Module talk:be-pronunciation

Questions

@Guldrelokk Hello. It's nice to see you working on this module. I'd expect адход (adxód) to be /atˈxot/, not /aˈtxot/. What do you think? --Per utramque cavernam (talk) 10:00, 24 March 2018 (UTC)

Same thing for вакза́л (vakzál): /vaɡˈzaɫ/, not /vaˈɡzaɫ/?

Notice that Russian вокза́л (vokzál) is rendered as [vɐɡˈzal], and отхо́д (otxód) as [ɐtˈxot].

@Atitarev --Per utramque cavernam (talk) 10:25, 24 March 2018 (UTC)

Hi! The syllabification given is based on the main source I’m using, Беларуская граматыка (у дзвюх частках) М. Бірулы et al., published in 1985 by the Academy of Sciences of the BSSR. It is the most complete Belarusian grammar to date, AFAIK, and was meant to be authoritative. The theory it presents is exactly the same as in the Russian authoritative grammar of the time, Русская грамматика Н. Шведовой et al., published in 1980 by the Academy of Sciences of the USSR. It always puts the syllable break before a cluster of non-sonorants (which can precede a sonorant), that is, before a sonority fall (/т/ and /х/ are given the same sonority). It should be kept in mind that syllabification is always somewhat arbitrary for both languages; it doesn’t play a visible role as a phonological unit. Investigations carried out to find an articulatory basis of a Russian syllable proved fruitless; this isn’t necessarily so for Belarusian, of course. Note also that while you can easily find other accounts for a syllable in Russian, I’m not sure about Belarusian, which is simply much less studied. Guldrelokk (talk) 11:03, 24 March 2018 (UTC)

Russian pronunciation module takes a different approach, giving [spʲɪkˈtaklʲ] for спектакль against Беларуская граматыка спе-кта-къл’ and Русская грамматика по-дго-ло-ву (под голову, with a re-syllabified preposition). Guldrelokk (talk) 11:11, 24 March 2018 (UTC)

@Guldrelokk: All right; I'm not well versed in IPA or Slavic phonology so I defer to you!

Another thing: you've implemented the rule /г/ is a stop in /зг/, /жг/. But at згі́нуць (zhínucʹ), I'm still hearing a fricative /ɣ/ rather than a stop /ɡ/. --Per utramque cavernam (talk) 09:22, 25 March 2018 (UTC)

Yes, it should be /ɣ/. /ɡ/ is almost non-existent in modern and standard Belarusian, except for voiced positions, such as вакза́л (vakzál). --Anatoli T. ^{(обсудить}/^вклад) 09:28, 25 March 2018 (UTC)

З/г/інуць was based on one of the sources, others give different examples for /зг/. I’ve just checked forvo for another pronunciation, and yes, [zɣʲ]. I bet this is because of a prefix. If I can find sources confirming that /г/ remains [ɣ] after a prefix, we’ll have to edit pronunciations of words like згінуць manually, because parsing morphology is too much for a pronunciation module, and otherwise /зг/ is pronounced as [zɡ], as all the sources say and you can hear at forvo. Guldrelokk (talk) 09:42, 25 March 2018 (UTC)

@Guldrelokk: To me, the recording above is for the Russian мозги́ (mozgí), not Belarusian мазгі́ (mazhí). I would mostly use /ɣ/ (except for voiced cases), since /ɡ/ is lost in Belarusian. Could you link to your source? Also, if it's OK, please make a user page with a Babel table for languages you speak or are learning. --Anatoli T. ^{(обсудить}/^вклад) 09:52, 25 March 2018 (UTC)

Беларуская граматыка (cited above) says (page 45): «Выбухны [г] (парны звонкі з [к]) вымаўляецца толькі ў паасобных словах і ў спалучэнні [зг]». Слоўнік беларускай мовы Бірулы says so too, though I can’t quote it. Слоўнік беларускай мовы (клясычны правапіс) (2001) spells it мазґі. See also these links: 1 (apparently normative), 2, 3, 4 (in Russian). Guldrelokk (talk) 10:23, 25 March 2018 (UTC)

@Guldrelokk: I see. Thanks for the quotes. I can now see that this is based on the older pronunciation and spelling, which was the norm - Taraškievica - тарашкевіца or клясычны правапіс (modern spelling: класычны правапіс). Letter ґ has been completely abandoned in Belarusian, since the sound is gone as well. You can still implement Taraškievica, I have no issue with that but we need to make sure users know it's not the government Belarusian. Please note (if you don't know already) that the Belarusian Wikipedia is split into two versions by language codes "be" and "be-x-old", the latter is based on "Taraškievica". It's also favoured by the Belarusian political opposition and there is some revival of this orthography and pronunciation. Maybe letter ґ will also be restored. --Anatoli T. ^{(обсудить}/^вклад) 10:36, 25 March 2018 (UTC)

Spelling, but not pronunciation! Clearly the works published by the BSSR Academy of Sciences in the 1980s were not written by Tarashkevitsa proponents, were they? The other source in Belarusian I gave you is also not in Tarashkevitsa. I referred to the 2001 Слоўнік just to show an example of this pronunciation being reflected in spelling (since the official orthography does not include ґ indeed). After all, Tarashkevitsa is (primarily) an orthography, not an orthoepic norm. Guldrelokk (talk) 10:47, 25 March 2018 (UTC)

Well, ґ was completely abandoned for a good reason. In my opinion, [ɡ] has become very marginal in Belarusian and I suspect modern authors might have just used some older resources. I won't challenge this and won't stop you from developing the module based on your sources, since you have solid references and resources for Belarusian are rather scarce but I am relatively sure that Belarusian pronounce "мазгі́" with a /ɣ/, not a /ɡ/. This can always be reviewed if other references dispute your findings.--Anatoli T. ^{(обсудить}/^вклад) 11:04, 25 March 2018 (UTC)

As long as the actual pronunciation is concerned, forvo has мазгі (previously referred) and розгі. They were both uploaded by a124kun, a user from Belarus. He doesn’t seem to record any Russian words, but a lot of Belarusian ones in what sounds like a perfectly standard accent, particularly pronouncing /г/ right, that is, as a fricative consonant – but not in /зг/. Therefore, this pronunciation must be current at least for some speakers. Guldrelokk (talk) 11:16, 25 March 2018 (UTC)

In Commons, however, it is always [zɣ]: бразгат, бразгаць, бразготка, compare брызгліна by a124kun. On one hand, it is clearly bad to have an IPA transcription different from the audio below; on the other, I cannot confirm it’s standard. It isn’t at all hard to hear Ukrainian щововкнути pronounced as [ɕːovofknutɨ]; this is not to be reflected in a dictionary as a norm. Guldrelokk (talk) 11:42, 25 March 2018 (UTC)

You may consider making variants based on different norms or practices or just stick to the norm you feel comfortable with. If an audio is different from standard or belongs to a different standard, it may have a short note. It's all manageable. The Ukrainian module was developed on a norm, which is not too common these days. Both Ukrainian and Belarusian are currently heavily influenced by Russian, by each other and by Polish. --Anatoli T. ^{(обсудить}/^вклад) 11:54, 25 March 2018 (UTC)

Yes, Russian influence and L2 speakers were what I was referring to. I’ll keep it as it is for now and look for more reliable sources. Guldrelokk (talk) 12:29, 25 March 2018 (UTC)

Geminations with prefixes

I wonder if адся́чы (adsjáčy) is done correctly /aˈtsʲat͡ʂɨ/. I think it should be geminated, because it's a prefix + root boundary, something like /aˈtsʲːat͡ʂɨ/ or /aˈt͡sʲˈsʲat͡ʂɨ/.

It doesn't have to match Russian but please check Russian отсе́чь (otséčʹ) - [ɐt͡sʲˈsʲet͡ɕ]. The gemination is not marked explicitly in thhis entry because the Russian module uses data with prefixes with mandatory geminations, implemented by User:Benwing2 (aka User:Benwing).--Anatoli T. ^{(обсудить}/^вклад) 02:47, 25 March 2018 (UTC)

Right, Беларуская граматыка says /т/ has a “fricative element” in such cases or simply becomes /ць/. Guldrelokk (talk) 09:47, 25 March 2018 (UTC)

 Done, thank you. Guldrelokk (talk) 10:00, 25 March 2018 (UTC)

дошцы

The case with до́шцы (dóšcy) (dative or locative singular of до́шка (dóška)) surprises me. I'm not saying it's wrong and I see that the module does what you tell it to do but is /ʂ/ really assimilated to /s/ here? I would pronounce it regularly as /ˈdoʂt͡sɨ/. I wonder if the reference is based on some very old Belarusian pronunciation.

BTW, I am not fluent in Belarusian but this assimilation doesn't make sense from a Ukrainian or Russian point of view.--Anatoli T. ^{(обсудить}/^вклад) 10:23, 25 March 2018 (UTC)

It does make sense. It is still present in many Russian dialects; more than that, it were formerly a part of the Moscow dialect and the literary standard as well. You can read about that in «История русского литературного произношения XVIII-XX вв.» М. Панова. However, Беларуская граматыка does give [шц] for шц. That it is still the norm for Belarusian can be seen here, page 104; a pretty modern publication. Other pronunciations may surely exist. Guldrelokk (talk) 10:40, 25 March 2018 (UTC)

Hmm, what was I trying to say? Беларуская граматыка gives [шц] and the modern source [сц]. I chose the modern source. Guldrelokk (talk) 10:50, 25 March 2018 (UTC)

Great, thanks. It's also on page 80. I don't have the Russian book handy but I can relate to the dialectal pronunciation of Russian смеёшься (smejóšʹsja, “you're laughing”) [smʲɪˈjɵsʲːə] instead of the regular [smʲɪˈjɵʂs(ʲ)ə] by analogy with the Belarusian смяе́шся (smjajéšsja) but the consonant cluster is not the same as [ʂt͡s] in "дошцы". Anyway, keep up the good work. You seem to have a good approach to this work and the required skills. --Anatoli T. ^{(обсудить}/^вклад) 10:57, 25 March 2018 (UTC)

Audio files

Commons has a rather abundant collection of audio files for Belarusian: commons:Category:Belarusian pronunciation; this might help. --Per utramque cavernam (talk) 10:57, 25 March 2018 (UTC)

This is helpful, thank you. Guldrelokk (talk) 11:19, 25 March 2018 (UTC)

Secondary stress

@Guldrelokk Hello. If I'm not mistaken, мовазнаўства should be transcribed as /ˌmovaˈznau̯stva/. But when I write мо̀вазна́ўства (the notation we use for Russian, I think), the module outputs **/moˌvaˈznau̯stva/. Could you have a look at that? --Per utramque cavernam 18:59, 1 May 2018 (UTC)

Sure, it should be handled correctly now. Guldrelokk (talk) 19:19, 1 May 2018 (UTC)

Affricates in front of nasals/stops?

@Guldrelokk, Atitarev: At выклі́чнік, I'm hearing [vɨˈklʲiʂnʲik] instead of [vɨˈklʲit͡ʂnʲik]. Should it be in the module? Is it a Russianism? --Per utramque cavernam 20:02, 2 May 2018 (UTC)

@Per_utramque_cavernam, Guldrelokk: I am not 100% sure. It may be a feature of Belarusian as well. Note that Belarusian ручні́к (ručník, “towel”) is often pronounced as "рушні́к". There's even a Belarusian song called "Рушнікі́". Some people say it's influenced by the Ukrainian рушни́к (rušnýk, “towel”). I don't know. I think рушні́к (rušník, “towel”) is an alternative form (slounik has it). It seems Ukrainian may have this feature as well (чн->шн) in some cases.

Note that чн->шн shift is traditionally called "Moscow" or "Old Moscow" pronunciation (even if it's spread much wider and many of examples are no longer used in modern Moscow). Also, Russian ручни́к (ručník, “hand break”) doesn't have the meaning "towel" as in Belarusian and Ukrainian but the origin is the same - from "рука" in all languages, and к produces ч, not ш (originally), and it's never "шн" in Russian in this word (it's much more modern).

More digression: The word ручні́к/рушні́к has some significance for Belarusians. Traditional Belarusian towels are part of the national flag. In Russian, Belarusian towels are sometimes called "рушни́к"/"рушники́" (pl). --Anatoli T. ^{(обсудить}/^вклад) 23:56, 2 May 2018 (UTC)

@Per utramque cavernam, Atitarev: It is a non-standard pronunciation; as in Russian, it is not applicable to all words, so it should be added manually whenever common, for example, to ручні́к (ručník). Guldrelokk (talk) 18:09, 2 June 2018 (UTC)

Monosyllables

@Guldrelokk Hi. The module puts a stress mark in front of monosyllables, such as шоўк. I don't think that's desirable. Could you fix that? Per utramque cavernam 09:48, 15 July 2018 (UTC)

Done. It counted [u̯] as a syllable. Guldrelokk (talk) 15:18, 15 July 2018 (UTC)

сь

@Benwing2, PUC: You may find it interesting that recordings at све́жы (svjéžy) or пасве́дчанне (pasvjédčannje) demonstrate that palatalisation of some consonants can be pronounced leaning on Polish for some speakers. It sounds almost like [ˈɕvʲeʐɨ] to me, not [ˈsʲvʲeʐɨ]. Compare with Polish świeży /ˈɕfʲɛ.ʐɨ/ (/v/ is devoiced to /f/ in Polish after voiceless consonants). The same Polish word in Lithuania will be pronounced similar to Belarusian (the initial consonant). --Anatoli T. ^{(обсудить}/^вклад) 01:01, 5 July 2020 (UTC)

@Atitarev Maybe not so surprising since Belarusian palatalizes т and д to affricates even in spelling, and Belarus borders on Poland (and in fact half of Belarus was once part of Poland). Benwing2 (talk) 01:08, 5 July 2020 (UTC)

@Benwing2: True but ць, дзь are supposed to be [ts̪ʲ] and [dz̪ʲ], not like Polish [t͡ɕ] and [d͡ʑ] but you may hear these pronunciations. Some Belarusians and Polish claim they don't hear the difference. I beg to differ but it's a dialect continuum. On the other hand, Bryansk, Smolensk or Pskov Russian accents may be similar to eastern Belarusian pronunciations. They use [ts̪ʲ] and [dz̪ʲ], not like Polish. --Anatoli T. ^{(обсудить}/^вклад) 01:15, 5 July 2020 (UTC)

Annotations

@Benwing2: Are you able to add |ann=, please? Required at e.g. наві́ны (navíny) / навіны́ (naviný). --Anatoli T. ^{(обсудить}/^вклад) 02:34, 7 July 2020 (UTC)

bot run to add Template:be-IPA

@Atitarev I have this basically working. It needs a few minor tweaks; I'll run it tomorrow. Benwing2 (talk) 07:38, 30 July 2020 (UTC)

@Atitarev The same code should work, with a few changes, for Ukrainian. Benwing2 (talk) 07:39, 30 July 2020 (UTC)

@Benwing2: Thanks for that. There are cases where the syllabifications (syllable onsets) are not perfect or probably redundant geminations (e.g. Іллі́рія) but I think it's better to have this template in entries to see what may require fixes or little tweaks. Do you agree? --Anatoli T. ^{(обсудить}/^вклад) 07:52, 30 July 2020 (UTC)

@Atitarev Agreed, and done. Benwing2 (talk) 02:04, 31 July 2020 (UTC)

@Benwing2: Awesome job, thank you! --Anatoli T. ^{(обсудить}/^вклад) 02:06, 31 July 2020 (UTC)

diff

@Theknightwho: "ә" is not a Belarusian character. What terms or respellings are you using? @Benwing2: FYI. Anatoli T. ^{(обсудить}/^вклад) 00:50, 7 September 2022 (UTC)

@Atitarev I was attempting to generate the unstressed pronunciation for the phoneme on а (a), as using the Latin schwa was generating the error that it wasn't a Cyrillic character. I was hoping to bypass that with this, but it didn't work. Theknightwho (talk) 00:53, 7 September 2022 (UTC)

@Theknightwho: I am not sure that a schwa is part of the standard Belarusian pronunciation. "а" and "о" result in [a] in unstressed positions. Anatoli T. ^{(обсудить}/^вклад) 01:02, 7 September 2022 (UTC)

I'll defer to you. Theknightwho (talk) 01:27, 7 September 2022 (UTC)

The Cyrillic letter "ў" and IPA symbols "w" vs. "u̯"

@Atitarev, @Benwing2: How is the letter ў/ŭ supposed to be represented in the IPA transcription? Right now the module does a position dependent selection between the symbols "w" and "u̯". So that, for example, паўаўтамат (paŭaŭtamat) is transcribed into [pawau̯taˈmat]. Was this intended or is it a bug in the current code? I have added some words with "ў" to Module:be-pronunciation/testcases.

The Belarusian word воўк (voŭk) and the Ukrainian word вовк (vovk) transcribe into [vou̯k] and [wɔu̯k] respectively. There are also Polish words stół and szkoła, where letters "ł" transcribe into "w" (more pronunciation audio can be found here and here). The Belarusian phonology article in English Wikipedia uses the IPA symbol "w" for "ў". -Ssvb (talk) 12:02, 15 November 2023 (UTC)

And one more comparison: the Belarusian зноў (znoŭ) [znou̯] vs. the Ukrainian знову (znovu) [ˈznɔwʊ]. Now the Ukrainian transcription also uses the [w] symbol. Are Belarusian/Ukrainian/Polish transcriptions supposed to be using compatible notations? I mean, in the sense that the same IPA symbols sound approximately the same, while different symbols sound mostly different. -Ssvb (talk) 13:19, 15 November 2023 (UTC)

@Ssvb: Hi. I am not sure what's required. Are there any cases that are incorrect? All cases you added at test_w: on Module:be-pronunciation/testcases are passing. Anatoli T. ^{(обсудить}/^вклад) 22:05, 15 November 2023 (UTC)

@Atitarev: The tests are passing because I added them to document the current behaviour. But they are demonstrating the use of two symbols [w] and [u̯]. Is this correct? Or should they be all changed to either single [w] or single [u̯]? -Ssvb (talk) 22:15, 15 November 2023 (UTC)

@Ssvb: That I don't know. You can research and advise what you need to be changed and under what conditions, e.g. if the behaviour is similar to the Ukrainian "в" (in some positions), it can be modelled on the Ukrainian. Then add/correct test cases, so that we can review and @Benwing2 can make the change. I am less focused on Belarusian and you probably know more about the correct phonology. But I'd prefer to keep it simple and consistent, if possible, if there are no glaring errors, like ["э"] as "e", etc. Anatoli T. ^{(обсудить}/^вклад) 22:21, 15 November 2023 (UTC)

IPA symbol(s) for the Cyrillic letters "е" and "э"

@Atitarev: A recent diff (also referenced from this Beer parlour discussion) introduced the split of the old vowel [e] into two: [e] or [ɛ] depending on their positions in the word. Is this how it should be in the long run? A simple phonemic notation only needs just one symbol for this sound. While a more sophisticated phonetic notation probably needs a lot more symbols, similar to how it's done in Module:ru-pron and documented in this table. Was the "middle ground" approach with two symbols taken deliberately to make it more comfortable for the Cyrillic script users, who probably want to still see and easily distinguish the familiar letters "е" and "э" even in the IPA transcription? Or was it a bug and only renaming [e] into [ɛ] was intended? -Ssvb (talk) 20:27, 15 November 2023 (UTC)

To make it clear, I'm personally in favour of just a simple and practical phonemic notation, which would be easy to maintain and verify against the information from the orthoepic dictionaries. But I'm just an amateur. -Ssvb (talk) 20:28, 15 November 2023 (UTC)

@Ssvb: I didn't add that table at Appendix:Russian pronunciation but it seems more correct. I also favour a more simple approach, which is easier to verify and maintain. That diff you mentioned (["э"] was "e") was a bug that needed to be fixed. We have missed it. Anatoli T. ^{(обсудить}/^вклад) 22:10, 15 November 2023 (UTC)