Contents Türkic languages
Classification of Türkic languages
Indo-European, Arians, Dravidian, and Rigveda
Scythian Ethnic Affiliation
Foundation of the Scythian-Iranian theory
Türkic borrowings in English
Türkic in Romance
Alans in Pyrenees
Türkic in Greek
Chronology of Türkic languages
and linguistic contacts of early Türks
This posting presents a groundbreaking work of A.V. Dybo in Türkic linguistic, which remains as much behind times as only falls to the unloved stepchildren. In limbo for 70 years after a nearly complete wipeout of the Turkology and Turkologists from the Eurasian scientific scene, the Turkology is resurrecting in feeble and frequently uncertain steps that on the background of mediocrity are gigantic. The work of A.V. Dybo is a seminal step in the ongoing process. It breaks the barriers of preconceived notions and institutional limitations, brings expertise to a field notable for the lack thereof, and in process decimates primitive constructs propagated by inept scholars. It peeks beyond the pale of settlement instituted in the 20th-century scientific standards.
The work of A.V. Dybo is notable for many insights that make a travesty of the established postulates, the most important of them is her rejection of the linguistic tree model, the very backbone of the whole contemporary linguistics, which, applied imprudently and indiscreetly, brought about innumerable distortions and fictitious paradigms into histories of many peoples. Another is the expansion of the search field for Bulgarian lexicon beyond the Chuvash language; A.V. Dybo turns to the Hungarian Bulgarisms, and that opens a gate for inclusion in the field of the Russian and Danube Bulgarian, Ukrainian, Balkarian, Galich, Itil Tatar and Itil Finnic, in addition to Hungarian and Chuvash, which undoubtedly would uncover new horizons and lexical interactions. Another insights is the treatment of the Sogdian language, with a critical eye toward the previous etymologies, and with understanding of the historical processes and distinct ethnological settings, which brings layers of previously ignored connections between Sogdian and Indian languages, elucidating or replacing the notoriously miopic and frequently shallow Iranian etymologies. Most bold is the A.V. Dybo stipulation that (literary) "Sogdian does not fit into Linguistic Tree model, because no pra-Sogdian can be conjectured". This is equivalent to a discovery of a new species, defined as "a taxonomic group that can't breed with other taxonomic groups". A.V. Dybo stipulates a product of different linguistic groups, an indirect allusion to a Creole composed of a symbiosis of the pre-Indo-Iranian flexive and possibly Türkic agglutinative languages "... it does not seem possible to regard the Eastern Iranian group as a whole, even excluding Parachi and Ormuri, as a genetic grouping. Such a concept would imply the existence of an ancestral "Proto-Eastern Iranian" intermediate between the "Common Iranian" and the attested Eastern Iranian languages; but if one reconstructs the "Proto-Eastern Iranian" in such a way as to account for all the features of the group, it proves to be identical to the "Common Iranian" reconstructible as the ancestor of the whole Iranian family. It is therefore more plausible to conceive of Eastern Iranian as a "Sprachbund" or areal grouping of languages. In this case the members of the "Sprachbund" happen to be genetically related, but the special features which mark them out as a group result rather from centuries of contiguity, during which innovations will have spread from one language of the group to another and neighboring languages will have supported each other in the retention of shared features." In other words, the very concept of the Eastern Iranian group is as much a fiction as any Creole conjured to be a daughter of a mother tongue. And still another insight is the finding that "The most pervasive external influence on the "Eastern Iranian" (quotation marks added) has been that exerted by the neighboring Indian languages, as is most evident in the development of aspirates (Khotanese, Parachi, Ormuri) and retroflex consonants (Khotanese, Pashto, etc.). However, similar developments are found in Baluchi (North-Western Iranian), which is also spoken in close proximity to Indian. The Indian loanwords already are found in most "Eastern Middle Iranian" (quotation marks added) languages (but also in Parthian) and they increase in numbers in the modern period." A major advancement is a clear discrimination between the Persian, Indian, and Türkic loanwords, critically important for understanding historical processes in Middle and Central Asia, and for truthful attribution of the deceitfully entitled so-called Tocharian and Khotan Saka linguistic properties to their real sources.
Among the very few shortcomings of the A.V. Dybo analysis is the one precipitated by the faulty institutional historiography, which contrary to the normal historical methodology divides a continuous historical stream into disparate floating unrelated segments. That did not allow A.V. Dybo, and likewise her predecessors, to extend their historical vision beyond the linguistically insignificant events of the political perturbations, into the milestone events of the 1,200 Yin Shang Culture, 1,000 BC Karasuk Culture, 700 BC re-population of the Middle Asia, and 200 BC integration of the Si/Hi/Xi/Kumosi 奚 Mongolian tribes into the Eastern Hunnic state, which divided Mongolian for the next two millennia into a Turko-Mongolian Creole and a Mongolian proper. The marriages of the unrelated linguistic families definitely had a far more profound effect then a transfer of allegiance from one suzerain to another, resulting in distorted glottochronological analysis, already disfigured beyond recognition by the wrongful application of the Linguistic Tree model, as much as would be the case if the Haitian language was used to reconstruct the glottochronological history of the French. The fact that the glottochronological dating happen to conveniently match the little relevant political events should have raised flags on the applicability of the methods and models. Indirectly, the article addresses how the Halaf, Dzharmin, Hasun, North Ubeyd, and Kura-Arax archaeological cultures of the middle of the 4th millennium BC figure in the development of the Sogdian language, 15 hundred years before the arrival of the Indo-Iranians to the South-Central Asia (http://www.fas.harvard.edu/~iranian/Sogdian/).
Posting clarifications and comments are (in blue italics), and in bold highlighting of the author's text. Page numbers are shown at the beginning of the page in blue. The Times New Roman font used does not support all diacritics of the original text, for a special detail please refer to the original http://altaica.narod.ru/LIBRARY/xronol_tu.pdf PDF file. The author used a Russian character ú to denote phonetic stops, thus Herúcul is equivalent to Her'cul. The abbreviation Gen. (General) is used to render the case when the application of the term Comm. (Common) is misleading, like in the case of the Common Türkic which in the context of this work denotes not the common Proto-Türkic ancestor language, but is a collective term for the prevailing modern linguistical distribution, and refers to all branches of the Türkic family except for Ogur (aka Bulgar) branch. The adjective Türkic and the noun Türk are used to denote the global world of the Türkic community that includes Turkish and Turks as one of the constituents; Türk is a noun of which Türkic and Turkic are adjectival derivatives, it is needed for translation from Russian, which has four distinct designations for four phenomena (Türk/Türkic or Turkic and Turk/Turkish). The semantics of that terminology in English and Russian is a result of their national histories.
|Abbreviations of language names (work in progress, a number of abbreviations need to be added)|
Chronology of Türkic languages and linguistic contacts of early Türks
The Türkic language family is a gratifying material for application of different classification methods and linguistic chronology. First, it is large enough ( at least 35 modern languages can be handled as individual idioms and dialects and not less than six well-recorded ancient and medieval languages). Secondly, it is relatively young, and the history of the Türkic peoples is mostly well documented, so the genetic classification of the Türkic languages may be interpreted in terms of real history. The main difficulty in the classification and chronology of the Türkic languages lays in numerous contacts between highly similar and often mutually intelligible languages. In that regard, Turkologists developed several classifications of the Türkic languages. The most popular of them, to date, are classification by N.A.Baskakov (see pp. 767), which is primarily aimed at interpretation of historical sources that not always correlate well with actual linguistic properties, and classification based on proposals of A.N.Samoilovich (see pp. 768), now with some changes, which relies on phonetic and morphological features, although we can not be sure that those defined at a surface level properties are really the most relevant for the genetic classification, but certainly that classification corresponds to the particular linguistic reality, at least reflecting one of geographic assemblages of the Türkic languages during their history. In the previous volume of our publication, "Regional Reconstruction", we essentially used with minor modifications the latter classification.
Samoilovish Classification (omitted)
The following is an attempt to construct an absolute chronology for branching of the Türkic family of languages based on lexical statistics. Such attempts have been made before, but the 100 word list surveys so far were not yet compiled for absolute majority of the languages 224. Our lists assembled all relevant languages for which could be actually compiled from written sources 100 word lists. The 100 word lists were assembled from the Yakhontov - Starostin amended list of meanings, and were processed using the amended Starostin technique, with identifiable borrowings zeroed 225, in the Starling program.
224 The last work that we know of is Diachok M.T. Glottochronology
of the Türkic languages (preliminary
analysis), Science, University, 2001, Materials of the Second Scientific Conference,
Novosibirsk, 2001. pp. 14-16. In his brief communication the author does not provide 100
lists he compiled for the Turkish, Uzbek, Tatar, Chuvash, Salar, Tuvan, Yakut,
Khakas, Kazakh, Kyrgyz, Turkmen, and Azeri, so we can't judge the comparison methods. Note
that the conclusions are similar to our conclusions: early separation of Bulgar, Yakut
and Tuva-Tofalar branches; absence of genetic proximity between the Yakut and
Tuva-Tofalar group; general affiliation of the other Siberian languages to the "Western"
group. The other conclusions of the author's are fairly obscure.
The study examined two sets of 100 word lists. The first set assumed that for each language the lists may contain unlimited number of synonyms (or quasi-synonyms, the words that might be translated by the same English word). This premise firstly allows to avoid a problem of semantic match-mismatch for different interpretations of the Swadesh meanings within the same language, and secondly it avoids a slippery choice of the "most common" word (because the usage can differ much depending on a genre of the texts). It should be noted that such method, despite its obvious advantages, obviously includes in the computations numerous "noises". The resulting tree contains obvious cases of secondary linguistic convergence due to geographical contacts (p. 770).
The second list retained only those "synonyms" where the semantical differences can not be directly expressed with English equivalents. These, for example, are such cases as "thin" (for flat objects) and "thin" (about three-dimensional objects), or several entries for "that" and "this" in a three- or four-dimensional system of spatial deixis. The resulting family tree (p. 771) has more correlations with the classification that was adopted originally. These differences require analysis and explanation.
Another point that should be discussed in connection with the method of linguistic
classification by the "common innovations" is as follows. Defining the problem in the
framework of comparative historical linguistics, a linguist has to follow the "tree"
model depicting changes in the linguistic phenomena: the other models, like the "wave"
model, are excluded by the very
essence of the comparative procedure. However, in the situation for the recent separation
of linguistic idioms that retain mutual understanding between the carriers of
these idioms often brings confrontation with the wave-like phenomena, when an innovation extends
across the borders of the idioms. The obvious interpretation of such phenomena within the historical
and comparative linguistics must be staged as follows. If a certain innovation arose at the time of
a dialect being affiliated with some group (driven by the common political, economical or
geographical boundaries) common for that group, then
the study of reflexes of that innovation holds that original group as a node in the
family tree model; however there is a caveat in case when a part of the dialects of that group
of an archaic phenomenon shared with other dialectal groups. That case
would result in a different, older tree with different nodes that correspond to other, older
groups of the dialects. Thus, the linguo-graphical approach in that form should allow obtaining
a relative chronology of re-grouped dialects (or closely related languages).
The task of a comparativist in that case is to establish a relative chronology of different family trees. Like in other cases, this is achieved by ascertaining the hierarchy of the transition rules for the transition from a reconstructed state to a reflective state. Namely, in cases where a reflexive state can be presented as a result of an F1 rule on the result of a rule F2, the order of the rules is defined as F2, F1. Then, if those rules form nodes of a family tree, the family tree resulting from F2more archaic and genetically predating, and the family tree resulting form F1 should be considered to be secondary. Undoubtedly, in that case both deduced trees reflect a linguistic reality. We should also emphasize that all nodes of a family tree do not mark the origin time or the duration for the conjured linguistical unit, but the timing of the split of two descendant units. In the following, we will call the rules which may lead to the formation of nodes on the family tree "isoglosses", those isogloss that are relevant in the formation of the tree we will call "connected".
The first node of both family trees is branching of the Chuvash from other languages, usually defined as branching of the Bulgar group. In terms of the linguistic change process, in the phonetic this division is expressed with the following most prominent isoglosses: 1) vocalization of the Pra-Türk. *t- > Comm. Türk. *d- before voiced stops, *r and *r' (traces of such voicing found in Oguz and Sayan languages, but are absent in the Danubian-Bulgar borrowings in Hungarian), 2) development of Pra-Türk. *l', *lč, *r' into Bulg. *l, *č, *r in accordance with Comm. Türk. *š, *š, *z; 3) development of Bulg. *-δ- > j in the roots with r, and *- δ- > -r- in other contexts (chronologically, this rule, apparently, can be considered to be a last of the three, since it could be effective after congelation in Bulg. of *r è *r', compare Chuv. xujúr 'bark' < Pra-Türk. *Kar-δyr). 767
In the morphology is as following (not coordinated relative to each other) processes: a) Common Türkic and Bulgar used two different ways of grammaticalization of the plural indices: Bulgar used old postpositional pronoun *sayin 'all', the Common Türkic used one of the old collective name suffixes, *-lar (note that in both branches can be found traces of the old indicator of (paired?) plurality *- r'); b) in the paradigm of the nominal declension the Common Türkic replaced the old genitive suffix -(i)ŋ (with distribution after vowels and consonants of the end of the base) to affix -(n)iŋ, extracted from reinterpreted pronominal forms; the Bulgar, on the contrary, initially preserved -(i)ŋ, and later developed -(n)iŋ for possessive-nominal and noun paradigms, a base of which contains the same vowels as Personal pronouns; c) the Common Türkic lost participle of expected future with -s, retaining it only for the formation of negative forms of an uncertain future; the Bulgar, by contrast, retained this participle. On both trees that node dates around -30 - 0 BC. We would tie that date to the migration of the Huns from the western Mongolia to the west, through northern Eastern Turkestan (orig.: Xinjiang) to the southern Kazakhstan, to the Syr-Darya River in 56 BC. (The "Bulgar" was a political umbrella term for a number of predominantly Türkic tribes. Under "Bulgar language" are meant the indiscriminate linguistic traces attributed and studied so far. Most of the "Bulgar language" embedded in the Russian, modern Bulgarian, Ukrainian, Galician, Itil Tatar and Itil Finnic has not been studied yet. Bulgars united diverse tribes of Onogurs, Caucasian Huns, Suvars, Izgils, Oukals, Khazars, Akathyrs, and many others, mostly with obscure provenance. Suvars were recorded in the Sumarian cuneiforms, Izgils were relatively extensively recorded in the Chinese annals as a strongest tribe of the Hun confederation. The geographical disPersal of the Bulgarian constituent tribes points to an ancient and complex linguistical origin, the surviving traces of which mostly belong to the Middle Age period)
In connection with the dating of the Pra-Türkic language fragmentation arises a problem of dating the foreign loans into the Pra-Türkic (that is, those regular reflexes which we find both in the Bulgar and Türkic group). As evidence of Pra-Türkic contacts with neighboring languages can be cited a number of Chinese loanwords into the Pra-Türkic before the split of the Bulgar, and perhaps a few words that can be interpreted as borrowings from some Eastern-Iranian language, in phonetic developments reminiscent of the Saka (Hotan oasis?) language (under misnomer "Saka" is actually meant the Hotan language).
Chinese borrowings in Pra-Türkic
An overview of the suspected early Sinicisms in the Türkic languages can be found in Shervashidze 1989. Here we review only the most indisputable of them, and also borrowed into the Pra-Türkic (i.e., present in both Pra-Türkic and Bulgar branches). The Chinese parallels are cited by the database of the Chinese characters, compiled by S.A.Starostin (project "Tower of Babel", www.starling.rinet.ru).
For dating of the Sinicisms in the Pra-Türkic we used a refined periodization of
in Starostin 1989. According to our dating for fractionation of the Pra-Türkic should
be expected loans
from the Western Han language (3rd c. BC - turn of the eras) or Eastern Han (turn of the
eras - beginning of the 3rd c. AD). The following phonetic developments were considered.
The Eastern Han already makes it possible to assume a transition of the lateral into fricatives in front of the short vowels (Starostin 1989, 468-469), and even l > d in the initials (Starostin 1989, 469). The final transition of L > D was determined for early postclassical ancient Chinese (from the 3rd c. AD), and at the same time finally r, rh > l, lh (Starostin 1989, 476). The transition pr- > p- also occur finally in the early postclassical Archaic Chinese. The combinations of the Tr- type produced retroflexes still in the Western Han. The terminal -r already in the classical ancient Chinese gave -n (Starostin 1989, 711). The Lh > z'b in front of the short in the Eastern Han (Starostin 1989, 468), T > c' in front of the short in the Eastern Han.
1. *Alaču-k 'hut, shed, small yurt': Anc.Uig. alaču; Krh.-Türk. alacu (MK)
Kum. alacyq, Bashk. alasoq, KKalp. ylasyq, Kaz. lasyq, Kirg.
alacyq, Alt. alančyq, Khak. alačyx, Shor. alančyq; Tuv.,
Tof. alazy; Yak. alaha; (?) Chuv. las' 226. See SIGTYA
1997, 497-498. Compared with the late Anc. Ch. la-λia? 'country house', consisting of two
words: 1) 庐 Modern Ch.lü2, Middle Ch. lö, Anc. Ch. r(h)a 'hut, shelter; to lodge' Karlgren 0069, 2)
舍 Modern Ch.she3, Middle Ch. sa, Anc. Ch. *λia?-s 'to rest, stop;
lodging house'. Karlgren 0048 a-b. Dating: r- > l- in the early postclassical
Archaic Ch. , while
maintaining the non-labial finals, i.e. certainly before the late ancient Chinese; lateral
in front of short vowels beginning from the Eastern Han; the
common date is the beginning of the 3rd c. AD.
226 Perhaps Kypch. borrowing in Chuv. - vocalism is not quite clear
(In the framework of the
PCT, a part of Chuvashes
never left the N.Pontic. They also may be those Uraloids that abandoned the Amudarya
valleys at around 2,000 BC. The Chuvash and pre-Han Chinese form may be the original form
of the word that split geographically into western and eastern forms, pik/pek = > Bulg./Chuv.
pik, = > Sakha bi:, = > Yin prak, = > Middle Ch. paik, = > Uig. beg, Sarmat
Polish pan, etc., underwent palatization and voicing, and survived from the Paleolithic to
6. *Čerig 'army', Bulgar. sarak, Chuv. sam, Yak. serii. Is
being associated with the Middle Ch. sen: 戰 , Modern Ch. zhan4, Middle Ch.
sen, Anc. Ch. tar-s 'fight, fight, tremble, be
afraid' Karlgren 0147 r. Dating: by the development of the initial into affricate the Eastern
Han, i.e. the turn of the era - 3rd c. AD. But the final r > n in the classical
ancient Chinese (5th c. BC - 3rd c. AD) 227. Evolution of the root in Starostin 1989:
pre-classical ancient Chinese tarh > classical ancient Chinese tanh > Western Han
Eastern Han tjanh, čanh, postclassical ancient Chinese Chuv. It is a murky
227 Accordingly, all hypothetical early Sinicisms in Pra-Türkic with replacement
of n to r should be considered to be highly questionable.
13. *Kög 'motive' with Anc. Uig., Tuv. xög, Tof. xög, Yak.
küj, Chuv. kəvə, see SIGTYA 1997, 614. Class. Ch. 曲, Modern Ch.qu1, Middle Ch. khöuk, Anc. Ch. khok 'be curved, bent', from the Han appears the
meaning 'music, melody'; Karlgren 1213a. Dating: the upper
limit apparently is the postclassical ancient Chinese khok Starostin 686 (i.e.
before the 3rd c. AD inclusively).
The selection of the borrowed words outlines a range of cultural interaction: trade, in particular, metalwork, war, writing, art, luxury goods, and philosophical concepts. The dating of the borrowings by the phonetic features generally points to the 3rd c. AD. What explains this discrepancy with the glottochronological dating (turn of the eras)? If we assume that the glottochronological dating corresponds to the loss of the first word in the 100 word list, it should be thought that after that for at least another three centuries the divided Türks continued fairly active contacts, and their languages did not undergo significant phonetical changes. In particular, the above cited connected isoglosses do not have phonetic material for the dating of the loss of the first word, and likewise for the third word, but the second word we obtain the lower boundary: no earlier than the 4th c. AD (from the word *gümiü' 'silver'). It should be noted that the corresponding model of the historical situation that can be seen in the history of the Huns' division onto the Northern and Southern: the first separation and withdrawal of the Northern Huns to the west has occurred, as was stated above, in 56 BC, and it is believed that that group of the (Northern) Huns was destroyed by the Chinese and the remaining Huns came under Chinese rule, the second split of the Huns into the northern and southern groups happened in 48 AD, from that time the Northern Huns gradually shifted to the Western Mongolia and later to the East Turkestan, to Dzungaria, and in 155 AD they migrated to the East Kazakhstan and Jeti-su, where they lived till the 5th c. AD 228. It is likely that the first (Northern) group was not destroyed physically, but only lost its state, then the second (Second Northern) group merged with it, bringing with them the Chinese cultural borrowings.
Possible Eastern Iranian loans in Pra-Türkic (In the historical Perspective, the Eastern Iranian essentially boils down to Sogdian, since the contribution of the Tarim and Taklamakan desert oases, amounting to 26,000 isolated people, on 1,500,000 steppe nomads belonging to hundreds of self-contained tribes is conceptually undetectable)
1. *Dāna 'calf': Middle Kypch. (of 14th c.) tana 'calf', Chat. tana 'two-year old calf'; in new languages: Tur. dana 'calf', Gag. dana 'two-year-calf', Az. dana 'calf', Turk. tana 'calf' (without distinction by sex); Salar. tana 'heifer'; Krh. tana 'calf', Kum., Bank., tana 'two-year old bullock', Tat. tana 'two-year heifer' Nog., KKalp. tana 'heifer in the second year', Kaz. tana 'yearling bull calf', Kirg. tana 'heifer in the second year'; Chuv. tyna 'calf (overwintered)'. VEWT 460, Fedotov 2, 267 229, ESTYA 8. Bulgar. *tynag > Hungarian tino 'heifer' Gomb. 130 (voicelessness is not clear). A borrowing from Iran is suggested. (Avest. daenu 'female animal', cf. Skt. dhenu- 'cow' (from "milking"), Khor. (Horesm or Horesm-Türkic) dyn 'woman', Yazg. (Western Middle Indian/Hindi dialect) δang 'with children (of a woman, female)' < *dainu-kā, HSak. (Hotan oasis) dmü 'cow' Bailey 159, Rastorgueva-Edelman 2, 447). The most plausible in this case seems a borrowing from a language of Northern Iranian-speaking population, close to the Saka (i.e. Hotan), which is the only Iranian language that gives a word meaning 'cow', and where appears the first vowel i, which gave a Pra-Türkic source of the closed a reflexes (The direction of borrowing is presumptive, since all Türkic languages have a meaning of cow, but in Hotan the cow is an exception in respect to the Iranian languages. The fallacy of the presumption is rooted in a postulate that Pra-Altaic languages originated in a single spot, and emanated from their origin following a language tree model. The spread of the Türkic languages as they entered the historical period makes the underlying postulate unsound. The work of A.V.Dybo serves as a philological corroborating evidence confirming the fallacy of the Altaic homeland postulate).
2. *Dorak 'cheese': see ESTYA VIII (in print). Traditionally it is unjustifiably identified with Krh.-Türk. tar 'butter-free churned cream', Yak. tar 'frozen yogurt', which also phonetically fits the Chuv. tora(x) 'yogurt' Fedotov 2, 253 and Mong. tarag, with SH tarah 'yogurt' (which then was borrowed from Mong. to Tuv. and Manchu). For this last dish is reconstructed the name PT *tar-aq 'type of sour milk', perhaps a common word with Mong., but also possibly a loanword into Mong. from the ancient Türkic. The base torak does not coincide with the *tar-aq neither phonetically nor semantically; records start with AbûH and Chag., Khalaj tuorāq, Turk. doraq, Tur. dial. torak and dorak; Chuv. tora, twara; Dan. Bulg. Loanword in Hungarian tura 'cheese', all meaning 'kind of cottage cheese, cheese'. See the discussion of etymologies in ESTYA 8, Doerf. 3, ¹ 1195, VEWT 490. This base may be a borrowing from Middle Iran. *Tura-ka, compare Av. türi- 'curded milk', which has IE etymology (see Ab. 3, 319; of Iranian forms also compare Saka (Hotan oasis?) (doubtful) ttüra 'cheese' Bailey 132, see also 124 tav- 'to go sour' Osset. turae 'fatty soup. "(However, the second Türkic base can also be Iranism, compare Av. tayuri- 'kind of bread' Barth. 647, Zoroaster, Pahlavi *ter, but a semantic similarity is much worse, documentation is worse). All Iranian traces are Northern Eastern Iranian. The borrowing would take place before the global Türkic voicing. (A little sanity check would not hurt. All "Northern Eastern Iranians" were either intermingled or immediately adjacent to the steppes, interfacing many, but not all Türkic tribes. Philologically, all of the Türkic tribes have that word for a product made of a soured milk, but only a tiny sampling of Indo-Europeans have it, and among Iranian languages it appears only at a late Middle Iranian. The source of the product also must not be ignored. Nomads carried milk in bags made from animal intestines, the final product depended on what milk and what intestines were used, and the particular microflora of different animals. That variety required extensive location specific linguistic nomenclature, the buttermilk soured in the camel guts was different from that soured in the goat guts even if the starting milk was from the horse milk. The agriculturists had little access to that linguistic palette, and were borrowing only of a pidgin type terms, generic in nature. The philological obstacle is that most of the nomadic technology and associated vocabulary is either lost or not properly documented. The idea that pastoral nomads who live on milk products would need to borrow a milk product word from the agriculturists, and not the other way around, is somewhat funny. Where we do have a documented source, it is from the nomads to the sedentarists: Herodotus reported that the IE Greeks did not have cheese, and learned it from the Scythians, with appropriate source name ascending to Tyran. Not from the Persians, whom they intimately knew, and learned a few things from, but from the Tyran Scythians)
228 These dates are given per: Klyashtorny S.G. Hun
state in the east./ /History of the
Ancient World. Decline of ancient societies. Moscow, 1989; Kryukov M.V., Perelomov L.S.,
Sofronov M.V., Cheboksarov N.N. Ancient Chinese in the era of centralized empires.
1983, Chapter 2: The ancient Chinese and their neighbors: racial and ethnic
characteristics, pp. 56-104. But the parallel proposed there Mong. cinoo 'calf' does not
exist: it is a 'wolf'.
Note that these both loans belong to the area of dairy farming and cattle husbandry, which seems quite natural for the early Türkic-East-Iranian contacts (cf., for example, main semantic fields that included Iranian loans into the Finno-Ugric languages) (These Türkic-East Iranian contacts must have been before the 12th c. BC, i.e. before the East Iranians ever existed, since the first Jungs/Hunyi/Xunyu/Xianxun that reached Shang were already fully formed nomadic pastoralists that had nothing to learn about milk products from the agricultural people who are even now noted for their lactose intolerance, which distinguishes the Melanoids and Polynesians, Indo-Iranians, and Chinese, see map on Wikipedia).
3. (?) *Dura 'tower, fortification, four-sided log or stone building': from Krh.-Uig., Gen.Türk., Including Siberian (without Chuv., Yak., Tuv.). See SIGTYA 1997, 486. Loanword in Mong. tura 'fortress city', West-Bur. tura 'chalet'. Altaic etymology was not found. In principle is possible a borrowing from the languages of the Northern Iranian-speaking people: cf. Saka (Hotan oasis?) ttaura 'wall' Bailey 134 (from the verb tau- 'cover'), Osset. tyrg 'porch, carport, balcony, yard' Abaev 3, 341. For Saka (Hotan oasis?) could be assumed a Turkism, but Osset. can not be a Turkism, compare to the same Skr. torana 'gates' (The reference to Ossetic is presumptive, since before Abaev composed his dictionary, Ossetes lived for millenniums with Türkic people, their vocabulary is more Türkic then Iranian, and they consist of three distinct tribes of totally different origin. The Digor tribe, which was a base of the Abaev's studies, in Modern Times is reported to be Türkic-speaking, and Digors are connected with the Türkic Tuhsi tribe, otherwise known as Tokhars and Toksi. The Ossetian paradigm is a subject of disputes that involves the Scytho-Iranian Theory which is based on the Osseto-Iranian connection. Any philological reference to Ossetian must be based on unbiased credible studies).
4. Perhaps to the same ancient borrowings can be connected the title recorded in the ancient Türkic: Orkh.-Türk., Enis.-Türk. tarqan, pl. tarqat, Anc. Uig. runic tarqan, pl. tarqat, Man., Budd. tarqan 'title or position; component of proper names'; Proto-Bulg. ΤΑΡΚΑΝΟΣ 'title'; Krh.-Türk. tarxan 'pagan name for the emir' MK; Chag. tarxan 'social estate freed from taxes and endowed with a number of other privileges; one of the Juchi ulus tribes' (rather, Chag. is a reverse borrowing from Mong.), the later Türk. forms most likely are also Mongolisms, reflecting in the semantics either later Mong. reduction of meaning, or the Kipchak Khanate (orig.: Horde) socio-political realitiy: Türk. tarxan 'privileged class; figurative: spoiled' = Az. dial. tarxan 'spoiled' Bashk. tarxan 'privileged class', Kirg. darkan 'smith', KKalp. darqan 'free, at large' Yak. darxan 'important, worthy'; Chuv. torxan 'proper name; privileged class; a component of a number of names of supernatural beings' (the latter meaning may ascend to Pra-Türk.). Fedotov, 2, 239-240, EDT 539, ESTYA 1980, 151-153 (In pre-Mongolian times the title “tarkhan” was widely used in the Eastern Europe, Central Asia, and in the Middle Asia, including Sogd, which saw a number of rulers entitled “Tarkhan”; the notion that it is re-importation from Mongolic, or borrowing from Iranian, is not serious).
By an ingenious hypothesis of E. Pulliblank, this title may be reflected in the title of the Huns' ruler Shanyu (Anc. Ch. reading of the corresponding Ch. rendering: *dan-γwaγ) - Pulleyblank 1962, 91. Türk. > Mong. darkan: Middle-Mong. darkan 'title, combined with certain privileges' SH 32, Written Mong. darkan 'master, estate, free from taxes' Kow. 1676 Khalkha darxan 'social estate; master; smith', Bur. darxa(n) 'smith, craftsman, Personally free' Kalm. darxn 'smith; class' Ramst. KWb 78. Numerous hypotheses about the origin of the word, see Doerfer 2, 879, ESTYA 1980, 152-153. The most plausible seems the Iranian etymology, advanced by V.I. Abaev: Sogd. trχ'n [tarxan] Gharib 9644 'title' (in particular, of Prince Devashtich, see Freiman 1962, 42-45), Saka (Hotan oasis?) ttar-kana- 'title' (I think, however, Türkic), Osset. tærxon 'judicial court', Skr. tark- 'to solve, hypothesize, judge' with Indo-European etymology (*tlk-, WP I 744), Ab. 3, 276-277 (Without reviewing Türkic etymologies, this Abaev's assertion is little credible: since the bulk of the usage, and the spread is associated with the Türkic history, the first unbiased option should be Türkic etymology; which gives the root and sufficient quantity of derivatives. Any assertion that studies only a preferential option is a sham).
All Iranian traces belong to the eastern group (Indicating a classical case of the borrowing, unless it can be proved as an ingenious dialectal innovation unconnected with the Iranian languages). Into the Türkic languages the word should be borrowed early enough, because the Mong. borrowing from the Türk. still reflects the initial sonority (a consequence of Genaral Türkic, but not the Bulgar Türkic, sonorousness of the *t- of any origin in the position before the combination -rk- - see Dybo 2003), subsequently eliminated everywhere except the Oguz and Sayan groups (cf. the Khitan dalayan Menges RE 155). The Hung. tarchan 'olimjudex' (Abaev, citing Munkácsy), judging by the semantics, is taken directly from the Alanian (of which we know nothing, and can fantasize all Abaev wants, in addition calling it “Scythian”). Perhaps Anc. Uig. terken 'queen', Krh.-Uig. terken 'appeal to the ruler; a Queen', Horezm-Türk. tārkān 'princess' (Khosrow and Shirin) EDT 544 is an adaption of the Eastern-Iranian *tarkanaya- or *tarkanf: relative adjective or feminine gender (In this insane etymological report, an Eastern-Iranian imaginary relative adjective like a “*whose” is equated with a Türkic title for Queen or even for a Shanyu King. The whole asinine quality of the tarkhan example demonstrates the extent which in the first half of the 20th century the fabrication of the facts on the ground reached in extreme cases. The sequence is even more ridiculous then the “translation” of the Ioan Tsets' Alanian phrase. It is hard to second-guess the motive, but one possibility is a support for a doctrine that the Middle Asia was Turkified only after a Moslem conquest, that before the conquest it was a pristine Iranian, or in a larger picture Indo-European domain. Not that that is very non-asinine, but at least no party involved can be accused of hearing voices ).
5. (??) Arju, arsu MK, IM 'jackal, hyena', arju-la- 'to stand as jackals' Late. Anc. Uig., MK (Cl. 200: "the -j- suggests a foreign (?) Sogdian origin") - Eastern Iran. *arsu 'bear' > HSak. (Hotan oasis) arm Bailey 8, Sogd. ššh Gharib 1770, Vanetsi (what is Vanetsi/Vaneti) yirz, Pashto yaz, Sarikoli uirh, Sanglechi xars, Shugn(i) yürš. Changing of the meaning at borrowing can be attributed to tabooisation (The Türks did not know tabooisation, and neither did Huns, as the Chinese annals state. The "Eastern Iranian" stands alone in this case, separate from the Iranian, indicating a non-Iranian word. As a Türkic word, apparently it is specific to Kashgar).
6. *Kumlak 'hops'. Indeed a Comm. Türk. and even Pra-Türk. word. With MK qumlaq (Kypchak); Middle Kypch., Chat. qumlaq 'plant, mix into the intoxicating mead' Houtsma; Pav.C. 436. In the new languages: Kypch. Krh. kumlak, kumlah, Tat. kumlak; Sib.-Tat. kumlak; Tum. 3C 152, Tat. dial., Bashk. kumalak; Nog. kylmak 230, Kaz., Kirg. kulmak, Mount. Alt. kumanak, Khak. humnah, Shor. kubanak; Chuv. hamla, see VEWT 299, EDT 628, ESTYA 2000, 138-139, Fedotov 2, 326. Bulgar. > Hungarian komló (see Gomb.BTLU 97, Gomb. 100-101, MNyTESz II 537). From Chuv. language was borrowed the Mari umla, əməlà 'hops' (see Räs.ČLČ 235).
The tentative Iranian source of the word is Alanian etymon Osset. xymoelloeg 'hops', it is etymologized in Ab. IV 261 as a compound *hauma-aryaka “Aryan haoma”. The etymology is phonetically flawless and semantically plausible, the names for the hops in the Caucasian, Türkic, Finno-Ugric, Slavic and Germanic languages are declared to be Alanisms (This is a best example of circular logics. The oddball word in Ossetian is decreed to be an Iranism, then the Scythian is Iranian, then the Iranian versions are borrowed into the Türkic languages. A break in a single link in that chain of presumptions reverses the presumed direction of the borrowing, turning it from a single oddball exception into a normal borrowing. The Abaev's lloeg = > aryaka = Arian is flawless? It is not even funny, all Ossetian elloegs are thus raised to pre-historic nationalistic pedigree. An adjective definition is used when there are alternate products. like Damask steel vs. regular steel, it indicates an import from an alien source, like Polish sausage, it screams “I am a borrowing, I am from these weird alien Arians!”). As correctly noted I.A. Shervashidze, this is the only plausible ancient Alanism in the Türkic languages. We shall note that in the other Iranian languages the reflexes for the name of the same plant fairly consistently retain the meaning of 'coniferous foliage, ephedra': *haumiā- (for some of the cited lexemes also is possible a reconstruction *hauma-) 'coniferous foliage, ephedra': Afg. uməm 'ephedra' (Asl.: 93), Sogd. (Al-Bîrûnî) hwm 'Pflanzenname', Persian houm < 'ephedra' (PRSl II: 734), Talysh hoəme 'hops, ivy' (Pireyko L.A. Talysh-Russian dictionary. M., 1976, p. 242), cf. Avesta haomya- 'zum Haoma gehörig'; *haumāna- 'ephedra': Afg. uman m. 'ephedra, coniferous foliage' (Asl.: 92), Munji yumana "ephedra", Yidgha uptepa 'ephedra'; *haumāka- 'ephedra': Wakhi (y)imúik 'ephedra', Òaj. dial. hyta 'ephedra'; *haumdčî (+ ak) 'ephedra'; Shugn(i) amojak 'ephedra' (Karamshoev 1: 91 231), Rushan. amojak 'ephedra', Huf. amojak 'ephedra' (see Stebl.-Kam. NKR Stebl.-Kam. Wakhi) (see Dybo 1999).
230 Nog. kupelek 'hops' may be new Ossetian borrowing, or simply a derivative of
kupe 'pot', cf. Az. Kün√-chich√ju 'fuchsia' (climbing ornamental plant, with hops-like flowers, but
larger, literally 'flower-pot').
Ephedra, a low shrub of the cypress family, does not look like hops, so the transfer of semantics is explained functionally: a transition to the hops to produce intoxicating beverages. But close to the Central Asian species of Ephedra are spread in the Caucasus and Northern Black Sea Coast (Trees and Shrubs of the USSR, 250), so it is unclear why would the Alans in such traditional and ritual area switch to another type of plant with a transfer of the name, but the Talysh form unconditionally supports the Ossetian etymology 232. The chronology of the Alan's phonetic transitions also does not contradict the supposition of borrowing into the Türkic (Under “Alanian”, of which we know absolutely nothing, and the same about its chronology, the author apparently refers to Ossettic).
The Türkic word generally has almost satisfactory Altaic etymology: PAlt. *k'iomo/V 'kind of fragrant edible plant', Mong. *kömeli 'kind of wild onion/garlic' Less. 487, Khalkha x(m(l 'Mongolian onion' TM *ximŋe-kte 'bird-cherry ' SSTMYA 1, 318, Jap. *kamira 'kind of garlic' (EDAL). The Bulgar form could serve as a source for Slavic and Germanic (cf. the collection of Bulgarisms in the Germanic languages, collected by V.A. Terentiev - Terentiev ST) forms, and the Ossetian word can also be a Bulgarism. In any case, it undermines the reliability of the "single Alanism" in the Gen. Türkic. Still, accepting the Iranian origin for the Pra-Türkic word (in that case the semantic side looks much better: the garlic and bird-cherry are compatible as sharp-smelling edible plants - and compare the parallel relation in the Indo-European (bird-cherry - wild garlic) - but the garlic and hops are not similar either externally nor functionally, in any case, the hops is definitely not used as a fragrant plant), then from the chronological point of view should be suggested a borrowing from the language of the Eastern Iranian carriers of the archaeological cultures in the Sayan-Altai region, which possibly close to the Alanian (i.e. Ossetian) 233.
231 In Zarubin 1960: 89 the meaning of this word is is also presented in the following
form: 'a shrub genus, which ash is used in the preparation of chewing tobacco (nas)'
Judging from glottochronological datings for the Iranian tree, borrowing from the Eastern Iranian languages into Pra-Türkic should have happened in a period after the split of the Saka and Alan branches (lexial-statistical dating 660 BC), already from the languages of Middle Iranian type, but possibly before the split of the Saka (Hotan oasis?) and Pashtun branches (lexial-statistical dating 290 BC).
Pra-Türkic and Tocharian (Tocharian is a name for Bactrian, and Bactrian-based languages, but in the following discourse the misnomer “Tocharian” is applied to the Kuchean (Tokh. A) and Arsi (Tokh. B) oases)
List of alleged Tokharisms in the Pra-Türkic proposed by A.Rona-Tash in a number of his
works seem to contain a number of questionable assumptions, we would prefer here here a
skeptical point of view, as proposed in Reinhart 1990 234. In addition to conventional
late borrowings, about which see below, almost all of these alleged loans are quite
reasonable Altaic etymology - see EDAL. Exception is the Turk. *bel'k 'five', which,
however, is unlikely in such a reconstruction (taking into account the Chuvash development, see SIGTYA 2002, 350) may be compared with Pra-Tokhar
***päns 'five' Adams ¹
2099 [Douglas Q. Adams A Dictionary of Tocharian B, Leiden
Studies in Indo-European 10,. Rodopi: Amsterdam – Atlanta, 1999, ISBN-10: 9042004355,
ISBN-13: 978-9042004351]. Very speculatively to the Tokharisms could be attributed the Pra-Türkic name
for the apple, because
it is clearly borrowed from some Indo-European, but not from the Iranian language,
preserving the distinction between r and l: *alma, Krh.-Uig. alma (MK) ~
Chuv. ulma (see SIGTYA 1997, 145) in IE *amel- 'fruit tree, its fruit'
(Anc. Greek ampelo-s f. 'grape vine', Germ. *amil-on- f. 'some fruit-tree', Gaul.
WP I 179, possibly also Hett. sham(a)lu, Luv. nom.-acc. sg. šamluwan-za 'apple' Ivanov-Gamkrelidze
639-640), > Skr. amra-m. 'Mango tree', Iran. *(a)marna-Stebl.-Kam. NKR 103-104, (
> Finno-Ugr. *omarna- > Fin. omena, Mord. umar SKES 429-430) 'apple',
*amru- 'pear' Stebl.-Kam. NKR 108. The expected source of the Türkic word would have to look like
- the reflex and the meaning, in principle, possible for the Tocharian. Unfortunately,
the Tocharian name for the apple is not recorded. Three fitonyms, reconstructed for the Pra-Türkic (with
a regular Chuvash parallel) and suspiciously resembling an Indo-European origin (but not Iranian, cf.
Centum reflex *- g'-) - *ekel 'acorn, pine cone' (cf. IE *aig'-il- 'a tree (a
type of oak), and its fruit' WP I 10), *elmen 'aspen, elm' (cf. IE *elem- 'tree
(elm)' WP I 151) and **ab(u)s-ak 'aspen, poplar' (cf. IE *ap[u]s 'aspen, poplar, ash,' WP
I 50, a comparison suggested in Róna-Tas TE) are limited in distribution - see SIGTYA
1997, 121, 126, 131 (which may be due to the limited spread of the plant at the current
Türkic territory - see above in the section on botanical terminology and the migrations
of the Türks), and thus have no solid Altaic etymologies. Phonetically, they could easily
ascend to Comm. Tokhar reflexes of the respective Indo-European bases (the breakdown
of the Comm. Tokhar into the A and B languages is glottochronologically dated by 20 BC),
but in the Tocharian monuments such reflexes were not found.
An interesting collection of potential Pra-Türkic loanwords into Comm. Tokhar published by A. Lubotsky, and S.A.Starostin (Lubotsky-Starostin; a large proportion of these etymological proposals repeat those suggested by A.Rona-Tash, but with the opposite direction of borrowing) (I.e from Türkic to Kuchean and from Türkic to Turfanian/Arsi oases): there are the following words (that do not have satisfactory Indo-European etymologies):
1. Tokh. A (Kuchean) kot, Tokh. B (Tarim) kait 'sun, day' < PTokh. *kaun(V) -:
(Anc. Uig. êüï 'sun, day', Türk. gün 'id.', etc.) < PAlt. *giojnu 'dawn,
daylight' ("It is
important that the combination of meanings of 'sun' and 'day' is quite unusual for the
Indo-European languages, which is a strong indication in favor of borrowing" - AL, SS).
A.V.Dybo cites the Proto-Altaic *t'öre for “soil,
dust” without stopping for elucidations, even though phonetically and semantically it is a clear
cognate of the Romance “terra” and its many variations and derivatives).
The last three words, witnessed only in the Tocharian B, may belong to another category
of the Turkisms in the Tocharian, about which see below. Note that here the borrowing,
first, predate the transition r' > z in Türkic, and secondly,
before the split of the Pra-Tokhar (i.e., before the 20 BC)
(Treating the Kuchean/Aksi language as a monolithic phenomenon that can be pinpoint
dated contradicts the composite nature of these languages, which are a historical
conglomerate of Bactrian, Sogd, Prakit and a number of other Indian languages, Türkic
Oguz, Türkic Ogur, Chinese, Tibetan, etc., assembled in a microcosm of the desert oases,
where the influences are disproportionally magnified because of the miniscule population
Predicted by all that dating the time and place of existence and diversification of the Türkic Proto-language seems quite well located in the large territory between the present Ordos and the southern Sayan-Altai, within the frame of the Loufan (Loufang 楼烦, 樓煩) archaeological culture, whose carriers were engaged in transhumant cattle breeding in a large territory, which neighbored Yuezhi, which demonstrates extensive connections with China and more modest connections with Pazyryk (ethnically Iranian) culture, and most likely they are connected with a ruling dynasty of Pazyryk origin they had for a certain period - see Shulga, 1999, see also this book , p. 393 (For Pazyryk Andronovans, we have so far only mt DNA for female lines of both males and females; the Y-DNA is not known yet; the mamas are from Mansi, Tuva, and Kazakh Scythians, Paleo-Sibirian and Altaian Scythians, N. Altaians, Teleuts, Shors (from craniology), Paleo-Sibirians and Asiatics (Mongoloids). No connection with Indo-Arians or Indo-Iranians whatsoever, neither in genetics, nor in thin-walled gracile osteology. One would strongly suspect that the Pazyryk people, who were living on milk products, also would not have the Indo-Arian/Indo-Iranian lactose intolerance. How the philologists made Teleuts, Shors, and Altaians speak Iranian is beyond comprehension. Even more puzzling is how philologists made them to uniformly abandon their native Iranian language that they supposedly received with their Teleut mother's milk, and switch in situ to a totally foreign Türkic without a trace of historical memory, including philological, osteological, and biological. Mind you, the biological studies were performed within a decade before the professor A.V.Dybo 2007 publication of this work, they instantaneously made the Scytho-Iranian theory a purely philological construct divorced from the factual evidence, and decimated all ethnological aspects of the theory. These philological miracles would have made the Circus of Soleil performances pale: it is easier to turn an elephant into an airplane than to turn cereal-eating Indo-Arians into pastoral Türks. In a country built around sobriety checkpoints, scientific roads do not require sobriety checks for the scholarly travelers. And again, wouldn't the science that reconstructs a language of long -dead peoples use a system of checks and balances, and do a check of the ethnological complex before making its conclusions an axiom?
In following the Iranian doctrine, A.V.Dybo missed on two cardinal points: the Tokhars/Tuhsi (Yuezhi in Anc. Chinese) were the same Ogur tribe as the Huns; moreover, depending on which tribe held the supreme title, the Tokhars/Tuhsi were the Huns, or the Huns were Tokhars/Tuhsi, as was the situation before Mode. The tribal fight for supremacy was endemic to Türkic society, it had never stopped. Secondly, the Pazyrykans were a contemporal branch of the Huns, or vice-versa. While the Pazyrykans were building kurgans in the Altai, the Huns/Jungs ets. were building kurgans in and around Yin China. They, Tokhars/Tuhsi, and Huns were politically distinct, ethnologically identical, ethnically nearly identical, diverse and divergent.)
Pra-Türkic and Samoyed (Nenets, people that speak Nenets languages object the use of the term Samoed as derisive, it comes from a Russian folk moniker Self-Eater)
As to the Nenets (Orig.: Samoyed) borrowing into Pra-Türkic, the most recent survey assessing reliability and tentative timing of the borrowing, see Helimski 2000, 301-312 235. We can note in this regard the following: E.A. Helimski evaluats as reliable borrowing into Pra-Türkic and General Türkic 7 etymologies (1. PS *kaətyə 'spruce' > PT *kady 'pine', 2. PS *ki, *kilz 'sable' > PT *kil 'sable', 3. PS *tyteŋ 'Siberian cedar' > OT *tyt 'larch', 4. PS *koəjə 'mountain, watershed' > OT *K(i)aja 'rock, mountain', 5. PS *käsa 'bark' > OT *kās 'bark', 6. PS *kacu 'blizzard' > OT *kād 'blizzard, storm', 7. PS *talä- 'steal' > *PT tala- 'rob'); the glottochronological dating of the Pra-Nenets (Orig.: Samoyed) divergence - 410 BC (splintering of Selkup), the breakup of other languages onto northern and southern groups - 160 BC, which, in general, is consistent with the possibility of borrowing from the Pra-Nenets (Orig.: Samoyed) or Pra-Southern Nenets (Orig.: Samoyed) into the Pra-Türkic (tentative localization applied to the 1st millennium BC - the eastern part of the Ob-Irtysh interfluvial - Helimski 2000, 17).
235 On impossibility of borrowing from the Ob-Ugric languages
into the Gen. Türkic of the word *degin 'squirrel' see SIGTYA 1997, 165. PT *kundur 'beaver' (SIGTYA 1997, 162) does not
unambiguously etymologize, and in principle can be borrowed from - or might as well be a source -
of the Pra-Ugr. *kuntz 'beaver' (MNyTESz 2, 127), which does not have further Finno-Ugric parallels. Currently,
the beaver is lightly spread east of the Urals (small
spot ranges in the area of the Ob-Irtysh river and further in the upper Enisei River and
the Lake Baikal - Mammals of the USSR, 281, map 140), it is rare further south (Murzaev 1966,
250: now is absent "almost in the whole of the Inner Asia, except for the upper Urungu and
possibly some tributaries of the upper part of the Black Irtysh, i.e. in Dzhungaria),
but (ibid.) such limited area should be viewed as a relic", so that we can not exclude
that that it was known in the mountain (and thus forest) regions of the Türks' ancestral
homeland. Taking a borrowing from Pra-Ugr. should apparently recognize
a singular character of the borrowing (perhaps occurring as a result of trade relations
within the east-west trade routes). The divergence of the Hungarian from the Ob-Ugric
is glottochronologically datedly by about 1,000 BC, the split of the Khanty and Mansi is
dated by 130
AD; can be considered a borrowing into the Pra-Türkic of the Pra-Ob Ugrian suffix forms: *kuntz-l' (in UEW 858
the opposite, a borrowing of suffix-less form into Pra-Ugr., but it is less likely culturally and historically).
However, the subsequent studies of the Altaian kinship did not confirm these loans,
because all these bases have quite convincing Altaic etymology (EDAL:
However, these are cases where the Nenets (Orig.: Samoyed) borrowing from a formal point of view still is more likely than the Altaic etymology:
4. PS *koəjə 'mountain watershed' > OT *K(i)aja 'rock, mountain'. The Nenets (Orig.: Samoyed) word has Uralic etymology (albeit not common Uralic: PU *kaδ'a > Hungarian hegy 'top, mountain'), the Türkic word seriously violates a standard pattern of development PAlt. *d > PT *d, demonstrating *j instead of *d (Who violates whom is a mute question: Are Scythians PAlt. or PT or what? Herodotus extensively used the Scythian word Caucas(us) in the 6th c. BC, Pliny 6.XIX stated that Scythians call Mount Caucasus Croucasis, which means “white with snow”, in both cases, separated by 650 years, the mountains were called by actual “kas”, not by a fictitious “*kaj”. In such cases, the theoretical “standard pattern” should be questioned, instead of the correctness of the Scythian pronunciation. In the word “Kazbek/Kazbeg” ~ Türkic “Prince Mountain”, the mountain is phoneticized “kaz” , which is consistent with “kas” , and is another fact on the gTur. The Kaskak in Altai mountains means “precipitous slope”, with a common Altaic kad~kaz “bluff, cliff”).
5. PS *käsa 'bark' > OT *kās 'bark'. Semantically, the
Türkic word deviates somewhat from the Altaic etymology, whhile the Nenets (Orig.: Samoyed)
word has a normal Uralic etymology (PFU *ko(v)skz 'bark'). The cases 1. PS *kaətyə 'spruce' > PT *kady 'pine' and 3. PS *tyteg 'Siberian cedar' > OT *tyt 'larch',
in addition to conflicting with the hypothesis about borrowing of a fact of the existence of
Altaic etymologies for the Türkic words,
though the paired, but phonetically and semantically quite correct, both offer the same semantic (and cultural)
oddity. Namely, accepting the hypotheses about the borrowing from Nenets (Orig.: Samoyed)
results that the Türks, adopting a name of a previously unknown to
them tree (cf. Terentiev, 1999, pp. 180: "The fact that in the Türkic languages this word
came to mean "pine" should not be confusing, because ... the Pra-Türks, living in the
steppe zone, were not sufficiently aware about the types of the conifers), for some
reasons borrowed for it a name of an entirely
different tree. As for the geographical distribution of the discussed trees, see above the relevant
section of this book; note that the cedar was witnessed on the southern slopes of the
Mongolian Altai (this is a southern limit of its distribution, cf. Murzaev
1966, 216), but the larch extends much farther to the east and south, in particular in the Tien Shan and Ordos,
i.e. much closer to the presumed location of the Pra-Türkic habitat - and most
likely was well known to them as a pine and spruce; in these circumstances a borrowing of the cedar
name to designate a larch, or fir name to designate a pine is unlikely
(Even much less
probable if the artificial prejudicious limitation on the original linguistic areal is
discarded. The Assyrians recorded an extensive Türkic lexicon in their neighborhood, not
in the “presumed location of the Pra-Türkic habitat” in the Loufang Ordos [A.D.
Mordtmann, “Über die Keilinschriften zweiter Gattung“, ZDMG XXIV, 1870, p. 50]).
With regard to the case 2. PS *ki, *kilz 'sable' > PT *kil 'sable', the Nenets (Orig.: Samoyed) word has no good Uralic etymology (cf. the new etymology PFU *kaδ'wç 'female', usually compared with PS *ki: PS *kejmā 'female' Janhunen 66, significantly more acceptable phonetically and semantically - Aikio 2002). According to zoological data, the sable inhabit the western and eastern Siberia, and the Far East, but to the Western Siberia the areal of its distribution comes fairly relatively (see map 75 in Mammals of the USSR, 137: a small territory north of the middle course of the Ob River), while the basins of the Enisei and Lena are included completely or even to the south of their sources. Generally, much more likely is the borrowing from the Pra-Türkic into the Pra-Nenets (Orig.: Samoyed) (with subsequent common Nenets (Orig.: Samoyed) transition *-l > *-j at the end of a syllable).
Thus, we have to abandon the hypothesis on the borrowings of the major terms for the “taiga realities” from the Pra-Nenets (Orig.: Samoyed) into the Pra-Türkic. It should be stated that in this regard also look much more doubtful the above cases 1. and 3.: with the absence of massive borrowings from the Pra-Nenets (Orig.: Samoyed) into the Pra-Türkic cultural lexicon is unlikely a borrowing of such words as "bark" and "cliff" (the first of which is even included into the 100-word Swadesh list, little permeable by definition).
The most recent discussion of the alleged borrowing from the Pra-Türkic to the Pra-Nenets (Orig.: Samoyed) was apparently published by V.A. Terentiev (Terentiev, 1999, pp. 182-194). Of the 80 reviewed etymologies, for which he suggests varying degrees of reliability, the following are credible semantically and phonetically, and to a some measure meet a criteria of cultural outcome 236:
236 Apparently, the following alleged borrowing from the Pra-Türkic
to the Nenets (Orig.: Samoyed) need comments. 1. For semantic reasons,
is questionable the possibility of borrowing the Nen. to' 'blanket' from the Türk. *don 'gown';
Kamas. ton 'fir coat', as V.A. Terentiev correctly pointed,
most likely the later borrowing came from some Siberian Türkic language. 2. PT *kap- 'capture'
> PS *kəpi- 'to yank'
- but compare the Ur. *kappV (Szin. 35), Comm. Nostratic root, so the borrowing
hypothesis is unnecessary. 3. PT **kalbuk
> *kasuk 'spoon' >
PS *kajwa 'spade,
shovel, oar' Janhunen 63 (Ngan., Nen., Kamas.). Compare the normal Uralic etymology of
the Nenets (Orig.: Samoyed)
word: UEW 170-171. 4. PT *Kugu 'swan' > PS *kukə (Janhunen 76-77) (Tundra
Nen., Tavgi); sooner should be suspected a borrowing of the TM *kuku/*huku SSTMYA 1, 426-427, 2,
336. 5. PT *ködeč (VEWT 286) 'vessel' does not exist, see Dybo,
The linguistic contacts,
the respective forms are divided into three different etymologies, among them is
Sib.-Türk.: Khak. ködes 'earthen pot, kettle' Shor. ködeš id.,
SUig. kodiš 'stove' ( <
*götič), ascending to the *göteč, which may have an Alt. etymology: TM *kota- 'tableware, cup, bowl' SSTMYA 1, 418. Accordingly, less likely becomes the borrowing into the PS *kyttiä > Ngan.
kita 'scoop', Forest. En. kide 'trough-like vessel', Tundra. Nen. hydya 'cup, bowl', possibly from PTM?. 6. *Küren 'ferret, weasel' >
Kamas. kürö 'weasel'.
Not justified is the Terentiev's assumption about an earlier meaning of the 'weasel' in the Türkic
languages, because the Hungarian borrowing from the Bulg., göreny, also means 'ferret', and
Pra-Türkic is reconstructed another word for ermine, *iars, see SIGTYA 1997, 163; phonetically
the Selkup and Kamas forms are poorly traced back to the PS state, but a separate borrowing from
is impossible. From Mong.? Why not compare the Hungarian directly with the Nenets (Orig.: Samoyed)?"
1. PT *jür 'hundred' > PS *jür 'hundred' Janhunen 50 (without Selkup), Cf. Helimski
Pra-Türkic and Pra-Enisei
Below, we focus on a reconstruction of Pra-Enisei phonetics, performed in Starostin 1982, and Pra-Enisei lexicon, reconstructed in Starostin 1995, as well as its expanded and updated for the latest literature version, published in the Internet at www.starling.rinet.ru 237. The split of the Pra-Enisei language by that data is glottochronologically dated by the turn of the eras, when developed two branches, "Ket" and "Kott". From the "Ket" separated in the 5th c. AD split a Pumpo-Kola language, then in the 12th c. AD diverged the Ket and Yug languages. The Kott Group in the 6th c. AD split into the Kott and Arin languages.
A number of loans of the historical times between the Enisei and Türkic Siberian languages are quite well known. The following is a list of possible borrowings into the Pra-Enisei and separate Enisei groups from the Pra-Türkic and possibly from some Türkic groups. 238
237 Compared with the "paper" publication, there are in particular the
updates and etymological discussions from the Werner 2002.
Pra-Eniseian loanwords (no later than the turn of the era):
Loans that have taken place not later than 6th c. AD (Only in "Kott" branch):
Loans of not later than 12th c. AD (Only in the Ket-Yug sub-branch):
Two potential Türko-Enisei contact words require a separate discussion.
2. PEn. *kun 'skunk' (KS 1995, pp. 242, Ket-Yug-Pumpokol, was proposed Sino-Caucasian etymology, there is another PEn. name rossomaha (wolverine, skunk)): cf. Türk. forms: Shor. kunu, Sib. Tat. kunu, Khak. kunu 'rossomaha', Bashk. qono, Krh. quna 'marten' VEWT 300, SIGTYA 1997, 162. Also compare Mator-Taigiy-Karagas kun'e 'ermine' Helimski 1997, 34. All of these words undoubtedly remind the Comm. Nostratic name for marten (PIE *keun-, PKartv. *kwenr-MSSNYA 346), but the Türkic word can not be its parallel, because likely such word is the PAlt. *kuren- > PT *Kuren 'ferret , weasel, marten' (see EDAL). Upon a closer examination it turns out that the Türkic forms can be divided: Sib. Tat., Shor. and Khak. kunu 'rossomaha' can be asserted to be a borrowing from the Enisei languages, and Bashk. qono and Krh. quna 'marten', both ascending to the *kuna, consider as a borrowing (apparently, initially in the Kypchak) from the well-known Anc. Rus. kouna 'marten; money unit'. The problem is the Nenets (Orig.: Samoyed) form: for the Enisei (or Siberian Türkic) borrowing it has a strange development of the meaning for the Uralic parallel to the Nostratic form, it is too isolated. As for the borrowing of the name rossomaha, compare PT *jebke 'rossomaha': Khak. jekpe, Yak. siegen, Dolg. hiegen, Tuv. čekpe VEWT 195 (has an Altaic etymology: PAlt. *zipe 'a kind of large predator': PMong *sibor 'leopard', PTM *sibige 'wolf, bear'), borrowed into Arin dzhip'ka 'marten', Kamas. djapka 'marten', Mator-Taigiy-Karagas djibke 'rossomaha' (Pallas) - see Helimski, Keto-Uralica, 248, also with a strange development of meanings. The modern areal of rossomaha (wolverine) fully covers the basins of the Enisei, Lena and more eastern and southern regions; in the Ob-Irtysh rossomaha (wolverine) is represented much worse (Mammals of the USSR, pp. 42-43, map 79). Thus, geographically most easily can be imagined a borrowing from the Enisei or Türk to the Nenets (Orig.: Samoyed) languages, but, judging from the unsettled meanings we deal with later reciprocal borrowing in the Sayan region, of rather taboo nature.
Note that the words "fish", "grind" and "willow" fell on the list of the pra-language borrowing both in Pra-Nenets (Orig.: Samoyed) and Pra-Enisei. Apparently, the borrowing between all three Proto-languages went on in the contact territory that includes the Enisei and Nenets (Orig.: Samoyed) ancestral homeland. The natural dating of these contacts would have been by the initial period of the Türks' advance toward the north-west, respectively, correlating with other datings and localizations, by the 2nd -1st centuries BC (Note that Shanyu Maodun in 203-202 BC subjugated to the Hun states the tribes in the Sayan, Altai and Upper Enisei - Eastern Turkestan 1992, pp. 118). In this connection, we can assert that the modern glottochronological dating of the Proto-Nenets (Orig.: Samoyed) divergence is somewhat too deep, may it should be returned to the traditional dating by the turn of the era (see, for example, Helimski 1982,45-46) (The premise that political control is equivalent to the linguistic distribution is doubtful; the premise that the Türkic people and languages were in the period prior to the 2nd c. BC completely absent from the enormous territory attributed to the Enisei and Nenets homelands is not any less doubtful; the spread of the Tele tribes in the initial period of written history contradicts these premises. Maodun Shanyu united the autonomous tribes across a large territory, but from the prior Chinese annals it is known that the tribes were decentralized politically and autonomous, not that their area was limited to certain locations known to the Chinese. An assertion that the forest foot hunters and nomadic pastoralists lived intersPersed, sharing the same territory but exploring different economical niches seems to be justified better.
The reference to the Anc. Rus. “kouna” 'marten also needs a
closer examination: the term appears as an attribute of the tax collection imposed by the
Avars and Khazars, and is likely a loanword into Early Slavic languages, adopted as a set
of related terms connected with tax collection from the foot hunter dependents; the Avars
and Khazars did not initiate the tax, they inherited it from the Huns and Bulgars when
they took political and economic control of the dependent tribes, but so did the
Huns and Bulgars, and before them Alans, Sarmatians, and Scythians; thus it is more
likely that the Slavic term ascends to the relations described by Herodotus in for the
Scythians and their dependents. The Rus annals described the procedure of the takeover in
clear terms: "Stop paying tribute to the Khazars, and pay it to us", changing only the
allegiance of the tributary tribe, but not the existing obligation).
COMMON TÜRKIC AND ANCIENT TÜRKIC EPOCH
The second level of the nodes on our two trees are quite different. In the first case
(unedited lists), the division into three branches can be defined as Yakut,
Siberian and the remainder, and the separation dates back to 160 AD. In the second case (edited
lists) the same dating appies to the division into four branches: Yakut, Sayan (Topa), Oguz in
a broad sense (with the inclusion of the ancient Türkic) and balance (conditionally it can be termed
central group) (to place Ogur in the same group with Oguz
downplayss the group that was most prominent in that historical period, and which
provides the bulk of the historical traces. This approach can't be productive). It should be noted that none of the
trees show a particular Yakut (Sakha)-Sayan unity, which is sometimes
suggested on the basis of two interesting, but
clearly not a "bound" phonetic features:
As for the more recent branching, the trees greatly differ, and for the “unedited” tree most of these differences can be is associated with the subsequent territorial contacts between the related languages (This conclusion already imbeds a prior status where the related languages cover immense territory, and extend their mutual influences across individual territories, presenting a linguistical continuum). Apparently, that can explain the Karakhanid-Türkic case, which on the “unedited” tree is separated from the ancient Türkic and is linked to Karluk-Kypchak group (which is historically correct, since Karakhanids were Karluk tribes, historically a part of of the Uigur tribes, who in turn were a division of the Tele tribes, whose Oguz language is assigned the spot of the “Ancient Türkic”; a closer examination would treat the Ogur group as a branch of the Ancient Türkic, labeling both Ogur and Oguz groups as Ancient Türkic time-wise, but genetically already as post r/z split); the case of the Turkmen-Salar unity, separated from the other Oguz languages and associated with the Kypchak group; the case of the Nogai-Kumyk and Tatar-Bashkir unities.
The third level of the “edited” tree is dated by 470 AD (a search for historical events possibly correlate this date with the migration of the Ashina clan to the southern Altai at about 460 AD, and subsequent political events). This first level includes a breakdown of "broad sense" Oguz group into the Ancient Türkic, Khalaj, and Oguz proper, secondly, the breakdown of the 'central' languages into Siberian (actually, Kyrgyz-Altai) and the Karluk-Kypchak groups (Essentially, the date of the 470 AD is a date of the Karluk-Oguz split, when the Karluks integrated with Uigurs, by 644 AD in the east Karluks are Uigur tribe Chinese annals, but the first record of the Karluks dates to 420 as a part of Ephthalites in Badakhshan province of Tokharistan. In 465 AD Karluks were a part of the Five Ogur coalition of Karluks, Kangars, Kalach/Alachin/Boma, and Kipchaks that invaded Media. The move of the Ashina tribe to the southern Altai is but one factor that may have affected Karluks. The spotty references to Karluks extend from Black Sea littoral to Mongolia, presenting a wealth of choices for glottochronological impacts. The Tokharistan and Caucasus Karluks melted with other tribes, and the Uigur Karluks left a distinct historical trace).
Then about 730 AD practically simultaneously the Oguz group split into the Western and Eastern Oguz, and the Kyrgyz group split with the Mountain Altai (the grouping happens relatively traditionally), this dating is very similar to the chronological framework in the formation of the Uigur Kaganate, and may reflect the related territorial and political developments (in particular, the demise in the operation of th the Great Silk Road; very close to that time is the first record of the codified the ancient Türkic literary language).
Apparently, approximately to that era belongs the emergence of a number of Sinicisms in the ancient Türkic, ascending to the Middle Chinese (see, for example, Shervashidze 1989), a special analysis require the following early Sinicisms that spread in the new Türkic languages. Were the borrowings in the Common Türkic in these cases (e.g., were they registered in the Yakut and/or Toba branches), into the smaller branches, or already into the ancient Türkic literary language?
Middle Chinese and Türkic
1. *Jaŋ 'image' (ESTYA 1989, 121-122), recorded from Anc. Uig., no Chuv., Yak. and Tuv. with the semantics of 'character', which may in this instance point to Mongolism. All forms of Siberian and Kirg. languages can also be explained as Mongolisms 239 (with a characteristic development of meanings 'custom, law, character'; later Mongolism with ½- are also in Kirg., Kaz., KKalp., Bashk.). Khalaj yang 'good practice' from Persian, about which see Doerf. 4, ¹ 1903. Thus, the reflex of the primary Chinese borrowing occurs only in the literary works. More widespread is a common derivative *jaŋ-lyg 'like' (which turns into a postposition "how"): monuments and languages of Middle Asia (Turk., Uzb., Nuig, Sug., KKalp., Kirg.), which likely received it through the Middle Asian literary languages. The source: Ch. 样 Modern yang4, Middle Ch. uaŋ, Anc. Ch. lags 'form, shape'. The corresponding development of the initial is dated by about 7th c. AD. Thus the borrowing could enter the ancient Türkic literary language at about 8th century (Since there are no earlier records, the direction of the borrowing and the timing are subjective. Could a Türkic word enter the Yin languages, and then be inherited by Chinese along with other technical and etiological lexicon? Probably that is not a case, but such possibility is not being refuted).
239 The Mongolian likely borrowed the word from the Turk. (interpretation of the initial
j- as a fricative).
2. *Jaŋ-gan 'elephant' (ESTYA 1989, 60; SIGTYA 1997, 156, recorded from the Anc.
Uig. jaŋa, then in the literary monuments, also exist Alt. jan (Verb. 64); Sug.
jaγan, jaŋan, Tuv. žan). All Siberian forms can be Mongolisms (Written Mong.
žaγan, from Turk.). The source: Ch. 象 Modern Ch. xiang 4, Middle Ch. zjaŋ, Anc. Ch. lhaŋ? (from Shih-ching 47.1, 47.2) 'elephant, ivory'. Phonetical dating: change of the initial lh to affricate points to the Eastern Han, i.e. borrowing took place not earlier than the 3rd c. AD, in the monuments and Mongolian borrowing was apparently reflected contaminated form with affixoids
(affixes that originally came from independent words) that already appeared
in the literary language.
240 The metric length of the units of the foot type is about 30 cm;
the metric length of the elbow in the Hinchzhou river basin is 45 cm
The Middle Iranian borrowings that fall to the era under discussion are associated with the Eastern Iranian languages, first of all it is Sogdian as a language of the people who were very actively engaged in trade and missionary work in those territories. The incipience of the Sogdian trading system dates back to the 2nd c. BC, the heyday of the Sogdian trading empire lasted from the end of 4th to the 7th century AD, at that time it was a dominanting force in the dissemination of the Chinese and Indian products across the whole of the Middle and Central Asia. Interestingly, according to historians, the main driving force of the Great Silk Road were then not just Sogdian, and mixed media Sogdian-Türkic culture often come from mixed families. By the 10 th c. the Sogdians were mostly displaced from the trade routes (see review by M.D.Bukharin of the book: E. de la Vaissiere. Histoire des marchands sogdiens. Paris. 2002. VDI ¹ 2, 2004, pp. 224-228).
Sogdian and Türkic
a) "literary" borrowings:
1. *Axšam 'evening': Krh., Uig. axšam MK in the new languages - Oguz, Karluk, Krh.,
Kum., KBalk., Kaz., KKalp., Kirg., i.e., the "south Muslim" zone, in the Tat.
ahsham 'prayer performed after sunset', likely from Turki (see ESTYA 1974, 207; contrary to EDT
96, not in the NE languages): the source is non-witnessed Sogd. *axšam, restored
from the Middle Pers. šām 'evening; dinner' < Late Av. xsafnya-, compare Sogd. axšap 'night' Gharib 2093 < Av. a-xšapan Bartholomae 548. From the Turk. it was borrowed to the Mong. asqun, asqan KWb 16, with a zero mark at Ramstedt, which means that Written Mong. form is not really
documented and is restored from the Kalm. asxən, most likely a new borrowing from
the neighboring Kypchak languages.
4. *Tamu-g 'hell': with Anc. Uig.. Man., Budd. tamu, Krh.-Uig. tamu MK, QB, form
tamug first appears in the QB verse: iki ev jaratty bu xalqqa qamuγ / biri aty uemaq biriniŋ
/ tamuγ (with transfer of morphological forms) 'he created two abodes for all these
people, one is named paradise, the other is named hell'; here syntactically possible is an accusative
definite case with -g, which moreover rhymes with the previous line. In Tfs already
is Dat. tamug-ka, the same is Chag.; Old Osm. tamu can phonetically be a reflex of tamug. In the new languages: *tamu: Alt., Tel., Shor. (P), Tuv. tamy; *tamug: Tat. tamug (an obvious
borrowing from Turki), Bashk. tamuq (also from Turki), Kaz. (P) tamyq (from Turki),
Uzbek, Uig. tamuγ. Chuv. tamúk < Kypch. < Turki. Tur. tamu, Turk. tamu equally probably can ascend to both forms. TMN 2, 568-569, Fedotov, 2, 169-170, VEWT 460, EDT 503.
Chuv. > Mar. tamyk. Anc. Uig. borrowing. into Mong. tamu. Turk. Most likely it is a loanword from Iran., likely Sogd. tam (the name with i-base, in the accusative case tmw)
Gharib ¹ 9588 (cf. also the Khorezm tam < Avest. tVmah), derived adjectives ('hell'): ¹ 9591
tamik, ¹ 9604 tamik, ¹ 9605 tamenu. Apparently, a form of accusative case served as a source
for *tamu, and one of the λ-derivatives served as a source for *tamug,
should be agreed with the Clauson suggestion about the borrowing of different forms into
the various Türkic dialects. In addition, all the Siberian
forms can be Mongolisms (Since the origin and dating of the
Avesta has not been settled, the underlying premise of Avesta's preeminence remains speculative).
9. Krh.-Uig. erej 'bliss, happiness' KB: Clauson (EDT 200) connects it with a number of
forms in Türkic medieval monuments and modern Türkic languages. But only the forms of
can ascend to the Sogd. ryz/ž 'want, desire, demand' Gharib ¹ 8673, 8676, with
less then ideal correlation of the meanings. The other forms associated with this EDT
article apparently have different origin: MK, Tfs erine 'happiness, pleasure,
twist of fate', as suggested Räsänen (VEWT 47), may be associated (as a normal verbal nouns) with Mong. *åï-
seek, aspire, want, ask' (BAMRS 4.437); found in a number of modern languages reflexes *yrys 'happiness, good luck',
most likely are linked with the
Türk. verbal root
*yr-, see ESTYA 1974, 666 -667, EDAL.
17. Anc.Uig. Man. and Budd. čaxš'apet 'commander' EDT 412 < Sogd.
čyšpδ, = Parth. čγš'byδ of Skr. siksapada Gharib ¹ 3315
(The čaxš'apet ~ chaxshapet is suspiciously close to
the common name for the “ten arrows” ~ shadapyt in the compound onŋ-shadapyt,
in Ch. Nu-shibi, which is known from the reconstructed Chinese phonetics).
The referenced to the vague "residual" Türkic languages,
undefined layers of the Persian, Sogdian sibling, and allusion to the undefined “really
early” ancient period leave only a single fact clear: that the Türkic and Sogdian both
shared this word, which was the objective of the analysis. The etymology of the Persian,
Sogdian, and Türkic adaptations is left out from the analysis).
b) "Casual and economic" borrowings::
22. OT *borè '' see ESTYA 1978, 196-197 (Middle Kypch. - KW, At-tuhfa, Middle Uig. - Bor.LT, IbnM,
Middle Oguz - Mel.AF). SIGTYA 1997, 338. Chuv. dial.
purúš 'duty' (Ashm. IX 305) not necessarily indicate the Pra-Türk. nature of the word, cf. Tat. burych, Bashk. burys, conversion of which can be Chuv. form. Other occurrences - all Oguz, Kypch., Karluk languages. Absent in Siberia. < = Sogd. pwrc < *partu-č < *prti 'loan' Gharib 8202,
cf. Avesta pVrVθa- 'Strafe' ('punishment',
seems quite remote semantically)) Barth. 892. Seemingly, this is only way for
etymologizing this Türkic word (Seemingly, since debts are
as old as the humanity, this is a stand-alone shared Sogdian/Türkic word).
23. Krh. maraz 'employee' hapax MK EDT 772, from Sogd. maraz Gharib ¹ 5418 the same.
The Türk. word from the East Iranian, represented by the Sogdian kne 'city' Gharib
4770 (Yagn. kant), HSak. (Hotan oasis) kantha- 'city' Bailey, 51 Osset.
kaent 'building' Ab. 1, 579, Pashto kandai 'quarter' (with dimunitive suffix). Other Iranian forms (Middle Pers.
kand as part of place names, Eastern
New Pers. kand 'village', Baluchi kant toponym, also Anc. Indian. (Panini) kantha-)
probably should be regarded as the East-Iranian borrowings, by the cultural and historical
reasons most likely Sogdian. Apparently, the Sogdian origin should be taken for Türk.
word (the Saka borrowing into the Türk. languages are structured differently, see below)
(The primacy of the Persian language has a historiographical
past, it was a first subject investigated by the Western European linguists in search for
their family tree and Urheimat. With time, many of their premises were disclaimed by
further studies. In respect to the Sogdian people and language, the study of the Amudarya
basin uncovered initial Uralic-type population predating 2000 BC, an arid depopulated
period extending from 2000 BC to 800 BC, and re-population of the area starting at ca 800
BC by two groups of the Timber Grave Kurgan Culture, one from the west and another from
the east. Both Timber Grave groups displayed "massive build" distinct from the gracile
features of the Mediterranean type, but the eastern retro-migrants brought along a
considerably heavier Mongoloid admixture then the western group. These were the people
that re-populated settlements and cities in the Amudarya basin; they created a symbiotic
society with the migrants from Bactria and Northern India. Persia, when it came about a
few centuries later, was separated from the Middle Asia by the Bactrian belt.
Accordingly, the Persian influences could impact the Middle Asia only sporadically and
far between, the language of the settlements could originate in the Bactria and Northern
India, provided a linguistical assimilation of the settled Timber Grave nomads. With that
background, an ancient Panini Indian word appears to be a suitable base for Sogdian/Türkic lexical development that produced all versions of Sogdian/Türkic kant/kent, and eventually was adopted into the Middle Persian,
Eastern New Persian, etc., and became an international word within the Middle Asia. The
"Ossetic" word belongs to the same circle, it is a Digorian/Tochar/Tuhsi word from the
Amudarya basin, it belonged to the Ogur family, and must have been shared by the Huns,
both Eastern and Western, Bulgars, Kangars, Uigurs, Usuns/Uisyns, Saka, Ephtalites, etc.,
i.e. all Türkic-speaking people of the Middle Asia. The compound Bactrian/Indian/Türkic
language extended into the Taklamakan and Tarim basin, bringing the kant/kent to its
oases, including the Khotan oasis. In the cities and villages the predominant language
apparently was Indian/Bactrian based, in the steppes apparently the Türkic-based was a
predominant language, but both were compound languages with intensive mutual influences.
The extent of mutual penetration was illustrated by Biruni, who described the Alan
language as half-Sogdian, half-Badjanak. On the other side, the first inscriptions of the
Türkic Kaganate were written in Sogdian, which apparently was not a mystery for the
241 This word (not preserved beyond the medieval Türkic literary languages) usually
is also attributed to a borrowing from Middle Pers. (see EDT 354) bor, which, however,
is known only with a meaning of "reddish", see CPD 19. Is it really a "Bulgar" parallel to
the Comm. Türk. *böza (ESGYA 1978,173-175)? (This appear to be quite a word. From Türkic word for "bagasse" böza, Tat., Tobol. boza, Chuv. peraqa = > New Greek τειπθρο and Hung sopro "bagasse", and Türkic bor "wine" = > Türkic baγ "vineyard" = > Türkic baγ "garden" = > Türkic baγcha "small garden" '= > Türkic baγcha "melon and watermelon garden", Chuv. pax "garden" and Sogd., Pehlevi, Middle Pers. and New Pers. baγ" garden "[Vasmer, 4, 315]).
28. Krh. baδič 'post for grape vine' hapax MK, EDT 300, cf. NPers. wayij, Taj.
wo'iš (these forms are clearly Sogdian borrowings, Sogd. word must be derived from the Av ascending to.
vaetay- 'vine' Barth. 1314). Sogd. form at Gharib not recorded, but it would be a most
likely source of the Türkic word from the phonetic point of view.
As can be seen from the above selection, the Sogdian borrowing into the Türkic languages are either borrowing into the ancient literary language, from where with the literary tradition they could spread to other languages, or local borrowings into the language spoken in the East Turkestan and usually designated as Karakhanid Türkic, or borrowings into the "Pra-central" language (thus, before the 479 AD; such can be considered the "nightingale"), or into the Pra-central and Pra-Oguz (as both at that time were in direct contact with the Sogdian trading empire; such can be considered "chigir", possibly "garden", "city", "duty," "paradise", "hell", "gift", "evening"). The loanwords cover the areas of agriculture, trade and exchange relations, and religious concepts (To put things to scale, the "Sogdian trading empire" by the 479 AD was already part and parcel of the Türkic states for 679 years, it was run by 34 consecutive generations of the Sogdian/Türkic scions, and the 479 AD was not an end of it: the Türkic embassy to the Byzantine court in the 552 AD was headed by the Sogdian Maniach who advocated a joint effort to circumvent the Persian infringements. The "Pra-central" language in this case is a euphemism for the Hun language, which was a Sogdian-impacted language otherwise called Ogur language, an offspring of the Timber Grave Tele language called Oguz).
Middle Persian and Türkic
Middle Persian is dated as existing from the 4th c. BC to the 7-8th cc. AD; but the actually representative texts, except for skimpy coin legends of the 2nd c. BC, chronologically correlate with the period of the Sasanid state (224-661 AD). Middle Persian was the liturgical language of the Eastern Manichaean church, which largely determines the nature of the borrowings.
a) literary borrowings (primarily from the Middle Pers. to Manichaean Uigur):
1. Runic (from the Tonyukuk inscription), Uig. Man. and Budd. betkeei 'scribe' EDT 304.
Ultimately from the Syr. petga 'tablet', from Greek πιθακιον. The word
is not found in the Middle Pers. and Sogd., but from the cultural and historical reasons most likely
was borrowed into the Türk. from the Middle Pers.
5. *Tan 'body': in this phonetic form it is in Anc. Türkic from Yrk Bitig, MK, Tfs,
Middle Kypch. In the new languages it does not occur. EDT 510 ("unusually early Iranian
borrowing"). New Pers. borrowing: Horezm-Türk. ten, Tur., Gag., Turk., Krh. ten, Tat., Bashk., KKalp., Kaz. tän, Kirg. ten, Uzbek tan [tän], Uig. tän. VEWT 473. The source of the Anc. Türkic (note that only the runic writing in Yrk Bitig provides the reading of the rear raw a, and not e) is attributed to the Middle Pers. tan, also compare Sogd. tan- in tan-par 'body' and other derivatives (Gharib 388) Saka (Hotan) ttani can not be a source with its meaning ('skin'). See Ab. 3, 261, alternative Bailey 122.
b) Commercial Borrowings:s:
1. *Bareun: Anc. Uig. Budd., bareyn, Krh.-Uig. bareyn KB, IbnM 'silk brocade'; further
Horezm-Türk. bareyn (Nahj al-Faradis) Kypch. Dictionary bareyn (Abû H.)
the same., see
EDT 357-358 (with obviously improbable Tocharian
(Kuchean) etymology), borrowed (based on the time
and place, and as for the Hungarian, also phonetically from the Danube Bulgar) to Old
Hungarian bārsony. In the new languages: Chuv. porzyn 'silk cloth',
from Chuv. (Late Itil Bulgar?) loanword to the Russian bursa 'Persian silk fabric'
Vasmer 1, 208-209, and to the Volga Finno-Ugric languages: Mari porsyn, parsyn, Udm.
burtchin and others (Fedotov 1, 447); from Middle Pers. abrešom ['plyš(w)m, Man.
'bryšwn] 'silk' CPD 4 Pers. abrešum < *abi-raiš- 'spin',
see Ab. I 132, Tsabolov 15-16 (be alert on Abaev's
uncorroborated "reconstructions", citation references are not corroborations);
the Pers. word was borrowed with the same meaning 'silk' into many Iranian languages. A New Pers.
word was borrowed into the Tur. ibiršim, NUig. (R: Taranchi) äbrišin 'silk (adj.)',
Middle Persian berišem. This word can formally be Pra-Türkic borrowing associated with the early stages of the Silk Road - in fact, already during Sassanids (from 224 AD), Persia was receiving an abundance of raw
silk, in the Sasanian Empire worked many manufactories for its processing, and already during Augustus
(27 BC - 14 AD) the silk, apparently through the Parthians, was reaching Rome
(In a chronological discourse, the silk and its name reached Rome in Parthian
transmission, and 2 centuries later Sasanids took over the trade, came up with Abaev's
reconstructed Persian name, and endowed their Türkic/Sogdian suppliers, but not the Latin
traders, with new shiny and sticky Persian terminology). Another
possibility is to assume a certain layer of "bazaar" borrowing into the early Bulgar (with phonetic adaptation) from the Middle Asian Türkic languages
(The most ancient mentioning of the Caspian Huns is from the middle of the 2nd c. AD by the ancient
writer Dionisius Periegetes (wrote 117-138 ) in poetic composition "Description of the
inhabited Earth" (Unns) living at the northwestern side of the Caspian Sea. His data
about location of tribes is authentic and agrees with ancient eastern data. Claudius
Ptolemy (160-180) noted Huns at the time of the Roman emperor Markus Aurelius. This data
agrees with the archeological data (L. Yablonsky) about a new wave of the Huns in the
Aral-Caspian basin that coincides with the recent defeat of the Northern Huns in the Middle
Asia. The Huns did not let the control of the Silk Road out of their hands, and the
spread of the word "bareyn" to the Middle Persian, Finno = Ugrian, and Slavic languages
should be attributed to the Huns, as asserted by A.V.Dubo. A.V.Dubo prudently does not
comment on the phonetical feasibility of the fanciful *abi-raiš = > abrešum
= > berišem = > *bareun = > bareyn. She also does not trace the Ahaemenid vocabulary
for the silk, which the Parthians inherited along with the Silk Road, with or without
introduction of their own lexicon; a peak into Parthian/Dahae/Tokhar/Tuhsi/Digor
terminology may also be pertinent and enlightening. The Ahaemenids used Semitic Assyrian
language for their bookkeeping, and the Silk Road taxation was an official business,
their terminology is also an indelible part of the objective bareyn etymology).
New Persian and Early Türkic
The beginning of the New Persian language is dated by the 9th c. AD. Here are examined such outwardly New Persian borrowings that are recorded in the monuments of the ancient period, i.e. not later than the Karakhanid period. It can be seen that all these loans are either narrowly localized, or entered already separated languages (correspondingly, the reflexes generally do not allow to restore the systemic prototypes, and the forms with asterisk notation are purely provisionary). Brings attention a strictly economic thematics of the borrowings.
1. *Akur 'stables, stall'. Krh.Uig. aqur 'stable' (MK, KB); EDT 1989, IM
ahur see Ras.VEWT 10a, where the word and its relevance traced back to Persian sources, see more
Bud. I 19, Ρ I 133, but OTD 49 < = ar.). SIGTYA 1997, 527; Gag., Osm., Az. axur,
Osm., Krh. axyr, Kum. axur, Eastern Türk. aγur, Taranchi
oqur VEWT 10. Pahlavi āxwarr CPD 14, Sogd. 'xwyr
= āxwer = > 'γwyr Gharib 2130 'stable'; NPers. āxur 'cowbarn'. Rather, the word
was borrowed from New Pers. in an already divided languages of the "southern Muslim" zone
(That loanword from the New Persian is historical impossibility, since the Kangars are known from the Herodotus' time as serving in the Persian currier service, Herodotus rendered their name
as Angareion; Scythians served in Persian army; both had to introduce their horse
terminology to the Ahaemenids and later to the Sasanid Persians; The nomadic
Parthians had to use their own horse terminology; another record long before the
incipience of the New Persian is the Greek myth about the Augean "stables", where "Aug" is clearly connected with the aqur for "stables" independently of the Sasanid or New Persian, and is at least a
millennium older then the Pehlevi record. The horse culture and the word has a clear
direction Türkic = > Persian, Türkic = > Greek and Türkic = >
Armenian. The linguistical localization has to do with the ancient Kangar and its sphere of cultural influence
rather then with the "South-Muslim Zone", and the period is more then a millennium
5. Late Ancient Uig. tana 'coriander seed', Krh.-Uig. tana id. MK, Middle Kypch.
EDT 515. From Pers. dāna. Compare Middle Pers. dānag 'seed corn' CPD 24.
(The smattering of New Persian correspondences into the Early Türkic, which starts with the Hunnic disPersion and ends with Karakhanid Uigur and Karluk, appear practically non-existent, a significant proportion turned out to be Pra-Türkic lexicon borrowed into the New Persian, and not from the New Persian as claimed; a few suspected cases only state the existence of correspondences and are left hanging without etymological analysis to demonstrate the direction)
New Indian languages and the Early Türkicic
Here we also have chosen such potential New Indian borrowing that are recorded in
the monuments of the Ancient Türkic period, i.e. not later then the Karakhanid period.
The areas of the thematic and
geographical distribution of the lexemes clearly indicate that we are dealing with a "bazaar"
borrowing associated with the territories encompassed by the Silk Road. The presence in
large volume of the Indian colonies and trading posts along the Silk Road was already
noted in the 2nd c. AD (East
Turkestan 1992, p.78); however, the surviving written records from these territories are either in
Sanskrit, or in the north-western Prakrit (Niyya or Gandhari; written in Kharoshti).
The start of the New Indian period is dated by the the 10th, at the earliest 9th c. AD (Masica
CP The Indo-Aryan languages. Cambridge, 1991. pp. 50-55, 472). However, this dating is
unlikely to be accurate, since the early period of Indian languages is very poorly
documented. Given the "marketplace" specifics, can be expected borrowings into the
Türkic languages originated from different New Indian dialects, but probably more frequent
would be the borrowings of the north-western type (i.e. west of
upper Indus, northern Afghanistan and Pakistan)n).
1. Krh.-Uig. čit 'embroidered silk' EDT 402, see TMN 3, 129, VEWT 103 (represented in
the Oguz (Turkish čit,-ti, Az. čit, Turk. čit, showing the original form of shortness of
a vowel), Kypchak, Karluk languages; absent in Siberia: Shor. shyt form 'thick (of
cloth)' is another word, ascending to Pra-Türk. *čyt 'network', see VEWT 112) As a
source of borrowing is cited Persian. čit 'chintz, cotton printed fabrics', borrowed from
the Ind. ultimately from Skr. citra 'motley' (cf. Avest. čithra) (see Ab. 1, 312,
Tsabolov 1, 241). Given the purely "commercial" nature of the borrowing, it can be
that the earliest Central Asian Türkic languages could receive it directly from the
New Indian (like the NPers., judging by the development of the combination -tr-), compare Hindi
Punjabi čit 'calico, muslin' and the like, Turner 261, which facilitates the interpretation
of the short vowel (the Persian length is usually well preserved in the Oguz borrowings)
(Though cultural re-importations are ubiquitous courses of development, the pra-Türkic
etymology "mesh, meshwork" indicates that it was a base for the consequent applications
for Skr., Av. "motley" and later derivatives. As a minimum, it is an equal
contender with the Skr. citra "motley").
3. Krh.-Uig. tarmaz, turmuz 'cucumber, cuke' MK. Clauson 550: "Unless this is a l.-w.
(loan-word) cognate to 2 turma it is prob. the Neg. Aor.Participle of
tar-, lit. 'Not
branching'." In reality, tarmaz is overt borrowing, cf. Pers. tarbuz 'watermelon, melon'
Stebl.-Kam. 79-80, ultimately from the Skr. trapusa, trapuşa 'cucumber,
trpra- 'salzig, scharf' (salty, spicy) (Iran. *trfra-
with Indo-European etymology, see Mayr. EWAIA I 665,
675). The Türkic word is semantically closer to Skr. than to the Persian; on the
other hand, the later forms given by Turner for the Skr. word look like Prakrit taüsa- n., Marathi
tavse n. 'Cucumis
sativus' Turner 5993, so can be assumed a "bazaar" New Indian borrowing from
some north-western languages with relaxation of the combination tr-of the type lahnda.
The variability of the vocalization and nasalization of the labial consonant allow to draw attention to one more
similar Indian name of
the melon-type plant: Skr. tumba- m. 'pumpkin Lagenaria vulgaris', with descendants
in the New
Indian languages like Hindi tomrā m. 'dried pumpkin' Turner 5868 (may be contamination of the two
loanwords?). In the new Türkic languages is seen only Kirg. darbyz; Uzbek tarvuz, Uig. ta(r)vuz 'watermelon', which are
borrowed either from the New Pers. or from Mong. 'watermelon' (Khalkha tarvas, Kalm.
tarvas [tarvys]), borrowed from Persian, see SIGTYA 1997, 138.
7. *Sart 'merchant, trader': Anc. Uig.
Man. sart (EDT 846), Krh.-Uig. sart (MK, QB);
Chag. sart 'Persian city-dweller not speaking Türkic' (Abush.) Middle Kipchak.
'city-dweller, commoners'; Tur. dial. sart (DD III 1192), Turk. dial. sart (TDGDS
156). Räs.VEWT 405; Meng. TLP 172; Áàðò. II, 4.2, 527-529; SIGTYA 2001, 336. The word, of
course, is not Common Türkic. Adapted ultimately from Sanskrit sārtha, compare Sogd.
8726 'caravan', Parth. sārt same. Apparently, it really got into the Turk. languages through the Middle Iranian (simultaneously,
judging by the uniform development of the meaning,
and not from the Middle and New Indian sources, judging by the reflex of the combination
-rt-), as a wandering word at the Central Asian markets.
Interlinguistic contacts in the early medieval Eastern Turkestan (East Turkestan, or Uiguristan)n)
Finally, we find fairly interesting group of borrowings, which characterizes a geographically narrowly the area of Eastern Turkestan and adjoining areas, namely, these are Türkic words recorded in the Middle Ages in the Central Asia, hapaxes of Mahmud Kashgari or late Ancient Uigur monuments without Türkic etymology or reflexes in the living Türkic languages. Such words might turn out to be Iranian or Tocharian (Kuchean) loanwords 242. Indeed, among them are the following categories.
Saka (Hotan oasis) 24343 = > Türkic
1. abamu Anc. Uig. 'infinite' (Cl. 12: "no doubt a l.-w.
(loan-word) (Indo-European?)") - HSak. (Hotan oasis) avama 'immeasurable' ( < pa-mā- 'measured' < *pati-mā- Bailey, 9 213, with the negation a-, not from Sogd. (*ptm't) and not from the Parth. (pdm't); from all possible sources Middle Iranian only Hotan-Sakas loses -t- in preverb .
242 In addition, among Mahmud Kashgari hapaxes are many specific Sinicisms, but
they are not reviewed here.
9. čür MK 'success, profit', čürle- 'benefit from smb.',
čürlet- 'make smb. steal from smb. still his property', čürlen-
'benefit on smb. account'
(Cl. 428:?) - HSak. (Hotan oasis) ssuru, suru 'goodies' (Acc.Sg. from
särā, sirā 'goodies, well-being') Bailey 400-402; NB
indirect base as a source of borrowing.ng.<10. čer MK 'al-waqt (time)' (in the example:
"Come at this time", i.e. "now") (Cl. 427:?) - HSak. (Hotan oasis)
tcira 'once', draicira 'triguptam' 'three times'
< *skar- 'cut' Bailey 140.
19. (?) (Known HSak. (Hotan oasis) borrowing): beekem
MK 'pennant of a silk piece or a wild bull tail' - HSak. (Hotan oasis)
baicakama 'yak or horse tail', Wakhan. bièkam id. (Cl. 295, Mgst.
IIFL 2516, Bailey 302, baica < *vaj- 'to weave', kama < *kahama 'hair', other Iranian
parallels are there; compare hypothesis of borrowing in the opposite direction Stebl.-Kam.
Wakh . 96 (That all Türkic nomadic troops fought under horse tail standard is known from
the European, Asian, and East Asian sources, the idea of a foreign borrowing is not
realistic; it is the other way around, the descendents of Türkic standards are still used
in military and heraldic fields in Europe and elsewhere;
Wakhans claim to be Radjputs who claim to be Ephtalites who were
an offshoot of the Huns, and descendents of the Kushans with a center in the Wakh yabgu;
on top of that the Wakhs live interspersed with Tajik
and Kirgiz people; and the horse tail standard is as endemic to Iranians as the horns to
the horses. If there is a remote possibility for an Iranian name for this battle
standard, it can only be a derisive moniker, like a "rag").t>.<20. zaranza 'saffron' MK (Cl. 989: "no doubt a
(loan-word)"),? Sogdian"). But compare Bailey 347
ysaramjsa 'saffron' < *zaranèi, compare Pers.
zarang, zirik 'carthamus tinctorius, from which is extracted a red dye." From HSak. (Hotan oasis)
ysar- 'be red', HSak. (Hotan oasis) word phonetically
corresponds exactly to the Türkic (ys = [z]). (The Sogdian lexical parallels
are absent; compare
Sogd. usparn 'saffron' Gharib 10,586).
244 Stebl.-Kam. NKR 53: "... the Pamir forms could not be borrowed from the
Türkic languages (which have forms with the second a short). The first part of
the statement is not disputed, because the early borrowing is justified by the Pamir palatalization, but the Turk.
necessarily require a second short vowel in the source form, since it can be supported by
the intra-Türkic synharmonism with the frontal raw with a vowel of the first
syllable, ascending to the a short.
25. šatu 'ladder': Anc. Uig. Man., Krh.-Uig. šatu (QB), Chat. šatu Sangl., Anc. Uig. > Relatively late inscribed Mong. šatu(n) 'ladder' (Doerfer TMN 3, 317): Written Mong. šatu(n) (L 754), Middle Mong. šatu (MA 407), Khalkha šat(an), Bur. šata, Kalm. šatə, Ordos šatu, Shira Yuig. šatə, KWb 351, MGCD 710 245. In the new Türkic languages it is represented by Tuv. čada ( < Buryat) NUig. šaty/u, which can be both inherited and Mongolism.sm.Like presumes EDT 867, the word apparently was borrowed from the Moddle Iran. form of the eastern type *šatu < *sritā (*sr- > *s in HSak. (Hotan oasis), Sogd., Pashto, > s- in Osset., u from the accusative case? compare HSak. (Hotan oasis), Sogd.). Traces of the borrowing source in Iranian languages: late Saka (Hotan oasis) šakye from *satV-ka Bailey 406, 511 as a gloss to Türk. catta [čatu, šatu]; in the "Zambasta Book" a prefix formation büşşäta f. < *abi-sritā. Compare Pashto šəl < sritā 'ladder', Yazg. xad < sritā, Osset. aşinä 'ladder' < *ā-srin-, ses 'wall' < *sraiša, Yidgha afsinγo 'ladder' < *abi-śrinaka. Bailey 300, Abaev I 76, 3 113-114.
Two southeastern Türkic words look like borrowings from some Eastern Iranian languages,
but they do not coincide with any recorded Middle Iranian language:ge:<1. üjme 'mulberry' MK (= New Uig.
üjme) (Cl. 27: "the j suggests that it is an Iranian
(loan-word)") = Sarikol yjma, ůjma Stebl.-Kam.
NKP 90 < *aizma-.
Tocharian = > Türkic (i.e. Kuchean oasis ~ Tokhar A and Turfanian/Arsi oasis ~ Tokhar B; no relation to the Tokhars whatsoever)
1. 1. oxaq MK 'juice of crushed apricots, used as a cooler drink' (Cl. 83: "prob. an
Iranian l.-w. (loan-word)") - apparently from Tokh. A (Kuchean),
oko 'fruit' Adams ¹ 656 < PIE *og-;
245 Contrary to EDAL, the Mong. forms should be excluded from
the Altaic etymology EDAL 2024: PAlt.
*sit'o 'ladder, grid wall': Mong. *sita, TM *sitki 'wall of the tent' SSTMYA 2, 99, Kor. *satari 'ladder' Nam 282, KED 870,
*sitəmi 'awning, blinds' JLTT 528, see Ram.SKE
217. The Mong. word falls out for two reasons: 1) looks rather strained the presumption
about Mong. borrowing in the ancient Uigur Manichaean text, 2) reconstruction *sita
for suggested Mongolian forms is phonetically impossible, and in general their reconstruction is
difficult: the reconstructed Proto-forms should have reasons for a-break in the first
syllable, and simultaneous for the short u of the second syllable (sufficiently well attested forms
from the MA and
Ordos); *situa would have given *šoto (cf. *činua 'wolf'), *sitaγu would
have given *šutuu (cf. *sibaγun 'bird').
Well-known Tocharian borrowings (i.e Kuchean oasis ~ Tokhar A and Turfanian/Arsi oasis ~ Tokhar B; no relation whatsoever to the Tokhars, Tuhsi, Bactrians, or their languages):
4. lešp 'snot': Anc. Uig. (Medical texts), MK
(Cl. 764: < Tokh. A (Kuchean), B
(Tarim) leüp same Adams 3046:
< Prakrit) (i.e. Prakit = >
Kucha and Turfan = > Anc. Uigur);
Interestingly, the contacts between Sinkiang (East Turkestan, i.e. Uiguristan) Türkic, Hotan-Saka (Hotan oasis) and Tocharian "worked" in all directions, as evidenced by the presence of other borrowing directions.
Türkic => Saka (Hotanian) (we cite only tentative borrowings not noted in Bailey 246)
1. (?) HSak. (Hotan oasis) chā 'length measure' Bailey 107 (compared with HSak. (Hotan oasis) chei 'branch' < *chaša-), in reality (for example) = Ch. tşi 'foot' - compare Mongor. DZā 'gap between thumb and middle finger', i.e. “open palm span”, an equivalent of a foot in the Chinese system of the length measures, from the Ch. (Synin) tsa, Ch. zhā 'palm span' SM 69, Ch-DWb II1262, ¹ 7477 (the sources for the Middle Ch. words were not found). The etymology can be justified with a detection of an intermediary source between the Chinese borrowing and HSak. (Hotan oasis) form. Or the direction of borrowing is opposite (from Iran. into Ch.)?
246 A number of Türkic loanwords into Saka
(Hotanian) is well known; mostly they are proper names, titles and ranks, names
for clothes, leather products, food and the like. Also compare Hovdhaugen, pp. 166, where
is alleged that the Türkic words in Hotan-Saka (Hotanian) are not borrowings,
but occasional words in connection with ostensibly weak interest of the Sakas to the contact with the Türks,
at the same time in the article are absent a number of well-known "household" borrowing from
into the Saka (Hotanian) like the yaragaka ( <
Türk. yargak 'leather'), yadama ( < Türk. yalma 'cloak') Bailey 258,
and others (Anthropological investigation of the Tarim basin
cemeteries shows biological admixture of the nomadic and sedentary people consistent with
the cultural symbiosis and genetical studies. Per Chinese annals, the very tribe named
Türk, and its dynastic clan Ashina, was a branch of the Saka tribe. If there were any
resident Sakas in the Hotan and environs, they were kinfolks and tribesmen of the Türks;
however, the Türkic admixture with the Hotan agriculturists most likely is of the Uigur,
and not Türk descent).
2. (?) HSak. (Hotan oasis) kaşa
'belt' Bailey 56 - Bailey attributes to the Iranian *kaša-. The full etymology of this
is as follows: Av. (j) kaša- m. 'armpit' Bartholomae 461; north-west. Gilan.
'hug', Lur. bunkash, kashbun 'hug' Jukovky II133, Talysh kəsh 'wing,
arm, sleeve', kəshə
'breast', keshəbyn 'armpit', Kurd, Kurmanji k'osh f. 'front of man from knees to the
waist' Bakaev, koš 'lap (skirt)' Kurdoev, Sorani koš 'knees, groin',
South Middle Pers. kaš, keš
'side, armpit' Abrahamian 169, Zor. p. dast-kaš 'arms under armpits' NPers.
'groin, armpit, chest, heart, hug', Taj. kash 'armpit'; Eastern HSak. (Hotan oasis)
kasa 'belt' Wakhi kal 'embrace', kalbən 'side, bosom, armpit' Pahalina Wakhi 210, Gr. -
Stebl.-Kam. 366, Mgst. IFL Π 525,247, Sogd. p-kšy 'side', Yagnob. kapaš, kepaš < *pəkaš <
*upa-kaša 'bosom' Andreev-Peschereva 270; Pashtu kše 'in' Mgst. EPsh 72,
Yidgha-Mungi avγuš 'bosom, hugs' Mgst. IFL II 144, Sanglechi-Ishkashim
kašviš 'armpit' Mgst. IFL II
400 ( < Pers.) Shugn(i) bijuγ 'armpit' < *api-kaša Zarubin 112, Sarikol
buxčo 'bosom' Pahalina Sar. 29, Bartang bijaw, bixčaw 'armpit,
bosom' Sokolova Bart. 89, 90. But the
meaning in the HSak. (Hotan oasis) looks far isolated
among other Iranian meanings, so can be offered a Türkic source for HSak. (Hotan oasis)
word (like in a number of other cases with the clothing names in HSak. (Hotan oasis)): Türk. *kel-č 'belt': Anc. Uig.
keš, MK keš 'belt', Kirg. kešene 'sash'
Chuv. kazan 'back, spine' VEWT 258, Cl. 752, Doerf. II 1697, ESTYA 1997, 60-61.
The Türk. base has a suitable Altaic etymology: PAlt. *ê'¸lå 'belt, waist': TM *xelgece 'waist;
span' > Evenk eŋene, Evenk eŋún 'waist' elgú 'span',
Negid. eŋene 'waist', Ulchi. xeŋgi
'waist', Oroch. xeŋgeje 'armpit' SSTMYA 2, 446, 458, Jap. *kəsi:
Anc. Jap. kosi, Middle Jap. kosi, Modern Jap. kòshi JLTT 458
(三尺 ?All “belt” ?). See APiPYaYa 76, 289.
247 Morgensterne believes that the Wakhan form was borrowed from
undetected Sanglech, like the Pers. dial. kalk 'side under armpit'.
5. (?) HSak. (Hotan oasis)
tadrrvā Bailey 102 'krosna, backing (?)' (“krosna/êðîñíà” =
manual loom with started work) in the following context: tta bura pvaicai cu
ttadrvā bāstadu padāmysa padāya pvaica tsvā sā 'these so many
which we had drawn upon the looms(?), in the first way one pvaica-covering came' KT II 76, 3-4.
Bailey produces the HSak. (Hotan oasis) word from *tantra,
Skr. tantra 'krosna, backing', but the semantics is not too convincing. The cover
appears repeatedly in combinations such as: yaragaka να pvaica 'cover for
(where yaragaka < Türk. yargak 'hide (pelt)'), siyam pvaicām jsa jsā yadama 'yalma with white
covers' (yadama < Türk. yalma 'cloak') Bailey 258, cf. in the same
place a number of of contexts with
the Türkic borrowings. It is possible to presume that ttadrvā is Türk.
tatyrga MK1489 (hapax - C1.
460) 'white tanned leather' (Arabic al-qadam 'white parchment, leather tablecloth'),
apparently related to tart- 'stretch', cf . another derivative from the sphere of leather terminology:
'scraper for leather' SIGTYA 1997, 361. (For the phonetic form of the HSak. (Hotan oasis)
word compare HSak. (Hotan oasis) anahid ipabhutti in accordance with the
Türk. ayag-alpaγut Bailey JRAS 1939).
Tocharian => Saka (Amazingly, neither “Tokharian” is Tokharian, nor “Saka” is Saka, but these are the misleading euphemisms of the modern Indo-European “philological science”, distinct from the sciences where things are called with what they are. The right title reads “Kuchean/Tarim => Hotan”, i.e. common elements within the Tarim/Taklamakan basin)
1. Interestingly HSak. (Hotan oasis) puka
'elbow' (Bailey 242: to the Iran. *pank- 'pierce' and Greek πνγών 'elbow' - ςηεν
in fact the latter is derived from the name of the fist, indicating a measure of length
from an elbow to a palm contracted into a fist). Obviously, it is taken from Tokh. A
poke, B pokai 'forearm, elbow' (= Skr. bahu-, Av. bazu- < PIE
*bhāg'hu-), see Adams ¹ 2191.
Saka => "Tocharian (The right title reads “Hotan => Kuchean/Tarim”, i.e. common elements within the Tarim/Taklamakan basin)
The Adams dictionary contains about 100 such loans, often Indian through the Saka (Hotan oasis) (Rephrasing this to an object-oriented speech, the Indian-based lexicon of the Hotan settlers spread to the settlers in the Kucha and Tarim oases; no relation whatsoever to the Tokhars, Tuhsi, Saka, or their languages).
Türkic => Tocharian (The right title reads “Türkic => Kuchean/Tarim”)
In Tokh. B (Tarim) is known a number of the Türkic (Uigur) personal names and titles, for example, ārslam 'Arslan' Adams ¹ 9, ālp Adams ¹ 375, er Adams ¹ 582, el < Adams ¹ 600, tārhkāne Adams ¹ 1526. We provide here nominal words that can be interpreted as Türkic borrowing already into Tocharian B (Tarim), after its separation from the Tocharian A (Kuchean) (After Tarim basin oases accumulated enough changes to be different from the morphed Kucha oasis vernacular. There is no evidence that Kuchean and Arsi oases ever spoke a common language; they might be as close as a borsch and potato salad, both made with the same potatoes, but with different other ingredients and processed differently. The family tree model's "separation" may be an illusory consequence of the model that A.V. Dybo herself dismissed from the beginning).
1. Tokh. B (Tarim) iprer, eprer, A
(Kuchean) eprer 'sky, air, firmament' Pouha 1940, Adams ¹ 568 ("Further relations are unknown. Certainly not ... from PIE *per- 'point'. Other proposed
connections, a borrowing from Middle Iranian abra- 'cloud, 'or a relationship with Breton
ebr 'heaven' founder either phonologically or morphologically (whence the final -r ·?), or
both") (What a great chance to peek into the lexicon of the
Alans in Amorica, who in the 5th-7th cc intermarried extensively with the Celtic Britons
in the Orica/Amorica, and who may have introduced elements of the Timber Grave lexicon into
the Bactrian/Sogdian language a millennia prior to that). Indeed, the expected source of the Indo-European Tokh. B (Tarim)
iprer - *nbh-ro-, the
zero stage of *nebh- 'sky, cloud': Skr. abhrā-, Av. avra- 'cloud', Greek
aphrys 'foam', Lat.
imber 'rain' Pokorny 316, but a normal reflex of *n in is Tocharian - *än > A
(Kuchean) än, B (Tarim) an (Burlak 123), and both Tocharian
forms can not be linked to ascend to the PIE root; moreover, the vowel
correspondencies between the Tocharian A
(Kuchean) and B (Tarim)
forms do not allow to reconstruction of the common Pra-Tokhar form, so one of them should be
taken as borrowed (this may be a
literary borrowing from A
(Kuchean) to B (Tarim), or
a use a live form from B (Tarim) by a scribe who was creating
a text in
a dead A
(Kuchean) language). Thus, the Tocharian forms do not have Indo-European etymology, and
do not ascend to a Pra-Tocharian state. But the Türkic has *ebren 'dome,
firmament, sky' >
MK evren 'dome', QB evren 'sky, firmament', Modern Turkish evren 'sky, world' C1. 13-14
(partially contaminated, but contrary to G.Clauson unrelated to the Türk. *ebren 'snake, dragon',
which has Altaic etymology - see EDAL 1605), apparently a derivation from *ebür- 'turn', EDT 1914, VEWT
34, ESTYA 1974, 498-500, Egorov 20. The Tocharian forms can be borrowed from the Türkic
with a replacement of a suffix (And this supposition neatly
falls in with the historically documented Amorica Alans, and explains the oddity of why
of all the Celtic languages only the Amorica Britons use this Türkic word).
2. (?) Tokh. B (Tarim) kakwār 'a type
of food' Adams ¹ 743 ("Etymology unknown") = Türk. kagur- 'to grill' with derivatives of
the type *kagur-ma, *kagur-dak and the like - all
epithets for fried foods - see ESTYA 1997, 175-177, C1. 612;
6. (?) Tokh. B (Tarim) miye 'some olive-type fruit?" = dimin. miyaske (Adams ¹ 2552:?). The context
that allows to judge the meaning of the word is the following: [wsā re nek]cy[e]ne cwi
'they were giving him at night miye and paşkārro'. The paşkārro presumably
is 'linseed' (Adams ¹ 1978).
Then we can assume that the miye is also a medicinal plant, and a good comparison for
it may be
a Türkic word meaning 'liquorice; goosefoot' *bynan: Chag. bijan, CCum.
(Kuman?) bujan, Tur. bojan,
mejan, Az. bijan, Turk. bujan, Tat. dial. myja, Kaz., Nog. myja, KKalp.
bojan, Kirg. myja,
Uzbek mija, NUig. buja, Chuv. majan Egorov 131. Turk. > Kalm.
bujā 'licorice'. This word
apparently ascends to PAlt. mina 'type of herbaceous plant' together with the Kor.
*minari 'celery, parsley'.
demonstrated groups of borrowings that proceded in all directions and various semantical fields
clearly show a situation of trilingual contacts that had to go on at a level of active
daily communication, which situation can be dated by the 7th-9th centuries AD
(Realistically speaking, neither a presumption that before the 7th. c. AD,
or after the 9th. c. AD existed a meaningful period without active daily contacts on a
family, social, trade, and intra-tribal communication is amiss. The symbiosis of local
sedentary population and their nomadic neighbors extends through the records that first
mention the Tarim basin settlements, and continued till present.)
MIDDLE TÜRKIC EPOCH
This period is addressed in most general terms, since most of the main points have already been covered in the previous book of our edition, “Regional Reconstructions”. Significantly, none of the ancient glottochronological trees reflect a node that could be associated with the Karluk group of languages. Indeed, as we have attempted to demonstrate in the first section of the book, the main phonetical and morphonological isoglosses that unite the Karluk group turn out to be secondary. Most probable is an existence at some stage of the Karluk-Kypchak community coupled by several common phonetical and morphonological (morphophonemical) processes. Subsequently, after a divergence of the languages, but with preserved continuity of the dialect situation, a part of innovations impacted Oguz together with Kypchak languages, but the relevant processes can not be considered to be genetically identical.
The further development of the Karluk-Kypchak tree also shows some interesting correlations with the historical facts. For example, the two subgroups that first branched from the rest were Caucasian Kypchaks (780 AD) and Karaims (900 AD), which pretty well corresponds with the two waves of Kypchak infiltration into the Eastern Europe. The divergence of the Karaims into the Trakai-Halich (Galicia) branch in one hand, and the Crimean branch on the other hand is dated by 1220 AD, which surprisingly accurately coincides with the date of the Mongol raid into the Northern Pontic steppes (1223), when Kypchaks were defeated and fled partly to the south, and partly to the east (to the future territory of Lithuanian Rus). Apparently, with the Mongol invasion can also be connected a bit later separation of the Balkar and Kumyk Languages (1300).
following is a quotation from a discourse about the divergence of the Kypchak languages
from the previous volume of
our Grammar: "The split of the Pra-Kypchak community onto the western and eastern
branches apparently started in the pre-Mongol era, and is associated with infiltration of
the Polovechens of the Rus chronicles, or the Kumans of the Western authors, into the steppes of
Europe. The remaining in the Asian steppes Kypchaks formed a base for the Kangly subgroup,
and the part that invaded the Northern Pontic formed a base for the Kuman subgroup. However, throughout the history
of the Kypchak-speaking tribes the centrifugal divergence was constantly leveled by the centripetal convergent
phenomena. The convergence of the eastern and western Kypchaks went on most intensively in
the Kypchak Khanate period, then in the contact zone between two main branches formed a
mixed language of
the Nogai type with predominance of the western elements. The beginning of forming individual languages
within the eastern and western branches belongs to the post-Kypchak Khanate era, apparently as a starting point should
be taken the turn of the 14th - 15th cc.
That is when the rebirth of the separate state entities, the Crimea, Astrakhan, Nogai, Uzbek, Kazakh
juzes and other Khanates. Within these Khanates went on the final
regroupping of the former tribal dialects.
A detailed analysis of the ethno-and linguogenesis for each individual modern Kypchak languages uncovers extremely complex and multi-component composition of the constituent ethnic elements, however for historical reasons in the west dominated one linguistic elements, and in the east dominated other elements. For example, during a rise of the Nogai Khanate, in the ancestor of the Nogai language the initial western features were thoroughly replaced the eastern features. The ancestors of the Kazan Tatars who separated from the Nogai Horde before the rise in the language of the eastern elements (incidentally, the Kazakhs still call the Tatars Noγaj, i.e. Nogais), and the Balkars fell into the in western Kypchak milieu, but retained some features of the eastern Kypchaks (in particular, the "dj" type). A final formation of the Tatar language went on after a formation of the Kazan Khanate in the middle of the 15th century. The Bashkir language, most likely an Oguz in its base, was subjected to repeated Kypchakization: in the pre-Mongol era, during the Kypchak Khanate period, and finally, in a relatively late period from the Tatar and Kazakh languages. Thus, the final formation of the Bashkir language belongs to the era of the Kazan Khanate, or even to a later period. In the Ural subgroup of the Kypchak languages are traced features of both western and eastern Kypchak languages (SIGTYA 2002,258-259).
This quite complicated picture has really little in common with the divergent, tree-like
development concept of the Türkic languages. But this picture also in some way correlates with
the tree obtained by linguostatistic method. As can be now seen, the last nodes of our
family tree from a historical perspective really date very believably, but these
datings are obviously connected with the history of the literary traditions, so it is not
excluded that they only characterize the connections between the dictionaries of the literary languages.
Namely, the separation of the Tatar language from the Eastern Kypchak group (1390) may be
associated with the formation of the so-called Khorezm-Türkic literary tradition, which is
geographically connected with the territory of the Kypchak Khanate, and as is known, the
contemporary Tatar literary language is partly a successor to that tradition. The branching of
the Uzbek language from the languages of the Nogai subgroup is dated by 1630, which more
or less corresponds to the final establishment of the Chagatai literary tradition.
Contents Türkic languages
Classification of Türkic languages
Indo-European, Arians, Dravidian, and Rigveda
Scythian Ethnic Affiliation
Foundation of the Scythian-Iranian theory
Türkic borrowings in English
Türkic in Romance
Alans in Pyrenees
Türkic in Greek