Help talk:IPA/Polish

Slender IPA palatalizations

I checked on Polish wiktionary that ɡʲ, kʲ and xʲ (all before i for g, k and (c)h) are literally the true IPAs on every letter are ɟ, c and ç, all of which are actually correct. Check it for reference. ApprenticeFan ^work 11:04, 31 August 2015 (UTC)

Both ⟨kʲ, ɡʲ, xʲ⟩ as well as ⟨c, ɟ, ç⟩ are correct transcriptions. There's nothing wrong with using the latter set, as long as you clearly state that these symbols represent post-palatal sounds (unless you're talking about something else - I'm not sure I understand you.) Peter238 (talk) 12:23, 31 August 2015 (UTC)

There was a discussion about that a while back. The distinction between the two is too subtle to call it anything more than arbitrary. The decision to use one set or another is, AFAIK, mostly convention. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:10, 31 August 2015 (UTC)

I thought that the distinction between those was nothing more than arbitrary. After all, ⟨ʲ⟩ says nothing about the amount of palatalization nor its exact phonetic realization. Peter238 (talk) 09:17, 2 September 2015 (UTC)

Phonetically speaking, yes. For dorsal consonants, ⟨ʲ⟩ tends to be used to reflect an advancement in the place of articulation forward toward the hard palate. It's not quite a "secondary" co-articulation like it is for labial or even coronal consonants. But there might be phonological, dialectal, historical, and even practical reasons to use one or another. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 21:32, 2 September 2015 (UTC)

@Aeusoes1: Actually, it'd be a good idea to use ⟨c, ɟ, ç⟩ instead of ⟨kʲ, ɡʲ, xʲ⟩. They're a bit easier to write and all descriptions of the phonology of Polish that I'm aware of analyze /c, ɟ, ç/ to be unambiguously phonemic along with /ɲ, tɕ, dʑ, ɕ, ʑ, j/ (the distinction between /ç/ and /ɕ/ is described as non-sibilant vs. sibilant). Plus, it's ⟨c, ɟ, ç⟩, not ⟨kʲ, ɡʲ, xʲ⟩ that are used by Jassem (2003) and Gussmann (2007).

Other soft consonants such as [tʲ] etc. do occur phonetically, but sources disagree on their phonemic status, which depends on whether you analyze /ɨ/ to be a separate phoneme like Rocławski (1976) or an allophone of /i/ like Gussmann (2007). Both analyses make some sense by the way. Mr KEBAB (talk) 05:47, 28 November 2017 (UTC)

Suggestions - Remove allophones

I think this article should mention phonemes, not allophonic variants which appear when a particular sound is followed by a specific sound. Specifically: 1) I'd remove the reference to `ɣ' in the `niechby' example. It's a voiced variant of `x'. 2) Similarly, I'd remove the palatalised xʲ in the 'hiacynt' example. Same as before, it's an allophone of plain `x' before a /i/ sounds. But it's the same phoneme. No need to complicate matters here; English too has this allophony in words such as `hard' and `huge' but in broad transcription the initial sound should be transcribed as /h/ in both cases. 3) Remove `ŋ'. It's a variant of 'n' before k/g (and not all speaker use this variant, anyway). 4,5) Remove palatalised k and g (examples Gienek and kierowca). I think you can very well use the /gjenek/ and /kjerowtsa/ transcriptions.

One can and should mention these allophonic variants in the notes, but in my opinion not in the main table. — L0rents (talk) 13:09, 6 June 2017

That's not what is done in case of other languages. We use a semi-narrow or broad phonetic transcription in these guides. Plus, IPA written inside the IPA-pl template is enclosed within phonetic brackets. Mr KEBAB (talk) 13:16, 6 June 2017 (UTC)

Suggestions - nasal vowels

Converning the nasal vowels ę ą, personally I don't like using the French-like ɛ̃ and ɔ̃. Of course it's a matter of convention, but I think those simbols are misleading. All modern analysis of the Polish nasal sounds say they are dipthongs with a first, non-nasal part which is /ɔ/ for ą and /ɛ/ for ę followed by a nasal glide w̃. So I'd prefer seeing kęs as /kɛw̃s/ and są as /sɔw̃/. I think it's much more accurate. For beginners learning Polish, I think it's also okay to suggest as an approximation to lose the nasality and substitute w̃ with `wn', so that the words are pronounced kełns and sołn. As to the English description, I don't like much the reference to French vin and son; these words sounds very different. Nevertheless, it's difficult to describe in words these sounds using only English sounds. Examples from Portuguese might be more accurate, but probably not really usefull for English speakers. — L0rents (talk) 13:09, 6 June 2017

French nasal vowels certainly don't sound "very different" but only somewhat different. Plus, the column is already called "English approximation". As to whether we should change [ɛ̃, ɔ̃] to [ɛw̃, ɔw̃], I'm not sure. I think it'd be a good idea. Mr KEBAB (talk) 13:16, 6 June 2017 (UTC)

Move discussion in progress

There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:17, 15 July 2017 (UTC)

palatalization

why is ɡʲ, kʲ, and xʲ listed but bʲ, dʲ, fʲ, lʲ, mʲ, pʲ, rʲ, tʲ, vʲ, wʲ not listed? LICA98 (talk) 06:49, 27 November 2017 (UTC)

It's my understanding that these consonants don't exist in Polish. There are a number of these that would be transcribed as [bj], etc (that is, as a consonant cluster with [j]). But [wʲ] definitely isn't a thing in Polish. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 16:46, 27 November 2017 (UTC)

@Aeusoes1: That's not correct. There's no [wʲ] (and most probably no [rʲ] either) in Polish, but [mʲ, pʲ, bʲ, tʲ, dʲ, tsʲ, dzʲ, tʃʲ, dʒʲ, fʲ, vʲ, sʲ, zʲ, ʃʲ, ʒʲ, lʲ] (I hope I've listed all of them) occur allophonically before /i/, some of them only in loanwords such as Toshiba [tɔˈʃʲiba]. They may be regarded as phonemes only if you consider /ɨ/ to be an allophone of /i/. Then, /wʲ/ must also be regarded as a phoneme, but that symbol just tells you that the correct phonetic realization of the following vowel is [i] instead of [ɨ]. Words such as weekend are pronounced [ˈwikɛnt], with a fully hard [w] and front [i]. The same applies to /rʲ/. Mr KEBAB (talk) 05:53, 28 November 2017 (UTC)

I was under the impression that the soft/hard pairing with /w/ was between [w] and [l]. What words would [wʲ] appear in? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 21:21, 28 November 2017 (UTC)

@Aeusoes1: As I said, there's no [wʲ] in Polish. There's /wʲ/ (note the slashes) when you regard /ɨ/ as an allophone of /i/ (so [ɨ]). Then, /wʲi/ is [wi] (as in weekend) and /wi/ is [wɨ] (as in łyk). Phonetically, that consonant is always hard regardless of the following vowel. Poles can't pronounce French [ɥi] and typically substitute [wi] for it, which is another proof that there's no phonetic [wʲ] in Polish. Mr KEBAB (talk) 17:04, 29 November 2017 (UTC)

Sorry, I'm not understanding. It sounds like you're saying there are analyses of Polish phonology where there are two phonemes, /wʲ/ and /w/, but the two are phonetically identical. That sounds nonsensical. What am I missing here? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:07, 29 November 2017 (UTC)

@Aeusoes1: As in Russian, /ɨ/ can be analyzed as an independent phoneme or as an allophone of /i/. The transcription /wʲ/ tells you about the backness of the following close vowel. Phonetically, the consonant itself is always hard. Mr KEBAB (talk) 17:12, 29 November 2017 (UTC)

@Aeusoes1: To put it another way, you can write /ˈwikɛnt, ˈwɨk/ or /ˈwʲikɛnt, ˈwik/ but not /ˈwikɛnt, ˈwik/ as that leads to loss of important information. /ˈwʲikɛnt, ˈwɨk/ is redundant and so also a bad transcription. Mr KEBAB (talk) 17:29, 29 November 2017 (UTC)

Sorry, but that still sounds like nonsense. There is no analysis of Russian that posits two phonetically identical phonemes. I've never heard of that anywhere for any language.

If anything, it sounds like /ɨ/ must be analyzed as a separate phoneme, since it would otherwise require such absurdities to make sense. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 19:06, 29 November 2017 (UTC)

@Aeusoes1: You're being too dismissive.

I've checked Gussmann (2007) (who does consider [i, ɨ] to belong to the same phoneme /i/) and on page 29 he transcribes weekend as [wʲikɛnt]. If there's any palatalization at all, it's so inaudible that I just can't detect it. On page 4, he says that the palatalized quality of [lʲ] is so negligible that he chooses to transcribe it [l]. It's probably the same with [wʲ]. There's a very small phonetic difference between it and [w]. Mr KEBAB (talk) 20:16, 29 November 2017 (UTC)

That's a bit of selective reading. Here's the transcription in context:

Another relevant phonotactic observation concerning [w] is that it cannot be followed by the vowel [i] but may be followed by [ɨ]. This restriction holds for the native vocabulary, e.g. łys-y [wɨsɨ] 'bald', zł-y [zwɨ] 'bad', but it also extends to recent loans; thus weekend, originally pronounced [wʲikɛnt] with the strongly non-Polish combination [wʲi], is heard more and more often now as [wɨkɛnt]. On the other hand, the lateral [l] cannot be followed by [ɨ] but readily combines with [i], e.g.: liść [liɕtɕ] 'leaf', dol-i [dɔli] 'fate, gen. sg.'.

So [wʲ] is not only considered "non-Polish" (Perhaps a foreign affective pronunciation akin to English speakers pronouncing a uvular rhotic in loanwords from French) but its pronunciation is deprecated for a pronunciation more in keeping with the phonotactic constraints that are normal for Polish. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 23:42, 29 November 2017 (UTC)

@Aeusoes1: It's selective because the bit about [wɨkɛnt] actually is nonsensical (apart from the fact that [wʲ] is non-Polish, which is obviously true). That pronunciation is non-standard. See [1], [2] and [3]. Mr KEBAB (talk) 05:16, 30 November 2017 (UTC)

Interesting. So the preference for [wʲikɛnt] among Polish language policymakers is akin to English Standard bearers preferring a uvular r in French loanwords or a tap/trill pronunciation in Spanish-language names like Graciela.

Going back to the original question, should we include the other palatalized consonants? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 14:06, 30 November 2017 (UTC)

@Aeusoes1: I'm sorry but that's not correct either. Gussmann is simply exaggerating with his 'more and more'. If there ever was such a tendency, it's long gone now. The only neutral pronunciation is [wʲikɛnt], [wɨkɛnt] carries humorous/dialectal (of traditional Warsaw dialect specifically, like lypa for lipa) connotations. Ask any other native speaker and he'll tell you the same. It's probably similar to pronouncing the English word pronunciation with /aʊ/ - it's just not standard in any way.

I'd say I'm ambivalent, but on second thought, no. There's no reason to complicate this guide. Mr KEBAB (talk) 14:27, 30 November 2017 (UTC)

Unless you didn't mean to say that the [wʲ] pronunciation is artifically standard (it's not, it's actually standard), then forget about the first part of my post. Mr KEBAB (talk) 14:28, 30 November 2017 (UTC)

If I'm understanding correctly, then, the pronunciation with [i] is standard, but the guy who thinks that [ɨ] and [ɨ] are allophones of one phoneme is twisting the data to fit with his preconceived notion about phonotactic constraints.

What about the other palatalized sounds? We've got them for Russian and if there are analyses that peg them as independent phonemes, that makes me think they are worthy of inclusion in our transcriptions. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:48, 30 November 2017 (UTC)

@Aeusoes1: Give me a few days to check my sources thoroughly. Mr KEBAB (talk) 10:01, 1 December 2017 (UTC)

Tiebar?

@Mr KEBAB: Affricate symbols in Template:IPAc-pl are with tiebars. LoveVanPersie (talk) 10:11, 10 January 2018 (UTC)

@LoveVanPersie: We can remove them. Mr KEBAB (talk) 06:06, 11 January 2018 (UTC)

why are there so many differences between transcriptions on Polish wiki and English wiki?

for example:

vowels before n, m are transcribed as nasal in Polish and non-nasal in English (zãmɛk / zamɛk)

ę, ą when they get n or m sound are transcribed as nasal in Polish but in English non-nasal (rɛ̃ŋka/rɛŋka)

and when they don't in Polish there is /w̃/ but in English just nasalized letter (kɛ̃w̃s/kɛ̃s)

/c/, /ɟ/, /ç/ are transcribed /kʲ/, /gʲ/, /xʲ/

consonants before j are palatalized in Polish but not English (pɔlʲit͡sʲja/pɔlʲit͡sja)

ia, ie are transcribed with /ʲj/ in Polish but without j in English (bʲjawɨ/bʲawɨ) LICA98 (talk) 23:32, 26 July 2018 (UTC)

You're right, the trascriptions used e.g. on the Polish wikidictionary (Wikisłownik) are different from the ones used on English Wikipedia. As far as I could determine, the transcription on Wikisłownik were produced automatically using a computer program which uses the system used in a certain reference [Ostaszewska, Tambor - Fonetyka i fonologia współczesnego języka polskiego, PWN, 2000]; in turn, this reference (a short booklet written for university lectures and which doesn't contain any original research) is based (if I'm not mistaken) on Słownik wymowy polskiej. Red. M. Karaś, M. Madejowa, Warszawa 1977. Among other things, in this system, ALL vowels are marked as nasalised (with a tilde) when followed by [n], [m], [ŋ]. Also, all consonants are marked as palatalized (with the j superscript) when followed by [j] or by [i]. Because these differences between the transcriptions in English/Polish Wikipedia are systematic, the two transcription systems carry the same amount of information and therefore amount to different but functionally equivalent conventions. However, if you ask me, this Ostaszewska-Tambor system is stupidly complicated, at odds with the IPA transcriptions in any other language I know, confusing (the ã in zamek has little to do with French nasal ã in, e.g., ans) and is very unsuitable for transcriptions in a dictionary. It makes things complicated in a totally unnecessary manner, and if I didn't know it comes from a printed book I'd think it was created by a crackpot. This in just my opinion, of course, but I'm happy English wikipedia uses a more sensible system of transcription.

L0rents (talk) 01:08, 20 December 2018 (UTC)

Dialectical pronunciations

Should there be a footnote about /ɦ/, which exists as a separate phoneme in some dialects? (I do not propose that we transcribe it, of course.) Double sharp (talk) 14:43, 5 January 2020 (UTC)

Would it be confusing for anyone if we didn't? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 18:13, 5 January 2020 (UTC)

labial palatalized consonants

My source (cited in the Polish phonology page, Swan 2002) says labial palatalized consonants are phonemes, not allophones, and it makes more sense as there are minimal pairs like pędź - pięć, zdrowe - zdrowie, bały - biały, mały - miały, and a near minimal pair szwie - szwed. So I suggest adding these to this page. JeanneAymonier (talk) 05:34, 19 February 2020 (UTC)

@Aeusoes1 and Kbb2: Can you possibly chime in on this (and on these edits to Polish phonology)? It seems the phonemic status of palatalized labials depends on the status of [ɨ] in the given analysis, but I can't make heads or tails. Nardog (talk) 20:57, 26 February 2020 (UTC)

I was under the impression that, phonetically, these were sequences (e.g. [mj] or [mʲj]) so that the contrast between mały miały could be something like /mawɨ/ and /mjawɨ/. If we've got sources that say that they are e.g. [mʲ] at the phonetic level, I'm game for changing how we transcribe Polish, even if [mʲ] is more frequently analyzed as /mj/. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 22:52, 26 February 2020 (UTC)

I don't have sources other than the one cited in the Polish phonology page, but the pronunciation for the word Mjanma is /ˈmjan.ma/, while the pronunciation for the word miały is /ˈmʲa.wɨ/ on Wiktionary. Same with Pjongjang, wjazd, objazd, etc., so it seems that /mj/ and /mʲ/ are phonetically different. JeanneAymonier (talk) 23:46, 26 February 2020 (UTC)

I have found out that Polish Reference Grammar By Maria Z. Brooks distinguishes /m/ and /m'/(palatalized m) etc. as phonemes. JeanneAymonier (talk) 00:38, 27 February 2020 (UTC)

I have listened to http://polfon.upol.cz/index.php? and I've figured out that ki+vowel is pronounced [kʲ] but pi+vowel is pronounced [pʲj] so I have changed my opinion. I'll revert the edit in the Polish phonology. page=homeJeanneAymonier (talk) 01:58, 27 February 2020 (UTC)

amigo - "g" as "h" !!??

Is this done??

--213.76.134.90 (talk) 22:19, 16 September 2020 (UTC)

is 's' really a voiceless dental fricative?

Copying what I originally posted in the general wiki help page:

I noticed that in both pages for Polish ortography and phonology, (and I just found out, also on this very page) 's' in its usual value is linked to voiceless dental fricative, which sounds weird to me, as that sound shouldn't be in Polish at all and I'd have expected it to be a Voiceless alveolar sibilant instead.

I checked in the code for the page and it seems that the links are automatically generated, so I didn't want to touch anything... Could someone with a bit more knowledge in the field tell me more, and correct either me or the link?

Thanks!--Stefano thf (talk) 14:57, 15 April 2021 (UTC)

No, the links are incorrect. Even if they are technically dental or denti-alveolar in articulation, the links they are going to is a class of sounds that have no sibilance, which is not true of this sound. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 03:17, 17 April 2021 (UTC)

Thank you for noticing this. Fixed in Module:IPA symbol/data, though the redirects and anchors are all over the place so they may also need refining. Nardog (talk) 05:13, 17 April 2021 (UTC)

Yes, [s z ts dz] are sibilant fricatives, which means that they're articulated with clenched teeth. Their being dental mostly means that they have no positional allophones approaching [ʃ ʒ tʃ dʒ], which they have in some non-Slavic languages (e.g. English and especially Dutch), at least when it comes to fricatives. And before you ask, yes, from the standpoint of someone who speaks a language with dental sibilants natively, the alveolar fricatives of English do have a slight [ʃ ʒ] flavor to them, especially in contact with /r/ (interestingly, Dutch has the same allophone after /s/, though it's even more [ʃ]-like). You can also hear it in the way [s z] blend into the following vowel. The transition is more dull in English. In Dutch, you can hear it in every position as their sibilants are way more Greek/Spanish (i.e. way more dull and probably slightly more weakly fricated) than they are English/German. Sol505000 (talk) 21:51, 18 April 2021 (UTC)