UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Zhong-shan phonology : a synchronic and diachronic analysis of a Yue (Cantonese) dialect 1980

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
UBC_1980_A8 C45.pdf
UBC_1980_A8 C45.pdf [ 29.85MB ]
UBC_1980_A8 C45.pdf
Metadata
JSON: 1.0100261.json
JSON-LD: 1.0100261+ld.json
RDF/XML (Pretty): 1.0100261.xml
RDF/JSON: 1.0100261+rdf.json
Turtle: 1.0100261+rdf-turtle.txt
N-Triples: 1.0100261+rdf-ntriples.txt
Citation
1.0100261.ris

Full Text

ZHONG-SHAN PHONOLOGY: A Synchronic and Diachronic Analysis of a Yue (Cantonese) Dialect by MARJORIE KIT MAN CHAN B.A., Univ e r s i t y of B r i t i s h Columbia, 1974 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES Department of L i n g u i s t i c s We accept t h i s thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September, 1980 © Marjorie K i t Man Chan, 1980 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of Brit ish Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the Head of my Department or by his representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of Brit ish Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Depa rtment ABSTRACT Zhong-shan i s a county i n Kwangtung Province i n southern China. What i s normally referred to as the "Zhong-shan d i a l e c t " i s the speech of S h i - q i , the administrative centre for the county. For the present thesis, data were c o l l e c t e d from native Zhong-shan speakers from Shi-qi and neighbouring v i l l a g e s where the speech can be equated with the S h i - q i , Zhong-shan d i a l e c t . The data e l i c i t e d consist of two main types: (1) c o l l o - q u i a l vocabulary, for which graphic representation (an the form of standard Chinese characters) do not e x i s t , and (2) a lexicon based on the reading of a standard word l i s t for Chinese d i a l e c t surveys (namely, the Fang-yan Diao-cha Zi-biao), which contains some 3,700 Chinese characters. The synchronic study, which used both sets of data, i s based on an amalgamation of Western struc- t u r a l i s t and Chinese ( t r a d i t i o n a l and modern) approaches. For the diachronic study, the d i a l e c t survey l i s t , arranged according to h i s t o r i c a l phonological categories, was indispensible. The diachronic study e s s e n t i a l l y mapped the pat- tern of correspondences of the d i a l e c t against the h i s t o r i c a l categories to which each word belonged. Against such a backdrop, i t i s possible to observe the development of a given d i a l e c t with respect not only to e a r l i e r s trata of the Chinese language, but also to other modern Chinese d i a l e c t s . Thus, i n Zhong-shan, some features may reveal certain mergers with reference to a p a r t i c u l a r stratum of the language, whereas other features may show survivals of yet older d i s t i n c t i o n s . References to previous studies on the Zhong-shan d i a l e c t are also made when differences between (.or among) data seem s i g n i f i c a n t . C r o s s - d i a l e c t a l l y , since the Can- tonese d i a l e c t i s the standard for the Yue d i a l e c t group to which Zhong-shan belongs, a comparison between Zhong-shan and Cantonese i s made throughout the study. Other southern Chinese d i a l e c t groups, such as Min and Hakka, are also c i t e d where relevant. The thesis i t s e l f i s divided into two main parts: the f i r s t part i s the synchronic study, and the second part the di a - chronic analysis. In order that the thesis may better serve future research endeavours, both the c o l l o q u i a l lexicon and the lexicon of character readings are included: the c o l l o q u i a l data appear . at the end of Part I, while the d i a l e c t word l i s t occurs at the end of Part I I . The reading of the characters i s recorded d i r e c t l y onto the format of the Fang-yan Diao-cha Zi-biao that the Chinese L i n g u i s t i c s Project at Princeton had prepared express- ly for fieldwork purposes. Immediately following the d i a l e c t survey material i s an index to the d i a l e c t material. The index i s likewise prepared by the Chinese L i n g u i s t i c s Project, with, the words arranged according to Pin-yin romanization. i v TABLE OF CONTENTS T i t l e Page i Abstract i i Table of Contents i v L i s t of Tables v i i L i s t of Charts v i i i L i s t of Maps i x Acknowledgement x Part I. Synchronic Study CHAPTER 0. INTRODUCTION 1 0.1. Terminology, T r a n s l i t e r a t i o n and Other Conventions 1 0.2. H i s t o r i c a l , Geographical and L i n g u i s t i c Setting 3 0.3. Data Base 6 0.3.1. The Informants—Biographical Sketch 6 0.3.2. Data-Collection Techniques 8 0.3.3. Additional Sources for Zhong- shan Data 11 Notes to Chapter 0 14 CHAPTER 1. PHONETIC DESCRIPTION AND PHONEMIC ANALYSIS. 17 1.1. I n i t i a l s 21 1.2. Finals 32 1.2.1. Nuclear Vowels 40 1.2.2. Medials 50 1.2.3. Endings 62 1.2.4. S y l l a b i c Nasals 63 1.3. Tones 63 1.3.1. Tonal System 64 1.3.2. Tone Sandhi 6 7 1.3.3. Tone Change 68 1.4. Combination of I n i t i a l s and Finals 77 1.4.1. Labia l D i s s i m i l a t i o n 77 V 1 . 4 . 2 . Syncope 78 1 . 4 . 3 . S e s q u i s y l l a b i c S t r u c t u r e s 79 1 . 4 . 4 . C o l l o q u i a l Versus L i t e r a r y Forms 95 Notes to Chapter 1 101 CHAPTER 2. SYLLABARY AND LEXICON 107 2 . 1 . S y l l a b a r y Arranged Accord ing to Modern Zhong-shan F i n a l s 110 2 . 2 . L e x i c o n of C o l l o q u i a l Terms Arranged Accord ing to Modern Zhong-shan F i n a l s 126 P a r t I I . D i a c h r o n i c Study CHAPTER 3 . ANALYSIS OF MODERN REFLEXES OF HISTORICAL CATEGORIES 161 3 . 1 . I n i t i a l s 164 3 . 1 . 1 . Reconst ructed Values of M idd le Chinese I n i t i a l s 164 3 . 1 . 2 . Zhong-shan Correspondences to Midd le Chinese I n i t i a l s 175 3 . 1 . 2 . 1 . B i l a b i a l s (LMC) 175 3 . 1 . 2 . 2 . L a b i o d e n t a l s CLMC) 180 3 . 1 . 2 . 3 . Denta ls (LMC) 183 3 . 1 . 2 . 4 . Denta l S i b i l a n t s (LMC) 191 3 . 1 . 2 . 5 . R e t r o f l e x e s (LMC) 197 3 . 1 . 2 . 6 . R e t r o f l e x S i b i l a n t s (EMC) and P a l a t a l s (EMC) 201 3 . 1 . 2 . 7 . V e l a r s (LMC) 216 3 . 1 . 2 . 8 . G u t t u r a l s (LMC) 232 3 . 2 . F i n a l s 241 3 . 2 . 2 . Modern Zhong-shan Correspondences to LMC F i n a l s 260 3 . 2 . 2 . 1 . Guo-she 261 3 . 2 . 2 . 2 . J i a - s h e 265 3 . 2 . 2 . 3 . Yu-she 266 3 . 2 . 2 . 4 . X i e - s h e 271 3 . 2 . 2 . 5 . Z h i - s h e 277 3 . 2 . 2 . 6 . X i a o - s h e 282 3 . 2 . 2 . 7 . L i u - s h e 283 v i 3.2.2.8. Xian-she 285 3.2.2.9. Shen-she 289 3.2.2.10. Shan-she 290 3.2.2.11. Zhen-she 297 3.2.2.12. Dang-she 304 3.2.2.13. Jiang-she 307 3.2.2.14. Zeng-she 309 3.2.2.15. Geng-she 311 3.2.2.16. Tong-she 317 3.3. : Tones • 319 3.4. Concluding Remarks 333 Notes to Chapter 3 334 CHAPTER 4. LEXICON (AS ARRANGED IN THE FANG-YAN DIAO-CHA ZI-BIAO) 337 Rhyme Groups: 1. Guo-she 1 339 2. Jia-she 7 345 3. Yu-she 14 352 4. Xie-she 28 366 5. Zhi-she 47 385 6. Xiao-she 67 405 7. Liu-she 81 419 8. Xian-she 92 430 9. Shen-she 110 448 10. Shan-she 115 453 11. Zhen-she 148 486 12. Dang-she 166 504 13. Jiang-she 180 518 14. Zeng-she 184 522 15. Geng-she 191 529 16. Tong-she 212 212 Index 565 Bibliography 655 v i i LIST OF TABLES Table 1. The I n i t i a l s i n Zhong-shan 21 Table 2. (a) Analysis of the Fi n a l s — C h a n 35 (b) Analysis of the Finals—Chao 36 (c) Analysis of the F i n a l s — E g e r o d 37 (d) Transcription of Finals i n (Macao) Zhong-shan—Ball 38 (e) Phonetic Transcription of the Fi n a l s i n Cantonese 39 Table 3. (a) Tonal System of Zhong-shan 6 5 (b) A Comparison of Cantonese and Zhong-shan Tones 66 Table 4. Coll o q u i a l Versus L i t e r a r y Forms: (a) Tone /13/ Aspirated I n i t i a l and /22/ Unaspirated I n i t i a l 98 (b) /a:ng/ and /ang/ Finals 98 (c) /ia:ng/ and /ing/ Finals 99-100 Table 5. Words with Alternate H i s t o r i c a l Tonal Categories: (a) Words L i s t e d as Ping-sheng 327 '(b) Words L i s t e d as Shang-sheng 327 (c) Words L i s t e d as Qu-sheng 328-329 (d) Words L i s t e d as Ru-sheng 329 Notes to Table 5 329 v i l i LIST OF CHARTS Chart 1. Syl l a b l e Structure of Chinese 17 Chart 2. Syl l a b l e Structure of Zhong-shan 19 Chart 3. (a) Analysis of the Vowels—-Chan 40 (b) Analysis of the Vowels—Chao 41 (c) Analysis of the Vowels—Egerod 41 Chart 4. The 36 I n i t i a l s of Late Middle Chinese 165-166 Chart 5. Early Middle Chinese I n i t i a l s 169 Chart 6. Middle Chinese I n i t i a l s i n the Fang-yan Diao-cha Zi-biao 172 Chart 7. Zhong-shan Correspondences to the Middle Chinese I n i t i a l s i n the Fang-yan Diao-cha Zi-biao 173-174 Chart 8. D i s t r i b u t i o n of the LMC Glides and Medials 244 Chart 9. The 16 Rhyme Groups 248 Chart 10. Late Middle Chinese Finals 249-252 Chart 11. Rhymes and Grades Within Each Rhyme Group 256-258 Chart 12. Zhong-shan Correspondences to the LMC Fi n a l s 259 Chart 13. (a) Zhong-shan Correspondences to the H i s t o r i c a l Tones 320 (b) Cantonese Correspondences to the H i s t o r i c a l Tones 320 Chart 14. Ru-sheng Correspondences i n Cantonese and Zhong-shan 323 Chart 15. Tonal Correspondences i n Zhong-shan (and Cantonese) 323 Map 1. Kwangtung Province Map 2. Zhong-shan County X ACKNOWLEDGEMENT The data for the present study of the Zhong-shan d i a l e c t was co l l e c t e d at various i n t e r v a l s , beginning with a short project i n the spring of 1977; the bulk of the data for the thesis, however, was gathered i n the spring of 1978. For these e l i c i t a t i o n s , the p r i n c i p a l informants were my parents, Chen Gui-hong P ^ L - ^ j ^ l ^ anc^ Yang Zhi-fang ^ / c V ^ f . They responded t i r e l e s s l y to my seemingly endless questions. To them, I owe a depth of gratitude that words cannot f u l l y express. Their willingness, patience and support have made t h i s experience a very personally rewarding one. I am grateful to the L i n g u i s t i c s Department at the Univer- s i t y of B r i t i s h Columbia for the academic t r a i n i n g that I have received, and for t h e i r supportive role i n my endeavours. Con- cerning the thesis-writing i n p a r t i c u l a r , I am indebted to my Committee members, Dr. M. Dale Kinkade and Dr. Sarah B e l l , for th e i r reading of an e a r l i e r draft and he l p f u l suggestions. I am also appreciative of Dr. Kinkade's assistance i n the f i n e r phonetic transcriptions noted i n the study. To my advisor, Professor E.G. Pulleyblank, I owe deep gratitude for his many constructive c r i t i c i s m s and suggestions throughout the thesis-writing, and for his c l a r i f i c a t i o n of v a r i - ous points that were unclear to me. His helpfulness, together with his quiet support and immense patience during that entire period, cannot be over-emphasized. I would also l i k e to thank Professor Jerry Norman, at the University of Washington, who kindly read through an e a r l i e r d r a f t and made a number of help f u l comments, p a r t i c u l a r l y with regard x i to Min d i a l e c t material. Also h e l p f u l i n my thesis endeavours has been Karl K. Lo, a native Zhong-shan speaker, who supplied a few of the forms that are recorded here. I am also grateful to my husband, Gary R. White, for his moral support, encouragement, and general a s s i s t - ance that f a c i l i t a t e d the writing and f i n a l type-up of the thesis. I wish to acknowledge with h e a r t f e l t thanks support from the University of B r i t i s h Columbia (U.B.C. Graduate Fellowship 1976-1978). I am also g r a t i f i e d that the University of B r i t i s h Columbia allowed me to use the Summer Research Grant (.1977) to attend the L i n g u i s t i c Society of America's 1977 Summer L i n g u i s t i c I n s t i t u t e . That year, the I n s t i t u t e was held at the University of Hawaii where special emphasis was on Asian and P a c i f i c languages. The study programme was also supported by the L i n g u i s t i c Society of America (LSA Fellowship 1977), to which I wish also to express my deep gratitude. Needless to say, although I owe much to a l l those who have guided me i n the thesis, I am s o l e l y responsible for whatever errors that remain. X I 1 Map 1. Kwangtuna Province. (A modification, of. Yue, 1979:2.) x i i i - 1 - PART I. SYNCHRONIC STUDY CHAPTER 0. INTRODUCTION Studies of. the Yue d i a l e c t s of Chinese have generally concentrated on standard Cantonese and Taishanese, while other di a l e c t s have received peripheral attention. To counterbalance th i s general trend, the Yue d i a l e c t which i s investigated i n the present thesis i s the Zhong-shan"*" <Xf d i a l e c t . The primary goal of t h i s study, however, i s to supplement e x i s t i n g works on the Zhong-shan d i a l e c t with a larger corpus of f i e l d data and a more detailed analysis of i t s phonology on a synchronic and diachronic l e v e l . The hope i s that both the raw data and the analysis w i l l contribute to future l i n g u i s t i c research. 0.1. Terminology, T r a n s l i t e r a t i o n and Other Conventions The term "Cantonese" has often been used to i d e n t i f y both the standard Cantonese d i a l e c t and the d i a l e c t group, thereby creating some confusion for those less f a m i l i a r with the l i n g u i s t i c s i t u a t i o n i n Kwangtung Province i n southern China. To eliminate th i s source of ambiguity, the d i s t i n c t i o n between the terms "Yue" and "Cantonese" drawn by Oi-kan Yue Hashimoto (1972a:1) i s adopted here: "Yue" i s used to refer to the d i a l e c t group, and "Cantonese" to the group norm. Yue i s the d i a l e c t group representative of Kwangtung Province, although found there are also other major Chinese d i a l e c t groups such as Hakka and Min. Cantonese, or Standard Yue, i s the d i a l e c t of the majority of people i n Canton and Hong Kong, located on the Pearl River Delta. "Standard Cantonese" w i l l be a term used only for further c l a r i f i c a t i o n - 2 - or emphasis. The new Pin-yin romanization system, which has been o f f i c i a l l y adopted i n the People's Republic of China to transcribe the national standard, pu--tong-hua ('common d i a l e c t ' , or what i s usually regarded as the Peking d i a l e c t of Mandarin), w i l l be used here for the t r a n s l i t e r a t i o n of Chinese terms, including personal and place names, with a few minor modifications. Hyphens w i l l be inserted between s y l l a b l e s within a word, and an occasional tone mark w i l l be used for disambiguation i n cases where the romanized form of several words, for example, would have been i d e n t i c a l except for tone. The four tones i n Mandarin are marked thus: '"' for l e v e l , " 1 r i s i n g , l v ' dipping ( f a l l i n g - r i s i n g ) , and , v ' f a l l i n g . Pin-yin i s used for Chinese personal and place names except for well-known geographical names, such as Canton and Hong Kong, for which the romanization established i n the China Postal Atlas w i l l be retained. Moreover, rather than attempt to over- standardize personal names to a single romanization system, the romanization that has already been established for the names of well-known Chinese l i n g u i s t s such as Yuen Ren Chao and Fang Kuei L i w i l l likewise be used here. The use of the terms "character" and "word" needs some c l a r i f i c a t i o n . The Chinese language distinguishes between what i s a "character" ( z i ^jl ) and what i s a "word" (c i ^s) ). A cha- racter i s simply the i n d i v i d u a l ideograph, which i s a monosyllable. Often a character constitutes a word; the character da , for instance, i s a word meaning "big". However, i t can also combine with the character xue ^ (which, as a monosyllabic word, means 'to study') to form the d i s y l l a b i c word da-xue ^ ^ / which -- 3 - means ' u n i v e r s i t y ' . In a grammatical a n a l y s i s , one should f u r t h e r in t roduce the l i n g u i s t i c term 'morpheme', which i s 2 l o o s e l y d e f i n e d as a "min imal meaningfu l u n i t " . U s u a l l y there i s a o n e - t o - o n e correspondence between a s y l l a b l e (or charac te r ) and a morpheme; t h a t i s , most morphemes i n Chinese are m o n o s y l l a - b i c , as e x e m p l i f i e d by da and xue above, which can now be analyzed as two m o n o s y l l a b i c morphemes c a r r y i n g the meanings ' b i g ' and ' t o s tudy ' r e s p e c t i v e l y . There a r e , n o n e t h e l e s s , a few r a r e cases of d i s y l l a b i c morphemes whose o r i g i n i s no longer known. The word h u - d i e tĝ JJ ' b u t t e r f l y ' i s a case i n p o i n t ; the f i r s t s y l l a b l e hu c o n t a i n s no meaning i n and of i t s e l f . Other t e r m i n o l o g i e s and convent ions w i l l be i n t r o d u c e d as they are met i n the f o l l o w i n g s e c t i o n s . 0 . 2 . H i s t o r i c a l , Geograph ica l and L i n g u i s t i c S e t t i n g . Dur ing the Han Dynasty (206 B . C . - 2 3 A . D . ) , what i s now the county (x ian ^j^. ) ^of. Zhong-shan was p a r t of Pan-yu county "I) % * I n T a n 9 t imes (618-907) i t became a p a r t of Dong- guan county jĵ  . I t was a t the beg inn ing of the Song Dynasty (960-1279) t h a t i t became e s t a b l i s h e d as a separate county c a l l e d X iang -shan county |j J, |tĵ  , now over e i g h t hundred years ago. In the f o u r t e e n t h year of the R e p u b l i c of China ( i . e . , 1925) , i n honour of Dr . Y a t - s e n Sun, the f a t h e r of the Chinese r e v o l u t i o n , the name of h i s b i r t h p l a c e was o f f i c i a l l y changed by the p r o v i n c i a l government of Kwangtung from X i a n g - s h a n to Zhong- shan VJ* based on the name which Dr . Sun adopted w h i l e seek ing p o l i t i c a l asylum i n Japan, "Zhong-shan" be ing the Chinese - 4 - pronunciation of "Naka-yama" ^ 4 Geographically , Zhong-shan i s one of the coastal counties of Kwangtung Province. I t i s situated west of the Pearl River delta and immediately north of Macao, thus partway between Canton and Macao (see Map 1). The county i s over 1,800 square kilometres i n area—70 kilometres long running north and south, and 35 k i l o - metres wide east and west. Within t h i s circumference, the county i s divided into nine administrative sections c a l l e d qu (3* . Shi- q i (.otherwise romanized as "Shekki") fa , the administrative centre and the largest town i n the county, i s located i n the f i r s t qu. The Zhong-shan d i a l e c t i s here i d e n t i f i e d and equated with Shi-qi speech, which i s considered the standard for the county. Besides S h i - q i , also entering into the present study i s the neigh- bouring v i l l a g e of Ku-chong jfjjj , situated i n the fourth qu immediately behind the boundary southeast of Shi-qi (see Map 2). A number of the more c o l l o q u i a l expressions c o l l e c t e d i n the present study r e f l e c t Ku-chong speech rather than that of the more educated townspeople of S h i - q i . In terms of i t s l i n g u i s t i c c l a s s i f i c a t i o n , Zhong-shan i s one of the Yue d i a l e c t s , consequently sharing many of the features found i n Cantonese. At the same time, the Shi-qi d i a l e c t i n China i s also influenced by the surrounding d i a l e c t s . Although small i n area, a l l three major d i a l e c t groups i n Kwangtung—-namely Hakka, Min and Yue—are spoken i n the Zhong-shan county. I t i s therefore not surprising to f i n d each d i a l e c t i n turn influenced by the others, i n a process which has been described by Egerod (1956:76) as "balkanization", using the term i n the l i n g u i s t i c sense of "the gradual merging of geographically close, etymologically far-removed - 5 - speech forms". Egerod (p.77) notes, for example, that the simpli- c i t y of Shi-qi tonal pattern, i n contrast to that i n Cantonese, renders the Shi-qi d i a l e c t closer to Min than to Yue with respect to the number of tones. Zhong-shan i s , i n fact, the only exception to the pattern of eight or more tones and the dichotomy of the Yin-ru tone which constitute two of the c h a r a c t e r i s t i c features of the Yue d i a l e c t s (Hashimoto, 1972a:44). Hence, i t would not be unreasonable to suggest balkanization as a possible factor i n the uniqueness of the Zhong-shan tonal system amongst the Yue d i a l e c t s . Although the Shi-qi d i a l e c t i s the standard for the county, i t i s predominant only within i t s own qu. In the remaining eight qu, Cantonese, Hakka, and various forms of Min constitute the 6 major d i a l e c t s . Hence, the Shi-qi d i a l e c t i s actually spoken i n a very li m i t e d area, i t s prominence f e l t only as a r e s u l t of i t s economic and p o l i t i c a l status. In t h i s p osition also, i t has stringent competition from Macao, which i s geographically and h i s - 7 t o r i c a l l y , though no longer p o l i t i c a l l y , a part of Zhong-shan county. The l i n g u i s t i c scene i n Macao has undergone d r a s t i c changes since J.D. B a l l ' s fieldwork before the turn of the century when the county was s t i l l c a l l e d "Xiang-shan" (or "Hong-shan", i n g B a l l ' s Cantonese t r a n s c r i p t i o n ) . According to B a l l (1897:550) the d i a l e c t spoken i n Macao was i d e n t i c a l to the Zhong-shan di a - l e c t , with exceptions a r i s i n g primarily from the desire of the educated class i n Shi-qi to emulate the more prestigious Cantonese forms. By mid-twentieth century, Egerod (p.3) observes that Stan- dard Cantonese has become the main d i a l e c t i n Macao. Neverthe- les s , the Cantonese spoken there retains a few traces of the Zhong- - 6 - shan d i a l e c t , such as the fusion of the p l a i n and l a b i a l i z e d velar stops as a r e s u l t of which Cantonese /kwa/ and /ka/, for instance, are both pronounced /ka/ i n Macao, with l a b i a l i z a t i o n l o s t . Meanwhile, because of low y i e l d i n a g r i c u l t u r a l production i n the past i n Zhong-shan"^, i t has been a t r a d i t i o n for the l o c a l people to seek a means of l i v e l i h o o d away from home. Among the places which attracted many immigrants was the Hawaiian Islands. As Chao (1948:49) 1 1 commented, the Chinese population i n Hawaii was predominantly speakers of the Zhong-shan d i a l e c t . Since the publication of Chao 1s a r t i c l e i n the middle of the century, i t i s possible that the i n f l u x of Chinese immigrants from Hong Kong, Taiwan, and other areas may have reduced the proportion of Zhong- shan speakers i n Hawaii. 0.3. Data Base. The study of Zhong-shan phonology i s based on data c o l l e c t - ed by the writer at various i n t e r v a l s from 1977 to 1980, the bulk, of which was gathered during the spring of 1978. Published works on the d i a l e c t have also been consulted, as w i l l be further e l a - borated subsequently. 0..3.1. The Informants--Biographical Sketch. Data were e l i c i t e d from two main informants: Chen Gui-hong fjL $h a n c ^ Y a n 9 Zhi-fang $ ' t n e writer's parents. Chen was born i n 1923 i n the v i l l a g e (cun ̂  ) of Ku-chong ^ fj ^ , where he acquired the early part of his education. He com- pleted grammar school i n S h i - q i , which i s west of the v i l l a g e , about half an hour's walk away. His mother, who i s l i v i n g with - 7 '- the family, was born i n the v i l l a g e (xiang ) of Yuan-feng walking distance from the town (see Map 2). Chen attended busi- ness college i n Hong Kong, where he also studied Mandarin, English and Japanese. He t r a v e l l e d between Ku-chong and Hong Kong several times before immigrating to Vancouver, B r i t i s h Columbia i n 1952 to j o i n his parents. Chen's father, who was born i n Ku-chong, spoke Zhong-shan and a few words of English, while Chen's mother only knows the Zhong-shan d i a l e c t . Contact with fellow v i l l a g e r s from Ku-chong i s maintained through friends, r e l a t i v e s and annual gatherings organized by residents of the "Ku-chong Home", a house open to Ku-chong v i l l a g e r s who wish to drop i n from time to time, or who need a place to stay. Yang was born i n 19 27 i n S h i - q i , where she received four or f i v e years of education before i t was disrupted by the invasion of China by Japan. She continued to l i v e i n S h i - q i , working there as a nurse during the war years. Yang l i v e d i n Ku-chong for several years before spending two or three years i n Macao and Hong Kong, a r r i v i n g i n Vancouver, B r i t i s h Columbia i n 1958. She has had exposure to Mandarin and Japanese. Her knowledge of English has been acquired informally, primarily through contact with cus- tomers i n a small, family-operated grocery store. to a much more limited extent since she l e f t Ku-chong at the age of four or f i v e and was then exposed to Cantonese i n Macao and Hong Kong, and i n Canada subsequently. , north of S h i - q i , also approximately half an hour's The writer herself also served as an informant, although - 8 - p.. 3 . 2 . D a t a - C o l l e c t i o n Techniques A p r e l i m i n a r y se t of da ta was c o l l e c t e d i n the s p r i n g of 1977 i n the attempt to e l i c i t c o l l o q u i a l versus l i t e r a r y read ings of the same c h a r a c t e r s based on those found i n Cantonese ( i . e . , i n Hashimoto, 1 9 7 2 a : 1 6 9 - 1 7 0 ) . I t was recogn ized by the w r i t e r t h a t such an approach has i t s l i m i t a t i o n s , the pr imary one b e i n g t h a t a d i s t i n c t i o n between l i t e r a r y and c o l l o q u i a l r e a d i n g of some words found i n Zhong-shan but not p resent i n Cantonese would be missed . Given l i m i t e d t i m e , however, i t was n e v e r t h e l e s s a c o n - v e n i e n t means to o b t a i n a s i z e a b l e l i s t w i thout r e s o r t i n g t o an e x t e n s i v e l e x i c o n . The method used was as f o l l o w s : a character ; , was f i r s t read by Chen, f o l l o w e d by a d i s c u s s i o n of whether or not there e x i s t s a c o l l o q u i a l c o u n t e r p a r t . Some e f f o r t was a l s o made subsequent ly to produce l i t e r a r y versus c o l l o q u i a l forms, a l though not by any s y s t e m a t i c or c o n s i s t e n t approach. The major task of e l i c i t i n g Zhong-shan data was, however, based on the Fang-yan D i a o - c h a Z i - b i a o ^ -̂ /l) ^ ^- jfc^ 'A Table of Charac te rs f o r D i a l e c t Surveys ' ( h e r e a f t e r r e f e r r e d to as the " d i a l e c t survey l i s t " f o r s h o r t ) . That source p r o v i d e s a s tandard l i s t of j u s t over 3,700 c h a r a c t e r s arranged a c c o r d i n g to t r a d i t i o n a l Chinese p h o n o l o g i c a l c a t e g o r i e s , and i s used f o r com- p a r i s o n s among the Chinese d i a l e c t s and f o r s t u d i e s of the h i s t o - r i c a l phonology of a p a r t i c u l a r d i a l e c t . A l though d i f f e r e n t e d i t i o n s of t h i s survey l i s t e x i s t , they are e s s e n t i a l l y the same, w i t h but minor v a r i a t i o n s i n the cho ice of c h a r a c t e r s and format . The e d i t i o n used i n the p resent study i s p u b l i s h e d by P r i n c e t o n U n i v e r s i t y (1970) , and i s p a r t of the Chinese L i n g u i s t i c P r o j e c t a t P r i n c e t o n . Th is e d i t i o n , e s p e c i a l l y arranged f o r f i e l d w o r k - 9 - purposes, i s based on the 1955 character l i s t of the same name compiled by the Research Inst i t u t e of L i n g u i s t i c s , Chinese Academy of Sciences i n Peking. The survey l i s t provided data for the study of the phonological system of Zhong-shan, as well as y i e l d i n g the necessary material for a diachronic analysis of the d i a l e c t . In preparing for the d i a l e c t survey on Zhong-shan, each character on the l i s t was assigned a number consisting of two parts separated by a dash with the part preceding the dash in d i c a t i n g reference to the page number i n the d i a l e c t survey l i s t , and the part following the dash that of the character's p o s i t i o n 12 h i n the column on a given page. The word duo ^ 'many', for example, i s assigned the number "1-1" to specify page one, and the f i r s t character on the page. S i m i l a r l y , tuo ftfcj 'to drag along' i s assigned the number "1-2" since i t i s the second charac- ter on page one i n the survey text. (See Chapter 4.) After each character had been given a number, the order of the characters i n the survey l i s t was randomized. What the informant actually saw. i s the character, the number assigned to i t , and, where provided by the compilers, such information as one- word d e f i n i t i o n s , part of speech, environments i n which the charac- ter may occur, and alternate s p e l l i n g , for the purpose of a s s i s t - ing the informant i n r e c a l l i n g the character and/or making the correct choice for a character with multiple pronunciations. A p a r a l l e l case i n English would be to i d e n t i f y "export" as a verb or noun i n order to e l i c i t the form with the stress pattern sought. The d i a l e c t survey l i s t was read by Chen and recorded on a r e e l - t o - r e e l tape recorder. The writer transcribed the data phonemically during the e l i c i t a t i o n and used the tape for double- - 10 - checking afterwards. Questions concerning some of the f i n e r phonetic d i s t i n c t i o n s were brought to the attention of M.D. Kin- kade i n the L i n g u i s t i c s Department at the University of B r i t i s h Columbia. The data c o l l e c t e d was then compared with the Cantonese 13 forms given i n Hashimoto (1972a) . A l i s t of characters was drawn for double-checking pronunciation. Each of these characters was accompanied by additional information to f a c i l i t a t e r e c a l l , or to avoid confusion with other characters which graphically appear quite s i m i l a r . Such information usually consists of d e f i n i t i o n s (in English or Chinese) and the combination of these characters with others to form p o l y s y l l a b i c words. Variant forms co l l e c t e d often r e f l e c t d i a l e c t a l influence from Cantonese, esp e c i a l l y i n those cases where the informant was uncertain of the pronunciation. It i s l i k e l y that some of the more l i t e r a r y words have received Cantonese pronunciation as a r e s u l t of contact with Cantonese speakers i n Hong Kong and Vancouver. This may account for some of the differences found between the present set of data and that obtained by Chao i n Sh i - q i . For the second set of data, Yang served as primary i n f o r - mant, with Chen joi n i n g i n on some of the occasions. The second task was stimulated by the observation of many gaps i n Chao's repertoire of Zhong-shan s y l l a b l e s , including d i s t i n c t i o n of tone. The aim was to f i n d c o l l o q u i a l Zhong-shan words to f i l l i n as many of these empty spaces as possible. The writer soon found that to simply ask whether such-and-such a s y l l a b l e exists i n Zhong-shan produced few responses. The next approach was for the writer her- s e l f to systematically go through each potential s y l l a b l e , - 11 - p a r t i c u l a r l y those l e f t blank i n Qiao's study, i n order to f i n d a word which would l a t e r be confirmed by Yang as a word spoken i n Shi - q i , i n both Shi-qi and Hong Kong, or s t r i c t l y as a Hong KOng colloquialism. The majority of the words suggested by the writer were i d e n t i f i e d by Yang as Shi-qi expressions. Yang was frequently able to elaborate on the meaning of a word suggested by the writer, and sometimes produced additional meanings or usages for the word or s y l l a b l e . These were necessarily informal sessions, conducted whenever the opportunity arose. Chen's mother, who i s now i n her seventies, understands Cantonese, but speaks only Zhong-shan. A few d i s t i n c t i v e expres- sions which she uses are also recorded for t h i s study. 0.3.3. Additional Sources for Zhong-shan Data As far as the writer i s aware, to date only three works have been published which contain f i e l d material on the Zhong-shan d i a l e c t . Of these, "Zhong-shan fang-yan" k % ~$ ('Zhong-shan dialect') (1948) by Y.R. Chao, and portions of The Lungtu Dialect (1956) by S. Egerod contain data on the Zhong-shan d i a l e c t spoken i n S h i - q i . A t h i r d source i s an a r t i c l e by J.D. B a l l i n 1897 e n t i t l e d "The Hong Shan or Macao d i a l e c t " , which describes the Zhong-shan d i a l e c t as spoken i n Macao, allegedly i d e n t i c a l to Shi- q i speech. Shi-qi forms which d i f f e r from those found i n Macao are recorded by B a l l i n his footnotes. A comparison with Chao's a r t i c l e reveals more differences between B a l l ' s Macao data and Chao's Shi-qi data than were reported by B a l l . At le a s t some of these differences may have resulted from sound changes during that i n t e r v a l between B a l l ' s c o l l e c t i o n - 12 - of data on the. Zhong-shan d i a l e c t and Chao's fieldwork i n Sh i - q i , a. period of at lea s t t h i r t y or forty years. B a l l ' s a r t i c l e i s primarily of h i s t o r i c a l s i g n i f i c a n c e ; i t s usefulness for compara- t i v e purposes i s rather l i m i t e d . Tone indicators, for instance, are omitted by B a l l i n his tran s c r i p t i o n s . Furthermore, lacking an international alphabet by which to transcribe with greater precision and conciseness the phonetic sounds of the d i a l e c t at that time, B a l l had to struggle with the inadequacies of the English, language and the limi t a t i o n s of the English alphabet. Consequently, he resorted at times to circumlocutory phrases to describe a p a r t i c u l a r sound. This i s es p e c i a l l y true when he attempted to explain the manner i n which some of the vowel clus t e r s 14 and diphthongs were a r t i c u l a t e d . B a l l ' s choice of format i n his syllabary Ci.e., repertoire of the s y l l a b l e s i n the dialect), also presents a weakness: by using Cantonese s y l l a b l e s to show Zhong-shan counterparts, the structure of Zhong-shan phonology i s obscured. Sounds which, contrasted h i s t o r i c a l l y and were preseryed i n Zhong-shan are not readi l y discerned i n B a l l ' s syllabary i f they merged i n Cantonese. In such, cases, only one of the sounds i n Zhong-shan •iwast presented i n the syllabary and the other merely entered i n the footnote as .exceptions:,-, Egerod's contribution rests primarily cm his f a i r l y detailed phonetic description of the Shi-qi d i a l e c t , as well as the provision of background information on the d i a l e c t and the region. The most valuable and most frequently c i t e d source on the Zhong-shan d i a l e c t i s Chao's a r t i c l e on the phonology of the Shi- q i d i a l e c t . Chao's material i s based mostly on data e l i c i t e d from one informant i n Shi-qi i n 1929.. Supplemental material was - 13 - gathered on another informant i n Hawaii ten years l a t e r . Although te r s e l y written, Chao's a r t i c l e i s a comprehensive work containing a description of the phonological system of Zhong-shan, a s y l l a - bary, a l i s t of some exceptions to the sound changes i n the modern d i a l e c t on the basis of h i s t o r i c a l phonological c l a s s i f i c a t i o n s , and a sample dialogue. Besides the aforementioned works, there i s also a Sino- Portuguese glossary compiled by Guang-ren Yin and Ru-lin Zhang c i r c a mid-eighteenth, century i n Macao which i s of h i s t o r i c a l i n t e r e s t . The glossary, consisting of 395 l e x i c a l items, i s part of the text Ao-meri J j - l u e f*\'$£j (-'Record of Macao') of which, several editions ex i s t today. Studies which have been con- ducted on the glossary include CR. Bawden (1954)., R.W. Thompson (.1959), W. Boltz (.1977), and M. Chan (forthcoming). - 14 - Notes to Chapter 0. .1. Also transcribed as "Chungshan" i n Wade-Giles romanization, and as "Hbng Shan" according to Cantonese pronunciation. 2. For a more detailed discussion of Mandarin morphological structure, see, for example, Kratochvil (.1968:55-88). 3. The h i s t o r i c a l information i s derived from various sources: Hsu, 1970:540; Zhao, 1955:135-136; Xie, 1933:100; and Scenery Publishers' "Zhong-shan qiao-xiang feng-guang" r}? ^ | ^ ('Scenic spots of Chungshan'—English t i t l e provided) (no date—c.1975?). 4. Much of the geographical and l i n g u i s t i c information presented i n t h i s section i s thanks to the description of the county by Egerod (1956:3). (Note that Egerod refers to Zhong-shan as a " d i s t r i c t " . ) 5. Unless noted otherwise, references to Egerod are to his book, The Lungtu Dialect (1956). For Egerod (1956), only page reference w i l l be given hereafter. 6. Hashimoto (1972a:8) notes that although Zhong-shan i s predo- minantly Yue-speaking, a quarter of i t s inhabitants speak a Hakka d i a l e c t while a number speak a Southern Min d i a l e c t . 7. For example, i n Zhao (1955:136) i t i s stated that Macao, while belonging to Xiang-shan county, was leased to Portugal i n the Sino-Portuguese treaty of 1888 i n the thirteenth year of Emperor Guang-xu )\j \M • (See also Bu, 1977:1; Hua-qiao Zhi, 1964:1). 8. A l l subsequent references to B a l l w i l l be to his 1897 a r t i c l e ; therefore, only page number w i l l be s p e c i f i e d hereafter. 9. Karl Lo, a native Zhong-shan speaker, received his elementary - 15 - education i n Macao during the 1940's, at which time the dia- l e c t of i n s t r u c t i o n i n his classes was Zhong-shan, not stan- dard Cantonese (Lo., personal communication) . The emergence of Cantonese as the main d i a l e c t must have been more or less contemporaneous with Egerod's fieldwork. 10. Hsu (1970:540) attributes poor a g r i c u l t u r a l output to i n f e r - t i l e s o i l , whereas the p i c t o r i a l magazine on "Zhong-shan qiao- xiang feng-guang" (n.d.:17) claims that "(t)here are several m i l l i o n mu of f e r t i l e a l l u v i a l sandy land i n Ch.ungsh.an county ... (b)ut i t s annual grain output reached only 200 to 250 k i l o s i n the past because of i t s backward agriculture and shabby water conservancy". (.'Mu1 i s a Chinese land-measure of area.) 11. Unless otherwise indicated, a l l reference to Chao pertains to his publication on Zhong-shan phonology (1948); therefore, only page number w i l l be provided i n subsequent references. 12. After the survey was conducted, a few words were added to the survey l i s t using l e t t e r s or additional numbers. This w i l l be discussed i n greater d e t a i l l a t e r . 13. Unless s p e c i f i e d otherwise, a l l subsequent c i t i n g s of Hashi- moto are of Hashimoto (1972a). For t h i s work on Cantonese which i s c i t e d throughout the present study for comparative purposes, only page number w i l l be given subsequently. 14. An excerpt from the description of the sound / o i / , as i n the word gai / k o i / 'ought', i s c i t e d here as an example: "An approximation of i t may be got by pronouncing the two Canton-:. ese sounds ko ( & ) and y i (_3» ) rapidly together. Having now got th i s sound, then s l i g h t l y open the l i p s further and - 16 - pronounce i t a l i t t l e more open, at the same time taking great care not to get too near to the open koi sound of the Canton- ese c£jĵ  ) • In t h i s way and by l i s t e n i n g to a native from Macao pronouncing i t , i t i s possible to a r r i v e at the sound of the word "f$ x , koi i n Macao." (B a l l , p.509). (Underlining of the transcriptions mine.) - 17 - CHAPTER 1. PHONETIC DESCRIPTION AND PHONEMIC ANALYSIS The phonology of the Zhong-shan d i a l e c t w i l l be described according to an amalgamation of American s t r u c t u r a l i s t and Chinese ( t r a d i t i o n a l and modern) approaches. Since i t i s the l a t t e r approach which i s unfamiliar to most Western l i n g u i s t s , some of the terminologies and conventions based on Chinese a n a l y t i c a l categories w i l l be introduced below. F i r s t of a l l , the s y l l a b l e i s analyzed by Chinese l i n g u i s t s into two major parts: the i n i t i a l and the f i n a l (cf. Chart 1). The i n i t i a l i s simply the beginning consonant whose absence con- s t i t u t e s the "zero i n i t i a l " . The f i n a l consists of an optional medial, which i s a glide or v o c a l i c segment and i s ignored i n Chinese rhyming practices. The rhyming part - ( i . e . , the rhyme)•con- s i s t s of an obligatory s y l l a b i c segment, the nucleus, followed by an optional ending, which i s a glide, or a nasal or stop conson- ant. Also considered part of the f i n a l by some Chinese l i n g u i s t s i s the tone, which i s usually described i n terms of p i t c h contour and duration. Other l i n g u i s t s simply regard tone as a supraseg- mental feature and treat i t separately. Chart 1. S y l l a b l e Structure of Chinese. S Y L L A B L E FINAL RHYME INITIAL MEDIAL NUCLEUS ENDING - 18 - This analysis of the Chinese s y l l a b l e i n terms of an i n i t i a l - f i n a l dichotomy i s an inter p r e t a t i o n of a t r a d i t i o n which has i t s source i n a method of " s p e l l i n g " used since late Han, or second century A.D., c a l l e d fan-qie fg^ ̂ jj , l i t e r a l l y , 'turning and c u t t i n g 1 . I t i s a means of obtaining the pronunciation of a character by using two other characters, the f i r s t of which bears the same i n i t i a l ( e s sentially the non-rhyming portion), and the second the same rhyme, or f i n a l . The character dong 'east', for example, i s defined phonetically by the gloss de and hong faj- : that i s , d(e) - (h) png = dong. The f an-qie method works well when the f i n a l of a word contains only the rhyme, but breaks down when the f i n a l includes a medial 1, which i s not consistently found i n the second character i n the fan-qie. While fan-qie i s involved i n grouping words which contain the same rhyme, the niu £.3- (.'knot') p r i n c i p l e i s used i n categori- zing words which have the same i n i t i a l ; that i s , words containing the same i n i t i a l which are grouped together belong to the same niu (Pulleyblank, 1977). Beyond the d i v i s i o n of a character into i n i t i a l and f i n a l , the further decomposition of the f i n a l into smaller component parts such as nucleus and ending i s the i n f l u - ence of Western approach, which analyzes a s y l l a b l e i n terms of a li n e a r sequence of sounds arranged according to t h e i r temporal pos i t i o n i n a spoken chain. The canonical shape of the s y l l a b l e i n Zhong-shan (Chart 2) can be analyzed using the terms introduced above. The i n t i a l i n Zhong-shan i s a consonant which, when omitted, constitutes the "zero" i n i t i a l , and i s assigned the phoneme /0/ ( i . e . , the n u l l element, for convenience of reference; the zero i n i t i a l i s omitted - 19 - Chart 2. Syllable Structure of Zhong-shan. T (I) + (M) + N + (E) Examples: £T r 55, L t e a i IJ J / t + i + a: + 55 ng /= 55 /tia:ng / ' n a i l Lkwax J / k + w + a + _ 5 5 /= /kwa 5 5/ 'melon' - U ? r 5 1 i IWAin J / w + — + a: + n 5 1 / = /wa:n 5 1/ 'yet' [ J A i k 2 ] / j + — + a: + k 2 /= / j a : k 2 / 'eat' t [ixp 2J / — + — + i + 2 P /= / i p 2 / ' l e a f 3- In J / — + — + ng + _ 1 3 /= / n g 1 3 / ' f i v e ' where T = Tone I = I n i t i a l (Consonant or "zero" i n i t i a l ) M = Medial . ( / i / or /w/) N = Nucleus (Vowel or s y l l a b i c nasal) E = Ending (Glide, stop, or nasal) i n actual t r a n s c r i p t i o n s ) . With regard to the f i n a l , the f i r s t element i s the medial, which i s an optional segment. Of central importance i s the nucleus, which i s a vowel or a s y l l a b i c nasal i n Zhong-shan. The ending i s optional, consisting of a gl i d e , stop or nasal consonant. The tone, indicated by the use of superscripts, following the s y l l a b l e , i s treated as a suprasegmental feature over the entire s y l l a b l e . With regard to conventions for indi c a t i n g tone, Y.R. Chao has devised two systems, one using tone l e t t e r s and the other .... - 20 numerals. Both systems continue to be used by Chinese l i n g u i s t s . In the f i r s t , tone l e t t e r s graphically represent the p i t c h height and the contour of a tone. The tone l e t t e r s for the various l e v e l tones, for instance, are "1 1 -\ -I J", with a gradual descent i n pitch height represented by the gradual lowering of the horizontal stroke. Tonal length i s also graphically represented. A long . l e v e l tone, for example, i s indicated by a comparatively longer horizontal stroke than a short l e v e l tone (e.g., "1 versus 1 ) . Short l e v e l tones are generally used for s y l l a b l e s ending i n a stop consonant. The second system for tonal t r a n s c r i p t i o n i s a numerical one which uses the scale of "1" to "5" wherein "1" indicates the lowest pitch and "5" the highest. Thus, a high l e v e l tone may be represented by "55", for instance, and a high f a l l i n g tone by "51", and so forth. A single number i s used for denoting short s y l l a - bles, e s p e c i a l l y relevant to the t r a n s c r i p t i o n of s y l l a b l e s with stop consonant endings. For typographical and comparative reasons, i t i s t h i s l a t t e r system of tone notation which w i l l be employed i n the present study. Transcription of consonants i s b a s i c a l l y i n accordance with the International Phonetic Alphabet (IPA), with a few modifi- cations. Aspiration, for example, i s indicated by the d i a c r i t i c ['] (e.g., [p'] instead of IPA [phj) . The segment I t s J represents an a f f r i c a t e rather than a sequence of two segments; I t s ' ] i s the aspirated counterpart. - 2 1 - 1 . 1 . I n i t i a l s The consonants i n Table 1 represent the inventory of i n i t i a l consonants i n broad phonetic t r a n s c r i p t i o n . Those con-, sonants accompanied by a dash. "-" occur i n s y l l a b l e - i n i t i a l p o s ition only. Table 1 . The I n i t i a l s i n Zhong-shan. Plosive Unasp*. Asp'd. Nasal Lateral F r i c a t i v e Glide B i l a b i a l p p i - m w Labiodental f - Dental t t'~ n 1- Alveolar t s - t s ' - j Velar k k'- q G l o t t a l (?) h- The broad phonetic t r a n s c r i p t i o n i n Table 1 above i s e s s e n t i a l l y i d e n t i c a l to a phonemic inventory that can be arrived at by applying the p r i n c i p l e s of complementary d i s t r i b u t i o n and phonetic s i m i l a r i t y . Thus, Table 1 also serves as the phonemic system of Zhong-shan i n i t i a l s . For typographical reasons, /ng/ w i l l be used to represent the velar nasal [q] i n the phonemic system and should be considered a unit phoneme. Likewise, / t s / and / t s ' / constitute unit phonemes. In the following paragraphs i s a more detailed discussion of some of the phonemes and t h e i r allophones. The phonemes /p/, / t / , /ts/.and /k/ are unaspirated, and generally quite weak, voiceless plosives, often perceived phonetically as [bj , [dj , {d4z] - 22 - and [g], and sometimes even with s l i g h t voicing. Note that t h i s observation d i f f e r s from that made by Chao (p.51) who asserts that, unlike Mandarin, the di a l e c t s of Cantonese, Zhong-shan and Wu are a l i k e i n having strong, unaspirated stops i n s y l l a b l e - i n i t i a l p o s i t i o n . There i s strong aspiration i n the aspirated s e r i e s , p a r t i c u l a r l y apparent i n the careful enunciation of in d i v i d u a l characters containing them. Some lip-rounding accompanies the i n i t i a l s when they occur before the high rounded vowel [u:J (e.g., ̂ [p w, p < W / t s W , k w, h W ] ) . I t was further observed that s l i g h t a f f r i c a t i o n may it 5 22 occur, as i n the word bu tf? 'step', phonetically [p ux. J . There i s some disagreement among l i n g u i s t s concerning the nature of the s i b i l a n t s i n Zhong-shan. Chao (p.51) feels that there i s c l e a r l y no p a l a t a l i z a t i o n i n the Shi-qi d i a l e c t , i n contrast to his observation of strong p a l a t a l i z a t i o n i n Cantonese. Egerod (p.13), on the other hand, posits a p a l a t a l series for both Cantonese and Sh i - q i , with the phonemes /c/, /ch/ and /s/ 3 s' representing the phonetic segments [c J, [c J and [s] respectively. B a l l ' s analysis may constitute a compromise i n that complementary d i s t r i b u t i o n can perhaps be in f e r r e d from his t r a n s c r i p t i o n s : i n general, B a l l ' s / t s / occurs with back vowels while his /ch/ occurs with front ones, which suggests that the p a l a t a l /ch/ i s the 2 res u l t of assimilation to the following front vowel. . ' ..It was observed i n the writer's data that Chen's pronun- c i a t i o n tends to be s l i g h t l y more p a l a t a l than Yang's, but i s nonetheless not as strongly p a l a t a l i z e d as found among some speak-7 ers of Cantonese. In her study of Cantonese, Hashimoto (p.88) 23 - describes the s i b i l a n t s i n her i d i o l e c t as dental i n a r t i c u l a - t i o n , but adds that the place of a r t i c u l a t i o n of t h i s series of sound, ranges from dental to p a l a t a l among Cantonese speakers, with a tendency for some degree of p a l a t a l i z a t i o n preceding the high front vowels. Hashimoto (p.l7,fn.8) further c i t e s an obser- vation made by D.C. Lau at the University of London who found that there i s a greater tendency f o r male speakers than female speakers to p a l a t a l i z e . This would concur with our observations on Chen and Yang's speech with regard to the r e l a t i v e degree of p a l a t a l i z a t i o n of the s i b i l a n t s e r i e s . As noted by both Chao and Egerod, the i n i t i a l nasal con- sonants are often accompanied by homorganic stops: [m3, n^, n~̂ ] . However, i n the case of a non-labial nasal followed by the high, back vowel [u:], i t was found that the homorganic stop may be replaced by l a b i a l [b] i n a n t i c i p a t i o n of the following l a b i a l segment, as i n one of the repetitions i n the pronunciation of the word nu Ijp 'anger', which was rendered [n u: ] by Chen. At times, instead of the nasal consonant being accompanied by a homorganic stop, the i n i t i a l segment becomes i n fact a prenasalized stop, as i n Chen's pronunciation of the word men 'door'. I t 51 i s [mum J i n broad phonetic t r a n s c r i p t i o n ; i n narrow phonetic t r a n s c r i p t i o n , however, the word should be recorded as { mbu:n^ ij. The alternation between /n/ and /!/ observed i n the speech of a number of Hong Kong speakers i s not found i n Zhong-shan, which maintains a c l e a r d i s t i n c t i o n between these two i n i t i a l s . None of the words which had the h i s t o r i c a l Ni yf<J 3 (*n-) i n i t i a l are pro- nounced with the l a t e r a l [11 i n Zhong-shan. There are only three cases i n the survey of an In] corresponding to the L a i - 24 - 13 i n i t i a l : (.1) the character l i a n g fa which i s pronounced [n0ot) ] 13 with the meaning of 'two (173-12) and [loan ] when i t i s used as a measure word to mean 'a t a e l ' (173-13); (.2) the character l i n g 4f| C201-7) for ' c o l l a r ' and 'to apply for which has the l i t e r a r y 13 13 reading of I l i n ] , while a c o l l o q u i a l reading of [neairj ] only 13 applies to the f i r s t meaning, and that of [leain J only to the second meaning; and (3) the character l i (1.13-13) which i s a c l a s s i f i e r for small, round objects such as seeds, grain, pearls, 5 buttons, etc., and i s pronounced InBp J. Although any explanation of why these three characters have acquired a nasal i n i t i a l can only be speculative, i t i s possible that the /n/ i s a r e s u l t of borrowing i n a l l three cases. Two of the three words, liang and l i n g , can be found i n the very useful c r o s s - d i a l e c t a l reference source, the Han-yu Fang-yan Zi-hui y& 7f "if % (.'Chinese Dialect S y l l a - bary', editted by the B e i j i n g Daxue Zhong-guo Yu-yan Wen-xue-xi Yu-yan-xue Jiao-yan-shi (.1962)—hereafter referred to as "Zi-hui" for short, with page reference included only). There i s only one entry for the word l i a n g . Since the Zi-hui does not provide glosses, one can only suppose that the two meanings of the word, 'two' and ' t a e l ' , are combined i n the single entry i n that s y l l a - bary. While most of the d i a l e c t s have the l a t e r a l 111 as the r e f l e x of the h i s t o r i c a l Lai i n i t i a l , a few of the d i a l e c t s represented therein show an [n] i n i t i a l . In Amoy (a Southern Min dialect), two forms contain an In] i n i t i a l , and one an 11J i n i t i a l , p a r t i a l l y r e f l e c t i n g s t y l i s t i c differences: .[lion] i s . l i t e r a r y 4 whereas [niu]. and Inp] are c o l l o q u i a l readings. However, [n] and [1] are i n complementary d i s t r i b u t i o n i n Amoy, with In], occurring - 25 - 5 before nasalized vowels and [q] and [1] elsewhere. Although there may be contact between Zhong-shan and Amoy, or some other si m i l a r Southern Min d i a l e c t , stronger evidence for possible d i a l e c t a l influence comes from Fu-zhou, a Northeastern Min d i a l e c t (data provided by J. Norman). Not a l l Fu-zhou speakers maintain a phonemic d i s t i n c t i o n between /n/ and / l / ; however, among those who do, 'two' i s [naq 2] , while ' t a e l ' i s I - l i o r ) ] . ^ The /n/-~> / l / s p l i t i n Fu-zhou not only serves the same function as that found i n Zhong-shan, but the same assignment of i n i t i a l to meaning occurs i n the two d i a l e c t s . There are no data on the Long-du d i a l e c t ("Lungtu" i n Wade-Giles romanization), which i s spoken i n the second qu i n Zhong-shan, and i s regarded by both. J. Norman and N. Bodman as a Northeastern Min d i a l e c t . Neverthe- les s , one would suspect that Zhong-shan borrowed the / n / ~ / l / s p l i t for the two meanings of the word l i a n g xfc) from Long-du—if such a d i s t i n c t i o n occurs i n that d i a l e c t — o r from Fu-zhou or another Northeastern d i a l e c t where the same d i s t i n c t i o n i s found. It might also be noted that although Hashimoto (p.570) records a l a t e r a l i n i t i a l f o r both meanings of the character l i a n g 1$) , whein i t i s useid to mean ' two1 , the character i s pronounced i n the Yang-shang tone 1 2 4 ] , while i t s use to mean 'tae l ' i s rendered in the Yin-shang tone I 3 5 J . Hashimoto Cp-668) suggests that since lj.ang as a measure word i s a c o l l o q u i a l term, i t s Yin-shang tone may be the product of either changed tone phenomenon ( i . e . , an o r i g i n a l tone changed to a high r i s i n g I 3 5 J ) , or influence from the c o l l o q u i a l layer. The evidence thus far demonstrates, ..firstly,. that the two meanings of the character l i a n g do diverge i n t h e i r suscepti- - 26 - b i l i t y to external influences, with. Zhong-shan and Fu-zhou show- ing an alternation of i n i t i a l s , while Cantonese exhibits a tonal d i f f e r e n t i a t i o n . Secondly, and more importantly with regard to Zhong-shan, given the paucity of exceptions i n the modern r e f l e x of the Lai i n i t i a l i n Zhong-shan and the evidence from Fu-zhou, i t would be l o g i c a l to suggest that the reading of l i a n g which has the /n/ i n i t i a l i n Zhong-shan i s l i k e l y to have been the one borrowed into the d i a l e c t , whereas the one with the / l / i n i t i a l i s the native form. A p a r a l l e l case may be argued for the character l i n g , which has both an / l / and an /n/ i n i t i a l i n the c o l l o q u i a l reading 13 13 i n Zhong-shan: the form [leanj J means 'to apply for' and [neain J ' c o l l a r ' . Again, there are several d i a l e c t s i n the Zi-hui (p.255) which show an [n] i n i t i a l for the character l i n g . Of the southern Chinese d i a l e c t s recorded i n the Zi-hui, only Chao-zhou (Southern 52 Min) shows an [n] i n i t i a l for the character: [nia J (with no other form given). In general, the / n / ^ / l / d i s t i n c t i o n i s not very well maintained i n Southern Min d i a l e c t s . I t i s noted that such a d i s - t i n c t i o n i s likewise not always maintained i n Northeastern Min. It i s therefore conceivable that the nasal i n i t i a l for the c o l l o - q u i a l reading of ' c o l l a r ' i s a borrowing from one of the Min dia- l e c t s . Chao's data on the Zhong-shan d i a l e c t show a l a t e r a l .. i n i t i a l for the c o l l o q u i a l reading of the character l i n g 4%. r which 13 7 i s transcribed by Chao phonemically as / l i a : n g /. I t i s s i g n i f i - cant that Chao leaves the s l o t for the potential s y l l a b l e / n i a : n g x 3 / i n his syllabary empty, which quite strongly suggests that the 13 pronunciation of / l i a : n g / for the character l i n g encompasses - 27 - both, the meanings of ' c o l l a r 1 and 'to apply f o r ' . The present data on Shi - q i , which y i e l d s a nasal form, would lead to the proposal that the borrowing may be a recent one. In contrast, the character l i a n g xfi) i n Chao's data agrees with the present corpus i n haying both a l a t e r a l and a nasal i n i t i a l , which probably r e f l e c t s an e a r l i e r borrowing. The character l i %.JL [n^p^j , which i s a measure word f o r small, round objects, only has the /n/ i n i t i a l i n Zhong-shan. I t i s l i k e l y that t h i s i s also a borrowed form, although the writer has no concrete c r o s s - d i a l e c t a l data on which to support t h i s claim. The general observations made above on the Min di a l e c t s and the r a r i t y of /n/ as a r e f l e x of the Lai i n i t i a l i n Zhong-shan would suggest a sim i l a r borrowing i n t h i s t h i r d and l a s t case of exceptions to the Lai i n i t i a l i n the Zhong-shan data. Moreover, although Hashimoto (p.513) records [lsp ] for the character, i t i s noteworthy that Huang (1970:425)., for example, posits only an /n/ i n i t i a l for the same character, carrying the meaning of 'measure of seeds, grains, etc.' i n his Cantonese Dictionary. The word does not occur elsewhere i n the dictionary, or more precisely, i t does not appear i n the section under / l / . (Huang transcribes the word as nap, i n Yale romanization.) S.L. Wong (.1954)_ likewise records the word L i under the s y l l a b l e [nsp] for Cantonese. However, Wong does note that the character i s also read s i m i l a r to that of the h i s t o r i c a l l y homophonous word, l i j£ . 5 (.113-12) ; that i s , l i has an alternate pronunciation of [ l B p J. One can, however, assume that the reading of the word l i %JL with the /n/ i n i t i a l i n Cantonese i s the primary one i n Wong's Chinese syllabary based on the pronunciation i n Canton. Given Wong's decision to record l i . %JL. only, under .the /n/ i n i t i a l category, and Huang's recording of the word only with an /n/ i n i t i a l , i t appears that not only Zhong-shan, but Cantonese speakers as well, who do not normally "confuse" /n/ and /!/ i n i t i a l s would neverthe- less pronounce the word l i . JJC^L with a nasal i n i t i a l . The i n i t i a l /h/ i s a g l o t t a l f r i c a t i v e . Chao (p. 51). how- ever states that when /h/ precedes the back vowels /u/ and /o/, there i s some velar or uvular f r i c a t i o n , r e s u l t i n g i n a sound which i s almost .[;xj . This allophonic v a r i a t i o n was not observed i n the speech, of Chen and Yang to any s i g n i f i c a n t degree. g The zero i n i t i a l /#/ , according to Chao (p.51), i s a smooth onset before high vowels, but i s usually a g l o t t a l stop [jj before mid and low vowels. This allophonic d i s t r i b u t i o n of the zero i n i t i a l was not consistently observed i n the present data. The zero i n i t i a l occurs before f i n a l s beginning with a nuclear vowel, as well as before s y l l a b i c nasals. In the representation of a s y l l a b l e containing a zero i n i t i a l , the n u l l symbol "ft" i s actually never transcribed. Egerod records a phonemic g l o t t a l stop /?/ rather than a zero i n i t i a l , which he posits before vowels and s y l l a b i c nasals. Egerod makes no note of allophonic v a r i a t i o n s . Before discussing the semi-vowel i n i t i a l s , i t should be made clear that the phonemic analysis of both Egerod and the writer distinguishes between vowels and glides and, moreover, allows the l a t t e r to serve as i n i t i a l s . Chao, on the other hand, does not d i f f e r e n t i a t e glides from vowels i n his analysis. He treats [j] as an allophone of the phoneme/i/, and [wj as that of /u/. His rationale i s founded on the s t r i c t application of the p r i n c i p l e of complementary d i s t r i b u t i o n ; since whether the segment - 29. T / i / , for instance, i s the vowel {ij or the semi-vowel {jj can be predicted on the basis of i t s po s i t i o n i n the s y l l a b l e and/or what rhyme or f i n a l accompanies i t , Chao combines the front vowel and p a l a t a l glide under the phoneme / i / . The corresponding back vowel [u] and the l a b i a l glide {wj are incorporated under the phoneme /u/ on s i m i l a r grounds. Since vowels and glides are i n complementary d i s t r i b u t i o n i n Zhong-shan, contrasting them phonemically would inevitably create redundancies i n the system. Egerod's d i s t i n c t i o n of glides and vowels i s probably prompted by the desire to set up the vowel clus t e r /ua/, which he records phonetically as [ox>] or [ D : J . Set- ti n g up the phonemes /w/ and /u/ allows such minimal pairs as /kwan/ and /kuan/, which are phonetically [kwerij and [kom] respec- t i v e l y (using our phonetic t r a n s c r i p t i o n s ) . Although the present analysis does not have such an objective i n mind, i t i s i n the treatment of medials that the recognition of vowels versus glides becomes a p a r t i c u l a r l y important issue, as we s h a l l see l a t e r . Chao also d i f f e r s from Egerod and the writer i n that he does not treat the glides as i n i t i a l segments. They are not d i s - cussed i n the section on i n i t i a l s . Chao (pp.51,53) refers to them e x p l i c i t l y as medials ( j i e - y i n ^ 'medial sounds'). The > / 22 phoneme /u/ i n the word hua /ua / (in Chao's t r a n s c r i p t i o n , 22 phonetically [WA: ] ) , i s analyzed by Chao as a medial, whereas both Egerod and the writer would analyze that segment as an i n i t i a l . Although on the whole, there i s h i s t o r i c a l basis for analyzing a glide i n such a position i n the modern d i a l e c t as a medial pre- ceded by the zero i n i t i a l , i t i s nonetheless a very costly step i n that i t would substantially increase the number of f i n a l s . - 30 - At the same time, many of these f i n a l s would actually have a very low functional load. F i n a l s such as IvoinT., IwDikJ, [jun] and I q 0 q j , for example, would only occur with the zero i n i t i a l . Chao solves the problem by putting the two glides (his / i / and /u/ phonemically) i n brackets i n his chart of i n i t i a l s and, in his syllabary, he posits these two medials under what i s i n : fac t the category of i n i t i a l s . Thus, although Chao analyses them as medials, he i s actually treating them as i n i t i a l s without actually acknowledging i t . Only i n the case of his f i n a l s / i a / , /ia:ng/ and /ia:k/ does Chao treat / i / as a medial i n his syllabary, the reason being that in these f i n a l s , medial / i / does co-occur with, most of the i n i t i a l s and hence carry a heavy functional load. However, by allowing' / i / to occupy both i n i t i a l and medial po s i t i o n i n his syllabary, Chao also creates c e r t a i n redundancies; a s y l l a - ble such as * / i i a / , for example, would not be possible i n the J d i a l e c t , but i t could be generated from the combination of " i n i t i a l + f i n a l " i n h i s syllabary. In the present analysis, a si m i l a r problem i s avoided by stating the following r e s t r i c t i o n : the two medials, /w/ and / i / , never co-occur with the zero i n i t i a l . (The two semi-vowel i n i t i a l s are /w/ and / j / . ) . According to t h i s co-occurrence r e s t r i c t i o n , the phoneme /w/ i n the s y l l a b l e /kwa/, fo r example, i s analyzed as a medial whereas, i t i s an i n i t i a l i n the s y l l a b l e /wa/. The phoneme / i / i s a medial i n the s y l l a b l e /kia:k/, for instance, while / j / i n the s y l l a b l e such as /ja:k/ i s an i n i t i a l . (The phoneme / i / can only occur s y l l a b l e - i n i t i a l l y when i t i s the nuclear yowel Ii,:].) The reasons f o r choosing an asymmetrical pair of phonemes (/w/ and / i / ) . for the medials w i l l be discussed - 31 - l a t e r . Returning now to the discussion of glide i n i t i a l s , /w/ i s simply the back, rounded semi-vowel [wj, while / j / i s usually the front glide I j ] . Before the front rounded vowel [ 0 ] , however, the glide assimilates to the rounding of the following segment, and thus becomes the front, rounded semi-vowel [qj. The only occur- rence of t h i s regressive assimilation i n the present data i s i n the s y l l a b l e / j 0 j / , phonetically [i[0uj , with the nuclear vowel af f e c t i n g both the preceding and the following segment. With regard to a comparison of the i n i t i a l s i n Zhong-shan and Cantonese, i n general, there i s very l i t t l e difference between the two d i a l e c t s . One observation discussed e a r l i e r was the d i f f e r - ence i n degree of p a l a t a l i z a t i o n present i n t h e ^ s i b i l a n t s i n the two d i a l e c t s . Another, concerns the nasal i n i t i a l s . Unlike Zhong-shan and some of the other Yue d i a l e c t s (e.g., S i - y i d i a l e c t s ) , Canton^ ese does not have homorganic stops accompanying the a r t i c u l a t i o n of nasals. Another difference that i s found i n Cantonese i s the presence of homorganic glides preceding the high vowels. Thus, the Zhong-shan s y l l a b l e [u:n] would be [wu:n] i n Cantonese. Like- wise, Zhong-shan [ix] corresponds to Cantonese [ j i x ] , and Zhong- shan [yx] i s Cantonese [qyx]. Moreover, Cantonese i s often treated as containing a phonological contrast between a p l a i n versus a l a b i a l i z e d velar i n i t i a l ( i . e . , /k/ versus /kw/, and /k 1/ versus /k'w/). The l a b i a l i z e d counterparts are treated i n Zhong-shan as a sequence of a velar i n i t i a l plus a l a b i a l medial to correspond to the non-labial medial / i / . The absence of a medial / i / i n Cantonese, combined with a heavier functional load of the l a b i a l - ized i n i t i a l s i n Cantonese make the option of treating /kw/ and - 32 - /k'w/ as i n i t i a l segments a p r a c t i c a l one for Cantonese. 1.2. Finals. A broad phonetic t r a n s c r i p t i o n of the f i n a l s i n Zhong-shan, together with a phonemic analysis, i s presented i n Table 2 (a) on page 35. In the table, the phonemic renditions of the medials and s y l l a b l e nuclei are placed i n the left-most column, while the endings are recorded across the very top. J u s t i f i c a t i o n for the pa r t i c u l a r phonemic solution proposed here w i l l be detailed l a t e r . In the meantime, for the purpose of comparison, Chao and Egerod's treatment of the f i n a l s are re-interpreted i n Tables 2 (b). and (c) i n order to f i t into the format of Table 2 C a ) . The f i n a l s i n B a l l ' s study are given i n Table 2 (d). For cross-dia- l e c t a l comparison with Cantonese, the f i n a l s i n Cantonese, based on Hashimoto (p.90), are placed i n Table 2 (e). To conform with, the format established here for Zhong-shan, the l a b i a l feature of the l a b i a l i z e d velar i n i t i a l s i s treated as a medial i n the table. Such a treatment w i l l also prove useful i n the l a t e r comparison of h i s t o r i c a l reflexes of various rhyme groups. Egerod's data cause some d i f f i c u l t y since he did not provide a syllabary for Shi-qi and his corpus of data on the d i a l e c t i s extremely limited. I t w i l l be assumed here that the f i n a l s found i n Chao's study are also present i n Egerod's. More- over, some segments which Egerod has analyzed as "vowel c l u s t e r s " have been redi s t r i b u t e d i n Table 2 (_c). : his /aa/ c l u s t e r i s regarded as a simple "vowel nucleus", while the clusters / i a / , /0a/ and /ua/ as well as the sequences /wa/ and /waa/ (_with /w/ o r i g i n a l l y part of the velar consonant c l u s t e r ) , are considered a combination of - 33 - "media l + vowel n u c l e u s " , based on h i s t o r i c a l and c r o s s - d i a l e c t a l d a t a . I t i s more d i f f i c u l t to know the p r e c i s e phonet i c va lue J . D . B a l l in tended i n h i s t r a n s c r i p t i o n of Macao Zhong-shan. In some i n s t a n c e s , there i s an attempt to t r a n s c r i b e the words p h o n e t i c a l l y : f o r example, h i s vowels / i / and / ! / correspond to [ i ] and [ i f ] r e s p e c t i v e l y i n modern ( S h i - q i ) Zhong-shan. However, /'£/ i s a l s o used f o r a sound which i s now the f r o n t g l i d e [ J J . In o ther s i t u a t i o n s , i t ^ a p p e a r s to be s imply an o v e r - d i f f e r e n t i a - t i o n of forms. The i n i t i a l s / t s / and / c h / , mentioned e a r l i e r , are t r a n s c r i b e d as though they may be c o n t r a s t i v e i n the d i a l e c t . I t i s l i k e l y t h a t such a c o n t r a s t between a d e n t a l and an a l v e o l a r (or p a l a t a l ) s e r i e s — a l s o found i n a number of e a r l i e r works on Cantonese ( i n c l u d i n g those by B a l l ) — i s i n f a c t an a r t i f i c i a l c r e a t i o n which became q u i t e w ide ly accepted s i n c e i t had f u n c t i o n a l v a l u e . The o v e r - d i f f e r e n t i a t i o n i s based on a d i s t i n c t i o n found i n the n a t i o n a l s tandard wherein the d e n t a l s i b i l a n t s i n Cantonese g e n e r a l l y correspond to the d e n t a l / p a l a t a l s e r i e s i n Mandar in , and the Cantonese a l v e o l a r s (or p a l a t a l s , as the case may be) t h a t of the r e t r o f l e x s e r i e s i n the Mandarin d i a l e c t . I t i s p r i m a r i l y f o r the purpose of a i d i n g Cantonese speakers i n l e a r n i n g the n a t i o n a l d i a l e c t t h a t the two s e r i e s had o f t e n been kept apar t i n Cantonese, even though such a d i s t i n c t i o n i s founded on pedagog ica l r a t h e r than l i n g u i s t i c grounds (see, f o r example, Chao (.1947:18- 19); Yuan (.1960:183), e t c . ) . The f a c t t h a t a s e r i e s of d e n t a l versus p a l a t a l s i b i l a n t s i s phonemic i n Shun-de, which i s one of the Yue d i a l e c t s but i s by no means r e p r e s e n t a t i v e of the d i a l e c t - 34 - family as a whole, may have contributed to the o r i g i n a l d i f f e r e n - t i a t i o n recorded i n e a r l i e r writings on Cantonese. Some of B a l l ' s f i n a l s also appear to be overly d i f f e r e n t i - ated; for example, B a l l transcribes the f i n a l s of the three words ming ^ (.200.-2) 'name', j i n g fcfa (200-3) 'quick-witted', and l i n g ^ (201-8) ' h i l l ' i n three d i f f e r e n t ways, as shown i n (.1) below. The f i n a l s of a l l three words are pronounced £ea:y] i n the c o l l o q u i a l reading i n Zhong-shan. H i s t o r i c a l l y , these f i n a l s not only belong to the same rhyme group ( G e n g ^ ^ ) ^ , but also to the same rhyme (Qing ^ ) and the same grade ( I I I ) 1 ( \ There i s no reason to expect that these f i n a l s recorded by B a l l should have re f l e c t e d a contrast i n t h e i r pronunciation s u f f i c i e n t to warrant three d i f f e r e n t means of t r a n s c r i p t i o n . (1) ..Ball Chan .Gloss ^ /meeng/ /mia:ng 5 1/ ([mea:n 5 x]) 'name' 55 ,,rj_ 55-L 5 /tsieng/ /tsiasng /([tseaxq ]) 'quick-witted' ^ /leng/ / l i a : n g 1 3 / ( [ l e a i n 1 3 ] ) ' h i l l ' In the recording of B a l l ' s f i n a l s i n Table 2 (d), some of the d i a c r i t i c a l marks that he used i n his a r t i c l e are not very d i s t i n c t i n the microfiche of his a r t i c l e from which a photocopy was made for t h i s study. As a r e s u l t , i t i s possible that the writer may have recorded a circumflex {*) where i t should have been an umlaut ("), a grave accent (>) instead of a circumflex, etc. An attempt i s nevertheless made to be consistent. F i n a l s /ing/ and / i k / are found i n Shi-qi speech only. Macao and the surrounding areas use /ang/ and /ak/ instead (phonetically {BIJJ and Ink] i n modern speech). - 35 - T a b l e 2 ( a ) . A n a l y s i s o f t h e F i n a l s — C h a n . \ e n d i n g voweTS. nucleuses - j -w -m -n -ng "P - t - k i i i i i w i im i i n if) i i p i i t i k y y * y i n y i t u u : ux j u i n un U I t uk 03 I- 0U_ 0n 0 t o 3 X 31 j ow o :n o i n D i n D i p D l t D i k a BJ e n BQ Bp B t Bk a: A : A I j A I W Aim A i n A i n A I p A l t A i k 1 medial • vowel nucleus i o 0 31} 0Dk i a : ea.i- e a: w eaim eai.y ea,i;p ea i t ea i k w i w i k wa WBJ wBn WBQ WBt WBk wa: WA : WA I j WA i n WA i g WA I t WA i k s y l l a b i c nasal m ? - 36 - Table 2 (b). Analysis of the Finals—Chao. N e n d i n g v o w e r v n u c l e u s ^ - i -u -m -n -P ~ t - k i i i i i w i i m i i n e q i i p i i t e k y y x y i n y i t u u x u : j u i n U l t CE 031 n 031 q 03 i k O e y o w eri o n et o k o: D i =>* j D : m D i n D i n D i p D i t D i k a A W •em Bn Bn Bp B t B k a: A I A X j A : w A i m A i n A i n A I p A I t A i k m e d i a l * v o w e l n u c l e u s i a : i A : i a i i ) i a i k wa WB j w e n WBt wa: WA I WA I j W A i n WAin WA I t WA i k s y l l a b i c n a s a l m J ] * - 37 - Table 2 (c) . Analysis of the FihaTs--KEgerod. sending vowePv nucleus^ - j -w -m -n -q -P.. - t - k i i : i i w iim i i n eq ixp ixt ek y V- y xn y i t u u X ux j u xn OJ} u I t ok 0 oey oen (Bt a a j aw am an aq ap at ak aa a i axj axw aim a in a xq a xp a i t a i k medial+ vowel nuc] Leus i a ea eaq e ak 0a (BO oeoq oeok ua oo oon Doq DOt DOk wa wa j wan waq wat waa wa i wax j wain wax q wa i t wa i k syllabic nasal m i r) - 38 - Table 2 Cd) . Transcription of Finals i n (Macao) Zhong-shan-- B a l l """-̂ ^ ending vower\^ nucleus ~ ~2- -u - i i -m -n -ng -p - t -k * [ i ] ( i ) , (ing) . (ik) [ i i i • . 1 : . - s X / s / 1 U l f f l xn xp x̂ t [u] u ung uk [yx] ii ii iin ut [ U l ] u • u un lit [ U l j uu / / U U l [ 0,ux] uo uou [oex ,ceo] 6 6 oii ong ok loi] 0 0 on ong ot ok I D : ] 6" A f O l [owj V 0 b [ B ] a • a i au am an ang ap at ak [ A X ] • a • // a ax au am an ang ap at ak medial+ vowel nucleus I e a x ] e e eng ek lea x J i e xeng xek [ e a x J ye yeu yem yep [ eax] A ee eeng leaxj ya (ya) yam syllabic nasal m Modern Zhong-shan equivalents are recorded on the left-most column. Zhong- shan [q] = Ball's /ung/. > - 39 - T a b l e 2 (e) . Phohe t i c T r a n s c r i p t i o n o f t h e F i n a l s i n C a n t o n e s e . Nending vowelv n u c l e u s \ y \J v - i - y - u . -m - n - 0 - p - t - k i i n i k i : i x i : u i x m i xn i x-p i x t y* V- y xn y i t u UI) uk UI U I Y U X 1 uxn UX-t e e i e xn e xk e V e i 0 0y 0n 0 t cs ce: cexrj ce x k 0 ou D 1 D X V 3X1 o xn O I t ) o x t oxk B B l BU Bm Bn Bl) Bp B t Bk A I A : A l l AXU A i m A xn AXt) AXp A X t AXk medial+ vowel nuc! .eus W I w i n w i k wi x wi x W D A wo x. wo x q WD xk WB WB1 WBn WBq WBt WBk WA I WA X WA X 1 WA xn WA X 0 W A I t WAlk s y l l a b i c nasal m 1 0 • - 40 - 1.2.1. Nuclear Vowels Of the f i n a l s , the nuclear vowels w i l l be discussed f i r s t . In the present analysis of Zhong-shan, they form a t h r e e - t i e r , six-vowel system with a phonemic contrast of length i n the low vowels only, as diagrammed i n Chart 3 (a). It may be observed that, generally, tense and lax vowels—or long and short vowels— sharing the same vowel height (high, mid, low) serve as allophones of the same phoneme. The chart also includes an analysis of the combination of "medial + vowel". Comparative charts of Chao and Egerod's vowel system for Zhong-shan are shown i n Charts 3 (b) and (c) respectively. As i n Table 2, part of t h e i r phonemic system i s re-interpreted i n terms of "medial + vowel" for which Chart 3 Ca) serves as the model. Chart 3 (a). Analysis of the Vowels—Chan. Phonetic Representation: I i i u: u 0 ce: rzi Phonemic System: i y u 0 o a/a: A : Medial + Vowel: tea:! = / i a / [03] = / i o / IW B3 = /wa/ [WAX.]' = /wa:/ - 41 - Chart 3 ( b ) . Analysis of the Vowels—Chao. Phonetic Representation: y* u: ce i Medial + Vowel; [ia:] = / i a / . [we] = /ua/ [ W A:] = /ua:/ Phonemic System: i y u CE o/o: a/a: Chart 3 (c). Analysis of the Vowels—-Egerod. Phonetic Representation: 0 l : e u: o Medial + Vowel: [ea] = / i a / [ceo] = /0a/ [ob]/[oi] = /ua/ [wa] = /wa/ [wax] = /waa/ Phonemic System: i y u a/aa - 42 - There are three phonemic h i g h vowels i n Zhong-shan: / i / , /y/ and /u/. The phoneme / i / i s pronounced Ii:.] i n s y l l a b l e - f i n a l position, and when i t i s followed by the l a b i a l glide /w/f the l a b i a l o r dental n a s a l ( i . e . , /m/ o r /n/) , o r the correspond- i n g s t o p s C/p/ and / t / ) . Chao describes the vowel as being almost the cardinal I i ] . Before the stop endings /p/ and / t / , there i s some tendency towards pronouncing the vowel with a schwa o f f - g l i d e thus: I i 9 p ] and I i 9 t ] , f i r s t noted by Chao (p.53). Sometimes i t i s a case o f the laxing o f I i : J , r e s u l t i n g i n the pronunciation o f the word j i e 'knot', f o r instance, as I k i ' t 2 ] (phonemically 2 / k i t /).. At other times, the o f f - g l i d e i s further lowered, p r o - ducing the segment [ i e ] , as observed i n the pronunciation o f the word j i e 'outstanding' as I k i e t 2 J . The phonetic description above i s based on Chen's pronunciation. In general, however, the vowel i s simply a l o n g I i : ] . Egerod (p.14) records the l o n g vowel Ii:. J without elaboration. C r o s s - d i a l e c t a l l y , the si t u a t i o n i s quite d i f f e r e n t : what i s merely a l o w - l e v e l schwa o f f - g l i d e found i n some / i p / and / i t / f i n a l s i n Zhong-shan i s a f u l l vowel i n other d i a l e c t s such as Mandarin, Southern Min and Hakka. The: character j i e 'knot' , 35 32 f o r example, i s pronounced [tcsie ] i n Mandarin, I k i a t ] i n Amoy 21 11 (Southern Min), and I k i e t ] i n M e i - x i a n (Hakka) ( Z i - h u i , p.34). This f i n a l i s reconstructed by Karlgren as Ancient Chinese ( o r "Anc." f o r short) * i a t i n D i v i s i o n I I I , and * i e t i n D i v i s i o n IV; and by Pulleyblank as Late Middle Chinese (.LMC). * i a t f o r both grades, o r d i v i s i o n s . Cross-dialectal data f o r what are the / i m / and / i n / f i n a l s i n Zhong-shan p a r a l l e l ' the above observations, r e f l e c t e d i n the h i s t o r i c a l reconstructions (e.g., LMC *iam and - 43 - *ian). Similar c r o s s - d i a l e c t a l and h i s t o r i c a l observations as above can be made of f i n a l s which correspond to Zhong-shan / y t / and /ut/ f i n a l s i n which a schwa o f f - g l i d e i s only heard sporadic- a l l y . Usually, the two f i n a l s / y t / and /ut/ are pronounced l y i t j and [u:t], without a perceptible g l i d i n g e f f e c t . Again, i t was Chao (p.53) who f i r s t made the observation of a schwa o f f - g l i d e , while Egerod simply recorded a long l y i ] and a long [ u i j , with no discussion of allophonic v a r i a t i o n s . The phoneme / i / i s pronounced [i] before velars. (The preferred IPA symbol for [i] i s [i].) Chao also analyzes t h i s vowel as / i / , which he describes as being quite open before /o/ and /k/, almost becoming an [ e ] , which i s the broad phonetic form he uses for the vowel. Egerod also transcribes / i / as [e] i n the same environment. This vowel i s i d e n t i c a l to the one found i n Cantonese, usually transcribed as either Ie] or [ 1 ] . A further point must also be added concerning the special complementary d i s t r i b u t i o n found i n the s y l l a b l e /ing/. In a l l the tones except the r i s i n g one, the s y l l a b l e has a smooth onset, and begins simply with the vowel [ 1 ] , as i n the word ying jfe 55 'distinguished 1, which i s phonemically /ing / and i s pronounced 55 . ]. In contrast, the s y l l a b l e i n the r i s i n g tone begins with the front on-glide [ j ] , as i n the words ying ^ 'shadow' and ying 'to r e f l e c t ' , both of which are /ing / phonemically, but are 13 i n fact pronounced [jig ], with an i n i t i a l p a l a t a l on-glide. The reason for t h i s phenomenon i s not clear to the writer. Pulley- blank (personal communication) speculates that i t may be the r e s u l t of the laxness of {1] combined with some c h a r a c t e r i s t i c of the r i s i n g tone, since a l l other f i n a l s with / i / i n the r i s i n g - 44 - tone have the tense, long [ i x ] , and are produced with a smooth onset before the nuclear vowel; for example, y i 'chair' i s / i 1 3 / , phonetically [ i x 1 ^ ] , and yan 'to conceal* i s /im 1^/, 13 phonetically lixm J . The phoneme /y/ i s pronounced somewhat more open than cardinal ly] i n Chen's speech, which agrees with Chao's observa- ti o n of the vowel. In Zhong-shan th i s vowel i s always long, occurring as a f i n a l either alone or followed by a dental. How- ever, r e c a l l that the f i n a l / y t / i s sometimes pronounced [y t ] , with an o f f - g l i d e following the vowel. The phoneme /u/ i s pronounced [uxj i n s y l l a b l e - f i n a l p o s ition and when i t occurs before / j / and the dentals, /n/ and / t / . The vowel i s actually s l i g h t l y more open than cardinal [uj, a point also made by Chao (p. 53). . As mentioned e a r l i e r , the f i n a l /ut/ i s sometimes pronounced [u t j . In the environment before the velar endings, /ng/ and /k/, the phoneme /u/ i s somewhat more open than lu] (= luj or the preferred IPA symbol ]>J), which i s i d e n t i c a l to i t s Cantonese counterpart, variously transcribed as I.oJ . IuJ and lu] i n Cantonese. In both Egerod's and the present study of Zhong-shan, what i s transcribed here as the high, lax vowel [u] i s analyzed as the allophone of /u/, i n complementary d i s t r i b u t i o n with the corresponding tense, long vowel [ux]. Chao, however, treats our - [u] as /o/, moreover regarding the analysis of th i s vowel as /u/ in Cantonese being due to the influence of English and German, wherein the l e t t e r "u" i n the English word "put", for example, i s , according to Chao, extremely open and quite close to cardinal [o] (p.53, fn.9). Whether or not Chao's surmise of influence from - 45 - English and German i s true, the analysis of the high, lax, back rounded vowel [u] as /u/ i n the f i n a l s provides congruity to the pattern established e a r l i e r i n the analysis of the high, lax, front unrounded vowel [ i ] . as / i / i n the f i n a l s [ i n ] and [ i k ] (see Chart 3 (a)). Furthermore, the f i n a l s [ug ] and [ukj correspond cross- d i a l e c t a l l y and h i s t o r i c a l l y to [ i n ] and [ i k ] , and should be s i m i l a r l y analyzed. The symmetry between tre a t i n g lax [ i ] as an allophone of / i / and lax [u] as an allophone of /u/ can be b r i e f l y elaborated as follows. H i s t o r i c a l l y , the Late Middle Chinese f i n a l s *-im, * - i p , * - i n , * - i t y * - i j ( < * - i a j ) and *-iw gave the lax f i n a l s -am, -ap/ - s i n ; -at/ -aj # and -aw, now [em], [ep ] and so fort h i n modern Zhong-shan and Cantonese. Such a development did not occur before *-q (and the corresponding Ru-sheng ending, * - k ) . LMC f i n a l * - i n resulted i n -en (now [ i n ] (or [eg] i n alternate transcriptions) i n the two Yue d i a l e c t s ) . As a comparison, LMC f i n a l * - i g gave Mandarin - i n . In the case of the Late Middle Chinese f i n a l *-og (*-ok i n Ru-sheng), i t remains -og i n Zhong-shan and Cantonese, now phonetically [ug] (or [og] i n alternate t r a n s c r i p t i o n s ) . In Mandarin, LMC f i n a l *-og y i e l d s -ug. Thus, the symmetry of -eg and -og (and -ek and -ok) j u s t i f i e s p a r a l l e l treatment of [ i ] and [T J ] on h i s t o r i c a l grounds. The same -eg~ - o g p a r a l l e l l i s m i s observed i n l i t e r a r y Min and appears to be a general southern development. In the north, the p a r a l l e l l i s m between the LMC f i n a l s * - i q and *-og i s observed i n the Mandarin reflexes of - i q and -ug corresponding to the two h i s t o r i c a l f i n a l s . Thus, the tense and lax d i s t i n c t i o n , i n conjunction with h i s t o r i c a l and cross- - 46 - d i a l e c t a l arguments supports a phonemic analysis which recognizes the p a r a l l e l l i s m or symmetry that exists between the vowels [ i l . 12 and l u j , and treats them accordingly. Turning now to the mid vowels, phonemically. there are two mid vowels: /</>/ and /o/. Regarding the f i r s t phoneme, /</>/ i s pronounced [ce:] as a s y l l a b l e - f i n a l segment. There are very few words i n Zhong-shan and Cantonese with the f i n a l Ice:] , and these may be vestiges of an e a r l i e r layer of the language as suggested by Pulleyblank (personal communication). Chao (p.53) chooses phonemic symbol /oe / for Ice:] when i t occurs i n s y l l a b l e - f i n a l position. Before the endings /ng/ and /k/, Chao notes that /ce / tends to break into a [ 0 0 ] c l u s t e r , which i s treated phone- mically as / i o / i n the present analysis. (We w i l l return to t h i s point later.) Egerod posits the vowel c l u s t e r /0a/, which, he transcribes phonetically as [ceo] , occurring with an i n i t i a l con- sonant only, or before /q/ (= /ng/) and /k/. In the present phonemic analysis, /0/ i s recorded as [cei] s y l l a b l e - f i n a l l y , and elsewhere ( i . e . , before the endings /n/, / t / and /j/) as a rounded vowel which i s s l i g h t l y more open and centralized than [ 0 ] . The vowel [ 0 ] i s recorded as the broad phonetic form i n Table 2 (.a). Chao (p. 52) describes the same vowel as a rounded, mid central [el, which i s the same as i t s Cantonese counterpart. Chao analyzes his Ie] as /o/ phonemically, adding an umlaut above the phoneme ( i . e . , /6/) i n order to f a c i l i - tate recognition and r e c a l l of i t s actual phonetic value. More- over, Chao (p.52,fn.8) makes the comment that t h i s vowel i s trans- cribed by Daniel Jones as [ce] , which i s the same phonetic form suggested by Kinkade for the vowel i n question. Egerod (p.14) also - 47 - hears the vowel Ice] i n his Zhong-shan data, analyzing i t phonemic- a l l y as On a preceding page however, Egerod (p.12), i n describing the phonological system of standard Cantonese, records the phoneme /0/ which he states i s I0] before /n/, / t / and / j / . Recent studies on Cantonese (.e.g., Kao, 1971; Hashimoto, 19.72a; Cheung, 19.72) have v a c i l l a t e d between selecting the form Ice] or [0] for transcribing the vowel. The phoneme /of i s pronounced [D:J except i n the environ- ment before the l a b i a l glide /w/ where i t i s pronounced somewhat lower than cardinal l o ] . Whereas i n the present analysis length i s non-contrastive except i n the low vowels, Chao treats short .'[.o.J as /o/ phonemically, and long ID:.] as /o:/, thereby i m p l i c i t l y considering length phonemic i n Zhong-shan. He i s not consistent, however, since he also subsumes his mid central vowel Ie] under the phoneme /o/, and provides no corresponding long phoneme to his / C B/, which only has the long Ice:] as i t s basic phonetic shape. It i s observed that the vowel ID:] sometimes s p l i t s into loa] (in which the i n i t i a l segment of the b i v o c a l i c nucleus i s s l i g h t l y lower than cardinal ID] and the second somewhat more centralized than cardinal [a.]}. Chao (p.53) describes the break- ing of the vowel ID:] (his phoneme /o:/) as [OD] , p a r t i c u l a r l y notable aft e r velars and laryngeals. Egerod (p.14). sets up a 13 vowel c l u s t e r /ua/, with. I DO] as the p r i n c i p a l phonetic form, presumably occurring aft e r velars, laryngeals and / f / , since i t s allophone I D:], which i s enclosed i n round brackets, occurs most commonly after dentals, s i b i l a n t s and l a b i a l s with the exception of / f / . Concerning the allophone [oil, Egerod states that i f i t i s considered a phoneme, i t should be written /o/. This would - 48 - p a r a l l e l his treatment of the vowel I ox,] as /o/ on the preceding page for Standard Cantonese (p.12). It should be observed that i n Cantonese there i s a contrast between IkJ and [kw] (and between the aspirated series, I k ' J and Ik'wjX before the vowel I D : ] , namely a f f e c t i n g the f i n a l s [ o i l , I D :..rjJ and [31k]. Such a contrast i s not found i n Zhong-shan. Consider, for example, the Cantonese pronunciation of the word ge \fy ' ( . c l a s s i f i e r ) ' , which i s {k 3: 4 4.]. 1 4,. and that of guo j$l 4 4 'to cross', which i s Ikwsx *].. The pronunciation of these two words i s merged i n Zhong-shan such that both ge and guo are pro- 22 nounced Iks': J . The contrast found i n Cantonese between the two sample words i s r e f l e c t e d h i s t o r i c a l l y i n the d i s t i n c t i o n between Kai-kou V2 (.'open mouth', associated with the absence of l i p - rounding), and He-kou ^ & (.'closed mouth', associated with the presence of lip-rounding), reconstructed by Pulleyblank as LMC *ka and *kwct respectively. As noted above, Chao (p. 52) describes the breaking of the vowel [01] i n Zhong-shan into the sequence [osj a f t e r velars and laryngeals. While one can say that Zhong- shan does, not have a Kai-He d i s t i n c t i o n , Chao suggests that one might equally take the vowel i n Zhong-shan as o r i g i n a l l y between Kai and He. Egerod's treatment of [ 3 : ] , and even more so that of his c l u s t e r [3t>] as /ua/ rather than /o/, as i n our analysis, or /o:/, as i n Chao's, i s obviously an attempt, however i m p l i c i t , to r e f l e c t the He-kou category. There are two low vowels i n the data, contrasting i n length and tenseness. Short /a/ i s phonetically a very lax vowel, des- cribed by Chao (p.52) as one raised towards [e], except before /w/, where i t i s a low, central vowel for which he adopts the symbol - 49 - [ A ] . In our broad phonetic t r a n s c r i p t i o n i n Table 2 (a), the vowel [BJ i s used throughout the set of /a/ f i n a l s . Kinkade per- ceives short /a/ i n the data as simply the low, back vowel [a], although he observes that the vowel occasionally s h i f t s to midway between Is] and lae] . Egerod simply records a. .low central .[&.•] with- out any comment concerning allophonic v a r i a t i o n s . This vowel i s i d e n t i c a l to the corresponding short, low vowel i n Cantonese. Long /a:/ i s a low, central vowel, transcribed here using the symbol [A:]. Chao's phoneme /a:/ i s also [A:] i n Zhong-shan. The same vowel i s also found i n Cantonese. Hashimoto, for instance, transcribes the long, low vowel i n Cantonese as [A : ] . Egerod . records his long, low central vowel as [a:], which he treats phonemically as the vowel c l u s t e r /aa/. Chao (p.52) notes that before stop endings, the difference i n vowel length, between the two low vowels /a/ and /a:/ i s not sharp; for the most part, i t i s the vowel quality which d i f f e r e n t i a t e s them. Egerod (p.14) makes the observation that i n his s y l l a b l e /jaa/, the phonetic form [ja:.J and Ijaei] occur i n free v a r i a t i o n . This i s not noticed i n the speech of the informants used i n the present study wherein only the low, central [AI] i s found. Note i n Table 2 (a) that a l l the vowels i n s y l l a b l e - f i n a l position are long i n Zhong-shan. While there i s a phonemic con- t r a s t between /a/ and /a:/ i n non-final position, such a contrast does not e x i s t i n s y l l a b l e - f i n a l p o s i t ion. Thus, to mark length i n the l a t t e r case would be redundant, and i s therefore omitted in such an environment (e.g., ma ^ 'mother' i s /ma55/, phonetic- 55 a l l y [JIA i ] . In the discussion of nuclear vowels i n th i s section, no - 50 - account i s taken of the influence which the medials may have on them. The following section on medials w i l l include allophonic variations of the nuclear vowels r e s u l t i n g from assimilation of certai n features of the medials, and conversely, of the e f f e c t of the nuclear vowels on the medials. 1.2.2. Medials Two medials are proposed f o r the Zhong-shan data, namely / i / and /w/. Although i t might have been more symmetrical to posit either a pair of medial vowels or a pair of medial gl i d e s , the choice of the vowel / i / and the semi-vowel /w/ i s based on synchronic data, supported by c r o s s - d i a l e c t a l and h i s t o r i c a l e v i - dence. Synchronically, for example, vocalic medial / i / i s a vowel i n Zhong-shan, phonetically [e] before /a/, and [0J before /o/. The glide /w/ i s i n fact the semi-vowel [w] i n Zhong-shan. Thus, phonetically, the two medials i n Zhong-shan are not symme- t r i c a l with respect to vowel qua l i t y . Diachronically, there are also good reasons for di s t i n g u i s h - ing the v o c a l i c nature of the two Zhong-shan medials. Medial / i / i n Zhong-shan i s derived from the LMC medial * i , while medial /w/ has i t s source i n medial *w. Thus, h i s t o r i c a l l y , the two medials also d i f f e r i n v o c a l i c q u a l i t y . Pulleyblank (1970-71), for example, analyzes the four grades i n h i s t o r i c a l Chinese phonology i n terms of differences found i n the medials, since Chinese rhyming practices do not take either the Kai-He d i s t i n c t i o n s or the various grades into consi- deration. Pulleyblank reconstructs both medial vowels and medial gl i d e s . Evidence for such a d i s t i n c t i o n i s found i n Cantonese and Zhong-shan. The word guan '|p (134-1) ' governraent o f f i c i a l ' , 55 for example, i s LMC *kuan, Zhong-shan [ku:n ] and Cantonese 5 3 [kuin ], wherein the medial i n Grade I of the Shan LU rhyme group (He-kou series) i s reconstructed as the vowel *u. In modern Zhong-shan and Cantonese, the nuclear vowel *a was l o s t , r e s u l t - ing i n medial *u becoming the new nuclear vowel which was length- ened to compensate for the l o s t segment. In contrast, the word guan (138-7) 'to shut', which i s found i n Grade II of the same rhyme group and reconstructed as 55 53 LMC *kwam, i s [kwAin ] i n Zhong-shan and [kwAin ] i n Cantonese. This time the reconstructed medial i s a g l i d e , *w, which i s pre- served as a glide medial i n both present day Zhong-shan and Can- tonese. In the modern reflexes of many di a l e c t s of Chinese, the two reconstructed forms, *kuan and *kwa:n, have merged (e.g., both, words 'government o f f i c i a l ' and 'to shut' are pronounced guan i n Mandarin). In the Yue d i a l e c t s such as Zhong-shan and Cantonese, however, one can see that the d i s t i n c t i o n between the two forms has been maintained. On the basis of Zhong-shan and Cantonese, i t i s evident that the reconstruction of a medial vowel versus a medial gl i d e i s important i n h i s t o r i c a l phonology. Karlgren also distinguishes a "vocalic u" as opposed to a "consonantal w" i n the two words discussed above, reconstructing guan '̂f as Anc: *kuan and guan ĵ£J as Anc. *kwan. (Note that Karlgren also recon- structs a q u a l i t a t i v e l y d i f f e r e n t nuclear vowel.) In the Zhong- shan data, i t i s therefore reasonable to posit a medial /w/ which 15 arises from an h i s t o r i c a l medial *w. Medial / i / i n Zhong-shan, i n contrast, has i t s source i n an h i s t o r i c a l * i . On the basis of Pulleyblank'. s reconstructions, - 52 - while medial * j i s l o s t i n Zhong-shan, medial * i i s s t i l l present i n the d i a l e c t . It i s maintained as the medial vowel / i / i n the c o l l o q u i a l layer of Grades III and IV words i n the Geng rhyme group. In the corresponding l i t e r a r y layer, medial * i had become the nuclear vowel / i / i n Zhong-shan. Consider the word j i n g t (199-6) 'mirror', which occurs i n Grade II I , Geng )$L rhyme group. Jing /4JfLi i s reconstructed by Pulleyblank as LMC *kiajn (..= Anc. 22 *kiang). I t i s /kia:ng / i n the c o l l o q u i a l layer of Zhong-shan and / k i n g 2 2 / i n the l i t e r a r y layer. Now, contrast j i n g /jjftj with geng ^ (192-8) 'watches of 55 the night', which i s /ka:ng / i n the c o l l o q u i a l reading, and 55 /kang / i n the l i t e r a r y layer. No medial i s present i n the Zhong-shan forms for geng , which i s a Grade II word. Pulley= blank reconstructs i t with a medial * j : LMC * k j a i j n (.= Anc. *kong) . Thus, i n terms of Pulleyblank.'s reconstructions, Late Middle Chinese medial * j . i s l o s t i n Zhong-shan, as demonstrated by the word geng . In contrast, medial * i i s preserved i n the d i a l e c t either as a medial vowel or nuclear vowel, as shown i n j i n g . The preservation of the LMC glide medial *w as a medial /w/ i n Zhong-shan, and the p a r a l l e l preservation of a v o c a l i c medial i n Zhong-shan corresponding to the LMC medial * i lends h i s t o r i c a l support for not pos i t i n g the same vo c a l i c q u a l i t y to the pa i r of medials i n Zhong-shan. Cr o s s - d i a l e c t a l l y , medial * i i s also preserved as a medial segment i n some d i a l e c t s . Regarding the word qing ^ £ (.201-1) 44 ' l i g h t (e.g., i n weight)', for example, i t i s pronounced [tcsliai) .] i n Wen-zhou (a Wu d i a l e c t ) . In Nan-chang (a Gan d i a l e c t ) , the 42 same word has a c o l l o q u i a l reading [te'ian ] and a l i t e r a r y read- - 53 - ing [to.'in ] . In Mei-xian (a Hakka dialect) , qing has a 44 44 c o l l o q u i a l form [ k i a q ] and a l i t e r a r y form I k i n J . (Zi-hui, p.257,) The medial / i / i n the f i n a l s /iong/ and /iok/ i n Zhong- shan i s also derived from medial * i i n Late Middle Chinese. The two f i n a l s are reconstructed by Pulleyblank as *ian and * i a k respectively, (= Anc. *i,ang and * i a k ).. The medial i s also pre- served i n most Chinese d i a l e c t s . The word lia n g ftj (174-18) 'bright' (cf. Zi-hui, p.232), for example, has the s y l l a b l e I l i a y J i n (Peking) Mandarin and a number of other Mandarin d i a - l e c t s , Su-zhou (Wu), Chao-zhou (Southern Min), and the l i t e r a r y layer of Amoy (Southern Min). The s y l l a b l e [ l i a n ] i s found i n Nan-chang (Gan), Mei-xian, and the c o l l o q u i a l layer of Amoy. >• 22 Zhong-shan pronounces the word lia n g faj as [loon ] , which i s 22 phonemicized as / l i o n g /. Pulleyblank (1977) proposes that i n Cantonese c e r t a i n LMC f i n a l s (viz., Grade I f i n a l s i n the Dang % rhyme group), f i n a l s *-ag and *-tik rounded to -og and -ok, just as *-a rounded to -o (in the Guo ^ rhyme group). He fur- ther.postulates that i n Grades III and IV (of the Dang rhyme group), * - i a g and *-iak also rounded to - i o g and-iok as an i n t e r - mediary stage, aft e r which the p a l a t a l medial became fused with the following vowel by a process of umlaut, r e s u l t i n g i n the f i n a l s -cag and -ask. Zhong-shan data would suggest the following development: the vowel *a i n the LMC f i n a l s *-Q, *-aq and *-ak rounded to :D, The same process affected the *-d i n the f i n a l s * - l a g and *'-.iak. In the l a t t e r case with medial * i , the medial underwent a lower- ing and rounding to -0 as a r e s u l t of the following back, rounded - 54 - mid vowel -o. The result is the present Zhong-shan finals, [0oql and l0o.k] . Cantonese, on the other hand, may be a case of the lower- ing and rounding of medial *i to -ce. The nuclear vowel is subse- quently lost, causing compensatory lengthening of -ce to - c e : , yielding present day Cantonese finals IceiijJ. and {cerkj . Liang rftj , for example, is pronounced I ice i n J in Cantonese. This lowering of medial *i in the Dang ^ rhyme group in Cantonese parallels the lowering of medial *i to -e in the Geng rhyme group. What is proposed here is that in Cantonese, medial *i lowered to -e in the finals *-iajn and. *r-iajk. In this case, rounding of the medial does not occur since there is no condition for such regressive assimilation. However, as in the previous case, the nuclear vowel becomes lost, and -e becomes the nuclear vowel. It is lengthened in compensation for the loss of the original nuclear vowel *-a. The finals that emerge in Cantonese are -e:g and -eik- Thus, the colloquial reading of jing^Jb is 44 . • [keiq ] in Cantonese. The nuclear vowel *-a is not lost in Zhong-shan; jing is pronounced [keaxn J (phonemically , . 22 /kiarng / , as noted earlier). Turning now to a synchronic analysis of the interaction between the medials and other segments in the syllable, medial /w/ will be discussed first. In the present data, the only two initials that occur with medial /w/ are /k/ and /k'/. Egerod and Chao, however, also include the velar nasal as another initial which can precede the labial medial. The cluster /ngw/ is merely mentioned by Egerod. In Chao's article it occurs only in the word 51 wan .jf.fi . (137-U4) 'obstinate', phonemicized by as /ngua:n / . - 56 - l a b i a l segment i n Shi-qi depended on the i n d i v i d u a l speaker, although Macao shows a.complete loss of /w/ after velars. Addi- t i o n a l s y l l a b l e s containing a l a b i a l velar cluster recorded by B a l l for Shi-qi i s l i s t e d i n (2) below. B a l l states that these words are spoken by some people i n S h i - q i , and attributes that to the imitation of Cantonese speakers. The asterisk (.*) marks the column ...... which B a l l regards to be the pronunciation of some Shi- q i speakers. Regarding the c h a r a c t e r ^ " , i t i s used i n Cantonese to mean 'to be t i r e d ' , and i s a word which i s not used c o l l o q u i a l l y 22 i n Mandarin. Zhong-shan has i t s own c o l l o q u i a l word /na:j /, with, b a s i c a l l y the same meaning of 'to be t i r e d ' . (In the l i s t below, "Cant." i s "Cantonese", and "Zh." i s "Zhong-shan".) (2) B i l l , ' Cant.* s Macao . Sample Word Modern Cant. Modern Zh. Gloss 1. kwe ke '%, ? ? ? 2. kwing kang fa 35 [ k w 11) J •p 'to b o l t ' 3. kwit k i t *&'\ ? ? ? 4. kwo ko * r 5 3 i [wo:. ' j [koi J 'spear' 5. kwok kok i i Ikwo z.k̂ ] Ik a i k 2 J 'nation' 6. kwong kong to IkwDiq J Ikoxn J 'bright' 7. k'wong k' ong [k ' o i n 2 1 ] Ik'oxn 5 1] 'crazy' 8. kwGii koii [ k u xj 3 3] I k u i j 2 2 ] 'to be t i r e d ' 9. k' woii k' ou I k ' u i j 3 5 ] Ik'uxj ] 'middleman' 10. kwu ku [kux.53] [kux 5 5] 'father's s i s t e r ' 11. k' wu k'u ** Ikux 5 3] ? 'wheel' 12. kwun kun f Ikui.n 5 3] I k u i n 5 5 ] 'government o f f i c i a l ' - 55 - Wan fify^ occurs i n the He-kou series of the Shan ih rhyme group. Apparently, i t i s the only common word from that set where the s y l l a b l e /ngwa:n/, or the potential s y l l a b l e /ngwa:k/ can occur. In consulting the p r i n c i p a l informants as well as another Zhong- shan speaker concerning the word wan 7 ^ , i t was found that none of them preserve a l a b i a l segment i n the word: they simply gave /nga:n 5 1/. For the present data at least, i t w i l l be concluded that the only i n i t i a l s which can occur with medial /w/ are /k/ and /k'/. The sequence /ngw/ i s assumed to have merged with the plain / v e l a r nasal. In other words, the Kai-He d i s t i n c t i o n has been l o s t a f t e r the velar nasal i n the Shan rhyme group. The word wan ffy i s i n fact the only word present i n the d i a l e c t survey l i s t which can be reconstructed as LMC *nwa.in. In contrast, there are a few common words with the s y l l a b l e /nga:n/ from LMC *D.ja:n and a small number of indigenous (characterless) words i n Zhong-shan also containing the s y l l a b l e /nga:n/. I t i s therefore not surprising to f i n d that the word wan has l o s t i t s l a b i a l segment, and has merged with, the more common s y l l a b l e without l a b i a l i z a t i o n . In Cantonese (as i n Mandarin), the word wan has l o s t i t s velar nasal i n i t i a l , but has maintained the l a b i a l segment, and i s pronounced r  2 1 i 1WA in J- Concerning the clusters /kw/ and /k'w/, these combine only with the nuclear vowels / i / (found solely i n the rhyme / i k / ) , /a/ and /a:/. Gaps i n the f i n a l such as the non-occurrence of /*wa:m, *wam, *wap/ and so forth are the r e s u l t of l a b i a l d i s s i m i - l a t i o n , which i s a phenomenon found i n many di a l e c t s of Chinese. According to B a l l ' s data, the loss or preservation of the - 57 - The l i s t i n (2) prompts a low^level phonetic observation that might otherwise have been l e f t unstated. In modern (Shi-qi). Zhong-shan and Cantonese, some lip-rounding occurs i n the pro- 17 nunciation of i n i t i a l s before the rounded vowel _ L u z J . I t i s possible that t h i s feature i s absent i n Macao, as r e f l e c t e d i n B a l l ' s recording of numbers (.10) to (12) i n the l i s t above. A si m i l a r lack of lip-rounding on the i n i t i a l s may be interpreted from B a l l ' s t r a n s c r i p t i o n of the Macao forms i n (8) and (.9). Numbers (4). to (.7) show a presence of l a b i a l i z a t i o n which i s not eyident i n the Zhong-shan speech recorded by Chao, Egerod and the writer, although i t i s s t i l l preserved i n Cantonese. That i s , there i s a phonemic contrast i n Cantonese between /k/ and /kw/ and between /k.'/ and /k'w/ before the vowel I D Z J , as we have men- tioned e a r l i e r , exemplified by the words g_e and guohjffi^ . Egerod (.p. 3) makes the observation that although the d i a l e c t spoken i n Macao has now s h i f t e d to the standard Yue dia- l e c t , Cantonese, there are s t i l l a few vestiges of the l o c a l d i a - l e c t , among which i s the fusion of /k/ and /kw/. In studying the Sino-Portuguese glossary of c i r a 1750, nevertheless, i t can be strongly substantiated that although the medial /w/ was l o s t a f t e r /k/ when followed by the mid back vowel which i s now ID:.] , the 'li.-. d i s t i n c t i o n between the presence or absence of a l a b i a l a f t e r the velar stop was s t i l l maintained before low vowels at the time that the three hundred and ninety-five Portuguese entries were trans- cribed into Chinese. While the characters ge , ge gffi^ and ge (Cantonese [ k o i ] ) , were used most frequently to represent Portuguese /ko/, the characters guo ^ and guo jyffi^. (.Cantonese [kwDz.]) were nonetheless used for the same purpose, thereby - 58 - suggesting a loss of d i s t i n c t i o n between /k/ and /kw/, at l e a s t i n that p a r t i c u l a r environment. As i n the (Shi-qi). Zhong-shan d i a l e c t today, however, the Sino-Portuguese data show that t h i s d i s t i n c t i o n was preserved before low vowels at that time, r e f l e c t - ing either the speech of the transcriber only, or the eighteenth century Macao d i a l e c t i n general. In the glossary, the characters j i a , j i a and j i a were consistently used to transcribe Portuguese words which do not contain l a b i a l segments. A l l three characters belong to the Kai-kou series of the J i a 4 £ s L rhyme group and are pronounced IkAiJ i n both present day Cantonese and (Shi-qi). Zhong-shan. I t i s therefore s i g n i f i c a n t that gua CtkwAiJ i n both Cantonese and Shi-qi Zhong-shan), which i s the only He-kou word from the J i a rhyme group, should be used to transcribe the word quatro 'four' (entry number 328 i n the glossary). In turn, quatro i s the only Portuguese word reconstructed i n the glossary that contains a velar i n i t i a l followed by a l a b i a l segment. Assuming that the transcriber's speech r e f l e c t s the n: " at the time, one can quite safely conclude from the observation above that, with, respect to low vowels, a Kai-He d i s t i n c t i o n was s t i l l preserved i n the eighteenth century Macao data. Turning now to a''.discussion of medial / i / , the medial com- bines with one of two possible nuclear vowels: /a:/ or /o/. When i t combines with /a:/, medial / i / may follow l a b i a l s (except / f / and /w/), dentals, s i b i l a n t s , velar stops and laryngeal /h/. Re-interpreting Chao's data from th i s perspective, his medial / i / i s phonetically somewhat more open than the vowel {ij (cf. fn.16 of the present chapter). Egerod and the writer transcribe t h i s sound as l e j . At the same time, medial / i / i s sometimes heard as - 59 - a p a l a t a l [j] . Thus, the word ji n g 'neck', for example, may 13 13 Be pronounced Tkeaig J or [kjain J i n free v a r i a t i o n . I t i s only i n the environment a f t e r laryngeal /h/ that p a l a t a l i z a t i o n of the medial does not occur. As suggested by Pulleyblank upon observing t h i s phenomenon i n the data, i t may be a case of the d i a l e c t attempting to produce a p a l a t a l medial to correspond to the l a b i a l one, thereby creater greater symmetry i n the system. The f a c t that medial / i / i s never pronounced as a p a l a t a l g l i d e when i t follows /h/ i s probably due to the e f f o r t s of the speakers not to p a l a t a l i z e /h/, since the r e s u l t of p a l a t a l i z i n g /h/ would be perceived by the native ear as being quite aberrant from the o r i g i n a l laryngeal i n i t i a l . In general, Yang tends to produce more p a l a t a l [j] medials than Chen. At the same time, i t i s also Yang who pronounces the medial / i / vowel higher than Chen. The more open pronunciation of the vowel by Chen may be an influence from Cantonese i n which the Zhong-shan sequence I ear], o r i g i n a t i n g from the sequence ,vi,v V. "medial / i / + nuclear vowel /a:/", corresponds to the Cantonese long (nuc1earI vowel I e : ] . In the environment before /o/, medial / i / i s lowered and rounded as a r e s u l t of regressive assimilation. In t h i s environ- ment, medial / i / i s pronounced [0]. The sequence / i o / only occurs i n the f i n a l s /iong/ and /iok/, phonetically [00^] and [00k] respectively. These two f i n a l s are treated by Chao as /aeng/.and /aek/. Note that Egerod also treats the f i n a l [OBI.] as the vowel clus t e r /0a/. The writer analyzes t h i s f i n a l as phonemic /0/ rather than the c l u s t e r / i o / . Chao treats i t phonemically as /CE/. D i s t r i b u t i o n a l l y , the f i n a l s /iong/ and /iok/ occur with - 60 - dentals, s i b i l a n t s and velar stops. I n i t i a l /h/ combines only with, the /iong/ f i n a l . Present i n both. Chao and Egerod' s data are the s y l l a b l e s lj0oql and I j 0 D k J , which correspond to [jaxgj and IjDikJ i n the present data. B a l l (p.531) records the s y l l a b l e /yong/, examplified by the word yang ^ (193-5) 'ocean1, for both Cantonese and Macao, but records /yong/ for Shi-qi (fn.120). Presumably, the corresponding stop ending r e f l e c t s the same pattern, although i t was not made e x p l i c i t i n B a l l . The present Zhong-shan data and modern Cantonese concur with B a l l ' s observa- t i o n : Zhong-shan has the s y l l a b l e s Ijonj] and [joik] while Canton- ese has [uosir)] and [qcexk] . Based on B a l l ' s observations and the writer's data, one could suggest that s y l l a b l e s [jong] and [joik] are the indigenous forms i n the Shi-qi speech, and that the s y l l a b l e s [j0orjj and [J03kJ recorded by Chao and Egerod are due to Cantonese influence. The writer, however, would prefer to suggest that the difference observed between the present data and those of Chao and Egerod i s s t r i c t l y a case of a sub-dialectal v a r i a t i o n . F i r s t of a l l , Chao (p.51) e x p l i c i t l y describes the i n i t i a l segment of his s y l l a b l e /icEng/ phonetically as the unrounded front g l i d e [ j ] . Thus Chao's s y l l a b l e appears to be mid-way between the s y l l a b l e found i n the present Zhong-shan data and that i n Cantonese. Phonetically, Chao's s y l l a b l e i s [J0oq], whereas the Cantonese s y l l a b l e i s [xifflin ] . Secondly, except for t h i s s y l l a b l e Ij0onj, and the s y l l a - ble with the corresponding stop ending, the phonological system recorded i n Chao and Egerod i s i d e n t i c a l to the present Zhong-shan one. There i s no reason to expect such a unique borrowing. - 61 - Third and l a s t l y , there are a few c o l l o q u i a l terms i n Chao's a r t i c l e that show s l i g h t phonological variations between his informant's speech and the speech of the writer's informants. In the d e i c t i c word ' t h i s ' , for example, Chao records the variant 22 55 55 forms of /ko / and /ko / whereas the present data show /ku /. (On one occasion, nonetheless, the writer heard Chen's mother use 55 /ko / for 'this'.) In the reading of is o l a t e d words, Chao's informant also pronounces some words with an i n i t i a l or f i n a l that d i f f e r s from the pronunciation of the writer's informants. The word ku ^ (.16-24) ' b i t t e r ' , for instance, i s recorded by Chao 13 13 as /k'u / and by the writer as /hu /. In terms of the o v e r a l l pattern of correspondences to the h i s t o r i c a l phonological catego- r i e s , however, the data from Chao, Egerod and the writer are i n agreement. In summary, there i s evidence of low-level, sub-dialectal variations that, i n general, do not a f f e c t either the Zhong-shan phonological system on the whole, or the d i a l e c t ' s h i s t o r i c a l correspondences. One can see no reason to suppose that there has been an is o l a t e d case of substituting a Cantonese s y l l a b l e for a Zhong-shan one. The conclusion, therefore, i s that among some Zhong-shan speakers, the s y l l a b l e [joon] i s used, whereas others use [ j o i n ] . One would suspect, given Chao's observation of an ... i n i t i a l IjJ that the following segment i s probably s l i g h t l y less rounded than cardinal J>]. Cr o s s - d i a l e c t a l l y , the s y l l a b l e [jog] i s also found i n other d i a l e c t s . The word yang p| (173-8) 'bright' (Zi-hui, p. 236) has the s y l l a b l e [jon] (or [iaq]) i n Nan-chang (Gan), Mei- xian (Hakka), and Amoy (S. Min). The Mandarin d i a l e c t s generally - 62 - pronounce the word as Jax.n'l (transcribed i n the Zi-hui as {iag]). Of the d i a l e c t s represented i n the Zi-hui, only Cantonese has the s y l l a b l e [ u 3 i >j] (.[jcenj i n the Zi-hui) . Fu-zhou (_N. Min) has the s y l l a b l e fyon] for yang ^ . Both yang 1 s ^ and f-|j are recon- structed as LMC * j i a g . 1.2.3. Endings The inventory of endings present i n Zhong-shan i s i d e n t i - c a l to that found i n Cantonese. There are three nasals: /m/, /n/ and /ng/; three corresponding stops: /p/, / t / , /k/; and two g l i d e s : /w/ and / j / . The stops are unreleased, and often p r e g l o t t a l i z e d i n Zhong-shan. The ending /k/ i s moreover often simply reduced to a g l o t t a l stop. Syllables with stop endings usually occur i n l e v e l tones only. Exceptions r e s u l t from changed tones, to be elaborated upon l a t e r . Chao notes that a f t e r long nuclear vowels ( i . e . , i n the endings IA I W ] and [i:w]), /w/ i s more open.than a f t e r short nuclear vowels (as i n the endings [ew] and [ow]). The glide / j / i s phonetically [j] with one exception: following the front rounded vowel / 0 / , / j / i s phonetically the rounded semi-vowel [q] . Thus, the word r u i ,gjp^ (61-15) 'stamen' 13 13 / J 0 j /r for example, i s phonetically [^0^ J, i n which both the preceding and the following semi-vowel segments are rounded to {uj as a r e s u l t of assimilation to the rounding of the nuclear vowel. - 63 - 1.2.4. S y l l a b i c Nasals There are two s y l l a b i c nasals i n Zhong-shan which c o n s t i - tute the only segment i n the s y l l a b l e , and a f f e c t a very r e s t r i c t e d lexicon. These two segments are /m/ and /ng/, pronounced ImJ and [gj respectively. The s y l l a b l e /m/ only occurs c o l l o q u i a l l y , as * 51 ' the negative marker B "gr /m / 'not' (also found i n Cantonese). Words with the s y l l a b l e /ng/ include: wu /ng"*"3/ ' f i v e ' , wu ^ / n g 5 1 / 'Wu (a surname; a Chinese d i a l e c t group to which the Shang- i < 22 hai d i a l e c t belongs) ', and wu •fQ' /ng / 'to r e a l i z e ' . The s y l l a - ble /ng/ occurs i n a l l but the high l e v e l tone. Words possessing s y l l a b i c /ng/ h i s t o r i c a l l y bore a velar nasal i n i t i a l , and are reconstructed as LMC *nua. Some of the d i a l e c t s s t i l l preserve the velar nasal i n i t i a l , as exemplified by the pronunciation of 331 the word wu f̂- i n such d i a l e c t s as Su-zhou (Wu): [gsu J , 23 42 33 Shuang-feng (Xiang): [qu ], Mei-xian: [gu ], Xia-men: [rjo J , 35 242 Chao-zhou: Igo ], and Fu-zhou: [qua ] (Zi-hui, p.94). 1.3. Tones Pitch v a r i a t i o n which i s used to d i f f e r e n t i a t e the meanings of words i s c a l l e d tone. In i t s function as a tonal phenomenon, pitch, c a r r i e s r e l a t i v e value only, with the height and range of the pitch dependent upon such factors as the age and sex of the speaker. Chao's numerical system, which i s used here for trans- c r i b i n g the tones i n Zhong-shan, i s based on the subdivision of the normal pi t c h range of a speaker into f i v e p i t c h l e v e l s , with "5" designating the highest pitch and "1" the lowest. The tones i n Chinese may be s p e c i f i e d by a single pitch l e v e l or the move- ment of the pitch from one l e v e l to another. (In the case of the - 64 - t h i r d tone i n Mandarin, for instance, which has a f a l l i n g - r i s i n g contour, a m u l t i - d i r e c t i o n a l movement i s involved.) Besides d i f f e r e n t i a t i n g l e x i c a l meaning, tone also d i f f e r - entiates grammatical (morphological) meaning. It i s important to bear i n mind t h i s second function, to which we w i l l l a t e r return. For the present, however, the focus w i l l be on the f i r s t function, i n continuation of the phonological analysis of Zhong-shan. 1.3.1. Tonal System In t r a d i t i o n a l , h i s t o r i c a l analysis, there are four tones, or sheng i n Chinese: Ping-sheng ^ Fg^ 'even tone', Shang- sheng _H_ jfe 'ascending tone', Qu-sheng f-^_ 'departing tone', and Ru-sheng >v^jF 'entering tone'. While s y l l a b l e s containing the f i r s t three tones end i n a long vowel, a g l i d e , or a nasal, Ru-sheng sy l l a b l e s end i n a stop consonant ( i . e . , /p/, / t / or /k/) and i s sometimes referred to as "checked" s y l l a b l e s . As a r e s u l t of the abrupt closure at s y l l a b l e - f i n a l p osition, Ru-sheng s y l l a - bles normally are l e v e l i n pitch and shorter i n duration than t h e i r non-checked counterparts. In terms of western, s t r u c t u r a l l i n - g u i s t i c theory and analysis, the Ru-sheng i s i n complementary d i s t r i b u t i o n with l e v e l p i t c h tones and should not be i s o l a t e d and treated as phonemically d i s t i n c t . For diachronic studies and d i a l e c t a l comparisons, i t i s nonetheless convenient to distinguish s y l l a b l e s ending i n /p,t,k/ from those containing other endings. Or more generally, the adoption of the Chinese terms i s useful for analyzing the modern reflexes of these h i s t o r i c a l Chinese tones both within a given d i a l e c t and across d i a l e c t s . T r a d i t i o n a l l y , the tones i n Chinese are further divided - 65 - into "upper" (yin , or shang J, ) and "lower" (yang J% , or xia "f ) r e g i s t e r s , which r e f l e c t the nature of the h i s t o r i c a l i n i t i a l s , to be discussed l a t e r . Suffice i t to say for now that the Yin-Yang s p l i t was taken into consideration by Chao i n his tonal analysis of Zhong-shan. The numerical values o r i g i n a l l y assigned by Chao for Zhong- shan, and subsequently by Egerod as well, are adopted here with the minor modification of taking into account the shorter duration of the Ru-sheng by assigning to i t single numbers, as opposed to double numbers for the other tones. In the enumeration of Zhong- shan tones according to the Chinese approach, there are a t o t a l of six tones i n the d i a l e c t ; but according to western phonemic analysis, there are only four of them. The four phonemic tones i n Zhong-shan are given i n Table 3 (a) below. The present enume- rat i o n of the tones as Tone 1 through Tone 4 rather than the t r a d i t i o n a l terms as used by Chao, recorded i n the table for com- parative purposes, i s primarily to avoid confusion with the h i s t o r i c a l tones. Table 3 (b) i s a comparison of Zhong-shan tones with the Cantonese ones transcribed by Hashimoto. T r a d i t i o n a l names for the Cantonese tones are used i n the table. Table 3 (a)• Tonal Sys Chan Tone 1: 55 (or 5*) Tone 2: 51 Tone 3: 35 Tone 4: 22 (or 2*) C * Used for checked ;em of Zhong-shan. Chao Yin-ping f£ -f- /Yin-ru f^T 7 v : 55 Yang-ping f-§ f : 51 Shang : 35 Qu £ /Yang-ru \% ̂  : 22 syllables only.) - 66 - Table 3 (b). A Comparison of Cantonese and Zhong-shan Tones. Cantonese Zhong-shan Yin-ping HA 55/53 55 Tone 1 Yang-ping l l f 21/22 51 Tone 2 Yin-shang 35 \ r 1 3 Tone 3 Yang-shang ?&>. 24 J Yin-qu ft'* 44 ] V 22 Tone 4 Yang-qu n i 33 J Shang Yin- r u Xft^ 5 5 Tone 1 Zhong Yin- r u f ft 7^ 4 \ r 2 Tone 4 Yang-ru 3 J Concerning Zhong-shan tones, Tone 1 i s high and l e v e l i n non-checked f i n a l s . Both B a l l (.p. 510) and Egerod (p. 14) observe that the Zhong-shan upper even tone i s s l i g h t l y lower i n pi t c h than i t s Cantonese counterpart. Chao (p.54) notes that i n Zhong- shan, the s t a r t i n g p i t c h of the Yin-ping (Tone 1) i s s l i g h t l y lower than the Yang-ping (Tone 2). I t i s therefore possible to record Tone 1 as /44/. However, for the sake of greater phonemic contrast, /55/ i s used by Chao, Egerod and the writer to represent the phonemic tonal value of the Yin-ping tone. /5/ i s used for Tone 1 i n those Ru-sheng, or checked, f i n a l s which Chao terms "Yin-ru". Because Ru-sheng s y l l a b l e s end i n a stop consonant, these s y l l a b l e s are r e l a t i v e l y short i n com- parison to those with open s y l l a b l e s or those ending i n a nasal or g l i d e . /5/ p a r a l l e l s the high l e v e l /55/ tone i n that i t s pit c h height i s s l i g h t l y lower than the Shang Yin-ru, or the - 67 - high upper entering tone of Cantonese (which Hashimoto and others transcribe as /5/). Tone 2 i s a high f a l l i n g tone, transcribed by Chao and the writer as /51/. Tone 3 i s recorded phonemically as /13/, although Chao observes that i t i s i n fact closer to .1121 . Again, for the pur- pose of maximizing phonemic d i s t i n c t i o n , /13/ was chosen instead. It was noticed i n the present data that Tone 3 was produced with, a minute dip to the tone; that i s , there i s a s l i g h t f a l l before the r i s e . B a l l (.p. 510) also describes the tone i n Macao as a tone which "descend(s) a short space—beginning at the same p i t c h of voice as the Cantonese ~j» ^ , ha hou* (or Yin-qu—Hashimoto' s /44/), lower r e t i r i n g voice, i t f a l l s a l i t t l e lower at i t s end than the Cantonese T ^" , ha p'ing Cor Yang-ping—Hashimoto's /21/)„, lower even tone". Tone 4 i s a mid-low tone, recorded by Chao as /22/. I t i s here recorded as /22/ i n non-checked s y l l a b l e s . The shorter /2/ i s used for stop endings. 1.3.2. Tone Sandhi It i s noted by Egerod (p.14) that one tone sandhi phenome^ non i n Zhong-shan operates i n the following manner: when two or more low l e v e l tones ( i . e . , /22/) occur i n a sequence, the f i r s t tone i s pronounced s l i g h t l y higher than the succeeding, phonemic- a l l y i d e n t i c a l tone(s), and may be transcribed phonetically as ^ / s i kon / 'work, a f f a i r ' would be 33 22 phonetically [ s i : ko:n ]. The same sandhi phenomenon i s observed i n the present data. - 68 - Egerod a l s o notes t h a t i n n o n - f i n a l p o s i t i o n ( i . e . , i n environments not p reced ing open junc tu re or pause ) , the r i s i n g tone (Tone 3) on ly e x h i b i t s a s l i g h t r i s i n g contour , o r even a low l e v e l tone , p h o n e t i c a l l y [12] o r [11] . The p resent Zhong-shan data agree w i t h Egerod 's o b s e r v a t i o n s , w i t h the a d d i t i o n a l remark t h a t the r i s i n g tone i n such environments tends to be s h o r t e r i n d u r a t i o n than when i t occurs before a pause, such as a t the end of a c l a u s e o r i n s e n t e n c e - f i n a l p o s i t i o n . 1 . 3 . 3 . Tone Change "Tone change", o r b i a n - y i n ftfc^ ~fl 'changed t o n e ' , r e f e r s to the m o r p h o l o g i c a l and s y n t a c t i c use of tone d i s t i n c t from i t s l e x i c a l f u n c t i o n . U n l i k e Cantonese, which has a r i c h d i s t r i b u - t i o n of s y l l a b l e s e x h i b i t i n g the tone change phenomenon s e r v i n g v a r i o u s purposes , the grammatical use of tone i n Zhong-shan i s very l i m i t e d . There are two changed tones i n Zhong-shan, as there are i n Cantonese: a h igh l e v e l /55/ and a lengthened , h i g h r i s i n g / 3 5 / , which we w i l l here term " M o d i f i e d Tone 1" and " M o d i - f i e d Tone 3" r e s p e c t i v e l y . Only M o d i f i e d Tone 3 i s d e s c r i b e d by Chao and Egerod. I t i s a h i g h r i s i n g tone which Egerod t r a n s - c r i b e s as / 3 5 / . The examples t h a t Chao and Egerod g i v e i n v o l v e s y l l a b l e s which o r i g i n a l l y had Tone 3 or Tone 4. A l though exper imenta l s t u d i e s need to be conducted on the b a s i c and changed tones i n Zhong-shan befo re one can d e s c r i b e i n more d e t a i l and w i t h g r e a t e r p r e c i s i o n the tone change s i t u a t i o n , p e r c e p t u a l l y , a t l e a s t , the m o d i f i e d tones are s l i g h t l y longer i n d u r a t i o n than t h e i r b a s i c forms, and are more prominant i n the speech of the female in formants than the male i n f o r m a n t . More - - 69 - oyer, i t has been observed i n the present data that the other tones i n the d i a l e c t also can undergo tone change. In represent- ing bian-yin, the o r i g i n a l tone i s given f i r s t , with the changed tone following, separated from the o r i g i n a l by a comma. The charac- ter undergoing the tone change i s also marked by an asterisk (e.g., chi / j a : k 2 / 'to eat', / j a : k 2 , 3 5 / 'to have eaten' (eat + PERFECTIVE)). Modified Tone 1 does not d i f f e r s i g n i f i c a n t l y from the regular Tone 1 i n terms of pi t c h height. It i s the increase i n s y l l a b l e length which i s most prominent. B a l l (p. 511). describes the Modified Tone 1 i n the Macao (Zhong-shan) d i a l e c t as only . . j .'<: s l i g h t l y higher i n p i t c h l e v e l than the regular Tone 1. The Modi- f i e d Tone 1 i s , i n fact, of the same p i t c h height as the ordinary Yin-rping tone i n Cantonese ( r e c a l l i n g that the Zhong-shan Yin- ping has been noted as being lower i n pitch than the Cantonese one). Tone 1 words which underwent tone change w i l l be indicated using 55 55* an asterisk (e.g., l a / l a : j / 'to p u l l ' , 25L * / l a : j / 'to have pulled' (pull + PERFECTIVE)). One of the functions of bian-yin uses Modified Tone 1 to indicate f a m i l i a r i t y , or c o l l o q u i a l speech. Such changes are highly unpredictable. Zuo fl'jte 'yesterday' / t s o k 2 / and r i |3 'day' 2 / j a t /, for example, combine to form the word 'yesterday', which 2 2 5 i s c o l l o q u i a l l y pronounced /tsiok j a t ' /, with a v o c a l i c change i n the f i n a l of zuo 0̂  and a tone change i n r i # • These changes r e f l e c t c o l l o q u i a l , d a i l y usage, whereas the pronunciation 2 2 of zuo-ri Qty Q /tsok j a t / r e f l e c t s a more l i t e r a r y form. 51 Likewise, ming $Q ' l i g h t ' /ming / combined with r i 0 2 'day' / j a t / i s highly l i t e r a r y for 'tomorrow'. The Zhong-shan - 70 - c a s u a l word f o r ' tomorrow' i s based on the combinat ion of ming p lus zao J^- ' e a r l y ' / t s o w 1 3 / to produce / m i n 5 1 ' 5 5 t s o w 1 3 / . In t h i s c a s e , the tone change on ming $fl i s accompanied by a change 19 xn the a r t x c u l a t x o n of the n a s a l endxng as w e l l . The word ya j^L ' t o p r e s s ' has the formal p r o n u n c i a t i o n of / a : t 2 / , which would be used i n terms such as y a - l i JjjjjL f) 2 2 ^ ' p r e s s u r e ' / a : t l i k / . C o l l o q u i a l l y , the word i s / a : t ' / t o 2 mean ' t o p r e s s ' , and / a : t / f o r the p a s s i v e meaning of ' t o be p r e s s e d ' , as i n y a - z h u J^L ( ' .press + c e a s e ' ) . In the word y i ^jj^ ' s i s t e r of one 's w i f e or m o t h e r ' , / i 5 1 / , when i t occurs i n d i r e c t address , i t combines w i t h a §U (or i t s 2 2 2 2 5 1 5 5 graph ic v a r i a n t /a /) to form /a i ' / . (A i s an empty p r e f i x a l form which serves to prevent m o n o s y l l a b i c i t y i n a p p e l l a t i o n s . ) In Zhong-shan, a - y i i s 'mother ' s younger s i s t e r ' . The tone change to M o d i f i e d Tone 1 i s o b l i g a t o r y . ' M o t h e r ' s o l d e r s i s t e r ' i n Zhong-shan i s y i - m a / i ma / , w i t h no tone change i n y i . In words such as H^JL - f i n a l l y ' /saw mi / , tone 5 5 13 5 5 change i s o p t i o n a l , wi th , /saw mi ' / e q u a l l y p e r m i s s i b l e . A & /t, 2 2 . 1 3 , . i n , 2 2 , 5 5 . synonym, \ / h a w mx / xs u s u a l l y pronounced /haw 13 5 5 mi ' / i n the speech of the Zhong-shan in fo rmants i n the s tudy , 33 a l though Hashimoto (p. 97) i n d i c a t e s the o p t x o n a l i t y of Ih-ew 2 4 33 2 4 5 5 mej J versus [hnw mej ' J . Note a l s o t h a t i n Zhong-shan, both s y l l a b l e s undergo tone change. We t u r n now to the more s y s t e m a t i c grammatical f u n c t i o n of tone change, a l l of which i n v o l v e M o d i f i e d Tone 3 . One f u n c t i o n of M o d i f i e d Tone 3 , f o r i n s t a n c e , i s to i n t e n s i f y the a d j e c t i v e i n a r e d u p l i c a t i n g , m o n o s y l l a b i c a d j e c t i v e i n which the tone change - 71 - occurs on the f i r s t member of the reduplicating p a i r . Thus, 13 'good', for example, i s hao Jj$- /how /, and 'very good 1 i s 1 ̂  ^5 13 22 /how ' how /. Correspondingly, 'big' i s da ^ / t a : j /, j 22 35 22 while 'very big' i s 7 ^ * 7 ^ / t a : j ' t a : j /. As already stated, the derived tone i s longer, containing a high r i s i n g con- tour. Besides i n t e n s i f i c a t i o n , a reduplicated adjective also has the meaning of 'however + ADJECTIVE 1; thus, 7 ^ * 7^- / t a : j ' 22 ta: j / also has the meaning of 'however big', as i n the follow- ing clause C'NEG" = negative): ( 3 ) K* K & h $ (-> ' .22,35 ^ .22 , 55 13 . 22 ta:j ta: j tu mow j ung CBIG BIG STILL NEG-HAVE USE) 'However big ( i t i s ) , ( i t ' s ) s t i l l no use.' In the case of an adjective modified by the character hao /how13/, meaning 'very', emphasis can be placed on the word 'very' by means of tone change to i n t e n s i f y the adjective follow- ing i t . Thus, the phrase 'very f a r ' ^ 3 " l i b - /how 1 3 y n 1 3 / can be \ 13 35 13 stressed by rendering i t as Jfo*/how ' yn /, denoting even greater distance. The character j_i / k i 1 3 / , on the other hand, has the meaning of ' f a i r l y , rather', and only i n the modified tone does i t acquire the meaning of 'very', serving to in t e n s i f y the adjective. 'Quite good 1, for example, i s 3|;4V/ki 1 3 how13/, while / k i ' how /, with the changed tone, means 'very good, stupen- dous'. A rough t r a n s l a t i o n a l equivalent of sentence (4), for example, would be 'It sure tasted good!' or 'It sure was good!' ("S" = sentence, "PRT" = pa r t i c l e . ) - 72 - (4) ° . Yl3,35 u 13 . ,2 . 22 k i how ja:k l a (VERY GOOD EAT S-PRT) 'It was very good (to eat).' Another regular function of the bian-yin i s to compensate for a deleted s y l l a b l e containing a high tone /5/ or /55/; that i s , the pitch of the tone i s absorbed by the immediately preced- ing s y l l a b l e when the s y l l a b l e which o r i g i n a l l y c a r r i e d i t was deleted. One case of such a usage i s when there i s a change i n the tone of a verb. The tone change serves to mark p e r f e c t i v i t y 55 i n l i e u of the post-verbal perfective marker /p'ow / i n Zhong- 55 55 55 shan. Variant forms of /p'ow / are /how / and /ow /. The perfective marker may sometimes be the repeat ..of the f i n a l on the verb i n which the repeated f i n a l c a r r i e s the high pitch, of Tone 1. j. 22 In the case of qu 'to go' /hy /, for example, the perfective 22 55 22 55 of the verb may be /hy p'ow / (or i t s var i a n t s ) , or /hy y /. 22 35 P e r f e c t i v i z a t i o n through tone change would y i e l d /hy ' /. In Zhong-shan, the marker of p e r f e c t i v i t y *|J«L (or *fi. ) /tso / i s also used, but i t i s considered a more formal or p o l i t e form bor- rowed from Cantonese. In analyzing the p e r f e c t i v i z a t i o n of verbs by means of tone change, one theory that has been advanced for Cantonese i s likewise proposed here, namely, that the tone change i s a r e s u l t of copying the high pitch of the perfective marker onto the end of the preceding s y l l a b l e , which i s the verb. When the perfective marker i s deleted, a trace remains i n the compensatory lengthening of the preceding s y l l a b l e , with a concomitant r i s e i n the contour of that s y l l a b l e at i t s end point. There are a few cases of tone change connected with the - 73 - 5 deletion of jjL 'one' / j a t /. Three cases of y i - d e l e t i o n and concomitant tone change w i l l be discussed. As i n p e r f e c t i v i z a - t i o n by means of bian-yin above, the tone change here also serves a compensatory r o l e . The f i r s t case of y i - d e l e t i o n involves t r i - s y l l a b i c phrases i n which, the f i r s t s y l l a b l e i s a monosyllabic verb, the second yjL ~-~ , and the t h i r d the reduplication of the monosyllabic verb. The deletion of y i results i n the f i r s t verb form acquiring Modified Tone 3. Thus, 'to take a look' kan-yi-kan i / 22 5 22 ^jj -~ ̂ jj , for instance, i s /hon j a t hon / (.'look one look'). 5 The deletion of 'one' / j a t / produces the long, high r i s i n g tone . i 2235 22 on the f i r s t s y l l a b l e y i e l d i n g & * j|j /hon ' hon /. As observed by Kwok (.1971:50). for Cantonese, the resultant r e d u p l i - cated verb form denotes a b r i e f duration of action denoted by the verb. In his analysis of Mandarin, Chao (1968:204) refers to this reduplication as the "tentative aspect of verbs". In the second case of Modified Tone 3 a r i s i n g from the 5 deletion of 'one' / j a t /, the singular occurrence of an action 5 13 can be s p e c i f i e d using the expression y i - x i a ^- e"f* / j a t ha / 'one time/occasion'. CXia nrf (or ~f ) i s a c l a s s i f i e r for the number of occurrences of an action.) When an action occurs once, the number 'one' i s usually omitted. Thus, kan y i - x i a ^ *"f 22 5 13 'to look once' /hon j a t ha / (.'look one time'), becomes 22 35 13 /hon r ha /, with a tone change on the verb. The meaning of 5 13 / j a t ha / i s sometimes extended so that i t does not necessarily always haye the l i t e r a l meaning of 'one time/occasion', as shown i n (.5).. (A c i r c l e " O " i s used to designate a c o l l o q u i a l word with, no written character associated with i t . "CL" = c l a s s i f i e r . ) - 74 - <5> a ' ^ V a S ' f 13 ^ 5 5 n 22 hon ha ..sin syn (LOOK+ONE CL BEFORE PLAN) 'We'll see.'/'Wait and see f i r s t . ' D' Zl 13 13,35 . 13 how yn ' ha (GOOD FAR+ONE CL) 'It's kind of f a r . ' / ' I t ' s quite far (contrary to expecta- tion) . ' c* 13,35 , 13 . 22 51 . .22 . . 5 ja:ng ha tsaw m k i tak (SHADOW+ONE CL THEN NEG REMEMBER ABLE) 'One moment l a t e r (I) (already) forgot!' ( i . e . , i n an instant, very quickly) , , 51 . .55* . 1 3 * 22 22 k'y ha:j ha tsaw ha:m (HE/SHE TOUCH CL THEN CRY) 'As for him, just brush against (him) l i g h t l y and (he) c r i e s . ' In the t h i r d case of compensatory tone change a f t e r y i - deletion, the change of Modified Tone 3 occurs when the second y i —~ i s omitted i n expression containing " y i + CL + y i + CL", such as yi-kuai-yi-kuai — "~ 'piece by piece' ( i . e . , one piece after another) / j a t 5 f a : j 2 2 j a t 5 f a : j 2 2 / ('one piece one p i e c e 1 ) . Deletion of the second y i yi e l d s — J$k* £5&>/jat f a : j ' fasj /, This tone change i s a general one af f e c t i n g any reduplicated c l a s s i f i e r i n the above environment. Note that i f a c l a s s i f i e r i s simply reduplicated, the expression has the meaning of "every + CL". In Zhong-shan the f i r s t member of such a reduplicated p a i r J ^ 2 2 2 2 does not undergo tone change. Thus, / f a : j f a : j / 33 22 (phonetically I f A i j f A i j 1 due to tone sandhi noted e a r l i e r ) means 'every piece'. (Contrast t h i s with Cantonese 44 35 44 I f A i j f A i j ] 'every piece'.) There are also some cases of compensatory tone change due - 75 - to s y l l a b l e deletion that i s highly i d i o s y n c r a t i c and hence non- productive. The expression ^ / k i 1 3 t o ^ / 'how "much, how many', for example, i s often reduced to simply ^ * / k i 1 3 ' " ^ / . Thus, the question 'How much did i t cost?' may be uttered as (6a). or (6b). ("Q" = question.) (6) a. * * 1 3 J ? 5 5 5 1 B. 1 3 2 2 k i . . t o t s i n ma:] a (HOW-MUCH MANY MONEY BUY Q-PRT) 'How much did (it) cost?' 13 22 k i ' t s i n " " " ma:j a (HOW-MUCH MONEY BUY Q-PRT) 'How much did (it) cost?' Tone change can also serve to replace the word dou ^ 55 'also, likewise' /tu / as a marker of inc l u s i o n (INCL). (Note that dou ^jjj* has the same meaning i n Cantonese that i t has i n Zhong-shan, but i n Mandarin dou means ' a l l , altogether'.) The sentence 'I'm going, too', for example, may be uttered as 7 (a) or (b) (7) a. & 1 3 lf55 * 2 2 ngo tu hy (I ALSO GO) 'I'm going, too.' nTo13'35 hy2 (I+INCL GO) 'I'm going, too.' The r e s u l t a t i v e or s e r i a l verb 'to return, give back (some- * ̂/ 1 3 5 5 1335 thing)' O ife-- /pia:ng fa:n / i s often reduced to /pia:ng ' /, as i n (8) . ngo pia:ng fa:n k'y l u (I GIVE RETURN HIM/HER S-PRT) 'I have returned (it) to him.' - 76 - b - & 1 3 . ° * 1 3 , 3 5 . 5 1 n ° 2 2 ngo pia:ng ' k'y l u (I GIVE+BETURN HIM/HER S-PRT) 11 have returned (it) to him. 1 There are also at least a couple of tone changes to Modi- f i e d Tone 3 which are highly i r r e g u l a r . The verb you 'by 5 1 (someone—in passive constructions)' /jaw / i n the changed tone means 'to allow, to be up to (someone to decide)', as i n sentence (.9) • (9) ** >L o i y ; . 5 1 , 3 5 . , 5 1 , 2 2 jaw k'y l a (ALLOW HIM/HER/IT S-PRT) 'Let i t be./Don't bother with it./Leave i t alone. 1 \- 51 Another example i s the word mang "̂ j 'blind' /ma:ng / 5 1 3 5 which, i n the Modified Tone 3 /ma:ng ' /,..is used to describe some action done b l i n d l y , fervently or p e r s i s t e n t l y , as i n ( 1 0 ) . < 1 0 > * 5 1 r 5 1 , 3 5 ° / 2 2 ^ 2 k'y ma:ng ' kam ja:k (HE BLINDLY THUS EAT) 'He kept on eating ( f u r i o u s l y ) . ' The l a s t example involves interrogatives containing the c o l l o q u i a l word dian 'how' / t i m 1 3 / . In Cantonese, a c o l l o - 1 **** q u i a l expression such as 'how, i n what way' i s 3 5 3 3 1 3 [ti:m qceirj ]. In Zhong-shan, /tim / undergoes tone change to the high, r i s i n g Modified Tone 3 for the same expression: ,u - jo o r 2 2 1 3 /tim ' Jong /. Likewise, other combinations with /tim / for interrogatives r e s u l t i n a tone change i n the word; for example, 'why' i s / t i m 1 3 ' 3 5 k a : j 1 3 / ; O / t i m 1 3 ' 3 5 t s i 1 3 / means 'how, by what means'. - 77 - 1.4. Combination of I n i t i a l s and Finals Phonemically, including the zero i n i t i a l , there are eighteen i n i t i a l s i n Zhong-shan. Of the f i n a l s there are a t o t a l of seventy, i n which tonal d i s t i n c t i o n s have not yet been taken into account. When these are also considered, the combination of the i n i t i a l s and f i n a l s , together with tonal d i f f e r e n t i a t i o n , y i e l d s a sum of at lea s t 1,600 d i f f e r e n t s y l l a b l e s i n the Zhong- shan d i a l e c t . This figure i s , of course, less than the absolute potential number, which, may be the re s u l t of natural gaps here and there, occurring at random i n a language, a r i s i n g from a "defective" d i s t r i b u t i o n of some segments; that i s , there are pot e n t i a l l y permissible combinations that are not actualized i n the d i a l e c t . Some of the gaps are systematic, and may be the consequence of mergers i n the d i a l e c t which are phonologically conditioned, a point which w i l l become clearer l a t e r . Other gaps are the re s u l t of co-occurrence r e s t r i c t i o n s . The most wide- spread case of co-occurrence r e s t r i c t i o n among the Chinese d i a l e c t s i s l a b i a l d i s s i m i l a t i o n , which i s also c h a r a c t e r i s t i c of Zhong- shan. I t i s a topic that w i l l be discussed next. Besides the phenomenon of l a b i a l d i s s i m i l a t i o n , we w i l l also examine that of syncope, the deletion or loss of a phonolo- g i c a l segment r e s u l t i n g i n the contraction of a word. The p a r t i a l reduction of a word creating a sesquisyllabic structure w i l l also be studied, with a discussion of the d i s t i n c t i o n between c o l l o q u i a l versus l i t e r a r y forms i n the d i a l e c t concluding t h i s chapter. 1.4.1. Labial D i s s i m i l a t i o n There are several environments i n which the l a b i a l - 78 - d i s s i m i l a t i o n process occurs i n Zhong-shan. They are e s s e n t i a l l y the same environments as those found i n Cantonese, In both d i a - l e c t s , l a b i a l i n i t i a l s (including velars followed by medial /w/, a l t e r n a t i v e l y analyzed i n Cantonese as l a b i a l i z e d v e l a r i n i t i a l s ) cannot occur together with l a b i a l consonantal endings. Nor can the l a b i a l glide i n i t i a l co-occur with the l a b i a l glide ending. Thus, the following are examples of impermissible s y l l a b l e s i n both Zhong-shan and Cantonese: [*mAi.p, *p'i:p, *moim, *fo:m, *kwi:.m, *k'wAip, *wo:p, *wew] . There i s also a pr o h i b i t i o n of front rounded vowels either preceded by a l a b i a l i n i t i a l , or followed by a l a b i a l ending. The s y l l a b l e s [*0w, *0p, *yip, *p0, *p'y:], for instance, do not occur i n either Zhong-shan or Cantonese. 1.4.2. Syncope A few examples w i l l be c i t e d . The c o l l o q u i a l expression for 'what', for instance, i s % ^ /mat5 j a 1 3 / which, when short- 51 ened by syncope, becomes /mia /. The negative imperative 'don't' /m51 hou 1 3/ (.'not good') i s sometimes reduced to /mow13/. k 24 (The alternative term of [mej ] i n Cantonese i s not used i n 22 2 Zhong-shan.) 'Twenty' i s er-shi -\ /ngi sap / ('..two ten') , which can be syncopated to if /ngap 2/. This syncopated form i s more frequently used for figures between twenty-one and twenty- nine i n c l u s i v e than for the number 'twenty' by i t s e l f ; for example, u- -* 2 55 'twenty-three' i s rendered - f /ngap sa:m / f r o m er-shi-san i - 22 2 55 S^— ~T /ngi sap sa:m / ('two ten t h r e e 1 ) . (In Zhong-shan, -* -J, 55 2 ' t h i r t y ' san-shi Ju. T /sa:m sap / (.'three ten') often becomes simply . / s a 5 1 / , since the preceding process of syncope, i f - 79 - 2 applied to the number ' t h i r t y 1 , would have yielded /sap /, lead^ ing to confusion with the same s y l l a b l e already being used to mean 'ten' .) There are some words for which only the syncopated form seems to have survived. The d e i c t i c words 'here' and 'there' 5 5 appear to be a fusion of O /ku / 'th i s ' and the diminutive- p a r t i c l e er / n g i 5 1 / to y i e l d 0 / k u j 5 5 / 'here', and O 55 51 55 /nu / 'that' plus /ngi / combine to form O /nuj / 'there'. This i s p a r a l l e l to Mandarin zhei 'jX, for 'this' and na f)^ for 'that', with the corresponding locative terms zher yg\^j(.zhei Vl\ + er 'tfij ) for 'here' and ner (na J)J + er ) for 'there' . In 55 55 Zhong-shan, /kuj / and /nuj / also mean 'this thing/matter' and 'that thing/matter' respectively. The same p a r t i c l e er /ngi"* 1/ mentioned above was 5 5 probably also fused i n the interrogative word O /naj / 'where', 5 5 as was postulated by our informant, Chen, i n which O /naj / i s the syncope of na ^ /no 5 1/ 'where' (which only occurs i n l i t e r a r y • • in 51 usage) and the diminutive p a r t i c l e er /ngi /. (Compare thi s with Mandarin nar (na ^ + er ) 'where'.) 1.4.3. Sesquisyllabic Structures While the examples i n the preceding section show the reduc- t i o n of two s y l l a b l e s to a single one by syncope, there are other s y l l a b l e s i n Zhong-shan where th i s merger i s only p a r t i a l l y accom- plished, r e s u l t i n g i n a "sesquisyllabic structure", a term used by J. Matisoff (.1973:86) to refer to morphemes i n Proto-Austro- A s i a t i c (a l i n g u i s t i c superstock which includes, among i t s members, Mon-Kmer and Viet-Muong) which were "a s y l l a b l e and a h a l f " i n - 80 - 20 length : "(t) hat i s , the prevoealic consonant was often preceded by-a ' p r e - i n i t i a l ' consonant, as i n the modern Cambodian words psaa 'market', tkiam 'jaw', ckat 'dog', kr}aok 'peacock'". In Zhong-shan, there are a few cases of words which can be said to contain sesquisyllables. Such words i n the d i a l e c t have consonant clusters which are separated by an epenthetic schwa. From a synchronic l e v e l , these sesquisyllabic structures can be analyzed as reductions of f u l l form which are p o l y s y l l a b i c i n o r i g i n . This process of schwa-reduction i s moreover r e s t r i c t e d to words i n which the second s y l l a b l e undergoing reduction contains the i n i t i a l / l / . The schwa-reduction process actually affects a very l i m i t e d vocabulary. With the exception of a small handful of c o l l o q u i a l expressions found i n the data thus f a r , other mani- festations of t h i s phenomenon involve onomatopoeic types of words. Both the f u l l and reduced forms are found i n Zhong-shan, with the f u l l form e s s e n t i a l l y c i t a t i o n forms. I t i s the reduced forms that are normally used i n d a i l y , conversational speech. Given the s y l l a b l e structure that normally does not admit to a sesquisyllabic structure, the synchronic analysis very naturally seeks to derive the reduced form from a . f u l l , p o l y s y l l a b i c one. Diachronically, however, there are arguments for proposing o r i g i n a l i n i t i a l consonant clusters for at least some of the sesquisyllabic forms i n Zhong-shan; that i s , some of the sesquisyllabic forms can be associated with words which have been reconstructed i n Old Cor Archaic). Chinese as containing i n i t i a l consonant clusters. Consider, f i r s t , the c o l l o q u i a l word for 'corner-(e.g., of a room)_1 . The characters usually given for i t i n Cantonese and Tai-shan (another Yue dialect) are ji.fl . They are pronounced 4 5 35 1 5 35 Ikoxk l o i k t 'Bw J i n Cantonese and [kok Iok hau J i n Tai-shan. The same term i s also found i n Zhong-shan. The c i t a t i o n form for 2 5 51 i t i s Ikax.k l o i k t'-ew J. C o l l o q u i a l l y , the word i s usually reduced to I k 3 l o : k 5 t ' B w 5 1 ] (,/klok5 t'aw 5 1/). Hashimoto (.1972b: 34) expresses doubt concerning the connection between the second • ti s y l l a b l e and the character luo (.169-21) , which i s normally pronounced i n a low tone i n Zhong-shan, as i t i s i n Cantonese. Hashimoto suggests that both, the f i r s t and the second s y l l a b l e i n fact stand for the character j i a o jj) (.183-20) , and further pro- poses that j i a o jl) i s derived from a h i s t o r i c a l * k l - c l u s t e r i n proto-Yue. (Tou £1^ i s simply a word-formative s u f f i x (Kratoch- v i l , 1968:68), and i s not relevant to the present discussion,) Hashimoto's evidence for claiming that the f i r s t two s y l l a b l e s both represent the character j i a o i s based on fan-qie: j i a o $J has two d i f f e r e n t fan-qie's which demonstrate that i t has both a *k- and an *\i i n i t i a l . Accordingly, Hashimoto hypothesizes that when i n i t i a l c lusters f e l l out of usage, the only means whereby these clusters could survive was by having a s y l l a b l e inserted between the two consonants i n the c l u s t e r . (Yang (.1971) , for example, referred to such an i n s e r t i o n — t h a t of a v o c a l i c element between adjacent consonants—as a "process of anaptyxis".) To account for the tonal difference between the f i r s t and second s y l -4 5 21 35 lable i n the word Ikaxk laxk t'BW ' J 'corner', Hashimoto suggests that the high tone of the second s y l l a b l e i s a r e s u l t of tone change, since such a change i s not uncommon i n reduplicated s y l l a b l e s — o r , i n t h i s case, pseudo-reduplicated s y l l a b l e s — i n the Yue d i a l e c t s . Various Chinese phonologists have likewise reconstructed an i n i t i a l consonant cluster for j i a o |2) , which i s a Grade II word. Jiao jt) has been reconstructed by Fa-kao Chou as Archaic Chinese *krewk, for instance, and by Pulleyblank as Old Chinese *krak . The Zhong-shan sesquisyllabic form for 'corner' would therefore lend support for some kind of * k l - or *kr- i n i t i a l consonant c l u s t e r for j i a o i n Old Chinese, a cl u s t e r which had survived i n protp-Yue, and apparently i n e a r l i e r forms of other 55 " 35 In '*£ d i a l e c t s as well, as witness Ika l s r J |*J }%• i n (Peking) Man- darin for 'a hidden corner' (Hashimoto, 1972b:33-34). The t r i - s y l l a b i c " f u l l " form for the word 'corner' i n Zhong-shan i s b a s i c a l l y a c i t a t i o n form which expands the sesquisyllable to f i t the more common CVC s y l l a b l e structure i n the d i a l e c t . The pro- 21 cess of "dimidiation" , which p a r a l l e l s the h i s t o r i c a l process proposed by Hashimoto, i s used i n such situations as teaching the word to a c h i l d , or repeating the word slowly to an i n q u i s i t i v e and persistent l i n g u i s t . It should be noted that despite the f u l l form given by Hashimoto for the word 'corner' i n Cantonese, McCoy (1966:185,fn. 27) i n fact argues that t h i s word i s one of a very rare number of words i n Standard Cantonese which possess atonic s y l l a b l e s . McCoy indicates such s y l l a b l e s by using the tone d i a c r i t i c [°]. The nuclear vowel i n these s y l l a b l e s i s also reduced to a schwa and linked to the following s y l l a b l e by a hyphen to show close junc- ture. Of the words i n his data, McCoy found only two which con- 4 35 22 t a i n an atonic s y l l a b l e : [ks° - loxk t'B W 1 'corner' i s one 33 33 of them, and [hem b a ° - I A X - O ] ' a l l , completely' i s the other. (.We w i l l return to the second word later.) In a l l p r o b a b i l i t y , McCoy's atonic s y l l a b l e , which has - 83 - close juncture with the following f u l l s y l l a b l e , i s what we are treating here as a sesquisyllable. However, McCoy seems to deny the existence of the f u l l , c i t a t i o n form; he notes that the forms, containing neutral tone s y l l a b l e s c i t e d i n d i c t i o n a r i e s are record- ed "as i f " they were pronounced with, f u l l tones. The fact that both f u l l and reduced forms can be e l i c i t e d from the present informants indicates at least that both variants do e x i s t and are not mutually exclusive. As suggested e a r l i e r , the c i t a t i o n form i s l i k e l y an attempt to expand the sesquisyllabic form to conform with the more orthodox s y l l a b l e structure of the d i a l e c t . Another c o l l o q u i a l expression i n Zhong-shan which ihas a i sesquisyllabic structure i s the word 'knot'. The c i t a t i o n form 5 5 ^ 5 i s [k'ixt l i : t ], and the reduced form [k' lix.t ]. Hashimoto (1972b: 37, fn..'19; 38,fn.26) notes that K. Whitaker (1952:47-48), i n her d i s s e r t a t i o n on the "Characterization of the Cantonese d i a - l e c t with, special reference to i t s modified tones", proposes an i n i t i a l * k l - c l u s t e r for the word 'knot' by making a connection between the l i t e r a r y word for 'knot' j i e (133-1) [kixt^J 2 (./kit / i n Zhong-shan) and the c o l l o q u i a l word for 'knot' i n 4 4 Cantonese, which has the variant forms of [ k i i t ] and [ l i x t ]• Hashimoto (1972b:33). , however, rejects Whitaker's proposal based on the alternative pronunciation of the c o l l o q u i a l word for 'knot 1 on the premise that the l a t t e r has no fan-qie evidence to support her claim. Hashimoto proposes, instead, a connection between the word x i e k*$. [sixt^] 'to t i e ; a b r i d l e ' and the c o l l o q u i a l word [li x t J. She bases her claim on the observation that the character *f!!i|_, for example, was supposed to have had an * i _ i n i t i a l accord- ing to fan-qie s p e l l i n g . (However, the fan-qie for xie js--t£ i n - 84 - both Chou (1973) and the C i - h a i d i c t i o n a r y , f o r i n s t a n c e , shows an * s - i n i t i a l o n l y . ) Nonethe less , on the b a s i s of her r a t h e r tenuous f a n - q i e ev idence , Hashimoto p o s i t s an * s l - c l u s t e r f o r the word x i e Zhong-shan ev idence , on the other hand, would tend to support W h i t a k e r ' s c l a i m of a * k l - c l u s t e r . Whi le Cantonese a l t e r n a t e s between a /k/ and an / ! / i n i t i a l f o r the c o l l o q u i a l word ' k n o t ' , i n Zhong-shan the same word has a d i s y l l a b i c s t r u c - 5 5 5 tu re / k ' i t l i t / , or the s e s q u i s y l l a b i c form of / k l i t / . I f one can assume t h a t the words ' c o r n e r ' d i s c u s s e d above, and the present word ' k n o t ' underwent a s i m i l a r process i n the breakdown of i n i t i a l consonant c l u s t e r s , there i s s u f f i c i e n t b a s i s f o r s u g - g e s t i n g t h a t the word j i e / k i t 2 / ' k n o t ' o r i g i n a l l y bore some k i n d of * k l - c l u s t e r . H i s t o r i c a l l y , the r e c o n s t r u c t i o n of a * k l - c l u s t e r f o r j j e l£ , a Grade IV word, has been suggested by P u l l e y b l a n k (.196 2) , then as a * k i - c l u s t e r . P u l l e y b l a n k d i s t i n g u i s h e s between c l u s t e r s i n Old Chinese which c o n t a i n * - l - and those which c o n t a i n * - r - : * - r - occurs w i t h Grade I I words ( e . g . , j i a o ) , and * - i - w i t h Grade IV words ( e . g . , j i e j& )• Very b r i e f l y , P u l l e y b l a n k was mot ivated to p o s t u l a t e a * k l - c l u s t e r f o r such Grade IV words as j i (154-16) ' l u c k y ' / k a t 5 / (Middle Chinese * k j i t ) and j i e ~tj (154-17) ' t o i n v e s t i g a t e ' / k ' i t 2 / (Middle Chinese * k ' j i t ) . . i n o . order to e x p l a i n why they d i d not have p a l a t a l i n i t i a l s i n M idd le Ch inese , whereas Grade I I I words, such as z h i JL (48-12) ' b ranch ' 55 / t s i / (Ear l y M idd le Chinese *t<?ia) d i d develop p a l a t a l i n i t i a l s . I t was the presence of * - l - i n the Old Chinese forms of Grade IV words such as jjL a and j i e ' f S which b locked p a l a t a l i z a t i o n a t - 85 - the Middle Chinese stage. 5 In the modern Zhong-shan form of / k ' l i t / for 'knot', one s t i l l needs to account for the presence of an aspirated i n i t i a l instead of the unaspirated one i f an association i s to be made between the Zhong-shan c o l l o q u i a l term for 'knot' and the l i t e r a r y 2 one of j i e i/£ / k i t /. There i s also the question of the difference i n tone between the two forms. Neither poses a major d i f f i c u l t y . A number of c o l l o q u i a l (and l i t e r a r y ) words which are pronounced with an aspirated stop i n i t i a l i n Zhong-shan have been reconstructed with p l a i n stops. Words containing the Jian |L» (*k-) i n i t i a l , \ 22 for example, include gua jf%\* (45-1) 'to hang' /k'wa / and guang J^i, (180-1) 'to s t r o l l ' /k'wa.ng22/. Cantonese has [kw A: 4 4J and 33 IkwAiy J respectively for the two words. Mandarin, i t may be noted, likewise showsunaspirated i n i t i a l s for gua and guang . It i s therefore not implausible to consider the reconstruct t i o n of a p l a i n * k l - c luster despite the c o l l o q u i a l word 'knot' i n Zhong-shan showing aspiration on the i n i t i a l segment. Concerning the problem of a tonal difference i n the two forms for 'knot', one might f i r s t look at the c o l l o q u i a l form for 2 5 51 'corner'. In /kok Iok t'aw /, the f i r s t s y l l a b l e preserves the b 2 tonal value of the word ji a o ji| /kok /. Presumably, the second 5 s y l l a b l e /Iok / i s the r e s u l t of a tone change such that i t o r i -2 5 5 g i n a l l y bore tone /2/ ( i . e . , /Iok ' / ) . In the word 'knot' / k ' i t l i t 5 / , neither s y l l a b l e has preserved the Zhong-shan r e f l e x of the tone for j i e / k i t 2 / , namely tone /2/. I t i s possible that the word 'knot' may have undergone an intermediary stage during which only the second s y l l a b l e was at f i r s t affected by tone change (as i n the case of the word 'corner'), and i t i s only - 86 - subsequently that the f i r s t s y l l a b l e was also thus affected. As noted e a r l i e r , besides the word 'corner 1, McCoy (1966: 185,fn.27> also mentions the Cantonese word ' a l l / completely', 33 33 which i s phonetically Ihem ba° - IAI-O ] . The same word i s 33 33 33 transcribed as Ihem P A : I A I I J ] . by Hashimoto (p. 333). The phoneticization of Huang's (1970:394) t r a n s c r i p t i o n of 'altogether, 33 33 33 a l l t o l d ' i s Ihem p A i n I A I J J ] . In Zhong-shan, t h i s word ' a l l ' 51 51 51 51 is pronounced /ham pa la:ng / i n c i t a t i o n form and /ham 51 pla:ng / i n d a i l y speech. In narrow phonetic t r a n s c r i p t i o n , /p/ assimilates the voicing of the preceding segment and i s pronounced 5 1 a 51 l b ] : [h-em b I A I H J . Except for tonal differences, McCoy's form for Standard Cantonese and the corresponding Zhong-shan form are i d e n t i c a l . It i s i n s t r u c t i v e to add that two other Yue dia- l e c t s recorded by Hashimoto (p.19) likewise appear to contain sesquisyllables for the word ' a l l ' : Hashimoto records Ihorn2 plan 1] for ' a l l ' i n Teng-xian, and [Sham °plag] i n Tai-shan. Regarding the portion of the c o l l o q u i a l word ' a l l 1 which < : 51 i s the sesquisyllable (e.g., /pla:ng / i n Zhong-shan, or the 33 33 corresponding Cantonese sequence [ p A i 1A:O J)/ Pulleyblank (personal communication) suggests l i n k i n g i t to the l i t e r a r y word fan fa (109-1) ' a l l ' , derived from LMC *ff\aim<EMC *buam. (Fan 51 21 i s pronounced [ f A i n ] i n modern Zhong-shan and If A M I J i n Cantonese.) Pulleyblank points out that fan f\j i s the phonetic in a character such as fan (109-7) (/fa:n 5 1// EMC *b uam)/ which had been used to transcribe the word 'Brahma'. Fan ) \ i s also the phonetic i n lan , which i s LMC *lam, and modern Cantonese It i s not inconceivable that the word fan ]3u once bore a [ l A i m ] - 87 - consonant c l u s t e r * b l - or *br-. In Zhong-shan, the presence of 51 a velar nasal ending i n the modern r e f l e x /pla:ng / can be ea s i l y explained i n terms of l a b i a l d i s s i m i l a t i o n , changing *-m to - n i n the presence of the l a b i a l i n i t i a l *b-. In the s y l l a b l e s / p a 5 1 l a : n g 5 1 / and the sesquisyllable / p l a : n g 5 1 / , the tone i s /51/ thus corresponding to the same tonal r e f l e x that i s i n the l i t e r a r y 51 word fan )\> /fa:n /. Of the Yue di a l e c t s mentioned above, only i n Zhong-shan has the regular tonal r e f l e x of fan ̂  been preseryed. 33 33 The tones i n Cantonese [p A x 1A:I) ], for example, dxd not pre- 21 serve the Yang-ping [21] tone of fan [ f A i n ]• What i s s t i l l not e n t i r e l y clear i s the role of the s y l l a -51 51 51 ble /ham / i n Zhong-shan /ham plarng /. I t i s possible that the nasal segment /m/ serves to preserve the voicing of the h i s - t o r i c a l voiced *b- i n i t i a l , when voicing was no longer phonologic- 51 23 a l l y d i s t i n c t i n the i n t i a l segment /p/ of the s y l l a b l e /pla:ng / The observation by McCoy and l a t e r by the writer that /p/ i s phone- t i c a l l y [bj i n the modern c o l l o q u i a l word ' a l l ' may not be fortu- 51 33 itous a f t e r a l l . The s y l l a b l e /ham / C [ h - e m ] i n Cantonese) may be regarded as an i n t r i n s i c part of the h i s t o r i c a l reconstruction of the word fan : i t bears the voicing feature which would otherwise have been l o s t when *b- no longer contrasted with the corresponding voiceless l a b i a l stops i n the phonological system of the d i a l e c t . Another c o l l o q u i a l word i n Zhong-shan which can be associ- ated with a word which once contained an i n i t i a l consonant c l u s t e r 22 22 51 22 51 i s /ka l a wa:ng /, or /k l a wa:ng / 'crosswise; at ri g h t 33 angles to'. Long-du has a si m i l a r form for 'crosswise': /kaa 33 33 laa waaq / (tone numerals are used instead of Egerod's tone - 88 - d i a c r i t i c s , and /q/ - [IJ]) O O \% (Egerod, p.91). . Jv Norman (personal communication) does not recognize t h i s word as a regular Min colloquialism, and can only presume that i t i s borrowed from 24 Zhong-shan. Ignoring tonal differences between the two d i a l e c t a l forms for the moment, one can propose that the p o l y s y l l a b i c form of the word 'crosswise' i n both d i a l e c t s contain traces of an o r i g i n a l i n i t i a l consonant cl u s t e r . More precisely, the complete 22 22 51 33 33 forms of Zhong-shan /ka l a wa:ng / and Long-du /kaa laa waaq 3 3/ can be connected to the word heng , which i s a Xia i n i t i a l , Grade II word reconstructed by F.K. Chou, .for example, 22 as Archaic Chinese *grwang. The sesquisyllable / k l a / then pre- 51 serves the i n i t i a l . *gr- cluster, and /wa:ng / i s simply the s y l l a b l e - f i n a l of the word. Regarding i n i t i a l *g- i n the cl u s t e r , phonologists have observed that at lea s t some of the words containing the Xia i n i t i a l i n Middle Chinese had common origins with velar stops, as r e f l e c t e d i n the reconstruction of i n i t i a l *g- i n Archaic Chinese by Chou for heng jfll , for example, and an aspirated *g--- by Karlgren. In terms of h i s t o r i c a l phonology, the Zhong-shan po l y s y l l a b i c form for 'crosswise' thus lends support for the reconstruction of an e a r l i e r i n i t i a l c l u s t e r which contains a yelar element. More generally, the Zhong-shan p o l y s y l l a b i c form provides evidence for an i n t i a l consonant c l u s t e r for the word Turning now to the question of the tonal values recorded for the word 'crosswise' i n Zhong-shan as well as i n Long-du, one observation should f i r s t be made concerning heng (Zhong-shan 51 33 /wa:ng /, Long-du /waaq / ) . Both d i a l e c t a l forms are regular reflexes of the h i s t o r i c a l Yang-ping tone with respect to t h e i r respective pattern of correspondences to the h i s t o r i c a l phonolo- g i c a l system. Thus, the tones i n a l l three s y l l a b l e s i n the Long- 33 33 33 du form /kaa laa waaq / are completely regular. In Zhong- shan, while / 5 l / i s the regular r e f l e x of Yang-ping, /22/ i n the 22 22 22 s y l l a b l e s /ka l a / and the sesquisyllable / k l a / i s not. One possible explanation for the exceptional tonal r e f l e x i s that 51 51 /ka / and / l a /, which would be the expected reflexes for 51 Zhong-shan, are extremely rare. In the present data, /ka / only occurs i n a c o l l o q u i a l , characterless word which refers to a leg- 51 astride, standing po s i t i o n , while / l a / occurs i n the combination 51 55 it: / l a a:j / Q meaning ' d i r t y 1 , and as a sentence-final p a r t i - c l e . In contrast, tone /22/ counterparts of these s y l l a b l e s are much more numerous, and can be found i n regular l e x i c a l items. 51 51 Perhaps the r a r i t y of s y l l a b l e s /ka / and / l a / motivated the tone change to /22/ i n Zhong-shan. No firm conclusions w i l l be attempted at t h i s time. In the next word, Zhong-shan has two c i t a t i o n forms and one corresponding sesquisyllabic form. To describe an instrument or object that i s very straight, or a road or route that i s very 22 5 5 straight or d i r e c t , Zhong-shan uses the phrase /tim pat l a t / 22 5 5 %, O O , or /tim pat pat /. The sesquisyllabic form i s 22 5 22 /tim pl a t /. For the f i r s t s y l l a b l e /tim /, Cantonese likewise 33 has the word [ti:m J meaning 'straight', and coined .the character to represent that c o l l o q u i a l word. 5 What interests us here i s the sesquisyllable / p l a t / which one would l i k e to associate with the word b i 3p (153-12) 'a writing brush; straight, d i r e c t ' , which has been reconstructed - 90 - w i t h an i n i t i a l * p l - c l u s t e r . K a r l g r e n , f o r example, r e c o n s t r u c t s b i -jt as A r c h a i c Chinese * p l i s t ; Chou r e c o n s t r u c t s i t as * p l i w 9 t . A * p l - c l u s t e r h a s : a l s o been p o s t u l a t e d by P u l l e y b l a n k (1962:111) f o r b i , which i s a Grade I I I word be long ing to the s o - c a l l e d chong -n iu J[ £3- (..'double k n o t ' ) or Grade I I I/XV doub le t f i n a l s . X i e - s h e n g evidence i n Midd le Chinese shows contac t of phonet i c "̂p w i t h both i n i t i a l *p- as i n b i / p a t 5 / and * 1 - i n l u ̂jp 2 (.162-10) ' a law' /10t / . P u l l e y b l a n k a l s o brought the w r i t e r ' s a t t e n t i o n to h i s t o r i c a l ev idence of b u - l u ^ ^ (Zhong-shan 5 2 VK /pat IfSt /) be ing used to render the p r o n u n c i a t i o n of b i jE" . I t would appear t h a t i n Zhong-shan, the i n c l u s i o n of j$i / t i m 2 2 / meaning ' s t r a i g h t ' was used to r e i n f o r c e the meaning of / p l a t 5 / as the a s s o c i a t i o n of b i Jjjf to the * p l - i n i t i a l c l u s t e r became opaque a f t e r the genera l l o s s of i n i t i a l c l u s t e r s i n Ch inese . E v i d e n t l y , s y l l a b l e s c o n t a i n i n g / p i / then became viewed as some s o r t of onomatopoeic s y l l a b l e . One should suspect 22 5 5 22 5 t h a t of the two c i t a t i o n forms / t i m pat l a t / and / t i m pat 5 pat / , the l a t t e r i s a more recent i n n o v a t i o n , where the o r i g i n a l 5 5 5 meaning of /pat l a t / or / p l a t / i s complete ly obscured . Somewhat more t e n t a t i v e i s the l i n k a g e betwen the Grade I I I , He-kou word juan (142-13) / k y n 1 3 / or / k u n 1 3 / ' a r o l l ' and the c o l l o q u i a l word f o r ' s o m e r s a u l t ' s u p p l i e d by K a r l L o , 55 55 13 *L another Zhong-shan speaker : /kwa:n l a : n taw / O O >y i n 55 13 c i t a t i o n form and / k l a : n taw / i n reduced form. A l though no i n i t i a l * k l - c l u s t e r i s u s u a l l y r e c o n s t r u c t e d f o r the word juan ( e . g . , K a r l g r e n : A r c h a i c Chinese *kiwan) due to the absence of x i e - s h e n g e v i d e n c e , i t might be observed t h a t an i n i t i a l * k l - was hypothes i zed by P u l l e y b l a n k (1962:126) f o r the word. In t h i s - 9.1 - case, Pulleyblank treats *-.!- as a derivational i n f i x based on Wulff's theory of a morphological i n f i x * 1 . 2 5 Thus, juan 'turn around' i s Middle Chinese *kiwan, whereas juan ^ ' r o l l ' i s Middle Chinese *kiwen < Old Chinese *kwl5n. 55 55 55 If Zhong-shan /kwa:n la:n / and /kla:n / can be linked to the word juan , i t i s int e r e s t i n g that the c i t a t i o n form would be the one that i n fact preserves the l a b i a l medial. In this case, one would expect that the c i t a t i o n form and the sesqui- s y l l a b i c form both evolved simultaneously from the dimidiation of the i n i t i a l c l u s t e r , with the c i t a t i o n form preserving the He-kou feature. Conceivably, at an e a r l i e r stage, the sesqu i s y l l a b i c form had i n i t i a l *kw 3l-. The f i n a l i n the Zhong-shan c o l l o q u i a l expression s t i l l presents a problem since the regular Zhong-shan re f l e x for juan i s /yn/. The writer w i l l leave t h i s problem for future investigation. F i n a l l y , there i s also the l i t e r a r y word g_e f[%T (170-10) 2 'armpit' /kok /, which Karlgren has reconstructed as Archaic Chinese *klak. A * k l - c l u s t e r has also been postulated by other phonologists, sometimes with the h i s t o r i c a l l y homophonous word ge ^- (170-7) 'each' /kok / representing ge and others i n th i s set of Grade I words. Thus, L i (1974:251) and Pulleyblank (.1962:119), for example, both posit an o r i g i n a l * k l - i n i t i a l c l u s t e r for g_e ̂  : Archaic Chinese *klak by L i , and Old Chinese *kiak (since revised as *klak) by Pulleyblank. C o l l o q u i a l data from a number of Chinese d i a l e c t s (cf. esp e c i a l l y Yang, 1971) strongly supports the postulation of an i n i t i a l consonant clu s t e r for ge . Consider f i r s t the Cantonese example. Included among ^he various c o l l o q u i a l forms for 'armpit' i n Cantonese i s that of [ k A i k * * lAik - 1 t - B j J J ] (Hashimoto, pp.242, 330), for which Hashimoto uses the characters ge-le-di Wh/fc* • P a r a l l e l to the case of the word 'corner', one can postulate that 4 5 the f i r s t two s y l l a b l e s , i.IkAik J and IlAi.k ], both represent the word ge , which o r i g i n a l l y bore a * k l - c l u s t e r i n i t i a l , as reconstructed by Karlgren and others. (The s u f f i x d i ̂  , which means 'underside, base', does not enter into the reconstruction of the word g_e j ^ ^ o " . ) Also p a r a l l e l to the case of the word 'cor- ner' i s the tone change i n the second s y l l a b l e : the basic tone of the character l e faf) i s Zhong Yin-ru /4/ ( i . e . , { l A i k 4 ] ) . As argued i n the word 'corner', the character used to represent the second s y l l a b l e i s e s s e n t i a l l y a dummy element with both the f i r s t and the second s y l l a b l e having been derived from the f i r s t character h i s t o r i c a l l y . Dong-guan, another Yue d i a l e c t , has a c o l l o q u i a l form for 'armpit' which i s very close to the Cantonese form; Ikak lak haj M/̂  "f (from Yang, 1971—no tone marks provided) . (The s u f f i x xi a "p means 'below, under', and does not pa r t i c i p a t e i n the reconstruction.) The same arguments used i n the Cantonese form for postulating a * k l - c l u s t e r can be applied here. Zhong-shan also has a c o l l o q u i a l form for 'armpit' which reveals an i n i t i a l * k l - c l u s t e r . The Zhong-shan form, however, had undergone a t t r i t i o n of the f i n a l /k/ ending and p i t c h - r a i s i n g 55 55 22 /-\ ~r" to a high tone y i e l d i n g /ka l a ha / O O r as the c i t a t i o n 55 22 form, and / k l a ha / as the sesquisyllabic form. I t i s in s t r u c - t i v e to know that the Fu-zhou form (supplied by Norman). , l i k e the Zhong-shan one, shows loss of the stop ending, and i s phonetically [ k o 2 2 r o u ^ a 2 4 2 ] }%T~^ • Since i t i s at y p i c a l for Zhong-shan - 93 - to lose i t s stop endings, i t i s possible that the a t t r i t i o n of f i n a l /k/ i n the Zhong-shan form may be due to influence from Fu-zhou or another Northeastern Min d i a l e c t . On the basis of comparisons with, the Cantonese, Dong-guan and Fu-zhou forms, one can be quite confident that the Zhong-shan form likewise repre- sents an i n i t i a l consonant c l u s t e r for the word g_e jj^j- . The writer i s actually not the f i r s t to use p o l y s y l l a b i c , c o l l o q u i a l forms to support the reconstruction of a * k l - c l u s t e r for g_e . Among the pioneers to do so i s Yang (1971) who has not only c o l l e c t e d such forms from a number of Chinese d i a l e c t s , but has also included cognates i n other East Asian languages i n order to support arguments for an e a r l i e r existence of a * k l - cl u s t e r i n the word g_e . 'armpit'. Thus, cognates which Yang found for the word 'larmpit' include kliak i n Khmer (.Cambodian) and ke1ek i n Malay. Pulleyblank (personal communication) also proposes that 55 the c o l l o q u i a l , p o l y s y l l a b i c word i n Mandarin for 'armpit' [ka 5 5 t§\ WUD ] may be i d e n t i f i e d with the word ge jffir (_*klak) or a closely related form. For instance, i n what Pulleyblank c a l l s his Type B s y l l a b l e s (for our present purposes, they are Grade III and IV s y l l a b l e s which he reconstructs as ..containing medial * - i - i n the Kai-kou series i n LMC), the development of Old Chinese *klak i s *klak > *k-^.iajk > k-tcsiajk > k-tcsi (reconstruction h i s ) . A velar for the i n i t i a l consonant and an a f f r i c a t e (palatal or retroflex) for the second s y l l a b l e seems to be quite prevalent among the Mandarin d i a l e c t s . The same pattern is, also found i n Su-zhou (Wu) . (See, e.g., Yang, 1971; Han-yu Fang-yan Ci-hui (.'A c o l l e c t i o n of Chinese d i a l e c t a l words') by the same editors as the - 94 - Zi-hui (hence, hereafter ' Ci-hui' for short) (1964:195);.) Concluding the subject of ge , i t should be noted that although the Zhong-shan c i t a t i o n form / k a 5 5 l a 5 5 h a 2 2 / merely adds to Yang's already long l i s t of p o l y s y l l a b i c forms i n various 55 Chinese d i a l e c t s for 'armpit', the sesquisyllabic form / k l a / i s s i g n i f i c a n t i n that i t may attest to an e a r l i e r stage of the evo- l u t i o n of the consonant cl u s t e r . In the present as well as pre- viously discussed reduced forms, one may regard the schwa inserted between the i n i t i a l consonants i n the sesquisyllable as the f i r s t stage i n the anaptyxis process; that i s , the ses q u i s y l l a b i c struc- ture may represent the stage at which adjacent consonants i n s y l l a - b l e - i n i t i a l position were separated by means of schwa-epenthesis, thus creating a sesquisyllable. Subsequent changes led to tone alterations i n some cases, and a t t r i t i o n of segments i n others. The breaking up of the i n i t i a l consonant c l u s t e r , nonetheless, i s captured and preserved i n the sesquisyllabic form. As observed i n McCoy's recording of the c o l l o q u i a l word ' a l l ' and Hashimoto's recording of the same word i n Teng-xian and Tai-shan, i t appears that the sesquisyllable exists i n other Yue d i a l e c t s besides Zhong-shan, and may i n fact be present i n non- Yue d i a l e c t s also. The existence of the c i t a t i o n form, which i s the form normally recorded i n fieldwork, i s then a further progres- sion away from the i n i t i a l c l u s t e r s . As suggested e a r l i e r , the c i t a t i o n form e s s e n t i a l l y represents an o r i g i n a l sesquisyllable which had been made to conform to the more pre v a i l i n g CVC structure. - 95 - 1.4.4. Co l l o q u i a l versus L i t e r a r y Forms Among many Chinese d i a l e c t s , there i s a d i s t i n c t i o n between what are recognized as c o l l o q u i a l , or bai \;D , forms and what are regarded as l i t e r a r y , or wen , forms of words which are phono- l o g i c a l l y and semantically related. Generally speaking, the ... :c l i t e r a r y forms r e f l e c t influence from the standard language. In the case of a d i a l e c t which i s influenced by both a regional and a national standard, or one which incorporates forms from one or more dia l e c t s at d i f f e r e n t stages i n i t s history, several forms for a word may e x i s t side by side, r e f l e c t i n g the various sources or layers of borrowing. One example i s the Min d i a l e c t i n southern China i n which one often encounters as many as three layers of pronunciation for a given word. As i n Cantonese, Zhong-shan has two major categories of c o l l o q u i a l morphemes. The f i r s t category consists of purely c o l l o - q u i a l terms which generally lack character representation, with the exception of those special characters coined for Cantonese which can also be used i n Zhong-shan to represent c e r t a i n character- less words. A number of these c o l l o q u i a l Zhong-shan terms are included i n the lexicon i n Chapter 3li The second major category of c o l l o q u i a l morphemes consists of characters which have both a c o l l o q u i a l and a l i t e r a r y reading. The two forms can normally share the same syntactic environment. The c o l l o q u i a l form i s normally used i n d a i l y or informal speech, while the l i t e r a r y form only occurs i n formal speech or i n the reading of written texts. In general, what i s analyzed as the c o l l o q u i a l reading r e f l e c t s an e a r l i e r layer of the Chinese lan- guage, and the l i t e r a r y reading a l a t e r layer. - 96 - The d i s t r i b u t i o n of the c o l l o q u i a l versus l i t e r a r y forms i s , as Hashimoto (p.118) emphasizes, not necessarily exclusive. This i s p a r t i c u l a r l y the s i t u a t i o n with regard to p o l y s y l l a b i c words which may u t i l i z e a c o l l o q u i a l form i n one combination and a l i t e r a r y form i n another. Thus., c o l l o q u i a l l y , a 'provincial however, i s cheng-shi ^ j 1 /sing"""1" si" 1"""/; here, the l i t e r a r y pronunciation of cheng i s used instead. There i s no known general rule governing the choice of a c o l l o q u i a l rather than a l i t e r a r y form (or vice versa) i n a p a r t i c u l a r combination. There are also morphemes which do not have s t y l i s t i c alternants, with one form »£. 55 serving a l l functions. The morpheme qing (2-0-7) / t s ' i n g / 'clear, pure', for instance, only has the l i t e r a r y form regardless of usage or combination, whereas the morpheme jin g (202-1) 1 3 /kia:ng / 'neck' only has a c o l l o q u i a l form. c o l l o q u i a l and l i t e r a r y forms, i t i s obvious that the term " c o l l o - q u i a l " does not apply s p e c i f i c a l l y and solely to "spoken" forms, any more than does the term " l i t e r a r y " apply only to "written" forms. Hashimoto (p.119) contends that although the d i v i s i o n of items i n terms of c o l l o q u i a l versus l i t e r a r y i s based primarily on an actual s t y l i s t i c d i s t i n c t i o n between the two sets of forms, t h i s d i v i s i o n should be considered a phonological one since items not only con- form to s t y l i s t i c d i s t i n c t i o n s , but they also follow c e r t a i n patterns of phonological behaviour. I t i s the phonological c r i - t e r i o n which i s p a r t i c u l a r l y important both i n proposing general statements about the h i s t o r i c a l development of Cantonese and Zhong- Given the non-exclusiveness i n the d i s t r i b u t i o n of the - 97 - shan, and i n formulating reconstructed forms for proto-Yue. The h i s t o r i c a l significance of the c o l l o q u i a l versus l i t e r a r y readings w i l l be c l a r i f i e d i n the second part of t h i s study. For the present, three main types of c o l l o q u i a l verus l i t e r a r y forms w i l l be shown i n t h i s section. In the f i r s t set, there i s an alternation of i n i t i a l s and tones. The morpheme containing an aspirated i n i t i a l together with tone /13/ i s the c o l l o q u i a l form, while the morpheme with the unaspirated form i n tone /22/ serves as the l i t e r a r y counterpart. Examples are presented i n Table 4 (a). The second set of alternants r e f l e c t i n g c o l l o q u i a l and l i t e r a r y readings i s /a:ng/ versus /ang/ f i n a l s . (This includes the corresponding stop f i n a l s . ) The c o l l o q u i a l forms contain f i n a l /a:ng/ and the l i t e r a r y forms f i n a l /ang/. A l i s t of such alternations are given i n Table 4 (b). The t h i r d and f i n a l set shows an alternation between /ia:ng/ and /ing/ f i n a l s (and between the corresponding stop f i n a l s ) . The /ia:ng/ f i n a l i s used i n c o l l o q u i a l speech and the / i n g / f i n a l i s used for the reading pronunciation. A l i s t of such s t y l i s t i c alternations i s presented i n Table 4 (_c) . Syllables with the corresponding stop consonant endings are also included i n the three tables. - 98 - Table 4 (a)• Co l l o q u i a l versus L i t e r a r y Forms: Tone /13/ Aspirated I n i t i a l and /22/ Unaspirated I n i t i a l , Number Word Co l l o q u i a l L i t e r a r y Gloss 5-10, / t s ' o 1 3 / / t s o 2 2 / 'to s i t ' 6-9 29-12 96-9 / t s ' o j 1 3 / / t s o j 2 2 / 'to be located, at' 95-13, J* / f a : m 1 3 / /ta:m 2 2/ ' l i g h t ' 134-18 igft / t ' y n 1 3 / / t y n 2 2 / 'broken' 155- 10, j/t /k'an 1 3/ /kan 2 2/ 'near' 156- 2 Table 4 (b). Col l o q u i a l versus L i t e r a r y Forms: /a;ng/ and /ang/ Fi n a l s . Number Word Col l o q u i a l L i t e r a r y Gloss 184-21 t i /ha:ng 5 5/ a /hang 5 1/ ' constant' 192-4 4. 55 /sa:ng / , 55 -/sang / ' raw' 192-5 55 /sa:ng / / 55 , /sang / 'animal' 192-8 55 /ka:ng / /kang 5 5/ 'watches of the night 192-13 -? •5 55 ,h /ha:ng / /hang 5 5/ 'to pervade' 192-14 ft /ha:ng / /hang 5 1/ 'to walk' 192-16 /ma:ng13/ /mang 1 3/ 'f i e r c e ' 192-18 /la:ng / / l a n g 1 3 / 'cold' 195-5 f 55 /tsa:ng / /4. 55 , /tsang / 'to struggle' 195-7 22 /ts'a:ng / /*. 55, /tsang / 'to open the eyes' 195-8 to 55 /ka:ng / /kang 5 5/ 'to farm' a In the name of a v i l l a g e i n the f i r s t qu i n Zhong-shan county /ha: 55 .13, ng mi /. k In the name of a neighbouring v i l l a g e •Z P 55 .13 , : ^ /ha:ng mi /. - 9 9 - Table 4 (c) . C o l l o q u i a l versus /ing/ F i n a l s . Number Word Col l o q u i a l 1 9 7 - 1 4 5 1 /p'iarng / 1 9 8 - 3 /kia:ng / 1 9 8 - 1 9 9 - 1 5 1 / • 1 3 , /ja.ng / /pia.ng / 1 9 9 - 2 / • 2 2 . /pxasng / 1 9 9 - 3 2 2 /mia:ng / 1 9 9 - 4 . . . 2 2 , /kia:ng / 1 9 9 - 6 / i • 2 2 , /kia:ng / 2 0 0 - 2 / • 5 1 . /mia:ng / 2 0 0 - 3 5 5 /tsia:ng / 2 0 0 - •9 , 4 - . • 5 1 , /ts 'xa:ng / 2 0 0 - •19 / • 5 5 , /sia:ng / 2 0 0 - •20 ^ » , . 5 1 / /sia:ng / 2 0 0 - •21 / • 5 1 / /sxa:ng / 2 0 1 - •1 / u - 5 5 , /hia:ng / 2 0 1 - •5 * / • 5 1 . /Da:ng / 2 0 1 - •7 /nia:ng / / - . • 1 3 / /lxa:ng / 2 0 1 - •8 / T I 3 / /lxarng / 2 0 1 - •9 / 4 . • 1 3 / /tsxa:ng / 2 0 1 - - 1 0 *A 2 2 /tsViarng / 2 0 2 - - 5 2 2 / t s i a r n g / 2 0 2 - - 7 / • 2 2 , /sxa:ng / 2 0 2 - - 9 2 2 /tsia:ng / 2 0 3 - - 1 0 • • f t / s i a : k 2 / L i t e r a r y Forms: /ia:ng/ and Lit e r a r y Gloss /p'xng / • f l a t ' . 5 5 / /kxng / 'to fear, a f r a i d ' / • 1 3 / /xng / 1 shadow1 / p i n g 2 2 / •handle' 2 2 /ping / 1 s i c k 1 /ming 2 2/ • l i f e ' ... 22, /kxng / 'to respect, be c a r e f u l ' / k i n g 2 2 / 'mirror' , . 5 1 / /mxng / 'name' 5 5 / t s i n g / 'sharp, quick-witted' . • 5 1 . /ts 1xng / 'clear sky' / • 5 5 / /sxng / 'sound, voice' / • 5 1 , /sxng / 'accompli shed, ent i r e ' . . 5 1 . /sxng / •c i t y ' / k ' i n g 5 5 / ' l i g h t (e.g., i n weight)' . . 5 1 / /xng / 'to win' / l i n g 1 3 / ' c o l l a r ' / l i n g 1 3 / 'to apply for' / l i n g 1 3 / • h i l l ' /a. • 1 3 / /tsxng / •a well' - 2 2 / /ts Vxng / 'to i n v i t e ' / * . • 2 2 / /tsxng / 'clean' / • 2 2 / /sxng / 'surname' ix. • 2 2 / /tsxng / 'upright' / s i k 5 / 'to be fond of' - 1 0 0 - 2 0 3 - 1 6 1 2 / t s i a . k / / t s i k 2 / 'to burn 1 2 0 3 - 1 7 * / t s ' i a . k 2 / / t s ' i k 2 / 'red' 2 0 4 - 8 / p ' i a : n g 5 5 / / . • 5 5 , /P ing / 'to r i s k (e.g. one's l i f e ' 2 0 4 - 9 / i • 5 1 -a /p'ia:ng / /p'xng / 'bottle' 2 0 4 - 1 6 /tia:ng / / t i n g 5 5 / • n a i l ' 2 0 4 - 2 0 / t 1 i a : n g 5 5 / / t ' i n g 5 5 / 'to l i s t e n , hear, obey' 2 0 4 - 2 3 % / * . ! - 5 1 , b /t'ia:ng / / f i n g 5 1 / 'pavilion' 2 0 5 - 2 0 5 - 2 3 % / l i a : n g / /lxa:ng / / i - 5 1 , / l i n g / / i - 5 1 , / l i n g / 'efficacious' 'remainder, plus, over' 2 0 5 - 8 \ / t s ' i a : n g 5 5 / / t s ' i n g 5 5 / 'blue, black' 2 0 . 5 - 1 1 KL / • 5 5 , /sia:ng / , . 5 5 , /sing / 'smelly (of f i s h , blood)' 2 0 6 - 1 /tia:ng / / t i n g 1 3 / ' top;.(classifier for hat)' 2 0 6 - •7 $i /ts'ia:ng / / • 1 3 / /sing / 'to awake' 2 0 6 - •8 fa / * - • 5 5 , /tia:ng / / t i n g 5 5 / 'to n a i l ' 2 0 6 - 9 /tia:ng / / t i n g 2 2 / 'to subscribe' 2 0 6 - •10 5 5 /t'ia:ng / / f i n g 5 5 / •to l e t ' 2 0 6 - •11 . . . 2 2 . /tia:ng / / t i n g 2 2 / 'to f i x , to order' 2 0 7 - •25 2 /sia:k / / s i k 5 / • t i n ' Additional morphemes not i n the d i a l e c t survey l i s t : 1 . / l i a : n g 5 1 / / l i n g 5 1 / '(a kind of f i s h ) ' 2 . ^ / t s i a : k 2 / / t s i k 2 / 'straw mat' 3 . / l i a : k 2 / C / l i k 5 / 'to drip, to t r i c k l e ' Only i n the combination /jaw p'ia:ng / 'child/brought to a second marriage'. In such combinations as £dLi •§* / t i tVia:ng / ' the ground i n front of the house i n the v i l l a g e s used for drying grains, etc.' . c In the names of two v i l l a g e s , }$L / t a : j 2 2 l i a : k 2 / and / s a J l i a : k / (from Chao). - 101 - Notes to Chapter 1 1. The segment c a l l e d the "medial" was probably f i r s t extracted from the f i n a l by Bernhard Karlgren, whose reconstruction of Chinese remains the groundwork fo r most studies on h i s t o r i c a l Chinese phonology i n t h i s century. 2. B a l l ' s system of tr a n s c r i p t i o n i s not phonetic, nor i s i t t r u l y phonemic since there are a number of redundancies.X. However, rather than introduce another set of bracketing notation that would be purely ad hoc, B a l l ' s system w i l l be treated as i f i t i s phonemic. 3. The character n i yfcj has conventionally been used by Chinese phonologists to represent a class of words (of which i t i s a member) containing an i n i t i a l which has been reconstructed as *n. The names of other i n i t i a l s are derived i n a sim i l a r manner—a member containing a certa i n i n i t i a l i s used to represent the group possessing the same i n i t i a l . 4. A discussion of " c o l l o q u i a l " and " l i t e r a r y " readings of characters i s presented i n section 1.5. 5. This complementary d i s t r i b u t i o n of the dental nasal and the l a t e r a l i n i t i a l s i n Amoy was brought to the writer's atten- t i o n by Pulleyblank. 6. The semi-circle, which may or may not be underlined, i s used by some t r a d i t i o n a l Chinese phonologists to mark tone. The presence of underlining indicates the Yang r e g i s t e r , and i t s absence the Yin r e g i s t e r . The semi-circle at the bottom-left corner of the tr a n s c r i p t i o n designates the Ping tone, top- l e f t corner the Shang tone, top-right corner the Qu tone, and bottom-right corner the Ru tone. The names of the tones - 102 - refer to h i s t o r i c a l phonological c l a s s i f i c a t i o n s , while the Yin-Yang dichotomy r e f l e c t s reflexes of h i s t o r i c a l i n i t i a l s . These terms w i l l be further elucidated i n t h i s study at the appropriate place. To conform with the present study, Chao's numerical system for tones rather than his tone l e t t e r s i s used for t r a n s c r i b - ing h i s Zhong-shan data. The velar nasal i s transcribed as /ng/ for typographical reasons. The zero i n i t i a l /#/ should not be confused with the vowel /0/ , which w i l l be introduced l a t e r , since the zero i n i t i a l never occurs i n the transcriptions. Moreover, they do not occur i n the same environment. As i n the case of the names of the i n i t i a l s (see footnote 3 above), the character used to represent a given rhyme group also belongs to that group. Thus, Geng i s both the name used for the rhyme group and i s i t s e l f a member of that rhyme group. Furthermore, a rhyme group i s composed of a number of "rhymes", which also possess names established by convention. As i n the above cases, the character bearing the name of the rhyme i s also a member of that rhyme. The f i n a l s h i s t o r i c a l l y exhibit a four-way contrast of deng ^ 'grade, d i v i s i o n ' (Karlgren translates the term as ' d i v i - sion', and Pulleyblank translates i t as 'grade'). P r i o r to Pulleyblank's (1970-71) theory of the system of the four grades, i n which the contrast among the grades i s i n terms of phonological d i s t i n c t i o n s i n the medials, reconstruction of the grades had been accounted for p a r t i a l l y i n terms of d i s t i n c t i o n s i n the medial and p a r t i a l l y i n terms of d i s t i n c - - 103 - tions i n the rest of the f i n a l . The system of the grades w i l l be explained i n greater d e t a i l i n chapter 2. For the present, l e t i t s u f f i c e that the grade system i n Pulleyblank.'s (1970-71) analysis concerns a four-way d i s t i n c t i o n of the medials involving the presence or absence of p a l a t a l i z a t i o n . 11. The Zi-hui i s not s t r i c t l y phonetic. Length, for example, i s not marked, nor are glides consistently kept d i s t i n c t from vowels. Moreover, i n i t i a l glides which can be predicted from features i n the following vowel may be omitted, as i n the various Mandarin d i a l e c t s . 12. Thanks i s due to Pulleyblank for explaining the symmetry . i between I i ] and [ u ], and hence providing additional support for the phonemic analysis proposed here i n which they are treated as corresponding high vowels, / i / and /u/ respectively, 13. The sequence [ob] i s presented as the p r i n c i p a l form for the c l u s t e r /ua/ i n Egerod's vowel chart, with l o i j included i n round brackets. However, only the sequence [UD] occurs i n the discussion of the various phonemes i n the d i a l e c t . This l a t t e r form i s probably a typographical error, and should i n fact be [ob] , which only appears on the vowel chart. 14. Unless stated otherwise, Cantonese pronunciations follow the phonetic t r a n s c r i p t i o n used by Hashimoto. A few minor changes, are made i n order that her t r a n s c r i p t i o n system conform with the present one; e.g., Hashimoto's glides [1], lyj and [u] are transcribed as [ j ] , IH] and Iw] respectively, and her [k] and [k'] are transcribed as {kw] and Ik'w]. w Regarding [kw] and Ik'w], whether one wishes to consider them as consisting of a sequence of sounds or a single l a b i a l i z e d - 104 - segment i s b a s i c a l l y a phonological decision. In terms of h i s t o r i c a l phonology, i t i s advantageous to have the l a b i a l element separate from the i n i t i a l and treated as a medial. Kai-He d i s t i n c t i o n s do not a f f e c t the h i s t o r i c a l i n i t i a l s but they do the medials, such that the l a b i a l element occurs in He-kou series, for example, whereas the p a l a t a l element (e.g., Zhong-shan medial / i / ) occurs i n the Kai-kou s e r i e s . Synchronically, i n terms of minimizing the t o t a l number of i n i t i a l s and f i n a l s , i t i s i n fact more economical for both Cantonese and Zhong-shan to treat the l a b i a l segment as part of the i n i t i a l by creating an extra pair of i n i t i a l s , namely l a b i a l i z e d velars [kw] and [k'w]. Zhong-shan also has to contend with a p a l a t a l segment which has a very d i f f e r e n t d i s t r i b u t i o n from /w/. Medial / i / co-occurs with less than half of the rhymes ( i . e . , f i n a l minus the medial), but with almost a l l the consonantal i n i t i a l s . In t h i s l a t t e r case, i t i s more economical to treat the p a l a t a l medial as part of the f i n a l than to create a p a l a t a l i z e d series of i n i t i a l s , which would double the present inventory of i n i t i a l 3 * Thus, an alternate solution to what has been proposed here that would be more sympathetic to a synchronic analysis would be to create a l a b i a l i z e d velar series for Zhong-shan, and recognize one medial only—namely, the v o c a l i c medial / i / — which would be part of the f i n a l . A comparable solution i s i n fact adopted i n setting up the Zhong-shan syllabary and the lexicon of c o l l o q u i a l Zhong-shan terms i n chapter 3. The segments [kw] and Ik'w] are treated there are part of the i n i t i a l . - 105 - 15. Obviously, among h i s t o r i c a l Chinese phonologists, there i s not always consensus i n interpretations and reconstructions. This i s a case i n point. J. Norman (personal communication), for instance, does not believe, as do Karlgren, Chao and Pulleyblank, that there was a d i s t i n c t i o n of medials between guan j|J and guan ^ . Norman proposes that guan ^* (*kuan). passed through a stage *kuon and then became [ku:n] i n Can-, tonese and Zhong-shan. Guan j|ifj (*kuan) , on the other hand, Norman feels did not evolve i n the same way because the nuclear vowel was front and hence much more re s i s t a n t to rounding. Thus, Norman posits the difference between the two words i n terms of a difference i n the vowel rather than a difference i n the medial. 16. Chao's medial / i / i n his s y l l a b l e s / i a / , / i a i t j / and /ia:k/ i s the f u l l vowel [ i ] . These s y l l a b l e s are transcribed i n the present study as / j a / , /ja:ng/ and /ja:k/ respectively, with / j / analyzed as a f u l l glide i n i t i a l . 17. Norman pointed out that lip-rounding of the i n i t i a l before [u:] also occurs i n Peking Mandarin and probably i n most other Chinese d i a l e c t s . 18. Sheng also has the dictionary d e f i n i t i o n of 'sound, voice'. 19. Chao (p.68), on the other hand, records the phrase 'tomorrow morning' as ^ ^ [men.51 t s i :w55 tsow 1 3] , with ming 0$ 51 pronounced [men ] i n t h i s context. 20. The writer i s indebted to David Strecker for bringing to her attention the concept of the "sesquisyllable" discussed i n Matisoff's a r t i c l e . Another term for the phenomenon i s that of the "minor s y l l a b l e " , which i s used i n England (Pulley- - 106 - blank, personal communication), 21. Pulleyblank informs the writer that the term "dimidiation" was used by Broodberg to re f e r to his theory that such rhym- ing compounds as kun-lun ^ , for the Kunlun Mountains i n Tibet, were derived from monosyllables; thus, kun-lun f v ̂  , for example, would be derived from *klun, etc. 22. Except for r e t a i n i n g the following i n McCoy,'s t r a n s c r i p t i o n system: schwa [ a ] , symbol for the neutral tone [°], recording of a voiced [b],and use of a hyphen to indicate close junc- ture, his data has been phoneticized based on Hashimoto 4s system. A couple of variant forms for 'corner' recorded by McCoy (1966:185,fn.27) are: [ k a ° - l D i k 3 5 ] and [ka°- l o x k 4 ] . 51 23. The i n i t i a l suggestion for the s y l l a b l e /ham / i n Zhong-shan was made by Pulleyblank. 24. Because Long-du i s spoken i n the Zhong-shan county and i s moreover i n the qu, just west of Shi-qi (Qu II on Map 2) , not only does the phonological system of (Shi-qi) Zhong-shan have the unique status of forming the Long-du l i t e r a r y pronuncia- tions , but many Shi-qi forms have also been incorporated into the c o l l o q u i a l layer of the Long-du d i a l e c t . 25. Chinesisch und T a i , by K. Wulff, Copenhagen: Levin and Munksgaard, 1934. - 107 - CHAPTER 2 . SYLLABARY AND LEXICON A syllabary representing the repertoire of Zhong-shan sy l l a b l e s i s presented i n the following section. Wherever possi- ble, a character i s used to represent a given s y l l a b l e . Single underlining of a character indicates c o l l o q u i a l reading, and double underlining s p e c i f i e s l i t e r a r y reading. C o l l o q u i a l words with no character counterpart (or only coined characters to repre- sent them) are marked using a c i r c l e " O " , except for onomatopoeic sy l l a b l e s and foreign loans. The l a t t e r are distinguished by the use of an "X" i n the syllabary since what i s normally a rare or impermissible sequence can occur i n an onomatopoeic expression or borrowed term. However, i n the lexicon of c o l l o q u i a l Zhong-shan words that follow the syllabary, a c i r c l e i s used to represent any characterless word, including onomatopoeic s y l l a b l e s and loan- words . In both the syllabary and the lexicon, medial /w/ i s treated as part of the i n i t i a l i n order to reduce the combinatory p o s s i b i l i t i e s of i n i t i a l s and f i n a l s . Thus, besides /k/ and /k'/ there are also the clusters /kw/ and /k'w/. Medial / i / continues to be treated as part of the f i n a l . Besides the words e l i c i t e d by the writer, the syllabary also includes some words recorded by Chao. Co l l o q u i a l terms entered into the lexicon from Chao's a r t i c l e are also given e x p l i - c i t acknowledgement. In the d i a l e c t survey l i s t , also, the source of data i s likewise s p e c i f i e d i f no confirmation i s made by the writer that a given word i s also used by her informants or i s pro- nounced i n the same way by them. For convenience of comparison, Chao's data are treated as - 108 - i f they were transcribed using the present phonemic system. Hence, the f i n a l [uk], for example, i s transcribed /uk/ for both the present data and Chao's. In Chao's o r i g i n a l phonemic system, [uk] would have been transcribed as /ok/. (Note that i n the present system /ok/, i n turn, i s phonetically [axk].) Besides the ease of comparison, using a single phonemic system has another advantage: i t f a c i l i t a t e s collapsing of two sets of data with no r e a l loss i n information. Where the two sets of data d i f f e r , as i n the case of the s y l l a b l e [jsin] versus [jcexrj]., the present form of [joxg], phonemically /jong/, i s used i n the syllabary. S i m i l a r l y , since the present data do not show the i n i t i a l c l u s t e r /ngw/, no space i s a l l o t t e d for i t i n the syllabary. As regards the lexicon, i t s primary purpose i s to assign meaning to the many c i r c l e s and X's scattered throughout the syllabary. Other items included merely serve as a sampling of Zhong-shan c o l l o q u i a l usage. No attempt i s made here to pro- duce an extensive lexicon. I t should also be made cl e a r that a number of the expressions, p a r t i c u l a r l y the more vulgar or abusive terms, are not actually used by the writer's informants, but are expressions that can be found i n the d i a l e c t and are used by the less educated. In terms of format, the lexicon i s arranged according to the Zhong7Shan f i n a l s presented i n Table 2 (a). The lexicon begins with the set of f i n a l s containing the high, front nuclear vowel / i / ( i . e . , / i / , /iw/, /im/, et c . ) , followed by the set of f i n a l s containing /y/, and so forth. The following symbols and abbreviations appear i n the lexicon: - 109 - r>^ = r e p e t i t i o n of the s y l l a b l e that occurs at the left-most column S = sentence Q = question CL = c l a s s i f i e r PRT = p a r t i c l e s i . = slang l i t . = l i t e r a l l y esp. = especially s.o. = somone s.t. = something Moreover, i f no characters are assigned to a p o l y s y l l a b i c word, .'<• i t w i l l be assumed that the word contains characterless words, with or without the exception of the repeated s y l l a b l e from the left-most column. Thus, there w i l l be no case of "0~", " ^ O O " , etc., actually recorded i n the lexicon; they w i l l simply be implied. - 1 1 0 - 2 . 1 . S y l l a b a r y A r r a n g e d A c c o r d i n g t o M o d e r n Z h o n g - s h a n F i n a l s i y u 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p ' m f ** *& | if f # $ o # •* £ t t ' n 1 o x </o ft it hi f if «t A IS fti & t s t s ' s *« 0 tin- fa 0L & i*t» 'X jjb $ fc & A & A $ fg tit §& k k " n g h 0 A* * % -£6 o A 4 o * 4 - # i & * f /il Jk* a ^ 4" li» » ti t k w k ' w w j - I l l - m i 0 o a 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p ' m f ft 1k & i& ik h % t t ' n 1 0 o *$ I i rt #f *f 4 ft tt. o a o ft O o O O O O t s t s ' s o X ft it i l •fc 41 £L i x -ft IL & ^ » + % o k k ' n g h 0 o o % 0 V\ o o % 0 ML WK * 0 % *fH f f -f k w k ' w w ' $ & it * i& ** 1 * j 1 i |j f̂" -e> /IL - 1 1 2 - i a u j 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p * m f X o o O O O o * ft i i & t t ' n 1 % o o *fe ii i t i l Rt i l o # O t s t s ' s i t < f i t l f k k ' n g h 0 o o hi o •1 1 g$ 0^ *t o % 1— / k w k ' w w j - 113 - Vs. a j a;,j 55 51 13 22 55 51 13 22 55 51 13 22 p p.' m f X ft & & i & o i i * . i i i i f o < # Ik * *£ t i ft t t' n 1 ft & £> ft t i t » * i f t « f tt 4rt 0 & 4# 4# 48 o o II ts t s ' s t f 4 o A A & fa 1 i t o $ f £ t m f t 4 i o H BI k k' ng h 0 W a # i *̂  It) t & & % o K'h A o | © S to M ̂ ft ;S o £ ° ft &i 4& kw k'w w i £ * ^ IS) £ ft j - 114 - i w o w a w 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p ' m f if fja. ik % 4 Is o 4t o fa #8> o ±> o ti f o o t t ' n 1 flfc \'L % IP o % M, o ft 3 ft AS i i f 1 o f % t s t s ' s i '£) B i l M 4ft * i is * 1 i t * '£ A f ? j i ft A i f *\ £ & # & {i £ k k ' n g h 0 & # IL it *. f t 4 ^ ^ t ft 0 ft 0 ^ Sit -6 ;& k w k ' w w j 4k # 1 ^ - 1 1 5 - a; w i a ; w ira 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 P P ' m f & fa to 46J -Jtfe t t ' n 1 I f O V % k xk \k ft % / f k & t s t s ' s •*» ^ | $ o 41 % & » -ft £ ik f l k k ' ng h 0 X o & (IL tk rt i o | o o o % %r *f o ii 4 t k w k ' w w j - 1 1 6 - am am a :m 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p ' m f X t t ' n 1 O X O O 0 O O O O o fa ;f &t iii $ * t il i i i l o If :^ t s t s ' s ^ -3T >M1.- f ft # t #1 i f *) k k ' ng h 0 A % # O vfy O 0 O it o ° & k w k ' w w j 1 4 & & o - 117 - i a ; j n i n y n 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p ' m f i t 4 $ 01 ° f t f n 1 II * * O 4f te. ft. III tfty t s t s ' s i i . $ ft % ® A J & 3 t & k k ' n g h 0 f t a* « & i t t ft m 1 £ ik ri k w k ' w w j - 1 1 8 - u n jzSn o n 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p " m f * * t % £ # in n yk n t t' n 1 / k W 0 ts ts' s i £ >| o k fa h o ill 4 «ft /'II k k ' n g h 0 f f % 1 "ft 1 If It 1 * % fh k w k ' w w j - 1 1 9 - a n a : n i n g 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 5 5 5 1 1 3 2 2 p p ' m f 1 1 * * * * M M f 4 w tit ° afc »H if) ^ t JCJ ^ I O t t ' n 1 % o £ ^ £" o o •hi? %V Ar/~ 5? ^ jfc « ft *a ® « « T JS i l ft a 4 t s t s ' s -I ft i m ?t / f - % 4 / i 'At J « % tk ik i_ si n k k' n g h 0 ^ f it ft & *t 4 t *)t « ft w & | l C /fi W T Ft kw k'w w ^ £® $ * 9 *£. & ftS 'it f & & ^ . . . j & A. & if - 120 - u n g o n g a n g <>N? v\ 55 51 13 2 2 55 51 13 2 2 55 51 13 2 2 p p' O % * i t if) /m -J- X M Ai m t \\ 0 k f 0 t t ' t i t >t * * n O i 1 I o ti o 1 i t ft t s * f If t s " s *} fa 0 1 0 i HP 8 o k 'A i - f i r i t k' $ 0 o n g h & i f an I o 0 1 ik H x2 0 0 B o kw 0 k'w w * * j ft * 1 9 4 - 121 - a ; n g l o n g i a : n g 55 51 13 22 55 51 13 22 55 51 13 22 p p' m f o \ h%. o -f X I % t f n 1 o O L & $fu o i i t s t s ' k s i » if i o M 4 /» *t ^* *l 4 « $ ft *8 l i i £ k k' n g h 0 JL o o I ^ 1 51 31, 1 t- ft I $ 1 41 0 kw k'w w O X 0 Bi o y[c j 4 11 - 1 2 2 - i p o p a p a : p i a : p 5 2 5 2 5 2 5 2 5 2 P P ' m f t %& 0 o O t ' ft 0 n U o 1 t s & # 1 f i t t s ' . 'S fab i f . + to k X k ' t J - l o # 0 h A-a 4*. 0 t 0 k w k ' w w j o - 1 2 3 - i t y t u t o t a t 5 2 5 2 5 2 5 2 5 2 5 2 p p' m f O t t' n 1 o IK o H o If X g?u a t 0 o o 0 t s t s * s ~ KK ° q A o & \& 1 il 1 k k ' n g h 0 o H o % & ¥*• S i A X 44 ;4 X k w k ' w w h ii j - B - 1 2 4 - a;t l a : t ik uk ok ak 5 2 5 2 5 2 5 2 5 2 5 2 p p' m f /V 0 X X h « o « 0 ^ o If i l t t' n 1 & X ft & it as 0 0 * « $ o o o ^ ts t s ' s it X i t f W ;f 31') I* k k" ng h 0 O X * 1 $ ft 4- o f Jl. i/E. kw k'w w ft o o j - 1 2 5 - a ; k i o k i a : k ( S y l l . ) 5 2 5 2 5 2 5 . 5 5 1 1 3 .. 2 2 p P' m f O f o t X t ' 0 n • 1 o o o i t t s f f t s ' & o 8 4 * k fa k ' ng h ft 0 X k w k ' w o 0 w j *1J - 126 - 2.2. Le x i c o n of C o l l o q u i a l Terms Arranged A c c o r d i n g to Modern Zhong- shan . F i n a l s . . 55 P i o i 55 55 / p i p i / = a w h i s t l e . , . 55 p ' l o = worn. . 55 mi IL = t o s l e e p . . 51 mi \L = / f a : t 2 m i 5 1 mung 2 2/ ^ ) = to have a dream. ..55 t i 55 = some (Chao). (More f r e q u e n t l y : /naj /.) ,..51 t i 0 / t i 5 1 t i 5 1 t a r n 5 1 t a m 5 1 / = s a i d o f s.o. who i s qu i c k - f o o t e d o r nimble. . 55 n i = Q-PRT. . 55 l i o / l i 5 5 t s a j 1 3 / (~ Ah ) = a s m a l l basket; a bamboo- type c o l l a n d e r or s i e v e . o / l a : w 5 1 l i 5 5 / ($Jjp ~ ) = a wired s t r a i n e r w i t h . long handle. / k i 5 5 l i 5 5 k w a r j 5 5 l a : j 5 5 / ("^ ~ * ) = d e p i c t s sound of speech t h a t i s u n r e l a t e d and unconnected (Chao). l i 2 2 m = tongue. t s i 5 1 0 = to poke s.o. wit h f i n g e r (from \% / t s i 1 3 / ? ) . o / s a p 5 t s i 5 1 t s i 5 1 / {}$. ) = very moist. o / t s i 5 1 t s i 5 1 t s a 5 1 t s a 5 1 / = t a l k a t i v e , g o s s i p y . o / t s i 5 1 t s i 5 1 t s a m 5 1 t s a m 5 1 / = m u f f l e d sound of v o i c e s . , • 55 t s ' 1 $ = t o s t i c k ; / t s ' i 5 5 saw 1 3 t s ' i 5 5 k i o k 2 / ( ~ % ^ jjtfV) = very s t i c k y ( l i t . , s t i c k s to hands, and f e e t . 4. i-51 t s 1 1 = s m a l l pond.,., (Note: / t ' o n g 5 1 / ^ = l a r g e r pond; / t s ' i 5 5 t'ong / (;&*,# ) = pond ( n e u t r a l ) . 9$ / s i 5 1 man 5 1/ (<~$ ) = t a l k ; / s i 5 1 man 5 1 t o 5 5 k o 2 2 f a i n / ( ~ j yjh ) = m o r e t a l k t h a n r i c e " very t a l k a t i v e . #t / t ' i n 5 5 s i 5 1 / ( ^ ) = the weather. - 127 - s i " A / s i 1 3 f a t 5 / C~ % ) = rectum. i 1 3 JL / s i 1 3 s i 2 2 £ / s i 2 2 kwa:n 5 5/ (/~ #j ) = because. k i 5 5 A ' u 1 3 k i 5 5 / (.jjt ^ ) = hungry. 51 51 51 51 51 k i O / k i l i ka:w la:w / = noisy t a l k i n g — a s of a group. k i 2 2 - f £ j / k i 2 2 sam 5 5/ ( ~ / C / ) = memory (Also: / k i 2 2 s i n g 2 2 / = to turn around (e.g., head) (Chao). .55 _ ,, 55 .55 .13, \ •, j. , • ngx O /tsa:ng ngi nga:j / (f ^ O ) = almost ( i . e , lack just a l i t t l e b i t ) . h i 1 3 = to l i f t up. i 5 1 j?p / i 5 1 k a 5 5 / ( ~/>° ) = es p e c i a l l y , even more. iw . 13 miw *ky /miw 1 3 t s t f j 1 3 miw 1 3 s i t 2 / (~ "% ~ k ) = to sneer. 4 . . • 51 t' I W hk = CL for long, slender objects (e.g., necklace, thread, street, e t c . ) ; also used for trees. . 55 niw fa* 55 55 55 = slender; /niw pa:ng pa:ng / = very slender. o 5 R 55 1 3 5 5 5 /niw„ nana 7 = precariously; / k ' i tak niw nang / ( <AE \% ^ O ) = to be standing precarious- ly (e.g., on a high ledge). , . 55 liw o = to poke finger i n corners, small holes, etc. . . 22 tsiw = to chew. . , . 55 ts ' I W o = to pry open; to dig (e.g., a d i t c h ) . . , . 55 k' iw o = to cross the legs; to scoop out (e.g. with spoon). . , . 13 k' iw o = to cross the legs. o = coincidental, extraordinary, unexpected. /k'iw 1 3 t'aw 5 1/ ( <~ Ik ) = of the g a r l i c family, used for p i c k l i n g . 13 hiw o = to understand. , . 22 hiw o = to st i c k out (e.g., l i d of an opened t i n can). (Also: /hia:w 2 2/). - 128 - im %fk = straight; / t i m 2 2 pa£ 5 l a | 5 / ° 22 = v e § y straight (Also: /tim pat pat /, /tim p l a t / ) . -JfÛ  = to mind, to be fussy, to d i s l i k e . 2 22 /pa:k im / = peevish (referring to children only) i n O = diaoer. (Also: /niw 2 2 p ' i n ] 3 / (>fic ~ ) , / s i 1 3 p ' i n ± 3 / ). 55 55 55 55 = where. (Also: /min naj /,,/naj /, /min /ts'y / ^ ).. (Phonetic variant: /pin / ) • ri 55 13 fljj /min tsow / ( ̂  ) = tomorrow. /how 1 3 t ' i n 5 5 / ( <k-y~ ) = clear day. O /ngaw 5 1 n i n 5 5 / ($ "*".)-•= cow's milk (Chao). f | / t s i n 5 5 h o n 5 1 t s i n 5 5 l a : n g 1 3 / («~ £ ~ ' t" ) = shivering cold, to have c h i l l s . *t /t'aw 5 1 s i n 5 5 / (Sit ^ ) or / t s i a : n g 2 2 s i n 5 5 / (.IE. r~- ) = just a while ago, just a moment ago. lh\ / s i n 2 2 k a j 5 5 / ( ) = castrated chicken. ^ (Buti one,who castrates chicken i s c a l l e d : /im k a j 5 5 low 1 3/ JfctttiL)- 0 = to l i f t up (e.g., mat), to peel o f f . /min 5 1 t s ' o n g 5 1 / ( ~ # w ) = bed. ing O = to toss c a r e l e s s l y . O / t i w 2 2 t i w 2 2 f i n g 2 2 / ( i f $ ^ ) = to be hanging loosely or precariously. 0 / t i n g 1 3 t a j 2 2 / = fussy (e.g., to describe the way someone eats). = to hold; to bring. - 129 - k i n g 5 1 O / k i n g 5 1 l i n g 5 1 kam 5 1 lam 5 1/ or / k i n g 5 1 l i n g 5 1 kung . lung / = of noise from stamping feet, mov- ing^ furniture, ̂ ^ t c . (7\lso: / k l i n g klam / or / k l i n g klung /) . iP s i p 2 ^ / t ' i n 5 5 s i p 2 / ).= l i g h t n i n g . ^ = to wedge, to s l i d e under a narrow opening. k i p 5 jj£ / p ' i 5 1 k i p 5 / (JL^ ) = suitcase. (Hashimoto £p. 349) acknowledges McCoy's analysis that [ki:p ] for 'bag, suitcase 1 i s a loan word from English 'grip'.) 5 5 13 ngip O = to blink; /ngip nga:n / (~ eK__) = to blink; i n a wink of the eye ( i . e . , very q u i c k l y ) . 5 13 0 /ngip s i / = stingy. O = to bite.Ce.g., of small i n s e c t s ) . i p 2 = to preserve i n s a l t or vinegar; by extension, means: to treat c r u e l l y a daughter-in-law or children of the husband's e a r l i e r marriage. i t 5 fy^ = to pinch; to tear up; to peel. mit ^•,5 /,. .5. ...5 .13 . 22 .55 . .13, . ̂  r> t i t O / ( 3 a t ) t : L t S 1 kam na] tsaj / ((.-) ^ ^ v$ 0 ) = a very tiny l i t t l e b i t , a very small amount. O / t i t 2 t i t 5 tow 1 3/ (j0L~/& ,.|) = to f a l l down. 5 5 5 l i t 0 = to be i n a knot (as i n wrestling); / k ' i t l i t / = a knot (Also: / k ' l i t / ) . t s i t O = to t i c k l e ; to squeeze out. 2 22 t s ' i t O = tp^dote on, to l i k e (e.g., a c h i l d ) , /fa:n jan t s ' i t / K ^ ) = lovable (of a c h i l d ^ (But., older-generation use the phrase: /fa:n jan tsang / ( VL> >^ % ), l i t . , ' d e t e s t a b l e ' ) . k i t 0 = to jab, to pierce. k ' i t 5 O (see: / l i t 5 / ) . (ki) . - 130 - |& / n g i t 2 t'aw 5 1/ C~*ft) = sunshine. ik 5 5 O / t i k sik / = small, cute and elegant. O = to make a check mark ' </ ' . kjj* = the 'hand' r a d i c a l . 5 5 5 O /t'iw t ' i k / = fussy, c r i t i c a l . = to take s.t. along. 5 5 5 5 5 5 O /pik l i k pa:k la:k / or / p l i k pla:k / = crackling sound—of f i r e , or s t r i n g of fir e c r a c k e r s . O = to yank; to straighten up (e.g., clothes). /m51 s i k 5 s i n g 2 2 / (»&'•*'fl.) = naive; to be misbe- having. O = to clog, to jam. O / k ' i k 5 l i k 5 k'ap 5 l a p 5 / or / k ' l i k 5 k?lap 5/ = odds and ends; c l a t t e r i n g sound—as of dishes. y ;jt = he, she, i t ; / k ' y 5 1 t i 2 2 / (~ •*€*) = they. O = to keep warm by putting over low heat. O / h y 5 1 tow 1 3/ M£'J ) = to worry (Chao) . / t o 5 5 y 1 3 sgij13/ = a l o t of r a i n . yn = crooked. = to r o l l i n s.t. (e.g. f l o u r , d i r t ) ; to coat. = to handle (esp. a l o t or too much). = to crawl through. 1 3 = c l a s s i f i e r for^kooks (also pronounced /kun / ) . (Also use: /pun / — a more-modern term). - 131 - yt t y t 5 o = to protrude; to pout. t ' y t 5 o /wa:t 2 t ' y t 5 t ' y t 5 / (>f ~ ~ ) = very slippery. i y t 2 o / l y t 2 h i 1 3 j a t 5 kaw 5 1 l a w 5 1 / form a clump. (~&- *% O ) = to t s y t 5 O = to give a k i s s . t s y t o '= a d u l l , gnawing pain. k y t 2 = a^piece that i s one-half or less; l i m .22 22 /ta: j pun 3 0 = thick (of l i q u i d ) . (Cantonese: [ k i : t ].) 2 2 2 55 K5*) ngyt 0 /ngyt / or /ngyt kung / (̂  J ) = the moon. u p u 1 3 £ / p u 1 3 / . ^ / p u 1 3 t s ' 0 j 5 1 / (~&iL) = an axe (But: 1 /ts'j2$j 3 J-/ ) = a hammer). P ' u 5 5 /P 'u 5 5 k o j 2 2 / (~S.) = luggage. p ' u 5 1 = to f l o a t ; /la:w 5 1 p ' u 5 1 / (#$~) = to f i s h or scoop out; to l i f t up (e.g., table) by two or more people. 55 22 55 -c mu 0 /a mu / (It ~ ) = woman servant (Chao) . mu 1 3 kfy /a22 mu 1 3/ (ft ~ ) = mother (Also: ^ a 2 2 ma13/.) (Chen's mother c a l l s her mother /a t s i a /ft . - W i ) . / i o 1 mu 1 3/ (̂ &-~) = mother's ojLder s i s t e r Colder term; a more modern one i s : / i ma /Jtfc. k$g ) - mu 2 2 ^ = fog; /mu22 s 0 j 1 3 / (~*)<-) = dew. t u 5 5 %f = s t i l l , yet; also. 55 55 55 nu O = that; /nu naj / = those; there. l u 5 5 " f * A u 5 5 k u 5 5 l u 5 5 l u 5 5 / = to talk i n an i n d i s t i n c t manner (Chao). 2 2 0 = inchoative, S - f i n a l marker. (Also: / l u /.) t s u 2 2 O /Verb + t s u 2 2 s i n 5 5 / (Verb + ~/b») = V - f i r s t . = to save, to c o l l e c t . O = t h i s ; / k u 5 5 t s a n 2 2 / O f£ ) = and then. #D / t s i 5 5 ma 5 1 k ' u 5 1 / (/£/& — ) = sesame paste. O /ts'jzfok'u 5 1 §gaj 1 3 / 5 ^ ^ r ) = young c h i l d ' s penis (Also: /tsaj tsaj /.) woman; )L / h u 5 5 n i o n g 5 1 / (~i£j = a married / h u 5 5 n i o n g 5 1 t s a j 1 3 / (~4M3-) = a g i r l . 0 = to throw o f f (.e.g., blanket), ftf = t o Paste; / t s i o n g 5 5 u 5 1 / (.{JJL—) = paste, = d i r t y ; /uDO tsow 3 D l a : t z t ' a : t z / | f 0 O ) = d i r t y . A 55 51 51 vS? /u ham ham / = very dark and murky; / u 5 5 t ' i n 5 5 ha:k 5 t i 2 2 / {~jL&i&) = very dark- as of pending storm. ^ /haw 2 2 p u j 2 2 t a j 1 3 / (^ jL^/fo ) = at the back, behind ^ /muj 5 5 t s a j 1 3 / = a purchased maid. O = there, at that place. 0 = a kind of clam (Chao). >f / l u j 5 1 kung 5 5/ ( r 3 ' A ) |,thunder; /ha:ng 5 1 l u j 5 1 / iff <~ ) or /hiong l u j / («fo ~ ) = to thunder. O = .here, at t h i s place. O = H e y l ung O = sound of h i t t i n g or hammering; sound of f a l l i n g object h i t t i n g a surface. O = a surge of fragrance or odour. O /p'ung 5 5 p'aw 2 2/ = f l u f f y . - 133 - O = t o s t i c k s . t . f i r m l y onto t h e ground. 51 51 "jL O / t ' u n g p'o / "4r ) = shaman o r medium ( f e m a l e ) ; / t ' u n g 5 1 k u n g 5 5 t s a j 1 3 / 43-) = shaman o r medium (male). (Most shamans a r e women. They a r e n o t found i n e v e r y v i l l a g e , a l t h o u g h b o t h Ku-chong and Yuan- f e n g had one.) O = b u r n t . = s t i l l , y e t . (More f r e q u e n t l y use: /wa:n /J^.) • 22 22 O / t s ' u n g ha / (<-̂  "F ) = j u s t below, t h e n e x t one down. $fc^ = f o o d t o be e a t e n w i t h r i c e . 22 55 -n; V* /a kung / (sk ~ ) = p a t e r n a l g r a n d f a t h e r ; 55 /Surname + kung / = m a t e r n a l g r a n d f a t h e r . = t o push. O = t o f e r t i l i z e ( C h a o — h e a l s g , . t h i n k s t h a t i t may be a sound change from ,/ung / yjfc. ) . u t 5 5 55 *b O = sound o f a h o r n ; / p u t p u t t s 1 i a / ('v"v'Jf ) = c a r ( o l d term used by v i l l a g e c h i l d r e n ) . <c . , . 2 ^ .22 , , .51 . 22. ^ j - j l / ? * x ^ , / p ' u t t a : ] k'waj s m / {^K'K^ ) = t o be a matchmaker ( l i t . , t o f a n w i t h a l a r g e p a l m - l e a f fan) . 5 5 55 '-0/ / k u t k u t s i a : n g / rfr ) = sound o f s w a l l o w i n g . uk O = d u s t y ; /puk t s ' a n / ('-'J^ ) = d u s t . Hit = t o p u t one' s head down on one ' s arms. "̂K = t o t o p p l e ; t o be t u r n e d u p s i d e down. 5 22 Jkt 0 = t o p r e d i c t , t o guess; /muk t s u n g / (.<*- j ) t o guess c o r r e c t l y . 0 = t o p r i c k . - 134 - c 55 5 nuk O = to move, unsteady; / n i nuk / - unsteady, wobbly (e.g., of legs of a table). 5 suk l u k 5 j$!o = to r o l l ; a r o l l , a loa f . 0 / f a 5 5 l i 5 5 l u k 5 l u k 5 / (.#> 0 ~ ~ ) = very bright and c o l o u r f u l . l u k 2 yjjfa = to scald. ts'uk 5 O /ts'uk 5 h i 1 3 t s ' uk 5 h i 1 3 / (~,££~î  ) = to be bunched up. O = rancid (e.g., of s p o i l t meat, r i c e , e t c . ) . 22 5 /a suk / (i. ) = father's younger brother; a term of address for s.o. younger than one's father; father (used by some, e.g., Chen), when a male i s "adopted" by his father's deceased older brother who i s c h i l d l e s s — i n order that the l a t t e r may have p o s t e r i t y ) . j u k 5 = to move; /juk 5 j u k 5 kung 2 2/ = to wriggle about • ( i . e . , can't s i t s t i l l ) . tjzS55 O = to scold. l e i 5 5 = to remove food from inside mouth with tongue. l e i 5 1 / t ' i n 5 1 1 0 5 1 / (fJ7~) = fresh-water s n a i l s . 13 tsjz* O = to eat up. s 0 5 1 O / s 0 5 1 s 0 5 1 s i a : n g 5 5 / ) = sound of running water. ' 51 51 51 51 O / s i s i sjzJ sfi / = depicts (rude) manner of guzzling down food. k 0 5 5 O = to r o l l up. 13 k0 O = to r o l l up (e.g., cigarette, s c r o l l , mat, etc.) k 0 2 2 JfrfL = to saw (s.t. small). k'izT 1 fn = eggplant; /fa:n 3 k'szTV (& ~ ) = 0 5 1 0 = S - f i n a l p a r t i c l e of assertion. tomato. - 135 - O /tam 2 2 t 0 j 5 5 / ( J l ~ ) = shabby. ")<)L. = to steam-cook; of smoke going up the chimney. P&_ / t s ' 0 j 5 5 haw 1 3 sa:w 2 2/ V ) = to sing (But,in modern usage, i t mean to whistle, esp. at g i r l s ) . 22 22 O /10n ts0n / = clumsy, troublesome, nuisance. ^ /tsjrfn 5 5 s i 5 5 / ( 'v if ' i , ) = v e r m i c e l l i . O (See: /10n 2 2/) O / t i n 5 5 t i n 5 5 ts'jrfn 2 2 ts'tfn22/ ) = insane, crazy. = to s l i d e down. » sound of snoring. o 3P5 = C L f ° r trees. )JL /P'o 5 5 t s ' o j 2 2 / (-I-) = spinach. „ A 22 51 « „, m /a p'o / ) = paternal grandmother; 51 /Surname + p'o / = maternal grandmother; / k a 5 5 p ' o 5 1 / C & . ~ ) = mother-in-law. _ 55 55 51 O /mo s i t'o / = slow, to be a slow-poke. 1* ^ ^ . , 55 u 51 51 13 .13 ., 2, ^ = to (touch; /mo ha:m ngy su ma] t'a:p / (^fyfij* ,$kJLO) = touch salted f i s h , count r i c e bin--constantly counting what one has, how much money one i s making, etc. Jji / t o 5 5 f a : n 5 1 / ) or / t o 5 5 f a : n 5 1 s a : j 2 2 / ("'W'&J = thank you (normally said by rec i p i e n t of a favour). i & J / t ' o 5 5 h a : j 5 1 / = sl i p p e r s . 51 22 0 /t'o l a j / = to involve, to implicate. $ = to rub with the fingers. •jS / l o 5 5 s o 5 5 / (~*jL) = wordy. 55 51 0 / l o jaw / = derriere. O / t a 1 3 ha:m 2 2 l o 2 2 / tfl O ^ ) = to yawn. O / l o haw / (~ = denotes f e e l i n g i n throat afte r eating o i l y , f r i e d food. to axe. = an axe. / 22 "{© = t h i s (Chao, variant form: /ko /) . •jlS = possessive s u f f i x ; CL (general) ; t h i s (Chao and B a l l ) . 0 = to rinse. O = to eliminate (vulgar term). = Q-PRT (phonetic variant: / a 5 5 / 0 ^ ). °$ = Oh! Ah! = nodding p a r t i c l e . oj O = an exclamation of disgust or exasperation. j ) £ / f a w 5 1 f o j 5 5 / ( # L ~ ) or /t'aw 5 1 t s y 5 5 t ' o j 5 5 / ) = Drop dead! rfri = a long time. = to come. (Cantonese: [ I B J 2 1 ] ) / t s ' o j 5 5 k o 2 2 n i 1 3 / (~ O'ftf. ) = Fie on you! 0 = to leave i t up to s.o. 22 22 O /ngoj soj / = i r r i t a t i n g (e.g., as of a canker sore). 53 O = to carry an infant on the back. (Cant.: [me: , ] ) . - 137 - 22 51 Je = to love; t o want; to need; /oj s i / C~ 0f ) = sometimes. ow = to cook (in water); a cooking pot. 55 55 O = perfective s u f f i x (Variant forms: /how /, /ow /) = suds. O / f a : t mow / (^~) (Chen, suggests i t came from English 'mold') !to become moldy 1. ^ = not have. = a case, envelope; a s u i t (of clothing). = to mix. tyjp /low 5 1 l o w 5 1 l u k 5 l u k 5 / = hard-working (of one who r t o i l s ) . O = to pour i n and out; to pour through a funnel; a funnel. 22 5 • ^ i /tsow f a t / = a cooking stove—has an opening i n the front for putting i n f u e l , and openings on top of the stove for a wok, k e t t l e , etc. The ones used i n Ku-chong and some houses i n Shi-qi had a large hole towards the front of the stove f o r the wok, and two smaller holes i n the back for k e t t l e s , pans, etc. /fat / i s Jjf ?) . = to shake; /ngow 5 1 ngow 5 1 nap 2/ = wobbly, shaky. )fr /m51 how 1 3/ (A ~ ) or /mow13/ (fr o m : 4 ^ ) = don't 2 (negative imperative). (Zhong-shan also use /mok / 0 (See /p'ow55/) om O = CL for leafy vegetables (e.g., spinach), clumps of grass, bushes, etc. - 138 - on ^ = straw (Also used for f u e l i n the cooking stove.) 22 , 13 /sow kon / (^ ̂ ) = (straw) broom. , 55 51 = to put, to place; /on jan / = (^ A-) = mother- in law. ong 51 ^, 51 a. / m a J X f o n g ' V t # K ~ ) = c a n d y , = c r a z y . O = s t a l e , •̂ t = t o s h a k e o u t . o p ^ = c o m p a t i b l e . o t O = t o s c o l d , t o . r e p r i m a n d . ok j* = to h i t hard on the head. O = to chop (e.g., wood). A i 5 55 55 jwl = to peel (e.g., s h e l l ) ; /mok fa sang / (~ MJ £.) = to s h e l l peanuts; to be the t h i r d party (si.) = don't (negative imperative); /mok2 f i 5 5 / (̂  ̂  ) = could i t be possible that . . . (Apgroxima£c|ly simi l a r to the Cantonese use of [m f u g ]»%3aL) • / j o n g 2 2 mok2/ ^) or / s i o n g 2 2 mok2/ (J£, ^ ) = approximately. = to carry with head or shoulder. - 139 - t'ok^ O = to explore 5 Iok O = to extract teeth ... }%r /kok 2 Iok 5 t'aw 5 1/ or /klok 5 t'aw 5 1/ ( $ ̂  ) = corner. 2 = inchoative S - f i n a l marker (variant: /ok /). . O = to support s.o. ph y s i c a l l y . O = to quiz o r a l l y . O = to knock on the head. 5| = to raise the head, or t i l t i t up. a O /ham 5 1 p a 5 1 l a : n g 5 1 / or /ham 5 1 p l a : n g 5 1 / = a l l (.&).. p ' a 3 i O = to paddle (e.g., canoe). 55 . . ma ty- = twin. O = /ma 5 5 k u 5 5 i n 5 5 / (/-O^!) = cigar; /ma 5 5 k u 5 5 i n 5 5 t a j 2 2 / (~0 ' A £ ^ ) = cigarette butt; 55 55 (/ma ku / i s borrowed from the l a s t two s y l l a - bles of Portuguese 'tobaco'—term used i n Shi-qi.. Observe that the Portuguese word was recorded i n the Sino-Poirtuguese glossary using the Chinese w o r d s ^ " 5 5 / Pronounced i n Zhong-shan today as /t a : j ma ku / ) . 2 2 ma 0 = a pause-PRT at the end of a clause, ta jft = dozen (from English 'dozen'). t a 5 1 O = s t i l l , yet CAlso: /wa:n51/)J^,) . t a 1 3 jft / t a 1 3 l i 1 3 / = to take care of a place .(e.g. , a room), to keep a place clean. n a 5 5 = a scar. O = to s t i c k ; / t s ' i 5 5 n a 5 5 / (J>$o ~ ) = sticky; /m51 n a 5 5 ka:ng 5 5/ ("%^0) = don't connect (as of unrelated t o p i c s ) . Iok 2 ts'ok 2 k'ok 5 ngok 2 - 140 - 22 na = an i n t e r j e c t i o n — H e r e ! (Chao).. l a 5 5 O / k a 5 5 l a 5 5 h a 2 2 / (~ O T ) or / k l a 5 5 h a 2 2 / = armpit ( f l£-) . = S-PRT—exclamatory p a r t i c l e of affirmation (from / l o k 2 a 5 5 / *>£-*f=i ) . RI Ri RR *• l a 3 X O / l a a:j / (~ = d i r t y . vjfc. = S - i n i t i a l p a r t i c l e — T h e r e ! 13 l a O = pungent, alkaline taste. l a 2 2 = a crack (e.g., i n the wall) . t s a 5 1 O = / t s i 5 1 t s i 5 1 t s a 5 1 t s a 5 1 / = yakkety yak, talka- t i v e , gossipy. 22 22 22 tsa O = yet (Informants claim i t i s from /mi a / 0 > . 55 ts'a O = bad. 22 22 £l< sa O = to loosen (from /san / ?) . k a 5 5 O / k a 5 5 ma 1 3 s i 2 2 / = an "X". / k a 5 5 / or / k a 5 5 h a 1 3 / (~ "F ) (Chao has / k a 5 5 ha 2 2/) = now. (There i s a saying that goes:4* *^ »w-*1)L"lr H| "tjj. < fiP : Fo^^han uses , Canton uses ^ ( i . e . , [ j i x k A : ] ) . 22 55 > ^ = elder s i s t e r ; e^g., / t a : j ka / ) = eldest s i s t e r ; /ngi ka / ~ ). = second eldest s i s t e r , etc. (Also used: / t s i / k% i n thc^same-^nviron- ment for 'eldest s i s t e r ' , but /ngi t s i / ) i s only used i n salutations i n l e t t e r s ) . 51 ka O = standing position with-legs .astride. k ' a 5 5 O /k'a:w55 k ' a 5 5 / = to make an "X". O / k ' a 5 5 k ' a 5 5 s i w 2 2 / ( IL) = to laugh. k ' a 5 1 £-(; / k ' a 5 1 t s a : t 2 / H f ) = cockroach. n g a 5 5 ^ /nga 5 5 t s a j 1 3 / if & ) = a baby; 22 55 J . / t a : j nga / (?̂ -'~' ) = big baby—but used for the eldest of one's younger brothers; / • 2 2 55/ /ngi nga / brothers, etc. n g i 2 2 5 5 (^ ~ ) = next eldest of one's younger - 141 - 2 2 2 2 O /nga tsa / = awkward,, bulky. O /nga 2 2 t s a 2 2 / (~ fllf) = said of s.o. who hogs space food, the road, etc. 55 A' *f*J = Q-PRT (phonetic variant: /o /»4\ ) (Chao) . w[3 = Q-PRT to e l i c i t r e p e t i t i o n of information (.Chao) » ^ = S - f i n a l PRT—affirmative exclamation. (Chao) . aj *JL /maj 5 5 m a j 5 5 s i w 2 2 / 0 ~ ~ ? L ) = to gri n . = to cram; /maj 5 5 s y 5 5 / (~ ) = bookish; to cram. O = then (e.g., i n conditional.Jsentences) . O = precisely, exactly. O = rotted Xe.g., wood); useless (e.g., p e o p l e ) — from / f a j / jfc ?) (Chao). O = to hand s.t. up to s.o. Jj? /a t'aj / (^L^) = term of address used by the older generation to a boy (more formal). 5 3 O = some (Cantonese: [ t i : ] ) • 55 13 O /naj tsaj / (~ /(V) = small amount; 5 55 / j a t naj / (— ~ ) = a small amount (post-verbal) 5 55* / j a t naj / (w~<*) = a l l (pre-verbal) . O = where O / n a j 2 2 ka:k 2/ {~j[%r) = adhering s t r i c t l y to rules and regulations. |\j = fat t y or greasy (of food) . = to place (Zhong-shan usually use: /on / ty , n / f o n g 2 2 / ^ , or / t a t 5 / O ) . = son, c h i l d ; diminuative s u f f i x . " = said of food that i s hard on the digestion. 1 55 5 O /saj t s i k / = a side glance. - 142 - it- / m 5 1 s a J 1 3 / (*«'~) = don't need to, don't have to. O / t o 5 5 k a j 1 3 / C$t~ ) = cunning (from / t o 5 5 k a j 2 2 / 'many schemes'?).. (Chao). \% / k ' i n g 5 5 k a j 1 3 / (>fl~) = to chat. O / l o n g 5 1 k ' a j 5 5 / (4jL~ ) = daring. O = to take a very small b i t e . O = to b i t e using the top and bottom front teeth (e.g., to eat melon seeds). J$z = vagina (Chao) . O /nga:n 1 3 h a j 1 3 h a j 1 3 / ( 0 & ~ ~ ) = expectantly, eagerly (Chao). 5 22 = to be; / j a t haj / (.~- ^ ) = or, or else. 22 22 £ O /aj na:w / (.<"<,' If! ) = noisy and bustling. 22 22 O /aj f a j / = worried, unsettled f e e l i n g . 22 22 O /aj tsaj / = st u f f y (e.g., of a room). ^ = to be able. O = bad, naughty. ft eJQj = to badger. a: j O = mile (from English 'mile'). iJ£ = to be close or near; with, along with, close to. 2 51 jifL = aspect marker of completion (e.g. /ja:k ma:j / n <i >~= to eat up the rest) . y = remaining l i q u i d and substance at the bottom of a bowl or cup. O = slow, pokey. 22 55 Jrfl /siw na:j / (̂  ~ ) = wife of a young master. )0 / n a : j 5 1 n a : j 1 3 / ( ~ ) = mothjr-in^law (used i n r i c h families; otherwise /on jan /Q A- )• O = t i r e d ; /nga:n 1 3 n a : j 2 2 / (SjL^ ) = sleepy. - 143 - 22 na:j O = to be connected to. 51 51 55 l a : j O / l a : j t a : j / = sloppy. l a : j 2 2 = to pour l i q u i d (e.g., sauce) over s.t, 55 L * t s ' a : j = to knead; to press on top of s.o. 13 ts ' a : j = to step on, to tread on. sa:j = to waste. 22 ka:j = to saw (something large). 1 3 J k'a:j ffy = to take to s.o. = a section (e.g., of an orange). n g a : j 5 5 ^ / n g a : j 5 5 t s i 1 3 / (~ 3~) = very poor people who l i v e near Shi-qi; e.g., such women are c a l l e d : / n g a : j 5 5 t s i 1 3 p ' o 5 1 / ( ~ - V $ ) . n g a : j 1 3 O / t s a : n g 5 5 n g i 5 5 n g a : j 1 3 / (#" O ~ ) = almost (.i.e., lack just a l i t t l e bit). . 5 5 ha:j O = to brush against l i g h t l y , to touch l i g h t l y . 51 ha:j O = itchy. 51 51 51 k'wa:j <*jjf|'\ = bad, michievous, vill a n o u s ; /k'wa:j jan / ( ~ / / 0 = v i l l a i n s . k'wa:j 2 2 $L /k'wa:j 2 2 t s i 1 3 / (.*>4*) = chopsticks. aw paw 5 5 = to squat. 51 paw tf> = to s t i c k out i n a lump; a swelling. 22 22 22 paw 0 /law paw / = clumsy. p'aw 2 2 /p'aw 2 2/ or /p'ung 5 5 p'aw 2 2/ = f l u f f y . maw55 / t i 2 2 maw55/ (^0'°) = ru f f i a n s , b u l l i e s . taw 5 1 0 /low 1 3 taw 5 1/ (£ ~) = father (si.) taw 2 2 O = to be near to; /taw 2 2 h i 1 3 k'an 1 3/ l~ or /taw 2 2 t s y 2 2 / (~&) = to be near. - 1 4 4 - = to rest. O = to unwrap. = to be angry. O = to c l u s t e r around, to hover over. O = plant from which the Annamese use the leaves to r o l l betel-nuts to chew (Chao). ^ = overcoat. = to cover over. •frxh = to p u l l up (e.g., pants); to l i f t up. 2 2 = CL for keys ( i . e . , a bunch) (Also: /nang /• O = to do s.t. taking advantage of circumstances, <Q = CL for stones, etc. ( i e . , a lump). /kaw 2 2 n i n 5 1 / ( ^ ) = l a s t year. O = to watch clo s e l y . O /aw 5 5 l a w 5 1 / = dirty , a:w /na:w k'a:w / {~&K) = to argue, to quarrel i.% = wrinkled. 0 /sa:w 5 5 k o n 5 5 / (.~#Ll ) = to air-dry ( i . e . , as opposed to drying i n the sun). 0 = to grab. 51 51 O /la:w ka:w / = hectic, m a hurry; / k i 5 1 l i 5 1 ka:w 5 1 la:w 5 1/ or / k l i 5 1 kla:w 5 1/ = noisy t a l k i n g — a s of a group of people. 0 /k'a:w55 k ' a 5 5 / = to make an "X". ft? = to scratch; /nga:w ha:j / = to scratch an i t c h . = to rake (e.g., leaves, grassjl °$^= to b i t e . 2 2 O = to enjoy (e.g., talking) (Also: /how /•£$-). 51 51 :i O /ha:w p'o / (~ dp ) = a loose woman; 51 51 51 /ha:w t'an t'an / = said of a loose woman i n reference to her mannerism, behaviour, etc. 13 22 2 22 O = j o i n t ; /saw a:w / ^ ) = wrist; /kiok a:w / = ankle. am = pump. 55 13 /f^ = to pump; /tarn s0j / {r^j^~) = to pump up water. 51 51 51 51 O / t i t i t a i r v t a r n / = said of s.o. who i s quick-footed or nimble. 13 35 ]$i^_= to throw, to throw away; /tarn ' / = to excel. 22 55 = to hang down; /tarn tgij / = shabby. ft? = to trample; /tarn 2 2 saw 1 3 tarn 2 2 k i o k 2 / (~ ^ jĴ P ) ^ = to stamp the feet up and down (as i n a tantrum). = a pool of water. «jg /t'am 5 1 t'am 5 1 t s y n 2 2 / (<~~ ) = to c i r c l e about. = to lure. O / t a j 1 3 t a k 5 nam 5 5/ (iMf-') = to be able to endure or withstand; 55 13 /nam tsaj / (~"14-) = s.o. who i s w i l l i n g to do favours, etc. for people. O = s o f t . to think. 0 = soaked. 55 51 ^,...51 * 55 51 51 0 = to shade over; /lam p'ung t'aw / tXL. = term for the r a d i c a l mi an " • > 7 " 'a r o o f . O = to collapse j ^ v = to p i l e up. O / t s i 5 1 t s i 5 1 tsam 5 1 tsam 5 1/ = muffled sound of voices. - 146 - O = a callous. O = to sprinkle (e.g., seasoning, f l o u r ) . ^,^/kam 5 5 t s i 1 3 / (.^4-) or /kam 5 5 j o n g 2 2 / C~ <flo = thus, i n t h i s way. 51 51 51 51 51 51 0 / k i l i kam lam / or / k l i klam / = denotes noisy state of a f f a i r s , . O / k i n g 5 1 l i n g 5 1 kam 5 1 lam 5 1/ or / k l i n g 5 1 klam 5 1/ =» noise from stamping feet, moving furniture, dropped objects, etc. /kam 2 2 t o 5 5 / l~ ) = t h i s much, that much. A'am 5 1 l o 5 1 / ) = spider. O = to lean with forearms resting on a surface (e.g., a r a i l i n g ) . 51 51 55 O A'am k'am ts'ia:ng / = denotes doing s.t. i n a big hurry.or f r a n t i c a l l y . {2 = a l i d , a cover; to put a cover or l i d over s.t. O = to slap i n the face. O = toothless. = to whisper; to grumble; /ngam 5 1 ngam 5 1 ts'am 5 1 ts'am / = grumbling sound. O /ham 5 1 p a 5 1 l a : n g 5 1 / or /ham 5 1 p l a : n g 5 1 / = a l l (fKi) . O = to bump against. O = to press down on. O = to f i s h s.t. out of one's pockets. a :m = tender beef; / n g 1 3 f a 5 5 na:m 1 3/ {Z.tw) = side pork. = to hold close to the body (e.g., as of s.t. very dear or precious); to hold s.t. large and bulky (e.g., a p i l e of clothes). = to step over. | /mun51 ts'a:m 1 3/ i f ] " ) = door-step, threshold. - 1 4 7 - $L / l a : p 2 1 ^ a : p 2 ts'a:m 1 3/ or / f a n 2 2 k i 5 5 2 2 y 1̂ >'a:m / ( ̂ |£ ~> ) = dustpan. (See also: /fan / O = bran for feeding pigs. ~ J u s t ' correct; /nga:m55 nga:m 5 5 s i n 2 2 / (~~ $lL) = just r i g h t , exactly. O /p'uk 5 nga:m 1 3 t s ' y 5 5 / or /p'uk 5 k a : j 5 5 t s ' y 5 5 / = to be i n a rush. ^Note: ̂ not an abusive term as i n Cantonese [ p ' u k k A i j ] 'Drop dead!'). . 22 22 (s /ha:m s i n / = telephone. O = to move up and down ( i . e . , standing and squatting i n succession). an I / t a : j 2 2 p a n 2 2 t s i o n g 2 2 / = elephant. 2 2 2 0 /pan ts'ot / = stubborn; sucker ( s i . ) . gill) = to sleep (Chao)—usually / m i 5 5 / (. ;jjL) i n Zhong- shan. j£ = ash u s e d to bury human waste, which i s l a t e r ^ used as f e r t i l i z e r ; / f a n 2 2 k i 5 5 / (~ ̂  ) = a shallow bamboo pan; / f a n 2 2 k i 5 5 ts'a:m 1 3/ (-$ %t ) = dustpan. O = to put s.t. down very roughly; to s i t down hard and abruptly. O = to pace; / f a n 5 1 s i o n g 3 1 f a n 5 1 l o k 2 / (~Jt~/S*) = to pace up and down. 0 / f a n 5 1 f a n 5 1 t s a n 2 2 / ( — % ^ ) - shivering, shak- ing; / l a : n g 1 3 f a n 5 1 f a n 5 1 / (;̂  ) = shivering cold. 0 / f a n 2 2 haw 2 2/ C ^ f c j = to move backwards. = to exasperate. ^ / l a n 5 5 s i o n g 2 2 l a n 5 5 l o k 2 / ( = to chatter incessantly (Chao). / t s ' a n 5 1 / , / i n 5 5 t s ' a n 5 1 / (>M^) or /puk 2 t s ' a n 5 1 / = dust. - 148 - O = to shuffle the feet Cas of old people). 55 55 55 ^ /ngan / or /ngan tsan / - small, stunted i n ^~ growth. ^ = to j i g g l e the feet while s i t t i n g down; /ngan 2 2 k i o k 2 f a 5 5 / l~ftif&). = the j i g g l i n g of the feet up and down while seated (considered impolite). a:n O = to brush or wave away; to dust. O = to p u l l down. O = to climb, to hold to, to hang on (Cha^ wonders i f i t i s an i n i t i a l change from /p'a:n /, with the same meaning) . 4" / j a 1 3 ma:n 5 1/ {f$~ ) = peevish (referring to c h i l d - ^ ren only); 55 51 /tiw ma:n / (.3 ) = unreasonable, s p o i l t , peevish. ^ / t a r n 5 5 t a r n 5 5 / ) = only. jffijj^ = a stand, a s t a l l (e.g., for f r u i t s , f i s h , etc.) (Chao). a& 55 55 2 sMi = to crawl, to creep; /la:n s i kat / = Get out! 22 55 22 55 0 /la:n l i la:n tsaw / = raggedy. 22 51 13 0 /la:n t'aw t'iarng / = s.o. who t r i e s to pre-vent trouble—extended to mean s.o. who wants to be the f i r s t to do s.t. O = quaint (of people). 0 /k'a:n 5 5 k'a:n 5 5 pow 5 5 pow 5 5/ ( — = pots and pans. 0 / k i 1 3 sap 2 nga:n 5 5/ (^xi^~) = said of someone who i s very, very old. = to crush by r o l l i n g . O /kwa:n 5 5 l a : n 5 5 taw 1 3/ (~ O-^ ) or / k l a : n 5 5 taw 1 3/ = a somersault (from Lo; Yang gave: / f a : n 5 5 kam 5 5 taw 1 3/ 0^ ). - 149 - ang pang 22 mang" 55 mang 13 fang tang 51 22 nang 55 22 nang ts'ang 22 sang 55 k 1 ang 55 k 1 ang 22 51 ngang 55 hang" hang 13 ang 51 i . 5 5 5 5 = to lean against; /aj aj pang""" pang"""/ C$Li^ / v "' % ' ). = describess .o. who always leans against s.t. 22 22 = to p u l l , to yank, 55 55 22 O /mang k'ang siong / (~ Oi% ) = a sour, puckered expression. |$L = bold, f i e r c e . 0 = to punch. O = to take care of (e.g., children); to take s.o. some place. O = to knock s.o. on the head. Q = a knob. 55 55 43 5 5 5 55 O /niw _ nang / = precarious; / k 1 I tak niw nang / (4l 1 f i'l"") = to be standing precariously (e.g., on a high ledge) . 51 55 iWi O = a knot, a j o i n t ; / t s * a : j nang / {%. ~ ) = a knot i n a piece of wood. O = said of thread hanging out. 22 55 O /nang kow / (~ & ) = to be on tip- t o e s . O = to moan and groan continuously; to complain. ||| = to blow the nose; /sang 2 2 p i 2 2 k o 5 5 / (~|p^[ ) = I to sneer. O = to l i g h t l y snap against s.t. (e.g., chinaware or glassware to f i n d out i f i t i s of good or poor q u a l i t y ) . 0 = to knock (e.g., door). 0 = powerful. 51 51 55 O /ngang ngang sia:ng / (•—/%. ) = moaning sound. O = to grumble or complain continuously. O = taut. O = to press against; /ang 1 3 k i o k 2 / (r? dtf) = to press against the foot Ce.g., as with a! piece of - 150 - 13 55 pebble i n the shoe); /ang sam / (~"/o>) = (s.t.). pressing the heart-—worried. O = to slam (e.g., door). O = to carry a heavy load of things. ^ = to wait; to l e t (Also: / t a n g 1 3 / ^ ). . a:ng O /niw 5 5 pa:ng 5 5 pa:ng 5 5/ (Jf'h^1^) = very slender. O = Bang!—sound of gunfire, f i r e c r a c k e r , slamming of the door, etc.; to slam (the door). 55 55 flfl O /p'a:ng hoj / [~ rfl ) = to divide up. O = to stretch s.t. across or over s.t. else. 55 2 O /ma:ng kiok / (~A*|') = leggings (Chao). $L ~ powerful; /ma:ng13 kwaj 1 3/ (~* %>) = a l i v e l y ghost. O = to kick o f f (e.g., blanket). Q = wool, yarn (from French 'laine'?). 55 55 A jt O /la:ng tsung / ) = alarm clock. O / l a : n g 5 1 l a : n g 5 1 s i a : n g 5 5 / (rJ~ ) = denotes doing s.t. very quickly and e f f i c i e n t l y . O = to s t u f f . • 5 3 - to give b i r t h (Cantonese: [sow ] ) ; /haw 2 2 sa:ng 5 5/ (4^.^) = the young people. = to scour. 51 13 51 2 O = to get i n the way; /ka:ng saw ka:ng kiok / (~% ~ k^f ) = to get underfoot, to get i n the way. 13 22 O / t a ka:ng / ) = a hold-up, a highway robbery. O = h i l l s i d e , mountainside (Chao; he thinks th_at perhaps i t i s a sound change from |§0 /kong /. 'mountain ridge'). = a cooking pot. - 1 5 1 - 5 5 kwa:ng O = to lock up, to imprison. i 5 1 kwa:ng o = to slam (e.g., door). 1 3 kwa:ng 0 = stem Ce.g., of a f r u i t ) (Also: / t a j 2 2 / ^ ) . , , 5 5 k'wa:ng ! a frame, a framework. 0 = a hoop. k'wa:ng 0. = to e n c i r c l e ; a c i r c l e (as to walk i n a c i r c l e ) . 2 2 k1wa:ng ** = to bump against; to tear one's o u t f i t by catch-ing i t on s.t. o = sound of metal. 5 1 wa: ng 0 51 51 /wa:ng man / = peevish (of a c h i l d ) ; cranky. , 5 1 , . 22 . 2 2 5 1 , , 2 2 , 2 2 /watng.^ / ,. /ka l a wa:ng / or /k l a l a wa:ng / = crosswise. 2 2 wa:ng A /ka:w 5 5 i 1 3 warng 2 2/ = the horizontal bars connecting the legs of a chair. ap t a p 5 o = a small earthen j a r or jug (Chao). u / t a p 5 t a j 5 5 t'aw 5 1/ l-sifoUi) = to lower the head. . 2 tap ik = to s t r i k e or h i t l i g h t l y (as of raindrops, or l i g h t hammering). f a p 5 0 = to cave i n , to collapse; lowered, collapsed. 5 ' " nap GO = dented; /nap 5 nap 5 kung 1 3 kung 1 3/ (-~~^ j§ ) rough or bumpy surface, warped. nap 2 yfe = sticky, moist (e.g., of fingers, body); / n i 2 2 n i 2 2 nap 2 nap 2/ (jIL S«Si — ' ) = sticky and moist (of body or body parts). l a p 5 O = to ,!^step into (e.g., water). t s a p 5 i t , = a small amount that can be picked up with the * fingers {e.g., of sand, r i c e , e t c . ) , a pinch of. sap 5 c$ /sap 5 s u j 2 2 / (rs&fy) = miscellaneous; /sap 5 t s a j 1 3 / (^4^-) = a grocery store. kap 5 O /kap 5 t s a j 1 3 / fa ) = a frog. - 152 - k'ap O = to slow-boil or stearn-cook—from sound of slow b o i l i n g water r a i s i n g and lowering the l i d of the pot. 2 2 22 0 /k'ap k'ap tsan / = to shiver. 5 t a. u j-j. • 4 . 1 / 5 55 5 .22, ngap *A^= to chatter i n d i s c r e e t l y ; /ngap sa:m ngap s i / (^js_~* CD ) = to make thoughtless comments. ngap 2 to nod the head. hap 5 <s" = to close the eyes; to doze, ft 5 ap = to cover (e.g., with a medical or herb patch). / * a: p 2 ta:p O = to request s.o. to do one a favour (Also: /t'okV-tb ) • t'a:p 2 O = a container; /maj 1 3 t'a:p 2/ ( i L ^ ) = r i c e bin. 2 na:p O = to catch on s.t. l a : p 2 i& = to c o l l e c t ; / l a : p 2 n a : j 1 3 / (~#) ) = a hoarder. O = to push up (e.g., sleeves). tsa:p ^.^/tsa:p / t s i k / (.MA'] ) = to lean sideways. 2 sa:p 'AlSL = t o c o o k by b o i l i n g . 0 = to sprain (e.g., ankle). 2 nga:p "î L = t o f o l d up or r o l l up (e.g., sleeves). 2 2 51 ha:p 0 /ha:p p'a / (~ ̂ ) = chin. a:p 2 = to pressure or force s.o. to do s.t. j a : p 2 ^ = to wave the hand. at 5 5 5 13 mat <Li /mat / or /mat ja / (r'J-i ) = what (Chao suggests perhaps sound change f r o m ^ /mat /).; /mat5 k a : j 1 3 / ) = why. f a t 5 0 = to sweep up (.e.g., onto the dustpan). O = a p i l e (e.g., clothes). O = to place, to put. O = to glare at s.o. O = to s l i p down. O = to dip into sauce. O = anxious, uneasy. JtL = to come o f f . O /wa:t 2 l a t 5 l a t 5 / (Jj| ) = very smooth. O = to erase or blot out Ce.g., wrong characters). O /m51 juk 5 t s a t 5 / ("4r^f' v) = not move at a l l (to denote something that i s d i f f i c u l t to move). j|l = a cork; to cork. 13 2 2 O /haw tsat tsat / (<X ^~ ) = to stammer. O = to hesitate; nervous; a j o l t . *<] = to pierce. O = to put the hair up; to get up (vulgar term); to lean over with derriere up. ^/U = to s l i d e over sideways i n a seated position; / s i 1 3 f a t 5 ngat 5 ngat 5/ ) = derridre always moving—can't s i t s t i l l . 5 5 5 0 /ngat ngat tsa t / = not open-minded. £Jfc = blunt, dull-edged. O A'wat 2 t'aw 5 1 l u 2 2 / C^SSlS) = a dead-end street. = kernel, p i t (.of a f r u i t ) ; /ngazn 13 wat 2/ (8fL~) = eye, eyeball; /nga:n 1 3 wat 2 t i n g 2 2 t i n g 2 2 / ( J / L ~%~^ 2\ = f $ a r i n § in t e n t l y , dec|g i n thought (Also: / t i n g h i j a t p'un siong / / 3 L i&-~ O ^ 0 ) • /nga:n 1 3 wat 2 t a t 2 t a t 2 / ( i i ^ O O ) = to glare at s.o. l3 = the sun. - 154 - a: t o p'a:t O = a small amount of l i q u i d . 5] 51 2 2 51 2 O / P ' i l i p'a:t l a : t z / or / p ' l i p ' l a : t V = sound of water being splashed (as by young children) 2 ta : t O = to f a l l (from a great height). = to toss c a r e l e s s l y . ^ = CL for locations. f a : t 2 0 / f a : t 2 h a 1 3 t ' a : t 2 s i a : n g 5 5 / ) = sound of s l i p p e r s . na:t 2 = to have a l i g h t burn (e.g., on the hand from cooking, or by a l i t c i g a r e t t e ) . l a : t 2 O / l a : t 2 t ' a : t 2 / = d i r t y . O = to v i s i t s.o. 2 sa:t O = to keep badgering s.o. for s.t. (as of a c h i l d ) . nga:t = stench of urine. O to s l i d e back and forth i n the seat. 5 5 51 £ ha:t O /ha:t t s 1 i / (r^ *jf J ~ a sneeze. a : t 5 = to press. 2 2 2 a:t = to be pressed; /a:t l i k / fj ) = pressure. ak mak5 /mak5 t'aw 5 1/ (~S-t.) = trade-mark (from English 'mark'); good-looking (Chao). 2 tak 0 = to n a i l . 5 55 55 5 5 nak O / n i n i nak nak / = describes walking on t i p -toes, as on high heels. O = great, great grandchild. 2 -nak 0 = to tread on, to step on. 22 2 0 / n i nak / = sti c k y . 5 5 5 lak %b /lak k'ak / = jerky (of speech); rough (of surface) - 155 - 5 rJr" 5 5 sak fjg- /sak ak / = a hiccup. = great grandchild. ngak 5 ofo = to cheat. 2 ak = S - f i n a l PRT (Chao). 2 22 2 2 k'wak O /nga:ng k'wak k'wak / = denotes s.t. very hard and s t i f f (e.g., of a piece of stal e bread). 2 wak O = to swing one's arms (while walking). a :k 2 ma:k O = to open (e.g., mouth, eyes). 2 fa:k O = to whisk, to beat (e.g., eggs); to swing one's arms back and forth. 5 5 5 5 5 ta:k 0 / t i k t i k ta:k ta:k / = c l i c k i n g sound of high heel shoes. t'a:k 5 O /ma:n13 t'a^k 5 h a ^ 5 / (0jfe> ~ ) = night-time (Also: / j a ma:n / ^ Qfij ) . l a : k 5 0 / p i k 5 l i k 5 pa:k 5 l a : k 5 / or / p l i k 5 p l a : k 5 / = crack- l i n g sound—of f i r e , or s t r i n g of fir e c r a c k e r s . l a : k 2 $/) /mok2 ts'a:k 2 l a : k 2 / (.&) 0 ) = bard (to waist). 51 51 2 2 O / l i l i la:k la:k / = low r u s t l i n g sound of paper or si m i l a r sheets of material. 2 sa:k O = to chop (something large). Q = a piece (e.g., f i e l d ) . 2 ka:k Q = to sieve. $r = abnormal or strange (of person). a:k 2 = a bracelet. 2 4 k'wa:k el* = to make a loop (e.g., i n sewing); to wrap around and t i e with s t r i n g or cloth; to la t c h . j a : k 2 o'U = to eat; / j a : k 2 i n 5 5 / ) = to smoke (cigarettes, etc.) - 156 - long = to peck; to h i t against a surface with a small, sharp object (e.g., a small chip of glass).. iok = to chop—a singular action, as to chop s.t. i n h a l f . = to chop repeatedly to small pieces, as to make minced meat. 2 2 5 / t s i o k j a t ' / (/̂  8 ) = yesterday. i a 55 13 = beer (from English word). (Also: /pia tsaw / .~ 1% ) • / s a : n 5 5 p ' i a 5 1 / (ti* — ) = h i l l s i d e . = to carry s.t. that hanejg from the arm (e.g., a purse) (Cantonese: Iwaxn ] ) . = what (from /mat5 j a 1 3 / - t i f f ). = slanted; / w a : j 5 5 w a : j 5 5 m i a 1 3 m i a 1 3 / """̂  ) = to denote s.t. that i s askew or slanted. 13 13 / l i a f i a / = sloppy, d i r t y . 13 13 22 / t i a / or / t i a tiw / = to be s p o i l t or pampered. (See: / f i a 1 3 / ) = subordjnative s u f f i x (Chao) (Zhong-shan usually uses /ko /.) = a part of a tree, e.g., a branch. = stool (vulgar term). = to stagger. i a : w = to toss away. 55 55 13 O /lia:w sung low; / O' O v&) = term for north-ern Chinese p e o p l e — a c t u a l l y from lao xiong {JLf^-i) 'elder brother", a p o l i t e term of address. O = to grab. 51 51 55 O /kia:w kia:w sia:ng / (~ f̂-e, ) = denotes noisy rowdiness. O = to s t i c k gut (e.g., l i d of an opened t i n can). (Also: /hiw / ) . ia:m O = a s l i g h t cough, or the sound of a l i g h t cough.. ,ia:ng 13 * O /pia:ng fa:n/ (r* 1&>) = to give back. O = to throw s.t. at s.o. 0 = the back (of a c h a i r ) . jfJi /jaw 5 1 p ' i a : n g 5 1 t s a j 1 3 / (jlfc^te") = a c h i l d brought to a second or subsequent marriage by a woman; / t ' o 5 5 jaw 5 1 p ' i a : n g 5 1 / (#b;»£~) = to bring such a c h i l d to a subsequent marriage. (Uncomplimen- tary terms). 0 = a piece or a sheet (e.g., paper, land). 0 = sound (e.g., of a drum). O /tia:ng / or /tia:ng fong / (.-»7> ) = a place. 0 /Verb + t i a : n g 2 2 t i 2 2 / (Verb + ~ = to have just V'd. ^» = an open space, an area of l e v e l ground; 51 51 /wo t'ia:ng / (.^^ ) = area for threshing grain (Chao). 22 51 / t i t'ia:ng / (>€J^) = area i n front of the house for drying grains, etc. i n the v i l l a g e . O = to hide. $#J = pretty. - 158 - kiarng 3 ĵ̂  /kiazng / or /kia:ng sia:k / (~ <ta ) = to be very careful with, to handle with great care and gentleness. k ' i a z n g 5 1 O / k ' i a : n g 5 1 / or / l i a : n g 5 1 k ' i a : n g 5 1 / (yf ~ ) = auspicious. 13 22 51 13 t'ia:ng 0 /la : n t'aw t'iarng / = s.o. who does s.t. f o r another person which he finds embarrassing or disagreeable to do. xa:p ti a : p O = to taste a l i t t l e , ( a s to try i t ) . k i a : p 5 O /ngo 2 2 tow 2 2 k i a : p 5 k i a : p 5 s i a : n g 5 5 / £'] ~ ~ ) = hungry to the extent that the stomach i s pro- testing, i.;e., very hungry. i a : t p i a : t 2 O / l a : n 2 2 p i a : t 2 p i a : t 2 / ('/{ĵ  ~ ~* ) = soft and mushy. p ' i a : t O / p ' i a : t p ' i a : t / = derriere (speaking to a young c h i l d ) . 2 p ' i a : t O = a mass of soft, mushy substance; / l a : n 2 2 p ' i a : t 2 p ' i a : t 2 / 1%$""* ) = soft and mushy; / p ' i a : t 2 s i 1 3 / {^M-) = said of s.o. who s i t s there and does not f e e l l i k e moving. (Very uncom- plimentary term). t i a : t 2 O / l a : n 2 2 t i a : t 2 t i a : t 2 / #|0 ~ ~ ) = soft and mushy. i a : k p'ia:k O = to throw around car e l e s s l y (e.g., one's clothes) 2 tia:k O = t o chase s.o. away. l i a : k 2 ;i& / t a : j 2 2 l i a : k 2 / (A~ ) and / s a j 2 2 l i a : k 2 / (i.©~) /'-̂ ~ = names of v i l l a g e s (Chao) . 5 13 5 -lia:k O = smart, clever; /la:n l i a : k / = smart alecky. t s V i a : k 2 = pa i n f u l ; /t'aw 5 1 t s ' i a : k 2 / ( S J l ^ ) = to have a - 159 - " headache. 13 2 = upset, hurt; to be mad at; /nga:n ts ' i a : k / r J) = to be envious. = to love (e.g., a child) (Can also be used to abuse s.o. by intending the opposite; i . e . , by the use of sarcasm.) (Syllabic). 51 55 >* P£ = not; /m koj / lr/> ) = please; thank you (after s.o. does a p o l i t e deed, such as handing one a cup of tea, a bowl of r i c e , etc.) ffi / n g 1 3 tsok 2 low 1 3/ = undertaker. (Also: /kun 5 5 t s ' o j 5 1 low 1 3/ kf & > • [ e ] 0 S- or c l a u s e - f i n a l PRT. (Chao). [ e j ] fX^= exclamation of affirmation (Chao). [el O = S - f i n a l PRT. (Chao). 0 = S - f i n a l PRT. (Chao). [a] = Q-PRT for yes-no Q's and negative Q's. (Chao). a 0 /mung51 k'wa 5 1 k'wa 5 1/ ($$^~~') = very blurry. - 160 - /pa:k 2 j a 5 5 / (& ~ ) = father (term of address y = to poke with a long object (e.g., a closed umbrella) - 161 - PART I I . DIACHRONIC STUDY CHAPTER 3. ANALYSIS OF MODERN REFLEXES TO HISTORICAL CATEGORIES. Bernhard Karlgren, a pioneer i n applying Western l i n g u i s t i c methods to the study of the h i s t o r i c a l Chinese sound system, established two stages i n the language: "Ancient Chinese" and "Archaic Chinese". As outlined i n his Compendium (.1954:212), Karlgren designates "Ancient Chinese" to be the language of c i r c a 600 A.D., as c o d i f i e d i n the Qie-yun "gift rhyme dictionary, which he i d e n t i f i e s as the d i a l e c t spoken i n Chang-an, Shen-si. "Archaic Chinese", on the other hand, refers to the language spoken i n the He-nan region during the f i r s t Zhou centuries (from 1028 B.C.), based partly on the rhymes i n the S h i - j i n g ('Book of Odes') and other early manuscripts, and p a r t l y on xie-sheng characters (compounds containing a " r a d i c a l " (or " s i g n i f i c " ) and a "phonetic"). Following Karlgren, most Chinese phonologists continue to reconstruct Ancient Chinese as a single stage i n the language based on the Qie-yun. Pulleyblank, who prefers the'terms "Middle" and "Old" Chinese to Karlgren's "Ancient" and "Archaic" Chinese, further subdivides Middle Chinese into "Early Middle Chinese" (EMC) , and "Late Middle Chinese" (LMC). Pulleyblank (.1977:12) emphasizes that the s h i f t from EMC to LMC i s not simply a case of h i s t o r i c a l evolution of the language, but represents a major s h i f t i n d i a l e c t base. Pulleyblank (1970:204, 1977:4) considers "Early Middle Chinese" to be the language of the Qie-yun, which he iden- t i f i e s as the standard Mandarin of the Northern and Southern Dynasties i n the s i x t h century i n the courts of Lo-yang\ and - 162 - Nan-jing, and "Late Middle Chinese" the language of the rhyme tables, representing the speech i n Chang-an during the middle and l a t t e r part of the Tang dynasty (618-907 A.D.). Although h i s t o r i c a l l y the two languages do not constitute a continuum i n the evolution of Chinese, i n formulating deriva- t i o n a l rules, Pulleyblank (1970-71:204, 1977:12) nevertheless treats Late Middle Chinese as though i t has developed from Early Middle Chinese. He j u s t i f i e s t h i s treatment on the following grounds: since the two languages are clo s e l y related and t h e i r phonological categories are on the whole quite congruent, Late Middle Chinese must have descended from a language the phonological system of which was very si m i l a r to that of Early Middle Chinese. Hence, i t makes sense to est a b l i s h rules to derive the former from the l a t t e r even though t h i s does not quite exactly r e f l e c t h i s t o r i - c a l r e a l i t y . According to Pulleyblank, i t i s Late Middle Chinese which formed the basis for both Sino-Japanese (Kan-on) and Sino-Korean. Moreover, with the exception of Min, a l l the modern Chinese dia- l e c t s can be traced back to Late Middle Chinese; that i s , the phonological system of the modern d i a l e c t s — i n c l u d i n g both l i t e r a r y and c o l l o q u i a l l a y e r s — c a n be explained on the basis of the phono- 2 l o g i c a l categories of Late Middle Chinese. The primary source for the reconstruction of Late Middle Chinese i s the rhyme tables, which have evolved from e a r l i e r rhyme d i c t i o n a r i e s which, i n turn, are expansion and revisions of the Qie-yun of Fa-yen Lu, completed i n 601 A.D. The rhyme tables have e s s e n t i a l l y extracted the information contained i n the rhyme dic t i o n a r i e s on the pronunciation of the language, and have /•; - 163 - systematically arranged the information i n the form of a g r i d . Thus, tones and f i n a l s , for example, are placed along the v e r t i c a l axis, while i n i t i a l s are arranged along the horizontal one. Although these rhyme tables do not contain phonetic descriptions of the sounds i n the language, they do contain descriptive terms for the place and manner of a r t i c u l a t i o n of the phonological cate- gories, of which some have to be interpreted. Kai-kou versus he- kou, for instance, has been understood to be a d i s t i n c t i o n of the absence of l a b i a l i z a t i o n versus i t s presence i n the f i n a l s . Other terms include ya-yin % % 'back-tooth sound' for velars and cun- yin % ' l i p sound' for l a b i a l s . The term qing \ 'clear' i s used for i n i t i a l s reconstructed as p l a i n , unvoiced segments, and ci-qing )%, }\ 'second clear' for the aspirated ser i e s . Zhuo or 'muddy' i n i t i a l s are usually assumed to be voiced i n i t i a l s , although Pulleyblank (1970-71:210-211) reconstructs such i n i t i a l s as p l a i n consonants accompanied by voiced aspiration i n Late Middle Chinese which have evolved from e a r l i e r voiced i n i t i a l s (e.g., the Ding ^ i n i t i a l i s reconstructed as LMC *tfi-< EMC. *d-). In discussing the reconstructed values of h i s t o r i c a l Chinese phonological categories, Pulleyblank's reconstruction of Early and Late Middle Chinese w i l l be adopted; the following sec- t i o n w i l l therefore be a b r i e f description of Pulleyblank's system. No attempt w i l l be made to discuss t h e o r e t i c a l issues and contro- versies connected with c e r t a i n reconstructions. Following the introductory description of Middle Chinese phonology, we w i l l proceed to analyze Zhong-shan reflexes of these Middle Chinese sounds according to the organization of the phonological categories found i n the d i a l e c t survey l i s t , the Fang-yan Diao-cha Zi-biao. . - 164 - The aim of the chapter i s to demonstrate the general pattern of correspondences of modern Zhong-shan to h i s t o r i c a l categories. As a r e s u l t , rather than try to account for a l l the exceptions present i n the data on modern Zhong-shan, there w i l l be an e f f o r t to discuss only some of these i r r e g u l a r sound changes. Emphasis w i l l be placed on the regular pattern of correspondences. As i n the synchronic analysis, the subsections w i l l focus on the main categories of i n i t i a l s , f i n a l s and tones. S p l i t s and mergers of various phonological categories i n Zhong-shan w i l l be discussed, as well as some of the more intere s t i n g problems and exceptions observed. The compilation of Zhong-shan data for the present diachronic analysis i s placed i n Chapter 4. 3.1. I n i t i a l s 3.1.1. Reconstructed Values of Middle Chinese I n i t i a l s In the rhyme tables representing Late Middle Chinese, there were o r i g i n a l l y t h i r t y i n i t i a l s which were l a t e r increased to t h i r t y - s i x . These i n i t i a l s are l i s t e d i n Chart 4 on the f o l - lowing two pages. The l a t e r s i x i n i t i a l s are marked by asterisks (*). As we have already seen i n Chapter 1, the characters repre- senting the d i f f e r e n t i n i t i a l s serve as the names of these i n i t i a l s . The reconstructed values assigned by Pulleyblank (.1977:64) are also given i n the chart. In the set of t h i r t y - s i x i n i t i a l s , the l i g h t l i p sounds (qing cun-yin), or dental l a b i a l s , arose out of the corresponding l a b i a l s e r i e s , or the heavy l i p sounds (.zhong cun-yin) . The Feng %~ (LMC *ffi-) i n i t i a l had s p l i t from the Bing j L CEMC *pf\-> . i n i t i a l , and the Wei 4^- (LMC *v-) i n i t i a l from the Ming B$ Chart 4 . The 36 I n i t i a l s of Late M i d d l e Ch inese . Qing Zhuo (c lear ) (2nd -c lear ) (muddy)(not -c lear -not -muddy)(Clear) (muddy) Qing C i - q i n g Zhuo Bu-qing-bu- -zhuo CUN-YIN sound) LABIALS: a) Zhong (heavy) Bang Pang B ing / Ming m LABIODENTALS: b) Qing %t ( l i g h t ) F e i * Fu* Feng1* ^ f « p ) | ^ ; f « p « ) ^ f « b) Wei* SHE-YIN (tongue sound) DENTALS: a) She-tou£&Jt Duan (tongue head) t Tou Ding N i RETROFLEXES: b) She -shang ^ J ; Zh i Che (tongue up) t r ĵjfc t̂r* Cheng Niang* CHI-YIN Si % ( f r o n t - t o o t h sound) DENTAL SIBILANTS: a) Ch i - tou&SS. J i .ng k t : ( tooth head) RETROFLEX SIBILANTS: b) Zheng-chi i fe} Zhao ( t rue tooth) 8§ ts Qing :> t s ' Chuan I t 5' Cong ^ t s f i Chuang* X i n X i e fCP s ^ sfi Shen Chan I 5 ty'1 (Chart 4. cont'd.) Qing Ci-qing BAN-SHE-YIN DENTAL: (half tongue sound) BAN-CHI-YIN RETROFLEX: % ** (half tooth sound) ?i YA-YIN VELARS; Jian (back tooth 7 7 sound) HOU-YIN GUTTURALS: Ying Xiao (throat *"/ 7 sound) * Not included i n the o r i g i n a l 30 i n i t i a l s . Ri lz) r Qun kfi Xia xn Yu »4TV # - 167 - (EMC *m-) i n i t i a l . In the case of i n i t i a l s F e i $ (LMC *f-) and Fu (LMC *f-) / although they are distinguished i n the series of * f t h i r t y - s i x i n i t i a l s as having arisen from the Bang (EMC *p-) and Pang ffi (EMC *p'-) i n i t i a l s respectively, Pulleyblank regards that d i s t i n c t i o n as almost c e r t a i n l y just a h i s t o r i c a l one based on fan-qie d i s t i n c t i o n s i n the Qie-yun and reconstructs both F e i and Fu as LMC * f - . Pulleyblank (1970-71:217-218) argues that the d i s t i n c t i o n between Fei and Fu had i n fac t been t h e o r e t i c a l from the st a r t ; they were kept d i s t i n c t p r e c i s e l y because they had arisen from d i f f e r e n t Early Middle Chinese i n i t i a l s . The Chuang AL i n i t i a l (LMC *(t ) ? f i - ) , according to Pulley- blank, was not phonemically d i s t i n c t from i n i t i a l Chan-'T^ (*gfi-). I t i s placed i n the muddy a f f r i c a t e column i n the Yun-jing, under the dental s i b i l a n t Cong (LMC *tsfi-) and corresponds i n Grade II to EMC *d^-but i n Grade III to EMC *?•<-. The Chan i n i t i a l , which i s placed i n the muddy f r i c a t i v e column, under Xie (LMC *sfi-) i s mainly confined to Grade III where i t corresponds to EMC -d?. Since *d? and *f had merged i n LMC as the f r i c a t i v e *gfi, the authors of the rhyme tables had only the fan-qie spellings to di s t i n g u i s h them and mistakenly assigned the o r i g i n a l a f f r i c a t e to the f r i c a t i v e column and vice versa. Before high vowels ( i . e . , i n Grade III and Grade II of the Zhi rhyme group), the pronunciation of both i n i t i a l s was mostly as a f r i c a t i v e , while before non-high vowels ( i . e . , i n Grade II i n other rhyme groups), the pronunciation was mostly as an a f f r i c a t e ; but there i s evidence of free v a r i a t i o n i n both cases (1970-71:223). A further small complication i s that there i s also a voiced r e t r o f l e x f r i c a t i v e i n i t i a l *\ i n EMC Occurring only i n two words i n the Zhi rhyme group. This would - 168 - have been pronounced *sn-.in LMC. I t was placed as Grade II of i n i t i a l Chan i n the Yun-jing. The Yu (LMC x#-) , or zero, i n i t i a l Pulleyblank regards as probably a weak velar f r i c a t i v e y ( l i k e the "zero" i n i t i a l i n Mandarin), which would perhaps produce less confusion i f omitted altogether i n transcribing reconstructed forms. Although there i s a phonemic d i s t i n c t i o n between i n i t i a l s Ying <(j and Yu '&\ during Late Middle Chinese, i t had generally disappeared between late Tang and Yuan times. Vestiges of the d i s - t i n c t i o n could nevertheless be discerned i n the phonemic contrast between Yin- and Yang-ping tones. (Pulleyblank, 1970-71:227). Since the d i a l e c t survey includes i n i t i a l s which Pulley- blank has reconstructed as part of the phonological system of Early Middle Chinese, i t i s also necessary to discuss t h i s e a r l i e r layer of the language. Pulleyblank (1977:80) posits thirty-nine i n i t i a l s , which are shown i n Chart 5 overleaf. He uses some of the same characters from the standard l i s t of the t h i r t y - s i x Late Middle Chinese i n i t i a l s where possible, and supplements these with additional characters to cover those d i s t i n c t i o n s which are not present i n the l a t e r stage. Early Middle Chinese i s reconstructed as a diasystem which accounts for both the northern and southern d i a l e c t areas. Where the values reconstructed for the two groups d i f f e r , i t i s the southern d i a l e c t a l form which i s enclosed i n parentheses i n the chart. Whereas Pulleyblank (1977:12) regards Late Middle Chinese as seemingly closer to the northern variety of Early Middle Chinese, i t i s the modern southern Chinese d i a l e c t s which he suggests have closer correspondences to cert a i n d i s t i n c - tions i n the southern branch of Early Middle Chinese. - 169 - Chart 5. Early Middle Chinese I n i t i a l s . LABIALS: DENTALS: Bang Pang Bing Ming b 0fl m Duan Tou Ding Ni -fa n Yun L a i RETROFLEXES: Zhi Che Cheng Niang fa t r 4&tr« j f-dr j ^ n r (t) (t<) (d) (n) DENTAL Jing Qing Cong SIBILANTS: $j| t s ; f f t s ' ^ d z t z ) PALATALS: Zhao Chuan Chan Ri $8 t<? j f tQ< ^ d? ® Qi (?) RETROFLEX Zhuang Chu Chuang SIBILANTS : | i t ? : ^ ts « ^ j d z. ( $ Xin /O's Xie Shen Shen Yang (or Yi) Shan S i h 5 / f ^ * VELARS; Yl Jian Qi Qun Xiao Xia GUTTURALS: Ying * #(ft) (. ) : Southern d i a l e c t a l form. What eventually became the muddy i n i t i a l s i n Late Middle Chinese were f u l l y voiced consonants i n Early Middle Chinese. As one may observe from the reconstruction of a g l o t t a l stop for the Ying ^ i n i t i a l i n both Charts 4 and 5, t h i s p a r t i - cular i n i t i a l had remained stable during the two stages of Middle Chinese. - 170 - The reconstructed value of *x- for the Late Middle Chinese i Xiao i n i t i a l projects back to *x- for the northern d i a l e c t and *h- for the southern one, while that of *xfi- for the Late Middle Chinese Xia j^. i n i t i a l goes back to * - for the northern form and * f i - for the southern. The Early Middle Chinese Yang % (Yi vk i n the d i a l e c t survey) and Yun i n i t i a l s are merged as the Late Middle Chinese Yu 9$\ (*#-, or zero) i n i t i a l . In the rhyme tables of Late Middle Chinese, what were Early Middle Chinese Yang and Yun i n i t i a l s are i n complementary d i s t r i b u t i o n i n Late Middle Chinese: Yun occurs i n Grade III and Yang i n Grade IV. y A Pulleyblank reconstructs Y i % as the true zero i n i t i a l for Early Middle Chinese, which consists of a closed class of two e n c l i t i c p a r t i c l e s , y i %, and yan -^j . They are subsumed under the Yu °H'J i n i t i a l i n Late Middle Chinese. As mentioned above, the Early Middle Chinese p a l a t a l s i b i l a n t s became merged with, t h e i r r e t r o f l e x counterparts, y i e l d - ing the Late Middle Chinese "true front-tooth' i n i t i a l s . The two Early Middle Chinese series are i n complementary d i s t r i b u t i o n i n Late Middle Chinese: the p a l a t a l series occurs i n Grade III rhymes, while the r e t r o f l e x series i s found i n Grade IV. I t i s proposed by Pulleyblank (.1970-71:219) that the complementary d i s t r i b u t i o n came about when medial was l o s t i n words which bore an Early Middle Chinese r e t r o f l e x s i b i l a n t i n i t i a l . It should be noted that the Ni j/£j and Niang i n i t i a l s i n the two stages of Middle Chinese are combined i n the survey l i s t , probably as a r e s u l t of accepting the proposal put forth by Y.R. Chao (.1940:210) that these two i n i t i a l s were merely i n - 171 - complementary d i s t r i b u t i o n , and not actually phonemically d i s t i n c t . Pulleyblank (1970-71:214-216), on the other hand, argues for a phonemic d i s t i n c t i o n between the two i n i t i a l s which, by the Mongol period (Yuan dynasty, 1279-1368) was on the verge of disappearing. The Ri l3 i n i t i a l i s reconstructed by Pulleyblank as a p a l a t a l nasal * i > - i n Early Middle Chinese which became, i n Late Middle Chinese, a lax, voiced r e t r o f l e x * r - , as i n the modern Mandarin pronunciation of that i n i t i a l . The s h i f t of the Ri i n i - t i a l from p a l a t a l to r e t r o f l e x i s part of the more general s h i f t i n the language, as witness the merger of the p a l a t a l s i b i l a n t s with those of the r e t r o f l e x series i n Late Middle Chinese. The velar i n i t i a l s remained stable from Early to Late Middle Chinese. It i s important to r e a l i z e that, on the whole, the d i a l e c t survey l i s t maximizes ce r t a i n d i s t i n c t i o n s found i n the two stages of the language, and not others. The f i n a l r e s u l t i s a t o t a l of forty i n i t i a l s i n the d i a l e c t survey. These i n i t i a l s are shown in Chart 6 on the next page. In the chart, Late Middle Chinese serves as the base. In some cases, however, i t i s necessary to make i t clear that a certain series only occurs i n one stage of the language. For the sake of c l a r i t y , the p a r t i c u l a r stage—LMC or EMC—is therefore s p e c i f i e d at the same time that the c l a s s i f i - cation of the series i s given; f o r example, Labiodentals (LMC), Retroflex S i b i l a n t s (EMC), and P a l a t a l S i b i l a n t s (EMC). The l a t t e r two sets of i n i t i a l s , EMC Retroflex ... and P a l a t a l S i b i l a n t s , subsequently merged as LMC Retroflex S i b i l a n t s . Since the diachronic study of Zhong-shan i s based on the l i s t of characters prepared i n the Fang-yan Diao-cha Zi-biao, the - 172 - Chart 6. Middle Chinese I n i t i a l s i n the Fang-yan Diao-cha Zi-biao. BILABIALS: (LMC) Bang Pang Bing Ming 1 P P' £ Pft Aft * LABIODENTALS: (LMC) DENTALS: (LMC) DENTAL i: SIBILANTS: (LMC) RETROFLEXES (LMC) RETROFLEX SIBILANTS: CEMC) PALATALS : (EMC) Fei * f ( * P ) Duan Zhi & » t r Zhuang & tg (>LMC tg-II) Zhao 8 ^ t 9 0 LMC tg: -III) Fu Tou i t * ' Q|ng :,% t s ' CHE ttl.tr' Chu ^ tg' (>LMC tg'-H) Feng £ f ft Ding * tft (<d) Wei (cm) Ni 3/6 n Cong Xin 4*c.tsfi /O' s (<dz) Cheng /Niang^ ( tdr Chuang Shan /Ŵ dzL it» g (>LMC (>LMC (t)gft g-II) -ID Chuan y to* (>LMC tg' -III) Shen Chan d? (>LMC gft (>EMC g -III) - III Xxe If sft 1 <5z) L a i Shen Ri ft* B i> (> LMC gft (-> LMC r ) -III) VELARS: J'ian Qi Qun Y i Xiao Xia (LMC) j£,k $ k< jpfkfl ( t o g - X \ GUTTURALS: Ying • (LMC) f i ? *B»I# (III; <:EMC w IV c EMC j»A j ). phonological categories set up i n that survey w i l l be the launching ground for the study of modern Zhong-shan reflexes to the h i s t o r i - c a l system. Thus, Chart 6 shows the Middle Chinese i n i t i a l s found i n the survey l i s t , while Zhong-shan reflexes of these i n i t i a l s are summarized i n Chart 7.on the following two pages. - 173 - .Chart 7. Zhong-shan Correspondences to the Middle Chinese I n i t i a l s i n the Fang-yan Diao-cha Zi-biao. ZHONG-SHAN INITIALS COLLOQUIAL LITERARY M.C. InitiaJ?^,,^ Ping > Shang Qu, Ru Ping Oblique Ui Bang P — *c U H Pang P' s § I . I Bing P' P P' P r~1 m Ming m C/) Fei P/P' (L MC ) B I0 D E N 1  Fu Feng P' f < Wei i n Duan t Ui Tou t' a EH Ding f t f t Q Ni/Niang n Lai 1 Ui EH Jing ts LM C)  SI BI LS  Qing t s ' i LM C)  SI BI LS  Cong ts ' ts t s ' ts EH Xin s W Q Xie ts ' 1 ts t s ' ts W W X Zhi ts — H a cx, Che ts i HH1 o EH - 1 Cheng ts' ts t s ' ts - 174 - (Chart 7. cont'd) ZHONG-SHAN INITIALS ^ ^ ^ ^ ^ M.C. I n i t i a l ^ ^ COLLOQUIAL LITERARY Ping, Shang Qu, Ru Ping Oblique (E MC ) RE T.  S IB IL AN TS  Zhuang ts (E MC ) RE T.  S IB IL AN TS  Chu t s ' (E MC ) RE T.  S IB IL AN TS  Chuang ts' ts t s ' ts (E MC ) RE T.  S IB IL AN TS  Shan (Si) s (E MC ) 1 PA LA TA LS  | Zhao ts (E MC ) 1 PA LA TA LS  | Chuan ts ' (E MC ) 1 PA LA TA LS  | Shen Shen Chan s (E MC ) 1 PA LA TA LS  | Ri n g ~ j (L MC ) 1 VE LA RS  1 Jian k (L MC ) 1 VE LA RS  1 Qi h k' (L MC ) 1 VE LA RS  1 Qun k' k •v k' . k (L MC ) 1 VE LA RS  1 Y i ng LITERARY AND COLLOQUIAL; Kai-Kou He-Kou M.C. I n i t i a T ^ ^ s ^ I II j III | IV Tong, I | IlijIIlj IV Other I | i i f i l i | : i v (L MC ) .V EL AR S Xiao . Jfc. h ~ f (L MC ) .V EL AR S X ia h ~ w I (L MC ) 8 1 GU TT UR AL S!  Ying 0 0 j 0^ w I (L MC ) 8 1 GU TT UR AL S!  Yu (III) j I (L MC ) 8 1 GU TT UR AL S!  Yu (IV) 0~j j • - 175 - 3.1.2. Zhong-shan Correspondences to Middle Chinese I n i t i a l s 3.1.2.1. B i l a b i a l s (LMC) The regular correspondent to the Bang ^ (*p-).' i n i t i a l i s /p/, e.g., 8-19 /pa / 'to take' 15-23 ft / 13 / /PU / 'to mend' 40-1 ft /pu: / 'a cup  1 77-1 / • 13 / /piw / 'to display' 186-1 /pak 5/ 'north' There are nonetheless about a dozen words which are pro- nounced with an aspirated /p 1/ i n i t i a l . Most of these exceptions can be explained i n terms of analogical readings of more common characters; that i s , readings based on characters which, appear graphically s i m i l a r . In the case of the word po (4-18) 'lame', 55 the c o l l o q u i a l reading of /paj / has preserved the regular Bang 13 correspondent while the l i t e r a r y form, /p'o /, has acquired an aspirated i n i t i a l . The word bao (.183-1) 'to s t r i p o f f , also contains the aspirated /p'/ i n i t i a l for the l i t e r a r y reading of the word, and 2 i s pronounced /p'ok /. In the c o l l o q u i a l layer, however, i t i s 5 2 pronounced /mok / or /mok /, with an /m/ i n i t i a l . One other word f*. 2 i n the data containing an /m/ i n i t i a l i s bo 'SJ (196-6) /ma:k / 'to break open with hands'. The alternative form has i n i t i a l 2 /p'/ and i s read /p'ia:k /. Zhong-shan and Cantonese do not d i f f e r much with regard to the modern reflexes of the Bang i n i t i a l . Generally, the same word i n the two d i a l e c t s would exhibit the same correspondent. - 176 - There are few differences. One such difference i s i n the word 22 bao jfo (73-16) 'leopard', which i s phonetically IpAi-w ] i n 44 Zhong-shan, and [ p ' A i w ] i n Cantonese. In thi s case, Zhong-shan has preserved the regular r e f l e x of /p/. The Pang C*p'-)' i n i t i a l i s usually pronounced /p'/ in modern Zhong-shan, as i t i s i n Cantonese. Examples are: (2) 5-22 XiiL / P ' ° 2 2 / 'broken' 17-9 Aft /p'u 2 2/ *a store' 111-16 *9 /p'an 1 3/ 'personality' 131-4 ^ / p ' i n 2 2 / 'a s l i c e ' 194-5 /p'a:k 2/ 'to clap' It should be noted that although many of the Middle Chinese i n i - t i a l s enter the same general correspondence patterns i n Zhong-shan as i n Cantonese, with regard to i n d i v i d u a l words the two d i a l e c t s may show some differences, as already observed with the Bang i n i t i a l . In t h i s case, the word pin £a i n (2) can be c i t e d . Zhong-shan has an aspirated / p ' / for the word, whereas i t i s un- 35 aspirated i n Cantonese: [pen ]. In addition to the regular r e f l e x of /p'// there i s also a small handful of words which has /p/ as the correspondent to the Pang i n i t i a l i n Zhong-shan (e.g., bo iJL (3-10) 'glass' /po 5 5/. Hashimoto (p.630) c i t e s the word pou (82-18) 'to s p l i t ' as containing an [f] i n i t i a l i n modern Cantonese f o r which she could give no explanation. In Zhong-shan, the same word simply 13 contains the regular /p'/ i n i t i a l : /p'aw /. Zhong-shan reflexes of the Bing jji (*pft-< EMC *b-) i n i t i a l f a l l into the following pattern according to h i s t o r i c a l - 177 - tonal categories: Ping-sheng words have the aspirated /p'/ i n i t i a l , with no d i f f e r e n t i a t i o n of c o l l o q u i a l or l i t e r a r y pronunciation. Qu- and Ru-sheng words, i n both the c o l l o q u i a l and l i t e r a r y layers, contain the p l a i n /p/ i n i t i a l . Shang-sheng words, however, sub-^ divide according to whether they are c o l l o q u i a l (C.) or l i t e r a r y (L.) forms--the former i s aspirated while the l a t t e r i s not. ;- Examples from the various tones are presented i n (3). (3) a. Ping-sheng: 3-11 /p'o / 'an old woman' b. Qu-sheng: 35-3 | ^ / p a : j 2 2 / 'to destroy' c. Ru-sheng: 194-8 & /pa:k 2/ 'white' d. Shang-sheng / p ' i ] 22 49-5 ^ ' i 1 3 / C . 'a q u i l t ' 205-18 $u /ping /L. 'also' There are several points which may be brought up at th i s time. F i r s t of a l l , i n Zhong-shan a word with a muddy i n i t i a l , such as the Bing i n i t i a l , and containing Middle Chinese Ping-sheng has Yang-ping /51/ as the regular correspondent. In more general terms, the Yang re g i s t e r i s a r e f l e x of the muddy i n i t i a l s , while other i n i t i a l s occur with the Yin r e g i s t e r , tone /55/ i n Ping- sheng words i n Zhong-shan. Secondly, with, regard to the h i s t o r i c a l Shang-isheng reflexes, not only do the i n i t i a l s d i f f e r i n the modern form i n Zhong-shan, but the tones as well. The c o l l o q u i a l form has tone /13/, whereas the tone of the l i t e r a r y form has merged with the - 178 - regular Zhong-shan r e f l e x of the h i s t o r i c a l Qu-sheng, namely tone /22/. Lastly, the general pattern of correspondence to words with i n i t i a l Bing i n Zhong-shan and Cantonese i s e s s e n t i a l l y the same. However, whereas Cantonese has [f] as the only other cor- respondent to the Bing i n i t i a l i n a few i s o l a t e d cases, Zhong-shan has / f / as well as /h/ as i r r e g u l a r correspondences, as exemplified i n (4). (4) 7-13 ^ /faw 2 2/ 'a port' 119-13 ^ / f a : n 2 2 / 'a section 1 212-2 j|L /hung 5 1/L. 'a s a i l " (/p'ung51/C.) For the word bu ij£ , which has an alternate pronunciation of fou in Mandarin, Pulleyblank informs the writer that i t i s a c o l l o q u i a l word not found i n the Guang-yun, a rhyme dictionary which i s an enlarged version of the Qie-yun (published i n 1007 A.D. and i s s t i l l extant). Tang dynasty sources use bu ^ for what i s e v i - dently t h i s word, bu 1̂ , which i s described as a d i a l e c t word of Wu ^ or Chu jjj^ A* 2 1 2 1 Peng yt- i n (4) above i s pronounced [fug ] or Ip'ug J i n Cantonese. The form with the /p 1/ i n i t i a l i n both Zhong-shan and Cantonese represents the c o l l o q u i a l reading. Cantonese [fj corresponds to Zhong-shan [h] before high back vowels (.luj and [u:]), which accounts for the laryngeal i n i t i a l i n the Zhong-shan form for peng as opposed to the labiodental i n i t i a l found i n „ ^ 3 Cantonese. The modern Zhong-shan r e f l e x of the Ming 0$ (*m-) i n i t i a l i s /m/. Examples are shown on the next page. - 179 - 34-3 1 /ma:3 / 1 to buy 1 53-5 /mi / 'beautiful 129-7 ik /mm / 'to sleep' 196-8 /ma:k2/ 'wheat' 215-13 2 /muk / 'wood' While the Ming i n i t i a l i s usually pronounced with the b i l a b i a l nasal i n Zhong-shan and Cantonese, i n three cases i n the data the Ming i n i t i a l has /n/ as i t s modern r e f l e x i n the two Yue di a l e c t s . These three exceptions are l i s t e d i n (_6) , showing the Zhong-shan pronunciation only. (6) 47-15 / n i 5 1 / 'to f i l l ' 49-8 pj- / n i 2 2 / 'the ends of a bow" 72-10 Jjgf /na,:w51/ 'an anchor' With regard to mi and mi jpf above, Hashimoto (:p.631) proposes that these two words are possible vestiges of the d i s t i n c t i o n of the so-called chong-niu ^ 0- ("double knot') pairs of i n i t i a l s , or Grade III/IV doublet i n i t i a l s , since the two words belong to the Grade IV counterpart of these doublets for which some p a l a t a l feature has been suggested. In the Zhong-shan data, mo ^§ (194-11) 'a f i e l d path' i s 2 pronounced /pa:k /, with a /p/ i n i t i a l . The expected i n i t i a l i s i 3 /m/, as i n Cantonese. [mek ] i s recorded for Cantonese i n the Zhong-hua Xin Zi-dian ('Chinese new dictionary') (1977:606) , which agrees with Chen's observation. Hashimoto (p.590) records the word as [met ]. B a l l (p.524) records /mak/ for mo f § (which would be /ma:k/ as the modern, corresponding form) for Zhong-shan, con- t r a s t i n g with his recording of /mak/ ( i . e . , [mek]) for Cantonese. - 180 3.1.2.2. Labiodentals (LMC) There are two regular correspondents to the Fei ffi (*f- <EMC *p-) i n i t i a l i n Zhong-shan: /h/ before the high back vowel /u/, and / f / elsewhere, as exemplified i n (7). The /h/ r e f l e x before /u/ affects the Yu T&s , L i u %J and Tong l | f ^ rhyme groups. Recall that p a r t i c u l a r l y i n the environment before the high back vowel [u:J , /h/ i s d i s t i n c t l y a l a b i a l i z e d f r i c a t i v e £ h w ] . Names of the rhyme groups are included i n round brackets i n (7a). (7) a. Before /u/: 23-1 55 /hu"/ 'a sage' (Yu) 89-7 /hu 2 2/ 'wealth' (Liu) 218-4 /hung 5 5/ 'wind' (Tong) 220-4 * /huk 5/ 'happiness' (Tong) Elsewhere: 46-8 / f a j 2 2 / 'to abrogate' 65-10 / f i 5 5 / •to f l y ' 109-8 / f a : t 2 / 'laws' 164-11 / f a n 1 3 / 'powder' 165-7 / f a t 2 / 'not' Note, however, that the L i u rhyme group has both /u/ and /aw/ as modern Zhong-shan reflexes: /h/ occurs before the former, 2 2 as i n fu ^ /hu / above, and / f / before the l a t t e r , as i n fou £ (.87-23) 'not' /faw 1 3/. In addition to i n i t i a l s / f / and /h/ as regular correspon- dents of the Fei i n i t i a l , there are three words, l i s t e d i n (8), which are pronounced with a b i l a b i a l stop as a r e f l e x . A l l three words i n (.8) come from the Yu rhyme group. Of these, two have the - 181 - aspirated /p 1/ i n i t i a l , and one the p l a i n /p/ i n i t i a l . 18) 25-4 25-5 'dried meat to begin' 25-6 an axe Recalling that the Fei i n i t i a l had developed out of the EMC Bang (*p-) i n i t i a l , the apparent exception can actually be analyzed as forms which have preserved the e a r l i e r i n i t i a l . Hashimoto (p.36) observes that a l l the Yue d i a l e c t s have labiodental f r i c a t i v e s corresponding to the LMC labiodentals, with one exception: the . Chen-cun variety of Shun-de has [p'] as the regular r e f l e x . In the case of the Min d i a l e c t s , a b i l a b i a l stop i s i n fact a regular r e f l e x of t h i s same set of LMC i n i t i a l s . • Fu ^ , for instance, 51 has the c o l l o q u i a l reading of [pu ] and the l i t e r a r y counterpart 51 52 of [hu ] i n Xia-men. The same word i s recorded as [pou ] i n f 31 has a c o l l o q u i a l form of {p'uo ] 31 and a l i t e r a r y one of [xu J. (Zi-hui, p. 77..) I t i s reconstructed as *p- i n Proto-Min by Norman (.1969:260) . In Cantonese, [f] (and only [f]) i s the regular correspon- dent to the Fei i n i t i a l . The singular exception i n Hashimoto's 35 data i s fu , which i s pronounced [p'ow ] (p.425). The Fu (*f- < EMC *p'-) i n i t i a l i n Zhong-shan follows the same d i s t r i b u t i o n a l pattern recorded for the F e i i n i t i a l : /h/ before /u/, and / f / elsewhere. There are no exceptions i n the data. Cantonese likewise has [f] corresponding to the Fu i n i t i a l , with no exceptions observed i n Hashimoto's data. also does not deviate from the pattern occurring i n the F e i Correspondence to the Feng f- (*ffi-C EMC *b-) i n i t i a l - 182 - i n i t i a l — i n either Zhong-shan or Cantonese. Nevertheless, there are two exceptions present i n the Zhong-shan data which have an alternation between / f / ~ / h / and /p 1/ corresponding to a l i t e r a r y - c o l l o q u i a l contrast. As already mentioned, the b i l a b i a l stop can be analyzed as traces of an e a r l i e r stage i n the evolution of the i n i t i a l . C9) 85-5 /faw 5 1/L., /p'u 5 1/C. 'to f l o a t ' 87-24 ^ /hu 2 2/ 'a wife', but /p'u 1 3/ i n : jjty J$? / s a n 5 5 p ' u 1 3 / 'daughter-in-law' Cantonese also shows a l i t e r a r y - c o l l o q u i a l d i s t i n c t i o n for the words i n (9) above, r e f l e c t e d i n a I f J ^ t p ' ] contrast (Hashimoto, pp.488,632). Another exception to [f] as the regular correspondent to the Feng i n i t i a l i n Cantonese c i t e d by Hashimoto (pp .>577,632) is the l i t e r a r y word fu #| (180-5) 'to bind' I p o i k 4 ] . The word 2 i s /fok / i n Zhong-shan, with the regular labiodental r e f l e x . Note, however, that B a l l (p.520), on the other hand, records /pok/ for Zhong-shan, and /fok/ for Cantonese. Chao's data agrees with the present Zhong-shan form. The correspondent to the Wei (*v-< EMC *m-) i n i t i a l i n Zhong-shan has merged with that of the Ming i n i t i a l . The r e f l e x of both h i s t o r i c a l i n i t i a l s , i s /m/. Again, r e c a l l that the Wei i n i t i a l had i n f a c t developed out of the Ming i n i t i a l . I t i s therefore to be expected that the Min d i a l e c t s , for example, would show [mj as a regular r e f l e x of the Wei i n i t i a l rather than iv] or [wj, as i s the case i n the northern Chinese d i a l e c t s . Examples of the Zhong-shan/correspondent to the Wei i n i t i a l are shown i n (10). 183 - (10) 66-8 y£j 85~7 it 146-8 178-18 -jy 220-14 8 / -13 / /mi / /maw / /ma:t2/ /mong 5 1/ /muk2/ 'a t a i l ' 'to scheme' 'stockings' 'to forget' 1 eyes' In the Zhong-shan data on W e i - i n i t i a l words, the only exception appears to be the word wan (145-3) 'to draw back', 13 which has a /w/ i n i t i a l : /wa:n /. Hashimoto (p.632) c i t e s the same word as the only exception to [mj as the regular r e f l e x of the Wei i n i t i a l i n Cantonese. The word wan ffijj also has the 24 l a b i a l glide i n i t i a l i n Cantonese and i s pronounced {wAin ] . 3.1.2.3. Dentals (LMC) The Zhong-shan (and Cantonese) correspondent to the Duan (*t-) i n i t i a l i s / t / , with only a small handful of exceptions Examples of the regular correspondence i n Zhong-shan are shown i n (11): 1-1 \ / 4 . 55 -/to / 'many' 37-4 to /*. -55/ 'to bow the head' 42-10 ** / 4 - -22, 'opposite to' 81-1 ... 22/ /tiw / 'to f i s h (with hook 186-4 /t a k 5 / 'to obtain' Exceptions containing an aspirated dental stop are l i s t e d i n (.12) (12) 16-3 gi. / t ' u 1 3 / 'animal stomach' 37-5 ±Jv^-y<t'aj 5 1/ 'a dike' I \ 55 67-16 p# /t\'ow / 'garrulous' - 184 - 69-5 /t'ow 1 3/ 'to pray' 79-8 $|i / t i w 5 5 / "to engrave' (/t'iw 5 5/C.) As i n Cantonese (Hashimoto, p.632), the word du (.16-3) 'animal stomach1 i n Zhong-shan has merged with the word du (16-7) for 'human stomach', which i s written with the same charac- ter . In the word ti. j f t jr (.37-5) , observe that i t has an aspirated i n i t i a l i n conjunction with Yang-ping tone /51/ i n Zhong-shan. Pulleyblank (personal communication) states that many di a l e c t s have readings for t h i s word which imply the muddy Ding Ĵ L (*tfi-) i n i t i a l rather than the Duan i n i t i a l . (See, for example, Zi-hui, p. 56.) The Zhong-shan reading of aspirated / t ' / i n dao i s based on tao P/7 (67-19) 'to desire', which has the Tou j j ^ . (*t'-) i n i t i a l . The source of aspiration for dao ^ (69-5) i s not cl e a r . Diao jQp. i n (12) above has a variant form /t'iw / which i s s t r i c t l y c o l l o q u i a l . The same exceptions noted above for the Duan i n i t i a l are also found i n Cantonese (Hashimoto, pp.632-633). The words duo (137-1) 'to gather up' and duo $ f (.137- 2 2) 'to weigh a thing i n the hand' are both / t s y t / i n modern Zhong- 22 shan, with an alternate reading of /ts0j /. An a f f r i c a t e i n i t i a l i s also found i n modern Cantonese; for example, i t i s recorded i n 4 the Zhong-hua Xin Zi-dian as [ t s y i t ]. Pulleyblank advises that there i s an alternate reading i n the Guang-yun that can be recon- structed as LMC *tryat (< EMC *trwiat). The regular r e f l e x of LMC *tryat, with the Zhi #tt (*tr-) i n i t i a l , i s / t s y t / i n Zhong-shan. 2 \4i Hence, the Zhong-shan form of / t s y t / for duo Jjfc i s derived from an alternate reading and i s not an exception to the Duan i n i t i a l . The /n/ correspondent to the Duan i n i t i a l i n the word - 185 - niao Jlj (80 -6 ) 'a b i r d ' i n Zhong-shan (and Cantonese) i s an excep- ti o n widespread among the Mandarin d i a l e c t s . The anomalous i n i t i a l i n the modern di a l e c t s can be explained i n terms of an avoidance of homonymy with a taboo word. The regular correspondent to the Tou 2 S C*t'-) i n i t i a l i s / t ' / i n Zhong-shan, as i n Cantonese. Zhong-shan examples are given i n (.13) . 1-2 / t ' o 5 5 / 'to drag along 37 -6 • » 55 'a ladder' 6 7 - 1 9 55 /t'ow D D/ "to deserve' 1 0 8 - 3 / t ' i p 2 / 'a card' 2 0 7 - 1 7 / t ' i a : k 2 / 'to kick' There are several exceptions containing the unaspirated dental stop corresponding to the Tou i n i t i a l . These are l i s t e d i n (14) below. (14) 9 4 - 5 gfc / t a : p 2 / 'to tread' 1 3 0 - 1 0 fjfc / t i n 1 3 / 'protruding (belly) ' 204 -22 / t i n g 5 5 / 'a (sand) bank' Hashimoto (p.633) suggests that the l i t e r a r y word ta ( .94-5), 3 [ t A i p ] i n Cantonese, i s due to the analogical reading of the word ta t|" (94-7) 'a stack ( c l a s s i f i e r for paper) ', which belongs to the Ding i n i t i a l . A similar proposal cannot be put forth for Zhong-shan since the i n i t i a l f or ta i s aspirated i n modern 2 Zhong-shan: /t'!a:p /. The Zhong-shan pronunciation of p l a i n / t / i n taffi may be due to Cantonese influence. The lack of aspira- t i o n i n both t i a n and tin g ;"J above i s probably the r e s u l t of - 186 - analogical r e a d i n g s — t i a n to dian ^ (.130-9) 'a statute' ('found i n common words such as: zi-dian 'Jr ' dictionary') , and ting tp ding J (.204-15) 'an i n d i v i d u a l ' or ding (204-16) 'a n a i l ' . The pattern of correspondence of the Ding (*tn- < EMC *d-) i n i t i a l p a r a l l e l s that of the Bing i n i t i a l i n Zhong-shan (.and Cantonese) . Middle Chinese Ping-sheng words have the / t ' / i n i t i a l i n the modern Zhong-shan d i a l e c t , with no l i t e r a r y - c o l l o - quial d i s t i n c t i o n . In contrast, the h i s t o r i c a l Qu- and Ru-sheng words are unaspirated i n the modern form. With regard to the Middle Chinese Shang-sheng words, there i s a d i s t i n c t i o n of l i t e r a r y versus c o l l o q u i a l layer: the i n i t i a l i n the l i t e r a r y '. reading i s unaspirated and the tone i s merged with the Zhong-shan ref l e x of the h i s t o r i c a l Qu-sheng; the i n i t i a l of the c o l l o q u i a l form i s aspirated and the tone i s /13/. An example of each i s given i n (15). Dan (95-18) i n (15d) has a c o l l o q u i a l versus a l i t e r a r y reading i n Zhong-shan. (The same s t y l i s t i c d i s t i n c t i o n for the word i s also found i n Cantonese.) (.15) a. Ping-sheng: 81-17 jJLlJ /t'aw 5 1/ 'head' b. Qu-sheng: 54-12 / t i 2 2 / 'the earth' c. Ru-sheng: 132-10 $L / t i t 2 / 'to f a l l down* d. Shang-sheng: 95-18 /t'a:m /C. ' i n s i p i d ' 95-18 yf^ /ta:m 2 2/L. ' i n s i p i d ' - 187 - Aside from exceptions r e s u l t i n g from the unpredicted pre- sence or absence of aspiration on the i n i t i a l (e.g., the Ru-sheng word ta (94-7) /t'a:p / 'a stack' has i n i t i a l / t ' / instead of / t / ) , there are also a couple of words which show an a f f r i c a t e as a re f l e x of the Ding i n i t i a l : (.16) 215-19 i% /tsuk 2/ 'a c a l f 215-20 /tsuk 2/ 'a ditch' The regular Zhong-shan correspondent to the Ni j/Li (*n-) i n i t i a l i n the d i a l e c t survey i s /n/. Recall that Ni and Niang "jflL. (*nr-) i n i t i a l s are combined under the Ni i n i t i a l i n the survey l i s t with no in d i c a t i o n of which word i n fact belongs to the Ni and which to the Niang i n i t i a l i n the rhyme tables which d i s t i n - guish t h i r t y - s i x i n i t i a l s . The merging of the two Middle Chinese i n i t i a l s does not present a problem i n analyzing the Zhong-shan data since the modern r e f l e x of both h i s t o r i c a l i n i t i a l s i s /n/, as shown i n (17). (.17) a. Ni i n i t i a l : 14-16 ^ /nu 5 1/ 'a slave' 2 94-8 /na:p / 'to give' 184-13 |fo /nang 5 1/ 'can, may' b. Niang i n i t i a l : 20-6 /ny 1 3/ 'a woman' 88-1 /naw 1 3/ 'a knot' 104-10 / n i p 2 / 'to tread' (also / s i p 2 / ) A l l the exceptions to the regular correspondence arise from words with the Niang i n i t i a l . In the case of the word - 188 - . 55 nian jUj (.101-9) 'sticky' /nim /, although i t does have /n/ as the regular r e f l e x of the Niang i n i t i a l , when i t enters into the combination nian-mi ^ jjp- 'glutinous r i c e ' , the term i s / t s i m 5 5 13 maj / i n Zhong-shan, with / t s / as the correspondent to the Niang i n i t i a l . (The same alternation of i n i t i a l s occurs i n Cantonese v i s - a - v i s t h i s word.) Another N i a n g - i n i t i a l word which has / t s / as the r e f l e x i n Zhong-shan i s nian ,&JL. (124-9, 125-8) '(.a stone r o l l e r for husking 13 grain)'. The a f f r i c a t e i n i t i a l for the word, which i s / t s i n / i n Zhong-shan, i s probably due to the analogical reading of the character zhan $^ (124-16) 'to open 1, which occurs i n a number of common p o l y s y l l a b i c words i n the language. The word nie (.104-10) has two variant forms i n Zhong- 2 2 IL shan: /nip / and / s i p /, while nie jpi. (104-8) ' (a surname) ' has 2 only the form of / s i p /. In discussing a si m i l a r phenomenon of I s ] corresponding to the Niang i n i t i a l i n Cantonese, Hashimoto (p.637) suggests that the reading of a s i b i l a n t i n i t i a l may be based on the analogical reading of the more common word she 4 2 (.105-3) (phonetically [si:p ] i n Cantonese and I s i i p ] i n Zhong- shan), which contains the EMC Sh^n ^ (.*<?- < LMC * s - ) i n i t i a l . It may be noted that B a l l (p.525) records nie ̂  as /nip/ for both Zhong-shan and Cantonese. The word nong >J[ (222-8) 'thick (of l i q u i d s ) , dark (.of 51 t i n t s ) ' i s /nung /, except i n reference to strong teas, soup and other l i q u i d s , i n which case the word i s c o l l o q u i a l l y pronounced 51 • • /jung /. (A sim i l a r alternation of nasal and p a l a t a l i n i t i a l exists i n Cantonese with regard to t h i s word.) The three words i n (.18) on the next page exhibit only a - 189 - p a l a t a l / j / corresponding to the Niang i n i t i a l . The f i r s t and t h i r d word i n 0-8) likewise have a p a l a t a l glide i n Hashimoto's data for Cantonese. >-ll $ " 2 2 (18) 112-11 g /jam / 'to rent' 170-19 jM / j o n g 2 2 / 'troubled' 174-17 ^ / j o n g 2 2 / 'to brew' 1* v ti In the case of ren (or -iin) & • Peking l i n for |fj i s also i r r e g u l a r . The alternate pronunciation of re*n for the word i n Mandarin found i n some sources (e.g., Mathews' Chinese-English Dictionary) would correspond to the p a l a t a l i n i t i a l i n the Zhong- shan and Cantonese forms for ren ^ . The / j / r e f l e x may be based 2 2 on the analogical reading of ren ft (.113-2) 'duty' /jam /. An alternate reading for niang J$j__ i n (.18) above can be found under i n i t i a l Ri 0 (EMC *i>-> LMC *r-) . Since / j / i s one of the regular reflexes of the Ri i n i t i a l i n Zhong-shan, the read- ing of niang with the p a l a t a l i n i t i a l i n Zhong-shan may be based on the R i - i n i t i a l reading. I t i s possible, however, that the p a l a t a l glide i n i s o l a t e d occurrences of N i a n g - i n i t i a l words i s a vestige of the o r i g i n a l d i s t i n c t i o n between the Ni (*n-) and Niang C*nr-) i n i t i a l s , as Pulleyblank suspects. Probably s i g n i f i - cant i s the observation that a l l exceptions i n the Zhong-shan data pertaining to the two i n i t i a l s a r i s e with respect to the Niang i n i t i a l , with the cases of the p a l a t a l glide as a r e f l e x of the Niang i n i t i a l being p a r t i c u l a r l y noteworthy. Pulleyblank (1970:214-216) provides evidence from the Tibetan hP'ags-pa s p e l l i n g to propose that *rir- became *n j - ( i . e . , * i > - ) i n front of - i , and that the d i s t i n c t i o n between * n i - and *i>i- - 190 - survived u n t i l early Yuan (.1279-1368) even i n Mandarin. Given the case of the Ri (EMC * i > - ) i n i t i a l , i n which * i > - yielded / j / as one of i t s correspondents i n Zhong-shan, one would expect / j / to be a regular r e f l e x of the pa l a t a l nasal * i > - a r i s i n g from the Niang i n i t i a l . Thus, i t may be the case that some words with i n i t i a l / j / descending from the Niang i n i t i a l are traces of the o r i g i n a l d i s t i n c t i o n between the Ni and Niang i n i t i a l . The L a i C * l - ) i n i t i a l has / l / as the regular r e f l e x i n Zhong-shan. Examples are given i n (.19) . (19) (-20) 3-17 / i t 5 1 / 1 a mule' 18-25 / l u 5 1 / 'a donkey' 69-15 /low / •old' 129-15 * / l i n 5 1 / 'to p i t y ' 189-6 il / l i k 2 / 'strength' There , are five exceptions i n the data: 3-19 p /wo55/ 'lines on the finger' 39-10 22 /tai / 'to belong to' 113-13 /nap 5/ 'a grain; ( c l a s s i f i e r ) 173-12 / n i o n g 1 3 / 'two' 201-7 4k / • 13 / n /nxa:ng /C. 'co l l a r ' (/ling 1 3/L.) The pronunciation of the word luo jffi with a /w/ i n i t i a l i n Zhong- shan may be based on the analogical reading of the common word wo (4-17) 'a nest', which i s a He-kou word, reconstructed i n LMC as *?wa, with the Ying f/p (*?-)' i n i t i a l . The writer has;no explanation to o f f e r for i n i t i a l / t / i n the word l i _ aside from observing that jJL does occur as a phonetic i n a few words - 191 - h i s t o r i c a l l y containing a dental stop i n i t i a l . Di \(a moun- ta i n tree l i k e the cherry).', for example, i s h i s t o r i c a l l y homopho- nous with d i % (38-6) 'younger brother' / t a j 2 2 / . One could, at th i s point, make a couple of suggestions, including possible ana- l o g i c a l readings. The writer w i l l , nevertheless, leave further speculations for the present. L i ||L i s also recorded by Chao as 22 /ta j /. Cantonese likewise demonstrates an xrregular dental 33 stop i n i t i a l for the word (..[tuj J ) . The l a s t three exceptions i n (20), with i n i t i a l /n/, have already been discussed i n Chapter 1.1 i n which d i a l e c t a l borrow- ings was postulated for the rare occurrence of /n/ as a r e f l e x of the Lai i n i t i a l i n Zhong-shan. 3.1.2.4. Dental S i b i l a n t s (LMC) Zhong-shan, l i k e Cantonese, has / t s / as the correspondent to the Jing ^ (*ts-) i n i t i a l . Examples are: (21) 14-23 / t s u 5 5 / 'to rent' 38-8 ffi / t s a j 5 5 / 'to crowd' 124-11 H / t s i n 1 3 / 'to cut with scissors' 175-2 / t s i o n g 2 2 / 'a general' 207-22 £1 / t s i k 5 / 'to spin' There are four words i n the data, l i s t e d i n (22), which have / t s ' / corresponding to the Jing i n i t i a l . (22) 6-5 /ts'o / 'to push down' 69-19 /ts'ow 2 2/ 'to bathe' * 22 71-3 ĵ jL /ts'ow / 'easily provoked' 101-15 jfjfl! / t s