Viking101
New member
From: USA
Registered: 2008-08-18
Posts: 7
Is there a good guide that lists when kanji become voiced? For example, the on-readings for 済 are サイ and セイ. However, in the word 経済, it is read けいざい. This happens to a lot of kanji when they are placed after certain sounds. Is there a guide that explains the rules for when these changes occur?
jajaaan
Member
From: America
Registered: 2009-11-14
Posts: 115
Any kanji not in initial position may become voiced. I'm not a specialist in Japanese linguistics, but in many languages, there is a phonological phenomenon whereby an unvoiced sound can assimilate the voicing of a neighboring sound. Since Japanese syllable structure is basically (C)V(n), you're pretty much guaranteed to have a voiced sound preceding every unvoiced sound. Exceptions abound, however, and I'm not claiming to know all the complexities of Japanese phonology, so don't take as the only answer.
magamo
Member
From: Pasadena, CA
Registered: 2009-05-29
Posts: 1039
This is called 連濁 (れんだく). If you google for the term, a lot of informative websites will show up. I might have posted about it a few times somewhere on this form, but I guess these wikipedia articles are more than enough:
http://en.wikipedia.org/wiki/Rendaku
http://ja.wikipedia.org/wiki/%E9%80%A3%E6%BF%81
If you don't like to read it, just remember k, s, t, and h often become g, z, d, and b respectively when they're the first consonants of the second components in two-word/two-kanji compounds, e.g., 株式会社 (かぶしき + かいしゃ = かぶしきがいしゃ). 経済 is a "s -> z" example. The Wikipedia articles give several rules when they don't get voiced.
If you want to understand the whole thing about 連濁, you need deep knowledge of phonology of classical Chinese and classical Japanese, and Japanese linguistics in general.
There are similar things when it comes to readings of kanji compounds. For example, I'm pretty sure you'll wonder why 因縁 is pronounced いんねん rather than いんえん and why 作曲 is さっきょく instead of さくきょく as your Japanese gets better. These phenomenons are called 連声 (れんじょう) and 促音便 (そくおんびん) respectively. As is the case with 連濁, there are general rules. But they're not very practical. 連声 is kind of universal across human languages so probably you'll pick it up naturally. As for 促音便, I explained it briefly when someone asked about pronunciation of Japanese counters:
http://forum.koohii.com/viewtopic.php?pid=64121#p64121
Anyway, I also think it's better to memorize each pronunciation as you come across it and not to rely on rules. It's wrong to say there are no rules because there are. But, as always, they're not practical just like any other complicated grammar rules.
Last edited by magamo (2010 January 29, 8:53 pm)
magamo
Member
From: Pasadena, CA
Registered: 2009-05-29
Posts: 1039
yudantaiteki wrote:
magamo wrote:
It's wrong to say there are no rules because there are.
There are some guidelines, and some rules that describe specific situations where rendaku does not occur. But there is no set of rules that will allow you to predict accurately when rendaku will occur or not occur just given the two words/morphemes. Even in linguistic texts you will not find them.
Per wikipedia:
連濁は必ず起きるわけではなく、さまざまな条件によって阻止される。しかしどの条件にも例外があり、また条件をクリアしていても、偶発的に連濁の起きない語は多い。そのため、少なくとも現代語では、連濁が起こるかどうかは完全に予測できるものではない。
That only says there is no set of rules that can predict whether 連濁 occurs with 100% accuracy. But there are rules as to when it tends to occur such as the one I just posted and the "ng" sound rule in classical Chinese. The exception rules listed on the Wikipedia article is meant to be applied to words that should trigger 連濁 according to the rough when-it-can-happen rules. There still remain exceptions, but there are rules.
Edit: I did a little googling and found a nice website that explains 連濁, though it's written in Japanese:
http://members.e-omi.ne.jp/ichhan-h/ren … nts%29.htm
As the above site says, it's a pretty hard problem to explain 連濁. To predict if 連濁 occurs in a word fairly accurately, you at least need to know the original sound of kanji in question in classical Chinese, when the word in question was born, and when-it-doesn't-occur rules (The above website gives a lot more rules than you can find on the Wikipedia articles). As you can see on the website, this isn't still enough. So learning all the known rules isn't practical at all, but there are (incomplete) rules and some people are studying the phenomenon.
Last edited by magamo (2010 January 29, 10:17 pm)
magamo
Member
From: Pasadena, CA
Registered: 2009-05-29
Posts: 1039
Viking101 wrote:
Thanks for the help. I was hoping for an easy formula. Since it's Japanese, I should have guessed it would be more complicated than that.
The simplest explanation I know that works fairly well for native speakers is:
1. It only applies to the consonant k, s, t or h in the second component following a kanji whose reading ends with m, n, or ng in classical Chinese,
2. It only applies to a noun that sounds like one word rather than a compound of two words.
These rules are not bad to explain when a word gets voiced because of 連濁. But the first rule requires knowledge of phonology of classical Chinese, and the second rule requires a keen intuition native/near-native speakers have. So they're again not practical to learners...
You need to learn phonology of classical Japanese and how Japanese sounds have changed in order to describe the first rule more accurately. You also need deep knowledge of dialects. To rewrite the second rule in a more accurate, linguistic way, you need grammar, etymology, and frequency of each word in modern Japanese. I don't know all the rules either.
Since this is not about grammar per se, you can't expect simple grammatical explanations. If someone establishes a set of rules that explain 連濁 almost perfectly, it would require deep knowledge of the history of the Japanese language and etymology of Japanese words.
Last edited by magamo (2010 January 29, 10:42 pm)
