Help talk:IPA/Serbo-Croatian

Dž

Do SC distinguish an affricate dž from a stop-fricative sequence dž ? Most Slavic languages make a distinction in the orthography, but that wouldn't be possible in this case. kwami (talk) 14:56, 20 June 2009 (UTC)

Yes it does (e.g. nadživjeti "to outlive", compound of nad "over" + živjeti "live" where subsequent affrication would obscure the meaning), and also between the sequence /nj/ and /ɲ/ which are both spelled as 'nj' (e.g. in injekcija "injection"), and also /lj/ and /ʎ/ (of which some orthographies claim there are no phonetically ambiguous examples of, but these have been disputed recently in some chemistry terminology such as metiljantarni [1]). --Ivan Štambuk (talk) 11:47, 21 June 2009 (UTC)

Additional letters and transcriptions

Bulgarian

• [dʒ] = дж = dzh
• [j] = ӣ = i, y
• [ɲ] = нь = ny
• [ʃ] = ш = sh
• [ts] = ц = ts
• [tʃ] = ч = ch
• [ʒ] = ж = zh
• [ə]/[ɤ] = ъ = a/ă

Macedonian

• [dz] = ѕ = dz
• [ɟ] = ѓ = ǵ/gj
• [c] = ќ = ḱ/kj

That's a basic overview of the differences between them and Serbo-Croatian. BalkanFever 08:46, 17 July 2009 (UTC)

Would Slovenian be too divergent?

So, the Macedonian C's, the Bulgarian schwa, plus orthography.

We could put in extra switches. Or we could have several templates (SC, Bulg, Slov.) which all link to this IPA key. kwami (talk) 09:02, 17 July 2009 (UTC)

(ec) Not really AFAIK. It contains [ə] which is written <e> (so is [ɛ]) but that's about it. Other points:

• in Sl and Mk <lj> generally represents a [lj] sequence rather than [ʎ]
• in Mk and Bg, [v] = в = v , rather than [ʋ]
• in Sl, Mk and Bg, there is no [tɕ] or [dʑ]. BalkanFever 09:13, 17 July 2009 (UTC)

Hmm, I think it would be easier with different templates (e.g. Macedonian pronunciation: [{{{1}}}] etc.) leading to the same page, WP:IPA for South Slavic rather than one complicated template. BalkanFever 09:20, 17 July 2009 (UTC)

Slovenian

• e, o, ə
• dz
• ɣ
• tone differences?

kwami (talk) 11:53, 18 July 2009 (UTC)

Tone marks are, according to de Bray's Guide to South Slavic, grave accent for short falling, acute accent for long rising and some form of inverted breve for long falling. BalkanFever 04:38, 19 July 2009 (UTC)

Okay, pretty similar, but it would require an extra column in the tone table for the grave and the SH tones missing in Slovenian. kwami (talk) 06:40, 19 July 2009 (UTC)

vowels

Vowel examples need some reworking because of the lengths and tones. In some of the SC examples the vowel is short where on the corresponding English word it's long and vice versa. Also is the problematic <ije> sequence which is pronounced as [ie] only when it is a jat reflex (and not in words such as dijeta "diet"), and it has dual pronunciation when following /n/ and /l/ of both palatalizing and nonpalatalizing (e.g. mlijeko as [mljeko], [mʎeko] and even as [mlieko] according to some grammars which in practice is never spoken tho). --Ivan Štambuk (talk) 08:12, 12 November 2009 (UTC)

request

I've templated all unformatted IPA I could find in the March 12 dump, with just 9 articles to go. 3 of those are SC: Mirijevo, Novaci (Ub), Vukašin Brajić. Could s.o. here template them with IPA-sc, and fix up the transcription if needed? I'd do it myself, but I have no idea what the accents would be. — kwami (talk) 10:43, 21 September 2010 (UTC)

Done. No such user (talk) 14:09, 21 September 2010 (UTC)

Thanks! — kwami (talk) 20:45, 21 September 2010 (UTC)

Tone

How to hell do we use the rising and falling tone marks? Are those caron+circumflex, or breve+upside-down breve? Using combining marks is rather inconvenient, but it seems to be the only especially with symbols like ɔ or ɛ. When I use combining breve with ̆ɔ it looks good in the edit window (DejaVu Sans Mono for me):

̆ɔ 

ɔ
...but look offset on the text (Vector skin+Arial+Firefox). No such user (talk) 14:21, 21 September 2010 (UTC)

I think maybe your edit-window font doesn't display it properly. What you coded here is the breve over a space, followed by the bare vowel. (And, BTW, for IPA it should be an angular caron for rising tone, ‹ ̌›. Below the edit window the order is bottom tone ə̏, rising ə̌, falling ə̂. The breve is for a short vowel.) Here's what I get for those vowels, w/ and w/o IPA templating:

• with {{IPA}}: ɛ̂ ɛ̌ ɔ̂ ɔ̌
• without: ɛ̂ ɛ̌ ɔ̂ ɔ̌

— kwami (talk) 20:51, 21 September 2010 (UTC)

Thanks. Your examples look correct both in my edit window and in the page proper. I tend to forget that the combining marks must come after the letter, which is rather counter-intuitive (because when typing, one usually first hits a prefix key, then a letter). Not sure why DejaVu shows the first example above "correctly" for me (although I typed it incorrectly). However, it also shows your examples well. Still, it would be convenient if we had combining diacritics in the symbol palette when editing. No such user (talk) 07:40, 22 September 2010 (UTC)

Mid or open-mid?

The article Serbo-Croatian phonology seems to say that /e/ and /o/ are realized as mid vowels, and uses the symbols [e] and [o] everywhere. This IPA guide, however, uses the symbols [ɛ] and [ɔ], suggesting they're in fact open-mid vowels. Which is right? — AdiJapan 08:07, 30 May 2011 (UTC)

They're mid, or somewhere in-between (see the space at Serbo-Croatian phonology#Vowels). I think that we should really switch to using <e> and <o> symbols everywhere because 1) they're not wrong 2) we already have a mixture across articles, so we could standardize to a simpler version and 3) we often attach rising/falling IPA diacritics, and there are no precomposed <ɛ> and <ɔ> glyphs with diacritics, so it is quite impractical to type. More opinions are welcome, we could initiate an informal RFC on WP:PHONETICS No such user (talk) 11:04, 30 May 2011 (UTC)

What does the literature typically use? I imagine it's e and o, since they're typographically easier, but I haven't checked. — Ƶ§œš¹ _{[aɪm ˈfɹ̠ˤʷɛ̃ɾ̃ˡi]} 14:51, 30 May 2011 (UTC)

There is a mixture. Moren uses <ɛ>, <ɔ> and I've seen <e>, <o>. I concede that <ɛ> and <ɔ> are more accurate representations, but the burden on practicality is just too much for my taste. As far as I know, Spanish and Hebrew have essentially the same vowel quality, but WP:IPA for Spanish and WP:IPA for Hebrew use <e> and <o>. No such user (talk) 06:43, 1 June 2011 (UTC)

If it's just a matter of ease of typing, we can always clean up with a bot or AWB. And we could of course add them pre-combined to the edit window; the only question is how many languages can we do that for before it becomes overwhelming.

Personally, I like to use the more precise letters, but I'm not familiar with SC so I don't really have an opinion. — kwami (talk) 07:28, 1 June 2011 (UTC)

Is there still a desire to add vowel–tone combos to the edit window? Figure we might want, ɛ́ ɛ̀ ɛ̂ ɛ̌ ɔ́ ɔ̀ ɔ̂ ɔ̌ ɪ́ ɪ̀ ɪ̂ ɪ̌ ʊ́ ʊ̀ ʊ̂ ʊ̌ ə́ ə̀ ə̂ ə̌, not just for this lang but for others too.

I'm finding a fair amount of <e, o>, and it is more accessible as well as being in the phon. article, so I'm switching over. — kwami (talk) 05:55, 26 January 2012 (UTC)

Needs tone

Could s.o. add tone to Belgrade? — kwami (talk) 07:06, 5 October 2011 (UTC)

And Kosovo#Name, Serbia, Serbs, please? — kwami (talk) 05:45, 26 January 2012 (UTC)

Done. Still having problems with those pesky diacritics though. No such user (talk) 16:17, 31 January 2012 (UTC)

If you want to continue (there are lots of articles without tone, though most aren't very important), you can tell me and I can type them up. I created a keyboard to enter the diacritics directly, so it's easy for me. — kwami (talk) 22:43, 31 January 2012 (UTC)

Also Bosniaks. Serbian Christmas traditions (multiple, many with just the {{IPA}} template). And Hudi Bitek looks like it might be a mix of systems. — kwami (talk) 08:54, 3 February 2012 (UTC)

Labiodental fricative vs labiodental approximant

In standard croatian language, the letter V is labiodental fricative, and not labiodental approximant. That means, it's articulated with both lower lip and upper teeth! Ivan.milicic3510 (talk) 18:49, 3 February 2012 (UTC)

If you have a source for that, and people prefer it, then it would be a more straightforward letter to use. It contradicts our phonology article, though, and v does not behave as a fricative. It doesn't become [f] before p, t, k, for example; likewise, f does not become [v] before b, d, g. That is, SC v and f do not form a fricative pair the way z and s do, or ž and š. — kwami (talk) 04:28, 4 February 2012 (UTC)

Yes, it is fricative only on surface, as Móren pointed out; actually, [v] and [ʍ] [ʋ] are allophones, and/or in complementary distribution: the former in syllable onset, and the latter in coda position (or along these lines). Personally, I don't have a preference for either. No such user (talk) 22:40, 4 February 2012 (UTC)

Russian /v/ is similar. While it can be devoiced itself, it's transparent to voicing assimilation, making it act a lot more like a sonorant.

As long as we're consistant, I'm fine with either, but I'm not very familiar with the language's phonology. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 01:00, 5 February 2012 (UTC)

Um, I don't think so (concerning Russian): IIRC e.g. ковточка 'sweater' is ['koːftɐt͡ɕkə]; Russian also exhibits final devoicing: здоров 'healthy' [zdə'rof]. No such user (talk) 09:51, 6 February 2012 (UTC)

Yes, those are examples of /v/ devoicing. How does that contradict what I said? — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 00:17, 7 February 2012 (UTC)

You said that it's "transparent to voicing assimilation", which I interpreted as "does not devoice in contact with voiceless consonants", which I then disproved. Are you saying I did not understand you well? No such user (talk) 14:33, 7 February 2012 (UTC)

No such, do you mean [w̥], or [ʋ̥]? I didn't realize it devoiced at all. But if the voiced allophone truly is [v], then I don't know why we'd transcribe it [ʋ]. (And also, if it's [v], why did so many of our articles have [v̞]?) And if it devoices, we should indicate that too, IMO. — kwami (talk) 01:54, 5 February 2012 (UTC)

I meant [ʋ] (aka [ʋ]), sorry for confusion (typed from a device without that glyph, so I mis-copied). I admit I don't follow you 100%, but I'll try to answer: it never devoices to [f]. When it is in "strong" position, it is not as fricative as e.g. Russian one, hence it could be written [v̞] (e.g. visok ['v̞isok], 'high'); when in "weak" position, it is [ʋ] (e.g. čavka ['t͡ʃaːʋka] 'jackdaw', krov [kroʋ] 'roof'). In our articles, pronunciations including [v̞] were inserted mostly by a couple of friendly IPs, though I'd argue it's too much level of detail. Personally (maybe due to influence of my mother tongue), I don't perceive much difference between [v] and [ʋ], so I'm not sure if my allophone observations are quite accurate, but you get the gist.
I guess, if you ask a native speaker, "what's the 28th letter of the alphabet?", you'd get the answer: [və]. However, since it behaves like an approximant, I think it's slightly better to stick to [ʋ], lest we get questions like "how it's possible to pronounce [vk] in ['t͡ʃaːvka]?" No such user (talk) 09:34, 6 February 2012 (UTC)

You mean the devoicing ring was the typo? It does not devoice to [ˈt͡ʃaːʋ̥ka]?

Since I'd normally read [v̞] to be an approximant, I'd take it to be essentially the same as [ʋ]. Are you saying it's a light fricative, halfway between a fricative and an approximant? I agree that's too precise a transcription. We don't bother with it in Spanish, for example, though we do describe it in a footnote, as well as in the phonology article. — kwami (talk) 15:15, 6 February 2012 (UTC)

Yes, basically it's a light fricative.

Now I'm confused -- who introduced that devoicing ring? My only typo in this thread was [ʍ] instead of [ʋ]... Ah, I see what you mean: [w̥] is an alt symbol for ʍ. Yes, that was a typo. No such user (talk) 15:25, 6 February 2012 (UTC)

Oops, my bad.

So, it's a light fricative in initial position, and an approximant in coda position? I'll add a note. — kwami (talk) 15:36, 6 February 2012 (UTC)

That's likely my OR. Bruce Morén (2005), Consonant-Vowel Interactions in Serbian: Features, Representations and Constraint Interactions (PDF), Center for Advanced Study of Theoretical Linguistics, Tromsø says,

Second, [v] is usually described as a phonetic fricative in Serbian. However, its

distribution and phonological behavior are consistent with that of a sonorant, not an obstruent - e.g. it does not participate in voicing alternations with [f]. Therefore, I assume, with Browne (1993), Baric¤ et al. (1997) and Miller-Ockhuizen and Zec (2003)

that it is a phonological sonorant (i.e. /V/) despite its surface fricative property.

No such user (talk) 16:44, 6 February 2012 (UTC)

In that case, since this is a phonetic transcription, I propose replacing it with ⟨v⟩ in initial position. We can leave ⟨ʋ⟩ in coda position. Or is it phonetically [v] in all positions? If so, IMO we should replace it with more accurate (and more accessible) ⟨v⟩ across the board, and leave the adherent behaviour to a footnote. — kwami (talk) 17:00, 6 February 2012 (UTC)

It is the same phoneme throughout, so we should use the same symbol, whichever it is. I slightly favor [ʋ], because it is phonologically accurate, and phonetically good enough, and because it is used by Moren, Browne etc. There is a note in the phonology article and this key, so that shouldn't be a problem. No such user (talk) 08:40, 7 February 2012 (UTC)

Yes, Brown uses ⟨ʋ⟩, but Morén uses ⟨v⟩. Also, this is a phonetic key, and we don't use the same letter throughout for other phonemes. "Phonologically accurate" is also a theoretical claim, whereas phonetically accurate is an empirical claim, and phonetically it sounds like it's basically [v]. — kwami (talk) 11:40, 7 February 2012 (UTC)

Tone descriptions

According to the now removed (though more thorough, I guess) pronunciation guide in Serbo-Croatian phonology, 'è' isn't a "short vowel with rising tone" (which would mean that the tone rises within the short vowel and probably stays more or less level outside), it's a short vowel with low tone that rises on the succeeding syllables (the reverse of what the current description suggests). Analogically for 'ȅ'. Native speakers' attention appreciated. 89.231.125.218 (talk) 00:28, 24 May 2012 (UTC)

è is called "short rising", though it's actually short vowel with high tone that stays (about) the same on the succeeding syllables. Here's a rather succinct overview, [2] which should be entered in the article. Full text here (free), and this article seems spot on, but it's preview only. We really need some sound recordings, though, which is something which I've planned to do for several years. No such user (talk) 07:22, 24 May 2012 (UTC)

Problem

The help page isn't correct. Though you acknowledged "t" and "d" as dental consonants and "nj" and "lj" as alveolo-palatal ones, there are still some problems. Consonants "s", "z" and "c" are also dental (first and second - sibilants; third - affricate) and in no way alveolar. This is supported by all language textbooks issued by Board for Standardization of the Serbian Language. 178.221.193.240 (talk) 17:45, 3 March 2013 (UTC)

Since it's ineffective to cite these books because you can't see them, take a look at this. I cite:

• зубници (дентали): д, т, з, с и ц; on English: with teeth (dental): d, t, z, s, c

178.221.193.240 (talk) 17:52, 3 March 2013 (UTC)

It is not even disputed. It is just a matter of convention which level of detail we want to record in the phonetic key for a given language, and we generally don't want too fine details there: our IPA keys are rather broad and indicate only general phonemes and a general indication of pronunciation for non-expert audience. Too many IPA diacritics would be both impractical to type and a bit frightening for the general reader. For example, our /e/ should really be an /ɛ˰/; /l/ is an /ɫ/; /dʒ/ should be /ɖʐ/, etc. The practical solution is to use simplified, more general glyphs. Aeusoes1 made a practical compromise [3] by linking to more appropriate (dental) articles while retaining the simpler (alveolar) glyphs.

However, you're welcome to add a note to Serbo-Croatian phonology#Consonants about dental nature of the plosives. No such user (talk) 07:57, 4 March 2013 (UTC)

OK, but I just found it misleading that some phonemes link to the other ones just for sake of practicity. However, additions to the phonology article would solve the probs. Thanks, 77.46.254.117 (talk) 13:02, 4 March 2013 (UTC)

(edit conflict) Now I see that I didn't read your first post carefully, and was answering something else, obviously.

Yeah, I think you're right. SC /s/ and /ts/ are dental as well. We do have an article Voiceless dental sibilant, but s̪ does not redirect there (yet), and the article looks stubby. On the other hand, Voiceless dental affricate is a red link; it is questionable if this article should exist, as no language contrasts /ts̪/ from /ts/ (and AFAIK no language contrasts /s/ from /s̪/ as well, maybe Basque). I'll try to update Serbo-Croatian phonology when I find some time, but for more general approach I'd rather leave the comment to our more seasoned linguist editors. No such user (talk) 13:13, 4 March 2013 (UTC)

I didn't realize that voiceless dental sibilant existed. One of the items on my to-do list is to merge the dental and alveolar consonants into one page, though the relevant discussion seems to be that there would still be separate tables. Keep in mind, also, that a number of linguistic descriptions gloss over the dental/alveolar distinction in descriptions for languages that don't distinguish the two. So SC s could actually still be alveolar. — Ƶ§œš¹ _{[ãːɱ ˈfɹ̠ˤʷɪ̃ə̃nlɪ]} 16:22, 4 March 2013 (UTC)

IPA h/х

Isn't the Serbo-Croatian/Bosnian/Montenegrin h/х doesn't pronounce as Scottish loch with an /x/? But it sounds like in /h/ as hue in Received Pronunciation? ApprenticeFan ^work 13:35, 23 May 2014 (UTC)

Well yes, BCS /x/ is maybe somewhat closer to [x] than to [h], but generally has less friction than e.g. Russian or German /x/... or Scottish, for that matter. As the text says, "English approximations are ... only intended to give a general idea of the pronunciation", and many English speakers won't know how Scottish loch is exactly pronounced anyway. Why don't we keep it simple, then? This page is conceived as a rough guide for general readers, and more advanced and nuanced stuff should be put into Serbo-Croatian phonology. That's why I'm opposing burdening it with too much detail of marginal relevance. No such user (talk) 15:17, 23 May 2014 (UTC)

I feel a great to chime in here.

Serbo-Croatian has no such sound as /x/. This mistake is probably so common because the cyrillic alphabet uses 'x' where the BCS latin alphabet uses 'h'. The latin and cyrillic characters in BCS have a one-to-one correspondence, in this case h/х.

BCS 'h' (latin) and 'x' (cyrillic) are in fact the voiceless palatal fricative [ç]. Check out Laut & ach Laut for the difference, which is an important distinction in German, even though it is an allophone. The Ich-Laut is a voiceless palatal fricative [ç], and the ach-Laut is a voiceless velar fricative [x].

I recommend using 'h' since it is so similar to the 'h' of English and other Germanic languages. It is often pronounced with no audible difference from the English. For general usage, it is more obvious to English readers and more accurate than /x/. 109.60.96.17 (talk) 20:40, 10 May 2015 (UTC)

Nonsense, linguists won't fooled by such superficial orthographic things. It's rather that you see 'h' of Standard Croatian and 'h' of English and do not perceive any difference. However, in both languages [x] and [h] are not distinct, so you won't automatically hear the difference. In [4], it is clearly a velar fricative. Yes, it sounds soft, unlike the typical ch of Scottish and German and like [ç] of German 'ich', but it lacks the palatal element in the sound. If you listen to the sound at voiceless velar fricative, you'll that the prototypical velar fricative actually sounds smooth, like the 'h' of Serbo-Croatian. --JorisvS (talk) 08:17, 11 May 2015 (UTC)

Move discussion in progress

There is a move discussion in progress on Help talk:IPA which affects this page. Please participate on that page and not in this talk page section. Thank you. —RMCD bot 16:18, 15 July 2017 (UTC)

ć, č

it says "Many speakers in Croatia and some in Bosnia have no distinction between /t͡ɕ/ and /t͡ʃ/ (⟨ć⟩ and ⟨č⟩) or between /d͡ʑ/ and /d͡ʒ/ (⟨đ⟩ and ⟨dž⟩)"

but it doesn't say whether they pronounce it as /t͡ɕ/ or /t͡ʃ/...? LICA98 (talk) 01:45, 29 October 2017 (UTC)

@LICA98: I think that in the vast majority of the merging regions the resulting sounds are hard, or between hard and soft. So I'd say that they merge to /dʒ/ and /tʃ/. However, I've heard of Croatian dialects that have only /dʑ, tɕ/, but that's probably not very widespread. I'm not sure. Mr KEBAB (talk) 21:10, 6 November 2017 (UTC)

Right. However, do note that "canonic" ⟨dž⟩ and ⟨č⟩ are actually retroflex, as in Polish, /dʐ/ and /tʂ/ respectively. "Soft" merging into /dʑ, tɕ/ along with fronting /ʃ/→[ɕ], /ʒ/→[ʑ] is actually a feature of Central Bosnian dialects, and is a stereotypical feature of Sarajevo accent. I suppose it's audible in this iconic Top Lista Nadrealista sketch [5]. No such user (talk) 21:51, 6 November 2017 (UTC)

@No such user: There's some debate on retroflexes and whether they can even be laminal. ⟨dʒ, tʃ, ʃ, ʒ⟩ are still a correct transcription, as IPA doesn't comment on their softness, it just calls them postalveolar. Our articles on them present incomplete information. One day I'll try to fix that. Mr KEBAB (talk) 13:38, 7 November 2017 (UTC)

About devoicing and voicing

"Voicing contrasts are neutralized in consonant clusters, so that all obstruents are either voiced or voiceless depending on the voicing of the final consonant, though this process of voicing assimilation may be blocked by syllable boundaries."

Should we add this rule into the help page or not transcribe devoicing and voicing? LoveVanPersie (talk) 20:45, 17 February 2018 (UTC)

But there is a syllable boundary in Zubčić (Zub-čić) and Izetbegović (I-zet-be-go-vić). Now I'm not sure if these are pronounced /ptʃ, db/ or /btʃ, tb/. For now, I'll remove the IPA from Filip Zubčić. Mr KEBAB (talk) 19:02, 19 February 2018 (UTC)

Those are /ptʃ, db/, although you might expect some pushback from well-meaning natives who might have an idea that the writing system is fully phonemic, no exceptions. As evidenced particularly in onomastics, there are exceptions, particularly on morpheme boundaries (Izetbegović < Izet + beg); Zubčić is probably result of historical spelling, Zupčić is a more common form of the surname. No such user (talk) 20:48, 19 February 2018 (UTC)

[ŋ] and [ɱ]

I have some questions about "/n/ is velar [ŋ] before /k, ɡ/, as in stanka [stâːŋka]", cited from Serbo-Croatian phonology.

1. Does the rule include /x/? It's "before velar consonants" on this help page.

2. Does it happen even through there is a word boundary? Same question about /m/ → [ɱ] and "/t͡s, f, x/ are voiced [d͡z, v, ɣ] before voiced consonants".

Thanks! LoveVanPersie (talk) 15:00, 1 March 2018 (UTC)

/ʎ/

Is /ʎ/ precisely the same as SC ⟨lj⟩/⟨љ⟩? According to the IPA pulmonic consonant chart with audio, it seems more like the Spanish ⟨ll⟩. Would it be more accurate with a “velarized” sign ([ʎ̴])? --S. Jevtić (talk) 18:43, 30 June 2019 (UTC)