Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x What is Sino-Tibetan? Snapshot of a Field and a Language Family in Flux Zev Handel* University of Washington Abstract Sino-Tibetan is one of the great language families of the world, containing hundreds of languages spoken by over 1 billion people, from Northeast India to the Southeast Asian peninsula. The best-known languages in the family are Chinese, Tibetan, and Burmese. Although the existence of the family has been recognized for nearly 200 years, significant progress in reconstructing the history of the family was not achieved until the latter half of the twentieth century. In recent decades, this progress has accelerated, thanks to an explosion of new data and new approaches. At the same time, a number of interesting controversies have emerged in the field, centered on such issues as subgrouping and reconstruction methodology. 1 Introduction Sino-Tibetan (ST) is one of the great language families of the world, with the second-largest number of speakers after Indo-European (Matisoff 1991: 469; van Driem 1997: 455). In terms of geographic range, typological variation, and historical development, the family presents a complex and challenging picture to the historical comparative linguist. While the existence of the family has been recognized for some time (e.g. Leyden 1808; see Matisoff 1991 and van Driem 1997 for overviews of the history of Sino-Tibetan linguistics), progress in understanding its history and in reconstructing its ancestral sources has in general lagged behind that of many other language families. In recent years, the field of ST studies has been gripped by considerable ferment. Rapid progress has been made in some areas, while major scholarly controversies have erupted in others. In the last decade, a number of significant books and monographs have been published, including Ting and Sun (2000, 2001), Thurgood and LaPolla (2003); van Driem (2001); Gong (2002); Matisoff (2003), and Sagart et al. (2005), in addition to numerous important journal articles and book chapters. In this survey article, I will provide an overview of some recent work in the field, while focusing on a few issues that may be of particular interest to the nonspecialist. While a survey of this length will of necessity be incomplete, © 2008 The Author Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 423 I hope it will convey some of the flavor of the current excitement of work in the field, and also highlight areas where ST studies might contribute to, or benefit from, more general methodologies of comparative linguistics and historical reconstruction. The ST hypothesis, upon which the existence of the family is based, can be formulated in a number of ways. It proposes that a number of Asian linguistic groupings, notable among them those including the Chinese, Tibetan, and Burmese languages are genetically related.1 These genetically related languages together constitute the ST family.2 The family’s hypothetical most recent common ancestor is termed Proto-Sino-Tibetan (PST). Estimates of the number of languages in the family vary, but most scholars place the number between 200 and 300 (see, for example, Matisoff 2003: 3). Some languages have been discovered only recently, and it is likely that more languages will be discovered in coming years. There remains considerable disagreement about the time period and homeland of PST. The linguistic diversity of the family suggests that its time depth is comparable to that of Indo-European, or about 6000 years (Matisoff 1991: 470). Attempts to associate the PST homeland with specific Neolithic cultures attested in the archeological record would push the date back even further, possibly as far back as 9000 years (van Driem 2001: 410). The ancestral homeland of PST speakers has been placed in the Himalayas (Matisoff 1991), Sichuan province in southwest China (van Driem 2001), the Yellow River valley of northern China (Thurgood 2003; Sagart 2005a), and central China between the Yellow and Yangtze Rivers (Bellwood 2005). Wherever their place of origin, the earliest ST communities were probably agriculturally based, and successful enough that increasing population densities led to significant expansion (Bellwood 2005). SinoTibetan speakers now occupy much of East Asia, peninsular Southeast Asia, and parts of South Asia, having absorbed or displaced (or, in same cases, coexisted with) the peoples whose territories they invaded. This process continues today, especially in China, where the expansion of the Chinesespeaking Han ethnic majority is eroding the linguistic and cultural identities of minority peoples, including many speakers of non-ST languages. While the ST hypothesis is widely accepted, the overall membership of the ST family, the subgrouping of its constituents, and the broader connections to the other major language families of Southeast Asia are all controversial. I will discuss each of these issues below. It must also be acknowledged that the existence of the ST family (like that of many other proposed language families) has not yet been definitively proven, although significant progress has been made toward the reconstruction of the vocabulary of the ancestor language (see, for example, Peiros and Starostin 1996; Gong 2000 [2002] and the appendices of reconstructed roots in Benedict 1972 and Matisoff 2003). Alternative hypotheses to explain convergent features of, for example, Chinese and Tibetan, bear consideration. © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 424 Zev Handel 2 The Scope of Sino-Tibetan The history of ST, as a family of related languages recognized as such by Western philologists and linguists, dates back several hundred years. Detailed descriptions of the different hypotheses concerning the nature of the family, and their associated nomenclatures, can be found in van Driem 1997, 2005 (in English), and in Ting and Sun 2000 (in Chinese). Determinations about the scope of the family are intimately bound up with the other families in the region, notably Hmong-Mien (aka Miao-Yao), Tai-Kadai (aka Kra-Dai), Austroasiatic, and Austronesian.3 One of the first modern descriptions of the family, that of Li (1937), defines ST as consisting of Chinese, Tibeto-Burman (TB; the set of languages descended from the most recent common ancestor of Tibetan and Burmese, presumed to exclude Chinese), Tai-Kadai, and Hmong-Mien. This view of ST is essentially the same as that of most linguists in China today (Ting and Sun 2000). In contrast, most Western linguists today believe that Tai-Kadai and Hmong-Mien should be excluded from the ST family (Benedict 1972; Matisoff 1991; van Driem 1997; Thurgood and LaPolla 2003). In their view, the notable lexical and typological similarities among Chinese, Tai-Kadai, and Hmong-Mien are due to contact-induced convergence. The example of Vietnamese, a Mon-Khmer language that became monosyllabic and tonal under Chinese influence, demonstrates the plausibility of this view. The exclusion of a group like Tai-Kadai from the ST family does not, however, preclude the possibility of an older, more distant genetic relationship between the two families. I will discuss various proposals for macro-connections like this below. While there is disagreement about the constitution of the family and its wider genetic connections, there is widespread agreement on the identification of particular languages as belonging to Chinese, TB, Hmong-Mien, Tai-Kadai, etc. Thus for any particular definition of ST, its membership in terms of individual languages can be readily elaborated. In this article, I will adopt the modern conservative or narrow view of ST, excluding Tai-Kadai and Hmong-Mien. Defined this way, the membership of the family includes: • Chinese (aka the Sinitic languages), spoken primarily in the eastern half of the modern Chinese polity; • over 100 languages, reflecting tremendous linguistic diversity, spoken in the northeastern Indian provinces of Arunachal Pradesh, Assam, Nagaland, Meghalaya, Manipur, and Mizoram, including Bodo, Garo, Tiddim Chin, Mizo (aka Lushai), Meitei, Karbi (aka Mikir), Apatani, Kokborok, and Angami; • many languages of the Himalayas, including Newari and the Kiranti languages of Nepal, Dzongkha of Bhutan, Lepcha of Sikkim, and Tibetan; • the Qiangic and rGyalrongic languages spoken primarily in the southwestern Chinese province of Sichuan; © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 425 Table 1. Words for ‘five’, ‘fish’, and ‘I’ in selected Sino-Tibetan languages. ‘five’ ‘fish’ ‘I’ WT WB OC Bd CN Bai Gq Jg PTB l-na ña na nâ nâ na *Ana? *Bn(r)a *Ana ba ná an nâu nàu nò nv33 nv55 no31 nè35 tfr 53 m i53 nø35 mà31na33 na55 nai33 *l/b-na *s-nya *na WT, written Tibetan; WB, written Burmese; OC, Old Chinese; Bd, Bodo (Northeast India) (data from Bhat 1968); CN, Chang Naga (Northeast India) (data from Weidert 1987); Bai, Jianchuan Bai (Yunnan Province, Southwest China) (data from Sun et al. 1991); Gq, Guiqiong (Sichuan Province, Southwest China) (data from Sun et al. 1991); Jg, Jinghpaw (Burma) (data from Sun et al. 1991); PTB, Proto-Tibeto-Burman reconstructions (data from Matisoff 2003). • Burmese and a number of minority languages of Burma,4 such as the Karen languages and Jinghpaw (aka Kachin);5 and • many minority languages of southwest China and northern Thailand, including Yi, Lahu, Akha, Dulong, and Bai. In most cases, a language can be identified as ST if its words for ‘five’, ‘fish’, and the first-person singular pronoun ‘I’ are all near-homophonous and derivable from earlier forms similar to ña. Consider the examples in Table 1. Thus, when a newly discovered language like Rouruo (spoken by about 2100 people in China’s Yunnan province) comes to light, the following lexical items from the Guoli dialect can be used to initially classify the language as TB: ñu55 ‘I’, ño33 ‘five’, and ño33 ‘fish’ (see Sun et al. 2001). It is a relatively simple matter to come up with extensive lists, involving basic vocabulary, of apparent cognates among languages like Chinese, Tibetan, and Burmese (see Gong 1995 [2002], 2000 [2002], and Matisoff 2003 for examples). It is also fairly straightforward to identify patterns of regular sound correspondence. However, when these patterns are subjected to greater scrutiny, inconsistencies emerge. What appear to be prefixal elements, although they lack obvious morphological function, come and go capriciously. In contrast to the regularity of place and manner of articulation correspondences, there is considerable variability in voicing and aspiration correspondences. Aspirates alternate with non-aspirates, stops with nasals, and voiced obstruents with voiceless obstruents. In Benedict 1972, which despite its faults is recognized as having ushered in the modern era of ST studies, Benedict provides a chart of initial consonant correspondences. The two lines for velar initials are reproduced in Table 2. These correspondences were mocked by Roy Andrew Miller as an insult to the comparative method in his 1974 review of Benedict’s book. Most Sino-Tibetanists have dealt with this variation by making two assumptions: one, that some of this reflects natural variation in the parent language;6 two, that PST had a rich system of derivational prefixes and a © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 426 Zev Handel Table 2. Velar correspondences for Tibeto-Burman as given in Benedict 1972. TB Tibetan Kachin7 Burmese Garo Lushei8 *k *g k(h) g k(h)~g g~k(h) k(h) k k(h)~g g~k(h) k(h) k basic word structure that has been described as ‘sesquisyllabic’ by Matisoff and ‘iambic’ by Sagart.9 The derivational prefixes disappeared in many daughter languages, but not without in many cases influencing the articulation of the following consonant. One of the great challenges in ST linguistics is to uncover the various morphological processes operating in the different languages and subgroups, thus identifying root morphemes. This will, on the one hand, allow us to better explain what is now simply identified as ‘variation’ and, on the other hand, permit a more precise specification of sound laws and sound correspondences.10 3 Subgrouping of Sino-Tibetan A source of recent controversy in the field has been the proper subgrouping of the family. Van Driem (2003: 283) identifies 37 branches of the family, each consisting of closely related languages; the arrangement of these branches into higher-order groupings is one of the crucial tasks facing Sino-Tibetanists.11 Subgrouping remains difficult for a number of reasons. In the absence of a complete reconstruction of PST, it is difficult to identify shared innovations with certainty. Many languages remain insufficiently documented. Complex migration histories and areal convergence, both within ST and between ST languages and languages belonging to other families, can make it difficult to recognize genetic relationships. In addition to these objective difficulties, scholars have applied different methodological approaches to the identification of meaningful subgroups. Among the methodologies that have been employed are core vocabulary cognate counts (e.g. Benedict 1976), lexical analysis (Baxter 1995), shared innovations (Burling 1983; see Thurgood 2003 for additional examples), typological similarities, morphosyntax (e.g. van Driem 1997), and various hypotheses about migration patterns informed by archeological evidence (e.g. van Driem 2001). Finally, we can cite agnosticism as a methodological principle, for example, in Matisoff ’s purely geographical ‘Kamarupan’ grouping or his identification of Bai and Tujia as isolates within the family. Unfortunately, it has often been the case that these different methodological approaches have produced conflicting results, instead of mutual reinforcement. The examples of subgrouping schemes given here indicate some of the diversity of opinion in the field. For simplicity, only high-level nodes (one level below TB) are provided. © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 427 Fig. 1. The Sino-Tibetan family according to Benedict (1972). 1. Benedict (1972). It is interesting to observe from Figure 1 that Benedict declined to give a traditional tree structure for TB. Instead of a series of branchings, representing historical splits, his TB subgroups radiate out from Kachin (aka Jinghpaw), which he viewed as central to the family both geographically and typologically. The lines of the tree represent a degree of relationship (whether the result of common origin or of contact) rather than historical origin. 2. Matisoff (1991, 2003). See Figure 2. 3. Thurgood (2003). Sino-Tibetan consists of TB and Chinese. TibetoBurman is subgrouped as follows: • Lolo-Burmese branch (made up of Burmish and Loloish); • Bodic branch (including Tibetan); • ‘Sal’ languages (Bodo-Konyak-Jinghpaw) (see Burling 1983); • Kuki-Chin-Naga branch; • Rung branch (including Kiranti, rGyalrong, Dulong, Qiangic, and others); and • Karenic branch. Thurgood also lists several ‘small subgroups’, such as Tani (aka AborMiri-Dafla), as well as ‘unsubgrouped languages’ such as Bai. 4. Sun Hóngkai (2000). Sino-Tibetan consists of Chinese, TB, HmongMien, and Tai-Kadai. The TB subgrouping listed here is adapted from Ting and Sun 2000: 105: Tibetan-Himalayish group, Bodo© 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 428 Zev Handel Fig. 2. The Sino-Tibetan family according to Matisoff (1991, 2003). Fig. 3. The Sino-Tibetan family according to van Driem (1997). Naga-Chin group, Qiang-Jinghpaw group, Lolo-Burmese group, and Karen group. 5. van Driem (1997). See Figure 3. For other subgrouping schemes, see the list given in LaPolla 2001: 226, footnote 3. The description of TB found in Bradley 1997 is particularly useful, although it does not deal with Chinese. The subgrouping schemes of Thurgood and van Driem are worth considering in greater detail. In a series of articles (van Driem 1997, 2001, 2002, 2005, 2007), as well as in two reviews of Thurgood and LaPolla 2003 (van Driem 2003, 2004), van Driem has forcefully rejected the traditional bifurcation of the family into a Sinitic branch and a TB branch.12 Because he views Chinese as a node well within the family tree, he refers to the language family as a whole as TB rather than ST. [van Driem uses © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 429 the terms ‘truncated Tibeto-Burman’ (2002) and ‘pinioned Tibeto-Burman’ (2005) to refer to the TB of the traditional view, that is, a TB exclusive of Chinese.] van Driem’s claim rests on three legs. First, he argues that the traditional bifurcation was never anything more than impressionistic. Chinese appeared to be radically different in many respects from the other languages in the family in large part, because early reconstructions of Old Chinese, principally that of the revered Swedish Sinologist Bernhard Karlgren (as codified in Karlgren 1957), were significantly flawed, and thus failed to reveal many of its lexical, phonological, and morphological characteristics. Looking instead at advances in Old Chinese reconstruction in recent decades by such scholars as Baxter, Sagart, Pulleyblank, and Starostin, van Driem (2002: 89) says: ‘the new face of Old Chinese has given the language a decidedly more Tibeto-Burman appearance and, in fact, made Old Chinese look like just another Tibeto-Burman language, closer to the Bodic languages in countenance and certainly far less eccentric from the mainstream Tibeto-Burman point of view than, say, Gongduk or Toto.’13 van Driem argues that there never was compelling evidence for the relegation of Chinese to a TB sister node, and implies that had mid-twentieth century scholars like Benedict had access to better reconstructions of Chinese, it would never have occurred to them to bifurcate the family. There are, he says, no shared innovations on the TB side that would justify treating it as a separate node under ST. Second, van Driem (1997) argues that there are specific morphosyntactic and lexical affinities between Chinese and Limbu, one of the Kiranti languages of Nepal. Apparent vestiges of early Chinese morphological paradigms are associated with Kiranti morphological alternations (van Driem 1997: 463–71), and a number of lexical correspondences between Chinese and Limbu are listed (van Driem 1997: 471–84). This evidence is provided in support of his ‘Sino-Bodic’ hypothesis, which posits that Chinese and Bodic (the branch in which the Kiranti languages are found) form a subgroup within TB, called Sino-Bodic or Northern TB, based on ‘shared lexical isoglosses and vestiges of a shared morphology and morphosyntax’ (van Driem 1997: 461). Finally, van Driem argues that his overall subgrouping scheme is supported by archeological evidence reflecting migration patterns of the ancestors of the modern languages (van Driem 2001, 2007). Matisoff (2000), in a rejoinder to van Driem (1997), dismisses the Sino-Bodic hypothesis as an example of what he calls ‘neosubgroupitis’. For Matisoff, the core of van Driem’s argument is merely a collection of Old Chinese and modern Limbu forms that bear a superficial resemblance. Matisoff sees no persuasive evidence of regular sound correspondences, shared innovations, or sets of lexical items that lack cognates in other TB branches. [Indeed, van Driem does not explicitly claim that the shared features he identifies are Sino-Bodic innovations, acknowledging that they © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 430 Zev Handel may be shared retentions from Proto-Tibeto-Burman (PTB) (van Driem 1997: 465). His arguments, therefore, make a stronger case for eliminating a separate Chinese (or Sinitic) branch within Sino-Tibetan, and subsuming Chinese within TB, than for establishing a new Sino-Bodic subgroup.] As for van Driem’s morphosyntactic argument, Matisoff dispenses with it in a footnote (p. 367, note 34), noting that what van Driem refers to as the ‘PTB agreement system’ is a chimera. ‘TB agreement markers have long been recognized as having developed largely through the post-verbal cliticization of the personal pronouns.’ There is in fact considerable evidence in support of the cliticization hypothesis, although Matisoff ’s contention that the issue is long settled in the field is too dismissive (see LaPolla 1992: 298). In fact, the reconstruction of an ST proto-system of inflectional morphology is another area of controversy, one that has significant ramifications not only for subgrouping issues but also for the persuasiveness of the Sino-Tibetan hypothesis as a whole. While it is relatively easy to reconstruct a number of derivational processes found throughout the family (such as a causative *s- prefix; see LaPolla 2003 and Sagart 2006a for other examples), it is not at all clear whether PST had complex paradigms of verbal inflection or was primarily analytic. The modern members of the family are typologically all over the map, so to speak. Chinese, Bai, and the Loloish languages are characterized by tones, predominantly monosyllabic morphemes, and a lack of inflectional morphology. The languages of the Himalayas, most notably the Kiranti languages of Nepal, have rich and complex systems of verbal agreement. The former set of languages is located in East and Southeast Asia, in close contact with typologically similar languages in the Hmong-Mien and Tai-Kadai families (as well as Vietnamese). The latter are in contact with Indo-European languages with significant inflectional morphology. It is therefore possible to argue, as van Driem has, that the parent language had Kiranti-like paradigms, which became eroded in the eastern languages of the family. It is also possible to argue that the parent language was more isolating, and that the Western languages developed complex morphologies after splitting off. LaPolla (1992) has forcefully advocated the latter position, arguing that similarities in verb-agreement systems found in different TB branches are not remnants of an earlier proto-system but instances of new systems independently forming through the grammaticalization of pronouns. As elaborated in LaPolla 1994, such developments are attributable in part to the tendencies for parallel ‘drift’ that inhere in genetically related languages because of the perseverance of typological similarities – an idea originally put forward by Sapir. A third possibility that must be considered is that the parent language did have a significant amount of inflectional morphology, but that little of it survives in common across the main branches, making it impossible to reconstruct. Under this scenario, the Kiranti verbal paradigms and the isolating typology of Chinese and Loloish languages might both be innovations. © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 431 Even if one does not accept van Driem’s arguments for a particularly close relationship between Chinese and the Himalayish languages of TB, van Driem’s criticism of the bifurcation of the family, with Chinese an isolated taxon, must still be taken seriously. Matisoff argues that his subgrouping (found, for example, in Matisoff 1991, 2000, 2003) follows a ‘healthily agnostic approach’ (2000: 356) given the current state of knowledge, but this in itself does not justify placing Chinese outside of TB instead of making it a sister node on the same level as the seven other subgroups in his schema. Indeed, it is possible to argue that the bifurcation is in part a reflection of the scholarly interests and administrative divisions in academic institutions, where one tends to have either scholars devoted primarily to Chinese studies, or scholars devoted primarily to TB or ‘minor Asian language’ studies.14 van Driem’s main objection is the lack of shared innovations in what he calls ‘truncated Tibeto-Burman’ to justify its position beside Chinese in a ST stammbaum. He makes this clear in his review comments on Thurgood’s chapter on subgrouping in Thurgood and LaPolla: ‘no evidence has ever been adduced to support the hypothesis that truncated “TibetoBurman” (i.e. Tibeto-Burman minus Sinitic) shares common innovations that would define it as a coherent branch vs. Sinitic’ (2003: 283). Indeed, Thurgood himself acknowledges the importance of shared innovations in any subgrouping scheme (2003: 5) and justifies his internal TB subgroups partly on this basis, but his justification for the Chinese/TB bifurcation is terse: ‘the Sino-Tibetan family consists of two major subgroups, Chinese and Tibeto-Burman. By and large the distinction between the two is unambiguous and widely accepted. . . .’ (2003: 6). In support of Thurgood’s statement, it is possible to identify shared innovations in TB that set it apart from Chinese. For example, Gong (1995 [2002]) reconstructs PST with two distinct vowels *a and *e. The distinction is preserved in Old Chinese, but the vowels merge to *a in TB. Phonological mergers are, of course, among the types of sound changes that can positively be identified as innovations rather than retentions.15 Overall, the subgrouping issue in ST studies is plagued by a familiar Catch-22. Until a more complete reconstruction of PST exists, it is difficult to accurately identify shared innovations. Yet, without a clear subgrouping, it is difficult to properly weight and evaluate data from the daughter languages in order to refine PST reconstructions. Of course, it is possible that improved analysis of existing data will make it possible to identify more shared innovations even in the absence of a PST reconstruction, since certain types of changes (e.g. phonological mergers, grammaticalization of lexical items) are recognized as being unidirectional. It is also possible that more advanced computer-assisted statistical techniques will permit scholars to propose more reliable subgroupings based on the distribution of cognate sets across languages. Most likely, it will take several more decades of trial and error to sort through and test competing hypotheses. In the end, it © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 432 Zev Handel may not be possible to reduce the lower-order nodes of TB below a half-dozen or so (the Indo-European family presents us with a similar problem), but it should be possible to resolve the question of the position of Chinese within the family, especially as internal reconstructions of Old Chinese continue to be refined. 4 Broader Affinities of Sino-Tibetan The question of the position of Chinese within the ST family is intimately tied up with the broader question of the inter-relations among the language families of Southeast Asia (Sino-Tibetan, Tai-Kadai, Hmong-Mien, Austronesian, and Austroasiatic), not least because some proposals about such relations have hypothesized that Chinese has a closer relationship to one or another of these families than to TB, a position that naturally precludes the embedding of Chinese within a sub-branch of TB. Many macro-connections involving ST have been proposed. Sagart (1994) noted what might be common morphological prefixes in Austronesian, TB, Old Chinese, and Austroasiatic, and argued in particular for a genetic relationship between Chinese and Austronesian. He has recently argued (Sagart 2005b) that Tai-Kadai belongs within the Austronesian family, and thus is genetically related to Sino-Tibetan as well. Two Chinese linguists, Zhèng-Zhang Shàngfang (2000) and Pan Wùyún (1995, 2005), have argued that Sino-Tibetan (broadly construed to include TB, Chinese, Tai-Kadai, and Hmong-Mien) is related to Austronesian and Austroasiatic, forming a super-group termed Sino-Austric or Yangtzean. The ancestor of the proposed super-family encompassing all the major language families of Southeast Asia was called Proto-East Asian by Starosta (2005). More far-flung connections have also been proposed, such as between ST and Indo-European (Pulleyblank 1995) and among Sino-Tibetan, North Caucasian, and Yeniseian (Starostin 1991). None of these proposed macro-connections or super-families have been widely accepted, but some remain valid hypotheses. While proposals like those just listed, relating recognized language families in various larger genetic supergroups, have been circulating for a long time (e.g. Wilhelm Schmidt proposed in 1906 that Austronesian and Austroasiatic form an ‘Austric’ group), a number of proposals have recently gained new currency. While still controversial, they bear consideration. The most exciting and linguistically sophisticated proposal is that of Sagart (2005a), proposing a genetic connection between Austronesian and ST. (In earlier publications Sagart had argued that Chinese was genetically related to Austronesian but not to TB.) The proposal is supported by 75 lexical comparisons (showing regular sound correspondences) between Chinese and Austronesian and the identification of four shared morphological processes (all involving affixation) between ST and Austronesian. As an example of the latter, Sagart identifies the Proto-Austronesian ‘instrumental/beneficiary focus prefix *Si-’ with © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 433 Table 3. Three proposed Austronesian/Chinese cognate sets from Sagart 2005a. 1. 2. 3. Vocabulary item PAN or PECL brain female breast to shoot *punuq *nunuH1 *panaq Old Chinese *anu? *bno? *ana? ‘crossbow’ PAN, Proto-Austronesian; PECL, Proto-East-Coast-Linkage (the East Coast Linkage is a major branch within Austronesian that includes Malayo-Polynesian). the causative *s- in PST, arguing that both descend from an earlier valency-increasing prefix. What is particularly striking about Sagart’s cognate proposals is that they appear to reveal a regular Austronesian correspondence for the Chinese ‘Type A/B’ distinction. Old Chinese is known to have had two distinct syllable types, one of which conditioned phonological developments generally characterizable as palatalization. The precise nature of the distinction remains uncertain. Among the possibilities that have been proposed are the presence or absence of a segmental feature, a prosodic distinction, a vowel length distinction, a phonation type distinction, and a laryngealization or pharyngealization distinction. No clear-cut correlation of this distinction with any reconstructable feature of TB has ever been found. Sagart’s hypothesis is that bisyllables in the proto-language remained polysyllabic in ProtoAustronesian, but were reduced to monosyllables in PST through the loss of the first syllable. The Chinese syllable type (A or B) was conditioned by the voicing of the original initial consonant of the lost syllable.16 For example, consider the three comparisons in Table 3. (Chinese syllable type is indicated by a raised a or b.) Voiceless *p- has conditioned the development of syllable type A, while voiced *n- has conditioned the development of syllable type B. One difficulty with Sagart’s comparisons, or indeed with any comparisons of putative cognates between Old Chinese and Proto-Austronesian, is inherent in the different syllabic structures of morphemes in the two languages. On the one hand, the dozens of distinct pre-syllables in ProtoAustronesian roots (such as *pu-, *nu-, and *pa- in Table 3) correspond to zero (or, at best, a binary Type A/B distinction) in Old Chinese; on the other, the consonant clusters found in initial and coda position in the Old Chinese syllable correspond to simple initials in Austronesian. As a result, the number of possible correspondences is magnified and the possibility of chance resemblances is increased, reducing the statistical reliability of the proposed cognates. It remains to be seen how Sagart’s proposal for a Sino-Tibetan-Austronesian family will fare.17 It may be noted, however, that if his ideas are correct, © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 434 Zev Handel they might lend further support to the traditional ST bifurcation, since the loss of the A/B distinction can be seen as a shared innovation within the TB branch.18 5 Major Recent Work and Methodological Issues As mentioned earlier, the seminal work on Sino-Tibetan in the modern era was Benedict 1972. Benedict’s PTB reconstructions relied primarily on data from five key languages, each well-attested and each from a different subgroup. These languages were Tibetan, Burmese, Lushai (know usually referred to as Mizo), Garo, and Kachin (now usually referred to as Jinghpaw). Benedict argued that the major features of the ancestor language were preserved in one or more of these key languages, and that therefore the broad outlines of a reconstruction could be achieved without the need to first work out reconstructions for each TB subgroup. Benedict called this technique ‘teleoreconstruction’. Benedict supplemented this basic approach with supporting data from dozens of other languages, although much of this data was incomplete and of poor quality. Benedict compared his PTB reconstructions to the Old Chinese reconstructions of Bernhard Karlgren in order to reconstruct PST roots. The Tibeto-Burmanist who has worked most closely in the Benedict tradition has been James Matisoff, whose recent 2003 publication comprehensively lays out his current thinking on TB and ST comparison and reconstruction, as well as presenting hundreds of reconstructed lexical items and thousands of supporting forms. The interested reader can find here a discussion of variability in TB and numerous hypotheses concerning various derivational processes in the language. Matisoff is able to make use of data that are more numerous and reliable than that available to Benedict, and is also able to take advantage of many subgrouplevel reconstructions that have been published since 1972. (For more on these reconstructions, see below.) For a critical review of Matisoff ’s book, see Sagart 2006b. A different approach is found in the work of Gong Hwang-cherng, an expert in Chinese, Tibetan, and Tangut linguistics. His approach to ST comparison, refined in publications from 1980 to the present, has been to focus on four languages: reconstructed Old Chinese, written Tibetan, written Burmese, and written Tangut. In 2002, a volume of his collected works in ST studies was published, including articles in both English and Chinese. He follows in the long tradition of scholars who have worked primarily with the great written languages of the family. Thurgood and LaPolla 2003 is the first book that attempts to provide a general overview not just of the ST family but of many of its member languages. Most of the chapters are mini-grammars of, or brief introductions to, a representative sample of these languages. Given the size of the family, the book is inevitably incomplete in its coverage, but will give the © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 435 interested reader an excellent sense of the typological variation found across the family. As these and other recent major publications (Ting and Sun 2000, 2001; van Driem 2001) reveal, Sino-Tibetanists are still grappling with fundamental methodological problems as they try to adapt the comparative method and other traditional techniques of historical linguistics to the particular typological and historical characteristics of the family. Because few inflectional paradigms can be identified across the family, and it has therefore proven difficult to reconstruct an inflectional morphology for the proto-language, the identification of shared morphological paradigms (especially irregular inflectional paradigms) cannot be relied on to prove the existence of the family.19 This fact means that the establishment of a genetic relationship must rely on shared derivational morphology and lexical comparisons.20 But given a sufficient degree of contact, it is possible for both derivational morphology and core vocabulary to be borrowed.21 The problem is exacerbated by the existence of many languages with monosyllabic (or sesquisyllabic) roots, simple syllable structure, and compact phonological inventories. The possibility of chance resemblances leading to the false identification of cognates, especially when sound correspondences are imprecise and morphological patterns are not fully understood, is relatively high. Through an ongoing process of experimentation and debate, these methodological hurdles may be overcome. At the same time, another major current in ST studies is underway, namely, ongoing descriptive and analytical work, especially on the lesser-studied languages of Northeast India, the Himalayas, and western China. Many excellent Sino-Tibetanists are at work in the field. Descriptions of new languages, as well as improved descriptions of known languages, continue to be published at a rapid rate.22 The accumulation of so much new data over the last several decades has allowed scholars to achieve detailed mid-level reconstructions of recognized ST subgroups. Among these valuable works are Bradley 1978 (Proto-Loloish), French 1983 (Proto-Northern Naga), J. Sun 1993 (Proto-Tani), Mazaudon 1994 (Proto-Tamangic), Opgenort 2005 (Proto-Kiranti), and Joseph and Burling 2006 (Proto-Bodo-Garo). Much recent excitement has been engendered by study of the complex Qiangic languages, which will likely provide many new insights into the history of ST as a whole. (Two of the many recent publications on Qiangic that can be cited as examples are Evans 2001 and LaPolla and Huang 2003.) As these trends continue, the traditional prominence held by the great written languages of the family will be diminished. The resulting shape of the ST family, its internal contours, and external relations, as well as its reconstructed lexicon, phonology, morphology, and syntax, will no doubt be markedly different from today’s default assumptions. It is even conceivable, though unlikely, that the ST hypothesis itself will lose favor. © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 436 Zev Handel 6 Conclusion Sino-Tibetan linguistic research is an important and increasingly vibrant academic endeavor. In conjunction with archeological, historical, geographic, and genetic studies, it is helping to enrich our understanding of the human cultures and migrations that have shaped the Asian continent. Within the field of ST linguistics, the work of historical comparativists and that of descriptive fieldworkers have been mutually galvanizing, increasing our knowledge of contemporary languages – many spoken by small and decreasing numbers of speakers, and many unknown or poorly described until recent decades – and improving our understanding of historical changes and developments. At the same time, the methodological insights gained from historical comparative work on the ST languages are proving valuable to linguists working on other language families whose typologies, complex histories, and inter-relationships make them less amenable to strict application of the traditional comparative method than has proven to be the case for families like Indo-European and MalayoPolynesian. Every language family and linguistic area is a living laboratory, yielding unique insights into language contact, development, convergence, and divergence. While the ultimate outcome of current and future work in the field – including even the fate of the ST hypothesis itself – is unknown, one thing that is clear is that research into this large, diverse, and fascinating group of languages will continue to be of immense interest and importance well beyond the confines of the field. Acknowledgements I would like to thank Guillaume Jacques for helpful comments. I am also indebted to two anonymous reviewers who provided extensive and thoughtful comments, corrections, and suggestions. I am of course solely responsible for any remaining errors of fact or judgment. Short Biography Zev Handel’s research is concerned with all aspects of the history of the Chinese language, and can be broadly described as falling into three categories: (i) the reconstruction of the earliest stages of Chinese pronunciation; (ii) comparison of Chinese with the TB languages; and (iii) historical/ comparative work on the Chinese dialects. His papers have appeared in such journals as Journal of Chinese Linguistics, Cahiers de Linguistique – Asie Orientale, Language and Linguistics, and Linguistics of the Tibeto-Burman Area. In 2002, he won the International Association of Chinese Linguists’ Young Scholar Award for his presentation Northern Min tone values and the reconstruction of ‘softened initials’ at the International Association of Chinese Linguistics conference in Nagoya, Japan. This year his book-length monograph, © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 437 Old Chinese Medials and Their Sino-Tibetan Origins: A Comparative Study, will be published by the Institute of Linguistics, Academia Sinica, Taiwan. He is currently researching the adaptation of Chinese characters in the earliest writing systems of Japanese, Korean, and Vietnamese. He holds a BA in mathematics from Harvard and a PhD in Chinese from the University of California, Berkeley, and teaches at the University of Washington in Seattle. Notes * Correspondence address: Zev Handel, Department of Asian Languages and Literature, University of Washington, Box 353521, Seattle, WA 98105, USA. E-mail: zhandel@u.washington.edu. 1 While Chinese is typically conceived of as a language comprising numerous dialects, linguistically it is more accurately described as a highly diversified but closely related group of languages, similar to the Romance languages in scope and time depth. These Chinese languages can be categorized into subgroups, often called major dialect groups. Languages in different subgroups are mutually unintelligible, and some of the subgroups also boast considerable internal diversity. Some scholars refer to Chinese as Sinitic in order to emphasize the fact that it is a family rather than a single language. In this article, I will follow common practice by referring to the different spoken forms of Chinese as ‘dialects’. 2 As we will see below, some scholars refer to this same family as Tibeto-Burman. These scholars reject the term Sino-Tibetan, because they associate it with a particular hypothesis about the internal structure of the family. For the purposes of this article, I will use the term ‘Sino-Tibetan’ as just defined, without prejudging its internal composition. I will use the term ‘Tibeto-Burman’ to refer to a taxon that, according to the dominant subgrouping hypothesis, exists within the Sino-Tibetan family. 3 Tai, of which Thai (or Siamese) is the best-known member, is one of five subgroups in the Tai-Kadai family. The others are Kam-Sui, Hlai, Be, and Kra. 4 The military junta running Burma has changed the official name of the country to Myanmar. I will use the more familiar name Burma in this paper. 5 Many TB languages are referred to by multiple names or spellings. Jinghpaw, for example, may be called Kachin, Jingpo, Jinghpo, or Jingpho. See Matisoff 1996 for an extensive crossreferenced list of TB language names, and Bradley 1997 for a list of languages, alternate language names, speaker populations, and locations. 6 For a fully developed discussion of variation within word families and its relationship to semantics, see Matisoff 1978. 7 This is another name for the Jinghpaw language. 8 This name (also written Lushai) is now considered pejorative, and has been replaced by Mizo. 9 This refers to a two-syllable phonological structure in which the first syllable is reduced. English words like ‘terrain’ and ‘below’ illustrate this kind of structure. Such words are common, for example, in Modern Burmese and Jinghpaw. The phonological reduction of the first syllable is manifested in its simple Ce- structure, where aspiration and voicing contrasts are neutralized, so that only a few consonants can fill the C slot. Some sesquisyllabic forms might be indivisible roots; others might result from the reduction of the first syllable of compounds; still others from prefixation of a monosyllabic root with vowel epenthesis. In many cases, there is insufficient evidence to determine the etymology of sesquisyllabic words. 10 For example, one variational pattern that recurs in TB roots is alternation between *i and *ya, as in *mik~*myak ‘eye’. (Here y represents a palatal glide [ j], following customary practice in TB studies.) Is this the last vestige of a long-dormant morphological process? (As an analogy, one might imagine that in descendants of modern English the distinction between the simple past and the present perfect will be neutralized. Where irregular verbal paradigms currently exist, some future dialects will select one form for the new past tense, other dialects the other form. So in one daughter language we end up with ‘I sang’, in another ‘I sung’. Similarly with © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 438 Zev Handel ‘rang’/‘rung’, ‘swam’/‘swum’, etc. The historical linguist might be able to reconstruct an original *a/*u alternation without being able to identify its origin, other than to say it represents dialect mixture or proto-variation.) Or does it reflect dialectal or sociolinguistic differences in pronunciation? Or is it a vestige of layering resulting from contact and borrowing at different time depths among languages within the family? 11 Not all scholars agree with the specifics of van Driem’s 37-branch classification, but the subgrouping problem presents similar challenges regardless of any particular breakdown of the lowest-order branches. 12 ‘Traditional’ is here a relative term; van Driem (1997) stresses that the bifurcation is a midtwentieth-century artifact. 13 To this list of scholars, we might also add the Chinese historical phonologists Jin Lixin and Zhèng-Zhang Shàngfang. 14 In fact, at many of the International Conferences on Sino-Tibetan Languages and Linguistics, sessions are clearly delineated as being primarily ‘Chinese’ or primarily ‘Tibeto-Burman’. 15 Handel (1998) notes another possible innovation in TB, the loss of *-r- after *s- before low or back vowels. Chinese preserves this *-r-. Of course, one or two apparent shared innovations is not as persuasive as a full set of lexical, phonological, and morphological isoglosses that all support the establishment of a separate taxon in the family tree. As Guillaume Jacques has pointed out (personal communication, 17 August 2007), it is possible to find distinctions in other branches of the ST family that do not appear to be reflected elsewhere. Any such distinction could theoretically be projected back to the proto-language, thus, providing grounds for grouping all other branches together in one taxon marked by an innovative merger. As an example Jacques notes that the Proto-rGyalrong distinction between velarized and non-velarized vowels seems not to exist in any other branch of ST. 16 This is something of an over-simplification of a more complex correspondence pattern. Sagart himself describes it as a ‘statistical correlation’. See Sagart 2005a for details. 17 Sagart’s earlier proposals for a Chinese–Austronesian connection were widely criticized, in part because they entailed a refutation of the genetic relationship between Chinese and TB. 18 Unless, of course, a feature that correlates with the Chinese Type A/B distinction is discovered in a TB language. Zhèng-Zhang Shàngfang (2000: 54–5, 2003: 182) has put forward evidence for a correlation with vowel length in Dulong, and Starostin (1991) has made a similar claim for vowel length in Mizo (Lushai). But these correlations fall short of the degree of regularity necessary for a persuasive claim of correspondence, and, as Norman (1994: 400) notes of Starostin’s claim, ‘still precariously depend on the validity of his Sino-Lushei comparisons, many of which appear highly speculative at best’. 19 The classic example of this type of shared irregular paradigm is the comparison of Latin est/sunt (the third-person singular and plural forms of the verb ‘to be’) with German ist/sind, which is generally recognized as sufficient evidence to establish the genetic relationship between the two languages. Similar examples in ST are not obviously apparent. Recently, Guillaume Jacques (2007) has identified an irregular pronominal paradigm shared by Chang (a Naga language spoken in Northeast India) and Taoping (a dialect of southern Qiang spoken in Sichuan), TB languages sufficiently distant geographically as to make contact-induced convergence highly unlikely. If this or other paradigms can be shown to be more widespread, it would constitute powerful evidence in support of the ST hypothesis. 20 There is quite a bit of shared derivational morphology in the family. For an overview, see LaPolla 2003. 21 For example, many core lexical items were borrowed into northern varieties of Middle English from Norse, and a number of these survive in modern English, among them verbs give, get, and run and pronominal forms they, them, and their (Thomason and Kaufman 1988: 292–5). Middle English later borrowed a number of derivational morphemes from French, many of which became sufficiently productive to be affixed to Germanic roots, such as -able in speakable and knowable, and -age in bondage. See Thomason and Kaufman 1988 for further discussion of intense language contact situations. 22 The Central University for Nationalities Press has recently begun publishing a series titled New found minority languages in China (Zhongguó xin faxiàn yuyán yánjiu cóngshu ). A number of English-language reviews of books in the series have appeared in recent issues of Linguistics of the Tibeto-Burman Area. © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 439 Works Cited Baxter, William H. 1995. A stronger affinity . . . than could have been produced by accident: a probabilistic comparison of Old Chinese and Tibeto-Burman. The ancestry of the Chinese Language, ed. by William S.-Y. Wang, 1–39. Berkeley, CA: Project on Linguistic Analysis. Bellwood, Peter. 2005. Examining the farming/language dispersal hypothesis in the East Asian context. The peopling of East Asia: putting together archaeology, linguistics and genetics, ed. by Laurent Sagart, Roger Blench and Alicia Sanchez-Mazas, 17–30. London: Routledge. Benedict, Paul K. 1972. Sino-Tibetan: a conspectus. Contributing editor James A. Matisoff. New York, NY: Cambridge University Press. ——. 1976. Sino-Tibetan: another look. Journal of the American Oriental Society 96.2.167–97. Bhat, D. N. Shankara. 1968. Boro vocabulary (with a grammatical sketch). (Deccan College Building Centenary and Silver Jubilee Series, 59.) Poona, India: Deccan College Postgraduate and Research Institute. Bradley, David. 1978. Proto-Loloish. (Monograph Series, 39.) Copenhagen, Denmark/London: Scandinavian Institute of Asian Studies. ——. 1997. Tibeto-Burman languages and classification. Tibeto-Burman languages of the Himalayas (Papers in Southeast Asian Linguistics, 14), ed. by David Bradley, 1–72. Canberra, Australia: Department of Linguistics, The Australian National University. Burling, Robbins. 1983. The Sal languages. Linguistics of the Tibeto-Burman Area 7.2.1–32. van Driem, George. 1997. Sino-Bodic. Bulletin of the School of Oriental and African Studies 60.3.455–88. ——. 2001. Languages of the Himalayas (2 vols.). Leiden, The Netherlands: E.J. Brill. ——. 2002. Tibeto-Burman replaces Indo-Chinese in the 1990s: Review of a decade of scholarship. Lingua 112.2.79–102. ——. 2003. Review of Thurgood and LaPolla 2003. Bulletin of the School of Oriental and African Studies 66.2.282–4. ——. 2004. Review of Thurgood and LaPolla 2003. Journal of Asian Studies 63.4.1127–8. ——. 2005. Sino-Austronesian vs. Sino-Caucasian, Sino-Bodic vs. Sino-Tibetan, and TibetoBurman as default theory. Contemporary issues in Nepalese linguistics, ed. by Yogendra Prasada Yadava, Govinda Bhattarai, Ram Raj Lohani, Balaram Prasain and Krishna Parajuli, 285–338. Kathmandu, Nepal: Linguistic Society of Nepal. ——. 2007. The Diversity of the Tibeto-Burman language family and the linguistic ancestry of Chinese. Bulletin of Chinese Linguistics 1.2.211–70. Evans, Jonathan. 2001. Introduction to Qiang phonology and lexicon: synchrony and diachrony. Tokyo, Japan: ILCAA, Tokyo University of Foreign Studies. French, W. T. 1983. Northern Naga: a Tibeto-Burman mesolanguage. New York, NY: City University of New York. Gong, Hwang-cherng [Gong Huángchéng] . 1995 [2002]. The system of finals in Proto-Sino-Tibetan. The ancestry of the Chinese language. Collected papers on SinoTibetan linguistics, 79–124. Taipei, Taiwan: Institute of Linguistics (Preparatory Office), Academia Sinica. ——. 2000 [2002]. Cóng Yuánshi Hàn-Zàngyu dào Shànggu Hànyu yijí Yuánshi Zàng-Mianyu de yùnmu yanbiàn [Vocalic development from Proto-Sino-Tibetan to Old Chinese and Proto-Tibeto-Burman]. Collected papers on Sino-Tibetan linguistics, 213–41. Taipei, Taiwan: Institute of Linguistics (Preparatory Office), Academia Sinica. ——. 2002. Hàn-Zàngyu yánjiu lùnwén jí [Collected papers on SinoTibetan linguistics]. (Language and Linguistics Monograph Series, C2-2.) Taipei, Taiwan: Institute of Linguistics (Preparatory Office), Academia Sinica. Handel, Zev. 1998. The medial systems of Old Chinese and Proto-Sino-Tibetan. Berkeley, CA: University of California dissertation. Jacques, Guillaume. 2007. A shared suppletive pattern in the pronominal systems of Chang Naga and Southern Qiang. Cahiers de Linguistique Asie Orientale 36.1.61–78. Joseph, U. V., and Robbins Burling. 2006. The comparative phonology of the Boro-Garo Languages. Manasagangotri, Mysore, India: Central Institute of Indian Languages. © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd 440 Zev Handel Karlgren, Bernhard. 1957. Grammata Serica recensa. Bulletin of the Museum of Far Eastern Antiquities 29.1–332. LaPolla, Randy J. 1992. On the dating and nature of verb agreement in Tibeto-Burman. Bulletin of the School of Oriental and African Studies 55.2.298–315. ——. 1994. Parallel grammaticalizations in Tibeto-Burman: Evidence of Sapir’s ‘drift’. Linguistics of the Tibeto-Burman Area 17.1.61–80. ——. 2001. The role of migration and language contact in the development of the SinoTibetan language family. Areal diffusion and genetic inheritance: problems in comparative linguistics, ed. by Alexandra Y. Aikhenvald and R. M. W. Dixon, 225–54. Oxford, UK: Oxford University Press. ——. 2003. Overview of Sino-Tibetan Morphosyntax. The Sino-Tibetan languages, ed. by Graham Thurgood and Randy J. LaPolla, 22–42. London: Routledge. LaPolla, Randy J., and Chenglong Huang. 2003. A grammar of Qiang: with annotated texts and glossary. Berlin, Germany: Mouton de Gruyter. Leyden, John. 1808. On the languages and literature of the Indo-Chinese nations. Asiatick Researches 10.158–289. Li, Fang-kuei. 1937 [1973]. Languages and dialects of China. Journal of Chinese Linguistics 1.1.1–13. Matisoff, James A. 1978. Variational semantics in Tibeto-Burman: the ‘organic’ approach to linguistic comparison. Philadelphia, PA: Publications of the Institute for the Study of the Humanities. ——. 1991. Sino-Tibetan linguistics: present state and future prospects. Annual Review of Anthropology 20.469–504. ——. 1996. Languages and dialects of Tibeto-Burman. (STEDT Monograph, 2). Berkeley, CA: Center for Southeast Asia Studies, University of California. ——. 2000. On ‘Sino-Bodic’ and other symptoms of neosubgroupitis. Bulletin of the School of Oriental and African Studies 63.3.356–69. ——. 2003. Handbook of Proto-Tibeto-Burman: system and philosophy of Sino-Tibetan reconstruction (UC Publications in Linguistics, 135.). Berkeley, CA: University of California Press. Mazaudon, Martine. 1994. Problèmes de comparatisme et de reconstruction dans quelques langues de la famille Tibéto-Birmane. Paris, France: Université de la Sorbonne Nouvelle dissertation. Miller, Roy Andrew. 1974. Sino-Tibetan: inspection of a conspectus. Journal of the American Oriental Society 94.2.195–209. Norman, Jerry. 1994. Pharyngealization in early Chinese. Journal of the American Oriental Society 114.3.397–408. Opgenort, Jean Robert. 2005. A grammar of Jero: with a historical comparative study of the Kiranti languages. Leiden, The Netherlands: Brill. Pan Wùyún . 1995. Duì Huá-Ào yuxì jiashuo de ruògan zhichí cáiliào [Some evidence in support of the Sino-Austric hypothesis]. The ancestry of the Chinese Language, ed. by William S.-Y. Wang, 113–44. Berkeley, CA: Project on Linguistic Analysis. ——. 2005. Hàn-Tái guanxi cí zhong de tóngyuán céng tàntao [An investigation of the cognate layer in Sino-Tai related words]. Hànyushi yánjiu: Jìniàn Li Fangguì xiansheng bainián míngdàn lùnwén jí [Essays in Chinese historical linguistics: Festschrift in memory of Professor Fang-Kuei Li on his centennial birthday], ed. by Pang-Hsin Ting and Anne O. Yue, 27–55. Taipei, Taiwan: Institute of Linguistics, Academia Sinica. Peiros, Ilia, and Sergej A. Starostin. 1996. A comparative vocabulary of five Sino-Tibetan languages. Melbourne, Australia: University of Melbourne Department of Linguistics and Applied Linguistics. Pulleyblank, Edwin G. 1995. The historical and prehistorical relationships of Chinese. The ancestry of the Chinese language, ed. by William S.-Y. Wang, 145–94. Berkeley, CA: Project on Linguistic Analysis. Sagart, Laurent. 1994. Old Chinese and Proto-Austronesian evidence for Sino-Austronesian. Oceanic Linguistics 33.2.271–308. ——. 2005a. Sino-Tibetan-Austronesian: An updated and improved argument. The peopling © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd Sino-Tibetan: a Snapshot 441 of East Asia: Putting together archaeology, linguistics and genetics, ed. by Laurent Sagart, Roger Blench and Alicia Sanchez-Mazas, 161–76. London: Routledge. ——. 2005b. Tai-Kadai as a subgroup of Austronesian. The peopling of East Asia: Putting together archaeology, linguistics and genetics, ed. by Laurent Sagart, Roger Blench and Alicia Sanchez-Mazas, 177–81. London: Routledge. ——. 2006a. On intransitive nasal prefixation in Sino-Tibetan languages. Cahiers de Linguistique Asie Orientale 35.1.57–70. ——. 2006b. Review of Matisoff 2003. Diachronica 23.1.206–23. Sagart, Laurent, Roger Blench and Alicia Sanches-Mazas (eds.) 2005. The peopling of East Asia: Putting together archaeology, linguistics and genetics. London: Routledge. Starosta, Stanley. 2005. Proto-East Asian and the origin and dispersal of the languages of East and Southeast Asia and the Pacific. The peopling of East Asia: Putting together archaeology, linguistics and genetics, ed. by Laurent Sagart, Roger Blench and Alicia Sanchez-Mazas, 182–97. London: Routledge Curzon. Starostin, Sergej A. 1991. On the hypothesis of a genetic connection between the Sino-Tibetan languages and the Yeniseian and North Caucasian languages. Dene-Sino-Caucasian languages, ed. by Vitaly Shevoroshkin, 12–41. Bochum, Germany: Brockmeyer. Sun Hóngkai , Chénglóng Huáng , and Máocao Zhou . 2001. Róuruòyu yánjiu [A study of Rouruo]. Beijing, China: Zhongyang Mínzú Dàxué Chubanshè [Central University for Nationalities Press]. Sun Hóngkai (ed.) et al. 1991. Zàng-Mianyu yuyin hé cíhuì [Tibeto-Burman phonology and lexicon]. Beijing, China: Zhongguó Shèhuì Kexué Chubanshè [Chinese Social Sciences Press]. Sun, Jackson. 1993. A historical-comparative study of the Tani (Mirish) branch in TibetoBurman. Berkeley, CA: University of California dissertation. Thomason, Sarah, and Terrence Kaufman. 1988. Language contact, creolization, and genetic linguistics. Berkeley/Los Angeles, CA: University of California Press. Thurgood, Graham. 2003. A subgrouping of the Sino-Tibetan languages: the interaction between language contact, change, and inheritance. The Sino-Tibetan languages, ed. by Graham Thurgood and Randy J. LaPolla, 1–21. London: Routledge. Thurgood, Graham, and Randy J. LaPolla (eds). 2003. The Sino-Tibetan languages. London: Routledge. Ting Pang-Hsin [Ding Bangxin] , and Hóngkai Sun . 2000. Hàn-Zàngyu yánjiu de lìshi huígù [Retrospective history of Sino-Tibetan studies]. (HànZàngyu tóngyuáncí yánjiu [Cognate words in Sino-Tibetan languages], 1). Nanning: Guangxi Mínzú Chubanshè [Guangxi Nationalities Press]. ——. 2001. Hàn-Zàng, Miáo-Yáo tóngyuáncí zhuantí yánjiu [Special topics in Sino-Tibetan and Hmong-Mien comparison]. (Hàn-Zàngyu tóngyuáncí yánjiu [Cognate words in Sino-Tibetan languages], 2). Nanning: Guangxi Mínzú Chubanshè [Guangxi Nationalities Press]. Weidert, Alfons K. 1987. Tibeto-Burman tonology: A comparative account. (Current Issues in Linguistic Theory, 54.) Amsterdam, The Netherlands/Philadelphia, PA: John Benjamins. Zhèng-Zhang Shàngfang . 2000. The phonological system of Old Chinese. Cahiers de Linguistique Asie Orientale Monograph, 5. English translation by Laurent Sagart. ——. 2003. Shànggu yinxì [The sound system of Old Chinese]. Shanghai, China: Shànghai Jiàoyù Chubanshè [Shanghai Education Press]. © 2008 The Author Language and Linguistics Compass 2/3 (2008): 422–441, 10.1111/j.1749-818x.2008.00061.x Journal Compilation © 2008 Blackwell Publishing Ltd