Proto-Turkic/Phonology

Welcome to the Proto-Turkic pronunciation. More general explanation about pronunciations can be found at the previous lesson.

There are no */ʃ/ and */z/ phonemes in Proto-Turkic, instead, the phonemes are developed from *ĺ and *ŕ. The original phoneme *ń survives nowhere in modern Turkic languages, but for historical reasons (e.g. *ańaŕ are derived from a lost root *ań-, c.f. *ań-gɨr- > Kazakh аңызақ) it is reconstructed. In intervocalic and word-final positions, /b/, /d/, and /ɡ/ was pronounced as allophones /β/, /ð/, and /ɣ/, respectively. Such allophones can be known from the descendants, like Turkish dağ ← *tag "mountain", deve ← *tebe "camel", and ayak ← *adak "foot". Additionally, /k/ are also pronounced as /q/ when adjacent to back vowels.

All vowels contrast short and long vowels. Long vowels are indicated by macrons at top of vowels, so the long vowels include ⟨ā⟩, ⟨ạ̄⟩, ⟨ē⟩, ⟨ẹ̄⟩, ⟨ī⟩, ⟨ï̄⟩, ⟨ö⟩, ⟨ȫ⟩, ⟨ü⟩, and ⟨ǖ⟩ orthographically. Also, some words are contrasted by vowel length, for example *at "horse" vs. *āt "name". Only a number of Turkic languages that use inherited vowel length, for example Turkmen, Khalaj, or Yakut; note that Kyrgyz vowel lengths are formed through compensatory lengthening processes.

Proto-Turkic words are prototypically stressed in the last syllable (adak "foot", tabïĺgan "rabbit"), like almost all today's Turkic languages. However, the negator suffix -ma/-me (placed between the stem and the tense ending), including its irregular formation in the present tense -maŕ, -meŕ are always unstressed (kẹlmedi "(he) not came", not kẹlmedi); although in Turkish even in Old Turkic the suffix -mez (< -meŕ) is always stressed, while in Chuvash makes these negative suffixes became uniformly stressed. These suffixes will explained in the lessons 3: Basics (for -meŕ) and 8: Verbals (for -me).

In a verb clause sentence, the main stress is always on the item before the predicate. If the predicate is at the beginning, the stress is on the predicate. Depending on this, the location of the items can of course change. E.g; The difference between Bẹ ebke tǖn kẹltim and Ebke tǖn bẹ kẹltim ("I came home at night.") is that the main stress is on tǖn in the first one and bẹ in the second, as in Ural-Altaic languages.^[1] (Underlined items are predicates.)

The main stress in the noun clause is always in the predicate. E.g; Bẹ ebe (turur) ("I am good."). However, if it is expressed together with the verb turur, it becomes a verb clause and since turur can only come after the predicate in noun clauses, the stress will automatically be in the same item. E.g; ebe (turur) bẹ or ebe turur bẹ. (never ebe bẹ turur. If you make that sentence, it will mean ebe bẹ turur "good is me")

There is no special stress in monosyllabic words except adverbs.

Such phonemes include d-, g-, l-, ĺ-, m-, ŋ-, r-, and ŕ-.^[2] In Proto-Turkic language, g- cannot begin a word. Occasionally, proponents of a genetic Altaic grouping will reconstruct g- at the beginning of a word in Proto-Turkic, since there is a g at the beginning of the word in Proto-Mongolic. These reconstructions are generally considered incorrect, and the g- form is an Oghuz realization.

Whether there can be d- at the beginning of a word is a matter of debate. In some words d- is accepted at the beginning of the word, for example in the number *dȫrt (also reconstructed as *tȫrt). But otherwise, word-initial d- is a marginal phoneme.

The sounds o and its vowel-harmony variants can only appear in the first syllable of a root. For example, the reconstruction *dạmor is incorrect. The accepted reconstruction is *tạmïr.

Next Lesson: Pronouns and numbers

1. ↑ https://dergipark.org.tr/tr/download/article-file/31986 page 273
2. ↑ [1] The author, however, do not include d- and g- but included n- (except in *ne "what").