CC BY-NC-ND
23
Gong Xun
School of Oriental and African Studies, University of London; minus273cn@gmail.com
Chinese loans in Old Vietnamese with a sesquisyllabic phonology 1
While consonant clusters, taken broadly to include presyllables, are commonly hypothesized for Old Chinese, little direct evidence is available for establishing the early forms of specific words. This essay examines a hitherto overlooked source: Old Vietnamese, a language substantially attested in a single document, which writes certain words, monosyllabic in modern Vietnamese, in an orthography suggesting sesquisyllabic phonology. For a number of words loaned from Chinese, Old Vietnamese provides the only testimony of the form of the Vietic borrowing. The small list of currently known sesquisyllabic words of Chinese origin attested in this document includes examples of both words with a secure initial Chinese cluster and words with plausible Vietic-internal prefixation.
Keywords: Old Chinese language, Old Vietnamese language, historical reconstruction, sesqui-syllabic words, prefixal morphology.
1. Context
Old Vietnamese2 is a Vietic language of which substantial attestation is limited to a single document, the JSfife Phat thuyet Dai bao phu mau an trong kinh (Foshuo Dabao
fumu enzhdngjing, "Sutra explained by the Buddha on the Great Repayment of the Heavy Debt to Parents", henceforth Dai bao). The language of this document preserves Proto-Vietic sesquisyllabic phonology. In Old Vietnamese, we find words of Chinese origin like if *s-kuong 'mirror', borrowed from Chinese M *kix^H.3 This paper examines whether words like
1 I would like to acknowledge the generous support of the European Research Council for supporting this research, under the auspices of 'Beyond Boundaries: Religion, Region, Language and the State' (ERC SynergyProject 609823 ASIA). This paper would have remained a mere idea, were it not for Huang Shiqi and Masaaki Shimizu who took the trouble to send or borrow me the necessary literature. Nathan Hill gave precious advice and comments throughout the drafting of this essay. Finally, I would like to thank the anonymous reviewers, Laurent Sagart, Masaaki Shimizu, Dao Huy Linh, Guillaume Jacques, Alexis Michaud, Mark Alves, Nguyen Minh-Chau, Wolfgang Behr, Tran Trong Duong and Jonathan Smith for their comments which have considerably improved this essay.
2 On the term Old Vietnamese see §1.3. In this essay, Old Vietnamese is transcribed with reconstructed initial, in roman type, and projected Modern Vietnamese rime in italic type. For example, *plai 'heaven' denotes that I reconstruct Old Vietnamese initial *pl-, and that the Modern Vietnamese cognate of the word is trai.
3 Chinese (and Sino-Vietnamese) are transcribed in several different ways depending on the context. In proper names and terms of cultural nature, the Chinese is rendered in pinyin, in Sino-Vietnamese or both, depending on the specific context to which the term pertains.
Chinese sources of loanwords under consideration are transcribed in a modified version of Baxter's Middle Chinese transcription (1992). Most notably, the transcription of this paper uses medial -i- instead of -j-, and the vowel o is written л. Other sounds are transcribed in an IPA-like fashion: among Baxter's alternative orthographies for vowels, x, £, i are preferred to ae, ea, +. Retroflex stops are written f, fh, c[, ц, retroflex sibilants tp, tph, d,, p, z,, palatal sibilants tq, tqh, d%, p, q, Finally, we have j for y, j for ng, ? for ' and у for h.
Finally, Chu Nom orthography is transcribed first in Sino-Vietnamese followed by Middle Chinese transcription: ШЖ (pha tan < *phaH sanX). Unencoded Chu Nom characters are represented with Unicode ideographic
Journal of Language Relationship • Вопросы языкового родства • 16/2 (2017) • Pp. 55-72 • © The authors, 2017
this support the reconstruction of initial consonant clusters in Old Chinese. More specifically, it investigates whether Chinese loans in Old Vietnamese that exhibit sesquisyllabic phonology confirm or challenge Baxter and Sagart's reconstruction of Old Chinese (2014), which employs data from Vietic languages (notably Vietnamese and Ruc) to support Old Chinese clusters.
This section introduces the context of the paper, beginning with a general introduction (§1.1) to the difficulties of reconstructing Old Chinese initial consonant clusters. §1.2 discusses the importance of Vietic languages, which offer an important source of evidence for them in cases where they might have disappeared without a trace. The source Dai bao is introduced in §1.3, and the phonetic interpretation of its sesquisyllabic orthography in §1.4.
1.1. Initial consonant clusters in Old Chinese
While Middle Chinese and contemporary Chinese dialects have a simple syllabic canon, scholars have long hypothesized that Old Chinese has a more complex phonotactics, especially on the left periphery of the syllable. Scholars from Henri Maspero (1930) reconstruct initial consonant clusters like *pr- or *sm-, while recent reconstructions (Pan Wuyun 2000, Baxter and Sagart 2014) also hypothesize a sesquisyllabic phonology with initial minor syllables. Baxter and Sagart (2014), for example, reconstruct Old Chinese forms such as *k.te? for || *t$eX "paper" and *ma.lat for § *%et "tongue".
In this paper, par abus de langage, we call both initial clusters like *kt- and the sequence of an initial minor syllable followed by the initial consonant like *mdl- 'consonant clusters'. I follow the convention of Baxter and Sagart (2014) in referring to the first components, such as *k- or *ma-, as preinitials, and distinguish them as consonantal (*k-) and syllabic (*ma-).
While various kinds of evidence point to consonant clusters in Old Chinese, it is very difficult for scholars to agree on the clusters to reconstruct for precise etymons. One of the reasons for this uncertainty is that, under the type of evolution that languages of China and Mainland Southeast Asia usually undergo, different phonotactic types of clusters have almost intrinsically different rates of survival in descendants. A case in point is Old Tibetan, in which most consonants can take preinitials s- and d-. However, the existence and identity of preini-tials is much better preserved before sonorants than stops: in a typical modern Kham dialect (for example the dialects of Derge and Batang, cf. Skal-bzang 'Gyur-med and Skal-bzang Dby-angs-can, 2002), Old Tibetan ng-, sng- and dng- remain distinct as _^-4, and but k-, sk-and dk- all merged into k-. If one were reconstructing Old Tibetan without the benefit of orthographic forms, it is likely that *g-, *sq- and *dq- would be reconstructed correctly, but *k-, *sk- and *dk- would be very hard to distinguish.
This asymmetry in the preservation of clusters before obstruents and sonorants explains the curious situation of Old Chinese reconstructions, where there is a clear gradient to the levels of consensus among different reconstruction of different phonotactic types of OC clusters. For example, recent reconstructions of Old Chinese agree on either one of two theories about *sm- type initial consonants (see Mei 2012, Sagart & Baxter 2012). Given that there is an implicational hierarchy to the effect that languages with *sm- type initial consonants usually have *sk- type initial consonants (Goad 2011), Old Chinese likely had *sk- type initial consonants as well. However, authors cannot agree on which particular OC words have *sk-type consonants (see Gong & Lai 2017 for a brief account).
description characters such as LU or H. For example, the sequence "LIS" represents a character composed horizontally of I and S.
4 designates the high tone, and _ the low tone, in tonal modern dialects of Tibetan.
In order to better understand both the phonological structure of Old Chinese and the reconstruction of individual Old Chinese words, we need to search for more direct evidence of consonant clusters.
1.2. Does Vietic evidence support Old Chinese clusters?
From the very beginning of research on Vietnamese words of Chinese origin, it has been noted that certain words borrowed from Chinese exhibit lenition of initial consonants (Maspero 1912: 19-39, Wang Li 1948: 71). An example is Chinese *kixmH 'sword', which is borrowed as Vietnamese guom [y-]. This contrasts with the absence of initial lenition in the Sino-Vietnamese pronunciation of the word, kiem.
Vietnamese belongs to the Vietic branch of Austroasiatic languages. Modern Vietic language exhibit the whole typological spectrum from monosyllabic Chinese-like Viet-Muong languages5 to completely sesquisyllabic languages such as the Chut (Ruc-Sach) cluster. Haudricourt (1965) first explained Vietnamese spirant initials like v-, d- [z- < 5-], g- [y-] and ras reflecting a process of lenition caused by lost Proto-Vietic presyllables. According to this theory (cf. also Thompson 1976: 1131-1133, Ferlus 1976, 1982), Proto-Vietic presyllables, still preserved almost intact in conservative Vietic languages like Ruc, disappeared in all modern Viet-Muong languages. Their former presence triggered lenition in Mainstream (Hanoi-Saigon) Vietnamese, but not in Muong varieties nor the "Haut-Annam" (Maspero 1912) or "heterodox" dialects of Vietnamese, such as Vinh (Ferlus 1991) or Quang Binh (Michaud, Ferlus and Nguyen 2015).
Table 1: Lenition in Vietnamese
Proto-Vietic Vietnamese Muong4 Ruc6
*p- 'four' bon ([b-]) ponW po:n
*CVp- 'lime' vôi polW kapu:l
*s- 'hand' tay thajW si:
*CVs- 'snake' ran thapW pasi:yi
As predicted by this theory of the origin of Vietnamese lenition, for Chinese words just as for native words, conservative Vietic languages have a cognate with a prefix: Vietnamese guom [y-] 'sword' is cognate to Ruc tdkidm. Pulleyblank (1981: 281-286), the first focused treatment of Chinese loans in Vietnamese showing lenition, considers the possibility that the preinitial was present in the original Chinese form 'of great interest' but 'not easy to pursue further without additional information to enable one to determine the kind of clusters involved'.7
5 This essay follows the terminology of Hayes (1992), now in general usage. Vietic designates the larger group, which includes notably Chut (Ruc-Sach). Viêt-Mwong, a term which formerly often comprises all Vietic languages, designates only the languages spoken by the ethnic Vietnamese and Muong groups. For the terminological questions, cf. Michaud, Ferlus and Nguyên (2015: 126).
6 Ruc forms are cited from Nguyên, Trân and Ferlus (1988). 'Muong' designate Khen Muong (Hoa Binh province), transcribed by Milton and Muriel Barker, and cited from Thompson (1976).
7 Maspero (1912: 21-23 et passim) deems Vietnamese words with lenition to be a 'dérivé récemment formé' of their counterparts without lenition in Muong and Vinh-type Vietnamese dialects. Wang Li (1948) includes lenition in what he calls the process of Vietnamization Viet hoa), which Nguyên Tài Cân (1979) defines as "development under a different path than that of Sino-Vietnamese readings" (diên biên theo mot con duong khac voi cach doc Han Viet).
Baxter and Sagart (2014) were the first to systematically use loanword material in Vietic languages to reconstruct consonant clusters in Old Chinese. By comparing the Vietic data with Chinese loans in the Kra-Dai language Lakkia, Baxter and Sagart (2014: 36-37, 93-97) give a compelling argument that preinitials in these words do come from the Chinese source. Their examples are II *t$eX 'paper', borrowed as Vietnamese giay [z-], Ruc kdcdj, Lakkia khjei3, and M *dznk 'bandit', borrowed as Vietnamese giac [z-], Ruc kdcnk, Lakkia kjaks. The oldest sizeable set of Chinese loans in Vietic and Lakkia, just as their counterparts in Tai and Kam-Sui, all result from the same event of massive borrowing which follows the Qín-Hán conquest of the region corresponding to modern Guangdong, Guangxi and Vietnam. For a Chinese word to receive a Vietic prefix and be then borrowed into Lakkia would be extremely implausible. Even if the number of correspondences is limited, we have here an important argument for the Chinese origin of the preinitials.
In the system of Baxter and Sagart (2014), Vietic preinitials are understood as reflecting genuine Old Chinese preinitials, thus I *t$eX 'paper', Vietic *k- < *k.te?; M *dzak 'bandit', Vietic *k- < *k.dz?ak.
While there are a great number of words of Chinese origin with a softened initial in Vietnamese, only a few of them can be found in conservative Vietic languages as Ruc. In order to further evaluate Baxter and Sagart's use of Vietic data in Old Chinese reconstruction, we need additional sources that attest to Chinese loanwords with preinitials.
1.3. Old Vietnamese and Dai báo
We find many cases where a word of Chinese origin shows a lenited modern reflex in Vietnamese, but is not attested in conservative Vietic languages such as Ruc and Thavung. We know that the Proto-Viet-Muong form must have contained a preinitial, but it is unclear which.
The preponderance of older Chinese loans in Vietnamese and Viet-Muong but not other Vietic languages is not surprising given the linguistic history of Vietic. As shown by Ferlus' work (2010) on the Khmer names of the duodecimal year cycle (ift^ dizhi, Dia Chi 'earthly branches'), the Vietic branch of Austroasiatic had already split into subbranches by the time the early layer of Chinese words arrived. Chinese words entered the ancestor of today's Viet-Muong, and were then borrowed into other languages. It is not surprising that lexical Sinicization was followed by phonological Sinicization: the group that originally received the Chinese loans, which also contains the largest number of them, was then monosyllabicized under the weight of Chinese influence. Middle Vietnamese, as attested in Alexandre de Rhodes' dictionary (1651), had already lost all sesquisyllabic preinitials, and retained only a few relictual initial clusters such as bl-.
This paper continues Baxter and Sagart's work (2014) on Vietic borrowings in Old Chinese by examining a remarkable document, which greatly improves our knowledge of Viet-Muong historical phonology. The document, ffit^l^SMSIM Phat thuyet Dai báo phu mau an trong kinh ("Sütra explained by the Buddha on the Great Repayment of the Heavy Debt to Parents", henceforth Dai báo), is held in the Société asiatique, Paris. It is a version of a popular Chinese apochyphon more commonly known under the title ^SMS^flM Fumu Énzhong Nánbaojing, Phu mau an trong nan báo kinh ("Sütra on the Difficulty of Repaying the Heavy Debt to Parents"), in which the Chinese text is accompanied by a vernacular translation (called
Maspero, Wáng Li and Nguyen Tái Can essentially postulate a non-Neogrammarian unconditional split, the condition of which is convincingly explained by Haudricourt, Thompson and Ferlus as Proto-Vietic presyllables. As a terminological convenience, we can understand Vietnamization as morphological prefixation within Proto-Vietic or Proto-Viet-Muong, which developed into Vietnamese lenition.
giai am in Vietnam) in a rudimentary form of Chu Nom, where vernacular words are written with Chinese characters and modified versions thereof.
The language of the vernacular translation in Dai bao is clearly Viet-Muong. I propose to call this language Old Vietnamese, as the language reflected in Dai bao shows an important archaism: certain words that are monosyllabic in Modern Vietnamese are written with two Chinese characters (digraphic orthography), the first of which corresponding to a preinitial in conservative Vietic languages and proto-Vietic: ran 'snake', Middle Vietnamese ran 'cobra; anguis' (Rhodes 1651: 636), is written (pha tan < *phaH sanX).
The traditional periodization schemes of Vietnamese distinguish a period Archaic Vietnamese (Nguyen Dinh-Hoa, 2009) or Old Vietnamese (vietnamien ancien, Ferlus, 2010). I believe that calling the language of the Dai bao is justified by the fact that there is a fundamental divide between, on the one hand, Archaic Vietnamese as traditionally defined as well as the language of Dai bao, and, on the other hand, Middle Vietnamese of Annan Yiyu (^^titt, An Nam Dich Ngu) and Rhodes (1651). For example, 'snake' was borrowed from Archaic Vietnamese to Old Khmer as masan (Ferlus 2010: 9). The Archaic Vietnamese form clearly sides with the form in the Dai Bao *p-san , and is distinguished from Middle and Modern Vietnamese ran. Old Vietnamese, as defined in this essay, can be regarded as the last stage, uniquely attested, of Archaic/Old Vietnamese as hitherto defined in the literature.
Apart from the Dai bao, an extensive text, Old Vietnamese is attested in a much smaller scale in two sources: remnants in Modern Chu Nom usage (cf. §3.1) and proper names attested in Chinese-language inscriptions, most importantly the Ho Thanh Mountain Ninh
Binh province) inscription (Shimizu, Le & Momoki 2005).
Dai bao was first brought to scholarly attention when a copy of it was sent to the Han Nom Institute in 1979; Nguyen Ngoc San (1982) wrote about the digraphic orthography of Dai bao, which he interpreted as clusters and preglottalized consonants. Shimizu Masaaki (1996) made the first systematic study of Dai bao's digraphic orthography, in which he collected a corpus of 24 glyph-word pairs exhibiting cluster orthography, as well as 47 glyph-word pairs exhibiting sesquisyllabic orthography. Hoang Thi Ngo's candidate thesis (1996), later published as Hoang (1999), is the first and only transcription of the text into modern Vietnamese.
Dai bao shows a great number of cases of digraphic orthography, where a word, monosyllabic in Modern Vietnamese, is written with two Chinese characters (two-character orthography), or in a composite character made of two different characters (composite-character orthography). Following Shimizu (1996), we classify Old Vietnamese digraphic orthography into two phono-tactic classes:
• Consonant-liquid cluster orthography: The word transcribed is strictly monosyllabic and has a consonant-liquid initial cluster: CRVC; the first character transcribes the initial consonant C; the second character transcribes the medial RVC.
For example, the word trbi 'heaven', MViet blbi 'ceo; c^lum' (Rhodes 1651: 45), is written © (E+M ba le < *pa lieiH) in Dai bao. I reconstruct Old Vietnamese *plbi, the first character E ba < pa transcribes the initial consonant *p, the character M le < *lieiH transcribes the remaining part of the syllable *loi.
• Sesquisyllabic orthography8: the word transcribed is sesquisyllabic C-CVC. The first character transcribes the presyllable C-; the second character transcribes the initial and the rime CVC.
8 This type is called disyllabic construction so-onsetsu kozo) in Shimizu (1996).
For example, the word ran 'snake', MViet ran 'cobra; anguis' (Rhodes 1651: 636), is written (phâ tân < *phaH sanX). I reconstruct Old Vietnamese *p-san, cf. Ruc pdsi:p. The first character 55 phâ < *phaH transcribes the preinitial *p-, and the second character f tân < *sanX transcribes the remaining part *san.
In this study, we focus our attention on the second kind of digraphic orthography, which relies on Vietic and Old Vietnamese sources. The first kind of digraphic orthography, which transcribe consonant clusters of the TR- type, will be deferred to another discussion along with other sources of evidence.
Concerning the dating of the Dai Bâo, the document held in the Société asiatique itself was printed at the initiative of Trinh Quan iPfg around 1730 (Shimizu 1996: 84). The text conspicuously avoids the character ^[J loi, the name of the first emperor (c. 1384-1433) of the Lê dynasty (1428-1788). Mainly based on this taboo, Shimizu (1996, 2015) dates the text to the beginning of the Lê dynasty (XVth century). On the other hand, Nguyên Tài Can (2008) suggests that the text is likely to be copied from a Trân-dynasty (1225-1400) precedent, by pointing to a number of less conspicuous taboo practices in the text9.
A number of texts are dated, with more or less certainty, to the Trân dynasty. Trân Trong Duong (2011) pointed out that, compared to known texts dated to the Trân dynasty, Dai Bâo contains at least an order of magnitude more cases of two-character orthography. For example, compared with the text Chi nam ngoc âm giâi nghîa, which shows the most cases
of orthography in two characters apart from the Dai Bâo, there are 16 cases of two-character orthography for a total length of 15000 characters; in the Dai Bâo, on the other hand, Trân Trong Duong counts 103 cases of two-character orthography for a total length of 4942 characters. Independently of Trân Trong Duong's work, I checked the list of words showing sesquisyllabic orthography in Early Vietnamese texts given in Trân and Nguyen (2007). Among the 62 examples, 50 are attested only in Dai bâo. The result also shows that there is a qualitative difference between the language of Dai bâo and that of later texts. As Nguyen Quang Hong (2008: 127-144) and Trân Trong Duong (2011) observe, the text should be dated before Trân dynasty, likely to XIIth century.
1.4. The reconstruction of Old Vietnamese preinitials in this paper
The reconstruction of Old Vietnamese preinitials in this paper roughly follows that of Shimizu (1996). I reconstruct the following Old Vietnamese preinitials: *p-, *t-, *k-, *ç-, *s-, ?-. We note that the list of Old Vietnamese preinitials coincides with the more numerous preinitials in Michel Ferlus's reconstruction of Proto-Vietic (2007): *p-, *t-, *k-, *c-, *s-, *?a. In the following list, I provide the spellers and corresponding Proto-Vietic initial for each Old Vietnamese preinitial.
Old Vietnamese *p- < Proto-Vietic *p-
Spellers: ^ (ba < *pa), E (ba < *pœ), 55 (phâ < *phaH)
Example: *p-san 'snake' 55f (phâ tân < *phaH sanX) = ran, Middle Vietnamese ran 'cobra; anguis' (Rhodes 1651: 636)
Vietic cognates: Ruc pdsi:p 'snake'
Old Vietnamese *t-, cf. Proto-Vietic *t-
Speller: # (da < *ta)
9 Tran Trong Duong (2010) disagrees with Nguyen Tai Can's analysis, and takes the taboo characters under question to be normal graph variants at the period.
Example: *t-mai 'shoulder' (da mai < *ta mej) = vai, Middle Vietnamese pai 'ombros:
humerus' (Rhodes 1651: 65) Vietic cognates: Thavung ?api:p / ?apir 'shoulder'
Old Vietnamese *k- < Proto-Vietic *k-
Spellers: ^ (kha < *khaX), ^ (ca < *kaH)10
Example: *k-pa 'remember' (kha nhu < *khaX p.nX) = nha, Middle Vietnamese da or nha
'lembrarse; recordor' (Rhodes 1651: 175) Vietic cognates: Ruc kdpd: 'to remember'
Old Vietnamese < Proto-Vietic *c-
Spellers: ^ (xa < *t$hx), # (xa < *$xX)
Example: *q,-mang 'hear' ^^ (xa mang < *t$hx marfX), (xa mang < maqX) = mang, obsolete in Modern Vietnamese, but cf. Kieu l.535 ^fg^^MfM Mang tin xiet noi kinh hoang 'How he was scared after hearing the news!' Middle Vietnamese cf. mang tin fama; fama' (Rhodes 1651: 450) Vietic cognates: Ruc camar 'to hear, listen'
Note: The Chinese character ^ has two alternative readings: MC *kin = Mandarin ju, Sino-Vietnamese cu; MC *t$hx = Mandarin che, Sino-Vietnamese xa. Nguyen Huu Vinh et al. (2009: 764) and Tran Trong Duong (2012) read ^ as cu < *kin, reconstructing *k-mang. This reading is to be rejected considering the alternative orthography and Vietic cognates.
Old Vietnamese *s-, cf. Proto-Vietic *s-Speller: ^ (tu < *si)"
Example: *s-pui 'merry, joyful' © tu bdi < *si pwni) = Vietnamese vui, Middle Viet-
namese pui 'allegre; hilaris' (Rhodes 1651: 74) Vietic cognates: Ruc tupuj 'merry, happy'
Old Vietnamese *?-, comparable to Proto-Vietic *?a Speller: M (a < a)
Example: *a-poi 'early' M^ (a bdi < *a pwni) = Vietnamese voi 'hasty', Middle Vietnamese
poi 'cousa apressada; properus' (Rhodes 1651: 71). (No known cognates in conservative Vietic languages)
Shimizu (1996) reconstructs *6- (or *p-), *ph-, *d- (or *t-), *l-, *s-, *kh-, *s- and *?-. My reconstruction of Old Vietnamese preinitials differs from his treatment in the following respects:
• Voiceless unaspirated stops, akin to other Vietic languages and reconstructed Proto-Vietic, are preferred to implosives.
• I do not admit the difference between *p- and *ph-. Although two different sets of spellers are used with bilabial stops, with unaspirated stops (^ ba < *pa and E ba < *px) and with aspirated stops (5$ pha < *phaH), the same behaviour is seen in *k-, with
10 This character, a simplified form of fi (ca < *kaH) 'one', is a graphical variant of other simplified forms of fi in currency in East Asia, such as Simplified Chinese ^ (ge) or Japanese ■ (ka). All derive from one half of the bamboo component in fi.
11 Nguyen Ngoc San (1982) and Tran Trong Duong (p.c.) propose that the speller ^ (tw < *si) should be read as *t- instead of *s-. This is unlikely, since the sound change t < *s in syllable initial did not happen before lenition, as ä < *t gives Middle Vietnamese d [ö] in lenition, but t < *s gives r.
unaspirated speller ^ (ca < *kaH) as well as aspirated speller ^ (kha < *khaX). In both cases, there is no pressing evidence to support the distinction, which does not conform to usual Austroasiatic phonological patterns, as in those of Khmer or reconstructed proto-Vietic.
• I do not consider the word (la da < *la txrfX) 'rock', which would be *l-ta according to Shimizu (1996) and other studies, as a case of an Old Vietnamese sesquisyllabic word. It does not cause lenition in Modern Vietnamese: da not *da; it survived in forms of Vietnamese later than Old Vietnamese: we find (first character la < *la, second character Nom with # da < *ta as phonetic component) in H^ff^- Quoc am thi tap; more strikingly, we find Middle Vietnamese la da 'lagea de pedra, lapis planus' (Rhodes 1651: 390).
2. Some sesquisyllabic Old Vietnamese words borrowed from Chinese
In this section, I study some words of Chinese origin from Dai bao. I collect Old Vietnamese words in two-character orthography from two studies of the document (Shimizu 1996, Tran and Nguyen 2007), and select the words given in the two studies for which the Chinese origin appears secure to me.
The quoc ngu reading of the text of Dai bao is based on Hoang Thi Ngo (1999), who provides the only complete transcription so far available. I report all cases where my judgment differs from that of Hoang Thi Ngo (1999), Shimizu (1996) or Tran and Nguyen (2007).
Three examples will be discussed not in this section, but in subsequent sections. Section 3 discusses if *s-kwcmg 'mirror', which bears important consequences on several aspects concerning Old Chinese reconstruction in general and Baxter and Sagart's reconstruction (2014) in particular. Section 4 treats two words with Old Vietnamese preinitial *?-, which could be a case of Vietic-internal prefixation.
2.1. EK *p-Quc 'to spread, to apply' < — ik 'to wipe'
(1) ek ^ m
*p-Quc nhung phan spread PL powder
'apply different kinds of powder' (Dai bao 8a-1)
Original: (^Aftffi ) '(Women live in this world,) apply themselves thickly
with rouge and powder... '
I follow Tran and Nguyen (2007, #37) in reading EK (ba-thuc < *pa-$ik) as xuc 'to anoint, to rub, to apply', Middle Vietnamese xuc 'untar, ungir; ungo' (Rhodes 1651: 896). Hoang Thi Ngo (1999: 147) has suc 'force', likely merely an alternative spelling of the same word.
The Old Vietnamese reconstructs to *p-Quc. The Old Vietnamese preinitial *p- fails to cause lenition in and yields x- ([s]) in Modern Vietnamese. This behaviour is probably regular, and observed in *k-^a > xa 'far' (§ 2.3) and *7-q,udng > xuang 'chant' (§ 4).
The Chinese original of E-K *p-Quc is ^ *$ik 'to wipe' < OC (Baxter-Sagart) *lak. The Old Vietnamese form supports an initial cluster *pl- or *pal- in Old Chinese.
Under the Baxter-Sagart system, a syllabic preinitial is lost before *1- in pre-Middle Chinese, while a non-syllabic preinitial prevails over *1-: cf. ^ *thwat < *l?ot < *ma-l?ot 'peel off'; % *pher < *ph?eq < *[p.l]?eq 'frank words'. The preinitial needs to be syllabic *pa- under the
Baxter-Sagart system. Old Vietnamese data support a revised Baxter-Sagart reconstruction of K *$ik 'to wipe' as *pa.lak.
2.2. *t-pen 'edge' < S *pen 'edge, side'
(2) £? & m
chang co *t-pen coi
NEG have edge border
'does not have edges or borders' (Dai bao 44b-3)
Original: '(all the Buddhas of lands as countless as specks of
dust) without borders'
I read (da-bien < ta-pen) as Modern Vietnamese ven ' (river)bank; near', cf. Middle Vietnamese uen sou 'borda do rio; ripa fluminis, melius' (Rhodes 1651: 865)12. My reading differs from Tran and Nguyen (2007, #79) and Hoang Thi Ngo (1999: 147), both reading bien, the Sino-Vietnamese reading of the same etymon S *pen 'edge, side'. Compared with bien, or other words deriving from the Chinese etymon, such as ben 'side, edge', the reading ven is preferable in order to account for the presyllable, which triggers lenition: Old Vietnamese *t- should cause lenition of initial *p into Vietnamese v. Also note the locution ven coi survived in the glossing register of Modern Vietnamese: in Thieu Chuu's Han-Viet tu dien (1942) the Chinese character *d%we 'frontier, border' is glossed as ven coi.
The Old Vietnamese is reconstructed as *t-pen. As is explained above, Old Vietnamese *t-p- yields a lenited initial in Modern Vietnamese v-.
The Chinese original is S *pen 'edge, side', Baxter-Sagart *p?e[n]. The rime correspondence, where Chinese e-like vowel is rendered as Vietnamese e [e] puts the word among the oldest borrowings from Chinese to Vietnamese, cf. forms preserving Old Chinese r-: sen 'lotus' < *kr-, borrowed from Chinese 2 *len, Baxter-Sagart *k.[r]?e[n]; rem 'curtain', borrowed from Chinese ^ *liem, Baxter-Sagart *rem.
Chiang Chia-lu (2011: 106) proposes another etymology for ven, namely from Chinese t^/m 'edge' *jwen < Baxter-Sagart *lon. As she herself recognizes, it is highly improbable for Middle Chinese j- < Old Chinese *l- to be rendered in Vietnamese by v-, an anomaly she explains as a late borrowing from an unspecified Southern Chinese dialect. Analysing the word as coming from S *pen 'edge, side', which is furthermore supported by the Chu Nom orthography S and Bn^, avoids these formal problems.
The Old Vietnamese form *t-pen for the Chinese word S *pen 'edge, side' supports an initial cluster *tp- or *tap- in Old Chinese.
Under the Baxter-Sagart system, syllabic and non-syllabic preinitials exist with the same consonantism, which has different treatments in Late Old and Middle Chinese. A syllabic pre-initial is lost before a voiceless stop, cf. A *kiuwX < *[k]u? 'nine', which could derive from *ta.ku? (Baxter and Sagart 2014: 155). On the other hand, a non-syllabic alveolar preinitial prevails over a grave initial: cf. ^ *t$uwX < *tu? < *[t.p]a? 'peel off'. The preinitial in S *pen needs to be syllabic *ta- under the Baxter-Sagart system. Old Vietnamese supports a revised reconstruction under Baxter-Sagart system as S *ta.p?e[n].
12 Guillaume Jacques (p.c.) points to a difficulty in this etymology: the word is spellt uen in Rhodes, with the letter v- (probably [w-]) instead of ^ ([|3-]), the predicted outcome of a p- with lenition. However, as Haudricourt (1974) notes, there is a fluctuation between letters p and v in Rhodes. For example, pai 'ombros; humerus' (Rhodes 1651: 65), annotated 'alij vai'.
2.3. ^^ *k-Qa < ^ 'far'
(3) # ^
di *k-Qfl
go far
'go far' (Dai bao 15b-1)
Original: ) '(the moral debt of parents missing you when you) travel far'
(4) & M & ^ ang na cang *k-$a father mother more far
'the parents become still more distant' (Dai bao 36b-5)
Original: '(Parents look for a wife for their son;
after they obtained another woman,) the parents become on the contrary distant from their son.'
(5) # H m ?
dang chiu chang *k-$a
will suffer NEG far
'will suffer it; it's not far' (Dai bao 43a-2)
Original: '(We will suffer the bitter repayment in the three
ways;) we will suffer it not far away'
The orthographies ^^ (kha da < *khaX jx), ^^ (kha xa < *khaX $x) and ^^ (kha xa < *khaX tqhx) transparently spell the Modern Vietnamese word xa 'far', Middle Vietnamese xa 'longe; distans' (Rhodes 1651: 879), as Shimizu (1996), Tran and Nguyen (2007, #90, #91) and Hoang Thi Ngo (1999: 153 et passim) observe.
The Old Vietnamese form of this word reconstructs to *k-$a. The Old Vietnamese preini-tial *k- fails to cause lenition in and yields x- [s] in Modern Vietnamese. This behaviour is probably regular and observed in *p-quc 'to spread, to apply' > xuc 'to spread, to apply' (§ 2.1) and *7-q,uang > xuang 'chant' (§ 4). The orthography ^^ (kha da < *khaX jx) might indicate an alternative pronunciation with lenition *k-^a, which did not survive into Modern Vietnamese.
The Chinese original is ^ 'far'13. The Chinese character, which contains the component M 'cowrie shell, wealth', is generally used to write a word that has the sense 'buy or sell on credit, defer payment'. This economic sense must be original, as it is the only one attested in pre-Han and Han texts such as Rites of Zhou Ml® Zhouli and Book of Han MM Hanshu (fascicles 24, 91 and 99), as well as the definition given in the Shuowen tt^.
The sense 'far' is attested much later. Its first occurrence according to Hanyu Da Cidian (Luo Zhufeng et al. 1993) is in Baopuzi ^^^ (Book of the Master Who Embraces Simplicity), a text ascribed to ^^ Ge Hong (283-343), who led his whole life in ^^ Jurong, in the Jiangzuo (Lower Yangtse Basin) region. The subsequent attestations are from Southern Dynasties literature.
The word ^ 'far' is likely to derive from ^ *tpx 'to defer payment', via the following semantic development 'defer payment' > 'delay, postpone'14 > 'late'15 > 'far'. The word is likely
13 The Southern provenance of the Chinese word suggests another possibility, namely that the Chinese word
is borrowed from Vietnamese. This possibility can be dismissed, as the original Vietic word for 'far' is Vietnamese ngai - reduced to a secondary synonym of xa, as in the synonymic locution xa ngai - and Ruc cara:j.
14 Cf. Tao Qian 1%^'s poem fP^HW^KWW (~ 400 CE): "I'm slowly waiting for the autumn harvest, but the crops look meagre and poor; the harvest will be delayed long."
a Southern dialect word which entered the literary language during the period of the cultural dominance of the Jiangzuo region over the rest of China (roughly 317-619).
We finally note that the word ^ has a third sense in Literary Chinese, namely 'many', which might have been borrowed into Vietnamese as xe (Tran Trong Duong, 2012b).
The late date and derived sense suggest that Old Vietnamese *k-^a is borrowed from a Southern dialect form that corresponds with Middle Chinese Barring unexpected later prefixation, however, the *k in the source dialect to the Vietnamese form can be projected back in Old Chinese, and supports an initial cluster *kl- or *kal- in Old Chinese. Under the Baxter-Sagart system, the Old Vietnamese data suggest that the reconstruction of ^ currently *1A, should be revised to *ka.lA.
2.4. *k-cieng < # *tsierX 'well'
(6) x X M LU№
trong *k-cieng trong bep chon da
in well in stove place stone
'to wells, to stoves or the milling-places' (Dai bao 35a-3,4)
Original: '(when the children) go to neighbours east and
west, or to wells, stoves, pestles or mills'
I follow Tran and Nguyen (2007: #45) in reading (ca-chinh < *kaH-tpierH) as gieng 'well', Middle Vietnamese gyeng 'pogo de agoa pera beber; puteus aquae ad potum' (Rhodes 1651: 283). Hoang Thi Ngo (1999: 165) reads chieng, with the voiceless initial ch- based on the spelling ^ chinh < *t$ierH. This reading should be rejected for two reasons. First, no word chieng 'well' exists in Modern Vietnamese, a fact indirectly admitted by Hoang herself, who explicitly glosses this form as 'gieng'. Second, Old Vietnamese *k-cieng regularly results in modern gieng by lenition.
The Old Vietnamese reconstructs to *k-cieng. Old Vietnamese *k-c- yields lenited initial in Modern Vietnamese gi-. Compare for example M *dzak 'bandit', where Vietnamese giac derives by lenition from a form akin to Ruc kdcnk.
The Chinese original is clearly ^ tsierX 'well'. The Old Vietnamese form supports an initial cluster *kts- or *kats- in Old Chinese. Specifically, under the Baxter-Sagart system, the current reconstruction *C.tseq? can be safely refined to *k.tseq?.
2.5. B — if *k-kam?? <? ff, if *tpim 'pin, needle'
(7) ® If B ~ B—i dut gan ruot lam *k-kam?? break liver intestines make pin
'breaks liver and intestines into pin-sized pieces' (Dai bao 41b-5)
Original: ^KfM '(the work of bringing children up) breaks the liver and intestines
(of parents) into inch-sized pieces'
As is the case for 'edge', we have again the situation where we need to choose a reading among different Vietnamese words meaning 'pin, needle', with similar pronunciation, and which can all be written with Chu Nom characters based on the Chinese characters ff, if *tpim.
15 Compare the locution ^ffi *<pX-tshiok 'later or sooner, long or short': "Today is dif-
ferent from long ago; (naturally, whether the period of mourning is) long or short is different" Book of Song, fascicle 15 i^föj— = Töngdian, fascicle 82).
Judging from dictionaries of later Chu Nom usage, the major possibilities are gam, kim and Sino-Vietnamese chdm. Another word, ghim, is usually written ^ in later Chu Nom, but the orthography ff, $f is not unimaginable.
Tran Trong Duong (p.c.) suggests that the character HXi should be read as tram 'hundred', with lam tram interpreted as 'make into a hundred pieces'. This is an unlikely interpretation, as tr- does not show any sign of being confounded with ch- until very late. In 'modern' Chu Nom, tram is written in as # (ldm < *lim) or with characters such as M with the phonogram
The Vietnamese words concerned has been hitherto analysed as deriving from Chinese ff, if *t$im 'pin, needle', Baxter-Sagart *t.[k]am. However, Baxter and Sagart (2017) propose that the words kim and ghim derive rather from Chinese ^ *kim 'metal, bronze', Baxter-Sagart *k(r)[a]m.
Tran and Nguyen (2007: #45) interpret the composite character HXi as a case of di-graphic orthography, composed of X (ca < *kaH) and $f (chdm < *t$im). Tran and Nguyen's reading chdm does not conform with their orthographical interpretation. If their orthographical interpretation is correct, the Modern Vietnamese reading would be one showing lenition, namely gam (or ghim), which would derive from Old Vietnamese *k-kam.
However, this reconstruction *k-kam is problematic on both internal and external grounds. Internally, sesquisyllabic languages typically do not admit the homorganic stops as the preinitial and as the initial: Ferlus' Proto-Vietic (2007) contains no syllables with *k-k- or *t-d-. Externally, the Baxter-Sagart reconstruction *t.[k]am is supported by Lakkia them1.
Another possible interpretation of the orthography is an early case of the dau ca (the ca X sign, cf. Nguyen Quang Hong, 2012), an orthographic device that indicates that the character is a specifically Nom character, not to be read a la chinoise. If HXi is a case of dau ca, all the non-Sino-Vietnamese readings, namely gam, kim (and probably ghim) are possible, with no implications on Old Chinese preinitials.
3. ^ *s-kuong < M *kixyH 'mirror': Old Vietnamese preinitials and Old Chinese morphology
In this section, I examine the Old Vietnamese word if *s-kwong, borrowed from Chinese M *kix^H 'mirror', Baxter-Sagart *C.qraq?-s, which touches on several questions of interest in the reconstruction of Old Chinese. The Old Vietnamese form is reconstructed in §3.1. The reconstructed Vietnamese form enables the Baxter-Sagart form to be refined to *s.qraq?-s, which, however, reveals an internal inconsistency in the Baxter-Sagart reconstruction of Old Chinese, discussed in §3.2. The unique value of this loanword is discussed in §3.3, as, regardless of the phonological interpretation, this word offers an example of an Old Chinese *s-stop cluster which is supported both by morphology with well-known Tibeto-Burman correlates as well as directly transcribed with s- in foreign borrowing.
3.1. Reconstruction of the Old Vietnamese form
(8) m ^ m x m
*s-kuang sang bo trong gia
mirror clear put in stand
'A clear mirror is put on a stand.' (Dai bao 12a-5)
Original: 'A clear mirror covers the dressing table.'
I agree with Hoang Thi Ngo (1999: 151) and Tran and Nguyen (2007: #115) in reading if tu-cuong < *si-kiar) as guong 'mirror', Middle Vietnamese guong 'espelho; speculum' (Rhodes 1651: 301). This reading is further supported by the fact that the glyph if survived into later Chu Nom usage, spelling the word guong 'mirror'.
The Old Vietnamese reconstructs to *s-kuong. Old Vietnamese *s-k- corresponds with lenited Modern Vietnamese g- [y-].
This word is borrowed from Chinese ^ *kixrH 'mirror'. The correspondence between Chinese qu ^ and Vietnamese bang ^ (ngang and huyen) tones is characteristic of older loans (cf. Alves 2016: 271). The Vietnamese data supports a cluster *s(a)k- or *s(a)q- in Old Chinese.
3.2. Can we reconstruct Baxter-Sagart *s.qraq?-s?
On account of the Vietnamese initial lenition, Baxter and Sagart (2014) reconstruct *C.qraq?-s for this word. The Old Vietnamese data enables the refinement of this reconstruction into *s.qraq?-s.
This emendation, as attractive as it seems, is not compatible with the Baxter-Sagart reconstruction of Old Chinese as it currently stands. Initial *s.q(?)- is supposed to yield Middle Chinese s-, while *s.q(?)r- should yield Middle Chinese tp- (Baxter and Sagart 2014: 137-138). The amended reconstruction *s.qraq?-s would have become Middle Chinese *tpixrH and not kixrH.
Further examination reveals that there is a more general problem with *C.q- clusters in the Baxter-Sagart reconstruction of Old Chinese. Baxter-Sagart *C- stands for a preinitial that "cannot be identified because it has been lost in all the pronunciations under consideration" (Baxter and Sagart 2014: 168). According to this definition, whenever there is a *C- in Baxter-Sagart, it can only stand for one of the preinitials reconstructed in the Baxter-Sagart OC phonology: *N-, *m-, *s-, *p-, *t-, *k-:
• *N.q- and *m.q- give Middle Chinese j- < *g;
• *s.q- gives Middle Chinese s-;
• *p.q- has no examples but should give Middle Chinese p-, cf. *p.k- which gives p- and *p.qh- which gives ph-;
• *t.q- gives Middle Chinese tp- < *t-;
• *k.q- probably did not exist, but would be, indeed, the only one that could give Middle Chinese k-.
Baxter-Sagart *C.q- cannot give Middle Chinese k-, except in the unlikely case where *C is *k.
On the other hand, Baxter and Sagart (2014) explicitly reconstruct *C.q- in order to account for the word ^ kixrH 'mirror'. The velar initial in Middle Chinese kixrH and Vietnamese guong is explained by a sound change by which "that uvulars shifted to velars after a tightly attached nonnasal preinitial: *C" (Baxter and Sagart 2014: 101).
Once we become cognizant of the fact that *C stands for any initial in Baxter-Sagart's notation, we notice that *C.q- clusters are reconstructed twice, with different Middle Chinese reflexes. They are reconstructed once with specific preinitials, for which cases the preinitial prevails, and once in *C, for which cases the Middle Chinese reflex is fronted k-. For a hypothetical Old Chinese *s.qraq?-s, the first reconstruction would yield Middle Chinese t§i&rH, while the second would give kixrH. How the two classes should be reconstructed in a Baxter-Sagart-like system remains a problem to be solved.
3.3. ^ *kixr/H 'mirror' as an example of Old Chinese *s-stop cluster
Baxter and Sagart (2014: 101, 168) reconstruct an initial *q- and a qu-tone (*-s) derived from shang-tone (*-?) for ^ *ki&rH < *s.qraq?-s 'mirror', as they consider this word as being in the same family as ^ *?ixrX < *qraq? 'shadow'. The refinement of the reconstruction into *s.qraq?-s enables us to also refine its morphological derivation. Independently of the specific forms reconstructed, ^ *ki&rH 'mirror' is one of the strongest examples of *s-stop clusters in Old Chinese.
A morphological prefix *s- in Old Chinese has been hypothesized since Conrady (1896). As the recent argument between Mei (2012) and Sagart & Baxter (2012) shows, there is broad agreement on the fact that the function of *s- would correspond to that in other Sino-Tibetan languages. However, due to different suppositions on the Middle Chinese reflexes of *s-, there are few words where scholars can actually agree on the reconstruction of *s-.
One of the well-attested functions of *s- is that of deriving an instrumental noun. Rgyal-rongic languages have a sA- prefix16, that notably derives instrumental nouns from the verb roots. In Zbu Rgyalrong, the prefix se- derives from ke-twe? 'to open' the instrumental noun se-thwe? 'key'. The Zbu form belongs to a relic formation (Sun 2004: 293-294, Gong 2018: 211) which shows the age of prefix. Similarly, in written Tibetan, s- can derive instrumental nouns from verb roots: from nod 'to receive', snod 'container, vessel'; from nyan 'to listen', snyan 'ear (honorific)' (cf. Zhang Jichuan 2009: 246, Gong Xun 2017: 154-155, Jacques to appear). Baxter and Sagart (2014: 101) reconstruct *s- with an instrumental value. A good example (Sagart & Baxter 2012: 50-51) is M *d^in 'hoe', reconstructed as *s-l<r>a, derived from ^ *(fia < *l<r>a 'remove', as a hoe is an instrument of removal.
Based on the Old Vietnamese data and the Sino-Tibetan comparative evidence, we can reconstruct the noun ^ *kixrH < *s.qraq?-s 'mirror' as derived from the verb ^ *?ixrH < *qraq?-s 'to reflect', which is itself a denominal from ^ *?ixrX < *qraq? 'shadow > reflection'. It is relevant to note that the formation has a parallel in the Late Middle Chinese/Early Mandarin period. During the Song dynasty, when the syllable kixrH was taboo, being the personal name of the grandfather of the first emperor Zhao Jing (MC *tfieuX kixrH), mirrors were then called Mi zhaozi, MC *tpieuH-tsiX. This formation is entirely parallel to what is proposed here for ^ *kixrH < *s.qraq?-s, as i *zi, MC tsiX was the dominant instrumental noun suffix, added to M zhao, MC *tpieuH 'to shine on', later 'to reflect'.
The word ^ *ki&rH for 'mirror' is a Warring States period (475-221 BC) neologism replacing original ^ *kxmH, coined inside Warring States period Chinese and not inherited from an older state of the language. It can be regarded as one of the safest examples both of a *s-stop cluster in Old Chinese and an example of the instrumental value of the *s- prefix in Old Chinese.
4. *?-: template-filling dummy prefix?
We find two words of Chinese origin showing the preinitial *?-, written ^ (a < a): ^^ *?-lo 'kiln, oven' < ffi *lu 'oven'
(9) « m m tm
*?-lo do doc sot
kiln red poison burning
'red kiln and burning hot poison' (Dai Bao 29a-3,4)
Original: ^ffi^m 'kiln of fire and burning poison'
16 This prefix is, to my knowledge, first described in Lin Xiangrong (1993: 162), cf. also Sun 1998: 142, Yan-muchu 2005, Jacques 2008: 332-333, Lai 2017: 158 etc.
Both Tran and Nguyen (2007, #6) and Hoang Thi Ngo (1999: 161) read (a-lo < *a-luH) as lo 'kiln, oven', Middle Vietnamese lo 'fogao, forno; fornax' (Rhodes 1651: 417). This word is borrowed from Chinese ffi *lu 'oven' < Baxter-Sagart *[r]?a. Note that the word is not from the earliest layer of Chinese loans in Vietnamese, as the initial Middle Chinese *l- < Old Chinese *r-, is reflected as l- in this word but r- in the earliest Chinese loans in Vietnamese.
№h uong 'to chant' < O *t$ha^H 'to chant'
(10) # ^ tieng *2-,uong rang sound chant this 'They chant as follows' (Dai Bao 22b-4,5) Original: KSOW 'They loudly chant as follows'
I follow Shimizu (1996) in reading the Sino-Vietnamese xuong 'to chant'. Hoang Thi Ngo (1999: 158) reads het 'to cry', usually written phonetically as ^ (hat < *hat) or ^ (hiet < *hiat). This reading has no argument in its favour except that xuong is infrequent in contemporary Vietnamese. Concerning this point, cf. xuong ten, 'nomear alguem como por rol; vocare nomen alicuius tanquam e catalogo' (Rhodes 1651: 900), which should be emended into xuong ten 'chant the name-list'. This word is borrowed from Chinese O *t$ha^H 'to chant' < Baxter-Sagart *ma-thaq-s.
Both borrowings show the preinitial *?-, which has not been hypothesized for Old Chinese. We also note that they show signs of rather late borrowing from Chinese: for *?-, uong 'chant', the phonological correspondence is of the newest type, and completely coincides with the Sino-Vietnamese pronunciation. We further notice that in the So dialect of Thavung (Su-wilai 1996), monosyllabic Thai-Lao verbs are borrowed with a prefix ?a-: ld:k 'to drag along the ground' is adapted as ?ald:k. For *?-,uong 'to chant', at least, it is attractive to suppose a similar prefix *?-, a dummy prefix that fills a sesquisyllabic phonological template. The problem of the prefix *?- in Old Vietnamese and other Vietic languages deserves further study.
5. Conclusion
This paper is a preliminary examination of Old Vietnamese loans from Chinese attested with sesquisyllabic orthography in Phat Thuyet Dai Bao Phu Mau An Trong Kinh.
Old Vietnamese data is one of the most eloquent pieces of evidence that argues for the presence and determines the identity of *Ci(a)C2- initial clusters in Old Chinese. It enables the proposal of a few changes to Baxter and Sagart's reconstruction of Old Chinese.
• In two cases, where Baxter and Sagart reconstruct an unknown preinitial *C- based on lenition in Modern Vietnamese, Old Vietnamese enables the identification of the unknown preinitial: ^ tsieqX 'well', Baxter-Sagart *C.tseq?, modified to *k.tseq?; M kixqH 'mirror', Baxter-Sagart *C.qraq?-s, modified to *s.qraq?-s.
• In two cases, where Baxter and Sagart reconstruct Old Chinese forms without preini-tials, Old Vietnamese suggests that a preinitial likely existed in Old Chinese: ^ $ik 'to wipe', Baxter-Sagart *lak, modified to *pa.lak; ^ pen 'edge, side', Baxter-Sagart *p?e[n], modified to *ta.p?e[n].
The Old Vietnamese word *?-lo 'kiln, oven' and №h uong 'to chant' are possible examples of a prefix *?-, internal in Vietic, which is also attested in the conservative Vietic language So Thavung (Suwilai 1996).
The word if *s-kuong < ^ kixr/H 'mirror', which permits a revised Baxter-Sagart reconstruction as *s.qraq?-s, is important in two different ways. First, it points to the fact, hitherto ignored, that *C.q- clusters are reconstructed twice in the Baxter-Sagart system, with different reflexes in Middle Chinese and other sources. Second, this loan in Old Vietnamese provides one of the strongest examples both of an Old Chinese initial consonant cluster of the *s-stop type, and of the Old Chinese prefix *s-, which derives here an instrumental noun from a verb.
References
Alves, Mark J. 2016. Identifying Early Sino-Vietnamese Vocabulary via Linguistic, Historical, Archaeological, and
Ethnological Data. Bulletin of Chinese Linguistics 9: 264-295. Baxter, William H. 1992. A Handbook of Old Chinese Phonology. Berlin: Mouton de Gruyter. Baxter, William H., Laurent Sagart. 2014. Old Chinese: A new reconstruction. Oxford University Press. Baxter, William H., Laurent Sagart. 2017. Old Chinese reconstruction, a response to Schuessler. Diachronica 34(4): 559-576.
Chiang, Chia-lu. 2011. Yuènân hànzîyïn de lîshï céngcî yânjiù (S^^^^^M^S^W^). Doctor thesis, National Taiwan Normal University.
Ferlus, Michel. 1976. Du nouveau sur la spirantisation ancienne en vietnamien. Bulletin de la Société de Linguistique de Paris 71: 305-312.
Ferlus, Michel. 1982. Spirantisation des obstruantes médiales et formation du système consonantique du
vietnamien. Cahiers de linguistique — Asie orientale 11: 83-106. Ferlus, Michel. 1991. Le dialecte vietnamien de Vinh. 24th International Conference on Sino-Tibetan Languages and
Linguistics, October 1991, Bangkok. Available online at: https://halshs.archives-ouvertes.fr/halshs-00922737. Ferlus, Michel. 2007. Lexique de racines Proto Viet-Muong (Proto-Vietic Lexicon). Ms., accessed from the SEAlang Mon-
Khmer Etymological Dictionary <http://sealang.net/monkhmer/dictionary/>. Ferlus, Michel. 2010. Le cycle khmer des douze animaux : histoire d'un contact ancien entre Vietnam et Cambodge.
Cahiers de linguistique - Asie Orientale 39: 3-19. Goad, Heather. 2011. The Representation of sC Clusters. In: Marc van Oostendorp, Colin Ewen, Elizabeth Hume,
Keren Rice (eds.). The Blackwell Companion to Phonology: 898-923. Oxford: Blackwell Publishing. Gong, Xun, Yunfan Lai. 2017. Consonant clusters. In: Rint Sybesma et al. (ed.). Encyclopedia of Chinese language and
linguistics: 665-672. Leiden: Brill. Gong, Xun. 2017. Jiäröng-yuzu yuyan yu Shànggu Hànyu de Xingtàixué (SSrofflroB^iïlroW^i?)- In: Park Haeree, Chen Shaoxuan (eds.). Gûwénzî yü hànyû lîshï bïjiào yïnyùnxué (î^^^^œMÎ fc^^^?) [Paleography and Chinese Historical and Comparative Phonology]: 132-154. Shanghai: Fudan Daxue Chubanshe. Gong, Xun. 2018. Le rgyalrong zbu, une langue tibéto-birmane de Chine du Sud-ouest. Une étude descriptive, typologique et
comparative. Doctoral thesis, INALCO. Haudricourt, André-Georges. 1965. Les mutations consonantiques des occlusives initiales en môn-khmer. Bulletin
de la Société de Linguistique de Paris 60(1): 160-172. Haudricourt, André-Georges. 1974. Hai Chû B, trong cuôn tù diên cûa A-lêch-xan-do Rôt. Ngôn Ngû 1974(4): 37-38. Hayes, La Vaughn H. 1992. Vietic and Viêt-Muong: a new subgrouping in Mon-Khmer. Mon-Khmer Studies 21: 211-228.
Hoàng, Thi Ngo. 1996. Nghiên cûu vê chû Nôm và tiêng Viêt qua bân giâi âm Phât thuyêt dai bâo phu mâu ân trong kinh.
Candidate Thesis. Hanoi: Viên nghiên cuu Han Nôm. Hoàng, Thi Ngo. 1999. Chû Nôm và tiêng Viêt qua bân giâi âm Phât thuyêt Dai bâo phu mâu ân trong kinh. Hanoi: Nxb Khoa hoc Xa hôi.
Jacques, Guillaume (Xiàng, Bölin). 2008. Jiärongyü yânjiû Beijing: Minzu chubanshe.
Jacques, Guillaume. Forthc. Fossil nominalization prefixes in Tibetan and Chinese. To appear in: Bulletin of Chinese Linguistics.
Lai, Yunfan. 2017. Grammaire du khroskyabs de Wobzi. PhD thesis, Université Sorbonne Nouvelle Paris 3. Lin, Xiangrong. 1993. Jiärongyü yânjiû Chengdu: Sichuan minzu chubanshe.
Luo, Zhufeng et al. 1993. Hànyû dà cidiän (Rii^iiÄ). Shanghai: Shanghai cishu chubanshe.
Maspero, Henri. 1912. Études sur la phonétique historique de la langue annamite. Bulletin de l'Ecole française d'Extrême-Orient 12: 1-124.
Maspero, Henri. 1930. Préfixes et dérivation en chinois archaïque. Mémoires de la Société de Linguistique de Paris 23(5): 313-327.
Mei, Tsu-lin. 2012. The Causative *s- and Nominalizing *-s in Old Chinese and Related Matters in Proto-Sino-Tibetan. Language and Linguistics 13(1): 1-28.
Michaud, Alexis, Michel Ferlus, Minh-Châu Nguyên. 2015. Strata of standardization: the Phong Nha dialect of Vietnamese (Quâng Binh Province) in historical perspective. Linguistics of the Tibeto-Burman Area 38(1): 124-162.
Nguyên, Ngoc San. 1982. Gop vài y kiên vê âm dâu tiêng Viêt cô qua cû lieu Nôm trong sach «Phât thuyêt dai bâo phu mâu ân trong kinh». Ngôn ngù 1982(3): 34-41.
Nguyên, Quang Hong. 2001. Khâi luân van tu hoc chù Nôm. Hanoi: NXB Giao duc.
Nguyên, Ta Nhi. 1987. Loi danh dâu ca trong chù Nôm. Tap chi Hân Nôm 1987(1): 35-38.
Nguyên, Tài Cân. 1979. Nguôn gôc và quâ trinh hînh thành câch doc Hân Viêt. Hanoi: NXB Khoa hoc Xâ hôi.
Nguyên, Tài Cân. 2008. Mot sô vân dê vê ngành Nôm hoc. Ms. (Hôi nghi Nôm hoc, April 2008, Temple University).
Nguyên, Phu Phong, Tri Dôi Trân, Michel Ferlus. 1988. Lexique Vietnamien-Ruc-Français: parler d'une minorité ethnique des montagnes de Quâng Bînh, Vietnam avec notes ethnographiques et introduction linguistique. Paris: Sudestasie.
Nguyên, Hùu Vinh, Dâng Thê Kiêt, Nguyên Doân Vuong, Lê Vân Dâng, Nguyên Vân Sâm, Nguyên Ngoc Bich, Trân Uyên Thi. 2009. Tk Diên Chù Nôm Trich Dân: Dictionary of Nôm characters with excerpts. Westminster: Viên Viêt-Hoc.
Pan, Wuyun. 2000. Hànyu lîshïyïnyùnxué (KiiM^^S^). Shanghai: Shanghai jiaoyu chubanshe.
Pulleyblank, Edwin G. 1981. Some notes on Chinese historical phonology. Bulletin de l'École française d'Extrême-Orient 69: 277-288.
De Rhodes, Alexandre. 1651. Dictionarium annamiticum, lusitanum et latinum. Rome: Sacrae Congregationis de Propaganda Fide. Copy held at the Bayerische Staatsbibliothek, accessed at: http://mdz-nbn-resolving.de/ urn:nbn:de:bvb:12-bsb10521857-2.
Sagart, Laurent, William H. Baxter. 2012. Reconstructing the *s- Prefix in Old Chinese. Language and Linguistics 13(1): 29-59.
Shimizu, Masaaki. 1996. Kanbun-Chunomubun taiyaku Bussetsu Daiho Bumo Onjugyo ni miru chunomu ni tsuite = Ningen-kankyogaku 5: 83-104.
Shimizu, Masaaki, Lê Thi Liên, Shiro Momoki. 2005. A Trace of Disyllabicity of Vietnamese in the 14th Century -Chu Nom Characters Contained in the Inscription of Hô Thành Mountain. Kobeshi Gaikokugo Daigaku Gaikokugaku kenkyu 64: 17-49.
Shimizu, Masaaki. 2015. A reconstruction of Ancient Vietnamese initials using Chù Nôm materials. NINJAL research papers 9: 135-158.
Skalbzang 'Gyur med (Gésâng Jumiân), Skalbzang Dbyangs can (Gésâng Yângjïng). 2002. Zàngyu Fàngyân Gàilùn Beijing: Minzu chubanshe.
Sun, Jackson T.-S. 1998. Nominal morphology in Caodeng rGyalrong. Bulletin of the Institute of History and Philology, Academia Sinica 69(1): 103-149.
Sun, Jackson T.-S. 2004. Verb-stem variations in Showu rGyalrong. In: Lin Ying-chin et al. (eds.). Studies on Sino-Tibetan Languages: Papers in Honor of Professor Hwang-Cherng Gong on His Seventieth Birthday: 269-296. Taipei: Academia Sinica.
Suwilai, Premsrirat. 1996. Phonological characteristics of So (Thavung), a Vietic language of Thailand. Mon-Khmer Studies 26: 161-178.
Suwilai, Premsrirat. 2000. So (Thavung) Preliminary Dictionary. Salaya & Melbourne: Institute of Language and Culture for Rural Development, Mahidol University and the University of Melbourne.
Thiêu Chùu. 1942. Hân-Viêt Tu Diên. Hanoi: Duoc Tuê xuât bân.
Thompson, Laurence C. 1976. Proto-Viet-Muong Phonology. In: Philip N. Jenner and Laurence C. Thompson and Stanley Starosta (eds.). Austroasiatic Studies: 1113-1204. Honolulu: The University Press of Hawaii.
Trân, Trong Duong. 2010. Vê dâu vêt chù Nôm ky huy trong sâch 'Phât thuyêt'. Ms. (Tap chi Vân hoa Nghê An 7).
Trân, Trong Duong. 2011. Phât thuyêt co phâi dich phâm Nôm cùa thê ki XII? Ngôn ngù 2011(4): 31-47.
Trân, Trong Duong. 2012a. Mot sô tk gôc Hân co câu truc CCVC qua ngù liêu tha Nôm trong Quôc âm thi tâp. Presentation made at the Hôi thâo ky niêm 1 nâm ngày mât GS Nguyên Tài Cân.
Trân Trong Duong. 2012b, Tù nguyên cùa XE và cac diêp thùc cùa no. In: Thông bâo Hân Nôm hoc nam 2010-2011: 557-562. Hanoi: Thê gioi.
Trân, Uyên Thi, Hùu Vinh Nguyên. 2007. Ai vë âuac, ai xoâ âwçc? Dâu vêt âm viêt cô: tù song tiêt và phu âm kép.
Presentation made at the Hôi Nghi Quôc Tê vê Tiêng Viêt, Viên Viêt-Hoc, Westminster. Wang, Li. 1948. Hànyuèyu yanjiù Lingnan xuebao 9(1): 1-96.
Yânmùchù. 2005. Shilùn Jiârôngyu zhong sa de yufâ gongnéng ji xingshi (iifà^^in^sa Aba
Shifan Gaodeng Zhuanke Xuexiao Xuebao 22(1): 91-92. Zhang, Jichuan. 2009. Zàngyu cizu yânjiù: gudài zàngzu ruhé fêngfù fàzhan tàmende cihuî "fritMR^i^g^MffiWiiC). Beijing: Shehuikexue Wenxian Chubanshe.
Гун Сюнь. Китайские заимствования с полуторасложной структурой в старовьетнамском языке
В настоящее время практически не остается сомнений в том, что в древнекитайском языке присутствовали начальные сочетания согласных, включавшие в себя пресилла-бы; однако прямых указаний на то, какие конкретные формы реконструировать для тех или иных слов, сохранилось очень немного. В данной статье вводится в научный обиход новый источник, ранее остававшийся неисследованным — старовьетнамский язык, достаточно подробно зафиксированный в одном сохранившемся тексте, где ряд слов, которые в современном вьетнамском языке являются односложными, записан в орфографии, предполагающей для них изначальный полуторасложный характер. Для некоторых слов, заимствованных из китайского, только варианты их записи в старовьетнамской орфографии позволяют определить изначальную форму, в которой они были заимствованы. Небольшой список известных на сегодняшний день полуторасложных слов, зафиксированных в старовьетнамском тексте и имеющих китайское происхождение, включает как лексемы, для которых можно надежно реконструировать в китайском сочетания согласных, так и слова, где вероятна префиксация уже на вьетнамской почве.
Ключевые слова: древнекитайский язык, старовьетнамский язык, историческая реконструкция, полуторасложные слова, префиксальная морфология.
