Contents Türkic languages
Ogur and Oguz
Alans and Ases
E.Pulleyblank Eastern Hun Language
O.Pritsak Onomasticon of Western Huns W.B.Henning Xiongnu are Huns
L.Gumilev Language of Huns
Kisamov N. The Hunnic Oracle
Tekin T. Hsiung-Nu Language
Vovin A Hsiung-Nu Language
Taskin V. Hsiung-Nu Language
Doerfer G. On Hunnic and Turkic (snippets)
Gmyrya L. Caspian Huns
|The Hun language|
E.G. PULLEYBLANK (1922 - 2013)
The Hun (E.Pulleyblank: Hsiung-nu) Language
THE CONSONANTAL SYSTEM OF OLD CHINESE: PART II
Asia Major, 1962/3, 9:59-144 and 206-265
This posting complements the work of L.Potapov, it is a portion of E.G. Pulleyblank “very old” work, where he advanced an idea that Huns were Ket-speaking. Though that idea did not really take root, some scholars (A.Vovin, for example) kept pursuing it. As an attempt to tie Ket with Hinnic, the work did not produce much, and what was advanced brought up criticisms, but it stimulated further research, and contained very interesting facts, which mostly were left without a follow-up. The list of 278 Hunnic words as a concise and scholarly tabulation still awaits its time, but the few examples, given in the work with an eye toward non-Türkic provenance in reality achieve the opposite, and are definitely most educational. Any additional comments and clarifications are shown in blue italics. E. Pulleyblank used confusing Chinese terminology and confusing terminology used in the Scytho-Iranian theory; the posting gives English terminology with appended original terminology. Some diacritics were retained, but many were dropped to minimize font conflicts. Chinese hieroglyphic spelling was also dropped, with  indicating placeholders. Page numbers are shown at the beginning of the page. A reference PDF file is here. A work by O.Pritsak on onomasticon of European Huns, who also were a subject of various speculations, is here.
In his complement to E.G. Pulleyblank's work, A.Vovin thoroughly debunked all E.G. Pulleyblank's postulations given in this Appendix (Vovin A. 2000, Did the Xiong-nu Speak a Yeniseian Language?). Then A.Vovin turned around and expressed support for his conclusions.
The phrase is concise and even in modern Türkic is readily recognizable. Semantically, it may sound obsolete, some expressions are not in common use any more, but the grammar is intact, and since we can easily catch the phrase “The zex would have xyzed the door wide” with only literary familiarity with Mr. Zex and the xyz action, so the whole phrase sounds to the Türkic speaker undoubtedly natural “If you do that, you' d get it”:
The Hunnic phrase with Türkic and English rendition:
In modern Turkish, the second line of the phrase is practically the same:
Any linguist would observe the amazing continuity of the vocabulary and grammatical
The Modern Turkish replaced the verb tiligan (tiligar) with a different root, çık, the only substantial modification in the 2,000-year old phrase. Other then that, each grammatical form and each word is known from the new and old Türkic doctionaties, and from the common speach..
Most interesting is the homophonic message of the poem, completely missed by the non-Türkic-speaking investigators, the ancient Chinese as well as the modern scholars. Pugu is not only a title/rank of Liu Yao, pug/buk is also poop. In Türkic tilekk (Turkish dilek, ref. Old Türkic Dictionary, 1969, Leningrad, Science, p. 560) is "to wish", with affix -gan it becomes tiligan = having wished (past participle, 3rd person singular, perfect tense, ref. Mahmud Kashgari, 1960. Turky suzlar devoni (Devon lugotit turk). vol. 1, p.412. Tashkent.), Süčy tiligan = Army commander has wished (Russian "ďîćĺëŕâřčé").
Pugu qüitudan has 2 homophonic forms:
Thus, the poem relays three messages:
And finally, melodically the verse follows a five-syllable metrical pattern, or pentameter, typical for the Türkic ancient poetry (Khatipov Gosman, "Shigyr tozeleshe", Tatar Publishing, Kazan 1975,. pp. 108-135)
Analysis ©2010 A. Mukhamadiev
For the philologists E.Pulleyblank and A.Vovin not to recognize an obviously Türkic phrase, readily recognizable by a putative trader in a Tashkent, Baku, or Istanbul bazaar, at least in a philologically most rudimentary form, is a disgrace, it can only be explained by their personal notions unbecoming to their profession. In science, as in other contentious fields, is a clear demarcation between scientific exploration and scientific fraud.
An improved understanding of the phonology of Chinese in the Han period offers an opportunity to re-examine the question of the affinities of the Huns (E.Pulleyblank: Hsiung-nu), the most important neighbors and rivals of the Chinese at that period. There are many more Chinese transcriptions from Hunnish (E.Pulleyblank: Hsiung-nu) than from any other foreign language before the coming of Buddhism. Most of the words are proper names or titles whose exact significance is unknown but there are also a number for which the meaning is indicated. There have naturally been attempts to identify these with known words in other languages but the degree of generally acknowledged success has been meager.
The most prevalent view nowadays, at least in the west, is probably that the Huns
were ancestors of the Turks. Apart from the fact that the Chinese historians expressly say
that they were, which unfortunately has little evidential value in itself since genuine
Huns (E.Pulleyblank: Hsiung-nu) known by that name had
long since disappeared when the Türks (E.Pulleyblank: T'u-chüeh) come on the
scene in the middle of the sixth century A.D., the main support for this theory has been
the evident connection between the Hunnic (E.Pulleyblank:
Hsiung-nu) word ch'eng-li "heaven" and Turkish tängri
(Huns did not disappear at all, but are known to endure as
cohesive tribes for another millennia, in different locations and in different groups, and
then continue as identifiable seoks for another half of a millennia almost well into the
Modern Times. Hun tribes lived parallel to the Türkic and Tele superethnoses. A second
point, about "Tängri", the pronunciation is very variable, from "Tura" is Chuvash, "Tanri"
in Turkish to "Tengri" in Bulgarian and "Dinger" in Sumerian. E.Pulleyblank did not need
to limit the comparison to a single anecdotal example. It is the breadth, not the depth
that leads to conclusions.)
Pelliot (1944) has however shown that variation and instability of the word in Turkish where it yielded a pronominal form tärim, and also in Mongol, make it quite likely that it was ultimately a loanword in those languages. The attempts to furnish Turkish or for that matter Mongol or Manchu, etymologies for other Hunnic (E.Pulleyblank: Hsiung-nu) words have not been very satisfying (in the eyes of the beholder). Before discussing individual words it may be worthwhile to consider what can be discovered about Hun (E.Pulleyblank: Hsiung-nu) phonology by examining the whole body of Hun (E.Pulleyblank: Hsiung-nu) transcriptions. I have gathered some 190 probable Hun (E.Pulleyblank: Hsiung-nu) words for the Former Han period from the Shih-chi and Han-shu, 57 more from the Hou Han-shu and 31 from the Chin-shu. Most of them are proper names or titles. A few are words for which a meaning is given and these have all been studied by previous investigators. (A respectable dictionary of 278 well-garbled words)
The first point that strikes one is the large number of (inferred reconstructed) words beginning with l. There are seventeen in the Former Han material, i.e. nearly nine per cent of the total (including some which are probably *hl or *vl). In the Later Han material there are three and in the Chin material two. We must note further that there are also words in Old Chinese *δ, which is likely to have been used for foreign l in the Former Han period (if there were any). In all Altaic languages words with initial r- are totally absent and words with l- are rare, consisting mainly of onomatopes and obvious loanwords. This evidence does not of course absolutely (and it does not prove anything for that matter) prove that the Hun (E.Pulleyblank: Hsiung-nu) language had words beginning with l- and r-, since the Chinese, in transcribing, might have left out an initial vowel sound (and made inferred reconstructions misleading and useless. The presence of vl- , where "v" is a vowel in Türkic is unquestionable). Occasional examples of this from later times could be found. But we have only to contrast the position with what we find in the transcriptions of the T'o-pa Wei period, from which initial l- is conspicuously absent (Bazin 1950), to see that the frequency of l- must raise serious (is adjecive "serious" too strong for all uncertainties?) doubts about the possibility of connecting Hunnic (E.Pulleyblank: Hsiung-nu) with any Altaic language.
The second point that is equally opposed to the phonology of Altaic languages is the
evidence for initial consonantal clusters. There are at least fourteen cases and perhaps
more among the Former Han transcriptions which the Chinese word used for the beginning of
the transcription probably had an initial cluster at that period. Here again the evidence
is not conclusive in individual cases, since an initial vowel might have been omitted or
Chinese -l- as the second element of a cluster might have represented foreign -r- or
at the end of the syllable by metathesis, as in "klek-kuən=*ken-kuən=*Qïrqun or
*Qïrqur for the later Kirghiz (p. 133 above). There is at least one clear case however in which an
initial cluster seems proved and that is in the name Hun (E.Pulleyblank:
Hsiung-nu) itself. It was shown in the
first part of this article that Hun (E.Pulleyblank: Hsiung-nu)
< M. hioŋ-nou probably went back to *flöŋ-nah
and was equivalent to the Greek Φροΰνοι
(Modern Türkic languages preserved both articulations for the word Hunnu,
"Hünnu" and "Sünnu", both widely used in the scientific literature,
and both could be spelled in Chinese with an initial "cluster". The
Chinese hieroglyphs were selected not for presumed perfect phonetical representation,
but for the derogatory semantical value “malicious slave”).
It was also shown that the royal clan name of the
Hun (E.Pulleyblank: Hsiung-nu) Luan-ti [aka Luandi,
挛鞮 , 挛 = tangled/entwined/crooked, 鞮 = leather shoes] could be
reconstructed as *vlan-teh or *vlon-teh, being no doubt
connected with the name of the Hun (E.Pulleyblank: Hsiung-nu)
(Found in the literature
Suylyanti/Sulianti 虚连题 [虚 = 虍 (“hu”)+ 丘 = hollow/vain/false, 连 =
join/connect/continuous/even, 题 = forehead/title/headline theme], Sui-ian, Xiu-ian, Kuyan for Hun's tribe that eventually became their dynastic
clan have a reliable translation and hieroglyph semantics as "hare" = Türkic "Kuyan". Is
it scholarly kosher not to mention other opinions?).
It may be further noted that according to the Shik-chi 110.0245.1, the Hun (E.Pulleyblank: Hsiung-nu) ruler held court every year in the fifth month (i.e. at mid-summer) at Lung [ ] City, where he sacrificed to his ancestors, heaven and earth, and the spirits (or the spirits of heaven and earth) (Han-shu 94A.0596.1 has [ ]). The Hou Han-shu 119.0907.1 says, "It is the custom of the Huns (E.Pulleyblank: Hsiung-nu) to have three lung sacrifices per year. On the day wu [ ] of the first, fifth and ninth months they always sacrifice to the god of heaven". The word Lung also occurs in the inscription composed by Pan Ku in honor of the general Tou Hsien, whom he praises for having "burned the Lung court of Lao-shang (the successor of Mao-tun)" [ 老上單于] (Hou Han-shu 53.0746.1). Some of the Chinese commentators try to explain Lung as meaning "dragon", saying that the Hun (E.Pulleyblank: Hsiung-nu)'s chief god was a dragon. Takigawa was however no doubt right in concluding that lung was a foreign word of unknown meaning and had nothing to do with "dragon" (see Shiki kaichü köshö 110, p. 23). For this reason it was sometimes written with the addition of the grass radical or the bamboo radical. Now we have independent reason for reconstructing an initial cluster in this Chinese word M. lioŋ < "vloŋ (see p. 137 above). In the foreign original we once again probably have to do with a word related to the name Hun (E.Pulleyblank: Hsiung-nu) (Dragom can't be excluded, because it was carried on for millennia, even into Europe, and to ask the Sky to bless a drago standard is no different from asking Jesus to bless a drago standard, the standard should not be easily confused with God for the sake of an argument).
We must therefore reckon at least with clusters of the type fr-, vr- (or perhaps bilabial φr-, βr-) in Hunnish (E.Pulleyblank: Hsiung-nu) and it is furthermore likely that clusters also occurred where Chinese has *kl-, *gl-, *tl-, *dl-, possibly also where Chinese has *hl-.
This includes the word for "heaven". The correct reading of the first character [ ] is not quite certain. The only pronunciation indicated in the Kuang-yün is M. thaŋ but the gloss of Yen Shih-ku, the commentator of the Han-shu implies M. daŋ. Furthermore the Chi-yün indicates an alternative taŋ. (It may be noted that if the usual reading of the character in th- is the correct one, this is a rare case of an aspirated surd stop in a Hun (E.Pulleyblank: Hsiung-nu) transcription in the Shih-chi or Hanshu) Whichever reading we adopt, we are left with a Middle Chinese supradental stop which implies a thl-, dl- or tl- cluster in the Han period. This in turn would probably point to *tr- or *dr- in Hun (E.Pulleyblank: Hsiung-nu). The fact that no -r- appears in the initial of Turkish täŋgri, etc., does not prove that it did not exist in the Hun (E.Pulleyblank: Hsiung-nu) word, since such a cluster, being contrary to the phonology of those languages, would have had to be eliminated either by simplifying to a simple dental stop or introducing an extra syllabic vowel.
The other general features of Hun (E.Pulleyblank: Hsiung-nu)
phonology which can be deduced from the Chinese
transcriptions do not give such definite evidence against Altaic connections but do not
agree closely with either Turkish or Mongol (What about Chuvash,
or European runiform and Greek-alphabet inscriptions, they do not matter?).
A notable feature of the Hun (E.Pulleyblank: Hsiung-nu) transcriptions is the absence of aspirated surds. We find voiced and unvoiced stops: *k, *g, *t, *d both initially and medially but very rarely *kh, *th. The only exceptions I have found in the Former Han transcriptions are in the doubtful case of the word for "heaven" and [ ] M. khiuŋ-lio "yurt", for which variant spelling with M. k- or g- are also found. We also find the affricate *tsh. Here we probably have to reckon with the fact that the Chinese dental affricate may have been used for an unfamiliar palatal affricate. This feature of Hunnic (E.Pulleyblank: Hsiung-nu) phonology would argue strongly against its having been a form of (modern) Mongolian, since the surds of (ancient) Common Mongol are considered to have been strongly aspirated. This does not seem to have been true of (modern) Turkish however and, to this extent, Hunnic (E.Pulleyblank: Hsiung-nu) phonology would be close to Turkish than Mongol.
In transcriptions of the Later Han period we begin to find cases of *kh- and we find *th also in the transcriptions of the Chin-shu. These may.perhaps reflect increasing penetration and admixture with the Eastern Ηu, that is the Syanbi (aka spelled Xianbi, E.Pulleyblank: Hsien-pi) and Wu-yüan ( 服务员 “service person”, aka Wu-huan or Wu-wan) [ ] (or [ ]) M. *ou-hwan < *ah-hwan=Avar, who probably spoke a Mongolian type of language. It was the Syanbi (E.Pulleyblank: Hsien-pi) who became dominant on the steppe after the collapse of the Hun (E.Pulleyblank: Hsiung-nu) empire in the second century A.D (Syanbi own language has not been reliably established, supposedly it was Mongolic or as previously thought Tungusic, nor was established their language of administration. Probably, the ancient Mongolic should be viewed as a branch of Tungusic already strongly impacted by Türkic, along the lines of the Slavic branch of Baltic. Eynically, Syanbi were predominantly Türkic since they included the Türkic tribes of Türks, Tele, Huns, Toba, and Usuns. It is unknown what the Avars spoke in their kitchens, but every inscription ascribed to the European Avars reads in Türkic).
In the Hun (E.Pulleyblank: Hsiung-nu) transcriptions it is noteworthy that we find only *b- initially, never *p-. Medially the situation is reversed, we find only *-p- never *-b-. This is like both Old Turkish and Mongolian, though initial *p- is supposed to have existed in the latter at an early stage (Poppe 1960) (Applying a Indo-European scale to non-Indo-European type of language may have a value, but not applying a well-known peculiar scale is an omission; "b"/"m" alternation is a well-known phenomenon of the Türkic languages, and it would skew the rationalization in "*b-"-"*p-" comparisons if the Hun language was an "m" language. Among the life Türkic languages, Chuvash and Kazakh stand alone with 30-100 times lesser frequency of initial "*b-" and "*m-" then the other Türkic languages, which can be read as their closer affiliation with the ancient Hunnic language, and an order of magnitude more suitable for comparisons than the modern Turkish.)
The existence of initial n- in Hun (E.Pulleyblank: Hsiung-nu) distinguishes it from Old Turkish (though not, perhaps, from ρre-Turkish). Initial n- exists in Mongolian (The initial n- in miniscule quantities exist in all modern Türkic languages, except for Tele-derived S.Altaian, here no educated judgment can be made without comparing the numbers and ratios).
It seems clear from this investigation that Hun
(E.Pulleyblank: Hsiung-nu) was not closely like any form of
Turkish or Mongol ("any form" seems to be misrepresentation,
since all Türkic languages except Turkish were excluded from the comparisons)
which we know of and is unlikely to have been Altaic at all. Of the
other neighboring language groups, we can quickly rule out Sino-Tibetan. Hunnish
(E.Pulleyblank: Hsiung-nu) was
clearly a polysyllabic language and the presence of only two series of stops as compared
to the three of Chinese and Tibetan would also distinguish it from any but a very evolved
or aberrant form of Sino-Tibetan. No evidence of genetic connections has been shown with
Indo-European, though there are quite likely to have been mutual borrowings between the
Huns (E.Pulleyblank: Hsiung-nu) and the neighboring Tokharians
(under "Tokharians" are meant the religious languages of the
settled populations in the Kucha and Karashar oases, arbitrarily projected a millennia
backward to artificially bring them into the Hunnic period) and perhaps also with the Iranians farther west.
(Cf. Maenchen-Helfen 1945 (2). The proposals made there cannot be regarded as certain
however) (Under under-defined "Iranians" most likely are meant
Sogdians/Chorasmians). There is however one other little known, quite distinct language group that must
be taken into account, namely the Yenissei family, of which the only surviving member is
Kettish, otherwise known as Yenissei-Ostyak.
This language was studied in the middle of the last century by the Finnish linguist Castren, who also recorded and studied the related Kottish of which he was able to find only five living speakers and which has since become extinct. By his time the Arins and Assans, of whose languages a few words had been recorded by eighteenth century travelers, had died out or been absorbed into the surrounding peoples. Since Castren's day a little more material on Kettish has been published, notably that collected by Donner. Some progress has been made in the analysis of this meager published material, especially by Lewy and Bouda who have analyzed the complicated morphology of the verb in Kettish and Kottish. A phonological analysis of Kettish is still very much needed however. Still less has there been a systematic comparative study of Kettish and its extinct relatives. Though the material is scanty, it seems likely that some correspondences could be established which would enable us to reconstruct some features of earlier, common, stages of the language. (For general information and bibliography see Jakobson 1952 and 1957.)
Much of the attention that has been paid to these languages so far has been with a view to showing genetic relationships with other language groups. Some scholars have indeed persuaded themselves that the Yenissei languages are a northern branch of the Sino-Tibetan family. Both phonologically and morphologically the contrast between Kettish and Kottish on the one hand and Chinese or Tibetan on the other could scarcely be greater. This could of course be the result of divergent evolution if it could be shown that there was an underlying basic vocabulary in common but the efforts that have been made to show this have so far been unconvincing. There are a few striking word comparisons but these may well be explicable by early borrowings, especially if it should turn out that the Yenisseians and the Chinese were once contiguous.
As far as Hunnic (E.Pulleyblank: Hsiung-nu) is concerned, the Yenissei languages must come into question since Ligeti 1950 showed that the Hu word for "boot" recorded in Chinese texts as [ ] M. sak-dok < *sek-δak closely resembled the Kettish sagdi (Imbazk dialect), shagdi (Bakhta dialect) "boot".
The Hun (E.Pulleyblank: Hsiung-nu) word for "boot" may be ultimately Iranian in origin if the etymology suggested by Professor Sir Harold Bailey is correct. He has kindly supplied me with the following note.
"Iranian sak-, participle saxta-, and sak, participle satta-. Avestan saxta- in
xvainisaxta- "finely equipped"; Sogd. Manichaean ptsγt-. participle ptsγttyy •patsaγdai
(-ē) "prepared, adorned"; Sogdian in New Persian loanwords
asaγdah, pasaγdah "prepared";
Ossetic säxtäg "a fastening"; Zoroastrian Pahlavi saxtax in the phrase mochak saxtak "boot
of prepared (leather)"; New Persian saxtiyan "leather tanned with sumac (morocco
leather)". See W. B. Henning, BSOAS 13, p. 644. With -a- occur Zoroastrian Pahlavi saxt
"prepare, equip", saxt and saz "equipment"; Pahlavi asp zēn
sachēt "he equips the horse
with a saddle"; New Persian asp-ra zēn saxtan "to harness a horse with a saddle";
Armenian loanword an-saxt "not equipped", saxteal "harnessed"; Ossetic sadzun, saγd,
saxta "put in, put up, put on"; Sogd. Manich. pts'k *patsak "preparation"
(Alternate opinion is that Scythian riders introduced the Türkic
term for the boots, which was absorbed into many European and Middle Eastern languages:
zapata, sapog, etc).
"These Iranian words attest a base sak- and sak- indicating "prepare" and in specialized senses "build, equip, harness, fasten" and in reference to mochak "boot" a specialized form of leather. A word saxtak, or saγdaγ could have existed in the second century B.C. meaning "equipment", from which specialized meanings could have derived. The Hunnic (E.Pulleyblank: Hsiung-nu) so-to from *sak-dak "boot" might be this same Iranian word."
[Though the specialized meaning "boot" is not attested in Iranian, it might well have developed in some Scythian dialect and thence been borrowed into Hunnic (E.Pulleyblank: Hsiung-nu). The Huns (E.Pulleyblank: Hsiung-nu) certainly had cultural features, such as the worship of the sword (see below) which were probably derived from the Scythians. Besides the words meaning "boot" in Hunnic (E.Pulleyblank: Hsiung-nu) and Yenisseian, it is possible that we should see the same root in Mongol saγadaq Turkish sadaq "bow case, quiver" (see Boodberg 1936, p. 174). E.G.P.J.
It is difficult to compare the phonology of Hunnic (E.Pulleyblank: Hsiung-nu) as revealed by Chinese transcriptions of the Han period with the imperfectly known Kettish and Kottish of two millennia later. At present Kettish admits of initial r- as little as Altaic. Initial dusters are also excluded. On the other hand initial l- occurs. (Kottish d'-, occurring initially, is in complementary distribution to l, occurring medially and finally.) These exclusions may however be the result of a long evolution in contact with Altaic neighbors. Roughly the same evolution has occurred in Chinese, perhaps from the same cause (Talking about ethnical evolution, Pulleyblank was up to something: during their millennium-long stay in Siberia, the tribes of cattlemen, call them Huns or Turks or Tele, absorbed the Kets, and by the time they reached Gansu they did have Kets with them, whether in dissolved condition, or with some chunks still floating between them. After spending a millennium+ with the Kets, intermarrying their women, absorbing their men, it would be a miracle if no traces of the Kets survived. Thus, though the nomadic cattle-breeding economy and the language is Turkic, linguists and geneticists find traces of the Ket admixture.)
Phonology is not helpful therefore but in terms of vocabulary a number of further comparisons between Hunnic (E.Pulleyblank: Hsiung-nu) and the Yenisseî languages can now be brought forward.
Ku-t'u [ ] "son" (N.Bichurin phoneticized it Gudu)
The full title of the Hun (E.Pulleyblank: Hsiung-nu)
ruler is given in the Shih-chi and Han-shu as Ch'eng-li
ku-t'u shan-yü. The first word, which is explained as' meaning "heaven", has already been
discussed. The last, which is the abbreviated title by which the Hun
(E.Pulleyblank: Hsiung-nu) rulers were
normally referred to, will be discussed below. The middle word ku-t'u < M. kou-dou <
*kwah-δah is said to mean "son" and Ch'eng-li ku-t'u is said to mean "son of Heaven",
like the Chinese T'ien-tzu. It would be very natural for the Huns
(E.Pulleyblank: Hsiung-nu) to have borrowed
this Chinese title for their own supreme ruler, as did the Kushans later; but in what
language can we find a word; like this meaning "son"? Nothing at all similar can be found
in Turkish or Mongol ("soul" fits the semantics, since the soul comes from Tengri Sky,
in Türkic soul is "kut", and "Sky-[given] soul" is a standard expression, corresponding to
the Chinese "Son of Sky". Müller misunderstood it as "majesty"). Shiratori
found the word guto "son" in Tungusic and this was the mainstay of his opinion that the Huns
(E.Pulleyblank: Hsiung-nu) belonged to that linguistic group; but the
phonetic resemblance is at best vague even to the Modern Chinese form and it fades away
when we reconstruct the probable Han dynasty value. Others have thought that the Chinese
were wrong in what they said about the word's meaning and F. W. K. Müller suggested
that ku-t'u might stand for Turkish qut "majesty", often found in titles. This too
becomes less plausible in terms of the Han dynasty form. The Chinese certainly had the
vowel a in both syllables and the initial d- of the second j syllable is shown by its
hsieh-sheng series to have come from *δ-, which is likely to have been used for foreign
in Former Han times. The foreign original ought therefore to be something like *kwala
(In modern Türkic languages “male child” or “son” is ogul,
yvăl, ool, og'ul, ulan, ul, ül, ůgil, ool, uol, oolax, uul, half of which have
obvious cognates of k/g-V-l, and since phonological arguments are not
applicable [A.Vovin, 2000], there is no need to turn to exotic Old Japanese and Tungusic. The g is silent
in half of the Türkic languages, so ool = oĝol,
etc. No need to go to the exotic dead Arin or phantom *Proto-Yeniseian, the simple English ulan
will do fine, it is a kin of the Hunnic “son”, no pun intended).
The ordinary word for "son" in Kettish is fyp (Sym dialect), hyep (Imbazk), hyp (Bakhta) and this word was also found in Kottish as fup. In the extinct Arin language however we find the word bîkjál "son" (also bikjaljá "daughter") where bi appears to be a prefix added to nouns of relationship: bjapp "father", Kottish op, Kettish ob, op; bjamja "mother", Kon. ama, Kett. am; bamagál "brother"; bikhei "lord", Kott. hii, Kett. qyj (Donner), etc. The word -kját is no doubt cognate to Kettish qalek' "younger son, grandson" (Donner) and probably also to falla "son" in the extinct Pumpokol dialect. Here we have then a word for "son" which agrees very closely with the hypothetical Hunnic (E.Pulleyblank: Hsiung-nu) form required by the Chinese. Being a word for a fundamental human relationship, it is unlikely to be a loanword in Yenisseian and unless it is an extraordinary coincidence it creates a presumption that the Huns (E.Pulleyblank: Hsiung-nu) belonged to that language group. (For extinct Yenisseian languages, see Klaproth 1823.) (Its a quick presumption. Before jumping to presumptions, a preamble would be appropriate, explaining how the traditional foot forest hunters became the world-renown riders, and managed to overpower any other pretenders on leadership of the tribes of fast-moving riders, which managed to use their superb mobility to subjugate the slow-moving and stationary Ket population in their (Ket's) own backyard. These are hard facts that underlie the philological development, not the other way around.)
Chüeh-t'i [ ] "horse"
At the beginning of the chapter on the Hun (E.Pulleyblank: Hsiung-nu) in the Shih-chi occurs the following passage, "The livestock which they keep in large numbers are horses, cattle and sheep. Their uncommon livestock are camels ([ ]), asses and mules, chüeh-t'i, t'au-t'u [ ] and t'o-hsi [ ]" The last three words are clearly non-Chinese.
The Shuo-wen says (literally) that chüeh-t'i means the offspring of a stallion and a mule. This is of course an impossibility and the text should no doubt be emended to read "a mule which is the offspring of a stallion and a she-ass", that is a "hinny" (Turkish katır, inatçı, terlik. "Katir" fits nicely into "chüeh-t'i)". This definition does not agree with the commentary of Hsü Kuang quoted in the Shih-chi chi-chieh who says that a chüeh-t'i is a superior type of horse of the northern barbarians. A hinny is, in fact, a weak, inferior type of mule (Mules are preferred to horses for draft duty, and that's why they were used in mines, mills, and as best beast of burden during American Civil War, when to win the war the Union Army used 50,000 mules for supply trains). Egami Namio (1948, pp. 180 if.) has gathered a number of other passages in which the word chüek't'i occurs and shows that they only make sense in terms of Hsü Kuang's definition. He thinks that the chüek't'i was in fact the larger western horse in contrast to the ordinary Mongolian type that was normal both in China and in Mongolia (Mule is a mule, not a western or Mongolian horse. Since Chinese had their only supply from the eastern nomadic cattlemen, they did not have much of a choice, though breeding to order can't be excluded).
One passage which illustrates very well that the cküeh-t'i was a prized, superior sort of
animal and incidentally shows that it was already known in China before the unification of
China by Ch'in in 221 B.C. is found in a memorial of Li Ssu to the King of Ch'in contained
in Shih-chi 87.0215.1. Li Ssu, who came from Ch'u, was defending himself against ministers
of Ch'in who wished to expel all the foreign advisers ([ ] "guests") at the Ch'in court
(Ch'u = Chui valley inhabited by Chui tribes, later Chumi, Chue,
and Chumul ?).
By way of illustration he pointed to the many precious things which Ch'in obtained from
beyond its borders. "If things must be produced in Ch'in before they are permitted", he
said, ". . . the women of Cheng and Wei would not fill your inner palace and fine
chüeh-t'i would not supply your stables without."
Other passages in Shih-tsu (quoted in T'ai'p'ing yü-lan 773-5b) and Huai-nan-tzu II. refer to chüeh-t'i as the best kind of horses for drawing a carriage. In the biography of Tsou Yang in Shih-chi 83.0209.1, in another passage which seems to indicate the knowledge of this animal in late Chou times, it is said that the King of Yen served chüek-t'i to Su Ch'in to eat, the implication being that it was a mark of special favor. The nomads of course commonly ate horseflesh and the larger chüek-t'i was probably better eating than the wiry Mongolian pony (E.Pulleyblank is no horsemeat gourmet: with herds of thousands, Huns could afford to keep dedicated meat and milk and riding and draft herds, and would not eat a beast of burden when they have foals. But for foals, the type of foal would not make much difference. The trick is that Chinese could not breed mules, they were dependent on fresh supplies from the Huns, so they only had old mules to eat, a rare delicacy in their eyes). No doubt the comparative rarity of the animal would also lead to its being considered a delicacy.
Though Egami's opinion about what was meant by the chüeh-t'i seems very acceptable, his view, taken originally from Shiratori, that the word could be connected with the Mongol word külüsün "sweat" because" the western horses were later known as "blood sweating" horses seems very-far-fetched. But there does not seem to be any other suitable word in" Turkish or Mongolian to compare with M, kwet-dei < "kwet-deh (sic!).
On the other hand we find in Yenisseian that the ordinary word for horse is Ket. kus. Kot. hus, pi. huchan, Pumpokol kut, kus. There are other words in which a -t in Pumpokol corresponds to -s or -sh in other dialects, e.g.: "eye" Ket. des, Kot. tish, Pumpokol dat; "house" Ket. k'uos, xus, Kot. kush, Pumpokol hukut (For the mule subject these examples are irrelevant, but jumps in the eyes the closeness of the Ket “house” with Türkic derivatives: Serbian “kosh”, Ukrainian “hata” - the correspondences are amazing); "one" Ket. husä, kuhşäm, Kot. hûcha, Pumpokol chuta; "stone" Arin. khes, Pumpokol kit (it seems that Kot. shish, Ket. tyes probably comes from a different root, since Pumpokol had a second word chys "stone" evidently related to them). From this it would appear that we should be justified in supposing an earlier form in Yenisseian something ri like *kuti, the final -i having caused the palatalization of the preceding t in some dialects before disappearing. This gives a reasonably good equivalent for Chinese *kwet-deh. To agree with the Chinese a form *küti would be better than *kuti, cf. *koh-kwet = Mongol kökül (p. 259 below) (All these inferred reconstructions and comparison of inferred reconstructions totally unrelated to the word “mule” steps over a dead body of the Turkic real “katır” and guides us to the refinement of the envisaged Ket word unrelated to the subject of drawing a carriage. The words "t'au-t'u" and "t'o-hsi" are not analyzed even in this fanciful content. Notably, all three borrowings indicate a presence of Turkic "at" = "horse", supplied with a definition, as it should be rightfully expected).
Chiek [ ] "stone"
With the Yenisseian word khes, kit "stone" may be compared the name Chieh [ ] M. kiat <
*kat applied to an important branch of the Hun
(E.Pulleyblank: Hsiung-nu) in North China in the fourth century
From this people, who constitute one of the so-called "five barbarians", i.e. the
Huns (E.Pulleyblank: Hsiung-nu), the Chieh, the Syanbi (E.Pulleyblank:
the Kangars (E.Pulleyblank: Ch'iang) and the Ti, who vied with one another in
setting up short-lived dynasties at this period, came Shih Lo [ ] the founder of Later Chao (reigned A.D. 319-334). His Chinese surname means "stone". There is good reason to
think that the Chieh were not properly part of the Huns
(E.Pulleyblank: Hsiung-nu) but had more western,
Indo-European affinities (T'ang 1955, p. 416). They had however entered China as part of
the Southern Huns (E.Pulleyblank: Hsiung-nu), and were therefore sometimes referred to as Huns
(E.Pulleyblank: Hsiung-nu). The Chin-shu 104.1354.4 says of Shih Lo, "He was a Chieh from Wu-hsiang
district in Shang-tang country (in southern Shansi).
His ancestors came from the separate tribe of the Huns (E.Pulleyblank: Hsiung-nu), Ch'iang-ch'ü [ ] M. khiag-gio" (see T'ang 1955, pp. 414 ff.). Ch'iang-ch'ü is probably just another form of K'ang-chü [ ] M M. khaŋ-kio. The Kangars (E.Pulleyblank: K'ang-chü) were of course an important people in Sogdiana in the Han period. They later gave their name to Samarkand but in the Former Han period were centered around Tashkend (Tashkent, from Türkic “tash” = “stone” + “kend” = “village, city”). The Ch'iang-ch'ü group in the Hun (E.Pulleyblank: Hsiung-nu) were presumably a part of the Kangars (E.Pulleyblank: K'ang-chü) people who had at some time been captured and incorporated by the Hun (E.Pulleyblank: Hsiung-nu) (This phrase does not make sence: if Ch'iang-ch'ü spelling is just another form of K'ang-chü, they are one people, they can't be a part of themselves as a whole). Now it happens that Tashkend (Tashkent) was later known in China as Shih Kuo "Stone Country" and people from there who came to China took the same surname Shih "Stone". Tashkend (Tashkent) itself means "Stone City" in Turkish. This is usually regarded, following Marquart (1901, p. 155), as simply a Turkicization of the earlier Chach, but this does not account for the Chinese name which is long before the region became Turkish (For a scholar this should be an indication of the fallacy of the Turkicization hypothesis. Also, "became Turkish" and "became Türkic" are two different scientific categories).
The Kangars (E.Pulleyblank: K'ang-chü) people are usually thought of (or rather were heavily promoted) as Iranian but they had close links with Ta-yüan (= *Taxwār, Tochari) and the Ases/Tochars (E.Pulleyblank: Yüeh-chih) and they shared the title hsi-hou = yabgu with the latter (Ases/Tochars) and the Usun (E.Pulleyblank: Wu-sun). It is quite likely therefore that they too were Tocharian (Here Tocharian does not mean sedentary Kucheans, but supposedly Iranian-lingual phantom nomads; the confused terminology serves to muddle the picture) in origin and that they moved into Sogdiana as part of the same westward movement that brought the Ases/Tochars (E.Pulleyblank: Yüeh-chih) (under Iranian-lingual nomadic concept. In fact Asses and their kyshtyms Tuhs Tochari belonged to the local, Middle Asian Tele superfamily) and then the Tochari spilling over the Pamirs (coming of the Kuchean religious missionaries is not a “spilling”). In this case we may look in Tocharian (i.e. reputably Iranian) for an interpretation of their name. It happens that there is a word kānka- in Tokharian (i.e. reputably Iranian) A about which Sir Harold Bailey has kindly given me the following note.
Tocharian A dialect: kānk- in the plural kankan; kānkuk in singular and instrumental singular kānkukyo. The context of kānk- in 264a2 is knkan wasirssān, that is “the kānk-
1. consisting of vajra - stones
The context preceding kānkan (stone -an) contains a list of weapons, including Bud. Sanskrit cakra- "discus", tomara- "lance", sakti- "spear", trisula- "trident", bhindipala-"missile". The vajra- (stones - missile) in battle scenes in Bud. Sanskrit texts is a missile used, for example, by Yaksa goblins. In the present context /// sānāri is a plural from /// sānār-. It may have lost an initial syllable, It could possibly have come from Bud. Sanskrit asani- (Pali asani) "a missile", sometimes identified with vajra- (stones - missile) but occurring also in the Sanskrit compound vajrāsani- (stones - missile-sani- ). This word came into Khotanese (older) asuni- and (later) asū'na (s = z).
The word kānkuk occurs twice,
1. 24bı vājär kānkuk srivas panont şotreyäntu..., that
is "the vajra (stones - missile) kānkuk
(stone -uk) srivatsa- sign, splendid ( = auspicious) marks ..." (the vajra
(stones - missile) and
srivatsa- as auspicious signs are known),
The above contexts seem to assure a Tocharian (Kuchean?) word kank- meaning "stone". A corresponding word in Tocharian Β (Karasharian)dialect has not yet been pointed out.
(As a conclusion, as all the H. Bailey examples indicate, the conjecture that kank- meaning “stone” at least is not corroborated by anything, if not outright tenuous, it may never be corroborated. Unlike the “tash” = “stone” which is a solid fact)
To support this interpretation of kānk- as "stone", note that Tibetan translate vajra- as rdo-rje (dorje), explained as rdoyi rje "prince of stones". The dorje in stylized thunderbolt shape has a place in Buddhist ritual. The Indian missile vajra- corresponds in the Avesta to vazra- which is an ordinary human weapon, identified in later texts as a club, the New Persian gurz. In later texts however vajra- is used foe a hard stone and especially for the diamond. B. Laufer, The Diamond, discusses the history of the use of vajra. There are two Sanskrit compounds: vajra-gulaka "a kind of weapon (gulaka- "ball"), and vajra-kila (kila- "wedge").
E. Sieg and after him W. Couvreur proposed hesitatingly to see in kankuk an Indian plant name kanguka- “a kind of panic grass”, but there is no justification for this view.
Although a connection in other Indo-European languages can easily be proposed, an etymology is not useful in assigning a meaning to kānk-. But if kānk- does mean "stone", a connection would tie with Old Slavonic kamy, gen. sing. kamene, Russian ęŕěĺíü “stone”, from 'kŕ-men- as we have Greek "anvil", Lithuanian akmuo, ashmuo "stone". Old Ind. asman-, Avestan asman-. That is, a base ak- : k- and, with enlargement -ā-, also k-ā-. The Sanskrit word asani- "missile, thunderbolt" is derived the same ak-, as also is Avestan asan "stone".
I have no other satisfactory example of the suffix -uk in an indigenous Tocharian (Kuchean) A dialect word. There is aşuk "āyata-, extended", possibly from as-uk. In knuk "neck" it has been proposed to see the cognate of Middle High German knock "neck" but it might be rather kn-uk.
The probability that kank- in Tocharian (Kuchean) means "stone" and the indications from totally separate evidence that the name of the Kangars (E.Pulleyblank: K'ang-chü) people had something to do with "stone" makes it extremely likely that it is this Tocharian (Kuchean) word which enters into their name. It further makes it probable that the alternative name Chieh < M. kiat is the Hun (E.Pulleyblank: Hsiung-nu) word: for "stone", related to the later Yenissei words quoted above.
Chach, the old name of Tashkend (Tashkent), found for example in Chinese transcription as
[ ] M. cia\-cie, [ ] M. cia\-ciet, [ ] M. cia\-jie, and in Manichaean Sogdian ch'ch'ny=*chachane
"a native of Chach" (Henning 1940 p. 9), might also be connected with another word for
"stone" in Yenisseian,* namely Ket. tyes. Kot. sis, Pumpokol cys. This would presumably
be a relic of the Huna occupation of Sogdiana in the fifth and sixth centuries. The
name Kangars (E.Pulleyblank: K'ang-chü) is no doubt
also connected with [ ] M. kham\-kiat (for *Kamkar?), the
name by which the capital of Shih" (Tashkend (Tashkent)) is mentioned in connection with the setting
up of an administration in the region by the Chinese in A.D. 658 (see Chavannes 1903, p.
141), and to the name Kankar given to the lower Yaxartes in Ibn Chordadhbih (Marquart
1898, p. 5, n. 5, also 1914, p. 168). Marquart also suggested, connections with the
Kängaräs mentioned in the Orchon inscriptions and with Kaγγar, Κaγkar, said in
Konstantinos Porphyrogennctos to be a name of honor given to the three leading tribes of
the Badjanaks-Besenyos (E.Pulleyblank: Pechnegs) (Moravcsik 1958, II, p. 145); but these further questions cannot be pursued
here (Because the story goes in a wrong direction,and doctrinal
Iranians start speaking plain Türkic?) .
Animal milk and its products did not form part of the normal diet of the agricultural Chinese in ancient times and do not today; but they became known when the Chinese came into contact with pastoral nomads of the steppe and they even enjoyed a certain vogue at some periods, especially from Han to T'ang. With the foreign products came foreign names and since the dominant steppe people at the time of their introduction were the Huns (E.Pulleyblank: Hsiung-nu) the names were naturally also taken from the Huns (E.Pulleyblank: Hsiung-nu) language. In some cases there is direct testimony to this. In others it can be inferred with a high degree of probability.
The milk culture of the steppe has been alluded to or described by many writers since Herodotus (Book IV. 2) mentioned the use of mare's milk among the Scythians. Zemarchus, the Byzantine envoy to the Turks at the end of the sixth century, referred to the sweet wine not made from grapes that he was offered at the court of the kagan, which must have been koumiss (Chavannes 1903, p. 232). William of Rubruck, who visited the Mongol court in the twelfth century, gives a detailed account of cosmos (i.e. fermented mare's milk, kumiss) and its preparation, of the preparation of butter and the dried curd or gruit (=Turkish qurut) which they kept for winter provisions. He also refers to a drink called aira made of sour cow's milk (i.e. Turkish airan or Mongol airaγ).
What he has to say agrees substantially with what the eighteenth and nineteenth century travelers relate, the most notable difference being the absence of any clear reference to distillation of a strong spirit from kumiss, the Turkish araki, Mongol araki. The method of preparing the superior "black cosmos" served to the lords is not very clear. He says, "They churn then the milk until the thicker parts go straight to the bottom, like the dregs of wine, and the pure part remains on top, and it is like whey or white must. The dregs are very white, and they are given to the slaves and they provokes (sic) much sleep. This clear (liquor) the lords drink, and it is assuredly a most agreeable drink and most efficacious." The translator notes, "These dregs are called bossa by the Kalmucks - see Pallas, I, 511" (Rubruquis, p. 67). The bossa described by Pallas are the lees of distillation produced in making araki, but there is nothing in the process described by Rubruck to suggest distillation. It seems to be some other method of removing some of the milk solids from the fermented mare's milk.
It was in any case just about this time that the knowledge of the new invention of
distillation was spreading eastwards as well as westwards from the Arab world. Li
Shih-chen, the author of the great Ming pharmacopoeia, the Pen-Cs'ao kang-mu, says of
hsiao-chiu "brandy", "It is not an ancient method. The method was first started in
the Yuan period." He also quotes the name a-la-chi [ ], i.e. araki, from the Yin-shan
cheng-yao. The Yin-shan cheng-yao was produced by the Food and Drug Department at the Yuan
court and was presented to the throne in 1331. It contains a recipe for a-la-chi wine
which reads as follows, "Its taste is sweet and pungent. It is very hot and has much tonic
property. It is good for reducing chill, strengthening resistance (?) and expelling cold
Take spoiled (nu ? wine and boil it. Collect the "dew" and it will make a-la-chi."
In view of the fact that distillation began in the Arab world it is easy to accept the etymology for araki that is usually proposed, namely that it is from Arabic 'araq which originally meant "juice" or "sap" (Yule 1901) (E.Pulleyblank dives into ethnology and forgets that Hun and Türkic dynasts were famous for their love of alcohol long before Arabs burst onto the stage. Araki did not have to start as a distilled drink, the higher sugar content of the horse milk makes it much stronger drink then from the cow milk).
It is clear (sic!) that the Chinese word lao [ ], which appears already in the second century B.C. to refer to the typical milk drink of the nomads can have nothing to do with arrack, raki, araki, etc., as Karlgren thought (1926, p. 138). Lao is always a milk product and may or may not be alcoholic. Arrack, etc., always means a distilled spirit and only becomes specially associated with kumiss among the nomads who naturally used kumiss as the main material from which to make it. (Poor old naïf Karlgren... fell for plain phonetic resemblance...)
* * *
Excluding the as yet unknown araki, we find foreign names for the following typical nomad milk products in the Han period: (1) milk, (2) sour milk (3) butter (4) fermented mare's milk, (5) dried curd.
The Chinese had, of course, a word for "milk", namely [ ] ju < M. njou < *no. It means "nipple", like Tibetan nu-ma, hence "to suckle" and more generally "to care for the young", as well as "milk". The modern colloquial [ ] nai < M. nəi', implying an earlier *nə is no doubt a variant form of the same word. Since the Chinese did not use animal milk however, it is not surprising to find that they sometimes used a foreign word for "milk" in this sense. (Funny, really E.Pulleyblank knows that the Chinese moms did not produce milk, and needed foreigners to educate them?) The word [ ] tung occurs already in the Mu Tien-tzu chuan in the phrase "milk of cows and mares" given as present to King Mu by a barbarian tribe. Assuming this passage to be genuine, it must date from before 296 B.C., that is well before the Han period. The next occurrence is in the chapter on the Huns (E.Pulleyblank: Hsiung-nu) in the Shih-chi. The Han envoy Chung-hsing Yüeh, who went over to the Huns (E.Pulleyblank: Hsiung-nu), is reported to have urged the Shan-yü, "When you get Chinese food, throw it all away to show that it is not as suitable and pleasant as tung-lao [ ]." (The Han-shu reads chung [ ].)
The word [ ] appears in the Shuo-wen, where it is defined as [ ] "milk fluid" (no spelling for the word? Why do we need it then?). In Chapter 6 of Lieh-tzu (ca. A.D. 300) we find the expression tung used of human milk. "At your birth there was too little vital fluid-& your mother's womb and too much milk in your mother's breast." (Graham 1960, p. 128-29.) Such usage is rather rare but we do find tung as a ti variant of ju "milk" sometimes in early translations of Buddhist sutras (see T.5J: (I.1615), T.i67oa and b).
What appears to be a variant of the same word appears in the expression t'ung ma
which appears in Han-shu 19A.0353.2.
In the Table of Officials and Ministers it is recorded that in the first year of T'ai-ch'u (104 B.C.) of the Emperor Wu the title of the prefect (ling) of the Household Horses [ ] was changed to T'ung-ma. The commentary of Yin Shao (2nd C. A.D.) says, "He has charge of the milk horses ([ ]). He takes their milk and prepares it by "shaking" (t'ung). It tastes sour but can be drunk." The commentary of Ju Chun (third century) says, "He is in charge of the milk horses. One makes a bag ([ ] M. kaəp-tu) of skin with a capacity of several tou and fills it with mare's milk ([ ]). One "shakes" it (t'ung) and takes the fat off the top and hence they are called t'ung horses." In translating t'ung as "shake" I follow the usual interpretation, which is necessary to make sense of Ju Chun's commentary; but it makes a curious epithet for horses. The commentators are evidently taking [ ] as equivalent to [ ] "move". This interpretation of t'ung is made explicit in the commentary of Li Ch'i [ ] to another reference to t'ung-ma which occurs in the Monograph on Ceremonial and Music - where it is stated that t'ung-ma wine ([ ]) shall be given to certain officials. Li Ch'i says, "They make wine of mare's milk. They beat and shake it ([ ]) and so it is made." Yen Shih-ku adds, "[ ] has the sound of tung [ ]."
The expression chia-tou < M. kaəp-tu is not found elsewhere and does not appear in any dictionary. It must be related to Mongol xabtaγa(n) "bag, pouch, purse, pocket" and to Turkish qaptirγai "ein grosser tiefer Sack". In Rubruquie we find captargac "square bag for putting unconsumed food" (p. 65). The Chinese word would imply *klep-toh with an initial cluster but it is uncertain whether the initial cluster would have still survived for Ju Chun in die third century A.D. (Just another example that the cluster idea does not hold the water. Russian "kapterka" also belongs here, without any clusters).
Though the Yü-p'ien defines the character [ ] as meaning [ ] tung "move", it does not appear to occur in this sense in any other context. The Shuo-wen defines it as meaning "to push and pull" ([ ]) and uses it, evidently in this sense, in the definition of the character [ ] "to repress". In Huai-nan-tzu occurs the phrase [ ], to which the commentator remarks, "ting-tung is like 'up and down' " [ ], It means 'seeking what is convenient and profitable' ". This is evidently an expressive binome, not to be analyzed into separate components (Looks that E.Pulleyblank expects that as in English, "a milk" and "to milk" in Chinese have to be the same word, and is puzzled that it is not. "To milk" is movement, up and down, and suck as far as the money are concerned, etc. It can also be semantically used like "dough" in "bring the dough", or "produced" like in "freshly produced milk or beer".)
It seems evident that a much simpler explanation than that given by the commentators would be to take [ ] as simply another writing for [ ] "milk". It should be noted that there is also a word [ ], homophonous with [ ], defined in the Kuang-yün as [ ] i.e. kumiss.
Since [ ] has two readings M. duŋ and M. duŋ' and
[ ] has several readings, M. tuŋ', toŋ'
(Kuang-yün) (the latter quoted from the Tzu-lin), tuŋ',
tonŋ' (added by the Chi-yün), we
have a variety of possibilities for the hypothetical Han dynasty pronunciation that is
implied. This is in itself evidence that the word was non-Chinese in origin. Apart from
the variation between voiced and unvoiced initial, we have the form in M. t which, in a
native Chinese word ought to imply an earlier *tl-. It probably does not do so here
however. In the Shih-chi chi-chieeh we find a fan-ch'ieh spelling implying M.
tioŋ', with a
pure dental initial.
This would be contrary to Middle Chinese phonology, which would only permit of ti or ci not ti-. The best explanation of the variant forms seems to be that a pronunciation something like [toŋ] or [toŋ] or [doŋ] was intended. Since it was a foreign word, a stable Chinese pronunciation was not established at the beginning and the Chinese pronunciation was probably corrected at later stages by renewed reference to the original. This would account for the new readings in Middle Chinese -oŋ < earlier *-uŋ after Old Chinese *-oŋ had become Middle Chinese *-uŋ. After the yodization of long vowels it would appear that -iuŋ was a thought by some to be a better equivalent than either of the short vowelled forms, hence the readings M. tiouŋ' or M. tiouŋ' (After all Chinese, with all the multitude of their languages, are unable to come up with their own word for moving a spout of a bag or a nipple up and down. Now, poor poor Chinese. Like English, they should fossilize a word in the 2nd BC, and use that form until the American Revolution, to escape suspicions of plagiarism).
Neither Turkish, nor Mongol, nor, as far as I can discover, Tungusic provides a possible original for this word. The existing Kettish word for "milk" is mâmel (= mâm "breast, nipple" + ul “water”). In the extinct dialects however we find Pumpokol den, Arin teŋul. Kottish used the Turkish word süt for "milk" but retained the word ten "nipple". It ii evident that Arin teŋul is a compound formed in the same way as Kettish mâmel from *ten + Arin kul "water" (see Bouda 1957) (In Türkic "kul" is a body of water, river, etc. Also a borrowing from Arin? Actually, the Türkic “kul” started as a watery gorge, ravine, and developed into “lake later”). It is also clear that we have the same correlation between words meaning "milk" and "breast" that we find in Chinese.
Chinese -ŋ could very well represent foreign -n, especially after a back vowel. (The fact
that so many of the readings are in the oblique tone, implying *-ŋh
or *ŋ', may be
connected with the fact that the foreign original had something different from Chinese
-ŋ.) The only substantial discrepancy between the Chinese and the Yenissei forms is
therefore in the vowel. As it happens there is a sixth extinct dialect, that of the
or "Sable" Ostyaks, who were located in the early eighteenth century on the Elogui River,
somewhat east of the Kotts. Only a handful of words are recorded. Among them however are a
number in which ö corresponds to e in other dialects, e.g.:
öedh "sable", Ket. ead',
eati, eadi; ös "god, heaven", Ket.
es, ech, Arin es; shösh "stream", Ket. ses, shesh, Kot.
ök "thunderstorm". Ket. ekŋ "thunder";
ölugh "river name" = elogui. The word for "milk" is
not recorded but these correspondences would support a reconstruction of *ton. We have no
direct evidence of course as to whether *e or *δ was original but an unrounding and
fronting of *δ to e would be a very common type of phonetic change. A hypothetical
would provide as close a correspondence as possible to the Chinese forms
(and guarantee that Chinese would not distort it beyond
recognition and will thoroughly document it). The variation
between; initial *t and *d is found again in t'i-hu discussed below. In the recorded
Yenissei languages we find two dental stops, a weak (lenis) - Ket d, Kot. t - and a strong
(fortis) - Ket. t, Kot. t'. It would seem that the weak; member of the pair, whatever its
exact phonetic characteristics may have been in the Hunnic
(E.Pulleyblank: Hsiung-nu) language, sounded sometimes
more like Chinese t and sometimes more like d
(All we have
to do now is to connect the Chinese trivial verb with a noun of a remote minor language of
stationary taiga foot hunters - and then - voila, Ket-speaking Huns).
"Sour milk, curds, kumiss"
While the word tung "milk" has not survived in current usage in Modern Chinese, the other word associated with it in Chung-hsing Yüeh's advice to the Shan-yü, lao, has done so. The dictionary definitions suggest a rather vague application to various milk products such as "cream", "cheese" "kumiss" and extended to milky substances such as [ ] "almond tea". Its proper meaning is however "soured or fermented milk", i.e. the curdled milk of cows or ewes or the fermented but curdless milk of mares, namely kumiss. (Dried curd is also called kan-lao, see below.)
Li Shih-chen in the Pen-ts'ao kang-mu says "Lao can be made from the milk of water buffaloes, ch'in cattle (i.e. ordinary domestic cattle), yaks, ewes, mares and camels. For use in medicine one considers cow's lao to be best. It is only because cow's milk is also most plentiful." He goes on to give a recipe for its preparation quoting from the Yin-shan cheng-yao. First the milk is boiled, then cooled and the cream skimmed off. After that a small quantity of old lao is added and the milk is sealed over with paper and allowed to ferment. This evidently refers to the making of yoghurt.
The ordinary word for a drink made from soured cow's milk in Turkish is airan, a word found in all dialects. In Mongol the word is airaγ. No doubt the two words are ultimately connected, in what way must be left to Altaists to decide. The Classical Mongol form of airaγ is ayiraγ from which we could reconstruct a proto-Mongol *ayiraγ. This is phonetically quite close to Chinese lao < M. lak < *hlak. It should be noted further that airaγ means not only soured, curdy, cow's milk but can also be used, like lao, for soured mare's milk or kumiss. (As far as one can judge from the dictionaries this is not true of Turkish airan.) (Funny, E.Pulleyblank is so bravely dismissing modern Kipchak, Tele, Kazakh, Uigur, etc. airans as not true to the dictionary standard. With the modern mass-production and mass-marketing that kind of anarchy will be wiped out. China will buy mare milk in Argentina, and flood the world with real dictionary-compliant Turkish airan).
Since we have ruled out the possibility that Hunnic
(E.Pulleyblank: Hsiung-nu) was a form of Mongol, we must
suppose that airaγ was borrowed from the same source as Chinese
lao and was not itself the
word on which the Chinese loan was based. For the hypothetical Hunnic
(E.Pulleyblank: Hsiung-nu) form we should
prefer to reconstruct a monosyllable, something like *yrak or *grak. Mongol would have
been unable to handle the initial cluster and would have had to change it into something
like *aytray. In looking for a modern form in Yenissei we have to reckon with the fact
that initial *r has disappeared and initial clusters have been simplified. It is possible
therefore to compare with the supposed *Orak, *ytak the Kottish uk "Milchsuppe", Kettish
tık, uok (Imbazk) "Suppe, fliissiger Brei". Donner defines this word as "Mehl-suppe" but
the surviving Kets are a hunting and fishing people, not herdsmen, so that a word properly
applicable to milk products would be likely to be transferred in meaning, if it survived,
to some similar product used as a substitute (see also the next section on "butter").
Phonetically the equation must be considered possible. The rounding of the vowel from *a to uo
(for which Joki writes o) or u could be regarded as the effect of the lost voiced back velar or uvular
which we have to suppose in the Hunnish (E.Pulleyblank: Hsiung-nu) (What a logic: first, we determined that there are no Türkic words in Hunnic; then we
verified that lao = airaγ
which is Türkic airan and which describes a milk product; then we said that it can't be,
and ventured to find a soup that would sound like airan, but made by hunting and fishing
people, not herdsmen, i.e those would apparently milk sables and minks. The last stroke is
to declare a total victory. Another one of those 278 is thoroughly dethroned.).
su [ ] "butter"
(Fast forward: the following illustration first declares cheese to be butter; next we survey neighbors for butter; incorrectly announce that none have similar; then we turn to Kets, and their word is obviously no good; lastly, we find a Ket with resembling phonetics, for rendered butter-like fat, and claim it a a suitable primogenitor for the Chinese term. Examples from modern Turkish lexicon: cream öz; creamer sütlük; creamery süthane.)
The word su is defined by Mathew's Chinese English Dictionary as "cheese, flaky, crisp, short". It is however clear that it originally and properly meant "butter". The nomadic method of preparing it from milk did not consist of churning cream separated from whole milk. Instead the whole milk was first soured and then shaken or beaten in a leather bag as described by Ju-chun (see above) and, in later times, by Rubruck and Pallas. In the Pen-ts'ao kang-mu, the sixteenth century pharmacopoeia, Li Shih-chen says, "Su is what forms floating on the surface of lao. Men today very often confuse it with white sheep fat but the two must be distinguished." Sometimes instead of su alone we find su yu "su oil". The butter lamps used in Buddhist temples are called su teng. (Seems clear: "su yu" is cream/butter (oil, fat) = Turkish "tereyağı"; butterfat "süt kaymağı", the source is "lao" = scalded milk or buttermilk, and "su" is cream from milk or buttermilk)
The character [ ] appears in the Shuo-wen where it is defined simply as su lao [ ]. We occasionally find this, or lao-su, in texts as an equivalent for lao (or perhaps meaning the whole soured milk still containing the butter fat?).
Instead of the special character [ ] we quite often find the homophonous [ ] in the same
meaning (see for example variants in T-5 [ ] (I. I6IC)). The latter is a common character
in transcriptions of the Han period, and there can be no doubt that we have to do with a
loanword in Chinese. The Middle Chinese pronunciation was sou from an earlier *sah. None
of the words for "butter" in neighboring languages bears a close resemblance to this.
Mongol has tosu(n), Turkish yaγ (actually, "süt yaγ", an
exact prototype for Chinese "su yu"), Tibetan
mar, Tocharian (Kuchean) A. shalyp, B. shalype, shalywe. I do
not find a word for "butter" recorded for Kettish. Kettish kajax, kajag was a loanword
from (Türkic) Sogot (Soyot/Soiot ?)
or Yakut. There is however' a Kettish word so (Donner 1955, p. 83) defined as
"aus Fischdarmen gekochtes Fett". The same word occurs in the expression
(Donner's narrow transcription) "Multbeerenfett (Multbeeren mit Fisch-fett vermischt)"
(It is obvious that Kets shared with the Türks the word şhuδ
= su/sub/suv = water, liquid).
The modern Ketts are a hunting and fishing people, not herdsmen, and presumably make
little use of dairy products but it would appear that so means a rendered, butter-like,
fat (as opposed to ky't (Donner-Joki), kyt, kyt, kyet (Castren) "fat" in
general - Kottish also had special word t'empu for "melted fat", as well as kit "fat"). It
seems very likely that this Kettish word is the descendant of the word which in Hunnic
(E.Pulleyblank: Hsiung-nu) meant "butter" and was borrowed
into Chinese. It is uncertain whether we should suppose
that the Yenissei word was originally more like [sa], or whether the Chinese *sah was already
somewhat rounded towards M. sou when it was used for the Hun
(E.Pulleyblank: Hsiung-nu) word. Compare its use in *sah-gleats (p. 220 above).
4. t'i-hu "clear kumiss", "clarified butter" ("whey")
We have noted that Mongol airaγ means both sour cow's milk and sour mare's milk or kumiss and that the same is true of Chinese lao. There is however a special word for kumiss in Mongol, chige(n). With this I propose to compare Chinese [ ] dei-hou < *deh-gah, or [ ] tei-hou.
The first form occurs in the Shuo-wen where it is defined as [ ] "of lao, the pure or fine [kind]". The phonetic correspondence to the Mongol word is good, Proto-Mongol *tί < *ti or *ti became chi in Common Mongol (Poppe 1960), so we are justified in reconstructing a form *tigä(n) or *tiga(n) as the ancestor of chige(n). On the other hand the Mongol *t- would probably have been strongly aspirated, whereas the Chinese form shows no *th but the same vacillation between *t and *d that we found in the word for "milk". The assumption of a common borrowing from Hunnic (E.Pulleyblank: Hsiung-nu) seems to be the best way to account for this.
There is a further complication with regard to the meaning of the word in Chinese. By far the commonest use of the word t'i-hu is in Buddhist texts where it means clarified butter or an even more refined oil obtained from "butter". The semantic link seems to be in the idea of something that is clear. Mare's milk, which contains little butter fat, forms a clear liquid when it is fermented, unlike cow's or sheep's milk, which becomes curdy and thick. Boiled butter is dear as compared to fresh butter. In the Pen-ts'ao kang-mu a definition for t'i-hu is quoted from an earlier work, which is different from those found elsewhere. "T'i-hu is the liquid ([ ] 'broth') of lao", i.e. whey. This would be quite incompatible with the definition of t'i-hu as either kumiss or clarified butter except for the fact that it is again a translucent liquid which is referred to.
To complete the argument it would be desirable to find a Yenisseian cognate, preferably one having the idea of "clear" or "pure", but I am unable to quote one. (E.Pulleyblank does not mention another incarnation of this diluted airan/whey drink, "tahn", in its many phonetical variations, a clear cognate of the same t'i-hu in Chinese rendering).
5. mi-lo "dried curd"
The usual name in Chinese for the dried curd or hard cheese, that is, the gruit
referred to by Rubruck which the Tartars put away for winter food or used as
provisions on a journey, is kan-lao "dried lao". We find however another word which
appears to mean the same thing in the Ch'ang-yang fu of the first century B.C. poet Yang Hsiung (Han-shu).
Speaking of the victories of the Chinese over the Huns
(E.Pulleyblank: Hsiung-nu) under the
Emperor Wu, he says, "We destroyed their wagons ([ ] biwən-uən) and ruined their
yurts ([ ] khiuŋ-lio), drove away their camels and burned their [ ] M. mek-lwa (or -lie)." The word M. mek-lwa
(or -lie), which is not found elsewhere, is explained by
the commentator Chang Yen as meaning "dried lao". He adds, "They use it for the mother
(=ferment) of lao. By burning it one destroyed their means of livelihood."
It was no doubt correct to say that the nomads used dried lao to start the fermentation but we may surmise that the threat to their livelihood was more direct. The stores of dried curd were the nomads' winter food supply and to destroy them was equivalent to burning the grain supplies of an agricultural people.
A different interpretation of the word is given by another commentator, Chang Chi, who says that it was the name of a mountain. Perhaps there was such a mountain but it is obviously nonsense to take the word in that sense here
There is no word for curd or cheese in neighboring languages which shows any resemblance to M. mek-lwa that I can discover. The word throughout Turkish dialects is qurut (whence Rubruck's gruit) and Mongol appears to use the same word xorot. I can find no word for "cheese" recorded for Yenisseian. Though the search for an etymology reaches a negative conclusion, it is interesting to find that the Hunnnu (E.Pulleyblank: Hsiung-nu) resembled the later nomads in their basic economy (Quite a discovery).
Some Hun (E.Pulleyblank: Hsiung-nu) Titles
1. Shan-yü [ ]
It would not be surprising to find that the title of the supreme ruler of the Huns (E.Pulleyblank: Hsiung-nu) reappeared in later nomad empires. Indeed it would be more surprising if it did not. Sir Gerard Clauson has recently suggested that it is to be recognized in yabgu which we find among the Türks (E.Pulleyblank: T'u-chüeh) in the T'ang period. This is impossible to accept on several grounds. Shan-yü < M. jien-hiou < *den-hwah is phonetically quite unsatisfactory as an equivalent for yabgu even if we reconstruct as an early Turkish word with an initial δ, Chinese would have used -m or -p to represent the labial consonant, never -n. Moreover a good Han dynasty transcription of yabgu exists in hsi-hou, found among the Usun (E.Pulleyblank: Wu-sun), Ases/Tochars (E.Pulleyblank: Yüeh-chih) and Kangars (E.Pulleyblank: K'ang-chü), but not the Huns (E.Pulleyblank: Hsiung-nu), and probably of Tocharian (Kuchean)- (8th c. AD Kucha and Karashar oases) origin (see p. 95). Later the title occurs among the descendants of the Ases/Tochars (E.Pulleyblank: Yüeh-chih) in Bactria and it was probably borrowed by the Turks from there. A fuller discussion must be left for another occasion.
Nevertheless shan-yü did not vanish and we can, I think, see in it the ancestral form of
another title that reappears among the Turks and Mongols and was also known farther west,
namely tarqan, tarxan, etc. This is one of the titles which have "Mongol" plurals in
(tarqat) and which, according to Pelliot (1915), must have been borrowed by the Türks
from their Jujan (E.Pulleyblank: Juan-juan) predecessors.
He pointed out also that the spelling with -x- in
Kashgari was a characteristic of words of foreign origin (1944, p. 176, n. 2). Whether or
not Pelliot was right about the immediate source of the word in Turkish, the ultimate
source was no doubt the Hun (E.Pulleyblank: Hsiung-nu).
Phonetically the correspondence is good. The use of
Chinese -n for foreign -r is regular in the Han period. The Chinese initial
*d- would not yet have been palatalized in the second century B.C. when the transcription first
The use of Chinese hw- for a foreign back velar or uvular γ or G has been discussed above. The Hun (E.Pulleyblank: Hsiung-nu) word lacks the final -n which we find in the Turkish but we shall find other examples of this in qaγan, qatin and tegin.
Moreover we find the title tarqan without its final -n on the coins of the Hephthalite ruler of Afghanistan in the seventh century, Nezak Tarxan. In Greek script we find either TAPKA or TAPAKA. It is generally supposed that Nezak got his title from the Turks but it is quite probable that the title was already known among the Hûna in Afghanistan before the arrival of the Turks, just as the title tegin was known among the Hephthalites in Gandhara (see below). (For western references and bibliography see Moravcsik 1958, II, p. 299.)
In Chinese transcription in the T'ang period the Turkish title tarqan appears in a variety of forms: [ ] M. dat-kan, [ ] M. dat-kwan. The last of these means "advanced official" or "official with direct access" in Chinese and it has been suggested that this was the etymology of the Turkish word. Pelliot, while not rejecting the suggestion out of hand, remarked, "Toute la question est de savoir si nous avons affaire a une reelle identite etymologique ou a une transcription d'erudits basee sur une simple analogie phonetique et semantique" (1944, p. 176, n. 2). The latter must certainly be correct. Ta-kuan is not such a common term in Chinese that one would expect it to be borrowed as a title by foreigners.
It is of course not at all surprising that what had been the supreme title under the Hun (E.Pulleyblank: Hsiung-nu) should have declined in status and become merely a high-ranking officer among the Turks. Mongolian doruγa (a form which agrees even better with the Hunnic (E.Pulleyblank: Hsiung-nu) original than does the Turkish and may have been borrowed directly) has declined even further coming to mean no more than the holder of certain privileges. We may compare the fate of khan in the modern Middle East where it has become no more than "mister" (E.Pulleyblank must have had an inside track to Chingiz-Khan or Reza Khan, the future first Shah of the Pahlavi dynasty if they felt like "misters" respectively one and two millennia after the Huns. This is another dear philological contribution to the task of rewriting the history. One of unacknowledged alternate suggestions was that "shan-yü" was simply a Chinese rendition of the word "Khan/Kaan/Kagan" of the time. Shanyu is "respected house" or " esteemed house", equivalent with Huyui "protecting a house", where house is expressed with hieroglyph "tszya" = "family, clan, house" = "extended family" = "country").
2. T'u-ck'i [ ] ti
The title given to the Crown Prince of the Hun
(E.Pulleyblank: Hsiung-nu) was Left T'u-ch'i Prince. Next below
him came the Right T'u-ch'i Prince. According to Han-shu 94A, t'u-ch'i meant
or "worthy" and the titles are sometimes translated and written Left and Right Hsien
Prince. The same word t'u-ch'i occurs as the apρelation of a shan-yü and of a Hun
queen (Han-shu 94). In meaning it may be compared with Turkish bilgä "wise" which also
appears in royal titles. The word itself on the other hand must be the ancestral form of
Turkish tegin, tigin "prince", having been borrowed as a title without its original
semantic content. The phonetic values of the characters are M. dou-gii < *dah-ge(δ) (on
the transcription value of the second character see p. 124 above). The addition of -n and
the plural in -t both speak in favor of a borrowing into Turkish through Mongolian type of language
(Pelliot 1915) (Or simply Türkic possessive affix).
In the T'ang period the correct transcription of the Turkish title tegin was [ ] M. dək-giən, but through graphic corruption the second character normally appears in texts as [ ] M. lək. This same corrupt appears in an earlier form of the title [ ] M. diək-lək (read - giən) which we find in the Pei-shih in a passage on Gandhara based on Sung Yün's account of his visit there early in the sixth century. In the Lo-yang ch'ieh-lan chi we find correctly [ ] M. diək-giən (some texts have instead [ ] over the heart radical but this must again be a case of graphic corruption - no such character is found in dictionaries, see Lo-yang ch'ieh-lan chi chiao-chu, p. 318). According to Sung Yün, the Hephthalites appointed a Ch'ih-chin to rule in Gandhara after they conquered it and as S. Levi pointed out, the existence of the title tegin in Gandhara is firmed by the Rajatarangini, which speaks of a ruler there called thakkana (Chavannes 1903, p. 225, n. 3; cf. also Marquart 1901, pp. 211-12). R. Ghirshman (1948, pp. 109 ff.), finding the appearance of a Turkish title among the (to him) Iranian Hephthalites embarrassing, was at pains to explain away Sung Yün's testimony. Reading Ch'ih-le (Tele) instead of Ch'ih-chin, he proposed to see in it a transcription of Tshavla, Zabul. But M. diək-lək can hardly have anything to do with Zabul which we find in Chinese transcription as [ ] M. dzau. Ghirshman's idea was clearly influenced by the belief that Ch'ih-le, which, it is true, is one of the earlier spellings of T'ieh-le (Tele) (see p. 230 above), could be a transcription of Tölis, Tölis being thought to have some resemblance (surely very remote!) to Tshavla; but T'ieh-le (Tele) and its earlier forms have nothing to do with Tölis, though the view, consecrated by Chavannes, dies hard (see Pulleyblank 1956).
The only reasonable interpretation of Sung Yün's testimony, corroborated by the Rajatarangini, is that the title tegin was used by the Hephthalites long before the appearance of the Turks.
This is not the place for a full scale discussion of the affinities of the Hephthalites
but a few remarks can be made. I am not at all convinced by the arguments which have been
made in recent times to show that they were Iranian (Ghirshman, op. cit.; Enoki 1951, 1952
and 1959). That there should be Iranian elements in their empire is only to be expected
since the subject population must have been predominantly Iranian. Much more' significant
are the evidences of Altaic connections in the ruling Hephthalites themselves. Besides
the title tegin, one may point in the first place to their proper ethnic name, Hephthalite
being a dynastic appellation. In the Liang-shu they are referred to as the country of [ ]
M. hwaət. As has long been recognized, this name must be the same as the[ ] M. hwat of
Hsüan-tsang, which, as Yule and Marquart have shown, is to be identified with the
city known to the Arabic geographers as War-waliz, Wal-walig, Walig, al-Waliga, i.e. the later Kunduz.
In the Hsin T'ang-shu this city is referred to as [ ] M. -a-hwan' City (the Chiu T'ang-shu has [ ] M. at-hwan'). These forms imply an ethnic War or Awar which can scarcely be separated from the Ούάρ κοί Xouννi of Theophylactus Simocatta, the Ούαρ-χώνΐται of Menander Protector and the Avars of Europe. Still earlier the same name occurs as [ ] M. Ou-hwan < *ah-hwan, one of the two divisions of the Eastern Hu in the Han period (the other being the Syanbi (E.Pulleyblank: Hsien-pi) ). The phonetic identity is perfect and there are very good supporting arguments in favor of a connection between the peoples.
The second part of the name War-waliz was supposed by Marquart to be composed by the reduplication of War with addition of the Iranian suffix -ich. I wish to suggest that it was rather the Altaic word for "city", Turkish baliq, Mongol balgasun < *balaka-sun (Poppe 1960, p. 122). The majority of the Arabic spellings quoted by Marquart have -iğ which would be a normal Arabic way of writing a foreign -g. (The spelling in -ix is more difficult to explain in this way.) On this interpretation War-waliz is "City of the Awar" like Chinese A-huan ch'eng. In some of the Arabic forms it would appear that the ethnic is omitted and we have simply Walig or al-Waliga "the city". It is surely not a coincidence that the later Iranian name Kunduz means in fact "Citadel".
The word baliq, *bálaka-sun is not specifically Turkish or Mongolian and the Arabic spellings are hardly sufficiently explicit to indicate definitely one or the other, but there are other reasons for favoring an identification of the Hephthalite-Awar as probably Mongolian rather than Turkish. The *ah-hwan were a division of the Eastern Hu, closely akin to the Syanbi (E.Pulleyblank: Hsien-pi) . According to the very detailed account of the two peoples in the Wei-shu (quoted in the commentary of the San-kuo-ckih toei-chih 30.1003.4; it is also the basis for the account in Hou Han-shu 120) they spoke the same language. Now Pelliot has shown very convincingly that parts of the Syanbi (E.Pulleyblank: Syanbi (Ch. Dingling) (E.Pulleyblank: Hsien-pei)) spoke a Mongolian language and the Wu-huan - Avars should therefore have done so also. (If a part of Syanbi spoke Mongolian, does that oblige Wu-huan - Avars to speak that part, or could they, together with Hephthalites speak the other part? E.Pulleyblank's line of reasoning is not convincing. Instead, his identification of Avars with Eastern Hu Wu-huans (aka Uhuans 乌桓) brings together eastern and western records, an achievement of a tremendous scale.)
Schlegel 1892 noted long ago that according to the Chinese accounts, the Hephthalite
married women wore the characteristic conical headdress of the Mongols which was adopted
as a lady's fashion both in mediaeval Europe, where it was known as the hennin, and in
China, where it was known as the [ ] ku-ku (with other spellings), from Mongol kökül.
Still earlier, what must be this same headdress is described in the account of the Wu-huan
in the Wei-shu, where it is given this very name [ ] M. ku-kwet (the first character has
a number of other readings: ku\, kious\, giou\; the Old Chinese vowel was *o/ō). The
identification of this word with kökül has already been proposed by Egami 1951. Earlier
Shiratori discussed the description of the woman's headdress but did not recognize the
Mongolian name (As much Mongolian as Türkic; Russian borrowing is
"kokoshnik", where -shnik is an affix, the headdress is attested from pre-Mongol time).
We have in this evidence an important cultural and linguistic link between the Wu-huan of Manchuria and the Hephthalite-Awar of Afghanistan, and between both and the Mongols. Historically there is nothing in the least difficult about such an hypothesis. The westward movement of the Eastern Hu after the collapse of the Hun (E.Pulleyblank: Hsiung-nu) empire in the middle of the second century A.D. is well attested, and various Syanbi (E.Pulleyblank: Hsien-pi) groups appear in Kansu (Gansu; it is an accepted spelling. However, E.Pulleyblank may be up to something: Kansu would tend to be etymologized in Türkic a la Tengri-Tag and Ala-Tau as apparent Khan-Su = Khan's River) from the third century onward. One of these, the T'u-yü-hun; established a state in the Ch'ing-hai-Tsaidam region and extended its influence out into Sinkiang, where they were next-door neighbors of the Hephthalite empire. We are told in the Liang-shu that the people of Hua (i.e., the Hephthalites) were illiterate and that their language could only be understood when interpreted by the men of Ho-nan (i.e. the T'u-yü-hun). This statement can be most easily understood if we suppose that the Hephthalites spoke a language which was the same as that of the T'u-yü-hun, or closely akin to it, therefore a Mongolian dialect (see Pelliot 1921). They no doubt represent a westward extension of the same great movement of the Eastern Hu which brought the T'u-yü-hun to Ch'ing-hai.
The Chionites, Xoννi, Hûna with whom the Hephthalite-Awar are closely associated in Afghanistan are, in my opinion, likely to have been of Hun (E.Pulleyblank: Hsiung-nu) origin. They may either have come independently of the Hephthalite-Awar or in association with them. After the Syanbi (E.Pulleyblank: Hsien-pi) overthrow of the Hun (E.Pulleyblank: Hsiung-nu) empire we are told that many Hun (E.Pulleyblank: Hsiung-nu) became incorporated into the Syanbi (E.Pulleyblank: Hsien-pi) and there must have been a good deal of mixing up of Hun (E.Pulleyblank: Hsiung-nu) and Eastern Hu peoples in the following centuries without an immediate complete loss of identity. Such questions as the immediate provenance of the European Huns and Avars, the relation between the "Pseudo-Avars" and the "True Avars", the connections with. the Jujan (E.Pulleyblank: Juan-juan), etc., must be left aside for the present.
The Hephthalite-Awar were not the only Altaic people to use the title tegin before the
Turks. It was also known among the (Türkic) T'o-pa
where we find the form [ ] M. diək-giən (Sung
shu 95, Nan Chi shu 57). The T'o-pa are usually regarded, since the studies of Boodberg
and Bazin, as Turkish but, while there were certainly Turkish elements among
them - possibly. mainly of Tele (Ch. Dingling) (E.Pulleyblank: Ting-ling)
origin - they were regarded as Syanbi (E.Pulleyblank: Hsien-pei) by.the Chinese
and many of the general features of their language that have been noted could be as
readily interpreted as Mongolian as proto-Turkish (Gabain 1950, 55 pp.). The
identification of the leading group as Turkish leans, it seems to me, rather heavily on
titles like tegin which are not really Turkish at all. The whole question needs
re-examination. (Identification of Dinlins, Modern Chinese Pinyin 丁零 Dinglings,
with Gao-gui/Kao-ku, Modern Chinese Pinyin 高車 Gaoche, and then with Tele, Modern Chinese Pinyin
鐵勒 Tiele, and of Tele with the Türkic tribes and states to modernity is well
established, there is no doubts about their Türkic-linguality).
The title qaγan, xaγan was transcribed in the T'ang period, when it was the supreme title of the Turks, as [ ] M. kha'-han. Earlier it had been used among the T'u-yü-hun and the Jujan (E.Pulleyblank: Juan-juan). Its earliest occurrence is in the story of how the two brothers T'u-yü-hun and Jo-lo-hui agreed to separate. As the Sung-shu 96.1653.1 puts it, "Lou [the envoy of the younger brother to T'u-yü-hun] was glad. He bowed and said, "Ch'u k'o-han [ ]". The barbarian words ch'u k'o-han mean in the language of Sung, 'Be it so, sire ([ ]))' ". The word I have translated as "sire" is an expression used to refer to the Chinese emperor. It would appear however that among the Turks the title qaγan was originally not confined to the supreme ruler but was used, rather like English "lord", as a general title of respect.
"The word ch'u < M. chio' was first interpreted by Pelliot 1921, p. 329, as Mongol chi "you (sing.)", taking Chinese erh as the second person pronoun. This makes the sentence read somewhat awkwardly and, since erh can also be used as a synonym of [ ] "so, like this", he later changed his opinion and interpreted it as Mongol jōb or ja (T'oungpao 29, 1932, p. 261). The same word occurs as the title of one of the rulers of Jujan (E.Pulleyblank: Juan-juan), called Ch'u k'o-han and the Chinese historian explains ch'u as meaning wei [ ] "yes" , strange as this may seem as a royal name.
In a passage which appears in the T'ung-tien and not elsewhere the social organization of the Türks (E.Pulleyblank: T'u-chüeh) is described. The original date of composition is not known but one may surmise that it is comparatively early, perhaps from the sixth century when the Turks first became known to the Chinese. After listing various offices which existed among them, the text reads, (in Liu Mau-ts'ai's German translation) "Manchmal wurde der Posten Fu-lin (alttürkisch: böri(n)) Khagan [ ] errichtet; Fu-lin bedeutet Wolf. Der Titel sollte auf die Mordsucht ausspielen. Es gab auch Khagane die im Range niedriger standen als der Ye-hu (Yabgu). Es kam auch vor, dass grosse zuhausbleibende, also nicht amticrende Familien sich gegenseitig I Khagan [ ] nannten. Die T'u-küe sagten für den Raum (oder das Haus [ ]) I [M. ywi\] (alttürkisch- äb~äv). Der Titel bedeutete also Raum- (oder Haus-) Khagan." (Liu Mau-ts'ai, 1958, PP- 4o8-99.)
The Hun (E.Pulleyblank: Hsiung-nu) word Hu-yü < M. hou\-hiou < *hwax-&wah which I wish to see as the origin of qaγan also seems to have had the same kind of meaning. This title occurs only in Han-shu 94B.0602.1. Because several Crown Princes had died during the time of Wu-chu-liu Shan yü (8 B.C.-A.D. 13) it was thought that the title Left Wise Prince was unlucky and the title hu-yü was given to a certain prince instead. The Han-shu says, "The dignity of hu-yü was the most honorable. He was to become shan-yü in succession." Thereafter we hear no more of this title.
I have given phonetic arguments above for the belief that hu-yü, representing a foreign
original something like *γwaγwa or *GaGa, could be the original behind Turkish
(which in the vacillation between q and x, it will be noted, shows the same evidence as
tarqan/tarxan of being a loanword in Turkish). It is tempting to try to compare also Kett.
ky, pl. kykŋ "prince", Kott. hiji, hije, pl. hijan, hikŋ)
but in the absence of more knowledge of Yenisseian phonology this could be no more than
conjecture. (So, E.Pulleyblank takes "Ye-hu" = "Yabgu", for
whatever reasons transmitted by Chinese in a form apparently phonetically way out of the
canonical phonetical conversion laws, and transforms it following the correct laws, gets a
true Chinese phonetics *γwaγwa or *GaGa, and comes to a conclusion that in reality
phonetically "yabgu" is "qaγan". On top of that, because Chinese vacillated in
rendering the Türkic title, E.Pulleyblank detects an evidence that it is a loanword in
Türkic. Can you hold my hat, please, while I try to wipe my tears?)
4. O-chih, qatun
As stated above (p. 89) I propose to identify the title of the consort o£ the Hun (E.Pulleyblank: Hsiung-nu) rulers, [ ] M. •at-cie < *•at-teh, with the corresponding title among the Turks, qatun/xatun. There are a number of complications In the first place there are uncertainties about both characters in this. expression. The fact that the second character sometimes appears as [ ] M. *tei < *teδ is not particularly serious, since it would have little effect on the transcription value in the Former Han period. The variant readings of the first character are a more difficult problem. Besides the normal M. •at in the sense of "to obstruct" ([ ]) various special readings appear in dictionaries. The reading M. •io < *•ah, which occurs only in the reduplicative binome [ ]. •io-yo can be ignored, as also can the reading M. •iat < *•at, found in the astrological expression [ ] M. tan-•iot. Besides these we are also given Μ. •ien, •en, and according to the commentators Chang Shou-chih and Yin Shih and to the Kuang-yün these readings are proper to the Hun (E.Pulleyblank: Hsiung)-title. On the other hand Ssu-ma Cheng in the Shih-chi so-yin says, "The old pronunciation was M. hat-tei [ ]" (This old "Hatei" is a closest and direct reflection of the later Türkic "Hatun/Hatin").
The readings given by the Kuang-yün seems to be based on a supposed connection between
the Hun (E.Pulleyblank: Hsiung-nu) title and the word
[ ] M. -•en-cie "safflower". We find this
stated in a letter by the historian Hsi Cho-ch'ih (4th c. A.D.). Writing about the
safflower which, he says, is used by the northern barbarians as a cosmetic, he adds,
"The Hsiung-nu word for 'wife' is O-chih. It means that they are lovely as yen-chih 'safflower'
" (T'ai-p'ing yü-lan 719.3a). This is of course no more than a popular etymology. The
word Yen-chih "safflower" is also the name of a mountain (written [ ] M. *ien-cie or [ ]
M. •ien-gii) near Shan-tan in Kansu (Gansu) from which safflower was obtained. (Shih-chi
110.0246.2, Han-shu 94A.9597.2). The Hsi-ho chiu-shih also cited in the T'ai-p'ing
yü-lan 719.3a, quotes an old song purporting to express the lament of the Hun
(E.Pulleyblank: Hsiung-nu) at
being driven out of Kansu (Gansu). "If we lose our Ch'i-lien mountains, it will make our livestock
not breed; if we lose our Yen-chih mountain, it will make our wives (fu-nü) lose their
beauty." If this song was known to Hsi Cho-ch'ih it may have suggested to him the
connection between the words O-chih and Yen-chih. In his day M. •at-cie and M. •en-cie or
•ien-cie would have sounded much alike, especially if final -t in North China had already
begun to weaken to a spirant -δ. In the former Han period however they would have been
quite different, not only in the first syllable but also in the second, since [ ] goes
back to *keh but [ ] goes back to *teh. It is to be suspected that the special readings of
[ ] in
O-chih arose only out of the later pun and have no real authority behind them.
One of the spellings of the Yen-chih mountain is the same as for *Argi, the old name of the Tocharian kingdom at Karashahr. In spite of the fact that the safflower is red, not white, and is used to produce a red dye, it seems to me likely that *•ean-keh or **•an-keh "safflower" comes from a As/Tochar (E.Pulleyblank: Yüeh-chih) word related to Kuchean (E.Pulleyblank: Tocharian) A. arki, B. arkwi "white". The Indo-European root primarily means "shining, bright", cf. Greek < αργος "bright", αρργυρος "silver".
A variant spelling for O-chih is found in the Lun-heng: [ ] M. •ien-dei. This is interesting in confirming the dental medial consonant. It would appear to support the readings in -n of the first syllable but the evidence is less strong than it appears since there is very frequent graphic confusion between [ ] and [ ] M. •ou < *•ah. We find, for instance, the place name [ ], known in Later Han as [ ] (see p. 105 above) and also spelt [ ], has a variant [ ] in the Lü-shih ch'un-ch'iu (see Tz'u-t'ung p. 0084; cf. Pelliot 1936, pp. 266 ff.).
Finally, there seems no reason to give Ssu-ma Cheng's testimony less weight than that of the other commentators. His reading M. hat-tei not only confirms the absence of an -n in the first syllable but also has initial *h rather than the glottal stop, seeming to support the opinion that the foreign word had a consonantal opening and not merely a vowel.
Turning to qatun/xatun, which again seems from its form to be a loanword in Turkish, we find that it is commonly supposed to be a derivative of Sogdian γwt'ynh *xwaten "queen". The fact that in a Sogdian text we find γwt'ynh and γ'ttwnh = xatun together, the one referring to the queen, the other to the first ranking concubines (Benveniste 1940, 6 1.165), does not in itself absolutely rule out this etymology, since the word might have been reborrowed from Turkish into Sogdian without being recognized. Nevertheless it can at any rate be said not to support the Sogdian etymology. Moreover the phonetic correspondence is far from exact and the meaning is not quite the same either. Both the Turkish qatun and the Hun (E.Pulleyblank: Hsiung-nu) o-chih would seem to have meant originally simply "wife" rather than "queen" whereas γwt'ynh can only mean "queen". (The word "katun" is also spelled and articulated "khatun", "khotun", "khotan", "khatin", and the like. "Katun" was a title for a spouse of Khan or Kagan, and so was named her enclave or estate. Eastern Europe is full of topology that carries that name, indicating the location of the Quinn estate in the Late Antique times, when the Alans, Huns, Avars and Bulgars controlled these territories - see Wikipedia. Unbeknown to E.Pulleyblank, the status of the Katun was of the same level as of the supreme Kagan. Moreover, in the gynocratic Türkic societies the title to the country's land belonged to the Katun clan; the Kagan ruler was an elective position akin to CEO in modern world; like a CEO, theoretically he could be dismissed from the position, and in usual practice he was simultaneously dispatched to the other world. A Kagan could have a number of wives and concubines, but only one Katun. In times of trouble, or when princes were underage, the Katun clan could take over the rule of the country, producing such famous names as Massagetan Tomaris, Alanian Boarix, and Kharka (Kreka in Priscus), a wife of Attila. Ref Yu.Zuev 2000). This Türkic word originated with Huns, because they brought it to Eastern and Central Europe, and if Sogdians learned it, it was from the Huns or Türks, not the other way around. Usuns and Kangars had a similar social organization as the Huns, and they could have passed the term to Sogdians. At any rate, the Sogdian letters are 3/4 of a millennia later than Chinese records, that gives a priority to the Huns.)
Still further doubt is thrown on the Sogdian connection by the fact that what must clearly be related forms of the same word are found already among the T'u-yü-hun and the T'o-pa. By the former the consort of the ruler was called [ ] M. khak-tsuən (Chou-shu 42,23403), by the latter [ ]. kha'-suən (Nan Ch'i shu 57,1755-4). Though Sogdians were present on the western borders of China at this period, it was as traders rather than in any military or political sense and it seems doubtful whether the nomadic rulers would have turned to them for titles.
Boodberg thought that M. kha'-suən represented an original like *qasun. Bazin suggested
instead an interdental fricative *qaθun. A more likely interpretation for both T'o-pa and
T'u-yü-hun forms would be *qachun.
We have noted the difficulty which the Chinese had in the Forrmer Han period in representing foreign palatals. By the fourth and fifth centuries Chinese had of course palatal affricates but there were no syllables in -uən or -iwən with palatal initials and even much later we still find recourse to dental affricates in such Buddhist transcriptions as [ ] M. tsuən-na=Cunda (T.363).
If we could suppose that *qachun developed out of an earlier *qachi-. with the addition of a suffix, we could account for the medial -ch- by the normal palatalization of *ti in Mongolian. This would lead us back to an earlier *qati- which is very close to the presumed Hunnic (E.Pulleyblank: Hsiung-nu) original behind *•at-teh. Mongolists must decide whether *qachun < *qachi- is possible. The same problem of the vowel in the second syllable is found in Turkish qatun/xatun, borrowed either independently from Hun (E.Pulleyblank: Hsiung-nu), or from a proto-Mongol form earlier than *qachun, since it preserves -t-. Here it may be noted that an alternation between i and u is frequent in Old Turkish as for instance in qatun/qatin "become hard' (von Gabain 1950, pp. 49, 327). Many of the later dialect forms of the word do in fact show -in rather than -un and there are also related forms without final -n like qât "junges Weib, Frau" (Abakan dialect), qaday "Gattin, Weib, alte Frau" (Tuv. slov.), gadi "alte Frau, Omama", gade "Schwägerin, Schwester" (Anatolia, Söz Derl.) (Çağatay 1961, p. 17).
The "Hun (E.Pulleyblank: Hsiung-nu) Couplet"
Some mention should be made of the famous couplet in the "Chieh" (Kangar) language found in Chin-shu 95.1331c (cf. Wright 1948, p. 344). If our theory about the Chieh (that Chieh stands for Kangars) is correct (p. 247), we should expect their language to be either the Kangars (E.Pulleyblank: K'ang-chü) variety of Kuchean (E.Pulleyblank: Tocharian) or proper Hun (E.Pulleyblank: Hsiung-nu), since they might have lost their original tongue while living in the east (Kangars are known to be west of Hunnic state and north of Usuns in Jeti-su, north and east of lake Balkhash along Syrdarya/Yaxartes/Seyhun. E.Pulleyblank must be referring to a branch of Kangars living separately from their state, in the Hun state in the east). We should not expect any form of Turkish or Mongolian (That is a strange statement, Kangars were Hun's allies as far as China was concerned, and maintained friendly relations with Usun and Hun states both on political, and personal level. Later Kangars were in Tele and Kipchak confederations, and a branch of Kangars under a name of Badjanaks/Becenyo/Patsinaks was undoubtedly Türkic-speaking). On the other hand on the supposition that the Hun (E.Pulleyblank: Hsiung-nu) spoke Turkish a number of attempts have been made to interpret the couplet in terms of Turkish (in recent times we may note the attempts of Ramstedt 1922, Bazin 1948, and Gabain 1949). None of these interpretations can be considered very successful since all do more or less violence to the phonetic values of the Chinese characters and to the explanation given in the accompanying Chinese text.
The couplet as explained in Chinese consists of four words:
Beyond remarking that -ŋ is a common
verbal ending in Yenisseian, especially Kottish, I shall not, at least for the present,
attempt to add to the list of suggested reconstructions.
My study of some of the Hun (E.Pulleyblank: Hsiung-nu) words appearing in Chinese transcription leads to the following conclusions:
(i) the evidence for the existence of initial r and l and initial clusters in Hun (E.Pulleyblank: Hsiung-nu) makes it most unlikely that it was an Altaic language (These initial liquids and initial clusters need to be rechecked, and questionable cases removed);
(2) a number of words for which the meaning is given or can be inferred correspond quite closely to words of the same or similar meaning in the Yenissei languages - among them the words for "son", "milk", "stone" may be especially noted as being unlikely to be loanwords in Yenisseian (These words need to be tabulated and classified by confidence in recorded information and by confidence in possible alternatives, with valuation of factors);
(3) certain Hun (E.Pulleyblank: Hsiung-nu) titles (and also the words for "heaven", "sour milk", and "kumiss") can be traced later in Mongolian or Turkish or both. The simplest hypothesis to explain these facts is that the Hun (E.Pulleyblank: Hsiung-nu) spoke a language of the Yenissei family and that the Mongolians and Turks who followed them as masters of the eastern steppes inherited elements of culture and political organization, with the corresponding names. (This scheme does not fit in the known history, conflicts with archeological record, and is ethnologically impossible)
This hypothesis, based on linguistic evidence, must be tested by reference to other types of evidence, particularly archaeological (cf. Jettmar I952).
θδğŋγşāáäēə ï öō üūû Türkic
Contents Türkic languages
Ogur and Oguz
Alans and Ases
E.Pulleyblank Eastern Hun Language
O.Pritsak Onomasticon of Western Huns W.B.Henning Xiongnu are Huns
L.Gumilev Language of Huns
Kisamov N. The Hunnic Oracle
Tekin T. Hsiung-Nu Language
Vovin A Hsiung-Nu Language
Taskin V. Hsiung-Nu Language
Doerfer G. On Hunnic and Turkic (snippets)
Gmyrya L. Caspian Huns