狂ったブログ

A large percentage of Japanese vocabulary comes from Chinese (more than half?). As I’ve heard it, the original Japanese language had no writing system (or it was lost to history) and when the Chinese writing system was adopted during the first millenium AD, a large quantity of Chinese vocabulary was brought with it. In addition being used for those words borrowed from Chinese, kanji (漢字, Chinese characters) were also mapped to native Japanese words according to best semantic fit. As a result of this, most kanji have two or more common pronunciations or readings in Japanese, usually one each of the two main types of readings: on’yomi (音読み) and kun’yomi (訓読み). On’yomi is considered the “Chinese” reading, and they are mostly used in words derived from borrowed Chinese words. Kun’yomi is considered the “Japanese” reading and is usually used for words that derive from older Japanese words that pre-date the absorption of so many Chinese words into the language. In many instances there are multiple common on’yomi for the same character, due to the fact that Chinese words using the same kanji were borrowed during different dynasties and from different regions (some kanji dictionaries break these down). Probably more common, though, are characters with multiple common kun’yomi, often drastically different from each other, indicating that they were different Japanese words for which that character was considered the best semantic fit.

One consequence of all of these borrowed words is a preponderance of synonyms (and, because of the fairly limited phoneme set of Japanese, a preponderance of homonyms as well), particularly in pairs. Generally speaking, given a synonym pair of one on’yomi word and one kun’yomi word, the on’yomi word is seen as more formal or katai (固い, or hard) and is primarily used in writing, news reports, and official communications (e.g., a bank teller serving a customer or an announcer at a train station). Along with various other factors, this means that learning written or formal Japanese is in many ways like learning a different language from informal spoken Japanese.

Also like most languages, old Japanese often had multiple distinct semantic values for one lexeme, so many lexemes from old Japanese ended up having several Chinese characters assigned to each of them. This means that in modern Japanese often several words with similar pronunciation but completely different kanji are etymologically connected. Long before I found out about any of this, I remember noticing an apparent connection in pronunciation and meaning for certain words despite the different kanji. In particular I remember 始める (hajimeru, to begin) and 初めて (hajimete, for the first time). There was also the strangely coincidental pronunciation of 湖 (mizuumi, lake) which seemed a lot like the concatenation of 水 (mizu, water) and 海 (umi, ocean), although my teacher at the time insisted that it was nothing more than a coincidence.

I first learned about the phenomenon a few years back, and have since been noticing likely etymological connections, particularly the obvious ones where there are several words with the same pronunciation and the same basic meaning (e.g., 合う, 会う, 逢う; 切る, 斬る; 診る, 見る, 観る). Unfortunately I’ve never seen a Japanese->English dictionary that includes notes about etymology, so they’ve primarily just been suspicions. The other day, however, I discovered that the Japanese->Japanese dictionary in my denshi jisho (電子辞書, electronic dictionary) has etymological notes for a lot of these connections, and it has enabled me to go exploring through all of my past suspicions about these odd little “coincidences” in Japanese word pronunciations. One interesting pair I’ve been able to confirm is 捨てる (suteru, to throw away) and 廃る (sutaru, to collapse, go to ruin). I was also glad to see that 雷 (kaminari, thunder) comes from 神鳴り (kami and nari, i.e., God crying out), even though the character 雷 itself has nothing to do with God and is composed of the character parts for rain and a rice field (it is probably the Chinese word for thunder, and the pronunciation is probably quite different).