The genetic position of Chinese

Encyclopedia of Chinese Language and Linguistics Volume 2 De–Med For use by the Author only | © 2017 Koninklijke Brill NV General Editor Rint Sybesma (Leiden University) Associate Editors Wolfgang Behr (University of Zurich) Yueguo Gu (Chinese Academy of Social Sciences) Zev Handel (University of Washington) C.-T. James Huang (Harvard University) James Myers (National Chung Cheng University) For use by the Author only | © 2017 Koninklijke Brill NV ENCYCLOPEDIA OF CHINESE LANGUAGE AND LINGUISTICS Volume 2 De–Med General Editor Rint Sybesma Associate Editors Wolfgang Behr Yueguo Gu Zev Handel C.-T. James Huang James Myers LEIDEN • BOSTON 2017 For use by the Author only | © 2017 Koninklijke Brill NV Typeface for the Latin, Greek, and Cyrillic scripts: “Brill”. See and download: brill.com/brill-typeface. ISBN 978-90-04-18643-9 (hardback, set) ISBN 978-90-04-26227-0 (hardback, vol. 1) ISBN 978-90-04-26223-2 (hardback, vol. 2) ISBN 978-90-04-26224-9 (hardback, vol. 3) ISBN 978-90-04-26225-6 (hardback, vol. 4) ISBN 978-90-04-26226-3 (hardback, vol. 5) Copyright 2017 by Koninklijke Brill NV, Leiden, The Netherlands. Koninklijke Brill NV incorporates the imprints Brill, Brill Nijhoff, Global Oriental and Hotei Publishing. All rights reserved. No part of this publication may be reproduced, translated, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission from the publisher. Authorization to photocopy items for internal or personal use is granted by Koninklijke Brill NV provided that the appropriate fees are paid directly to The Copyright Clearance Center, 222 Rosewood Drive, Suite 910, Danvers, MA 01923, USA. Fees are subject to change. This book is printed on acid-free paper and produced in a sustainable manner. For use by the Author only | © 2017 Koninklijke Brill NV 297 Genetic Position of Chinese Wilhelm, Andrea, “Bare Nouns and Number in Dëne Sųłiné”, Natural Language Semantics 16/1, 2008, 39–68. Yang, Rong, “Chinese Bare Nouns as Kind-denoting Terms”, Ruling Papers 1, 1998, 247–288. Jo-Wang Lin Genetic Position of Chinese 1. In t roduct i on Although no linguist has ever claimed that Chinese was a language isolate, its exact relationship was and is still controversial. The main reason behind this controversy is the limited quantity of morphology in Chinese. As Meillet put it: si l’on est en présence de langues qui n’ont presque pas de grammaire, si presque toute la grammaire proprement dite tient en quelques règles de position relative des mots, comme dans certaines langues d’Extrême-Orient ou du Soudan . . . alors la question des parentés de langues est pratiquement insoluble, aussi longtemps qu’on n’aura pas prouvé de critères qui permettent d’afijirmer que les langues de ce type sont issues les unes des autres et que les ressemblances de vocabulaire qu’elles offrent ne sont pas dues à des emprunts. In the case of languages which have almost no grammar, if almost all the grammar properly speaking is limited to some rules of relative ordering of words, as in certain languages of the Far East or Sudan . . . the question of genetic relatedness is practically insoluble, as long as we have no criteria permitting us to conclude that languages of this type are originated from a common origin and that the lexical similarities that they show are not due to borrowing. (1982:97) Although some morphology is reconstructed for Old Chinese (see section 6), the absence of irregular flexional paradigms, which are viewed as the most reliable type of evidence for genetic relationship, is a serious impediment against proving a genetic relationship between Chinese and any other language, especially if the other languages in question have a limited morphological system too. Reconstructible afijixes are typically limited to one consonant belonging to a limited set (only nasals, s, and voiceless unaspirated stops), so that resemblances with other languages could easily be coincidental. As Meillet pointed out, similarities in vocabulary have limited value by themselves, unless one can distinguish inherited vocabulary from borrowings. Up to now, there is no universally accepted theory regarding the genetic status of Chinese. Chinese is generally recognized to belong to a Sino-Tibetan family, but the exact subgrouping and extension of this family is not agreed on by all scholars. The three currently accepted models are Greater Sino-Tibetan (including Chinese, Tibeto-Burman, Kra-Dai, and Hmong-Mien), Bifurcate Sino-Tibetan (with two main branches, Chinese and Tibeto-Burman, and excluding Kra-Dai and Hmong-Mien), and Multifurcate Sino-Tibetan. Additionally, several superfamilies including Chinese have been proposed, in particular SinoTibeto-Austronesian and Sino-Caucasian. 2 . Gr e a t e r S i n o - T i b e t a n o r In d o - C h i n e s e This theory, still widely accepted, especially by mainland Chinese scholars (for instance Xíng 2000), views Chinese as one of the main Sino-Tibetan Chinese Tibeto-Burman Kra-Dai Hmong-Mien Figure 1 For use by the Author only | © 2017 Koninklijke Brill NV Genetic Position of Chinese subgroups of a family including Tibeto-Burman, Kra-Dai (→ Tai-Kadai), and → Hmong-Mien (Miáo-Yáo 苗瑤). As Van Driem (2005) has shown, it originates from Leyden’s (1808) antiquated “Indo-Chinese” model of classiijication of East Asian languages, but its prominence is due to Li’s (1937) influential article on the classiijication of the languages of China. In this theory, the isolating, tonal, and monosyllabic character of Chinese, Kra-Dai, HmongMien, and some Tibeto-Burman languages such as Lolo-Burmese is an archaic feature going back to proto-Sino-Tibetan, and → TibetoBurman languages that depart from this model (such as → Tibetan, → Rgyalrong, or Kiranti) are innovative. Related words can be found in all four groups of languages, including basic items such as numerals, some body parts, and some common verbs. The Greater Sino-Tibetan hypothesis views this common vocabulary as being common inheritance, though it acknowledges the existence of marginal loanwords in some cases. The weakness of this theory is the fact that, as Meillet pointed out long ago, the isolating and monosyllabic character of these languages is no proof of their relatedness, and that the common vocabulary, even if it includes basic items, could be due to borrowing. Haudricourt’s (1954a and 1954b) articles showed that the tonal systems of Vietnamese and Chinese were secondary developments, and therefore that the presence or absence of tone was irrelevant for proving language relationships. Besides, the highly opaque and irregular morphology of languages like Tibetan is not likely to be a recent development out of an earlier isolating stage, otherwise we would expect to ijind strictly regular paradigms. In some cases, we have objective evidence that some of the common vocabulary among Chinese, Kra-Dai, and Hmong-Mien is the result of borrowing. See Table 2. 298 In the ijirst two examples, we observe that the initial consonant found in Thai languages is closer to the Middle Chinese form than to Old Chinese. It indicates that the Chinese word was borrowed probably even later than the Hàn dynasty. As for the word ‘iron’, the Tai form resembles Old Chinese more than Middle Chinese, and this word would appear to be a good candidate for a cognate, as Chang (1972) has proposed. However, given the fact that iron melting did not exist before the 6th century BCE in East Asia (Wagner 1993), it is unlikely that the word ‘iron’ could have existed in the common ancestor of Chinese and Tai. The example proves that even words whose phonetic form is closer to Old Chinese than to Middle Chinese in Tai languages can be borrowings. Incidentally, the Tibetan form for ‘iron’ is lcags < *lhjaks, showing a case of (probably indirect) contact between Old Chinese and proto-Tibetan. 3 . B i fur c a t e S i n o - T i b e t a n An alternative influential model is Benedict (1972) and Matisoff ’s (2003) view of SinoTibetan (see Figure 3), which excludes Kra-Dai and Hmong-Mien, considering their common vocabulary as the exclusive result of borrowing. Benedict explicitly takes Chinese and Karen apart from the other languages, mainly because of their SVO word order, which differs from the more common verb-ijinal word order of other languages and families such as Tibetan and Lolo-Burmese. Matisoff reintegrates Karen within his “Tibeto-Burman” subgroup, but still gives Chinese a pre-eminent place as the ijirst branch off of the Sino-Tibetan family. Bifurcate Sino-Tibetan is the theory accepted by most specialists of Sino-Tibetan historical linguistics, for instance Bradley (1997), Sagart (1999), and Thurgood (2003). Table 2 ten bronze iron ิบ ทอง เ ล็ก proto-Tai (Li 1977) Middle Chinese Old Chinese *sip D *dɔːŋ A *hlek D dʑip duwŋ thet *gip *lˁoŋ *lhˁik For use by the Author only | © 2017 Koninklijke Brill NV 299 Genetic Position of Chinese Surprisingly, as Handel (2008) noticed, the bifurcate Stammbaum of Sino-Tibetan is taken for granted by these scholars, and few have attempted to prove the existence of a “TibetoBurman” node. The number of cognates between Chinese and other languages is indeed limited; for instance, Jacques (2005) only found 94 (by contrast, Japhug has about 120 cognates with Tibetan and more than 300 with Tangut), of which only 15 belong to Swadesh’s 100 basic word list. However, this is by itself no proof that Chinese is the ijirst branch of the family. Some languages such as Sulung appear to share even fewer cognates with the rest of Sino-Tibetan than Chinese, and the lack of cognates could be due to lexical innovations in Chinese that replaced much of the original inherited vocabulary. Only a set of common innovations, whether phonological, lexical or morphological, could prove that all languages apart from Chinese belong to a common node. Matisoff and Benedict’s reliance on word order as proof of the special status of Chinese is not compatible with the oft-repeated (but poorly supported: see Djamouri et al. 2007 for a conflicting viewpoint) idea that proto-SinoTibetan was SOV: if non-Chinese languages have preserved the proto-Sino-Tibetan word order, and only Chinese has changed, there is no basis to argue that they form a coherent subgroup. In any case, typological features like word order are not easily reconstructible and constitute weak evidence for subgrouping. Handel (2008) cites only one example of a Tibeto-Burman innovation: the merger of protoSino-Tibetan *a and *ə everywhere but in Chinese, as suggested by Gong’s (1995) comparative work. If veriijied, this would be an important piece of evidence in favor of the bifurcate SinoTibetan hypothesis. However, there is clear evidence for a different treatment of the vowels corresponding to *a and *ə in some non-Chinese languages. For instance, in → Tangut (See Table 4). Old Chinese *a typically corresponds to Tangut -ji or -e (as do Japhug and Tibetan -a), but OC *ə corresponds to Tangut -u, even though Japhug and Tibetan do not appear to have preserved this contrast. The only counterexample 395 ŋwe 牽 2.07 ‘cow’ corresponding to Chinese *ŋˁwə 牛 is probably not a real cognate, since cattle herding is late in East Asia, appearing after 2000 BCE, a date too late to be contemporaneous with proto-Sino-Tibetan. Alternatively, one could explain this example by supposing that *a and *ə merge in Tangut after labiovelars. The contrast left no trace in closed syllables, where Chinese *əC and *aC correspond to the same sets of Tangut rhymes. Detailed investigation of other non-Chinese Sino-Tibetan languages may reveal similar traces of the contrast. The Chinese-centered way of approaching the classiijication of Sino-Tibetan is not immune to bias; since Chinese is better known that any other language of the family, it is easier to notice distinctions in Chinese not found in other languages. This bias can be illustrated by noting that, from a Rgyalrong-centered point of view, one could propose an analogous (and clearly incorrect) alternative theory. To see this, consider that Rgyalrong languages are known to have a contrast between plain and velarized vowels, which is best preserved in Zbu (Sun 2000; → Rgyalrong). This contrast can clearly be reconstructed to proto-Rgyalrongic, but cannot be clearly explained as coming from lost clusters or ijinal consonants, since Rgyalrong languages are phonologically quite conservative. The only language outside of Rgyalrong that shows a trace Table 4 Reference Tangut 5203 4046 4681 1338 躅 瑩 艘 側 .wji 1.67 khie 1.09 nju 1.03 dzu 1.01 Meaning Japhug axe bitter ear love tɯ-rpa tɯ-rna Tibetan OC kha rna mdza 斧 *pˁaʔ 苦 *kʰˁaʔ 耳 *nəʔ 慈 *dzˁə For use by the Author only | © 2017 Koninklijke Brill NV Genetic Position of Chinese 300 Sino-Tibetan Tibeto-Burman Chinese Baic Kamarupan Karenic Himalayish Qiangic Jingpo-NungishLuish Lolo-BurmeseNaxi Figure 3. Matisoff ’s (2003:5) Sino-Tibetan Stammbaum. of the contrast is again Tangut, as the examples in Table 5 show. Tangut has -o or -ow corresponding to protoJaphug *-am and *-aŋ and -a corresponding the rhymes *-aˠm and *-aˠŋ with velarized vowel. No other language seems to show the same split. One could use here the same argument as for *a and *ə, and suggest that Rgyalrong and Tangut form the ijirst branch of Sino-Tibetan, while lumping all other languages in a subgroup. This is obviously wrong for other reasons, but it illustrates the fact that one or two proposed innovations are of limited value for establishing a subgrouping. Even if we cannot exclude the possibility that the bifurcate Sino-Tibetan hypothesis might be shown to be valid eventually, accepting it as proven without evidence is highly detrimental to comparative Sino-Tibetan. It implies that only features found in both Chinese and in some other language can be reconstructed for protoSino-Tibetan, thus over-valuing Chinese data and discouraging comparativists from the study of more conservative (and more endangered) languages. In such a research paradigm, any feature common to Tibetan, Rgyalrong, and Kiranti would be deemed irrelevant to the reconstruction of Sino-Tibetan, being viewed as merely a “Tibeto-Burman” innovation. 4 . Mult ifurcat e Si no-Ti b eta n Multifurcate Sino-Tibetan is a more agnostic vision of the subgrouping of this family, proposed by many scholars from Klaproth (1823) to Van Driem (2005). This is not an actual model of subgrouping but simply an acknowledgment of our ignorance of the Stammbaum of SinoTibetan and a decision to postpone language classiijication until the languages are better described and their historical phonology better analyzed. Van Driem proposes about 30 basic branches in Sino-Tibetan that can be argued to be monophyletic (i.e., to have an exclusive single common ancestor). It is of course unlikely that Sino-Tibetan had so many as 30 primary branches, but the sorry state of our knowledge of Sino-Tibetan historical phonology does not allow for a more reijined phylogeny for the time being. Alternative views on Sino-Tibetan classiijication such as Sino-Bodic (Van Driem 1997) are not strongly supported by the evidence at hand but still deserve as much potential consideration from Sino-Tibetan specialists as the bifurcate Sino-Tibetan theory. Postponing language classiijication is not necessarily an obstacle to studying the correspondences between languages or even reconstructing proto-Sino-Tibetan. Language families such as Indo-European and Algonquian have been successfully reconstructed without focus on language classiijication, and the exact Stammbaum of these families is still controversial among specialists (see Gamkrelidze et al. 1990 and Gray et al. 2003 concerning Indo-European and Goddard 1979 and Proulx 1980 concerning Algonquian). For use by the Author only | © 2017 Koninklijke Brill NV 301 Genetic Position of Chinese Table 5 Reference Tangut 1099 2584 3443 39 1391 5528 975 壬 兢 數 慣 胎 蝦 廠 dow 2.47 dzow 1.54 po 1.49 kowr 2.82 ba 1.17 bar 1.80 par 1.80 Meaning Japhug bear bridge uncle tooth deaf drum to freeze ndzom tɤ-βɣo tɤ-mɢom tɤ-mbɣo tɤ-rmbɣo jpɣom 5. M ixed L ang u a ge Hy p oth esi s Aside from the three major hypotheses presented above, some scholars have proposed that Chinese was to some extent a mixed language whose vocabulary was partially borrowed from non-Sino-Tibetan languages. Benedict (1972:197) even claimed that “the Chou people might be regarded as the bearers of a S[ino-]T[ibetan] language, which became fused with, or perhaps immersed in, a non-ST language spoken by the Shang people”. Although such a strong hypothesis seems highly unlikely in view of the continuity between → Shāng 商 and Zhōu 周 Chinese (the political transition occurred in the 11th century BCE), the idea of Chinese as a mixed language is still influential. Schuessler (2007:5), in particular, argues for a large Austroasiatic substratum in Old Chinese and Starostin (2008) proposed that many “Altaic” loanwords can be detected in Old Chinese. (Independently of the validity of the Altaic family, some of his comparisons involve Chinese words that have potential cognates in the other ST languages. For instance, *tsrˁuʔ 爪 ‘claw’ is better compared with Japhug tɯ-ndzrɯ ‘claw, nail’ rather than with Turkic or Tungusic.) 6 . T h e M orp hol ogi ca l E vi d ence f or a Sino-T ib eta n F a m i l y The exclusive evidence on basic vocabulary as a proof of genetic relationship is problematic, as suggested by Meillet, and morphology is generally considered to be a better proof of genetic relatedness, even though the shortness of grammatical elements increases the risk of OC 熊 *ɢwəm < *ndzam < *-kpaŋ < *-mɴɢam < *-mbaˠŋ < *-mbaˠŋ < *lpaˠm 冰 *prəŋ accidental similarities (cf. Callaghan 1977, who pointed out the fortuitous resemblance between some Miwok and Indo-European inflectional paradigms). The reconstructible morphological system of Old Chinese is limited to a few derivational afijixes (Sagart 1999; Pulleyblank 2000), not all of which are equally well established—indeed, some are highly controversial. Nevertheless, this remnant morphology is of critical importance for assessing the genetic position of Chinese. One of the most interesting morphological processes found in Chinese is the alternation between Middle Chinese voiced and voiceless stops, where the voiced counterpart is an intransitive verb, and the voiceless one a transitive verb. This process has been studied by many scholars such as Downer (1959:263), Zhōu (1962) and more recently Sagart (2003). The list in Table 6, taken from Sagart (2003), illustrates this phenomenon (the correspondence of k- to the voiced fricative ɣ- instead of g- is expected: g- and ɣ- occur in complementary distribution in Middle Chinese). (I present here an IPA interpretation of Baxter’s typeable MC transcription.) The voicing alternation observed here is wellknown in other Sino-Tibetan languages, as has been noticed by August Conrady. The only modern language group in which this process in still productive is Rgyalrong: in Japhug Rgyalrong, it applies to at least one loanword χtɤr from Tibetan gtor (see Jacques 2011). The prenasalization derives anticausative (not passive) verbs, expressing an action taking place spontaneously without an agent. Most For use by the Author only | © 2017 Koninklijke Brill NV genetic position of chinese 302 Table 6 Middle Chinese meaning (tr.) Middle Chinese meaning (it.) pjet 别 pæjH 敗 twanH 斷 tɕet 折 kenH 見 kejH 繫 to separate to defeat to cut to break, bend to see to attach bjet 别 bæjH 敗 dwanH 斷 dʑet 折 ɣenH 現 ɣejH 系 to take leave to be defeated to be cut to be bent to appear to be attached transitive meaning intransitive meaning ftʂi prɤt qɤt χtɤr melt (tr) break, cut separate scatter ndʐi mbrɤt ɴɢɤt ʁndɤr melt (it) be broken, cut to be separated be scattered Table 7 Sino-Tibetan languages have some trace of this process, though in some languages phonetic changes have rendered it difijicult to distinguish from the effects of the causative preijix. Few of the Chinese pairs have direct cognates in the other Tibeto-Burman languages. In Tibetan, for instance, one ijinds only the pair N-bye-d (past phye) ‘to open, to separate’ vs. N-bye (past bye) ‘to be opened, to get separated’ potentially cognate to pjet / bjet 别 (though the ijinal -t in Chinese is not explained), and a few examples like N-khyig (past b-kyig-s) ‘to tie’ cognate to kejH 繫 without intransitive counterpart. A phenomenon comparable to anticausative prenasalization is also found in Hmong-Mien (Downer 1973). However, by contrast to “TibetoBurman” languages, where few pairs of verbs related to Chinese can be found, all of Downer’s examples are related to Chinese (Sagart 2003). For those who do not subscribe to the idea that Hmong-Mien belongs to Sino-Tibetan (almost all Western linguists), this implies that anticausative prenasalization never was a productive process in Hmong-Mien, and that what appears to be a morphological alternation is a mere artefact: verbs have been borrowed in pairs. Note that the anticausative alternation in non-Chinese Tibeto-Burman languages cannot be argued to have been borrowed from Chinese (or vice-versa), since there is almost no lexical overlap. Sagart (2003) suggests that the voicing alternation of Middle Chinese originates from a nasal preijix *N- in Old Chinese. This idea is based on the fact that according to Downer, the voicing alternation in Hmong-Mien goes back to a nasal preijix in proto-Hmong-Mien that voiced the initial stops. According to Sagart, in Chinese, when preijixed to a voiceless unaspirated stop, *N- causes it to voice, becoming a Middle Chinese voiced stop; when preijixed to an aspirated stop, it leaves no trace in Middle Chinese, but some examples in Hmong-Mien show the presence of the preijix with aspirated initials. Aside from anticausative prenasalization, we do ijind in Chinese several afijixes that have potential cognates in the rest of Sino-Tibetan: the ubiquitous Sino-Tibetan causative *s- (Mei 1989), the denominative *s-, the oblique nominalization *s- (cf. Japhug Rgyalrong sɤ-), the nominalization sufijix *-s, perhaps also the applicative *-t (see Sagart 2004:73). Most of the traces of morphology in Chinese are, however, difijicult to interpret and quite controversial (see for instance Branner 2002). Chinese does not preserve clear evidence of irregular flexional morphology, unlike other Sino-Tibetan languages, where we ijind common irregular pronominal paradigms (Jacques 2007) For use by the Author only | © 2017 Koninklijke Brill NV 303 genetic position of chinese and verbal stem alternations (Jacques 2010) in several unrelated branches. A worthwhile direction for future research, however, would be to investigate whether the traces of vowel ablaut in Old Chinese (Behr 1998) could be relatable to stem alternation in Rgyalrongic languages (Sun 2000; Sun 2004). A few scholars still doubt the validity of SinoTibetan, such as Roy Andrew Miller and Christopher Beckwith; however, the morphological commonalities between Chinese and the other languages, especially the anticausative prenasalization, make it highly implausible that such a relationship could be explained exclusively as resulting from contact. 7. M acrocompa r i sons Chinese and other Sino-Tibetan languages have not been neglected in the quest for superfamilies. Sino-Tibetan has been proposed to be related to Indo-European (mainly based on putative morphological resemblances; cf. Pulleyblank 1965), Ienissean, North-West Caucasian, and Nakh-Daghestanian (Starostin 1984; SinoCaucasian); even additionally in some versions Athabaskan, Burushaski and Basque (Dene-Caucasian), Austronesian (Sagart 2005, Sino-Austronesian), Mayan languages, and other attempts that barely deserve to be mentioned. We will discuss here only two of these proposed superfamilies: Sino-Austronesian and Sino-Caucasian. 7.1. Sino-Austronesian According to the Sino-Austronesian hypothesis in its current deijinition (Sagart 2005), SinoTibetan and Austronesian share a common ancestor. The Kra-Dai family, excluded from Sino-Tibetan, is considered to be a heavily relexiijied subbranch of Austronesian; Hmong-Mien and other families of East Asia are regarded as being unrelated to Sino-Tibetan. This hypothesis is based on about 60 lexical comparisons that follow regular correspondences, of which ten belong to Swadesh’s basic vocabulary list. The morphological evidence is not lacking either, such as the Austronesian goal focus marker *-en compared to the Sino-Tibetan nominalizing sufijix -n. However, since no irregular morphology is reconstructed in protoAustronesian, the comparisons exclusively rest on regular afijixal patterns; no common irregular alternations can be found between Sino-Tibetan and Austronesian. Sino-Austronesian started as a direct comparison between Chinese and proto-Austronesian, and non-Chinese Sino-Tibetan languages have been only marginally investigated. If SinoAustronesian is valid, we should ijind considerably more common vocabulary and afijixes between these languages and Austronesian. Grander classiijication schemes based on Sino-Austronesian have been proposed, such as Starosta’s (2005) Sino-Tibetan-Yangtzean that includes Sino-Tibetan, Austronesian, KraDai, Hmong-Mien, and Austroasiatic, but these hypotheses have not yet been thoroughly investigated. Opponents of the Sino-Austronesian hypothesis generally accept the validity of some of the Chinese-Austronesian comparisons, but interpret them as resulting from contact rather than from genetic inheritance (Peiros et al. 1984). 7.2. Sino-Caucasian The idea of a relationship among Sino-Tibetan, Ienissean, Burushaski, and the two northern Caucasian families (North-West Caucasian and Nakh-Daghestanian) was proposed by Starostin (1984; 2004). Starostin compares his reconstructed proto-Sino-Tibetan (Peiros et al. 1996) with his proto-Ienissean and proto-North Caucasian, itself a controversial proto-family, since the relationship between North-West Caucasian and Nakh-Daghestanian is not considered to be proven by most specialists. In view of the complex morphology that all ijive families exhibit (if, in the case of SinoTibetan, one takes Rgyalrong rather than Chinese or Lolo-Burmese as representative), it comes as a surprise that comparative morphology is relatively absent from Starostin’s work: we ijind no comparison of irregular paradigms between the languages under investigation. If a deep genetic connection exists among For use by the Author only | © 2017 Koninklijke Brill NV genetic position of chinese Sino-Tibetan, Ienissean, and Caucasian languages, we would not necessarily expect to ijind hundreds of common lexical items (we ijind much fewer valid cognates between Chinese and Rgyalrong for instance), but morphological evidence should be present. As a comparison with another superfamily, the very promising SiouanYuchi hypothesis (Rankin 1998) mainly rests on common irregular paradigms, and barely more than 20 convincing etymologies can be found between Yuchi and Siouan-Catawban. Starostin’s work certainly deserves to be carefully reviewed by experts on all ijive language families; his Sino-Tibetan reconstruction, though not free from defects, constitutes the best available database on comparative Sino-Tibetan. At the same time, Starostin adopts a very mechanistic approach to language comparison, whereby a proto-phoneme is posited for nearly every correspondence, resulting in the proliferation of arcane proto-symbols. The comparisons between Sino-Tibetan and proto-North Caucasian are based on only one or two common phonemes per etymon, and it remains to be demonstrated whether the similarities detected by Starostin are anything more than change resemblances. Besides, the hypothesis of a common ancestor between Sino-Tibetan and North Caucasian beg the questions of the Urheimat of this hypothetical proto-language and of the historical scenario needed to explain the spread of this putative family. Finally, since Rgyalrong languages and Ienissean appear to be the only two groups of languages in Eurasia to share the typological feature of being verb ijinal with a mainly preijixing morphology, it is intriguing that this fact (and the potential resemblances of the verbal preijixes of Rgyalrong and Ienissean) has played no role in Starostin’s hypothesis. 8 . Conclusion The evidence in favor of the inclusion of Chinese within a Sino-Tibetan family excluding Kra-Dai and Hmong-Mien is overwhelming. The exact subgrouping of this Sino-Tibetan family however is very controversial, and it is unlikely that a satisfactory answer will appear, until the historical 304 phonology and morphology of the main SinoTibetan languages has been better investigated. The current underdeveloped state of comparative Sino-Tibetan is due to various factors, especially language contact, which has blurred the distinction between cognates and loanwords, as well as poor understanding of the morphology. The study of comparative Sino-Tibetan morphology has been hampered by the fact that the branches that are best known, Lolo-Burmese and Chinese, have relatively simple morphology. A better integration of conservative SinoTibetan languages such as Rgyalrong, Tibetan, Dulong/Rawang, and Kiranti in the Sino-Tibetan reconstruction model is likely to solve longstanding problems regarding word families and irregular correspondences. Rgyalrong and Kiranti present an important quantity of common inflexional and derivational morphology (Jacques 2012), and since these two groups are considerably divergent lexically and located far away from each other, it is probable that this morphology represents proto-Sino-Tibetan inheritance. In this view, Chinese (and many other Sino-Tibetan languages) would have lost their formerly complex verbal systems; DeLancey (2010) presents an account of how such a drastic change could have taken place. Larger superfamilies such as Sino-Austronesian and Sino-Caucasian deserve to be investigated and reevaluated as new data from recently described Sino-Tibetan languages becomes available. If these hypotheses are valid, additional supporting evidence should come to light in these previously unknown languages. Bibliography Behr, Wolfgang, “Preliminary Notes on the Forms of Old Chinese Qualitative Ablaut”, paper presented at the Troisième Colloque International sur la Grammaire du Chinois Ancien, Centre de Recherches Linguistiques sur l’Asie Oriëntale, Paris: École des hautes études en sciences sociales, 1998. Benedict, Paul K., Sino-Tibetan: A Conspectus, Cambridge: Cambridge University Press, 1972. Bradley, David, “Tibeto-Burman Languages and Classiijication: Tibeto-Burman Languages of the Himalayas”, in: David Bradley, ed., Papers in Southeast Asian Linguistics 14, Canberra: Paciijic Linguistics, 1997, 1–72. For use by the Author only | © 2017 Koninklijke Brill NV 305 genetic position of chinese Branner, David Prager, “Common Chinese and Early Chinese Morphology”, Journal of the American Oriental Society 122/4, 2003, 706–721. Callaghan, Catherine A., “An ‘Indo-European’ Type Paradigm in Proto Eastern Miwok”, in: Katheryn Klar, Margaret Langdon and Shirley Silver, eds., American Indian and Indoeuropean Studies: Papers in Honor of Madison S. Beeler, The Hague: Mouton, 1980, 31–41. Chang Kun, “Sino-Tibetan ‘Iron’: *Qhleks”, Journal of the American Oriental Society 92/3, 1972, 436–446. DeLancey, Scott, “Language Replacement and the Spread of Tibeto-Burman”, Journal of the Southeast Asian Linguistics Society 3/1, 2010, 40–55. Djamouri, Redouane, Paul Waltraud and John Whitman, “Reconstructing VO Constituent Order for Proto-Sino-Tibetan”, paper presented at the 18th International Conference on Historical Linguistics, Montréal, UQAM, 2007. Downer, Gordon B., “Derivation by Tone-Change in Classical Chinese”, Bulletin of the School of Oriental and African Studies 22/2, 1959, 259–290. Downer, Gordon B., “Strata of Chinese Loanwords in the Mien Dialect of Yao”, Asia Major 18/1, 1973, 1–33. Driem, George van, “Sino-Bodic”, Bulletin of the School of Oriental and African Studies 60/3, 1997, 455–488. Driem, George van, “Tibeto-Burman vs Indo-Chinese: Implication for Population Geneticists, Archeologists and Prehistorians”, in: Laurent Sagart, Roger Blench and Alicia Sanchez-Mazas, eds., The Peopling of East Asia: Putting Together Archaeology, Linguistics and Genetics, New York: RoutledgeCurzon, 2005, 81–106. Gamkrelidze, Tamaz V. and Vjacheslav V. Ivanov, Indo-European and the Indo-Europeans, The Hague: Mouton de Gruyter, 1995. Goddard, Ives, “The Evidence for Eastern Algonquian as a Genetic Subgroup”, Algonquian Linguistics 5, 1979, 19–22. Gong, Hwang-cherng, “The System of Finals in ProtoSino-Tibetan: The Ancestry of the Chinese Language”, Collected Papers on Sino-Tibetan Linguistics, Taipei: Academia Sinica, 1995, 79–124. Gray, Russell D. and Quentin D. Atkinson, “LanguageTree Divergence Times Support the Anatolian Theory of Indo-European Origin”, Nature 426, 2003, 435–439. Handel, Zev, “What is Sino-Tibetan? Snapshot of a Field and a Language Family in Flux”, Language and Linguistics Compass 2/3, 2008, 422–441. Haudricourt, André-Georges, “De l’origine des tons en vietnamien” [On the origin of Vietnamese tones], Journal Asiatique 242, 1954a, 69–82. Haudricourt, André-Georges, “Comment reconstruire le chinois archaïque” [How to reconstruct Archaic Chinese], Word 10/2–3, 1954b, 351–364. Jacques, Guillaume, “Jiāróngyǔ yǔ shànggǔ Hànyǔ 嘉绒语与上古汉语” [Rgyalrong and Old Chinese], paper presented at the International Conference on Old Chinese Phonology, Fùdān Dàxué 复旦 大学, 2005. Jacques, Guillaume, “A Shared Suppletive Pattern in the Pronominal Systems of Chang Naga and Southern Qiang”, Cahiers de Linguistique Asie Orientale 36/1, 2007, 61–78. Jacques, Guillaume, “A Possible Trace of Verb Agreement in Tibetan”, Himalayan Linguistics Journal 9/1, 2010, 41–49. Jacques, Guillaume, “Argument Demotion in Japhug Rgyalrong”, in: Katharina Haude and Gilles Authier, eds., Passive and Ergativity, The Hague: Mouton de Gruyter, 2011. Jacques, Guillaume, “Agreement Morphology: The Case of Rgyalrongic and Kiranti”, Language and Linguistics 13/1, 2012, 83–116. Klaproth, Julius, Asia polyglotta, Paris: A. Schubart, 1823. Li, Fang-kuei, “Languages and Dialects of China”, Journal of Chinese Linguistics 1/1, 1973, 1–13. Li, Fang-kuei, Handbook of Comparative Tai, Honolulu: University Press of Hawai’i, 1977. Leyden, John, “On the Languages and Literature of the Indo-Chinese Nations”, Asiatic Researches 10, 1808, 158–289. Matisoff, James A., Handbook of Proto-Tibeto-Burman: System and Philosophy of Sino-Tibetan Reconstruction, Berkeley: University of California Press, 2003. Mei, Tsu-lin, “The Causative and Denominative Functions of the s- Preijix in Old Chinese”, in: Proceedings of the 2nd International Conference on Sinology, Taipei: Academia Sinica, 1989, 33–51. Meillet, Antoine, Linguistique historique et générale [Historical and general linguistics], Paris: Champion, 1982. Peiros, Ilia and Sergej A. Starostin, “Sino-Tibetan and Austro-Thai”, Computational Analyses of Asian and African Languages 22, 1984, 123–128. Peiros, Ilia and Sergej A. Starostin, A Comparative Vocabulary of Five Sino-Tibetan Languages, Melbourne: University of Melbourne Department of Linguistics and Applied Linguistics, 1996. Proulx, Paul, “The Linguistic Evidence on Algonquian Prehistory”, Anthropological Linguistics 22/1, 1980, 1–21. Pulleyblank, Edwin, “The Indo-European Vowel System and the Qualitative Ablaut”, Word 21, 1965, 86–101. Pulleyblank, Edwin, “Morphology in Old Chinese”, Journal of Chinese Linguistics 28, 2000, 26–51. Rankin, Robert, “Siouan, Yuchi and the Question of Grammatical Evidence for Genetic Relationship”, presidential address, Society for the Study of the Indigenous Languages of the Americas, 1998. Sagart, Laurent, The Roots of Old Chinese, Amsterdam: Benjamins, 1999. Sagart, Laurent, “Sources of Middle Chinese Manner Types: Old Chinese Prenasalized Initials in HmongMien and Sino-Tibetan Perspective”, Language and Linguistics 4/4, 2003, 757–768. Sagart, Laurent, “The Chinese Names of the Four Directions”, Journal of the American Oriental Society 124, 2004, 69–76. For use by the Author only | © 2017 Koninklijke Brill NV Glides, Phonological Status Sagart, Laurent, “Sino-Tibetan-Austronesian”, in: Laurent Sagart, Roger Blench and Alicia SanchezMazas, eds., The Peopling of East Asia: Putting Together Archaeology, Linguistics and Genetics, New York: RoutledgeCurzon, 2005, 177–181. Sagart, Laurent, Roger Blench and Alicia SanchezMazas, The Peopling of East Asia: Putting Together Archaeology, Linguistics and Genetics, New York: RoutledgeCurzon, 2005. Schuessler, Axel, ABC Etymological Dictionary of Old Chinese, Honolulu: University of Hawaii Press, 2007. Starosta, Stanley, “Proto-East Asian and the Origin and Dispersal of the Languages of East and Southeast Asia and the Paciijic”, in: Laurent Sagart, Roger Blench and Alicia Sanchez-Mazas, eds., The Peopling of East Asia: Putting Together Archaeology, Linguistics and Genetics, New York: Routledge Curzon, 2005, 182–198. Starostin, Sergej A., “Гипотеза о генетических связях синотибетских языков с енисейскими и северокавказскими языками” [A hypothesis on the genetic relationships of the Sino-Tibetan languages with the Yeniseian and the North Caucasian languages], in: I.F. Vardu, Лингвистическая реконструкция и древнейшая история Востока [Linguistic reconstruction and the ancient history of the East], Moscow: Institute of Orientalistics of the USSR Academy of Sciences, 1984, 19–38. Starostin, Sergej A., “Sino-Caucasian”, 2004, starling. rinet.ru. Starostin, Sergej A., “Altaic Loans in Old Chinese”, in: Alicia Sanchez-Mazas, Roger Blench, Malcom Ross and Ilia Peiros, eds., Past Human Migrations in East Asia, London: Routledge, 2008, 254–262. Sun, Jackson T.-S., “Parallelisms in the Verb Morphology of Sidaba rGyalrong and Lavrung in rGyalrongic”, Language and Linguistics 1/1, 2000, 161–190. Sun, Jackson T.-S., “Verb-Stem Variations in Showu rGyalrong”, in: Ying-chin Lin, Fang-min Hsu, Chunchi Lee, Jackson T.-S. Sun, Hsiu-fang Yang and Dahan Ho, eds., Studies on Sino-Tibetan Languages: Papers in Honor of Professor Hwang-Cherng Gong on His Seventieth Birthday, Taipei: Academia Sinica, 2004, 269–296. Thurgood, Graham, “A Subgrouping of the SinoTibetan Languages: The Interaction between Language Contact, Change, and Inheritance”, in: Graham Thurgood and Randy J. LaPolla, eds., The Sino-Tibetan Languages, London: Routledge, 2003, 1–21. Wagner, Donald B., Iron and Steel in Ancient China, Leiden: Brill, 1993. Xíng Gōngwǎn 邢公畹, Hàn-Táiyǔ bǐjiào shǒucè 汉台 语比较手册 [Comparative handbook of Sinitic and Taiwanese languages], Běijīng 北京: Shāngwù 商务 印书馆, 2000. Zhōu Fǎgāo 周法高, Zhōngguó gǔdài yǔfǎ, gòucí biān 中國古代語法，構詞編 [Ancient Chinese grammar: Morphology], Taipei 台北: Zhōngyāng Yánjiūsuǒ 中央研究所, 1962. Guillaume Jacques 306 Glides, Phonological Status The status of the medial onglides (also known as prenuclear or prevocalic glides), i.e., [j], [w] and [ɥ], within the Mandarin syllable has been debated fervently over the past century. Evidence has been applied from phonotactic constraints, morphophonemic processes, poetic alliteration and rhyme, reduplicative language games, and speech errors to argue for competing proposals, which include the grouping of the medial glide with the rime, the placement of the medial glide within the onset, treatment of the medial glide as a secondary articulation of the initial, asymmetric placement of front and back glides, and models which allow for indeterminacy or variation between speakers or dialects. 1. In i t i a l—Fi n a l T h e o r y In traditional Chinese philology, placement of the medial glide is with the rime, forming a unit known as the ijinal (also known as “rime projection”)—thus the syllable is divided into an initial consonant and a ijinal, as shown in Fig. 1. The arrangement assumes that the medial glide enjoys a closer afijiliation with the rime than with the initial consonant—a claim that is supported by some evidence from Mandarin phonotactics. Zhèng (2001) notes that that across the Chinese dialects both medial glide and nasal coda can displace an underlying midvowel while initial consonants cannot, which “justiijies the traditional grouping of the medial as part of the ijinal” (Zhèng 2001:36), while Yip (2002:60) cites a phonotactic constraint σ Initial Final Medial C G Rime Nucleus Coda V C (C=consonant; G=glide; V-vowel) Figure 1. The initial-ijinal model. For use by the Author only | © 2017 Koninklijke Brill NV