Weak lips
A recent post by Matt ponders some historical Japanese phonology, particularly the loss of (some instances of) intervocalic /m/ and the fusion of /au/ to /oo/ (written <ou> おう) in modern Japanese, though only rarely pronounced such. I’d guess that the change was, put simply, [ɑmu] > [ɑ̃ũ] > [ou] (> [o:]).
In fact, through my study of Japanese as a second language, even before I started to learn about language change and common paths of phonological change, I noticed a certain type of pattern of sound change in Japanese occurring both in history and synchronically. Namely, weakening of labial stops. In addition to the the above change, also mentioned in this languagelog post by Bill Poser, there is also the well-known p > ɸ > f changes from Classical to Old to Modern Japanese, which gives kawa from kapa (river), and lies dormant in compounds like tabi-bito ‘traveler’ (didn’t get that? Look up rendaku). p is also preserved in non-intervocalic environments like geminates and post-nasally. And of course, I need not mention the loss of w before all vowels but ɑ.
The synchronic grammar also expresses labial weakening in various verbal contractions. Those familiar to most learners of Japanese are kereba → kerya (hypothetical form) and tewa/dewa → cha/ja (gerundive + TOP). There is also reba → rya (an alternative for the hypothetical suffix) and ewa → aa (as in kore wa → koraa ‘this TOP’, boku wa → bokaa ‘1sg TOP’)).
Now, there’s a lot of synchronic phonology that gets very complicated with all of this, especially the historical loss of p, which likes to reappear when least convenient. There’s not only an h~p alternation in several Sino-Japanese words (方法 hoo-hoo, 文法 bun-poo), but carry-overs from when t-final SJ loans were pronounced in compounds without an epenthetic vowel. For instance, 一 has pronunciations iʧi and itsu, perhaps from underlying (or original) it. In compounds, say with 分 hun, what was originally 一分 it-pun has become modern ip-pun. In fact, in this year’s LSA annual meeting, Japanese phonologist TATEISHI Koichi presented a paper on the representation of underlying /p/ in modern Japanese, presenting data from accent position on words that lost intervocalic p. In the question period I brought up a pattern seen on a popular TV shop in Japan, where a made-up classifier hee is used to count how “wow”-worthy something is (from the exclamation hee!). When used with numerals that create geminate consonants, the h becomes not a geminate p as might be expected, but actually a geminate h: juhhee from juu-hee (instead of juppee).
Nice! I probably had actually read that LL post, since I think I was reading that far back, but I’d forgotten it completely. Pity, since it explained the very thing I was posting about in much greater detail. (shame)
It’s funny. I only remember the LL post because I once attempted to explain it to someone without referencing it, and got it totally wrong and was called on it. Getting stuff wrong is really good motivation to get it right the next time.
Especially on a blog archived for all time.
It occured to me today that there’s one odd caveat to the general “labials get weaker” rule: the preferance for a /w/ between two /a/s, even to the extent of adding it if necessary. I know that 可愛い and 可哀想 are ateji and don’t prove anything etymologically, but I know a couple of people who pronounce 場合 bawai.
Yeah, as soon as you mentioned the w between two a’s the 場合 case came to mind. I brief search of Google-Japan gives several discussions among Japanese people as to why this pronunciation exists. Some suggest hypercorrection, but I don’t think that’s right. Most see the problem with the long /a/ sound, which is indeed rare in the native vocab (I can only think of kaasan), and it’s also across a morpheme boundary in this case. Also - I wasn’t aware that 可愛い and 可哀想 were ateji, though it makes sense. (可愛 is a word in Chinese, though for all I know it could have been borrowed from the Japanese ateji - that would be pretty funny if true)
IIRC かわいい is supposed to come from かはゆし which in turn comes from かほはゆし (顔映し), which originally meant something like “embarrassing [to look at]” and is parallel to other words like 目映し (まばゆし) and 面映し (おもばゆし). I don’t know if 可愛 existed in Chinese and was applied because it made sense once かわいい had evolved into meaning “cute, appealing”, or if the Chinese borrowed it from Japanese, though…
I was under the impression that かわいそう was an offshoot of かわいい back when the latter still mostly meant “pitiful, unbearable to look at”, but I don’t remember where I got that idea.
I am told you also get geminate h’s in Japanese tautomorphemically in loanwords, e.g. [mahha] for German “Mach” and [jihhi] (where [j] is the affricate) for German “sich.”
Kevin: Yes, that’s true, and I’ve also heard [beççjo:] (i.e., べっひょう) for 別表 [beppyo:] ’separate table,’ though that may have been…uh, a slip of the tongue, perhaps? That, too, though, is across a morpheme boundary.