Be Seen and Not Heard
I mentioned last week that the Japanese language is based on written expression, while English is based on verbal expression. Since both languages are obviously both spoken and written, I suspect I'll have to explain myself a bit. After living more than half my life in Japan, though, I have come to feel that this is a crucial difference not only between two cultures, but between two different ways of looking at the world. And the worldview, obviously, forms a significant but not always clearly visible part of any literature.
I suspect most people reading this blog will be familiar with English, so let me just point out that we use maybe a hundred characters in writing. Japanese has an enormous number of characters (say, fifty katakana, fifty hiragana, all the characters used in English, and several thousand kanji... although the total number of kanji actually in the Japanese dictionary is probably ten times that). Even though Japanese has far, far more characters than English, the English language is capable of producing many more sounds. There is considerable difference of opinion about just how many sounds (phonemes) English actually uses, but it looks about a dozen vowels and twice that many consonants. Japanese has fifty kana characters, but it is a matrix of vowels and consonants, providing a very specific selection.
There are only five vowels listed in the kana table, and while there are unquestionably a few more phonemes, the Japanese cultural desire for uniformity (not to mention nationwide TV networks) is making sure that most people use a pretty small set of sounds in speech. In spite of the more characters available, the number of possible sounds is far smaller than that of English. (Which also explains, in part, why many Japanese have such trouble with English prouniciation.)
So, English has a few letters and lots of possible sounds, and Japanese has an enormous quantity of characters but is a bit, shall we say, phoneme deprived. This causes problems, even for the Japanese.
To give an example, when I was studying Japanese, Professor Katō gave us a sheet of paper written in kana, and asked us to figure out the right kanji. It was a short conversation between two people, and they were talking about 'nōgyō' (のうぎょう). He left the room and we talked about it, recognizing that nōgyō could be written 農業 and would mean agriculture. With dictionaries in hand we whizzed through the translation, quickly finding all the kanji to fill out their conversation about what they grew and how. Pretty easy, actually, although there were of course of lot of words we didn't know.
Problem is, when the prof came back and gave us the answer, it wasn't talking about agriculture at all. It was talking about the Noh theater (能業), and all the words we looked up were wrong. They were homophones in Japanese (and, if your write them in kana, homonyms). If you're having trouble following all this, think about the word 'right.' When you hear that word and are asked to spell it, what do you write? There are four immediate possibilities: right, write, rite and wright. The Japanese kana corresponds to phonetic symbols in the dictionary, and the correct characters in Japanese (a combination of kana and kanji) corresponds to English spelling.
If you leave Japanese only in the phonetic symbols (kana), or speak it, you end up with a massive vocabulary (remember, thousands of kanji) being expressed in only a very few characters (or phonemes). And that's where the problem comes in.
There are a huge number of homophones in Japanese, especially when using a pair of kanji with the on-yomi (音読み) pronunciation. Depending on the specific kanji used, the meaning can be completely (or subtly) different, but it's impossible to tell by listening to it.
A temple in the city here is named 安国寺, and I was curious if it were pronounced Yasukunidera or Ankokuji, so I asked one of the monks. He replied, quite seriously, that it was both. Modern administrative Japan demands a 'proper' reading, and so Ankokuji is in the books, but the monks, he said, use both and have no doubt in their minds that they are equally valid.
So, first of all a Japanese author can blur the meaning of a word by writing it in kana instead of kanji. This can be very useful not only in making it difficult to pin down the precise meaning (which Japanese authors dearly love to prevent us from doing), but also because writing in hiragana alone gives the text itself a softer touch than mixed hiragana-kanji text, or the katakana-kanji text used to indicate robots speaking.
In everyday Japanese conversation, there are commonly cases where a person will say something, and the other person won't understand it immediately because of the homophone, and ask which kanji is meant. In written Japanese this problem is much less common, because the actual kanji are visible and show clearly what is meant.
Unlike English, where the spoken word is almost always understood clearly, in Japanese it is often essential to see the written characters before the meaning can be confirmed.
As if that isn't enough, though, Japanese writers go even further. They can take a Japanese kanji or kanji pair (called a kango; 漢語) and put the pronunciation next to it in tiny characters called ruby text. Normally ruby characters are used to show the pronunciation of unusual characters (specialized jargon, for example), or to tell the reader how to pronounce a character's name. In literature, however, they can also be used to add a totally different, additional meaning to the kanji... the example on the wiki page shows the characters for Tokyo, with the same pronunciation written above in ruby characters. It would also be possible for the author to write, say, ビッグみかん (Big Mikan) instead of とうきょう or トウキョウ.
This means that a translator has to take two meanings into account: Tokyo, and the nickname Big Mikan (as in New York, the Big Apple). If it was something said by the one of the characters, it usually indicates that the character actually said "Big Mikan," and the reader is getting an explanation that the author means Tokyo. Unfortunately, though, not always... sometimes it is impossible to tell which meaning is supposed to be primary, and whi secondary, and the translator has to find a way to express both meanings in the translation.
Text that plays with multiple meanings, possible in a wide variety of ways, makes the translator's job so much harder. It also makes it a heck of a lot more rewarding when you come up with a really good way to capture the full meaning of the source in English!