drammock
commented
6 years ago hmm, I'd argue this is a case of "add new inventory(s) for Turkish from more recent source(s)"... but maybe that's what you meant by "update" (i.e., not that we should change the existing SPA/UPSID inventories, which "update" could also mean)
From Treysi (see also _turkish folder in squibs):
In Turkish velar stops, lateral liquids, non-lateral liquids and vowels have always been sources of exciting debates. The two Turkish phonological inventories in Phoible have slight problems that are not really different analyses issues. I am sending you 3 different chapters on Turkish phonology. All of these are written by some of the most influential Turkish linguists and are a lot more recent than the sources the contributors of Turkish have used.
Inventory Nuclear Turkish (SPA) states that Turkish has alveolar stops, fricative and nasals. However, Turkish has no such alveolar sounds. These are dental in Turkish as you can see from the chapters I'm sending you. (I have seen these Turkish sounds being classified as "alveolar" in other sources but those sources do not consider "dental" sounds as a distinct category.)
Inventory Nuclear Turkish (UPSID) has the velar fricative /ɣ/ as a phoneme. This is kind of problematic because the phone [ɣ] is NOT a sound in Standard Turkish. (Some dialects do have such a consonant but not the Standard dialect.) This phoneme is also a big source of debate in Turkish linguistics and I rewrote this section of the e-mail multiple times because it is a very complicated issue.
My personal opinion: I think the source of much of the confusion is that there is a distinct letter that is matched with /ɣ/ and Turkish's orthography is highly transparent so the fact that there is a letter that does not correspond to an actual sound causes problems. I believe assigning the /ɣ/ phoneme a sound from IPA chart is misleading. I wish a linguist who did not know Turkish or anything about Turkish orthography mapped out the Turkish phonological inventory but to my knowledge no such inventory exists. It would have been interesting to see this hypothetical linguist's analysis of the so-called /ɣ/ phoneme in Turkish.
Anyways, back to the the function of the phoneme /ɣ/... Sometimes it has no function/sound, sometimes it lengthens the preceding vowel and sometimes it is pronounced as a glide. So the problem with the Inventory Nuclear Turkish (UPSID) is that since there is no footnote or no comment about the /ɣ/, it gives the wrong impression to people who see that inventory. The sources I have sent you also consider /ɣ/ as a distinct phoneme but they always note that it is not a traditional phoneme.
Kornfilt states it is "a phonemic consonantal segment which remains unpronounced in the standard language." Goksel & Kerslake state it "lacks a corresponding ‘consonantal’ sound in standard Turkish." Erguvanli states it "is an abstract phonemic unit which does not surface phonetically in standard Turkish as a fricative."
I never got the chance to read Swift (1963) or Lees (1961) [the sources for the Inventory Nuclear Turkish (UPSID)] so I cannot tell you from first hand that they also did include such a note. However, Erguvanli states (on page 39) that "Thus, Lees includes [ɣ] as an abstract phonological unit in the phonetic inventory, which doesn't surface as a true consonant in standard Turkish."
That was a lot of explanation. I just wanted to explain why I thought the current Turkish inventories on Phoible are slightly problematic as a native speaker of Turkish and a Turkish linguist. Feel free to disregard all that. I really like talking about this stuff and could go on forever. The chapters I am attaching to this e-mail are very informative (and interesting, I think). They are from the following books:
Erguvanlı Taylan, E. (2015). The phonology, and morphology of Turkish. Istanbul, Turkey: Bogazici University Press.
Göksel, A., & Kerslake, C. (2004). Turkish: A comprehensive grammar. NY: Routledge.
Kornfilt, J. (1997). Turkish grammar. NY: Routledge.
A new phonological inventory for StandardTurkish according to Erguvanli or Goksel & Kerslake (their inventories are almost identical) would have a lot of similarities with the two inventories on Phoible. The consonants are very similar to Inventory Nuclear Turkish (UPSID). There are only 3 differences. (1) These chapters consider /s/ and /z/ dental fricatives. (2) According to these chapters, the lateral liquid phonemes are the palatal and velar ones. (3) Both of these chapters do not consider the glottal stop as a phoneme as it is only used by elder generations and most people in Turkey do not use it at all. The vowels in these chapters are more similar to those in Inventory Nuclear Turkish (SPA) but the mid front phonemes are closed not open. Moreover, the only long vowel phonemes according to these chapters are /a:/, /i:/, /e:/ and /u:/.