Quick viewing(Text Mode)

The Phonology, Meaning, and Origin of the Epithet Ḥarya ~ Ārya in East Asia

The Phonology, Meaning, and Origin of the Epithet Ḥarya ~ Ārya in East Asia

The earliest Chinese words for ‘the Chinese’: The phonology, meaning, and origin of the epithet ḥarya ~ ārya in

Christopher I. Beckwith

Abstract The first signs of Chinese ethnic self-consciousness coincide with the beginnings of political philosophy in the Warring States period. The peoples who spoke Old Chinese adopted an unprecedented appellation for themselves and also began referring to their many states collectively as if they were, or should be, a unified country with a unified imperial ancestry, the Hsia 夏 (Old Chinese *Ḥarya). Most strikingly, this new self appellation, which eventually became the autonym for ‘Chinese’, is a widespread epithet of Central Eurasian ruling peoples as far west as Persia and continued to be used into the Early Middle Ages. Analysis of the data shows that in the Warring States period the Chinese acquired this new term and idea from the Central Eurasian population of the state of Chao.1

Résumé Les premiers signes chinois d’une prise de conscience ethnique coïncident avec le début de la philosophie politique dans la période des Royaumes combattants. Les peuples locuteurs du Chinois archaïque adoptaient une appellation sans précédent, Hia 夏 (Chin. arch. *Ḥarya), faisant ainsi allusion aux États collectivement, comme s’ils constituaient une unité, partageant le même patrimoine ancestral et impérial. Cette nouvelle appellation qui devenait l’autonyme pour « Chinois » fut utilisée comme qualificatif très répandu par des peuples dominants, de l’Eurasie centrale vers l’ouest jusqu’à la Perse, encore dans le haut Moyen Âge. L’analyse des données révèle que dans la période des Royaumes combattants, les Chinois ont acquis cette nouvelle appellation et son idée directrice des peuples Centre-Eurasie de l’État de Tchao.

1 Hsia 夏~ Hua 華 ‘the Chinese’: (ca. 390 to ca. 312 bc),2 and other contemporaneous Texts and traditional internal reconstruction Warring States period texts contain the earliest datable usages of the Classical Chinese words used as generic The origins of Chinese ‘ethnic’ or ‘national’ con- terms for the dominant inhabitants of the early Chinese sciousness, whether or not connected to political unifica- cultural area, which comprised many states large and tion, have long been mysterious. The Tso chuan 左傳 small during the period when the texts were composed. The words do not occur in the sense ‘(the) Chinese’ in any earlier texts,3 including the Ch’un ch’iu 春秋, the 1 I am indebted to the Japan Foundation for supporting part of the genuine early chronicle on the basis of which the Tso research and writing of this paper with a Short Term Fellowship in Tokyo (summer, 2013). I would like to thank E. Bruce Brooks, Yanxiao chuan was later written. He, Gisaburo N. Kiyose, Victor Mair, and Andrew E. Shimunek for The usual generic term is Hsia 夏 MSC (Modern corrections and suggestions for improvements, and Wolfgang Behr for Standard Chinese) xià, which as a common noun in Clas- sending me his article on Hsia 夏 (Behr 2007). I also thank the Acad- sical and modern Chinese means, literally, ‘summer’.4 It emy of Korean Studies (Seoul) for inviting me to present an earlier version of this paper in Seoul in 2013. I am of course responsible for any errors. Abbreviations: Bax. (Baxter 1992); C: any consonant; 2 Dating by E. Bruce Brooks and Taeko Brooks . Historic Sinological Reconstruction (the traditional system based on 3 Mair (2013). medieval rhyme tables); JDB (Omodaka et al. 1967); LMC: Late Mid- 4 Karlgren (1957: 28). According to Mair (2013: 5, 7-8), Hsia 夏 dle Chinese; Pul. (Pulleyblank 1991); MSC: Modern Standard Chinese is first attested in the Bronze Inscriptions and early Classical texts in in spelling; Sch. (Schuessler 1987); Sta. (Starostin 1989); Tak. the meaning “large, grand; variegated”, and is later written with an (Takata 1988); V: any vowel. For forms reconstructed (by anyone) additional “sun” (日) radical to express the meaning “summer”. It from Chinese characters via the traditional method, tone marks are con- occurs in its usual form, meaning ‘summer’, in the earliest authentic verted to subscript numerals and the mark (✩) is added. Internal deri- historical work from the Spring and Autumn–Warring States period, the vation is marked by < (‘from’); → and ← mark direction of borrowing. Ch’un ch’iu ‘Spring and Autumn Annals’.

Journal Asiatique 304.2 (2016): 231-248 doi: 10.2143/JA.304.2.3186091

99263_JAS_2016-2_06_Beckwith.indd 231 25/01/17 12:38 232 Christopher I. Beckwith

is also the name of the legendary first dynasty of Chinese “The Master said, “The rude tribes of the east and north history, traditionally dated from the end of the third mil- have their princes, and are not like the States of our great lennium to the middle of the second millennium bc.5 In land which are without them.” However, since all of the the Tso chuan, Hsia 夏 without the explicit collective 諸夏 ‘Chinese (peoples, states)’ had princes, the point plural marker chu 諸 MSC zhū sometimes means ‘Hsia was surely that they were not unified under one prince. Dynasty’ or things or people related to it, so it is not This is supported by another famous early example, from always generic. However, as 諸夏, in the collective plu- the Tso chuan, where the Chinese are compared unfa- ral, it always means ‘the Chinese (peoples, states)’,6 as it vorably to all foreign peoples: usually does in the unmarked form 夏 as well. The text 吾聞之,天子失官,學在四夷。 once explicitly contrasts the 諸夏 ‘the Hsia (peoples or states)’7 with the foreign peoples of the north and west.8 I have heard that if a Son of Heaven has lost (the way of In one instance it is used in contrast to the semi-Sinified good) governance, (he should) study it among the four I.11 southern states, such as Ch’u 楚 and Yüeh 越.9 The 夷 explicitly plural 諸夏 occurs also in the Lun yü 論語 Here the “four I” refers to the foreign peoples (I ‘Analects’, but in a late interpolation in which Confucius MSC yí) in the four quarters outside Chinese territory. 夏 says: Hsia ‘Chinese’ also occurs in other Warring States texts. The more specific expression 諸夏之國 ‘the states 夷狄之有君,不如諸夏之亡也。 of the Chinese (peoples)’ occurs twice in the Hsün tzu 荀子 (early to mid-3rd century bc),12 in which it is used The Ti 狄 and I 夷, who have a lord, are unlike the Chi- nese (諸夏), who do not have one.10 to explicitly distinguish the assemblage of Chinese states from the states of the foreign peoples of the four quarters The passage is ambiguous (it is either a positive or a outside Chinese territory. negative comparison), and the word 君 ‘lord, ruler, A synonymous term that occurs less frequently and in prince’ can be understood as either singular or plural. A far fewer texts, but is unambiguous, is Hua 華 MSC huá, frequent reading is that although the foreigners do have which as a common noun normally means, literally, a lord (ruler) or lords, they are not as good as the Chi- ‘flower’.13 Unlike most other texts, in which Hua 華 nese, who are without a lord or lords. Legge translates it: rarely occurs, in the 左傳 Tso chuan it occurs about as frequently as Hsia 夏, both as an explicit collective plural 5 Keightley (1999: 248) ends his discussion of the date of the (諸華 MSC zhū huá ‘the Chinese [peoples, states]’)14 or establishment of the subsequent fully historical dynasty, Shang, by not- alone,15 usually for the specific purpose of distinguishing ing that based on the dating of “astronomical events it is possible to conclude that the first year of Cheng Tang would have been 1554.” Mair (2013: 30), summarizing the present state of knowledge, says, 11 Tso chuan, Chao-kung 17. Texts and Legge from the Chinese “we cannot find any evidence that the word Xià in any of its various Text Project .) On the erroneous ancient variant of this senses, much less as the name of a dynasty or state, existed during the passage see Beckwith (2009: 74-75 n. 64). Shāng period. I have not even been able to ascertain that the word Xià 12 Hsün tzu 18.5; 18.14. occurs in the Western Zhōu B[ronze ]I[nscription]s in any of its later 13 Loewe (1999: 994) translates 諸華 (zhū Huá) as “the many senses. In any event, there is no evidence that it was employed during states blessed with elegance”. Mair (2013: 27) supports similar etymo- the Western Zhōu as the name of a dynasty that was supposed to have logical proposals, and concludes, “Xià was very much in evidence preceded the Shāng. Xià comes to be used as the name of an ancient during the Warring States period as a ubiquitous epithet (viz., ‘grand’) dynasty only in Warring States texts, a good thousand years after the for the peoples and cultures of the … Central Plains” (Mair 2013: 31). alleged Xià Dynasty”. Nevertheless, regardless of the possible etymological origins and mean- 6 It so occurs six times in the Tso chuan according to Serguei ings of the graphs as Chinese common nouns—hua 華 and hsia 夏, Zinine’s database . usually ‘flower’ and ‘summer’ respectively by Classical Chinese 7 Tso chuan V/21.5 (Legge 1893/1985: 179, 180), IX/13.3 (Legge times—in the proper name usage discussed here they are unquestion- 1893/1985: 456, 458b), IV/1.1-2 (Legge 1893/1985: 123-124). Loewe ably variant transcriptions of one and the same originally foreign epi- (1999: 993) translates 諸夏 zhū xià as “the many Xia”. thet, which appeared no earlier than the Warring States period and early 8 Tso chuan IV/1.1-2 (Legge 1893/1985: 123-124). Hua 華 is used became the autonym for ‘(the) Chinese’, as shown in this paper. On the in the same way. The even more explicit 諸夏之人 ‘the people(s) of later expression Hua-Hsia 華夏 see Note 15. the Chinese (states)’ also occurs in the Tso chuan (Ai-kung year 20) 14 In Chinese, as in Old Tibetan, unqualified ethnonyms do not but *諸華之人 does not occur. distinguish among ‘people, nation, state’, etc. in singular or plural; any 9 Tso chuan IX/13.3 (Legge 1893/1985: 456, 458b), which or all of these senses may be intended. recounts the defeat of the 諸夏 by the king of Ch’u. 15 It so occurs four times in the Tso chuan; see Serguei Zinine’s 10 Lun yü, 八佾 5. Brooks and Brooks (1998: 127) identify this database . In one instance in Tso chuan quotation as an interpolation dating to ca. 310 bc. The positive inter- IX/26.10 both 夏 and 華 are used together in a synonym compound, pretation is thus more likely. There are no passages with Hsia 夏 in the Hua-hsia 華夏 MSC huáxià ‘the Chinese (peoples)’, translated by sense ‘Chinese’ in the Analects. The few passages which do have the Legge (1872/1985: 521, 526b) as “the States [of the north].” It might word, in the sense ‘Hsia dynasty’, are dated by Brooks and Brooks be thought to suggest consciousness of the two dialect forms (v. infra), (1998) to 342 bc or later, well within the Warring States period. Cf. which could reflect two different parts of the region where the ‘­Chinese’ Note 69. as a whole then lived. However, this is the only occurrence of the

99263_JAS_2016-2_06_Beckwith.indd 232 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 233

Chinese from foreigners such as the Jung 戎, who are dialects or languages. This hypothesis is confirmed by the often explicitly plural (諸戎 MSC zhū róng ‘western for- hard data—the loanwords for ‘the Chinese, China’ in eigners [peoples, states]’).16 There are two instances neighboring foreign languages—as shown in detail below, where Hua 華 is used without the explicit plural marker following consideration and revision of reconstructions but the context clarifies that the plural is meant. The con- based on the traditional quasi-linguistic method.20 text of the first passage is a minister trying to convince The appellation Hsia 夏 MSC xià has two traditional his lord not to go off on a foolish military campaign Middle Chinese readings. In view of the unanimous for- against the Jung (peoples or states), arguing that this eign evidence for the final vowel *-a (see below), one would weaken his state’s dominant relationship with Middle Chinese reading may be reconstructed as EMC ✩ ✩ ✩ ✩ other Hua peoples or states, which the lord would then not γɨa2 ~ γia2 (Pul. 334 γaɨ2 / γɛː2); Bax. 796 hæx ✩ be able to aid. The minister says, “The Chinese (諸華 [ γæ2]) from a traditional Old Chinese *γraʁ (Bax. 796 zhū huá ‘Chinese [peoples, states]’) will certainly rebel. *g/Hraʔ [sic]; Sta. 562 *g(h)rāʔh [sic]; Sch. 663 *gəraʔ) 17 ✩ ✩ The Jung are just beasts. If we capture the Jung we will and the other reading as EMC γɨa3 ~ γia3 (Pul. 334 ✩ ✩ 21 lose the Chinese (Hua 華). Would it not be best to pre- γaɨ3 / γɛː3) from a traditional OChi *γrah ~ *γras. vent that?”18 The other example occurs in a famous pas- However, which Old Chinese “reading” underlies the sage spoken by a Jung lord contrasting the Chinese with Middle Chinese or modern readings as an appellation is his own peoples, the 諸戎 zhū róng (explicitly plural). He unknown, and it is also unknown which other readings says, might have escaped being recorded in the original Ch’ieh yün or the later rhyme books based on it. The unanimous 我諸戎飲食衣服,不與華同,贄幣不通,言語不達 。 testimony about the lack of a coda applies equally to 華 Our Jung peoples’ drinks, food, and clothing are different (discussed below). The latest research suggests that the from those of the Chinese (Hua 華), we do not exchange ‘rising tone’ of Middle Chinese developed in monosyl- ceremonial silk gifts, and our languages are not mutually labic morphemes that in Late Old Chinese had been intelligible.19 disyllabic and had ended in a vowel.22 The ancient and early medieval loanwords and tran- These examples show that by the fourth century bc scriptions into other languages all have a word-final open ethnic Chinese had distinguished themselves from the for- syllable with the vowel -a; none have a word-final coda. eigners they had come into increasingly intense contact Since Hua 華 MSC huá has no coda in Old Chinese even with, and had begun referring to themselves by a single according to the traditional reconstruction, and the tran- appellation. Because the term is written with two unre- scriptions 夏 and 華 occur interchangeably in Old Chi- lated characters that have no semantic connection to each nese texts, the underlying word transcribed by both char- other or to their inherited denotation, they undoubtedly acters in the sense ‘the Chinese’ (to be distinguished transcribe a foreign word, like many other well-known from other words with other meanings transcribed by the examples of transcriptions and loanwords written in same characters) did not, therefore, have a coda. Correct- this way. The two forms thus represent either different ing the reconstruction based on 華 and the external hard ­Chinese dialect pronunciations of it or independent bor- data for the vowel and against a coda, the word Hsia 夏 rowings of variant pronunciations of it in different foreign ‘Chinese’ may be provisionally reconstructed for early Warring States period Old Chinese as *γra. It is thus ­compound in the entire Tso chuan text; moreover, it does not occur in other pre-Han Dynasty texts, based on a search using the Chinese Text identical to the preliminary reconstruction of Hua 華 Project online as a database (downdating that site’s chrono- below, with the exception of the usually assumed labial logical attribution of the I-lin 易林 to the Western Han Dynasty period). feature. The Tso chuan example must therefore be removed from the corpus of genuine Warring States period forms. Hua-hsia 華夏 does become fairly common in the late Han Dynasty period, so its occurrence in the Tso chuan is clearly a late interpolation or a copyist’s mistake for 諸夏. This textual error is yet another example of the pressing need 20 Traditional HSR works are cited below. Recent HSR publica- for critical editions (q.v. Beckwith 2012a: 156-157) of the major pre- tions (e.g., Behr 2007; Baxter and Sagart 2014) have deviated increas- modern Chinese texts. ingly from earlier work, but still not in the direction of scientific com- 16 There are several examples of 諸戎 ‘the Jung (collective plural)’ parative-historical linguistics. in the text, notably in IX/14.1 (Legge 1893/1985: 460, 463-464). 21 CY 4 (departing tone) rhyme 37, CY 3 (rising tone) rhyme 32. 17 I.e., unlike the Chinese, they are foreign and thus hardly human. The putative *ʔ coda (corresponding to the Middle Chinese rising tone) 18 Tso chuan IX/4.6 (Legge 1893/1985: 422, 424a). of recent HSR Old Chinese reconstructions is impossible for that lan- 19 Tso chuan IX/14.1 (Legge 1893/1985: 460, 464a). Legge’s guage, as shown at length in Beckwith (2008b) and elsewhere; both translation reads: “Our drink, our food, our clothes are all different loanwords and internal evidence show it was an uvular *ʁ (like modern from those of the Flowery States; we do not interchange silks or other French and German r) or velar [ɣ], at least in some dialects (Kiyose articles of introduction with their courts; their language and ours do not and Beckwith 2008; Beckwith 2007a). admit of intercourse between us and them”. 22 Beckwith (2014); Kiyose and Beckwith (forthcoming).

99263_JAS_2016-2_06_Beckwith.indd 233 25/01/17 12:38 234 Christopher I. Beckwith

The character Hua 華 has two readings in the earliest 2 loanwords and transcriptions in preserved redaction of the Ch’ieh yün 切韻.23 Taking scientific linguistic reconstruction note again of the unanimous hard-data evidence for final *-a, 華 MSC huá ‘flower; Chinese’ may be traditionally It is normal, throughout the world, for people to bor- ✩ ✩ reconstructed as Middle Chinese γwɨa1 ~ γwia1 (Pul. row their names for their neighbors from their neighbors, 128: ✩γwaɨ / ✩γwɛː ; Bax. 763: “hwæ” [i.e., ✩γwæ])24 either directly or indirectly. It is extremely rare to invent from Old Chinese *γwra ~ *γrwa (Bax. 763: OChi *wra; an ethnonym for other nations (or even one’s own nation) Sta. 561: *wra; Sch. 243: *wǝra); 華 MSC huà ‘place using one’s own language, though it is common to name (not “China”)’ may be traditionally reconstructed folk-etymologize loanwords, including foreign ethno- ✩ ✩ ✩ as Middle Chinese γwɨa3 ~ γwia3 (Pul. 129: γwaɨ3 / nyms. Among the well-known ancient to early medieval ✩ 25 γwɛː3) from a Late Old Chinese *γwrah ~ *γrwah. Chinese names for historical Central Eurasian peoples, A corrected provisional Old Chinese reconstruction of none are verifiably pure Chinese creations; all are loan- both readings of Hua 華 would be *γwra- ~ *γrwa- or words or transcriptions. Because of the already plentiful the like. number of character variants even in Old Chinese— However, the hard data below, including examples of including those with insulting or humorous meanings— retention of disyllabic morphemes even in Late Old the Chinese sometimes deliberately chose homophonous ­Chinese,26 establish that Hua 華 had no coda (final con- characters with bad meanings for disliked foreigners, but sonant), and also that it was disyllabic, as was Hsia 夏,27 it is nevertheless demonstrable that they were transcrip- thus suggesting a preliminary traditional-based recon- tions first of all. The discovery of countless variant writ- struction OChi *γara ~ *γarwa, etc. Although determin- ings of Old Chinese words in the excavated ancient texts ing the phonetic value of the first syllable vowel as *a from the Warring States through the Han period has would normally be problematic, considerable external shown us that Chinese of the day wrote primarily accord- evidence exists in this case. However, the labial feature ing to pronunciation.28 Similarly, the names for ‘(the) attested from Middle Chinese in Hua 華 cannot be recon- Chinese; China’ in the languages spoken in Antiquity structed back to Old Chinese, as shown below. around the Chinese are borrowed names in those lan- guages: they are loans from Chinese, not etymological creations of the non-Chinese-speaking peoples.29 The foreign borrowings of the word that had become the Chinese autonym took place in the Old Chinese period, in Antiquity. The Old Chinese word for ‘Chinese’ was 23 CY 1/2 (even tone) rhyme 34 and CY 4 (departing tone) rhyme thus borrowed by Proto-Tibetan, as well as by Proto-­ 37. The original Ch’ieh yün (601) is lost. Although the 706 manuscript Japanese—perhaps when it was still spoken on the Asian used here (Lung 1968), an expanded version of the work, is the earliest continent—and by Proto-Puyo-Koguryo (Puyŏ-Koguryŏ), preserved redaction by several centuries, it is virtually never used in the 30 HSR system. As a traditionalist notes, “most Middle Chinese readings or possibly already by Common Japanese-Koguryoic. are taken from the Guǎngyùn, which is still the most convenient rhyme 廣韻 book to use” (Baxter 1992: 40), though the Kuang yün Guǎngyùn Old Tibetan rgya from pre-Old Tibetan *γaryá is dated 1007-1008. 24 Baxter’s transcription of Early Middle Chinese *γ (a voiced ‘China, Chinese’ onset) is “h”. It is doubly misleading because the voiced onset *γ- did become unvoiced h- in Late Middle Chinese and Mandarin, transcribed The Old Tibetan name Rgya རྒྱ་ ‘China, Chinese (peo- by him also as “h”. This later reading of 華 is recorded precisely in ple), Chinese (adjective)’ seems never to have been a Tibetan segmental script as hwa (Tak. 308), reflecting LMC [hwa] ~ mystery to anyone. Surprisingly, Jäschke does not even [χwa], with the expected unvoiced onset. It is clear from the Ch’ieh give the form Rgya ‘China’, which is the name for the yün, as well as transcriptional and loan evidence, that the final vowel was a (Beckwith 2008b), and remains [a] in Mandarin, and also that country in most Classical Tibetan literature, and exclu- there was a palatal element in the Old Chinese word onset, as shown sively in Early Old Tibetan (which language was not yet below. known in Jäschke’s day). He gives only Rgyanag, a very 25 The earlier consonantal reflex of the attested Early Middle late name, which he explains as meaning “the ‘black ­Chinese tone is thought to have been *-s. Baxter (1992: 763) has two extent’.” He similarly gives the etymology of Rgyagar readings for Hua 華: MSC huā < MChi ✩xwæ (i.e., ✩[χwæ]—cib) < OChi *hwra and MSC huá < MChi ✩hwæ (i.e., ✩[γwæ]—cib) < OChi ‘India’ as “the ‘white extent or plain’,” taking rgya- as *wra. However, the first equates 華 with 花 huā, MChi ✩xwæ (i.e., the root of rgyaba ‘to be wide, extensive; width, extent’, ✩[χwæ]—cib) in his system, which is listed in the CY in the same rhyme as 華 huá, ‘flower’. As common nouns, 華 and 花 are now syn- onyms meaning ‘flower’. 28 See e.g. Pai (2012). 26 Beckwith (2014). 29 Beckwith (2005a; 2006/2007; 2009: 355-362, 375-383) and 27 Cf. Schuessler’s (1987: 243) 華 OChi *wǝra and his 夏 OChi Kiyose and Beckwith (forthcoming). *gəraʔ (Sch. 663). 30 Beckwith (2007a).

99263_JAS_2016-2_06_Beckwith.indd 234 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 235

etc., and -gar as a form of the Old Tibetan root dkar- of Tibetans moved from the Ch’inghai-Kansu Sprachbund the adjective dkarpo ‘white’.31 The etymologies seem so area to south Central Tibet, from which region they even- semantically reasonable that they have never been seri- tually (from about the late sixth century on) established ously questioned by anyone, including other dictionary the Tibetan Empire. makers, with the exception of Rgyagar ‘India’, which is While the syllable rgya in these names is thus not never spelled *rgya dkar, nor is dkar- ‘white’ ever analyzable semantically or etymologically within spelled *gar-, so the traditional explanation has rightly Tibetan, it certainly can be analyzed phonologically. Like long been suspected to be a folk etymology.32 However, many native Tibetan words, it underwent regular metath- in Old Tibetan texts, the only name for ‘China, Chinese’ esis in pre-Old Tibetan, in this case from either *grya ~ is Rgya, just as the only name for ‘India, Indian’ is *γrya or *gyra ~ *γyra, both of which contain disallowed Rgyagar. Like the Old Tibetan names of all other peoples onset clusters.35 Old Tibetan texts record foreign words and states of the day (including Bod ‘Tibet’ itself),33 the with a voiced velar fricative onset using one of two pos- Tibetan word Rgya ‘Chinese, China’ is simply a name. sible transcriptions: the Tibetan letter ག , which in As it is a foreign name for a foreign country and people, Early Old Tibetan had at least two allophonic values, [g] it cannot have a genuine Tibetan-internal etymology, and and [ɣ]; or the Tibetan letter འ <ḥ>,36 which in Early Old all the etymologies proposed so far, whether by Tibetans Tibetan had the value [ɣ].37 The name Rgya is thus sim- or others, are folk etymologies.34 Chronologically the ply a borrowing, direct or indirect, of a Late Old Chinese word must have been borrowed after the meaning of the pronunciation of an actual Chinese name for ‘China’,38 word within Chinese had shifted from an epithet to an the earlier form of which within Chinese itself could the- ethnonym meaning ‘the Chinese’ but before the Proto-­ oretically have been *gVrya ~ *γVrya or *gVyra ~ *γVyra.39 As with the reconstruction of the Old Chinese 31 Jäschke (1881: 105-106). He also gives Rgyaser ‘Russia’—a form, the specific vowel of the first syllable of *gVrya ~ modern neologism—without explanation, probably because it is easily *γVrya is not determinable purely from the Tibetan loan explainable, in the same way, as ‘the yellow [ser] extent or plain’. But form alone, but it is fortuitously well attested in the Chi- rgya never means ‘plain’ in Early Old Tibetan, and the derivation of the name Rgya(-) from rgyaba is simply an old folk etymology. nese, Japanese, and Koguryo transcriptions of the variant 32 Unfortunately, the misinformation is still regularly cited in form Kara, which reflects *γara, from *γarya, and in the scholarly works. On the phonemic distinction that explains the lack of Old Tibetan transcriptions of the T’u-yü-hun forms of the such a misspelling in Early Old Tibetan see Beckwith (2006a). word (see below), solidly giving us Proto-Tibetan *garyá 33 Beckwith (2005a). Note that the Old Tibetan name Spu is also a ~ *γaryá.40 Middle Chinese shows that the onset was the name, and as a name hardly means ‘body-hair’ or the like, despite countless folk etymologies. 34 Behr (2007: 736-737), in a lengthy analysis of the Chinese word Hsia 夏, reconstructed by him as “*N-kkra-q/s”, discusses a 35 For the phonology of Old Tibetan syllable onsets and codas see relationship with “written Tibetan” rgya in the meaning ‘extended, Beckwith (2006a). plain,’ etc., which “is well-attested in the exonyms for China, India, 36 The letter འ is today most frequently—but misleadingly for Old , etc., in Tibetan.” He makes no reference to Old Tibetan, in Tibetan—transcribed as an apostrophe < ’ >. which language rgya means only ‘China, Chinese’, and cites only 37 For example, the Old Tibetan Annals (line 133) records the Jäschke (1881) and a dictionary of Modern Tibetan. However, the ­Chinese word ho 河 MSC hé < MChi ✩γa ‘river’ (Pul. 122) as ག་ , word Rgya ‘China, Chinese’ occurs frequently in the Old Tibetan representing [ɣa] in the place name Ho chou 河州 (Beckwith 1993: 63) Annals and the Zhol Inscription (ca. 764 ce)—i.e., in early Imperial and in other foreign words, such as the name Uighur ~ Uyghur—Old Old Tibetan (Beckwith and Walter 2016)—so it is firmly attested for Turkic [huyɣur]—recorded in Late Old Tibetan as Hoyoḥor ཧོ་ཡོ་འོར་ almost a millennium and a half. Arguing that modern dictionaries also [hʊyʊɣʊr] (PT 1283 line 558) as well as Hoyohor ཧོ་ཡོ་ཧོར་ [hʊyʊhʊr]; define the Tibetan word as ‘beard’, so the word for ‘China’ might see Ligeti (1971). mean ‘the bearded ones’, he cites a Chinese internal “Bezug zu 胡 38 For studies of early East Asian and Central Eurasian names and hú < *hu < *kka ‘Bart’,” a quotation from the Huai nan tzu, and epithets and their Chinese transcriptions, see inter alia Beckwith comparisons with the names of ancient Germanic peoples. He then (2005a; 2007a; 2014). discusses the proposal of Chin Li-hsin that the complex onset rgyV- in 39 Early Old Tibetan had no zero-onset words; an onset (conso- Tibetan is the result of secondary epenthesis of “rgyV- < *s-yV-” nant) was required. Even the glottal stop does not occur in native Early corresponding to Old Chinese “*l(l)-”, and cites several comparisons Old Tibetan words; the letter for it is used strictly to transcribe foreign of “Written Tibetan” and reconstructed Old Chinese to buttress Chin’s words (Beckwith 2006a). Many errors exist in the literature about Early idea. In short, Behr etymologizes the names Hsia and Rgya and argues Old Tibetan phonology because of the widespread lack of awareness of that they ultimately go back to the same ‘Proto-Sino-Tibetan’ common the importance of dating texts, and of critical edition, q.v. Note 15. noun meaning ‘beard’. Yet this would require the respective inherited 40 Less possibly, *gyVra ~ *γyVra, because palatalized onsets such words for ‘beard’ to be fortuitously (and equally oddly) chosen thou- as *gy and *γy are both generally thought not to have existed in Old sands of years later by both Old Chinese speakers and Proto-Tibetan Chinese. The other theoretically possible reconstructions, *gayra ~ speakers as their respective names for ‘Chinese, China’, among other *γayra, with the first syllable peak *ay (and similarly, forms with the difficulties. Also not with Mair (2013: 9-11), who follows Behr long vowel *ā) can be ruled out as sources of the loanform in Tibetan (2007), and a number of Tibetologists who cite the usual folk etymol- because the syllable length would have attracted the accent and caused ogies. Hill, quoted by Mair (2013: 13), rightly throws cold water on deletion of the final vowel instead. They would most probably have the ‘beard’ idea. become pre-OTib *ger or the like.

99263_JAS_2016-2_06_Beckwith.indd 235 25/01/17 12:38 236 Christopher I. Beckwith

fricative *γ, not the stop *g,41 as shown also by the Early Kara (written ✩kala [可羅], ✩kalak [加烙], etc.47), in Old Tibetan transcription of the name or epithet of the both Chinese and Japanese sources.48 Early Japanese-­ T’u-yü-hun royal clan, Ḥaźa འ་ཞ་ [ɣaʑa], from a loan- Koguryoic languages did not have a phonemic distinction form *[ɣa.rya] ~ *[ɣa.lya].42 In view of the earlier loan between voiced and unvoiced stops in word-initial posi- of the same word, Rgya, from *γarya (with *r instead of tion, so Old Japanese and Old Koguryo words are written *l), the possible ancestors of the T’u-yü-hun donor form indiscriminately with Chinese characters pronounced in to Tibetan are reduced to one: *γarya [ɣa.rya], as actu- Middle Chinese with word-initial voiced stops, unvoiced ally attested in later direct transcriptions, q.v. further stops, voiced fricatives, or unvoiced fricatives. Thus the below. well-attested Old Japanese word Kara ‘China’—which like the Old Tibetan word for ‘China’ is a loan from Old Japanese and Old Koguryo Kara from Old Chinese (not Middle Chinese)—could reflect an χ γ Old Chinese dialect *γara ‘China, Chinese’ Old Chinese donor *kara, *gara, * ara, or * ara plus any number of additional Chinese phones lost in transmis- Common Japanese-Koguryoic, the ancestor of Old sion. Japanese and Old Koguryo, is now known to have been Comparing the possible Old Chinese reconstructions spoken on the Asian mainland in the vicinity of the Old based on the Old Tibetan form Rgya ‘China, Chinese’ Chinese speaking area. It may have been spoken further with those based on the Old Japanese form Kara ‘China, to the south in earlier times, but speakers of both branches Chinese’, gives *γarya ‘China, Chinese’ (which is faith- are firmly attested in late Antiquity in the area of Liao- ful to the Old Tibetan data) or *γara ‘China, Chinese’ hsi (in or slightly further west of what is now Liaoning (which is faithful to the Japanese and Korean area data). Province), which is thus the proximal homeland of the These names clearly represent forms of one and the same attested language family.43 The Old Japanese name for spoken Old Chinese word. Because early Tibeto-Burman ‘China (sometimes including )’ is Kara (written languages, including Proto-Tibetan and Old Tibetan, had ✩kala [可羅] ~ [伽羅], ✩kalaŋ [可良],44 etc.). The Old complex syllable onsets as well as simple onsets (both Koguryo form of the same name is also Kara (written CCV and CV forms are reconstructable to early stages of ✩kalak [加烙]) ‘T’ang [唐] China’.45 In addition, the Tibeto-Burman), it is understandable that the Old Tibetan principalities in the south central coast region of the form preserves several Old Chinese phonetic constituents Korean Peninsula for much of the Korean ‘Three King- of the name in its onset. By contrast, native Old Japanese doms’ period (from late Antiquity to the Early Middle words do not allow clusters in any position, i.e., in Ages) were known collectively as Mimana 任那46 or word-initial or word-final position, or intervocalically between two syllables. Old Japanese and Old Koguryo both tend to preserve foreign disyllabic forms, so their 41 When MChi [ɣ] became devoiced [h] ~ [χ] in Late Middle form of the name for ‘China, Chinese’49 preserves the ­Chinese, it came to be regularly transcribed as ཧ , as seen by many examples in Takata (1988). vowels of the donor form, but would thus seem to have 42 A donor form *[ɣar.ya] ~ *[ɣal.ya] is acceptable in Early Old simplified its syllable structure from an alien and disal- Tibetan, and would have been borrowed as *འར་ཡ་ <ḥar.ya> ~ *འལ་ཡ་ lowed sequence *Ca.CCa [ɣa.rya]50 to *CaCa [ɣa.ra]— <ḥal.ya>. The syllable boundary also supports the explanation of the pronounced in Japanese [ka.ra]—to accord with the pho- γ depalatalization of * arya *[ɣa.rya] to attested kara in Japanese-Kogu- notactic constraints of Pre-Old Japanese and the related ryoic. 43 Kiyose and Beckwith (2008). See Beckwith (2007a) on Old Archaic Puyo-Koguryo languages that dominated the Koguryo, the relationship of the Puyo-Koguryoic languages to the entire region until well into the Middle Ages. However, ­Japanese-Ryukyuan languages, their possible relationships to other the Old Chinese donor was in both cases necessarily a ­languages, and archaic northeastern Middle Chinese. Northeastern dialect, which has already been shown to 44 Omodaka et al. (1967: 228 ff.); cf. Martin (1987: 438). 45 In the Samguk Sagi it occurs in only one attestation, written [加火] ✩kaχwa ‘T’ang 唐 (i.e., China)’, but the second character is a graphic error for 烙 MChi ✩lak (the right side of the character is miss- ing), as is clear because 烙 is one of the forms of the second syllable 47 See Note 45. of the name Kara (Beckwith 2007a: 74) in the Koryŏ Sa and elsewhere. 48 The Chinese clearly did not recognize that Kara was the local The transcription in the Koguryo materials discussed in Beckwith pronunciation of the same word they used as their own name, ‘China, (2007a) must therefore be corrected to read Kara [加烙] ✩kalak Chinese’, because over time the respective pronunciations of the word ‘China’. Old Japanese did not allow final syllable codas, so the final had changed beyond recognition. They thus reborrowed Kara as a new ✩k, ✩ŋ, etc. in the Middle Chinese-based transcriptions of this word foreign word from one of the local languages. were not pronounced in Japanese. The same may be true of Koguryo 49 Old Koguryo ✩Kara is apparently a literary holdover from an too—there are no attested disyllabic words ending in a stop (the only identical Archaic Koguryo form, like some other Old Koguryo words attested codas in disyllabic words are nasals and ✩r)—but further (Beckwith 2007a). research is needed. 50 As noted, the Old Tibetan loanform of the T’u-yü-hun epithet, 46 LOC *mī́ma-nà (Kiyose and Beckwith, forthcoming). The mod- as well as the direct transcription discussed below, rules out the possi- ern Sino-Korean reading is Imna. bility *CaC.Ca (*γar.ya).

99263_JAS_2016-2_06_Beckwith.indd 236 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 237

have depalatalized many palatalized onsets.51 This fact, from a length contrast within Old Chinese, with Hua 華 plus the attested T’u-yü-hun forms *γarya ~ γarya- deriving from OChi *γāryá and Hsia 夏 from OChi [ɣa.rya] and other material,52 indicates that the steppe *γaryá.56 Either form would have been borrowed by Pro- zone languages long retained the complex second-sylla- to-Japanese, pre-Old Japanese, or Archaic Koguryo ble onset, and that the simplification of the cluster in speakers with a devoiced onset and without the palatal Northeast Asia apparently occurred within Chinese. The feature *y, giving in both cases *kara, the attested Old Japanese-Koguryoic data is thus faithful to a Chinese Japanese and Old Koguryo form, as discussed above.57 dialect donor form *γara that derives in turn from earlier However, the speakers of the latter two languages cer- *γarya. tainly borrowed a spoken Chinese dialect form of the Hsia 夏 and Hua 華 rhymed, both in attested Late name (regardless of how it was written), and only much Middle Chinese and in Old Chinese, with a word-final later wrote it phonetically in Old Japanese and Old *a, and both had a voiced velar fricative onset, written Koguryo, so it seems not to be possible to determine the here *γ [ɣ]. The only difference in attested Late Middle donor dialect category—i.e., whether *γārá or *γará.58 Chinese between their monosyllabic forms is that Hsia Both the y and the r in Old Tibetan Rgya, a loan from 夏 has a non-labialized velar onset, whereas the velar Old Chinese *γarya, are specifically Old Chinese in onset of Hua 華 is labialized. The voiced onsets of both date—as is the entire form, which could not have been words underwent canonical devoicing in Late Middle borrowed from Middle Chinese. If a word *γarya was Chinese times, becoming attested h- [h] ~ [χ].53 The fact borrowed as a loanword into early Old Tibetan in Middle that Hua 華 has retained its labiovelar feature from Mid- Chinese times, it would have been written འ་ཞ་ <ḥaźa> dle Chinese right down to the present explains why the (or ག་ཞ་ ), and in fact precisely such a form is so onset was not palatalized in Mandarin, unlike the onset written, as discussed above. It is arguable that the Old of Hsia 夏, which was palatalized in New Mandarin Chinese dialect which transmitted the word to Proto-­ times. Tibetan could already have undergone monosyllabiciza- There is no information that would indicate whether tion within Old Chinese, producing *γrya. However, the the labial feature of Hua 華 is older than Middle ­Chinese, onsets of many native Tibetan words also underwent and it is not even certain that the feature existed in all metathesis well before Old Tibetan times,59 and the Old dialects of Middle Chinese. In Old Chinese the sequence Tibetan form retains the *r of Old Chinese *γarya ~ * appears to have alternated with *o at different times *γrya, a marker of a specifically Old Chinese loan, so the in different dialects (and perhaps in early Middle Chi- word should have been borrowed into Tibetan in late nese dialects as well),54 so the segment *w might be Antiquity, before the more or less final monosyllabiciza- viewed as an artifact of a transcriptional form with the tion of Chinese morphemes that took place in the third to vowel *o, giving an alternate reconstruction of Hua 華 fourth centuries ce, at which time formerly distinct disyl- as Late Old Chinese *γora.55 However, in the fourth cen- labic and monosyllabic morphemes merged phonetically tury bc, in the dialects spoken by the authors of the Tso chuan and other Classical texts, the word written Hua 華 is clearly just a dialect variant of the same word that is 56 Loss of the first vowel did not make the two words total hom- also written Hsia 夏, because the two characters are used onyms because the long vowel in Hua 華 OChi *γārya would seem to interchangeably (and absolutely the same semantically) have followed the regular shift from Old Chinese *ā ~ *â to Middle Chinese ✩ɔ ~ ✩ʊ, which then underwent breaking, perhaps during the to transcribe the word for ‘the ­Chinese’, as shown above. monosyllabicization period, producing the labial feature in the Middle The Middle Chinese labialization thus seems to be a Chinese word’s onset. However, my assumption is that breaking later development best explained as having originated occurred with long vowels, so perhaps Pulleyblank’s (1991) reconstruc- tion of Middle Chinese with a length contrast is right and the rounded Middle Chinese vowel was long. Cf. the problematic first vowel of the 51 Beckwith (2007a). The already palatalized onsets continued to Old Japanese form of tora ‘tiger’ (Kiyose and Beckwith 2008). be palatalized in the Central dialect and eventually developed a phone- 57 These languages did not have phonemic vowel length. As noted, mically distinctive retroflex articulation in the standard language. the *y is not attested either in the Middle Chinese readings borrowed 52 See below on the word Hu 胡, one of the most frequent Late Old by the Japanese along with the characters in the Early Middle Ages, or Chinese words for foreigners of the north and west, including the in the Old Tibetan transcriptions of Middle Chinese, but these borrow- ­Hsiung-nu. ings and transcriptions are both over a millennium later than the 53 Takata (1988: 308-309, 華 hwa; 1993: 320 [line 132] and many ­Warring States period Old Chinese word. other instances) gives the Old Tibetan transcription of 華 as ’hwa (i.e., 58 For preliminary attempts to characterize the major Old Chinese [ɦhwa] ~ [ɣhwa]); 夏 is unattested in either corpus. dialects see Beckwith (2006b, 2008b). 54 This little-discussed question is still very far from resolution, 59 Old Tibetan is first attested in Tibetan script in 649-650 ce, the despite much belief to the contrary. For detailed discussion of dialect date of the first contemporaneous entry in the Old Tibetan Annals, the forms of one Early Old Chinese word that exemplifies the problem see unique manuscript copy of which is thought to date to the ninth century Beckwith (2006b). (K. Iwao, p.c., 2013). The Zhol Inscription, ca. 764 ce (Walter and 55 The other choice, *γaro, is not possible because of the unani- Beckwith 2010) is the earliest datable text. The word Rgya ‘China, mous testimony that the final vowel was *a. Chinese’ occurs in both texts, as does the word ḥaźa ‘T’u-yü-hun’.

99263_JAS_2016-2_06_Beckwith.indd 237 25/01/17 12:38 238 Christopher I. Beckwith

in the Central dialects of Chinese,60 and any remaining and transcriptions without a palatal y in their onsets) or Old Chinese consonant clusters were simplified. This later palatalizations or depalatalizatons within Chinese. accords well with historical data on the Later Han In sum, the early loanwords (OTib Rgya, OJpn ✩Kara, Dynasty period migration of early Ch’iang 羌 or ­Tibeto- OKog ✩Kara) and the Old Chinese transcriptions Burman peoples from the Ch’inghai-Kansu region south- (Hsia 夏~ Hua 華)64 and the reconstructions of their Old westward into the Tibetan Plateau proper, where they ­Chinese forms, all attest to a spoken word *γaryá ~ are said to have taken refuge with the Fa Ch’iang 發羌, *γāryá ‘Chinese, China’, which came into use shortly a people whose Chinese-transcribed name Fa 發 has before the fourth century bc in the Classical period of been identified onomastically with Bod ‘Tibet’.61 It is Old Chinese, when it is first attested. unknown if the Proto-Tibetan speakers borrowed the word directly from the Old Chinese speakers, or via an intermediary such as the T’u-yü-hun (see below), but 3 old Chinese *Ḥarya and Chung-kuo 中國 whether or not the distinction between the two dialect the ‘Middle Realm’ forms within Chinese was still purely one of vowel length at the time of the original borrowing (into The word Hsia 夏 *γaryá ~ Hua 華 *γāryá is the ­Proto-Tibetan directly or into an intermediary language), ­earliest known self-appellation of the dominant people or it appears that the Tibetan loanword must be from the peoples of the Warring States, who use it in texts from Old Chinese dialect with a short vowel, namely Hsia 夏 that period to distinguish themselves, ‘the Chinese’, from OChi *γaryá, because the long vowel *ā of the other others near them who had different cultures or languages dialect form, Hua 華 OChi *γāryá would have attracted and were called by different names. The appellation the accent in Tibetan, making the presumed loanform appears in Chinese at exactly the same time as the appear- *γārya and resulting in loss of the second vowel rather ance of the name Chung-kuo 中國 MSC zhōngguó, than the first vowel during monosyllabicization. The a purely Chinese term that means ‘Middle Realm’, ­Proto-Tibetan form was thus borrowed from the short- ‘­Central Region’, or the like and eventually came to be vowel dialect form, Hsia 夏 OChi *γarya ~ *γaryá, not interpreted as meaning ‘the Middle Kingdom’. In all 華 OChi *γārya ~ *γāryá.62 three of its earliest attestations (according to traditional The traditional Early Middle Chinese nucleus catego- dating), which are in the Shih ching 詩經 ‘Book of rization of the unitary rhyme that includes the Early Songs’, the reference of 中國 is unambiguously singular, ­Middle Chinese readings of both Hsia 夏 and Hua 華— not plural. In the 孟子 Mencius and the Tso chuan 左傳, namely, Pulleyblank’s ✩aɨ ~ ✩ɛː and Baxter’s ✩æ63— two of the most important Warring States political texts, reflects an underlying palatal feature somewhere in that the contextual number reference of 中國 is often ambig- rhyme, but the attested final vowel of the word from uous,65 but sometimes it is explicitly singular, whereas it Antiquity to the present is /a/, falsifying the HSR recon- is never explicitly plural. That means the translation ‘the structions. The palatal y present in the onset of the Old Central States’ or the like, popular among scholars since Tibetan form also must represent a palatal feature of Old Legge’s time, is specifically ruled out by the data. Early Chinese, and has nothing to do with Middle Chinese (in Chinese speakers thus simultaneously developed a unify- which the onsets of both words are attested in loanwords ing appellation for themselves which later became their ethnonym, and a geographical appellation meaning 60 Beckwith (2014), Kiyose and Beckwith (forthcoming). ‘­Middle Realm’ that later became their national name, 61 Beckwith (1993: 7). even though they were not a united people and did not 62 The vowel length distinction between dialect forms of the same have a united nation or state. The fact that they adopted word—the example of Hua 華 and Hsia 夏 is one of many—was a such unprecedented terms and ideas reflects the recogni- typical distinctive feature of Chinese from Early Old Chinese times at least until the monosyllabicization (and other major changes) that pro- tion at the time that there was no united kingdom or 66 duced Proto-Mandarin, the spoken language underlying Early Middle empire, but there should have been one. Chinese (Beckwith 2014; Kiyose and Beckwith forthcoming); cf. Beckwith (2006b, 2008b). 63 This is not to say, however, that the HSR vowels are correct. As 64 V. infra on Hu 胡 MSC hú < OChi dial. *γará < *γaryá, and on noted, the consistent quality of the final vowel in all attested forms of འ་讱་་་ γarya-. both words, /a/, right down to the present, is remarkable and important. 65 For those unacquainted with Classical Chinese grammar, nouns The point is that the HSR system, which is wholistic in nature, indicates themselves are not marked overtly for number, which must be estab- a palatal feature in the syllables included in the rhyme in question. lished by an explicit pluralizing marker—such as 諸 zhū, another Baxter (1992) is consistent in reconstructing all examples (as far as ­quantifier, or a phoronym, in Old Chinese—or by context. For a critical I have noticed) from the respective rhyme as ✩æ. Pulleyblank (1991) study of number and phoronyms (classifiers, measures, etc.) in Modern divides the unitary Early Middle Chinese rhyme, as given in the attested Standard Chinese see Beckwith (2007b). redaction of the Ch’ieh yün text (see Notes 21 and 23), into at least two 66 As shown in this paper, the adoption of these terms and concepts rhymes, in two categories: one category consisting of an ‘earlier’ did not happen in isolation. It must be stressed that according to the rhyme ✩aɨ and a ‘later’ rhyme ✩ɛː. Chinese historical record what became the Chinese culture zone was

99263_JAS_2016-2_06_Beckwith.indd 238 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 239

The words Hsia ~ Hua ‘Chinese’ and Chung-kuo do However, the native appellation of Bactrian—the lan- not occur in the Ch’un ch’iu ‘Spring and Autumn guage of Bactria, or Tokharistan, the core of the later Annals’, a genuine historical text that ends in 479 bc, the Kushan Empire—is now known from the Rabatak year of the death of Confucius.67 That provides a clear Inscription of Kanishka (ca. 115 – 128 ce)72 to have terminus ante quem for the appearance of the terms. been Αρια Aria, i.e., Arya (or Ārya), because in the They do of course appear in the Tso chuan, as discussed inscription the term Aria is used explicitly to refer to the above, but it is actually a fourth century text (ca. 390 to Bactrian language, in contradistinction to the Greek ca. 312 bc).68 In the Analects (Lun yü) the word Hsia language.73 This corresponds to Old Persian Ariya [arya] occurs four times in the sense ‘the dynasty before the and Eastern Old Indic Ārya, which are also used in native Shang’ (q.v. infra), and one time, explicitly plural, in the works to refer to the respective languages (not scripts), sense ‘the Chinese’, as noted above, but all of these are as well as to the respective peoples.74 All three agree in later Warring States period additions to the text, contem- turn with the reconstruction of the early Chinese auto- poraneous with the Tso Chuan and other works from that nym *Ḥarya, written Hsia 夏 or Hua 華. They also agree period.69 These terms and the concepts behind them are with the Old Tibetan term Rgyagar ‘India, Indian’, from thus innovations of the Warring States period Chinese. an earlier *Ḥaryaḥwar or the like, a regular loanform, However, the term *Ḥarya cannot be a Chinese word via Old ­Chinese, of an Old Indic *āryāwar (Sanskrit etymologically. It must be a loanword, as shown by its Āryāvarta) ‘home of the Āryas, India’, as discussed two strictly phonetic transcriptions, Hsia 夏 *γaryá ~ below.75 Hua 華 *γāryá, as well as by other data given below. The word Ta 大 MSC dà ‘big, great’ in premodern Chinese texts is used as an epithet marking large dynastic realms as ‘great kingdoms’, or ‘empires’. The T’ang 4 ta Hsia 大夏 Dynasty realm is called Ta T’ang 大唐 ‘Great T’ang’, ‘Great ĀRYA, the ĀRYA Empire; Bactria’ the empire is called Ta Ming 大明 ‘Great Ming’, and so forth. The usage is regularly applied to Hsia 夏 is also well known as the significant part of foreign peoples, including the Ta Yüeh-chih 大月氏 the traditional Chinese name of Bactria, Ta Hsia 大夏 ‘Great Yüeh-chih,’ i.e., ‘the Yüeh-chih Empire’. Chinese ‘Great Hsia’, an important Central Asian country in Ta Hsia 大夏 thus corresponds literally to ‘Great Ārya’, Antiquity and the Middle Ages that is mentioned in or ‘the Great Āryas’, but it actually signifies ‘the Ārya ­Chinese historical records from the late second century Empire’, and in this particular instance refers explicitly bc onward.70 This name has often been thought to repre- to Bactria or a realm that included it. sent the application of one of the Chinese people’s own However, it has been generally overlooked that a names for themselves, Hsia 夏, to a foreign nation, which ­century before the first firm attestation of the term Ta has accordingly been etymologized as ‘Great China’.71 Hsia in the sense ‘Bactria’, it occurs in the Lang-yeh Inscription of Ch’in shih huang ti, the ‘First Emperor’, never isolated, from the beginnings of its written history during the erected in 219 bc, the second year of the unified Chinese Shang Dynasty, ca. 1200 bc, down to modern times (Beckwith 2009). 67 Brooks and Brooks (2007). 68 Brooks and Brooks (2007). Hsia 夏 in Ta Hsia 大夏, a transcription of a foreign word (cf. 69 The one passage in which Hsia ‘the Chinese’ occurs is a late ­Beckwith 2010b), though the choice of transcribed characters was no interpolation (Brooks and Brooks 1998: 127), as discussed above. doubt ­modeled on the already existing folk etymology of the much Brooks and Brooks (2007) note that the Analects (LY) dates to ca. earlier Ta Hsia 大夏 . “0479-0249. The text is accretional; it consists of a core of genuine 72 Robert Bracey (n.d.) . by a long tail of material deriving from the successor Confucian School 73 Sims-Williams (1997). Both languages, Greek and Bactrian, are of Lu. Those late and invented sayings interact with other advocacy written in Greek script, so it is clear that the Bactrian language is texts known to be from the 04th century (the early Gwandz; LY 12-13) meant, not the script it is written in. and the early and middle 03rd century (Sywndz and Jwangdz; LY 74 It seems that the expression 夏言 ‘Chinese’ is more commonly 17-19).” Chung-kuo appears three times—but not Hsia or Hua—in used than 華言 ‘Chinese’ in early texts relating to northern peoples. the Shih ching, which is dated to “c0460-c0320. … Some late additions The account of the Türk in the Chou shu, an early medieval text, has to the Shr repertoire created a classical sanction for ideas which were 夏言. The Pei shih account of Paekche also has 夏言. The phrase new in Warring States political theory (such as the universality of rule 華言 seems to be later—e.g., it appears in the Ch’i-tan kuo chih in in Shr 205). By the late 04th century, the repertoire had been standard- reference to Sung Chinese (p.c., A.E. Shimunek, 2015). This should be ized at 300 poems”. (Ibid.) investigated very carefully. 70 Shih chi 123. 75 Āryāvarta is regularly explained (Monier-Williams 1899: 152) 71 Although this peculiar idea should surely have been questioned, as referring to a particular region of northern India, but the texts cited it seems to be supported by a later parallel, Ta Ch’in 大秦 ‘Great are later than the probable date of the loanword into Proto-Tibetan— Ch’in’, the traditional name for the Roman Empire from the Han period perhaps the first or second century ce, the early Eastern Han period, on, which was similarly folk-etymologized as meaning ‘Great China’. when Normative Buddhism is attested in China (Zürcher 2007: But the name Ch’in 秦 in this name is undoubtedly at base like 23 ff.).

99263_JAS_2016-2_06_Beckwith.indd 239 25/01/17 12:38 240 Christopher I. Beckwith

realm, the Ch’in Empire.76 In the text, the name Ta Hsia speaking Scythian-Saka peoples, who were in the process 大夏 unambiguously refers to the major geographical of being replaced politically in the area of the Ordos and region to the north of the Chinese in the Central Eurasian Mongolia by the Hsiung-nu.81 This is confirmed in four steppe zone. The text gives the extent of the ruler’s ways by the Chinese sources. ­conquests as follows: First, the earliest datable reference to Ta Hsia 大夏 ‘Great *Ḥarya ~ ārya’ in the sense ‘Hsia dynasty’ is in 皇帝之土 西涉流沙 南盡北戶 東有東海  the Tso chuan, in a legend projected into highest antiquity 北過大夏  but actually dating to the fourth century like the rest of ‘The land of the emperor:/ in the west it crosses over the the text. In it, the mythical Chinese culture-founder Yao flowing sands/ in the south it goes all the way to the end is said to have separated two quarreling brothers, sending of those whose doors face north/ in the east it includes the one of them to Ta Hsia 大夏,82 which is identified in the Eastern Sea/ in the north it extends across Great *Ḥarya 77 commentarial tradition with the northerly region of what (Ta Hsia).’ is now Shanxi province83—i.e., the region occupied by Because of its explicit location, it is accepted that much of the state of Chao in the Warring States period. Ta Hsia in this particular text does not refer to Bactria; This story is one of the most important passages in the but it also does not refer solely to “the area between the Tso chuan, as it legitimizes the target region as a part of Yellow River and the Fen 汾 River in present western the Chinese world, even though at the time it was still Shanhsi 山西”, i.e., part of the former state of Chao, the mostly non-Chinese in culture, and the rulers themselves First Emperor’s birthplace, though it necessarily included were at least partly non-Chinese in origin. that area.78 The text specifies that the locations men- Second, despite the Shih chi’s reputation for vague or tioned are all conquests lying beyond the home territory misleading attributions of foreign peoples’ origins, the of the Chinese speaking peoples, and all are suitably vast very beginning of its famous account of the Hsiung-nu geographical regions, not minor realms or provinces. states, “As for the Hsiung-nu, their progenitor was a 84 Ta Hsia in the inscription thus refers to a generic ethno- descendant of the lineage of the rulers of Hsia 夏.” geographical region corresponding to one of the four Taken in the traditional understanding, i.e., the legendary ­cardinal directions around the Central Realm of the first dynasty of “China” in later Chinese historiography, ­Chinese. They are specifically non-Chinese regions, so it is of course not literally true and makes little sense. But Ta Hsia ‘Great *Ḥarya’ here refers to the Northern it is correct from the viewpoint of the early myth of the Steppe zone79 of Central Eurasia—so to say, “Central Hsia Dynasty created in the Warring States period, since Eurasia par excellence”. In 219 bc, Ch’in shih huang ti the previous rulers of the Hsiung-nu home territory must had not yet sent his general Meng T’ien north to attack be the Ta Hsia ‘Ārya Dynasty’ or ‘Ārya Empire’. That the steppe peoples in the Ordos,80 but in 222 Ch’in had the latter name is given to the vast northern realm tar- already defeated and conquered Chao’s far northern rem- geted by Ch’in shih huang ti confirms the correctness of nants, which were in the steppe zone. Accordingly, con- the Shih chi statement. That is, the ancestors of the peo- quest of part of Ta Hsia—perhaps the most important ple of Ta Hsia 大夏 ‘Great *Ḥarya ~ Ārya’ necessarily part for Ch’in shih huang ti—was equated by the ruler’s descended from rulers called Hsia 夏 *Ḥarya ~ Ārya who officials as tantamount to having conquered it all; the lived in the steppe zone region known to the ­Chinese as same no doubt applies to the other directions as well. Ta Hsia 大夏, not from rulers living in Bactria, which Ta Hsia 大夏 ‘Great *Ḥarya’ is thus used in this inscrip- tion to refer to the realm of the salient ethnolinguistic 81 region to the north of the Chinese, the home of the Cen- For the archaeology see Di Cosmo (1999). The Hsiung-nu ­confederation included Sakas, but the ethnolinguistic affinities of the tral Eurasian nomads, whose physical culture is known Hsiung-nu ruling clan are uncertain. Cf. Note 90. from archaeology to have been virtually identical (at the 82 Tso chuan, Chao-kung 1 (Legge 1893, II: 585): “Anciently, the time of Ch’in shih huang ti) to that of the North Iranic- emperor Kaou-sin had two sons, of whom the elder was called Oh-pih, and the younger Shih-ch‘in. They dwelt in K‘wang-lin, but could not agree, and daily carried their shields and spears against each other. The 76 This was the ruler’s 28th year (Kern 2000: 25 n. 43); Kern does sovereign emperor (Yaou) did not approve of this, and removed Oh-pih not discuss the inscription dates. to Shang-k‘ëw, to preside over the star Ta-ho…The ancestors of Shang 77 Text from Kern (2000: 33), q.v. for his translation. followed him [in Shang-k‘ëw], and hence Ta-ho is the star of Shang. 78 Kern (2000: 33 n. 77) cites Chavannes (Mémoires historiques, [Yaou also] removed Shih-ch‘in to Ta-hëa, to preside over the star Sin 2: 148-149), whom he follows for both conclusions and adds that, (? in Orion). The descendants of T‘ang (Yaou) followed him, and in accordingly, “Ta-hsia is already mentioned in Tso chuan”. Ta-hëa served the dynasties of Hëa and Shang.” Legge’s “Ta-hëa” is 79 This is from the Chinese perspective; from a world-geography Ta-hsia. standpoint it is the “Eastern Steppe”. 83 Wu (1994: 87 n. 109). 80 In 215 bc (Di Cosmo 1999: 892, 964), q.v. for the earlier history 84 Shih chi 110: 匈奴,其先祖夏后氏之苗裔也 Text from of Central Eurasian peoples north of the Chinese. ­Chinese Text Project; translation by Giele (2010: 237).

99263_JAS_2016-2_06_Beckwith.indd 240 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 241

was only identified two centuries later as Ta Hsia (or subjects as “non-aryas”. This suggests that the term arya perhaps a part of Ta Hsia). was a way of marking the ruling people as a kind of Third, the same Shih chi chapter on the Hsiung-nu “chosen people”. That is precisely the way the Chinese reveals that the Hsiung-nu at the time of T’ou-man first use it in Warring States texts. *Tumen, founder of the Hsiung-nu empire and father of the Fourth, during and after the Han Dynasty period famous Mo-tun *Baγatur, were vassals of the Yüeh-chih the generic term used by the Chinese to refer to the 月氏 ‘the *Tokwar ~ *Tukwar kings’,85 who therefore ­Hsiung-nu and other foreigners of the north and west is actually ruled the Eastern Steppe at the time. At the time Hu 胡 MSC hú from MChi ✩γɔ (Pul. 126) from tradi- of Ch’in shih huang ti that political order was still in tional OChi *γrâ.89 This may be straighforwardly recon- place—it was overthrown only in ca. 176 bc by Mo-tun structed as OChi *γará, a regularly depalatalized north- (r. 209-174 bc) and his son, who defeated the Yüeh- eastern Chinese dialect form of *γaryá ‘noble chih86—so the Chinese term Ta Hsia ‘Great *Ḥarya’ (ones)’—the very same word earlier transcribed as Hsia must have referred, as an epithet, to the Eastern Steppe 夏 OChi *γaryá ~ Hua 華 OChi *γāryá,90 which came empire of the Yüeh-chih. The use of *Ḥarya as an epithet to mean ‘the Chinese, China’ and was borrowed into is fairly clear from Western examples, such as its use by Japanese and Koguryo as Kara from this dialect pro- Darius I, who uses Old Persian arya in its early meaning nunciation, *γará. ‘noble’ or ‘royal’—i.e., being “of arya [noble] descent”, After having been defeated and driven West by the as he says of himself in the Naqš-i-Rustam Inscription Hsiung-nu and then by the Wu-sun, the Yüeh-chih 月氏 (DNa, lines 13-14). For the Chinese too, the adopted for- ‘Tokwar ~ Tukwar kings’ invaded Bactria, where the eign term was hardly an ethnonym meaning ‘Yüeh-chih’, Greeks called them Tokharoi ‘Tokhars’ (or ‘Tokhari- or ‘Iranian’ or anything of the kind. Because it meant ans’). The attested Bactrian form of the name is actually ‘noble, noble one(s)’ or the like, and the early users of Τοχοαρ- Tokhwar- / Toχwar- ~ *Tukhwar- / Tuχwar, the term knew that, it could be used by many different corresponding very closely to the ancient Chinese tran- peoples, who spoke distinctive or unrelated languages scription 月 OChi *Tokwar /Tukwar and the early medi- and had different ethnic identities,87 to refer to them- eval transcription T’u-huo-lo 吐火羅 MChi ✩Toχwala selves in a positive way that at the same time contrasted representing *Toχwara ~ *Tʊχwara.91 They thus became their own “noble” nature with the base nature of the the rulers of what became known in Classical Arabic as other peoples under their sway or outside their territory. Ṭukhāristān (Ṭuχāristān). The Shih chi version of the Foreign descriptions say that early Central Eurasian peo- story specifies that Bactria—the land conquered by the ples, e.g., the Scythians and the Koguryo, viewed their Yüeh-chih—was known as Ta Hsia 大夏 ‘Great Ārya; subjects as “slaves”,88 and in late antique to medieval the Ārya Empire’. However, it would have made perfect Indic and Iranic texts the dominant people refer to their sense if the first Chinese envoy to reach them in that location, Chang Ch’ien, had called the Yüeh-chih realm 85 Beckwith (2009: 380-383), where I reconstruct this name as Ta Hsia because it was the epithet of the Eastern Steppe *Tokwar-kē ~ *Tukwar-kē (literally, the ‘Tokwar/Tukwar kings’), but realm of the same people before their migration, and have unaccountably omitted one key datum: OJpn *tukï ~ *tuku- because the Chinese were at the time aiming to forge an ‘moon, month (月)’ (JDB 461, 464; Martin [1987: 554] has “*tukiy < alliance with them—that was the express purpose of *tuku-Ci”, but the combining form *tuku- only occurs in compounds Chang’s mission. It would thus have been politically sen- before *y-) < *tukïy ~ *tukuy < PJpn *tukwɨr (cf. Whitman 1990). The word clearly corresponds to the Old Chinese transcription of the sitive to use the same ‘imperial’ title which the Chinese attested Bactrian name, Tokwar, as OChi dial. *tukwar ~ tokwer (etc.), had formerly used. Either possibility is left open by the ending up as Mandarin Yüeh 月 MSC yuè < attested MChi ŋgwar Rabatak Inscription of Kanishka, two-centuries later, (Takata 1988: 372-373; cf. traditional MChi ✩ŋuat [Pul. 388]) < OChi dial. *tukwar ~ *tokwer < Early OChi *nokwét (Beckwith 2006b); OJpn *tukï is thus a borrowing from Old Chinese. Note that *e in my 89 HSR reconstructions ignore their own overwhelming evidence Old Chinese reconstructions is not the same as HSR *e. that the 魚 ‘fish’ rhyme in the Shih ching (Starostin 1989: 561-564, his 86 Beckwith (2009: 84-85). class XIII) includes a liquid; many of the words not so reconstructed 87 Beckwith (2016). by HSR are attested as Old Chinese loanwords into one or more of the 88 Beckwith (2007a). They are called “slaves” from the viewpoint neighboring languages with a liquid. of the peripheral peoples, not the Central Eurasians themselves. In fact, 90 The name Hsiung-nu is unrelated to the word Hu, and—pace they were not actual “slaves”, and are not even “viewable” as slaves. de la Vaissière (2005)—it is also unrelated to the name Hun The Central Eurasian Culture Complex from its earliest attestation (­Shimunek et al. 2015). Hu means precisely ‘Central Eurasians’, by down to early Modern times was characterized by a strictly hierarchical birth or heritage. socio-political structure founded on the comitatus relationship and 91 See Note 85 on this name. Hsüan Tsang visited Balkh in 628 or belief in Heavenly God. No one was “free”, but no one was a “slave” 630 ce (Beckwith 2012: 64) and transcribes the second syllable of the ✩ (in the modern sense) either, unlike in the peripheral societies of the word with the character huo 火 MSC huǒ < EMC xwa2 (Pul. 135), i.e., Greeks and Chinese, which practiced true slavery and therefore misin- [χwa]. The many attempts to connect the name Ta Hsia to the name terpreted Central Eurasian culture in this respect. Tokhar- ~ *Tukwar ~ *Toχwar (etc) are misguided.

99263_JAS_2016-2_06_Beckwith.indd 241 25/01/17 12:38 242 Christopher I. Beckwith

which refers to the ruler’s own language—the Iranic lan- loanword Rgyagar96 ‘India’, evidently from an intermedi- guage evidently adopted by the *Tokwar/Tukwar after ary Old Chinese donor form *Ḥaryaḥwar, from an early their migration to Bactria92—as Αρια Aria ‘the noble Indic *āryāwar, corresponding to Sanskrit Āryāvarta (one)’, but because neither Classical Greek script nor its ‘home of the Āryas, India’. According to the eastern Bactrian adaptation indicates any laryngeal fricative ­Eurasian data, that onset [ɣ] was still articulated in the onsets, and the Greek vowel α transcribes both short and Eastern Steppe region at the time the word was first bor- long /a/, Αρια equally represents *arya ~ *ārya ~ *γarya rowed into Old Chinese and was still so pronounced there ~ *harya, etc. in the Early Middle Ages, as attested by the T’u-yü-hun forms of the word discussed below. The absolutely earli- est attested Eastern Old Indic (i.e., Sanskrit) texts only 5 *Ḥarya ~ ĀRYA ~ arya: date to the early centuries ce,97 and thus do not directly History of transmission in the East transcribe the pronunciation of the word half a millen- nium or more beforehand,98 but according to current The Chinese evidence on the term Ta Hsia 大夏 ­theory the Old Indic long vowel represents the regular ‘Great *Ḥarya; the *Ḥarya Empire’ reveals that the epi- outcome of a lost laryngeal onset. In view of the com- thet Hsia 夏 *Ḥarya ~ Ārya was used during the Warring pletely consistent eastern Eurasian data from the earliest States period by the non-Chinese northern and western times on, when the Eastern Steppe was ruled by Eastern rivals of the state of Chao as well as the non-Chinese Scythians, the Sakas, it is significant that the word people of Chao itself. The word is used explicitly to refer is attested in the name of one of the Scythian kings, to them in the Tso chuan legend about the two brothers ­Ariapeithes (r. ca. 490-460 bc), father of King Scyles separated by Yao, and in the Lang-yeh Inscription, as (i.e., ‘King Scythes’),99 and that the Scythians first entered discussed above. The term’s highly distinctive, culturally Western Asia from the East.100 When joined with addi- involved meaning related rulers to the hierarchical order tional evidence specifically from the Sakas and other of the Central Eurasian Culture Complex, which was Central Eurasian peoples,101 it may thus be hypothesized characterized by an ideal imperial realm focused on a that the Scythians and Sakas, who referred to themselves single ruler,93 as the Chinese of the time explicitly state as the ‘royal Scythians’ and the ‘Saka kings’ (i.e., the (v. supra). ‘royal Sakas’), both used the word *Ḥarya (based on the The remaining scientific problem94 is the Chinese attested forms from the Eastern Scythian-speaking realm form’s onset, *γ [ɣ] ~ *Ḥ, which appears to be missing north of the Chinese Warring States)102 ~ Arya (taking the from the Kushan inscriptional form of the Bactrian word, Greek transcription traditionally, as being onsetless and Aria; the Achaemenid inscriptional form of the Old short) in its original epithet sense, ‘noble one(s)’, as did ­Persian word, Ariya [arya]; and the Eastern Old Indic the Iranic and non-Iranic peoples who evidently borrowed form, Ārya. Analysis of this problem reveals the history it from them. of the word’s transmission. Old Tibetan Rgya ‘Chinese; China’ corresponds to The attested onset γ [ɣ] agrees perfectly with the inde- the Old Chinese transcription now read Hsia 夏 ‘­Chinese, pendently reconstructed onset of Old Indic Ārya, the China’ as well as to Hsia in Ta Hsia 大夏 ‘Great Ārya’, laryngeal *Ḥ,95 which is directly attested in the Tibetan which is both the name given to Bactria or to a larger realm including Bactria, as well as to the pre-Hsiung-nu name of the imperial Yüeh-chih steppe empire, the 92 This assumes that they were not originally Iranic speakers; the point is controversial. 93 Beckwith (2010a). 96 The word is earliest attested in the Khri Lde Srong Brtsan Tomb 94 The non-apparent semantic issue is addressed below. Inscription (in Phyongs-rgyas, south-central Tibet), line 29 (Old Tibetan 95 The controversy over the Indo-European “laryngeals” is far Documents Online ) to be dated to the death of the honoree, ­Hittite, transcribed <ḥ>. In most theories one laryngeal is considered to 815 ce, and the Sino-Tibetan Treaty Inscription (Lhasa, explicitly dated have lengthened an adjacent vowel a without coloring it and then dis- 823), East face line 14 (Old Tibetan Documents Online ). seem to be inexplicable vowels (including long vowels) in inherited 97 Pollock (2009: 1, 39); Beckwith (2015: 245). words. However, it is now generally believed that Old Indic Ārya and 98 The fragments of Western Old Indic date to the fourteenth Old Iranic Arya are not Indo-European words by origin, and are prob- ­century bc, but the word ārya is not attested in them. ably not even “Indo-Iranian”; thus, not with Mallory and Adams 99 Szemerényi (1980). 100 (1997: 213), among others, who reconstruct it back to a PIE *h4erós ~ Diakonoff 1985: 93-97. 101 *h4eri̯os ‘member of one’s own (ethnic) group, peer, freeman, (Indo- Beckwith (2010a; 2016). Aryan) Aryan’. However, their “as if PIE” reconstruction to explain 102 Classical Greek script and Old Persian script are both mute on the vowel length agrees perfectly with the data from Eastern Eurasia this issue because Greek does not record non-sibilant fricative onsets, presented here, thus also strengthening the case for borrowing. while Old Persian does not record the voiced dorsal fricative onset *γ-.

99263_JAS_2016-2_06_Beckwith.indd 242 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 243

­southern part of which realm Ch’in shih huang ti con- foreign words, such as the name Źaŋźuŋ (Zhangzhung), a quered according to the Lang-yeh Inscription. Inherited pre-Old Tibetan borrowing of a foreign name that is also from the Proto-Tibetan loanform of Hsia 夏 OChi *γaryá recorded in the early New Persian geographical work ,r’ngrng, read Rângr(u)ng رانگر نگ in the developed sense ‘Chinese, China’, Old Tibetan Ḥudūd al-‘Ālam as Rgya also attests that the Eastern Iranic word was heard from *ryaŋryuŋ ~ *räŋrüŋ.107 In the case of the name by the Chinese in China as *γaryá. Since Old Tibetan Ḥaźa [ɣaźa], the -źa thus derives from earlier *-rya or Rgya- in Rgyagar ‘India, Indian’ also corresponds pre- *-lya. In view of the other attestations of the word arya cisely to Old Indic ārya- in the early name Āryāvarta, as in eastern Eurasia, it is clear that the ruling clan of the noted, it directly supports Indo-Europeanists’ reconstruc- T’u-yü-hun actually called themselves Ḥarya, i.e., ‘Ārya, tion of Old Indic ārya as *Ḥarya (i.e., *[ɣarya]), accord- the noble(s)’, pronounced [ɣaźa] by the Tibetans, who ing to the laryngeal theory.103 Thus, the identity of ārya took it to be their name. The phone [ź] also tells us that ~ *Ḥarya ~ *γarya as the Eastern Eurasian form of the the Tibetans did not hear a foreign [ɣara], which they word attested in Western Eurasia as arya is confirmed. would have borrowed as [ɣara] and written འ་ར་ <ḥa.ra> As mentioned, there is significant additional support or possibly ག་ར་ . The phonetic value represented for the pronunciation of the word ārya with a voiced by the Tibetan letter འ has long been known to have laryngeal onset in the Eastern Steppe in late Antiquity been [ɣ].108 It is well attested in Imperial Old Tibetan as and into the Early Middle Ages. The normal Old Tibetan the onset of the usual spelling of the genitive suffix after appellation of the T’u-yü-hun (Old Tibetan Thogon vowel stems, i.e., -ḥi འི་ -γi ([ɣi], usually transcribed ཐོ་གོན་), a Serbi-speaking104 steppe people who migrated -‘i)—versus གྱ་ -gyi after consonant stems—and occurs southwest and settled in the Ch’inghai-Kansu region with very high frequency throughout the ­language. between 307 and 313 ce,105 is recorded in Early Old The T’u-yü-hun came from the Eastern Steppe, Tibetan from the mid-seventh century ce as Ḥaźa འ་ཞ་, the very region that had formerly been called Ta Hsia phonetically Old Tibetan [ɣaźa].106 This word reflects a ‘Great *Ḥaryá’ in Chinese. After they migrated south- known, specifically Tibetan sound change that took westward in late Antiquity into northwestern China, the place in pre-Old Tibetan and Early Old Tibetan, namely area of Kansu, and into the Koko Nor region of Amdo ~ that from *ryV or *lyV to Old Tibetan źV. The phono- Ch’ing-hai, the ruling clan, who referred to themselves tactics of early Tibetan did not allow the sequences as *Ḥarya [ɣarya], eventually came into contact with *ryV- and *lyV-, so the Tibetan speakers regularly the early Tibetans, who took this to be their name, but repaired them by converting them to źV, as in inherited because of Tibetan-internal phonological rules pro- Tibeto-Burman words (e.g., OTib bźi ‘four’ from pre- nounced and wrote it Ḥaźa འ་ཞ་ [ɣaźa]. The ruling clan OTib *blyi from Common Tibeto-Burman *bli) and also of an Eastern Steppe people was thus still calling itself *γarya ~ *Ḥarya in the Early Middle Ages, late enough that the early Old Tibetan speakers, when they encoun- 103 Mallory and Adams (1997: 213). ce 104 On the Serbi-Mongolic languages, and the position of T’u-yü- tered them by ca. 600 , learned the word, and recorded 109 hun among them, see Shimunek (forthcoming). it in their earliest dated texts. 105 Molè (1970: xii). 106 First identified by Pelliot (1912); cf. Molè (1970: 73 n.22). The word is recorded in Chinese also. The Sung shu 宋書 reads: 107 See Beckwith (2012b), which also discusses the Chinese tran- 西北諸雜種謂之為阿柴虜 ‘the mixed peoples of the northwest call scription of the same name. them the A-ch’ai slaves’ (Shen 1974, 96: 2370). The transcription 108 Jäschke (1881: xvi) records forms from a modern Khams dia- A-ch’ai MSC àchái ~ āchái (Pul. 23: EMC ✩ʔa; 47: ✩dʐaɨj/✩dʐɛːj) is lect that have this value. thus a pronunciation of the term by foreign intermediaries, evidently 109 If the early Old Tibetan speakers had encountered someone who representing a foreign *arǰa. There is another transcription of the same pronounced the name *arya, as it is usually thought to have been pro- word, applied to an important ruler a few generations later (Shen 1974, nounced in the West—i.e., [ʔa.rya], with a glottal stop onset as in 96: 2371), A-ch’ai 阿豺 MSC àchái ~ āchái (Pul. 23: EMC ✩ʔa; 47: English pronunciation and a maximized second syllable onset—they ✩dʐəɨj/✩dʐɛːj). These are patently two Chinese transcriptions of the would have pronounced it *Aźa and written it as ཨ་ཞ་ <ʔaźa> in the same T’u-yü-hun word, which is clearly not the actual name of the two Old Tibetan Annals. In Imperial Old Tibetan, native Tibetan words rulers so called, but foreign versions of their shared title or epithet. On could not begin with a vowel, or even a glottal stop (Beckwith 2006a), the well-attested ‘entering tone’ reading (attested as final coda -r) of 阿 so the Tibetans could conceivably have heard *[ʔaźa] but pronounced in Middle Chinese see Takata (1988: 304; 2000 n. 20); cf. Beckwith it [ɣaźa]. However, we would then have to explain why Old Tibetan (2016: 42 n. 16). Although this word is certainly related to the one texts do not transcribe any other zero-onset (or glottal stop-onset) for- underlying the Old Tibetan transcription, the differences are striking. If eign words with the onset ḥ འ [ɣ] (rather than with ʔ ཨ [ʔ], as attested). the early Tibetans had encountered such a form they could very easily Clearly, even though the glottal stop onset was foreign to Tibetan the have written the name *ཨ་讗་ [ʔa.rǰa] or *ཨར་ཇ་ [ʔar.ǰa]. But they did Tibetans nevertheless heard the difference and transcribed the respec- not do so. Both Old Tibetan and Middle Chinese have a voiced fricative tive phones correctly, e.g. ʔan ཨན་ (Old Tibetan Annals version I, line word-onset phoneme, so the Tibetans (unlike the Chinese) got their 288) the ‘Chinese envoy (rgya’ï pho nya)’ ʔan da lang, presumably form, Ḥaźa འ་ཞ་, directly from the T’u-yü-hun themselves. For “Impe- with the well-known surname An 安 MSC ān < MChi ✩ʔan (Pul. 24), rial Old Tibetan” see Note 109. and the Chinese official ʔu ling ꍴ་ལིང་ (Sino-Tibetan Treaty Inscription

99263_JAS_2016-2_06_Beckwith.indd 243 25/01/17 12:38 244 Christopher I. Beckwith

Furthermore, at the end of the seventh century the Old rgyalpo ‘king’ from rgyal- ‘royal’ from *ḥarya-l- (with Tibetan Annals three times records the name of a non-­ an unidentified suffix-morpheme *l-). Some of these lan- Tibetan place in Mdosmad, the northeastern part of the guages are unattested before the Early Middle Ages, empire which included the former realm of the T’u-yü- assuring us that half a millennium or more earlier, in Old hun, transcribing the name Rgyamśïgar རྒྱམ་ཤྀ་གར་ (line Chinese times, the word *Ḥarya ~ ārya could only have 113; an obviously Tibetanized form); Ryamśïgar 讱མ་ཤྀ་གར་ been borrowed by the Chinese as an epithet in its early (line 206); and ḥaryamśïgar འ讱མ་ཤྀ་གར་ (line 222). These meaning ‘noble one(s); royal’, in the sense that one so are all different attempts at transcription of the unfamiliar characterized belonged to the legitimate ‘noble, royal’ non-Tibetan pronunciation heard by the scribe or scribes. class, the ‘chosen people’ who could or should rule over The third transcription begins with the exact string everyone else. ḥarya-, which is strikingly transcribed in the “technical” In view of the data discussed above, it must be way Sanskrit came to be written in Tibetan script, as ­concluded that the forms *Ḥarya ~ ārya ~ arya represent འ་讱་་་ [ɣa.rya-], explicitly marking the onset of the second one and the same epithet: Old Persian arya = syllable as ry-. This can hardly be anything other than the Western Scythian Aria = Bactrian Aria = Old Chinese native T’u-yü-hun pronunciation of the epithet of the *Ḥarya (→ Old Tibetan Rgya and Old Japanese Kara ­T’u-yü-hun rulers, the very same word borrowed (as a ‘Chinese’) from Eastern Scythian *Ḥarya = Old Tibetan loanword) by the pre-Old Tibetan speakers and trans- Ḥaźa from *Ḥarya from T’u-yü-hun Ḥarya = Old Indic formed according to Early Old Tibetan phonology into Ārya from *Ḥarya = Old Tibetan Rgya- ‘Ārya-’ in ḥaźa, the usual Old Tibetan name of the T’u-yü-hun. Rgyagar (← Old Chinese *Ḥaryaḥwar = early Indic This confirms the above analysis that the word was pro- *Āryāwar = Sanskrit Āryāvarta ‘home of the Āryas; nounced Ḥarya [ɣa.rya] by the T’u-yü-hun themselves.110 India’) = Old Tibetan rgya-l- ‘royal’ from *Ḥarya.

6 what did Ḥarya ~ ĀRYA mean? 7 ta Hsia and the state of Chao

When the Chinese chose their first appellation for Little attention has been paid to the state of Chao 趙, themselves in the late fifth or early fourth century bc, despite the fact that Ch’in shih huang ti was half Chaoish,­ they chose 夏 ~ 華 *Ḥarya ~ Ārya. Why did they choose having been born and raised in Han-tan, the capital of that particular word? At least in the beginning, it could Chao, as the son of a Chaoish princess. The state’s lone not yet have meant ‘Chinese’, or ‘China’, per se. ‘China’ claim to fame in traditional accounts of early China is did not yet exist, except perhaps in the imagination, and that King Wu-ling of Chao (r. 325-299 bc)112 forced the in the West and India the word ārya has no connection subjects of his “Chinese” state to learn and adopt several to China.111 foreign customs: horse-riding (foreign animal and skills), The use of this word by many different ancient to wearing trousers (foreign clothing practices), and shoot- early medieval peoples, many of whom were not speak- ing with the compound bow on horseback (foreign style ers of Iranic or Indic languages, shows it could hardly of warfare). These allowed Chao to conquer deep into the have been borrowed as an ethnonym with the meaning steppe zone to the north and west, greatly enlarging the ‘Iranian’ or ‘early Vedic Indian immigrants from Central state, and at the same time to defeat rivals to the south.113 Asia’. It must still have been an epithet meaning ‘noble, But the usual presentation of these changes in the sources noble ones; royal’ or the like, thus explaining its usage appears to be a rewriting of history, perhaps in order in Early Old Persian, Young Avestan, Old Chinese, Old to ensure Chao’s continued acceptance as a legitimate Indic, T’u-yü-hun, Tokharian A, Old Turkic, and ­Serbi- ­Chinese state;114 but whatever its motivation, the tale Mongolic. To this list can now be added Old Tibetan obscures one of the most important events in the entire Warring States period. The story would have us believe that the king forced his people to adopt foreign customs, south face line 27, given also in Chinese characters as Yü Ling 於陵 to the extent that they dressed, lived, and fought not as ✩ ✩ MSC yú líng < MChi ʔɨă liŋ (Pul. 380, 195). Chinese but as Central Eurasians who lived outside 110 I inadvertently discovered this material in the Old Tibetan Annals when I had thought the manuscript of this paper was finished. It confirms my analysis, but it also suggests the possibility that the early Tibetans may have borrowed some Late Old Chinese words via the 112 Lewis (1999: 602). T’u-yü-hun rather than directly from the Chinese. Further careful study 113 Di Cosmo (1999: 951-952, 960-961). should clarify the question. 114 The stories giving a *ḥarya pedigree to the earlier Spring and 111 The etymological approach is also falsified by the existence of Autumn state of Chin 晉, which had been located in Shanxi, necessar- two distinctive transcriptions with unrelated semantics in Chinese, the ily date to the Warring States period, when the idea of a Chinese Hsia usual sign of a loanword or transcription of a foreign word. Dynasty was invented, or even later.

99263_JAS_2016-2_06_Beckwith.indd 244 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 245

Chao, and thus outside the Chinese oikoumene. That is shifted to and culture—surely a quite the exact opposite of the many attested examples of cul- normal thing to do for people of aristocratic heritage. ture shift throughout the long history of Chinese-Central They secured the place of their Hsia 夏 *Ḥarya ~ ārya Eurasian historical contact, in which foreign peoples ancestry within Chinese tradition by a little creative ­living among Chinese usually ended up adopting Chinese ­reinterpretation of their ancestral realm, Ta Hsia 夏 culture. Most such examples also record that internal ‘Great *Ḥarya; the Ārya Empire’, as a Hsia 夏 *Ḥarya struggles took place for and against the shift. “dynasty”. And when shortly thereafter their own Reanalyzing the Chao story gives a principled expla- prince—who was, appropriately, half-Chaoish—became nation of the historical and linguistic data and provides a king of Ch’in and later Ch’in shih huang ti ‘the First scenario to explain how the Chinese could have adopted Emperor of Ch’in’, uniting the Hsia 夏 *Ḥarya ‘noble a loanword, *Ḥarya, as the epithet that became their peoples’, the merger was as much a victory of the Central autonym. Eurasian Culture Complex as it was of Chinese civiliza- The people of Chao lived on the frontier of the north- tion. ern steppe and the southern agricultural zone. The state thus included both non-Chinese steppe-zone people and Chinese agricultural-zone people, so the ancestors of the 8 from non-Chinese ‘Hsia Empire’ people of Chao necessarily included both peoples as to Chinese ‘Hsia Dynasty’ well. They must have been bilingual, at least in part, but some of the people of Chao who dressed and lived and In the Warring States period, Chinese-speaking peo- fought as Central Eurasians certainly did so because they ple thus first began to use the word Hsia 夏 *Ḥarya ~ already were Central Eurasians and wanted to keep the Ārya to refer not only to themselves but to their own traditional Central Eurasian way of life. Pressure from assumed ancestors, a hypothesized, romanticized, the partly Central Eurasian ruling people of Chao is imagined people who preceded the historical Shang surely what resulted in the Chinese-speaking people of Dynasty. According to later traditional conceptions, the Chao adopting some of their customs and beliefs. Hsia must have had a monarchist system of government, When Chao was first created, by partition of the making their realm a Hsia Dynasty or Empire, therefore: Spring and Autumn state of Chin 晉 during the early Ta Hsia 夏 ‘Great *Ḥarya; the Ārya Empire’. However, Warring States period, it was accepted as a Chinese state no evidence datable before the Warring States has been (as were the other two states formed out of Chin), but it found in which the name Hsia 夏 appears in reference to must already have had a large Central Eurasian popula- such a people or dynasty—no inscriptions, texts, or any- tion, as well as a large Chinese population, and some thing else.115 Since the name is not recorded either in the among both peoples must necessarily have been bilin- meaning ‘the Chinese’ or in the meaning ‘a Chinese gual; this dichotomy would only have increased as Chao dynasty that preceded the Shang’ for almost the entire expanded further to the north. first millennium of written Chinese texts, Hsia *Ḥarya in Reinterpreting the traditional story, many of those both senses can only be a creation of the Warring States among the Chao ruling class who were of Central period in which it appears de novo. Nevertheless, it is ­Eurasian heritage wanted to shift to Chinese language crucially important for understanding the development of and culture, but the conservatives among them— Chinese ideas about themselves in this period. Those certainly including King Wu-ling himself—wanted to who adopted the word *Ḥarya ~ ārya for themselves and retain their steppe warrior traditions, including their their ancestors would have known that it meant ‘noble(s)’ self-image as *Ḥaryas ‘noble ones’, which word they if they were already using it for themselves in their other must have used for themselves, as they must have language, a steppe zone language, which they certainly referred to their homeland as ‘Great *Ḥarya; the Ārya called *Ḥarya ~ ārya ‘the noble one’, the same way so Empire’. Although the traditional story about the revolu- many Central Eurasian peoples used the term. Since they tion of King Wu-ling reverses the directionality of the referred to themselves, too, as *Ḥarya ~ Ārya, their impetus for the historical merger of Chinese language ancestors were of course *Ḥarya ~ Ārya also. After their and values with Central Eurasian state traditions and language and culture shift to Chinese, they kept calling practical customs, it nevertheless attests to the general themselves *Ḥarya ~ Ārya, and naturally kept referring directionality of the change—from Central Eurasians to to their ancestors as *Ḥarya ~ Ārya. It thus appears Chinese—and to the fact that the merger did take place. that when Chao achieved great-power status among the Moreover, the linguistic data can only be interpreted Warring States, the usage of this term spread from the according to the revisionist version, in which people of Chao Chinese speakers to the other speakers of Chinese, Central Eurasian ancestry effectively kept on claiming that they were Central Eurasian nobles even after they 115 See note 4.

99263_JAS_2016-2_06_Beckwith.indd 245 25/01/17 12:38 246 Christopher I. Beckwith

so that the Chinese as a whole came to call themselves References 夏 *Ḥarya ~ Ārya. Since they also referred to the people of Ta Hsia 夏 ‘Great *Ḥarya; the Ārya Empire’ (who are Adams, Douglas Q. 1999. A Dictionary of Tokharian B. mentioned for the first time in the Tso chuan) as their Amsterdam: Rodopi. ancestors, logically that made the people of Ta Hsia the Alekseyev, A.Yu. 2005. Scythian Kings and ‘Royal’ ­Burial-Mounds of the Fifth and Fourth Centuries bc. In: first Hsia 夏 *Ḥarya ‘Chinese’, whose ‘dynasty’ or David Braund, ed. Scythians and Greeks: Cultural Inter­ ‘empire’—here is the creative innovation—had preceded actions in Scythia, Athens and the Early Roman Empire. the Shang (the first fully historical Chinese dynasty). Exeter: University of Exeter Press, 39-55. Calling these imagined Chinese people Hsia 夏 *Ḥarya Baxter, William H. 1992. A Handbook of Old Chinese ~ Ārya made the Warring States rulers of the Chinese ­Phonology. Berlin: Mouton de Gruyter. (= Bax.) “Āryas and the descendants of Āryas”, exactly as Darius Baxter, William H. and Laurent Sagart 2014. Old Chinese: I says of himself in the Naqš-i Rustam Inscription. A New Reconstruction. Oxford: Oxford University Press. The alternative to this scenario (i.e., a late Spring and Beckwith, Christopher I. 2005a. On the Chinese Names for Autumn ~ early Warring States influence via Central Tibet, Tabghatch, and the Turks. Archivum Eurasiae Medii Eurasia) is belief that the legendary Chinese dynasty Aevi 14: 5-20. known as Hsia actually existed. As noted by scholars, that —, 2006a. The Sonority Sequencing Principle and Old Tibetan Syllable Margins. In: C.I. Beckwith, ed., Medieval Tibe- is not supported by a shred of authentic pre-Warring to-Burman Languages II. Leiden: Brill, 45-55. States evidence. To the contrary, both historical and lin- —, 2006b. Old Tibetan and the Dialects and Periodization of guistic data tell us that there certainly was a Ta Hsia Old Chinese. In: C.I. Beckwith, ed., Medieval Tibeto-­ 大夏 ‘Great *Ḥarya; Ārya Empire’ in the Warring States Burman Languages II. Leiden: Brill, 179-200. period, and some of its people did participate prominently —, 2006/2007. The Frankish Name of the King of the Turks. in the struggles among the Warring States, in the process Archivum Eurasiae Medii Aevi, 15: 5-11. spreading use of the epithet Hsia 夏 *Ḥarya ~ ārya. —, 2007a. Koguryo, the Language of Japan’s Continental In sum, Chinese speakers ultimately adopted the epi- ­Relatives: An Introduction to the Historical-Comparative thet ārya ‘noble (ones)’ from Eastern Scythian speakers Study of the Japanese-Koguryoic Languages, with a Prelim- who pronounced it *Ḥarya, and they used it for them- inary Description of Archaic Northeastern Middle Chinese. Second Edition. Leiden: Brill. selves in the same way that the Scythians, Persians, and —, 2007b. Phoronyms: Classifiers, Class Nouns, and the other Central Eurasian peoples did at about the same ­Pseudopartitive Construction. New York: Peter Lang. time, in the mid-first millennium bc, and that other Eur- —, 2007c. On the Proto-Indo-European Obstruent System. asian peoples did from Antiquity down into the Middle Historische­ Sprachforschung 120: 1-19. Ages. The basic idea, a fundamental element of the —, 2008b. Old Chinese Loans in Tibetan and the Non-unique- Central Eurasian Culture Complex, is clearly the Scyth- ness of ‘Sino-Tibetan’. In: C.I. Beckwith, ed., Medieval ian-Saka rulers’ view of themselves as a ‘chosen peo- ­Tibeto-Burman Languages III. Halle: IITBS GmbH, 161-201. ple’—the ‘royal Scythians’ and the ‘Saka kings’—in —, 2009. Empires of the Silk Road: A History of Central contrast to other peoples with whom they were in con- ­Eurasia from the Bronze Age to the Present. Princeton: flict or over whom they ruled.116 When this term and Princeton University Press. accompanying viewpoint had been adopted by the —, 2010a. A Note on the Heavenly Kings of Ancient Central Eurasia. Archivum Eurasiae Medii Aevi 17 (2010) 7-10. ­Chinese speakers of the disunited Warring States as a —, 2010b. Old Chinese Loanwords in Korean. In: Sang-Oak way of referring to themselves as a distinctive people Lee, ed., Contemporary Korean Linguistics: International and culture—and thus ideally united—it developed Perspectives. Seoul: Thaehaksa, 1-22. semantically to become the first autonym for “the Chi- —, 2012a. Warriors of the Cloisters: The Central Asian Ori- nese”, and the name for their earliest ancestors. The gins of Science in the Medieval World. Princeton: Princeton word in later times, increasingly written in its alternate University Press. form Hua 華, eventually became the traditional ethno- —, 2012b. On Zhangzhung and Bon. In: Henk Blezer, ed., nym of the Chinese people used down to modern times. Emerging Bon. Halle: IITBS GmbH, 164-184. Hua 華 remains an element of the official names of both —, 2014. The Aramaic Source of the East Asian Word for the first Republic of China, Chung-hua min-kuo 中華民國 ‘Buddhist Monastery’: On the Spread of Central Asian (which continues in Taiwan) and the second, the People’s Monasticism in the Kushan Period. Journal Asiatique­ 302.1: 111-138. Republic of China, Chung-hua jen-min kung-ho-kuo —, 2015. Greek Buddha: Pyrrho’s Encounter with Early Bud- 中華人民共和國 . dhism in Central Asia. Princeton: Princeton University Press. —, 2016. The Pronunciation, Origin, and Meaning of A-shih-na 阿史那 in Early Old Turkic. In: István Zimonyi and Osman 116 Beckwith (2010a). Karatay, eds., Central Eurasia in the Middle Ages: Studies

99263_JAS_2016-2_06_Beckwith.indd 246 25/01/17 12:38 The phonology, meaning, and origin of the epithet ḥarya ~ ārya in East Asia 247

in Honour of Peter B. Golden. Wiesbaden: Harrassowitz, Civilization to 221 b.c. Cambridge: Cambridge University 39-46. Press, 587-650. Beckwith, Christopher I. and Michael L. Walter 2016. Dating Ligeti, L. 1971. À propos du “Rapport sur les rois demeurant and Characterization of the Old Tibetan Annals and the dans le Nord”. In: Études tibétaines dédiées à la mémoire Chronicle. In: Charles Ramble and Hanna Havnevik, eds. de Marcelle Lalou. Paris: Librairie d’Amérique et d’Orient, Festschrift for Per Kvaerne. Oslo. 166–189. Behr, Wolfgang 2007. Xià: Etymologisches zur Herkunft des Loewe, Michael 1999. The Heritage Left to the Empires. ältesten chinesischen Staatsnamens. Asiatische Studien/ In: Michael Loewe and Edward L. Shaughnessy, eds., The Etudes asiatiques 60.3: 727–754. Cambridge History of Ancient China: From the Origins of Brooks, E. Bruce and A. Taeko Brooks 1998. The Original Civilization to 221 b.c. Cambridge: Cambridge University Analects: Sayings of Confucius and His Successors. New Press, 967-1032. York: Columbia University Press. Lung, Yü-ch’un (龍宇純), ed. 1968. 唐寫全本王仁昫刊謬補 —, 2007. An Overview of Selected Classical Chinese Texts. 缺切韻校箋 [T’ang hsieh ch’üan pen Wang Jen-hsü k’an Amherst: Warring States Workshop . Checked on Decem- wen ta-hsüeh. ber 31, 2013. Mair, Victor 2013. Was there a Xia Dynasty? Sino-Platonic Carling, Gerd 2009. Dictionary and Thesaurus of Tocharian Papers 238. Philadelphia: University of Pennsylvania. A, Volume 1: A-J. Wiesbaden: Harrassowitz. Martin, Samuel 1987. The through time. Diakonoff, I. M. 1985. Media. In: Ilya Gershevitch, ed. The New Haven: Yale University Press. Cambridge History of Iran, Vol. 2. Cambridge: Cambridge Mallory, J.P. and D.Q. Adams, eds. 1997. Encyclopedia of University Press, 36–148. Indo-European Culture. London: Fitzroy Dearborn. Di Cosmo, Nicola 1999. The Northern Frontier in Pre-Imperial Molè, Gabriella 1970. The T’u-yü-hun from the Northern Wei China. In: Michael Loewe and Edward L. Shaughnessy, to the Time of the Five Dynasties. Serie Orientale Roma eds., The Cambridge History of Ancient China: From the XLI. Roma: Istituto Italiano per il Medio Ed Estremo Origins of Civilization to 221 b.c. Cambridge: Cambridge ­Oriente. University Press, 885-966. Monier-Williams, Monier 1899. A Sanskrit-English Diction- Freu, Jacques 2003. Histoire du Mitanni. Paris: L’Harmattan. ary, Etymologically and Philologically Arranged, with Giele, Enno, trans. 2010. The Hsiung-nu, Memoir 50. In: ­Special Reference to Cognate Indo-European Languages. ­William H. Nienhauser, Jr., ed. Ssu-ma Ch’ien, The Grand New Edition. Oxford: Clarendon. Reprinted: Delhi: Moti- Scribe’s Records, Vol. IX. Bloomington: Indiana University lal Banarsidass, 1988. Online edition, 2008 . Jäschke, Heinrich August 1881. A Tibetan-English Dictionary, Omodaka, Hisataka (澤瀉久孝), et al. 1967. 時代別国語大辞 with Special Reference to the Prevailing Dialects. London: 典, 上代編 [Jidaibetsu kokugo daijiten, jōdaihen]. Tokyo: The Secretary of State for India in Council. Sanseido. (= JDB). Keightley, David N. 1999. The Shang: China’s First Historical Pai Yü-lan (白於藍) 2012. 戰國秦漢簡帛古書通假字彙纂 . Dynasty. In: Michael Loewe and Edward L. Shaughnessy, Zhanguo Qinhan jianbo gushu tongjiazi huizuan. Fu-chou: eds., The Cambridge History of Ancient China: From the Fu-chien jen-min ch’u-pan-she. Origins of Civilization to 221 b.c. Cambridge: Cambridge Pelliot, Paul 1912. Les noms tibétains des T’ou-yu-houen et University Press, 232-291. des Ouigours. Journal Asiatique sér. 10, tome 20: 520-523. Kern, Martin 2000. The Stele Inscriptions of Ch’in shih-huang: Pollock, Sheldon 2006. The Language of the Gods in the Text and Ritual in Early Chinese Imperial Representation. World of Men: Sanskrit, Culture, and Power in Premodern New Haven: American Oriental Society. India. Berkeley: University of California Press. Kiyose, Gisaburo N. (清瀬義三郎則府) 2001. 邪馬壹國の言 Pulleyblank, Edwin G. 1991. Lexicon of Reconstructed Pro- 語を論じ原日本語の故地におよぶ [Yamatö koku no nunciation in Early Middle Chinese, Late Middle Chinese, gengo o ronji gen-Nippongo no kochi ni oyobu]. 語源研究 and Early Mandarin. Vancouver: UBC Press. (= Pul.) [Gogen Kenkyū] 5: 202-244. Schuessler, Axel 1987. A Dictionary of Early Zhou Chinese. Kiyose, Gisaburo N. and Christopher I. Beckwith 2008. The Honolulu: University of Hawaii Press. (= Sch.) Origin of the Old Japanese Twelve Animal Cycle. Arutaigo Shen, Yüeh (沈約) 1974. 宋書 [Sung shu]. Peking: Chung-hua kenkyū – Altaistic Studies 2: 1-18. shu-chü. —, forthcoming. Apocope of Late Old Chinese *a. Shimunek, Andrew Eric forthcoming. Languages of Ancient La Vaissière, Étienne de 2005a. Huns et Xiongnu. Central Southern Mongolia and North China: A Historical-Compar- Asiatic Journal 49.1: 3-26. ative Study of the Serbi or Xianbei Branch of the Ser- Legge, James 1893/1985. The Chinese Classics. Vol. V, The bi-Mongolic Language Family, with an Analysis of North- Ch’un ts’ew with the Tso chuen. Taipei: Southern Materials eastern Frontier Chinese and Old Tibetan Phonology. Center (reprint of the Shanghai 1935 revised edition). Tunguso-Sibirica. Wiesbaden: Harrassowitz. Lewis, Mark Edward 1999. Warring States Political History. Shimunek, Andrew Eric and C.I. Beckwith, Jonathan N. Wash- In: Michael Loewe and Edward L. Shaughnessy, eds., The ington, Nicholas Kontovas, and Kurban Niyaz 2015. The Cambridge History of Ancient China: From the Origins of Earliest Attested Turkic Language: The Chieh 羯 (*Kɨr)

99263_JAS_2016-2_06_Beckwith.indd 247 25/01/17 12:38 248 Christopher I. Beckwith

Language of the Fourth Century A.D. Journal Asiatique —, 2000. Multilingualism in Tun-huang. Acta Asiatica, Bulletin 303.1: 143-151. of the Institute of Eastern Culture 78: 49-70. Sims-Williams, Nicholas 1997. New Findings in Ancient Walter, Michael L. and Christopher I. Beckwith 2010. The Afghanistan—the Bactrian documents discovered from the Dating and Interpretation of the Old Tibetan Inscriptions. Northern Hindu-Kush. . Whitman, John 1990. A rule of medial *-r- loss in pre-Old Starostin, Sergei A. 1989. Реконструкция древнекитайской Japanese. In: Philip Baldi, ed., Linguistic Change and фонологической системы [Rekonstrukcija drevnekita- Reconstruction Methodology. Berlin: Mouton de Gruyter, jskoj fonologičeskoj sistemy]. Moscow: Nauka. (= Sta.) 511-545. Szemerényi, Oswald J.L. 1980. Four Old Iranian Ethnic Wu, Fu-chu (吳福助) 1994. 秦始皇刻石考 [Ch’in shih huang Names: Scythian—Skudra—Sogdian—Saka. Österreichis- k’e shih k’ao]. Taipei: Wen-shih-che ch’u-pan-she. chen Akademie der Wissenschaften, Philosophisch-Histor- Zürcher, Edmond 2007. The Buddhist Conquest of China: The ische Klasse, Sitzungsberichte, 371 Band. Vienna: Verlag Spread and Adaptation of Buddhism in Early Medieval der Österreichischen Akademie der Wissenschaften. China. Third Edition. Leiden: Brill. Takata, Tokio (高田時雄) 1988. 敦煌資料による中国語史 の研究 [Tonkō shiryō ni yoru Chūgogugoshi no kenkyū] (A Historical Study of the Chinese Language Based on ­Dunhuang Materials). Tokyo: Sobunsha. (= Tak.)

99263_JAS_2016-2_06_Beckwith.indd 248 25/01/17 12:38