Hypothesis: Gibberish has vowel harmony

One thing I have noticed about my father's pronunciation of Japanese proper nouns is that he seems to prefer harmonizing the vowels. For instance (pardon the simplified phonemic transcription):

  • "Yamanote" → /yEm@'noto/
  • "Ikebukuro" → /ikib@'kjuro/ (< OG /ik@'bjuk@ro/)
  • "Asakusa" → /@'suk(@)sa/
  • "Narimasu" → /nEr@m@'su/
  • "Ginza" → /'ginza/

Because he knows less Japanese than an extra from Gone with the Wind, these names have no etymological meaning for him -- they're gibberish, basically. And it seems that he prefers to have only one type of vowel/diphthong* per stress group, if that's the right term, and differing types in neighboring stress groups.

Thus, for example, /ikib@'kjuro/ divides into a front-vowel part before the stress, and a back/central-vowel/dipthong part afterwards.

Schwas seem able to substitute for any vowel, and may be involved in postpone stress where necessary for harmonic reasons (e.g. "Narimasu", where if the stress was on the second-to-last syllable as expected you would have two different kinds of vowels in the same stress group.) They may also be rapid-speech expression of underlying "real" vowels: for example, the /u/ in his "Asakusa" -- why has it been moved up, if not to (1) harmonize with the following /u/ expressed as a schwa (if at all) between /k/ and /s/; (2) contrast with the press-stress /a/ or schwa; or (3) both?

The final unstressed /a/s are interesting because they seem to to be an exception, but I would explain them away as imported English word-final habits: they have exactly the same sound as the end of a word like "rapture" or "boxer" in Australian English. (The final /o/ in /ikib@'kjuro/ may be a similar phenomenon, although the only example I can think of offhand is "bureau".)

In other news, I accidentally bought a ladies' (as in -wear) umbrella to stay dry on the way home today, and it was so small I felt like a gnome holding a toadstool.

* All those /o/s on the right are really a diphthong [oU] like at the end of "no" or "grow"

Popularity factor: 13


Kamikaze = "kamakazi"
Karaoke = "kari-oki"
Hara kiri = "harri carri"

I always saw this as Americans' difficulty in saying hard "e" at the end of words and hard "a" in the middle.


Americans? I don't think other English speakers pronounce these words any better. Here's a couple of citations from the OED:

859 Times 18 Aug. 10 These officers no longer perform hari-kari, or in other words disembowel themselves, rather than survive the disgrace of admitting foreigners. 1862 HOLMES Hunt after Captain in Old Vol. of Life (1891) 58 He will very commonly consent to the thing asked, were it to commit hari-kari.

(The Times, of course, is the London paper.)

Matt: Has it ever occurred to you that you may in fact be a gnome holding a toadstool?


I say this with the toughlove of someone who's been there all too often, but the correct spelling is diphthong.

That said, this would make a cool paper if you could get enough raw material.


Language Hat: Those are the kind of questions that keep a person up at night...

Matt: I noticed the "Asukasa" phenomenon before. First my mother said it, and I didn't correct her, because what's the point? It's just a nonsense word to her. But when I heard other English speakers in Japan keep calling it "Asukasa", I began to halfway expect the English subway announcements to say it!


My parents couldn't stop saying "Takiyama" instead of "Takayama"...


(Just for the record, my dad and I are both Australian, although his Australian English accent is not that strong [he doesn't quite say "die" for "day", etc] and mine is even weaker [my only big Australian giveaway is not pronouncing the "r" in words like "car" (while simultaneously not being nasal enough to pass as a Bostoñero)].)

Anon-0: Hmm, those are good points too. Influence from the phonemics of the speaker's native language is obviously pretty important, maybe so much so that it obliterates the need for a vowel harmony hypothesis... Even exceptions like "yamanote" -> "yamanoto" could be explained as interference from well-known Japanese proper nouns like "Yamamoto". But I still have the sneaking suspicion that "vowel harmony" and/or "rhyme" would be the last-ditch way to handle unusual sounds. I need more data!

LH: thanks for the OED info!

And yes, that is my most secret fear.

Scott: Ha ha ha ha ha! I don't know what you're talking about! Why, that's how I HAVE spelled it! See?!


Morgan: Yeah, I don't correct my father every time (because that would make it impossible for him to tell me anything about what he did that day), but I do occasionally, because I want him to be able to ask other people (perhaps those whose English is not so great) for help if he needs directions or something, and I don't know how much deformation of famous place names is permissible before they become unrecognizable to native speakers. ("Excused me, Mr Noy Yarker! Are you knowing where the Breedway is?")

SHK: Valuable extra data point! (Coughthatunderminesmythesiscough). Thanks.


Hi there,
I found your blog by searching for "don't look at the sword" on Google as I was researching as to why nobody ever looks at the samurai that is holding the sword (they end up getting killed as they are looking at the wrong thing.) Will pass the address on to some people I know are interested in Japanese culture.

Gaijin Biker:

What is the linguistic term for breaking a syllable into two? Like saying To-ki-yo or Ki-yo-to instead of Tokyo or Kyoto.


You sure your old man didn't have an ancestor who fought at Gallipolli and came back with a chronic case of vowel harmony?

The kyo CCV sequence is the only one that Japanese has that English doesn't, so English speakers do exactly what Japanese speakers do to English consonant clusters: break them apart with vowels. Kyu, of course, works just fine in English.



...I'm sorry.


Thanks, anon-1, that's very kind!

Joel: Now that you mention it there is that family legend about Gallapalla...

Is there actually a technical term for that process of consonant-separating, though? It seems like there must be. Resyllablization or something?

Morgan: ... (p.s. I'm jealous of your job)



Vowel "epenthesis" is the term you're looking for, I believe.

I've been thinking: Does Japanese /kya/ occur in native words? That's another Japanese CCV cluster that English doesn't have, but as far as I know it only appears in Japanese words borrowed from English words beginning with ca- as in cabbage, camping, and cancel.


Thanks Joel!

As far as I know /kya/ doesn't appear in any native Japanese words except for some onomatopoeia/etc. and as a contraction in forms like "shinakya". (If it was a native sound, it would probably have its own kana row.) I guess being non-native it isn't in many place names (or personal names) and the opportunities for non-speakers to mangle it are fewer.

Comment season is closed.