Syllabic /n/ (and other things) in Japanese singing

[eldin raigmore]

When Yoko Takahashi sings "Zankoku na Tenshi no Teeze" ("The Cruel Angel's Thesis", the opening theme song of Neon Genesis Evangelion), I notice that the first word of the first line is four syllables:
/za n ko ku/
and the first word of the second line is also four syllables:
/ʃo u ne n/
and later the word hohoenderu is six syllables:
/ho ho e n de ru/.

Are those instances of the /n/ phoneme, syllabic in Japanese, in any context other than song?

I also noticed many vowel-clusters had a syllable-boundary between the vowels; in other words, a preference for hiatus as opposed to polyphthong.
Something similar happened to (at least some of) the geminated vowels; rather than being pronounced as long vowels, they were pronounced as if a syllable boundary were between the two occurrences of the vowel.

Am I wrong?
If so, what's right?
If I'm not wrong, what's the explanation?

[hr][/hr]

While I'm thinking of it;
I liked the anime very much; but I found the final episode (episode 26) very disappointing.
How did other viewers feel?

[Ahzoh]

That's because the /n/ is moraic, so it sounds like it's in a syllable of its own.
https://en.wikipedia.org/wiki/Mora_(lin ... )#Japanese

[eldin raigmore]

That's because the /n/ is moraic, so it sounds like it's in a syllable of its own.
https://en.wikipedia.org/wiki/Mora_(lin ... )#Japanese

Thanks.
I might have had a vague and tenuous recollection it might be something like that.
But I wasn't sure (until hearing the song) that it was actually attested in the pronunciation.

I still don't know how perceptible that effect is in prose, especially in allegro speech.
In fact, how perceptible is it in spoken (as opposed to sung) poetry?

Do you know?

[Ebon]

Not an expert, but syllabic n and separate vowels as opposed to diphthongs or long vowels are far more pronounced in singing than in speaking in my experience. It's particularly obvious in the -tai form. I've never heard anyone say the vowels separately in speech, but it happens all the time in singing.

I don't know about how this is done in poetry, though.

[sangi39]

I noticed a similar thing in a few songs by BabyMetal, Yousei Teikoku and Wagakki Band and a couple of intro songs to anime. I think Ahzoh and Ebon might be right in that it's kind of just highly enunciated moraic /ɴ/. It seems to appear, at least to me, mostly where it would fit the rhythm better, almost like the use of "banishéd" as opposed to "banished" in Shakespeare's works (I think?... "hence banishéd is banished from the world").

[Sumelic]

Yes, in singing different morae are often split up into distinct syllables, but this is not that common in ordinary spoken Japanese in my experience. Another thing you may hear in singing is people giving a separate syllable for the first half of geminate consonants, so something like "itte" might be pronounced as three syllables "i-it-te." In speech, the preceding vowel is usually a bit lengthened before a geminate (opposite to the pattern we get in e.g. Italian; however, I've read that geminate consonants do tend to shorten the following vowel in Japanese), but I don't think it's common at all to add an extra syllable for it like this.

I think even in singing, a diphthong may be used if it's necessary for the rhythm. But I can't think of an example at the moment.

[Chagen]

I have most definitely noticed this in many Japanese songs. For instance this one, at 0:23 where they sing kanra kanra to warai yamanu haratsudzumi, the beginning is very obviously [ka.ɴ.ɾa ka.ɴ.ɾa]. I have always found it very odd that this moraic nasal is uvular in Japanese, though. The phoneme was introduced from constant influx of Chinese loanwords, yet there are no uvular consonants in Chinese and pretty much none in ALL of Southeast Asia IIRC.

Also in my experience geminates in japanese songs are usually pronounced with a short vowel and a slight pause, while long vowels are two short vowels in hiatus. Not in the song I posted, but then again they're rapping. In another song by the same group (different rapper), the geminates are pretty much pronounced the same as short consonants, but then again, AO's voice is so weird (and cute!) I wonder if he's not speaking the standard dialect.

Also I'm using this topic signal-boost TAMAONSEN because they're fucking awesome.

[KaiTheHomoSapien]

Love Evangelion

I've been listening to Japanese music for years now and I've noticed a lot of what you've noticed. It does sometimes seem that in song, they often pronounce each mora as a separate syllable, thus /n/ is syllabic; diphthongs are pronounced as two sounds, e.g. "nai" as /'na.i/ as opposed to /naj/; vowels before geminates are pronounced as two syllables, e.g. "atte" pronounced /'a.a.te/, and long vowels are pronounced as two syllables, e.g. "hontou" as /ho.'n.to.[w]o/

I also notice that the particle "wo" (in speech pronounced /o/) is actually pronounced as /wo/ by most singers in many instances. And that what is /ɾ/ in speech, like at the beginning of stressed syllables with <r> is pronounced exactly as /l/. I know one song where the singer pronounces "warai" as /wa.'la.i/

I kind of think Japanese sounds nicer when sung than when spoken.

[qwed117]

Shhh, but English does this too. And I bet more langurges do too.

[Lao Kou]

Shhh, but English does this too. And I bet more languages do too.

I liken Japanese "n" to French "e" in song and poetry.

[Squall]

The phoneme is not /n/. It is /N/. It gets assimilated before other consonants (it is /m/ before /p b/).
It is /n/ only before /t d/ and affricates.

This is how I perceive the first sentence of that song.
/za ũ 'koku na'tẽshi no'jo on'i/

[clawgrip]

Japanese accent operates at the mora level, but moraic n is normally not syllabic. Outside of certain colloquial mergers, adjacent vowels are kept distinct, but there is a smooth transition between them, making them sound diphthongy. Even so, the two (or more) vowels are still very clearly identifiable in most cases.

People do not normally speak in distinct morae unless they are spelling a word out for someone. So 進行 shinkō may be pronounced shi-n-ko-u if a person is purposefully spelling it out for whatever reason.

Poetry, and verse in general, operates at the mora level. A haiku, for example, is not 5-7-5 syllables, it's 5-7-5 morae. Modern songs follow this same method. I suspect this tendency was reinforced by the writing system, which also operates at the mora level.

As people have mentioned, geminate stops/fricatives in song are usually pronounced with an echo vowel in place of the geminate, mainly because this allows the singer to continue singing, where a geminate would force the singer to go silent for a moment in order to delay the release of the consonant, and this is aesthetically undesirable.

wo is also often pronounced with the /w/ in song, which it never is in normal speech unless, as above, someone is spelling something out. I get the feeling this is more common when wo appears last in a line or on some sort of stressed beat or what have you. If wo is prominent, pronouncing the /w/ helps keep it distinct from the preceding vowel.

[eldin raigmore]

Thanks, everyone! I think I now know more than I expected to!