ec2-54-163-210-170.compute-1.amazonaws.com | ToothyWiki | RecentChanges | Login | Advent calendar | Webcomic
The bulk of the Japanese written language is formed of several thousand 'pictogram' characters which share a common root with the Chinese equivalents (note that common root means similar meanings but definitely not similar pronounciation).
Each Kanji has multiple meanings, often with a different pronunciation for each meaning. The meaning (and pronunciation) are not indicated in any way by the Kanji unless you are lucky enough to have Furigana to help (which usually only happens in children's books).
In combination, Kanji have yet more meanings (and pronunciations) often completely disconnected from the meanings of the individual Kanji.
"Abandon all hope, ye who enter here."
Kanji are normally keyed in by typing out the phonetic pronunciation and selecting the appropriate kanji from a list. Some dictionaries also allow you to select graphical elements the kanji contains until the set of possibilities is narrow enough for you to pick it out - this is useful if you don't know the phonetic pronunciation.
- Over-reliance on the computer writing the kanji for you naturally leads to a situation of being unable to remember how to write the kanji if you don't have a handy list popping up. The Japanese have an entertaining term for this which AlexChurchill discovered today: ワー プロ馬鹿 "WAAPURO-baka", literally "word processing idiocy". And the Kanji for "baka" individually mean "deer" and "horse"...
Paper-based dictionaries also need a way for you to be able to look up kanji you can't pronounce. There are several methods for doing this. The traditional method of identifying which of the 200ish radicals it contains has always been confusing for kanji containing several radicals, and become even more so as characters change over time and no longer actually contain the radical under which they're still indexed. Another one (the four-corner method?) involves classifying the strokes in the corners of the kanji to generate an ID number. todo: list ways, name them
[Jeffrey's Kanji Lookup] provides a number of ways to look up kanji. The [SKIP code] is very good for (non-Japanese) beginners, not requiring you to know radicals. You do need a very basic knowledge of stroke count - enough to know that a square is three strokes not four, basically. And a little trial and error - but rather less than most other Kanji index methods.
- Uhh - could you list any other 'obvious?' or 'trivial' bits of knowledge like that please? --Vitenka
- BwaHaHaHa. I never claimed it was obvious or trivial - I chose that example to indicate that I indeed meant "very basic but non-zero". Well, I learnt my entire knowledge of Kanji stroke count from pages 1-8 of Let's Learn Kanji (ISBN 4770020686 ) (which I recommend, incidentally). Fortunately, Jeffrey's lookup linked above accepts certain wildcards. So you can do a SKIP search for 1--4 or 2-12-* . --AC
The Sony Clie TJ25 (as well as the more expensive models) has recognition for handwritten kanji (stroke order doesn't matter). Coupled with a [free dictionary], it becomes a rather powerful tool. It costs a little over 90 quid plus postage from [amazon.co.jp]. (MoonShadow knows this because SunKitten just got a piece of targeted Spam from a guy who's offering to import them at 150 quid each and he googled for the model name).
A request for those who can get their computers to write Kanji:
I need a copy of the Ai (love) Kanji which is roughly 500 pixels high. Net searches have turned up copies up to about 100 pixels in size but I really don't want this to look pixellated. The only way I can think of doing this properly would be to get a true type font, write it out at the appropriate size and then screen capture it. This would, unfortunately, require more knowledge than I have about how you actually type kanji.
- The [dictionary] linked above gives you codes in most of the encoding systems when you look up a kanji. The easiest way in this case would be to type "ai" into the readings box or "love" into the English tags box under "cull search"; we get loads of results back, which - annoyingly - aren't sorted by any useful criteria; the one we want is actually a little way down in the results list. We get:
Encodings: JIS 3026, EUC B0A6, Kuten 1606, Shift-JIS 88A4, Unicode 611B
- (I *think* that's the right one; could also be 慕 for all I know - I'm only going by the English tags here). We can then grab the unicode value and wrap it in an HTML entity:
- I can't actually read that myself at work, since I don't have a kanji font, but you should be able to. - MoonShadow
Could someone send me such an image? Or, as an alternative, simply send me the html code for the ai kanji as I could then, theoretically, do the above myself?
- If you can send me the largest one you have now, I can spend a little time on it in Illustrator, and you can have a bitmap pretty much any size you like back. - Tsunami
Are you doing a picture of [Gaara] or something? I'll help if someone can tell me how to do a screen grab in Windows :-(. Ah, figured that out. How shall I send it to you? My e-mail address is (on a previous revision) --Mjb67
- Thanks for that. Did you get my mail? --K
- Ah, yes you did. Thanks again --K
See Also: Hiragana, Katakana, Furigana