The oldest records of a Turkic language, the Old Turkic Orkhon inscriptions of the 7th century Göktürk khaganate, already show characteristics of the Eastern branch of Turkic, and reconstruction of Proto-Turkic must rely on comparisons of Old Turkic with early sources of the Western branches, Oghuz and Kypchak, as well as the Oghur branch (Bulgar, Chuvash, Khazar). Because early attestation of these non-Eastern languages is much more sparse, reconstruction of Proto-Turkic still rests fundamentally on East Old Turkic of the Göktürks.
Proto-Turkic exhibited vowel harmony, a feature sometimes also ascribed to the so-called Proto-Altaic, distinguishing vowel qualities e, i, o, u vs. ë, ï, ö, ü besides a, as well as two vowel quantities.
The consonant system had a two-way contrast of stop consonants (fortis vs. lenis), k, p, t vs. g, b, d, with verb-initial b- becoming h- still in Proto-Turkic. There was also an affricate consonant, č; at least one sibilant s; and sonorants m, n, ń, ŋ, r, ŕ, l, ĺ with a full series of nasal consonants.
The sounds denoted by ń, ĺ, ŕ refer to palatalized sounds and have been claimed by Altaicists to be direct inheritances from Proto-Altaic. The last two can be reconstructed with the aid of the Oghur languages, which show /r, l/ for *ŕ, *ĺ, while Common Turkic has *z, *š. Oghuric is thus sometimes referred to as Lir-Turkic and Common Turkic as Shaz-Turkic.
However, an alternate theory holds that Common Turkic is closer to the original state of affairs, and reconstructs Proto-Turkic *z, *š. The glottochronological reconstruction based on analysis of isoglosses and Sinicisms points to the timing of the r/z split at around 56 BCE–48 CE. This, as A. V. Dybo puts it may be associated with:
the historical situation that can be seen in the history of the Huns' division onto the Northern and Southern [groups]: the first separation and withdrawal of the Northern Huns to the west has occurred, as was stated above, in 56 BC,...the second split of the (Eastern) Huns into the northern and southern groups happened in 48 AD.
Dybo suggests that during that period, the Northern branch steadily migrated from western Mongolia through southern Xinjiang into the north's Dzungaria and then finally into Kazakhstan's Zhetysu until the 5th century.
- Antonov, Anton; Jacques, Guillaume (2012). "Turkic kümüš 'silver' and the lambdaism vs sigmatism debate". Turkic Languages. 15.2: 151–170.
- Gyula Décsy, The Turkic Protolanguage: A Computational Reconstruction (1998).
- Edward J. Vajda, review of Décsy (1998), Language (2000), 473–474. https://muse.jhu.edu/journals/language/summary/v076/76.2.vajda04.html
- Gerard Clauson, Etymological dictionary of pre-thirteenth-century Turkish, Oxford, Clarendon Press (1972).
- Vilhel Gronbech, Preliminary Studies in Turkic Historical Phonology (Uralic & Altaic), RoutledgeCurzon (1997), ISBN 0-7007-0935-5.
- András Róna-Tas, 'The Reconstruction of Proto-Turkic and the Genetic Question', in The Turkic Languages, edited by Lars Johanson and Éva Csató, Routledge Language Family Descriptions, London: Taylor & Francis, Inc. (1998), ISBN 0-415-08200-5, pp. 67–80.