The Turkic languages

The Turkic languages are spoken from the Balkans to the Kolyma and from the Arctic forest to the Persian Gulf, by something over two hundred million people. They are close enough to one another that a speaker of one can often follow a neighbouring language without study. That closeness is the first fact about the family. The family spread very far and split comparatively late.

The first connected texts date from the eighth century and are in Old Turkic. The written record is continuous from then on.^[1] What came before is reconstructed. The reconstruction has no text to check itself against.

The first split

The family divides once at the root and then repeatedly within one of the halves.

Oghuric, also called Lir-Turkic, keeps r and l where the rest of the family has z and š: the numeral that is tokuz, nine, in the inscriptions is tăχăr in Chuvash. One Oghuric language lives.

Common Turkic, also called Shaz-Turkic, is everything else, and it includes all the branches below.

The two are far enough apart that the split must be older than any written Turkic. How much older it is has been estimated and is not known.

One proposal puts the split at the beginning of the era. It assumes a Pre-Turkic stage that ended in the first centuries AD. Two branches then stood side by side: Proto-Bulgar with r and l, and Proto-Turkic with z and š.^[2] The oldest Turkic loanwords in Proto-Samoyedic, taken over around the beginning of the era, still show r and l where later Turkic has z and š. Examples are Proto-Samoyedic yür, ‘hundred’, against Turkic yüz, and kil, ‘winter’, against Turkic kïš.^[3]

The record in order

The stages below are stages of writing, not of speech. Each is a body of texts with a date. What happened between the stages is inferred.

Before the eighth century. Turkic appears mostly as names and titles inside foreign texts: in the Chinese annals, in Sogdian, in Greek. Whether any of it is Turkic at all is argued case by case. The Xiongnu question is the largest of these arguments. From about the beginning of the Christian era neighbouring peoples recorded Turkic words. One example is the place name suvlag, ‘watering place’, in early Chinese sources, whose suffix is Turkic.^[4] The exception to names and titles is a group of documents in Bactrian from Afghanistan, the earliest dated to the first half of the seventh century, which contain Old Turkic phrases.^[5]
Old Turkic, from the eighth century. The runiform inscriptions of the Orkhon, the Yenisei, Talas and the Altai, and the Irk Bitig.^[1]
Old Uyghur, ninth to fourteenth century. The Buddhist, Manichaean and Christian literature of Turfan and Dunhuang, the largest body of pre-modern Turkic by volume.
Karakhanid, eleventh to thirteenth century. The first Turkic of an Islamic state: the Kutadgu Bilig and the dictionary of Mahmud al-Kashgari, which surveys the dialects of the Turkic world from the inside.^[6] The Kutadgu Bilig calls its own language türkçe.^[7] The language has few Persian loanwords but many calques, Turkic words that carry Persian meanings.^[8] Whether Karakhanid is the last stage of Old Turkic or the first of Middle Turkic is a matter of convention. Its grammar is close enough to Old Uyghur for one description to serve both. By contrast, the twelfth-century Atabatu ’l-Haqa’iq, though composed in the Karakhanid realm, belongs to Middle Turkic.^[9]
Volga Bulgar, thirteenth to fourteenth century. Gravestone inscriptions in Arabic script: the only Oghuric language on record besides Chuvash. The Danube Bulgars left a few sentences in Greek letters, of the ninth or tenth century, in an idiom that stands apart from the rest of the early record.^[10]
Kipchak of the Golden Horde and of Mamluk Egypt, thirteenth to fifteenth century. The Codex Cumanicus, compiled for missionaries and merchants, and the grammars written in Cairo for a Turkic-speaking army.
Khwarezmian Turkic and then Chagatai, fourteenth to nineteenth century. The literary language of Central Asia, from which modern Uzbek and modern Uyghur both descend.
Old Anatolian Turkish and then Ottoman, thirteenth to twentieth century. The western line, and the one that acquired a state that lasted six centuries.
The modern standards, from the nineteenth century. Written languages made deliberately, most of them in the Soviet Union and the Turkish Republic, out of dialects that had not been written before.

Three periods are usually distinguished. Old Turkic ends with the Mongol invasion of the thirteenth century. In Middle Turkic, four groups that are distinct by people and by country soon take shape as written languages: Eastern Turkic, Kipchak, Bulgar and Oghuz. Modern Turkic begins in the middle of the nineteenth century, when the as yet unwritten languages of High Asia were first described.^[11]

A second division counts five stages. Pre-Turkic runs to the beginning of the era. Proto-Bulgar and Proto-Turkic follow from about the first to the sixth century. Old Turkic and Old Bulgar belong to the sixth to the eleventh century, Middle Turkic and Middle Bulgar to the eleventh to the sixteenth, and New Turkic and Chuvash to the time after that.^[12]

Oghuric

Chuvash — on the middle Volga, the only living Oghuric language, and the one Turkic language a speaker of Turkish cannot follow at all.
Volga Bulgar — extinct by the fourteenth century, known from gravestones; the ancestor or a close relative of Chuvash.
Danube Bulgar — the language of the Bulgars who crossed the Danube in 679 under Asparukh and founded a state in what is now Bulgaria. The last dated text is of 907. The material is small: animal-year names, numerals, titles and personal names in the Old Russian copy of the list of Bulgar khans, runic and Greek inscriptions on the gold vessels of Nagy-Szent-Miklós, titles and names in Greek inscriptions and Byzantine sources, and a few loanwords in Church Slavonic. It is still enough to show an r and l language.^[13]
Khazar — the language of the Khazar khaganate is usually placed here on the strength of a handful of words and titles. The evidence does not carry more than a guess. Runiform inscriptions found in that country and further west were probably cut in the second half of the first millennium. They have been attributed to the Khazars, the Avars, the Pechenegs, the Bulgars and others. The readings proposed for them are unsatisfactory and doubtful.^[5]
Hunnic and Avar — claimed for this branch by some and denied by others; neither people left a text.

Common Turkic: Oghuz, the southwestern branch

Turkish — Anatolia and the Balkans, the largest Turkic language by speakers.
Azerbaijani — the Republic of Azerbaijan and north-western Iran, in two standards.
Turkmen — Turkmenistan, northern Iran and Afghanistan.
Gagauz — Moldova; Christian speakers of an Oghuz language.
Qashqai, Afshar and the other Oghuz varieties of Iran — tribal languages of the Zagros and the Iranian plateau.
Khorasani Turkic — north-eastern Iran, between Azerbaijani and Turkmen.
Salar — in Qinghai in China: an Oghuz language carried east in the fourteenth century and surrounded since by Chinese, Tibetan and Mongolic.
Old Anatolian Turkish and Ottoman — the written stages behind modern Turkish.
Pecheneg — usually placed with Oghuz, on names alone.

Common Turkic: Kipchak, the northwestern branch

Tatar and Bashkir — the Volga and the southern Urals.
Kazakh, Karakalpak and Nogai — the steppe from the Caspian to the Altai.
Kyrgyz — Tian Shan; classed with Kipchak, with the Siberian group, or between the two.
Karachay-Balkar and Kumyk — the northern Caucasus.
Crimean Tatar — Crimea, with dialects that lean Kipchak in the north and Oghuz in the south.
Karaim, Krymchak and Urum — the languages of Jewish and Christian communities of Crimea, Lithuania and Ukraine, written in Hebrew and Greek letters.
Siberian Tatar and Baraba — western Siberia.
Cuman — extinct; the language of the Codex Cumanicus.
Mamluk-Kipchak and Armeno-Kipchak — extinct written forms, in Cairo and in Lviv.

Common Turkic: Karluk, the southeastern branch

Uzbek — Uzbekistan and the neighbouring states, the second Turkic language by speakers.
Uyghur — Xinjiang; the modern language of the Uyghurs, descended from the Karluk-Chagatai line rather than directly from the language of the Uyghur khaganate.
Ili Turki and Aini — small languages of Xinjiang, the second with a Persian vocabulary on a Turkic grammar.
Karakhanid and Chagatai — the written stages behind both.

Common Turkic: the Siberian branch

Sakha (Yakut) and Dolgan — the Lena and the Taimyr, the furthest north any Turkic is spoken and the most divergent Common Turkic there is.
Tuvan and Tofa — the Sayan mountains.
Khakas, Shor, Chulym — the middle Yenisei and its tributaries; the country of the Yenisei Kirghiz.
Altai, northern and southern — the Altai mountains, two languages under one name.
Fuyu Kyrgyz — in Manchuria: a few speakers, carried there in the eighteenth century.
Old Kyrgyz — the language of the Yenisei epitaphs, claimed as the ancestor of this group.

Arghu

Khalaj — central Iran. It keeps sounds the rest of Common Turkic lost, which is why it is given a branch of its own rather than counted among the Oghuz languages that surround it.
Argu — the dialect of the Argu country around Balasaghun, known from the eleventh-century dictionary of Kashgari. It merges the old palatal nasal ñ with n and has -dum for the first-person past.^[14] Several of its features are shared with Khalaj.^[14]

Points of disagreement

Three questions run through every classification of this family.

The first is where the boundaries fall. Kyrgyz, Khalaj and the Siberian languages sit at the edges of their branches, and a classification that follows sound changes does not always agree with one that follows shared vocabulary.

The second is which of the peoples of the steppe spoke Turkic at all. The Xiongnu, the Huns, the Avars, the Pechenegs and the Khazars are each claimed for the family and each denied it. In every case the evidence is a short list of names and titles in a foreign transcription. On the Xiongnu, Talat Tekin holds that they spoke an archaic Turkic, in his terms Proto-Turkic or even Pre-Turkic. He rests the claim on the Chinese statement that the Göktürks descended from the Xiongnu, on a few glossed words, on a couplet recorded in the early fourth century, and on court titles of the Tuoba. He grants that the language material that could prove the claim is almost nothing.^[15] He reads the couplet as Turkic, ‘go out against the army, seize Bukuk’. In its first verb he finds l where later Turkic has š, and takes this as the sign of an l language.^[16] Read that way, the language of the couplet would stand on the r and l side of the first split, not beside the z and š Old Turkic of the Göktürks.^[17] A title proves little. The Avar ruler is called kaganus in Hungary, yet the source of that title appears not to have been Turkic, and other nations borrowed it too.^[18] The Rouran are the one case where the argument has moved recently. It moved away from Turkic.

The third is the family’s own ancestry: whether Turkic, Mongolic and Tungusic descend from one language or have grown alike through a thousand years of contact. The two answers use the same evidence and weigh it differently. No text will settle the question, since the contact and the descent would both lie before writing. Turkic shares many words, grammatical units and typological traits with Mongolic and, to a far lesser degree, with Tungusic, Korean and Japanese.^[18] Much of the copying between Turkic and Mongolic took place before either people began to write. Therefore the term Altaic is useful at least for an area of contact. A relationship by descent between Turkic and Mongolic is possible, but serious research into its nature is still lacking.^[18]

The spread of the family

Turkic spread across a continent in a few centuries and kept its shape. The reason is the way it was carried: by mobile people whose political units absorbed and released populations, over open country where distance was crossed on horseback. Languages carried that way stay in contact with one another and drift apart slowly.

The consequence is that the family is a poor guide to descent. A people that spoke Turkic in the ninth century need not be the children of a people that spoke Turkic in the second. The argument from language to ancestry is the one the ethnogenesis debate turns on.

References

^ a b Kül Tegin inscription
^ Tekin 2003, p. 9, 11
^ Tekin 2003, p. 15
^ Erdal 2004, p. 2
^ a b Erdal 2004, p. 4
^ Dīvān Lughāt al-Turk
^ Kutadgu Bilig
^ Dankoff 1983, p. 10–11
^ Erdal 2004, p. 8–9, 21
^ Erdal 2004, p. 5
^ Erdal 2004, p. 5–6
^ Tekin 2003, p. 14
^ Tekin 2003, p. 28–29
^ a b Erdal 2004, p. 16
^ Tekin 2003, p. 12–14
^ Tekin 2003, p. 12–13
^ Tekin 2003, p. 13, 16
^ a b c Erdal 2004, p. 3

Sources

Literature

Peter B. Golden, An Introduction to the History of the Turkic Peoples (1992)
Lars Johanson; Éva Á. Csató (ed.), The Turkic Languages (1998; 2nd ed. 2021)
Gerard Clauson, An Etymological Dictionary of Pre-Thirteenth-Century Turkish (1972)
András Róna-Tas, Hungarians and Europe in the Early Middle Ages (1999)
Emre-E. Yavuz, The Gokturk Empire: 2nd Edition (2025)
Marcel Erdal, A Grammar of Old Turkic (2004)
Talât Tekin; Mehmet Ölmez, Türk Dilleri: Giriş (2003)