kanji - Apparently unrelated 音読み?

For most 漢字 with multiple 音読み, there appears to be a fairly reasonable link between the sounds, and fairly minor drift, such as with 聞 being モン or ブン.

Some 漢字, however, appear to have completely unrelated 読み, the prime example being 回 with カイ and エ. If it was just that one it could be readily dismissed as an historical accident, but the カイ/エ 読み is shared with the also visually unrelated 会 and friends.

So the question is; what's going on here?

Update: As l'électeur points out, this might be explained by the size of China, but if so: Why don't more 漢字 have these apparently unrelated 読み?

Answer

Taking up the chance to talk about "Unrelated readings" because I get the chance to talk about both Japan and China's phonologies in the past.

It's well known that even amongst 音読み, there are the types from different Chinese regions and times.

漢音：Hanyin, from the original middle-chinese pronunciations

呉音：Wuyin from SE China (state of Wu)

唐音：Tangyin from Tang China around the late 1500s

慣用音：These are actually Japanese corruptions of Chinese readings, not used anywhere in China. コク for 石 is one example, which was corrupted from ジャク (the approximation of Hakka's pronunciation, shak)

There are others, usually taking name from the Chinese state that they originate from.

General sense

When Japanese first borrowed Chinese character sounds （漢音）, Chinese had a slightly different phonology to now. Firstly, /h/ was /φ/ (same sound as ふ nowadays, which never changed). Usually, Chinese's /x/ (pinyin: h) was borrowed as /k/, but sometimes as /w/. Examples of this are 会 (ZH: hui JP: kai, e) and 絵 (ZH: hui JP: e). Tones weren't borrowed at all.

They also copied more closely the ~ao vs ~ou distinctions in Chinese (compare 道 and 豆 ZH: dao, dou JP: dou, tou). Japanese dictionaries kindly list pronunciations of kanji that have changed. Dictionaries tell you that 道's reading of どう actually used to be だう.

There are too many rules to list all of them out. Common ones are those above, and final 〜う instead of ~ng found in Chinese varieties (such as 同{どう} ZH: tong).

Both Japanese and Chinese (taking Chinese as one language) have since changed, but there are enough varieties of Chinese that it's usually clear to see which language / dialect is the divergent one (like how Mandarin has no -p, -t, -k final consonants anymore)

In the purest sense

"Unrelated" on'yomi are almost always going to be 慣用音. 良 has a fair few, such as ら, which sound nothing like the Chinese liang used today. This is because they actually diverged from the Chinese loanwords into something distinctly Japanese, and is also why they're so uncommonly used in actual words (I can't think of any examples of 良 as ら other than in names).

Other reasons

One thing to note is that sometimes the Chinese itself was unclear when Japan borrowed the sounds, even without different languages or bad transliterations. There are kanji that have readings with /m/ initials, and some with /b/ (文聞無望亡 to list a few). These are all taken from the same Chinese regions and time, but people probably couldn't say which it sounded more like, and use both /m/ and /b/ readings these days (and Chinese has all 3 starting with /w/).

So, the reason why kanji are unclear are usually:

1. Japanese borrowed them badly in the first place because it didn't have the right sounds to say them properly (eg 花{か} ZH: hua)


2. Japanese borrowed them properly, but then its phonology changed and now doesn't resemble as much (eg 絵{え} from ゑ ZH: hui, said like "hway")


3. Chinese was in a transitional phase when Japanese borrowed them, so Japan took a few pronunciations all from the same Chinese dialect (eg 聞, read as ぶん and もん)


4. Japanese borrowed properly, and just corrupted them (eg 良{りょう} as ら)