zompist bboard

Because of some discussion in the 'The "How do you pronounce X" Thread' thread, and a song that I've been listening to, I've been thinking about singing pronunciation and I'm curious about how it is in various languages.

Here's the context:

Qxentio wrote:

linguoboy wrote:

Imralu wrote:A lot of singers turn the HAPPY vowel into something like [e] or even [E] and pronouncing it as [ i ] seems weird to me, like if the person's singing is not amazing, that will make it sound worse.

How did this start?

Could it be related to the lack of happy tensing in Southern American English? I can easily see lowering the [ɪ] to [e] to make it more salient in singing.

Sorry to butt into the discussion so late, but I find the phonetics of singing absolutely fascinating. Lots of things happen that don't happen in any other registers of speech.
For example, I love it when in this song, the singer goes: /loun.le loun.le, laun.liː/. Two totally different pronunciations of the same word.

__________________________

And then I've been listening a lot to this song in Swahili: Madikei - Aliumbwa saa ngapi and I've noticed a LOT of final i's pronounced as e's.

The first part of the chorus goes like this ... all the red e's are normally /i/. It's interesting to here a very clear /i/ in one of the ngapi's with the other short, cut off /e/s (which I've indicated with a grave accent).

Madikei wrote:Aliumbwa saa ngapè
Ui dem aliumbwa saa ngapi
Kwane aliumbwa saa ngapè
Girl nafanya tu mpaka mi namtamane

And a part of the first verse (from about 45 seconds in) contains a lot of the same words but a very exaggerated /e/ pronunciation at the end of the lines where all of these "should" be /i/.

Swali ni aliumbwa saa ngape
Kwani na mimi nilikuwa wape
Swali ni aliumbwa saa ngape
Kwani nilikuwa nimeenda wape

And straight after that, all the lines end in /i/ ... in English words adapted to Swahili phonology with a final epenthetic /i/: ... steki steki ... keki keki ... feki feki ... cheki ... bleki bleki.

So, I'm wondering if this switch of final /i/ to /e/ is from the influence of English singing styles or if it's part of something also in Swahili culture ... or even a more universal thing for saliency of the more open vowels. That would make sense in English where, in most dialects, there is no final short [e] or [E], but Swahili does have final /e/'s and switching /i/ to /e/ potentially results in more confusion.

__________________________

So, my question and the reason for starting the thread, is to see if others have information about strange pronunciations that occur in the singing in English or other languages. I know that tones are typically ignored in singing in tonal languages, but aside from that, I don't really know anything about how people sing in other languages. I'd love to see examples with links to songs where things are pronounced differently from how they would be spoken.

I’m not sure I’d call what Hebrew has strange, but it definitely has pronunciations that aren’t as common in speech as they are in songs.

First, it has register-based pronunciation differences (mostly in proclitic conjunctions/prepositions) which are used to effect in some styles of singing but not others. As an example: the sequence meaning “and when (I did…)” is pronounced vekshe- [vekʃe-] in ordinary speech, and in songs sung by young and hipstery people whose songs are trying to capture a certain “street-style”, or to sound colloquial and unpretentious. On the other hand, the same sequence is pronounced ukhshe- [uχʃe-] in formal speech-registers, and in songs based in more traditional musical styles, or in songs intended to sound poetic, grandiose, dramatic, or formal.

Then there are also pronunciation differences not based on register, but on heritage and/or musical tradition, for example: laxing/centralization of vowels, uvular |r| [ʁ], elision of glottals /h ʔ/, and a very fortitioned pronunciation of fricatives, which characterizes Western-style, young, non-traditional music and speech. This stands in opposition to the alveolarization of |r| [ɾ], the pharyngeal pronunciation of |ħ ʕ| as [ħ ʕ] (instead of the usual, ‘standard’ [χ ʔ]), along with unlaxed vowels and unfortitioned fricatives, characteristic of modern Eastern-style music. In addition, traditional songs are usually sung in the liturgical pronunciation of the songs’ region of origin, which (especially the Yemenite) can be quite unintelligible to untrained ears.

Below are some examples, graded according to perceived register, more or less.


1: Gila Beshari – Ahavat Hadasa (The Love of Hadassah)
‘traditional Eastern’ style: poetic language, high-register proclitics, Yemenite liturgical pronunciation
- sound changes of Yemeni Arabic transferred wholesale onto Hebrew
- nobody speaks like this, but it’s still used for traditional Yemenite songs and liturgy 2: Avraham Fried – Ashreynu (We are Fortunate)
‘traditional Western’ style: poetic language, high-register proclitics, Lithuanian liturgical pronunciation
- sound changes of Lithuanian Yiddish transferred wholesale onto Hebrew
- nobody speaks like this, but it’s still used for traditional Ashkenazi songs and liturgy 3: Shoshana Damari – Tsarikh Letsaltsel Pa’amayim (Ring the Bell Twice)
‘older Eastern’ style: formal language, high-register proclitics, former standard pronunciation
- retention of glottals and pharyngeals; alveolar |r|; no reductions; |ē| is [ej] in open syllables; etc.
- now only used in speech by some elderly people (or else as an affectation expressing Eastern heritage), but it’s still common in Eastern-style songs 4: Sarit Hadad – Ani Me’akhelet Lekha (I Wish that You)
‘modern Eastern’ style: standard language, high-register proclitics, standard/formal pronunciation
- retention of [h] and of some [ʔ]; alveolar |r|; tense vowels; no pharyngeals; no reductions; etc.
- quite a lot of people speak like this, with or without high-register forms, in unmarked or formal settings, but not as many as the below 5: Yuval Dayan – Ad SheTakhazor (Until You Return)
‘modern Western’ style: standard language, high-register proclitics, standard/colloquial pronunciation
- elision of most glottals; uvular |r|; laxed vowels; no pharyngeals; no reductions; etc.
- most people speak like this, with or without high-register forms, in both unmarked and formal settings 6: Gal Malka ft. Statik – Ba Lah Lirkod (She Wanna Dance)
‘street’-style: slangy language, low-register proclitics, colloquial pronunciation
- elision of all glottals; uvular |r|; laxed vowels; no pharyngeals; all sorts of reductions; etc.
- most people younger than about 50 speak like this in informal settings *warning: awful song*


I don’t know if that’s the kind of thing you meant, or if the differences are clearly audible to non-speakers, but I think it’s interesting.

This is honestly just a one-off example in Malayalam, and I've noticed things like this in other songs in other languages, too, but I always thought this particular example was a little weird. There's an old movie song where, at the beginning (in the first line), a vowel + nasal combination is replaced by just a nasalized version of the vowel, but then two lines later (and possibly even the second time in the first line), it's pronounced normally:

[n̪iːˈɾaːɖuʋãː], [n̪iˈɭəjil n̪iːˈɾaːɖuʋaːn] (or perhaps [n̪iːˈɾaːɖuʋãː] again),
'To bathe, to bathe in the Nila...'
[n̪iːjɛn̪ˈd̪eː ˈʋajgi ʋən̪n̪u], [puːn̪d̪iŋgəˈɭeː]?
'Oh, dear moon, why have you come (so) late' [what I've translated as 'dear moon' here is a commonly used phrase in romantic Malayalam songs and probably refers to a girl, not the actual moon, although I'm not a native speaker so maybe I'm wrong about that]
[n̪iːˈɾaːɖuʋaːn], [n̪iˈɭəjil n̪iːˈɾaːɖuʋaːn] (or perhaps [n̪iːˈɾaːɖuʋãː] again )
'To bathe, to bathe in the Nila?' [i.e. why didn't you come take a bath in the river earlier?]

The first time, he's clearly singing [n̪iːˈɾaːɖuʋãː] with a nasalized vowel, which is definitely not the way this word is normally pronounced. The second time in the same line, it's harder to tell because the vowel is also pronounced longer, so there's naturally a longer delay between the onset of the vowel and the point where the singer would be pronouncing his [n] anyway. However, in the third line, he clearly sings [n̪iːˈɾaːɖuʋaːn] at the beginning of the line, which is the usual pronunciation of this word. The second time in that line, again, it's harder to tell so who knows.

There's nothing, as far as I can tell, to be remarked upon pronunciation-wise in most of merengue, salsa, cumbia, tango, flamenco, reggaeton, or pop/rock music in Spanish, other than the fact stress is sometimes placed in an unusual syllable to fit the music or the rhythm. For example, you can hear cuando pronounced [kwanˈdo] ("cuandó") in Alex Syntek's El Camino at 1:19.

• Te esperaré cuando puedas venir
  [t espeɾaˈɾe kwanˈdo pweðas βeˈniɾ]
  I'll wait for you to be able to come
  (note: non-standard grammar in the Spanish line)

Individual singers might have quirks though. The flamenco singer Camarón de la Isla for instance loved to make line-final vowels into diphthongs often. For example, in his song Como el agua you can hear the following at 0:48.

• Como el agua clara
  [ˈkomo el ˈaɣwa ˈklaɾau]
  que abaja del monte,
  [ke aˈβaha ðel ˈmontei]
  así quiero verte
  [ˈasi ˈkjeɾo ˈβeɾtei]
  de día y de noche.
  [de ˈði.a .i ðe ˈnotʃei]
  Like clear water
  that runs down the hill,
  I want to see you this way
  day and night.
  (note: abajar is a non-standard verb)

Note how así is pronounced "[ˈasi]" instead of [aˈsi] because of the rhythm, too.

They do this thing in traditional Japanese singing, where they pronounce each mora as a separate syllable. In contemporary music they either sing "normally", or have some lines in the traditional style and some in the "normal" style. Here's an example with the word kissaten (café): Normally it's pronounced [kis.sa.teɴ] or [kis.sa.tẽː], but in the traditional style it would be sung [ki.is.sa.te.ɴː]. As you can see, the first half of the long consonant is sung as a syllable of its own, by adding a copy of the previous vowel. And then that final nasal is sung as a syllable of it's own. I've also noticed that the particle wo which is normally pronounced [o], because Japanese has lost [w] before all vowels but /a/, is sometimes sung with a faint [w]-sound.

In Welsh, alongside the expected spelling pronunciations of diphthongs (though now increasingly widespread in Welsh anyway) you have the pronunciation of ei eu (the third person possessive pronouns, normally /i~e:/) as /aj/. In fact, this may also occur in other words with the same (spelt) diphthong.

Yng wrote:In Welsh, alongside the expected spelling pronunciations of diphthongs (though now increasingly widespread in Welsh anyway) you have the pronunciation of ei eu (the third person possessive pronouns, normally /i~e:/) as /aj/. In fact, this may also occur in other words with the same (spelt) diphthong.

I was surprised to hear plural -au pronounced [aɨ] in some Ffa Coffi Pawb songs. Usually their pronunciation is more colloquial than that.

Wouldn't know if it's standard in singing Korean, but when PSY sings Gangnam Style, there are points where he lowers /i/ quite a bit. I perceive the /i/ in 사나이 /sʰanai/ as closer to /ɛ/ or even /ɛɚ/ than /i/, and it also sounds to me like he's monophthongizing the /ai/ in 스타일 /sʰɯtʰaiɭ/. (To be fair to him, I don't perceive him as lowering all instances of /i/ where it occurs.)

Kyo (京), the vocalist of Dir en grey and Sukekiyo, pronounces things in ways that sound strange sometimes, but a big part of it is simply that he makes some weird sounds to begin with. But even in regular singing, especially in old songs like this one, the /ɾ/ is an acutal [l] at times. I know he's not the only one who pronounces it like that and that it happens in speech as well, but still, it sounds pretty dramatic considering how Japanese doesn't usually have any true L-sounds. The う vowel occasionally being either a full-blown or [ɯ] doesn't seem uncommon overall in sung Japanese, or ふ sounding more like [hu] than [ɸɯ̟ᵝ]. I've also noticed おう /o̞ː/ and えい /e̞ː/ pronounced either as diphthongs or as [o̞u] and [e̞i], reflecting how they're written; I guess this is just an example of what Qwynegold already mentioned earlier in this thread, but I don't think it only happens with that kind of exaggerated pronunciation.

The Finnish pop singer Sanni sometimes pronounces /ɑ/ as [a~ɐ̟] instead of the usual [a̠~ɑ~ɑ̝~ɐ̠], making it sound almost like /æ/, but also sometimes pronounces /æ/ as [a~æ̠]; this song demonstrates both. Some others do this, too, but I don't remember any that do it as much or as clearly as her even though I'm sure they exist. It's pretty common to pronounce /y/ as an actual [y] rather than [y̠~ʉ̟] when singing, and rarely the opposite of backing it to [ʉ] happens as well. Fronting of /u/ also happens, but never all the way to [ʉ] and it's not unique to singing; Wikipedia seems to claim this and a bunch of other phonetic details only exist in Estonian, but that's not true.

Kazakh singing tends to sound like it's articulated more clearly than speech, eg. /e/ and /i/ being [e~e̞] and instead of [ɘ~ɘ̟~ə̟~ə] and [ɪ~ɨ~ɨ̞~ɘ] or whatever like they usually are in speech. Not that I'd have ever heard Kazakh in person or spent much time listening to recordings of it being spoken, though, and Wikipedia isn't always fully accurate when it comes to phonology, so there might be more to it but I feel like it's more obvious than most languages with its "clearer" pronunciation when sung.

StrangerCoug wrote:I perceive the /i/ in 사나이 /sʰanai/ as closer to /ɛ/ or even /ɛɚ/ than /i/, and it also sounds to me like he's monophthongizing the /ai/ in 스타일 /sʰɯtʰaiɭ/. (To be fair to him, I don't perceive him as lowering all instances of /i/ where it occurs.)

Is it always after /a/? Because that's a phenomenon called "diphthong harmonisation" found, for instance, in West Germanic languages like Old English and Modern German.

linguoboy wrote:

StrangerCoug wrote:I perceive the /i/ in 사나이 /sʰanai/ as closer to /ɛ/ or even /ɛɚ/ than /i/, and it also sounds to me like he's monophthongizing the /ai/ in 스타일 /sʰɯtʰaiɭ/. (To be fair to him, I don't perceive him as lowering all instances of /i/ where it occurs.)

Is it always after /a/? Because that's a phenomenon called "diphthong harmonisation" found, for instance, in West Germanic languages like Old English and Modern German.

To the extent I've paid attention.

Edited to add: I believe it is. I've read along with the hangul while listening to the song in another tab, and I don't notice any lowering when /i/ does not directly follow /a/ (edited again to add: except in English <lady>, and even though I may not hear [iː] there, it still registers in my brain as /iː/ since it's, for lack of knowledge of what the linguistic term is, a "slangy" pronunciation).

So, my question and the reason for starting the thread, is to see if others have information about strange pronunciations that occur in the singing in English or other languages. I know that tones are typically ignored in singing in tonal languages, but aside from that, I don't really know anything about how people sing in other languages. I'd love to see examples with links to songs where things are pronounced differently from how they would be spoken.

As far as I know, doesn't happen in Polish, unless the singer specifically overpronounces things, or the song is explicitly regional. Apart from the fact that sung Polish is in many cases pronounced more carefully and lacks most contractions and assimilations of your average spoken Polish.

Concrete examples: (selected quite randomly)

Alicja Majewska — Odkryjemy miłość nieznaną

All phonemes pretty much have their standard values, along with the realization of final /ɛ̃/ as oral [ɛ] — maybe except the trill /r/ ([ɾ] in spoken Polish) and the /ɨ/ (which has a more reduced value in spoken Polish, closer to [ɘ]). The /u/ in „czuć” sounds fronted to me, but I'm not sure it's a specific enunciation, or assimilation to the following palatal. Also, /a/ seems to be the front [a] rather than the central [ä], but it wouldn't sound out of place in spoken Polish either.

Also, if I'm hearing correctly, there's something weird going with final /a/'s in the first two stanzas — for me it sounds like in „da” and „portowa” it's backed to [ɑ], versus its standard realization in „prawda” and „słowa”, which breaks the rhyme a little bit. Actually, I wonder if it's some weird reflex of the Old Polish long „a” that has been retained in Majewska's personal dialect…


Paweł Kukiz — Odpłyniesz wielkim autem

Quite standard pronunciation, even with some reductions (final /ɨ ɛ a/ → [ə ɛ̈ ɐ]?). Also, it sounds like this particular singer tends to enunciate /a/ as long front nasal [ãː].


Organek — Ultimo

The vowel sounds seem all over the place, /ɛ/ becomes [e], /ɔ/ becomes [o], /a/ becomes [a] and [ɑ] in random places — due to the personal style, I think.


Grzegorz Turnau — Pompa

All vowels have their standard values with little to no reduction or variation. Regarding consonants, I'm under impression /tʂ ʂ ʐ/ here are close to palatalized [tʃ ʃ ʒ].

Actually, singers are trained in Poland to drop nasalization of word-final vowels because it "sounds better" and is an accepted alternative pronunciation anyway.

My personal theory with /i/ > [e] is that restricts your voice more. You can generally sing higher if you take [e] over . But it's an unfounded hunch I have as a person who sings a lot.

gestaltist wrote:Actually, singers are trained in Poland to drop nasalization of word-final vowels because it "sounds better" and is an accepted alternative pronunciation anyway.

That's why I've written “standard values”.

I don't know if this is to do with my growing up in Albania and being surrounded by (mostly mid-west) American English rather than my native British English, but when I'm singing songs that I don't know were written by someone with a British accent, I'll suddenly get the cot-caught merger and dark l for no clear reason, and my sister does the same, but we stay non-rhotics unless we've heard it sung otherwise. I'm assuming it is to do with the fact that we grew up speaking '4 languages': American with friends, Albanian in public, English with my mum, and Swedish with my dad, and so it has never felt unnatural for me to do an American accent, and still doesn't that much (although I'm aware its much worse than it used to be), butI would be interested if anyone does anything similar.

Well, as we've discussed here, many British singers use an American or half-American accent - e.g. American vowels combined with non-rhoticity, which is similar to what you describe.

Some research shows that this process isn't entirely conscious:
http://aut.researchgateway.ac.nz/bitstr ... ibsonA.pdf

Qwynegold wrote:They do this thing in traditional Japanese singing, where they pronounce each mora as a separate syllable. In contemporary music they either sing "normally", or have some lines in the traditional style and some in the "normal" style. Here's an example with the word kissaten (café): Normally it's pronounced [kis.sa.teɴ] or [kis.sa.tẽː], but in the traditional style it would be sung [ki.is.sa.te.ɴː]. As you can see, the first half of the long consonant is sung as a syllable of its own, by adding a copy of the previous vowel. And then that final nasal is sung as a syllable of it's own. I've also noticed that the particle wo which is normally pronounced [o], because Japanese has lost [w] before all vowels but /a/, is sometimes sung with a faint [w]-sound.

It's not just traditional singing, though, is it? I've heard it in plenty of anisong. It seems to me that there are complex rules to do with sentence stress as to when it's used. E.g. in this song there are lots of syllabified んs but the ましょう at the end of each line is only pronounced as two syllables: https://www.youtube.com/watch?v=LN-x3IRjYZc

As for wo with [w], I've heard it applied to お as well, e.g. in 'yukou' here https://youtu.be/bYCRTmGonlM?t=34. The kind of singing style in this song is something that seems particular to very genki singers, with overemphasised and diphthongised vowels and vocal fry.

RE British English singing pronunciation, I can confirm it's very normal to sing with a slight American accent. In fact, if you speak SSBE or anything near RP as I do, it sounds incredibly stilted/forced to sing in my 'normal' accent. An interest case of this is Marina Diamandis, a Welsh pop singer who has a weird mix of singing pronunciation, Welsh Englishisms, and specific singing style: https://youtu.be/ZNE-7DyLJ5w?t=49. In this example her pronunciation of 'years' stands out a lot.