Lol dr. Jones, if I am a relatavist, then you are a generalizer.  Did you read my last post?  I said several times that there are instances in which certain keywords are better.  My point is some here are a bit too eager to throw out keywords like "rouse" and "fold" which I believe to have already been obsessively considered by Heisig.  The trick to finding a successful keyword is selecting one that is a marriage between the specific and vague.  It has to be specific enough to make sense and yet vague enough to blanket all, or at least, most of the primitive's possible meanings.  And yes, "bone fracture" is too specific, and "fold" is BETTER.  LOL

Last edited by Dragg (2008 March 20, 5:34 pm)

ファブリス wrote:

PS: that's not what I meant by "ambiguous keywords", btw. I was referring to keyword which point to different possible meanings, but only one of those meanings is related to the character (as in "wonder").

That's one of the reasons why this site has become an invaluable companion to RtK.  Even when I inteded to think of my own story for a specific kanji, I would start by looking through the shared "stories" to see if anyone had posted notes about such ambiguities.  That saved me in more than a few cases.  Sometimes I went with the "wrong" meaning for the sake of wordplay, but at least I was (for the most part) conscious of when I was doing that.

DrJones wrote:

If the keywords are wrong, you fail at any attempt to predict new words from the keywords you know, and you have to relearn its meaning.

Of course if you try to predict new compounds based on knowing a few existing Japanese words that use the same kanji you'll probably fail too, or at the very least not get the nuances right...

For the plain and straightforward words, my suggestion would be to learn the Japanese *first* (ie as spoken language, and if written, kana only); that way you attach the right concepts to おきる and  おこす and so on. Then you can do RTK later and you have a handy memory key for remembering how to write these words you already know.  (But then I'm biased, because that's what I'm doing :-)) Learning all the kanji up front seems backwards to me, even if it is what Heisig recommends; and it tends to lead you down the path of thinking that kanji have a 'meaning' in and of themselves, which I disagree with.

pm215 wrote:

Learning all the kanji up front seems backwards to me, even if it is what Heisig recommends; and it tends to lead you down the path of thinking that kanji have a 'meaning' in and of themselves, which I disagree with.

Wait... what? Kanji do have a meaning in and of themselves. They are all either pictographs or ideographs, meaning that they either represent what they mean with a picture or represent the idea. So... what do you mean by kanji don't have a meaning in and of themselves?


I mean half of the reason kanji are so interesting is because, when one sees a kanji in a word, they can immediately draw assumptions. This leads to later be able to guess meanings and even readings (though that's easier in Chinese) just because of the kanji.

pm215 wrote:

I mean that (as with any language) the spoken language is primary, and Japanese deals in words, not in individual kanji. So there is 注ぐ (to pour), and 注意 (caution), 注射 (injection), 注文 (order), and so on. You can assign a 'meaning' to the individual kanji based on looking at common themes in the meanings of the words they turn up in, but that's a secondary thing and not actually very important.

I would say that in the case of words formed from "on" readings, the written language is primary.

Going back 1500 years. . .
The Japanese spoken language exists, but without writing.

Then. . .
Chinese symbols (kanji) are imported and assigned to write Japanese words. The Chinese symbol meaning is matched to the Japanese pronunciation. For example "dog" is written 犬 and pronounced "inu". This is the "kun" reading.

A little later. . .
Many new words are formed in Japanese by combining kanji. In this case it is more efficient to use the short (mostly one syllable) Chinese pronunciations of the kanji. These are the "on" readings.

The new words formed from "on" readings would be chosen according to the meaning of the kanji as the pronunciations by themselves had no meaning.

Suppose we want to make a word for "shipyard".

This is a place where ships are created.
This is a 所 where 船 are 造.
造船所 （ぞうせんじょ）

So the word is formed by choosing kanji with appropriate meanings, combining them, and then using the pronunciation that results.

The examples you give above: 注意 (caution), 注射 (injection), 注文 (order), were probably formed in the same way, i.e. by choosing kanji with suitable meanings and then using the pronunciation that results, not the other way round.

DrJones wrote:

The ambiguity on keywords is easily solved by just adding a context after the keyword. For example, for "Trunk" I have "Trunk (body)" and the problems went away.

True. I had given up doing that mostly because I would be tampering with Heisig's keywords and I expect many users will complain, or say that my choice of extra information is wrong, and that something else would be better ;-)  That wink is not meant as a critic by the way but just to say the same will happen to me. So I would have to put like a "original keyword" option and I'd rather keep it like it is in that case.

laner36 wrote:

That sounds really difficult to implement but what about this: have a little box above all the stories that had alternate meanings from say KANJIDIC.  For example for 折: "fold; break; fracture; bend; yield; submit".  What do you think?

That would be part of a discussion area (like an extra tab) on each kanji page where people can discuss anything related to the characer such as Transtic's excellent post on "tassel". It's in my to-do box... can't work on it until I finish refactoring the site's code ..

I was under the impression that Chinese, like Japanese, has a lot of words which are kanji compounds. So I'm not convinced that the "meaning" of a character is necessarily a well defined thing there either. (But I don't know much about Chinese at all, so I'm open to correction.) And of course the 'meaning' of some characters has drifted over time, as the way they have been used has changed.

Katsuo wrote:

learning the kanji first and giving each one a name is a great help for subsequently learning vocabulary.

Mmm; I just wouldn't get too hung up on whether that 'name' is really the 'meaning' of the kanji or not, I suppose.

I knew I would find it eventually!

Heisig does, in fact, address this problem of so called "wrong" keywords. All you have to do is read the introduction to Lesson 31, which I will not type out in length here. (Pages 307-308 4th edition).

Like it could be said of the mathematician Ramanujam (http://en.wikipedia.org/wiki/Srinivasa_Ramanujan) - that he had a deep personal friendship with every integer! I wish the same should be true of every member of this forum and the kanji.

A good idea is to use this book as a method of -introducing- oneself to the kanji.
There's lots of 'lots' more behind the surface. Of course. Just like people.
I'm using this book to breeze through the entire set, before I can talk with the gakusho kokugo/kanji jitens. So ... this is just like, maybe ... saying 'Hi!'

I will use everything at my disposal to forge deep friendships with every single one of them in the time to come. So, what face of a kanji is revealed at first sight is not -terribly- important for me. I -know- it is not the only face ... maybe not even 1/100ths of the true import of the character.

my 2 JPY

It's debates like this that REALLY make me want to learn more Kanji, because this is really fascinating stuff but it's not really possible to understand the debate or contribute to it fruitfully unless one can actually read Japanese or Chinese, which I'm not really there yet.

From my very limited perspective, though, I'll propose this. So far, the most useful definition I've heard is that a Kanji is a "nexus of associations".

From this point of view, it's misleading to think each Kanji has "a" meaning, or even several "meanings." What happens is that when Japanese folk see a kanji, it triggers a cascade of associations in their minds. The precise meaning is only made determinate by context: without it, kanji often don't mean anything clear at all.

So, for example , when a Japanese reader sees 本, she's reminded of a number of loosely associated concepts at once: truth, origin, book, main, real. In isolation, "本" doesn't *mean* any one of these things, but it does *suggest* all of them.

It's easy to see how the meanings of 本 are metonymically related, how they can slide into one another and form a kind of family of meanings brought together by associations that are lateral, almost poetic.

This, I think, is why Japanese people talk about given kanji having a kind of "taste", and about 言霊: the "spirit of words."

For someone on Heisig frame 675, it's a sobering realization: here I am struggling like hell to memorize the tip of the semantic iceberg. Cuz, of course, it's insane to think a single keyword can even begin to encompass the psychoanalytic free-play of laterally associated meanings so many kanji seem to get up to in their free time. But hey, you gotta start somewhere.

I'm not a semiotician or a linguist, so this is probably not a very precise definition, from an academic point of view...but it works for me!

Last edited by AndamanIslander (2008 April 09, 3:55 am)

I'm no expert either, but what you say sounds quite reasonable to me.

Some kanji do have a narrow enough range of associations that you can say they have a specific meaning, like 目 "eye" AFAIK basically means eye, 子 "child" basically means child.

However, even the most basic kanji don't have a single meaning. 日 "day" can also mean sun (and even Japan), 木 "tree" can also mean wood.


With regard to radical+phonetic 漢字 (the majority of them, it would seem), the radical part is the semantic component while the phonetic is (or, in many cases, originally was) indicative of the pronunciation.

However, it seems apparent that the choice of phonetic was not entirely arbitrary. Having worked through RTK1, many phonetics do appear to have some relationship to the meaning of the entire kanji, suggesting that where possible, it was based on a word that sounded similar but also had some semantic relationship that could aid recall.

For example, 枯 "wither" is made of "tree" and "old". The pronunciation /ko/ of 枯 is the same as that of 古, hence its use as phonetic. However, the combination of tree+old does suggest the meaning "wither". 恩 "grace+gratitude" is made of "cause" and "heart". The pronunciation of 恩 /on/ is similar to 因 /in/, but the use of "cause" suggests that the grace shown is the cause of the response of gratitude.


I think we've strayed off topic, but it's been an interesting thread to read.