Talk:Modern Greek phonology

Consonants
"The series of voiced plosives can be analysed as sequences of nasals and voiceless plosives, [b] = /mp/, [d] = /nt/, [g] = /ng/. Again, this corresponds to the orthographic spelling (using digraphs <μπ, ντ, γκ>)."

/ng/ is _not_ "a sequence of a nasal and a voiceless plosive"; if someone wants to dispute "g"-s status as a phoneme, they need to claim that "[g] = /nk/" (or /ŋk/ if Greek had an "ŋ" phoneme, but it doesn't). --Adolar von Csobánka (Talk) 23:50, 3 February 2007 (UTC)


 * Yep, that should have been /nk/, certainly. Fut.Perf. ☼ 23:53, 4 February 2007 (UTC)

Em, you are wrong. The ŋ phoneme surely exists, in words such as άγχος (/'aŋxos/). BionicMadness (talk) 07:23, 20 May 2018 (UTC)

u
This might not be the right place to ask this, but is there a vowel, or a diphthong in ancient or modern greek which makes an "oo" sound? The Plutach and Pluto pages use "ou" but does that make the same "oo" sound that the "u" would make in english?167.206.69.62 (talk) 15:10, 20 November 2008 (UTC)Dali


 * In Modern Greek, the letter combination <ου> (omikron-upsilon, transliterated as ou) is indeed an [u] sound, roughly like English oo in root. In early classical Greek, there was a stage when this letter combination was still pronounced more like [o:] or [ou] (the oa in boat); in its place, the letter <υ> (upsilon) alone had the value [u]. During classical Greek, words with <υ> changed to a pronunciation like German ü, and later to [i]; while words with the <ου> combination raised to the present [u] value. Fut.Perf. ☼ 09:44, 22 November 2008 (UTC)

Sandhi
I have removed a passage that was inserted in the "Sandhi" section:


 * This phenomenon is not a part of standard taught rules and may vary according to dialect, speed and formality of speech. Proper, sandhi-less pronunciation is often voluntary and requires attention and slow speech. Since sandhi is highly variable depending on circumstance, to the degree of being completely absent in formal speech, foreign speakers who are not comfortable with everyday usage may just avoid it altogether and prefer pronunciation "by the book" instead.

I'm afraid this passage was simply wrong. The description of sandhi may come as a surprise to native speakers, as it is a low-level and low-awareness phenomenon, and many native speakers may not be aware that they are actually doing this – but they are. All of them, all the time. I have never, ever heard a Greek pronounce [ton patera] literally, with a fully unvoiced [t] and a fully alveolar segmental [n]. Yes, it's not part of "taught rules" - because it doesn't need to be, because everybody does it automatically. All the linguistic descriptions I've read agree in describing this as an automatic, obligatory process in natural speech.

The advice to non-native learners that they should just omit the sandhi realisation seems to me to be very poor advice indeed, as it will lead to a very foreign accent rather than a proper "formal" one. The implication that the sandhi-less pronuciation is the more "proper" one is quite misleading too, in my view.

Fut.Perf. ☼ 09:58, 7 January 2009 (UTC)

==Palatalized pronunciation [for "li" and "ni"] presupposes the presence of yet another vowel after the palatalized consonant and its following /i/. If there is no subsequent second vowel, palatalization does not occur.==

In the pronunciation of a good many Greeks, it does indeed. My experience is that these palatalized pronunciations (without another following vowel) are somewhat out of favor (particularly in the case of "li") and may be recessive. Kostaki mou (talk) 04:53, 18 February 2009 (UTC)

ευ
How exactly is "ευ" pronounced today, is it /eu/, /ew/, /ev/, /ef/? Dan Pelleg (talk) 11:15, 14 March 2009 (UTC)


 * It's [ev] (like in English l ev el) in most words, but [ef] when another voiceless consonant follows. Fut.Perf. ☼ 15:20, 14 March 2009 (UTC)
 * thanks. So ευρώ is /evˈro/? Dan Pelleg (talk) 15:47, 14 March 2009 (UTC)
 * Yes. Fut.Perf. ☼ 15:59, 14 March 2009 (UTC)
 * Ta! [[image:SMirC-wink.svg|17px]]

Is the υ part of the syllable coda, or onset? That is, is the vr of Ευρώπη pronounced the same as the vr of Λαβράνδα ? kwami (talk) 00:00, 31 July 2009 (UTC)


 * Yes it is.Kostaki mou (talk) 01:00, 31 July 2009 (UTC)


 * Okay, good to know. I just edited the Greek place names article with that assumption. kwami (talk) 01:02, 31 July 2009 (UTC)

Word-initial clusters
I don't know for sure how many places the chart for the word-initial consonant clusters is wrong, but I did see that it doesn't have /fθ/ as in φθινοπωρο or /xθ/ as in χθές. Maybe someone with access to the work cited can double check that its been copied accurately here. 97.116.171.22 (talk) 12:51, 22 March 2012 (UTC)
 * Thanks for pointing this out. Unfortunately, the presentation in the article is in accordance with the cited source (Joseph & Philippaki-Warburton 1987), which omits /xθ/ and /fθ/. Probably, the authors were thinking in terms of conservative Δημοτική, where these would generally be absent in favour of /xt, ft/ (χτες, φτινόπωρο). However, I think any observer of present-day Standard Modern Greek would agree that the /xθ, fθ/ variants are today a fully established part of spoken Standard Modern Greek. Indeed, J & P-W are somewhat inconsistent in this respect insofar as in their grammar chapters they repeatedly include examples involving the word χθες, transcribed with /xθ/.
 * We'll probably just have to find a more up-to-date source to fix this. Fut.Perf. ☼ 13:24, 22 March 2012 (UTC)
 * Update: Holton/Mackridge/Philippaki-Warburton (1997: ch.1.2.1) have a much better treatment of phonotactics. I'll rework the section on that basis as soon as I find the time. Fut.Perf. ☼ 09:43, 8 November 2012 (UTC)

Geminates
Does Modern Greek have geminate consonants? Ancient Greek does; if Modern Greek doesn't, this is a significant difference that should be mentioned. — Eru·tuon 22:55, 30 June 2012 (UTC)
 * Modern Greek doesn't. There is already a sentence about it at the end of the "consonants" section, but perhaps it could be expanded a bit. Fut.Perf. ☼ 23:29, 30 June 2012 (UTC)
 * Only Cypriot Greek has geminate consonants -Myrmecium (talk) 13:57, 19 August 2018 (UTC)

Consonant clusters that begin with r
The table is missing ρ- clusters. I'm very new to Greek, so I might be missing something. I'm getting my example from the verb έρχομαι/ήρθα. In a song ένα τσιγγανάκι είπε / Νότης Σφακιανάκης, the verse goes: Θα 'ταν όλα ίδια, ίδια με το χθες μα ήρθες και με καις. To me is sounds that he does pronouce ρθ each with the usual sound of the letter, as a consonant cluster. --Itaj Sherman (talk) 19:48, 7 November 2012 (UTC)
 * Oh, I suppose they would formally consider the ρ belonging to the previous syllable. And they do state 5 possible codas for most native words. Is that misleading? --Itaj Sherman (talk) 20:10, 7 November 2012 (UTC)
 * According to the normal phonological principle of "maximal onset", the syllable division would certainly be [ir.θes], not [i.rθes]. There are no [rθ] or similar clusters word-initially, so they wouldn't be found in a word-internal syllable onset either. Fut.Perf. ☼ 21:35, 7 November 2012 (UTC)

Rewrite/improve
I've added some recent material under the 'Further reading' heading if anybody feels like improving this article. Arvaniti (2007a) would be a good starting point. — Lfdder (talk) 21:35, 9 April 2013 (UTC)

I've made quite a few changes/additions, but there's still a lot more that could be said. There's nothing on phonotactics or acquisition, and there's very little about phonological processes. The section on sandhi is a bit of a hit and miss. — Lfdder (talk) 19:15, 12 April 2013 (UTC)


 * Thanks for all this work. I'm afraid I find the new consonants section a bit too abstract and not optimally structured though. Especially the sentence in the beginning, "The table below is adapted from Arvaniti (2007a, p. 7), who does away with the entire palatal series, both affricates and the velar nasal /ŋ/" is rather problematic on several counts. First, it doesn't really make sense to tell the reader that somebody "does away with" something as long as the reader has no prior idea that such elements exist in the first place. What "both affricates" and what palatal series are those? Also, I'm not sure how /ŋ/ found its way in there. Has anybody ever proposed there was a phonemic /ŋ/? (apart from the IP that added it to this article in an unexplained edit some time ago, which unfortunately slipped through.) I'd tend to think it's pretty obviously not a phoneme in Greek.
 * The passage also now relies a bit too exclusively on Arvaniti. I'm not saying it's not a good source, but for instance I see no strong reason to emphasize her personal opinions, like in saying that she is reluctant to accept phonemic affricates. The same point, that the status of the affricates is doubtful, can also be sourced to the standard reference grammar by Holton/Mackridge/Philippaki-Warburton. Why was the reference to that work removed completely?
 * I'd also like to see the connection to the spelling system reintroduced. Not that it's centrally a part of what a phonological analysis is about to the expert reader, of course, but it just serves to make the material much easier to understand for the lay reader if they can connect the phonological symbols to the letters by which they will typically know these sounds. Fut.Perf. ☼ 22:48, 15 April 2013 (UTC)
 * Thanks for the feedback. I don't think that the opening paragraph is particularly problematic. It's not meant to hold a lot of weight, really; it's just a by-the-way of sorts. I've clarified 'affricates' and removed /ŋ/.
 * >I see no strong reason to emphasize her personal opinions, like in saying that she is reluctant to accept phonemic affricates
 * That was to explain why they're missing from the table.
 * >Why was the reference to that work removed completely?
 * I've removed it 'cos the page wasn't specified so I couldn't verify it. I'd be fine with putting it back in (with the page number).
 * >I'd also like to see the connection to the spelling system reintroduced ...
 * Help:IPA for Greek does an all right job connecting the letters with the sounds, I think. I'm not convinced that whole section would be worth reinstating. How about including the table from Arvaniti 1999, p. 2?
 * >... make the material much easier to understand for the lay reader if they can connect the phonological symbols to the letters by which they will typically know these sounds ...
 * Well, that's only if that particular lay reader speaks Greek. ;P — Lfdder (talk) 23:26, 15 April 2013 (UTC)

Vowels
I find the section on vowels somewhat misleading. The system described seems to be overly idealised - perhaps Arvaniti is enamoured with the idea of a simple 5 cardinal system? - but looking at other sources (and listening to spoken Greek), the situation is much more complicated. I'm no expert on Greek though I am a linguist by training and I really struggle to see (and hear) any /e/ vowels and even simply looking at the vowel chart, the nearest vowel symbol seems to be /ɛ/, not /e/. Similarly, to me /o/ seems to be much closer to /ɔ/. Akerbeltz (talk) 11:13, 16 October 2013 (UTC)
 * Well, yes, the front one is certainly closer to cardinal [ɛ] than to cardinal [e], and the same may very well be true for the back one too (keep in mind though, as a German native speaker, that IPA cardinal [e] and [o] are still not quite as close as our German /e/ and /o/). But in any case, in a five-vowel system like this, it is conventionally legitimate in phonemic notation to use the simpler 〈e〉 and 〈o〉 symbols for any of the roughly "mid" vowels, no matter what their precise phonetic realization is. Fut.Perf. ☼ 16:40, 16 October 2013 (UTC)


 * The phonetic realisation is different, but how is the situation "much more complicated"? Do you mean that the realisation tends to vary? — Lfdder (talk) 17:21, 16 October 2013 (UTC)


 * It may be legitimate in a phonology textbook but this is supposed to be an overview over Greek phonology to people who are NOT phonologists and for such, it's questionable suggesting that the vowel is more /e/ than /ɛ/. And with "more complicated" I mean that there seems to be, at best, a rather wide range of vowel allophones but I can't see any mention of that. Anywhere on the page. Akerbeltz (talk) 00:25, 17 October 2013 (UTC)
 * I don't think anybody would object if you changed the symbol to /ɛ/. There's some things about vowel allophony in Arvaniti 2007. — Lfdder (talk) 01:38, 17 October 2013 (UTC)
 * I'd actually prefer continuing with using /e/ and /o/. The argument about our readers not being phonologists can be turned the other way round: "e" and "o" are simply much more familiar and easier to understand, and most readers don't even know the difference between "e" and "ɛ" anyway. Plus, there might be additional confusion because when they see "ɛ" in this context, they might think it's simply the Greek letter and might wonder why we are using a Greek letter for this one vowel but not for the others. Our IPA transcriptions throughout Wikipedia, as documented at Help:IPA for Greek, also use "e" and "o". Fut.Perf. ☼ 06:38, 17 October 2013 (UTC)
 * I'm bowing out at this stage. I've raised the issue and I still think it's not sensible but I don't have the time, expertise or sources to argue about this particular one and I can feel this turning into one of "those" debates :) Just a minor point in parting - the IPA transcriptions for languages on Wikipedia have been known to be wrong. Akerbeltz (talk) 10:00, 17 October 2013 (UTC)

Palatals
This sentence needs clarification: "Greek possesses palatals [c ɟ ç ʝ] that are distinct [from /k g x ɣ/, presumably, rather than from each other] before /a/, /o/ and /u/, but in complementary distribution with velars before front vowels /e/ and /i/ [What is the conditioning? Presumably this means that velars don't occur before /e/ and /i/. Is that correct?].[9] [ʎ] and [ɲ] occur as allophones of /l/ and /n/, respectively, in CJV (consonant–glide–vowel) clusters, in analyses that posit an archiphoneme-like glide /J/ that contrasts with the vowel /i/ [If you're going to mention this, it needs explanation. What does /J/ sound like, what effects does it have? In what way does it contrast with /i/?].[10] All palatals may be analysed in the same way. " Linguistatlunch (talk) 22:02, 31 March 2014 (UTC)
 * Well, yes, distinct from velars. Yes, only the palatals occur before /e i/ -- a clarification there would be a welcome addition. /J/ does not occur phonetically; it palatalises the preceding consonant. I don't remember what the source said about it contrasting with /i/ specifically -- I'll look at it later. Thanks for your comments. — lfdder 22:26, 31 March 2014 (UTC)
 * Are these allophones always palatal in Standard Modern Greek, or may they regionally be found articulated as alveolo-palatals or (especially in Cyprus) palato-alveolars, as in some of the traditional dialects (including Cypriot Greek)? (For example, και might be pronounced as [tɕe] or [tʃe] instead of [ce].) My understanding is that most traditional dialects are now effectively extinct, almost extinct or moribund, but varying articulations such as these might reasonably be retained and carried over into Standard Modern Greek (unless they are perceived as excessively divergent). --Florian Blaschke (talk) 18:48, 6 November 2015 (UTC)
 * Well, the two main dialects that are "famous" for shifting the palatal phonemes to affricates or post-alveolars (like the example you mentioned, where και instead of /ce/ becomes /tɕe/ or /tʃe/) are Cretan Greek and Cypriot Greek. Those two dialects face no danger of getting extinct whatsoever, at least not in the near future, so any attempt to retain their phonemes would be unnecessary. -Myrmecium (talk) 13:48, 19 August 2018 (UTC)

Where is ξ?
Seriously, ξ shows up in the example text, but has no demonstration that it's a phoneme above, let alone how to pronounce it. "Everyone knows how to pronounce the missing Greek letter(s)." Yes, well same with English, but the English Phonology page goes into great detail covering all of the phonology of English. --66.235.63.70 (talk) 18:46, 23 July 2015 (UTC)
 * That's because "ξ" is not a phoneme, but a letter, and this article is not about the writing system but about the sound system. Those are two different things. The sound value of "ξ" happens to be not a phoneme of its own but simply a cluster of the phonemes /k/ and /s/. (The information on how to pronounce specific letters and letter combinations of the alphabet, including those like "ξ" and "ψ", doesn't belong in this article but in that on the Greek alphabet. Fut.Perf. ☼ 20:10, 23 July 2015 (UTC)

Accented on the third-from-last syllable
The example given ("monks": "των καλόγερων") i think that it maybe wrong, it's still pronounced : "των καλογέρων" even in modern Greek, though i don't know if it is just a local variant that remained which does not respond with the Athenian way of speaking today the modern language. — Preceding unsigned comment added by DiogenesLaertios (talk • contribs) 19:43, 5 August 2017 (UTC)

Basically, katharevousa and formal forms of standard Greek (and maybe some dialects, but I am not entirely sure) follow the rule that says "the third-from-the-last syllable cannot be accented if the last one is long". Informal (Colloquial) forms of the language tend to disregard that "rule" BionicMadness (talk) 07:33, 20 May 2018 (UTC)

External links modified (February 2018)
Hello fellow Wikipedians,

I have just modified 6 external links on Modern Greek phonology. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20160303222353/http://www.kent.ac.uk/secl/ell/staff/amalia-arvaniti/docs/Illustrations%20%20of%20the%20IPA%20-%20Modern%20Greek.pdf to http://www.kent.ac.uk/secl/ell/staff/amalia-arvaniti/docs/Illustrations%20%20of%20the%20IPA%20-%20Modern%20Greek.pdf
 * Added archive https://web.archive.org/web/20131211020607/http://www.kent.ac.uk/secl/ell/staff/amalia-arvaniti/docs/Greek%20Phonetics%20-%20The%20State%20of%20the%20Art.pdf to http://www.kent.ac.uk/secl/ell/staff/amalia-arvaniti/docs/Greek%20Phonetics%20-%20The%20State%20of%20the%20Art.pdf
 * Added archive https://web.archive.org/web/20160303233211/https://dl.dropboxusercontent.com/u/11236687/TopintziBaltazaniISTAL2011.pdf to https://dl.dropboxusercontent.com/u/11236687/TopintziBaltazaniISTAL2011.pdf
 * Added archive https://web.archive.org/web/20130515110353/http://icgl7.icte.uowm.gr/Adaktylos.pdf to http://icgl7.icte.uowm.gr/Adaktylos.pdf
 * Added archive https://web.archive.org/web/20160303224302/http://www.kent.ac.uk/secl/ell/staff/amalia-arvaniti/docs/Greek%20wh-questions%20and%20the%20phonology%20of%20intonation.pdf to http://www.kent.ac.uk/secl/ell/staff/amalia-arvaniti/docs/Greek%20wh-questions%20and%20the%20phonology%20of%20intonation.pdf
 * Added archive https://web.archive.org/web/20070302095746/http://www.ling.ohio-state.edu/~gdanelis/Variationstypologie.pdf to http://www.ling.ohio-state.edu/~gdanelis/Variationstypologie.pdf

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 08:34, 3 February 2018 (UTC)

Stress mark missing in /apo/?
In the transcription of "The North Wind and the Sun", /apo/ has no stress mark despite the second syllable of από being stressed. Is this an oversight or is it meant to show that the word από is unstressed when reading the text? &mdash; 89.197.103.111 (talk) 15:51, 23 May 2018 (UTC)
 * In Greek some words, usually those with one to two syllables, despite having a stress mark, will sound like having no stress at all. This is the case with από, as well. Από will sound like it is part of the following word. Example: Greeks when they want to say "από εκεί" or "από 'κεί" will sound like /apoε'ki/ and /apo'ki/ respectively and not /a'po ε'ki/ and /a'po 'ki/ -Myrmecium (talk) 13:28, 19 August 2018 (UTC)

Greek indeed has an /ŋ/ phoneme
I have seen throughout the talk section of this article that the /ŋ/ phoneme does not exist in Greek. However, I have to disagree with this opinion. This phoneme is usually pronounced in cases like "άγχος" or "ελέγξω" where we say /'aŋxos/ and /e'leŋkso/ respectively. Trying to pronounce them as /'anxos/ or /e'lenkso/ sounds unnatural to a native speaker (and it is difficult to pronounce as well). I would like someone to explain to me how this idea of /ŋ/ not existing in Greek was formed. -Myrmecium (talk) 14:11, 19 August 2018 (UTC)
 * That phone exists as a positional allophone of the /n/ phoneme, but not as a phoneme on its own. There are no minimal pairs like /ˈanxos/ and /ˈaŋxos/, or /ˈnomos/ and /ˈŋomos/, which would display a contrast between /n/ and /ŋ/. --Theurgist (talk) 17:48, 19 August 2018 (UTC)
 * Oh, my bad. I thought the people [who said that /ŋ/ isn't a Greek phoneme] meant that /ŋ/ doesn't exist at all in Greek. -Myrmecium (talk) 19:20, 19 August 2018 (UTC)

Non-Retracted /z/ exists as well
In the consonant section I have noticed that there is only mention of /s/ being variably retracted, but this also happens to /z/ in Greek, which can be like English "zoo" or close to "measure" depending on the environment. Similar mention should be shown in the affricate clusters, like in the IPA help section. — Preceding unsigned comment added by EndlessReading (talk • contribs) 01:26, 11 April 2022 (UTC)

Vowel length distinction
Hi there, I apologize if there's a self-evident answer to my question that I'm overlooking, but wouldn't ζω and ζώο be a minimal pair for vowel length, at least for the /o/ phoneme? HAmitySandwich (talk) 22:18, 21 June 2023 (UTC)
 * Hi, thank you for your post. Unfortunately, the talk page of this article is for discussions about how to improve this article based on reliable sources, and not for general discussions/inquiries related to the topic (WP:NOTFORUM). But since it is a really cool question, I suggest you bring it to Reference desk/Language. I'm sure the regulars of the reference desk will be delighted to discuss it. –Austronesier (talk) 18:07, 22 June 2023 (UTC)
 * It's completely on topic here as it's about something we could potentially mention in the article. Nardog (talk) 18:12, 22 June 2023 (UTC)
 * If we can dig up something from a good source, sure. Basically, we need to see if the old question whether a morphophonemic like-vowel sequence can be treated as a single phonemic segment has been discussed somehwere in the literature about Modern Greek phonology. –Austronesier (talk) 18:15, 22 June 2023 (UTC)
 * (ec) I agree it's a valid question. The answer, I think, is that ζώο would be considered two separate syllables, hence not a single (long) /o:/ phoneme. In support of this, one could argue that either of the two "halves" of the o sound can be the carrier of phonological stress (compare ζώο with stress on the penultimate, versus εννοώ 'I mean', with stress on the ultimate. Not sure if there's an actual minimal pair, but there might be.) Haven't got a reference handy though. Fut.Perf. ☼ 18:24, 22 June 2023 (UTC)