There are a bunch of corpora out there ripe for the picking. (Analyzing. Whatever.)
http://forum.koohii.com/showthread.php?tid=2509

Frequency-based learning has some usefulness early on, in that it makes some material more accessible, but in the end, you gotta get em all.

I just checked how many unique kanji there are in two lists, going to 80% coverage.

First list was a kanji-only word list, so it had 1537 unique kanjis.

Then I checked the newspaper word list, and I get 949 unique kanjis.

It is higher than the 555 for 80% kanji coverage, which I expected. I'll need to check against other corpuses though.

To OP, I posted a similar thoughts on the KO2001 threads and others. However, I argued that one should take frequency lists one step further:

Kanji frequency lists should be ordered for learning in an intuitive way. This can be stepped radical order ala Heisig or onyomi ala Movie Method. Most using RTK will not argue with me on this one.

Word frequency lists should also be organized in an intuitive way. Those that have done KO2001 via the books or iKnow lists might agree with me here.

The benefit to this systematic learning is it gets you into native material quicker. Using frequency lists just allows more "bang for your buck". No, you're not going to be 80% literate, but you looking up less meanings for native sources.

For the above, I would even throw on grammar concepts (say JLPT level 4, 3 and 2 via Kanzen Master for example). These example sentences I would also encourage being ordered in an intuitive way, such as Tae Kim attempts to do with his page.

Long point short: Frequency based lists (Kanji, Vocabulary, Grammar) are best if organized further into an easier to learn order.

Now, on to the next part: How to spread this out? The problem with systematic learning is the higher you want to get, the more time you spend for each additional percentage. This is true with Kanji, Vocabulary and Grammar.

My personal opinion: Studies of each should be broken up into 100 study hour groups. By personal experience, 100 study hours equates to 1000 kanji or 1000 new vocabulary words or 500 example grammar sentences (3-4 sentences per grammar point/sub point).

In addition, these are not separate. Going through grammar sentences you may (ok will) find kanji not in your first 1000 and have vocabulary not in you vocabulary list. Like one would do when mining sentences, add in new kanji and words and grammar as you come across them in these formal studies.

So, doing 100 hours of each gets you upto JLPT3 level (300 study hours total). Doing an additional 100 each should net you upto JLPT 2 level or close to it (600 study hours) with the exception of vocabulary.

So what does that mean now? Well, out of about 700 sentences from Zettai Kareshi ep 01, I used 300 of them. In those there were 10 new kanji (2500 and above in RTK 3) and 110 new words (iKnow Core 6000 words). On top of that were about 20 words not in the Core list. So by having a base line of RTK and Core 2000 vocabulary I'm looking up only a few new words.

Of course, ZK is a romance comedy based set in a Gourmet Sweets company. When I go through Rookies, it'll have words dealing with baseball, thugs and school. Likewise Monster Parent will have more legal terms, etc, etc. So these will not

I don't think one has to get them all. But one will need larger passive vocabulary knowledge. Now, I could take the time (400 more hours) to systematically go through Core 6000. I won't as when I attempted it it bored me to tears the first week. It seems more fun now to activate Core 6000 words as they come along via sentence mining. However, I would not be at that level had I not invested the prior 600 hours systematically (and yes, I did have fun doing that).

Long point short: Study systematically for the first 80 to 90 percent of frequency based material.

Aijin Wrote:To me it makes more sense to do it at the same time

I wonder if you know that this entire site was created for and is dedicated to the idea of learning the kanji writing and English meaning completely independently of any Japanese language influence. Just comments like this make me wonder if you stumbled across these forums by accident and became intrigued by the many avid and successful Japanese learners, without realizing most of them owed their success (and motivation) to this very concept.

mentat_kgs Wrote:@ahibba
I don't know your source but it looks far from reliable. If you cite from where you took it it might help.

First, ideogram is not a very good word for it.

First, where did you read the word "ideogram"?

Second, these are not my words, I quoted it from Heisig's introduction to Japanese In MangaLand:

http://www.nanzan-u.ac.jp/~heisig/pdf/Br...0Kanji.pdf

scuda Wrote:I wonder if there is a threshold of word coverage that gives just enough understanding to pick up all the new words entirely through context, without needing to rely on a dictionary. As a native english speaker, I picked up the vast majority of my vocabulary via reading many books as I grew up, and I never used the dictionary to learn words.

I've read a few papers on this topic, and it seems the consensus among researchers of second language acquisition is that you need about 95% coverage to guess the meanings of unfamiliar words well enough to develop your vocabulary through reading without relying heavily on dictionaries. The research is mainly about English as a foreign language, but I think it could give you a rough idea.

The caveat is that their research methods aren't sophisticated. The typical method is:

1. Take arbitrary text and leave out X% of words at random.
2. Ask native speakers to fill in the blanks and see how well they guessed the left out words.

Apparently native speakers have much deeper knowledge of the language, so they'd be much better at guessing than the average learner. 95% could be a little conservative.

There is another catch when it comes to frequency based learning, that is, they don't consider how well you know each word. For example, most of the Japanese high school students know the most frequent 1000~2000 English words. Some excellent students know 5000+ words. But probably 99% of them don't know what "What's up?" means. Of course they'd say they know "what," "is," and "up." They also know the contraction "what's." But still they fail to understand the very basic phrase. "I don't get it." is another *difficult* expression. I don't think more than 0.1% of Japanese high schoolers would understand what "You can't get away with it!" means. "We made it!" "They had pulled it off!" "I should get going." "The train pulled in the station." etc. are all difficult phrases. To put it simple, 70% Japanese word coverage would assume you understand the Japanese equivalent of "She's got knocked up. I kid you not." because all words are listed as "frequent."

Frequency lists of idioms and whanot would help a little, but still learners would get tripped up by very simple sentences like "That's the way it is." If you made a frequency based list of phrases that don't match up with literal translations, then the list would be unacceptably huge. I wouldn't be surprised if 95% coverage list of frequent anti-literal translation expressions was larger than your average dictionary for native speakers; native speakers don't need natural expressions in their dictionaries while those natural phrases are not natural to you at all.

Frequency lists can cover words pretty well, and coverage and stuff can make things look easier. But you should learn them deeper than that. I think lists are valuable to teachers, but it might be better for learners to focus more on how well they know each word than on how many words they know.

magamo Wrote:I don't think more than 0.1% of Japanese high schoolers would understand what "You can't get away with it!" means. "We made it!" "They had pulled it off!" "I should get going." "The train pulled in the station." etc. are all difficult phrases.

Neither me!

But I think if I read them in context, I'll understand most of them.

I think "We made it" = we were succeed?

"I don't get it" = I don't undestand?

But "You can't get away with it!", "They had pulled it off!", "I should get going." "The train pulled in the station." are very difficult to me.

You can't get away with it = you can't cheat/deceive them?

Your English seems very good. I wish I know who did you manage to reach this level?

Nukemarine Wrote:Word frequency lists should also be organized in an intuitive way. Those that have done KO2001 via the books or iKnow lists might agree with me here.

Ironically, when I did the KO2001 list on smart.fm, it wasn't organized intuitively. The list order goes like this:
one
one o'clock
one (thing)
1st of the month
two
february
two o'clock
two (things)
2nd of the month

I found that order extremely irritating and hard to learn, I thought it would have been better to separate those out and group them. I.E. one, two, three, four then one o'clock, two o'clock, three o'clock, etc..

Nukemarine Wrote:Long point short: Study systematically for the first 80 to 90 percent of frequency based material.

I agree, get to 80% or 90% then perhaps start studying more organically after that point, and just do sentence mining etc..

I question the point of studying grammar though. Unless you're planning on trying to output japanese right away, there doesn't seem to be much point in studying the grammar directly. I'd rather just avoid outputting japanese at all, and mine lots of sentences, and let my sense for grammar develop naturally.

BTW, to get to 80%, it seems like that would require 100 hours on kanji and 200 hours on words. Twice as much on words. But, it seems to me that the 1000 kanji would be covered in the first 50 hours, and then studying the words is started after that, while SRSing the kanji daily to retain them.

magamo Wrote:Frequency lists can cover words pretty well, and coverage and stuff can make things look easier. But you should learn them deeper than that. I think lists are valuable to teachers, but it might be better for learners to focus more on how well they know each word than on how many they know.

I think learning idioms from lists might be awkward, I'm inclined to think that it would be better to pick them up through lots of exposure to the language.

But, if 80% coverage of an idiom frequency list was relatively small, it could be worth it.

Not sure how you would train the depth of understanding of each word, or if you should. I would expect that the more words you know, the more you can understand of each word, because words are defined by context to each other, so if you know lots of words, and read lots of sentences to see their actual usage, then with enough exposure you'll understand each word deeper.

Unless there's a faster way to train for depth of understanding a word? Reading multiple definitions of a word from a dictionary sounds awkward, would it be worth it?

Although I agree with RTK before learning jukugo it should be noted that kanji are not ideographic. It is a pitfall that all beginners of Japanese/Chinese fall into, I know because I too fell into the same trap but you learn as you go.

When one sees 花、they say　はな. That implies a phonetic connection, not an ideographic one. The whole ideograph theory (at least pure ideographic) has been proven false by linguistic research. All languages are alike, *all* languages are all based on sound. It's just that you have been so immersed in your own mother tongue that when you go to marvelous orient, you think you are witnessing something exotic and unique, when in fact it is just the same shit, but with a different smell.

Ok the above is still arguable, and is argued by people today I'll give that, but now lets move on to the actual problem at hand.

The problem now is you have people saying that, because of this, you can learn all of Japanese using Kana only, (heck you can, but you'd be stupid to do so).
The bad thing about RTK though is that it makes users think they can instantly jump onto learning readings separate from actual words. If, and only if, there was a nice clean, 100% logical system of readings and words then perhaps this would be an ok method.

But a language is not something that is invented one night in an office with a bunch of academics. Rather, an intricate mix of vuglar, written, spoken, standardisation that was all piled up over time to form what we see today.

RTK keywords have been alright, but I think I would have survived without them, I still have to look up the ACTUAL meaning of a word when I come across it even when I know both the characters through RTK. The only thing RTK has done is given me an idea as to what it may mean given the context of the sentence etc, but once I learn the actual word from a dictionary the keywords become further useless to me and I build the meaning of the character closer around other contextual situations to which it is used in.

I do admit, learning etymologies for English words has helped me build my English vocabulary and has allowed me to guess words that are new to me. But when I read some scholarly article and they use some bullshit formalities which I have never heard of before, knowing bene- from benevolent or sense-, from sensitive etc has not helped at all with the new word, and I would always consult my handy Oxford English Dictionary.

Mind you I'm still at a 2nd year University level in Japanese, and I'm already saying this now. I noticed a lot of JLPT2/1 level uses here saying they found keywords life saving, so I'm still puzzled at that. But each to their own.

The main benefit of RTK for me was merely writing them with confidence and not forgetting. Once I get out of the way with keywords I'm jumping straight on to pure Japanese. I don't think the issue of frequency lists is a huge matter anymore anyway, RTK has shown us otherwise anyway, the people who have completed RTK and have gone from there know this very well. I am still half way in RTK and have been half way for about half a year already.

Edit sorry I wrote this at like 5am last night

welldone101 Wrote:

Aijin Wrote:To me it makes more sense to do it at the same time

I wonder if you know that this entire site was created for and is dedicated to the idea of learning the kanji writing and English meaning completely independently of any Japanese language influence. Just comments like this make me wonder if you stumbled across these forums by accident and became intrigued by the many avid and successful Japanese learners, without realizing most of them owed their success (and motivation) to this very concept.

I didn't stumble upon this site too randomly. Since I want to teach Japanese at American Universities I often quiz the students in the Japanese classes at my college about how they learn, etc. One of them mentioned this method, and when I tried looking it up online I found this forum before actually finding out about the method itself
So, I'm slowly learning about this learning method, bear with me

Quote:Unless there's a faster way to train for depth of understanding a word? Reading multiple definitions of a word from a dictionary sounds awkward, would it be worth it?

I don't think there's any shortcut to depth of understanding for words. It can take many years of exposure to truly grasp all the associations a word has, all the situations it's used in, what group of people mainly use that word, what other words sound similar to it and might cause a subconscious influence in a reader's mind, etc.

For example, someone learning Japanese as a second language might find that both 暴く　and 明かす both translate roughly to 'to reveal' in English and use them accordingly. It would take a lot of exposure to the usage of those two verbs for the person to realize their differences, that 暴く is sensationalist, always used by the media, and that it taints the significance and seriousness of what is 'being revealed' if you try to use it to explain the revealing of something meaningful.
Dictionaries both give 青白い and 淡い the same translation, but they don't explain that 淡い is a poetic term and has a very different feel and evocation of images than 青白い.
A lot of foreigners I've noticed have trouble with the differences between 空く (すく) and 空く (あく) since they have the same meaning and are written identical. Not many foreigners would know, if they're reading aloud, which they should pronounce 空く is, I imagine, as you'd practically have to live in Japan to know all the situations those two words are used in.

I wish there were dictionaries that explained all the subtleties to a word and all it's uses and associations rather than just it's translation. But, I don't think there are any that in-depth yet, so nothing can substitute for a good decade of exposure to the culture

ahibba Wrote:I wish I know who did you manage to reach this level?

Um, random native/near-native speakers? So far I've made about 130 posts on this forum, and most of the expressions in my posts are sort of direct quotes from my SRS. I can't remember all the sources of the sentences you mentioned, but I vaguely remember mining some of them:

pull it off: Gene (Outlaw Star)
I should get going: Tomoyo (Clannad)
the train pulled in: probably Konata (Lucky Star) but possibly someone from Eureka Seven.

The senses/actual words I mined might be a little different. They're all common phrases so it's hard to remember when and where I first heard them.

scuda Wrote:I think learning idioms from lists might be awkward, I'm inclined to think that it would be better to pick them up through lots of exposure to the language.

But, if 80% coverage of an idiom frequency list was relatively small, it could be worth it.

Not sure how you would train the depth of understanding of each word, or if you should. I would expect that the more words you know, the more you can understand of each word, because words are defined by context to each other, so if you know lots of words, and read lots of sentences to see their actual usage, then with enough exposure you'll understand each word deeper.

Unless there's a faster way to train for depth of understanding a word? Reading multiple definitions of a word from a dictionary sounds awkward, would it be worth it?

Idioms are just an aspect of language. There are a lot of things learners should work on such as collocation, nuance and intonation. Also, I guess it'd make you crazy to deepen your understanding of each word by memorizing a dictionary. Besides you need to know way more than a dictionary says. Look up a few simple words in a dictionary of your mother tongue and see how superficial their explanations are.

They're useful, but I don't think dictionaries would deepen your knowledge as much as exposure does. After all, reading a dictionary is part of your reading in the target language unless you're using a bilingual dictionary.

Edit: Reading a dictionary would be nice as long as you can enjoy it. I like reading books, and a dictionary is no exception. YMMV

Aijin Wrote:

Quote:Unless there's a faster way to train for depth of understanding a word? Reading multiple definitions of a word from a dictionary sounds awkward, would it be worth it?

I don't think there's any shortcut to depth of understanding for words. It can take many years of exposure to truly grasp all the associations a word has, all the situations it's used in, what group of people mainly use that word, what other words sound similar to it and might cause a subconscious influence in a reader's mind, etc.

For example, someone learning Japanese as a second language might find that both 暴く　and 明かす both translate roughly to 'to reveal' in English and use them accordingly. It would take a lot of exposure to the usage of those two verbs for the person to realize their differences, that 暴く is sensationalist, always used by the media, and that it taints the significance and seriousness of what is 'being revealed' if you try to use it to explain the revealing of something meaningful.
Dictionaries both give 青白い and 淡い the same translation, but they don't explain that 淡い is a poetic term and has a very different feel and evocation of images than 青白い.
A lot of foreigners I've noticed have trouble with the differences between 空く (すく) and 空く (あく) since they have the same meaning and are written identical. Not many foreigners would know, if they're reading aloud, which they should pronounce 空く is, I imagine, as you'd practically have to live in Japan to know all the situations those two words are used in.

I wish there were dictionaries that explained all the subtleties to a word and all it's uses and associations rather than just it's translation. But, I don't think there are any that in-depth yet, so nothing can substitute for a good decade of exposure to the culture

Exactly, and that is why learning readings separately from words is useless. It is best to learn words as one in the context they are used. And you are a Japanese person? why are you on these forums! ahha

Nukemarine Wrote:Well, out of about 700 sentences from Zettai Kareshi ep 01, I used 300 of them. In those there were 10 new kanji (2500 and above in RTK 3) and 110 new words (iKnow Core 6000 words). On top of that were about 20 words not in the Core list. So by having a base line of RTK and Core 2000 vocabulary I'm looking up only a few new words.

Nukemarine, I don't want to hijack the thread, but is there a deck or list of those sentences I could get a copy of？ 実は大ファンです。

Aijin Wrote:I wish there were dictionaries that explained all the subtleties to a word and all it's uses and associations rather than just it's translation.

There is no need for such dictionaries. In my language, we usually use one word for the words "see", "watch", "look", "view" but when I read English in context I understand the subtle meaning of each word. No dictionary can explain this without examples, and examples is all what you need. Google is the best source for examples.

BTW, did you understand Heisig's idea of learning the kanji seperately?

yukkuri_kame Wrote:is there a deck or list of those sentences I could get a copy of？

1. Go to d-addicts.com
2. Download Zettai Kareshi ep1 + English subtitles + Japanese subtitle.
3. Use sub2srs to generate the deck.

Quote:Exactly, and that is why learning readings separately from words is useless. It is best to learn words as one in the context they are used. And you are a Japanese person? why are you on these forums! ahha

It's possible to get by without readings, by just memorizing words as complete units, but it'd make life a lot, lot harder. If you don't know readings you won't be able to use a real Japanese dictionary, read personal/place names, new words without looking them up, etc (I know you're not advocating not learning the readings at all, I am just stressing the importance of them).
I think words should be learned along with readings. For example, if you're learning a simple character like 食 then 食う, 食べる and 食欲 should be learned with it, I think Then if you see a word like 飲食 or 食事 or 悪食 the first time you can read it right off the bat without looking it up.

Oh, I'm here to learn more about the different ways that people learn Japanese as a foreign language. My goal is to be a kick-ass teacher of Japanese, and I can't do that unless I know what ways will best help my students learn and retain

Quote:There is no need for such dictionaries. In my language, we usually use one word for the words "see", "watch", "look", "view" but when I read English in context I understand the subtle meaning of each word. No dictionary can explain this without examples, and examples is all what you need. Google is the best source for examples.

I don't think only examples works for the more complex situations. I had a pretty clear understanding of English vocabulary before I moved to the United States, but when I was actually immersed in the culture 24/7 and heard words in thousands of situations, contexts, etc, the depth of understanding of those words really unfurled tenfold.

Aijin Wrote:I wish there were dictionaries that explained all the subtleties to a word and all it's uses and associations rather than just it's translation. But, I don't think there are any that in-depth yet, so nothing can substitute for a good decade of exposure to the culture

Ha ha. It's like asking a perfect private detective equipped with future gadgets who can give you all sort of info about your bride/bridegroom. I'm not saying it's impossible because it's a moving target or too complicated. People in here are in love with the language. It's just their sweethearts can't be described by words.

Different people speak differently. Learning language is acquiring your own language. My Japanese can't be described by linguists. My English is unique and that's why I love it.

Are you learning something others speak? That's boring if you ask me. I'm learning something I speak.

Language is beyond words, I think.

I agree. But hey, sometimes a little private investigating pays off.

Language is beyond words, and that's why I love it. Like any art-form it's a unique, personal experience and can't be truly felt by anyone but yourself.

Oh no I still think one should learn readings, but not utterly separate from everything else.

Aijin, out of interest, has your question been answered as to why many of us learn Kanji via our native language first?

On topic, I think Magamo hit in on the head. Knowing tons of words will not help you when it comes to how natives put those words to use. If all you know are definitions (or just one sample sentence) then you'll likely use the words as you would in your native language. At that point, natives can tell pretty much what country or area you are from. Ergo, drown yourself in native material.

On the same note, you can't understand native sentences if you don't understand the words or grammar. And it's hard to understand some of the words or grammar if you don't understand kanji. To get out of this part quick (since it's not natural) we like to be systematic. But at a point, you can't reasonably cover every eventuality.

Course, then comes the part of how to best cover native material. That's multiple threads in of of itself.

Aijin Wrote:It's possible to get by without readings, by just memorizing words as complete units, but it'd make life a lot, lot harder. If you don't know readings you won't be able to use a real Japanese dictionary, read personal/place names, new words without looking them up, etc (I know you're not advocating not learning the readings at all, I am just stressing the importance of them).
I think words should be learned along with readings. For example, if you're learning a simple character like 食 then 食う, 食べる and 食欲 should be learned with it, I think Then if you see a word like 飲食 or 食事 or 悪食 the first time you can read it right off the bat without looking it up.

What's fortunate is that readings are automatically learned with words. You can't (as in, it's impossible to) actually spend your time 'only learning words' and 'not learning readings,' because every time you learn a word you've learned a reading of the kanji. You'd have to be pretty dense to not understand, after learning the word 食事, that ショク is one of 食's readings. The point of learning this way is that it turns what's usually thought of as a two-step process into a single step.

Aijin Wrote:I wish there were dictionaries that explained all the subtleties to a word and all it's uses and associations rather than just it's translation. But, I don't think there are any that in-depth yet, so nothing can substitute for a good decade of exposure to the culture

Magamo Wrote:They're useful, but I don't think dictionaries would deepen your knowledge as much as exposure does. After all, reading a dictionary is part of your reading in the target language unless you're using a bilingual dictionary.

Yeah, I wasn't serious about actually relying on a dictionary for much. I still think exposure is the only way to pick up the idioms and actual usage of words. I was just curious if maybe people actually had ideas for short-cutting exposure, but I highly doubt there's any way to.

We don't really need frequency lists. Even if we just choose random words from texts instead, we'll end up learning the most frequent words first simply due the laws of probability.

Also, we're using the text that we actually want to read as the source of vocabulary. So we're going to learn words that are much more likely to be used again in things that we are interested in.

Words that we are ready to learn with very little effort (for example because we know their kanji well enough) will naturally be more noticeable. It's efficient to learn them even if they are not high frequency.

Important words are not always the same thing as frequent words. When reading a sentence, we are able to focus on the words that will allow us to make sense of the sentence and ignore some of the "glue" around them.

In the end, I think that getting vocabulary from the material you're reading is a more efficient way to improve your comprehension fast. Frequency lists can still be useful when you lack inspiration or if they come with ready to use sample sentences. But fundamentally they are not the best way to learn.