Talk:Turkish phonology

Comments

Thanks for creating this article. Atilim Gunes Baydin 22:18, 24 April 2007 (UTC)

Different phonemes, or allophones?

The section Turkish phonology describes /c/, /ɟ/ and /ɫ/ as separate phonemes next to /k/, /g/ and /l/ – at least that is the impression I get from reading the text and the use of the notation "/.../". Would it not to be better to describe [k] and [c] as complementary allophones of a single phoneme /k/, and so on?  --Lambiam 15:19, 4 December 2008 (UTC)

As the article states, they're in complementary distribution in native words but contrast in loanwords. So they're marginal phonemes, maybe but still differentiated. AlexanderKaras (talk) 22:46, 5 June 2010 (UTC)

I have put these sounds in parentheses, as they only contrast in loanwords. Jɑυмe (xarrades) 19:44, 10 December 2011 (UTC)

Phoneme consistency

The consonant chart uses kʲ, gʲ, while some examples use c, ɟ. For consistency's sake, I'm going to change the latter to the former. I don't think there's any reason for these to be inconsistent, but should the latter or the former be used? (Personally I prefer the former because it shows the relationship between the two k variants. ʙʌsʌwʌʟʌ ^{spik ʌp!} 10:28, 6 February 2011 (UTC)

Actually I'm going with the latter since WP:IPA for Turkish supports it. ʙʌsʌwʌʟʌ ^{spik ʌp!} 10:33, 6 February 2011 (UTC)

Either is fine, but you're right. We should be consistent. — Ƶ§œš¹ _{[aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]} 15:54, 6 February 2011 (UTC)

Velarization

According to the article,

"Some Arabic loanwords pronounced with a velarized /tˠ/ in Turkish[citation needed] do not conform to vowel harmony either. For example saat-e ('to the clock/hour'), seyahat-e ('to the trip'), istirahat-e ('to rest')."

I'm skeptical about this. These words in Arabic have plain /t/; they are not velarized. Why would they be in Turkish? Unless I see a source, I might go ahead and delete it. Thoughts? AlexanderKaras (talk) 20:12, 22 February 2011 (UTC)

I was the one who wrote "velarization" and was not sure of it myself, feel free to correct it. As mentioned in the text, some Arabic-origin words ending in -t that do not conform to vowel harmony in Turkish are: saat (ساعة), seyahat (سياحة), istirahat (إستراحة), hakikat (حقيقة) (you can confirm this with the Turkish Language Institute dictionary, for example [1]). Some other Arabic-origin words ending in -t that do conform to vowel harmony are hattat (خطّاط), sakat (سقط), vasat (وسط). Arabic words ending with -d have been imported with a -t into Turkish and they also conform to vowel harmony, such as fesat (فساد), cellat (جلّاد), hasat (حصد). It appears from the Nisanyan etymological dictionary that most Arabic words that do not obey the vowel harmony words end with ة, (Tāʾ marbūṭa). There are exceptions though, such as hayat (حياة) and rahat (راحة), I am not sure why. Words ending with the Arabic letter ط (Ṭāʾ) appear to be imported as a normal T sound in Turkish. These are my observations, I cannot provide a source that explains all this, although there are Turkish texts that mention that foreign words that tend to be pronounced with a "light" T are followed by a front vowel (for example here, at the end of the section titled "Büyük Ünlü Uyumunun Özel Durumları"). I am not a linguist, so if you could use this information appropriately to improve the text, it will be great! --İnfoCan (talk) 18:08, 28 February 2011 (UTC)

/tˠ/ is palatalized, not velarized. And if there are two different pronunciations of the letter ‹t›, this should be mentioned (with sources) first in the section on Consonants. CapnPrep (talk) 22:39, 28 February 2011 (UTC)

In Arabic, emphatic /tˠ/ is velarized/pharyngealized, not palatalized. — Ƶ§œš¹ _{[aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]} 23:28, 28 February 2011 (UTC)

I guess you mean ط‎ /tˤ/, but as İnfoCan said, this one gives rise to a normal /t/ in Turkish, with no effect on VH. The question is whether the Turkish ‹t› (corresponding to ة) that triggers front vowel harmony should be analyzed as palatalized /tˠ/. As ʙʌsʌwʌʟʌ says below, this does not appear to be phonetically motivated. CapnPrep (talk) 00:16, 1 March 2011 (UTC)

I don't think there are two different pronunciations of <t> in Turkish. Also, it seems that the palatalized /tˠ/ in Arabic, when borrowed into Turkish actually follows VH, while the plain dental/alveolar /t/ sometimes doesn't. We need some good sources on this. Does Geoffrey Lewis say anything on this? ʙʌsʌwʌʟʌ ^{spik ʌp!} 22:56, 28 February 2011 (UTC)

There was a discussion among the old Arabic grammarians on how to pronounce the ta-marbuta. Some said it should be pronounced with the normal sound of any "a" in the language, others said it should have a fronted sound (probably something like [æ] or [ɛ]). Some dialects of Arabic, notably Syrian/Lebanese use such a pronunciation still today (though only in the dialect, and not in all environments). It's just a guess, but this could theoretically be the reason why the Turks felt that ta-marbuta should trigger front vowels. This would also explain the exception "hayat", though not "rahat". A fronted pronunciation (indeed [e]) is also used in Persian, but this seems to be a relatively young development. — Preceding unsigned comment added by 93.206.139.188 (talk) 00:27, 18 February 2016 (UTC)

It has stuck as a historic artifact Erkin Alp Güney :

Consonants of Istanbul Turkish according to Redhouse

	Labial		Dental		Alveolar		Palato- alveolar		Palatal		Velar		Glottal
Nasal		m		n̪		nˠ				(ŋʲ)		ŋ
Stop	p	b	t		tˠ	d	t͡ʃ	d͡ʒ	c	ɟ	k	ɡ
Fricative	f	v	s	z			ʃ	ʒ			x	ɣ	h
Approximant						ɫ				j
Trill						r

18:37, 19 July 2019 (UTC)

Harf-e ?

Continuing with the topic of vowel harmony, of the Arabic-origin Turkish words harf (حرف), zarf (ظرف) and sarf (صرف), the first one does not conform to vowel harmony (harf-e, 'to the letter'), while the other two do (zarf-a, 'to the envelope', sarf-a, 'to the expenditure'). It would be good if an explanation could be found for this usage. --İnfoCan (talk) 19:18, 28 February 2011 (UTC)

From an Arabic point of view, there is actually a difference inasmuch as sarf and zarf have an initial laryngealized plosive, whose laryngeality is (or: may be) extended to the rest of the word: [sˁɑrˁfˁ], [ðˁɑrˁfˁ]. In harf, this is not the case: [ħɑrf]. But I wonder if that's the reason or whether it's just coincidental. Why would harf take a front vowel anyway? Is that expected?

It might be caused by "harf" being pronounced with front /a/ in Persian and Ottoman. — Preceding unsigned comment added by Erkinalp9035 (talk • contribs) 17:15, 22 March 2018 (UTC)

Sezer stress

Proper names (of both places and foreign people) follow a different stress pattern, known in the linguistics literature as Sezer stress (after the discoverer of the pattern, Engin Sezer). In this lexical domain, stress occurs on the antepenult if the penult is light and the antepenult is heavy, and otherwise on the penult. The weight of the final syllable is irrelevant.

Okay, cool. Newbie crit coming. a) Do Turkish names not follow Sezer stress? i.e. are they always word-final? b) Now I've unravelled - I think - the description, wouldn't it be clearer to split the rule description into two halves: first, that two-syllable words have word-initial stress; second, that three-syllable words obey the rule specified. Afraid I can't come up with a more elegant way of stating that rule.

Turkish proper names are very complex. Some initial, some Sezer pattern, some final. Local dialects even show a reversed Sezer pattern. 78.162.115.186 (talk) 14:24, 7 March 2018 (UTC)

I'm asking because Wikipedia is pretty high up in Google rankings for 'Sezer stress' and so it's more productive for me to ask the questions here than to delve into lit for a language I'm not familiar with. 87.194.30.190 (talk) 20:16, 10 July 2011 (UTC)

9th vowel?

There's a discussion here among Turkish speakers as to whether there is significant allophony for /e/, [e̞] and [æ]. Some argue it sounds nothing like English /æ/ (that the distinction is more like [e] [ɛ]), others that it does. Also that there are lexicalized exceptions in Istanbul which would make it incipiently phonemic. Came up because an editor wanted to add it to the Turkish IPA key, but it should really be worked out here first.

Last post:

There is general rule for when and how to pronunce each e. Most e's are closed apart from the rule below:

In the same syllable, if the e is follwed by: r, l, m or n. The "e" is open.

Ex: Sen, the "e" is followed by "n", therefore pronunced: sæn.

Let's make it accusative:

Seni, the syllables are: Se-ni, the "e" is not follwed by an "n" in the same syllable. Thus: closed.

There are a few exceptions of course. For instance the word "renk" is pronunced with a closed e, although it is followed by an "n".

The editor bringing this to my attention,[2] User:Amateur55, says,

The mentioned exceptions only exist in the standard Istanbul dialect. For instance, in my northern dialect, I pronounce kendi, elli, renk with a /æ/ sound, while they are all exceptions to the described general rules in Istanbul dialect.

Certainly s.t. we should take into account if we can justify it. And, does the same thing happen w other mid vowels? — kwami (talk) 09:03, 13 August 2011 (UTC)

First, I never noticed something similar with the other mid-vowels.

Second, there's definitely a /æ/ vowel in Turkish. The users in the given page use /e/ and /ɛ/ for the closed e sound, which is represented by /e̞/ in the current Turkish vowel chart. The correct IPA representation of that sound is the one here, not there. — amateur (talk) 11:20, 13 August 2011 (UTC)

But how do we know, when one says it's the English 'a' sound, and another says it's nothing like that sound? — kwami (talk) 20:23, 13 August 2011 (UTC)

Well, we know that the closed e sound is /e̞/ and not /e/ or /ɛ/, this is confirmed by our references in the article.

For the open e sound, one user in the given page claims that sound does not exist at all, seven others and me assert that it exists and is /æ/. No one says that it's /ɛ/. If that's not enough, then we have to find a RS or determine the sound by listening.

Besides, in addition to the given rules above, the "-mez" suffix has an open e. — amateur (talk) 15:53, 14 August 2011 (UTC)

maxguncel and macrotis say there's only one /e/ sound. Seraphim says there are two, but that the open one is [ɛ] and that [æ] sounds "very rude" to Turkish ears. Says they're like French é and ê. Eline0909 and Rallino agree there are two sounds but don't specify what they are. Only Black4blue says he agrees with you that it's [æ]. So that's two votes for [æ], not seven; one vote for [ɛ]; two for an undisclosed distinction; and two for no distinction. Not that such things should be decided by a vote: Reality is not a democracy. Still, it's hardly a sweeping endorsement for [æ]. — kwami (talk) 23:42, 14 August 2011 (UTC)

Speakers who pronounce close e as [e] pronounce open e as [ɛ], ones who pronounce close e as [e̞] pronounce open e as [æ]. One-thirds of vowel height difference. Erkinalp9035 talk) 18:14, 7 March 2018 (UTC)

The Redhouse grammar of Ottoman Turkish (1884) counts 11 short (phonetic) vowels and 8 long, but these are not phonemes. Since the Redhouse analysis predates the Latin alphabet for Turkish, it's an interesting piece of evidence. I wonder whether (a) the phonetic analysis is still valid for modern Turkish; (b) how the phonetic analysis corresponds to the phonemic analysis; (c) how it might be influenced/biased/distorted by the Ottoman writing system and the English phonetic system. --Macrakis (talk) 01:37, 15 August 2011 (UTC)

Dale Chock posted a link to a ref on my talk page: Göksel and Kerslake, Turkish: a comprehensive grammar, 2005, p. 10. "All" vowels but /o a/ have lowered allophones word-finally: [ɪ ʏ ɛ œ ʊ] (they don't give an example for /ɯ/). ([e ø o], BTW, are close mid, not mid.) In addition, /e/ is [æ] before coda /m n l r/ with some lexical exceptions for some people. Just what amateur said above. There was another ref, Balpinar, Turkish phonology, morphology, and syntax, pp. 38–39,[3] which had something very different: /i u/ are [i u] before ğ, [ɪ ʊ] elsewhere. /e/ is something like [ej] before ğ, [ɛ] elsewhere. Some speakers have [æ] in "certain environments", such as in genç, mendil, Mehmet. /a/ is [ə] in the first syllable and [ʌ] elsewhere. Some speakers have [a], but that is much less common. — kwami (talk) 09:18, 17 August 2011 (UTC)

Don't Celia Kerslake and Aslı Göksel say that 〈a〉 is /ɑ/, not /ɐ/? Then that /ɑ/ has allophones in certain environments.

"/a/ is [ə] in the first syllable and [ʌ] elsewhere." What? [ə] would be misunderstood by Turkish speakers as a mispronounced /ɯ/ or mispronounced /ɛ/. Erkinalp9035 (talk) 17:49, 12 March 2018 (UTC)

Added Redhouse vowel table translated into currently used phonology terms and IPA. Erkin Alp Güney 19:11, 16 July 2019 (UTC)

Moved to talkspace after an edit war: Erkin Alp Güney 20:30, 16 July 2019 (UTC)

Long vowels of Istanbul Turkish according to Redhouse

	Front		Back
	unrounded	rounded	unrounded	rounded
Close	i	ʏ	ɯ	u
Open	ɛ	œ	ɑ	o

Short vowels of Istanbul Turkish according to Redhouse

	Front		Central	Back
	Unrounded	Rounded	Unrounded	Unrounded	Rounded
Close	i	y			u
Close-mid	e		ɯ		o
Open-mid		ɔ̈			ɔ
Open	æ		ɐ	ɑ

The source is outdated and you're misusing it. We have more accurate, detailed, and up-to-date sources available. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:37, 16 July 2019 (UTC)

Turkish phonology has not changed much since late 18th century except for vowel harmony and consonantal assimilation of few suffixes, hence still relevant. Erkin Alp Güney 20:34, 16 July 2019 (UTC)

Proof that current Istanbul Turkish phonology is not that different than Abdulhamid era: https://www.youtube.com/watch?v=Wefs6EDVSOQ (then-monarch speaking about a dethrone attempt) Erkin Alp Güney 20:46, 16 July 2019 (UTC)

Again, you are misusing the source. I'm sorry, you will need to find more up-to-date resources to back up your claims. Until then, this table should stay out of the article. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 21:31, 16 July 2019 (UTC)

If that's true, finding a recent reliable source that supports your claim shouldn't be difficult. Nardog (talk) 21:59, 16 July 2019 (UTC)

References

Sound of Ğ (Soft g)

It seems like there is a debate on whether this phoneme has a sound or not. While many sources claim that it has no sound and used only for determining the vowel length, I think it does have a sound, but not a velar approximant [ɰ] as it is said in the article. I think it's a sound articulated far back in the mouth, something like a uvular approximant [ʁ̞]. Since it's a very slight sound, it might be realized as having no sound, but I think it does. I'm not a linguist or anything, but I (at least) know that the sound is not articulated anywhere near the velum. — amateur (talk) 20:43, 30 August 2011 (UTC)

Local dialects do articulate as in Language of Azerbaijan. TRT accent do not. 78.162.115.186 (talk) 14:27, 7 March 2018 (UTC)

Turkish terminal "-r" sounding like "-rzh" or "-rsh"

[I am copying the following from Wikipedia:Reference_desk as it is relevant to this article and may inspire further development. --İnfoCan (talk) 22:09, 7 May 2012 (UTC) ]

Some Turkish speakers tend to pronounce words ending with "-r" like "-rzh" or "-rsh". Not everyone does it, and it seems to occur mainly when the word ending with "-r" is the last word of a sentence, while you hear a clear "r" sound if another word follows it.

You can hear examples of this on this YouTube video:

At :03 seconds the speaker pronounces kelimeler as "kelimelerzh", and at :40 seconds pronounces onlar as "onlarzh" or "onlarsh".

Here is another example from a YouTube video I found mentioned in a forum site :

at :11 seconds, and again at :59, the speaker pronounces hayir as "hayirzh". A at 4:31 she reads three phrases in which she pronounces bir as "birzh"- then repeats the phrases pronouncing it "bir".

I couldn't find anything on this phenomenon other than inquires of it by learners of Turkish, and curiously most Turks are not aware of it, some even deny it strongly. Is there any formal description of this? To make the question more generic, how would you describe/label this kind of R sound? Has such a phenomenon been observed in other languages? --İnfoCan (talk) 16:20, 3 May 2012 (UTC)

I don't have an answer, but t's also discussed briefly at the alveolar trill's talk page: Talk:Alveolar_trill#The_Turkish_Final_.22r.22 (though the sound is an alveolar flap, not trill, according to our article on Turkish phonology). ---Sluzzelin talk 17:25, 3 May 2012 (UTC)

"R" has interchanged with a fricative at various times and places. In West Slavonic, before a front vowel, /r/ became a fricativised trill in Czech (written '"ř") and a fricative /ʒ/ in Polish (written "rz"). On the other hand, in North Germanic, final "-r" results from an original "-s", apparently via "-z". --ColinFine (talk) 17:36, 3 May 2012 (UTC)

My assumption when I met a Turkish speaker was that it was tied in with word-final consonant devoicing, where word-final b d g v z become p t k f s. r seems to have been classified more like z than like l or n, so it devoices as well. In the process, it also becomes more sh-like, likely because a devoiced r is harder to produce than a normal r.

Our page on Turkish phonology seems to confirm this. Lsfreak (talk) 18:13, 3 May 2012 (UTC)

Why is it that, for a language spoken by 10s of millions of people, in a top 20 world power, the Wikipedia article dedicated to it has no mention of the phenomenon most noticeable by non-speakers, 12 years after a note has been written about it in the discussion page? — Preceding unsigned comment added by 151.48.169.118 (talk) 14:08, 9 April 2024 (UTC)

Take another look. It's covered in the Consonants section by the first subnote of the fourth note under "Phonetic notes". Largoplazo (talk) 15:40, 9 April 2024 (UTC)

Why /c/ and not /k/

Why is it that in the Turkish language article the pronunciation of Türkçe is [tyrcˈtʃe] and not [tyrkˈtʃe]. Which rule states that? — Preceding unsigned comment added by 109.245.68.133‎ (talk • contribs)

We may have to ask User:Amateur55, who changed the k to a c back in December. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 20:22, 9 June 2012 (UTC)

It is explained in the article:

In native Turkic words, the velar consonants /k, ɡ/ are palatalized to [c, ɟ] (similar to Russian) when adjacent to the front vowels /e, i, œ, y/. Similarly, the consonant /l/ is realized as a clear or light [l] next to front vowels (including word finally), and as a velarized [ɫ] (dark l) next to the central and back vowels /a, ɯ, o, u/. These alternations are not indicated orthographically: the same letters ⟨k⟩, ⟨g⟩, and ⟨l⟩ are used for both pronunciations. amateur (talk) 19:16, 22 December 2012 (UTC)

It is not [c], which is articulated between the mid-body of the tongue and the dome of the hard palate. It is a palatalized k, [kʲ], but that is not the same as [c]. N. Pharris (talk) 06:27, 17 February 2022 (UTC)

düğün (wedding)

Before editing (and probably enhancing) the article, I have a question. Is the pronunciation of düğün really ['dü:jün] or do some Turkish dialects have the ğ all silent (['dü:n])? To me the latter would be dubious, especially in respect to frequent misunderstandings likely to be expected in fast speech. dün = "yesterday" !! Hence, düğün ought to be one of the few "ğ-words" which apparently do not have an alternative pronunciation. (unlike soğuk (either ['so:uk] or (more rarely, but used) ['so:ωuk]) -andy 77.190.18.144 (talk) 00:41, 5 July 2013 (UTC)

It's definitely not ['dü:jün]. It's silent (and lengthens the vowel) in the standard dialect. — Lfdder (talk) 00:50, 5 July 2013 (UTC)

It is /dʏ.'ʏn/. Erkin Alp Güney 09:22, 1 February 2020 (UTC)

The stress in these words is on the ultimate syllable: /so(:)ˈuk/, /dy(:)ˈyn/. I think that in standard (Istanbul) speech there is no perceptible lengthening of the vowel, except perhaps when reciting a poem. The article states that the insertion of a [j] glide in düğün is mandatory in fast speech to distinguish it from dün, but I think this claim is bogus. To start, the second syllable of /dyˈyn/ has (tonal) stress, so you can hear the tone going up. Even in very fast speech in which the two syllables are contracted to one, there is still an audible contrast between /dy:n/ (< /dyˈyn/) and /dyn/.  --Lambiam 11:52, 8 April 2019 (UTC)

Rhythm Section has very confusing sentence

The Section "Rhythm" contains these three sentences "Turkish is usually considered a syllable-timed languages. Stressed syllabled and unstressed are not different so much. Pitch and stress are very important in Turkish."

The meaning of the embolded sentence is not clear. — Preceding unsigned comment added by Boaby (talk • contribs) 14:47, 25 October 2014 (UTC)

Some nouns may be distinguished by their stresses only. But unstressed sylllables do not get shortened nor a vowel reduction occurs (at least in Anatolian Turkish). Hence syllable-timed (actually mora-timed as long vowels trigger an extra timing unit). Erkinalp9035 (talk) 06:20, 15 March 2018 (UTC)

The difference is small (phonetically). And yet it is important (semantically).  --Lambiam 12:00, 8 April 2019 (UTC)

"saving space"

/t͡ʃ, d͡ʒ/ are affricates, not plosives. They have nevertheless been placed in the table in that manner to save space.

Haha, what? This isn't a printed encyclopedia, you know. You don't have to "save space" on the Internet :D — Preceding unsigned comment added by 83.30.144.127 (talk) 21:24, 25 July 2016 (UTC)

An affricate is essentially a stop with a fricative release, so it does make sense to "save space" in that manner. Mr KEBAB (talk) 22:17, 25 July 2016 (UTC)

Perhaps "plosive" could be changed to "oral occlusive". Votedaisy (talk) 01:17, 31 July 2016 (UTC)

Phonotactics: complex onsets in not very careful speech

The phonotactics section states: Turkish only allows complex onsets in a few recent English, French and Italian loanwords; such as Fransa, plan, program, propaganda, strateji, stres, steril and tren. Even in these words, the complex onsets are only pronounced as such in very careful speech. How are these onsets articulated in non-careful speech? Are consonants dropped, or are epenthetic vowels inserted? And what are the rules for what and where the consonants or epenthetic vowels are dropped or inserted? At the very least, it would be good to have IPA for the example words in non-careful speech. --Votedaisy (talk) 00:57, 13 August 2016 (UTC)

Epenthesis of something between schwa and "ı", as in TRT accent. We never drop consonants. But complex clusters are increasingly pronounced without simplification in private broadcasters. Erkinalp9035 (talk) 14:47, 7 March 2018 (UTC)

The word kral ("king") is an old loan, and thus an exception to the recency of loanwords with complex onsets. It is pronounced as if it is spelled kıral, a spelling one may even occasionally encounter in print (and in the surname of Turkish filmmaker Erden Kıral). Which vowel plays the epenthetic role is not easily predictable, as seen in Turkish kulüp from French club and pilaki from Greek πλακί (plakí).  --Lambiam 12:21, 8 April 2019 (UTC)

Palatals or palatalized velars?

Despite widespread confusion, these are not the same thing, you know. What little Turkish I've paid enough attention to has, at most, palatalized velars [kʲ gʲ], in keeping with the presence of [lʲ], not palatals [c ɟ]. (This guy barely fronts his /k/ at all; that's not palatalization, it's just the Western Romance [k̟g˖].) Palatalized velars are also found in Russian, Polish, and notably northern incl. Standard Greek; palatal plosives are very rare in Europe, found in Latvian (ķ ģ) and Hungarian (ty gy; soundfiles here) and quite possibly nowhere else.

So, unless evidence for palatal plosives in any kind of Turkish is forthcoming soon, I'm going to change the text.

David Marjanović (talk) 07:25, 18 September 2016 (UTC)

As a native Turkish speaker, I can assure you that they are palatal stops. I always pronounce the velar stops as [c ɟ] before front vowels. And I'm not a speaker of a specific dialect, I speak Standard Turkish and this is the way how we pronounce them. — efekankorpez (talk) 22:50, 1 December 2017 (UTC+3)

As far as your own speech is concerned, that may be true. But I agree that the example speaker does not have [c], nor even probably [kʲ], but just a slightly protracted version of [k]. So would you say the speaker speaks a strange kind of Turkish? I'd rather think that the phoneme /c/ ranges between [c]~[kʲ]~[k̟]. — Preceding unsigned comment added by 2.201.0.62 (talk) 00:07, 20 March 2019 (UTC)

Perhaps there is also some variation between rounded and unrounded vowels. This speaker pronounces "kelimeler" (beginning of the video) with a palatalised sound, [kʲ] or [c] (still not sure it's really [c], though). However, she clearly says "güle güle" (at the very end) without such palatalisation; that's just [ɡ], at most slightly protracted. — Preceding unsigned comment added by 2.201.0.62 (talk) 00:25, 20 March 2019 (UTC)

I can agree with David. Modern Turkish, unlike Azerbaijani and Hungarian, uses [gʲ kʲ], not [ɟ c]. Erkin Alp Güney 19:20, 18 December 2021 (UTC)

Final /h/

@Mr KEBAB: When they mean 'final /h/', do they mean at the end of syllables or words? — AWESOME meeos ！ * ([ˈjæb.ə ət məɪ])) 01:40, 5 September 2017 (UTC)

@Awesomemeeos: Have you checked the source yourself? I can't do that right now. Mr KEBAB (talk) 01:48, 5 September 2017 (UTC)

Velar and palatal nasal as allophones

Contrary to what's being claimed in the article, in standard Turkish /n̪/ is never realized as [ɲ ŋ]. Using these "allophones" is very often considered uneducated speech. One should also note that those speakers who use the sound /ŋ/ also often realize word-final /n̪/ as [ŋ]. So this allophone should be regarded as dialectal and non-standard. As for /ɲ/, I have never heard this sound being used in any dialect of Turkish. There is no distinct palatal nasal in Turkish like the palatal nasals of French and Italian. — efekankorpez (talk) 22:50, 1 December 2017 (UTC+3)

ə~ɯ or ɪ̈?

Kılıç and Öğüt describe /ɯ/ as having a huge allophonic variation in the [ə ~ ɯ] area.

I have read the whole article, the huge variation is for perception. For pronunciation, exact formant values are given, which corresponds to [ɪ̈]. Erkinalp9035 (talk) 20:10, 7 June 2018 (UTC)

Superscript numbers

In the consonant phonemes table there are some consonants listed with superscript numbers, e.g. Fricative voiceless f³, voiced ʒ³. What do these numbers signify? — Preceding unsigned comment added by Tuluqaruk (talk • contribs) 10:37, 6 December 2018 (UTC)

Found it myself, it refers to the footnotes dirctly beneath the table. --Tuluqaruk (talk) 10:43, 6 December 2018 (UTC)

Claims made in article rhotic consonant

The article says that some accents of Turkish are non-rhotic much in the same way as English or German, i.e. that /r/ is dropped or vocalized syllable-finally. However, I'm not sure about this. Might this be restricted to the ending -iyor (and perhaps some other individual cases)? Or is it really a general phenomenon in some accents? Anyway, this should either be added to the present article or the other article should be corrected. — Preceding unsigned comment added by 2.201.0.62 (talk) 23:34, 19 March 2019 (UTC)

It is definitely also heard in bir, realized as [bi:]. I don't know if this can be generalized to all word-final or even syllable-final /ɾ/s.  --Lambiam 12:27, 8 April 2019 (UTC)

Preconsonantal /v/

According to our article, preconsonantal /v/ in words like "sevda" or "evlilik" would have to be an actual fricative [v]. Clearly it is usually not a fricative, but "sevda" is pronounced something like [seʋda] or [seβda] or perhaps even [seʊ̯da]. At any rate, [sevda] is not the only pronunciation, nor probably a common one. — Preceding unsigned comment added by 2.201.0.62 (talk) 23:54, 19 March 2019 (UTC)

Merge Turkish, Azerbaijani and Gagauz phonology pages into Oghuz Turkic phonology

Merge because most scientifice articles on Oghuz language phonologies cover Istanbul Turkish and Azerbaijani and Gagauz have very similar phonologies. We have better have all it in one place and note the differences rather than scattering across three. Erkin Alp Güney 08:14, 26 October 2019 (UTC)

That's not a good reason to merge phonology pages. Plus, there are no separate phonology pages for Azerbaijani and Gagauz. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 00:26, 27 October 2019 (UTC)

I too see no compelling reason for a merger. Nardog (talk) 00:35, 27 October 2019 (UTC)

Azerbaijani and Gagauz have no separate phonology pages, but they do have page-length phonology sections.Erkin Alp Güney 08:07, 27 October 2019 (UTC)

As they should. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:27, 27 October 2019 (UTC)

Azerbaijani and Gagauz are both Oghuz Turkic languages, and have similiar phonological inventories, but they are also two different languages. There is no way that they should be merged into one page that is strictly, by definition, a sub-branch of the Turkic language family, not a "language". If that is the case, then therefore you might as well merge Norwegian with Swedish and Danish as "North Germanic phonology". It just does not make any logical sense. Fdom5997 (talk) 18:46, 27 October 2019 (UTC)

Compound word stress

"If in Turkish any accent on the non-initial component is lost it is quite different from English."

Not lost, but severely weakened. Much weaker than e.g. secondary stress in English, but not indifferent to neighboring syllables. Erkin Alp Güney 09:13, 1 February 2020 (UTC)

Stress in compounds with multisyllable first element

The section on coumpounds does not seem to treat compounds of whose first element has (at least) two syllables and whose first element has the stress as well. Do(es) the stress(es) remain on the same syllable(s) as where they are when the first element is not part of a compound? More in particular I am wondering on which syllable(s) any stresses in TR “yeşilgöz” [cf. https://sozce.com/nedir/342934-yesilgoz] occur, and if that is different for the nou the family name Yeşilgöz?Redav (talk) 00:51, 29 November 2023 (UTC)

Yeşilgöz as a family name (and also a place name, the phrase "yeşil göz" wouldn't really be used in a non-proper way) keeps the word-final stress found in the first element, "yeşil". So the stress is in the second syllable of the compound.

Also, an example similar to that is listed in the article:

Some surnames have non-final stress: Erdoğan, Erbakan, İnönü, Atatürk (compound) GMFinnegan (talk) 18:20, 30 November 2023 (UTC)

To call it "stress" though is misleading. The accent on the second syllable of "Atátürk" is really a pitch accent not a stress accent. There is a slight rise in pitch, but the second syllable isn't louder. Kanjuzi (talk) 10:50, 1 December 2023 (UTC)

/ɑ/

Moved from User talk:Nardog

It's already quite firmly established with sources in the existing version of the Turkish phonology article that the vowel in question patterns with the other back vowels, and considered back in a phonemic sense in Turkish. – anlztrk (talk | contribs) 14:23, 2 October 2024 (UTC)

Can you name the sources that phonemicize it as /ɑ/? Nardog (talk) 14:25, 2 October 2024 (UTC)

Currently the relevant section reads:

/a/ has been variously described as central [ä] and back [ɑ], because of the vowel harmony. For simplicity, this article uses the diacriticless symbol ⟨a⟩, even in phonetic transcription. /a/ is phonologically a back vowel, because it patterns with other back vowels in harmonic processes and the alternation of adjacent consonants (see above). The vowel /e/ plays the role as the "front" analog of /a/.

So the existing sources both say the phoneme in question is either /ä/ or /ɑ/, but never a front /a/. Furthermore, they explicitly mention that in the vowelharmonic alternation that vowel acts as the back counterpart of the front /e̞/.

This irregularity wouldn't be a big deal if it were limited to the Turkish phonology article itself. However, as a result of the article on the phonology using the symbol ⟨a⟩ instead of ⟨ä⟩ 'for simplicity' [sic], the Turkish IPA key and the countless articles that use Turkish IPA do so too. While Turkish doesn't have the sound [a], it does have a similar sound in [æ], which in Turkish is a positional allophone of /e̞/. So the mere act of removing them runs the risk of suggesting a pronunciation that would be perceived as an entirely different phoneme by Turks. Using the symbol [ɑ̈] for the phone and /ɑ/ for the phoneme avoids this issue entirely. – anlztrk (talk | contribs) 14:39, 2 October 2024 (UTC)

That is not how phonemicization works. English /ʌ/, for instance, is very rarely narrowly [ʌ] but is so phonemicized because that is what people have come to associate and identify the phoneme with. [ä] is most often broadly transcribed with ⟨a⟩, e.g. for German, Hebrew, Italian, Korean, Japanese, and Spanish. One reason is, as the Handbook of the IPA (p. 28) says, "The term 'broad' sometimes carries the extra implication that, as far as possible, unmodified letters of the roman alphabet have been used ... Under this definition a transcription of English hideout as /haidaut/ would be broad, while /haɪdaʊt/ would not be". We must simply follow how reliable sources most commonly phonemically represent a given phoneme, because that's what Wikipedia does—it restates what reliable sources say, and not present original research or novel syntheses. And I see /a/ is phonemicized with ⟨a⟩ in e.g. Zimmer & Orgun (1999), Kornfilt (1997), Göksel & Kerslake (2005), Csató & Johanson (2015), Comrie (1997), and Yavuz & Balci (2011). Unless you show it is more commonly phonemicized with ⟨ɑ⟩ by reliable sources, I see no reason to switch. Nardog (talk) 15:15, 2 October 2024 (UTC)

@Nardog I'm sorry but that's just not true. If all Wikipedia did was to strictly restate, the platform would be entirely made up of quotations and nothing else.

Replacing ⟨a⟩ with ⟨ɑ⟩ isn't original research, because it's not research at all. It's simply a presentation of the existing information already present in the article in a more precise and less misleading way. Changing the symbol distorts nothing, and it synthesizes or creates nothing new. If such clarification were to be banned, tens of thousands of entire articles would be needed to be entirely rewritten or removed. – anlztrk (talk | contribs) 15:36, 2 October 2024 (UTC)

What should motivate us to deviate from the notational convention that successfully works for the vast majority of reliable sources? –Austronesier (talk) 16:10, 2 October 2024 (UTC)

The vast majority of reliable sources have in depth explanations accompanying that convention, clarifying the precise sound the notation /a/ represents, which the reader would naturally be expected to have also read. As the article here serves as the direct source of the pronunciation key located at Help:IPA/Turkish, which is a creation of Wikipedia itself, it also indirectly leads to mispronunciation of countless Turkish proper names on Wikipedia articles where that key is used, which could be read by people who are not necessarily familiar with Turkish phonology but are familiar with the International Phonetic Alphabet.

The IPA key at Help:IPA/Turkish is what should be fixed. To fix that, first this article should be altered. – anlztrk (talk | contribs) 18:32, 2 October 2024 (UTC)

No, we summarize and paraphrase sources.

There can be occasions where it comes down to a judgment call, e.g. for an underdocumented language about which only sources that do not use IPA exist, but when it comes to languages as widely spoken and documented as Turkish, phonemes are like proper nouns: if reputable sources consistently refer to something as /a/ and we refer to it as /ɑ/, readers will be confused. Nardog (talk) 00:46, 4 October 2024 (UTC)

References

Palatal and palatalized are not the same

The Turkish "soft" 'k' phoneme in words like 'küçük' (commonly, but erroneously transcribed as /cyˈtʃyc/) is not a true palatal sound, it is a palatalized plosive (at best) or just a fronted velar plosive. I do not understand why this misconception is perpetuated. 2001:4C4E:2484:DA00:1E30:5AEE:462D:875E (talk) 10:43, 18 December 2024 (UTC)