Module talk:pl-IPA

Definition from Wiktionary, the free dictionary
Jump to navigation Jump to search

/a/ vs /ä/[edit]

@Korn: /ä/ is only used when it is necessary to distinguish between /a/ and /ä/ as phonemes or in narrow transcriptions where the difference is necessary to point out. Your edit makes the transcriptions at galaktyka pointlessly ugly. --WikiTiki89 15:44, 30 April 2018 (UTC)[reply]

@Wikitiki89 I've always opposed this imprecision generally and it, correct me if I'm wrong, has no history in en.Wiktionary's Polish entries. Aesthetics aside, Polish <a> is central, even in palatal contexts. Korn [kʰũːɘ̃n] (talk) 15:47, 30 April 2018 (UTC)[reply]
@Korn: Right, /a/ can represent either a central or front vowel, depending on context. In the context of Polish, it's a central vowel. As for a history of using /a/ in en.wikt Polish entries, staying with the example galaktyka, the pronunciation section there has used /a/ since it was first added to the entry in 2008 (diff), and has been that way ever since until just now when you changed the output of this module. I'm going to revert your change until you can demonstrate that /ä/ actually has a history of being used consistently in our Polish entries. --WikiTiki89 15:56, 30 April 2018 (UTC)[reply]

⟨y⟩[edit]

On a different note, ⟨y⟩ is universally pronounced with the sound transcribed [ɪ̞] or [ɘ̘] (cf. Ukrainian [ɪ] for the same vowel) , why are we transcribing it as ⟨ɨ⟩? Korn [kʰũːɘ̃n] (talk) 16:39, 30 April 2018 (UTC)[reply]

Perhaps because it's a phonemic transcription and phonemic transcriptions typically avoid diacritics. — Eru·tuon 23:34, 3 May 2018 (UTC)[reply]
@Korn: Right, basically same reason as above. Do you have any sources to support [ɪ̞] or [ɘ̘] anyway? We've always used /ɨ/, and so does Polish Wiktionary. --WikiTiki89 14:39, 4 May 2018 (UTC)[reply]
Personally only from experience, but copying the sources Wikipedia uses: Bronisław Rocławski (→ISBN), Wiktor Jassem (page 106), Edmund Gussmann (Page 2 (Arabic)). All of these transcribe as ⟨ɨ⟩ out of Slavistic tradition, but in terms of data there is no disagreement whatsoever that ⟨y⟩ is a centralised close-mid sound, or lower than near-close anyway. While I totally disagree about conflating [a] and [ä], that one I get, they're sort of close, close enough that the IPA couldn't be bothered to install a non-combined glyph for [ä]. But this case here...There would need to be made a pretty good argument as for why we should basically deceive our users about the pronunciation of ⟨y⟩. Korn [kʰũːɘ̃n] (talk) 15:18, 4 May 2018 (UTC)[reply]

Brackets missing[edit]

@Wikitiki89, Erutuon, IvanScrooge98, Esszet At argot, {{pl-IPA|ar'go}} is producing a form with neither brackets nor slashes around it, which puts it in CAT:IPA pronunciations with invalid representation marks. Can this be fixed? —⁠This unsigned comment was added by Mahagaja (talkcontribs) at 13:11, 1 February 2019 (UTC).[reply]

@Mahagaja: If I search hastemplate:"pl-IPA" insource:/\{\{pl-IPA\|/, it looks like the template expects the user to supply a word with slashes around it. That doesn't make sense to me (it shouldn't be up to the editor to choose the bracketing, because most editors have no way to know whether the module will generate a phonemic or phonetic transcription), and I think the module can automatically supply slashes when they are missing. — Eru·tuon 21:25, 1 February 2019 (UTC)[reply]
@Erutuon: Especially since the template does automatically provide slashes when nothing is specified in parameter 1, e.g. at użądlenie where {{pl-IPA}} is used without a parameter and the slashes are there. —Mahāgaja · talk 21:29, 1 February 2019 (UTC)[reply]
@Mahagaja: Okay, done. Some transcriptions may have square brackets, so the module checks for those as well as for slashes before automatically adding slashes. — Eru·tuon 21:32, 1 February 2019 (UTC)[reply]

remove allophones and correct palatalized velar consonants[edit]

@Korn, Wikitiki89, Erutuon, IvanScrooge98, Esszet Since there are slashes around the transcriptions, they should be phonemic transcriptions. According to the Polish phonology page, ki gi chi before vowels other than i are /kʲ ɡʲ xʲ/, not /kʲj ɡʲj xʲj/, while pi bi fi wi mi ri li before vowels are /pj bj fj vj mj rj lj/. Also, voicing of consonants before voiced consonant is not reflected on the module, and rz after voiceless consonants should be treated differently from ż in that position. JeanneAymonier (talk) 04:17, 5 March 2020 (UTC)[reply]

@Tweenk You're right about voiced consonants, I just tried to add a "voice" section (sorry, I don't know what it's called, whatever the "devoice" section is called), but my browser (I think) isn't rendering affricates properly, and I don't know how to do loops anyway. I guess you're also right about "ż" (I'm far from fluent, so I can't think of any examples off the top of my head), but I don't know how to do that, either. I personally think pi, bi, wi, and fi are truly palatalized and there is a /j/ after palatalized velar stops (again, I'm far from fluent), but you'd better talk to a fellow native speaker about that. Esszet (talk) 00:39, 8 March 2020 (UTC)[reply]

Updating the module[edit]

@Shumkichi Here is where can update the IPA for Polish.

@Shumkichi If you write out the changes you want to make, I'm sure we can find someone to update them. It was adding /j/ to palatalized consonants, right? anything else? Vininn126 (talk) 12:50, 21 August 2021 (UTC)[reply]
@Vininn126 It's hard to come up with something just like that but the template seems to have a problem with ę and especially ą. For example, "są" is /sɔ̃/ but should be /sɔm/ (or, if you want to be really petty, [sɔ̃m] but let's ignore this one as it's too detailed); "wąż" should be /vɔ̃w̃ʂ/ and "mężczyzna" /mɛ̃w̃ʂˈt͡ʂɨz.na/. I get the reason for not including the nasalised semi-vowel where necessary but the thing is that nasal vowels NEVER occur alone in Polish in any context, they are always assimilated in some way to the surrounding consonants. I don't think it'd be too petty to show all the allophones of ą and ę as we do so with /n/, for example it's /pɔŋk/, not /pɔnk/. It's just antiquated to treat ą and ę as pure nasal vowels but we've already discussed it. I don't know what else; we should remember about being consistent with /j/: so any soft consonant followed by /j/ + any vowel other than /i/ should be of the following form: /Cʲj/, e.g. /'kʲjɛ.dɨ/, but when followed by /i/, it should be reduced to /Cʲ/, e.g. /'kʲi.nɔ/. Shumkichi (talk) 04:30, 22 August 2021 (UTC)[reply]
EDIT: I've got another one: in "cyngiel", for example, /n/ is incorrect. It's incorrectly displayed as /ˈt͡sɨn.ɡʲɛl/ while it should obviously be /ˈt͡sɨŋ.ɡʲjɛl/. [ŋ] occurs as an allophone of /n/ before every /k/, /g/, and /x/, no matter if it's within the same syllable or at syllable boundaries. Shumkichi (talk) 06:55, 22 August 2021 (UTC)[reply]
@Shumkichi So essentially, word final ą goes to /ɔm/, and before retroflex sibilants nasals goes to /Vw̃/. How do you feel we should approach nasals before /s/ and /z/, given that they tend to allow both the /Vw̃/ and the /Vn/ pronunciation? And then palatals before /a/, /ɛ/, /ɔ/, /u/ acquire /j/, and nasals assimilate? I think they talked about nasal assimilation above, might want to give that a read. Vininn126 (talk) 10:32, 22 August 2021 (UTC)[reply]
"before /s/ and /z/, given that they tend to allow both the /Vw̃/ and the /Vn/" - really? I mean, maybe, but I can't think of any such word that would be realised as /Vn/. I would pronounce both "wąs" and "kąsać" with /ɔ̃w̃/ and "kęs" with /ɛ̃w̃/ but maybe both pronunciations are in free variation. I will read up on it. Shumkichi (talk) 11:01, 22 August 2021 (UTC)[reply]
This is based on what I've heard. Go to forvo and listen to the recordings of język, one guy even does both pronunciations in the same recording. Also I just realized something - how narrow should we be doing these transcriptions? This conversation came up on the discord recently with English pronunciation and I'm wondering how much we should. Specifically word final -ą, I wonder if /ɔ̃w̃/ would suffice. Vininn126 (talk) 11:04, 22 August 2021 (UTC)[reply]
@Shumkichi Oh, also adding voiceless r, right? Should we add other voiceless consonants such as voiceless /n/, /l/, or /w/? (Such as pleśń, motocykl, or plótł). Or would that be too precise? Vininn126 (talk) 09:04, 24 August 2021 (UTC)[reply]

Palatalization and nasals[edit]

@Vininn126 @Shumkichi @Tashi I wanted to ask what do you think about removing the palatalization and changing the way <ą> and <ę> are transcribed in the module.

As I've already mentioned replying to Vininn on my talk page, I think we could replace /Cʲj/ with /Cj/. Though the /Cʲj/ transcription reflects its pronunciation, I think replacing it with /Cj/ could make the transcriptions easier to read (/mjɛt͡ɕ/ takes less time to read than /mʲjɛt͡ɕ/ (especially since "ʲ" is a small symbol and can require some focus to read on smaller displays)) and could bring more consistency with other parts of module, as it currently doesn't seem to palatalize words where palatalization is represented by a <j> (f.e. objawiać). Since transcriptions generated by the module are phonemic, I don't think making the change would make the transcriptions any less accurate, as both /pjɛs/ and /pʲjɛs/ represent [pʲjɛs].

I also want to suggest a change of how the nasal vowels are transcribed with the module. Since they're not really nasal vowels, I think we could use either the /Vw̃/ (which is a pretty common way to describe them, but I'm pretty sure it would also require us to update how the module deals with <n> before fricatives (as "kęs" [kɛw̃s] and "sens" [sɛw̃s] are rhymes)) or the /Vŋ/ (which doesn't add an additional phoneme and avoids the trouble of dealing with <n> before fricatives as "kęs" would be /kɛŋs/ and sens would be /sɛns/) way of transcribing them. Though it's not really an issue (the transcriptions are after all phonemic and /ɔ̃/ and /ɛ̃/ stand for [ɔw̃] and [ɛw̃] anyway), I think making the change would make the transcriptions more accurate without sacrificing their readability.

I've already tried to implement the suggested changes yesterday, though they were rightfully reverted, as I didn't discuss them before doing so. Mazab IZW (talk) 00:19, 22 November 2021 (UTC)[reply]

@Mazab IZW, well, Shumkichi recently got banned. I'm going to be pinging @Hythonia as our other resident Pole, as well as @KamiruPL. Also tagging @BigDom and @Hergelei.
I'm gonna explain what happened here: I recently added the /j/ to soft consonants, and many other (including non-Polish editors) agreed with the decision, but I also left the /ʲ/'s, at the advice of one Shumkichi back when I was working on the rhymes pages. Now I personally wouldn't be opposed to removing the palatalization markers as you did before, as they are more phonetic than phonemic it seems to me, and the main thing would be the /j/'s. If we do so, we're gonna have to mass update the rhymes we've been adding and change their names, or at least I'd like to. Perhaps someone would be willing to make a bot to do that for us. We would also have to update the test cases on both the main module page and the module's sandbox but that's trivial.
As for the nasals - your choice again wasn't exactly wrong, but I think the exact implementation was. Polish nasals appear only at the end (except ę) and before sibilants, otherwise they assimilate. Your original change had it as a nasal vowel + ...what was it, a velar nasal? Which is not the standard - they are mostly a nasal vowel + a nasal w, so /ɛ̃w̃/, /ɔ̃w̃/, however these are also not exactly phonemic, but they are basically almost always pronounced that way. This is a change I personally see as less important. If we implemented it, we'd again have to mass update the rhymes and test cases.
I should also mention that whenever changes like this are made, many problems start popping up, so get ready to start fixing those :p Vininn126 (talk) 09:52, 22 November 2021 (UTC)[reply]
Ah, I wasn't aware that removing the palatalization markers would cause problems/inconsistencies with the rhymes. Yeah, this is something we'll need to take care of if we do end up making the change (though it seems pretty time-consuming to do manually).
And I agree the change with the nasals is not really that important. As for my edit, I used the velar nasal as I remember reading somewhere that [w̃] is just its positional variant occurring word-finally and before fricatives. Though it's entirely possible that I messed something up so in this case thank you for reverting! Mazab IZW (talk) 14:44, 22 November 2021 (UTC)[reply]
I don't think that transcribing kęs as /kɛw̃s/ looks ok. It's pronounce more like /kɛns/. Tashi (talk) 16:14, 22 November 2021 (UTC)[reply]
@Tashi I have noticed that also with /s/ and /z/. język is often /'jɛn.zɨk/. It seems to only be in front sz, ż, and ch. Vininn126 (talk) 17:06, 22 November 2021 (UTC)[reply]
@Vininn126 As for język being often pronounced /'jɛn.zɨk/, I checked out some YouTube videos I found after searching "język polski" and "język polski lekcje" (1, 2, 3, 4 - with timestamps) and they all seem to be /'jɛw̃.zɨk/. Maybe the /n/ pronunciation is either dialectal or perhaps a form of hypercorrectness? Because aside from one Forvo recording, I can't really recall hearing it anywhere, especially in movies/everyday speech. (Though as I said before, I'm just talking about my own experience. If you have any examples of it being pronounced with a /n/, feel free to share them :) ) Mazab IZW (talk) 17:34, 22 November 2021 (UTC)[reply]
Native speakers will also not assimilate in slow, careful speech. In fast speech /ns/ and /nz/ are more popular. Vininn126 (talk) 17:42, 22 November 2021 (UTC)[reply]
@Vininn126 Ahh, I see. Perhaps leaving them as /ɛ̃/ and /ɔ̃/ is a better idea since the pronunciation doesn't seem to be completely consistent.
Anyway thank you so much for all your help! Mazab IZW (talk) 17:58, 22 November 2021 (UTC)[reply]
@Tashi That's interesting, the audio samples from both Forvo and Polish Wiktionary seem to be [kɛw̃s]. This also seems to be the case for other words like dąsać or wąs. I'm not sure if I remember ever encountering the [kɛns] pronunciation either, I often hear Polish speakers pronounce /ns/ (even when it's spelled <ns>) as [w̃s] (sens, balans). Perhaps the [ns] pronunciation is a dialectal thing? Mazab IZW (talk) 17:18, 22 November 2021 (UTC)[reply]
Hmmm... in slow pronunciation, I could pronounce it with [w̃s] but in fast speech I guess it's always [ns]. If it help, I may record a longer speech and then try to figure it out. I have some dialectal features in my speech but I don't think that pronouncing /ns/ is dialectal. I'd say it's rather a part of the standard Polish. I might be wrong though. Tashi (talk) 10:36, 23 November 2021 (UTC)[reply]
@Tashi And how about removing the /ʲ/? And what would you say is the best transcription for nasals before /ʂ/ and /ʐ/? Vininn126 (talk) 17:59, 22 November 2021 (UTC)[reply]
Also /x/, but more importantly, what your thoughts on the n assimilating to /ŋ/ before k and g, as I have updated? I read a fairly convincing argument sent by Shum, but I'm curious what you think. Vininn126 (talk) 18:02, 22 November 2021 (UTC)[reply]
@Mazab IZW, @Tashi I think we can probably remove the /ʲ/. As for nasals, my vote is to leave them as they are for now, with the potential exception of having them assimilate with /s/ and /z/. Vininn126 (talk) 10:52, 25 November 2021 (UTC)[reply]
@Vininn126 Yeah, I agree with that. As for the nasals though, I'm unsure about having them assimilate with /s/ and /z/ as both the [w̃s]/[w̃z] and [ns]/[nz] pronunciations seem to be used in Polish. Mazab IZW (talk) 19:23, 25 November 2021 (UTC)[reply]
@Mazab IZW Surjection helped me remove the palatalization. Vininn126 (talk) 19:24, 25 November 2021 (UTC)[reply]