Module talk:sk-IPA

A simple resource

Apart from Slovak phonology, I found this: https://www.101languages.net/slovak/orthography.html with important points on foreign words, exceptions and respellings. Anatoli T. ^{(обсудить}/^вклад) 05:57, 16 May 2023 (UTC)

Palatalisation is expected for d, t, n and l before e and i

@Erutuon, @Benwing2: Hi,

I have added failed cased to Module:sk-IPA/testcases where normally a palatalisation is expected for consonants "d", "t", "n" and "l" before vowels "e" and "i" (there are exceptions in loan words and some other cases in the link I gave in my previous post).

So, deti is /ˈɟɛci/, not /ˈdeti/, etc. Anatoli T. ^{(обсудить}/^вклад) 06:03, 25 May 2023 (UTC)

I think this module needs a major update F. V. Lorenz (talk) 23:24, 12 July 2024 (UTC)

Respellings for ireegular pronunciations

@Železná päsť: Hi. Are you able to/planning to add respellings for irregular pronunciations? E.g. some special letters or symbols? E.g. vtedy to be pronounced [ˈftɛdi], not [ˈvcedi]. Anatoli T. ^{(обсудить}/^вклад) 00:00, 11 September 2024 (UTC)

Yes, on the current version of the template, the IPA can be produced with {{sk-IPA|/'ftedi/}} although in the future, the template should deduce a more precise pronunciation. Železná päsť (talk) 02:04, 11 September 2024 (UTC)

@Železná päsť: Thanks. I am aware of the manual input with {{sk-IPA|/'ftedi/}}. A lot of pronunciation modules employ respellings when there alternative symbols - Russian, French, Bulgarian, etc but I don't know what a symbol could be to force an alternative pronunciation, which is irregular.

For example, Czech {{cs-IPA|neobyčejný|ne?obyčejný}} would give [ˈnɛobɪt͡ʃɛjniː], [ˈnɛʔobɪt͡ʃɛjniː] where "?" produces [ʔ] and Bulgarian {{bg-IPA|ўе́бсайт}} gives [ˈwɛpsɐjt] ("ў" produces "w"), a letter never used in Bulgarian.

I don't think the module can be smart enough to know how to automatically pronounce telefón. In Czech it's easier with ti/di/ni, you just respell them with ty/dy/ny for loanwords (which can be applied to Slovak as well) but there are more palatalisation cases in Slovak and cases where the palatalisation should not be used.

CC @TomášPolonec, @Benwing2. Anatoli T. ^{(обсудить}/^вклад) 02:30, 11 September 2024 (UTC)

I would have called that phonetic extension instead of respelling, so that could be easily implemented by adding to the "phonetics" table in the template. My answer was only in response to the example that you mentioned in your comment, because earlier it wasn't possible to produce the required IPA. And in addition to that, I don't speak this language fluently, I have barely learned anything while editing entries on this website. Železná päsť (talk) 02:56, 11 September 2024 (UTC)

@Železná päsť: You can work on the module, even if you don't know the language perfectly. The module is getting better but still need to find a way to cater for exceptions with phonetic respellings and add additional symbols when native means are unavailable. You can't mix Slovak letters with phonetic symbols, like 'ɛ', unless you add some handling for it.

A symbol, such "^" after a consonant, could be used to indicate no palatalisation, {{sk-IPA|t^el^efón}}. (I am not insisting on this method, you may want to add handling for 'ɛ', e.g. {{sk-IPA|tɛlɛfón}} as a phonetic respelling).

Russian бренд (brɛnd) is respelled with {{ru-IPA|phon=брэнд}}. Anatoli T. ^{(обсудить}/^вклад) 04:12, 11 September 2024 (UTC)

@Atitarev@Železná päsť I just saw all the changes and reverts. Is the module perfect? No, but with the hard type T- D- N you can just overide the results with capital letters, we just need to pay attention with the foreign words and certain exceptions. For me more problematic is the issue with -v- like bravčovina or -ev in cirkev ‎ which you can't override manually. Chihunglu83 (talk) 07:32, 11 September 2024 (UTC)

@Chihunglu83, @Železná päsť, @Fenakhay, @Benwing2:

Chihunglu83, do you mean it's pronounced [ˈt͡sirkew], not [ˈt͡sirkef] and [ˈbrawt͡ʃɔvina], not [ˈbravt͡ʃɔvina]? I think it's just a flaw of the module but, anyway, a "w" could be used for respelling [w], which should not be used as a synonym for "v" in respellings.

I noticed @Železná päsť made wrong edits on those entries.

Correcting to use {{IPA|sk|[ˈt͡sirkew]}} and {{IPA|sk|[ˈbrawt͡ʃɔvina]}}. Anatoli T. ^{(обсудить}/^вклад) 08:43, 11 September 2024 (UTC)

@Atitarev Unfortunately I do not have a good feeling about User:Železná päsť. You can see in the history of their talk page that they've been blocked for creating major errors through module edits and then either refusing to fix the problems or throwing a tantrum when their erroneous changes were reverted. Last time this happened (a few months ago) I told them they need to use sandbox modules rather than doing their trial-and-error coding directly in the production space, and they refused to do so and said they would instead leave Wiktionary entirely. IMO unless this user agrees to use sandbox modules for testing, they should be blocked from the module space as they really don't know what they're doing and it seems they refuse to learn. Benwing2 (talk) 08:53, 11 September 2024 (UTC)

@Benwing2: I see your point. You've got the skills to fix it, the phonology is not too complicated, even if you use what's easily available. Anatoli T. ^{(обсудить}/^вклад) 09:31, 11 September 2024 (UTC)

There's a perfect example, when someone reverts my edits and all they said is "infinite loop" without explaining where it was. There weren't slovak entries on the module error page and I was looking at it after I published the edits. Maybe there is a problem with wiktionary instead. How do you expect people to learn when you don't teach, morons. Železná päsť (talk) 18:09, 11 September 2024 (UTC)

@Železná päsť: That revert was made by @Fenakhay. He doesn't know Slovak but he knows some Lua. Better ask him about that revert. I am not so great with Lua at all. Anatoli T. ^{(обсудить}/^вклад) 00:06, 12 September 2024 (UTC)

I didn't even reply to your comment, good to see that everyone is paying attention. Železná päsť (talk) 00:17, 12 September 2024 (UTC)

@Atitarev /w/ is fine for me, I also see /u̯/ here and there. I assume /ʋ/ is the best and it is included in our IPA - special letter. Probably needs concensus. Chihunglu83 (talk) 09:11, 11 September 2024 (UTC)

@Chihunglu83: Agree with displaying /ʋ/, letter "w" could be used as a respelling tool. If it's consistent (using one standard or source), we could just apply it. Anatoli T. ^{(обсудить}/^вклад) 09:32, 11 September 2024 (UTC)

@Chihunglu83, @Benwing2: After my module edit on diff (["w"] = "ʋ"), I have applied {{sk-IPA|cirkew}} on cirkev to get [ˈt͡sirkeʋ] and {{sk-IPA|brawčovina}} on bravčovina to get [ˈbraʋt͡ʃɔvina]. Anatoli T. ^{(обсудить}/^вклад) 09:39, 11 September 2024 (UTC)

Hi @Benwing2: Hi, would you be able to add a respelling form, as in the Czech module:

detstvo is respelled as "dectvo" to become [ˈɟet͡stvɔ] (with a tie).

{ "detstvo", "ˈɟet͡stvɔ", "dectvo"},

@Chihunglu83: I have added a handling for the palatalisation of "l", e.g. leto as [ˈʎetɔ], following the same rule as t, d, and n. Anatoli T. ^{(обсудить}/^вклад) 10:18, 11 September 2024 (UTC)

@Chihunglu83: does it break the case for chlieb? Should it be [xʎi̯ep] or [xli̯ep] Anatoli T. ^{(обсудить}/^вклад) 10:22, 11 September 2024 (UTC)

I didn't make "wrong" edits on those entries; the link to 101languages in the first comment said

When "v" stands at the end of the syllable, it is pronounced as non-syllabic u (bilabial approximant/u̯/), with the exception of the position before "n" or "ň", for example, kov /kou̯/ (metal), kravský /krau̯skiː/ (cow - adjective), but povstať /pofstatʲ/ (uprise) because the v is not at the end of the syllable ( po-vstať ), hlavný /hlavniː/ because "v" stands before "n" here

so it isn't something that I just invented because it sounds right. You can't blamer me when other people lie about the language's phonology. Železná päsť (talk) 18:07, 11 September 2024 (UTC)

@Železná päsť: Wrong edits I pinged you on had modules errors. You saw errors but you left them. You can't take offense when people point you out to your errors. You won't survive here if you remain aggressive and refuse to change. Some people evolve but I don't know about you. Language and technical skills may not be enough.

Those cases you describe can now be handled by respelling with "w", e.g. on kov use {{sk-IPA|kow}} to get [kɔʋ] (the correct notation). The other cases are handled automatically. Anatoli T. ^{(обсудить}/^вклад) 23:58, 11 September 2024 (UTC)

No, I didn't see modules errors on the pages that I edited, so you must be delusional. The "voiced labiodental approximant" that you mentioned is supposed to be the correct notation for "v" in the word "kov" but wikipedia said it is used in the word "voda" transcribed as [ˈʋo̞dä] and wikipedia also clearly states that the letter "v" at the end of a word isn't pronounced the same. Evidently, everyone on wikipedia fails to understand slovak phonology, so just delete the whole module, no loss. Železná päsť (talk) 00:15, 12 September 2024 (UTC)

@Železná päsť: The errors are all in the history, you can't hide. You can still see them, if you look at your edits, just one of your numerous errors is here: this revision. They were all in Cat:E. So who is delusional? Anatoli T. ^{(обсудить}/^вклад) 00:35, 12 September 2024 (UTC)

Who is delusional? Let me answer that question: you are. Because the edit history on those pages is out of sync with the current version of the module. Simple, your comment just proved that you're delusional. I'm not here to "survive" with the rabble of the internet. I'm gone. Železná päsť (talk) 00:46, 12 September 2024 (UTC)

@Atitarev, Benwing2, Chihunglu83:Hello, thanks for tagging me, I haven't been active for some time. The official rules for pronunciation and transctiption with all the details can be found in this book: KRÁĽ, Ábel. Pravidlá slovenskej výslovnosti. 3. ed. Martin : Matica slovenská, 2016. 423 pp. ISBN 978-80-8128-159-4. I don't have access to it at the moment, but there are exact rules with lists of exceptions in it, which could be useful for upgrading the module. The thing is, that the pronunciation often differs based on whether a word is a loanword or native (especially in regards to palatalisation), which the module can't automatically identify. And then there are other exceptions like pointa pronounced as /'po.enta/, etc. So while a lot of the "irregular" pronunciations covered in this discussion so far could be actually done automatically (like dcéra, chlieb, cirkev...), unless we add a "nativeness" parameter, automatic palatalisation of the d/t/n/l+e/i is basically useless because of the huge amount of loanwords where it would create false soft spellings (as in telefón , which should be /'telefo:n/, internet /'internet/, fonetika /'fonetika/...).

As for the specific concerns mentioned: in Slovak phonetics, /(a/e/o)u̯/ is more common to mark the pronunciation of the -av/-ev/-ov/-ou endings. And chlieb is pronounced /xʎi̯ep/. --TomášPolonec (talk) 14:05, 11 September 2024 (UTC)

@TomášPolonec: Thank you for your response. Although not everyone here knows enough Slovak. Handling of loanwords and using respellings is nothing new. perhaps adding |loan=y parameter would be a shortcut for respellings like {{sk-IPA|foNeTika}}.

I have corrected the test case for chlieb as [xʎi̯ep] when I figured out. "-ie-" palatalises consonants just like "i" does. Anatoli T. ^{(обсудить}/^вклад) 00:41, 12 September 2024 (UTC)

Hello @TomášPolonec, @Chihunglu83, @Benwing2: I have just added a "^" symbol to force syllabic as in {{sk-IPA|šmr^nc}} on šmrnc or in {{sk-IPA|čl^n}} on čln, resulting in [t͡ʃl̩n] and [ʃmr̩nt͡s].

Please add a respelling parameter in the test case module. Anatoli T. ^{(обсудить}/^вклад) 01:14, 12 September 2024 (UTC)

@TomášPolonec: Pity this book is not available online and I can't find downloads.

@Chihunglu83: Hi, you need to be careful as well with {{sk-IPA}} won't be able to guess how to pronounce the Slovak joint venture without a respelling, as you did in diff. This was fixed by Železná päsť in the next edit.

Slovníkový portál Jazykovedného ústavu Ľ. Štúra SAV has the respelling - [džojnt venčr]. Anatoli T. ^{(обсудить}/^вклад) 04:47, 12 September 2024 (UTC)

@Atitarev: Sure, I think that adding a |loan= parameter would solve the issue with the loanwords. As I have said, with the rules given in that book, all the native words could be handled automatically (most of the "irregularities" are just assimilation or consonant clusters) and most of the loanwords as well (given that the |loan= parameter is introduced). Unadapted or semiadapted borrowings are usually the only cases where the pronunciation is unpredictable. I could get the book, scan it and share it with all of you via the Wiktionary Discord server (which, as I've just found out, I am no longer a member of :)). --TomášPolonec (talk) 12:25, 12 September 2024 (UTC)

@TomášPolonec: |loan= is an idea ATM, yet to be implemented, so we can continue with full respelling until that happens. If you share the book with us, it might be used for the module improvement and for adding those irregular readings. I can't commit, since my skills in Lua are low but I can help adding test cases and entries.

I wonder how detailed are entries in

politika in Slovníkový portál Jazykovedného ústavu Ľ. Štúra SAV. Ortograficko-gramatický slovník slovenčiny says [-t-] but nothing about "l". So is politika pronounced [ˈpɔlitika] or [ˈpɔʎitika]. I think the former, as the entry says but I am not sure. The entry polícia has [ˈpɔʎiːt͡si̯a]. I have questions about how "l" is consistently pronounced in loanwords. https://forvo.com/search/literat%C3%BAra/ sounds like [ˈʎiteratuːra] to me. Anatoli T. ^{(обсудить}/^вклад) 23:11, 12 September 2024 (UTC)

@Atitarev: I can look at the module later, as I've already worked with Lua before (mainly on Module:sk-verb), and make an improved version in my sandbox.

The dictionaries usually don't mention the [l], since [ʎ] is becoming rare in modern Slovak. In loanwords unmarked palatalisation never occurs, especially in those from Latin or Greek. Even let's say in Chinese names where the group [-ci-] would appear, it's marked as ťi (see e.g. w:sk:Ťing Chaj-pcheng). So it is [ˈpɔlitika], [ˈpɔliːt͡sia] (not [ˈpɔliːt͡si̯a], but I can't find a way to force a hiatus) and [ˈliteratuːra] (which in what's on the recording, that's just a regular [l]). I must say though that I speak with a soft Central Slovak dialect, so I tend to (incorrectly!) pronounce these consonants as palatalised :) --TomášPolonec (talk) 00:40, 13 September 2024 (UTC)

@TomášPolonec: Hi. I have just created lôžko and checked sound recordings. Is there really no devoicing in Slovak or it's just lacking in the module? It should be [ˈlu̯oʃkɔ], not [ˈlu̯oʒkɔ], right?

E.g. otázka is [ˈɔtaːska], not [ˈɔtaːzka] per entry and Wikipedia.

Re: hiatus. Can we add a respelling? Anatoli T. ^{(обсудить}/^вклад) 01:21, 16 September 2024 (UTC)

@Atitarev: Most of the irregularities in pronunciation mentioned in this discussion (at least those not concerning palatalisation) are because of assimilation: vtedy [ˈftedi], <v> as [ʋ] (cirkev, bravčovina, kov...) and your examples should also be [ˈlu̯oʃkɔ] and [ˈɔtaːska]. That's what I was saying when I mentioned that the module needs a major upgrade because otherwise the automatic transcription is simply wrong in a lot of cases where it could be avoided (sound assimilation at any position in a word presenting as either voicing or devoicing is super common).

Certainly, maybe a dot could be used? Not sure. --TomášPolonec (talk) 09:57, 16 September 2024 (UTC)

@TomášPolonec: Thanks. Voicing and devoicing should be handled as regular changes, so an overhaul is required but there are many pronunciation modules that already handle that. So maybe "borrowing" is easier than writing from scratch. Respellings should only be required when words are really irregular.

In vtedy [f] is a regular predictable pronunciation (devoicing) but [t] is not (lack of palatalisation), so a respelling should be "vTedy", not "fTedy" if the module worked correctly. Anatoli T. ^{(обсудить}/^вклад) 22:40, 16 September 2024 (UTC)