Empires of the Silk Road: A History of Central Eurasia From the Bronze Age to the Present (87 page)

46.
The word for ‘wheel, chariot’ in Chinese is written
, which was originally a pictograph. It has two readings in Middle Chinese, *t∫a and *kü, the second of which can be reconstructed for Old Chinese as either *klâ or *krâ, from theoretical Early Old Chinese *kelé ~ *kolé ~ *karé ~ *kore ~ etc. Because assumed Early Old Chinese *o and *we ~ *wa merged later within Chinese, the form *kolé is not distinguishable from *kwelé, which is itself clearly a form of the Indo-European word for ‘wheel’. The wheel was introduced to China as a part of the chariot—in Early Old Chinese the one word has both meanings—so *kolé ~ *kwelé appears to be the correct form. However, the Old Tibetan word for ‘wheel, circle’,
hkorlo,
derives regularly from *kwerlwe ~ *kewrlew ~ *kwerlo ~ *korlew (etc.) in pre-Old Tibetan, which language has the same problem of the indistinguishability of earlier *o and *we ~ *wa. Of these forms, the Proto-Tibetan form *kwerlo corresponds perfectly to Proto-Indo-European *kweklo ‘wheel’ with the exception that Tibetan has
r
instead of the second PIE *k. This interesting anomaly may be due to an Old Chinese intermediary—syllable-final *γ was evidently phonetically close to [
] (the standard French and German pronunciation of /r/) at some point in Old Chinese, because it was perceived as /r/by Common Japanese-Koguryoic speakers, who borrowed OChi
‘fowl, bird’ as *tewr (Beckwith 2007a: 138, where *tawr should be corrected to *tewr), among other examples. The Old Tibetan form thus presupposes an Old Chinese donor form *kwe
lo, from *kweγlo, from Early Old Chinese *kweklo ‘wheel, chariot’. The fact that both Chinese readings of
have level tone, not rising tone (the normal reflex of Old Chinese syllable-final *γ) shows that within Central Old Chinese both syllables continued to be analyzed as open syllables, so *γ was still perceived as the onset of the second syllable (*γlo), not as the coda of the first syllable (i.e., not as *kweγ), but in the donor dialect, *γ was perceived as the coda of the first syllable (*kweγ) and thus shifted to *kwe
and was heard as *kwer by the Proto-Tibetans. It is also possible that the Tibetans borrowed the word directly from early Indo-European speakers and remodeled the first syllable on the basis of the Tibetan verbal root √kor ‘to turn; revolve’ (and its causative √skor ‘to turn, encircle’), which could correspond to an o-ablaut form of PIE *(s)ker ‘to turn, bend’ (Wat. 78). In any event, the Proto-Chinese word for ‘wheel, chariot’ is certainly an Indo-European loanword and appears to be reconstructible as *kweγlwe ~ *kweγlo, from PIE *kweklo. Nevertheless, much more work is needed on this topic.

47.
According to the current received view among Sinological linguists (as distinguished from archaeologists, who are no longer so parochial), Chinese culture developed essentially as an island surrounded by wasteland inhabited by wild animals and
barbarians.
The only foreign influences admitted are from the south: much discussion focuses on the putative contributions of the Hmong and Proto-Miao-Yao to the early Chinese. But Chinese culture was certainly by far the most advanced in the East Asian region at that time, regardless of its language or languages. The consequent unbelievability of such speculations—particularly the proposed direction of some of the loaning (from the Miao-Yao, etc., to the Chinese)—seems not to have been noticed. With the exception of E. G. Pulleyblank, contemporary Sinological linguists accept the existence of only a single Indo-European word (a loanword) in Old Chinese: the Tokharian word for ‘honey’. Their refusal to look at the now undisputed archaeological evidence, or to attempt to relate it to the linguistic evidence, is baffling.

48.
The
Mischsprache,
or ‘mixed language’, is supposedly a language that is so mixed its genetic ancestry is unclear, unlike a creole, the ancestry of which is clear. The
Mischsprache
theory has once again been disproved (Beckwith 2007a: 195–213; Mous 1996). In a book on the world’s only putative example, Ma’a (or Mbugu), argued by Thomason and Kaufman (1988) to be a mixed language, Mous (2003) now waffles on the issue, claiming that Mbugu speakers do not simply use the Ma’a “language” as a code—exactly like English Romani, who speak English and also, for the sake of secrecy and ethnic solidarity, a register of English with a heavy admixture of Romani, as shown very clearly by Thomason and Kaufman (1988)—they actually speak two languages, Mbugu and Ma’a, which share one grammar. This is a step backward. The clear statement of Mous’s (1996) article remains the simple, unvarnished truth about Mbugu and about Ma’a, which must be seen not as a language but as a register of Mbugu.

49.
It has been argued that because the Sintashta-Petrovka chariot seems to be a ritual model and would probably be unstable if actually driven, it is therefore a Central Eurasian “imitation” of Near Eastern chariots, thus proving that the horse-drawn war chariot was invented in the ancient Near East, not Central Eurasia. But this argument depends on two highly questionable points: it is argued, first, that the early chariot developed and spread in this way because of its “prestige value” rather than because of its practical military use and, second, that unrecognizably crude Ancient Near East portrayals of indeterminate beasts hitched to two-wheeled vehicles represent chariots (Littauer and Crouwel 2002: 45–52). For the Sintashta-Petrovka chariot one would assume the opposite to be more likely: ritualized objects used in burials would seem to be based on practical things long used in a culture, whether or not of local origin, whereas uncomprehended foreign objects, when placed in burials, would presumably tend to be the actual practical things (such as the chariots buried at Anyang), not ritualized versions of them. This problem should, however, be addressed by archaeologists.

50.
Herodotus says that “the nomad Scythians inhabiting Asia, being hard-pressed in war by the Massagetae, fled away across the river Araxes to the Cimmerian country (for the country which the Scythians now inhabit is said to have belonged of old to the Cimmerians)…. And to this day there are in Scythia Cimmerian walls, and a Cimmerian ferry, and there is a country Cimmeria and a strait named Cimmerian. Moreover, it is clearly seen that the Cimmerians in their flight from the Scythians into Asia did also make a colony on the peninsula where now the Greek city of Sinope has been founded; and it is manifest that the Scythians pursued after them and invaded Media, missing their way; for the Cimmerians ever fled by way of the coast, and the Scythians pursued with the Caucasus on their right till they came into the Median land, turning inland on their way” (Godley 1972: 210–213; cf. Rawlinson 1992: 299–300). Godley (1972: 213, n. 1) comments on Cimmeria: “The name survives in ‘Crimea’ “; his “strait named Cimmerian” is (literally) the ‘Cimmerian Bosphoros’ “. Although some of this account appears to be due to late Persian stories, the archaeological record largely supports Herodotus.

51.
The view presented here was written before I knew about and read the article of de la Vaissière (2005d), who emphatically rejects the consensus among Central Eurasianists that the Huns were unconnected to the Hsiung-nu. The case has been made for the Hsiung-nu having been Iranians (Bailey 1985: 25 et seq.), Kets (Pulleyblank 2000; Vovin 2000), or others. Certainly they were at least strongly influenced culturally by the Sakas, the eastern branch of the Northern Iranians, and it is quite possible that the name Hsiung-nu is actually a transcription of a form of the old North Iranian ethnonym *skuδa ‘archer’, but further work on Old Chinese reconstruction is required in order to either confirm or disprove this hypothesis. Although de la Vaissière makes the strongest, most convincing case so far in favor of the connection, there unfortunately remain many problems that he does not resolve. Most significantly, he does not discuss the phonology of the Chinese transcription. But that is after all the key issue. The evidence in that regard indicates that the Old Chinese form of the name began with an initial cluster, among other major differences from the Middle Chinese form. However, because de la Vaissière introduces some data not previously utilized in the arguments back and forth, his argument calls for examination in depth. See also endnote 52.

52.
It is accepted that at least some syllables with *sC(C)- onset clusters became reduced to the simple onset *?- by Early Middle Chinese times at the latest. It is probable that all such onset clusters became so reduced in the Central Dialect, and quite possibly even the simple onset *s- also became *?- (it certainly did so in some cases). The change was complete in the Central dialects by Middle Chinese times, but *s still existed in some positions in the early centuries
AD
because it shows up in Chinese transcriptions of Indic terms, as has been shown by Pulleyblank (1984) and others. It is thus probable that Old Chinese *s was still preserved in the early Western Han period. In view of the phonology of the name Hsiung-nu, it is unlikely that the onset did not begin with *s; the beginning of the first syllable thus had the shape *sV- or *sCV-; because the capital of the Western Han was in Ch’ang-an, it is probable that the official dialect of Chinese at the time had oralized nasals (e.g., Mo-tun represents *Baγtur, not *Maγtur). Accordingly, the transcription now read as Hsiung-nu may have been pronounced *Soγdâ, *Soγlâ, *Sak(a) dâ, or even *Skla(C)da, etc. See
appendix B
.