漢語) Jiào wǒ shùxué de lǎoshī yǐjīng tuìxiū-le. teach I math REL teacher already retire Perf. “The teacher who taught me math has already retired.”

(8)

Genitive phrase: 他們的房子非常漂亮。 (現代漢語) Tāmen de fángzi fēicháng piàoliang. they GEN house very beautiful “Their house is very beautiful.”

(9)

Adjectival phrase: 乾乾淨淨的衣服。 (現代漢語) Gàn-gān-jìng-jìng de yīfú clean (reduplication) ASSO cloth “clean clothes”

(10)

Nominal phrase: 北京市的古跡很多。 (現代漢語) Běijīng shì de gǔjī hěn duō. Beijing city ASSO historic-site very many “There are many historic sites in the city of Beijing.”

(11)

Adverbial phrase: 她喫惊地看着我們。 (現代漢語) Tā chījīng de kàn-zhe wǒmen. she surprise DE look-PROG we “She looked at us in surprise.”

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

496

In (11), “Adv + de + V” is a verbal phrase with the verb as its head. Although the syntactic category of the adverbial phrase is not the same as that of “Adj. + de + N,” they all share the syntactic schema of having a modifier marked by de preceding the head constituent. If the head noun is already identified in the discourse context, the structure “modifier + de” may occur without the head, as mentioned previously. This headless construction occurs widely for relative clauses and genitive, adjectival, and associated phrases, but it is less common for adverbial phrases to appear with the verb head, as illustrated below: (12)

(a) 這是我昨天買的(車)。 (現代漢語) Zhè shì wǒ zuótiān mǎi de (chē).3 this be I yesterday buy REL car “This is the car that I bought yesterday.” (b) 我買了一輛最貴的(車)。 Wǒ mǎi-le yī-liàng zuì guì de (chē). I buy-PERF one-CL most expensive ASSO car “I bought a most expensive car yesterday.” (c) 北京的(古跡)很多。 Běijīng de (gǔ-jī) hěn duō. Beijing ASSO historic-site very many “There are many historical sites in Beijing.” (d) 說什麼呢吞吞吐吐地。 Shuō shénme ne? tūn-tūn-tǔ-tǔ de. say what QU hesitant DE “What did you say? (You spoke) hesitantly.” (Contemporary Chinese)

The above functions are syntactically characteristic of the particle de, but all the earlier three particles (i.e. zhī, zhě, and suǒ) lacked them. The free-standing “modifier + de” shows the following syntactic hierarchy for the de phrase: (13)

[(Modifier + de) + head].

In the above structure, the modifier plus de first forms an immediate constituent. This constituent then combines with the head to establish a more complex syntactic construction.

3

The parentheses mean that the word contained in them is absent.

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Functions of the Structural Particle De

497

In many cases, the use of de is obligatory. Briefly, if the modifier is a relative clause, a reduplicated or quantified adjective, an alienable possessor, an associative phrase, or one of the majority of quantified adverbs, this particle must be used to link the modifier to its head. Otherwise, the construction will be ill-formed, as illustrated below. (14)

Relative clause: (a) 那是我寫的書。 (現代漢語) Nà shì wǒ xiě de shū. that be I write REL book “That is the book I wrote.” (b) *那是我寫書。 * Nà shì wǒ xiě [ that is I write

(15)

] shū. book

Reduplicated adjectives: (a) 這是一件乾乾淨淨的衣服。 (現代漢語) Zhè shì yī-jiàn gàn-gān-jìng-jìng de chènshān. this be one-CL clean-clean ASSO shirt “This is an entirely clean shirt.” (b) *這是一件乾乾淨淨襯衫。 *Zhè shì yī-jiàn gàn-gān-jìng-jìng [ that be one-CL clean-clean

(16)

] chènshān. shirt

Alienable possession: (a) 這是我的電腦。 (現代漢語) Zhè shì wǒ de diànnǎo. this be I GEN computer “This is my computer.” (b) *這是我電腦。 *Zhè shì wǒ [ ] diànnǎo. this is I computer

(17)

Associative phrase: (a) 這家賓館的價錢很合理。 (現代漢語) Zhè-jiā bīnguǎn de jiàqián hěn hélǐ. this-CL hotel ASSO price very reasonable “The price of this hotel is very reasonable.”

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

498

(b) *這件賓館價錢很合理。 * Zhè-jiā bīnguǎn [ ] jiàqián hěn hélǐ. This-CL hotel price very reasonable (18)

Adverbial quantifier: 4 (a) 他十分仔細地檢查了一遍。 (現代漢語) Tā shífēn zǐxì de jiǎnchá-le yī-biàn. he very carefully DE check-PERF one “He checked it over very carefully.” (b) *他十分仔細檢查了一遍。 *Tā shífēn zǐxì [ ] jiǎnchá-le yī-biàn. he very carefully check-PERF one-CL

Note that there are two factors determining whether the particle de is grammatically obligatory or optional. First, idiomatic or fixed “modifier + noun” phrases tend to be incompatible with de, for example, gē zhǐ dāo (lit. “cut paper knife”) “paper-cutting knife” does not need the particle de though the modifier is a verb phrase because it is a fixed term for a kind of tool. Second, “modifier + head” phrases without de may refer to lexicalized items whose meanings are not derivable from their components, but the corresponding phrases may have the sum of the meanings of their two components. For instance, bái-cài (lit. white-vegetable) means “cabbage,” but bái-de cài (lit. white DE vegetable) means “white paper.”

20.5 Demonstratives and Structural Particles A given grammatical domain recruits its morphological markers only from a limited set of lexical fields (Traugott 1994). Both cross-linguistic and diachronic evidence shows that demonstrative and interrogative pronouns are often candidates for being grammaticalized into morphological markers for relative clauses and other modifying phrases, e.g. that, which, where, and when in English. Similarly, in the history of Chinese grammar, several demonstrative or interrogative pronouns in different historical periods developed into relative clause markers, genitive markers, and so on. For example, the demonstrative pronouns in Contemporary Chinese, zhè and nà, are also on their way to evolving into structural particles (see Shi and Li 1998 for details). In what follows, we focus on the historical changes.

4

In the written language, adverbial phrases are associated with the heads by the Chinese character 地, but it shared the same source as the structural particle 的.

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Demonstratives and Structural Particles

499

20.5.1 The Structural Particle Zhī The changes in the functions of the structural particles can be better understood through a comparison of the current de and the corresponding zhī in Old Chinese. As was pointed out above, before the emergence of de, there was another form – zhī – to mark relative clauses, associate phrases, and genitive phrases that survived until the eleventh century AD, as illustrated below. (19)

Genitive phrases: 夫子之文章可得而聞也。 (論語 公冶長) Fūzǐ zhī wénzhāng kě dé ér wén yě. Confucius GEN article can get and read PRT “Confucius’ articles can be understood just from listening.” (Lun Yu, Gong Ye Chang, 500 BC)

(20)

Relative clauses: 其猶穿窬之盜也與? (論語 陽貨) Qí yóu chuān yú zhī dào yě-yǔ? he also go-through wall ZHI theft PRT “Is he not like a thief who climbs over the wall?” (Lun Yu, Yang Huo, 500 BC)

(21)

Associate phrases: 殊有确然之志。 (世說新語 識鑒) Shū yǒu què rán zhī zhì. definitely have firm so ASSO ambition “(He) certainly has a firm ambition.” (Shi Shuo Xin Yu, Shi Jian, AD 450)










According to Djamouri and Waltraud (1997), zhī was used exclusively as a demonstrative pronoun in the oracle bone inscriptions of the Shang Dynasty (from the sixteenth to the eleventh century BC), and its demonstrative usage coexisted with its grammatical function for several centuries. Example (22) illustrates the demonstrative usage of zhī in the third century BC: (22)




鄭人醢之三人也。 (左傳 襄公十五年) Zhèng rén hǎi zhī sān rén yě. Zheng people mince this three people PRT “The Zheng people minced these three men.” (Zuo Zhuan, Xiang Gong Shi Wu Nian, 550–400 BC)

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

500

20.5.2 The Structural Particle Gè Additionally, there was a historical structural particle gè that developed in parallel with the development of the particle de. The former is still used in many southern dialects today. Gè originally meant a length of bamboo, an important material for ancient utensils and instruments. Thus it developed into a classifier around the fifth century AD (Wang 1989: 18‒40, for details, see Chapter 16). Furthermore, this classifier began to be grammaticalized as a marker for modifier phrases and competed with the demonstrative dǐ after Late Medieval Chinese. The modifiermarking usages of gè are illustrated below. (23)

Genitive phrases: 你個骨是弃骨。 (張協狀元) Nǐ gè gǔ shì qì gǔ. you GEN bone be abandon bone “Your bone is abandoned (i.e. you are a humble man.).” (Zhang Xie Zhuang Yuan, AD 1200)

(24)

Adverbial phrases: 你丈夫想是真個不在家了。 (沈小霞相會出師表) Nǐ zhàngfū xiǎng shì zhēn gè bù zài jiā le. you husband guess be real GE not at home PERF “Your husband, I guess, is really not at home.” (Shen Xiao Xia Xiang Hui Chu Shi Biao, AD 1650)

Around the seventeenth century AD, gè developed into a relativizer. This usage is still preserved in the Wu dialect and many southern Mandarin dialects. The following examples are from the Shanghai dialect and the Jinhu dialect in Hubei Province: (25)

這是修鐘錶個工具。 (上海話) Tzeq zy sioe tzonpio ku kondjiu. these be repair clock REL tool “These are the tools that are used to repair clocks.” (Shi 2002b, Shanghai dialect)

(26)

我看家做莊家個人。 (金湖方言) Wǒ kānjiā zuò zhuāngjiā gè rén. I see do farm REL person “I saw one person who is a farmer.” (Wang 1991, Jinhu dialect)

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Demonstratives and Structural Particles

501

In addition to its classifier usage, gè was a demonstrative pronoun. Its demonstrative pronominal usage coexisted for several centuries with its classifier usage. Example (27) illustrates the demonstrative pronoun occurrence of gè in Medieval Chinese: (27)




咄哉個丈夫! (白居易 自詠) Duō zāi gè zhàngfū! aggressive PRT this man “How aggressive this man is!” (Bai Ju Yi, Zi Yong, AD 850)

In Mandarin Chinese, however, gè has become exclusively a general classifier, having lost its demonstrative, genitive, associate, and relativizer usages. However, all its usages in Modern Chinese have been preserved in the Wu dialect and some southern Mandarin dialects.

20.5.3 The Demonstrative/Interrogative Dǐ To begin with its diachronic development, let us examine the function of the demonstrative dǐ (i.e. the earlier form of the particle de) before it developed into the structural particle de. Similar to the two earlier structural particles zhī and gè, dǐ was originally used as a pronominal item, either a demonstrative or an interrogative pronoun, in Medieval Chinese, as illustrated in (28) and (29):




(28)

底事歸郎許。(劉希夷 江南曲) Dǐ shì guī láng xǔ. this matter belong youth promise “This matter belonged to the youth’s promise.” (Liu Xi Yi, Jiang Nan Qu, AD 700)

(29)

缘底名愚谷? (王維 愚公谷) Yuán dǐ míng yú gǔ? for what call fool valley “For what (reason) was it called Fool Valley?” (Wang Wei, Yu Gong Gu, AD 750)




Synchronic evidence from Mandarin Chinese also exhibits morphosyntactic similarities of demonstrative pronouns and the structural particle de. As pointed out previously, de is obligatory in linking a relative clause or a genitive phrase to the head noun. However, if a head is already modified by a demonstrative pronoun, either zhè or nà, in Contemporary Chinese, de will become optional, as illustrated below:

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

502 (30)

(a) 他就是剛才在這裡喫飯 (的) 那人。 (現代漢語) Tā jiù shì gāngcái zài zhèlǐ chīfàn (de) nà rén. he exactly be just at here eat-meal that person “He is exactly the person who just ate here a while ago.” (b) 我 (的) 這輛車已經跑了八萬公里。 Wǒ (de) zhè-liàng chē yǐjīng pǎo-le bā-wàn I DE that-CL car already run-PERF eighty-thousand gōnglǐ. kilometer “My car has already run 80,000 kilometers.” (Contemporary Chinese)

If there is no demonstrative in the two examples above, the use of the particle de is obligatory. This fact shows that the demonstrative can function like a structural particle in certain contexts, revealing that there might be a functional similarity of these two word classes. It makes it possible for the structural particle de to be diachronically derivable from its original demonstrative usage. Both the cross-linguistic and Chinese data support the proposal that demonstrative/interrogative pronouns are particularly suitable for grammaticalization into the structural particle for marking relative clauses and associate and genitive phrases. As analyzed above, this process of grammaticalization happened at least three times in the history of Chinese.

20.5.4 Formation of the Classifier System To understand the causal relationship between the classifier system and the emergence of the structural particle de, let us briefly review the development of the classifier system. In Chapter 16, we offered a detailed discussion of the development of the classifier word class. According to Shi and Li (2001: 282‒304), the development of classifiers underwent the following stages in Chinese. Stage 1. In Old Chinese, numeral phrases did not require any classifier, and, in reality, the classifier word class had not yet come into existence. Nouns could be directly modified by numerals, as in English. Example (30) illustrates a noun phrase composed of a numeral and a head noun at this stage. (31)




三人行, 必有吾師焉。 (論語 述而) Sān rén xíng, bì yǒu wǒ shī yān. three person do definitely have my teacher among “Among three people who work together, one of them must be my teacher.” (Lun Yu, Shu Er, 500 BC)

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Demonstratives and Structural Particles

503

Stage 2. In the first century BC, the form “numeral + Classifier + noun,” the paradigm of numeral phrases in Contemporary Chinese, started to appear, as illustrated in (32): (32)




其富人至有四五千匹馬。 (史記 大宛列傳) Qí fù-rén zhì yǒu sì-wǔ-qiān pǐ mǎ. its rich-people most have four-five-thousand CL horse “Its richest people have four or five thousand horses.” (Shi Ji, Da Wan Lie Zhuan, 100 BC)

From the first century AD to the eighth, the new structure of numeral phrases illustrated above grew steadily, but was still less common than the old numeral phrases without classifiers. In other words, classifiers were largely absent from numeral phrases, and they were always optional at this stage. Stage 3. The turning point in the development of the classifier system happened around the tenth century AD, a period in which a classifier was generally needed to mark numeral phrases. By the fourteenth century AD, the classifier system had been firmly established at the expense of the old “number + noun” numeral form. Since then, the numeral modifier and the head noun must be linked by a proper classifier; otherwise, the structure will become ungrammatical, as illustrated below: (33)

(a) 我有三本書。 (現代漢語) Wǒ yǒu sān-běn shū. I have three-CL book “I have three books.” (b) *我有三書。 *Wo you san shu. *Wǒ yǒu sān [ ] shū. I have three book

(Contemporary Chinese)

Table 20.2 provides a general picture of the development of the classifier system. As Table 20.2 shows, from the eighth century AD to the twelfth was the key period for the development of the classifier system. The same period also happened to be when the particle de made its first appearance (around the eighth century AD), and it developed quickly after that. According to Shi and Li (1998), one critical change in sentence structure was responsible for the emergence of the classifier system (for details, see Chapters 7 and 8). The earlier structure of a simple sentence was as follows:

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

504

Table 20.2 The rise and decline of the old and new numeral phrases

(34)

Title of sample text

Time

Without classifier

With classifier

Shi Shuo Xin Yu Dun Huang Bian Wen Zhu Zi Yu Lei Lao Qi Da

AD 450 AD 800–1000 AD 1200 AD 1300

88% 80% 29% 1%

12% 20% 81% 99%

Subject + verb + object + XP.

The XP position, which can be called the “second predicate,” could be occupied by many kinds of syntactic category, including intransitive verbs, adjectives and “number + classifier” phrases. We have provided examples in which the XP is a “number + classifier” phrase. Due to the fusion of the verb and the XP, which occurred first in constructions with an elided object, the XP position was gradually eliminated (for a fuller discussion, see Chapter 7). As a response to this change, the “number + classifier” phrase was moved to precede the head noun, where it was used as a modifier of noun phrases. It was in this context that certain measure words became grammaticalized into a new syntactic category – the classifier, which eventually became obligatory in numeral expressions. Initially, the “numeral + classifier + noun” phrase could appear only in postverbal position, while classifiers were still optional for numeral expressions. After the classifier underwent further grammaticalization, it was limited to preverbal position, signaling the completion of the development of the classifier system. Note that the development of the demonstrative dǐ into the structural particle was parallel to that of classifiers in two major respects. First, dǐ began to appear as a structural particle when the new “numeral + classifier + head” form was gradually established. Second, in the thirteenth century AD, both of the significant developments, the firm establishment of the classifier system and the completion of the grammaticalization process from the demonstrative dǐ to the particle de, had reached their completion point.

20.5.5 Commonalities of Classifiers and the Particle De From a synchronic point of view, classifiers and the particle de also share crucial morpho syntactic properties. First, both the classifier and the particle de can be considered grammatical markers in the following schema: (35)

modifier + grammatical marker + head.

In other words, both the classifier and de function to connect a modifier to its head.

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Demonstratives and Structural Particles

505

Second, there is a complementary distribution between the type of modifier marked by the classifier and the type of modifier marked by de: only numeral phrases can cooccur with classifiers, but any other modifier (except for numerals) can co-occur with de. Third, in contexts in which the head noun is understood, the structural particle de and its preceding modifier can stand alone; e.g. tā chī de “he eat-de” could refer to tā chī de miànbāo “he eat de bread” (i.e. the bread he ate). Similarly, the “numeral + classifier” phrase can appear without its head if the head is identifiable from the context; for instance, sān zhāng “three + CL” may stand for sān zhāng zhǐ “three + CL + paper.” Fourth, in certain contexts the occurrence of the classifier renders the occurrence of de optional, and it functions like the demonstrative. As discussed above, de is obligatory if an adjective modifier is reduplicated or quantified by a degree word. However, if the head noun is preceded by a classifier phrase, the particle de becomes optional, as illustrated in (36): (36)

(a) *嶄新嶄新襯衫 (現代漢語) *zhǎnxīn zhǎnxīn [ ] chènshān brand-new brand-new shirt (b) 嶄新嶄新一件襯衫 zhǎnxīn zhǎnxīn yī-jiàn chènshān brand-new brand-new one-CL shirt “a brand-new shirt” (c) 嶄新嶄新的襯衫 zhǎnxīn zhǎnxīn de chènshān brand-new brand-new ASSO shirt “a brand-new shirt.”

(Contemporary Chinese)

If the “numeral + classifier” phrase is removed from (36b), as in (36c), the particle de becomes obligatory. The above phenomenon shows that there is a certain grammatical similarity of the classifier and the structural particle. Therefore we hypothesize that the grammaticalization of the demonstrative dǐ into the structural particle de was stimulated by the emergence of the classifier system. More supporting examples come from the Kaiping dialect, which belongs to the Yue family (Cantonese). In this dialect, a proper classifier that matches the head noun also expresses the possessive relation,5 similar to the particle de in Mandarin Chinese (Yue 1998), as illustrated below. There are hundreds of classifiers in this dialect, and each of them can be used as a structural particle. In other words, this dialect lacks a fixed structural particle. 5

There are many classifiers, each of which is applied to different types of noun.

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

506 (37)

(a) 我件帽 (開平方言) ŋɔ kin mou I CL hat “My hat” (b) 我本書 ŋɔ kyn ʃy I CL book “My book”

(Kaiping dialect)

Although there are no historical documents available for us to determine when the genitive use of the classifiers in the Kaiping dialect occurred, as illustrated in (36), logically it should have happened after the establishment of the classifier system, because the genitive function is a semantic extension of classifiers. Since classifiers and de are functionally similar and even interchangeable in certain contexts and their diachronic developments are parallel to each other, we propose the following hypothesis for the emergence of de. The emergence of de as a morphosyntactic marker was motivated by the new structure of numeral phrases: “numeral + classifier + head noun.” This grammatical structure is an instantiation of a general grammatical principle for Chinese: “modifier + grammatical marker + head noun.” The emergence of the numeral phrase made possible the development of this general syntactic principle in Chinese grammar. Another possibility to consider is the development of the morphological marker de for modifiers other than numerals as the extension of the analogy.

20.6 Grammaticalization of the Structural Particle De As pointed out above, the particle de has multiple functions, but they did not occur at the same time. Historically, it took four centuries or even longer to develop all these functions. From the sequence of the emergence of these functions, we can determine in what contexts the demonstrative dǐ was first grammaticalized and how it underwent further development into other functions. Roughly speaking, the development can be divided into the following three stages. Stage 1. A text composed around the eight century AD, Dun Huang Bian Wen, which reflected the vernacular language at the time, is the earliest text in which the particle de (spelled dǐ in the following examples) was attested. At the time, it was used mainly as a relativizer to link a relative clause with the head noun, as illustrated in (38) and (39). This reveals the point at which the demonstrative dǐ was motivated to develop into a structural particle.

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Grammaticalization of the Structural Particle De

507




(38)

解說昨夜見底光。 (敦煌變文 频婆沙罗) Jiěshuō zuóyè jiàn dǐ guāng. explain last-night see REL light “(He) explained the light that they saw last night.” (Dun Huang Bian Wen, Pin Po Sha Luo, AD 800–1000)

(39)

已滿今生發底願。 (敦煌變文 維摩詰所說經) Yǐ mǎn jīnshēng fā dǐ yuàn. already satisfy this life make REL wish “(I) have realized all the wishes that I made in this life.” (Dun Huang Bian Wen, Wei Mo Jie Suo Shuo Jing, AD 800–1000)




As we have mentioned repeatedly, verbs and adjectives are treated as the same word class in Chinese grammar, synchronically and diachronically. In Chapter 21, we will see that adjectival modifiers were treated as a sort of relative clause. Thus it is unsurprising that, in the beginning, the particle also linked an adjectival clause with its head noun, as illustrated in (40): (40)




相勸直論好底事。 (敦煌變文 無常經) Xiāngquàn zhí lùn hǎo dǐ shì. persuade only mention good REL thing “(You should) mention only good things when persuading others.” (Dun Huang Bian Wen, Wu Chang Jing, AD 800–1000)

A characteristic function of the particle de deserves special attention. At this earliest stage, the head noun of the relative clause (VP) could be omitted, and the “VP + de” behaved like a nominal phrase to refer to a thing, as illustrated below: (41)




此是俗門作底。 (敦煌變文 維摩詰所說經) Cǐ shì sú-mén zuò dǐ. this be ordinary-people make REL “This is what ordinary people made.” (Dun Huang Bian Wen, Wei Mo Jie Suo Shuo Jing, AD 800–1000)

As mentioned previously, the above function distinguished the particle de from the earlier two relativizers – zhī and zhě: the former could never omit its head noun but the latter could never be followed by a head noun. The double functions of the new particle de came from analogy with classifiers, and all classifiers could either link a numeral modifier with a head noun or appear without the head noun to refer to a thing. This fact demonstrates that despite having the same lexical sources, grammatical morphemes might possess different functions because they came into

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

508

existence in different circumstances, namely different grammatical systems at different periods. Stage 2. In Zu Tang Ji, a text composed in the tenth century AD, the particle de developed two new functions. First, it could link an adverbial phrase, typically a reduplicated or complex form, with the verb head, a function that the former zhī and zhě lacked, as illustrated in (42) and (43):




(42)

和尚驀底失聲便唾。 (祖堂集 百丈和尚) Héshàng mò dǐ shīshēng biàn tuò. monk sudden DI lose-voice then spit “The monk suddenly spat without speaking.” (Zu Tang Ji, Bai Zhang He Shang, AD 950)

(43)

裴相公有一日微微底不安。 (祖堂集 黄檗和尚) Péi Xiànggong yǒu-yī-rì wéi-wéi dǐ bù’ān. Pei scholar one-day slight-slight DI uneasy “Mr. Pei was slightly uneasy one day.” (Zu Tang Ji, Huang Bo He Shang, AD 950)




Second, the verb head could be omitted, and the “adverbial + de” phrase could be used as the predicate alone, as illustrated in (44) and (45). In this case, the adverbial was usually a reduplicated form of an adjective or adverb.




(44)

舉措悉皆索索底。 (祖堂集 岩頭和尚) Jǔcuò xījiē suǒ-suǒ dǐ. behavior all fumble-fumble DI “They were all a-fumble.” (Zu Tang Ji, Yan Tou He Shang, AD 950)

(45)

莫終日口密密底。 (祖堂集 仰山和尚) Mò zhōngrì kǒu mì-mi dǐ. Do-not all-day mouth talkative-talkative DI “Don’t keep talking all day long.” (Zu Tang Ji, Yang Shan He Shan, AD 950)




Historically, there were parallel developments in the structures of verbal and nominal phrases, as is perfectly shown in the above examples. The same grammatical morpheme could be used in both NPs and VPs, linking the relative clause or the adverbial phrase to their head, either nominal or verbal, where both heads could be omitted (for a related discussion, see Chapter 21).

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Grammaticalization of the Structural Particle De

509

Stage 3. Around the twelfth century AD, all the major functions of the particle de in Contemporary Chinese had developed. First, the modifier could be a nominal phrase, as illustrated below: (46)

生絹底衣服, 漸漸底都曹破。 (萬秀娘仇報山亭兒) Shēng juàn dǐ yīfú jiànjiàn-dǐ dōu cáo-pò. raw silk DI clothes slowly all wear-out “All the raw silk clothes were worn out.” (Wan Xiu Niang Chou Bao Shan Ting Er, AD 1200)

(47)

我家中有八十歲底老母。 (萬秀娘仇報山亭兒) Wǒ jiā zhōng yǒu bāshí-suì dǐ lǎomǔ. I family in have eighty-year DI old mother “I have a mother who is eighty years old.” (Wan Xiu Niang Chou Bao Shan Ting Er, AD 1200)

Importantly, the particle de further developed into a genitive marker, one of the most important functions in Contemporary Chinese, as illustrated below: (48)

你如今必竟是我底丈夫。 (萬秀娘仇報山亭兒) Nǐ rújīn bìjìng shì wǒ dǐ zhàngfū. you now after-all be I GEN husband “Now you become my husband after all.” (Wan Xiu Niang Chou Bao Shan Ting Er, AD 1200)

(49)

大官人牽了萬秀娘底馬。 (萬秀娘仇報山亭兒) Dà Guānrén qiān-le Wànxiù Niáng dǐ mǎ. important official lead-PERF Wanxiu lady GEN horse “The gentlement led Lady Wanxiu’s horse.” (Wan Xiu Niang Chou Bao Shan Ting Er, AD 1200)

Additionally, at this stage, the particle de developed into a sentence-final marker, functioning to emphasize the truth of the statement, as illustrated in (50), or to indicate an action in the past, as illustrated in (51): (50)

惹著他不是輕放手的。 (碾玉觀音) Rězháo tā bùshì qīng fàngshǒu dǐ. provoke he easily let-go EMP “He won’t easily let you go if you provoke him.” (Nian Yu Guan Yin, AD 1200)

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

Structural Particles

510 (51)

喫他的酒才來的。 (错斩崔宁) Chī tā de jiǔ cái lái dǐ. drink he GEN wine just come EMP “(The man) came just after he drank his wine.” (Cuo Zhan Cui Ning, AD 1200)

From the twelfth century AD to the present day, the particle de has continued undergoing some adjustments in its functions, but no further major change has happened. This point was also when the classifier system had completed its development, so that the proper classifier was required to connect a numeral phrase to its noun head. This was not a coincidence because there was a cause–effect relationship between the two events, as we argued above; that is, the emergence and establishment of the classifier system motivated and gave impetus to the development of the structural particle. Once again, this proves that the grammar evolved in a regular, consistent, and systematic fashion.

https://doi.org/10.1017/9781108921831.020 Published online by Cambridge University Press

21 Word Order and Relative Clauses 21.1 Introduction Among all forty-five Greenbergian universals, the correlation “VO ⸧ N REL” is the most regular and is almost free of exceptions because Chinese is the only counterexample.1 However, the Chinese language has historically actually obeyed this universal correlation, in which the relative clause also follows the head noun. This chapter addresses the typological change in relative clauses from postnominal to prenominal position over time. We also propose that the constituent order of the relative clause and the head noun is not directly related to the verb and its arguments, as is generally assumed in the literature, but is correlated with the ordering of the VP and PP instead. The diachronic evidence of Chinese shows that the order change of the PP from postverbal to preverbal position was the critical motivation for the change in question. Additionally, our analysis suggests that the Greenbergian universal correlations neither represent an optimal state of any language nor function as an impetus for any language to develop from an inconsistent state to a consistent one. Rather, they are more likely to represent the default settings of any language structure, and every language can customize them according to its needs. Conversely, there is no pressure on any language to return to the default settings. In our view, no language identified as being “inconsistent” should develop toward a purely consistent system; in other words, the so-called inconsistent languages are normal and inevitable. Greenberg (1966a) suggested that a number of other features tend to correlate with the relative order of verbs and objects. In fact, however, they merely reflect statistical tendencies, and few of these correlations are free of exceptions. Among them, the following correlation is the strongest universal: (1) If a language is SVO, the relative clause follows the head noun, namely, SVO ⸧ N REL.

1

The symbol “⸧” means “entail” in mathematic logic.

511

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

512

Of the 645 sample languages that Dryer (1992) investigated, Chinese is the only exception to the above implicational universal.2 Although there has been debate over whether Chinese has undergone a word order change from SVO to SOV (Li and Thompson 1974, Sun and Givón 1985), the unmarked word order of Chinese has always remained SVO throughout history. It is true that, in Modern Chinese, the relative clause always precedes the head noun, which is necessarily connected by the relativizer de,3 as illustrated in (2) (cf. Section 20.6). (2)

這是我看書的桌子。 (現代漢語) Zhè shì wǒ kàn shū de zhuōzǐ.4 this be I read book REL table “This is the table where I read books.” (Contemporary Chinese)

From the perspective of Contemporary Chinese, furthermore, the Chinese language is a remarkable exception to another of the Greenbergian universals; that is, Universal 24 in Greenberg (1966a) which states, (3)

If the relative expression precedes the noun either as the only construction or as an alternate construction, either the language is postpositional, or the adjective precedes the noun or both.

Chinese has always been prepositional throughout history,5 and, more importantly, adjectives in Modern Chinese do not behave like relative clauses; for example, they can directly modify the head noun without necessarily using the morpheme de, as in dà shù “big tree”; in contrast, relative clauses must be marked by this morpheme. In addition, Chinese stands as the only counterexample to another Greenbergian universal correlation regarding the structure of comparison (Dryer 1992). Greenberg (1966a) identified VO order and “adjective–marker–standard” order in comparative 2

3

4 5

Dryer (2005) reported that, in the 756-language sample, only Chinese (including its dialects), Bai (a Tibetan-Burman language spoken in the Yunnan Province of China) and Amis (an Austronesian language spoken in the Taiwan Province of China) have the VO and REL N patterns. In Bai and Amis, these patterns should be the effect of Chinese influence. This grammatical marker, which grammaticalized from a demonstrative in Late Medieval Chinese, has multiple functions in Contemporary Chinese, such as the possessive marker (to indicate a possessive relationship between two nouns) and the associative marker (to connect a phrasal modifier to the head noun). In the examples of this chapter, the part in bold font refers to the head and the underlined part to the relative clause. Li and Thompson (1981: 356) argued that there are postpositions in Chinese; for example, in zài zhuōzǐ shàng “on table above,” shàng is regarded as a postposition in their analysis. However, this word can always be replaced by its corresponding disyllabic compound word shàngmiàn “upper cover,” which is a full noun even in Contemporary Chinese. Thus the number of the words of this type is very limited, and they can be better treated as nouns.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Introduction

513

structures (e.g. John is taller than Tom) and OVorder and “standard–marker–adjective” order as the most common patterns. Chinese matches neither of these two correlations; consider the following example of Modern Chinese (for details, see Chapter 11): (4)

約翰比湯姆高。 (現代漢語) Yuēhàn bǐ Tāngmǔ gāo John COMP Tom tall Subj Marker Standard Adj “John is taller than Tom.” (Contemporary Chinese)

In the above comparative example, on the one hand, the “Marker + Standard” phrase precedes the adjective, which distinguishes Chinese from other VO languages. On the other hand, in the Chinese comparative structure, the marker precedes the standard, which differs from the OV languages in which the order is the reverse (see Dryer 1992 for details). However, the Chinese language historically was actually not an exception to the above universal correlations; in other words, it perfectively reflected the universal correlations of the constituent orders of the relative clause and the head noun, as well as in the comparative structures. Specifically, from Old to Medieval Chinese, as an SVO language, Chinese had the order “N REL” and the structure “Adjective + Marker + Standard” (cf. Chapter 11), as illustrated in (5) and (6), respectively:




(5)

客來早者並得佳設。 (世說新語 雅量) Kè lái zǎo zhě bìng dé jiā shè. guest come early REL all receive delicious food “All the guests who came early received delicious food.” (Shi Shuo Xin Yu, Ya Liang, AD 450)

(6)

闾丘冲优于满奋、郝隆。 (世說新語 品藻) Lǘ Qiūchōng yōu yú Mǎn Fèn, Hǎo Lóng Lü Qiuchong better than Man Fen Hao Long “Lü Qiuchong is better than Man Fen and Hao Long.” (Shi Shuo Xin Yu, AD 450)




The constituent order change in the relative clause and the comparative structure was motivated by the same cause, and this chapter focuses on the diachronic change in the order of the relative clause and its head noun, only briefly discussing the comparative structure. The marker in the comparative structure, yú, was a preposition with multiple functions at the time; thus the order of the phrase “Marker Standard” (i.e. the preposition

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

514

phrase) reflects the correlation between VO and adposition. As discussed in Section 7.9, Chinese grammar underwent such a systematic change: from Middle Chinese onward, all non-resultative PPs had to occur in preverbal position, although most could appear only in postverbal position in Old Chinese. As a result, the yú PP in the old comparative structure disappeared from sentence-final position, and the bǐ PP emerged in preverbal position. Additionally, we demonstrate that the historical change in the position of PPs directly triggered the shift in the constituent order of the relative clause and its head noun. This finding challenges all previous beliefs that the order of the relative clause and the noun is determined by the order of the verb and its argument (i.e. VO or OV). In other words, the present analysis aims to prove the following implicational universals, which will likely come as a great surprise to researchers in linguistic typology: (7)

SVO ⸧ VP PP ⸧ NP REL.

As discussed in detail in Section 7.9, there is robust empirical evidence to justify the above proposal. Both adverbial phrases (e.g. PPs) and relative clauses are marked by the same de form that grammaticalized from the demonstrative in Late Medieval Chinese. Additionally, their applications are governed by the same grammatical rule. That is, the change in position of PPs from postverbal to preverbal position is analogous to the shift of the relative clause from postnominal to prenominal position.

21.2 Changes in the Constituent Order The grammar of any language or the same language at different times represents a unified system, and no features exist in isolation, implying that every feature is related to others. The order “N REL” in Chinese history was connected to other grammatical features at the time. To understand the syntax of the relative clauses in history, therefore, we require a global picture of the related word orders at the time, which are surprisingly distinct from those in Contemporary Chinese, in which modifiers always precede their heads. When the head is a noun, the types of modifiers include adjectives, “number + classifier” phrases, PPs, demonstratives, and genitives, as well as relative clauses; when the head is a verb, the modifiers can be adverbials, PPs, and time and locative phrases; and when the head is an adjective, the modifiers might be degree words or other intensifying adverbials. These constituent orders across grammatical categories can be well explained by the principle of cross-category harmony, by which modifiers tend either to precede their head or to follow it (Hawkins 1983: 97). However, this consistent constituent order in Contemporary Chinese actually distinguishes Chinese from other SVO languages and makes it more similar to the features of SOV languages. According

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Changes in the Constituent Order

515

to Greenberg (1966a), Li and Thompson (1981), and Dryer (2007), the correlations with the relative position of verb and object are as in Table 21.1.6 Contemporary Chinese has all four features in the right-hand column that typically occur with OV languages. At the same time, Chinese has prepositions and its auxiliaries precede the verb, which are two typical features of VO languages, as Li and Thompson (1981) pointed out. Additionally, Li and Thompson (1981: 18) reported the following seven features of Chinese grammar that are supposed to be associated with OV languages:7 (8) (a) (b) (c) (d) (e) (f)

relative clauses precede the head noun, certain adverbials precede the verb, PP precedes the verb, postpositions exist, OV sentences occur, aspect markers follow the verb.

In the subsequent section, in fact, we will see that only feature (9e) has remained the same throughout history, and all six of the other features emerged in the language after the seventh century AD. To illustrate these diachronic changes, we select five types of constituent order in Chinese history that are different from the corresponding structures in Contemporary Chinese.

21.2.1 Titles, Professions, and Locations In Modern Chinese, a surname must precede the title, such as Zhāng Xiānshēng “Zhang Master” and Wáng Nǚshì “Wang Madam,” the order of which is the reverse of the English expressions. However, there were two alternatives in Old Chinese; for instance, in Analects (c. 500 BC), the honorific title zǐ “Master” could either precede the surname Table 21.1 Word order correlations with VO and OV

6 7

VO languages

OV languages

Noun + relative clause Noun + adjective Verb + adverb Verb + PP

Relative clause + noun Adjective + noun Adverb + verb PP + verb

Only the features relevant to the present analysis are listed here. I have reordered the list for the purpose of convenience in understanding the present analysis.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

516

(e.g. Kǒng Zǐ “Kong Master” and Zēng Zǐ “Zeng Master”) or follow it (e.g. Zǐ Gòng “Master Gong” and Zǐ Zhāng “Master Zhang”). Similarly, as with professional designations, in Contemporary Chinese, surnames always precede professional titles, as in Wáng Jiàoshòu “Wang Professor” and Lǐ Lǎoshī “Li Teacher.” In contrast, in Old Chinese, the constituent order was the opposite. For example, in Zhuangzi (c. 300 BC) there were Páo Dīng “Butcher Ding,” Yì Qiū “Chess-player Qiu,” and Jiàng Shí “Craftsman Shi,” which have the same order as in English. Note that, if the modifiers were adjectives, the order was always the opposite: Adj. + professional title, as illustrated in (9): (9)




弈秋, 通國之善弈者也。 (孟子 告子上) Yì Qiū, tōng guó zhī shàn yì zhě yě. Yi Qiu whole country POSS skillful chess-play -er PRT “Yi Qiu is a person who is skillful at playing chess in the whole country.” (Meng Zi, Gao Zi Shang, 300 BC)

If the modifier was a surname, the order was “professional title + surname”; if the modifier was an adjective, the order was “Adj + professional title.” The same contrast is found in English, e.g. Professor Smith and an excellent professor. Moreover, the same constituent order occurred in naming locations. In Modern Chinese, place names always precede political units, such as Běijīng shì “Beijing City.” However, in Old Chinese, the order was the opposite. For example, in the texts of the Zuo Zhuan (550–400 BC), there were Chéng Yǐng “City Ying” and Chéng Pú “City Pu.” Similarly, in Old Chinese, we find expressions such as hé zhōng “river middle” and gǔ zhōng “valley inside,” in which normal nouns precede locative nouns. In Contemporary Chinese, the constituent order is exactly the opposite. Apparently, the order of the modifier and the head noun in Old Chinese was different from that in Contemporary Chinese. It remains unclear what motivated the change in the above constituent orders within nominal phrases. 21.2.2 Head Noun and “Num + CL” Phrases In Contemporary Chinese, the “Num + CL” phrase always precedes the head noun, and a proper classifier must be used to connect the number to the head noun, such as sān běn shū “three Cl book” and wǔ gèrén “five Cl person.” However, this grammatical structure was not firmly established until the thirteenth century AD. According to Zhang (2001), in Early Old Chinese the number modifier generally followed the head noun instead, as illustrated in (10):

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Changes in the Constituent Order (10)

517

賜彤弓一, 彤矢百, 馬四匹。 (西周銘文) Cì tóng gōng yī, tóng shǐ bǎi, mǎ sì pǐ. bestow red bow one red arrow hundred horse four CL “(He) was bestowed with one red bow, one hundred red arrows, and four horses.” (Wang 1989: 25, Xi Zhou Ming Wen, 1100 BC)

In Old Chinese, the situation changed: numbers generally preceded the head noun; although classifiers were mostly not needed, if there was a classifier, the “Num + CL” phrase had to occur in postnominal position, namely following the head noun: (11)




子產以帷幕九張行。 (左傳 昭公十三年) Zǐ Chǎn yǐ wéimù jiǔ-zhāng xíng. Zi Chan take certain nine-CL leave “Zi Chan took nine curtains and left.” (Zuo Zhuan, Zhao Gong Shi San Nian, 550–400 BC)

Classifiers grammaticalized from ordinary nouns over time – a change that can be traced back to as early as the first century BC (Wang 1989: 18). It was in postnominal position that ordinary nouns gradually developed into classifiers, and, once they were fully grammaticalized, “Num + Cl” phrases started to occur before the head noun (Shi 2016: 676‒697).

21.2.3 Adjectival Modifier and Head Noun When used as a modifier, in Contemporary Chinese, the adjective must precede the head noun, regardless of the structural complexity (i.e. weight) of the adjectival phrase. If the adjective is simple, the associative morpheme de is optional in connecting the modifier and the head, as in dà shù “big tree” and dà de shù “big ASSOC tree.” If the adjective is quantified or intensified by reduplication, degree words, and the like, the associative morpheme de becomes obligatory to connect the adjectival modifier and the head, as illustrated in (12): (12)

好好的衣服 hǎohǎo de yīfú 很貴的車 hěn guì de chē “quite good clothes” “very expensive car”

*好好衣服 *hǎohǎo [ ] yīfú *很貴車 *hěn guì [ ] chē

However, there were two alternative orders in Old Chinese: adjectival modifiers could either precede or follow the head noun, as illustrated in the following examples. In these cases, the adjectival phrases in postnominal position were often marked by the relativizer zhě, signaling that Old Chinese treated complex adjectival modifiers as a type of relative clause.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

518




(13)

良庖歲更刀。 (莊子 養生主) Liáng páo suì gēng dāo. fine butcher year replace knife “A fine butcher replaces his knife every year.” (Zhuang Zi, Yang Sheng Zhu, 300 BC)

(14)

今夫士之高者乃稱匹夫。 (戰國策 齊策) Jīn fū shì zhī 8 gāo zhě, nǎi chēng pǐfū. now PRT intellectual ASSO excellent REL then call ordinary-man “Now excellent intellectuals are called ordinary men.” (Zhan Guo Ce, Qi Ce, 450–200 BC)

(15)

不如視諸王最賢者立之。 (史記 呂太后本紀) Bù rú shì zhū-wáng zuì xián zhě lì zhī. no as see dukes most virtuous REL make-king him “It is not as good as to find the one who is most virtuous among many dukes and to make him king.” (Shi Ji, Lü Tai Hou Ben Ji, 100 BC)







When adjectives preceded the noun, grammatical markers were not needed to associate them. When adjectives followed the head noun, the morpheme zhī optionally occurred between them, as illustrated in (15), but the morpheme zhě was necessary to mark the postnominal adjectival modifiers, as illustrated in (16). In Old Chinese, the morpheme zhī was the most widely used grammatical marker in adjectives, genitives, and relative clauses – a very common cross-linguistic phenomenon (Aristar 1991). Note that the morpheme zhī would occur between the head and the modifier regardless of whether it was “modifier–head” or “head– modifier” order, although the former was more common than the latter. According to Dryer (1992), adjectives across languages are treated as relative clauses; thus the morpheme zhě was also an obligatory marker for postnominal relative clauses – a point that we discuss in great detail in Section 21.3.

21.2.4 Adverbial Phrases in Sentence-Final Position In Modern Chinese, adverbial phrases always precede the matrix verb, e.g. bái lái-le “in-vain come PERF” and fēicháng xǐhuān “extremely love.” However, in the history of

8

The morpheme zhī in Old Chinese, which grammaticalized from a demonstrative, has multiple functions, similar to the later marker de, which is used to indicate an associative or possessive relationship between a modifier and its head noun. It often introduced a postnominal modifier, a function that the later de lacked.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Changes in the Constituent Order

519

Chinese, adverbial phrases appeared relatively freely in sentence-final position, as illustrated in the following examples:




(16)

仲尼不為已甚者。 (孟子 離婁下) Zhòng Ní bù wéi yǐ shèn zhě. Zhong Ni not benefit already much PRT “Zhong Ni (Confucius) did not act very much.” (Meng Zi, Li Lou Xia, 300 BC)

(17)

後箭射人深。 (祖堂集 藥山和尚) Hòu jiàn shè rén shēn. latter arrow shoot person deeply “The latter arrow shot a person deeply.” (Zu Tang Ji, Yue Shan He Shang, AD 950)




In Contemporary Chinese, however, adverbial phrases are not allowed to appear in sentence-final position, and they must occur somewhere in preverbal position, either at the beginning of a sentence or between the subject and the predicate.

21.2.5 The Prepositional Modifier and Its Head Noun A PP can be used as a modifier to delineate the referents of the NP or the VP. In English, PPs can only follow the head NP, such as the chair in the classroom or the cook in the kitchen. In contrast, in Contemporary Chinese, the PP always precedes the head;9 when modifying a nominal head, PPs behave like relative clauses, which must be marked by the relativizer de, as illustrated in (19): (18)

(a) 在教室裡的學生 (現代漢語) zài jiàoshì lǐ de xuéshēng in classroom inside REL student “the students in the classroom” (b) *在教室裡學生 *zài jiàoshì lǐ [ ] xuéshēng in classroom inside student (Contemporary Chinese)

In Old Chinese, however, the PPs used as modifiers usually followed the head noun, which was marked by the relativizer zhě – the same marker for postnominal relative clauses (we discuss this issue in detail in Section 21.3):

9

If a PP expresses the resultative state of the verb, it still follows the VP, a point we will discuss in Section 21.3.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

520




(19)

晉人以宋五大夫在彭城者歸。 (左傳 襄公元年) Jìn rén yǐ Sòng wǔ dàfū zài Péng chéng zhě guī. Jin people take Song five official in Peng city REL return “The Jin people took the five officials of the Song country who were in Peng City and left for home.” (Zuo Zhuan, Xiang Gong Yuan Nian, 550–400 BC)

(20)

下士與庶人在官者同祿。 (孟子 萬章下) Xià shì yǔ shù-rén zài guān zhě tóng lù. low intellectual and ordinary-people in government REL equal salary “Low-level intellectuals and ordinary people who are serving the government have equal salaries.” (Meng Zi, Wan Zhang Xia, 300 BC)

(21)

諸侯子在關中者儘集栎阳为卫。 (史記 高祖本紀) Zhū-hóu zǐ zài Guānzhōng zhě jiē jí Lìyáng wéi Those-dukes sons in central-Shaanxi REL all meet Liyang be wèi. guard. “The sons of those dukes who lived in central Shaanxi all met in Liyang to be guards.” (Shi Ji, Gao Zu Ben Ji, 100 BC)







That is, when used as modifiers, PPs and relative clauses were treated as having the same grammatical status at the time, both following the head noun and being marked by the relativizer zhě. From the Modern Chinese period to the present day, both have been restricted to prenominal position. In summary, the constituent orders in Old Chinese were: (a) noun–adjective, (b) verb– adverb and (c) verb–PP; historically, more importantly, there was also noun–relative clause, as discussed below. That is, the constituent orders of many structures in Old Chinese were systematically different from those in Contemporary Chinese. Considering all four of the major VO and OV language features listed in Table 21.1, it is clear that the Chinese language historically was much more consistent with VO languages than Contemporary Chinese is. In other words, the grammatical system of Contemporary Chinese fundamentally deviates from the typical properties of SVO languages. The focus of this chapter is on the motivation of the historical typological change from “NP-REL” to “REL-NP.”

21.3 The Relative Clause and the Head Noun in Old Chinese Before discussing the historical syntax of relative clauses in Chinese, we should clarify the working definition of the concept of a “relative clause” used in the present analysis.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

The Relative Clause and the Head Noun in Old Chinese

521

The markers and structures of relative clauses vary from one language to another; some languages take finite declarative sentences as their relative clauses, but relative clauses in other languages might be nonfinite verbs, which is similar to nominalization, as in Turkish (for details, see Comrie 1981: 142). Chinese does not have morphological markers to distinguish finite from nonfinite clauses; hence we must use a functional definition of relative clauses, as suggested by Comrie (1981: 143): A relative clause then consists necessarily of a head and a restricting clause. The head in itself has a certain potential range of referents, but the restricting clause restricts this set by giving a proposition that must be true of the actual referents of the over-all construction. According to the above definition, Modern Chinese has only one type of relative clause, which can be formulized as “REL de N.” The morpheme de is the only relativizer necessary to associate the relative clause with the head noun, and relative clauses must precede the noun: (22)

這是我種的樹。 (現代漢語) Zhè shì wǒ zhòng de shù. these be I plant REL tree “These are the trees that I planted.” (Contemporary Chinese)

(23)

種樹的人來了。 (現代漢語) Zhòng shù de rén lái le. plant tree REL people come PERF “Those tree-planting people have come.” (Contemporary Chinese)

21.3.1 Positions of Relative Clauses The above two examples actually represent two subtypes of relative clause in Old Chinese for which distributions differed. The type of relative clause in (22) took the order in (24a), but the type in (23) adopted the order in (24b): (24) (a) NP + REL + zhě, (b) REL + zhī + NP. In the pattern of (24a), relative clauses followed the noun, which was necessarily marked by the clause-final relativizer zhě. In the pattern of (24b), relative clauses instead preceded the noun, which was necessarily associated with the marker zhī. There was

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

522

a basic division of labor between the two patterns: roughly, the relative clause in (24a) expressed a real event that had already happened, and that in (24b) referred to a statement about functions, properties, or other characteristics. Therefore the two types of relative clause, as exemplified in (22) and (23), respectively, actually corresponded to two different structures in Old Chinese: (22) corresponded to (24a), and (23) corresponded to (24b). The major division of labor between the two relativizers zhī and zhě was that the head noun of the zhě clause had to be the subject of the clause, which is typically a human noun, but the head noun of the zhī clause could be the object of the clause, which might be a patient argument, a temporal expression, a locative argument, an instrument argument, and so on. According to Croft (1990), many languages, including English, have nonfinite forms (participles) that function as relative clauses, allowing only subject NP heads, as in the child playing in the bedroom. The zhě clause in Old Chinese behaved like these nonfinite relative clauses in English. The distinctive distributions of the zhī and zhě clauses were related to their lexical sources. Both zhī and zhě were demonstratives in Old Chinese, but zhī always preceded a head noun and zhě typically followed a head noun. According to Heine and Kuteva (2002: 113‒115), relative clause markers are also frequently derived from demonstratives. There is a cross-linguistic tendency: if a language has a noun before a demonstrative, then it has a noun before a relative clause (Hawkins 1983: 84). This tendency implies that, if a language has a noun after a demonstrative, then it has a noun after a relative clause, which explains why the two relativizers differed syntactically.




(25)

之二臣者, 甚相憎也。 (呂氏春秋 慎勢) Zhī èr chén zhě, shén xiāng zēng yě. this two minister PRT very each-other hate PRT “The two ministers hate each other very much.” (Lü Shi Chun Qiu, Shen Shi, 250 BC)

(26)

北山愚公者, 年且九十。 (列子 湯問) Běi shān yú-gōng zhě, nián qiě jiǔ-shí. north mountain foolish-man this age near ninety “The foolish man in the north mountain is nearly ninety years old.” (Lie Zi, Tang Wen, 400 BC)




As (25) shows, the demonstrative zhī always preceded its head noun, but the demonstrative zhě often occurred after its head noun when used as an anaphor coindexing the preceding noun. The example in (26) is the typical structure of copular sentences in Old Chinese, in which the demonstrative zhě was used as an anaphor to coindex the subject

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

The Relative Clause and the Head Noun in Old Chinese

523

(Wang 1989: 65). These distinctive contexts for their grammaticalization into relativizers were responsible for the different distributions, prenominal or postnominal. In Early Old Chinese, there were no overt relativizers to mark a relative clause and a VP was simply placed prior to the head noun to delineate the referents of the noun (Chen 1956: 133, Gao 1987: 283, Serruys 1981: 356, Zhang 2001):10 (27)

王省武王成王伐商圖。 (西周銘文) Wáng xǐng Wǔ wáng Chéng wáng fá Shāng tú. King look-at Wu king Cheng king expedition-against Shang picture “The king inspected the picture of King Wu and Cheng leading an expedition against the Shang country.” (Xi Zhou Ming Wen, 1100 BC)

New patterns for relative clauses came into being around the fifth century BC. At that time, the morpheme zhě could function as a complementizer to introduce an object clause, as illustrated in (28): (28)




漢魏父子兄弟接踵而死於秦者百世矣。 (戰國策 秦策) Hàn Wèi fùzǐ xiōngdì jiēzhǒng ér sǐ yú Qín zhě Han Wei father-son brothers many then die in Qin COM bǎi shì yǐ. hundred generation already “It has been hundreds of generations since many fathers, sons, and brothers of the Han and Wei countries died in the Qin country.” (Zhan Guo Ce, Qin Ce, 450–200 BC)

The morpheme zhě in Old Chinese behaved exactly like that in English, which has multiple functions, such as demonstrative, relativizer, complementizer, and anaphora; a comprehensive discussion of its various usages is beyond the scope of this chapter. In what follows, I concentrate on the syntax of its relativizer use.

21.3.2 The Syntax of the Zhě Relative Clause The syntax of relative clauses in any language can be examined from two perspectives: first, what role the head noun plays in the clause, such as subject or object, and second, which noun of a declarative sentence can be relativized. As mentioned previously, in Old Chinese, the head noun can only be the subject of the relative clause. However,

10

According to Greenberg (1966a), this pattern is commonly found in verb-final languages. However, Old Chinese was a typical SVO language. In Section 21.5, we argue that the relative order of verb and object has no direct effect on the order of the relative clause and noun.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

524

various nouns in different syntactic positions can be relativized. Keenan and Comrie (1979) presented the well-known accessibility hierarchy as follows: (29) Subject > Direct Object > Indirect Object > Oblique Object > Genitive Noun > Object of Comparison. The above formula indicates that in any language the primary relative clauseforming strategy must apply to the top segment of the hierarchy. In other words, if it applies to any position in the hierarchy, it must also apply to all of the positions above this position. In Old Chinese, the postnominal relative clause marked by zhě could apply to the nouns in all of the syntactic positions listed in (30) (Zhu 1983, Wang 1989). (a) Subject. This structure is the most common “NP + REL” structure from Old to Medieval Chinese. In the following examples, the parts in bold font are the subjects, and the underlined parts are the relative clauses marked by the relativizer zhě. Since these clauses describe a real event that has already happened, they can occur only in postnominal position, marked by zhě. In these cases, the prenominal relative clause marked by zhī could not be used since it could not introduce a relative clause portraying a concrete and real event, as mentioned previously. We saw one such example in (5); two more examples follow:




(30)

齊人之井飲者相捽也。 (莊子 列御寇) Qí rén zhī jǐng yǐn zhě xiàng zuó yě. Qi people POSS well drink REL each-other seize PRT “The Qi people who drank at the well seized each other.” (Zhuang Zi, Lie Yu Kou, 300 BC)

(31)

客新有從山東來者曰蔡澤。 (史記 范睢蔡澤列傳) Kè xīn yǒu cóng Shāndōng lái zhě yuē Cài Zé. Guest recently have from Shandong come REL call Cai Ze. “The guest who recently came from Shandong is called Cai Ze.” (Shi Ji, Fan Sui Cai Ze Lie Zhuan, 100 BC)




Throughout the history of the Chinese language, there has been a special type of sentential structure in which a “subject + predicate” phrase can be used directly as the predicate of a higher-level sentence without being connected by any verbal element, such as a copula, a structure that is rarely attested in other languages (Zhu 1982: 100). In (32), mù “tree” is the main subject, which is modified by the relative clause shí fán zhě “whose fruits are too numerous,” and the predicate is another clause consisting of a subject (i.e. zhī “branch”) and a predicate (i.e. bì pī “must crack”):

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

The Relative Clause and the Head Noun in Old Chinese (32)

525




木實繁者枝必披。 (戰國策 秦策) Mù shí fán zhě zhī bì pī. tree fruit numerous REL branch must crack “The tree whose fruits are too numerous must have cracks in its branches.” (Zhan Guo Ce, Qin Ce, 450–200 BC)

(b) Object. According to my investigation, a relative clause that did not describe a concrete event could also be used in the postnominal position, marked by the relativizer zhě; for instance, the relative clause in (33) states the possibility “who is likely to give birth to a boy.” In other words, the scope of the zhě relative clause is broader than that of the zhī one: the former can introduce either a real clause or a subjunctive clause in postnominal position, but the latter can introduce only a subjunctive clause in prenominal position. Occasionally, the possessive marker zhī can occur between the head noun and the zhě relative clause, forming a pair of markers “zhī . . . zhě” for the relative clause, as illustrated in (34) and (35), respectively. This is further evidence of the cross-linguistically intrinsic relationship between genitives and relative clauses (Aristar 1991).




(33)

求婦人宜子者進之。 (戰國策 楚策) Qiú fùrén yí zǐ zhě jìn zhī. seek woman likely-bear boy REL send she “(They) sought women who were likely to give birth to a boy and sent them (to the emperor).” (Zhan Guo Ce, Chu Ce, 450–200 BC)

(34)

樂毅複以兵平齊城之不下者。 (史記 樂毅列傳) Yuè Yì fù yǐ bīng píng Qí chéng zhī bù xià zhě. Yue Yi again use army quell Qi city POSS not surrender REL “Yue Yi again took his armies to quell the Qi cities that had not surrendered.” (Shi Ji, Yue Yi Lie Zhuan, 100 BC)

(35)

(嚴仲子) 遊求人可以報俠累者。 (史記 刺客列傳) (Yán Zhòngzǐ) yóu qiú rén kěyǐ bào Xiá Lèi zhě. Yan Zhongzi travel seek person able revenge Xia Lei REL “(Yan Zhongzi) traveled around to seek the person who could avenge Xia Lei.” (Shi Ji, Ci Ke Lie Zhuan, 100 BC)







Historically, a noun that is the complement of a copular structure can also be modified by a zhě relative clause, as illustrated in (36). Note that there was no copular verb in Old

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

526

Chinese; hence the sentence-final particle yě was obligatory to mark the copular structure. Between the subject and the complement was a pause, and the subject was often coindexed by the anaphoric pronoun zhě (i.e. one of the multiple functions of the pronominal element zhě at that time). The copular verb shì, which is between the subject and the complement in Contemporary Chinese, did not come into being until the first century BC (Wang 1989: 183). (36)




晉士之送葬者歸以語史趙。 (左傳 昭公十一年) Jìn shì zhī song zàng zhě guī yǐ yǔ Shǐ Zhào. Jin people GEN attend funeral REL return with tell Shi Zhao “The Jin people who attended the funeral returned to tell Shi Zhao.” (Zuo Zhuan, Zhao Gong Shi Yi Nian, 550–400 BC)

(c) Oblique object. It seems that the zhě relative clause can be applied to any position in terms of the accessibility hierarchy identified by Keenan and Comrie (1979). The noun in PP, namely the object of a preposition, can also be modified by a zhě relative clause. In (37), for instance, the head noun bīnkè “guest” is the object of the preposition jí “with,” which is modified by the zhě relative clause “who knew the mission”:




(37)

太子及賓客知其事者皆白衣冠而送之。 (史記 刺客列傳) Tàizǐ jí bīnkè zhī qí shì zhě jiē bái yī guàn yǐ prince with guest know the mission REL all white clothes hat with sòng zhī. send he “The prince, together with his guests who knew the mission, all wore white hats and clothes and saw him (i.e. the hero) off.” (Shi Ji, Ci Ke Lie Zhuan, 100 BC)

(38)

殯諸伯有之臣在市側者。 (左傳 襄公三十年) Bìn zhū Bó Yǒu zhī chén zài shì cè zhě. hold-wake-for along Bo You POSS courtier live city side REL “(They) held a wake for his attendants who lived at the side of the town.” (Zuo Zhuan, Xiang Gong San Shi Nian, 550–400 BC)




In (38), the zhě relative clause modifies a genitive phrase that is the object of the preposition zhū “among.” It reveals another important rule governing the application of the zhě relative clause: if the head noun was preceded by another modifier, such as an adjective, a possessor, or the zhī relative clause, only the postnominal relative clause marked by zhě was allowed. The head in (39) is a genitive phrase and hence cannot be modified by a preverbal relative clause; more examples of this type are analyzed as follows.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

The Relative Clause and the Head Noun in Old Chinese

527

21.3.3 Usage of the Zhě Relative Clause In Contemporary Chinese, the relative clause marked by de is restricted to occur in prenominal position, regardless of whether the head noun has other attributives, as illustrated in (39). In fact, all attributives must precede the head noun, unlike the grammar of Old Chinese, which allowed various types of modifier to follow the head noun, as discussed in Section 21.2. (39)

這是我昨天買的一本很有趣的書。 (現代漢語) Zhè shì wǒ zuótiān mǎi de yī-běn hěn yǒuqù de shū. this be I yesterday buy REL one-CL very interesting ASSO book “This is a very interesting book that I bought yesterday.” (Contemporary Chinese)

In (39), all three modifiers – an adjectival phrase, a numeral phrase, and a relative clause – precede the head noun shū “book,” but this structure was not allowed in Chinese before the fifth century AD. Once there was another modifier in prenominal position, regardless of the grammatical category to which it belonged, the relative clause marked by zhě was the only remaining choice; in other words, the prenominal relative clause marked by zhī was inapplicable in this context. This rule can be illustrated by the following examples. First, in (40) the head noun dú “calf” was already modified by an adjective, so a prenominal relative clause marked by zhī could no longer be used, and the zhě relative clause in postnominal position was the only option: (40)




南門之外有黃犢食苗道左者。 (韓非子 內儲說上) Nán mén zhī wài yǒu huáng dú shí miáo dào-zuǒ south gate GEN outside have yellow calf eat seedling road-left zhě. REL “Outside the South Gate, there is a yellow calf who eats the seedlings on the left side of the road.” (Han Fei Zi, Nei Chu Shuo Shang, 300 BC)

Second, in (41), the head noun fù “woman” was preceded by a verbal modifier (equivalent to the past participle in English), so the relative clause could be introduced only by the relativizer zhě in postnominal position: (41)




出婦嫁於鄉里者, 善婦也。 (戰國策 秦策) Chū fù jià yú xiānglǐ zhě, shàn fù yě. divorce woman marry in neighborhood REL virtuous women PRT “The divorced woman who is married (by someone) in the neighborhood is a virtuous one.” (Zhan Guo Ce, Qin Ce, 450–200 BC)

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

528

Third, if the head noun was preceded by a possessor modifier, once again, the zhě relative clause in the postnominal position was the only option for relativizing the head noun, as illustrated in (42): (42)




諸侯之賓問疾者都在。 (左傳 昭公二十年) Zhūhóu zhī bīn wèn jí zhě duō zài. duke POSS guest inquiry illness REL many alive “The dukes’ guests who inquired about his illness are mostly alive.” (Zuo Zhuan, Zhao Gong Er Shi Nian, 550–400 BC)

Fourth, if there was already a relative clause in prenominal position, another relative clause could be used only in postnominal position, which was marked by the relativizer zhě. In (43), a relative clause (i.e. Shí Huáng suǒ fēng “that the emperor wrote”) precedes the head noun (i.e. shū “the imperial edict”); hence another clause (i.e. cì gōngzǐ Fúsū zhě “to bestow upon Prince Fusu”) can only follow the head noun: (43)




丞相斯陰謀破去始皇所封書賜公子扶蘇這。 (史記 秦始皇本紀) Chéngxiàng Sī yīnmóu pòqù chief-minister Si secretly destroy Shí Huáng suǒ fēng shū cì gōngzǐ Fúsū zhě. first emperor SUO write imperial-edict bestow son-of-emperor Fusu REL “Chief Minister Si secretly destroyed the imperial edict that the emperor wrote to bestow upon Prince Fusu.” (Shi Ji, Qin Shi Huang Ben Ji, 100 BC)

In fact, neither the zhě relative clause nor the zhī relative clause could be recursive within the NP; that is, neither of them could be used twice there. In Contemporary Chinese, in contrast, the de relative clause can be recursive in prenominal position, indicating that it can be used more than once within an NP, as illustrated in (44): (44)

這是我買的日本出產的車。 (現代漢語) Zhè shì wǒ mǎi de Rìběn chūchǎn de chē. this be I buy REL Japan make REL car “This is the car that was made in Japan and that I bought.” (Contemporary Chinese)

From the point of view of language universals, we can see that Old Chinese actually obeyed the heavy-constituent principle, by which longer, more complex modifiers often follow the head, whereas other lighter modifiers precede the head. Hawkins (1983: 98‒106) offered a similar pragmatic explanation for the heaviness serialization principle, which defines a preference for “lighter” modifiers to occur leftward in their

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Typological Change in the Relative Clause Structure

529

respective phrases, while “heavier” modifiers tend to be placed rightward. Preposed long modifiers make it more difficult to determine the head; once the head has been identified, a subsequent long modifier (e.g. relative clause or adjectival phrase) is more easily processed. Thus category harmony and heaviness shift aid language users in unpacking the constituent structure. This concept resembles Greenberg’s concept of dominance in its effect of complementing harmony: heavier modifiers follow the noun even if the harmonic order is modifier–noun. The rank of heaviness is as follows: relative clause > genitive noun > adjective > demonstrative/pronoun. That is, the grammar of Chinese historically exhibited more consistency with certain language universals. The robust empirical evidence provided above is sufficient to demonstrate that the postnominal relative clause not only was well attested but also played a key role in the grammar. In the next section, I attempt to find the answer to why this type of relative clause disappeared.

21.4 Typological Change in the Relative Clause Structure Roughly after the sixth century AD, postnominal relative clauses gradually became outdated, and the Chinese language underwent a typological change from “N REL” to “REL N.” This change was by no means an isolated case but was related to a development of the grammar around this time. From the beginning of the Medieval Chinese period, the language underwent a systematic change: all PPs and adverbials that did not express the resultatives of the predicate verbs disappeared from sentence-final position, while their functions were taken over by a set of new functional words that grammaticalized in preverbal position. Regarding constituent orders, the syntactic position of PPs and other adverbials underwent a development parallel to that of relative clauses, which can be formulated as follows: (45)

(a) VP + (PP/Adverbial)non-resultative > (PP/Adverbial)non-resultative + VP, (b) NP + Relative Clause > Relative Clause + NP.

In my view, the above two changes could not be coincidental but reveal an intrinsic relationship between them. Actually, they represent two outcomes of a global change. Specifically, relative clauses function to delineate the referents of NP, and PPs/adverbials serve to narrow the potential referents of the VP. For example, read the newspaper in the library refers to an event different from read the newspaper in the kitchen, and come early and come late express two different events. The grammatical relationship between an adverbial phrase (including a PP) and the head verb is the same as that between a relative clause and its head noun. This fact is supported by solid

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

530

morphological evidence in the history of Chinese. Since Late Medieval Chinese, both adverbial phrases (including clause-like forms) and relative clauses have been marked by the same morphological morpheme, de, which grammaticalized from a demonstrative around the seventh century AD. This finding indicates that adverbial phrases and relative clauses are regarded as the same grammatical category at a quite abstract level in Chinese. In what follows, let us first discuss the systematic change in Chinese grammar and then illustrate the morphological evidence with examples.

21.4.1 The New Information-Organizing Principle As we saw in Chapter 7, the structures of declarative sentences underwent a fundamental change in Medieval Chinese, which caused the redistribution of PPs (i.e. adjuncts). It is observed in the following two syntactic configurations: (46)

Subj + PPnon-resultative + VP > Subj + PPnon-resultative + VP.

The motivation for this schematic change was the establishment of the resultative construction, which in turn triggered the occurrence of the new principle of information structure (for a fuller discussion, see Section 7.9). Under this principle, only PPs denoting resultatives, such as the end point or recipient, remained in postverbal position, whereas PP adjuncts expressing non-resultative features could no longer stay in the original postverbal position and thus became extinct. Note that a group of verbs grammaticalized in the first verbal position of a serial verb structure and assumed the function of the old grammatical markers. That is, the schematic change did not involve any movement at all. This process started as early as 100 BC and was not completed until AD 1800, lasting nearly 2,000 years. The position changes of these PPs in sentence-final position to delineate the preceding VPs is critical for the present analysis. In Old Chinese, as in English and other VO languages, PPs usually appeared in sentence-final position to modify the preceding VPs in various ways. As typological studies (Greenberg 1966a, Comrie 1981: 86‒103, Croft 1990: 208‒213) have shown, it is universal that SVO languages allow PPs to occur freely in sentence-final position, modifying the predicate. That is, the grammar of Old Chinese was perfectively consistent with this language universal. The PPs could express the agent for passives, the standard for comparison, the instrument for action, the location for events, and so on. However, all of these PPs could appear only in preverbal position except for those expressing some types of resultatives of verbs (for a fuller discussion, see Section 7.9).

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Typological Change in the Relative Clause Structure

531

(47)

Comparative structure: 而越大於少康。 (左傳 哀公元年) Ér Yuè dà yú Shǎokāng. but Yue big than Shaokang “However, the Yue country is bigger than the Shaokang country.” (Zuo Zhuan, Ai Gong Yuan Nian, 550–400 BC)

(48)

Locative structure: 公會齊侯於城濮。 (左傳 定公十年) Gōng huì Qí hóu yú chéng Pú. Duke meet Qi marquis in city Pu “The duke met the marquis of the Qi country in Pu City.” (Zuo Zhuan, Ding Gong Shi Nian, 550–400 BC)

(49)

Instrumental structure: 五畝之宅, 樹之以桑。 (孟子 梁惠王) Wǔ mǔ zhī zhái, shù zhī yǐ sāng. five acre GEN house plant it with mulberry “Houses and grounds covering five acres are planted with mulberries.” (Meng Zi, Liang Hui Wang, 300 BC)










Obviously, PPs in passives, comparatives, instrumentals, and locatives do not express any resultative senses for the verb. Under the impact of the new information-organizing principle, all of these functional constructions were grouped together to form a construction schema and underwent the same path of change (for details, see Chapter 7). However, only a few PPs that express the resultative state of the action remain in postverbal position (Chao 1979: 176‒180), as illustrated in (50): (50)

約翰寄了一個包裹給瑪麗。 (現代漢語) Yuēhàn jì-le yī-gè bāoguǒ gěi Mǎlì. John mail-PERF one-CL parcel to Mary “John mailed one parcel to Mary.” (Contemporary Chinese)

In example (50), “Mary” is the destination (end point) of the “mail” action, a type of resultative state. Thus the structure in (50) remains grammatical in Contemporary Chinese. Although some locative PPs can occur in either the preverbal or postverbal position, they produce two very different interpretations as a result of the analogizing effects of the resultative construction. The distribution of locative PPs is an ideal window through which to see how the information-organizing principle works in Modern Chinese

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

532

grammar. Some locative PPs can occur in either preverbal or postverbal position, but they produce completely different interpretations.

21.4.2 The Grammaticalization of the Marker De Motivated by the emergence of the classifier system, the demonstrative dǐ grammaticalized as an associative particle to connect modifiers to its head noun (for details, see Chapter 20).11 In addition to “Num + CL” phrases, in Chinese, the modifiers of nouns could be adjectives, genitive nouns, and relative clauses. As mentioned previously, in Old Chinese, numbers directly preceded their head nouns without any intervening classifiers. When, in Medieval Chinese, the classifier system emerged, some of which came into being as early as Late Old Chinese, a proper classifier had to be used to associate a number with its head noun. Consequently, the structure of numeral NPs underwent the following change: (51)

Num + N > Num + CL + N.

In Late Medieval Chinese, by analogy, the demonstrative dǐ was triggered to develop into a morpheme to connect all of the other types of attributive, except for numbers, to the head noun. This new morpheme, de (a phonologically reduced form of the demonstrative dǐ), took over the functions of both zhī and zhě, but it was restricted to occurring in prenominal position. The earliest examples show that the new marker de first connected a relative clause with its head noun, as illustrated in (52) and (53), and then was extended to introduce other types of attributive:

11




(52)

石頭上坐底僧, 若是昨來底後生。 (祖堂集 石頭和尚) Shítóu shàng zuò de sēng, ruò shì zuó lái de stone on sit REL monk like-be yesterday come REL hòushēng. young-man “The monk who is sitting on the stone looks like the young man who came yesterday.” (Zu Tang Ji, Shi Tou He Shang, AD 950)

(53)

將飯與闍黎喫底人還有眼也無? (祖堂集 丹霞和尚) Jiāng fàn yǔ shélí chī de rén hái yǒu yǎn yě wú? take food give monk eat REL person still have eye PRT lack “Did the person who fed the monks still have eyes or not?” (Zu Tang Ji, Dan Xia He Shang, AD 950)




The full phonological form of the demonstrative dǐ was [tiei] with a rising tone in Medieval Chinese. Due to the phonological reduction of the grammaticalization, it became [tǝ] without tonal value in Modern Chinese.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Typological Change in the Relative Clause Structure

533

Beyond the usages in nominal phrases, the marker de could also connect a complex adverbial phrase, such as a reduplicated form, to its verb head (i.e. predicate), as illustrated in (54) and (55).12 This indicates that in Chinese the structures “Modifier + Head,” regardless of whether the head is nominal or verbal, are treated the same, at least morphologically.




(54)

忽然堂堂底坐。 (祖堂集 鲁祖和尚) Hūrán tang-táng de zuò. suddenly impressive-impressive ADV sit “(He) is impressively sitting there suddenly.” (Zu Tang Ji, Lu Zu He Shang, AD 950)

(55)

但自家這個意思長長底新。 (朱子語類卷十四) Dàn zìjiā zhè-gè yìsī cháng-cháng de xīn. only self this-CL idea long-long ADV fresh “Just keep your own ideas fresh for a long time.” (Zhu Zi Yu Lei, Chapter 14, AD 1200)

According to He (2007), in the texts of Early Modern Chinese (roughly from AD 1200 to 1400), the morpheme de had to be used to associate the adverbial phrase with the predicate verb when the adverbials were a reduplication of adjectives or a complex VP, as illustrated in (56) and (57), respectively. In the following examples, the underlined parts are the adverbial phrases and the parts in bold font are the predicates, with the former parts functioning to modify the latter:




(56)

原來你深深地花底將身兒遮。 (元刊雜劇 閨怨佳人拜月亭) Yuánlái nǐ shēn-shēn de huā-dǐ jiāng shēnér zhē. so you deep-deep ADV flower-under take body cover “So you covered your body under the flowers.” (Yuan Kan Za Ju, Gui Yuan Jia Ren Bai Yue Ting, AD 1300)

(57)

逃席的走來家。 (元刊雜劇 詐妮子調風月) Táo xí de zǒu-lái jiā. run-away banquet ADV go-come home “The one who ran away from the banquet went home.” (Yuan Kan Za Ju, Zha Ni Zi Tiao Feng Yue, AD 1300)




More importantly, in Contemporary Chinese the rule regarding the usage of the marker de is equally applied to both nominal and verbal modifiers (Lü 1999). For example, the 12

Note that in Modern Chinese, the morpheme de introducing an adverbial phrase is written with a different Chinese character. However, both morphemes possess the same phonological form in Modern Chinese and historically originated from the same lexical source.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

534

morpheme de is optional when the nominal and verbal modifiers are simple adjectives, but it is obligatory when the modifiers are “degree word + adjective” phrases, regardless of whether the head is a noun or a verb (cf. Chapter 20), as illustrated in (58) and (59), respectively:13 (58)

(a) 很好的朋友 (現代漢語) Hěn hǎo de péngyǒu. very good ASSO friend “a very good friend.” (b) *很好朋友 *Hěn hǎo [ ] péngyǒu very good friend (Contemporary Chinese)

(59)

(a) 很好地完成了任務。(現代漢語) Hěn hǎo de wánchéng-le rènwù. very good ADV finish-PERF task “(They) have finished the task very well.” (b) *很好完成了任務。(現代漢語) *Hěn hǎo [ ] wánchéng-le rènwù very good finish-PERF task (Contemporary Chinese)

In traditional Chinese linguistics, PPs are regarded as adverbials because they share the functions and distributions of ordinary adverbs (e.g. Chao 1979: 330, Zhu 1982: 151, Lü 1979: 64). Indeed, PPs used as adverbials are not marked by any morphological forms. Undoubtedly, however, the change in PP-adverbials from the postverbal (sentence-final position) to preverbal position was parallel to the change in relative clauses from postnominal to prenominal position. This diachronic evidence proves that the relationship between the noun and relative clause is the same as that between the verb and the PP-adverbial; in other words, the relative order of O and V bears no direct relationship to the order of the noun and relative clause, as many typological linguists have assumed (Greenberg 1966a, Dryer 1992, Croft 1990: 47‒49). That is, my analysis reveals that there is no correlation between the two constituent orders: VO and N + REL, and, at least in Chinese, the typological change from “V + PP” to “PP + V” motivated the shift from “N + REL” to “REL + V,” an issue to which I return in subsequent sections. 13

In English, an adjective usually needs to have a suffix -ly (e.g. quick versus quickly) added when used to modify a verb. In Chinese, however, a vast majority of adjectives can be directly used to modify a verb without adding any suffix, e.g. kuài pǎo “quick run.”

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Typological Change in the Relative Clause Structure

535

21.4.3 Consistency of Old Chinese with VO Languages According to Dryer (2003), the “REL N” order in VO languages is exceedingly rare cross-linguistically; the only attested instances are Bai and the Chinese language, both of which are Sino-Tibetan (p. 26).14 Bai is a minority language spoken in the province of Yunnan in China. It is reasonable to assume that the “REL N” order in Bai arose out of intensive language contact with Modern Chinese. If the Chinese and Bai cases are not considered, therefore, we have identified the only exception-free universals among all forty-four of the Greenbergian universals: (60) If VO, then N REL. Due to a series of internal changes over history, Chinese gradually deviated from the most typical features of VO languages. This fact calls into question the assumption that language contact is the only possible factor causing deviation from the so-called universals (Lehman 1973, 1978). In the above analysis, we also briefly discussed the comparative structure in Old Chinese, which was also consistent with other VO languages, as predicted by Greenberg (1966a). In the literature, once again, the comparative structure of Modern Chinese creates the only counterexample for the related Greenbergian Universal 23. This deviation is in fact also due to the diachronic change that was responsible for the order change of noun and relative. Li and Thompson (1974) listed seven major typological features of Contemporary Chinese that are supposed to belong to OV languages. Thus far, the robust empirical evidence in the previous sections has demonstrated that Chinese historically exhibited the following three features which are typically consistent with VO languages: To further support my finding that the Chinese language was historically much more consistent with SVO languages, we will briefly discuss the differences between Old and Table 21.2 Word order changes in Old and Modern Chinese

14

Old Chinese

Modern Chinese

(a) Relative clauses follow the head noun (b) PPs follow the verb (c) Certain adverbials follow the verb

Relative clauses precede the head noun PPs precede the verb All adverbials precede the verb

Mallinson and Blake (1981: 285), in their 150-language sample, found only one other VO language with exclusively prenominal RCs, Palauan (Malayo-Polynesian, Austronesian). It is not clear why Dryer (2003) did not take this case into account. For VO languages that have both prenominal and postnominal RCs, see Mallinson and Blake (1981: 285), Comrie (1981: 141), and Keenan (1985b: 144), among others.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Word Order and Relative Clauses

536

Modern Chinese with respect to the remaining four features of Li and Thompson (1981: 18) that are not listed above, as follows. Feature (d) listed in (8): “Postpositions exist.” The so-called postpositions refer to shàng in zài zhuōzi shàng “on the table above.” In our view, this shàng is not truly an adposition because it can always be exchangeable with the full compound noun shàngmiàn “upper surface.” Additionally, unlike other prepositions that can be used independently, the so-called postpositions are always paired with another preposition. Even a word such as shàng is viewed as a type of adposition, which did not exist until Modern Chinese. Feature (e) listed in (8): “OV sentences occur.” The so-called OV sentences in Contemporary Chinese refer mainly to the two following patterns: the bǎ construction and the construction with a patient noun between the subject and the verb. However, these syntactic structures did exist in Old Chinese: the former emerged around AD 700, and the latter occurred around AD 1300. The earliest examples of the bǎ construction did not appear before the eighth century AD (for details, see Wang 1989: 266‒271), and one of the earliest examples is illustrated in (61). The construction with an unmarked patient noun between the subject and the verb came into being as late as the thirteenth century AD, as illustrated in (62): (61)




莫把杭州刺史欺。 (白居易 戲醉客) Mò bǎ Hángzhōu cìshǐ qī. Do-not BA Hangzhou governor cheat “Don’t cheat the governor of Hangzhou!” (Bai Ju Yi, Xi Zui Ke, AD 850)

(62)

我昨日冷酒多喝了。 (老乞大) Wǒ zuórì lěng-jiǔ duō hē le. I yesterday cold-wine much drink PERF “I drank too much cold wine yesterday.” (Lao Qi Da, AD 1300)

Feature (f) listed in (8): “Aspect markers follow the verb.” In Contemporary Chinese, there are three aspect suffixes – -le (perfect), -zhe (progressive), and -guo (experiential) – which follow the verb to indicate the state of progression of the action of the matrix verb (cf. Chapter 13). However, all three of these aspect markers were actually introduced into the language after the tenth century AD (Wang 1989: 102). In Old Chinese, the aspect markers behaved like auxiliary verbs that preceded the verb, as illustrated in (63); the word yǒu “have” is a perfect marker, and in (64) the word cháng, which grammaticalized from the verb “taste,” is an experiential marker:

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Typological Change in the Relative Clause Structure

537




(63)

女子有行, 遠父母兄弟。 (詩經 蝃蝀) Nǚzǐ yǒu xíng, yuǎn fùmǔ xiōngdì. girl have marry far parents brothers “The girl has got married and lives far away from her parents and brothers.” (Shi Jing, Di Dong, 1000–600 BC)

(64)

俎豆之事, 則嘗聞之矣。 (論語 衛靈公) Zǔdòu zhī shì, zé cháng wén zhī yǐ. ceremony-sacrifices ASSO matter but ever learn it before “But I learned the ceremony to offer sacrifices before.” (Lun Yu, Wei Ling Gong, 500 BC)




Feature (e) listed in (8): “Genitive nouns precede the head.” Throughout history, genitive nouns always preceded the head: the genitive marker in Old Chinese was zhī, and in Modern Chinese it is de, as illustrated in (65) and (66), respectively.




(65)

夫子之文章可得而聞之。 (論語 公冶長) Fūzǐ zhī wénzhāng kě dé ér wén zhī. master POSS article can obtain and hear it “What the Master teaches can be understood just by listening.” (Lun Yu, Gong Ye Chang, 500 BC)

(66)

我的車修好了。 (現代漢語) Wǒ de chē xiū-hǎo le. I POSS car fix-good PERF “My car has been fixed.” (Contemporary Chinese)

Thus we can say that the genitive structure is the only exception, diachronically and synchronically, if this feature is truly correlated with VO order. In fact, however, the correlation between genitive phrases and VO languages is quite weak; for example, English as an SVO language is another exception, e.g. John’s car,15 in which the genitive noun precedes its head. Perhaps there is actually no correlation between genitive phrases and VO order. Based on the above analyses, it is safe to say that the Chinese language was historically much more consistent with VO languages than Contemporary Chinese is. Over the past approximately 3,000 years, Chinese has undergone a global 15

In English, animateness plays a role in determining the relative order of nouns and possessives: possessives follow the head in a construction with an inanimate possessor, as in the corner of the box, but precede it in a construction with an animate possessor, as in John’s books.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

538

Word Order and Relative Clauses

change from a “typologically consistent” language to a “typologically inconsistent” language in terms of the Greenbergian universals, and this change has many important theoretical implications (the next section is a more complete discussion of this issue).

21.5 Development toward Consistency In this section, I turn to exploring some important theoretical implications related to the empirical evidence above, highlighting three issues: (a) the interaction between the cross-category harmony principle and the heaviness serialization principle, (b) the problem with the Greenbergian parameters of the relative order between verb and object, and (c) the Greenbergian universals as best viewed as the default setting of languages.

21.5.1 Cross-category Harmony Principle The diachronic developments in the history of Chinese discussed above might reflect the cross-category harmony principle proposed by Hawkins (1983: 97), which states the following: “modifiers either all tend to precede their heads or all tend to follow, which has been related to processing, to the need to distinguish heads of phrases from non-head modifiers.” Although the changes in Chinese caused the language to systematically deviate from the features of VO languages, the effect left Chinese extremely harmonic in conformity with the order of modifier/head. Except for the sentential level VO remaining unchanged, all of the other modifiers changed from the position after the head to the position before the head, involving both nominal and verbal phrases. As pointed out previously, the modifiers included relative clauses, adjectives, “number-classifier” phrases, PPs, and adverbial phrases. The functions of the two morphemes zhī (introducing possessives, relative clauses, and adjectives in prenominal position) and zhě (introducing relative clauses and adjectives in postnominal position) were eventually taken over by the single morpheme de. Unlike zhī and zhě, the morpheme de could be recursive; it could be used more than once within a single phrase. However, respect for cross-category harmony was achieved at the expense of violating the heavy-constituent principle, by which longer, more cumbersome dependent elements are often postposed even if head-last order is expected (Rankin 2003). As a result, some awkward constructions in Modern Chinese are grammatical but are not favored typologically or from the perspective of language processing.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Development toward Consistency

539

(67)

湯姆的最貴的那輛車。 (現代漢語) Tāngmǔ de zuì guì de nà-liàng chē. Tom POSS most expensive REL that-CL car “The most expensive car of Tom’s.” (Contemporary Chinese)

(68)

在約翰昨天買的德國製造的車上。 (現代漢語) Zài Yuēhàn zuótiān mǎi de Déguó zhìzào de chē shàng. in John yesterday buy REL German make REL car above “In the car made in Germany that John bought yesterday.” (Contemporary Chinese)

In (67), the head noun chē “car” is modified by a genitive phrase, an adjectival phrase, and a numeral phrase, and the morpheme de is used twice. In the attributive position of (68), the object chē “car” of the preposition zài “in” is preceded by two relative clauses that are marked by the morpheme de, but this structure was not allowed in Old Chinese. Dryer (2005) stated that relative clauses tend not to occur between a noun and an adposition, so prepositional languages tend to have structures of the form “P + N + REL” rather than “P + REL + N.” Furthermore, Frazier (1979) argued that structures of the form “P + REL + N” present a problem for the parser since “P + REL + N” might be too long a unit, and “P + REL” does not form a semantic unit. Similarly, Hawkins (1983) offered a pragmatic explanation for the heaviness serialization principle that defines a preference for “lighter” modifiers to occur earlier in their respective phrases, while “heavier” modifiers tend to be placed later. Preposed long modifiers make it more difficult to determine the head; once the head has been identified, a subsequent long modifier (e.g. relative clause or adjectival phrase) is more easily processed. However, the diachronic evidence in Chinese shows that the cross-category harmony principle (reflected in Contemporary Chinese) is somehow in conflict with the heaviness serialization principle (obeyed by Old and Medieval Chinese), although they are supposed to aid people in unpacking the constituent structure.

21.5.2 Greenbergian Parameters There are always certain exceptions, to a greater or lesser degree, to all of the so-called language universals thus far stated (Vennemann 1974, 1976, Lehman 1973, Hawkins 1983: 59‒60, Dryer 1992). As we saw above, if Old Chinese is considered, it seems that the correlation “VO ⸧ N REL” is free of exceptions. However, a critical question remains unanswered: what is the intrinsic relationship between these two types of constituent order? From Old to Modern Chinese, the word order changes from harmony

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

540

Word Order and Relative Clauses

to disharmony with the Greenbergian universals are theoretically significant in at least the following two aspects. Language changes are admittedly accidental and unpredictable and are thus responsible for violations of the so-called synchronic universals (Dryer 1992). However, our research experiences show that languages develop in a highly regular and systematic fashion and serve as an ideal window for discovering and evaluating universal correlations. Dryer (1992) correctly criticized the proposals to explain Greenbergian universals for facing a major theoretical problem, which is that there is no external motivation for asserting that languages are “consistent,” or serialize “naturally,” or are “harmonic.” There is no pretheoretical hierarchy that these descriptions reflect, and no independent principle from which they derive their putative explanatory power. All we can know about such assertions is that they are largely consistent with the data on which they are based.Our analysis shows that universal correlations can be observed to operate only in the realm of diachrony. Under these circumstances, diachronic investigation is necessary in searching for an explanation for Greenbergian universals. Since Greenberg (1966a), the relative order of V and O has been treated as the most important parameter in seeking cross-linguistic correlations (e.g. Hawkins 1983, Dryer 1991, 2007). The relation between the verb and the object has been generalized as “head and adjunct” (Comrie 1981: 96), “head and dependent” (Bartsch and Vennemann 1972: 131‒9), “mother node and non-mother node” (Hawkins 1984), and “operand and operator” (Hawkins 1983: 136). Superficially, the verb–object order is crucial at some level, but no proposals can explain why (Dryer 1992). Thus Hawkins (1983: 70) decided to abandon verb–object orders as a prime typological indicator. Over the past half century or so, a number of cross-linguistically recurrent correlations have been found among the orderings of different syntactic constituents. Having examined a sample of 625 languages, Dryer (1992) found the following property clusters – languages tend to have either A or B as follows. To my knowledge, no one has clarified the relations among the various features in each cluster. Assuming that the order of verb and object is the parameter for determining Table 21.3 The recurrent correlations of VO and OV languages A

B

Verb + object Verb + manner adverb Noun + relative clause Adposition + noun phrase VP + PP

Object + verb Manner adverb + verb Relative clause + noun Noun phrase + adposition PP + VP

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Development toward Consistency

541

the order of other features, there are at least two possibilities for each cluster (actually many more); as for Cluster A, the possible relation might be as follows: (69)

(a) Verb/Object ⸧ Verb/Manner Adverb ⸧ Verb/PP ⸧ Noun/Relative Clause; (b) Verb/Object ⸧ Verb/Manner Adverb; Verb/Object ⸧ Verb/PP; Verb/Object ⸧ Noun/Relative Clause; Verb/Object ⸧ Verb/PP.

In (69a), there is a sequence: Verb/Object directly determines Verb/Manner Adverb and so forth; Noun/Relative Clause is determined directly by Verb/PP but indirectly by Verb/Object. In (69b), every feature is determined by Verb/Object, but they bear no relation to each other. However, the diachronic evidence in Chinese strongly suggests such a feature hierarchy: (70)

Old Chinese: Verb/PP ∩ Verb/Adverbs ⸧ Verb/Relative Clause,16 Modern Chinese: PP/Verb ∩ Adverb/Verb ⸧ Relative Clause/Verb.

Semantically, the relationship between verbs and adverbials is similar to that between nouns and relatives. Hawkins (1983: 132) observed that diachronic links are plausible only for word order pairs that are sufficiently closely related semantically. In addition to the semantic relationship, the same morphological morpheme de has been used to associate adverbial phrases with the VP and to connect the relative clause to the head noun. According to our observations, the relative order of the verb and object has no apparent influence on the order of the other features. At least in Chinese, VO order directly determines the order of adposition and noun because all prepositions in Chinese were historically grammaticalized from ordinary verbs; even in Modern Chinese, many prepositions still can function like verbs. Keenan (1985b) claimed that verb-final languages heavily favor preposed relative clauses. According to the grammaticalization theory of Hopper and Traugott (2003: 60), it is assumed that word order changes can have a profound effect on grammar because different phrasal orders are typically associated with VO and OV languages. VO languages tend to be prepositional; adjectives, relative clauses, and possessives follow the noun; the auxiliary precedes the main verb; and the question particle marking yes/no questions occurs in the initial position in the clause. VO languages tend to show the order in reverse. Hopper and Traugott admitted that there are no ideal VO and OV languages. Although OV instances arose in Medieval Chinese, Chinese did not undergo a word order change from VO to OV (Sun and Givón 1985). Even if we hypothesize that Chinese truly underwent such a change from VO to OV, it fails to explain why Modern Chinese is still prepositional and the auxiliary always precedes the main verb. 16

The symbol “∩” means “and” in mathematic logic.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

542

Word Order and Relative Clauses 21.5.3 Greenbergian Universals

In my view, the Greenbergian universals, or any other correlations in linguistic typology, can be best viewed as representing the default setting of any software (e.g. Microsoft Word), and every language can customize them in response to internal or external motivations. Once the settings have been changed, they cannot automatically return to the original state. In other words, this default setting represents the most natural state of a grammar. However, it is not the most optimal setting and does not serve as a synchronic principle to drive an “inconsistent” language to develop toward a “consistent” one because there is no such distinction between “consistent” and “inconsistent” languages. In any particular period, within a language, grammatical features are harmonic with each other in one way or another, forming an integrated system. If a language did not change over time, its grammar would likely be closest to the natural state predicted by the universals (if the universals are correctly generalized). In reality, however, every language must change over time; hence no language is truly consistent in terms of the Greenbergian universal correlations. In other words, these socalled universals represent the starting system of a language evolution rather than the destinations toward which a language strives to develop. Some languages reserve more features of the default system, but others might dramatically deviate from them because of the particular paths of development that they have undergone. As we have seen, Old Chinese was much more harmonic with SVO languages, but it underwent a series of dramatic changes, which caused Contemporary Chinese to manifest exceptions in many aspects. The grammar of Contemporary Chinese represents another type of harmonic system, but there is no sign in Contemporary Chinese that it is returning to the state of historical Chinese. This fact challenges the hypothesis of Vennemann (1976), which defined the principle of natural serialization, stating that time provides an impetus for languages to conform to a consistent ordering of operators with their operands, regardless of the constructions that appear; hence the Greenbergian universals are viewed as the end point of language development. Exactly the opposite, however, is found in the evolution of Chinese grammar. Compared with Contemporary Chinese, the historical Chinese language was much more consistent with typological universals. The fewer counterexamples that a universal correlation has, the more likely it is that it represents one of the default settings of a language, which defines a starting point from which a language develops. The Greenbergian universals are best viewed as the most natural states of a language, but any features can be altered due to unavoidable language development. All languages are always in the course of evolution, so no languages are completely consistent with any typological universal.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

Development toward Consistency

543

A new type of structural hierarchy has been identified in the history of the Chinese language. Although it has been assumed that VO languages strongly imply the N REL order, in fact the order of noun and relative is directly influenced by the relative constituent order of N and PP. Beyond their semantic similarity, nominal and verbal modifiers are marked by the same morphological morpheme de, providing direct morphological evidence for the present analysis. Although it has long been assumed that the relationship between the verb and the object is the same as that between the adjunct and the head, this assumption has never been seriously considered, and no scientific explorations have yet been conducted. The present analysis sheds new light on this issue. Specifically, my analysis invalidates the assumption that the two consistent word order types, adjunct–head (operator–operand, OV) and head–adjunct (operand– operator, VO), represent the most natural states for a language, and pressure to conform to these two ideal types is sufficient to initiate language change. On the contrary, in my view, the universal correlations represent only the natural states from which a language evolves, and since all languages are always in the course of change, no languages can entirely preserve the natural state. According to Comrie (1981: 92‒101), more than half of the world’s languages violate rules of consistency. More specifically, some languages appear to be quite stable in combinations that violate the canonical types, such as Persian, which is a head–adjunct language that nonetheless has stable final order and has been in this position for several centuries. As Comrie (1981: 86‒103) stated, typological theory does provide an explanation for why, given the appearance of a typological inconsistency, the language should strive to drift back into consistency, but it provides no explanation for why the inconsistency should have arisen in the first place – this fact is especially surprising given the alleged degree of pressure toward typological conformity. Thus “inconsistent” or “nonharmonic” languages would be an inevitable result of the inherent nature of change.

https://doi.org/10.1017/9781108921831.021 Published online by Cambridge University Press

22 Conclusions

22.1 The Momentum of Language Change Language development receives impetus from momentum, an inertial power gained from the pre-existing system. A language is an ever-changing body that is energized by the collective cognition of the language community. Every speaker is subject to the impetus gained by this ever-changing body. Factors related to language acquisition, pragmatic inference, and language contacts cannot cause any change unless they are in accordance with the momentous development that the language is undergoing. Once a language gains momentum, it will develop in that direction for many centuries, and no individual can stop or alter its course. This point of view is illustrated by the following two cases. First, in Late Old Chinese, momentum was gained that forced all non-resultative elements to be moved from the postverbal to preverbal position. This change lasted more than a millennium and involved many major syntactic constructions, such as the passive, comparative, locative, instrumental, and ditransitive constructions. During this process, a set of ordinary verbs that occurred in the first verb position of a serial verb construction were stimulated to develop into grammatical morphemes, taking over the functions of the former ones that had been in postverbal position. Consequently, the texture of the grammar was fundamentally altered. The grammaticalization processes of ordinary verbs must take place in specific contexts that are subject to the influence of semantic suitability, frequency, and pragmatic inference. As a whole, however, any concrete change must be in line with the general development of the grammatical system. Meanwhile, any deviation from the direction of development could not survive and was eventually abandoned. For example, in Late Medieval Chinese, the disposal construction had two major markers, jiāng and bǎ, but the former could occur only in postverbal position and the latter only in preverbal position. Clearly, the preposition phrases in the disposal construction were nonresultative; consequently, the jiāng construction was abandoned, and only the bǎ construction is still used today.

544

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

The Momentum of Language Change

545

Second, considering the way of introducing an agent, the passive structure has undergone three major changes: in Old Chinese, no passive morphemes could introduce an agent noun in preverbal position; in Medieval Chinese, it was optional for passive morphemes to introduce an agent noun in preverbal position; and in Modern Chinese, it became obligatory for passive morphemes to introduce an agent noun in preverbal position. This momentum moved steadily in a given direction. This course determined the fates, lifespans, functions, and lexical sources of passive morphemes. Historically, there were nearly twenty passive morphemes with different lexical sources and numerous passive structures with different functions. These phenomena appear extremely complex but became truly regular from the perspective of the developmental momentum. Our research indicates that none of these changes occurred in isolation or were simply caused by accidental factors. These facts call into question three widely known hypotheses about factors involved in language change; that is, (a) children’s language acquisition, (b) pragmatic inference, and (c) language contact, all of which assume that there is a static language system that may be changed by accidental factors. This is not the case. All individuals must be influenced by the momentous power of the ever-changing language from the moment they are born. Therefore the big picture is that the evolution of the grammatical system is regular, is systematic, and has a direction, and each characteristic is illustrated as follows. By “regular,” we mean that grammatical changes are often free of exceptions and must proceed in line with certain rules. When the information-structuring principle of the predicate was established, all related constructions had to change, which involved the locative, passive, instrumental, comparative, and ditransitive constructions. In this case, for example, the comparative construction could not remain unchanged. The term “systematic” here has two senses. First, the overall properties of the grammatical system at the time play a crucial role in motivating and controlling an individual change, which involves the recruitment of lexical sources, the life span and function of a grammatical morpheme, the concrete structure of a grammatical device, the direction and fate of a grammatical form, and so forth. Second, only pragmatically induced changes that are consistent with the trend of the grammatical development can survive and develop into a new grammatical category. Note that the meaning of directionality here is different to grammaticalization theory, which defines a cline along which a lexical item develops toward a grammatical marker. The expression “has a direction” here refers to the fact that the grammatical system develops in a direction for hundreds of years and even millennia, and this development is not interrupted by so-called children’s language acquisition or different generations’ parameter settings. Considering the presence of an agent noun, for example, the passive markers in preverbal position have undergone three stages: they could not introduce the

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

546

Conclusions

agent in Old Chinese, they could optionally introduce the agent in Medieval Chinese, and they must introduce the agent in Modern Chinese. There is no fluctuation with respect to this feature of the passive structure.

22.2 Changes in Cluster As we analyzed in this book, there have been hundreds of changes in Chinese grammar in history, and none of them happened in isolation. We use the term “change in cluster” to describe the characteristics of the evolution of the grammatical system. This term can be illustrated by a Chinese proverb, “Pull one hair and you move the whole body,” which means that a slight move in one part may affect the whole situation and that a minor step may have major consequences. This reveals the most essential property of the diachronic system: all parts form an integral whole, similar to the human body or an ecological circumstance, where, if any individual part changes, it will affect the most closely related parts and even the whole system. Thus far, no theoretical frameworks in historical linguistics can predict and explain the model called “language change in cluster,” a phenomenon that has happened repeatedly in the history of Chinese grammar. There are several hypotheses available in the literature and their predictions of the fashions of grammar changes are briefly stated here. According to the principle-parameter hypothesis, the set of principles remains stable and the parameters may change due to the settings of different generalizations (Lightfoot 2013). This hypothesis predicts that the development of a grammatical system proceeds abruptly, and a parameter change can be applicable only to similar structures, such as those consisting of the modifier and the head. However, a cluster of changes usually includes various types of construction or grammatical morphemes that bear the least structural assemblies, such as the emergence of the copula shì and the typological change in the syntax of wh- words. More importantly, there is no evidence that any change is abrupt in pace over changes in generations. According to the usage-based model, grammatical innovations are most likely to happen to high-frequency words, and a grammaticalization process is accompanied by an increase in token and type frequency (Bybee 2013). One key issue that must be resolved here is that, in many cases, the innovative form violates certain existing grammatical rules; for example, the emergence of the resultative construction broke the rule governing the co-ordinate verb construction. How could this new form come into existence, mature, and finally replace the former grammatical rule? No change is a matter only of its own usage, such as repetition and entrenchment of cognition, but must be motivated and guided by the overall properties of the grammatical system. Additionally, this model can only predict that any change in grammar occurs in isolation

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Changes in Cluster

547

and cannot explain why change typically occurs in a cluster of constructions rather than individually. In grammaticalization theory, the major factor in initiating a change is pragmatic inference (Hopper and Traugott 2003: 71‒98). Theoretically speaking, every language user may have their own pragmatic inferences; thus the question is, what factor determines which pragmatic inference will be recognized by the whole language community and finally become a new grammatical device? More crucially, this hypothesis can only predict that any new grammatical form will happen individually and that each change has its own pragmatic inference. According to our research, only pragmatic inferences that are favored by the overall properties of the grammar in the particular period can cause a change in a grammatical device in a morphosyntactically specific context. It is true that any grammatical form must develop out of a proper context, unlike lexemes, which can be coined independently of any linguistic construction. The necessary condition guarantees that the new form is to a great extent structurally consistent with the existing grammatical system. The history of Chinese grammar encompasses several major clusters of changes, from which we can see that the developments are highly regular. These clusters are summarized as follows. Cluster 1: the emergence of the copula shì. The copula shì grammaticalized from its original demonstrative usage, in which it was often used as an anaphor between two nominal phrases. This development was encouraged by the SVO word order at the expense of the old copular construction that was marked by the sentence-final particle yě; in turn, the new copular construction marked by shì further strengthened the feature of word order. Later, the copula shì developed into a focus marker as an analytic device to replace the old method of highlighting by shifting the constituent order. Consequently, the old focus constructions disappeared; specifically, movement of three types of element ceased to be used, including wh- words, pronouns in negative construction, and the displacement of nouns. Furthermore, wh- words exhibited their inherent focus features by fusing with the focus morpheme shì, which resulted in the replacement of all interrogative pronouns. Cluster 2: the emergence of the resultative construction. The emergence of the resultative construction may have been the most important event in the whole history of Chinese grammar, as it fundamentally changed the texture of the grammar and distinguished the system of Contemporary Chinese from that of Classical Chinese. This change was enabled by two high-level conditions, one intermediate-level condition,

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

548

Conclusions and three low-level conditions. At the high level, the disappearance of the conjunction ér created the serial verb construction in which two verbs occurred adjacently, and the disyllabification tendency fused two monosyllabic units that co-occurred frequently. At the intermediate level, the co-ordinate verb construction required that only two or more transitive verbs could precede an object and no other element could appear between the transitive verb and the object. This condition channeled the process from which the resultative construction developed. The establishment of the resultative construction was responsible for various changes, mainly including the disappearance of the co-ordinate verb construction; the emergence of the disposal, verb-copying, and SOV constructions; the emergence of the aspect system; the formation of verb reduplication; the boundedness of the predicate structure; and the formation of the negative system in Contemporary Chinese. Furthermore, due to its high frequency and wide scope of usage, the resultative construction assimilated other constructions, such as the ditransitive and some verb-plus-object constructions. Cluster 3: the establishment of the principle governing the information structure of the predicate. This principle was established by analogy with the resultative construction, which required that only preposition phrases that expressed some kind of resultative could stay in postverbal position and all others that were non-resultative could appear only in preverbal position. Under the operation of this principle, many types of sentential structure underwent fundamental changes in constituent order, including locatives, comparatives, instrumentals, and ditransitives. During this change, a set of verbs in the first verb position of a serial verb construction became grammatical morphemes to take over the functions of the older forms. Under the operation of this series of changes, the relative clauses that formerly followed the head noun were restricted to prenominal position. Consequently, the constituent order within a nominal phrase developed from two types in Old Chinese, either head–modifier or modifier–head, into one “modifier– head” order. Cluster 4: the establishment of the classifier system. The establishment of the classifier system was the outcome of the combined powers of the disyllabification tendency and the information-structuring principle of the predicate. The emergence of classifiers added a new word class to the grammar and profoundly affected the structure of nominal phrases: first, the numeral modifier had to be associated with its head by

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Patterns of Grammatical Change

549

a classifier; second, by analogy, the demonstrative dǐ was triggered to develop into an associative morpheme to mark relative clauses, genitive phrases, and the adjectival phrases. The emergence of the classifier system made the old demonstratives unsuitable for the new nominal structures; at the same time, the most general classifiers zhè and gè developed into proximal demonstratives in the northern and southern dialects respectively. Additionally, the classifier system motivated the grammaticalization process of the plural -men from a classifier and the diminutive marker -er from the concept “child.”

22.3 Patterns of Grammatical Change The patterns of grammatical change are far richer than those identified in the literature and expected by any framework. According to our research, these patterns include the following: Type 1: the disappearance of grammatical forms. The history of Chinese grammar consists of two essentially related sides: the appearance and disappearance of grammatical forms. Grammatical change is similar to the metabolism of any living creature, with new forms being constantly added while old forms are abandoned. There is no language that just increases the number of its grammatical forms without removing old ones. To reveal the regularity of grammatical development, studying the disappearing phenomena is equally important because it may create a condition for the emergence of a new form (e.g. the disappearance of the conjunction ér made it possible for the verb and the resultative to become fused) and may control the course of grammaticalization (e.g. the development of the resultative construction was heavily influenced by the co-ordinate verb construction, and the disappearance of the latter was caused by the former). Type 2: replacement of the old by new forms. Some historical grammatical forms simply disappeared without any replacement, such as the verb conjunction ér, which meant that the grammatical category no longer existed in the language. However, it is common for the grammatical category to remain the same while the form changes, which may involve replacing markers and altering structures. For example, the comparative construction underwent the replacement of the marker from yú to bǐ and the repositioning of the standard phrase from following the predicate to preceding the predicate.

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

550

Conclusions Type 3: innovation in grammatical category and form. Historically, there were many innovations that did not exist before, such as the disposal construction, the verb-copying construction, the classifier system, the plural morpheme -men, and the diminutive particle. Type 4: grammaticalization from lexical sources. A majority of grammatical markers developed out of lexical items. For example, the perfective aspect -le came from the verb “finish,” the plural marker men from the classifier for enumerating families, and the diminutive -er from the concept “child.” Type 5: phonological derivation. Not all grammatical markers must originate from a lexical source; phonological derivation is another way for them to develop. This has often happened in two complementary categories within a functional domain. The exemplar case is the proximal and distal demonstratives, where the proximal part developed out of a general classifier and the distal part was phonologically derived from the proximal. Type 6: change in function. The same marker may acquire a new function over time. For example, the passive bèi could not introduce an agent for the first 600 years or so when it grammaticalized from the verb of suffering, but it later acquired that function. Type 7: obligatory constituents. For the passive structure, for instance, the presence of the agent phrase was originally optional, but it became obligatory in Contemporary Chinese by analogy with the disposal construction. From the very beginning, the presence of the patient noun in the disposal construction was obligatory to make the construction well formed. Type 8: change in syntactic category. In Old Chinese, the category of aspect was expressed by a set of auxiliary verbs, but since the beginning of Modern Chinese it has been expressed by a set of verb suffixes. Type 9: change in the function of syntactic position. This change does not involve any overt formal feature. For example, in Medieval Chinese, a rule took place: a bare noun in the subject position was automatically assigned the feature definite, whereas a bare noun in the object position was assigned the feature indefinite. In turn, this rule had a profound effect on the emergence of the disposal construction and the extension of wh- words to universal and indefinite reference. Type 10: the borrowing of a grammatical category. This change typically happened in some isolated dialects and is similar to loanwords. The purely innovated grammatical categories tended to be borrowed by

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Drivers of Language Change

551

isolated dialects. In the vast area of the northern dialect, the diminutive marker -er grammaticalized from the word ér “child” in Medieval Chinese. For geographic and social reasons, it could not spread to many isolated dialects. However, these dialects simply borrowed this new grammatical category by assigning it a phonological form, such as a high-pitched tone, a reduplicated form, or a nasal consonant. The above list represents the major types of grammatical change but is not exhaustive.

22.4 Drivers of Language Change From one perspective and to a certain extent, language is like any biological phenomenon, with any current state evolving from its former state. However, every human being undergoes the stages from infant to toddler, adolescent, and adult. Even when a person becomes an adult, the infant from whom the adult grew is still alive, but with a different shape. Similarly, Old Chinese, which existed more than three millennia ago, is still alive today and has evolved into Contemporary Chinese. Only a language that has ceased to be used can be regarded as dead; for example, Sumerian, the oldest-known written language, which dates back to at least 3500 BC but only lasted until 2000 BC, is a dead language. This metaphor is somehow misleading because it implies that a language in the early stages might be simplistic and destined to die. However, the grammatical system of a language at an earlier stage is by no means simpler than that at a later stage. A language will exist as long as it is used as a spoken language. The life of a language consists of two parts: the cognition of the language community and the symbolic system for communication. The cognition of the language community determines the openness of human language, and the symbolic system renders the closeness of human language. Once again, this is something like the human body. To maintain life, humans must interact with the outside world, such as absorbing food and inhaling oxygen; meanwhile, the organs in their bodies need to work co-operatively, such as blood circulating and the heart beating. Therefore the motivations and mechanisms of the evolution of the grammar fall into two types, both outside and inside the grammatical system. Theoretically, any phenomena that can be the object of human cognition, which might be social, natural, and even imaginary, may affect the language system. The history of Chinese shows that there has been a massive inertial force of cognition within the language community that has overridden any influence from individual pragmatic inference or language acquisition across generations. Meanwhile, many changes are triggered simply by certain internal factors. Although the reduplicated

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Conclusions

552

forms of adjectives and nouns were already widely used in Old Chinese, verb reduplication did not come into existence until the fourteenth century AD, when the establishment of the resultative construction and the emergence of the aspect markings co-operatively created a new syntactic slot between the verb and the object. Due to outside cognitive activities and inside systematic factors, any language is always in the course of evolution and many developments happen necessarily rather than possibly. Changes in grammar, far from being random, independent, or unpredictable, are rather regular, systematic, and predictable.

22.5 Construction Schema Historically, the structures of passives, locatives, comparatives, instrumentals, and ditransitives have changed dramatically, but these changes happened in a consistent way, from which we can see that the construction schema was at work in the evolution of grammar. In Section 7.9, we discussed the formation of the principle of action–resultative ordering, which required that all non-resultative preposition phrases and adverbials disappear from postverbal position. These events resulted in a series of changes in sentence structures involving all the constructions mentioned above. As we saw previously, various passive morphemes and structures have existed throughout history. From a different perspective, this chapter aims to identify a new type of grammatical form, called the “construction schema,” that was at work throughout history, an essential concept for capturing the regularity of grammar development. In all major theoretical frameworks, functional constructions, such as passives, comparisons, and ditransitives, are regarded as the largest linguistic unit or maximal schema (e.g. Chomsky 1957, Langacker 1987, Goldberg 1995, Croft 2001). In empirical studies of language, the maximal forms in linguistic typology are word orders, such as SVO and SOV (e.g. Greenberg 1966a, Comrie 1981, Croft 1996). For instance, Chomsky (1957), Goldberg (2006), and Croft (2001) specifically used the passive as their largest construction (“mostly schematic,” in Croft’s words). They portrayed the passive construction in English: (1)

(a) NP2 – AUX + be + en – V – by + NP1 (Chomsky 1957: 43), (b) Subj AUX VPpp (PPby) (Goldberg 2006: 5), (c) SBJ be-TNS VERB-en by OBL (Croft 2001: 5).

In the evolution of the Chinese language, however, we believe that the passive is not the most schematic construction and that a type of linguistic form, termed the “construction schema,” exists between these functional constructions and basic word orders.

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Construction Schema

553

It is abstracted from a concrete construction, such as the passive, locative, and comparative, on the basis of their shared function and form. The evidence from the history of Chinese clearly shows that a construction schema has been at work in determining the diachronic changes in Chinese passives. Within a given period, the construction schema determines what must happen, what cannot happen, and in which direction the language might develop. A construction schema is the outcome of a general form of human cognition – schematization (Hintzman 1986, Rimé et al. 1990). This type of cognition is at work at all levels of linguistic units, including syllables, words, phrases, and sentences. This concept has been applied to language acquisition (e.g. Bybee 1985, 1995, Tomasello 2003, Goldberg 2006: 220). According to Langacker (2013: 167‒173) and in cognitive linguistics, grammar is regarded as a schema with different levels of internal complexity, in contrast to generative linguistics and OT theory, which concern rules or constraints. Recently, an increasing number of researchers have applied this concept to historical investigations, claiming that grammaticalization, especially constructionalization, is related to schematization (Noël 2007, Traugott 2007, Trousdale 2008, Traugott and Trousdale 2013: 13‒16). In the literature, the most abstract schema is not beyond the functional construction; for instance, in Traugott and Trousdale (2013), the “macro construction” refers to the “give-gerund” (e.g. Tom gave John a push). In addition, all these studies suggest that schematization works in a “bottom-up” fashion, namely how a construction is created via high frequency or entrenchment. However, in this chapter, we discuss how schematization works in an “top-down” fashion – guiding the direction of grammar evolution. Kiparsky (2013) provided an extensive investigation of passives in numerous languages, arguing for a “null theory” in line with OT-based lexical decomposition grammar (Stiebels 2002, Wunderlich 1997, 2006), a base-generated syntax that eliminates NP movement. His central viewpoint was that there is no specific construction exclusively for the passive and that the syntax of the passive is parallel to other syntactic structures of the same abstract schema. Within Kiparsky’s framework, the passive clauses in a language have no passive-specific syntactic properties, and the syntax is predictable from the active sentences of the language and from the argument structure of passive predicates. Therefore he concluded that “a learner who knows the grammar of active sentences of a language can predict the syntax of other diatheses” (Kiparsky 2013: 7). In what follows, we demonstrate that the passive and many other active clauses in Chinese actually belong to the same construction schema and that they underwent parallel developments in history. Construction schemas are universal, reflecting the nature of human language. For instance, the most general pattern of English may be abstracted as follows:

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Conclusions

554 (2)

The English passive: be + V-ed (past participle) + by NP.

In the generativist tradition, all components of the passive structure are meaningless, generated from so-called deep structures (typically an active form) through a set of formal rules (e.g. Chomsky 1957: 61‒84). However, this English passive pattern is by no means universal; for example, the passive in Japanese does not have an auxiliary (copula) and uses the preposition ni to introduce the agent in preverbal position. These differences are determined by the overall properties of individual grammars. Langacker (2013: 121‒126) argued that no parts of the English passive are exclusively designed for the passive only and that all of them are meaningful. In his theory, only the copula and the past participle are the central parts; the by-phrase is marginal, closely related to its locative usage. Thus the concrete construction is licensed by a more abstract schema: (3)

AUX + V-ed + PRE + NP.

“Vi” could be either a present participle or a past participle. That is, the following sentences share the same construction schema as passives: (4)

(a) (b) (c) (d)

The window was broken by Tom. Bill was sleeping by the table. They had finished their homework in school. John is reading newspapers in the dining room.

Within a synchronic system, it may sometimes be difficult to determine which constructions are grouped to form a schema. As indicated in the next section, however, it is easy to identify how many concrete constructions a construction schema includes in the evolution of a grammatical system. Langacker (2013: 167) first proposed the term “constructional schema.” In his model, linguistic units, from simple words to complex constructions, are by nature symbolic assemblies that can be either specific or schematic. In our definition, in contrast, construction schemas refer only to those schematized from clauses or sentences. In the present analysis, a construction schema is understood as an abstract template that treats passive clauses and many other active clauses as members – a linguistic level that is between a functional sentential structure (e.g. passive, ditransitive) and so-called basic word orders (e.g. SVO, SOV). (a) Word order refers to the linear arrangement of the three basic constituents – subject, verb, and object. Within a given language, the basic word order is stable, and there is usually some variation due to pragmatic factors. Word

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Construction Schema

(b)

(c)

555

order is a matter of convention and hence does not express any semantic or pragmatic values. The construction schema refers to the linear orderings involving all elements composing a sentence, such as the three basic constituents, adjuncts (e.g. preposition phrases), morphological markers, and word classes. It is schematized from functional constructions, such as passives, ditransitives, and resultatives, as a clause-level form with an identifiable meaning and form. Functional construction refers to types of clause structure with different semantic or pragmatic values, such as passives, ditransitives, resultatives, topicalization, and questions. In this chapter, the term “construction” refers to clause structures, excluding words, morphological markers, and phrasal structures (for the definition of construction, see Goldberg 1995: 1‒5 and Croft 2001: 14‒28).

The identification of this linguistic form, namely the construction schema, will greatly benefit researchers in historical linguistics, as shown in the following analysis. Grammatical schemas may exist at phrasal levels or syntactic levels. For example, due to the emergence of the classifier system and the associated particle de, the schema of the nominal phrase changed from “modifier + head” to “modifier + grammatical marker + head.” The establishment of the resultative construction created a syntactic schema “V + X + O,” where the X element is intransitive and refers to the state of the progression of the verb. It replaced the former schema “V1 + V2 + O,” where the two verbs had to be transitive and each governed the object individually. The informationstructuring principle of the predicate caused a series of structural changes whose combined outcome was the formation of two subtypes of syntactic schema: (a) “Subj VO PPresultative” and (b) “Subj PPnon-resultative VO.” Having identified the existence of the construction schema, we can see that the evolution of morphosyntactic history becomes highly regular, and the motivations for changes in the passive can be easily explained. This construction schema involves at least five major functional constructions: passive, comparative, instrumental, locative, and disposal. They were grouped together by the new information-organizing principle: all non-resultative preposition phrases had to occur in preverbal position, which caused the following three changes. First, all prepositions that introduced non-resultatives, mainly including yǐ (instrumental) and yú (passive, locative, comparative), disappeared from preverbal position. Second, many verbs grammaticalized in the first verbal position of a verbal serialization, expressing the functions of the passive. Third, the preposition phrases that introduced an agent noun were restricted to preverbal position. As a result, the passive morpheme jiàn was abandoned because it never developed the

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

556

Conclusions

function of indicating an agent, bèi evolved a new function of introducing an agent, and the wèi . . . suǒ pattern was strengthened because it was able to introduce an agent in preverbal position even in the Old Chinese period. Moreover, in the new construction schema, preposition phrases in preverbal position had to be followed by a nominal phrase to make the structures well formed. In particular, the disposal construction emerged around the eighth century AD, and the patient noun needed to occur to have a legitimate construction. They were functionally complementary, and the newly formed passive patterns required an agent noun to be grammatical. This requirement had the following consequences. First, the passive bèi quickly declined in use because it always optionally introduced an agent noun, although this passive pattern was used for more than 2,300 years. Second, the chī passive existed only for a very short period and disappeared entirely because it was never fully developed and was subject to elimination under the strong trend. Third, the newly grammaticalized passive morphemes, namely, jiào, ràng, and gěi, that were introduced into the language in the last 200 or 300 years shared the same syntax: the presence of an agent noun is obligatory in order to make the structure well formed. This usage is typologically peculiar. Additionally, the lexical sources for grammaticalization to become passive morphemes were extremely diverse in Modern Chinese, and there were at least sixtynine different ones among the 930 dialects, including “allow,” “call,” “give,” “wait,” “ask for,” and “want.” All these verbs, which are typically followed by objects, frequently occur in the first verbal position of a serial verb construction structure, the context responsible for the change. Thus their grammaticalization paths appear very natural if the internal construction network is taken into consideration. Finally, the length of the life spans of grammatical morphemes and the selection of lexical candidates for certain grammatical domains are determined largely by the construction schema in a particular period. Throughout history, the life spans of different passive morphemes varied greatly; for instance, bèi has been used for more than 2,000 years, but chī existed for only roughly 600 years. In fact, their fates were determined by the construction schema as well as by their abilities to adapt to the overall grammatical property. The selection of lexical items in grammaticalization was also by no means accidental and was often determined by the overall grammatical property. The so-called frequency and pragmatic inferences in triggering grammaticalization, however, are entirely independent of the grammatical system. From the perspective of the construction schema, it is safe to say that no changes are random or accidental. Beyond diachronic investigation, we can also see how the grammar of a language operates consistently. Additionally, the traditional distinction

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

Construction Schema

557

between passive and active disappears; thus all theories based on this distinction need to be reconsidered. The identification of a construction schema also has profound implications for theoretical linguistics. Since the inception of generative linguistics, it has been assumed that every passive sentence is transformed from its corresponding active sentence. Chomsky (1957: 43) explained it as follows: (5)

If S1 is a grammatical sentence of the form NP1 – AUX – V – NP2 then the corresponding string takes the form NP2 – AUX + be + en – V – by + NP1 is also a grammatical sentence.

This hypothesis faces many empirical problems; for instance, many active sentences do not have corresponding passive forms, and the opposite is also true. This theory is powerless to explain why the passive form in a particular language has undergone fundamental changes. Our analysis has proven that the so-called passive and active forms actually belong to the same construction schema, which makes the grammar considerably simpler. We believe that we will gain a deeper understanding of grammatical evolution when our research agenda includes construction schemas from crosslinguistic perspectives.

https://doi.org/10.1017/9781108921831.022 Published online by Cambridge University Press

REFERENCES

Aboh, Enoch. 2006. If we see focus, you go left and I go right! Paper presented at the International Conference on Bantu Grammar. SOAS, London. Adams, Marilyn Jager and Allan Collins. 1979. A schema-theoretic view of reading. In Roy O. Freedle (ed.), New directions in discourse processing, 1–22. Norwood: Ablex. Aikhenvald, Alexandra Y. 2000. Classifiers: a typology of noun categorization devices. New York: Oxford University Press. Aissen, Judith L. 1992. Topic and focus in Mayan. Language 68: 43–80. Albridge, Edith. 2010. Clause-internal wh- movement in Old Chinese. Journal of East Asian Linguistics 19: 1–36. Allan, Keith. 1977. Classifiers. Language 53: 285–311. Aoun, Joseph and Y.-H. Audrey Li. 1993. Wh- elements in-situ: syntax or LF? Linguistic Inquiry 24: 199–238. Aoun, Joseph, Norbert Hornstein, and Dominique Sportoche. 1981. On some aspects of wide scope of quantification. Journal of Linguistics Research 13: 67–95. Ariel, Mira. 1990. Accessing noun-phrase antecedents. London: Routledge. Aristar, Anthony Rodrigues. 1991. On diachronic sources and synchronic patterns: an investigation into the origin of linguistic universals. Language 67: 1–33. Baker, Carl L. 1970. Notes on the description of English questions: the role of an abstract Q morpheme. Foundations of Language, 6: 197–219. Baker, Mark. 1988. Incorporation: a theory of grammatical function changing. Chicago: The University of Chicago Press. Barber, Charles. 1995. The English language: a historical introduction. Cambridge: Cambridge University Press. Bartsch, Renate and Vennemann Theo. 1972. Semantic structures: a study in the relation between semantics and syntax. Frankfurt am Main: Athenaum. Barz, Richard and Anthony Diller. 1985. Classifiers and standardisation: some South and SouthEast Asian comparisons. In David Bradley (ed.), Language policy, language planning and sociolinguistics in South-East Asia, 155–184. Canberra: Pacific Linguistics. Behaghel, Otto. 1909. Beziehungen zwischen Umfang und Reihenfolge von Satzgliedern. Indogermanische Forschungen 25: 110–142. Beida. 1995. Hanyu fangyan cihui (The dialect lexicon of Chinese). Beijing: Yuwen Press. Bisang, Walter. 1999. Classifiers in East and Southeast Asian languages: counting and beyond. In Jadranka Gvozdanovic (ed.), Numeral types and changes worldwide, 113–185. Berlin: Mouton de Gruyter. Bouton, Lawrence F. 1973. Some reasons for doubting the existence of a passive transformation. In Braj B. Kachru, Robert B. Lees, Yakov Malkiel, Angelina Pierangeli, and Sol Saporta (eds.),

558

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

559

Issues in linguistics: papers in honor of Henry and Rene Kahanée, 70–84. Champaign: University of Illinois Press. Bresnan, Joan. 2001. Lexical-functional syntax. Oxford: Blackwell. Brinton, Laurel J. and Elizabeth C. Traugott. 2005. Lexicalization and language change. Cambridge: Cambridge University Press. Brown, Lucien and Jaehoon Yeon. 2015. Handbook of Korean linguistics. Chichester: WileyBlackwell. Brown, Penelope and Stephen C. Levinson. 1978. Politeness: some universals in language usage. Cambridge: Cambridge University Press. Bruening, Benjamin. 2007. Wh- in-situ does not correlate with wh- indefinites or question particles. Linguistic Inquiry 38: 139–166. Bussman, Hadumod. 1996. Routledge dictionary of language and linguistics. London: Routledge. Butt, Miriam and Tracy Holloway King (eds.). 1996. LFG-Workshop proceedings of the first LFG conference. Rank Xerox Research Centre, Grenoble, August. Bühler, Karl. 1934. Sprachtheorie: die Darstellungsfunktion der sprache. Jena: Fischer. Bybee, Joan. 1985. Morphology: a study of the relation between meaning and form. Philadelphia: Benjamins. 1995. The semantic development of past tense modals in English. In Joan Bybee and Suzanne Fleischman (eds.), Modality in grammar and discourse, 503‒507. Amsterdam: John Benjamins. 2003. Mechanisms of change in grammaticization: the role of frequency. In Brian D. Joseph and Richard Janda (eds.), Handbook of historical linguistics, 602‒623. Oxford: Blackwell Publishing Ltd. 2006. Frequency of use and the organization of language. Oxford: Oxford University Press. 2007. Diachronic linguistics. In Dirl Geeraerts and Hubert Cuyckens (eds.), Handbook of cognitive linguistics, 945–987. New York and Oxford: Oxford University Press. 2013. Usage-based theory and exemplar representations of construction. In Thomas Hoffmann and Grameme Trousdale (eds.), The Oxford handbook of construction grammar, 49‒69. Oxford: Oxford University Press. Bybee, Joan and Clay Beckner. 2015. Usage-based theory. In Bernd Heine and Heiko Narrog (eds.), The Oxford handbook of linguistic analysis, 827‒856. Oxford: Oxford University Press. Bybee, Joan, Revere Perkins and William Pagliuca. 1994. The evolution of grammar. Chicago: The University of Chicago Press. Cao, Guangshun. 1986. Zutangji zhong di que le zhe (On the particles di, que, le and zhe in Zutangji). Zhongguo Yuwen 3: 192–202. 1995. Jindai Hanyu zhuci (Particles in Modern Chinese). Beijing: The Yuwen Press. Cao, Guangshun and Yu Xiaorong. 2000. Zhonggu yijing zhong de chuzhishi (The disposal construction in the translated sutras in Medieval Chinese). Zhongguo Yuwen 6: 555–563. Cao, Zhiyun (ed.). 2008. Hanyu fangyan dituji (yufa juan) (The maps of the Chinese dialects (grammar volume)). Beijing: Shangwu yinshuguan (The Commercial Press). Cardona, George. 1965. Gujarati reference grammar. Philadelphia: University of Philadelphia Press. Chafe, Wallace L. 1976. Givenness, constrastiveness, definiteness, subject, topics, and point of view. In Charles N. Li (ed.), Subject and topic, 25–55. New York: Academic Press. Chao, Yuen-Ren. 1968. A grammar of spoken Chinese. Berkeley and Los Angeles: University of California Press. 1979. Hanyu kouyu yufa (A grammar of spoken Chinese). Translated by Lü Shuxiang. Beijing: Shangwu Yinshuguan (The Commercial Press).

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

560

References

Chen, Mengjia. 1956. Yinxu buci zongshu (A survey of Yin oracle bone inscriptions). Beijing: The Science Press. Chen, Zhongmin. 2002. Lun Guangzhouhua xiaocheng biandiao de laiyuan (The origin of the tone sandhi of the diminutive in Cantonese). In Pan Wuyan (ed.), Dongfang yuyanxue (Eastern linguistics), 167–181. Shanghai: The Eastern Publisher. Cheng, Lisa L.-S. 1991. On the typology of wh- questions. Doctoral dissertation. MIT. 2003. Wh- in-situ. Glot International 7: 103–109. Cheng, Lisa L.-S. and James C.-T Huang. 1996. Two types of donkey sentences. Natural Language Semantics 4: 121–163. Cheng, Lisa L.-S., C.-T James Huang, Y.-H. Audrey Li, and C.-C. Jane Tang. 1993. Three ways to get passive. MS, University of California, Irvine. Cheng, Lisa L.-S., and Johan Rooryck. 2002. Types of wh- in-situ. MS, Leiden University. Cheng, Lisa L.-S., and Rint Sybesma. 1999. Bare and not-so-bare nouns and the structure of NP. Linguistic Inquiry 30: 509–542. Cheung, Samuel H.-N. 1977. Perfect particles in the Bian Wen language. Journal of Chinese Linguistics 5: 55‒74. Chierchia, Gennaro. 1998a. Plurality of mass nouns and the notion of “semantic parameter.” In Susan Rothstein (ed.), Events and grammar (Studies in Linguistics and Philosophy 70), 53–103. Dordrecht: Kluwer. 1998b. Reference to kinds across languages. Natural Language Semantics 6: 339–405. Chomsky, Noam. 1957. Syntactic structures. The Hague: Mouton. 1964. Current issues in linguistic theory. The Hague: Mouton. 1965. Aspects of the theory of syntax. Cambridge: MIT Press. 1976. Conditions on rules of grammar. Linguistic Analysis 2: 303–351. 1995. The minimalist program. Cambridge: MIT Press. 2000. Minimalist inquiries: the framework. In Roger Martin, David Michaels, and Juan Uriagereka (eds.), Step by step: essays in minimalist syntax in honor of Howard Lasnik, 89–155. Cambridge: MIT Press. Chomsky, Noam and Morris Halle. 1968. The sound pattern of English. New York: Harper and Row. Chu, Zexiang, Liu Jingsheng, Long Guofu, Tian Hui, Ye Guilin, and Zheng Xianzhang. 1997. Hanyu cunzaiju de lishixing kaocha (A diachronic investigation of the existential sentences in Chinese). Gu Hanyu Yanjiu (Studies in Classical Chinese) 4: 13–20. Clark, John and Colin Yallop. 1992. An introduction to phonetics and phonology. Oxford: Blackwell. Claudi, Ulrike. 1985. Zur entstehung von genussystemen: Überlegungen zu einigen theoretischen aspekten, verbunden mit einer fallstudie des zande. Hamburg: Helmut Buske. Clements, George N. 1990. The role of the sonority cycle in core syllabification. In John Kingston and Mary Beckman (eds.), Papers in laboratory phonology 1: between the grammar and physics of speech, 283–333. Cambridge: Cambridge University Press. Cole, Peter and Gabriella Hermon. 1998. The typology of wh- movement: wh- questions in Malay. Syntax 1: 221–258. Comrie, Bernard. 1981. Language universals and linguistic typology. Chicago: The University of Chicago Press. 1988. Topics, grammaticalized topics, and subjects. In Shelley Axmaker, Annie Jaisser and Helen Singmaster (eds.), General session and parasession on grammaticalization, 265–279. Berkeley: Berkeley Linguistics Society. Corbett, Greville. 2004. Number. Cambridge: Cambridge University Press. Chung, Sandra. 1976. An object-creating rule in Bahasa Indonesia. Linguistic Inquiry 7: 41–87.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

561

Croft, William. 1990. Typology and universals. Cambridge: Cambridge University Press. 1994. Semantic universals in classifier systems. Word 45:145–171. 1995. Intonation units and grammatical knowledge. Linguistics 33: 838–882. 1996. Typology and universals (2nd edition). Cambridge: Cambridge University Press. 2001. Radical construction grammar: syntactic theory in typological perspective. New York: Oxford University Press. 2004. Cognitive linguistics. Cambridge: Cambridge University Press. Crystal, David. 2008. A dictionary of linguistics and phonetics (6th edition). Oxford: Blackwell. Dai, Qingxia. 1998. Zang Mian yuzu yuyan yanjiu (Studies in the Tibetan-Burmese language family). Kunming: Yunnan Minzu Chubanshe. Dalrymple, Mary. 2001. Lexical-functional grammar (Syntax and Semantics, Vol. 34). London: Academic Press. Dayal, Veneeta. 2014. Questions and the vP phase in Hindi–Urdu. MS, Rutgers University. Déchaine, Rose-Marie. 1993. Serial verb constructions. In Joachim Jacobs, Arnim von Stechow, Wolfgang Sternefeld, and Theo Vennemann (eds.), Syntax: an international handbook of contemporary research, 799–925. Berlin and New York: De Gruyter Mouton. DeLancey, Scott. 2018. Evidentiality in Tibetic. In Alexandra Aikhenvald Y. (ed.), The Oxford handbook of evidentiality, 580–594. Oxford: Oxford University Press. Diessel, Holger. 1999a. Demonstratives: form, function, and grammaticalization. Amsterdam and Philadelphia: John Benjamins Publishing Company. 1999b. The morphosyntax of demonstratives in synchrony and diachrony. Linguistic Typology 3: 1–49. 2006. Demonstratives. In Keith Brown (ed.), Encyclopedia of language and linguistics, 430‒435. Amsterdam: Elsevier. Dik, Simon C. 1980. Studies in functional grammar. New York: Academic Press. Ding, Shengshu. 1935. Shi foudingci “fu,” “bu” (An explanation of the negative markers fu and bu). Qingzhu Cai Yuanpei xiansheng liushiwu sui lunwenji (A paper collection in celebration of Cai Yuanpei at the age of sixty-five), 967–996. Beijing: Guoli Zhongyang Yanjiu Yuan Lishi Yuyan Yanjiusuo. Ding, Shengshu, Lü Shuxiang, Li Rong, Sun Dexuan, Guan Xiechu, Fu Jing, Huang Shengzhang, and Chen Zhiwen. 1961. Xiandai Hanyu Yufa Jianghua (Lectures on modern Chinese grammar). Beijing: Shangwu Yinshuguan (The Commercial Press). Djamouri, Redouane and Paul Waltraud. 1997. Les syntagmes prépositionnels en yu et zai en chinois archaïque. Cahiers de linguistique-Asie orientale 26: 221–248. Dong, Hongyuan. 2009. Issues in the semantics of Mandarin questions. Doctoral dissertation. Cornell University, New York. Dryer, Matthew S. 1991. SVO languages and the OV: VO typology. Journal of Linguistics 27: 443–482. 1992. The Greenbergian word order correlations. Language, 68: 81–138. 2003. Word order in Sino-Tibetan languages from a typological and geographic perspective. In Graham Thurgood and Randy LaPolla (eds.), Sino-Tibetan languages, 43–55. Richmond: Curzon Press. 2005. Relationship between the order and verb and the order of relative clause and noun. In Martin Haspelmath, Matthew S. Dryer, David Gil, and Bernard Comrie (eds.), The world atlas of language structures, 390–93. Oxford: Oxford University Press. 2007. Word order. In Tomonthy Shopen (ed.), Language typology and syntactic description, Vol. I. Clause structure, 61–131. Cambridge: Cambridge University Press. Duanmu, San. 2007. The phonology of Standard Mandarin. Oxford: Oxford University Press. Duguine, Maia and Aritz Irurtzun. 2014. From obligatory wh- movement to optional wh- in-situ in Labourdin Basque. Language, 90: e1–e28.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

562

References

Durand, Jacques. 1990. Generative and non-linear phonology. London: Longman. Ehlich, Konrad. 1979. Verwendungen der Deixis beim sprachlichen Handeln. Bern: Lang. Faarlund, Jan Terje. 1990. Syntactic and pragmatic principles as arguments in the interpretation of runic inscriptions. In Jacek Fisiak (ed.), Historical linguistics and philology, 165–186. Berlin: Mouton de Gruyter. Fanselow, Gisbert. 2017. Partial movement. In Martin Everaert and Henk van Riemsdijk (eds.), The Blackwell companion to syntax, 437–492. Oxford: Blackwell. Feng, Chuntian. 1990. Shilun jiegou zhuci “di (de)” de yixie wenti (Some issues on the structural particle “di(de)”). Zhongguo Yuwen 6: 448–453. Feng, Shengli. 1996. Prosodically constrained syntactic change in Early Archaic Chinese. Journal of East Asian Linguistics 4: 323‒371. 2014. Historical syntax of Chinese. In C.-T. James Huang, Y.-H. Audrey Li, and Andrew Simpson (eds.), The Oxford handbook of Chinese linguistics, 537–575. Oxford: John Wiley and Sons, Inc. Frajzyngier, Zygmunt. 1987. From verb to anaphora. Lingua 72: 155–168. 1995. On sources of demonstratives and anaphors. In Barbara Fox (ed.), Studies in Anaphora, 169‒203. Amsterdm: John Benjamins Publishing Company. Frazier, Lyn. 1979. Parsing and constraints on word order. University of Massachusetts Occasional Papers in Linguistics 5: 177–198. Gao, Ming. 1987. Zhongguo guwenzixue tonglun (A Conspectus of Chinese ancient characters). Beijing: Wenwu Chubanshe. Geeraerts, Dirk and Hubert Cuyckens. 2007. The Oxford handbook of cognitive linguistics. Oxford: Oxford University Press. Givón, Talmy. 1979. On understanding grammar. New York: Academic Press. 1984. Syntax: a functional–typological introduction, Vol. 1. Amsterdam: John Benjamins. 1990. Syntax: a functional–typological introduction, Vol. 2. Amsterdam and Philadelphia: John Benjamins. 1991. Serial verbs and the mental reality of “event”: grammatical vs. cognitive packaging. In Elizabeth Traugott and Bernd Heine (eds.), Approaches to grammaticalization, Vol. I: Focus on theoretical and methodological issues, 81–127. Amsterdam: John Benjamins Publishing Company. Gnanadesikan, Amalia Elizabeth. 1997. Phonology with ternary scales. Doctoral dissertation, University of Massachusetts, Amherst. Goldberg, Adele E. 1995. Constructions: a construction grammar approach to argument structure. Chicago: The University of Chicago Press. 2006. Constructions at work: the nature of generalization in language. New York: Oxford University Press. Greenberg, Joseph. 1966a. Some universals of grammar with particular reference to the order of meaningful elements. Joseph H. Greenberg (ed.) Universals of languages (2nd edition), 73–113. Cambridge: MIT Press. (ed.) 1966b. Universals of language (2nd edition). Cambridge: MIT Press. 1978. How does a language acquire gender markers? In Joseph H. Greenberg, Charles A. Ferguson and Edith A. Moravcsik (eds.), Universals of human language, Vol. 3: Word structure, 47–82. Stanford: Stanford University Press. Greenberg, Joseph, Charles A. Ferguson and Edith A. Moravcsik (eds.). 1978. Universals of human language, Vol. 3: Word structure. Stanford: Stanford University Press. Grimshaw, Jane and Armin Mester. 1988. Light verbs and θ-marking. Linguistic Inquiry 19: 205–232. Grinevald, Colette. 2000. A morphosyntactic typology of classifiers. In Gunter Senft (ed.), Systems of nominal classification, 50–92. Cambridge: Cambridge University Press.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

563

Guo, Xiliang. 1997. Guanyu xici “shi” chansheng de shidai he laiyuan de jidian renshi (The date and source of the copula shi revisited). In Hanyu shi lunwen ji (A collection of papers on historical Chinese linguistics), 106–123. Beijing: The Commercial Press. Haiman, John. 1985. Natural syntax: iconicity and erosion. Cambridge: Cambridge University Press. Hale, Ken and Samuel J. Keyser. 1993. On argument structure and the lexical expression. In Richard Kayne, Raffaella Zanuttini, and Thomas Leu (eds.), An annotated syntax reader: lasting insights and questions, 312–327. Oxford: Wiley-Blackwell. Harris, Alice C. and Lyle Campbell. 1995. Historical syntax in cross-linguistic perspective. Cambridge: Cambridge University Press. Hashimoto, Mantaro. 1976. Language diffusion on the Asian continent. Computational Analysis of Asian and African Languages 3: 49–65. 1984. The Altaicization of northern Chinese, MS, National Inter-university Research Institute of Asian and African Languages and Cultures, U. Haspelmath, Martin. 1993. A grammar of Lezgian. Berlin: Mouton de Gruyter. Hawkins, John A. 1983. Word order universals. New York: Academic Press. 1984. Modifier–head or function–argument relations in phrase structure? Lingua 63: 107–138. Hayward, Richard J. 1990. Notes on the Aari language. In Richard J. Hayward (ed.), Omotic language studies, 425–93. London: School of Oriental and African Studies. He, Wei. 1989. Huojia fangyan yanjiu (A study in the Huojia dialect). Beijing: Shangwu Yinshuguan (The Commercial Press). 2004. Zuozhuan xuci yanjiu (Studies on the functional words of the Zuo Zhuan). Beijing: Shangwu Yinshuguan (The Commercial Press). 2005. Shi Ji yufa tedian yanjiu (Studies in the grammar of the Shi Ji). Beijing: Shangwu Yinshuguan (The Commercial Press). 2007. Hanyu yufashi duandai zhuanshu bijiao yanjiu (A comparative approach to the history of Chinese grammar with special reference to particular books in different periods). Kaifeng: Henan University Press. 2012. Zuozhuan yufa yanjiu (Studies in the grammar of the Zuo Zhuan). Kaifeng: Henan University Press. Heffner, Roe-Merrill S. 1950. General phonetics. Madison: University of Wisconsin Press. Heine, Bernd, Ulrike Claudi, and Friederike Hünnemeyer. 1991. Grammaticalization: a conceptual framework. Chicago: The University of Chicago Press. Heine, Bernd and Tania Kuteva. 2002. World lexicon of grammaticalization. Cambridge: Cambridge University Press. 2007. The genesis of grammar: a reconstruction. New York: Oxford University Press. Heine, Bernd and Heiko Narrog. 2011. The Oxford handbook of grammaticalization. Oxford: Oxford University Press. Heine, Bernd and Mechthild Reh. 1984. Grammaticalization and reanalysis in African languages. Hamburg: Buske. Hengeveld, Kees. 1992. Non-verbal predicate: theory, typology, diachrony. Berlin and New York: Mouton de Gruyter. Hintzman, Douglas L. 1986. “Schema abstraction” in a multiple-trace memory model. Psychological Review 93: 411–428. Hoffmann, Thomas and Graeme Trousdale. 2013. Handbook of construction grammar. Oxford: Oxford University Press. Hogg, Richard M., Norman Blake, Roger Lass, Suzanne Romaine, Robert Burchfield, and John Algeo. 2001. The Cambridge history of the English language. Cambridge: Cambridge University Press.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

564

References

Holm, John. 1988. Pidgins and creoles, Vol. 3: Theory and structure. Cambridge: Cambridge University Press. Hopper, Paul and Elizabeth Closs Traugott. 2003. Grammaticalization. Cambridge: Cambridge University Press. Horvath, Julia. 1986. Focus in the theory of grammar and the structure of Hungarian. Dordrecht: Foris. Hu, Chirui. 2005. Dongjie shi de zao qi xingshi jiqi panding biaozhun (The earliest forms of the resultative construction and the criteria for judging them). Zhongguo yuwen 3: 214‒225. Hu, Yushu (eds.). 1981. Xiandai hanyu (Modern Chinese). Shanghai: Shanghai jiaoyu chubanshe. Huang, Borong (ed.). 1996. Hanyu fangyan yufa leibian (The classification collection of the grammars of the Chinese dialects). Qingdao: The Qingdao Press. Huang, Dinghua. 1963. Minnan fangyan li de yiwen daici (Interrogative pronouns in the southern Min dialect), Zhongguo yuwen 4: 299–307. Huang, James C.-T. 1982. Logical relations in Chinese and the theory of grammar. Doctoral dissertation, MIT. 1994. On lexical structure and syntactic projection. In Proceedings of the Third International Symposium of Chinese Languages and Linguistics, 45‒48. Taipei: Academia Sinica. 1997. On lexical structure and syntactic projection. Chinese Languages and Linguistics 3: 45–89. 2006. The macro-history of Chinese syntax and the theory of change. Invited lecture presented at Workshop on Chinese Linguistics, University of Chicago. Huang, James C.-T., Y.-H Audrey Li, and Yafei Li. 2009. The syntax of Chinese. Cambridge: Cambridge University Press. Huang, James C.-T., Y.-H. Audrey Li, and Andrew Simpson. 2014. The Oxford handbook of Chinese linguistics. Oxford: John Wiley and Sons, Inc. Huang, Nansong. 1994. Shilun duanyu zizhu chengju suoying jubei de ruogan yufa fanchou (On the grammatical categories of determining clauses as independent sentences). Zhongguo yuwen 6: 441–447. Huang, Shuan-Fan. 1978. Historical change of prepositions and the emergence of SOV order. Journal of Chinese Linguistics 6: 212–242. Huang, Xiaohui. 1992. Xiandai Hanyu chabi geshi de laiyuan ji yanbian (The origin and development of the superior comparative construction in Modern Chinese). Zhongguo yuwen 3: 213–224. Hudson, Richard. 1992. So-called “double objects” and grammatical relations. Language 68: 251–276. Hulst, Harry van der. 1984. Syllable structure and stress in Dutch. Cinnaminson, NJ: Foris Publications. Hyman, Larry M. and Comrie, Bernard. 1981. Logophoric reference in Gokana. Journal of African Languages and Linguistics 3: 19–37. Jakobson, Roman and Linda R. Waugh. 1979. The sound shape of language. Bloomington: Indiana University Press. Jayaseelan, Karattuparambil A. 1996. Question-word movement to focus and scrambling in Malayalam. Linguistic Analysis, 26: 63–83. Jenks, Peter. 2018. Articulated definiteness without articles. Linguistic Inquiry 49: 501–536. Jespersen, Otto. 1922. Language. London: Allen and Unwin. Jiang, Lansheng. 1989. Beidong guanxici “chi” de laiyuan chutan (A basic study of the lexical source of the passive morpheme “chi”). Zhongguo yuwen 5: 370–377. 1995. Shuo “me” yu “men” tongyuan (On the same origin of “me” and “men”). Zhongguo yuwen 3: 259–273.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

565

1999a. Chusuo ci de lingge yongfa yu jiegou zhuci “di” de youlai (The genitive usage of locative words and the origin of the structural particle “di”). Zhongguo yuwen 2: 83–93. 1999b. Cong yuyan shentou kan Hanyu binishi de fazhan (The development of analogy in the Chinese language as seen from language infiltration). Zhongguo shehui kexue (Social sciences in China) 4: 169–179. Jiang, Shaoyu. 2005a. “Gei” ziju, “jiao” ziju biao beidong de laiyuan (The lexical sources of the passive morphemes “gei” and “jiao”). Yuyanxue luncong (Linguistic studies) 1: 159–177. 2005b. Jindai Hanyu yanjiu gaiyao (A survey of research on Modern Chinese). Beijing: Peking University Press. Jiang, Shaoyu and Cao Guangshun. 2005. Jindai Hanyu yufa shi yanjiu zongshu (A survey of studies about grammar of Modern Chinese). Beijing: Shangwu Yinshuguan (The Commercial Press). Katz, Jerrold J. and Paul M. Postal. 1964. An integrated theory of linguistic descriptions. Cambridge, MA: MIT Press. Kayne, Richard S. 1994. The antisymmetry of syntax. Cambridge, MA: MIT Press. Kayser, Alois (eds.). 1993. Nauru grammar. Yarralumla and Canberra: Embassy of the Federal Republic of Germany. Keenan, Edward L. 1985a. Passive in the world’s languages. In Timothy Shopen (ed.), Language typology and syntactic description, 243–81. Cambridge: Cambridge University Press. 1985b. Relative clauses. In Timothy Shopen (ed.), Language typology and syntactic Description, Vol. 2, 141–170. Cambridge: Cambridge University Press. Keenan, Edward and Bernard Comrie. 1979. Data on the noun phrase accessibility hierarchy. Language 55: 333–351. Kemenade, Ans van and Bettelou Los (eds.). 2006. The handbook of the history of English. Oxford: Blackwell Publishing Ltd. Kemenade, Ans van and Nigel Vincent (eds.). 1997. Parameter of morphosyntactic change. Cambridge: Cambridge University Press. Kenstowicz, Michael. 1994. Phonology in generative grammar. Oxford: Blackwell. Kidwai, Ayesha. 2000. XP Adjunction in universal grammar: scrambling and binding in Hindi/ Urdu. Oxford and New York: Oxford University Press. Kiparsky, Paul. 1979. Metrical structure assignment is cyclic. Linguistic inquiry 10: 421‒41. 1992. Analogy. In Keith Brown (ed.), The encyclopedia of language and linguistics, 56‒61. Amsterdam: Elsevier Science. 2011. Grammaticalization as optimization. In Diane Jonas, John Whitman, and Andrew Garrett (eds.), Grammatical change: origins, nature, outcomes, 14–51. Oxford: Oxford University Press. 2013. Towards a null theory of the passive. Lingua 125: 7–33. Kirsner, Robert S. 1976. On the subject of pseudo-passives in Standard Dutch and the semantics of background agents. In Charles N. Li (ed.), Subject and topic, 387–415. New York: Academic Press. Kiss, Katalin É. 1987. Configurationality in Hungarian. Budapest: Akadémiai Kiadó. Kong, Lingda. 1994. Yingxiang Hanyu juzi zizhu de yuyan xingshi (The linguistic forms affecting the independence of the sentence in Chinese). Zhongguo Yuwen 6: 434–440. Kornfilt, Jacklin. 1997. Turkish. London: Routledge. Kuteva, Tania, Bernd Heine, Bo Hong, Haiping Long, Heiko Narrog, and Seongha Rhee. 2019. World lexicon of grammaticalization (2nd edition). Cambridge: Cambridge University Press. Kytö, Merja and Päivi Pahta. 2016. The handbook of English historical linguistics. Cambridge: Cambridge University Press. Lakoff, Robin. 1971. Passive resistance. Chicago Linguistics Society 7: 149–162.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

566

References

Langacker, Ronald W. 1977. Syntactic reanalysis. In Li, Charles N. (ed.), Mechanisms of Syntactic Change, 57–139. Austin: University of Texas Press. 1982. Space grammar, analyzability, and the English passive. Language 58: 22–80. 1987. Foundations of cognitive grammar, Vol. 1: Theoretical prerequisites. Stanford: Stanford University Press. 1988. Usage-based model. In Brygida Rudzka-Ostyn (ed.), Topics in cognitive linguistics, 127–161. Amsterdam: John Benjamins. 1991. Foundations of cognitive grammar, Vol. 2: Descriptive application. Stanford: Stanford University Press. 1994. Conceptual grouping and constituency in cognitive grammar. Linguistics in the Morning Calm 3: 149–172. 1999. Grammar and conceptualization. Berlin: Mouton de Gruyter. 2008. Cognitive grammar: a basic introduction. New York: Oxford University Press. 2013. Essentials of cognitive linguistics. New York: Oxford University Press. Langendonck, Willy van. 2007. Iconicity. In Dirk Geeraerts and Hubert Cuyckens (eds.), The Oxford handbook of cognitive linguistics, 394–418. Oxford: Oxford University Press. Larson, Richard. 1988. On the double object construction. Linguistic Inquiry 19: 335–391. Lass, Roger. 1976. English phonology and phonological theory. Cambridge: Cambridge University Press. Lassadi, Boutheina. 2003. Optional wh- movement in French and Egyptian Arabic. Cahiers linguistiques d’Ottawa, 31: 67–93. Lehman, Winfred P. 1973. A structure principle of language and its implications. Language 49:42–66. 1978. The great underlying ground-plans. In Winfred P. Lehman (ed.), Syntactic typology, 3–55. Austin: University of Texas Press. Lehmann, Christian. 1995. Thoughts on grammaticalization. Munich: Lincom Europa. Levin, Beth and Malka Rappaport Hovav. 1990. Wiping the slate clean: a lexical semantic exploration. Cognition 41: 123–155. Levin, Juliette. 1985. A metrical theory of syllabicity. Doctoral dissertation, MIT. Li, Audrey Y. H. and Shi Yuzhi. 2000. Hanyu liangci xitong de jianli yu fushu biaoji “men” de fazhan (The establishment of the classifier system and the development of the plural “-men”). Contemporary Linguistics 2: 27–36. Li, Charles N. (eds.). 1975. Word order and word order change. Austin: University of Texas. (ed.). 1977. Mechanisms of syntactic change. Austin: University of Texas Press. Li, Charles N. and Sandra A. Thompson. 1974. Historical change of word order: a case study in Chinese and its implications. In John M. Anderson and Charles Jones (eds.), Historical linguistics, 199–217. Amsterdam: North-Holland. 1976. Subject and topic: a new typology of language. In Charles N. Li (ed.), Subject and topic, 457–489. New York: Academic Press. 1977. A mechanism for the development of copula morpheme. In Charles N. Li (ed.), Mechanism of syntactic change, 419–444. Austin: University of Texas Press. 1981. Mandarin Chinese: a functional reference grammar. Berkeley: University of California Press. Li, Charles N. and Shi Yuzhi. 1997. Lun Hanyu tibiaoji dansheng de jizhi (On the mechanism for the emergence of aspect markers in Chinese). Zhongguo yuwen 2: 82–96. Li, Fanggui. 1993. Certain phonetic influences of the Tibetan prefixes upon the root initials. Bulletin of the Institute of History and Philosophy 49: 15–36. Li, Ping. 1987. Shi shuo xin yu he Bai yu jing zhong de dongbu jiegou (The resultative construction in Shi shuo xin yu and Bai yu jing). Yuyanxue luncong 14: 129–157.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

567

Li, Rulong. 2001. Hanyu fangyanxue (The dialectology of Chinese).Beijing: The High Education Press. Li, Xinkui, Huang Jiajiao, Shi Qisheng, Mai Yun, and Chen Dingfang. 1995. Guangzhou fangyan yanjiu (Research on the Guangzhou dialect). Guangzhou: Guangdong renmin chubanshe. Li, Xuping and Walter Bisang. 2012. Classifiers in sinitic languages: from individuation to definiteness marking. Lingua 122: 335–355. Li, Yafei. 1993. Structural head and aspectuality. Language 69: 480–504. Li, Yin. 2018. The diachronic development of the Chinese passive: from the wei . . . suo passive to the long passive. Language 94: e74–e98. Li, Yuming. 1996. Lingshu guanxi yu shuangbinju fenxi (Genitive relation and the analysis of the ditransitive construction). Yuyan jiaoxue yu yanjiu (Language Teaching and Research), 3: 62–73. Li, Zhenhua and Zhou Changji. 1999. Hanzi gujin yinbiao (Phonological dictionary of Chinese characters from ancient to modern). Beijing: Zhonghua Shuju. Light, Timothy. 1979. Word order and word order change in Mandarin Chinese. Journal of Chinese Linguistics 7: 149–180. Lightfoot, David. 2013. Grammatical approaches to syntactic change. In Brian D. Joseph and Richard D. Janda (eds.), The handbook of historical linguistics, 495–508. Oxford: Blackwell. Lin, Jonah T.-H. 2001. Light verb syntax and the theory of phrase structure. Doctoral dissertation, University of California, Irvine. 2014. Light verbs. In James C.-T. Huang, Y.-H. Audrey Li, and Andrew Simpson (eds.), The Oxford handbook of Chinese linguistics, 73–99. Oxford: John Wiley and Sons, Inc. Lin, Jo-Wang. 1996. Polarity licensing and wh- phrase quantification in Chinese. Doctoral dissertation, University of Massachusetts at Amherst. 2014. Wh- expressions in Mandarin Chinese. In James C.-T. Huang, Y.-H. Audrey Li, and Andrew Simpson (eds.), The Oxford handbook of Chinese linguistics, 180–207. Oxford: John Wiley and Sons, Inc. Lin, Sue. 2018. Zaoqi Wuyu zhishici “ge” (The demonstrative “ge” in the Early Wu dialect). Fangyan 2: 221–230. Lipák, Anikó. 2002. On the syntax of wh- items in Hungarian. Glot International 6: 293–301. Lipka, Leonhard. 1990. An outline of English lexicology: Lexical structure, word Semantics, and word-formation. Tübingen: Niemeyer. Liu, Jian and Jiang Shaoyu. 1990. Jindai Hanyu yufa ziliao huibian (Tang Wudai juan) (A collection of grammatical material (from the Tang to the Five Dynasties)). Beijing: Shangwu yinshuguan (The Commercial Press). 1992. Jindai Hanyu yufa ziliao huibian (Songdai juan) (A collection of grammatical material (Song dynasty)). Beijing: Shangwu yinshuguan (The Commercial Press). 1995. Jindai Hanyu yufa ziliao huibian (Yuan Ming juan) (A collection of grammatical material (Yuan and Ming dynasties)). Beijing: Shangwu yinshuguan (The Commercial Press). Liu, Linjun. 2009. A cognitive approach to topic constructions in Beijing Mandarin. Doctoral dissertation, University of Manitoba, Winnipeg. Liu, Yuehua, Pan Wenyu, and Gu Ye. 2001. Shiyong xiandai Hanyu yufa (Practical Modern Chinese grammar). Beijing: Shangwu yinshuguan (The Commercial Press). Lombardi, Linda. 1999. Positional faithfulness and voicing assimilation in optimality theory. Natural Language and Linguistic Theory 17: 267–302. Lord, Carol Diabe. 1989. Syntactic reanalysis in the historical development of serial verb constructions in languages of West Africa. Doctoral dissertation, University of California, Los Angeles. 1993. Historical change in serial verb constructions. Amsterdam and Philadelphia: The John Benjiamins Publishing Company.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

568

References

Lu, Jianming. 1988. Xiandai Hanyu zhong shuliangci de zuoyong (The function of classifiers in Contemporary Chinese). In Lu Jianming (ed.), Xiandai Hanyu Jufa lun (On the syntax of Contemporary Chinese), 120–134. Beijing: Shangwu Yinshuguan (The Commercial Press). Lü, Shuxiang. 1963. Xiandai Hanyu danshuang yinjie wenti chutan (An approach on monosyllabic and disyllabic words in Modern Chinese). Zhongguo yuwen 1: 10–22. 1979. Xiandai Hanyu yufa yanjiu Wenti (Issues in research on Modern Chinese grammar). Beijing: Shangwu Yinshuguan (The Commercial Press). 1984. Hanyu yufa lunwenji (Collected papers on Chinese grammar). Beijing: Shangwu Yinshuguan (The Commercial Press). 1985. Jindai hanyu zhidaici (Pronominal words in Modern Chinese). Shanghai: The Xuelin Press. (ed.). 1999. Xiandai hanyu babai ci (The eight hundred words of Contemporary Chinese). Beijing: Shangwu Yinshuguan (The Commercial Press). Lucy, John A. 2000. Systems of nominal classification: a concluding discussion. In Gunter Senft (ed.), Systems of nominal classification, 326–41. Cambridge: Cambridge University Press. Ma, Zhen and Lu Jianming. 1996. Xingrongci zuo jieguo buyu qingkuang kaocha (An investigation into the resultative uses of adjectives). Hanyu xuexi 1: 3–7. Mallinson, Graham and Barry Blake. 1981. Language typology. Amsterdam: North-Holland. Mayerthaler, Willy. 1980. Ikonismus in der Morphologie. Zeitschrift für Semiotik 2: 19–37. 1988. Morphological naturalness. Ann Arbor, MI: Karoma. Mei, Tsulin. 1980. Sisheng bieyi de shijian cengci (The history of the tone shifts in differentiating meanings). Zhongguo Yuwen 6: 427–443. 1981. Xiandai Hanyu wanchengmao jushi he ciweide laiyuan (The origin of the perfective aspect construction and the perfect suffix in Modern Chinese). Yuyan yanjiu 1: 65–77. 1988. Ciwei “di,” “de” de laiyuan (The origin of the suffixes “di” and “de”). Shiyusuo jikan (The Bulletin of the Institute of History and Philosophy) 1: 141–172. 1991. Cong Handai de dong sha dong si lai kan dongbu jiegou fazhan – jian lun zhonggu shiqi de shishou guanxi de zhongli hua (The historical development of the verb–resultative construction, with a note on the neutralization of the pre-verbal agent/patient distinction in Medieval Chinese). Yuyanxue luncong 16: 112–36. Moravcsik, Edith A. 1978. Reduplicative constructions. In Joseph Greenberg (ed.), Universals of human language, Vol. 4, 297–334. Stanford: Stanford University Press. 2010. Explaining language universals. In Jae Jung Song (ed.), The Oxford Handbook of Linguistic Typology, 66–83. Oxford: Oxford University Press. Moro, Andrea. 2006. Copula sentences. In Martin Everaert and Henk van Riemsdijk (eds.), The Blackwell companion to syntax, Vol. 2, 1–23. Oxford: Blackwell. Mycock, Louise. 2007. Constituent question formation and focus: a new typological perspective. Transactions of the Philological Society 105: 192–251. Noël, Dirk. 2007. Diachronic construction grammar and grammaticalization theory. Functions of Language 14: 177–202. Norman, Jerry. 1988. Chinese. Cambridge: Cambridge University Press. Ohta, Tatsuo. 1987. Zhongguoyu lishi wenfa (The historic grammar of Chinese). Translated by Jiang Shaoyu and Xu Tianhua. Beijing: Peking University Press. Park, Insun. 1992. Constituency problems in the auxiliary verb construction in Burmese. Paper presented at the Twenty-Fifth Annual International Conference on Sino-Tibetan Languages and Linguistics, University of California, Berkeley. Parker, Stephen G. 2002. Quantifying the sonority hierarchy. Doctoral dissertation, University of Massachusetts, Amherst. Peirce, Charles Sanders. 1955. Philosophical writings of Peirce. New York: Dover.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

569

1974. Collected papers of Charles Sanders Peirce, vols. 1 and 2: Principles of philosophy and elements of logic. Cambridge: Harvard University Press. Perlmutter, David M. 1978. Impersonal passives and the unaccusative hypothesis. Proceedings of the annual meeting of the Berkeley Linguistics Society 38: 157‒189. Peyraube, Alain. 1986. History of the passive constructions in Chinese until the 10th century. Journal of Chinese Linguistics 17: 335–372. 1997a. On word order and word order change in Pre-archaic Chinese. Chinese Languages and Linguistics 4: 105‒124. 1997b. On word order in Archaic Chinese. Cahiers de linguistique Asie orientale 26.1: 3‒20. 2017. Periodization. In Rint Sybesma, Wolfgang Bether, Yueguo Gu, Zev Handel, C. T. James Huang, and James Myers (eds.), Encyclopedia of Chinese language and linguistics, Vol. 5, 346–349. Leiden: Brill. Pike, Eunice. 1954. Phonetic rank and subordination in consonant patterning and historical change. Miscellanea phonetica 2: 25–41. Plank, Frans. 1979. Exklusivierung, reflexivierung, identifizierrung, relationale Auszeichnung zu einem semantisch-pragmatischen thema. In Inger Rosengren (ed.), Sprache und Pragmatik Lunder Symposium, 330–54. Lund: CWK Gleerup. Pulleyblank, Edwin G. 1995. Outline of classical Chinese grammar. Vancouver: University of British Columbia Press. Qian, Nairong. 1997. Shanghaihua yufa (The grammar of the Shanghai dialect). Shanghai: The Shanghai People Press. 2014. Shanghai fangyan dingzhi zhishici “ge” (The definite demonstrative “ge” in the Shanghai dialect). Fangyan 1: 14–20. Qiao, Quansheng. 2000. Jin fangyan yufa yanjiu (Studies in the grammar of the Jin dialect). Beijing: The Commercial Press. Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985. A comprehensive grammar of the English language. London: Longman. Rankin, Robert L. 2003. The comparative method. In Brian D. Joseph and Richard D Janda (eds.), The handbook of historical linguistics, 183–212. Oxford: Blackwell Publishing Ltd. Ratcliff, Robert R. 2005. Bukhara Arabic: a metatypized dialect of Arabic in Central Asia. In É. Á. Csató, B. Isaksson, and C. Jahani (eds.), Linguistic convergence and diffusion: case studies from Iranian, Semitic, and Turkic, 141–159. London: Routledge. Rimé, Bernard, Pierre Philippot, and Cisamolo, Daniela. 1990. Social schemata of peripheral changes in emotion. Journal of Personality and Social Psychology 59: 38–49. Rizzi, Luigi. 1990. Relativized minimality. Cambridge: MIT Press. Ross, John Robert. 1967. Constraints on variables in syntax. Doctoral dissertation, MIT. Saeed, John I. 1987. Somali reference grammar. Wheaton: Dunwoody. Schachter, Paul. 1973. Focus and relativization. Language 49: 19–46. Selkirk, Elisabeth. 1984. On the major class features and syllable theory. In Mark Aronoff and Richard T. Oehrle (eds.), Language sound structure: studies in phonology presented to Morris Halle by his teacher and students, 107–136. Cambridge: MIT Press. Senft, Gunter. 1996. Classificatory particles in Kilivila. New York and Oxford: Oxford University Press. Serruys, Paul L.-M. 1981. Toward a grammar of the Shang bone inscriptions. Proceedings of the International Conference on Chinese Studies, 313–364. Taipei: Academia Sinica. Shao, Jingmin. 1996. Xiandai Hanyu yiwenju yanjiu (Studies in interrogative sentences in modern Chinese). Shanghai: Huadong Shifan Daxue Chubanshe (The Eastern China Normal University Press).

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

570

References

Shao, Jingmin and Zhao Chunli. 2005. Zhishi ba ziju he shengyin bei ziju jiqi yuyong jieshi (The causative disposal construction and absent passive construction: a pragmatic approach). Hanyu xuexi 4: 11–18. Shen, Jiaxuan. 2006. “Wang Mian sile fuqin” de shengcheng fangshi (The derivation of “Wang Mian died his father”). Zhongguo yuwen 4: 291–300. Shi, Youwei. 1984. Guanyu “dong + you” (On the “verb + have” construction). Yuyanxue luncong 13: 13–25. Shi, Yuzhi. 1995. Lun Hanyu de dayinjie jiegou (On the structure of macro-syllables in Chinese). Zhongguo Yuwen 3: 230–240. 2002a. The establishment of Modern Chinese grammar: the formation of the resultative construction and its effects. Amsterdam: John Benjamins Publishing Company. 2002b. Liangci, zhishi daici he jiegou zhuci (Classifier, demonstrative, and structural particle). Fangyan 2: 117‒126. 2004. Han Ying shuangbin jiegou chabie de gainianhua yuanyin (The motivation of conceptualization for the differences in double-object construction between Chinese and English). Waiyu jiaoxue yu yanjiu (Foreign Language Teaching and Research) 36: 83–89. 2010. Hanyu yufa (Chinese grammar). Beijing: Shangwu yinshuguan (The Commercial Press). 2016. Hanyu yufa yanhuashi (The evolutionary history of Chinese grammar). Nanchang: Jiangxi Education Press. Shi, Yuzhi and Charles N. Li. 1998. Lun zhuci “zhi” “zhe” he “de” de xingti (The rise and decline of the structural particles “zhi,” “zhe” and “de”). Zhongguo shehui kexue (Social Sciences in China) 6: 165–180. 2001. Hanyu yufanhua de licheng (A history of grammaticalization in Chinese). Beijing: Peking University Press. Shibatani, Masayoshi. 1985. Passives and related constructions: a prototype analysis. Language 61: 821–848. Shimura, Rioji. 1967. Chugo Hango no gobo to goi (A grammar of Medieval Chinese). In Ushijima Tokuji et al. (eds.), Chugoku bunka sosho 1 (Series on Chinese culture, Vol. 1), 212–241. Tokyo: Taishukan. Shopen, Timothy (eds.). 1985. Language typology and syntactic description. Cambridge: Cambridge University Press. Siewierska, Anna. 1984. The passive. a comparative linguistic analysis. London: Croom Helm. 1991. Functional grammar. London: Routledge. Simpson, Andrew. 2017. Bare classifier/noun alternations in the Jinyun (Wu) variety of Chinese and the encoding of definiteness. Linguistics 55: 305–331. Simpson, Andrew and Tanmoy Bhattacharya. 2003. Obligatory overt wh- movement in a whin-situ language. Linguistic Inquiry 34: 127–142. Simpson, Andrew and Priyanka Biswas. 2016. Bare nominals, classifiers, and the representation of definiteness in Bangla. Linguistic Analysis 40: 167–198. Simpson, Andrew, Hooi Ling Soh, and Hiroki Nomoto. 2011. Bare classifiers and definiteness. Studies in Language 35:168–193. Smith, Neil. 1981. Consistency, markedness and language change: on the notion “consistent language.” Journal of Linguistics 17: 39–54. Song, Jae Jung (eds.). 2010. The Oxford handbook of linguistic typology. Oxford: Oxford University Press. Song, Wenhui, Luo Zhengjing, and Yu Jingchao. 2007. Xiandai Hanyu beidong ju shishi yinxian de jiliang fenxi (A quantitative analysis of the presence and absence of agents in the passives of Contemporary Chinese). Zhuoguo yuwen 2: 114–124. Spencer, Andrew. 1995. Morphological theory. Oxford: Blackwell Pulishers Ltd.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

571

Stiebels, Barbara. 2002. Typologie des argumentlinkings: Ӧkonomie und Expressivität. Berlin: Akademie. Sun, Chaofen. 1996. Word order change and grammaticalization in the history of Chinese. Stanford: Stanford University Press. Sun, Chaofen and Talmy Givón. 1985. On the so-called SOV word order in Mandarin Chinese: a quantified text study and its implications. Language, 61: 329–351. Sun, Xixin. 1997. Hanyu lishi yufa conggao (A collection of papers on historical Chinese grammar). Shanghai: Hanyu dacidian chubanshe. Tai, James H-Y. 1985. Temporal sequence and Chinese word order. In John Haiman (eds.), Iconicity in syntax, 49–72. Amsterdam: John Benjamins. Tang, Yuming. 1987. Han Wei Liuchao beidongshi luolun (A basic study on the passives of the Han, the Wei, and the Six Dynasties). Zhongguo yuwen 3: 216–222. 1988. Tang zhi Qing de beidongju (Passives from the Tang Dynasty to the Qing Dynasty). Zhongguo yuwen 6: 459–468. Tang, Yuming and Zhou Xifu. 1985. Shilun Qin–Han beidongshi de fazhan (On the development of the passive in the Qin and Han dynasties). Zhongguo yuwen 4: 281–285. Thompson, Sandra A. and Robert Longacre. 1985. Adverbial clauses. In Timothy Shopen (ed.), Language typology and syntactic description. Vol. 2, 171–234. Cambridge: Cambridge University Press. Tomasello, Michael. 2003. Constructing a language: a usage-based theory of language acquisition. Cambridge: Harvard University Press. Traugott, Elizabeth C. 1982. From propositional to textual meanings: some semantic and pragmatic aspects of grammaticalization. In W. P. Lehmann and Y. Malkiel (eds.), Perspectives on historical linguistics,245–271. Amsterdam: John Benjamins. 1994. Grammaticalization and lexicalization. In Keith Brown (ed.), The encyclopedia of language and linguistics, 1481–1486. Amsterdam: Elsevier Science. 1997. Lexicalization and gramaticalization. MS, Stanford University. 2007. The concepts of constructional mismatch and type-shifting from the perspective of grammaticalization. Cognitive Linguistics 18: 523–557. 2011. Grammaticalization and mechanisms of change. In Bernd Heine and Heiko Narrog (eds.), The Oxford Handbook of Grammaticalization, 44–52. Oxford: Oxford University Press. Traugott, Elizabeth and Bernd Heine. 1991. Approaches to grammaticalization, Vol. 1: Focus on theoretical and methodological issues. Amsterdam: John Benjamins Publishing Company. Traugott, Elizabeth C. and Graeme Trousdale. 2013. Constructionalization and constructional changes. Oxford: Oxford University Press. Traunmüller, Hartmut. 1996. Sound symbolism in deictic words. Paper presented at Fonetik 96, Swedish Phonetics Conference, Nässlingen, May. Trousdale, Graeme. 2008. Constructions in grammaticalization and lexicalization: evidence from the history of a composite predicate construction in English. In Graeme Trousdale and Nikolas Gisborne (eds.), Constructional approaches to English grammar, 33–67. Berlin: Mouton de Gruyter. Trousdale, Graeme and Nikolas Gisborne (eds.). 2008. Constructional approaches to English grammar. Berlin: Mouton de Gruyter. Tsao, Feng-Fu. 1990. Sentence and clause structure in Chinese: a functional perspective. Taipei: Student Books. Tsujimura, Natsuko. 1999. The handbook of Japanese linguistics. Oxford: Blackwell Publishing Ltd. Vennemann, Theo. 1974. Analogy in generative grammar: the origin of word order. Proceedings of the Eleventh International Congress of Linguistics, Bologna, 79–83.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

572

References

1976. Categorial grammar and the order of meaningful elements. In Alphone Juilland (ed.), Linguistic studies offered to Joseph Greenberg on the occasion of his sixtieth birthday, 615–634. Saratoga, CA: Anma Libri. Vincent, Nigel. 1979. Word order and grammatical theory. In Jürgen M. Meisel and Martin D. Pam (eds.), Linear order and generative theory, 1‒22. Amsterdam: John Benjiamins Publishing Company. Voisin, Sylvie and Alice Vittrant. 2012. Pluriactionnalité temporelle en birman et wolof: les cas de repetition extrême. Bulletin de la Société linguistique de Paris 107.1: 353–377. Wahba, Wafaa A.-F. B. 1991. LF movement in Iraqi Arabic. In James C.-T. Huang and Robert May (eds.), Logical structure and linguistic structure, 253–76. Dordrecht: Kluwer. Wang, Guosheng. 1991. Daye Jinhu hua de “de”, “ge” he “de ge” (The particles “de,” “ge” and “dege”). Zhongguo yuwen 3: 211–215. Wang, Li. 1943(1985). Zhongguo xiandai yufa (Modern Chinese grammar). Beijing: Shangwu Yinshuanguan. 1980. Hanyu shigao (A history of the Chinese language). Beijing: Zhonghua Shuju. 1978. Tongyuan zi lun (On cognates in Chinese). In Yuyanxue lunwen ji of Wang Li (The collected linguistic papers of Wang Li), 533–546. Beijing: The Commercial Press. 1989. Hanyu yufa shi (The history of Chinese grammar). Beijing: Shangwu Yinshu Guan. (ed.). 1997. Tongyuan zidian (A dictionary of Chinese cognates). Beijing: The Commercial Press. Wang, Tongshang and Shi Yuzhi. 2019. Jindai Hanyu “chi” zi beidongju xingshuai de yuanyin (The motivations for the rise and decline of the “eat” passive in Modern Chinese). Gu hanyu yanjiu (Research on the Ancient Chinese Language) 2: 40–50. Wang, William S.-Y. and Chaofen Sun (eds.). 2015. The Oxford handbook of Chinese linguistics. Oxford: Oxford University Press. Watanabe, Akira. 1992. Wh- in-situ, subjacency, and chain formation (MIT Occasional Papers in Linguistics 2). Cambridge, MA. Watters, John. 1979. Focus in Aghem. In L. Hyman (ed.), Aghem grammatical structure (Southern California Occasional Papers in Linguistics 7), 138–189. Los Angeles: Department of Linguistics, University of Southern California. Wei, Pei-chuan. 1999. Lun xianqin hanyu yunfu de weizhi (On positions of the operator in Old Chinese). In A. Peyraube and S. Chaofen (eds.), Linguistic essays in honor of Mei Tsu-Lin: studies on Chinese historical syntax and morphology, 259–297. Paris: École des hautes études en sciences sociales. Woodworth, Nancy. 1991. Sound symbolism in proximal and distal forms. Linguistics 29: 273–299. Wu, Fuxiang. 2003. Zai lun chuzhi shi de laiyuan (On the source of the disposal construction). Yuyan yanjiu 3. 1–14. Wu, Yunji. 2000. Hunan fangyan de daici (Pronominal words in Hunan Province). Changsha: Hunan Normal University Press. Wunderlich, Dieter. 1997. Cause and the structure of verbs. Linguistic Inquiry 28: 27–68. 2006. Argument hierarchy and other factors determining argument realization. In I. Bornkessel, M. Schlesewsky, B. Comrie, and A. D. Friederici (eds.), Semantic role universals and argument linking: Theoretical, Typological, and Psycholinguistic Perspectives, 1–25. Berlin: De Gruyter. Xu, Dan. 2006. Typological change in Chinese syntax. Oxford and New York: Oxford University Press. Xu, Jialu (ed.). 1995. Wenbai duizhao zhuzi jicheng (The translation of famous ancient books into Contemporary Chinese). Guangxi: Guangxi Education Press.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

References

573

Xu, Jie. 1999. Liangzhong baoliu binyu jushi ji xiangguan jufa lilun (Two constructions of preserving objects and the related theories). Contemporary Linguistics 1:16–29. 2004. Yuyishang de Tongzhi Guanxi yu Jufashang de Shuangbinyu Jushi (The semantic relationship of co-reference and the syntactic construction of double objects). Zhongguo yuwen 4: 302–313. Xu, Jie and Li Ying-Che. 1993. Focus and two non-linear grammatical categories: negation and question. Zhongguo yuwen 3: 81–92. Xu, Liejiong. 2015. Topic prominence. In William S.-Y. Wang and Chaofen Sun (eds.), The Oxford handbook of Chinese linguistics, 393–404. Oxford: Oxford University Press. Xu, Liejiong and Terence Langendeon. 1985. Topic structures in Chinese. Language 61: 1–27. Yan, Sen. 1993. Lichuan yangyan yanjiu (Studies in the Lichuan dialect). Beijing: The Social Science Document Publisher. Yang, Bojun. 1981. Gu Hanyu Xuci (Functional words in Classic Chinese). Beijing: Zhonghua Shuju. Yang, Bojun and He Leshi. 2001. Guhanyu yufa jiqi fazhan (The grammar of Classical Chinese and its development). Beijing: Yuwen Chubanshe. Yang, Chengkai. 1996. Hanyu yufa lilun yanjiu (Studies in Chinese grammar theory). Shenyang: Liaoning Education Press. Ye, Guoquan and Tang Zhidong. 1982. Xinyi fangyan de bianyin (The tone sandhi in the Xinyi dialect). Fangyan (Dialect) 1: 47–51. Yen, Sian L. 1986. The origin of the copula shi in Chinese. Journal of Chinese Linguistics 14.2: 227–241. Yuan, Jiahua. 2001. Hanyu fangyan gaiyao (A survey of Chinese dialects). Beijing: Philology Press. Yue, Anne O. 1998. Zhi in pre-Qin Chinese. T’oung pao 84.4‒5: 239‒292. 2015. The Yue language. In William S.-Y. Wang and Chaofen Sun (eds.), The Oxford Handbook of Chinese Linguistics, 173–188. New York: Oxford University Press. Yue-Hashimoto, Anne. 1993. The lexicon in syntactic change: lexical diffusion in Chinese syntax. Journal of Chinese Linguistics 2: 213–254. 1995. Guangdong Kaiping fangyan de “de” zi jiegou (The “de” construction in the Kingping subdialect of Cantonese). Zhongguo yuwen, 4: 289–297. Zec, Draga. 1988. Sonority constraints on prosodic structure. Doctoral dissertation, Stanford University . Zhang, Bojiang. 2001. Lun “ba” ziju de jushi yuyi (On the meaning of the “ba” construction). Language Research 1: 28–40. 2006. Guanyu “suoqu lei shuangbinyu” (On the so-called “take”-type double object). Yuyanxue luncong 33: 298–312. Zhang, Huiying. 1997. Hanyu fangyan daici yanjiu (Studies on pronominal words in Chinese dialects). Fangyan 2: 88–96. 2001. Hanyu fangyan daici yanjiu (Studies in pronominal words in Chinese dialects). Beijing: The Yuwen Press. Zhang, Ning. 1998. The interactions between construction meaning and lexical meaning. Linguistics 36: 957–980. Zhang, Yujin. 2001. Xizhou Hanyu yufa yanjiu (Studies in the grammar of the West Zhou Dynasty). Beijing: Shangwu yinshuguan (The Commercial Press). Zhou, Chiming. 1958. Hanyu de shidongxing fuhe dongci (Causative verb compounds in Chinese). Wen shi zhe 4: 175–226. Zhu, Dexi. 1982. Yufa jiangyi (Lecture notes on grammar). Beijing: Shangwu Yinshuguan (The Commercial Press). 1983. Zizhi he zhuanzhi: Hanyu mingcihua miaoji “de, zhě, suǒ he zhī” de yufa gongneng he yuyi gongneng (Self-referred and other-referred: the grammatical and semantic functions of the four nominalizers “de, zhě, suǒ and zhī”). Fangyan [Dialects] 1: 16–31.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

574

References

Zhu, Minche. 1958. Xianqing liang Han shiqi de dongbu jiegou (The resultative construction in the period of Pre-Qin and Two Hans). Yuyanxue Luncong Vol. 2, 17–30. Shanghai: Xinzhi chubanshe. 1990. Zhuzi yulei zhong de dongci buyu – jian tan dongci houzhui (The resultative constructions and verbal particles in Zhuzi yulei). In Wang Li xiansheng jinian lunwen ji (A collection off papers in honor of Wang Li), 240‒251. Beijing: Shangwu yinshu guan. Zsiga, Elizabeth C. 2013. The sounds of language: an introduction to phonetics and phonology. Oxford: Wiley-Blackwell. Zwicky, Arnold M. 1982. Stranded to and phonological phrasing in English. Linguistics 20: 3‒57.

https://doi.org/10.1017/9781108921831.023 Published online by Cambridge University Press

PRIMARY SOURCES OF TEXTS

Bai Yu Jing Yi Zhu (百喻經譯註, The Sutra of a Hundred Parables), Beijing: Zhonghua Book 2008. Bao Pu Zi Nei Pian Jiao Shi (抱樸子內篇校釋, The Master of Preserving Simplicity), Beijing: Zhonghua Book Company, 1985. Bei Qi Shu (北齊書, The Northern Qi Dynasty Book), Beijing: Zhonghua Book Company, 1972. Cao Yu Xi Ju Xuan (曹禺戲劇選, Selected Plays by Cao Yu), Beijing: People’s Literature Publishing House, 1977. Cao Zhi Ji Zhu (曹植集註, Cao Zhi’s Poems), Beijing: Zhonghua Book Company, 1979. Chun Qiu Gong Yang Zhuan Zhu Shu (春秋公羊傳註疏, Gongyang Commentary on Spring and Autumn Annals), Beijing: Peking University Press, 1999. Chun Qiu Zuo Zhuan Zhu (春秋左傳注, The Zuo Tradition), Beijing: Zhonghua Book Company, 1990. Chu Yao Jing Zhu Shi Yu Bian Xi (出曜經注譯與辨析, Sutra of the Appearance of Light), Beijing: The Religious Culture Press, 2019. Da Hui Pu Jue Chan Shi Yu Lu (大慧普覺禪師語錄, Quotations of Zen Master Dahui Pujue), Shanghai: The Oriental Press, 2018. Dun Huang Bian Wen Ji (敦煌變文集, Anthology of Dunhuang Bianwen), Beijing: People’s Literature Publishing House, 1957. Er Nü Ying Xiong Zhuan (兒女英雄傳, Heroes and Heroines), Beijing: People’s Literature Publishing House, 2018. Er Shi Nian Mu Du Zhi Guai Xian Zhuang (二十年目睹之怪現狀, Strange Situations Witnessed over 20 years), Beijing: People’s Literature Publishing House, 1959. Gong Sun Long Zi Jiao Shi (公孫龍子校釋, The Works of Kung-Sun Lung-Tzu), Shanghai: Shanghai Classic Publishing House, 2001. Guan Han Qing Xi Ju Ji (關漢卿戲劇集, Guan Hanqing’s Drama Collection), Beijing: People’s Literature Publishing House, 1976. Guo Yu (國語, The Historical Documents of Nations), Shanghai: Shanghai Classic Publishing House, 1978. Han Fei Zi Ji Jie (韓非子集解, Han Feizi), Beijing: Zhonghua Book Company, 1998. Han Shan Shi Zhu (寒山詩註, The Poetry of Han-shan), Beijing: Zhonghua Book Company, 2000. Han Shu (漢書, History of the Former Han), Beijing: Zhonghua Book Company, 1960. Hong Lou Meng (紅樓夢, Dream of the Red Chamber), Beijing: People’s Literature Publishing House, 1990. Hou Han Shu (後漢書, History of the Latter Han), Beijing: Zhonghua Book Company, 1964. Hua Jian Ji Jiao (花間集校, Among the Flowers), Beijing: People’s Literature Publishing House, 1981. Huai Nan Zi Jiao Shi (淮南子校釋, Huai Nan Tzu), Beijing: The Peking University Press, 2013.

575

https://doi.org/10.1017/9781108921831.024 Published online by Cambridge University Press

576

Primary Sources of Texts

Jin Dai Han Yu Yu Fa Zi Liao Hui Bian – Song Dai Juan (近代漢語語法資料彙編 – 宋代卷, A Collection of Grammatical Materials in Modern Chinese – Song Dynasty Volume), Beijing: The Commercial Press, 1992. Jin Dai Han Yu Yu Fa Zi Liao Hui Bian – Tang Wu Dai Juan (近代漢語語法資料彙編 – 唐五代卷, A Collection of Grammatical Materials in Modern Chinese – Tang and Five Dynasties Volume), Beijing: The Commercial Press, 1990. Jin Dai Han Yu Yu Fa Zi Liao Hui Bian – Yuan Dai Ming Dai Juan (近代漢語語法資料彙編 – 元 代明代卷, A Collection of Grammatical Materials in Modern Chinese – Yuan and Ming Dynasties Volume), Beijing: The Commercial Press, 1995. Jin Ping Mei Ci Hua (金瓶梅詞話, Jin Ping Mei Poetry and Prose), Beijing: People’s Literature Publishing House, 1981. Jing Ben Tong Su Xiao Shuo (京本通俗小說, Popular Stories from the Capital Edition), Shanghai: Shanghai Classic Literature Press, 1955. Jing De Chuan Deng Lu (景德傳燈錄, The Records of the Transmission of the Lamp in the Chingte Era), Shanghai: Shanghai Bookstore Publishing Company, 2010. Lao Can You Ji (老殘遊記, Old Can’s Travel Notes), Beijing: People’s Literature Publishing House, 2020. Lao Qi Da (老乞大, Old Qida), Shanghai: The Sanlian Publishing House, 2002. Lao Zi Dao De Jing Zhu (老子道德經註, Tao Te Ching by Lao Tzu), Beijing: Zhonghua Book Company, 2008. Li Ji Yi Zhu (禮記譯註, Book of Rites), Shanghai: Shanghai Classic Publishing House, 2004. Lie Zi (列子, Lieh Tzu), Beijing: Zhonghua Book Company, 2007. Liu Zhi Yuan Zhu Gong Diao (劉知遠諸宮調, Liu Zhiyuan’s Palace Tunes), Beijing: People’s Literature Publishing House, 1978. Liu Zu Tang Jing (六祖壇經, The Sixth Patriarch’s Altar Sutra), Kong: The Phoenix Press, Hong 2010. Lun Heng Jiao Shi (論衡校釋, On Balance), Beijing: Zhonghua Book Company, 1990. Lun Yu Yi Zhu (論語譯註, The Analects of Confucius), Beijing: Zhonghua Book Company, 1980. Meng Zi Yi Zhu (孟子譯註, Mencius), Beijing: Zhonghua Book Company, 1988. Mo Zi Jiao Zhu (墨子校註, Mo-tse), Beijing: Zhonghua Book Company, 1993. Piao Tong Shi (樸通事, Park Tongshi), Shanghai: Shanghai Sanlian Publishing House, 2002. Qi Min Yao Shu (齊民要術, Essential Skills for Benefiting the People), Beijing: Zhonghua Book Company, 2015. Qian Long Da Zang Jing (乾隆大藏經, Tibetan Tripitaka), Beijing: The Heritage Press, 2005. Quan Song Ci (全宋詞, The Complete Collection of Song Dynasty Ci Poems), Beijing: Zhonghua Book Company, 1964. Quan Tang Shi (全唐詩, The Complete Collection of Tang Period Poems), Beijing: Zhonghua Book Company, 1999. Quan Yuan Xi Qu Zhang Xie Zhuang Yuan (全元戲曲張協狀元, The Complete Collection of Plays in the Yuan Dynasty – The Champion of Zhang Xie), Beijing: People’s Literature Publishing House, 1999. Ru Lin Wai Shi (儒林外史, Unofficial History of Confucian Scholars), Beijing: People’s Literature Publishing House, 1981. San Guo Yan Yi (三國演義, Romance of the Three Kingdoms), Beijing: People’s Literature Publishing House, 1973. San Guo Zhi (三國志, Records of the Three Kindoms), Beijing: Zhonghua Book Company, 1964. San Xia Wu Yi (三俠五義, Three Heroes and Five Righteousnesses), Beijing: People’s Literature Publishing House, 2001. Shang Shu Yi Zhu (尚書譯註, The Book of History), Shanghai: Shanghai Classic Publishing House, 2004.

https://doi.org/10.1017/9781108921831.024 Published online by Cambridge University Press

Primary Sources of Texts

577

Shi Ji (史記, The Records of the Grand Historian), Beijing: Zhonghua Book Company, 1959. Shi Jing Yi Zhu (詩經譯註, Book of Songs), Shanghai: Shanghai Classic Publishing House, 2012. Shi Shuo Xin Yu Jiao Jian (世說新語校箋, A New Account of Tales of the World), Beijing: Zhonghua Book Company, 1984. Shui Hu Zhuan (水滸傳, Outlaws of the Marsh), Beijing: People’s Literature Publishing House, 1997. Shui Jing Zhu (水經註, Commentary on the Waterways Classic), Beijing: Zhonghua Book Company, 1987. Song Yuan Xiao Shuo Jia Hua Ben Ji (宋元小說家話本集, Vernacular Novels in the Song and Yuan Dynasties), Jinan: Qilu Publishing House, 2000. Wang Fan Zhi Shi Jiao Zhu (王梵志詩校註, Wang Fanzhi’s Peoms), Shanghai: Shanghai Classic Publishing House, 2010. Wen Bai Dui Zhao Zhu Zi Ji Cheng (文白對照諸子集成, Collection of Masterpieces), Nanning: Guangxi Education Press, 1995. Xi Xiang Ji (西廂記, Romance of the West Chamber), Beijing: People’s Literature Publishing House, 1998. Xi You Ji (西遊記, Journey to the West), Beijing: People’s Literature Publishing House, 1980. Xin Kan Da Song Xuan He Yi Shi (新刊大宋宣和遺事, The Legacy of Xuanhe of the Great Song Dynasty), Shanghai: Shanghai Classic Publishing House, 1958. Xing Shi Heng Yan (醒世恆言, Stories to Awaken the World), Beijing: People’s Literature Publishing House, 1956. Xun Zi Ji Jie (荀子集解, Xun Kuang), Beijing: Zhonghua Book Company, 1988. Yan Shi Jia Xun (顏氏家訓, The Family Instructions of Master Yan), Beijing: Zhonghua Book Company, 2007. Yan Zi Chun Qiu Ji Shi (晏子春秋集釋, The Spring and Autumn Annals of Master Yan), Beijing: Zhonghua Book Company, 1962. Yi Jing (易經, Book of Changes), Beijing: Peking Joint Publishing Company, 2019. Yi Li Zhu Shu (儀禮註疏, Rites and Ceremonies), Beijing: The Peking University Press, 1999. Yuan Kan Za Ju San Shi Zhong (元刊雜劇三十種, Thirty Dramas Published in the Yuan Dynasty), Shanghai: The Commercial Press, 1958. Yue Fu Shi Xuan (樂府詩選, The Folk Poetry Selection), Beijing: People’s Literature Publishing House, 1953. Zhan Guo Ce (戰國策, Strategies of the Warring States), Beijing: Zhonghua Book Company, 2006. Zhu Zi Yu Lei (朱子語類, Classified Conversations of Master Zhu), Beijing: Zhonghua Book Company, 1986. Zhuang Zi Ji Shi (莊子集釋, South China Classic), Beijing: Zhonghua Book Company, 1985. Zu Tang Ji (祖堂集, Ancestral Hall Collection), Beijing: Zhonghua Book Company, 2008.

Electronic Corpora The Scripta Sinica Corpus of Classical Chinese full texts, Academia Sinica, Taiwan (http:// lingcorpus.iis.sinica.edu.tw/ancient). The Chinese Language Corpus, Centre for Chinese Linguistics, Peking University, Beijing (http:// ccl.pku.edu.cn:8080/ccl_corpus). The Chinese Language Corpus, Beijing Language University, Beijing (http://bcc.blcu.edu.cn).

https://doi.org/10.1017/9781108921831.024 Published online by Cambridge University Press

INDEX

accompanying feature, 164, 247, 314 action–resultative, 127, 163–173, 234, 245, 260, 264, 265, 275, 279, 282, 361, 404–407, 552 action–resultative ordering, 164 adposition, 274, 514, 536, 539, 541 affirmation, 12, 368–371, 373 agent noun, 175, 181, 185, 188, 195–200, 241–243, 260, 386, 544–546, 555, 556 analogy, 21–23, 29, 47, 57, 103, 139, 167, 171–174, 180, 193, 195, 199, 200–202, 206, 242, 290, 296, 300, 332–334, 339, 340, 351, 386–388, 400, 404, 418, 492–495, 506–508, 532, 548–551, 565 anaphor, xix, 10, 19–21, 44–46, 144–149, 229, 522, 547 assimilation, 567 associative particle, xix, 172, 394, 532 auxiliary, xix, 41, 57–61, 74, 120, 171, 180, 181, 182, 186–188, 197, 198, 201, 326, 327, 346, 348, 361, 362, 364–367, 369, 370, 372, 373, 374, 375, 379, 381, 382, 383, 384, 536, 541, 550, 554, 568 auxiliary verb, xix, 57–61, 74, 180, 181, 348, 364–367, 372, 568 Bai, 51, 55–57, 67, 182, 188, 219–221, 224, 229, 238, 242, 256, 277, 284, 289, 291, 310, 349, 360, 385, 433–436, 501, 508, 512, 533–536, 566, 575, 577 bare noun, 20, 151–157, 216, 219, 227–230, 231, 299, 422–427, 480, 550 borrowing, 126, 205–208, 352, 355–357, 359, 550 boundedness, 25, 214, 381, 387, 388–395, 548

Cantonese, 83, 221, 275, 354, 356, 357, 370, 415–417, 423, 424, 428, 443, 446, 459, 460, 480, 486, 493, 505, 560, 573 causative, xix, 71, 118–120, 150, 301, 570, 573 change in cluster, 546 Chomsky, 4, 27, 29–32, 48, 175–177, 461, 552, 554, 557, 560 cleft structure, 14 cline, 69–71, 75, 108, 127–129, 355–357, 464, 482–484, 545 clitic, 14, 73, 109, 327, 356, 464 cliticization, 36, 449, 456, 458, 464–466 cluster, 9, 69–71, 139, 338, 494, 540, 546–548 cognate, 448–450 cognitive linguistics, xvi, 1, 2–5, 8, 323–325, 553, 559, 562, 566 comparative, 167–169, 170, 171, 180, 193, 201, 227, 234, 260, 263, 273, 274, 275, 276, 277, 278, 279, 280, 281, 282, 284, 285, 286, 287, 288, 289, 291, 292, 392–395, 512–514, 535, 544–546, 549, 553, 555, 563, 564, 569, 570 comparative construction, 276, 279–281, 285, 287, 288, 404–407, 545, 549, 564 competition, 172, 243, 245, 326, 341, 415 complementizer, 16, 151, 481–483, 523 compound, 51–55, 68, 73, 74, 75–77, 78, 81, 83–87, 89–96, 99, 106, 107, 109, 114–116, 124, 127, 130, 134, 140, 157, 158, 185, 225, 233, 251, 269, 293, 321, 331, 334, 338, 339, 343, 363, 366, 370, 384, 398–402, 476, 481, 483, 484, 512, 536 compounding, 51, 67, 75, 78, 86, 88–90, 128, 225, 333, 351, 400–403, 483 conceptualization, 296, 315, 318, 319, 320, 323, 324, 370, 379, 482–484, 566, 570

578

https://doi.org/10.1017/9781108921831.025 Published online by Cambridge University Press

Index conjunction, xix, 23, 62–66, 138, 267–270, 329–332, 333, 348, 546–549 connective, xiii, 65–77, 87, 90, 138 constituent order, 23, 148–150, 164, 172, 206, 274, 297, 331–333, 367, 377, 398, 402–404, 511, 513–517, 539, 543, 547–549 construal, 3, 195–197, 200, 264 construction grammar, 1, 8, 227, 323–325, 559, 561, 562, 563, 568 construction schema, 531, 552–557 continuous aspect, 147, 350, 351 co-ordinate, 10, 41, 62, 66, 75, 90, 127, 138, 269, 305–307, 313, 326, 331–334, 343, 347, 351, 545–549 copula, xix, 10–15, 17–26, 28, 29, 34–39, 46, 48, 49–61, 126, 149–152, 176–178, 210–213, 329, 378, 414, 524, 545–548, 554, 563, 566, 573 copular construction, 10, 12–16, 17, 18–26, 50, 373, 377, 378, 547 copular particle, 11–13, 22, 149 copular verb, 24, 25, 47, 49, 376–379, 525 cross-category harmony principle, 514, 538–540 definite article, 412, 441 definiteness, 151, 152, 153–164, 218–222, 234, 239, 299, 345, 410, 420–427, 430, 432, 436–438, 470, 480, 559, 564, 570 definiteness assignment, 151, 154, 410 degree word, 266, 285, 390, 491, 505, 534 delimitative aspect, 265, 305, 329, 355 diminutive, xix, 327, 355, 358, 359, 478, 482–487, 548–551, 560 direct object, xx, 113, 170, 173, 208–210, 218, 223, 232, 260, 263, 296–300, 303, 305–312, 313, 314, 320 disposal construction, 9, 10, 23, 62, 86, 88, 114, 153–155, 171–174, 193–202, 203, 204, 205, 206, 208, 210, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 236, 237, 238, 239, 240, 241, 242, 243, 244, 246, 252, 253, 296, 298, 299, 306–309, 312–314, 327, 344, 386, 387, 410, 420, 423–427, 430, 437, 544, 550, 556, 559, 570, 572 disposal marker, xix, 151, 154, 173, 194, 203–205, 213, 224, 237, 242–244, 296, 300, 380

579

distal demonstrative, xiv, 211, 212, 359, 418–421, 439, 441–444, 446–449, 454–456, 459–463, 465–467 disyllabic word, 85, 483 disyllabification, 71, 76–77, 78, 79, 80–90, 93, 95, 97–99, 101, 103, 104, 124, 127, 134, 137–139, 225, 256, 269, 334, 337, 339–342, 351, 400–403, 452, 483, 546–549 ditransitive construction, 58, 173, 174, 208–210, 223, 225, 232, 246, 262–265, 294–295, 296, 297, 298, 299, 300, 301, 303, 304, 305, 306, 308, 310, 311, 312, 313, 314, 315, 317, 320, 322–325, 567 emphasis, 10–12, 23, 37, 143, 226 English, xiii, xviii, 1, 4, 5, 6, 8, 10–12, 16, 24, 28–33, 34, 40, 44, 54, 55, 56, 61, 62, 65, 66, 71, 85, 98, 106–108, 111–113, 120, 123, 126, 142–144, 148, 153, 156–158, 160, 161, 165, 168, 176–181, 190–192, 197, 199, 204, 212, 225, 233, 237, 243, 273, 286–289, 292, 294–295, 300, 301, 303, 304, 306, 307–311, 312–321, 323–325, 327–331, 345, 352, 359–362, 363, 364, 368–371, 373, 379, 386, 387, 394, 397, 405, 412, 419, 425, 435, 440–443, 452–454, 461, 467–469, 470, 473, 478, 479–482, 498, 502, 514–517, 519, 522–524, 527, 530, 534, 536–538, 551–555, 558, 559, 560, 563, 565, 566, 567, 569, 570, 571, 574 equal comparison, 276, 281 existential construction, 160, 161 experiential aspect, xix, 341, 353, 366, 381–384, 387 focus, xix, 9, 11, 12, 13–18, 23, 26, 28, 29, 33–42, 43–48, 49–57, 60, 102, 115, 122–124, 136, 142, 143–145, 147–152, 157, 181, 203, 209–211, 212, 222, 226, 253, 301, 346, 354, 376, 387, 396, 397, 407–409, 442, 450, 459, 464–466, 498, 520, 547, 558, 564, 568 French, 5, 18, 34–37, 56, 204, 442, 566 frequency, 3, 14, 18, 25, 29, 44, 47, 52, 57, 67, 94, 109, 124–128, 129, 130, 132–134, 136–139, 156, 160, 179, 184, 189, 200, 251, 269, 288, 294, 303, 334, 340, 350, 377, 381, 391, 398–401, 413, 419, 426–428, 448–450, 464, 473, 481, 488, 494, 544–546, 548, 553, 556, 559

https://doi.org/10.1017/9781108921831.025 Published online by Cambridge University Press

580

Index

functional word, 14, 280, 439, 488 fusion, 52–55, 62, 71, 73–75, 76–77, 78, 89–98, 100–103, 105, 108, 109, 124, 125, 127, 128, 131–141, 194, 203, 226, 229–231, 236, 242, 252, 253, 335–339, 374, 383, 426, 433, 448, 462, 464, 465, 504 future, 361–363 future tense, 361–363

inflection, 128, 413, 447, 462, 477, 483–484 information structure, 142, 143, 164, 172, 193, 279, 298, 299, 361, 404–407, 530, 548 instrumental construction, 296, 297 intensifier, 115, 289 interrogative pronoun, xxi, 11, 494, 501 inversion, 20, 29, 40, 42–44, 170, 204, 209, 214, 218, 231, 375–377

Gan dialect, 415–417, 443, 446 gender, 441, 562 generative linguistics, 2–5, 27, 33, 48, 61, 189, 553, 557 genitive, xx, 51, 104, 205, 359, 397, 434, 436–438, 455, 472–474, 488–490, 493–496, 498–502, 506, 509, 526, 529, 532, 537, 539, 549, 565 German, 5, 46, 369, 379, 539 given information, 15, 44, 142, 144, 148, 150, 226 glottal, 448, 460–463, 468 grammatical borrowing, 356 grammatical category, 3, 9, 11, 160, 172, 246, 265, 326, 327, 345, 355–357, 358–361, 369, 396, 401, 408, 421, 438, 448, 449, 471, 486, 493, 527, 530, 545, 548–551 Greenbergian universal, 511–513, 542

Japanese, 12, 30, 81, 179, 207, 240, 273, 364, 554, 571 Jin dialect, xiv, 357, 415–418, 443, 458, 459, 461, 464–466, 569

Hakka dialect, 356, 370, 415–417, 443, 446 head noun, 3, 205, 206, 393, 397–400, 402, 403, 416–418, 421, 422, 431–433, 472–474, 478–481, 489–493, 496, 501–508, 511–536, 539, 541, 548 high pitch, 142, 486 historical linguistics, xvi, 1, 7–9, 200, 205, 333, 411, 417, 546, 555, 559, 565, 567, 571 hypotaxis, 70, 73 iconicity, 36, 164, 440, 466, 467, 468 inchoative aspect, 350, 351 indefinite, 31–33, 143, 145, 151–162, 173, 176, 203, 209, 215–219, 221–223, 228, 230–232, 236, 240, 244, 247–249, 253, 256, 296, 298–300, 422–424, 426, 436, 441, 470, 471, 475, 482, 550 indefinite article, 160, 441, 482 indirect object, xx, 173, 263, 296, 297, 299–301, 303–318, 320, 323–325 inferior comparison, 286, 287 infix, 95–97, 110, 118

Korean, 12, 144, 207, 484, 559 language acquisition, 3, 383, 385, 544–546, 551, 552–554, 571 language contact, 5, 29, 88, 126, 178, 203–205, 207, 239–242, 290–292, 293, 535, 545 lexical source, 14, 199–201, 208, 211, 328, 346, 369, 411, 435, 439–441, 453, 475, 485, 489, 492, 533, 550, 564 lexicalization, 78, 84–87, 95, 125–129, 133, 135, 303, 400–402, 571 light verb, 38, 48, 55, 213 linguistic typology, 8, 167, 177–178, 204, 514, 542, 552, 560, 570 logic form, 61 long passive, 184, 186, 187–190, 567 Mandarin, xiv, xviii, 1, 31, 43–45, 55, 80, 83, 87, 88, 98, 104, 173, 175–177, 190, 193–196, 202, 219–221, 226, 236, 240, 245, 266, 270–272, 291–293, 305, 314, 319, 327, 354, 356, 357, 358–361, 363, 365, 370, 385, 411, 413, 415–417, 420–432, 436–438, 442, 443, 444, 447, 448, 454–456, 458, 462, 471–473, 476–482, 484, 500–502, 505, 561, 566, 567, 571 mechanism, 4, 7–9, 18, 27, 101, 135, 152, 177, 251, 326, 329, 340, 368, 397, 409, 411, 439, 440, 468, 469, 566 Min dialect, xiv, 16, 51–54, 369–371, 379, 415–417, 418, 443–446, 447, 458, 461, 462, 564 minority language, 535 modality, 345 Mongolian, 239, 293 mood, 11, 143, 267

https://doi.org/10.1017/9781108921831.025 Published online by Cambridge University Press

Index nasal sound, 486 natural laboratory, 2–5, 61 negation, 11, 12, 35–38, 51, 57, 60, 116, 165, 368–374, 377, 379–382, 404, 573 negative marker, 12, 129, 267, 368, 371, 374, 376, 378–381 neutral tone, 82, 128 new information, 15, 44, 142–144, 148, 157, 201, 226, 530–531, 555 nominal reduplication, 407, 410 nominalization, 521 nominalizer, xx, 51, 56, 392, 481–484, 489 nonreferential, 246–249, 253, 254, 256 northern dialect, 271, 355, 370, 415–417, 442–444, 445, 456–458, 466, 551 nucleus, 69–73, 122, 374, 433, 446–449, 450–455, 456–469, 483 onomatopoeia, 440, 453, 467 onset, 53, 79, 140, 374, 430, 444–449, 450–466, 468 parataxis, 73, 75 passive construction, 62, 173, 177, 178, 201, 313, 314, 424–427, 552, 570 passive marker, xix, xx, 179, 181–183, 185, 196, 200, 201, 241–243, 491 passivization, 309, 310, 311 patient noun, 100, 114, 118, 147, 153–155, 172, 194, 198, 203, 207–209, 213–219, 222–227, 230–242, 249, 253, 271, 303, 307, 310, 344, 386, 425, 536, 550, 556 perfective aspect, xx, 12–15, 101, 208, 265, 270, 272, 327, 332–334, 338, 340, 351, 352–359, 364, 365, 368–371, 377–380, 382–385, 387, 392, 550, 568 periodization, 5–6, 7 phonological change, 78, 81, 401, 405, 413 phonological derivation, 356, 359, 439–443, 446, 451, 453, 464, 468, 550 phonological reduction, 96, 139, 208, 329, 338, 366, 375, 401, 464–466, 485, 488, 532 phonological representation, 81–83, 86, 89, 93, 127, 139, 485 phonology, 5, 6, 9, 84, 133, 265, 324, 416, 442, 467, 475, 485, 560, 561, 562, 566, 569, 574 plural, xiv, xx, 104, 246, 327, 356, 358, 359, 386, 387, 423, 431–433, 440, 441, 446–449, 470, 471–487, 548–551, 566 possessive marker, 512, 525

581

postverbal position, 162 potential form, 75, 95–99, 110, 128–132, 134, 136–138 potential resultative, 128 pragmatic inference, 3, 544, 545, 547, 551 prefix, 146, 361, 434 preposition phrase, xx, 12, 17, 24, 40, 42–44, 57–59, 122, 147–150, 166, 167, 170–173, 179, 223, 231, 252, 264, 274, 278–279, 315, 344, 514 preverbal position, 24, 27, 32, 36–38, 55, 57–59, 88, 91–96, 100, 143, 144, 147–159, 164–174, 175, 176, 180–190, 192, 193–202, 203, 209–211, 214–219, 223, 226–228, 239–243, 267, 275, 278–280, 296–299, 300, 307, 308, 309, 312, 314, 348, 361, 367, 409, 420–430, 437, 480, 504, 511, 514, 519, 529, 530, 534, 544–549, 554, 555, 556 principle of action–resultative ordering, 163–165, 167–173, 234, 245, 260, 264, 265, 275, 279, 282, 361, 404–407, 552 principle of definiteness assignment, 410 progressive aspect, xx, 194, 344, 358–361, 366 pronoun, xx, xxi, 15, 16, 18, 25, 34–38, 43–45, 69, 76, 144–148, 178, 197, 212, 219–221, 277, 308–311, 314, 374–377, 412, 430, 436, 442, 448, 471–478, 499–502, 526, 529 proximal demonstrative, xiv, 173, 413–416, 418–421, 442, 444–447, 449, 454–469 quantifier, 27, 32, 33, 46, 225, 232, 252, 289–291, 382, 473, 475, 498 reanalysis, 62, 78, 87–89, 91, 92, 94, 96, 100–103, 140, 186, 203, 236, 239, 291–294, 327, 330–332, 338–341, 342, 344, 348, 349, 356, 373, 401, 425, 563, 566, 567 reduplication, 4, 9, 79, 85, 151, 158–160, 246, 265–272, 295, 304–306, 329–331, 355, 359, 387, 389, 392–395, 404–410, 486, 491, 495, 517, 533, 548, 552 referential, 214, 252–254 relation grammar, 4 relative clause, xx, 3, 35–37, 145, 172, 205, 206, 358, 377, 397, 412, 489, 492, 496–499, 501, 506–510, 511, 512, 513, 514, 520, 521, 522, 523, 524, 525, 526, 527, 528, 529, 532, 534, 539, 541, 561 relativizer, 39, 50–52, 55, 206, 488, 495, 500, 501, 506, 512, 517–529

https://doi.org/10.1017/9781108921831.025 Published online by Cambridge University Press

582

Index

resultative construction, xx, 9, 62, 66–68, 69, 76, 78–81, 84–96, 98, 100, 101, 103–105, 106, 107, 108, 109, 110, 111, 112, 114–119, 120, 121, 122–129, 131–141, 144, 164, 167, 173, 180, 193, 230, 234, 235–238, 245, 251–253, 255, 257, 259, 266, 267–270, 271, 272, 287–292, 296, 298–305, 307–310, 314, 330, 331, 332–334, 339, 341, 344–346, 348, 350, 351, 361, 367, 368, 372, 381, 383, 387, 392, 395, 401, 424–427, 482, 530, 531, 541–543, 546, 547, 548, 549, 552, 555, 564, 566, 568, 570, 574 semantic primitive, 411 semantic suitability, 212, 238, 281, 350, 353, 419, 493–495, 544 sentence-final position, 11, 14, 16, 17, 30, 42–44, 55–57, 58, 122, 143, 144, 146, 149, 150, 165, 168, 180, 196–199, 209, 260, 265, 274–280, 285, 296, 312, 327, 514, 518, 519, 529–531, 534 sentential particle, 30 separable resultative, 68–73, 76, 86–92, 93–98, 101, 107, 116, 120, 121, 122–125, 128, 129, 133–141, 338–341 serial verb construction, 62, 66–67, 171, 193–196, 205, 208, 237, 241, 244, 267–270, 282–284, 302–304, 331, 348, 388, 389, 544, 548, 556 short passive, 187–190 singular, xiv, xx, xxi, 423, 428, 432, 440–443, 449, 470–473, 477–481, 483 sonority degree, 453 sonority sequencing principle, 451 SOV language, 12, 14, 18, 34, 42, 204, 205–208, 239, 242, 273, 293, 376 structural particle, 267, 358, 393, 488, 489, 492–495, 498–507, 510, 562, 565, 570 subject–predicate construction, 18, 143, 145 suffix, 13, 80, 86, 118, 146, 220, 293–295, 326–328, 331–333, 342–362, 367, 369, 379, 389, 433, 483, 484, 534, 568 superlative comparison, 277 SVO language, 10–12, 14, 27, 42, 92, 168, 197, 204, 215, 273–275, 367, 513, 523, 537 syllable, 78, 79, 80, 81, 83, 96–98, 103, 104, 319, 357, 374, 401, 418, 444, 447, 450–454, 457, 484, 569

syntactic position, 27, 121, 142–144, 151–158, 160–164, 181, 185, 203, 218, 228, 234, 252, 265–267, 269, 271, 299, 304–307, 330, 350, 409, 410, 427, 480, 529, 550 Tibetan, 239–242, 366, 409, 512, 535, 561, 566, 568, 576 tone value, 89, 265, 338, 357, 471, 477 topic, xx, 3, 9, 15–17, 20, 21, 27–29, 43–46, 100, 131, 142–148, 149–152, 153–157, 223, 226, 248, 310–313, 409, 420, 423, 426, 428, 470, 559, 565, 566, 567 topicalization, 15, 16, 28, 40, 44–46, 61, 85–89, 114, 144–147, 148, 156, 164, 173, 223, 229, 247–250, 310–314, 555 transitivity, 29, 48, 56, 111–113, 213–215, 238–241, 249, 251, 279, 342 twin-syllable words, 452 typological change, 24–26, 27, 28, 29, 32, 38, 46, 54, 55, 56, 57, 60, 62, 66, 151, 203, 489, 511, 520, 529, 534, 546 typology, 434, 558, 560, 561, 562, 563, 565, 566, 568, 570, 571 universal correlation, 3, 273–275, 511–513, 542 universal grammar, 565 universal reference, 32, 157–160, 164–167, 267, 407, 408, 410 verb co-ordination, 62–64, 66–68, 78–80, 89–93, 99, 116, 117, 124, 133, 134, 139, 251, 269, 331–333, 339–341, 348, 351 verb reduplication, 266, 267–272, 355, 407 verb-copying construction, 9, 113, 174, 203, 246–260, 303, 387, 426, 550 verb–object construction, 307 weight, 2, 290, 303, 397, 517 wh- in situ, 24–26, 27, 28, 29, 30, 32, 33, 34, 35, 46–49, 52, 54, 55, 56, 57, 60, 61, 560, 561, 570 wh- movement, 24–26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 40, 43–52, 54, 55, 56, 57, 60–61, 558, 560, 561, 566, 570 wh- word, 16, 17, 24, 28, 29–32, 39–54, 55–61, 145, 148, 151, 157–159, 164 word formation, 251, 350–352, 481, 483

https://doi.org/10.1017/9781108921831.025 Published online by Cambridge University Press

Index word order, xx, 3, 10–11, 12, 14, 17, 18, 21–26, 28, 34, 42, 47, 51, 55, 60, 76, 107, 140, 144, 180, 193, 194, 203–208, 215, 239–243, 259, 334, 338–341, 376, 377, 512, 539, 541–543, 547, 554, 561, 562, 566, 567, 569, 571 word order change, 25, 51, 60, 144, 193, 204–209, 239, 334, 340, 512, 541, 566, 567, 569

Wu dialect, 91, 194, 237, 305, 343, 355, 415–417, 428, 442, 445, 446, 500, 501, 567 Xiang dialect, xiv, 370, 415–417, 443, 446–447 Yue dialect, xiv, 415–417, 441–444, 459

https://doi.org/10.1017/9781108921831.025 Published online by Cambridge University Press

583

https://doi.org/10.1017/9781108921831.025 Published online by Cambridge University Press