I have employed my own system of transcription for several reasons. First, the full complement of phones represented in IPA are not found in the Turkic languages, so there is no need to employ potentially confusing symbols. Most Turkic languages do not, for example, distinguish between aspirated and voiceless stops, so any attempt at a narrow transcription would fail. Second, most Turcological literature employs a modified version of the the Uralic Phonetic Alphabet or Americanist phonetic notation, so I am better able to match the symbols used in my sources by employing the system described here.
Because not all potential sounds are found in the Turkic languages, not all cells in the following charts have been filled out. Even when there is a symbol that could be used to express a given sound, it may not be used if there is no phonemic distinction made in a given language. For the most part, I have followed orthography. If, for example, a language's orthography distinguishes velar and uvular sounds, then that is expressed in the database. If not, then the two phones will be represented with a single grapheme, usually the velar one.
|Stop||p b||t d||ṭ ḍ||ť ď||k g||q ɢ||'|
|Affricate||pf bv||ts dz||č ǰ||čʼ ǰʼ|
|Fricative||φ β||f v||θ ð||s z||š ž||ṣ ẓ||šʼ žʼ||ś ź||x ɣ||ḥ||h|
The schwa /ə/ represents a highly reduced and centralized vowel; it often indicates a lax pronunciation of /ï/.
The above charts are supplemented with a few additional symbols and diacritics as needed. Doerfer's notation is employed for a few languages; see this section for further details.
- Other glides not shown on the chart: labio-velar: w, labio-palatal: ẅ
- Acute accent: ◌́ (ḱ, ź) - indicates palatalization, applied only to coronal and velar sounds
- Grave accent: ◌̀ (à, ǜ) - indicates low tone (sometimes described as "pharyngealization"), usually when describing the Sayan languages.
- Breve: ◌̆ (ŏ, ï̆)- indicates a short or weakly articulated vowel. In Chuvash, the breve represents the poorly described reduced vowels /ă/ and /ĕ/.
- Superscript h: ʰ (ʰp, tʰ) - indicates aspiration, only indicated when aspiration and voicing are contrastive features or when pre-aspiration occurs, as in Salar and Western Yugur.
- Small caps: d, z, ü̆, δ, γ - indicates a partially voiced sound. Used in the Uralic Phonetic Alphabet, but avoided here because of display issues (the delta and gamma should render as small caps, but do not in all browsers). Pretty much only found in Doerfer, represented by regular capital letters if necessary.
Gerhard Doerfer (1920-2003) was an incredibly important Turcologist and Altaicist. Thanks to his work we have documentation of many otherwise understudied languages of Iran. Because I have employed many of his works as sources, I include here a brief discussion of his system of transcription. Further information about his data collection methods are available here.
Like many Turcologists, Doerfer employs a phonetic notation based on the Uralic Phonetic Alphabet and Americanist phonetic notation. Consonants tend to be indicated in a fairly straightforward way, although it is worth noting that he employs a left arrowhead below to indicate uvular /ɢ/, e.g. "g͔", presumably because he and the UPA employ small caps to indicate a semi-voiced consonant. Uvular /q/ is always expressed as "q".
Doerfer's vowels are somewhat confusing, as shown in the two vowel charts reproduced below. The first represents the vowel space as a tetrahedron, the second as a cube.
Many vowels are followed by superscript numbers. Doerfer describes their meanings as follows:
- etwas offen - slightly open/lowered
- etwas geschlossen - slightly close/raised
- etwas vorn - slightly fronted
- etwas hintern - slightly backed
- etwas labialisert - slightly labialized/rounded
- etwas delabialisert - slightly delabialized/unrounded
Doerfer also employs the schwa "ə" to indicate a highly reduced vowel, something that is not standard in the Uralic Phonetic Alphabet. Length is indicated with a macron (ā, ǖ), while mid-long vowels are indicated with a raised dot (ä˙, o˙) following the letter. Nasals are indicated with a tilde (ǟ̃, ĩ). Occasionally, diacritics will not stack correctly. This is an issue with the display font. Note that Doerfer indicates the palatal glide with /y/ rather than /j/.
For the most part, I have taken forms provided in Doerfer's works as-is. The transcription scheme used in this database makes the following changes:
- g͔ is transcribed as ɢ
- Superscript numbers following vowels are dropped.
Василий Васильевич Радлов, a.k.a. Wilhelm Friedrich Radloff (1837-1918), was a major figure in early Turkology. His magnum opus, Опыт словаря тюркских нарѣчій / Versuch eines Wörterbuches der Türk-Dialecte, employs a rather unusual system of Cyrillic transcription that is not always obvious. Although I have not yet used data from this work, I have included details of his transcription system as the volume in which it is laid out is not easily located. Further information on Radloff and this work can be found here.
- а - as in German. Local transcription /a/, IPA /a~ɑ/
- ɷ - as in English all. Local transcription /å/, IPA /ɒ~ɔ/
- ӓ - as in German Berg, French frère. Local transcription /ä~e/, IPA /ɛ/
- ɷ̈ - as in English learn, between /ä/ and /ö/. Local transcription /ĕ~ə/, IPA /ɜ~ɵ/
- е - as in German sehen. Local transcription /e/, IPA /e/
- э - as in English man. Local transcription /ä/, IPA /æ/. Only found in Trakai Karaim.
- о - as in German Woche. Local transcription /o/, IPA /o/
- ӧ - as in German Mörder. Local transcription /ö/, IPA /ø/
- ө - weakly pronounced o with the tongue root lowered???. Local transcription /ŏ/, IPA /ö/. Only found in Volga dialects
- ӫ - palatal version of ө. Local transcription /ö̆/, IPA /ʏ~ø̈/
- ы - not like the sound this character represents in Russian. Local transcription /ï/, IPA /ɨ~ɯ/
- ы̵ - as in French que. Local transcription /ə~ï̆/, IPA /ɘ/
- і - as in German sich. Local transcription /i/, IPA /i/
- і̣ - as in English before. Local transcription /ĭ/, IPA /ɪ~ɘ/. Only found in Volga dialects
- у - as in German Bund. Local transcription /u/, IPA /u/
- ӱ - as in German Küche. Local transcription /ü/, IPA /y~ʏ/
- ў - Only found word-finally in Taranchi dialects. Appears to be a semivowel. Local transcription /w/, IPA /w/
Long vowels are indicated with a macron.
|Stop||п б||т д||ȷ, ȷ̉||к г||k|
|Affricate||ц ӡ||ч џ|
|Fricative||ф в||с з||ш ж||ꜧ||х ҕ||һ|
- The character w represents the labiovelar glide /w/.
- Palatalization is indicated with an apostrophe over the character: п̓, т̓. This appears to indicate a palatalized rather than palatal consonant.
- The symbol ђ is used only for Lutsk Karaim, where is represents the palatal sound found before i in words like tiš~kiš and til~kil. It is unclear how this differs from the other palatal sounds he describes.
- The symbol б̱ is used only for the Abakan dialect word-finally, where it represents a sound with labial closure but some nasal release. Perhaps IPA /b̃/, /bⁿ/?
- The symbol г̄ represents "a sonorous k-sound that I found only in the northern Altai. The tongue flap on the soft palate is so weak at the production of this sound, that the outgoing air produces a sound which consists, as it were, of a series g-g-g." (No clue here...it's likely not a uvular trill since that sound exists in German. Maybe it's pharyngeal?)