Re: 90%

According to wikipedia:
"Approximately 60% of the words contained in a modern Japanese dictionary is estimated to consist of kango". Kango being words taken from chinese or constructed on chinese roots. The same article also mentions ateji words like 世話 and 面倒 which use onyomi readings but are not chinese, and Japanese created kanji such as 働 which is assigned an onyomi when used in compound words like 労働. I'm not sure if these are included in the 60%. I guess I was exaggerating when I said 90%, but I still think that for all intensive purposes when word frequency is taken into account, the number of onyomi words actually used is much higher, particularly in writing (not that 60% is low anyway). Take into account there are perhaps 500,000 japanese words in existence (I read that somewhere), including archaic, obsolete or very obscure words. I never meant to say that 90% of the words in any given sentence are onyomi words. Of course not, the core of the language (verbs, particles etc) that forms most of the grammatical base are mostly wago words. And as we know something like 90% (figure pulled out arse) of the time we we just reuse the same small base of core vocabulary (maybe 2000 words or so). It's in the remaining 10%, the content words, that the the rest of our passive and active vocabulary comes into play, and it's here that I think the balance shifts significantly to kango and increasing English derived katakana words, particularly when moving away from everyday topics. Open up a newspaper, all the international organizations, economic forums and so on are all big onyomi compounds, 世界貿易機関、国連児童基金、北大西洋条約機構、etc. And it makes sense too. Whenever a new word needs to be coined in whatever field, people draw upon the commonly used readings rather that the obscure uncommon ones. Being able to read something something like 最高経営責任者 or 停止衛生 after just 3 months of study is a pretty good thing if you ask me.

Onyomi are also frequently used in names, particularly boys' first names, しょうた、りゅういち、たろう、しょうへい but also occasionally in girls names and last names. They are also frequently used in place names　東京、京都、阿蘇 etc.

nadiatims Wrote:the number of onyomi words actually used is much higher, particularly in writing (not that 60% is low anyway).

Care to elaborate?

Asriel Wrote:Heck, nukemarine put together a spreadsheet just to show those common

Link to this?

I was always under the impression that kango was kind of like French/Greek/Latin-derived words in English: numerous, but not as heavily used in usual conversation as the Anglo-Saxon/Scandinavian words that are the core of the English language.

Anyway, I have some actual data that speaks to the issue.

From "Principles of Japanese Discourse," Maynard at p.15:

Quote:"Casual speech, for instance, shows restricted use of kango while an academic thesis uses kango more abundantly. According to the National Language Research Institute study quoted in Yutaka Miyaji (1985:99), the proportional distribution of different kinds of vocabulary in written Japanese (magazines) and in spoken Japanese show the following results.

"Yamatokotoba and kango appear in approximately 54% and 41% of the total vocabulary, respectively, in magazines, while in spoken Japanese Yamatokotoba is more dominant (occurring approximately 72% of the total) than kango, which appears in only 24% of the total vocabulary. (The remaining 4% falls into Western words and mixed vocabulary.)"

I'd wager that since the study was done in 1985, that 4% has ballooned to ~10%, and I'm being conservative there, but I would *not* wager that there has been a kango revolution. While the increase in use of loan words may change, I doubt that traditional language patters re: the choice of wago over kango, or vice-versa depending on the situation, would change too much.

By the way, this is a fascinating book for all of you word nerd types. (I count myself among them.)

brianobush Wrote:

Asriel Wrote:Heck, nukemarine put together a spreadsheet just to show those common

Link to this?

I wish I could find it again. I think it had a list of the kanji with the readings listed, and the uncommon ones with a * next to it.

What I do have, however, is the common most common readings with all the kanji of that reading included, so here's that

KanjiDevourer, in the post below me, found a spreadsheet which has all the information I had, plus more!! So definitely check that one out.

It seems like the one I had was more compact, but then again, maybe that was modifications that I made. Still, adding the main onyomi into your RtK stories in one way or another should probably be a good idea. At least, it wouldn't be detrimental, I can't imagine.

This link https://spreadsheets.google.com/ccc?key=...l=en#gid=0 I got from Katsuo from this thread http://forum.koohii.com/showthread.php?tid=6294 on Integrating the readings into the RTK story, maybe that's what you're looking for.

I agree with the point the author of that youtube upload makes, but I agree his advice is not really meant for the well informed members of this forum. And I will try to put in some onyomi into my stories - even if it won't always work, it might help a bit. Now that I think of it, maybe it could make making up a story easier, since the onyomi could add new persona and make it more lively..

In terms of learning readings, he's spot on. Learning them in words is the fastest method, killing two birds with one stone and giving you a decent grasp on which readings to select when encountering a new word.

Now let's learn some KEENJI!

Yamatokotoba 72% in conversation vs 24% for kango.
Yamatokotoba 54% in magazines vs 41% for kango.

Until I saw those stats, I thought on-yomi first was more important, but now I'd think it'd make more sense to learn kun-yomi and its associated vocabulary first, to be honest.

Otherwise, it's like learning patriarch before learning dad, if you get my drift.

It's not an either-or thing; some kanji have more common on-yomi, and some kanji have more common kun-yomi. Some kun-yomi on the Jouyou list are extremely rare. I don't think it makes sense to focus exclusively on either since you need both to read almost anything.

rich_f Wrote:Yamatokotoba 72% in conversation vs 24% for kango.
Yamatokotoba 54% in magazines vs 41% for kango.

Until I saw those stats, I thought on-yomi first was more important, but now I'd think it'd make more sense to learn kun-yomi and its associated vocabulary first, to be honest.

Otherwise, it's like learning patriarch before learning dad, if you get my drift.

You don't read kanji when you're speaking.
Magazines -- well, you got me on that one.
All I can say is that instead of learning kunyomi, just learn the words themselves! No point in learning 楽 as たの without learning 楽しい.

I would quite often get made fun of by my girlfriend for using words like 完売 instead of 売り切れ because it made me sound like too much of a 北大生.

harhol Wrote:This 90% figure seems wildly speculative. Are there any valid sources which confirm it?

I don't think that learning the most common reading of a kanji by incorporating it into your story while doing RTK1 would ever be detrimental. However, it's the kind of thing that you only really think about after you've finished RTK1. If someone had suggested I incorporate onyomi readings when I first did RTK1, I'd have demanded to know what a yomi was and why I needed to be on it.

http://forum.koohii.com/showthread.php?tid=3216&page=1

This thread talked about the scan of Japanese wikipedia (25,000 pages) and the kanji and compound occurences. It took 874 kanji to reach 90%, while 2000 kanji got you up to 98.8%. Now, it might take a Mecab integration with the scan to get a decent Onyomi coverage chart but something tells me we'll still get near that 90%. Only thing left is find out how many words it takes to get 90% coverage (Again, a mecab integration with the scan might be able to do that).

As for integrating onyomi, I did it after RTK and it seemed to work. Reason being I could group the kanji together and create connected stories with the onyomi. Doing during RTK may not be useful as the onyomi will be thrown at you randomly, hence using the Movie Method is better as it says to learn all the primitives first then learn kanji grouped by onyomi. Problem with that is you're pretty much stuck making up ALL your stories as I don't know any comprehensive resource for the Movie Method with stories.

yudantaiteki Wrote:It's not an either-or thing; some kanji have more common on-yomi, and some kanji have more common kun-yomi. Some kun-yomi on the Jouyou list are extremely rare. I don't think it makes sense to focus exclusively on either since you need both to read almost anything.

Exactly. I have no idea why you would focus on 1 kind of reading or the other when it is pretty easy to learn the one that is most common. If you learn the readings within vocab this will happen naturally.

Another 90% for those who take comfort in numbers:
In Katsuo’s kanji list 90% of the Joyo kanji have only one common on-yomi.

That doesn’t mean it is easy to learn these readings, nor does it mean you can read 90% of a text once you mastered them. But it means, you can learn them (with whatever method) and it would be helpful.

Although one thing to remember is that saying the *result* is helpful is not the same thing as endorsing a particular method. To use an extreme example, knowing every word in the dictionary would be very helpful, but that doesn't mean it's a good idea to sit down and start studying the dictionary from page 1. In the same way, of course it's helpful to know the on-yomi of many kanji, but that doesn't mean it's a good idea to try to cram them out of context.

Two years ago I saw the don't learn kanji guy give a 45 minute speech (all in Japanese) on preschool English education. I'm 90% sure that his crazy accent on youtube is just his marketing gimmick. He's the creator of Genki English and he puts a whole lot of energy into... being genki ...and marketing that persona (which made him a ton of money apparently).

He also speaks Shikoku dialects fluently which really impressed the audience.

I don't bother to learn either/or with readings. When I learn readings for a kanji, I learn the major on/kun readings at once as vocabulary, in context. (That is, I'm a KO/KIC zombie, and it has served me well.)

What I *was* pointing out was the fallacy behind the premise of obsessing over onyomi to the exclusion of kunyomi. The numbers suggest to me that coming at the language with -0- knowledge of any words, and just finishing RTK, it's more profitable to learn kunyomi first, and only onyomi when there are no kunyomi for a kanji.

Considering that you're learning Japanese, which is not just a written, but also a *spoken* language, forgetting that almost 3/4 of what you'll want to say is Yamatokotoba-based makes obsessing over onyomi to the exclusion of kunyomi even sillier.

(Think about verbs for a second. How many non-する verbs are kango? Most are yamatokotoba/wago. Life without verbs would _____ .)

I'm *not* saying don't study onyomi. What I am saying is if you're going to do something as reckless as only study one reading for each kanji, then if you want to be able to communicate in a basic fashion, then yes, I think emphasizing kunyomi over onyomi is more useful up-front and more beneficial, *if* you were going to just learn one reading like that.

...which I would not recommend in the first place.

Ideally, I would grab a non-existent frequency dictionary as well, but sadly, a properly researched one does not exist. (Yes, there's a file of some newspaper frequency stuff, but that only measures the frequency of words you see in a newspaper, not the real language you need to know.)

@yudantaiteki
Why should anyone cram on-yomi without context?

In this thread alone you find two good alternatives:
- together with vocab (also used in Heisig II)
- integrate the reading in the story (e.g. Movie Method)

And you can always limit yourself to the most frequent kanji or the first groups of Heisig II - no obligation to do all the Joyo or all the dictionary. But don’t forget the kun-yomi …

Can you really study kunyomi out of context like we're talking about with Onyomi and Kanji? Most the lists I see give you the kunyomi with the okurigana attached, meaning you basically have a vocabulary word. Those same lists however only give the onyomi, with some lists giving sample words though.

Personally, the best systematic kunyomi learning that I've experienced is the Cangy sorted vocab lists via 2k1KO order. Even those give some context such as an English definition at the minimum though usually with an example sentence.

I've found in the beginning learning it from context is the best way, then vocab later on(once you've reached that intermediate stage for kanji). A combination of vocab+context+a lot of reading usual fixes the problems for kanji readings quite naturally.

someone in this thread Wrote:From "Principles of Japanese Discourse," Maynard at p.15:

"Casual speech, for instance, shows restricted use of kango while an academic thesis uses kango more abundantly. According to the National Language Research Institute study quoted in Yutaka Miyaji (1985:99), the proportional distribution of different kinds of vocabulary in written Japanese (magazines) and in spoken Japanese show the following results.

"Yamatokotoba and kango appear in approximately 54% and 41% of the total vocabulary, respectively, in magazines, while in spoken Japanese Yamatokotoba is more dominant (occurring approximately 72% of the total) than kango, which appears in only 24% of the total vocabulary. (The remaining 4% falls into Western words and mixed vocabulary.)"

Yamatokotoba occur more frequently but I would argue they draw from a smaller sample of commonly used words, the grammatical base (verbs etc) of the language as well as words required in everyday conversation. But I think while kango make up a smaller percentage of the words appearing in any one sentence, they are drawn from a larger pool of words. Some 60% of the total japanese vocabulary is thought to be kango according to the wikipedia article I posted earlier. Let's suppose the average educated Japanese adult has passive knowledge of some 50,000 words. Now if only 20,000 of those words (being conservative) can be read using common onyomi or derivatives, that is still a huge chunk of vocabulary that can become immediately more accessable very early in the Japanese learning game for almost no extra work. Multiple members of this forum have reported success learning common onyomi during or directly after RTK using methods such as:
- Sound primitive method: just add an extra primitive to each story while doing RTK.(which is what I did)
- Movie method
- learn onyomi via vocab (heisig II)

I understand very well that you need to learn kunyomi as well, but this seems less easy to do in a systematic and therefore highly efficient way for a number of reasons: there are more of them, they are harder to make mnemonics for, they tend to appear with okurigana etc. If I could go back in time, I would probably go through heisig 2 and learn kunyomi by aggressively flashcarding the vocab, with the aim of being able to read a decent chunk of the words in the language before the aggressive reading stage. Now I can read most things anyway I don't need to go back and do this, but I expect it would have sped up the process.

Womacks23 Wrote:Two years ago I saw the don't learn kanji guy give a 45 minute speech (all in Japanese) on preschool English education. I'm 90% sure that his crazy accent on youtube is just his marketing gimmick. He's the creator of Genki English and he puts a whole lot of energy into... being genki ...and marketing that persona (which made him a ton of money apparently).

He also speaks Shikoku dialects fluently which really impressed the audience.

Perhaps he is very fluent, but his accent is well audible. I am not very far into Japanese yet, but his pronunciation just isn't good, even after so many years. His "san" is.. well, like "keenji". Not that I despise him for it or anything, but I hope to be able to speak Japanese with as least accent as possible and also with the right stress-timing. And I think that can be done with a little effort towards listening and speaking.

He's a preschool teacher trying to sell materials for preschool children, so it wouldn't really do his cause much good if he dressed in a suit and went around talking like a dour salaryman. I'm sure he could do a serious, non-exaggerated accent if he wanted to.

Go watch some of his other videos. His Japanese is not good. That's fine, he has his priorities. But his advice is pretty much tantamount to "I didn't study seriously and I can get by kind of O.K, so you shouldn't either". After 13 years, I'd expect much better especially from someone marketing themselves as some kind language education guru.

Eh, there's too much sniping about accents imo. Khatz, those business people and now this guy... everyone has a terrible accent apparently. All that really matters is that you can communicate and that you don't sound like a bumbling idiot. This guy is confidently presenting to rooms of adults (not easy in itself) and has them in the palm of his hand, making them laugh, etc. He's clearly capable of more than just getting by kind of okay.

Virtually no one speaks a non-native language without a distinguishable foreign accent. (The only people I've heard do so with English were Scandinavians who had been living here for a decade or more, and even then English was effectively a native language for them.) I see no reason why this wouldn't hold for non-native Japanese speakers, but of course you're welcome to upload a recording of yourself if you want to show us how it should be done.

harhol Wrote:Eh, there's too much sniping about accents imo. Khatz, those business people and now this guy... everyone has a terrible accent apparently. All that really matters is that you can communicate and that you don't sound like a bumbling idiot. This guy is confidently presenting to rooms of adults (not easy in itself) and has them in the palm of his hand, making them laugh, etc. He's clearly capable of more than just getting by kind of okay.

Virtually no one speaks a non-native language without a distinguishable foreign accent. (The only people I've heard do so with English were Scandinavians who had been living here for a decade or more, and even then English was effectively a native language for them.) I see no reason why this wouldn't hold for non-native Japanese speakers, but of course you're welcome to upload a recording of yourself if you want to show us how it should be done.

agreed! Accent is the last thing people should worry about, the best we can do is listen/match from what we hear/read in japanese. The rest time will either fix it up or not. But accent has nothing to do with fluency in a language. Fluency is much more important. I know people complain about big accents that some people have but that's something we cannot always fix up.