Module talk:ru-pron
- Archives
Brackets as invalid IPA characters
[edit]@Wyang, Benwing, Atitarev This module is generating errors in CAT:IPA pronunciations with invalid IPA characters because it apparently is using square brackets inside transcriptions rather than only at the right or left edge. See [1] for an example. Can this be fixed, please? —Mahāgaja (formerly Angr) · talk 07:55, 11 January 2018 (UTC)
- @Mahagaja The original code was too complicated for me to touch. I fixed it using an ad hoc method for now. Wyang (talk) 08:46, 11 January 2018 (UTC)
- @Wyang: Thanks! —Mahāgaja (formerly Angr) · talk 08:50, 11 January 2018 (UTC)
- @Wyang: I found where the two pronunciations were generated and did a cleaner fix. — Eru·tuon 10:43, 11 January 2018 (UTC)
- @Erutuon Oh thanks! Wyang (talk) 11:26, 11 January 2018 (UTC)
- @Erutuon Now it's causing module errors, e.g. антенна радиолокационного дальномера. —Mahāgaja (formerly Angr) · talk 21:23, 11 January 2018 (UTC)
- @Mahagaja: Reverted my edit that caused it. — Eru·tuon 21:53, 11 January 2018 (UTC)
- @Erutuon Now it's causing module errors, e.g. антенна радиолокационного дальномера. —Mahāgaja (formerly Angr) · talk 21:23, 11 January 2018 (UTC)
- @Erutuon Oh thanks! Wyang (talk) 11:26, 11 January 2018 (UTC)
Incorrect stress mark placement indicating syllable break between [z] and [n] in узнать, обознаться, сознание, etc.
[edit]Should be [ʊˈznatʲ], [ɐbɐˈznat͡sːə]; not [ʊzˈnatʲ], [ɐbɐzˈnat͡sːə]. Tetromino (talk) 18:57, 29 January 2018 (UTC)
- @Tetromino That can fixed manually by adding a slash between the vowel and following з in the call to
{{ru-IPA}}
, but to fix it automatically is hard; words beginning with раз- and из- should have a syllable division after the з, not before it, and there's no easy way to automatically handle these and a zillion comparable cases (e.g. наткну́ться (natknútʹsja) should arguably be [nɐˈtknut͡sːə] as it's a combination of на- and ткну́ться, but understandably the algorithm doesn't split it that way automatically. Benwing2 (talk) 06:10, 22 May 2018 (UTC)
Pronunciation of various words beginning with и
[edit]Going through и in Ivanova. Please comment.
- изжо́га (izžóga): Ivanova appears to say жж or ж. We just have жж.
- иллюзио́н (illjuzión), иллюзиони́ст (illjuzioníst), иллюзиони́стский (illjuzionístskij), иллю́зия (illjúzija), иллюзо́рный (illjuzórnyj): Ivanova says лл or л. We have лл for иллюзия, л for иллюзорный.
- иммигра́нт (immigránt), иммигра́нтка (immigrántka), иммиграцио́нный (immigraciónnyj), иммигра́ция (immigrácija), иммигри́ровать (immigrírovatʹ): Ivanova says мм or м. We have м.
- импрессиони́зм (impressionízm), импрессиони́ст (impressioníst): Ivanova says сс or с. We have с.
- ингредие́нт (ingredijént): Ivanova says only [иэ́]; we have [ие́] followed by [иэ́].
- инду́сский (indússkij): Ivanova says сс or с; we have just сс, by phonetic rule.
- иннова́ция (innovácija), инновацио́нный (innovaciónnyj): Ivanova says only нн; we have only н.
- интервьюе́р (intervʹjujér): Ivanova says [интэрвьюэ́р] and [интэрвъюэ́р]; we have [интэрвьюэ́р] and [интэрвью́ер]. Possibly the variant in в is systematic (see исподло́бья below).
- инъе́кция (inʺjékcija): Ivanova says [иньекция] and [инъекция] (in that other); we have just [инъекция] (as per the spelling).
- ирреденти́зм (irredentízm): Not in Ivanova, but ruwikt says [дэ] while we say [де].
- исподло́бья (ispodlóbʹja): Ivanova says [исподло́бья] and [исподло́бъя]; we have just [исподло́бья] (as per the spelling).
- исчисле́ние (isčislénije): Ivanova says [щи] and [щчи]; we have only [щчи].
- исчи́слить (isčíslitʹ): Ivanova says [щи] and [щчи]; we have only [щи].
- Benwing2 (talk) 05:59, 22 May 2018 (UTC)
- @Atitarev, Cinemantique, Wikitiki89, Wanjuscha, Per utramque cavernam, Guldrelokk Benwing2 (talk) 05:59, 22 May 2018 (UTC)
- @Benwing2: My replies are after "-" in the order of preference
- изжо́га (izžóga): Ivanova appears to say жж or ж. We just have жж. - жж
- иллузио́н (illuzión), иллузиони́ст (illuzioníst), иллузиони́стский (illuzionístskij), иллю́зия (illjúzija), иллюзо́рный (illjuzórnyj): Ivanova says лл or л. We have лл for иллюзия, л for иллюзорный. - л or лл. Never heard "иллу-"
- иммигра́нт (immigránt), иммигра́нтка (immigrántka), иммиграцио́нный (immigraciónnyj), иммигра́ция (immigrácija), иммигри́ровать (immigrírovatʹ): Ivanova says мм or м. We have м. - м
- импрессиони́зм (impressionízm), импрессиони́ст (impressioníst): Ivanova says сс or с. We have с. - сс or с
- ингредие́нт (ingredijént): Ivanova says only [иэ́]; we have [ие́] followed by [иэ́]. - [ие́] or [иэ́]
- инду́сский (indússkij): Ivanova says сс or с; we have just сс, by phonetic rule. - сс or с
- иннова́ция (innovácija), инновацио́нный (innovaciónnyj): Ivanova says only нн; we have only н. - only нн
- интервьюе́р (intervʹjujér): Ivanova says [интэрвьюэ́р] and [интэрвъюэ́р]; we have [интэрвьюэ́р] and [интэрвью́ер]. Possibly the variant in в is systematic (see исподло́бья below). - all four, -ер is less common
- инъе́кция (inʺjékcija): Ivanova says [иньекция] and [инъекция] (in that other); we have just [инъекция] (as per the spelling). - [иньекция] or [инъекция]
- ирреденти́зм (irredentízm): Not in Ivanova, but ruwikt says [дэ] while we say [де]. - [дэ]
- исподло́бья (ispodlóbʹja): Ivanova says [исподло́бья] and [исподло́бъя]; we have just [исподло́бья] (as per the spelling). - both
- исчисле́ние (isčislénije): Ivanova says [щи] and [щчи]; we have only [щчи]. - [щчи] or [щи]
- исчи́слить (isčíslitʹ): Ivanova says [щи] and [щчи]; we have only [щи]. - [щчи] or [щи]
- --Anatoli T. (обсудить/вклад) 06:51, 22 May 2018 (UTC)
- @Atitarev Thanks! My mistake for writing иллу-; I meant иллю-, the corresponding spelling in English was interfering. Benwing2 (talk) 13:19, 22 May 2018 (UTC)
Pronunciation of various words beginning with э
[edit]I went through э in Ivanova. Please comment.
- эдельве́йс (edelʹvéjs): Ivanova appears to say [ве́] or [вэ́], we have it in the opposite order. Not so important, may only be an artifact of the way she specifies the pronunciation and may not be significant.
- экземпля́р (ekzempljár): Ivanova says [зе] or [зэ], we have only [зэ].
- экспре́сс (ekspréss): Ivanova appears to say [ре́] or [рэ́], we have it in the opposite order. Not so important, see above.
- экспре́ссия (ekspréssija): Here we have [ре́] or [рэ́], opposite to the order in экспре́сс. Presumably they should be consistent? Also, Ivanova says [сс] or [с], we say only [с].
- экспре́сс-диагно́стика (ekspréss-diagnóstika): Here we have only [ре́]; I assume this is an error.
- экспрессиони́зм (ekspressionízm): Here we have only [ре]; I assume this is an error again? Ruwikt says [ре] or also [рэ].
- электроста́нция (elektrostáncija): Ivanova lists secondary stress on элѐктро-, we don't have it.
- эми́ссия (emíssija): Ivanova says [сс] or [с], we say only [с].
- эпиде́рма (epidérma), эпиде́рмис (epidérmis): Ivanova says only [дэ́], we have [дэ́] or [де].
- эполе́т (epolét): We have this as a masculine singular, with genitive plural either эполе́тов or эполе́т. Ivanova lists this as feminine singular эполе́та with genitive plural эполе́т; she also lists masculine singular эполе́т and genitive plural эполе́тов, but marks them as dated.
- эре́кция (erékcija): Ivanova says [ре́] or [рэ́]; we say only [рэ́].
- эссе́ (essé): Ivanova says [ссэ́]; we say either [сэ́] or [ссэ́].
Thanks!
- @Atitarev, Cinemantique, Wikitiki89, Wanjuscha, Per utramque cavernam, Guldrelokk Benwing2 (talk) 21:07, 10 June 2018 (UTC)
- Given how variable the е/э pronunciation is in general, I would say that whenever Ivanova says both exist both deserve to be added. экспресс-диагностика (ekspress-diagnostika) and экспрессионизм (ekspressionizm) are wrong. электростанция (elektrostancija) has a secondary stress. I would definitely read эполе́т (epolét) as the genitive of эполе́та (epoléta), whatever it may be. сс in экспре́ссия (ekspréssija) and эми́ссия (emíssija) is normally pronounced long, as it is a strong position for geminates; I also have never heard /эсэ́/. Guldrelokk (talk) 21:27, 10 June 2018 (UTC)
@Benwing2: Also, what’s with the IPA of эполе́т (epolét)? It begins with [ɨ] instead of [ɪ]. Guldrelokk (talk) 21:40, 10 June 2018 (UTC)
- @Guldrelokk Not sure what you're referring to ... it appears as [ɨ] for me. Benwing2 (talk) 21:44, 10 June 2018 (UTC)
- @Benwing2: Initial unstressed и- and э- are never distinguished and are both pronounced [i~ɪ]; this merged /i/ becomes [ɨ] after prepositions ending in hard consonants: эполет [ипале́т], к эполету [кыпале́ту]. Guldrelokk (talk) 21:49, 10 June 2018 (UTC)
- @Guldrelokk: I see. Let's see what others say; if they agree, I can implement this. Note that Cinemantique borrowed Module:ru-pron into ruwikt but made a change to render all unstressed э (initial or not) as [ɛ], based on some source or other. Hence игра́ть [ɪˈɡratʲ] but эгре́т [ɛˈɡrɛt]. This seems strange to me but it's what ruwikt does. Benwing2 (talk) 21:56, 10 June 2018 (UTC)
- Initial э- is one of the major source of errors for Russian schoolchildren, who cannot tell it from и-. This is so unambiguous that I doubt any disagreement whatsoever can be found in any sources. Guldrelokk (talk) 22:00, 10 June 2018 (UTC)
- @Benwing2: Please implement initial "э" as [ɪ], rather than [ɨ] but it will change to [ɨ] after prepositions, just like the initial "и" per Guldrelokk. I don't agree with ruwikt's implementation but we need to fix this too. [ɨ] was incorrect, sorry, my mistake.
- I'm OK with Ivanova's assessment but palatalised экземпля́р and эре́кция sound weird to me. I don't object if add palatalised as alternatives. --Anatoli T. (обсудить/вклад) 22:33, 10 June 2018 (UTC)
- @Atitarev OK, will do. I think it already changes [ɪ] to [ɨ] after prepositions and such. Benwing2 (talk) 22:41, 10 June 2018 (UTC)
- @Atitarev, Guldrelokk Done. Benwing2 (talk) 23:49, 10 June 2018 (UTC)
- @Atitarev, Guldrelokk What about элѐктроэне́ргия? Does the second э also change to [ɪ]? If so, does this apply after all vowels within a word (stressed and unstressed)? Benwing2 (talk) 00:52, 11 June 2018 (UTC)
- Yes; yes. Ikanye eliminates the difference between any unstressed /э/ and /и/, which then follow the usual /и/ allophony pattern: [ы] after hard consonants, [и] otherwise. Guldrelokk (talk) 01:00, 11 June 2018 (UTC)
- @Guldrelokk Implemented. Benwing2 (talk) 01:25, 11 June 2018 (UTC)
- Yes; yes. Ikanye eliminates the difference between any unstressed /э/ and /и/, which then follow the usual /и/ allophony pattern: [ы] after hard consonants, [и] otherwise. Guldrelokk (talk) 01:00, 11 June 2018 (UTC)
- @Atitarev, Guldrelokk What about элѐктроэне́ргия? Does the second э also change to [ɪ]? If so, does this apply after all vowels within a word (stressed and unstressed)? Benwing2 (talk) 00:52, 11 June 2018 (UTC)
- @Atitarev, Guldrelokk Done. Benwing2 (talk) 23:49, 10 June 2018 (UTC)
- @Atitarev OK, will do. I think it already changes [ɪ] to [ɨ] after prepositions and such. Benwing2 (talk) 22:41, 10 June 2018 (UTC)
- Initial э- is one of the major source of errors for Russian schoolchildren, who cannot tell it from и-. This is so unambiguous that I doubt any disagreement whatsoever can be found in any sources. Guldrelokk (talk) 22:00, 10 June 2018 (UTC)
- @Guldrelokk: I see. Let's see what others say; if they agree, I can implement this. Note that Cinemantique borrowed Module:ru-pron into ruwikt but made a change to render all unstressed э (initial or not) as [ɛ], based on some source or other. Hence игра́ть [ɪˈɡratʲ] but эгре́т [ɛˈɡrɛt]. This seems strange to me but it's what ruwikt does. Benwing2 (talk) 21:56, 10 June 2018 (UTC)
- @Benwing2: Initial unstressed и- and э- are never distinguished and are both pronounced [i~ɪ]; this merged /i/ becomes [ɨ] after prepositions ending in hard consonants: эполет [ипале́т], к эполету [кыпале́ту]. Guldrelokk (talk) 21:49, 10 June 2018 (UTC)
Pronunciation of various words beginning with де-
[edit]I've gotten as far дем-.
- деви́чник (devíčnik): Ivanova says only [шн], we have [шн] and [чн].
- девяностоле́тие (devjanostolétije), девяностоле́тний (devjanostolétnij): Ivanova says no secondary stress, we have [но̀с], as with ruwikt.
- дегусти́ровать (degustírovatʹ), дегуста́ция (degustácija), дегуста́тор (degustátor): Ivanova says only [де], but ruwikt, Reznichenko and my paper dictionary all allow [дэ], so we currently have [де] and [дэ].
- де́за (déza) (from дезинформа́ция (dezinformácija)): We have [де́], ruwikt has [дэ́]. Not in Ivanova.
- дезактиви́ровать (dezaktivírovatʹ): We have [де] and [дэ] (the latter marked as less common), ruwikt has only [де], not in Ivanova.
- дезинтегра́ция (dezintegrácija): Ivanova says only [дэзынтэ]; we have [дэзынтэ] and [дезынтэ].
- дезинфе́кция (dezinfékcija), дезинфици́ровать (dezinficírovatʹ): Ivanova says [дезинф]; we have [дезынф].
- дезинформа́ция (dezinformácija): Ivanova says "[допуст. дэзын]"; I don't quite know whether that is intended to mean [дезин]/[дэзын], or something else. We have [дезын] or [дэзын]; ruwikt has the same.
- дезоксирибонуклеи́новый (dezoksiribonukleínovyj): We have the pronunciation exactly as the headword, with no secondary stress and no [э]. Ruwikt says [дэ] and [рѝбо].
- дека́да (dekáda): Ivanova says only [де]. We have [де] and [дэ], the latter marked as less common.
- деко́р (dekór): Ivanova says only [дэ]. We have [дэ] and [де], the latter marked as less common.
- де́ксель (dékselʹ): Not in any sources I can find. We have [е] throughout.
- дематериализа́ция (dematerializácija): Not in any sources I can find except ruwikt. We have unstressed [де], ruwikt has secondary-stressed [дэ̀].
- де́мбель (démbelʹ): Not in any sources I can find except ruwikt. We have [де́], ruwikt has [де́] or [дэ́].
- демиу́рг (demiúrg): Not in Ivanova. Avanesov says [дэ]. We have [де] or [дэ], which agrees with ruwikt.
- демобилиза́ция (demobilizácija), демобилизова́ть (demobilizovátʹ): Ivanova says only [де]. We have [де] or [дэ].
@Atitarev, Cinemantique, Wikitiki89, Wanjuscha, Per utramque cavernam, Guldrelokk Any comments welcome. Benwing2 (talk) 05:16, 11 June 2018 (UTC)
- Zarva: "дезоксирѝбонуклеи́новый [дэ]".--Cinemantique (talk) 05:29, 11 June 2018 (UTC)
- "дембель" in [Yeskova 2015].--Cinemantique (talk) 05:43, 11 June 2018 (UTC)
@Benwing2:
- деви́чник (devíčnik): both [шн] and [чн].
- девяностоле́тие (devjanostolétije), девяностоле́тний (devjanostolétnij): secondary stress
- дегусти́ровать (degustírovatʹ), дегуста́ция (degustácija), дегуста́тор (degustátor): both, as with the rest with prefix де-
- де́за (déza) (from дезинформа́ция (dezinformácija)): both
- дезактиви́ровать (dezaktivírovatʹ): both, as with the rest with prefix дез-
- дека́да (dekáda): ours is correct
- деко́р (dekór): per Ivanova
- де́ксель (dékselʹ): both
- дематериализа́ция (dematerializácija): both
- де́мбель (démbelʹ): per Cinemantique
- демиу́рг (demiúrg): both
- демобилиза́ция (demobilizácija), демобилизова́ть (demobilizovátʹ): both
--Anatoli T. (обсудить/вклад) 10:48, 11 June 2018 (UTC)
Outstanding failed cases
[edit](moved from Module talk:ru-pron/Archive 1)
@Benwing2: There are some outstanding cases. This one seems to be based on straightforward rules, which I explained:
- в Япо́нии /v‿jɪˈponʲɪɪ/
Let me know if you need a refresher. I know you're busy with other things, just a friendly reminder that some of these need to be fixed eventually. I don't agree with some other cases. --Anatoli T. (обсудить/вклад) 07:01, 22 June 2018 (UTC)
- @Atitarev I do remember this one and it should be easy to fix. I'll try to get to it soon. Benwing2 (talk) 07:07, 22 June 2018 (UTC)
- @Benwing2: Thanks. For "о̀ргкомите́т" /ˌorkːəmʲɪˈtʲet/ to work, you need a "gem=y" parameter, not sure it's available in the test module. --Anatoli T. (обсудить/вклад) 07:09, 22 June 2018 (UTC)
- @Atitarev It is in fact available and I added it. Not sure whether this particular test does any good but at least it doesn't signal an error now. Benwing2 (talk) 07:21, 22 June 2018 (UTC)
- @Atitarev Fixed в Япо́нии. Benwing2 (talk) 16:25, 22 June 2018 (UTC)
- @Benwing2: Thanks. For "о̀ргкомите́т" /ˌorkːəmʲɪˈtʲet/ to work, you need a "gem=y" parameter, not sure it's available in the test module. --Anatoli T. (обсудить/вклад) 07:09, 22 June 2018 (UTC)
з before palatal м
[edit](Notifying Atitarev, Cinemantique, KoreanQuoter, Useigor, Wanjuscha, Wikitiki89, Stephen G. Brown, Per utramque cavernam, Guldrelokk): Ivanova says that in пла́зменный (plázmennyj) the з is optionally palatalized. Does this apply to all зм + palatal vowel? Some examples of such words:
- безвозмездный
- безмерный
- безмятежный
- взметнуть
- змея, змей, змий, змеевик, etc.
- измельчать
- измена
- измять
- оргазмический
- размежевать
- размер
- соизмеримый
- чрезмерный
Benwing2 (talk) 03:55, 29 June 2018 (UTC)
- @Benwing2: This applies to all sequences of dental + soft labial (зверь (zverʹ), дверь (dverʹ), сменить (smenitʹ) and so on also here), including over boundaries with prepositions. It’s a part of the dated pronunciation. Guldrelokk (talk) 04:04, 29 June 2018 (UTC)
- I think this pronunciation is dated and we don't have to show it.--Cinemantique (talk) 04:08, 29 June 2018 (UTC)
- (Notifying Benwing2, Cinemantique, KoreanQuoter, Useigor, Wanjuscha, Wikitiki89, Stephen G. Brown, Per utramque cavernam, Guldrelokk): : I think the "dated" part varies from speaker to speaker and word to word, e.g. perfectly OK in измена (izmena), зверь (zverʹ). We can include this as optional. It now sounds funny or old but /z/ in words with the suffix -изм (-izm) was also palatalised. --Anatoli T. (обсудить/вклад) 04:37, 29 June 2018 (UTC)
- I support adding (j) or something like this, but not for -изм (-izm): it was a common pronunciation, but now it’s not, and throughout its life it was heavily proscribed, unlike the dental-labial assimilation which was preferred. Guldrelokk (talk) 04:42, 29 June 2018 (UTC)
- @Guldrelokk: No кли́зьма от катакли́зьма (klízʹma ot kataklízʹma), LOL? --Anatoli T. (обсудить/вклад) 04:48, 29 June 2018 (UTC)
- Jokes aside, I have to make an effort NOT to palatalise ни́зменность (nízmennostʹ). --Anatoli T. (обсудить/вклад) 04:50, 29 June 2018 (UTC)
- Since it sounds like it's dated in most cases, I won't show it unless people specifically want it. Benwing2 (talk) 06:11, 29 June 2018 (UTC)
- @Guldrelokk, Atitarev, Cinemantique Benwing2 (talk) 06:13, 29 June 2018 (UTC)
- It’s not that dated, people still speak so, especially old people. Guldrelokk (talk) 06:35, 29 June 2018 (UTC)
- @Benwing2: I guess we want it, pls. --Anatoli T. (обсудить/вклад) 06:38, 29 June 2018 (UTC)
- It’s not that dated, people still speak so, especially old people. Guldrelokk (talk) 06:35, 29 June 2018 (UTC)
- Jokes aside, I have to make an effort NOT to palatalise ни́зменность (nízmennostʹ). --Anatoli T. (обсудить/вклад) 04:50, 29 June 2018 (UTC)
- @Guldrelokk: No кли́зьма от катакли́зьма (klízʹma ot kataklízʹma), LOL? --Anatoli T. (обсудить/вклад) 04:48, 29 June 2018 (UTC)
- I support adding (j) or something like this, but not for -изм (-izm): it was a common pronunciation, but now it’s not, and throughout its life it was heavily proscribed, unlike the dental-labial assimilation which was preferred. Guldrelokk (talk) 04:42, 29 June 2018 (UTC)
- (Notifying Benwing2, Cinemantique, KoreanQuoter, Useigor, Wanjuscha, Wikitiki89, Stephen G. Brown, Per utramque cavernam, Guldrelokk): : I think the "dated" part varies from speaker to speaker and word to word, e.g. perfectly OK in измена (izmena), зверь (zverʹ). We can include this as optional. It now sounds funny or old but /z/ in words with the suffix -изм (-izm) was also palatalised. --Anatoli T. (обсудить/вклад) 04:37, 29 June 2018 (UTC)
гетеросексуальный and similar words
[edit](Notifying Atitarev, Cinemantique, KoreanQuoter, Useigor, Wanjuscha, Wikitiki89, Stephen G. Brown, Per utramque cavernam, Guldrelokk): I noticed that ruwikt lists both [се] and [сэ] as possibilities, with [се] first, while we have only [сэ]. (Note, I changed гетеросексуальный itself, but not any other such words.) Ivanova seems to agree; she has [допуст. сэ], which implies [се] and [сэ], by all words beginning with секс- except for секс itself, which simply says [сэ]. Words such as гетеросексуальный and транссексуальный have no indication concerning the се, which would imply only [се], but I assume this is an oversight. Is she correct in that all words with -секс- allow both [се] and [сэ], and is секс a special case where only [сэкс] is possible, or can it also be pronounced [секс]? Thanks! Benwing2 (talk) 21:56, 8 July 2018 (UTC)
- Also, (1) is it [гѐтэро-] or [гетэ̀ро-] or both (and if both, which should be listed first), (2) is it [тра̀нс-] or unstressed [транс-] or both (and if both, which should be listed first), (3) does трансс- have long с? Ivanova says only [гетэ̀ро-], [тра̀нс-] and isn't clear about the pronunciation of double сс. Benwing2 (talk) 22:04, 8 July 2018 (UTC)
- @Benwing2: [секс] sounds weird in all words but some people do pronounce it this way - rare, I think. Make [гетэ̀ро-] the only possibility, if you wish. I pronounce [гѐтэро-] but I now wonder if it's standard. We should use [тра̀нс-] and [s] is geminated. --Anatoli T. (обсудить/вклад) 22:28, 8 July 2018 (UTC)
- @Atitarev I trust your intuitions. If you say [гѐтэро-], probably others do as well. Note that ruwikt is inconsistent: Under гетеросексуал they have only [гетэ̀ро-] and put [секс] before [сэкс], but in гетеросексуальность they have both [гетэ̀ро-] and [гѐтэро-], and [сэкс] before [секс]. Benwing2 (talk) 22:33, 8 July 2018 (UTC)
- @Benwing2: TBH, I never heard [гетэ̀ро-]. On [тра̀нс-], please use the form without the secondary stress as well, sorry, changed my mind :) (The shorter the word, the more likely - no secondary stress is used but it's seldom intuitive enough.) --Anatoli T. (обсудить/вклад) 23:02, 8 July 2018 (UTC)
- @Atitarev Thanks. With secondary stress I've gathered that if syllables X (where the secondary stress might go) and Y (where the primary stress is) are separated by 3 or more unstressed syllables, then X usually gets secondary stress, whereas if there are 2 unstressed syllables, X sometimes gets secondary stress, and if there are 0 or 1 syllables then secondary stress is rare. This is different from English, which doesn't like sequences of 3 or more unstressed syllables (cf. àntidìsestàblishmentárianìsm with 4 secondary stresses and no sequences of 3 or more unstressed syllables) but occasionally tolerates sequences of 3 unstressed syllables (as in ìnternàtionalizátion). Benwing2 (talk) 23:12, 8 July 2018 (UTC)
- @Benwing2: Thanks, I need some coffee to follow this :) Anyway, транссексуа́льный (transsɛksuálʹnyj) is leaning towards using a secondary stress but трансге́ндерный (transgéndɛrnyj) is not. BTW, are you interested in working on Appendix:Russian pronunciation with my assistance? No rush. --Anatoli T. (обсудить/вклад) 23:23, 8 July 2018 (UTC)
- @Atitarev Sure, I'll work on that page. Benwing2 (talk) 23:47, 8 July 2018 (UTC)
- @Benwing2: Thanks, I need some coffee to follow this :) Anyway, транссексуа́льный (transsɛksuálʹnyj) is leaning towards using a secondary stress but трансге́ндерный (transgéndɛrnyj) is not. BTW, are you interested in working on Appendix:Russian pronunciation with my assistance? No rush. --Anatoli T. (обсудить/вклад) 23:23, 8 July 2018 (UTC)
- @Atitarev Thanks. With secondary stress I've gathered that if syllables X (where the secondary stress might go) and Y (where the primary stress is) are separated by 3 or more unstressed syllables, then X usually gets secondary stress, whereas if there are 2 unstressed syllables, X sometimes gets secondary stress, and if there are 0 or 1 syllables then secondary stress is rare. This is different from English, which doesn't like sequences of 3 or more unstressed syllables (cf. àntidìsestàblishmentárianìsm with 4 secondary stresses and no sequences of 3 or more unstressed syllables) but occasionally tolerates sequences of 3 unstressed syllables (as in ìnternàtionalizátion). Benwing2 (talk) 23:12, 8 July 2018 (UTC)
- @Benwing2: TBH, I never heard [гетэ̀ро-]. On [тра̀нс-], please use the form without the secondary stress as well, sorry, changed my mind :) (The shorter the word, the more likely - no secondary stress is used but it's seldom intuitive enough.) --Anatoli T. (обсудить/вклад) 23:02, 8 July 2018 (UTC)
- @Atitarev I trust your intuitions. If you say [гѐтэро-], probably others do as well. Note that ruwikt is inconsistent: Under гетеросексуал they have only [гетэ̀ро-] and put [секс] before [сэкс], but in гетеросексуальность they have both [гетэ̀ро-] and [гѐтэро-], and [сэкс] before [секс]. Benwing2 (talk) 22:33, 8 July 2018 (UTC)
- (Notifying Benwing2, Cinemantique, KoreanQuoter, Useigor, Wanjuscha, Wikitiki89, Stephen G. Brown, Per utramque cavernam, Guldrelokk): : Guys, I really want to know the source of [гетэ̀ро-]. I only ever heard [гѐтэро-]. Also, if the community thinks we should allow [секс] as an alternative pronunciation, so be it. --23:57, 8 July 2018 (UTC)
- The sources are Yeskova 2015, Zarva 2001, Kalenchuk/Kasatkina 2001 (гѐтеро "в непринуждённой речи" - ?), Reznichenko 2003 (гѐтеро "в беглой речи" - in fluent speech?), Ivanova 2005. The order Е vs. Э in Ruwiki doesn't matter, I just put Е first.--Cinemantique (talk) 02:41, 9 July 2018 (UTC)
- @Cinemantique, Benwing2: @Cinemantique, great, thank you. I would translate "в непринуждённой речи" as "in casual speech" and "в беглой речи" as "in fast speech". --03:01, 9 July 2018 (UTC)
- @Atitarev, Cinemantique (Anatoli, didn't get your ping because your signature is missing) I fixed up the гетеросекс- words to have [гетэ̀ро-] followed by [гѐтэро-] but kept -секс- as only [сэкс]. If you guys really want it changed to list [секс] as well, I will do so. Benwing2 (talk) 03:49, 9 July 2018 (UTC)
- @Benwing2, Cinemantique: I replied, so no reply was missed. Re: [гѐтэро-] vs [гетэ̀ро-] IMO, what was in "fast/casual speech" ([гѐтэро-]) is now more common but keep it as is, it's IMO only without any evidence. --Anatoli T. (обсудить/вклад) 03:55, 9 July 2018 (UTC)
- @Atitarev, Cinemantique (Anatoli, didn't get your ping because your signature is missing) I fixed up the гетеросекс- words to have [гетэ̀ро-] followed by [гѐтэро-] but kept -секс- as only [сэкс]. If you guys really want it changed to list [секс] as well, I will do so. Benwing2 (talk) 03:49, 9 July 2018 (UTC)
- @Cinemantique, Benwing2: @Cinemantique, great, thank you. I would translate "в непринуждённой речи" as "in casual speech" and "в беглой речи" as "in fast speech". --03:01, 9 July 2018 (UTC)
- The sources are Yeskova 2015, Zarva 2001, Kalenchuk/Kasatkina 2001 (гѐтеро "в непринуждённой речи" - ?), Reznichenko 2003 (гѐтеро "в беглой речи" - in fluent speech?), Ivanova 2005. The order Е vs. Э in Ruwiki doesn't matter, I just put Е first.--Cinemantique (talk) 02:41, 9 July 2018 (UTC)
зж
[edit](moved from Talk:бздеть)
(Notifying Atitarev, Cinemantique, Useigor, Wikitiki89, Stephen G. Brown, Guldrelokk, Fay Freak, Tetromino, Per utramque cavernam): Do the forms бзжу, пизжу, спизженный, etc. have ӂӂ as a possible pronunciation, i.e. IPA(key): [bʑːu], IPA(key): [ˈspʲiʑːɪn(ː)ɨj], etc.? I'm guessing yes based on other verbs, but you can't find vulgar verbs like this in any of the pronouncing dictionaries. Benwing2 (talk) 04:53, 12 January 2019 (UTC)
- @Benwing2: Definite yes for for пизжу́, спи́зженный and probably yes for "бзжу". I just haven't heard it. --Anatoli T. (обсудить/вклад) 12:34, 12 January 2019 (UTC)
- @Benwing2 Yea, for many speakers any position which is written -зж- is pronounced palatalized (written /ʑː/ here) – an often case of such a cluster which one hears pronounced so is the 1st person singular present of verbs. It can just be assumed automatically. Speakers who have the habit to pronounce (/ʑː/) pronounce it when they read aloud something written -зж-. Fay Freak (talk) 12:09, 13 January 2019 (UTC)
- Does this also apply across prefix boundaries like изжа́рить? I can't find any evidence it does. Benwing2 (talk) 13:39, 13 January 2019 (UTC)
- @Benwing2 Neither it is pronounced /ʐː/ in изжа́рить but /zʐ/ in the unpalatalized pronunciation, when it is pronounced averagely careful, so … it can happen in fast speech but it is not a usual speech we want to describe, or it cannot be recommended. Like хо́дит (xódit) can be pronounced [ˈxoɪt] instead of [ˈxodʲɪt] by lazy tongues but we don’t mention this. I find it dubious how изжива́ть (izživátʹ) is given [ɪʐːɨˈvatʲ] instead of [ɪzʐɨˈvatʲ] but maybe ask others – I can only say that I consider [ɪzʐɨˈvatʲ] normative for изжива́ть (izživátʹ), not [ɪʐːɨˈvatʲ] or [ɪʑːɨˈvatʲ]. Others should answer if this pronunciation across boundaries [ʐː] is somewhere (in Moscow dialect?) fashionable or acclaimed. It doesn’t seem proper to me. Didn’t ever get the idea of pronouncing it that way, and I speak Russian everyday. (Also the first syllable seems /i/ and not /ɪ/ to me, who should know the difference from the use of ɪ in German, but I have ignored this here. We read in Vowel reduction in Russian#High vowels that “Two high vowels /u/ and /i/ are usually thought to undergo no reduction” and it is true that they don’t, it is thought and done so – only because one collected some speech samples for which it is untrue one cannot derive an opposite rule). Fay Freak (talk) 15:32, 13 January 2019 (UTC)
- [ɪʐːɨˈvatʲ] is standard pronunciation. See Avanesov, §14.3.--Cinemantique (talk) 16:15, 13 January 2019 (UTC)
- @Benwing2 Neither it is pronounced /ʐː/ in изжа́рить but /zʐ/ in the unpalatalized pronunciation, when it is pronounced averagely careful, so … it can happen in fast speech but it is not a usual speech we want to describe, or it cannot be recommended. Like хо́дит (xódit) can be pronounced [ˈxoɪt] instead of [ˈxodʲɪt] by lazy tongues but we don’t mention this. I find it dubious how изжива́ть (izživátʹ) is given [ɪʐːɨˈvatʲ] instead of [ɪzʐɨˈvatʲ] but maybe ask others – I can only say that I consider [ɪzʐɨˈvatʲ] normative for изжива́ть (izživátʹ), not [ɪʐːɨˈvatʲ] or [ɪʑːɨˈvatʲ]. Others should answer if this pronunciation across boundaries [ʐː] is somewhere (in Moscow dialect?) fashionable or acclaimed. It doesn’t seem proper to me. Didn’t ever get the idea of pronouncing it that way, and I speak Russian everyday. (Also the first syllable seems /i/ and not /ɪ/ to me, who should know the difference from the use of ɪ in German, but I have ignored this here. We read in Vowel reduction in Russian#High vowels that “Two high vowels /u/ and /i/ are usually thought to undergo no reduction” and it is true that they don’t, it is thought and done so – only because one collected some speech samples for which it is untrue one cannot derive an opposite rule). Fay Freak (talk) 15:32, 13 January 2019 (UTC)
- Does this also apply across prefix boundaries like изжа́рить? I can't find any evidence it does. Benwing2 (talk) 13:39, 13 January 2019 (UTC)
(Notifying Atitarev, Cinemantique, Useigor, Wikitiki89, Stephen G. Brown, Guldrelokk, Fay Freak, Tetromino, Per utramque cavernam): Comments? Is assimilation of зж across a prefix boundary non-normative? What about e.g. сжечь (sžečʹ), расши́рить (rasšíritʹ)? Benwing2 (talk) 15:38, 13 January 2019 (UTC)
- Such assimilation is not normative or even not in use.--Cinemantique (talk) 16:15, 13 January 2019 (UTC)
- @Cinemantique: I think Benwing2 means the assimilation [зж] > [жж], which is definitely in use and also normative, as demonstrated by your link to Avanesov. Do you mean something else? Guldrelokk (talk) 16:42, 13 January 2019 (UTC)
- I thought Benwing2 meant the assimilation [zʐ] > [ʑː] "across a prefix boundary".--Cinemantique (talk) 17:11, 13 January 2019 (UTC)
- @Cinemantique: I think Benwing2 means the assimilation [зж] > [жж], which is definitely in use and also normative, as demonstrated by your link to Avanesov. Do you mean something else? Guldrelokk (talk) 16:42, 13 January 2019 (UTC)
- I strongly disagree. Pronunciation like [zʐ] is completely unnatural and is not unconsciously produced by any native speaker in existence. ‘Careful’ pronunciation also lacks vowel reduction and final obstruent devoicing – this is simply reading out words letter-by-letter and not Russian speech. Native speakers are unaware of living phonological processes, but 100% of studies demonstrate that sibilant clusters with final retroflex always assimilate in natural speech. This is also prescribed by all prescriptive authorities, like Avanesov (Орфоэпический словарь русского языка, 4 ed., p. 675):
Согласные [с] и [з] называются свистящими в отличие от согласных [ш], [ж], [ч] — шипящих. Сочетание свистящего согласного с последующим шипящим произносится всегда как двойной шипящий согласный. Это сочетание встречается на стыке разных морфологических частей слова, а также предлога и следующего слова.
сш и зш: [рашшытый] (расшитый), [рашшырил] (расширил), [биешшу́мный] (бесшумный), [вазро́шшый] (возросший), [вле́шшый] (влезший), [шшу́мъм] (с шумом), [6иешша́пки] (без шапки);
жж и зж: [жжок] (сжёг), [жжал] (сжал), [ижжа́рил] (изжарил), [6иежжалъсный] (безжалостный), [жжа́ръм] (с жаром).
- сжечь and изжа́рить are universally pronounced with [ʐː]. For this reason, no letter sequences should be automatically converted to [ʑː]. Guldrelokk (talk) 16:35, 13 January 2019 (UTC)
These cases зж, сжь, сш, сщ are unlikely to be come out (normatively) separate. Though I see differences, and that’s even in Moscow: I have checked how Кровосто́к pronounces it in the last three albums, it being Moscow rap:
- Lua error in Module:quote at line 2959: Parameter 1 is required. – clearly assimilated, but syllable boundary, so maybe it is to be written [ʐʐ] and not [ʐː], i. e. the first syllable ends with [ʐ] as the second begings with [ʐ]
- Lua error in Module:quote at line 2959: Parameter 1 is required. – [s] glides into [ʂ] through a syllable boundary
- Lua error in Module:quote at line 2959: Parameter 1 is required. – [s] glides into [ɕ] through a syllable boundary, but the [s] was already quite postalveolar – however it did not begin palatalized
- Lua error in Module:quote at line 2959: Parameter 1 is required. – [s] glides into [ʂ] through a syllable boundary
The album Студень didn’t even have a syllable boundary of such a cluster. It might come out that normatively there isn’t even a rule because the cluster is too rare across morphological boundaries for speakers to assume a norm. But then [ɪzʐɨˈvatʲ] is a viable pronunciation too. However the statement “произносится всегда как двойной шипящий согласный” oversimplifies phonetic reality. Встречается, but by far not всегда or universally. The свистящие do assimilate in their articulatory position to the following шипящие, that is the tongue is in position to go over to the next fricative (so [s] there is not the same as [s] elsewhere) but they are still distinct – like vowels are coloured by the surrounding consonants. Look also at исче́знуть (isčéznutʹ). The pronunciation given at исче́з (isčéz) [ɪˈɕːes] is sub-standard. [ɪɕˈt͡ɕes] can exist but [ɪsˈt͡ɕes] is clearly normal. That Avanesov claims the same for ч is totally strange. While [rɐˈʐːɛt͡ɕ] is not wrong (though not the standard) [ɪˈɕːeznʊtʲ] is mispronounced or at the best dialectal at best imho, of course the case is different when another alveolar plosive is in between the two fricative phonemes. Non-Slavic foreigners and lazy people would like to have less consonant clusters so they would rather see [ɪˈɕːeznʊtʲ] even when [ɪsˈɕːeznʊtʲ] should be attained, so academics start to see it to titillate underperformers, since they live on their being taught by the system too. I heard phonetic professors claiming that German Anfang is actually pronounced [ˈaɱ.faŋ] instead of [ˈan.faŋ], even claiming “people just can’t pronounce the alveolar nasal in such a position”. Neophilia of academia, plus being situated in certain places where the universities are makes academics biased. They want to inculcate us norms that have always seen different because academic career demands to write new things lead them to write exceptions to rules and because they don’t want to go to remote Sibirian villages or some Old Believers, or Germany to find out how people speak here. Nor do they look how it was before the statist system tried to impose everything up to the pronunciation: How was the norm of Saint Petersburg in tsarist times? Before the Bolsheviks have purged the nobles more people likely tried to assimilate not. We might find multiple layers of pronunciation: More Sovietic pronunciations trying to conform to underperformers being still strong and popular, but revived free pronunciation that is maybe even rarer but nonetheless more standard. Actually pronunciation is politically slanted. Fay Freak (talk) 17:24, 13 January 2019 (UTC)
- Now this sounds like a conspiracy theory. [ɪˈɕːeznʊtʲ] and [ɪɕˈtɕeznʊtʲ] are attested in running speech by phoneticians, *[ɪsˈɕːeznʊtʲ] is not. This has nothing to do with ‘neophilia’. Songs are not an example of running speech.
- Contrary to what you assume, pre-revolutional norm had much more living assimilation processes, including the reverse sibilant assimilation: мужско́й, for example, was pronounced [мус(:)ко́й] and is transcribed so by the contemporary linguists. See Панов, История русского литературного произношения XVIII-XX вв.. The likbez seriously affected the norm, bringing it closer to the orthography. Guldrelokk (talk) 17:53, 13 January 2019 (UTC)
- Here is the relevant section. Abbreviations: ЙЛ — Lundell J. A. Etudes sur la prononciation Russe, Upsala, 1890; ДУ 11 — Ушаков Д. Н. Русское правописание, Москва, 1911; ПШ — Шаров П. Г. Образцовое русское произношение, Москва, 1914; ВЧ 15 — Чернышев В. И. Законы и правила русского произношения, Петроград, 1915; РК — Кошутић Р. Грамматика руског језика, Петроград, 1919 and so on. Are these Soviet scholars?
- Is Daniel Jones a Soviet scholar? He collected his data in pre-revolutional St. Petersburg, and like all observers, he noticed the assimilation. Here is his IPA example from 1912 with [ʃ̇ṫ͜ʃ̇i`tɑt:sə] (modern IPA [ɕt͡ɕiˈtät͡sːə]) for счита́ться. He further published a book on Russian phonetics in 1923. Guldrelokk (talk) 18:27, 13 January 2019 (UTC)
Another good indication of the actual pronunciation are mistakes children and badly educated adult people do: ищез, исчу. These people do not get grants for ‘writing new things’ – they struggle to write correctly and fail because there is no difference in pronunciation between исчез and *ищез; the way it is written must be memorised. Guldrelokk (talk) 19:02, 13 January 2019 (UTC)
- No reason to assume that running speech, as you call it, is what forms the norms: What the linguists write about pronunciation is only a mention and not the use and suffers the bias I have formulated and the bias of running speech since the “running speech” that those populares collect contains all kinds of tramps, mountebanks and showmen instead of the most careful speech, then they generalize just one even though there are many we should count, like Alfred Kinsey to find out “the sexual behaviour of the human male/female” particularly asked convicted pedophiles and generalized it. So we find claims like about Postalveolar Fricatives in Slavic Languages as Retroflexes “From the four languages examined, Polish and Russian were shown phonetically to have retroflex fricatives, as supported by phonological evidence. Bulgarian, on the other hand, has a non-retroflex, palatoalveolar fricative” […] “If Keating’s claim (1991) is right that Serbian also has a retroflex fricative (recall the introduction), this might refute the hypothesis, as Serbian is a South Slavic language and should have a palato-alveolar according to this hypothesis.” How can we even consider that a language has only one or the other? They calculate the median value though people can have varying pronunciation inside one of the languages, so it is for some [ʒ] and for others [ʐ], like /r/ is varying according to speaker (well-known in Germany) as [r] and [ɾ] [a] [ʀ] and [ʁ] (the last happens in Russia too somehow regularly though being less standard than elsewhere). I just don’t believe we have one pronunciation only. The premise has been wrong and academics found what they wanted to find: They asked if the sound is [ʒ] or [ʐ] and the answer was only one of two if they calculated the mean value. The same method has been applied with this assimilation thing. Teachers often tell their pupils that something is pronounced “thus” even though it is much more complicated. The question then is how the norm, the standard (even if multiple standards or a norm-ladder, an accepted pronunciation, another accepted one, unaccepted ones) can be found out. Is it some phoneticians instead of the doxa?
- It is dubious that you’ll rely on secondary sources so much that you even base your claims upon them contrary to primary evidence of the Russian language that you have yourself by having perceived Russian around you. Surely [ɪsˈɕeznʊtʲ] exists, I pronounce it like that, and is there are a reason not to? No, so others do too. Actually we know a priori that [ɪsˈɕeznʊtʲ] exists because of the desire to distinguish and it is not falsified. One does not need to have read any Russian phoneticians to know how Russian is pronounced and he who hasn’t read them knows Russian pronunciation better than he who has if the latter adopts the hearsay in phonetic books, the mere catchphrases of “Russian is pronounced X” instead of considering the whole picture: It’s not a conspiracy, it is hivemind. Contemporary linguists of Tsar’s times had the same structural problems – the populist credo had also been widespread before, hence the Bolsheviks could win. People wanted everything to be average and to avoid inequality, inequal pronunciation – of course not only Russians.
- The reason people struggle with writing is that [s] does not equal [s] as I have said: There is some assimilation and some already choose a different sign though strictly it is still alveolar – assuming it can be that strict and their oral cavities aren’t special: it might even be intentional, you underestimate people’s desires to write non-standard. Fay Freak (talk) 19:07, 13 January 2019 (UTC)
- I do not need to rely on secondary sources to know that *[ɪsˈɕeznʊtʲ] can exist only as an overpronunciation, never occurs in unconscious speech and should not be represented in the dictionary, just like [kod] for код (kod), though surely exists as an overpronunciation, should not be represented. I can rely on secondary sources to falsify your claim that the natural observation that the assimilation always occurs in a given context has anything to do with politics or ‘neophilia’. Your commentary about the ‘populist credo’ is almost ridiculous: Šarov, for example, was a strictly prescriptive phonetician, and his book Образцовое русское произношение, as can be deduced already from its title, reflects the ‘exemplary’ theatrical pronunciation.
- I certainly do not overestimate my former classmates’ desire to get good scores at school: they struggled to write correctly, but were unable to, simply because many orthographic differences have no counterpart in the pronunciation. Сч ~ зч ~ щ at morpheme boundaries were one of the many troubles they had. I might had it myself as well – I do not remember now. Guldrelokk (talk) 19:24, 13 January 2019 (UTC)
- What’s unconscious speech? I would think it’s something one does when one sleeps. Otherwise I don’t see why the conscious speech and the “overpronunciation” should not be represented, or why natural speech is to be given as normative, that’s the naturalistic fallacy. I thought conscious speech, unnatural speech is exactly the pronunciation that should be represented, and common less conscious ones then since the space is there, but unfortunately Russian conscious speech seems to be understudied. For German entries Theaterdeutsch is most high-prestige and therefore given as standard, but is it now “too conscious”? Modern Standard Arabic is in a way required to be always conscious since if it is not what is spoken is a dialect and gets a different language code. – Some assimilations are more required than others and we could think more about how the alleged norms are constructed by us. But maybe there lack the investigations across regions and social groups to make more statements about unconscious and conscious speech separately. So how is -зж- across morpheme boundaries pronounced unconsciously and consciously, also palatalized? Since in words like визжа́ть even conscious speech palatalizes, whereas I did not know this about conscious speech over syllable boundary and myself who always speaks consciously and unnaturally speak there without assimilation, nor do I know how it is commonly in the so-called unconscious speech. Only I know that is like that in even less careful speech. It is just news to me that Russian-Wiktionarians think that not the most conscious speech has to be given as the top-tier. Fay Freak (talk) 19:56, 13 January 2019 (UTC)
- @Guldrelokk, Atitarev Thanks for the Avanesov citations. Based on this I'll assume the current assimilation of сж, зж -> [жж] is correct. In terms of rendering зж and жж as [ʑː], one thing that's possible is to have the pronunciation module automatically render [ʑː] as a second, additional pronunciation (listing [ʐː] first) everywhere *except* across prefix boundaries. The code already knows the concept of a prefix boundary and handles it specially (in particular, geminated consonants are normally pronounced double across a prefix boundary, but often not elsewhere), and is generally quite accurate in knowing where these boundaries occur. If I did this, I would probably add a parameter
|zhpal=y
to force the inclusion of [ʑː] as a secondary pronunciation and correspondingly|zhpal=n
to force it to not occur, for exceptional situations or in case the module wrongly predicts a prefix boundary. Some questions:- Do you agree with doing this?
- Should it apply to written жж as well as зж? I assume so; there aren't so many occurrences of written жж but all of them (e.g. вожжа́ (vožžá), дро́жжи (dróžži), жужжа́ть (žužžátʹ), можжеве́льник (možževélʹnik), сожже́ние (sožžénije), verb forms like жжёшь (žžošʹ), разожжём (razožžóm)) have [ʑː] listed as a possibility.
- Is it ever appropriate to list the [ʑː] pronunciation first, before [ʐː]? Currently езжа́ть (jezžátʹ) and derived verbs generally list [ʑː] first, but most other cases list [ʐː] first. I'm guessing that for speakers who use [ʑː], they use it everywhere it's possible to do so, and for speakers who don't use it, they never use it, and it doesn't vary depending on the given lemma; but I may be wrong. Benwing2 (talk) 19:42, 13 January 2019 (UTC)
- @Benwing2: Yes, that would be appropriate. Should work for both ‹зж› and ‹жж› inside morphemes.
- On the third point: unfortunately, it does vary: e. g. I only say е[ӂӂ]ать, дро[ӂӂ]и, but [жж]ёт and мо[жж]евельник. Probably the data on the distribution is unavailable, so it still makes more sense to use some fixed order. Any order would be correct. Many handbooks suggest that [жж] is the more common pronunciation (although I cannot confirm that); [ӂӂ] is favoured by prescriptivists like Avanesov. Guldrelokk (talk) 19:54, 13 January 2019 (UTC)
- @Benwing2 The speakers who pronounce -зж- as [ʑː] do not necessarily. pronounce -жж- so too. The former is more well-known. And no, I don’t think it is appropriate to list a [ʑː] pronunciation first since it is thought to be more standard to pronounce [ʐː] even if the former is more common. Fay Freak (talk) 19:56, 13 January 2019 (UTC)
- @Guldrelokk I have now implemented this and removed the manual ӂӂ pronunciations. Benwing2 (talk) 01:21, 14 January 2019 (UTC)
Error
[edit]@Benwing2: the Module is causing an error at паяльник (pajalʹnik), писклявый (piskljavyj), and трухлявый (truxljavyj). —Mahāgaja · talk 15:11, 14 May 2019 (UTC)
- @Mahagaja I don't see the errors. Benwing2 (talk) 15:19, 14 May 2019 (UTC)
@Benwing2: The entries are being put into CAT:IPA pronunciations with invalid IPA characters. If you open the Pronunciation section, you'll see a notice saying "invalid IPA characters ()" where what's between the parentheses is U+200E, the LEFT-TO-RIGHT MARK. It seems to be right at the beginning, between the opening square bracket and the first letter. —Mahāgaja · talk 15:26, 14 May 2019 (UTC)- Never mind, the problem was in the entries themselves, not in the module. —Mahāgaja · talk 15:28, 14 May 2019 (UTC)
Adding links to rhymes
[edit]I've been trying to expand the Russian Rhymes section, but it seems like the easiest way to automate it is to make the Rhymes pages Categories and automatically add an entry to a rhyme Category based on the IPA. Is there a reason this hasn't been implemented yet? --Michael Ehart (talk) 20:41, 30 March 2020 (UTC)
Word-final [a]
[edit]I was comparing the Russian and English versions of the Wiki Russian Phonology page. According to (only) the Russian version there are three allophones of unstressed [a] in the word-final position:
- /æ/ if the preceding consonant is soft
- /ʌ/ otherwise if the second-to-last syllable is stressed
- /ə/ otherwise (is unconditionally what ru-pron currently chooses)
(Some examples where stress is on the first syllable and recorded pronunciations end with the æ sound are below. Limited examples because I picked them from sibilant-stem feminine-form category, those are easy to spot)
- чукча
- дача
- мульча
- чаща (no audio but on [[2]]/youglish)
- чача (no audio but [[3]]/youglish. You can hear that both "a" sounds sound the same. Since the first one is æ the second one should be æ as well)
@Benwing2 What are your opinions about transcribing the last a of the examples above with æ?
(магараджа on Forvo doesn't undergo this change; it seems this doesn't happen for ʐ?)
About the the second case: The Russkaya Fonetika page claims what ru-pron transcribes as [ɐ] can also be represented with [ʌ], but the use of [ʌ] as I described above requires [ʌ]. Ru-pron currently never uses [ʌ] and introducing it might cause confusion since the Russian Phonology page doesn't describe ʌ/ɐ distinction. (Apart from these problems, I believe this is quite important since all those two-syllable CVCV proper noun diminutives ending with A, e.g. Вова would be mispronounced by someone reading the IPA. A(nother) word that I can hear well whose ending doesn't sound like Schwa in the wiktionary audio is намного.)
What do you think? I believe even [ɐ] would be better than schwa in the second case, if consistency is more important than strictly following the wiki. (Otherwise the best/only option is [ʌ]) Serios3723 (talk) 14:35, 27 May 2020 (UTC)
- @Atitarev Can you comment on this? /ʌ/ (which is properly an unrounded back sound) is sometimes written when [ɐ] is intended, and [ɐ] is somewhat similar to [ə], but [æ] is quite different. BTW @Serios3723, ж is not a soft sound in Russian, so to the extent there is an [æ] after soft sounds, it would not be expected to occur after ж. Benwing2 (talk) 23:39, 31 May 2020 (UTC)
- @Benwing2, Serios3723: I am a bit a loss about what is being asked. Which representations seems to be incorrect? The difference between [ɐ] and [ʌ] is very light but [ɐ] is considered the Moscow pronunciation, the recordings may not reflect that. [æ] doesn't occur in post-tonal positions. The Forvo examples of ча́ща (čášča) are exaggerated (Irina_abeja) and not all are native speakers. You can try Google translate for чаща, which is probably loser to how it's actually pronounced. [ˈt͡ɕæɕːə] is accurate. --Anatoli T. (обсудить/вклад) 23:59, 31 May 2020 (UTC)
- ку́ча (kúča) has a good audio. Does it sound like [ˈkut͡ɕə] to you? It sounds accurate. --Anatoli T. (обсудить/вклад) 00:02, 1 June 2020 (UTC)
- It sounded like [ˈkut͡ɕæ]. Maybe it's a slight diphthongization from æ to ə (Or most probably I just hear incorrectly)? Anyways, I now agree with you that the best transcription is /ə/. @Benwing2 @Atitarev thanks, it seems ru-pron is accurate and the Русская фонетика is misleading. Serios3723 (talk) 07:45, 1 June 2020 (UTC)
- ку́ча (kúča) has a good audio. Does it sound like [ˈkut͡ɕə] to you? It sounds accurate. --Anatoli T. (обсудить/вклад) 00:02, 1 June 2020 (UTC)
- @Benwing2, Serios3723: I am a bit a loss about what is being asked. Which representations seems to be incorrect? The difference between [ɐ] and [ʌ] is very light but [ɐ] is considered the Moscow pronunciation, the recordings may not reflect that. [æ] doesn't occur in post-tonal positions. The Forvo examples of ча́ща (čášča) are exaggerated (Irina_abeja) and not all are native speakers. You can try Google translate for чаща, which is probably loser to how it's actually pronounced. [ˈt͡ɕæɕːə] is accurate. --Anatoli T. (обсудить/вклад) 23:59, 31 May 2020 (UTC)
Pronunciation of unstressed adjectival "-ое" as [əjə]
[edit]Does anyone have any documentation of this? Any articles, papers, etc. explaining how and why it happens? It's not mentioned on the Russian or English Wikipedia pages on Russian phonology, as far as I can tell. BirdValiant (talk) 01:29, 13 July 2020 (UTC)
- @BirdValiant: What is your understanding how it should be pronounced? Google's text-to-speech for бе́лое (béloje) https://translate.google.com.au/#view=home&op=translate&sl=ru&tl=en&text=белое is how I would pronounce it. (if you click on the audio button a second time, it will say slower) Some people don't reduce, so you get closer to [ˈbʲeləje]. Also check out но́вое (nóvoje), which has an audio.
- @Benwing2: FYI. --Anatoli T. (обсудить/вклад) 03:19, 13 July 2020 (UTC)
- @Atitarev: The typical pronunciation of unstressed e is as [ɪ], so pronunciation as ə (which is more usual for а/о type vowels) deserves some kind of explanation. BirdValiant (talk) 04:21, 13 July 2020 (UTC)
- @BirdValiant: Final unstressed "e" is seldom pronounced as [ɪ], unless it's part of a collocation, such as "по́ле бо́я". In isolation, по́ле (póle) is pronounced [ˈpolʲe] (more common these days), [ˈpolʲə] (slightly dated) or [ˈpolʲɪ] (when followed by other words). Traditional references will support [ˈpolʲə]. When designing the module, a lot of sources were used but there's little on this particular topic. You can also look through Module talk:ru-pron/Archive 1. @Benwing2: do you remember any sources? --Anatoli T. (обсудить/вклад) 04:35, 13 July 2020 (UTC)
- @Atitarev, BirdValiant I don't remember very well. Maybe Zaliznyak or Avanesov? User:Cinemantique would know better. Benwing2 (talk) 04:43, 13 July 2020 (UTC)
- @Benwing2: Thanks. I have provided a recording with list of words with the final "-е". I can't find the discussion. Here's my recording:
--Anatoli T. (обсудить/вклад) 04:53, 13 July 2020 (UTC)Selected Russian words ending in unstressed -е: (file) - @BirdValiant, Benwing2: Here's the script of the above audio file: User:Atitarev/recording and the talk page User talk:Atitarev/2020/recording. Back then, while researching the pronunciation, I have provided my own recording with the best attempt to make a natural unbiased pronunciation. --Anatoli T. (обсудить/вклад) 04:58, 13 July 2020 (UTC)
- Yes, Avanesov, §16.5.--Cinemantique (talk) 06:15, 13 July 2020 (UTC)
- Thanks, Cinemantique. --Anatoli T. (обсудить/вклад) 06:22, 13 July 2020 (UTC)
- Thanks @Atitarev: for the audio file and @Cinemantique: for the citation. This information should be added to the phonetics Wikipedia articles in both languages, with the Avanesov citation. Otherwise the mismatch in the effort required to learn these endings and the lack of anyone saying them, will surely lead to disappointment. BirdValiant (talk) 08:52, 13 July 2020 (UTC)
- @Cinemantique: Could you provide the citation to the Avanesov work again? The link has died and I never copied down the actual citation, so I can't find it again. I don't known which one of Avanesov's works it was. BirdValiant (talk) 20:21, 2 June 2022 (UTC)
- @Benwing2: Thanks. I have provided a recording with list of words with the final "-е". I can't find the discussion. Here's my recording:
- @Atitarev, BirdValiant I don't remember very well. Maybe Zaliznyak or Avanesov? User:Cinemantique would know better. Benwing2 (talk) 04:43, 13 July 2020 (UTC)
- @BirdValiant: Final unstressed "e" is seldom pronounced as [ɪ], unless it's part of a collocation, such as "по́ле бо́я". In isolation, по́ле (póle) is pronounced [ˈpolʲe] (more common these days), [ˈpolʲə] (slightly dated) or [ˈpolʲɪ] (when followed by other words). Traditional references will support [ˈpolʲə]. When designing the module, a lot of sources were used but there's little on this particular topic. You can also look through Module talk:ru-pron/Archive 1. @Benwing2: do you remember any sources? --Anatoli T. (обсудить/вклад) 04:35, 13 July 2020 (UTC)
- @Atitarev: The typical pronunciation of unstressed e is as [ɪ], so pronunciation as ə (which is more usual for а/о type vowels) deserves some kind of explanation. BirdValiant (talk) 04:21, 13 July 2020 (UTC)
Hard "л" as [ɫ]
[edit]@Benwing2: Hi. Russian and Ukrainian are the only ones not using [ɫ] for an unpalatalised hard "л". Bulgarian, Belarusian, Macedonian have basically the same sound and use the symbol. Do you think it's worth unifying and making use of [ɫ] for Russian and Ukrainian as well? BTW, ruwiki uses it. --Anatoli T. (обсудить/вклад) 02:25, 22 July 2020 (UTC)
- @Atitarev Yup, I do think it's a good idea. It's easy to implement so I'll go ahead and do it. Benwing2 (talk) 02:40, 22 July 2020 (UTC)
- @Benwing2: Thank you! --Anatoli T. (обсудить/вклад) 02:43, 22 July 2020 (UTC)
- @Benwing2: I guess the test cases will need some cleanup. I can put it on my to-do list. --Anatoli T. (обсудить/вклад) 02:44, 22 July 2020 (UTC)
- @Atitarev Yup, they will. Benwing2 (talk) 02:44, 22 July 2020 (UTC)
- @Benwing2: I'll fix them AFTER you implement this, so that I can see better, which ones to fix. I can see that many cases like "роди́лся", "вы́гладьте", "не пла́чьте", "лу́чший", "ло́же" and "гёрлфренд", etc. already use the symbol, nothing to change. Pls let me know if you disagree or need more info on any other failed cases. --Anatoli T. (обсудить/вклад) 03:13, 22 July 2020 (UTC)
- @Atitarev It is implemented now. If you want, I can fix them myself using a script; it will take a while to do it manually since there are over 50 cases that need to be fixed. Benwing2 (talk) 03:23, 22 July 2020 (UTC)
- @Atitarev I went ahead and fixed the testcases using a regexp in vim. Benwing2 (talk) 03:30, 22 July 2020 (UTC)
- @Atitarev It is implemented now. If you want, I can fix them myself using a script; it will take a while to do it manually since there are over 50 cases that need to be fixed. Benwing2 (talk) 03:23, 22 July 2020 (UTC)
- @Benwing2: I'll fix them AFTER you implement this, so that I can see better, which ones to fix. I can see that many cases like "роди́лся", "вы́гладьте", "не пла́чьте", "лу́чший", "ло́же" and "гёрлфренд", etc. already use the symbol, nothing to change. Pls let me know if you disagree or need more info on any other failed cases. --Anatoli T. (обсудить/вклад) 03:13, 22 July 2020 (UTC)
- @Atitarev Yup, they will. Benwing2 (talk) 02:44, 22 July 2020 (UTC)
- @Benwing2: I guess the test cases will need some cleanup. I can put it on my to-do list. --Anatoli T. (обсудить/вклад) 02:44, 22 July 2020 (UTC)
- @Benwing2: Thank you! --Anatoli T. (обсудить/вклад) 02:43, 22 July 2020 (UTC)
allophones of e
[edit]@Atitarev If as a native speaker you don't see any difference between the e in цель and цех then you might want to investigate the wikipedia page "Russian phonology" Dngweh2s (talk) 04:25, 3 August 2020 (UTC)
- @Dngweh2s: I don't want to spend time convincing someone about my point, I only work at the English Wiktionary and I have plenty of work to do here. Someone may have a referenced quote for their claim at Wiktipedia. As I said in Wiktionary:Beer_parlour/2020/July#Russian_vowel_allophone, we're may be a bit more phonemic in this aspect, anyway. And if there IS a difference, then it's hardly noticeable and it's not worth arguing about. Also, in my opinion, the audio recording at цель (celʹ) has a too narrow [e], that's not how I pronounce it and how you would normally hear on TV, IMO. --Anatoli T. (обсудить/вклад) 04:39, 3 August 2020 (UTC)
- @Atitarev @Dngweh2s: What about initial э before a soft consonant? The best example of this would be эти vs это. This raising is noted in Russian Grammar (1980) in Table 3. Link As a non-native, I'm not sure how well this reflects current usage, but the э in эти seems to be more closed than that in это. 1998alexkane (talk) 18:47, 6 August 2020 (UTC)
- (Notifying Benwing2, Cinemantique, Useigor, Guldrelokk, Fay Freak, Tetromino, PUC): It may be worth reviewing this after all. @1998alexkane, Dngweh2s. --Anatoli T. (обсудить/вклад) 04:38, 7 August 2020 (UTC)
- FWIW, I do pronounce цель and цех differently (the vowel in цех is lower); maybe it's Saint Petersburg pronunciation (where I grew up) vs. the Moscow standard? Tetromino (talk) 17:30, 7 August 2020 (UTC)
- @Tetromino, Benwing2, Cinemantique: Thanks, @Tetromino (you should have pinged). I think it's worth addressing. I am sure @Benwing2 will be able to implement this, maybe we just need to make new test cases. For me цель and цех sound the same or the difference is too small to notice but эти and это are little more noticeable. Also pinged @Cinemantique, perhaps you can do it in ruwiki as well, if no objections. What do Avanesov and other authors say about this? --Anatoli T. (обсудить/вклад)
- @Tetromino, Benwing2, Cinemantique, Atitarev: Here's a couple excerpts from from Avanesov 1984 [4][5] and Jones and Ward 1967 [6]. It looks like both agree that the "э" in "эти" is /e/, but disagree with the "е" in "цель", with Avanesov saying it is a slightly advanced /ɛ/ and Jones and Ward saying it is a retracted /e/. 1998alexkane (talk) 21:29, 13 October 2020 (UTC)
- @Tetromino, Benwing2, Cinemantique: Thanks, @Tetromino (you should have pinged). I think it's worth addressing. I am sure @Benwing2 will be able to implement this, maybe we just need to make new test cases. For me цель and цех sound the same or the difference is too small to notice but эти and это are little more noticeable. Also pinged @Cinemantique, perhaps you can do it in ruwiki as well, if no objections. What do Avanesov and other authors say about this? --Anatoli T. (обсудить/вклад)
- FWIW, I do pronounce цель and цех differently (the vowel in цех is lower); maybe it's Saint Petersburg pronunciation (where I grew up) vs. the Moscow standard? Tetromino (talk) 17:30, 7 August 2020 (UTC)
- (Notifying Benwing2, Cinemantique, Useigor, Guldrelokk, Fay Freak, Tetromino, PUC): It may be worth reviewing this after all. @1998alexkane, Dngweh2s. --Anatoli T. (обсудить/вклад) 04:38, 7 August 2020 (UTC)
- @Dngweh2s, Tetromino, Atitarev, 1998alexkane: Although this is an old discussion, I wanted to pipe in here because I'd started a similar discussion on Russian е. The way I've always seen it is that these stressed vowels between differing levels of palatisations are almost diphthongs, where, for example, the е in свет drops down from the palatalised position to reach an [ɛ], and the е in, for example, же́нщина, rises from the unpalatalised position to reach the [nʲɕː] in [ˈʐɛnʲɕːɪnə]. The latter is already accounted for in the IPA, but the former is not, which is what was arguing. · • SUM1 • · (talk) 01:21, 20 May 2022 (UTC)
Final obstruent devoicing
[edit]I stumbled upon this paper [7] that argues that final obstruent devoicing in Russian is not really complete (and my intuition as a native speaker from central Russia is not against it); could this have implications for the IPA in this module? Draco argenteus (talk) 20:29, 25 September 2021 (UTC)
волшебство
[edit]@Benwing, Atitarev, Tetromino The IPA transcription for волшебство places the accent in the middle of the suffix -ство, which violates both the morphological/etymological structure of the word and the universal phonological tendency to maximize the onset. Is this a mistake in the module (like the one that affected узнать, discussed earlier), or do Russians really break up the suffix into /st/ and /vo/? Martin123xyz (talk) 07:25, 15 October 2021 (UTC)
- The IPA transcription generated by the module is indeed wrong; [stvo] in волшебство is pronounced as one syllable. Tetromino (talk) 16:45, 16 October 2021 (UTC)
Support The Nicodene (talk) 07:13, 25 October 2021 (UTC)
Stressed е not followed by consonant + front vowel or by palatalised consonant should be ɛ?
[edit]I'm not sure if this has been addressed yet, as it would take a long time to thoroughly check, but I didn't see anything after a brief check.
I noticed this when I realised the IPA I was reading out to myself on Wiktionary was different from the audio files provided (and speech I've heard elsewhere). Someone had also written out this more accurate pronunciation on the Wikipedia article for the National anthem of Russia. Russian phonology also seems to make a vague mention of this rule.
In short, in words like свет, цвет, сове́т, челове́к, компле́кт, сде́лать, бе́гать, полде́ла, се́но, etc. I think the [e]'s should be [ɛ]'s. Does not apply to unstressed instances like село́ or слепо́й, front vowels after the consonant like ве́тер or вре́мя or palatalised consonants like семь or смотре́ть. But it's fairly unequivocal that the е in свет and семь is not pronounced the same when you listen to the audios. Just compare the е in the audio of все with the identically-"e"-IPA-annotated caissier, fussiez, fassiez and fissiez.
I did encounter some instances where all these conditions were met but it was debatable whether the е was being pronounced as an [e] or [ɛ], like кре́пкий or кре́пость (probably due to the palatalised consonants later on), but they were the minority, whereas right now it seems the majority cases are not represented in the IPA. · • SUM1 • · (talk) 00:49, 13 December 2021 (UTC)
- @SUM1: This has been already described at Module_talk:ru-pron#allophones_of_e above. @Benwing2 might be able to do it. The difference between [e] or [ɛ] as in э́ти (éti) ([e]) vs э́то (éto) ([ɛ]) is very light and subtle and is often ignored. We are using broader transcriptions at the moment. --Anatoli T. (обсудить/вклад) 21:54, 13 December 2021 (UTC)
- @SUM1, Atitarev IMO, if we are to start representing this level of detail, we should consider splitting the pronunciation into two, as e.g. with Spanish; cf. tinieblas. However, if we were to do that, I'm not sure how we'd represent vowel reduction in the phonemic (broad) transcription. Benwing2 (talk) 06:40, 14 December 2021 (UTC)
- @Benwing2: Thanks. I am not insisting on the change. The difference is indeed very light and may even depend on speakers and realisations. I am highlighting because of the interest above. As for the representation, we are already quite phonetic in the Russian API in vowel reductions, consonant assimilations, voicing and devoicing, silent letters, it's just another detail to that. Our transcription is phonetic, isn't it? The phonemic would be showing sounds as they are originally. --Anatoli T. (обсудить/вклад) 23:26, 14 December 2021 (UTC)
- @Atitarev Judging by that conversation, it didn't address what I was talking about. I was talking about stressed Russian е preceded by palatalisation and followed by non-palatalisation. Е essentially starts palatalised, so any palatalisations afterwards (or it being unstressed) keep it so, while non-palatalisations, as I'm arguing, when it's stressed, removed that palatalisation by the end of it, which is not reflected yet in the IPA. With е, the preceding consonant is palatalised or a [j], as in [svʲɛt], so the vowel glides from a raised one to a lower one. It doesn't stay at [e]. On the other hand, the discussion above concerned е when followed by palatalisations. I think, in цель for example, it's fine as it is now, as an [ɛ], because it starts in that position but raises by the end of it to meet the palatalised consonant. That's what I'm arguing should happen with the [e]'s I'm talking about. They already show as palatalised from the start (or with a [j] at the start), and they should end in the lower [ɛ] position to meet the unpalatalised consonant when stressed. · • SUM1 • · (talk) 00:34, 20 May 2022 (UTC)
- @Benwing2: Thanks. I am not insisting on the change. The difference is indeed very light and may even depend on speakers and realisations. I am highlighting because of the interest above. As for the representation, we are already quite phonetic in the Russian API in vowel reductions, consonant assimilations, voicing and devoicing, silent letters, it's just another detail to that. Our transcription is phonetic, isn't it? The phonemic would be showing sounds as they are originally. --Anatoli T. (обсудить/вклад) 23:26, 14 December 2021 (UTC)
Stressed а after non-palatalisation and before л should be [ɑ]
[edit]I'd like to see number 3 of Module talk:ru-pron/Archive 2#fronting of я, ю only between soft consonants implemented.
Па́лка is pronounced [ˈpɑɫkə], not [ˈpaɫkə]. · • SUM1 • · (talk) 01:50, 20 May 2022 (UTC)
Unstressed [ɨ] and [ʉ] should be lowered
[edit]Unstressed [ɨ] (as it currently stands, unstressed ы and э) should be lowered in the form of [ɨ̞], because it is not identical to stressed [ɨ]/ы. Such is reflected on w:Russian phonology (in more detail, but this change will get it significantly closer to that in one step), and such level of detail with diacritics in Wiktionary IPA templates is already employed in Template:fi-pronunciation or Template:es-IPA, for example. Unstressed [ʉ] (as it currently stands, ю before a palatalised consonant or у between palatalised consonants) should also be lowered in the form of [ʉ̞]. Such is also reflected on w:Russian phonology.
Lines 303 and 304 of the module should have the respective vowel characters (the 2nd and 3rd) replaced with [ɨ̞], and line 343 should also have the respective vowel character replaced with [ʉ̞]. · • SUM1 • · (talk) 12:28, 28 October 2022 (UTC)
Recent edit broke the module, but how?
[edit]@Theknightwho, Benwing2, Erutuon: the module has started displaying Cyrillic "е" instead of Latin "e" in IPA transcriptions, when |1=
of the template has "ѐ" with a grave accent to mark secondary stress (for example, [8]). It looks like "ѐ" in |1=
is also (1) putting the secondary stress mark right before the vowel instead of before the onset consonant at автовеломотогонкам, (2) failing to mark palatalization after a consonant at вероисповедания, and (3) failing to add /j/ after a vowel at военно-технический. The only recent edit to the module was this one, but I don't see how that could have caused the error. Any ideas? —Mahāgaja · talk 08:04, 13 July 2023 (UTC)
"ѝ" is causing problems too, e.g. [9]. —Mahāgaja · talk 08:22, 13 July 2023 (UTC)
- @Mahagaja It was the latest change to Module:ru-translit by User:Theknightwho. I've reverted it for now. Benwing2 (talk) 08:46, 13 July 2023 (UTC)
- Also, that change added a whole lot of not-so-well-documented logic; I'm concerned that adding all that stuff to a translit module like this that may be invoked 1,000 times or more will lead to additional memory errors. Do we really need it just for an extremely rare edge case? Benwing2 (talk) 08:53, 13 July 2023 (UTC)
- @Atitarev Benwing2 (talk) 08:54, 13 July 2023 (UTC)
- @Benwing2 It's all behind an if statement, so doesn't add anything intensive the vast majority of the time. Plus using the
string
library is an order of magnitude faster than theustring
one, and neither cause much memory impact as they both involve strings or objects not declared as variables. I'll investigate and fix it. Theknightwho (talk) 13:48, 13 July 2023 (UTC)- @Theknightwho In my experience, adding code to a heavily require()d module increases memory usage even if it's not used. I had that experience with adding code to Module:table pushing various pages into OOM situations. Benwing2 (talk) 18:03, 13 July 2023 (UTC)
- @Benwing2 @Mahagaja I've fixed the issue - it was because I had forgotten about Cyrillic ѐ (and ѝ) being composed characters. The new part I added decomposes the text to NFD and then recomposed it to NFC afterwards, but happens after ѐ and ѝ were already decomposed earlier on, which meant that they were left in a state that the main substitution function couldn't deal with; it now just uses the decompose function in Module:ru-common instead. I've also added much fuller explanation of how the new code works, as it's not difficult to understand but isn't very intuitive at first glance.
- Aside from the really rare stuff, it also fixes a bug where capital Ё wasn't handled properly in the special cases, as the logic was only written for ё.
- I did some quick speed tests, and the speed/memory differences are negligible. Impossible to tell which is better either way, and I've been looking at this kind of thing a ton lately for the parser. Theknightwho (talk) 18:08, 13 July 2023 (UTC)
- @Theknightwho Removing your code gains almost a MB on el (which is enough to not push it over the limit), and removing the single line in el inside of
{{multitrans}}
that includes two references to Russian terms gains almost 2 MB total. Removing the Russian lines in c gains you close to 1 MB. So it's not as negligible as you're making it out to be. In general I am skeptical of your "it's negligible" claims because we're seeing consistently increasing memory usage (which is why we have 35 items in CAT:E when we formerly had none). I haven't observed new items going into CAT:E from the changes I'm making (which are mostly not even to core modules), so I assume it has to be your changes. I should also add, rather than wait for this discussion to play out, you went ahead and restored your code and then presented it as a fait accompli; this sort of attitude is frustrating for me to deal with (and in the longer run may bite you in the ass as you continue to lose the faith of others). Benwing2 (talk) 10:13, 14 July 2023 (UTC)- @Benwing2 Sure, but (picking at random) I see a 200KB reduction on lo and a 100KB increase on teacher. Even with c, the page is running out of memory further down the page now than it was before the changes, which is a fairer comparison than just removing the Russian outright. One of the most difficult problems with Lua memory is that it fluctuates wildly and unpredictably due to the garbage collect bug, and I've seen it increase after removing a bunch of code entirely. I will check (because I'd like to know for certain), but I think Lua handles functions in a similar way it does strings (meaning they're shared between environments in memory), which means they're not being duplicated tons of times for each invoke.
- I haven't made any significant changes to the core modules recently because I've been working on other things, and there were 36 entries in CAT:E just before I made the change (which I checked so I could monitor). I don't really see how this could be down to me. Theknightwho (talk) 10:46, 14 July 2023 (UTC)
- As a test, I previewed a few pages with 2,500 lines of random code added to the if-block on pages where it doesn't get used, and the result was that c ran out of memory at the same place, teacher increased by 100KB, lo increased by 250KB and el dropped to about 100KB under the memory limit. It's essentially random. Theknightwho (talk) 11:18, 14 July 2023 (UTC)
- @Theknightwho Removing your code gains almost a MB on el (which is enough to not push it over the limit), and removing the single line in el inside of
- @Theknightwho In my experience, adding code to a heavily require()d module increases memory usage even if it's not used. I had that experience with adding code to Module:table pushing various pages into OOM situations. Benwing2 (talk) 18:03, 13 July 2023 (UTC)
- @Benwing2 It's all behind an if statement, so doesn't add anything intensive the vast majority of the time. Plus using the
- @Atitarev Benwing2 (talk) 08:54, 13 July 2023 (UTC)
- Also, that change added a whole lot of not-so-well-documented logic; I'm concerned that adding all that stuff to a translit module like this that may be invoked 1,000 times or more will lead to additional memory errors. Do we really need it just for an extremely rare edge case? Benwing2 (talk) 08:53, 13 July 2023 (UTC)
There's a slight issue with the pronunciation of маои́зм (maoízm) and its derivatives:
At the moment, this necessitates the input {{ru-IPA|мɐо̂и́зм}}
, which isn't ideal. The same issue crops up elsewhere, too, such as with Аомы́нь (Aomýnʹ), авиаотря́д (aviaotrjád), аорта́льно-по́чечный (aortálʹno-póčečnyj) etc. Tagging @Benwing2 @Atitarev. Theknightwho (talk) 00:14, 24 December 2023 (UTC)
- @Atitarev Is this general enough to be implementable as a rule? I have doubts about авиаотря́д (aviaotrjád) in particular since it's авиа- + отря́д. Benwing2 (talk) 00:36, 24 December 2023 (UTC)
- @Theknightwho, @Benwing2: Yes, маои́зм (maoízm), таои́зм (taoízm), Аомы́нь (Aomýnʹ), Аоя́ма (Aojáma) (and their derivatives) are mostly pronounced without a reduction but it's hard to set a rule around it. It's to do with some foreign words with a group of vowel in a row, where vowel reduction would sound a little weird.
- сте́рео (stéreo) has no reduction but стереофони́ческий (stereofoníčeskij) does.
- There are no reductions on авиаотря́д (aviaotrjád) or аорта́льно-по́чечный (aortálʹno-póčečnyj). Anatoli T. (обсудить/вклад) 01:56, 26 December 2023 (UTC)
- Also, ра́дио (rádio) (non-reduced), doesn't spread onto радиоста́нция (radiostáncija) (reduced) but can be both ways on радиоаппара́т (radioapparát). Anatoli T. (обсудить/вклад) 03:29, 26 December 2023 (UTC)
- @Atitarev @Benwing2 I think the rule is that there’s no reduction if it would result in *[ɐɐ] when there should be a diphthong - this explains маои́зм [mɐoˈizm], таои́зм [tɐoˈizm], Аомы́нь [ɐoˈmɨnʲ], Аоя́ма [ɐoˈjamə], авиаотря́д [ˌavʲɪɐoˈtrʲat], аорта́льно-по́чечный [ɐorˌtalʲnə ˈpot͡ɕɪt͡ɕnɨj]. With regards to радиоаппара́т, I wonder if the reason it can go two ways is because the first pronunciation should be [ˌradʲɪɐəpɐˈrat], but since that’s not very clear, it sometimes becomes [ˌradʲɪoəpɐˈrat]? It could also be to do with it being spelled ⟨оа⟩, not ⟨ао⟩ like the others, but I’m not sure.
- The pairs сте́рео [ˈsʲtʲerʲɪo], стереофони́ческий [sʲtʲɪrʲɪəfɐˈnʲit͡ɕɪskʲɪj] and ра́дио [ˈradʲɪo], радиоста́нция [rədʲɪɐˈstant͡sɨjə] seem easier to explain: reduction only happens if it’s not at the end of the word. Theknightwho (talk) 18:21, 28 December 2023 (UTC)
- @Atitarev What happens with северноалта́йский (severnoaltájskij) and южноамерика́нский (južnoamerikánskij)? Theknightwho (talk) 19:15, 28 December 2023 (UTC)
- @Theknightwho: These are regular, pronounced with a reduction. Anatoli T. (обсудить/вклад) 20:56, 28 December 2023 (UTC)
- @Atitarev Could you give a transcription? I expected северноалта́йский [sʲɪvʲɪrnoɐɫˈtajskʲɪj] and южноамерика́нский [jʊʐnɐəmʲɪrʲɪˈkanskʲɪj] or [jʊʐnoəmʲɪrʲɪˈkanskʲɪj]. Theknightwho (talk) 21:00, 28 December 2023 (UTC)
- @Theknightwho: [ˌsʲevʲɪrnɐɐɫˈtajskʲɪj], [ˌjuʐnɐɐmʲɪrʲɪˈkanskʲɪj] Anatoli T. (обсудить/вклад) 00:13, 29 December 2023 (UTC)
- @Theknightwho I think we'll need to manually specify cases like маои́зм (maoízm), although I can fix the vowel quality of the а before о̂. Benwing2 (talk) 03:01, 29 December 2023 (UTC)
- @Benwing2 @Atitarev Sure. That being said, I’m confused what [ɐɐ] is supposed to imply, since it’s suppposed to be a phonetic transcription. Is it [ɐː]? It feels really odd to count them as separate syllables if they’re the same vowel. e.g. the WP article Aritarev refers to mentions сообража́ть [sɐɐbrɐˈʐatʲ], but it definitely sounds like [sɐːbrɐˈʐatʲ]: . Theknightwho (talk) 19:36, 1 January 2024 (UTC)
- @Theknightwho: Long vowels are basically missing in Russian. It's [sɐɐbrɐˈʐatʲ], not [sɐːbrɐˈʐatʲ]. Anatoli T. (обсудить/вклад) 21:37, 1 January 2024 (UTC)
- @Atitarev Phonemically, sure, but I don’t see how that could exist in a phonetic transcription. Theknightwho (talk) 21:39, 1 January 2024 (UTC)
- @Theknightwho: It's phonetic. Please try listening to
. Can you hear two [ɐɐ] in a row? Anatoli T. (обсудить/вклад) 21:45, 1 January 2024 (UTC)кооперация: (file) - Also, try listening to those words in Google Translate audio (twice!). The 2nd time is always slower and clearer. GT imitates the native Russian pronunciation well (mostly). Anatoli T. (обсудить/вклад) 21:48, 1 January 2024 (UTC)
- @Atitarev I can’t hear any phonetic break between the vowels, even though I know they’re separate phonemically. It’s phonetically indistinguishable from a long vowel, isn’t it? Theknightwho (talk) 21:51, 1 January 2024 (UTC)
- @Theknightwho: A glottal stop is normally not inserted between separate vowels but if you insert a light glottal stop, you won't sound too unnatural. Two identical sounds in a row sound different from long vowels to me. It's called hiatus.
- I am trying to think of examples in other languages. Maybe Italian alcool in a careful speech?. It's two [o]'s, not a long vowel.
- @Benwing2 is already familiar with this feature, maybe he can explain better. Anatoli T. (обсудить/вклад) 22:00, 1 January 2024 (UTC)
- @Theknightwho A sequence of two identical short vowels is not the same as a single long vowel, even phonetically. Ladefoged has a section about this in his authoritative book on phonetics, concerning a language that has short vowels, long vowels and various sequences of identical short vowels and short + long and long + short vowels. Generally, two short vowels in a row are longer than a single long vowel and there are two phonetic impetuses associated with two short vowels. Benwing2 (talk) 22:00, 1 January 2024 (UTC)
- I hear two vowels rather than just a single long one, but I do wonder how much our knowledge of morphology influences our perception. If only we could wipe the slate clean and listen to things with a fresh ear. What do spectrograms say? PUC – 22:16, 1 January 2024 (UTC)
- @PUC: Thanks. The French coopération with two [ɔ]'s in a row, is another good example. Anatoli T. (обсудить/вклад) 22:27, 1 January 2024 (UTC)
- @Benwing2 @Atitarev Thanks - that’s useful to know. I feel like displaying [ɐ.ɐ] would be beneficial, since it’s non-obvious. Theknightwho (talk) 00:34, 2 January 2024 (UTC)
- @Theknightwho: Perhaps, I don't know, since we don't normally insert dots between syllables. If we do, it will be a significant change. @Benwing2: what do you think? Anatoli T. (обсудить/вклад) 00:40, 2 January 2024 (UTC)
- @Atitarev I just mean where clarity would be helpful, as opposed to everywhere. Theknightwho (talk) 00:41, 2 January 2024 (UTC)
- @Theknightwho: I don't have any objections. Anatoli T. (обсудить/вклад) 00:43, 2 January 2024 (UTC)
- @Atitarev @Theknightwho I think either we should put the syllable dividers everywhere or nowhere. Benwing2 (talk) 00:51, 2 January 2024 (UTC)
- In this case I think it's not necessary because in IPA, it's not normal to write two vowels in a row to indicate a long vowel, so this can only be interpreted as a sequence of two vowels. Benwing2 (talk) 00:52, 2 January 2024 (UTC)
- @Benwing2, @Theknightwho: Let's leave as is then. Anatoli T. (обсудить/вклад) 00:53, 2 January 2024 (UTC)
- @Atitarev @Theknightwho I think either we should put the syllable dividers everywhere or nowhere. Benwing2 (talk) 00:51, 2 January 2024 (UTC)
- @Theknightwho: I don't have any objections. Anatoli T. (обсудить/вклад) 00:43, 2 January 2024 (UTC)
- @Atitarev I just mean where clarity would be helpful, as opposed to everywhere. Theknightwho (talk) 00:41, 2 January 2024 (UTC)
- @Theknightwho: Perhaps, I don't know, since we don't normally insert dots between syllables. If we do, it will be a significant change. @Benwing2: what do you think? Anatoli T. (обсудить/вклад) 00:40, 2 January 2024 (UTC)
- @Benwing2 @Atitarev Thanks - that’s useful to know. I feel like displaying [ɐ.ɐ] would be beneficial, since it’s non-obvious. Theknightwho (talk) 00:34, 2 January 2024 (UTC)
- @PUC: Thanks. The French coopération with two [ɔ]'s in a row, is another good example. Anatoli T. (обсудить/вклад) 22:27, 1 January 2024 (UTC)
- @Atitarev I can’t hear any phonetic break between the vowels, even though I know they’re separate phonemically. It’s phonetically indistinguishable from a long vowel, isn’t it? Theknightwho (talk) 21:51, 1 January 2024 (UTC)
- Also, try listening to those words in Google Translate audio (twice!). The 2nd time is always slower and clearer. GT imitates the native Russian pronunciation well (mostly). Anatoli T. (обсудить/вклад) 21:48, 1 January 2024 (UTC)
- @Theknightwho: It's phonetic. Please try listening to
- @Atitarev Phonemically, sure, but I don’t see how that could exist in a phonetic transcription. Theknightwho (talk) 21:39, 1 January 2024 (UTC)
- @Theknightwho: Long vowels are basically missing in Russian. It's [sɐɐbrɐˈʐatʲ], not [sɐːbrɐˈʐatʲ]. Anatoli T. (обсудить/вклад) 21:37, 1 January 2024 (UTC)
- @Benwing2 @Atitarev Sure. That being said, I’m confused what [ɐɐ] is supposed to imply, since it’s suppposed to be a phonetic transcription. Is it [ɐː]? It feels really odd to count them as separate syllables if they’re the same vowel. e.g. the WP article Aritarev refers to mentions сообража́ть [sɐɐbrɐˈʐatʲ], but it definitely sounds like [sɐːbrɐˈʐatʲ]: . Theknightwho (talk) 19:36, 1 January 2024 (UTC)
- @Theknightwho I think we'll need to manually specify cases like маои́зм (maoízm), although I can fix the vowel quality of the а before о̂. Benwing2 (talk) 03:01, 29 December 2023 (UTC)
- @Theknightwho: [ˌsʲevʲɪrnɐɐɫˈtajskʲɪj], [ˌjuʐnɐɐmʲɪrʲɪˈkanskʲɪj] Anatoli T. (обсудить/вклад) 00:13, 29 December 2023 (UTC)
- @Atitarev Could you give a transcription? I expected северноалта́йский [sʲɪvʲɪrnoɐɫˈtajskʲɪj] and южноамерика́нский [jʊʐnɐəmʲɪrʲɪˈkanskʲɪj] or [jʊʐnoəmʲɪrʲɪˈkanskʲɪj]. Theknightwho (talk) 21:00, 28 December 2023 (UTC)
- @Theknightwho: These are regular, pronounced with a reduction. Anatoli T. (обсудить/вклад) 20:56, 28 December 2023 (UTC)
- @Theknightwho: авиаотря́д (aviaotrjád) is regular - [ˌavʲɪɐɐˈtrʲat]. Anatoli T. (обсудить/вклад) 20:57, 28 December 2023 (UTC)
- @Atitarev So no diphthong? Hmm. Theknightwho (talk) 21:00, 28 December 2023 (UTC)
- @Theknightwho: w:Russian_phonology#Vowel_mergers has a phrase about words like коопера́ция (kooperácija) [kɐɐpʲɪˈrat͡sɨjə]. Anatoli T. (обсудить/вклад) 21:04, 28 December 2023 (UTC)
- In a whole sentence it is uttered fast enough that there can be no [ɐː]. The speaker for our сообража́ть at least tries to make two vowels (“two phonetic impetuses”), I hear it in the record, but from experience the masses just say a single short [ɐ].
- The amount of vowel reduction given by Wiktionary always seemed rejectable, the difference between [ˌavʲɪɐɐˈtrʲat] авиаотря́д. [ˌradʲɪɐɐpɐˈrat] and [ˌradʲɪoəpɐˈrat] радиоаппара́т is mysterious to me. Trailing interfix ⟨о⟩ such as in радиоаппара́т may or may not be [ɐ] or stay [o] according to the speaker’s desire for clarity, i.e. adherence to the literary language and the word’s character as internationalism (i.e. in my milieu these words shall not deviate too much from German, unlike pure Russian words, where things like си-ди́ (“CD”) do not exist in favour of цеде́шка), but ⟨о⟩ generally never becomes [ə] except when attempting to imitate Muscovite speech or some complicated environments. [əpɐˈrat] is impossible for me also in compounds, so like аппара́т is given [ɐpɐˈrat] so is радиоаппара́т [ˌradʲiɐɐpɐˈɾat] or [ˌradʲioɐpɐˈɾat]. Yes, [i] and not [ɪ], coincidentally like the vowel in German Radio, obviously I can very well distinguish [i], [ɪ], and [ɨ]. There must be subtle differences between the pronunciations of original Russian settlements (“primary formation”) and those of Russified Germans and Central Asians without these being marked or discernible as non-native to anyone without rare phonetic training in multiple Russian regiolects. Ok, you don’t make
{{a|Kazakhstan}}
, but the pronunciation sections don’t become wrong if they are sometimes closer to that than to Moscow (which is recently also relatively hard to enter in comparison to all other places with subtly deviating pronunciations that could over time appear more representative), they are good enough if they hit any apparently native pronunciation, there is no optimization beyond “good enough” (especially for Russians …). Fay Freak (talk) 00:08, 2 January 2024 (UTC)- @Fay Freak What you mention about the speaker’s desire for clarity is something I’m familiar with in English, yes - a lot of words with schwa reduction lose it if the speaker is enunciating for clarity: compare ensure and insure, which are nominally homophones, but most native speakers would probably reject that analysis. I feel like something similar may be occurring here. Theknightwho (talk) 00:32, 2 January 2024 (UTC)
- This is a good example. Due to pressure from neighbouring languages, speakers also radicalize: back to the phonemic values, even if it means—paradoxically—more similarity to the advantaged language rather than the targetted standard language, for greater desire of clarity. Like one knows beyond language that outlanders have to prove themselves, put in even more effort than the average to be part of the community. These expatriate communities have only grown due to the current events, and later our “standard” pronunciations will be rarefied like RP against MLE, with teaching materials all around the world presenting pronunciations that aren’t encountered outside of Рублёвка. Fay Freak (talk) 00:44, 2 January 2024 (UTC)
- @Fay Freak: I have added the casual pronunciations for both сообража́ть (soobražátʹ) and сообрази́ть (soobrazítʹ) where double vowels are pronounced as one.
- It's a case where it does happen but this shortening doesn't apply to other cases with double vowels like коопера́ция (kooperácija), сооте́чественник (sootéčestvennik), etc. Let's not make a rule out of it, it's case by case. Anatoli T. (обсудить/вклад) 00:51, 2 January 2024 (UTC)
- @Fay Freak What you mention about the speaker’s desire for clarity is something I’m familiar with in English, yes - a lot of words with schwa reduction lose it if the speaker is enunciating for clarity: compare ensure and insure, which are nominally homophones, but most native speakers would probably reject that analysis. I feel like something similar may be occurring here. Theknightwho (talk) 00:32, 2 January 2024 (UTC)
- @Atitarev So no diphthong? Hmm. Theknightwho (talk) 21:00, 28 December 2023 (UTC)
- @Atitarev What happens with северноалта́йский (severnoaltájskij) and южноамерика́нский (južnoamerikánskij)? Theknightwho (talk) 19:15, 28 December 2023 (UTC)
- Also, ра́дио (rádio) (non-reduced), doesn't spread onto радиоста́нция (radiostáncija) (reduced) but can be both ways on радиоаппара́т (radioapparát). Anatoli T. (обсудить/вклад) 03:29, 26 December 2023 (UTC)
more narrow transcriptions?
[edit](copied from https://en.wikipedia.org/wiki/Russian_phonology)
Hard /t, d, n/ are laminal denti-alveolar [t̪, d̪, n̪]; (they don't have a separate symbol from dental consonants)
/ts, s, sʲ, z, zʲ/ are dental [t̪s̪, s̪, s̪ʲ, z̪, z̪ʲ]
Theres a lot of other stuff that could be added to make the transcriptions more narrow but these are the simplest, so if people are open to more narrow transcriptions then this is a start.
Zbutie3.14 (talk) 02:30, 2 August 2024 (UTC)