INFORMATION TO USERS The most advanced technology has been used to photo­ graph and reproduce this manuscript from the microfilm master. UMl films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are re­ produced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. These are also available as one exposure on a standard 35mm slide or as a 17" x 23" black and white photographic print for an additional charge. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

UMI University Microfilms International A Bell & Howell Information Com pany 300 Nortfi Z eeb Road, Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600

Order Number 9011202

Middle Mandarin phonology: A study based on Korean data

Kim, Yoimgman, Ph.D. The Ohio State University, 1989

U'M'I 300 N. Zeeb Rd. Ann Arbor, MI 48106

MIDDLE MANDARIN PHONOLOGY:

A Study based on Korean data

DISSERTATION

Presented in Partial Fulfillment of the Requirements for

the degree Doctor of Philosophy in the Graduate

School of the Ohio State University

by

Youngman Kim, B.A., M.A.

*****

The Ohio State University

1989

Dissertation Committee: Approved by

F .S . Hsueh

T.Y. Li

Brian Joseph Dept, of East Langs. & Lita. To My Mother

11 ACKNOWLEDGMENTS

I owe the first and greatest debt of gratitude to

Professor F. S. Hsueh for his guidance and encouragement throughout this dissertation research. I am also grateful to the other members of my dissertation committee.

Professors Brian Joseph and T. Y. Li, for their suggestions and comments. I wish also to thank Professor Timothy Light for helping me initiate this research.

Further acknowledgment is due to the professors, secretaries, and fellow students of the Department of East

Asian Languages and literatures for their help and moral support, and especially to Dana Bourgerie for so many things. A special debt of gratitude is extended to

Professor S. H. Kang, who graciously sent me valuable source materials and his publications. Most of all, I wish to thank my family for their love, support, and prayers during my graduate studies.

Ill VITA

January 26, 1951 ...... Born - Seoul, Korea

1977 ...... B.A., Korea University, Seoul, Korea

1980 ...... M.A., Dept, of East Asian Languages and Literatures, The Ohio State University, Columbus, Ohio

1981-1986 ...... Teaching Associate, Dept, of East Asian Languages and Literatures, The Ohio State University,- Columbus,- Ohio

1986 - 1988 ...... Instructor, East Asian Studies Program,Denison University, Granville, Ohio

FIELDS OF STUDY

Major Field: Chinese Linguistics

Studies in Chinese Phonology, Dr. F.S. Hsueh

Chinese Syntax, Dr. Timothy Light

Chinese Literature, Dr. T.Y. Li

General Linguistics, Dr. Brian Joseph

IV TABLE OF CONTENTS

DEDICATION . . 11

ACKNOWLEDGMENTS iii

VITA iv

LIST OF TABLES viii

CHAPTER

I. INTRODUCTION ...... 1

1.0 Introduction ...... 1

1.1 The relation between Mandarin and Pekingese 2

1.2 The background of Chinese historical phonology ...... 5

1.3 Sources on Chinese historical phonology 7

1.3.1 Phonetic series ...... 8 1.3.2 The rhyme book tradition 9 1.3.3 The rhyme table tradition 12 1.3.4 Transcriptional evidence 17

1.4 The task of the present study .... 19

Notes ...... 24

II. THE KOREAN SOURCES OF MANDARIN CHINESE . . . 28

2.1 The invention of the Korean alphabet 28

2.2 Transcription of Mandarin pronunciation 32

2.2.1 The transcriptional sources . . 32 2.2.2 The three types of pronunciation notations ...... 34 2.3 NoQoltae and Pakt'onosa ...... 38

2.3.1 Introduction ...... 38 2.3.2 A brief review of previous studies . 42

N o t e s ...... 50

III. THE PHONOLOGY OF MODERN P E K I N G E S E ...... 54

3.0 Introduction ...... 54

3.1 The syllable in Chinese ...... 55

3.2 The initial consonants of Modern Pekingese 58

3.3 The finals of Pekingese ...... 63

3.3.1 The m e d i a l s ...... 63 3.3.2 The e n d i n g s ...... 68 3.3.3 The vowels ...... 70

3.4 R-suf fixation ...... 76

3.5 The tones of Modern Pekingese ...... 78

N o t e s ...... 83

IV. THE INITIAL SYSTEM OF MIDDLE MANDA R I N ...... 85

4.1 The letters of Hunmin c h c n a ' u m ...... 85

4.2 The method of pronunciation notation . . . 86

4.3 The initial consonants of Middle Mandarin 96

4.3.1 Former voiced obstruent initials . 96 4.3.2 Labials ...... 100 4.3.3 Alveolar and retroflex sibilants . 103 4.3.4 The retroflex initial consonant /r/ 115 4.3.5 Gutturals ...... 117 4.3.6 Summary ...... 120

Notes ...... 123

V. THE FINAL SYSTEM OF MIDDLE MANDARIN ...... 126

5.0 Introduction ...... 126

5.1 The rhyme-sets Zhen. Han, Shan, and Xian 138

vi 5.1.1 The Zhen rhyme-set ...... 139 5.1.2 The Shan rhyme-set ...... 146 5.1.3 The Xian rhyme-set ...... 150 5.1.4 The Han r h y m e - s e t ...... 154'

5.2 The rhyme-sets Dona. Gena, and Yang . . . 160

5.2.1 The Dona rhyme-set ...... 161 5.2.2 The Yana rhyme-set ...... 165 5.2.3 The Gena rhyme-set ...... 173

5.3 The rhyme-sets Pin. Tan, and Y a n ...... 183

5.3.1 The Pin rh y m e - s e t ...... 184 5.3.2 The Tan and Yan rhyme-sets .... 188

5.4 The rhyme-sets Xiao, vao. and You .... 193

5.5 The rhyme-sets Ge, Ma, and Z h e ...... 202

5.6 The rhyme-sets Zhi, Pi. Jie. and Hui . . . 210

5.6.1 The Jie rh y m e - s e t ...... 210 5.6.2 The Hui rh y m e - s e t ...... 214 5.6.3 The Zhi and ûi r h y m e - s e t s ...... 216

5.7 The rhyme-sets Yu and M o ...... 222

5.8 Summary ...... 226

N o t e s ...... 231

VI. CDNCLUSIPN ...... 235

BIBLIPGRAPHY ...... 246

Vll LIST OF TABLES

TABLE PAGE

1. The Structure of the Chinese syllable .... 56

2. The initial consonants of Modern Pekingese 60

3. The phonemic representations of the MP initials 62

4. The syllbic medials of Modern Pekingese . . . 67

5. The syllabic endings of Modern Pekingese . . 69

6. The phonetic forms of the MP vowels ...... 70

7. The phonemic representations of the MP finals 75

8. The MP finals in a phonetic transcription . . 75

9. The four tones of MP ...... 79

10. The 17 basic consonants of Hunmin chona'um 86

11. The consonant system of Middle Korean .... 87

12. The vowel system of Middle Korean ...... 88

13. The initial consonants of HWZY ...... 93

14. The Korean vowel letters used in HMCUYH . . . 94

15. The Korean endings used in HMCUYH ...... 94

16. The initial consonants of MM ...... 96

17. Transcription of the former voiced initials 99

18. The labial initials in PN and PP ...... 101

19. Transcription of the initial Wei ...... 102

20. Transcription of the dentalveolar sibilants 111

viii 21. Transcription of the Dy initials ...... Ill

22. Transcription of the Py initials ...... 112

23. Transcription of the Pr[l] initials ...... 112

24. The emergence of the zero i n i t i a l ...... 119

25. The phonemic representations of the MM initials 122

26. The vowel phonemes of OM ...... 127

27. The final types of the sixteen rhyme classes . 130

28. The rhymes of ZYYY ...... 131

29. The finals of ZYYY ...... 132

30. The rhyme-sets of HWZY ...... 133

31. The finals of HWZY ...... 134

32. The finals of DYTJ ...... 135

33. Transcription of the Zhen rhyme-set words . . . 140

34. Transcription of the Shan rhyme-set words . . . 147

35. Transcription of the Xian rhyme-set words . . . 152

36. Transcription of the Han rhyme-set words . . . 157

57. Transcription of the Dona rhyme-set words . . . 162

38. Transcription of the Yana rhyme-set words . . . 167

39. Transcription of the Gena rhyme-set words . . . 177

40. Transcription of the Pin rhyme-set words . . . 184

41. Transcription of the words the Tan and Yan rhyme-sets ...... 189

42. Transcription of the words of the Xiao. Yap, and You rhyme-sets ...... 196

43. Transcription of thewords of the Ge, Ma. •and Zhe rhyme-sets ...... 204

44. Transcription of the Jie rhyme-set words . . . 211

IX 45. Transcription of the Hui rhyme-set words . . . 215

46. Transcription of the words of the Zhi and Qi r h y m e - s e t s ...... 218

47. Transcription of the words of the Yu and Mo r h y m e - s e t s ...... 224

48. The finals of MM ...... 228

49. The phonemic representations of the MM finals . 230 CHAPTER I

INTRODUCTION

1.0 Introduction

The Noaoltae ( ^ % "K Laoqida 'Good Old Chinese') and the Pakt'onosa ( iâ, ^ Putongshi 'Interpreter Pak') were popular textbooks of Mandarin Chinese of unknown authorship dating from the fourteenth century in Korea (Zhu

1958 and Min 1964). Ch'oe Se-jin ( tt ) f a prominent

Korean philologist of Chinese in the early part of the sixteenth century, added Korean transcription and translation to the Chinese texts of the two books about the year of 1515 A.D. These revisions are known as the Ponvok

Noaoltae ( isHt ^ Fanyi Laoqida 'Translation of the

Noaoltae ' ) and the Ponvok Pakt'onasa ( ^ ^ -ÿ- Fanyi

Putongshi 'Translation of the Pakt'onasa'). respectively

(Nam 1972).

The history of Mandarin can be divided into three stages: Old Mandarin, Middle Mandarin, and New Mandarin, or

Modezm Pekingese (Hsueh 1975:13). The aim of the present study is to reconstruct the sound system of Middle Mandarin of the early sixteenth century based on the Ponvok Noaoltae

1 2

and the Ponvok Pakt'onasa and their later revisions, and to

reassess the historical development of Mandarin Chinese.

That is, first, I shall examine the Korean transcription in

the two books and their later revisions in order to

interpret the sound system of Middle Mandarin as represented

by the Korean transcription. Secondly, I shall endeavor to

explicate the sound change from Old Mandarin to Middle

Mandarin and from that stage to Modern Pekingese.

However, some introductory remarks of a general nature are necessary as a prelude to the main discussion

about the phonological history of Mandarin. In this

chapter, therefore, I shall start with a discussion of the relation between Mandarin and Pekingese, and then review briefly the background of Chinese phonological study. In addition, I shall outline the overall plan of this study.

1.1 The relation between Mandarin and Pekingese

The Chinese language ( ’Ÿ Ê Zhongguohua) is also known as Hanyu ( the 'Han language'), for it is spoken by the Han nationality, which makes up ninety-four percent of China's population at present. The different varieties of the Chinese language are conventionally referred to as

"dialects". The Chinese language is usually classified into eight major dialect groups based primarily on a comparison of their sound structure: Northern dialect (Mandarin), Wu dialect, Xiang dialect, Gan dialect, Kejia dialect (Hakka), 3

Yue dialect (Cantonese), Northern Min dialect. Southern Min dialect.^ These "dialects" in their spoken form may differ from one another to the point of being mutually unintelligible. For example, Mandarin and Cantonese differ from each other as much as two separate Romance languages do.^ On the other hand, on the grounds that the Chinese

"dialects" are spoken in the same country and that they share the same writing system (i.e., Chinese characters), they are commonly called "dialects" instead of "languages".

Mandarin, now spoken by more than seventy percent of the Chinese people, can be divided into four major dialects.

They are Northern Mandarin, Northwestern Mandarin,

Southwestern Mandarin, and Lower Yangzi Mandarin; and their representative localities are Peking, Taiyuan, Chengdu, and

Nanjing, respectively (Yuan 1960:24 and Egerod 1967:100-

107). However, Mandarin, in its narrow sense, has also been used conventionally as a term for Northern Mandarin. This conventional usage shall also be kept hereafter in this study, unless it is necessary at certain points to distinguish "Mandarin" as a major dialect group from

"Northern Mandarin" as a major dialect of Mandarin, for this study shall treat mainly the history of Northern Mandarin phonology.

Peking has been the capital of China since the Jin dynasty (1115-1234 A.D.) moved the capital to Peking in 1153

A.D., then called Yanjing ( ^ ) (Wang Li 1980:37). 4

Peking has also been the cultural center of China since the beginning of the Yuan dynasty (1279-1368 A.D). The dialect of the capital of the Yuan dynasty, i.e., Peking, then called Dadu ( ), is referred to as the early stage of

Mandarin. During the Ming period (1368-1644 A.D.) Nanjing was the capital initially, but the capital was moved back to

Peking only fifty years later. The stage of Northern

Mandarin about the late Ming period is considered to be

Middle Mandarin. During the period of the Qing dynasty

(1644-1911 A.D.) Mandarin was generally referred to as

Guanhua ( the 'official language'), and the norm for the pronunciation of Mandarin was based on the court speech of the capital Peking (Pulleyblank 1984:2).

After the imperial era of China, the Nationalist government designated Northern Mandarin, i.e., Pekingese, as the National language of China, and called it Guoyu ( ^ |§- the 'national language'). In 1955 the government of the

People's Republic of China also proclaimed Pekingese to be the standard Chinese language, but called it Putonghua

( ^ )h. t'h the ' common language ' ). Putonghua embodies the pronunciation of the general Peking dialect, the grammar of

Northern Mandarin, and the vocabulary of modern Chinese vernacular literature.^ This dialect, though called by different names, is now taught in schools in both Mainland

China and Taiwan. Since both Putonghua and Guoyu are based on the Peking dialect, they are identical except for some 5 differences in certain areas of pronunciation and vocabulary. Therefore, the term "standard Chinese language" generally refers to Putonghua, Guoyu, or Northern Mandarin as a whole.

In short. Mandarin has been the most prestigious dialect of the Chinese language for seven centuries.

Pekingese, or the Peking dialect, has set the norm for the pronunciation of Mandarin throughout its history. Thus,

Pekingese is appropriately considered the standard form of

Mandarin as far as pronunciation is concerned. Besides, the bulk of studies of Mandarin phonology has centered on

Pekingese due to its abundant philological and phonological records as well as its status as the official dialect.

Consequently, what we shall attempt in this study of the history of Mandarin phonology is a consideration of the diachronic phonological development of Pekingese.

1.2 The background of Chinese historical phonology

Native study of the Chinese language started early in

China, and developed into an important branch of traditional scholarship. It also handed down a large number of written records, which provide many kinds of useful information on the history of the Chinese language. Traditional literature dealing explicitly with the language can be conveniently divided into two broad groups: 6 (1) Lexicographical works, which include lexicons and

glosses on the classics;*

(2) Phonological works, which include rhyme books and

rhyme tables.

Historical phonology has been the mainstream of

Chinese historical linguistics. Our present knowledge of

Chinese historical phonology is based primarily on the traditional phonological works. In addition, there is much written material from which phonological information can be inferred, viz., (1) phonetic series, (2) poetic rhymes, and

(3) transcriptional evidence (Baxter 1978:1-3).

Until the early centuries A.D. the pronunciation of a character was indicated by a homophonous character. When no homophone existed, it was necessary to rely on near­ homophones. In this kind of situation, the invention of the fangie ( tj] ) spelling was a major breakthrough in the analysis of a syllable. A fanqie spelling indicates the pronunciation of a character by using two other, presumably known, characters. The first, or upper character represents the initial consonant; and the second, or lower character represents the final and its tone. For example, the word ^

(Middle Chinese [tung]) is spelled ^ , i.e., [t(ak) +

(h)ung]. The invention of the fanaie method was probably influenced by Sanskrit phonology as early as about the second century (cf. Hsueh 1975:13 and Ramsey 1987:121). 7

This method later came into common use, and facilitated the

discussion of Chinese syllabic structure.

The rhyme book Oievun (601 A.D.) by Lu Fa-yan

( fè Q ) is generally regarded as the starting-point of

reference for all subsequent phonological studies in the

native tradition. The native tradition of phonological

studies continued to develop largely in the form of rhyme

books (or rhyme dictionaries) and rhyme tables, and

eventually created the school of 'older pronunciation'

during the Qing dynasty.^ The scholars of the 'older

pronunciation' school took the Oievun system as the

'contemporary pronunciation' ( 'f' ^ jin-yin), and worked at

the reconstruction of the 'older pronunciation' ( ÿ ^ gu-

yin), i.e.. Archaic Chinese (ca. early first millennium

B.C.) phonology (Hashimoto 1965:2).

In the present century, a modern linguistic approach

has also been applied to the field of Chinese historical

phonology since the eminent Swedish Sinologist, Bernhard

Karlgren first (about 1920) applied the comparative method

of Western historical linguistics to the reconstruction of

Ancient Chinese sound system as represented by the Oievun.

1.3 Sources on Chinese historical phonology

In order to lay the groundwork for our discussions in

the remainder of this study, it may be helpful at this point to discuss the following aspects of the history of Chinese phonology: (1) phonetic series, (2) the rhyme book tradition, (3) the rhyme table tradition, and (4) transcriptional evidence.

1.3.1 Phonetic series

The unit of the Chinese writing system is the character, or ^ ( '^ ). A character, or zL, is composed of three elements: the graph ( ^ xing 'shape'), the pronunciation ( ^ yin 'sound'), and the meaning ( ^ yi

'definition'). A character, or a graph represents a basic meaning, and is pronounced as a syllable which is identical with the spoken form for the basic meaning.

At the outset of their development, Chinese characters were created for objects or ideas represented by specific speech sounds.* However, as Chinese characters could not be created on an ideographic basis alone,

'phonetic compounds' began to be created already in their early stage of development, and continued to increase. Most of the Chinese characters belong to the category of phonetic compounds. ’’

A phonetic compound consists of two components: a

'radical' (also called 'signifie' or 'classifier') and a

'phonetic'. For example, the character (ran 'to burn) was also used to mean 'thus, or so'. Later, a radical H

(huo 'fire') was added to the original character, or a phonetic, ( ran ), thus creating a new character, ( ran 9

'to burn') in order to differentiate (ran 'to burn')

from ( ran ' thus ' ). The new character ( ran ' to

burn') consists of a radical, (huo 'fire') and a

phonetic, (ran).

A set of characters which share the same phonetic is

called a phonetic series. Karlgren arranged phonetic

compounds according to phonetic series in the Grammata

Serica (1940). Phonetic series are one of the two main

sources of information about Archaic Chinese side by side

with the poetic rhymes in the Book of Odes ( ff’ Shi-iina

ca. 6th cent. B.C.). However, "because of sound changes

since Archaic Chinese, the characters in a phonetic series

often show little resemblance in [their] modern

pronunciation" (Baxter 1977:2).

1.3.2 The rhyme book tradition

A rhyme book ( ^ ^ yunshu) is a rhyming dictionary

in which characters are arranged by tones and finals into a

large number of rhyme groups. The rhyme book Oievun was

compiled by Lu Fa-yan in 601 A.D., but the pronunciation

standard for the rhyme book was set by Lu himself and his

eight co-compilers some twenty years earlier. The rhyme

book became the most authoritative and fruitful source for

the study of Ancient Chinese phonology. The Oievun, its

subsequent revisions, and other rhyme dictionaries modelled after it formed the rhyme book tradition. 10

The organization of the Oievun may be summarized as follows:*

1. Tone groups. About twelve thousand characters are classified into four main parts, one for each of the four tones of Ancient Chinese: pina-shena ( jp- ^ 'even tone'), shana-shena ( _h. ^ 'rising tone'), au-shena ('going tone'), and ru-shena ( \ ^ 'entering tone'). Section one and section two of pina-shena respectively have 26 rhymes, and 28 rhymes.® Shana-shena. au-shena. and ru-shena. have

51 rhymes, 56 rhymes, and 32 rhymes, respectively. Thus, the book comprises 193 rhymes in total.

2. Rhymes. Within each tone group characters are subdivided into rhyme groups, which are conveniently identified by their first character. For example, the Dona

^ rhyme is the rhyme group of which the first character is dona ^ . All the characters in the same rhyme group rhyme with each other. But a single rhyme may include characters with different finals, for the final includes not only the rhyme base of the syllable, but also the medial before the main vowel.

3. Homophone groups. Within each rhyme group, characters which are completely homophonous are grouped together in homophone groups. Under the first character in each homophone group, the pronunciation is indicated by a fanqie spelling consisting of two characters: the former 11 represents the final, and the latter represents the final and its tone.

The original of the Oievun is not extant. A number of revisions of this rhyme book were made during the Tang-

Song period, e.g., Kanmiu buoue Oievun ( -hj ^ f^ 5;^ ^/7 M ),

Da Song chonaxiu Guanovun ( iZ ^ Hf' ^ ^ , or Guancrvun ),

Jivun ( g I ), etc. Among these revisions the Guanovun

(1008 A.D.) preserves the phonological categories of the original Oievun almost entirely intact, and hence has been the chief source of information. The Guanovun has 206 rhymes, while the original Oievun has only 193; but this probably has no phonological significance."

Karlgren reconstructed the sound system underlying the Oievun. He applied the comparative method to the results of Chinese phonological tradition. That is, he first listed the initial and final categories as set up in the rhyme tables (see Section 1.2.3), and then compared them with modern dialectal and Sino-Xenic (namely, Sino-Japanese,

Sino-Korean, and Sino-Vietnamese) pronunciations. Thus, he set up a system of 47 initials since he distinguished palatalized (or 'yodized') labials, gutturals, and liquids from their corresponding non-palatalized initials on account of different fanaie initials for these corresponding series of initials. He also set up a complicated system of 15 vowels. Karlgren's analysis was done on a phonetic basis.

In the recent decades Karlgren's reconstruction has been 12

revised in the light of phonemic theory. Chao Y. R.

(1940) pointed out that the palatalized and non-palatalized

series of initials were in complementary distribution, being

conditioned by the presence or absence of the medial i in

Karlgren's system. It is now generally agreed that the

initials of the Oievun consist of 36 consonant phonemes,

including the "zero" initial. The 15 vowels have also been phonemicized, but a consensus on the vowel system within a phonemic framework is yet to be reached (Hsueh 1985:50, note

9). Karlgren called the stage of the Chinese language represented by the Oievun Ancient Chinese, and assumed

Ancient Chinese to be the spoken language of Ch'angan

( ^ ), the capital of the Sui and the Tang dynasties.

However, it is now generally accepted that the Oievun is a codification of literary pronunciations of different areas arranged by the compilers. That is, the Oievun does not represent a stage of any one spoken dialect, but the national literary standard as recorded by Lu Fa-yan, et al." Fulleyblank (1984:3) calls this literary standard pronunciation Early Middle Chinese (EMC).

1.3.3 The rhyme table tradition

Actual pronunciations in the North during the Tang period (618-907 A.D.) must have differed considerably from the Oievun literary standard, which incorporated 13 pronunciations from other areas of the country (Bodman

1967:15). The language changed in time, and people must

have begun to puzzle over the sound values of fanqie

spellings and the differences between some Oievun

pronunciations and the contemporary pronunciations. Under

the circumstances, a new phonological study, influenced by

Sanskrit phonology, started with classifying the initial

consonants into the representative characters called zimu

( , also called shengmu 'mother of initials',

hereafter initial characters). After having added the ways

of analyzing and classifying the finals ( ^ yunmu

'mother of finals'), this study developed into a

full-fledged phonological tradition called denovun-xue

( ^ ^ 'the study of graded rhymes', or 'the study of

rhyme tables')./*

A rhyme table ( ^ ^ dengyuntu) is a phonological

chart, which is marked off by intersecting horizontal and

vertical lines into a grid. Each syllable of the Oievun

tradition is placed in a square of the grid, so that the

pronunciation of each syllable can be ascertained by means

of its abscissa and ordinate locations. Dengyuntu, usually

translated as 'rhyme table', may mean either (1) a

phonological chart in the form of a grid, or (2) a book

which consists of a number of such phonological charts. It may also mean (1) and (2) as a whole. Hereafter, I shall

refer to (2) as a rhyme table book when there is a need to 14 differentiate (2) from (1).

Although the organization of one rhyme table book may differ from that of another slightly, the general structure of rhyme table books can be summarized as follows:

1. The thirty-six initial characters are lined up horizontally on top of each table, and lead twenty-three vertical rows since on top of some rows two initial characters are placed." The twenty-three rows are classified into five (or seven) columns, which correspond to places of articulation to some extent. Within each of these columns, the initial consonants are arranged into four phonation types. The difference between stop versus fricative is treated as subcategories of places of articulation."

2. Each table is divided into four horizontal sections corresponding to the four traditional tonal categories. Each tonal section is further subdivided into four 'divisions' ( ^ deng, also called 'grades'), which are referred to as division 1, division 2, division 3, and division 4, respectively. Thus, each of the sixteen horizontal rows is assigned for one Oievun rhyme. The four divisions together apparently had genuine phonological distinctions of some kind. However, their phonetic qualities have been a much debated problem. It is generally assumed that, for the stage of the language represented by 15 the rhyme tables, divisions 3 and 4 had a high front vowel, while divisions 1 and 2 did not."^

3. A number of rhymes form a 'rhyme class' ( ^ s h e ).

Some of the rhyme classes have no words in division 2 except under the sibilant initials, and fill only division 1 and/or division 3; while the others fill the second division, and optionally fill the other divisions. The former classes are marked as the 'inner series' ( nei-zhuan), and the latter as the 'outer series' ( wai-zhuan).^ Within each rhyme class, the contrast between the 'rounded' ( ^ iJ he-kou 'closed -mouth') and the 'unrounded' ( ^ 0 kai-kou

'open-mouth'), if appropriate, is set up depending on the presence or absence of the semi-vowel /w/ in the medial position.

The Yunling ( ) is the oldest extant rhyme table book. The evolution of the rhyme tables comprises three phases, which are represented in order by (1) the

Yunling and the Oivinlue ( -t ^ ), (2) the Sisheng dengzi ( IS ^ ^ ^ hereafter SSDZ) and the Oievun zhizhangtu ( t# hereafter QYZZT), and (3) the Jingshi zhengvin Oievun zhinan ( ^ '<57 a| ).

The Yunling seems to reflect the period when only thirty zimu. or initial characters were established since the labial fricatives ( ^ ^ qingchun-yin) are not distinguished from their corresponding stops { ^ ^ zhongchun-yin). The thirty-six initial characters were 16 explicitly named in the SSDZ for the first time. In the

Yuniina the 206 Guanovun rhymes are grouped into forty-three zhuan ( literally 'turns') according to their phonological characteristics; hence, the book consists of a set of forty-three rhyme tables. In the SSDZ the forty- three zhuan are regrouped into sixteen she, or rhyme classes.

In the past it was a common assumption in the field of the history of Chinese phonology that the rhyme tables were formed on the basis of some rhyme books; hence, the sound system underlying the rhyme tables was not explored separately from Ancient Chinese phonology, i.e., the sound system of the Oievun. Nevertheless, recently scholars in this field have affirmed that the rhyme tables were based upon the sound system of a real language, and they identify the language represented by the rhyme tables with standard

Chinese of the late Tang and early Song period (Fulleyblank

1970-71, 1984 and Hsueh 1978, 1985). Fulleyblank calls the language represented by the rhyme tables Late Middle Chinese

(LMC). The division of the Ancient Chinese era into two stages, namely. Early Middle Chinese and Late Middle

Chinese, offers a new view on the periodization of the history of the Chinese language and a prospect of refining upon the lineage of the standard Chinese language."* 17

1.3.4 Transcriptional evidence

As is remarked above, the rhyme books and the rhyme tables provide major information on the history of Chinese phonology. Also, it has been a standard method to compare traditional phonological categories with their reflexes in modern Chinese dialects and corresponding Sino-Xenic pronunciations for the reconstruction of the EMC and the LMC sound systems. Yet, this method of comparative reconstruction is subject to some limitations. To be specific, the traditional terminology for describing the phonological characteristics of Chinese syllables is basically classificatory rather than articulatory or acoustic. Therefore, the classificatory terms used in the native phonological tradition can only be interpreted phonemically. Moreover, our present understanding of the pedigree of modern Chinese dialects alone is rather insufficient to carry out any complete comparative reconstruction of a pre-modern stage of the Chinese language

(Hashimoto 1978:4).

Yet, there are other kinds of source materials, from which we can infer phonetic information on pre-modern

Chinese. Such source materials in general consist of the following:

1. Foreign words in Chinese texts: Chinese characters are sometimes used to transcribe foreign words, for example, Buddhist technical terms. If the 18

pronunciations of the original words are known, in many

instances they make it possible to reason out the

contemporary pronunciations of the Chinese characters used

to transcribe them.

2. Chinese loanwords in foreign languages: Chinese

words were borrowed early into other East Asian languages,

which were culturally in close contact with the Chinese

language. Since Chinese loanwords of this kind were

naturalized into the borrowing languages in their early

stages, the loanwords have undergone the sound changes that

have occurred in the borrowing languages separately from the

original Chinese words. Therefore, we have to do much

research into the early stages of both Chinese and the

borrowing languages before we undertake any thorough

comparative reconstruction based on those words.

3. Sino-Xenic characters : China's neighboring

countries— namely, Japan, Korea, and Vietnam— borrowed

Chinese characters to write their languages before they

created their own writing systems. The pronunciations of borrowed characters were adapted to the native Japanese,

Korean, and Vietnamese sound systems, and were often subject

to the sound changes of those languages. Since the bulk of

Sino-Xenic characters were borrowed during the Tang period,

Sino-Xenic pronunciations are mainly based on the Tang

standard language. Therefore, Sino-Xenic pronunciations provide valuable information for explicating the 19

classificatory terms used in the traditional Chinese

phonological literature (Fulleyblank 1984:62).

4. Transcriptions of Chinese texts in foreign

scripts: Some Chinese texts were transcribed in foreign

alphabets, e.g., Tibetan, Uighur, hP'ags-pa, Korean, and,

more recently, roman.Transcriptions of this kind are an

indispensable addition to the conventional source materials

for the history of Chinese phonology. For example,

Hashimoto (1965) and Cheng T. F.(1976) examine the

hP'ags-pa transcription of the thirteenth century in detail,

and apply their findings to the study of Ancient Chinese

phonology. Also, it is this kind of transcriptional

evidence that is most relevant to the present study, for the main source material for this study, viz., the Ponvok

Noaoltae (PN) and the Ponvok Pakt'onasa (PP) and their

revisions, include transcriptions of Chinese texts in the

Korean alphabet.

1.4 The task of the present study

In the preceding section, we have touched upon three

of the five key stages in the history of the Chinese

language in chronological order; namely. Old Chinese (or

Archaic Chinese, represented mainly by the Book of Odes, ca.

sixth cent. B.C.), Early Middle Chinese (or Ancient Chinese,

represented by the Oievun 601 A.D.), and Late Middle Chinese

(represented by the rhyme tables, ca. 10th cent. A.D.)^^ 20

Old Mandarin (represented by the Zhonovuan vinvun 1324

A.D.)f and Modern Pekingese. The period of Mandarin can be subdivided into three stages: Old Mandarin (also called

Early Mandarin), Middle Mandarin, and New Mandarin, or

Modern Pekingese (Hsueh 1975:13). By virtue of a great deal of careful research and reconstruction of EMC by many previous scholars, perhaps all that is in our ken has been explicated concerning EMC except for some remaining problems. However, EMC is not based on any single spoken dialect, but is a national literary standard language of eclectic nature. On the other hand, LMC is considered to be based on the standard language of the late Tang and early

Song period. Therefore, the two stages of the Middle

Chinese period need to be reconstructed separately, although their phonological categories are congruent in general

(Pulleyblank 1984:xiv).

We may also turn our attention to a new perspective in tracing back the phonological history of standard

Chinese. That is, I believe that it is reasonable to ground our starting-point of reference on Modern Pekingese, a living language of which we have a solid knowledge, and trace back the phonological history of standard Chinese (cf.

Hsueh 1975:16). Therefore, the phonological history of

Mandarin shall be the scope of the present study. What I shall attempt to do is to examine and explain the changes in 21

the sound systems of Old Mandarin to Modern Pekingese,

through Middle Mandarin.

Old Mandarin is represented by the rhyme book

Zhonovuan vinvun ('Pronunciations and rhymes of the central

plains' hereafter ZYYY).^^ This book was compiled by Zhou

De-qing ( ^ ) in 1324 A.D. based on the language

spoken around Peking in the Yuan dynasty, and was used as a

standard reference book for versification of the vernacular

poems known as qu ( ). There has been a considerable

amount of research on this book including the following

notable works : Zhao Y. T. 1936, Lu Z. W. 1946, Hattori and

Toodoo 1958, Stimson 1966, and Hsueh 1975 (see

Bibliography). While the earlier studies on the phonology

of ZYYY were done largely on a phonetic basis, Stimson

(1962) first tried to work out the phonemic system of ZYYY,

and Hsueh (1975) succeeded in a thorough phonemic

interpretation. Therefore, the phonology of Old Mandarin is well explained phonemically although we are still not clear

about some of its phonetic details. On the other hand,

enough information is available for the phonology of Modern

Pekingese, for it can be and has been studied directly on

the basis of the contemporary native speakers' speech.

There exists, however, a need for research on the

phonology of Middle Mandarin. Middle Mandarin can be

identified with the stage of Pekingese of the late Ming

period, about the sixteenth and seventeenth centuries. A 22 group of textbook-like rhyme books were compiled in the late

Ming and early Qing period. For example^ the rhyme book

Yunlue huitono ( ## ^y Bi Gong-chen ( f )

from Shandong province can be chosen to represent Middle

Mandarin (Hsueh 1975:13). Lu Z. W. published a series of

papers on some rhyme books of this kind (see Bibliography),

but his studies were mainly speculations on the sound values

of the reflexes of traditional phonological categories on a

phonetic basis. The phonemic system of Middle Mandarin is

yet to be examined, and should be set up separately from

those of Old Mandarin and Modern Pekingese, no matter how

similar the former may turn out to be to the latter.

In this study, however, I will use the Ponvok

Noaoltae (hereafter PN) and the Ponvok Pakt'onasa (hereafter

PP) and their later revisions as the primary source material, for these books provide extensive phonetic data

representing the sound system of Middle Mandarin. These

books include two types of pronunciation notations

transcribed in the Korean alphabet under each Chinese

character. The one on the lower left-hand side was a normative pronunciation according to a rhyme book called the

Sasong to'nahae (1517 A.D.). The other one on the lower right-hand side was the contemporary pronunciation in colloquial Pekingese in the early sixteenth century. I aim at reconstructing the phonology of Middle Mandarin based primarily on the contemporary pronunciation of Chinese 23 characters in PN and PP, with occasional reference to other contemporary works. 24

CHAPTER I

NOTES

1. As for the geographical spread and phonological traits of the different dialect groups, see Yuan 1960, Egerod 1967, and Dong 1968: 10-11.

2. According to Chao Y . R ., these dialect groups are as far apart as Dutch and English, French and Spanish, or French and Italian (Chao 1976:24,87,97,105). However, such comparisons are impressionistic only.

3. For a discussion of sociolinguistic factors of Putonghua, see Lehmann 1975:10-40.

4. For example: (1) The Erva ( ^ /f/±. ) of the Western Han dynasty (206 B.C. to 23 A.D.) is the oldest lexicon arranged by semantic categories. (2) The Fanovan { f) ) of Yang Xiong ( tfi,53 B.C. to 18 A.D.) is the first dialectal lexicon. (31 The Shuowen 1iezi ( ^ ca. 100 A.D.) of Xu Shen ( Pf ) is the first paleographical lexicon. (4) The Shiming ( ^ ) of Liu Xi ( ^/j second cent. A.D.) is the earliest etymological lexicon.'"' (5) The Jinadian shiwen ( A 583 A.D.) of Lu De-ming ( ft; ) is an important glossarial work on the Confucian classics. See Wang Li 1980:5-7. 9 5. For a discussion about the emergence and development of phonological research during the Qing period, see Elman 1982.

6. The Chinese characters can be classified into 'six types based on the principles of character formation' ( /t ^ liu-shu). They are (1) 'pictographs ' ( ^ xiangxing), (2) 'simple ideographs' ( ^zhishi),(3) 'compound ideographs' ( ^ huiyi), (4) 'phonetic compounds' ( ^ xingsheng, or xiesheng), (5) 'derivative characters' ( II ^ zhuanzhu), and (6) 'loan characters' ( jiajie). These Chinese terms for the 'six principles' were originally named by Xu Shen ( f-f 55? to 149 A.D.), and their English translations are quoted from Chao Y. R.'s Mandarin Primer (1948:68).

7. The great increase of the phonetic compounds is attested by the following facts cited from De Francis (1984): 25

(1) Phonetic compounds constituted 34 percent of the characters inscribed on Oracle Bones of the Shang dynasty (16th c. B.C. - 1066 B.C.). (2) As high as 97 percent of the characters in the Kanaxi dictionary of 1716 are phonetic compounds. For detailed evidence of the importance of the phonetic compounds, see De France 1984:74-115.

8. This explanation of the organization of the Oievun is adapted from Baxter 1977:8-9. Certain conventional terms which may be unfamiliar to Western linguists are used in this description of the organization of the Oievun. For the definitions of such terms for the Chinese syllabic structure, refer to Section 2.1.

9. The division of the pina-shena part into two sections does not reflect any tonal difference between the two sections; but it was simply an editorial convenience since the pina-shena part had more characters than any other tone.

10. For a list of principal revisions of the Oievun. see Chang 1974:74.

11. Pulleyblank (1984:135) states "This was merely a case of separating the kai-kou and he-kou parts of certain rhymes and did not in itself imply any increase in distinct homophone groups (Pulleyblank 1985:135)." Also see Dong 1968:84-88 and Baxter 1977:16.

12. For a detailed discussion of the Oievun language, see Chapter 5 of Dong 1968 and Chang 1974. For a phonemic revision of the Karlgren's reconstruction, see Chao 1941 and Martin 1953.

13. This new interpretation of the identity of the language represented by the Oievun is mainly based on two kinds of evidence: Lu Fa-yan's own preface to the Oievun and poetic rhymes during the Tang period. See Martin 1953:4, Denlinger 1961:2-5, Chang 1974:67, Wang Li 1980:49, and Pulleyblank 1984:3, 130.

14. Ramsey (1987:122-123) remarks the difference between Sanskrit and Chinese syllabic analyses as follows: "If Chinese phenologists had followed through with the Sanskrit-style analysis, which they had begun in the initial consonants, they would have segmented the rest of the syllable into vowel units and, in many cases, a final consonant (i.e., -m, -n, -ng, -p, -t, -k). But they did not." This statement is, I believe, misleading without some qualification. Chinese phenologists also indeed segmented 26

the final into smaller units, i.e., the medial, the main vowel, and a syllabic ending; however, they did not mark them in terms of alphabetic symbols, but instead marked sub­ final segments by means of the format of a rhyme table.

15. This list of thirty-six initials is different from the set of thirty-six initials which can be derived from phonemic analysis of the fanqie upper characters in the Oievun. The former arose later than the latter, and hence the former probably reflects a later stage of the language. See Baxter 1977:21.

16. The places of articulation and their literal translations and English equivalents are as follows: (1) chun-vin ( ^ 'lip sounds', or 'labials'), (2) she-vin # 'tongue sounds', or 'alveolar stops', (3) va-vin ( if- ^ 'back-tooth sounds', or 'velars'), (4) chi-vin ( lÈf % 'front-tooth sounds', or 'alveolars affricates and alveolar fricatives ' ), and ( 5 ) hou-vin ( ^ ' throat sounds', or 'gutturals'), the two initials lai and B ri respectively ate classified as ban-she ( ^ $ 'half tongue sound', or semi-alveolar stop') and ban-chi ( ^ ^ 'half front-tooth sound', or semi-alveolar fricative'). The four phonation types are as the following: . (1) qing ( 'clear', or 'voiceless'), (2) ci-aina ( 'second clear', or 'voiceless aspirated'), (3) zhuo ( y p 'muddy', or 'voiced aspirated, or murmured'), (4) bu-ainq- bu-zhuo ( /f- ' not-clear-not-muddy ', or 'voiced'), later called ci-zhuo ( 'second muddy', 'voiced'). Cf. Hashimoto 1965 and Pulleyblank 1984:63-70.

17. See Baxter 1977:18-19 and Pulleyblank 1984:73-79. However, this view seems to be based on the assumption that there is one single type of rhyme tables, in which the words of any division must have the séune vowel. See Hsueh 1985 for a discussion of the phonemic structure of the syllabic finals of rhyme tables.

18. Luo (1933) states , based on Karlgren's reconstruction of the Oievun. that the inner series seem to have higher vowels, while the outer series seem to have lower vowels. For a new interpretation of the meaning of the inner/outer distinction, see Hsueh 1985.

19. By the standard Chinese language here I mean the nationally accepted norm for pronunciation in China in any given period.

20. During tl^ Yuan (Mongol) dynasty, at first, the Uighur script ( ^ 7C weiwu) was borrowed to write Mongolian; later, the hP'ags-ba script ( /\ /e- 6, ba-siba) was created between 1260 and 1269 A.D., and was used officially for 27 writing Mongolian and transcribing Chinese. See Yu 1974:11- 15 and Hashimoto 1978:63-76.

21. The first term in each case— first used by French scholars, and later by Pulleyblank— seems to be more appropriate, and has become prevalent. The second term in each case was first used by Karlgren, and is well-known.

22. For a detailed information on ZYYY and a discussion about the phonology of ZYYY, see Stimson 1966 and Hsueh 1975. CHAPTER II

THE KOREAN SOURCES OF MANDARIN CHINESE

2.1 The invention of the Korean alphabet^

Korea has been in close contact with China in various fields ever since prehistoric times. It is probable that

Chinese characters began to be used in Korea before the end of the first century A.D.^ From that time Chinese characters were used for writing the Korean language until the invention of the Korean alphabet in the mid-fifteenth century. However, since the Korean language is very different from the Chinese language grammatically, it must have been difficult for Koreans to write their language by means of Chinese characters.^ Koreans' realization of such difficulty alone would have motivated them to create their own writing system.

Moreover, in the early part of the Yi dynasty,

Koreans' realization of such difficulty met with positive factors which promoted the invention of the Korean alphabet

(Ledyard 59-75). The Yi dynasty (1392-1910) had close relations with the Ming dynasty (1368-1644) of China in many aspects of culture, both being Confucian states.

28 29

Consequently, there was a great need to translate Chinese

books into Korean. At that time, some Koreans studied their

neighbors' languages and scripts, including the Mongolian

hPags-pa alphabetic script (see Chapter 1 Fn. 20). In

addition, some Korean scholars were then familiar with

Chinese rhyme books and rhyme tables, and were able to apply

their knowledge of traditional Chinese phonology to the

study of the Korean sound system.

Under such circumstances. Great King Sejong ( ^ ,

the fourth monarch of the Yi dynasty) had a keen interest in

phonological studies, and acquired a good knowledge of

traditional Chinese phonology. He was particularly

concerned about the necessity for a new phonetic writing

system to write the Korean language. It was he who first

announced the Korean alphabet in 1443, and entitled it

Hunmin chona'um ( gUj ^ jf ^ 'Correct Sounds for

Instructing the People').* The alphabet was also called

Chong ' um ( jH ^ ' Correct Sounds ' ) or Onmun ( ^

'Vernacular Script'). Initially, Hunmin Chong'um was

composed of 28 basic letters: 17 single consonants and 11 vowels. Three consonant letters and one vowel letter among

them became obsolete. The remaining 24 letters (14

consonants and 10 vowels) are used currently, and these are generally called Han'gul ( ^ , 'Great Writing") (see

Section 4.1). 30

Hunmin Chona'um was intended to be used for a triple purpose at the outset, and was used so in the Yi dynasty: first, to write the Korean language; second, to write Sino-

Korean pronunciation; third, to transcribe the Chinese language.* Accordingly, the creation of Hunmin Chona'um gave rise to three types of phonological materials as follows :

1. Korean phonology. The Hunmin Chona'um (HC) and the Hunmin Chona'um haerve ( ^ îH. ^ , HCH) were originally published in a single binding in 1446, but are usually treated as two separate books (see Bibliography).

HC is a primer for Hunmin chona'um. whereas HCH is a learned commentary. HCH expounds the rationale and the design of the Korean alphabet as well as the Korean sound system at the time of the invention of the alphabet.

2. Sino-Korean pronunciation. Sino-Korean characters generally keep their basic, original meanings, but their pronunciations conform to the sound system of the Korean language. At the beginning of the Yi dynasty there were different layers of Sino-Korean pronunciation due to differences in borrowing dates, Chinese dialectal origins, and subsequent sound changes of Korean (cf. Koono 1968).

Therefore, the Tong'auk chonc'un ( ^ ^ 'Correct

Rhymes of the Eastern Country', 1447, TC) by Sin Suk-chu

( 7 ^ - ^ r 1417-1475) et al. was compiled by order of

Great King Sejong for the purpose of standardizing Sino- 31

Korean pronunciation. The editors compiled TC by

transliterating the fanqie spellings in the Guiin vunhui

iuvao ( "ÿ ^7 ^ ^ ^ • 1297, hereafter GJYHJY) into the Korean alphabet, and by applying their knowledge of Chinese

rhyme tables (Ledyard 1966:313, Hashimoto 1977:106, and Yu

1981a:45). The organization of TC is different from that of

Chinese rhyme books of the Oievun tradition (see Section

1.2.2). In TC, characters are first classified into rhyme

categories, and then are arranged in order of representative

initial characters, and finally are arranged in order of the

four tones of Middle Chinese.*

3. Transcription of Mandarin pronunciation. The

Honomu chonc'un vokhun ( ^ ^ g//| 'Correct Rhymes of

the Hongwu Era with Transliterated Pronunciation Notations',

1455, hereafter HHCUYH) and the Sasong t'ona'ao ( ^ ^

'Thorough Study of the Four tones', ca. 1450, hereafter

SSTG) were published. SSTG was compiled as a reference book

for HMCUYH.^ Since the pronunciation notations in HMCUYH and SSTG are relevant to those in the major sources of the present study, namely, the Ponvok Noaoltae (PN) and the

Ponvok Pakt'onasa (PP), we shall discuss those books and the relation between those books and PN and PP in the next section. 32

2.2 Transcription of Mandarin pronunciation

2.2.1 The transcriptional sources

The Honowu zhenovun ( ^ jf ^ 'Correct Rhymes of

the Hongwu Era', 1375, hereafter HWZY) was compiled by Yue

Shao-feng ( ^ ) et al. under the auspices of the

first emperor of the Ming dynasty in an effort to specify

the standard pronunciations of Chinese characters. Thus,

HWZY was considered the most authoritative book for learning the standard Chinese pronunciations of Chinese characters in the Ming period.

Sin Suk-chu et al. started transliterating the fanqie spellings in HWZY by order of Great King Sejong about 1445.

Since the organization (i.e., the arrangement of characters) of HWZY basically followed the Oievun tradition, representative initial characters were not listed in HWZY.

For this reason, the Korean editors first sorted out 31 representative initial characters from the upper characters of the fanqie spellings ii* HWZY. Then, they transcribed into the Korean phonetic notation the 31 representative initials and the 76 rhymes of HWZY, and added two types of pronunciation notations (the "correct" pronunciation and the

"vulgar" pronunciation) next to each fanqie spelling in

HWZY.® Thus, the editors succeeded in completing the Korean version of HWZY with transliterated notations after some ten 33

years' meticulous work in 1455, and this book was entitled

Honaiau chona'un vokhun.’

HMCUYH kept intact the original organization of HWZY,

which followed the Oievun tradition. It must have been

inconvenient for Koreans to use HMCUYH at the time when they

were familiar with Chinese rhyme tables, and used the Korean

alphabet as a means of transcribing Chinese pronunciation.

Therefore, Sin Suk-chu compiled SSTG as a reference book to

HMCUYH in the following manner:

(1) He deleted the definitions of characters;

(2) classified the characters into rhyme categories;

(3) within a rhyme category, arranged characters first

in order of representative initial characters, and second in

order of the four tones of Middle Chinese;

(4) added a "correct" pronunciation notation in front

of the first character in each homophone group, and in

addition added "vulgar" pronunciation notations to some

characters. SSTG is not extant. However, Ch'oe Se-jin

reprinted the original "introduction" ( j^ij 'general

examples') to SSTG in his Sasong t'onchae, and referred to

SSTG in his preface to the Sasona t'onahae. This gives us a

pretty good idea about the arrangement of SSTG. The basic

organization of SSTG must have been the same as that of TC.

Ch'oe Se-jin ( 14787-1543) was a prominent

Korean linguist of both Chinese and Korean, and an expert in

Pekingese in the early sixteenth century.^ He compiled the 34

Sasona t'onahae ( V3 ^ 'Thorough Explanation of the

Four Tones', 1517, hereafter SSTH) on the basis of SSTG. He kept the general organization of SSTG, but revised and expanded SSTG in the following manner:

First, he deleted some characters and added some new ones;

Second, he added the definitions of characters on the basis of GJYHJY;

Third, he added the "present vulgar" pronunciation notations to some characters, and in addition added to some characters some "alternative" pronunciation notations based on the Menaau vunlue ( 'Abridged Rhymes of

Mongolia'), GJYHJY, or ZYYY. SSTH became a regular Chinese-

Korean dictionarv.“

2.2.2 The three types of pronunciation notations

As is mentioned in the preceding section, two types of pronunciations transcribed in the Korean alphabet are recorded in HMCUYH and SSTG. One type is called the

"correct" pronunciations ( £ ^ ), and the other type is called the "vulgar" pronunciations ( ^ - In SSTH, in addition to these two types of pronunciations, there is also another type of pronunciations, which is called the "present vulgar" pronunciations ( )• As a rule, a "correct" pronunciation notation is added to the first character in each homophone group. A "vulgar" pronunciation notation is 35

added to a character which was pronounced differently, in

mid-fifteenth century Northern Mandarin, from its "correct"

pronunciation. A "present vulgar" pronunciation notation is

added to a character which was pronounced differently, in

early sixteenth century Northern Mandarin, from both its

"correct" and "vulgar" pronunciations. The three terms

"correct" pronunciation, "vulgar" pronunciation, and

"present vulgar" pronunciation respectively shall be

abbreviated to CP, VP, and PVP, and shall be referred to as

such hereafter in the remainder of this study.

In SSTG and SSTH, we can find some statements

regarding the respective natures of the three types of pronunciation notations and the relation between them." It

is said that the CP is in accordance with the pronunciation specified by the fan-aie spellings in HWZY. The compilers of HMCUYH and SSTG recorded the VP in order to specify the contemporary pronunciations of characters in Northern

Mandarin about the mid-fifteenth century. Furthermore,

Ch'oe Se-jin recorded the PVP according to the popular pronunciations of characters as he heard in the early sixteenth century. The same CP notations and the same VP notations are recorded in both HMCUYH and SSTG. The CP and the VP notations in SSTH are almost identical with the corresponding ones in HMCUYH and SSTG except that Ch'oe Se-

jin deleted the letters for the ending consonants of the entering tone words. 36

It is generally known that HWZY is based to a great extent on Lower Yangzi Mandarin (presumably, the Nanjing dialect in particular) at the beginning of the Ming dynasty, but it is still open to question whether the sound system underlying HWZY was based on a single dialect or more than one dialect. Moreover, the compilers of HWCUYH and SSTG set up the CP by transcribing the fanqie spellings in HWZY probably with reference to GJYHJY. That is, the CP was arrived at mainly by examining rhyme books rather than by direct description of speech sounds. Thus, we may also question the degree of accuracy of the CP in representing the sound system underlying HWZY. These two questions are still debatable issues. Therefore, until these issues are settled, the CP glosses may be somewhat difficult to use as reliable phonetic data for the study of the historical development of Northern Mandarin pronunciation.

On the other hand, we may expect that the VP and the

PVP notations in HMCUYH, SSTG, and SSTH provide us with phonetic information about the sound system of Mandarin about the late fifteenth century. However, contrary to our expectation, Kang S. H. (1973 and 1974) points out that the

VP and the PVP notations in HMCUYH, SSTG, and SSTH do not give sufficient phonetic information about the then sound system of Mandarin for the following two reasons:

First, the VP notations are not remarkably different from their corresponding CP notations since the former were 37 obviously Influenced by the latter. For instance, both the

CP and VP notations transcribe the same initial consonant system in the same manner.

Second, since the VP and the PVP notations are added only to a small number of characters, they do not amount to sufficient phonetic data.

Nevertheless, in this regard, we are fortunate enough to find extensive records of the VP and the PVP notations in the Ponvok Noaoltae (PN) and the Ponvok Pakt'onosa (PP).

All the Chinese characters in these two books are consistently annotated with both the VP's and the PVP's.

Each character in these books is annotated with a VP on its lower left-hand side, and also with a PVP on its lower right-hand side. Thus, the VP and the PVP notations in these two books are more extensive in comparison with those in SSTH. More importantly, since the PVP notations in PN and PP were recorded by Ch'oe Se-jin according to the pronunciation of Pekingese as he himself heard in the early sixteenth century, I believe, therefore, that they are the most accurate phonetic data for the sound system of Middle

Mandarin. 38

2.3 Noaoltae and Pakt'onasa

2.3.1 Introduction

The Nolaoltae { ^ zl 'K. Laoqita 'Good Old Chinese' ) and the Pakt'onasa ( Putongshi 'Interpreter Pak') were popular textbooks of colloquial Chinese in the early part of the Yi dynasty." These two books and their later editions were, due to their similarity, not only mentioned and published together in the Yi dynasty, but also have often been treated together in modern studies of them.

Following this tradition, we shall also deal with them together in this study, although they are two separate books.

Although the authors and the dates of publication of the Noaoltae and the Pakt'onasa are unknown, it is probable that these two books were first written sometime in the latter half of the fourteenth century by some unknown

Northern Chinese, and were soon imported into Korea as textbooks of colloquial Chinese." The language of these books is Northern Mandarin of the Yuan period. The Noaoltae is written in dialogue, and deals with trade and other practical matters which travelling merchants experienced."

The Pakt'onasa is written in narrative, and deals with social, cultural, and literary items extensively.

The original of the Noaoltae is not extant, but its revision by Ch'oe Se-jin is available at present (see 39

Bibliography). Ch'oe's revision of the Noaoltae keeps the

language of the original Noaoltae intact except for a few particles, which Ch'oe updated (Kang S. C. 1985:25-26). In addition, the Noaoltae revised by Ch'oe is printed as the

Chinese text of PN. That is, the language of the original

Noaoltae is identical with the Chinese text of PN for all but a few words. In like manner, we may suppose that the language of the original Pakt'onasa is identical with the

Chinese text of PP except for a limited number of words, although the original Pakt'onasa does not survive.

Being the most authoritative and most popular textbooks of colloquial Chinese, the Noaoltae and the

Pakt'onasa were later revised or translated and annotated into Korean a number of times in the Yi dynasty. Thus, the

Noaoltae, the Pakt'onasa. their later revisions and Korean translations formed a series of pre-modern Korean textbooks of spoken Chinese." Among this series of books, the following three versions are particularly relevant to the study of the diachronic sound changes of Mandarin, for the pronunciation notations of the following respectively represent Mandarin Pronunciation of three different periods in chronological order:"*

1, The Ponvok Noaoltae and the Ponvok Pakt'onasa.

Ch'oe Se-jin translated and annotated the Noaoltae and the

Pakt'onasa shortly before the publication of his SSTH

(1517). His translations were also entitled simply Noaoltae 40

and Pakt'onasa. respectively. Ch'oe also wrote the Ponvok

Noqoltae Pakt/onqsa_BpinnYe, (

'Introduction to PN and PP'), which consists of a number of

guidelines for the readers of his PN and PP concerning the

pronunciation of Chinese and Korean of the period." Nam

(1972) proposed to rename Ch'oe's translations Ponvok

Noaoltae ( ^ ^ ^ & X. ) Ponvok Pakt'ongsa ( f f ^ ) after the title Ponvok Noaoltae Pakt'onasa pomnve

(hereafter PNPP) in order to distinguish Ch'oe's

translations from the original Noaoltae and the original

Pakt'ongsa on the one hand, and from the later revisions of

Ch'oe's translations on the other hand. Since then it has

been a convention to call Ch'oe's translations respectively

Ponvok Noaoltae (PN) and Ponvok Pakt'ongsa (PP), which we

also shall follow in this study. PN consists of two volumes, and is extant in its entirety. PP consists of

three volumes, but only the first volume of it is extant

(see Bibliography).

2. The Noaoltae onhae and the Pakt'ongsa onhae.

The Noaoltae onhae ( ^ ^ ■k, Laoqita yanjie

'Vernacular Exposition of the Noaoltae. hereafter NO) and the Pakt'ongsa onhae ( ^ ^ ^ Putongshi yanjie

'Vernacular Exposition of the Pakt'onasa. hereafer PC) are written with the same methods of translation and pronunciation notation, and also are printed in the same 41

format. NO was published in 1670 by an unknown author. PC

was published in 1677, and according to I Tam-myong's

( ^ ) preface to PC, this book was written by Pyon

Som ( ^ ^ ), Pak Se-hwa ( ; ^ ^ ) et al. The Chinese

texts of NO and PO are identical with those of PN and PP,

respectively, but the translation and the pronunciation

notation of PN and PP are revised in NO and PO. NO and PO

are extant entirely, and their reprints are readily

available (see Bibliography). For this reason, NO and PO

have been used as the source material more often than any

other edition of this series.

3. The Pakt'ongsa sinsok onhae.

The Noaoltae sinsok ( ' New Interpretation of

the Noaoltae'. 1761) and the Pakt'ongsa sinsok (

$ff 'New Interpretation of the Pakt 'onasa'. 1765) are recensions of the Chinese character texts of PN and PP (or those of NO and PO), respectively. Hence, the language of

NS and PS is Northern Mandarin of the early Qing period.

Kim Ch'ang-jo ( ) et al., the authors of PS, also translated and transcribed PS, and published the Pakt'onasa sinsok onhae ( ^ ^ff ^ ^ ^'Vernacular Exposition of the Pakt'ongsa sinsok'. 1765, hereafter PSO). PSO survives in its entirety. The Noaoltae sinsok onhae. known only by its title, is lost.

In PN and PP, every Chinese character is marked with a VP on its lower left-hand side and a PVP on its lower 42 right-hand side. In addition, every Chinese sentence is followed by a Korean translation. This arrangement of the

Chinese text, the Korean translation, and the pronunciation notations is kept in the later revisions of PN and PP.

However, while the VP notation of PN and PP was kept intact, their PVP notation was revised first in NO and PO, and again in PSO. In terms of the corresponding time, the pronunciation notation of PN and PP probably represents the pronunciation of Northern Mandarin of the early sixteenth century; the pronunciation notation of NO and PO represents the pronunciation of Northern Mandarin of the mid­ seventeenth century; the pronunciation of PSO represents the pronunciation of Northern Mandarin of the mid-eighteenth century. In the present study, we shall endeavor to reconstruct the sound system of Middle Mandarin on the basis of the pronunciation notation of PN and PP. Also, we shall make use of the pronunciation of NO and PO and that of PSO for tracing the sound changes from Middle Mandarin to Modern

Pekingese.

2.3.2 A brief review of previous studies

There have been numerous studies on the Noaoltae. the

Pakt'onasa. and their Korean versions in recent decades.

Such studies can be divided into three broad groups as follows: (1) bibliographical studies, (2) phonological studies on the basis of the pronunciation notations in the 43

Korean alphabet, (3) grammatical and lexical studies of the

Chinese texts.The phonological studies consist of two types of research. One comprises studies of the history of

Korean phonology, the other studies of the history of

Chinese phonology. Since the present study is primarily concerned with Chinese phonology, we shall pay more attention to the latter type of research. Among the previous studies relevant to the history of Chinese phonology, the following seem to be more important, and hence a few comments shall be made on them here.

1. Hu Ming-yang (1963). The author considers that the PVP of NO and PO represents the sound system of

Pekingese of the early sixteenth century," He replaces the letters of the Korean alphabet used to represent the initials and the finals of the Chinese syllables with phonetic symbols in IPA, and thereby gives phonetic notations in IPA to Chinese characters. This article is the first attempt to make use of the Korean pronunciation notations of Chinese characters for the study of the sound system of Middle Mandarin. It is meritorious in that the author arranges the syllable finals systematically according to the structure of the final, i.e., the types of the components of the final, rather than merely following the rhymes of a rhyme book. However, apart from the listing of the finals, this article makes no effort in analyzing the 44 phonemic system of the Pekingese vowels of the period represented by the PVP notation of NO and PO.

2. Kang Sin-hang (1974, 1978a, and 1973b). The author first points out that the pronunciation notations in

PN and PP, NO and PO, and PSO respectively represent

Mandarin of three different periods in chronological order.

Kang S. H. (1974) gives phonetic notations to Chinese characters (as they are supposed to have been pronounced in

Pekingese of the early sixteenth century) on the basis of the PVP's in PN and PP. Kang S. H. (1978a) gives phonetic notations to Chinese characters (as they were supposedly pronounced in Pekingese of the mid-seventeenth century) on the basis of the PV pronunciation notations in NO and PO.

Kang S. H. (1978b) gives phonetic notations to the characters (as they were supposedly pronounced in Pekingese of the mid-eighteenth century) on the basis of the PVP's in

PSO. In these three articles, the author chooses phonetic symbols over the Korean letters used to represent the initials and the finals of Chinese syllables, and thereby converts the pronunciation notations in the Korean alphabet into the phonetic notations of the IPA. The author does not arrange the Chinese syllable finals according to the types of the combination of the components of the finals.

Instead, he merely presents characters according to the rhyme categories of a contemporary rhyme book or rhyme table book. Kang S. H. (1974) gives his phonetic notations in IPA 45

to the characters together with Lu Z. W.'s (1947a and 1947c) phonetic notations of the same characters respectively based

on the Yunlue vitona ( M > 1442) and the Yunlue huitono ( , 1642). Kang S. H. (1978a and 1978b) gives his phonetic notations in IPA to the characters, and

in addition cites Lu Z. W.'s (1947b) phonetic notations based on the Chonadina Sima Wenaona Denovun Tuiina ( ^ ^

W f ( i ^ - 1606). 3. Kang Sil-chin (1985). The author deals with the pronunciations of PN, PP, and their later revisions in the third part of his dissertation in the following manner:

First, he converts the pronunciation notations for the

Chinese characters in the Korean alphabet into phonetic notations in IPA.

Second, he arranges the Chinese characters according to the rhyme categories of HWZY, cites the CP's (of HMCUYH), the VP's (of PN and PP), and the PVp's (of PN and PP) of the characters, and compares these pronunciations.

Third, he examines the changes in the PVP's from PN and

PP to NO and PO, and to PSO.

Fourth, he draws up a table for the initial consonants in which the Korean letters used for the Mandarin initial consonants and their phonetic notations are listed, and a table for the finals in which the Korean letters used for the finals and their phonetic notations are listed. This study is probably the most comprehensive among the studies 46 which relate the pronunciation notations to the history of

Mandarin. However, this study is also done on a phonetic basis with little reference to the phonemic system of Korean and that of Mandarin Chinese.

Generally speaking, the following criticisms may be made of the previous studies:

First, the previous studies pay too little attention to the phonemic system of Korean and that of Mandarin.

Their main concern seems to be merely transliterating the pronunciation notations in the Korean alphabet into the corresponding phonetic notations in IPA. They choose the corresponding IPA symbols for the Korean letters with little reference to the phonemic system of the vowels of Middle

Korean (12th - 16th cent.).“ Moreover, they say nothing about the phonemic systems of Mandarin of different historical periods. Thus, although they give Chinese characters phonetic notations in IPA, these may be questionable, and remain to be interpreted phonologically.

In order to figure out accurately the sound values of the

Korean alphabet used for transcribing Middle Mandarin, in my opinion, the phonemic system of Middle Korean and that of

Middle Mandarin should be given careful consideration. By adopting the phonemic approach, we shall be better prepared to describe and to account for the phonetic and phonological interference (which could have occurred in transcribing

Mandarin pronunciation in the Korean alphabet) on the one 47 hand, and the diachronic sound changes of Mandarin on the

other hand.

Second, the previous studies insufficiently analyze the

Chinese syllable. To be specific, they speculate on the phonetic approximation of the whole final as represented by the Korean pronunciation notation, but neglect to examine the medials, main vowels, and endings separately. However,

since assimilation is prevalent within the final, the phonetic realizations of finals should be examined further in respect of the medial, the main vowel, and the ending.

In this way, we shall be in a better position to determine the phonetic value and phonemic status of segments in a final, and also to describe the historical development of individual segments. In order to achieve this goal, I believe we should take the hierarchical structure of the

Chinese syllable as the basic pattern for phonological analysis of the Chinese language, and examine and explain the transcriptional data strictly according to the basic pattern.

Third, the previous studies fall short of describing and explaining the diachronic changes of Mandarin systematically. They speculate on the sound system of

Middle Mandarin in phonetic terms on the basis of the Korean transcriptional data, but say that they (e.g., Kang S. H. and Kang S. C.) expect that future studies will make use of the Korean pronunciation notations in order to help 48 elucidate the history of Mandarin phonology.I fully agree with them on this statement of a common purpose.

However, they do not explain why their studies cannot achieve that purpose, nor suggest how such prospective studies are to be carried. I think, in this regard, that the crucial shortcoming in the previous studies is that they have not compared systematically the sound system of Middle

Mandarin (MM) which they have arrj.ved at with that of a certain key stage in the history of the Chinese language.

Of course, the previous studies frequently refer to the sound classes (i.e., representative initials and rhyme categories) of rhyme books and rhyme tables and sometimes their reflexes in MP. However, the point here is that they have not described explicitly the phonemic structure of any stage of Mandarin, nor have they specified the sound correspondences by drawing a set of a diachronic rules between MM and any other stage of Mandarin.

We certainly know more about the phonology of Modern

Pekingese (MP), which is directly based on the contemporary native speakers' speech, than any other key stage of the

Chinese language. Moreover, MP follows MM lineally and closely. On the other hand. Old Mandarin (OM) is probably related to MM in a similar manner, and in addition we have a pretty solid knowledge of OM phonology (see Section 1.3).

Therefore, in my opinion, it is most logical to compare MM phonology with MP phonology first, and with OM phonology 49 next. By adopting a strictly phonemic analysis along the line of the Chinese traditional analysis of the syllable, we shall be able to elucidate the entire phonological history of Mandarin systematically. 50

CHAPTER II

NOTES

1. This section has benefited from Ledyard (1966), a comprehensive study on this topic, to which the reader is referred for more extensive discussion on this topic.

2. This fact is inferred from a remaining epitaph on a tombstone named Chomch'an-bi ( ) written in Lishu ( ^ ^ ) style of Chinese characters. See Pyon 1978. The majority of Sino-Korean loanwords may be dated around 600 A.D. See Karlgren 1954:215-216.

3. In order to reduce the difficulties of writing Korean by means of the Chinese characters, Koreans developed their own methods of using Chinese characters for the writing of Korean as early as the late sixth and early seventh centuries. Among such methods hvanach'al ( ^ ) and idu ( Ê tt ) were most widely used. In hvanoch'al texts, Chinese characters are used as phonograms in Korean word order. In idu texts, Chinese texts are rearranged to conform to Korean syntax, and Korean particles are inserted. For a detailed discussion of pre-alphabetic writing systems of Korean, see Ledyard 1966:29-57 and Yu 1981b:26-36.

4. In the section for the 12th month of Sejong's 25th year (1443) of the Belong sillok ( ^ ^ 'Annals of the Sejong Era'), the following statement'appears :

& 4 %% jc =: f/\ dE "In this month His Highness personally created the 28 letters of the vernacular script (onmun)." It is an open question whether Sejong invented the Korean alphabet by himself, or he directed certain scholar officials to invent it (of. Ledyard 1966:97-99 and Yu 1981b:51). I suspect that Sejong must have been the principal inventor of the Korean alphabet, though he probably consulted with certain scholar officials about this matter.

5. See Lee 1972:8-9 and Yu 1981b:44-45. Nowadays, the Korean alphabet is used for writing the Korean language and Sino-Korean pronunciation, but is seldom used for transcribing the Chinese language.

6 . For a thorough study of the Tong'auk chona'un. see Yu 1981a. 23 representative initial characters and 91 rhymes are established in TC (ibid., pp. 35-39). Yu states that 51 the organization of TC follows that of the Menggu yunlue ( ^ 'Abridged Rhymes of Mongolia) (ibid., p. 234). Both the Yunlue vitono and the Yunlue huitong of the Ming dynasty also follow the same organization.

7. Yu (1974:153-156) argues that HMCUYH and SSTG were published as two parts of a single book. Kang S. H. (1973) states SSTG was probably completed together with HMCHYH in 1455. However, SSTG and HMCUYH are treated as two separate books due to their different natures.

8 . "Vulgar" pronunciation glosses are added to some of the characters which were pronounced differently in Northern Mandarin from their "correct" pronunciations.

9. For comprehensive discussion of HMCUYH, see Yi 1959, Chong 1972, and Ch'oe 1975.

10. Ch'oe Se-jin compiled dictionaries, glossaries, and also wrote textbooks. As for his life and scholarly achievements, see Ledyard 1966:324-330, Pak T. G. 1974, Ch'en Z. F. 1981:111-121, and Yu 1981b:101-108.

11. For comprehensive discussion of SSTH, see Kang S. H. 1973.

12. SSTH has two introductions: the one has three paragraphs, and the other has twenty-six paragraphs. The following statement in Paragraph 2 of the 26-paragraph-long introduction to SSTH specifies the criterion of "correct" pronunciation : ^ IE. 1'^ ^ pA E a S â ? "The correctness and the vulgarity of pronunciation are determined solely according to HWZY."

As for the definition of the "vulgar" pronunciation, we find the following statement in Paragraph 1 of the introduction to SSTG:

"As for those [pronunciations] which are popularly used in the contemporary pronunciation, but are not congruent with rhyme books and rhyme tables; a "vulgar" pronunciation gloss is written under a fanqie spelling, character by character."

The following statement regarding the "vulgar" and the "present vulgar" pronunciation glosses is found in Paragraph 8 of the 26-paragraph-long introduction to SSTH; 52

e & 4^ . e ^ 3^ 7 ^ ^ P^' ' t ^'f ^ 4 ^ 4 C # ,

e ^ « : # ® Î4S. « • SI ^ îf f i *" * ¥ i- « •

"Among the glosses, those which are called the "vulgar" pronunciations are merely those which are originally marked as the "vulgar" pronunciations in SSTG. Those which are called the "present vulgar" pronunciations are those I mark as vulgar pronunciations at the present time. Some of the "present vulgar" pronunciations are written, while some of them are not written. This does not mean that some [characters] have "present vulgar" pronunciations, while the others do not. I gloss them as [I] have heard."

13. The meaning of the title "Nogoltae, or Laoqita" is not clear. However, it may possibly mean literally either "Great China" or "the old Chinese"; for ^ ^ is the same as ^ , and means either China or Chinese. See Yang 1957:197 and Kang 1985:15-17. All our sources agree that the title "Pakt'ongsa, or Putongshi" means literally an interpreter surnamed Pak.

14. Historically, the titles of these two books are mentioned for the first time in the section for the 5th year (1423) in the Selono sillok ( 4ËI ^ 'Annals of the Sejong Era') to the effect that they should be printed together with certain other Chinese books. Since then the two books are mentioned a number of times in the Selong sillok and the Sonalona sillok ( /K ^ ^ Wl 'Annals of the Songjong Era'). See Ogura (1920) 1964:556-558.

15. Min (1964) states that the Nogoltae was written by an unknown Mongol between the years of 1382-1385 A.D. , a Korean Buddhist monk's courtesy name, is found on page 65 of volume 1 of the Pakt'ongsa onhae. Zhu (1957) infers from this name that Pakt'ongsa was written between the years of 1346-1368 A.D. See Zhu 1957:71, Min 1964, Dyer 1983:3, and Kang S-C. 1985:22.

16. See Dyer (1983) for an English translation of the Nogoltae.

17. See Kang S-C. (1985:31-40) for a list of this series of books and bibliographical information on them. The Nogoltae and the Pakt'ongsa were also translated into Japanese, Manchu, and Mongolian as textbooks for interpreters in the Yi dynasty. See Ogura (1920) 1964:559, Cota 1953:13, and Dyer 1983:8-9. 53

18. As pointed out by Kang S. H. (1974, 1978a, 1978b) and Kang S. C. (1985:38).

19. Besides, Ch'oe also wrote such glossaries as the ^ Tanlahae (%È $1^0: ) / the Noaoltae chimnam ( :^ 'Z Â ) r and the Pakt'onosa chimnam~ ) ' The Ponvok Noaoltae Pakt'onosa chimnam was publis^^ together with these glossaries in a single binding as a reference book under the title of Noaoltae Pakt'onosa chimnam for the use of the reader of PN and PP. See Bibliography.

20. A list of previous studies on the Noaoltae. the Pakt'onasa. and their Korean versions is given in Kang S. C. (1985:4-8), to which the reader is referred. Also, see the bibliography of the present study.

21. Hu (1963:185) seems to assume that the "present vulgar" pronunciation notations in PN and PP are kept intact in NO and PO, but actually they were revised in NO and PO. Thus, it is expected that the "present vulgar" pronunciation of PN and PP represents the sound system of Pekingese of the early sixteenth century, whereas that of NO and PO represents the mid-seventeenth century.

22. According to Lee K. M. (1961 and 1972), the history of the Korean language can be divided into three principal periods: Ancient Korean (Proto-Korean - 11th cent.). Middle Korean (12th - 16th cent.), and Modern Korean (17th - present).

23. When I discovered this suggestion— to reassess Mandarin historical phonology by means of the Korean pronunciation notations as phonetic data— in Kang S. H.'s article (1980a) first in 1984, I as a Korean student of Chinese linguistics felt a sense of duty to follow Prof. Kang's open suggestion. This was then reinforced by my academic advisor Prof. Hsueh, who approved of my dissertation research on this topic. CHAPTER III

THE PHONOLOGY OF MODERN PEKINGESE

3.0 Introduction

There have been nuraeroue stuiies on the sound system of Modern Pekingese (MP) in modern Western linguistic terms, starting with Chao (1934), Hartman (1944), and Hockett

(1947). Chao (1968:18-56) provides a useful introduction to

MP phonology. Among recent studies on this topic, Cheng

(1973) treats MP phonology in a generative framework; Hsueh

(1980a) analyzes the phonemic structure of Pekingese syllable finals, employing the practice of rhyming as a key factor; and Hsueh (1986) presents a thorough analysis of the

MP sound system.

In this chapter, we shall summarize the important aspects of MP phonology, so that we may set up a firm basis for reference in the remainder of this study. In our discussion of MP phonology, we shall adopt the following four criteria of native speakers' intuition in phonemic analysis proposed by Hsueh (1986:4-6, 9-10): (1) the traditional analysis of the Chinese syllable, (2) traditional classifications of the medial and the ending,

54 55

(3) rhyming practice, and (4) alliteration.

Since this study is intended for general linguists as well

as for Chinese linguists, it is appropriate here to

introduce the traditional terminology used to describe the

syllable structure of Chinese before we start our discussion

of MP phonology.

3.1 The syllable in Chinese

Since the syllable in Chinese constitutes a morpheme,

the monosyllable functions as the basic building block.

Most of the monosyllabic morphemes are words by themselves

in Chinese. There are also words consisting of two or more

syllables in Chinese, and their number has been on the

increase. However, the number of such words is still rather

limited in the total vocabulary of Modern Chinese (Hsueh

1980a:494), and even those words are largely made up of monosyllabic morphemes.^ Thus, it holds true that Chinese

is by and large "a monosyllabic language— a language in which every syllable has a meaning" (Chao 1968:139).

In addition, the Chinese character, which is the

basic unit of the Chinese writing system, also represents a monosyllabic morpheme, and its pronunciation is identical with the spoken form for the monosyllabic morpheme (cf.

Section 1.2.1). Also, native speakers of Chinese are prone

to accept the syllable as the smallest pronounceable unit 56

intuitively. In short, the Chinese syllable is the basic

phonological and morphological unit.

The hierarchical structure of the Chinese syllable

can be illustrated as the following diagram;

Table 1 - The structure of the Chinese syllable

Tone

Initial Final

Medial Rime base

Nucleus Ending

As shown above, a Chinese syllable consists of three parts: the syllabic initial { ^ ^ shengmu), the syllabic

final ( a| ^ yunmu), and the tone of the syllable

shengdiao). The tone is on the suprasegmental level, while the initial, and the final are on the segmental level. The tone is a lexically distinctive pitch contour. The initial

slot may be occupied by a consonant or zero. The final is the remainder of the syllable after the initial consonant, and consists of the medial ( ^ yuntou, or yp ^ jieyin) and the rhyme base ( ^ yunji). The medial is either

zero or a semivowel. The rhyme base consists of the main vowel ( yunfu, or ^ ^ ^ ^ zhuyao-yuanyin, also called the nucleus) and the ending ( fl yunwei, or

shousheng). The ending may be a semivowel, a consonant, or zero. 57

The components on the segmental level can also be delineated by the following formula (Hsueh 1980a:495):

#(C)(M)V(E)#

The four letters in the above formula represent respectively the initial, the medial, the nucleus, and the ending. The nucleus (or main vowel) is an essential element for syllable formation; while the other components in parentheses are optional elements, and play only a differentiating role.

The tone is also an essential component of a syllable, and thus can be said to be riding on the nucleus.

Syllables which share the same rhyme base, i.e., the same vowel and the same ending, can rhyme with one another.

Within the syllable, there are subsyllabic bonds between components. For example, the rhyme base and the final function as coherent units. In this way, the relation between components is hierarchical within the syllable

(Walton 1983:34). Light (1976) pays attention to the rhyme base, and argues for treating various subsyllabic portions as coherent units and analyzing the Chinese syllable in terms of constituent structure. Hsueh (1980a) regards the rhyme base as a key unit in analyzing the sound structure of

Chinese, and presents the phonemic structure of the

Pekingese syllabic finals on the basis of the common rhyming practice. 58

Since the monosyllable has always been taken as the

basic unit in Chinese, the structure of the Chinese syllable must be taken into consideration in analyzing the

phonological system of Chinese. Furthermore, since the monosyllable has always been the primary subject of study in

the traditional studies of rhyme books and rhyme tables, and

in the modern phonological analyses of Chinese as well, it

captures the vestiges of Chinese sound change. Therefore,

in the present study, the syllable shall also be kept as the

largest unit for analysis, and the traditional, hierarchical

segmentation of the Chinese syllable shall be followed.

3.2 The initial consonants of Modern Pekingese

It is generally accepted that there are twenty-one initial consonants in MP. In addition, there is the "zero" initial, which is commonly assumed to mean absence of an initial consonant. According to Chao (1968:20), the "zero" initial has a slight obstruction of the airstream in the vocal tract in the form of a velar or uvular approximant,

[ y ] or in the majority-type pronunciation. A large minority of speakers pronounce it as a glottal stop or a pure vocalic beginning. A small minority of speakers replace it by the velar nasal [ 7) ]. Phonemically it may be logical and convenient to omit the "zero" initial, but for diachronic discussions, it would be advantageous to include it in the guttural series of our phonological analysis.^ 59

In order to reflect their phonetic realties closely, the following terminology will be used for the categories of articulation of the consonants in Table 2:

(1) The term approximant is used here to accommodate both the liquids and the zero initial.

(2) Labials are divided into bilabials and a labiodental.

(3) I choose the term "dentalveolar" instead of just

"dental" (cf. Chao 1968:22) or "alveolar" (of. Pulleyblank

1984:42) for the MP initial sounds [t, t**, n, 1, ts, ts\, s] because they tend to be pronounced a little further front in the oral tract when compared with the alveolar sounds of

English [t, d, n, s, z, 1, r], but not so far front as the interdental sounds of English [&, Ï ].

(4) The term dentalveolar sibilants, following Chao

(1968:22), accommodates the dentalveolar affricates and fricative.

(5) The term "guttural" is to include both the velars

[k], [k**], [ 7) ] ; the uvular [% ]; and the zero initial pronounced as [ y ], [ 6 ]t [? ], or a vocalic beginning.^

The velars of MP tend to be pronounced somewhat further back in the oral tract in comparison with the velars of English.

The term guttural is suitable as a cover term for both the velar and the uvular.

The following is a list of the MP initial consonants in a broad phonetic transcription: 60

Table 2 - The initial consonants of Modern Pekingese

Manner Stops Affricates Frica­ Nasals Approxi- -asp +asp -asp +asp tives mants

Place

Bilabials P P‘‘ m Labiodentals f

Dentalveolars t t“ n 1

Dentalveolar ts ts“ s sibilants Retroflexes tgh P % (orX) Palatals tq tqh ? Gutturals k k“ X {Yor

The palatals are in complementary distribution with the dentalveolar sibilants, the retroflexes, and the velars; the palatals occur only before high front vowels or glides, namely, i (or [j]) or iu (or [y]), while the others do not occur in such environments. Historically, the velars were first palatalized before high front vowels or glides; thus the palatals emerged as allophones of their corresponding velars. Later, in the former half of the Qing period, high front vowels or glides were deleted after the retroflexes.

Subsequently, the dentalveolars were also palatalized before high front vowels or glides (cf. Hsueh 1986:33-34).

Consequently, the palatals of both the velar and the dentalveolar origins merged, and this process is referred to as jian-tuan hunru ( ^ /J] 'jtc K 'the mixture of the sharp. 61 or the dentalveolars, and the round, or the velars') (Wang

Li 1980:124).

Due to the fact that the palatals are simultaneously in complementary distribution with the dentalveolars, the retroflexes, and the velars, it has been an issue of much debate how to interpret the palatals at the phonemic level.*

There are two viable solutions to this problem. On the one hand, Chao (1934 and 1968) and Hsueh (1980a and 1986) affirm that the palatals are allophones of the velars according to native speaker's intuition in addition to historical reasons. Native speakers' intuition in this matter is proved by alliteration and alternative readings.^ On the other hand, Cheng C. C. (1973:40) treats the palatals as underlying segments. You R. J. et al. (1980) also treat the palatals as autonomous phonemes on account of phonetic dissimilarity between the palatals and the other series.

It is difficult to determine which one of the above two solutions are right, for each is justifiable on good grounds; the former emphasizes native speakers' intuition, whereas the latter emphasizes phonetic dissimilarity.

However, since phonemic analysis is, in its most fundamental sense, an analysis of the native speakers' reaction to speech sounds, we prefer to treat the palatals as allophones of the velars (cf. Chao 1934:385, 1968:21 and Martin

1957:223). This choice is also congruent with the present 62

study, which treats mainly of the diachronic phonology of

Mandarin.

Now we can set up the initial consonant phonemes of

MP according to the above discussion.

Table 3 - The phonemic representations of the MP initials

Manner Stops & Affricates Frica­ Nasals Approxi —asp +asp tives mants

Place

Labials P f m Dentalveolars t e n 1

Dentalveolar c c" s sibilants Retroflexes cr cr** sr r

Gutturals k k** h

SOURCE: Hsueh 1980a: 496

The above table includes the following arrangements :

(1) We classify /p, ph, m, f/ simply as labials because these phonemes act together as a single natural class in MP.

In MP, phonemically labials occur only before the medial

/-j-/ or /-w-/, but not before /-jw-/; though phonetically the medial /-w-/ is not very noticeable when it precedes a low vowel (Hsueh 1975:38).

(2) We rewrite [ts, X , /] as /c, h, r, jzl/, respectively.

(3) 'h' is also used as a symbol for the feature aspiration when it is attached to stops or affricates. 63

(4) 'r' is also used as a symbol for the feature

retroflexion when it is attached to the sibilants.

(5) [t<^, tqih, 9 ] are treated as allophones of /k, k*',

h// respectively.

(6) The phonemes represented by a combination of two or

three symbols are not consonant clusters, but single

phonemes.

3.3 The finals of Pekingese®

Most past phonological analyses of the MP finals put undue emphasis on the phonetic impressions of indvidual

finals, neglecting their phonological properties in the whole system. Such an approach has often obscured the division and the interaction of the components of a final

(i.e., the medial, the main vowel, the ending, and the rhyme base), and therefore has resulted in a vowel system which is not congruent with native speakers' intuition reflected in their practice of rhyming.

3.3.1 The medials

In the rhyme tables of the Ming and Qing period the

Chinese syllables are classified into the following four types : kai-kou ( IJ 'open mouth'), ai-chi ( ^ ^ 'level mouth ' ), he-kou { \J ' closed mouth ' ), and cuo-kou ( xj

'pursed mouth') (cf. Zhao Y. T. 1957:161-165). There are also two pairs of contrastive terms in traditional Chinese 64 phonology. One pair is kaikou-hu ( ü '^'f^'open breath') vs. hekou-hu ( 'closed breath' ); kai-kou and ai-chi belong to kaikou-hu. whereas he-kou and cuo-kou belong to hekou-hu. The other pair is hono-vin ( ^ 'wide sounds') vs. xi-vin ( ^ 'narrow sounds'): kai-kou and he-kou belong to hona-vin. whereas ai-chi and cuo-kou belong to xi-vin.

Phonetically, the syllable medials of the four types are [^] (zero), [j], [w], and [i(], respectively, [j], [w], and [i|] are the IPA symbols for palatal, labial-velar, and labial-palatal semivowels, respectively.^ We shall use the symbols "j" and "w" to represent the features palatalization and labialization, respectively. Consequently, [j], [w], and [!|] respectively can be represented as /j/, /w/, /jw/ phonemically. In this way, we can also account for the two dichotomies, i.e., kaikou-hu vs. hekou-hu and hona-vin vs. xi-vin. The former represents a contrast between the absence and the presence of labialization, whereas the latter represents a contrast between the absence and the presence of palatalization (Hsueh 1980a:498).

There has been disagreement about how to transcribe the semivowels phonetically and/or phonemically. Chao

(1968) uses the symbols i, u, and iu (=[y])" to represent the four medials. In addition, the symbols "i, u, and Ü (or y)" are used for both phonetic transcription and phonemic representation of medials and vowels in the 65

phonological analyses of MP done in the framework of

generative phonology, e.g., Robert Cheng (1966) and Cheng

C. C. (1973).® Moreover, in these generative analyses of MP

phonology, the semivowels which occur in the medial and the

ending positions are mistakenly considered to be [+vowel] at

the phonemic level. As a matter of fact. Lu Z. J. (1985:

36) does derive the semivowels which occur in the medial

position from the underlying phonemes /i, u, y/ by a rule,

stating the rule as follows:

A vowel becomes a semivowel when it occurs as the

medial, i.e.:

V V [-vowel] / (C) ___ V ({^}) [sic]

However, this "rule" does not specify any concrete phonetic

or phonological change for the following reason:

"Vowel" is neither a phonetic feature nor a phonological

feature. It is simply a major natural class of segments.

Even if we assumed "vowel" to be a feature, this "rule" would be ambiguous at best; since "[-vowel]" would mean

consonants [-vocalic, +consonantal], liquids [+vocalic,

+consonantal], or semivowels [-vocalic, -consonantal] (cf.

Chomsky and Halle 1968). Thus, I doubt if Lu (ibid.)

actually means by the formula "V — » [-vowel]" that a vowel

("V") becomes a nonvowel ("[-vowel]"). If "[vowel]" were

replaced by [vocalic] in the rule, the rule would mean that

a vowel (tvocalic, -consonantal) becomes a semivowel 66

(-vocalic, -consonantal).® Therefore, I suspect that the inadequacy of Lu's "rule" is due to the wrong assumption that underlying high vowels occur in the medial and the ending positions.

As is stated above, it has been a common practice in the study of MP phonology to treat the semivowels, which occur in the medial and the ending positions, as vowels at the phonemic level and to transcribe them as the high vowel symbols i, u, li/y both phonetically and phonemically. But there is no justification for this confusing practice since this practice fails to capture the phonetic and phonemic differences between the semivowels [j, w, ] from the high vowels [i, u, li/y] by assigning the same vowel symbols [i, u, and u/y] for both (cf. Pulleyblank 1984:46).

Therefore, we are obliged here to make our standpoint clear on this practice so as to straighten out this problem.

In terms of structure, "only the main vowel represents the peak of a syllable, the part that is responsible for a syllable to be a syllable " (Hsueh 1980a:495). That is, at the phonemic level only the main vowel is [+syllabic], whereas the other components of a syllable are [-syllabic].

Chomsky and Halle (1968:354) propose that the feature

Syllabic replace the feature Vocalic. [tsyllabic] segments are those constituting a syllabic peak, i.e., vowels, syllabic liquids, and syllabic nasals; all remaining segments, i.e., obstruents, glides, nonsyllabic liquids, and 67 nonsyllabic nasals are said to be [-syllabic].

Consequently, it is the semivowel, not the vowel, that can occur in the medial position of a syllable at the phonemic level.

In the remainder of this study, the phonemic representations of the three semivowels will be /j, w, jw/ where j" and "w" stand for the features palatalization and labialization, respectively. We shall transcribe the three semivowels that occur in the medial position as [ j, w, t]] phonetically. In this way, we can use appropriate symbols for the three semivowels, and avoid insufficiently differentiating them from the high vowels i, u, u/y both phonetically and phonemically.

In sum, the four types of the medial can be described as follows;

Table 4 - The syllable medials of Modern Pekingese (of. Hsueh 1980a:498)

Traditional Phonemic Distinctive Phonetic classification representation features transcription

Kai-kou [—pal, —lab]

Qi-chi /j/ [+pal, -lab] [ j] He-kou /w/ [-pal, +lab] [w]

Cuo-kou /jw/ [+pal, +lab} [V] (Phonetically, we use the symbol 0 to denote the absence of a sound; phonemically, we use the symbol ^ to denote the zero initial, the zero medial, or the zero ending.) 68

3.3.2 The endings

There is a strong tendency to coarticulate the segments within the final, i.e., the medial, the main vowel, and the ending, as a coherent unit. Acoustically, diphthongs composed of two segments and triphthongs composed of three segments can occur in the final position, and there is no clear boundary between their constituent segments, e.g., [ai], [ei], [ja«], [wai ], etc.'°

However, since the ending is more closely tied to the main vowel than the medial is, the main vowel and the ending form the rhyme base (cf. Hsueh 1980:498). This strong cohesion of the main vowel and the ending is supported by the synchronic fact that regressive assimilation is dominant over progressive assimilation in MP finals as pointed out by

Cheng C. C. (1973).

Therefore, the main vowel and the semivowel ending are coarticulated as a cohesive unit and pronounced in tne form of a diphthong, e.g., [ai], [am], etc. (Hsueh

1980a:498). In this way, when the semivowels /j/ and /w/ occur in the ending position, they are phonetically realized as [i3 and [#], respectively. [i] denotes a lax vowel situated near the meeting-point of the areas i-e-i; [#] denotes a lip-rounded lax vowel near the meeting-point of the area u-o-i.“ This process can be formalized as follows : 69

f -consonantal] ___ f+vocalic]/ V ___ # L-vocalic > ^

(# denotes a syllabic boundary; a syllable boundary also functions as a morpheme boundary for most Chinese morphemes.)

In sum, the following endings occur in the syllable ending position:

Table 5 - The syllabic endings of Modern Pekingese

Traditional Phonemic Phonetic classification representation transcription

Zhi-yin { ^ ) /^/ /

Shou-yin ( 4%^^ ) /]/ [t] Shou-wu ( 0^ ) /w/ [p]

Di-e ( “II ) /n/ [n]

Chuan-bi ( '^ 4 ) /{)/ [ij]

In addition, "-r" occurs in the ending position of the r- suffixed syllables and the syllables representing such words

^ / and zt. ( see Section 2.4).

In spite of the fact that a main vowel and a lax vowel ending form a diphthong phonetically, we affirm that the main vowel and its following ending can be recognized as two separate segments, and thus two separate phonemes should be set up. Hsueh (1980a) states the reason for accepting the vocalic ending as a separate segment as follows:

In MP it is still recognizable and its recognition greatly clarifies the vowel system. Moreover, classification of syllables into different types according to syllable ending is also a time-honored practice in Chinese tradition. (p.499) 70

In addition, it will be found that certain vocalic endings

are reflexes of consonantal endings in the history of

Mandarin.

3.3.3 The vowels

Phonetically, the following vowels occur in Modern

Pekingese:

Table 6 - The phonetic forms of the MP vowels

front central back

high 1 I i/y ^ ui/u 3 mid-high e/(6) 2T/o

mid-low ^/(æ) A/(3 ) low a

(The phonetic values of the MP vowels in this table are cited from Hsueh (1986:59). ü and Ï in the source are here replaced by the corresponding IPA symbols y and w, respectively. A slanting line marks a contrast of roundedness between a pair of vowels : the vowel on the left- hand side of a slanting line is [-round], whereas the one on the right-hand side is [+round].)

[9] is the only central vowel. It is also a unique high vowel in that it is slightly higher than the mid vowels, but slightly lower than the other high vowels. As a rule, [9] occurs only before a non-zero ending as follows:

% [hSl] K [hen] ^ [hag]

[kwai] jQ [wan] A [wàg] 71

In contrast/ the mid vowels [e,2r, o] occur before the zero ending, e.g., ^ij [pje], ySf [haH, ^ [two], etc. However, since words such as , etc. are more often than not transcribed as [jou], [hou], etc., it seems that the mid vowel [o] also occurs before a non-zero ending. But in reality, the [o] in [jou] and [hou] is phonetically much higher than and quite different from the [o] in [two] (cf.

Hsueh 1986:60-61). Therefore, some people, e.g., Hartmen

(1944), have identified it with [@]. Thus, ^ and become [j@#] and [hd ], respectively. The generalization that the mid vowels do not occur before a non-zero ending holds true. Hsueh (ibid.) considers [9 ] to be a "semi- phonemic" symbol in that [@], as a lax vowel, has at least two allophones, [%] or [u], due to regressive assimilation, e.g., ic [wan] -» [w;n], ^ [wag] -*[wug].

[1 ] and [2 ] ere called apical vowels: [j] occurs before the dentalveolar sibilants [ts, ts**, s], whereas [ % ] occurs before the retroflexes [tp, t ^ ‘‘, p, %].^ As to the phonetic qualities of [%] and [%], Chao (1934) states as follows:

In the Chinese syllables [tgi], [ts'i], [gt], [%1], [tsi], [ts'i], [si], there is a vowel which is a vocalized prolongation of the preceding consonant, and is understood to be present when these syllables are written as Æ , f , 8 , f i ^ in the National Phonetic Script. This is therefore a way of representing actual sounds by zero symbol. (p.374) (Underlining mine) 72

To recapitulate, there are "actual sounds" following the dentalveolar sibilants and the retroflexes, and they are vowels with coloring from the preceding consonant. In addition, it has been proved that those syllables consist of a consonant and a high vowel acoustically (Li K. P.

1964:41). Thus, we consider that [j] and [%] respectively follow the dentalveolar sibilants and the retroflexes. This solution also means that those syllables are not vowelless syllables, and thereby allows the general principle to be satisfied that the syllables must have a vowel.

[i], [u], and [y] are often considered single vowels, and these vowels are taken in the form of /i/, /u/, and/y/ as underlying vowels in generative studies of MP phonology

(see Sections 3.3.0 and 3.3.1). On the contrary, we consider the [i], [u], and [y] which are not followed by another vowel to be sequences of two phonemes which are realized as single segments at the surface phonetic level.

To be specific, [i], [u], and [y] are phonetic realizations of the phonemic compounds /ji/, /wi/, and /jwi/, respectively (Hsueh 1980a and 1986). The treatment of [i],

[u], and [y] as two sequential phonemes is necessary for the following two reasons (ibid.):

First, the four-way classification (i.e., kai-kou, qi-chi, he-kou, and cuo-kou) and the two dichotomies (i.e., kaikou-hu vs. hekou-hu and hongyin vs. xi-yin) are applicable to all Chinese syllables (see Section 3.3.1). 73

Therefore, even when a final is said to be made up of only

[i], [u], or [y], the final must still contain an on-glide responsible for its classification.

Second, such words as ([i]=/ji/), f ([fi]=/sri/), and $ ( [tcpy]=/k jwi/) do rhyme together in folk literature and in the 'Thirteen Tracks' ( -j- shi-san-zhe) tradition, and thus imply that they share by definition the same rhyme base."

In sum, we posit the unrounded high central vowel /i/ as the only underlying high vowel. The basic phonetic value of /i/ is [ut] (Hsueh 1986:63). When /i/ constitutes a whole final after a dentalveolar consonant or a retroflex, its phonetic value is [i] or [%], respectively. When /i/ follows the medial /j/, /w/, and /jw/, and precedes /n/ or the zero ending, it becomes, as a result of assimilation,

[i], [u], and [y], respectively. In all other cases, we consider [a] to be the phonetic value of /i/.

The vowels of MP show a three-way contrast in height.

We posit three vowel phonemes, one for each of the three levels of vowel height; namely, the high vowel phoneme /±/, the mid vowel phoneme /9/, and the low vowel phoneme /a/.^*

Syllables with a zero ending, e.g., ^ [i] /ji#/, ^ [je] /j

#/, and ^ [ja#] /ja^/, show the three-way contrast in vowel height; while a two-way contrast between the high vowel and the low vowel occurs in syllables with a non-zero ending. 74 e.g., [tqin] /kjin/ and [tcien] /kjan/ (Hsueh

1980a:501).

When /9/ alone constitutes a final, it is phonetically realized as [f], e.g., ^ [kf], ^ [tgV],

ê [syj, etc. Thus, we take [%r] as the basic phonetic form of /d/. /0 / can also be phonetically realized as [e] or [o] due to assimilation. For example, /ja/ [je],

^ /w9/ [wo], etc.

The basic phonetic value of the low vowel phoneme /a/ is [a], e.g., /ta/ [ta], /a/ can, also due to assimilation, be pronounced as [a] or [£], e.g., ^ /laj/

[lai], f /jan/ [j n ] .

We have so far discussed the three vowel phonemes, i.e., /i, 9 , a/, and their phonetic realizations with respect to allophonic distributions. We have also dealt with the medials and the endings in the preceding two sections. Therefore, we are now in a position to examine the whole system of finals of MP. 75

Table 7 - The phonemic representations of the MP finals

' ! Endind a 1 w i " 9 Vowel i a a i a i a ! i a i a medial

— i a a : ii ai iw aw in an i^ ai)

— i- 1i 1a fiai) ' iiw law 1 iin ian iin• J ian • ^

-w- wi wa wa ' wi1 wai I win wan wij wag. 1 1 i -iw- iwi iwa liwin iwan iwiç

SOURCE: 1986:78

Table 8 - The MP finals in a phonetic transcription

Endino 0 i w n n vowel i a a i a i a i a i ■' a Medial 1

-o- zr a 81 al 8 n an 3p ao 1 (]3o) -i- i ie iu iûO in # n in 1*9

A A —w — u WOWA W8l wai w8n wan w#n wùn i

(yô) -iw- . Y _ 4 e _ vn ycen iu^ 1

SOURCE: cf. Hsueh 1986:71 76

3.4 R-suffixation

As we all know, there is little morphophonemic variation in Chinese word formation.But in MP there is a morphophonemic process, which is called retroflex suffixation (hereafter r-suffixation). R-suffixation is one of the most characteristic features of MP when compared with other dialects or stages of the Chinese language.^®

The suffix "r" is derived from such words as (li

'therein' e.g., , 9 (ri 'day' e.g., ^ /L ) / and

(er 'child' e.g., ^ ) (Chao 1948:46). The suffix "r"

"does not occur after a clearly definable class of forms, but occurs rather by choice or convention" (Hsueh 1980a:

505). However, the suffix "r" is mostly attached to nouns, and in this case it often adds a diminutive connotation to the meaning of a root morpheme. Thus, the suffix "r" is generally known as a diminutive noun suffix, and represented by the character ^ (or J(^).

Hsueh (1975:91, 1980a, and 1986) explains the emergence of the suffix "r" as follows;

At the time of the Zhonavuan vinvun (1342), words like ^ , jS , n. had the reading /ri/ (plus tone), but by Xu Xiao's ( ^ ) time (1606), a metathetical change must have rendered this syllable into /ir/, and this unique enduing thereby came into existence. ... When the word that had this ending became a diminutive noun suffix, it lost its vowel through phonetic fusion. Consequently, a large number of derived syllables with /r/ as ending appeared. (1980a:504) 77

Hartman (1944) transcribes ^ /U ('song') and -f:^

('root') as [kaar] /keer/ and [kàr] /ker/, respectively, and thus attributes the distinction between the two syllables to the difference in the lengths of the vowels in question

(ibid., pp. 33-36).^^ But this solution is ad hoc since he allows length for one vowel only, but does not for the others in MP. On the contrary, we believe, following Hsueh

(1980a), that the distinction between the two syllables lies rather in the vowel height. ^ [ky] and ;ji^[kan] are phonemicized as /ka/ and /kin/ respectively according to our vowel system of MP (see Section 3.3.3); thus, /Ü should be /kar/, while ^ J L (/kinr/) should be /kijr/, due to the changes of syllable ending /n/ to /j/ before the suffix "r"

(see below). Likewise, the syllable ^ [âr], which is in cl&ar contrast with /ar/, will be identified as one with a high vowel."* Therefore, /ir/ and [ar] respectively are the phonemic and the phonetic transcriptions of the citation of the suffix "r ". This solution makes it possible to describe the morphophonemic changes caused by r-suffixion in terms of phonological rules.

When the suffix "r" is added to a root morpheme, the vowel of the suffix "r" is deleted, and the syllable for the root morpheme and the /-r/ form a r-suffixed syllable. Chao

(1968:51) points out that among the native speakers of MP, the "old generation" and the "young generation" pronounce 78

r-suffixed syllables differently. Hsueh (1980a and 19 86)

states that there are three versions of r-suffixion: two

versions among the "old generation" and one version among

the "young generation". He describes the three versions of

morphophonemic changes in terms of phonological rules.

The first version of r-suffixion, can be described by

the following two rules:

(1 ) ^ ^ j (M)i _ r (M ^ w)

(2 ) n j / __ r

The second version of r-suffixion includes a new rule as

follows in addition to the above two rules:

(3) jzi j / a __ r

In the third version, the following rule occurs in

addition to the above three rules.

(4) ij 9 / j(w) __ r (restricted to syllables of the 3rd and the 4th tones)

3.5. The tones of Modern Pekingese

The tones in Chinese are phonemic, i.e., they

contrast lexically in a given phonological environment. In

MP there are four basic tones and a neutral tone.^° The

four tones are numbered as 1st, 2nd, 3rd, and 4th tones, and

are represented by the symbols ( which indicate the shapes of the pitch contours): — , j , ^ , respectively. The

four tones of MP are usually described in a five-point scale as follows: 79

Table 9 - The four tones of MP

Tone Chinese name Description Pitch Graph

1st tone FÉ f inping high-level 55 1

2nd tone P# f- yangping high-rising 35 / 1 3rd tone Jz. ^ shangsheng low-dipping 214 \/ 4 th tone ^ qusheng high-falling 51 N SOURCE : Chao 1968:26

When syllables are juxtaposed, certain changes of tones occur. This phenomenon is called tone sandhi. Tone sandhi is, on the whole, regressive assimilation, i.e., the tone of a syllable is affected by the tone of the following syllable. In MP there are tone sandhi rules as follows;

(1) When a third tone syllable is followed by any tone except another third tone syllable, the third tone changes into the half-third tone, which has the pitch contour 21.

214 21 / __ T (T = any tone except the third) E.g., ma che ma che ^ ^ 'horse vehicle'

(2) When a third tone syllable is followed by another third tone syllable, the third tone of the preceding syllable changes into the second tone.

214 35 / __ 214

E.g., qi ma qi ma ^ 'at least'

(3) When a second tone syllable is preceded by a syllable with either the first tone or the second tone, and 80 is followed by a syllable with any one of the four tones; it changes into the first tone.

35 55 /| 55j __ T (T=any tone except neutral)

E.g., san nian ji san nian ji H. ^ 'the third grade'

(4) When two syllables with the fourth tone occur in sequence, the fourth tone of the preceding syllable takes on the pitch contour 53.

51 -» 53 / 51

E.g., da shu — » da shu 'Mj ' big tree '

As shown above, the most prominent tone sandhi rules involve the third tone. The third tone has two tonemic variants. One has two allotones: the original pitch contour

214 and the pitch contour 21 derived by tone rule (1). The other has the pitch contour 35 derived by tone rule (2), and this pitch contour is not distinguishable from that of the second tone; thus, tone rule (2 ) is morpho-phonemic in nature (Chao 1968:27 note 22 and Cheng 1973:43).

The domain of the tone in Chinese has been a debated problem. The debate centers around the question whether the tone belongs to a segmental unit or a syllable. Chao's

(1968) view on this question is as follows:

... a tone, which is primarily the pitch pattern of the voiced part of the syllable, so that, if the initial is voiced, the tone begins with the initial and spreads over the whole syllable, while, if the initial is voiceless, the tone is spread over the final only. (p. 19) (Underlining mine) 81

On the other hand, Hsueh (1980a) affirms that the

tone goes with the vowel for the reason that "a syllable by

definition must have a vowel", and elaborates as follows:

... a toneme can also be extended phonetically beyond a vowel to cover other compatible elements in the syllable, and this fact should not be allowed to confuse its phonemic status as a supraseomental element of the vowel. (p. 496) (Underlining mine)

Chao's (ibid.) view is that the tone spreads over the voiced part of the syllable phonetically. Hsueh (ibid.) assigns the domain of the tone to the vowel at the phonemic

level, yet he admits that the tone can spread beyond the vowel at the phonetic level. Thus, it appears that the two views are congruent with each other at the phonetic level.

William Wang (1967:35)'also maintains that the domain of the tone is over the entire voiced part of the syllable phonetically. In sum, the tone spreads over the voiced part of the syllable phonetically, but it is considered to belong to the vowel phonemically.

Unlike tone, stress is not phonemic in Chinese. The four basic tones of MP occur mostly in stressed syllables, while the neutral tone occurs mostly in unstressed syllables. This fact brings about an overgeneralization that, "when a syllable is stressed, it has a tone, but when it is unstressed, its tone becomes neutral" (Cheng C. C.

1973:57). However, Chen C, Y. (1984) refutes this view aptly by pointing out the following two phenomena; First, there are weak-stressed syllables which can trigger tone sandhi. 82

Second/ there are also certain grammatical morphemes which lack the tone at both the underlying and the surface levels.

Therefore, weak stress and the neutral tone do not always interact as cause and effect.

We have thus far discussed the characteristics of MP phonology, and thereby have explained our theoretical stance on the important aspects of MP phonology. We shall take the theoretical framework of MP phonology presented in this chapter as a firm basis of comparison and as a point of reference in analysis of MM phonology in the ensuing chapters. 83

CHAPTER III

NOTES

1. The distinction between a morpheme and a word is elusive in Chinese since most of the morphemes make words by themselves. When there is a need to differentiate a compound word from its component morphemes, I shall call the former a word and the latter morphemes.

2 . .Due to the presence of the "zero" initial, in mian'ao ( 'wadded jacket') the n does not link with ao, as would link English "ran out". See Chao 1968:20.

3. The term "guttural— a translation of the Chinese term houyin ( ^ 'throat sounds')— was first used by Karlgren, and is now well-known in the field of Chinese phonology.

4. For further discussion of this matter, see Chao 1934:384-385, Cheng 1973:36-40, and Hsueh 1986:32-34.

5. For examples, see Cheng C. C. 1973:39 and Hsueh 1986:33.

6 . This section benefits greatly from Hsueh 1980a and Chapters 2 and 5 of Hsueh 1986. Among various analyses of the MP finals that I have consulted, I believe Hsueh's analysis is the most sound; for his analysis is viable in both traditional Chinese and modern Western phonological frameworks, and also is satisfactory for both synchronic and diachronic descriptions.

7. is the lip-rounded form of [j] articulatorily, and a single segment acoustically." (Li K. P. 1964::41) Ladefoged (1975) specifies the places of articulation of the semivowels as follows: Because a semivowel is a kind of approximant, it can be considered to have a particular place of articulation just like any consonant. ... [j] is a palatal approximant, and [w] is a labial-velar approximant. The semivowel [i|] is a labial-palatal approximant. (p. 206)

8 . In our system of MP finals, however, the symbols i, u, and y are used to represent phonetic fusions of a semivowel and a main vowel, but they are not used for phonemic representation. See Section 2.3.3.

9. I think, at any rate. Lu Z. J. (1985) should have given some qualification as to what he means by [vowel] in order to give any meaning to this rule. 84

10. See Li K. P. 1964:42-45. Here I replace Li's transcriptions /ai, ei, iau, uai/ by [al, ei, jaO, wac].

11. See The Principles of the International Phonetic Association. 1949 edition (reprinted 1963), p. 8 .

12. The symbols [i] and [%] were first used by Karlgren, and are now used generally in studies of Chinese phonology.

13. The 'Thirteen Tracks' is a rhyme chart for folk songs, which has been popular around Peking since the time of the ending of the Ming dynasty or the beginning of the Qing dynasty. Therefore, the 'Thirteen Tracks' represents the native speakers' feeling of rhyming. See Hsueh 1986:16-17. The word ^ ([u]=/wi/) forms a separate rhyme from the words f , and ^ in the "Thirteen Tracks" notwithstanding the former has the same rhyme base /i#/ as the latter. As for the explanation of this phenomenon, see Hsueh 1980a:503.

14. The establishment of the three phonemic vowels takes vowel height to be predominant over phonetic similarity in phonemicizing the MP vowels. However, it is affirmed that the three phonemic vowels— denoted by three central vowel symbols— represent the psychological reality of the native speakers of Mandarin. As for the justification for this solution and the means for arriving at it, see Hartman 1944, Hsueh 1980a, and Light 1976:20-32.

15. Walton (1983) argues that the Chinese sound pattern must be analyzed in terms of surface phonetic structure rather than deep phonological structure on account of the relative lack of morphophonemic alternation in Chinese.

16. There are two noticeable morphophonemic processes in M P : the one is tone sandhi rule (2) (see Section 2.5) and the other is r-suffixion.

17. The fact that there is a distinction between the syllables ^ // and is also attested by Chao 1968:51-52.

18. On the contrary, in previous studies of MP phonology the vowel of the syllable ^ has often been mistakenly considered to be a mid vowel, e.g., Hartman (1944), Robert Cheng (1966), and Hashimoto (1970).

19. For examples and detailed discussion of these rules, see Hsueh 1980a:504-510 and 1986:84-85. CHAPTER IV

THE INITIAL SYSTEM OF MIDDLE MANDARIN

In this chapter, we shall reconstruct the initial consonants of Middle Mandarin (MM) as represented by the PVP notation of PN and PP. We shall examine the PVP notations mainly in comparison with the VP notations, and occasionally with reference to other related materials as well as our knowledge of the phonology of Modern Pekingese (MP).

However, we shall first introduce the letters of Hunmin chong'um . or the Korean alphabet, and the method of transcription in Hunmin chong'um of Chinese pronunciation at the beginning of this chapter before we begin to examine the pronunciation notations of the initials in this chapter and the finals in the following chapter.

4.1 The letters of Hunmin chong'um

Hunmin chong'um is composed of twenty-eight basic letters; seventeen consonants and eleven vowels. The seventeen basic consonant letters and their sound values can be presented as follows:

85 86

Table 10 - The 17 basic consonants of Hunmin chong'um

Manner Stops Frica­ Nasals Approxi­ & Affricates tives mant s -asp +asp

Place

Labials 8 [p] J l [ph] J [m]

Alveolars C [t] c [th] L [n] E [1 ]

Alveolar 7\ [ts] K [tsh] A [s] A [zl sibilants

Velars 1 [k] 1 [kh] [X] o' Û Glottals s r?i [fi]/zero

NOTE: Ô [•?] was not used for writing Korean, but only for transcribing Chinese pronunciation and writing Sino- Korean pronunciation (Lee 1972:27). The underlined letters became obsolete, and are not used for writing Modern Korean. Consequently, at the present time, the letter "o " stands for zero when it occurs in the initial position, and for [t|] when it occurs as an ending.

The explanation of the shapes of consonant letters is

found in HCH 1.2.0. One basic shape was adopted for each of the seven articulatory categories of consonants used in

Chinese rhyme tables (see Chapter 1 Fn. 16) in the

following manner:

~I [k] depicts the outline of the tongue blocking the throat. L [n] depicts the outline of the tongue touching the upper palate. □ [m] depicts the outline of the mouth. A [s] depicts the outline of the incisor 0 [fi] depicts the outline of [a cross section] of the throat. ... 2, [1] and A [z] likewise depict the outline of the tongue and incisor. (Translation cited from Ledyard 1966:229).

The letter for an aspirated stop or affricate is formed by adding a stroke to the basic letter shape for its category 87 of consonants. The letters for double consonants or consonant clusters are made by reduplicating a basic letter shape or combining basic letter shapes.^

The consonants of Middle Korean (12th - 16th cent.) of the fifteenth century can be presented as follows:

Table 11 - The consonant system of 15th century Korean

Manner NasalsObstruents Approxi- -asp +asp +tense mants

Place

Bilabials ti /p/ % /ph/ BH /p'/ 0 /m/

Alveolars C /t/ c /th/ cc/t'/ i- /n/ /I/

Alveolar 7\ /c/ A /ch/ M /c'/ sibilants (ts) (tsh) (ts')

A /s/ /V\ /s'/ A /z/

Velars 7 /k/ 7 /kh/ 77 /k'/ 6 /%/

o /x/ d o /X'/

Glottals /fi, zero/

SOURCE ! cf. Ho 1965:370 and Ledyard 1966:185a.

The eleven basic vowel letters consist of seven monophthongs and four diphthongs as follows:

Monophthongs: ^ [A], _ [ i ] / | [i], _L [o], [a], T [u], ./ [a].

Diphthongs: [jo], /. [ja], 77 [ju], :| [j5].

([A] merged into [a] before the mid-eighteenth century, and thus the letter " ' " is not used for writing Modern Korean.) 88

The vowel system of Middle Korean of the fifteenth

century can be schematized as follows:

Table 12 - The Korean vowel system of the 15th century

front central back

-round -round +round

high _ 1 < u

mid high -i a

mid-low

low

SOURCE: of. Moon Y-S. 1974: 179

The explanation of the shapes of vowel letters is

found in HCH 1.3. Three basic shapes were formed in the

following manner:

' [A]î the roundedness of the outline is a depiction of Heaven. __ [i]: the flatness of the outline is a depiction of Earth. I [i]: the uprightness of the outline is a depiction of Man. (Translation cited from Ledyard 1966:234-235)

The other eight basic vowel letters were created by reusing the three basic shapes in the following manner:

(1) Monophthongs: J_[o], /«[a], T[u], and »/ [a] are formed by attaching » to either _ or / . 89

(2) Diphthongs: ^ [jo], |; [ja], r: [ju], and :j [ja] are formed by adding another « to , j-! » , . , respectively.

Graphically, when * is added to , it takes the form of a short vertical line; when ' is added to | , it takes the form of a short horizontal line. Thus, the vowel letters in the above are written as j_ , j- , — , -j , jL f , "jT ,

^ , respectively. More vowel symbols for diphthongs and triphthongs are created by combining some of the eleven basic vowels.

In addition. Middle Korean of the fifteenth century has two semivowels, which reuse basic vowel letters, and more diphthongs and triphthongs are formed as follows:

Semivowel: I /j/ J - r T /w/

Diphthongs: 'I/aj/, J /ij/, ij/oj/, H /aj/, rf/uj/,

/e j/,

/wa/, r-j /wê/

Triphthongs: jif/joj/, /=| /jaj/, ,Tj / juj/, /jaj/,

ij^/waj/, rt//w9j/

In terms of vowel quality, the low vowels [/\], [a], and [o] are treated as positive; the high vowels [i], [a], and [u] as negative; and the only high front vowel [i] as neutral (see Table 12). This classification of vowels plays a decisive role in description of the vowel harmony of

Middle Korean and in formation of the shapes of vowel 90

letters. In Middle Korean vowel harmony, positive vowels

(or low vowels) agree with positive vowels only, whereas

negative vowels (or high non-front vowels) agree with

negative vowels only. Graphically, a positive vowel can be

combined with another positive vowel or a neutral vowel, but

not with a negative vowel in creating vowel letters, and

vice versa, when » is put above , the vowel thus

formed j_ [o] is positive; whereas when » is put below _

, the vowel thus formed T [u] is negative. When \ appears

on the right side of j , the vowel thus foirmed f- [a] is

positive; whereas when » appears on the left side of / ,

the vowel thus formed [a] is negative.

4.2 The method of pronunciation notation

A Korean syllable consists of three parts: the

initial, the vowel, and the ending.^ The structure of the

Korean syllable can be delineated by the following formula:

#(C)V(E)#

The three letters in the above formula respectively represent the initial consonant, the vowel, and the ending.

The initial may be a consonant or zero. The vowel may be a monophthong, a diphthong, or a triphthong. The ending may be a consonant or zero. The vowel is the essential element

for the formation of a syllable, while the initial and the ending are optional elements. 91

The Korean language is abundant in polysyllabic words. In addition, even monosyllabic words become polysyllabic when suffixes are added to them. However, unlike other polysyllabic languages, Korean is written syllabically, i.e., a syllable is the basic unit in Korean writing, even though it has an alphabetic writing system.^

Thus, a Chinese character, which is read as a syllable, is also transcribed as a Korean syllable. On the other hand, due to the structural difference between the Chinese and

Korean syllables (of. Section 3.1), the following rules of correspondence obtain in the transcription of a Chinese syllable with the Korean alphabet and in the deciphering of such a transcription:

(1) There is a one-to-one correspondence between the

Korean initial and the Chinese initial.

(2) When a Chinese syllable ends in a semivowel or zero, the Korean vowel describes the phonetic realization of the Chinese final (i.e., the medial and the nucleus, and the ending).

(3) When a Chinese syllable has a consonant ending; the

Korean vowel describes the phonetic realization of the

Chinese medial and nucleus, and the Korean consonant ending corresponds to the Chinese consonant ending.

We are now ready to explain how the Hunmin chona'um letters are used for the transcription of Mandarin pronunciation. I shall list the Hunmin chona'um letters 92 used for pronunciation notations in HMCUYH, with a view to using them as a point of reference in our examination of the

PVP notations in PN and PP, for the following reasons;

First, HMCUYH is the first book in which Chinese character pronunciations are transcribed in the Korean alphabet, and as such it exemplifies the method of Korean transcription of Chinese pronunciation for all subsequent pronunciation notations.

Second, the number of distinctions made in HMCUYH is overall greater than those in the PVP notations of PN and PP and their revisions (see Section 2.2.2).

I shall classify the Hunmin chona'um letters used for the pronunciation notations in HMCUYH according to the three parts of the Korean syllable.

The thirty-one representative characters for the initials of HWZY, their corresponding Korean initial consonants of the CP and VP notations in HMCUYH, and their reconstructed sound values are arranged in Table 13. Table

13 includes the following arrangements:

(1) The traditional quan-zhuo ( ^ 'muddy') initials are reconstructed as murmured sounds or voiced aspirated stops.* There is no IPA diacritic for murmur. Ladefoged

(1975:127) uses the symbol [b^] to denote the murmured bilabial stop. In studies of Chinese historical phonology, following Karlgren, a murmured sound is conventionally denoted by a voiced stop followed by " f" (the symbol for 93 weak aspiration), e.g., g'-, d'-, etc. In this study, however, we shall use a voiceless stop followed by "H", e.g., pH, tH, etc., to denote a murmured sound.^

(2) The initial consonants in parentheses are considered to have existed in the sound system underlying

HWZY, although they were on the verge of disappearing (Chou

1989:101-106 and 134-141).

Table 13 - The initial consonants of HWZY

Manner Obstruents Nasals Approxi- -voice +voice -voice +voice mants -asp +asp Place % $ Labials d P JL ph ( UU pH) Ü m ? V

-/M. Oentalveolars c t c th ( CC tH) L. n £ 1

# ' t /C' Dentalveolar /; ts ^ tsh ( /9^ tsH) /' s ( AÂ sH) sibilants ag # $ 3 Retroflex tg ^ tgh ( tpH) (/Y\ pH) ^ ? sibilants

1 0 s. Is. Velars 7 k 7 kh (77 kH) 0 X ( ot XH) (à (] )

Glottal (Ô ? ) 0 0 94

The following is a list of the Korean vowel letters

used in HMCUYH and their sound values in IPA:

Table 14 - The Korean vowel letters used in HMCUYH

I i _ i T u -j 9 j_o j - a

TT ju (ja) je 1= ja T-j wa -Lj- wa

-I ij rj Uj H aj J-fj waj

7T| juj ^1 ( j9j) jej jaj

Tpj(juje) jue

SOURCE: cf. Kang S-H. 1973:11. NOTE: The vowel letter " " is a combined form of " " [ja] and " I " [j], and the vowel letter " iN " is a combined form of " 7T " [ju] and " " [ja]. They are expected to denote [jaj] and [juj9], respectively. However, " ^ " [ja] is used to denote [je] /je/ in Mandarin, and in addition [juja] does not fit the syllable patterns of Chinese and Korean. Therefore, we transcribe " " as [jej], and " ird " as [jue] in IPA.

The Korean endings used in HMCUYH and their corresponding Chinese endings are as follows:

Table 15 - The Korean endings used in HMCUYH

tl p a m a , ? w

C t L n ^ (i)^

1 k

Ô ? 95

(1) The endings of the entering tone words are denoted

by a [p], C [t], and 7 [k] in the CP notation; but they

are denoted by o’ [?] only in the VP notation except for

the ending of the HWZY Yao ^ rhyme, which is marked by &

[w].

(2) In the CP notation, ? [w] denotes the ending of the Xiao ^ , Yao , You rhymes.

(3) The consonant letter " a " denotes the retroflex approximant initial /^/ in Mandarin when it occurs in the syllable initial position (see Table 13 in this section).

However, in the VP notation, " a " is put in the ending position of the Zhi ^ rhyme only, which has the unrounded high central vowel /i/ in Mandarin. In this case, it is equivalent to the zero ending, but it shows that the preceding nucleus is an apical vowel, i.e., [%] or [%]. 96 4.3 The initial consonants of Middle Mandarin

The initial system of MM is similar to those of ZYYY and MP, but noticeably different from those of LMC and HWZY.

Thus, we shall account for the characteristics of the initial system of MM in comparison with those of LMC and

HWZY. The initial consonants of MM as represented by the

PVP notation of PN and PP are presented in a broad phonetic transcription in the following table:

Table 16 - The initial consonants of MM

Manner Stops & Affricates Frica- Nasals Approxi- -asp +asp tives mants

Place

Labials W p Kph ^ f Om Sv

Oentalveolars C t 2 th L n 2 1

Dentalveolar ts /r tsh / s sibilants

Retroflexes ^ tg /\ tgh A g A % (or r)

t; t;h ; j

Gutturals ~Tk 1 kh o' X o j6

4.3.1 Former voiced obstruent initials

The traditional 'muddy' ( quan-zhuo) initials of EMC are reconstructed as voiced sounds, and those of LMC are reconstructed as murmured sounds.* The murmured 97 obstruent (i.e., stop, affricate, and fricative) initials of

LMC were all devoiced before OM. In the syllables of the MC

level tone with murmured stop or affricate initials, the initials merged with their voiceless aspirated counterparts, and the tone concomitantly became the Yang-level tone. In the syllables of the MC oblique (i.e., rising, going, and entering) tones with murmured stop or affricate initials, the initials merged with their voiceless unaspirated counterparts. This process can be formulated as follows

(cf. Hsueh 1975:37);

H -->^h/ [+stop ] (1) I elsewhere

NOTE: denotes the finals of the Yang-level tone syllables derived from the MC level tone syllables.

The following statement is found in the preface to HMCUYH:?

i ÿ , ^ f -'A è-;t, iL ht K i» . "As for the words with muddy initials, [the muddy initials of] the level tone [words] are similar to voiceless aspirated sounds, [those of] the rising, going, and entering tone words are similar to the voiceless unaspirated sounds, this is how they are used popularly."

Notwithstanding such descriptions, the murmured obstruent initials of HWZY— represented by the initial characters

^ /pH/, ^ /fH/, 'E/tnl, ^/tsH/, ^>/sH/, ^/tsH/, 98

/sH/, /kH/, j£. /xH/— are kept and transcribed in the

CP and VP notations.

Ch'oe Se-jin describes the process of devoicing and

tonal split as follows in the section about the

"Distinctions of Voicing and Tone" in PNPP:

(1) The muddy initials of the level tone syllables became voiceless aspirated sounds, and the tone became the Yang-level tone.

(2) The muddy initials of the rising, going, and entering tone syllables became voiceless unaspirated sounds.

The tones are not changed except for the rising tone words, which sound like the going tone words with voiceless unaspirated sounds.

The tonal changes described in the above can be formulated as follows:

F°^ / H,R ( / elsewhere (2)

04 I jj prl«ing { f “ / elsewhere (3) F°\ F°^, F°^, and F°* respectively stand for the finals with the 1st, 2nd, 3rd, and 4th tones derived from the MC non­ entering tones. "H" and "R" respectively stand for the murmured initials and the resonants (nasals and approximants). 99

The following table illustrates how the muddy initials are transcribed in the VP and PVP notations in PN and PP.

Table 17 - Transcription of the former voiced initials initials jr 2# BE

aa yj cc /y^ ^ w tt oo VP pH fH tH tsH sH tgH ph kH xH

Level g c A A ^ A 9 ^ tone ph f th tsh s tgh p kh x PVP

Oblique tones pfttsstppkx

NOTE: VP and PVP respectively stand for the "vulgar" pronunciation notation and the "present vulgar" pronunciation notation. The initials of the VP notation represent the initial system of HWZY, while those of the PVP notation represents the initial system of early sixteenth Northern Mandarin.

The above table demonstrates the following changes in transcription of the former voiced obstruent initials from the VP notation to the PVP notation:

(1) In the VP notation, the muddy stop or affricate initials ( jk , f » and ^ ) are respectively transcribed as [pH, tH, tsH, tpH, and kH]. In the PVP notation, however, they are respectively transcribed as [ph, th, tsh, tgh, and kh] in words of the level tone, and as [p, t, ts, tg, and k] in words of the oblique tones. 100

(2) In the VP notation, the muddy fricative initials

( ^ f f f^ / and {£ ) are respectively transcribed as

[fH, sH, gH, xH]. In the PVP notation, they are

respectively transcribed as [f, s, g, x], irrespective of

tonal differences.

4.3.2 Labials

The labial ( ^ chun-yin 'lip sounds') initials

p, ph, pH, !î/j m/ of EMC split into two series of

LMC, i.e., the bilabial ( ^ ^ zhong-chun-yin 'heavy

lip sounds') initials / p, ph, ^ pH, 9^ m/ and the

labiodental ( ^ ^ qing-chun-yin 'light lip sounds')

initials / f, ^ fh, fH, f;^v/ of LMC. The labial

initials became labiodental before the labio-palatal medial

/-jw-/ first,® and then the feature palatal /-j-/ was lost after the labiodentals. This process can be formulated as

follows (of. Hsueh:1975:39 and Chou 1989:18):

B Bf/ ___ jw (4)

j — » Bf ___ w (5)

Rule (4) precedes rule (5). B: labial initials. Bf: labiodental initials.

The bifurcation of the labial initials had no sooner occurred than the merge of /f/ and ^ /fh/; since phonetically distinction between plain and aspirated

fricatives is hardly plausible, or it could have been the 101 case that the opposition between /f/ and /fh/ was artificial from the beginning (Wang Li 1980:115). Later,

^ /fH/ also merged into /f/ through devoicing. Thus,

/ fl= f, ^ fh, ^ fH/ have merged completely before OM, and only two labiodental initials, namely, /f / and /v/ remained in CM.’

The bilabial stop initials / ^ p, ph, ^ pH, 3^ m/ were transcribed as [ Ü p, 3l ph, B j j pH, m ] . The letters for the labiodental initials were created by putting the diacritic "g" under the Hunmin chona'um letters for the corresponding bilabial initials. In the CP and VP notations, the three labiodental fricatives / # f, ^ fH,

v/ appear, and they are respectively transcribed as

[ % f, “ÿ fH, ^ V]. In the PVP notation, only two labiodental initials / f , v/ appear, and they are respectively transcribed as [ W f, ? v]. The following table shows what labial initials occur and how they are transcribed in PN and PP.

Table 18 - The labial initials in PN and PP

Initials t K VP H p X ph HU pH cr m If fH f V Initials $

PVP tl p JC ph 0 m ^ f ? V 102

The initial Wei is transcribed as follows in the

VP, PVP, and PVP[1] notations;

Table 19 - Transcription of the initial Wei

Characters wei wu wu ÀT wen M wen

a VP [ vwi ] ■f [vu] t- [vu] ê [vin] j>— [vin]

a PVP f) [wi] T [u] f [vu] 4- [vin] -t [vin]

tl PVP[1] -r [U] f [un] f [un]

PVP[1]: The PVP notation revised in PN (1670) and PP (1677).

We can infer from the above changes in transcription that

the pronunciation of the initial Wei underwent the

following changes:

(1) The Korean letter for the initial /v-/ " # " is a

combined form of the latter " Û " [m] and the diacritic for

fricatives "o "• In the PVP notation, [v-] and the zero

initial are mixed up.

(2) Phonetically, the initial Wei is [v-] in the VP

notation; [v-] or zero in the PVP notation; and zero in the

PVP[1] notation. That is, the initial /v-/ was in the

process of merging into the zero initial.

Therefore, the Wei initial was probably pronounced as or

close to the labial-velar approximant [w] in early sixteenth

Northern Mandarin.

V — 5^ / # ___ ( 6 ) 103

4.3.3 Alveolar and retroflex sibilants

4.3.3.1 The three palatal series of EMC, namely, the palatalized alveolar stops (Dy) f^oty, tyh, tyH/, the palatal sibilants (Py) / ^ c, § cy, cyh, ^ sy, syH/, and the retroflex sibilants (Pr[l]) / if cr, ^ crh, ^ crH,

^ sr, srH/ gradually coalesced from EMC to OM, through

LMC, and finally merged into the retroflex sibilants of

Mandarin (Pr[3]) /cr, crh, sr, r/ by the time of ZYYY

(1324).^ (Here, "y", "h", "H", "c", and "r" respectively stand for [+palatal], [+aspirated], voiced aspirated (or murmured), the affricate [ts], and [+retroflex]. When

[+palatal] is a feature of an initial consonant, it is conventionally denoted by "j" in studies of Chinese historical phonology. However, "y" denotes the feature

[+palatal] of an initial here because in this study we use the symbol "j" to represent the feature [+palatal] of a medial or the palatal medial itself (see Section 3.3.1).)

In the rhyme tables, the Py initials occur before the finals of division 3, while the Dy and Pr[l] initials occur before the finals of both divisions 2 and 3. The process of this great coalescence can be summarized as follows (Chou

1989:19-20):

(1) The Pr[l] initials before the finals of division 3 lost their medial /-j-/ (/-y-/ in her transcription) before 104

LMC. Then, the Pr[l] and the Py initials lost contrast, and formed the new retroflex series of LMC, namely, (Pr[2])

/ cr, ^ crh, ^ crH, sr, srH/.

(2) Towards the end of LMC, the Dy initials were affricated, and merged into Pr[2]. Thus, the Dy and Pr[2] initials formed the retroflex series of Mandarin (Pr[3])

/cr, crh, sr, r / .

This two-step development can be formulated as follows (cf.

Hsueh 1975:43 and Chou 1989:20):

Step One: Pr[l], Py > Pr[2] Dy > Dr

j-*çrf/Pr[l] (7)

y r/ (8)

Step Two: Pr[2], Dr > Pr[3]

t c/ r ( 9 )

In respect of the emergence of the retroflex initials of Northern Mandarin, Paragraph 3 of the "introduction" to

SSTG describes the merge of the palatalized alveolar stops

( ^ Ji. she-shang 'tongue-up') into the retroflex sibilants

( I E . ^ zheng-chi 'upright incisor') as follows

H ^ Si i 1 ^ 1 .^ . at#r 105

"In general, the palatalized alveolar stops (Dy series) are

[made when] the blade of the tongue touches the palate; thus, they are difficult to pronounce, and naturally merge into the retroflex sibilants (Pr[2]). Therefore, in GJYHJY

(1297) / ty, f^tyh, tyH, ny/ are assigned to / f;?, cr, ^ crh, crH, sr/; /!^uy/ alone merged into

in the current pronunciation in China, thus and /_^ny/ are merged, and are not separated in this book

(HMCUYH). At present ty, tyh,7^tyH/ (Dy) are combined with / cr, ^ crh, ^ crH/ (Pr[2j), and ny/ is combined with / ^ n/."

This description proves that the palatalized alveolar stops

(Dy series) merged into the retroflex sibilants (Pr[2] series) completely no later than the time of GJYHJY (1297).

That is, the retroflex sibilants of Mandarin (Pr[3j) were formed before the fourteenth century.

4.3.3.2 It is generally accepted that the three palatal series of EMC merged into the retroflex series of Mandarin completely by the time of ZYYY. However, there is some disagreement as to the phonetic nature of the retroflex initials of OM and MM.^ Lu Z. W. (1946:43-44) asserts that

"ts + 1 " should not occur unless clear evidence existed.

Thus, he reconstructs [tqi] before high front vowels or glides, and [tp] elsewhere. It seems that Lu Z. W. regards

[tql and [tg] as the allophones of the initial /cr/ (in our 106

transcription) of ZYYY, though he does not employ the

concept of a phoneme explicitly. Similarly, Dong T. H.

(1968:61-62) states that it is unnatural that retroflex

sounds should occur before [i]. Thus, he reconstructs /tj,

tj', j'/ as the retroflex initials of ZYYY, and adds that

they sound like [t^, t^', çf] before [i], and [tg, tg', g]

elsewhere. It seems that he regards [t^] and [tg] as the

allophones of /tj/, though he does not use the term phoneme

or allophone.

4.3.3.3 Now, let us examine how the dentalveolar and retroflex sibilants are described in our source materials.

Paragraph 5 of the introduction to SSTG explains the transcription of the dentalveolar sibilants ( ^ chi-tou

'incisor head') and the retroflex sibilants ( jP ^ zheng- chi 'upright incisor') as follows:

^ ^ ^'1 f è , «C* f # A A

i ^ i>l ^ Jt B ^ ^ i ^'1 > 7 1 1 % h T ^ # â À ' >■'k^ t..

"As for the sibilants, the dentalveolar sibilants /s, c, ch/ are [made when] the tongue is raised and touches [the rear gum of] the front-teeth, and thus they sound light; the retroflex sibilants /sr, cr, crh/ are [made when] the tongue is curled and touches the palate, and thus they sound heavy. 107

Since the alveolar sibilants [in the language] of our

country [ /\ s, /\ ts, ^ tsh] are [produced] between [the

places of articulation] for the [Chinese] dentalveolar

sibilants and retroflex sibilants, there is no difference

between the dentalveolar sibilants and the retroflex

sibilants in Hunmin chona'um. In order to differentiate

them, / /' , ^ , f' I will denote the dentalveolar sibilants

/s, c, ch/, and i \ , \ t will denote the retroflex

sibilants /sr, cr, crh/."

In addition, the following description of the

Mandarin sibilants is found in the "notes" to HCH:

U X ‘i ‘■t K"- W 2 | - , aj ±.ZI ( £ . Û) I ±21 tfti 4 f I M

^ J. £| 'I' 'i '-h W z(- .

"For the dentalveolar sibilants (of Mandarin) sound lighter

than the (alveolar sibilant) sounds of the language (of our

country), they are made when the tip of the tongue touches the upper back of the upper incisors. For the retroflex

sibilants (of Mandarin) sound heavier than the (alveolar

sibilant) sounds of the language (of our country), they are made when the tip of the tongue touches the rear gum of the

lower incisors."

From the above descriptions we can infer the following

facts: 108 (1) At the time of SSTG (ca. 1446-1450) there were only two types of sibilants in Mandarin: the dentalveolar sibilants and the retroflex sibilants (Pr[3]).

(2) The sound values of the dentalveolar and retroflex sibilants of Northern Mandarin of the mid-fifteenth century must have been very similar to those of MP. However, the retroflex sibilants had two types of allophones: the palato- alveolar [tj] type and the retroflex [tg] type (cf. Kang

S. H. 1983).

4.3.3.4 The alveolar sibilants are the only series of sibilants phonemically in the Korean language, though they are palatalized before [i] phonetically. The Hunmin chong ' um letters [As, 7\ ts, A tsh] with both the left and right strokes of equal length denote the Korean alveolar sibilants. In the VP notation, the dentalveolar sibilants

/c, ch, s/ are represented by the letters with an elongated left stroke / ^ , À /, while the retroflex sibilants

/cr, crh, sr/ are represented by the same letters with an elongated right stroke / A , ^ , \ /. However, in the PVP notation, the dentalveolar and retroflex sibilants are not transcribed differently, for only the basic letters with both the left and right strokes of equal length [ A ts, A tsh, A s] are used to represent both of them. The question of why the two series of Mandarin sibilants are transcribed indiscriminately in the PVP 109 notation is open to debate. Here, I shall explain briefly why the change in transcription of the Mandarin sibilants occurred in the PVP notation. Although there were only two series of sibilants phonemically in OM and MM; there were probably three (or possibly four) types of sibilants phonetically in OM and MM. In the VP notation, which is conservative, the two Mandarin sibilants were differentiated. When Ch'oe Se-jin was writing PN and PP, he must have been able to differentiate the Mandarin sibilants.

At the same time, he as a linguist must also have understood that the distinctions of the MM sibilants were very difficult for Korean learners of spoken Mandarin to make.^

Facing such a dilemma, he simply used only the basic letters for the Korean alveolar sibilants for transcribing all the sibilants of MM for pedagogical expediency in the PVP notation. However, the two series of Mandarin sibilants could still be differentiated phonemically according to the

VP notation, which was added to all characters in PN and PP side by side with the corresponding PVP notation.^*

4.3.3.5 Let us now examine how the words with the dentalveolar sibilant initials and those with the retroflex initials are transcribed in the VP and PVP notations in PN and PP. Table 20 exemplifies how the Mandarin dentalveolar sibilants developed from the EMC initials Ds / ^ c, ch,

cH, fC.' s, sH/ are transcribed. Tables 21, 22, and 110

23 respectively exemplify how the reflexes of the three palatal series of EMC, which coalesced into the Mandarin retroflex series— namely, Dy / ty, tyh, tyH/, Py

/ ÿ cy, cyh, ^ cyH, sy, ^ syH/, and Pr[l] / ^t cr, crh, ^ crH, ^ sr, srH/— are transcribed in the VP and PVP notations. Ill

Table 20 - Transcription of the dentalveolar sibilants

zi ci /O' ci ,2'si qu XU Î* XU

A A 4 4 4 VP ^ 4 e tsi tshi tsHi Si tshju tsHju s ju sHju .i. A n PVP T -TT IT tshi ts ju s ju

jiu le jiu ^ aiu jin qin jin &T xin

//j 4! VP * tsiw tshiw tsHiw siw tsin tshin tsHin sin

a 1 ^1 PVP T ? 4» tsiu tshiu tsiu siu tsin

Table 21 - Transcription of the Dy initials

zhi ;/^ chi zhi zhu chou zhen ^ chen chen wi Kl % % % VP ^ t/i t/hi t/Hi t/ju t/hiw t/in t/hin t/Hin

4 ¥ ¥ ¥ ¥ PVP tji t/hi t/i t/ju t/hiu t/in t/hin t/hin 112

Table 22 - Transcription of the Py initials

il zhi ^ chi it'shi Jf^shi ^ zhu ^ chu ^ shu :^shu

JL â- A. A 4- -%' 4- M VP £ k tgi tghi t^ju tphju gju pHju /s A ir ir PVP it pju

^ zhou ' 3 chou ^shou '^sHou ^ zhen shen ■^ shen /^chen

% tl VP % 0 V 1' tfiw tghiw giw gHiw tgin gHin pin pHin

^ d . A l PVP ^ 4 ¥ ¥ L— tgiu tghiu piu piu " pin II tphin

Table 23 - Transcription of the Pr[l] initials

ÿ shi shi kl chu zhu shu ^ zhou ^ c h o u ^ shou ^ s h e n

A a A > A VP ^ 1 ? f 0 ù. pHl tghu tgHu pu tpiw tpHiw giw pin

A A S. A A A A > L. PVP 1 T T T* gl tpu tphàu piu pin 113

The pronunciation notations in the above tables reveal the following phenomena:

(1) In the VP notation, the shapes of the Hunmin chong'urn letters for the Korean alveolar sibilants are modified in two ways: the letters with the left stroke elongated , Â " denote the Mandarin dentalveolar sibilants, while the letters with the right stroke elongated

" \ ' K ' denote the Mandarin retroflex sibilants. In the PVP notation, however, only the letters with the original shape (both the leftward and rightward strokes of equal length "7\, K > A ") represent both the dentalveolar and retroflex sibilants of Mandarin.

(2) The Ds initial consonants occur before both palatal and non-palatal vowels or medials. Thus, it is possible that there are two allophone types of the Ds series: the palato-alveolar type [t^, tcjfh, cy] before [i] or [j] and the dentalveolar type [ts, tsh, s] before [%]. However, since I do not have clear evidence to prove this postulation, I transcribe the Ds initial consonants only with [ts, tsh, tsH, s, sH].

(3) In the PVP notation, the zhi-si rhyme words with the Dy initial consonants such as ^ zhi, ^ chi and zhi are transcribed with the letter compound " J" [i]; while the Zhi-si rhyme words with the Py initial consonants such as ^ zhi, ^ chi and shi are transcribed with the letter compound " z " ' , which denotes [%] or [i]. Thus, the 114

Dy initial consonants occur only before a palatal vowel or medial, i.e., [i] or [j]. The Py initial consonants occur before both palatal and non-palatal vowels or medials. The

Pr[l] initial consonants occur only before a non-palatal vowel or medial. In view of the fact that [tp, tph, g] are phonetically incompatible with the palatal vowel or medial

[i] or [j], we can infer from this phenomenon that chronologically the Pr[l] series became retroflex first, the

Py series second, and the Dy series third.

(4) We find the record in the "introduction" to SSTG that the Dy series merged into the Pr[2] series and formed the retroflex series of Mandarin retroflex sibilants (Pr[3]) before the fourteenth century (see Section 4.3.3.1).

However, we also find the record in the "notes" to HCH

(1446) that the retroflex sibilants had two types of allophones: the palato-alveolar [tj] type and the retroflex

[tg] type (see Sections 4.3.3.2 and 4.3.3.3). Thus, I transcribe the reflexes of the Pr[2] (i.e., Py and Pr[l]) series only with the retroflex sounds [tg, tgh, g] for both the VP and PVP notations. However, I transcribe the reflexes of the Dy series with the palato-alveolar [tJ] type for the VP and the PVP notations since they occur only before a palatal vowel or medial, i.e., [i] or [j]. In this way, I accept that the retroflex intial consonants [tg, tgh, g] ocurred both palatal and non-palatal vowels or medials in

OM and MM, although they occur only before non-palatal 115 vowels or medials in MP. Furthermore, a corollary of this reconstruction is that chronologically the retroflexion of initial consonants preceded the deletion of [i ] or [j].

Thus, the retroflex initials caused the deletion of [i] or

[j] by progressive assimilation. This process can be formulated as follows:

j — » / # r ___ (10)

4.3.4 The retroflex initial consonant /r/ The retroflex series of Mandarin consists of four retroflex sounds, namely, /or, crh, sr, r/. Since we have dealt with /cr, crh, sr/ among them in the preceding section, suffice it here to discuss /r/ only, /r/ is developed from the initial consonant of EMC and LMC denoted the representative initial character Ri B . This initial consonant of EMC is reconstructed as a palatal nasal consonant: e.g., Li Rong (1956:125-126) reconstructs it as

[A], and Wang Li (1980:50) reconstructs it as [nz]. Hsueh

(1975:28) represents it phonemically by /ny/ (/nj/ in his transcription. This initial consonant became /nr/ in LMC by

Rule (8) y r / ___ (in Section 4.3.3.1), and subsequently became /r/ in OM by the following rule (Hsueh

1975:44):

n o /# ___ r (11) 116

We also accept /r/ as the phonemic value of the ^ initial consonant of OM and MM. However, phonetically we reconstruct the M initial was probably [j] or [%"] in OM and

MM. This initial consonant is transcribed by the letter

Then, one may question why the letter " A " for

Korean /z/ was used for transcribing Mandarin /r/. My answer to this question is as follows:

(1) In Korean [r] and [1] are the allophones of the phoneme /I/ in the following manner:

/I/ [r] / # ___ V ___ V

/I/ [1] / ___ #

Thus, both [r] and [1] are transcribed by the same Korean letter "2 ", which represents /I/. Since this letter is chosen to transcribe the Mandarin dentalveolar approximant

[1] / ^ 1/, the Mandarin retroflex approximant [%;] /r/ needs to be transcribed by a different symbol in order to differentiate the two different phonemes of Mandarin.

(2) The initial Ri is an approximant sound of the Py series of EMC, and the approximant sound of the Pr[2] series of LMC after the merge of the Py and Pr[l] series. The

Pr[2] series are transcribed by the Korean alveolar sibilant letters with the right stroke elongated. Thus, the Korean letter for a voiced alveolar sibilant "a " was chosen to 117

transcribe the Mandarin approximant initial consonant of the

Pr[2] series.

The Zhi-si rhyme words with the initial consonant /r/

such as ^ ^ and are transcribed as follows:

CP ^1 [ji] -4. VP [rn PVP f [n] -> PVP[2] I , A [f-ii] or I , [U] PVP[2] stands for the PVP notation revised in PSO (1765).

As is shown above, first the initial consonant of these words became retroflex, and then their palatal vowel became apical by progressive assimilation from CP to PVP. Second, the retroflex consonant and the apical vowel switched their positions by metathesis. These words can be reconstructed phonemically as /ri/ for OM as represented by the CP and VP notations. They can be reconstructed as /ri/ for the early part of MM as represented by the PVP notation of PN and PP

(ca. 1515), but as /ir/ for the late part of MM as represented by the PVP[2] notation of PSO (1765). In addition, these words appear with the zero initial in DYTJ

(1606) (see Section 3.4). Thus, we believe that this syllable underwent metathesis during the sixteenth and early seventeenth centuries.

4.3.5 Gutturals

The guttural series consists of the velar sounds and the glottal sounds (see Section 3.2 and Table 13 in Section 118

4.2). The changes made from the VP notation to the PVP notation in transcription of the gutturals represent two types of coalescence as follows:

First, the murmured initials Qun /kH/ and Xia 13L

/xH/ respectively merged into the voiceless initials Jian

I. /k/ (or Xi /kh/) and Xiao 3 ^ /x/ through the devoicing process (see Section 4.3.1).

Second, the three initials Ying ^ /?/, Yu l/, and

Yi / merged into the zero initial /^/ before the time of PN and PP (see below).

In the VP notation, the three initials Ying ,

Yu ^ , and Yi ^ are respectively transcribed as " j " /?/,

" 0 " f/, and " 6 " /()/. But the actual sound value of the initial Yi was /{if/ already at the time of SSTG (ca.

1450) according to the following description in the

"introduction" to SSTG: M II , ^ J# ÿ I I t , t 5 IS; , i!] 0 ^ I'l $ à , V. % ±. "In the present rhyme book (HWZY), the words formerly with the initials Yi or Yu are mostly mixed up. Here, Yu

0 ^ is written as " o " ( l

" à ” (/g/) under each character according to the old rhyme books in order to differentiate them."

However, the three initials are all transcribed as

" 0 " /^/ in the PVP notation. This merger of the three 119 guttural initials can be expressed by the following two rules :

(12) (13)

This merger of the three guttural initials can be illustrated as follows:

Table 24 - The emergence of the zero initial

Initials Yingj^% Yu Yi & È Examples & you yu you # y u niu yu

%' 'TT i f r i & VP [7iw] [?ju] [iw] [ju] [niw] [gju] SI f- PVP [iu] [ju] [iu] [ju] [niu] [ju]

Some of the syllables which originally had the initial /g-/ do not have /^/ in the initial position, but the initial consonant /n/. The character ^ niu is one of them. The word ^ niu appears together with the word you in

ZYYY, and it reads /jiw/ in that dialect. The reading

/njiw/ probably came from another related dialect, and became the standard. The Korean data show that the change occurred as early as the beginning of the Ming dynasty.

In MP, the palatal initial consonants [t^, t^h, <^] occur as allophones of the velar initial consonants /k, kh. 120

h/ before a high front vowel or medial, i.e., [i], [j], or

[y] (see Section 3.2). In both the VP and PVP notations,

the Chinese velar initial consonants are all transcribed

with the corresponding Korean velar consonants, but not with

the Korean alveolar sibilants. For example, in the PVP

notation, words such as jin, £ jian, jie,

qing, ^ jia, and qi are transcribed as " ^ " [kin],

" [kjen], " 7=1" [kje], " 7j " [khig], " /j: " [kja], and

" fj " [khi], respectively. These notations show that the

Mandarin velar initial consonants were not palatalized

before a palatal vowel or medial yet at the time of the PVP

notation (ca. 1515). Later, in the PVP[2] notation, i.e.,

the PVP notation revised in PSO (1765), some of the Mandarin

words with a velar initial and a palatal vowel or medial are

transcribed with a Korean alveolar sibilant, e.g., ^ "

[tsi] ([t^i]), ^ " 4 " [tsja] ([t

([t^ig]), etc.; while some of them are still transcribed with a Korean velar initial, e.g., ^ " /j " [ki], â "7^"

[kjen], ^ " [kin], etc. According to these PVP[2]

notations, the Mandarin velars were in the process of

palatalization in front of a palatal vowel or medial in the

eighteenth century (Kang S. H. 1978b: 83-84).

4.3.6 Summary

We have found that the prominent changes from the

initial consonants represented by the VP notation to those 121

of the initial consonants represented by the PVP notation

are (1) the devoicing of the former voiced obstruent initial

consonants, accompanied by a process of tonal split and

regrouping, and (2) the coalescence of the three guttural

initials /?-, o-, ^ -/ into the zero initial /0-/. These two

changes coincide with the major differences between the

initial system of HWZY and that of ZYYY. That is, the

differences between the initial system of the VP notation

and that of the PVP notation reflect the dialectal

differences between Lower Yangzi Mandarin and Northern

Mandarin in the OM period (cf. Kang S. H. 1973 and 1974).

Phonemically, there existed only two series of

sibilants in OM and MM, i.e., the dentalveolar sibilants /c,

ch, s/ and the retroflex sibilants /cr, crh, sr/. The

dentalveolar sibilants /c, ch, s/ were pronounced as (ts,

tsh, 8]. The retroflex sibilants /cr, crh, sr/ were

pronounced as [tJ, tj^, /] before [j] or [i] due to

regressive assimilation, and as (tp, tph, g] elsewhere.

Therefore, there were at least three types of sibilants

phonetically.

Besides, the initial /v-/ was in the process of merging with the zero initial at the time of the PVP

notation, and it completely merged with the zero initial

before the time of PVP[1] notation of PN (1670) and PP

(1677). The Mandarin velar initial consonants were not

palatalized before a palatal vowel or medial, i.e., [i]. 122

[j], or [y], yet at the time of the PVP notation of PN and

PP (ca. 1515).

In sum, on the basis of our discussion of the initial

system of MM, we can set up the initial consonant phonemes

of MM as follows :

Table 25 - The phonemic representations of the MM initials

Manner Stops & Affricates Frica- Nasals Approxi- -asp +asp tives mants

Place

Labials p ph f m (v)

Dentalveolars t th n 1

Dentalveolar c ch s sibilants

Retroflexes cr crh sr r

Gutturals k kh h gf

NOTE: The symbols "c, h, r" respectively denote [ts], [+aspirated], and [+retroflex]. /v-/ is put in parentheses since it was on the verge of diappearing during the MM period. 123

CHAPTER IV

NOTES

1. A complete explanation of the shapes of the Hunmin chong'um letters is presented in HCH. For an English translation of HCH, see Ledyard 1966:228-260.

2. In studies of Korean phonology, the three parts of a Korean syllable are called conventionally as the initial, the medial, and the final, which are respectively the literal translations of the traditional terms: ch'o-song ( , ji/J ^ ) f chung-song ( ), and chong- song ( ^ ) • However, the terms medial and final are used for different meanings in studies of Chinese phonology including this study (see Section 3.1). Therefore, in order to avoid any possible confusion of these terms, in this study we call the three parts of the Korean syllable the initial, the vowel, and the ending. The relation between the terms for Korean syllabic parts and those for Chinese syllabic parts is as follows: (1) The Korean initial corresponds to the Chinese initial. (2) The Korean vowel corresponds to the whole final of a Chinese syllable with a semivowel ending or zero ending, or it corresponds to the medial and the nucleus only of a Chinese syllable with a consonant ending. (3) The Korean ending corresponds to the consonant ending of a Chinese syllable.

3. The practice of writing syllabically is common in Chinese, Japanese, Korean, and Vietnamese, irrespective of their different writing systems. See Ledyard 1966:213.

4. A murmured sound is made "when the vocal cords are only slightly apart: they can still vibrate, but at the same time a great deal of air passes through the glottis." (cited from Ladefoged 1975:122)

5. Although voicing was a distinctive feature in EMC and LHC, it is not a distinctive feature in Mandarin. Thus, here we may use either the letters for voiced stops or those for voiceless stops. We use the symbol "H" to denote murmur for typographic convenience.

6. See Chapter 1 Fn. 16 and Pulleyblank 1984:123, 163. Cf. Martin 1953:16.

7. In addition, there are similar descriptions of the devoicing of the muddy initials and the concomitant tonal 124 split in the "introduction" to SSTG and in the "introduction" to SSTH.

8. In the rhyme tables such as the Sishena denazi and the Oievun zhizhana tu. the two series of labials appear in complementary distribution; the labiodental initials occur only in division 3 of the rounded type, while the bilabial initials occur elsewhere. See Hsueh 1975:38.

9. The split of the labial initials created a new series labiodental affricates, i.e., /pf, pfh, pfH, mv/. These affricates became /f, fH, v/ by the process of de­ af frication, and suseguently devoicing occurred. Thus, only /f, v/ remined. (Hsueh: personal communication)

10. The process of this coalescence is much too complex to be dealt with thoroughly here. For further discussion of the process of this coalescence in phonetic terms, see Wang Li 1980:116. Also, for one in phonemic terms, see Hsueh 1975:41-43.

11. The same description also appears in the introduction to HMCUYH.

12. The phonetic nature of the retroflex series of EMC and LMC is also controversial. For instance, Wang Li (1980:50) reconstructs the Pr[l] series of EMC as /tJ, tJ', dj', J/, while Pulleyblank reconstructs them as /tp, tg', d%, p, ^/. However, the phonetic values of the retroflex series of EMC and LMC are not our imminent concern here since we are primarily concerned with the diachronic change of Mandarin pronunciation in this study.

13. Kang S«C. (1985:80) states that Korean phenologists (of the Chinese language) underdifferentiated the Mandarin sibilants. A corollary of my view here is that at that time Korean phenologists of the Chinese language could differentiate the dentalveolar and retroflex sibilants, although the other Koreans may have underdifferentiated them.

14. Hu (1963:186) states that the transcriptional change occurred in the PVP notation since the transcribers did not understand the distinction between the dentalveolar sibilants and the retroflex sibilants. However, in my opinion, the transcribers did not have to print the dentalveolar and the retroflex sibilants differently in the PVP notation since the two series of sibilants are printed differently in the matching VP notation. If they really had not understood the distinction between the two series of sibilants, they probably would not have printed the two series differently in the VP notation, either. On the 125 whole, the PW notation was designed to show contemporary pronunciation, and was suitable for pedagogical use; while the VP notation was kept to show conservative pronunciation, and was useful for phonological study.

15. The graph "d " in Middle Korean occurred in between vowels, and its sound value is reconstructed as [z] /z/. When the consonant was lost in Korean after the mid­ sixteenth century, the graph " a " became obsolete, and it is not used in writing Modern Korean.

16. Wang Li (1958:109) reconstructs both [tj', tj'', / , J ] and [tg, ts', g, %] for the retroflex initials of ZYYY. He considers that the [tj] type was in the process of changing to the [tg] type at the time of ZYYY, and.states that the change was probably was completed only after the fifteenth century (ibid., p. 116) CHAPTER V

THE FINAL SYSTEM OF MIDDLE MANDARIN

5.0 Introduction

In this chapter, we shall reconstruct the finals of MM on the basis of the PVP notation of PN and PP in

comparison with the CP and VP notations, and with occasional

reference to other relevant contemporary materials. Since

MM is the intermediate stage between OM and MP, we shall

also make reference to our knowledge of OM phonology and MP

phonology in order to help determine the phonetic and

phonemic status of the MM finals. In addition, when there

is a need to compare the OM finals with their previous

forms, we shall make reference to the final system of LMC as

represented by the rhyme tables.

In examining the pronunciation notations, we shall take into consideration the sound system of Middle Korean and the method of pronunciation notation, so that we can decipher the sound values of the Hunmin chono'um letters used for transcribing Mandarin pronunciation and account for possible "phonic interference"^ and transcriptional expediencies (see Section 4.1).* Due to the structural

126 127 difference between the Chinese and Korean syllables, the medial, nucleus, and semivowel ending of a Chinese syllable are transcribed together as the vowel of a Korean syllable, although the Chinese consonant ending corresponds to the

Korean consonant ending (see Section 4.2). For this reason, it is both logical and practical to examine the Korean transcriptions of the Chinese whole finals, so that we may determine the phonetic and phonemic status of the finals of both OM and MM.

We accept that Hsueh's (1975) reconstruction of the

OM vowel system represents the nuclear vowels of LMC, OM, and MM at the phonemic level accurately. Hsueh's (ibid.) phonemic reconstruction of the OM nuclear vowels consists of four vowel phonemes, which have a two-way contrast in height and a three-way contrast in depth as shown in Table 26.

Table 26 - The vowel phonemes of OM

Front Central Back

High i

Low e a o

SOURCE: Hsueh 1975:54. NOTE: Phonetically, /e/ and loi may be relatively higher than /a/. Hsueh (1986:93) replaces loi by /3/. /D/ is more appropriate than loi for representing the low back vowel phoneme, which is phonetically realized closer to [o] than to [o]. However, we keep using the symbol loi for typographical convenience. 128

As for the final system of LMC, OM, and MM, we shall take the following results of previous research as our points of reference:

(1) The phonemic reconstruction of the final system of the rhyme tables by Hsueh (1982 and 1985) (see Tables 27).

(2) The phonemic representations of the finals of ZYYY as reconstructed by Hsueh (1975) and revised by Hsueh (1986)

(see Tables 28 and 29).

(3) The phonemic representations of the finals of HWZY as reconstructed by Chou (1989) (see Tables 30 and 31).

(4) Lu Z. W.'s (1947b) phonetic reconstruction and

Hsueh's (1983) phonemic reconstruction of the finals of Xu

Xiao's ( ^ ) Chonadina Sima-Wenaona Denovun-tuiina i ^ Ü ® 1606 hereafter DYTJ), which represents, in the form of rhyme tables, the pronunciation of Northern Mandarin of the early seventeenth century (see

Table 32).

Both Hsueh's (1975 and 1986) phonemic reconstructions of the finals of the rhyme tables, ZYYY, and DYTJ, and

Chou's (1989) phonemic reconstruction of the finals of HWZY are arrived at on the basis of the following two working rules (Hsueh 1975:51 and Chou 1989:98):

(1) Only one nuclear vowel is allowed for one rhyme (in

ZYYY) or one rhyme-set (in HWZY). 129

(2) Different rhymes (which share the same ending in

ZYYY) or different rhyme-sets (which share the same ending or homorganic stop endings) must contrast in nuclear vowels.

We adopt these two working rules for our phonological analysis of the finals of Mandarin of any stage.

Furthermore, since we shall examine the phonetic details of the finals of MM extensively in this chapter, we need to restate Working Rule (1) in the above as the following to suit our phonetic analysis and description of the finals of

Mandarin of any stage:

(3) Phonetically different vowels, as long as they are allophones of one nuclear vowel phoneme, may be allowed for one rhyme (in ZYYY) or one rhyme-set (in HWZY). 130

Table 27 - The final types of the sixteen rhyme classes

Inner-A:

Tone Zhi Yu Liu Shen

1 -wiT),k -wi# -iw 2 3 -jwij),k -j(w)iy -jwi# -jiw -jim,p 4

Inner-B:

Guo Zena Dana

1 -(w)o# -(w)op,k -(w)og,k 2 3 -j(w)o# -j(w)o;i,k -j(w)og,k 4

Outer-A:

Gena Jiana Xie Xiao Shan Xian Jia

-(w)oj -ow -(w)on,t "Oni/p ————— 2 -(w)ap,k. -ag,k -(w)aj -aw -(w)an,t -am,p -(w)a# 3 -j(w)a/i,k ------j(w)aj -jaw -j(w)an,t -jam,p -j(w)a# 4 -j(w)eji,k ---- -(w)ej -ew -(w)en,t —jem,p —————

Outer-B:

Zhen

1 -(w)in,t 2 [-(w)an,t] 3 -j(w)in,t

SOURCE: Hsueh 1985:47-48 NOTE: The numbers on the right side mark divisions. Inner-A, Inner-B, Outer-A, and Outer-B denote the final- patterns. The symbol "y" for the palatal medial in Hsueh's transcription is replaced by '' j ". 131 Table 28 - The rhymes of ZYYY (1) Dong-zhong m -ij (11) Xiao-hao $k -ow (2) Jiang-yang -or) (12) Ge-ge -og)

(3) Zhi-si (13) Jia-ma -a^ (4) Qi-wei -ij (14) Che-zhe fii. -e^ (5) Yu-mo -wi^ (15) Geng-qing

(6) Jie-lai -a j (16) You-hou t ft -iw (7) Zhen-wen At -in (17) Qin-xun ft f -im (8) Han-shan —an (18) Jian-xian -am (9) Huan-huan -on (19) Lian-xian a # -em (10) Xian-tian k % -en SOURCE: Hsueh 1975 and 1976 NOTES: (a) The rhymes are numbered according to their original order in ZYYY. (b) Hsueh (1975) originally reconstructed the rhyme bases /-or)/, /-ag/, and /ew, aw, ow/ respectively for rhymes 1, 2, and 11, but he (1986:93 and 105 Fn. 2) later revised them as shown in this table and Table 29. (c) The symbol for the palatal semivowel or the feature Palatal, viz /y/ is replaced here by /j/. 132 Table 29 - The finals of ZYYY

~p — j —w

-i- -e- -a- -o- -i- -a- -i- -o-

(3) (14) (13) (12) (4) (5) (16) (11)

i a O ij aj iw ow

-j- f5)_ je ja jij jaj jiw jow

-w- wi wa wo wi j waj wiw wow

-jw- jwi jwe jwo

-n -m

-i- -e- -a- -o- -i- -e- -a- -i- -e- -o-

(7) (10) (8) (9) (17) (19) (18) (1) (15) (2)

in an im am er) or)

- j- jin jen jan jim jem jam je)) jor)

-w- win wan won wig we g wo g

-jw- jwin jwen jwig jweg

SOURCE: Hsueh 1975:74 and 1986:93 NOTES: See Notes to Table 28. 133 Table 30 - The rhyme-sets of HWZY

Level, Rising, Going Enteri

1) Dong ^ Dong ^ Song ^ (1) Wu 2) Zhi Zhi Zhi 3) Qi Ji f. Ji 4) Yu ÏU If Yu Mo Mu ^ Mu 5) f i $ 6) Jie Jie Tai & 7) Hui « Hui ^ Dui fÈ<. Zhen Zhen ^ Zhen ^ Zhi f 8) % (2) Han Han Han (3) He 9) ? 10 Shan S i Chan Jian (4) Xia 11 Xian Xian f t San % (5) Xi

12 Xiao f Xiao H Xiao

13 Yao Qiao Xiao % 14 Ge Ge M Ge 15 Ma Ma Ma

16 Zhe Zhe Zhe

17 Yang Yang ^ Yang (6) Yao ^

18 Geng Geng Jing % (7) Mo

19 You t You ^ You r 20 Qin ft Qin ^ Qin (8) Qi # 21 Tan Gan Kan Èfj He ^ /O'* (9) 22 Yan 1 Yan Yan (10) % É NOTE: The numbers of the rhyme-sets and those of entering tone rhymes are the same as their original numbers in HWZY. 134

Table 31 - The finals of HWZY

-w i -i- -i- -i- (16 ^ (151 (14V (191 (121 (131 (31 (61 i- a o iw ow aj

- j- ji je ja jiw jew jow jij jaj 5 7 —w — wi wa wo wi j waj

-C ~ jw- jwi jwe

n -m -9 -i- -i- f8) I 111 (101 (91 (20 1 (221 (211 (1) (181 (171 in an on im am °0

- j“ jin jen jan jim jem jam jej) jog

—w— win wan won WiQ wej) wo g

- jw- jwin jwen jwitj jwe^

t -p -k -i- -i- -i- (5) (41 (31 (8) (101 (91 (11 (71 (61 -t- it at ot ip ap ek ok

-i- jit jet jat jip jep jap jek jok.

—W— wit wat wot wik wek wok

-jw- jwit jwet jwik jwek

SOURCE: Chou 1989: 207 NOTE: See Note to Table 30. 135

Table 32 - The finals of DYTJ

Xu Xiao Lu Z-W. Hsueh Xu Xizo Lu Z-W. Hsueh —————————r——————— ———————————————————————————————————— Tpnq ^ ^ u, iu

: o,(w,.,--- 1- ^ A ! ------( K f ) Zhuo ^ Zhi i ic (j)(w)e# . I U6 ■^r’ (j)(w)i# (t? # ) iue

( __ -t- ) y (u iu?) Zhen ------^ an Xie in ( j ) (w)in ^ ai u(a)n : iai (jfW)aj ( A . / ^ ) iun i à p - t ______------Lei ^ an ei (iei) i&n (j)(w)an ( 1% 1 % . ) uei (yei) {w)ij ^ ^ uan ------{ § ai| ) iugn Xiao ------*u (?) Dang (%à ) iftu (?) (j/W)aw AT) (»g?) m g (j,w)ag Guo ^ _ uan (uttn?) 0 (?X^f) (liudg) _ i: (j)(w)o# — — rr------Liu ^ ------9u (?) ^ K i7^ ti) i9u (?) (j,w)iw A------1 A ( 3^ ) u A (j,w)a#

SOURCE: Lu Z-W. 1947b:189-190 and Hsueh 1983:161. NOTES: The rhyme labels are those of the thirteen rhymes of DYTJ, and their corresponding 'Thirteen Tracks' rhyme labels are given in parentheses for comparison. 136 There are two closely-related dialectal strata or speech variants in the development of Mandarin pronunciation, mainly in respect of the development of the former entering tone syllables. One is called the

"colloquial pronunciation" ( ll? If* ^ kouyu-yin), and the other is called the "literary pronunciation" ( ^ ^ dushu-yin). The former and the latter respectively originate from Northern Mandarin and Lower Yangzi Mandarin, but both gradually became coexisting speech variants of

Northern Mandarin. The basic phonological structure of the

"colloquial pronunciation" is, however, almost identical with that of the "literary pronunciation", except that the latter retained for some time the entering tone category in the form of the glottal stop ending /-?/(Hsueh 1975:134).^

HWZY represents mainly Lower Yangzi Mandarin of the

OM period, while ZYYY represents Northern Mandarin of the OM period. HWZY follows the "literary pronunciation", while

ZYYY follows the "colloquial pronunciation" with heavy borrowing from the "literary pronunciation". MP follows the

"literary pronunciation" but with heavy borrowing from the

"colloquial pronunciation" (Hsueh 1975:134 and Chou

1989:281-283). Besides, in the sound system of HWZY, the entering tone category with the endings -p, -t, and -k remained, and also the former voiced obstruent initials remained partially (Chou 1989:101-106). On the other hand, in the sound system of ZYYY, the entering tone category and 137 the former voiced obstruent initials are comlpletely lost

(see Section 4.3.1). Notwithstanding such discrepancies between HWZY and ZYYY, these two books both provide important phonological evidence of the OM final system (cf.

Chou 1989).

There are seventy-six rhymes in total in HWZY. A rhyme in HWZY, in accordance with its meaning as used in the

Oievun. represents the syllables with the same rhyme base and the same tone (cf. Section 1.3.2). Hereafter, we shall call a set of rhymes with the same rhyme base but with different tones in HWZY by the term "rhyme-set".* There are twenty-two rhyme-sets in HWZY. Twelve of them consist of three rhymes each, one rhyme for each of the level, rising, and going tones; while the remaining ten rhyme-sets consist of four rhymes each, one rhyme for each of the level, rising, going, and entering tones. We shall call each rhyme-set by the character representing the rhyme with the level tone (see Table 30).

The term "rhyme" used in ZYYY has a different meaning from that of the same term used in the Oievun. Syllables of the same rhyme in the Oievun must share both the same rhyme base and the same tone. However, syllables of the same rhyme in ZYYY must share the same rhyme base, but they may have different tones (of. Hsueh 1975:18, 51). Thus, the term "rhyme" in ZYYY is roughly equivalent to the term

"rhyme-set", which we use here to refer to the rhymes with 138 the same rhyme base but with different tones in HWZY.

Moreover, since the twenty-two rhyme-sets of HWZY are very similar to the nineteen rhymes of ZYYY,^ the finals of OM and their reflexes in MM can be classified systematically according to either the rhyme-sets of HWZY or the rhymes of

ZYYY.

On the other hand, if we arrange the three types of

Korean pronunciation notations in the same manner, it will be convenient to examine the sound correspondences between the "present vulgar" pronunciations with the "correct" and

"vulgar" pronunciations of the same words (see Section

2.2.2). In this respect, the CP notations of the OM finals can be classified conveniently according to the rhyme-sets of HWZY since the CP notation represents the sound system underlying HWZY. Therefore, in the following discussion of the MM final system, we shall arrange the MM finals according to the rhyme-sets of HWZY. In addition, we shall group the rhyme-sets according to the types of syllable endings.

5.1 The rhyme-sets Zhen. Han. Shan, and Xian

These four rhyme-sets in HWZY have the alveolar nasal stop ending /-n/ for the rhymes with a non-entering tone or the alveolar stop ending /-t/ for the rhymes with the entering tone. The endings of the entering tone words of the rhyme-sets Zhen. Han. Shan, and Xian are transcribed 139

uniformly as " c " [-t] in the CP notation, as " 5" " [-?] in

the VP notation, and as zero in the PVP notation. According

to this transcriptional change, the alveolar stop ending was

deleted completely no later than the time of the PVP

notation (or the beginning of the sixteenth century). This

change can be formulated by the following two rules in

sequence ;

t -+ ? / V ___ # (14)

? — ^ / V ___ # (15)

The Zhen rhyme-set— which comes from the Zhen rhyme

class of LMC— is reconstructed with the high central vowel

/-i-/ as its nucleus (Chou 1989:151). The other three

rhyme-sets Han. Shan, and Xian— which are all developed from

the LMC Shan rhyme class— are respectively reconstructed with the low vowel phonemes /-o-/, /-a-/, and /-e-/ as their

nuclei (ibid., p. 154).

5.1.1 The Zhen rhyme-set

The pronunciation notations of the words of the Zhen

rhyme-set are illustrated in the following table. 140

Table 33 - Transcription of the Zhen rhyme-set words

fX jin ^ min 1^ zhen È7 qi 3 ri

CP [kin] ^ [min] ^ [tgin] g [khit] cr[?it]

VP ÿ [khi?] -^i[^i?]

PVP ■[a ?( [khi] Aj [%i]

% gen ben <èK dun ^ jun ^ chun

CP ■h [kin] [pun] -^■[tun] ^ [kjun] [tghjun]

VP -C [pin]

PVP ^[pin]

ge f gu ^ bu chu

CP c* [kit] Ï [kut] [put] ^ [tghjut]

VP f [ki?] t [ku7] [Pu?] 4 [tshju?]

PVP ± [ki] T [ku] T [Pn] [tghju]

NOTE; The IPA transcriptions on the right side of Korean pronunciation notations are basically the phonemic transliterations of the notations. However, we use them as phonetic data for the OM and MM pronunciations of the corresponding Chinese characters. Therefore, we put them in brackets (for phonetic transcriptions), instead of slanting lines (for phonemic transcriptions), in this and all the following tables exemplifying the Korean pronunciation notations. 141

1. The finals of such words as jin, ^ bin,

jin, zhen, $J/ yin are transcribed as " J " [-in] in the

CP, VP, and PVP notations. We take [-in] and /-jin/ respectively as the phonetic and phonemic forms of this final type since the phonemic compound /ji/ is phonetically realized as [i] in Mandarin (see Section 3.3.3).

This type of final remains unchanged in MP except when it occurs after retroflex initials. The medial /-j-/ after retroflex initials was deleted during the period between MM and MP. Thus, in MP, words like (É zhen

([t^in]/crjin/ in OM and MM ► [tg@n] /crin/ in MP) and shen ([^in]/srjin/ in OM and MM — » [gan] /srin/ in MP) share the same final [-an] /in/ with words like gen ([k@n]

/kin/ in OM, MM, and MP) and hen ([xan] /hin/ inOM, MM, and

MP). This change is represented by Rule (10) j — / / # r __ (in 4.3.3.5).

2. The finals of such words as gen, zhen, and

en are transcribed as "XT " [-in] in all the three pronunciation notations. The nucleus of these words in OM and MP is often transcribed as the mid central vowel [j]

(e.g., Dong T. H. 1968 and Cheng C. C. 1973). However, the nucleus of these words is transcribed consistently with the letter " ", which denotes the high central unrounded vowel

/i/ in Middle Korean and Modern Korean. Thus, I consider that the nucleus of this final must be a high vowel with perhaps a phonetic quality somewhat higher than but similar 142 to schwa, namely, [a] (see Section 3.3.3). Therefore, I reconstruct [-5n] and /-in/ respectively as the phonetic and phonemic forms of this final.

The finals of the words with labial initials such as

ben, pen, men, and "SC wen are transcribed as

" %"" [-un] in the CP notation, but they are transcribed as

" C " [-in] in the VP and PVP notations, [-un] and [-+n] are phonetic realizations of /-win/ and /-in/, respectively.

Thus, from this transcriptional change we can infer that labialization became weak after labial initials, in a phonetic sense, before the time of the VP notation (or the mid-fifteenth century) (cf. Hsueh 1986:30). But the labial medial /-w-/ remains at the phonemic level. This change can be formulated by the following rule:

w ^ / B (16 )

3. The finals of such words as gun, dun, sun, wen are transcribed as " 27 " [-un] in all three pronunciation notations. We take [-un] and /-win/ respectively as the phonetic and phonemic forms of this final since the phonemic compound /wi/ is phonetically realized as [u] in Mandarin (see Section 3.3.3).

Phonemically, this final type is kept unchanged in

MP. Phonetically, however, when this final type is followed by the zero initial, it is usually pronounced as [wan] in

MP, e.g., 7^ wen; whereas, judging by the prounciation 143

notations, it was pronounced as [un] in OM and MM. To

explain, the phonemic compound /wi/ is phonetically realized

in two ways by the following pair of rules:

/wi/ — [wa] / # __ E (E^ o)

/wi/ — [u] / elsewhere

I consider that these rules are valid for the whole period

of Mandarin. Thus, the word '/flz. wen was pronounced as [?un] when it maintained the glottal initial; but the word has been pronounced as [wan] since the initial became zero.

Since the glottal initial consonant merged into the zero

initial before the time of PN and PP (cf. Section 4.3.5), the word wen was probably pronounced as [wan] at the time of the PVP notation. To the contrary, the word -jan. wen

is transcribed as [un] in the PVP notation just as in the CP and VP notations. This discrepancy between the probable pronunciation and the pronunciation suggests that [un] and

[wan] must have been pronounced very similarly. Thus, I presume that these two phonetic forms were both phonetic realizations of the same phonemic form /win/.

4. The finals of words like ^ jun, ^ xun, ^ chun, and S yun are transcribed as " S " [jun] in all three pronunciation notations. In OM this type of final is the cuo-kou /-jw-/ type of the ZYYY Zhen-wen rhyme /-in/, and in MP it is the cuo-kou /-jw-/ type of the Ren-chen rhyme /-in/ except for the words with a retroflex initial, 144 which joined the he-kou type of the Ren-chen rhyme due to the loss of the semivowel /-j-/. Hence, the phonemic value of this final must also have been /-jwin/ in MM.

In order to figure out the phonetic value of this final, we must first determine the phonetic value of "'IT "

[ju] in MM. In PVP, words such as lé you and you of the You rhyme-set are transcribed as " " [iu], while words such as 09 yu and yu of the Yu rhyme-set are transcribed as " " [ju]. Phonemically, the former notation represents

/jiw/, while the latter represents /jwi/. One may postulate the following possibilities for the phonetic value of /jwi/ in MM: [iu], [ju], [j*], [y]. Among these possibilities I choose [j«] or [y] for the following reasons;®

(1) [iu] is ruled out since it is transcribed with a different pronunciation notation, viz., " ^

(2) If /jwi/ was pronounced as a compound sound in MM, it is plausible that it was realized as [ j*]. Since [-«■] is a high central rounded vowel, while [u] is a high back rounded vowel, [»] is closer to [j] than [u] in articulatory distance.^

(3) If /jwi/ was a single sound phonetically in MM, it was probably pronounced as [y].

However, we do not have definite evidence to choose [ju] or

[y]. In addition, /-jwi-/ is consistently transcribed with

" TT ", which is pronounced as [ju] in both Middle Korean and

Modern Korean. Thus, we transcribe the phonetic form of 145

/-jwi-/ in MM as [ju].

5. Due to the complete loss of the entering tone ending, former entering tone words of the Zhen rhyme-set acquired new rhyme bases as follows:

(1) The final type " J " [-it] in the CP notation

(e.g., ^ [kit], [pit], [tshit], and [sit]) changed to " j " [-i] in the PVP notation.

(2) The final type " J " [-ut] in the CP notation

(e.g., [kut], [put], and ^ [ut]) changed to " T "

[-U] in the PVP notation.®

(3) The word [kit] (/kit/) in the CP notation became [k+] (/ki/) in the PVP notation. When /i/ occurs alone as a whole final after a guttural initial, it is phonetically realized as [w], which is the basic phonetic form of /i/. In MP, the phonetic and phonemic forms of this word are [kÿ-] and /k^/, respectively. This change shows that the nucleus of this word was lowered after the complete loss of its ending. This change in vowel height can be formulated as follows:

i / G # (17)

6. In sum, the Zhen rhyme-set has the final types as follows :

(1) With the non-entering tone words, [-in] (/-in/),

[-in] (/-jin/), [-un] (/-win/), and [-jun] (/-jwin/) occur in both the CP, VP, and PVP notations.® 146 (2) For the entering tone words [-it] (/-it/), [-it]

(/-jit/), [-ut] (/-wit/), and [-jut] (/-jwit/) occur in the

CP notation, while the ending [-t] (/-t/) is replaced by

[-7] (/-?/) in the VP notation. These final types are changed to [i] /-i/, [-i] (/-ji/), [-u] (/-wi/), [-ju]

(/-jwi/) in the PVP notation.

5.1.2 The Shan rhyme-set

The Shan rhyme-set in HWZY and its corresponding rhyme in ZYYY, or the Han-shan rhyme, have been both reconstructed with the low central vowel phoneme /-a-/.

Their reflexes in MP also have /-a-/ as their nucleus.

Therefore, it is probable that words of the Shan rhyme-set also had /-a-/ in MM as their nucleus. Now, let us examine the Korean notations of the words of the Shan rhyme-set. 147

Table 34 - Transcription of the Shan rhyme-set words

if dan jian guan wan fan

CP ft [tan] ^[kjan] ^ [kwan] -K!" [vwan] [ f wan ]

VP 'Ë [kjen] [kwan] ^^[fan]

PVP ft [wan]

^ da xia # fa ^ wa shua

CP ^ [tHat] t:^[xat] ^ [fwat] ^ [vwat] ^'"[fwat]

VP Ÿ [tHa ] ÿ[

PVP [ta] t|=[kja] 8f[fa] -2f-[wa] [gwa]

1. The finals of such words as dan, man, lU shan, Ian are transcribed as " " [-an] in all three pronunciation notations. We take [-an] and /-an/ respectively as the phonetic and phonemic forms of this final type.

2. Since there was not any mid or low front vowel phoneme in Middle Korean (see Table 12), " -j " [g] (/9/ in

Korean) was the closest vowel in Middle Korean to transcribe

[£] or [e] in OM and MM. Consequently, the Korean diphthong

" =J " [j3]— which is graphically made up of " | " [j] and

" -j " [9]— was chosen to transcribe [j£] or [je]. 148

The finals of such words as jian, yan, and

xian are transcribed as " " [-jan] in the CP notation, but they are transcribed as " ^ " [jen] in the VP and PVP notations. These words are from the LMC /-an/-type words with guttural initials. These words first became palatalized as shown in the CP notation, and later vowel- fronting took place as shown by the transcriptional change from the CP notation to the VP notation.

^ j / G ___ a (18)

a — > e / j ___ n (19)

In OM /-an/ and /-en/ are separate rhyme bases.In contrast, in DYTJ, these two rhyme bases of OM merged into one, namely, the rhyme base of the Shan rhyme. Hence, we know that the two rhyme bases of OM merged into one before the seventeenth century. In short, initially /-an/ after

/-j-/ became /-en/ by Rule (19) (see above), and later /-a-/ and /-e-/ became allophones of /-a-/ as formulated.

e, a -•» a / ___ n (20)

Therefore, we take [jan] and /jan/ respectively as the phonetic and phonemic values of the final type " ^ " [-jan] for the CP notation. We take [-jan] as the phonemic representation of the final type " ^ " [-jen] for MM.

3. The finals of such words as guan, wan, and wan are transcribed as " -‘■f'" [-wan] in all the three 149 transcriptions. Thus, we take [-wan] and /wan/ as the phonetic and phonemic forms of this final type.

When this final type occurs after labiodental consonants Fei /f-/ (e.g., fan and fan) and Feng

/fH-/ (e.g., fan and fan), it is transcribed as

" " [-wan] /-wan/ in the CP notation, but as " " [-an]

/-an/ in the VP and PVP notations. These words belong to the Yuan rhyme in division 3 of the Shan rhyme class, and the Yuan rhyme is reconstructed with the final type /-jwan/ in LMC. This final type first lost /-j-/ by Rule (5) j -» ^ / Bf ___ w (in Section 4.3.2.) between LMC and OM, and then lost /-w-/ by Rule (16) w — > ^ / B ___ (in Section

5.1.1) between OM and MM. However, words with the Wei /v-/ initial, e.g., 3^ wan and wan, are transcribed with the final type " " [-wan] /-wan/ in all three transcriptions.

This shows that theses words lost the initial /-v/ by Rule

(6) V -+ 0 / # ___ (in Section 4.3.2).

4. The word jian is from the LMC /-an/-type words.“ However, this word is transcribed with the Korean diphthong [ja] in the CP notation. Similarly, the word xia is transcribed with the Korean monophthong " j- " [a] in the CP notation, but with the Korean diphthong " " [ ja] in the VP and PVP notation. This transcriptional change is represented by Rule (18) ^ — j / G ___ a (in this section). 150

5. Due to the complete loss of the entering tone

ending, the rhyme-base /-at/ of the entering tone words of

the Shan rhyme-set changed to /-a/.

6. In sum, the words of the Shan rhyme-set have the

following final types:

(1) With the non-entering-tone words, [-an] (/-an/) and

[-wan] (/-wan/) occur in the CP, VP, and PVP notations. In addition, [-jan] (/-jan/) occurs in the CP notation, while

[-jen] (/-jan/) in the VP and PVP notations.

(2) With the entering tone words, [-at] (/-at/) and [-wat]

(/-wat/) occur in the CP notation, while the ending [-t]

(/-t/) is replaced by [-?] (/-?/) in the VP notation. These final types are changed to [-a] (/-a/), [-wa] (/-wa/), and

[-ja] (/-ja/) in the PVP notation.

5.1.3 The Xian rhyme-set

The nucleus of the Xian rhyme-set of HWZY and that of its corresponding rhyme of ZYYY, viz., the Xian-tian rhyme, have been both unanimously reconstructed with a low front vowel." The Han-shan /-an/ and Xian-tian /-en/ rhymes of

ZYYY coalesced into the Shan /-an/ rhyme of DYTJ. To explain, the nuclei of the words from the Han-shan and Xian- tian rhymes became the allophones of the low vowel phoneme

/-a-/ in DYTJ, and they were respectively pronounced as [a ] and [£] as formulated by the following rule: 151

/a/ — [$] / (w) ___ n

/a/ — [8 ] / j(w) ___ n

Therefore, the nucleus of the Xian rhyme-set words must be a low front vowel, namely, [&] /e/ in OM, and [£] /a/ in MM.

On the other hand, there was not an appropriate vowel in Hunmin chona'um to transcribe a low front vowel, for there was not a mid or low front vowel phoneme om Middle

Korean. The only low vowel letter in Hunmin chona'um was

" t " [a], but this letter was used to transcribe the nucleus of the Shan rhyme-set. Therefore, the next closest vowel letter " -f " [3] was chosen to transcribe a low front vowel of OM and MM. In short, the vowel letter '' ~f " [9] in the Korean notation denotes [£] /e/ in OM and MM when it is used to transcribe the nucleus of the Xian rhyme-set words

(cf. Section 5.1.2 no. 2). In some previous studies the letter " "/ " was often interpreted as a non-front vowel symbol in their decoding of the Korean notations of the Xian rhyme-set words into IPA," but this practice is misleading since the nucleus of the Xian rhyme-set was a front vowel.

Now, let us examine the pronunciation notations of the words in the Xian rhyme-set. 152

Table 35 - Transcription of the Xian rhyme-set words

I, jian ^ tian mian ^ yuan chuan

CP ^ [kjen] ^ [thjen] ^ [mjen] 4^[gjuen] [tgHjuen]

VP [t&hjuen]

PVP ^ [juen] [tphjuen]

jie W( tie -t/7 q i e ^ jue 5^ shuo

CP ^ [kjet] [tjet] ^ [ts^jet] ^[tsHjuet] ^ [gjuet]

VP r [kje ] ^^[tje ] ^[t^je ] ^ [t s H j u e ] ^ [ g j u e ]

PVP ^ [kje] 5f[tje] M [ts^'je] W[tsjue] ^[fjue]

NOTE: " =1 " [ja] and " 7T^ " [juja] are respectively transcribed here as [je] and [jue] in IPA (see the above explanation, no. 2 in this section. Table 14 in Section 4.2, and Section 5.1.2 no. 2).

1. The finals of such words as ^ jian, ^ tian, and ^ mian are all transcribed with " £ " [-jen]. We take

[-j£n] and /-jen/ respectively as the phonetic and phonemic forms of this final type.

2. The finals of such words as yuan, ^ guan, and chuan are all transcribed as " " [-juen]. Since

" 7^ " is graphically a sequence of " if " [ju], " =j " [ja], and " L " [n], it is supposed to be transliterated into IPA as [-jujen]. However, I supect that the semivowel " [ " [j] 153 is added to " - j " [9] to show that " =| " [jg] represents the front vowel [e] since " " [3] alone is not suitable to represent the low front vowel [e]. In addition, since the

Chinese syllabic pattern includes one medial per syllable, and we use a maximum of two symbols ("j" and "w" (or "ju")) for one medial; we transcribe " as [-juen] instead of

[-jujen].

3. After the loss of the entering tone ending, the rhyme base /-et/ of the former entering tone words of the

Xian rhyme-set changed /-e/, which later became the rhyme base of the Zhuo rhyme /-e/ of DYTJ.

(1) The final type " ^ " [-jet] (/-jet/) in the CP notation of the words such as jie, tie, and tjj qie became " ^ " [-je] (/-je/) in the PVP notation.

(2) The final type " 7T^ " [-juet] (/-jwet/) in the CP notation of such words as ^ que, jue, and shuo became " " [-jue] ([-jwE] /-jwe/) in the PVP notation.

4. In sum, the words of the Xian rhyme-set have the following final types:

(1) With the non-entering tone words, [-jen] ([-j£n]

/-jen/) and [-juen] ([-jwen] /-jwen/) occur in both the CP,

VP, and PVP notations. The phonemic representations of these finals in MM are /-jan/ and /-jwan/, respectively.

(2) With the entering-tone words, [-jet] (/-jet/) and 154

[-juet] ([-jw£t] /-jwet/) occur in the CP notation, while

the ending [-t] (/-t/) is replaced by [-?] (/-?/) in the VP

notation. These final types are changed to [-je] ([-jE]

/-je/) and [-jue] ([-jwg] /-jwe/) in the PVP notation.

5.1.4 The Han rhyme-set

The Han rhyme-set of HWZY consists of all the words

of the Huan rhyme (which has the final type /-won,t/) in

division 1 of the 'rounded' type and the words with a

guttural initial consonant of the Han ^ rhyme (which has

the final type /-on,t/) in division 1 of the 'unrounded'

type, both of the Shan rhyme class in LMC. The Han rhyme-

set of HWZY is reconstructed with the final types /-(w)on,t/

(Chou 1989:156). This rhyme-set corresponds to the Huan-

huan rhyme and the words with a guttural initial consonant

of the Han-shan rhyme of ZYYY, and they are respectively

reconstructed as /-on/ and /-an/ (Hsueh 1975:53-54).

The words from the Huan and Han rhymes of LMC are

respectively transcribed with the Korean vowels " " [w@]

and " " [9] in the CP notation. This transcription

suggests that the nucleus of the HWZY Han rhyme-set was

probably close to the mid-central vowel [a] phonetically.

Thus, I reconstruct the lower-mid back vowel [o] as the phonetic value of the nucleus of the Han rhyme-set words of

HWZY, and loi as its phonemic value (cf. Table 26). 155

In the VP and PVP notations, the words from the LMC

Huan rhyme are still transcribed with the Korean vowels

" r| " [w9] and " -| " [a]. For this reason, we are prone to the assumption that the nucleus of these words remained [o ]

/o/ in the VP and PVP notations. Nevertheless, we find an important description about the phonetic quality of the nucleus of these words in the section of the Han rhyme-set in SSTH. ^ f f # y " ir g:, “f-/I K , ^ ” f- ", t K i fè-i- ^ 4-4^ t T ,

"In the even, rising, and going tone categories, the words transcribed with the vowel " -j " [3] are pronounced as " j~ "

[a] not only in the vulgar pronunciation and in the Meneau vunlue but also in the present vulgar pronunciation.

However, in order to avoid the complexity of letter symbols,

I do not write the vulgar pronunciation under each character, but simply follow SSTG."

From this description we can infer that the nucleus of the non-entering tone words from the LMC Huan rhyme was actually

" j- " [a] already in the mid-fifteenth century. Therefore, we take [a] and /a/ as the phonetic and phonemic forms of these words for the VP and PVP notations. This change of vowel-fronting (which occurred from LMC to ZYYY) can be expressed by the following rule (Hsueh 1975:49, 63, and Chou 1989:26-27): 156

o — a / C ___ (E) (21)

(E=[-back]; If C = G, then E / p, t)

The words with guttural initiais from the LMC Han

rhyme are reconstructed with the final type /-an/ (Hsueh

1975). In addition, they are not only transcribed with the

final type " " [-an] /-an/ in the VP and PVP notations,

but also pronounced with the same final type in MP. That

is, these words maintained [-an] /-an/ as their final type

from OM through MP. Therefore, I take [a] and /a/ as the

phonetic and phonemic values of the nucleus of these words

for the VP and PVP notations.

On the contrary, in the CP notation these words are

transcribed with the final type " ^ " [-an] ([-jn] /-on/).

I consider that this final type existed in the sound system

underlying HWZY. In other words, the difference between

[-on] /-on/ (in the CP notation) and [an] /-an/ (in the VP

and PVP notations) in transcription of the finals of the words with a guttural initial from the LMC Han rhyme does

not reflect a diachronic sound change in Northern Mandarin, but it reflects a dialectal difference between the sound

systems of HWZY and ZYYY. To be specific, from LMC to HWZY, a different condition, namely, C = G was put on Rule (21) o — a / C ___(e ). Thus, the nucleus of these words remained as /-o-/ in the CP notation. 157 Let us now examine the pronunciations of the finals of the Han rhyme-set words.

Table 36 - Transcription of the Han rhyme-set words

gan guan ^ ban ge tuo

CP ^ [kan] ^ [kwgn] [pw3n] ^ [k9t] ^ [tw9t]

VP ^it [kan] " t! [pan] ÿ [ka?] [twa?]

PVP " " " [ka] S [to]

1. Words such as gan, ^ kan, ^ an, and ^ han are from the LMC Han rhyme with a guttural initial. The finals of these words are transcribed with " ^ " [-3-n] in the CP notation, and with " " [-an] in the VP and PVP notations. The final type of these words is [-3n] /-on/ for the CP notation, and [-an] /-an/ for the VP and PVP notations.

2. Words from the LMC Huan rhyme such as guan,

duan, ^ suan, and ^ huan are all transcribed with

" 7^" [w3n] in all three notations. We take [-won] and

/-won/ as the phonetic and phonemic forms of this final type for the CP notation, but [-wan] and /-wan/ for the VP and

PVP notations (see the above discussion). 158

3. The words with labial initials from the LMC Huan rhyme such as ^ ban, ^ij pan, 7^ man are transcribed with

" " [-wan] ([-wjn] /-won/) in the CP notation, and with

" -[ " [-an] ([-an] /-an/) in the VP and PVP notations (see the above discussion). This transcriptional change caused by the weakening of labialization of the he-kou medial /-w-/ after the labial initials by Rule (16) w — 5> ^ / B ___ (in

Section 5.1.1) and the fronting of the nuclear vowel by Rule

(21) o — > a / ___ (E) (in this Section).

4. Following the loss of the entering tone ending, the finals of the former entering tone words changed as follows :

(1) The finals of the words with guttural initials such as ge and 7^ ke are transcribed as " ^ " [-at] ([-Jt]

/-ot/) in the CP notation, as " d " [-3?] ([-Jf] /-o?/) in the VP notation, and " " [-@] ([-3] /-o/) in the PVP notation. Since these words have the final type [-f] /-d/ in MP, I suspect that the vowel [-3] was raised to [-f] following the deletion of the entering tone ending.

[3] — [f] / G ___ # o 3 / G ___ # (22)

(2) The finals of such words as tuo, ^ bo, 7^ huo are transcribed with " 7^" [-wat] ([-wot] /-wot/) in the

CP notation, as " " [-wa?] ( [-wo^ /-wo^) in the VP notation. Although the final type [-wo] /-wo/ is expected. 159 curiously we find the final type [-o] /-o/ in the PVP notation. This phenomenon reflects a wide-spread sound change in Mandarin. The distinction between /-wo#/ and

/-o#/ is lost by either (1) the addition of /-w-/ to /-o#/ or (2) the loss of /-w-/ before /-o#:

gf — > w /C o # (C / G) (23)

w — > ^ /C o # (C G) (24)

In general, case (1) occurs in Northern Mandarin and

Northwestern Mandarin, and coincides with the "colloquial pronunciation"; whereas case (2) occurs in Lower Yangzi

Mandarin and Southwestern Mandarin, and coincides with the

"literary pronunciation". The changes seen in ZYYY and DYTJ follow case (1 ), while those in HWZY and the Yunlue huitona follow case (2).^*

The final type " j_" [-o] /-o/ in the PVP notation follows case (2), which incidentally is an example of the

"literary pronunciation". In addition, the retention of the entering tone ending [-?] /-?/ in the VP notation is also a typical example of the "literary pronunciation". In this way, the VP and PVP notations follow mainly the "literary pronunciation" in transcription of the entering tone words, notwithstanding the fact that the VP and PVP notations respectively represent mid-fifteenth century and early sixteenth century Northern Mandarin (cf. Section 2.2.2)

Such transcriptional evidence suggests that the "literary 160 pronunciation" was prevalent due to dialectal borrowing in

Northern Mandarin at that time.

5. In sum, words of the Han rhyme-set have the following final types:

tllV With the non-entering tone words, [-3n] (/-on/) and

[-won] (/-won/) occur in the CP notation, but they are changed to [-an] (/-an/) and [-wan] (/-wan/) in the VP and

PVP notations.

(2) With the entering tone words, [-jt] (/-ot/) and

[-wTt] (/-wot/) occur in the CP notation, but the ending is changed by [-?] (/-?/) in the VP notation. In the PVP notations, these final types became [-3] /-o/ after guttural initials, and [-o] /-o/ due to the neutralization of /-o#/ and /-wo#/.

5.2 The rhyme-sets Dona. Gena, and Yana

These three rhyme-sets in HWZY have the velar nasal stop ending /-%/ for the rhymes with a non-entering tone or the velar stop ending /-k/ for the rhymes with the entering tone. The endings of the entering tone words of these three rhyme-sets are transcribed as " 7 " [-k] in the CP notation, as " o’ " [-?] (or " y " [-W] for the Yao ^ rhyme words) in the VP notation, and as zero in the PVP notation. This change in transcription of the entering tone words can be summed up by the following rules. 161 k -■> 9 / V ____ # (25)

? _-► ^ / V ____ # (15)

k — w / V ____ # (26)

Rule (25) precedes Rule (15) (first appeared in Section

5.1), and these two rules occur in the "literary pronunciation". Rule (26) occurs in the "colloquial pronunciation".

5.2.1 The Dona rhyme-set

The Dona rhyme-set comes from the LMC Tona rhyme class, which has only two types of finals: the /-wig,k/- type for division 1 and /-jwig,k/-type for division 3. In the sound system of HWZY, LMC /-jwiT),k/-type syllables, except for the syllables with guttural initials, are in the process of losing their palatal medial /-j-/ and thus merging into the /-wig, k/-type syllables with the same initials (Chou 1989:159-164). 162

Now, let us examine the pronunciation notations of the Dong rhyme-set syllables.

Table 37 - Transcription of the Dona rhyme-set words

gong ^ feng kong zhong ^ qiong

-f- CP [kung] •f'O [fug] f [khjug] f [tfjag] f [kHjug]

2u 4- VP ■f [khug] T [tpug] à [khjug]

PVP f "

du zu Yl zhu ^ qu îE» yu

Sc. CP T [tHuk] ^ [tsjuk] [tgjuk] [khjuk] ? [gjuk] 4- VP f [tHu?] O [tsju ] f" [tgju?] ■ f [khju?] % [gju?]

c: PVP T [tu] T* [tsu] ^ [tgu] [khju] [ju]

/S 1. The words from the LMC /—'wig/-type, e. g. / ^ gong, tong, ^ peng, and yfi are transcribed uniformly with the final type " "J” " [-ug] (/-wig/) in all three pronunciation notations.

The words from the LMC /-jwig/-type with labiodental initials such as M feng (LMC /fjwig/) and feng (LMC

/fHjwig/) are also transcribed in the same way as the LMC

/-wig/-type words. This fact shows that the palatal medial

/-j-/ preceded by the labiodental initials was lost before 163

the time of HWZY by Rule (5) j / Bf ___ w (in Section

4.3.2).

2. The words from the LMC /-jwiirj/-type with

guttural, dentalveolar sibilant, and retroflex sibilant

initials, e.g., ^ gong, ^ kong, cong, zhong, and j-1 chong, are transcribed with the final type " J" " [-jujj]

(/-jwig/) in the CP notation, but they are transcribed with

the final type " J* " [-ug] (/-wig/) in the VP and PVP

notations. This transcriptional change shows that the

palatal medial /-j-/ preceded by gutturals, dentalveolar

sibilants, and retroflex sibilants was also deleted before

the VP notation. Thus, most of the /-jwig,k/-type syllables

lost the palatal medial /-j-/, and thus joined their

/-wig,k/-type counterparts before the mid-fifteenth century.

We can sum up the process of the deletion of the palatal

medial /-j-/ in the Dona rhyme-set words by the following

rule:

j — f / ___ wig,k (27)

Accordingly, the retroflex sibilant initials of the Dona

rhyme-set words must have had only the [tp]-type at the time

of the VP notation, for the retroflexion of initial

consonants preceded the deletion of the palatal medial /-j-/

(of. Section 4.3.B.5). However, some words with guttural

initials, e.g., ^ qiong, ^ xiong, rong, and qu maintain the final type " ‘g ’ " [-jug] (/-jwig/) in all three 164

pronunciation notations. These exceptions incidentally

seems to support the claim that a phonological change

spreads lexically and gradually (cf. Chen, M. and W. Wang

1975).

3. The deletion of the palatal medial /-j-/ affects

the entering tone words, too. Due to the loss of the

entering tone ending, the former entering tone words of the

Dona rhyme-set get two new final types: " T " [-u] (/-wi/)

and " 7T " [-ju] (/-jwi/).

(1) The finals of the words from the LMC /-wik/-type words, e.g., ^ ku, -f' du, h bu, wu, are transcribed

as " " [-uk] (/-wik/), " % " [-u?] (/-wi?/), and " T "

[-U] (/-wi/) respectively in the CP, VP, and PVP notations.

In addition, the /-jwik/-type words with labial and

labiodental initials, e.g., Ü (LMC /mjwik/), (LMC

/fjwik/), and ^ (LMC /fHjwik) lost the palatal medial

/-j-/ before HWZY (cf. no. 1 in this Section), and therefore the finals of these words are transcribed in the same way as those of the /-wik/-type words.

(2) Words from the LMC /-jwik/-type are transcribed with the final type " 7^ " [-juk] (/-jwik/) and " ^ " [-ju?]

(/-jwi?/) respectively in the CP and VP notations. In the

PVP notation, some of these words with dentalveolar and retroflex sibilant initials, e.g., ^ zu, tT zhu, ^ chu, are transcribed with the final type " T " [-u] (/-wi/) (cf. 165

no. 2 in this Section), while the other words, e.g., (ÿ

qu, yu, su, are transcribed with the final type

" 7T " [-ju] (/-jwi/). To sum up, the deletion of the palatal medial /-j-/

in the /-jwig,k/-type words started first in the words with

labial initials before the time of HWZY, spread gradually through words with initials of different types, and finally affected the words with guttural initials. Thus, most of the /-jwig,k/-type words merged with their /-wiy,k/-type counterparts except a limited number of residues, e.g.,^ qiong, rong, qu, su, at the time of PN and PP

(or the early sixteenth century).

(1) With the non-entering tone words, [-ujj] (/-wig/) and [-jug] (/-jwi^/) occur in the CP, VP, and PVP notations.

(2) With the entering tone words, [-uk] (/-wik/) and

[-juk] (/-jwik/) occur in the CP notation, but the ending

[-k] (/-k/) is replaced by [-?] (/-?/) in the VP notation.

These final types are changed to [-u] (/-wi/) and [-ju]

(/jwi/) in the PVP notation, respectively.

5.2.2 The Yana rhyme-set

The Yang rhyme-set of HWZY consists of the words from the Jiang /-ag,k/ and Dang /-(j)(w)og,k/ rhyme classes of

LMC. The LMC Jiang rhyme class /-ag,k/ merged into the Dang rhyme class /-( j ) (w)o/),k/ by the following rule of vowel- 166 backing in late period rhyme-table books such as SSDZ and

QYZZT (Hsueh 1985:46 and Chou 1989:28):

a o / g ,k (28)

Thus, the nucleus of the HWZY Yana rhyme-set words is reconstructed as the low back vowel phoneme /-o-/ (Chou

1989: 164-167). Also, the nucleus of the ZYYY Jiana-vana rhyme, which corresponds to the HWZY Yana rhyme-set, is also reconstructed as the low back phoneme f-o-l (Hsueh 1982 and

1986).

The notations of the Yana rhyme-set words include the

Korean vowels " t '' [a], " ij-” [wa], and " j= " [ja] in the

CP notation. That is, the nucleus of the Yana rhyme-set words is transcribed consistently with the Korean vowel

" j- ", which is the only low vowel phoneme /a/ in Middle

Korean (see Table 12 in Section 4.1). For this reason, the vowel letter " f- " is used to transcribe both a low central vowel, [a], and a low back vowel, This, therefore, does not contradict the fact that the nucleus of the HWZY

Yana rhyme-set and that of its corresponding rhyme in ZYYY, i.e., the Jiana-vana rhyme, have been both reconstructed as the low back vowel phoneme /-o-/. 167

Now, let us examine the pronunciation notations of

the Yana rhyme-set words.

Table 38 - Transcription of the Yana rhyme-set words

. 1%. gang ■p dang wang zhuang /<_; gwang

CP Ÿ [kag] ÿ[tag] ÿ ■[vag] Ÿ [tgwa ] [kwag]

VP 1t "

æ /* PVP [wag] A

7-^ jiang T zhang ft tuo fh pu È & he

CP Ÿ [kjag] Ÿ [tgjag] [thak] 4" [Phak] ^ [xHak]

VP S [thaw] Ÿ [Phaw] [xHaw]

PVP 5 .[tho] Î: [pho] ^ [xa]

^ ge % zhuo ^ xue ^ yao ^ que

CP [kak] ^[tgwak] ^[xHjak] 4" [jak] [khjak]

eh VP I [ka?] ^ [tgwaw] ÿ[xHjaw] f [ jaw] ^ [khaf]

PVP ^ [ka] ^ [tgo] -2- [xjo] -2. [jo] ^ [ka]

1. Words from the LMC /-og,k/-type such as gang,

^ dang, and ^ cang are transcribed with the final type 168 " J~ " [-ag] in all three pronunciation notations. We consider [-ag] and /-og/ to be the phonetic and phonemic forms of this final type for OM; but [-^^] /-ag/ for MM, for only one low vowel phoneme, viz., /a/, occurs before a non­ zero ending in MM and MP.

The words with labial initials from the LMC /-ag,k/- type, e.g., bang and ^ bang are also transcribed with the same final type in all three pronunciation notations.

This shows that the LMC /-ag,k/-type words with labial initials merged into the words from the LMC /-og,k/-type words before the time of HWZY by Rule (28) a -9» o / ___^,k

(in this section).

2. Words with the Wei /v-/ initial from the LMC

/-jwo,k/-type, e.g., t wang, wang, ^ wang are transcribed as " ^ " [vag] in the CP and VP notations, but it is changed to " " [wag] in the PVP notation. We take

[v&g] and /vwog/ respectively as the phonetic and phonemic forms of the final type of these words for the CP and VP notations, but [-w4g] and /-wag/ for the PVP notation. When these words had the initial /v-/ in OM, the labial medial

/-w-/ was very weak phonetically, and so they were transcribed in Korean as [vag].‘® After the deletion of the initial /v-/ by Rule (6) v — ^ / # _ (in Section 4.3.2), the labial medial /-w-/ of these words was pronounced distinctively, and these words were, therefore, transcribed as [wag]. 169 3. The words with Pr[l] initials from the LMC

/-jog/-type (e.g., zhuang, chuang, and ^ shuang) and the words with Pr[l] initials fromthe LMC

/-ag/-type (e.g., ^ chuang and ^ shuang) are both transcribed as " J" " [-ag] ([-ag] /-og/) in the CP notation.

This reveals that the former type words and the latter type words coalesced before OM in the following way;

(1) The final type /-jog/ lost its medial by Rule (10) j -T» ^ / r ___ (in Section 4.3.3.5), and became /-oi)/.

(2) The final type /-ag/ also became /-og/ by Rule (28) a — > o / g ,k (in this Section).

The final type of these words are changed to " " [-wag]

([wa^] /-wag/) in the VP and PVP notations. This transcriptional change shows that the final type /-og/ preceded by retroflex initials acquired the labial medial

/-w-/ before the time of the VP notation.

^ — w / r_____ og (29)

4. The words with guttural initials from the LMC

/-wo /-type (e.g., ^ guang, ^ huang, and huang) and those from the LMC /-jwog/-type (e.g., ^ kwang, ^ wang, and 5 . wang) are transcribed with the final type " "

[-wag] in all three pronunciation notations. These words have the same pronunciation in ZYYY and MP. We consider

[-wag] and /-wog/ to be the phonetic and phonemic forms of this final type for OM, but [-wag] and /-wag/ for MM and MP 170 since phonemically only one low vowel occurs before a non­ zero ending in MM and MP. The pronunciation notation of this final type shows that the /-jwo)J/-type words merged into the /-wog/-type words before OM by the loss of the palatal medial /-j-/.

i — ^ / G ___ wo/), k (30)

5. The words from the LMC /-jo/|/-type (e.g., qiang, zhang, and ^ yang) and the words with guttural initials from the LMC /-ag/-type (e.g., -/X jiang, ^ qiang, ^ xiang, and ^ yang) are transcribed with the final type " / " [-jag] in all three pronunciation notations. This phenomenon reveals that the /-ag/-type first acquired the palatal medial /-j-/ by Rule (18) fi -4» j / G a (in Section 5.1.2), and later its nuclear vowel merged to f-o-l by Rule (30) a — *■ o f g ,k

(in this Section). In this change. Rule (18) must have preceded Rule (30) since /-a-/ is a conditioning factor in

Rule (18).

The words with Py initials from the LMC /-jog/-type

(e.g., ^ zhang, g chang, and shang) keep the palatal medial /-j-/ in all three transcriptions, while the words with Pr[l] initials from the LMC /-jop/-type (e.g., zhuang, chuang, and 0 shang) are without the palatal medial /-j-/ in all three transcriptions (see no. 3 in this 171 section). This phenomenon confirms that the Pr[l] initials became the [tp-] type before the time of HWZY.

6 . In the CP notation, the entering tone words of the Yang rhyme-set, namely, the Yao rhyme words, are transcribed uniformly with the velar stop ending " 7 " [-k] in line with the "literary pronunciation". In the VP notation, most of them are transcribed with the semivowel ending " ^ " [-w] (cf. Table 15) in line with the

"colloquial pronunciation" (e.g., ^ [xHj«iw] and ^

[ ja.w] ), while some of them are transcribed with the glottal stop ending " T " [-?] (e.g., ^ [ka?] and ^ [kha?]) in line with the "literary pronunciation". This fact reveals that, although the "literary pronunciation" was prevalent for the former entering tone words due to dialect borrowing, the Yao rhyme words were pronounced mostly in line with the

"colloquial pronunciation" in Northern Mandarin about the mid-fifteenth century. In the PVP notation, however, all the Yao rhyme words are transcribed with " " [-o#] and

" y " [-9#], in line with the "literary pronunciation".

This flip-flop in transcription proves that the "literary pronunciation" became more popular than the "colloquial pronunciation" even in Northern Mandarin about the beginning of the sixteenth century.

(1) Words from the LMC /-ok/-type (e.g., ^ tuo, bo, ^ mo, and ^ luo) are transcribed with the final types " -I*" [-ak] (/-ok/), " ^ " [-aw] (/-ow/), and " " 172

[-0 ] (/-o/), respectively in the CP, VP, and PVP notations.

However, there are some exceptions. The word he appears with the final type " -j " [-g] (/-o/) in the PVP notation.

The word ^ ge appears with the final types " ^ " [--J?]

(/-O'?/) and " -f " [-j] (/-o/) respectively in the CP and PVP notations.

(2) Words from the LMC /-wok/-type (e.g., zhuo and

shuo) appear with the final types " [-wak] (/-wok/),

" " [-waw] (/-wow/), and " j_" [-o] (/-o/) respectively in the CP, VP, and PVP notations.

(3) Words from the LMC /-jok/-type (e.g., ^ jiao, ^

zhao, ^ xue, ^ yao) are transcribed with the final types " ^ " [-jak] (/-jok/), [-jaw] (/-jow/) and [-jo]

(/-jo/) respectively in the CP, VP, and PVP notations. The word if que, however, appears with the final type " "

[-9?] (/-o?/) and " -j " [-3] (/-o/) in the VP and PVP notations.

7. To sum up, the following final types occur in the words of the Yana rhyme-set:

(1) With the non-entering tone words, [-ag] ([-#g]),

[-wag] ([-wag]), and [-jag] ([-jag]) occur in the CP, VP, and PVP notations. We take /-og/, /-wog/, and /-jog/ respectively as the phonemic forms of these final types for

OM, but /-ag/, /-wag/, and /-jag/ for MM.^ 173 (2) With the entering tone words, [-ak] ([-(

[-wak] ([-w#k] /-wok/), and [-jak] ([-jak] /-jok/) occur in the CP notation, [-aw] ([-aw] (/-ow/), [-waw] ([-w a w ]

/-wow/), [-jaw] ([-jaw] /-jow/), and [-3?] ([-D?] /-o?/) occur in the VP notation. [-3] ([-3] /-o/), [-jo] (/-jo/), and [-0 ] (/-o/) occur in the PVP notation.

5.2.3 The Gena rhyme-set

The HWZY Gena rhyme-set consists of the words from the LMC Gena rhyme-class (which has the final types

/-(w)a]j,k/, /-j(w)a/),k/, and /-j(w)e^,k/) and the LMC Zena rhyme-class (which has the final types /-(w)o^,k/ and

/-j (w)oyj,k/. First, the LMC Gena and Zena rhyme classes coalesced into one rhyme class, and thus the two rhyme classes appear as a single rhyme class in SSDZ and QYZZT.

This change can be formulated as follows (cf. Hsueh 1985:45 and Chou 1989:168):

o -* a / j(w) k (31)

Second, the nuclear vowels of the combined rhyme class coalesced into one vowel, and thus the combined rhyme-class formed a single rhyme-set before the Mandarin period, namely, the Gena rhyme-set of HWZY or its corresponding ZYYY rhyme, viz., the Gena-aina rhyme. The HWZY Gena rhyme-set and the Gena-aina rhyme of ZYYY are both reconstructed with 174 the low front vowel phoneme /-&-/ as their nucleus (Hsueh

1975:54 and Chou 1989:168).

In ZYYY/ some words of the Gena-aina rhyme also appear in the Dona-zhona rhyme (which corresponds to the

HWZY Dona rhyme-set). For example, words from the LMC

A A /-wo/l,k/-type (e.g., & hong, hong), /-wa^,k/-type

(e.g.f^hong, hong), and /-jwa/j,k/-type (e.g., ^ xiong) belong to the Gena-aina rhyme, but they also appear in the

Dona-zhona rhyme. This phenomenon suggests that the words of the Gena-aina and Dona-zhona rhymes shared the same nuclear vowel phoneme in OM. The ZYYY Dona-zhona rhyme and the HWZY Dona rhyme-set have been both reconstructed with the high central vowel /*/ as its nucleus (see Section

5.2.1). More importantly, the HWZY Gena rhyme-set words are transcribed with the Korean high vowels (i.e., " | " /i/,

" — " /i/, and " T " /u/) as their nuclear vowels. In addition, the HWZY Dona rhyme-set words are transcribed with the Korean high back vowel " t " [u]. Therefore, I consider that the HWZY Gena rhyme-set must be reconstructed with the high vowel /i/ as its nucleus.""

Consequently, we can postulate the following sound changes in order to explain this phenomenon:

First, the different vowels of the combined Gena-zena rhyme class coalesced into one as a result of assimilation to the palatal ending /-j/. 175

Second, some time later, the nucleus of the Gena rhyme- set (or the Gena-aina rhyme) was raised from /e/ to /i/, again conditioned by the palatal ending as the Korean notations show.

Third, still later, the feature [+palatal] was lost, and only then, the ending became plain velar. Consequently, the

Dona and Gena rhyme-sets joined into one as the VP and PVP notations show.

Therefore, we now can formulate the formation of the HWZY

Gena rhyme-set (or the ZYYY Gena-aina rhyme) from the combined rhyme class of the LMC Gena and Zena rhyme classes by the following two rules of vowel-fronting and vowel- raising before the palatal nasal ending /-^/ or the palatalized velar stop ending /-k/:

a, o -4» e / ___ ^ ,k (32)

® — i / rj fk (33 )

In the CP notation, the Gena rhyme-set words are transcribed with the following final types: " " [-ig],

" -J" [-ijg], " [-ujg], " rrj" [-jujg], " -j " [-ik], " W " [-tjk], " 7 " [-Ujk], and " " [-jujk] (see Table 39 below). Upon close examination of these Korean transcriptions, we find out that the semivowel " j " [j] appears between the main vosel and the syllable ending except when the main vowel is " j " [i]. All these Korean 175 transcriptions are possible forms according to the Korean syllable structure. However, except " J " [-ig] and " -/ "

[-ik], the other six forms represent uncommon sequences of segments in Korean, and thus they are seldom used in Korean writing. Therefore, I have no doubt that the six letter compounds are used ad hoc for transcribing final types of the Gena rhyme-set words.

Yet, the six transcriptions of Chinese final types

(i.e., [-ijg], [-ujg], [-jujg], [-ijk], [-ujk], and [-jujk]) do not fit the Chinese syllable pattern, (C)(M)V(E); since only one segment (i.e., a semivowel, a consonant, or the zero ending) can occur in the syllable ending position.

Therefore, I suspect that the palatal semivowel " | " [j] is used to denote some additional phonetic quality of the A syllabic ending /-]]/ or /-k/. In this case, the semivowel

/j/ is used to add the feature [+palatal] to a velar ending."

Therefore, I consider that the syllable endings of the HWZY Gena rhyme-set words represented by the CP notation are palatal consonants.^® Coincidentally, this idea is in agreement with Hashimoto's (1970b) reconstruction of the endings of the LMC Gena rhyme class as palatal consonants, namely, the palatal nasal stop /-/I/ for the non-entering tone words and the palatal stop /-c/ for the entering tone words. In this study, following Hsueh (1985), we accept that both the LMC Gena and Zena rhyme classes have the 177 palatal endings, and replace the palatal stop ending /-c/ by

the palatalized velar stop ending /-k/.

Now, let us examine the Korean pronunciation notations of the Gena rhyme-set words.

Table 39 - Transcription of the Gena rhyme-set words

T ding 3^ ming ^ deng .'i. sheng peng

CP o [tig] d* [mig] [tijg] 6 [gijg] ^ [pHijg]

VP f [tig] ^ [gig] T [pHug]

PVP Î [&A?)]

f K heng yong qing li iii bei

CP [xHujn] ^ [ j u i ] [kjui ] ^ [lik] “J [Pijk]

a( VP [xlig] ] 1 [ki ] T [li?] T [ p i * ]

PVP T ZXUTJ] - ( [li] H [pa]

ge 1$ dei ^ zhai f to guo

CP ^ [kijk] [tijk] ^ [t^Hijk] ^/[pHijk] ^ [kujk]

VP ^ [kij?] ^[tij?] V[tfHij?] ^[pHij?] ^ [kuj?]

PVP 7f [kj@] [tij] [tgaj] "T [Pu] "rf [ku3] 178 1. Words from the LMC /-ja^,k/-type (e.g., ^ bing, .A ming, and % zheng), /-jep,k/-type (e.g., -J ding, ^

jing, and ^ ling), and /— jo)i,k/-type (e.g., ^ cheng,

m sheng, and xing) are transcribed with the final type

" J " [-ig] in all three pronunciation notations. Since the

phonemic compound /ji/ is phonetically realized as [i] in

Mandarin (see Section 3.3.3); we take [-i^] and /-jip/

respectively as the phonetic and phonemic forms of this

final type for the CP notation (see the above discussion),

but [-ig] and /-jiij/ for the VP and PVP notations.

Among the words with this final type, the words with A t retroflex initials from the LMC /-ja^,k/ and /-joyi,k/-types

lost the palatal medial /-j-/ by Rule (10) j — 0 / # r ___

(in Section 4.3.3.5) during the period between the PVP

notation and MP, e.g., ^ cheng ([tshig] /crhjig/ in OM and

MM — » [tghag] /crhig] in MP and ^ sheng ([gig] /srjig/ in

OM and MM — J- [gag] /srig/ in MP). A 2. Words from the LMC /-a^^k/-type (e.g., ^ geng

and leng) and /-o^,k/-type (e.g., ^ deng and ) are

transcribed with the final type " " [-ijg] in the CP

notation, but with the final type " ^ " [-ijj] in the VP and

PVP notations. We take [-i^] and /-±p/ as the phonetic and

phonemic forms of the final type in the CP notation, but

[-ig] and /-ig/ as the phonetic and phonemic forms of the

final type in the V? and PVP notations. The nuclear vowel

/i/ in these two final types is, in all likelihood, [3] 179 phonetically (see Section 3.3.3 and Section 5.1.1 no. 2).

A The words with retroflex initials from /-aji,k/-type, e.g., ^ zheng and sheng, are trnscribed with the final types " " [-ijg], " -g" " [-ig], and " ^ " [/\^] repectively for the CP, VP, and PVP notations. The vowel letter " % "

[A] in Korean notation denotes the Mandarin high vowel /i/ only when this is preceded by either a dentialveolar or retroflex sibilant intial; thus, it probably denotes [i] or

[1] when it occurs alone as a whole final, or [a] before a non-zero ending.

Words like peng and ^ meng are transcribed with the final type " ^ " [-ig] in the VP and PVP notations; while the word peng is transcribed with the final type

" "J" " [-ug] in the VP and PVP notations. We consider that these two final types both have the same phonemic representation /-wi^/. These words are with a labial A initial and from the LMC /-a^,k/-type. Labialization became weak after labial initials by Rule (16) w — f / B ___ (in

Section 5.1.1) before the time of VP notation. The final type " ^ " [-ig] shows the weakening of labialization of the labial medial /-w-/ after labial initials, while the final type " 'T " [-ug] shows the occurrence of labialization. In this way, these two different transcriptions show that phonetically the labial medial /-w-/ and the zero medial i-^-l are not distinguishable after labial initials. 180

3. The word heng from the LMC /-wa^,k/-type is

transcribed with the final type ” " [-uj^] in the CP

notation, but this final type is changed to " J" " in the VP and PVP notations. We take [-ujî] and /-wi^/ as the phonetic

and phonemic forms of this final type for the CP notation,

and [-ug] and /-wig/ for the VP and PVP notations.

4. Words from the LMC /-jwa^,&/-type (e.g., ^

xiong and yong) and /-jweji,k/-type (e.g., qing and

^ ying) are both transcribed with the final type " /gf "

[-jujg] /-jwip/) in the CP notation. In the VP and

PVP notations, the words from the LMC /-jwa^,k/-type are

transcribed with the final type " "g" " [-jug] (/-jwig/),

while the words from the LMC /-jwejj ,k/-type are transcribed

with the final type " J " [-ig] (/-jig/). That is, the

labio-palatal medial /-jw-/ in the words from the LMC A /-jwe^,k/-type lost the feature [+labial] (/-w-/) before the

time of the VP notation. A \ 5. Entering tone words from the LMC /-jak/, /-jek/,

and /-jok/-types, e.g., shi, di, ji, ^ ji, and

jj li, are transcribed with the final types " _j " [-ik]

([-ik] /-jik/, " J " [-i?] (/-ji?/) and " I " [-i] (/-ji/)

respectively in the CP, VP, and PVP notations. These words

retain the final type [-i] /-ji/ in MP; except for the words with a retroflex initial, which lost the palatal medial 181 /-j-/ by Rule (10) j — ^ / # r ___ (in Section 4.3.3.5)

during the period between the PVP notation and MP, and thus

have the final type [-%] /-±/ in MP.

\ 6 . Entering tone words from the LMC /-ak/-type

(e.g., ^ ke, zhai, and § bai), /-jak/-type (e.g..

It zhi and shi), /-ok/-type (e.g., te, bei, and

^ mo), and /-jok/-type (e.g., zhi) are transcribed

uniformly with the final types " " [-ijk] ([-ik] /-ik/)

and " y " [-ij?] ([-i?] /-i?/) respectively in the CP and

PVP notations. These words, however, have five different

final types in the PVP notation as follows:

(1) Words such as ^ ke, bei, and ^ se have the final type " -j " [9]. This change is in accordance with the

"literary" pronunciation. We take [-3] and /-o/ as the phonetic and phonemic forms of this final type. The nuclear vowel /i/ was lowered to /o/ after the loss of the entering tone ending.

(2) Words cuch as ge, ke have the final type

" =[ " [-je] ([-j£] /-je/). These words have a guttural initial, and come from the LMC /-ak/-type. These words first obtained the palatal medial /-j-/ by Rule (18)

^ — j / G ___ a (in Section 5.1.2), and later the nucleus

/a/ changed to /&/ due to a vowel-fronting shift after the palatal medial /-j-/.

(3) Words such as dei, zei, and hei have the final type " ^| " [-ij] (/-ij/) in the PVP notation. 182

This change is in accordance with the "colloquial" pronunciation.

(4) The word zhai has the final type " ^ " [-aj]

(/-aj/). This change follows the "colloquial pronunciation".

(5) The final of the word ^ bo is transcribed as 'V

" [-U] /-wi/. This notation shows the strengthening of labialization after the deletion of the entering tone ending.

7. The word ^ guo (from the LMC /-wok/-type) and

(from the LMC /-jwok/-type) are transcribed with the final types " if " [-ujk] ([-uk] /-wik/) and " ~d " [-uj?]

([-u?] /-wik/) respectively in the CP and VP notations.

These words are transcribed with the final type " " [-wa]

([-wD] /-wo/) in the PVP notation. Although /-wo/ and /-o/ are generally neutralized in Mandarin, the contrast between,

/-wo/ and /-o/ is usually maintained after guttural initials in MP (see Section 5.1.4 no. 4). /-wo/ is transcribed here by " -pj " [wa] since "t" [w] (a negative vowel) and " m " [o]

(a positive vowel) can not be combined graphically (see

Section 4.1).

8 . In sum, the Gena rhyme-set words have the following final types;

(1) With the non-entering tone words, [-ig] ([-i/l]

/-jiy/), [“ijgl /-i/>/), [-uj^] ([-up] /-wiyj/), and 183

[-jujg] ([-ju/l] /-jwijj/) occur in the CP notation, but the ending [-^] (/-/J/) is changed to [-^] (/-?;/) in the VP notation. In the PVP notation, [-ig] (/-jig/), [-ig]

(/-ig/), [-ug] (/-wig/), and [-jug] (/-jwig/) occur.

Besides, the final type " ^ " [-/\g] ([-9g] /-ig/) occurs after retroflex initials in the PVP notation.

(2) With the entering tone words, [-ik] ([-ik] /-jik/),

[-ijk] ([-ik] /-ik/), [-ujk] ([-uk] /-wik/) occur in the CP

A A notation, but the ending [-k] (/-k/) is changed to [-?]

(/-■?/) in the VP notation. After the complete loss of the entering ton ending, [-i] (/-ji/), [-a] ([-0] /-o/), [-j3]

([-jàj /-je/), [-ij] (/-ij/), [-aj] (/-aj/), [-u] (/-wi/), and [-wa] ([-wo] /-wo/) occur in the PVP notation.

5.3 The rhyme-sets Pin, Tan, and Yan

These three rhyme-sets in HWZY are reconstructed with the bilabial nasal stop ending /-m/ for the words with a non-entering tone or the bilabial voiceless stop ending /-p/ for the words with the entering tone. The endings of the entering tone words of these three rhyme-sets are transcribed uniformly as " W " [-p] in the CP notation, as

" Ô " [-?] in the VP notation, and as zero in the PVP notation. This change in transcription of the entering tone endings of these three rhyme-sets can be formulated by the following two rules in sequence: 184

p ? / V # (34)

f ^ / V ___ # (15)

Rule (15) appeared in Sections 5.1 and 5.2.

5.3.1 The Pin rhyme-set

The Pin rhyme-set— which comes from the LMC Shen rhyme class (i.e., /-jim,p/-type)— is reconstructed with the high central vowel /i/ as its nucleus (Chou 1989:173).

The ZYYY Pin-xun rhyme— which corresponds to the HWZY Pin rhyme-set— is also reconstructed with the high central vowel phoneme /i/ as its nucleus (Hsueh 1975:60).

Table 40 - Transcription of the Pin rhyme-set words

fp ^ zan shen /S' zen jin gin

CP [tgim] o" [gim] ^ [tsim] ^ [kirn] '■q [tsim]

VP ^ [tgin] é- [fin] ■ a [tsim] 7 ^ 1 [kin] ^ [tsin]

A PVP [tgAn] ^ [gAn] a [tsAm] % [tsin]

shen ^ shen & ji k li X ru

CP y [gim] ^ [fHim] ^ [kHip] ^ [lip] lj [%ip]

VP ^ [gin] [kHi?] % [li?] -nr [%ju7]

PVP li [fin] â [gAm] 7/ [ki] 2/ [li] ■‘r [%ju] 185

1. The finals of such words as ^ zan and ^ shen are transcribed as " ~a " [-im], " — " [-in], and " iL " [-An]

([-an] /-in/) respectively in the CP, VP, and PVP notations.

These words are transcribed with the ending " □ " [-m] in the CP notation, but the ending is changed to " t- " [-n] in the VP and PVP notations. This transcriptional change shows that the bilabial nasal stop ending [-m] merged into the alveolar nasal stop ending [-n] in Mandarin before the time of the VP notation (or the mid-fifteenth century).

m -» n / ___ # (35)

The Korean vowel letter " » " [A] represents the

Mandarin high vowel /i/ only when this is preceded Mandarin sibilants in the PVP notation, the final type of these words was probably pronounced as [-an] at the time of the PVP notation (or the early sixteenth century) (see Section 5.2.3 no. 2). However, the final of the word zen is transcribed with the ending [-m] in all three pronunciation notations; i.e., " ■g- " [-im] (/-im/) in the CP and VP notations, and as " ^ " [-Afn] ([-&m] /-im/). Since the word

zen forms a compound word, ^ (zenmo 'how'), together with another word, ^ ma; it is plausible that the bilabial nasal stop ending [-m] of the word zen remained unchanged in the VP and PVP notations due to assimilation to the initial [-m] of the following word me. 186

2. The finals of such words as ^ jin, qin, and

'j'^ shen are transcribed as " J " [-im] (/-jim/) in the CP

notation, but as " J_ " [-in] (/-jin/) in the VP and PVP notation. This transcriptional change is represented by

Rule (34) m — >■ n / ___ # (see no. 1 above). However, the final of the word ^ shen is transcribed as " J " [-im] (/-jim/) in the CP and VP notations, and as " ^ " [-Am] ([-am] /-im/) inthe PVP notation. Since the word ^ shen occurs in a compound word, ^ ^ (shenmo 'what'), the ending [-m] remained unchanged in th VP and PVP notations probably due to assimilation to the initial [m-] of the following word ^ me.

The words with a retroflex initial and this final type lost the palatal medial /-j-/ by Rule (10) j — / / # r ___ (in Section 4.3.3.5) during the period between the PVP noatation and MP, e.g., zhen ([tgin]

/crjin/ in MM — [tsan] /crin/ in MP) and shen ([pin]

/srjin/ in MM — [pan] /srin/ in MP.

3. The finals of the entering tone words, e.g., ji, /k. ji, ^ ji, ‘-f" shi, are transcribed as " J " [-ip]

(/-jip/), " J " [-i?] (/-ji?/), and " I " [-i] (/-ji/) respectively in the CP, VP, and PVP notations. These words, except for the words with a retroflex initial, retain the same final type in MP as in PVP.

The words with a retroflex initial and this final type, e.g., ^ zhi and -f- shi lost the palatal medial 187

/-j-/ by Rule (10) j ^' / # r ___ (in Section 4.3.3.5),

and thus have the final type [-%] /-i/ in MP.

The word ru is transcribed as " " [%ip], "^ " 0 [^ju?], and " " [gju] respectively in the CP, VP, and PVP notations. We take [jip] and /rjip/ as the phonetic and phonemic forms of this word for the CP notation; [jju?] and

/rjwi?/ for the VP notation; and [Jiu] and /rjwi/ for the

PVP notation. First, this word acquired the labial medial

/-w-/ before the time of the VP notation; this word underwent an exceptional change probably because it is a

"taboo" word. Second, the glottal stop ending [-?] became zero before the time of the PVP notation. Presumably, the initial of this word became [|^-] phonetically, and subsequently the palatal medial /-j-/ was deleted during the period between PVP and MP. Thus, the phonetic and phonemic forms of this word are [fu] and /rwi/ in MP.

4. In sum, the Pin rhyme-set has the final types as follows ;

(1) With the non-entering tone words, [-im] (/-im/) and

[-im] (/-jim/) occur in the CP notation. These two final types are changed to [-in] (/-in/) and [-in] (/-jin/) respectively in the VP and PVP notations.

(2) With the entering tone words, [-ip] (/-jip/) and

[-i?] (/-ji?/) respectively occur in the CP and VP notations. Due to the loss of the entering tone ending. 188 this final type is changed to [-i] (/-ji/) in the PVP notation.

5.3.2 The Tan and Yan rhyme-sets

The Tan and Pin rhyme-sets in HWZY both come from the

IiMC Xian rhyme class, which has the final types /-om,p/,

/-am,p/, /-jam,p/, and /-jem,p/. Division 1 (/-om,p/-type) and division 2 (/-am,p/) of the LMC Xian rhyme class formed the Tan rhyme-set, and in addition the words from division 3

(/-jam,p/ with labiodental initials which lost the palatal medial /-j-/ before the Mandarin period (of. Section 4.3.2) also joined the HWZY Tan rhyme-set. The Yan rhyme-set comes from division 3 (/-jam,p/) and division 4 (/-jem,p/) of the

LMC Xian rhyme class.

The Tan rhyme-set and the Yan rhyme-set of HWZY are reconstructed with the final types /-am,p/ and /-jem,p/, respectively (Chou 1989:173-179). The ZYYY Jian-xian rhyme

(which corresponds to the HWZY Tan rhyme-set) and the ZYYY

Lian-xian rhyme (which corresponds to the HWZY Yan rhyme- set) are reconstructed as /-am/ and /-em/, respectively

(Hsueh 1975:60). 189

Now, let us examine the Korean pronunciation notations of the words of the Tan and Yan rhyme-sets.

Table 41 - Transcription of the words of the Tan and Yan rhyme-sets

^ gan tan H san & jian iX. qian

CP t7 [kam] ^ [tHam] 'it [sam] ^ [kjam] [khjam]

VP Ÿ [kan] ^[tHan] ^[san] ^ [kjen] [khjan]

PVP ^[than] lir [san] [khjen]

da yè fa la «Z. he f jia

CP ^ [tap] [fap] ^ [lap] ^ [ x H a p ] [kjap]

VP ^[ta?] f [fa?] Ÿ [la?] ^ [XH3?] t [kjaf]

PVP 4 [ta] ri/'[fa] Zf[la] Pf [xH ] 71: [kja]

k qian t' nian yan # jie Ê ye

CP ^ [khjem] ^[njem] ^ [ jem] [tsjep] i [jep]

VP ^[khjen] Ll[njen] ^ [ jen] [tsje?l i [je?]

PVP ^[tsje] *f[je]

NOTE: The Korean diphthong " =j ’’ [j@] is transcribed as [je] in IPA here (see Section 5.1.2 no 2. and Section 5.1.3). 190

1. Words from the LMC /-om,p/-type (e.g., ^ gan,

^ kan, -S. san, and ^ Ian) and words from the LMC

/-am,p/-type (e.g., shan) are transcribed with the final type " J- " [-am] (/-am/) in the CP notation, but this final type is changed to " J* " [-an] (/-an/) in the VP and PVP notations. This change of syllable endings is represented by Rule (34) m n / ___ # (in Section 5.3.1).

The word ^2, (LMC /fHjwam/ in division 3 of the

Xian rhyme class) also has the same final types (i.e., " h "

[-am] (/-am/) in the CP notation and " " [-an] (/-an/) in the VP and PVP notations). This phenomenon reveals that in this word first the feature [+palatal] was deleted by Rule

(5) j — > 0 / Bf ___ w (in Section 4.3.2), and subsequently the labialization of the labial medial /-w-/ was weakened by

Rule (16) w — > / B__ (in Section 5.1.1). Thus, the final of this word became homophonous with that of the words from division 2 /-am/-type in HWZY.

The words with a guttural initial from the LMC

/-am,p/-type (e.g., ^ jian and ^ xian) are transcribed with the final type " " [-jam] (/-jam/) in the CP notation, but this final type is changed to " ^ " [-jen] in the VP and PVP notations. This transcriptional change reflects two sound changes as follows;

First, the LMC final type /-am,p/ became palatalized by Rule

(18) ^ -* j / G ___ a (in Section 5.1.2) before HWZY.

Second, the nucleus /-a-/ changed to /-e-/ by Rule (19) 191 a -» e / j ___ n (in Section 5.1.2) before the time of the

VP notation.

Thus, these words have the same final type as that of the words from the LMC /-jem,p/ (see no. 2 below)

2. Words from the LMC /-jam,p/-type (e.g., qian,

j^chan, and ^ shan) and words from the LMC /-jem,p/-type

(e.g., ^ nian, ^ jian, ^ yan) are transcribed with the final type " ^ " [-jem], but this final type is changed to

" [-jen] in the VP and PVP notations. We take /-jem/ as the phonemic representation of this final type for OM; but

/-jan/ for MM since phonemically only one low vowel /a/ occurs in MM (cf. Section 5.1.2 no. 2 and Section 5.1.3).

The words with a retroflex initial and this final type lost the palatal medial /-j-/ by Rule (10) j > ^ / # r ___ (in

Section 4.3.3.5) during the period between the PVP notation and MP, e.g., ([tghjan] /crhjan/ in MM — > [tphan]

/crhan/ in MP) and ^ shan ([pjen] /srjan/ in MM --y [^an]

/sran/ in MP).

3. Entering tone words from the LMC /-op/-type

(e.g., ^ da, na, and za), /-ap/-type (e.g.,'^/J zha, and cha), and /-jap/-type with labiodental initials

(e.g., fa and ^ fa) are transcribed with the final types " J' " [-ap] (/-ap/), " ^ " [-a?] (/-a?/), and " "

[-a] (/-a/) respectively in the CP, VP, and PVP notations.

The words with a guttural initial from the LMC 192 /-op/-type such as ke and ^ he are transcribed with the final types " " [-ap] (/-ap/), " j " [-3 ?] ([-?/]

/-0 7 /), and " -( " [-a] ([-3] /-o/) in the CP, VP, and PVP notations. This final type became [-/] /-a/ in MP by Rule

(22) o — 9 / G __ # (in Section 5.1.4). The words with a guttural initial from th LMC /-ap/-type such as ^ jia and

qia are transcribed with the final types " " [-jap]

(/-jap/), " / " [-ja?] (/-jar/), and " /= " [-ja] (/-ja/) in the CP, VP, and PVP notations. This final type became palatalized by Rule (18) yi — » j / G___ a (in Section

5.1.2) before the OM period.

4. Entering tone words from the LMC /-jap/ and

/-jep/-types, e.g., que, she, tie, and ye, are transcribed with the final types " ^ " [-jep] ([-j&p]

/-jep/), " :d " [-je?] ([-j£?] /-je?/), and " " [-je]

([-je] /-je/) in the CP, VP, and PVP notations. The words with a retroflex initial and this final type first lost the palatal medial by Rule (10) j — > ^ / # r ___ (in Section

4.3.3.5), and susequently the nucleus of these words was raised as formulated by the following rule, e.g.,

([tf jg] /crje/ in MM —► [tgr] /era/ in MP) and ■yiy ([gj&]

/srje/ in MM -» [§Y] /sra/ in MP).

e — 0 / r ___ # (36)

5. In sum, the words of the Tan and Yan rhyme-sets have the following final types; 193 (1) With the non-entering tone words, [-am] {/-am/),

[-jam] (/-jam/) and [-jem] ([-jam] /-jem/) occur in the CP notation, [-an] (/-an/) and [-jen] ([-jen] /-jen/) occur in the VP and PVP notations.

(2) With the entering tone words, [-ap] (/-ap/), [-jap]

(/-jap/), and [-jep] ([-j6p] /-jep/) occur in the CP notation. [-a?] (/-a?/), [-;??] ([-37] /-o?/), [-ja?]

(/-ja?/), and [-je?] ([-j£‘?] /-je?/) occur in the VP notation, [-a] (/-a/), [-a] ([-3] /-o/), [-ja] (/-ja/), and

[-je] ([-j&] /-je/) occur in the PVP notation.

5.4 The rhyme-sets Xiao. Yao. and You

These three rhyme-sets in HWZY have the semivowel ending /-w/. The HWZY Xiao and Yao rhyme-sets come from the

LMC Xiao rhyme class. Division 1 (/-ow/-type) and division

2 (/-aw/-type) of the LMC Xiao rhyme class formed the HWZY

Yao rhyme-set; division 3 (/-jaw/-type) and division 4

(/-ew/-type) of the same rhyme class formed the HWZY Xiao rhyme-set. The You rhyme-set comes from the LMC Liu rhyme class, which has only two final types, i.e., the /-iw/-type for division 1 and the /-jiw/-type for division 3 (see Table

27).

The Xiao. Yao. and You rhyme-sets of HWZY are reconstructed with the final types /-jew/, /-(j)ow/, and

/-(j)iw/, respectively (see Tables 30 and 31 and Chou

1989:180-185, 205). The Xiao and Yao rhyme-sets of HWZY 194

correspond to the 'narrow sounds' ( ^ ^ xi-yin) and the

'wide sounds' ( ^ hong-yin) of the Xiao-hao rhyme of

ZYYY. The You rhyme-set corresponds to the You-hou rhyme of

ZYYY. The 'narrow sounds' and 'wide sounds' of the ZYYY

Xiao-hao rhyme are reconstructed with the final types /jow/

and /-(w)ow/f respectively; the ZYYY You-hou rhyme is

reconstructed with the final types /-(w)iw/ and /-jiw/ (see

Tables 28 and 29 in Section 5.0 and Hsueh 1986:93).

The endings of the words of the rhyme-sets Xiao, Yao, and You are transcribed as " ? " [-w] in the CP and VP notations, and as " i- " [-o] (for the Xiao and Yao rhyme-

sets) or " T " [-U] (for the You rhyme-set) in the PVP notation. The following explanation of the transcription of these endings are found in a paragraph about " ®ore endings in PNPP: ^ If ^ ir ^ f il- ? , I V â £ ; 2 , ^ ’ ? * e T J- ,

Ï X > J L , % g g & #

"In the present transcription, » » [tHjaw] becomes " ^ "

[thjao]; " ^ " [tpHiw] becomes " " [tshiu]; [tsHjaw] becomes " ^ " [t^jo] or " ^ " [t^jao]; " 7^ " [tsaw] becomes

" ^ " [tso] or " ^ " [tsao].^ Originally, " g " [w] is not the sound " T " [u] or " " [o], and " " [w] is not the 195 sound " _u " [o] or " jl, " [jo]. However, " g " [w] in the

Xiao and Yao rhyme-sets is pronounced like " p " [u]

[Judging from pronunciation notations, this is probably a misspelling of " j _ ” [o]]; " g " [w] in the You rhyme set is pronounced like " f " [u]; " ^ " [w] in the Yao rhyme is pronounced like " _L " [o] or " iL" [jo]. Therefore, words with the endings " » , % " are now transcribed with " J_ ,

iL, T " also according to the pronunciation of each rhyme, and are written as such in order to facilitate beginning learners' study of these words."

We can infer from this explanation that the Korean letters

" f f ^ f .1. , T ", which correspond to the endings of the

Xiao, Yao, You rhyme-sets, and the Yao rhyme, all transcribe the same ending in Mandarin, i.e., [-w] /-w/. 196

Now, let us examine the Korean pronunciation

notations of the words of the Xiao. Yao. and You rhyme-sets.

You rhyme-sets

jiao & biao xiao fS. Zhao T liao

CP ^ [kjew] ^ [pjew] [sjew] /y [tgjew] ^gf [liew]

VP Ÿ [kjaw] g" [pjaw] ^[sjaw] Ÿ [tejaw] ^ [liaw]

PVP 'é [kjao] Ü [pjao] jf [sjao] l^[tgjao] 5^ [liao]

d (vj gao 7] dao T zao ^ jiao ^6 qiao

CP ^ [kaw] Ÿ [taw] [tsaw] ÿ [kjaw] - ÿ [khjaw]

VP ft

PVP Zt [kao] ^ [ t a o ] ^ [tsao] Ï [kjao] "^[khjao]

Cr kou zou hou k. jiu ^ shou

CP f [kiw] ■f* [tsiw] -# [xHiw] ÿ [kiw] V [giw]

VP

PVP ^ [kiu] f [tsiu] f [Xiu] T [kiu] ^ [giu]

NOTE: The Korean diphthong " ^ " [jaj is transcribed as [je] in IPA here (see Section 5.1.2 no. 2 and Section 5.1.3). 197

1. The Xiao rhyme-set words— which come from the LMC

/-jaw/-type (e.g., qiao, ^ biao, j" shao, and

zhao) and /-ew/-type (e.g., o2j. jiao, ^ diao, ^ niao,

and 7 liao)— are transcribed with the final type " ^ "

[-jew] ([-j£w] /-jew/) in the CP notation. This final type

is changed to " " [-jaw] in the VP notation, and is

transcribed as " " [-jao] in the PVP notation. Both

pronunciation notations transcribe the same pronunciation in

Mandarin, viz., [-jaw] (see the discussion in the above).

In the Zhonazhou vinvun ( Hi ^ f^ 1351) of the

Yuan dynasty, words of this final type became homophonous with the words with guttural initials from the LMC /-aw/-

type (see no. 2 below and Li X. K. 1983:183). Thus, we

consider [-jaw] and /-jow/ to be the phonetic and phonemic

forms of this final type in Northern Mandarin of the OM period. This implies that we consider /-jow/ to be the

final type of the 'narrow sounds' of ZYYY Xiao-hao rhyme.

Therefore, we consider that the transcriptional change from

" " [-jew] ([-jaw] /-jew/) in the CP notation to " J= "

[jaw] ([-jaw] /-jow/) in the VP notation reflects a vowel- backing shift before the labial-velar off-glide, which can be formulated as follows:

e -4» o / ___ w (37)

The Xiao rhyme-set words, except those with a retroflex

initial, retain the same phonetic value [-jaw] in MP as 198 represented in the VP and PVP notations. We take /-jow/ as the phonemic representation of the final type of the Xiao rhyme-set words in OM, but /-jaw/ as that in MM and MP since phonemically only one low vowel, viz., /a/, occurs before a non-zero ending in MM (of. Table 32 in Section 5.0).

The words with a retroflex initial of the Xiao rhyme-set lost its palatal medial /-j-/ by Rule (10) j — ^ 0 / r ___ (in Section 4.3.3.5) after the PVP notation, and thus have the final type [-4tw] /-aw/ in MP, e.g.,

( [tg jaw] /crjaw/ in MM — ^ [tgaw] /craw/ in MP) and ''y

([sjfiw] /srjaw/ in MM --»► [saw] /sraw/).

2. Words from the LMC /-ow/-type (e.g., ^ gao,

77 dao, mao, cao, and lao) are transcribed with the final type " ^ " [-aw] in the CP and VP notations, and this final type is written as " jiT" [-ao] in the PVP notation. These two notations transcribe the same pronunciation in Mandarin, viz., [-a,w]. We take /-ow/ as the phonemic representation of this final type in OM, but

/-aw/ as that in MM and MP."

Words from the LMC /-aw/-type, except those with a guttural initial, (e.g., è? bao, li\ zhao, and 0 " chao) are also transcribed in the same way as those from the LMC

/-ow/-type. We can infer from this phenomenon that the LMC

/-ow/ and /-aw/ types coalesced before the Mandarin period, and thus formed the HWZY Yao rhyme-set. This change can be 199 formulated by a vowel-backing shift before a labial-velar semivowel ending /-w/ as follows (Chou 1989:184):

a — o / ___ w (38)

We have also dealt with similar rules of vowel-backing, namely. Rule (28) a — ^ o / ___^,k (in Section 5.2.2) and

Rule (37) e — » o / ___ w (37) (in this section). We now can combine these three rules into a general rule of vowel- backing as follows:

a, e — > o / ___ E (E = [+back]) (39)

The final type of the words with a guttural initial from the LMC /-aw/-type are transcribed as " [-jaw] in the CP and VP notations, and as " jf " [-jao] in the PVP notations. These two notations both transcribe the same pronunciation in Mandarin, viz., [-jAw]. We take /-jow/ as the phonemic representation of this final type in OM, but

/-jaw/ as that in MM (see no. 1 above). This phenomenon reveals that the LMC /-aw/-type words with a guttural initial acquired the palatal medial /-j-/ by Rule (18) o ► j / G ___ a (in Section 5.1.2) before the Mandarin period. Furthermore, we can infer that this final type was palatalized before the LMC /-ow/ and /-aw/-types had coalesced by Rule (38) a —> o / ___ w (in this Section), for the nuclear vowel /-a-/ is a conditioning factor in Rule

(18), which formulates the palatalization of this final 200

type. Later, the nucleus of this final type changed to the

low back vowel /-o-/ by Rule (38) a — > o / ___ w (in this

section). This final type [-j^w] /-jow/ became homophonous

with that of the Xiao rhyme-set words (which are from the

LMC /-jaw/ and /-ew/-types) in the VP and PVP notations and

also in the Zhonazhou vinvun (1351) of the Yuan dynasty. On

the basis of these facts, we consider that the words of the

HWZY Xiao rhyme-set, or the 'narrow sounds' of the ZYYY

Xiao-hao rhyme, must be reconstructed with the final type

/-jow/ (see no. 1 above).

3. The finals of the words from the LMC /-iw/-type

(e.g., Ü kou, ^ dou, ^ tou, ^ zou, and hou) are

transcribed as " ? " [-iw] in the CP and VP notations, and as " % " [-iu] in the PVP notation. These two notations both transcribe the same pronunciation in Mandarin, viz.,

[-iw] /-iw/. We consider that the nuclear vowel of this

final type is a high vowel with a phonetic quality higher than but similar to schwa, namely, [a] (see Section 3.3.3 and Section 5.1.1 no.2).

The words with a labiodental initial or a Pr[l]

initial from the LMC-type are also transcribed in the same way as those from the LMC /-iw/-type. To explain, words

such as ^ fou (LMC /f jwiw/) and fu (LMC /fHjwiw/)

first lost the feature [+palatal] by Rule (5)

j — > ^ / Bf ___ w (in Section 4.3.2), and subsequently the

labial medial /-w-/ became weak phonetically after a labial 201

initial by Rule (16) w — ^ / B ___ (in Section 5.1.1).

Thus, the final type of these words became [-iw] /-wiw/ before the Mandarin period, and it is transcribed in the

Korean notations in the same way as the final type [-iw]

/-iw/."

The words with a Pr[l] initial from the LMC /-jiw/- type, e.g., zhou (LMC /orjiw/), chou (LMC /crHjiw/), and ^ shou (LMC /srjiw/), lost the palatal medial /-j-/ by

Rule (7) j ^ / Pr[l] ___ (in Section 4.3.3.1) before the

LMC period. Therefore, these words have the same final type

([-iw] /-iw/) as those of the words from the LMC /-iw/-type in all three Korean pronunciation notations.

4. Words from the LMC /-jiw/-type, except those with a labiodental or Pr[l] initial, e.g., ^ jiu, ^ jiu, shou, and xiu, are transcribed with the final type

" è " [-iw] in the CP and VP notations, and this transcription is changed to " " [-iu] in the PVP notation.

We take [-iw] and /jiw/ as the phonetic and phonemic forms of this final type in all three pronunciation notations.

5. The words of the Xiao, Yao. and You rhyme-sets have the following final types:

(1) With the Xiao rhyme-set, [-jew] ([-j6w] /-jew/) occurs in the CP notation, but this final type is changed to

[-jaw] ([-j&w] /-jow/) in the VP notation.

(2) With the Yao rhyme-set, [-aw] ([-&w]) and [-jaw]

([-jAw]) occur in all three pronunciation notations. We 202

take /-ow/ and /-jow/ as the phonemic representations of

these final types in OM, but /-aw/ and /-jaw/ as their

phonemic representations in MM and MP.

(3) With the You rhyme-set, [-iw] (/-iw/) and [-iw]

(/-jiw/) occur in all three pronunciation notations.

In sum, the final types [-jew] ([-jew] /-jew/), [-aw]

([-ÜW] /-ow/), [-jaw] ([-jaw] /-jow/), [-iw] ([-iw]), and

[-iw] (/-jiw/) occur in the CP notation. [-jew] ([-jew]

/-jew/) is changed to [-jaw] ([-jaw] /-jow/) in the VP notation, which obtatins a different phonemic status, viz.,

/-jaw/ in MM. Therefore, the final types [-aw] ([-aw]

/-aw/), [-jaw] ([-jaw] /-jaw/), [-iw] (/-iw/), and [-iw]

(/-jiw/) occur in MM. Our reconstruction here implies that according to the VP, the words of the ZYYY Xiao-hao rhyme had the final types /-(w)ow, -jow/ in Northern Mandarin of the OM period; whereas, according to the CP notation, the words of the HWZY Xiao and Yao rhyme-sets had respectively the final types /-jew, -(j)ow/ in the sound system of

HWZY."

5.5 The rhyme-sets Ge, Ma, and Zhe

These three rhyme-sets of HWZY are developed from the

LMC Guo and Jia rhyme classes, both with the zero ending and low nuclei (/-o-/ for the Guo rhyme class and /-a-/ for the

Jia rhyme class). The Guo rhyme class (i.e., the /-(w)o/ and /-j(w)o/-types) and the Jia rhyme class (i.e., the 203 /-(w)a/ and /-j(w)a/-types) appear as a combined rhyme class in the late period rhyme table books such as SSDZ, QYZZT, and QYZN (cf. Martin 1953:3 and Hsueh 1975:60). According to this phenomenon, the Guo and Jia rhyme classes apparently coalesced by a vowel-fronting shift as follow:

o — > a / j(w) ___ # (40)

We have dealt with a similar rule of vowel-fronting in

Section 5.2.3, namely. Rule (31) o — > a / j(w) ___^,k. We now can combine these two rules into a single rule of vowel- fronting as follows:

o — > a / j(w) ___ (E) E = [-back] (41)

The Ge rhyme-set comes from the LMC /-(w)o/-type

(division 1 of the Guo rhyme class). The Ma rhyme-set comes from the LMC /-(w)a/-type (division 2 of the Jia rhyme class). The Zhe rhyme-set comes from the LMC /-j(w)o/-type

(division 3 of the Guo rhyme class) and /-j(w)a/-type

(division 3 of the Jia rhyme class) (see Tables 27, 30, and

31). 204

Let us now examine the Korean transcriptions of the words of the £e. Ma, and Zhe rhyme-sets.

Table 43 - Transcription of the words of the Ge, Ma, and Zhe rhyme-sets

CP ^ [k3] [t5] [13] 4 [tHa] [?3]

VP If o^^tHa] cf[?a]

PVP 5 . [to] & [lo] [ta] [a]

6^ ke 1Ê. zuo A ma H da /lé pa

CP [khw9 ] ^ [tsHwa] ^[mwa] ^1' [ta] [pha]

VP oj- [ma]

PVP -2. [kho] 2L [tso]

^ jia \iL hua n qie ^ Che xue

CP 7|= [kja] i|-[xwa] [kHje] ^ [tghje] fJ[xHjue]

VP

PVP [khje] ,t=j [xjue]

NOTE: The Korean vowel letters " =j " [ja] and " " [jujs] are transcribed here as [je] and [jue] in IPA, respectively (see Section 5.1.2 no. 2 and Section 5.1.3). 205

1. The Ge rhyme-set words from the LMC /-o/-type

(division 1 of the 'unrounded' type) and /-wo/-type

(division 1 of the 'rounded' type), both of the Guo rhyme

class, are respectively transcribed with the Korean vowels

" -j " [d] and " r-j " [w3] in the CP notation. This

transcription suggests that the nucleus of the HWZY Ge

rhyme-set was probably close to the mid central vowel [a] phonetically. On the other hand, the HWZY Ge rhyme-set and

its corresponding rhyme in ZYYY, the Ge-ae rhyme have been both reconstructed with the low back vowel phoneme /o/ as their nucleus (see Tables 29 and 31, Hsueh 1975:61, and Chou

1989:191). Therefore, we accept /o/ as the phonemic representation of the nucleus of the HWZY Ge rhyme-set, and postulate the lower-mid back vowel [j] as its phonetic value

(see Table 26 and Section 5.1.4).

2. The Ge rhyme-set words from the LMC /-o/-type are all transcribed with the final type " -j " [-a] ([-3] /-o/) in the CP notation. However, this final type is changed in the VP and PVP notations in three different ways as follows:

(1) Words with a guttural initial, e.g., ^ ge, ^ ke, and jtj he retain the same final type " -j " [-a] ([-a]

/-o/) in the VP and PVP notations. We take [-J] and /-o/ as the phonetic and phonemic forms of this final type in OM and

MM. This final type changed to [- 9 / G ___ # (in Section 5.1.4) during the period between MM and MP. 206

(2) Words with a dentalveolar or dentalveolar sibilant initial, e.g., ^ duo, nuo, ^ zuo, and J^luo, retain the final type " " [-3 ] ([-3 ] /-o/) in the VP notation, but it is changed to " _L " [-0] (/-o/) in the PVP notation. It is likely that these words acquired the labial medial /-w-/ by Rule (23) ^ > w / C____ o# (in Section

5.1.4) during the OM period. Thus, in the PVP notation, these words share the same final type [-0] (/-(w)o/) with the words from the LMC /-wo/-type (see no. 3 below). In addition, words like e (LMC /rjo/) and ^ wo (LMC / t jo/) are also transcribed in this way, i.e., ([-j] /-o/) in the CP and VP notations, and [-0] (/-o/) in the PVP notation.

(3) Words with a dentalveolar stop initial such as ta, 7^ da, and na are transcribed with the final type

" " [-a] ([- ] /-a/) in the VP and PVP notations.

According to this transcription, these words obtained the final type [-a] /-a/, which is retained in MP, in Northern

Mandarin no later than the time of the VP notation (or the mid-fifteenth century). In ZYYY, the word 4^ ta (LMC

/tho/) appears in the Ge-ae rhyme (/-o/), where it is expected, but the words da (LMC /tHo/) and na (LMC

/no/) appear the Jia-ma rhyme (/-a/). The word jm ta (LMC

/tho/) must have acquired the labial medial /-w-/ by Rule

(23) frf — > w / C ___ o# (C f G) (in Section 5.1.4), and thus has the final type /-wo/ in ZYYY. On the other hand, since 207

the words da, na, etc. had their nuclear vowel

changed from to / to /a/ exceptionally, they did not acquire

the labial medial ; thus, they have the final type /-a/ in

ZYYY.

3. The Qe rhyme-set words from the LMC /-wo/-type,

e.g., ^ guo, ^ duo, po, and ^ he, are transcribed

with the final type " pj " [-wa] ([-w3] /-wo/) in the CP and

VP notations, but it is changed to " ^ " [-o] (/-(w)o/) in

the PVP notation. We consider that this transcriptional

change reflects the neutralization of /-wo/ and /-o/ in

Mandarin (see Section 5.1.4 no. 4).

The final of the word ma (LMC /mwo/) is

transcribed as " 7^ " [-w3] ([-wJ] /-wo/) in the CP notation,

but as " " [-a] ([-&] /-a/) in the VP and PVP notations.

This obviously underwent an exceptional change probably

because the word is sort of an interjection.

4. The Ma rhyme-set words, which are developed from

the LMC /-(w)a/-type have the following three final types, which remain unchanged in MP:

(1) The words, except those with a guttural initial,

from the LMC /-a/-type, e.g., ^ da, pa, ^ ma, and

5^'sha, are transcribed with the final type " " [-a]

([- ] /-a/) in all three pronunciation notations.

(2) The words with a guttural initial from the LMC

/-a/-type, e.g., ^ jia, jia, ^ ya, 'p xia, are

transcribed with the final type " j: " [-ja] (/-ja/) in all 208 three pronunciation notations. These words acquired the

palatal medial /-j-/ by Rule (18) ^ — j / G ___ a before

the Mandarin period. However, the guttural initials of

these words were not palatalized yet at the time of the PVP

notation, for they are transcribed with Korean velar or

glottal consonant letters.

(3) The words from the LMC /-wa/-type, e.g., /i\ gua,

wa, 'g. shua, and hua, are transcribed with the

final type " j|>" [-wa] ([-wA] /-wa/) in all three pronunciation notations.

5. The words from the LMC /-j(w)a/-type, e.g., 'jjà qie, jie, ^ che, ^ she, and ^ ye, are transcribed with the final type " " [-je] ([-je] /-je/) in all three pronunciationn notations. This transcription reveals that

the LMC /-j(w)a/-type words underwent a vowel-fronting shift as follows before the Mandarin period.

a — > e / j(w) ___ # (42)

We have dealt with a similar vowel-fronting shift rule in

Section 5.1.2, namely. Rule (19) a e / j ___ n. We now can combine these two rules into one rule as follows:

a — > e / j(w) ___ (E) (E = [-back]) (43)

Words with this final type, [-j2] /-je/, still retain the

same final type in MP. However, between MM and MP, the words with a retroflex initial first lost the palatal medial 209

/-j-/ by Rule (10) j — ► ^ / #r ___ (in Section 4.3.3.S),

and later the nucleus of these words became [-a-] by Rule

(36) e — d / #r ___ (in Section 5.3.2), e.g., ^ che

([tghj£] /crhje/ in OM and MM -y» ([tshf] /crha/ in MP) and

$ ([gjë] /srje/ in OM and MM --7 [gar] /sra/). The final of the word xue (LMC /xjwo/) is

transcribed as " ,7^ " [-jue] ([-jw6 ] /-jue/) in all three pronunciation notations. The nucleus of this word in LMC

/-o-/ became /-e-/ in OM by a two-step change as follows before the Mandarin period: first by Rule (41)

o — > a / j(w) ___ (E) (in this section), and subsequently by Rule (43) a e / j(w) ___ (E) a (in this section). We

speculate that this final type was phonetically realized, in

all likelihood, as either [-y£] or [-jw6 ] in OM and MM (see

Section 5.1.1 no. 4). The phonetic and phonemic forms of

this final type in MP are [y£] and /-jwa/, respectively, due to the appearance of the high front rounded vowel [y] at the phonetic level and the raising of the low front vowel the mid central vowel /-3-/ by rule (36) e — > 9 / ___ # (in

Section 5.3.2) after the time of the PVP notation.

6 . To sum up, the following final types occur with

the Ge, Ma, and Zhe rhyme-sets:

(1) With the Ge rhyme-set, [-a] ([-3] /-o/) and [-wa]

([-wD] /-wo/) occur in the CP notation; [-9] ([-3] /-o/). 210

[-a] ([-&] /-a/), and [-wa] ([-wJ] /-wo/) occur in the VP

notation; [-a] ([-0 ] /-o/) and [-o] (/-(w)o/) occur in the

PVP notation.

(2) With the Ma rhvme-set. [-a] ([-a] /-a/), [-ja]

(/-ja/), and [-wa] /-wa/) occur in all three pronunciation

notations.

(3) With the Zhe rhyme-set, [-je] ([-jë] /-je/) and

[-jue] ([-jw6 ] /-jwe/) occur in all three pronunciation

notations.

5.0 The rhyme-sets Zhi, QL, Jie, and Hui

These four rhyme-sets of HWZY are developed from the

LMC Zhi and Xie rhyme classes, both with the palatal

semivowel ending /-j/. The LMC Zhi rhyme class has words

only in division 3, which has the /-j(w)ij/-type; whereas

the LMC Xie rhyme class has words in all four divisions, which respectively have the final types /-(w)oj/, /-(w)aj/,

/-j(w)aj/, and /-(w)ej/ (see Table 27 in Section 5.0).

5.6.1 The Jie rhyme-set

The HWZY Jie rhyme-set comes from division 1 of the

'unrounded'type (the /-oj/-type) and division 2 (the

/-(w)aj/-type), both of the LMC Xie rhyme class. The Jie rhyme-set in HWZY and its corresponding rhyme in ZYYY, viz., the Jie-lai rhyme, have been both reconstructed with the 211 final types /-(w)aj, -jaj/ {see Tables 28, 29, 30, and 31; and Hsueh 1975:66 and Chou 1989:194).

Let us now examine the finals of the Jie rhyme-set words.

Table 44 - Transcription of the Jie rhyme-set words

Ht. gai ^ lai # jie M- ya '1'^ kuai

CP 7f| [kaj] H/f [laj] 7/j [kjaj] a#[jaj] JJj [khwaj]

VP [kjej] dfj [aj]

PVP 2/| [lej]

NOTE: The Korean diphthongs " 4| " [3)] and " ^ " [ja] are transcribed here as [ej] and [je] in IPA, respectively (see Section 5.1.2 no. 2 and Section 5.1.3).

1. Words from the LMC /-oj/-type (e.g., gai,

'g tai, zai, ^ ai) and the words with a non-guttural from the LMC /-aj/-type (e.g., pai, ^ mai, ^ zhai) are transcribed with the final type " N " [-aj] (/-aj/) in all three pronununciation notations. This phenomenon reveals that the LMC /-oj/-type merged into the LMC /-aj/- type by Rule (21) o — > a / C ___ (E) (in Section 5.1.4) before the OM period. This final type remains unchanged in

MP.

The final of the word ^ lai (LMC /-oj/) is transcribed as " tf " [-aj] (/-aj/) in the CP and VP 212

notations, but as " " [-ej] ([-£j] /-aj/) in the PVP

notation. The nucleus of this word can be phonetically

realized as [-£-] by a rule of assimilation as follows:

V > [+front] / ___ [+front] (see Hsueh 1986:64). Since

there was not any mid or low front vowel phoneme in Middle

Korean, " j " [g] (/a/ in Korean) was the closest vowel in

Middle Korean to transcribe [£] or [e] in OM and MM (see

Table 12 in Section 4.1 and Section 5.1.2 no. 2). Whereas

the other words with the same final type /-aj/ could also be pronounced with the same final type [-£j] in casual speech;

the final of this particular word was perceived differently probably because this word has an approximant initial, namely, the dentalveolar liquid /I/, and in addition

frequently appears as a directional verb in resultative verb compounds.

2. Words such as ^ jie, kai, ^ xie, and ^ ya are transcribed with the final type " H " [-jaj] (/-jaj/) in the CP notation. These words have a guttural initial, and are from the LMC /-aj/-type. The final type /-jaj/ in the CP notation shows that these words acquired the palatal medial /-j-/ by Rule (18) ^ — j / G ___ a (in Section

5.1.2) before the OM period. The final type /-jaj/ of these words bifurcates in the VP and PVP notations as follows:

(1) Words such as jie (LMC /kaj/), kai (LMC

/khaj/), ^ xie (LMC /xHaj/) appear with the final type 213

" =|j " [-j®j] ([-j j] /-jaj/) in the VP and PVP notations. These word must have underwent a vowel-shift by Rule (43) a — e / j(w) ___ (E) (in Section 5.5). This reveals that the nuclear vowel of these words moved further front and up, but it did not change to a different phoneme; these words still rhymed with words with the final type /-aj/ and

/-way/. After the time of the PVP notation, the final type

[-j6 j] /-jaj/ lost the palatal ending /-j/ due to palatal dissimilation between the palatal medial and ending, and thus it became [-j€] /-j)/ in MP.

(2) The words ^ ja (LMC //jaj/) and -j^ (LMC /^aj/) are transcribed with the final type " ^ " [-aj] (/-aj/) in the VP and PVP notations. These words had the zero initial at the time of the VP notation (see Section 4.3.5). These words must have lost the palatal medial /-j-/ due to labial dissimilation between the palatal medial and ending before the time of the VP notation/"" Based on these cases of palatal dissimilation, we now can write the following rule of palatal dissimilation:

j -tf / / j V ___ ( / (w) V j (44)

3. Words such as 3^ guai, kuai, ^ wai, and

huai are transcribed with the final type " dfj " [-waj]

(/waj/) in all three pronunciations. These are the words with a guttural initial from the LMC /-waj/-type. 214 4. To sum up, the Jie rhyme-set words have the final types as follows:

(1) [-aj] (/-aj/), [-waj] (/-waj/), and [-jaj] (/-jaj/) occur in the CP notation.

(2) [-aj] (/-aj/), [-waj] (/-waj/), and [-jej] ([-j£j]

/-jaj/) occur in the VP notation.

(3) [-aj] (/-aj/), [-ej] ([- j] /-aj/), [-waj]

(/-waj/), and [-jej] ([-j£j] /-jaj/) occur in the PVP notation.

5.6.2 The Hui rhyme-set

The HWZY Hui rhyme-set is developed from division 1, division 3, and division 4 of the 'rounded’ type of the LMC

Xie rhyme class (i.e., the /-woj/, /-jwaj/, and /-wej/- types, respectively) and division 3 of the 'rounded' type of the LMC Zhi rhyme class (i.e., the /-jwij/-type). The HWZY

Hui rhyme-set and the 'rounded' type of the ZYYY Oi-wei rhyme, which corresponds to the HWZY Hui rhyme-set, have been both reconstructed with the final type /-wij/ (see

Tables 28, 29, 30, and 31 in Section 5.0; Hsueh 1975:70-71 and Chou 1989:196). 215

Table 45 - Transcription of the Hui rhyme-set words

gui dui ^ zui sui K bei

CP [kuj] rl [tuj] 4) [tsuj] [sHuj] -tI [puj]

VP ti II II II [Pij]

PVP [suj]

1. Words from the LMC /-woj/-type (e.g., dui,

zui, and ^ hui), /-jwa j/-type (e.g., zhui, shui, and ^ lei), the LMC /-we j/-type (e.g., sui), and

/-jwij/-type (e.g., ^ gui and gui) are transcribed with the final type " t | " [-uj] (/-wij/) in all three transcriptions. From these transcriptional data we can infer that the four LMC final types /-woj, -jwaj, -wej,

-jwij/) formed the OM final type (/-wij/) through the following changes:

First, the nucleus of the LMC /-jwaj/-type underwent a vowel-fronting shift by Rule (43) a — » e / j(w) ___ (E)

(E = [-back]) (in Section 5.5), and thus the LMC /-jwaj/,

/-wej/-types merged into the LMC /-jwej/.

Second, the nuclei /-o-/ and l-e-l in syllables with the palatal ending /-j/ underwent a vowel-raising shift, which can be formulated as follows:

e, o -•> i / ___ j (45) 216

Third, the /-jwij/-type lost the palatal medial /-j-/ by

Rule (44) j ^ / ___ (w) V j (in Section 5.6.1).

Therefore, the HWZY Hui rhyme-set has only one final type, viz., /-wij/, and this final type is retained unchanged in

MP as well as in the VP and PVPnotations.

2. The final of the word ^ bei (LMC /pwoj/) and

mei (LMC /mjij/) are transcribed as " t| " [-uj]

/-wij/) in the CP notation, but as " -|" [-ij]

(/-wij/) are transcribed in the VP and PVP notations. This transcriptional change reveals that the labial medial /-w-/ in these words was weakened phonetically by Rule (16) w -4»" ^ / B ___ (in Section 5.1.1) before the time of the VP notation.

3. In sum, the words of the Hui rhyme-set have the final type [-uj] (/wij/) in all three pronunciation notations. This final type is transcribed as [-ij] (/-wij/) in words with a labial initial in the VP and PVP notations.

5.6.3 The Zhi and fli rhyme-sets

The Zhi and gi rhyme-sets in HWZY come from the LMC

/-jij/-type (division 3 of the 'unrounded' type of the Zhi rhyme class) and /-jaj, -ej/-types (divisions 3 and 4, respectively, of the 'unrounded' type of the Xie rhyme class. The HWZY Zhi rhyme-set is reconstructed with the final types /-i/ and /-ji/; the HWZY fii rhyme-set is reconstructed with the final type /-jij/ (see Tables 30 and 217

31 in Section 5.0. and Chou 1989:193). The Zhi rhyme-set in

HWZY corresponds to the Zhi-si rhyme and part of the

'unrounded' type of the Oi-wei rhyme in ZYYY. The ûi rhyme-

set in HWZY corresponds to part of the 'unrounded' type of

the Oi-wei rhyme in ZYYY. The ZYYY Zhi-si rhyme is

reconstructed with the final type /-&/; the ZYYY Oi-wei

rhyme is reconstructed with the final types /-(w)ij, -jij/

(see Tables 28 and 29 in Section 5.0 and Hsueh 1975:67-72).

In the CP notation, the Zhi rhyme-set words with

sibilant initials are transcribed with the Korean consonant

letter " A " in the ending position. In such a case, the

letter " ,1^ " is equivalent to the zero ending in Mandarin,

but shows that the preceding nuclear vowel is an apical

vowel, namely [%] between a dentalveolar initial and the

nucleus /&/ or [%] between a [tg-]-type retroflex initial

and the nucleus /i/.

In the PVP notation, when the Mandarin high central

nuclear vowel /-i-/ immediately follows a dentalveolar or

retroflex initial, the nuclear vowel is represented by the

Korean vowel " % " [/^]. In this case, the phonetic value in

Mandarin which the Korean vowel denotes is dependent on the

structure of the syllable in which it occurs; i.e., [%]

after a dentalveolar initial in a syllable with the zero

ending, [j] after a [tg-]-type retroflex initial in a

syllable with the zero initial, and [9] before a non-zero

ending. 218 Now, let us examine the Korean pronunciation notations of the words of the Zhi and gi rhyme-sets.

Table 46 - Transcription of the words of the Zhi and Oi rhyme-sets

^ zi È zhi zhi mi M - bei

CP L [tsi] \\ [tgi] \| [t/i] 4 [mi] f/i(/[pHi]

VP 5* [tsj] ^[t§l] A* [t/i] ■2| [mij]

PVP ^ [tsA] [tfA] M [t/i] -1 [pij]

/+N ''Si $ qi a ji ni f. li

CP *1 [vi] [kHi] H [kjej] L^j [njej] [Ijej]

VP rj [ki] [ni] 2/ [li]

PVP 4 [wi] 7/ [khi]

NOTE: The Korean letters " 2T " [iz], " a' '' [iz], " » " [/\], and " " [j9j] are respectively transcribed here as [1,2], [i], [A], and [jej] in IPA. (see the above discussion. Table 15 in Section 4.2, Section 5.1.2 no. 2, Section 5.1.3, and Section 5.2.3 no. 2).

1. Words with a dentalveolar initial from the LMC

/-jij/-type (e.g., ^ zi, ci, j| zi, and 0 si) are transcribed with the final type " — " [-i] (/-&/) in the CP notation. This final type is written as " %"" [-%] (/-&/) in the VP notation, and as " % " [ a] ([-1] /-±/) in the PVP 219

notation. From this transcription we can infer that the LMC

final type /-jij/ after a dentalveolar initial became the

/-i/-type before the time of HWZY. To explain, the LMC

final type /-jij/ first lost the palatal ending /-j/ by Rule

(44) j » ÇJ / j V ___ (in Section 5.6.1), and subsequently

lost the palatal medial /-j-/ as formulated by the following rule:

j / Ds ___ (46)

2. Words with a Pr[2] initial from the LMC /-jij/- type (e.g., ÿ zhi, ^ chi, shi, and shi ) are transcribed with the final type " ( " [-i] (/-ji/) in the CP notation. This final type is changed to " [-1] (/-&/) in the VP notation, and is transcribed as " * " [-/\] ([-i]

/-il) in the PVP notation. These words lost the palatal ending /-j/ by palatal dissimilation as represented by Rule

(44) j — > o / j V ___ (in Section 5.6.1) before the time of

HWZY, and thus has the final type /-ji/ in the CP notation.

Later, these words also lost the palatal medial /-j-/ by

Rule (10) j — » j/ / # r (in Section 4.3.3.5) and thus have the final type /-i/ (which is identical with the rhyme- base of the ZYYY Zhi-si rhyme) in the VP notation.

3. Words with a Dy initial from the LMC /-jij/-type

(e.g., ^ zhi, chi, and ^ zhi) and words with a Py initial from the LMC /-jaj/-type (e.g., iât shi and ^ shi) are transcribed with the final type " | " [-i] (/-ji/) in 220

the CP and PVP notations, and this final type is written as " J " [-i] (/-ji/) in the VP notation. We can explain this phenomenon by the following sound changes:

First, the LMC final type /-jaj/ merged into the LMC

/-jij/-type by Rule (43) a - » e / j(w) ___ (E) (E = [-back]

(in Section 5.5) first and subsequently by Rule (45)

e — ^ i / ___ j ( in Section 5.6.2).

Second, the LMC final type /-jij/ lost the palatal ending /-j/ by Rule (44) j ^ / j V ___ (in Section

5.6.1). The palatal medial /-j-/ is not deleted in these words in the PVP notation, but it is lost between the period of the PVP notation and MP by Rule (10) j — > / # r ___ (in

Section 4.3.3.5). This phenomenon supports our hypthesis that the Dy initials became the [tg-]-type later than the

Pr[2] initials (see Section 4.3.3.5).

4. Words with a non-sibilant initial from the LMC

/-jij/-type (e.g., ji, ^ qi, ci, ^ xi) are transcribed with the final type " j " [-i] (/-ji/) in all three pronunciation notations. The words lost the palatal ending /-j/ by palatal dissimilation as represented by Rule

(44) j — ^ ^ / j V ___ (in Section 5.6.1).

Words with a labial initial from the LMC /-j(w)ij/- type, e.g., ttl bi, bei, ^ pi, ^ fei are also transcribed with the final type " j " [-i] (/-ji/) in all three pronunciation notations. We consider that words with a labial initial initial have the labial medial /-w-/ at the 221 phonemic level. However the labial medial /-w-/ after a labial initial is phonetically realized very weakly or not perceptible. Therfore, some words with a labial initial from the LMC /-jwij/-type are transcribed differently in the

VP and PVP notations as follows:

(1) The final of the word mi is transcribed as

" -I" [-ij] (/-ij/) in the VP and PVP notations. According to this transcription, this word lost the palatal medial

/-j-/ instead of the palatal ending /-j/, whereas all the other words in this rhyme-set lost the palatal ending /-j/.

(2) Words like ^ bei, bei, and fei are transcribed with the final type " -I " [-ij] in the PVP notation.

(3) The word wei is transcribed as " ^ " [uj]

(/wij/) in the PVP notation.

5. Words from the LMC /-ej/-type (e.g., di, ^ mi, ^ qi, and ^ xi) and some words from the LMC /-jaj/- type (e.g., 2 ji, 'j/JZ ni, J L li) are transcribed with the final type " =<( " [-jej] (/jij/) in the CP notation. This final type is changed to " / " [-i] (/-ji/) in the VP and

PVP notations. This phenomenon can be explained as follows:

First, the LMC /-jaj/ and /-ej/-types merged together, and formed the final type /-jej/ by a vowel-fronting shift as represented by Rule (43) a — > e / j(w) (E)

(E = [-back]) (in Section 5.5). 222

Second, the low front nuclear vowel /-e-/ was raised to

the high nuclear vowel /-&-/ by Rule (45) e — i / ___ j

(in Section 5.6.2), and thus the /-jej/-type merged into the

/-jij/-type as shown in the CP notation.

Third, the final type /-jij/ lost the palatal ending

/-j/ by palatal dissimilation as represented by Rule (44)

j — y gf / j V ___ (in Section 5.6.1). Thus, the final type

/-ji/ was formed as shown in the VP and PVP notations, and

it is retained unchanged in MP.

6 . In sum, the words of the Zhi and j^i rhyme-sets

have the following final types;

(1) With the Zhi rhyme-set, [-i] (/-i/) and [-i]

(/-ji/) occur in the CP and VP notations. [i,î] ([-i]

/-i/) and [-i] (/-ji/) occur in the PVP notations. Besides,

some words with a labial initial have the final type [-ij]

(/-ij/) or [-uj] (/-wi-j/) in the VP and/or PVP notations.

(2) With the ÛÎ rhyme-set, [-jej] (/-jij/) occurs in

the CP notation. This final type is changed to

[-i] (/-ji/) in the VP and PVP notations.

5.7 The rhyme-sets Yu and Mo

The Yu and Mo rhyme-sets in HWZY both come from the

LMC Yu rhyme class, which has only two final types, both with the labial medial /-(j)w-/ and the high central nuclear vowel /-i-/, i.e., the /-wi/-type for division 1 and the 223

/-jwi/-type for division 3 (see Table 27 in Section 5.0).

Words from the LMC /-jwi/-type formed the Yu rhyme-set.

Words from the LMC /-wi/-type and the words with a Pr[l] or

labiodental initial from the LMC /-jwi/-type formed the Mo

rhyme-set.

The HWZY Yu rhyme-set and the 'unrounded' type of the

ZYYY Yu-mo rhyme, which corresponds to the HWZY Yu rhyme-

set, have been both reconstructed with the final type

/-jwi/. The HWZY Mo rhyme-set and the 'rounded' type of the

ZYYY Yu-mo rhyme, which corresponds to the HWZY ^ rhyme- set, have been both reconstructed with the final type /-wi/

(see Tables 28, 29, 30, 31 in Section 5.0, Hsueh 1975:67, and Chou 1989:201). 224

Now, let us examine the Korean pronunciation

notations of the words of the Yu and Mo rhyme-sets.

Table 47 - Transcription of the words of the Yu and Mo rhyme-sets

‘Sj ju ^ nu qu È. zhu ru

CP [kju] T [nju] 4 [tshju] [tsju] 4 [%ju]

VP

PVP T

gu ^ fu %. ®u chu

CP -f [ku] f [nu] 4- [fu] -4- [su] 4 [tghu]

VP II II II It

PVP If II It It

1. The words of the Yu rhyme-set (e.g., ^ ju, nu, qu, zhu, and -ip ru) are transcribed uniformly with the final type " 7T " [-ju] in all three pronunciation notations. We have already determined that the phonetic value of this final type in MM is either [-j*] or [-y] and that its phonemic representation is [-jwi] (see Section

5.1.1 no. 4). (However, we transcribe the phonetic form of

this final type simply as [ju] due to a lack of clear 225 evidence to choose [-ju] or [-y], and also for typographical

convenience.) The phonemic status of this final type has

been /-jwi/ throughout the Mandarin period. The phonetic

value of this final type is [-y] in MP. Among the words

with this final type, those with a retroflex initial lost

the palatal medial /-j-/ by Rule (10) j --^ ^ / # r ___ (in

Section 4.3.3.5) during the period between the PVP notation

and MP, e.g., ^ ([tpju] /crjwi/ in OM and MM ([tgu]

/crwi/ in MP) and ([gju] /srjwi/) in OM and MM — ([gu]

/srwi/ in MP).

2. Words of the Mo rhyme-set (e.g., ^ gu, ^ nu,

^ fu, ^ su, and ^ chu) are transcribed uniformly with the final type " T " [-u] (/-wi/) in all three pronunciation notations. In addition to the words from the LMC /-wi/- type, the words with a Pr[l] initial or a labiodental

initial from the LMC /-jwi/-type are also transcribed with this final type. This phenomenon reveals that the LMC

/-jwi/-type words with a Pr[l] or labiodental initial lost the palatal medial /-j-/ before the Mandarin period. For example, the words Sit) zhu (LMC /crHjwi/) and shu (LMC

/srjwi/) lost the palstal medial /-j-/ by Rule (10)

j / # r ___ (in Section 4.3.3.5), and have the final type /-wi/ in Mandarin. The words '5C fu (LMC /fHjwi/ and

^ wu (LMC /vjwi/) lost the palatal medial /-j-/ by Rule

(5) j — 5> p / Bf ___ w (in Section 4.3.2), and have the final type /-wi/ in Mandarin. 226 3. In Slim, the words of the Yu and Mo rhyme-sets have the following final types:

(1) With the Yu rhyme-set, [-ju] (/-jwi/) occurs in all three pronunciation notations.

(2) With the Mo rhyme-set, [-u] (/-wi/) occurs in all three pronunciation notations.

5 .8 Stimmary

We have now finished a long discussion about the syllable finals of OM and MM as represented by the CP, VP, and PVP notations. We have analyzed in detail the finals of

OM and MM as transcribed in Hunmin chona'um in accordance with a strictly phonemic analysis along the line of the traditional, hierarchical segmentation of the Chinese syllable. We have also compared the three types of Korean pronunciation notations of the pre-modern Mandarin pronunciations of Chinese characters in order to investigate the diachronic phonological development of the Mandarin syllable finals.

Thus, we have arrived at a complete reconstruction of the MM finals on the basis of the PVP notation of PN and PP.

In the phonemic system of the MM finals, there are four medials /-^-, -j-, -w-, -jw-/, four nuclei /-i-, -e-, -a-,

-0-1, and five endings /-j^, -j, -w, -n, -%/. The MM vowel system is in the process of restructuring from the four- vowel system of OM to the three-vowel system of MP. 227

Specifically, the three-way contrast in depth of the OM low vowel phonemes was lost in front of a non-zero ending during the MM period (see Sections 5.1.2, 5.1.3, and 5.1.4), whereas the three-way contrast was maintained in front of the zero ending during the same period (see Section 5.5). There are thirty-six final types in MM at the phonemic level. The remarkable reduction of the forty-seven final types in the sound system of ZYYY to the thirty-six final types of MM was caused by the merger of the bilabial nasal stop ending /-m/ to the alveolar nasal stop ending /-n/ and the loss of the three-way contrast in depth of the OM low vowel phonemes before a non-zero ending (cf. Table 29 in

Section 5.0 and Table 49 in this Section). The thirty-six final types of MM was further reduced to the thirty-four final types due to the coalescence of the low front vowel

/e/ and the low back vowel /o/ of MM into the mid vowel Iq I of MP during the period between MM and MP.

Our reconstruction of the MM finals are presented in

Tables 48 and 49. In Table 48, the phonetic and phonemic forms of the MM finals and their PVP notations are arranged according to the rhyme-sets of HWZY. In Table 49, the phonemic representations of the MM finals are arranged according to the phonemic structure of the final, i.e., the types of the components of the final. 228 Table 48 - The finals of MM

(1) Dono " 7 " [-ug] /-wig/. " % " [-]ug] /-jwip/ Wu* u » [-U] /-wi/. ir " [-ju] /-jwi/

(2) Zhi II ^ II [-If -Î] /-i/ " / " [-1 ] /-ji/

(3) QL " 1 " [-i] /-ji/

(4) IR " 7T " [-ju] /-jwi/

(5) Mo " T " [-U] /-wi/

(6) Jie " H " [-aj] /-aj/, " il " [-Zj] /-aj/,

" ^ " [-waj] /-waj/ "4/ " [-jaj] /-jaj/

(7) Hui " '1" [-uj] /-wij/. " -I " [-ij] /-wij/ (8 ) Zhen [-in] /-in]. " il " [-in] /-jin/,

" r " [-un] /-win/. " ^ " [-jun] /-jwin/

Zhi* [-i] /-i/, I " [-i] /-ji/, " r " [-U] /-wi/. " TT [-ju] /-jwi/

(9) Han " J- " [-an] /-an/. " " [ -wan ] /-wan/ He* [-3] /-o/, -L " [-0 ] /-(w)o/ " 4 " (10) Shan " it " [-an] /-an/. " ib " [-wan] /-wan/,

" %[" [-j£n] /-jan/

Xia* " [-a] /-a/, o|-" [-wa] /-wa/,

[-ja] /-ja/

(11) Xian " i " [-j£n] /-jan/ " " [-jw£n] /-jwan/

Xie* " 4 " [-j£] /-je/, " 7p| " [-jwô] /-jwe/ (12) Xiao " [-jao] /-jaw/

(13) Yao " J:" [—4o] /—aw/, " " [-jao] /-jaw/

(14) Ge " -i " [-3] /-o/. J_" [-0 ] /-(w)o/ 229

(15) m t [-A] /-a/. " F " [-ja] /-ja/.

4 [-wa] /-wa/

(16) Zhe 4 [-j£] /-je/, " 7T=f " [-jwg] /-jwe/ (17) Yang J- [-&#] /-a%)/, " -jl'" [-w&g] /-wa^/.

/ [-jaij] /-ja/;/ Yao* ■i [-J] /-o/. " Æ " [-jo] /-jo/. J- [-0 ] /-(w)o/

fl8 ) Gena ? [-ig] /-i#/, " J " [-i/j] /-ji%/, T 6 [-ug] /-wi/J/, " J " i-jug] /-jwir)/ Mo* 1 [-i] /-ji/, " -/" [-7] /-o/.

[-jf] /-je/. " -/ " [-ij] /-ij/ H [-aj] /-aj/. " T " [-U] /-wi/

[-W3] /-wo/

(19) YOU f [-iw] /-iw/. " T " [-iw] /-jiw/ (20) gin r [—in] /—in/, " J. " [-in] /-jin/

gi* 1 [-i] /-ji/ (21) Tan [-an] /-an/

He* [-a] /—a/, ' -j " [-0] /-o/

(22) Yan %! [-j£n] /-jen/

Ye* F [-ja] /-ja/. " ^ " [-jS] /-je/ NOTE: (a) The numbers of the rhyme-sets are the same their original numbers in HWZY (see Table 30 in Section 5.0). (b) The entering tone rhymes of HWZY are marked with an asterisk, (c) The phonetic value of the final type /-jwi/ in MM is either [-j#] or [-y], although there is no clear evidence to choose one over the other. However, we transcribe the phonetic form of this final type as [-ju] for typographical convenience (see Section 5.1.1 no. 4). 230

Table 49 - The phonemic representations of the MM finals

Endina i w n V Vowel i e a o i a ± a i a t a Medial

-flj— i a o ii ai iw aw in an in an

-i- ii 1® ia 1o iai iiw iaw 1in ian iin ian

-W- wi- wa wo wii wai win wan win wan

-iw- iwi iwe iwin iwan iwin 231 CHAPTER V

NOTES

1. Weinreich (1953) termed both phonetic and phonological interference "phonic interference".

2. Middle Korean has more high vowel phonxmes but fewer low vowel phonemes than OM and MM. Thus, i ineral Mandarin high vowels are sufficiently different: , i, but Mandarin low vowels are frequently underdifferentiated in the pronunciation notations. Cf. Tables 12 and 25.

3. For a phonological explanation of the "colloquial pronunciation" and the "literary pronunciation", see Hsueh 1978 and Chou 1989.

4. The term "rhyme-set" is borrowed from Chou 1989:148.

5. As for the corresponcence between the twenty-two rhyme- sets of HWZY and the nineteen rhymes of ZYYY, see Dong T. H. 1968:72.

6 . The phonemic compound /-jwi-/ in MP is mostly realized as a single sound phonetically, viz., [y]. However, /-jwi-/ in some words are still pronounced as [j*], e.g., xiong [fjeg] /hjwing/ and ^ yong [ j*g] /jwi^/.

7. In the Xiru ermu zi. three types of "u" are recorded, namely, the "deep u" [u], the "central u" [*], and the "secondary u" [)*]• See Lu Z. W. 1947:123-124. Cf. Hu M. Y. 1963:188.

8. The word y'k, mei is transcribed with the final [ut] /-wit/ in the CP notation, but it joined Hui-tui rhyme /-ij/ in the "Thirteen Track" since it changed following the "colloquial pronunciation".

9. Following the Guanovun practice, the Sasona t'onahae has the Zhen rhyme-set divided into two matching rhyme-sets; words with the high front vowel [i] or the palatal medial [j] are classified as the Zhen rhyme-set ( Zhen. ^ Zhen, ^ Zhen. ^ Zhi), while words with the high non­ front vowels [i] and [u] are classified as the Wen rhyme-set ( X. Wen. Wen. ^ Wen, ^ Wu). Thus, there are four more rhymes in SSTH than in HWZY, i.e., there are eighty rhymes in SSTH.

10. /-an/ and /-en/ are respectively the rhyme bases of the Han-shan and Xian-tian rhymes of ZYYY, and of Han and Xian rhyme-sets of HWZY. 232

11. The word ^ jian belongs to the Shan rhyme in division 2 of the unrounded type of the Shan (L rhyme class of LMC, and the Shan rhyme had the final type /-an/ in rhyme tables.

12. For example, Dong (1968:66) reconstructs the nucleus of the Xian rhyme-set of HWZY as [e], and Chou (1989:155) reconstructs it as /e/. Lu Z-W. (1946:50) reconstructs the nucleus of the Xian-tian rhyme of ZYYY as [g], and Hsueh (1975:53) reconstructs it as /e/.

13. E.g., Hu m . Y. (1963), [f] : Kang S. C. (1985), [a].

14. In MP, the contrast between /-wo/ and /-o#/ still remain partially. The words with labial initials lost the contrast (cf. Rule (15) in Section 5.1.1). The words with dentalveolar, dentalveolar sibilants, and retroflex sibilant initials of both types became the rounded type "Cwo#" in MP. The words with guttural initials retain the contrast, but have quite a few exceptions. For example, the words such as ^ ge (LMC /kwo#/) and jfnf he (LMC /xHwo*) respectively^ became homophones with the words ^ ge (LMC /ko#/) and (LMC /xHo#/), and are pronounced as [gy] /ga/ and [h?T] /h@/ in MP. For further discussion on the loss of the contrast between /-wo#/ and /-o#/, see Chou 1989:258-262.

15. On the other hand, the low front vowel phoneme /e/ of OM and MM is transcribed with the Korean vowel " -j " [a]. See Section 5.1.2 no. 2. and Section 5.1.3.

16. However, we consider that the labial medial /-w-/ remained in these words at the phonemic level. See Section 5.1.1 no . 2.

17. Phonemically, only one low vowel, viz., /-a-/, occurs before a non-zero ending in MM. Cf. Table 32.

18. However, Chou (1989:168) the nucleus of the HWZY Gena rhyme-set as the front low vowel, and she formulates the formation of the HWZY Gena rhwie-set as follows : a,o -» e / JI ,R

19. The semivowel [j] is used to denote additional phonetic features such as [+front], [thigh] and [tpalatal] of other segments in Korean transcription of Chinese. Another example of using the semivowel " I " [j] to denote a phonetic quality is the transcription of the Mandarin low front vowel /e/ with the Korean vowel letter " 4 " [js]. In this case, " I " [j] adds the feature [tfrcnt] to " -j " [ a ] . See Section 5.1.2 no. 2 and Section 5.1.3. 233

20. In the VP notation, the semivowel [j] appear immediately before the glottal stop ending [--?] only in the transcriptions of the entering tone words. In the PVP notation, the semivowel [j] appear alone in the ending position of the entering tone words which changed in accordance with the "colloquial" pronunciation. Coincidentally, the MP reflexes which changed in accordance with the "colloquial" pronunciation from the entering tone words of the Gena rhyme-set retain the palatal semivowel ending /-j/ after the complete loss of the stop ending. Thus, it is probable that the glottal stop ending /-j?/ in the VP notation is palatalized.

21. The "literary" reading and the "colloquial" reading were two competing pronunciations for the former entering tone words, and the "literary" reading won in most cases eventually. Therefore, a shift from the "colloquial pronunciation" to the "literary pronunciation" during the time between the CP notation and the VP notation, (cf. 5.1.4 no. 4).

22. This final type occurs in the 'wide sounds' of the ZYYY Xiao-hao rhyme. Since we reconstruct the 'narrow sounds' of the same rhyme with the final type /-jow/ (see no. 1 and the following discussion), and in addition only one nuclear vowel phoneme is allowed for one rhyme by Working Rule (1) (in Section 5.0), we consider that the nucleus of the 'wide sounds' of the ZYYY must be the low back vowel phoneme /o/. Hsueh (1975:62-66) postulate three low nuclear vowels (/e, a, o/) for the ZYYY Xiao-hao rhyme, and reconstructs the Xiao-hao rhyme with the final types /-(w)ow, -(w)aw, -jaw, -jew/. Hsueh (1986:93 and 105 Pn. 2) revises his former reconstruction of the ZYYY Xiao-hao rhyme; he posits only one low nuclear vowel /-o-/ for the Xiao-hao rhyme, and rewrites the final types of the Xiao-hao rhyme as /-(w)ow, jow/. Stimpson (1977, 1978) posits the low central vowel /a/ as the nucleus of the ZYYY Xiao-hao rhyme, writes the final types of the Xiao-hao rhyme as /-(w)aw, -jaw/, and accounts for the three-way contrast in the Xiao-hao rhyme by differences in the medials. See Hsueh 1975:62-66, 1986:93. Also, cf. Stimpson 1977, 1978; and Ting P. H. 1980.

23. After the murmured initial /fH-/ of the word 'J%- fu became /f-/ through devoicing, the two words ^ fou and ^ fu became homonyms with the same pronunciation [fiw] /fwiw/. Later, labial dissimilation between the medial /-w-/ and the ending /-w/ caused the deletion of either the medial or the ending in these words. Thus, the word ^ fou lost the medial /-w-/, and became [faw] /fiw/ in MP; while the word fou lost the ending /-w/, and became [fu] /fwi/ in MP. 234

24. The CP notation is based on the sound system of HWZY, which is based to a great extent on Lower Yangzi Mandarin in the beginning of the Ming dynasty. Since the VP and PVP notations represent Northern Mandarin, particularly, Pekingese, of the mid-fifteenth and early sixteenth centuries, respectively; these two pronunciation notations are in line with the diachronic development of the sound system of ZYYY after the time of ZYYY. Although the sound systems of HWZY and ZYYY by and large correspond to each other; the former is somewhat conservative in comparison with the latter. Some of the discrepancies between HWZY and ZYYY are attributed to dialectal differences, although most of them are caused by diachronic sound changes. For further discussion on the relation between HWZY and ZYYY, see Sections 2.2.2 and 5.0.

25. This phenomenon is, apparently, an exception to Rule (23) 0 — w / C ___ o# (C / G). The initial /#-/ of these words changed to the zero initial before the time of the PVP notation (see Section 4.3.5). This change seems to be a borrowing from another related dialect, in which the a different version of Rule (23), viz., f ^ w / (C) ___ o# has occurred.

26. In MP, these words have the final type [-ja] /-ja/. This MP pronunciation also shows that these words did not undergo a vowel-fronting shift, but only underwent a palatal dissimilation. However, the final type /-ja/ in MP shows the deletion of the palatal ending /-j/, whereas the final type /-aj/ in the VP and PVP notations shows the deletion of the palatal medial /-j-/. CHAPTER VI

CONCLUSION

In this study, we have reconstructed the phonology of

Middle Mandarin (ca. 16th and 17th centuries, MM) as represented by the "present vulgar" pronunciation (PVP) notation of the Ponvok Noaoltae (PN) and the Ponvok

Pakt'onasa (PP). Now, we shall summarize what we have discussed in the previous chapters, and recapitulate some of the findings we have made in the course of the discussion.

On the basis of a careful review of source materials, we started with the assumption that the PVP notations in PN and PP are accurate phonetic data for the sound system of

Middle Mandarin. In examining the pronunciation notations, we carefully compared the phonemic system of Middle Korean with that of Middle Mandarin in order to determine accurately the sound values of Middle Mandarin as transcribed in the Korean alphabet. In analyzing the sound system of Middle Mandarin, we kept the syllable as the largest unit for phonological analysis, and followed the traditional, hierarchical analysis of the Chinese syllable.

235 236

In order to investigate the diachronic phonological development of Mandarin, we made a comparison between the

"correct" pronunciation (CP) notations, the "vulgar" pronunciation (VP) notations and the "present vulgar" (PVP) notations, and also made reference to other contemporary sources as well as our knowledge of Late Middle Chinese

(LMC), Old Mandarin (OM), and Modern Pekingese (MP).

However, this statement should not imply in any sense that our reconstruction of the sound system of MM is arrived at by manipulating our knowledge of OM and MM. We have indeed made every effort to reconstruct MM phonology as an independent sound system solely on the basis of the PVP notation of PN and PP.

In Chapter I, we presented the aim and scope of this study in the light of the phonological history of Chinese.

While there has been a considerable amount of research on the sound systems of OM and MP in modern Western linguistic terms, there exists a need for thorough research on the sound system of MM. Since Chinese source materials give phonological information in classificatory terms, they can only be interpreted phonemically. In contrast, Korean transcriptional sources offer phonetic data for the sound system of pre-modern Mandarin. Therefore, we made it our goal to reconstruct MM phonology on the basis of the Korean transcription of MM pronunciation, and thereby trace the diachronic phonological development of Mandarin. 237

In Chapter II, we surveyed the source materials which include the transcriptions in Hunmin chona'um (or the Korean alphabet) of the pre-modern Mandarin pronunciations of

Chinese characters. There are three types of pronunciations :

1) the "correct" pronunciation (CP), which represents the sound system of the Honowu zhenavun (HWZY);

2) the "vulgar " pronunciation (VP), which represents mid­ fifteenth century Northern Mandarin;

3) the "present vulgar" pronunciation (PVP), which represents the early sixteenth century Northern Mandarin.

We found out that the PVP notations in the PN and PP are the most extensive and accurate phonetic data for the sound system of Middle Mandarin. Thereafter, we introduced our source books, viz., PN and PP, and reviewed previous studies which dealt with the PVP notation in relation to the history of Chinese phonology. We proposed that a strictly phonemic analysis along the line of the Chinese traditional analysis of the syllable be adopted for both reconstructing MM phonology and explaining the diachronic phonological changes of Mandarin.

In Chapter III, we explained our theoretical stance on the important aspects of MP phonology. We take the syllable as the largest unit for phonological analysis of

Chinese, and follow the traditional, hierarchical segmentation of the Chinese syllable. In particular, we 238 adopted the phonemic analysis of MP proposed by Hsueh (1986) as the firm foundation of phonological analysis of Mandarin in the ensuing chapters, for it is solid in both traditional

Chinese and modern Western phonological frameworks, and also is viable for both synchronic and diachronic descriptions.

We treat the palatal initial consonants as allophones of the velar initial consonants. There are four medials, namely,

-W-, -jw-/. There are five endings, namely,

/-frf/ -jf -w, -n, -#/. Only the main vowel is [+syllabic] at the phonemic level. We posit three vowel phonemes, one for each of the three levels of vowel height; namely, the high vowel phoneme /&/, the mid vowel phoneme /a/, and the low vowel phoneme /a/. The tone is considered to belong to the vowel phonemically, although it spreads over the voiced part of the syllable phonetically.

Chapter IV consisted of two parts. In the first part, we introduced the letters of Hunmin chonc'um and the sound system of Middle Korean. In addition, we explained the method of transcription in Hunmin chonc'um of Chinese pronunciation by identifying the Hunmin chong'um letters used for the CP notation with the corresponding Chinese sounds they represent.

In the second part, we reconstructed the initial system of MM as represented by the PVP notation (see Tables

16 and 25). We compared the initial consonants represented by the VP notation (which has the same transcription of the 239 initial consonants as that of the CP notation) with those

represented by the PVP notation. Thus, we found that the

prominent changes from the former to the latter are as

follows:

(1) The former voiced obstruent initial consonants were

devoiced, accompanied by the tonal split of the Middle

Chinese (MC) even tone into the MP first and second tones

and the regrouping of the MC rising tone into the MP third

and fourth tones.

(2) The two guttural initials /?-, g -/ coalesced, and

joined the zero initial /^-/.

These two changes reflect the major differences between the

initial system of HWZY and that of the Zhonovuan vinvun

(ZYYY). Hence, the common assumption that the VP notation

represents the mid-fifteenth century Northern Mandarin must

be renounced as far as the initial system is concerned.

In addition, we also discovered the following

phenomena :

(3) The OM initial consonant /v-/ was in the process of merging into the MP zero initial /^-/ during the MM period.

(4) Phonemically, there was only one series of

retroflex initial consonants in OM and MM, i.e., /cr, crh,

sr, r/; but there were two types of allophones, i.e.,

[t/, t/h, J, J ] before [j] or [i], and [tp, tfh, g]

elsewhere. 240

(5) The palatal initial consonants— which are treated as allophones of the velar initial consonants before [j] or

[i] in MP— were not developed yet in MM.

In Chapter V, we reconstructed the final system of MM as represented by the PVP notation. We discussed in detail the phonetic values of the MM vowels. On the whole, Middle

Korean has more high vowels but fewer low vowels in comparison with OM and MM (of. Tables 12 and 26). For this reason, in the Korean pronunciation notations, MM high vowels are sufficiently differentiated, while MM low vowels are frequently underdifferentiated. Consequently, it was not a simple matter at all to determine the phonetic values of MM low nuclear vowels only on the basis of the pronunciation notations. In such cases, in order to help determine the phonetic values of Mandarin low vowel phonemes, we had to make the best use of the vowel systems of Middle Korean and MM as well as our knowledge of LMC, OM, and MP phonology.

In this way, we presented a full description of the

MM finals at both the phonetic and phonemic levels. In MM, there are four medials -j-, -w-, -jw-/, four nuclei

/-i-, -e-, -a-, -o-/, and five endings /-jS, -j, -w, -n, -j)/.

The forty-seven final types in the sound system of ZYYY are reduced to thirty-six final ti»pes in the sound system of MM as represented by the PVP notation of PN and PP (cf. Table

29 in Section 5.0 and Table 49 in Section 5.8). This 241 remarkable reduction of final types was caused by the merger of the bilabial stop ending /-m/ into the alveolar nasal stop ending /-n/ and the loss of the three-way contrast in depth of the OM low vowel phonemes before a non-zero ending.

The thirty-six final types in MM are further reduced to the thirty-four final types in MP due to the coalescence of /-e/ and /-o/ (the Mie-xie and Suo-po rhymes of the 'Thirteen

Tracks', respectively) into /-a/, which occurred during the period between MM and MP (cf. Table 7 in Section 3.3.3,

Table 32 in Section 5.0, and Table 49 in Section 5.8).

The OM vowel system reconstructed by Hsueh (1975) consists of four vowel phonemes, viz., /i, e, a, o/ (see

Table 26); while the MP vowel system consists of three vowel phonemes, viz., /i,d , a/. The MM vowel system is in the process of restructuring from the four-vowel system to the three-vowel system. Specifically, the three-way contrast in depth of the OM low vowel phonemes was lost in front of a non-zero ending before the time of DYTJ (1606). For example, the Han-shan /-an/, Huan-huan /-on/, and Xian-tian

/-en/ rhymes of ZYYY were in the process of coalescence at the time of the VP notation. First, /-o-/ completely merged to /-a-/ by a vowel-fronting change, namely. Rule (21) o — >> a / __ (E) (E = [-back]; If C = G, then E ^ p, t) (in

Section 5.1.4) before the time of ZYYY (1324) and the VP notation (ca. 1450). Second, /-a-/ between the palatal medial /-j-/ and the alveolar nasal ending /-n/ became /-e-/ 242 by a vowel-fronting change before the time of the VP notation. Subsequently /-a-/ and /-e-/ lost their phonemic contrast in this context, and respectively became the conditioned allophones [-%-] and [-£-] of /-a-/. Thus, the three low vowel phonemes /o/, /a/, and /e/ merged into /a/ before the alveolar nasal ending /-n/.

All the words of both the Dona-zhona /-ig/ and the

Geng-aina /-eji/ rhymes of ZYYY are transcribed with Korean high vowels in both the VP and PVP notations. This transcriptional evidence reveals that the two rhymes merged into one rhyme (e — ^ i / ji) before the time of the VP notation (ca. 1450). Furthermore, this transcriptional evidence suggests that the nuclear vowel of the ZYYY Gena- aina rhyme was probably a high vowel phoneme. This suggestion is also supported by the fact that most of the

Geng-aina rhyme words with the labial medial /-w-/ are also found in the Dona-zhona rhyme in ZYYY (cf. Hsueh 1975:53-54 and 59).

The words of the Zhen rhyme-set /-in/ are transcribed with four different Korean vowels in contrast. Their phonetic values are [i, i, u, ju], which respectively correspond to the high vowel phoneme /i/ preceded by the four types of the medial f

[i, u, y] in MP. This phonological characteristic of the

high vowel phoneme /i/ brings about an interesting

phenomenon. That is, in Mandarin, the words of the same

rhyme share the same nuclear vowel phoneme, which may be phonetically realized quite differently due to phonetic assimilation. Such phonetic realizations of the same

Mandarin phoneme can be so different from each other that they are transcribed with different Korean letters.

There are two closely-related dialectal strata or speech variants in the development of Mandarin pronunciation, mainly with regard to the development of the former entering tone syllables. One is called the

"colloquial" pronunciation, and the other is called the

"literary" pronunciation. The basic phonological structure of the former is, however, almost identical with that of the latter, except that the latter retained the entering tone category for some time in the form of the glottal stop ending /-?/ (of. Hsueh 1978).

In the CP notation, the MC entering tone words are transcribed with the stop endings [-p, -t, -k]. In the VP notation, the entering tone endings are transcribed uniformly with the glottal stop ending [-?] according to the

"literary" pronunciation except for the Yao rhyme words, which are mostly transcribed with the labio-velar semivowel 244 ending [-w] according to the "colloquial" pronunciation. In the PVP notation, the former entering tone words are transcribed without the glottal stop ending, but most of them are transcribed according to the "literary" pronunciation. HWZY represents the "literary" pronunciation, while

ZYYY represents the "colloquial" pronunciation with heavy borrowing from the "literary" pronunciation. MP follows the

"literary" pronunciation, but with heavy borrowing from the

"colloquial" pronunciation (Hsueh 1975:134 and Chou

1989:281-283). Historically, the national capital of the

Ming dynasty (1368-1644) was moved from Nanjing to Peking in

1421, and there was a flow of migrants from the South to the

North China Plain in the early fifteenth century. As a result of such political and social influence, the northerners probably adopted the "literary" pronunciation as the new speech norm, although they probably carried over some features of the "colloquial" pronunciation into the new speech norm. Therefore, Chou (1989) hypothesizes, on the basis of dialectal evidence, that the shift from the

"colloquial" pronunciation to the "literary" pronunciation occurred in Northern Mandarin due to dialect borrowing at that time. As is stated above, the VP and PVP notations already followed the "literary" pronunciation with borrowing from the "colloquial" pronunciation. From this fact, we can infer that the shift from the "colloquial" pronunciation to 245 the "literary" pronunciation must have occurred in Northern

Mandarin no later than the mid-fifteenth century. This inference confirms that the dialectal shift occurred at that time. BIBLIOGRAPHY

This bibliography is divided into two sections: 1 . Pre-modern works 2. Modern works. Since it is inconvenient to cite pre-modern works by author and date (both being often uncertain), they are arranged by title.

1. Pre-modern works

Chonoding Sima Wenaona Dencvun tulina % $J % ca. 1606. By Xu Xiao ^ . Gulin vunhui ^ ^ at ^ . ca. 1292. By Huang Gong-shao ^ 1" • Not extant.

Gulin vunhui 1uvao ^ j M ^ / 30 juan. 1279. By Xiong Zhong-ju . Abridgement of above. Edition of 1883, Huainan Shuju.

Guangyun , 5 juan. 1008. Compiled by Chen Peng-nian (961-1017) and others. Reprint of the Jiaozheng Songben Guangyun. 1970. Taipei; Yiwen Yinshuguan.

Hongmu chong'un vokhun 7^ ^ ^ ^ #!(l . 1455. Sin Suk-chu ÿ ^ et al. A Korean phonetic transcription is added to each homonym group in the Hongwu zhengvun.

Hongwu zhengyun . 1375. By Yue Shao-feng et al. Xylograph of 1548.

Hunmin chong'um ^ . 1446. By Sejong Original text of the Korean alphabet.

246 247

Hunmin chong'um haerye ^ll ^ . 1446. By Chong In-ji ^ / Sin Suk-chu ^ , et al. Explanation of the Hunmin chong'um.

Hunmong chahoe W ^ ^ . 1527. By ch'oe Se- j in ^ -j# Edition of 1912, Kwangmunhoe.

Jivun ^ fI , 10 juan. 1039. By Ding Du T / C (990-1053) et al. Edition of the Wanyou wenku, 1937, (Reprint of a Japanese edition of 1838). Shanghai; Commercial Press.

Jinadian shiwen ^ ^ , 30 juan. 583 A.D. By Lu De-ming pè (ca. 554-642).

Jinashi zhenavin Oievun zhin^ )£. ^ ^ ^ 1336. By Liu Jian. reprint of the edition of 1497 in Denavun minazhu wu zhona. 1972. Taipei; Taishun Shuju.

Kanmiu buaue Oievun jlj ^ , 5 juan. 706. By Wang Ren-xu 3L fh. ^$7 . Manuscript copy formerly preserved in the Imperial Palace, Peking, published by Tang Lan ^ , 1947, Peking. Reprint 1964, Taipei; Guangwen Shuju.

Kanaxi zidian È ^ . 1716. Edited by Zhang Yu-shu

Libu vunlue ^ * 1037. By Ding Du ^ et al.

Meneau zivun ^ , 2 juan. 1308. By Zhu Zong-wen ^ ^ >C • Manuscript copy in the British Museum, reproduced in Luo Chang-pei and Cai Mei-biao, Basiba zi vu Yuandai Hanvu. 1959. Peking; Kexue Chubanshe.

Noaoltae ^ . ca. 13th century. Authorship unknown. Not extant. Textbook of Mandarin Chinese. Ch'oe Se-jin's revision is reprinted together with the Noaoltae onhae and the Pakt'onasa onhae.

Noaoltae onhae ^ , 2 juan. 1670. Author unknown. Revision of the Ponvok Noaoltae of Ch'oe Se- jin, ca. 1515. Reprint of the edition of the Kyujangkak reprint series, 1938, Seoul. Reprint Seoul; A-se-a munhwasa, 1973. Taipei; Lianjing Chuban Shiye gongsi, 1978.

Noaoltae sinsok 'z "K. , 1 juan. 1761. Ed. by Pyon Hon ^ ^ et al. Revision of the Chinese text of the Noaoltae onhae: its Korean transcription is deleted. Also known as Sinsok Noaoltae 248 Noaoltae sinsok onhae ^ ca. 1761. Authorship unknown. Not extant.

Pakt'onasa . ca. 13th century. Authorship unknown. Not extant. Textbook of Mandarin Chinese.

Pakt 'onasa onhae , 3^uan. 1677. By Pyon Som Jâ / Pak Se-hwa iÊT # et al. Revision of the Ponvok Pakt'onasa of Ch'oe Se-jin, ca. 1515. Reprint of the edition of the Kyujangkak reprint series, 1938, Seoul. Reprint Seoul: A-se-a munhwasa, 1973. Taipei: Lianjing Chuban Shiye Gongsi, 1978.

Pakt'onasa sinsok ^ , 1 juan. 1765. By Kim Ch'ang-jo ^ et al. Revision of the Chinese text of the Pakt'onasa onhae; its Korean transcription is deleted. Reprint, 1972. Nagoya: Saika Shorin.

Pakt'onasa sinsok onhae ^ ^ % , 3 juan. 1765. By Kim Ch'ang-jo ^ ^ et al. Translation of the Pakt'onasa sinsok.

fPonvokl Noaoltae / 2 juan. ca. 1515. By Ch'oe Se-jin * * Korean transcription and translation of the Noaoltae. Reprint vol. 1, Seoul: Tae-je-gak, 1974. vol. 2, Seoul: A-se-a Chulp'ansa, 1980.

Ponvok Noaoltae, Pakt'onasa pomnve^ p , 1 juan. By Ch'oe Se-jin. Repr. Seoul: A-se-a Munhwasa, 1973. See Noaoltae Pakt'onasa chimnam.

(Ponvok) Pakt'onasa , 3 juan. ca. 1515. By Ch'oe Se-jin. Korean transcription and translation of the Pakt'onasa. Reprint vol. 1, Kyongbuk University, 1956 and Seoul: Tae-je-gak, 1974. Vols. 2 and 3 are not extant.

Qi vin lue HC ^ . ca. 1162. Juan 36 of the Tongzhi iji , 200 juan, an encyclopedia by Zheng Qiao (1108-1166). Reprinted in Denavun minazhu wu zhona. 1972. Taipei; Taishun Shuju.

Oievun '^57 / 5 juan. 601 A.D. By Lu Fa-yan ^ ^ ë . Manuscript fragments reproduced in Liu Fu, 1936, Shi vun huibian . Reprint 1963, Taipei.

Oievun kao ## 4 . 1842. By Chen Li ^ f# (1810- 1882). 249

Oievun kao waipian Written 1879-80, published, 1884. By Chen Li 7£ . Reprinted together with the Oievun kao in Yinvunxue conashu. Taipei: Guangwen Shuju, 1966

Oievun zhizhana tu f? ^ ^ Exact date and author unknown. According to Zhao Yin-tang (1957: 94-107), the work was composed between 1176 and 1203. (1) Edition of the Sibu Congkan, reproducing a Southern Song handcopy. (2) Edition based oijjyie text of the Siku quanshu. By Yan Shi-hui ^ K . 1930. Chengdu. The latter is reproduced in Denavun minazhu wu zhona, 1972. Taipei: Taishun Shuju.

Renzi xinkan Libu vunlue i. . 1252. By Liu Yuan

Sasona t'ona'ao . ca. 1446-1450. By Sin Suk-chu ]P . Not extant. Revised and expanded as next.

Sasona t'onahae 13 ^ • 1517. By Ch'oe Se-jin.

Shuowen 1iezi ^ ^ . ca. 100 A.D. By Xu Shen (died ca. 120 A.D.). Modern reprint 1972, Hong Kong: Zhonghua Shuju.

Sishena denazi 3 ^ ^ . Date and authorship unknown, probably Northern Song. Reproduced in the Denavun minazhu wu zhona. 1972. Taipei: Taishun Shuju.

Sishena Oievun biao (57 ^ ^ A. - By Jiang Yong f (1681-1762). ^

Taniahae % ^ , 1 juan. Repr. Seoul: A-se-a Munhwasa, 1973. See Noaoltae Pakt'onasa chimnam.

Tang vun ^ ^ , 5 juan. Two versions: 720 and 751. By Sun Mian ^ . Fragmentary manuscript reproduced in Liu Fu 1936, Shivun huibian 4- Jgfi ^ M e • Reprint 1963, Taipei. ' ^ ^

Tonaauk chona'un K ^ . 1447-48. By Sin Suk-chu et al. Chinese character dictionary for the standardized Sino-Korean pronunciation.

Wufana vuanvin Ï.A • Between 1624 and 1672. Originally by Fan Teng-feng and revised by Wei Yun-tai

Wuvin 1ivun ^ ^ . 1211. By Han Dao-zhao . 250

Xiru ermu zi ® ^ â ^ . 1626. By Nicholas trigault. Reprint 1957. Peking: Wenzi Gaige.

Yanshi iiaxun ^ ^ g/)| . By Yan Zhi-tui (531-595).

Yiaie ling vinvi ^

Yuniino % . Author and original date unknown. Extant text comes from a Southern Song edition by Zhang Lin- zhi , with preface dated 1161 and 1203, copied and reprinted in Japan. See Long Yu-chun 1960. Yunlue huitor^ ^ . 1642. By Bi Gong-chen % ^ • Reprinted together with the Yunlue vitona Taipei: Kuangwen Shuju. 1962.

Yunlue vitona ^ % . 1442. By Lan Mao ^ . See Yunlue huitona.

Zhonovuan vinvun ^ ^ "If. . 1324. By Zhou De-qing /f) ^ ~ In Hattori Shiro and Toodoo Akiyasu, 1958. See also Zhao Yin-tang 1956.

2. Modern works

Abbreviations

BIKP Bulletin of the Institute of History and Philology, Academia Sinica

BMFEA Bulletin of the Museum of Far Eastern Antiquities, Stockholm

CH Chindan Hakpo

CHHP Tsinghua Journal of Chinese Studies (Qinghua Xuebao)

HJAS Harvard Journal of Asiatic Studies

UAL International Journal of American Linguistics

JAOS Journal of the American Oriental Society 251

JCL Journal of Chinese Linguistics

JCLTA Journal of the Chinese Language Teachers Association

JL Journal of Linguistics

L g . Language /t-' YCHP Yenching Journal of Chinese (Yanjing Xuebao)

YYYJ Yuyan Yanjiu

ZGYW Zhongguo Yuwen

References

Anderson, J.M. 1973. Structural aspects of language change. London: Longman.

Allen, W.S. 1964. On one-vowel systems. Lingua 13, 111- 24.

Arisaka, Hideyo. 1957. Kokugo on'inshi no kenkyu [Studies on Japanese historical phonology]. Rev. and enl. by Kindaiti Kyoosuke. Tokyo: Sanseidoo. (1st ed. 1944)

Bach, Emmon and Robert T. Harms (eds.). 1968. Universals in linguistic theory. New York: Holt, Rinehart, and Winston.

Baldi, Philip and Ronald N. Werth (eds.). 1978. Readings in historical phonology. University Park, Pa.: The Pensylvania State University Press.

Baxter, W.H., III. 1977. Old Chinese origins of the Middle Chinese chongniu doublets: a study using multiple character readings. Ph.D. dissertation, Cornell Univ.

1987. Review of Middle Chinese by E.G. Pulleyblank 1984. HJAS 47.635-656.

Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart, and Winston. Reprint, The University of Chicago Press, 1984.

Bodman, Nicholas C. 1967. Historical linguistics. In T.A. Sebeok ed, 1967, pp. 3-58. 252

Bynon, Theodora. 1977. Historical linguistics. Cambridge University Press.

Chafe, W.L. 1968. The ordering of phonological rules. U A L 34.115-136.

Chang, Kun. 1968. Ancient Chinese phonology and the Ch'ieh-vun. CHHP, n.s. 10.61-82.

Chao, Yuen-ren. 1934. The non-uniqueness of phonemic solutions of phonetic systems. BIHP 4.363-97. Reprint in Joos 1957, pp. 38-54.

______. 1941. Distinctions within Ancient Chinese. HJAS 5.203-233.

1968. A grammar of spoken Chinese. University of California Press.

1975. Aspects of Chinese sociolinguistics Stanford University Press.

Chen, Chung-yu. 1984. Neutral tone in Mandarin: Phonotactic description and the issue of the norms. JCL 12.299-333.

Chen, Matthew Y. 1976. From Middle Chinese to Modern Peking. JCL 4.113-277.

Chen, M. and W. Wang. 1975. Sound change: actuation and implementation. Lg. 51.255-281.

Chen, Zhi-fan. 1981. Lun Choi Se-jin zai Chaoxian yuwen he Hanyu fangmian de gongxian [Discussion on Ch'oe Se- jin's contribution to the Korean and Chinese languages]. Minzu yuwen lunwenji [Collected papers on national (minority) languages], pp. 111-191. Peking: Zhongguo Shehui Kexue Chubanshe.

Chen, Zhi-wen. 1965. Guanyu Beijinghua li er-hua de laiyuan [About the origin of r-suffixion of Pekingese]. ZGYW 1965; May.396-370, 412.

Cheng, Chin-chuan. 1973. A synchronic phonology of Mandarin Chinese. (Monographs on Linguistic Analysis, 4) The Hague: Mouton.

Cheng, Robert. 1966. Mandarin phonological structure. JL 2.135-158. 253

Cheng, Tsai-fa. 1976. Early Mandarin as revealed in hP'ags-pa transcription and related materials. Ph.D. dissertation. University of Wisconsin.

Cheun, Sang-buom. 1975. Phonological aspects of Late Middle Korean. Ph.D. dissertation. University of Indiana.

Ch'oe, Yong-ae. 1975. Honowu zhenavun yanjiu [A study of Honowu zhenavun1. Doctoral dissertation. National Taiwan University.

Chomsky, Noam and Morris Halle. 1968. The sound pattern of English. New York: Harper and Row.

Chong, Yon-ch'an. 1972. Honomu chonoun vokhun ui yon'gu. Seoul: IIcho-gak.

Chou, Ghi-zhen. 1985. On the nature of the rusheng words in the Yunlue huitono. JCLTA 20, no. 2.41-63. ______. 1989. Hono-wu zheno-vun: its relation to the Nanjing dialect and its impact on Standard Mandarin. Ph.D. dissertation. The Ohio State University.

DeFrancis, John. 1984. the Chinese language: Fact and Fantacy. Honolulu: University of Hawaii Press.

Denlinger, Paul B. 1961. Chinese historical linguistics: the road ahead. JAOS 81.1-7.

Ding, Bang-xin. (See Ting, Pang-hsin)

Ding, Sheng-shu. 1958. Gujin zijin duizhao shouce [A handbook of correspondences between Ancient and Modern pronunciations]. Peking. Reprint 1966, Hong Kong: Taiping Shuju.

Dong, Shao-wen. 1958. Yuin changshi [General knowledge of phonetics]. 2nd printing, 1964, Peking: Wenhua Jiaoyu Chubanshe.

Dong, Tong-he. 1968. Hanyu yinyunxue [Chinese phonology]. Taipei: Guangwen Shuju. Rev. and enl. ed. of Zhongguo yinyunshi [History of Chinese phonology]. 1954. Taipei.

Dragunov, A.A. 1930. The hPags-pa script and Ancient Mandarin. Izvestija Akaderoii Nauk SSSR, 7th series, Otdelenie gumanitarnyx nauk 9.627-647, 10.775-797. Tr. into Chinese by T'ang Yu, 1959, Basiba-zi yu gu Hanyu. Peking: Kexue Chubanshe. 254

Dyer, Svetlana Rimsky-Korsakoff. 1983. Grammatical analysis of the Lao Ch'i-ta with an English translation of the Chinese text. Faculty of Asian Studies Monographs: New Series, no. 3. Canberra: Faculty of Asian Studies, Australian National University.

Egerod, Soren. 1967. Dialectology. In T.A. Sebeok ed. 1967, pp. 91-129.

Elman, Benjamin. 1982. From value to fact: the emergence of phonology as a precise description in late imperial China. JAOS 102.493-500.

Fodor, J.A. and J.J. Katz eds. 1964. The structure of language: Readings in the philosophy of language. Englewood Cliffs, N.J.: Prentice-Hall.

Forrest, R.A.D. 1948. The Chinese language. London: Farber and Farber.

Halle, Morris. 1962. Phonology in generative grammar. Word 18.54-72. Reprinted in Fodor and Katz 1964, pp. 334-352.

1964. On the bases of phonology. In Fodor and Katz 1964, pp. 324-333. Reprinted in Makkai (ed.) 1972, pp. 393-400.

Halliday, M.A.K. 1959. The language of the Chinese Secret historv of the Mongols. Publication of the Philological Society, vol-. 17. Oxford: Blackwell.

Hartman, Lawton M., III. 1944. The segmental phonemes of the Peiping dialect. Lg. 20.28-42.

Hashimoto, Mantaro J. 1965. Phonology of Ancient Chinese. Ph.D. dissertation. The Ohio State University. Reprinted as nos. 10 and 11 of Study of Language and Cultures of Asia and Africa Monograph Series, Tokyo, 1978 and 1979.

1970a. Notes on Mandarin phonology. In Studies in general and Oriental linguistics, ed. by R. Jakobson and S. Kawamoto eds., pp. 207-220.

. 1970b. International evidence for Ancient Chinese palatal endings. Lg. 46.336-365.

1977. Current developments in Sino-Korean studies. JCL 5.102-125. 255

1981. Review of The lexicon in phonological change (ed. by William S-Y. Wang). Lg. 57.1.183-191.

Hattori, Shiro, and Toodoo Akiyasu. 1958. Chuuaen on'in no kenkyuu, kyoohon hen [Study of the Zhonovuan vinvun. revised text]. Tokyo: Konan Shoin.

Ho, Ung. 1965. Kug-o um-un-hak [Korean phonology]. Seoul: Chong-um-sa.

Hock, Hans H. 1975. Current trends in historical linguistics. Paper presented at Kentucky Foreign Language Conference.

Hockett, Charles F. 1947. Peiping phonology. JAOS 67.253- 267.

Hsueh, F.S. 1975. Phonology of Old Mandarin. (Janua linguarum, 179) The Hague: Mouton.

______. 1976. Seven centuries of rhyming in Chinese poetry: a linguistic interpretation. Paper presented at the annual meeting of the Chinese language Teachers Association, New Orleans, La.

1978. Lun ru-sheng zi zhi yanhua guilu [Discussion on the diachronic rules of the entering tone words]. In Qu Wan-li Xiansheng qizhi rongqing lunwenji [A festschrift for Dr. Qu Wan-li on his seventieth birthday], pp. 407-433.

1980a. The phonemic structure of Pekingese finals and their r-suffixation. BIHP 51.491-514.

1980b. Lun Zhi Si yun de xingcheng yu yanjin [A Discussion on the formation and development of the Zhi Si rhymes]. Zhongguo Shumu Jikan, vol. 14, no. 2.53- 75.

. 1982. Lun yinbian yu yinwei jiegou de guanxi [The relation between sound change and phonological structure]. YYYJ 1982, no. 2.11-17.

1983. Xu Xiao's Revised Rhvme-Tables: a bold renovation- In Chu; Coblin, and Tsao (eds.). Papers from the 14th International Conference on Sino-Tibetan Languages and Linguistics, 1981, pp. 151-170. Taipei.

1985. Shilun dengyun-xue zhi yuanli yu nei/wai zhuan zhi hanyi [The theoretical basis of the rhyme- tables and the meaning of their inner/outer distinction]. YYYJ 1985, no. 1.38-56. 256

. 1986. Beijing yinxi jieshi [Analysis of Pekingese phonology]. Peking: Bejing Foreign Language Institute Press.

1987. Historical phonology and dialect study: some examples from the Pingdu dialect. In A fest­ schrift for Dr. Wang Li on his eightieth birthday, pp. 221-243. The Chinese Language Society of Hong Kong.

Hu, Ming-yang. 1963. Laoaida vaniie he Putonashi vaniie zhong suo jian de Hanyu Chaoxianyu duiyin [Corresponding sounds between Chinese and Korean as shown in the Noaoltae onhae and Pakt'onasa onhae1. ZGYW 1963.3. (no. 124).185-192.

Hyman, Larry M. 1975. Phonology: theory and analysis. New York: Holt, Rinehart, and Winston.

Im, tong-sok. 1982. Chaoxian yixue kao [A study of translation studies in the Yi dynasty]. Doctoral dissertation. National Taiwan Normal University.

Jakobson, Roman, C. Gunnar Fant, and Morris Halle. 1951. Preliminaries to speech analysis: the distinctive features and their correlates. MIT Press. 5th printing, 1963.

Jakobson, Roman, and Morris Halle. 1956. Fundamentals of language. (Janua linguarum, 1) The Hague: Mouton.

Jakobson, Roman and Kawamoto Shigeo (eds.). 1970. Studies in general and Oriental linguistics. Tokyo: TEC.

Jeffers, Robert J. and U s e Lehiste. 1979. Principles and methods for historical linguistics. MIT Press.

Joc3, Martin (ed.). 1957. Readings in linguistics I. University of Chicago Press.

Kang, Sil-chin. 1985. Laoaida Putonashi vaniiu [A study of Noaoltae and Pakt'onasa!. Doctoral dissertation, National Taiwan Normal University.

Kang, Sin-hang. 1973. Sasona t'onahae yon'gu [A study of the Sasona t'onahae!. Seoul: Sin-a-sa.

1974. Ponvok Noaoltae. Pakt'onasa chi umgye. [The phonological system of the Ponvok Noaoltae and Ponvok Pakt'onasa1. CH 38.123-150. 257

. 1978a. Noaoltae. Pakt'onasa onhae nae cha-um chi umgye [The sound system of the transcriptions in the Noaoltae onhae and Pakt'onasa onhae1. Tongbang Hakchi 18.61-86. Seoul: Yonse University.

• 1978b.1 3 /0 0 . Pakt'onasarajcx; onasa sinsoKsinsok onnaeonhae nae cna-umcha-um uiu i umgye [the sound system of the transcriptions in the Pakt'onasa sinsok onhae1. Haksulwon Nonmunjip 18.79 102.

. 1980a. Hwa-um kvemona onhae nae cha-um ui umgye [The sound system of the transcription in the Hwa-um kvemona1. Tongbanghak-chi 23 & 24.167-192. Seoul: Yonse University.

1980b. Yiju Chaoxian ziliao lueji jindai Hanyu yuyinshi [A sketch of the phonological history of Mandarin Chinese based on Korean data]. BIHP 51.515-534.

1983. Ch'ium gwa Han'aul p'yogi [Sibilants and Korean transcription]. Kug-o-hak 12.13-34.

Karlgren, Bernhard. 1915-26. Etudes sur la phonologie chinoise. Trans, into Chinese by Y-R. Chao, F-K. Li, and C-P. Luo, Zhongguo yinyunxue yanjiu. Changea: Shangwu Yinshuguan, 1948. 2nd edition Taipei, 1962.

1940. Grammata Serica: script and phonetics in Chinese and Sino-Japanese. BMFEA 12.1-471.

1954. Compendium of phonetics in Ancient and Archaic Chinese. BMFEA 26.211-367

. 1957. Grammata Serica recensa. BMFEA 29.1-332

Keiler, A.R. (ed.). 1972. A reader xn historical and comparative linguistics. New York: Holt, Rinehart, and Winston.

Kenstowicz, Michael J. and Charles Kisserberth W. 1979. Generative phonology. New York: Academic Press.

Kim, Chin-wu. 1968. The vowel system of Korean. Lg. 44.516-527

______. (ed.). 1978. Papers in Korean linguistics. Proceedings of the symposium on Korean linguistics, (University of Hawaii, 1977). Columbia, S.C.: Hornbeam Press. 258

Kim, Kong-on. 1978. Vowel system of Korean revisited. In Kim C. W. ed. 1978, pp. 75-83.

Kim-Renaud, Youngkey. 1978. The syllable in Korean phonology. In Kim C. W. ed. 1978, pp. 85-98.

Kim, Wan-jin. 1976. Noaoltae onhae yon'gu [A study of the Noaoltae onhae1. Seoul; Han'guk Yon'guwon.

King, Robert D. 1969. Historical linguistics and generative grammar. Englewood Cliffs, N.J.: Prentice- Hall.

Kiparsky, Paul. 1968. Linguistic universals and linguistic change. In E. Bach and R. Harms ed., Universals in linguistic theory, pp. 171-202. Also in Keiler 1972, pp. 338—367.

Koono, Rokuroo. 1968. Choosen kanjion no kenkyuu [A study of Sino-Korean pronunciation]. Tenri: Tenri Jihoosha.

Labov, William. 1981. Resolving the Neogrammarian controversy. Lg. 57.267-308.

Ladefoged, Peter. 1975. A course in phonetics. New York: Hartcourt Brace Jovanovich.

1980. What are linguistic sounds made of? Lg. 56.485-502.

Ledyard, Gari Keith. 1966. The Korean language reform of 1446: the origin, background, and early history of the Korean alphabet. Ph.D. dissertation. University of California, Berkeley.

Lee, Ki-moon. 1961. Kug-o-sa kaesol [An outline of the history of the Korean language] Seoul: Minjung Sogwan.

1972. Kug-o umunsa yon'gu [A study of Korean phonological history]. Seoul: Han'guk Munhwa Yon'guso, National Seoul University. Reprint as Kug-o- hak ch'ongso, no. 3 by Kug-o-hak-hoe, 1977, Seoul: T'ap Ch'ulp'ansa.

Lehiste, Use. 1970. Suprasegmentals. MIT Press.

Lehmann, Winfred P. ed. 1975. Language and linguistics in the People's Republic of China. Austin; University of Texas Press. 259

Li, Charles N. and Sandra A. Thompson. 1981. Mandarin Chinese: a functional reference grammar. Berkeley: University of California.

Li, Fang-kuei. 1966. The zero initial and the zero syllabic. Lg. 42.300-312.

1971. Shangguyin yanjiu [A study of Archaic Chinese phonology]. CHHP, n.s. 9.1-61.

. 1973. Language and dialects. JCL 1.1-13.

Li, Kung-pu. 1964. Machine recognition of Mandarin monosyllables. (Project on Linguistic Analysis no. 3) The Ohio State University Research Foundation.

Li, Rong. 1951. Oievun yinxi [The sound system of the Oievun1. 2nd ed., 1956. Peking: Kexue Chubanshe.

Li, Xin-kui. 1963. Guanyu Zhongyuan yinyun yinxi de jichu he ru-p'ai-san-sheng de xingzhi [Comments on Zhao and Zeng 1962]. ZGYW 1963.4.275-281.

1981. Yuniino yanjiu [A study of Yunlinol. YYYJ 1.125-166. Wuhan: Huazhong Gongxueyuan,

1983. Zhonovuan vinvun yinxi yanjiu [A study of the sound system of the Zhonovuan vinvun1. He-nan Sheng, China: Zhongzhou Shuhuashe.

Light, Timothy. 1976. The Chinese syllabic final: Phonological relativity and constituent analysis. Cornell University East Asia Papers, no. 9.

Long Yu-chun. 1960. Yuniino jiaozhu [Emendation of the Yunlinol. Taipei: Yiwen Yinshuguan.

Lu, Zhi-ji. 1985. Notes on the structure of finals and the vowel system of Mandarin Chinese. JCLTA 20, no. 2. 33- 40.

Lu, Zhi-wei. 1946. Shi Zhonovuan vinvun [Explication of the Zhonovuan vinvun1. YCHP 31.35-70.

______. 1947a. Ji Lan Mao Yunlue vitono [Remarks on Lan Mao's Yunlue vitono1. YCHP 32.161-168. Reprint Lu Z. W. 1971, vol. 2. pp. 223-230.

1947b. Ji Xu Xiao Chonodino Sima Wenoono Denovun tulino [Remarks on Xu Xiao's rime table]. YCHP 32.169-196. Reprint Lu Z. W. 1971, vol. 1. pp. 187- 214. 260 . 1947c. Ji Bi Gong-chen Yunlue huitona [Remarks on Bi Gong-chen's Yunlue huitona1. YCHP 33.105-113. Reprint Lu Z. W. 1971, vol. 2. pp. 213-221.

1947d. Jin Ni-ge Xiru ermu zi suo ji de yin [The sounds transcribed in Nicholas Trigault's Xiru ermu zi1. YCHP 33.115-128. Reprint Lu Z. W. 1971, vol. 2. pp. 231-244.

. 1948a. Ji Wufana vuanvin [Remarks on the Wufana vuanvin1. YCHP 34.1-13. Reprint Lu Z. W. 1971, vol. 2. pp. 245-257.

1948b. Guoyu ru-sheng yanbian xiaozhu [Remarks on the development of the entering-tone of the 'National Language']. YCHP 34.21-28. Reprint Lu Z. W. 1971, vol. 2. pp. 259-266.

1971. Hanyu yinyunxue lunji [Collected papers on Chinese phonology]. 2 vols. Hong Kong: Chongwen Shudian.

Luo, Chang-pei. 1933. Shi nei/wai zhuan [An explanation of the inner/outer distinction]. BIHP 4.209-226.

1963. Luo Chang-pei yuyanxue lunwen xuanji [Selected papers of Luo Chang-pei in linguistics]. Ed. by Zhongguo Kexueyuan Yuyan Yanjiusuo. Peking; Zhonghua shuju.

Martin, Samuel E. 1953. The phonemes of Ancient Chinese. Supplement to JAOS, no. 16.

1957. Problems of hierarchy and indeterminacy in Mandarin phonology, BIHP 29.209-229.

Martinet, Andre. 1952. Function, structure, and sound change. Word 8.1-32.

Moulton, William G. 1962. Dialect geography and the concept of phonological space. Word 18.23-32.

Mei, Tsu-lin. 1970. Tones and prosody in Middle Chinese and the origin of the rising tone. HJAS 30.86-110.

______. 1977. Tones and tone sandhi in 16 th century Mandarin. JCL 5.237-260.

Min, Yong-kyu. 1964. Nogoltae pyonui [Discussion about the Noaoltae1. Inmun Kwahak 12. Seoul: Yonse University. 261

Moon, Yang-soo. 1974. A phonological history of Korean. Ph.D. dissertation, The University of Texas at Austin.

Nakano, Miyoko. 1971. A phonological study in the 'Phags- pa script and the Mena-ku Tzu-vun. (Faculty of Asian Studies: Oriental Monograph Series, no. 7) Canberra: Australian National University Press.

Nam, Kwang-u. 1969. Choson (Yi-jo) Hanchaum yon'gu [A study of Sino-Korean pronunciation of the Yi dynasty. ______. 1972. Sinbalgyon in Ch'oi Se-jin cho Ponvok Noaoltae kwon sang ul pogo [Upon seeing the first volume of the newly discovered Ponvok Noaoltae of Ch'oi Se-jin]. Kug-o Kungmunhak, no. 56 and 57. Seoul.

Oota, Tatsuo. Rookittai no gengo ni tsuite [About the language of the Nogoltae]. Chuugoku Gogaku Kenkyuukai Ronshu (vol. 1) 1953, pp. 1-14. Nara, Japan.

Ogura, Simpei. 1920. Choosen gogakushi [History of the Korean language]. Rev. and enl. by Koono Rokuroo as Zootei hochuu Choosen gogakushi [Expanded and annotated edition of the History of the Korean language] Tokyo: Tookoo Shoin, 1964.

Pak, Pyong-ch'ae. 1971. Ko-dae Kug-o ui yon'gu, umun p'yon [A study of ancient Korean]. Seoul: Korea University

Pak, T'ae-gwon. 1974. Choi Se-jin yon'gu [A study of Choi Se-jin]. Pusan: Ch'inhaksa.

Peking University. 1962. Hanyu fangyin zihui [A list of Chinese characters with dialectal pronunciations]. Peking: Wenzi Gaige Chubanshe.

Pike, Kenneth L. 1948. Tone languages. University of Michigan Press. 9th printing, 1977.

Postal, Paul M. 1968. Aspects of phonological theory. New York: Harper and Row.

Pulleyblank, E.G. 1970-71. Late Middle Chinese. Asia Major 15.197-239 and 16.121-168.

______. 1978. The nature of the Middle Chinese tones and their development to Early Mandarin. JCL 6.173- 203.

1984. Middle Chinese: a study in historical phonology. The University of British Columbia Press.

Pyon, Tae-sop. 1978. Han'guksa yoron [Important discussions about Korean history]. Seoul: Samyongsa. 262

Ramsey, S, Robert. 1987. The languages of China. Princeton University Press.

Schane, Sanford A. 1973a. Generative phonology. Englewood Cliffs, N.J.; Prentice-Hall.

1973b. [back] and [round]. In Stephen R. Anderson and Paul Kiparsky, eds., A Festschrift for Morris Halle, pp. 174-184. New York: Holt, Rinehart, and Winston.

Scovel, Thomas. 1964. A distinctive feature analysis of the phonemic segments of Mandarin. M.A. Thesis, The Ohio State University.

Sebeok, Thomas A. et al. (éd.). 1967. Linguistics in East Asia and South East Asia. Current Trends in linguistics, vol. 2. The Hague: Mouton.

Shibatani, Masayoshi. 1973. The role of surface phonetic constraints in generative phonology. Lg. 49.87-106.

Stanley, Richard. 1973. Boundaries in phonology. In S.R. Anderson and P. Kiparsky eds., A Festschrift for Morris Halle, pp. 185-206. New York: Holt, Reinhart, and Winston.

Stimpson, Hugh M. 1962a. Phonology of the Chunavuan vinvun. CHHP, n.s. 3.114-159.

1962b. Ancient Chinese -p, -t, -k endings in the Peking dialect. Lg. 38.376-384.

1966. The Jongyuan in yunn. New Haven, Conn.; Yale University Far Eastern Publications.

1977. Review of Phonology of Old Mandarin by F.S. Hsueh. Lg. 53.940-944.

1978. Old Mandarin dialects. Old Pekingese, and the rhymes of the Zhonovuan vinvun. In T'ang T'ing- chih et al. eds.. Proceedings of Symposium on Chinese linguistics, 1977, Linguistic Institute of the linguistic Society of America, pp. 221-232. Taipei: Xuesheng Book Co.

Suyama, Nobuo. 1973. Haku tsuii genkai.- Ro Kittai oenkai goi sakuin [A vocabulary index for the Pakt'oncsa onhae and the Noaoltae onhae1. Nagoya : Saika Shorin. Tang, Yu. 1932. Er yin de yanbian [The diachronic change of the sound er]. BIHP 2, no. 4.457-467. 263

Ting, Pang-hsin. 1980. Review of F.S. Hsueh's Phonology of Old Mandarin. JAGS 100.94.

1981. Yu Zhonovuan vinvun xiangguan de ji zhong fangyan xianxiang [Some dialectal phenomena related to the Zhonovuan v i n v u m . BIHP 52.619-650.

Toodoo, Akiyasu. 1952. Kanwa no seritsu katei kara mita Seiiu iimokushi [the Xiru ermu zi as seen through the formative process of Mandarin]. Toohoogaku 5.99-122.

______. 1960. Ki-, tsi- no koogaika wa juuhasseiki ni okotta [The palatalization of ki- and tsi- occurred in the eighteenth century]. Chuugoku Gogaku 94.1-3.

1964. Development of Mandarin from 14th c. to 19th c. Acta Asiatica 6.31-40.

Ts'ou, Benjamin K. 1976. A single character index to the Noooltae onhae. Computational analysis of Asian and African languages, no. 4. Tokyo: National Inter- University Research Institute of Asian & African Languages & Cultures.

Walton, A. Ronald. 1983. Tone, segment, and syllable in Chinese: a polydimensional approach to surface phonetic structure. Cornell University East Asia Papers, no. 32.

Wang, John C. and Hsueh, F.S. 1973. The Lin-ch'i dialect and its relation to Mandarin. JAGS 93.2.136-145.

Wang, Fu-shi. 1963. Beijinghua yunmu de ji ge wenti [Some problems of the Pekingese syllabic final]. ZGYW 1963.2. (no. 123).115-124.

Wang, Li. 1935. Zhongguo yinyunxue [Chinese phonology]. Reprinted as Hanyu yinyunxue [Phonology of the Chinese language]. Peking: Zhonghua Shuju, 1956.

______. 1958. Hanyu shi gao [A draft history of the Chinese language]. 3 vols. Peking: Kexue Chubanshe. Revised edition, 1980, Peking: Zhonghua Shuju.

1979. Xiandai Hanyu yin fenxi zhong de ji ge wenti [Some problems in the analysis of modern Chinese sounds]. ZGYW 1979.4.281-286.

Wang, William S.Y. 1967. Phonological features of tone. U A L 33.93-105. 264

1969. Competing sound change as a cause of residue. Lg. 4.9-25.

1979. Language change: a lexical perspective. Annual Review of Anthropology 8.353-371.

Weinreichf Uriel. 1953. Languages in contact. New York: Linguistic Circle of New York. Reprint 1964, The Hague: Mouton.

Wu, Zhong-qi. 1980. Shi lun Putonghua yuyin de qubie tezheng ji qi xianghu guanxi [A preliminary study of distinctive features and their correlations in Standard Chinese]. ZGYW 1980.5 (no. 158).321-327.

Yang, Lien-sheng. 1957. Laoaida Putonashi li de yufa yuhui [The grammar and vocabulary as in Laoaida and Putonashi1. BiHP 29.197-208. Reprinted in Laoaida vaniie Putonashi vaniie. Taipei; Lianjing, 1978.

Yi, Sung-nyong. 1959. Honomu chonaun vokhun ui yon'gu. [A study of the Honomu chonaun vokhun1. CH 20.115-179.

Yi, Tong-lim. 1968. Honomu chonaun vokhun kwa Sasona t 'onohae ui pigyo yon'gu [A comparative study between the Honomu chonaun vokhun and the Sasona t'onohae1. Tongguk Taehakkyo Nonmunjip, vol. 5. Seoul: Tongguk University.

Ying, Yu-kang. 1970. Honawu zhencvun shengmu yinzhi zhi niding [Reconstitution of the initial sounds of the Honawu zhenovunl. Zhonghua Xueyuan 6.1-35. Zhongguo Wenxue Yanjiusuo, National Zhengzhi University, Taiwan.

You, Ru-jie, Qian Nai-rong, and Gao Zheng-xia. 1980. Lun Putonghua de yinwei xitong [A discussion of the phonemic system of Mandarin]. ZGYW 1980.5.328-334.

Yu, Chang-gyun. 1974. Mona'go unnvak kwa Sasona t'ono'oo ui yon'gu [A study of the Mona'ao unnvak and the Sasona t 'ona'ao1. Seoul: Hyungsol Ch'ulp'ansa.

______. 1981a. Tona'auk chona'un yon'gu [A study of the Tong'guk chong'un], revised edition. Seoul: Hyongsol Ch'ulp'ansa.

1981b. Sin'go Kug-o-hak-sa [A history of the Korean language]. Seoul: Hyongsol Ch'ulp'ansa.

Yuan, Jia-hua et al. 1960. Hanyu fangyan gaiyao [An outline of the Chinese dialects]. Peking: Peking University. 265

Zhao, Xia-qiu and Zeng Qing-rui. 1962. Zhonovuan vinvun yinxi de jichu he ru-pai-san-sheng de xingzhi [The foundation of the phonological system of the Zhonovuan vinvun and the nature of the split of the entering tone into the other three tones], ZGYW 1962.7.312-324.

Zhao, Yin-tang. 1936. Zhonovuan vinvun vanliu TA study of the Zhonovuan vinvun1. 2nd printing, 1956, Shanghai.

1957. Dengyun yuanliu [Origins of rhyme table phonology]. Shanghai: Shangwu yinshuguan.

Zhu, De-xi. 1958. Laooida 1anlie Putonoshi vanlie shuhou [A note on Noooltae onhae and Pakt'onosa onhae1. Beijing Daxue Xuebao 1958.2.69-75.