• Author / Uploaded
• rivetrenuck

Citation preview

•

ABC CHINESE DICTIONARY SERIES Victor H. Mair, General Editor The ABC Chinese Dictionary Series aims to provide a complete set of convenient and reliable reference tools for all those who need to deal with Chinese words and characters. A unique feature of the series is the adoption of a strict alphabetical order, the fastest and most user-friendly way to look up words in a Chinese dictionary. Most volumes contain ~raphically oriented indices to assist in finding characters whose pronunciation is not known. The ABC dictionaries and compilations rely on the best expertise available in China and North America and are based on the application of radically new strategies for the study of Sinitic languages and the Chinese writing system, including the first clear distinction between the etymology of the words, on the one hand, and the evolution of shapes, sounds, and meanings of characters, on the other. While aiming for conciseness and accuracy, seties volumes also strive to apply the highest standards of lexicography in all respects, including compatibility with computer technology for information processing.

ABC Etymological Dictionary of Old Chinese Axel Schuessler

Other titles in the series

ABC Chinese-English Dictionary (desk reference and pocket editi()ns) Edited by John DeFrancis ABC Dictionary of Chirme Proverbs Edited by John S. Rohsenow ABC Chinese-English Comprehensive Dictionary Edited by John DeFrancis

A Handbook of 'Phags-pa Chinese W. South Coblin

University of Hawai'i Press Honolulu

.. CONTENTS

PREFACE .................................................................................................................. xi ACKNOWLEDGMENTS ........................................................................................... xiii ARRANGEMENT OF THE DICTIONARY ................................................................. xv SYMBOLS AND ABBREVIATIONS ........................................................................ xvii © 2007 University of Hawai'i Press

All rights reserved Printed in the United St:ates of America 12 11 10 09 08 07

Library of Congress

6 5 4 3 2 1

Cataloging~in-Publication

data

Schuessler, Axe!. ABC etymological dictionary of old Chinese I Axe! Schuessler p. cm. -

(ABC Chinese dictionary series)

Includes bibliographical references and index. ISBN-13: 978-0-8248-2975-9 (cloth : alk. paper) ISBN-10: 0-8248-2975-1 (cloth : alk. paper) 1. Chinese language-Etyrnology-Dict:ionaries·-English

I. Title.

1.1.1 1.2 1.2.1 1.2.2 1.2.3 1.2.4 1.2.5 1.2.6 1.2.7 1.2.8 1.3 1.3.1 1.4 1.4.1 1.4.2 1.4.3 1.4.4

OLD CHINESE AND ETYMOLOGY ........................................................... I Chinese ..................................................................................................... 1 Sources of Old Chinese ............................................................................... 1 Old Chinese and its linguistic neighbors ........................................................ 1 Chinese and Sino-Tibetan ............................................................................ 2 Tibeto-Burman languages ............................................................................ 3 Miao-Yao .................................................................................................. 3 Tai -Kadai, Kam-Tai .................................................................................. 3 Austroasiatic ............................................................................................ 4 Vietnamese ............................................................................................... 4 "Northern" Austroasiatic ............................................................................ 5 Sun1mary .................................................................................................. 5 Old Chinese dialects .................................................................................. 6 Rural dialects ............................................................................................ 7 The study of Old Chinese etymology ............................................................ 7 Approaches to word families and cognates .................................................... 8 Approaches to etymology through the graph .................................................. 9 Identification of cognates ............................................................................ 9 The present approach ................................................................................ I 0

2 2.1 2.1.1 2.1.2 2. 1.3 2.2 2.3 2.3.1 2.3.2 2.4 2.4.1 2.4.2 2.4.3 2.4.4 2.5 2.5.1 2.6 2.6.1 2.6.2 2.7 2.8

MORPHOLOGY AND WORD DERIVATION ........................................... l2 Grammatical relations in Old Chinese ......................................................... 12 Word order ............................................................................................. 12 Word class .............................................................................................. 12 Derivation and word class ......................................................................... 14 Types of derivations and allofams ....... ,..................................................... 14 Sino-Tibetan morphology ........................................................................... 15 The nature of Sino-Tibetan affixation ......................................................... 15 Sino-Tibetan morphemes ............................................................................ 16 Morphemes in Old Chinese ........................................................................ 17 Historical layers of morphemes in Old Chinese ........................................... 17 Suffixes in Old Chinese ............................................................................. 17 Sino-Tibetan prefixes in Old Chinese ........................................................... 18 I nfixation ................................................................................................ 19 Parallel roots and stems ............................................................................ 20 Parallel stems of 'swell' ............................................................................ 20 Austroasiatic morphology in Old Chinese ................................................... 22 Austroasiatic infixes in Old Chinese ........................................................... 22 Austroasiatic word families in Old Chinese ................................................. 23 Expressives, reduplication ......................................................................... 24 Non-morphological word formation ............................................................ 25

I 1.1

IL Series

PL128l.S38 2007 495.1'321-dc22

2005056872

University of Hawai'i Press books are printed on acid-free paper and meet the guidelines for permanence and durability of the Council on Library Resources.

Camera-ready copy prepared by the author.

Printed by IBT Global V

CONTENTS

CONTENTS

25 25

5.6

26 26 27 27

5.8 5.8.1 5.8.2 5.8.3 5.8.4

3 3.1 3.2 3.2.1 3.2.2 3.2.3 3.2.4 3.3 3.3.1 3.3.2 3.3.3 3.4 3.5

MC TONES AND THEIR OLD CHINESE EQUIVALENTS ......................... 29 Middle Chinese tone A (pfngsheng ¥~) ..................................................... 29 Middle Chinese tone B (shangsheng J:.~): phonology ..................................... 30 Tone B from Sino-Tibetan *-? ..................................................................... 30 Tone B for Tibeto-Burman final *-k ........................................................... 31 ST *-?in closed syllables ........................................................................... 32 Tone B for foreign final -I] ........................................................................ 32 Tone B as morpheme ................................................................................ 33 Tone B (I): terms for body p~rts and humans ............................................... 33 Tone B (2): coverbs and particles ............................................................... 34 Tone B (3): independent pronouns .............................................................. 34 Middle Chinese tone C (qusheng *!t): phonology ......................................... 35 Tone C: later OC general purpose morpheme ................................................ 36

5.8.5

4 4.1 4. 1.1 4.2 4.2.1 4.3 4.3.1 4.3.2 4.3.3 4.4 4.4.1 4.5 4.5.1 4.6 4.6.1

Tones B, C, and voicing: direction and diathesis ........................................... Direction and diathesis ............................................................................. Direction and diathesis in Old Chinese ........................................................ Tone C (qusheng $;~): two morphological functions .................................... The Sino-Tibetan sources. of tone C ............................................................. Tone C (1): exoactive derivation ................................................................. Tone C: exoactive - extrovert, ditransitive ............................................... Tone C: exoactive- transitive, causative I putative .................................... Residue ................................................................................................... Tone C (2): exopassive derivation ............................................................... Exopassive as a transitive verb ................................................................... Tone B (shangsheng J:.~): endoactive derivation .......................................... Tone B: endoactive nouns .......................................................................... Voicing of the initial consonant: endopassive derivation ................................ Residue .................................................................................................

5 5. 1

INITIAL CONSONANTS .......................................................................... De voiced initials ...................................................................................... Devoicing of ST initial *z-> MC s-............................................................ Sino-Tibetan *s-prefix ............................................................................... Causative s-prefix >Middle Chinese s- ........................................................ Causatives-prefix> MC voiceless initial. .................................................... Iterative s-prefix >MC s-, ~-, voiceless initial.. ........................................... Nouns with *s- >MC s-, voiceless initial. .................................................... Devoicing and PTB *r-.............................................................................. ST and PCH *k- ....................................................................................... Other sources of devoicing .......................................................................

2.8.1 2.8.2 2.8.3 2.8.4 2.9 2.10

Re-analysis .............................................................................................. Backformation, re-cutting ......................................................................... Metathesis ............................................................................................... Convergence ........................................................................................... Meaning and sound ................................................................................... Semantic extension ...................................................................................

5. 1.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.4

5.5

vi

38 38 40 41 42 42 43

44 45 45 46 46 48 48 50

51 51

51 52

52 52 53 54

55 56 56

5.7

5.8.6 5.8.7

5.9 5.9.1 5.9.2 5.9.3

5.9.4 5.10 5.10.1 5.10.2 5.10.3 5.10.4 5.1 0. 5 5. 11 5. 12 5.12.1 5.12.2 5.12.3 6 6. 1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.2.3 6.3 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.4.5 6.5 6.5.1 6.5.2

6.6 6.7 6.8 6.9

MC initial x- from voiceless acute initials ................................................... 57 MC affricates from *s + stop consonant. .................................................... 58 Aspirated initial consonants ...................................................................... 58 MC aspiration: loss of ST pre-initial. .......................................................... 58 MC aspiration: causative .......................................................................... 59 MC aspiration: iterative ............................................................................ 59 MC aspiration: auxiliary verbs .................................................................. 60 Aspiration: outward and I or forceful motion .............................................. 60 Aspiration: hollow, empty ......................................................................... 61 Aspiration in foreign words ....................................................................... 61 Aspiration from PCH consonant clusters .................................................... 61 MC tsh- from *k-s- .~nd *s-?- ...................................................................... 61 MC ts 11 - from s +voiceless sonorant. .......................................................... 62 MC ts 11j- from OC clusters *k-hl-................................................................ 62 MC aspiration from other types of PCH initial clusters ................................. 63 Reflexes of Mon-Khmer affricates in Chinese .............................................. 63 MK c, j = MC affricates ........................................................................... 63 MK c, j ==MC retroflex affricates .............................................................. 63 MK er-, jr- =MC retroflex affricates ......................................................... 64 MK c, j ==MC tsj-, tj-, etc. from OC *t(r)j- ............................................... 64 MK c, j ==velar initials k, g ...................................................................... 65 MC?- from foreign kl-type clusters ............................................................ 65 Nasal initials ........................................................................................... 65 IJ(w)- ~ nw- ............................................................................................ 65 Chinese m- for TB and foreign b- ............................................................... 65 Austroasiatic nasal infix........................................................................... 67 FINAL CONSONANTS ............................................................................. 68 Final *-k ................................................................................................. 68 k-extension ............................................................................................ 69 Suffix -k: distributive ............................................................................... 70 Final -t ................................................................................................... 70 Nouns with final -t ................................................................................... 70 Final *-t, *(t)s with grammatical words ...................................................... 72 Final -t =foreign final-s ........................................................................... 72 Final -s ................................................................................................... 72 Final -n .................................................................................................. 72 Final -in I -it ........................................................................................... 72 Final *-un from *·UIJ ................................................................................ 74 Nominalizing suffix -n ............................................................................. 74 Final -n with verbs .................................................................................... 75 Pronominal final -n .................................................................................. 76 Final -1) ................................................................................................... 76 Final ·I) as a morpheme: terminative ........................................................... 76 Final ·I) and open syllables ......................................................................... 76 Final stop consonant- nasal ...................................................................... 77 Dissimilation with labial finals -pI -m ...................................................... 77 OC final -i ............................................................................................. 78 Absence of final consonant after long vowel.. .............................................. 79

vii

' CONTENTS

CONTENTS

7 7.1 7.1.1 7.1.2 7.1.3 7.1.4 7.1.5 7.2

7.2.1 7.2.2 7.3 7.4 7.5

7.6 7.6.1 7.6.2 7.7 7.7.1 7.7.2

7.7.3 7. 7.4 7.7.5

8 8.1 8. 1. 1 8.1.2 8.1.3 8. 1.4 8.1.5 8.2

8.2.1

8.2.2 8.3 9 9. 1 9.1.1 9. 1.2 9.1.3 9.1.4 9.2 9.2.1 9.3 9.4

10 10.1 10.1.1

OLD CHINESE AND FOREIGN *r .......................................................... 80 OC *r as reflected in MC initial consonants ................................................ 80 MC initial!- ............................................................................................ 80 OC voiceless *r-....................................................................................... 81 MC retroflex initials ................................................................................. 81 MC initial 'borrowed from'.

< [~'~which has been retained in OC, but has merged with ST *a in TB. There are occasional hints, however, that the TB proto-language might also have made this distinction (STCp. 183, n. 482). Innovations unique to CH do not establish a fundamental split in the ST family, they

1.2.4 Tai-Kadai, Kam-Tai Kam-Tai languages (KT) are not related to Chinese and ST (Oiti Qlngx.ia 1991 ). Massive lex.ical exchanges in both directions between Chinese and Tai, from OCto more recent dialects, have led some investigators to conclude otherwise. In the distant past, people speaking these languages likely Iived in areas as far north as the Yangtze River basin. For example, the ruling family of the ancient state of Chu ~had the clan name xi6ng ff~ 'bear', but in the Chu language the name was ml~, which is the KT word for 'bear'. Today, though, KT people live farther to the south in Gu£mgxf, Gulzhou, and southern Humin (Pulleyblank 1983: 429ff). There have been significant exchanges of vocabulary in both directions between MK (including Viet-Muong) and Tai languages; Lao especially has many loan words from its Vietnamese neighbor. Tai languages also have relatively recent Khmer loans, an earlier layer of loans from M on, and loans from an even older Northern AA language that today is represented by Khmu and that the Tai must have overlain at some early date (Ferlus 1978:

2

3

c

1.2.5 - 1.2.6

OLD CHINESE AND ETYMOLOGY

OLD CHINESE AND ETYMOLOGY

13-14, n.22). Consequently, some alleged Chinese-Tai lexical sets turn out to be spurious, the Tai words being loans from Khmer or elsewhere. For example, Siamese suanA 1 'garden' has nothing to do with~ ywin Ill *wan 'garden', but is borrowed from Khmer swna /suugn/ 'care for, look after> flower I pleasure garden'. Or consider Tai JuaiJ 'royal'; it reflects the Khmer word lu;;,IJ 'king' and has no (direct?) connection with Chinese --t huang .:§! 'august, royal'. An example of a KT word in OC is chin 8 *dran 'farm, farmyard', from Tai: Siamese aanA2 < *rtanA' Kam-Sui (PKS) *hra:n I 'house'.

1.2.5 Austroasiatic The Austroasiatic (AA) language family is unrelated to STand Chinese. AA languages fall into two major groups: Munda (exclusively on the Indian subcontinent); and Mon-Khmer (MK) scattered over Assam, Southeast Asia, and SW China and includes the Mon language in Burma, and Khmer in Cambodia. As only MK languages could have left traces in China, the terms MK and AA are often synonymous here. AA loans have been identified in TB l"anguages such as Lepcha (Forrest 1948) and in languages in Assam such as the Tani group (J. Sun LTBA 16:2, 1993: 165); AA lexical material is also encountered in Lushai (in this dictionary), in the TB Kanauri-Almora language Raji (Shanna 1990, vol. III, part Il: 170-228), as well as transparent Khasi loans in Mikir. MK influence in Old Chinese and ST has also received some attention (Shorto 1972; Ferlus 1998; LTBA 22:2, 1999: 1-20; Schuessler 2003; 2004; studies by Norman and Mei). Languages from at least two AA branches or layers have contributed to prehistoric and perhaps early historic Chinese: an early Viet-Muong language similar to Vietnamese (that may be called 'Viet-Yue') (§ 1.2.6) and a language (or languages) in the Yellow River basin that shows affinities to the modern Khmer and Khmu branches of MK, and on occasion also to Mon (§ 1.2.7). Purely historical and philological considerations also point to the prehistoric and early historic presence of AA in parts of northern China. The ancient Yf ~people, who lived in the east from the Shandong peninsula south to the Yangtze, were probably AA (Pulleyblank 1983: 440ft). The ancient Yue lf!& people in ZhejHing were certainly AA; the place Langye fJtE.ffi~ in Shandong was their traditional cultural center ( Yue jue shu; Eberhard 1968: 414ft). Under the year 645 BC, the Zuozhuim quotes a line from the famous Y1]ihg where we find the AA word for 'blood', huang ifu *hmal) (PAA *mham or the like) substituted for the usual ST etymon xue .rtil (Mei 1980). The deliberations in which context this line is quoted and apparently understood by all participants took place north of the Yellow River in today's Shanxi. Huang cannot have been a CH innovation, rather it must have been a survival from an earlier substrate language that was replaced by a ST layer, i.e., 'Chinese' as we know it. When pursuing OC and TB I ST etyma down to their apparent roots, one often seems to hit AA bedrock, that is, a root shared with AA. 1.2. 6 Vietnamese In addition to the significant influx of Chinese loans from antiquity to more recent times, Vietnamese has incorporated a large contingent of Tai words (Maspero 1912: 115). A language close to Vietnamese was spoken in SE China as late as the Han period by the ancient Yue ~-people (Yue OC *wat, the 'Viet' in Vietnam); it left a residue ofViet-Yue words in the modern Min dialects in Fujiim province (see articles by Norman and Mei, also quoted in Schuessler 2004). Early Chinese commentators have stated that the words zhi ifL 'epidemic' and sou JJl 'dog' are from the ancient Yue language (Pulleyblank 1983: 438f), but these might 4

1.2.7 - 1.2.8

have come from "northern" AA instead (see §1.2.7); Han period scholars merely noted the &imilarity with the Yue words of which they happened to be aware. Unlike the later M1n dialects, QC does not include many words that compel us to conclude that the source was specifically Vi et-Yue.

1.2. 7 "Northern" Austroasiatic An AA substrate ("AA-OC") contributed a significant number of AA words as well as fragments of AA morphology to prehistoric and subsequent CH (§2.6; §5.10). MK words gradually trickled from a substrate into mainstream ST-based OC over hundreds or thousands of years, so that layers and various MK sources can be discerned. The earliest, prehistoric layer of AA items is already encountered in the language of the first written records, the OB (1250-1050 BC); OC borrowings from this remote past occasionally do not agree very closely with MK phonologically (though in a regular fashion). For example, ch!.i ~ *tshro 'hay' vs. PMonic *ks:)::JY (cf. below), hU f.,T£ *hlii? 'tiger' vs. PMK *kla?. More "recent" items (found in BI, Sh1}ing, and then later texts) agree more closely with AA forms, e.g., cuo 1 ~ *tshoih 'hay' vs. PMonic *ks:>Jy (cf. above), jiang :?'I *krol) 'river' vs. PMonic *krooiJ. Many such OC words appear to be very similar to Khmer. This does not mean that the MK substrate was Khmer, but only that Khmer happens to have preserved (and I or scholars happen to have provided) data that provide suggestive comparisons with OC, just as the great number of Tibetan -OC comparative sets reflect more on the availability of Tibetan data, but not necessarily on a close historic relationship. 1.2.8 Summary The OC lexicon has many sources (Schuess.ler 2003). A few sample ST vs. non-ST words fol1ow, tci provide an impression (for details and explanations consult the dictionary entries): Animals: ST words: 'ox' gang !t!N), 'dog' quan j(, 'rhinoceros' xf ~,'horse' ma ,~,'fowl' yan ~(quail), 'louse' shf£il,\,, 'muntjac' jl~ Non-ST words: 'elephant' xiimg ~, 'dog' gou .jl:fl, 'buffalo' sl .7[',, 'chicken' jr~t, 'tiger' hti 'pig' tuan ~.'pig' shl ~,'small deer' zhl ~ Body parts: ST: 'head' ymin /C, 'head' shou §, 'eye' mu §,'hair' shan :B~. 'fem. breast' ru ~L, 'bitter I liver' xfn $, 'forehead' e ~''blood' xue IDl Non-ST: 'gall' dan ~l, 'forehead' sang~], 'blood' huang ._g Others: ST: 'root' b~m 7-Js:. 'forest' lfn .1*, 'firewood' xfn if, 'house' jia ~, 'temple' zong 'day' rl B, 'year' nian if,, 'breath' xf }i~L 'eat I meal' can ~ Non-ST: 'root' gen ;flt 'forest' lu ii, 'palace' gong 0 ';§, 'farm' chan Jll, 'temple' miao !lm, 'moon' (goddess) heng-e ~s93l 'year' r(m f,"2:, 'breath' ql ~, 'eat/ meal' xiang ~~ Numerals and grammatical words generally are ST: 'two' er=. 'copula' wei it, 'behind' hou f~ Of uncertain provenance: 'Wood' mu*· 'mountain' shan ilJ. 'flower' hua ::{t

m.

*·

5

1.3

OLD CHINESE AND ETYMOLOGY

OLD CHINESE AND ETYMOLOGY

1.3 Old Chinese dialects Languages which are spread over large areas and mountainous terrain naturally develop regional varieties; stratified societies also exhibit differences in speech along class lines. The OC language of the Shang and Zhou period and subsequent Classical Chinese was a standardized written language without noticeable regional flavors. The Chinese script would have hidden differences in pronunciation that might have existed, just as today B 'day' is read ri in Mandarin, jut in Cantonese. Yet one catches a few glimpses of language variation within OC when comparing the Shl]ihg W'#m rimes, the phonetic series and the later Middle Chinese (MC) as reflected in the Qic~yiin t)]~~ dictionary (AD 60 1) as well as modern dialects. First, in the Qieyun and modern dialects, as well as in the OC phonetic series there are certain words with the OC analogue rimes *-el] such as ming ::6 'name' which had in the Shijrng the rime *-in. The ST rimes *-il) I *-ik became either *-el] I *-ek or *-in I *-it in OC; which way a word went depended presumably on the dialect. Thus we find for ST *-i!] I *-ik the OC rime *-el) I *-ek: mfng ~ 'name', rrrfng P,~ 'to sound', ming 1fP 'order', sheng ±.'live'; but xrn 'firewood', ji ~ *tsit 'masonry'.

if

1-1

ST

Later South

QYS/MC

Shijing

not

--

--

*]);} > bu

*]);} > bu

*m a

*ma

--

not not have not have name

*r-miiJ

dark

*mil)

night green

*mial)

1'

1'

*ma>wU #.\

---

--

*mal)7>wang

*me!) > mfng iS

*m in iS

~

*meiJ > mfng

~·

*me!) > ming

~A:

~

*meiJ > mfng

~

(MK mal))

*ma.I]

*meiJ >mfng

*C-seiJ

*tshaiJ

*tshei] > qfng ~

*ts heiJ > q fn g 111

*tsharJ > cang ht

*tshalJ > cang ~

green

a

mother

*mo

*mo? >mu

go-between

MK dm::lj

'*m~>mei J1li1;

*m~?

>mu

a

:

*m~>mei J1li1;

1.3 . 1 - 1.4

have the word wang ~L *mal]? for 'not have, there is no'; only near the end of the estern Zhou period is it replaced by the familiar ~ *ma. These and similar phenomena suggest a language that is far from uniform, but we cannot ,fell whether these are individual preferences, or class or regional distinctions, nor if the latter, from which regions.

wu

Rural dialects Additional phonological oddities in OC may also be the result of dialectal differences. MC ~,nd, by backward projection, OC, has multiple phonological correspondences for what one surmises ought to be a single OC phonological configuration. Words with rare and unusual features typically have meanings with a rustk or vulgar flavor. We will, therefore, for now call tnis strain (or strains) 'Rural' as opposed to 'Standard', i.e., literary OC. The following phonological peculiarities may be identified as Rural: {1) OC voiceless initials *r-, *1-, and *n- are normally reflected in MC coronal t\ sj-, and, the case of *r-, in MC th-, th- (§5.1). However, in a few words such a voiceless continuant has yielded MC x-, xj-, and its equivalents in modern dialects. This unexpected development to a guttural initial is found in words that relate to ordinary, especially rural, life; they include words for: beard, to face I toward, ribs (of a horse), to know, to vomit, to rear animals, stupid, roar, tiger, pig ( §5 .6). To differentiate the two developments of voiceless initials, we will write OCM *lh-, *nh-, *rh- for MC th-, sj-, etc., but OCM *hn-, *hl-, *hr- when it is the ~spiration that survives as MC x-. Of course, voiceless *hi]-, *hm-, and *hw- regularly yield MC x-, thus any voiceless initial that shows up as x- in MC is written in OCM with the *hpreceding the sonorant. (2) Standard OC and foreign initial *1- (>MC ji-), or *l in the initial, have in some words merged with *r- (>MC 1-). This might be another Rural feature; examples in §7.3 include: salt, turtle, grain I to sow, bamboo. The *1 =Rural OC *r equation is often encountered in loans from non-ST languages, e.g., eel, splint hat, barrier I bolt, descend, frost; or the confusion of laterals may be due to the late date of borrowing in either direction. (3) Some non-ST words with initial *kl- have MC initial t- which may have been *tl- in OC. Such words include: Carry

tJi

dan [tom] 'to carry on the shoulder' AA: Khmu? kl2m 'carry on the shoulder'

For more examples and comments, see §8.2.1. (4) MC initial gj- and U- stand in a few correspondence sets for a foreign initial r, or r in combination with labial or velar consonants (§7.1.4). The semantic range .of such items conjures up a rural sphere: farm, pheasant, old man, to fall, bamboo, sickle, wrist, etc. (5) Some modern southern dialects have in their colloquial layers the vowel a for standard ?·This trend seems to be foreshadowed in some OC words which have the vowel a also for foreigneor i; seeTable 1-1 above, and§ll.l.3.

Secondly, OC labial-initial syllables of the type *P~ and *Po merged into *P;) in the Shi}ihg dialect(s) and the phonetic series, but remained distinct in the Qieyun and modern dialects (Baxter 1992); for example, we have the Mandarin readings mei !£ij 'each' vs. mu fB: 'mother' (same phonetic, same Sh1)ihg rime). Finally, a strain of OC must have retained ST *main the meaning 'not' because it is preserved in modern southern dialects, but does not exist in Shang and Western Zhou texts, apart from an occasional occurrence in classical texts. Table 1-1 illustrates these and additional differences within OC. Choice of words in individual texts often shows particular preferences that may be due to dialects. For example, in theZuozhuan :b:::{$ we find the interrogative xr~ *ge 'how' instead of he i6J *gai. In some chapters of the Shiijing if#~ the words for 'you' and 'your' are ru #£. and nai J'J respectively; in others, the word for both 'you' and 'your' is Pm. Later texts replace words common in earlier ones, e.g . , the OB, BI, and some parts of the Shl)ihg and

The study of Old Chinese etymology A Chinese word may have one of several origins: (I) It can have been inherited from the hypothetical Sino-Tibetan proto-languag~ when it has cognates among the related TibetoBurman languages. (2) It can be a loan from another language, or can have survived from an earlier substrate (Miao-Yao, Kam-Tai, Austroasiatic I Mon-Khmer). (3) It can be the result of

6

7

er

1.4.1

OLD CHINESE AND ETYMOLOGY

OLD CHINESE AND ETYMOLOGY

internal innovation, i.e., word derivation by morphology, internal borrowing from dialects, or phonological change. A word is usually assumed to be genetically related to another because of transparent or impressionistic phonological and semantic similarity. The range of sound alternations within an OC wf will be suggested throughout the introductory sections. Members of a wf, i.e., 'allofams' (Matisoff's term, alias 'cogeners') typically differ in tone, initial voicing (e.g., *kens 'to see' 3~ *gens 'appear'), and I or the Middle Chinese division (deng ~. i.e., vocalism, e.g., MC kill) vs. kjaiJ; see §9.1). Occasionally, they also differ in the vowel, in initial consonant(s) or final consonant. Since much concerning STand Chinese morphology is still not well understood, the terms 'wf' and 'allofam' are often fuzzy but conventional catch-all categories. For example, it seems obvious that the wordsjiim ~ *krihns 'look at' and !an 'I; *ram? 'to see' are related, but what the difference in later tones and the presence I absence of an initial *kmight have entailed is so far a matter of speculation. On the other hand, we can confidently state that zhi *~ *t;;lkh or *t::lks, literally 'something that has been woven', is a regular exopassive derivation from zlld~ *t~k 'to 'heave'. We consider bothjian and Jan, and zhfand zhi to be allofams in their respective word families. 1.4.1 Approaches to word families and cognates Investigators have differed significantly over the range of sound alternations within a word family. Karlgren (1933) allows for a broad range: a word family could have a final of the type -K, -T, or -P, etc. in conjunction with the initial consonant type K-, T-, N-, or P-, etc. where Tincludes any acute initial consonant, i.e., any which is not a guttural or labial. For instance, his wf with items 242-262 (1933: 69) has a root T-K and includes the following words (Karlgren's 'archaic Chinese', i.e., OC; in parentheses OCM): yang ~ *diaiJ (*la!]) 'light' 3~ zhao time'~~ xfng £ *sieiJ (*se!]) 'star'

s:g *~iog (*tau) 'bright' 3~ zhOu. *tiog (*trukh) 'day

1.4.2 - 1.4.3

dialect interference, or to convergence in which the initial n- would be associated with 'soft', just as words with initial gl~ typically suggest something 'gliding, glossy' in English (§2.9). 1.4.2 Approaches to etymology through the graph The above approaches start with the OC word while the graph that writes it is of secondary concern. However, approaching etymology from the other end by emphatic reliance on the graph is fraught with the danger of misinterpretation or overinterpretation. This leads occasionally to "strained explanations of loangraph meanings as semantic extensions" (Qiu Xigui 2000: 287); it has been suggested, for example, that Iii *'wheat' and Jai *'to come' are the same word ('wheat' is the cereal that 'came' from abroad), but the two are unrelated. Boodberg (1937: 339-341) went so far as to suggest that even graphic elements that are nearly universally recognized as semantic and not phonetic play a phonological and etymological role; for example, he believes that graphs written with the element zi § derive from a root *BDZi,..., *BSI:zi El 'self', biM- 'nose',xi}~, 'breathe'. The traditional source for the interpretation of ancient graphs is Xu Shen's Shuowen jiezi (SW) of ea. AD 150. But this is explicitly a dictionary of graphs, not words; it often describes a graph, which is not the same as an etymological explanation. For example, the SW (and also GSR 1166c) explains jiao 1;~ [kau 8 ] 'burn on a pyre of crossed logs' as cognate to jiao j.t [kau] 'to cross'. But the definition 'burn on crossed logs' could well have been suggested by the graphic element 'to cross'; therefore the word may have had just the meaning 'to burn' and be related to relevant TB items, but not to 'to cross'. Xu Sht:m also was unaware of the earlier forms of graphs as they are known today from the OB inscriptions; he was inadvertently misled by the graphic forms available at his time. Thus he explains the left element in the graph for she ~-j 'to shoot' as sben ~'body'; the shen element, however, goes back to the OB image of a bow with an arrow (Qiu Xigui 2000: 55±). We study the phonetic series and composition of graphs with interest because they often offer etymological clues, but two words are not a priori assumed to be etymologically related just because they share a phonetic element. In the end, every one of the above approaches contributes to interesting discoveries.

In this proposed wf, the OC initials, as understood today (Baxter), are *1-, *t-, *s-; the vowels are *a, *au (OCB *-aw), *e, *u; the finals are *-I), *-k, *-V (vowel). The TB cognate for yang is *laiJ (e.g., WB laiJ 8 'be bright'), zhOu is clearly cognate to WT gdugs (< g-duk-s) 'midday, noon'. These two TB items are certainly not related. Therefore, Karlgren's phonological parameters are much too broad. Cognates usually share the same rime and initial consonant type. However, in many instances an obvious cognate has a different final or rime, or initial variation outside the normal spectrum. LaPolla (see §6) has dedicated a study to ST rimes and finals. In order not to go off in all directions, investigators prefer to keep to a given rime and allow the initials to vary, or keep to one category of initials and then allow for variations in finals. Wang Ll (1958: 542-545) provides examples for both approaches: same initial but different rimes (such as the negatives with initial *m-), and same rime but different initials (such as rime *-a!) 'bright'). Or note a wf proposed by Pulleyblank (1973: 121) (traditional MC forms in parentheses): r6u (flzj::lU) ~~ ruan ~ (flzjwan 8 ) ~< nuo '1'1 (flzju, nz:jwanB, nuanC) 3< ruo ~~ (flzjak), all meaning 'soft', but he has not included ren ff (nzj~mB) 'soft'. w ang Ll (1982) splits this particular group into one with a tendency toward final velars, and one with final dentals. Thus the set ru tiE (nzjwo 8 ) 3~ ruan g;g (nzjwanB) 3< nen, nun ~ (p. 571) is distinct from r6u (nzj~u) 3~ ruo ~~ (nzjak) 3~ rau ~ (nzjuk) 'meat, flesh' (p. 236). As long as we do not know more about oc morphology, we cannot tell if distinctions in this wf are due to morphological derivation,

1.4.3 Identification ofcognates Beside morphological patterns which are discussed throughout the introduction, the following considerations also help in the identification of etymological connections (seealso §2.10). Matisoff's Conclusion to his HPTB (pp. 535-542) could be quotes here in full as well. Semantic parallels strengthen the case for the identification of etymological relations. For example, since jihg ~ 'capital city' also means 'mound, hill', it is likely that qiii Ji 'village, town' is also the same word as the homophone qiii 'mound, hill'. Settlements are often built on higher ground. Cognates from related TB languages sometimes help identify connections within Chinese. For example, 'naked' luo tl *roi?, and cheng :fj *dreiJ are probably cognate to such forms as PTB *groy > WT sgre-ba, and WT sgren-mo 'naked', Lushai (eenR 'bare', respectively. As the TB items derived from the same root, Chinese forms may have as well (cheng from PCH *(d)roi-1) ?). On the other hand, the correct identification of cognates is sometimes impeded by one or another type of interference or obstacle, as follows. Etymological investigation is hampered or helped by the investigator's native language and

8

9

*

*

1.4.4

OLD CHINESE AND ETYMOLOGY

OLD CHINESE AND ETYMOLOGY

culture. A native speaker of Chinese would with little hesitation, and probably correctly, equate ya 5f 'sprout' with ya 3f 'tooth', while this connection might not be self-evident to speakers of European languages. The composition of a Chinese characterinterferes occasionally with the semantic understanding of the word behind it (see also §1.4.2). Yu ~~'leftover, rest' is usually thought to mean originally 'food leftovers' because it is written with the radical shf'to eat'. Yet the radical may have been chosen because concrete food leftovers were easier to represent graphically than the abstraction 'remainder, rest'. Thus 'food leftovers' is merely one semantic extension of the word. The Chinese writing system is not alphabetic, although a phonetic element in the majority of graphs provides some clue for a word's OC sound. But there is disagreement on some details of OC reconstruction, especially about the initial consonants. Depending on whose OC system one follows, one may arrive at startlingly different etymologies; for example, wei M£ (MCjiwi) 'to be' is reconstructed *r~d by Li Fang Kuei who relates this then to WT red-pa 'to be', but reconstructed *wjij by Baxter, which turns out to be related to PTB *w~y 'to be'. Our investigations are based on Baxter (1992), many uncertain details notwithstanding. Variant forms are common occurrences in dialects, i.e., bifi B 'colloquial' vs. wen 'literary' forms, such as Mandarin col. ta 1-(g 'he, she, it' vs. lit. tu a 'other'. These are lexically two different words but historically one and the same etymon, no ablaut morphology derived one from the other. This phenomenon is so ubiquitous in China that one might expect this to have occurred already in ancient and archaic times. Subjective judgment slips into etymological consideration easily because of the monosyllabic nature of the words (countless words have the syllable structure CV) and the often diffuse and fuzzy field of meanings that Chinese words and graphs have accumulated over millennia. Even when the meaning is specific or when the syllable structure is complex, it is occasionally difficult to decide what is related to what. Two illustrations: (!) T6u !I] *do 'head' agrees exactly with TB-PL *du 'head' (PLB *u = PTB, ST *o). But it agrees equally well with a MK etymon: note Khmer /dool/ 'head'; a MK final consonant is often lost in OC after a long vowel, hence the equation is also perfect. Which is related to Chinese? Are both CH and PL descended from MK? (2) CMn ~ *sran or *sren 'to produce' strikes one as the obvious cognate of WT srel-ba 'to raise, bring up'. The Chinese word even has a counterpart with initial *m- in the word miiin ~!JG *mran 'give birth', thus forming a well-known ST pair *s- (transitive I causative)- *m- (intr.). But then Khmer has a word /samraal/ (i.e., *s-m-raal) 'to give birth', derived from riila /r{i;}]/ 'to increase, ... distribute, propagate'. On the one hand, Chinese is closely related to TibetoBurman; on the other, Mon-Khmer provides a possible etymology for both OC words, i.e., a root from which the items in question could be derived, while there is no TB counterpart to mifm. Is the Chinese wf ST or AA? Or do both Chinese and WT go back to the same area etymon?

Y:.

to the whim of a writer (see more in §12.1.2). Thus zhur{t: OCM *tui (GSR 57 5; OCB 'a bird' was selected to write 'to be' which could have been due to all kinds of mental processes and associations. MCjiwi points to OC *wi 'to be'; PTB *w~y (or *wi?) 'to be' confirms this. Thus the comparative method as well as MC point to OC *wi and nothing else, ,,1 sgyur-ba 'cause to change, transform' tr. (Beyer 1992: 116). This function includes the formation of verbs from nouns, e.g., JP mji1] 33 'name' > j;;J 3l.mji1]3J 'to name' (more JP examples in Dai I Wu 1995: 95). Later in PCH and OC, causatives were formed with the exoactive tone C (§4.3), and eventually in OC and later, simply by word order (§2.1.2). It has been suggested that causativity and intensive I iterative are two aspects of one original morpheme as, for example, in Indo-European languages where both are expressed by the same suffix *-eyo- (W. Lehmann 1993: 168), as in Greek words ending in -eo like 'make tremble' (Palmer 1980: 266), or Vedic lndic patayati 'flies about' vs. pi'itayati 'causes to fly' (A. MacDonelll916: 195), with the Skt. iterative I causative morpheme -ya-. However, for practical purposes, the two functions are distinct in CH. Prefix *mis a ST morpheme which tends to mark introvert I intr. verbs, which contrast with ST *s- for causative I extrovert verbs (Matisoff LTBA 15:1, 1992). It survives in the rare initial MC dij(LH z-) from earlier *m-1-. This m-prefix, apparently already unproductive in PCH, is very old because in OC we usually observe direct TB cognates with this morpheme, even in contrasting pairs (see above; §2.3 .2, Table 2-4; §8.1.4).

2.4.4

qf -t; *tshit < *snhit? 'seven', lit. 'two again' PTB *snis 'seven' < ST *nis 'two'

J< (huaiB) *hm:::li?

Fire

Frost Suffix-t k t 1 b"ects·itis . . ST morpheme productive in Proto-Chinese, which typically mar s na u~a o ~ , , ' 1s a found redundantly , rfn *h Wit , 'bl oo d, , yue n *IJot moon often attached to nouns, e.g., xue, lill (§6.2.1).

2.4.4

Infixation

OC and ST had no infixation. An exception may be OC medial *r which derives causative verbs (§7.5), e.g., chli ~lli [thut) *-rut 'to expel'< cha tl:\ [t§hut] 'to come out'. It is not certain whether this *r was an infix or prefix. If the source was ST, it may be related to the r-prefix as found in WT; alternatively, some MK languages have a causative r-infix. Since chuwas prob. 19

2.5 - 2.5.1

MORPHOLOGY

MORPHOLOGY

OCM *k-hlut, it is difficult to imagine what a cluster with *r might have sounded like. 2.5 Parallel roots and stems A difficulty in TB and ST historical linguistics and reconstruction is the frequent existence of two or more stems or words with similar meaning and similar phonological shape (examples below). They are difficult to reconcile by known phonological or morphological processes, even though they look like members of a wf. This is also the case in other language families in the area, including MK about which Shorto (1973: 375) observes: ..... it is noticeable that within the general range of meanings encountered in each of the three series [i.e., word families that he discusses] there is only limited agreement between the specific meanings assigned to a given variant or derivate in different languages. This implies a marked tendency to semantic shift in phonaesthetic series, a tendency which would itself help to explain the retention of variants as distinct lexemes in individual languages." Pending further insights, we will assume parallel roots and stems for ST (Shorto's variants and derivates) and individual branches.andlanguages to account for the occurrence of not quite homophonous stems which are distributed rather erratically over CH and TB languages (Matisoff 1978, esp. p. 21 ). Since CH has often inherited these parallels from the ST parent language, their etymology cannot be uncovered within OC. As in MK, parallel stems often differ in vowels (a - e/i), in the presence or absence of medial *r, or both. Examples include (also in §11.1.2-3; throughout Matisoff 1978): ST *sriiJ 'live' ( ~ sheng 2 ~) - ST *sraiJ 'Ii ve' (WB hraiJ) ST *sal) 'clear' (WT sal) ) oc an, ~*?an 'calm' ST *(r)wa 'rain' ( ~ yu 3 ffi) ST *ka 'solid' (--+ gu 1 ~) ST *traiJ (?)'stretch' (--+ zhiing 1 WB *kruik 'meet with' JP kru?55 < kruk55 'to sprout' JP kro(55 < krok55 'to hatch' OC chU 4 *hruk, *rhuk 'nourish' OC eh~~ *k-hla? 'dwell' ST *lap(s) 'foliage, generation' (--+ ye4 ~) ST *l(j)am 'blaze' (--+ yan 2 ~)

*)

=if

1'f) if)

-ST *sil) 'alive, green'(-+ qfng 1 - ST *sal) 'alive, green' ( ~ cang 3 - ST *sel) 'clear' (-~ qing 2 rM) - OC yan 1 *~ *?ens 'be at ease • - ST *(r)we 'rain' ( ~ yun 2 ~~7Jl.J() -ST *kar 'solid' (~gu 1 1l!!J) - WT thaiJ 'id.' - WT khug-pa - khugs-pa 'to find, get' -JP ku?55 < kuk55 'id. • , - kou ~ *khokh 'chick' ( ~ g~ 14 ~) - ju 5 :If® *kuk 'nourish •, \ - ju 2 @ *kah 'id.' - WT rabs 'I ineage' - ST *warn 'blaze' (-) yan 3 ~)

These variants cannot be the result of early ST dialectal evolution, because they do not correspond in a systematic way to later branches of ST; also, on occasion, both are found in the same language, frequently CH. We may speculate that Ur-ST complex sound clusters might have been responsible, thus ST *sril) vs. *k-srii]? > *sring vs. *k-siiJ; or ST *lwam ? > *lam - *warn. Eventually, simple explanations may be found.

2.5.1

Parallel stems of 'swell'

Parallel stems can be quite numerous and present a picture of meaning I stem distributions that looks rather boundless and chaotic. However, considering the overall semantic field of a parallel stem (related or not) can sometimes clarify CH and ST etymologies and even 20

*bu

*bo

Lush. pukL swell, sprout' 'swollen' WB phoB 'swollen' WB phuB 'to swell, bud' WB pu 'bulge in middle'

WT 'bo-ba 'to

2.5.1

*bur or *bru

*bun

WT 'bur-ba 'prominent,

fen J:Jt *f:>Qns 'swell' Lushai tiL-puunH 'increase' (water) AA-Khmer biina I puun/'to amass, accumulate, heap'

bud, unfold' Chep. pyur- 'bulge, swell' ? WB phrui B'fat, swell' AA-Khm bura /pur/ 'swell up' WT 'bur-ba 'rise, be

Lushai po?L 'protrude' WB phuc 'protuberance'

prominent, bud, unfold' WB pruB 'protuberate' fen Jf *lr.m 'wellprominent, bud, unfold' set' (fruit)

WT 'bo-ba 'to

WT 'bu-ba, 'bus

swell, sprout' WB :l-phu 8 'bud, swell...'

'open' (flower) JPpu55 'to bloom, bud' 3~ pu 33 'flower'

JP bo 33 'head'

WT dbu' 'head'

fen :11t *b;)n 'big (head), horned'

WT spo 'summit' WT 'bog 'small

fu ~- *bu? 'big, hill'

fen :J:l *f:>Qn 'mound, big'

WT 'bur-ba 'rise, be

hillock' pou ~ *b() 'assemble, all' WB po 8 'plentiful, numerous' pei .ey.

*ph~

'large' fu $ *bu? 'big'

? WB phruiB 'fat, swell. up'

fen :J:l *b;)n 'big, mound'

connections that would not be obvious if one attempted to study a putative ST root in , or study the lex:icon of only a single language. For illustration, let us consider the of stems 'Swell' in some detail (Tables 2-6, 2-7; see also-+ chu 1 for additional "''"'f•ti.nn and considerations). We will not complicate the picture with too many AA data, Khmer /book! 'swelling mass', pora /baaor/ 'swell up, rise, bulge', /pur/ 'to swell up, , /-pull 'to swell', /-puul! 'rnass, heap, pile', /-puuk! 'mass, mound, group', etc. Representative data about 'Swell' in the two tables list distinct ST stems or roots in the >"vJuu•uo, while the rows represent similar meanings. The latter are randomly distributed over Jar-looking roots and stems (labial stop initial, vowel o or u, without or with final nasal or ~-r): bo, bu, bur, bun ('Swell' Table 2-6), poiJ, puiJ, pom, pum ('Swell' Table 2-7). (This list is ,~ot exhaustive.) Each of the eight sterns (four in each table) means, in at least one language, >'to swell, swollen', or a transparent semantic derivative of 'swell' ('bubble', 'thigh'). The stem I meaning distribution shows, for ex:ample, that JP bum31 'hill', JP bum55 'swell', CH peng *~m 'luxuriant' belong to the same stem and are genetically related in spite their rather different meanings, whereas JP phum 31 'lie down • is outside the semantic range unrelated. Conversely, CH fiJ $ 'big hill' probably is not cognate JP bum 'hill', etc. the table in another column. Looking at this distribution from another angle: for 'hill' used the sterns *pol) and *bu, JP *pum; 'protrude': JP *poiJ, Lushai and WB *po, WT and *bur- *pru; 'assemble, amass': JP *poiJ, Lushai *puiJ, WB *pum, OCM *bo.

ill

Jt

21

2.6 - 2.6.1

MORPHOLOGY

MORPHOLOGY

2-7

*pOI)

*pUIJ

*pom

*pum

swell

Lush. p:)::!I]R 'swollen' WT 'p"oiJs 'buttocks' WB p 11 0IJ 'bubble'

NNaga *pu:IJ 'swelling > breast, flower' Lushai pui)H 'increase, assemble' AA-Khmer puna /puiJ/ 'bulge, swell'

Lushai puamH 'to swell'

JP bum 55 'swell'

bloom, bud

WT 'bo!]s-ba

'roundness'

NNaga *pu:IJ 'swelling > breast, flower'

luxuriant, abundant

feng .$ *p 11 olJ 'densely growing' beng ~ *p61J? 'luxuriant'

feng ~ *p"uiJ 'abundant' WB prui!] 8 'full, abundant'

big 11 numerous

feng ~m *boiJ 'great'

heap, amass, assemble

JP phoiJ31 'amass' WT bOI] 'volume,

protrude 11 fat, stout hill; mountain

2.6

bulk'

*ban 'luxuriant'

peng 1t *oom 'luxuriant'

JP phu!]33 'crowd, herd' WT p11 ul)-po 'heap' Lushai VUIJH 'heap, mound' Lushai pu!JH 'increase, assemble'

xiang (sjal]) *snal] 'belt, sash, horse's belly-band' OKhmer dniiy ldnaaj/ 'retainer, king's servant' (->Tai thanaaj 'attorney, representative') -> OC dai. (t~iC) *t~h 'to carry on the head, bear, support'

*d-m-g

Khmer with agentive m-infix: dhmaya /tmfigjl 'agent, representative'> 'marriage go-between' ->mei fr~ (mu~i) *mg 'marriage go-between, match-maker' frJ!t, (divine match-maker:) 'god of fecundity' oc reinterpreted the *m- as the root initial, as did Khmer in the next word:

ren ~ (flzjgm 8 ) *ngm? 'year' MC 1- (Sagart 1999: 111-120 'infix'), or *1- > MC ji- or d-, e.g., M and. zhang-Jang !1$~ 'cockroach', kan-tan:l:)(~[k~mB-d()mB] *kMm?-!Qm?? 'pit', hfm-dlm [g~m 8 -d()m 8 ] *g~m?-l~m? 'kind of Water lily'; (d) partial reduplication, involving only the initial (shuangsheng ~§.alliteration), e.g., cencrft.&_ [t~h()m-t~ha] *tshr~m-tshrai 'uneven, irregular'; (e) partial reduplication involving only the vowels as in a s:;himing ex.pression; the alternations are typically between *e I *~or *i I *u, e.g., zhlinzhuiin ill$$ *tren?-tron? 'toss and turn', or xishuai !IL~!R$ *srit-srut 'cricket'.

i?&N

Type (c) is also common in Tai; Li Fang Kuei (1977: 93) quotes Siamese ma-JajA2'damage, destroy'; because this word is probably borrowed from a Chinese form (note ~ huf6 11!.!}1 *hmai 'destroy') the /11 is here an 'infix'. Historically, such forms probably originated as a type of reduplication where the first syl1able is reduced to a vowel, and the rime is repeated with an initial /-. Siamese words like m eel]- ma-leeqA2 'insect', and met- Jet- ma-JetD2S 'grain' are ambiguous because the first syllable is also an animal I plant prefix. Aslian (AA) 24

2.8- 2.8.2

also have an 1-infix for lower creatures. This shows that such forms do not always from earlier consonant clusters by 'dimidiation', but reflect bisyllabic forms. gart (1999: 118ft) has concluded that in some Min and Jln dialects, bisyllabic reduplicaorms with the meanings intensive I durative I distributive are formed with the second starting with Ill, similar to the Tai forms, as in Fuzhou tsing55 'to stare' > tsP1-Jing55 fixedly'. This dialectal /11 may be the continuation of an earlier OC *r-infix (div. II) ibid.); note the OC word for 'to stare': cheng ~ MCthel) < *thraiJ. This implies that in words, the OC cluster with *r > div. II either was, or is descended from, a bisyllabic

Non-morphological word formation Re-analysis

. languages (commonly "dialects") sometimes have word forms that are not the m..,, ..,.,.v•4• analogues of standard Mandarin or MC forms. These near homophones are not result of phonological change, but of re-etymologizing, either occasioned by taboo derations or by folk etymology. For example: Cantonesejiti2.th11u45j3I

~m •sun' is re-etym~logized as 'the hot one' 3I ~'earth'= MC duoA2 'plaster, mud' is re-etymologizing of the analogue of Mand. tu± =MC thuoBl 'earth';

gushl i'i')($ 'story' has in southern dialects been re-analyzed as the Mand. equivalent of gushi i;$ 'old affair', e.g., M-FuzhOu ku31f44.J~y242; qian ~H- 'lead' n. < kh-for MCjiwiin has been re-etymologized in most Mand. and Wu dialects as connected with 'hard, strong' qian ~=MC khifn; qiiinbl ~H-* 'pencil' is called in some dialects, such as Wuhan and JUtn'ou, the Mand.

equivalent of yangbl 1$ i: lit. •western I foreign pen', which may perhaps be an apt re-etymologizing of the original MCjiwiin 'lead' (as in K-Meixian ian11-pitll);

jingfl. MC kjBIJC'mirror' (CV C). A case in point is PTB *b-I]a? 'five' (variant of *1-I]a) > Lushai pal)a > poiJ.

Five

Here the intermediate sesquisyllabic stage of a form is attested. Such a form is not found in the few other cases; they all involve a putative ST *r or *m pre-initial: To steal

PTB *r-ku, WT rku-ba 'to steal', WB khui 8 vs. Lushai ruk(or variant of rak ?)

Sell

PTB *r-I)a 'to sell' vs. WB roiJB 'to sell'

Control

ST *m-IJa 'to drive, control', WT mlJa '-ba 'might, dominion', yu flED *I)ah 'drive a chariot, direct' vs. WB mo1J 8 'threaten, drive away'~~ :)-moiJB'driving'

Smell

ST

*m-n~m

?, PTB *m-nam 'to smell', WT mnam-pa 'smell of' vs. wen

llfJ *m~n 'to smell'; in this stem the final dental nasal -n also could have been the result of labial dissimilation. Some of the words with medial OC *r may also be the result of this process (§6.1; §7 .7 .3).

2.8.4

Convergence

Occasionally two etyma have coalesced in OC into one word or word family. This can even involve items from different language families (CH has attached the final *-k): 26

MORPHOLOGY

2.9 - 2.10

se§ (tJ;:~k) [~tk, s :pk] *sr~k (< s;:~r-k) ( 1) 'col or, col or of face, appearance, countenance' < ST: Lushai saarH < saar 'prismatic colors' ~f saarR I sarhR 'healthy looking, rosy, ruddy' (2) 'good looks I charms of women, sexual pleasures' 'fear'; 'keep in the mouth'> 'dissatisfied, resentment'; 'rise'> ; 'cold'> 'grieved'; 'thick'> 'generous'; 'white'> 'understand'; 'black'> 'evil'; 'get the of'> 'be able'; 'carry, bear'> 'endure'; 'heart'> 'mind'; 'gall'> 'courage'; 'sun'> tday'. Metaphors become new words. Chinese shares many metaphors with other languages, luding English. Occasionally, the semantic leap of a suspected extension is difficult to follow; note the ST 'hear'> 'ask'. A few verbs with the meanings 'to think, to say' or other abstractions are y semantic extensions of verbs 'to be, to do, to act, to go'. Something similar is ved in many languages, such as Engl. the saying went like this: "... ", or he went on and In TB languages: WT byed-pa, byas 'to make, fabricate, do' > zes byas-pa 'thus said, so 27

2.10

MORPHOLOGY

called' (i.e., marks direct discourse); WT mclri-ba 'to come, go'> 'to say' (marks direct discourse); Lushai tiL I ti?L < tii? I ti? 'to do, perform, act, work; act towards; say; to think, consider, feel, wish'; Mandarin Chinese~ wei3 ~ 'do, act' in the phrase ylwei f)~ 'to consider ... , to think' (lit. 'take something to be, take something for'). These considerations make an etymological link between the following pairs likely: fei ~F is not shl ~ is, this you 't; right side zuo left side mie will not I cannot fu ~ cannot, will not sf j§'j to act, manage wei ~'lt to be hul -m_ should be

ti. Ml

>be wrong >be right >to honor, appreciate >to disapprove >despise >resist >to think,~, >to think 'lt >be kind

,..

3

MIDDLE CHINESE TONES AND THEIR OLD CHINESE EQUIVALENTS

.IfZ!I

fti

Formally, wei ~ *wa(t)s 'to tell, call' seems to fit the "irrealis" pattern (§6.2.2), looking like a derivation from you l j *wJ? 'there is, have', but it is not clear if there is an etymological connection and how the semantic leap came about.

A.

' and LHan had three tones: tone A (pingsheng ='even' or 'level' tone), tone B tnq~mer.rgL!f ='rising' tone), and tone C (qiisheng -*~='falling' or 'departing' tone), = according to traditional Chinese phonological analysis, toneD (riisheng ' tone) for words which end in a stop consonant (p, t, k), i.e., this short-stopped type was toneless. These tonal categories are projected back to OC where tone A is to have been an open syllable or one ending in a nasal, tone B marked a syllable with a stop in the final (or a glottolized syllable}, and tone C a syllable with final *-sI *-h. A and_D are usually left unmarked as this causes no ambiguity. OC probably had no "in the later sense but instead segmental phonemes. Nevertheless, we will here apply the 'tone' also to OC in the sense of "later tonal category" for the sake of clarity and to arguments about their OC phonetic nature. Because MC tones are projected back to ones in LHan, subsequently examples will often be cited in simpler LHan forms. ·All three tones can belong either to a root or stem, or play some morphological role. The common morphological tones are Eshiingsheng and C qilsheng, which together with voicing form a derivational system which marks direction and diathesis (§4). The between allofams in the three different tones is exemplified by the following parawhere the form in tone A is the simplex, the derivation in tone B is endoactive (§4.5), . the derivation in tone C is exopassive (§4.4) (LHan after the graph):

zhr ~ tsa, *t;) zhl

lf:

tsaB, *t;)? tsac, *t;)h

'to go, proceed' 'foot' ( yu ~ [waA] 'sacrifice With prayer for rain'

'yii is a widely attested ST word; it can be set up as ST *wa? (with final glottal stop) on strength of Kuki-Chin and Chepang forms in addition to MC. This rules out the possibility 'rain' is a tone B derivation from 'sacrifice for rain', which would also be semantically ·~:l:'''"'uo""'"· However, elsewhere tone B can be a morpheme which creates or marks endoac. verbs or words (§4.5); 'to rain' fits this pattern, so that tone B may have been felt to be this rather than part of the root. 'Sacrifice for rain' was then created as a back ation by removal of the alleged suffix tone B. 28

29

3.2 - 3.2.1

TONES

TONES

Additional tone A derivations include the following items (LH forms after the character): Slave

xi ~ > xf ~

gee ge

'to be bound' 'slave, captive'

Wine-master

jiu >qiu

rim

tsiuB dziu

'wine' 'wine-master'

Writing slip

bian ~ >pian ~

pianB phi an

'narrow' 'writing slip'

Tally

fu Pit > fu t--1

buoC buo

'to adjoin' 'a tally'

bing ill[ > pian ,!§31:

bei]B ben

'side by side' 'two (horses) side by side'

siB si

'to display' 'corpse, persona tor of a dead'

Side by side

rm

shr fJ;( > shf

Corpse

F

-

he ~ [gek] 'kernel fruit' - hai ~ [gd] 'bones, skeleton'.

He is a ST etymon: WT rag 'fruit stone, bead', Mikir rak 'fruit stone'; therefore he was the original form from which hai was derived.

Middle Chinese tone B (shangsheng 1:~): phonology 3.2 MC tone B probably derives from an OC *-?.In some modern dialects tone Bends with glottal constriction (Branner 2000: 119) -note M1n-Songyang puf1? ('measure for books' ben :zt;:; Branner 2000: 344). Glottal stop after nasal codas is also shown by variants like xliJt [sei 8 ] [senB] 'to wash', as well as Shi]ihg rimes such as *-an? I *-a? (Shi 30 1), *-uiJ? I *-u? (Shi 264,7); some rimes confirm that the phoneme in question was a stop consonant: *-ap I *-am? (Shi 265, 3) and *-et I *-en? (Shi 265,5). The glottal feature can, however, appear elsewhere, e.g., in the middle of a syllable (Sagart 1999: 132, n. 1: Xiaoyi dialect in Shanxi). Tone B seems to be a weakened variant of final-kin some words (§3.2.2).

3.2.1 Tone Bfrom Sino-Tibetan *-? Tone B can be part of the root. In some words it goes back to the ST level because some OC open syllable words with tone B correspond to Kuki-Chin and Chepang words, which are also reconstructible with a final glottal stop (Chepang still has final -?).The first several items in the list below are taken from Ostapirat (LTBA 21:1, 1998: 238±) with WB and LHan forms added. The agreement in final*-? is particularly persuasive because the first seven items have been selected without Chinese in mind. Tiddim and Lushai tones sometimes split according to vowel length or timbre (hence Lushai tones F(alling), R(ising), L(ow) < *-1'); the corresponding Tiddim Chin tone is 1; LHan forms follow the graph: Gloss Bird Child Water

Tiddim va: 1 ta:l tu:i 1

Lushai vaF faF < faa? tuiR

Chepang wa? eo? ti? 30

me1·R me1·R k"aL I k 11 akF rua?L tu?L thiR fakF kuaR < kua? zaR I zatL

meil mei 1 xa: 1 gua? 4 tu? 4 si: 1

me? hme? kha? wa? w~i?

wei huo ku

ffi

mui 8 k huaiB =5 khaB waB yu shu t~j dzo 8 shf f,t 'to eat' TB *m-lyak 'to lick'

PCH *s-l;Jk-s > sl f,t 'to feed' 'make I let lick, feed' TB *s-I yak

4-2

simplex - introvert

s-suffix -extrovert-causative

Limbu

thUI)-

t 11 Ul)S-

OC /PCH

*?;1m? > yln j'lj; 'to drink'

*?;;~m -s

OC/PCH

*m-l;1k > shf it *m-l;;~k 'to eat'

*s-!;Jk-s > sl

'to drink'

'give to drink'

> yln J1l( 'give to drink'

it

'give to eat, feed'

Historically, three phenomena have partially or completely coalesced in OC: ( 1) semantic introvert I extrovert; (2) grammaticaltransitivity distinctions; ( 3) grammatical active I passive distinctions. The result of these partial mergers is the intersecting pattern of endoactive I exoactive I endopassive I exopassive derivational categories (Table 4-3 ). These terms apply to morphologically marked derivational categories of new words; in the OC language, they are independent of grammatical I syntactic behavior and word class; see §2.1.3 (Beard 1998: 44ft). This system constitutes the 'youngest' morphological layer which was still productive or at least transparent in OC. Introvert and extrovert (endo- I exo-) are semantic categories that are readily apparent from • a word's meaning. In introvert words, the action is directed toward the subject, or happens to or within the subject (to buy, to watch, to grow); in extrovert words the action originates in or with the subject and is directed out and away to a necessarily external object (to sell, to show). Introvert verbs often tend to be intransitive, extrovert verbs transitive or causative. Transitive words like 'to buy' were originally perhaps purely introvert, while 'to sell' was purely extrovert. 'To grow' was intransitive or 'middle voice', whereas 'to sow' (derived from 'seeds' n.) was grammatically transitive I causative. The extrovert and valence increasing categories overlap to a large extent: 'to sow' is extrovert beside being transitive ('sow seeds'); and 'to seW is ditransitive beside being extrovert; zhfmg ~'to grow' is introvert beside being intransitive. 38

4.1

two concepts overlap completely in words that imply 'giving', as in 'to drink' (introvert, tive) >'give someone something to drink' (extrovert, ditransitive, i.e., valence increa, see Table 4-2. Thus direction and transitivity have merged in OC. Furthermore, the active I passive distinction has been added to the mix. Active and passive with familiar lE (incl. English, Latin) grammatical concepts. The agent of an active is its subject: 'he is growing', 'he is buying something'. In a sentence with a passive verb the object of the active counterpart becomes the subject of the passive; the agent (tr. is placed in a prepositional phrase ('he heard this' >passive 'this was heard by him'). The terms 'endoactive' (Japanesejidoshi ~!WJ~Il), also 'intransitive') and 'exoactive' @;t\fJ~"l. also 'transitive') describe pairs of introvert I extrovert verbs in Japanese mar (Lewin 1990: 118f.; Shibatani 1990: 115; also Takashima 1996), for example: naru 'to become' intr. yurumu 'to become loose' intr.

nasu 'to form' tr. yurumeru 'to loosen' tr.

gh nearly identical with the familiar 'intransitive' and 'transitive', the Japanese en dove category includes verbs that can take an object, therefore the term 'intransitive' is not e.g., kado-wo deru 'walk out the door' ('door' is the obj. in Japanese). The terms already been used for Chinese (Sag art 1999: 133) where endoactive (introvert) verbs also be tr. like mai 'to buy (something)'. Exoactive verbs are the extrovert, transitive I causacounterparts (mfd 'sell something to someone'). '-'ll'uu,, .. ,.ve' is similar to the 'middle voice' of some IE languages: the action was conas operating in or on the subject; in the exoactive, the verbal action was directed sfrom the subject (L. R. Palm er, The Latin Language, 1987: 262 ). H. H. Hock (1986: provides the example 'to grow' intr. for one of the specific uses of the lE middle voice; CH zhling ~ 'to grow' as illustration for endoactive derivation ( §4.1.1). In accordance with this system, we add the terms 'endopassive' and 'exopassive'. In exove derivations, the agent (i.e., an external, 'exo-' element) is still available as in a typical ve sentence: 'it was heard by him'< 'he heard it'. Endopassive fits the definition of ive verbs:" ... refer to a resultant state without any indication of an agent" (Sadler and 1998: 223); the endoactive verb ching fi: 'be long' is a stative verb ('be in a state of stretched< be stretched'). We could have used the familiar labels 'middle' for 'endoactive', 'active' or 'causative' for 've', 'passive' for 'exopassive'. However, the IE middle voice drifts toward the passive, the CH category remains active. Furthermore, IE languages make no clearly marked t- extrovert distinction. In the end, the labels endo- I exo- and active I passive describe well the nature of the CH categories. They also allow us to reserve the endo- I exonology for word derivation, and continue to use 'active' and 'passive' for grammatical on ships. These categories apply to derivations from a stem or other word, i.e., words fall into one of four_categories only when compared with their simplex. Many words have MC tone B, like ~ (tjaiJ 8 ) 'to grow' where the tone marks the word as endoactive because it contrasts and is derived from, the simplex zhang 5:R (tjal)) 'to stretch'. But huo !)( (xuaB) 'fire' is 'ved; tone B marks nothing. Although underived words do not belong to such a category, occasionally feel that a word inherently does, as sl7"E 'to die' with its typical endomeaning (see §4.5). Or wen~ 'to hear (something)' tr. could be seen as inherently

39

4.1.1

DIRECTION and DIATHESIS

DIRECTION and DIATHESIS

introvert. But these words are morphologically unmarked and their categorization would in many cases be arbitrary and add nothing to our understanding.

4.1.1

Direction and diathesis in Old Chinese

The OC morphological changes which mark direction and diathesis have left traces in later MC tones and voiciNg of the initial consonant of a word; the tones, possibly also voicing, go back to earlier segmental affixes in OC or PCH. The morphemes are ( 1) Tone C ( 1) (MC qusheng *~'ask someone about something' ditr.) (§4.3); (2) Tone C (2) (MC qusheng *~ mai

meB 'to buy' meC 'to sell' < 'to give someone something to buy'

Give

shou 5t > shou 1§t

dzu 8 'receive, accept' ctz.uc 'to give' ditr.

Show

shi fl > shi ;J\

giB 'to look at' giC 'to show' ditr. < 'make someone look at'

Teach

xue* > xiEW

g:1k, *gruk 'to learn' gauc, *grukh 'to teach'

Ask

wen~ > wen F~=~~

mun 'to hear about' tr. mu ne !to ask about something', 'ask someone' tr.

Wife

qf~ >ql ~

tshei 'consort, wife' ts 11eiC 'give a wife to', 'give as a wife' tr.

'}
tlng ~~

theiJ 'to listen' theiJC 'to listen to, obey' 43

4.3 .2 Reply Tell Overturn

DIRECTION and DIATHESIS

da~ >dui ~M

tgp 'respond to' tu;:Js, *t:}ps < *tups 'to reply' (to a person)

yu ~g >yii ~g

IJta 8 'to speak' IJtac 'to tell someone'

fu

>

-11 fu 1'1

Aid

you ki >you ti{ti

Before

xian >xian

phuk, *phuk 'to overturn, overthrow' tr. phuC, *phukh 'to overspread, provide cover for someone' wuB, wuC 'right (side), be to the right' wuc tr. 'to aid, support, assist someone'

7'G 7'G

sen 'to go in front, ahead, before, former' senC ( 1) tr. 'to walk before' (in order to protect)> 'take care of someone, attend'; (2) caus. 'to put first'

1Jt 1Jt

DIRECTION and DIATHESIS

4.3.3 - 4.4

Residue remains the inevitable residue where the exact function of tone C is not obvious: perhaps follows the pattern of 'feed' (§2.3.2). xiaT

> xia T

gaB 'to descend, down, below' gaC tr. 'descend, fall'

Tone C (2): exopassive derivation tone C [2] < *-s/-h) is the passive of a transitive or exoactive word. As in incl. endopassive, the subject is the object of the exoactive I transitive counterpart; in to introvert endopassive, exopassive is extrovert in the sense that there is an explicit or

,...,,,Q.,iVP. (MC

outside agent which acts on the subject. Usually, passive agrees in meaning with languages, although in OC we are dealing with word derivation, not grammar (ex. 2); ssive derivation is often a noun, literally meaning "the thing that has been verb-ed" ·opposed to endoactive where the noun is "the thing that is doing the verb-ing"). Exopassive counterpart to exoactive. Contrast the simplex ex. 1 with the exopassive ex. 2; the is still in the extrovert column because the former (outside) subject is still available in ...'..,,,..,,"'~itional yu phrase as agent, i.e., there is still 'the outside' which hears the sound.

After

hou >hou

goB or yo 8 'be behind, after', 'to follow' goC QT yoc ( 1) tr. 'be behind someone, attend, support'; (2) caus. 'to put after'

Attend

shi '!~ > shi f~

dz;}B 'to depend on' dz;}C 'to wait upon someone, attend'

nyu ft >nyu ft

l)a 8 'woman, wife' n. l)Gc 'give (as) a wife' tr.

. (2) sheng wen yti wai ~1*1 'f Ji- (sound I be heard [passive] I preposition I outside) '(The

khf.oB 'to take' tr. khf.oC 'to take a wife'

sentence, the passive verb wen looks like a grammatical form as if OC still had tion. But an inflectional suffix *-s should change other verbs in a predictable way as which is not the case in OC. Furthermore, in OC grammar, passives are regularly created rn active verbs by word order alone (§2.1.2, ex. 6). However, example (2) does suggest an Iier inflectional origin of the morpheme. More passives:

Woman

Take a wife qu Jf:SZ

>qu~

In this last example, we might have expected that tone C makes the derivation qu extrovert, whereas it is still introvert ('take', not 'give'). However, the key was apparently the person as indir. object. After all, the meanings of derivations are by their very nature unpredictable.

4.3.2

mstruments ')sound is heard outside' [Shi 229, 5].

Tone C: exoactive- transitive, causative I putative

When applied to nouns or stative verbs, tone C effects an increase in transitivity (valence), while extroversion is more coincidental (LHan after the graph): Sow

zhong ~~ > zhong 1]!

tsoiJ 8 'seeds, cereals' n. tsoiJC 'to sow' tr.

Hit center

zhong cp >zhong i=f:l

tuiJ 'be in the middle, middle, inside' tulJC 'hit the center, hit, attain'

Repair

shan > sh~m

dzan 8 'be good (at)' tr. (intr.) dzanC caus. (make good:) 'repair, put in order' tr.

Love

hao iff >hao fjf

hou 8 'be good' sv.

Hate

.:g #,g

'M e;t.::J'

Grasped

wen lifj >wen~

mun 'to hear about' (simplex) munC 'be heard about, renowned, fame'

zhf ¥t\

>zhi~

tsip, *t;:Jp 'to hold, grasp' tsic, *t;:Jps 'be grasped, seized'

jln 3Wi

dzin 8 'exhaust' > jln

xiaT >xia T

gaB 'descend, below' [Shijing] gaC 'be put down'

71! dzinC 'be exhausted, destroyed'

Back n.

bei ::1 G >bei ~

pQk, *p:}k 'north' pQc, *p~kh 'what is turned north (passive): the back'

Measured

duo J}t >du )jt

dok , *dak 'to measure, plan' doe, *dakh 'be measured'> 'to regulate'

houC put. (consider I treat someone as good:) 'to love' tr.

>wu ML':J\

?ok, *?ak 'be evil' sv. ?oc, *?akh put. (treat as bad:) 'to hate'

yuan ~ >yuan ~

wonB 'far away, distant' sv. wane put. (treat as far away:) 'keep at a distance'

I

Far

( 1) WO wen qf sheng 3Xrnl~~ (I I hear I his I voice) 'I hear his voice' [Shi "199, 3].

, such an exopassive derivation can occasionally be a noun (see 'cloth' and 'carriage' §4.1.1 ), for exam pie: na k:P'J *ngp < *nup 'to bring into, > nei *ngts < *nfips 'what is entered >inside'

pg

zhf 44

L

*tQ 'to go' > zhi

$

*t~h 'what is gone to> goal, purpose'

45

4.4.1 - 4.5

4;4.1

DIRECTION and DIATHESIS

Exopassive as a transitive verb

An exopassive can become a new active verb, for example:

DIRECTION and DIATHESIS

verb zh.iing (intr.) can, of course, be used causatively ( §2.1.2 ex. 15) and as an (§2.1.2 ex. 14). zhang ~ > zhang

?ui 'to overawe, terrorize', i.e., 'frighten someone' tr. ?uiC 'to fear' tr., 'be afraid' intr. yang {qJ

1)01)

'be high'

l)tol)B

(look up'

jOI) jol)B

'to lift, raise' 'nourish, raise, support'

-m

yang >? yang

::1

\01)

the exoactive word has an introvert counterpart which in many pairs has tone B. This could be a CH innovation because of the inherent introvert meaning of the word, and in to create morphologically marked introvert I extrovert pairs (LHan after the graph):

That the passive form of a verb can flip over into a new transitive verb could therefore be an old ST phenomenon; it also may speak in favor of a single origin of the two functions of C. Other examples (LHan after the graph): Yield

t;ol)B

;R

.m

maJesty' [Shi 272]. The tr. verb wei m\(; 'to scare' turns passive with tone C wei ¥!:"I am scared [of Heaven]"; wei is actually a new transitive verb 'to fear', thus "I fear Heaven". What remains of the passive is the switch in the direction of the action, i.e., the new subject is still the object of simplex 'to scare'. This is similar to what has been observed in Kuki-Chin languages. In Tiddim and Lushai the second verb fof'tn can itself become a new transitive verb (Geilich 1 170). An example from Lushai:

4.5

maiil > mai Jf

me 8 'to buy' meC 'to seW < 'to give someone something to buy'

shou ~ > shou t~

dzuB 'receive, accept' dzuC 'to give'< 'to give someone something'

shi:fj > shi 7}\

giB giC

shan~

>shan .*ff

dzanB, *dan? 'be good (at)' tr. (intr.) dzanc, *dans caus. (make good) 'repair, put in order' tr.

haoPf >haotif

hou 8 'be good • sv. houc put. (consider I treat someone as good) 'to love' tr.

'to look at' 'to show' < 'make someone look at something'

note 'to drink' in §4.3. B may have been added to some intr. verbs because of their implicit endoactive nature it came to mark change of state:

,.... . . . . . .,J . . .

4.5 Tone B (shangsheng J: ~): endoactive derivation Endoactive (MC tone B < *-?)is active, i.e., the agent is the subject as in an exoactive word; the derivation is also introvert, i.e., the action takes place in the subject or is directed toward the subject. Sometimes, valence is decreased, but introvert forms can also be tr. verbs (endoactive 'to buy something' tr.). Again, it seems that two earlier phenomena have coalesced, valence decreasing, one introvert; both are the opposite of the exoactive tone C (§4.3). An introvert meaning is obvious in 'to buy I receive something'. In zhang 'to grow' intr. (ex. 2), the action takes place in the subj., no outside agent acts on it; this contrasts with the tr. simplex zhiing 'to stretch something' (ex. 1) where the direction from the subj. outward the obj.: ( 1) zhiing wo gong ~flt"3 (stretch I our I bows) 'we drew our bows' [Shi 180, 4 ]. (2) sheng er zhang, zhang er dii ~mi~~ mi* (give birth I particle I to grow I to grow I part./ '(the living things) ... are born and grow, they grow and become big .. .' [Ltishi chunqiu 3, 5].

die ('Change , Rise Sit ·Dwell

si 7£ gai ;:!)( ql /EB zuo ~ chul$&

siB kaB khf;)B dzuoi 8 tshaB

'to die' PTB *si (not *si?) 'to die' 'to change' 'to rise' 'to sit' 'to dwell' PI'B *gla

words remain in which the role of tone B is not clear, for example: 'Expand

shf

:tit!i

> shl 5-JH

soi 'to expand, spread out, expose' soi 8 'to extend, spread'

zheng ~ > zheng 71
dvang ~ ~

tal) 'be equal of, rank' tal) 8 'category, party' ('equal in rank')

Brother

d'i ~

deiC 'order, sequel' deiB 'younger brother'

Offspring

-T

>d'i~

¥Z Low ground bei !If. Dry

kan 'dry'> hi:m

gan

:!:$. ¥}

> bl A turn

z~li

> zai In addition, see ziJ

:f£1.

....

iJ
chang ~

tal) 'make long, stretch' bie 53U

ptat 'to divide, separate' btat 'be different' sv.

jia ~ > xhi jJ(

kep 'be on both sides, press between' gep 'be narrow' sv.

jie ~~ >xie $

ke 8 'to loosen' ge 8 'be slack, idle, careless' sv.

sv. is a well-defined word class whose members can freely function as adjectives. Many verbs are, however, intr. and even tr., but they still agree more or less with the definition of a 'stative verb'. jian 5i > xian fYi\

kenC 'to see, visit' gene 'to appear' intr. (get oneself seen)

jiang ~ > xiang Pit:

bl)c 'to descend' g:Jl) 'to submit' intr. (get oneself down)

11 >qian rlf

tsiam 'extinguish' (=put water on a fire) dziam 'go into water, wade' (get oneself into w.)

jiao 3(: >yiao ~

kau 'to cross, mix with, exchange' gau 'mixed, confused' (get mixed)

Jlao ""

tsia u 'to roast, burn' dziau 'firewood, to burn' intr. (get to burn itself)

Go into water jian

._ m .,

> q~au

t~

'1~ .. ,

49

4.6.1

DIRECTION and DIATHESIS

Attach

zhu ]i! >shU.

tsok 'to attach, apply' dzok 'be connected, attached to' (get to be attached)

Apply

zhu6~ >zhu6 ~

tok 'to place, put, apply' 4ok 'to come in contact with' (be placed)

Set

zhi :][ > zhf 11.[

t~c 'to set, place' 4~c 'hold upright, a pole' (got set and held)

Half

ban >pan

-*

Ill¥

pane 'half' banC 'separate from, rebel'

5

INITIAL CONSONANTS

Because MC initial consonants and other phonemes are projected back to (nearly) identical ones in LHan, subsequently exam pies will often be cited in simpler LHan forms. Most of the OC- TB or foreign initial consonant equations are straightforward: m-= m-, p -ph- b = p- ""ph-"" b-, etc. For unusual correspondences with laterals and semivowels, see §7§ I 0. Voiced initials are discussed in other co~texts:

The endopassive nature of the following two examples is not obvious. Yet 'to imitate, follow an example' and 'to learn' are introvert processes and can be felt to be passive. The choice of endopassive voicing may be due to the inappropriate extrovert flavor of the exopassive. Set example jiao ~ > xiao ~:5({5&

kauc 'teach, set an example' gauc 'ij;nitate, follow (an example)'

Get insight jue Jl: >xue ~

bk 'to wake up' intr., 'get insight' g:Jk 'to learn'

Two derivations behave exopassively because the subject of the simplex is presumably available: Destroy

b!ti ~)1: > b!li J&

pas 'to ruin, destroy' bas 'to go to ruin, be defeated'

Ruin

huai f~ >huai ~

kueiC 'to destroy, ruin' gueiC 'to be ruined'

4.6.1 Residue This residue may constitute a subclass of verbs that are derived from nouns (Baxter and Sagart 1998:47 with additional examples). The meaning of some items implies introversion (LHan after the graph): b" Back p~c ""' 'the back' eJ ;:1~ F.l >bei ~

roC 'turn the back on' {introvert)

Morning

zhao M > cMo Jj!}j ltlJ

tau 'morning' 4au 'go to court, morning tide' (introvert)

Shore

bfn ~ >pfn t!iJi

pin 'shore, bank' bin 'river bank, be on the edge, close to'

Dike

dr ~:lf! >tr

te 'bank, dike' de 'to raise, to take up' (tr.)

Pole

zhf, de >zhf

tstk 'a pole' dztk 'to plant, raise, establish' (tr.)

Finger

zhlt~ ~~ shl iJf.

ki 8 'to point to, finger' giB I giC 'to see, look' I 'to show'

m

m m

50

morphological voicing of the initial (endopassive), see §4.6; MC initial d- from OC *d- and *1-, see §8.1.1; MC initial zj- from OC *s-1- (§8.1.2), *s-j- (§9.4), and *s-w- (§ 10.1.2); MC initial dzj- (LHan z-), the rare initials, from OC *m-1-, see §8.1.3; MC initial ij- (LHan di-), the common initial from OC *d-, but also from *j-, see §9.3; and on the initials in GSR 41 3 ~. see §8 .1. 5.

5.1

Devoiced initials

Some voiceless initials are discussed in connection with *r- (§7.1.2; 7.1.4) and OC *1- (§8.2.1). Devoicing of sonorants is common within ST wfs and in OC. Many of the following MC initials reflect an earlier voiceless sonorant; examples can be found in §5.2.2 to §5.6. MC th-from *lh-, *nh-, rarely from *rh-. MC sj- from *lh, *nh-, *hj-(?); MC ij-never represents OC *sj-! MC x- from *hi], *hiJ, *hw, also *h-; rarely from acute initials (§ 1.3.1; §5.6). MC !h- occasionally from *rh-, otherwise from *thr- (§7.1.2). MC tshj- occasionally from *k-hl- (§5.9.3), otherwise from *th-. Other languages indicate that the loss of consonants like *s-, *k-, *p-, *?- were the cause of 0C voicelessness, by first devoicing the following sonorant, then disappearing. Devoiced initials eo-occur in phonetic series with the voiced counterparts *1-, *-n, *-r, *m-, *IJ-, *w-. Because they do not mix in phonetic series with other types of consonants likes-, k-, t-, we must assume devoiced sonorants already for pre-literate Chinese, i.e., at least 1250 BC; thus 'fire' ~ huo 1 MC sMC zj- reflects an original s-prefix in the combinations *s-1- (§8.1.2), *s-j- (§9.4), and *s-w(§10.1.2). But the voiced ST root initial *z- is devoiced in OCto *s-, for example: 51

5.2 - 5.2.2

INITIAL CONSONANTS

Small

ST *zi: xi f.i!E (sieiC) [seC] *seh or [seiC] *si'h 'thin, small' PTB *ziy > West Tib. zi·'very small'; WB se 8 'small, fine'; Kachin zi 'small'

To act, do

ST *zg: sf I§J (st) [sig] *sg 'be in charge, manage' WT mdzad-pa ( < m-za-t) 'to do, act'~~ bzo 'work, labor'

Sleep

ST *k-zim: qln

(I) MC s- < OC *s-, and MC~- from *sr-, from a PCH s-prefix, it occurs mostly before the

initials *n-, *1)-, and *r-, occasionally also in front of others; see §5.2.1. (2) MC zj- before OC *1-, *j-, and *w-; see §8.1.2. (3) MC voiceless sonorants in OC; see §5.2.2-3; §5.6. ( 4) Perhaps MC dental affricates ts-, etc. from earlier s +stop consonant; see §5. 7. The s-prefix is of ST origin. It survives"in OC as a non-productive morpheme which once formed causatives (Mei 1985: 334-343; 1989; Baxter and Sagart 1998: 53) (§5.2.1 and §5.2.2), iteratives (§5.2.3), and common nouns (§5.2.4). See §2.4.3 for the history of this prefix.

5.2.1 Causatives-prefix > Middle Chinese sIn some MC forms which have preserved *s-, the original causative I transitive function is still detectable (LHan after the graph): sal)c, *smal)h 'to lose, destroy' < wang L [mUGIJ] *maiJ 'to lose'

Revive

sii ~

so, *siJa 'torevive' < wu

Fear

sii ~

so, *sl)a 'to tremble, fear' ~~ wu "5i£l [IJaC] *IJakh 'unexpectedly come across'

Tremble

suo*

sok, *siJak 'tremble, fear' < e ~~~~ [IJokJ *IJak ·scared, tremble, fear'

Entice

xu i/Jil

Expel

·XIang

.l.,P,-

~ ~

'Jt,g:jl: [IJaC]

*l)ah 'to awake'

sut, *sjut, ? 'to entice' is prob. a caus. of you~ 'to entice' (under~ shu 4 ~1fc1Mt) (?) siaiJ, *snaiJ 'to remove, expel' < rang [naiJ] *nal) 'to remove, steal'

tltet

5.2.2 Causative s-pre]zx > MC voiceless initial As already suggested, the voiceless initial consonant represents an early layer in which a ST or PCH causative *s- has been lost after devoicing the following sonorant: Neglect

huang 1ft hUG!), *hmal) 'neglect, reject' < wang t: [muaiJl *maiJ 'Jose'

Let go

shf~

sit, *lhit tr. 'to lose, fail, neglect, let go, err' < yi {X~ Uit] *lit intr. 'to escape, retire'

Transport

shU lMtr

so, *lho (or *hjo ?) 'to transport, transmit' < yu ~Jltr:hlli [jo] *lo or *jo? 'pass over' 52

5.2.3

Injure

shang 11

Let go

yi ~ [sak] ~ [jak] *lak 'unfold, unravel' > devoicing ea us.: shi f~ [sak] *lhak 'unloose, put away, let go' >s-suffix caus.: she 'i/J!J_ [saC] *lhakh 'let go, liberate, pardon' > s-prefix caus.: cli)f; [tshiak] *s-lhak? 'to expel'

*lhal) 'to injure, hurt', tr. or caus. of < yang ~ [jol)] *laiJ 'be sick' In at least one word, an s-causative has been added to the already devoiced initial (§5.9.2):

$! [tshimB] 'sleep' WT gzim-pa 'sleep' (§5.9.1)

5.2 Sino-Tibetan *s-prefix The STand PCH s-prefix is reflected in several MC initials:

Lose

INITIAL CONSONANTS SGI),

Some causative derivations with s-prefix have in addition tone C which also marks causative (exoactive; §4.3).

5.2.3 Iterative s-prejiX >MC s-, h voiceless initial PCH *s- forms iteratives and intensives. While 'intensive' is often a subjective perception, iterative can readily be identified when it refers to repeated action or movement. Some examples of iteratives follow; the first item is particularly illuminating: Flicker

shan /7C] LH [samB] *!ham? ('flicker') 'to twinkle, time of a short glance, moment'. 'Flicker' is a repetitive phenomenon. MC s-, voiceless initial A ST *s-I devoicing is also conspicuous in nouns, but the function I meaning of the *s- is not clear. We may speculate that nouns with iterative meaning, such as 'Fire', 'Comb', 'Brush', 'Sieve' (in §5.2.3) served as a paradigm. Dai and Wu (1995: 99) include examples from JP, e.g., sj33 'to die' > tf~-'3-sj33 'a dead person' (LHan after the graph): 54

5.3

Archery hall

xie~j

zia c, *s-lakh 'archery hall' (initial as in §8.1.2)

Beard

xu ~Ji~

sio, *sno 'beard' fYfB *sno(w)- *sno(t) 'mouth, vessel'

Frost

shuang ;ffi ~Ol], *sral) 'hoarfrost' < liang Y~ *raJ) 'cold'

Bean

shu ~

suk, *nhuk 'bean' PLB-M *(s-)nok 'bean' > WB nok

Fire

huo j(

huoiB, *hmgi? 'fire' PTB *mey, OTib. smye; PL *C-mi2 'fire'

Louse

shfEfu

~it, *srit < ST *srik 'louse' PTB *s-rik

Scorpion

chai

If

thas, *rhilts 'scorpion'< 11 .lias, *rats 'sharp'

In 'Louse' and 'Scorpion', the *s-looks like the common TB and SE Asian "animal prefix." However, it occurs rarely with animals in OC; furthermore an equivalent to the word PTB *sya 'animal, meat', the source of the *s-, does not exist in Chinese, therefore the initial in these creatures is probably also this ST nominal prefix. 5.3 Devoicing and PTB *rVoicelessness corresponds in some CH words to PTB *r- which can also show up as OC medial *-r- (ShfXiangdong 2000: 208-210) (§7.4): Grasp

shejj

sap, *nhep 'to grasp, gather up' WT rffab-rfiab-pa 'to seize or snatch together'

Gums

shen ~iJI

sin B, *nhin? 'the gums' PTB *s-nil > WT rfiil

Peaceful

tuo ~

~

sfiil 'gums'

thoi 8 , *nh6i? 'peaceful' WT mal 'to rest'~~ mnal 'to sleep'

In light of the frequent PTB (WT) *r-- *s- doublets, however, we suspect that the cause of OC devoicing was *s- as in other instances (not *r-): Sky To desire Gums Excrement To pull I throw down Body hair Sleep

WT rmu - smu ~ dmu WT rkam-pa ~ skam-pa WT rfiil ~ sfiil WT rkyag-pa - skyag-pa WT rfiil-ba ~ sfiil-ba PTB *(r-)mul- *(s-)mul PTB *(r-)miy- *(s-)miy

The vacillation between pre-initial *r- and *s- is particularly common with nasal root initial words. The cognates below ('Shame', 'Long-necked') demonstrate, first, that this uncertainty is also found within Chinese, and second, that OC medial *-r- often derives from an earlier prefix or pre-initial *r-, not from an 'infix' *-r- which is typologically unlikely in TB languages and therefore probably also unlikely in OC (but see §7.5): Shame Long-necked

xiu ltf. qian ll

[siu] *snu 'shame'>~: niu '!If [vuk] *nruk < *r-nuk 'ashamed' [khen] - [kan] *khen- *kren < *h/s-ken- *r-ken 'long-necked'

All things considered, OC medial *-r- derives from earlier *r-, even where TB cognates have initials-; and OC aspiration I voicelessness I *s- derives from earlier pre-initial *s-, even where TB cognates have pre-initial r-. 55

r

5.4 - 5.5

!

INITIAL CONSONANTS

INITIAL CONSONANTS

When the initial consonant is a stop or affricate, pre-initial *r- and *s- show up as aspiration in MC; see §5.8 below. 5.4 ST and PCH *kVoicelessness in an OC initial can correspond to ST or foreign *k- (TB d-, g-) in the initial configuration. The TB prefixes d- and g-can also correspond to OC medial *-r- (§7.4). Hull vb.

ch6ng

Excrement

shr.l*

Swallow

tiin

Evil

te Jli

Covet House Shore

:ff

if

tan~ shl~ hrdm:

LH sol), *lhol) 'to hull grain with a pestle' S. *klooiJ 'to hull rice' siB, *!hi? 'excrement, dung' PTB *k-liy 'excrement', Kanauri kli, PL *7/k(l)e2, WB khyeB th:;,n 'to swallow' PT *kl-: S. kl.finAI 'to swallow', Ahom k(l)en th:;,k, *hn;}k 'evil, wrong' WT na,t'black', gnag 'black, wicket'

~e snag 'ink'

th:;,m, *rhgm 'be covetous' ~e bin~ l::>m, *g-~m ? 'covetous' PTB *d-rum> WT drum-pa 'to desire', WB krum 'to pine' sit, *lhit? 'house, hall' WT gzi 'ground, residence, abode'

3E

lE

LH kwin 'potter's wheel' < yun

?] [win] 'even, uniform'

gu] )l

kui 8 'ghost'< wei ~ [?ui) 'overawe, terrorize'

Country

gu6 ~

kugk 'country'< yu ~ [wtk] 'boundary, region'

Additional examples may include~ qiao ~ 'hole'; ~ qii1 ~ 'a blanket'; ~ gou fi 'drain, irrigation canal'. This morpheme is not ST. Perhaps AA was the source; note that Khmer has a prefix kwhich forms nouns (Jenner and Pou 1982: xxxix-xl); or note PVM *k-ve:l 'village' < *ve:l 'to return' (Fer! us). 5.5 Other sources of devoicing Other sources of devoicing involve non-ST words where the function, if any, is outside the OC system. LH thou, *lhau 'to pour water, wash' PMiao *?leu2A 'to pour'

Iron

tH! •

Food

xUing ~~ SUIJ(BfC), *hnal)?/h 'bring food to' MY: Anc. Miao izonC(Wang Fushi) 'cooked rice, food', Yao: Biao Min 1)81]5, Mien (Chiang Rai dial.) l)haaiJ5

thet, *!hit or *lhet 'iron' Tai: S. lekDIS 'to double, to rise high' ~; 'increase, numerous' :1:~ appears to be a derivation from < deng !r [t:}l)] *t~l) 'to rise, ascend, mount, raise'

Add

Twist

if

khu~t. *khut 'dig in the ground, cave, hole' PTB: *r-ko-t 'dig'

Dig

kii

Rob

kou 70C

Girdle

xie, qie. get, khet, *get, 'khet 'girdle' WT rked-pa 'the waist, loins, the middle'

Body

qii~

khfo, *kho 'body, person' PTB *(s-)kuw > WT sku id.

Leather

kuo~

khuak, *khwak 'leather' WT skog-pa'-kog-pa 'shell, peel, rind'

5.8 Aspirated initial consonants Only one out of six MC words with velar and labial initials has an aspirated surd stop. Many other MC aspirated words probably had complex OC or ST initials which were not necessarily aspirated; others fall into certain semantic categories which OC apparently marked with aspiration. Therefore aspirated stops and affricates are not of ST origin, but are CH innovations. Several different sources of MC aspirated initial consonants can be distinguished: i

khianB, *khen? 'to send to, send away' WT skyel-ba 'to send'

Get rid

kh+aB, *kha? 'get rid, eliminate' skyag-pa. bskyags 'to lay out, expend'~ skyag ""rkyag 'dirt, excrement' (unless this ex. belongs to §5.8.5)

Dismember

phuk, *pbak 'dismember' PLB *pak - *?pak > Lahu ph§7 'unfasten, dismantle'

5.8.2

MC aspiration: causative

In some words, the aspiration may perform the same causative function as *s-and devoicing (§5.2.1-2) and may in fact be the trace of a lost s-prefix. The aspiration existed already in OC because qiang H~ (tshjal)) 'to beg' (~ jiiing U:t:f) writes 'tinkle' in Shijfng 27 4 which must likely have been simply *tshal), not *s-tsal). Aspiration also forms causatives in TB languages, e.g., Tiddirn Chin: kaiJ 'to rise' > kllaiJ 'to raise' (Geilich 1994: 171).

H

Bright

qiao

Half

pan ?#IJ Jt phone, *phans 'to cleave, divide' (i.e., 'make half') ~ ban [pane] *pans 'half'

Overturn

fu

As expected, there remains a residue where aspiration defies explanation; it includes common words like~ ku ?f [khaB] 'bitter'. In light of the secondary nature of most aspiration, words like ku may perhaps go back to an original C+h cluster that is known from Mon-Khmer languages.

5.8.1

MC aspiration: loss ofSTpre-initial

Some Middle Chinese aspirated words correspond to TB (and I or WT) initials with s- or rprefix, also to Lolo-Burmese *?-. This is parallel to devoicing (§5.2-3), although it is only a trend (Sino-Tibetan prefixes are volatile and change from language to language). Cliff

k~tn

1lJi

LH k~mc, *kh~ms 'cliff, bank, step' PTB *r-ka[:]m id. 58

5.8.3

LH kheu, *khiau 'bleached white (of bones)' (i.e., 'made white') ~~ jHio ~ [keuB] *kiau? 'bright'

-¥

•

(1) Aspiration due to loss of pre-initial (§5.8.1)

(2) Initial MC tsh-from OC or PCH *k-s- and *s-7- (§5.9.1-2) (3) Initial MC tshj-from OC *k-hl- (§5.9.3) (4) Sound-symbolic aspiration (§5.8.5-§5.8.6) (5) For sources of MC initial th-, see §5.1) ·

khoC, *khoh 'to rob, robber' PTB *r-kuw > WT rku-ba 'to steal'

Send

jf ~~ [tsek] *tsek < *;_s-kek? 'to twist, spin' appears to be a deri vate from < xl{*~ [keC] *keh 'to bind, tie up, attach'

This metathesis is rare, though. Several often cited examples have alternative explanations. 'Carpenter' ~ jHtng 1ft [dzial)C] is supposed to be cognate to WT sbyoiJ-pa, sbyal)s 'to train, exercise, practice', but the CH word has a better etymology, q.v. The same goes for c6ng iti (under~ jii ~)and the words writ.ten with the phQnetic ~ zu $(also under~ zao J'l) which Bodman had associated with WT words with initial st-, etc. (cited under~ tun '41~). This throws some doubt on the metathesis theory.

5.8.2 - 5.8.3

phuk, *phuk 'to overturn, overthrow I repeatedly' caus./iter. ~ fu fl [puk] *puk 'double, lined' (garment) ~~ fu 11! [buk] *buk 'to come back, return, restore'

MC aspiration: iterative

Words that suggest repetitive motion are often aspirated, which parallels the iterative I intensive function of the s-prefix and devoicing (§5.2.3 ). Gnaw Strike Cut open

ken • ko llP ku ~U

Pass over Stride Notches Cut through Splinter

kua ~ kur Hi ql ~ qie ~ que~

kh;)nB OCM *khem? 'to gnaw' khoBfC, *khO?/h 'to strike, attack' khuac, *khwah 'cut open, cut to pieces' ~ hwi • [yua] *grwa 'to cleave' khuac, *khrwah 'to step over, pass over' khyeB, *khwe? 'a stride' khes, *khets 'script notches' khet, *khet 'to cut, cut through' khuet, khyat, *khwet 'to break, splinter' 59

5.8.4 - 5.8.5 t~ J-i,~

Disperse Break Brush off

pf po fu

Pull off

tiao {I~

5.8.4

m

INITIAL CONSONANTS

p"tai, *phai 'divide, disperse' p 11 aic, *phaih 'to break (into pieces)' p 11 ut, *phut 'to brush off, shake off' (iterative) < fu ~ [put] *put 'clear away (dense vegetation)' theu, *lhiu 'pull leaves off branches' (iterative) < tiao {~ [deu] *JiG 'branch, to be extending (branches)'

INITIAL CONSONANTS

5.8. 6

Vault Pit Pit Empty Hollow Mouth

MC aspiration: auxiliary verbs kan

WJ

kh:::lm, *khgm 'to endure, equal to' < hin [y~m] *ggm 'have [hold] in the mouth'

15-

Able

ke

r:iJ

khaiB, *khai? 'to bear, can, be able' < he 1 {PJ~ [goiB] *gai? 'to carry'

Able

ke

:5'[

kh:::lk,

Willing

ken

-ll:.

kh:::li)B, *kMI]? 'be willing'

5.8.5

Fl

*kh~

Aspiration: outward and I or forceful motion

Lungs Breathe Sigh Breath Anger Stench Spit Spit Spit Cough Weep Weep Throw away Leave

Hole Basket Crotch Hollow Hole Husk Hollow

'be capable, able, can, conquer, vanquish'

Words that connote breathing, exhaling, an emphatic outward gesture or forceful motion such as 'spitting', 'ejecting', 'motion away', 'hewing', 'hitting', 'beating' tend to be aspirated. This is a sound-symbolic feature in CH although the origin may have been s-intensives: fei H$ chou 1Jl xi ·I~ ql: ~

LH p 11 uas, *ph as ( < *ph os ?) 'lungs' tshu 'sound of an ox breathing' hts. kh:::ls 'sigh' k11ts 'breath' 3~ kai 'I'*" lk11:::ls] 'be angry' fen ?J; p 11 unBfC 'anger, angry' pregnancy'

Reflexes ofMon-Khmer affricates in Chinese 5.10 MK languages have typically one set of affricate consonants which are transcribed c, j, eh, and jh following Indic conventions. Two sets, one sibilant ( ts, dz), one perhaps palatal (ts, dz), can be reconstructed for PMK (Ferlus MKS 7, 1978: 1-38). This may explain the multiple correspondences with OC, but this issue requires further study. Since OC had only dental affricates, these had to be substituted for the foreign palatals.

5.10.1

MK c,j=MCaffricates

To ford

jl ~if!f [tseiC] *tsfh or *tsaih 'to ford, cross' AA: M on inscr. cnis [cn!Zis] > cnih 'a ghat, place of access to river ... , landing place'< cis [c~s] 'to go down, descend'

To polish

cuo ~ [tshai] *tshai 'to rub, polish'